This may be the author’s version of a work that was submitted/acceptedfor publication in the following source:

Zare, Ali, Nabi, Md Nurun, Bodisco, Timothy, Hossain, Md Farhad, Rah-man, Md Mahmudur, Chu Van, Thuy, Ristovski, Zoran, & Brown, Richard(2017)Diesel engine emissions with oxygenated fuels: A comparative study intocold-start and hot-start operation.Journal of Cleaner Production, 162, pp. 997-1008.

This file was downloaded from: https://eprints.qut.edu.au/108014/

c© Consult author(s) regarding copyright matters

This work is covered by copyright. Unless the document is being made available under aCreative Commons Licence, you must assume that re-use is limited to personal use andthat permission from the copyright owner must be obtained for all other uses. If the docu-ment is available under a Creative Commons License (or other specified license) then referto the Licence for details of permitted re-use. It is a condition of access that users recog-nise and abide by the legal requirements associated with these rights. If you believe thatthis work infringes copyright please provide details by email to qut.copyright@qut.edu.au

License: Creative Commons: Attribution-Noncommercial-No DerivativeWorks 2.5

Notice: Please note that this document may not be the Version of Record(i.e. published version) of the work. Author manuscript versions (as Sub-mitted for peer review or as Accepted for publication after peer review) canbe identified by an absence of publisher branding and/or typeset appear-ance. If there is any doubt, please refer to the published source.

https://doi.org/10.1016/j.jclepro.2017.06.052

Accepted Manuscript

Diesel engine emissions with oxygenated fuels: A comparative study into cold-startand hot-start operation

Ali Zare, Md Nurun Nabi, Timothy A. Bodisco, Farhad M. Hossain, M.M. Rahman,Thuy Chu Van, Zoran D. Ristovski, Richard J. Brown

PII: S0959-6526(17)31222-2

DOI: 10.1016/j.jclepro.2017.06.052

Reference: JCLP 9802

To appear in: Journal of Cleaner Production

Received Date: 15 February 2017

Revised Date: 20 April 2017

Accepted Date: 7 June 2017

Please cite this article as: Zare A, Nabi MN, Bodisco TA, Hossain FM, Rahman MM, Van TC, RistovskiZD, Brown RJ, Diesel engine emissions with oxygenated fuels: A comparative study into cold-start andhot-start operation, Journal of Cleaner Production (2017), doi: 10.1016/j.jclepro.2017.06.052.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service toour customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, and alllegal disclaimers that apply to the journal pertain.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Cold-Start

Higher NOx (up to 13.9%)

Lower PN (up to 73%)

Lower PM2.5 (up to 89%)

Lower PM1 (up to 91%)

Hot-Start

vs.vs.

˟ Higher NOx (up to 125%)

˟ Higher PN (up to 17 times)

˟ Significant increase in

nucleation mode particles

Compared to hot-start, during

cold-start

• using diesel showed

lower NOx (up to 15.4%), PN (up

to 48%), PM1 (up to 44%) and

PM2.5 (up to 63%)

• using biodiesel showed

significantly higher NOx

(up to 94%), PN (up to 27

times), PM1 (up to 7.3 times)

and PM2.5 (up to 5 times)

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

1

Diesel engine emissions with oxygenated fuels: A comparative study into cold-start and 1

hot-start operation 2

3

Ali Zarea,*, Md Nurun Nabib, Timothy A. Bodiscod, Farhad M. Hossaina, M.M Rahmana,c, 4

Thuy Chu Vana,e, Zoran D. Ristovskia,c, Richard J. Brown 5

6

a Biofuel Engine Research Facility, Queensland University of Technology (QUT), QLD, 4000 7 Australia 8

b School of Engineering and Technology, Central Queensland University, WA 6000, Australia 9 c International Laboratory for Air Quality and Health, Queensland University of Technology 10

(QUT), QLD, 4000 Australia 11 d School of Engineering, Deakin University, VIC, 3216 Australia 12

e Vietnam Maritime University (VMU), Haiphong, Vietnam 13

14

Abstract 15

As biofuels are increasingly represented in the fuel market, the use of these oxygenated fuels 16

should be evaluated under various engine operating conditions, such as cold-start. However, 17

to-date quantification has been mostly done under hot-start engine operation. By using a 18

custom test designed for this study, a comparative investigation was performed on exhaust 19

emissions during cold- and hot-start with diesel and three oxygenated fuels based on waste 20

cooking biodiesel and triacetin. This study used a six-cylinder, turbocharged, after-cooled 21

diesel engine with a common rail injection system. The results during cold-start with diesel 22

showed lower NOx (up to 15.4%), PN (up to 48%), PM1 (up to 44%) and PM2.5 (up to 63%). 23

However, the oxygenated fuels during cold-start showed a significant increase in NOx (up to 24

94%), PN (up to 27 times), PM1 (up to 7.3 times) and PM2.5 (up to 5 times) relative to hot-25

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

2

start. The use of oxygenated fuels instead of diesel during hot-start decreased the PN, PM2.5 26

and PM1 (up to 91%) while, during cold-start, it only decreased PM1 and PM2.5 at some 27

engine operating modes and increased PN significantly up to 17 times. In both cold- and hot-28

start, the use of oxygenated fuels resulted in an increase in NOx emission. For cold-start this 29

was up to 125%, for hot-start it was up to 13.9%. In comparison with hot-start, the use of 30

oxygenated fuels during cold-start increased nucleation mode particles significantly, which 31

are harmful. This should be taken into consideration, since cold-start operation is an 32

inevitable part of the daily driving schedule for a significantly high portion of vehicles, 33

especially in cities. 34

Keywords: Cold-start; biodiesel; fuel oxygen content; PM; PN; particle size distribution. 35

36

1. Introduction 37

This study investigates the influence of oxygenated fuels on exhaust emissions during cold-38

start, with a comparison to hot-start. During cold-start, the engine temperature is sub-optimal 39

owing to lower temperatures of three interrelated thermal masses: the main engine block, the 40

engine coolant and the lubricant oil (Roberts, Brooks et al. 2014). Driving when the engine 41

temperature is lower than its optimal value influences engine performance and emissions 42

(Roberts, Brooks et al. 2014, Roy, Calder et al. 2016). For example, the low temperature of 43

the engine cylinder wall was reported to be a reason for higher emissions and fuel 44

consumption (Roberts, Brooks et al. 2014). A study by Cao (2007) demonstrated that a cold 45

engine block, which can cause incomplete combustion, can significantly affect emissions. 46

Additionally, the higher viscosity of the engine lubricant due to its low temperature increased 47

friction losses and decreased thermal efficiency (Roberts, Brooks et al. 2014). The friction 48

losses during cold-start increased up to 2.5 times, compared to when the engine is warmed up 49

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

3

(Will and Boretti 2011). To overcome high friction losses and to maintain brake power 50

output, more fuel must be injected into the cylinder during cold-start, regardless of the 51

ambient temperature (Will and Boretti 2011, Khair and Jääskeläinen 2013). A study showed a 52

13.5% increase in fuel consumption during cold-start compared to hot-start (Samhaber, 53

Wimmer et al. 2001). 54

Emissions during cold-start are a significant proportion of an engine’s total emissions 55

(Sakunthalai, Xu et al. 2014, Reiter and Kockelman 2016). Nam (2008) estimated that 56

particulate matter (PM) emissions during the cold-start phase of the LA92 Unified Driving 57

Cycle, Phase 1, can consist of up to 30% of total PM emission from that cycle. This is despite 58

Phase 1 only being a 12% and 21% proportion of the entire cycle distance and duration, 59

respectively. They also showed that PM emission from Phase 1 is 7.5 times higher than phase 60

3, which is a hot-start phase with the same driving schedule as Phase 1. Another report 61

showed that PM and hydrocarbons (HC) were several times higher during the first three 62

minutes of cold-start operation compared to the time the engine was warmed up; more than 63

40% of total emissions were related to those first three minutes (Bielaczyc, Merkisz et al. 64

2001). Lee et al. (2012) used the FTP test and demonstrated that nitrogen oxides (NOx), CO 65

and HC emitted from a cold engine, in comparison to a warmed up engine, can be up to two, 66

three and four times higher, respectively. 67

There are different definitions for cold-start in the literature. Reiter and Kockelman (2016) 68

reviewed some of them. Based on the literature, this study considers cold-start to be an 69

engine starting after an overnight soak (engine-off time) and persisting until the engine 70

coolant temperature reaches 70°C, inline with EU Directive 2012/46/EU. This investigation 71

also studies a condition in which the engine coolant temperature is above 70 °C but the 72

engine lubricant temperature is still sub-optimal. 73

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

4

A significant proportion of daily driving starts and finishes while the engine temperature is 74

still below its regular operating level (Reiter and Kockelman 2016). A study of the driving 75

patterns of 55 French cars, which included 1000 trips (71 000 km representing 1260 hours), 76

under real conditions showed that one third of the journeys were completed before the engine 77

coolant and lubricant exceeded 70°C (André 1991). It should be noted, however, that the 78

cold-start time and distance depend on emission species (André and Joumard 2005, Reiter 79

and Kockelman 2016). Andre and Joumard (2005) studied and modeled excess emission 80

during cold-start based on a survey of 39 European laboratories. The data they used were 81

obtained from 1766 vehicles and 35 941 measurements. They estimated an average distance 82

of 5.2 km for the cold-start distance, the average distance at which emissions (CO, CO2, HC 83

and NOx) stabilised, at an ambient temperature of 20 °C. It should be mentioned that most of 84

these reports are based on conventional petro-diesel. 85

Over the last few decades, biofuel has displaced some conventional fuels, but petro-diesel is 86

still the most common type of diesel fuel. Using biofuel instead of fossil fuels is relevant due 87

to the adverse health effects, environmental degradation and the impact on global warming 88

caused by fossil fuels. These negative points contributed the European Union’s directive to 89

increase the use of renewable biofuels to 10% by 2020 in order to offset fossil fuel usage (EU 90

Directive 2009/28/EC). The commitment to reduce fossil fuel use was reinforced when 195 91

countries agreed to limit global warming in the Paris agreement (December 2015, Paris 92

climate conference (COP21)). 93

The use of biodiesel instead of diesel has some advantages in reducing air pollution, such as 94

lower PM (Rahman, Pourkhesalian et al. 2014), particle number (PN) (Nabi, Zare et al. 95

2016), CO (Ahmed, Hassan et al. 2014, Ruhul, Abedin et al. 2016), HC (Rahman, Hassan et 96

al. 2014, Sanjid, Masjuki et al. 2014, Sanjid, Kalam et al. 2016) and CO2 (Zare, Nabi et al. 97

2016) emissions; however, they can potentially lead to more toxic emissions that should also 98

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

5

be considered (Stevanovic, Miljevic et al. 2013, Hedayat, Stevanovic et al. 2016). Among the 99

different types of biofuel, waste cooking biodiesel has the potential to be used as a fuel due to 100

advantages such as global availability, close properties to diesel and low price (Kulkarni and 101

Dalai 2006). Doroda et al. (2003) studied the effect of waste olive oil, as a fuel, on engine 102

emissions. They reported that CO and CO2 emissions decreased by up to 58.9% and 8.6%, 103

respectively. While, NO2 emission showed an increasing trend. A recent study from our 104

research group demonstrated that using waste cooking biodiesel, instead of diesel, can 105

decrease PM, PN, CO (at higher loads), HC, and CO2 emissions; however NOx emissions 106

showed an increasing trend (Zare, Nabi et al. 2016). 107

Different fuel properties have been used to interpret the changes in exhaust emissions, such 108

as fuel oxygen content (Rahman, Pourkhesalian et al. 2014). The presence of long-chain alkyl 109

esters, which have two oxygen atoms per molecule, is an influential factor that distinguishes 110

biofuels from conventional fossil fuels. However, the oxygen content in biofuels is related to 111

the fatty acid profile such as carbon chain length and unsaturation level (Pham, Bodisco et al. 112

2014). Since the presence of oxygen in fuel can reduce emissions, a low volume of a highly-113

oxygenated fuel additive can significantly reduce emissions. 114

Triacetin [C9H14O6]—a triester of glycerol acetic acid—could be introduced as an additive to 115

biodiesel (Casas, Ruiz et al. 2010). Since glycerol is a byproduct of the biodiesel 116

transesterification process, its production will increase proportionally with an increase in the 117

production of biodiesel. Consequently, this can reduce the price of this feedstock. However, 118

there are some limitations to using glycerol directly as a fuel due to its physical and chemical 119

properties (Gupta and Kumar 2012). The solution could be triacetin, a glycerol-derived 120

product, which is produced by the acetylation process of glycerol and acetic acid (Lapuerta, 121

Rodríguez-Fernández et al. 2015). Mixing this highly-oxygenated additive with biodiesel 122

increases fuel oxygen content, viscosity and the density of the blend, while the heating value 123

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

6

and cetane number of the blend decreased (Casas, Ruiz et al. 2010). There is limited 124

published research on triacetin as a fuel in engines (Nabi, Zare et al. 2015, Zare, Bodisco et 125

al. 2015, Hedayat, Stevanovic et al. 2016, Nabi, Zare et al. 2016, Zare, Nabi et al. 2016, Zare, 126

Bodisco et al. 2017). A recent study by Zare et al. (2016) showed that, by increasing the share 127

of triacetin in the blend, the PM and PN emissions decrease while NOx and HC emissions 128

increase. 129

Due to the increasing share of biofuels in the fuel market, the use of these oxygenated fuels 130

should be evaluated under different engine operating conditions such as cold- and hot-start. 131

This quantification is essential to evaluate the possible use of these alternative fuels in the 132

market. As already discussed, cold-start engine operation is an inevitable part of the daily 133

driving schedule for a considerably high portion of vehicles in cities (André 1991). However, 134

most studies have only focused on hot-start engine operation. 135

This paper investigates the influence of oxygenated fuels on exhaust emissions during cold-136

start, with a comparison to hot-start, using four fuels with oxygen contents ranging from 0% 137

to 13.57%, based on diesel, waste cooking biodiesel and triacetin as an additive. 138

A thorough search in the literature could not find any comparative study into exhaust 139

emissions during cold-start and hot-start using triacetin as a fuel additive with waste cooking 140

biodiesel. However, some results during hot-start have been investigated in a recent study by 141

our research group (Zare, Nabi et al. 2016). More detailed information about this study is 142

discussed in Section 3 in a critical comparison of results. 143

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

7

2. Experimental facilities 144

2.1 Engine specifications and Test setup 145

In order to achieve the stricter emission standards of EURO IV-VI, engines are typically 146

equipped with emission-control devices. As such the emissions can also be dependent on the 147

type of after-treatment devices used. To avoid this, and to gain better insight into the actual 148

engine dependent emissions, it was decided to use a EURO III engine. Here, different after-149

treatment systems can be mentioned. For example, Exhaust Gas Recirculation (EGR) and 150

Selective Catalytic Reduction (SCR), which could be used to abate the excess NOx. Also, a 151

Diesel Particulate Filter (DPF) can be used to abate the excess PM emissions, which can 152

consequently decrease the PN emission depending on the type of DPF (Giakoumis, 153

Rakopoulos et al. 2012). However, the engine used in this study had no after-treatment 154

system. 155

This experimental investigation used a heavy duty, six-cylinder, turbocharged and after-156

cooled diesel engine with a common rail injection system, as shown in Table 1. The engine 157

was coupled to an electronically-controlled water brake dynamometer which controls the 158

engine load. This engine research facility has the ability to program different test cycles that 159

can run automatically using custom software developed by DynoLog. Bodisco et al. (2013) 160

contains more specific information about the engine used and experimental facilities. 161

Figure 1 is a schematic diagram of the test setup. To measure emissions, the exhaust was 162

sampled after the exhaust manifold through a 50 cm long stainless steel tube. To measure 163

CO2 and NOx emissions, a fraction of raw exhaust gas was directed to a CAI-600 CO2 and a 164

CAI-600 CLD NO/NOx analyser through a copper tube fitted with a water trap and a HEPA 165

filter. The remainder of the sampled exhaust gas was diluted with HEPA-filtered ambient air 166

at room temperature in a mini dilution tunnel. CO2 was measured from the diluted exhaust 167

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

8

with a SABLE CA-10 CO2 analyser. The dilution ratio was calculated based on the raw CO2 168

and diluted CO2 from the CAI-600 and SABLE CA-10 CO2 gas analysers, respectively. The 169

results indicated that the dilution ratio ranged between 10 and 27, depending on the engine 170

speed and load. 171

From the diluted exhaust, PM2.5 was measured with a DustTrak TSI 8530. In addition, a 172

DMS500 MKII Fast Particulate Spectrometer was used to measure PN and the particle size 173

distribution. PM1 was calculated from this data with the reinversion utility in the DMS Series 174

Data Presentation Utilities (Version 7.42). PM2.5 and PM1 refer to the mass of particles with 175

diameters of 2.5 and 1 µm or less, respectively. 176

177

Table 1 Test engine specifications 178

179

180

Model Cummins ISBe220 31

Cylinders 6 in-line

Capacity 5.9 L

Emission standard Euro III

Bore × stroke 102 × 120 (mm)

Maximum torque 820 Nm @ 1500 rpm

Maximum power 162 kW @ 2500 rpm

Aspiration Turbocharged

Fuel injection High pressure common rail

Compression ratio 17.3:1

Dynamometer type Electronically-controlled water brake dynamometer

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

9

181

Figure 1 Schematic diagram of test set-up 182

183

2.2 Fuel selection 184

This study used four fuels: diesel (D100); waste cooking biodiesel (B100); the third fuel, 185

which contained 60%, 35% and 5% (by volume) of diesel, waste cooking biodiesel and 186

triacetin, respectively (denoted as D60B35T05); and, B92T08 (92% waste cooking biodiesel 187

and 8% triacetin, by volume). For each fuel, a miscibility test was conducted at room 188

temperature for 96 hours and no phase separation occurred. The tested fuels—D100, 189

D60B35T05, B100 and B92T08—had ascending oxygen contents of 0%, 6.02%, 10.93% and 190

13.57%, respectively. As shown in Table 2, increasing the oxygen content is associated with 191

a decrease in lower heating value (LHV), which consequently decreases engine power. It can 192

also be seen from the table that the fuel density and kinematic viscosity (KV) increases with 193

fuel oxygen content. Kinematic viscosity can have a negative effect on fuel atomisation and 194

evaporation during combustion (Giakoumis, Rakopoulos et al. 2012). 195

Calculations for the listed values for D60B35T05 and B92T08 in Table 2 are based on pure 196

substance compositions. Fuel technical specifications for D100, B100 and T100 can be found 197

in Appendices A1, A2 and A3, respectively. More information about triacetin can be found in 198

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

10

a study by Casas et al. (2010). The reader can also refer to a recent publication from our 199

research group for more information about waste cooking biodiesel (Islam et al., (2015). 200

201

Table 2 Fuel properties (in columns of ascending oxygen content) 202

Fuel D100 D60B35T05 B100 B92T08 T100

O (wt%) 0 6.02 10.93 13.57 44.00

C (wt%) 85.1 80.46 76.93 74.73 49.53

H (wt%) 14.8 13.47 12.21 11.74 6.42

Kinematic Viscosity@40°C (mm2/s) 2.64 3.66 4.82 5.06 7.83

LHV (MJ/kg) 41.77 38.92 37.2 35.57 16.78

HHV (MJ/kg) 44.79 41.74 39.9 38.15 18.08

Density@15°C (g/cc) 0.84 0.866 0.87 0.893 1.159

Cetane number 53.3 53.24 58.6 55.11 15

203

2.3 Design of experiment 204

After reviewing the custom tests used in the literature (Armas, Hernández et al. 2006, 205

Rakopoulos, Dimaratos et al. 2010, Giakoumis, Rakopoulos et al. 2012, Tian, Xu et al. 2013, 206

Tan, Ruan et al. 2014, Rakopoulos, Rakopoulos et al. 2015, Zare, Nabi et al. 2016), a custom 207

test was designed to investigate the engine behaviour at different engine speeds and loads. 208

The rationale behind the design of this custom test was to conduct a fundamental comparative 209

study into cold- and hot-start. 210

As an adaptation of the custom drive cycle introduced in Zare et al. (2016), Zare et al. (2017) 211

introduced a cold-start custom drive cycle to investigate the performance characteristics of 212

oxygenated fuels during cold-start. The present work extends this to consider the effect of 213

oxygenated fuels on emissions during cold-start. 214

Figure 2 shows that the custom test had three main parts, each related to one speed; (a) 1472, 215

(b) 1865 or (c) 2257 rpm, including three engine loads; 0%, 25% and 50%. The speeds and 216

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

11

loads are based on the European Stationary Cycle (ESC) (according to Euro III legislation, as 217

the engine used in this has a Euro III emission certification). The transient parts (acceleration 218

and load increase) are based on the Supplemental Emissions Test (SET) (according to US 219

EPA 2004 emission standards). 220

221

Figure 2 Custom test designed for this study at (a) 1472, (b) 1865 and (c) 2257 rpm 222

2.4 Experimental procedure 223

The cold-start tests were conducted each day, after an overnight engine-off time, with a 224

consistent ambient temperature of 21-23 °C. Hot-start tests were conducted after at least one 225

hour engine warm up and the results had the same trend as the results from ESC, which had 226

some similar operating modes. The measured engine speed during each test were 1472 ± 7, 227

1865 ± 9 and 2257 ± 18 rpm. These small variations, less than 1%, demonstrate the engine 228

speed stability during the tests. The repeatability test for the cold-start experiment was 229

conducted three times with D100, and the repeatability of the results were confirmed by 230

calculating the average, standard deviation (SD) and coefficient of variation (CoV) for 231

different parameters. For example, Table 3 shows the average, SD and CoV of the three 232

repeats for engine speed, indicated torque and indicated mean effective pressure (IMEP). As 233

can be seen from the statistical analysis of the test repeatability in Table 3, CoV (SD/Average 234

600

900

1200

1500

1800

2100

2400

0

25

50

0 200 400 600 800

Sp

eed

(rp

m)

Load

(%

)

Time (s)Load Speed

(c)(a) (b)

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

12

x 100) is less than 1% in most of the cases, however the maximum CoV was 1.95%. The low 235

CoV clearly demonstrates the repeatability of the tests. 236

237

Table 3 Statistical analysis of the test repeatability 238

Engine speed (rpm) Indicated torque (N.m) IMEP (MPa)

rpm@load% Average SD CoV Average SD CoV Average SD CoV

1472@25 1471.19 1.83 0.12 242.36 1.65 0.68 517.67 3.52 0.68

1472@50 1472.50 1.87 0.13 447.97 0.63 0.14 956.83 1.34 0.14

1865@25 1860.83 0.99 0.05 258.61 2.32 0.90 552.37 4.97 0.90

1865@50 1862.43 0.83 0.04 438.38 6.37 1.45 936.35 13.60 1.45

2257@25 2246.25 1.41 0.06 221.62 4.33 1.95 473.35 9.25 1.95

2257@50 2247.91 2.35 0.10 381.39 1.36 0.36 814.62 2.91 0.36

239

240

Fuel lines were cleaned after each fuel change. In addition to CAI-600 CO2, CAI-600 CLD, 241

SABLE CA-10 CO2, DustTrak TSI 8530 and DMS500 MKII, a Testo 350IR gas analyser and 242

MAHA MPM-4M particulate measuring instrument were used to minimise any faulty 243

measurement from emission instrumentation at low redundancy. 244

For each fuel, the three parts of the custom test—(a) 1472, (b) 1865 and (c) 2257 rpm—were 245

used serially. Hence, at cold-start the engine coolant and oil temperatures increased during 246

the custom test. For example, at the first mode of the custom test (1472 rpm) the engine was 247

fully cold, while at 2257 rpm the engine was partially warmed up. The average engine 248

coolant temperature at 1472@25, 1472@50, 1865@25, 1865@50, 2257@25 and 2257@50 249

(with the unit of rpm@% of load) were 34, 39, 61, 65, 83 and 87 °C, respectively. The 250

average engine coolant temperature during hot start testing was higher than 90 °C. 251

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

13

EU Directive 2012/46/EU defines cold-start operation to be when the engine coolant 252

temperature is below 70°C. The result showed that the coolant and lubricant temperatures 253

within the custom test at 1472 and 1865 rpm were under 70°C while, at 2257 rpm, these 254

temperatures were above 70°C. Hence, the final part of the custom test at 2257 rpm cannot be 255

strictly classified as cold-start according to the EU Directive 2012/46/EU. However, at this 256

engine speed the engine lubricant temperature is sub-optimal, in contrast to the coolant 257

temperature. Thus, the results at 2257 rpm show the effect of engine sub-optimal lubricant 258

temperature, rather than coolant and lubricant temperatures. The time lag between the optimal 259

temperatures of coolant and lubricant has been reported previously (Jarrier, Champoussin et 260

al. 2000, Li, Andrews et al. 2009). 261

3. Results and discussion 262

This section studies engine exhaust emissions using four fuels that vary in oxygen content 263

from 0% to 13.57% during cold-start at different engine operating modes with a custom test 264

designed for this study, and compares the results with the hot-start results. The parameters 265

evaluated in this study are NOx, PN, particle size distribution, PM2.5 and PM1. 266

3.1 Nitrogen oxides 267

Figure 3 (a, b) shows the brake specific NOx (BSNOx) during cold-start and its variation, 268

compared to hot-start, at the steady-state modes of the custom test for the tested fuels. For 269

fixed oxygen content, the BSNOx, during cold-start, decreases with an increase in engine 270

speed or load. For example, with D100 under 25% engine load when engine speed is 271

increased from 1472 to 1865 rpm, the BSNOx decreased from 4.98 to 4.72 g/kWh. At 1472 272

rpm, by increasing the engine load from 25% to 50%, the BSNOx with D100 decreased from 273

4.98 to 4.84 g/kWh. This happened at 1472 and 1865 rpm, where the engine temperature is 274

lower, compared to 2257 rpm. As mentioned before, the engine coolant and oil temperatures 275

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

14

increase with engine speed through the custom test. At 2257 rpm, which is the last part of the 276

custom test when the engine is partially warmed up, BSNOx displayed similar behaviour to 277

hot-start and increased with engine load, in contrast to cold-start. During hot-start, BSNOx 278

increases by decreasing the engine speed or by increasing the engine load. NOx emission is 279

influenced by different factors, such as: fuel properties, fuel injection strategy, in-cylinder 280

temperature and residence time under high temperature (Giakoumis, Rakopoulos et al. 2012). 281

These influential factors can reinforce or cancel the other’s effect. For example, higher 282

injected fuel and consequent higher combustion temperature could be a reason for NOx 283

increasing with engine load during hot-start. However, during cold-start, within the custom 284

test, increasing the engine load is associated with an increase in injected fuel quantity, owing 285

to higher engine load demand. This leads to higher NOx formation. Similarly, increasing the 286

engine load is also associated with a decrease in injected fuel quantity, owing to increased 287

engine temperature and decreased friction losses. This leads to lower NOx formation. Thus, 288

the effect of cold-start is superior and BSNOx decreased with engine load. As mentioned 289

above, more fuel is injected during cold-start in order to overcome high friction losses and 290

maintain brake power output (Khair and Jääskeläinen 2013). This could explain the 291

decreasing trend between BSNOx and engine load at 1472 and 1865 rpm. During the custom 292

test, the engine undergoes temperature increases through speeds: 1472, 1865 and 2257 rpm. 293

The decreasing trend between BSNOx and engine load at 1472 and 1865 rpm is due to the 294

fact that at these two engine speeds, the engine is still working under cold-start condition and 295

the BSNOx is affected by it. However, at 2257 rpm the engine is partially warmed up and the 296

friction losses are lower than the other two speeds. Therefore, the engine behaviour is similar 297

to the hot-start condition at some points. 298

Figure 3 (a) shows that, during cold-start, the oxygenated fuels emit higher NOx in 299

comparison to D100. This is similar to hot-start (Zare, Nabi et al. 2016). According to the 300

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

15

literature, biodiesel use can lead to more NOx (Mueller, Boehman et al. 2009, Sun, Caton et 301

al. 2010, Giakoumis, Rakopoulos et al. 2012, Rahman, Masjuki et al. 2014, Imdadul, Masjuki 302

et al. 2017). A reason could be the higher adiabatic flame temperature of biodiesel compared 303

to diesel (Sun, Caton et al. 2010), attributed to the presence of double bonds in biodiesel 304

molecules (Sun, Caton et al. 2010). In addition, lower in-cylinder radiative heat transfer with 305

biodiesels, because of their lower PM formation (similar to this study), could be another 306

factor that leads to a higher adiabatic flame temperature (Mueller, Boehman et al. 2009, Sun, 307

Caton et al. 2010). 308

Figure 3 (b) shows the variation of BSNOx during cold-start, compared to hot-start. BSNOx 309

with D100 during cold-start is lower than hot-start. The range of variation at 1472 and 1865 310

rpm is between 11% and 15.4%, while at 2257 rpm—when the engine is partially warmed 311

up—the variation is less than 2%. The lower combustion temperature during cold-start could 312

be a reason for the lower NOx emission with D100 compared to hot-start. Decreased NOx 313

emission is expected with lower combustion temperature, as at cold-start (Dardiotis, Martini 314

et al. 2013), because thermal NOx formation is significantly influenced by the combustion 315

temperature. Due to the presence of more C-C and C-H bonds in diesel than biodiesel, the 316

combustion temperature could be more influential on NOx formation during diesel 317

combustion than during biodiesel combustion (Omidvarborna, Kumar et al. 2015). 318

In contrast to D100, the oxygenated fuels used in this study had a higher BSNOx during cold-319

start than hot-start. During the first mode of the custom test, when the engine is cold, BSNOx 320

with D60B35T05, B100 and B92T08 were 55%, 94% and 54% higher than that of hot-start, 321

respectively. The minimum BSNOx variation at 1472 and 1865 rpm with oxygenated fuels is 322

around 31% higher than hot-start, while at 2257 rpm the variation range is less than 10%. 323

Roy et al. (2016) reported the reason for higher NOx emission during cold-start as the low 324

temperature NOx (prompt NOx), which is a non-thermal mechanism that could occur during 325

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

16

the combustion of biodiesel when the mixture is subjected to the diffusion flame—when OH 326

and oxygen are present (Mueller, Boehman et al. 2009). Other influential factors could be the 327

higher injected fuels (Roy, Calder et al. 2016) and the earlier start of injection during cold-328

start (Kegl 2008, Giakoumis, Rakopoulos et al. 2012). 329

During cold-start, increased engine temperatures through the custom test are associated with 330

a decrease in the range of higher BSNOx with oxygenated fuels, compared to D100. For 331

example, at the first mode of the custom test (1472 rpm under 25% load) the BSNOx 332

emissions with D60B35T05, B100 and B92T08 are 76%, 125% and 71% higher, 333

respectively, than D100. The variations at the last mode of the custom test (2257 rpm under 334

50% load), when the engine is partially warmed up, are less than 7%. This can be confirmed 335

as the result showed that, during cold-start, the range of BSNOx variation with oxygenated 336

fuels is—in comparison with D100—higher than during hot-start. For example, in 337

comparison with the abovementioned range during cold-start, the highest BSNOx variation 338

during hot-start is 13.9% higher than D100, related to B92T08 with 13.57% oxygen content; 339

however, this variation is less than 10% at most of the modes. 340

341

Figure 3 (a) Brake specific NOx emissions during cold-start and (b) its variations compared to hot-342 start vs. fuel oxygen content, at different modes of the custom test (25% and 50% load at 1472, 1865 343 and 2257 rpm) for all the tested fuels: 0% O2 6.02% O2 10.93% O2 13.57% O2 344

345

4

6

8

10

12

14

-1 4 9 14

BS

NO

x (g

/kW

h)

Oxygen content (%)

1472@25 1472@50 1865@25

1865@50 2257@25 2257@50

(a)

-20

0

20

40

60

80

100

120

-1 4 9 14

BS

NO

x in

crea

se

com

par

ed t

o h

ot-

star

t (%

)

Oxygen content (%)

1472@25 1472@50 1865@25

1865@50 2257@25 2257@50

(b)

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

17

3.2 Particle number 346

Particulate matter has an adverse impact on the environment due to its chemistry, physics and 347

toxicology. In recent years, particle number (PN) emission—a count of the individual 348

particles—has been a focus of attention. For example, the EU Commission included PN in 349

the Euro 5b and Euro 6 emission standards for both light- and heavy-duty vehicles. 350

Figure 4 (a, b) shows the brake specific particle number (BSPN) during cold-start and its 351

variation, compared to hot-start, at steady-state modes of the custom test for the tested fuels. 352

As with hot-start (Zare, Nabi et al. 2016), BSPN during cold-start decreases with engine load 353

for all the tested fuels within the custom test at the three engine speeds. For example, at 1472 354

rpm, under the cold-start condition, and by increasing the engine load from 25 to 50%, the 355

BSPN with D100 decreases from 4.93E+14 to 3.18E+14 #/kWh. PN during hot-start 356

increased by increasing the engine speed for all the tested fuels (Zare, Nabi et al. 2016). 357

During cold-start, the BSPN with D100 displayed a similar trend to hot-start. BSPN with 358

oxygenated fuels had a higher value at 1472 rpm than the two other speeds, 1865 and 2257 359

rpm. This result, contrary to hot-start, could be due to the cold-start effect. BSPN emission is 360

influenced by different factors (e.g. engine temperature, engine speed/load and fuel 361

properties), which can reinforce or cancel the effect of one another under different conditions. 362

During cold-start, the engine coolant and oil temperatures increase with engine speed through 363

the custom test. At 1472 rpm, the engine was completely cold and the BSPN emission was 364

highest. An increase in the engine speed, which could potentially increase the BSPN, 365

coincided with an increase in engine temperature, which could potentially decrease the BSPN 366

toward its hot-start value. Here, the latter effect was superior. 367

As shown in Figure 4 (a), oxygenated fuels during cold-start at all the modes of the custom 368

test had a higher BSPN than D100. For example, at 1472 rpm under 25% and 50% engine 369

load, where the engine was fully cold, oxygenated fuels such as B100 rather than diesel 370

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

18

increased the BSPN up to 15 and 17 times, respectively. This is in contrast to the hot-start 371

results: using the oxygenated fuels instead of D100 decreased PN by up to 73% (Zare, Nabi 372

et al. 2016). During hot-start, an increase in the oxygen content of the fuels decreased PN. 373

The presence of oxygen and the lack of aromatic hydrocarbons in the tested oxygenated fuels 374

may explain these trends (Rahman, Pourkhesalian et al. 2014, Tan, Ruan et al. 2014, Zare, 375

Nabi et al. 2016); however, during cold-start other factors were more influential and caused a 376

higher particle count than with oxygenated fuels compared to D100. This is explained further 377

in Section 3.3. 378

Figure 4 (b) shows the variation of BSPN during cold-start in comparison to hot-start. The 379

BSPN of D100 at cold-start is lower than at hot-start at different modes of the custom test. 380

For example, at 1472 rpm under 25% and 50% engine load, BSPN with D100 during cold-381

start is 42% and 48% lower than during hot-start. This variation decreases as the engine 382

temperature increases through the custom test. On the contrary to D100, for the oxygenated 383

fuels the BSPN during cold-start is higher than hot-start at different modes of the custom test. 384

For example, BSPN for B100 during cold-start could be up to 27-times higher than hot-start. 385

A reason for this contradictory result could be the higher viscosity of the biodiesel during 386

cold-start, which would significantly degrade fuel atomisation and evaporation (Armas, 387

Hernández et al. 2006). 388

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

19

389

Figure 4 (a) Brake specific PN during cold-start and (b) its variation compared to hot-start vs. fuel 390 oxygen content, at different modes of the custom test (25% and 50% load at 1472, 1865 and 2257 391 rpm) for all the tested fuels: 0% O2 6.02% O2 10.93% O2 13.57% O2 392

393

3.3 Particle size distribution 394

Here, the PN size distribution analysis can help to better explain the PN behaviour with 395

different fuels under cold-start and hot-start. The nucleation and accumulation modes are the 396

two main categories of particle size distribution classified by the particle’s diameter. The 397

nucleation mode comprises particles with a diameter of 3–30 nm. Nucleation mode particles 398

are sensitive to sampling systems and dilution parameters (Ristimäki, Keskinen et al. 2005, 399

RÖnkkÖ, Virtanen et al. 2007). They consist of soluble organic fraction (SOF) and sulphates 400

and are mainly formed during cooling and dilution (Vaaraslahti, Keskinen et al. 2005, Lähde, 401

Rönkkö et al. 2010), usually through homogenous and heterogeneous nucleation mechanisms 402

(Meyer and Ristovski 2007, He, Ratcliff et al. 2012). The other category, the accumulation 403

mode, contains the particles but with diameter sizes of 30–1000 nm; it consists of the 404

agglomerated soot and adsorbed volatile materials (Kwon, Lee et al. 2003, Tan, Ruan et al. 405

2014). The agglomeration of nucleation mode particles which are formed from the 406

condensation of volatile materials can contribute to the formation of accumulation mode 407

particles (Peckham, Finch et al. 2011). 408

Figure 5 (a-f) illustrates the PN size distribution for all the tested fuels during cold-start 409

(indicated by solid lines) and hot-start (indicated by dotted lines) at all modes of the custom 410

0

2E+15

4E+15

6E+15

8E+15

1E+16

-1 4 9 14

BS

PN

(#

/kW

h)

Oxygen content (%)

1472@25 1472@50 1865@25

1865@50 2257@25 2257@50

(a)-50

950

1950

2950

-1 4 9 14

BS

PN

incr

ease

co

mp

ared

to

ho

t-st

art (

%)

Oxygen content (%)

1472@25 1472@50 1865@25

1865@50 2257@25 2257@50

(b)

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

20

test. According to the count median diameter (the peak points on the graphs), at almost all the 411

modes of the custom test for the three tested oxygenated fuels during cold-start, a 412

significantly higher portion of the particles are in the nucleation mode than particles in the 413

accumulation mode. This can be observed by comparing the particles in the nucleation and 414

accumulation modes of each coloured solid line (e.g. the green line related to B100) in any 415

sub-diagram of Figure 5. The nucleation mode particles play a key role in the total particle 416

number. The nucleation mode particles during cold-start are significantly higher than hot-start 417

(dotted lines in the figure), when particles are mostly in the accumulation mode. D100 fuels 418

have a higher mass emission as well as lower emissions of volatile materials (Pourkhesalian, 419

Stevanovic et al. 2014) that are building blocks for the nucleation mode. Therefore, in 420

contrast to the oxygenated fuels, most of the particles with D100 are in the accumulation 421

mode during both cold- and hot-start at all the modes of the custom test (except 1472 @25 422

under cold-start), as there is an insufficient amount of volatile material to form strong 423

nucleation mode particles. The accumulation mode particles here play a key role in the total 424

particle count. During cold-start, the number of particles in the accumulation mode is lower 425

than hot-start. 426

Nucleation, condensation and coagulation are the three processes related to the development 427

of particles in the exhaust system that can change particle mass, number, size distribution and 428

chemical composition (Ning, Cheung et al. 2004). Nucleation is related to the formation of 429

new particles, especially fine particles. This gas-to-particle conversion process occurs when 430

nanoparticles form from volatile materials (e.g. hydrocarbons) in a locally supersaturated 431

zone. Condensation is also a gas-to-particle conversion process that occurs when volatile 432

materials condense on existing particles; however, this process leads to an increase in 433

particulate size. In contrast to nucleation and condensation, coagulation occurs when existing 434

particles come into contact with each other and these smaller particles form larger particles. 435

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

21

This process, which is affected by the Brownian force, decreases particle number (mainly 436

from smaller particles) and increases particle size (Ning, Cheung et al. 2004). 437

In this study, the temperature of the exhaust line during cold-start was lower than at hot-start 438

but increased through the custom test. During cold-start, the low temperature of the exhaust 439

line lowered the temperature of hot exhaust gas in the system. During this cooling, the 440

volatile substances nucleated homogenously into nanoparticles (Kasper 2005). Similarly, 441

Ning et al. (2004) reported that the cooling of exhaust gas results in the nucleation and 442

condensation of volatile materials. Decreasing the exhaust gas temperature leads to increased 443

saturation vapour pressure and saturation ratio of the volatile materials, resulting in 444

nucleation and the formation of liquid particles. That could be why particles increased in the 445

nucleation mode during cold-start rather than hot-start, as shown in Figure 5. Another factor 446

that reinforces the higher nucleation during cold-start is a high amount of volatile materials 447

during cold-start. The volatile organic compounds during cold-start have been reported to be 448

higher than during hot-start (Tsai, Chang et al. 2012, George, Hays et al. 2014). For example, 449

George et al. (2014) reported a significant increase (up to a factor of 24) in volatile organic 450

compounds during cold-start, compared to hot-start. 451

Since the presence of volatile materials can increase the probability of nucleation, excessive 452

quantities of volatile compounds from the combustion of biodiesel (Surawski, Miljevic et al. 453

2011, Pourkhesalian, Stevanovic et al. 2014) could be a reason for the higher particle number 454

of the oxygenated fuels in the nucleation mode in comparison to D100. To further support our 455

observations, we refer to the study of Hedayat et al. (2016), who used exactly the same 456

engine and exactly the same fuel. Their experiments, that used ESC (which has similar engine 457

operating modes to the custom test), have shown a significant increase in the oxygenated 458

volatile organic compounds (building blocks of the nucleation mode) with the use of 459

biodiesels. This can be seen in Figure 5 by comparing the solid lines in the nucleation mode. 460

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

22

With D100, the number of particles in the nucleation mode during cold-start is slightly higher 461

than hot-start while, with oxygenated fuels, the difference is significant. The higher particle 462

volatility of the biodiesel compared to diesel may be the explanation. In addition, the 463

presence of volatile materials can increase the chance of condensation as well as nucleation. 464

As mentioned above, condensation can increase the size of particles. This may be another 465

contributing factor for bigger particles in oxygenated fuels, compared to D100, during cold-466

start at 1472 rpm under 25% load, where the cold-start test begins. This can be seen by 467

comparing the count median diameters of the solid lines in Figure 5 (a). However, during hot-468

start, the oxygenated fuels had smaller count median diameters when compared to D100 469

(Zare, Nabi et al. 2016). 470

During hot-start, the use of biodiesel, instead of diesel, increases the nucleation mode 471

particles—this is owing to an increase in volatile materials. The observed decrease in the 472

accumulation mode particles is due to a lack of aromatic hydrocarbons (Tan, Ruan et al. 473

2014). Rahman et al. (2014) reported that using pure biodiesel instead of diesel decreased PN 474

emissions, principally from the particles with a mobility diameter of greater than 100 nm, in 475

contrast to nanoparticles (below 50 nm). Figure 5 also shows that, within the custom test by 476

increasing the engine speed, which leads to a higher engine temperature (higher engine 477

lubricant and coolant temperatures), the size of particles increased and moved toward the 478

accumulation mode. 479

The use of oxygenated fuels during cold-start has a noticeable disadvantage, it increases 480

nucleation mode particles. These small particles are harmful to health as they can penetrate 481

deep into lungs and cause inflammation at the sites of deposition (Cheung, Polidori et al. 482

2009, Vaughan, Stevanovic et al. 2015, Stevanovic, Vaughan et al. 2017). In addition, the 483

adverse effect of particles increases as their size decrease (Lapuerta, Armas et al. 2008). The 484

harmful effect of using oxygenated fuels during cold-start should be considered since cold-485

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

23

start operation is an inevitable part of the daily driving schedule for a significantly high 486

proportion of vehicles in mass-populated cities. 487

488

489

490

491

492

493 Figure 5 PN size distribution during cold-start (solid lines) and hot-start (dotted lines) within the 494 custom test at (a) 1472@25, (b) 1472@50, (c) 1865@25, (d) 1865@50, (e) 2257@25 and (f) 495 2257@50 for all the tested fuels: 0% O2 6.02% O2 10.93% O2 13.57% O2 496

497

0

1

2

3

4

1 10 100 1000

dN

/dlo

gDp

(#

/cm

^3)

Particle size (nm)

x107

(a)

0

2

4

6

8

1 10 100 1000

dN

/dlo

gDp

(#

/cm

^3)

Particle size (nm)

x107

(b)

0

2

4

6

8

1 10 100 1000

dN

/dlo

gDp

(#

/cm

^3)

Particle size (nm)

x107

(c)

0

2

4

6

8

10

1 10 100 1000

dN

/dlo

gDp

(#

/cm

^3)

Particle size (nm)

x107

(d)

0

1

23

4

56

1 10 100 1000

dN

/dlo

gDp

(#

/cm

^3)

Particle size (nm)

x107

(e)

0

2

4

6

8

1 10 100 1000

dN

/dlo

gDp

(#

/cm

^3)

Particle size (nm)

x107

(f)

10 100

10 10010 100

10 100 10 100

10 100

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

24

3.4 Particulate matter 498

Figure 6 (a-d) shows the brake specific PM1 (BSPM1) and PM2.5 (BSPM2.5) during cold-start 499

and their variations, compared to hot-start at the steady-state modes of the custom test for the 500

tested fuels. As with hot-start (Zare, Nabi et al. 2016), Figure 6 shows that, during cold-start, 501

BSPM2.5 and BSPM1 emissions increase at most of the modes of the custom test with 502

increasing engine load. This is because increasing the engine load leads to a greater mass of 503

injected fuel, lower AFR, higher combustion temperature, longer diffusion combustion 504

duration and lower oxygen availability—all contributing to lower soot oxidation in the 505

expansion stroke and, consequently, increased PM. However, during cold-start, PM 506

formation caused by load increase can be affected by other factors as well. 507

PM forms in the fuel-rich zone under high temperature decomposition during combustion 508

(Giakoumis, Rakopoulos et al. 2012). This mainly occurs in the core region of the fuel spray. 509

The presence of oxygen, in addition to the fuel type, aids the soot oxidation process. Apart 510

from the intake air effect, fuel oxygen can prevent higher PM formation by reducing the local 511

fuel-rich zone within the core region of the sprayed fuel. It has frequently been reported that 512

PM formation is highly influenced by fuel oxygen content (Giakoumis, Rakopoulos et al. 513

2012, Rahman, Pourkhesalian et al. 2014, Zare, Nabi et al. 2016). The result during hot-start 514

indicated that an increase in the oxygen content of the fuel decreases PM emission (Zare, 515

Nabi et al. 2016). Compared to the fuel physical properties—such as cetane number and 516

viscosity—the fuel chemical composition, especially fuel oxygen, has a greater influence on 517

PM reduction (Rahman, Pourkhesalian et al. 2014). 518

During hot-start, D100 had higher BSPM1 and BSPM2.5 than the oxygenated fuels; however, 519

the PM difference between D100 and the oxygenated fuel for BSPM2.5 is higher than BSPM1. 520

For example, at the first mode of the custom test (1472@25), the use of D60B35T05, B100 521

and B92T08 reduced BSPM2.5 by 68%, 80% and 88%, respectively, compared to D100, 522

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

25

BSPM1 reduced by 20%, 76% and 81%, respectively. This difference is related to the 523

particles with diameters 1–2.5 µm, these have a greater effect on PM emission owing to their 524

higher mass because the greater the size of a particle, the greater its mass. The difference can 525

be explained because the size of the particles emitted with D100 is larger than that of the 526

other tested fuels in this study, as shown in Figure 5. For example, Figure 5 (f) shows that, 527

under 50% engine load at 2257 rpm, during hot-start, the accumulation mode count median 528

diameters for the tested fuels with 0%, 6.02%, 10.93%, and 13.57% oxygen content are 65, 529

56, 48 and 42 nm, respectively. Increasing the fuel oxygen content decreases the size of the 530

emitted particles (Zare, Nabi et al. 2016). Thus, the smaller size of particles with oxygenated 531

fuels is related to the higher volatility of biodiesel particles. 532

During cold-start, lower engine temperature—in addition to the fuel’s properties—also 533

affects the size and number of particles and, hence, PM emission. Figure 6 (a) shows that, 534

during cold-start, the use of oxygenated fuels instead of D100 only decreases BSPM2.5 at 535

some of the modes. For example, at the first three modes of the custom test, the use of 536

D60B35T05 instead of D100 could decrease BSPM2.5 up to 40%, while using B92T08 537

instead of D100 increased BSPM2.5 up to 28%. It can be seen in Figure 6 (c) that oxygenated 538

fuels did not successfully reduce BSPM1 at the first three modes of the custom test during 539

cold-start, when the engine temperature was low. These results contrast with hot-start results, 540

where oxygenated fuels could decrease both BSPM1 and BSPM2.5. This is because, during 541

cold-start, the oxygenated fuels emit a significantly higher number of particles than D100, as 542

shown in Figure 5. These small particles increase the BSPM1 and BSPM2.5 and dominate the 543

effect of the fuel oxygen content. Owing to the small size of these particles, which are mostly 544

in the nucleation mode, they are more influential on BSPM1 than BSPM2.5 because BSPM1 545

counts the mass of particles smaller than 1 µm, and BSPM2.5 counts the mass of particles 1–546

2.5 µm as well as PM1. 547

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

26

Figure 6 (b) and (d) show that B100 and B92T08 during cold-start had significantly higher 548

BSPM2.5 and BSPM1 relative to hot-start. This could be explained by the significant increase 549

of PN emission during cold-start. By comparing Figure 5 (b) and (d), it can be seen that 550

BSPM1 had more sensitivity (higher variation) to cold-start than BSPM2.5. For example, at 551

1472 rpm under 50% load with B100 and B92T08, BSPM1 during cold-start is, respectively, 552

7.3 and 6.6 times higher than hot-start, while that for BSPM2.5 was five and four times higher. 553

The cause of higher variation in BSPM1 than BSPM2.5 is the higher number of particles with 554

a smaller size in nucleation mode during cold-start than hot-start, as shown in Figure 5. These 555

small particles have more effect on BSPM1 than BSPM2.5, because BSPM1 refers to the mass 556

of particles of 1 µm or less in diameter, compared to BSPM2.5, which refers to the mass of 557

particles of 2.5 µm or less in diameter. 558

In contrast with B100 and B92T08, D100 showed lower BSPM2.5 and BSPM1 during cold-559

start than hot-start, as shown in Figure 6 (b) and (d). The variation could be up to 63% and 560

44% for BSPM2.5 and BSPM1, respectively. This could be because, during cold-start, D100 561

emitted a lower number of particles than hot-start. However, the greater quantity of small 562

particles during cold-start affected BSPM1 more than BSPM2.5, and that explains a lower 563

variation of BSPM1 compared to BSPM2.5. 564

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

27

565

566

Figure 6 (a and c) Brake specific PM1 and PM2.5 during cold-start and (b and d) their variations 567 compared to hot-start vs. fuel oxygen content, at different modes of the custom test (25% and 50% 568 load at 1472, 1865 and 2257 rpm) for all the tested fuels: 0% O2 6.02% O2 10.93% O2 13.57% 569 O2 570

571

4. Conclusion 572

This study evaluated the effect of oxygenated fuels on exhaust emissions during cold-start, 573

with a comparison to hot-start, using four fuels with oxygen contents ranging from 0 to 574

13.57%, based on diesel, waste cooking biodiesel and triacetin as an additive. This research 575

used a custom test to evaluate the exhaust emissions during cold-start. The engine used in this 576

investigation was a six-cylinder, turbocharged, after-cooled diesel engine with a common rail 577

injection system. The cold-start tests were conducted every day after an overnight engine-off 578

time. The parameters evaluated in this study were NOx, PN, particle size distribution, PM2.5 579

and PM1, and the following conclusions were drawn: 580

• A comparative study between cold-start and hot-start when diesel was the tested fuel 581

showed that during cold-start NOx, PN, PM1 and PM2.5 were up to 15.4%, 48%, 44% 582

0

0.03

0.06

0.09

0.12

0.15

-1 4 9 14

BS

PM

2.5(g

/kW

h)

Oxygen content (%)

1472@25 1472@50 1865@25

1865@50 2257@25 2257@50

(a)

-100

0

100

200

300

400

500

600

-1 4 9 14

BS

PM

2.5in

crea

se

com

par

ed t

o h

ot-

star

t (%

)

Oxygen content (%)

1472@25 1472@50 1865@25

1865@50 2257@25 2257@50

(b)

0

0.1

0.2

0.3

-1 4 9 14

BS

PM

1(g

/kW

h)

Oxygen content (%)

1472@25 1472@50 1865@25

1865@50 2257@25 2257@50

(c)

-50

150

350

550

750

950

-1 4 9 14

BS

PM

1in

crea

se

com

par

ed t

o h

ot-

star

t (%

)

Oxygen content (%)

1472@25 1472@50 1865@25

1865@50 2257@25 2257@50

(d)

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

28

and 63% lower, respectively. However, the oxygenated fuels during cold-start showed 583

significant increases in NOx (up to 94%), PN (up to 27 times), PM1 (up to 7.3 times) 584

and PM2.5 (up to 5 times), when compared to hot-start. 585

• Compared to diesel, oxygenated fuels showed significantly higher NOx (up to 125%) 586

during cold-start, while during hot-start that increase was up to 13.9%. 587

• Using oxygenated fuels instead of diesel during hot-start decreased the PN (up to 588

73%), PM2.5 (up to 91%) and PM1 (up to 89%), while during cold-start it only 589

decreased PM1 and PM2.5 at some engine operating modes and PN significantly 590

increased up to 17 times. 591

• Compared to hot-start, using oxygenated fuels during cold-start increased the 592

nucleation mode particles significantly, which is harmful. 593

The results should be taken into consideration, since cold-start operation is an inevitable part 594

of the daily driving schedule for a significantly high portion of vehicles, especially in cities. 595

5. Acknowledgement 596

This research was supported by the Australian Research Council Linkage Projects funding 597

scheme (project number LP110200158). The author would also like to acknowledge: the 598

software developer, Mr. Andrew Elder from DynoLog Dynamometer Pty Ltd; laboratory 599

assistance from Mr. Noel Hartnett; Peak3 Pty Ltd for assistance with measuring instruments; 600

Eco Tech Biodiesel (Dr. Doug Stuart) for the supply of waste cooking biodiesel; Mr. James 601

Hurst for providing copyediting and proofreading service; and Dr. Svetlana Stevanovic and 602

Dr. Meisam Babaie for their support and guidance. 603

604

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

29

6. References 605

Ahmed, S., M. H. Hassan, M. A. Kalam, S. M. Ashrafur Rahman, M. J. Abedin and A. Shahir 606 (2014). "An experimental investigation of biodiesel production, characterization, engine 607 performance, emission and noise of Brassica juncea methyl ester and its blends." Journal of 608 Cleaner Production 79: 74-81. 609 André, J.-M. and R. Joumard (2005). "Modelling of cold start excess emissions for passenger 610 cars." INRETS Report, No. LTE 509: 239. 611 André, M. (1991). In actual use car testing: 70,000 kilometers and 10,000 trips by 55 French 612 cars under real conditions, SAE Technical Paper. 613 Armas, O., J. J. Hernández and M. D. Cárdenas (2006). "Reduction of diesel smoke opacity 614 from vegetable oil methyl esters during transient operation." Fuel 85(17–18): 2427-2438. 615 Bielaczyc, P., J. Merkisz and J. Pielecha (2001). A method of reducing the exhaust emissions 616 from DI diesel engines by the introduction of a fuel cut off system during cold start, SAE 617 Technical Paper. 618 Bodisco, T. and R. J. Brown (2013). "Inter-cycle variability of in-cylinder pressure 619 parameters in an ethanol fumigated common rail diesel engine." Energy 52: 55-65. 620 Cao, Y. (2007). Operation and Cold Start Mechanisms of Internal Combustion Engines with 621 Alternative Fuels, SAE Technical Paper. 622 Casas, A., J. R. Ruiz, M. a. J. s. Ramos and A. Pérez (2010). "Effects of triacetin on biodiesel 623 quality." Energy & Fuels 24(8): 4481-4489. 624 Cheung, K. L., A. Polidori, L. Ntziachristos, T. Tzamkiozis, Z. Samaras, F. R. Cassee, M. 625 Gerlofs and C. Sioutas (2009). "Chemical characteristics and oxidative potential of 626 particulate matter emissions from gasoline, diesel, and biodiesel cars." Environmental science 627 & technology 43(16): 6334-6340. 628 Dardiotis, C., G. Martini, A. Marotta and U. Manfredi (2013). "Low-temperature cold-start 629 gaseous emissions of late technology passenger cars." Applied energy 111: 468-478. 630 Dorado, M., E. Ballesteros, J. Arnal, J. Gomez and F. Lopez (2003). "Exhaust emissions from 631 a Diesel engine fueled with transesterified waste olive oil." Fuel 82(11): 1311-1315. 632 George, I. J., M. D. Hays, R. Snow, J. Faircloth, B. J. George, T. Long and R. W. Baldauf 633 (2014). "Cold temperature and biodiesel fuel effects on speciated emissions of volatile 634 organic compounds from diesel trucks." Environmental science & technology 48(24): 14782-635 14789. 636 Giakoumis, E. G., C. D. Rakopoulos, A. M. Dimaratos and D. C. Rakopoulos (2012). 637 "Exhaust emissions of diesel engines operating under transient conditions with biodiesel fuel 638 blends." Progress in Energy and Combustion Science 38(5): 691-715. 639 Gupta, M. and N. Kumar (2012). "Scope and opportunities of using glycerol as an energy 640 source." Renewable and Sustainable Energy Reviews 16(7): 4551-4556. 641 He, X., M. A. Ratcliff and B. T. Zigler (2012). "Effects of gasoline direct injection engine 642 operating parameters on particle number emissions." Energy & Fuels 26(4): 2014-2027. 643 Hedayat, F., S. Stevanovic, A. Milic, B. Miljevic, M. N. Nabi, A. Zare, S. Bottle, R. J. Brown 644 and Z. D. Ristovski (2016). "Influence of oxygen content of the certain types of biodiesels on 645 particulate oxidative potential." Science of the Total Environment 545: 381-388. 646 Imdadul, H. K., H. H. Masjuki, M. A. Kalam, N. W. M. Zulkifli, M. Kamruzzaman, M. M. 647 Shahin and M. M. Rashed (2017). "Evaluation of oxygenated n-butanol-biodiesel blends 648 along with ethyl hexyl nitrate as cetane improver on diesel engine attributes." Journal of 649 Cleaner Production 141: 928-939. 650 Islam, M. A., M. M. Rahman, K. Heimann, M. N. Nabi, Z. D. Ristovski, A. Dowell, G. 651 Thomas, B. Feng, N. von Alvensleben and R. J. Brown (2015). "Combustion analysis of 652 microalgae methyl ester in a common rail direct injection diesel engine." Fuel 143: 351-360. 653

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

30

Jarrier, L., J. Champoussin, R. Yu and D. Gentile (2000). Warm-up of a DI diesel engine: 654 experiment and modeling, SAE Technical Paper. 655 Kasper, M. (2005). Sampling and measurement of nanoparticle emissions for type approval 656 and field control, SAE Technical Paper. 657 Kegl, B. (2008). "Biodiesel usage at low temperature." Fuel 87(7): 1306-1317. 658 Khair, M. K. and H. Jääskeläinen (2013). Diesel Fuel Injection. DieselNet Technology 659 Guide, DieselNet. 660 Kulkarni, M. G. and A. K. Dalai (2006). "Waste cooking oil an economical source for 661 biodiesel: a review." Industrial & engineering chemistry research 45(9): 2901-2913. 662 Kwon, S.-B., K. W. Lee, K. Saito, O. Shinozaki and T. Seto (2003). "Size-dependent 663 volatility of diesel nanoparticles: chassis dynamometer experiments." Environmental science 664 & technology 37(9): 1794-1802. 665 Lähde, T., T. Rönkkö, A. Virtanen, A. Solla, M. Kytö, C. Söderström and J. Keskinen (2010). 666 "Dependence between nonvolatile nucleation mode particle and soot number concentrations 667 in an EGR equipped heavy-duty diesel engine exhaust." Environmental science & technology 668 44(8): 3175-3180. 669 Lapuerta, M., O. Armas and J. Rodriguez-Fernandez (2008). "Effect of biodiesel fuels on 670 diesel engine emissions." Progress in energy and combustion science 34(2): 198-223. 671 Lapuerta, M., J. Rodríguez-Fernández and R. García-Contreras (2015). "Effect of a glycerol-672 derived advanced biofuel –FAGE (fatty acid formal glycerol ester)– on the emissions of a 673 diesel engine tested under the New European Driving Cycle." Energy 93, Part 1: 568-579. 674 Lee, D.-W., J. Johnson, J. Lv, K. Novak and J. Zietsman (2012). Comparisons between 675 Vehicular Emissions from Real-World In-Use Testing and EPA MOVES Estimation, Texas 676 Transportation Institute College Station, TX. 677 Li, H., G. E. Andrews, D. Savvidis, B. Daham, K. Ropkins, M. Bell and J. Tate (2009). 678 "Study of thermal characteristics and emissions during cold start using an on-board 679 measuring method for modern SI car real world urban driving." SAE International Journal of 680 Engines 1(1): 804-819. 681 Meyer, N. K. and Z. D. Ristovski (2007). "Ternary nucleation as a mechanism for the 682 production of diesel nanoparticles: Experimental analysis of the volatile and hygroscopic 683 properties of diesel exhaust using the volatilization and humidification tandem differential 684 mobility analyzer." Environ Sci Technol 41(21): 7309-7314. 685 Mueller, C. J., A. L. Boehman and G. C. Martin (2009). An Experimental Investigation of the 686 Origin of Increased NOx Emissions When Fueling a Heavy-Duty Compression-Ignition 687 Engine with Soy Biodiesel, SAE Technical Paper. 688 Nabi, M. N., A. Zare, F. M. Hossain, M. M. Rahman, T. A. Bodisco, Z. D. Ristovski and R. 689 J. Brown (2016). "Influence of fuel-borne oxygen on European Stationary Cycle: Diesel 690 engine performance and emissions with a special emphasis on particulate and NO emissions." 691 Energy Conversion and Management 127: 187-198. 692 Nabi, N., A. Zare, M. Hossain, M. M. Rahman, D. Stuart, Z. Ristovski and R. Brown (2015). 693 Formulation of new oxygenated fuels and their influence on engine performance and exhaust 694 emissions. Proceedings of the 2015 Australian Combustion Symposium, The Combustion 695 Institute Australia and New Zealand Section. 696 Nam, E. (2008). Analysis of particulate matter emissions from light-duty gasoline vehicles in 697 Kansas City, US Environmental Protection Agency. 698 Ning, Z., C. Cheung and S. Liu (2004). "Experimental investigation of the effect of exhaust 699 gas cooling on diesel particulate." Journal of Aerosol Science 35(3): 333-345. 700 Omidvarborna, H., A. Kumar and D.-S. Kim (2015). "NOx emissions from low-temperature 701 combustion of biodiesel made of various feedstocks and blends." Fuel Processing 702 Technology 140: 113-118. 703

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

31

Peckham, M. S., A. Finch, B. Campbell, P. Price and M. T. Davies (2011). Study of particle 704 number emissions from a turbocharged gasoline direct injection (GDI) engine including data 705 from a fast-response particle size spectrometer, SAE Technical Paper. 706 Pham, P. X., T. A. Bodisco, Z. D. Ristovski, R. J. Brown and A. R. Masri (2014). "The 707 influence of fatty acid methyl ester profiles on inter-cycle variability in a heavy duty 708 compression ignition engine." Fuel 116: 140-150. 709 Pourkhesalian, A., S. Stevanovic, F. Salimi, M. Rahman, H. Wang, P. Pham, S. Bottle, A. 710 Masri, R. Brown and Z. Ristovski (2014). "Influence of fuel molecular structure on the 711 volatility and oxidative potential of biodiesel particulate matter." Environmental science & 712 technology 48(21): 12577-12585. 713 Rahman, M., A. Pourkhesalian, M. Jahirul, S. Stevanovic, P. Pham, H. Wang, A. Masri, R. 714 Brown and Z. Ristovski (2014). "Particle emissions from biodiesels with different physical 715 properties and chemical composition." Fuel 134: 201-208. 716 Rahman, M. M., M. H. Hassan, M. A. Kalam, A. E. Atabani, L. A. Memon and S. M. A. 717 Rahman (2014). "Performance and emission analysis of Jatropha curcas and Moringa oleifera 718 methyl ester fuel blends in a multi-cylinder diesel engine." Journal of Cleaner Production 65: 719 304-310. 720 Rahman, S. M. A., H. H. Masjuki, M. A. Kalam, M. J. Abedin, A. Sanjid and S. Imtenan 721 (2014). "Effect of idling on fuel consumption and emissions of a diesel engine fueled by 722 Jatropha biodiesel blends." Journal of Cleaner Production 69: 208-215. 723 Rakopoulos, C. D., A. M. Dimaratos, E. G. Giakoumis and D. C. Rakopoulos (2010). 724 "Investigating the emissions during acceleration of a turbocharged diesel engine operating 725 with bio-diesel or n-butanol diesel fuel blends." Energy 35(12): 5173-5184. 726 Rakopoulos, D. C., C. D. Rakopoulos and E. G. Giakoumis (2015). "Impact of properties of 727 vegetable oil, bio-diesel, ethanol and n-butanol on the combustion and emissions of 728 turbocharged HDDI diesel engine operating under steady and transient conditions." Fuel 156: 729 1-19. 730 Reiter, M. S. and K. M. Kockelman (2016). "The problem of cold starts: A closer look at 731 mobile source emissions levels." Transportation Research Part D: Transport and Environment 732 43: 123-132. 733 Ristimäki, J., J. Keskinen, A. Virtanen, M. Maricq and P. Aakko (2005). "Cold temperature 734 PM emissions measurement: Method evaluation and application to light duty vehicles." 735 Environmental science & technology 39(24): 9424-9430. 736 Roberts, A., R. Brooks and P. Shipway (2014). "Internal combustion engine cold-start 737 efficiency: A review of the problem, causes and potential solutions." Energy Conversion and 738 Management 82: 327-350. 739 RÖnkkÖ, T., A. Virtanen, J. Kannosto, J. Keskinen, M. Lappi and L. Pirjola (2007). 740 "Nucleation mode particles with a nonvolatile core in the exhaust of a heavy duty diesel 741 vehicle." Environmental science & technology 41(18): 6384-6389. 742 Roy, M. M., J. Calder, W. Wang, A. Mangad and F. C. M. Diniz (2016). "Cold start idle 743 emissions from a modern Tier-4 turbo-charged diesel engine fueled with diesel-biodiesel, 744 diesel-biodiesel-ethanol, and diesel-biodiesel-diethyl ether blends." Applied Energy 180: 52-745 65. 746 Ruhul, M. A., M. J. Abedin, S. M. A. Rahman, B. H. H. Masjuki, A. Alabdulkarem, M. A. 747 Kalam and I. Shancita (2016). "Impact of fatty acid composition and physicochemical 748 properties of Jatropha and Alexandrian laurel biodiesel blends: An analysis of performance 749 and emission characteristics." Journal of Cleaner Production 133: 1181-1189. 750 Sakunthalai, R. A., H. Xu, D. Liu, J. Tian, M. Wyszynski and J. Piaszyk (2014). Impact of 751 cold ambient conditions on cold start and idle emissions from diesel engines, SAE Technical 752 Paper. 753

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

32

Samhaber, C., A. Wimmer and E. Loibner (2001). Modeling of engine warm-up with 754 integration of vehicle and engine cycle simulation, SAE Technical Paper. 755 Sanjid, A., M. A. Kalam, H. H. Masjuki, M. Varman, N. W. B. M. Zulkifli and M. J. Abedin 756 (2016). "Performance and emission of multi-cylinder diesel engine using biodiesel blends 757 obtained from mixed inedible feedstocks." Journal of Cleaner Production 112, Part 5: 4114-758 4122. 759 Sanjid, A., H. H. Masjuki, M. A. Kalam, S. M. A. Rahman, M. J. Abedin and S. M. Palash 760 (2014). "Production of palm and jatropha based biodiesel and investigation of palm-jatropha 761 combined blend properties, performance, exhaust emission and noise in an unmodified diesel 762 engine." Journal of Cleaner Production 65: 295-303. 763 Stevanovic, S., B. Miljevic, N. C. Surawski, K. E. Fairfull-Smith, S. Bottle, R. Brown and Z. 764 Ristovski (2013). "Influence of oxygenated organic aerosols (OOAs) on the oxidative 765 potential of diesel and biodiesel particulate matter." Environmental science & technology 766 47(14): 7655-7662. 767 Stevanovic, S., A. Vaughan, F. Hedayat, F. Salimi, M. Rahman, A. Zare, R. Brown, R. 768 Brown, H. Wang and Z. Zhang (2017). "Oxidative potential of gas phase combustion 769 emissions-An underestimated and potentially harmful component of air pollution from 770 combustion processes." Atmospheric Environment. 771 Sun, J., J. A. Caton and T. J. Jacobs (2010). "Oxides of nitrogen emissions from biodiesel-772 fuelled diesel engines." Progress in Energy and Combustion Science 36(6): 677-695. 773 Surawski, N. C., B. Miljevic, G. A. Ayoko, S. Elbagir, S. Stevanovic, K. E. Fairfull-Smith, S. 774 E. Bottle and Z. D. Ristovski (2011). "Physico-chemical characterisation of particulate 775 emissions from a compression ignition engine: the influence of biodiesel feedstock." 776 Environmental Science & Technology 45(24): 10337-10343. 777 Tan, P.-q., S.-s. Ruan, Z.-y. Hu, D.-m. Lou and H. Li (2014). "Particle number emissions 778 from a light-duty diesel engine with biodiesel fuels under transient-state operating 779 conditions." Applied Energy 113: 22-31. 780 Tian, J., H. Xu, A. Ghafourian, D. Liu, C. Tan and S.-J. Shuai (2013). "Transient emissions 781 characteristics of a turbocharged engine fuelled by biodiesel blends." SAE International 782 Journal of Fuels and Lubricants 6(2): 457-465. 783 Tsai, J.-H., S.-Y. Chang and H.-L. Chiang (2012). "Volatile organic compounds from the 784 exhaust of light-duty diesel vehicles." Atmospheric Environment 61: 499-506. 785 Vaaraslahti, K., J. Keskinen, B. Giechaskiel, A. Solla, T. Murtonen and H. Vesala (2005). 786 "Effect of lubricant on the formation of heavy-duty diesel exhaust nanoparticles." 787 Environmental science & technology 39(21): 8497-8504. 788 Vaughan, A., S. Stevanovic, L. E. Morrison, A. M. Pourkhesalian, M. Rahman, A. Zare, B. 789 Miljevic, F. Goh, V. Relan and R. V. Bowman (2015). "Removal of organic content from 790 diesel exhaust particles alters cellular responses of primary human bronchial epithelial cells 791 cultured at an air-liquid interface." Journal of Environmental & Analytical Toxicology 5(5): 792 1. 793 Will, F. and A. Boretti (2011). "A new method to warm up lubricating oil to improve the fuel 794 efficiency during cold start." SAE international journal of engines 4(1): 175-187. 795 Zare, A., T. Bodisco, N. Nabi, M. Hossain, M. M. Rahman, D. Stuart, Z. Ristovski and R. 796 Brown (2015). Impact of Triacetin as an oxygenated fuel additive to waste cooking biodiesel: 797 transient engine performance and exhaust emissions. Proceedings of the 2015 Australian 798 Combustion Symposium, The Combustion Institute Australia and New Zealand Section. 799 Zare, A., T. A. Bodisco, M. N. Nabi, F. M. Hossain, M. M. Rahman, Z. D. Ristovski and R. 800 J. Brown (2017). "The influence of oxygenated fuels on transient and steady-state engine 801 emissions." Energy 121: 841-853. 802

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

33

Zare, A., T. A. Bodisco, M. N. Nabi, F. M. Hossain, Z. D. Ristovski and R. J. Brown (2017). 803 "A comparative investigation into cold-start and hot-start operation of diesel engine 804 performance with oxygenated fuels during transient and steady-state operation." Fuel Under-805 review. 806 Zare, A., M. N. Nabi, T. A. Bodisco, F. M. Hossain, M. M. Rahman, Z. D. Ristovski and R. 807 J. Brown (2016). "The effect of triacetin as a fuel additive to waste cooking biodiesel on 808 engine performance and exhaust emissions." Fuel 182: 640-649. 809

810

811

812

813

814

815

816

817

818

819

820

821

822

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

34

7. Appendices 823

The following Appendices A1, A2 and A3 show the fuel properties of D100, B100 and T100, 824

respectively. 825

826

7.1 A1. Automotive Diesel –Xtra Low Sulfur (D100) 827

828

Provided by: 829

CALTEX REFINERIES (Qld) Ltd 830 Wynnum, QLD 4178, Australia 831 832

833

834

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

35

7.2 A2. Waste cooking biodiesel (B100) 835

836

Provided by: 837

ECOTECH BIODIESEL 838 SPECCHECK LABORATORIES P/L 839 Mittagong, NSW 2575, Australia 840 841 METHOD TEST SPECIFICTION RESULT UNITS

EN12662 Total contamination 24 max 11.1 mg/kg

ASTM D6584 Free glycerol 0.02 max 0.01 %

ASTM D6584 Total glycerol 0.25 max 0.089 %

ASTM D6584 Monoglycerides 0.8 max 0.246 %

EN14103 Ester content 96.5 min 97.4 %

ASTM D2068 Filter blocking tendency (B100) 2 max 1.05 -

ASTM D7501 Cold soak filterability 360 max 201 sec

ASTM D6304 Moisture content 500

312 ppm

ASTM D664 Total Acid Number 0.8 max 0.25 mgKOH/g

ASTM D93 Flash point 120 min >130 °C

EN14110 Alcohol content 0.2 max 0.02 %

EN14112 Oxidation stability 9 min 9.52 h

ASTM D130 Copper Corrosion 1 max 1A -

ASTM D4350 Carbon Residue (10% res) 0.3 max 0.108 %

ASTM D6371 Cold filter plugging point - - -2 °C

ASTM D874 Sulphated Ash 0.02 max <0.01 %

ASTM D1160 Distillation temp @90% rec 360 max 316 °C

842

Fatty acid

C16:0 Palmitic acid

C18:0 Stearic acid

C18:1 (cis) Oleic acid

C18:1 (trans) Oleic acid

C18:2 Linoleic acid

11.2

3.5

64.4

2.7

18.3

SFAs

MUFA

PUFA

14.72

67.03

18.25

843

844

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

36

7.3 A3. Triacetin (T100) 845

846

Provided by: 847 REDOX Pty Ltd 848 Richlands, QLD 4077, Australia 849 850

851

852

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

1. Waste cooking biodiesel and triacetin (highly oxygenated additive) were used

2. Diesel during cold-start had lower NOx, PN, PM1 and PM2.5 than hot-start

3. Oxygenated fuels during cold-start had lower NOx, PN, PM1 and PM2.5 than hot-start

4. Oxygenated fuels during hot-start decreased PN, PM2.5 and PM1

5. Oxygenated fuels during cold-start increased PN and nucleation mode particles