conference proceedingstweb.cjcu.edu.tw/conference/2015_07_03_10_32_45.579.pdfadita evalina fitria...
TRANSCRIPT
Conference Proceedings
June 27-29, 2015 Seoul, South Korea
SICASE
Seoul International Conference on Applied Science and Engineering
SICBENS
Seoul International Conference on Biological Engineering & Natural Science
SICASE
Seoul International Conference on Applied Science and Engineering
ISBN 978-986-89536-5-9
SICBENS
Seoul International Conference on Biological Engineering & Natural Science
ISBN 978-986-89536-3-5
1
Content
General Information ................................................................................................................ 3
Conference Organization ........................................................................................................ 5
SICASE International Committee Board ........................................................................................... 5
SICBENS International Committee Board ........................................................................................ 6
Special Thanks to Session Chairs ........................................................................................... 7
Conference Schedule ................................................................................................................ 8
Natural Sciences Keynote Speech ......................................................................................... 10
Tutorial Course ...................................................................................................................... 12
Oral Sessions –June 27th
, 2015 .............................................................................................. 13
Civil Engineering ................................................................................................................................ 13
SICASE-735 ............................................................................................................................... 14
SICASE-748 ............................................................................................................................... 23
SICASE-777 ............................................................................................................................... 24
SICASE-828 ............................................................................................................................... 33
Computer & Information Science I / Electrical and Electronic Engineering ............................... 42
SICASE-727 ............................................................................................................................... 43
SICASE-388 ............................................................................................................................... 63
SICASE-776 ............................................................................................................................... 71
SICASE-818 ............................................................................................................................... 80
Biological Engineering / Biomedical Engineering / Fundamental Science .................................... 88
SICASE-701 ............................................................................................................................... 90
SICASE-729 ............................................................................................................................... 95
SICBENS-1292 .......................................................................................................................... 97
SICBENS-1300 ........................................................................................................................ 106
SICASE-732 ............................................................................................................................. 107
SICASE-829 ............................................................................................................................. 108
Oral Sessions –June 28th
, 2015 ............................................................................................ 129
Material Science & Engineering ...................................................................................................... 129
SICASE-771 ............................................................................................................................. 131
SICASE-772 ............................................................................................................................. 140
SICASE-734 ............................................................................................................................. 151
SICASE-740 ............................................................................................................................. 164
SICASE-753 ............................................................................................................................. 175
SICASE-755 ............................................................................................................................. 183
Environmental Sciences .................................................................................................................... 192
SICASE-708 ............................................................................................................................. 194
2
SICASE-805 ............................................................................................................................. 203
SICASE-843 ............................................................................................................................. 204
SICASE-840 ............................................................................................................................. 211
SICBENS-1326 ........................................................................................................................ 221
Computer & Information Science II ............................................................................................... 234
SICASE-723 ............................................................................................................................. 235
SICASE-703 ............................................................................................................................. 243
SICASE-759 ............................................................................................................................. 252
Mechanical Engineering ................................................................................................................... 272
SICASE-704 ............................................................................................................................. 273
SICASE-722 ............................................................................................................................. 279
SICASE-820 ............................................................................................................................. 286
SICASE-819 ............................................................................................................................. 294
Poster Presentation .............................................................................................................. 302
Poster Session (1) ............................................................................................................................... 302
Biomedical Engineering / Biological Engineering / Natural Science .................................. 302
SICASE-745 ............................................................................................................................. 305
SICASE-797 ............................................................................................................................. 309
SICBENS-1287 ........................................................................................................................ 318
SICBENS-1306 ........................................................................................................................ 320
SICBENS-1330 ........................................................................................................................ 322
SICBENS-1334 ........................................................................................................................ 323
SICBENS-1333 ........................................................................................................................ 325
SICBENS-1336 ........................................................................................................................ 327
SICBENS-1303 ........................................................................................................................ 329
Poster Session (2) ............................................................................................................................... 331
Civil Engineering / Electrical and Electronic Engineering / Environmental Sciences
/Fundamental & Applied Sciences / Material Science and Engineering ............................ 331
SICASE-798 ............................................................................................................................. 333
SICASE-803 ............................................................................................................................. 343
SICASE-796 ............................................................................................................................. 353
SICASE-779 ............................................................................................................................. 361
SICASE-765 ............................................................................................................................. 362
SICASE-823 ............................................................................................................................. 366
SICASE-743 ............................................................................................................................. 380
3
General Information
Venue
Courtyard by Marriott Seoul Times Square
442 Yeongdeungpo-dong 4-ga, Yeongdeungpo-gu, Seoul, 150-798 South Korea
Phone: +82 2 2638 3000
Fax: +82 2 2638 3001
Registration Time
4th
Floor, Courtyard by Marriott Seoul Times Square
Saturday, June 27, 2015 08:30-17:00
Sunday, June 28, 2015 08:30-15:00
Poster Area
Hallway, 4th Floor, Courtyard by Marriott Seoul Times Square
Mounting Presentation Removal
Poster Session (1)
Saturday, June 27, 2015 13:20-13:30 13:30-14:30 14:30-14:40
Poster Session (2)
Saturday, June 27, 2015 14:50-15:00 15:00-16:00 16:00-16:10
Poster Session (3)
Saturday, June 27, 2015 16:20-16:30 16:30-17:30 17:30-17:40
Floor Map
4
Oral Instructions
- Total length of each session is 90 minutes. The specific length of each presentation
depends on the number of presentations in that session.
- All presentation rooms are equipped with a screen, an LCD projector, and a laptop
computer installed with PowerPoint software. You will be able to insert your USB flash
drive into the computer and double click on your presentation to open it in PowerPoint.
We recommend that you bring two copies of your presentation in case of one fails. You
may also link your own laptop computer to the projector cable, however if you use your
own Mac please ensure you have the requisite connector.
Poster Instructions
- Materials Provided by the Conference Organizer:
1. X-frame display & Base Fabric Canvases (60cm×160cm)
2. Adhesive Tapes or Clamps
- Materials Prepared by the Presenters:
1. Home-made Poster(s)
- Requirement for the Posters:
1. Material: not limited, can be posted on the canvases
2. Size: smaller than 60cm*160cm
A Polite Request to All Participants
Participants are requested to arrive in a timely fashion for all addresses, whether to their own, or to
those of other presenters. Presenters are reminded that the time slots should be divided fairly and
equally between the number of presentations, and that they should not overrun. The session chair is
asked to assume this timekeeping role.
Organizer
Higher Education Forum (HEF)
Tel: + 886 2 2740 1498〡www.prohef.org
5
Conference Organization
SICASE International Committee Board
Khalid M. Mosalam, University of California, Berkeley
Chueerat Jaruskulchai, Kasetsart University
M. Cheralathan, SRM University
J N Bandyopadhyay, Indian Institute of Technology Kharagpur
S. Dhar, University of Calcutta
Poongothai Shankar, Annamalai University
Amit Agrawal, Indian Institute of Technology Bombay
Cheng Li, The Hong Kong Polytechnic University
T.M. Indra Mahlia, University of Malaya
Kunal Ghosh, Indian Institute of Technology Kanpur
Narayanan Kulathuramaiyer, University of Sarawak Malaysia
Arup K. Sarma, Indian Institute of Technology Guwahati
Suresh K Bhargava, School of Applied Sciences
Rapeepan Pitakaso, Ubonratchthani University
Jie Liu, Carleton University
Carlos Alejandro Figueroa, Plasmar Tecnologia
Poul Vaeggemose, VIA University College Denmark
Maha M. O. Khayyat, Umm Al-Qura University
Kant Kanyarusoke, Cape Peninsula University of Technology
Paramita Bhattacharjee, Jadavpur University
Nathalia Devina Widjaja, Binus International University
C. M. Khalique, North-West University,South Africa
6
SICBENS International Committee Board
C H OH, National University of Singapore
Lim Koon Ong Pk, Universiti Sains Malaysia
Parin Chaivisuthangkura, Srinkharinwirot University
Zulkifli Hj. Shamsuddin, Universiti Putra Malaysia
Avinash C. Sharma, GGS Indraprastha University
E. Shaji, University of Kerala
A. M. Chandra, University of Calcutta
Chien-Cheng Chang, National Taiwan University
Stephen Redmond, The University of New South Wales
J. Gao, The University of Hong Kong
P.K.Barhai, Birla Institute of Technology
Singh, Gauri Shankar, Indian Institute of Technology Roorkee,
Mohd. Nasir Mohd. Desa, Universiti Putra Malaysia
Aliakbar Roodbari, Shahroud University of Medical Sciences
Olivia J Fernando, Annamalai University
K. Padmakumar, Kerala University Campus
John Hearne, Royal Melbourne Institute of Technology
Hamed M. El-Shora, Mansoura University
Huabei Jiang, The University of Florida
7
Special Thanks to Session Chairs
Fumihiko Isada Kansai University
Ya-Fen Lee Toko University
Chua Bee Seok Universiti Malaysia Sabah
Yonglei Tao Grand Valley State University
Moon J. Lee University of Florida
Tatiana Vladimirovna Kolpakova Transbaikal State University
Vinoth Jagaroo Boston U. School of Medicine & Emerson College
Hwee Ling Lim The Petroleum Institute
Adita Evalina Fitria Universitas Indonesia
Minnie M. Liangco Pampanga State Agricultural University
Prapat Pentamwa Suranaree University of Technology
Gigi Lam Tung Wah College
Seonho Choi Bowie State University
Pratuang Dinnaratna Kasetsart University
Yanhua Wu Nanyang Technolocial University
8
Conference Schedule
Saturday, June 27, 2015
Oral Presentation
Time Information
08:30-17:00 Registration
09:00-10:30 Civil Engineering Room 2
10:30-10:45 Tea Break
11:00-12:15 Natural Sciences Keynote Speech (SBPA) Room 1
12:00-13:00 Lunch Time 5F
13:00-14:30 Computer & Information Science I / Electrical and Electronic
Engineering
Room 2
14:30-14:45 Tea Break
14:45-16:15 Tutorial Room 1
16:15-16:30 Tea Break
16:30-18:00
Economics Room 1
Biological Engineering / Biomedical Engineering /
Fundamental Science
Room 2
9
Saturday, June 27, 2015
Poster Session ,4F Pre-Function Area
Time Information
13:30-14:30 Poster Session (1)
Biomedical Engineering / Biological Engineering / Natural Science
15:00-16:00
Poster Session (2)
Civil Engineering / Electrical and Electronic Engineering / Environmental
Sciences / Fundamental & Applied Sciences / Material Science and
Engineering
Sunday, June 28, 2015
Oral Presentation
Time Information
08:30-15:00 Registration
09:00-10:30 Material Science & Engineering Room 2
10:30-10:45 Tea Break
10:45-12:15 Environmental Sciences Room 2
12:00-13:00 Lunch Time 5F
13:00-14:30 Computer & Information Science II Room 2
14:30-14:45 Tea Break
14:45-16:15 Mechanical Engineering Room 2
10
Natural Sciences Keynote Speech
Room 1
2015/06/27 Saturday 11:00-12:15
Topic:Probabilistic Analysis of Volcanic Ash Dispersion over Korea
Keynote Speech:Lee, Sungsu, Ph.D
Professor
School of Civil Engineering, Chungbuk National University, 1
Chungbukdae-gil, Cheongju, Rep. of Korea
In 2010, a series of volcanic eruption of Mt. Eyjafjallajokull in Iceland put a hold on the air
travel in the continent of Europe and affects all over the world. Besides this direct impact,
numerous indirect effects were found from the aftermath. In January of 2011, another series of
eruption from Mt. Shinmoedake, Kyushu, Japan also alerted Japan and the world. With these
series of events, the reports from Chinese seismologists since 2002 have drawn an attention and
initiated discussions on the possibility of the eruption of Mt. Baekdu which is located on the
border of North Korea and Machuria, China. It was found that Mt. Baekdu has history of
eruptions more than 20 times since 10th
century and its strongest one in 10th
century was
believed to be about 1000 times that of Mt. Eyjafjallajokull.
Alerted from the series of events and the rediscovered facts about Mt. Baekdu, Korea has
thought of potential threats from volcanos, that is, not only Mt. Baekdu but also volcanos in the
vicinity of Korea, knowing that there are more than 10 Chinese volcanos within 1000km from
Seoul and tens of Japanese ones. If any of those volcanos erupts, the most effects on Korean
peninsula will come from ash dispersion and this study focuses on the probabilistic analysis of
volcanic ash dispersion both in the air and on the land.
Thousands of hypothetical eruptions from Mt. Baekdu are assumed and the ash from the vent is
simulated by a well-known numerical model employing particle tracking method based on
lagrangian scheme. It is assumed that each eruption starts at midnight and lasts for 24 hours on
every day from Jan. 1st , 2008 to Dec. 31
st, 2013. The meteorological conditions including wind
and pressure on each day are borrowed from reanalysis data of NCEP, which represents the
real weather conditions. With these enormous amount of computed outputs, the probabilistic
estimation of airborne and fallout ashes has been carried out over Korea and adjacent regions in
11
North-East Asia. The results show that once it happens, the ash can sweep the part of Korean
peninsula and travel across the East Sea to Japan and to the Pacific.
Therefore, it is very essential to establish the international collaboration and the cooperation in
the research of volcanic risk in North-East Asia.
This work was supported by the grant from Natural Hazard Mitigation Research Group funded
by Ministry of Public Safety and Security, South Korea.
Educational Background
B.S. : Seoul National University, Dept of Navar Architecture, Rep. of Korea (1986)
M.S. : Seoul National University, Dept of Navar Architecture, Rep. of Korea (1988)
Ph.D : Colorado State University, Ft. Collins, CO, USA, Dept of Civil Engineering
(1997)
- Dissertation : Large Eddy Simulation of Flow past a Square Cylinder using Finite
Element Method
- Adviser : Prof. Bogusz Bienkiewicz
Work Experience
Researcher, Agency for Defense Development, Rep. of Korea (1988-1993)
- Major Experience : Computational Fluid Dynamics for Rotating Machinery,
Design of Propellers
Research Assistant, Dept of Civil Engineering , Colorado State University
(1994-1997)
Research Associate, National Oceanic and Atmospheric Administration
(NOAA)/CIRES, Boulder, CO, USA (1998-1999)
- Major Experience: LES for Stratified Flow around Bodies
Research Professor, Dept of Mechanical Engineering, Inha University, Rep. of Korea
(2000)
Professor, School of Civil Engineering, Chungbuk National University, Rep. of Korea
(2000-Present)
Director of Volcanic Disaster Preparedness Research Center (2012-).
Field of Interests
Computational Fluid Dynamics and its application to Turbulent Wind Flow
Large Eddy Simulation of Turbulent Flow around Buildings and Moving Bridges
Typhoon Simulation using Statistical/Stochastic Approaches
Structural Vulnerability to Wind
Wind Damage Analysis and Mitigation
Micro-Meteorology for Urban Region
12
Tutorial Course
Room 1
2015/06/27 Saturday 14:45-16:15
Speaker: Sounghwa Walker, M.Ed.
Topic: Engaging the audience through empowering
nonverbal communication
Background of the speaker
Entered the teaching profession in 1994; CA teaching
credential in Math
Taught English (Korea) & Mathematics (Southern California, USA)
Taught various education courses (University of California, Riverside, USA)
Presented papers at numerous international conferences
Conducted a workshop (45 min.) at the 2015 national conference on Future Educators
Association, Washington, DC
Educational Psychology, PhD in progress
(Faculty Advisor: Dr. Keith Widaman, Distinguished Professor)
Instructor: Graduate School of Education at UC Riverside, USA
1. Introduction (10 min.): Purpose of the presentation (Knowing who we are)
2. Nonverbal Communication - NVC (35 min.)
a) What is it & why should we care?
b) Relationships between NV communication & presentation
c) Group activities & volunteers
3. Improving Nonverbal Communication Skills (35 min.)
a) Practical application of NVC skills
b) Engaging the audience
c) Empowering NVC
4. Closing (10 min.)
13
Oral Sessions –June 27th
, 2015
Civil Engineering
Session Chair: Ya-Fen Lee
SICASE-735
A New Conceptual Sustainable Water Resources Research
Choo, Tai Ho︱Pusan National University
Yoon, Hyeon Cheol︱Pusan National University
Yun, Gwan Seon︱Pusan National University
Kwak, Kil Sin︱Pusan National University
SICASE-748
Measuring Financial Liquidity of Construction Contractors in Hong Kong
Tsang, Yick Tat︱City University of Hong Kong
SICASE-777
Evaluating the Rescue Performance of Flooding Disaster Based on Residents’ Perception
Ya-Fen Lee︱Toko University
Yun-Yao Chi︱Chang Jung Christian University
Ching-Hui Peng︱Chang Jung Christian University
SICASE-828
Fatique Resistance of Warm Mix Asphalt Using Bayat Natural Zeolite and Elastomeric
Polymer Bitumen
Ani Tjitra Handayani︱College of Technology STTNAS, Indonesia
Bagus Hario Setiaji︱Diponegoro University
Ani Prabandiyani︱Diponegoro University
2015/6/27 Saturday 09:00-10:30 Room 2
14
SICASE-735
A New Conceptual Sustainable Water Resources Research
Choo, Tai Ho
Department of Civil and Environmental Engineering, Pusan National University,
South Korea
Yoon, Hyeon Cheol*
Department of Civil and Environmental Engineering, Pusan National University,
South Korea
Yun, Gwan Seon
Department of Civil and Environmental Engineering, Pusan National University,
South Korea
Kwak, Kil Sin
Department of Civil and Environmental Engineering, Pusan National University,
South Korea
Abstract
A paradigm about the water resources management is changing the operation management to
the nation water security. At this stage, the sustainable water resources are a matter of great
interest. There are many methods to secure the water resources. Among others, constructing a
dam is direct and active. In the Korea, however, the dam construction is more and more
difficult because of the decrease of a suitable location, the opposition of community Residents
and environment group and negative public opinion. Therefore, a new conceptual Multipurpose
Regulating Dam is suggested in the present paper. The Multipurpose Regulating Dam can
provide securing water resources, flood control, hydroelectric power, eco-environmental
instream flow and recreation as well as the use of function of existing dam. The Multipurpose
Regulating Dam saves the construction and compensation costs due to the use of conventional
river channel, and has an advantage economically. Also, it will improve negative eyes from the
local residents. The research sequence introduces basic concepts for the Multipurpose
Regulating Dam construction, and the propriety Evaluation, the decision of dam type and the
verification of operation rule are approached sequentially. The verification is applied to Yeong
Cheon Dam(water supply only) in the Korea.
15
The appropriate operation scenario is selected after establishing the various scenarios by
HEC-Ressim modeling. The results show securing water resources of 14 million ton and the
flood control of 15.4 million ton in comparison with solo operation of existing dam. If the
current research is advanced and continuous, it will be a good alternative as new water
management method.
Keywords: eco-environmental instream flow, flood control, hydroelectric power, Multipurpose
Regulating Dam, water resources
1. Introduction
To cope with and adapt to various water-related disasters in response to the crisis according to
climate change is a common concern of all countries (UNWWDR 4, 2002). In last June 2013,
Ganges River and its branch were flooded due to localized heavy rains in the Hindustan
(northern area of India), and more than 5,000 people were killed or disappeared caused by the
occurrence of large-scale landslide. Also, it was estimated that in November of that year, while
the Philippines were hit hard by a super typhoon „Haiyan‟, approximately 8,000 casualties were
occurred, whereas, since a drought was deepened due to a record-heat wave in China, over 600
million people in 13 castles such as Hunan suffered from a drinking water shortage, and the
economic loss of about 940 billion won, such as the crops damage was occurred. Thus, if water
is excessive or deficient, it gives much damage to human. However, because the available
water resources have the restriction of time and space, the preserved amount of water is limited.
Therefore, the establishment of the response measures in accordance with a climate change,
leading to the increase of the intensity and incidence frequency of flooding and drought, at the
same time the that distribution of the sufficient water is conducted by maximally utilizing the
available water resources is essential. In fact, the most direct and positive method is to secure
the water resources by means of the structures such as a dam, a dammed pool for irrigation,
and Estuary Bank. In order to compensate the temporal and spatial imbalance of water
resources, most of methods are compensating the water shortage of a period of water shortage
by adjust the peak discharge and by storing the surplus water resources, using a dam in a flood
season (Chung et.Al, 2010).
But in the Korea these methods require the long period from planning to construction, and need
the selection of the right place for construction, and suffer from the opposite of many
environmental groups and the local residents owing to the large-scale civil engineering projects.
In addition, if exceeding the marginal capacity of the structure, there is also the case that the
utility becomes a dead dog. Even though this drawback, many countries around the world are
securing still the water resources through the methods by the various structures, such as dams
and barrage, and that is to say, it indirectly shows the point that these methods are the most
effective than other alternatives.
16
But the reality is faced with a very difficult situation to build a large dam. First, it is the
tendency that the suitable places to build a dam are gradually decreased, and because the
negative perception of local residents or environmental groups, and the public opinion against
the large structures (particularly, dam) for water resources are very strong, and the problem
losing hometown, and the cultural property issues along with astronomical compensation
amount should be solved, it is the difficult situation to promote the business. Therefore, to
develop the water resources projects, the improvement direction should be sought based on the
current situation diagnosis.
Therefore, the current study proposed a new conceptual “Blue Reservoir”. Accordingly, first,
the authors have established the definition of the Blue reservoir, and they have suggested the
matters, operation and the application method of model that shall be considered to install the
Blue reservoir, and they have also presented the results applying it to the actual river basin
through a case study.
2. The Definition of Blue Reservoir
The Blue reservoir is basically installed in online at the channel in the direct downstream of the
dam having various purposes such as a multipurpose dam, water supply dam, and a
hydroelectric dam. In other words, it has the similar form to the existing balancing reservoir
dam. But there are many differences between the Blue reservoir and the existing balancing
reservoir. The existing balancing reservoir played a role that evenly discharges it to the
downstream over 24 hours after storing the water discharging through the Peak-load power at
the multipurpose dam and hydroelectric dam. Whereas a Blue dam provides the water bowl for
the security of the water resources in the watershed along with flood control effect by utilizing
the discharge flow of the flood period meaninglessly flowing to downstream. Also the Blue
dam can uniformly discharge the flow rate for ecological environment preservation to the
downstream beyond the concept of the flow rate for maintaining a river, and it can help for the
restore integrity of a river and the safety security of ecosystem by preventing the cold-weather
damage.
And it has many right functions that the demand and supply through self-electrical generation
is possible and can use also the stored water for the local recreation, and it has the advantage
that can establish its own operation rules by reflecting the operation rules of the existing dam.
Above all, the Blue dam has the advantage that can minimize the environmental damage,
because the construction costs become cheaper and the construction period can be shortened by
utilizing the existing channel as it is. Wolff and Gleick(2002) have suggested that will provide
potentially very positive opportunity if the harmony with nature is formed by using the dam
that the environment is not almost damaged. Furthermore, it is expected that the Blue dam can
present the solutions for large water structure with the opposite of the environmental groups
and the local residents, and the social negative perception due to the loss of hometown and
17
compensation issues, and the cultural property issues, and it is considered that it will be the
good alternative in the current situation that the right place of a dam is gradually decreased.
3. Research Area
In this study, the Blue reservoir was installed in the downstream of Yeongcheon dam which is
a water supply dam located at the riverside of Geumhogang River in Gyeongsangbuk-do as
shown in Fig. 1.
The Yeongchen dam is located at Jayang-myeon in Yeongcheon-si, Gyeongsangbuk-do, the
height of the dam is 42m, the length is 300m, and total storage capacity is 96.4 million ,
This dam was constructed for the purpose of industrial and domestic water supply, and the
irrigation water and river maintenance water supply for the downstream area of the dam. In
addition, unusually the Yeongcheon dam is using the water supplied from Imha dam located in
the other river basin, and supply water through a raceway to other region, and is producing the
electricity of 6GWh at the small hydraulic plant with the station capacity of 1 MW.
4. The Research Process
The installation technique of Blue reservoir proposed in this study consists of the validation
steps such as the evaluation of installation possibility, the capacity and operation decision, and
its flow chart is shown in Fig. 2.
18
Whether the installation is possible should be determined by conducting the evaluation for the
security of water quantity if the Blue reservoir will be installed at the downstream of the dam.
In other words, it is evaluated by calculating the water quantity discharged in excess in flood
seasons.
The position of the Blue reservoir must determine by considering the aerial photograph and the
land utilization, and the capacity of the reservoir is decided by quantitatively presenting
through the hydrologic statistic techniques targeting the morphological analysis through GIS
(geographic information system) and the flow rate directly discharged from the original dam
into the river.
The storage capacity of the Blue reservoir and the operation rules are finally determined by
reviewing whether the improvement of the river flow condition in the Blue reservoir and the
increase of additional water quantity for electricity generation are possible, while maintaining,
as it is, the water supply and the electricity discharge of the original dam through a simulation.
5. Installation Guide of Blue Reservoir
As a result comparing the annual water supply and the operating performance data, it was
analyzed that the security of water of 133M per year is possible as shown in Fig. 3.
19
The annual maximum discharge per the duration among the discharge flow discharged from the
original dam into the river is consisted as the input data, and the probability discharge per the
frequency-duration was calculated through the process such as the basic statistics, preliminary
analysis, parameter estimation and the goodness of fit test. In this study, in order to calculate
the optimum storage capacity by considering the amount of dredging, the water level-storage
capacity curves of 30.4, 34.8, and 39.8 M corresponding to the duration of 7 days for
100-year frequency, the duration of 6 days for 200-year frequency, respectively, and the
duration of 10 days for 200-year frequency were calculated.
The Blue reservoir‟s effects are analyzed by using HEC-ressim model. The water level (Daily)
of the reservoir among the operation performance data of the original dam was designated as a
calibration target, and its period is from 2001 to 2013. The spill occurred in the dam basin and
the inflow (Daily) through the raceway are the upstream boundary conditions, and the water
supply and the electricity discharge are the downstream boundary conditions, and the water
level of the original dam was calibrated by the overtropping discharge through the sluice, that
is, a parameter. The electricity discharge and the water supply were built as input data, and the
model was calibrated with the discharge flow by the gate operation as the parameter. The
observations of the original dam and the simulated hydrological curve are the same as Fig. 4.
As a result conducting the linear regression analysis using this result, the coefficient of
determination was analyzed as 0.9977 (Fig. 5).
20
The Blue reservoir is focused on whether the original dam can additionally supply the water for
river maintenance and the water quantity for electricity generation by securing the excessive
flow rate of the downstream while maintaining the operation of the dam, as it is. In order to
determine the most ideal capacity and the additional discharge, 12 scenarios were composed as
shown in Table 1.
The flow-duration variability was reviewed by year in order to determine the optimal volume
and operation of the Blue reservoir, and low flow (Q533) of scenario 9 that added 10% was
determined since the scenario can provide the additional water supply in comparison with other
scenario also in 2008, 2009 and 2010 corresponding to the drought year. To compare with the
installation effect of Blue reservoir, the result of unilateral operation of Yeong Cheon Dam and
the scenario 9 simulation result are summarized as shown in the table 2. The flow-duration in
the river is improved because Existing water supply maintains at the same time and low
flow(Q355)of downstream in the river increased about 8%.
21
In addition, the peak discharge flow corresponding to the flood control benefit was reduced as
much as about 8.5/sec as 13-year average, and it was analyzed that the electricity generation
discharge corresponding the power generation benefit increased as much as 92.5 M in
comparison with the unilateral operation. This is because the power discharge is achieved
through small hydro power plant when the water for agricultural use and the water for river
maintenance will be discharged, and the Blue reservoir also is the same. Finally, the 13-year
average water quantity of about 11.1 M was additionally secured.
6. Conclusion
In this study, the authors have proposed the Blue reservoir installation technique of new
concept in order to secure the water resources, and the simulation of the YeongCheon dam
established in HEC-ResSim was calibrated, and the effect of the proposed technique was
analyzed by applying to a model.
The scenario 9 can supply the additional water (water for agricultural use + water for
maintaining) of 12-year average 6.85M into the downstream river, at the same time, the
water quantity for the hydroelectric power generation was increased by using of 49.21M, and
it was simulated that the water of the average 9.87 M can be secured in addition over the
entire period. It is considered that the Blue reservoir proposed in this study will contribute to
the vitalization of the local economy by better river water supply and the algae blooms
occurrence reduction, the flood damage reduction of the local residents in downstream, and the
offer of recreation using the secured reservoir by means of the contribution to the increase in
the water bowl, the increase of the water quantity for hydroelectric power generation which is
the clean energy, and the supply of the additional water in the water that has flown into the
existing river.
22
7. Acknowledgement
This research was supported by a grant (11-TI-C06) from Advanced Water Management
Research Program funded by Ministry of Land, Infrastructure and Transport of Korean
government.
8. Reference
[1] Chung, G.H., Ryu, G.H. and Kim, J.H.(2010) Optimization of Multi-reservoir operation
considering water demand uncertainty in the Han river basin. Journal of the Korean
Society of Hazard Mitigation, KOSHAM, 10(1), 89-102.
[2] UNISDR(United Nations International Strategy for Disaster Reduction Secretariat). 2011.
Global Assessment Report on Disaster Risk Reduction. Geneva, UNISDR.
[3] Korea Water Resources Corporation (2013) Dam Operation Manual.
[4] Wolff, G. and Gleick, P. H.2002. The soft path for water. P.H. Gleick(ed.) The Worldd‟s
Water : The Biennial Report on Freshwater Resources, 2002-2003. Washington DC, Island
Press, pp. 1-32.
23
SICASE-748
Measuring Financial Liquidity of Construction Contractors in Hong Kong
Tsang, Yick Tat
Department of Civil and Architectural Engineering,
City University of Hong Kong
Abstract
Construction business is sensitive to the fluctuation of its socio-economic environment. The
financial liquidity is one of the key concerns in maintaining the survival and competitiveness of
a construction contractor due to the highly capitalized and risky nature of construction industry.
This research investigated the financial liquidity of construction contractors in Hong Kong
based on the published financial data of listed construction companies between 1992 and 2010.
The liquidity was empirically modelled by means of factor analysis based on a number of
financial ratios. The time trend of the liquidity was further analysed by correlating with a pool
of identified driving determinants such as macroeconomic performance, construction market
demand, tender price, construction productivity, construction cost, etc., by means of bivariate
cross-correction analysis. The findings served the basis of understanding and predicting the
business well-being of construction contractors in reacting to the contextual change and
provided an insight to the corporate managements, clients and creditors in monitoring the
contractors‟ financial health.
Keywords: Construction contractor, liquidity, factor analysis, cross-correlation analysis
24
SICASE-777
Evaluating the Rescue Performance of Flooding Disaster Based on Residents’
Perception
Ya-Fen Lee
Department of Leisure and Recreation Management, Toko University, Taiwan
Yun-Yao Chi, Ching-Hui Peng
Department of Land Management and Development, Chang Jung Christian University, Taiwan
Abstract
Under the global climate change, costal urban in Taiwan will face the threat of flooding in the
future, whose nature and condition differs from the fire. Thus, flooding rescue is a major issue
in the disaster prevention and protection. A sample of residents in the high-potential area of
flooding in Tainan, Taiwan provided information about their perception in influence factors of
flooding rescue performance. In this study, a model to evaluate the rescue performance of
flooding disaster based on residents‟ perception was proposed and the illustration of rescue
performance was presented as well.
Key words: global climate change, flooding, rescue performance, structural equation modeling
1. Introduction
Due to the influence of the global climate change in recent years, the flooding victims and
property loss are increasing constantly with the increasing occurrence frequencies of flooding
all around the world. Taiwan locates on the Pacific climate zones and the typhoon-induced
heavy rainfall occurs every year. In recent years, the serious flooding disaster often happens,
especially in the coastal lowlands in Taiwan. For example, the torrential rainfall induced by
Typhoon Morakot in August, 2009, initiated a flooding disaster and submerged a lot of coastal
cities in southern Taiwan, which caused many flooding victims trapped to wait rescued. The
flooding rescue experiences during typhoon Morakot are unforgettable for firefighters in
Taiwan and the related issues of rescue performance of flooding disaster begin to be concerned
deeply.
The Fire Bureau, Taiwan owns many considerable experiences of fire rescue tasks. However,
the rescue nature and condition of flooding disaster differs from that of fire disaster. It is not
familiar with flooding rescue tasks for firefighters. When the serious flooding attacks coastal
lowlands, Taiwan in the future, the amount and distribution of rescued victims will be over
than fire event, which will be a large Challenget in flooding disaster prevention and protection
25
for firefighters.
Therefore, a practical performance evaluation is necessary to judge the rescue performance of
flooding disaster.
Edwards(1991) pointed that the coincidence of individual ability and expected ability for a
work can be used to express the work performance. In this study, the rescue performance of
flooding disaster based on urban residents‟ perception was evaluated. The urban residents
living in high potential flooding area are sampled to examine their perception and attitude in
flooding disaster rescue. The structural equation modeling is then used to develop the model to
evaluate the rescue performance of flooding disaster. The study results can supply a reference
on the prevention and protection of flooding disaster in the future.
2. Literature Review
The Article 2 at the disaster prevention and protection act, Taiwan formulates the hazards
caused by any of the following disasters: windstorm, flooding, earthquake, drought, frost, and
debris flow is called a disaster. It is a rule that the disaster prevention and protection units
should usage completely all kinds of supplies and facilities for disaster prevention and
protection. Fritz(1961) and Mileti(1999) indicated that a disaster is a hazard occurrence, which
destroys seriously the social operation and causes loss of life, property and environment.
Therefore, the disaster prevention and protection is not only the rescue of firefighting unit, but
also use well the resources of human, equipment and facilities to mitigate loss of life and
property in the shortest time (Yeh, 2003). Chen(2002) pointed out that flooding prevention and
protection affairs includes the prevention propaganda of heavy rain and strong winds to
residents before disaster, remove of collapsed tree and dropped banner, disaster area control,
victims trapped rescue, Search for missing people, pumping of basement in water, and
transportation of food and water, which need the implement and assistance of firefighters. 柯
志明(2006) took Taipei, Taiwan as an example and indicated that the influence factors of
flooding rescue performance includes the training frequency, extent and content of disaster
prevention and protection, performance assessment, unit character, management operation,
commanding and coordination of rescue workers, grasp of disaster situations, information
delivery and supports from other units. Hu(2005) studied the rescue performance of disaster
rescue system of Nantou County Government, Taiwan during Typhoon-Mindulle-caused
flooding disaster, 2004, and pointed out some shortcomings such as lack of dedicated unit,
apply procedure complicatedly of helicopters, weak role of national army and police on rescue
system, and poor communicate of official and unofficial between Organizations. Yang(2009)
emphasized that the information system should be planned synthetically to construct the
governmental information integration platform for delivering the disaster situation efficiently
and systemically based on the rescue problems of Typhoon-Morakot-caused flooding disaster.
Yang(2009) also denoted that the works of disaster rescue can be connected by information,
26
mobilization and inter-organizational cooperation based on the literatures.
3. Method
The purpose of the study was to analyze residents‟ perception toward influence factors of
rescue performance of flooding disaster and to develop a practical evaluation model of rescue
performance of flooding disaster. The residents in the high-potential area of flooding in Tainan,
Taiwan were sampled to explore their perception in influence factors of flooding rescue
performance. First, the initial questionnaire was developed based on a review of literatures and
through experts‟ interview and was revised by the content validity index. The content validity
index (CVI) proposed by Waltz and Bausell (1981) was analyzed to test the fit of items to the
domain of content represented in an instrument for developing the formal questionnaire in this
study. The four rating scale (Lynn, 1986) was used to response the proportion of experts who
are in agreement about items, which includes a rating of 1 = not relevant, 2 = somewhat
relevant, 3 = quite relevant, and 4 = very relevant. The ratings of 1 and 2 are considered
“content invalid,”whereas ratings of 3 and 4 are considered to be “content valid”(Lynn,
1986). The CVI of each item can be calculated by the proportion agreement of experts on each
item. An item can be retained as the calculated CVI is greater than the threshold of 3.5. After
executing the CVI, the content of questionnaire was ensured and the formal questionnaire was
developed. In this study, the 5-point scale including 1 = highly disagree, 2 = disagree, 3 =
somewhat agree, 4 = quite agree and 5 = agree was used and the mean of each item in the
questionnaire were calculated to obtain the agreement level of responds.
The reliability scale test was carried out for all sections in order to assessing the internal
consistency of variables. According to document of Babbie (1992), the value of Cronbach‟s
alpha was classified based on a reliability index in which 0.90 –1.00 is very high, 0.70 - 0.89 is
high, 0.30 – 0.69 is moderate, and 0.00 – 0.29 is low. Next, target residents were sampled to
examine their perception in influence factor of rescue performance for a given flooding disaster.
Finally, the structural equation modeling was used to develop an evaluation model of rescue
performance of flooding disaster based on residents‟ perception.
4. Results and Discussion
4.1 Instrument Development
Five experienced fire captains in Taiwan were interviewed in 2013 years to develop the initial
questionnaire for evaluating the resuce performance of flooding disaster. The interview results
were listed as follow.
1. As the predicted rainfall exceeds the standard rainfall before a typhoon occurs, the related
departments should composite an anti-disaster preparedness and response group
immediately for disaster prevention and protection. At this time, the firefighters play the
27
major roles on the disaster rescue works in a flooding disaster area.
2. Before disaster, the pre-deployment are conduct by the fire Bureau, Taiwan. The rescue
workers, lifeboat and equipment are pre-arranged in a potential disaster area. The disaster
deliver system should be ensured the function. The order about people evacuation should be
implemented by district office staffs and polices. As a disaster is announced and delivered
by resident, the condition of flooding disaster should be delivered correctly, the rescue work
should be implemented at once, and the police and traffic units should be noticed to
blockade the roads in a disaster area for avoiding others to enter. The victims should be
rescued instantly and send to a hospital for saving their life.
3. During disaster, the residents in a disaster area will be evacuated to many locations. A rescue
team is consisted of two firefighters and one lifeboat, and about twenty victims can be
rescued by each team one day. Of course, many teams can implement the rescue works at
the same time, and firefighters and volunteer fire can be replaced to rescue. Under the
different disaster circumstance (e.g. flooding range, flooding depth, flooding velocity,
dangerous level, victims‟ amount and time pressure) and full preparation rate of different
equipment, the rescue performance of flooding disaster has a great variation.
4. Nodaway, there are seven fire battalions comprising 53 fire detachments. The responsible
area of every fire detachment is determined by an administrative district, which will not be
adjusted based on a disaster domain. A total of 481 firefighters and 116 volunteer fires are
distributed in the potential flood disaster area in Tainan, Taiwan, which is not enough to
execute the rescue works of flooding disaster. About 200 human are needed in Tainan,
Taiwan for disaster prevention and mitigation.
5. The rescue equipment of flooding disaster in Tainan, Taiwan includes lifeboat, speedboat,
hovercraft, inflatable, water scooter, pump, command vehicles and ambulances. The full
preparation rate of above equipment is higher than 95%.
As noted in the literature review and interview records, the initial questionnaire is consisted of
three sections including rescue dispatch, rescue mobilization, and Inter-organizational
cooperation comprising a total of 9 factors, and a total of 20 items. The detailed contents of
initial questionnaire in this study depict in table 1.
Next, the content validity index (CVI) was used to revise the items of initial questionnaire. The
analytical results of CVI show all items in the initial questionnaire are greater than 3.5 (see
Table 1) and all items can be retained to form the formal questionnaire in this study.
4.2 Data Collection and Sample Characteristics
The final questionnaire was conducted with the households living in the flooding-prone
districts in Tainan, Taiwan. Only a respondent with over than 20 years old in a household was
chosen randomly to explore the perception of rescue performance of flooding disaster. In this
28
study, a total of 350 samples were distributed during July to October, 2014, and 226 samples
were obtained with a valid response rate of 67.1%. Of 266 respondents, majority of
respondents are men with a rate of 57.0% and got married with a rate of 68.9%. 42.1% had a
bachelor diploma and 23.7% are aged between 51-60 years old. The resident period of
responds are divided approximately into two groups, that is, 11 to25 years and over than 40
years. About 24.3 % of responds had suffered over than 5 experiences of flooding and the
average flooding records are 2.3.
4.3 Descriptive Information
In this study, the mean of each item in the questionnaire were calculated to obtain the
agreement level of respondents. As shown in table 1, the agreement level of items ranges in
3.975 to 4.277, which revealed responds commonly agree with the items. Among the section of
“rescue dispatch”, the highest and lowest agreement levels were the items of “Check rescue
equipment (X2)” and “Inquire ability of disaster (x4)” respectively. Similarly, among the
section of “rescue mobilization”, the highest and lowest agreement levels were the items of
“Announce and delivery system of disaster (x6)” and “Training of rescue groups (x12)”
respectively. Among the section of “Inter-organizational cooperation”, the highest and lowest
agreement levels were the items of “Follow command and contact from disaster response
center (x18)” and “Circumstance understanding degree to flooding -prone area (x15)”
separately.
4.4 Analytical Results of Structural Equation Modeling
In order to assessing the internal consistency of variables, the Cronbach‟s alpha proposed by
Babbie (1992) was used to calculate the alpha of each section. As shown in Table 1, the
resulting alpha values ranged from 0.550 to 0.738, which fall into the classification of moderate
and high. That is to say, the alpha values in this study were within the acceptable range to
assure reliability.
The structural equation modeling was used to develop the rescue performance evaluation of
flooding disaster based on 235 respondents‟ opinion. Table 2 listed the results of goodness-of –
fit of the proposed model, which shows the proposed model pass through the verification. The
weights of items in the proposed model were listed in table 3.
4.4 Illustration of Application
Table 3 presented illustrated application of the proposed model to calculate the rescue
performance of flooding disaster. The rescue performance value of flooding disaster can be
calculated by summation of score of each item. The score of each item was the multiply of
weigh and valuation of each item. The illustrated rescue performance value was 74.39, which
was greater than 70. If a threshold value is assumed to be 70, the illustrated case can be
29
regarded as good rescue performance.
5. Conclusion
The need of modern people for the service of firefighters is increasing. Thus, the rescue ability
of disaster should be strengthen and stressed for firefighter department to fit the need of people.
Under the change of global climate, the rescue of flooding disaster is major issue for firefighter
department in Taiwan. Because residents is the principal part of rescue works, the residents in a
high potential disaster area were investigated to develop an evaluation model of performance
rescue of flooding disaster using the structural equation modeling. An illustration was used to
present the calculation procedure of rescue performance of flooding disaster as well. The
proposed model should be revaluated revised through actual implementation.
30
31
References
C. E. Fritz, Disasters in R. K. Merton and R. Nisbet (Eds.), Contemporary Social Problems
(New York, 1961).
D. S. Mileti, (1999), Disasters by Design : a reassessment of natural hazards in the United
States (Washington, 1999).
C.H. Yeh, Proposing a Fire Rescue Management Strategy for High-rise Superstructure, Master
32
thesis of Graduate Institute of Civil and Disaster Prevention Engineering at National
Taipei University of Technique, Taiwan (2003).
C.Y. Chen, A study of Fire Department Rescue about Keelung River Flood-The Example of
Taipei City Fire Department Rescue Nari Typhoon, Master thesis of Department of Fire
Science at Central Police University, Taiwan (2002).
S.W. Hu, Master thesis of EMBA of Master Program in Urban Governance at Tunghai
University, Taiwan (2005).
Y.N. Yang, Research on Disaster Rescue System: The 88 Flood Case in Taiwan, Journal of
Public Administration 32, 143-169 (2009).
C.F. Waltz and R.B. Bausell, Nursing research: Design, statistics, and computer analysis.
Philadelphia:F. A. Davis. (1981).
M.R. Lynn, Determination and quantification of content validity, Nursing Research 35,
382-385 (1986).
E. Babbie, The practice of social research. California: Wardsworth Publishing Company
(1992).
J.R. Edwards, Person-job fit: A conceptual integration, literature review, and methodological
critique, International review of industrial and organizational psychology (1991).
33
SICASE-828
Fatique Resistance of Warm Mix Asphalt Using Bayat Natural Zeolite and
Elastomeric Polymer Bitumen
Ani Tjitra Handayani 1Doctoral Candidate at Civil Engineering Department, Diponegoro University, Indonesia
Bagus Hario Setiaji
Civil Engineering Department, Diponegoro University, Indonesia
Ani Prabandiyani
Civil Engineering Department, Diponegoro University, Indonesia
Abstract
The development of increasingly dense traffic and weather changes greatly affect the quality of
the pavement which could cause premature damage. Polymer modified asphalt is used to reduce
the early damage and increase the durability of pavement to various damage such as fatigue,
cracking due to temperature changes. All this time Polymer asphalt mixing process is using hot
asphalt metode. According to the name the hot mixing asphalt requires heating at a certain
temperature which is quite high on the Asphalt Mixing Plant (AMP), also requires a certain
density. Some countries have developed asphalt mixture known as Warm Mix Asphalt by mix
and compaction temperatures lower than hot asphalt mixtures using added material. Research
using experimental methods with polymer bitumen as a binder in the warm mixed asphalt with
the added material of natural zeolite, thus obtained mixture quality hot asphalt mixture
equivalent but processed by low mixing and compaction temperatures. Levels of 1% natural
zeolite materials added to the warm mix asphalt with asphalt binder polymer in addition to
reduce the temperature of mixing / compaction temperature of 10 ° C on a hot asphalt mixture of
pure asphalt binder without zeolite from 156oC becoming 146oC also have greater fatigue
resistance, of hot mix asphalt with pure asphalt binder.
Keywords: Fatigue Resistance, Zeolite as a Material added, Asphalt Polymers
1. Introduction
1.1 Background
Polymer modified asphalt is used to improve the durability of pavement to various damage such
as fatigue, permanent deformation, cracking due to temperature changes, [1].
34
But unfortunately polymer modified asphalt mixing process, this time using the method of Hot
Mix Asphalt. Hot asphalt mixture using a polymer modified asphalt intended to increase
resistance to a variety of damages, require the mixing temperature in Asphalt Mixing Plant,
which is higher around 175oC [2]. Mixing temperature is the temperature where the bitumen has
a viscosity of 170 ± 20 cSt viscosity whereas for asphalt compaction temperatures between 280 ±
30 cSt [3]. High temperature of mixing and compaction in the process of making asphalt
mixture resulted in a lot of fuel, exhaust emissions are large and high costs [4].
Some countries have developed asphalt mixture with mixing and compaction temperatures lower
than hot asphalt mixture known as Warm Mix Asphalt. Warm mix asphalt technology has
become the main method in the use of road construction in the United States and the world, it is
because the fuel savings, environmentally friendly [8]. Warm Mix Asphalt, produced in AMP
and compacted in the field while maintaining the required workability so it can be spread and
compacted well [5]. Classification of warm mix asphalt based on the technology, can be
classified into three types; based on chemical uses, water uses and process using water and
additive[5]. Most research warm mix asphalt using pure asphalt binder with the added ingredient
of synthetic zeolite. In this research, the material used is asphalt binder polymer elastomer types
Starbit E-55and the added materials used are natural zeolite Bayat. Natural zeolite serves as a
material added by utilizing the water it contains, can increase the volume of bitumen/asphalt
during mixing and compaction [7]. Using elastomeric polymer bitumen as the bonding material,
While the bonding material using asphalt elastomeric polymer, with with expectations of getting
a mixture that has a great fatigue resistance that can be mixed and compacted with a bunch of
lower temperature than the hot mix asphalt with pure asphalt binder that is often used in
Indonesia.
1.2 Objectives
The main objectives of this research include: (1) examine the effect of zeolite as an added
material to develop warm mix asphalt using asphalt polymer with the added material of natural
zeolite (2) assess the optimum zeolite content that meets the specifications Marshall (3) assess
the fatigue resistance of asphalt hot mix asphalt with polymer binder at the mixing temperature
GCC and compaction temperatures lower than hot mix asphalt binder pure asphalt penetration 60
/ 70.
2. Experimental Work
2.1 Materials and Mix Proportion
This warm mix asphalt research using materials added natural zeolite, types of mineral
mordenite sourceds from Bayat, Central Java, Indonesia, in powder form through sieve no.400.
35
Natural zeolites are used because they are cheap and Indonesia is one of the countries with the
largest deposits of natural zeolites [11].
Before use the zeolite first processed by chemical activation methods to clean the surface of the
pore, to dispose impurity compounds, and re-measure the location of the atoms that are
exchanged, so that the zeolite pores can be cleaned to increase the water content to be used, In
this research, the polymer modified asphalt was a polymeric bitumen elastomeric. Based on the
ability of Elastomeric polymer bitumen are more resistant to deformation after receiving a load
on its surface, will stretch the surface and return to its original shape after the load is lost. In
this research, Elastomeric polymer bitumen/asphalt used was asphalt Starbit E-55. To get a
comparable mix of warm asphalt mixture using elastomeric polymer bitumen without zeolite
[10]. The aggregate was from Subang, West Java, with the combined aggregate gradation to
wear layers.
2.2 Details of Experimental Tests.
The first thing to do is to test the materials to be used, the aggregate, Elastomeric polymer
bitumen/asphalt (Starbit E-55) and natural zeolite, penetration bitumen 60/70. Asphalt testing
performed on elastomeric polymer bitumen and penetration bitumen 60/70. Tests conducted on
two kinds of aggregate, the coarse aggregate and fine aggregate, which are both taken from
Subang, West Java. Before making an asphalt mixture, natural zeolite as an ingredient added,
activated in advance using chemical activation. Then Marshall experiments conducted to obtain
the optimum bitumen content in asphalt content between 5% to 7%, with increases of 0.5% each,
on any mixture containing zeolite content of 0%, 1% and 3% were mixed at temperatures of
mixing and compaction temperatures, according to the results of testing of elastomeric polymer
bitumen 0% zeolite is between 173 ° C - 179 ° C, whereas, the compaction temperature between
159 oC - 165 oC. In this study for normal mixture (without zeolite) for comparison with the
binder is a mixture of pure bitumen 60/70 penetration mixing temperature was set to 156 ° C, and
compaction temperatures of 142 oC.
Elastomeric polymer asphalt mixture characteristics were evaluated using Marshall [10] on each
mix asphalt Starbit E-55 with levels of natural zeolite (0%, 1%, 3%). Based on the experiment
Marshall obtained the optimum bitumen content of 5.8% (0% zeolite content), 5.86% (1%
zeolite content) and 5.95% (3% zeolite content).
Blend asphalt mixture Starbit E-55, with the zeolite content variation tested with Marshall. Then
obtained mixture that produces values that meet the specifications. The connection of viscosity
and temperature of the elastomeric polymer bitumen with material added natural zeolite can not
be done, because the process of mixing zeolite done by the dry method. Aggregate heated first,
then the natural zeolite and asphalt poured simultaneously. If seen by viscosity and is associated
only with the mixing and compaction temperatures, as if the addition of natural zeolite as
36
material added cannot reduce the temperature of mixing and compaction.This is in accordance
with research which states that viscosity measurements and conditions in this way only suitable
for conventional asphalt, but not suitable for warm mix asphalt testing [12]. To find out the
fatigue resistance of warm mixed asphalt using asphalt binder Starbit E-55, which combined
natural zeolite as an added ingredient, testing the mixture with mixing and compaction
temperatures lower than normal temperature (176oC / 162oC). Mixing / compaction temperature
used was 166oC / 152oC; 156oC / 142oC; 146oC / 132oC. For comparison hot mix asphalt with
pure asphalt binder mixed and compacted at a temperature of 156oC / 145oC.
Analysis will be conducted on the experimental results, which include testing of all the
properties of the mixture are planned. The nature of the mixture and volumetric done by testing
on a variety of mixing and compaction temperatures, with several levels of zeolite materials
added, and compare the properties of each of the pure asphalt binder and asphalt mixture with a
polymer binder.
3. Results and Discussion
3.1 Aggregate Tests performed on coarse aggregate, fine aggregate and gradation, the results
can be seen in Table 1, Table 2 and Figure 1.
37
Based on Table 1 the results of testing of coarse aggregate and Table 2 the results of testing of
fine aggregate, and Figure 1 gradation test results, the results showed that all three test
materials that will be used to make the mixtures according to the specifications. Zeolite using
chemical activation method test result produces moisture content of 18.99%. Table 3 shows the
results of testing 60/70 penetration bitumen, bitumen of penetration 60/70 looks that meet the
specifications [3] so that the material consists of aggregates, asphalt and zeolites can be used to
make asphalt mixture. In this study, the effect of the use of warm mix asphalt binder to be
analyzed based on the results of the Marshall test. Mix asphalt, penetration 60/70, used as a
benchmark mix from warm mix asphalt with asphalt binder polymer and additive materials
zeolite levels 1% and 3% which used to reduce the temperature. Effect of natural zeolite levels
on asphalt mixture Starbit E-55 as shown in Table 3.
38
In Table 1, it appears in general that, with using the polymer bitumen as a binder in hot mix
asphalt with added zeolite materials as a medium to reduce the temperature, make VIM value
increased due to the water vapor produced from heating the zeolite causing the air cavity
between the aggregate grains encased in asphalt increased, while the value of the stability
increased on the condition of 1% zeolite content, but the content of zeolite for more than 1% lead
to decreased stability values. Thus, the mixture can be selected to decrease the temperature of
mixing and compaction temperatures are elastomeric polymer modified asphalt mixture with
zeolite content passes 1% as shown in Table 5.
To determine the effect of the use of polymer bitumen as a binder in the warm mixed asphalt to
the fatigue resistance, can be seen in Table 5, Table 6 and Figure 2.
39
Based on Figure 2 and Table 6 and Table 7 shows that the strain rate is inversely proportional to
the amount of loading up to collapse, the greater the strain given, Causes, the number of loading
cycles will be shorter. This is because the magnitude of the applied strain causes the greater the
voltage generated to maintain the strain. The load received in the mixture will be greater, thus
have an impact on
In general, asphalt mixture with asphalt binder Starbit E-55 and 1%n natural zeolite as a material
added had pretty good resistance to fatigue compared with the asphalt mixture with pure asphalt
binder penetration 60/70. Can be seen from the average number of repetitions of loading to
collapse in asphalt mixture of normal Starbit 66367.5 higher than in the mixture of asphalt binder
asphalt, penetration 60/70 size 39097.5.
40
The average load repetitions to collapse on the warm mixed asphalt, asphalt mix Starbit E-55
with 1% zeolite higher than the mixture asphalt penetration 60/70.
4. Conclusions
Based on the test results, the influence of the polymer as a binder asphalt on warm mix asphalt
with natural zeolite as a material added, can be concluded, that the asphalt binder polymer can be
used in hot mix asphalt to the required temperature for mixing / compaction 10oC lower than the
required temperature at the warm mix asphalt with asphalt binder 60/70 penetration. Based on
the average number of repetitions, loading to collapse on asphalt mixture of normal Starbit
66367.5 higher than the asphalt mixture with asphalt binder 60/70 penetration 39097.5. The
average load repetitions to collapse on the warm mix asphalt blend asphalt Starbit E-55 with 1%
zeolite higher than the asphalt mixture 60/70 penetration.
5. Acknowledgement
The authors highly appreciate the financial support from Sekolah Tinggi Teknologi Nasional
Yogyakarta (STTNAS Yogyakarta). The support of the Building Material Laboratory staffs at
the Faculty of Engineering, Pusat Penelitian dan Pengembangan Jalan, Bandung,Indonesia and
for their assistance in conducting the experimental works are also gratefully acknowledged.
Disclosure
The paper has not been previously published, is not currently submitted for review to any other
journal, and will not be submitted elsewhere before a decision is made by this journal.
References
[1] Yildrim, Y. 2007.Polimer Modified Asphalt Binders, Journal Construction & Building
Materials, Vol.21 No.1. pp.66-72,2007.
[2] Indonesia, Kementrian Pekerjaan Umum. Spesifikasi Umum Revisi 2. Jakarta: Ditjen Bina
Marga, 2010.
[3] Asphalt Institute. Mix Design Methods for Asphalt Concrete and Other Hot Mix Types.
MS-2 Sixth Edition Lexington: Asphalt Institute, 1993.
[4] D‟Angelo, J., Harm, E., Bartoszek, J., Baumgardner, G., Corrigan, M., Cowsert, J.,
Harman,T., Jamshidi, M., Jones,W., Newcomb, D., Prowell, B., Sines, R., dan Yeaton, B.
2008, „Warm Mix Asphalt: European Practice‟, Publication FHWA-PL-08-007. FHWA,
U.S Depertmen of Transportation, February 2008.
[5] Gierhart, D. Warm Mix Asphalt-what is it and how can we benefit?. In SE States Pavement
Association Management and Design Conference Southeast Pavement Preservation
Partnership Meeting.Lexington, 2009.
[6] Cho, Doong Woo. Development of Warm Mix Asphalt Technology. In Joint Work Shop
Between KICT and IRE October 29-30, 2012. Bandung.IRE, 2012.
41
[7] Prowell, B.D, Hurley,G.C and B. Frank. Warm Mix Asphalt. Best Practices, National
Asphalt Pavement Association, Lanham:NAPA, 2011.
[8] Hurley, G. dan Prowell, B.D. Evaluation of Potential Processes for Use in Warm Mix
Asphalt, Journal of the Association of Asphalt Paving Technologists. Proceedings of the
Technical Sessions, 2006 Annual Meeting. Vol.75, pp. 41-90, 2006.
[9] Hurley, G. C. dan Prowell, B. D. 2009, Evaluation of Aspha-Min Zeolite for Use in
Warm-mix Asphalt, NCAT Report 05-04, National Center for Asphalt Technology, Auburn
University, Auburn, Alabama
[10] American Association of State Highway and Transportation Officials 2012. Resistance to
plastic flow of bituminous Mixtures Using Marshall Apparatus. AASHTO 2008. Standard
Specifications for Transportation Materials and Methods of Sampling and Testing,
Washington, DC.
[11] Sugih, P. 2008, Evaluasi Memanfaatkan Sumber Daya Alam Daerah di Sumatera Barat,
Media Litbang Provinsi Sumatera Utara, Vol.5, No.4, pp.246-254.
[12] Zaumanis. M. 2010. Warm Mix Asphalt Investigation Master of Science. Thesis, Riga
Technical University.
42
Computer & Information Science I / Electrical and Electronic
Engineering
Session Chair: Yonglei Tao
SICASE-727
Pw/GAKE: An Efficient and Provably-Secure Password-based Authenticated Group-Key
Exchange Protocol
Wuyang Cai︱Dalian University of Technology
Yuan Tian︱Dalian University of Technology
SICASE-388
The Newest Double-Population Particle Swarm Optimization
Wenjing Liu︱Inner Mongolia University of Science & Technology
Jianguo Wang︱Inner Mongolia University of Science & Technology
Zhijie Zhang︱Inner Mongolia University of Science & Technology
Da Lv︱Bao Tou College
SICASE-776
Learning Issues in Opportunistic Software Development
Yonglei Tao︱Grand Valley State University
SICASE-818
Active Low Pass Filter with Transmission Zero Using OTA-TURC
Sorapong Wachirarattanapornkul︱King Mongkut’s Institute of Technology Ladkrabang
(KMITL)
2015/6/27 Saturday 13:00-14:30 Room 2
43
SICASE-727
Pw/GAKE: An Efficient and Provably-Secure Password-based
Authenticated Group-Key Exchange Protocol
Wuyang Cai
Software School, Dalian University of Technology, P.R. China
Yuan Tian
Software School, Dalian University of Technology, P.R. China
Abstract
The authenticated group key exchange protocol is one of the most fundamental security
components in network environment, particularly in contemporary wireless mobile/ad hoc
networks. The password-based key exchange protocol accepts quite a lot of research because of
its practical advantages, however, most schemes are for 2-party cases. In this paper a
highly-efficient, password-based authenticated group-key exchange protocol named Pw/GAKE
is constructed which security is based-on the Decisional Diffie-Hellman Problem‟s hardness.
During this protocol‟s session all group members have only one-round message transmission
and only need to carry out exponential and fast hash computation(no public-key schemes are
required), making it particularly suitable to small-scale groups in modern wireless/ad hoc
networks. The protocol‟s security is proved in random-oracle model by reducing its security to
Abdella-Pointcheval‟s 2-party key-exchange protocol (SPAKE)‟s security.
Keywords: Password-based Authentication, Group Key Exchange, Decisional Diffie-Hellman
Problem, Provable Security.
1. Introduction
Authenticated group-key exchange protocol is one of the most critical security infrastructures
for modern networks, particularly for wireless mobile and ad hoc networks. In comparison with
traditional 2-party key-exchange mechanisms, group-key exchange is subjected to some new
requirements and properties together with more complicated security proof[1-4,17-20]. For
example, the realistic group communication environment is usually dynamic in which group
members can join or leave, as a result the group-key needs to be updated correspondingly and
the communication complexity is relatively more demanding than 2-party protocols.
Furthermore, a reliable group-key exchange protocol must have a rigorous security proof in
terms of contemporary cryptography theory. So far lots of group-key exchange protocols have
been proposed, some by cryptography but more by application community, however, only few
44
of them have satisfactory security proofs. Most schemes have only heuristic analysis which is
quite insufficient for secure applications, e.g., the famous Ateniese-Steiner-Tsudik protocol[2]
is found to have lots of security flaws[19]. Fortunately Bresson, et al. in [3] established a
widely acceptable group-key exchange security model which is similar as the semantic
(computational indistinguishability) security model established by Bellare-Rogaway for 2-party
key exchange protocols.
Authentication is one of the necessary requirements for the group-key exchange protocol. In
contrast to the long random secret number, the short and easy-to-remember password is an
efficient and popular solution to authentication in practice. Intuitively speaking, a
password-based solution is secure if it prevents any dictionary attacks and as a result the
relatively most efficient attack is (provably) on-line password guessing.
Besides the 2-party password-based authenticated key-exchange protocols[5-11] there are also
solutions to the case of group communications[1,12-15,20], among which most are in the
random-oracle model and based-on some Diffie-Hellman hardness variants. Some of them
have O(N) while others have O(1) communication complexity[13] (where N is the group size).
The widely accepted security model for the password-based authenticated group-key exchange
protocol is recently established in [12]. In [20] a O(1) communication complexity solution is
constructed in standard-model which computation complexity is much higher than
random-oracle solutions. Compared with the 2-party protocol design, there are more challenges
to the efficient group solutions.
2. Contributions
In this paper we construct an efficient and provably-secure password-based authenticated
group-key exchange protocol based-on the decisional Diffie-Hellman hardness, named
Pw/GAKE hereafter, in the random-oracle model. This construction is an interesting extension
of the Abdalla-Pointcheval‟s 2-party protocol SPAKE[7] in the group scenario([7] proposed
two versions, SPAKE 1 and SPAKE 2, which slight differences are not essential to our
Pw/GAKE‟s construction so they are simply called SPAKE hereafter). Pw/GAKE‟s security is
proved in terms of security notion in [12]. More interestingly, this security proof is a reduction
from Pw/GAKE‟s security to SPAKE‟s security (i.e., any attack against Pw/GAKE is reduced
to an attack against SPAKE).
As for the performance, in addition to the high computation efficiency a random-oracle-based
solution can generally have, Pw/GAKE needs no complicated cryptographic computations and
its only operations come from hash and power calculations on the underlying (mathematical)
group. The typically related works are [12,13]. [12]‟s prototype comes from [2] and [13] is
actually a variant of the Bermester-Desmedt scheme. All these works (including ours) are
based-on essentially the same computational hardness hypothesis. Pw/GAKE and [12] have
45
O(N) communication complexity but Pw/GAKE needs only one-round message
communication (during which the message transferred from the first to the last group member)
and one broadcast, however, [12] needs one-round communication plus N peer-to-peer
authentications between each group member and the group controller. With respect to
computation complexity, each group member in Pw/GAKE only needs O(N) power, one hash
and one message authentication code(just another hash) operation while each member in [12]
needs O(N) power, O(1) hash (but O(N) for the controller) and one zero-knowledge proof
operation. [13] has only O(1) communication complexity but it needs (besides the random
oracle) the ideal cipher scheme. Each member in [13] needs O(1) power, O(1) hash and O(1)
ideal cipher encryption/decryption operations. In practice the ideal cipher‟s implementation is
much more complicated than the random oracle. In addition, [12-13] need three random oracles
but Pw/GAKE needs only one.
It‟s worthwhile to indicate that a theoretical solution with O(1)-communication complexity is
not always efficient in realistic network environments. Its real efficiency depends on the
underlying media access control mechanism[21]. For example, in the collision-free network
(e.g., TDMA channel) such solution can reach its ideal performance. However, in the colliding
network (e.g., wired/wireless Ethernet or CSMA channel) the actual communication
complexity depends on the channel‟s overhead: the higher the overhead is, the closer the
performance is to the O(N) solution, as a result in such environment the solutions with O(1)
and O(N) communication complexity usually have no significant differences.
Organization Section 3 presents necessary notions, hardness hypothesis and the security
model. Section 4 presents Pw/GAKE‟s construction. For the sake of space all details and proofs
are put in the appendices.
Notations x||y represents the concatenation of the string x and y. |x| is the bit-size of the binary
string x. If X is a set, |X| represents X‟s cardinality and a←$X represents that a is generated
from X at random with uniform distribution. Poly (k) represents a given polynomial in the
complexity parameter k.
P.P.T. means “probabilistic polynomial time”,PPT
means “P.P.T. indistinguishability”,... FDP
means “perfect indistinguishability”(i.e., having the same probabilistic distribution
function(P.D.F)). All algorithms are represented via pesudo programming language where
/*...*/ presents the comments.
46
3. Computational Hardness Hypothesis and Security Model
3.1 Group Diffie-Hellman Problem and the Hardness Hypothesis
q is a k-bit prime, G is a (cyclic) group of order q and g is a generator of G.
Definition 3.1(Diffie-Hellman Configuration) Let N be an interger greater than 2, x1,…,xN∈Zq,
z(j,i)∈G for 1≤j≤i≤N, ρ∈G. The expression ij xxx ...ˆ...1 represents ijj xxxx ..... 111 . An expression
ξ≡(z(j,i):1≤j≤i≤N; ρ) is defined as a true G-DDHP(N) configuration on (x1,…,xN) if for each j
and i we have z(j,i)=ij xxx
g...ˆ...1
and ρ=Nxx
g......1
; ξ is defined as a false G-DDHP(N) configuration
on (x1,…,xN) if for each j and i we have z(j,i)=ij xxx
g...ˆ...1
but ρ is a random element in G.
Definition 3.2(Hardness of Decisional Diffie-Hellman Problem, or G-DDHP(N) Hardness for
simplicity) Let k be the complexity parameter, N=poly(k), Y=(G,g)q(k) be the ensemble of
prime order (cyclic) groups, g be a (public) generator of G. For any P.P.T. algorithm A, consider
the following experiment)()(
, kExp NDDHPG
AG
:
x1,…,xN←$Zq;
z(j,i)←ij xxx
g...ˆ...1
, 1≤j≤i≤N;
b←$0,1;
ρ0←Nxx
g......1
; ρ1←$G;
/* so (z(j,i),1≤j≤i≤N; ρ0) is a true G-DDHP(N) configuration on (x1,…,xN) while (z(j,i),1≤j≤i≤N;
ρ1) is the false configuration on (x1,…,xN).*/
d←A((z(j,i),1≤j≤i≤N; ρb));
output(b d);
Define A‟s advantage)()(
, kAdv NDDHPG
AG
≡|2P[1)()(
, kExp NDDHPG
AG ]-1|
= |P[x1,…,xN←$Zq; ξ←(x1,…,xN)‟s true G-DDHP(N) configuration: A(ξ)=1]
- P[x1,…,xN←$Zq; ξ←(x1,…,xN)‟s false G-DDHP(N) configuration: A(ξ)=1]|
(The (x1,…,xN) in the above two probabilistic expressions represent the same group of random
numbers). The G-DDHP(N) hardness on Y means that for any P.P.T. algorithm A its advantage
)()(
, kAdv NDDHPG
AG
is negligible in k. The upper-bound of the advantage for all A is notated as
)()( kAdv NDDHPG
G
.
47
For our Pw/GAKE‟s security proof(theorem 4.1) the following statement is needed.
Theorem 3.1[3,16]
k、N、Y=(G,g)q(k) have the meaning in definition 3.2. If the decisional
Diffie-Hellman problem is hard on Y, then the G-DDHP(N) problem is also hard on Y.
Concretely: )()( tAdv NDDHPG
G
≤ O(N)))(( 3NOtAdv DDHP
G .
3.2 Security Model
This subsection presents the security model for password-based authenticated group-key
exchange protocol. This model is in computational indistinguishability style and is essentially
the same as that established in [12,13]. Furthermore, in case of two members this model is
reduced to the traditional 2-party password-based authenticated key exchange protocol‟s
seurity mode(e.g., that established in [7]).
A password-based authenticated group-key exchange protocol Π consists of three transactions
(sub-protocols): pwGrpKey, pwJoin and pwRemove. k denotes the complexity parameter,
A=(A1,A2) denotes the two-phase attacker and )((.), kExp pwGKS
denotes an environment providing
A with the following interfaces which are modeled as oracles accessed by A to describe A‟s
passive/active attacking capabilities.
GrpSetup(aux, gid): create a group and assign a set of secret information (including the
password) for group members. aux represents any public information assigned by A, e.g., the
members‟ identities, and gid is the group identity. Furthermore we suppose that every occurred
group transaction is associated with a unique sid and the global transaction identity is gid||sid.
Execute(opcode, param): run a protocol‟s transaction‟s (i.e., one of pwGrpKey, pwJoin and
pwRemove) instance where param is the set of running parameters assigned by A. For example,
Execute(“pwGrpKey”, gid) generates a group-key on the group gid, Execute(“pwJoin”, (gid,U0))
adds a member U0 to the group gid. Execute returns the transaction‟s transcripts (message flow)
to A. Actually this interface describes passive attacks from A(in particular this provides all
information for A to launch its dictionary attacks).
Send(U, M): send the attacker-crafted message M(always including gid||sid as one of its
components) to the member U and return the response of U to A. this interface describes A‟s
active attacks and w.l.o.g. the calling to Send can be regarded as the on-line password-guessing
attacks.
KeyReveal(gid, sid): return the group-key, denoted by Kg(gid||sid), of the group gid to A
generated during its transaction sid and label the transaction gid||sid as “disclosed”. If the
transaction gid||sid has finished then the returning value is null. This interface describes the
48
known-key attacks and the protocol‟s forward-security.
Corrupt(gid, sid, U): return to A with all instantaneous states of the transaction sid and
permanent secrets of the member U in group gid( if the password can vary with transaction
instances then the password is regarded as an instantaneous state component, otherwise it is a
permanent secret). Furthermore, all transactions on group gid participated by U at and after the
time when Corrupt is called will be labeled as “disclosed” and U is labeled as “corrupted”.
This interface describes A‟s intrusion and known-key attacks (i.e., A collects passwords from
those groups other than the target. We note that some proposed protocols can really resist such
attack but not explicitly present it).
Let Δ denote the candidate password dictionary. Consider the following experiment for the
protocol Π:
1))(, kExp conspwGKS
A
/*for key-consistency property*/
gid*||sid*←AO(.,.)
(k,Δ);
/*where O(.,.) represents the interfaces accessible by A, e.g.,
O(“GrpSetup”,s) means A calls GrpSetup(s), O(“Send”, (a,M))
means A calls Send(a,M), etc. */
if (in the transaction gid*||sid* all members normally terminate
∧there exist members Ui and Uj such that their output K*i≠K*j
∧A doesn‟t corrupt Ui and Uj)
/* where K*i is the group-key output by Ui*/
then output(1);
else output(0);
Define A‟s advantage )(, kAdv conspwGKS
A
≡P[)(, kExp conspwGKS
A
=1].
2))(, kExp indpwGKS
B
, B=(B1,B2): /*for key-security property*/
(St, gid*||sid*)←B1O(.,.)
(k,Δ);
/*w.l.o.g., suppose the transaction gid*||sid* already finished.*/
if( in the transaction gid*||sid* all members normally terminate
∧all members output the same group-key K*g )
then
Z0←K*g; Z1←$0,1
|Z0|; b←
$0,1;
d←B2O(.,.)
(St,Zb); /*Zb is called B‟s test-key hereafter.*/
if (transaction gid*||sid* is not labled as “disclosed”)
output(b d);
else halt;
49
else halt;
Define B‟s advantage )(, kAdv indpwGKS
B
≡|P[)(, kExp indpwGKS
B
=1|～halt]-1/2|.
The upper-bound of such advantages for all attacking algorithms are notated by
)(kAdv conspwGKS
and )(kAdv indpwGKS
respectively. The protocol Π is defined as semantic
secure (or secure for simplicity), if 1)for any P.P.T. attacker A, the advantage )(, kAdv conspwGKS
A
is negligible, and 2)for any P.P.T. attacker B, we have )(, kAdv indpwGKS
B
≤O(qs)/|Δ|+έ(k) where
έ(k) is negligible in k and qs is the number of queries for B to call its oracle Send(.).
4. Pw/GAKE
Let q be a prime number and its bit-size |q|=k be the complexity parameter. G is a (cyclic)
group of order q and g is G‟s (public) generator. H:0,1+→0,1
L(k) is a random
oracle,∑=(MKG,Tag,Vf) is a UF_CMA secure MAC scheme and f is a pesudo-random
function.
N is the number of group members. For a given group G=U1,…,UN, Ui means it‟s i-th
member which has a public element Mi∈G. All Ui share the password pw.
The adjacent member identifier i and i+1 means that the member Ui and Ui+1 are neighbors in
some specific sense(only meaningful in communication but not in security). s represents a
unique transaction identifier. Each group member Ui has its own specific state including a
secret random number xi. The expression ij xxx ...ˆ...1 means ijj xxxx ..... 111 . To avoid lengthy
expressions we hereafter often use the set of words to express the concatenation of these words
(the meaning is always clear).
Protocol Pw/GAKE:
U1 randomly generates x1←$Zq and a new transaction identifier s, stores its state (x1,
pwxMg 1
1
)
and sends U2 with the following message:
(init, s, pwx
Mg 11
)
On receiving the message from U1, U2 verifies that s is fresh, randomly generates x2←$Zq,
stores the state (x2,pwx
Mg 22
) and sends U3 with the message
50
(init, s, pwx
Mg 11
||pwx
Mg 22
||pwxx
Mg 221
)
(Note that pwxx
Mg 221
=pwxpwpwx
MMMg 21121 )/)(( and all
pw
iMcan be pre-computed).
In general, Ui(1≤i≤N-1) randomly generates xi←$Zq, stores its state (xi,
pw
i
xMg i
) and sends Ui+1
with the message
(init, s, pwx
Mg 11
||…||pw
i
xMg i
||pw
i
xxMg i...1
||pw
i
xxxxMg iij 11 ...ˆ...
:1≤j≤i-1||pw
i
xxMg i
1
...... 11
)
When UN receives the following message
(init, s, pwx
Mg 11
||…||pw
N
xMg N
11
||pw
N
xxMg N
1
... 11
||pw
N
xxxMg Nj
1
...ˆ... 11
:1≤j≤N-2||pw
N
xxMg N
2
...... 21
)
from UN-1 it does:
verify that s is fresh;
xN←$Zq;
K(N)←NN xpw
N
pw
N
xxMMg )/( 11
... 11
; K a(N)←f(K(N)||1);
Kg(N)←H(U1||…||UN||pwx
Mg 11
||…||pw
N
xMg N
||pw||K(N));
erase xN;
tN←Tag(Ka(N), N||s||pwx
Mg 11
||…||pw
N
xMg N
);
broadcast the following message
(grpkey, s, pwx
Mg 11
||…||pw
N
xMg N
||pw
N
xxxMg Nj ...ˆ...1
:1≤j≤N-1, tN);
On receiving the above broadcasted message each Ui(1≤i≤N-1) does:
K(i)←iNNi xpw
N
pw
N
xxxxMMg )/)(( 11 ...ˆ...
; erase xi;
Kg(i)←H(U1||…||UN||pwx
Mg 11
||…||pw
N
xMg N
||pw||K(i));
/* Kg(i) is the group-key generated by Ui, pw
i
xMg i
comes from Ui‟s state and other pw
j
xMg j
‟s
come from the message. */
Ka(i)←f(K(i)||1);
Verify Vf(Ka(i), N||s||pwx
Mg 11
||…||pw
N
xMg N
, tN)=1;
51
Broadcast (end, s, Tag(Ka(i), i||s||pwx
Mg 11
||…||pw
N
xMg N
));
On receiving the broadcasted messages (end, s, tj) from Uj(j≠i,N) each Ui(1≤i≤N) verifies that
Vf(Ka(i), j||s||pwx
Mg 11
||…||pw
N
xMg N
, tj)=1. If the verification is true for all j≠i,N the Ui outputs
“accepted, Kg(i), U1,…,UN” otherwise outouts “rejected”.
Note that when N=2 the Pw/GAKE is just the protocol SPAKE2(or SPAKE1 if pw is not
included in the group-key‟s computation formula).
It‟s not hard to check that Pw/GAKE has all features overviewed in section 2. In summary,
the design of Pw/GAKE mainly for small-size group applications is reasonable and practical.
Let Y=(G,g)q(k) be an ensemble of prime order (cyclic) groups on which DDHP is hard.
N is any given constant. Other symbols, e.g., H, ∑, f, have the same meaning as in before.
Theorem 4.1 Pw/GAKE is secure under the above conditions.
5. References
[1] M.Manulis, Security-Focused Survey on Group Key Exchange Protocols, Tech. Report
06-27, Horst-Goertz Institute, Ruhr University of Bochum, Mar 2006.
[2] G.Atenses, M.Steiner, G.Tsudik New Multiparty Authentication Services and Key
Agreement Protocols, IEEE Journal of Selected Areas in Communications, 18(4):628-639,
2000
[3] E.Bresson, O.Chevassut, D.Pointcheval Provably-Secure Authenticated Group
Diffie-Hellman Key Exchange. ACM Transactions on Information and System Security,
16(3):421-454, 2011.
[4] Y.Kim, A.Perrig, G.Tsudik, Tree-based Group Key Agreement, ACM Transactions on
Information and System Security, 7(1):60-96, 2004.
[5] P..Mackenzie, The PAK Suite: Protocols for Password Authenticated Key Exchange, Tech.
Report 2012-46, DIMACS, 2012.
[6] P. Mackenzie, More Efficient Password Authenticated Key Exchange, In D.Naccache ed:
RSA Conf‟01, LNCS Vol.2020, 351-377, 2001.
[7] M.Abdella D.Pointcheval Simple Password-Based Encrypted Key Exchange Protocols, In
Menezes ed: Topics in Cryptology, LNCS Vol.3376, 191-208, 2005.
[8] M.Bellare, D.Pointcheval, P.Rogaway Authenticated Key Exchange Secure Against
Dictionary Attacks, Proc. Eurocrpt‟00, LNCS Vol.1807, 139-155, 2000.
[9] R.Gennaro, Y.Lindell, A Framework for Password-based Authenticated Key Exchange,
Proc. Eurocrypt‟03, LNCS Vol.2656, 524-543, 2003.
[10] R.Canetti, S.Halevi, J.Katz et al, Universally Composable Password-based Key Exchange,
Eurocrypt‟05, 2005.
52
[11] Feng Deng-Guo, ChenWei-Dong Modularized Design and Analysis of Password-based
Security Protocol, Science in China Series E, 2007,37(1):223-237
[12] E.Bresson, O.Chevassut, D.Pointcheval A Security Solution for IEEE 802.11 Ad hoc Mode:
Password-authentication and Group-Diffie-Hellman Key Exchange, International Journal
of Wireless and Mobile Computing, Special Issue on Security of Computer Networks and
Mobile Systems, 2(1), 4-13, 2007.
[13] M.Abdalla, E.Bressen, O.Chevassut et al, Password-based Group Key Exchange in a
Constant Number of Rounds, In Proc. 9th
International Workshop on Theory and Practice
in Public Key Cryptography(PKC‟06), LNCS Vol.3958, 427-438, 2006.
[14] R.Dutta, R.Barua, Password-based Encrypted Group Key Agreement, International Journal
of Network Security, 6(1):30-41, 2013.
[15] E.Bressen, O.Chevassut, D.Pointcheval, Group Diffie-Hellman Key Exchange Secure
Against Dictionary Attacks, Proc. ASIACRYP‟02, LNCS vol.2501:497-514, 2002
[16] .E.Bressen, D.Pointcheval, The Group Diffie-Hellman Problems, Proc. SAC‟02, In
K.Nyberg ed: LNCS Vol.2595, 325-338, 2002.
[17] Hyun-Jeong Kim, Su-Mi Lee, Dong Hoon Lee, Constant-round Authenticated Group Key
Exchange for Dynamic Groups, In P.J.Lee ed: Advances in Cryptology‟04, LNCS
Vol.3329, 245-259, 2004.
[18] J.Katz, M.Yung, Scalable Protocols for Authenticated Group Key Exchanges, In D.Boneh
ed: Advances in Cryptology‟03, LNCS Vol.2729, 110-125, 2003.
[19] O.Pereria, J-J Quisquater Some Attacks on Authenticated Group Key Agreement Protocols,
Journal of Computer Security, 22(4):555-580, 2012.
[20] J-M Bohli, M.L.G.Vasco, R.Steinwandt Password-Authenticated Constant-Round Group
Key Establishment with a Common Refernce String, Public Key Cryptology, LNCS
Vol.3958, 2006.
[21] K.Pahlavan, P.Krishnamurthy Principles of Wireless Networks: A Unified Approach(5th
ed),
Prentice-Hall Inc., 2005.
6. Appendices
6.1 Some Auxiliary Lemmas
For the security proof in section 4 we present some auxiliary conclusions which also may be of
independent values.
For a given P.P.T. algorithm BN consider the experiment)()(*
, kExp NDDHPG
BG N
(called
G*-DDHP(N) problem hereafter):
x1,…,xN←$Zq;
b←$0,1;
53
ρ0←Nxx
g......1
; ρ1←$G;
d←BN(1x
g ||…||Nx
g || ρb);
output(b d);
Define BN‟s advantage)()(*
, kAdv NDDHPG
BG N
=|P[)()(*
, kExp NDDHPG
BG N
=1]-1/2| and its upper-bound for
all BN is notated as)()(* kAdv NDDHPG
G
. Note that G*-DDHP(2) is just the decisional
Diffie-Hellman problem(DDHP).
Δ is a public dictionary. For a given P.P.T. algorithm CN consider the experiment
)()(*
),,( kExp NDDHPMG
CG N
(called G*M-DDHP(N) problem hereafter)：
x1,…,xN←$Zq; r←
$Δ;
M1,…,MN←$G;
b←$0,1;
ρ0←r
N
xxMg N.....1
; ρ1←$G;
d←CN(M1||…||MN, Δ, rx
Mg 11
||…||r
N
xMg N
|| ρb);
output(b d);
Define CN‟s advantage )()(*
),,( kAdv NDDHPMG
CG N
=|P[)()(*
),,( kExp NDDHPMG
CG N
=1]-1/2| and its
upper-bound for all CN is notated as )()(*
),( kAdv NDDHPMG
G
. When empasizing time complexity,
we use the computation time t instead of k as the variable.
Lemma 6.1 r∈Δ、x1,..., xN∈Zq, r, x1,..., xN are unpublic random numbers which are
independent each other, M1,..., MN are public, R is random and DDHP is hard on Y, then 1x
g
||…||Nx
g ||Nxx
g.....1
PPT
1x
g ||…||Nx
g ||R and
rxMg 1
1
||…||r
N
xMg N
||r
N
xxMg N.....1
PPT
rx
Mg 11
||…||r
N
xMg N
||R
Concretely:
)()(* tAdv NDDHPG
G
≤))1(()1(* OtAdv NDDHPG
G
≤))(( NOtAdv DDHP
G
54
)()(*
),( tAdv NDDHPMG
G
≤))(()(* NOtAdv NDDHPG
G
≤))(( 2NOtAdv DDHP
G
Proof (1)We construct the solver BN-1 for G*-DDHP(N-1) problem from the solver BN for
G*-DDHP(N) problem. For the input V=v1||…||vN-1||ρ where vi=ix
g and ρ which is either
11...... Nxxg or a random element on G, BN-1 does:
xN←$Zq;
d←BN(v1||…||vN-1||Nx
g ||Nx
);
return(d);
Then)()(*
, tAdv NDDHPG
BG N
≤))1(()1(*
, 1OtAdv NDDHPG
BG N
and furthermore)()(* tAdv NDDHPG
G
≤
))1(()1(* OtAdv NDDHPG
G
. By induction on N we finally have )()(* tAdv NDDHPG
G
≤
))(( NOtAdv DDHP
G
(2) Let CN be the P.P.T. solver for G*M-DDHP(N) problem, from CN we construct the solver
BN for G*-DDHP(N) problem. For the input V=v1||…||vN||ρ where vi=ix
g and ρ is either
Nxxg
......1
or a random element on G, BN does:
M1,…, MN←$G; r←
$Δ;
d←CN (M1||…||MN,Δ,rMv 11 ||…||
r
NN Mv||
r
NM,);
return(d);
So)()(*
),( tAdv NDDHPMG
G
≤))(()(* NOtAdv NDDHPG
G
and by induction on N we get
)()(*
),( tAdv NDDHPMG
G
≤))(()(* NOtAdv NDDHPG
G
≤))(( 2NOtAdv DDHP
G .
Lemma 6.2 r∈Δ、x1,..., xN+1∈Zq, r、x1,..., xN+1 are unpublic random numbers which are
independent each other, M is public, R is random and DDHP is hard on Y, then
R||rxxxx
Mg NNj 11 ...ˆ...
:1≤j≤N-1... FDP
rx
Mg N 1
||rxxxx
Mg NNj 11 ...ˆ...
:1≤j≤N-1.
(hereafter we use the notation ||uj: 1≤j≤n instead of u1||...||un to avoid the lengthy expressions)
Proof Let h=11...... Nxx
g , xiyi=1 mod q, then 11 ...ˆ... Nj xxx
g =jy
h and 1Nx
g =Nyy
h...1
so
55
R||rxxx
Mg Nj 11 ...ˆ...
:1≤j≤N-1= R||ry
Mh j
:1≤j≤N-1
rxMg N 1
||rxxx
Mg Nj 11 ...ˆ...
:1≤j≤N-1=ryy
Mh N...1
||ry
Mh j
:1≤j≤N-1
Since ryy
Mh N...1
=ryyy
Mh NN )( 11...
and yN is probabilistically independent with y1,…,yN-1,
R||ry
Mh j
:1≤j≤N-1 has the same distribution as ryy
Mh N...1
|| ry
Mh j
:1≤j≤N-1, which ends
the proof.
Lemma 6.3 r∈Δ、x1,..., xN+1∈Zq, r、x1,..., xN+1 are unpublic random numbers which are
independent each other, M is public, R is random and DDHP is hard on Y, then
rxxx
Mg Nj 11 ...ˆ...
:1≤j≤N-1||RPPT
rxxx
Mg Nj 11 ...ˆ...
:1≤j≤N-1||rxxx
Mg NN 111...
Proof. Let D be the P.P.T. distinguisher for rxxx
Mg Nj 11 ...ˆ...
:1≤j≤N-1||rxxx
Mg NN 111...
and
rxxx
Mg Nj 11 ...ˆ...
:1≤j≤N-1||R, we construct the solver E for G-DDHP(N-1) problem(defined in
sec.3.1)from D. For G-DDHP(N-1)‟s given input η=z(j.i): i≤j≤N-1; ρ where z(j,i)=ij xxx
g...ˆ...1
and ρ is11...... Nxx
g or a random element on G, E does:
xN, xN+1←$Zq; M←
$G; r←
$Δ;
return(D(M,rxx
MNjz NN 1)1,( :1≤j≤N-1||
rxMN 1
));
Let)()(*
),( tAdv NDDHPG
G
be the upper-bound of the advantage to P.P.T.-distinguish
rxxx
Mg Nj 11 ...ˆ...
:1≤j≤N-1||rxxx
Mg NN 111...
and pwxxx
Mg Nj 11 ...ˆ...
:1≤j≤N-1||R, it‟s easy to get
that )()(*
),( tAdv NDDHPG
G
≤ ))(()1( NOtAdv NDDHPG
G
. Furthermore by theorem 3.1 we get
)()(*
),( tA d v NDDHPG
G
≤ O(N)))(( 3NOtAdv DDHP
G .
6.2 Party Password-based Authenticated Key-exchange Protocol SPAKE
Suppose two participants U1 and U2 share the password pw, H is a random oracle,
∑=(MKG,Tag,Vf) is a secure(i.e., unforgeable against chosen message attack or UF_CMA
forshort) message authentication code (MAC) scheme, f is a pesudo-random function. The
2-party key exchange protocol SPAKE between U1 and U2 is presented in Fig.1[7].
56
It‟s easy to see that K1=K2=gxy
and sk1=sk2=H(U1||U2||X*||Y*||pw||g
xy). Actuallly Fig.1 is an
enhanced version of SPAKE 2(SPAKE1 is the same as SPAKE2 except that pw is not explicitly
included in sk1 and sk2‟s derivation.) where, in comparision to the version in [7], we add a
message U1→U2：s, t1 and two MAC operations in order to make SPAKE have explicit
consistency property([7] only proved SPAKE‟s key-security property). Such enhancement
obviously doesn‟t impact SPAKE‟s security. According to [7] both SPAKE 1/2‟s security rely on
the computational Diffie-Hellman problem‟s hardness(SPAKE 1‟s security directly relies‟ on
S-PCCDHP‟s hardness, more details can refer to [7]‟s section 4 and 5).
It‟s not hard to prove the enhanced SPAKE‟s consistency property: Let X2*, Y1*, t1*, t2* be the
message components received by U2 and U1 respectively. The fact that both U1 and U2
normally terminate implies that Vf(Ka1, 2||s||X*||Y1*,t*2)=1 and Vf(Ka2, 1||s||X2*||Y*,t*1)=1.
Since X* and Y* are random and A corrupts neither U1 nor U2(so A doesn‟t known the key Ka1
and Ka2), it must be true that X2*=X* and Y1*=Y*(otherwise the attacker can successfully forge
at least one new message/authentication-code pair, i.e., (2||s||X*||Y1*,t*2) or (1||s||X2*||Y*,t*1),
for the MAC scheme ∑, contradicting ∑‟s UF_CMA security) therefore (sk1, U1,U2)=(sk2,
U1,U2) with a (negligible) exception probability upper-bounded by )(_ kAdv CMAUF
, i.e.,
),,(2/1 se
conspwGKS
SPAKE qqtExp
≤ )(_ tAdv CMAUF
where t、qe、qs are respectively the attacker‟s computation time and number of queries for the
oracle-Excute() and oracle-Send().
Note that from theoretical perspective in the (enhanced) protocol SPAKE it‟s unnecessary to
explicitly include the random session identifier s, because X* and Y* are already random
numbers so X*||Y* can play the same role as s. In practice, however, verifying the fact that s is
not repeated is usually much easier than verifying the same fact on X* and Y*(which are group
elements) so the protocol explicilty include the session id and we do the similiar thing in our
group protocol‟s design in the following.
57
6.3 Security Proof of Pw/GAKE
Theorem 4.1 Pw/GAKE is secure under the above conditions.
Proof At first we can get the consequence of consistency almost verbatimly following the
argument on SPAKE’s consistency at the end of section 6.2:
),,(/ se
conspwGKS
GAKEPw qqtExp
≤O(N) )(_ tAdv CMAUF
In the following we focus on proving Pw/GAKE‟s security. To emphasize the number of
members in the group, N, during the protocol‟s running we use the notation Pw/GAKE(N). The
critical point in the proof is to establish a relationship between Pw/GAKE(N+1)‟s security and
Pw/GAKE(N)‟s security.
Let AN+1=(AN+1,1,AN+1,2) be the P.P.T. attacker against Pw/GAKE(N+1), we construct an attacker
AN=(AN,1, AN,2) against Pw/GAKE(N) based-on A N+1. Let g(k) and m(k) be respectively the
upper-bound of number of groups attacked by AN+1 and the number of transactions in each
attack. Let M1, M2, ..., MN be public parameters acquired by AN. AN performs the following
steps 1)-5):
1)generate gid0←$
1,...,g(k), sid0←$
1,...,m(k)(i.e.,guessing AN+1,1‟s output gid*||sid*) and
xN+1←$Zq;
2)call AN+1 with parameters M1, M2,..., MN,1Nx
NMand simulate the environment
)()1(/ kExp indpwGKS
NGAKEPw
to interact with AN+1: if the transaction initiated by AN+1( via a request to
58
one of the environment‟s interfaces) is not that of gid0||sid
0,，AN creates group members U1,
U2, ..., creates a group and assigns its members with a shared password pw (a new password for
each new transaction) and runs Pw/GAKE(N+1) to response the request to)()1(/ kExp indpwGKS
NGAKEPw
from AN+1. Obviously such simulations for)()1(/ kExp indpwGKS
NGAKEPw
on these transactions are perfect.
Each time AN+1 queries H, AN queries it‟s own oracle H with the same parameters and responses
AN+1 with the returning values.
If AN+1 outpus gid*||sid*≠gid0||sid
0 then AN halts. Otherwiase AN outputs gid*||sid*(i.e.,
gid0||sid
0). Denote M(k)≡g(k)m(k) and note that P[gid*||sid*=gid
0||sid
0]≥1/M(k).
3)AN forwards the test-key(refer to sec.3.2) Zb provided by )()(/ kExp indpwGKS
NGAKEPw
to AN+1
and then proceeds.
4)When AN+1 initiates its request for Pw/GAKE(N+1)‟s transaction gid0||sid
0, AN initiates
the corresponding request for the oracle of its own(i.e., calls the same interface of
)()(/ kExp indpwGKS
NGAKEPw
with the same input). If the output ξ returned by the oracle is not
Pw/GAKE(N)‟s grpkey-message, then AN forwards ξ to A N+1 and then sends the message η
responsed by AN+1 to ξ‟s target; If ξ is Pw/GAKE(N)‟s grpkey-message, i.e., ξ=(grpkey,
gid0||sid
0, σ1||…||σN||ωj,N:1≤j≤N-1, tN), then AN randomly generates R0←
$G and sends AN+1
with the following message ξ1
0(to simualte the communication UN→UN+1):
ξ10= (init, gid
0||sid
0, σ1||…||σN||R0||ωj,N:1≤j≤N-1||λN-1)
where λN-1 is the component in the last init-message (init, gid0||sid
0, σ1||…||σN-1|| λN-1||
ωj,N-1:1≤j≤N-2||ωN-1,N-1) of Pw/GAKE(N) obtained by AN;
If AN+1 initiates the request for Pw/GAKE(N+1)‟s grpkey-message and ξ is the same message as
the above, then AN randomly generates R1,R2←$G and forwards the following message ξ
11 to
AN+1:
ξ11= (grpkey, gid
0||sid
0, σ1||…||σN||R1||
1
,Nx
Nj:1≤j≤N-1||R2, t)
5)At last A N outputs whatever is output by A N+1.
Without loss of generality, suppose )()1(/ kExp indpwGKS
NGAKEPw
does not halt so )()(/ kExp indpwGKS
NGAKEPw
doesn‟t halt either. In the event that all members participating the transaction gid0||sid
0 normally
terminate with the same key and gid0||sid
0 is not labeled as “disclosed”, it‟s obvious that
59
σi=pw
i
xMg i
, 1≤i≤N-1; ωj,N =pw
N
xxxMg Nj ...ˆ...1
:1≤j≤N-1
with the exception probability of O(N)/2L(k)
where L(k) is the key-size. Denote
)(!),1(/ kExp indpwGKS
ANGAKEPwsimulated N
‟s message flow as Ψ1 and the real
)(!),1(/ kExp indpwGKS
ANGAKEPw N
‟s
message flow as Ψ0. Summarize the above arguments, (from AN+1‟s perspective) the only
difference between Ψ0 and Ψ
1 is:
The counterpart of ξ1
0 in Ψ0 is ξ
00 = (init, gid
0||sid
0,
pwxMg 1
1
||…||pw
N
xMg N
||pw
N
xxMg N.....1
||
pw
N
xxxMg
Nj ...ˆ...1
:1≤j≤N-1||pw
N
xxMg N
1
..... 11
);
The counterpart of ξ11 in Ψ
0 is ξ
01 = (grpkey, gid
0||sid
0,
pwxMg 1
1
||…||pw
N
xMg N
||pw
N
xMg N
11
||pw
N
xxxMg Nj
1
...ˆ... 11
:1≤j≤N-1||pw
N
xxxMg NN
1
... 111
, t) where MN+1=1Nx
NM;
All other corresponding messages in Ψ0 and Ψ
1 have the same distribution.
For further analysis, we need the following lemma which indicates that, from AN+1‟s
perspective, AN‟s test-key Zb is P.P.T. indistinguishable from AN+1‟s test-key.
Lemma 6.4 H is a random oracle, pw∈Δ, x1,..., xN+1∈Zq, pw, x1,..., xN+1 are all independently
random and unknown, g, M1,..., MN+1,U1,...,UN+1,pwx
Mg 11
,…, pw
N
xMg N
11
are public(butNxx
g...1
,
11... Nxxg are unknown), DDHP is hard for Y, then
H(U1||…||UN||pwx
Mg 11
||…||pw
N
xMg N
||pw||Nxx
g...1
)
is P.P.T. indistinguishable from
H(U1||…||UN+1||pwx
Mg 11
||…||pw
N
xMg N
11
||pw||11... Nxx
g )
Proof Denote h0≡H(U1||…||UN||pwx
Mg 11
||…||pw
N
xMg N
||pw||Nxx
g...1
), h1≡H(U1||…||UN||UN+1||
pwxMg 1
1
||…||pw
N
xMg N
||pw
N
xMg N
11
||pw||11... Nxx
g ), h*N≡pwx
Mg 11
||…||pw
N
xMg N
|| pw||Nxx
g...1
, h*R≡
pwxMg 1
1
||…||pw
N
xMg N
||pw||R where R is an arbitrary random number. Denote the upper-bound
of the advantage of the P.P.T.-distinguisher for h0 and h1 as )(, tAdv indSK
MH
and the upper-bound
60
of the advantage of the P.P.T.-distinguisher for h*N and h*R as )(tAdv indSK
H
where t is the
computation time. The proof consists of the following two steps:
(1)Claim: Under the above conditions we have h*N
PPT
h*R .
Let D be a P.P.T.-distinguisher for h*N and h*R, we construct a P.P.T. distinguisher E via D to
distinguish 1x
g ||…||Nx
g ||Nxx
g...1
and 1x
g ||…||Nx
g ||R. i.e., E efficiently solves the
G*-DDHP(N) problem(sec.6.1). For the given input (1x
g ,…,Nx
g , ρb) where b∈0,1, E
performs:
randomly generate r, M1,...,MN;
return(D(rx
Mg 11
||…||r
N
xMg N
||r||ρb));
Obviously)(, tAdv indSK
DH
≤))(()(*
, NOtAdv NDDHPG
EG
and furthermore by lemma 6.1 we have
)(tAdv indSK
H
≤))(()(* NOtAdv NDDHPG
G
≤))(( 2NOtAdv DDHP
G . As a result h*N
PPT
h*R .
(2) By the above fact we get
h0 = H(U1||…||UN||pwx
Mg 11
||…||pw
N
xMg N
||pw||Nxx
g...1
)
PPT
H(U1||…||UN||pwx
Mg 11
||…||pw
N
xMg N
||pw||R)
H(U1||…||UN||UN+1||pwx
Mg 11
||…||pw
N
xMg N
11
||pw||R)
PPT
H(U1||…||UN||UN+1||pwx
Mg 11
||…||pw
N
xMg N
11
||pw||11... Nxx
g ) = h1
and it‟s easy to derive the inequality
)(, tAdv indSK
MH
≤ 2 ))1(( OtAdv indSK
H
≤ 2))(( 2
, NOtAdv DDHP
EG
which ends lemma 6.4‟s proof.
Now by lemma 6.1‟s second part we get
pwxMg 1
1
||…||pw
N
xMg N
||pw
N
xxMg N.....1
||pw
N
xxxMg
Nj ...ˆ...1
:1≤j≤N-1||pw
N
xxMg N
1
..... 11
)
PPT
pwx
Mg 11
||…||pw
N
xMg N
||R0||pw
N
xxxMg
Nj ...ˆ...1
:1≤j≤N-1||pw
N
xxMg N
1
..... 11
)
61
i.e., ξ10
PPT
ξ0
0 .
Furthermore by lemma 6.2-6.3 we get
pwxMg 1
1
||…||pw
N
xMg N
|| R1||pw
N
xxxMg Nj
1
...ˆ... 11
:1≤j≤N-1|| R2
PPT
pwx
Mg 11
||…||pw
N
xMg N
||pw
N
xMg N
11
||pw
N
xxxMg Nj
1
...ˆ... 11
:1≤j≤N-1|| R2
PPT
pwx
Mg 11
||…||pw
N
xMg N
||pw
N
xMg N
11
||pw
N
xxxMg Nj
1
...ˆ... 11
:1≤j≤N-1||pw
N
xxxMg NN
1
... 111
i.e., ξ11
PPT
ξ0
1. Combining all the above analysis we get
)(!),1(/ kExp indpwGKS
ANGAKEPwsimulated N
PPT
)(!),1(/ kExp indpwGKS
ANGAKEPw N
and concretely
| P[)(
!),1(/ kExp indpwGKS
ANGAKEPwsimulated N
=1]-P[)(
!),1(/ kExp indpwGKS
ANGAKEPw N
=1]|
≤ O(N)))(( 3NOtAdv DDHP
G + O(N)/2
L(k), so
)(),(/ tAdv indpwGKS
ANGAKEPw N
= | P[)(),(/ tAdv indpwGKS
ANGAKEPw N
=1]-1/2 |
≥ (1/M(k))| P[)(
!),1(/ kExp indpwGKS
ANGAKEPwsimulated N
=1]-1/2 |
≥ (1/M(k))| P[)(
1),1(/ tAdv indpwGKS
ANGAKEPw N
=1]-1/2 |
–(1/M(k))| P[)(
!),1(/ kExp indpwGKS
ANGAKEPwsimulated N
=1]- P[)(
1),1(/ tAdv indpwGKS
ANGAKEPw N
=1]|
≥ (1/M(k))()(
1),1(/ tAdv indpwGKS
ANGAKEPw N
–O(N)))(( 3NOtAdv DDHP
G - O(N)/2
L(k) )
equally
)()1(/ tAdv indpwGKS
NGAKEPw
≤ M(k))()(/ tAdv indpwGKS
NGAKEPw
+ O(N)))(( 3NOtAdv DDHP
G +O(N)/2
L(k)
This is the relationship between the security of Pw/GAKE(N+1) and Pw/GAKE(N). By
recursion on N upto N=2, we get the desired relationship bewteen the security of
Pw/GAKE(N+1) and SPAKE:
)()1(/ tA d v i n dp w G K S
NG A K EPw
≤
62
M(k)N-1 )()2(/ tAdv indpwGKS
GAKEPw
+))((()( 3
21 jNOtAdvNkM DDHP
G
j
Nj + +O(N)/2L(k)
)
≤ M(k)N-1 )(tAdv indpwGKS
SPAKE
+ O(M(k)N-1
N)()))((( 3 kMNOtAdv DDHP
G +1/2
L(k) )
It is already proved in [7] that SPAKE 1/2 are secure under CDHP‟s hardness hypothesis, in
particular, SPAKE is secure under DDHP‟s hardness hypothesis. As a result, for any constant N
the right-hand-side of the above inequality is always negligible in k, which proves
Pw/GAKE(N)‟s security.
Note1: Theorem 4.1 requires DDHP‟s hardness on Y=(G,g)q(k), a stronger condition than that
of the original protocol SPAKE but still a widely used and practical hypothesis.
Note2: In the full version paper concrete efficiency analysis is done for the protocol‟s
instantiations with some specific MAC schemes and DDHP-hard groups.
63
SICASE-388
The Newest Double-Population Particle Swarm Optimization
Wenjing Liu
Inner Mongolia University of Science & Technology, China
Jianguo Wang*
Inner Mongolia University of Science & Technology China
Zhijie Zhang
Inner Mongolia University of Science & Technology, China
Da Lv
Bao Tou College, China
Abstract
In the iteration process of the particle swarm optimization algorithm, there exist the poor
diversity of population and premature phenomenon. In order to increase the diversity of
population, a new double-population particle swarm optimization (DP-PSO) that based on the
cooperative double-population is proposed. The population will be divided into two parts: local
optimization and global optimization (add the characteristic of the local optimization). For
global optimization, to be using a self-adapting strategy to adjust the accelerating factors which
ensure the swarms to avoid the local extreme and improve the local and global search ability.
Experiments show that the proposed DP-PSO method has been significantly improved in both
convergence speed and computational accuracy.
Keywords：Particle swarm optimization；Double-population；Dynamic adjust； Accelerating
factor
1. Introduction
Particle swarm optimization (PSO) [1] algorithm is proposed in 1995 by James Kennedy and
Russell Eberhart, the basic idea is influenced by the research on the behavior of birds and they
used the model of biological group by the biologists Frank Heppner. Due to the method is
simple and easy to implement, the less parameters and it is not dependent on the issues such as
information characteristics , currently PSO has achieved to success in the applications such as
64
the function optimization, engineering design and optimization, electric power system [2] , the
robot control [3] and other fields.
However, PSO is also easy to be trapped in local optimum and it has the slow convergence
speed in the late stage and the weak accuracy. Therefore, scholars have proposed many
improved methods, for example Shi and Eberhart introduce the inertia weight w in PSO and
they proposed using an adaptive inertial weight to balance the global convergence and the
convergence speed in the evolutionary process, this evolution equation was called standard
PSO algorithm by scholars [4], [5]. Inertial weight w make particle keep moving, make it have
the trend of expanding searching space, and have the ability to explore new areas. A larger w
has the global convergence ability, and a smaller w has a strong ability of local convergence.
Along with the increase of number of iterations, inertial weight should constantly decrease, so
that particle swarm algorithm has a strong ability of global convergence in initial stage, and
later it has a strong ability of local convergence. Therefore, Shi put forward an adjusting
strategy that uses a linearly decreasing inertia weight, inertial weight should meet:
Where Max-number is the truncated generation, t is the current iteration and w is decreasing
from 0.9 to 0.4. For experimental results of the test, the effect is good. Also, there are many
other improvement methods about w[6],[7]. Clerc introduces a convergence factor in the
update formula to control the trend of convergence in the algorithm, the experimental results
show that this method has better convergence, but for some functions，it cannot reach the
global optimal in the prescriptive iteration times [8]. In the PSO algorithm, c1 and c2 are used to
adjust the particle, s best position and its companions
, best position to influence the movement
of particles. Ratnaweera, A and Halgamuge, S. K. apply linearly decreasing method to adjust
the accelerating factor for the balance of local and global optimization [9]. Then, according to
the change trend of c1 and c2 in different periods, Lin. M. uses cosine function to adjust
accelerating factor [10].The experimental results show that this method has obtained the certain
effect.
In this paper, to consider the influence of the population diversity and the accelerating factors
in the algorithm, an improved particle swarm optimization based on a double-population is
proposed and this paper uses a self-adapting strategy to adjust the accelerating factors in global
optimization. The experimental results show that the proposed DP-PSO is stable and effective;
it performs high accuracy and fast convergence.
65
2. The Standard Particle Swarm Optimization
Particle swarm optimization algorithm is similar to other evolutionary algorithms, it also uses
the concept of group and evolution, and it performs operation on the basis of the fitness value
of the individual (particles). PSO algorithm regards every individual as a volume-less particle
(a point) in the n-dimensional search space, every particle has a certain velocity in this space,
while in PSO each particle (individual) is updated according to its own flying experience and
the flying experience of group. The ith particle is represented as Xi=(xi1,xi2,… …,xin).
Vi=(vi1,vi2,… …,vin) is the velocity of the ith particle. The best previous position (the position
giving the best fitness value) of the ith particle is recorded and represented as Pi=
(pi1,pi2,… …,pin). The index of the best particle among all the particles in the population is
represented by the symbol g. The particles are manipulated according to the following
equation:
where w is the inertia weight, c1 and c2 are two accelerating factors, and r1 and r2 are two
random functions in the range [0, 1]. All particles in PSO are kept as members of the
population through the course of the run (a run is defined as the total number of generations of
the evolutionary algorithms prior to termination). It is the velocity of the particle which is
updated according to its own previous best position and the previous best position of its
companions. The particles fly with the updated velocities [11].
3. New Double-population Particle Swarm Optimization
3.1 Improved Accelerating Factors
Accelerating factors c1 and c2 represent the weight of acceleration which will pull the particles
to the local optimal location and the global optimal location. Through the research, it founds
that algorithm in ideal condition, all the particles should be as much as possible to fly over the
whole search space, so that it can increase the diversity of particles in the initial stage, and in
the end of the search, the particles with fast velocities and accuracy converge to the global
optimal solution. In this paper we analyze the influence of the accelerating factors to the
algorithm, we propose a strategy which uses arctan function to adjust the accelerating factors to
make the balance between global search and local search. The accelerating factors equation of
c1 and c2 should meet:
66
Where, c1start and c2start are initial value of c1 and c2, c1end and c2end are final value of c1 and c2,
Tmax is the maximum iterating times, e is adjustment coefficient which control the decay of
the curve. Here we let Tmax=1000, c1start=2.75, c1end=1.25, c2start=0.5, c2end=2.25, e=6and make
comparisons with comparing LPSO、CPSO. The results are depicted in Fig.1.
The whole optimization process of DP-PSO is depicted as follows:
In the initial stage, c1 should be larger so that it is easy to find the optimal value and the
decreasing speed of c1 should be fast in order to improve the convergence speed; and c2 should
be smaller in order to reduce the influence of other particles and improve the global search
ability.
In the late stage, the decreasing speed of c1 should slow down to improve stability; and c2
should be gradually increased so that it can improve the ability of social information sharing
and improve the local searching [12].
3.2 The Strategy of Population Division
When the particle swarm optimization algorithm was used, all particles are following the
global optimal individual to update their positions in the iteration process, that would make the
particles gather around a point, so that the diversity of population fall rapidly, it causes
premature phenomenon. In order to increase the diversity of population, we apply cooperative
double-population of coevolution, where the population will be divided into two populations
and then use two different evolution models. The detailed strategy of population division is
described as follows:
67
Firstly, it supposes the population has N particles, the population is composed of two parts: Q1
(local optimization) and Q2 (global optimization), Q1 has n1 particles and Q2 has n2 particles,
N=n1+n2; and then use different evolution model for the two parts respectively, where the
population of Q1 uses the equation of the rapid convergence to improve the local optimization
ability; The population of Q1 uses the equation which has global search ability; finally, we use
the equations (4) and (5) to adjust the accelerating factor. The evolution equations are depicted
as follows:
The inertia weight w is employed to control the impact of the previous velocity on current
velocity, and influence the balance between global and local exploration abilities. The role of
Q1 is quickly searching in local area. A smaller inertia weight tends to facilitate local
exploration to fine-tune the current search area, and it has strong local convergence ability, its
task is searching in the area between the current global best position and individual best
position (nearby the global best position), where accelerating the convergence; The role of Q2
is global search around a wide range, when the particles find the new best position, through the
information exchange between particles, Q1 will be guided to the new best position. Due to the
group cooperation, it is not only increasing the diversity of population, but also improving the
efficiency of optimization.
4. Simulation
In this paper, we use three classic functions: Sphere, Rastrigrin and Griewank to test the
standard particle swarm optimization (PSO), LPSO, CPSO and DP-PSO and make
comparisons, the detailed results are depicted in Table 1.
Parameters Setting: the population size N = 30, the maximum iterating times Tmax = 1000, the
dimension is 20, for the single-peak function Sphere function, in the evolution equation
accelerating factor take c1 = c2 = 2, Rastrigrin and Griewank function use equations (4) and (5),
c1start=2.75，c1end=1.25，c2start=0.5，c2end=2.25, adjustment coefficient e = 6, all algorithms are
using linear self-adapting inertia diminishing strategy, every test runs 50 times, taking the
minimum, the maximum and the average as test results.
68
Table 2 gives the optimal results of three test functions for four algorithms. Table 1 data show
that, for Griewank function, the optimization precision of these methods is very close, it shows
that these methods have the strong global optimization ability. Performing in Sphere (one-peak
function) and Rastrigrin (multi-modal function), the optimization precision of DP-PSO is better
than other methods, it illustrates that DP-PSO do not only have the strong global optimization
ability, but also have a better local optimization ability. For one-peak function, the better
performance of DP-PSO is due to the contribution of the convergence of group Q1. For
multi-modal function, the performance of DP-PSO benefits from the global search of group Q2.
Fig.2 are the curves which fitness value change following the iteration times, in the Fig.2, we
can see from the graphs that DP-PSO has higher optimizing efficiency and accuracy than other
methods.
69
5. Conclusion
PSO algorithm is a kind of high efficient and novel intelligent optimization algorithm, and its
operation is simple, the performance is well, also it need further improvements. This paper
proposes a strategy based on the cooperative double-population of co-evolution, and we use a
self-adapting strategy to adjust the accelerating factors. In the strategy of dividing populations,
although the evolution equation of each group is generally particle swarm evolution equation
(only different parameter selection), but different parameters settings increase the diversity of
the group, make the particles have defined roles, and improve the searching efficiency.
Through the tests of the classical functions, it shows that the performance of DP-PSO is
evidently better than other PSO algorithms.
6. Acknowledgment
Project supported by the National Natural Science Foundation of China (Grant No.21366017 ) ,
Inner Mongolia department of application and research and development (Grant No.20130302)
and Inner Mongolia University of Science and Technology Innovative Start-up Foundation
(Grant No.2014QDL024).
7. References
1. Kennedy, J., Eberhart, R.C.: Particle swarm optimization. Proceeding of IEEE International
Conference on Neural Networks, 1942-1948 (1995)
2. Zeng J., Liu Z., He S., Yang H.: Study of the reactive power optimization based on
sub-region particle swarm algorithm. Power System Protection and Control, Vol.40, No.1,
Jan.1 (2012)
3. Liu L., Cai Z.: Multi-robot cooperative localization based on particle swarm optimization.
Journal of Central South University (Science and Technology), Vol.42, No.3, Mar (2011)
4. Shi, Y.H., Eberhart, R.C.: Modified particle swarm optimizer. Proceedings of the IEEE
Conference on Evolutionary Computation, ICEC, p 69-73 (1998)
5. Shi, Y.H., Eberhart, R.C.: Parameter selection in particle swarm optimizer. IEEE
70
International Conference on Evolutionary Computation. Alaska, May 4-9 (1998)
6. Cheng G., Li M., Zhang X.: Particle swarm optimization algorithm based on local reach for
inertia weight. Application Research of Computers, Vol.28, No.3, Mar (2011)
7. Zhao Y., Xu J.: Modified particle swarm optimization algorithm. Computer Engineering and
Applications, 46(2):31-33(2010)
8. Clerc, M., Kennedy, J.: The Particle Swarm-explosion, Stability, and Convergence in a
Multidimensional Complex Space. IEEE Trans. on Evolutionary Computation, 6(1): 58-73
(2002)
9. Ratnaweera, A., Halgamuge, S.K.: Self-organizing hierarchical particle swarm optimizer
with time-varying acceleration coefficients. Evolutionary Computation, 8(3): 240-255
(2004)
10. Lin M., Liu F., Jiang X.: A particle swarm optimization method with dynamic adjustment
accelerating factor. Computer Knowledge and Technology, Vol.5, No.34, December, pp.
9816-9818 (2009)
11. Zeng J., Jie J., Cui Z.: Particle swarm optimization. Beijing Science Press (2004)
12. Chen H., Fan Y., Deng S.: A dynamic accelerating factor adaptive particle swarm
optimization algorithm. Journal of China University of Petroleum (Edition of Social
Sciences), 34 (6):173-176 (2010)
71
SICASE-776
Learning Issues in Opportunistic Software Development
Yonglei Tao
School of Computing and Information Systems
Grand Valley State University, USA
Abstract
Opportunistic software development is a method of software development in which developers
meld together software pieces they found, rather than starting from scratch. It is a practical
method for problem solving and also offers great educational value. Using opportunistic
software development in a classroom setting exposes students to programming tasks they are
not prepared for. Students must learn what they need to know at the time they need to use and
effective learning is crucial for successful completion of their tasks. As such, it creates a
realistic educational context for students to pursue self-directed learning. However, one cannot
assume that desired results simply happen when students are placed in such a situation. Our
experience shows that careful guidance from the instructor is indispensible to ensure students
meet their learning objectives in an orderly manner.
Keyword: Opportunistic Software Development, Self-Directed Learning, Computer Science
Education
1. Introduction
As technologies evolve, software developers in the future are more likely to work with systems
they cannot possibly develop from scratch or use diverse tools that they were never taught in
their former education. It is important that computer science educators make it possible for
students to gain experience in using, modifying, and combining software components available
on the Web [1].
Opportunistic software development is a method of software development in which developers
meld together software pieces they found, rather than starting from scratch. Most often the
developer finds unrelated software components that weren‟t designed to work together but
provide functionality needed in a new system [5]. Opportunistic software development
emphasizes speed and ease of development over code robustness and maintainability; therefore,
it is useful for writing code to prototype, ideate, and discover [2]. Opportunistic software
development is reported in the literature as being effective for problem solving [3] and also
having great educational value [4, 13].
72
The challenge for computer science educators is twofold. They must teach a significant level of
specialized, rapidly evolving knowledge that often requires quick and cost-effective updates.
At the same time, they strive to improve the quality of students‟ learning experience [9]. Using
opportunistic software development principles in computer science education not only
encourages students to develop solutions that cross the boundaries of different technologies [4],
but also stimulates students to make a conscious effort to learn what they need to use [13]. As a
consequence, it creates a realistic educational context for computer science educators to meet
the challenge in a balanced way.
In this paper, we describe our approach to guiding students through opportunistic software
development for non-trivial class projects. Opportunistic software development provides a
unique opportunity for students to pursue self-directed learning. But one cannot assume that
desired learning objectives simply happen when students are placed in such a situation. Our
experience shows that careful guidance from the instructor is still indispensible, especially
when students are exposed to tasks they were not fully prepared for.
This paper is organized as follows. Section 2 introduces related work. Section 3 discusses an
approach to guiding students in opportunistic software development and section 4 describes our
experience with using it for programming projects in two courses. Section 5 summarizes the
results of a survey taken by participating students. Finally section 6 concludes this paper.
2. Related Work
Learning is the central issue in education. Research has shown that most learning in the
workplace takes place outside of formal training and, given the swiftly changing nature of the
field, computer science graduates, more than most workers, need to be able to learn new topics
outside of organized classes [7]. As such, computer science educators have devoted significant
attention to enhance students‟ learning skills. A large number of reports about learning issues
in computer science education can be found in the literature. Many of them focus on
encouraging students to participate in self-directed learning [14-16].
Self-directed learning or active learning is known as a process by which individuals take the
initiative, with or without the assistance of others, in identifying their learning needs, locating
resources for learning, choosing appropriate learning strategies, and evaluating learning
outcomes [6]. As an instructional method, self-directed learning allows engagement in the
learning process and requires students to think about and be actively involved in meaningful
learning activities [10].
Students are highly motivated when pursuing self-directed learning. Self-directed learning is
goal oriented in which students view problems as challenges, not obstacles [12].
73
Studies have shown that students can also master difficult subjects more effectively via
self-directed learning activities [11].
Student leaning can be improved when instructors move from traditional lecture-style
instruction to a student-centered, interactive instruction. In such a situation, the teacher
becomes a facilitator, guiding and coaching, rather than directing and lecturing, students learn
by doing [8]. Clearly, self-directed learning does not relieve educators of assisting students in
the learning process; instead it requires an innovative way to meet the challenge.
3. Our Approach
As computer science instructors, we tend to teach students to be proficient in a particular
technology before letting them to apply the knowledge in a non-trivial project. In opportunistic
software development, however, students must face tasks they are not fully prepared for.
Students have to analyze the needs of a new system, search the Web for relevant components to
meet their needs, and adapt or modify the selected components to build the intended system. In
this process, students are often exposed to technologies they have never been taught; they must
learn what they need to know at the time they need to use it.
In opportunistic software development, functional requirements determine development tasks
which in turn determine the learning objectives for students. Clearly, learning is not random; it
depends on one‟s view about what to do and how to proceed in the development process.
We consider a two-phase process as critical to ensure students meet their learning objectives in
an orderly manner.
(1) Breakdown system requirements to identify user tasks
(2) Develop each task opportunistically and incrementally
A task is an activity that the user performs to achieve a certain goal. It describes a small piece
of functionality for the intended application. Since students are new in the subject area,
guidance from the instructor is necessary. We use a Web application project as an example to
describe our approach below.
A web application is often viewed as consisting of several layers built with different
programming tools, for example, HTML, PHP, and MySQL. Traditional teaching methods
attempt to familiarize students with each of these tools in isolation before giving them an
opportunity to use. While appropriate in a rigorous web programming course at the junior or
senior level, doing so would divert too much attention from the primary topics in courses that
are not entirely devoted to web programming.
74
Alternatively we consider a web application as a collection of partitions with each crosscutting
the layers of the application to fulfill a user task, for example, login, form filling, data
retrieving, and file uploading. Each partition is a functional unit that can be created, tested, and
run independently. And its limited functionality leads to a short development cycle.
Incremental development and rapid iteration allow students to focus on a small number of new
concepts in each development cycle and also give them the ability to verify each unit of
functionality as soon as it is built.
Our intention is to have students learn web programming as they develop a practical web
application. Students work on various partitions over a number of development cycles. At the
beginning of each cycle, we help students find a small example to illustrate key programming
concepts in an application-independent manner. Students then build an initial version of the
partition under consideration according to the given example. Such a version is executable but
accomplishes only partial functionality. Students must explore further to find what they need to
do and how they can accomplish their objectives. Often they search the Web for pieces of code
with the desired functionality, copy them to their programs, and modify them to meet the
requirements. In this process, knowledge is partially and incrementally obtained as
opportunities – the needs to meet requirements and the availability of relevant materials – arise.
Successful discovery and use of new knowledge allow students to gradually extend their
programs to completion.
4. Experiment and Observation
We have experimented with using opportunistic software development for programming
projects in two different classes. One project was to build a Web application to support the
coop program of the engineering school in our university, which involved only a small group
of sophomore students. And the other one was to develop a multimodal user interface for an
online ticketing application. It involved fifteen students in a graduate class.
4.1 A Web Application Project
Here we illustrate our approach through activities in creating a login form. Since it was the first
task for students to do in this project, we had more involvement in the preparatory work than
that for other tasks.
We began with introducing to students a simple login form that was made of an HTML
document and a PHP script with limited functionality, that is, validating the user input against a
hardcoded data. We also discussed relevant programming concepts and demonstrated how to
run the login form from a browser.
75
Then we guided students through the current development cycle in a stepwise manner. We
asked students to perform adequate testing on each version of their programs before moving on
to the next step. Given below are steps that we guided them to take for building a login form:
Creating an application-specific login form on the basis of the given example.
Adding code to handle a forgotten password; that is, allowing the user to submit an email
address and receive a valid password via email from the system administrator.
Adding code to validate the user input using the information in a password table of a
MySQL database.
Students wrote code via cut-and-paste and in-place modification from pieces of code they
found on the Web. In this process, what they learned is more than programming concepts. For
example, a student combined normal password processing and forgotten password handling
into one PHP script, but found the default button behaved inconsistently in different browsers.
Solving that problem relied on certain system-specific knowledge that students could never be
prepared for. Ultimately, a solution was found with the instructor‟s assistance. Clearly, it is a
realistic scenario and working through such scenarios helps students develop problem-solving
skills.
Students could produce surprising results at times. For example, a student found a piece of
code with a JavaScript statement to bring up a message box for error notification. He
successfully integrated it into his PHP code, making his program more robust, even though
using any programming tool beyond HTML, PHP, and MySQL was not expected. Such a case
is not unusual when students engage in opportunistic learning.
4.2 A Multimodal User Interface Design Project
In this project, a multimodal user interface was defined as one allowing the user to interact with
an application via either touch or voice. Ideally, it should be a cohesive integration of a GUI
(Graphical User Interface) and a VUI (Voice User Interface).
Students had to face challenges at multiple levels. A multimodal user interface should provide
the required functionality and also guidance for the user to achieve their goals. While students
understood the visual design for a user interface, they were not familiar with issues and
techniques of voice interface design. In addition, they never had any exposure to speech
recognition technologies in their previous courses.
A speech recognition engine is available in almost every programming platform, including
Microsoft Visual Studio, Java (such as CMU Sphinx), Android, as well as Google Web Toolkit.
Since students had different programming backgrounds and preferences, we allowed them to
76
choose the tools they wanted to use. We also provided an overview on various speech
recognition technologies to assist them in making an appropriate decision.
At the beginning, students conducted a functional decomposition to identify user tasks that the
intended user interface must support. Due to the complexity of this project, students selected a
relatively simple task to begin with and proceeded to implement it according to opportunistic
software development principles.
We asked students to develop this project by incorporating voice-activated control into a
graphical user interface, rather than building two user interfaces independently. As such,
students took three steps to achieve the goals for each task:
Develop a graphical user interface
Add voice-activated control
Address issues of integration between touch and voice capabilities
Code examples about how to use speech recognition engine can be found on the Web. Often
they include very limited functionality, for example, two or three screen elements and a simple
grammar for speech recognition. Nonetheless, they demonstrate the basic structure of such a
program. Students searched the Web to find an appropriate example and modified it
incrementally to achieve the required functionality. Obviously, they must obtain a thorough
understanding on the existing code before making any modification. Hence, students had to
learn what they needed to know, including concepts and techniques for speech recognition,
natural language understanding, and dialog management.
Students met diverse difficulties, especially with implementing voice-activation control and
resolving integration issues, such as how to design a grammar to reduce ambiguity, how to
modify configuration files to adjust the confidence level for speech recognition, and how to
handle errors of various categories. Difficulties they had often motivated them to learn new
knowledge, exchange ideas with their classmates, and work out creative solutions.
5. Student Feedback
We conducted an evaluation survey to collect feedback from students about their experience
with opportunistic software development.
Many students considered the learning experience they had as the most important benefit.
Students felt that they had learned more than they would do in the conventional way. Listed
below are a few comments from students:
77
“Building this application was a learning experience in itself.”
“By figuring out on our own, we learn more because we had to figure it out each time we
wanted to implement something new, rather than just coping out notes that were used in class.”
“When I first started I had a brief idea what needed to be accomplished, but honestly I had
minimal experience with website and database applications. I think learning these applications
in a work atmosphere helped me way more in the long run.”
Students also considered as beneficial for them to choose technologies that they wanted to use,
even though those technologies were not previously taught. As a result, students were highly
motivated to acquire and apply new knowledge as illustrated by comments below:
“Working with new techniques provided us with a great benefit because we were able to
use what we wanted to, and not what someone had taught us.”
“It was my first voice user interface I ever did. It gave me a great experience with speech
recognition technology through challenges I had.”
On the other hand, it did take a considerable amount of time for students to learn a new
technology. Due to the lack of a broad view about existing technologies, some of them did not
consider that they had made the best choice for themselves at the beginning. Given below are
some comments in this regard.
“We learned more while doing research, but we also used valuable time that could have
been better spent on making our application the best.”
“I wish I had looked into other options more carefully before making the decision (about
which technology to use).”
In addition, students were proud of what they had accomplished. Some of them posted their
projects on the Internet to show what they were able to do.
Clearly, it is beneficial for students to follow the opportunistic software development principles
when developing a practical project. However, it is a challenge for the instructor to help
students learn what they need to use without diverting too much attention from the primary
course objectives.
6. Summary
In most software engineering methods, programming activities are viewed as being systematic,
starting from high level goals that are in turn decomposed into achievable actions via a
successive refinement process. Deviating from conventional approaches, several recent studies
emphasize the opportunistic nature of the programming process [2-5].
78
Our experience indicates that software development often involves a mixture of top-down and
opportunistic activities. As described above, a key element of our approach is to identify user
tasks. Using a top-down strategy first not only helps students cope with software complexity,
but also creates a foundation for them to develop their programs in an opportunistic manner. A
strictly opportunistic process is neither realistic nor feasible in a classroom setting.
Opportunistic software development encourages just-in-time learning of new knowledge and
skills [2]. It makes it less compulsory for the instructor to motivate students. As a result, the
focus of teaching shifts from preparing students for what they must know to guiding them
through the development process. Since students have to explore an unfamiliar domain
themselves, they may run into a wide range of obstacles. Using an opportunistic approach
requires greater attention to individual students, which may be problematic for a large class.
Learning is the central issue in education. Computer science instructors are especially
concerned with students‟ learning skills due the rapidly changing nature of this field. Our initial
use of an opportunistic approach has proven to be a positive learning experience for students.
We plan to continue our effort of incorporating this approach into our curriculum. We also plan
to investigate several ethical and pedagogical issues that may arise when using this approach,
such as whether it is appropriate to use code found from the Web for personal or commercial
purposes, and what adverse effects it may have on students‟ programming habits.
7. References
[1] Cohen, J.: Updating Computer Science Education, Communications of ACM. Vol. 48, No. 6,
June 2005, PP.29-31.
[2] Brandt, J, et al.: Two Studies of Opportunistic Programming: Interleaving Web Foraging,
Learning, and Writing Code. Proceedings of ACM CHI „09, Boston, Massachusetts, April
2009, pp 1589-1598.
[3] Robillard, P.: Opportunistic Problem Solving in Software Engineering. IEEE Software,
November/December 2005, pp. 60-67.
[4] Obrenovic, Z., et al.: Stimulating Creativity through Opportunistic Software Development.
IEEE Software, November/December 2008, pp. 64-70.
[5] Ncube, C., et al.: Opportunistic Software System Development: Making Systems from
what‟s Available. IEEE Software, November/December 2008, pp. 38-41.
[6] Knowles, M.: Self-Directed Learning. Follet, Chicago, 1975. p. 18.
[7] Boustedt, J., et al.: Students‟ Perception of the Differences between Formal and Informal
Learning. ICER ‟11. Aug. 8-9, 2011, Providence, Rhode Island, USA, pp. 61-68.
[8] Towhidnejad, M., et al.: Transforming Engineering and Science Education through Active
Learning. Proceedings of IEEE Frontiers in Education (FIE), 2014, pp.859-863.
79
[9] Longstreet, C. S. and Copper, K.: Curriculum Design Factors for Constructing Serious
Educational Game: a Learning Objective Centric Approach. Proceedings of Computer
Games: AI, Animation, Mobile, Multimedia, Educational and Serious Games (CGAMES),
2014, IEEE.
[10] Zepke, N. and Leach, L.: Improving Student Engagement: Ten Proposals for Action.
Active Learning in Higher Education, 11:167-177, 2010.
[11] Tsai, F. S., et al.: From Boxes to Bees: Active Learning in Freshman Calculus.
Proceedings of Global Engineering Education Conference (EDUCON), 2013, Berlin,
Germany, IEEE, pp.59-68.
[12] Swart, A. J. and Sutherland, T.: Co-operative Learning versus Self-directed Learning in
Engineering: Student Preferences and Implications. Proceedings of IEEE Frontiers in
Education (FIE), 2014, pp. 1466-1470.
[13] Tao, Y. and Plouff, C.: An Opportunistic Approach for Learning Web Programming.
Proceedings of IEEE FIE ‟10 (the 40th annual Frontiers In Education conference),
Arlington, Virginia, Oct. 27-30, 2010.
[14] Stephen, F. and Ruben, R.: Constructive Pedagogy Meets Agile Development - a Case
Study. Proceedings of 2010 Seventh International Conference on Information Technology,
IEEE Computer Society, April 12-14 2010, Las Vegas, Nevada, USA.
[15] Edward, G. and Carolyn, M.: Student-Generated Active-Learning Exercises. Proceedings
of ACM SIGCSE ‟09, March 3-7, Chattanooga, Tennessee, USA, pp 81-85.
[16] Barry, E. F., et al.: Teaching OO Methodology in a Project-Driven CS2 Course.
Proceedings of ACM OOPSLA ‟05, Oct. 16-20, 2005, San Diego, California, USA, pp
338-343.
80
SICASE-818
Active Low Pass Filter with Transmission Zero Using OTA-TURC
Sorapong Wachirarattanapornkul
Faculty of Engineering, King Mongkut‟s Institute of Technology Ladkrabang
(KMITL),Thailand
Abstract
This paper describes a active low pass filter with transmission zero by using the passive
element which 2 types of uniformly distributed RC line (hereafter called URC) such as one
three-capacitive layers URC and one double capacitive layers URC (hereafter called TURC and
DURC with respectively) and active element as two operational transconductance amplifier or
OTA. The simulation results by PSPICE and MATLAB as shown a steeper rate in cut-off, and
this has low sensitivity in the pass band of active and passive elements to confirm this circuit
idea.
Keyword: active low pass filter, transmission zero, TURC, OTA
1. Introduction
The uniformly distributed RC line or URC is one of electronic devices that applied in filter
circuit by using URC as the lumped RC circuits. [1]-[8] And URC has many structure such as
lumped RC, VLSI circuit and distributed RC elements in thin-film technology is built using
smaller substrate area, and less parasitic problems at high frequency. Nowadays, URC elements
have three types of discrimination based on the capacitive layers such as, URC [1]-[2] has one
capacitive layer, DURC [3]-[4] has double capacitive layers and TURC [6]-[8] has three
capacitive layers. The structure of the general URC consists of layer of conductors, resistive
layer and dielectrics can be sandwiched together in many permutations.
The general of filter circuits are some disadvantages as ripple magnitude in the pass band or a
slow rate at stop band. Thus we are to propose this circuit by combination the DURC, TURC
and OTA which has a steeper rate in cut-off and the minimum attenuation in the stop band.
81
2. DURC and TURC
For basically to circuit symbol of the DURC and TURC in this part as show in Fig. 1
Fig. 1 as shows the passive DURC and TURC symbol circuits, by Fig.1 (a) as show DURC
circuit symbol you can see DURC components has double capacitive layers (CD1 and CD2) one
resister layer (RD1), then Fig.1 (b) as show as TURC circuit symbol with has three capacitive
layers (CT1, CT2 and CT3) and one resister layer (RT1) For analysis the admittance parameters
[Yij] of the DURC and TURC in Fig.1 by set one port of DURC and TURC to connect ground
as show in eq.(1) and eq.(2) with respectively.
[
] [
( )
( )
( ) ( ) ] [
] (1)
Where *( ) ( )+,
,
, ,
√ ,
and are the values of total resistance and capacitive of DURC, and is the complex
frequency.
[
] [
( )
( )
( )
( )
( )
( )
( )
( )
] [
] (2)
82
Where *( ) ( )+,
*( ) ( )+,
* ( ) +,
,
,
, ,
√ ,
and are the values of total resistance and capacitive of TURC, and is the complex
frequency.
3. OTA
The OTA or operational transconductance amplifier is one of active building block circuit used
to active device, Basically OTA has 3 port as input port in voltage has 2 ports ( ( ) and
( ), one output port in current ( o) and transconductance gain (gm) to change current
output. By the circuit symbol of OTA as shown in Fig. 2, and the characteristic of OTA as
shown in equation (3).
gm( ( ) ( )) (3)
Where gm
at
is bias current of OTA
4. The Proposed of Active Low Pass Filter with Transmission Zero Circuit
The proposed of active low pass filter with transmission zero circuit using the combination of a
TURC and DURC are shown in Fig.2. For the circuit design structure has 2 sections as, the
first section is a DURC and a TURC are cascade, the second section is active element as two
OTA in amplifier circuit for adjust gain frequency response.
83
The transfer function of the proposed circuit in Fig.3, is given as:
T( ) −α α
+sinhP P
−α α
+sinhP P
K
, −(α +α K)-
α −α α
γ+sinhP P
δ− −α α
co +sinhP P
α K
α −sinhP P
(4)
Where γ β
,
δ β
( ),
K gm
gm , β
,
,
,
3
,
√𝑠 , √𝑠 ,
, ,
84
In Fig. 4 as shown as the frequency response of the proposed circuit when varies K with K is
ratio between transconductance gains of OTAs in circuit (K gm
gm ⁄ ), you can seen in
response has two transmission zeros in stop band which has the results in greater slope. And the
peak of transmission zeros in the stop band is depend on gm
gm ⁄ .
5. Simulation Results
The simulation result in frequency response by MATLAB and PSPICE of the proposed circuit
in Fig.3, and transfer function in equation as (4). We choose values of the circuit elements as
follow:
00KΩ,
U 8pF,
U 00pF, (5)
0. 8 ,
0. 8 , 0. ,
gm gm 0. 8
85
Fig. 5 as shown as the frequency response of the proposed circuit with compare MATLAB and
PSPICE, you can seen that the frequency response of MATLAB and PSPICE as conformable.
And the sensitivity response of the active and passive element of the proposed circuit as shown in
Fig. 6, as low sensitivity in the passband.
86
6. Conclusion
The active low pass filter with transmission zero using OTA-TURC for propose and discuss.
This circuit has four parameters as; , , , and K gm
gm ⁄ for used to adjustment
in magnitude of transmission zero, and steeper slope. In improve to sensitivity analysis of
active (OTA) and passive (TURC and DURC) element of the proposed circuit has low in the
passband. Finally This proposed circuit will be suitable to fabrication by LSI process.
Acknowledgement
This work is supported by Faculty of Engineering, King Mongkut‟s Institute of Technology
Ladkrabang (KMITL).
References
[1] M.S. Ghausi/J.J. Kelly, “Introduction to Distributed Par ameter Networks with Application
to Integrated Circuits,” Holt, Rinehart and Winston, INC. pp.-136-141
[2] M. Teramoto, S. Sudo, Y.suzuki, “On the Design of the Active Low Pass Filer using
Double Layer Uniformly Distributed RC Line,” JIC-CSCC’95, 1995
[3] K. Janchitrapongvej, et.al “Notch Tunable Filters using Double Layer Uniformly
Distributed RC Line,” 1998 IEEE APCC/ICCS, Vol 2, pp.590-592.
[4] S. Sorapong, J. Kanok, M. Teramoto, and S. Sudo, “Notch Frequency Adjustable Active
Filer Using Uniformly Distributed RC Line,” IEEE Asia Pacific Conference on Circuits
and Systems Proceedings, China, p.568-570. Dec 2000
87
[5] S. Sorapong, J. Kanok, et.al, “On The Realization of Active Filter with Notch
Characteristics using Multilayer Uniformly Distributed RC Line,” IFCS2002 The
International Conference on Fundamentals of Electronics Communications and Computer
Sciences, pp.14_10-14_13, Japan, March 2002.
88
Biological Engineering / Biomedical Engineering /
Fundamental Science
Session Chair: Vinoth Jagaroo
SICASE-701
Experimental and FEM Study of Red-Blood Cells
Zhao Yong︱Alfaisal University
Mohamed Tarek Abdelaty︱Alfaisal University
Ibrahim Nabeel Muhssen︱Alfaisal University
Muhammad Usman Zafar︱Alfaisal University
Nourah Abdulaziz N Alrubaiq︱Alfaisal University
Arwa Abdullah S Alanqary︱Alfaisal University
Samah F. AlAbbasi︱Alfaisal University
Randa AlNounou︱Kingdom of Saudi Arabia
SICASE-729
Neurocognitive Systems and Biomedical Engineering: Optimization through the
Phenotype-Marker Approach and Neuroinformatics
Vinoth Jagaroo︱Boston U. School of Medicine & Emerson College
William J. Bosl︱University of San Francisco
SICBENS-1292
Single and Mixed Microbial Biosorbents and the Influence of Their Varying Dosages on
Removal of Lead (Pb) in Multi-Metal Solutions
Adeline Su Yien Ting︱Monash University Malaysia
Carrie Siew Fang Sim︱Monash University Malaysia
SICBENS-1300
Obesity Crusade through Structural Study on FTO and Its Inhibitor Design
Yonggui Gao︱Nanyang Technological University
Joel D.W.Toh︱Institute of Molecular and Cell Biology
Lingyi Sun︱National University of Singapore
Esther C. Y. Woon︱National University of Singapore
2015/6/27 Saturday 16:30-18:00 Room 2
89
SICASE-732
Case Study: Experimental of New Propeller Design 2 Blades Type and Observation
Cavitation of Tailing Thai Boat
Jaruphant Noosomton︱Kasetsart University
Sarinya Sanitwong Na Ayutthayaa︱Kasetsart University
SICASE-829
Model and Algorithm of Violator Detector Network problem: Application on Traffic
Network
Amber Shafiq︱Korea Advanced Institute of Science & Technology
90
SICASE-701
Experimental and FEM Study of Red-Blood Cells
Zk Abdelaty, Muhammad Usman Zafar, Nourah Abdulaziz N hao Yong*,
Mohamed TareAlrubaiq, Arwa Abdullah S Alanqary
College of Engineering, Alfaisal University, Kingdom of Saudi Arabia
Ibrahim Nabeel Muhssen, Samah F. AlAbbasi
College of Medicine, Alfaisal University, Kingdom of Saudi Arabia
Randa Al Nounou
Department of Pathology, King Faisal Specialist Hospital & Reseach Center,
Kingdom of Saudi Arabia
Abstract
Blood cells are composed of 99.9% erythrocytes, red blood cells. Erythrocytes transport
oxygen to all cells of the body through narrow blood capillaries. The mechanical properties and
deformability of erythrocytes are therefore vital in maintaining their ability to perform their
functions. Many factors and diseases are known to change erythrocytes in both shape and
mechanical properties, such as: sickle cell anaemia, and carbon monoxide. The effect of such
changes on the functionality of red blood cells in vivo however remains everyone‟s educated
guess. In this study, finite element analysis was used to model, analyse, and visualise the
functionality of red blood cells in a number of realistic biological settings. Experimental results
for cell properties and biological settings were obtained from literature and experimental
measurements using a nano-indentation testing. The commercial FEA package ANSYS was
used on a parallel cluster to perform the analysis. Statistical data are to be obtained from FEA
results to help determine the probabilistic effect of several biological factors on the
performance of red blood cells in the body.
Key words: red blood cells; finite element method; nano-indentation; smoking
1. Introduction
Blood consists of plasma and blood cells. Blood plays a very important role in keeping our
bodies alive, it serves as a medium for transporting nutrients and wastes to and from body cells.
Moreover, it plays a major role in defending our body from pathogens.
Red blood cells (erythrocytes) account for 99.9% of blood cells. Their principal function is the
transport of oxygen from the lungs to respiring tissues, and carbon dioxide from tissues back to
91
the lungs. This is achieved through the presence of haemoglobin, a conjugated metallo-protein
(Ashton, 2013). The non-nucleated erythrocyte is unique among human cells in that the plasma
membrane, its only structural component, accounts for all of its diverse antigenic, transport,
and mechanical characteristics (Mohandas & Gallagher, 2008). The ability of cell membrane to
bind and transport oxygen is what makes its physiological function possible (Zuk,
Targosz-Korecka, & Szymonski, 2011).
Erythrocytes health is related to the mechanical properties of the cells. One of the unique
mechanical features of the normal red cell membrane is its high elasticity, which enables the
cell to rapidly respond to applied fluid stresses in the circulation (Mohandas & Gallagher,
2008). Erythrocytes abnormalities are the most common inherited disorders (Mohandas &
Gallagher, 2008). Many of those disorders lead to a decreased elasticity which may lead to
haemolysis (rapturing of cell membrane) thus causing a decrease in the number of erythrocytes.
This is due to the decrease in the ability of erythrocytes to pass repeatedly through the
microcirculation. A decrease in erythrocyte deformability may therefore lead to tissue death
and necrosis. Hereditary blood disorders are also common in Saudi Arabia; in the premarital
screening and genetic counselling (PMS&GC), results showed that 2.7 out of each examined
1000 were positive for sickle cell diseases & 42.4 out of each examined 1000 were carriers for
sickle cell diseases (Memish, Owaidah, & Saeedi, 2011).
According to (Mohandas & Gallagher, 2008), biophysical studies showed that the following 3
features are the primary regulators of the ability of the cell‟s capacity to undergo the necessary
deformations:
(1) Geometry of the cell, particularly cell surface area to volume ratio;
(2) Cytoplasmic viscosity determined by intracellular haemoglobin concentration; and
(3) Membrane deformability: determined by the protein spectrin.
Extensive research has been done in studying the mechanical properties of red blood cells, and
their ability to deform and change in shape. Researches have also looked at the effect of
different drugs [e.g. (Zuk et al., 2011)], and storage conditions on those properties.
The interaction between erythrocytes, and chemicals and molecules in our blood affect the
membrane elasticity. Membrane proteins interact with so-called xenobiotics (eg, drugs) from
the blood and facilitate their transportation within the circulation. Some drugs which bind with
the cell membrane can modify the structure and elastic properties of red blood cells, and may
eventually cause haemolysis of the erythrocyte (Zuk et al., 2011). In (Zuk et al., 2011), the
effect of two drugs used in asthma treatment was tested and it was found that they cause a
decrease in the ability of erythrocytes to bind to oxygen and transport it.
92
One of the molecules that bind to the haemoglobin of erythrocytes is carbon monoxide (CO).
CO has much higher affinity for haemoglobin than oxygen (Varma, Mulay, & Chemtob, 2009).
This account for its high toxicity, when haemoglobin binds to carbon monoxide (CO),
carboxyhemoglobin (COHb) is formed. Despite the fact that carbon monoxide poisoning is
prevalent, no previous research was done on the effect of carbon monoxide binding to the
erythrocytes on their mechanical properties. Carbon monoxide poisoning is highly severe,
especially for the organs with the highest demand for oxygen such as the brain and heart, which
are more vulnerable to injury (e.g. myocardial infarction, arrhythmias ,etc.) (Lippi, Rastelli,
Meschi, Borghi, & Cervellin, 2012).
Pollution is one of the most significant factors that increase the level of CO in the atmosphere.
Another important factor is smoking. Smoking tends to produce carbon monoxide, thus
smokers tend to have a higher level of CO in their blood. In Saudi Arabia, smoking is one of
the major problems. (Al-Nozha et al., 2009) showed that overall prevalence of smoking among
Saudis was 12.8%. Males (1555) were significantly smoking more than females (662) with a
prevalence of 18.7% and 7.3% respectively.
In this study, the effect of carbon monoxide on mechanical properties of erythrocytes shall be
analysed. This study will help in better evaluating the risks of exposure even to a low level of
CO.
2. Nano-Indentation Experiment
Different methods exist for studying the mechanical compliance S of soft tissues, beside
nanoindentation, such as: AFM probing (McNally & Borgens, 2004), soft substrate stretching
(Brown, 2000), optical tweezers, (Dao, Lim, & Suresh, 2003) and magnetic-bead attachment
(Wang, Butler, & Ingber, 1993). In this study however the nanoindenter currently available at
Alfaisal University shall be utilized to calculate the mechanical compliance of red blood cells.
The advent of nanoindentation, or depth-sensing indentation, has provided a method for
determining time-dependent mechanical properties on a size scale compatible with tissue
dimensions in controlled, physiological environments. (Gupta, 2008) Depth-sensing
indentation involves the application of a controlled load or displacement to an indenter,
inducing local deformation of the surface beneath the indenter tip. Both the load and the
displacement are monitored simultaneously during loading and unloading of the indenter, and
the resulting load-displacement curve shall be used to characterize the mechanical compliance
of the surface using the following formula.(Gupta, 2008; Zhang, 2008)
( − ) ( )
( − ) ( ) ,
93
where is the force applied to the RBC specimen, E is the Young's modulus of the
specimen which is 1/S, is the Poisson's ratio (often assumed to be 0.5), is the half
opening angle of the indenting tip, and is the indentation depth. (Zhang, 2008) From the
above equation, and are provided by the nanoindentation measurements, is a
known property of the indenter‟s head used, is assumed, and S can therefore be determined.
Different types of heads are commonly used in nanoindentation, such as: pyramidal heads
which include the Berkovich, Vickers and Knoop indenters. Other heads commonly used are:
spherical, conospherical and cylindrical (flatpunch) indenters. Figure 1 shows three different
types of commonly used indenters. Due to the smaller contact area, pyramidal indentation is
less sensitive than conospherical and flatpunch indentation to surface irregularities and
asperities. Conospherical tips are therefore recommended for testing most polymerical and
biological materials. (Gupta, 2008) In this research, a conospherical indenter shall be used.
3. Finite Element Modelling
In an attempt to better visualize the effect of carbon monoxide on the mechanical compliance
and the functionality of RBC, a finite element model for the passage of a red blood cell in a
thin capillary shall be analyzed using the finite element package ABAQUS/CAE. The model
will be based on the values for mechanical compliance computed from experimentation.
4. Modelling of RBCs in Capillaries
5. Conclusion
6. References
Al-Nozha, M. M., Al-Mazrou, Y. Y., Arafah, M. R., Al-Maatouq, M. A., Khalil, M. Z., Khan,
N. B., … Nouh, M. S. (2009). Smoking in Saudi Arabia and its relation to coronary artery
94
disease. Journal of the Saudi Heart Association, 21(3), 169–76.
doi:10.1016/j.jsha.2009.06.007
Ashton, N. (2013). Physiology of red and white blood cells. Anaesthesia & Intensive Care
Medicine, 14(6), 261–266. Retrieved from
http://www.sciencedirect.com/science/article/pii/S1472029913000787
Brown, T. D. (2000). Techniques for mechanical stimulation of cells in vitro: a review. Journal
of biomechanics, 33(1), 3–14. Retrieved from
http://www.ncbi.nlm.nih.gov/pubmed/10609513
Dao, M., Lim, C. T., & Suresh, S. (2003). Mechanics of the human red blood cell deformed by
optical tweezers. Journal of the Mechanics and Physics of Solids, 51(11-12), 2259–2280.
doi:10.1016/j.jmps.2003.09.019
Gupta, S. (2008). Analytical and Numerical Nanoindentation Studies of Compliant
Biomaterials and Soft Tissues (pp. 1–52).
Lippi, G., Rastelli, G., Meschi, T., Borghi, L., & Cervellin, G. (2012). Pathophysiology, clinics,
diagnosis and treatment of heart involvement in carbon monoxide poisoning. Clinical
Biochemistry, 45(16), 1278–1285. Retrieved from
http://www.sciencedirect.com/science/article/pii/S0009912012002731
McNally, H. A., & Borgens, R. Ben. (2004). Three-dimensional imaging of living and dying
neurons with atomic force microscopy. Journal of neurocytology, 33(2), 251–8. Retrieved
from http://www.ncbi.nlm.nih.gov/pubmed/15322383
Memish, Z. A., Owaidah, T. M., & Saeedi, M. Y. (2011). Marked regional variations in the
prevalence of sickle cell disease and β-thalassemia in Saudi Arabia: findings from the
premarital screening and genetic counseling program. Journal of epidemiology and global
health, 1(1), 61–8. doi:10.1016/j.jegh.2011.06.002
Mohandas, N., & Gallagher, P. G. (2008). Red cell membrane: past, present, and future. Blood,
112(10), 3939–48. doi:10.1182/blood-2008-07-161166
Varma, D. R., Mulay, S., & Chemtob, S. (2009). Carbon Monoxide: From Public Health Risk
to Painless Killer. In R. C. Gupta (Ed.), Handbook of Toxicology of Chemical Warfare
Agents (pp. 271–292). Retrieved from
http://www.sciencedirect.com/science/article/pii/B9780123744845000201
Wang, N., Butler, J. P., & Ingber, D. E. (1993). Mechanotransduction across the cell surface
and through the cytoskeleton. Science (New York, N.Y.), 260(5111), 1124–7. Retrieved
from http://www.ncbi.nlm.nih.gov/pubmed/7684161
Zhang, W. L. (2008). Design, Modeling, Fabrication, and Characterization of a MEMS Device
for Measuring the Mechanical Compliance of a Biological Cell.
Zuk, A., Targosz-Korecka, M., & Szymonski, M. (2011). Effect of selected drugs used in
asthma treatment on morphology and elastic properties of red blood cells. International
journal of nanomedicine, 6, 249–57. doi:10.2147/IJN.S15802
95
SICASE-729
Neurocognitive Systems and Biomedical Engineering: Optimization through
the Phenotype-Marker Approach and Neuroinformatics
Vinoth Jagaroo
Boston U. School of Medicine & Emerson College
William J. Bosl
University of San Francisco
Abstract
This paper highlights the relevance of the phenotype-marker approach in neuroscience to
biomedical engineering for the brain sciences. In behavioral neuroscience, a rapidly emerging
approach to resolving the multidimensional nature of neurocognitive processes is that of
neurocognitive phenotypes – the precise delineation of neural circuits and their componential
role in behavior, and mapped across a genome-to-phenome pathway. This approach is critically
enabled by neuroinformatics – informatics geared to the discovery of relationships between
complex data sets in the neurosciences. Together, phenotypes and neuroinformatics
(phenomics) define a new framework for neuroscience discovery. The central point of the
paper is that continuing advances in biomedical engineering pertaining to the brain become
clinically useful if articulated within this new relational framework.
Neuroscience has achieved remarkable breakthroughs in recent decades. It is now possible to
conceive of realistic frameworks to discover measureable neural correlates of behavior and
behavioral disorders. The recent US National Institute of Mental Health Strategic Plan calls
for the development of new ways of classifying psychopathology based on observable behavior
and neurobiological measures. The goal of this strategy is to discover the fundamental “units
of behavior” – cognitive phenotypes -- that can be reliably measured and used not only for
classification of psychopathological disease but also for mapping neurophysiological measures
to these units of behavior. The emphasis on cognitive and neural phenotypes is central to this
strategy.
Conventional interpretation of brain research data inevitably aligns with one or more
operational models. For example, functional imaging studies heavily adopt the notion of
cortical centers/modules within a larger network. While valid, this model has also come
under criticism for over-reliance on a distributed nodal-network perspective. Complex
dynamical models of the brain involving nonlinear system properties (drawn from physics and
96
engineering), or immunohistochemical parcellations of the brain (a perspective drawn from
brain oncology), are examples of models that have been less amenable to neuroscientists.
A comprehensive understanding of the brain requires the delineation of multiple levels of
neural complexity and how these map to behavioral measures or cognitive phenotypes, and
hence the powerful new trajectory of the phenotype-marker approach.
The paper: (a) first provides an overview of phenomics. (b) As it central illustrative example,
it then examines the area of brain-computer interface – examples of signal interpretation from
experimental neural implants and from portable EEG headsets are considered, first outside the
neural marker framework and then within it, to show how interpretation and engineering
efficacy are aided by a neurophenotype approach. (c) Some current developments in
neuroinformatics and neurocognitive markers are reviewed with the aim of addressing their
relevance to brain related biomedical engineering initiatives. Special attention is given to the
use of machine learning algorithms to find patterns in multidimensional neurophysiological
measures and map them to behaviorally defined phenotypes.
The paper may be of relevance to neuroscientists, neuro-engineers, and data scientists,
especially those interested in the exciting frontier of neural systems interface with computers
and machines.
97
SICBENS-1292
Single and Mixed Microbial Biosorbents and the Influence of Their Varying
Dosages on Removal of Lead (Pb) in Multi-Metal Solutions
Adeline Su Yien Ting*, Carrie Siew Fang Sim
School of Science, Monash University Malaysia, Malaysia
[email protected]; [email protected]
Abstract
In this study, the fungal isolate Saccharicola bicolour (Sb) was applied as mixed biosorbent
with a metal-tolerant bacterium Stenotrophomonas maltophilia (Sm). The influence of
adsorbent dosages (0.10, 0.50, 1.00, 1.50 g) of the mixed biosorbent (Sm+Sb) was also tested
and compared against respective single biosorbents (Sm, Sb) to determine their efficacy in
removing Pb in multi-metal solutions. Results revealed that single biosorbent Sm was the most
effective in adsorbing Pb, followed by mixed biosorbent (Sm+Sb) and Sb single biosorbent.
Biosorbent dosage as low as 0.10 g was sufficient to remove Pb effectively, with highest
biosorption at this dosage reported for single application of Sm (8.712 mg g-1
), followed by
mixed biosorbent of Sm + Sb (4.946 mg g-1
) and by single biosorbent of Sb (2.389 mg g-1
). The
biosorption process for all three biosorbents complied with the pseudo-second order kinetic.
The FTIR-analysis indicated that functional groups detected in most biosorbents include
hydroxyl, amino, methyl, alkene, carbonyl, and carboxyl groups. Although single biosorbent
Sb has the most number of spectra peaks (11 peaks) compared to Sm+Sb (10 peaks) and Sm (8
peaks), efficacy was not significantly enhanced.
Keywords: biosorbent dosage; mixed biosorbent; Saccharicola bicolour; single biosorbent;
Stenotrophomonas maltophilia; Pb removal
1. Introduction
Lead (Pb) in the environment is of great concern due to their toxicity to living organisms. It is a
non-essential metal ion, often discharged from mining and welding industries, and has extreme
effects on cardiovascular, nervous, immune and reproductive systems [1, 2]. Throughout the
years, physicochemical processes have been employed to remove Pb and other metals from the
environment, which include ion exchange, reverse osmosis and chemical precipitation [3, 4].
These methods however, were less effective in low metal concentrations. As an alternative,
biosorption using biological materials are sought resulting in various bacterial (Bacillus,
Pseudomonas) and fungal (Aspergillus, Penicillium) species established as biosorbents [5].
Early studies performed were mostly on the use of single biosorbents. However, in this study,
mixed microbial biosorbents were used to mimic the various microorganisms in the natural
environment [6], and how they may interact with one another and influence biosorption.
98
In this study, the biosorbents were tested in varying dosages as biosorbent dosages are known
to influence metal biosorption attributed to availability of binding sites on the surface of
biosorbents. In addition, we determined the mechanism of biosorption of the biosorbents via
kinetic modelling and examined their surface properties using the FTIR approach. The novelty
in our approach is the use of Saccharicola bicolour and Stenotrophomonas maltophilia, in
which the former is a new fungal endophyte, recently discovered to tolerate several key metals
[7]; and the latter a well-known bacterium with tolerance to Cu [8]. We also attempted this
study using multi-metal solution where it mimics the natural wastewaters. Hence, this paper
reports the influence of biosorbent dosages of mixed biosorbents (Sm+Sb) and their respective
single biosorbents (Sm, Sb) on the removal of Pb in multi-metal solutions.
2. Materials and Methods
Isolate Preparation and Biosorbent Development
The bacterial isolate S. maltophilia (Sm) used in this study was isolated from river sediment of
the Penchala River, Kuala Lumpur, Malaysia [8] while the fungal isolate S. bicolour (Sb) was
isolated from Phragmites, a plant used to treat leachate [7]. Both Sm and Sb were maintained
on Nutrient Agar (NA, Merck) and Potato Dextrose Agar (PDA, Merck), respectively, at room
temperature (25±2oC). To generate biomass, Sm was inoculated to 250 mL of Nutrient Broth
(NB, Merck) while Sb was cultured in 250 mL Potato Dextrose Broth (PDB, Merck). After 3-5
days of incubation, biomass of Sm was retrieved by centrifugation (9000 rpm, 7 min) while for
Sb, biomass was obtained via filtration using Whatman No. 1 filter paper. Both biomass were
autoclaved (121oC, 20 min), oven-dried, powdered, sieved (mesh size ~ 0.08 cm) and stored in
sterile bottles at room temperature (25±2oC) until use.
Biosorption Study
The biosorbents were prepared and assigned as Treatments 1 (T1), 2 (T2), 3 (T3) where T1:
Sm only, T2: Sb only, and T3: Sm+Sb (mixed-biosorbent). The dosage for each treatment
(biosorbent) was varied from 0.1, 0.5, 1.0 to 1.5 g biosorbent, and inoculated to 15 mL of
multi-metal solution (containing 100 ppm each of Pb, Cu, Zn, Cd). The initial pH was adjusted
to pH 5 and incubated with agitation (150 rpm, 480 min). At every sampling period (15, 30, 60,
120, 240, 360, 480 min), the solutions were filtered with filter paper and the collected filtrates
were analysed using Atomic Absorption Spectroscopy (AAS) [Agilent Technologies 240
Series AA, air-acetylene flame]. The settings were 15mA, with wavelength 217.0 nm for Pb.
The biosorption capacities (Q) were then calculated based on [9] as follows:
99
Kinetic Modelling
To predict the rate controlling mechanisms, pseudo-first and pseudo-second order kinetic
models were used. Pseudo-first order kinetic involves the following equation:
Log (qe-qt)=log qe-
.
where qe: metal adsorbed at equilibrium; qt: metal adsorbed at particular time; k1: rate constant;
t: time
Pseudo-second order kinetic is expressed as:
t
where t/qt: contact time divided with adsorption at particular time; ƙ2: rate constant; qe:
adsorption
For pseudo-first order, plots of log (qe-qt) against t should give a linear relationship of which k
and qe were derived from the slope and intercept of the graph, respectively. Similarly, for plots
of pseudo-second order, t/q against t should give straight lines. The adsorption, qe and rate
constant, k2 were derived from the slope and intercept, respectively. Compliance to pseudo-first
order model suggests that neither boundary layer nor external resistance occurs at the
beginning of the biosorption process while a good fit to pseudo-second order model indicates
that biosorption is limited by chemisorption [10].
FTIR Analysis
The functional groups present on the surfaces of all the biosorbents were analysed using
Fourier Transform Infrared Spectroscopy (FTIR) [Thermo Scientific NicoletTM
iSTM
10] via the
Attenuated Total Reflectance (ATR) method. The dried and powdered biosorbents (as prepared
in previous sections), was directly introduced into the ATR diamond crystal. The IR
absorbance data were then obtained with a total of 16 scans at resolution of 4 cm-1
, for
wavenumbers ranging from 600-40000 cm-1
[11]. Spectra with percentage transmittance
against wavenumbers were generated for the single and mixed biosorbents. Functional groups
represented by the peaks generated were then identified.
100
Statistical Analysis
The obtained data were analysed using ANOVA (Analysis of Variance). The mean and
standard deviations were compared with Tukey comparisons (HSD(0.05)) using the Statistical
Packaging for the Social Science (SPSS) software version 20.0.
3. Results and Discussions
Biosorption of Pb by Biosorbents and the Influences of Dosages
The single biosorbent Sm (T1) adsorbed Pb more effectively than mixed biosorbent (T3:
Sm+Sb) with 1.111 to 8.712 mg g-1
removed by the former, compared to lower levels of 0.573
to 4.946 mg g-1
of Pb adsorbed by the latter (Figure 1). The inferior performance by mixed
biosorbent (T3: Sm+Sb) may have been attributed to the presence of Sb in the biosorbent, as
the single biosorbent Sb (T2: Sb) removed only 0.551 to 2.389 mg g-1
of Pb. All three
biosorbents however, showed a similar trend where Pb was removed most effectively when
0.10 g biosorbent was used. At this dosage, single biosorbent of Sm (T1) adsorbed 8.712,
followed by mixed biosorbent (T3: Sm+Sb) with 4.946 and T2 (Sb) with 2.389 mg g-1
of Pb
(Figure 1). Increase in biosorbent dosages to 0.50, 1.00 and 1.50 g resulted in a significant
decline in amount of Pb adsorbed (except for T1: Sm), with lowest biosorption occurring at
1.50 g for T2: Sb (0.551 mg g-1
) and T3: Sm+Sb (0.573 mg g-1
). For T1: Sm, the lowest
biosorption was recorded when 1.00 g of biosorbent was used instead (1.111 mg g-1
) (Figure
1).
Our observations here suggest that biosorbents of the bacterial (Sm) origin may have superior
biosorption potential than fungi (Sb), presumably due to the negatively-charged surfaces of
bacteria that were capable of binding to cations [12]. Bacteria also have outer membranes
which are rich in negatively-charged phosphate groups which aids in metal binding. The
addition of Sb to Sm (mixed biosorbent) was antagonistic for the biosorption of Pb, as lower
levels were removed compared to use of single biosorbent Sm. This may be a result of ligands
and functional groups of the two organisms competing for metal cations thus interfering with
biosorption. It is also possible that the antagonistic interaction observed in this study is between
a combination of fungal (Sb) and bacterial (Sm) isolate, as other studies have shown that
mixtures of bacteria-bacteria biosorbents were able to remove more metals than their single
biosorbents [13]. For fungal biosorbents, biosorption is better when applied as mixed
biosorbent (T3: Sm+Sb) compared to single biosorbent (T2: Sb).This has been reported by [14]
although the mixtures used in their study were of fungal-fungal origin. Further investigations
using more combinations of bacterial-fungal and fungal-fungal as mixed biosorbents are
recommended. There may also be an influence of metal-metal interaction between the metals in
the multi-metal solution, where metal preference for binding may occur depending on the type
of biosorbent used. In reference to biosorbent dosages, lower dosages applied were more
101
beneficial than higher dosages. One of the possible reasons is that aggregation of biomass
occurs when larger dosages are used resulting in lesser surface area for biosorption [15].
Kinetic Modelling for Biosorbents
Biosorption of Pb to Sm (T1), Sb (T2) and Sm+Sb (T3) is a rate-limiting step, and occurs as a
result of exchange of electrons between biosorbent and metal cations [10]. The compliance to
pseudo-second order was evident with higher R2 values derived compared to comparisons with
pseudo-first order (Table 1). In addition, the calculated qeq values were similar to
experimental qeqex values derived from pseudo-second order kinetic, suggesting a good fit to
pseudo-second order kinetic (Table 1).
Functional Groups Analysis via FTIR
There were approximately 8, 11 and 10 peaks observed in the spectra for T1 (Sm), T2 (Sb) and
T3 (Sm+Sb), respectively (Table 2, Figure 2). For T1 (Sm) peaks at wavenumbers 3265.88
cm-1
, 2924.87 cm-1
, 1633.36 cm-1
, 1519.94 cm-1
, 1455.20 cm-1
, 1393.04 cm-1
, 1223.66 cm-1
and
1065.54 cm-1
were detected, indicating possible functional groups of hydroxyl or amide,
methyl, alkene or carbonyl, aromatic alkene or nitro, methyl and carboxyl respectively (Table
2). As for T2 (Sb), a total of 11 peaks could be seen in this single biosorbent spectrum with
102
wavenumbers 3265.53 cm-1
, 2922.38 cm-1
, 2852.60 cm-1
, 1743.37 cm-1
, 1625.73 cm-1
, 1548.76
cm-1
, 1376.63 cm-1
, 1311.60 cm-1
, 1236.37 cm-1
, 1151.59 cm-1
and 1027.30 cm-1
which
suggested possible functional groups of hydroxyl or amide, methyl, carbonyl, alkene or
carbonyl, aromatic alkene or nitro, methyl, nitro and carboxyl groups respectively (Table 2,
Figure 2). In mixed biosorbents, though there were differences in the number of peaks
observed; hydroxyl, amino and carboxyl groups are consistently present in all the biosorbents
(single and mixed biosorbents) suggesting primary roles in binding of Pb (Table 2, Figure 2).
4. Conclusion
To conclude, the single biosorbent of Sm was superior in removing Pb, followed by the mixed
biosorbent (Sm+Sb) while single biosorbent of the fungal origin (Sb) was the least effective.
Biosorption was attributed to the functional groups present on the biosorbents, which were
generally similar for single and mixed biosorbents. The biosorbent dosage that was most
effective in removing Pb was 0.10 g (for all three biosorbents) for 15 ml of multi-metal
solution.
Acknowledgements
This project supports a main project funded by the Ministry of Higher Education (MOHE)
under the FRGS scheme (FRGS/2/2013/STWN01/MUSM/02/2). The authors are also grateful
to Monash University Malaysia for the facilities to conduct the study.
103
104
References
[1] L. Järup, Hazards of heavy metal contamination, British Medical Bulletin, 2003, 68:
167-182.
[2] World Health Organization, 2006.
105
[3] A. Malik, E. Grohmann, Environmental Protection Strategies for Sustainable Development,
2011, Springer, New York.
[4] R. Mohee, A. Mudhoo, Bioremediation and Sustainability: Research and Applications,
2012, John Wiley & Sons, Canada.
[5] R. Boopathy, Factors limiting bioremediation technologies, Bioresource Technology, 2000,
74: 63-67.
[6] H. Mikeskova, C. Novotny, K. Svobodova, Interspecific interactions in mixed microbial
cultures in a biodegradation perspective, Applied Microbiology and Biotechnology, 2012,
95: 861-870.
[7] C.S.F. Sim, W.S. Tan, A.S.Y. Ting, Endophytes from Phragmites for metal removal:
evaluating their metal tolerance, adaptive behaviour and biosorption efficacy,
Désalination and Water Treatment, 2015, DOI: 10.1080/19443994.2015.1013507.
[8] A.S.Y. Ting, C.C. Choong, Utilization of non-viable cells compared to viable cells of
Stenotrophomonas maltophilia for copper (Cu(ii)) removal from aqueous solutions,
Advances in Environmental Biology, 2009, 3: 204-209.
[9] S. Zafar, F. Aqil, I. Ahmad, Metal tolerance and biosorption potential of filamentous fungi
isolated from metal contaminated agricultural soil, Bioresource Technology, 2007, 98 (13):
2557–2561
[10] J. Febrianto, A.N. Kosasih, J. Sunarso, Y.H. Ju, N. Indraswati, S. Ismadji, 2009.
Equilbrium and kinetic studies in adsorption of heavy metals using biosorbent: a summary
of recent studies, Journal of Hazardous Materials, 2009, 162: 616-645.
[11] M. Iqbal, A. Saeed, S.I. Zafar, FTIR spectrophotometry, kinetics and adsorption isotherms
modelling, ion exchange and EDX analysis for understanding the mechanism of Cd2+
and
Pb2+
removal by mango peel waste, Journal of Hazardous Materials, 2009, 164: 161-171.
[12] P. Sanasi, J. Kader, B.S. Ismail, S. Salmijah, Sorption of Cr(VI), Cu(II), Pb(II) by growing
and non-growing cells of a bacterial consortium, Bioresource Technology, 2006, 97:
740-747.
[13] R. Gourdon, S. Bhande, E. Rus, S.S. Sofer, Comparison of cadmium biosorption by
gram positive and gram negative bacteria from activated sludge, Biotechnology Letters,
1990, 12: 839-842.
[14] R. Pan, L.Cao, R. Zhang, Combined effects of Cu, Cd, Pb, and Zn on the growth and
uptake of consortium of Cu-resistant Penicillium sp. A1 and Cd-resistant Fusarium sp.
A19, Journal of Hazardous Materials, 2009, 171 (1-3):761-766.
[15] A. Ö zer, G. Gürbüz, A. Ç alimli, B.K. Körbahti, B.K., Biosorption of copper (II) ions on
Enteromorpha prolifera: application of response surface methodology (RSM), Chemical
Engineering Journal, 2009, 146: 377–387.
106
SICBENS-1300
Obesity Crusade through Structural Study on FTO and Its Inhibitor Design
Joel D.W.Toh
Institute of Molecular and Cell Biology, Singapore
Lingyi Sun, Esther C. Y. Woon
Department of Pharmacy, National University of Singapore, Singapore
Yong-Gui Gao
Institute of Molecular and Cell Biology, Singapore
School of Biological Sciences, Nanyang Technological University, Singapore
The AlkB family of nucleic acid demethylases are of intense biological and medical interest
because of their roles in nucleic acid repair and epigenetic modification. However their
functional and molecular mechanisms are unclear, hence, there is strong interest in developing
selective inhibitors for them. FTO protein, known as fat mass and obesity-associated protein,
belongs to the AlkB family. As the first mRNA demethylase that has been identified, the FTO
gene has been implicated in causing obesity, which is relevant to many diseases in humans.
Here we report the identification of key residues within the nucleotide-binding sites of the
AlkB subfamilies that likely determine their substrate specificity. We further provide proof of
principle that a strategy exploiting these inherent structural differences can enable selective and
potent inhibition of the AlkB subfamilies. This is demonstrated by the first report of a
subfamily-selective and cell-active FTO inhibitor 12. The distinct selectivity of 12 for FTO
against other AlkB subfamilies and 2OG oxygenases shall be of considerable interest with
regards to its potential use as a functional probe. The strategy outlined here is likely applicable
to other AlkB subfamilies, and, more widely, to other 2OG oxygenases.
107
SICASE-732
Case Study: Experimental of New Propeller Design 2 Blades Type and
Observation Cavitation of Tailing Thai Boat
Sarinya Sanitwong Na Ayutthaya*
Kasetsart University, Faculty of International Maritime Studies, Thailand
Jaruphant Noosomton
Kasetsart University, Faculty of International Maritime Studies, Thailand
Abstract
This paper presents to new design of propeller 2 blade types of low-speed and observation
cavity flow around ship propellers experimentally. Testing of different types new model
propellers and local type propellers are performed in this study. The first a 2-blades, skew angle
18°, Developed area 35 call model-A (B2-18-35) and model-B a local type propeller which
commonly used by Thai fisherman. The cavitation phenomenon of model A and B at 850 rpm
were compared. It can be observed that cavitation of model A was less than model B at both
speeds. Surprisingly, cavitation at face of blade was observed for model-B, causing adverse
effect to performance of the propeller.
Keywords: Developed Area, Skew Angle, Cavitation Number ()
108
SICASE-829
Model and Algorithm of Violator Detector Network problem: Application on
Traffic Network
Amber Shafiq
Department of Industrial and Systems Engineering, Kaist
Abstract
This paper utilizes the flow interception model with the objective of locating detectors on
transportation network to maximize the expected benefit to catch the violators who are threat
for public safety. We install the detectors where peoples or vehicles come from the residential
area near candidate location or flow passing by candidate location. Model was then solved
using two approaches: a mathematical approach and a greedy heuristic approach. The results
show that the improved greedy heuristic gives a solution that is very close to optimum in terms
of the number of detectors needed to intercept a given percentage of the vehicle flow. Several
numerical experiments are conducted to illustrate the computational efficiency and solution
quality of the proposed heuristic algorithm.
Keywords: Facilities planning and design, Flow interception problem, Traffic Network, Greedy
Heuristic, Genetics Algorithm
Introduction
In most facility location problems, the task is to locate facilities on a network to service a set of
customers located on the vertices of the network. The objective is to optimize the cost of
facility location or the cost of assigning customers to facilities, or both. Different variants of
facility location problems arise as a result of problem-specific characteristics such as the type
and number of facilities to locate and the type and quality of service required. Flow
interception problems form a distinct class where customers are not located on the vertices of
the network. Instead, they flow on the network between their respective origins and
destinations. The problem is to install detectors to optimize the expected benefit by catching
the violators on a traffic network. A violator can be caught by a detector at any node/candidate
location on its path and the benefit may depend on the position of the facility, flow and covered
length of the vehicle on the link before detection.
Traditional location_allocation models, such as the maximal covering location model (MCLM)
and the p-median model, aim to locate network facilities to optimally serve demand expressed
as weights at nodes [1_4]. Nowadays, many customers purchase services as part of routine
pre-planned trips, i.e., the daily commute to and from home and the workplace, instead of
109
making a special-purpose trip to obtain a service. Such facilities include convenience stores,
gas stations, ATM machines, drugstores, laundries and restaurants. Thus, as the purchasing
behavior changes, there are cases where demand in a network is now expressed as flows, rather
than nodes. To solve these types of facility locations in a network where demand is not
expressed at nodes, but is exerted by traffic flowing between origins and destinations, Hodgson
[5] and Berman [6] present the flow interception problem (FIP) and developed a heuristic
greedy algorithm to solve the FIP. The basic problem of FIP [5,6] is to locate m facilities to
intercept as much flow as possible from a given set of pre-existing flows on the network. It
assumes the ``interception'' occurs if a flow passes through at least one facility. The focus is on
maximizing the total consumption of the service by ``flow-by'' customers traveling on
preplanned paths (e.g. daily commute). Based on the basic FIP, they also published a series of
studies for a class of FIP [7_9].
Real-time traffic information is sampled by detectors to monitor traffic status and to develop
control strategies. The effectiveness of a traffic surveillance system depends on not only the
accuracy of the sampled information but also the coverage over the transportation network.
However, implementing these new technologies usually requires large investment. Accuracy
and coverage are often two conflicting objectives due to limited resources: collecting high
quality information usually relies on sophisticated and expensive technologies and thus limited
budget would restrict the number of installations; on the other hand, due to the limited effective
range of most sensors, complete coverage over a network usually requires dense installations.
To balance this trade-off, intensive studies have been conducted to determine efficient and
reliable deployment of surveillance systems. Lam and Lo (1990) proposed a heuristic approach
to select locations for traffic volume count sensors in a roadway network. Yang et al. (1991)
conducted a robust analysis on the utility of traffic counting points for traffic O–D flow
estimation. Yang and Zhou (1998) proposed a sensor deployment framework to maximize such
utilities. This framework has been extended to accommodate turning traffic information
(Bianco et al., 2001,2006), existing installations and O–D information content (Ehlert et al.,
2006), screen line problem (Yang et al., 2006), time-varying network flows (Fei et al., 2007;
Fei and Mahmassani, 2008), railcar inspection under potential sensor failures (Ouyang et al.,
2009), and unobserved link flow estimation (Hu et al., 2009). The emergence of advanced
traffic sensing technologies also enabled flow estimation for individual O–D paths. For
example, Gentili and Mirchandani (2005) investigated guidelines for locating advanced traffic
sensors that are able to read both a vehicle‟s identification and its route information. Castillo et
al. (2008) proposed a mathematical model to determine the optimal locations of vehicle plate
scanning sensors for path flow reconstruction.
110
Based on the above theory we want to locate detectors on traffic network to maximize the
expected benefit by catching the maximum violators. The violators are mixed with common
public or vehicles volume from any origin to destination.
Expected benefit depends on the vehicle flow volume and length or speed covered by the
violator travelling on assigned origin like home to destination as workplace or simply passing
through the detector location. The detector location is assigned as downstream and upstream
location of the nodes on the routes covered the OD pairs.
Literature Review
Dimensional variation is inherent to any manufacturing process. Therefore, it is important to
model and predict how it propagates through the processes. Fast and accurate evaluation of
inherent process variation can be critical in determining the final dimensional variation of an
assembled product and in selecting robust product/process designs. Several models have been
proposed in the past to predict how variation propagates during assembly. Initial approaches
were focused on rigid part assembly using either the Root Sum Squares (RSS) method or
Monte Carlo Simulation. Detailed review and discussion can be found in Lee and Woo(1990)
and Chase and Parkinson(1991).
Significant research has been conducted on fixture fault diagnosis, including the use of
principal component analysis (PCA) and designated component analysis (DCA). PCA is a
systematic methodology to diagnose fixture faults by extracting the principal components from
the measured data on the subassembly (Hu and Wu 1992). Because of its entire dependence on
the measurement data, the root cause of principal components extracted may not always have a
physical meaning. To address this problem, the method of designated component analysis
(DCA) was developed by Liu and Hu (2005). This approach first defines mutually orthogonal
dimensional variation patterns (called diagnostic vectors) for a part, each of which is associated
with one or more fixture faults. Then, actual measurement data are decomposed into the
components of the diagnostic vectors, disclosing the major components and thus the associated
fixture faults. DCA is a powerful method for simple problems, but it is not always obvious how
to define physically meaningful, mutually orthogonal patterns for complex part geometries.
More physics-based diagnostic vectors, which are only constrained to be linearly independent
with each other, were derived by Ceglarek and Shi (1996). The diagnostic vectors are defined
by simple geometric analyses of rigid body movement for each fixture fault. The same
diagnostic vectors were used for diagnosis of multiple fixture faults using the least squares
method by Apley and Shi (1998).This diagnosis approach was adopted by Camelio, Hu, and
Yim (2005) for the fixture fault diagnosis in the assembly of complaint parts, where the
diagnostic vectors corresponding to fixture faults were obtained by finite element analyses
(FEA). Furthermore, Camelio, Hu, and Yim (2005) also presented a methodology to determine
111
the optimal locations of sensors on the parts by applying the effective independence (EfI)
method so that different diagnostic vectors can readily be distinguished from each other.
At station level, Liu et al.(1996) and Liu and Hu(1997) proposed a model to analyze the effect
of deformation and springback on assembly variation by applying linear mechanics and
statistics. Using finite element methods (FEM), they constructed a sensitivity matrix for
compliant parts of complex shapes. The sensitivity matrix establishes the linear relationship
between the incoming part deviation and the output assembly deviation. Long and Hu(1998)
extended this model to a unified model for variation simulation by considering part variation
and locating fixture variation. Shiu et al (1997) presented a simplified flexible beam
representation of body structures. Huang and Ceglarek(2002) presented a discrete cosine-
transformation (DCT) based decomposition method for modeling and control of compliant
assemblies form error. The method decomposes the dimensional error field into a series of
independent error modes. At multi-station level, Lawless et al(1999) proposed a method called
Variation Driver Analysis using time series analysis. The method is based in tracking the
characteristics of individual parts as they pass through multiple stations using autoregressive
models. Comparatively, little research has been done in multi-station systems considering
compliant, non-rigid parts. Liu and Hu(1997) developed a model to evaluate the spot weld
sequence in sheet metal assembly. This model considered a process where welding was carried
out in multiple stages. Chang and Gossard (1997) presented a graphic approach for
multi-station assembly of compliant parts. However, there are no analytical models for
variation analysis in multi-station compliant assembly systems.
It is critical to develop realistic models for sheet metal assembly process that consider
compliant parts and also include the station-to-station interaction in multi-station assembly
systems. Such models can be quite useful during both the design and launch stage of the
manufacturing system. During design, such models can be used to predict product variation so
that changes in parts or processes can be made early. During the launch of the manufacturing
system, such models can aid in the diagnosis of root causes of variation. The purpose of this
paper is to present a methodology for modeling the impact of part and tooling variation on the
dimensional quality on a multi-station assembly system with compliant sheet metal parts and
study how variation propagates from different subassemblies to the final product by defining a
geometric variation correlation measure.
PCA for Pattern Identification
The application of PCA in variation pattern identification for assembly systems was first
proposed by Hu and Wu. These patterns often are closely related to rigid body motions due to
fixture failure and can be diagnosed by comparing them with a set of pre-assumed hypothetical
variation patterns based on knowledge of the product/process. The purpose of PCA is to
112
transform p correlated variables . . . into new uncorrelated
variables . . . , as defined in the following linear transformation:
From a geometric perspective, this is a coordinate rotation in the dimensional space. In Eq.
(1), the transformation matrix consists of p rows of mutually orthogonal vectors called
eigenvectors. The ith row of matrix B, described as , . . . - , is the ith eigenvector
of the covariance matrix obtained in the form
Where is the ith eigenvalue of the covariance matrix , and is an identity matrix. If one
wish to transform a set of variable by a linear transformation , whether is
orthogonal or not.
Example:
The data of 8 metal sheets is given:
Mean of the Principle Components:
113
In the figure x-axis, y-axis and z-axis are measured in mm and z-axis gives us the mean value
of the 8 principle components of the given data.
Mean of the Data (Using pc and Eigen Vectors)
In the figure x-axis, y-axis and z-axis are measured in mm and z-axis gives us the mean value
of the data obtained from pc‟s and Eigen vectors while going to backward direction.
Designated Component Analysis
The calculation of a principal component is a totally data-driven process, i.e., the measurement
data determines the eigenvectors and eigenvalues. The pattern compounding effect of PCA
0 5 10 15 200
10
20
0
1
2
3
4
5
05
1015
20
0
5
10
15
20
0
5
10
15
114
actually makes the interpretation of variation patterns difficult. The ability to distinguish less
significant patterns is also affected by the compounding effect. Because of this, only when a
variation pattern is very significant (usually the first eigenvector), we can approximate and
interpret it to a known pattern such as rigid body motion. The second and other eigenvectors
are usually hard to interpret.
Unlike PCA, in which we first estimate variation patterns statistically and then compare them
with physically known patterns, DCA first defines variation patterns with known physical
interpretations, and then measure their statistical significance. A basic requirement for those
known patterns (or vectors) is that they must be orthogonal to each other. As a result, the
designated components can be detected for existence, and also be decomposed, removed from
the original data domain, and analyzed and monitored individually. Furthermore, the
correlation among these designated components can be analyzed and utilized for diagnostic
purposes.
This is where DCA differs from the other process-oriented methods and methods based on
factor analysis.
With the same format as used for the definition of eigenvectors in Eq. (1), the designated
patterns (vectors) are denoted as the following obtained using Gram-Schmidt
orthonormalization:
After designating mutually orthogonal variation patterns, their corresponding designated
components can be calculated using the following equation:
Or
Where W are the designated components, D are the designated variation patterns and Y is the
given data of 8 sheets.
115
DCA Based Fault Diagnosis
The percent contribution of each pattern, , to the total variation can be calculated individually
as a proportion of the measurement data variation explained by each pattern. The total variation
of the system is defined as the sum of the variance of all patterns. The effect of each pattern is
calculated as the effect of the estimated variance of each significant pattern in the total
variation of the significant patterns. Therefore, the contribution of each pattern is obtained as
follows:
( 𝑠 𝑠 )
( )
This equation characterizes the contribution ratio of to system variation.
Percentage Contribution of Each Designated Pattern Significance Level
According to the significance level of each variation pattern we can remove the rigid body
motion of Tz and By .
Gram-Schmidt Orthonormalization
In mathematics, particularly linear algebra and numerical analysis, the Gram–Schmidt process
is a method for orthonormalising a set of vectors in an inner product space, most commonly the
Euclidean space Rn. The Gram–Schmidt process takes a finite, linearly independent set S =
v1, …, vk for k ≤ n and generates an orthogonal set S′ = u1, …, uk that spans the same
k-dimensional subspace of Rn as S.
We define the projection operator by
where denotes the inner product of the vectors u and v. This operator projects the vector v
orthogonally onto the line spanned by vector u.
The Gram–Schmidt process then works as follows:
116
The sequence u1, ..., uk is the required system of orthogonal vectors, and the normalized vectors
e1, ..., ek form an orthonormal set. The calculation of the sequence u1, ..., uk is known as Gram–
Schmidt orthogonalization, while the calculation of the sequence e1, ..., ek is known as Gram–
Schmidt orthonormalization as the vectors are normalized.
To check that these formulas yield an orthogonal sequence, first compute ‹ u1,u2 › by substituting
the above formula for u2: we get zero. Then use this to compute ‹ u1,u3 › again by substituting the
formula for u3: we get zero. The general proof proceeds by mathematical induction.
Geometrically, this method proceeds as follows: to compute ui, it projects vi orthogonally onto
the subspace U generated by u1, ..., ui−1, which is the same as the subspace generated by v1, ...,
vi−1. The vector ui is then defined to be the difference between vi and this projection, guaranteed
to be orthogonal to all of the vectors in the subspace U.
The Gram–Schmidt process also applies to a linearly independent countably infinite sequence
vii. The result is an orthogonal (or orthonormal) sequence uii such that for natural number n:
the algebraic span of v1, ..., vn is the same as that of u1, ..., un.
117
If the Gram–Schmidt process is applied to a linearly dependent sequence, it outputs the 0
vector on the ith step, assuming that vi is a linear combination of v1, ..., vi−1. If an orthonormal
basis is to be produced, then the algorithm should test for zero vectors in the output and discard
them because no multiple of a zero vector can have a length of 1. The number of vectors output
by the algorithm will then be the dimension of the space spanned by the original inputs.
Geometric Variation
Geometric variation is obtained using the following formula:
∑ ( 𝑠 )
Where C‟values(significance level of each designated pattern) are the weights and mode shapes
are the corresponding designated component and „i‟ is the total number of mode shapes which
is 5 in our case.
Geometric Variation of Weighted Mode Shape
This figure tells us about the geometric variation of all 5 weighted mode shapes. In the figure
x-axis, y-axis and z-axis are measured in mm and z-axis gives us geometric variation of the 5
mode shapes.
0
5
10
15
20
05
1015
20
0.01
0.012
0.014
0.016
0.018
0.02
118
Variation Pattern of Original Data
In the figure x-axis, y-axis and z-axis are measured in mm and z-axis gives us the mean value
of the 8 sheet data.
Error Variation Pattern between Geometric Variation and Original Data
This error is calculated using following formula
Where geometric variation is the sum of weights of 5 mode shape variation and data matrix is
the 8 sheet data.
In the figure x-axis, y-axis and z-axis are measured in mm and z-axis gives us the error value
between geometric variation of mode shapes and the given data.
We also calculated the error percentage as
0 510
15 20
0
5
10
15
20
0
2
4
6
8
10
12
0 2 4 6 8 10 12 14 16 18
0
5
10
15
20
-15
-10
-5
0
119
(
) 00
Where Error is the value calculated in the above formula and data mean is the mean of the 8
sheets data. Error% has the value -0.0100.
Proposed Methodology
Multi-Station Assembly Model
The automotive body assembly process will be used to develop the methodology for modeling
and analyzing dimensional variation propagation in multi-station systems. However, the
developed model can be generalized into other multi-station assembly process with compliant
parts such as shipbuilding or furniture manufacturing.
An automotive body assembly process is a multi-leveled hierarchical process, in which sheet
metal parts are joined together to form a subassembly. During the assembly process, each part
or subassembly becomes an input for the subsequent stations. While parts move from one
station to another, dimensional variation of the parts and subassemblies propagates through the
system. The station-to-station interactions cause an increase or sometimes decrease of the
dimensional variation. An assembly process can be considered a discrete-time dynamical
system, where the independent variable time can represent the station location. Then, a state
space representation can be developed to illustrate the part deviation. In station k, part
deviation after assembly operations is function of the input parts deviation and major mode
designated patterns. Then, part deviation can be calculated by the state equation, if function f is
known. The objective of the model is to define the appropriate functions f.
( ) ( ( ) ( )) ( )
where X(k) represents the designated components in the assembly at station k, D(k) the
designated patterns of major mode shapes at station k and Y(k) the measurement data of sheet
metal parts variation at station k.
Problem:
How do we find the function “f” (which is the function of designated patterns)?
Following methods are used to develop “f”
1- PCA gives the designated patterns in the form of Eigen vectors.
2- DCA defines the designated patterns with known physical interpretation of the metal
sheets or from the human knowledge using previous data.
3- Define the weighted linear combination of the physical known patterns of the metal sheets
and estimate the weights using least square method.
120
Methodology
Based on the concept of geometric covariance (Camelio, Hu, and Marin 2004), it can be
assumed that the deformation vector of the assembly components can be decomposed in the
contribution of different deformation patterns. Therefore, the part variation vector can be
modeled as a linear combination of patterns, , each of which may represent a
potential variation pattern of the parts, as will be explained in next section. Then, the
deformation vector for the parts can be written as
. . .
Where
[
]
whose rows are the number of input sources and columns are the number of variation patterns.
xi (i = 1, 2, …, q) are the coefficients of the linear combination and represent the weight of the
each pattern in u. The dimensions of the vectors, bi and u, are both assumed to be p × 1; x
vector dimensions are q × 1; and B dimensions are p × q. p corresponds to the number of input
sources or key control characteristics (KCC).
The assembly variation vector, .
=
=[
𝑠 𝑠 𝑠 𝑠
]
.
different from those of the parts variation vector, u and is the sensitivity matrix obtained
using finite element analysis. The value of m here corresponds to the number of measurement
points on the assembly.
121
The measurement vector, y, in Figure 1 is simply modeled as the sum of the assembly variation
and noise.
Given that the matrices S and B are constant and known, Eq can be simplified as follows:
. . . .
often called the diagnostic matrix is defined as an m × q matrix. The matrix D represents the
effect of the different deformation patterns of the components on the assembly. It should be
noted from Eq. that the measurement vector, y, has the same dimension, m, as the assembly
variation vector, v. These m points of measurement must be carefully selected and thus the
sensitivity matrix, S, must be defined such that . yields v at those m selected points. The
columns of the matrix D, often called the diagnostic vectors, may be interpreted as the effect of
deformation pattern represented by each diagnostic vector, . , on the assembly
variation measurement.
The components of the vector x may accordingly be interpreted as the weights of those
components existing in the measurement y. From the problem of the present study may be
interpreted as finding the weights of diagnostic vectors x from the assembly measurement data
y given the diagnostic matrix D. In other words, the diagnostic problem is to estimate the
amount of variation in the measured data that can be explained by each designated pattern.
To construct the diagnostic matrix, , as . and to reduce its number of rows on to the
desired number of measurements, , we used the optimal sensor placement method. The
measurement data are obtained online from the measurement sensors, which
provide , for the j-th assembly and i-th station.
For each measured part, a measurement vector is obtained. The corresponding weight
vector, , is estimated using the least-squares method, as
( )−
( )
122
where is the least-squares estimator of . The estimator differs from the real
contribution because the noise effect is included. This approach requires the matrix to be
invertible. In other words, must be of full rank and will be look like
The percent contribution of each pattern , , to the total variation can be calculated
individually as a proportion of the measurement data variation explained by each pattern. The
total variation of the system is defined as the sum of the variance of all patterns. The effect of
each pattern is calculated as the effect of the estimated variance of each significant pattern in
the total variation of the significant patterns. Therefore, the contribution of each pattern is
obtained as follows:
( ) (
)
The function diag is defined as the diagonal of a matrix. The trace function is defined as the
sum of the variances of all significant components. If the k-th diagnostic vector (or equivalently
the k-th assembly pattern) is found to be significant, the k-th part variation pattern, , is
regarded as a significant source for the assembly variation.
Flow Chart of the Assembly Variation for One Station:
123
Results of Step 1:
The data was given according to the 2 steps of plate assembly in figure below
124
Redline shows the welding line and we have to assemble 3 plates in 2 steps.
PCA Analysis:
In the figure x-axis, y-axis and z-axis are measured in mm where x and y-axiz shows the grid
points and z-axis gives mean value of the principle components of the given data in step 1.
0
5
10
15
0
5
10
15-0.1
-0.05
0
0.05
0.1
0.15
125
DCA Based Fault Diagnosis:
Percentage Contribution of Each Designated Pattern Significance Level
Using DCA, after observing the significance level of each variation pattern we see that Rx is
cause of fault.
Least Square Analysis:
Using least square analysis, after observing the significance level of each variation pattern we
see that Rx is cause of fault.
126
Results of Step 2:
PCA Analysis:
In the figure x-axis, y-axis and z-axis are measured in mm where x and y-axiz shows the grid
points and z-axis gives mean value of the principle components of the given data in step 1.
DCA Based Fault Diagnosis:
Percentage Contribution of Each Designated Pattern Significance Level
Using DCA, after observing the significance level of each variation pattern we see that Ry is
cause of fault.
0
5
10
15
0
5
10
15-0.1
-0.05
0
0.05
0.1
127
Least Square Analysis:
Using least square analysis, after observing the significance level of each variation pattern we
see that Ry is cause of fault.
Conclusion
We defined multistage assembly model and test it with designated component analysis. Also,
the results are compared with Least square analysis. It concludes that multistage assembly can
detect the faults in assembly welding more accurately than the single stage assembly.
References
Apley, D. and Shi, J. (1998). “Diagnosis of multiple fixture faults in panel assembly.” Journal
of Mfg. Science and Engg. (v120), pp793-801.
Liu, Y.G. and Hu, S.J. (2005). “Assembly fixture fault diagnosis using Designated Component
Analysis.” ASME Journal of Mfg.Science and Engg. (v127, n2), pp358-368.
Strang, G. (1988). Linear Algebra and Its Applications, 3rd ed.Harcourt College Publishers.
Hu, S. J., and Wu, S. M., 1992, „„Identifying Sources of Variation in Automobile Body
Assembly and Using Principal Component Analysis,‟‟ Transaction of NAMRI/SME XX,
pp. 311–316
J A.Cam and Yim,H. J. (2006). “Identification of Dimensional Variation Patterns on Complaint
Assemblies.” Jounal of Manufacturing Systems. (v25,n2). pp. 65-76
Lee, W. J., and Woo, T. C., 1990, „„Tolerances: Their Analysis and Synthesis,‟‟ ASME J. Ind.,
112, pp. 113–121.
Chase, K. W., and Parkinson, 1991, „„A Survey of Research in the Application of Tolerance
Analysis to the Design of Mechanical Assemblies,‟‟ Research in Engineering Design, No.
3, pp. 23–37.
Hu, S.J. and Wu, S.M. (1992). “Identifying root causes of variation in automobile body
assembly using Principal Component Analysis.” Transactions of NAMRI/SME (v20),
pp311-316.
Liu, Y.G. and Hu, S.J. (2005). “Assembly fixture fault diagnosis using Designated Component
Analysis.” ASME Journal of Mfg. Science and Engg. (v127, n2), pp358-368.
Ceglarek, D. and Shi, J. (1996). “Fixture failure diagnosis for autobody assembly using pattern
recognition.” ASME Journal of Engg. for Industry (v118), pp55-66.
128
Apley, D. and Shi, J. (1998). “Diagnosis of multiple fixture faults in panel assembly.” Journal
of Mfg. Science and Engg. (v120), pp793-801.
Camelio, J.; Hu, S.J.; and Yim, H. (2005). “Sensor placement for effective diagnosis of
multiple faults in fixturing of compliant parts.” ASME Journal of Mfg. Science and Engg.
(v127,n1),pp68-74.
Liu, S. C., Hu, S. J., and Woo, T. C., 1996, „„Tolerance Analysis for Sheet Metal Assemblies,‟‟
ASME J. Mech. Div., 118~1!, pp. 62–67.
Shiu, B., Ceglarek, D., and Shi, J., 1997, „„Flexible Beam-Based Modeling of Sheet Metal
Assembly for Dimensional Control,‟‟ Transactions of NAMRI/ SME, 25, pp. 49–54.
Huang, W., and Ceglarek, D., 2002, „„Mode-based Decomposition of Part Form Error by
Discrete-Cosine-Transform with Implementation to Assembly and Stamping System with
Compliant Parts,‟‟ CIRP Ann., 51~1!, pp. 21–26.
Lawless, J. F., Mackay, R. J., and Robinson, J. A., 1999, „„Analysis of Variation Transmission
in Manufacturing Processes-Part I,‟‟ J. Quality Technol., 31~2!, pp. 131–142.
S.Charles Liu, S.Jack Hu (1997), “Variation Simulation for deformable Sheet Metal
Assemblies Using Finite Element Methods”. J. Manuf. Sci. Eng. 119(3), 368-374.
Minho Chang, David C Gossard(1997), “Modeling the assembly of compliant, non-ideal
parts”.Computer –Aided Design. 29(10), pp 701-708.
Mantripragada, R., and Whitney, D. E., 1999, „„Modeling and Controlling Variation
Propagation in Mechanical Assemblies Using State Transition Models,”IEEE Trans. Rob.
Autom., 115~1!, pp. 124–140.
129
Oral Sessions –June 28th
, 2015
Material Science & Engineering
Session Chair: Adita Evalina Fitria
SICASE-771
Interfacial Reactions of the Cu/Sn/Ni Sandwich Couples
Tzu-Ting Huang︱National Taiwan University of Science and Technology
Shin-Jung Chai︱National Taiwan University of Science and Technology
Yee-Wen Yen︱National Taiwan University of Science and Technology
SICASE-772
Dissolution Behavior of Ni Substrate and Ni3Sn4 Phase in Molten Lead-Free Solders
Yen Wei Chang︱National Taiwan University of Science and Technology
Meng Han Guo︱National Taiwan University of Science and Technology
Yee Wen Yen︱National Taiwan University of Science and Technology
SICASE-734
Development of A Method to Design An Electric Activated NiTi-Based Shape-Memory
Wire Actuator
Benjamin Fleczok︱Ruhr University Bochum
Dieter Kreimeier︱Ruhr University Bochum
SICASE-740
An Investigation Using Self-Sensing to Estimate the Lifetime of Shape Memory Actuators
Christian Rathmann︱Ruhr University Bochum
Benjamin Fleczok︱Ruhr University Bochum
Dieter Kreimeier︱Ruhr University Bochum
2015/6/28 Sunday 09:00-10:30 Room 2
130
SICASE-753
Effect of Carbon Nanotube Composition to Enhance Wear Resistance and Self –
Lubrication Cylinder Liner Based on CNT – Al2O3 + 13% TiO2 Nanocomposite by Cold
Spraying Method
Adita Evalina Fitria︱Universitas Indonesia
Aulia Fitriani︱Universitas Indonesia
Desak Putu Dewi Chrisnanda Indrayuni︱Universitas Indonesia
Ernitha Yuniar Ghaisani︱Universitas Indonesia
SICASE-755
Development of Leucite Glass-Ceramics for Non-Metallic Dental Product
Siti Mazatul Azwa Bt Saiyed Mohd Nurddin︱Mineral Research Centre
131
SICASE-771
Interfacial Reactions of the Cu/Sn/Ni Sandwich Couples
Tzu-Ting Huang, Shin-Jung Chai, Yee-wen Yen*
Department of Material Science and Engineering,
National Taiwan University of Science and Technology, Taiwan (R.O.C.) *[email protected]
Abstract
The 3D IC structure has replaced the traditional 2D advanced micro-electronic packaging. The
Cu/Sn/Ni sandwich structure is commonly used in the solder joint. The interfacial reactions,
including the surface morphology, intermetallic compound (IMC) formation, phase
transformation and Sn layer consumption in the Cu/Sn/Ni sandwich couple with different Sn
layer thicknesses at, 5, 10 and 15 µm, were investigated in this study. The reaction couples
were reflowed at 250oC for 1, 3 and 5 min, and then aged at 180
oC for 300 h. The results
indicated that at the same reflowing condition, the thick Sn layer comparing with the thin Sn
layer caused the longer time to consume the Sn layer. In the condition of reflowing for 3 and 5
min, the phase transformation occurred lately with the increase the Sn layer thickness. When
the reaction couple was reflowed for 1 min, the Cu3Sn phase thickness was decreased with the
increase of the Sn layer thickness. No Ni atoms in Cu3Sn were found during the aging process
for 300 h. Thus, the (Cu, Ni)6Sn5 might be the good diffusion barrier for Ni atoms to diffuse
toward the Cu3Sn phase.
Key words: 3D IC; Cu/Sn/Ni sandwich couple; interfacial reaction; diffusion barrier
1. Introduction
The 3D IC structure has replaced the traditional 2D structure in advanced micro-electronic
packaging. 3D IC packing uses solders to connect the electronic devices and substrates. After
high-temperature aging the solder will form intermetallic compounds (IMC) at the interface.
IMC formations affect interface morphology and negatively impacting the electronic products
reliability. These solder joints are microscopic in size, giving IMC formation great influence on
solder joints [1]. Electronic products containing lead components are prohibited from sale due
to environmental protection legislation. Lead-free solders, based on Sn and added with Ag and
Cu (SAC), have replaced the traditional Sn-Pb solders. SAC has good mechanical properties,
wettability and creep resistance [2-3].
The most common solder combination is Cu/solder/Ni in the flip-chip. Au on the substrate will
rapidly fuse into solder during reflowing process. On the circuit board side, organic surface
preservative (OSP) will be moved during reflowing.
132
This movement causes the Ni layer and Cu layer to react directly with the solder form the
Cu/solder/Ni sandwich structure [4-7]. With respect to different thickness of Sn, from 5 to 15
µm, we discuss the sandwich structure interfacial reactions at different reflowing temperatures.
We observe IMC formation under these experimental conditions.
2. Experimental Procedures
100 µm Cu thickness was used as the substrate in this study. The substrate was grinded and
polished before electroplating. A Sn layer was electroplated onto the Cu, followed by
electroplating a Ni layer onto the Sn. This formed the Cu/Sn/Ni sandwich structure. The Sn
electroplating solution contains Deionized water 500 ml, Solderon Acid HC 233 ml, Solderon
Tin 60 ml, Solderon ST-380 antioxidant 80 ml, Solderon ST-380 primary 4 ml and Solderon
ST-380 secondary 20 ml. The Ni electroplating solution contains NiSO4.6H2O 82.5 ml, FX-510
3 ml, DI water 64.5 ml, NiCl2.6H2O 2.25 g and H3BO3 6 g. The reaction couples were prepared
by electroplating 5, 10 and 15 µm-thick Sn layers onto Cu foil. Sn electroplating was carried
out at 5 ASD (A/dm2), while 10 µm-thick Ni was performed at 60°C for 3 min
The reaction couples, without heat treatment, were mounted in epoxy and subjected to a
metallographic grinding-and-polishing process. Optical Microscopy (OM) and Scanning
Electron Microscopy (SEM, HITACHI TM-1000) were used to examine the interface. The
reaction couples, Cu/Sn(5 µm)/Ni, Cu/Sn(10 µm)/Ni and Cu/Sn(15 µm)/Ni, were first placed
in a furnace at 250°C for 1, 3 and 5 min for reflowing. The samples were cooled and then
grinded and polished again. The experimental samples were then aged at 180°C for 3, 9, 20, 80,
100, 150, 200 and 300 hours. The intermetallic compounds (IMC) were then observed using
energy energy dispersive x-ray spectrometer (EDS) mixed binary phase diagram and Dual
Beam-Focused Ion Beam (DFIB) for compositional analysis and interfacial morphology. IMC
thickness were observed using SigmaScan Pro5.0. Five data points from each aging condition
were collected to calculate the average thickness.
3. Results and Discussion
We focused on the interfacial reaction between Cu and Ni using Sn at different thicknesses.
Solder joint cross sections were observed. IMCs were formed at the interface between the
substrate, Cu, and coating layer, Ni based on Cheng [8].
3.1 Interfacial Reaction of Cu/Sn(5 µm)/Ni
Figure 1(a) presents Cu/Sn(5 µm)/Ni reflowing for 1 min at 250°C. IMC formed at interface of
Sn/Cu. The gray region composition is Cu-46.3 at.% Sn and Cu-25.7 at.% Sn in the dark gray
region. This phase is likely to be the Cu6Sn5 and Cu3Sn [9]. On the Ni/Sn side, the IMC
composition is Cu-40.7 at.%Ni-58.1 at.% Sn and is likely to be the Ni3Sn4 phase [10]. Thus the
Cu diffusion rate is faster than Ni, no Ni diffusion aoccured at Sn/Cu. Only scant Cu atoms
133
diffused into Ni3Sn4. (Ni, Cu)3Sn4 is written to present the solubility.
The IMC becomes obviously thicker after aging for 3 hours, as shown in Figure 1(b). Some Ni
atoms are observed diffusing into Cu6Sn5, and we present it as (Cu, Ni)6Sn5. The Ni solubility
in Cu6Sn5 is 0.24 at.%. On the Ni/Sn side, the Ni solubility becomes 1.5 at.% with the increase
in aging time. Table 1 shows the solubility of Ni and Cu in each phase.
As the aging time increased to 50 hours, as Figure 1(c) shows, it forms (Cu, Ni)6Sn5+(Ni,
Cu)3Sn4 on the Ni/Sn side. Comparing the Cu solubility in (Ni, Cu)3Sn4 at 20 and 50 hours, the
solubility becomes less as the aging time increases, as listed in Table 1. The main reason is (Ni,
Cu)3Sn4 changes into (Cu, Ni)6Sn5+(Ni, Cu)3Sn4 after aging to 50 hours, so some Cu solubility
turns into (Cu, Ni)6Sn5.
Before aging to 100 hours, Ni atoms solute into (Cu, Ni)6Sn5 increases as the aging time
increases. But after aging to 100 hours, the solubility drops to 6.9 at.%. As shown in Figure
1(d), (Cu, Ni)6Sn5 contact each other at both ends. Therefore, Ni tends to reduce in order to
balance the concentration. Figure 1(e) presents all (Ni, Cu)3Sn4 changes into (Cu, Ni)6Sn5 and
the solubility of Ni reduces to 2.8 at.% when aging to 150 hours. When aging to 300 hours, the
Sn layer is tollay replaced by (Cu, Ni)6Sn5 and the solubility of Ni increases with the aging
time in this stage. Overall, Cu3Sn becomes thicker and the Ni solubility increases with time.
134
135
The two other reaction couples, reflowed at 250°C for 3 min and 5 min and then aged at 180°C
for 300 hours , exhibited similar phase transformation results s reflowing for 1 min. On the
Sn/Cu side the IMC types are (Cu, Ni)6Sn5 and Cu3Sn. On the Ni/Sn side, (Ni, Cu)3Sn4 also
turns into (Cu, Ni)6Sn5 after extended aging, as Figure 2 shows. The solubility is tabulated in
Table 1 and it shows that Ni atoms increase with time.
Figure 3 shows a cross section of Cu/Sn(5 µm)/Ni reflowed for 5 min and aged for 300 hours.
When aging to 300 hours the Sn layer is totally replaced by (Cu, Ni)6Sn5. The white dashed
line region is the Cu3Sn phase. The literature indicates that Cu3Sn should exhibit a columnar
structure, but we cannot find corresponding results from the FIB image [11]. It is speculated
that Cu3Sn formation in this study is different from that in the literature.
In summary, Cu/Sn(5 µm)/Ni forms same the IMC at different reflowing time, as Figure 4
shows. The only difference is no Ni solubility on the Cu/Sn side occurs when reflowed for 1
min. (Cu, Ni)6Sn5 is considered a diffusion barrier because no Ni solutes into Cu3Sn when aged
to 300 hours. The IMC formation under which the reflowing conditions and aging times
completely consumed the Sn layer are listed in Tables 2 and 3.
Figure 5 shows the relationship between the Cu3Sn thickness and reflowing time. When aged to
300 hours the thickest Cu3Sn was produced on the condition of reflowing for 1 min and then is
5 min. The reason is that (Cu, Ni)6Sn5 forms faster as the reflowing time increases, inhibiting
Cu3Sn growth [12].
136
However, the Sn layer is consumed earlier at 5 min than 3 min, and it makes Cu3Sn grow faster
in the late. It can be expected that if we prolong the aging time, the thickness at 5 min reflow
will be greater than at 1 min.
137
3.2 Interfacial Reaction of Cu/Sn(10 µm)/Ni and Cu/Sn(15 µm)/Ni
Similar results are obtained when solder is reflowed for 1, 3 and 5 min in the two reaction
couples. The IMC formation at different stages is shown in Table 2. Ni atoms solute into (Cu,
Ni)6Sn5 and also increase with aging time. (Cu, Ni)6Sn5 is also considered a diffusion barrier
because no Ni solutes into Cu3Sn when the aging is extended to 300 hours. In Cu/Sn(15
µm)/Ni, it is worth noting that the Sn layer was not completely consumed after aging to 300
hours and the Cu3Sn thickness becomes greater at reflowing for 5 min than 1 min.
3.3 The Comparison of Three Reaction Couples at Different Reflowing Times.
For the reaction couples of Cu/Sn(5 µm)/Ni and Cu/Sn(10 µm)/Ni were reflowed for 3 and 5
min, the phase transformation stage was delayed as the Sn layer became thicker. In these two
reaction couples the phase transformation occurred early with increased reflowing time.
Conversely, Cu3Sn became thinner as the Sn layer getting thicker. When Sn and Cu reacted
with each other, Cu6Sn5 grew first. Therefore, a thicker Sn layer produces a the thinner Cu3Sn
layer, as shown in Fig. 6.
138
Conclusions
In Cu/Sn(5 µm)/Ni and Cu/Sn(10 µm)/Ni, the phase transformation occurred early with the
reflowing time. In the same reflowing time, Cu3Sn got thinner as Sn layer was thicker. In the
same thickness of Sn, Cu3Sn got thicker as the aging time increasing, and Cu6Sn5 is considered
as a diffusion barrier of Ni diffusing to Cu3Sn.
Acknowledgements
The authors acknowledge financial support from the Ministry of Education (MoE) Top
University Projects. The author is also thankful for the help from Mr. S. C. Laiw who works at
National Taiwan University of Science and Technology for FESEM-EDS operation and
analysis.
Reference
M. P. Tian, Y. W. Yen, Electronic Packaging Technology for Semiconductor Devices, Taipei:
Wu-Nan Book Inc, 2005, pp. 5-8.
Y. W. Liu, RoHS Green Directive: global environmental norms & lead-free soldering
technology, Taipei: Dotop Media CO., LTD., 2005.
J. W. Yoon, S. W. Kim, S. B. Jung, J. Electron. Mater, 392 (2005) 247-252.
Y. W. Yen, Y. C. Chiang, C. C. Jao, D. W. Liaw, S. C. Lo, C. Lee, Journal of Alloys and
Compounds, 509 (2011) 4595-4602
Q. P. Li. 2005, Cross-Interaction of Ni and Cu in Sn. Master dissertation, Taoyuan: Graduate
Institute of Chemical and Materials Engineering, National Central University.
S. C. Yang, 2010, Combined Effects of Cu Concentration, Solder Volume, and Temperature on
SnAgCu/Ni Interfacial Reaction, PhD dissertation, Taipei:, Graduate Institute of Materials
Science and Engineering, National Taiwan University.
A. Munding, H. Hubner, A. Kaiser, S. Penka, P. Benkart, E. Kohn, Wafer Level 3-D ICs
Process Technology, A. Munding, Springer Sci. & Bus. Media, (2008) 132
C. W. Chang, S. C. Yang, C. R. Kao, J. Electron. Mater., 36 (2007) 1455-1461
139
N. Saunders, A. P. Miodownik, 11 (1990) 278-287.
K. N. Tu, R. D. Thompson, Acta metal, 30 (1982) 947-952.
J. F. Li, P. A. Agyakwa, C. M. Johnson, Acta Mater, 59 (2011) 1198-1211.
Y. W. Wang, 2010, Adding Cu and Ni to Solders on the Mechanism of Retarding Micro Voids
and Cu3Sn, Master dissertation, Taipei: Graduate Institute of Materials Science and
Engineering, National Taiwan University.
140
SICASE-772
Dissolution Behavior of Ni Substrate and Ni3Sn4 Phase in Molten Lead-Free
Solders
Yen Wei Chang, Meng Han Guo, Yee Wen Yen*
Department of Material Science and Engineering,
National Taiwan University of Science and Technology, Taiwan (R.O.C.) *[email protected]
The dissolution behavior of the Ni or Ni3Sn4 phase molten lead-free solders in molten Sn,
Sn-3.0Ag-0.5Cu (SAC), Sn-0.7Cu (SC), Sn-58Bi (SB), and Sn-9Zn (SZ) at 240, 270, and 300
was investigated in this study. The experimental results show the dissolution rate of Ni in
the molten solders follows the order Sn> SAC> SC> SB> SZ. The dense and thick Ni5(Zn,
Sn)21 phase was formed at the SZ/Ni interface and could restrain the Ni atoms to dissolve into
the molten SZ. Dissolution rate of Ni3Sn4 phase in the molten solders follows the order Sn>
SAC> SC> SB at 240 , and SC> Sn> SAC> SB at 300. Massive spilling of the (Cu,
Ni)6Sn5 phase was observed in the SAC/Ni3Sn4 and SC/Ni3Sn4 interface. The SZ solder reacted
with the Ni3Sn4 phase to form the Ni5(Zn, Sn)21 phase at the interface. And then Zn atom
diffused toward the Ni3Sn4 phase to form the Ni3(Sn, Zn)4 phase.
Key words: dissolution behavior; Ni; Ni3Sn4 phase; molten lead-free solders
1. Introduction
Traditional tin-lead solder has excellent welding properties and reliability 1-2
. However, lead
solder, a very toxic metal, causes environmental pollution 3-5
. Most electronic manufacturers
have replaced traditional tin-lead solder to meet world-wide lead-free regulations (WEEE &
RoHS) 6. Many kinds of lead-free solders have been developed, such as pure Sn,
Sn-3.0Ag-0.5Cu, Sn-0.7Cu, and Sn-9.0Zn, the lead-free solder with the most potential.
Nickel is typically used in lead-free solder as a surface finish or diffusion barrier for the under
bump metallization (UBM) layer. The interface reactions between nickel and solder are slow
rate, which means it prevents the solder from having reactions with the substrate 7-8
. In the
electronic packaging industry, nickel and its intermetallic compounds have a significant issue,
in that it directly affects the solder and UBM contact reliability 9-12
.
Studies focused on lead-free solder and electroless nickel substrate dissolution exist 13-15
, but
no research exists that indicates the dissolution rate values. This study explored the morphology,
intermetallic compounds and dissolution reaction mechanism of Ni/ solder, and Ni3Sn4/ solder
141
at different reaction temperatures and times. It is expected that the results of this study can
provide the industry a reference for product reliabilities.
2. Experimental
This research is divided into two parts: the Ni/ solder dissolution rate, and the Ni3Sn4/ solder
dissolution rate. Nickel wire 0.63 mm in diameter is used as the substrate. The nickel substrate
was cleaned and polished using 1 μm aluminum oxide to produce a smooth surface before
experimentation. Five grams of Sn, Sn-3.0Ag-0.5Cu (SAC), Sn-0.7Cu (SC), Sn-58Bi (SB), and
Sn-9Zn (SZ) were used as the various lead-free solders. The solder was placed into a quartz
tube and the tube placed in a vertical tubular furnace to melt the tin solder alloy. The nickel
wire with flux (W-2326) was then dipped vertically into the molten solder. The reaction
temperatures were set at 240, 270, and 300. The reaction times were set at 10, 20, 30, 40, and
50 min. After the process was completed the reaction couples were quenched and mounted
using a mounting press. The mounted reaction couples were ground with abrasive paper
starting with number 400 and advancing sequentially through 600, 800, 1200, 2400 to 4000
until the interface was revealed. The specimens were then polished using 0.3 and 0.1 μm
aluminum oxide powder. After the metallographic procedures the specimens were inspected
using an optical microscope (OM; Olympus BX51M, Japan) to observe the surface and
reaction interface. A scanning electron microscope (SEM; Hitachi TM-3000, Japan) with
backscattered electron image (BEI) was used for intermetallic compounds morphology analysis.
The reaction phase compositions formed between the solder and substrate were determined
using energy dispersive spectrometer (EDS) and electron probe micro analyzer (EPMA, JEOL
JSM-8200, Japan). Each composition report was compared with the phase diagrams to identify
the reaction phase. An image analysis system, measuring the metal substrate radius, was
utilized to observe the dissolution behavior, the average intermetallic layer thickness and the
nickel dissolution behavior in the liquid lead-free solders.
The second part of these experiments used Ni3Sn4 intermetallic compound and solder to test the
reaction interface dissolution rate. Based on the Sn-Ni binary phase diagram 16
, pure Ni3Sn4
bulk was prepared from Ni-Sn alloy at 1200 for 72 h. The samples were quenched in ice
water and placed into a 500 furnace for 720 h for annealing to reach the homogenous phase.
The alloy was then cooled and sliced into pieces 6 mm in diameter with a thickness of 1.3 mm.
The Ni3Sn4 specimen surface was sputtered with Mo film as the atomic diffusion barrier. The
Mo film was used as a reference frame to observe the Ni3Sn4 and solder dissolution behavior.
The solder was set at 0.2 grams per sample. The Ni3Sn4/ Mo/ solder sample was sealed into a
quartz glass tube and placed into a furnace with reaction temperatures of 240, 270, 300 and
reaction times of 10, 20, 30, 40, 50 min.
142
3. Results and Discussion
3.1 Pure Ni and Solder
Figure 1 shows the relationship between the Ni/ tin-solder dissolving radius and reaction time
at 240, 270, 300. In the initial reaction the nickel substrate quickly dissolved into the molten
tin-solder. The nickel radius reduction with the amount of nickel dissolution was linear with
increased reaction time. As the reaction temperature increased, the greater the amount of nickel
substrate radius dissolved. The experimental solubility data was measured using the linear
equation Y = AX + B's. Wherein A represents the substrate dissolution rate slope. The nickel
and various tin-solder dissolution reaction data are shown in Table 1. These data show the
dissolution rate speed in order, as Sn> Sn-3.0Ag-0.5Cu> Sn-0.7Cu> Sn-58Bi> Sn-9Zn, which
means the temperature is proportional to the dissolution rate.
143
In each reaction temperature case, the different dissolution rates between the solder and nickel
substrate can be explained using Dybkov's equation 17
:
)( CCV
Sk
dt
dCs
Eq. (1)
The amount of solder used in this experiment is quite large so that the substrate concentration
in the liquid solder C does not change. It can be seen as zero or constant (C = constant). The
tin-nickel binary phase diagram provided Cs, nickel concentration in the tin solder, as 0.07
wt.% under 240, 0.08 wt.% under 270, and 0.13 wt.% under 300. That means when the
reaction temperature is higher, the nickel solubility is larger. When (Cs-C) is increased the
gradient concentration between the substrate and solder is increased. Based on the experimental
data, Pandat was used to calculate the tin-nickel-copper ternary systems for the cross section
diagram 18
.
144
Figure 2 shows a FE-SEM picture that includes (a) Sn, (b) Sn-3.0Ag- 0.5Cu (SAC), (c)
Sn-0.7Cu (SC), (d) Sn-58Bi (SB), and (e) Sn-9Zn (SZ). These five kinds of solder and nickel
substrate were reacted at 240for 40 min with the interface etched. The morphology from
Figure 4(e) shows the Ni5(Zn, Sn)21 intermetallic layer is thick and dense, which means there is
no tunnel for diffusion. It may be deduced that Ni5(Zn, Sn)21 is a good diffusion barrier layer.
Therefore, nickel substrate dissolution can be effectively inhibited using the Ni5(Zn, Sn)21
intermetallic layer.
As Table 2 shows the dissolution rates are Sn> Sn-3.0Ag-0.5Cu> Sn-58Bi> Sn-0.7Cu> Sn-9Zn
at 240, Sn> Sn-3.0Ag-0.5Cu> Sn-0.7Cu> Sn-58Bi> Sn-9Zn at 270 and 300. However, the
experimental results showed Sn-0.7Cu> Sn-58Bi at 240. The reason is different kinds of
intermetallic compounds were formed during the process. Sn-58Bi solders only formed Ni3Sn4,
but Sn-0.7Cu solder formed both (Ni, Cu)3Sn4 and (Cu, Ni)6Sn5 19-20
. Therefore, Sn-0.7Cu
solder consumed more nickel substrate to form intermetallic compounds than Sn-58Bi solder
did.
145
3.2 Ni3Sn4 and Solder
The reaction temperature was set at 240 and 300. The Ni3Sn4 substrates were reacted with Sn,
Sn-3.0Ag-0.5Cu (SAC), Sn-0.7Cu (SC), Sn-58Bi (SB) and Sn-9Zn (SZ) to observe the
dissolution behavior and interfacial morphology.
Figure 3(a) is a pure Sn and Ni3Sn4 reacted at 240 for 20 min for the liquid-solid interface
reaction. Figure 3(b) is a pure Sn and Ni3Sn4 reacted at 300 for 20 min. From both of these
figures the interface does not generate any intermetallic phase during the reaction. The
experiments show that as the reaction time increases the solder becomes diffused along the
Ni3Sn4 grain boundary.
146
Figure 8 shows the Sn consumption thickness from the Sn and Ni3Sn4 reaction plotted versus
time at 240 and 300. The figure shows that the Ni3Sn4 consumption thickness is enhanced
with the increase in reaction temperature. The Ni3Sn4 consumption thickness is enhanced
parabolically with the increase in reaction time. The Ni3Sn4 dissolution rate is flattened by
degrees, which means the Ni concentration on the solder side is gradually saturated.
Figure 4(a) shows Sn-3.0Ag-0.5Cu and Ni3Sn4 reacted at 240 for 20min as the liquid-solid
interface reaction. Figure 4(b) is Sn-3.0Ag-0.5Cu and Ni3Sn4 reacted at 300 for 50min.
Figure 4(a) shows some detached layers between the SAC solder and Ni3Sn4 at 240. The
layer compositions are Sn-20.5Ni- 30.9Cu using EDS analysis, which means it is the (Cu,
Ni)6Sn5 phase. However, the inter- metallic compounds or detached layers were not observed
between the solder and Ni3Sn4 at 300.
Figure 8 plots the reaction consumption thickness between Sn-3.0Ag-0.5Cu and Ni3Sn4 versus
time at 240 and 300. The figure shows that Ni3Sn4 consumption thickness is enhanced with
the increase in reaction temperature. The Ni3Sn4 dissolution rate is flattened because the Ni
concentration in Sn-3.0Ag-0.5Cu gradually reaches saturation. Because the interface generated
the (Cu, Ni)6Sn5 phase at 240 and the copper from the solder gradually dissolved into the
Ni3Sn4 as (Ni, Cu)3Sn4 phase at 300, these two features consumed the copper concentration
in the solder. Using the Pandat calculation the copper concentration lead to an increase in the
Ni Cs equilibrium concentration. Based on Dybkov's equation Eq.(1) we know that an increase
in the Cs equilibrium concentration will drive up the dissolution forces. Thus, the Ni3Sn4
phase is increased again, as Figure 8 shows.
Figure 5(a) shows Sn-0.7Cu and Ni3Sn4 reacted at 240 for 50 min. Figure 5(b) shows
Sn-0.7Cu and Ni3Sn4 reacted at 300 for 40 min. The figure shows that Cu atoms from the
solder diffused into the Ni3Sn4 and formed the (Cu, Ni)6Sn5 phase. The detached layers
between the Sn-0.7Cu solder and Ni3Sn4 were observed only at 300, was and did not occur at
240. The Cu concentration in the intermetallic compound layer was higher at 300 than at
240. This evidence shows that copper diffusion occurs faster with increased temperatures,
which may cause the (Cu, Ni)6Sn5 phase to detach easier.
147
Figure 8 is plotted the consumption thickness of reaction between Sn-0.7Cu and Ni3Sn4 during
versus time at 240 and 300. The figure shows that consumption thickness of Ni3Sn4
enhanced with the reaction temperature increased. The dissolution rate of Ni3Sn4 is flatten by
degrees, which means the concentration of Ni in solder side is gradually saturated. Compared
to the line chart, the quantity of Ni3Sn4 dissolved at 300 was times than 240. That means
it is easily increased the dissolution of Sn-0.7Cu solder with temperature raised.
Figure 6 is Sn-58Bi and Ni3Sn4 reacted at 300 for 20min as the liquid-solid interface
reaction. It shows that the intermetallic phase dissolved into solder side.
Figure 8 is plotted the consumption thickness of reaction between Sn-58Bi and Ni3Sn4 during
versus time at 240 and 300. The figure shows that consumption thickness of Ni3Sn4
enhanced with the reaction temperature increased. The dissolution rate of Ni3Sn4 was not
flatten by degrees, which means the concentration of Ni in solder side was not saturated. The
main reason is the solder contained 58 wt.% of Bi, which is the element not to react with
Ni3Sn4 substrate. Thus, only Tin solder and Ni3Sn4 substrate involved the dissolution reaction.
In other words, the diffusion of Ni into the solder has not reached saturation after 50 min.
Figure 7(a) is Sn-9Zn and Ni3Sn4 reacted at 240 for 10min. Figure 7(b) Sn-9Zn and Ni3Sn4
reacted at 300 for 10min. The figure shows that there were two different intermetallic layers
148
formed in 240 and 300. The layer beside the solder is composed of Sn-16.2 at.% Ni-81.3
at.% Zn, which is formed Ni5(Zn, Sn)21 intermetallic phase by EDS analysis. The other layer is
composed of Sn-35.0 at.% Ni-21.3 at.% Zn, which is formed Ni3(Sn, Zn)4 phase. In addition,
the intermetallic layer was getting thicker with time and temperature increased.
Figure 8 is plotted the consumption thickness of reaction between Sn-9Zn and Ni3Sn4 during
versus time at 240 and 300. The figure shows that consumption thickness of Ni3Sn4
enhanced with the reaction temperature increased.
Overlapped the consumption thickness of reaction between different solder and Ni3Sn4 versus
time, we can get the figure 8(a) and 8(b). The dissolution rate of Ni3Sn4 phase in the molten
solders follows the order Sn> Sn-3.0Ag-0.5Cu> Sn-0.7Cu> Sn-58Bi at 240, and Sn-0.7Cu>
Sn> Sn-3.0Ag-0.5Cu> Sn-58Bi at 300. Massive spilling of the (Cu, Ni)6Sn5 phase was
observed in the SAC/Ni3Sn4 and SC/Ni3Sn4 interface. The Sn-9Zn solder reacted with the
Ni3Sn4 phase to form the Ni5(Zn, Sn)21 phase at the interface. And then Zn atom diffused
toward the Ni3Sn4 phase to form the Ni3(Sn, Zn)4 phase.
4. Conclusions
The nickel substrate and tin solder dissolution rate increased linearly with the reaction time and
temperature. The rate order is Sn> Sn-3.0Ag-0.5Cu> Sn-0.7Cu> Sn-58Bi> Sn-9Zn. Sn-9Zn
solder shows that it can inhibit nickel substrate and solder dissolution. Dissolution rate
differences can be explained using Dybkov's equation, intermetallic compounds formation and
149
grain morphology. However, the experimental results showed Sn-0.7Cu> Sn-58Bi at 240
which did not coincide with the ideal value. The reason is different kinds of intermetallic
compounds are formed during the process. Sn-58Bi solders only formed Ni3Sn4, but Sn-0.7Cu
solder formed both (Ni, Cu)3Sn4 and (Cu, Ni)6Sn5. Therefore, Sn-0.7Cu solder consumed more
nickel substrate to form intermetallic compounds than Sn-58Bi solder did. The arrangement
between the grains and grain boundaries will also affect the dissolution amount and rate. The
more densely stacked the formed grain boundary the lower the dissolution rate will be.
The next part is focused on the dissolution rate between the Ni3Sn4 substrate and various
solders with the temperature set at 240 and 300. The other solders (excluding Sn-9Zn solder)
displayed faster dissolution rate with increased reaction time and higher temperature. The
dissolution rate for the Ni3Sn4 phase in molten solders follows the order Sn> Sn-3.0Ag-0.5Cu>
Sn-0.7Cu> Sn-58Bi at 240, and Sn-0.7Cu> Sn> Sn-3.0Ag-0.5Cu> Sn-58Bi at 300.
Massive (Cu, Ni)6Sn5 phase spalling was observed in the SAC/Ni3Sn4 and SC/Ni3Sn4 interface.
The Sn-9Zn solder reacted with the Ni3Sn4 phase to form the Ni5(Zn, Sn)21 phase at the
interface, with Zn atoms diffusing toward the Ni3Sn4 phase to form the Ni3(Sn, Zn)4 phase.
Reference
D. R. Frear, P. T. Vianco, Metall. Mater. Trans. A. 25A (1994) 1509
B. Trumble, J. Brydges, IEEE Int. Symp. Electr. (1998) 112-119
D. R. Smith, Flegal, J. A. R, J. Electron Mater. 30 (2001) 1293
C. Ernhart, S. Scarr, presented at the ACYF research conference, New Directions in Child and
Family Research (1991).
A. Z. Miric, and A. Girusd, Solder Surf. Mt. Tech. 10 (1998) 19-25
P. Zarrow, Circuits Assembly, August, 18-20 (1999)
C. E. Ho, Y. W. Lin, S. C. Yang, C.R. Kao, D. S. Jiang, J. Electron Mater. 35 (2006) 1017-1024
W. H. Tao, C. Chen, C. E. Ho, W. T. Chen, C. R. Kao, Chem. Mater. 13 (2001) 1051-1056
C. M. Liu, C. E. Ho, W. T. Chen, and C.R. Kao, J. Electron Mater. 30 (2001) 1152-1156
C. E. Ho, L. C. Shiau, and C.R. Kao, J. Electron Mater. 31 (2002) 1264-1269
S. W. Chen, Y. W. Yen, J. Electron Mater. 30 (2001) 1133-1137
C. E. Ho, Y. M. Chen, C. R. Kao, J. Electron Mater. 28 (1999) 1231-1238
A. Sharif, Y. C. Chan, M. N. Islam, M. J. Rizvi, J. Alloy Compd. 388 (2005) 75-82
M. N. Islam, Y. C. Chan, A. Sharif, M. O. Alam, Microelectron Reliab. 43 (2003) 2031-2037
Y. W. Yen, W. T. Chou, Y. Tseng, C. Lee, C. L. Hsu, J. Electron Mater. 37 (2008) 73-83
T. B. Massalski, Binary Alloy Phase Diagrams-CD-ROM, Materials Park, OH: ASM
International (1996)
C. E. Ho, Y. M. Chen, C. R. Kao, J. Electron Mater. 28 (1999) 1231-1237
C. N. Chiu, S. W. Chen, Personal Communication
S. K. Kang, R. S. Rai, S. Purushothaman, J. Electron Mater. 25 (1996) 1113-1120
150
W. T. Chen, C. E. Ho, C. R. Kao, J. Mater. Res. 17 (2002) 263-266
151
SICASE-734
Development of A Method to Design An Electric Activated NiTi-Based
Shape-Memory Wire Actuator
Benjamin Fleczok, Dieter Kreimeier
Ruhr University Bochum, Germany
Abstract
The use of NiTi shape-memory alloys (SMA) in actuators generates advantages for designing
robust, simple and lightweight applications. The SMA effect is based on a phase transformation
of the atomic lattice in response to stress, strain and temperature. The resulting crystallographic
configurations lead to a complex behavior revealing different electrical and mechanical
characteristics.
Due to these complex interdependencies of the SMAs, it is very difficult to design an actuator
for defined requirements. Therefore, Konstantin Lygin has developed a comprehensive, generic
product development method (PDM) [1] that combines the V-Model [2] and the
TRIZ-methodology based on the new contradiction matrix of D. Mann [3]. Lygin extends the
method by specialized shape-memory based “development modules” to provide the basis for an
adapted PDM. He arranges and validates these SMA modules with an R-phase spring actuator
and points out the high potential for an evaluable PDM for SMAs.
Based on [1], this work presents product development modules for the V-model to design
electrical activated SMA wire actuators. In this context, the V-model is extended by new
approaches. In detail, the pre-evaluation module for the input parameters is extended by a new
tool which includes the interdependence of the complex material behavior. In addition to this, a
new impact factor has been created to rate the interdependence of the specific parameters of the
NiTi-based SMAs. Furthermore, the system conception is extended by new parameters for the
TRIZ contradiction matrix. This work will also discuss general design rules and the resulting
potentials of the design guide for SMA wire actuators. It also reflects on the challenges of the
new PDM.
Keywords: shape memory alloy, product development method, TRIZ, pre-evaluation tool,
impact factor
1. Introduction / Motivation
Today smart materials (SM) like shape memory alloys (SMA) already play an important role in
the product development. SMs have a multitude of advantages for the design of cross-industry
152
technical products such as the energy saving potential of about 16TWha/a and the emission
saving potential of about 8t CO2/a [4]. Other advantages are the potential for lightweight,
miniaturization and the significant potential for reducing the complexity of technical actuator
assemblies [5, 6]. Despite these advantages, the global market forecast for smart materials [7]
show in relation to other actuation technologies that the impact of shape memory alloys is still
very low. Smart materials, in detail comparable applications for motors and actuators, have a
market value of USD25.4 billion with a compound annual growth rate (CAGR) of 15.8%
(between 2011 and 2016). [7] Electric motors (DC-, AC-, hermetic motors) for actuation have a
total market value of USD120.68 and a CAGR of 6.3% from 20013 to 2019. [8]
The advantages in comparison of the total market value demonstrate the problems of the
implementation of SMAs in technical products. This is based on the complex relationship
between stress, strain and temperature and, as a result, the electrical and mechanical
characteristics that limit the handling for engineers and scientists [9]. This is justified by the
fact that there is a lack in the knowledge of product design [10]. The SFB459 also describes the
increasing importance of a rising knowledge base [11].
Lygin is the first who designs a comprehensive and generic product development method for
ambient activated (AA) NiTi-based R-phase SMA spring-actuators (SA) [1]. Previous
approaches did not consider the comprehensive product development cycle with the system
integration or a domain specific design [12, 13]. The method from [1] based on the V-model is
a valid approach for a systematic product development for engineers as a comprehensive and
evaluable tool. With individual product development modules based on the AA R-phase, SA
engineers have the possibility to systematically develop a spring actuator in an evaluable
process. This work is an important step forward for the implementation of SMAs in technical
applications. Lygin validates his developed method successfully and claims this method to be
tested and revised for other NiTi-based SMA actuators.
153
2. Method – Pre-valuation
The original model is now described before the revised and extended approaches will be
introduced.
Lygins method is based on a V-model which contains specialized and adapted method modules.
The model starts with a specification and a pre-evaluation. An abort criterion is linked from the
pre-evaluation module to the specification. If the specification is finished, the pre-evaluation
starts with the network diagram. The five corners of Lygin‟s module includes the important
temperature marks of the R-phase of SMAs. The influence of the temperature is the most
important factor that limits all SMA product developments significantly. Based on the fact that
the temperature influences a lot of characteristics of SMAs, the new approach for EA SMA
WA is also based on these parameters. The new network diagram differs in that it considers the
material-specific transformation temperatures and the resulting interdependent pre-evaluation
tool.
154
These specific temperatures influence the reaction behavior of an actuator significantly. The
following temperatures are essential for the pre-evaluation network for electrically activated
SMAs:
a. Austenite start
b. Austenite finish
c. Martensite start
d. Martensite finish
e. Functional temperature array
The start and finish temperature of the austenite are important to split in the development
process. A lot of products have special temperature specifications like storage temperatures or
post paint temperatures of about 120°C [14]. Therefore, it is important to check the finish
temperatures for the actuator. If the actuator activates in a post painting process, it could lead to
a total damage of the entire product. Furthermore, the array of the martensite finish temperature
is important for the reset behavior and the potential frequency of the actuator, so it is necessary
to put these temperature on one corner. The hysteresis limits the frequency and causes a
different behavior for the bidirectional displacement and should be focused for high-precision
actuation tasks.
155
A lot of applications set the functional temperature array for example at -40° to +85°C
(automotive – fuel clap actuator / central locking actuator) or -20°C to +55°C for applications
in the interior of civil aircraft. If these requirements have not been covered by the temperature
array of the SMA, one abort criterion is fulfilled. The respective industries react in a very
conservative way in that they change the requirements in the specification book due to the
product reliability so that these defined arrays are static.
Thus the network diagram for the evaluation of an electrically activated SMA WA also reacts
in a very inflexible way if one specification temperature is not in the temperature array
although a solution may be possible. The pre-evaluation network diagram has a limited
reliability because of the complex interdependencies. It should be possible to solve some of the
resulting contradictions by an intelligent mechanical product design. If, for example, the
electrically activated SMA wire is out of the functional temperature array of the network
diagram, a second wire (only ambient-activated) in an antagonistic position prevents the
operating displacement at an ambient temperature of 120°C.
To get a small barrier to enter in the complex SMA technology, the development engineers
need a simple base for decisions based on one factor. Therefore, this work has developed a
network diagram with evaluable factors. “Impact factors” (IF) are calculated by an
interdependency matrix in Table 1. These factors help to rate the five influences in the
pre-evaluation so that the network diagram could give more valid information in an early
product development phase. The chosen interdependency parameters of the matrix are the most
important factors for the product development with SMAs.
To fill out the matrix, the influences or effects from the chosen parameters of the column have
to be rated along the parameters in the line. If the column parameter has an effect to the
parameter of the line, the corresponding field is marked with an “X”.
156
The sum of the “X” has to be built for the parameters in the line and column. The sum of the
line shows the influence on other parameters (A); the sum of the column shows the degree of
the affect from other parameters on the corresponding parameter (B). As a result of this, the
influence factor of a parameter is calculated as follows:
+
(∑ ( − ) )
(1)
The parameter “functional temperature array” is calculated as follows:
+
( − ) 0. (1.1)
The sum of column and line has to be built divided by the value of the interdependence
parameters (As, Af, etc.) minus 1. This matrix rates the interdependence in accordance to the
interdependence matrix of May.
157
With the calculation as in (1.1), the most important factor for the network diagram is the
functional temperature array with an impact factor of 0.54. The result of the entire matrix is
that the strongest impact is caused by the electrical power on an EA SMA WA. Other
important factors are the force, the actuation frequency and, as mentioned before, the
functional temperature array.
As a result of this study, the temperature array is marked by the value of 1 in the new network
diagram. Out of the ideal temperature array (blue mark), the values decrease from 1 to 0. So if
one requirement is not in the ideal array, the value decreases.
As can be seen in Figure 3, some requirements are out of the ideal range (blue array). The
value of austenite start (AS) and martensite finish (MS) are 0.7 and 0.95. But with the impact
factor of 0.38 each, the difference to the ideal result is relatively low. The total rated value of
the example is about 1.857. This is a percentage difference of 9.285% to the ideal so that the
resulting solution should be taken into account for the product development. The range
between the values 1.65 and 2 is defined as a realistic value to continue the product
development (difference about 27% from the ideal). This expert opinion has to be validated in
following studies.
3. Method - System Concept
After these first steps to pre-evaluate the requirement in the development cycle, the “system
conception” starts. Lygin defines and analyzes two new models for the system conception with
AA R-phase SMA SA in the module box. The innovation checklist (ICL) and the contradiction
matrix are the tools from the TRIZ method which he defines and adjusts for the development
with SMA. For an electrically activated SMA WA, the corresponding modules have to be
proven and adjusted for the specific requirements.
158
In the module analysis, the innovation checklist helps to define the exact present situation
systematically with all external influences on the problem. The ICL is a convenient tool for the
engineer to get an extensive view on the problem.
The “problem history” (PH) has been added to the original toolbox by B. Klein [18] and D.
Zobel [19]. The PH describes the development of the problem and “systems with comparable
problems” as well as “additional problems that need to be solved”. This point helps to get
information of an earlier state of the problem or an earlier try to solve the problem.
The object function is designed to split the problem into differentiated functions and is an
assistance for the contradiction matrix. The contradiction matrix is a powerful tool in the
product development. To build the matrix, the input parameters have to be defined. Lygin
declares 28 parameters as suitable for the AA SA. In this work, the parameters by D. Mann [3]
are described as follows:
159
The numeration of the Table is taken from the list by D. Mann. The grey lines are the result of
adjusting and extending the parameters for an EA SMA WA following Lygin. The parameters
are mainly expanded for the production so that a series production is automatically in the focus
of the product development. With these input parameters, the contradiction matrix could be
built up. The 40 inventive (architecture) principles that Altschuller defines are used for
solutions in the matrix.
160
Therefore, the principles have to be designed in detail for an EA SMA WA. As a result of
Orloff [20], the 40 principles are enough to solve most problems. In combination with the
contradiction matrix, the engineer has a good base to search for a possible solution. The
“system conception” includes the first steps to create an EA SMA WA with a standardized
pre-evaluation and the parameters for the TRIZ- tools to get the first concepts. After these steps,
the possible solutions have to be developed into a concrete solution with the help of a
morphological box. After substantiating, the solution has to be rated to get a concept. The first
module “system conception” in the development approach is closed by different concepts.
The domain-specific design, which is addressed in the modules of material mechanics,
mechanics, electrical engineering and information process, has to be processed to create the
subsystems and the actuator in detail. Following the “system integration”, the domain-specific
design is connected with the whole system.
4. Conclusion and Outlook
This work adjusts and expands the V-model for SMAs based on Lygin [1] for a systematical
approach to develop an SMA actuator in the module “pre-evaluation” and in the “system
conception” based on an electrically activated SMA wire actuator. The pre-evaluation includes
a new tool to get a valid input for the product development. Therefore, the complex behavior of
the material is reduced to the relevant parameters. Subsequently, an interdependence matrix has
been built up that includes a rating with defined “impact factors”.
161
As a result, a network diagram which comprises the temperature as a relevant parameter helps
to rate the chance to find a solution for the problem. This new procedure helps to consider the
complex interdependencies in an early development stage. Further studies should investigate a
software tool that supports the pre-evaluation.
The system conception is based on a combination of the V-model by VDI 2206 and TRIZ-tools.
The two TRIZ-tools of the “innovation checklist” and the “contradiction matrix” are proved
and adjusted for an EA SMA WA.
Furthermore, this work expands the input parameters for the contradiction matrix. Therefore,
32 parameters are focused compared to Lygin to get a higher significance on the “production
parameters”.
Further studies should focus on the validation of the adjusted approach. Therefore, the
pre-evaluation tool has to be proven on real applications and solutions. To get a valid output of
the module “system conception”, a standardized system to rate the solutions of the
morphological box should be built up. Possible solutions for a standardized rating of the
concepts from the morphological box are identified in the method of VDI 2221 [21] or in
Pahl/Beitz [22]. A specialized or defined module for a target system could help to rate the
identified solution in a standardized approach.
5. Acknowledgement
The authors acknowledge the funding of the hyProFGA project by the EU through
Ziel2.NRW-programm Automotive+Produktion.NRW.
6. Reference
[1] Lygin, K: Eine Methodik zur Entwicklung von Umgebungsaktivierten FG-Aktoren mit
geringer thermischer Hysterese am Beispiel der Heizungs- und Klimatechnik. Dissertation,
Ruhr-Universität Bochum. Bochum 2014.
[2] Gausemeier, J., Moehringer, S.: "New Guidline VDI 2206-A flexible procedure model for
the design of mechatronic systems" DS 31: Proceedings of ICED 03, the 14th
International Conference on Engineering Design, Stockholm. 2003.
[3] Mann, D.: et al. Matrix 2003: Updating the TRIZ contradiction matrix. CREAX press,
2003.
[4] Jochen, E., Jaeger, C., et. al: Investments in climate friendly Germany. Study carried out by
the Ministry for the Environment, Agriculture and Reactor-protection. (In German:
Investitionen für ein klimafreundliches Deutschland. Studie im Auftrag des
Bundesministeriums für Umwelt, Naturschutz und Reaktroschutz.) Potsdam 2008
[5] Butera, F., Coda, A., Vergani, G.: Shape memory actuators for automotive applications. In:
Nanotec IT newsletter. Roma: AIRI/nanotec IT; 2007. p. 12–6.
162
[6] Abrahamsson P., Møster E.: Demands on shape memory alloys from the application
designer‟s point of view. J Phys IV France 1997;07. C5-667-C5-72.
[7] McWilliams A.: Smart materials and their applications: technologies and global markets.
BCC Research Advanced Materials Report; 2011. p. 161.
[8] Transparency Market Research: Electric Motors Market (AC motors, DC motors, Hermetic
motors) - Global Industry Analysis, Size, Share, Growth, Trends and Forecast 2013 - 2019
[9] Mertmann, M., Vergani, G.: "Design and application of shape memory actuators." The
European Physical Journal-Special Topics 158.1 (2008): 221-230.
[10] Welp EG, Breidert, J.: Knowledge and method base for shape memory alloys.
Materialscience and Technology. (In German: “Materialwissenschaft und
Werkstofftechnik”) 2004;35:294–9.
[11] Eggeler, G.: Final Report of CRC 459 (in German: Abschlussbericht SFB 459 in
Bearbeitung), in progress. Bochum 2014
[12] Langbein, S.: Local configuration and partial activation of the shape memory effect to
produce smart device structures (In German: Lokale Konfiguration und partielle
Aktivierung des Formgedächtniseffektes zur Erzeugung smarter Bauteilstrukturen) PhD
Thesis, Ruhr-University Bochum, Germany 2010
[13] Czechowicz, A.: Adaptive and adaptronic optimization of shape memory actuator systems
for automotive applications (In German Adaptive und adaptronische Optimierungen von
Formgedächtnisaktorsystemen für Anwendungen im Automobil). PhD Thesis,
Ruhr-University Bochum, Germany 2012.
[14] Seiffert, U.: Vieweg Book. Handbook of Automotive Engineering. (in German. Handbuch
Kraftfahrzeugtechnik.) Springer-Verlag, 2011.
[15] Datasheets of Smartflex, Dy90. Memry GmbH. Bochum 2015
[16] Stork, H.: Construction, modeling and control of Formgedächtnisaktorsystemen. Progress
Report VDI, number 8: Measurement and Control Technology, 1997 (In German: Aufbau,
Modellbildung und Regelung von Formgedächtnisaktorsystemen. Fortschritt-Berichte
VDI, Reihe 8: Mess- Steuerungs- und Regeltechnik) No. 657. Buhl
[17] May, G., Chrobok, R.: Prioritization of the entrepreneurial project portfolio, in: Magazine
- Management + Organization, (In German: Priorisierung des unternehmerischen
Projektportfolios, in: Zeitschrift Führung + Organisation)70th year, issue 2, pp 108-114.
2001
[18] Klein, N.: TRIZ/ TIPS. TRIZ / TIPS. Methodology of Inventive Problem Solving.
Oldenbourg, Munich 2002. (In German: Methodik des erfinderischen Problemlösens).
Oldenbourg „Wissenschaftsverlag“, München 2002
[19] Zobel, D.: Triz for all. The systematic way of solving problems. (In German: Triz für alle.
Der systematische Weg zur Problemlösung) 3rd edition. Expert Verlag, Renningen 2011
[20] Orloff, M: "Development of TRIZ." Inventive Thinking through TRIZ: A Practical Guide
(2006): 280-306.
163
[21] VDI Guideline: Systematic Approach to the Design of technical Systems and Products -
VDI 2221, VDI Verlag 2005
[22] Feldhusen J., Grote, K.: Engineering design: a systematic approach. Vol. 157. Springer
Science & Business Media, 2007. (In German: Konstruktionslehre. Methoden der
erfolgreichen Produktentwicklung. Springer Vieweg Verlag
164
SICASE-740
An Investigation Using Self-Sensing to Estimate the Lifetime of Shape
Memory Actuators
Christian Rathmann*, Benjamin Fleczok, Dieter Kreimeier
Ruhr University Bochum, Germany *[email protected]
Abstract
Today‟s industrial applications require smarter actuator systems, which can provide more
information. To achieve this, additional sensors are often necessary and thus the complexity,
cost and size of actuator systems increase. In this context, shape memory actuators are a
promising alternative due to their integrated sensing function that is able to measure the
respective positioning and environment temperature. Moreover, it is possible to predict the
fatigue of an actuator system and thus the lifetime of this system by electrical resistance.
Therefore, it is possible to establish highly integrated but simple actuator systems. Furthermore,
shape memory actuators can enable entirely new maintenance concepts.
The purpose of this paper is to investigate the electrical resistance as a parameter to predict the
fatigue of shape memory actuators. Firstly, an overview of possible applications for the
integrated sensing function will be given. Secondly, various factors which influence the
resistance of shape memory actuators will be outlined. Based on this, experiments are
performed and analyzed. The results will be discussed in terms of fatigue prediction and
suitability for maintenance.
Finally, this work presents the potential of integrated sensor function for new or hybrid
products based on shape memory alloy actuator systems. Future studies should consider further
factors influencing the resistance and improving the reliability to predict the fatigue of a shape
memory alloy based actuator system.
Keywords: shape memory alloy, sensor, self-sensing, resistance, lifetime prediction, fatigue,
experiments
1. Introduction
The attitude towards technical systems has changed dramatically in recent decades. In the past,
many companies in the field of mechanical engineering were highly integrated along the entire
value chain. However, the speed of technological progress requires companies to specialize,
which reduces the vertical integration of firms and value add. This effect is intensified by
changes in market conditions.
165
Companies and consumers are no longer able and willing to deal holistically with products.
Despite this fact, the regulatory requirements regarding safety increases.
This leads to a changing product understanding. Products must now, first of all, be easy to use
but also safe. In the field of mechanical engineering, this means, for example, the
manufacturing companies want to exclusively provide the production of goods but not the
maintenance or fault diagnosis of machinery. This significantly increases the number of sensors
required and thus the complexity as well as costs.
A promising approach to face this challenge are shape memory alloy based actuator systems
(SMA-ASs). SMA-ASs can be used as an integrated sensor-actuator system with condition
monitoring capability. Thus SMA-ASs contribute to reduce system complexity. In addition,
very user-friendly and economical products can be offered that enable new business models.
Although SMAs are promising, some questions remain to be answered. One of the most
important issues for the use of SMA-ASs as an integrated actuator-sensor system is the
question of how to provide the detectable characteristics and a prognosis accuracy or reliability
of the measurements. This paper makes a contribution to answer this question.
The paper is organized as follows: Section 2 gives an overview of the SMAs, including the
shape memory effect, advantages and disadvantages as well as the application of SMA-ASs. It
presents the behavior of electrical resistance of shape memory alloys as well as important
aspects of the resistance and fatigue of SMAs. Section 3 describes the experimental design
and set up. The results are discussed in terms of fatigue prediction and maintenance potentials.
The potentials of integrated sensor functions and thus predictability of lifetime of SMAs for
innovative, new products based on the findings are highlighted. Finally, section 4 closes by
sum up the key findings and thus how the paper makes a significant contribution within the
research field.
166
2. Shape Memory Alloys
Shape memory alloys (SMAs) belong to the group of active materials with direct coupling for
actuation [1]. After being deformed, SMAs have the ability to remember their previously
imprinted shape. Furthermore, they have an actuator function and show super-elastic behavior.
Figure 1 illustrates the different shape memory effects (SME) for actuation. The shape memory
effect can be distinguished into a mechanically triggered (pseudoelastic) and thermally
triggered effect. In the first case, the material retains its shape immediately after unloading [2].
In the second case, the material retains its shape by supplying energy. This is possible by the
use of the ambient temperature or resistance-based heating as part of electrical activation.
The SME is based on the phase transformation from martensite, the low temperature phase
austenite, the high temperature phase, which exhibits a temperature hysteresis (ΔTh) due to
inner friction [3]. The starting point of the phase transformation is the austenite start
temperature (As). Below As there is only martensite structure. The phase transformation starts
as soon as As is exceeded. It is completed once the austenite finish temperature (Af) is reached.
At Af there is only austenite structure. If the temperature decreases below a certain temperature
(martensite start temperature, Ms), the retransformation from austenite to martensite begins.
The retransformation is completed as soon as the martensite finish temperature (Mf) is reached.
Between As and Af or Ms and Mf both material structures coexist. The phase transformation is
illustrated in Figure 2.
167
SMAs can be characterized by their high energy density, high displacement and forces, high
corrosion resistance, and good biocompatibility [1, 4, 5]. Besides this, SMAs have a high
potential for miniaturization, have no abrasion of material, are electromagnetically compatible,
are characterized by their simplicity, lightweight and noiselessness [6].
Furthermore, SMAs have the self-sensing effect, which can be used for control purposes [1, 7,
8]. However, the knowledge about these alloys is still very limited [9].
That is why industrial product development is still resource-demanding and the application
potentials of SMAs are not fully exhausted.
Depending on their specific application, the limited temperature range of SMAs reduces the
number of application fields [10].
2.1 Application of Shape Memory Alloy Based Actuator Systems
The first industrial application of SMAs was a hydraulic tube coupling used in the high
pressure hydraulic system of the F-14 Tomcat fighter aircraft of the US Navy in the year 1971
[11, 12]. Due to their unique properties, SMAs are interesting for a wide range of applications,
especially valves, locking and unlocking mechanisms or vibration damping applications [1, 13].
Therefore, SMAs are suitable for a wide range of actuator applications, which are illustrated in
Figure 3: Applications of shape memory alloy based actuator systems Figure 3. SMAs are
especially interesting within the following areas: biomedicine, electronics and robotics,
aerospace, automotive and rail, consumer and household appliance, telecommunications and
photography, heating and air conditioning as well as engineering [14].
2.2 Behavior of Electrical Resistance of Shape Memory Alloys and Implications
The electrical resistance of material is defined through electrical resistivity and conductivity
and the volume (length and cross-sectional area). The electrical resistance changes with
temperature. With increasing temperatures of SMAs, ions are excited and vibrate around their
168
equilibrium position. As a result, the electron flow is disturbed and, consequently, the electrical
resistance increases.
The resistance behavior of SMAs is significantly influenced by the structural conditions. As in
all metals, the resistance increases with increasing temperatures. When reaching As, the
resistance behavior changes. Between As and Af the resistance behavior depends on structural
conditions. The crystal lattice of the austenite structure is more compact than that of martensite
structure. As the electrical conductivity of a material improves with the increasing number of
atoms per volume, the resistance decreases although the temperature increases further until Af
is reached. Beyond Af the resistance rises with an increasing temperature again, because the
phase transformation is completed. However, the resistances below As and beyond Af are not
identical.
Regarding the positioning of SMA actuators using the electrical resistance, various studies
prove this ability [8, 7, 15, 16, 17]. [17] proves that the behavior of the electrical resistance is
an indicator of the start and end of the phase transformation and thus actuator activation as well
as an indicator of the load condition. Additionally, Czechowicz [18] concludes that by
measuring the resistance of the material, the monitoring of fatigue is possible.
It can be stated that one important feature of SMAs compared to conventional materials is the
capability to monitor fatigue and to monitor the functionality by measuring the displacement.
By measuring the resistance, it is possible to draw conclusions about material conditions and
material history as well as displacement. Thus, countermeasures can be taken if the resistance
decreases below a certain value and possible breakdowns of the system may be prevented.
3. Experiments Investigating the Electrical Resistance to Predict the Lifetime of Shape
Memory Alloy based Actuators
3.1 Experimental Set Up
Figure 4 shows the experimental set up which has been used to investigate the electrical
resistance to predict the lifetime of shape memory alloy based actuators. The experimental rig
is able to run five experiments simultaneously. In total 10 samples have been used to predict
the resistance behavior of SMAs.
The samples are made from SMA wires type Smartflex90°C by memry an SAES Getters
subsidiary with an activation temperature of As 90° C, a diameter of 0.15 mm and a length of
100 mm. The SMA samples are activated by a power supply unit (EA-PS 3016-20B, EA
Elektro-Automatik GmbH & Co. KG). Constant current pulse is used because it is easy to set
up and all cycles get the same activation. The activation time is set to 1 sec. heating with a
current of 0.6 A and the cooling time is set to 8 sec.. This current has been determined by
169
preliminary experiments as the optimum current for a full activation. Eyelets are crimped at the
SMA wire ends, which is necessary to fasten them in the experimental rig. The SMA wires are
fastened in a mounting, loaded on a weight, and guided through an axial plain bearing.
The SMA samples are loaded at a constant force of 400 MPa. The power supply unit is
connected with alligator clips to the SMA wire eyelets. A measuring amplifier (MGC+ system,
HBM GmbH) records the measurements, which includes voltage, current, force and
displacement. The data from the measurement amplifier is transmitted to a measurement
computer and recorded by specific software (Catman 4.0, HBM GmbH) as well as saved for a
further analysis. A sampling frequency of 5 Hz is chosen for the measurement data. This
frequency is sufficient for the future analysis. The program itself saves at every hour, because
otherwise the file size is too large. After the wires are activated, they are cyclically loaded until
they break.
3.2 Results
The chosen parameters of 1 sec. heating with 0.63A, 8 sec cooling with the force of 400 MPa
are constant until the end of the wire lifetime. The lifetime differs from wire to wire, based on
slight material differences within one batch. For a valid interpretation of the resistance
behavior to determine the lifetime, the curves do not bear references to the cycle but to the
failure of the wire. This enables a cycle-independent interpretation of the lifetime.
The shape of the ten curves reveals a reproducible behavior with the typical standard deviation,
which means a gauss distribution. For a better overview, the average of the curves has been
established and plotted in Figure 5. As shown in array 1 in Figure 5, the electrical resistance
rises linearly by about 4.5% percent. The increasing gradient of array 1 depends on the wire
temperature as a result of the heating energy and time, the mechanical stress and the actuation
170
frequency. All curves have the same behavior, which means there is a critical linear increase of
the resistance before the parabolic section 2 begins.
The difference in the electrical resistance from the first value of the parabolic curve to the end
is about 15%. If array 2 is reached, the NiTi-SMA wire is in the final third of its lifetime. In
this section of its lifetime, other parameters like the displacement change, too.
Figure 6: has shown that the resistance behavior is almost identical to the behavior of the
displacement. Array 1 and array 2 are the same as in Figure 5. In the beginning the NiTi wire
has shown a linear drift of its displacement. In array 2 the shape of the curve changes to a
parabolic form. These experiments have shown the relation between the effects, causing the
171
drift of the displacement and the changes in the electrical behavior. The resistance behavior
results from the phase transformation and the material defects increasing over lifetime.
As shown in Figure 7, the changes of the first and last cycle in the voltage are very high. There
is a difference about 20%. Curve “b” represents the voltage of the last cycle and shows also a
small increase in the gradient while the wire is activated. This effect is based on the
overheating process due to the rising defects in the material and the grid vibrations.
As a result of the experiments, it has to be noted that there are two sections in the lifetime of
the wire. With these defined arrays, the lifetime of the NiTi-SMA wire is predictable. These
two arrays can be mathematically described and used to limit a functional range of the NiTi
SMA wire.
4. Potential of the Integrated Sensor Effect of Shape Memory Alloys for Innovative
Products
Product innovation is important for companies to compete in today‟s business environment. In
this context, industrial product service systems (IPSS) are of particular interest for SMA-ASs,
as they combine products and services. IPSS serve to customize products in order to create new
value for customers [19] and to strengthen the market position of a company or to create a
competitive advantage. These products strengthen companies‟ positions in the market because
they are able to offer solutions to market challenges, for instance, the increasing demand for
holistical solutions due to the ongoing specialization of companies. Moreover, the result is a
win-win situation for customers and producers or suppliers [20]. In conclusion, IPSS are a way
to avoid cost pressure and are a chance to differentiate from competition.
172
SMA-ASs allow companies to offer IPSS for actuator solutions due to their unique properties,
e.g. the self sensing effect or integrated sensor function. Referring to SMAs, this involves
analyzing and interpreting the electrical resistance as an indicator for the condition of an
actuator system regarding its functionality or fatigue. With SMA-Ass, simple integrated
actuator sensor systems are possible, which offer new functionalities and are connected to a
service [17]. It is possible to monitor properties like:
System errors in case the SMA-ASs (in form of a valve) are responsible for a function in a
complex system like the functionality of coolant circuit within a tooling machine
Operation errors including unpredicted factors like rising environmental temperature
Actuator displacement including uncompleted cycles
Operation conditions like loading or environmental temperature
Damage detection like the fatigue of material
Through IT support, simple programs can be developed that automate condition monitoring.
This is very convenient and efficient as well as transparent. Compared to conventional
materials, SMAs have significant potentials which are still not fully used. Expensive sensors
can be saved. As a result, system complexity decreases and cost savings can be realized.
The economical evaluation of a technical draft of a service-oriented shape memory actuator
system is discussed in [21]. In [14] maintenance processes and the resulting business models
based on condition monitoring done by SMA-ASs are presented.
5. Conclusion and Outlook
The paper has investigated the resistance behavior of SMA-AS on the example of SMA-wires.
Furthermore, the results indicate that due to rising defects in the material and the grid
vibrations over lifetime, the activation has to be modified to enhance the lifetime of SMA-ASs,
which can be done by control loops. It has been possible to show that the resistance behavior is
sufficient to predict the lifetime of an SMA-AS. This fact makes SMA-ASs interesting as
simple and cost-efficient integrated sensor-actuator systems. Additionally, this results in a large
number of potential applications of SMA-ASs for IPSS and thus can help companies to
develop new business models.
To put SMA-sensoring into practice, more research is needed. Further studies should focus on
the predictability of fatigue and the functionality of the self-sensing effect regarding the
influence of actuator geometries, actuator volume, stress, activation energy and environmental
issues. Furthermore, investigations are required to develop proper electronics. These
electronics have to consider only the requirements of SMAs and low-cost condition monitoring.
If costs are considered early, SMAs will appear in a wide range of applications. Designing
solid, automated processes is important for providing services based on SMAs. These
173
processes have to be mapped by software. Finally, a long-term test of SMA-AS under real
conditions is necessary.
6. Acknowledgment
The authors acknowledge the funding of the hyProFGA project by the EU through
Ziel2.NRW-programm Automotive+Produktion.NRW.
7. Reference
[1] Lagoudus DC. 2008. Shape Memory Alloys – Modelling and Engineering Applications.
New York: Springer Science+Business Media. ISBN: 9780387476841.
[2] Lygin K. 2014: A method for developing environmental activated SMA-actuators with low
hysteresis on the example of heating and air conditioning (in German Eine Methodik zur
Entwicklung von umgebungsaktivierten FG-Aktoren mit geringer thermischer Hysterese
am Beispiel der Heizungs- und Klimatechnik. PhD Thesis, Ruhr-University Bochum,
Germany.
[3] Liu Y, Kohl M. 2003. Shape memory micro valves with high energy density (in German
Formgedächtnis-Mikroventile mit hoher Energiedichte), research centre Karlsruhe:
science reports (FZKA 6934).
[4] Jani JM, Leary M, Subic A, Gibson MA. 2014. A review of shape memory alloy research,
applications and opportunities, Materials and Design 56, pp. 1078–1113.
[5] Treppmann D. 1997. Thermo-mechanical treatment of NiTi – including solutions for
quality assurance and standardization of shape memory alloys (in German
Thermomechanische Behandlung von NiTi (mit Lösungsansätzen für Qualitätssicherung
und Normung von Formgedächtnislegierungen). Fortschrittsbericht VDI series 5 Nr. 462.
ISBN: 3183462052
[6] Mavroidis C, Pfeiffer C, Mosley M. 1999. Conventional Actuators, Shape Memory Alloys,
and Electrorheological Fluids". In: Automation, Miniature Robotics and Sensors for
Non-Destructive Testing.
[7] Musolff A. 2005. Shape memory alloys – experimental study and design of adaptive
structures (in German Formgedächtnislegierungen – Experimentelle Untersuchung und
Aufbau von adaptiven Strukturen). PhD Thesis, TU Berlin, Germany.
[8] Schiedeck F, Mojrzisch S. 2008. Improvement of SMA Actuator performance using self
sensing, Proceedings of the International Conference on New Actuators,
ACTUATOR2008-P104, pp. 891-894.
[9] Gümpel P (2004) Shape memory alloys – applications in mechanical engineering, medical
technology and actuators (in German Formgedächtnislegierungen – Einsatzmöglichkeiten
in Maschinenbau, Medizintechnik und Aktuatorik). Renningen: expert press. ISBN:
3816922937.
174
[10] Kohl, M. 2002. Developing micro actuators made of shape memory alloys (in German
Entwicklung von Mikroaktoren aus Formgedächtnislegierungen). Habilitation treatise.
University Karlsruhe.
[11] Schetky L. 1979. Shape-memory alloys, Scientific American 241 (74-82).
[12] M. Wayman, J. Harrison, The origins of the shape memory effect, Journal of Minerals,
Metals, and Materials 41 (99) (1989) 26–28.
[13] Otsuka K, Wayman CM. 1998. Shape Memory Materials. Cambridge: University Press.
ISBN: 052144487.
[14] Rathmann C, Remmetz T, Kreimeier D. 2015. Maintenance of Shape Memory Actuator
Systems - Applications, Processes and Business Models, Proceedings of the 7th CIRP
Conference on Industrial Product-Service Systems, CIRP S-15-00155(accepted).
[15] Schiedeck F. 2009. Development of a model for shape memory actuators for dynamically
controlled operation (in German Entwicklung eines Modells für Formgedächtnisaktoren
im geregelten dynamischen Betrieb), PhD Thesis, Gottfried Wilhelm Leibniz University,
Hannover, Germany. ISBN: 978-3-941416-23-9.
[16] Janocha H. 2008. Adaptronics and smart structures – basics, materials, design and
application. Berlin: Springer Press. ISBN 978-3-540-71965-6
[17] Herrera GA, McKnight GP, Gao X, Johnson N, Browne AL. 2011. Use of intrinsic
electrical resistance changes in shape memory alloys as robust actuators state and fault
detection sensors., Proceedings of the ASME Conference on Smart Materials, Adaptive
Structures and Intelligent Systems, SMASIS2011-5000 , pp.1-8.
[18] Czechowicz A. 2012. Adaptive and adaptronic optimization of shape memory actuator
systems for automotive applications (in German Adaptive und adaptronische
Optimierungen von Formgedächtnisaktorsystemen für Anwendungen im Automobil). PhD
Thesis, Ruhr-University Bochum, Germany. Shaker Press. ISBN: 9783844014334.
[19] Kaluza B, Blecker T, Gemüden HG. 2006. Wertschöpfungsnetzwerke. Festschrift für
Bernd Kaluza. Berlin: Erich Schmidt.
[20] Winkelmann, P. 2012. Sales concepts and controlling – instruments of integrated customer
relationship management (in German Vertriebskonzeption und Vertriebssteuerung. Die
Instrumente des integrierten Kundenmanagements). Vertriebskonzeption und
Vertriebssteuerung.
[21] Rathmann C, Czechowicz A, Meier H. 2013. An investigation of service-oriented shape
memory actuator systems for resource efficiency, Proceedings of the ASME Conference
on Smart Materials, Adaptive Structures and Intelligent Systems, SMASIS2013-3065 ,
pp.1-8.
175
SICASE-753
Effect of Carbon Nanotube Composition to Enhance Wear Resistance and
Self – Lubrication Cylinder Liner Based on CNT – Al2O3 + 13% TiO2
Nanocomposite by Cold Spraying Method
Adita Evalina Fitria Utami*
Department of Metallurgy and Materials Engineering, Universitas Indonesia
Aulia Fitrian
Department of Metallurgy and Materials Engineering, Universitas Indonesia
Desak Putu Dewi Chrisnanda Indrayuni
Department of Metallurgy and Materials Engineering, Universitas Indonesia
Ernitha Yuniar Ghaisani
Department of Metallurgy and Materials Engineering, Universitas Indonesia
Carbon nanotubes (CNTs) Alumina + 13% TiO2 nanocomposite with different volume fraction
of CNT had been to determine the effect of CNT distribution in the ceramic matrix. The
behavior of self-lubrication on the surface coating on the cylinder liner investigated. Alumina
13% TiO2 + CNT nanocomposite containing CNTs Maximum 4% which was conducted by
cold spraying on the surface of the cylinder liner coating. Nanocomposite prepared with
different compositions with a planetary ball mill mixing method. Effect of wear resistance,
friction and self-lubrication will be analyzed with a Vickers Microhardness, Surface Roughness,
Wear Tribometer Pin - on - Disc before and after lubrication, and Fourier Transform InfraRed.
The results obtained wear resistance increases with the addition of CNTs. Wear resistance and
microhardness by the addition of CNTs calculated in this study. SEM - EDX was being used to
observe the surface of the coated CNT - Alumina 13% TiO2 nanocomposite. All results show
the method of spraying cold and planetary ball mill significantly improve the distribution of
CNTs in the alumina matrix 13% TiO2 thereby increasing wear resistance and self-lubricant
effect on the nanocomposite.
Keyword: carbon nanotube, tribology, ceramic nanocomposite, self – lubrication, cold spraying
176
1. Introduction
Related to the issue of global warming continue to be socialized, educational institutions
should be able to apply the principles of green technology and green industry in each
experiment, for the of modern materials components manufacture has to be able to make a
principled long life components parts or reuseable parts. To improve the performance of the
components of the material, the use of nanocomposite materials such as CNT – Alumina
coating to provide a mechanism of self-lubricant and bridging mechanism in alumina matrix
nanocomposite[1]
. Component material used in this experiment is piston, especially on cylinder
liner, generally constituent material used is cast iron, referring to the desired components have
a mild nature, and have a high wear resistance is used aluminum alloy (Al-Si) AC4B and after
specimen will be provided by coating the surface with material nanocomposite MWCNT
(Multi-Wall Carbon Nanotube) in a matrix of Al2O3, where both MWCNT and alumina in
powder form. The reason for choosing cylinder liner and piston ring to be coated with
MWCNT - Alumina because these components often fail, especially failures due to wear and
tear, a condition that often occurs when the current operational state, part piston ring and
cylinder liner experienced a blow that is high enough so that the solution of this problem is
usually with the use of lubricants or lubricant is high enough to reduce the wear failure[2]
.
However, the use of lubricants is worth noting because if the lubricant experiencing deficiency
in the second part will be mutually scrape the surface of one another resulting in leakage of
combustion gas that ultimately will fail all the components of the machine.
To form the CNT - Alumina nanocomposite there are various methods were used that method
Bulk Material (Spark Plasma Sintering, Hot Pressed Sintering, and Compression & Hot Rolling)
and methods Surface Coating (Dip Coating, Thermal Spray, Electroless Deposition, Laser
Deposition, and Cold Spray)[3]
. In this experiment, the method used is cold spray, where this
method is a kind of surface coating with spraying powder particles (CNT - Alumina) to the
surface of the specimen, the advantages of this method is to minimize the occurrence of
oxidation, grain coarsening and phase change. Before doing this process a powder mixture
previously done so that when sprayed distribution between CNT - Alumina powders evenly[4]
.
CNT powder Preparation - Alumina beginning of the process of dispersing CNT powder using
a surfactant PEG (Polyethylene Glycol) and the mixing process was conducted using a
planetary ball mill which aims to avoid the CNT forming aggregates. CNT powder preparation
process becomes very important because given the characteristics of lightweight CNTs are very
easy to form aggregates that will determine the outcome of the process of cold spraying.
2. Experimental Methodology
The first step in this experiment is to perform observations of the structure of the powder to be
used, which is also a major factor in the success of the study. Because it ensures that the
177
structure of alumina powder and CNT we have good is the first step in an experiment. Alumina
and CNT material characterization using high resolution SEM and EDS.
After doing the weighing powders used in the ratio of 0% CNT = 0 grams; 100% Al2O3 = 50
grams; 2% CNT = 1gram; 100% Al2O3 = 49 grams; 4% CNT = 2 grams; 100% Al2O3 = 48
grams.
Dispersing Treatment
In this experiment, we did not do a pre-treatment to the CNT powders that we use, because the
CNT powder we use is the type of MWCNT[5]
(Multi Walled CNT). The aim of dispersing
treatment is to prevent the CNT nano veils which does not undergo agglomeration when being
used for the illustrations can be seen in the following figure.
178
This experiment used surfactant in the form of PEG or Polyethylene glycol 4000 by using an
organic solvent DMF or Dhymethyl Formadide as dispersing agentnya, because PEG is not
soluble in water and can only be dissolved in DMF[6]
.
So the first stage is carried out in a dispersing treatment is to mix as much as CNT + PEG with
the composition 1: 3 + DMF as a solvent with a ratio of CNT 2% (1 gram) + PEG (3 grams) +
DMF 15 mL; CNT 4% (2 g) + PEG (6 grams) + 20 mL DMF.
Planetary Ball Mill for parameters used ball mill that is 50 grams of a mixture of as-dispersed
CNT + Alumina as-received (with composition as in Table 3) was added 25 mL ethanol by using
a rotation speed of 300 rpm (or about 5 Hz) for 4 hours, and balls used during the ball mill is
alumina ball (where the hardness of the ball must not be less than the hardness particles stirred).
Reason uses 25 mL ethanol or dry milling classified types
because there are significant differences in density between the CNT and alumina, which for
CNT density is 2.1 g / cm3 and alumina density ranges between 3.75 - 3.95 g / cm3)[7]
, wherein
the density difference The use of wet milling when the number reached about 240 mL ethanol
both the powder separately (where CNT is above and alumina at the bottom of the container / jar),
and cannot be mixed at all. Then using dry milling the powder is expected to be mixed well.
Function than the ethanol itself in the process of ball mill is to prevent sticky powder on the walls
of the jar, besides ethanol was also volatile so that post heating is needed after the mixing process
does not require high temperature and long duration. For a number of alumina balls were used I
use judgment based manual product a planetary ball mill device issued by Nuoya Machinery[8]
.
After the powder mixing process is completed, then we observed powder before dispraying,
whether the result of dispersing CNTs in a matrix of Alumina we have been successful or not.
The observation that we do is by using SEM against one Alumina powder-CNT, where the
results showed that the powder had been dispersed[9]
, although not completely dispersed. In the
figure 24 seen some of CNT that terdispers on A, B, C, and D, can be seen below:
179
After doing business with a surfactant dispersing and ball mill then the next is the process of
coating specimen using cold spraying method. Temperature processes used around 250-280 ° C,
with other similar parameters for all specimens. Prior to spraying with a mixture of
CNT-alumina powder, specimens performed pre-treatment such as sand blasting and coating of
Ni-Al bond coat.
After all specimens have dispraying, the characterization can be done. Where the first
characterization done is metallography (both macrostructure and microstructure) and also has a
micro hardness testing alumina-CNT nanocomposite coating them. For both observation
macrograph and micrographic metallographic, observed part of the specimen is part of the
cross-section and made these
observations suggested the most first before damaging parts specimens in subsequent testing.
Specimens divided into two parts by using medium speed diamond saw and done very carefully
so that layers do not experience damage from a process in which the two parts of the first part to
metallographic observation, SEM, micro hardness and the second part is treated wear test and
surface roughness testing. Specimens for metallographic observation macrograph first specimen
forms to record data before any treatment was given, or in other words, still pure after spraying.
Macro observation carried 2x 8x and 20x magnification. Here are some macro photos
transversely observed can be seen in the image below.
180
3. Results and Discussion
The Result of Wear Test[10]
with Lubricant and Non – Lubricant
181
Result and Discussion FTIR
Based on these data, it can be analyzed that an increase in the level of violence along with the
addition of CNTs ranging from the addition of 0%, 2%, and 4% hardness tested using Vickers
Microhardness Testing tool.
C4-1
Name
Sample 011 By Administrator Date Tuesday, April 07 2015
Description
4500 4504000 3000 2000 1000
100
7374
76
78
80
82
84
86
88
90
92
94
96
98
100
cm-1
%T
2921.31cm-1, 73.78%T
2852.78cm-1, 80.73%T
1458.01cm-1, 88.27%T
1376.93cm-1, 92.74%T
721.80cm-1, 95.53%T
182
Significant results shown by the degree of wear produced with increasing CNTs need to be
known as the progressive increase in violence, the wear rate will increase as well, it is known
that the material which causes the level of hardness and wear increase is Al2O3 CNT material
while giving the effect of self - lubrication, the effect of self - such lubrication known based on
results of testing using FTIR (Fourier Transform Infrared) the device to test whether or not the
levels of hydrocarbons in the material.
Conclusion
1. The addition of CNT improve the properties of hardness and wear on the surface of the
cynlinder linear
2. Al2O3 effects the increase in violence commited cold spray material
3. CNT self-lubrication effect by storing lubricant in the casing CNT
Reference
[1] M.D, Bermudez., et al. 2011. “Abrasive wear under multiscratching of Polystyrene +
SWCNT nanocomposites. Effect of sliding direction and modification by ionic liquid”.
Applied Surface Science, 257, pp. 9073-9081.
[2] Ahmad, I., et al. 2010. “Wear resistant properties of multi-walled carbon nanotubes
reinforced Al2O3 nanocomposites”. WEAR, 269, pp. 71-78.
[3] Zhang, W., et al. 2013. “Anti-Friction, Wear-Proof and Self-Lubrication Application of
Carbon Nanotubes”. Rev.Adv.Mater.Sci.36 pp. 75-88
[4] Sofyan, T.Bondan., et al. 2006. “Modification of Microstructure of Low Pressure Die
Casting AC4B Alloy by Addition of 0.015wt.% Sr”. 9th International QIR Proceeding. pp
1-5.
[5] Sarkar, Soumya., et al. 2014. “Processing and Properties of Carbon Nanotube/Alumina
Nanocomposites : A Review”. Rev.Adv.Matter.Sci.37, pp 53- 82.
[6] Ernst Geckeler, Kurt., et al. 2011. “Carbon nanotubes: are they dispersed or dissolved in
liquids?”. Geckeler and Premkumar Nanoscale Research Letters 2011, 6:136.
[7] Al-Samarai, Riyadh A., Haftirman, Khairel Rafezi Ahmad, Y. Al-Douri. 2012. “Evaluate
the Effects of Various Surface Roughness on the Tribological Characteristics under Dry
and Lubricated Conditions for Al-Si Alloy”. Journal of Surface Engineered Materials and
Advanced Technology, Scientific Research, pp. 167-173
[8] Manual Product Planetary Ball Mill : Nuoya Machinery
[9] Nam Kim, Gwi., et al. 2013. Evaluation on The Mechanical Characteristic of Fabricated
Alumina-CNT Composites. 9th International Conference on Fracture & Strength of Solids
June 9-13, 2013, Jeju, Korea
[10] CHEN, Jinfeng., et al. 2014. “Investigation on Premature Failure of the Self- Lubricated
Piston Rings in Oil-free Compressor”. 22nd International Compressor Engineering
Conference at Purdue pp.1-11.
183
SICASE-755
Development of Leucite Glass-Ceramics for Non-Metallic Dental Product
Siti Mazatul Azwa Saiyed Mohd Nurddin*, Malek Selamat
Mineral Research Centre, Mineral and Geosciences Department Malaysia *[email protected]
Abstract
The objective of the study was to determine the potential of using Malaysian silica sand as the
SiO2 raw material in producing leucite (SiO2-Al2O3-K2O) glass-ceramics. The crystallization,
mechanical and biological properties of the glass-ceramic was studied. A starting glass
composition in the system of leucite was melted in an electric furnace, quenched in deionized
water and dry milled to obtain glass powder. The glass powders were ball milled and
compressed to form 13mm x 10mm pellet. The thermal analysis, phase composition,
microstructure, flexural strength and cytotoxicity of the glass-ceramics were investigated.
Thermal analysis showed that crystallization of the glass occurred at the range of 650°C and
850°C. The pellets were sintered at 650˚C, 700˚C, 750
˚C, 800
°C and 850
°C for 1.0 hour. The
crystallization depends on the temperature and time of sintering. At 700°C, leucite began to
form with minor phase of sanidine. The peak intensity increased as the temperature was
increased up to 850°C. For sintering time 3 to 12 hours, the peak intensity of leucite and
sanidine were increased but microcline was formed as a minor phase. The microstructure
analysis showed that the presence of dendritic leucite and prismatic sanidine. The leucite
glass-ceramics appeared translucent. The flexural strength values (80–175 MPa) were
comparable with commercial product (112-140 MPa). The in-vitro bioactivity results proved
that the leucite glass-ceramics sample can be classified as a bio-inert and non-cytotoxity
material and can be used for restorative dental products.
Keyword: silica, glass-ceramics, leucite, sanidine, dental, bioactivity, bio-inert, cyctotoxicity
1. Introduction
Silica sand is the main raw material for production of silicates glass. Glasses in general term
are not advanced materials but through refining in composition, heat treatment and
manufacturing, glasses can be transformed into a new class of material called glass-ceramics.
Glass-ceramics are polycrystalline materials that can be produced through controlled of
crystallization during the heat treatment process. Currently, the application of glass-ceramics
as biomaterial for dental products is growing. For example, leucite glass-ceramics are widely
used in dentistry as restorative dental material to fabricate dental inlays, crowns, bridges and
veeners prostheses. Leucite glass-ceramics can be produced by surface sintering of glass
powder.
184
This processing route involves melting, quenching, milling of glass frit, and sintering in order
to promote crystallization of glass-ceramics. Glass-ceramics based on leucite system show
exceptional biocompatibility, and good physical, chemical and mechanical properties [1, 2].
Since the 1990s, efforts to develop biomaterials for restorative dentistry have been
concentrated on producing metal-free systems. An important milestone in this respect was
reached in the development of glass-ceramics containg leucite (K2O -Al2O3- SiO2)[3].
Glass-ceramics have been successfully used for many years in dentistry to construct crowns
and fixed partial bridges due to the properties of high mechanical strength, chemical inertness,
wear resistance, aesthetics and low density. Presently, leucite, mica and lithium disilicate
glass-ceramics are widely used as restorative materials [4]. This material is particularly
suitable for fabricating single units such as as dental inlays crowns and veneers because its
special optical properties [5]. The aim of this project was to investigate the effect of using
natural Malaysian silica sand as the SiO2 raw material on the phase crystallization,
microstructure, flexural strength, in-vitro bioactive and cytotoxicity of leucite glass-ceramics.
2. Materials and Methods
Natural silica sand sample was taken from Terengganu. .The sample was wet screened and a
fraction 75< to <150 µm was prepared in order to comply with the requirement of traditional
glass production technology for glass melts. Chemical composition of the silica sand sample
was determined by X-ray fluorescence analysis (Rigaku, Japan). .The composition weight
percent of the starting glass selected for the study was; 64.2% SiO2, 16.1% Al2O3, 11.9% K2O,
5.1% Na2O, 1.7% CaO, 0.5% TiO2 and 0.5% LiO2. All oxides used were with purity in the
range of 90 to 99% from Merck, Germany; Unilab, Australia and Hamburg Chemical. The
batches were placed in plastic bottles and mixed using a Heldoph Reax 2 Shaker, Germany for
8 hours in order to make sure the raw materials were homogenized. Zirconia balls were used
as the media in the mixing process. Samples were melted in alumina crucibles in a
bottom-loading high temperature furnace (Modutemp, Australia) at 10°C/min to 1450°C and
held for 3 hours, and then quenched in cool water to produce glass frit. The glass powders
were milled (Retsch PM-400, Germany) for 1 hour and sieved to the required size of less than
75 µm. The crystallization temperature of the glass powder was studied using Differential
Thermal Analysis, DTA (Linseis, Germany) with a 10oC/min heating rate at temperature from
25˚C to 1100
˚C in a dry air atmosphere. The glass powder was cold pressed using a
laboratory hydraulic hand press (Carver, USA) to obtain green compact in the form of
cylindrical pellet of 13mm x 10mm. Pellets were heat treated in electric furnace (Termo
Temp, UK) at 700˚C, 750
˚C, 800
˚C and 850
˚C at a heating rate of 2
oC/min and 1.0 hour soaking
time to study the crystallization behaviour of glass. The effect of sintering time on
crystallization was also studied through five different soaking time at 3.0, 6.0, 9.0 and 12.0
hours.
185
XRD (D8 Advanced , Bruker, Germany) was used for the identification of the crystalline
phases in the glass-ceramics using Cu Kα radiation at a scan speed of 2°/min for 2θ from 10
° to
80°. Microstructure studies were done using field emission scanning electron microscopy
(FESEM) (Supra 40 VP, Germany). Before viewing with FESEM, glass-ceramics specimens
were embedded, polished and etched with 30% hydrofluoric acid for 7 seconds. The
specimens were platinum coated and viewed using the secondary electron detector. The
flexural strength of sintered glass-ceramics was measured using three points bending test with
bars of 20 mm x 5 mm x 5 mm (INSTRON, UK, 0.5 mm/min displacement).
The biocompatibility of glass-ceramics sample after heat treatment at 850oC for 9.0 hour was
examined by in-vitro bioactivity and cytotoxicity test. The in-vitro bioactivity test was
conducted in Kokubo's simulated body fluid, SBF, which contains almost the same inorganic
constituent as human body plasma. The sample was brought into contact with SBF fluid for
10 and 20 days. XRD was used to characterize the formation of apatite on the glass-ceramic
surfaces. Cytotoxicity of the leucite glass-ceramics was evaluated by testing on extracts of
leucite glass-ceramics according to ISO 10993-5:2009(E) and ISO 10993-12:2012(E).
American Type Culture Collection L-929 mouse subcutaneous connective tissue fibroblast
cells (Mus musculus, NCTC clone 929, CCL-1) was used in this test. Zinc sulphate at 240
µg/ml and complete growth medium were used as the positive and negative control,
respectively. The test material was tested in triplicate at concentration of 6.25, 12.5, 25, 50,
100 and 200 mg/ml
3. Results and Discussion
Chemical analysis result of the silica sand is presented in Table 1. The result reveals that the
natural silica sand sample is highly pure and contain very low concentrations of impurities.
The content of Fe2O3 is low. Generally, typical acceptable value of Fe2O3 in silicate glass
product is between 0.02 – 0.03% (Malaysian Standard MS 701:1981).
The XRD pattern as shown in Figure 1 identifies that the main mineralogical content of the
silica sand sample is quartz. Quenching the glass melts in water at room temperature resulted
in transparent and colourless glass frit.
186
The XRD pattern of the glass sample as shown in Figure 1 indicates the present of amorphous
glass phase and did not show any evidence of crystalline phase in the glass. The DTA result of
the quenched glass powder is shown in Figure 2. The DTA was used to identify the glass
transition and crystallization temperature of the glass sample. The temperatures behaviour of
a glass can provide information of glass-ceramics fabrication. According to Sooksaen et al.,
the glass transition value of SiO2-Al2O3-K2O glass system is at about 594˚C to 638
˚C [6]. The
DTA curves show that no clear endothermic peak was observed related to glass transition.
However, an exothermic peak associated with the crystallization temperature of glass to form
glass-ceramic was in the range of 650˚C to 850
˚C. Consequently, heat treatment was
performed in this study at 700˚C, 750
˚C, 800
˚C and 850
˚C to form the glass-ceramics, assuming
phase evolution was completed at each isothermal hold of 1.0 to 12.0 hours.
7
The glass-ceramics appearance after sintering at 700°C was slightly translucent and white.
Increasing the sintering temperature up to 800°C has transformed the samples become more
white and opaque. The XRD patterns of glasses that were heat-treated at 700°C, 750
˚C, 800
˚C
and 850˚C for 1.0 hour are shown in Figure 3. The results indicated the presence of cubic
leucite (ICDD: 00-038-1423) and sanidine (ICDD: 00-019-1227). It is also proved that the
amount of leucite and sanidine phase for each heat-treated glass composition was increased
with the increasing of heat treatment temperature.
187
The effect of sintering time on the crystallization is shown in Figure 4. As the sintering time
was increased, all peaks associated with cubic leucite and sanidine became stronger. However,
for sintering temperature at 700°C and 750
°C and soaked 6 hours, microcline (ICDD:
00-019-0932) started to form as a minor phase. While for 800°C and 850
ºC microcline started
to crystallize within 3.0 hours soaking time. The crystallization of microcline was previously
identified in leucite reinforced glass-ceramics in SiO2-Al2O3-K2O glass system [7].
188
Figure 5 shows the FESEM micrographs of glass-ceramics sample sintered at 850oC for 1.0
hour and 9.0 hours. The formation of dendritic cubic leucite and prismatic sanidine phases is
clearly observed. A previous study by Holand, et al. on crystallization mechanisms in
glass-ceramics showed that the early stages of bulk leucite growth have been due to dendrites
growing in preferred crystallographic directions [8]. The micrographs also show there were no
obvious micro cracks and the surface developed a glassy texture.
189
Figure 6 shows the graph of the flexural strength results of the glass-ceramics sintered at 700oC,
750oC, 800
oC and 850
oC and soaked at 1.0, 3.0, 6.0, 9.0 and 12.0 hours. The flexural strength
of leucite glass-ceramics was increased as the time and sintering temperature was increased.
This is probably due to the high volume of crystalline phases and the existence of prismatic
sanidine. The higher flexural strength, 175MPa was achieved for sample sintered at 850°C for
9.0 hours. However, the flexural strength decreased after soaking for 9.0 hours at all
temperatures. This is probably due to the glass-ceramics started to melt and part of crystallite
began to dissolve in the residual glass phase. The flexural strength values of glass-ceramics
samples were comparable with commercial product of which is 112-140 MPa [9].
Glass-ceramics for restorative dental applications must fulfil the standard tests for biomaterial
use, such as compatibility with the oral environment. Bioactivity on the surface of the dental
restoration must not occur. XRD patterns of sintered glass-ceramics before and after
soaking in SBF for several days are shown in Figure 8. Patterns show that there is no
presence of apatite layer on the surface of glass-ceramics sample after 20 days of immersion in
SBF that indicated the sample is inert bioactive materials.
190
In this study, the MTT assay is used to evaluate the cytotoxicity characteristics of the leucite
glass-ceramics and the yellow tetrazolium salt (MTT) is reduced in metabolically active cells to
form insoluble purple formazan crystals which are solubilized by the addition of a solvent
(dimethyl sulfoxide). Cell viability is qualified by colorimetric enumeration whereby a low
OD reading corresponds to low cell viability which is associated with a loss in mitochondrial
dehydrogenase activity. The test material leucite glass-ceramics did not inhibit the viability
of L929 cells at all concentrations following 24-hour treatment as shown in Table 1 and Figure
1. Both negative and positive controls performed as anticipated. In MTT assay, the well
with the highest absorbance indicates the highest cell viability.
4. Conclusion
The highly pure natural Malaysian silica sand can be used as SiO2 source for producing leucite
glass-ceramics without any further chemical upgrading. Sintering of SiO2-Al2O3-K2O glass
powder at 700°C to 850
°C for duration of 1 to 12 hours contributed to the crystallization of
cubic leucite and sanidine with minor phases of microcline. The leucite glass-ceramics has a
flexural strength that is comparable with commercial dental products. The absence of apatite
layer on the surface indicated that it is an inert bioactive material. The leucite glass-ceramics
is ranked as a non-cytotoxic in terms of in-vitro cellular response to human cell lines under the
prevailing test conditions.
5. Acknowledgements
The authors would like to thank the Director of Mineral Research Centre, Department of
Minerals and Geoscience Malaysia and all staff of Mineral Research Centre who were
involved and had contributed in this research work.
6. References
[1] Ahmad N., 2006. Production of high purity silica from Malaysian silica sand, Phd Thesis,
University of Leeds, UK.
191
[2] Cattel, M.J. Knowles, J.C. and Clarke, R.L. 2005. The crystallization of an aluminosilicate
glass in the K2O-Al2O3-SiO2 system, J. Dental Materials, 21, pp. 811-822.
[3] Rheinberger, V., 1997. Perspectives in dental ceramics. Glastech Ber. Glass Sci. Technol.,
70C:339-400.
[4] Apel, E., Hoen, C.V., Rheinberger, V. and Holand, W., 2007. Influence of ZrO2 on the
crystallization and properties of lithium disilicate glass-ceramics derived from a
multi-component system. Journal of the European Ceramic Society, 27: 1571-1577.
[5] Lee HH, K. M., Asaoka K., 1997. Influence of modification of Na2O in a glass matrix on
the strength of leucite-containing poercelains. Dental Materials, 16: 134-143.
[6] Sooksaen, P. Boonmee, J. Witpathomwong, C. Likhitlert, S. 2010. Effect of K2O/SiO2
ratio on crystallization of leucite in silicate based-glasses, J. Metals, Mater. Minerals.
20(1) pp. 11-19.
[7] Chen X, C. T., Wilson R, Hill R and Cattell M., 2010. Crystallization of high-strength
fine-sized leucite glass-ceramics. J Dent Res, 89: 1510-1516.
[8] Holand, W. and Beall, G.H. 2002. Glass-ceramic technology. American Ceramic Society,
Ohio, USA.
[9] El-Meliegy, E.M. and Noort, R. 2011.Glasses and glass-ceramics for medical applications,
Springer.
192
Environmental Sciences
Session Chair: Prapat Pentamwa
SICASE-708
The Distribution of Aroids along Hau River, Vietnam
Duong Minh Truyen︱University Sains Malaysia
Mashhor Mansor︱University Sains Malaysia
Amir Shah Ruddin︱University Sains Malaysia
Duong Minh Tho︱Cantho University
Nguyen Trung Toan︱Hong Bang International University
SICASE-805
Heavy Metals Contamination of Community Water Supply in Northeastern Thailand
Prapat Pentamwa︱Suranaree University of Technology
Tanisa Wongklom︱Suranaree University of Technology
Jitlada Shimmai︱Suranaree University of Technology
SICASE-843
Preparation of Sodium Aluminate from Alumina Waste from Petrochemical Industry
Vilaiwan Teankrajang︱Chulalongkorn University
Orathai Chavalparit︱Chulalongkorn University
Rewadee Anuwattana︱Technology Thailand Institute of Scientific and Technological
Researcher
SICASE-840
Electrochemical Regeneration of Chloride Contaminated NaY Zeolite Wastes from
Petroleum Refinery
Kavidsara Homchuen︱Chulalongkorn University
Orathai Chavalparit︱Chulalongkorn University
Rewadee Anuwattana︱Ministry of Science and Technology Thailand
2015/6/28 Sunday 10:45-12:15 Room 2
193
SICBENS-1326
Environmental Hazards of Radioactive Radiation from Granite Building Stones through
a Look into Burien’s Granite Mass in Isfahan
Siyamak Bagheriyan︱Islamic Azad University of Shahriyar
194
SICASE-708
The Distribution of Aroids along Hau River, Vietnam
Duong Minh Truyen*
School of Biological Sciences, University Sains Malaysia, Malaysia
Mashhor Mansor
School of Biological Sciences, University Sains Malaysia, Malaysia
Amir Shah Ruddin
School of Biological Sciences, University Sains Malaysia, Malaysia
Duong Minh Tho
Cantho University, Vietnam
Nguyen Trung Toan
Hong Bang International University, Vietnam
Abstract
The Araceae is a monocotyledonous family group of flowering plants and is one of the largest
families in the world after the orchids, grasses and sedges. The aroids play important roles in
human life such as supplement of food sources, ornamentals and medicine. The sampling stations
were selected based on the analysis of the main groups of population zonations and land uses
of the Hau River. A total 27 species found, belonging to 23 genera in Araceae family can be
divided into three main important groups such as ornamentals, weedy and crop cultivation.
Generally, Hau River recorded 16 ornamental species, 11 weedy species and 3 crops species.
Surprisingly, no endemic species is recorded in Hau River. Among the aroid species,
Colocasia esculenta been a common species at Hau River, used as food for the local
communities along Hau River.
Keywords: “Araceae”, “Hau River”, “distribution”, “plant – use”, “aroid species”
195
1. Introduction
The family Araceae is one of the common monocotyledonous flowering plants in the world and
has a total of 117 genera and more than 3790 species [1]. According Mansor, Boyce [2], Araceae
is also one of the largest families in the world after the orchids, grasses and sedges. Most of the
species are found in tropical areas. Sulaiman and Boyce [3] made the observation that the aroids
generally occur along the riverbanks, wetlands and close to humid areas where the
environmental conditions are wet, with high humidity and shaded sites.
There are several factors affecting the distribution of aroid plants. The first factor is growth of
human population on a large scale and the migration to new areas as population pressures build
up. With a high density population occupying the land, the land use changes invariably takes
place as natural ecosystems are converted or modified for agriculture and human settlements.
Secondly, aroid distribution is effected by their economic values as cultivated species. Species
with commercial values are cultivated or collected in the wild. This action changed their natural
habitats and distribution pattern. Finally, climate and topography influence aroid distribution
when the aroid species are introduced. Humid climate increases aroid growth while too dry
climate restricts species growth. Topography of flood plains and the low lands in Mekong Delta
are suitable for some species and their distribution expand for some particular species, namely
Colocasia esculenta. Nevertheless, mountains, hills and high lands are suitable for the climbing
group or special group (Amorphophallus).
Through the survey of aroids distribution in Hau Rivers for this study, the species presence and
the habitats in surveyed areas are recorded and assessed. Moreover, common species such as
Colocasia esculenta is expectedly found in all types of land-use as this species has a very
cosmopolitan distribution. This will provide a supplement to future research in Mekong Delta in
the next years.
2. Methodology
Hau River is located in Mekong Delta, a highly disturbed area, which is the lowest part of
Mekong River with an extremely high human population densities. Mekong Delta has about 18
million inhabitants (40,000 km2). The high population densities in Mekong Delta resulted in
the conversion of most of parts of the delta into agro – ecosystems. Hau River reflects a highly
modified environment. Rice fields cover almost 80% of the area.
The study has been conducted along Hau River, using land use maps and soil maps in order to
survey along riverbanks. Observation is the technique of visually assessing an area, and was
used for sampling along Hau River. This is based on an understanding of the ecology of the
area and noting the various features related to the objectives of the study. Based on the Mekong
196
Delta population map, can be divided into 3 zones (lowly populated, moderately populated,
highly populated) [4] and [5] (figure 1).
Selection of the sampling sites were based on the surveyed land uses. In each land use type, the
sampling points (1km x1km) were established. In 1km2 sampling area, subplots (1mx1m) were
recorded for aroid species presence taking into account species composition and distribution.
During the survey, photos were taken on all aroid species found in every study site for further
reference.
197
3. Results
3.1 Checklists of Aroids and Its Distribution in Hau River.
There are 27 species distributing along Hau River, belonging to 23 genera in Araceae family as
shown in Table 2. Generally, aroids can be divided into three main important groups such as
ornamentals, weedy and crops cultivation. Among aroid species found, three species are
identified as weedy and crops, namely Caladium bicolor, Colocasia antiquorum and Colocasia
esculenta. Hau River recorded 16 ornamental species, 11 weedy species and 3 crops species
based on endemistism and class groups. Surprisingly, no endemic species is recorded in Hau
River.
The numbers of aroids in each zone of Hau River are shown in Table 3. The highest number of
species recorded is at zone 2 Hau River with 22 species. Followed closely by zone 3 Hau River
(19 species), zone 1 Hau River (18 species).
198
Based on the results, 10 species were recorded in three zones of Hau River, 12 species in two
zones and 5 species in one zone. In zone 2, only four species found, namely Aglaonema
nitidum, Philodendron erubescens, Pseudocracontium lacourii and Scindapsus officinalis.
Only Syngonium macrophyllum was recorded in zone 1 in Hau River.
3.2 Species Occurrence for Landuse Types
The numbers of aroids in each landuse of Hau River are shown in Table 4. The highest number
of aroid species have been recorded at crops landuse with 19 species, followed closely by
residential areas landuse (17 species), plantation (16 species), others landuse (6 species),
unused land (5 species) and forest (2 species).
199
Based on the results, it is found out that only Dieffenbachia amoena can be found at all landuse
types as shown in Table 4. This been followed by Colocasia esculenta, recorded at 5 landuse
types along Hau River. In addition, 11 species been recorded at 3landuse types, 8 species at 2
landuse types. There were 6 species that were recorded at one landuse type. Meanwhile, out of
this number one species such as Aglaonema hybrid recorded at crops landuse, Philodendron
erubescens at residential areas, Pistia stratoides and Typhonium flagelliforme at other types of
landuse, Pseudocracontium lacourii at plantation and Spathiphyllum patinii at unused land
(Table 4).
3.3 Species Diversity Index between Different Populated Zones in Hau River
The species diversity index of the three zones in Hau River is shown in Table 5. The highest
number of Shannon index is shown in highly populated zone with 1.68, followed by lowly
populated zone with 1.46 and lastly at moderately populated zone with 0.76. In which, highly
populated zone is more diverse than lowly populated. Moderately populated zone is less
diverse than other zones.
200
On the other hand, the highest Evenness index is in moderately populated zone, followed by
lowly populated and highly populated zones, with 0.91, 0.79 and 0.77 respectively.
3.4 Species Diversity Index between Landuses in Hau River
Table 6 reflects the diversity between landuse types in Hau River. The diversity indices are not
too high in Table 7. However, the indices numbers show the diversity of each landuse in Hau
River. Meanwhile, crops landuse has the highest Shannon index (1.49). The second high
Shannon index is shown in residential area, followed by plantation with 0.79 and 0.62
respectively. Unused land and river have smaller Shannon indice (0.30 and 0.32 respectively).
Nevertheless, the Evenness indices are shown in Table 6. The highest Evenness index is at
unused land (0.90), followed by three types of landuse, namely crops, plantation and residential
areas with 0.80 values. Evenness index at the forest is 0.55 while at other types of landuse is
0.35
4. Discussion
In Vietnam, researches on Araceae are limited as there are not many scientists working on this
family. Until 2012, Matthews [6] published aroid checklist with 135 aroid species of 25 genera. In
comparison, 27 aroid species were found along Hau River, having 20% of the species in Vietnam.
The number of aroid species along Hau River does not represent completely the distribution of
aroids of the Mekong Delta. However, 27 species is quite numerous in the smaller area in the Hau
River. Furthermore, in the Vietnam aroid checklist, there are no endemic species found in Hau
River even though there are 40 endemic species in Vietnam [6]. Most endemic species in Vietnam
have been recorded in natural habitats such as forests and mountainous places in the upper parts of
201
Vietnam. For example, Arisaema claviforme, a new aroid species was found at the Hon Ba Nature
Reserve, Khanh Hoa province, further to the north [7]. However, in Mekong Delta, there are no
primary forests and hilly areas left. Although the forests are located in moderate populated zone in
Hau River, they were used mainly for tourisms. Many artifacts such as pagodas, stores and shops
have replaced the natural habitats in forests even in moderately populated zone. In highly
populated zones, there are no more natural places and the main landuses are agricultural fields and
residential areas. This explains why there are no endemic species recorded in the Hau River. In
addition, according to Toan, Sebesvari [8], the main feature is the people‟s activities along
riverbanks. Nowadays, the main activities in Mekong Delta are intensive farming and secondary
vegetation [8]. Because of human actions, the natural systems are replaced by agricultural
farming [9]. Cultivated crops had replaced wild plants. Native and endemic species cannot be
found in the Mekong Delta because of the loss of their wild habitats and changed environmental
conditions. In addition, Aroid species in Hau River are cultivated for humans‟ demands. Total area
of Mekong Delta is 40,548.2 km ² while the population of Mekong Delta is 17,330,900 people [10].
With high population, the demands to use and cultivate Araceae species in Vietnam are also high.
Aroids presences according to landuse types are different in Hau River. Species are widely
distributed as food crops, residential areas and plantation landuses. In Mekong Delta, most of
people activities focus on both sides of the rivers such as farming, fishing and construction [11].
Densely populated areas are close to big rivers. This is because the presence of big rivers makes
it convenient for agriculture, irrigation and drainage. Hence, people focus on changing wasteland
along riverbanks into cultivated areas. This explains why those landuses above are more diverse
than other types of landuse in Hau River. Moreover, the overall Shannon diversity index in Hau
River is 1.15. Although 27 aroid species recorded in Hau River, the environment is highly
disturbed by human population‟s increase and urbanization. Due to human disturbance, the
habitats in Hau River are no longer natural. That makes Hau River less diverse. Nevertheless,
overall Evenness index in Hau River is 0.80. Because the main landuses in Hau River are
agricultural fields, aroid species are planted in specific farming field due to the demands of human.
Between landuse types, in Hau River, the Shannon index in crops landuse type is the highest,
followed by residential areas with 1.49 and 0.79 respectively. In Hau River, people cultivate
aroid species for food crops, feeding animals, ornamentals and medicine. This explains why
cultivated places will be more diverse than other places in Hau River. In addition, ornamental
plants are often planted around housing areas. It makes the Shannon index in residential areas is
higher than other landuse types. Nevertheless, Evenness index in unused land is highest. This is
because in unused land, with little disturbance from human activities, wild species occur widely
and naturally, including weedy and cultivated plants.
202
5. Conclusion
There were 27 species in the Hau River. Out of this number, 16 species recorded as
ornamentals, 11 species as weeds and 3 species as crops. No endemic species is recorded in
Hau River. The overall Shannon index of Hau River is 1.15 while overall Evenness index is
0.80. Some common species adapt to different habitats from upper to lower zones of two rivers,
such as Colocasia esculenta. Furthermore, study the ecology and the distribution of aroids in
larger areas to record more endemic species for Mekong Delta, Vietnam.
6. Acknowledgements
I would like to express my gratitude towards Universiti Sains Malaysia for funding the
publication fee under the Individual Research Grant (1001/PBiologi/816217).
7. References
[1] Nauheimer, Global history of the ancient monocot family Araceae inferred with models
accounting for past continental positions and previous ranges based on fossils. New Phytol,
2012. 195(4): p. 938-50.
[2] Mansor, M., et al., The Araceae of Peninsular Malaysia. Academic imprint series / Penerbit
USM2012, Glugor , Pulau Pinang: Penerbit Universiti Sains Malaysia : Distributors MPH
Distributors. ix, 146 p.
[3] Sulaiman, B. and P.C. Boyce, Studies on Homalomeneae (Araceae) of Peninsular Malaysia I :
Homalomena asmae, a New Species from Perak. APG : Acta phytotaxonomica et
geobotanica, 2010. 60(3): p. 163-166.
[4] Khang;, N.D., A. Kotera;, and M. Yokozawa, Influence of Climate Change on Water Regime
and Rice Production in Mekong Delta. Vietnam Academy for Water Resources 2011.
[5] Hoi, T.N., Typical Flood and Flooded Areas in Mekong Delta. Vietnam Academy for Water
Resources, 2011.
[6] Matthews, P., Bibliography of work on Araceae in Vietnam. International Aroid Conference
2012.
[7] Ho, P.H., Cay co vietnam = An illustrated flora of Vietnam1991, California: Pham Hoang Ho.
v. <1 >.
[8] Toan, P.V., et al., Pesticide management and their residues in sediments and surface and
drinking water in the Mekong Delta, Vietnam. Science of the Total Environment, 2013. 452:
p. 28-39.
[9] Duong, L.T. and K. Cho, Farm-Management and Farming Systems-Development in Vietnam.
Journal of the Faculty of Agriculture Kyushu University, 1994. 39(1-2): p. 1-8.
[10] Wilby, A., et al., Arthropod diversity and community structure in relation to land use in the
mekong delta, vietnam. Ecosystems, 2006. 9(4): p. 538-549.
[11] Wood, J.F.a.J.L., Dry-Season Flood-Recession Rice in the Mekong Delta: Two Thousand
Years of Sustainable Agriculture? Asian Perspectives, 1999. 38(1).
203
SICASE-805
Heavy Metals Contamination of Community Water Supply in Northeastern
Thailand
Prapat Pentamwa *, Tanisa Wongklom, Jitlada Shimmai
School of Environmental Health, Institute of Medicine,
Suranaree University of Technology, Thailand *[email protected]
Abstract
This study aims to investigate the contamination of heavy metals from community‟s water
supply in Nakorn Ratchasima, Chaiyaphum, Buriram and Surin provinces of Northeastern
Thailand by evaluating specific substances of Cadmium (Cd), Lead (Pb), Iron (Fe), Zinc (Zn),
Manganese (Mn), Arsenic (As) and Mercury (Hg). The Cd, Cr, Pb, and Zn compounds were
quantified by Flame AAS and only As and Hg were analyzed by ICP-MS method. The human
health risk assessment for the consumer water communities was assessed according to the
USEPA (2005) by investigating non-cacinogenic risk and the cancer risk method. Most of
heavy metals (Cd Pb Fe Zn Mn) were in water quality acceptable standard except for Fe in tap
water has exceeded the standard by 6.3 percent. The most detected of ferrous (Fe)
contamination in surface water and tap water was found in Nakorn Ratchasima surface water
supplies of 1.77 and 1.66 mg/L, respectively, which is higher the Fe drinking water standard
of 0.5 mg/L. Thirty percent of Cd in ground water supply in both water source and home users
is higher than the standard. In Chaiyaphum, Burirum, and Surin provinces, Cd contaminations
in groundwater water supplies at water sources and home user was higher than the water
standard of 2 and 1.33, 2 and 2.6, and 2.3 and 1.6 times, respectively. This study showed
Pb Hg and As were not detected in any water samples. However, the hazard quotient (HQ)
average of targeted heavy metals are in the acceptable non-carcinogenic risk.
Keyword: Heavy Metals, Contamination, Water supply, Communities, Thailand
204
SICASE-843
Preparation of Sodium Aluminate from Alumina Waste from Petrochemical
Industry
Vilaiwan Teankrajang*
Department of Environmental Engineering, Chulalongkorn University, Thailand
Orathai Chavalparit
Department of Environmental Engineering, Chulalongkorn University, Bangkok, Thailand
Rewadee Anuwattana
Department of Environment and Resources, Technology Thailand Institute of Scientific and
Technological Researcher, Thailand
Abstract
In this research, alumina waste from the petrochemical industry is used to absorb moisture
from the air in laboratories. Nowadays, alumina waste increased, especially from
petrochemical industry which occurred about 37.5 tons / year. The most entrepreneurs didn‟t
seriously intend about the alumina waste management because the desiccant alumina with low
price made the amount of alumina waste disposal increased. An alternative environmental
management method of alumina waste is used to be the starting raw material for preparation of
sodium aluminate. The aim of this research is to determine the optimum conditions for
preparation sodium aluminate from alumina waste from petrochemical industry. Sodium
aluminate was prepared by the dissolution of alumina waste in a sodium hydroxide. The
amount of alumina waste and the concentration of sodium hydroxide were adjusted to the mole
ratio of sodium oxide and alumina (Na2O: Al2O3). The results showed that the optimum
concentration of sodium hydroxide was 6M. The mole ratio of Na2O: Al2O3 as 2:0.25, 2:0.75
and 2:0.9 has the XRD spectra similar to sodium aluminate standard graph and no interruption
from alumina waste. The ratio 2:0.75 has the highest percentage of sodium aluminate crystal
(93.62%). The structure of ratio 2:0.75 mostly consisted of orthorhombic and tetrahedral
crystals. The XRF data are approximately similar to the sodium aluminate standard. Then, the
optimum condition showed that 6M of sodium hydroxide activated with the mole ratio of Na2O:
Al2O3 was 2:0.75 as the best condition for preparation of sodium aluminate from alumina waste
from petrochemical industry.
Keywords: alumina waste; sodium aluminate; sodium hydroxide; waste recycling
205
1. Introduction
Activated alumina is manufactured from aluminium hydroxide by dehydroxylation that
produces a highly porous and surface area. It is used for a wide range of adsorbent and catalyst
applications. Activated alumina is used as a desiccant, it works by a process called adsorption.
The water in the air will attached to the alumina as the air passes through it [1]. Alumina is
used to absorb moisture from the air to control humidity in laboratories because the moisture
inside the building is a source of bacteria and fungi. Nowadays, alumina waste increased,
especially from petrochemical industry which occurred about 37.5 tons / year. The most
entrepreneurs didn‟t seriously intend about the alumina waste management because the
desiccant alumina with low price made the amount of alumina waste disposal increased. The
conventional disposal method is by incineration (with cement) and sanitary landfill. However,
the cost of landfills disposal and energy recovery in cement kiln are expensive. The waste
management was based on the 3Rs principles (reduce, reuse and recycle) to maximize the use
of resources which there were many options for this management. An alternative
environmental management method of alumina waste is used as raw material for preparation of
sodium aluminate.
Sodium aluminate is a white crystalline solid and prepared by reaction of sodium oxide (Na2O)
and alumina (Al2O3). Sodium aluminate is an important commercial chemical for various
applications. It is used as an adjunct to water softening systems, as a coagulant aid to improve
flocculation and a pH adjustment for wastewater treatment [2]. It was also a catalyst for
concrete hardening, a production of refractory bricks and a chemical flocculation. It can be
treated phosphate and fluoride in the wastewater [3].
The aim of this research is to determine the optimal conditions for preparation of sodium
aluminate from alumina waste from the petrochemical industry.
2. Material and Methods
2.1 Preparation of Material
The alumina waste used in the study was obtained from laboratories humidity control unit in a
petrochemical industry located in Mab Taput industrial state. The alumina waste sample was
average 5 mm in diameter. Chemical analysis of alumina waste can be achieved by X-Ray
Fluorescence (XRF), Phillip PW2400. The result showed in Table 1. It was dried at 105 oC
for 24 h using an air-circulating oven and ground using grinder and sieved to particles with a
top cut of 2 mm.
206
2.2 Preparation of Sodium Aluminate
Experimentation was to find the optimum condition for preparation of sodium aluminate from
alumina waste by varies the ratio of alumina waste and sodium hydroxide concentration. The
alumina waste and the concentration of sodium hydroxide were varied from 5-40 g and 1-5
mol/L, respectively. The alumina waste and sodium hydroxide were mixed into 1000 ml beaker.
This solution was heated at 60 ° C and stirred at 200 rpm for 30 minutes. The amount of
alumina waste and the concentration of sodium hydroxide were adjusted to the mole ratio of
sodium oxide and alumina (Na2O: Al2O3). Then, the mole ratio of sodium oxide and alumina
from the sodium aluminate standard was 2:1 and the concentration of sodium hydroxide was
varied at 2M, 4M, 6M, 8M and 10M. Finally, the molar ratio of sodium oxide and alumina was
varied at 2:0.1, 2:0.25, 2:0.5, 2:0.75, 2:0.9, 2:1, 2:1.5, 2:2, 2:2.5 and 2:3 by the optimal
concentration of sodium hydroxide [4]. The crystallization, chemical compositions and
morphology of sodium aluminate were analyzed by X-Ray Diffraction (XRD) method, X-Ray
Fluorescence (XRF) method and Scanning Electron Microscopy (SEM) method respectively.
3. Result and Discussion
For the varied concentration of sodium hydroxide which the mole ratio of sodium oxide and
alumina as 2:1, the results showed that the optimum concentration of sodium hydroxide was 6M.
For crystallization of sodium aluminate, the crystal sodium aluminate obtained from the
experiments was the molar ratio of sodium oxide and alumina as examined by the XRD spectra
as shown in Fig.1. From the XRD spectra can be observed that ratio 2:0.25, 2:0.75 and 2:0.9
similar to sodium aluminate standard graph and no interruption from alumina waste. Therefore, the
sodium oxide concentration and initial alumina concentration were an important role in the
crystallization of sodium aluminate [5,6].
207
The results of the percentage of crystallization were showed in Table 2. The highest percentage of
sodium aluminate crystal was 93.62 which obtained from sample with ratio 2:0.75. Moreover, it
obviously showed that the percentage of crystallization trend to decrease when the molar ratio
of sodium oxide and alumina decrease.
The chemical composition by XRF of mole ratio 2:0.75 is mostly similar to the sodium
aluminate standard which compared with other ratios.
208
The chemical composition by XRF is 2:0.75 (Na2O: Al2O3), the main composition of alumina
(50.5%) and sodium oxide (34.6%) which is almost close to the sodium aluminate as shown in
Table 3.
The fig.2 showed that the morphology of the sodium aluminate with ratio 2:0.75 was
investigated by using SEM The structure of sodium aluminate mostly consisted of orthorhombic
crystals which is approximately similar to the sodium aluminate standard.
209
4. Conclusion
The result of the study showed that alumina waste from laboratories control unit in a
petrochemical industry can be recycle as the raw material for preparation of sodium aluminate.
The optimum concentration of sodium hydroxide and the mole ratio of Na2O:Al2O3 was 6M
and 2:0.75 respectively. The result from physical and chemical compositions of sodium
aluminate prepared from alumina waste is similar to sodium aluminate standard.
5. Acknowledgments
This paper was financially supported by the 90th
anniversary of Chulalongkorn University fund
(Ratchadaphiseksomphot Endowment Fund), Thailand. The author was supported from TISTR
for the use of laboratory instruments.
6. References
M. Takbiri, K. J. Jozani, A. M. Rashidi, and H. R. Bozorgzadeh, Preparation of nanostructured
activate alumina and hybrid alumina-silica by chemical precipitation for natural gas
dehydration, Microporous and Mesoporous Materials, 2013, vol. 182, pp. 117-121.
V. Rayzman, I. Filipovich, L. Nisse, and Y. Vlasenko, Sodium aluminate from alumina bearing
intermediates and wastes, JOM, 1998, vol. 50, pp. 32–37.
R. J. Keller, and J. J. Len, Kirk-Othmer Encyclopedia of Chemical Technology, 2003.
W. G. Michael, and D. S. Darryl, Zeolites Synthesized from Class F Fly Ash and Sodium
Aluminate Slurry, American Ceramic Society, 1997, vol. 80, pp. 2449-2453.
Y. F. Zhang, Y. H. Li, and Y. Zhang, Phase diagram for the system Na2O–Al2O3–H2O at high
alkali concentration, Journal of Chemical and Engineering Data, 2003, vol. 48, pp. 617–
620.
210
C. Shaotao, Z. Yifei, and Z. Yi, Preparation of sodium aluminate from the leach liquor of
diasporic bauxite in concentrated NaOH solution, Hydrometallurgy, 2009, vol. 98, pp.
298-303.
211
SICASE-840
Electrochemical Regeneration of Chloride Contaminated NaY Zeolite
Wastes from Petroleum Refinery
Kavidsara Homchuen*
Department of Environmental Engineering, Chulalongkorn University, Bangkok, Thailand
Orathai Chavalparit
Department of Environmental Engineering,
Chulalongkorn University, Bangkok, Thailand
E-mail address: [email protected]
Rewadee Anuwattana
Environment and Resources Technology Department, Thailand Institute of Scientific and
Technological Research Ministry of Science and Technology Thailand, Thailand
Abstract
The aim of this study was to investigate the chloride desorption efficiency of NaY zeolites
waste from petrochemical factory using an electrochemical cell assembled with a Ti/Pt as
electrodes. Current density, initial pH of water and retention time periods were varied in the
range of 4 to 100 mA/cm2, 5 to 12 and 5 to 30 minutes, respectively. The results showed that
the chloride desorption efficiency could be achieved at pH 10, current density of 20 mA/cm2
and reaction time at 20 minutes. At the optimum condition, the chloride remove efficiency was
higher than 71% and chloride residual in regenerated zeolite was equal to 7.03 mg chloride/g
zeolite. The structure of regenerated NaY zeolites analyzed by XRD technique was closely to
commercial NaY zeolites. It was observed, the surface area and pore volume of the regenerated
NaY zeolites were increased compared to the NaY zeolite wastes. Therefore, it can be
concluded that NaY zeolites contaminated with chloride can be regenerated by using an
electrochemical process.
Keywords: NaY zeolite, Electrochemical Process, Hydrogen Chloride
1. Introduction
Petroleum refineries in Thailand have grown continually. Industrial wastes from this industry
are consisting of solid wastes those contained heavy most metal and oil contamination which
considered as hazardous waste; and it is difficult to manage [1, 2]. The NaY zeolite is used for
remove hydrogenchloride from hydrogen stream in gas separation unit of petroleum refineray
212
process such as Dichloromethane (CH2Cl2), tetrachloroethylene (C2Cl3H) and chlorobenzene
(C6H5Cl) [4, 5]. The reaction of organochloride conversion to ethane, benzene, methane and
hydrogen chloride, produced according to the reactions:
C2Cl3H + 4H2 → C2H6 + 3HCl (1)
C6H5Cl + H2 → C6H6 + HCl (2)
CH2Cl2 + 2H2 → CH4 + 2HCl (3)
These hydrogen chloride gases must be removed by NaY zeolite before treatment to prevent
corrosion and fouling. Corrosion is caused by hydrogen chloride (HCl) and some of hydrogen
chloride gases combine with ammonia to form ammonium chloride salt (NH4Cl), which causes
fouling in the process. Thus, spent NaY zeolite is generated in refinery. Spent NaY zeolite is
generated after adsorb organochlorides and hydrogen chloride gases [3].
Generally, most of wastes generated in refinery plant can be reused and recycled; however,
some waste types still disposed of in landfill, especially, spent NaY zeolite. In Thailand, Spent
NaY Zeolite from petroleum refinery is disposed in landfill because landfill is simple and
cost-effective method. However, this method has shown some disadvantages such as loss of
land use and high risk of groundwater contamination. Normally, spent zeolite is regenerated by
ion-exchange [6], ultraviolet (UV) photolysis [7], oxidation with ozone [8], and microwave
assisted regeneration [9]
Recently, electrochemical process can be applied for desalination of various materials such as:
concrete, sandstones, granite and tile [10]. Electrodialysis for chloride removal from the
chemical recovery cycle of a Kraft pulp mill was also study. The result showed that chloride
removal efficiency was higher than 60% [11]. The electrochemical removal of chlorides from
reinforced concrete in Switzerland, it studied to obtain information on the efficiency of
electrochemical chloride removal as a restoration technique for chloride contaminated concrete
was presented. It is shown that about 50% of the initial chloride content is removed within 8
weeks [12]
The aim of this study is to investigate the using of electrochemical process to remove hydrogen
chloride from spent NaY zeolite. The operation parameters were varied to find the optimum
condition such as: initial of current density, reaction time and initial pH. The chemical structure
of regenerated NaY zeolite was determined compared to pure zeolite in order to reuse and
recycle as adsorbent.
213
2. Materials and Method
2.1 NaY Zeolite Waste
The NaY zeolite was used for remove hydrogenchloride from hydrogen stream in gas
separation unit of petroleum refinery process. Spent NaY zeolite was disposal for landfill
method since it is non-hazardous. It was consistent with the result from Leachate solution by
inductively couple plasma-optical emission spectrometer (ICP-OES), It showed that heavy
metal content in zeolite waste is not more than Soluble Threshold Limit Concentration (STLC)
regulatory limits (Table I). The chemical composition of commercial zeolite and spent
contaminated zeolite using X-ray fluorescence (XRF). The result was represented in Table II.
2.2 Electrochemical Set-Up
Electrochemical batch reactor was conducted in a 50 cm3 (5 cm x 1 cm x 10 cm). A schematic
diagram of the experimental setup is shown in Fig. 1. The Ti/Pt electrodes with flat and
rectangular shapes were placed in parallel with a separation distance of 1 cm. The electrode
gap was kept constant in all experiment. The surface area of Ti/Pt electrodes was 25 cm2 (5 cm
x 5 cm). The NaY zeolite was put in space between Ti/Pt electrodes. The reactor was
connected in parallel to a digital DC power supply. A direct current was imposed by a
stabilized DC power supply (Maximum of voltage 60 V and current 5 A). All experiment was
conducted at the room temperature of 28oC.
214
2.3 Methods
These researches of various parameters were the levels of current density, initial pH levels and
retention time periods. The parameters were obtained the optimal desorption conditions. The
levels of current density were varied 4 to 100 mA/cm2. The initial pH of water was varied from
5 to 12 and retention time periods were tested for 5 to 30 min. The remaining solution and
spent zeolite were analyzed to determine chloride concentration according to EPA Method
300.0(A) (Ion Chromatography: IC) [13].
2.4 Analytical Method
Chloride content was measured using a Dionex™ ion chromatograph (IC). The chemical
composition of commercial zeolite and spent contaminated zeolite were analyzed by X-ray
fluorescence (XRF), PW2400 model. The heavy metal content in leachate solution was
analyzed by inductively couple plasma-optical emission spectrometer (ICP-OES, Perkin Elmer
plasma-1000). The solids were taken for phase analysis by X-ray diffraction (XRD, D8
Discover Bruker). The morphology was analyzed by Scanning Electron Microscopy (SEM,
JEOL-5800LV). The surface area, pore size and pore volume were analyzed by BET (ASAP
2020).
3. Results
3.1 Effect of Current Density
It is known that the current density can affect the chloride desorption, which in turn affects the
desorption efficiency. Ti/Pt was selected as electrodes, for the experiments to examine the
effect of current density. It was varied from 4 to 100 mA/cm2. The initial pH of 8.06 and
reaction time of 30 minutes were maintained constant. Result shown that the chloride residual
in regenerated zeolite decrease when the electrical resistance decrease. The lowest of electrical
resistance was obtained at current density 20 mA/cm2
(as shown in Fig. 2). This result was
215
consistent with the Fig.3: it had shown the variation of the chloride removal efficiency as a
function of the current density. The highest efficiency of removal was achieved by applying the
current density 20 mA/cm2. The chloride desorption was occurred to 76.2%.
3.2 Optimum Initial pH of Water
To determine the effect of initial pH on the desorption efficiency of chloride were examined
using water with initial pH values of 5 to 12. Since pH of the water was alkaline (pH 8.06), it
was adjusted to desired pH using H2SO4 and NaOH. In each batch experiment, the applied
current density and reaction time were set constant at 20 mA/cm2 and 30 minutes. Result
shown that the efficiency of chloride desorption increase when the initial pH of water decrease.
The maximum chloride desorption efficiency was 79.2% at the pH-adjusted acidic water (pH
5). The chloride residual in regenerated zeolite as pH value at 5 was 5.19 mg chloride/g zeolite
as shown in Fig. 4. Moreover, the efficiency of chloride removal as pH 8.06, 10 and 12 were
removed by 73.8%, 71.9% and 69.4%, respectively. Chloride residual in regenerated zeolite as
pH value at 8.06, 10 and 12 were occurred to 6.55, 7.03 and 7.64 mg chloride/g zeolite,
respectively.
216
3.3 Effect of Retention Time
As seen in Fig. 5, the efficiency of chloride removal raised with an increase of the reaction time
period. The applied current density 20 mA/cm2 and initial pH of water value at 5 produced the
faster removal rate for chloride within 15 minutes. The efficiency of chloride desorption at the
retention time as 15 minutes was 78.6%. Furthermore, chloride residual in regenerated zeolite
was occurred to 5.58 mg chloride/mg zeolite.
3.4 Characterization of Regenerated NaY Zeolite
From the previous experimental, pH-adjusted acidic water (pH 5) was shown the best condition
for chloride desorption. However, as shown in Table II, the result of characteristics of
regenerated zeolite by XRF analysis indicated that the structure of zeolite was changed when
used sulfuric acid as pH-adjusted acidic water. The structure of zeolite was destroyed, silica
and alumina composition were reduced. Due to acidification in water and sulfide ion from
sulfuric acid solution, compositions in zeolite were leached. Moreover, after regeneration,
sulfide ion was adsorbed on zeolite reached to 2.07%.
217
This result was consistent with the Table II, the area-volume-pore size of zeolite by BET
analysis pointed that the pore volume of zeolite (pH 5) had pore volume minimum because
sulfides blocked pore size of zeolite.
As shown in Fig. 6, all of electrochemical regenerated zeolites had crystalline phase, except pH
value at 5 because outer surface has destroyed whereas inner surface was still same when used
sulfuric acid as chemicals used for pH-adjusted water [14].
218
The result of XRD represents in Figure 7. All condition gave the same result. The structure of
spent zeolite is crystalline phase composition. Nonetheless, the structure of spent zeolite
regenerated by electrochemical with the alkaline water as electrolyte is most likely pure zeolite
which is consistent the case of the zeolite treated with sodium hydroxide solution, the proceeds
in a more complete conformation and produces better crystallinity [15].
Because the chloride ions were transported out of the zeolite and into the electrolyte (5)
therefore, the chloride ions in the NaY zeolite are significantly reduced. Thus, the hydroxyl
ions are generated which produced a high pH in water (8). Reaction (6) and (8) showed that
bubbles from the gases will be seen near both electrodes.
Anode (oxidation):
2 Cl− → Cl2 (g) + 2 e
− (5)
2 H2O (l) → O2 (g) + 4H+
(aq) + 4 e− (6)
Cathode (reduction):
2 Na+
(l) + 2 e− → 2 Na (l) (7)
219
2 H2O (g) + 2 e−
→ H2 (g) + 2 OH−
(aq) (8)
Na content in zeolite affected the absorption potential and structure of NaY zeolite (Al-ONa-Si)
because of the amounts of Na+ (NaOH) species [16]. Therefore, the mechanism is reversible
when use NaOH as chemical used for pH adjustment due to the Na was exchanged in process.
This result was made clear by Table II, the amount of Na2O and pore volume are increased
after regeneration and higher than in spent NaY zeolite [17]. It should be suggested that
alkaline water is suitable electrolyte for chloride desorption from chloride contaminated zeolite
in refinery plant, its structure was similar to the commercial grade.
4. Conclusions
Electrochemical process is an efficient method for chloride removal of NaY zeolite
regeneration. Results from testing on the three sets of electrodes show that optimum condition
can be achieved using Ti/Pt electrodes with current density of 20 mA/cm2 and the initial pH
value of 10 at retention time of 20 minutes. The efficiency of chloride desorption was 71% and
chloride residual in regenerated zeolite was 7.03 mg chloride/g zeolite. The surface area of
regenerated zeolite was 431.3 m2/g. Consequently, it can be concluded that electrochemical
process is an alternative option for regeneration of NaY zeolite wastes which can be recycle as
absorbent for polluted in gas stream or wastewater. Moreover, the regeneration process has
shown several advantages such as reduction of waste disposal by landfilling.
Acknowledgements
This paper was financially supported by the 90th
anniversary of chulalongkorn university fund
(Ratchadaphiseksomphot Endowment Fund), Thailand and the Ratchadaphiseksomphot
Endowment Fund 2015 of Chulalongkorn University (WCU-58-020-CC )
References
[1] Usapein P, Chavalparit O. “Development of sustainable waste management toward zero
landfill waste for the petrochemical industry in Thailand using a comprehensive 3R
methodology: a case study,” Waste Manag Res 32(6), 2014, pp. 509–518.
[2] Usapein P.,Chavalparit O. Evaluating the performance of 3R options to reduce landfill wastes
using the 3R indicator (3RI): case study of polyethylene factories in Thailand. J Mater Cycles
Waste Manag, 2014, DOI 10.1007/s10163-014-0336-9.
[3] Erfan M. Chloride removal in refineries: A review of catalytic removal of chlorides from
refinery streams and a critique of current analytical techniques for estimating chloride content.
Unicat Catalyst Technologies. Available at: http://www. digitalrefining.com/data/articles
/file/433058853.pdf,2014
220
[4] Lo´pez E., Ordo´n˜ez S., and Dı´ez F. “Deactivation of a Pd/Al2O3 catalyst used in
hydrodechlorination reactions: Influence of the nature of organochlorinated compound and
hydrogen chloride, Applied Catalysis B: Environmental 62, 2006, pp 57–65.
[5] Pinard L., Magnoux P., AyraulP. T and Guisnet M. Oxidation of chlorinated hydrocarbons over
zeolite catalysts 2. Comparative study of dichloromethane transformation over NaX and NaY
zeolites, Journal of Catalysis, Vol 221, Issue 2, 2004, pp. 662–665.
[6] Rahmani A., Samadi M., and Ehsani H. Investigation of clinoptilolite natural zeolite
regeneration by air stripping followed by ion exchange for removal of ammonium from aqueous
solutions. Iran. J. Environ. Health. Sci. Eng. 6, 2009, pp. 167-172.
[7] Singh, S. Regenerating Spent Zeolites with UV and UV/H2O2 To Enhance Removal of
Endocrine Disrupting Compounds. Environmental & Water Resources Engineering Masters
Projects, 2010, Paper 38.
[8] Alejandro S., Valdés H., and Manéro M. Oxidative regeneration of toluene-saturated natural
zeolite by gaseous ozone: the influence of zeolite chemical surface characteristics. J
Hazard Mater, 2014, pp. 212-274
[9] Eduard S. and Călinescu I. Microwave assisted regeneration of 3å molecular sieves used for
ethanol dehydration. U.P.B. Sci. Bull. 76, 2014, pp. 96-106
[10] L.M. Ottosen, C.D. Ferreira, I.V. Christensen, Electrokinetic desalination of glazed ceramic
tiles, J. Appl. Electrochem. 40 ,2010, 1161–1171.
[11]Rapp H.and Pfromm P. Electrodialysis for chloride removal from the chemical recovery
cycle of a Kraft pulp mill, Journal of Membrane Science 146, 1998, pp. 249-261.
[12]Elsener B., Molina M., and Böhni H. The electrochemical removal of chlorides from
reinforced concrete, Advances in Corrosion and Protection 35, 1993, pp. 1563–1565.
[13]JohnD.Pfaff.1993.EPA300; Determination of inorganic anions by ion chromatography.
Environmental monitoring systems laboratory office of research and development U.S.
Environmental Protection Agency. Cincinnati., Ohio
[14] Salamatinia, Zuhairi and Kamaruddin. “Regeneration and reuse of spent NaOH-treated oil
palm frond for copper and zinc removal from wastewater,” Chemical Engineering Journal.
156, 2010, pp. 141–145.
[15] Eckehart R., Peter K., Akos K. and Frank H. 1993. Method of preparing zeolite Y. US 5223240
A.
[16] Piwgaew W., Grisdanurak N., and Kumsaen T. “Modified zeolite NaY for HCl gas trapping
from natural gas feed,” The 4 KKUU international engineering conference, 2012.
[17] Pinard L., Mijoin P., Magnoux P.,and Guisnet M. “Oxidation of chlorinated hydrocarbons over
Pt zeolite catalysts 1-mechanism of dichloromethane transformation over PtNaY catalysts,”
Journal of Catalysis 215, 2003, pp. 234–244.
221
SICBENS-1326
Environmental Hazards of Radioactive Radiation from Granite Building
Stones through a Look into Burien’s Granite Mass in Isfahan
Siyamak Bagheriyan
faculty member of Islamic Azad University, Tiran Branch
Siyamak. bagheriyan@ gmail.com
Abstract
Used for main works of structures are building stones including rubble, malone and face stones
with metamorphic, sedimentary and igneous base. The mine groups of non-carbonate and hard
stones such as granite, gabbro, gneiss and hornfels often considered as igneous or metamorphic
rocks are available as granite in the market; however, lithologically they fall under other stone
types. In recent years, regardless of its hazards and environmental shortcomings supply and
demand have increased for these stone types. Because of their chemical composition and ore-
oriented texture, they send out radioactive radiations. Accordingly, WHO (World Health
Organization) is currently compiling some instructions regarding the proper use of these stones
in structural industry. On the other hand, it is possible that people develop diseases such as
silicosis due to high amount silica in these stones. This research focused on the processing and
using these stone through a look into Buien- Miyandasht`s granite mass in Isfahan province.
Key words: building stones, granite, environmental, disease, Isfahan, Buien`s mass
1. Introduction
Man`s use of building, decorative and face stones dates back to the ancient times, when they
utilized them in accordance with their needs and capabilities to process the stones. In doing so
Persians were fabulously talented. They would manage to use the stones of sedimentary,
igneous and metamorphic origin both situationally and locally. Bringing granite stones to work
in Mashhad and Hamedan enjoys an old background. Recently, particularly after Islamic
Republic Revolution, the country has explored, extracted and processed huge masses of granite
at hand. Henceforth, especially during 70s of solar calendar plenty of factories have applied
changes in the manufacturing systems and processed these decorative stones. Although the
granite stones are of greater strength and beauty when compared with other decoratives such as
carbonate rocks (travertine, porcelain), the environmental issues and hazards have
unfortunately been overlooked while processing. These stones contain the greatest amounts of
radioactive elements and heavy metals. They are much more dangerous than any other stones
particularly when used inside the buildings. On the other hand, the immense masses of
granitoid in the country offers a subtle perspective on discovering magmatic decorative stones.
As a result, environmental issues have to be taken into fine considerations.
222
2. Discussion
2.1 Building, decorative and face stone types
Building and decorative stones cover a variety of igneous, metamorphic and sedimentary
stones, each of which are affluently subdivided into categories. Recently, they have been
marketed under special trademarks and the scientific titles have been put aside.
Igneous rocks: the stones of this group come from huge rocks which are themselves either
external or internal. Another ranking for these stones is their degree of acidity on the basis of
chemical composition and silica (Sio2) in particular, which are acid, moderate, alkali and
ultra alkali igneous rocks. They differ in beauty, color and texture. Table l illustrates various
types of these stones.
Table 1: Different igneous stone types used as material and face stone in construction
Major class
Minor class
Name
Site in Iran
Internal igneous
rocks
acid
Granodiorite
Hamedan
Granite, Astaneh
Granodiorite
moderate
Synite, Tonalite
West of Khomein
alkali
gabbro
Kerman,
Uroumia
ultra alkali
Dunite, pyroxenite,
Harzborgyte
Esfandiyeh of
Kerman, Birjand
External igneous
rocks
acid
Rhyolite
West Azarbaijan,
Damavand
moderate
Andesite
Yazd
alkali
Basalt
Torbat-e-Jam
Destructive
igneous
Echo, lava
Agglomerate,Echo,Lapilli
Ghazvin, North
of Karaj
Metamorphic rocks: This group includes different stones of adjacent, regional and burial
metamorphic stones with igneous, sedimentary and metamorphic origin. Used as decorative,
building and face stones, the stones of this group are classified as carbonate and non-
carbonate.
223
Table2: Different metamorphic stones used as material and face stone in construction
Sedimentary rocks: These make the biggest class of decorative, building and face stones.
Compared with other types of igneous rocks, the carbonate ones are more popular in use since
they exist in great amounts, are soft and easy to process.
Sedimentary rocks are two types: destructive and chemical. Table 3 illustrates a commercial-
based classification.
Major class
Name
Description
Site in Iran
Carbonate
Pearl
They are known as
porcelain and crystal in
the market.
Azna and Aligoudarz`s
porcelain
Non- carbonate
Hornfles
They are known as
black granite in the
market.
Touiserkan of
Hamedan
Amphobolite
Dark green stones
known as green granite.
Birjand
Gneiss
The stones are so
similar to granite.
Yazd
224
Table3: A commercial- based classification of sedimentary rocks
2.2. Magmatic building stones of Iran
Huge supplies of magmatic (igneous) stones in Iran has now made it possible to discover,
extract and process different types of igneous stones (often commercially called granite). The
growth of industry in the country, development of stone industries and possibility of cutting
hard stones (stones much harder than steel) in particular allows an increase in processing hard
igneous stones. Naturally, along with the demand for modern industries and simultaneous
development in building industries and other related industries, the demand for exploring these
supplies has gone up. Consequently, the number of explosion permits shows a considerable
growth from 3 in 1993 to 40 in 2012 in Lorestan province. From among explored igneous rocks
in the country, which are now being processed, we can refer to granite (granite from
Hamadan`s Ganjnameh), granodiorite (Masouleh mine), serpentine (Mashhad mine), echo
(Northen Qazvin), basalt (Torbat-e-Jam), synite (Khorramdarreh), tonalite (Eastern khomein),
gabbro (Torbat-e-Heydarriyeh, Uroumia), andesite (Yazd), pyroxenite and amphiolite
(Birjand).
Major class
Name
Commercial
name
Description
Site in Iran
Carbonate
Limestone
Pearls various
colors
They come in
beige, white, light,
green, back and
red.
Gohareh of
Khorramabad,
Deh Bid of
Shiraz
Travertine
Travertine
They are of
different colors.
Mahallat
Annix Pearl
pearl
It has borders.
Yazd and
Kerman
Destructive
Conglomera
Color granite
Rarely used as face
stone.
Shahr-e-Kord
Sandstone
Red, black
granite
Mainly used as
Malone stone.
Alborz
Mounts, esp.
North of
Tehran
225
2.3. Acid igneous stones (Granite, Granodiorite)
Igneous masses situated in Sanandaj- Sirjan zone in the west of Iran are more or less seen in
intruding forms of oval and chain-like masses in a metamorphic line alongside northwest and
southeast Uroumia- Golpaygan provinces. According to Aghanabati (1993), the life span of
these intruding masses dates back to middle Jurassic, for on one hand these masses have cut
into Jurassic deposits and on the other hand lower cretaceous deposits sit over them with a
gradient of igneous- base. Sabzeh`ei(1997), clamis metamorphic zone complex around
Hamedan, Malayer, Broujerd, Aligoudarz, Bouien- Miyandasht granite masses like the
Southern Sanandaj-Sirjan zone are indicative of complexes generated during orogenisis. As a
result of laramid orogenisis some considerable intruding masses like Mahabad, Qorveh, Alvand,
Malayer, Arak, Broujerd- Aligoudarz, Golpaygan, Bouien- Miyandasht granite sites have been
started all with a main oval orientation parallel to one another and ultimately to Zagros
drift( Radfar, 1995). There are granitoid , pegmatitic veins containing granite and granodiorite
compositions.
2.4. Intruding masses of Bouien- Miyandasht
Granitoid masses of Sanandaj- Sirjan zone have intruded into southwest of Aligoudarz in
Bouien- Miyandasht region inside the Jurassic sandstones and shale. Most of these igneous
stones have been composed of middle- sized granule granite- biotite and granodiorite that get
porphyritic in some places and have been cut into by aplite veins. Tileh et al (1968) have
estimated the life span of these igneous masses, like the intruding ones of Hamedan and
Brujerd, back to late Triassic and early Cretaceous. However, regarding definite age of Alvand
mass in Hamedan, this mass possibly belongs to late Cretaceous i.e. Paleocene (Bloun & Broud,
1975). Based on studies, and dating though radiometry, Valizadeh (1978) claims) the granite
mass of Hamedan to be of about 70 million years old. Soheili et al believe that the intrusion of
granitoid masses into destructive metamorphic deposits and upper Triassic- lower Jurassic has
led into formation of adjacent metamorphic stones around itself with mineralogy including
srycite + chlorite + biotite + epidote + quartz + scapolite + calcite + feldspar + (albite- epidote
– hornfles) facies.
The main texture of granitoid stones of Bouien- Miyandasht are mainly granular and tend to be
tiny on marginal regions and coarse on center. In order of affluence, the major constructing
ores of these stones are quartz, alkaline feldspar, and minor ores of biotite, amphibole and
tourmaline, respectively. Rare ores include apatite, zircon, and decomposed ores like mica, clay,
chlorite, epidote, and eventually calcite. From mineralogical point of view these are considered
to be either granite or granodiorite. These masses have various compositions of light granite
and are poor in ores of iron, magnesium and dark granodiorite, rich in ores of hornblende and
biotite. Granules range from 2 to 6 millimeters in size and at times feldspar seems to be coarser.
Erosion, mechanical and chemical decomposition are evident especially along the fractures.
Petrological and petrographic studies show igneous rocks of granite mass of Bouien tend to sit
at the granitoid of S class zone. To firm this we can mineralogically refer to the presence of
226
ilmenite and andalusite crystals, basic biotite and masscovite, the absence of hornblende. These
granodiorite messes of S class contain high degrees of radioactive.
2.5. Granite rocks of Bouien region and radioactive
Of rocks, acid igneous rocks are high in radioactive. Due to their ionic radius and high ionic
charge, uranium, thorium and strontium are condensed inside these stones. The affluence rate
of uranium is illustrated in table 4 and that of ores in 5. As such xenotime, zircon and monazite
contain maximum amount of uranium and olivine minimum amount, and they are found in acid
igneous rocks and moderate alkaline ones.
Table 4: Uranium rate in major rocks and ores (gram each tone) (Clark and others, 1966, Nash
and others 1987)
Name Average uranium
rate gram per tone
Name Average uranium
rate gram per tone
Donate
0/02
Granodiorite
2/6
Echologite
0/2
Granite
4
Pyroxenite
0/7
Organic shale
10
Gabbro
0/9
Red shale
3
Diorite
2
Sandstone
1
Carbonates
2
Evaporative
0/1
227
Table 5: Uranium rate in silicate ores (Clark and others, 1966, Rogers, and Adams, 1967).
Ore
Uranium rate g/t
Ore
Uranium rate
g/t
Quartz
0/1-10
Monazite
500-3000
Feldspar
0/1-10
Titanite
10-700
Masscovite
2-8
Epidote
20-200
Biotite
1-60
Allanite
30-1000
Hornblende
0/2-6
Apatite
5-150
Pyroxene
0/01-50
Garnet
6-30
Olivine
0/05
Ilmenite
1-50
Xenotime
300-35000
magnetite
1-30
Zircon
100-6000
Biotite: About 5 to 20 percent of granitoid mass stones of Bouein- miyandasht are made up of
biotite. They turn to be brown when exposed to light and olive green at times. Also, it is
capable of making spectrum of different colors. The ore is usually shape less; however, it may
seem to have a form, sometimes. There can be seen faces along with its longitudinal cut.
Biotites have turned into curved forms due to tectonic pressures. Among heavy loads of zircon
with paleocroic trace, rutile, titanite and dark ores can be observed as well. These crystal are
strongly chlorite and retain their basic from. Chlorite is often gained from penitite
decomposition. While decomposing, biotites get filamentous.
Masscovite intrusion is occasionally observed and Piranjo (1992) relates this to a potash
metasomatism phase particularly if it accompanies plagioclases` alteration. The radioactive
228
(uranium) rate of these ores was microprobically analyzed. Table6 illustrates it. (Ahmadi khalaj,
2011).
Table 6: microprobic analysis of biotite, a granodiorite sample from granodiorite masses of
Sanandaj- Sirjan zone (adapted from Ahmadi khalaj, 2011)
Grd Grd Grd Grd Grd Grd نمونو سنگ
Grt Grt Grt Bit Bit Bit کانی
Pint
Namber
#1 #2 #8 #1 #2 #8
Sio2 86/36 65/36 16/37 Sio2 25/34 55/34 61/34
Ai2o3 46/21 43/21 42/21 Ai2o3 87/18 7/18 63/18
Mgo 92/1 9/2 75/2 Mgo 88/7 45/7 48/7
Feo 32/31 84/33 94/33 Feo 65/23 2/23 23/23
Tio2 83/8 1/8 89/3 Tio2 34/2 2 92/1
Mno 39/8 72/4 84/4 Mno 18/8 23/8 21/8
Cao 24/1 14/1 17/1 Cao 8 85/8 82/8
Na2o 83/8 82/8 81/8 Na2o 19/8 22/8 15/8
K2o 8 81/8 8 K2o 57/8 49/8 49/8
Cr2o3 81/8 86/8 87/8 Cr2o3 83/8 86/8 83/8
Nio 8 8 - Nio 83/8 82/8 8
Total 27/181 88/188 44/181 Total 2/95 97/94 78/94
Si 96/2 95/2 97/2 Si 66/2 68/2 69/2
Ai 83/2 83/2 82/2 AIIV 34/1 32/1 31/1
Mg 23/8 35/8 33/8 ∑ 4 4 4
Fe 11/2 27/2 27/2 AIVI 39/8 4/8 4/8
Ti 8 81/8 81/8 Mg 82/8 86/8 87/8
Mn 57/8 32/8 33/8 Fe 54/1 51/1 51/1
Ca 11/8 1/8 1/8 Ti 14/8 12/8 11/8
Na 8 8 8 Mn 14/8 12/8 81/8
K 8 8 8 Ca 8 8 8
Cr 8 8 8 Cr 8 8 8
Ni ∑ 9/2 9/2 91/2
Total 82/8 83/8 82/8 Na 83/8 83/8 82/8
K 85/8 84/8 84/8
TOT 78/7 78/7 77/7
229
Zircon: With specific crystal and optical features, like remarkability and antireflexion, this ore
exists in great loads in granitoid rocks of the zone. It is rarely found on its own but mainly in
biotites. A few of them can also be found in quartz and feldspar. This ore produces a strong
paleocroic circle around itself since it contains radioactive (including uranium, thorium). This
process is mainly evident at thin layers and it called metamictization; crystals surrounding
zircon ore. This is seen as burns in biotites of Bouien- Miyandasht. Furthermore, black and
brown spots are sometimes seen inside biotites.
Apatite: Apatite ores do not exist much in granitoid rocks of this zone. Besides, they are often
as pinnacle crystals and a little shapely in from inside the ores. Quartz and feldspar crystal are
common places where apatite often sits. Sometimes, dark spots are observed over them.
2.6. Economical investigation of granite building stones of Bouien- Miyandsht
The presence of hung granitoid masses offers a suitable perspective to explore magmatic
decorative stones all over the zone. Since physical characteristics of ores or the ore types range
in properties, using these stones is greatly limited. Their polishability totally depends on the
texture of ore type and this in turn is followed by factors such as magmatic composition,
density, cooling rate and environmental composure. In addition, the quality is unquestionably
related to dating factor. The greater age, the greater alteration and erosion and the more limited
use of ores. These stones are affected by tectonic actions and get factures. Then, masscovite,
sericite and iron oxide fill the fractures. (Ghasemi 1993). Differences in ore hardness, severe
alteration with inappropriate granulation, gaps and close fractures, low-grade color and form
and existence of xenoliths are all considered to be of remarkable factors disqualifying these
stones. They differ in quality on a scale of darkness, brightness and granule size. The west zone
of mass is of greater quality. Consequently, exploratory operations are mostly being offered for
these zones.
2.7. Building stones and their radioactive radiation effects on environment
Acid igneous rocks contain the highest amount of radioactives because of their chemical
composition and mineralogical texture. These materials that are mainly condensed into biotite
and zircon ores of these stones ore capable of high degrees of radioactive radiation. Biotite rate
of Bouien zone stones is about 5-20 percent. Sometimes, biotite in the form of xenolith makes
the stone get dark spots. Also, these radio actives may concentrate in zircon, apatite, and
ilmenite of these stones and they get high at times and visible at microscopic sections. High
rates of radioactives in granite stones has persuaded World Health Organization (WHO) to
decide on codification of instructions which encourage proper use of these stones in
construction industry (Aliyev and Zolotovitskaya, 2005). The research carried on construction
materials by Aliyev and zolotovitskaya in Azerbaijan indicates high rates of radioactives in
materials freely used in the kitchens and bedrooms. Recent studies are even indicative of
significant amounts of radium. Therefore, acidic igneous rocks (granites and granodiorites),
which contain zircon and biotite, are not proper to be used inside buildings for they send out
230
and radiate radon gas. A thorough investigation needs to be done from sanitary point of view in
as much as all construction materials are composed of ores and stones; an issue which
unfortunately has not yet been regarded. Nonetheless, many people and even business owners
are not still fully aware of the dangers and since there is little epidemiological evidence, no
clear regulations on construction industry have been predicted. Some right notifications on
building contamination and in particular radioactive radiation must be supplied. However, the
construction materials with granite structure are seriously hazardous if used in public places
particularly apartments. And the society knows almost nothing of this.
2.8. Uranium in granite rocks
Human body does not need uranium and chemically it has no poisonous effects on living
creatures (especially humans). The major hazards of this element relate to its radiation and the
most important effect of uranium on human body is developing different types of cancer.
Floating particles of radioactive are taken into human body through breathing. Uranium decay
produces the hazardous and cancerous radon gas, which can evaporate from uranium inside
granite rocks and if breathed, it can be fatal. After being released, radon starts to produce other
radioisotopes which are sucked by flouting particles in the air, eventually taken into lungs
through breathing and this causes cancer or other lung- related diseases (Shahab pour, 2008).
From among other products developed under uranium decay we can refer to thorium
radioactive and plutonium, each of which can have adverse effects on human body. The
primary side effects of exposure to uranium radiation are nausea, exhaustion and depression,
serious skin burn, ague and chronic harms of these radiations can be damages to different body
organs, genetic mutations ( such as children`s retardation) and cancer. There is some thorium in
biotite and zircon which is hazardous on both living things and ernvironment. Regardless of its
radiative dangers, it has no chemical toxins. Thorium`s effects and its radioactivity is the same
as other radioactive elements like uranium. Uranium Clark within granitoid rocks is higher and
under exogenous conditions tetratomic uranium turns into hex atomic one due to its instability.
When exposed to air, U3 O8 changes into uranyl ions (UO2 +2
) which are the most mobile
uranium compositions. Radioactive radiates in the forms of alpha, beta and gamma, each of
which affects human body differently. Some elements of radioactive such as radium and
thorium act like calcium and therefore, when absorbed by human body, they immediately
aggregate in the bones and start doing a damage to body (Dabiri, 2001)
231
Table 7: The results from chemical analysis of different granitoid samples in Bouien
B-M-1 B-M-2 B-M-3 B-M-4 B-M-5 B-M-6 B-M-7
Sio2 72.81 71.42 70.24 70.54 68.1 69.98 65.1
Ai2o3 13.20 11.95 14.21 13.56 15.19 16.02 17.98
Fe2O3 1.60 1.46 1.61 2.5 1.92 1.01 3.2
FeO 0.59 1.25 1.48 1.95 1.16 1.1 0.72
MgO 0.59 0.32 0.87 0.65 0.27 0.25 0.69
MnO 0.04 0.03 0.02 0.11 0.06 0.03 0.02
CaO 1.22 1.92 1.95 0.21 0.98 1.56 1.21
Na2O 2.51 3.01 3.50 2.54 3.87 3.14 3.68
K2O 4.20 3.99 4.20 3.57 4.01 4.02 3.89
P2O5 0.15 0.16 0.15 0.19 0.18 0.12 0.18
L.O.I 0.80 1.45 0.69 1.22 0.01 0.74 0.78
TOTAL 99.31 97.94 98.89 97.98 97.31 98.99 97.91
Table 8: Modal analysis of granitoid ores in Bouein
B-M-1 B-M-2 B-M-3 B-M-4 B-M-5 B-M-6 B-M-7 شماره کانیها
Q 51.1 39.8 42.9 40.1 45.3 37.1 33.2
Plg 21.2 27.1 27.9 30.4 20.8 40.2 31.9
Ort 17.5 26.1 16.7 27.9 10.9 14.2 21.9
Bio 5.9 3.1 7.9 1.1 18.9 51.1 9.2
Musc 0.5 - 0.1 - 1.3 - -
Turm - 2.1 3.0 0.2 2.1 - -
Opaq - 0.1 0.3 0.2 - 0.3 0.5
3. Conclusion
Huge masses of granite in the country like Sanandaj- Sirjan zone (e.g. Bouien- Miyandsht) has
offered opportunities to explore and process these supplies to be used as building and face
stones. These stones are granite and granodite that contain ores such as quartz, feldspar, biotite,
zircon and apatite. Radioactive rate and uranium rate in particular is high in acid igneous rocks
which is mainly located inside biotite and zircon ores. Accordingly, utilizing these granite
stones as construction materials especially inside apartments and kitchens increases the
radioactive radiation danger. From a microprobic analysis perspective on biotites and high
uranium rate, there seemingly is a great need to study radioactive contamination efficiently.
232
References
[1] Aliyev, Ch. S. (2005) Variation of subsoil radon concentration during earthquais
International workshop for : Recent GAEODYNAMICS, Georisk and sustainable
Developme in the Black save sea fo Caspian sea region, 3-6 july, baku.
[2] Aliyev, Ch. S., Zolotovitskaya, T.a. (2005) Rdiation safety of building materil
Azerbaijan , p459-461.
[3] Aliyev, F. S., Mammadova, M.A., (2005) Geological problems of fresh underground waz
resources in Azervaijan Rep. in the conditions of urbanization and global climate chang
the main sources of water supply of population,JournalScience,Baku State Univers
(BSUJS), No1, pp128-136, (in Azeri)
[4] Anonymous (2005) Silicosis, Learn the fast, National institute for occupational safety are
health (NIOSH), Cincinati, Oh., 20 pp.
[5] Appleton, Donald, J. (2005) Radon in air and water, Essentials of Medical Geolog Elsevier
Academic press, Chap. 10 , p227-262.
[6] Atabey, Eshref, (2005) Medical Geology (TibbiJeoloji), TMMOB jeolojiMuhendishOdasi
Publishers, Anakara, 194pp., (in Turkish)
[7] Bennet, M.R., Doyle, P. (1997) enrironmental geology (Geology and the human
environment), John Wiley & Sons, 512pp.
[8] Derbyshire, Edward (2005) Nturalaerosolic mineral dusts and human health, Essentials of
Medical Geology, Elsevier Acadimic Press, Chap. 18, p459-480
[9] Ghazban, freydoun, (2003), Environmental Geology, 1st print, Tehran university pablitions,
p.416
[10] HajAlilou, Behzad, Vosough, Bhram (2010) Medical geology, Payam-e- Noor University
[11] Konikow, Leonaed, F. and Glyny, Piere, D. (2005) Modeling gtoundwater flow and
quality, Essentials of Medical Geology , Elsevier Acadimic Press, Chap.31, p737-766.
[12] Mahdavi, pedram (2007), Medical Geology, Mine Expoloration and Geology organization
site of Iran
[13] Mehr partou, Leili (2007), cadmimum and Medical Geology, Mine Exploration and
Geology Organization site of Iran
[14] Mine exploration and geology organization of Iran (2007), Radiation from Buildings,
Mine Exploration site
[15] Moller, Lennart (Head editor), (2000) Environmental Medicine, Falth&Hassler AB,
Stodkholm , Swedin, 325pp.
[16] Weinstein, Philip, Cook, Angus (2005) Volcanic emissions and health, Essentials of
MEDICAL Geology, Elsevier Academic Press, Chap. 9, p203-226.
[17] www.medicalgeology.org
233
Table 7: The results from chemical analysis of different granitoid samples in Bouien
234
Computer & Information Science II
Session Chair: Seonho Choi
SICASE-723
Decision Supporting System Applicaton in Power Generation Technology Selection
Irwanto Walgito Pawiroijan︱PLN Indonesia
Zainal Arifin︱PLN Indonesia
SICASE-703
A Pilot Study of Real-time Sport Tactic Technique Record, An Example for Badminton
APP Application
Chia-Ping Huang︱Hsing-Wu University
Pai-Feng Lee︱Hsing-Wu University
Cheng-Chung Lu︱Hsing-Wu University
SICASE-759
Optimized Design of Geo-Temporal Query Based User Authentication Scheme for
Smartphones
Seonho Choi︱Bowie State University
Kun Sun︱College of William & Mary
Edward Jung︱Kennesaw State University
2015/6/28 Sunday 13:00-14:30 Room 2
235
SICASE-723
Decision Supporting System Applicaton in Power Generation Technology
Selection
Irwanto Walgito Pawiroijan
PLN Indonesia, Indonesia
Email address: [email protected]
Zainal Arifin
PLN Indonesia, Indonesia
Abstract
Energy planning is mostly relating to the selection of energy source and power generation
technology to produce the electricity. The focus of this paper is how to select the power
generation technology using logical, scientific, and objective method. A decision supporting
system (DSS) with intelligent algorithm is used to calculate the parameters of each technology.
The parameters are electricity price, quantity of supply, investment cost, construction period,
environmental impact, etc. By calculating these parameters, the DSS will obtain the score of
each technology and display it on a map. This map contains 4 zones which shows different
level of priority; the green zone (first priority), yellow (second priority), orange (third priority)
and red (fourth priority). With the DSS, the selection process will be rigorous and highly valid.
The company can choose the best option of power generation technology regarding to their
purposes based on scientific, logic and objective consideration.
Keyword: Decision Supporting System, DSS, Power Generation, Technology
1. Introduction
Energy planning is one of main concerns of electricity companies to maintain their electricity
supply. Energy planning is mostly relating to the selection of energy source and power
generation technology to produce the electricity. The focus of this paper is how to select the
power generation technology using logical, scientific, and objective method. A decision
supporting system (DSS) is used to calculate the parameters of power generation technologies.
The parameters are electricity price, quantity of supply, investment cost, construction period,
environmental issue, etc. By calculating these parameters, the DSS obtains the score of each
technology. The scores are displayed in a map which helps the management to select the right
technology easily.
236
2. Main Contents
2.1 Computer Server
The DSS is an interactive web based program which is installed in a computer server. It uses
windows server 2003 for operating system, php for programming code, apache for web server
and mysql for database. The web based program is selected because of its multi user ability and
wide range access through local area network.
2.2 Intelligent Algorithm
An intelligent algorithm is used to calculate the power generation parameters. First, it records
the weight of each parameter according to its impact to electricity system. Second, it records
the value of each parameter. Third, it calculates the parameter score by multiplying the
parameter weight with the value. Finally, it plots the position of each technology in a map
display based on its score.
2.3 Power Generation Parameters
The parameters of power generation technology are divided into 2 axes, which are Impact and
Uncertainty. Each parameter is determined its weight and value. The total weight of the
parameters is 100 and the value of each parameter is between 0 and 10. The details are shown
in Table 1 and Table 2.
237
2.4 Display Map
Display map is very important to plot the scores of each technology in a graph. It helps the
management to see and choose the best technology to be implemented in the company. The
map display consists of two maps, which are priority map and period map (figure 1 and figure
2).
3. Result
The DSS produces 2 display maps which are showing the position of each power generation
technology. The first map is a priority map which shows the priority level of power generation
technology. It is shown in figure 1. The map consists of 4 zones which shows different level of
priority; the green zone (first priority with high impact and low uncertainty), yellow zone
(second priority with high impact and high uncertainty), orange zone (third priority with low
impact and low uncertainty) and red zone (fourth priority with low impact and high
uncertainty). The management should choose the technology in the first priority because of its
highest benefit to the company.
From figure 1, it is shown that thermal and hydro power plants have big impact and low
uncertainty. The renewable energy such as wind, ocean, fuel cell, and solar cell has low impact
and high uncertainty. Based on this map, the management should choose the thermal and hydro
to produce the electricity.
238
The second map display is a period map which shows the implementation period of each power
generation technology. It is shown in figure 2. This map consists of 3 period phases with 10
years intervals. The period phases are Phase I (2015-2025), Phase II (2025-2035), and Phase III
(2035-2045). Each phase also consists of different priority of level zones.
From the period map, it is shown that thermal and hydro power plants have big impact. The
renewable energy such as wind, ocean, fuel cell, and solar cell has low impact. Based on this
map, the management should choose the thermal and hydro to produce the electricity and
implement them in phase I.
239
Priority map and period map have detail information. The management can access it by
clicking the icons on the map. The detail information of each technology will be displayed as
Table 3, 4, 5, 6, and 7.
240
241
The total score of impact parameters and uncertainty parameters are used to determine the
technology position in the priority map and period map.
The detail information also consists of a short description about definition of the technology,
how it works, classification, maturity of technology development, projects in the world,
technical analysis, economic analysis, environment impact, list of manufacturer, and reference.
242
4. Conclusion
The DSS is the best solutions to choose the power generation technology. It helps the
management to make a decision based on scientific, logic and objective consideration with
valid data and information.
References
[1] http://www.dfpni.gov.uk/eag-the-weighting-and-scoring-method
[2] http://www.eia.gov/forecasts/capitalcost/
[3] https://php.net/manual/en/
[4] http://www.appservnetwork.com
243
SICASE-703
A Pilot Study of Real-time Sport Tactic Technique Record, An Example for
Badminton APP Application
Chia-Ping Huang
Department of Information Management, Hsing-Wu University
Pai-Feng Lee
Department of Wealth Management, Hsing-Wu University
Cheng-Chung Lu
Department of Common Knowledge, Hsing-Wu University
Abstract
Due to rapid improvement and changing of the information technology, it results in the
popularity of the use of smart mobile devices with lots of relevant researches to be conducted.
Therefore much attention has been shown in sports industry to develop APP for realizing the
electronic and mobilized information platform applied for sports record analysis. The early
traditional method of record analysis was done manually which is not efficient and costs a lot
of labor and time. The currently conventional statistical information system is not able to
provide a flexible information platform with the correct and real-time data leading to a more
dominant match record analysis.
In order to compete with their rivals, the players of the Badminton have to regulate their
physical fitness and skills involved during a match. The information obtained through the
videotape record analysis is not accurate and usually cost a lot of labor and time. The main
focus of this research is to develop a real-time, flexible and lightweight badminton tactics
analyzer providing a tool platform for the coaches and players to collect lots of badminton
technical data with rapid match record analysis. A smart mobile device with an APP is aimed
to be built for analyzing the professional badminton scoring patterns, technology utilization
rate, serve and shot position through the nine rectangle-grid record mode.
The badminton tactics analyzer is aimed to provide the real-time data record involved during a
match for the coach, players and scouts. The complete recording and subsequent analysis of the
relevant badminton technical skill utilization rate such as serve, net shot, smash, passing shot,
push shot, attacking clear, kill, drive, drop shot...etc and the scoring patterns such as 'too late to
244
hit the shuttlecock', out, 'touch the net during a shot', misjudgment are able to be achieved
using the mobile analyzer with a custom-made APP. The results show that different identity
(such as coaches, scouts, players, recorders) can immediately understand the real-time match
record and possible batting tactics analysis using this professional badminton tournament data
recording analyzer with a supporting database display mode. It is verified that the lightweight
mobile analyzer with the required APP has the ability of reducing the complexity of the
originally tedious manual records.
Keywords：badminton、APP、tactics analyzer
1. Introduction and Research Background
Since badminton is a kind of openness sport, in order to compete with their rivals, the players
of the Badminton have to regulate their physical fitness and skills involved during a match. The
player and coach usually analyze the batting skills and tactics through the information obtained
from the videotape record of the match. This analysis can provide the batting strategy, such as
'time-span per shot', 'time-span between two shots' and 'time-consuming per game', for the
players to apply in a match against their rivals. Some important messages of the rivals and
games can also be obtained from the analysis. Theoretically, the intensity involved during the
confrontation between the two rivals is thought to be related to the 'time-span per shot' and
'time-span between two shots'. The energy consumption of the player is assumed to be relevant
to the 'time-consuming per game'. The aforementioned analytic data can be applied directly at
the training courses of all levels of badminton terms in order to increase the training
effectiveness and enhance the player's physical fitness. However, the data obtained through the
videotape record analysis is usually not accurate and cost a lot of labor and time, therefore, this
focus of this study is trying to develop an mobile information platform with the APP to enable
the function of real-time statistical sports record analysis
The innovative objective of this research is to invent a real-time, flexible, lightweight
badminton tactics analyzer which contains a smart mobilized device and an APP used for
analyzing the professional badminton scoring patterns, technology utilization rate, serve and
shot position through the nine rectangle-grid record mode. The analyzer has the ability to
compile a lot of badminton technical information with integrated fast record of match condition
provided for the coach and player to conduct further review and analysis.
2. Literature Survey
Basically, there are different method and technology model to carry out the match record,
tactical classification and advantage-disadvantage analysis for each ball game tournament in
order to meet the professional needs of different ball game. For example, a badminton teaching
and training system developed using Flash by Dr. Wu-Zhou Chen[1].
245
A visual analysis system applied for the table-tennis ball game was developed by the image
identification technique by Dr. Mao-Chun Yang[2]. A technical and tactical analysis model
applied for the first three shots involved during a table-tennis ball game was developed with the
Visual 2008 by Dr. Ming-Yue Wang[3]. A real-time technical analysis and match record for
the table-tennis ball game was developed using Visual C#.NET by Dr. Yu-Fen Chen[4]. These
research results have shown that the advanced information technology, used for the assistance
in the self-training, match records, compilation of the technical analysis data, is gradually
implemented into different types of sports ball game based on their user requirements. It is
indicated that the traditional manual ball game record mode has been gradually replaced by the
ones developed with E-tools (EXCEL/WORD), information tools (FLASH/ACCESS), web
techniques (Visual 2008/SQL) and M-tool (APP)..etc. Recent important sports ball game
record analysis system developed using adequate information tools are summarized in Table 1.
3. Issues analysis and Research Methodology
246
According to the aforementioned research results, the speed of sports ball games are so fast that
the time-span between the batting stroke and the return stroke is assumed to be around
milliseconds - several seconds. Therefore, the sports ball game is referred to be a fast-paced,
rapid-response exercise and discipline. Because the flight speed of a badminton ball after being
hit by the racket is normally higher up to around 230-260 km / hr, therefore, it is necessary to
grasp the information obtained through the videotape record analysis in order to master the
rival's game tactics mode during a match. It also requires a lot of ball game records and
analysis in order for statistically analyzing the technical condition of the plays. For
immediately seeking the tactics to beat the rivals, the coach needs the real-time ball game
analysis information. However, for the coaches, players and scouts, it requires quite a lot of
time and professional competence to grasp the statistical analytic information by watching the
ball game videotape record films. Because the understanding of the technical skills and tactics
of the rivals is the tips to be ever-victorious in a match, therefore, a quick, real-time, efficient
badminton tactics analyzer with the APP having the function of ball game record analysis, to
be developed to fulfill the requirement of the coaches, players and scouts, is needed and
currently one of the new trends of the information technology applications for the sports
industry.
Based on the foregoing analysis, the SDLC information system development mode is chosen to
conduct this research for the development of the ubiquitous badminton tactics analyzer. As
soon as the system requirements analysis is determined, the design of the relevant systematic
interface of U-analyzer is then followed. The database, developed for the analyzer integrated
with a APP, is planned to be validated by way of a field testing process. The research
procedures involved during this study is subdivided as shown in Figure 1.
The goal of this system development is let the traditional manual record analysis mode to be
replaced by a ubiquitous analyzer with real-time images operation mode integrated in a
PC-Tablet. The main development tools with related hardware and software are chosen to be
operated in the Android system (due to its larger market share) platform. The testing and the
validation of the U-analyzer is focused on the screen response speed and the touch range to be
met with the required operation specifications. The data access and transfer are designed to be
adjusted in accordance with the rules of the badminton ball game and the ways of the education
and training.
4. System Design and Verification Tests
(1) The User Interface and Process Definition Applied for the APP：
According to the characteristics and the rules involved during an openness badminton rally, the
preliminary design processes for the development of the user interfaces being operated in the
APP of the U-badminton tactics analyzer are shown on Figure 2.
247
2. Programming and the User Interface Design：
248
3. Programming and the Screen Description：
249
4. The Criterions for the Testing and Validation
(1) The basis of validation criteria：Based on the user identity, aim of the usability and
verification tasks to test the APP for badminton ball game record analysis, three evaluation
criterions are determined as 'learnability', 'efficiency' and 'effectiveness'. Two questionnaires
related to the system's ease of usability and satisfaction scale carried out in this study.
1. Learnability：It is evaluated, based on the time required to repeatedly work through the same
250
procedure, using the assessment formula.
2. Efficiency：The rapidity for the tasks to be fulfilled is estimated based on the total time to
complete all tasks.
3. Effectiveness：It is evaluated, based on the task performance showing success, partial
success or failure, using the assessment formula.
(2)The testing of the validation criteria：Four functions（tactics classifications ategory,
time-span per shot, time-span between two shots and time-consuming per game）of the APP are
tested by two badminton expert users and two badminton student-level users. The results of the
testing are summarized as the Table shown below：
5. Conclusions
1. The objective of this research is to develop a badminton tactics analyzer for the coaches,
players and scouts to carry out the real-time ball game data analysis.
2. The tactics analyzer with a customized information platform is capable of executing the
complete ball game record and subsequent data analysis, for example, the badminton technical
skill utilization rate such as serve, net shot, smash, passing shot, push shot, attacking clear, kill,
drive, drop shot...etc and the scoring patterns such as 'too late to hit the shuttlecock', out, 'touch
the net during a shot', misjudgment.
3. The evaluation （tested by three different user indentity）test results showed that the
real-time professional badminton tournament data record analyzer with a customized APP and
a supporting database display mode is capable of providing real-time match record data and
possible batting tactics analysis to the coaches and players and also significantly reducing the
chattiness of the traditional manual data record analysis.
4. The Preliminary objective evaluation data on the APP's learnability, efficiency, effectiveness,
usability and satisfaction are achieved in this study. It indicated that it is learnable and
convenient to execute the badminton ball game analysis using the APP. In order to obtain a
better and quick data analysis, it needs more practice for being familiar with the analyzer and
increasing the response speed.
251
The subsequent research in this study is aimed to develop a more appropriate presentation tool
platform（such as bigger touch screen mobile device with better screen resolution）in order for
achieving the optimum ball game record data analysis.
6. References
1. Jia-Jing Cai and Wu-Zhou Chen, ' a badminton teaching and training system developed
using Flash', 2009, Master Thesis of National Sport University.
2. Xuan-Jun Guo and Mao-Chun Yang, ' A visual analysis system applied for the table-tennis
ball game', 2008, Master Thesis of National Dong Hwa University.
3. Ming-Yue Wang, Ying-Jie Liu and Zhi-Quan Zheng, 'A technical and tactical analysis
model applied for the first three shots involved during a table-tennis ball game', 2012,
Journal of College sports, pp. 229-238.
4. Yu-Fen Chen and Cheng-Zhi Huang, ' A real-time technical analysis and match data record
for the table-tennis ball game player', 2008, International Symposium of Sports Pedagogy
and Sports Coaching in Taipei Physical Education College, Taipei.
5. V1Golf, 'V1Golf Victory one golf swing analysis software', 2013, http://www.v1golf.com/.
6. Valand, 'Visual swing golf digital imaging teaching software', 2013, www.wisinfo.com.tw/.
7. Ming-Chih Tsai and Ming- Guo He, 'A billiards image comparison system', 2010, Master
Thesis of Tatung University.
8. Zheng-LongYang, Jian-Ren Huang, Hong-Xiang Chen and Zhi-Hao Yu, ' Taijiquan
Competition Information Management System', 2003, Proceeding of the Taijiquan
International Symposium in 2003.
252
SICASE-759
Optimized Design of Geo-Temporal Query Based User Authentication
Scheme for Smartphones
Seonho Choi
Department of Computer Science, Bowie State University
Kun Sun
Department of Computer Science, College of William & Mary
Edward Jung
Department of Computer Science, Kennesaw State University
Abstract
As smartphones are gaining more popularity, it is becoming more critical to protect the data
and resources from malicious access. Many user authentication schemes were proposed and
studied for mobile devices. Biometric based, pattern based, and password based techniques are
some of the examples. We proposed a new authentication idea utilizing a series of binary
geo-temporal queries based upon user‟s episodic memory [16]. With the advancement of the
smartphone technology it became more than feasible to acquire and maintain some aspects of
the user‟s episodic memory such as locations and times where and when the user has been.
This technique has several strengths compared to the other user authentication techniques. It is
resistant to most of the attack types including the shoulder surfing attack, and its usability is
high due to the high memorability of such information. This scheme may be used either as a
primary authentication tool or as one of the techniques in multi-factor authentication. However,
we need to formulate its design process in such a way that various security requirements may
be specified and several parameters may be determined. In this paper, this design process is
formulated as an optimization problem and the solution approach is derived. By using this
algorithm, design time decision may be made on such parameters such as number of maximum
queries to be given out, the minimum number of hits needed for a successful authentication. On
Android 4.3, we built a prototype App, which provides user authentication to unlock a
Samsung Galaxy S4 smartphone with low power consumption.
Keywords: User authentication, Geo-Temporal Query, Smart Phone
253
1. Introduction
Mobile devices such as smart phones have been widely used for processing monetary
transactions and accessing sensitive data. The mobile device should be protected by an
authentication mechanism to prevent an unfriendly user who possesses the device even for a
short time from stealing valuable sensitive data. A number of authentication techniques have
been proposed and developed for mobile devices. Some of the most commonly used techniques
include conventional password based scheme, keystroke dynamics based technique,
pattern-based scheme, physiological biometric based techniques, etc.
As more users are utilizing mobile devices for carrying out various tasks including monetary
transactions, it is becoming more and more critical to protect them against malicious accesses.
A lot of user authentication techniques have been proposed and developed for this purpose
including physiological or behavioral biometric techniques, pattern-based scheme,
conventional password based scheme, etc.
However, most of the previously developed user authentication schemes suffer from various
limitations including the low resistance against some attack types such as shoulder surfing
attacks, low memorability of the required information to be used for authentication such as
passwords or phrases, high false negative and/or false positive rates, requirements for
cumbersome training steps needed for some biometric-based schemes, or dependence on
special hardware units such as fingerprint readers.
We proposed a new authentication idea utilizing a series of binary geo-temporal queries based
upon user‟s episodic memory [16]. With the advancement of the smartphone technology it
became very feasible to acquire and maintain some aspects of the user‟s episodic memory such
as locations and times where and when the user has been. This type of personal episodic
memory information may be stored along with their temporal information, and a series of
geo-temporal queries may be created and asked to a user. Even though various types of
geo-temporal queries may be generated we chose to focus on utilizing binary geo-temporal
queries for its simplicity and applicability. A user is given a series of binary queries and he/she
needs to choose true or false on each query.
Our geo-temporal query based scheme has the following characteristics [16]:
Usability: users can easily answer a number of true/false queries for authentication
purpose.
Attack resistance: it can protect against various attack types including the shoulder surfing
attack.
254
Quantification and Regulation of security risks: it should allow us to (probabilistically)
quantify the security risks such as the authenticity probability given user‟s response, expected
false positive and false negative rates, etc.
Implementation feasibility: our technique should be implementable on most mobile
platforms without any special HW requirements.
However, personal-history based techniques have an inevitable drawback of privacy disclosure
problem. For instance, in geo-temporal query technique queries include location and temporal
information, and an adversary may obtain correct history information once the true queries –
whose answers are true – are identified out of the false queries – whose answers are false –
among the queries observed during an attempted authentication process. To minimize the
privacy disclosure level we propose two pre-termination techniques for the query process,
success pre-termination and failure pre-termination.
However, we need to formulate its design process in such a way that various security and
privacy requirements may be specified and several parameters may be determined. In this
paper, this design process is formulated as an optimization problem and the solution approach
is derived on how the optimization problem may be solved at system design time. By using our
formulation and algorithm, design time decision may be made on such parameters such as
number of maximum queries to be given out, the minimum number of hits needed for a
successful authentication, the maximum number of misses that will automatically fail the
authentication, the true (or false) query generation probability, etc. On Android 4.3, we built a
prototype App, which provides user authentication to unlock a Samsung Galaxy S4 smartphone
with low power consumption.
The remainder of the paper is structured as follows. We present the threat model and
assumptions in Section 2. Section 3 presents a geo-temporal authentication framework. In
Section 4 details of our authentication scheme is presented. Initial prototype app developed on
Android platform is introduced in Section 5 with implementation details and issues.
Discussion and future works are presented in Section 6. Section 7 discusses the related works
on user authentication for mobile devices. Section 8 concludes the paper.
2. Threat Model and Assumptions
We consider someone as an attacker who has physical access to a smart phone at least for a
limited time. Even though an attacker has physical access to a phone, he/she may get access to
the resources in the smart phone only after successful authentication. An attacker is assumed to
have knowledge on various system parameter values. We assume the mobile devices are
equipped with GPS to record the location information. Most smart phones have GPS included.
255
GPS typically cannot function indoors; however, we won‟t require such fine-grained
geo-location information.
3. Geo-Temporal Authentication Framework
3.1 Overview
錯誤! 找不到參照來源。 shows the geo-temporal authentication process. User‟s location
history information is obtained by using the GPS component. The location coordinates are
periodically sampled and stored along with the sampling time information. Once the coordinate
values are obtained from the GPS, the coordinates and sampling time will be stored into the
Location History Base. When a user attempts to unlock the smart phone, the Query Generator
and Processor component will make use of the Location History Base to create a series of
geo-temporal queries to be answered by the user. Given a geo-temporal query, if the user‟s
answer is correct, then it is said that a hit occurred. Otherwise, it is said that a miss occurred.
Multiple queries may be challenged to allow for some operation mistakes or vague memory.
3.2 Geo-Temporal Data Engine
For our scheme to work we need to create and regularly update a geo-temporal database
containing the location history information sampled with time information. The necessary
entries in each record are sampling time, longitude, latitude, address (obtained by using the
Reverse Geocoding), and duration of stay.
There are tradeoffs between GPS (or other sensors) sampling period and energy consumption.
Using smaller sampling period will produce more accurate traces of user locations, while
consuming more energy. Less accuracy may be achieved with bigger sampling period, but less
power will be spent.
3.3 Authentication Query Generator and Processor
In this paper, we focus on using the binary geo-temporal query, whose answer should be either
true or false, and which contains both the location and temporal information in the query itself.
256
3.3.1 Scaled Geo-Temporal Granularities
There is a correlation between human memorability and location granularity. The finer the
location granularity is, the more difficult to remember.
If a user is given an address containing street number and name, he/she would have a more
difficult time in remembering compared to the case when he/she is challenged only with city
name. For example, it would be more difficult to answer “Did you visit 1234 Main Street, New
York, NY last May?” than to answer “Did you visit New York, NY last May?”
Also, the finer the temporal granularity is, the more difficult to recall. For example, consider
the following two queries: “Did you visit New York, NY, last May?” and “Did you visit New
York, NY, between May 12 and 14?” In this case, the latter query would be more difficult to
answer.
The relationship between location and temporal granularities may be represented by different
regions in the scaled X-Y graph as is shown in Figure 2. Out of the possible regions in the
graph, we will focus on those along the correlation line. These regions will have the similar
level of location and temporal granularities. “Did you visit Canada in 2009?” is an example
query in region 5, and “Did you visit Quebec, Canada, in May 2009?” in region 4. As is noted
in the figure, more privacy will be disclosed by the low-numbered regions. But, more accurate
authentication may be provided.
However, human memory tends to fade away as time goes by. They can remember recent
events more accurately compared to the ones in remote past. For example, the places visited
yesterday may be remembered more clearly along with temporal information than the ones
visited a month ago. This implies that we may utilize the queries in the low-numbered regions
257
only from the recent location history, and from the remote location history those in the
high-numbered regions may be used in the authentication process. In Figure 2 the
low-numbered regions are characterized with short memorability spans and the high-numbered
ones with long memorability spans. The temporal remoteness of the location samples needs to
be considered in generating the queries as well as the location and temporal granularities.
When queries are generated in our approach, the temporal and location granularities of queries
will be adjusted to reflect this correlation with the remoteness of the location history. We name
this technique as Scaled Geo-Temporal Granularities. The following shows the example types
of queries that are generated according to this principle:
Did you visit 1234 35-th Street, New York, NY at 3:15pm today? (Region 1 in Figure 2)
Did you visit 35-th Street, New York, NY the day before yesterday? (Region 2 in Figure
2)
Did you visit New York, NY 3 weeks ago? (Region 3 in Fgiure 2)
Did you visit NY state 3 last May? (Region 4 in Figure 2)
Did you visit Canada in 2010? (Region 5 in Figure 2)
3.3.2 False Query Generator
False queries need to be obtained and presented to a user along with the true queries. Simple
approach may be to generate totally random location and time. However, the locations used in
the false queries may not be too far from the locations that have been visited by a user. For
example, if an adversary acquired a phone in New York City and asked whether the user was in
California that morning, it would be easy to guess the answer. For this reason we adopted to
base the false query generation process upon the information stored in the Location Base, and
introduce some randomness to create false queries. The basic idea is to choose a location
sample from the Location Base as is done in generating the true query, and add/subtract
random numbers to/from the coordinates and sampling time of the chosen sample.
4. Authentication Scheme
Multiple queries may be challenged to allow for some operation mistakes or vague memory.
There is a constraint on the number of queries that can be given to a user. We will denote the
maximum number of queries that can be given as M. A series of geo-temporal queries will be
given until one of the following conditions met:
Authentication success: enough number of hits made after M queries are given.
Authentication failure: enough number of hits not made even after M queries are given.
We propose two pre-termination techniques to reduce the number of disclosed true queries:
success pre-termination and failure pre-termination. First, if the number of hits reaches a
pre-determined threshold H, then the query process terminates early even before M queries are
258
given out. Second, even before all of M queries are presented to a user, the query process may
be terminated if there is no possibility that the constraints will be met even if the remaining
queries are answered correct. Figure 2 shows the basic authentication algorithm.
It is assumed that H > M-H, which must be true for the most of the practical parameter values.
This means H > M/2.
Suppose that k queries have been presented to a user, and let h denote the number of queries
answered correct (hits) by the user. The number of remaining queries is M-k, and the maximum
number of hits that may result out of M queries (k already answered and M-k to be answered) is
h+(M-k) under the assumption that all of the M-k remaining queries are answered correctly. If
h+(M-k) < H, then there is no hope that H queries will be answered correct even after the
remaining queries are given and answered. This corresponds to the failure pre-termination case.
259
4.1 Security Constraints
We will design the authentication scheme in which a guarantee may be given in term of the
following security constraints:
Pauth : When an authentication is granted, a probability that a user is authentic should be
greater than or equal to this value.
PFP: The probability that a valid user is not authenticated should be less than or equal to
this.
PFN: The probability that an invalid user is authenticated should be less than or equal to
this.
Given these parameter values along with Mmax, we need to determine the values of M, H, Ptrue,
and Pfalse, that will be used in the actual query generation and challenge process.
Thus, we obtain the following optimization problem.
Optimization Problem 1: Find out values of M, H, Ptrue and Pfalse (1≤H≤M) satisfying the
given PAuth, PFP, PFN and Mmax conditions.
It is possible that there may be more than one candidate values for H. In this case the minimum
value will be chosen as H. Here, we are supposed to find minimum values for M (and H)
satisfying the security constraints. pf=P(Cc|A) is the probability that an authentic user makes a
mistake in answering a query. This probability may be experimentally obtained. This may vary
according to different factors such as different persons (with different memory capability),
260
different scales for Geo-Temporal Query Granularity settings. In this paper, we start with some
initial probability (e.g., 0.1), and dynamically adjust this value based upon the user‟s responses
whenever a user authentication succeeds. We use Exponentially Weighted Moving Average
algorithm to update this probability.
P(A) is the probability that a user is authentic. It is known that 18.13% of cell phones are lost or
stolen each year. So, we set this to be P(A)=1-0.1813=0.8187. This leads to P(Ac)=0.1813.
Suppose that M queries are presented to a user, and let h denote the number of queries actually
answered correct by the user. Then, 0≤h≤M holds. The following probabilities may be
obtained.
( | ) ( | ) ( )
( | ) ( )+ ( | ) ( ) (1)
( | ) ∑ ( )( )
−
(2)
( | ) ∑ (
)(
) ( )
−
(3)
The first probability is obtained by applying the Bayes Theorem. P(A|NM≥h) represents the
probability that a user is an authentic user when the user provides h or more correct answers to
M queries. Given P(A) and pf values, this is a function of M and h.
We assume the true queries whose answer should be “true” are randomly generated with a
probability Ptrue, and false queries with a probability of Pfalse = 1-Ptrue. These probabilities can
be known to an adversary. Given a query, the probability that an adversary provides a correct
answer is (Ptrue)2+(Pfalse)
2=1-2Ptrue+2(Ptrue)
2, and the probability for providing an incorrect
answer is 1-(Ptrue)2-(Pfalse)
2=2Ptrue-2(Ptrue)
2.
The number of queries to be asked to a user varies in [0, M]. PFP is the target threshold
probability for false positives which occur when a valid user is not authenticated. If this is
0.01, then it should be guaranteed that at most 1% of the cases a valid user is rejected (e.g., by
user mistakes). PFN is the target threshold probability for false negative when an invalid user is
authenticated. If this is 0.005, then it should be guaranteed that at most 0.5% of the cases will
result in authentication even though the user is not authentic. This may happen when an
adversary launches a random guessing attack. Mmax is an upper bound on M, and it is given at a
design time. This condition will be used as one of the constraints in the optimization process.
For an authentication to be granted for the user with (M, h) values – i.e., NM≥h to be used as an
authentication criteria, the following should hold.
261
P(NM≥h | Ac) PFN (4)
P(NM<h | A) = 1 - P(NM≥h | A) PFP (5)
P(A | NM≥h) ≥ PAuth (6)
We may choose the minimum value of M for which at least one integer h (1≤h≤M) exists
satisfying the above conditions. Then, we may denote them as M and H, respectively. Again,
for more than one candidate values for H we select the minimum value.
4.2 Privacy Metrics
In the authentication scheme, a series of queries are presented to a user based upon the location
history file; however, this may endanger user privacy by revealing locations frequently visited
or other location-related information. To overcome this limitation, we reformulate our
optimization problem as follows:
Optimization Problem 2: Find out values of M, H, Ptrue and Pfalse (1≤H≤M) satisfying the
given PAuth, PFP, PFN and Mmax conditions and minimizing the number of true queries exposed.
It is possible that there may be more than one candidate values for H. In this case the minimum
value will be chosen as H.
Suppose that an adversary tries to obtain as much location history information from the smart
phone as possible by launching a guess and check attack. Also, let‟s assume that k queries
have been given to the adversary so far. Then there are three possibilities:
Successful pre-termination condition: k≥H, H correct answers and k-H incorrect answers.
Failure pre-termination condition: k≥M-H+1, and there were M-H+1 incorrect answers
and k-M+H-1 correct answers. M-H represents a maximum number of queries that may be
missed before a failure pre-termination occurs.
Query continuation condition: the other cases except for the above two conditions.
We may divide the cases into the following three cases depending upon the values of k.
H ≤ k: Success pre-termination and failure pre-termination are mutually exclusive events.
That is, if both conditions hold at the same time, then we have k-H=M-H+1, which yields to
k=M+1. However, this is a contradiction.
M-H+1 ≤ k <H: Failure pre-termination condition and the query continuation are mutually
exclusive events. Again, failure pre-termination occurs when there are M-H+1 incorrect
answers and k-M+H-1 correct answers.
k ≤ M-H: Only query continuation is the possible state.
Let xi (1≤i≤k) denote a random binary variable which has a value of 0 when a user answers
wrong for the i-th query, and 1 when a user answers correctly. And, let yk=x1+x2+…+xk
denote a sum of x1 through xk, which represents the number of correctly answered queries out
of k queries. Then, we may trace the values of yk over the index values from 1 through k for
262
various cases including success pre-termination, success termination, failure termination as is
shown in Figure4 Through Figure 9. In these figures, it is assumed that M=7 and H=5
In Figure 4 success pre-termination case is shown. In the graph, at each k index value there are
two possible cases, either yk jumps up by one unit or stays the same as before. In this case the
user answers correctly to the first 5 queries without any misses, and the query process
pre-terminates only after 5 queries are given. The process terminates when yk=5 is reached,
which occurs for k=5 in this example.
Figure 5 shows a success termination case where all of the M queries are given out without
pre-termination. In this scenario 5 queries are answered correctly, but with 2 misses in between.
Because of the misses, the termination occurs at k=7 where yk=5 is reached.
263
Figure 6 shows a failure pre-termination case where the query process is terminated after
giving out 5 queries. 2 queries were answered correctly and 3 queries were answered
incorrectly, which yields to the condition where an authentication is impossible even if the user
answers the remaining 2 queries. Note that the failure termination occurs when the trace line
hits the Termination Line.
An example scenario for failure termination is given in Figure 7. The last query q7 must have
been answered wrong. Otherwise, a failure pre-termination should have occurred at an earlier
query response.
Different points that a trace line may end at k=5 is shown in Figure 8. In this specific scenario
there are 4 such points denoted as a, b, c, and d. The possible y5 values at a, b, c and d are 2, 3,
264
4 and 5 where y5=x1+x2+x3+x4+x5. Point a corresponds to a success pre-termination, and point
d corresponds to a failure pre-termination. At points b and c the query process will continue
without termination. It is again assumed that H=5 and M=7.
In Figure 9 all the possible termination points are shown for H=5 and M=7. We may divide
the entire area into three subareas, A, B, and C. Area A corresponds to the condition (k ≤ M-H),
and no termination is possible. Area B corresponds to (M-H+1 ≤ k <H), and either
continuation or failure pre-terminations are possible. Area C satisfies the condition (H ≤ k), and
continuation, failure pre-termination, failure termination, success pre-termination or success
termination is possible.
We want to find the expected number of disclosed true queries as a privacy metrics. As is
shown in Figure 9, the termination points in the (k, yk) graph may belong to three different
regions denoted as A, B, and C. We need to find the probabilities of the query process
terminating at these points.
Probabilities in Area C (H ≤ k)
When a k-th query is given and answered where H ≤ k, the possible range of yk values are
[k-(M-H)-1, H]. For example, this corresponds to the point d in Figure 8. The minimum value
k-M+H-1 corresponds to a failure pre-termination and occurs when exactly M-H queries are
answered wrong and k-1-M+H queries are answered correct to the queries, q1, q2, … and qk-1,
and the last query qk is answered wrong. If more than M-H queries are answered wrong to the
queries q1, q2, … and qk-1, then a failure pre-termination must have occurred before the k-th
query was given. The probability of this failure pre-termination happening at k is given as:
( )
( − −
)( )
− + (
) − − +
(7)
265
The probability of yk=k-(M-H) – e.g., point c in Figure 8 – may be obtained by considering the
cases where M-H queries are answered wrong and k-M+H queries are answered correct to the k
queries q1, q2, … and qk. This is given as
( ) (
− + )(
) −
( ) − +
(8)
By applying the similar argument, the probability of yk=k-(M-H)+n, where 0 ≤ n ≤ M-k, may
be obtained by considering the cases where M-H-n queries are answered wrong and k-M+H+n
queries are answered correct to the k queries q1, q2, … and qk. This is given as
( )
( − −
)( )
− − (
) − + +
(9)
When n=M-k, a success pre-termination occurs for k<M and a success termination occurs for
k=M. Hence, the probability of success termination when the k-th query is given and answered
is
( ) ( )(
) −
( )
(10)
Probabilities in Area B (M-H+1 ≤ k <H)
In this case only failure pre-terminations are possible. When a k-th query is given and answered,
the possible range of yk values are [k-(M-H)-1, k]. The minimum value k-M+H-1 corresponds
to a failure pre-termination and occurs when exactly M-H queries are answered wrong and
k-1-M+H queries are answered correct to queries, q1, q2, … and qk-1, and the last query qk is
answered wrong. If more than M-H queries are answered wrong to queries q1, q2, … and qk-1,
then a failure pre-termination must have occurred before the k-th query was given. The
probability of this happening is given as:
( )
( − −
)( )
− + (
) − − +
(11)
Probabilities in Area A (k ≤ M-H)
In this case there is no possibility of terminations. The query process will only continue in this
area.
266
Privacy Metric: Expected Number of True Queries Exposed
We will use the expected number of true queries that may be exposed to an adversary when
he/she attempts to be authenticated. It is assumed that an adversary has knowledge on Ptrue and
Pfalse values and employs a random guessing strategy in answering the queries. That is, he/she
will try to answer a query with “true” with the probability of Ptrue. Then, the expected number
of true queries that are given to the adversary may be found as follows from (7), (10) and (11):
∑ ( )
∑ ( ) ∑ ( )
−
− +
∑ ( − −
)( )
− + (
) − − +
∑ ( )(
) −
( ) ∑
− − +
( − −
)( )
− + (
) − − +
(12)
This E represents the expected number of true queries exposed to an adversary. This formula is
used in our Optimization Problem 2 definition as a metric for privacy disclosure.
4.3 Solution Approach for the Optimization Problem
Our Optimization Problem 2 may be refined by incorporating the constraints on PFP, PFN and
Pauth, and by having an optimization objective function E – given in (12) – to be minimized.
One simple approach would be to exhaustively search for optimal solution by trying out
different combinations of (Ptrue, M, H) values. Figure 10 shows this algorithm. Ptrue is a
continuous variable and needs to be quantized with a gap of Δ. In our example Δ was chosen to
be 0.001.
267
4.4 Example Optimization
Given the following parameter values, we applied an optimization algorithm to find (Ptrue, M,
H) values satisfying the probability constraints and minimizing the expected number of
exposed true queries to a potential adversary. The input parameters are Pauth= 0.9, PFP = 0.05,
PFN = 0.05, pf = 0.05.
The optimal (M,H) was found to be (11,9) when Ptrue=0.378 with an expected number of
disclosed true queries equal to 2.391. Please note that, when Ptrue=0.5, the E value would be
equal to 2.986.
The graph in Figure 11 shows the points on (M,H) plane meeting the probability constraints
and the z axis shows the expected number of true queries to be disclosed for each (M,H)
combination.
268
5. Implementation
We constructed a prototype app on Android 4.3, and ran it for an hour to measure power
consumption on Samsung Galaxy S4 smartphone. As was shown in previous research [15],
invoking GPS unit and obtaining coordinates consumes a significant amount of power. Hence,
we decided to invoke the GPS unit periodically with a fixed sampling period. Table 4 was
obtained only by running the tracking component of the app, which will use most of the energy
due to GPS samplings.
Figure 12 shows an example screenshot with a geo-temporal query. User is supposed to choose
either the “true” or “false” button. This process is repeated between m and M times depending
upon the hit & miss patterns of the user responses. Note that only one query is given each time
since the termination may occur at any time and we want to minimize the number of true
queries exposed.
6. Discussion & Future Works
Through experimentations, we need to validate the correlations used in generating the
geo-temporal queries between temporal remoteness of the sample, location granularity and
temporal granularity. Or, more adequate correlation models may be found through such process.
The key issue is how we design the memorability model and validate it. Also, we may need to
elaborate and corroborate the parameter selection for the false query generation process.
269
To increase the memorability, we may develop several schemes such as a pre-hinting technique
to remind users of the locations he/she is visiting or recently visited as a preparation for future
authentication queries.
As was shown in our and previous works, GPS consumes significant energy and it would be
vital to activate the unit only when necessary. To save energy for location sampling we may
invoke the unit only when the user is in move by utilizing other sensors such as accelerometer
sensor.
7. Related Works
Besides using a pre-agreed set of shared secrets, we can authenticate users by challenging them
with questions dynamically generated from a wider base of personal resources, such as
calendar [7], email [6], or twitter [8]. Nosseir et al. [7] propose to query the users based on
their personal history and develop a prototype using automatic generation of questions from
electronic Calendar information. Nishigaki and Koike [6] propose a user authentication using
user‟s email history to generate the questions. However, it has a serious problem on leaking
privacy, since the impersonators are able to read the mail content when they are trying to log
on as the legitimate user. Okamoto [8] proposes to use Twitter to collect simple, memorable
question/answer pairs. It requires the uses to manually input the knowledge base, so it is not
practical to be used on the smart phones.
Biometric authentication has been well-studied too. Biometric authentication techniques use
either physiological or behavioral features [3, 10]. Physiological biometrics utilizes static
physical features of humans that are known to be unique for each individual. These may be
based upon fingerprint [2], face patterns [11, 1], iris [4], or speech patterns [9], etc. Among
these the most common technique is to use fingerprints to authenticate users. Built-in
fingerprint readers have been integrated in iPhone 5S.
Behavioral biometrics utilizes the fact that people have distinct stable patterns on a certain
behavior, such as keystroke on a keyboard. Two of the most widely studied techniques for
personal computers are keystroke-based [12] and mouse movement based [13]. However, due
to the size limitations it is difficult to apply these techniques to smart phones. An approach was
proposed to provide authentication by capturing the finger movement patterns and comparing
them against the owner‟s patterns [5]. Also, biometric technique based upon personal gait was
proposed using accelerometer sensors [14]. The behavioral biometrics may provide usability
and attack resistance, but most of them utilize complex algorithms for learning and matching
the behavioral patterns and they are vulnerable to high rates of false positives and false
negatives. Also, if a user‟s biometric is stolen, it would be impossible to change it. Because of
270
these reasons behavioral biometrics are generally used in conjunction with a password/PIN
rather than used as a sole authentication mechanism.
8. Conclusions
We introduced and developed an authentication scheme utilizing a series of binary
geo-temporal queries based upon the users‟ location history information. Two pre-termination
techniques were also proposed to minimize the privacy leakage level. We formulated a design
time optimization problem with various security constraints and an objective function to
minimize the privacy disclosure level. The iterative solution approach was presented to the
optimization problem. A prototype Android app was developed and the implementation details
and issues were presented along with future works.
We believe that the proposed approach satisfies the design objectives mentioned in the
beginning of this paper: implementation feasibility, attack resistance, quantification and
regulation of security risks, and usability.
References
[1] J. Daugman. High confidence visual recognition of personsby a test of statistical
independence. Pattern Analysis and Machine Intelligence, IEEE Transactions on,
15(11):1148–1161, 1993.
[2] A. Jain, L. Hong, S. Pankanti, and R. Bolle. An identity-authentication system using
fingerprints. Proceedings of the IEEE, 85(9), 1997.
[3] A. K. Jain, R. Bolle, and S. Pankanti, editors. Biometrics: Personal Identification in
Networked Society. Kluwer Academic Publishers, 1998.
[4] S. Kurkovsky, T. Carpenter, and C. MacDonald. Experimentswith simple iris recognition
for mobile phones. In Proceedings of the 2010 Seventh International Conference on
Information Technology: New Generations, ITNG ‟10, pages 1293–1294, 2010.
[5] L. Li, X. Zhao, and G. Xue. Unobservable re-authentication for smartphones. In ISOC
Network and Distributed System Security Symposium (NDSS), February 2013.
[6] M. Nishigaki and M. Koike. A user authentication based on personal history: A user
authentication system using e-mail history. 47(3):945–956, 2006.
[7] A. Nosseir, R. Connor, and M. Dunlop. Internet authentication based on personal history - a
feasibility test. In Proceedings of Customer Focused Mobile Services Workshop at
WWW2005, 2005.
[8] M. Okamoto. Knowledge-based authentication using twitter can we use lunch menus as
passwords? 5(5).
[9] K. N. Stevens, C. Williams, J. Carbonell, and B. Woods. Speaker authentication and
identification: a comparison of spectrographic and auditory presentations of speech
material. The Journal of the Acoustical Society of America, 44:1596, 1968.
271
[10] J. Wayman, A. Jain, D. Maltoni, and D. Maio. An introduction to biometric authentication
systems. In J. ayman, A. Jain, D. Maltoni, and D. Maio, editors, Biometric Systems,
pages 1–20. Springer London, 2005.
[11] J. Zhang, Y. Yan, and M. Lades. Face recognition: eigenface, elastic matching, and neural
nets. Proceedings of the IEEE, 85(9):1423–1435, 1997.
[12] F. Bergadano, D. Gunetti, and C. Picardi. User authentication through keystroke dynamics.
ACM Transactions on Information and System Security (TISSEC), 5(4):367–397, 2002.
[13] N. Zheng, A. Paloski, and H. Wang. An efficient user verification system via mouse
movements. In Proceedings of ACM CCS2012, pages 139–150. ACM, 2011.
[14] D. Gafurov, K. Helkala, and T. Søndrol, Biometric Gait Authentication Using
Accelerometer Sensor. Journal of Computers, Vol. 1, No. 7, Oct./Nov. 2006.
[15] K. Lin, A. Kansal, D. Lymberopoulos, F. Zhao, Energy-accuracy trade-off for continuous
mobile device location. Proceedings of the 8th international conference on Mobile
systems, applications, and services. Pages 285-298. 2010.
[16] B. Choi, K. Sun, S. Choi, Cloud-Based User Authentication with Geo-Temporal Queries
on Smartphones. Proceedings of the 2nd
International Workshop on Security in Cloud
Computing, Kyoto, Japan, June 3, 2014.
272
Mechanical Engineering
Session Chair: Yanhua Wu
SICASE-704
Flow Topologies of Energetic Turbulence Structures in the Separated Flow Over an
Airfoil at Deep Stall
Yanhua Wu︱Nanyang Technolocial University
Hui Tang︱Hong Kong Polytechnic University
SICASE-722
The Fracture Prediction of Electromagnetic Steel for Shorten Reheating Process in
Furnace
Yu-Hsi Huang︱National Taiwan University of Science and Technology
Ching-Kong Chao︱National Taiwan University of Science and Technology
Li-Wen Wu︱China Steel Corporation
Chia-Hsuan Cheng︱National Taiwan University of Science and Technology
Zhen-Yi Fan︱National Taiwan University of Science and Technology
SICASE-820
Direct Current Compressor for Vehicle Air Conditioning System in Non-Electric Vehicles
Henry Nasution︱Universiti Teknologi Malaysia
Amirah Haziqah Zulkifli︱Universiti Teknologi Malaysia
Afiq Aiman Dahlan︱Universiti Teknologi Malaysia
Azhar Abdul Aziz︱Universiti Teknologi Malaysia
SICASE-819
In-Cabin Temperature Control and Multi-Circuit Bus Air-Conditioning Systems
Henry Nasution︱Universiti Teknologi Malaysia
Azhar Abdul Aziz︱Universiti Teknologi Malaysia
Amirah Haziqah Zulkifli︱Universiti Teknologi Malaysia
Afiq Aiman Dahlan︱Universiti Teknologi Malaysia
2015/6/28 Sunday 14:45-16:15 Room 2
273
SICASE-704
Flow Topologies of Energetic Turbulence Structures in the Separated Flow
over an Airfoil at Deep Stall
Yanhua Wu
Nanyang Technolocial University, Singapore
Hui Tang
Hong Kong Polytechnic University, China
Abstract
A special application of proper orthogonal decomposition (POD) as in Wu (2014) has been
applied to study the flow topologies of large-scale energetic turbulence structures in the
separated flow over an S805 airfoil designed for horizontal-axis wind turbine blade at deep
stall with a large angle of attack of 30. Two kinds of instantaneous turbulence structures are
observed to contribute significantly to the first POD mode. One is the complete or incomplete
large-scale vortex with stronger induced vertical upward flow. The other is the three-vortex
structure with one large-scale vortex of similar size to that existed in the first kind of energetic
flow structure and two small-scale vortices located at the leading- and trailing-edge shear
layers, respectively.
Keyword: Turbulence structure, proper orthogonal decomposition, airfoil, stall
1. Introduction
The rotating blades of engineering turbomachinery are often operated at large angles of attack
(AOAs) at off-design conditions. For example, Lee & Wu (2013a, 2013b, 2014)[1-3] found
that when a blade of a horizontal-axis wind turbine with a design tip-speed-ratio (TSR) of 6 is
operating at TSR=3, the AOA at the inboard section of the blade is as large as 30. In order to
further improve the aerodynamics of the wind turbine at off-design conditions, the improved
understanding of the airfoil at full stall conditions is needed. At very large AOAs, the boundary
layer is separated from the leading-edge of the airfoil and the flow is turbulent on the suction
surface. Although the aerodynamics loads (pressure, lift, and drag coefficients) have been
studied quite extensively in the past, the three dimensional instantaneous turbulence structures
inside the turbulent separated flow at full stall are still illusive. In this study, the state-of-the-art
tomographic PIV will be used to measure the three dimensional velocity fields within a volume
of the separated flow. The three-dimensional topologies of the instantaneous large-scale
structures which contain a significant amount of turbulent kinetic energy will be revealed using
a special application of the proper orthogonal decomposition as in Wu (2014)[4].
274
2. Experiments
Experiments were conducted in a wind tunnel with a test section of 0.4m0.4m with a
freestream turbulence intensity of about 0.4%. The side walls of the wind tunnel are made of
glass and the top and bottom walls are made of acrylic for optical access in the tomographic
PIV measurements to be described below. The airfoil used here is S805 designed by NREL for
the purpose of horizontal-axis wing turbines. The airfoil which covers the whole width of the
test section was mounted at the vertical center. The chord length of the airfoil was chosen to be
30 mm so that the Reynolds number of the flow is similar to that of the blade model studied in
Lee & Wu (2013a, 2013b, 2014)[1-3] to be Re=5500. The results presented in this study were
from a large angle of attack of 30 where the airfoil was in deep stall with massively separated
flow.
Due to the intrinsic three-dimensionality of the separated flow on the suction surface of the
airfoil at large angle of attack, Tomographic PIV (Tomo-PIV) was used to measure the velocity
fields within the volume of fluids. The experimental setup is shown in Figure 1. A 8mm-thick
laser light volume illuminate the flow at the suction side of the airfoil at its spanwise center.
The images were recorded by four 16-bit CCD camers with 105mm lenses and each camera‟s
CCD chip has 24562058 pixels whose size is 3.45m. The cameras were configured in such a
way that the backward-scattering of the laser light from the particles were captured yielding
similar particle intensities in the four recordings. To reduce the laser light reflection from the
airfoil surface, the airfoil was painted with Rhodamine B and band-pass filters (53210nm with
>95% transmission) were used for each camera. FastMART algorithm was applied in LaVision
Davis software to reconstruct the three-dimensional light intensity field of the particles.
Volume self-calibration was also implemented to improve the calibration accuracy.
Three-dimensional velocity vector fields were obtained by recursive volumetric
cross-correlation method. Each resulting velocity field has 1069211 (xyz) vectors with a
vector grid spacing of 0.76mm.
275
The total field of view is about 80mm70mm8mm. 500 statistically independent velocity
fields were obtained. The overall uncertainty of the current Tomo-PIV measurements is about
2% of the full-scale velocity.
3. Proper Orthogonal Decomposition
The principle and equations of POD can be found in many places in the literature and they are
only summarized below. In this study, well-established snapshot POD was used on the
three-dimensional fluctuating velocity fields measured by the current Tomo-PIV. In POD, any
instantaneous fluctuating velocity field u(x,t) can be decomposed into the form of
1
)()(),(n
nn xtatxu , where n(x) is deterministic spatial POD modes, and an(t) is random
temporal coefficients. In snapshot POD, the coefficients an(t) are obtained first by solving the
eigenvalue problem with a positive definite Hermitian kernel of the form
')'()',(),()( dttadxtxutxuta n
T
nn
, where the integrations are over the spatial domain
and a time interval T. For an ensemble of measured velocity fields, the spatial domain could be
the whole or part of the field of view and the time domain represents the ensemble or the
collection of samples of the velocity fields. The eigenvalues n are real and positive and form a
decreasing and convergent series. The POD modes can then be computed through the equation
T
n
T
nn dtadtuax 2/)()( . The turbulent kinetic energy is equal to half of the summation of the
eigenvalues. Since the POD modes are usually normalized to be orthonormal, the POD
coefficients are related to the eigenvalues by T
nn dtta 2
)( , which dictates that the instantaneous
fluctuating velocity field with a larger value of 2
na contributes more to the eigenvalue n.
Therefore, equation 5 provides a possible means to identify what are the instantaneous
fluctuating flow structures that contribute significantly to the first few dominant POD modes. It
is this equation that the current study is based on to study the topologies of the large-scale
turbulence structures contributing significantly to the first two POD modes. The turbulence
structures contributing significantly to other higher POD modes were not sought in this work,
however, since each of the rest modes accounts for a much smaller amount of kinetic energy.
Since the contribution to the eigenvalue n comes from 2
na , instantaneous velocity fields with
either positive an or negative an contain equivalently important information on the flow
structures relevant to the nth
POD mode.
276
4. Results
The first POD mode was found to account for 28%, or almost a third, of the total turbulent
kinetic energy in the current flow. A vector plot of the first POD mode in the x-y plane at the
center of the measurement fluid volume is presented in figure 2. It shows a single large-scale
counter-clock-wise vortex with strong induced flow. The induced flow on the right side of the
vortex appears stronger than that on the left side. This vortex is located quite far from the
leading edge of the airfoil, but relatively much closer to the airfoil‟s trailing edge.
An instantaneous fluctuating flow field which is a significant contributor to the first POD mode
is shown in figure 3. This field has a large magnitude of positive coefficient of the first POD
mode whose a1=2.06a1, where a1 is the root-mean-square (RMS) value of the coefficients of
the first mode of the ensemble of the flow fields measured. The major flow topology in this
field is a huge vortex having the same rotational sense as the one in the first mode. Upstream of
it near the leading edge shear layer are a few small scale and much weaker vortices. The
fluctuating flow close to the airfoil‟s suction surface is very weak, corresponding to a
low-turbulence region which is typical in the separated flow at deep stall. An examination of all
the eight fluctuating fields whose a1 values are more than twice a1 reveals that a well-defined
large vortex like the one in figure 3 is not always present in every field. However, the common
feature of the flow topology is the large region of strong upward flow like that on the right side
of the vortex shown in figure 3.
277
An example of the instantaneous flow field with a large magnitude of negative value of
a1=-2.05a1 is shown in figure 4. In contrast to the flow topology in figure 3, there exist three
major vortices in this field: one large-scale clockwise vortex, and two small-scale vortices
upstream. The small vortex located at the leading-edge shear layer seems in this field to couple
with another weaker vortex to form a vortex ring with induced vertical flow in between. The
small vortex in the trailing-edge shear layer is closer to the huge vortex and these two vortices
are apparently strongly interacting with each other. These three vortices are the most
commonly observed features in all the eight fields whose absolute values of the negative a1 are
larger than 2a1 .
278
5. Conclusions
Flow topologies of instantaneous turbulence structures contributing significantly to the first
POD mode have been revealed in the separated turbulent flow over a static airfoil for the
horizontal axis wind turbine at deep stall of AOA=30. Two kinds of flow structures can be
observed: (1) Single complete or incomplete counter-clock-wise large-scale vortex with strong
induced upward flow on the right side of it; (2) Three-vortex structure with one large-scale
vortex of similar size to the one in the first kind of flow structure and two small vortices
located on the leading- and trailing-edge shear layers each. The small vortex at the
trailing-edge shear layer has a stronger interaction with the large-scale downstream vortex.
6. Reference
Lee H.M and Wu Y. Experimental study of rotational effect on stalling. Chinese Physics
Letters, 30, 064703, 2013.
Lee H.M and Wu Y. An experimental study of stall delay on the blade of a horizontal-axis
wind turbine using tomographic particle image velocimetry. Journal of Wind Engineering
and Industrial Aerodynamics, 123, 56-68, 2013.
Lee H.M. and Wu. Y. A Tomo-PIV study of the effects of freestream turbulence on stall delay
of the blade of a horizontal-axis wind turbine, Wing Energy, in press.
Wu. Y. A study of energetic large-scale structures in turbulent boundary layer, Physics of
Fluids, 26, 045113, 2014.
279
SICASE-722
The Fracture Prediction of Electromagnetic Steel for Shorten Reheating
Process in Furnace
Yu-Hsi Huang*, Ching-Kong Chao, Chia-Hsuan Cheng, Zhen-Yi Fan
National Taiwan University of Science and Technology, Taiwan
Li-Wen Wu
China Steel Corporation, Taiwan
Abstract
Electromagnetic steel is heat to almost 1200 ˚C for soften slab before operated hot-rolling
process. To increase the production of ultra-thin electromagnetic steel sheet, the heating rates
need to increase. However, the improper increase in heating rate may induce fracture on
ultra-thin electromagnetic steel. Therefore, the proper heating rates are the important issue of
manufacturing ultra-thin electromagnetic steel. The effects of heating processes on the fracture
mechanism of ultra-thin electromagnetic steel slab are numerically studied in this study.
According to six material constants given by 100-900 ˚C, including on the Young‟s modulus,
Poisson‟s ratio, density, specific heat, thermal conductivity, and thermal expansion, the
investigation of the optimal heating rates is calculated from finite element numerical
calculation. The criterion of strain energy density is used to predict the fracture mechanism in
the paper. By means of determination of the strain energy density in the different heating rates,
the safety condition is design by the maximal strain energy density smaller than the critical
value in heating processes. In order to predict on the fracture of ultra-thin electromagnetic steel
during heating processes, the critical strain energy density is applied to determine the optimal
heating rates. The strain energy density are experimentally obtained from integral area of
stress-strain curves on various temperatures by means of ASTM E21. Hence, the results from
numerical calculation is concluded that the heating period could be shorten from 320 minutes
to 260 minutes by theory of critical strain energy density. Through this research, in addition to
be informed whether ultra-thin electromagnetic steel sheet by shortening the heating time to
improve productivity, reduce energy costs, the theoretical model also applies to other kinds of
steel for manufacturing various types of steel available and effective information.
Keyword: Electromagnetic Steel, Fracture, Optimal Heating Rates, Strain Energy Density,
Furnace
1. Introduction
280
Silicon steel (also called electromagnetic steel, ES), suggests that it is the amount of
silicon-containing steel, and has a special purpose excellent electromagnetic properties used on
electronic motors, power communications, instrumentation industries. In manufacturing
process, refining molten steel by continuous casting steel is formed, then uniform heat in
furnace (reheating furnace) before cutting steel. The silicon steel in furnace will heat up to a
target temperature by segments, respectively preheat zone, heating zone and soaking zones,
finally up to roughly 1120-1200 degrees Celsius. Silicon steel slab in the furnace using heat
energy from fuel combustion of heating spend energy consumption. If the process can shorten
the silicon steel slab in the furnace of time, not only can improve production of silicon steel,
but also can save the cost of manufacturing, to achieve less energy consumption.
In actual manufacturing process, the temperature distribution of silicon steel slab within the
furnace cannot be directly measured when reheating furnace operates and controls. The
temperature of the furnace can be given and used to predict the steel temperatures, and also
using the steel temperature out of furnace to determine on the temperature distribution by
experience, numerical analysis, fuzzy method for each furnace zone. Ko et al. [1] used of
material and energy conservation to obtain a dynamic mechanism model furnace, and identify
the discrete-time control model. Ditzhuijzen et al. [2] measured on each furnace zone
temperature as the predicted value, and had been predictive controller design.
Among the manufacturing process of electromagnetic steel, silicon steel slab in the furnace
heat before hot rolling. Because the heating rate of speed may cause thermal cracking damage,
the damage criterion using strain energy density (SED), is proposed by Chao et al. [3- 5], who
also used the SED theory on zirconium hydride dry place storing spent nuclear fuel drums
surface cracks and damage in order to predict crack growth trend. Sih and Macdonald [6]
proposed the principle of SED as a failure mechanism baseline forecast, which also referred to
the stress intensity factor and fracture toughness strength. Beltrami [9] proposed the SED when
the structure of the material beyond the critical SED, the material will have to break. SED is a
function of stress and strain and the critical SED can be determined using a tensile test by the
distribution area of the tensile stress-strain material. Hence, the critical SED predicted to
fracture can be applied to any material.
2. Research Methods
This study estimated the silicon steel slab heat temperature and time in the furnace, and the
fracture of silicon steel slab is predicted on thermal cracking damage by temperature difference
between surface and center temperature. A damage factors is proposed to design protecting
thermal rupture of steel slab during heating shorten the time. The theoretical model for silicon
steel can be also predicted on what the heating rates as ambient conditions by air cooling cause
fracture. In this study, the finite element method (FEM) is used to establish prediction model of
281
SED employed on thermal cracking damage accrued during the heating process for low and
high silicon steels. The FEM results is corresponding to the toughness in the temperature
gradient with the heating time. In order to design the best mode of heating temperature and
time, the fracture mechanism of silicon steel in the furnace is to understand the temperatures of
the surface and the center of difference whether steel slab causes impact on rupture of silicon
steel slab. The theoretical models can obtain the resulting optimization of process parameters to
establish the model so that all kinds of steel slabs make corresponding optimal design on
heating time. The following is the described on the theory of strain energy density theory.
Any piece of elements consists of a finite number of units and the strain energy density per unit
volume can be expressed as [6]:
0
ij
ij ij
dWd
dV
(1)
In which, ij and
ij are stress and strain tensors, respectively, and V is the volume and W is
the total energy stored in the element. In homogeneous isotropic elastic body and other terms,
the strain energy density can be expressed as:
2 2 21
2x y z x y y z z x
dW v
dV E E (2)
The Equation (2) can be separated as:
v d
dW dW dW
dV dV dV
(3)
The v
dW dV is representative to the volume change produced by the expansion of the
strain energy density (dilatational strain energy density); and the d
dW dV is representative
to the shape change of the resulting deformation strain energy density (distortional strain
energy density). Using this theoretical model studies on predicted silicon steel slab in reheating
furnace and the failure mechanism is provided.
3. Silicon steel Reheating Process in FEM Analysis
In this study, a finite element commercial software, Abaqus V6.13, is employed for silicon steel
reheating process to simulate numerical analysis. The reheating process time required 320
minutes in origin, the FEM analysis uses a larger time step in implicit integration method
282
(Implicit) to solve arithmetic. The actual dimensions of the silicon steel slab is length 8.5 m ×
width 1.25 m × thickness 0.25 m. Material properties of silicon steel slab is provided on the
density, Young's modulus, Poisson's ratio, specific heat, thermal conductivity and coefficient of
thermal expansion of different temperature characteristics. Due to the heating time for high
silicon steel is 320 minutes in reheating furnace the temperature boundary conditions is set up
on the high silicon steel slab with the initial temperature at 200 ° C to final temperature. The
low silicon steel slab is 200 minutes for actual process. The simply-supported boundary
conditions and the gravity of steel itself are considered to provide on the skid conditions. When
the shorten heating period is used on the upper and lower surfaces on silicon steel slab, the
optimal heating time is results in the loop verification.
Silicon steel slab model is set of grid, while the thickness direction of the silicon steel mesh
refinement embryo, mainly to increase the accuracy of the temperature variation in the
thickness direction, and the thickness direction of the grid convergence analysis carried out, as
shown in the grid 15-30 presents the results of convergence, convergence results of this study
have access to the thickness direction of 21 grid in analysis.
4. Silicon Steel Reheating Process in Failure Mechanism Analysis
4.1 Silicon Steel Reheating Process of Analysis Results
This study analyzed the model in temperature and SED resulting data are derived from the
model of the most critical locations on the longitudinal center point of thickness. First, the
process of reheating of high silicon steel is analyzed to the results indicating on that results of
temperature and SED are, as shown in Fig. 1(a), presented on central positions and the upper
and lower surfaces. Because of the mechanical properties of the steel, the resulting SED in the
upper and lower surfaces are shown on the maximum values, when heated to about 120
minutes and the temperature reaches about 715 degrees Celsius. However, when the
temperature rises to about 800 degrees Celsius, the SED began to decline, thus corresponding
to the characteristics of high silicon steel material factor because SED may be speculate the rise
and fall due to the phase transition. When the temperature reached about 1130 degrees Celsius,
also the highest temperature of the upper and lower surfaces, the SED continues to decrease
with the center and surfaces SED values gap narrowing.
Low silicon steel in actual reheating process is 200 minutes, close to reheating process of
carbon steel. In order to compare with the high silicon steel in 320 minutes, a low silicon steel
warming strategy 200 minutes, two different results of low silicon steel heating periods were
analyzed, as shown in Fig. 1(b). The SEDs of low silicon steel showed that the SED of 200
minutes heating period is significant increasing of more than 320 minutes. At about 900
degrees Celsius, the maximum SED is performed to the two heating periods for the same low
silicon steel.
283
The SED values in 200 minutes is compared to 320 minutes about four times the magnitude of
the difference. Moreover, the SED analysis of the characteristics shows that the high silicon
steel with is achieved to larger SED values than low silicon steel in the same 320-minute
heating period.
4.2 High and Low Critical Strain Energy Density of Silicon Steel
By Equation (1) of strain energy density, it shows the distribution of the area of stress and
strain of the available SED value and we want to predict the circumstances reheating process
silicon steel of the damage in this paper. Therefore, the critical SED value size silicon steel
must be determined on high temperature tensile test for high and low silicon steels. In this
study, in accordance with the temperature of the metal tensile test (ASTM E21) [8, 9], this
specification includes the definition of yield strength, tensile strength, elongation, and enhance
the metal temperature of the reduction of the area of the tensile test methods in ASTM E8. Figs.
2 were stretched 100 to 900 degrees Celsius of the level of silicon steel stress-strain
experiments graphics setting the test piece stretched to break. The tensile stress and strain
throughout the complete recording process, are draw in experiments.
284
When the SED is computing by Equation (2) obtaining the critical SED under various
temperatures, the critical SEDs of the high and low silicon steel slabs were 3.951 (MJ / m3) and
5.478 (MJ / m3) about 900 Celsius degree. This result can only predict the reheating process of
silicon steel during this damage does not occur. And between 100 and 500 degrees greater
critical strain energy density, so that the silicon steel is heated at a temperature zone we can use
more rapid temperature control strategy, has high strength and toughness heat crack and failure
conditions resistant.
The maximum SED obtained by FEM for the high and low silicon steels in shorten reheating
process is divided to the critical SED. The damage factor is proposed and shown in Fig. 3.
Various heating strategies and 280 minutes from 320 minutes every 10 minutes analogy of
different strategies is used to shorten the heating thermal stress analysis. The visible trend for
reheating process as time is shortened, the maximum SED values increased gradually. The ratio
is similar to the present study therefore concluded that high silicon steel reheating process can
be reduced to shorten the time to 240 minutes, reheating process is close to five times (22%) of
safe range. When reheating process shortened to 160 minutes to assume that the critical low
silicon steel of the SED, the SED exceeds the destruction of value 3.951 MJ/m3 which can be
determined reheating process shorten as 160 minutes for high silicon steel, are not successfully
reach the heating may cause steel slab destruction.
5. Conclusion
The proper heating rates are the important factors of manufacturing ultra-thin electromagnetic
steel sheet. According to the investigation of the optimal heating rates from numerical study, an
experimental setup will be designed and developed by means of the similarity theory. Hence, a
heating furnace system will be built to simulate the heat process of ultra-thin electromagnetic
steel sheet.
285
The effect of various heating rates on the fracture of ultra-thin electromagnetic steel sheet will
be verified by experimental method. However, this study forecasting model destruction
mechanism but ideal for ultra-thin electromagnetic steel sheet surface temperature under the
assumption that the establishment of optimum temperature control technology, but the furnace
device which is reachable target temperature is subject to a number of heat transfer parameters
impact. Therefore, the reliable heating processes of the heating furnaces will improve the
production of electromagnetic steel sheet and achieve energy saving requirements.
Acknowledgements
Thank the National Science Council (Ministry of Science, Project No.
NSC102-2622-E-006-032) and China Steel Corporation provide funding for research grants.
[1] H. S. Ko, J. S. Kim, and T. W. Yoon, Modeling and predictive control of a reheating
furnace, American Control Conference, 2000.
[2] V. Ditzhuijzen, D. Staalman, and A. Koom, Identification and model predictive control of
a slab reheating furnace, IEEE International Conference on Control Applications, 2002.
[3] C. K. Chao, J. Q. Chen, K. C. Yang, and C. C. Tseng, Creep crack growth on spent fuel
zircaloy cladding in interim storage, Theoretical and Applied Fracture Mechanics, 2007,
47(1), pp. 26–34.
[4] C. K. Chao, K. C. Yang, and C. C. Tseng, Rupture of spent fuel Zircaloy cladding in dry
storage due to delayed hydride cracking, Nuclear Engineering and Design, 2008, 238(1),
pp. 124–129.
[5] C. K. Chao, K. C. Yang, C. K. Chen, and C. C. Tseng, Effect of hydride embrittlement on
creep life of spent fuel cladding, Journal of the Chinese Institute of Engineers, 2011, 34(1),
pp. 107-121, .
[6] G. C. Sih , B. Macdonald ,Fracture Mechanics Applied To Engineering Problems-Strain
Energy Density Fracture Criterion, Engineering Fracture Mechanics, 1974, 6 .pp.361-386.
[7] E. Beltrami, AslleCondizioni di Resistenzadeicorpi Elastic, Rendiconto del RealeIstituto
Lombardo, Ser. II, Tomo XVIII, 1985.
[8] Instron ,2736 Series Manual Wedge Action Grips High Temperature (2736-015)
[9] Instron ,Guide to High Temperature Tensile Testing.
286
SICASE-820
Direct Current Compressor for Vehicle Air Conditioning System in
Non-Electric Vehicles
Amirah Haziqah Zulkifli, Henry Nasution*,
Afiq Aiman Dahlan and Azhar Abdul Aziz
Automotive Development Center, Faculty of Mechanical Engineering,
Universiti Teknologi Malaysia, 81310 Skudai, Johor, Malaysia
*Corresponding Author: [email protected]
ABSTRACT
Vehicle air-conditioning (AC) system is the highest auxiliary load in the non-electric vehicle.
Belt-driven compressor adds a small speed-dependent parasitic load even when not in use. This
study investigates the possibility of direct current compressor to retrofit into an existing vehicle
AC system in the non-electric vehicle. Electric-powered present offer potential of low energy
consumption than their mechanical counterparts: electrically operated compressor can reduce
engine idle time and offering near zero parasitic loads when not in use. The experiment is done
using a compact hatchback car on a roller dynamometer to emulate the level road condition. An
experimental period of one hour is made at temperature setting of 25°C at 30 km/h vehicle
speed. The results indicate that the direct current compressor offers fuel saving of 8.26% at 30
km/h vehicle speed with improves cabin temperature inside the vehicle for better thermal
comfort.
Keyword: Performance, Direct Current Compressor, Vehicle Air Conditioning, Fuel
Consumption, Energy Efficiency
1. Introduction
Vehicle air conditioning (AC) system is the highest auxiliary load in a vehicle after alternator
and power steering [1]. The usage of current belt-driven compressor can increase NOx
emissions as much as 15-100% [2]. Lambert and Jones in their study found that around 12-17%
increment of vehicle fuel consumption is for AC operation [3].
Furthermore, belt-driven compressor is connected directly to the engine crankshaft, causing the
compressor speed to be dependent on engine crankshaft speed. Increment of compressor speed
would increase the compressor power consumption and thus higher energy is needed to operate
the AC system [4-5]. To increase overall energy consumption, the compressor speed must be
able to be controlled independently from the engine speed. Accordingly, the use of electric
compressor is seen as a solution to the problem.
287
Direct current (DC) compressor for vehicle AC offers a benefit of decoupling the compressor
from the engine, decreasing parasitic load even when not in use. Decoupling the AC from the
engine can benefit to lower fuel consumption by downsizing the AC components as they no
longer require sizing for effective operation at low engine speed when idling [6].
In this manuscript, the feasibility study of DC compressor for vehicle AC system in the
non-electric vehicle is done. The DC is the best choice of electric compressor as it does not
need a power inverter that will be further down the energy-efficiency [4, 7]. Both the DC
compressor and belt-driven AC system performances are compared which both using the
original on/off controller inside the evaporator.
2. Coefficient of Performance
Coefficient of performance (COP) of a refrigeration cycle is defined as the ratio of the energy
removed at the evaporator to the energy supplied to the compressor. The general formula for
COP is:
com
e
W
Q
hh
hh
)(
)( COP
12
41 (1)
where h1, h2 (kJ/kg) are compressor inlet and outlet enthalpy respectively, h4 (kJ/kg) is
evaporator inlet enthalpy, Qe (kJ/kg) is refrigerating effect and Wcom is compression work.
3. Experimental Setup
Experiments are carried out on a Naza Suria, a compact hatchback vehicle with 1.2 cc gasoline
engine employing a DC compressor. Refrigerant HFC-R134a temperatures and pressures are
measured by a T-type thermocouples and pressure gauge, respectively. The temperature data
logger TC-08 from Pico Technology is used to log the temperature data into the computer.
Power input to the compressor is measured using a current transducer and logged into the
compressor. Specifications of the setup are given in Table 1.
Figure 1 shows the schematic diagram of the experimental setup. The DC compressor is
powered by the car battery which is charged by the alternator. The inverter acts as the speed
controller and placed under a front passenger‟s seat. Flow meter takes the refrigerant flow rate
so that COP can be acquired. Figure 2 shows the position of direct compressor inside engine
compartment. The DC compressor is placed behind a front bumper as there is no other place is
available because this is a compact car. Figure 3 shows the whole experimental setup of the
vehicle on the roller dynamometer.
288
Table 1. Specification for Experimental Setup
Equipment Specification
Compressor Type: 12V DC, Hermetic reciprocating
Displacement: 7.5cm3
Speed range 1800-3700 rpm
Pressure meter Range 0-100 psi low side and 0-300 psi
high side
Thermocouple Type: T
Range -200 to -350C
Output signal: mV
Accuracy: ±0.4% of full scale
PicoLog temperature
data logger
Voltage mode: ±60 mV
Accuracy: ±0.5C
Fuel digital data
logger
Ono Sokki DF2420
Accuracy: ±0.1mL
Figure 1. Schematic diagram of DC compressor
289
Figure 2. Compressor placed behind front bumper
Figure 3. Experimental setup of vehicle on roller dynamometer
4. Results and Discussion
Measured and calculated values were analyzed to define the performance of the DC
compressor. Performance comparisons were performed between belt-driven compressor and
DC compressor using evaporator on/off controller at vehicle speed of 30 km/h and 25C
temperature setting. Total duration for all tests was 5 hours, with a sampling frequency of 1
second. The ambient temperature is around 29-33C with relative humidity of 60-70%.
Comparison of the vehicle cabin temperature distribution is illustrated in Figure 4. The
refrigerator thermostat is set to provide 13C in the evaporator to have temperature inside
vehicle cabin of 25C. Figure 5 shows the evaporation temperature, respectively.
DC Compressor
290
Figure 4. Temperature distributions for one hour cooling time
Figure 5. Evaporation temperatures for one hour cooling time
Figure 6. COP of electric compressor and belt-driven compressor
291
Figure 7. Mass flow rate of electric compressor and belt-driven compressor
Figure 8. Cooling capacity of electric compressor and belt-driven compressor
Figure 9. Fuel consumption of vehicle
Figure 4-8 shows the temperature distribution, evaporation temperature, COP, mass flow rate
and cooling capacity for belt-driven compressor and various speed of direct compressor. It can
be seen that the temperature distribution of belt-driven compressor is lower by 9% because of
the higher cooling capacity of belt-driven compressor. The cooling capacity and mass flow rate
of the electric compressor increase as the compressor speed increase. Unfortunately, the COP
decrease in the compressor consumption is increasing with the increase of compressor speed.
292
Fuel consumption of compressor speed at various vehicle speeds is illustrated in Figure 9. The
usage of DC compressor has a big impact on vehicle fuel consumption. The usage of electric
compressor managed to get up to 19.08 km/l for 1800 rpm compressor speed, saving 8.26% of
fuel compared to the belt-driven compressor.
4. Conclusion
In this study, performance of the compressor operation during on/off mode of the DC
compressor and belt-driven compressor has been compared. The comparisons are made in
terms of cooling capacity, COP and fuel consumption.
The fuel consumption usage of DC compressor is seen as a solution to decrease fuel
consumption and increase the temperature control, although it is using on/off control at the
expense of lower cooling capacity. Lowering cooling load inside vehicle cabin is the solution
to the problem.
According to the analysis, the DC compressor operation provides higher COP and lower fuel
consumption at all vehicle speed and internal heat load. This is because the lower power
consumption of the compressor. It is also noted that the vehicle cabin temperature reached the
set value with DC compressor earlier than the belt-driven compressor at all vehicle speed.
The DC compressor offers the potential of the intelligent controller to be implemented so that
the thermal comfort of passengers can be increased under variety of environment condition. So,
it is suggested to substitute the current on/off controller with variable-speed controller so that it
can control the compressor speed according to the cooling load inside vehicle cabin.
ACKNOWLEDGEMENTS
The research work is supported partially by the Knowledge Transfer Program (KTP), Vote No.
R.J130000.7809.4L509 and the Automotive Development Centre (ADC), Universiti Teknologi
Malaysia (UTM). The assistance of the technicians involved, their guidance and assistance are
gratefully acknowledged.
REFERENCES
[1] Khayyam, H., Nahavandi, S., Hu, E., Kouzani, A., Chonka, A., Abawajy, J., Marano, V.
and Davis, S. 2011. Intelligent Energy Management Control of Vehicle Air Conditioning
via Look-Ahead System, Applied Thermal Engineering, 31(16), 3147–3160.
[2] Welstand, J. S., Haskew, H. H., Gunst, R. F. and Bevilacqua, O. M. 2003. Evaluation of the
Effects of Air Conditioning Operation and Associated Environmental Conditions on
Vehicle Emissions and Fuel Economy, SAE Technical Paper, No. 2003-01-2247.
[3] Lambert, M. A. and Jones, B. J. 2006. Automotive Adsorption Air Conditioner Powered by
293
Exhaust Heat. Part 1: Conceptual and Embodiment Design, Proceedings of the Institution
of Mechanical Engineers, Part D: Journal of Automobile Engineering, 220(7), 959-972.
[4] Nasution, H., Yamani, M. H. Z.and Sumeru, K. 2013. Performance Study of DC
Compressor for Automotive Air Conditioning System, Advanced Materials Research,
614-615, 674-677.
[5] Perang, M. R. M., Nasution, H., Latif, Z.A., Aziz, A. A. and Dahlan, A. A. 2013.
Experimental Study on the Replacement of HFC-R134a by Hydrocarbons in Automotive
Air Conditioning System, Applied Mechanics and Materials, 388, 111-115.
[6] Surampudi, B., Redfield, J., Montemayor, A., Gustavo R. and Gregory, O. 2006. Electric
Air Conditioning for Class 8 Tractors, SAE Technical Paper, No. 2006-01-01652006.
[7] Thomas, V. M., Meier, A. K., Gunda, S. G. and Wenzel, T. P. 2011. Cars are Buildings:
Building-Like Energy use in Automobiles, Transportation Research Part D: Transport and
Environment, 16(4), 341-345.
294
SICASE-819
In-Cabin Temperature Control and Multi-Circuit Bus Air-Conditioning
Systems
Henry Nasution*, Azhar Abdul Azis,
Amirah Haziqah Zulkifli and Afiq Aiman Dahlan
Automotive Development Center, Faculty of Mechanical Engineering,
Universiti Teknologi Malaysia, 81310 Skudai, Johor, Malaysia
*Corresponding Author: [email protected]
ABSTRACT
In-cabin temperature control and multi-circuit bus air-conditioning systems were developed.
The system work dependent on the cooling load either one or more circuits will automatically
be switched on. One circuit of the system can have been varying speed in order to fine-tune the
cooling response time as the cooling load varies inside bus cabin. Variable-speed compressor
provides potential on the load-matching capability, energy saving and thermal comfort of the
passengers inside bus cabin. The experiments were conducted on a test rig with temperature
setting of 22°C for the conditioned space to evaluate the energy usage of the system with
thermostat and fuzzy logic (FLC) controllers. The results indicate that the proposed technique
can save energy and improve indoor comfort significantly for multi-circuit bus air-conditioning
systems compared to the existing system. The experimental results also showed that more
energy savings can be obtained through the use of FLC when compared to the thermostat
controller.
Keyword: Performance, Bus Air Conditioning, Multi Circuit, Energy Saving, Fuzzy Logic
Control
1. Introduction
Bus air conditioning (AC) system plays an important role to achieve thermal comfort for all the
passengers inside the cabin. Its operation is high transient conditions as the cooling load varies
caused by several factors such as opening a door, changing of sun load through the side glass
windows and number of passengers on board. At the same time, the system must be compact
and work efficiently to provide thermal comfort at the lowest energy usage. The high-energy
requirement for climatic control gives potential reduction of energy consumption [1].
The existing bus AC use only one AC system which worked at full speed and controlled via a
thermostat controller to get the temperature needed. This control technique is not efficient in a
high transient condition, such as in building, bus and car, caused by a relatively large
compressor capacity compared to the smaller refrigeration system. The thermostat controller
295
will cause a very high inertia torque to the compressor when it is started suddenly from zero to
high speed, leading the system to behave in a prone manner and higher probability for the
compressor gets damaged quickly [2,3].
In this paper, a roof-top multi-circuit bus AC system has been proposed to overcome the
shortcomings of the existing system. This system can use more multiple AC circuits, which
shared the same evaporator surface or also knew as face-to face evaporator control. The
advantages of this system are its simple installation and maintenance at much lower energy
usage. This paper focuses on the development of the controllers for the multi-circuit bus AC
system to have better temperature control but with the lowest energy usage using fuzzy logic
control (FLC). The results will be compared to the existing system.
2. Principle of Multi-Circuit Bus Air-Conditioning
The arrangement of the current bus AC system components is shown in Figure 1. The
condenser, evaporator, receiver–drier and expansion valve is located on the roof of the bus
while the compressor is connected via belt-driven with the auxiliary engine. This system gives
full loads to one AC system, caused the temperature variation to be higher.
Figure 1. Conventional bus AC system
296
Figure 2 shows schematic diagram of the proposed multiple-circuit bus AC system. The
principle is to split the circuit to smaller unit and driven by a different compressor, which one
has the compressor capacity of half of the total system and the remainder for other compressor
capacities. This study will have the same compressor capacity for both systems as if one
becomes inoperative, other system will hold the load until repair is done.
Figure 2. Multi-circuit bus AC system
3. Performance and Energy Analysis
The coefficient of performance (COP) of AC system is the ratio of the energy removed at the
evaporator to the energy supplied to the compressor. The following is the general formula of
the COP:
com
e
W
Q
hh
hh
)(
)( COP
12
41 (1)
where h1, h2 (kJ/kg) are the enthalpy at the compressor inlet and outlet respectively, h4 (kJ/kg)
is enthalpy at the evaporator inlet, Qe (kJ/kg) is the refrigerating effect, and Wcom (kJ/kg) is the
compression work.
Calculation of energy saving is expressed in terms of saving in percentage unit, based on the
difference of energy consumption using thermostat controller and FLC. The equation is:
100)/(
)()/(
energyoffon
energyFLCenergyoffonSavingEnergy (2)
297
4. Controller Algorithm
FLC block diagram for controlling cabin temperature is shown in Figure 3. The block diagram
shows a two-input and one-output FLC where the input variables are temperature error (e) and
temperature rate-of-change-of-error (e). The output variable of the FLC is the frequency of
the supply current to the compressor electric motor (Z).
Figure 3. FLC algorithm block diagram
The three principal phases for FLC are fuzzification, inference mechanism and defuzzification.
Fuzzification process transforms the defined value into fuzzy value. Inference mechanism
determines the output value depending on the rules fixed to the actual process. Defuzzification,
on the other hand, transforms the fuzzy output into a defined value [4].
Table I showed the fuzzy associate map of the fuzzy controller method. The nine linguistic
variables marked with the following labels: hot (H), normal (N), cool (C), negative (NE),
normal (NO), positive (PO), slow (SL), normal (NM) and fast (FT), are used to characterize the
input and output variables. Triangular type (Figure 4) for the membership function is used
because it is the fastest [1]. (e) and (e) covered the range between -2°C and 2°C. Z is a range
between 0 to 5Vdc.
Table 1. Fuzzy associative map [1]
Z e
H N C
e
NE SL SL SL
NO SL SL SL
PO FT NM SL
(a) Input (b) Output
Figure 4. Triangular membership functions for inputs and output [1]
Defuzzification method adopted is by determination of the center of gravity or centroid.
Microsoft Visual Basic is used to build the control algorithm. This algorithm provides the
298
output variable of voltage signal to continuously control the compressor speed through the
inverter.
5. Experimental Setup and Testing Procedure
The schematic diagram of the experimental setup is shown in Figure 5. The system is made up
of the original components from bus AC system and arranged to emulate the actual bus.
Electric heater is laid in the main air duct between main cabin and evaporators to simulate the
cabin cooling load from the passengers. Latent heat is then achieved through the mixing of
outside air and cooled air from the evaporator.
Figure 5. Schematic diagram of the experimental rig
The parameter needed that is temperature, pressure and mass flow rate is measured at locations
indicated in Figure 5. The temperatures of the refrigerant into the system and the air
temperatures inside the cabin are detected by thermocouples. The thermocouples to detect the
temperature of refrigerant inside the system were inserted into the copper tubes. There were
five thermocouples inside the simulated passenger cabin, and nine pressure gauges to measure
the refrigerant circuit. The mass flow rate of the refrigerant was measured using flow meter for
R134a.
A control system is installed to control the compressor speed, consisting of a thermocouple
inside cabin, thermostat and FLC subroutine installed on a computer, an inverter and an electric
motor. The thermocouples detecting the temperature inside simulated cabin and sent to the
controller and computer. The output function of the error between the value of the monitored
temperature and the required set-point temperature is then emitted by the computer to the
inverter, which varied the electrical frequency supplied to the motor such that it is linearly
299
proportional to the control signal. The experiments were conducted with: variable speed,
thermostat and FLC at 22C setup temperature.
6. Result and Discussion
Steady-state cabin temperature and energy consumption at various frequencies is shown in
Figure 6. The test is done in a one-hour test period. The results indicate that the energy
consumption of the system is dependent on motor frequency. As the frequency increases, the
energy consumption increases while the temperature inside the cabin is lower.
Figure 6. Steady state cabin temperature, energy consumption, and COP for variable speed
compressor
Figure 7. Temperature responses for thermostat and FLC
Figure 7 shows the temperature response for all controllers as a function of a time interval. The
thermostat controller turns on the compressor when it reaches upper limit and turns off the
compressor when it reaches lower limit setup temperature. As the compressor turned off, the
cooling load increases and the time taken into the compressor to turn on is fast, allowing cabin
temperature to be attained at temperature setup. The FLC, on the other hand, controlled the
compressor motor speed to minimize the error between the setup temperature and current cabin
temperature. If the error is big, the compressor will be set at the highest speed to cool down the
y = 38.704x-0.171 R² = 0.9386
y = 0.0191x1.4793 R² = 0.9997
y = 9.8124x-0.181 R² = 0.796
Ener
gy (
kWh
), C
OP
Cab
in T
emp
erat
ure
(oC
)
Frequency (Hz)
Cabin Temperature (C) Energy (kWh) COP
Te
mp
era
ture
(oC
)
Time (minute)
Setpoint Thermostat FLC
300
cabin faster. It can be seen from the figure that the temperature for the FLC is nearer to the
temperature setup compared to the thermostat controller throughout the one-hour test period.
Figure 8 shows the energy consumption of the system with thermostat and FLC. The energy
consumption was taken from the beginning of the start of the motor to the end of the one-hour
period time. The energy saving calculated is expressed in terms of percentage saving based on
the thermostat controller. The FLC shows the most energy saving of 61.66% compared to the
thermostat controller.
Figure 8. Energy consumption for thermostat and FLC
7. Conclusion
The results of the bus multi-circuit AC system with variable speed, thermostat and FLC
controller have been obtained. The results indicate that the speed of compressor can be changed
by varying the frequency for the inverter. The power consumption and cabin temperature is
directly dependent upon the frequency for the inverter. As the frequency increase, steady state
of cabin temperature will decrease while the energy consumption increases.
This study indicates that the cabin temperature can be controlled by the inverter driven
compressor to obtained better temperature control and energy saving through the usage of the
bus multi-circuit AC system. The study also concluded that FLC is a better energy-saving
control for the proposed system.
ACKNOWLEDGEMENTS
The research work is supported partially by the Knowledge Transfer Program (KTP), Vote No.
R.J130000.7809.4L509 and the Automotive Development Centre (ADC), Universiti Teknologi
Malaysia (UTM). The assistance of the technicians involved, their guidance and assistance are
gratefully acknowledged.
數列1,
Thermostat, 4.51
數列1, FLC,
1.73
Ener
gy C
on
sum
pti
on
(kW
h)
Controllers Mode
301
REFERENCES
[1] Nasution, H., Energy analysis of an air conditioning system using PID and fuzzy logic
controllers, PhD Thesis, Universiti Teknologi Malaysia, Malaysia, 2006.
[2] Mansour, M. K. M., Development of roof-top bus air-conditioning system that responds to
cooling load variations in tropical countries, PhD Thesis, Universiti Teknologi Malaysia,
Malaysia, 2007.
[3] Kaynakli O. and Horuz, I. 2003.An Experimental Analysis of Automotive Air Conditioning
System, Int Comm Heat and Mass Trans, 30, 273-284.
[4] Aprea, C., Mastrullo, R. and Renno, C. 2004. Fuzzy Control of the Compressor Speed in a
Refrigeration Plant, International Journal of Refrigeration, 27 (6), 639-648.
302
Poster Presentation
Poster Session (1)
Biomedical Engineering / Biological Engineering / Natural Science
2015/6/27 Saturday 13:30-14:30 4F Pre-Function Area
SICASE-745
Swelling Characterization of Photo-Cross-Linked Gelatin Methacrylate Spherical
Microgels for Biomedical Applications
Jinmu Jung︱Chonbuk National University
Hyojae Kim︱Chonbuk National University
Kyoungin Kang︱Chonbuk National University
Jonghyun Oh︱Chonbuk National University
SICASE-797
Effect of Titanium Content on Structural and Sensing Properties of TbTixOy Sensing
Films for Electrolyte-Insulator-Semiconductor Sensors
Tung-Ming Pan︱Chang Gung University
Bo-Rong Peng︱Chang Gung University
Chih-Wei Wang︱Chang Gung University
Wu-Shiung Feng︱Chang Gung University
Lifeng Chi︱Westfälische Wilhelms-Universität
SICBENS-1287
Development and Evaluation of ELISA for Diagnosis of Cystic Echinococcosis
Aitbay Bulashev︱Seifullin Kazakh Agrotechnical University
Orken Akibekov︱Seifullin Kazakh Agrotechnical University
Gulmira Abulgazimova︱Seifullin Kazakh Agrotechnical University
303
SICBENS-1306
Application of Hydroxynitrile Lyase from Prunus Amygdalus Turcomanica in Kinetic
Resolution of Racemic Mandelonitrile
Dilek Alagöz︱University of Ç ukurova
Deniz Yildirim︱University of Ç ukurova
Seyhan Tükel︱University of Ç ukurova
SICBENS-1330
Immature Seed as a Novel Explant for Indirect Somatic Embryogenesis in a Woody Tree;
Dalbergia Sissoo Roxb
Hafiz Mamoon Rehman︱Chonnam National University
SICBENS-1334
Cross-Linked Enzyme Aggregates of Hydroxynitrile Lyase from Wild Appricot
Deniz Yildirim︱University of Cukurova
Dilek Alagöz︱University of Cukurova
Seyhan Tükel︱University of Cukurova
SICBENS-1333
A Method to Prepare Large Amount of Sg-6 Saponins from Wild Soybean Seeds
Mei Itabashi︱Chonnam National University
Ayaka Kurosaka︱Iwate University
Panneerselvam Krishnamurthy︱National Academy of Agricultural Sciences
Chigen Tsukamoto︱Iwate University
Tai-Sun Shin︱Chonnam National University
Seung Hwan Yang︱Myongji University
Gyuhwa Chung︱Chonnam National University
SICBENS-1336
The Cases Study and Species Selection(Butterfly, Plant) for the Butterfly Garden
Landscaping
Jinkwan Son︱National Academy of Agricultural Science, RDA
Siyoung Lee︱National Academy of Agricultural Science, RDA
Jongku Kim︱National Academy of Agricultural Science
Minjae Kong︱Relationship of Dankook Univ. and Rural Development Administration
Donghyeon Kang︱National Academy of Agricultural Science
304
SICBENS-1303
Response of Red Blood Cell ATPase Activity in Freshwater Fish(Oreochromis Niloticus)
Exposed to Copper in Differing Ca2+ Levels
Gülüzar Atli︱Cukurova University
Esin G. Kanak︱Cukurova University
Mustafa Canli︱Cukurova University
305
SICASE-745
Swelling Characterization of Photo-Cross-Linked Gelatin Methacrylate
Spherical Microgels for Biomedical Applications
Jinmu Jung, Hyojae Kim, Kyoungin Kang, Jonghyun Oh*
Division of Mechanical Design Engineering, Chonbuk National University, Korea *[email protected]
Abstract
This study reported the swelling characteristics of photo-cross-linked gelatin methacrylate
(GelMa) spherical microgels for biomedical applications. Spherical microgels were generated
in a microfluidic system consisting of a co-axial flow-focusing device and an ultraviolet (UV)
irradiation apparatus. At a low flow rate ratio (< 0.14), the 9 wt.% GelMa spherical microgels
were smaller than the 6 wt.% ones. In contrast, at a high flow rate ratio, the results were
reversed. The increased GelMa concentration improved the mechanical properties and
increased the swelling ratios. The possibility of bioencapsulation was demonstrated, with good
viability of 3T3 cells encapsulated in the spherical microgels.
Keyword: Microfluidic, hydrogel, emulsion, spherical microgel, flow focusing
1. Introduction
Polymer microspheres have been employed as a flexible platform for biological or medical
applications such as cell delivery, drug delivery, drug release, and diagnostics [1–3]. Therefore,
in this study, a flow-focusing, glass capillary-based microfluidic device was employed to
generate microdroplets made of GelMa hydrogel. Inexpensive and cell-responsive GelMa has
methacrylate groups along the gelatin backbone, which allows microdroplets to be cross-linked
under filtered ultraviolet (UV) irradiation. The proposed microfluidic system facilitated both
the generation and the polymerization of microdroplets without an additional process. GelMa
microdroplets were controlled by size with flow rates of between 50 and 300 μl/h. The swelling
behavior of cross-linked GelMa microspheres was studied by observing microspheres in oil
and phosphate-buffered saline (PBS). The possibility of bioencapsulation was investigated
through 3T3 cell encapsulation in a polymer microsphere.
2. Experiments
Figure 1 shows a schematic illustration of the co-axial flow-focusing device used to generate
droplets. The microfluidic device using circular glass capillaries could provide several benefits
for forming microdroplets, such as high chemical resistance, precise control of surface
wettability, and true three-dimensional co-axial flow-focusing geometry. The GelMa
prepolymer solution was defined as the dispersed phase, while the oil phase was the continuous
306
phase. As two immiscible fluids (GelMa prepolymer solution and oil) were brought into
contact at the co-axial channel, the dispersed phase (GelMa prepolymer solution) was
compressed and broken up into monodisperse, spherical droplets.
A suitable photo-polymerization process for forming a GelMa spherical microgel was
conducted using UV irradiation. After exiting from the outlet of the microfluidic flow-focused
device, the aqueous droplets formed with the GelMa prepolymer solution flowing out via a
Tygon tubing (0.3 mm ID and 0.25 mm thick, Saint-Gobain Performance Plastics Co., Akron,
OH, USA). In the presence of a water-soluble photo-initiator (Irgacure 2959), the droplets
passing through the tubing were exposed to filtered UV light (320– 500 nm) for 5 min
(OmniCure S2000, Lumen Dynamics, Mississauga, ON, Canada).
As shown Figure 2, greater swelling ratios were observed with the greater GelMa concentration.
As shown in Figure 3, as a result of the overall swelling in PBS for 24 h (from 0 to 24 h), the
mean swelling ratio of 6 wt.% GelMa spherical microgels was 94.1 ± 27.9%, while the 9 wt.%
GelMa spherical microgels showed a greater mean swelling ratio of 182.0 ± 22.4%.
The stiffness of the GelMa spherical microgels was investigated using compressive strength
tests. In Figure 3, compressive stress-strain curves were obtained for two different GelMa
concentrations (6 and 9 wt.%). From the stress-strain curves, the elastic modulus was
calculated as 1.35 ± 0.20 and 1.94 ± 0.31 kPa for the 6 and 9 wt.% of GelMa hydrogels,
respectively (p < 0.01).
In the NIH/3T3 cell viability test results as shown in Figure 4, they demonstrated that the
microdroplet generation and polymerization processes using GelMa were non-cytotoxic and
biocompatible in the microfluidic system.
Acknowledgement
This research was supported by Basic Science Research Program through the National
Research Foundation of Korea (NRF), funded by the Ministry of Education
(NRF-2013R1A1A2058875 and 2014R1A1A1006388).
References
[1] Mathiowitz E, Jacob JS, Jong YS, et al. Biologically erodible microspheres as potential oral
drug delivery systems, Nature, Vol. 386, 1997, pp.410–414.
[2] Radovic M, Vranjes-Duric S, Nikolic N, et al. Development and evaluation of 90Y-labeled
albumin microspheres loaded with magnetite nanoparticles for possible applications in
cancer therapy, Journal of Materials Chemistry, Vol.22, 2012, pp.24017–24025.
[3] Freiberg S, Zhu XX. Polymer microspheres for controlled drug release, International
307
Journal of Phamaceutics, Vol.282, 2004, pp.1–18.
308
309
SICASE-797
Effect of Titanium Content on Structural and Sensing Properties of
TbTixOy Sensing Films for Electrolyte-Insulator-Semiconductor Sensors
Tung-Ming Pan*, Bo-Jung Peng, Chih-Wei Wang, Wu-Shiung Feng
Chang Gung University, Taiwan *[email protected]
Lifeng Chi
Westfälische Wilhelm, Universität, Germany
Abstract
This paper describes the impact of titanium content on the structural properties and sensing
characteristics of TbTixOy sensing films deposited on Si substrates through reactive
co-sputtering. X-ray diffraction and Auger electron spectroscopy were used to study the
structural and chemical features of these films as functions of the growth conditions (Ti plasma
power from 80 to 120 W). The observed structural properties were then correlated with the
resulting pH sensing performances. The TbTixOy electrolyte-insulator-semiconductor device
prepared under a Ti plasma condition of 100 W exhibited the best sensing characteristics,
including pH sensitivity, hysteresis and drift. We attribute this behavior to the optimal titanium
content and annealing temperature in this oxide film forming a well-crystallized Tb2Ti2O7
structure and a smoother surface.
Keyword: electrolyte-insulator-semiconductor (EIS), plasma power, TbTixOy, pH sensitivity.
1. Introduction
An ion-sensitive field-effect-transistor (ISFET) device has attracted significant attention in the
field of chemical sensors because it offers advantages of very small size, rapid response time, high
input impedance, and low output impedance [1]. Besides, this device is compatible with a
standard complementary metal-oxide-semiconductor (CMOS) process [2]. The concept of ISFET
device is based on the theory of metal-oxide-semiconductor field-effect transistor (MOSFET),
where metal gate is substituted by the electrolyte and a reference electrode. To realize an ISFET,
the investigation of the insulator material is a very important part of the process, because it is
directly exposed to the buffer solution. As a result, the insulator film is responsible for the
detection of the specimens because a change in the pH concentration induces the variations in the
electrolyte–insulator surface potential, which gives rise to the alterations in the insulator–
semiconductor electric field interface, channel conductance and current modulation. To promote
the sensitivity of an ISFET device, the nanoscale ISFETs are developed [3-4]. However, the
traditional SiO2 gate oxide in most Si-based MOSFET devices has been scaled close to its
310
physical thickness limit. To reduce the gate leakage current, high dielectric-constant (high-κ)
materials are being considered as a possible replacement of SiO2 as the gate insulator to provide a
physically thicker film for the same electrically equivalent oxide thickness [5-6].
In recent years, various high-κ dielectric materials (e.g., HfO2, ZrO2, Ta2O5, TiO2, Gd2O3) [7-11]
have been used as pH-sensitive membranes in electrolyte-insulator-semiconductor (EIS) devices
due to their good sensing performance. Desirable high-κ dielectric materials must have good
thermodynamic stability in direct contact with silicon. Some articles, however, have reported the
formation of an interfacial layer between the high-κ dielectric and Si [5-6]. Rare-earth (RE) oxide
thin films are one of the promising candidates to substitute the conventional SiO2 films in
nanoscale CMOS device applications due to their high-κ values, high resistivities, and large
bandgap energies [12-13]. Among them, Tb2O3 (terbium oxide) film has received a lot of
attraction as a promising high-κ gate dielectric due to its potential properties, such as a high
conduction band offset, large band gap, and good thermal stability [14-15]. Although RE oxide
film has been proposed to be one of the promising gate insulator materials for future nanometer
CMOS technology, device instabilities associated with the moisture absorption due to oxygen
vacancies or defects in the film are still the major problems [16]. To eliminate the problem, the
incorporation of Ti or TiO2 into the RE dielectric films can result in improved physical and
electrical properties because it reduces the reaction of the dielectric film with water [17].
Nevertheless, to the best of our knowledge, the properties of TiOx incorporated Tb2O3 (TbTixOy)
and their performance in resulting EIS-based pH sensors have yet to be reported.
2. Experimental
The EIS structures featuring a TbTixOy sensing membrane were fabricated on 4-in. p-type Si
(100) wafers. Before the deposition of TbTixOy films, the wafers were cleaned using a standard
RCA process and then treated with 1% HF to remove the native oxide. A TbTixOy film (~70
nm thick) was deposited on the Si substrate through cosputtering from a terbium and a titanium
target in diluted O2 environment. The terbium rf-plasma power was 100 W, while the same for
titanium was varied from 80 to 120 W. Subsequently, samples were annealed in O2 ambient for
30 s through rapid thermal annealing (RTA) at 900 °C to form Tb2Ti2O7. A 400-nm-thick Al
film was deposited on Si for good electrical backside contact. The dimensions of the sensing
film were defined through a standard photolithography process using SU8-2005 (MicroChem,
U. S.) as a photoresist. EIS devices were then assembled on a printed circuit board using a
conductive Ag gel. A handmade epoxy package was employed to encapsulate the EIS sensor
and the Cu line.
The film structure of the TbTixOy films prepared under different Ti rf-plasma power conditions
was investigated through X-ray diffraction (XRD) analysis. The compositional depth profiles
of TbTixOy films were measured using and Auger electron spectroscopy (AES).
311
The pH sensitivity, hysteresis voltage, and drift rate of the TbTixOy sensing films were
determined by capacitance–voltage (C–V) curves of EIS devices. The C–V curves for different
pH buffer solutions (Merck Inc.) were measured through Ag/AgCl reference electrode
(commercial liquid-junction electrode) using a Hewlett-Packard (HP) 4284A precision LCR
meter.
3. Results and Discussion
We used XRD to explore the crystalline structures of the TbTixOy films prepared under
different Ti power conditions (Fig. 1). The significant difference in sharpness, intensity, and
orientation of the diffraction peak was observed, implying that the crystal growth and
crystallinity of the TbTixOy films are strongly affected by the content of titanium in the film.
The crystal structure of Tb2Ti2O7 film is face-centered cubic. In the 80 W power condition,
only one Tb2O3 (400) peak was found in the XRD pattern, suggesting an amorphous TbTixOy
structure with low Ti content. In particular, the film prepared under the 100 W condition
exhibited a stronger Tb2Ti2O7 2O3 (400) peak, indicating that
the Tb2Ti2O7 film had a preferred orientation along the (222) direction. We attribute this
behavior to the reaction of the Ti atom with Tb2O3 to form a better stoichiometry Tb2Ti2O7
structure. Furthermore, the intensity of the Tb2Ti2O7 (222) peak became low for the 120 W
312
condition. Therefore, the content of titanium in the Tb2O3 film was an important factor for the
growth of homogeneous crystalline Tb2Ti2O7 films.
Fig. 2 displays the AES depth profiles of Tb, Ti, and O elements in the TbTixOy films for
different Ti power conditions. Under the 80 W condition, the Tb, Ti, and O concentrations
versus thin film depth stayed constant in the bulk with an O content of roughly at 30 at.% and a
Tb:Ti ratio of roughly 3:1, as shown in Fig. 2(a). It is clear that the content of Ti increases from
~15 to ~30 at.% with increasing the Ti plasma power from 80 to 120 W. In the AES profile, the
O content in the film prepared at the 100 W condition is almost the same as that of the film at
the 80 W condition, while the Tb:Ti ratio of the 100 W condition is approximately 2:1.
Furthermore, the Tb:Ti ratio of the sample prepared under the 120 W condition became 1:1,
whereas the O content of this condition increased to about 40 at.%.
The flatband voltage (VFB) of an EIS device is the most important measurable parameter; it is
defined in terms of the following equation [18]:
ox
effSisol
refFBC
Q
qEV
0
(1)
where Eref is the potential of the reference electrode (Ag/AgCl here considered), ψ0 is the
surface potential, χsol
is the electrolyte-insulator surface dipole potential, ΦSi is the silicon
workfunction, Qeff is the effective charge induced in the semiconductor by the various types of
charges that may be present in the oxide, and Cox is the gate oxide capacitance per unit area. In
this equation, all terms are constant except for ψ0, which can influence the dissociation of the
insulator surface groups, resulting in the EIS device sensitive to the electrolyte pH. The
relationship between the surface potential and the pH of the tested solution can be described
using the combination of the site-binding theory and Gouy–Chapman–Stern model [18-19],
which include properties of the gate oxide and buffer solution. The pH sensitivity of the gate
oxide surface is given as
313
with,30320
q
Tk.
pH
B
s
int
difB
q
TCk.
2
30321
1
(2)
where is a dimensionless sensitivity parameter, which varies between 0 and 1, kB is
Boltzmann‟s constant, T is the absolute temperature, q is the electron charge, Cdif is the
differential capacitance, and int is the intrinsic buffer capacity. The is related to the intrinsic
buffer capacitance of the gate oxide and the ionic concentration of the buffer solution. These
two factors are determined by the sensing material, process, and electrolyte tested.
314
To evaluate the sensing performance of the TbTixOy EIS device, we recorded a set of C–V
curves at pH ranging from pH 2 to pH 12. Figs. 3(a)-(c) show the pH-dependence of one group
of C–V curves for the TbTixOy EIS devices for three Ti plasma power conditions. These
normalized C–V curves were shifted as a result of the change in the number of hydrogen ion on
the sensing membrane. The TbTixOy EIS sensor exhibited distorted C–V curves, as has been
reported in the literature [20], presumably because of the existence of interfacial traps in the
oxide. Fig. 4 demonstrates the pH-dependence of the reference voltage (the voltage required to
achieve a normalized capacitance of 0.5) of the TbTixOy EIS devices fabricated under three
plasma power conditions. The TbTixOy EIS device prepared at the 100 W condition exhibited a
larger sensitivity of 69.31 mV/pH than those of the other conditions. This result could be
attributed to the optimal titanium content in the film forming a well-crystallized Tb2Ti2O7
structure, as evidenced from the results in Fig. 1.
Apart from the sensitivity detection, hysteresis and drift characteristics are two other critical
performance features affecting the success of an EIS sensor. Hysteresis phenomena can be
described through a surface-site model by assuming a portion of the sites respond slowly to pH
changes because of imperfections in the sensing film, resulting in the formation of porous
structures [9, 21]. The interior sites of these porous defects can react with ions present in the
tested solution, thus causing a hysteresis response. Another possible cause is the interaction of
the ions in the tested solution with the responding sites along the boundaries of grains on the
sensing film [9, 21]. We subjected the TbTixOy EIS sensors to pH loop of 747107
over a period of 1500 s. The hysteresis voltage is defined herein as the difference between the
initial and terminal reference voltages measured in this cycle. Fig. 5(a) indicates that the
TbTixOy EIS device prepared under the 100 W condition had a small hysteresis voltage of 3
mV in the pH loop of 747107. In contrast, the TbTixOy EIS device fabricated at the
120 W condition exhibited a hysteresis voltage of ~6 mV higher compared with other
conditions. This could be due to the high titanium content in the TbTixOy film producing a
large number of crystal defects.
315
The drift characteristics of an EIS sensor can be explained by a hopping and/or trap-limited
transport mechanism [22-23] to affect the rate of hydration of the sensing film. Such an EIS pH
sensor is related to the penetration of ions from the electrolytes into the sensing film [24].
Charged electrolytes can penetrate the hydrated layer of a sensing film to cause a decrease in
the effective gate dielectric thickness [9]. The change in gate capacitance induces nonideal
effects during EIS pH sensing. On the basis of the model proposed by Jamasb et al. [25], the
REF) can be expressed as:
t
nfni
nfni
neffinvscREF etQQQ)t(V 1 (3)
where Qsc and Qinv stand for the space charge and inversion charge, respectively. ni and nf
represent the dielectric constant of the original sensing layer and the hydrated layer,
respectively. tn
associated with structural relaxation. represents the dispersion parameter whose value lies
0< <1. From equation (3), drift is related to the hydrated layer thickness of sensing film. Fig.
5(b) displays the drift characteristics of the TbTixOy EIS sensors prepared under three Ti
plasma conditions and measured in solutions of pH 7 for 12 h. The drift coefficient was
evaluated from the slope of the drift characteristics. The change in the reference voltage can be
described as VREF=VREF(t)–VREF(0). In general, a drift behavior is regarded as a superposition
effect of a chemical and an electrical change at the surface of the gate oxide [21]. The chemical
change indicates the hydration dispersive transport whose degree was determined by both the
hydration species and the material properties of the sensing film. The TbTixOy EIS device
fabricated at 100 W condition exhibited the best long-term stability (slope=0.118 mV/h); in
contrast, the EIS capacitor under 80 W condition featured a serious drift of 0.327 mV/h. The
lower drift effect of TbTixOy EIS device fabricated at the 100 W can be manifested as the lower
capacitance of hydrated layer due to the optimal titanium content in the TbTixOy film reducing
the crystal defects. Consequently, the overall capacitance of the combination of the hydrated
and unhydrated sensing layer was close to the capacitance of the original layer. The higher drift
rate of EIS device processed under 80 W condition essentially is due to the higher capacitance
value of the hydrated layer.
4. Conclusion
We fabricated high-κ TbTixOy sensing films on Si substrates through reactive rf sputtering under
three Ti plasma power conditions. The optimal condition was a Ti plasma power of 100 W; the
resulting sensing film exhibited the well-crystallized Tb2Ti2O7 structure (from XRD) and the
Tb:Ti ratio of about 2:1 (from AES) as a result of the optimal titanium content present in the
TbTixOy film.
316
The EIS device using the high-κ TbTixOy sensing film exhibited a high pH sensitivity of 69.31
mV/pH, a low hysteresis voltage of 3 mV, and a small drift rate of 0.118 mV/h. The high-
TbTixOy film is a promising sensing membrane for use in future biomedical engineering
applications.
Acknowledgments
We thank the Chang Gung Memorial Hospital and the National Science Council of the Republic
of China, Taiwan, for financially supporting this research under contracts CMRPD2C0031 and
NSC 98-2221-E-182-056-MY3, respectively.
5. References
P. Bergveld, Sens. Actuators B, vol. 88, p. 20, 2003.
T. Matsuo and M. Esashi, Sens. Actuators, vol. 1, p. 77, 1981.
J. Kong, N. R. Franklin, C. Zhou, M. G. Chapline, S. Peng, K. Cho, and H. Dai, Science, vol.
287, p.622, 2000.
Y. Cui, Q. Wei, H. Park, and C. M. Lieber, Science, vol. 293, p. 1289, 2001.
G. D. Wilk, R. M. Wallace, and J. M. Anthony, J. Appl. Phys., vol. 89, p. 5243, 2001.
H. R. Huff, and D. C. Gilmer, High Dielectric Constant Materials: VLSI MOSFET
Applications, Springer: Berlin, 2005.
V. Jankovic, and J. P. Chang, J. Electrochem. Soc., vol. 158, p. P115, 2011.
D. Sobczynska and W. Torbicz, Sens. Actuators, vol. 6, p. 93, 1984.
L. Bousse, S. Mostarshed, B. van der Schoot,and N. F. de Rooij, Sens. Actuators B, vol. 17, p.
157, 1994.
S. P. Pogorelova, A. B. Kharitonov, I. Willner, C. N. Sukenik, H. Pizem, and T. Bayer, Anal.
Chim. Acta, vol. 504, p. 113, 2004.
J. L. Her, M. H. Wu, Y. B. Peng, T. M. Pan, W. H. Weng, S. T. Pang, and L. Chi, Int. J.
Electrochem. Sci., vol. 8, p. 606, 2013.
J. Paivasaari, M. Putkonen, and L. Niinisto, Thin Solid Films, vol. 472, p. 275, 2005.
O. Engstrom, B. Raeissi, S. Hall, O. Buiu, M. C. Lemme, H. D. B. Gottlob, P. K. Hurley, and
K. Cherkaoui, Solid-State Electron., vol. 51, p. 622, 2007.
L. Petit, A. Svane, Z. Szotek, and W. M. Temmerman, Phys. Rev. B, vol. 72, p. 205118, 2005.
M. Fanciulli and G. Scarel, Rare Earth Oxide Thin Film: Growth, Characterization, and
Applications, Springer: Berlin, 2007.
Y. Zhao, K. Kita, K. Kyuno, and A. Toriumi, Appl. Phys. Lett., vol. 89, p. 252905, 2006.
R. B. van Dover, Appl. Phys. Lett., vol. 74, p. 3041, 1999.
P. Bergveld and A. Sibbald, Comprehensive Analytical Chemistry. v. 23. Analytical and
Biomedical Applications of Ion-Selective Field-Effect Transistors, Elsevier Science
Publishing Company Inc: New York, 1988.
R. E. G. van Hal, J. C. T. Eijkel, and P. Bergveld, Sens. Actuators B, vol. 24-25, p. 201,1995.
317
T. M. Pan and C. W. Lin, J. Phys. Chem. C, vol. 114, p. 17914, 2010.
L. Bousse, H. H. van den Vlekkert, and N. F. de Rooij, Sens. Actuators B, vol. 2, p. 103, 1990.
H. Scher and E.W. Montroll, Phys. Rev. B, vol. 12, p. 2455, 1975.
G. Pfister and H. Scher, Phys. Rev. B, p. 15, vol. 2062, 1977.
L. Bousse, N. F. de Rooij, and P. Bergveld, IEEE Trans. Electron Devices, vol. 30, p. 1263,
1983.
S. Jamasb, S. D. Collins, and R. L. Smith, IEEE Trans. Electron Devices, vol. 45, p.1239,
1998.
318
SICBENS-1287
Development and Evaluation of ELISA for Diagnosis of Cystic
Echinococcosis
Aitbay Bulashev*, Orken Akibekov, Gulmira Abulgazimova
Seifullin Kazakh Agrotechnical University *[email protected]
Echinococcus granulosus is the causative agent of cystic hydatid disease in domestic
herbivorous animals and man. Immunological tests used in the diagnosis of echinococcosis are
based on phenomena of precipitation and agglutination. In recent years different options of
highly sensitive methods, such as Enzyme-linked immunosorbent assay (ELISA) and Lateral
Flow Test are being introduced into practice. However, accurate diagnosis of the infection
requires highly specific antigens to be used in immunoassays.
The aim of our research was to develop and evaluate ELISA-test based on monoclonal
antibodies (MAb) for the identification of animals infected with cystic echinococcosis that
would allows to monitor this invasion in the endemic areas and detect infected herds prior to
slaughter.
Antigens prepared from hydatid cyst envelope (HE), hydatid cyst fluid (HF) and
excretory-secretory products (ES) of E. granulosus were tested in order to select as an
immunogen for hybridoma technique. SDS-PAGE electrophoresis of HE, HF and ES showed
the presence of 9, 11 and 5 protein bands with the molecular weights in the range of 8-205 kD,
12-38 kD and 19-63 kD, respectively. According to the results of immunoblotting protein
fractions of ES had the most expressed antigenic properties. In this regard, it was selected for
the immunization of Balb/c mice to get immune lymphocytes for hybridization with myeloma
cells.
Mab 3B1G10 specific to the protein of E. granulosus ES with molecular weight of 21 kDa
were selected for further research. Antobodies‟ titers in the culture and in ascites fluids were
equal to 1:1024 and 1:25600, respectively. Mab belong to IgM isotype and did not react with
similar antigens of heterogeneous parasites.
The principle of detecting specific antibodies in serum by sandwich-ELISA was based on the
selective sensitization of the solid phase with specific protein from the composition of ES-Ag
with the help of Mab 3B1G10. Sera samples were obtained from 108 cows of endemic areas at
slaughterhouse and were used to determine the diagnostic value of sandwich-immunoassay in
comparison with its indirect variant. According to the results of autopsy cystic echinococcosis
319
(cysts with protoscoleces in lungs and liver, calcified echinococcal foci, acephalocysts) were
found in 21 (19,4%) cows. Both variants of ELISA detected the presence of antibodies in
serum of 9 cows with hydatid cysts protoscoleces or 42,9% of the total number of infested
animals. In 2 heards with calcified foci (9,5%) and in 5 animals with acephalocysts (23,8%)
specific antibodies were not found by the compared tests. Cows with acephalocysts and
calcified foci showed positive results by indirect-immunoassay in 4 cases (19,0%), whereas
antibodies in sandwich-immunoassay were revealed only in the serum of 1 head with
acephalocysts (4,7%). Moreover, indirect-ELISA gave false-positive results in serological
testing 2 cows 1 of which had pulmonary tuberculosis and the other had no visible changes in
the internal organs. Thus, sandwich-ELISA on the basis of Mab 3B1G10 directed to E.
granulosus ES was more specific than its indirect option for diagnosis of cystic
echinococcosis.
320
SICBENS-1306
Application of Hydroxynitrile Lyase from Prunus Amygdalus Turcomanica
in Kinetic Resolution of Racemic Mandelonitrile
Dilek Alagöz*
University of Cukurova , Vocational School of Imamoglu, Turkey
Deniz Yıldırım
University of Cukurova, Vocational School of Ceyhan, Turkey
Seyhan Tükel
University of Cukurova, Faculty of Science and Letters, Department of Chemistry,
Turkey
Abstract
The kinetic resolution is generally defined as a process where the two enantiomers of a racemic
mixture are transformed to products at different rates. Recently, enzymatic kinetic resolution
has recently become an alternative to the traditional kinetic resolution since one of the
enantiomers of the racemate is highly selectively transformed to product, whereas the other is
left behind. However, one obvious limitation of kinetic resolution is the maximum theoretical
yield is limited to 50% [1].
Hydroxynitrile lyases (HNLs) have been widely used in yielding optically pure cyanohydrins
[2,3]. In this study, we focused on the kinetic resolution of (R/S)-mandelonitrile that is one of
the recent developments in the rapidly growing field of HNL). The kinetic resolution of
(R/S)-mandelonitrile was studied using an free and immobilized HNLs from Prunus amygdalus
turcomanica (PatHNL) seeds. In most cases, the reaction proceeded with 90, 90 and 93%
enantiomeric excess and 45, 44 and 41% yield for free PatHNL, PatHNL immobilized onto
Eupergit CM and PatHNL immobilized as cross-linked enzyme aggregates (CLEA),
respectively.
In summary, we can offer a new method to obtain (S)-mandelonitrile enantiomer by using
free and immobilized PatHNLs.
Keywords: kinetic resolution, hydroxynitrile lyase, enantiomer
321
Literatures
[1] Ghanem, A. and Aboul-Eneın, H.Y.,(2005). Application of lipases in kinetic resolution
of racemates, Chirality 17, 1–15.
[2] Yildirim, D., Tükel, S. S. and Alagöz, D.,(2014). Crosslinked enzyme aggregates of
hydroxynitrile lyase partially purified from Prunus dulcis seeds and its application for the
synthesis of enantiopure cyanohydrins. Biotechnology Progress, 30 (4), 818-827.
[3]Hanefeld, U., (2014). Immobilisation of hydroxynitrile lyases. Chem Soc Rev, 42 (15),
6203-6568.
322
SICBENS-1330
Immature Seed as a Novel Explant for Indirect Somatic Embryogenesis in a
Woody Tree; Dalbergia Sissoo Roxb
Hafiz Mamoon Rehman
Chonnam National University
Abstract
A competent and trouble-free method was scientifically developed for inducing indirect
somatic embryogenesis (ISE) and plantlet regeneration from immature seeds (ISs) as a novel
explant of Dalbergia sissoo Roxb. A two-step culture system was used to stimulate somatic
embryos. Callus formation and consequent plant regeneration of D. sissoo was obtained by
culturing the ISs on Murashige and Skoog (MS) medium supplemented with different growth
regulators. For callus formation, 0.6 mg L-1
2,4-dichlorophenoxyacetic acid (2,4-D) + 0.20 mg
L-1
kinetin was found the best as compared to the other two tested treatments i.e 2 mg L-1
Benzylaminopurine (BAP) + 0.5 mg L-1
naphthalene acetic acid (NAA) and 0.6 mg L-1
2,
4-D + 0.5 mg L-1
NAA. Embryogenic calli with premier rating were subjected to shoot
formation on a medium supplemented with 0.26 mg L-1
NAA + 2 mg L-1
BAA. This treatment
achieved very high shoot rate as compared to other treatments (0.26 mg L-1
NAA + 2.13 mg L-1
Kinetin or 0.5 mg L-1
indole acetic acid (IAA) + 2 mg L-1
BAP. Best rooting was obtained on
½ MS medium with 0.75 mg L-1
Indole Butyric acid (IBA). Complete plantlets were
acclimatized and then planted in the field with 80 % survival and normal phenotype. This is a
successful pioneering report on utilizing immature seed as an explant for ISE and in vitro
plantlet regeneration in D. sissoo Roxb.
Keywords: Roxb, Dalbergia sissoo, immature seeds, Somatic Embryogenesis and two step
culture system.
323
SICBENS-1334
Cross-Linked Enzyme Aggregates of Hydroxynitrile Lyase from Wild
Appricot
Deniz Yildirim*
University of Cukurova, Vocational School of Ceyhan, Turkey
Dilek Alagöz
University of Cukurova, Vocational School of Imamoglu, Turkey
Seyhan Tükel
University of Cukurova, Department of Chemistry, Turkey
Abstract
Hydroxynitrile lyases (HNLs) are powerful catalysts in the synthesis of enantiopure
cyanohydrins. Both alcohol and nitrile groups of cyanohydrins can be transformed to a various
range of key synthons for the preparation of pharmaceuticals, agrochemicals, and other many
biologically active compounds [1-3].
In this study, cross-linked enzyme aggregates of Prunus pseudoarmeniaca HNL
(PpaHNL-CLEAs) were prepared from wild appricot seeds and characterized in terms of lyase
activity. The optimal pH and temperature were determined as 5.5 and 25 °C, respectively for
mandelonitrile cleavage. The kinetic parameters, Km and Vmax, were 1.1 mM and 16 U/g CLEA
for mandelonitrile, respectively. PpaHNL-CLEA was also tested for enantiopure
(R)-mandelonitrile synthesis. After 72 h reaction time, (R)-mandelonitrile was obtained with
100% yield and 97% enantiopurity.
In conclusion, PpaHNL-CLEA can be an alternative biocatalyst in the preparation of
enantiopure cyanohydrins.
References
Yildirim, D., Tükel, S. S. and Alagöz, D., Crosslinked enzyme aggregates of hydroxynitrile
lyase partially purified from Prunus dulcis seeds and its application for the synthesis of
enantiopure cyanohydrins, Biotechnology Progress, 2014, pp. 818-827.
Alagöz, D., Tükel, S. S. and Yildirim, D., Purification, immobilization and characterization of
(R)-hydroxynitrile lyase from Prunus amygdalus turcomanica seeds and their applicability
for synthesis of enantiopure cyanohydrins, Journal of Molecular Catalysis B: Enzymatic,
2014, pp. 40-46.
324
Hanefeld, U., Immobilisation of hydroxynitrile lyases, Chemical Society Reviews, 2013, pp.
6308-6321.
325
SICBENS-1333
A Method to Prepare Large Amount of Sg-6 Saponins from Wild Soybean
Seeds
Mei Itabashi*, Panneerselvam Krishnamurthy, Tai-Sun Shin, Gyuhwa Chung
Chonnam National University *[email protected]
Ayaka Kurosaka, Chigen Tsukamoto
Iwate University
Seung Hwan Yang
Myongji University
More than 100 kinds of saponins have been reported in the seeds of soybean (Glycine max) and
wild soybean (Glycine soja). They are classified into group A, DDMP, and Sg-6 saponins by
their aglycone. Sg-6 saponins are further subclassed into group H, I, and J saponins which are
consisted of three different aglycones, soyasapogenol H, I, and J, respectively. Group A and
DDMP saponins have been reported that they show specific characteristics depending on their
chemical structures. Group A saponins show astringent taste, anti-obesity effect, while DDMP
saponins and their degraded group B and E saponins show inhibitory effect on cancer cells,
heaptoprotective and neuroprotective activities, and so on. However, there is no information on
physiological or biological activities of Sg-6 saponins, thus they may show diverse biological
activities. To evaluate the health benefits, a large amount of Sg-6 saponins is required. Since
group A and Sg-6 saponins elute very closely by HPLC with a reverse phase preparative
column, contamination of group A saponins was unavoidable. In this report, new effective
method to separate Sg-6 saponins from group A saponins is provided. A wild soybean
collection CW15419, which accumulate Sg-6 saponins in the seed, were obtained from
Chung‟s wild legume germplasm collection (CWLGC) at the Chonnam National University,
and were used as a material for this research. Saponins were extracted with 5-fold volume
(v/w) of 80% (v/v) methanol and the extracts were dried up with a rotary evaporator in
vacuum. Aliquot of the concentrated extract was weighed and mixed with 15-fold (v/w) of
1N-KOH solution to deacetylate group A saponins, and to the mixture 1/10 volume of
concentrated hydrochloric acid were added. The supernatant was removed to obtain precipitate,
which contains all saponin components. The precipitate was washed with 15-fold (v/w) of
0.1N-HCl. This process was performed three times to remove other components such as sugars,
isoflavones and anthocyanins. The obtained precipitate was dissolved with 15-fold (v/w) of
40% methanol containing 0.1% acetic acid. The supernatant was applied to a reversed phase
326
C-18 open column and saponins were collected by stepwise elution of 50, 60, 70% methanol
containing 0.1% acetic acid.
Fractions were collected and concentrated with a rotary evaporator in vacuum, and were
freeze-dried. By this procedure, six components (H- - - - , I- g and I-
of Sg-6 saponins were clearly separated from the other saponins. From 450g of material wild
- -6
saponins were separated, and the Sg-6 saponin composition (w/w, %) was H- -
H- - - - = 13.7: 19.4: 10.8: 8.0: 28.3: 19.8. The recovery was 62% which is
extremely higher than conventional HPLC approach for separation of Sg-6 saponins. This
report is the first one to describe how to prepare a large amount of Sg-6 saponins from the
seeds of soybeans and wild soybeans. This method can apply to obtain the other group A, B, E,
and DDMP saponins by using other variety of soybean or wild soybean as a material.
327
SICBENS-1336
The Cases Study and Species Selection (Butterfly, Plant) for the Butterfly
Garden Landscaping
Son, Jinkwan
National Academy of Agricultural Science, RDA, Republic of Korea
Lee Siyoung
National Academy of Agricultural Science,RDA, Republic of Korea
Kim, Jongku
National Academy of Agricultural Science, Republic of Korea
Kong, Minjae
Relationship of Dankook Univ. and Rural Development Administration, Republic of Korea
Kang, Donghyeon
National Academy of Agricultural Science, Republic of Korea
Abstract
We have been designing various types of parks in order to resolve all these different
environmental issues in urban areas. In terms of a few of those parks, they are built as spaces in
which living organisms can live, and they are named ecological parks, dragonflies parks and
others. Such organism habitats are considered places not only to improve biodiversity of the
urban areas but also to give people chances to observe living organisms. In the light of that, this
study carried out its investigation in order to present basic materials to help planning and
designing of a butterfly garden. Through literature reviews, the study categorized butterflies
into a group for breeding and another group for freeing. 21 species of butterflies were collected,
and there were 1 Order, 4 Family and 16 Gunus. As for those 21 species, Phengaris teleius,
Pseudozizeeria maha, Thecla betulae, Argynnis hyperbius, Dichorragia nesimachus, Hestina
assimilis, Polygonia c-aureum, Sasakia charonda, Vanessa cardui, Vanessa indica, Boloria
selene, Byasa alcinous, Papilio machaon, Papilio macilentus, Papilio protenor, Papilio xuthus,
Papilio bianor, Sericinus montela, Colias erate, Eurema hecabe, Pieris rapae were included.
The study extracted host plants and nectar plants of the 21 species.
328
The selected plants were 76 Taxa which contained 23 Family, 57 Gunus, 66 Species, 9 Variety
and 1 Forma. 76 Taxa were analyzed into Compositae 14 Taxa(18.4%), Leguminosae 13
Taxa(17.1%), Rutaceae 8 Taxa(10.5%), Cruciferae 7 Taxa(9.2%), Umbelliferae 5 Taxa(6.6%)
and Rosaceae, Violaceae, Verbenaceae as well as others. As far as this study believes, we
should increase good understanding on life styles of the butterflies before getting into
designing the butterfly garden. In addition, we must need this patch large enough to attract and
grow the butterflies and the plants to build the garden. Add to that, since each of the species
has every different taste in their favorite plants, we must take time conducting careful
investigations on such tastes. The study does not assume that it is proper to design the butterfly
garden near agricultural lands because any species harmful to agricultural crops might be
introduced. When the butterflies are successfully attracted, moths also possibly come, and that
is why the study argues for a good solid review on how to effectively block the moths. After all,
the study suggests that we must work on prevention of ecological disruptions by cleverly
controlling the numbers of individual butterflies and individual plants.
Acknowledgements
This study was carried out with the support of the Project No. PJ009492, National Academy of
Agricultural Science, RDA, Republic of Korea
Keyword: Gardening, Insect, Restoration, Creation, Conservation
329
SICBENS-1303
Response of Red Blood Cell ATPase Activity in Freshwater Fish
(Oreochromis Niloticus) Exposed to Copper in Differing Ca2+ Levels
Gülüzar Atli*
Cukurova University Vocational School of Imamoglu, Turkey
Esin G. Kanak, Mustafa Canli
Cukurova University Faculty of Science, Department of Biology, Turkey
Abstract
Status of the several biochemical processes may be reflected in fish blood known to be altered
by heavy metals. Ca2+
is one of the most important factors in water chemistry as it affects fish
physiology and metal toxicity. In this basis, it was attempted to investigate the Ca2+
impact on
Cu2+
toxicity in the response of red blood cell ATPase activity changes in Nile tilapia
Oreochromis niloticus after sublethal acute and chronic exposures. Each group contained six
fish exposed to different Ca2+
environments. Basal Ca2+
level of exposure water was 15 mg
Ca2+
/L and 30 and 90 mg Ca2+
/L were achieved after Ca2+
addition to this water. Control
(15-Ca, 30-Ca and 90-Ca) and Cu2+
exposed groups (15-Ca+Cu, 30-Ca+Cu and 90-Ca+Cu)
were exposed for 3 and 30 days. Metal exposed group was treated with 0.3 µM and 0.03 µM of
Cu (CuCl2.2H2O) for acute and chronic exposures, respectively and followed by ATPase
activity measurement. In acute exposure; Total-, Na+/K
+-and Mg
2+-ATPase activities decreased
at 30-Ca exposed group though Total- and Na+/K
+-ATPase activities increased at 90-Ca
exposed group in contrast to 15-Ca exposure. Cu2+
decreased the Total- and Mg2+
-ATPase
activities in both 30- and 90-Ca+Cu exposed groups when comparing to 15-Ca+Cu exposure.
Total- and Mg2+
-ATPase activities significantly increased at all Cu2+
exposed groups in
contrast to Ca2+
environment at 15, 30 and 90 mg/L concentrations. Nevertheless,
Na+/K
+-ATPase activity varied as increase and decrease at 30- and 90-Ca+Cu exposed groups
in contrast to 30- and 90-Ca exposed groups, respectively. In chronic exposure; decreased
activities of Na+/K
+-ATPase in 30- and 90-Ca exposed group and Total-ATPase in only 90-Ca
exposed group were observed in contrast to 15-Ca exposed group. When comparing to
15-Ca+Cu exposed group, Mg2+
-ATPase activity decreased though Na+/K
+-ATPase activity
increased in 30-Ca+Cu exposed group and also 90-Ca+Cu exposure increased Total- and
Mg2+
-ATPase activities. The increased activities of Na+/K
+-ATPase in 15-Ca+Cu exposed
group and also Total- and Mg2+
-ATPase in 90-Ca+Cu exposed group were recorded in contrast
to their own Ca2+
environments without Cu2+
treatment. Ca2+
and Cu2+
significantly affected
the ATPase activities separately according to increased Ca2+
conditions whereas activity
changes seemed to be different when comparing the Cu2+
effect in their own Ca2+
conditions.
330
Sensitivity and variability of ATPase activities indicated the importance of these enzymes
reflecting the response of osmoregulation system as significant biological indicators to evaluate
fish health under metal stress in different natural waters.
Keywords: ATPase, Blood, Calcium, Copper, O. niloticus
331
Poster Session (2)
Civil Engineering / Electrical and Electronic Engineering /
Environmental Sciences /Fundamental & Applied Sciences /
Material Science and Engineering
2015/6/27 Saturday 15:00-16:00 4F Pre-Function Area
SICASE-798
Effect of Principal Axis Orientation on Seismic Capacity of Shear Governed RC Columns
Wen-I Liao︱National Taipei University of Technology
Mustapha Jallow︱National Taipei University of Technology
Yafatou Sanneh︱National Taipei University of Technology
Martin R. Mendy︱National Taipei University of Technology
SICASE-803
Resistive Switching Characteristics of Sm2O3 and Lu2O3 Thin Films for Low-Power
Flexible Memory Applications
Tung-Ming Pan︱Chang Gung University
Han-Xuan Xie︱Chang Gung University
Somnath Mondal︱Chang Gung University
Keiichi Koyama︱Kagoshima University
SICASE-796
Structural and Electrical Properties of High-K Tm2O3 Gate Dielectrics for InGaZnO
Thin Film Transistors
Tung-Ming Pan︱Chang Gung University
Tung-Yu Wu︱Chang Gung University
Fa-Hsyang Chen︱Chang Gung University
Keiichi Koyama︱Kagoshima University
332
SICASE-779
Green Bikeway Planning for Sustainable Environment in Scenic Area
Yao-Tsung Ko︱Tunghai University
Ming-Shih Chen︱Tunghai University
Pei-Ling Wu︱Tunghai University
SICASE-765
Phthalide, a New Naturally Occurring Bioactive Compound from Trichoderma
Harzianum WuHQ75 against Chilli Anthracnose
Chotika Jeerapong︱Walailak University
Worrapong Phupong︱Walailak University
Warin Intana︱Walailak University
Patoomratana Tuchinda︱Mahidol University
Sopana Wongthong︱Nakhon Si Thammarath Rajabhat University
SICASE-823
Construction and Sequencing Analysis of Scfv Antibody Fragment Derived from
Monoclonal Antibody Against Norfloxacin (Nor155)
Jirawat Mala︱Chulalongkorn University
Sarintip Sooksai︱Chulalongkorn University
Tanapat Palaga︱Chulalongkorn University
Songchan Puthong︱Chulalongkorn University
Kittinan Komolpis︱Chulalongkorn University
Yoshinobu Kaneko︱Osaka University
SICASE-743
Properties and Microstructure of Cobalt-Based Alloy Layer by Laser Cladding
Yao-Sheng Yang︱Kao Yuan University
Yin-Fang Wei︱Kao Yuan University
Bo-Yong Chen︱Kao Yuan University
Yeou-Yih Tsai︱Kao Yuan University
333
SICASE-798
Effect of Principal Axis Orientation on Seismic Capacity of Shear Governed
RC Columns
Wen-I Liao*, Mustapha Jallow, Yafatou Sanneh, Martin R. Mendy
Department of Civil Engineering, National Taipei University of Technology, Taiwan *[email protected]
Abstract
This paper presents the results of shake table tests on six full-scale shear governed RC columns.
Two types of RC columns were designed for the shake table tests. One is the columns with an
effective height of 0.6 m and the other is 0.7 m. The cross section of all the experimental
columns is 0.3x0.3 m2 with 6-#6 longitudinal steel rebars. The stirrup arrangements for the
columns with a height of 0.7 m or 0.6 m are #2@10 cm or #3@10cm, respectively. To evaluate
the effect of the principal axis orientation on the seismic performance of RC short columns, in
the three specimens of each type the principal axes of cross-section are in the directions of 0, 22.5,
and 45 degrees to the excitation direction. The input ground motion for all tests is the
acceleration time history recorded in the 1999 Taiwan Chi-Chi earthquake. The response time
histories for the accelerations and displacements, as well as the force-displacement hysteretic
loops are presented in this study. Based on the experimental results, the shear strength and
ductility of columns with different principal axis orientations are further discussed and compared
with those results obtained by conventional design formula.
Keywords: RC column, Shear, Principal axis, Shake table test
1. Introduction
Obviously, the concept of the performance based design attracted the attention of many
researchers in the last decade and currently it is intended to be incorporated into the next
generation of the seismic design code for new design or retrofitted structures in some
developed countries such as U.S.A [1, 2] and Japan. The functionality of structures and
facilities housed in structures during or after an earthquake event depends greatly on the actual
seismic performance of structures. For this reason, the accurate simulated numerical hysteretic
model for the material or component behaviors plays an important role for the structural
analysis in the practical application. The numerical hysteretic models for the shear behavior or
the flexure-shear interaction of reinforced concrete members have been paid more efforts in the
past researches since it is really much more complicated compared with the pure flexure
behavior. Besides, shear failure could lead to reduction of the desired lateral strength, change
of the inelastic behavior, loss of the axial load capacity, and abrupt collapse of the building.
334
Numerous models for the prediction of the shear strength (Ang et al. 1989 [3], Watanabe and
Ichinose 1991 [4], Priestley et al. 1994 [5]), as well as the response of the shear force-shear
deformation relationship (Collins 1978 [6], Vecchio and Collins 1986 [7], Hsu 1988 [8], Pang
and Hsu 1996 [9]) under monotonic loading have been developed. Various hysteretic models
for the shear behavior have also been proposed (Ma et al. 1976 [10], Maruyama and Jirsa 1979
[11], Takayanagi and Schnobrich 1979 [12], Ozcebe and Saatcioglu 1989 [13]). In 2001, Lee
and Elnashai [14] proposed a hysteretic model considering the flexure-shear interaction. In
2007, Dönmez and Sözen [15] developed a numerical model which can estimate the maximum
force and displacement responses, as well as estimate all ranges of response of biaxially loaded
reinforced concrete members.
While many studies have been conducted on RC columns under uniaxial loads, few studies
have been conducted under biaxial loads. The biaxial loads, unlike uniaxial loads, account for
the fact that earthquakes forces can act on structures in any direction. Forces in any direction
on a given plane could be considered to be equivalent to biaxial loads acting along two
perpendicular directions. Thus, columns tested under biaxial loads are subjected to a more
realistic form of earthquake loads in comparison to columns under uniaxial loads. Therefore, a
series of studies, including the experimental scheme (shake table tests and reversed cyclic
loading tests) and the prediction of the simulated numerical model, is included in this long term
research project. This paper addresses the performance of reinforced concrete columns
subjected to biaxial dynamic loads. Six full-scale shear governed reinforced concrete columns
are designed for the shake table test to evaluate the effect of the principal axis orientation on
the seismic performance. The parameters that were varied in the six specimens are the ratio of
the biaxial loads and the amount of lateral ties in the column specimens. It is believed that the
test results would be greatly beneficial for the correct prediction of the numerical hysteretic
model for the shear behavior of reinforced concrete columns. The effect of the long-sustained
load on reinforced concrete columns [16] will be considered for the prediction of the simulated
numerical model. It is worthy of noting that the seismic reactive weights are simulated by the
added precast concrete mass block to facilitate the experimental setup.
2. Test Structure and Test Program
The designed experimental configuration and practical setup are shown in Figure 1 ~ Figure 3.
In order to enhance the experimental efficiency and reduce the error due to the construction of
components, the seismic reactive weight of 143kN is simulated by the assemblage of the added
precast concrete mass block. The practical experience shows that the setup, test and removal of
each specimen only take 1.5 days approximately. The head of the column is bolted with the
steel plates embedded in the mass block to assure the close assemblage of the concrete mass
block and the reinforced concrete column. The height of the reinforced concrete foundation is
50cm.
335
The foundation is designed to remain elastic when the failure of the reinforced concrete column
occurs. Two load cells are placed underneath the foundation to measure the vertical load and
base shear during the test. Furthermore, to prevent the abrupt falling of concrete mass block
due to the brittle shear failure of the reinforced concrete column, four steel columns are
equipped in the four corners of concrete mass block for supporting. The steel roller is designed
to locate between the concrete mass block and the top plate of steel column to not only reduce
the friction but also resist the overturning moment induced by the concrete mass block. In
addition, the reinforced concrete column could behave as a nearly perfect reverse curvature
model such that the following analysis is relatively simple.
Two types of reinforced concrete columns are designed for the shake table tests. The cross
section of all the tested columns is 0.3mx0.3m with 6-#6 longitudinal steel rebars. The stirrup
arrangements for the tested columns are #2@10 cm and #3@10cm. The average concrete
strength is 20502/ cmN . The average yielding stresses for #2 (stirrups of columns), #3 (stirrups
and ties of columns) and #6 (longitudinal steel rebars of columns) are 552cmkN / , 41
2cmkN /
and 532cmkN / , respectively.
336
For the first type of the experimental column, the effective height is 0.7m and the stirrup
arrangement is #2@10cm within the middle length of 30cm. In order to control the shear failure
occurs in the middle portion of the column, #3@10cm for the stirrup and additional one tie is
designed for the length of 20cm in the both ends. To evaluate the effect of the principal axis
orientation on the seismic performance, the principal axes of the cross section in the directions
of 0, 22.5, and 45 degrees to the excitation direction are assigned for the three specimens, SH1,
SH2 and SH3, respectively, as illustrated in Figure 4. For the second type of the reinforced
concrete column, the effective height is 0.6m and the stirrup arrangement is #3@10cm within
the middle length of 30cm. Similarly, in order to prevent the undesired failure, #[email protected] for
the stirrup and additional one tie is designed for the length of 15cm in the both ends. The
principal axes of the cross section in the directions of 0, 22.5, and 45 degrees to the excitation
direction are assigned for the three specimens, SH4, SH5 and SH6, respectively.
337
The arrangement of instrumentation for the first type of the experimental column is shown in
Figure 5. The inertia force induced by the concrete mass block is calculated based on the
measurement of the accelerometers. Since the reinforced concrete column behaves as a reverse
curvature model, the difference between the measurements of displacement transducer D1 and
D2 could reasonably represent the drift of the reinforced concrete column. The axial load and
the shear force undertaken by the reinforced concrete column could be measured by the load
cells underneath the foundation.
The input ground acceleration history for all tests is the E-W component of the TCU078 station
of the 1999 Taiwan Chi-Chi earthquake (denoted as TCU078EW). Corresponding to the input
338
ground motion, the test protocol is sequentially the peak ground acceleration (PGA) of 50gal,
200gal, and with an increment of 200gal or 300gal for next test run until the failure of the
specimen occurs.
3. Test Results
The test results indicate that for all specimens there is no obvious shear crack and concrete
spalling until the PGA of the input ground motion is greater than 1000gal. Besides, the acute
failure occurs when the PGA of the input ground motion reaches 1500gal or 1800gal. The
failure statuses of specimen SH1 and SH5 after test runs with PGA equal to 1500gal and
1600gal respectively are illustrated in Figure 6 and Figure 7, and from which the obvious shear
failures are observed.
339
340
The shear-drift hysteretic loops of the all specimens subjected to the input ground motion are
illustrated in Figure 8 ~ Figure 13 from which it is seen that the shear strength is degraded
abruptly due to the brittle shear failure. The comparisons of the different principal axes
orientation for the shear-drift relations of two types of specimens are shown in Figure 14 and
Figure 15. The tested ultimate shear strengths for all specimens are summarized in Table 1. For
the first type of the reinforced concrete column, the tested ultimate shear strength is 270kN
approximately. It is concluded that the effect of principal axis orientation on the performance
of the shear strength is very limited. Furthermore, for the second type of the reinforced
concrete column, the tested ultimate shear strength is 300kN approximately, which is relatively
higher than the result of the first type of the reinforced concrete column due to the more
conservative design for the stirrup arrangement.
341
Similarly, the shear strength is not affected significantly by the variation of the principal axis
orientation. For the first type and second type of the reinforced concrete columns, the nominal
shear strengths calculated by the design formula are equal to 224kN and 288kN, respectively.
4. Conclusions
Before the prediction of the 3-dimension constitutive model and numerical hysteretic model,
the appropriate experiment is necessary to be conducted. Therefore, a series of studies,
including the experimental scheme (shake table tests and cyclic loading tests) and the
prediction of the simulated numerical model, is included in this research program. This study is
the first stage of a research program to design and conduct the shake table test to discuss the
effect of the principal axis orientation on the seismic performance of reinforced concrete
columns such as the shear strength and hysteretic loops. The satisfactory test results are
obtained due to the deliberate design of the experimental configuration. It shows that current
design formula for estimating the shear strength is relatively conservative, and the effect of the
principal axis orientation on the shear strength is not significant. In the next stage, the cyclic
loading test will be conducted and the experimental results could be incorporated into the test
database. Based on the complete experimental data, the shear strength and hysteretic loops of
reinforced concrete columns with different principal axis orientations can be further discussed
and compared with the conventional design. As a consequence, the incredible 3-dimension
constitutive model and numerical hysteretic model of reinforced concrete columns for the
structural analysis in the practical engineering application is the final objective of this research
program.
5. Acknowledgements
Financial support by National Science Council (NSC) is acknowledged. Immense gratitude
must be conveyed to the laboratory support provided by National Center for Research on
Earthquake Engineering (NCREE).
6. References
International Building Code, International Conference of Building Officials（ICBO）, Whittier,
California, 2006.
ASCE07-02, Minimum Design Loads for Buildings and Other Structures, American Society of
Civil Engineering (ASCE), Reston, V.A., 2005.
Ang, B. G., Priestley, M. J. N., and Paulay, T. (1989), „„Seismic Shear Strength of Circular
Reinforced Concrete Columns,‟‟ ACI Structural Journal, Vol. 86, No. 1, pp. 45–59.
Watanabe, F., and Ichinose, T. (1991), „„Strength and Ductility Design of RC Members
Subjected to Combined Bending and Shear,‟‟ Proceeding of Workshop on Concrete Shear
in Earthquakes, University of Houston, pp. 429–438.
Priestley, M. J. N., Verma, R., and Xiao, Y. (1994), „„Seismic Shear Strength of Reinforced
342
Concrete Columns,‟‟ Journal of Structural Engineering, ASCE, Vol. 120, No. 8, pp. 2310–
2329.
Collins, M. P. (1978), „„Towards a Rational Theory for RC Members in Shear,‟‟ J. Journal of
Structural Engineering, ASCE, Vol. 104, No. 4, pp. 649–666.
Vecchio, F. J., and Collins, M. P. (1986), „„The Modified Compression Field Theory for
Reinforced Concrete Elements Subjected to Shear,‟‟ ACI Structural Journal, Vol. 83, No.
2, pp. 219–231.
Hsu, T. T. C. (1988), „„Softened Truss Model Theory for Shear and Torsion,‟‟ ACI Structural
Journal, Vol. 85, No. 6, pp. 624–635.
Pang, X. B. D., and Hsu, T. T. C. (1996), „„Fixed Angle Softened Truss Model for Reinforced
Concrete,‟‟ ACI Structural Journal, Vol. 93, No. 2, pp. 197–207.
Ma, S. H., Bertero, V. V., and Popov, E. P. (1976), „„Experimental and Analytical Studies on
the Hysteretic Behavior of Reinforced Concrete Rectangular and T-Beams,‟‟ Report No.
EERC 76-2, Earthquake Engineering Research Center, University of California, Berkeley.
Maruyama, K., and Jirsa, J. O. (1979), „„Shear Behavior of Reinforced Concrete Members
under Bidirectional Reversed Lateral Loading,‟‟ CESRL Report No. 79-1, University of
Texas at Austin.
Takayanagi, T., and Schnobrich, W. C. (1979), „„Non-Linear Analysis of Coupled Wall
Systems,‟‟ Earthquake Engineering and Structural Dynamics, Vol. 7, pp. 1–22, Takeda.
Ozcebe, G., and Saatcioglu, M. (1989), „„Hysteretic Shear Model for Reinforced Concrete
Members,‟‟ Journal of Structural Engineering, ASCE, Vol. 115, No. 1, pp. 132–148.
Lee, D. H. and Elnashai, A. S. (2004), „„Seismic Analysis of RC Bridge Columns with
Flexure-Shear Interaction,‟‟ Journal of Structural Engineering, ASCE, Vol. 130, No. 11,
pp. 1692–1703.
Dönmez, C. and Sözen., M.A. (2007), “Numerical Model for Biaxial Earthquake Response of
Reinforced Concrete,” Computer-Aided Civil and Infrastructure Engineering, Vol. 22, pp.
238-253.
Rey, L. E., Schultz ,A. E., and Welton, S. S. (2004), “Long-Term Effects on Response of
Reinforced Concrete Columns to Cyclic Loading,” Journal of Structural Engineering,
ASCE, Vol. 130, No. 9, pp. 1320-1332.
343
SICASE-803
Resistive Switching Characteristics of Sm2O3 and Lu2O3 Thin Films for
Low-Power Flexible Memory Applications
Tung-Ming Pan*, Han-Xuan Xie, Somnath Mondal
Chang Gung University *[email protected]
Keiichi Koyama
Kagoshima University, Japan
Abstract
In this study, we investigated the resistive switching (RS) behavior in flexible Sm2O3 and Lu2O3
resistive random access memories (ReRAM). The structural morphologies of the Sm2O3 and
Lu2O3 thin films depend on the lattice energy of the oxides. The dominant current conduction
mechanism in the oxide layer changes from electrode control Schottky emission in Sm2O3 to
bulk controlled space-charge-limited-current in Lu2O3. The barrier height extracted from
Schottky emission model is 0.96 eV in Sm2O3 thin film, whereas the activation energy of traps
calculated from the Arrhenius plots is about 0.23 eV in Lu2O3 thin film. In addition, the
Ni/Sm2O3/ITO flexible memory device shows promising RS behavior with very low power of
operation (~30 µW). The memory reliability characteristics of switching endurance and data
retention show promising for future memory applications. The filament conduction model is
adopted to describe the RS behavior in the Sm2O3 and Lu2O3 ReRAM devices. The improved RS
performance in Sm2O3 thin film is attributed to the different physical properties of the thin films.
Keyword: resistive switching (RS), resistive random access memories (ReRAM), Lu2O3, Sm2O3.
1. Introduction
The resistance random access memory (ReRAM) in various metal oxides [1-3], organic
molecules [4, 5], and graphene oxides [6-8] has been proposed as a promising future nonvolatile
memory (NVM) device due to its simple device structures, high speed operation and
complementary metal-oxide-semiconductor (CMOS) compatible fabrication process. The
demand for flexible electronics such as flexible displays and wearable computers is revived
because of their inherit merits of low cost, light weight, excellent portability and user friendly
interfaces over conventional bulk silicon technology [9]. However, a major challenge for flexible
electronic systems is the lack of good performance NVM devices fabricated at low temperature
[10, 11]. Although organic memory and other low temperature processed oxide based memory
show good flexibility, its performance cannot match that of conventional charge trap flash
memory.
344
Recently, flexible resistive memories have been reported in various amorphous oxides including
TiO2 [12], Al2O3 [13], ZnO [14], and InGaZnO [15] thin films. However, the large switching
power, deprived reliability, and poor understanding of the resistive switching (RS) behavior are
some of the fundamental issues which limit its application for high-density and low-power
flexible electronics.
Rare-earth (RE) metal oxides have shown promising characteristics as high-κ gate insulators in
advanced CMOS devices because of their high resistivity, large dielectric constants, wide
bandgaps, and good thermodynamic stability [16]. The recent studies show that RE oxide
materials, such as Yb2O3 [17], Gd2O3 [18], and CeO2 [19], are promising for RS material, though
there are very little reports on the memory performance on flexible substrate. The superior RS
behavior and room temperature deposition process allow it to be a possible functional material for
flexible electronics. Therefore, in this study we present samarium oxide (Sm2O3) and lutetium
oxide (Lu2O3) ReRAM devices on ITO coated polyethylene terephthalate (PET) substrate for
flexible memory applications. The conduction mechanisms in flexible Sm2O3 and Lu2O3 RS
films are analyzed. Different electrical conduction mechanisms of Schottky emission and space
charge limited current (SCLC) conduction were observed in Sm2O3 and Lu2O3 thin films,
respectively. The RS characteristics and the memory reliability behavior are also investigated.
Unlike other typical flexible resistive memory, better RS characteristics were found in terms of
lower switching power and good distribution of switching parameters. Filamentary conduction
model was adopted to link the RS behavior in both Sm2O3 and Lu2O3 ReRAM devices.
2. Experimental
Flexible Sm2O3 and Lu2O3 RS memory devices were fabricated on ITO coated PET substrates
(purchased from Sigma-Aldrich, Taiwan). The ITO/PET substrate was glued on a Si dummy
wafer with polyimide tape to mechanically support the flexible substrate during fabrication
process. The Sm2O3 and Lu2O3 thin films were deposited by radio frequency magnetron
sputtering system in argon-oxygen (3:1) medium at room temperature. In the sputtering process,
metal Sm and Lu targets were used for Sm2O3 (~23 nm) and Lu2O3 (~21 nm) thin films
deposition, respectively. The structural behaviors were characterized by X-ray diffraction
(XRD) method. The roughnesses of the films were measured by atomic force microscope
(AFM) and the chemical compositions of the films were examined by X-ray photoelectron
spectroscopy (XPS). Finally, a 50 nm-thick square shape (100×100 µm2) Ni metal top
electrode was deposited on the oxide films through shadow mask by thermal evaporation
technique. The electrical properties of the Sm2O3 and Lu2O3 memory devices were measured
by a semi-automated cascade measurement system equipped with Agilent E5260 high speed
semiconductor parameter analyzer.
345
3. Results and Discussion
3.1 Structural Characteristics of Sm2O3 and Lu2O3 Thin Films
Fig. 1 shows the XRD spectra recorded on the Sm2O3 and Lu2O3 ※lms. No diffraction peak was
observed from the Sm2O3 and Lu2O3 films deposited at room temperature, indicates that the
films remains in amorphous phase. However, the lower intensity peak from ITO substrate in
Sm2O3/ITO sample may be due to the slight thickness variation of the film. Figs. 2(a) and (b)
show the AFM images of Sm2O3 and Lu2O3 films, respectively. A significant difference in
surface morphology was observed in both films and the surface roughness seemed to depend
on the lattice energy of the materials. The lattice energy of -13,180 kJ/mol for Sm2O3 [20] is
lower than that of Lu2O3 (-13,871 kJ/mol) [21]. It was found that the surface of Lu2O3 film
became smoother than that of Sm2O3 film. The Rrms values of surface roughness obtained by
AFM observation were 6.16 and 1.87 nm for Sm2O3 and Lu2O3 films, respectively. This is
probably caused by the difference of lattice energy of these materials [20].
346
Figs. 3(a) and (b) show the core level Sm 3d and O 1s spectra with their appropriate peak
curve-fitting lines for the Sm2O3 thin film, respectively. As shown in Fig. 3(a), the Sm 3d line
suffers a shift of ∼1.5 eV to the lower binding energy, relative to the Sm2O3 reference position
(1084.3 eV) [22]. Other Sm 3d peak at lower binding energy of 1079.9 eV is attributed to the
existence of Sm ions in the oxide thin film [23]. Fig. 3(b) depicts that the O1s line at the
surface of Sm2O3 thin film consists of a low peak binding energy at 528.8 eV for Sm2O3 [24],
and a high peak binding energy at 531.4 eV, usually attributed to oxide defects or non-lattice
oxygen ions [25]. In addition, the XPS analyses on Lu2O3 thin film show less defects in the
oxide. Figs. 3(c) and (d) show the core level Lu 4d and O 1s spectra for the Lu2O3 thin film,
respectively. The Lu 4d line spectrum consist a higher peak binding energy at 196 eV for
Lu2O3 and a lower peak binding energy at 194.4 eV, usually depicted as Lu ions [26]. The O 1s
spectrum at the surface of Lu2O3 thin film consist two peaks binding energy at 529.2 eV for
Lu2O3 and the other at 531.4 eV for defect formation, as observed in Sm2O3 [25, 26].
347
3.2 Conduction Mechanism in Sm2O3 and Lu2O3 Thin Films
The asymmetric behavior in current voltage (I-V) characteristics of Sm2O3 film suggests the
conduction mechanism is interface limited and can be well fitted with Schottky emission model,
as depicted in Fig. 4. The Schottky emission can be expressed as [27]
02( / 4 )
* expB ox r
B
q qV tI A T
k T
(1)
where A* is the Richardson constant, T is the absolute temperature, φB is the effective barrier
height, V is the external applied voltage, q is the electron charge, ε0 is the permittivity of free
space, εr is the dynamic dielectric, tox is the oxide thickness, and kB is the Blotzmann constant.
The ln(I/T2) plot with V
1/2 should be straight line for Schottky type conduction and the barrier
height can be calculated from the intercept of the straight line. Fig. 4 shows the strait line
behavior of ln(I/T2) - V
1/2 plot for Sm2O3 thin film and a barrier height qφB of 0.96 eV was
obtained.
348
The double logarithmic plot of I-V data for Lu2O3 thin film shows two clearly distinguished
regions namely Ohmic (I∞V) and quadratic (I∞V2) in Fig. 5(a). Therefore, space charge limited
current (SCLC) is dominant in Lu2O3 thin dielectric. If a single trap level is assumed in the SCLC
conduction mechanism, the I-V relation can be described as [25]:
0I
ox
VI qn
t (2)
2
0 3
9
8II r
ox
VI
t (3)
where n0 is the effective free carrier density of traps in thermal equilibrium and µ is the electronic
mobility in Lu2O3 thin film. In region I, when the applied voltage across the oxide is small (< 0.4
V), the slope of the I-V characteristic is 1, which implies the Ohmic conduction because the
number of the injected electrons are lower compared to the thermally generated free electrons
density (n0) inside the oxide film. Thus, the current depends on the applied voltage and the
material conductivity. An electrically quasi-neutral state occurs due to weak injection of the
charge carrier at such a lower field. After a certain voltage of Vtr, a transition from Ohmic region
(I) to SCLC region (II) is observed. In region II, the injected carrier density increases rapidly
with increasing the applied voltage and exceeds the volume-generated free carrier density. These
injected carriers are partly captured by the traps in the Lu2O3 and others contribute to the total
current, namely, injected free electrons. This region is realized as the charge injection limited
and trap filling region until all the existing trapping centers in the Lu2O3 are occupied by injected
charge carriers. The transition voltage from Ohmic conduction to SCL conduction can be
expressed as [27]
2
0
0
8
9
oxtr
r
qn tV
(4)
expC t C
n t B
N E E
g N k T
(5)
where θ is the ratio of free to total carrier density, Nc is the density of state in the conduction band,
gn is the degeneracy of the energy state in the conduction band, Nt is the trap density, and Et and
Ec is the trap and conduction band energy level, respectively. Further increase in applied voltage,
more carriers will be injected from the injecting electrode and a space charge region appears near
the injecting electrode interface so that the injected excess carriers dominate the thermally
generated charge carrier and hence the current increases rapidly.
349
The trapping level in the Lu2O3 thin film can be determined from the Arrhenius plot of the
current for different electric fields. The activation energy (Ea) was calculated from the slope
ln(I)-1/T plot to be about 0.23 eV, as shown in Fig. 5(b). The transition from Ohmic to trap-filled
limited conduction mechanism at different temperatures is presented in the inset of Fig. 5(a) and
(b). It can be seen that the Vtr decreases with increasing temperature, suggesting that the thermal
generation of the carrier increases with temperature. Relatively lower voltage is required to fill
all the trap levels at higher temperature and hence Vtr decreases.
3.3 RS Behavior of Sm2O3 and Lu2O3 ReRAM Devices
Figs. 6(a) and (b) show the typical I–V RS characteristics of Sm2O3 and Lu2O3 ReRAM devices
after the standard electroforming process, respectively. After the electroforming process, the
memory device remains in low resistance state (LRS). The increase in applied positive bias on
memory cell, a sudden drop of conduction current at certain voltage (reset voltage, Vreset) is
observed and the devices transform to high resistance state (HRS). In contrast, an application of
negative bias results in a transition from HRS to LRS at certain set voltage (Vset) and this effect is
reproducible over several thousand of voltage sweeping cycles. These ReRAM devices can be
switched between two distinguished resistance state (HRS to LRS and vice versa), at a very low
power of ~30 µW (~0.3 V at 100 µA set/reset current) and ~100 µW (~1.0 V at 100 µA set/reset
current) for Sm2O3 and Lu2O3 ReRAM devices, respectively. The lower switching power is
believed due to the electron hopping conduction via oxide defects [10].
350
Figs. 7(a) and (b) show the endurance characteristics (write–read–erase–read cycle) of the
Sm2O3 and Lu2O3 flexible memory devices, respectively, under impulse voltages measured at
room temperature and 85 °C. Here, high switching voltage was applied since larger operation
voltage can induce larger fluctuation of the resistance state than lower voltage. The Sm2O3
memory device can be switched over 104 program/erase (P/E) cycle maintaining a memory
window of >102 at both room temperature and 85 °C. A small fluctuation in the resistance values
at both memory states is observed at elevated temperature and with increasing stress cycle, as
shown in Fig 7(b). This may be attributed to the generation and redistribution of oxide defects in
the switching material [10, 28]. Figs. 8(a) and (b) show the data retention characteristics of the
Sm2O3 and Lu2O3 flexible memory devices after cyclic measurement at both room temperature
and 85 °C. Good data retention of 105 s is obtained at room temperature, as shown in Fig. 8(a).
The retention characteristics at elevated temperature are fluctuating in both resistance states as
depicted in Fig. 8(b). The HRS value showed a continuous decay of resistance for few tenths of
second keeping a well resolved memory window of ~102 after 10
5 s of test. This can be correlated
well with the stress induced leakage current via generated defects in the oxide thin films [29].
4. Conclusion
In this work, the electrical conduction and RS behavior in the Sm2O3 and Lu2O3 thin films on
flexible PET substrate were studied for flexible NVM applications. The dominant current
conduction mechanism changes from bulk limited SCLC in Lu2O3 to interface controlled
Schottky emission in Sm2O3.
351
The presence of a shallow trap level at 0.23 eV below conduction band and the barrier height of
0.83 eV were estimated in Lu2O3 and Sm2O3 thin films, respectively. The filament conduction
model was successfully described for the RS behavior in the Sm2O3 and Lu2O3 thin films on
ITO/PET substrate. Improved RS behavior of very low power of operation (~30 µW) and small
distribution of switching parameters were obtained in Sm2O3 ReRAM device. The memory
reliability characteristics of switching endurance and data retention were promising for future
memory applications. The enhanced RS performance in Sm2O3 thin film was attributed to
different physical behaviors, mainly the surface roughness and the oxide defects of the thin film.
The superior RS characteristics in Ni/Sm2O3/ITO ReRAM device have great potential for future
nonvolatile flexible memory applications.
Acknowledgments
This work was supported by the National Science Council (NSC) of Republic of China under
contract no. NSC-102-2221-E-182-072-MY3.
5. References
P. Gonon, M. Mougenot, C. Vallee, C. Jorel, V. Jousseaume, H. Grampeix and, F. E. Kamel, J.
Appl. Phys. 107, 074507 (2010).
C.-Y. Lin, C.-Y. Wu, C.-Y. Wu, T.-Y. Tseng, and C. Hu, J. Appl. Phys. 102, 094101 (2007).
B. J. Choi, D. S. Jeong, S. K. Kim, C. Rohde, S. Choi, J. H. Oh, H. J. Kim, C. S. Hwang, K.
Szot, R. Waser, B. Reichenberg, and S. Tiedke, J. Appl. Phys. 98, 033715 (2005).
B. Cho, S. Song, Y. Ji, H.-G. Choi, H. C. Ko, J.-S. Lee, G.-Y. Jung, and T. Lee, IEEE Electron
Device Lett. 34, 51 (2013).
W. Bai, R. Huang, S. Member, Y. Cai, Y. Tang, X. Zhang, and Y. Wang, IEEE Electron
Device Lett. 34, 223 (2013).
S. Pinto, R. Krishna, C. Dias, G. Pimentel, G. N. P. Oliveira, J. M. Teixeira, P. Aguiar, E. Titus,
J. Gracio, J. Ventura, and J. P. Araujo, Appl. Phys. Lett. 101, 063104 (2012).
L.-H. Wang, W. Yang, Q.-Q. Sun, P. Zhou, H.-L. Lu, S.-J. Ding, and D. W. Zhang, Appl. Phys.
Lett. 100, 063509 (2012).
J. Liu, Z. Yin, X. Cao, F. Zhao, L. Wang, W. Huang, and H. Zhang, Adv. Mater. 25, 233
(2013).
S. Kim, H. Y. Jeong, S. K. Kim, S.-Y. Choi, and K. J. Lee, Nano Lett. 11, 5438 (2011).
C. H. Cheng, F. S. Yeh, and A. Chin, Adv. Mater. 23, 902 (2011).
J. W. Seo, J.-W. Park, K. S. Lim, S. J. Kang, Y. H. Hong, J. H. Yang, L. Fang, G. Y. Sung, and
H.-Ki Kim, Appl. Phys. Lett. 95, 133508 (2009).
H. Y. Jeong, Y. I. Kim, J. Y. Lee, and S. Y. Choi, Nanotechnology 21, 115203 (2010).
S. Kim and Y. K. Choi, Appl. Phys. Lett. 92, 223508 (2008).
S. Kim, H. Moon, D. Gupta, S. Yoo, and Y. K. Choi, IEEE Trans. Electron Devices 56, 696
(2009).
352
Z. Q. Wang, H. Y. Xu, X. H. Li, X. T. Zhang, Y. X. Liu, and Y. C. Liu, IEEE Electron Device
Lett. 32, 1442 (2011).
G. He and Z. Sun, High-k Gate Dielectrics for CMOS Technology (Wiley-VCH, Verlag, 2012),
p. 111.
H.-C. Tseng, T.-C. Chang, J.-J. Huang, P.-C. Yang, Y.-T. Chen, F.-Y. Jian, S. M. Sze, and
M.-J. Tsai, Appl. Phys. Lett. 99, 132104 (2011).
K. C. Liu, W. H. Tzeng, K. M. Chang, Y. C. Chan, C. C. Kuo, and C. W. Cheng,
Microelectron. Reliab. 50, 670 (2010).
C. Y. Lin, D. Y. Lee, S. Y. Wang, C. C. Lin, and T. Y. Tseng, Surf. Coat. Technol. 203, 480
(2008).
S. Ohmi, C. Kobayashi, I. Kashiwagi, C. Ohshima, H. Ishiwara, and H. Iwai, J. Electrochem.
Soc. 150, F134 (2003).
P. Darmawan, P. S. Lee, Y. Setiawan, J. C. Lai, and P. Yang, J. Vac. Sci. Technol. B 25, 1203
(2007).
Y. Uwamino, Y. Ishizuka, and H. Yamatera, J. Electron Spectrosc. Relat. Phenom. 34, 67
(1984).
G. Dufour, R. D. Karnatak, J. -M. Mariot, and C. Bonnelle, Chem. Phys. Lett. 42, 433 (1976).
B. F. Dzhurinskii, D. Gati, N. P. Sergushin, V. I. Nefedov, and Y. V. Salyn, Russ. J. Inorg.
Chem. 20, 2307 (1975).
S. Mondal, H.-Y. Chen, J.-L. Her, F.-H. Ko, and T.-M. Pan, Appl. Phys. Lett. 101, 083506
(2012).
V. I. Nefedov, D. Gati, B. F. Dzhurinskii, N. P. Sergushin, and Y. V. Salyn, Zh. Neorg. Khimii
20, 2307 (1975).
T. P. Juan, S. Chen, and J. Y. Lee, J. Appl. Phys. 95, 3120 (2004).
W. Shen, R. Dittmann, U. Breuer, and R. Waser, Appl. Phys. Lett. 93, 222102 (2008).
S. B. Chen, J. H. Lai, A. Chin, J. C. Hsieh, and J. Liu, IEEE Electron Device Lett. 23, 185
(2002).
353
SICASE-796
Structural and Electrical Properties of High-K Tm2O3 Gate Dielectrics for
InGaZnO Thin Film Transistors
Tung-Ming Pan*, Tung-Yu Wu, Fa-Hsyang Chen
Chang Gung University *[email protected]
Keiichi Koyama
Kagoshima University, Japan
Abstract
In this study, we studied the structural and electrical characteristics of high-dielectric constant
(high- Tm2O3 gate dielectrics for amorphous indium–gallium–zinc–oxide (a-IGZO) thin-film
transistor (TFT) applications. We employed X-ray diffraction to examine the growth directions
and crystallinity of the Tm2O3 films, atomic force microscopy to monitor the morphological
feature of the films, and X-ray photoelectron spectroscopy to analyze the chemical structure of
the Tm2O3 films after annealing at various temperatures. The high- Tm2O3 a-IGZO TFT
annealed at 400 °C exhibited a lower threshold voltage of 1.68 V, a larger field effect mobility of
11.8 cm2/V-s, a higher Ion/Ioff ratio of 3.9×10
7, and a smaller SS of 420 mV/dec., relative to those
of the systems that had been subjected to other annealing conditions. This result suggests the
Tm2O3 film featuring a higher dielectric constant and a smoother surface due to an increase of
the amount of oxygen sites.
Keyword: high-dielectric constant (high- Tm2O3, indium–gallium–zinc–oxide (a-IGZO),
thin-film transistor (TFT).
1. Introduction
Amorphous oxide semiconductors (AOSs) represented by amorphous indium–gallium–zinc–
oxide (a-IGZO) are expected for channel materials in thin-film transistors (TFTs) for flat, flexible,
and flat-panel displays because they have large field-effect mobilities and low-temperature
process [1-3]. In comparison with their hydrogenated a-Si counterparts, a-IGZO TFTs show high
electron mobility and good uniformity for low-temperature fabrication. However, the operating
voltage of a-IGZO TFTs is too high for practical use. The development of low-voltage operating
a-IGZO TFTs has become a priority. The use of gate insulators with high capacitance can reduce
the operating voltage [4]. The high gate capacitance is achieved by the use of thin gate insulators
or high-dielectric constant (high- ) materials. Thin gate insulator will lead to a high leakage
current owing to the occurrence of direct tunneling [5], thus causing high power consumption and
reduced device reliability.
354
As a result, high- gate oxide is better than thin gate oxide with a low-dielectric constant to
reduce the gate leakage current. Several high- gate dielectric materials, including Ta2O5 [6],
TiO2 [7], HfO2 [8], and HfLaO [9], have been widely investigated in a-IGZO TFT devices. It is
well-known that the performance and stability of TFT devices depends critically on the gate
dielectric and its interface with the active channel [10-11].
In recent years, rare-earth oxide materials, such as Nd2O3 [12], Gd2O3 [13], Pr2O3 [14], and La2O3
[15], have been extensively studied for applications of metal-oxide-semiconductor field-effect
transistor, resistive random-access memory, TFT, and flash memory devices due to their have
high values, large band gaps, low leakage currents, and high thermal stability [16-17]. Thin
Tm2O3 film can be considered for potential use as a high- gate dielectric material owing to its
high value, large conduction band offset and good thermal stability [18]. The structural and
electrical properties of Tm2O3 gate dielectrics have not yet been reported in the a-IGZO TFT
device. In addition, the threshold voltage (VTH) stability of TFT device under gate bias stress is
considered to be the most important issue for display application [11].
2. Experimental
The bottom-gate, top contact TFT devices were fabricated on SiO2/Si substrates at room
temperature and patterned for gate electrode by photolithography and dry etching. Sputter
deposition was carried out at a constant power of 500 W and a total pressure of 10 mtorr in
mixed atmospheres consisting of Ar and N2. Before the Tm2O3 deposition, a 50 nm-thick TaN
gate electrode was first deposited by reactive sputtering from a Ta target onto the SiO2/Si
substrate. The Tm2O3 dielectric film was then deposited with a thickness of ~55 nm by reactive
sputtering from the Tm target with diluted O2 (Ar/O2 = 20/5) at room temperature. The
deposition rate of the Tm2O3 dielectric film is ~0.2 Å /s. Prior to the deposition of the IGZO
active layer, post-deposition annealing (PDA) treatments were executed on the Tm2O3
dielectric for 10 min in O2 gas at various temperatures (200, 300 and 400 °C) to improve the
quality of dielectric film. An a-IGZO active layer with a thickness of 20 nm was deposited on
the Tm2O3 dielectric by radio-frequency (RF) sputtering from an a-IGZO target
(In2O3:Ga2O3:ZnO=1:1:1 mol %) at a RF power of 60 W in pure Ar ambient at a working
pressure of 10 mTorr. Then, the IGZO channel layer was annealed in N2 ambient for 10 min at
200 °C. Finally, an Al electrode with thickness of ~400 nm was thermally evaporated onto the
IGZO film through a shadow mask to define the source and drain electrodes. The channel
The crystalline structure of the Tm2O3 dielectric films annealed at various temperatures was
analyzed by using X-ray diffraction (XRD). The surface morphology and roughness of the
dielectric films were examined using atomic force microscopy (AFM). The composition and
chemical bonding in Tm2O3 films annealed at different temperatures were determined by X-ray
355
photoelectron spectroscopic (XPS). The dielectric constant of the dielectric films was
determined from the Al/Tm2O3/TaN capacitor device. The capacitance-voltage (C−V) curves
of Al/Tm2O3/TaN capacitors was performed at the frequency of 1 MHz using a
Hewlett-Packard (HP) 4285A LCR meter. The current-voltage (I−V) curves of the Tm2O3
a-IGZO TFTs were measured using a semiconductor parameter HP 4156C.
3. Results and Discussion
The crystalline structure of the Tm2O3 dielectric films after PDA at different temperatures was
examined by XRD measurements. All diffraction peaks of Tm2O3 films are indexed in Fig. 1.
Many small peaks are appeared in the XRD pattern, which confirms the amorphous nature of
the film. The significant differences in intensity and orientation of diffraction peaks are
attributed as different growth kinetics of the film materials at various temperatures. After
annealing at 200 °C, two weak Tm2O3 (222) and (440) diffraction peaks appeared. Besides the
above, the Tm2O3 (400) peak found for the sample annealed at 300 °C and 400 °C. The
intensity of diffraction peaks for the film annealed at 400 °C exhibited stronger than that at 300
°C, presumably because of the formation of a better Tm2O3 structure.
In addition to the structural and compositional studies, we also used AFM to investigate the
impact of the thermal annealing temperature on the surface roughness of the Tm2O3 films. The
effect of the moisture absorption on surface roughness should be another concern of
hygroscopic lanthanide oxide film application [19]. Figs. 2 (a)-(c) show the AFM images of the
Tm2O3 dielectric films annealed at 200, 300 and 400 °C, respectively. The surface roughness of
Tm2O3 film annealed at 200 °C has a rough surface (5.78 nm). The rougher surface may be due
to the oxygen vacancies (or defects) in the film producing the moisture absorption of Tm2O3,
leading to the nonuniform volume expansion of the film [22]. Furthermore, the surface
roughness clearly decreased with increasing the PDA temperature.
356
This result is attributed to the reduced number of oxygen vacancies or crystal defects in the
film due to the film densification, thus decreasing the surface roughness of the Tm2O3 film.
Figs. 3(a) and (b) shows the Tm 4d5/2 and O 1s spectra, respectively, for the Tm2O3 films
annealed at various temperatures. In Tm 4d5/2 and O 1s of spectra, each fitting peak is assumed
to follow the general shape of the Lorentzian-Gaussian function. We assigned the Tm 4d5/2
peak at 176.6 eV to the Tm atoms in Tm2O3 [20]. Fig. 3(a) depicts that the shift of the Tm 4d5/2
peak position to a higher binding energy increases upon increasing the PDA temperature. The
film upon thermal annealing can reduce the number of oxygen vacancies or crystal defects in
the film, thus causing the shift. The O 1s spectra of the Tm2O3 films after different annealing
treatment are shown in Fig. 3(b). These O 1s spectra are convoluted with two suitable peak
curve-fitting lines. In two sets of spectra, first peak represents the Tm–O bond (located at 529.4
eV) [21] and second peak represents the Tm–(OH) bond (located at 531.9 eV). The intensity of
the O 1s peak corresponding to Tm2O3 increased upon increasing the PDA temperature,
whereas that corresponding to Tm(OH)x decreased upon increasing the PDA temperature. This
finding suggests that the high-temperature annealing decreases the number of oxygen vacancies
or crystal defects in the film, leading to the reduction in the moisture absorption of Tm2O3 film.
357
Dielectric materials should ideally exhibit larger capacitance to enhance the drain current while
necessary charge to accumulate at a lower voltage [5]. Fig. 4(a) shows the C–V curves of the
Al/Tm2O3/TaN
Tm2O3 dielectrics annealed at 200, 300, and 400 °C is evaluated to be 8.3, 8.6, and 8.8,
2O3 dielectric increases with increasing the PDA temperature.
This result suggests that high PDA treatment makes the film densification reducing the number
of oxygen vacancies or crystal defects, causing the formation of a better Tm2O3 structure, as
evidenced by XRD analysis. Fig. 4(b) depicts the I−V characteristics of the Al/Tm2O3/TaN
capacitor devices after annealing at different temperatures. The leakage current of the
Al/Tm2O3/TaN device annealed at 400 °C exhibits smaller than other annealing temperatures.
This result may be attributed to reduction in the number of crystal defects in the film. These
defects act as trap sites to influence the dielectric leakage current and transistor operation.
Therefore, the capacitor after PDA at 200 °C has a higher leakage current.
Fig. 5(a) shows the drain current versus gate voltage (IDS-VGS) curves of the Tm2O3 a-IGZO
358
TFTs annealed at different temperatures. The field- FE) and threshold voltage
of the Tm2O3 TFTs with PDA treatment are calculated from the transfer characteristics. The
VTH of Tm2O3 a-IGZO TFT devices annealed at 200, 300, and 400 °C is determined to be 2.94,
2.03, and 1.68, respectively. The VTH value decreases with increasing the PDA temperature.
The annealing at a higher temperature reduces the number of oxygen vacancies or defects in
the film. The mobility is an important parameter in terms of the performance of TFTs. High
charge carrier mobility can be obtained when the drain current is high. The electron mobility of
the Tm2O3 a-IGZO TFT devices after PDA at 200, 300, and 400 °C is estimated to be 1.1, 5.6,
and 11.8 cm2/V-sec, respectively. The FE of a Tm2O3 a-IGZO TFT device increased with
increasing the annealing temperature. The high mobility is attributed to the smooth surface
between the Tm2O3 dielectric and IGZO channel, causing the low density of interface states at
the dielectric–IGZO interface and small amount of bulk traps in the dielectric. In contrast, this
low mobility is due to the scattering of electrons and surface roughness by interfaces of
dielectric–IGZO and crystal defects in the Tm2O3 film [23], respectively. The Ion/Ioff ratio of
the Tm2O3 a-IGZO TFT devices annealed at 200, 300, and 400 °C is 4.2×105, 4.5×10
6, and
3.9×107, respectively. After high-temperature processing, the Ion/Ioff ratio of such a Tm2O3 TFT
device is demonstrated to improve by decreasing the crystal defects of the film, thus enhancing
the electrical performance of a-IGZO TFT device. The subthreshold swing (SS) represents the
conducting behavior of the device when the gate voltage is smaller than the threshold voltage
and this can be determined by the following equation [24]:
ox
Bit
DS
GS
C
CqD
q
kT
Id
dVSS
3.2
)(log …………………….…..…(1)
where k is the Boltzmann‟s constant, T is the absolute temperature, q is the elementary charge,
Dit is the trap density at the dielectric/channel interface, CB is the bulk capacitance per unit area,
and Cox is the capacitance per unit area of the gate oxide. The Dit of the Tm2O3 a-IGZO TFT
devices after PDA at 200, 300, and 400 °C is extracted to be 7.88×1012
, 6.74×1012
, and
5.36×1012
cm−3
eV-1
[25], respectively. In the above equation, the effective interface trap state
density is related to the SS value. The SS value of the Tm2O3 a-IGZO TFT devices annealed at
200, 300, and 400 °C is 621, 523, and 420 mV/dec., respectively. The small value of SS may
be attributed to both the high gate capacitance density and the low interface charge density.
The low SS can be ascribed to the reduced Dit at the interface between the Tm2O3 film and
IGZO channel due to the low surface roughness. In contrast, the high SS is noticed in the lower
annealing temperature. The moisture absorption enhances the surface roughness of Tm2O3 film
to cause the degradation of SS value. Fig. 5(b) displays the drain current versus drain voltage
(IDS-VDS) output characteristics of the Tm2O3 a-IGZO TFTs with different PDA treatments,
measured at VDS varied from 0 V to 10 V and VGS-VTH varied from 0 V to 10 V in 2 V steps.
359
All TFT devices exhibit the field-effect modulation of the channel conductance and typical
normally-off type n-channel behavior. The Tm2O3 a-IGZO TFT after annealing at 400
exhibited the highest output current among the annealing condition.
4. Conclusion
In this paper, we proposed a high- Tm2O3 gate dielectric films for a-IGZO TFT applications. The
structural, morphological, and compositional properties of Tm2O3 dielectric films after PDA at
different temperatures were analyzed by means of XRD, AFM, and XPS, respectively. The
Tm2O3 a-IGZO TFT device annealing at 400 °C exhibited better electrical characteristics in term
of a low threshold voltage of 1.68 V, a high field effect mobility of 11.8 cm2/V-s, a low SS of 420
mV/dec. and a large Ion/Ioff ratio of 3.9×107. This result is due to the higher dielectric constant
and the smoother surface. This Tm2O3 dielectric film appears to be a very promising gate
dielectric for use in a-IGZO TFT applications.
Acknowledgments
This work was supporting by the National Science Council (NSC) of Taiwan under contract no.
NSC-102-2221-E-182-072-MY3.
5. References
K. Nomura, H. Ohta, A. Takagi, T. Kamiya, M. Hirano, and H. Hosono, Nature, vol. 432, p.
488, 2004.
E. Fortunato, P. Barquinha, and R. Martins, Adv. Mater., vol. 24, p. 2945, 2012.
S. D. Brotherton, Introduction to Thin Film Transistors, Springer, New York, 2013.
M. Houssa, High-k Gate Dielectrics, Institute of Physics, London, 2004.
G. D. Wilk, R. M. Wallace, and J. M. Anthony, J. Appl. Phys., vol. 89, p. 5243, 2001.
C. J. Chiu, S. P. Chang, and S. J. Chang, IEEE Electron Device Lett., vol. 31, p. 1245, 2010.
H. H. Hsu, C. Y. Chang, and C. H. Cheng, IEEE Electron Device Lett., vol. 34, p. 768, 2013.
L. Yuan, X. Zou, G. Fang, J. Wan, H. Zhou, and X. Zhao, IEEE Electron Device Lett., vol. 32,
p. 42, 2011.
N. C. Su, S. J. Wang, and A. Chin, IEEE Electron Device Lett., vol. 30, p. 1317, 2008.
S. Chang, Y. W. Song, S. Lee, S. Y. Lee, and B. K. Ju, Appl. Phys. Lett., vol. 92, p. 192104,
2008.
J. M. Lee, I. T. Cho, J. H. Lee, W. S. Cheong, C. S. Hwang, and H. I. Kwon, Appl. Phys. Lett.,
vol. 94, p, 222112, 2009.
T. M. Pan, J. D. Lee, and W. W. Yeh, J. Appl. Phys., vol. 101, p. 024110, 2007.
X. Cao, X. Li, X. Gao, W. Yu, X. Liu, Y. Zhang, L. Chen, and X. Cheng, J. Appl. Phys., vol.
106, p. 073723, 2009.
T. M. Pan and T. W. Wu, IEEE Electron Device Lett., vol. 29, p. 353, 2008.
Y. H. Lin, C.H. Chien, T. Y. Yang, and T. F. Lei, J. Electrochem. Soc., vol. 154, p. H619,
360
2007.
O. Engstrom, B. Raeissi, S. Hall, O. Buiu, M. C. Lemme, H. D. B. Gottlob, P. K. Hurley, and
K. Cherkaoui, Solid State Electron., vol. 51, p. 622, 2007.
M. Fanciulli and G. Scarel, Rare Earth Oxide Thin Film: Growth, Characterization, and
Applications, Springer, Berlin, 2007.
G. He and Z. Sun, High-k Gate Dielectrics for CMOS Technology, Wiley-VCH, Germany,
2012.
Y. Zhao, K. Kita, K. Kyuno, and A. Toriumi, Appl. Phys. Lett., vol. 89, p. 252905, 1984.
Y. Uwamino, Y. Ishizuka, and H. Yamatera, J. Electron Spectrosc. Relat. Phenom., vol. 34, p.
67, 1984.
Y. A. Teterin and A. Y. Teterin, Russ. Chem. Rev., vol. 71, p. 347, 2002.
Y. Zhao, M. Toyama, K. Kita, K. Kyuno, and A. Toriumi, Appl. Phys. Lett., vol. 88, p.
072904, 2006.
J. E. Dominguez, L. Fu, and X. Q. Pan, Appl. Phys. Lett., vol. 81 p. 5168, 2002.
S. M. Sze and K. K. Ng, Physics of Semiconductor Devices 3rd ed., John Wiley & Sons, 2007.
T. M. Pan, C. L. Chan, and T. W. Wu, IEEE Electron Device Lett., vol. 30, p. 39, 2009.
361
SICASE-779
Green Bikeway Planning for Sustainable Environment in Scenic Area
Yao-Tsung Ko*, Ming-Shih Chen, Pei-Ling Wu
Tunghai University *[email protected]
Abstract
Under the thought of sustainable development, how to reduce the environmental impact for
creating a healthy and natural eco-tourism is an important issue for every country. In recent
years due to the rapid growth of the population and vehicles, cause the traffic congestion, noise
and air pollution problems increasingly serious, especially in scenic area. Meanwhile, the
environmental awareness rising some green environmentalists begin to advocate the concept of
green transportation and low-carbon tourism. The bike is the most non-pollution and
energy-saving for our environment among all vehicles.This paper focuses on the green bikeway
planning and building for sustainable environment in scenic area. The National Scenic Area of
Sun-Moon Lake in Taiwan is taken as a research field. The main research methods and tools
include Field, Survey, Analytic Hierarchy Process (AHP), Geographic Information Systems
(GIS) and Google Earth software. Field Survey and AHP expert questionnaire are adopted to
obtain the relevant evaluation factors and criteria weights as inputing parameters in GIS syatem.
Finally, this study creates five best planning of the green bikeway network for Sun-Moon Lake
eco-tourism by GIS system and Google Earth calculation. We hope that the results of this study
can not only provide advices to the administrator of Sun-Moon Lake scenic area for planning
eco-tourism environment and creating sustainable development in the future, but also can make
the concept of low-carbon tourism truly become a concrete contribution to our environment.
Keywords: green bikeway, sustainable environment, eco-tourism, AHP, GIS
362
SICASE-765
Phthalide, a New Naturally Occurring Bioactive Compound from
Trichoderma Harzianum WuHQ75 against Chilli Anthracnose
Chotika Jeerapong, Worrapong Phupong*
School of Science, Walailak University, Thailand *[email protected]
Worrapong Phupong*
School of Science, Walailak University, Thailand
Warin Intana
School of Agricultural Technology, Walailak University, Thailand
Patoomratana Tuchinda
Department of Chemistry and Center of Excellence for Innovation in Chemistry (PERCH-CIC),
Faculty of Science, Mahidol University, Thailand
Sopana Wongthong
Department of Chemistry, Faculty of Science and Technology,
Nakhon Si Thammarath Rajabhat University, Thailand
Abstract
Phthalide (1) and tyrosol (2) were isolated from the broth culture of Trichoderma harzianum
WuHQ75. Their structures were elucidated by extensive analysis of spectroscopic data and
comparison with literature data. Phthalide (1) has been reported herein for the first time as a
new naturally occurring compound. Tyrosol displayed the growth inhibition against
Colletotrichum gloeosporioides and Colletotrichum capsici with MIC of 64 μg/mL and
phthalide showed the growth inhibition against C. gloeosporioides and C. capsici with 64 and
>256 μg/mL, respectively.
Keyword: Trichoderma harzianum, chilli anthracnose, phthalide, tyrosol
363
1. Introduction
Chilli is more important to the economy, including the consumption in the country as well as
export. It is used as spices and natural health product. However, it is almost damaged by
anthracnose disease caused by C. gloeosporioides and C. capsici[1] leading to a decrease in
chilli production. In previous studies, the use of biocontrol such as Trichoderma species can
manage this disease and reduce the residues in the environment[2,3]. An antibiotic is a
mechanism[4] of Trichoderma species to protect plant pathogenic fungi.
Preliminary of our research group, the crude extract of T. harzianum WuHQ75 showed the
potent growth inhibition against C. gloeosporioides and C. capsici with MIC of 32 μg/mL.
Therefore, we interested to isolate and elucidate the structure of bioactive metabolites from this
strain.
2. Main Body
2.1 Method
T. harzianum WuHQ75 was isolated from soil collected in Nakorn Si Thammarat province,
Thailand. It was grown on potato dextrose agar (PDA) for 7 days. The discs (5 mm) of growing
colonies were cut from the margin of each plate and inoculated into Erlenmeyer flask
containing 250 mL of potato dextrose broth (PDB). The stationary cultures were incubated for
27 days at room temperature. The cultures were filtrated through filter paper (Whatman No. 1)
to obtain culture filtrate. The combine culture filtrate (20 L) was exhaustively extracted with
ethyl acetate (EtOAc) and the organic layer was dried over Na2SO4 followed by evaporation
under reduced pressure to yield the extract (6.90 g). The crude extract was separated by column
chromatography on silica gel, eluting with ethyl acetatehexane and methanolethyl acetate
gradients to afford seven separated fractions (F1-F7).
Fraction F4 (506.7 mg) was separated by column chromatography on Sephadex LH 20, eluting
with 50% methanol-dichloromethane to provide 6 subfractions (A1-A6). Subfraction A6 (24.5
mg) was further separated by column chromatography, eluting with 100% dichloromethane to
give compound 1 (5.4 mg) as a white solid.
Fraction F6 (1.592 g) was separated by column chromatography on Sephadex LH 20, eluting
with 50% methanol-dichloromethane to provide 4 subfractions (B1-B4). Subfraction B3 (227.8
mg) was further separated by column chromatography on silica gel, eluting with 40% ethyl
acetate-hexane and further purified by column chromatography on silica gel, eluting with 20%
acetone-hexane to give compound 2 (145.4 mg) as a white solid.
The microdilution method was applied to analyze antifungal activity of pure compound by
using 96-well microliter plates[5]. The 2-fold dilution of samples, spore suspension and
364
RPMI-1640 medium were added into each well to give final concentration of sample as 256-2
(µg/mL) and this test was performed in quadruplicate. The mixture was incubated for 48 h at
room temperature. 0.18% Resazurin (10 µL) was added to each well and further incubated for
3-4 h. Amphotericin B (10 µg/mL) and 1% DMSO were used as a positive control and control,
respectively. A result was indicated by a color change from blue to pink. The lowest
concentration at which the unchange color, was the MIC (minimum inhibitory concentration)
2.2 Results and Discussion
The investigation of the crude extract led to the isolation of a new naturally occurring
compound, phthalide (1, 5.4 mg), and tyrosol (2, 145.4 mg) which was previously reported
from T. harzianum[6] and T. brevicompactum[7]. Their structures were elucidated by analysis
of spectroscopic data ( UV, IR, MS, 1D and 2D NMR) and comprarison with literature.
Tyrosol displayed the growth inhibition of C. gloeosporiodes and C. capsici with MIC of 64
μg/mL whereas phthalide displayed the growth inhibition of C. gloeosporiodes and C. capsici
with MIC of 64 and >256 μg/ml, respectively.
2.3 Acknowledgments and Legal Responsibility
We would like to thank the Utilization of Natural Products Research Unit and Tropical Fruit
Research Unit at Walailak University for the facilities. This work was supported by a Walailak
University Fund.
3. References
[1] P. P. Than, H. Prihastuti, S. Phoulivong, P. W. J. Taylor, and K. D. Hyde, Chilli
anthracnose disease caused by Colletotrichum species., Journal of Zhejiang University.
Science. B, Vol. 9, no. 10, 2008, pp. 764–778.
[2] A. Isolation and F. Biocontrol, Trichoderma asperellum , Identified as a Novel Fungal
Biocontrol Agent for the Control of Plant Pathogen, 2012.
[3] B. Hajieghrari, M. Torabi-giglou, and M. R. Mohammadi, Biological potantial of some
Iranian Trichoderma isolates in the control of soil borne plant pathogenic fungi, African
Journal of Biotechnology, Vol. 7, no. 8, 2008, pp. 967–972.
[4] F. Vinale, K. Sivasithamparam, E. L. Ghisalberti, R. Marra, S. L. Woo, and M. Lorito,
Trichoderma-plant-pathogen interactions, Soil Biology and Biochemistry, Vol. 40, 2008,
pp. 1–10.
[5] N. D. Anaruma, F. L. Schmidt, M. C. T. Duarte, G. M. Figueira, C. Delarmelina, E. A.
Benato, and A. Sartoratto, Control of Colletotrichum gloeosporioides (penz.) Sacc. In
365
yellow passion fruit using Cymbopogon citratus essential oil, Brazilian Journal of
Microbiology, Vol. 41, no. 1, 2010, pp. 66–73.
[6] P. K. Tarus, C. C. Lang‟at-Thoruwa, A. W. Wanyonyi, and S. C. Chhabra, Bioactive
metabolites from Trichoderma harzianum and Trichoderma longibrachiatum, Bulletin of
the Chemical Society of Ethiopia, Vol. 17, no. 2, 2003, pp. 185–190.
[7] A. Tijerino, R. E. Cardoza, J. Moraga, M. G. Malmierca, F. Vicente, J. Aleu, I. G. Collado,
S. Gutiérrez, E. Monte, and R. Hermosa, Overexpression of the trichodiene synthase gene
tri5 increases trichodermin production and antimicrobial activity in Trichoderma
brevicompactum., Fungal genetics and biology, Vol. 48, no. 3, 2011, pp. 285–96.
366
SICASE-823
Construction and Sequencing Analysis of scFv Antibody Fragment Derived
from Monoclonal Antibody against Norfloxacin (Nor155)
Jirawat Mala
Program in Biotechnology, Faculty of Science, Chulalongkorn University, Thailand
Tanapat Palaga
Department of Microbiology, Faculty of Science, Chulalongkorn University, Thailand
Songchan Puthong, Kittinan Komolpis, Sarintip Sooksai*
Institute of Biotechnology and Genetic Engineering, Chulalongkorn University, Thailand *[email protected]
Yoshinobu Kaneko
Yeast Genetic Resources Lab, Graduate School of Engineering, Osaka University, Japan
Abstract
Norfloxacin belong to fluoroquinolone antibiotics which have been approved for treatment in
animals. However, their residues in animal products can pose adverse side effects to consumer.
Therefore, the detection of these residues in different food matrices could be concerned. In this
study, a single chain variable fragment (scFv) that recognizes norfloxacin antibiotic was
designed and constructed. The cDNA was synthesized from total RNA of hybridoma cells
against norfloxacin (Nor155). Genes encoding VH and VL regions of monoclonal antibody
against norfloxacin (Nor155) were amplified and sequenced. DNA fragments of VH and VL
regions were approximately 402 bp and 363 bp, respectively. The scFv of Nor155 was
constructed by adding (G4S)3 as linker between VH and VL regions and subcloned into pPICZαA,
an expression vector of Pichia pastoris. The obtained recombinant expression plasmid named
pJM01. The sequence of pJM01 expression plasmid was confirmed by sequencing analysis. The
complementarity determining region (CDR) I, II, and III of VH and VL were specified in the
deduced amino acid sequence. Sequence comparison indicates that the scFv was a homolog of
the single chain antibody 12G9 with 78% of degree homology. Taken together, the results
obtained reveal a potential information of amino acid sequence that will be useful for further
modification and production of scFv antibody in methylotrophic yeast P. pastoris.
Key words: monoclonal antibody, norfloxacin, scFv, sequencing analysis, CDRs
367
Introduction
Norfloxacin has been reported as an antibiotic which is used to treat urinary bacterial infection in
animal [1, 2, 3]. However, the extensive abuse of this antibiotic has caused severe food safety
problems. Meanwhile, some research indicates that low-level doses of antibiotic for long periods
could result in bacteria resistance [1, 4, 5). The concern of antibiotic residue in animal products,
which may cause antibiotic resistance in pathogens, maximum residue limits (MRLs) for several
antibiotics have been established in many countries to protect consumers [6]. In the case of
norfloxacin, the MRLs were set depending on the types of targeting tissue [7]. In order to ensure
human food safety and the entire ecosystem security, various chromatography methods have
been developed for the determination of norfloxacin in different food matrixes [1, 8-10].
However, these instrumental methods are time consuming and costly, and sample preparations
are demanding. During the last two decades, various immunoassay methods have been
developed to detect fluoroquinolone (FQs) based on polyclonal antibody (PAb) and monoclonal
antibody (MAb) [ 6, 11-13]. Compared with instrumental analysis methods, immunoassay
methods have advantages of rapid, sensitive and high throughput characteristics. It is known that,
ELISA is the most suitable method for rapid screening large members of samples due to its
simplicity rapid, sensitive and inexpensive [14]. In ELISA, the most important component is
antibody which binds specific to the desired drug residues. To produce large quantities of
antibodies, many hybridoma cell cultivation systems have been investigated. However, PAbs
sometimes experience nonspecific reactivity and MAb preparation is extremely complicated,
highly expensive and depends on expertise [14], and the preparation of high quality antibodies is
still a bottleneck issue when establishing immunoassay methods [1]. Therefore, an alternative
production system is required. A promising method is a fermentation of antibody-producing
yeast.
Single-chain fragment variable (scFv) is the smallest unit of immunoglobulin (Ig) molecule and
still remains their function in antigen-binding activities. The structure of scFv consists of
variable regions of heavy (VH) and light (VL) chain, which are jointed together by a flexible
peptide linker [ 15-17] such as (Gly4Ser)3 [18] that can be easily expressed in functional form,
allowing protein engineering to improve the properties of scFv such as increase of affinity and
alteration of specificity [19]. The order of the domains can be either VH-linker-VL or
VL-linker-VH and both orientations have been applied [19]. This structure remains the original
specificity and full monovalent binding of the intact parent Ab [20].
To date, scFv antibodies have been successfully isolated and displayed as fragments in various
expression systems such as mammalian cell, bacteria (ref), plant (ref), insect cells (ref), and also
yeast (ref). However, the production of heterologous eukaryotic proteins, yeasts is frequently
used as the best-studied host for heterologous gene expression through genetic engineering [21].
368
During the last two decades, the methylotrophic yeast, Pichia pastoris has become popular and
successful host for expression of recombinant proteins [16, 22-24].
The advantage is that the target proteins can be expressed as secretory forms [25]. Moreover, the
production of target proteins can be increased by high-cell-density fermentation. As a yeast cell,
P. pastoris, can express proteins for clinic application without contamination by endotoxins [26].
Furthermore, they have the ability to perform many of the post-translational modifications
usually performed in higher eukaryotes e.g. correct folding, disulfide bond formation, O- and
N-linked glycosylation and processing of signal sequences [24, 26-29].
The Institute of Biotechnology and Genetic Engineering (IBGE), Chulalongkorn University,
Thailand was recently achieved in the production of MAb against norfloxacin [Personal contact
to Ms. Songchan Puthong]. However, the production of norfloxacin-specific recombinant
antibody and the expression of scFv fragments using yeast expression system have not been
reported. In this study, gene encoding for VH and VL of MAb against norfloxacin (Nor155) was
identified. The scFv antibody was designed and constructed for cloning and expression in the
methylotrophic yeast, P. pastoris in the future.
Materials and Methods
Strains, Plasmids, Culture Medium and Reagents
Pichia pastoris strain GS115 (Mut+), Escherichia coli strain TOP10F‟, pPICZαA expression
vector and Zeocin™ were purchased from Invitrogen (USA). E. coli TOP10F‟ was used for all
plasmid constructions. The growth medium, Luria-B ertani (LB) medium, used in shake ‼ask
experiments was consisted of 5 g/L, yeast extract, 10g/L tryptone peptone and 10 g/L NaCl and
10 mg/L ampicillin. All medium components except ampicillin were sterilized by autoclaving
together at 121°C for 15 min. Ampicillin was sterilized by 0.22 mm ※ltration and added to the
medium immediately prior to inoculation. Other molecular biological reagents were purchased
from Takara (Japan) or Invitrogen (USA).
Preparation of First-Stand cDNA
Total RNA extraction from 5×106 hybridoma cells against norfloxacin (Nor155) using a
NucleoSpin® RNA II (Macherey-Nagel) was carried out according to the manufacturer‟s
instruction. First strand cDNA coding for the variable heavy and light chains was synthesized
from the total RNA extract (approximately 1 µg RNA) by using a first-strand cDNA synthesis kit
(Fermentas).
Construction of scFv Antibody Gene
Here, the scFv antibody gene was mainly constructed by PCR amplification and
ligation-reaction (Fig. 1A) followed by:
369
(i) Amplification of VH and VL. The 1st cDNA fragments encoding the variable heavy chain (VH)
were amplified by using VHFwMH1 as forward primers and VHRwIgG1 as a reverse primer
while those encoding the variable light chian (VL) were amplified by using VLFwMk as a
forward primer and VLRwKc as a reverse primer (Table 1). Each PCR reaction contains: 2 µL of
1st cDNA, 20 pmol of 5‟ and 3‟ primers, 5 µL of 2.5 mM dNTPs, 5 µL of Taq polymerse buffer,
and 0.5 µL of i-Taq DNA polymerase (Intron Biotechnology). The final volume was brought to
50 µL with nuclease-free water. Cycling conditions were: initial melt at 94OC for 3 min followed
by 30 cycles of three-step program (94 O
C, 1 min; 45OC, 1 min; and 72
OC, 2 min). The reactions
were then held at 72OC for 10 min and cooled to 4
OC [30]. Then, the amplicons were
electrophoresed through 1% low-melting point agarose gel and visualized by staining with
ethidium bromide. Each of the amplified DNA fragments corresponding to the predicted size
was excised from the gel and was purified by using QIAquick PCR Purification Kit (QIAGEN).
The purified fragments then were ligated individually into the pGEM® T-easy vector (Promega)
and subsequently introduced into competent E. coli TOP10F‟. The blue/white screening was
used to select the positive clones on LB agar plates containing 100 µg/mL ampicillin, 50 µL
X-gal (20mg/ml) and 30 µL IPTG (0.1 M) respectively.
(ii) Construction of VH and VL gene with flexible polypeptide linkers. A 15 amino acid linker
encoding (Gly4Ser)3 was introduced behind the VH domain and in front of the VL domain to
produce a full-length scFv which was oriented as VH-(Gly4Ser)3–VL (Fig. 1A) by using specific
designed primer N1-N4 based on the VH and VL gene sequences as shown in the table 1. After
PCR, the VH-linker and linker-VL amplicons were separated and purified by 1% (w/v) agarose
gel electrophoresis. Afterward, the synthesized genes were cloned separately to the pGEM®
T-easy vector and was selected as above. Plasmid of the ampicillin-positive clones were
extracted and analyzed by restriction enzyme digestions and electrophoresis. The plasmids
containing the fragments of the specified sizes were confirmed by DNA sequencing. Finally, the
transformant clones with the accurate sequence were then used to generate the expression
plasmid.
(iii) Construction of recombinant scFv antibody expression plasmid. The VH-linker and
linker-VL clones having the BamHI–BglII restriction enzyme sites and EcoRI-KpnI linearized
pPICZαA vector were assembled by ligation reaction and then introduced into E. coli TOP10F‟
to generate the pPICZαA-scFv construct named pJM01 (Fig. 1B). The pJM01 was cloned in
frame with the α-factor secretion signal sequence open reading frame of the pPICZαA vector.
370
Verification of pJM01 Plasmid
To confirm the ligation-reaction of constructed pJM01 plasmid, the following three methods
were carried out.
(i) PCR analysis of positive transformants. Five zeocin-resistance transformant clones were
analyzed by using PCR with designed primers TJ01F/TJ01R, N1/N4 and PAOX1F/TJ03R,
respectively (Fig. 2A) based on the pJM01 sequences map as shown in the table1, which could
371
detect the ligation of the scFv antibody gene in to the pPICZαA, while pPICZαA vector was used
as negative control to compare with the positive clones (Fig. 2B). To obtain the DNA template,
plasmid DNA was extracted. PCR components and conditions were as follows: AmpliTaq Gold®
360 Master Mix (AB Applied Biosystem), 0.5µM of each primer, 1 cycle of 95oC – 5 min, 30
cycles of three-step program (95oC, 15 sec; 55
oC, 30 sec and 72
oC, 1.30). The reactions were
then held at 72oC for 7 min and cooled until used at 4
oC.
(ii) Restriction enzyme digestion. The positive clones which shown the specified size according
to the expected map were tested with EcoRI and KpnI, respectively.
(iii) Sequencing. To consider the scFv antibody gene was cloned in frame with the α-factor
secretion signal sequence. The positive clones, which shown the expected size from (i) and (ii)
methods were sequenced with BigDye® terminal sequencing kit and ABI PRISM
® 3130xl
Genetic Analyzer (Applied Biosystems, USA) using four oligonucleotide primers, TJ03F, TJ04F,
TJ05F, and TJ06R which were designed based on pJM01 sequence as shown in the table 1.
Identification of Complementarity Determining Region (CDRs) and Framework
Sequences (FR)
The obtained sequence was aligned by searching against the Genbank database for sequence
homology analysis. The CDR and framework sequences were identified by comparison with
IgG1 and light chain sequences database form GenBANK (NCBI) or published data.
372
Results and Discussion
In this study, we designed, constructed, and sequenced the pJM01 (scFv antibody) plasmid
derived from MAb against norfloxacin (Nor155) and will be transformed into methylotrophic
yeast P. pastorris for production of recombinant antibody in the future.
Monoclonal antibody against norfloxacin clone 155, a clone that showed the highest efficiency
of IC50 and lowest of %cross-reactivity, was selected. Type and subtype of heavy chain of MAb
against norfloxacin (Nor155) was performed with mouse monoclonal antibody isotyping reagent
(Sigma-Aldrich) and shown that it was IgG1 (data not shown). Type of light chain of MAb
Nor155 was performed by using rat anti-mouse kappa light chain-HRP (Serotec) and it was
identified as kappa light chain (data not shown) [Personal contact to Ms. Songchan Puthong].
Construction of scFv Antibody
Total RNA was extracted from MAb nor155, and the first chain cDNA was synthesized by
RT-PCR kit using total RNA as template. The VH and VL genes were amplified separately by
PCR using the 1st cDNA template, and the lengths of VH and VL DNA fragments were
approximately 402 bp and 363 bp, respectively (GenBank, accession no. KJ623260). These
fragments were used to construct the VH-linker and linker-VL fragments by adding the flexible
linker (Gly4Ser)3. Then, these fragments were cloned into pGEM® T-easy vector and identified
by restriction enzyme digestion, electrophoretic analysis and DNA sequencing (data not
shown).The lengths of VH-linker and linker-VL DNA fragments were approximately 426 bp and
380 bp, respectively. Afterward, the recombinant scFv (VH-(Gly4Ser)3–VL) fragment (800 bp)
was assembled with EcoRI/KpnI linearized pPICZαA to generate the recombinant plasmid
named pJM01, which expected size was 4354 bp (Fig. 1B).
Verification of pJM01 Plasmid
To improve and facilitate the expression of the antibody fragment, VH-linker and linker-VL DNA
fragments were inserted into the open reading frame of the P.pastoris expression vector
pPICZαA under the regulation of AOX1 promoter using the EcoRI (bp 1203-1213) and KpnI
(1241-1246) restriction sites. At the C-terminal, the constructs were fused to a c-myc epitope tag
and a hexahistidine-tag. Colonies were screened for insertion using PCR analysis and restriction
cleavage reactions. In Fig.3A, the plasmid of five zeocin-positive transformants was extracted.
The presence of scFv antibody in those plasmids was tested by PCR technique with N1 and N4
primers as shown in the table 1. These primers were designed for up-stream amplification of VH
and down-stream of VL gene. The result reveals the 800 bp of PCR product according to the right
size of predicted restriction map of linearized pJM01plasmid (fig. 2A). This information
indicates that all of five positive transformants have had a scFv antibody gene inserted into the
pJM01 expression plasmid. In addition, we concerned of the nucleotide sequence of scFv
373
antibody gene when it was cloned in frame with α-factor secretion signal in the pPICZαA vector
and translated to amino acid should be accurate.
For this reason, the plasmid of zeocin-positive transformant No.1 was selected to digest with
several restriction enzymes. In accordance with expected map, EcoRI, KpnI, and PmeI were used.
These enzymes showed a 1 position of enzymatic digestion that mean after digestion completely
and through electrophoresis the digested plasmid could show the linearized plasmid. Fig.3B
showed the result bands corresponding to the correct size of 4354 bp plasmid (lane 2, 3, and 5),
while lane 4 represented the using of EcoRI and KpnI digest the plasmid in the same time. It was
shown the 800 bp of scFv antibody gene when compared with the DNA marker. Moreover, we
also confirmed the plasmid of clone No.1 by PCR with primers as descripted in the table 1. Fig.
3C exhibits the positive bands approximately 1461 bp of PCR product (lane 2) by using
TJ01F/TJ01R primers, approximately 800 bp of PCR product (lane 4) by using N1/N4 primers
and approximately 1350 bp by using 5‟AOX1/3‟AOX1 primers (fig. 3D). Whereas, the using
these primers with pPICZαA vector as control, it showed approximately 700 bp (fig.3C, lane 1),
589 bp (fig.3D, lane1) and cannot detect any band by using N1/N4 primers (fig.3C, lane 3),
respectively.
Sequencing Analysis
The pJM01 plasmid of zeocin-positive transformant No.1 was sequenced with specified primers
(Table 1). The result shows the nucleotide sequence of scFv antibody gene, 800 nucleotides were
reached. Predicted translation of this sequence encoding 266 amino acids with a predicted
374
molecular weight of 32.67 kDa including a flexible amino acid linker of (Gly4Ser)3 as shown in
Fig. 4.
Identification of Complementarity Determining Region (CDRs) and Framework
Sequences (FR)
Complementarity determining regions (CDRs) are part of the variable chains in
immunoglobulins (antibodies), where these molecules bind to their specific antigen. As the most
variable parts of the molecules, CDRs are crucial to the diversity of antigen specificities. Within
the variable domain, CDR I, CDR II, and CDR III are found in the variable region of a
polypeptide chain, and CDR3 is the most variable [31-32].
The complementarity determining regions (CDRs) I, II, and III of the variable region of both
heavy chain (VH) and light chain (VL) were identified according to the method provided by
Kabat [33]. The predicted of CDRI, CDR II and CDR III of VH domain was located at position
29-37, 52-60, and 99-112 in the primary protein sequences, while the predicted of CDRI, CDR II
and CDR III of VL domain was specified at position 170-180, 193-200, and 235-243 as depicted
in Fig. 4, respectively.
The deduced amino acid sequence of the scFv antibody gene was aligned by searching against
GenBank database. Through the blastp program, the result showed the similarity of scFv
sequence with others which have been submitted in. Over 50 different scFv antibodies
homologous sequence were found (data not shown). The first-five of these was selected and
375
demonstrated. Each NCBI GenBank accession number is a sequence identified by blastp as a
highly homologous sequence to the query putative scFv antibody sequence. Comparative
analysis of the predicted amino acid sequence with the five known scFv sequences is shown in
Fig. 5 and summarized in the Table.2 including the accession number, definition, amount of
amino acid, max score, total score, query cover, e-value and identity, respectively. As a result,
the scFv antibody was found most closely related to the single chain antibody 12G9, partial (78%
identity) followed by anti-porcine sialoadhesin scFv antibody, partial (75% identity) and both of
scFv B4 anti-pectinase antibody, partial and anti-isoketal-adduct single chain variable fragment,
partial (74% identity), and scFv antibody, partial from Mus musculus (71% identity),
respectively.
376
Conclusion
We have successfully designed, constructed, and sequenced a single-chain variable fragment
(scFv) antibody derived from the monoclonal antibody against norfloxacin clone Nor155. The
CDRs of amino acids that contribute the binding activity of scFv antibody were predicted.
Various results have been supported that the pJM01 expression plasmid containing scFv
antibody gene.
Acknowledgments
This research was partly supported by the 90th
Anniversary of Chulalongkorn University Fund
(Ratchadaphiseksomphot Endowment Fund) and Overseas Research Experience Scholarship for
Graduate Student, Graduate School, Chulalongkorn University. The authors would like to thank
Prof. Harashima Satoshi, Graduate School of Engineering, Osaka University, Japan for helpful
discussion.
377
References
[1] Chen M., Wen K., Tao X., Xie J, Wang L., Li Y., Ding S. and Jiang H. Cloning, expression,
purification and characterization of a bispecific single-chain diabody against
fluoroquinolones and sulfonamides in Escherichia coli, Protein Expression and Purification.
100 (2014) 19-25.
[2] Donoghue D.J. Antibiotic residues in poutry tissues and eggs: Human Health concerns?
Poultry Science, 82 (2003) 618-621.
[3] Zhang H.T., Jiang J.Q., Wang Z.L., Chang X.Y., Liu X.Y.,Wang S.H., Zhao K. and Chen J.S.
development of an indirect competitive ELISA for simultaneous detection of enrofloxacin
and ciprofloxacin, Journal of Zhejiang University Science B (Biomedicine &
Biotechnology), 12-11(2011) 884-891.
[4] D.A. Franco, J. Webb, C.E. Taylor, Antibiotic and sulfonamide residues in meat:
implications for human health, J. Food Prot. 53 (1990) 178–185
[5] K. Zhu, J. Li, Z. Wang, H. Jiang, R.C. Beier, F. Xu, J. Shen, S. Ding, Simultaneous detection
of multiple chemical residues in milk using broad-specificity antibodies in a hybrid
immunosorbent assay, Biosens. Bioelectron. 26 (2011) 2716–2719.
[6] Manaspong C., Wongphanit P., Palaga T., Puthong S., Sooksai S. and Komolpis K.
Production and characterization of a monoclonal antibody against enrofloxacin, Journal of
Microbiology and Biotechnology, 23-1 (2013) 69-75.
[7] The European Agency for the Evaluation of Medicinal Products. Enrofloxacin (extension to
all food producing species): Summary report, 2002. EMEA/MRL/820/02-Final
[8] Jiang J., Zhang H., Liu J., Li J. and Wang Z. Development and optimization of an indirect
competitive ELISA for detection of norfloxacin residue in chicken liver, 2nd
IPCBEE,
6(2011) 460-463.
[9] M.I. Lopez, J.S. Pettis, I.B. Smith, P.-S. Chu, Multiclass determination and confirmation of
antibiotic residues in honey using LC–MS/MS, J. Agric. Food Chem. 56 (2008) 1553–
1559.
[10] G. Stubbings, T. Bigwood, The development and validation of a multiclass liquid
chromatography tandem mass spectrometry (LC–MS/MS) procedure for the determination
of veterinary drug residues in animal tissue using a QuEChERS (QUick, Easy, CHeap,
Effective, Rugged and Safe) approach, Anal. Chim. Acta 637 (2009) 68–78.
[11] Jinqing J., Haitang Z., Yonghua Q., Yongyao C. and Ziliang W. Production and
characterization of monoclonal antibodies against norfloxacin, Procedia Environmental
Sciences, 8 (2011) 529- 535.
[12] A.-C. Huet, C. Charlier, S.A. Tittlemier, G. Singh, S. Benrejeb, P. Delahaut, Simultaneous
determination of (fluoro) quinolone antibiotics in kidney, marine products, eggs, and
muscle by enzyme-linked immunosorbent assay (ELISA), J. Agric. Food Chem. 54 (2006)
2822–2827.
378
[13] D.G. Pinacho, F. Sanchez-Baeza, M.P. Marco, Molecular modeling assisted hapten design
to produce broad selectivity antibodies for fluoroquinolone antibiotics, Anal. Chem. 84
(2012) 4527–4534.
[14] Wang, S., Du, X., Huang, Y., Lin, D., Hart, P. and Wang, Z. Detection of deoxynivalenol
based on a single-chain fragment variable of the antideoxynivalenol antibody, Federation
of European Microbiological Societies Microbiology Letters. 272(2007) 214-219.
[15] Freya M.F., Javier E.V., Marta A., Leonardo C.H., Hanssel B., Ivonne R., Arturo G.,
Alberto C. and Jorge V. G. Very high expression of an anti-carcinoembryonic antigen
single chain Fv antibody fragment in the yeast Pichia pastoris, Journal of Biotechnology,
76(2000) 157-163.
[16] Cornelia M., Patrick S., Kurt B.H., Roman K. and Reto A.S. Production of functionalized
single-chain Fv antibody fragments binding to the ED-B domain of the B-isoform of
fibronectin in Pichia pastoris, Protein Expression and Purification, 21(2001) 156-164.
[17] Rozbeh J., Patrik H., Macro P., and Birgitta E.S. Construction of divalent anti-keratin 8
single-chian antibodies (sc(Fv)2), expression in Pichia pastoris and their reactivity with
multicellular, Journal of Immunogical Methods, 364(2011) 65-76.
[18] Louise M.E., Amanda J.T., Philip J.B. and Peter J.N. Expression of an anti-CD33
single-chain antibody by Pichia pastoris, Journal of Immunological methods, 305(2005)
135-151.
[19] Ahmad Z.A., Yeap S.K., Ali A.M., Ho W.Y., Alitheen N.B.M. and Hamid M. Review
article: scFv antibody; Principles and clinical application, Clinical and developmental
immunology, (2012) 1-15.
[20] Parker S.A., Diaz I.L., Anderson K.A. and Batt C.A. Design, production, and
characterization of a single-chain variable fragment (scFv) derived from the prostate
specific membrane antigen (PSMA) monoclonal antibody J591, Protein Expression and
Puriffication, 89 (2013) 136-145.
[21] Jing H., Li X., Hongde H., Xiaojin D. and Lingyun J. High-level expression of
Staphylococcus Protein A in Pichia pastoris and purification and characterization of the
recombinant protein, Protein Expression and purification, 90(2013) 178-185.
[22] Chong S., Xiao W.Y., Fei L. and Yan X. Impact of gene dosage on the production of lipase
from Rhizopus chinensis CCTCC M201021 in Pichia pastoris, Appl. Biochem Biotechnol,
169(2013) 1160-1172.
[23] Mehmet I., Dinesh A., Jayanta S. and Michael M.M. Enhancement of protein secretion in
Pichia pastoris by overexpression of protein disulfide isomerase, Biotechnology and
Bioengineering, 93(2006) 771-778.
[24] Cereghino J.L. and Cregg J.M. Heterologous protein expression in the methylotrophic yeast
Pichia pastoris, FEMS microbial, Rev. 24 (2000) 45- 66.
[25] Issaly N, Solsona O. and Joudrier P. Optimization of the wheat puroindoline – a
production in Pichia pastoris. Journal Appl. Microbiol, 90 (2001) 397-406.
379
[26] Diego G.N., Matias N.R., Martin B., Gaston E.O. and Miguel A.G. Cloning, expression
and optimized production in a bioreactor of bovine chymosin B in Pichia (Komagataella)
pastoris under AOX1 promoter, Protein Expression and Purification, 92(2013) 235-244.
[27] Morka K., Pietruszka J. and Berstenhorst S.M. Comparative expression of lipase CAL-A
in the yeasts Saccharomyces cerevisiae, Kluyveromyces lactis and Hansenula polymorpha
to investigate a possible host influence, Journal of Biotechnology, 191(2014) 176-186.
[28] Diego G.N., Martin B., Matias R. and Miguel A.G. Bioprocess and downstream
optimization of recombinant bovine chymosin B in Pichia pastoris under
methanol-inducible AOX1 promoter, Protein Expression and purification, 104(2014)
85-91.
[29] Yang Y.L., Kai X.Z., Ai K.H., Dan N.L., Li Z.C. and Shu D.X. High-level expression and
characterization of a thermostable xylanase mutant from Trichoderma reesei in Pichia
pastoris, Protein Expression and Purification, accepted vesion xxx(2014) xx-xx.
[30] Wang, Z., Raifu, M., Howard, M., Smith, L., Hansen, D., Goldsby, R. and Ratner, D.
Universal PCR amplification of mouse immunoglobulin gene variable regions: the design
of degenerate primers and an assessment of the effect of DNA polymerase 3‟ to
5‟exonuclease activity, Journal of Immunology Methods, 233(2000) 167-177.
[31] Klaus S.L., Henrik S., Niels H.H. and Mingyue H. A novel DNAseq program for enhanced
analysis of illumine GA II data: a case study on antibody complementarity-determining
regions, New Biotechnology, 29(2012) 271-278.
[32] Jooho R., Sung J.B., Youngsil S., Minjae K., Jae H.L., Songmi K., Yuno L., Keun W.L., Jin
K.K. and Myung H.K. Generation of a chickenized catalytic anti-nucleic acid antibody by
complementarity-determining region grafting, molecular Immunology, 63(2015)
513-520.
[33] Kabat E., Wu T. and Perry H. (1991). Sequences of Proteins of Immunological Interest,
5th edn. US Department of Health and Human Services, National Institutes of Health,
Bethesda, MD
380
SICASE-743
Properties and Microstructure of Cobalt-Based Alloy Layer by Laser
Cladding
Yao-Sheng Yang*, Yin-Fang Wei, Bo-Yong Chen, Yeou-Yih Tsai
Department of Electronic Engineering, Kao Yuan University, Taiwan *[email protected]
In the study, the Stellite1 cobalt-based alloy coatings were produced by laser cladding on a JIS
S45C medium carbon steel substrate with a transversal flow CO2 laser. Microstructure and
wear properties of the cladding coatings were studied by the XRD, SEM and wear testing. The
results show that the cladding coatings have better metallurgical bonding to the substrate, the
microstructure of cladding coatings consisted of equiaxed, columnar and cellular dendrite, and
the eutectic structure was little difference in composition of arm spacing among these
microstructures. The coating of the Stellite1 cobalt-based alloy process has hypoeutectic
microstructure comprising γ-austenite and M7C3 carbide. Wear resistance of the cladding
coatings has certain corresponding with the cladding parameter; the wear resistance reaches the
maximum value when laser power is 2KW.
Key Words :carbon steel、laser cladding、wear resistance、hypoeutectic