P

Se

Za

b

a

ARRA

KSPAP

1

tedOEtra2Tct&F

mogGp

h1

ARTICLE IN PRESSG ModelARTIC-640; No. of Pages 7

Particuology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Particuology

journa l homepage: www.e lsev ier .com/ locate /par t ic

ubmicron particle sizing by aerodynamic dynamic focusing andlectrical charge measurement

hongchao Tana,∗, Raheleh Givehchia, Alena Saprykinab

Department of Mechanical & Mechatronics Engineering, University of Waterloo, Ontario, CanadaDepartment of Mechanical and Manufacturing Engineering, University of Calgary, Alberta, Canada

r t i c l e i n f o

rticle history:eceived 5 September 2013eceived in revised form 14 January 2014

a b s t r a c t

Principles of a novel submicron particle sizing technology are first introduced followed by experimentalvalidation. The sizing was accomplished by coupling aerodynamic particle focusing and maximum ionmeasurement. Experimental results showed that the prototype could detect particle sizes down to 40 nm

Particuology (2015) 18: 105-111.

ccepted 16 January 2014

eywords:ubmicron particlearticle sizingerodynamic focusing

in diameter. Comparison between the prototype and a scanning mobility particle sizer using identicalpolydisperse particles showed that the measurements agreed well for the tested particles.

© 2014 Published by Elsevier B.V. on behalf of Chinese Society of Particuology and Institute of ProcessEngineering, Chinese Academy of Sciences.

mmtrscWoa(ftcnmal

aF

article charging

. Introduction

The majority of airborne particles are smaller than 1 microme-er in diameter in various environments (Whitby, 1978; Weimert al., 2009; Yao et al., 2013). Some of these particles are pro-uced by both condensation and fuel combustion processes (Jung,h, Noh, Ji, & Kim, 2006; Dijk, Gopal, & Scheepers, 2011; Wallace,mmerich, & Howard-Reed, 2004). Increasing amount of nanopar-icles with recent rapid development of nanotechnology has alsoaised the concerns of potential negative impact on human healthnd the environment (Chow & Watson, 2007; Mauderly & Chow,008; Oberdörster, Stone, & Donaldson, 2007; Wang & Pui, 2011).here are significant indications that submicron particles, espe-ially nanosized ones, are more toxic than larger ones because ofheir small sizes and large surface areas (Oberdörster, Gelein, Ferin,

Weiss, 1995; Oberdörster et al., 2007; Roduner, 2006; Shin, Pui,issan, Neumann, & Trampe, 2007).

Various technologies have been developed for the measure-ent of particle size distribution (PSD); a comprehensive overview

f micron and nanosized particle measurement methods was

Please cite this article in press as: Tan, Z., et al. Submicron particlemeasurement. Particuology (2014), http://dx.doi.org/10.1016/j.partic.2

iven by Sabbagh-Kupelwieser, Maisser, and Szymanski (2011).enerally speaking, optical measuring methods work for aerosolarticles larger than 100 nm; smaller ones are detected by electrical

∗ Corresponding author. Tel.: +1 519 888 4567x38718.E-mail address: tanz@uwaterloo.ca (Z. Tan).

cpcrLeeM

ttp://dx.doi.org/10.1016/j.partic.2014.01.002674-2001/© 2014 Published by Elsevier B.V. on behalf of Chinese Society of Particuology

ethods. An example of the latter is the well-known scanningobility particle sizer (SMPS). It is able to measure particle size dis-

ribution from 2 nm to a few microns with a high efficiency and highesolution with duration of between 3 and 5 min for a single mea-urement (Wang & Flagan, 1990). SMPS consists of an electrostaticlassifier with a differential mobility analyser (DMA) (Knutson &

hitby, 1975) and a condensation particle counter (CPC). Particlesf different sizes are classified based on their electrical mobilitynd they are counted by CPC using the light scattering techniqueKulkarni, Baron, & Willeke, 2011). Watson et al. (2011) comparedour scanning mobility particle sizers in field and showed thathey performed differently because of the differences in particleharging efficiency, CPC counting efficiency, diffusion losses, andon-ideal DMA transfer functions. In addition, SMPS usually costsore than other alternatives and it mainly aims at applications inlaboratory environment because of flammable and expendable

iquid used in operation and a vibration-free environment.There are also a few alternative technologies for submicron

irborne particle measurement. SMPS+E substitutes CPC with aaraday cup and an electrometer. It measures particle number con-entrations in the size range of 0.8 nm to 1.1 �m. Electrical lowressure impactor (ELPI) employs aerosol corona charging and aascade low pressure impactor to measure particle sizes in theange of 30 nm to 10 �m in real time (Keskinen, Pietarinen, &

sizing by aerodynamic dynamic focusing and electrical charge014.01.002

ehtimäki, 1992; Marjamäki, Keskinen, Chen, & Pui, 2000). How-ver, ELPI showed poor size resolution for sub-30 nm particles,specially at high concentration (Ouf & Sillon, 2009; Virtanen,arjamäki, Ristimäki, & Keskinen, 2001). Aerosol particle mass

and Institute of Process Engineering, Chinese Academy of Sciences.

ARTICLE IN PRESSG ModelPARTIC-640; No. of Pages 7

2 ology

arwCdhitt&

gtenatot(toK

(ccywibtmsectt

2

iMoaMH

dmtahtoRtcfi2

d

wr

C

wm1

d

c

d

wfirspied

didFtoct

g

�

wacopS

d

bpmt

n

Z. Tan et al. / Particu

nalyzer (APM) classifies particles according to their mass to chargeatio. APM consists of two rotating coaxial cylindrical electrodeshich rotate at the same angular velocity (Ehara, Hagwood, &oakley, 1996; Olfert & Collings, 2005). It could detect particlesown to about 50 nm. Fast integrated mobility spectrometer (FIMS)as been developed for measuring the particle size distribution

n the range of 15–170 nm for low particle number concentra-ion, and there was good agreement with SMPS for particles largerhan 20 nm in diameter (Kulkarni & Wang, 2006; Olfert, Kulkarni,

Wang, 2008).While it is tedious and challenging to list all existing technolo-

ies and devices for submicron particle sizing, it is safe to concludehat they can size the airborne submicron particles with differ-nt resolutions and accuracies; their lower limits vary from a fewanometers to tens of nanometers. Each of them has advantagesnd disadvantages in terms of accuracy, resolution, respondingime and cost as well as application environment. In general, nonef the conventional technologies allow for the optimal combina-ion of high efficiency, easy access, quick response and low costFriedlander & Pui, 2004). Therefore, there is still a need to improvehe performances of existing technologies or to develop alternativenes for the measurement of particle size distribution (Sabbagh-upelwieser et al., 2011).

The work in this paper is inspired by that of Tan and Wexler2007), who introduced another approach to sub-300 nm parti-le sizing by aerodynamic particle focusing coupled with diffusiveharging. However, there was an error in the theoretical anal-sis part leading to wrong algorithm. The Coulomb constantas missing in Eq. (5) in that paper. More importantly, the

nstrument based on their algorithm relies on experimental cali-ration against another standard instrument (e.g. SMPS) to obtainwo key coefficients for data processing. As such, the instru-

ent could not perform particle sizing independently. This papertarts with an alternative theoretical analysis followed by anxperimental evaluation of a prototype with state-of-the-art dataollection systems. In addition, error sources are identified andaken into consideration in the performance evaluation of the pro-otype.

. Theoretical

Considerable work has focused on aerodynamic particle focus-ng since it was first developed by Liu, Ziemann, Kittelson, and

cMurry (1995a, 1995b). In this type of device with multiple stagesf focusing orifices, particles of decreasing diameter are focusedlong the axis with downstream the focusing orifice (Vidal-de-iguel & de la Mora, 2012). Numerical works are also listed byeadrick, Schrader, and Michelsen (2013).

In our approach, the sizes of particles are determined by aero-ynamic particle focusing and the particle numbers by measuringaximum charge. In order to determine the maximum charge,

hese particles are passed through a charger where sufficientmount of ions are present. Aerodynamic particle focusing (APF)as been developed and employed for accurate particle charac-erization (Wexler & Johnston, 2001). A well designed focusingrifice could isolate particles down to a few nanometers (Phares,hoads, & Wexler, 2002). Large particles cross the center line dueo great inertia and small ones with low inertia do not cross theenter line. The optimally focused particle size by a focusing ori-ce can be described as (Mallina, Wexler, Rhoads, & Johnston,

Please cite this article in press as: Tan, Z., et al. Submicron particlemeasurement. Particuology (2014), http://dx.doi.org/10.1016/j.partic.2

000)

∗p = (dm

p )2

2�Cc, (1)

wfcm

xxx (2014) xxx–xxx

here � is the mean free path of the carrier gas; Cunningham cor-ection factor Cc can be calculated by

c = 1 + 1.66Kn, (2)

here Kn is the Knudsen number and the diameter of the opti-um focused particle d∗

p is described as (Dahneke, Hoover, & Cheng,982; Mallina et al., 2000; Middha & Wexler, 2003)

∗p =√

(1.657�)2 + (dmp )2 − 1.657�. (3)

The maximum size of the focused particles dmp is a function of

ritical Stokes number Stk∗:

mp = 18�df Stk∗

�pvf, (4)

here �p is the density of particle, � is the viscosity of gas, df is theocusing orifice diameter and vf is the average velocity in the focus-ng orifice exit plane. Gas velocity in the focusing orifice shouldeach sonic speed for enabling aerodynamic focusing of a certain-ize particle. The critical Stokes number Stk∗ is based on the gasroperties at the orifice throat. The value of Stk∗ has been numer-

cally determined to be between 1 and 2 (Mallina et al., 2000) andxperimentally determined to be around 2 (Phares et al., 2002). Itoes not depend on the pressure of the carrier gas.

The analysis above indicates that optimally focused particle size∗p depends on the mean free path of the carrier gas and the max-mum focused diameter dm

p . dmp varies with the focusing orifice

iameter, the gas viscosity and the gas velocity through the orifice.rom an engineering practice point of view, it is more challengingo alter the focusing orifice diameter or the gas velocity through therifice than to control the size of the optimally focused particles byhanging the gas mean free path, which can be done by adjustinghe upstream gas pressure.

The mean free path of the carrier gas can be calculated followingas molecular dynamics (Bird, 1994):

= �

P√

2M/�RT, (5)

here R is the universal gas constant, T is temperature, P is pressure,nd M is the molecular weight of the carrier gas. This equation indi-ates that the mean free path � of the carrier gas before the focusingrifice can be expressed in terms of its corresponding macro-scalearameters, temperature and pressure, which can be measured.ubstituting Eqs. (4) and (5) into Eq. (3) gives

∗p =

√√√√( 1.657�

P√

2M/�RT

)2

+(

18�df Stk∗

�pvf

)− 1.657�

P√

2M/�RT. (6)

A particle can also be charged with ions and the maximum num-er of ions nm that a particle can be charged is directly related to thearticle size. Therefore, the particle size can be correlated with itsaximum number of ions loaded. The maximum number of ions

hat a particle can be charged is described as (White, 1951)

m = d∗pkbT

2e2Keln

(1 + d∗

pKevi�e2Nit

2kbT

)+ E(d∗

p)2

4eKe

(3εr

2 + εr

) (t

t + �

),

(7)

sizing by aerodynamic dynamic focusing and electrical charge014.01.002

here the first term on the right hand side (RHS) accounts for dif-usion charging mechanism and the second term of RHS is for fieldharging mechanism. For submicron particles, diffusion chargingechanism is dominant and the last term in Eq. (7) can be ignored

ARTICLE IN PRESSG ModelPARTIC-640; No. of Pages 7

Z. Tan et al. / Particuology xxx (2014) xxx–xxx 3

tion o

is

n

wcCec&

imtdcdpm

N

wmflnbo

N

pa

ipt

3

3

fdpmwmt(c(bdiotfalo

3

Fig. 1. Experimental setup for evalua

n most cases. Then the maximum number of ions charged onto aubmicron particle can be described as

m = d∗pkbT

2e2Keln

(1 + d∗

pKevi�e2Nit

2kbT

), (8)

here vi is the mean thermal speed of ions (240 m/s at standardondition), kb is Boltzmann constant (1.38 × 10−23 J/K), Ke is aoulomb’s constant of proportionality (1/(4�ε0) = 9 × 109 Nm2/C2),

is the elementary unit of charge (1.6 × 10−19 C), and Ni is ion con-entrations and it is typically in the order of 5 × 1014 ions/m3 (Tan

Wexler, 2007).In practice, it is nearly unlikely to count the actual number of

ons charged on particles. This information is commonly deter-ined by measuring the current of a stream of moving particles

hat carry these ions. The number of ions carried by the aero-ynamically focused particles can be correlated to the electricalurrent produced by these ions, which can be detected by a Fara-ay cup connected with an electrometer. The measured current isroportional to the number of particles that reach the cup and theaximum charge on the focused particle, I = N∗

pQnme, and it gives

∗p = I

Qnme, (9)

here N∗p is the particle number concentration (#/m3) of the opti-

ally focused particles by focusing orifice and Q (m3/s) is the gasow rate in the focusing orifice. This equation indicates that theumber concentration of particles that were optimally focused cane determined by measuring the electrical current and the numberf ions (calculated by Eq. (8)). Substituting Eq. (8) into Eq. (9) gives

∗p = I

(d∗k TQ/2eK ) ln(1 + (d∗K v �e2N t/2k T)). (10)

Please cite this article in press as: Tan, Z., et al. Submicron particle

measurement. Particuology (2014), http://dx.doi.org/10.1016/j.partic.2

p b e p e i i b

Most of the parameters in this equation can be determined. Thearticle diameter is determined by aerodynamic particle focusingbove Eq. (6). The others are constants or operating parameters. The

pwdi

f the performance of the prototype.

on number concentration and charging time depend on the specificarticle charger. The device-specific information is introduced inhe next section.

. Experimental

.1. Experimental setup

Fig. 1 shows the schematic diagram of the experimental setupor evaluating the performance of a laboratory prototype that wasesigned following the principles introduced above. Feed aerosolarticles were generated using a constant output atomizer (TSIodel 3076) with a concentration 0.1 g/L of NaCl in fresh distilledater. Diffusion dryer (TSI Model 3062) was used to remove theoisture from the produced aerosols. Feed aerosols were then neu-

ralized by a neutralizer that employed a radioactive source 210PONRD, Model P-2031) and the size distribution of these feed parti-les were first characterized by a scanning mobility particle sizerSMPS, Model 3936, TSI Inc.). Then these particles were chargedy passing through a home-made corona charger before beingirected into a low pressure channel through a pressure reduc-

ng orifice. Finally they are passed through a focusing orifice wherenly particles with certain sizes are selectively focused to the aper-ure of a Faraday cup, which was connected to an electrometeror electric current measurement. Other particles were pumpedway by a vacuum pump. The size distribution was then calcu-ated using the equations introduced in last section. Specificationsf each component are elaborated in the sections that follow.

.2. Corona charger

Fig. 2 shows the schematic diagram of the home-made needle-

sizing by aerodynamic dynamic focusing and electrical charge014.01.002

late corona charger. The active electrode of the corona chargeras a gold needle with a sharp cone-shaped tip. The top of the nee-le was connected to a stainless steel rod mounted into a ceramic

solator. The internal diameter of the bottom stainless steel plate

ARTICLE IN PRESSG ModelPARTIC-640; No. of Pages 7

4 Z. Tan et al. / Particuology

(tipe

detinpeacdc

t

wtnswTu

ewtwsboi

nI

E

wnc

cpfiaorbtvctw

3

0uoib6a

tsKatTo

3

F6chTFcwo

tvfvtnecorona charger acquired the maximum charges for voltages greater

Fig. 2. Schematic diagram of the needle-plate corona charger.

Dc) was 37.5 mm. The distance between the plate and the needleip (ı) was 12.7 mm, allowing air to pass through the corona whileons were polarizing the surrounding air molecules. A high-voltageower supply (Glassman high voltage Model PS/EL30R01.5) wasmployed to supply an electric potential in the range of ±30 kV.

There have been various particle charging technologies andevices available. In this work the home-made corona charger wasmployed to show the feasibility of the principles introduced inheoretical section. However, it does not guarantee the best charg-ng performance because optimum particle charging technology isot the main scope of this paper. Preliminary trials showed thatositive charging was preferred in favor of the low particle lossffect in the corona charger (Saprykina, 2009). The supplied volt-ge must reach a level that is high enough to ensure maximumharging with least error input and to avoid the electrical break-own of the air. The corresponding charging time of this specificharger can be determined by

= Dc

Vg= 2ıD2

c

Q, (11)

here Vg is the air flow speed and ı is the distance betweenhe plate and the needle tip. It has been speculated that theeedle-plate corona charger like this could produce extra nano-ized particles (Tan & Wexler, 2007), and this may result in an errorhen it is employed to measure the sizes of nanosized particles.

herefore, particle generation in this charger was characterized tonderstand the errors resulted from the charger itself.

The experimental setup for the characterization of particle gen-ration in the corona charger is different from that in Fig. 1. Airas first passed through a high-efficiency particulate air (HEPA) fil-

er before entering the corona charger. The particle concentrationsith and without power supply were measured using the same

canning mobility particle sizer (SMPS, Model 3936). The difference

Please cite this article in press as: Tan, Z., et al. Submicron particlemeasurement. Particuology (2014), http://dx.doi.org/10.1016/j.partic.2

etween the particle concentration measurements with and with-ut power supply was considered as the level of particle generation,f any. In presentation of the result, the electric field at the tip of the

tcp

xxx (2014) xxx–xxx

eedle is used instead of the voltage applied to the corona charger.t can be calculated as follows (Florkowska & Wlodek, 1993).

(�) = 2U

ln(4ı/rt)1

2� + rt − (�2/ı), (12)

here � is the distance from the needle tip (� = 0 at the tip of theeedle), U is the applied voltage, rt is the radius of curvature of theorona needle (rt = 0.25 mm for the setup herein).

Another potential error source is the loss of particles in theorona charger by electrostatic precipitation effect. To quantify thearticle loss effect of the corona charger, if any, a correction coef-cient (k) was experimentally determined to take this factor intoccount. This correction coefficient is defined as the ratio of numberf particles entering and leaving the charger. To quantify this cor-ection factor, polydispersed sodium chloride particles producedy the same aerosol generation system (see Fig. 1) were passedhrough the corona charger at a voltage of 0 kV and at the requiredoltages for maximum charging of the particles. The correctionoefficient is used to relate the measured data by the new pro-otype to the actual particle number concentration independentlyithout employing another device.

.3. Aerodynamic particle focusing

A group of pressure reducing orifices (0.026, 0.022, 0.02, 0.018,.016, 0.014, 0.012, 0.01, 0.08, and 0.06 inches in diameter) weresed to adjust the pressure of the carrier gas before the focusingrifice. Stainless steel tube was used as a low pressure region. Sim-lar to what was done by Tan and Wexler (2007), the distance (L)etween the pressure reducing orifice and the focusing one was00 mm. This region is deemed long enough to obtain a well-mixederosol flow pattern.

The diameter of the focusing orifice was fixed at 3 mm and itshickness was 0.5 mm. Air flow reaches sonic speed if the down-tream pressure reaches 52.8% of the upstream pressure (John &eith, 2006). The focal point of the optimum sized particles wast the aperture of the Faraday cup. Particles of different sizes fromhe selected optimum ones are pumped away by a vacuum pump.herefore, theoretically only particles of the certain size are pickedut from the suspending gas and focused into the Faraday cup.

.4. Current measurement and maximum charging

Downstream the focusing orifice, the current coming to thearaday cup was measured by an electrometer (Keithley Model430) connected via coaxial cable to Faraday cup to measureurrents down to 10−17 A. The electrical current was generated byitting the charged particles into the metal grid of the Faraday cup.he distance between the focusing orifice and the aperture of thearaday cup was 14 mm. Trial tests showed that the focused parti-les were not observed when the distance was shorter than 14 mm,hile for the distances larger than 14 mm there was significant loss

f the focused particles.Preliminary tests were also conducted to make sure that par-

icles reached their maximum charging. Some basic experimentolt-ampere data was recorded to determine the required voltageor maximum particle charging. Results showed that the suppliedoltage of approximately 4 kV in the corona charger was requiredo satisfy the maximum particle charging. Current measured didot increase any further with higher voltage, which means a great-st ion concentration. As a result, the particles exiting through the

sizing by aerodynamic dynamic focusing and electrical charge014.01.002

han 4 kV. Since the particles reached maximum charging, Eq. (8)an be used to determine the number concentration of focusedarticles.

ARTICLE IN PRESSG ModelPARTIC-640; No. of Pages 7

Z. Tan et al. / Particuology

F

4

4

pvgwFern1s

cafest1dsptip

tgrSv3ng(iig

bTinasan

4

mvtcicwdHecc

sbfield in the corona charger to be intensified near the gold needle;however, its electrical intensity decreased by closing the plate. As

ig. 3. Generation of particles in the filtered air with a volume flow rate of 1.4 L/min.

. Results and discussion

.1. Checking nanoparticle generation in the corona charger

The numbers of particles generated by the corona charger byassing the filtered air were measured by SMPS at different appliedoltages or electric filed. It was found that nanoparticles wereenerated by corona charging section when the supplied voltageas higher than a certain threshold value for the corona charger.

ig. 3 shows the average size distribution of the nanoparticles gen-rated in the corona charger operated at different voltages. Theesults indicate that the threshold voltage value for generatinganoparticles was about 7 kV corresponding to an electric field of0,538 kV/m. The SMPS could not detect any particles when theupplied voltages were less than the mentioned threshold voltages.

A possible explanation of the generation of nanoparticles by theorona charger is that a corona charger has enough energy to initi-te gas-phase chemical reactions in the charger region, such as theormation of ozone from oxygen, which may lead to particle gen-ration (Romay, Liu, & Pui, 1994). The sputtering of metal from theurface or the erosion of the electrodes could also cause particleso be generated (Liu, Pui, Kinstley, & Fisher, 1987; Romay et al.,994). Unnecessarily high corona voltage intensifies the coronaischarges, which facilitate the expansion of the active coronao that it occupies additional space between the needle and thelate (Chang, Lawless, & Yamamoto, 1991). Therefore, the higherhe voltage, the greater the erosion or chemical reactions, which

Please cite this article in press as: Tan, Z., et al. Submicron particle

measurement. Particuology (2014), http://dx.doi.org/10.1016/j.partic.2

ncrease the size and number of the particles generated around theeaks.

as

Fig. 4. Size distributions of the sodium chloride aerosols

xxx (2014) xxx–xxx 5

The positive corona produces nanoparticles with peak diame-ers of approximately similar sizes, which increase slightly withreater electric fields. The number of nanoparticles generatedises dramatically with an increased electric field. Hernandez-ierra, Alguacil, and Alonso (2003) reported that the requiredoltages for a corona charger to generate ions are approximately.2 kV for a positive corona charger. For greater voltages, the ionumber concentration is approximately constant because of thereater electrical mobility and the electrostatic deposition lossHernandez-Sierra et al., 2003). Thus, in the present study, increas-ng the supplied voltages would not affect the concentration ofons generated; it changed only the concentration of nanoparticlesenerated.

Nanoparticles have special aggregation characteristics causedy the known van der Waals forces because of their small size.he attractive interactions between the nanoparticles are increas-ng when the particles are moving toward each other due to theiratural tendency (Hakim, Portman, Casper, & Weimer, 2005). Thislso helps explain the uncertainty and less accuracy in the mea-urement of the size distribution in this range in the work of Tannd Wexler (2007). Nonetheless, the generation of particles wasegligible when the operating voltage was below 7 kV.

.2. Checking nanoparticle loss in the corona charger

Fig. 4 shows the size distribution of sodium chloride aerosoleasured downstream of the corona charger at different charging

oltages. The feeding size distribution of NaCl aerosol correspondso the measurement at the applied voltage of zero in the coronaharger. This concentration includes those NaCl particles pass-ng through the charger as well as those generated in the coronaharger. The measured particle number concentration decreasedith the increase of supplied voltages because of the electrostaticeposition loss of nanoparticles (Hernandez-Sierra et al., 2003;uang & Alonso, 2011). As a result, although inconclusive, thisxperiment demonstrated for higher applied voltages, the coronaharger tended to filter more particles and so decrease the numberoncentration of feed aerosol.

In this case, the needle discharge electrode was the constantupplied voltage, but the plate was grounded. The differenceetween the voltages of the needle and plates caused the electrical

sizing by aerodynamic dynamic focusing and electrical charge014.01.002

result, the corona charger acted as an electrostatic precipitatoro that larger charged particles deposited electrostatically on the

at different voltages applied to the corona charger.

ARTICLE IN PRESSG ModelPARTIC-640; No. of Pages 7

6 Z. Tan et al. / Particuology

Fig. 5. Particle size cumulative distributions measured using the prototype (solidd

pac

ma4Td

k

i

4

mfptWdfltmo

tpomltrttta2

4

secdtsawfthdpnsm

asrwtItvi

5

icmpgn

A

SF

R

B

C

C

D

E

F

F

F

Hakim, L. F., Portman, J. L., Casper, M. D., & Weimer, A. W. (2005). Aggregationbehavior of nanoparticles in fluidized beds. Powder Technology, 160(3), 149–160.

ots) and the SMPS (dashed line).

late and lost. Increasing the voltage intensified the electrical fieldnd the filtration effect that caused more particles to be lost in theorona charger.

Considering the effect of aerosol generation above 7 kV, theaximum charge required at >4 kV and the particle loss effect

bove, the voltage supplied to the corona charger should not exceed kV for the setup. Therefore, the experiments were run at 4 kV.he corresponding correction coefficient for particle loss was thenetermined for this voltage only as follows.

= 1.99(dp − 3.74)[

�

2− arctan(0.49dp − 6.89)

]− 1.57. (13)

This equation shows that, for the positive corona charger, theres no significant particles loss for the particles larger than 60 nm.

.3. Size distribution comparison with SMPS

Fig. 5 shows the comparison between SMPS and the measure-ent using the new prototype, by considering different coefficients

or the charging efficiency and particle loss. The concentration ofarticles correlates with the results of SMPS for particles downo 40 nm in diameter more efficiently than the study of Tan and

exler (2007). The measured concentration for particles withiameter less than 70 nm is lower than the concentration in SMPSor particles of the same diameter. The error analysis was calcu-ated in Fig. 5, based on the errors of the parameters included inhe calculation formula for focusing diameter, ion concentration,

easured current, temperature and pressure before the focusingrifice and particle density.

There are other error sources in the measurement of prototypeo explain the difference between the SMPS and prototype. Smallerarticles are charged less effectively than the larger ones becausef their small cross section area (Flagan & Seinfeld, 2012). Further-ore, a lot of particles can be pumped away due to the boundary

ayer within the orifice (Phares et al., 2002). In addition, some par-icles could be lost due to deposition on the surface of the pressureeducing section because of the high electrical mobility preventinghem from having charges. Besides, there are some particles withhe optimal size passing close to the edge of the focusing orifice

Please cite this article in press as: Tan, Z., et al. Submicron particlemeasurement. Particuology (2014), http://dx.doi.org/10.1016/j.partic.2

hat are not transmitted since they are not able to reach the samecceleration as those in the center of the flow (Middha & Wexler,003).

H

xxx (2014) xxx–xxx

.4. Limitations and recommendations

There are several significant limitations associated with thistudy. First of all, alternative particle charging should be consid-red in future prototypes. Although the current home-made coronaharger served its purpose for proof-of-principle in this paper, itid complicate the calculation by requiring an extra correction fac-or to correct the particle loss in the charger. In addition, it washown that corona charger generates nanoparticles at higher volt-ge. Although it did not affect the current result because the voltageas chosen to be 4 kV which was lower than the minimum voltage

or nanosized particle generation, it may impact the operation ofhis device in the future work. Furthermore, this corona chargeras the very small active volume for ionization around the nee-le compared to the whole charge volume between the needle andlate (Yehia & Mizuno, 2012). As a result the corona charger wasot helpful for maximum charging the nanoparticles especially themaller ones and needs to be replaced by other alternative chargingethods.The prototype was not tested for particles smaller than 40 nm,

nd it is not safe to conclude whether the technology works forub-40 nm particles. This is limited by the availability of pressureeducing orifice with a smaller size. In addition, the prototypeorked by manual switching between different orifice nozzles. It

ook some time in the order of 10 s to complete a full measurement.n future product development, the more nozzles for finer resolu-ion, the longer measurement time will be. An automatic scanningalve will greatly improve the measurement frequency. More works needed to improve the lower limit of this device.

. Conclusions

Theoretical analysis and experimental evaluation showed thatt is feasible to measure the size distribution of airborne submi-ron particles by aerodynamic particle focusing and particle chargeeasurement. Despite the errors in the experimental parts, the

rototype could measure particles in the range of 40–300 nm in aood agreement with SMPS. More research shall be done by alter-ative particle charging and improved lower size detection limit.

cknowledgments

The authors acknowledge the financial support from Naturalcience and Engineering Research Council (NSERC) and Canadaoundation for Innovation (CFI).

eferences

ird, G. A. (1994). Molecular gas dynamics and the direct simulation of gas flows. NewYork: Oxford University Press.

hang, J.-S., Lawless, P. A., & Yamamoto, T. (1991). Corona discharge processes. IEEETransactions on Plasma Science, 19(6), 1152–1166.

how, J. C., & Watson, J. G. (2007). Survey of measurement and composition ofultrafine particles. Aerosol and Air Quality Research, 7, 121–173.

ahneke, B., Hoover, J., & Cheng, Y. S. (1982). Similarity theory for aerosol beams.Journal of Colloid and Interface Science, 87(1), 167–179.

hara, K., Hagwood, C., & Coakley, K. J. (1996). Novel method to classify aerosol par-ticles according to their mass-to-charge ratio – aerosol particle mass analyser.Journal of Aerosol Science, 27(2), 217–234.

lagan, R. C., & Seinfeld, J. H. (2012). Fundamentals of air pollution engineering. NewJersey: Dover Publications, Incorporated.

lorkowska, B., & Wlodek, R. (1993). Pulse height analysis of partial discharge in air.IEEE Transactions on Electrical Insulation, 28(6), 932–940.

riedlander, S. K., & Pui, D. Y. H. (2004). Emerging issues in nanoparticle aerosolscience and technology. Journal of Nanoparticle Research, 6(2–3), 313–320.

sizing by aerodynamic dynamic focusing and electrical charge014.01.002

eadrick, J. M., Schrader, P. E., & Michelsen, H. A. (2013). Radial-profile anddivergence measurements of combustion-generated soot focused by anaerodynamic-lens system. Journal of Aerosol Science, 58, 158–170.

ARTICLE IN PRESSG ModelPARTIC-640; No. of Pages 7

ology

H

H

JJ

K

K

K

K

L

L

L

M

M

M

M

O

O

O

O

O

P

R

R

S

S

S

T

V

V

V

W

W

W

W

W

W

W

W

Z. Tan et al. / Particu

ernandez-Sierra, A., Alguacil, F. J., & Alonso, M. (2003). Unipolar charging ofnanometer aerosol particles in a corona ionizer. Journal of Aerosol Science, 34(6),733–745.

uang, C. H., & Alonso, M. (2011). Nanoparticle electrostatic loss within corona nee-dle charger during particle-charging process. Journal of Nanoparticle Research,13(1), 175–184.

ohn, J. E. A., & Keith, T. (2006). Gas dynamics (3rd ed.). New Jersey: Prentice Hall.ung, J. H., Oh, H. C., Noh, H. S., Ji, J. H., & Kim, S. S. (2006). Metal nanoparticle gen-

eration using a small ceramic heater with a local heating area. Journal of AerosolScience, 37(12), 1662–1670.

eskinen, J., Pietarinen, K., & Lehtimäki, M. (1992). Electrical low pressure impactor.Journal of Aerosol Science, 23(4), 353–360.

nutson, E. O., & Whitby, K. T. (1975). Aerosol classification by electric mobility:Apparatus, theory, and applications. Journal of Aerosol Science, 6(6), 443–451.

ulkarni, P., Baron, P. A., & Willeke, K. (2011). Aerosol measurement: Principles, tech-niques and applications. New York: John Wiley & Sons.

ulkarni, P., & Wang, J. (2006). New fast integrated mobility spectrometer for real-time measurement of aerosol size distribution – I: Concept and theory. Journalof Aerosol Science, 37(10), 1303–1325.

iu, P., Ziemann, P. J., Kittelson, D. B., & McMurry, P. H. (1995a). Generating particlebeams of controlled dimensions and divergence: I. Theory of particle motion inaerodynamic lenses and nozzle expansions. Aerosol Science and Technology, 22,293–313.

iu, P., Ziemann, P. J., Kittelson, D. B., & McMurry, P. H. (1995b). Generating particlebeams of controlled dimensions and divergence: II. Experimental evaluation ofparticle motion in aerodynamic lenses and nozzle expansions. Aerosol Scienceand Technology, 22, 314–324.

iu, B. Y. H., Pui, D. Y. H., Kinstley, W. O., & Fisher, W. G. (1987). Aerosol charg-ing and neutralization and electrostatic discharge in clean rooms. Journal ofEnvironmental Sciences, 30(2), 42–46.

allina, R. V., Wexler, A. S., Rhoads, K. P., & Johnston, M. V. (2000). High speedparticle beam generation: A dynamic focusing mechanism for selecting ultrafineparticles. Aerosol Science and Technology, 33(1–2), 87–104.

arjamäki, M., Keskinen, J., Chen, D.-R., & Pui, D. Y. H. (2000). Performance evalu-ation of the electrical low-pressure impactor (ELPI). Journal of Aerosol Science,31(2), 249–261.

auderly, J. L., & Chow, J. C. (2008). Health effects of organic aerosols. InhalationToxicology, 20(3), 257–288.

iddha, P., & Wexler, A. S. (2003). Particle focusing characteristics of sonic jets.Aerosol Science and Technology, 37(11), 907–915.

berdörster, G., Gelein, R. M., Ferin, J., & Weiss, B. (1995). Association of particulateair pollution and acute mortality: Involvement of ultrafine particles? InhalationToxicology, 7(1), 111–124.

berdörster, G., Stone, V., & Donaldson, K. (2007). Toxicology of nanoparticles: Ahistorical perspective. Nanotoxicology, 1(1), 2–25.

lfert, J. S., & Collings, N. (2005). New method for particle mass classification –The couette centrifugal particle mass analyzer. Journal of Aerosol Science, 36(11),1338–1352.

lfert, J. S., Kulkarni, P., & Wang, J. (2008). Measuring aerosol size distributionswith the fast integrated mobility spectrometer. Journal of Aerosol Science, 39(11),

Please cite this article in press as: Tan, Z., et al. Submicron particle

measurement. Particuology (2014), http://dx.doi.org/10.1016/j.partic.2

940–956.uf, F. X., & Sillon, P. (2009). Charging efficiency of the electrical low pressure

impactor’s corona charger: Influence of the fractal morphology of nanoparti-cle aggregates and uncertainty analysis of experimental results. Aerosol Scienceand Technology, 43(7), 685–698.

Y

Y

xxx (2014) xxx–xxx 7

hares, D. J., Rhoads, K. P., & Wexler, A. S. (2002). Performance of a single ultrafineparticle mass spectrometer. Aerosol Science and Technology, 36(5), 583–592.

oduner, E. (2006). Size matters: Why nanomaterials are different. Chemical SocietyReviews, 35(7), 583–592.

omay, F. J., Liu, B. Y. H., & Pui, D. Y. H. (1994). A sonic jet corona ionizer for elec-trostatic discharge and aerosol neutralization. Aerosol Science and Technology,20(1), 31–41.

abbagh-Kupelwieser, N., Maisser, A., & Szymanski, W. (2011). From micro- to nano-sized particles: Selected characterization methods and measurable parameters.Particuology, 9(3), 193–203.

aprykina, A. (2009). Airborne nanoparticle sizing by aerodynamic particle focusingand corona charging (Master of Science thesis). Alberta, Canada: University ofCalgary.

hin, W. G., Pui, D. Y. H., Fissan, H., Neumann, S., & Trampe, A. (2007). Calibrationand numerical simulation of nanoparticle surface area monitor (TSI model 3550NSAM). Journal of Nanoparticle Research, 9(1), 61–69.

an, Z. C., & Wexler, A. S. (2007). Fine particle counting with aerodynamicparticle focusing and corona charging. Atmospheric Environment, 41(25),5271–5279.

an Dijk, W. D., Gopal, S., & Scheepers, P. T. J. (2011). Nanoparticles in cigarettesmoke; real-time undiluted measurements by a scanning mobility particle sizer.Analytical and Bioanalytical Chemistry, 399(10), 3573–3578.

idal-de-Miguel, G., & de la Mora, J. F. (2012). Continuously converging multistagefocusing lenses to concentrate aerosols at high Reynolds numbers. Aerosol Sci-ence and Technology, 46, 287–296.

irtanen, A., Marjamäki, M., Ristimäki, J., & Keskinen, J. (2001). Fine particlelosses in electrical low-pressure impactor. Journal of Aerosol Science, 32(3),389–401.

allace, L. A., Emmerich, S. J., & Howard-Reed, C. (2004). Source strengths of ultra-fine and fine particles due to cooking with a gas stove. Environmental Scienceand Technology, 38(8), 2304–2311.

ang, J., & Pui, D. Y. H. (2011). Characterization, exposure measurement and con-trol for nanoscale particles in workplaces and on the road. Journal of Physics:Conference Series, 304(1), 012008.

ang, S. C., & Flagan, R. C. (1990). Scanning electrical mobility spectrometer. AerosolScience and Technology, 13(2), 230–240.

atson, J. G., Chow, J. C., Sodeman, D. A., Lowenthal, D. H., Chang, M. O., Park, K.,et al. (2011). Comparison of four scanning mobility particle sizers at the fresnosupersite. Particuology, 9(3), 204–209.

eimer, S., Mohr, C., Richter, R., Keller, J., Mohr, M., Prévôt, A. S. H., et al. (2009).Mobile measurements of aerosol number and volume size distributions in analpine valley: Influence of traffic versus wood burning. Atmospheric Environment,43(3), 624–630.

exler, A. S., & Johnston, M. V. (2001). Real-time single-particle analysis. In P.A. Baron, & K. Willeke (Eds.), Aerosol measurement: Principles, techniques andapplications (pp. 365–385). New York: Wiley-Interscience.

hitby, K. T. (1978). The physical characteristics of sulfate aerosols. AtmosphericEnvironment, 12, 135–159.

hite, H. J. (1951). Particle charging in electrostatic precipitation. Transactions ofthe American Institute of Electrical Engineers, 70(2), 1186–1191.

sizing by aerodynamic dynamic focusing and electrical charge014.01.002

ao, J., Liu, W., Chen, W., Wang, G., Zeng, Y., Huang, Y., et al. (2013). Effect of atmo-spheric parameters on fine particulate concentration in suburban shanghai.Particuology, 11(1), 48–54.

ehia, A., & Mizuno, A. (2012). Expectation of ozone generation in alternating currentcorona discharges. Physics of Plasmas, 19(3), 033513.