University of Cincinnati

UNIVERSITY OF CINCINNATI

Date:______

I, ______, hereby submit this work as part of the requirements for the degree of: in:

It is entitled:

This work and its defense approved by:

Chair: ______

Composition and Formation Mechanism of Diesel Particulate Matter Associated with Various Factors from A Non-road Diesel Generator

A dissertation submitted to the

Division of Research and Advanced Studies of University of Cincinnati

In partial fulfillment of the requirements for the degree of

DOCTORATE OF PHILOSOPHY (Ph.D.)

in the Department of Environmental Engineering of the College of Engineering

2006

Fuyan Liang

B.S. (Environ. Eng.), Tsinghua University, P.R. China, 2000 M.S. (Environ. Eng.), Tsinghua University, P.R. China, 2002

Committee Chair: Mingming Lu, Ph.D. ABSTRACT

Diesel engines emit large quantities of fine particles to the atmosphere, on which numerous organic compounds are absorbed, such as PAHs, nitro-PAHs, and sulfur-containing PAHs. It is well established that exposure to PAHs and their derivatives may represent a high risk for human health. Considering its organic fractions and inhalable properties, diesel particulate matter (DPM) is considered a probable human carcinogen. The concerns of DPM health effects have simulated studies on DPM chemical composition and its formation mechanisms.

This dissertation presents the results of an experimental investigation into the composition and formation mechanism of DPM at various engine operating conditions and fuel sulfur contents. The influence of sampling methods also was examined. High volume dilution sampling and ultrasonic assisted extraction methods were developed for collection and extraction of DPM. Gas chromatography/mass spectrometry (GC/MS) with selective ion chromatogram (SIC), tandem mass spectrometry (MS/MS) with selected ion monitoring (SIM), and gas chromatography with atomic emission detection (GC/AED) were applied for determination of numerous hydrocarbons and organosulfur species in diesel emissions. The results of the comparisons between DPM extracts, diesel fuel, and engine oil indicate that diesel fuel, engine oil, and combustion process were the three major sources of alkanes, organic acids, and PAHs and alkylated PAHs in DPM, respectively. As expected, the distribution of organic compounds between gas and particle phase diesel emissions was directly correlated with their vapor pressures. Adsorption of gas phase organics onto diesel particles was the predominant mechanism controlling the formation of diesel aerosols. This research revealed strong effects from engine operation conditions and sampling methods on the organic composition and formation mechanism of DPM, as well as the compound distribution. Higher engine loads caused the increase in DPM emission rate, its elemental carbon fraction, and the heavier components, which were pyrogenic species, formed during combustion processes, and may present greater health risks. The comparison between the results for dilution method and stack sampling method demonstrated that the dilution process significantly increased the organic fraction of DPM through the condensation of gas-phase organics. The research outcomes provide important knowledge for understanding DPM nature and formation mechanisms.

ACKNOWLEDGEMENTS

First and foremost, I would like to express my deepest gratitude to my advisor Dr. Mingming Lu for her guidance, support, and friendship throughout my Ph.D. study. I am in eternal debt to her for the various ways in which she supported me. I am also very grateful to Dr. Tim C. Keener, Dr. M. Eileen Birch, and Dr. George Sorial for serving on my committee and for providing me with very valuable comments and feedback. Although they did not serve as members of my committee, I would like to thank Dr. Soon-Jai Khang and Dr. Sumana Keener for their advice and support.

I would like to acknowledge the National Institute for Occupational Safety and Health for the use of their instrument and facilities at the laboratory in Cincinnati, Ohio. In particular, I want to thank Dr. M. Eileen Birch for the loaning of the instrumentation used for my research, for her help, friendship, great advice, and technical support throughout the method development and sample measurements.

This work is dedicated to my parents and my brother in China. They have always been so supportive and loving. Thank you very much. Because having you in mind, I could finish this mission. I hope one day I can reward you for all the time I was far from you.

My acknowledgement is extended to my friends at UC: Zhangli Cai, Kai zhang, Qiang Zhang, Peng Jin, Qiuli Lu, Kessinee, and Phirun. I thoroughly enjoyed their friendship. To Zifei, for the help with sampling in cool and hot weathers, and for driving me wherever needed. To Jun, for his kindness and help with the GC/MS questions, problems, and skills.

But most especially this work is dedicated to my dearest Hao, who always showed me the power of unconditional love. With him I have walked most of my path here in Cincinnati. I cannot thank him enough for his love, his friendship, his support in the worst moments, his help with my work and for feeding me with the most delicious food for so many years. Without you I would have never accomplished this task.

Also I would like to thank the National Institute for Occupational Safety and Health for their financial support. TABLE OF CONTENTS

Chapter 1 INTRODUCTION...... 1

1.1 BACKGROUND & MOTIVATION ...... 1

1.2 LITERATURE REVIEW...... 4 1.2.1 Diesel Fuel Composition...... 4 1.2.2 Engine Oil Composition ...... 8 1.2.3 Overview of DPM Composition ...... 9 1.2.4 DPM Formation ...... 11 1.2.5 DPM Measurement Methods ...... 12 1.2.6 Effects of Testing Conditions on DPM...... 14

1.3 RESEARCH OBJECTIVES ...... 15

1.4 REFERENCES ...... 17

Chapter 2 EXPERIMENTAL DESIGN AND METHOD...... 24

2.1 EXPERIMENTAL DESIGN...... 24

2.2 SAMPLING METHOD...... 26 2.2.1 EPA Method 5...... 26 2.2.2 Natural Dilution ...... 28 2.2.3 High Volume Dilution Sampling...... 30

2.3 EXTRACTION METHOD ...... 35

2.4 ANALYTICAL METHOD ...... 37 2.4.1 OC/EC Analysis...... 37 2.4.2 Hydrocarbon Analysis with GC/MS...... 40 2.4.2.1 Instrument and Operating Parameters...... 40 2.4.2.2 Compound Identification and Quantification ...... 41 2.4.2.3 Standard Compounds...... 43 2.4.2.4 Quality Assurance and Quality Control...... 46

2.5 EXPERIMENT ON SOURCE CONTRIBUTION TO DPM COMPOSITION ...... 47

2.6 REFERENCES ...... 50

i Chapter 3 THE ORGANIC COMPOSITION OF DIESEL PARTICULATE MATTER, DIESEL FUEL AND ENGINE OIL OF A NON-ROAD DIESEL GENERATOR...... 56

3.1 INTRODUCTION ...... 56

3.2 EXPERIMENTAL METHOD ...... 57

3.3 RESULTS AND DISCUSSION ...... 57 3.3.1 Chemical Composition...... 58 3.3.2 Alkanes ...... 65 3.3.3 PAHs and Alkylated PAHs...... 67 3.3.4 Alkylbenzenes...... 69 3.3.5 Organic Acids ...... 70 3.3.6 Isomer Distribution ...... 71

3.4 CONCLUSIONS...... 73

3.5 REFERENCES ...... 73

Chapter 4 PHASE DISTRIBUTION OF SEMI-VOLATILE ORGANIC COMPOUNDS IN DIESEL COMBUSTION AEROSOLS...... 75

4.1 INTRODUCTION ...... 75 4.1.1 Gas/Particle Partitioning ...... 75 4.1.2 Partitioning Constant...... 76 4.1.3 Partitioning Mechanisms and Predictive Models ...... 77

4.1.4 Impact Factors on Kp Values ...... 79

4.2 EXPERIMENTAL METHOD...... 80

4.3 RESULTS AND DISCUSSION...... 81 4.3.1 Phase Distribution...... 81

0 4.3.2 Log Kp vs. log pL ...... 84

4.4 CONCLUSIONS...... 91

4.5 REFERENCES ...... 92

Chapter 5 FACTORS AFFECTING HYDROCARBON DISTRIBUTION IN DIESEL EMISSIONS ...... 96

ii 5.1 INTRODUCTION ...... 96

5.2 EXPERIMENTAL METHODS...... 99

5.3 RESULTS AND DISCUSSION...... 99 5.3.1 Effect of Engine Load...... 99 5.3.1.1 DPM versus Load ...... 99 5.3.1.2 Organic Composition versus Load...... 107 5.3.2 Effect of Sampling Method...... 121 5.3.2.1 Dilution Effect on DPM, OC and EC ...... 121 5.3.2.2 Dilution Effect on the Organic Composition of DPM...... 124 5.3.3 Effect of Fuel Sulfur Content...... 133 5.3.3.1 Effect of Fuel Sulfur Content on DPM, OC, and EC...... 133 5.3.3.2 Effect of Fuel Sulfur Content on the Organic Composition135

5.4 CONCLUSIONS...... 135

5.5 REFERENCES ...... 137

Chapter 6 DETERMINATION OF POLYCYCLIC AROMATIC SULFUR HETEROCYCLES IN DIESEL FUEL AND DIESEL PARTICULATE MATTER BY GAS CHROMATOGRAPHY WITH ATOMIC EMISSION DETECTION...... 142

6.1 INTRODUCTION ...... 142

6.2 EXPERIMENTAL METHOD ...... 144 6.2.1 Sampling and Extraction Method ...... 144 6.2.2 Analytical Method ...... 144 6.2.3 Analytical Standards ...... 145

6.3 RESULTS AND DISCUSSION ...... 147 6.3.1 AED Response Factor...... 147 6.3.2 PASH Speciation in Diesel Fuel...... 149 6.3.3 PASHs in Diesel Emissions...... 154 6.3.3.1 PASH Distribution between Gas and Particle Phase ...... 158 6.3.3.2 Total Organic Sulfur in DPM ...... 159 6.3.3.3 Effect of Fuel Sulfur on DPM PASH Distribution ...... 162

iii 6.3.3.4 Effect of Engine Load on DPM PASH Distribution...... 163

6.4 SUMMARY AND CONCLUSIONS...... 166

6.5 REFERENCES ...... 167

Chapter 7 CONCLUSIONS AND RECOMMENDATIONS...... 170

7.1 CONCLUSIONS...... 170

7.2 RECOMMENDATIONS...... 174

7.3 REFERENCES ...... 178

Appendix A CHROMATOGRAMS FOR SAMPLES ...... 179

Appendix B DATA FOR DIESEL FUEL COMPONENT...... 186

Appendix C DATA FOR DIESEL EMISSIONS ...... 188

Appendix D PASH DATA FOR DIESEL FUEL AND DIESEL EMISSIONS ...... 204

Appendix E DATA FOR PHENANTHRENE-ENRICHED FUEL...... 207

iv LIST OF TABLES

Table 1-1 Hydrocarbon specifications of some commercial diesel fuel oils [32]...... 5

Table 2-1 Performance characteristics of high volume dilution sampler...... 34

Table 2-2 The selected quantitative fragment ions for alkylated PAHs and alkylbenzenes .... 42

Table 2-3 GC/MS/MS segment program for 16PAHs ...... 43

Table 2-4 List of standard compounds and their chemical properties...... 44

Table 3-1 Organic compounds present in low sulfur diesel fuel, engine oil and diesel particulate matter...... 60

Table 3-2 Ratio of 1-MN/2-MN and 1-MPh/2-MPh in diesel fuel and DPM...... 72

Table 4-1 Vapor pressures for some alkanes, PAHs and alkanoic acids at 25 ºC (torr)...... 82

p 0 Table 4-2 Values of mr and br in the equation of log Kp = mr log L + br at 25 ºC for alkanes and PAHs based on the experimental data and model estimations...... 86

p 0 Table 4-3 Literature values of mr and br in the equation of log Kp = mr log L + br for gas/particle partitioning at urban areas...... 87

Table 4-4 Activity coefficients for n-alkanes, PAHs and alkanoic acids from literature...... 88

Table 5-1 DPM emission rate from the non-road diesel generator at 0-75 kW (LSDF, dilution sampling)...... 100

Table 5-2 OC and EC emission rate (in mg/m3 and mg/hr) from the non-road diesel generator at 0-75 kW (LSDF, dilution sampling)...... 102

Table 5-3 Comparison of OC and EC percentage in total DPM mass from this study with the results from other studies...... 104

Table 5-4 Percentage of insoluble, soluble, unidentified and identified fraction in total DPM mass at 0-75 kW (LSDF, dilution sampling)...... 105

v Table 5-5 The concentration of compounds in DPM collected with EPA Method 5, the natural dilution method, and the dilution method (DR=3.4) (LSDF)...... 125

Table 5-6 Fractions of organic compounds, EC, and non-carbonaceous materials in DPM at low and high sulfur diesel fuel (dilution sampling)...... 134

Table 6-1 GC/AED operating parameters...... 145

Table 6-2 List of thiophene standards and related information...... 146

Table 6-3 Sulfur response factors for three sulfur compounds. Results are based on five injections...... 148

Table 6-4 PASHs in low and high sulfur diesel fuels (LSDF and HSDF)...... 151

Table 6-5 PASHs in gas and particle phase diesel emissions (0 kW through 75 kW) when burning low and high sulfur diesel fuels...... 156

Table 7-1 Results for phenanthrene survivability experiment...... 176

Table 7-2 Literature results for contributions from unburned fuel and pyrosysnthesis to PAHs in diesel emissions...... 176

vi LIST OF FIGURES

Figure 1-1 Typical structure of diesel particulate matter (Kittelson, 1998 [59])...... 10

Figure 2-1 Diesel emission source (diesel generator and load simulator)...... 25

Figure 2-2 Schematic of an EPA Method 5 sampling train...... 27

Figure 2-3 Schematic of the natural dilution sampler for DPM collection...... 29

Figure 2-4 Dilution stack sampler schematic diagram designed by Hildemann et al. [15]...... 32

Figure 2-5 Schematic of high volume dilution sampler for particulate and gas phase diesel emissions...... 33

Figure 2-6 Thermogram for filter sample containing organic carbon (OC), carbonate (CC), and elemental carbon (EC). PC is pyrolytically generated carbon or ‘char.’ Final peak is methane calibration peak. Carbon sources: pulverized beet pulp, rock dust (carbonate), and diesel particulate (from reference [3])...... 39

Figure 3-1 Total ion chromatogram of diesel fuel and DPM (time in minutes)...... 59

Figure 3-2 Concentration and percentage of each category of compounds in low sulfur diesel fuel and DPM...... 63

Figure 3-3 Relative percentage of alkanes in total identified compounds for low sulfur diesel fuel and DPM...... 66

Figure 3-4 Chemical structures of isoprenoids present in diesel fuel and diesel emissions...... 67

Figure 3-5 Relative percentage of PAHs in total identified compounds in low sulfur diesel fuel and DPM...... 68

Figure 3-6 Relative percentage of alkylated PAHs in total identified compounds in low sulfur diesel fuel and DPM...... 68

Figure 3-7 H/C ratio vs. PAH molecular mass found by GC/MS in low sulfur diesel fuel and DPM...... 69

Figure 3-8 Relative percentage of alkylbenzenes in total identified compounds in low sulfur diesel fuel and DPM...... 70

vii Figure 3-9 Relative percentage of n-alkanoic acids from C6H12O2 to C18H36O2 in DPM and engine oil...... 71

Figure 3-10 Relative percentage of isomers in low sulfur diesel fuel and DPM...... 72

Figure 4-1 The distribution of n-alkanes in gas and particle phase diesel emissions (400 ppmw S, 0 kW, DR = 3.4)...... 81

Figure 4-2 The distribution of PAHs in gas and particle phase diesel emissions (400 ppmw S, 0 kW, DR = 3.4)...... 83

Figure 4-3 The distribution of alkanoic acids in gas and particle phase diesel emissions (400 ppmw S, 0 kW, DR = 3.4)...... 84

0 3 p Figure 4-4 Log Kp (m /µg) vs. log L (torr, @ 25 ºC) based on experimental measurements and estimations for n-alkanes in diesel emissions at 0-75 kW...... 85

0 3 p Figure 4-5 Log Kp (m /µg) vs. log L (torr, @ 25 ºC) based on experimental measurements and estimations for PAHs in diesel emissions at 0-75 kW...... 85

Figure 4-6 Sources contributing to the variability in mr and br values...... 90

Figure 5-1 Correlation between DPM mass concentration and diesel engine load (LSDF, dilution sampling)...... 101

Figure 5-2 The concentration of organic and elemental carbon versus diesel engine loads (LSDF, dilution sampling)...... 102

Figure 5-3 Percentage of organic compounds, elemental carbon, and non-carbonaceous materials in DPM at 0-75 kW (LSDF, dilution sampling)...... 103

Figure 5-4 Percentage of each identified category in total identified particle phase compounds at 0-75 kW (LSDF, dilution sampling)...... 106

Figure 5-5 The concentration of n-alkanes from n-decane (C10) to n-pentacosane (C25) in gas and particulate phase diesel emissions at 0-75 kW (LSDF, dilution sampling).... 108

Figure 5-6 The concentration of branched alkanes in gas and particulate phase diesel emissions at 0-75 kW (LSDF, dilution sampling)...... 110

Figure 5-7 The concentration of cycloalkanes from C13 to C25 in gas and particulate phase diesel emissions at 0-75 kW (LSDF, dilution sampling)...... 112

viii Figure 5-8 The concentration of EPA 16 priority PAHs in gas and particulate phase diesel emissions at 0-75 kW (LSDF, dilution sampling)...... 114

Figure 5-9 The concentration of alkylated PAHs in gas and particulate phase diesel emission at 0-75 kW (LSDF, dilution sampling)...... 117

Figure 5-10 The concentration of alkylbenzenes in gas and particulate phase diesel emission at 0-75 kW (LSDF, dilution sampling)...... 117

Figure 5-11 The concentration of alkanoic acids from C6H12O2 to C18H36O2 in diesel emissions at 0-75 kW (LSDF, dilution sampling)...... 118

Figure 5-12 Normalized emission rates for identified compounds expressed in terms of mass of compound in the exhaust/mass of compound from the fuel input (LSDF, dilution sampling)...... 120

Figure 5-13 Comparison of DPM mass concentration between EPA Method 5 and the dilution method with a dilution ratio of ~3.4 (LSDF)...... 122

Figure 5-14 OC and EC concentration in DPM collected with EPA Method 5 and the dilution method (LSDF)...... 123

Figure 5-15 Fractions of organic compounds, EC, and non-carbonaceous materials in DPM collected with EPA Method 5 and the dilution method (LSDF)...... 124

Figure 5-16 The percentage of individual alkanes in total identified (a) n-alkanes, (b) branched alkanes, and (c) cycloalkanes in DPM collected with EPA method 5, the natural dilution method, and the dilution method with a dilution ratio of ~3.4 (LSDF, 75 kW)...... 130

Figure 5-17 The percentage of individual aromatics in total identified (a) PAHs, (b) alkylated PAHs, and (c) alkylbenzenes in DPM collected with method 5, the natural dilution, and the dilution method with a dilution ratio of ~3.4 (LSDF, 75 kW)...... 131

Figure 5-18 The percentage of individual alkanoic acids in total alkanoic acids in DPM collected with method 5, the natural dilution method, and the dilution method with a dilution ratio of ~3.4 (LSDF, 75 kW)...... 133

Figure 5-19 Ratios of the concentration of DPM, OC, EC, and non-carbonaceous materials at high sulfur diesel fuel over the concentration at low sulfur diesel fuel (dilution sampling)...... 134

ix Figure 6-1 Structures of PASH standard compounds...... 147

Figure 6-2 Carbon (179 nm) and sulfur (181 nm) AED chromatograms for low sulfur (433 ppmw) diesel fuel...... 150

Figure 6-3 Carbon (179 nm) and sulfur (181 nm) AED chromatograms for high sulfur (2284 ppmw) diesel fuel...... 150

Figure 6-4 Possible alkylated PASH groups in low and high sulfur diesel fuels (LSDF and HSDF, respectively)...... 152

Figure 6-5 PASHs in low and high sulfur diesel fuels (LSDF and HSDF, respectively) expressed in (a) µg S/g DF, and (b) mg S/g TS in DF...... 153

Figure 6-6 Sulfur (181 nm) AED chromatograms of gas and particulate phase diesel emissions generated at 0 kW with low sulfur (433 ppmw) diesel fuel...... 155

Figure 6-7 Organosulfur compounds in gas phase and particulate phase diesel emissions at 0 kW and low sulfur diesel fuel...... 159

Figure 6-8 Total organic sulfur in particulate phase and gas + particulate phase diesel emissions for LSDF and HSDF at various load conditions...... 160

Figure 6-9 Conversion rate of fuel sulfur to the emitted organic sulfur for HSDF and LSDF at different engine loads...... 161

Figure 6-10 Organosulfur distribution in DPM under 0 kW and 75 kW for LSDF (L) and HSDF (H)...... 163

Figure 6-11 Sulfur (181 nm) AED chromatograms of gas and particulate phase diesel emissions generated at 75 kW with low sulfur (433 ppmw) diesel fuel...... 164

Figure 6-12 PASH concentration, expressed in terms of µg S/g DPM and µg S/kg DF, in DPM generated with HSDF vs. engine loads...... 165

x GLOSSARY OF ACRONYMS AND SYMBOLS

A Gas phase concentration, ng/m3 ASTM American Standard Testing Method 2 ATSP Specific surface area of suspended particles, cm /µg BC Black Carbon 0 br Intercept in a regression of log Kp vs. log pL BSTFA N,O-bis[trimethylsilyl]trifluoroacetamide

C1 Concentration of phenanthrene in normal fuel, mg/L

C2 Concentration of phenanthrene in phenanthrene-enriched fuel, mg/L CC Carbonate Carbon

CDF Compound concentration in diesel fuel, g/kg DF CF Carbonaceous Fraction

CH4 Methane CIC Compound Independent Calibration

CO2 Carbon dioxide

CPh Concentration of phenanthrene in a fuel used, mg/L CPM Condensable Particulate Matter 3 CS Compound concentration in diesel emissions (gas + particulate phase), µg/m DCM Dichloromethane DHHS Department of Health and Human Services DPM Diesel Particulate Matter DR Dilution Ratio EC Elemental Carbon EPA Method 5 Determination of Particulate Emissions from Stationary Sources

ER1 Mg compound emission/kg of fuel combusted, mg/kg DF

ER2 Mg compound emission/g compound in diesel fuel, mg/g F Particle phase concentration, ng/m3 3 F1 Concentration of phenanthrene in DPM for normal fuel, ng/m 3 F2 Concentration of phenanthrene in DPM for phenanthrene-enriched fuel, ng/m FID Flame Ionization Detection fom Weight fraction of organic matter on the TSP 3 FPh Concentration of phenanthrene in DPM, ng/m G/P Gas/Particle GC/AED Gas Chromatography with Atomic Emission Detection GC/MS Gas Chromatography with Mass Spectrometry HEPA High efficiency particulate air filter

xi HPLC High Performance Liquid Chromatography HSDF High Sulfur Diesel Fuel IARC International Agency for Research on Cancer 3 Kp Gas-particle partitioning coefficient, m /µg LSDF Low Sulfur Diesel Fuel 0 mr Slope in a regression of log Kp vs. log pL MS/MS Tandem Mass Spectrometry

MWom Mean molecular weight of the obsorbing organic matter, g/mol ND Not Detected NIOSH National Institute for Occupational Safety and Health NIST National Institute of Standards and Technology

NOx Nitrogen oxides (NOx = NO + NO2) 2 Ns Surface concentration of sorption sites for adsorbing surface, mol/cm OC Organic Carbon PAHs Polycyclic Aromatic Hydrocarbons PASHs Polycyclic Aromatic Sulfur Heterocycles PC Pyrolytically generated carbon 0 pL Vapor pressure of the pure compound at the temperature of interest, torr PLE Pressurized Liquid Extraction PM Particulate Matter ppmv Parts per million, by volume ppmw Parts per million, by weight

PUBDF The percentage of phenanthrene contributed by unburned diesel fuel in the total particulate phenanthrene emission, % PUF Polyurethane foam

Q1 Enthalpy of desorption from adsorbing surface, kcal/mol

Q2 Enthalpy of volatilization of the pure liquid, kcal/mol 3 QS Stack gas flow rate at standard conditions, m /hr R Gas constant, (= 8.314×10-3 kJ/mol·K = 0.00199 kcal/mol·K = 8.2×10-5 m3·atm/mol·K) RCD Respirable Combustible Dust

RDF Fuel consumption rate, L/hr RPD Relative Percent Difference RSD Relative Standard Deviation SFE Supercritical Fluid Extraction SIC Selective Ion Chromatogram SIM Selected Ion Monitoring

SO2 Sulfur dioxide

xii SO3 Sulfur trioxide 2- SO4 Sulfate SOF Soluble Organic Fraction SRM Standard Reference Material SS Size Selective SVOCs Semi-Volatile Organic Compounds T Temperature, K TBDS tert-butyl disulfide TC Total Carbon TIC Total Ion Chromatogram TOS Total Organic Sulfur TSP Total Suspended Particulate Matter, µg/m3 UHC Unburned hydrocarbons US EPA United States Environmental Protection Agency WHO World Health Organization γ Activity coefficient of a compound in the organic matter on the mole fraction scale µg Micrograms, 1 ×10-6 g

ρDF Density of diesel fuel, 800 g/L

MN Methylnaphthalene DMN Dimethylnaphthalene TMN Trimethylnaphthalene MPh Methylphenanthrene DMPh Dimethylphenanthrene MT Methylthiophene DMT Dimethylthiophene ET Ethylthiophene PT Propylthiophene BT Benzothiophene MBT Methylbenzothiophene DMBT Dimethylbenzothiophene TMBT Trimethylbenzothiophene TTMBT Tetramethylbenzothiophene DBT Dibenzothiophene MDBT Methyldibenzothiophene DMDBT Dimethyldibenzothiophene TMDBT Trimethyldibenzothiophene Ph45T Phenanthro[4,5-bcd]thiophene BN12T Benzo[b]naphtho[1,2-d]thiophene

xiii BN21T Benzo[b]naphtho[2,1-d]thiophene BN23T Benzo[b]naptho[2,3-d]thiophene Ph34T Phenanthro[3,4-b]thiophene BPh9,10T Benzo[b]phenanthro[9,10-d]thiophene DiAT Diacenaphthothiophene

xiv Chapter 1

INTRODUCTION

1.1 Background & Motivation

Non-road diesel engines are extensively used in various applications, such as construction, power generation, underground mines, agriculture, and so on. It is reported that in the United

States, 103,490 units of generator sets rated between 56 and 130 kW were sold from the years

1996 to 2000, which represented 11% of the total sales volume (905,000) of all non-road diesel equipment sales of that size during that period [1]. However, the emissions of diesel particulate matter (DPM) from non-road sources are significant [2-4]. According to the United States

Environmental Protection Agency (US EPA), 44% of total DPM and 12% of total NOx emissions from mobile sources are emitted from non-road diesel vehicles [5].

Roadway studies have shown higher DPM contributions than ambient measurements acquired at compliance monitoring sites. At a Manhattan bus stop in New York City, with idling buses and heavy bus traffic, DPM contributions were in the range of 13-47 µg/m3 [6]. In-vehicle emissions of DPM are responsible for a large portion of human exposure. Fruin et al. [7] measured average black carbon (BC) concentrations in Los Angeles, finding ~5 µg/m3 when no vehicles in front, and ~130 µg/m3 when following an urban transit bus making frequent stops.

Occupational DPM exposures are also high. DPM exposures for mineworkers range from

10 to 1280 µg/m3 [8-10]. DPM exposures for those working near diesel-powered forklifts averaged 31 µg/m3 BC as reported by Zaebst et al. [11]. In other job categories, average personal

1 BC exposures were 27 µg/m3 for mechanics and 5.2 µg/m3 for truckers [11]. These levels greatly exceeded the residential background BC concentrations of 1.1 µg/m3.

Exposures to diesel exhaust have resulted in adverse human health effects due to the hazardous composition of diesel emissions. Studies have indicated that occupational exposure to diesel exhaust may be related to lung cancer for bus garage workers, miners, and forklift workers, and others [3, 12]. DPM consists of a large number of organic compounds [12-14], among which some polycyclic aromatic hydrocarbons (PAHs), such as chrysene and benzo[a]pyrene, nitro-

PAHs and oxygenated PAHs are known or suspected human carcinogens [15-20]. It is well established that exposure to PAHs and their derivatives may result in a higher health risk [21].

This is due to the small size and toxic composition of DPM, as approximately 90% of the DPM mass is within the inhalable range (<1 µm) and the organic compounds tend to be adsorbed into those fine particles that penetrate deep into the lungs.

Although further characterization of organic speciation is on-going, the hazardous effects of DPM emissions have been confirmed and regulated. In 1988, the National Institute for

Occupational Safety and Health (NIOSH) first recommended that diesel exhaust be regarded as a potential occupational carcinogen [22]. In 1989, the International Agency for Research on

Cancer (IARC) declared that DPM was possibly carcinogenic to humans [23]. The U.S.

Environmental Protection Agency (EPA) has classified diesel exhaust as a probable human carcinogen [24]. The World Health Organization (WHO) [25], the California Environmental

Protection Agency [26], and the U.S. Department of Health and Human Services (DHHS) [27] have also identified DPM as a probable human carcinogen.

Considering the adverse health effects of DPM emissions, significant efforts have been made on studying DPM emissions from on-road diesel vehicles. However, there have been few

2 studies on the chemical composition of DPM emissions from non-road diesel engines, even though the use of non-road diesel engines has increased rapidly and the resultant DPM emissions are of significant quantity. Unlike diesel vehicles, diesel generators run at a fixed rpm, and the amperage and voltage it produces can vary with the load [28]. The difference in the operating mode can potentially result in different emission characteristics. Therefore, the aim of this study is to characterize the organic composition of DPM emissions, especially carcinogenic polycyclic aromatic hydrocarbons (PAHs) and polycyclic aromatic sulfur heterocycles (PASHs), from a non-road diesel generator, and to investigate the source of contributions from fuel, combustion, engine oil, and other possible sources.

In addition, it is essential to consider engine testing conditions in determining the chemical composition of DPM emissions. It is generally accepted that diesel emissions are influenced by the engine testing conditions: types of operation (steady-state or transient), fuel quality (high or low sulfur), and after-treatment devices. Application of new technologies for reducing particle mass emissions can also change the nature of the particles, such as increasing the volatile fraction [29]. The above-mentioned considerations promote the understanding on the impact of fuel composition, engine operation conditions, and emission control technologies on the chemical composition of DPM for developing effective emission control strategies. Therefore, this research also aims to study DPM composition and formation mechanisms as a function of fuel properties and engine operation conditions.

Research on emissions from diesel engines has shown that sampling conditions can alter measured DPM size distributions [30]. Many studies have shown that DPM remains in a state of flux for some time after it is emitted to the atmosphere. During this state, coagulation, adsorption, and condensation of organics and inorganics which are usually present in diesel exhaust with

3 significant quantities are continuously occurring. The fate of these condensable compounds is significantly affected by dilution of the exhaust stream. Dilution sampling is a technique that has been developed to examine the influence of rapid cooling and dilution on PM emissions from combustion systems. Although a dilution sampler cannot simulate the actual atmospheric mixing, it allows examination of the effects of dilution on PM emissions in order to better understand the

PM transformations that occur in the atmosphere. Another objective of this study is to investigate the effects of dilution on DPM chemical composition and compound distribution by comparing the composition of DPM collected with dilution and stack sampling methods.

1.2 Literature Review

1.2.1 Diesel Fuel Composition

Diesel fuel consists mainly of saturated and aromatic hydrocarbons [31, 32]. Paraffins, cycloparaffins, monoaromatics, diaromatics, polynuclear aromatics and sulfur compounds are the major compound types typically found in diesel fuel [33]. Their relative distribution depends on the feedstock and fuel processing schemes. Table 1-1 listed the composition of different types of diesel fuels. Sjogren et al. [13] studied the chemical composition and physical characteristics of ten diesel fuels and obtained similar results. Generally, diesel fuel consists of 65-85% saturates,

5-30% aromatics, and 0-5% olefins. But the percentage of each category will be different for different types of diesel fuel.

Alkanes: Normal, branched and cyclic alkanes are the most abundant components of diesel fuel, which account for about 65-85% or even higher (Table 1-1). In general, diesel fuels contain alkanes from C5 to C30. Song stated that long-chain alkanes with carbon numbers in the range of 10-20 are the major paraffinic components [31]. There can be some lighter (C9-) and

4 heavier (C20+) alkanes, but they exist in small quantities in most high-way diesel fuels. However, for different types of diesel fuels, the range of carbon numbers is different. According to statistics of the World Health Organization, C10-C28 alkanes are the dominant, saturated hydrocarbons for general diesel fuel (i.e., diesel fuel No. 2), C9-C16 for diesel fuel No. 1, and C10-

C30 for diesel fuel No. 4 [32]. Some other studies have reported similar results [34].

Table 1-1 Hydrocarbon specifications of some commercial diesel fuel oils [32]. Specification Diesel fuel a Kerosene b Distillated marine diesel c Diesel fuel d Saturates (volume%) 65-95 78-96 60-90 59.4-76.6 Aromatic (volume%) 5-30 4-25 5-40 23.4-39.6 Olefins (volume%) 0-10 0-5 0-5 0-1.0 a From CONCAWE (1985); fuel oil similar to diesel fuel (general). b From CONCAWE (1985, 1995); fuel oil similar to diesel fuel No. 1. c From CONCAWE (1985); fuel oil similar to diesel fuel No. 4. d From German Scientific Association for Petroleum, Natural Gas, and Coal (1991); three samples of diesel fuel (general).

Branched alkanes are important in diesel fuels due to their large percentage and the characteristics of some specific species as tracers, which are referred to as isoprenoids. The regular isoprenoids, which have a head-to-tail structure, are the most abundant of this class of compounds. These isoprenoids are naturally present in crude oil [35] and therefore would be expected to be found in diesel fuel. The isoprenoids can be used as tracers for diesel engine exhaust in conjunction with elemental carbon, hopanes, and steranes. The latter two compounds classed are common markers of engine oil. The ratios of pristine to heptadecane and of phytane to octadecane will help to identify the source of a fuel spill; furthermore, they can be used to estimate the age of an environmental contamination and the degree of elimination as these ratios increase during biological degradation [32].

5 Alkylated cycloalkanes are also present in diesel fuels. The percentage of cycloalkanes is relatively low compared with normal and branched alkanes. Alkylated cyclopentane, cyclohexane, and cycloheptane are common components [36, 37].

Aromatics: Aromatic compounds constitute 5-30% of automotive diesel fuel, 5-40% of marine diesel fuel, and 10-30% of diesel fuel No. 2 [38]. Alkyl benzenes are common components of diesel fuel. Polycyclic aromatic hydrocarbons (PAHs), e.g. naphthalene, phenanthrene, acenaphthene, acenaphthylene, fluorene and fluoranthene, are also present, as well as alkyl- and cycloalkyl- substituted homologues of these substances. Generally naphthalene and its methyl-substituted derivatives are predominant species. Trace amounts of heavy PAHs (3+- ring aromatic compounds) such as chrysene, pyrene, benzanthracene and perylene can also be present. The PAH content of diesel fuels varies widely.

Some species of PAHs are know or potential carcinogens. For instance, benzo(a)pyrene is proven to be related to lung cancer [39]. As a result, seven PAHs have been listed in the 1999

EPA National Toxics Inventory. Those PAHs include benz(a)anthracene, Benzo(a)pyrene,

Benzo(b)fluoranthene, Benzo(k)fluoranthene, Chrysene, Dibenz(a,h)anthracene, and

Indeno(1,2,3-cd)pyrene.

Sulfur Content: The sulfur content of diesel fuel depends on the source of crude oil. For most diesel fuels, sulfur content is 0.1-0.5% by weight. Currently, the U.S. EPA is proposing stringent regulations for sulfur content in diesel fuels. For highway diesel fuel, fuel sulfur content was reduced from 500 ppmv to 15 ppmv in June 1, 2006 [40]. For the non-raod applications, the current EPA regulation is 3400 ppmv, but it will be reduced to 500 ppmv in 2007 and further to

15 ppmv in 2010 [41].

6 A variety of sulfur compounds are present in diesel fuels, which include mercaptans, sulfides, disulfides, cyclic sulfides, alkyl sulfates, sulfonic acids, sulfoxides, sulfones and thiophenes [34, 42]. Among these compounds, benzothiophene, dibenzothiophene (DBT) and their alkylated homologue are the most abundant species in diesel fuel [43, 44]. Sulfur in diesel fuels is known to contribute to DPM emissions [45]. During combustion processes, fuel sulfur is oxidized to SOx, while the majority of which is SO2. In the exhaust tail pipe, when exhaust gases are cooled, sulfates, sulfuric acid, bound water and hydrocarbons with relatively low vapor pressures will condense onto soot and increase DPM growth. It is reported that the amount of sulfates adsorbed will affect the mass of hydrocarbons in the particles [46], including PAHs, due to the formation of heavy hydrocarbons in the condensed phase from the reaction of adsorbed sulfuric acid with organic compounds in the exhaust [47]. Neeft et al. [48] reported that 7-12% particulate reduction can be obtained with a reduction of fuel sulfur content from 0.2% to 0.05%.

Baranescu et al. [45] evaluated the effect of fuel sulfur on the brake specific particulate emissions on medium and heavy-duty trucks using three diesel fuels with 0.05%, 0.19%, and

0.29% sulfur. The results indicated that for an increase of 0.1% in fuel sulfur, particulates increased by about 0.025 g/bhp-h.

Fuel additives: Additives are chemicals introduced in very small proportions to the fuel to improve performance, enhance its desirable characteristics and to reduce the undesirable ones.

Some additives function by physical interactions with fuel components, while others are involved in chemical reactions. There are various additives designed to improve certain properties of diesel fuels, such as cold flow improvers (ethylene vinyl acetate, polyolefin ester, polyamide), antioxidants (2,6-di-tert-butyl-4-methyl-phenol), metal deactivators, corrosion

7 inhibitors (alkyl phosphate), dispersants (polyamides, amines), detergents (amines, amides and imidazones), lubricity agents, and demulsifiers [49-52].

1.2.2 Engine Oil Composition

Engine oil is one of the most important liquid lubricants. Engine oils are made from base oils and a set of additives. It is recognized that the performance of engine oils is sensitive to the properties of base oils and additives used. The known lubricating oil additives include oxidation inhibitors, detergents, dispersants, corrosion inhibitors, rust inhibitors, viscosity index improvers, depressants, and foam inhibitors. As mentioned above, some organic compounds in DPM may be emitted from the evaporation of engine oils. However, there have been limited studies on the composition of engine oils since the dominant components of engine oils are heavy compounds and complex additives.

Kohler and Heeb [53] characterized ageing products of ester-based synthetic lubricants by liquid chromatography with electrospray ionization mass spectrometry and by electrospray ionization (tandem) mass spectrometry. They found liquid chromatography-electrospray ionization mass spectrometry (LC-ESI-MS) to be a powerful technique to study the chemical composition and the molecular distribution of homologues. In their results, the mixed esters of

C5, C7 and C9 carboxylic acids and pentaerythritol were observed in new lubricants. In used jet engine oils, pentaerythritol tetraesters and dipentaerythritol hexaesters were formed. Netten and

Leung [54] studied the constituents of two jet engine lubricating oils and their volatile pyrolytic degradation products using GC/MS. They identified some esters of carboxylic acids and phosphoric acid. However, in both studies the identification of some compounds was done using the NIST library rather than standard compounds. Also the compounds were not quantified.

1.2.3 Overview of DPM Composition

Diesel particulate matter consists of highly agglomerated solid carbonaceous material and ash, as well as organic and sulfur compounds. Carbon in the fuel is mostly oxidized during combustion with the residue exhausted in the form of solid carbon. A small fraction of the fuel and evaporated lubricant oil escape oxidation and appear as soluble organic fraction (SOF). A major proportion of SOF in DPM comprises high-molecular-weight semi-volatile organic compounds which cannot be resolved by gas chromatography. This portion is composed predominantly of branched and cyclic hydrocarbons. The resolvable portion of SOF contains alkanes, PAHs, organic acids, and can also contain some hetero-atoms such as oxygen, nitrogen, and sulfur. The SOF varies with engine design and operating condition, ranging from less than

10% to more than 90% by mass [55]. Most of the sulfur in the fuel is oxidized to SO2, but a small fraction is oxidized to SO3 that leads to sulfuric acid and sulfate aerosol. Metal compounds in the fuel and lubricant oil lead to a small amount of inorganic ash. Figure 1-1 illustrates the structure of DPM.

Many studies have been conducted on the measurements of diesel emissions from on- road diesel vehicles. Rogge et al. [56] characterized the organic particulate matter in gasoline- and diesel-powered vehicle exhaust by GC/MS. In their study, more than 100 organic compounds were quantified, including n-alkanes, n-alkanoic acids, benzoic acids, benzaldehydes,

PAHs, oxy-PAHs, steranes, pentacyclic triterpanes, nitrogen-containing compounds, and others.

Among these compounds, n-alkanes composed the main portion (60.8%) of the identified resolved organic mass, and n-alkanoic acids constituted the second largest fraction. Schauer et al.

[57] measured both gas and particle phase emissions from medium duty diesel trucks using a

9 dilution source sampling system. Their results indicated that C1-C13 carbonyls accounted for 60% of the gas phase organic compound mass emissions. In the particulate diesel emissions, PAHs and alkylated PAHs were the major proportion (44%).

Figure 1-1 Typical structure of diesel particulate matter (Kittelson, 1998 [58]).

Among all the organic compounds identified in diesel emissions by various studies, ,

PAHs are important due to their carcinogenic and mutagenic properties and many efforts have been put on the concentration, formation and ambient behavior of PAHs in diesel particles. In the studies by Rogge et al. [56] and Schauer et al. [57], alkylated 2- to 3-ring PAHs were the most abundant PAHs in particle phase diesel emissions. Heavy PAHs with more than 4 rings, such as chrysene, benz[a]anthracene, and benzo[a]pyrene, were also present in particulate diesel emissions. Zielinska et al. [59] studied phase distribution of PAHs in diesel vehicle emissions, and found that semivolatile PAHs and nonvolatile 4- to 6-ring PAHs were present predominantly

10 on particles. They also found that the gas/particle phase distribution of semivolatile PAHs depended on the engine loads.

In addition to PAHs, nitro-PAHs (N-PAHs) and oxygenated PAHs (O-PAHs) have also been linked with increases in mutagenicity. Oxygenated PAHs have been identified in particles emitted from diesel and gasoline engines [60, 61]. Nitro-PAH are of particular interest due to their higher mutagenic (2 × 105 times) and carcinogenic (10 times) properties of certain compounds compared to PAHs [62]. Many nitro-PAHs have been found in both air particulate materials and diesel particulate matter [63, 64]. Typically nitro-PAHs are found at low µg/g even ng/g concentration levels in diesel particulate matter [63, 64].

1.2.4 DPM Formation

The knowledge of the nature and relative influence of processes affecting diesel exhaust components is critical to understand the important differences between diesel aerosol measured in the atmosphere and in the laboratory. Many processes affect the formation, behavior, and measurement of diesel aerosols, including condensation/adsorption, evaporation/desorption, coagulation, and chemical reactions. In addition to these mechanisms, dilution ratio and residence time can also affect particulate behavior.

Particle-to-Particle Processes [55]: Important particle-to-particle processes leading to losses in an exhaust and sampling system include the following processes. (1) Thermophoresis, which is the motion of a particle from asymmetrical forces that arise from a temperature gradient, is important for systems with long sampling lines and cooled exhaust that promote thermophoretic deposition. (2) Diffusion of particles may occur from Brownian motion or the movement of particles down a particle concentration gradient, and is important for particles

11 smaller than 0.05 µm in diameter. (3) Coagulation is the process of agglomeration from particle- to-particle collisions. Coagulation changes the size distribution of particles, but does not change the particle mass and compound concentration.

Gas-to-Particle Conversions [55]: The amount and fate of volatile organic compounds are important in characterizing diesel aerosols, since volatile organics are associated with the smallest diesel particles (nano-particles). Volatile organics may remain as gaseous organics or may be converted to particle phase and form the soluble organic fraction (SOF). Gas-to-particle conversion of volatile organics may occur through two paths: adsorption on existing particles or nucleation to form new particles. Adsorption of volatile organics onto diesel particles can occur at low vapor pressures of adsorbed compounds. The amount of organics adsorbed on diesel particles depends on the saturation ratios for the organic compounds, the surface area of diesel particles available for adsorption, adsorption energy, and the residence time. Nucleation of gaseous organic compounds occurs when saturation ratios are sufficiently high. Nucleation causes the formation of the nuclei-mode particles in diesel aerosols, which primarily consist of volatile materials. Nuclei mode is often observed when diesel exhaust is rapidly diluted.

Condensation is an extension of nucleation and adsorption, which requires supersaturation and existing particles.

1.2.5 DPM Measurement Methods

Three methods are routinely used in mines to determine the concentration of diesel particulate matter (DPM): the respirable combustible dust method (RCD), the size selective method (SS), and the elemental carbon method (EC) [65]. In the RCD method, the RCD is the amount of materials burned off from the silver membrane, which is used to collect respirable

12 dust, by controlled combustion. RCD is composed of all combustible materials collected on a filter, including EC and the soluble organic fraction of DPM. In the SS method, it is assumed that particles with diameters below 0.8 µm originate from diesel emissions. In the SS sampler, a personal sampling pump draws air through a cyclone followed by an inertial impactor with a 0.8

µm cut point. DPM, which is mostly smaller than 0.8 µm, is collected on a polyvinyl chloride filter (or glass fiber filter), and is determined gravimetrically. In the EC method, samples are collected with or without an inertial pre-selector to remove particles larger than 0.8 µm. The thermal-optical method (NIOSH 5040) is used to determine carbonaceous aerosols, speciation of

OC/EC under controlled temperature and atmosphere by continuously monitoring filter transmittance. EC is considered as a specific marker of occupational exposure to DPM since EC is a product of combustion and is composed of inert graphitic carbon.

For the characterization of DPM organic composition, analytical techniques such as gas and liquid chromatography with mass spectrometry (GC/MS and LC/MS) are applied. Both methods have adequate resolution and sensitivity to be used for the quantification of trace quantities of PAH in complex environmental matrices [20, 66-68]. Compared with LC/MS,

GC/MS is a more general technique for studying the organic composition of complex DPM matrix. The GC/MS is most sufficient in DPM analysis and is effective in isomer identification, and low detection limits are obtained with the GC/MS technique.

1.2.6 Effects of Testing Conditions on DPM

Several environmental or testing parameters influencing the chemical composition of diesel emissions have been identified: engine load condition, fuel sulfur content, and sampling condition which is primarily represented by dilution ratio.

13 Engine load has been reported to affect DPM mass concentration, size distribution, organic carbon and elemental carbon (OC/EC) distribution, morphology and microstructure [69,

70]. A study by El-Shoboksh [69] has shown that the particulate emission factor increased linearly with engine load, and the amount of submicron (0-0.4 µm) and coarse (9-10 µm) particles increased as well. Liu et al. [70] reported that the fractions of EC in DPM increased from 21% to 84% as engine load was increased from 0 kW to 75 kW.

Sulfur in diesel fuels is known to contribute to DPM emissions [56, 71-73]. Neeft et al. reported that 7-12% particulate reduction can be obtained with a reduction of fuel sulfur content from 0.2% to 0.05% [48]. A study by Baranescu et al. [45] indicated that for an increase of 0.1% in fuel sulfur, particulates increased by about 0.025 g/bhp-h. The effect of fuel sulfur content on

OC and EC distribution has been reported by Liu et al. [70].

Many studies have shown that the fate of the condensable organics and inorganics presented in diesel exhaust is significantly affected by atmospheric aging and dilution of the exhaust stream [74, 75]. It has been reported that the dilution ratio has a significant effect on the size distribution and the total number of particles emitted. Increasing the dilution ratio increases the concentration of ultrafine particles [30, 76]. The results from England et al. have indicated that traditional source testing methods (EPA 201A and EPA 202 in the paper) may significantly overestimate particulate emissions, especially the ultrafine condensable particle fraction [77].

1.3 Research Objectives

The overall goal of this study is to explore DPM chemical composition and the physical and chemical formation mechanisms, provide important knowledge for assessing the human health impact of DPM and other ultrafine particles, and establish criteria for engine design,

14 operation, and fuel reformulation. Based on this overall aim, the following subtasks will be conducted:

1. Characterize the chemical composition of DPM emissions from a non-road diesel generator and investigate source contributions from diesel fuel, engine oil, and the combustion process.

Organic components of DPM may be directly from the evaporation of diesel fuels and lubricant oils used, or may be generated during combustion processes. The DPM from a non- road diesel generator will be collected, and its organic composition will be characterized using

GC/MS instrument. Organic components of potential sources, including diesel fuels and lubricant oils used, will be instrumentally identified. Studies on DPM chemical composition can provide important information for assessing DPM health effects. The characterization of DPM composition, combined with the identification of source compositions, will help to identify products that are formed directly during the combustion process and to determine finger print components for DPM. In addition, this information will be helpful in fuel reformulation and emission mitigation technologies.

2. Investigate the effects of fuel properties, engine operation conditions, and sampling methodology on DPM composition and formation mechanisms.

It has been recognized that fuel properties, engine operation conditions, and sampling methodology can affect DPM mass concentration and size distribution. The chemical composition of DPM may be influenced by these parameters as well. In this study, diesel emissions from a non-road diesel generator will be collected with dilution and stack sampling method under varied engine load conditions (0-75 kW) and fuel sulfur contents (400 ppmw and

2200 ppmw). The organic composition of the collected samples will be analyzed with GC/MS

15 instrument and compared under different engine loads, fuel sulfur contents, and sampling methods. The results of DPM composition, especially organic components, as a function of engine loads will help to improve diesel engine operating conditions. Studying the effects of fuel sulfur on DPM composition will be helpful to promote the establishment of more stringent regulations for diesel fuel, improve fuel reformulation technologies, and therefore reduce DPM emissions. In addition, this process will be helpful for understanding the function of sulfur in organic condensation and particulate formulation. Investigating the effects of sampling methods on DPM composition will provide essential information for understanding the formation, condensation, and transportation of organic compounds, and therefore help in evaluating DPM contribution and exposure.

3. Study gas/particle partitioning of organic components in diesel emissions and investigate the mechanisms by which the organic composition is formed and distributed between gas and particle phases.

Gas and particle phase diesel emissions will be collected using dilution methods and will be instrumentally analyzed. The gas/particle phase distribution of organic compounds in diesel emissions will be investigated through correlating partitioning coefficients with vapor pressures of targeted compounds. The study of gas/particle partitioning of organic compounds in diesel emissions is helpful in understanding DPM formation mechanisms, especially the adsorption and absorption contributions in forming secondary diesel aerosol.

4. Develop analytical methods for organosulfur species in diesel emissions, identify and quantify polycyclic aromatic sulfur heterocycles (PASHs) in DPM, and investigate their distribution in DPM under varied conditions.

16 The study on polycyclic aromatic sulfur heterocycles in diesel emissions is promoted by the recognition of the potential carcinogenic and mutagenic properties associated with some

PASHs. However, it’s challenging to identify and quantify PASH species in DPM due to the complexity of the DPM matrix, as well as the instrumentation limitations. In this study, analytical method using gas chromatography with the atomic emission detection (GC/AED) will be developed and evaluated. PASHs in DPM will be identified and quantified; their concentrations and distributions in DPM will be investigated under varied engine operation conditions and fuel sulfur contents. This experiment provides additional information about DPM chemical composition, and the developed analytical method for PASHs in DPM can be applied not only to air pollution studies, but also to characterizations of other engineered materials.

1.4 References

1. United States Environmental Protection Agency, Nonroad Engines, Equipment, and Vehicles; Reducing Nonroad Diesel Emissions; Industry Characterization., Available at http://www.epa.gov/nonroad-diesel/2004fr/420r04007b.pdf (accessed June, 2006).

2. Leotz-Gartziandia, E.; Tatry, V.; and Carlier, P., Sampling and analysis of organic compounds in diesel particulate matter, Environ. Monit. Assess., 65(1-2): 155-163, 2000.

3. Kuusimaki, L.; Peltonen, Y.; Mutanen, P.; et al., Urinary hydroxy-metabolites of naphthalene, phenanthrene and pyrene as markers of exposure to diesel exhaust, Int. Arch. Occup. Environ. Health, 77(1): 23-30, 2004.

4. McDonald, J.D.; Zielinska, B.; Sagebiel, J.C.; et al., Characterization of fine particle material in ambient air and personal samples from an underground mine, Aerosol Sci. Technol., 36(11): 1033-1044, 2002.

5. United States Envrionmental Protection Agency, Nonroad Engines, Equipment, and Vehicles; Reducing Air Pollution from Nonroad Engines, http://www.epa.gov/nonroad, Accessed 2006.

6. Kinney, P.L.; Aggarwal, M.; Northridge, M.E.; et al., Airborne concentrations of PM2.5 and diesel exhaust particles on Harlem sidewalks: A community-based pilot study, Environ. Health Perspect., 108(3): 213-218, 2000.

17 7. Fruin, S.A.; Hui, S.P.; Jenkins, P.L.; et al., Fine particle and black carbon concentrations inside vehicles, in 10th Annual Conference of the International Society of Exposure Analysis. 2000: Monterey, CA.

8. Cantrell, B.K.; Rubow, K.L.; Jr. F. W. Winthrop; et al. Pollutant levels in underground coal mines using diesel equipment. in Proceedings of the 6th US Mine Ventilation Symposium. 1993. Littleton, CO: Society for Mining, Metallurgy, and Exploration, Inc. (SME).

9. Rogers, A.; and Whelan, B., Exposure in Australian mines, in HEI Communication # 7-- Diesel Workshop: Building a Research Strategy to Improve Risk Assessment (Stone Mountain, GA). Health Effects Institute: Cambridge, MA, 1999.

10. Saverin, R., German Pptash miners: cancer mortality, in HEI Communication #7--Diesel Workshop: Building a Research Strategy to improve Risk Assessment (Stone Mountain, GA). Health Effects Institute: Cambridge, MA, 1999.

11. Zaebst, D.D.; Clapp, D.E.; Blade, L.M.; et al., Quantitative determination of trucking industry workers' exposures to diesel exhaust particles, Am. Ind. Hyg. Assoc. J., 52(12): 529-541, 1991.

12. Sauvain, J.J.; Duc, T.V.; and Guillemin, M., Exposure to carcinogenic polycyclic aromatic compounds and health risk assessment for diesel-exhaust exposed workers, Int. Arch. Occup. Environ. Health, 76(6): 443-455, 2003.

13. Sjogren, M.; Li, H.; Rannug, U.; et al., A Multivariate statistical analysis of chemical composition and physical characteristics of 10 diesel fuels, Fuel, 74(7): 983-989, 1995.

14. Fernandes, M.B.; and Brooks, P., Characterization of carbonaceous combustion residues: II. Nonpolar organic compounds, Chemosphere, 53(5): 447-458, 2003.

15. Shi, J.P.; Mark, D.; and Harrison, R.M., Characterization of particles from a current technology heavy-duty diesel engine, Environ. Sci. Technol., 34(5): 748-755, 2000.

16. Reilly, P.T.A.; Gieray, R.A.; Whitten, W.B.; et al., Real-time characterization of the organic composition and size of individual diesel engine smoke particles, Environ. Sci. Technol., 32(18): 2672-2679, 1998.

17. Pal, R.; Juhasz, M.; and Stumpf, A., Detailed analysis of hydrocarbon groups in diesel range petroleum fractions with on-line coupled supercritical fluid chromatography gas chromatography mass spectrometry, J. Chromatogr. A, 819(1-2): 249-257, 1998.

18. Lee, S.L.; De Wind, M.; Desai, P.H.; et al., Aromatics reduction and cetane improvement of diesel fuels, Fuel Reformulation, 5: 26-31, 1993.

19. Adonis, M.; Martinez, V.; Riquelme, R.; et al., Susceptibility and exposure biomarkers in people exposed to PAHs from diesel exhaust, Toxicol. Lett., 144(1): 3-15, 2003.

18 20. Gratz, L.D.; Bagley, S.T.; Leddy, D.G.; et al., Interlaboratory comparison of HPLC- fluorescence detection and GC/MS: analysis of PAH compounds present in diesel exhaust, J. Hazard. Mater., 74(1-2): 37-46, 2000.

21. IPCS, Environmental Health Criteria 202. Selected nonheterocyclic Polycyclic aromatic hydrocarbons. 1998, pp. 1-883, WHO: Geneva, ISBM. 9241572027.

22. NIOSH, Carcinogenic effects of the exposure to diesel exhaust. 1990: Current Inteligence Bulletin 50. Dept. Health and Human Services, Pulication No. 88-116.

23. International Agency for Research on Cancer, IARC monographs on the evaluation of carcinogenic risks to humans: diesel and gasoline engine exhausts and some nitroarenes. Lyon, France, v. 46, p. 458, 1989.

24. EPA, Health assessment document for diesel emissions. 1990: Workshop Review EPA- 600/8-90/057A, Washington, DC.

25. World Health Organization, Diesel fuel and exhaust emissions-Environmental Health Criteria 171, Geneva, Switzerland, 1996.

26. California Environmental Protection Agency, Proposed identification of diesel exhaust as a toxic air contaminant: health risk assessment for diesel exhaust. Office of Environmental Health Hazard Assessment, Sacramento, CA, 1998.

27. U.S. Department of Health and Human Services, National Toxicology Program, 9th Report on Carcinogens, Research Triangle Park, NC, 2000.

28. Manual of the Generac diesel generator (Model SD080), Tri-state generator, http://www.powercompany.org/PDF/SD6080.pdf, accessed July 2005.

29. Burtscher, H., Physical characterization of particulate emissions from diesel engines: a review, J. Aerosol. Sci., 36(7): 896-932, 2005.

30. Lipsky, E.; Stanier, C.O.; Pandis, S.N.; et al., Effects of sampling conditions on the size distribution of fine particulate matter emitted from a pilot-scale pulverized-coal combustor, Energy Fuels, 16(2): 302-310, 2002.

31. Chemistry of diesel fuels, ed. C. Song, C. S. Hsu and I. Mochida, Taylor & Francis, New York, 2000, p. 18-19.

32. Diesel fuel and exhaust emissions, Geneva: World Health Organization, 1996, p 11-13.

33. Hsu, C.S., Diesel Fuels Analysis, in Encyclopedia of Analytical Chemistry. Wiley: New York. pp. 6613-6622, 2000.

34. Speight, J.G., Handbook of Petroleum Product Analysis, New York: Wiley-Interscience, 2002, p 177.

19 35. Wagner, T.; and Wyszynski, M.L., Aldehydes and ketones in engine exhaust emissions - A review, Proc. Inst. Mech. Eng. Part D-J. Automob. Eng., 210(2): 109-122, 1996.

36. Andresen, J.M.; Strohm, J.J.; and Song, C.S., Characterization and pyrolytic performance of hydrotreated derivatives from a light cycle oil for jet fuel applications, Abstr. Pap. Am. Chem. Soc., 218: 15-PETR, 1999.

37. Andresen, J.M.; Strohm, J.J.; and Song, C.S., Thermal degradation of naphthenic jet fuels in the pyrolytic regime and effects of hydrotreating, Abstr. Pap. Am. Chem. Soc., 216: 061-PETR, 1998.

38. Block, R.N.; Allworth, N.; and Bishop, M., Assessment of diesel contamination in soil, in Hydeocarbon contaminated soils, Vol. 1. Remediation techniques, environmental fate, risk assessment, analytical methodologies, regulatory considerations, Calabrese, E.J. and Kostecki, P.T., Editors. Lewis Publishers: Chelsea, MI. pp. 135-148, 1991.

39. Denissenko, M.F.; Pao, A.; Tang, M.S.; et al., Preferential formation of benzo(a)pyrene adducts at lung cancer mutational hotspots in P53, Science, 274(5286): 430-432, 1996.

40. US EPA, Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements, EPA 420-F-00-057, December 2000.

41. US EPA, Office of Transportation and Air Quality, Regulatory Announcement, EPA420- F-04-032, May 2004.

42. Breitkreitz, M.C.; Raimundo, I.M.; Rohwedder, J.J.R.; et al., Determination of total sulfur in diesel fuel employing NIR spectroscopy and multivariate calibration, Analyst, 128(9): 1204-1207, 2003.

43. Ma, X.L.; Sakanishi, K.Y.; and Mochida, I., Hydrodesulfurization reactivities of various sulfur compounds in diesel fuel, Ind. Eng. Chem. Res., 33(2): 218-222, 1994.

44. Shi, H.; Tayor, L.T.; Fujinari, E.M.; et al., Sulfur determination in heavy diesel fuel with open-tubular column supercritical fluid chromatography, LC GC-Mag. Sep. Sci., 16(3): 276-+, 1998.

45. Baranescu, R.A., Influence of fuel sulfur on diesel particulate emissions, SAE Technical Paper Series 881174, 1988.

46. Duran, A.; Carmona, M.; and Ballesteros, R., Competitive diesel engine emissions of sulphur and nitrogen species, Chemosphere, 52(10): 1819-1823, 2003.

47. Yu, F.Q.; Turco, R.P.; and Karcher, B., The possible role of organics in the formation and evolution of ultrafine aircraft particles, J. Geophys. Res.-Atmos., 104(D4): 4079-4087, 1999.

48. Neeft, J.P.A.; Makkee, M.; and Moulijn, J.A., Diesel particulate emission control, Fuel Process. Technol., 47(1): 1-69, 1996.

20 49. Garrett, T.K., Automotive fuels and fuel systems, Vol. 2: Diesel. London: Pentech Press: pp 334, 1994.

50. Owen, K.; and Coley, T., Automotive fuels reference book, Warrendale, PA, USA: Society of Automotive Engineers, Inc., 1995, p 963.

51. Pipenger, G., Making 'premium' diesel fuel, Hydrocarb. Process., 76(2): 63, 1997.

52. Henly, T.J.; Kulinowski, A.M.; and Roos, J.W., The role of fuel additives in formulating clean transportation fuels, chap. 4, in Designing Transportation Fuels for a Cleaner Environment, Reynolds, J.G. and Khan, M.R., Editors. Taylor & Francis: Philadelphia. pp. 61-78, 1999.

53. Kohler, M.; and Heeb, N.V., Characterization of ageing products of ester-based synthetic lubricants by liquid chromatography with electrospray ionization mass spectrometry and by electrospray ionization (tandem) mass spectrometry, J. Chromatogr. A, 2001.

54. Netten, C.V.; and Leung, V., Comparison of the constituents of two jet engine lubricating oils and their volatile pyrolytic degradation products, Applied Occupational and Envrionmental Hygiene, 15(3): 277-283, 2000.

55. Kittelson, D.B.; Watts, W.F.; and Arnold, M., Aerosol Dynamics, Laboratory and On- road Studies - Supplemental Report No. 2 to EPA Grant Review of Diesel Particulate Matter Sampling Methods. 1998.

56. Rogge, W.F.; Hildemann, L.M.; Mazurek, M.A.; et al., Sources of fine organic aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks, Environ. Sci. Technol., 27(4): 636-651, 1993.

57. Schauer, J.J.; Kleeman, M.J.; Cass, G.R.; et al., Measurement of emissions from air pollution sources. 2. C-1 through C-30 organic compounds from medium duty diesel trucks, Environ. Sci. Technol., 33(10): 1578-1587, 1999.

58. Kittelson, D.B., Engines and nanoparticles: A review, J. Aerosol. Sci., 29(5-6): 575-588, 1998.

59. Zielinska, B.; Sagebiel, J.; Arnott, W.P.; et al., Phase and size distribution of polycyclic aromatic hydrocarbons in diesel and gasoline vehicle emissions, Environ. Sci. Technol., 38(9): 2557-2567, 2004.

60. Choudhury, D.R., Characterization of polycyclic ketones and quinones in diesel emission particulates by gas-chromatography mass-spectrometry, Environ. Sci. Technol., 16(2): 102-106, 1982.

61. Konig, J.; Balfanz, E.; Funcke, W.; et al., Determination of oxygenated polycyclic aromatic hydrocarbons in airborne particulate matter by capillary gas-chromatography and gas-chromatography mass-spectrometry, Anal. Chem., 55(4): 599-603, 1983.

21 62. Durant, J.L.; Busby, W.F.; Lafleur, A.L.; et al., Human cell mutagenicity of oxygenated, nitrated and unsubstituted polycyclic aromatic hydrocarbons associated with urban aerosols, Mutat. Res.-Genet. Toxicol., 371(3-4): 123-157, 1996.

63. Bezabeh, D.Z.; Bamford, H.A.; Schantz, M.M.; et al., Determination of nitrated polycyclic aromatic hydrocarbons in diesel particulate-related standard reference materials by using gas chromatography/mass spectrometry with negative ion chemical ionization, Anal. Bioanal. Chem., 375(5): 381-388, 2003.

64. Zwirner-Baier, I.; and Neumann, H.G., Polycyclic nitroarenes (nitro-PAHs) as biomarkers of exposure to diesel exhaust, Mutat. Res. Genet. Toxicol. Environ. Mutagen., 441(1): 135-144, 1999.

65. Watts, W.F.J.; and Ramachandran, G., Diesel Particulate Matter Sampling Methods: Statistical comparison, 2000, Final Report Submitted to the Diesel Emission Evaluation Program, http://www.deep.org/reports/watts_final.pdf, Accessed 2006.

66. Poster, D.L.; de Alda, M.J.L.; Schantz, M.M.; et al., Development and analysis of three diesel particulate-related standard reference materials for the determination of chemical, physical, and biological characteristics, Polycycl. Aromat. Compd., 23(2): 141-191, 2003.

67. Wise, S.A., Standard reference materials (SRMs) for the determination of polycyclic aromatic compounds - Twenty years of progress, Polycycl. Aromat. Compd., 22(3-4): 197-230, 2002.

68. Wise, S.A.; Sander, L.C.; and May, W.E., Determination of polycyclic aromatic hydrocarbons by liquid chromatography, J. Chromatogr., 642(1-2): 329-349, 1993.

69. El-Shobokshy, M.S., The effect of diesel-engine load on particulate carbon emission, Atmos. Environ., 18(11): 2305-2311, 1984.

70. Liu, Z.F.; Lu, M.M.; Birch, M.E.; et al., Variations of the particulate carbon distribution from a nonroad diesel generator, Environ. Sci. Technol., 39(20): 7840-7844, 2005.

71. Dobbins, R.A.; Fletcher, R.A.; Benner, B.A.; et al., Polycyclic aromatic hydrocarbons in flames, in diesel fuels, an in diesel emissions, Combust. Flame, 144(4): 773-781, 2006.

72. Jang, M.; Kamens, R.M.; Leach, K.B.; et al., A thermodynamic approach using group contribution methods to model the partitioning of semivolatile organic compounds on atmospheric particulate matter, Environ. Sci. Technol., 31(10): 2805-2811, 1997.

73. Pankow, J.F., An absorption model of gas/particle partitioning of organic compounds in the atmosphere, Atmos. Environ., 28(2): 185-188, 1994.

74. Ahlvik, P.; Ntziachristos, L.; Keskinen, J.; et al., Real time measurements of diesel particle size distribution with an electrical low pressure impactor, SAE Technical Paper Series 980410, 1998.

22 75. Kittelson, D.B.; and Abdul-Khalek, I.S. Formation of nanoparticles during exhaust dilution. in Proceedings of 2nd ETH Workshop on Nanoparticle Measurement. August 1998. Zurich, Switzerlan.

76. Brown, J.E.; Clayton, M.J.; Harris, D.B.; et al., Comparison of the particle size distribution of heavy-duty diesel exhaust using a dilution tailpipe sampler and an in- plume sampler during on-road operation, J. Air Waste Manage. Assoc., 50(8): 1407-1416, 2000.

77. England, G.C.; Zielinska, B.; Loos, K.; et al., Characterizing PM2.5 emission profiles for stationary sources: comparison of traditional and dilution sampling techniques, Fuel Process. Technol., 65: 177-188, 2000.

23 Chapter 2

EXPERIMENTAL DESIGN AND METHOD

2.1 Experimental Design

In this study, a Generac diesel generator (1992, Model SD080, model No. 92A-03040-S) located at the UC Center Hill Research Facilities in Cincinnati, Ohio, was used as the non-road source of diesel particulate matter (Figure 2-1). This unit is powered by a Generac diesel engine which features direct-injection, turbocharging, and compression-ignition, and is rated at 80 kW,

60 Hz and 1800 rpm. Unlike diesel vehicles, diesel generators run at a fixed rpm and voltage and the amperage it produces can vary with the load [1]. The difference in operating mode can potentially result in different emission characteristics. A load simulator (Merlin 100 manufactured by SIMPLX) was used to simulate loads by applying steady-state banks of heaters to the generator at 0 kW, 25 kW, 50 kW and 75 kW, respectively.

Samples were collected on quartz filters or sorbents using EPA method 5 (Source Test

Sampling for Particulate) and the dilution method. The tests were performed under four engine load conditions with two different sulfur-containing diesel fuels (low sulfur diesel fuel [LSDF] with sulfur content about 400 ppmw, and high sulfur diesel fuel [HSDF] with sulfur content about 2200 ppmw; both are from Steve Krebs Oil Company, Inc.).

Figure 2-1 Diesel emission source (diesel generator and load simulator).

Two types of samples were collected: for the analysis of organic carbon and elemental carbon (OC/EC), and for compositional analysis with GC/MS. Prior to sampling, the quartz fiber filters were baked at 550°C for a minimum of 12 hours to reduce residual carbon levels associated with new filters. After sampling, the filters were put in a desiccator for 24 hours, weighed, and then stored in a refrigerator until the samples were extracted or OC/EC analyzed.

The collected samples for compositional analysis were extracted with solvent, filtered, and evaporated by blowing nitrogen to 0.5~1mL. The small amount of extracts was split into two separate fractions. One fraction was then derivatized with N,O- bis[trimethylsilyl]trifluoroacetamide (BSTFA, Pierce) to convert organic acids to their methyl ester analogues, which are able to be identified and quantified with GC/MS. The other fraction was directly analyzed with GC/MS and other instruments. If the concentrations of selected

25 compounds in the extracts are not high enough to be identified, extracts from several samples were combined.

For the analysis of diesel fuel and engine oil (Valvoline, SAE 10W-30), they were diluted with HPLC grade dichloromethane (DCM, Fisher Scientific) to some certain concentrations and analyzed. If organic acids were targeted compounds, the diluted samples were derivatized by adding BSTFA, and then analyzed with GC/MS.

2.2 Sampling Method

Two major types of sampling methods have been used for collecting PM emissions from combustion systems: source-level sampling and ambient-level sampling. In general, source-level sampling is currently the accepted approach for the total suspended particulate and PM10 measurements for stationary sources, while ambient level sampling using dilution is the accepted approach for mobile source particulate emission measurements. In this study, three sampling methods were used, including EPA Reference Method 5, high volume dilution sampling, and natural dilution sampling.

2.2.1 EPA Method 5

EPA Reference Method 5, Determination of Particulate Emissions from Stationary

Sources, which can be found in the US Code of Federal Regulations [2], is the most common method used to determine whether a source is in compliance with the established particulate emission standards. In EPA Method 5, the undiluted exhaust is sampled through a heated sampling probe. Particulate matter is withdrawn isokinetically from the source and collected on a glass fiber filter maintained at 120 ± 14°C. The particulate mass, which includes any material

26 that condenses at or above the collection/filtration temperature, is determined gravimetrically after the removal of uncombined water. The Method 5 sampling train is shown in Figure 2-2.

For the compositional analysis and OC/EC measurement, the following modifications were made to the EPA Method 5 sampling procedures: (1) particles were collected on quartz fiber filters instead of glass fiber filters, as quartz filters are required for thermal-optical analysis by NIOSH 5040 [3, 4]; (2) due to the maximum loading limitation of the OC/EC analysis, the sampling duration for DPM collection ranged from about 8 to 15 minutes, which is much shorter than a typical Method 5 sampling period of approximately 60 minutes; (3) in order to obtain enough mass for compositional analysis, samples were collected for 1 to 2 hours without too high of a pressure drop observed.

Figure 2-2 Schematic of an EPA Method 5 sampling train.

27 2.2.2 Natural Dilution

Natural dilution sampling is generally used for studies on on-road vehicle emissions.

Kittelson et al. [5-11] have conducted a series of studies on the sampling and analysis of particle emissions from diesel engines and gasoline vehicles. In most of these studies, a natural dilution sampling system, which is referred to as the University of Minnesota Mobile Emissions

Laboratory (MEL), was used. The MEL was installed on a truck, which allows the sampling of exhaust from both mobile source and stationary source. When sampling the exhaust from mobile source, the MEL followed the target vehicle and the sample was collected through the air intake which was located in front of the MEL. A suite of aerosol instrumentation was used in MEL to measure particle number concentration, size distribution, surface area, and gas concentration.

This natural dilution system was large-scale, and suitable for on-road emission measurements.

However, the big problem with this natural dilution system includes less control, poor repeatability and seasonal variation of sampling results.

The natural dilution sampler was built and used for DPM compositional analysis at the beginning of this study (Figure 2-3). A high volume blower (General Metal Works Model 2000) was used to draw air through the system. Quartz filters with a diameter of 90 mm were used to collect DPM. The sampling flow rate, which is adjustable from 250 L/min to 350 L/min, was measured by an orifice meter which was pre-calibrated by a Pitot tube with a PVC air duct

(length: about 3 feet, inner diameter: 2 inches) [12]. The sampler was set 1 m away from the exhaust gas outlet with the sampling inlet about 1.4 m from the ground.

Figure 2-3 Schematic of the natural dilution sampler for DPM collection.

The natural dilution sampler was similar to the dilution sampler mentioned in the previous section, except that the exhaust gas and the dilution air were drawn in through the same inlet with an unknown dilution ratio and were mixed naturally in the inlet duct. The flow rates of dilution air and the exhaust gas were not measured separately. The dilution ratio depends on the characteristics and the relative position of the exhaust outlet and the sampler inlet, which include their diameters, the distance between them, the direction, the height, and the exhaust flow rate.

The natural dilution sampler was used for characterization of DPM composition in the 29 atmosphere. With the natural dilution sampler, the actual DPM composition after it was emitted into the atmosphere can be obtained, while the developed dilution sampler can only simulate the real situation. However, natural dilution sampling was significantly affected by ambient conditions and cannot provide satisfactory repeatability. Therefore, when examining the effects from various factors, the developed dilution sampling was applied.

2.2.3 High Volume Dilution Sampling

It is challenging to sample particulate matter emissions from diesel engines due to the elevated temperatures of the exhaust gases and the complexity of DPM. Many studies have shown that DPM remains in a state of flux for some time after it is emitted to the atmosphere.

During this state, coagulation, adsorption and condensation of organics and inorganics, which are usually present in diesel exhaust with significant quantities, are continuously occurring. The fate of these condensable compounds is significantly affected by dilution of the exhaust stream.

Research on emissions from diesel engines has shown that sampling conditions can alter measured DPM size distributions [13, 14].

Dilution sampling is a technique that has been developed to examine the influence of rapid cooling and dilution on PM emissions from combustion systems. A dilution sampler rapidly mixes hot exhaust gases with the pretreated dilution air, cools down the exhaust gas temperature, and allows nucleation, condensation, and coagulation to occur. Although a dilution sampler cannot simulate the actual atmospheric mixing, it allows the examination of effects of dilution on PM emissions in order to better understand the PM transformations that occur in the atmosphere. In the design of a dilution sampling device, Hildemann et al. [15] suggested that the following elements must be considered :

30 1. The sampler should simulate atmospheric dilution as closely as practicable. To

achieve this, the emissions should be diluted and cooled to ambient temperature.

2. The sampler must be designed to minimize the contamination to the sample, which

requires the sampler materials to be inert to organics and stable to high temperature.

3. The sampler should provide enough residence time to allow condensation to occur.

The Code of Federal Regulations also has some requirements for exhaust gas sampling and analytical systems for emissions from diesel engines [16, 17]. One important requirement is that the sampling system should provide the dilution of the exhaust to a temperature less than

51.7 °C at the sample filter.

In addition to the dilution requirement for the sampling system to collect diesel emissions, high volume sampling is another necessity for compositional analysis of diesel emissions. The sampler should be able to collect enough materials in a practicable time for effective speciation analysis of organic compounds in diesel emissions, such as PAHs, nitro-compounds, and oxygenated compounds.

Hildemann et al. [15, 18, 19] designed a dilution stack sampler specifically intended to collect fine organic aerosol from combustion sources while minimizing sample contamination

(Figure 2-4). This sampler drew the stack exhaust at a flow rate of 30 L/min, and mixed the stack exhaust with preconditioned dilution air at a dilution ratio ranging from 25 to 100. In the dilution tunnel, the exhaust was mixed with dilution air under turbulent flow (Re = 10000) and was cooled to near-ambient conditions. The dilution tunnel can provide a residence time of approximately 2 seconds, and samples collected under this condition can be used for chemical analysis. Additional residence time was provided by the residence time chamber, after which samples were collected for particle size distribution measurements and chemical analysis.

Figure 2-4 Dilution stack sampler schematic diagram designed by Hildemann et al. [15].

Figure 2-5 Schematic of high volume dilution sampler for particulate and gas phase diesel emissions.

In this study, a high volume dilution sampler was designed for sampling both gas phase and particulate diesel emissions from a diesel generator (Figure 2-5). This sampler is similar in

33 principle to the dilution stack sampler designed by Hildemann et al., but some control devices are simplified and the dilution ratio is lower due to the limitation of experimental conditions.

Besides the sampling of particulate diesel emission, this sampler is designed to collect gas phase emissions as well, which is a supplement to Hildemann’s sampler. The sampler consisted of two stages. The first stage was a 90-mm quartz fiber filter (Millipore) for the collection of diesel particulate matter. The second stage consisted of polyurethane foam plugs (60 mm diameter) in combination with the 10 g of adsorbent resin XAD-4 (PUF/XAD/PUF “sandwich” cartridge) that was placed downstream of the filter to collect the gas phase organics. Diesel exhaust is drawn from the exhaust duct through the sample inlet probe (12.7-mm diameter, 1.22-m length), and then mixed with particle-removed ambient air in the dilution pipe, a 38.1-mm-inner diameter,

2.44-m-long PVC duct.

A high volume blower (General Metal Works Model 2000) is used to draw air through the system. The total flow rate (Qt) of this sampler is approximately 300 L/min and was measured with an orifice meter. The dilution air was filtered at the dilution air inlet using a high- efficiency particulate air filter (HEPA) that has successfully achieved a minimum particle removal efficiency of 99.97%. The flow rate of dilution air (Qd) was measured with a flow meter

(Dwyer). The performance specifications for the sampler are listed in Table 2-1.

Table 2-1 Performance characteristics of high volume dilution sampler. Parameter Range Sample flow 50-60 L/min Dilution air flow 150-240 L/min Total flow 200-300 L/min Dilution Ratio ~3.4 Residence time 0.5-1 second Reynolds number of flow through mixing duct 7000-11000, turbulent flow

34 A dilution ratio of 3.4, which was calculated from the two measured flow rates, was achieved with the sampler. Even though the dilution ratio is low relative to other studies [20, 21], the dilution sampler provides sufficient dilution air to maintain the exhaust stream at a temperature ranging from 17.2 °C to 31.7 °C, which is lower than the temperature (51.7 °C) required for sample filters when dilution sampling by the Code of Federal Regulations [16].

The DPM samples were taken at four different engine load conditions: 0 kW (idle condition), 25 kW, 50 kW, and 75 kW. The sampling duration for one filter was from 15 min (75 kW) to 1 hr (0 kW), depending upon DPM concentration which related with the engine load. The breakthrough tests were performed for the sampling duration under different load conditions, which indicated that no significant breakthrough was observed for filters and sorbents with the designated sampling duration. An amount from 15 mg (for 0 kW) to 30 mg (for 75 kW) of DPM was collected to perform the GC/MS analysis.

2.3 Extraction method

In this study, three extraction methods that are generally used for DPM extraction [22-26], shaking by hand, soxhlet, and ultrasonic assisted extraction, were tested and compared. Although several new extraction methods have been developed, such as pressurized liquid extraction

(PLE), supercritical fluid extraction (SFE), and subcritical water extraction [27, 28], their applications have been limited due to the high instrumentation and operation expense. Extraction of shaking by hand is not sufficient, especially for heavy compounds. Previous studies have indicated that soxhlet extraction needs long operation time (normally 24 hours) [25, 26], and thus the recovery for light compounds is low. The extraction recovery that can be obtained ranges from 67% to 116% [29-33]. The extraction recoveries from these three methods were also

35 measured in this study and were 68-87% (shaking by hand), 34-76% (soxhlet) and 87-98%

(ultrasonic assisted extraction) respectively. Regarding all these considerations (including operation complexity, operation time, the amount of solvent used and extraction recovery for all compounds with different volatility), the ultrasonic assisted extraction method was used in this study.

Two types of solvent are generally used, which are pure solvent and mixture solvent.

DCM is the common pure solvent used for extracting diesel particulate matter, since most organic compounds can be solved in DCM [24-26]. Mixture solvents, such as hexane/ether

(90:10) [26], Benzene/methanol (4:1) [23], DCM/benzene (70:30), DCM/hexane (20:80), and methanol/DCM (20:80) [25] are preferred due to their ability to extract many compounds with different structures and chemical properties. However, solvent exchange needs to be performed after evaporation due to the requirement of most analytical instruments during which some light compounds may be lost. Therefore, DCM was used as the extraction solvent in this study.

During the extraction, deuterated internal standards were spiked on filters to determine the extraction recovery [26]. In some studies, several categories of deuterated internal standards were used to measure the recoveries of compounds with corresponding chemical properties respectively [20, 34]. For instance, deuterated polycyclic aromatic hydrocarbons (PAHs), ranging in volatility from naphthalene-d8 to coronene-d12, were often used as the internal standards for PAH analysis. The extraction recovery can also be approximated by using one deuterated compound (mostly deuterated tetrocosane (n-C24D50)) as an internal standard [35]. In this study, naphthalene-d8 (Aldrich) and phenanthrene-d10 (Aldrich) were used to measure extraction recoveries for all identified compounds.

36 Before extraction and prior to sampling, the quartz fiber filters were baked at 550 °C for a minimum of 12 h to reduce residual carbon levels associated with new filters, cooled to ambient temperature, and weighed. After sampling, the filters were desiccated at 20 °C ± 5.6 °C and under ambient pressure for at least 24 hours and weighed. Then the filters were stored in the refrigerator until the extraction. Before extraction, the quartz fiber filters were spiked with a mixture of deuterated internal standards to measure the recovery. Samples were extracted in

DCM with sonication for 1 h. Extracts were filtered through a Teflon filter (Altec, pore size: 0.2

µm) with a vacuum filtration system to remove the insoluble fractions (such as soot), evaporated under a gentle stream of nitrogen to ~1 mL, and separated into two fractions. One fraction was first derivatized with N,O-bis[trimethylsilyl]trifluoroacetamide (BSTFA) to convert organic acids into their trimethylsilyl ester analogues, which can be easily identified and quantified with

GC/MS. Then, both fractions were analyzed with GC/MS.

2.4 Analytical Method

2.4.1 OC/EC Analysis

In this study, organic carbon and elemental carbon (OC/EC) analysis with NIOSH

Method 5040 were applied for DPM measurement.

DPM has been known to consist mostly of carbonaceous materials, about 70% to 88%, which is often classified as EC and OC [36]. EC and OC act in different ways due to their different optical, physical, chemical, and toxicological properties. As the core of diesel soot particles, EC is a byproduct of incomplete combustion consisting of carbon layers that are structurally similar to graphite. Particulate OC consists of liquid droplets and soot-associated organics. The OC fraction of DPM is a complex mixture of unburned diesel fuel, oil, and

37 numerous organic compounds including PAHs [37-39]. EC has been linked to dysrhythmia and cardiovascular diseases [40], while PAHs in OC fraction are reasonably anticipated to be human carcinogens [24, 41-45].

The contributions of EC and OC to DPM vary with a number of factors, such as the fuel type, engine type, duty cycle, engine maintenance, individual operators, the use of emission control devices, and the compositions of the lubricant oil. Some previous studies have reported that the OC/EC distribution varies with load conditions for new heavy duty diesel vehicles [45], military vehicles [46], and non-road diesel generators [36].

The thermal-optical method NIOSH 5040 is most commonly used in North America for

OC/EC analysis [3, 4]. In this method, speciation of OC and EC is accomplished through temperature and atmosphere control and by continuous monitoring of filter transmittance. Laser light passed through the filter allows continuous monitoring of filter transmittance. Because temperatures in excess of 850 °C are employed during the analysis, quartz-fiber filters are required. A punch from the sample filter is taken for analysis, and OC and EC are reported in terms of µg/cm2 of filter area. The total OC and EC on the filter are calculated by multiplying the reported values by the deposit area. In this approach, a homogeneous sample deposit is assumed.

Thermal-optical analysis proceeds essentially in two stages. In the first, organic and carbonate (if present) carbon are evolved in a helium atmosphere as the temperature is stepped up to about 850 °C. The evolved carbon is catalytically oxidized to CO2 in a bed of granular

MnO2, then reduced to CH4 in a Ni/firebrick methanator. CH4 is quantified by an FID. In the second stage, the sample oven temperature is reduced, an oxygen-helium mix is introduced, and the temperature is stepped (to about 940 °C). As oxygen enters the oven, pyrolytically generated carbon (PC) is oxidized and a concurrent increase in filter transmittance occurs (see Figure 2-6).

38 The point at which the filter transmittance reaches its initial value is defined as the "split" between OC and EC. Carbon evolved prior to the split is considered as OC (including carbonate), and carbon volatilized after the split is considered as EC. TC is the sum of OC and EC. Blank filters were analyzed for a background check. The mass concentration of organic compounds can be estimated from OC by a multiplicative factor of 1.2-1.4 to account for other elements, (e.g., hydrogen, oxygen, nitrogen, sulfur) in addition to carbon [47].

39 2.4.2 Hydrocarbon Analysis with GC/MS

2.4.2.1 Instrument and Operating Parameters

A number of analytical methods have been developed and applied to the quantification of organic compounds in diesel emissions. Among these, gas chromatography / mass spectrometry

(GC/MS), and high performance liquid chromatography (HPLC) coupled with fluorescence detection, are the most typical analytical methods. Both of these methods have adequate resolution and sensitivity to be used for the quantification of trace quantities of PAH in complex environmental matrices [24, 48-50]. However, co-elution of compounds where separation and especially quantitation of species within a peak cannot be overcome with HPLC, since for this technique compound identification still relies on retention indices. Thus, errors may be involved in compound recognition. GC/MS is a better choice for such complex matrices. The GC/MS method is more efficient and provides results for more isomers than the LC-fluorescence method

[51], and low detection limits are obtained with the GC/MS technique.

In this study, the Varian GC/MS system (Varian, CP-3800 GC, Saturn 2200 ion trap MS) equipped with a CP-8400 automatic sampler was used. The column used for the analysis was

CP-Sil 8 CB low Bleed/MS (30 m × 0.25 mm × 0.25 mm, equivalent of DB5-ms).

The chromatographic procedure for GC/MS can be described as follows: The injection

(0.5 µL) was in splitless mode at a temperature of 280 °C. The column temperature was first set at 40°C for 2 min, then raised to 200 °C at 10 °C/min, to 270 °C at 5 °C/min, to 300 °C at 10

°C/min, and held at 330 °C for 10 min. Helium was used as the carrier gas with a flow rate of 1.2 mL/min. The total analysis time was 45 min.

The transfer line between the chromatograph and the mass analyzer was maintained at

280 °C, and the conditions for the mass analyzer follow. The mass analyzer manifold and the ion

40 trap were set to 45 ºC and 170 ºC, respectively. The initial ionization mode for the ion trap was electron impact, and the filament emission current for that purpose was set to 10 µA. The filament and electron multiplier were delayed for 3.5 min to eliminate the solvent peak. For mass data acquisition, the threshold was 1 count.

2.4.2.2 Compound Identification and Quantification

Compound identification was performed by comparing the retention times and mass spectra of samples with standard reference compounds under the help of National Institute of

Standards and Technology (NIST) standard library. The method of selective ion search was used to enhance the identification of the PAH compounds.

Compound quantification was based on external standards. The response factors were determined by injections of standard compounds with multilevel concentrations using the same analytical method. Although GC/MS is used to analyze compounds in diesel exhaust, it is difficult to quantify all components with conventional total ion chromatogram (TIC), since some components exist in trace amount, such as PAHs, and they are mixed with numerous components that comprise the very convoluted exhaust. Therefore, in this study three techniques were used to quantify different categories of compounds, TIC, selective ion chromatogram (SIC), and tandem mass spectrometry (MS/MS).

TIC: TIC was used to quantify these compounds which have high concentrations in diesel emissions and do not have prominent fragment ions. Normal alkanes, branched alkanes, saturated cycloalkanes, and organic acids were quantified with TIC.

SIC: SIC is a post-analysis technique that operates on mass signals that the computer previously saved as a function of time. The analyst can search a previous chromatogram

41 selectively for a particular molecular or fragment ion and simultaneously reject the interferences to generate a SIC with signals that correspond only to the mass to charge (m/z) of interest [52].

This dual function enhances the detection and sensitivity. An important advantage is that the established procedures for producing the traditional gas chromatography/mass spectrometry

(GC/MS) spectrum remain unchanged.

In this study, SIC was used for the quantification of alkylated PAHs and alkylbenzenes, which have prominent ions and medium concentrations in diesel emissions. The selected fragment ions for each compound were listed in Table 2-2.

Table 2-2 The selected quantitative fragment ions for alkylated PAHs and alkylbenzenes. Compound Ions Compound Ions

MN 141, 142, 115 C1-B 91, 92

DMN 156, 141, 115 C2-B 106, 91

TMN / MEN 170, 155, 128 C3-B 105, 120,91

MPh / MAn 192, 191, 165 C4-B 119, 134, 91

DMPh / DMAn 206, 191, 189 C5-B 133, 105, 148

Hopane 191 C6-B 147, 119, 162

MS/MS: MS/MS is an approach that reduces the background caused by the complex matrix. All ions have been excluded, except for the selected ions (generally one to three ions for each compound), which can then be fragmented to produce a unique product ion mass spectrum.

Because of the ability of MS/MS to preferentially isolate the ions of the target compounds, the background noise is lower and higher sensitivities are possible due to the absence of interfering ions [53-55].

The MS/MS process is a pre-programmed mass data acquisition event. Selected ion monitoring (SIM) was used for the analysis of all sixteen EPA priority PAHs. An analysis

16PAHs with an ion trap can be further segmented, with each segment optimized for particular 42 PAHs in that window. The time segment when to assign the m/z search for a particular compound was based on the chromatographic elution time for the standard compound. The time segment and monitored ions for all 16PAHs are listed in Table 2-3.

Table 2-3 GC/MS/MS segment program for 16PAHs. Compound Name Abbr. MW # of Time Ions monitored(m/z) rings segment (min) Naphthalene Nap 128 2 4.0-12.0 101-103, 127-129 Acenaphthylene Acy 152 3 12.0-14.8 75-77, 125-127, 151-153 Acenaphthene Ace 154 3 14.8-15.5 75-77, 125-127, 152-154 Fluorene Flu 166 3 15.5-16.5 82-84, 138-140, 165-167 Phenanthrene Phe 178 3 16.5-20 74-76, 151-153, 177-179 Anthracene Ant 178 3 16.5-20 74-76, 151-153, 177-179 Fluoranthene Flt 202 4 20-25 88-89,100-102, 201-203 Pyrene Pyr 202 4 20-25 88-89,100-102, 201-203 Benzo(a)anthracene Baa 228 4 25-30 99-101, 112-114, 227-229 Chrysene Chy 228 4 25-30 99-101, 112-114, 227-229 Benzo(b)fluoranthene Bbf 252 5 30-35 111-113, 124-126, 251-253 Benzo(k)fluoranthene Bkf 252 5 30-35 111-113, 124-126, 251-253 Benzo(a)pyrene Bap 252 5 30-35 111-113, 124-126, 251-253 Indeno[1,2,3-cd]pyrene Ind 276 6 35-45 123-126, 136-139, 275-279 Dibenz(a,h)anthracene Dba 278 5 35-45 123-126, 136-139, 275-279 Benzo(ghi)perylene Bgp 276 6 35-45 123-126, 136-139, 275-279

2.4.2.3 Standard Compounds

Quantification was based on the response factors of standard compounds and the extraction recoveries of deuterated compounds. The standard compounds listed in Table 2-4 were injected into GC/MS system. All these standard compounds were dissolved in HPLC grade dichloromethane.

43 Table 2-4 List of standard compounds and their chemical properties. Compound name Formula Abbreviation MW n-Alkanes (Restek) n-Decane C10H22 C10 142 n-Undecane C11H24 C11 156 n-Dodecane C12H26 C12 170 n-Tridecane C13H28 C13 184 n-Tetradecane C14H30 C14 198 n-Pentadecane C15H32 C15 212 n-Hexadecane C16H34 C16 226 n-Heptadecane C17H36 C17 240 n-Octadecane C18H38 C18 254 n-Nonadecane C19H40 C19 268 n-Eicosane C20H42 C20 282 n-Heneicosane C21H44 C21 296 n-Docosane C22H46 C22 310 n-Tricosane C23H48 C23 324 n-Tetracosane C24H50 C24 338 n-Pentacosane C25H52 C25 352

Branched alkanes (Chiron)

Norfarnesane C14H30 C14 198

Farnesane C15H32 C15 212

Norpristane C18H38 C18 254

Pristine C19H40 C19 268

Phytane C20H42 C20 282

Cycloalkanes (Chiron) n-Heptylcyclohexane C13H26 C13 184 n-Octylcyclohexane C14H28 C14 198 n-Nonylcyclohexane C15H30 C15 212 n-Decylcyclohexane C16H32 C16 226 n-Undecylcyclohexane C17H34 C17 240 n-Dodecylcyclohexane C18H36 C18 254 n-Tridecylcyclohexane C19H38 C19 268 n-Tetradecylcyclohexane C20H40 C20 282 n-Pentadecylcyclohexane C21H42 C21 296 n-Hexadecylcyclohexane C22H44 C22 310

44 Table 2-4 (Continued) Compound name Formula Abbreviation MW n-Heptadecylcyclohexane C23H46 C23 324 n-Octadecylcyclohexane C24H48 C24 338 n-Nonadecylcyclohexane C25H50 C25 352

PAHs (Aldrich)

Naphthalene C10H8 Nap 128

Acenaphthylene C12H8 Acy 152

Acenaphthene C12H10 Ace 154

Fluorene C13H10 Flu 166

Phenanthrene C14H10 Phe 178

Anthracene C14H10 Ant 178

Fluoranthene C16H10 Flt 202

Pyrene C16H10 Pyr 202

Benzo(a)anthracene C18H12 Baa 228

Chrysene C18H12 Chy 228

Benzo(b)fluoranthene C20H12 Bbf 252

Benzo(k)fluoranthene C20H12 Bkf 252

Benzo(a)pyrene C20H12 Bap 252

Indeno[1,2,3-cd]pyrene C22H12 Ind 276

Dibenz(a,h)anthracene C22H14 Dba 278

Benzo(ghi)perylene C22H12 Bgp 276

Alkylated PAHs (Aldrich)

1-Methylnaphthalene C11H10 1-MN 142

2-Methylnaphthalene C11H10 2-MN 142

1,2-Dimethylnaphthalene C12H12 1,2-DMN 156

1,3-Dimethylnaphthalene C12H12 1,3-DMN 156

1,4-Dimethylnaphthalene C12H12 1,4-DMN 156

1,5-Dimethylnaphthalene C12H12 1,5-DMN 156

1,6-Dimethylnaphthalene C12H12 1,6-DMN 156

1,7-Dimethylnaphthalene C12H12 1,7-DMN 156

1,8-Dimethylnaphthalene C12H12 1,8-DMN 156

2,3-Dimethylnaphthalene C12H12 2,3-DMN 156

2,6-Dimethylnaphthalene C12H12 2,6-DMN 156

2,7-Dimethylnaphthalene C12H12 2,7-DMN 156

1,3,5-Trimethylnaphthalene C13H14 1,3,5-TMN 170

45 Table 2-4 (Continued) Compound name Formula Abbreviation MW

1-Methylphenanthrene (Ultra Scientific) C15H12 1-MPh 192

2-Methylphenanthrene (Ultra Scientific) C15H12 2-MPh 192

3,6-Dimethylphenanthrene (Ultra Scientific) C16H14 3,6-DMPh 206

Alkylbenzenes (Fisher)

Toluene C7H8 C1-B 92

C2-Benzenes C8H10 C2-B 106

C3-Benzenes C9H12 C3-B 120

C4-Benzenes C10H14 C4-B 134

C5-Benzenes C11H16 C5-B 148

C6-Benzenes C12H18 C6-B 162 n-Alkanoic acids (Fisher)

Hexanoic acid C6H12O2 C6 116

Heptanoic acid C7H14O2 C7 130

Octanoic acid C8H16O2 C8 144

Nonanoic acid C9H18O2 C9 158

Decanoic acid C10H20O2 C10 172

Undecanoic acid C11H22O2 C11 186

Dodecanoic acid C12H24O2 C12 200

Tridecanoic acid C13H26O2 C13 214

Tetradecanoic acid C14H28O2 C14 228

Pentadecanoic acid C15H30O2 C15 242

Hexadecanoic acid C16H32O2 C16 256

Heptadecanoic acid C17H34O2 C17 270

Octadecanoic acid C18H36O2 C18 284

Aromatic acids (Fisher)

Benzoic acid C7H6O2 122

2.4.2.4 Quality Assurance and Quality Control

In analytical procedure, QA-QC was assured by the following steps:

(1) Before processing any samples, a reagent solvent blank was analyzed to demonstrate that interferences from the analytical system, glassware, and reagents are under control. Each

46 time a set of samples is extracted or there is a change in reagents, a reagent solvent blank should be processed as a safeguard against laboratory contamination. The blank samples should be carried through all stages of the sample preparation and measurement steps.

(2) For each analytical batch (up to 20 samples), a reagent blank, matrix spike, and deuterated/surrogate samples were analyzed. The blank and spiked samples were carried through all stages of the sample preparation and measurement steps.

(3) Calibration curves were obtained from the standard compounds with the R2 value greater than 0.999. During the analysis of samples, one standard solution was injected to check the response stability of instruments. If the responses of standard compounds changed more than

5%, calibration curves were re-made and new response factors were used;

(4) All sample and standard solutions were sealed in vials and stored in refrigerators to prevent sample quality decay. After 3-6 months, several fresh standard solutions were prepared to check the quality of the stored solutions. If the concentrations of stored solutions changed, the stored solutions were discarded and new solutions were prepared;

(5) Sample containers, sampling equipments, and laboratory glassware were cleaned and stored in a clean environment to ensure no contamination from external sources.

2.5 Experiment on Source Contribution to DPM Composition

For the investigation of the quantitative contribution from a source, it requires data on the amount of a compound in DPM emissions and knowledge on either of the following two parameters: the amount of a compound formed during combustion processes, and the amount of a compound from a direct source, such as unburned diesel fuel and evaporated engine oil. The parameter expressed in the amount of a compound in diesel emissions divided by the amount of

47 this compound in the diesel fuel is referred to as survivability. While it is difficult to quantify the amount of a compound from combustion formation, due to the complexity and variability of the actual combustion processes, survivability measurement of compounds in diesel fuels is relatively simple and applicable.

Several studies applied a radiotracer technique to investigate the contributions of unburned diesel fuels and other sources, which are mainly pyrosynthetic sources, to compounds in diesel emissions [56-60]. The survivabilities for fluorene, 2-methylnathphalene, naphthalene, pyrene, and benzo(a)pyrene were reported in these studies. Based on the survivability of a compound and its emission rate in the diesel exhaust, the contribution from unburned diesel fuel and the combustion source can be determined. Tancell et al. [60] reported that the majority of benzo(a)pyrene (>80%) in the exhaust was derived from unburned fuel. A study by Rhead and

Pemberton [58] indicated that naphthalene in diesel emission was mainly from unburned diesel fuel (76.2%), while other sources only contributed 23.8%. In the study by Rhead and Hardy [57], the majority of fluoranthene in diesel emissions was from unburned diesel fuel. The contributions to pyrene and fluorene in diesel emissions from unburned diesel fuel and pyrosynthesis depended on engine speed and engine load.

A simpler method for estimating the contribution to a compound in DPM emission from the diesel fuel was developed by Hori et al. [61]. In their study, the compound of interest was added into the fuel, which is referred to as the compound-enriched fuel. By the comparison between the compound emission rate from the combustion of the enriched fuel and the emission rate from the normal fuel, the survivability of the compound could be calculated and thus the contribution to the compound in DPM from the fuel can be obtained.

48 In this study, similar method was used to study the contribution from unburned diesel fuel to the compound in particulate diesel emissions. From preliminary study and literature results, it was observed that phenanthrene was abundant in both diesel fuel and diesel emissions.

Therefore, phenanthrene was added into low sulfur diesel fuel, which is referred to as phenanthrene-enriched (Ph-enriched) fuel. The amount of phenanthrene added should be controlled so that this additive would not alter the combustion process in diesel engine.

Approximately 1% (by weight) of phenanthrene was added into diesel fuel [62]. Phenanthrene content in normal fuel and phenanthrene-enriched fuel was measured. DPM samples were collected for both fuels under the same engine operation conditions. Then the amount of phenanthrene in the two types of samples was measured by GC/MS. The following equations were used to calculate survivability of phenanthrene and the percentage of phenanthrene contributed by unburned diesel fuel in the total particulate phenanthrene emission:

(F − F ) × Q Survivability = 2 1 S ÷106 (2-1) (C2 − C1 ) × RDF

CPh × RDF × Survivability 6 PUBDF = ×10 (2-2) FPh × QS where F1 and F2 are the concentration of phenanthrene in DPM for normal fuel and

3 3 phenanthrene-enriched fuel (ng/m ), QS is the stack gas flow rate at standard conditions (m /hr),

C1 and C2 are the concentration of phenanthrene in normal fuel and Ph-enriched fuel (mg/L), RDF is fuel consumption rate (L/hr), PUBDF is the percentage of phenanthrene contributed by unburned diesel fuel in the total particulate phenanthrene emission (%), CPh is the concentration of

3 phenanthrene in a fuel used (mg/L), FPh is the concentration of phenanthrene in DPM (ng/m ), and 106 is the conversion factor to convert mg to ng. The assumption for Eq. (2-1) and (2-2) is

49 that the amount of phenanthrene from sources other than unburned diesel fuel, which are mainly pyrosynthetic formation, is not affected by the addition of phenanthrene.

Low sulfur diesel fuel with phenanthrene concentration of 198 mg/L was used as the normal fuel. After phenanthrene was added, its concentration in Ph-enriched fuel was 10000 mg/L. The experiment was performed at 0 kW and DPM samples were collected using high volume dilution sampler.

2.6 References

1. Manual of the Generac diesel generator (Model SD080), Tri-state generator, http://www.powercompany.org/PDF/SD6080.pdf, accessed July 2005.

2. Federal Register, Part II, Environmental Protection Agency. Amendments for testing and monitoring provisions. Final Rule, 40 CFR Part 60, 61, and 63; Tuesday, October 17, 2000. See http://www.epa.gov/ttn/emc/promgate.html for method details.

3. U.S. Department of Health and Human Services (DHHS), Public Health Service, Centers for Disease Control, National Institute for Occupational Safety and Health (NIOSH). In NIOSH Manual of Analytical Methods (NMAM); Cassinelli, M.E., O'Connor, P.F., Eds.; Second Supplement to NMAM, 4th ed.; Publication No. 94-113. Cincinnati, OH, 1998.

4. Birch, M.E.; and Cary, R.A., Elemental carbon-based method for monitoring occupational exposures to particulate diesel exhaust, Aerosol Sci. Technol., 25(3): 221- 241, 1996.

5. Kittelson, D.B.; Watts, W.F.; and Johnson, J.P., Particle Sampling Methodology: On-road and laboratory measurements of nanoparticles from Diesel engines. Presentation to the 11th CRC On-road Vehicle Emissions Workshop, San Diego, CA, March 26-28, 2001. Available at http://www.me.umn.edu/centers/mel/, accessed August 2006.

6. Kittelson, D.B.; Watts, W.F.; and Johnson, J.P., On-road nanoparticle measurements, 2003. Proc. 8th International Conference on Environmental Science and Technology, Lemnos Island, Greece, September 8-10, 2003, Vol. A, pp. 453-460. Athens: Global Network of Environmental Science and Technology (GLOBAL NESTGNEST) and the Department of Environmental Studies of the University of the Aegean.

7. Kittelson, D.B.; Watts, W.F.; and Johnson, J.P., Diesel aerosol sampling methodology - CRC E-43 technical summary and conclusions, 2002. Available at http://www.me.umn.edu/centers/mel/, accessed August 2006.

50 8. Kittelson, D.B.; Watts, W.F.; and Johnson, J.P., Nanoparticle emissions on Minnesota highways, Atmos. Environ., 38(1): 9-19, 2004.

9. Kittelson, D.B.; Watts, W.F.; Johnson, J.P.; et al., Fine particle (nanoparticle) emissions on minnesota highways, 2001. Presentation to 7th Diesel Engine Emissions Reduction (DEER) Workshop, Portsmouth, VA, August 5-9, 2001, available at http://www.me.umn.edu/centers/mel/, accessed August 2006.

10. Kittelson, D.B.; Watts, W.F.; Johnson, J.P.; et al., On-road exposure to highway aerosols. 1. Aerosol and gas measurements, Inhal. Toxicol., 16: 31-39, 2004.

11. Kittelson, D.B.; Watts, W.F.; Johnson, J.P.; et al., Gasoline vehicle exhaust particle sampling study, 2003. Proc. of 9th Diesel Engine Emissions Reduction Conference (DEER 2003), August 24-28, 2003, Newport, RI. Washington, DC: Department of Energy. Available at http://www.osti.gov/fcvt/deer2003/kittelsonpresentation.pdf, accessed August 2006.

12. Liu, Z.F., The carbon and sulfur speciation of diesel emissions from a non-road generator, M.S. thesis, University of Cincinnati, Cincinnati, OH, 2005.

13. Kittelson, D.B.; and Abdul-Khalek, I.S. Formation of nanoparticles during exhaust dilution. in Proceedings of 2nd ETH Workshop on Nanoparticle Measurement. August 1998. Zurich, Switzerlan.

14. Lipsky, E.; Stanier, C.O.; Pandis, S.N.; et al., Effects of sampling conditions on the size distribution of fine particulate matter emitted from a pilot-scale pulverized-coal combustor, Energy Fuels, 16(2): 302-310, 2002.

15. Hildemann, L.M.; Cass, G.R.; and Markowski, G.R., A dilution stack sampler for collection of organic aerosol emissions: design, characterization and field tests, Aerosol Sci. Technol., 10(1): 193-204, 1989.

16. Exhaust gas sampling and analytical system for gaseous emissions from heavy-duty diesel-fueled engines and particulate emissions from all engines. Electronic Code of Federal Regulations (eCFR), Title 40, Part 86.1310-2007.

17. Exhaust gas sampling and analytical system; diesel engines. Electronic Code of Federal Regulations (eCFR), Title 40, Part 86.1310-90.

18. Hildemann, L.M.; Markowski, G.R.; and Cass, G.R., Chemical composition of emissions from urban sources of fine organic aerosol, Environ. Sci. Technol., 25(4): 744-759, 1991.

19. Hildemann, L.M.; Markowski, G.R.; Jones, M.C.; et al., Submicrometer aerosol mass distributions of emissions from boilers, fireplaces, automobiles, diesel trucks, and meat- cooking operations, Aerosol Sci. Technol., 14(1): 138-152, 1991.

51 20. Schauer, J.J.; Kleeman, M.J.; Cass, G.R.; et al., Measurement of emissions from air pollution sources. 2. C1 through C30 organic compounds from medium duty diesel trucks, Environ. Sci. Technol., 33(10): 1578-1587, 1999.

21. Zielinska, B.; Sagebiel, J.; Arnott, W.P.; et al., Phase and size distribution of polycyclic aromatic hydrocarbons in diesel and gasoline vehicle emissions, Environ. Sci. Technol., 38(9): 2557-2567, 2004.

22. Duran, A.; de Lucas, A.; Carmona, M.; et al., Simulation of atmospheric PAH emissions from diesel engines, Chemosphere, 44(5): 921-924, 2001.

23. Farrarkhan, J.R.; Andrews, G.E.; Ishaq, R.; et al., Quantitative diesel particulate analysis using gas-chromatography mass-spectrometry, Proc. Inst. Mech. Eng. Part A-J. Power Energy, 207(A2): 95-106, 1993.

24. Gratz, L.D.; Bagley, S.T.; Leddy, D.G.; et al., Interlaboratory comparison of HPLC- fluorescence detection and GC/MS: analysis of PAH compounds present in diesel exhaust, J. Hazard. Mater., 74(1-2): 37-46, 2000.

25. Leotz-Gartziandia, E.; Tatry, V.; and Carlier, P., Sampling and analysis of polycyclic aromatic hydrocarbons (PAH) and oxygenated PAH in diesel exhaust and ambient air, Polycycl. Aromat. Compd., 20(1-4): 245-258, 2000.

26. McDonald, J.D.; Zielinska, B.; Sagebiel, J.C.; et al., Characterization of fine particle material in ambient air and personal samples from an underground mine, Aerosol Sci. Technol., 36(11): 1033-1044, 2002.

27. Hawthorne, S.B.; Grabanski, C.B.; Martin, E.; et al., Comparisons of Soxhlet extraction, pressurized liquid extraction, supercritical fluid extraction and subcritical water extraction for environmental solids: recovery, selectivity and effects on sample matrix, J. Chromatogr. A, 892(1-2): 421-433, 2000.

28. Pineiro-Iglesias, M.; Lopez-Mahia, P.; Vazquez-Blanco, E.; et al., Comparison between Soxhlet, ultrasonic and microwave assisted extraction of polycyclic aromatic hydrocarbons from atmospheric particulate matter, Fresenius Environ. Bull., 9(1-2): 17- 22, 2000.

29. Arditsoglou, A.; Terzi, E.; Kalaitzoglou, M.; et al., A comparative study on the recovery of polycyclic aromatic hydrocarbons from fly ash and lignite coal, Environ. Sci. Pollut. Res., 10(6): 354-356, 2003.

30. Ochsenkuhn-Petropoulou, M.; Staikos, K.; Matuschek, G.; et al., On-line determination of polycyclic aromatic hydrocarbons in airborne particulate matter by using pyrolysis/GC-MS, J. Anal. Appl. Pyrolysis, 70(1): 73-85, 2003.

31. Pandit, G.G.; Srivastava, P.K.; Sharma, S.; et al., Monitoring of persistent organic pollutants (POPs) in aerosols using HPLC, J. Liq. Chromatogr. Relat. Technol., 25(8): 1271-1281, 2002.

52 32. Shu, Y.Y.; Tey, S.Y.; and Wu, D.K.S., Analysis of polycyclic aromatic hydrocarbons in airborne particles using open-vessel focused microwave-assisted extraction, Anal. Chim. Acta, 495(1-2): 99-108, 2003.

33. Swartz, E.; and Stockburger, L., Recovery of semivolatile organic compounds during sample preparation: Implications for characterization of airborne particulate matter, Environ. Sci. Technol., 37(3): 597-605, 2003.

34. Schauer, J.J.; Kleeman, M.J.; Cass, G.R.; et al., Measurement of emissions from air pollution sources. 1. C1 through C29 organic compounds from meat charbroiling, Environ. Sci. Technol., 33(10): 1566-1577, 1999.

35. Rogge, W.F.; Hildemann, L.M.; Mazurek, M.A.; et al., Sources of Fine Organic Aerosol.2. Noncatalyst and Catalyst- Equipped Automobiles and Heavy-Duty Diesel Trucks, Environ. Sci. Technol., 27(4): 636-651, 1993.

36. Liu, Z.F.; Lu, M.M.; Birch, M.E.; et al., Variations of the particulate carbon distribution from a nonroad diesel generator, Environ. Sci. Technol., 39(20): 7840-7844, 2005.

37. Fernandes, M.B.; and Brooks, P., Characterization of carbonaceous combustion residues: II. Nonpolar organic compounds, Chemosphere, 53(5): 447-458, 2003.

38. Sauvain, J.J.; Duc, T.V.; and Guillemin, M., Exposure to carcinogenic polycyclic aromatic compounds and health risk assessment for diesel-exhaust exposed workers, Int. Arch. Occup. Environ. Health, 76(6): 443-455, 2003.

39. Sjogren, M.; Li, H.; Rannug, U.; et al., A multivariate statistical analysis of chemical composition and physical characteristics of 10 diesel fuels, Fuel, 74(7): 983-989, 1995.

40. Tolbert, P.E.; Klein, M.; Metzger, K.B.; et al., Interim results of the study of particulates and health in Atlanta (SOPHIA), J. Expo. Anal. Environ. Epidemiol., 10(5): 446-460, 2000.

41. Adonis, M.; Martinez, V.; Riquelme, R.; et al., Susceptibility and exposure biomarkers in people exposed to PAHs from diesel exhaust, Toxicol. Lett., 144(1): 3-15, 2003.

42. Lee, S.L.; De Wind, M.; Desai, P.H.; et al., Aromatics reduction and cetane improvement of diesel fuels, Fuel Reformulation, 5: 26-31, 1993.

43. Pal, R.; Juhasz, M.; and Stumpf, A., Detailed analysis of hydrocarbon groups in diesel range petroleum fractions with on-line coupled supercritical fluid chromatography gas chromatography mass spectrometry, J. Chromatogr. A, 819(1-2): 249-257, 1998.

44. Reilly, P.T.A.; Gieray, R.A.; Whitten, W.B.; et al., Real-time characterization of the organic composition and size of individual diesel engine smoke particles, Environ. Sci. Technol., 32(18): 2672-2679, 1998.

53 45. Shi, J.P.; Mark, D.; and Harrison, R.M., Characterization of particles from a current technology heavy-duty diesel engine, Environ. Sci. Technol., 34(5): 748-755, 2000.

46. Kelly, K.E.; Wagner, D.A.; Lighty, J.S.; et al., Characterization of exhaust particles from military vehicles fuelled with diesel, gasoline, and JP-8, J. Air Waste Manage. Assoc., 53(3): 273-282, 2003.

47. Gray, H.A.; Cass, G.R.; Huntzicker, J.J.; et al., Characteristics of atmospheric organic and elemental carbon particle concentrations in Los Angeles, Environ. Sci. Technol., 20(6): 580-589, 1986.

48. Poster, D.L.; de Alda, M.J.L.; Schantz, M.M.; et al., Development and analysis of three diesel particulate-related standard reference materials for the determination of chemical, physical, and biological characteristics, Polycycl. Aromat. Compd., 23(2): 141-191, 2003.

49. Wise, S.A., Standard Reference Materials (SRMs) for the determination of polycyclic aromatic compounds - Twenty years of progress, Polycycl. Aromat. Compd., 22(3-4): 197-230, 2002.

50. Wise, S.A.; Sander, L.C.; and May, W.E., Determination of polycyclic aromatic hydrocarbons by liquid chromatography, J. Chromatogr., 642(1-2): 329-349, 1993.

51. Schubert, P.; Schantz, M.M.; Sander, L.C.; et al., Determination of polycyclic aromatic hydrocarbons with molecular weight 300 and 302 in environmental-matrix standard reference materials by gas chromatography/mass spectrometry, Anal. Chem., 75(2): 234- 246, 2003.

52. Hoffmann, E.D.; Charette, J.; and Stroobant, V., Mass Spectrometry-Chromatography Coupling; Tandem Mass Spectrometry (MS/MS), in Mass Spectrometry Principles and Applications. Wiley: Paris. pp. 99-142, 1996.

53. Nicol, S.; Dugay, J.; and Hennion, M.C., Simultaneous determination of polycyclic aromatic dydrocarbons and their nitrated derivatives in airborne particulate matter using gas chromatography-tandem mass spectrometry, J. Sep. Sci., 24(6): 451-458, 2001.

54. Nicol, S.; Dugay, J.; and Hennion, M.C., Determination of oxygenated polycyclic aromatic compounds in airborne particulate organic matter using gas chromatography tandem mass spectrometry, Chromatographia, 53: S464-S469, 2001.

55. Sandra, P.; Beltran, J.; and David, F., Enhanced selectivity in the determination of triazines in environmental samples by benchtop CGC-MS-MS, HRC-J. High Resolut. Chromatogr., 18(9): 545-550, 1995.

56. Buchholz, B.A.; Mueller, C.J.; Martin, G.C.; et al., Tracing fuel component carbon in the emissions from diesel engines, Nucl. Instrum. Methods Phys. Res. Sect. B-Beam Interact. Mater. Atoms, 223-24: 837-841, 2004.

54 57. Rhead, M.M.; and Hardy, S.A., The sources of polycyclic aromatic compounds in diesel engine emissions, Fuel, 82(4): 385-393, 2003.

58. Rhead, M.M.; and Pemberton, R.D., Sources of naphthalene in diesel exhaust emissions, Energy Fuels, 10(3): 837-843, 1996.

59. Tancell, P.J.; Rhead, M.M.; Pemberton, R.D.; et al., Survival of polycyclic aromatic hydrocarbons during diesel combustion, Environ. Sci. Technol., 29(11): 2871-2876, 1995.

60. Tancell, P.J.; Rhead, M.M.; Trier, C.J.; et al., The sources of benzo[a]pyrene in diesel exhaust emissions, Sci. Total Environ., 162(2-3): 179-186, 1995.

61. Hori, S.; Sato, T.; and Narusawa, K., Effects of diesel fuel composition on SOF and PAH exhaust emissions, JSAE Review, 18(3): 255-261, 1997.

62. Mi, H.H.; Lee, W.J.; Chen, C.B.; et al., Effect of fuel aromatic content on PAH emission from a heavy-duty diesel engine, Chemosphere, 41(11): 1783-1790, 2000.

55 Chapter 3

THE ORGANIC COMPOSITION OF DIESEL PARTICULATE

MATTER, DIESEL FUEL AND ENGINE OIL OF A NON-ROAD

DIESEL GENERATOR

3.1 Introduction

Three sources may contribute to diesel emissions: combustion process, unburned diesel fuel, and engine oil. Elemental carbon, as well as some fractions of organic carbon, was generated during the combustion process. In addition to combustion (pyrosynthesis), the organic fractions of DPM may be resultant from unburned diesel fuel and engine oil. The composition of the diesel fuel has been identified and categorized into fractions of fuel input, such as 65-85% saturates, 5-30% aromatics, and 0-5% olefins [1]. The percentage may vary with manufactures, the mining locations, the refining processes and sulfur content. Sjogren et al. studied ten diesel fuels from different suppliers, and obtained considerably differed percentages for each category

[2]. High sulfur content in the diesel fuels has been associated with more DPM emissions [3].

More stringent regulations have been proposed to reduce sulfur content in the near future to reduce particulate emissions. Engine oil is also believed to be a possible source of DPM. It was reported that some organic compounds in DPM are emitted from the evaporation of engine oil, such as some short-chain alkanoic acids, hopanes and steranes [4, 5].

Despite the progress in DPM sampling and analysis, there are still gaps to fill, especially in the compositional correlations among DPM, diesel fuel and engine oil. The aim of this study is to compare the organic composition of DPM to diesel fuel and engine oil and seek the correlation

56 of DPM from petrogenic (raw materials such as diesel fuel and engine oil) and pyrogenic

(combustion) sources.

3.2 Experimental method

In this study, a Generac diesel generator (1992, Model SD080, model No. 92A-03040-S) was used as the non-road source of diesel particulate matter. A high volume sampler with a flow rate of approximately 300 L/min was used for sampling DPM with the sampling time 1 hour.

The DPM was taken at idle mode (0 kW) on quartz filters. The low sulfur diesel fuel (Steve

Krebs Oil Company, Inc., sulfur content: 433ppmw), engine oil from this generator, and DPM samples were analyzed with GC/MS. The detailed experimental setup, DPM sampling, extraction procedures, and instrumental analysis are detailed described in Chapter 2.

3.3 Results and discussion

In this study, the sample was taken at the exit of the exhaust pipe, and natural dilution was used. The DPM mass concentration is approximately 3.7 mg/m3. Compared with other studies, our DPM mass concentration is in the range of the reported results [3]. However, the

DPM emissions can vary significantly due to the size and type of diesel machines used, the sampling method, and the load reported. The mass concentration of DPM from a Jeep Bobtail

(engine speed: 1700 rpm, engine load: 9-87 kW) reported by Zielinska et al. is in the range of

1687-3800 µg/m3 [6]. Chan et al. reported that the DPM mass concentration from the Ford FSD

425T turbocharged DI diesel engine (four-cylinder, direct injection, 800-4000 rpm, maximum power output 52 kW) was from 0.2 mg/m3 to more than a hundred mg/m3 at zero load condition

57 [7]. In addition, the DPM emissions from diesel vehicles are often reported as mass per mile, which further complicates the comparison.

3.3.1 Chemical Composition

Figure 3-1 (a) and (b) shows the total ion chromatograms (TIC) of diesel fuel and DPM from GC/MS, respectively. The chemical composition of low sulfur diesel fuel, DPM and engine oil is listed in Table 3-1, and the compounds labeled “ND” (not detected) can be either below the detection limit or not present in the sample. Figure 3-2 summed up the compounds by categories, i.e. n-alkanes, branched alkanes, saturated cycloalkanes, PAHs, alkylated PAHs, alkylbenzenes, n-alkanoic acids and aromatic acids for the diesel fuel and DPM, and the relative fractions of each category were calculated. It is indicated that DPM contains higher fractions of PAHs, alkylated PAHs and n-alkanoic acids comparing with the diesel fuel, while the diesel fuel has more fractions of branched alkanes and cycloalkanes.

(a) Diesel fuel. Peaks 1–11: n-alkanes C10–C20

(b) DPM. Peaks 1–8: n-alkanes C14–C21; peak 9: Phe; peak 10: 2-MN; peak 11: 1-MN; peak 12: DMN; peak 13: TMN; peak 14: MPh; peak 15: DMPh.

Figure 3-1 Total ion chromatogram of diesel fuel and DPM (time in minutes).

59 Table 3-1 Organic compounds present in low sulfur diesel fuel, engine oil and diesel particulate mattera. Compounds Diesel fuel DPM composition Engine oil composition (µg/g) (µg/g) composition (µg/g) n-Alkanes n-Decane (C10) 12115 9.71 ND n-Undecane (C11) 11271 21.6 ND n-Dodecane (C12) 17149 31.3 ND n-Tridecane (C13) 28834 57.1 ND n-Tetradecane (C14) 25604 204 ND n-Pentadecane (C15) 27660 564 ND n-Hexadecane (C16) 23965 1164 ND n-Heptadecane (C17) 26082 2645 ND n-Octadecane (C18) 8727 1479 ND n-Nonadecane (C19) 4988 1257 ND n-Eicosane (C20) 2193 742 ND n-Heneicosane (C21) 1092 138 ND n-Docosane (C22) 756 10.3 ND n-Tricosane (C23) 220 7.42 ND n-Tetracosane (C24) 107 5.19 ND n-Pentacosane (C25) ND 3.57 ND Sum of n-alkanes 190763 8341

Branched alkanes

Norfarnesane (C14) 11469 12.8 ND

Farnesane (C15) 9719 28.4 ND

Norpristane (C18) 7992 135 ND

Pristine (C19) 5871 147 ND

Phytane (C20) 4775 185 ND Other branched alkanes 328578 7084 ND Sum of branched alkanes 368404 7592

Cycloalkanes n-Heptylcyclohexane (C13) 13144 11.1 ND n-Octylcyclohexane (C14) 11467 26.8 ND n-Nonylcyclohexane (C15) 10582 28.5 ND n-Decylcyclohexane (C16) 9135 43.3 ND n-Undecylcyclohexane (C17) 6207 38.8 ND n-Dodecylcyclohexane (C18) 2073 60.7 ND n-Tridecylcyclohexane (C19) 165 12.3 ND n-Tetradecylcyclohexane (C20) 25 9.82 ND n-Pentadecylcyclohexane (C21) ND 6.79 ND

60 Table 3-1 (Continued) Compounds Diesel fuel DPM composition Engine oil composition (µg/g) (µg/g) composition (µg/g) n-Hexadecylcyclohexane (C22) ND 4.29 ND n-Heptadecylcyclohexane (C23) ND ND ND n-Octadecylcyclohexane (C24) ND ND ND n-Nonadecylcyclohexane (C25) ND ND ND Sum of saturated cycloalkanes 52798 242

PAHs Naphthalene (Nap) 753 4.06 ND Acenaphthylene (Acy) 159 2.03 ND Acenaphthene (Ace) 85 3.04 ND Fluorene (Flu) 100 7.10 ND Phenanthrene (Phe) 247 77.1 ND Anthracene (Ant) 7.5 4.06 ND Fluoranthene (Flt) ND 25.4 ND Pyrene (Pyr) 5.0 67.0 ND Benzo(a)anthracene (Baa) ND 5.07 ND Chrysene (Chy) ND 5.07 ND Benzo(b)fluoranthene (Bbf) ND 3.04 ND Benzo(k)fluoranthene (Bkf) ND 2.03 ND Benzo(a)pyrene (Bap) ND 4.06 ND Indeno[1,2,3-cd]pyrene (Ind) ND 6.61 ND Dibenz(a,h)anthracene (Dba) ND 1.65 ND Benzo(ghi)perylene (Bgp) ND 3.04 ND Biphenyl 437 25.9 ND Sum of PAHs 1793 246

Alkylated PAHs 1-Methylnaphthalene (1-MN) 585 29.8 ND 2-Methylnaphthalene (2-MN) 2291 86.0 ND 1,2-Dimethylnaphthalene (1,2-DMN) 373 31.0 ND 1,3-Dimethylnaphthalene (1,3-DMN) ND 235 ND 1,4-Dimethylnaphthalene (1,4-DMN) 1540 ND ND 1,5-Dimethylnaphthalene (1,5-DMN) ND 134 ND 1,6-Dimethylnaphthalene (1,6-DMN) 1807 61.0 ND 1,7-Dimethylnaphthalene (1,7-DMN) 2548 24.8 ND 1,8-Dimethylnaphthalene (1,8-DMN) ND ND ND 2,3-Dimethylnaphthalene (2,3-DMN) ND ND ND 2,6-Dimethylnaphthalene (2,6-DMN) 1224 54.2 ND 2,7-Dimethylnaphthalene (2,7-DMN) 1837 48.7 ND

61 Table 3-1 (Continued) Compounds Diesel fuel DPM composition Engine oil composition (µg/g) (µg/g) composition (µg/g) Methylethylnaphthalene (MEN) 273 ND Trimethylnaphthalene (TMN) 12327 2641 ND 1-Methylphenanthrene (1-MPh) 242 1770 ND 2-Methylphenanthrene (2-MPh) 528 2168 ND Dimethylphenanthrene (DMPh) ND 3420 ND Sum of alkylated PAHs 25302 10977

Alkylbenzenes

Toluene (C1-B) 1377 10.7 ND

C2-Benzenes (C2-B) 12932 20.5 ND

C3-Benzenes (C3-B) 10003 14.9 ND

C4-Benzenes (C4-B) 9724 15.7 ND

C5-Benzenes (C5-B) 5538 6.15 ND

C6-Benzenes (C6-B) 5222 3.17 ND Sum of alkylbenzenes 44796 71.1 n-Alkanoic acids

Hexanoic acid (C6) ND ND ND

Heptanoic acid (C7) ND 17.8 37.4

Octanoic acid (C8) ND 39.6 55.9

Nonanoic acid (C9) ND 87.8 98.7

Decanoic acid (C10) ND 289 120

Undecanoic acid (C11) ND 354 100

Dodecanoic acid (C12) ND 591 1240

Tridecanoic acid (C13) ND 444 237

Tetradecanoic acid (C14) ND 1955 3400

Pentadecanoic acid (C15) ND 1237 1429

Hexadecanoic acid (C16) ND 1991 6109

Heptadecanoic acid (C17) ND ND ND

Octadecanoic acid (C18) ND 4.08 ND Sum of n-alkanoic acids 7012 12527

Aromatic acids Benzoic acid ND 226 495 a All of the compounds detected were quantified with standard compounds. ND: not detected, i.e. the concentration is lower than detection limit or the compound is not present in the sample.

62 100% Alkylbenzenes, 6.6% 90% (4.5% of DF) Unidentified 80% (31.6%) Alkylated PAHs, 3.7% (2.5% of DF) 70%

60% PAHs, 0.26% (0.18% of DF) 50% Cycloalkanes, 7.7% 40% Identified (5.3% of DF) (68.4%) 30% Branched alkanes, 53.9% (36.8% of DF) 20% n-Alkanes, 27.9% 10% (19.1% of DF) 0%

(a) Diesel fuel Aromatic acids, 0.65% 100% (0.02% of DPM) Insoluble n-Alkanoic acids, 20.2% 16% Unidentified (0.70% of DPM) 80% 95.9% (80.5% of DPM) Alkylbenzenes, 0.21% (0.007% of DPM)

60% Alkylated PAHs, 31.7% (1.1% of DPM) Soluble 84% PAHs, 0.64% 40% (0.02% of DPM) Cycloalkanes, 0.70% Identified (0.02% of DPM) 20% 4.1% (3.5% of DPM) Branched alkanes, 21.9% (0.76% of DPM) n-Alkanes, 24.0% 0% (0.83% of DPM) (b) DPM

Figure 3-2 Concentration and percentage of each category of compounds in low sulfur diesel fuel and DPM.

Approximately 70% (on mass basis) of the diesel fuel has been identified. Among the identified fractions, n-alkanes, branched alkanes, saturated cycloalkanes, PAHs, alkylated PAHs and alkylbenzenes account for 27.90%, 53.87%, 7.72%, 0.26%, 3.70% and 6.55% respectively. 63 This is consistent with several published studies [2, 8, 9]. For branched and cyclic alkanes, studies have indicated that compounds ranging from C10-C20 are the most abundant [10], and the

C10-C28 fraction ranges from 65-85% [1], and 72-96% for the C8-C24 fraction [2]. From Table 3-

1, the aromatic compounds in the low sulfur diesel fuel include naphthalene, fluorene, phenanthrene, and alkylated naphthalenes and phenanthrenes with a total percentage of about 4%.

Other studies indicated that volume fraction of PAHs in diesel fuels varies widely from 5% to

30%, even less than 5% for some high-quality fuels. Some heavy PAHs identified in DPM, such as fluoranthene, benzo(a)anthracene, chrysene, benzo(b)fluoranthene, benzo(k)fluoranthene, benzo(a)pyrene, indeno[1,2,3-cd]pyrene, dibenz(a,h)anthracene, and benzo(ghi)perylene, are not identified in diesel fuel. This is consistent with some other studies where those heavy PAHs are either not present or below detection level in diesel fuel [1, 2, 5].

1.3% of engine oil was able to be resolved by GC-MS, which are mainly organic acids. n-

Alkanoic acids from C6H12O2 to C18H36O2 are the main components and benzoic acid is the most important aromatic acid in the 1.3% of engine oil that is amenable to identification by GC/MS

(Table 3-1).

84% of DPM can be dissolved in the solvent. This is consistent with the 82-86 % of organic fraction obtained from thermal-optical analysis (NIOSH Method 5040, a standard method for organic and elemental carbon (OC/EC) measurement) [11].

The consistency of dissolved fraction with the fraction of organic compounds also indicated the high efficiency of the extraction method. In DPM, the identified organics account for 3.47% of total DPM mass, which falls into the range of other studies [5, 12, 13].

64 Among the identified fractions, n-alkanes account for 24.02%, branched alkanes 21.90%, alkylated aromatic hydrocarbons 31.67% and n-alkanoic acids 20.23%. This distribution is consistent with the DPM emissions from medium duty diesel trucks [5].

3.3.2 Alkanes

As suggested in Figure 3-2, alkanes, which include n-alkanes, branched alkanes and saturated cycloalkanes, account for a large fraction of the diesel fuel and the identifiable portion of DPM. The relative percentages of n-alkanes, branched alkanes and saturated cycloalkanes in both the diesel fuel and DPM were illustrated in Figures 3-3(a-c) respectively.

A series of n-alkanes from C10 to C25 have been identified in diesel fuel and DPM. From

Table 3-1, the concentrations of n-alkanes in DPM range from several hundred to several thousand µg/g while in diesel fuel the concentrations can reach as high as tens of thousands µg/g.

Figure 3-3(a) suggests that in the identified alkanes the most abundant species are distinct for diesel fuel and DPM. For diesel fuel, the most abundant identified n-alkanes are C13-C17, which is consistent with Schauer’s results [5]. While for DPM, the distribution represents a bell shape with C17 as the most abundant. The relative fractions of C17-C21 are much higher in DPM than in diesel fuel while the fractions of C10 to C16 are much less. The enrichment of aliphatic hydrocarbons with higher carbon number was also reported by Rogge et al. [13, 14]. The shift to heavier n-alkanes from diesel fuel to DPM can be the results of combustion synthesis, as shorter- chained alkanes are more likely to undergo complete combustion than the heavier ones.

65 8.0 7.0 Diesel fuel 6.0 DPM 5.0 4.0 3.0 2.0 1.0 0.0 Percentage of identified organics (%)

C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25 (a) n-Alkanes

1.4 1.6 1.4 1.2 Diesel fuel Diesel fuel DPM 1.2 1.0 DPM 1.0 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2

0.0 0.0 Percentage of identified organics (%) Percentageof identified organics(%) C14 C15 C18 C19 C20 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25 (b) Branched alkanes (c) Saturated cycloalkanes

Figure 3-3 Relative percentage of alkanes in total identified compounds for low sulfur diesel fuel and DPM.

For branched alkanes, the identified compounds are also known as isoprenoids. These compounds are naturally present in crude oil and can be potentially used as tracers for diesel engine emissions [5]. Branched alkanes phytane and pristane are produced by plankton from isoprenoids; both are relatively resistant to breakdown and are abundant in new oil sediments.

Phytane is more persistent than pristane, so the phytane/pristane ratio is sometimes used to age oil [15]. The ratio of pristane to heptadecane and phytane to octadecane for a given source of

66 diesel fuel is frequently distinctive enough to enable source identification in a spill investigation

[15]. These ratios also increase during biological degradation and are useful in evaluating the age of environmental contamination [15].

The structures of selected isoprenoid alkanes are shown in Figure 3-4. Figure 3-3(b) shows the distribution of the isoprenoids from C14 through C20 in diesel particulate matter compared with diesel fuel used for the generator. As can be seen in the figures, DPM contains a larger fraction of higher molecular weight compounds.

2,6,10-Trimethylundecane (Norfarnesane) 2,6,10-Trimethyldodecane (Farnesane)

2,6,10-Trimethyltridecane 2,6,10-Trimethylpentadecane (Norpristane)

2,6,10,14-Tetramethylpentadecane (Pristane) 2,6,10,14-Tetramethylhexadecane (Phytane)

Figure 3-4 Chemical structures of isoprenoids present in diesel fuel and diesel emissions.

Cycloalkanes are more toxic than alkanes or branched alkanes [15]. Figure 3-3(c) shows the similar distributions of saturated cycloalkanes from C13 through C22 in DPM and diesel fuel.

The shift toward higher molecular weight compounds is similar to that of n-alkanes.

3.3.3 PAHs and Alkylated PAHs

Figure 3-5 and Figure 3-6 illustrate the relative percentages of dominant PAHs and alkylated PAHs in diesel fuel and DPM. It can be seen that compared with diesel fuel, the relative fractions of PAHs and alkylated PAHs are significantly higher in DPM. 67 0.25 Diesel fuel 0.20 DPM

0.15

0.10

0.05

Percentage ofidentified organics0.00 (%) Flt Ind Flu Bbf Bkf Ant Pyr Nap Acy Ace Phe Baa Chy Bap Dba Bgp

Figure 3-5 Relative percentage of PAHs in total identified compounds in low sulfur diesel fuel and DPM.

12.0

10.0 Diesel fuel

8.0 DPM

6.0

4.0

2.0

Percentage of identified organics (%) 0.0 MN DMN TMN MPh DMPh

Figure 3-6 Relative percentage of alkylated PAHs in total identified compounds in low sulfur diesel fuel and DPM.

Many more PAH species are observed in DPM that are not present in diesel fuel, such as benzo(a)anthracene, chrysene, benzo(b)fluoranthene, benzo(k)fluoranthene, benzo(a)pyrene, indeno[1,2,3-cd]pyrene, dibenz(a,h)anthracene, and benzo(ghi)perylene, with the mass range of

228-278 u (Figure 3-7). More phenanthrene and pyrene are identified in DPM than the diesel fuel.

68 This result indicates that many PAH compounds may originate from the combustion process in the diesel engine rather than from direct evaporation of diesel fuel. These pyrogenic PAHs of greater carcinogenic potency are precursors to soot particle formation and account for 14% of total PAHs identified in DPM. Similar PAH distribution in DPM and diesel fuel was also reported by Dobbins [16].

1.00

0.80

0.60

0.40 Diesel fuel 0.20 DPM H/C Ratio of PAHs 0.00 100 150 200 250 300 Molecular Mass (u)

Figure 3-7 H/C ratio vs. PAH molecular mass found by GC/MS in low sulfur diesel fuel and DPM.

For identified alkylated PAHs, the dominant species are trimethylnaphthalene, methylphenanthrene and dimethylphenanthrene for DPM and dimethylnaphthalene and trimethylnaphthalene for diesel fuel. This also suggests the enrichment of higher molecular weight compounds in DPM. Consistent with Shauer et al. [5] and Koziel et al. [17], alkylated

PAHs are the most abundant identified species in our study.

3.3.4 Alkylbenzenes

Figure 3-8 represents percentages of alkylbenzenes from C1-benzene to C6-benzene in low sulfur diesel fuel and DPM. Alkylbenzenes account for 6.55% of the identified fraction in

69 diesel fuel and 0.21% in DPM. It is suggested that alkylbenzenes have been significantly consumed by the combustion processes. The dealkylation from alkylbenzenes and alkylated

PAHs and recombination of alkyl groups onto the reactants have altered the relative isomeric distributions of these two categories [18].

1.6 1.4 Diesel fuel 1.2 DPM 1.0 0.8 0.6 0.4 0.2 0.0 Percentageof identified organics(%) C1-B C2-B C3-B C4-B C5-B C6-B

Figure 3-8 Relative percentage of alkylbenzenes in total identified compounds in low sulfur diesel fuel and DPM.

3.3.5 Organic Acids

Both benzoic and n-alkanoic acids ranging from C6 to C18 have been identified in DPM and engine oil as listed in Table 3-1. The distributions of n-alkanoic acids in DPM and engine oil are also shown in Figure 3-9.

In DPM, organic acids account for 20% of the total identified organics with concentrations as high as several thousand µg/g. n-Dodecanoic (C12), n-tetradecanoic (C14), n- pentadecanoic (C15), and n-hexadecanoic acid (C16) are the most abundant species identified.

From Figure 3-9, similar distribution of n-alkanoic acids is observed in engine oil, which indicates that the organic acids from DPM may be originated from the evaporation of engine oil.

70 60

7 50 6 5 4 40 3 2 1 30 0 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 20 Engine oil 10 DPM Percentage of identifiedorganics (%) 0 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18

Figure 3-9 Relative percentage of n-alkanoic acids from C6H12O2 to C18H36O2 in DPM and engine oil.

3.3.6 Isomer Distribution

From the results, the distribution variation of isomers in diesel fuel and DPM was observed. The relative percentages of isomers for methylnaphthalenes and methylphenanthrenes in Ultra-low sulfur diesel fuel and DPM are shown in Figure 3-10 (a) and (b). Figure 3-10 indicates that the amount of 2-isomers is higher than 1-isomers in both diesel fuel and DPM. The theory of heat of formation can be used to explain the higher abundance of 2-isomers. Heat of formation of a compound refers to the heat evolved (heat of formation is negative) or absorbed

(heat of formation is positive) during the formation of one mole of a substance from its component elements. Higher heat of formation (positive) means more heat is needed to form this compound. That means it’s more difficult to form a compound with higher heat of formation.

The heat of formation is 32.10 kcal/mol for 1-methylnapthalene, 31.23 kcal/mol for 2- methylnapthalene, 51.23 kcal/mol for 1-methylphenanthrene, and 47.95 kcal/mol for 2-

71 methylphenanthrene (Heat of Formation calculated by WinMopac7.21 (PM3)). Therefore, 2- isomers are more stable and abundant in environment.

90 80 80 1-MN 70 1-MPh 70 2-MN 60 2-MPh 60 50 50 40 40 30 30

Percentage (%) 20 Percentage (%) 20 10 10 0 0 Diesel Fuel DPM Diesel Fuel DPM

(a) methylnaphthalenes (b) methylphenanthrene

Figure 3-10 Relative percentage of isomers in low sulfur diesel fuel and DPM.

Table 3-2 listed the ratio of 1-/2-isomers in diesel fuel and DPM. For the methylnaphthalenes, the isomer ratio in DPM is relatively consistent with that of the diesel fuel.

This can be an indication that the two isomers are consumed or formed at similar rates, and this may be caused by the close resemblance of the heats of formation between the two. For the methylphenanthrenes, however, higher concentrations are found in the DPM, which suggests that the two isomers may be pyrogenic. Higher 1-MPh/2-MPh ratio in DPM than in diesel fuel indicates 1-MPh is formed more than 2-MPh. The phenanthrene groups are more affected by combustion, which promotes the interest in studying phenanthrene enriched fuel.

Table 3-2 Ratio of 1-MN/2-MN and 1-MPh/2-MPh in diesel fuel and DPM. 1-MN/2-MN 1-MPh/2-MPh Diesel fuel 0.26 0.46 DPM 0.35 0.82

72 3.4 Conclusions

The chemical composition of low sulfur diesel fuel, engine oil and diesel particulate matter emitted from a non-road diesel generator has been investigated. A total of approximately

90 organic compounds are identified and quantified, including a series of n-alkanes, branched alkanes, saturated cycloalkanes, PAHs, alkylated PAHs, alkylbenzenes, n-alkanoic acids and aromatic acids. It is shown that the organic composition of DPM is mainly resulted from three sources: n-alkanes in DPM are likely from unburned diesel fuel, the main source of organic acids is from engine oil evaporation, and most PAHs and alkylated PAHs are likely formed during combustion. Compared with diesel fuel, DPM contains fewer fractions of alkanes and more PAH compounds. The enrichment of higher molecular weight alkanes, including n-alkanes, branched alkanes and cycloalkanes, is likely resultant from fuel combustion. The pyrogenic process also resulted in the loss of alkylbenzenes and the alteration of isomeric distribution of PAHs and alkylated PAHs.

3.5 References

1. Diesel fuel and exhaust emissions. 1996, World Health Organization, Geneva. p. 11-13.

2. Sjogren, M.; Li, H.; Rannug, U.; et al., A multivariate statistical analysis of chemical composition and physical characteristics of 10 diesel fuels, Fuel, 74(7): 983-989, 1995.

3. Saiyasitpanich, P.; Lu, M.M.; Keener, T.C.; et al., The effect of diesel fuel sulfur content on particulate matter emissions for a nonroad diesel generator, J. Air Waste Manage. Assoc., 55(7): 993-998, 2005.

4. Kawamura, K.; Ng, L.L.; and Kaplan, I.R., Determination of organic acids (C1-C10) in the atmosphere, motor exhausts, and engine oils, Environ. Sci. Technol., 19(11): 1082-1086, 1985.

5. Schauer, J.J.; Kleeman, M.J.; Cass, G.R.; et al., Measurement of emissions from air pollution sources. 2. C1 through C30 organic compounds from medium duty diesel trucks, Environ. Sci. Technol., 33(10): 1578-1587, 1999.

73 6. Zielinska, B.; Sagebiel, J.; Arnott, W.P.; et al., Phase and size distribution of polycyclic aromatic hydrocarbons in diesel and gasoline vehicle emissions, Environ. Sci. Technol., 38(9): 2557-2567, 2004.

7. Chan, S.H.; and He, Y.S., Measurements of particulate mass concentration using a tapered-element oscillating microbalance and a flame ionization detector, Meas. Sci. Technol., 10(4): 323-332, 1999.

8. Pal, R.; Juhasz, M.; and Stumpf, A., Detailed analysis of hydrocarbon groups in diesel range petroleum fractions with on-line coupled supercritical fluid chromatography gas chromatography mass spectrometry, J. Chromatogr. A, 819(1-2): 249-257, 1998.

9. Hsu, C.S., Diesel Fuels Analysis, in Encyclopedia of Analytical Chemistry. Wiley: New York. pp. 6613-6622, 2000.

10. Chemistry of diesel fuels. ed. Song, C.;Hsu, C.S.; and Mochida, I. Taylor & Francis: New York. p. 18-19, 2000.

11. Liu, Z.F.; Lu, M.M.; Birch, M.E.; et al., Variations of the particulate carbon distribution from a nonroad diesel generator, Environ. Sci. Technol., 39(20): 7840-7844, 2005.

12. McDonald, J.D.; Zielinska, B.; Sagebiel, J.C.; et al., Characterization of fine particle material in ambient air and personal samples from an underground mine, Aerosol Sci. Technol., 36(11): 1033-1044, 2002.

13. Rogge, W.F.; Hildemann, L.M.; Mazurek, M.A.; et al., Sources of fine organic aerosol. 2. Noncatalyst and catalyst equipped automobiles and heavy duty diesel trucks, Environ. Sci. Technol., 27(4): 636-651, 1993.

14. Rogge, W.F.; Hildemann, L.M.; Mazurek, M.A.; et al., Sources of fine organic aerosol.8. Boilers burning No. 2 distillate fuel oil, Environ. Sci. Technol., 31(10): 2731-2737, 1997.

15. Alkanes Entry, in Environmental Contaminants Encyclopedia, Irwin, R.J., Ed. National Park Service: Fort Collins,CO, 1997.

16. Dobbins, R.A.; Fletcher, R.A.; B. A. Benner; et al. PAHs in flames, diesel fuels and diesel emissions. in 30th International Symposium on Combustion, Work-In-Progress Poster Session. July 25-30, 2004. Chicago, IL.

17. Koziel, J.A.; Odziemkowski, M.; and Pawliszyn, J., Sampling and analysis of airborne particulate matter and aerosols using in-needle trap and SPME fiber devices, Anal. Chem., 73(1): 47-54, 2001.

18. Yang, J.; and Lu, M., Thermal growth and decomposition of methylnaphthalenes, Environ. Sci. Technol., 39(9): 3077-3082, 2005.

74 Chapter 4

PHASE DISTRIBUTION OF SEMI-VOLATILE ORGANIC

COMPOUNDS IN DIESEL COMBUSTION AEROSOLS

4.1 Introduction

4.1.1 Gas/Particle Partitioning

Diesel emissions consist of a large number of organic compounds [1-3], among which some polycyclic aromatic hydrocarbons (PAHs), such as chrysene and benzo[a]pyrene, are known or suspected human carcinogens [4-9]. Studies have indicated that occupational exposure to diesel exhaust may be related to lung cancer for bus garage workers, miners, and forklift workers etc. [2, 10]. It is well established that exposure to PAHs and their derivatives may result in a higher health risk [11]. Therefore, there is a necessity to understand the physical-chemical behavior of the carcinogenic and toxic organic compounds present in diesel exhaust. One of the important physical-chemical processes determining the behavior of organic compounds is the partitioning between the gas and particle phases. Such gas/particle (G/P) partitioning determines the distribution of a given compound between the two phases and is strongly related to human

0 -6 0.0 -8 exposure. With pure compound liquid vapor pressures ( pL ) in the range of 10 -10 Pa (~10 -

10-2 Torr), semi-volatile organic compounds (SVOCs), which include many of the PAHs and substituted PAHs, are present significantly in both gas and particle phases due to their intermediate volatilities. Therefore, SVOCs are of special interest in the G/P partitioning studies.

Volatile organic compounds (e.g., benzene) are mostly present in the gas phase, and low- volatility organic compounds are almost exclusively in the particulate phase.

75 The significance of understanding G/P partitioning for these organic compounds lies in the fact that their health effects are strongly related to the physical forms in which the compounds are present when exposure occurs. In addition, G/P partitioning is important in predicting the behavior of diesel emission tracers and the effectiveness of control technologies in removing diesel aerosols.

4.1.2 Partitioning Constant

There have been some theoretical and experimental studies in understanding the distribution of chemical species between the gas and particulate phases. Junge [12] proposed a linear Langmuir isotherm equation to describe gas/particle partitioning in the atmosphere. The partitioning of organic compounds between gas and particulate phases can be parameterized as follows [13, 14]:

F /TSP K = (4-1) p A

3 where Kp is a compound- and temperature-dependent partitioning coefficient (m /µg), F is the particle phase concentration of the compound of interest (ng/m3), A is the gas phase concentration (ng/m3), and TSP is the amount of total suspended particulate matter (µg/m3).

Generally, F represents the mass of a compound measured on a filter, and A represents the mass of a compound measured on a downstream adsorbent.

The constant Kp represents the sorbed/gaseous concentration ratio. Increasing Kp implies increasing partitioning to the solid phase. Multiplying Kp by TSP gives a ratio of the particle to gas phase concentrations:

Kp × TSP = F/A (4-2)

76 When Kp × TSP is greater than 1, the compound partitions primarily into the particle phase; Kp ×

TSP values less than 1 indicate partitioning primarily to the gas phase.

4.1.3 Partitioning Mechanisms and Predictive Models

Seinfeld and Pandis [15] summarized theories describing gas/particle partitioning of organic compounds. Dissolution and adsorption are the two most important mechanisms, which are similar to the absorption and adsorption mechanisms suggested by Pankow et al. [14, 16-21].

When organic species are already present in particles, organic vapors tend to dissolve in the particle phase organics. This is referred to as dissolution. The gas/particle partitioning of organic vapor species via dissolution is determined by the molecular properties of the species and the organic phase and the amount of the particulate organic phase available for the vapor phase compound to dissolve into. When an organic particulate phase is absent, adsorption is the mechanism to initiate gas/particle partitioning. The adsorption process involves two steps: the particle surface is partially covered by vapor molecules, and a monolayer is formed.

Occasionally additional layers can be formed following the monolayer, which can be regarded as the third step. The adsorption behavior of a surface is generally characterized by an adsorption isotherm. Langmuir isotherm is frequently used to describe the equilibrium adsorption for the formation of a monolayer. One of the most widely used isotherms is based on the BET theory, which is an extension of the Langmuir isotherm to include the adsorption of two or more molecular layers.

Pankow and co-workers have developed the modeling framework for the sorption processes in a series of studies [14, 16-21]. Two mechanisms describing the gas/particle partitioning of SVOCs have been recognized: adsorption onto the aerosol surface and absorption

77 into the aerosol organic matter. When the partitioning is dominated by adsorption, the partition coefficient is given by [14]:

(Q1 −Q2 ) / RT N s ATSPTe K p = 0 (4-3) 16 pL

2 where Ns is the number of adsorption sites (mol/cm ), ATSP is the surface area of the TSP

2 (cm /µg), T is the temperature (K), Q1 and Q2 are the enthalpies of desorption and volatilization, respectively (kcal/mol), R is the universal gas constant (= 8.314×10-3 kJ/mol·K = 0.00199

-5 3 0 kcal/mol·K = 8.2×10 m ·atm/mol·K), and pL is the sub-cooled liquid vapor pressure of the pure compound (torr). When partitioning is dominated by absorption into organic matter of aerosol,

Kp is given by [19]:

760 f om RT K p = 0 6 (4-4) MWomγpL10

where fom is the fraction of organic matter on the TSP, MWom is the mean molecular weight of the organic matter (g/mol), and γ is the activity coefficient of the absorbate in the organic matter.

Usually, it is difficult to experimentally determine the predominant mechanism in the environment. In any given situation, both adsorptive and absorptive partitioning may exist simultaneously. The observed value of Kp will then contain contributions from both mechanisms so that for a given compound [19]:

(Q1 −Q2 ) / RT 1 ⎡ N s ATSPTe 760 f om RT ⎤ K p = 0 ⎢ + 6 ⎥ (4-5) pL ⎣ 1600 MWomγ10 ⎦

78 The actual relative importance of adsorptive and absorptive partitioning in a given situation will

(Q1 −Q2 ) / RT depend on the values of the various parameters in Eq. (4-5); increasing ATSP and e will favor adsorption; increasing fom, decreasing MWom, and decreasing γ will favor absorption.

Regardless of the relative importance of adsorptive vs. absorptive partitioning, the partition coefficient is inversely proportional to the sub-cooled liquid vapor pressure. Taking the logarithm of Eq. (4-5) gives the linear relationship shown in Eq. (4-6):

0 log K p = mr log pL + br (4-6)

where mr and br (expressed as Eq. (4-7)) are the slope and intercept of the linear regression, respectively.

(Q1 −Q2 ) / RT N s ATSPTe 760 f om RT br = + 6 (4-7) 1600 MWomγ10

0 A plot of log Kp vs. log pL for a series of compounds may give a slope mr = -1 if br remains constant for the compounds of interest, which occurs only when γ is equal for each compound in the absorbed phase and particle size effects are negligible. However, γ may vary depending on the absorbing aerosol [21]. Therefore, values calculated from Eq. (4-5) are often verified with field measurements. The deviations from mr = -1 are possibly attributed to sampling artifacts, non-exchangeable material within particles, or non-equilibrium partitioning.

4.1.4 Impact Factors on Kp Values

In general, gas/particle partitioning of organic compounds is a function of a number of physical and chemical factors. First of all, the state of the compound in the emission (e.g. vapor, adsorbed on the surface of particles or absorbed into particles) implies the possible partitioning

79 mechanisms which directly determine the Kp values to be estimated from Eq. (4-3) or Eq. (4-4).

The second important factor is the physical and chemical properties of the compound which may

0 include the vapor pressure of the compound ( pL ), the reactivity and stability of the compound represented by its activity coefficient (γ), and the affinity of the compound for the particle’s organic matrix which is parameterized by Q1 and Q2 in Eq. (4-5). The amount of particles, which is expressed in terms of the number of adsorption sites (Ns) and the surface area of the TSP

(ATSP), and the composition of particles, which is described by the fraction of organic matter (fom)

[22], are essential for determining Kp when the dominant partitioning mechanism is adsorption and absorption, respectively. Pankow [16] has reported that low TSP level increases the deviation of calculated log[(F/TSP)/A] from the estimated Kp from Eq. (4-3). In addition, the collection methodology and sampling conditions, i.e. sampling temperature (T) [16], relative humidity (RH) [23], and filters (quartz or Teflon) [24], and sampling duration [25] may also influence the partitioning.

4.2 Experimental Method

The experiments were performed on a Generac diesel generator with burning low sulfur containing diesel fuel (400 ppmw S). The samples were collected using the developed dilution method (DR = ~3.4), with particle phase emissions collected on quartz fiber filters and gas phase emissions collected by the PUF/XAD/PUF cartridge. A detailed description of the experimental setup and procedures, sampling methodology, and extraction and analytical methods has been described in Chapter 2.

80 4.3 Results and Discussion

4.3.1 Phase Distribution

The gas phase and particle phase n-alkanes emitted from the diesel generator at 0 kW are shown in Figure 4-1. Overall, the amount of n-alkanes in gas phase is much higher than the amount in particle phase. Volckens and Leith [26] have reported that when the value of F/A is greater than 1, the compound is primarily present in the particle phase; the value less than 1 indicates partitioning primarily to the gas phase. The experimental results in this study are consistent with the theory: lighter alkanes (carbon number between 10 and 19) with F/A values in the range of 0.0-0.41 are predominantly in the gas phase, while heavier alkanes (carbon number higher than 20), which have the F/A values between 1.48 and 31.2, are primarily present in the particle phase. Table 4-1 listed the vapor pressure for some alkanes and PAHs at 25 ºC, which is close to the sampling temperature. From Figure 4-1 and the vapor pressure listed in

Table 4-1, alkanes occur primarily in gas phase when the vapor pressures are higher than 10-4.36 torr and mainly in particulate phase when the vapor pressures are lower than 10-4.84 torr.

250000 Gas phase ) 3 200000 Particle phase

150000

100000

50000 Concentration (ng/m

0 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25

Figure 4-1 The distribution of n-alkanes in gas and particle phase diesel emissions (400 ppmw S, 0 kW, DR = 3.4).

81 Table 4-1 Vapor pressures for some alkanes, PAHs and alkanoic acids at 25 ºC (torr). 0 0 Compounds Log pL Ref. Compounds Log pL Ref. n-Hexadecane (C16) -3.16 [27] Fluoranthene (Flt) -4.07 [27] n-Heptadecane (C17) -3.27 [27] Pyrene (Pyr) -4.24 [27] d n-Octadecane (C18) -3.72 [27] Benzo(a)anthracene (Baa) -5.71 [21, 23, 28] n-Nonadecane (C19) -4.36 [27] Chrysene (Chy) -5.42 [27] a n-Eicosane (C20) -4.84 [27] Benzo(b)fluoranthene (Bbf) -7.12 [15] a n-Heneicosane (C21) -5.18 [27] Benzo(k)fluoranthene (Bkf) -7.13 [15] a n-Docosane (C22) -5.70 [27] Benzo(a)pyrene (Bap) -7.33 [15] a e n-Tricosane (C23) -6.43 [23] Dodecanoic acid (C12) -5.48 [28] b n-Tetracosane (C24) -6.58 [29] Tetradecanoic acid (C14) -5.84 [30] c n-Pentacosane (C25) -7.08 [29] Pentadecanoic acid (C15) -6.12 [30]

Fluorene (Flu) -2.24 [27] Hexadecanoic acid (C16) -7.11 [30]

Phenanthrene (Phe) -2.98 [27] Heptadecanoic acid (C17) -7.08 [30]

Anthracene (Ant) -3.01 [27] Octadecanoic acid (C18) -8.39 [30] a @ 20 ºC. b,c @ 31 ºC. d The vapor pressure was estimated from the vapor pressures at other temperatures listed in references [21, 23, 28]. e @ 15 ºC.

Figure 4-2 represents the PAH distribution between gas phase and particle phase diesel

0 -3.01 emissions at 0 kW. As Figure 4-2 shows, 2- and 3-ring PAHs ( pL > 10 torr) are

0 -4.07 predominantly present in gas phase emission, 4- to 6-ring PAHs ( pL < 10 torr) are primarily

0 -5.71 present in particulate phase with 5- and 6-ring PAHs ( pL < 10 torr) occurring exclusively in particulate phase. The correlation between the gas/particle distribution of PAHs and their vapor pressures indicates that the relative abundance in the particulate phase increases as the vapor pressure decreases. Comparing G/P distributions for PAHs and alkanes, it has shown that the boundary vapor pressure between gas and particle phase differs for PAHs and alkanes, which suggests that the G/P partitioning depends not only on the vapor pressure, but also on some other physical-chemical properties of the compound. These physical-chemical properties include the reactivity and stability of the compound (γ), and the affinity of the compound for the particle’s organic matrix (Q1 and Q2). These parameters, especially the activity coefficient γ, determine the 82 differed br values according to Eq. (4-7). Higher activity coefficients for PAHs than those for alkanes were reported by Jang et al. [21], which may be a confirmation for the prediction in this study.

10000 ) 3 Gas 200 8000 150 Particle 100 6000 50 0 4000 Bkf Ind Bbf Chy Bgp Baa Bap 2000 Dba Concentration (ng/m 0 Flt Ind Flu Bbf Bkf Ant Pyr Nap Acy Ace Phe Baa Chy Bap Dba Bgp

Figure 4-2 The distribution of PAHs in gas and particle phase diesel emissions (400 ppmw S, 0 kW, DR = 3.4).

Figure 4-3 shows the concentrations of alkanoic acids in both gas and particle phase

0 -5.48 diesel emissions at 0 kW. Alkanoic acids with a carbon number between 6 and 12 ( pL > 10 torr) are primarily in gas phase, and particulate phase contains more fractions of high-molecular weight acids with the vapor pressures lower than 10-5.48 torr. It has been expected that the gas/particle partitioning of alkanoic acids differs from that of alkanes due to the differences between their physical-chemical properties, such as polarity and activity. The experimental results have indicated that alkanoic acid partitions less into particle phase than the alkane and

PAH with similar vapor pressure. For instance, tetradecanoic acid (C14) has similar vapor pressure as n-docosane (C22) and benzo(a)anthracene (Baa) (Table 4-1), the ratio of particle phase concentration/gas phase concentration (F/A) for the former is lower than that for the latter two: the F/A value is 1.60 for tetradecanoic acid, while 5.91, and 5.51 for n-docosane and benzo(a)anthracene, respectively. The reactivity and stability of the compound and the affinity of

83 the compound for the particle’s organic matrix are the two important factors affecting G/P partitioning. It has been found that alkanoic acids have similar activities coefficients as n-alkanes

(~1.0). Therefore, the affinity, which may be related to the polarity of the compound, is the possible reason leading to the difference in their partitioning behaviors of alkanoic acids and n- alkanes.

12000

) Gas phase 3 10000 Particle phase 8000

6000

4000

Concentration (ng/m 2000

0 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18

Figure 4-3 The distribution of alkanoic acids in gas and particle phase diesel emissions (400 ppmw S, 0 kW, DR = 3.4).

Though the above discussions are based on the data at 0 kW, similar results and conclusions can be obtained for other load conditions.

0 4.3.2 Log Kp vs. log pL

3 The partitioning coefficient (Kp, m /µg) was calculated from the measured gas and particle phase concentrations of a compound (A and F, ng/m3) and the TSP concentration (µg/m3)

0 in diesel emissions according to Eq. (4-1). Log Kp values were plotted vs. log pL (torr, @ 25 ºC) for n-alkanes and PAHs (Figure 4-4 and Figure 4-5, respectively) and the linear regressions were

2 illustrated in the two figures as well. The slope (mr), intercept (br), and regression coefficient (R )

84 values for the regressions were given in Table 4-2. Table 4-3 summarized the slopes and intercepts reported by other studies [17, 26, 31-33].

-1 Measured Kp at 0 kW

-2 Measured Kp at 25 kW

Measured Kp at 50 kW

/µg) -3 3 Measured Kp at 75 kW (m p Expected Kp -4 (Adsorption+Absorption) Log K Expected Kp (Adsorption) -5 Expected Kp (Absorption)

-6 -6 -5 -4 -3 -2 0 Log pL , torr

3 0 Figure 4-4 Log Kp (m /µg) vs. log pL (torr, @ 25 ºC) based on experimental measurements and estimations for n-alkanes in diesel emissions at 0-75 kW.

-1

Measured Kp at 0 kW -2 Measured Kp at 25 kW

Measured Kp at 50 kW

/µg) -3 3 Measured Kp at 75 kW (m p -4 Expected Kp

Log K (Adsorption+Absorption) Expected Kp (Adsorption) -5 Expected Kp (Absorption)

-6 -6 -5 -4 -3 -2 0 Log pL , torr

3 0 Figure 4-5 Log Kp (m /µg) vs. log pL (torr, @ 25 ºC) based on experimental measurements and estimations for PAHs in diesel emissions at 0-75 kW.

85 0 Table 4-2 Values of mr and br in the equation of log Kp = mr log pL + br at 25 ºC for alkanes and PAHs based on the experimental data and model estimations. n-Alkanes PAHs 2 2 mr br R mr br R Experimental data 0 kW -0.83 -7.46 0.99 -0.64 -6.14 0.92 25 kW -0.82 -7.69 0.97 -0.62 -6.58 0.95 50 kW -0.68 -7.81 0.97 -0.61 -6.81 0.94 75 kW -0.66 -8.03 0.95 -0.63 -7.12 0.94 Model estimations Adsorption -1.00 -7.53 NAa -1.00 -7.50 NA Absorption -1.00 -7.59 NA -1.00 -7.62 to -8.19 b NA Adsorption + Absorption -1.00 -7.26 NA -1.00 -7.27 to -7.45 b NA a b Not available since mr and br values were obtained from Eq. (4-3), Eq. (4-4), and Eq. (4-5). The value of br differed for different compounds due to the varied values of activity coefficient (γ).

0 These figures and tables have suggested that the correlation of log Kp vs. log pL was statistically significant, which was also reflected by the high R2 values (R2 > 0.92). The slopes of

0 log Kp vs. log pL plots range from -0.66 to -0.83 for n-alkanes, and from -0.61 to -0.64 for PAHs, which are close to the expected value of -1. The corresponding intercepts br range from -7.46 to -

8.03 and from -6.14 to -7.12 for n-alkanes and PAHs, respectively. The slopes and intercepts determined in this study are in the range reported by other studies (Table 4-3) although some differences are observed. The dissimilarities between the results in this study and in other studies may be caused by different aerosol type and sampling method. In this study, diesel aerosols were collected from a direct source, while in most other studies, ambient aerosols were collected.

Sampling method (high-vol or not), filter type (glass fiber, Teflon, or quartz fiber filter), and sampling condition (temperature and humidity) may also result in the dissimilarities between studies.

86 0 Table 4-3 Literature values of mr and br in the equation of log Kp = mr log pL + br for gas/particle partitioning at urban areas. a Compound type Temperature (ºC) Location mr br Reference Alkanes NA Portland -0.70 to -1.15 -6.05 to -8.51 Hart (1989) cited in [17] Alkanes 5 Denver -0.86 -7.29 Foreman and Bidleman [34]b Alkanes 37 Los Angeles -1.03 -8.68 Liang et al. [29]c Alkanes 25-32 Athens -0.28 to -0.57 -4.27 to -5.87 Mandalakis et al. [31]d PAHs 5 Portland -0.69 to -1.06 -5.74 to -7.35 Hart (1989) cited in [17] PAHs 7 Portland -1.09 -7.42 Ligocki and Pankow [35]e PAHs 37 Denver -0.76 -6.71 Foreman and Bidleman [34]b PAHs 6-8 Baltimore Harbor Tunnel -0.57 to -0.76 -6.05 to -6.61 Benner et al. [36]f PAHs 6-10 Osaka, Japan -0.86 to -1.17 -6.92 to -9.05 Yamasaki et al. [13]g PAHs -10 to 3.7 Chicago -0.34 to -1.04 -3.48 to -8.29 Cotham and Bidleman [37]h PAHs NA Chicago -0.64 -3.47 Simcik et al. [38]i PAHs NA Lake Michigan -0.67 -3.71 Simcik et al. [38]i PAHs NA Columbia -1.18 -9.06 Foreman and Bidleman [39]j PAHs 25-32 Athens -0.47 to -0.70 -4.52 to -5.97 Mandalakis et al. [31]d PAHs in diesel exhaust aerosols at 5 ºC -0.67 to -1.06 -4.96 to -10.4 Volckens and Leith [26]k a In most of these studies, samples were ambient aerosols. b Sampling method: high-vol sampling (0.35-0.5 m3/min) with glass fiber filter (GFF)/PUF/PUF. c Aerosols were generated in a Teflon smog chamber under controlled conditions. Samples were collected on GFF/PUF/PUF at a flow rate of ~70 L/min. d Sampling method: high-volume air sampler, GFF/PUF; sampling duration: 12 or 24 h; sampling volume: 900 or 450 m3. e Sampling method: high-volume sampling with GFF/GFF/PUF/PUF or Teflon membrane filter (TMF)/TMF/PUF/PUF; flow rate: 100-190 L/min; face velocity: 30-60 cm/s. f Samples were collected in exhaust rooms of the Baltimore Harbor Tunnel. Sampling method: high-volume sampling with TMF/PUF or GFF/PUF; flow rate: ~0.8 m3/min. g Sampling method: high-volume air sampler with GFF/PUF/PUF; flow rate: 0.75-0.8 m3/min; linear velocity on the GFF: 30.6-32.6 cm/s. h Sampling method: high-volume sampling with GFF/GFF/PUF/PUF; flow rate: 0.5 m3/min; sampling volume: 260-390 m3. i Sampling method: high volume air sampler equipped with GFF/PUF. j Sampling method: high volume air sampler (1.0 m3/min) with GFF/PUF/PUF/PUF. k Four types of sampler were used to collect diesel exhaust aerosols at a flow rate of 4.2 L/min for each sampler: (1) XAD denuder/Fiberfilm PTFE coated GFF/XAD denuder, (2) Fiberfilm PTFE coated GFF/Fiberfilm PTFE coated GFF/XAD denuder, (3) electrostatic sampler/XAD denuder, and (4) Zeflour PTFE filter/XAD denuder/XAD denuder.

87 Kp values can also be estimated from Eq. (4-3), (4-4), and (4-5). In the estimation of Kp values, the values for some parameters in the equations were selected based on some literature.

Pankow [14] presented that for simple physical sorption, Ns is not very compound-dependent,

-10 2 -10 2 and is of the order of 4 × 10 mol/cm . Thus 4.0 × 10 mol/cm was selected for Ns in this study. Regarding the ATSP for urban aerosol, Corn et al. [40] have measured by the BET method and obtained the values in the range of 0.019-0.031 cm2/µg for particulate matter collected in

Pittsburgh, Pennsylvania. Calculations based on particle-size distribution data reported by

2 Whitby et al. [41] for Pasadena, California give an ATSP of about 0.05 cm /µg. Since the data of

Corn et al. [40] was obtained by direct surface area measurements, a value of 0.025 cm2/µg was selected for ATSP by Pankow [14] and was used in this study as well. The value of (Q1-Q2) has been reported to be in the range of +2 to +4 kcal/mol for PAHs by Yamasaki et al. [42] and was selected as +3 kcal/mol in this study. The values of fom and MWom depend on the type of particulate matter and fom = 0.5 and MWom = 300 g/mol were assumed in this study based on the literature data [26]. When studying the partitioning of compounds with different polarities and functional groups, activity coefficient (γ) should be considered since γ differs from compound to compound [21]. Table 4-4 listed activity coefficients for n-alkanes, some PAHs, and alkanoic acids summarized from the literature.

Table 4-4 Activity coefficients for n-alkanes, PAHs and alkanoic acids from literature. Compounds γ Reference Compounds γ Reference n-Alkanes 1.2 Jang et al. [21] Pyrene 4.3 Jang et al. [21] Fluorene 1.3a Chandramouli et al. [43] Benzo(a)anthracene 4.8 Jang et al. [21] Phenanthrene 2.8 Jang et al. [21] Chrysene 4.8 Assumedb Anthracene 2.8 Jang et al. [21] Alkanoic acids 1.1 Jang et al. [21] Fluoranthene 4.3 Jang et al. [21] a The activity coefficient of fluorene was assumed the same as that of fluorene-d10 considering their similar physical-chemical properties. b The activity coefficient of chrysene was not available and was assumed the same as benzo(a)anthracene in this study.

88 0 The estimated log Kp values were also plotted vs. log pL in Figure 4-4 and Figure 4-5 and the calculated mr and br values were listed in Table 4-2. The estimated Kp values for alkanes are similar for adsorption, absorption, and combination mechanisms, while for PAHs the difference among the three mechanisms is evident. As described above, the activity coefficient is important in gas/particle partitioning of the compound, and increasing the activity coefficient can cause the partitioning to favor absorption. As Table 4-4 shows, the activity coefficient is near 1.2 for n- alkanes while it increases from 1.3 for 3-ring PAHs to 4.8 for 4-ring PAHs. As a result of the increased activity coefficient for heavy PAHs, adsorption may become the dominant mechanism for low-vapor pressure compounds, which is also indicated by Figure 4-5: the measured Kp values approach the Kp values estimated from the absorption equation when γ is low.

Table 4-2 shows that the measured mr and br values deviate from the theoretical values, which might be caused by the sampling bias and the variability of compound properties [17, 44].

The effects from various sources on the measured values of mr and br were discussed by Pankow and Bidleman [17] and were summarized as follows. (1) The variability in the compound-to- compound differences in (Q1-Q2) could affect both mr and br values. (2) The event-to-event variability in ATSP and T will not affect mr while it can cause variability in br of about ±0.5 and

±0.15, respectively. (3) When some fractions of sorbing compounds are bound and cannot exchange with the gas phase, the measured log Kp values will be higher, the mr value is shallower, and br value is greater than the expected values at full equilibrium. (4) The particle-to- gas desorption of compounds will result in a steep mr and a low br. (5) The gas-to-particle adsorption of compounds will result in a shallow mr and a high br. (6) Within a given sampling event, if ambient concentrations (e.g. A, F, and TSP) increase and/or ambient temperature (T) decreases, the measured value of mr will become more shallow and br will be increased;

89 inversely, if ambient concentrations decrease and/or T increases, the measured value of mr will steepen and br will be lowered. The six types of effects have been illustrated in Figures 4-6(a-f).

Figure 4-6 Sources contributing to the variability in mr and br values.

(a) Variability in the value of mq (the slope of a linear regression of Q1 against Q2 for a group of compounds sorbing on a specific type of particulate matter); (b)

variability in ATSP and T; (c) non-exchangeable effects; (d) desorption kinetics; (e) adsorption kinetics; (f) changing concentrations or changing T within a given event. (Pankow and Bidelman [17])

Figure 4-5 and Table 4-2 have presented shallow mr and high br values for PAHs under the measurement conditions compared with predicted values. Based on the study by Pankow and

Bidleman [17], gas-to-particle adsorption of PAHs and non-exchangeable effects may occur during the sampling events. However, regarding the partitioning of n-alkanes, the shallow mr and low br values match none of above-mentioned six situations. Another study by Volckens and

Leith [44] has indicated that when particle phase mass is measured erroneously as gas phase, a shallow mr and low br would be observed for the measurement, which matches the partitioning

90 results for n-alkanes in this study. This result suggests that the filter breakthrough for n-alkanes may occur during the sampling events due to the high load of n-alkanes and/or long sampling time. It is possible that gas-to-particle adsorption for PAHs and filter breakthrough for n-alkanes occur simultaneously, since the concentration of n-alkanes in diesel exhaust is much higher than that of PAHs.

From Figure 4-4 and Figure 4-5, it also has been observed that for both PAHs and n- alkanes, the regression lines are almost parallel for four load conditions and move downward as the load increases. According to the discussion by Pankow and Bidleman, this result means that

2 the temperature T increases and the surface area of the TSP (ATSP, cm /µg) decreases as the load increases, which is true since the combustion temperature is higher and the particle size is larger at higher loads.

4.4 Conclusions

The gas and particulate phase diesel exhausts from a non-road diesel generator were collected with the developed dilution method under different load conditions, the gas/particle partitioning of compounds was investigated, and the measured partitioning coefficients were compared with the estimated values from adsorption, absorption, and combination mechanisms.

The results have indicated that the amount of compounds in gas phase is much higher than the

-4.36 amount in particle phase. Light n-alkanes with the vapor pressures higher than 10 torr (C10-

C19) occur primarily in gas phase emissions, and heavy n-alkanes with vapor pressures lower

-4.84 0 -3.01 than 10 torr are mainly in particulate phase. 2- and 3-ring PAHs ( pL > 10 torr) are

0 -4.07 primarily present in gas phase emission, while 4- to 6-ring PAHs ( pL < 10 torr) occur primarily or exclusively in particulate phase.

91 The G/P partitioning coefficient Kp of a compound could be related to its vapor pressure

0 0 pL by the equation: log Kp = mr log pL + br. The linear regression results have indicated that the

0 correlation of log Kp vs. log pL is statistically significant. For n-alkanes, the slopes (mr) of log Kp

0 vs. log pL plots range from -0.66 to -0.83 and the intercepts (br) range from -7.46 to -8.03. The shallow mr and low br values compared with estimated values suggest that the filter breakthrough for n-alkanes may occur during the sampling events. For PAHs, shallow mr (-0.61 to -0.64) and high br (-6.14 to -7.12) values have been observed, which indicates gas-to-particle adsorption of

PAHs and non-exchangeable effects may occur during the sampling events. Engine loads affect

G/P partitioning of compounds; the regression lines are almost parallel for four load conditions and move downward as the load increases. This result may relate with the increased temperature and the decreased surface area of the TSP with loads.

4.5 References

1. Fernandes, M.B.; and Brooks, P., Characterization of carbonaceous combustion residues: II. Nonpolar organic compounds, Chemosphere, 53(5): 447-458, 2003.

2. Sauvain, J.J.; Duc, T.V.; and Guillemin, M., Exposure to carcinogenic polycyclic aromatic compounds and health risk assessment for diesel-exhaust exposed workers, Int. Arch. Occup. Environ. Health, 76(6): 443-455, 2003.

3. Sjogren, M.; Li, H.; Rannug, U.; et al., A Multivariate statistical analysis of chemical composition and physical characteristics of 10 diesel fuels, Fuel, 74(7): 983-989, 1995.

4. Adonis, M.; Martinez, V.; Riquelme, R.; et al., Susceptibility and exposure biomarkers in people exposed to PAHs from diesel exhaust, Toxicol. Lett., 144(1): 3-15, 2003.

5. Gratz, L.D.; Bagley, S.T.; Leddy, D.G.; et al., Interlaboratory comparison of HPLC- fluorescence detection and GC/MS: analysis of PAH compounds present in diesel exhaust, J. Hazard. Mater., 74(1-2): 37-46, 2000.

6. Lee, S.L.; De Wind, M.; Desai, P.H.; et al., Aromatics reduction and cetane improvement of diesel fuels, Fuel Reformulation, 5: 26-31, 1993.

92 7. Pal, R.; Juhasz, M.; and Stumpf, A., Detailed analysis of hydrocarbon groups in diesel range petroleum fractions with on-line coupled supercritical fluid chromatography gas chromatography mass spectrometry, J. Chromatogr. A, 819(1-2): 249-257, 1998.

8. Reilly, P.T.A.; Gieray, R.A.; Whitten, W.B.; et al., Real-time characterization of the organic composition and size of individual diesel engine smoke particles, Environ. Sci. Technol., 32(18): 2672-2679, 1998.

9. Shi, J.P.; Mark, D.; and Harrison, R.M., Characterization of particles from a current technology heavy-duty diesel engine, Environ. Sci. Technol., 34(5): 748-755, 2000.

10. Kuusimaki, L.; Peltonen, Y.; Mutanen, P.; et al., Urinary hydroxy-metabolites of naphthalene, phenanthrene and pyrene as markers of exposure to diesel exhaust, Int. Arch. Occup. Environ. Health, 77(1): 23-30, 2004.

11. IPCS, Environmental Health Criteria 202. Selected nonheterocyclic Polycyclic aromatic hydrocarbons. 1998, pp. 1-883, WHO: Geneva, ISBM. 9241572027.

12. Junge, C.E., Basic considerations about trace constituents in the atmosphere as related to the fate of global pollutants, in Fate of pollutants in the Air and Water Environments. Part I., Suffet, I.H., Ed. Wiley: New York. pp. 7-26, 1977.

13. Yamasaki, H.; Kuwata, K.; and Miyamoto, H., Effects of ambient temperature on aspects of airborne polycyclic aromatic hydrocarbons, Environ. Sci. Technol., 16(4): 189-194, 1982.

14. Pankow, J.F., Review and comparative analysis of the theories on partitioning between the gas and aerosol particulate phases in the atmosphere, Atmos. Environ., 21(11): 2275- 2283, 1987.

15. Seinfeld, J.H.; and Pandis, S.N., Organic atmospheric aerosols, in Atmospheric chemistry and physics: from air pollution to climate change. Wiley: New York. pp. 700-765, 1998.

16. Pankow, J.F.; and Bidleman, T.F., Effects of temperature, TSP and percent non- exchangeable material in determining the gas-particle partitioning of organic compounds, Atmos. Environ., 25A(10): 2241-2249, 1991.

17. Pankow, J.F.; and Bidleman, T.F., Interdependence of the slopes and intercepts from log- log correlations of measured gas-particle partitioning and vapor pressure. 1. Theory and analysis of available data, Atmos. Environ., 26A(6): 1071-1080, 1992.

18. Pankow, J.F.; Storey, J.M.E.; and Yamasaki, H., Effects of relative humidity on gas/particle partitioning of semi-volatile organic compounds to urban particulate matter, Environ. Sci. Technol., 27(10): 2220-2226, 1993.

19. Pankow, J.F., An absorption model of gas/particle partitioning of organic compounds in the atmosphere, Atmos. Environ., 28(2): 185-188, 1994.

93 20. Pankow, J.F., An absorption model of the gas/aerosol partitioning involved in the formation of secondary organic aerosol, Atmos. Environ., 28(2): 189-193, 1994.

21. Jang, M.; Kamens, R.M.; Leach, K.B.; et al., A thermodynamic approach using group contribution methods to model the partitioning of semi-volatile organic compounds on atmospheric particulate matter, Environ. Sci. Technol., 31(10): 2805-2811, 1997.

22. Mader, B.T.; and Pankow, J.F., Study of the effects of particle-phase carbon on the gas/particle partitioning of semivolatile organic compounds in the atmosphere using controlled field experiments, Environ. Sci. Technol., 36(23): 5218-5228, 2002.

23. Storey, J.M.E.; Luo, W.; Isabelle, L.M.; et al., Gas/solid partitioning of semivolatile organic compounds to model atmospheric solid surfaces as a function of relative humidity. 1. Clean quartz, Environ. Sci. Technol., 29(9): 2420-2428, 1995.

24. Mader, B.T.; and Pankow, J.F., Gas/solid partitioning of semivolatile organic compounds (SOCs) to air filters. 3. An analysis of gas adsorption artifacts in measurements of atmospheric SOCs and organic carbon (OC) when using Teflon membrane filters and quartz fiber filters, Environ. Sci. Technol., 35(17): 3422-3432, 2001.

25. Baek, S.O.; Goldstone, M.E.; Kirk, P.W.W.; et al., Phase distribution and particle-size dependency of polycyclic aromatic hydrocarbons in the urban atmosphere, Chemosphere, 22(5-6): 503-520, 1991.

26. Volckens, J.; and Leith, D., Comparison of methods for measuring gas-particle partitioning of semivolatile compounds, Atmos. Environ., 37(23): 3177-3188, 2003.

27. Pankow, J.F.; Isabelle, L.M.; Buchholz, D.A.; et al., Gas/particle partitioning of polycyclic aromatic hydrocarbons and alkanes to environmental tobacco smoke, Environ. Sci. Technol., 28(2): 363-365, 1994.

28. Jang, M.; and Kamens, R.M., A predictive model for adsorptive gas partitioning of SOCs on fine atmospheric inorganic dust particles, Environ. Sci. Technol., 33(11): 1825-1831, 1999.

29. Liang, C.K.; Pankow, J.F.; Odum, J.R.; et al., Gas/particle partitioning of semivolatile organic compounds to model inorganic, organic, and ambient smog aerosols, Environ. Sci. Technol., 31(11): 3086-3092, 1997.

30. Tao, Y.; and McMurry, P.H., Vapor pressures and surface free energies of C14-C18 monocarboxylic acids and C5-dicarboxylic and C6-dicarboxylic acids, Environ. Sci. Technol., 23(12): 1519-1523, 1989.

31. Mandalakis, M.; Tsapakis, M.; Tsoga, A.; et al., Gas-particle concentrations and distribution of aliphatic hydrocarbons, PAHs, PCBs and PCDD/Fs in the atmosphere of Athens (Greece), Atmos. Environ., 36(25): 4023-4035, 2002.

94 32. Goss, K.U.; and Schwarzenbach, R.P., Gas/solid and gas/liquid partitioning of organic compounds: Critical evaluation of the interpretation of equilibrium constants, Environ. Sci. Technol., 32(14): 2025-2032, 1998.

33. Naumova, Y.Y.; Offenberg, J.H.; Eisenreich, S.J.; et al., Gas/particle distribution of polycyclic aromatic hydrocarbons in coupled outdoor/indoor atmospheres, Atmos. Environ., 37(5): 703-719, 2003.

34. Foreman, W.T.; and Bidleman, T.F., Semivolatile organic compounds in the ambient air of Denver, Colorado, Atmos. Environ., 24A(9): 2405-2416, 1990.

35. Ligocki, M.P.; and Pankow, J.F., Measurements of the gas/particle distributions of atmospheric organic compounds, Environ. Sci. Technol., 23(1): 75-83, 1989.

36. Benner, B.A.; Gordon, G.E.; and Wise, S.A., Mobile sources of atmospheric polycyclic aromatic hydrocarbons: A roadway tunnel study, Environ. Sci. Technol., 23(10): 1269- 1278, 1989.

37. Cotham, W.E.; and Bidleman, T.F., Polycyclic aromatic hydrocarbons and polychlorinated biphenyls in air at an urban and a rural site near Lake Michigan, Environ. Sci. Technol., 29(11): 2782-2789, 1995.

38. Simcik, M.F.; Franz, T.P.; Zhang, H.X.; et al., Gas-particle partitioning of PCBs and PAHs in the Chicago urban and adjacent coastal atmosphere: States of equilibrium, Environ. Sci. Technol., 32(2): 251-257, 1998.

39. Foreman, W.T.; and Bidleman, T.F., An experimental system for investigating vapor particle partitioning of trace organic pollutants, Environ. Sci. Technol., 21(9): 869-875, 1987.

40. Corn, M.; Montgomery, T.L.; and Esmen, N.A., Suspended particulate matter: seasonal variation in specific surface areas and densities, Environ. Sci. Technol., 5(2): 155-158, 1971.

41. Whitby, K.T.; Husar, R.B.; and Liu, B.Y.H., The aerosol size distribution of Los Angeles smog, J. Colloid Interface Sci., 39(1): 177-204, 1972.

42. Yamasaki, H.; Kuwata, K.; and Kuge, Y., Determination of vapor pressure of polycyclic aromatic hydrocarbons in the supercooled liquid phase and their adsorption on airborne particulate matter, Nippon Kagaku Kaishi, (8): 1324-1329, 1984.

43. Chandramouli, B.; Jang, M.; and Kamens, R.M., Gas-particle partitioning of semivolatile organics on organic aerosols using a predictive activity coefficient model: analysis of the effects of parameter choices on model performance, Atmos. Environ., 37(6): 853-864, 2003.

44. Volckens, J.; and Leith, D., Effects of sampling bias on gas-particle partitioning of semivolatile compounds, Atmos. Environ., 37(24): 3385-3393, 2003.

95 Chapter 5

FACTORS AFFECTING HYDROCARBON DISTRIBUTION IN

DIESEL EMISSIONS

5.1 Introduction

Emissions from diesel engines represent an important source of particles in ambient air.

Non-road diesel engines contribute to approximately 44% of total diesel particulate matter (DPM) emissions and 12% of NOx emissions from mobile sources nationwide [1]. DPM contains numerous organic components, among which polycyclic aromatic hydrocarbons (PAHs) and their oxygenated and nitrated derivatives are highly hazardous [2-7]. Diesel particulate matter has been regarded as a potential occupational carcinogen by several government agencies, such as the U.S. Environmental Protection Agency (EPA) [8] and the National Institute for

Occupational Safety and Health (NIOSH) [9].

It is generally accepted that diesel particulate emissions are influenced by the engine testing conditions: types of operation (steady-state or transient) [10-14], fuel quality (high or low sulfur) [12, 15, 16], sampling conditions [17], and after-treatment devices [18]. Therefore, it is important to understand the impact of different fuels, engines, and emission controls on the chemical composition of particles for developing effective emission control strategies.

Effect of Engine Load: Engine load is an important factor that affects the combustion process, diesel exhaust temperature, and the composition of diesel particulate matter. It has been reported that engine load has effects on DPM mass concentration, size distribution, organic carbon and elemental carbon (OC/EC) distribution, morphology, and microstructure. El-

96 Shoboksh [10] studied the exhaust particle size distribution and particulate emissions for a single cylinder test diesel engine running at 0.25, 0.50, 0.75, and full load and at a constant speed of

600 rpm. The particulate emission factor (g particles/kg fuel) has been shown to increase linearly with engine load. For particle size distribution, submicron (0-0.4 µm) and coarse (9-10 µm) particles were found to increase, while the intermediate size showed insignificant variation with engine load. Liu et al. [19] reported that the relative contributions of OC and EC vary significantly with engine load. The fractions of EC over DPM increase with increasing load from

21% at 0 kW to 84% at 75 kW.

Effect of Fuel Sulfur Content: Sulfur in diesel fuels is known to contribute to DPM emissions [16, 20-23]. It is reported that the amount of adsorbed sulfates affects the mass of hydrocarbons in the particles [21], including PAHs, due to the formation of heavy hydrocarbons in the condensed phase from the reaction of adsorbed sulfuric acid with organic compounds in the exhaust [23]. Neeft et al. reported that 7-12% particulate reduction can be obtained with a reduction of fuel sulfur content from 0.2% to 0.05% [22]. Baranescu et al. [20] evaluated the effect of fuel sulfur on the brake specific particulate emissions on medium and heavy-duty trucks using three diesel fuels with 0.05%, 0.19%, and 0.29% sulfur. The results indicated that for an increase of 0.1% in fuel sulfur, particulates increased by about 0.025 g/bhp-h. The effect of fuel sulfur content on OC and EC distribution has been reported by Liu et al. [19].

Due to the above mentioned effect of fuel sulfur content on DPM emissions, the U.S.

EPA has proposed more stringent regulations for sulfur content in diesel fuels. For highway diesel fuel, fuel sulfur content was reduced from 500 ppmv to 15 ppmv in June 1, 2006 [24]. For non-road applications, the current EPA regulation is 3400 ppmv, but it will be reduced to 500 ppmv in 2007 and further to 15 ppmv in 2010 [24].

97 Effects of Sampling Conditions: Two major types of sampling methods have been used for collecting PM emissions from combustion systems: source-level sampling and ambient-level sampling. In general, source-level sampling is currently the accepted approach for the total suspended particulate and PM10 measurements for stationary sources, while ambient level sampling using dilution is the accepted approach for mobile source particulate emission measurements. The particulate mass, size distribution, and chemical composition differ for the two methods due to the aerosol processes such as coagulation, condensation, and nucleation occurring during dilution. Many studies have shown that the fate of the condensable organics and inorganics presented in diesel exhaust is significantly affected by atmospheric aging and dilution of the exhaust stream [25, 26]. It has been reported that the dilution ratio has a significant effect on the size distribution and the total number of particles emitted. Increasing the dilution ratio increases the concentration of ultrafine particles [27, 28]. The results from England et al. have indicated that traditional source testing methods (EPA 201A and EPA 202 in the paper) may significantly overestimate particulate emissions, especially the ultrafine condensable particle fraction [29].

Most of the above-mentioned studies focused on the impacts of each factor on PM mass concentration and size distribution. However, the effects of engine load conditions, fuel sulfur contents, and sampling methods on DPM chemical composition, especially individual organic compound, have not been well studied. The purpose of this study is to investigate the concentration and distribution of many types of organic compounds variated with engine load conditions, fuel sulfur contents, and sampling conditions.

98 5.2 Experimental Methods

The experiments were performed under four engine load conditions (0 kW, 25 kW, 50 kW and 75 kW) to investigate load effect; two different sulfur-containing diesel fuels, which were low sulfur diesel fuel (LSDF 400 ppmw S) and high sulfur diesel fuel (HSDF, 2200 ppmw

S), were used for studying fuel sulfur impact; and three sampling methods (EPA method 5, natural dilution, and developed dilution method with a dilution ratio of 3.4) were applied for the investigation of the influence of sampling conditions. The detailed experimental setup and procedures have been described in Chapter 2.

5.3 Results and Discussion

5.3.1 Effect of Engine Load

5.3.1.1 DPM versus Load

DPM was collected using dilution sampling method (dilution ratio (DR) = 3.4) under four engine load conditions (0 kW, 25 kW, 50 kW and 75 kW) when burning low sulfur diesel fuel

(400 ppmw S). The total DPM mass was determined by the gravimetric method, and the DPM mass concentration was simply obtained by dividing the total mass with the total sample volume that was corrected to standard conditions (1 atm, 293 °K). The DPM emission rate was calculated by multiplying DPM mass concentration by stack gas flow rate.

The DPM mass concentration and emission rate for four load conditions were listed in

Table 5-1. When the load is increased from 0 kW to 75 kW, the DPM mass concentration increases from 5.01 mg/m3 to 16.88 mg/m3, and the emission rate increases from 1.32 g/hr to

6.45 g/hr. The similar results have been reported by Burtscher et al., which indicated that emission factors for diesel engine increased with increasing load [30]. This has to be expected

99 because higher load for diesel engines means lower air/fuel ratio and less complete combustion.

It is expected that higher DPM mass concentration should be obtained by dilution method than

EPA Method 5 since dilution method allows some organics to condense on particles. However, compared with a study by Saiyasitpanich et al. [16], which used EPA Method 5 as the sampling method, the DPM mass concentration measured in this study is close to their results. This study also applied EPA Method 5 to measure DPM mass concentration and obtained lower concentration than Saiyasitpanich’s results. The lower DPM mass concentration in this study may be attributed to DPM mass from probe wash. In this study, due to the consideration of correlating OC, EC and compositional results to DPM mass, DPM from probe wash was not included for both dilution method and EPA Method 5. If this fraction is included, the DPM mass concentration should be higher than the reported results. The DPM mass concentration as a function of the applied engine load can be plotted as a straight-line (Figure 5-1), which is consistent with Saiyasitpanich’s study.

Table 5-1 DPM emission rate from the non-road diesel generator at 0-75 kW (LSDF, dilution sampling). Engine Fuel Stack gas Air/fuel ratio DPM DPM load consumption flow rate (mass/mass concentration emission (kW) rate (L/hr) (m3/hr) basis) (mg/m3) rate (g/hr) 0 5.0 263 85 5.01 ± 0.58 1.32 ± 0.15 25 10.0 286 46 7.35 ± 1.13 2.10 ± 0.32 50 15.0 341 37 10.79 ± 1.57 3.68 ± 0.53 75 19.5 382 32 16.88 ± 1.59 6.45 ± 0.61

100 20 y = 0.1562x + 4.1526 ) 18 3 2 16 R = 0.9546 14 12 10 8 6 4

Mass concentration(mg/m 2 0 0255075 Load (kW)

Figure 5-1 Correlation between DPM mass concentration and diesel engine load (LSDF, dilution sampling).

The OC and EC on quartz filters were measured using NIOSH method 5040 [31] and the emission rates for four load conditions were summarized in Table 5-2. The results indicate that both OC and EC emission rates increase with load. When engine load is increased from 0 kW to

75 kW, OC emission rate increases by twice, while EC emission rate increases by a factor of about 13. The regression results for OC and EC emission rates also indicate that EC emission rate increases more significantly with load than OC emission rate (Figure 5-2). Figure 5-2 suggests that the OC emission rate increases linearly with engine load, while EC emission rate increases exponentially. The greater increase in EC emission rate is the result of the higher fuel usage, lower air/fuel ratio, and higher temperature at higher loads, as EC is a product of incomplete combustion.

101 Table 5-2 OC and EC emission rate (in mg/m3 and mg/hr) from the non-road diesel generator at 0-75 kW (LSDF, dilution sampling). Engine load OC emission rate EC emission rate OC emission rate EC emission rate (kW) (mg/m3) (mg/m3) (g/hr) (g/hr) 0 3.64 ± 0.49 0.63 ± 0.09 0.96 ± 0.13 0.17 ± 0.02 25 4.43 ± 0.65 1.55 ± 0.26 1.27 ± 0.18 0.44 ± 0.07 50 5.38 ± 0.78 3.77 ± 0.54 1.84 ± 0.27 1.29 ± 0.18 75 6.42 ± 0.91 8.42 ± 0.88 2.45 ± 0.35 3.22 ± 0.34

10 9 OC

) 8 3 EC 7 y = 0.0373x + 3.5694 6 R2 = 0.9963 5 4 3 0.0347x Concentraiton (mg/m 2 y = 0.6436e R2 = 0.9993 1 0 0255075 Load (kW)

Figure 5-2 The concentration of organic and elemental carbon versus diesel engine loads (LSDF, dilution sampling).

Figure 5-3 represents the fractions of EC, organic compounds, and non-carbonaceous materials in DPM. The amount of organic compounds was calculated from the amount of OC by multiplying a factor of 1.2 (according to the study by Pierson and Brachaczek [32]) to account for the hydrogen, oxygen and other non-carbon elements that are associated with OC mass but were not measured by the thermal/optical reflectance technique directly. The carbonaceous fraction (CF) of total DPM was estimated by

CF = (EC + 1.2 × OC)/DPM (5-1)

102 where the EC, OC, and DPM refer to the mass of EC, OC, and DPM collected, respectively.

100%

80% Non-carbonaceous materials EC 60%

Organic compounds 40% (1.2*OC) Percentage in DPM 20%

0% 0 255075 Load (kW)

Figure 5-3 Percentage of organic compounds, elemental carbon, and non-carbonaceous materials in DPM at 0-75 kW (LSDF, dilution sampling).

As Figure 5-3 shows, the emissions were almost entirely carbonaceous materials, with CF estimated to be 1.00, 0.93, 0.95, and 0.96 for 0 kW, 25 kW, 50 kW, and 75 kW, respectively.

The remaining non-carbonaceous materials may include sulfates, nitrates, metals, and other trace elements and ions, although further analysis is required to confirm this hypothesis. A closer examination suggests that the relative fractions of EC and OC change dramatically with change in engine load. OC is the major species for low load conditions and its fraction decreases with load, while EC is the major species for high load conditions and its fraction increases with load.

Therefore, it is necessary to account for varying DPM composition while estimating DPM health impacts and emission source profile. The EC and OC emissions as a percent of total DPM during this study are consistent with those reported by others as shown in Table 5-3.

103 Table 5-3 Comparison of OC and EC percentage in total DPM mass from this study with the results from other studies. Source Cycle Fuel EC OC Shah, [33] Cold start/idle, HHDDTa ULSD (< 15 ppm S)b 17.3 ± 12.3 72.7 ± 14.8 CARBc creep 36.8 ± 13.6 60.1 ± 14.1 CARBc transient 67.8 ± 7.2 28.8 ± 9.1 CARBc cruise 61.0 ± 7.3 33.0 ± 15.0 Kelly, [34] 1500 rpm and idle Diesel (220 ppm S) 13.4 75.5 1500 rpm and 27% load 2.58 73.6 1700 rpm and 23%, 9% loadd 17.9 52.8 1700 rpm and 23% load 29.7 56.2 1700 rpm, 93%, 23% load 35.3 51.6 Rogers, [35] Running at fixed rpm for each Diesel (380 ppm S) 20 ± 8 60 ± 30 measurement Shi, [2] 1600 rpm and 25% load Disel (427 ppm S) 24.58 ± 3.47 57.66 ± 8.28 1600 rpm and 50% load 35.88 ± 3.20 44.67 ± 8.40 1600 rpm and 100% load 33.87 ± 3.85 24.53 ± 11.87 2600 rpm and 25% load 35.16 ± 1.48 49.89 ± 8.85 2600 rpm and 50% load 47.30 ± 2.62 28.87 ± 2.80 2600 rpm and 100% load 51.52 ± 1.58 24.46 ± 10.16 Schauer, [36] MDDT, Hot start FTPe CA RFDf 30.8 ± 3.6 19.7 ± 1.6 Lowenthal, [37] Central Business District Diesel No. 2 43.3 ± 20.1 35.4 ± 17.8 This study 1800 rpm and idle Diesel No. 2 12.60 ± 1.80 72.55 ± 9.78 1800 rpm and 25% load (400 ppm S) 21.05 ± 3.54 60.20 ± 8.84 1800 rpm and 50% load 34.92 ± 5.00 49.86 ± 7.23 1800 rpm and 75% load 49.91 ± 5.21 38.05 ± 5.39 a HHDDT: Heavy Heavy-Duty Diesel Truck (gross vehicle weight > 33,000 lb). b ULSD (< 15 ppm S): ultra-low sulfur diesel with sulfur content less than 15 ppm. c CARB: California Air Resources Board. d Decelerations during the middle of the text. e MDDT: Medium Duty Diesel Trucks; FTP: Federal Test Procedure. f CA RFD: California reformulated diesel.

Material balances that describe the chemical composition of the organic mass detected by

GC/MS are shown in Table 5-4. The total DPM mass can be divided as soluble organic fraction

(SOF) and unsolvable fraction by extraction using DCM as solvent. The unsolvable fraction is mainly elemental carbon (EC), often resulting as a byproduct of incomplete combustion, and non-carbonaceous materials such as metals. The soluble organic fraction can be subdivided into identified and unidentified organics by GC/MS, the latter of which mainly consists of the

104 unresolved complex mixture and those compounds that could not be separated as discrete peaks.

From Table 5-4, the SOF is reduced from 84.3% to 52.5% as load is increased from 0 kW to 75 kW, which is consistent with the variation in organic fractions from OC/EC results. The decrease in identifiable fraction was observed as load increasing. Small fractions of identifiable organics in DPM (1.63-8.39%) were also observed by other studies [36, 38, 39].

Table 5-4 Percentage of insoluble, soluble, unidentified and identified fraction in total DPM mass at 0-75 kW (LSDF, dilution sampling). 0 kW 25 kW 50 kW 75 kW Insoluble fraction 15.7 22.4 34.0 47.5 Soluble organic fraction 84.3 77.6 66.0 52.5

Unidentified fraction 75.9 72.2 64.1 50.8 Identified fraction 8.39 5.41 1.95 1.63 n-Alkanes 2.36 1.75 0.60 0.42 Branched alkanes 2.68 1.78 0.56 0.46 Saturated cycloalkanes 0.11 0.09 0.07 0.04 PAHs 0.07 0.04 0.03 0.02 Alkylated PAHs 2.21 1.17 0.20 0.15 Alkylbenzenes 0.02 0.03 0.23 0.27 n-Alkanoic acids 0.90 0.51 0.21 0.21 Aromatic acids 0.03 0.04 0.05 0.07

The identified compounds were classified into eight categories (Table 5-4), and the relative percentage for each category in the identifiable mass for four load conditions was shown in Figure 5-4. The normal alkanes, branched alkanes, and alkylated PAHs are the major species in the identified compounds, and the remaining small percentage is attributed to saturated cycloalkanes, PAHs, alkylbenzenes, and organic acids. As Table 5-4 shows, the percentage in total DPM mass for each category (except alkylbenzenes and aromatic acids) decreases with

105 increasing load, which results from the greater increase in EC proportion. However, the variation of relative percentage with load is different for each category. The relative percentage for alkanes (56.60-66.54%) and n-alkanoic acids (9.55-13.55%) does not significantly vary with load. In contrast, a great dependence on engine load was observed for aromatics, including PAHs, alkylated PAHs, alkylbenzenes, and aromatic acid. As engine load is increased from 0 kW to 75 kW, the percent increases from 0.29% to 16.97% and from 0.37% to 4.29% for alkylbenzenes and aromatic acids, respectively, while it decreases from 26.36% to 9.53% for alkylated PAHs.

The load dependence or independence for different species may be related to the sources of these compounds. According to the study by Dobbins et al. [40], the load independent compounds are most probably from petrogenic sources (raw materials such as diesel fuel and engine oil) and load dependent compounds are from pyrogenic sources (from combustion process).

100% 90% Aromatic acids 80% n-Alkanoic acids 70% Alkylbenzenes 60% Alkylated PAHs 50% PAHs 40% Cycloalkanes 30% iso-Alkanes 20% 10% n-Alkanes Percentage in total identified compounds 0% 0255075 Load (kW)

Figure 5-4 Percentage of each identified category in total identified particle phase compounds at 0-75 kW (LSDF, dilution sampling).

106 5.3.1.2 Organic Composition versus Load

Alkanes were identified and classified as normal alkanes, branched alkanes, and saturated cycloalkanes. Alkanes account for a large percentage of the identified compounds and are mainly sourced from diesel fuels. Normal alkanes with the carbon number ranging from C10 through C25 are the major components of diesel fuel, and the selected branched alkanes are known as the isoprenoids, which are naturally present in crude oil and have the potential to be used as tracers for diesel engine exhaust.

Normal Alkanes: Even though the percentage of n-alkanes is almost stable under different load conditions, the absolute concentrations of n-alkanes (ng/m3 flue gas) in diesel emissions vary significantly with load as shown in Figures 5-5(a) and (b).

For the particulate phase (Figure 5-5(a)), a bell-shaped distribution for n-alkanes from

C10 to C25 was observed with C17-C20 as the most abundant n-alkanes for all load conditions. This bell-shaped distribution is related with the similar distribution of n-alkanes in diesel fuel [41], since diesel fuel is the major source for n-alkanes in DPM. From Figure 5-5(a), it was also observed that the concentrations of n-alkanes at 0 kW and 25 kW are much higher than at 50 kW and 75 kW, even though the fuel consumption rate linearly increases with load. The flue gas temperature at different load (113 ºC, 178 ºC, 259 ºC, and 310 ºC for 0 kW, 25 kW, 50 kW, and

75 kW, respectively) may be the answer. Flue gas temperature affects the gas/particle phase partitioning of compounds, which is dominated by the adsorption or absorption [42-44]. At higher loads, the higher flue gas temperature resulted in a higher sampling temperature due to the limitation of dilution ratio and residence time in the dilution tunnel, and therefore large percentage of unburned compounds cannot condense on particles and remain in the gas phase.

Similar results were reported by Zielinska et al. [45].

107 18000 16000 0kW

) 25kW

3 14000 50kW 12000 75lW 10000 8000 6000

Concentration (ng/m 4000 2000 0

C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25 (a) Particle phase

350000 0 kW 300000 25 kW ) 3 250000 50 kW 75 kW 200000

150000

100000 Concentration (ng/m 50000

0 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25

(b) Gas phase + particle phase

Figure 5-5 The concentration of n-alkanes from n-decane (C10) to n-pentacosane (C25) in gas and particulate phase diesel emissions at 0-75 kW (LSDF, dilution sampling).

Figure 5-5(b) shows that the total n-alkane emissions (gas phase + particle phase) are much higher than particle phase emissions, which suggests that most emissions are from the gas phase. Compared with the particle phase, the most abundant n-alkanes in the total emissions have shifted to lighter n-alkanes (the n-alkanes with smaller carbon number). This is to be expected, since during the dilution process, the flue gas is cooled down and the lighter compounds are

108 condensed onto particle phase from gas phase. However, the change in flue gas temperature does not significantly affect the total emissions for heavier compounds since they exist mainly in particle phase.

From Figure 5-5(b), two types of trends can be observed regarding the concentrations of n-alkanes versus load. For n-alkanes with a carbon number from 10 to 14, an increase followed by a decrease is indicated, which represents a bell-shaped distribution. For n-alkanes with a carbon number higher than 14, the concentration increases with load. The different effects of load on n-alkanes with a different carbon number are probably due to the coexistent influence from varied temperature and fuel consumption rate. As the load is increased from 0 kW to 75 kW, the combustion temperature and fuel consumption rate increase. Higher combustion temperature can result in an increase of the conversion fraction of n-alkanes [46, 47] according to the kinetics reaction mechanism of combustion, which was built based on thermal decomposition, reaction with O2, propagation, radical decomposition, and isomerization [46]. Therefore, as the load increases, the unburned fuel will be reduced due to the higher conversion fraction and the increased volatilization at higher temperatures, which leads to the reduction of the compound concentration in DPM. On the other hand, the compound input was increased since more fuel was introduced into the engine at high load conditions. This will cause an increase in the compound concentration in DPM. The combination of temperature and fuel input effects led to the two types of trends for the concentrations of n-alkanes versus load. When fuel input is the dominant impact factor, the concentration increases with load. The bell-shaped distribution may be caused by the shift of the dominant impact factor: from 0 kW to 25 kW, the increase in fuel input is dominant and results in the increase in compound concentration, while from 25 kW to 75 kW, the temperature predominates over fuel input and a concentration decrease is the result.

109 Branched Alkanes: Highly branched isoprenoid alkanes normally occur in crude oil [48].

It is known that branched alkanes are more resistant to biological degradation than straight-chain alkanes, which makes them valuable for fingerprinting oil sources. The distributions of branched isoprenoid alkanes from C14 to C20 in diesel particulate matter and total diesel emissions (particle phase and gas phase emissions) are presented in Figures 5-6(a) and (b), respectively.

12000

) 0 kW 3 10000 25 kW 8000 50 kW 6000 75 kW

4000

Concentration (ng/m 2000

0 C14 C15 C18 C19 C20

(a) Particle phase

600000 0 kW

) 500000 3 25 kW

400000 50 kW 75 kW 300000

200000

Concentration (ng/m 100000

0 C14 C15 C18 C19 C20

(b) Total (gas phase + particle phase)

Figure 5-6 The concentration of branched alkanes in gas and particulate phase diesel emissions at 0-75 kW (LSDF, dilution sampling).

110 In diesel particulate matter, the selected isoprenoids C18-C20 are the major species which have much higher concentrations than C14 and C15 (Figure 5-6(a)). The result indicates that the most abundant species for branched alkanes has shifted to the alkane with a larger carbon number (C20) compared with normal alkanes (C18). The possible reason is that branched alkanes are more active due to their side chains compared with n-alkanes having the same carbon number.

This can also be observed from the lower boiling points and shorter GC/MS retention times of branched alkanes compared with n-alkanes having the same carbon number.

Figure 5-6(b) represents the total branched alkane emissions (gas phase and particle phase emissions) for engine loads from 0 kW to 75 kW. The concentration of branched alkanes in total emissions significantly increases compared with particle phase emissions, especially for

C14 and C15 (increasing by 150-240 times). The most abundant branched alkane in total emissions has shifted to lighter alkanes compared with the particle phase, which is similar to the result observed for n-alkanes.

As the engine load is increased, the concentrations of branched alkanes C18-C20 increase for both particle phase emissions (Figure 5-6(a)) and total emissions (Figure 5-6(b)). For branched alkanes C14 and C15, the load-resulting variation in their concentrations shows less dependence on the engine load. The difference between the engine load effect on branched alkanes C18-C20 and the effect on C14 and C15 is possibly due to the relatively higher stabilities of

C18-C20 compared with the stabilities of C14 and C15. The load-resulting variation in the concentration for less stable branched alkanes C14 and C15 depends on many factors, such as fuel input, combustion consumption, and combustion generation, while for branched alkanes C18-C20, the variation dependents primarily on fuel input.

111 1200 0kW 1000

) 25kW 3 50kW 800 75kW 600

400 Concentration (ng/m 200

0 C13C14C15C16C17C18C19C20C21C22C23C24C25

(a) Particle phase

30000 0 kW 25000 )

3 25 kW

20000 50 kW 75 kW 15000

10000

Concentration (ng/m 5000

0 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25

(b) Total (gas phase + particle phase)

Figure 5-7 The concentration of cycloalkanes from C13 to C25 in gas and particulate phase diesel emissions at 0-75 kW (LSDF, dilution sampling).

Saturated Cycloalkanes: Alkanes are the major components of the diesel fuel. In general, alkanes in the diesel fuel are mostly straight-chain and branched alkanes, but mono- cycloalkanes are also present as an important fraction [49]. Figures 5-7(a) and (b) represent the distributions of saturated cycloalkanes from C13 to C25 in particle phase emissions and total emissions. Figures 5-7(a) and (b) show that the distribution of saturated cycloalkanes is similar

112 to that of normal alkanes, even though the concentration of saturated cycloalkanes is much lower than the concentration of n-alkanes. The most abundant cycloalkanes in total emissions are C13-

C15. This distribution is similar to the distribution of n-alkanes in total emissions, and is related with the distribution of cycloalkanes in the diesel fuel.

PAHs and Alkylated PAHs: Many studies have been performed on the formation and ambient behavior of PAHs and their substituted homologues due to their carcinogenic and mutagenic properties [11, 40, 45, 50-54]. The research efforts have indicated that fuel aromaticity, PAH accumulation in lubricant oil, lubricant oil combustion, and cold start behavior influence PAH emissions [2, 11, 17, 30, 39, 45, 51, 53, 55]. Engine load is an important factor which can influence combustion temperature, air/fuel ratio, and consequentially influences PAH formation and behavior [11, 45]. Most of these studies are focused on on-road vehicle emissions.

However, limited studies have been conducted for emissions from non-road diesel generators, which have different operating modes and emission characteristics from on-road vehicles.

In this study, more than 30 PAHs and alkylated PAHs have been identified and quantified in particle phase and gas phase emissions from a non-road diesel generator under different load conditions. The distributions of EPA 16 priority PAHs in particle phase emissions and total emissions and their variation with load are illustrated in Figures 5-8(a) and (b).

113 1000 900 0kW

) 800 25kW 3 700 50kW 600 75kW 500 400 300

Concentration (ng/m 200 100 0 Nap Acy Ace Flu Phe Ant Flt Pyr Baa Chy Bbf Bkf Bap Ind Dba Bgp

(a) Particle phase

30000 0 kW 300 ) 25000 3

) 25 kW 3 20000 50 kW 200 75 kW 15000 100

10000 Concentration (ng/m 0 Concentration (ng/m 5000 Baa Chy Bbf Bkf Bap Ind Dba Bgp

0 Nap Acy Ace Flu Phe Ant Flt Pyr Baa Chy Bbf Bkf Bap Ind Dba Bgp

(b) Total (gas phase + particle phase)

Figure 5-8 The concentration of EPA 16 priority PAHs in gas and particulate phase diesel emissions at 0-75 kW (LSDF, dilution sampling).

Figure 5-8(a) shows that phenanthrene, fluoranthene, and pyrene are the major PAHs in diesel particulate matter. As the engine load is increased, the variation of PAH concentration is not closely related to load conditions, especially for 2- and 3-ring PAHs. Some other studies have reported that the relation between PAH concentration and the engine load differs for different PAH species [11, 17, 45, 53]. Jones et al. studied the effects of air/fuel ratio (up to 3.94)

114 on the PAH content in diesel particles and found that higher air/fuel ratio, which is the case for low load conditions, reduced high molecular weight PAHs [11]. However, since the air/fuel ratio of the combustion in diesel engines is much higher than the studied range by Jones et al. and the combustion condition is more complex, the results from this study are not consistent with Jones’ results. Waldenmaier et al. has shown that with load increasing the concentration decreased for benz[a]anthracene and chrysene while it increased for fluoranthene [17]. A study by Williams et al. [55] suggested that when the air/fuel ratio was in the range of 20-60, PAH emissions in DPM were high at low load, decreased at mid-load (having the air/fuel ration of about 35), and increased at higher load. The examined air/fuel ratio in this study is in the range of 32-85, and the result is consistent with the work of Williams et al.

As Figure 5-8(b) shows, the major PAHs in total emissions are some lighter PAHs (2- and 3-ring PAHs) instead of 3- and 4-ring PAHs. As the engine load is increased, the concentration increases for all 16 PAHs. However, the mechanisms leading to the increase may differ between 2- to 4-ring PAHs and 5- and 6-ring PAHs. Two main mechanisms were recognized in explaining the occurrence of PAHs in diesel emissions [40, 50, 53]. PAHs in diesel emissions can result directly from fuel PAH survivals through the combustion process, which is referred to as survival pathway or petrogenic origin. On the other hand, PAHs in the exhaust emissions can be formed during combustion processes by decomposition, dealkylation, and recombination, which are called pyrosynthetic pathway or pyrogenic origin. Both mechanisms are dependent on compounds as well as the combustion conditions, including engine load, temperature, and air/fuel ratio. Most studies have indicated that pyrosynthetic pathway represented the majority of PAH formations in the DPM emission [53, 56, 57]. However, Tancell et al. [58] found that benzo[a]pyrene in the fuel, which had survived combustion, was the major

115 source of benzo[a]pyrene in the exhaust. Williams et al. [55] concluded that for engine conditions used in their work, 2- to 4-ring PAHs in the exhaust particulate were primarily unburned fuel components.

According to the study by Williams et al. [55], 2- to 4-ring PAHs are expected to be from fuel PAH survivals. The increasing concentration with loads for 2- to 4-ring PAHs in this study may have resulted from the increased amount of unburned fuel at high loads. This conclusion is based on two factors: 2-to 4-ring PAHs are present in the diesel fuel, and at high loads, the low air/fuel ratio causes the amount of unburned fuel increased, resulting in the increase of unburned

PAHs. However, 5- and 6-ring PAHs are believed to be pyrogenic species, which are of greater carcinogenic potency [59] and contribute directly to soot particle formation [60], since they are not present in diesel fuels. As the load is increased, the pyrosynthetic process is promoted and consequently the concentration of pyrogenic PAHs increases. This conclusion has to be examined by further studies.

Figures 5-9(a) and (b) present distributions of alkylated PAHs in particle phase emissions and total emissions, and variations with loads in their concentrations. Emissions of alkylated

PAHs are relatively high compared with emissions of unsubstituted PAHs, which is consistent with other studies [39]. For particle phase emissions, alkylated PAHs are reduced by 76% as the load is increased from 0 kW to 75 kW. Similar results have been reported by Jensen and Hites

[61]. When gas phase emissions are included, major alkylated PAHs are shifted from heavier species (trimethylnaphthalenes, methyl- and dimethylphenanthrenes) to some lighter species

(dimethylnaphthalenes).

116 30000 300000 ) ) 3 3 0kW 25000 250000 0 kW 25kW 25 kW 20000 200000 50kW 50 kW 15000 150000 75kW 75 kW 10000 100000 5000 50000 Concentration (ng/m Concentration (ng/m 0 0 MN MN TMN MPh TMN MPh DMN DMN DMPh DMPh (a) Particle phase (b) Total (gas phase + particle phase)

Figure 5-9 The concentration of alkylated PAHs in gas and particulate phase diesel emission at 0-75 kW (LSDF, dilution sampling).

Alkylbenzenes: Alkylbenzenes are common components in diesel fuels, as well as in diesel emissions. As the engine load is increased, the fuel input is increased, which results in the increase of alkylbenzene emissions (Figures 5-10(a) and (b)). A second source for alkylbenzene emissions may be the generation from the decomposition, dealkylation, and recombination from other compounds.

12000 350000 ) ) 3 3 0kW 0 kW 10000 300000 25 kW 25kW 250000 8000 50 kW 50kW 200000 75 kW 6000 75kW 150000 4000 100000 2000 50000 Concentration (ng/m Concentration (ng/m 0 0

C1-B C2-B C3-B C4-B C5-B C6-B C1-B C2-B C3-B C4-B C5-B C6-B (a) Particle phase (b) Total (gas phase + particle phase)

Figure 5-10 The concentration of alkylbenzenes in gas and particulate phase diesel emission at 0-75 kW (LSDF, dilution sampling).

117 Alkanoic Acids: As known from Chapter 3, alkanoic acids in diesel emissions are from engine oil used. Their distribution and variation with load are different from the fuel sourced compounds (Figures 5-11(a) and (b)).

10000 9000 0kW ) 3 8000 25kW 7000 50kW 6000 5000 75kW 4000 3000 2000 Concentration (ng/m 1000 0 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18

(a) Particle phase

100000 90000 0 kW ) 3 80000 25 kW 70000 50 kW 60000 50000 75 kW 40000 30000 20000 Concentration (ng/m 10000 0 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18

(b) Total (gas phase + particle phase)

Figure 5-11 The concentration of alkanoic acids from C6H12O2 to C18H36O2 in diesel emissions at 0-75 kW (LSDF, dilution sampling).

For alkanoic acids from C6H12O2 to C10H20O2, the concentrations increase with load in both particle phase and total emissions, while decrease with load for some heavier acids

(C14H28O2 to C18H36O2). This is related to the combustion temperature, which can influence the

118 evaporation of engine oil components. At higher loads, the higher combustion temperature greatly increases the evaporation of lighter compounds which have lower boiling points, while it has less influence on the evaporation of heavy compounds. The reduction of heavy acid emissions at higher loads is caused by the increase in the flow rate of stack gas at higher loads.

Compound Emission Rate Normalized to Fuel Usage: The emission rate of identified compounds in total diesel emissions (gas phase + particle phase) at 0-75 kW was calculated based on the data of compound concentrations in emissions, the stack gas flow rate, the fuel usage rate, and compound concentrations in the diesel fuel according to Eq. (5-2) and (5-3). The results were expressed as mg compound emission/kg of fuel burned and as mg compound emission/g compound in diesel fuel using the method of Williams et al. [55].

CS × QS ER1 = (5-2) RDF × ρ DF

ER1 ER2 = (5-3) CDF

Where:

ER1 = mg compound emission/kg of fuel combusted, mg/kg DF

3 CS = compound concentration in diesel emissions (gas + particulate phase), µg/m

3 QS = stack gas flow rate, m /hr

RDF = fuel usage rate, L/hr

ρDF = diesel fuel density, 800 g/L

ER2 = mg compound emission/g compound in diesel fuel, mg/g

CDF = compound concentration in diesel fuel, g/kg DF

119 Figure 5-12 shows the emission rate for several categories of compounds, expressed in terms of mass of compound in the exhaust/mass of compound in the fuel; also shown are the total unburned hydrocarbons (UHC).

1.40 0.30 PAHs 1.20 0.25 Alkylated 1.00 PAHs 0.20 0.80 n-Alkanes 0.15 0.60 Branched alkanes 0.10 0.40 Cycloalkanes / kg fuel combusted

G total UHC emission UHC total G 0.05 /compound g from fuel 0.20 Alkylbenzenes Mgcompound in emission 0.00 0.00 0255075 0255075 Engine load (kW) Engine load (kW)

(a) Each category of compounds (b) Total unburned hydrocarbons (UHC)

Figure 5-12 Normalized emission rates for identified compounds expressed in terms of mass of compound in the exhaust/mass of compound from the fuel input (LSDF, dilution sampling).

As Figure 5-12 suggests, total unburned hydrocarbons are 0.015-0.024 wt% of the fuel input, and the UHC emission rate varies with engine loads, following the “U” type trend which shows high emissions at low load, decreasing at 50 kW, and increasing at 75 kW. Similar survivabilities and trends are observed for n-alkanes, branched alkanes, cycloalkanes, and alkylated PAHs. However, unsubstituted PAHs and alkylbenzenes behave differently. A higher proportion of 2- to 4-ring PAHs (0.02-1.0 wt%) survive the combustion process, and 5- and 6- ring PAHs are formed during the combustion process. Total PAH emissions are 0.12-0.13 wt% of PAHs from the diesel fuel. The fraction of PAHs in the emission to PAHs in the fuel increases

120 with loads. Compared with a study by Williams et al. [55], the normalized emission rate of

PAHs and UHC is relatively low. This may be caused by different sampling methods. In the study by Williams et al., the exhaust gases were cooled by means of a water cooled condenser and glass fibre filters were used. Figure 5-12 also shows that the emission rate for alkylbenzenes increases with loads. Higher normalized emission rates for PAHs and alkylbenzenes suggest the production of PAHs and alkylbenzenes during the combustion process, especially at high load conditions.

5.3.2 Effect of Sampling Method

5.3.2.1 Dilution Effect on DPM, OC and EC

In this study, three sampling methods described in Chapter 2 have been used, i.e. EPA

Method 5, natural dilution, and dilution method with a dilution ratio of about 3.4. However, the

DPM mass, OC, and EC concentration expressed in terms of mg/m3 are not available for natural dilution method, since the actual exhaust gas volume collected and the dilution ratio cannot be measured for natural dilution. Therefore, the DPM mass, OC, and EC concentration are compared between EPA Method 5 and dilution method (DR=3.4) only.

Figure 5-13 indicates that DPM mass concentration linearly increases with loads for both methods, which is consistent with other studies [16, 19]. Compared with Method 5, higher DPM mass concentration is obtained with the dilution method. This is a result of the continuous adsorption, nucleation, coagulation, and condensation of the large quantities of condensable organics and inorganics during dilution processes. England et al. [29] compared PM2.5 mass measured by the dilution method and the in-stack method (EPA 201A / 202) for the gas-fired boiler and the gas-fired process heater. They found that the amount of filterable particulate

121 collected using EPA method 201A was lower than the particulate collected with the dilution method, which is similar to the result in this study. However, large amounts of condensable particulate matter (CPM), which was believed to be mostly sulfates and chlorides [29], were present in the exhaust and could be measured by EPA method 202. If this fraction was included, the total filterable matter plus CPM mass was much higher than the total PM2.5 mass obtained with the dilution method. Since particle condensation mechanisms are dependent on both vapor concentration of compounds in the emission and the exhaust temperature, the dilution method simulates conditions that more closely represent true atmospheric condensation conditions compared with traditional source testing methods and is recommended as the standard reference method for the measurement of automotive and stationary emissions.

) 20 3 Dilution (DR=3.4) 16 EPA Method 5

DPM mass concentration (mg/m 0 0255075 Load (kW)

Figure 5-13 Comparison of DPM mass concentration between EPA Method 5 and the dilution method with a dilution ratio of ~3.4 (LSDF).

From Figure 5-13, it is also observed that the difference of DPM mass concentration between EPA Method 5 and the dilution method is highest at 75 kW. At higher loads, the exhaust gas temperature, as well as the CPM mass, is higher. Therefore, the DPM mass 122 concentration is significantly increased at high loads with the dilution method due to the condensation of the large quantities of CPM during dilution processes.

Figure 5-14 illustrates the OC and EC concentration in diesel emissions collected using

EPA Method 5 and the dilution method under different load conditions. Multiple measurements were made at each load condition and sufficient reproducibility is evidenced by the figure.

Figure 5-14 indicates that the EC concentrations are similar for EPA Method 5 and the dilution method under each load condition. This has to be expected as EC is the carbonaceous carbon that is produced by the incomplete combustion and is not affected by the sampling condition, i.e. the sampling temperature, which is the major difference between the two methods.

12 OC-EPA Method 5 )

3 10 OC-Dilution (DR=3.4) 8 EC-EPA Method 5 EC-Dilution (DR=3.4) 6

Concentration (mg/m 2

0 0255075 Load (kW)

Figure 5-14 OC and EC concentration in DPM collected with EPA Method 5 and the dilution method (LSDF).

The organic carbon collected with the dilution method is much higher than that collected with EPA Method 5, especially at high loads (Figure 5-14). The fraction of organic compounds in DPM, which is calculated from 1.2×OC/DPM, is from 87% (0 kW) to 46% (75 kW) for the dilution method, while it is from 62% (0 kW) to 9% (75 kW) for EPA Method 5 (Figures 5-15 (a)

123 and (b)). The higher organic fraction in DPM collected with dilution method is caused by the condensation of the condensable organic particulate matter and is consistent with the higher

DPM mass concentration. Liu et al. [19] reported that at the collection temperature of 25 °C ± 3

°C, the organic compounds account for 27-75% of the DPM mass, as opposed to 9-62% at the collection temperature of 120 °C ± 14 °C. The result from this study is similar to the study by

Liu et al. [19].

100% 100%

80% 80% Non-carbonaceous materials 60% 60% EC 40% 40% Organic compounds

Fraction in DPM in Fraction 20% Fraction in DPM 20% (1.2*OC) 0% 0% 0255075 0255075 Load (kW) Load (kW)

(a) EPA Method 5 (b) Dilution method (DR = 3.4)

Figure 5-15 Fractions of organic compounds, EC, and non-carbonaceous materials in DPM collected with EPA Method 5 and the dilution method (LSDF).

5.3.2.2 Dilution Effect on the Organic Composition of DPM

As mentioned above, the actual exhaust gas volume and the dilution ratio for natural dilution method are not measurable; therefore, the concentration of individual compounds in particulate diesel emissions is expressed in terms of µg/g DPM for the three sampling methods

(Table 5-5). The concentration of identified compounds in DPM is compared among three sampling methods at low (0 kW) and high (75 kW) load conditions.

124 Table 5-5 The concentration of compounds in DPM collected with EPA Method 5, the natural dilution method, and the dilution method (DR=3.4) (LSDF). Compoundsa Concentration (µg/g DPM) 0 kW 75 kW Method 5 Natural Dilution Method 5 Natural Dilution dilution method dilution method n-Alkanes n-C10 5.3 9.7 116 1.1 3.8 24 n-C11 16 22 164 1.5 5.6 47 n-C12 21 31 222 1.6 7.4 55 n-C13 26 57 337 3.2 12 79 n-C14 48 204 500 5.2 14 116 n-C15 79 565 1085 7.2 18 203 n-C16 159 1164 2206 11 30 308 n-C17 219 2645 3189 14 46 441 n-C18 375 1479 3851 20 67 516 n-C19 427 1257 3415 23 92 560 n-C20 384 742 2769 28 81 528 n-C21 276 138 1921 25 69 404 n-C22 162 10 1041 15 40 225 n-C23 107 7.4 574 9.9 33 146 n-C24 75 5.2 421 6.5 17 113 n-C25 64 3.6 324 5.7 14 128 Total n-alkanes 2444 8341 22137 176 552 3893

Branched alkanes

C14 7.6 13 154 3.0 6.5 61

C15 16 28 269 6.0 6.8 91

C18 104 135 1065 68 89 635

C19 123 147 1204 86 107 697

C20 152 185 1454 95 116 760 Other branched alkanes 923 7084 22894 326 288 2295 Total branched alkanes 1325 7593 27040 584 613 4538

Cycloalkanes

C13 2.8 11 21 0.57 0.90 8.7

C14 4.4 27 46 0.80 1.5 23

C15 9.1 29 101 1.6 2.6 50

C16 12 43 168 3.0 5.0 70

C17 21 39 209 4.0 7.5 86

C18 32 61 150 5.8 13 82

C19 24 12 138 7.9 15 63

125 Table 5-5 (Continued) Compounds Concentration (µg/g DPM) 0 kW 75 kW Method 5 Natural Dilution Method 5 Natural Dilution dilution method dilution method

C20 17 9.8 78 7.0 9.6 42

C21 13 6.8 54 4.0 5.5 28

C22 11 4.3 45 3.0 4.1 17

C23 8.5 0.00 31 2.6 2.8 14

C24 7.2 0.00 26 1.9 2.0 10

C25 5.5 0.00 13 1.6 1.9 7.1 Total saturated cycloalkanes 168 242 1079 44 71 500

PAHs Nap 1.5 4.1 35 0.60 3.2 7.4 Acy 1.2 2.0 17 0.20 1.6 4.3 Ace 1.3 3.0 33 0.20 3.7 11 Flu 4.4 7.1 32 0.40 5.3 11 Phe 23 77 224 4.9 30 70 Ant 1.8 4.1 17 0.40 3.3 5.8 Flt 16 25 78 5.0 20 42 Pyr 28 67 146 9.7 35 56 Baa 7.8 5.1 21 3.8 8.3 12 Chy 7.8 5.1 29 4.3 11 16 Bbf 3.0 3.0 5.9 1.4 3.0 7.5 Bkf 2.7 2.0 5.0 1.2 1.5 6.2 Bap 4.9 4.1 11.1 2.1 5.3 10 Ind 3.8 6.6 8.2 1.7 3.7 7.4 Dba 1.8 1.7 2.9 0.94 1.8 2.5 Bgp 2.8 3.0 6.4 1.8 2.9 6.1 Total PAHs 112 220 673 39 139 275

Alkylated PAHs MN 33 116 340 6.0 19 51 DMN 125 589 2076 25 99 394 TMN 757 2914 5625 34 123 422 MPh 1167 3938 7135 27 76 274 DMPh 1032 3420 6268 30 95 362 Total alkylated PAHs 3114 10977 21445 122 412 1503

Alkylbenzenes

C1-B 6.8 11 41 4.3 27 198

126 Table 5-5 (Continued) Compounds Concentration (µg/g DPM) 0 kW 75 kW Method 5 Natural Dilution Method 5 Natural Dilution dilution method dilution method

C2-B 13 21 72 8.9 50 358

C3-B 11 15 43 16 79 464

C4-B 12 16 47 35 148 826

C5-B 5.6 6.2 23 26 105 554

C6-B 3.0 3.2 11 13 55 278 Total alkylbenzenes 51 71 236 104 464 2678 n-Alkanoic acids

C6 0 0 10 1.5 23 45

C7 3.6 18 27 6.8 62 134

C8 8.1 40 57 10 94 168

C9 18 88 125 20 140 256

C10 45 289 398 26 198 318

C11 55 354 478 19 133 184

C12 89 591 786 28 183 217

C13 93 444 582 11 54 69

C14 425 1955 2503 22 125 170

C15 281 1238 1559 9.7 50 62

C16 474 1991 1899 10 62 84

C17 25 0.0 5.2 4.2 2.5 3.1

C18 11 4.1 10 1.7 5.2 5.3 Total of n-alkanoic acids 1528 7012 8441 170 1133 1716

Aromatic acids Benzoic acid 58 226 305 35 430 677

Total identified compounds 8799 34682 81354 1274 3814 15780 a The abbreviation in the table represents the corresponding compounds in Table 5-5. For n-alkanoic acids,

C6-C18 abbreviations represent the alkanoic acids from C6H12O2 to C18H36O2.

Table 5-5 indicates that the concentration of most compounds is highest for the dilution method while lowest for EPA Method 5 under both low and high load conditions. As the exhaust gas temperature is 113 ºC, 178 ºC, 259 ºC, and 310 ºC for 0 kW, 25 kW, 50 kW, and 75 kW, respectively, it is expected that the dilution process has the greatest effect on compound 127 condensation at 75 kW. Therefore, the compound concentration and distribution are compared for the three methods at 75 kW in the following discussion. At 75 kW, the concentration ratio of individual compounds collected with the dilution method over those collected with EPA Method

5 is in the range of 2.6-53.3. While the concentration ratio of individual compounds collected with the natural dilution method over those collected with EPA Method 5 is mostly less than

10.0 and is between 10.0 and 18.7 for several compounds. These results suggest that the developed dilution method with a dilution ratio of 3.4 is more effective in converting organics from the gas phase to the particle phase than the natural dilution method with a sampling distance of about 1.0 m. This means the dilution ratio for the natural dilution with a sampling distance of about 1.0 m is less than 3.4. The lower sampling temperature for the developed dilution method than the natural dilution also confirms this prediction. Brown et al. [27] studied the relationship between the dilution ratio and the sampling distance and found that the dilution ratio increased with the sampling distance. At the sampling distance of 1.0 m, the dilution ratio was about 10, which is different from the result in this study. This may be explained from the fact that the dilution ratio not only depends on the sampling distance, but also depends on diameters and directions of the exhaust outlet and the sampling inlet.

Kittelson et al. [62] has reported that as the engine exhaust is diluted and cooled, the soluble organic fraction (SOF) may move from the gas phase to the particle phase by two paths: the adsorption on existing particles or the nucleation to form new particles. The condensation can be viewed as an extension of the nucleation and adsorption. Both adsorption and nucleation processes are influenced by the saturation ratio for the various organic species, which is defined as the partial pressure of the gaseous organic species divided by the vapor pressure of the same organic species. The partial pressure of the volatile components in the hot exhaust decreases with

128 dilution. The vapor pressure of those components is a function of temperature, which also decreases with dilution. Since the vapor pressure-versus-temperature relationship is nonlinear, the saturation ratio reaches a maximum at some dilution level. Kittelson et al. [62] has reported that the saturation ratio has the highest value for dilution ratios of about 5 to 50. Thus, the strongest force driving gas to particle conversion occurs in the critical dilution ratio range (5 -

50:1). The dilution ratio for the developed dilution method is about 3.4 and falls in the range where the saturation ratio increases with the dilution ratio. Therefore, the concentration of compounds is the highest for the developed dilution method, lower for the natural dilution, and lowest for EPA Method 5.

Distribution of Alkanes: The dilution process affects not only the absolute concentration of compounds but also the compound distribution, since the dilution effect on the adsorption and condensation differs for light and heavy compounds. Therefore, the percentage of individual n- alkanes in the total emitted normal alkanes is calculated and the result is compared between the three sampling methods (Figure 5-16(a)). Figures 5-16(b) and (c) present the distribution of identified branched alkanes and cycloalkanes in DPM collected at 75 kW with the three methods.

As Figure 5-16(a) shows, the n-alkane distribution is similar for the three methods, which indicates that the percentage of individual n-alkanes is highest for mid-molecular weight n- alkanes, while lower for low- and high-molecular weights. However, compared with EPA

Method 5, the most abundant n-alkane has shifted to the left side of the figure (from C20 to C19), which represents lower molecular weight compounds. The percentage of n-alkanes with lower molecular weights (the carbon number less than 18) is highest for the dilution method, lowest for

EPA Method 5, and falls in between for the natural dilution method. The percentage of n-alkanes with higher molecular weights (the carbon number greater than 20) is in reverse order for the

129 three methods. These results indicate that the formation of lower molecular weight compounds in

DPM is more significantly increased by the dilution process than the formation of higher molecular weights.

) 18 16 EPA Method 5 14 Natural dilution 12 Dilution (DR=3.4) 10 8 6 4 2 0 Percentage ofindividual n-alkanes(%

C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25 (a) n-Alkanes

d 40 25 35 EPA 20 Method 5 30 anes (%) Natural 25 15 dilution EPA

20 cycloalk Dilution Method 5 10 15 Natural (DR=3.4) alkanes (%) 10 dilution Dilution 5 5 (DR=3.4)

0 Percentage of 0 Percentage of individual branche

C14 C15 C18 C19 C20 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25 (b) Identified branched alkanes (c) Cycloalkanes

130 The result from the comparison of cycloalkane distribution among the three methods is similar to the result for normal alkanes, while for the selected branched alkanes no significant difference is observed among the three sampling methods.

Distribution of PAHs, Alkylated PAHs, and Alkylbenzenes: The distributions of

PAHs, alkylated PAHs, and alkylbenzenes for EPA Method 5, the natural dilution and the developed dilution method are illustrated in Figures 5-17(a-c).

) 30 (a) 25

Percentage of individual PAHs (% 0 Flt Ind Flu Bbf Bkf Ant Pyr Nap Acy Ace Phe Baa Chy Bap Dba Bgp 35 40 (b) (c) 30 35 30 25 25 20 20 15 15 10 10 alkylbenzenes (%) alkylated PAHs (%) Percentage ofindividual Percentage ofindividual 5 5 0 0 MN DMN TMN MPh DMPh C1-B C2-B C3-B C4-B C5-B C6-B

EPA Method 5 Natural dilution Dilution (DR=3.4)

131 Although Table 5-5 has indicated that the concentration of all these aromatics is highest for the dilution method and lowest for EPA Method 5, the distribution of these aromatics is similar and the most abundant species are the same for the three sampling methods. A detailed examination of the relative abundance for each compound has suggested that among the compounds which have similar structures and chemical properties, larger fractions of lower molecular weights are collected with the dilution method while it is opposite for EPA Method 5.

For instance, among 16 EPA priority PAHs, 2- and 3-ring PAHs account for a higher percentage of total PAHs for the dilution method (39.9%) than for EPA Method 5 (17.5%), while 5- and 6- ring PAHs account for a lower percentage for the dilution method (14.4%) than for EPA Method

5 (23.3%).

Distribution of Alkanoic Acids: Figure 5-18 represents the distribution of alkanoic acids in DPM at 75 kW for method 5, the natural dilution method, and the dilution method with a dilution ratio of ~3.4. Similar to the above-mentioned result, which suggests that the abundance shifts to the lower molecular weights for the dilution method compared with EPA Method 5, the percentage of alkanoic acids with carbon numbers from 6 to 10 in total particulate alkanoic acids is 37.57% for EPA Method 5, increases to 45.64% for the natural dilution, and further increases to 53.67% for the developed dilution method. On the contrary, the proportion of alkanoic acids with higher molecular weights (C12H24O2 to C18H36O2) increases for dilution method compared with EPA Method 5.

132 c 20 EPA Method 5 16 Natural dilution Dilution (DR=3.4) 12

8 acids (%)

Percentage of individual alkanoi 0 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18

5.3.3 Effect of Fuel Sulfur Content

5.3.3.1 Effect of Fuel Sulfur Content on DPM, OC, and EC

As illustrated in Figure 5-19, the concentration of DPM collected at high sulfur diesel fuel (2200 ppmw S) is approximately 1.2 times the concentration of DPM collected at low sulfur diesel fuel (400 ppmw S). Higher DPM concentration at higher fuel sulfur content have been reported by other studies [16, 19]. Figure 5-19 also indicates that in the DPM at high sulfur diesel fuel, the organic compounds are about 1.3 times as high, and non-carbonaceous materials are 1.7-2.8 times as high, as those at low sulfur diesel fuel. The higher emission of organic compounds at higher fuel sulfur content is consistent with the result reported by Wall and

Hoekman [63], and is due to the increased nucleation of organic compounds by the sulfuric acid

[64]. The higher non-carbonaceous material emission may result from the increase of sulfate fractions at higher fuel sulfur content.

133 3.0

2.5 Non-carbonaceous materials 2.0 OC 1.5 DPM 1.0 (HSDF /LSDF)

Concentration ratio EC 0.5

0.0 0255075 Load (kW)

Figure 5-19 Ratios of the concentration of DPM, OC, EC, and non-carbonaceous materials at high sulfur diesel fuel over the concentration at low sulfur diesel fuel (dilution sampling).

Table 5-6 lists the fractions of organic compounds, EC, and non-carbonaceous materials in DPM at low and high sulfur diesel fuel. The contribution of organic compounds and non- carbonaceous materials to DPM is higher for the high sulfur diesel fuel than for the low sulfur diesel fuel due to the above-mentioned occurrence of larger amounts of sulfuric acid and sulfate.

Table 5-6 Fractions of organic compounds, EC, and non-carbonaceous materials in DPM at low and high sulfur diesel fuel (dilution sampling). Load Fraction of organic Fraction of EC Fraction of non- (kW) compounds (1.2×OC) carbonaceous materials LSDF HSDF LSDF HSDF LSDF HSDF 0 87.1% 90.5% 12.6% 9.0% 0.3% 0.5% 25 72.2% 76.2% 21.0% 15.1% 6.7% 8.8% 50 59.8% 63.1% 34.9% 27.8% 5.2% 9.1% 75 45.7% 49.4% 49.9% 40.1% 4.4% 10.5%

134 5.3.3.2 Effect of Fuel Sulfur Content on the Organic Composition

Compared with low sulfur diesel fuel, the concentration of organic compounds in DPM collected at high sulfur diesel fuel is 1.2-3.0 times as high, which may result from the increased nucleation. However, regarding the fraction of each category and the distribution of the identified compounds, no significant difference is observed between low and high sulfur diesel emissions. Consistent with this result, the distribution of these compounds in high sulfur diesel fuel is similar to their distribution in low sulfur diesel fuel. Similar enhanced nucleation for these organics by sulfuric acid is another hypothesis and should be further investigated.

5.4 Conclusions

The engine load significantly affects the concentration of DPM, OC, EC, and individual compounds, and also affects the relative distribution of these compounds. As the load is increased, the concentration of DPM, OC, and EC increases, as well as the concentration of individual compounds in total diesel emissions (gas phase plus particle phase). However, the fraction of OC in DPM decreases with loads. In the identified particle phase compounds, alkanes account for more than 50% and this fraction does not significantly vary with loads; PAH and alkylbenzene fractions increase with loads; while the fraction of alkylated PAHs decreases.

When normalized to the diesel fuel usage, the emission rate increases with loads for most pyrogenic species (PAHs and alkylbenzenes), while it decreases with loads for most petrogenic species (alkanes, alkylated PAHs).

The sampling method is an important factor that affects the concentration and distribution of organics in DPM. The developed dilution method with a dilution ratio of about 3.4 is more effective in converting organics from the gas phase to the particle phase than the natural dilution

135 with a sampling distance of about 1.0 m, and both methods are more effective than the source sampling method (EPA Method 5). With the developed dilution method, DPM concentration increases 40-80% compared with EPA Method 5, OC concentration is 2.7-11.6 times as high as that for EPA Method 5, and EC concentration is similar for the two methods. The dilution process not only affects the concentration of organic compounds, but also affects their distribution. Compared with the source sampling method, the most abundant species has shifted to those compounds with lower molecular weights for the dilution method, which suggests that the dilution process favors the formation of light compounds more than the formation of heavy compounds.

Higher fuel sulfur content results in higher emissions of DPM, OC, and non- carbonaceous materials. As the fuel sulfur is increased from 400 ppmw to 2200 ppmw, the emission of DPM, OC, and non-carbonaceous materials increases to 1.2, 1.3, and 1.7-2.8 times respectively, while the EC emission is not significantly affected. Although the emission of individual organic compounds in diesel exhaust at high sulfur diesel fuel increases to 1.2-3.0 times as that at low sulfur diesel fuel, their relative distribution does not vary with fuel sulfur contents since the distribution of these compounds in low and high sulfur diesel fuels is similar.

136 5.5 References

1. United States Envrionmental Protection Agency, Nonroad Engines, Equipment, and Vehicles; Reducing Air Pollution from Nonroad Engines, http://www.epa.gov/nonroad, Accessed 2006.

2. Shi, J.P.; Mark, D.; and Harrison, R.M., Characterization of particles from a current technology heavy-duty diesel engine, Environ. Sci. Technol., 34(5): 748-755, 2000.

3. Reilly, P.T.A.; Gieray, R.A.; Whitten, W.B.; et al., Real-time characterization of the organic composition and size of individual diesel engine smoke particles, Environ. Sci. Technol., 32(18): 2672-2679, 1998.

4. Pal, R.; Juhasz, M.; and Stumpf, A., Detailed analysis of hydrocarbon groups in diesel range petroleum fractions with on-line coupled supercritical fluid chromatography gas chromatography mass spectrometry, J. Chromatogr. A, 819(1-2): 249-257, 1998.

5. Lee, S.L.; De Wind, M.; Desai, P.H.; et al., Aromatics reduction and cetane improvement of diesel fuels, Fuel Reformulation, 5: 26-31, 1993.

6. Adonis, M.; Martinez, V.; Riquelme, R.; et al., Susceptibility and exposure biomarkers in people exposed to PAHs from diesel exhaust, Toxicol. Lett., 144(1): 3-15, 2003.

7. Gratz, L.D.; Bagley, S.T.; Leddy, D.G.; et al., Interlaboratory comparison of HPLC- fluorescence detection and GC/MS: analysis of PAH compounds present in diesel exhaust, J. Hazard. Mater., 74(1-2): 37-46, 2000.

8. EPA, Health Assessment Document for Diesel Emissions, Workshop Review EPA- 600/8-90/057A, EPA, Washington, DC, 1990.

9. US Department of Health and Human Services, Public Health Service, Centers for Disease Control, National Institute for Occupational Safety and Health, DHHS (NIOSH), Current Intelligence Bulletin No 50 - Carcinogenic Effects of Exposure to Diesel Exhaust, Publication No. 88-116, Cincinnati, OH, 1988.

10. El-Shobokshy, M.S., The effect of diesel-engine load on particulate carbon emission, Atmos. Environ., 18(11): 2305-2311, 1984.

11. Jones, C.C.; Chughtai, A.R.; Murugaverl, B.; et al., Effects of air/fuel combustion ratio on the polycyclic aromatic hydrocarbon content of carbonaceous soots from selected fuels, Carbon, 42(12-13): 2471-2484, 2004.

12. Liang, F.; Lu, M.; Birch, M.E.; et al. Effects of engine load and fuel sulfur on the organosulfur content of diesel particulate matter. in proceedings of the 98th A&WMA Annual Conference and Exhibition. June 21-24, 2005. Minneapolis, MN.

13. Lindgren, M.; and Hansson, P.A., Effects of transient conditions on exhaust emissions from two non-road diesel engines, Biosyst. Eng., 87(1): 57-66, 2004.

137 14. Zhu, J.Y.; Lee, K.O.; Yozgatligil, A.; et al., Effects of engine operating conditions on morphology, microstructure, and fractal geometry of light-duty diesel engine particulates, Proc. Combust. Inst., 30: 2781-2789, 2005.

15. Corro, G., Sulfur impact on diesel emission control - A review, React. Kinet. Catal. Lett., 75(1): 89-106, 2002.

16. Saiyasitpanich, P.; Lu, M.M.; Keener, T.C.; et al., The effect of diesel fuel sulfur content on particulate matter emissions for a nonroad diesel generator, J. Air Waste Manage. Assoc., 55(7): 993-998, 2005.

17. Waldenmaier, D.A.; Gratz, L.D.; Bagley, S.T.; et al., The influence of sampling conditions on the repeatability of diesel particulate and vapor phase hydrocarbon and PAH measurements, SAE Technical Paper Series 900642, 1990.

18. Burtscher, H., Physical characterization of particulate emissions from diesel engines: a review, J. Aerosol. Sci., 36(7): 896-932, 2005.

19. Liu, Z.F.; Lu, M.M.; Birch, M.E.; et al., Variations of the particulate carbon distribution from a nonroad diesel generator, Environ. Sci. Technol., 39(20): 7840-7844, 2005.

20. Baranescu, R.A., Influence of fuel sulfur on diesel particulate emissions, SAE Technical Paper Series 881174, 1988.

21. Duran, A.; Carmona, M.; and Ballesteros, R., Competitive diesel engine emissions of sulphur and nitrogen species, Chemosphere, 52(10): 1819-1823, 2003.

22. Neeft, J.P.A.; Makkee, M.; and Moulijn, J.A., Diesel particulate emission control, Fuel Process. Technol., 47(1): 1-69, 1996.

23. Yu, F.Q.; Turco, R.P.; and Karcher, B., The possible role of organics in the formation and evolution of ultrafine aircraft particles, J. Geophys. Res.-Atmos., 104(D4): 4079-4087, 1999.

24. US EPA, Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements, EPA 420-F-00-057, December 2000.

25. Ahlvik, P.; Ntziachristos, L.; Keskinen, J.; et al., Real time measurements of diesel particle size distribution with an electrical low pressure impactor, SAE Technical Paper Series 980410, 1998.

26. Kittelson, D.B.; and Abdul-Khalek, I.S. Formation of nanoparticles during exhaust dilution. in Proceedings of 2nd ETH Workshop on Nanoparticle Measurement. August 1998. Zurich, Switzerlan.

27. Brown, J.E.; Clayton, M.J.; Harris, D.B.; et al., Comparison of the particle size distribution of heavy-duty diesel exhaust using a dilution tailpipe sampler and an in-

138 plume sampler during on-road operation, J. Air Waste Manage. Assoc., 50(8): 1407-1416, 2000.

28. Lipsky, E.; Stanier, C.O.; Pandis, S.N.; et al., Effects of sampling conditions on the size distribution of fine particulate matter emitted from a pilot-scale pulverized-coal combustor, Energy Fuels, 16(2): 302-310, 2002.

29. England, G.C.; Zielinska, B.; Loos, K.; et al., Characterizing PM2.5 emission profiles for stationary sources: comparison of traditional and dilution sampling techniques, Fuel Process. Technol., 65: 177-188, 2000.

30. Burtscher, H.; Kunzel, S.; and Huglin, C., Characterization of particles in combustion engine exhaust, J. Aerosol. Sci., 29(4): 389-396, 1998.

31. Birch, M.E.; and Cary, R.A., Elemental carbon-based method for monitoring occupational exposures to particulate diesel exhaust, Aerosol Sci. Technol., 25(3): 221- 241, 1996.

32. Pierson, W.R.; and Brachaczek, W.W., Particulate matter associated with vehicles on the road. 2, Aerosol Sci. Technol., 2(1): 1-40, 1983.

33. Shah, S.D.; Cocker, D.R.; Miller, J.W.; et al., Emission rates of particulate matter and elemental and organic carbon from in-use diesel engines, Environ. Sci. Technol., 38(9): 2544-2550, 2004.

34. Kelly, K.E.; Wagner, D.A.; Lighty, J.S.; et al., Characterization of exhaust particles from military vehicles fuelled with diesel, gasoline, and JP-8, J. Air Waste Manage. Assoc., 53(3): 273-282, 2003.

35. Rogers, C.F.; Sagebiel, J.C.; Zielinska, B.; et al., Characterization of submicron exhaust particles from engines operating without load on diesel and JP-8 fuels, Aerosol Sci. Technol., 37(4): 355-368, 2003.

36. Schauer, J.J.; Kleeman, M.J.; Cass, G.R.; et al., Measurement of emissions from air pollution sources. 2. C-1 through C-30 organic compounds from medium duty diesel trucks, Environ. Sci. Technol., 33(10): 1578-1587, 1999.

37. Lowenthal, D.H.; Zielinska, B.; Chow, J.C.; et al., Characterization Of Heavy-Duty Diesel Vehicle Emissions, Atmos. Environ., 28(4): 731-743, 1994.

38. McDonald, J.D.; Zielinska, B.; Sagebiel, J.C.; et al., Characterization of fine particle material in ambient air and personal samples from an underground mine, Aerosol Sci. Technol., 36(11): 1033-1044, 2002.

39. Rogge, W.F.; Hildemann, L.M.; Mazurek, M.A.; et al., Sources of fine organic aerosol. 2. Noncatalyst and catalyst-equipped automobiles and heavy-duty diesel trucks, Environ. Sci. Technol., 27(4): 636-651, 1993.

139 40. Dobbins, R.A.; Fletcher, R.A.; Benner, B.A.; et al., Polycyclic aromatic hydrocarbons in flames, in diesel fuels, an in diesel emissions, Combustion And Flame, 144(4): 773-781, 2006.

41. Liang, F.Y.; Lu, M.M.; Keener, T.C.; et al., The organic composition of diesel particulate matter, diesel fuel and engine oil of a non-road diesel generator, J. Environ. Monit., 7(10): 983-988, 2005.

42. Jang, M.; Kamens, R.M.; Leach, K.B.; et al., A thermodynamic approach using group contribution methods to model the partitioning of semivolatile organic compounds on atmospheric particulate matter, Environ. Sci. Technol., 31(10): 2805-2811, 1997.

43. Pankow, J.F., An absorption model of gas/particle partitioning of organic compounds in the atmosphere, Atmos. Environ., 28(2): 185-188, 1994.

44. Pankow, J.F., An absorption model of the gas/aerosol partitioning involved in the formation of secondary organic aerosol, Atmos. Environ., 28(2): 189-193, 1994.

45. Zielinska, B.; Sagebiel, J.; Arnott, W.P.; et al., Phase and size distribution of polycyclic aromatic hydrocarbons in diesel and gasoline vehicle emissions, Environ. Sci. Technol., 38(9): 2557-2567, 2004.

46. Dagaut, P., On the kinetics of hydrocarbons oxidation from natural gas to kerosene and diesel fuel, Phys. Chem. Chem. Phys., 4(11): 2079-2094, 2002.

47. Ristori, A.; Dagaut, P.; and Cathonnet, M., The oxidation of n-hexadecane: Experimental and detailed kinetic modeling, Combust. Flame, 125(3): 1128-1137, 2001.

48. Atlas, R.M., Microbial degradation of petroleum hydrocarbons - An environmental perspective, Microbiol. Rev., 45(1): 180-209, 1981.

49. Granata, S.; Faravelli, T.; and Ranzi, E., A wide range kinetic modeling study of the pyrolysis and combustion of naphthenes, Combust. Flame, 132(3): 533-544, 2003.

50. Dobbins, R.A.; Fletcher, R.A.; B. A. Benner; et al. PAHs in flames, diesel fuels and diesel emissions. in 30th International Symposium on Combustion, Work-In-Progress Poster Session. July 25-30, 2004. Chicago, IL.

51. Collier, A.R.; Rhead, M.M.; Trier, C.J.; et al., Polycyclic aromatic compound profiles from a light-duty direct- injection diesel-engine, Fuel, 74(3): 362-367, 1995.

52. Courtois, Y.; Molinier, B.; Pasquereau, M.; et al., Influence of the operating-conditions of a diesel-engine on the mutagenicity of its emissions, Sci. Total Environ., 134(1-3): 61-70, 1993.

53. Rhead, M.M.; and Hardy, S.A., The sources of polycyclic aromatic compounds in diesel engine emissions, Fuel, 82(4): 385-393, 2003.

140 54. Zielinska, B.; Sagebiel, J.; McDonald, J.D.; et al., Emission rates and comparative chemical composition from selected in-use diesel and gasoline-fueled vehicles, J. Air Waste Manage. Assoc., 54(9): 1138-1150, 2004.

55. Williams, P.T.; Bartle, K.D.; and Andrews, G.E., The relation between polycyclic aromatic compounds in diesel fuels and exhaust particulates, Fuel, 65(8): 1150-1158, 1986.

56. Marr, L.C.; Kirchstetter, T.W.; Harley, R.A.; et al., Characterization of polycyclic aromatic hydrocarbons in motor vehicle fuels and exhaust emissions, Environ. Sci. Technol., 33(18): 3091-3099, 1999.

57. Rhead, M.M.; and Pemberton, R.D., Sources of naphthalene in diesel exhaust emissions, Energy Fuels, 10(3): 837-843, 1996.

58. Tancell, P.J.; Rhead, M.M.; Trier, C.J.; et al., The sources of benzo[a] pyrene in diesel exhaust emissions, Sci. Total Environ., 162(2-3): 179-186, 1995.

59. Larsen, J.C.; and Larsen, P.B., Chemical carcinogens, in Air Pollution and Health, Hester, R.E. and Harrison, R.M., Editors. The Royal Society of Chemistry. pp. 33-56, 1998.

60. Dobbins, R.A.; Fletcher, R.A.; and Chang, H.C., The evolution of soot precursor particles in a diffusion flame, Combust. Flame, 115(3): 285-298, 1998.

61. Jensen, T.E.; and Hites, R.A., Aromatic diesel emissions as a function of engine conditions, Anal. Chem., 55(4): 594-599, 1983.

62. Kittelson, D.B.; Watts, W.F.; and Arnold, M., Aerosol Dynamics, Laboratory and On- road Studies - Supplemental Report No. 2 to EPA Grant Review of Diesel Particulate Matter Sampling Methods. 1998.

63. Wall, J.C.; and Hoekman, S.K., Fuel Composition Effects on Heavy-Duty Diesel Particulate Emissions, SAE Technical Paper Series 841364, 1984.

64. Shi, J.P.; and Harrison, R.M., Investigation of ultrafine particle formation during diesel exhaust dilution, Environ. Sci. Technol., 33(21): 3730-3736, 1999.

141 Chapter 6

DETERMINATION OF POLYCYCLIC AROMATIC SULFUR

HETEROCYCLES IN DIESEL FUEL AND DIESEL

PARTICULATE MATTER BY GAS CHROMATOGRAPHY WITH

ATOMIC EMISSION DETECTION

6.1 Introduction

Organic sulfur compounds are the most abundant non-hydrocarbon constituents in petroleum. The sulfur content of petroleum fuels contributes to the formation of sulfur dioxide

(SO2), which causes both acid deposition and poisoning of the catalytic converters in vehicles.

An increase in fuel sulfur also results in increased diesel particulate emissions. Fuel sulfur reduction has been mandated by US Environmental Protection Agency (EPA) in the near future.

A large fraction of the organic sulfur in diesel fuels occurs in aromatic structures, especially as alkylated homologues of polycyclic aromatic sulfur heterocycles (PASHs). It was reported that benzothiophene (BT), dibenzothiophene (DBT) and their alkylated homologues are the most abundant organosulfur compounds in diesel fuels [1]. Recently, increasing interest has been focused on PASHs for several reasons. Specifically, some PASHs have been reported for their potential mutagenic and carcinogenic properties [2, 3]; some PASHs, especially alkylated

DBTs, are difficult to remove in the desulfurization process for production of low sulfur fuels [4,

5]; and some PASHs can be potential indicators of the origin and maturity of crude oil [6].

142 Given the above-mentioned roles of sulfur compounds, efforts have been placed on the chemical characterization of PASHs in crude oil, diesel fuel, and other petroleum products [1, 7,

8]. However, there have been few studies on the organosulfur content of diesel particulate matter

(DPM), which is regarded as a carcinogen by the EPA [9] and the National Institute for

Occupational Safety and Health (NIOSH) [10]. PASHs in DPM can originate directly from the diesel fuel or be generated by the combustion process. They are adsorbed on the DPM, which includes a high number of ultrafine particles, and therefore penetrate into the deep lung. Thus, the characterization of PASHs and other particle-borne organic compounds is important and necessary for evaluating and controlling any adverse health effects associated with DPM exposure. Studies on the PASH content of DPM are difficult because their concentrations are low, there is a great variety of compounds present [11], and quantitative determination of individual PASH isomers in the complex mixture is difficult.

The identification and quantification of individual PASH require selective and sensitive methods of detection. Gas chromatography with the atomic emission detection (GC/AED) is a powerful technique that offers high-resolution separation of components in a complex matrix and highly selective spectrometric detection. Its application to the analysis of complex matrices, including petroleum products, has been demonstrated [12-15]. The AED is an element-selective and universal detection that provides relatively constant elemental response factors for different compounds [12-14], which makes a compound independent calibration (CIC) possible. CIC is highly useful because it minimizes the number of analytical standards required and permits quantification of compounds for which no standards exist. CIC is particularly attractive when dealing with highly toxic chemicals because relatively nontoxic surrogates can be used for instrument calibration.

143 In this paper, PASHs in two different sulfur-containing diesel fuels (low sulfur diesel fuel

[LSDF] and high sulfur diesel fuel [HSDF]) and the resultant DPM were identified and quantified by GC with sulfur-selective atomic emission detection. The distribution of PASHs in

DPM was investigated under different fuel sulfur and engine load conditions. For the convenience of description, low-molecular weight or lighter PASHs were defined as the PASHs with one or two rings, and high-molecular weight or heavier PASHs were defined as three-, four- or five-ring PASHs. The precision of a CIC was evaluated with a calibration solution containing several PASHs having different structures and molecular weights.

6.2 Experimental method

6.2.1 Sampling and Extraction Method

Sampling and extraction methods were described in Chapter 2.

6.2.2 Analytical Method

A 6890 GC equipped with a G2350A AED (Agilent Technologies, Palo Alto, CA) was used for quantification of the sulfur components examined in this study. Table 6-1 lists the

GC/AED operating conditions. The HP-5 MS, column used in this study, and equivalent columns such as DB5-MS have been proven to be effective for PASH separation in several studies [2, 6,

13]. Both carbon (179 nm) and sulfur (181 nm) selective modes were monitored for all samples.

Sulfur determination was based on external and internal standards. For the external standard calibration, the average response factor of a standard solution containing three sulfur compounds at different concentrations was used. As an internal standard, a solution of t-butyl disulfide

144 (TBDS) in DCM was spiked into tested fuels and DPM extracts, which were determined to be free of TBDS. Tentative compound identities are based on retention time.

Table 6-1 GC/AED operating parameters. GC Conditions Injection port temperature 280 °C AED transfer line 310 °C GC capillary column 30 m x 0.25-mm id x 0.25-µm film HP-5 MS Oven temperature program 40 °C to 300 °C at 10 °C /min, hold 10 min Column flow 1.3 mL/min Carrier gas Helium Injection 20:1 split for fuels; splitless for DPM extracts Injection volume 1 µL AED Parameters Reagent gases Oxygen, 50 psi; hydrogen, 45 psi Makeup flow 66 mL/min Cavity temperature 300 °C Carbon emission line 179 nm Sulfur emission line 181 nm

6.2.3 Analytical Standards

Standard compounds listed in Table 6-2 were prepared in dichloromethane. Thiophene and alkylated thiophenes were from Fisher Scientific, and other PASHs were obtained from a laboratory in Germany (Dr. Jan T. Andersson, Institut für Anorganische und Analytische Chemie,

Correns-Str. 30, D-48149 Münster, Germany). Figure 6-1 shows the structures for the PASH standards. As a check of method performance, the following National Institute of Standards and

Technology (NIST) standard reference material (SRM) also was used: SRM 2724b, sulfur in diesel fuel oil (426.5 ppmw total sulfur).

145 Table 6-2 List of thiophene standards and related information. No. Compound Abbreviation Formula MWa Bp or Mpb

1 Thiophene T C4H4S 84 84.4 °C (Bp)

2 2-Methylthiophene 2-MT C5H6S 98 113 °C (Bp) c 3 3-Methylthiophene 3-MT C5H6S 98 114 °C (Bp738 )

4 2,3-Dimethylthiophene 2,3-DMT C6H8S 112 142-144 °C (Bp)

5 2,5-Dimethylthiophene 2,5-DMT C6H8S 112 134 °C (Bp740)

6 2-Ethylthiophene 2-ET C6H8S 112 132-134 °C (Bp)

7 2-Propylthiophene 2-PT C7H10S 126 157.5-159.5 °C (Bp)

8 Benzothiophene BT C8H6S 134 221-222 °C (Bp)

9 2-Methylbenzothiophene 2-MBT C9H8S 148 51-52 °C (Mp)

10 3-Methylbenzothiophene 3-MBT C9H8S 148 125-127 °C (Bp25)

11 5-Methylbenzothiophene 5-MBT C9H8S 148 66-67 °C (Bp0.6)

12 3,5-Dimethylbenzothiophene 3,5-DMBT C10H10S 162 125-126 °C (Bp14); 118- 118.5 °C (Bp9)

13 2,3,5-Trimethylbenzothiophene 2,3,5-TMBT C11H12S 162 145-146 °C (Bp15)

14 2,3,7-Trimethylbenzothiophene 2,3,7-TMBT C11H12S 162 143-144 °C (Bp15)

15 2,3,4,7- 2,3,4,7- C12H14S 176 167.5-168 °C (Bp14) Tetramethylbenzothiophene TTMBT

16 Dibenzothiophene DBT C12H8S 184 332-333 °C (Bp)

17 4-Methyldibenzothiophene 4-MDBT C13H10S 198 >334 °C (Bp)

18 4,6-Dimethyldibenzothiophene 4,6-DMDBT C14H12S 212 ~340 °C (Bp) d 19 2,4,6-Trimethyldibenzothiophene 2,4,6-TMDBT C15H14S 226 NA

20 Phenanthro[4,5-bcd]thiophene Ph45T C14H8S 208 139-140 °C (Mp)

21 Benzo[b]naphtho[1,2-d]thiophene BN12T C16H10S 234 103-104 °C (Mp)

22 Benzo[b]naphtho[2,1-d]thiophene BN21T C16H10S 234 232-234 °C (Mp)

23 Benzo[b]naptho[2,3-d]thiophene BN23T C16H10S 234 158-159 °C (Mp)

24 Phenanthro[3,4-b]thiophene Ph34T C16H10S 234 82.5-83.5 °C (Mp)

25 Benzo[b]phenanthro[9,10- BPh9,10T C20H12S 284 129-130 °C (Mp) d]thiophene

26 Diacenaphthothiophene DiAT C24H12S 332 285-286 °C (Mp) a Molecular weight. b Bp: boiling point; Mp: melting point. c Bp738 is the boiling point at the pressure of 738 mm Hg. d Not available.

146

Figure 6-1 Structures of PASH standard compounds.

6.3 Results and discussion

6.3.1 AED Response Factor

Atomic emission detection should permit a compound independent calibration with surrogate reference compounds that contain the element of interest. Results with errors of a few

147 percent have been reported when the compounds to be quantified had elemental compositions, retention times, and concentrations similar to the reference compounds [12-15].

In this study, three organosulfur compounds (Table 6-3) that are most abundant in diesel fuel were chosen as external calibration standards. The sulfur response factor was calculated as the average of the three compounds. The sulfur response factor was not influenced significantly by differences in molecular weight or retention time; the deviation from the average value is about 5%.

Table 6-3 Sulfur response factors for three sulfur compounds. Results are based on five injections. Compound MWa Formula Response factorb RSDc (%) (n-5)

DBT 184.255 C12H8S 54.06 1.8

4-MDBT 198.282 C13H10S 52.73 0.7

4,6-DMDBT 212.309 C14H12S 52.54 1.1 aMolecular weight. bArea count per nanogram sulfur. cRelative standard deviation.

The total sulfur in diesel fuel was estimated by integrating the sulfur response from 7 min

(the beginning of the sulfur emission) to 25 min for the fuel and 32 min for DPM. The integration endpoint was the point where the sulfur emission fell to baseline and remained stable.

With the average response factor obtained for the three sulfur compounds, the calculated sulfur content of NIST SRM 2724b is 406 ppmw, which is a 4.9% deviation from the certified value of

426.5 ppmw. Based on the AED results, the sulfur contents of the diesel fuels (LSDF and HSDF) used in this study are 433 ppmw and 2284 ppmw respectively. These results are in good agreement with results (400 ppmw S for LSDF and 2200 ppmw S for HSDF) reported by another laboratory (OKI Analytical, Cincinnati, Ohio, 45212, USA) using ASTM D1552-03 [16] for the analysis. The relative percent difference (RPD) is 3.9% for LSDF and 1.9% for HSDF. Therefore,

148 the accuracy of this approach for total sulfur in the diesel fuels was verified through analysis of a

NIST standard and comparison with results obtained for the two fuels by an ASTM method.

6.3.2 PASH Speciation in Diesel Fuel

As was done in other studies, both carbon and sulfur selective modes were used to identify the compounds. Carbon (179 nm) and sulfur (181 nm) AED chromatograms for LSDF and HSDF are shown in Figure 6-2 and Figure 6-3, respectively. The total sulfur content and concentration of individual sulfur compounds are listed in Table 6-4. Samples were screened for twenty six different compounds, with 13 (T, 2-MT, 3-MT, 2,3-DMT, 2,5-DMT, 2-ET, 2-PT,

BN12T, BN21T, BN23T, Ph34T, BPh9,10T, and DiAT) not detected in diesel fuels. In Table 6-4,

3-MBT and 5-MBT were identified based on chemical standards. However, coelution may exist for 3- and 4-MBT, and for 5- and 6-MBT, which was reported by Depauw et al. [17].

The chromatograms for both LSDF and HSDF indicate a large number of alkylated

PASH isomers exist in the two fuels. Although we don’t have standards for each of the peaks, an isomer grouping method can be used to categorize these compounds into several groups according to their retention times and other properties [6]. The following possible groups can be assigned based on this comparison: C1-BTs (12.0-13.0 min), C2-BTs (13.5-14.5 min), C3-BTs

(14.6-15.8 min), C4-MBTs (16.8-17.2 min), C1-DBTs (18.8-19.6 min), C2-DBTs (19.6-20.9 min), and C3-DBTs (21.0-22.0 min). These groups account for more than 50% of the total sulfur in diesel fuel, with C2-DBTs as the largest fraction for LSDF and C3-BTs for HSDF (Figure 6-4).

Compared to previously reported chromatograms [6] for crude oil, the grouping pattern of diesel fuel resembles the middle fraction of crude oil.

149

Retention time

Figure 6-2 Carbon (179 nm) and sulfur (181 nm) AED chromatograms for low sulfur (433 ppmw) diesel fuel.

Retention time

Figure 6-3 Carbon (179 nm) and sulfur (181 nm) AED chromatograms for high sulfur (2284 ppmw) diesel fuel.

150 Table 6-4 PASHs in low and high sulfur diesel fuels (LSDF and HSDF). Compound Formula Sulfur contenta (µg S/g DF) ± std (δ; n = 3) LSDF HSDF Total organic sulfurb S 433±9 2284±12 T to 2-PT (7 compounds)c NDd ND

BT C8H6S 1.34±0.05 8.21±0.17

2-MBT C9H8S 0.11±0.02 7.68±0.14 e 3-MBT C9H8S 3.21±0.09 8.39±0.12 f 5-MBT C9H8S 2.86±0.13 15.29±0.45

3,5-DMBT C10H10S 3.85±0.16 18.71±0.42

2,3,5-TMBT C11H12S 8.28±0.24 63.56±2.46

2,3,7-TMBT C11H12S 10.72±0.19 75.91±2.78

2,3,4,7-TTMBT C12H14S 4.32±0.11 24.36±0.56

DBT C12H8S 15.23±0.29 83.99±1.92

4-MDBT C13H10S 21.22±0.52 66.78±1.64

4,6-DMDBT C14H12S 10.95±0.31 20.81±0.35

2,4,6-TMDBT C15H14S 4.34±0.15 6.63±0.24 g Ph45T C14H8S 2.08±0.13 3.03±0.12 h BN12T to DiAT (6 compounds) C16H10S ND ND a For both fuels (LSDF and HSDF), three sets of samples were injected. b Total sulfur was calculated by integrating the entire sulfur response as one peak. Integration began where the sulfur emission line increased from baseline or at the beginning of the first peak (7 min for diesel fuel). The final integration point was designated after the sulfur emission returned to baseline and remained stable (25 min for diesel fuel). The standard temperature program indicated in Table 6-1 was used. c The 7 compounds include compound No. 1 to No. 7 listed in Table 6-2. d ND (not detected) indicates result below the minimum detectable level (MDL). MDLs are based on results for a test mixture (Agilent part 8500-5067). The specification for sulfur 181 is 2 pg/sec. The MDL for our laboratory was typically about 0.5 pg/sec, or 1 pg S for a peak having a 2-sec width (at half height). e The coelution between 3- and 4-MBT may occur. f The coelution between 5- and 6-MBT may occur. g This peak is assigned as Ph45T based on the retention time of a standard compound. It is possible that this peak also contains a C3-DBT because C3-DBTs can h also elute in this region. Currently, we do not have a C3-DBT standard to examine this possibility. The 6 compounds include compound No. 21 to No. 26 listed in Table 6-2.

151 60

50 C3 -DBTs

40 C2 -DBTs

C1 -DBTs 30 C4 -BTs

20 C3 -BTs

C2 -BTs 10 Percentage oftotal sulfur (%) C1 -BTs 0 LSDF HSDF

Figure 6-4 Possible alkylated PASH groups in low and high sulfur diesel fuels (LSDF and HSDF, respectively).

For most diesel fuels, the sulfur content is 0.01-0.5% by weight. Currently, the EPA is proposing stringent regulations for the sulfur content of diesel fuels. For highway diesel fuel, sulfur content will be reduced from 500 ppmv to 15 ppmv on June 1, 2006 [18]. For non- road applications, the current EPA regulation is 3400 ppmv, but it will be reduced to 500 ppmv in 2007 and further to 15 ppmv in 2010 [19]. As the total sulfur content of diesel fuel is reduced from high to low, the distribution of organosulfur compounds also changes, as illustrated in Figure 6-5(a) and (b). In Figure 6-5(b), the concentration was normalized by total sulfur content in diesel fuel, which is different from Table 6-4.

152 90 80 70 (a) LSDF (TS=433 ppm) 60 HSDF (TS=2284 ppm) 50 40 30 20 Sulfur(µg S/g DF) 10 0 60

50 (b) LSDF (TS=433 ppm) 40 HSDF (TS=2284 ppm)

10 Sulfur (mg S/g TS in DF) 0 BT DBT DiAT Ph45T Ph34T 2-MBT 3-MBT 5-MBT BN12T BN21T BN23T 4-MDBT BPh9,10T 3,5-DMBT 4,6-DMDBT 2,3,5-TMBT 2,3,7-TMBT 2,4,6-TMDBT 2,3,4,7-TTMBT

Figure 6-5 PASHs in low and high sulfur diesel fuels (LSDF and HSDF, respectively) expressed in (a) µg S/g DF, and (b) mg S/g TS in DF.

Figure 6-5(a) shows that the concentrations of all the sulfur compounds are higher in

HSDF than in LSDF. The most abundant PASH species in both the low and high sulfur diesel fuels are DBT, 4-MDBT, 2,3,5-TMBT, and 2,3,7-TMBT. In LSDF, 4,6-DMDBT is also one of the major PASHs, whereas its relative abundance in HSDF is low. Figure 6-5(b), in which the concentrations of compounds are expressed in terms of mg S/g total sulfur (TS) in DF, shows that the relative abundance of the compounds is different for the two fuels. In the HSDF, the relative abundance of the lighter compounds is higher, while in LSDF, the abundance of heavier compounds is higher. Based on the chromatograms (Figure 6-2 and Figure 6-3), it also can be

153 seen that there are some lighter sulfur compounds (presuming compounds with shorter retention time are relatively lighter compounds) that are major components of HSDF, which is not the case for LSDF. This is consistent with the results of desulfurization in that lighter compounds are more easily removed from the fuel [5, 20]. It is also reasonable to expect that with the future fuel sulfur reduction, the organosulfur species remaining in the fuel will be more inclined toward the heavier (more than 3 rings) side.

6.3.3 PASHs in Diesel Emissions

In diesel emissions, sulfur compounds exist in both the gaseous and particulate fractions.

Sulfur dioxide (SO2) is a priority air pollutant regulated under the National Ambient Air Quality

2- Standards. Sulfate (SO4 ) is thought to play an important role in fine particle formation and the nucleation of organic compounds, including polycyclic aromatic hydrocarbons (PAHs), reportedly through formation of heavy hydrocarbons from reaction between adsorbed sulfuric acid and organic compounds in the exhaust [21-23]. Gas-phase sulfur compounds (SO2, SO3) and

2- particulate sulfate (SO4 ) in diesel exhaust have been investigated, but there have been limited studies on the organosulfur species in DPM and the influence of fuel sulfur on this fraction.

Although this fraction is of small percentage of sulfur emissions from diesel engines, there is a necessity to study these organosulfur species, their concentrations and distribution, since these organosulfur species are hazardous, some of which are carcinogenic, sorbed on DPM and transferred into the lung with DPM.

Both gas and particulate diesel emissions were collected and the extracts were injected into the GC/AED system. Samples were screened for carbon 179 and sulfur 181 and the sulfur

181 AED chromatograms for the extracts are shown in Figures 6-6(a-c). A total of twenty three

154 sulfur compounds were identified and quantified in either gas phase or particulate phase extract.

The concentrations of identified sulfur compounds are listed in Table 6-5, the total organic sulfur content (TOS) for each sample was calculated and summarized in Table 6-5 as well.

Retention time

Figure 6-6 Sulfur (181 nm) AED chromatograms of gas and particulate phase diesel emissions generated at 0 kW with low sulfur (433 ppmw) diesel fuel. (a) Collected with PUF, (b) collected with XAD, and (c) collected on filter.

155 Table 6-5 PASHs in gas and particle phase diesel emissions (0 kW through 75 kW) when burning low and high sulfur diesel fuels. Compound Formula Sulfur content (µg S/g DPM) ± std (δ; n = 3)a Gas phase emission at LSDF L0-Gb L25-G L50-G L75-G Total organic sulfurc S 775±35 786±27 450±26 400±30 d T C4H4S ND ND ND ND

2-MT C5H6S 5.56±0.62 10.15±0.89 3.42±0.33 5.41±0.39

3-MT C5H6S 9.90±0.78 19.01±1.25 6.10±0.52 10.37±0.89

2,3-DMT C6H8S 3.86±0.45 5.32±0.63 2.09±0.29 4.03±0.44

2,5-DMT C6H8S 4.17±0.54 6.68±0.56 2.45±0.28 4.59±0.38

2-ET C6H8S 4.03±0.49 6.82±0.61 2.72±0.32 4.13±0.26

2-PT C7H10S 1.73±0.28 4.35±0.52 1.54±0.27 2.76±0.37

BT C8H6S 9.56±0.82 7.76±0.67 3.96±0.36 4.18±0.43

2-MBT C9H8S 9.52±0.71 5.11±0.43 3.67±0.41 4.41±0.51 e 3-MBT C9H8S 12.20±0.88 9.22±0.79 5.88±0.51 5.87±0.55 f 5-MBT C9H8S 4.90±0.52 2.24±0.33 2.06±0.28 2.14±0.30

3,5-DMBT C10H10S 9.11±0.63 9.34±0.86 6.49±0.56 5.25±0.49

2,3,5-TMBT C11H12S 23.31±1.57 17.87±1.49 12.59±0.97 10.92±0.93

2,3,7-TMBT C11H12S 31.36±2.05 23.86±1.84 16.96±1.27 14.92±1.22

2,3,4,7-TTMBT C12H14S 2.28±0.33 1.85±0.26 1.38±0.18 0.85±0.17

DBT C12H8S 22.82±1.55 13.00±1.06 10.76±0.93 7.68±0.68

4-MDBT C13H10S 9.86±0.81 5.29±0.55 5.91±0.56 5.31±0.47

4,6-DMDBT C14H12S 3.59±0.42 ND ND ND Compound No. 19 to 26 listed in Table 6-2 ND ND ND ND

Particle phase emission at LSDF L0-P L25-P L50-P L75-P Total organic sulfur S 679±32 734±27 819±25 631±35 Compound No. 1 to No. 14 listed in Table 6-2 ND ND ND ND

2,3,4,7-TTMBT C12H14S 1.48±0.15 1.96±0.28 1.53±0.25 1.63±0.22

DBT C12H8S 2.84±0.23 2.68±0.32 1.77±1.89 1.51±1.75

4-MDBT C13H10S 9.07±0.68 7.33±0.88 4.71±0.55 3.55±0.41

4,6-DMDBT C14H12S 10.30±1.08 12.40±1.03 11.57±1.21 11.34±1.09

2,4,6-TMDBT C15H14S 13.37±1.22 34.50±2.94 27.69±2.49 22.38±2.23 g Ph45T C14H8S 6.42±0.75 22.39±2.19 14.86±1.78 11.68±0.94

BN12T C16H10S 0.99±0.37 2.37±0.36 4.09±0.83 5.58±0.63

BN21T C16H10S 1.13±0.25 2.87±0.43 3.83±0.75 5.11±0.82

BN23T C16H10S 0.95±0.22 1.74±0.31 2.23±0.41 4.69±0.76

Ph34T C16H10S ND ND ND 2.25±0.50

BPh9,10T C20H12S ND ND ND ND

DiAT C24H12S ND ND ND ND

156 Table 6-5 (Continued) Compound Formula Sulfur content (µg S/g DPM) ± std (δ; n = 3) Particle phase emission at HSDF H0-P H25-P H50-P H75-P Total organic sulfur S 1585 1599 1349 923 Compound No. 1 to No. 12 listed in Table 6-2 ND ND ND ND

2,3,5-TMBT C11H12S 2.81 1.93 0.98 1.85

2,3,7-TMBT C11H12S 3.44 3.40 2.00 1.25

2,3,4,7-TTMBT C12H14S 14.28 12.67 8.52 2.49

DBT C12H8S 30.29 25.18 11.21 4.41

4-MDBT C13H10S 32.39 26.86 19.81 11.39

4,6-DMDBT C14H12S 22.91 29.28 15.31 8.25

2,4,6-TMDBT C15H14S 28.08 42.40 27.20 18.56

Ph45T C14H8S 19.05 34.78 17.80 9.55

BN12T C16H10S 1.89 3.22 5.54 7.68

BN21T C16H10S 1.69 3.06 4.61 6.86

BN23T C16H10S 1.50 2.98 4.33 6.49

Ph34T C16H10S ND 0.47 2.48 0.48

BPh9,10T C20H12S ND ND ND 0.81

DiAT C24H12S ND ND ND ND

Total organic sulfur in gas phase at HSDF 1768 1654 1282 1077 a For DPM at LSDF, three sets of samples were collected and injected, but for DPM at HSDF only one set of samples was collected. b L or H indicates low or high sulfur diesel fuel was used; 0 through 75 is the engine load (kW); G or P indicates gas phase or particle phase. c Total sulfur was calculated by integrating the entire sulfur response as one peak. Integration began where the sulfur emission line increased from baseline or at the beginning of the first peak (7 min for DPM). The final integration point was designated after the sulfur emission returned to baseline and remained stable (32 min for DPM). The standard temperature program indicated in Table 6-1 was used. Notes d-g are the same as those for Table 6-4.

Compared with the chromatograms of diesel fuels, the chromatograms of diesel emission extracts (Figure 6-6(a-c)) show more sulfur peaks which include the peaks with the retention times shorter than 12 min in PUF and XAD extracts and the peaks with retention times longer than 22 min in DPM extracts. The large hump in Fgiure 6-6(c) represents a complex sulfur compound matrix in DPM. From Table 6-5, C3-BT, DBT, C1-DBT and C2-DBT, which are abundant in diesel fuels, are still predominant in diesel emissions. However, some lighter

157 compounds in gas phase, such as C1-T, BT, and C1- and C2-BT, and some heavier compounds in particulate phase, such as C3-DBT and PH45T, are also important and account for large percentages in diesel emissions. The occurrence of light compounds in diesel emissions indicates the possibility of decomposition of the sulfur compounds in diesel fuels during the combustion process. On the other hand, some heavier PASHs which do not exist in diesel fuels, such as

BN12T, BN21T, BN23T, Ph34T and BPh9,10T, are present in DPM with high fractions, especially at high load conditions. These PAHs appear to be generated through the pyrogenic pathway during the combustion process. These larger PASHs may increase the toxicity of DPM, as some three-, and especially four- and five-ring species, are known to be mutagenic [24, 25].

6.3.3.1 PASH Distribution between Gas and Particle Phase

Figure 6-7 presented the organosulfur compounds in both gas and particulate phase diesel emissions at 0 kW when burning low sulfur diesel fuel. Although the organosulfur species are different in gas and particulate phase diesel emissions, their concentrations are comparable which are in the range of 0-35 µg S/g DPM, the total organic sulfur contents in gas and particulate phases are comparable as well (400-820 µg S/g DPM). Thiophene, benzothiophene, and their alkylated homologue are primarily present in gas phase, while dibenzothiophene, its alkylated homologue, and other high-molecular-weight sulfur compounds (4- and 5-ring PASHs) are present predominantly in DPM. 7-ring PASH diacenaphthothiophene was not detected in diesel emissions, however, heavier PASHs, which might include 6- or 7-ring PASHs, could exist in DPM as shown by Figure 6-6(c). Due to the lack of standard reference compounds, those heavier PASHs were not measured in this study. Similar gas/particle distributions of organosulfur compounds were obtained for other engine load conditions.

158 35

30 L0-Gas 25 L0-Particle 20 15

Sulfur (µg S/g DPM) S/g (µg Sulfur 5 0 BT DBT 2-ET 2-PT 2-MT 3-MT 2-MBT 3-MBT 5-MBT Ph45T BN12T BN21T BN23T 4-MDBT 2,3-DMT 2,5-DMT 3,5-DMBT 2,3,5-TMBT 2,3,7-TMBT 4,6-DMDBT 2,4,6-TMDBT 2,3,4,7-TTMBT

Figure 6-7 Organosulfur compounds in gas phase and particulate phase diesel emissions at 0 kW and low sulfur diesel fuel.

The results have shown that the organosulfur compounds occurring in gas phase diesel emission are mostly light ones which are less toxic compared with those heavier one present in

DPM, therefore, when discussing the concentrations and distributions of PASHs in the following context, only the particulate phase PASHs will be referred to.

6.3.3.2 Total Organic Sulfur in DPM

Results for the total organic sulfur (TOS) in particulate phase and gas + particulate phase diesel emissions for LSDF and HSDF under different engine loads are shown in Figure 6-8. As evidenced by these figures, fuel sulfur content has a significant effect on the concentrations of total organic sulfur in diesel emissions. When fuel sulfur is increased from 433 ppmw to 2284 ppmw, the total organic sulfur in DPM is increased by a factor of 1.34 at 0 kW. At high load conditions, the total organic sulfur content in DPM is less dramatically affected by fuel sulfur, increased by a factor of 0.46 at 75 kW. As reported, when fuel sulfur was reduced from 0.2% to

159 0.05%, a 7-12% reduction in particulate mass can be obtained [26]. Relative to DPM mass, the amount of total organic sulfur content is much more affected by the change in fuel sulfur content.

Figure 6-8 also indicates that engine loads have some effects on total organic sulfur content in both gas and particulate phase diesel emissions, especially when burning HSDF. As the engine load increasing, the total organic sulfur content in gas + particulate phase diesel emissions decreased.

3600 3300 3000 at HSDF (G+P) 2700 2400 2100 at LSDF (G+P) 1800 1500 1200 at HSDF (P) 900

Sulfur (µg S/g DPM) 600 300 at LSDF (P) 0 0255075 Engine Loads (kW)

Figure 6-8 Total organic sulfur in particulate phase and gas + particulate phase diesel emissions for LSDF and HSDF at various load conditions.

When examining the effects of fuel sulfur and engine load, it is convenient and reasonable to compare the conversion rate of sulfur. In this study, the conversion rate is defined as the mass (µg) of organic sulfur in diesel emissions (or DPM) per gram of sulfur input from the fuel. Figure 6-9 represents the conversion rate of fuel sulfur to organic sulfur in diesel emissions for LSDF and HSDF at different engine loads.

160 900 800 700 at LSDF (G+P) 600 500 at HSDF (G+P) 400 at LSDF (P) 300 Conversion rate 200 g S in DPM/ g S input) at HSDF (P) µ ( 100 0 0255075 Engine loads (kW)

Figure 6-9 Conversion rate of fuel sulfur to the emitted organic sulfur for HSDF and LSDF at different engine loads.

As shown in Figure 6-9, for LSDF, only 0.069-0.079% of fuel sulfur is converted to (or remains as) organic sulfur in diesel emissions, among which 47-65% is in particulate phase. This means more than 99.9% of fuel sulfur is converted to inorganic sulfur, including gaseous sulfur

2- dioxide (SO2), sulfur trioxide (SO3), and sulfuric acid (H2SO4) and particulate sulfate (SO4 ).

Liu et al. [27] has reported that the conversion of fuel sulfur to total particulate sulfur was in the range of 0.12-0.50%, among which 50-100% occurs in the form of particulate sulfate. This result means 0-50% of particulate sulfur, which is 0-0.25% of fuel sulfur input, may be in the form of organic sulfur, which is consistent with our results. For high sulfur diesel fuel, the conversion of fuel sulfur to organic sulfur in diesel emissions is even lower, ranging from

0.027% to 0.043%. A possible reason for the lower recovery is that higher sulfur diesel fuel contains a higher proportion of lighter sulfur components, and these lighter sulfur compounds are

2- mostly combusted and converted to SO2 and SO4 .

Figure 6-9 also indicates that the conversion of fuel sulfur to particulate organic sulfur is not well correlated with engine load and the trend is different for LSDF and HSDF. For LSDF,

161 the conversion of fuel sulfur to particulate organic sulfur is higher at 50 kW and 75 kW than at 0 and 25 kW, while for HSDF, the situation is opposite. Some studies have reported that the sulfur conversion rate to particulate sulfur is not strongly correlated with engine load, which is consistent with our results [27]. The conversion of fuel sulfur to total organic sulfur in diesel emissions, including both gas phase and particulate phase, is basically higher at low loads than at high loads. The possible reason is that at high loads, the combustion temperature is high, which

2- causes the fuel sulfur being more completely converted to SO2 and SO4 .

6.3.3.3 Effect of Fuel Sulfur on DPM PASH Distribution

Fuel sulfur content not only affects the concentrations of organosulfur compounds in diesel emission, but also affects their distribution. Table 6-5 has shown that as fuel sulfur is increased from 433 ppmw to 2284 ppmw, the concentration of individual compounds in DPM increases by 2-9 times. The effect of fuel sulfur on the distribution of organosulfur in DPM is similar for each load condition. Figure 6-10 shows the PASH distributions in DPM for LSDF and

HSDF with 0 kW and 75 kW as examples. A significant influence from fuel sulfur content on the distribution of sulfur compounds in DPM has been observed. With low sulfur diesel fuel, the distribution curves are sharp, with 2,4,6-TMDBT as the most abundant PASH; while for DPM from HSDF, the curves are broader (more uniformed distribution), especially at 0 kW. In addition to 2,4,6-TMDBT, DBT, 4-MDBT, and 4,6-DMDBT are all major PASH products. This distribution suggests that DPM generated from HSDF contains more PASHs with lower molecular weights relative to LSDF. The presence (Figure 6-10) of heavier PASHs (i.e.,

BPh9,10T) was observed at maximum load in DPM from HSDF, but not from LSDF. This may

162 indicate a greater possibility of forming higher molecular weight PASHs when the sulfur concentration is higher.

35 30

25 L0 20 H0 15 L75 10 H75

Sulfur(µg S/g DPM) 5 0 DBT DiAT Ph45T Ph34T BN12T BN21T BN23T 4-MDBT BPh9,10T 2,3,5-TMBT 2,3,7-TMBT 4,6-DMDBT 2,4,6-TMDBT 2,3,4,7-TTMBT

Figure 6-10 Organosulfur distribution in DPM under 0 kW and 75 kW for LSDF (L) and HSDF (H).

6.3.3.4 Effect of Engine Load on DPM PASH Distribution

With both fuels, engine load affects the distribution of organosulfur compounds due to the variations in fuel usage rate, excess oxygen and combustion temperature, etc. Figures 6-11(a- c) illustrate the sulfur 181 AED chromatograms for the diesel emission extracts from PUF, XAD, and filter fractions. Comparing the chromatograms at 0 kW and 75 kW, engine load does not significantly affect organosulfur compounds in gas phase diesel emissions, while alters the organosulfur species and their distribution in particulate phase. In particulate phase diesel emissions, an obvious shift to heavier sulfur compounds is observed at high load, with most of the lighter compounds being absent.

163

Retention time

Figure 6-11 Sulfur (181 nm) AED chromatograms of gas and particulate phase diesel emissions generated at 75 kW with low sulfur (433 ppmw) diesel fuel. (a) Collected with PUF, (b) collected with XAD, and (c) collected on filter.

As the engine load was increased from 0 kW to 75 kW, three different types of PASH distributions in DPM at LSDF and HSDF were observed: increasing with load, decreasing with load, and bell-shaped distribution. In Figures 6-12(a-c), DPM at HSDF was used as an example, and the concentrations were expressed in terms of µg S/g DPM and µg S/kg DF (normalized to diesel fuel usage).

164 35 10 2,3,5-TMBT 30 8 25 2,3,7-TMBT 6 20 2,3,4,7-TTMBT 15 4 10 DBT 5 2 4-MDBT Sulfur (µgS/kg DF) Sulfur(µg S/g DPM) 0 0 0255075 0255075 Engine load (kW) Engine load (kW)

(a) Decreasing with load

50 14 12 40 4,6-DMDBT 10 30 8 2,4,6-TMDBT 20 6 4 10 2 Ph45T Sulfur (µg S/kg DF) Sulfur(µg S/g DPM) 0 0 0255075 0255075 Engine load (kW) Engine load (kW)

(b) Bell-shaped distribution

10 2.5 BN12T 8 2.0

6 1.5 BN21T 4 1.0 BN23T 2 0.5 Sulfur (µg S/kg DF) Sulfur (µgS/g DPM) 0 0.0 0255075 0255075 Engine load (kW) Engine load (kW)

Figure 6-12 PASH concentration, expressed in terms of µg S/g DPM and µg S/kg DF, in DPM generated with HSDF vs. engine loads.

For PASHs with lower molecular weights (MW ≤ 198), such as 2,3,5-TMBT, 2,3,7-

TMBT, 2,3,4,7-TTMBT, DBT, and 4-MDBT (2,3,5-TMBT and 2,3,7-TMBT were not found in

DPM from LSDF), the concentration was reduced as load increased (Figure 6-12(a)). These

165 PASHs also are present in diesel fuels. As the load was increased, the fuel input increased, which should result in a concentration increase for these lighter PASHs. However, as load increased, the combustion temperature also increased and a greater fraction of the lighter sulfur compounds was completely burned and converted to SO2 and SO3. As a result, a lower amount of these

PASHs remained in the particulate fraction.

For higher molecular weight PASHs (MW ≥ 234), which include BN12T, BN21T,

BN23T, Ph34T and BPh9,10T, the concentration increased with load (Figure 6-12(c)). These heavier PASHs were not present in diesel fuels and are possibly formed during combustion. As the load increased, so did the combustion temperature and fuel input, which may promote formation of higher molecular weight PASHs.

For intermediate molecular weight PASHs (198 < MW < 234), such as 4,6-DMDBT,

2,4,6-TMDBT and Ph45T, a bell-shaped distribution was observed (Figure 6-12(b)). As load was increased from 0 kW to 75 kW, the concentration first increased and then decreased, with the highest concentration observed at approximately 25 kW. These compounds are present in diesel fuels, and this distribution may be the result of competition between the increase in fuel input and combustion decomposition. As load increased, more fuel input led to a larger amount of

PASHs in the DPM, while the increased combustion decomposition lowered the PASH concentration. The combination of both effects resulted in the distribution shown in Figure 6-

12(b).

6.4 Summary and Conclusions

PASHs in LSDF and HSDF and the corresponding DPM emissions were determined by

GC/AED. Results for calibration standards, NIST SRM 2724b, and fuels indicated the sulfur

166 response factor in this application is relatively independent of analyte structure. As expected, the

PASH concentration and distribution in diesel fuels depends on the sulfur content. The most abundant sulfur species are DBT, C1-DBTs, and C3-BTs in both diesel fuels. In low sulfur diesel fuel, heavier PASHs such as DMDBT are also higher in concentration, which indicates the LSDF contains a larger fraction of heavier components than the HSDF.

In additions to C3-BT, DBT, C1-DBT, and C2-DBT which are abundant in diesel fuel, the most abundant PASHs in diesel emissions also include some lighter organosulfur compounds in gas phase, such as C1-T, BT, and C1- and C2-BT, and some heavier PASHs in particulate phase, such as C3-DBT and PH45T. Some heavier PASHs (e.g., BNTs) not found in diesel fuels, which are more hazardous, are present in DPM and may be generated through the pyrogenic pathway during the combustion process. 0.027-0.079% of fuel sulfur has been converted to emission organic sulfur. The organic sulfur content and PASH concentration and distribution in DPM are affected by both fuel sulfur and engine load. Higher fuel sulfur led to the following two consequences: higher total organic sulfur in diesel emissions and more PASHs with lower molecular weights in DPM. Engine load is an important factor affecting PASH distribution. With engine load increases, the dominant PASH species in DPM shift to higher molecular weight.

Three types of distributions were observed as the engine load was increased from 0 kW to 75 kW.

The concentrations decreased for lighter PASHs and increased for heavier PASHs as load was increased from 0 kW to 75 kW. For intermediate molecular weight PASHs, a bell-shaped distribution was observed, with the highest concentrations at 25 kW.

6.5 References

1. Ma, X.L.; Sakanishi, K.Y.; and Mochida, I., Hydrodesulfurization reactivities of various sulfur compounds in diesel fuel, Ind. Eng. Chem. Res., 33(2): 218-222, 1994.

167 2. Mössner, S.G.; and Wise, S.A., Determination of polycyclic aromatic sulfur heterocycles in fossil fuel-related samples, Anal. Chem., 71(1): 58-69, 1999.

3. Jacob, J., Sulfur Analogues of Polycyclic Aromatic Hydrocarbons (Thiaarenes), Cambridge University Press, Cambridge, UK, 1990, p 70.

4. Shafi, R.; and Hutchings, G.J., Hydrodesulfurization of hindered dibenzothiophenes: an overview, Catalysis Today, 59(3-4): 423-442, 2000.

5. Song, C.S., An overview of new approaches to deep desulfurization for ultra-clean gasoline, diesel fuel and jet fuel, Catalysis Today, 86(1-4): 211-263, 2003.

6. Hegazi, A.H.; Andersson, J.T.; and El-Gayar, M.S., Application of gas chromatography with atomic emission detection to the geochemical. investigation of polycyclic aromatic sulfur heterocycles in Egyptian crude oils, Fuel Process. Technol., 85(1): 1-19, 2003.

7. Breitkreitz, M.C.; Raimundo, I.M.; Rohwedder, J.J.R.; et al., Determination of total sulfur in diesel fuel employing NIR spectroscopy and multivariate calibration, Analyst, 128(9): 1204-1207, 2003.

8. Speight, J.G., Handbook of Petroleum Product Analysis, New York: Wiley-Interscience, 2002, p 43.

9. US Environmental Protection Agency (EPA) Health Assessment Document for Diesel Engine Exhaust. Prepared by the National Center for Environmental Assessment, Washington, DC, for the Office of Transportation and Air Quality; EPA/600/8-90/057F, 2002. Available from: National Technical Information Service, Springfield, VA; PB2002-107661.

10. NIOSH, Current Intelligence Bulletin 50 - Carcinogenic Effects of Exposure to Diesel Exhaust, US Department of Health and Human Services, Public Health Service, Centers for Disease Control, National Institute for Occupational Safety and Health, DHHS (NIOSH), Cincinnati, OH, NIOSH Publication no. 88-116, 1988.

11. Liang, F.Y.; Lu, M.M.; Keener, T.C.; et al., The organic composition of diesel particulate matter, diesel fuel and engine oil of a non-road diesel generator, J. Environ. Monit., 7(10): 983-988, 2005.

12. Andersson, J.T., Some unique properties of gas chromatography coupled with atomic- emission detection, Anal. Bioanal. Chem., 373(6): 344-355, 2002.

13. Becker, G.; and Colmsjö, A., Gas chromatography atomic emission detection for quantification of polycyclic aromatic sulfur heterocycles, Anal. Chim. Acta, 376(3): 265- 272, 1998.

14. Andersson, J.T.; and Schmid, B., The atomic-emission detector in gas-chromatographic trace analysis - some studies on the performance and applications, Fresenius J. Anal. Chem., 346(4): 403-409, 1993.

168 15. Pedersen-Bjergaard, S.; Asp, T.N.; and Greibrokk, T., Determination of sulfur-containing and chlorine-containing compounds using capillary gas-chromatography and atomic emission detection, Anal. Chim. Acta, 265(1): 87-92, 1992.

16. American Society for Testing and Materials (ASTM), Annual Book of ASTM Standards, vol. 05.01, ASTM, Conshohocken, PA, 2005.

17. Depauw, G.A.; and Froment, G.F., Molecular analysis of the sulphur components in a light cycle oil of a catalytic cracking unit by gas chromatography with mass spectrometric and atomic emission detection, J. Chromatogr. A, 761(1-2): 231-247, 1997.

18. US EPA, Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements, EPA 420-F-00-057, December 2000.

19. US EPA, Office of Transportation and Air Quality, Regulatory Announcement, EPA420- F-04-032, May 2004.

20. Schulz, H.; Böhringer, W.; Ousmanov, F.; et al., Refractory sulfur compounds in gas oils, Fuel Process. Technol., 61(1-2): 5-41, 1999.

21. Shi, J.P.; and Harrison, R.M., Investigation of ultrafine particle formation during diesel exhaust dilution, Environ. Sci. Technol., 33(21): 3730-3736, 1999.

22. Duran, A.; Carmona, M.; and Ballesteros, R., Competitive diesel engine emissions of sulphur and nitrogen species, Chemosphere, 52(10): 1819-1823, 2003.

23. Yu, F.Q.; Turco, R.P.; and Karcher, B., The possible role of organics in the formation and evolution of ultrafine aircraft particles, J. Geophys. Res.-Atmos., 104(D4): 4079-4087, 1999.

24. Pelroy, R.A.; Stewart, D.L.; Tominaga, Y.; et al., Microbial mutagenicity of 3-ring and 4- ring polycyclic aromatic sulfur heterocycles, Mutat. Res., 117(1-2): 31-40, 1983.

25. McFall, T.; Booth, G.M.; Lee, M.L.; et al., Mutagenic activity of methyl-substituted tricyclic and tetracyclic aromatic sulfur heterocycles, Mutat. Res., 135(2): 97-103, 1984.

26. Neeft, J.P.A.; Makkee, M.; and Moulijn, J.A., Diesel particulate emission control, Fuel Process. Technol., 47(1): 1-69, 1996.

27. Liu, Z.; Lu, M.; Keener, T.C.; et al. The sulfur speciation of diesel emissions from a non- road generator. in Proceedings of the 98th A&WMA Annual Conference & Exhibition. June 21-24, 2005. Minneapolis, MN.

169 Chapter 7

CONCLUSIONS AND RECOMMENDATIONS

7.1 Conclusions

Diesel particulate matter (DPM) has been regarded as a potential occupational carcinogen and the hazardous health effect of DPM emissions has been confirmed and regulated.

Considering the adverse health effect of DPM emissions, significant efforts have been put on studying DPM emissions from on-road diesel vehicles. However, there have been few studies on the chemical composition and formation mechanisms of DPM from non-road diesel engines, especially as a function of fuel properties and testing conditions, even though the use of non-road diesel engines has increased rapidly and the resultant DPM emissions are of significant quantity.

In this study, GC/MS with selective ion chromatogram (SIC) and tandem mass spectrometry (MS/MS) with selected ion monitoring (SIM) were developed and applied for the identification and quantification of PAHs. The successful measurement of PAHs in the complex

DPM matrix indicates that SIC and MS/MS techniques are effective and provide satisfactory selectivity and low detection limit for measuring trace components in complex matrices.

Diesel particulate matter from a non-road diesel generator burning low- and high-sulfur diesel fuel was collected using the dilution method and the stack sampling method at variated engine load conditions. The DPM chemical composition was instrumentally determined and the results were compared under variated fuel sulfur contents, engine loads, and sampling methods.

The chemical composition of possible sources contributed to DPM emissions was also characterized and compared with DPM composition. These experiments were conducted to

170 provide important information for understanding DPM formation mechanisms and assessing its health effects.

The research findings suggest that DPM contains 52-84% soluble organic fraction, among which only a small percentage can be resolved by GC/MS. In the GC resolvable fraction, alkanes, PAHs and alkylated PAHs, and organic acids, which are attributed to the unburned diesel fuel, the combustion process, and the evaporated engine oil, respectively, are predominant species. This result suggests that DPM control strategies should emphasize the control of fuel properties, and engine designs, as well as the engine oil quality. The comparison between the composition of the diesel fuel and DPM has indicated that DPM contains more fractions of higher molecular weight compounds, which are likely resultant from pyrogenic processes.

The experimental findings have indicated that the gas/particle distribution of organic compounds in diesel emissions is significantly related with their vapor pressures. Light

-4.07 compounds with vapor pressures higher than 10 torr, such as n-alkanes C10-C18 and 2- and 3- ring PAHs, are predominantly present in the gas phase emission, while heavy compounds with

-4.84 vapor pressures lower than 10 torr, such as n-alkanes C20-C25 and 5- and 6-ring PAHs, occur primarily or exclusively in the particulate phase. The G/P partitioning coefficient Kp of a

0 compound could be related to its vapor pressure pL by the following equation: log Kp = mr

0 0 log pL + br. The linear regression results have indicated that the correlation of log Kp vs. log pL is statistically significant. The shallow slopes of the linear regression results compared with estimated values suggest that the gas-to-particle adsorption of compounds is the major mechanism. The filter breakthrough for n-alkanes and non-exchangeable effects for PAHs may occur during the sampling events. The understanding of G/P partitioning for organic compounds in diesel emissions helps to evaluate the health effects of diesel emissions, which are strongly

171 related to the physical forms in which the compounds are present when the exposure occurs. In addition, studies on G/P partitioning of compounds in diesel emissions are helpful in understanding DPM formation mechanisms, especially the adsorption and absorption contributions in forming secondary diesel aerosol.

The applied research examined and confirmed the effects of fuel properties, engine operation conditions, and sampling methods on DPM emission rate, its chemical composition, and/or DPM formation mechanism. Increased engine loads result in higher emission rate of DPM, organic carbon (OC), and elemental carbon (EC). However, the fraction of OC in DPM decreases with loads. DPM formation mechanisms are also affected by engine loads. Higher engine load increases the pyrogenic formation of compounds while it weakens the role of petrogenic pathway. The sampling method is an important factor, affecting the concentration and distribution of organics in DPM. Compared with the stack sampling method, the dilution process results in an increase of 40-80% in DPM emission, and significantly increases the amount of individual organic components of DPM due to the continuous adsorption, nucleation, coagulation, and/or condensation of the large quantities of condensable organics and inorganics during dilution processes. The dilution process also alters the distribution of those organic compounds, causing the enrichment of low molecular weight components. This result suggests that the dilution process favors the formation of light compounds more than the formation of heavy compounds. The study of fuel sulfur effect on diesel emissions has indicated that when the fuel sulfur content is increased from 400 ppmw to 2200 ppmw, DPM emissions are increased by

20%, and an increase between 20-200% is obtained for individual compounds presented in diesel particulate matter. This result suggests that sulfur in diesel exhausts may act as a nucleation core for all components presented in diesel exhausts. However, the relative distribution of these

172 organic components does not significantly vary with fuel sulfur contents, which may relate to the similar composition of low and high sulfur diesel fuel. These findings will help to improve diesel engine designs, promote the establishment of more stringent regulations for diesel fuels, improve fuel reformulation technologies, and therefore reduce DPM emissions. In addition, investigating the effects of sampling methods on DPM composition will provide essential information for understanding the formation, condensation and transportation of organic compounds, and therefore helps to evaluate DPM contribution and exposure.

Polycyclic aromatic sulfur heterocycles (PASHs) are of interest due to their potential mutagenic and carcinogenic properties. However, studies on PASH content in DPM are difficult since the concentration of PASHs is low in DPM, there is a great variety of compounds present, and the quantitative determination of individual PASH isomers in the complex mixture is difficult. This study successfully applied gas chromatography with the atomic emission detection

(GC/AED) and compound independent calibration (CIC) techniques on the measurement of

PASHs in DPM, using optimized temperature program and specific reagent gases. The research also provides a potential, highly selective, and effective analytical technology for measuring organosulfur species in other complex sample matrices.

PASHs in low and high sulfur diesel fuel and the corresponding DPM emissions were determined by GC/AED. Results from fuel analysis have indicated that the PASH concentration and distribution in the diesel fuel vary with fuel sulfur contents. The amount of PASHs in high sulfur diesel fuel is higher than in low sulfur diesel fuel. However, low sulfur diesel fuel contains relatively more fractions of heavy PASHs, which are possibly more hazardous. This result suggests that more stringent regulations on fuel quality and further fuel sulfur reduction are required.

173 Experimental findings on PASHs in DPM indicate that 0.027-0.079% of the fuel sulfur has been converted to the emitted organic sulfur. DPM contains more complex PASH species than diesel fuels, with benzothiophene, dibenzothiophenes, and their alkylated homologue as the major species. Some heavier PASHs (e.g., BNTs) not found in diesel fuels are present in DPM.

They are more hazardous and may be generated through the pyrogenic pathway during the combustion process. Both fuel sulfur contents and engine loads affect the formation of PASHs in

DPM. Higher fuel sulfur content leads to higher PASH emissions. The effect of engine loads on the PASH formation in DPM is more complex, and three types of distributions are observed as the engine load is increased from 0 kW to 75 kW: decreasing with loads for lighter PASHs, increasing with loads for heavier PASHs, and bell-shaped distributions for intermediate molecular weight PASHs.

7.2 Recommendations

Although the research findings have provided important information for DPM health effect assessments, its formation mechanisms, and control strategies, more research is needed on

DPM chemical composition, especially as a function of DPM size distribution, quantitative contribution to DPM components from various sources, and sampling methodology and characterization of nano-sized DPM. The following studies are recommended:

1. The investigation on the quantitative correlation of diesel emissions from petrogenic and pyrogenic sources is recommended.

This research has identified the three sources of DPM emissions and their contributions to DPM chemical components; however, the quantitative contribution to a compound from each source has not been well studied. This research has preliminarily studied the survivability of

174 phenanthrene using procedures described in Chapter 2. The selection of phenanthrene as the additive is due to its abundance in both diesel fuel and diesel emissions. The amount of phenanthrene added (~1% by weight) is determined from the consideration that this additive would not alter the combustion process in diesel engine. The survivability value for phenanthrene calculated from Eq. (2-1) using the data listed in Table 7-1 is 0.0115%. The survivability value is considered to differ with the type of compounds, type of engines, operating condition, collection condition, and other factors [1, 2]. From Eq. (2-2), the unburned diesel fuel contributes to the major proportion (61.2%) of phenanthrene recovered in DPM, while other sources of phenanthrene, presumably pyrosynthetic formation, is 38.8% of the recovered phenanthrene in DPM. Table 7-2 summarized literature results for survivabilities and sources of

PAHs in diesel emissions. The relative importance of pyrosynthesis and unburned fuel greatly depends on PAH species and engine operating conditions.

Both radiotracer method and the method used in this study can be applied for investigations on source apportionment of PAHs in diesel emissions. Especially, the radiotracer method can also be applied for studying engine oil contributions.

From the experiment for Ph-enriched fuel, it is also found that the addition of phenanthrene affects aromatic emissions from diesel engines. For Ph-enriched diesel fuel, the amount of aromatics, especially alkylated naphthalenes and phenanthrenes and alkylbenzenes, in particulate diesel emissions has been increased to 1.3-3 times as high as for normal fuel.

However, the emission of alkanes and alkanoic acids was not affected by the addition of phenanthrene in fuel. Further investigations on combustion processes and aromatic formation mechanisms when burning Ph-enriched fuel are recommended.

175

Table 7-1 Results for phenanthrene survivability experiment. Fuel type Phenanthrene Phenanthrene Fuel Stack gas concentration in concentration in consumption flow rate diesel fuel (mg/L) DPM (ng/m3) rate (L/hr) (m3/hr) Normal fuel 198 709 5 263 Ph-enriched fuel 10000 22213 5 263

Table 7-2 Literature results for contributions from unburned fuel and pyrosysnthesis to PAHs in diesel emissions. Compounds Survivability Contribution from unburned fuel or pyrosynthesis Phenanthrene 0.0115% 61% from unburned fuel, 0 kW (this study) Most PAHs 0.05-0.2% [3] Naphthalene 0.48% [4]; 24% from unburned fuel [4]; Pyrosynthesis is the major 0.35-1%a [2] source [2]. Benzo(a)pyrene 0.04% [5] 80% from unburned fuel [5] Fluorene 0.87% [6]; Unburned fuel is the major source except for low-speed and 0.2-1.3%a [2] low-load condition [2]. Pyrene 0.17 [6]; At low load, pyrosynthesis is the major source; at high load, 0.2-0.4%a [2] pyrosynthesis and unburned fuel contribute almost equally[2]. Fluoranthene 0-0.1%a [2] Pyrosynthesis is the major source [2]. a The data was estimated from literature results.

2. Measurements of nitro-PAHs and oxygenated PAHs, which are highly carcinogenic and mutagenic, carbonyl compounds, and metals are recommended for the integrality of DPM chemical composition and the accuracy of DPM health effect assessments.

3. In addition, characterization of DPM from the combustion of a biodiesel blend, which is the mixture of biodeisel blended with petroleum diesel at some level, and the comparison of

DPM composition between biodiesel and normal petroleum diesel are also recommended.

Biodiesel is a domestic, biodegradable, and renewable fuel for diesel engines, derived from natural oils like vegetable oil or animal fat. Due to the increasing interest in the use of biodiesel,

176 the U.S. Environmental Protection Agency has conducted a comprehensive analysis of the emission impact of biodiesel using publicly available data. This investigation shows that the use of biodiesel in diesel engines results in substantial reduction of unburned hydrocarbons, carbon monoxide, and particulate matter [7]. Detailed compositional analysis of DPM from biodiesel blend will help to assess its health effects and to instruct the proper application of biodiesel.

4. The investigation on sampling methods for ultrafine and nano-sized diesel particles, and the investigation on DPM size-dependent composition are also recommended.

Measurements of ultrafine and nano-sized diesel particles are challenging, since much longer sampling time is required in order to get enough materials. High-throughput sampling method for ultrafine and nano-sized diesel particles is a necessity. The composition of diesel particles varies with their size, which leads to the size-dependent toxicity. Studying the size-dependent composition of diesel particles will be essential for a better understanding of their health effects and the development of effective emission control strategies.

5. The implementation of efforts using mathematical modeling to predict DPM emission and its chemical composition at specific conditions is recommended. This research has correlated

DPM composition with source contribution, engine operation condition, and sampling condition.

Combining the information from this research through some mathematical methods should deduce a predicting model, which will significantly contribute to estimating DPM exposure, evaluating DPM health effects, and developing effective control strategies.

It is hoped that the present applied research will contribute to the assessment of DPM health effect, the establishment of regulations on fuel quality and DPM exposure, the evolution of engine design and operations, and the development of DPM control technologies.

177 7.3 References

1. Williams, P.T.; Abbass, M.K.; Andrews, G.E.; et al., Diesel particulate emissions: The pole of unburned fuel, Combust. Flame, 75(1): 1-24, 1989.

2. Rhead, M.M.; and Hardy, S.A., The sources of polycyclic aromatic compounds in diesel engine emissions, Fuel, 82(4): 385-393, 2003.

3. Abbass, M.K.; Andrews, G.E.; and Williams, P.T., The influrence of diesel fuel composition on particulate PAH emissions, SAE Technical Paper Series 892079, 1989.

4. Rhead, M.M.; and Pemberton, R.D., Sources of naphthalene in diesel exhaust emissions, Energy Fuels, 10(3): 837-843, 1996.

5. Tancell, P.J.; Rhead, M.M.; Trier, C.J.; et al., The sources of benzo[a]pyrene in diesel exhaust emissions, Sci. Total Environ., 162(2-3): 179-186, 1995.

6. Tancell, P.J.; Rhead, M.M.; Pemberton, R.D.; et al., Survival of polycyclic aromatic hydrocarbons during diesel combustion, Environ. Sci. Technol., 29(11): 2871-2876, 1995.

7. United States Environmental Protection Agency, Draft Technical Report, A Comprehensive Analysis of Biodiesel Impacts on Exhaust Emissions, EPA420-P-02-001, October 2002.

178 Appendix A

CHROMATOGRAMS FOR SAMPLES

A.1 GC/MS Total Ion Chromatogram (TIC) for Diesel Fuel and Diesel Emissions

kCounts RIC all No.2DF-1000--4-29-2004.SMS 600 Diesel fuel

500

400

300

200

100

0 MCounts RIC all L0-Fillter.SMS 1.25 DPM

1.00

0.75

0.50

0.25

0.00 MCounts RIC all L0-PUF-Dilute.SMS Gas phase diesel 0.75 emission: PUF fraction

0.50

0.25

0.00 MCounts RIC all L0-XAD-Dilute.SMS Gas phase diesel 1.25 emission: XAD fraction 1.00

0.75

0.50

0.25 0.00

Figure A.1 Total ion chromatogram of diesel fuel and DPM (time in minutes).

179 A.2 GC/MS Selected Ion Chromatogram (SIC) for Diesel Fuel and Diesel Emissions

kCounts Ion: 128 all No.2DF-1000--4-29-2004.SMS 7 6 Naphthalene Ion: 128 5 4 3 2 1 0 kCounts Ion: 152 all No.2DF-1000--4-29-2004.SMS 1.50 Acenaphthylene Ion: 152 1.25 1.00 0.75 0.50 0.25 0.00 kCounts Ion: 154 all No.2DF-1000--4-29-2004.SMS 1.00 Ion: 154

0.75 Acenaphthene

0.50

0.25

0.00 kCounts Ion: 166 all No.2DF-1000--4-29-2004.SMS 1.5 Ion: 166 Fluorene 1.0

0.5

0.0 11 12 13 14 15 16 minutes 0 kCounts Ion: 178 all No.2DF-1000--4-29-2004.SMS Phenanthrene 1.25 Ion: 178 1.00 Anthracene 0.75

0.50

0.25

0.00 Counts Ion: 202 all No.2DF-1000--4-29-2004.SMS 200 Ion: 202

150 Pyrene 100

0 18 19 20 21 22 23 minutes Figure A.2 Selected ion chromatogram (SIC) for PAHs in diesel fuel (time in minutes).

180 000 kCounts Ions: 141+142+115 all No.2DF-1000--4-29-2004 2-MN Ions: 141+142+115 40 1-MN

0 kCounts Ions: 156+141+115 all No.2DF-1000--4-29-2004

C2-N 4 Ions: 156+141+115

30 2, 3 5

20 6, 7 1

10 8 9

0 kCounts Ions: 170+155+128 all No.2DF-1000--4-29-2004 C3-N Ions: 170+155+128

0 12 13 14 15 16 17 18 19 minutes kCounts Ions: 192+191+165 all No.2DF-1000--4-29-2004 2.5 Ions: 192+191+165 2.0

1.5 C1-178

1.0

0.5

kCounts Ions: 206+191+189 all No.2DF-1000--4-29-2004 Ions: 206+191+189 2.0 C2-178

1.5

1.0

0.5

19.5 20.0 20.5 21.0 21.5 22.0 minutes

Figure A.3 Selected ion chromatogram (SIC) for alkylated PAHs in diesel fuel (time in minutes).

181 0 kCounts Ions: 91+92 all No.2DF-1000--4-29-2004.SMS Toluene Ions: 91+92 12.5 10.0

7.5

5.0 2.5

0.0 kCounts Ions: 106+91 all No.2DF-1000--4-29-2004.SMS 20 2 Ions: 106+91 3 C2-B 15

10 1

0 kCounts Ions: 120+105+91 all No.2DF-1000--4-29-2004.S 50 7 Ions: 120+105+91 40 3 C -B 30 3 4,5,6 8 20 2 10 1

0 kCounts Ions: 134+119+91 all No.2DF-1000--4-29-2004.S Ions: 134+119+91 8-10 12 25 6 11 20 C4-B 15 2 5 10 3 4 7 5 1 0 456789 minutes 0 kCounts Ions: 148+133+105 all No.2DF-1000--4-29-2004 15.0 Ions: 148+133+105 12.5 C5-B 10.0 7.5 5.0 2.5 0.0 kCounts Ions: 162+147+119 Ions: 162+147+119 all No.2DF-1000--4-29-2004 15.0 12.5 C6-B 10.0 7.5 5.0 2.5 0.0 9.5 10.0 10.5 11.0 11.5 12.0 12.5 minutes

Figure A.4 Selected ion chromatogram (SIC) for alkylbenzenes in diesel fuel (time in minutes).

182 Counts Ion: 128 all 0kW-74.18mg4-4-2004.SMS 250 Ion: 128 200 Naphthalene

150

100

0 9.0 9.5 10.0 10.5 11.0 11.5 12.0 12.5 minutes 0 kCounts Ion: 152 all 0kW-74.18mg4-4-2004.SMS Acenaphthylene Ion: 152 2.0

1.5

1.0

0.5

0.0 kCounts Acenaphthene Ion: 154 all 0kW-74.18mg4-4-2004.SMS Ion: 154 1.5

1.0

0.5

0.0 kCounts Ion: 166 all 0kW-74.18mg4-4-2004.SMS Ion: 166 3 Fluorene

0 14.5 15.0 15.5 16.0 16.5 17.0 minutes

kCounts Ion: 178 all 0kW-74.18mg4-4-2004.SMS 30 Phenanthrene Ion: 178 25 Anthracene 20 15 10 5 0 18.0 18.5 19.0 19.5 20.0 20.5 minutes

kCounts Ion: 202 all dpm_zifei.sms 6 Pyrene Ion: 202 5 4 Fluoranthene 3 2 1 0 21.5 22.0 22.5 23.0 23.5 24.0 minutes

Figure A.5 Selected ion chromatogram (SIC) for PAHs in DPM (time in minutes).

183 0 kCounts Ions: 141+142+115 all 0kW-74.18mg4-4-2004.S 2-MN 12.5 Ions: 141+142+115 10.0 1-MN

7.5

5.0

2.5

0.0 kCounts Ions: 156+141+115 all 0kW-74.18mg4-4-2004.S 4 20 C2-N 5 Ions: 156+141+115

15 2,3 6,7

10 1 8 9 5

0 kCounts Ions: 170+155+128 all 0kW-74.18mg4-4-2004.S 25 Ions: 170+155+128 20

15 C3-N

0 13 14 15 16 17 minutes 0 kCounts Ions: 192+191+165 all 0kW-74.18mg4-4-2004.S

12.5 Ions: 192+191+165

10.0 C1-178

7.5

5.0

2.5

0.0 kCounts Ions: 206+191+189 all 0kW-74.18mg4-4-2004.S Ions: 206+191+189 7.5 C2-178

5.0

2.5

0.0 20 21 22 23 24 minutes

Figure A.6 Selected ion chromatogram (SIC) for alkylated PAHs in DPM (time in minutes).

184 0 kCounts Ions: 91+92 all 75kW-50.66mg4-5-2004.SMS

3 Toluene Ions: 91+92

0 kCounts Ions: 106+91 all 75kW-50.66mg4-5-2004.SMS

6 C2-B 2 Ions: 106+91 5 3 4 3 1 2 1 0 3.54.04.55.05.56.06.5 minutes 0 kCounts Ions: 120+105+91 all 75kW-50.66mg4-5-2004.S C3-B 60 7 Ions: 120+105+91 50 40 8 30 1 2 3,4,5 6 20 10 0 6.57.07.58.08.59.09.5 minutes 0 kCounts Ions: 134+119+91 all 75kW-50.66mg4-5-2004.S C4-B 8,9,10 13 40 6 11 Ions: 134+119+91

30 12 20 2 3 4 5 7 10 1 0 8.0 8.5 9.0 9.5 10.0 10.5 11.0 minutes 0 kCounts Ions: 148+133+105 all 75kW-50.66mg4-5-2004. 20 Ions: 148+133+105 C5-B 15

0 kCounts Ions: 162+147+119 all 75kW-50.66mg4-5-2004. Ions: 162+147+119 20 C6-B 15

0 9.5 10.0 10.5 11.0 11.5 12.0 12.5 minutes

Figure A.6 Selected ion chromatogram (SIC) for alkylbenzenes in DPM (time in minutes).

185

Appendix B

DATA FOR DIESEL FUEL COMPONENT

Table B.1 Organic compounds present in low and high sulfur diesel fuela, in µg/g DF. Compound LSDFb HSDFc Compound LSDF HSDF n-Alkanes n-Decane (C10) 12115 4388 n-Octadecane (C18) 8727 11288 n-Undecane (C11) 11271 13163 n-Nonadecane (C19) 4988 6038 n-Dodecane (C12) 17149 25688 n-Eicosane (C20) 2193 1425 n-Tridecane (C13) 28834 32888 n-Heneicosane (C21) 1092 1088 n-Tetradecane (C14) 25604 36375 n-Docosane (C22) 756 450 n-Pentadecane (C15) 27660 36413 n-Tricosane (C23) 220 338 n-Hexadecane (C16) 23965 37913 n-Tetracosane (C24) 107 225 d n-Heptadecane (C17) 26082 27863 n-Pentacosane (C25) ND ND Sum of n-alkanes 190763 235650

Branched alkanes

Norfarnesane (C14) 11469 9936 Pristine (C19) 5871 4045

Farnesane (C15) 9719 7639 Phytane (C20) 4775 3203

Norpristane (C18) 7992 6939 Other branched alkanes 328578 287624 Sum of branched alkanes 368404 319386

Cycloalkanes n-Heptylcyclohexane (C13) 13144 12837 n-Tetradecylcyclohexane (C20) 25 225 n-Octylcyclohexane (C14) 11467 10025 n-Pentadecylcyclohexane (C21) ND ND n-Nonylcyclohexane (C15) 10582 8885 n-Hexadecylcyclohexane (C22) ND ND n-Decylcyclohexane (C16) 9135 7450 n-Heptadecylcyclohexane (C23) ND ND n-Undecylcyclohexane (C17) 6207 5480 n-Octadecylcyclohexane (C24) ND ND n-Dodecylcyclohexane (C18) 2073 3471 n-Nonadecylcyclohexane (C25) ND ND n-Tridecylcyclohexane (C19) 165 337 Sum of saturated cycloalkanes 52798 48710

PAHs Naphthalene (Nap) 753 1425 Benzo(a)anthracene (Baa) ND ND Acenaphthylene (Acy) 159 250 Chrysene (Chy) ND ND Acenaphthene (Ace) 85 287 Benzo(b)fluoranthene (Bbf) ND ND Fluorene (Flu) 100 525 Benzo(k)fluoranthene (Bkf) ND ND

186 Compound LSDFb HSDFc Compound LSDF HSDF Phenanthrene (Phe) 247 425 Benzo(a)pyrene (Bap) ND ND Anthracene (Ant) 7.5 50 Indeno[1,2,3-cd]pyrene (Ind) ND ND Fluoranthene (Flt) ND 25 Dibenz(a,h)anthracene (Dba) ND ND Pyrene (Pyr) 5.0 50 Benzo(ghi)perylene (Bgp) ND ND Biphenyl 437 1079 Sum of PAHs 1793 4116

Alkylated PAHs 1-MN 585 3475 1,8-DMN ND 0 2-MN 2291 5650 2,3-DMN ND 1225 1,2-DMN 373 650 2,6-DMN 1224 4425 1,3-DMN ND 3425 2,7-DMN 1837 4725 1,4-DMN 1540 925 TMN 12327 13025 1,5-DMN ND 125 1-MPh 242 893 1,6-DMN 1807 2250 2-MPh 528 3280 1,7-DMN 2548 2725 DMPh ND 2675 Sum of alkylated PAHs 25302 49473

Alkylbenzenes

Toluene (C1-B) 1377 2059 C4-Benzenes (C4-B) 9724 10021

C2-Benzenes (C2-B) 12932 13258 C5-Benzenes (C5-B) 5538 7645

C3-Benzenes (C3-B) 10003 10689 C6-Benzenes (C6-B) 5222 5790 Sum of alkylbenzenes 44796 49462 a All of the compounds detected were quantified with standard compounds. b Low sulfur diesel fuel with sulfur content 400 ppmv. c High sulfur diesel fuel with sulfur content 2200 ppmv. d ND: not detected, i.e. the concentration is lower than detection limit or the compound is not present in the sample.

187 Appendix C

DATA FOR DIESEL EMISSIONS

C.1 Data for Diesel Emissions

Table C.1 Parameters for the non-road diesel generator. Engine load Fuel consumption rate Stack gas flow rate Air/fuel ratio (mass/mass (kW) (L/hr) (m3/hr) basis) 0 5.0 263 85 25 10.0 286 46 50 15.0 341 37 75 19.5 382 32

Table C.2 DPM mass, OC fraction, and EC fraction in diesel emissions. Concentration (mg/m3) Fraction in DPM Engine Organic compounds Non-carbonaceous load DPM OC EC EC (1.2×OC) materials High volume dilution method, low sulfur diesel fuel 0 kW 5.01±0.58 3.64±0.49 0.63±0.09 87.1% 12.6% 0.3% 25 kW 7.35±1.13 4.43±0.65 1.55±0.26 72.2% 21.0% 6.7% 50 kW 10.79±1.57 5.38±0.78 3.77±0.54 59.8% 34.9% 5.2% 75 kW 16.88±1.59 6.42±0.91 8.42±0.88 45.7% 49.9% 4.4%

High volume dilution method, high sulfur diesel fuel 0 kW 6.05 4.56 0.54 90.5% 9.0% 0.5% 25 kW 9.35 5.94 1.41 76.2% 15.1% 8.8% 50 kW 13.29 6.99 3.69 63.1% 27.8% 9.1% 75 kW 20.08 8.27 8.04 49.4% 40.1% 10.5%

EPA Method 5, low sulfur diesel fuel 0 kW 2.74 1.41 0.58 61.7% 21.2% 17.2% 25 kW 5.08 2.19 1.73 51.8% 34.1% 14.2% 50 kW 7.48 2.45 3.90 39.3% 52.1% 8.6% 75 kW 9.38 0.71 7.89 9.2% 84.1% 6.7%

188 Table C.3 Composition of DPM collected with EPA Method 5, in µg/g DPM. Compound 0 kW 75 kW Compound 0 kW 75 kW n-Alkanes n-Decane (C10) 5.3 1.1 n-Octadecane (C18) 375 20 n-Undecane (C11) 16 1.5 n-Nonadecane (C19) 427 23 n-Dodecane (C12) 21 1.6 n-Eicosane (C20) 384 28 n-Tridecane (C13) 26 3.2 n-Heneicosane (C21) 276 25 n-Tetradecane (C14) 48 5.2 n-Docosane (C22) 162 15 n-Pentadecane (C15) 79 7.2 n-Tricosane (C23) 107 9.9 n-Hexadecane (C16) 159 11 n-Tetracosane (C24) 75 6.5 n-Heptadecane (C17) 219 14 n-Pentacosane (C25) 64 5.7 Sum of n-alkanes 2444 176

Branched alkanes

Norfarnesane (C14) 7.6 3.0 Pristine (C19) 123 86

Farnesane (C15) 16 6.0 Phytane (C20) 152 95

Norpristane (C18) 104 68 Other branched alkanes 923 326 Sum of branched alkanes 1325 584

Cycloalkanes n-Heptylcyclohexane (C13) 2.8 0.57 n-Tetradecylcyclohexane (C20) 17 7.0 n-Octylcyclohexane (C14) 4.4 0.80 n-Pentadecylcyclohexane (C21) 13 4.0 n-Nonylcyclohexane (C15) 9.1 1.6 n-Hexadecylcyclohexane (C22) 11 3.0 n-Decylcyclohexane (C16) 12 3.0 n-Heptadecylcyclohexane (C23) 8.5 2.6 n-Undecylcyclohexane (C17) 21 4.0 n-Octadecylcyclohexane (C24) 7.2 1.9 n-Dodecylcyclohexane (C18) 32 5.8 n-Nonadecylcyclohexane (C25) 5.5 1.6 n-Tridecylcyclohexane (C19) 24 7.9 Sum of saturated cycloalkanes 168 44

PAHs Naphthalene (Nap) 1.5 0.60 Benzo(a)anthracene (Baa) 7.8 3.8 Acenaphthylene (Acy) 1.2 0.20 Chrysene (Chy) 7.8 4.3 Acenaphthene (Ace) 1.3 0.20 Benzo(b)fluoranthene (Bbf) 3.0 1.4 Fluorene (Flu) 4.4 0.40 Benzo(k)fluoranthene (Bkf) 2.7 1.2 Phenanthrene (Phe) 23 4.9 Benzo(a)pyrene (Bap) 4.9 2.1 Anthracene (Ant) 1.8 0.40 Indeno[1,2,3-cd]pyrene (Ind) 3.8 1.7 Fluoranthene (Flt) 16 5.0 Dibenz(a,h)anthracene (Dba) 1.8 0.94 Pyrene (Pyr) 28 9.7 Benzo(ghi)perylene (Bgp) 2.8 1.8 Sum of PAHs 112 39

Alkylated PAHs 1-MN 8.2 1.9 1,8-DMN NDa ND

189 Compound 0 kW 75 kW Compound 0 kW 75 kW 2-MN 25 4.1 2,3-DMN 1.3 ND 1,2-DMN 7.3 0.74 2,6-DMN 11 2.6 1,3-DMN 46 12 2,7-DMN 7.9 ND 1,4-DMN 5.3 ND TMN 757 34 1,5-DMN 30 6.3 1-MPh 481 11 1,6-DMN 12 3.5 2-MPh 686 16 1,7-DMN 5.3 ND DMPh 1032 30 Sum of alkylated PAHs 3114 122

Alkylbenzenes

Toluene (C1-B) 6.8 4.3 C4-Benzenes (C4-B) 12 35

C2-Benzenes (C2-B) 13 8.9 C5-Benzenes (C5-B) 5.6 26

C3-Benzenes (C3-B) 11 16 C6-Benzenes (C6-B) 3.0 13 Sum of alkylbenzenes 51 104 n-Alkanoic acids

Hexanoic acid (C6) 0 1.5 Tridecanoic acid (C13) 93 11

Heptanoic acid (C7) 3.6 6.8 Tetradecanoic acid (C14) 425 22

Octanoic acid (C8) 8.1 10 Pentadecanoic acid (C15) 281 9.7

Nonanoic acid (C9) 18 20 Hexadecanoic acid (C16) 474 10

Decanoic acid (C10) 45 26 Heptadecanoic acid (C17) 25 4.2

Undecanoic acid (C11) 55 19 Octadecanoic acid (C18) 11 1.7

Dodecanoic acid (C12) 89 28 Sum of n-alkanoic acids 1528 170

Aromatic acids Benzoic acid 58 35 a ND: not detected, i.e. the concentration is lower than detection limit or the compound is not present in the sample.

190 Table C.4 Composition of DPM collected with natural dilution method and high volume dilution method, in µg/g DPM. Compound Natural dilution method High volume dilution method 0 kW 25 kW 50 kW 75 kW 0 kW 25 kW 0 kW 75 kW n-Alkanes n-Decane (C10) 9.7 14 11 3.8 116 86 42 24 n-Undecane (C11) 22 26 11 5.6 164 123 77 47 n-Dodecane (C12) 31 31 16 7.4 222 179 99 55 n-Tridecane (C13) 57 47 21 12 337 273 123 79 n-Tetradecane (C14) 204 159 33 14 500 364 167 116 n-Pentadecane (C15) 565 229 51 18 1085 783 261 203 n-Hexadecane (C16) 1164 440 73 30 2206 1414 459 308 n-Heptadecane (C17) 2645 981 101 46 3189 2287 655 441 n-Octadecane (C18) 1479 865 129 67 3851 2625 896 516 n-Nonadecane (C19) 1257 724 179 92 3415 2509 948 560 n-Eicosane (C20) 742 506 133 81 2769 2293 836 528 n-Heneicosane (C21) 138 291 126 69 1921 1578 637 404 n-Docosane (C22) 10 124 76 40 1041 949 354 225 n-Tricosane (C23) 7.4 49 44 33 574 420 252 146 n-Tetracosane (C24) 5.2 27 21 17 421 320 140 113 n-Pentacosane (C25) 3.6 25 18 14 324 213 165 128 Sum of n-alkanes 8341 4537 1045 552 22137 16417 6112 3893

Branched alkanes

Norfarnesane (C14) 13 12 10 6.5 154 136 81 61

Farnesane (C15) 28 19 18 6.8 269 293 164 91

Norpristane (C18) 135 119 102 89 1065 1088 876 635

Pristine (C19) 147 134 124 107 1204 1197 940 697

Phytane (C20) 185 148 139 116 1454 1317 1067 760 Other branched alkanes 7084 2835 418 288 22894 13958 2823 2295 Sum of branched alkanes 7593 3267 812 614 27040 17988 5950 4538

Cycloalkanes n-Heptylcyclohexane (C13) 11 4.5 2.0 0.90 21 14 14 8.7 n-Octylcyclohexane (C14) 27 7.8 3.5 1.5 46 36 32 23 n-Nonylcyclohexane (C15) 29 15 7.2 2.6 101 82 72 50 n-Decylcyclohexane (C16) 43 23 14 5.0 168 122 102 70 n-Undecylcyclohexane (C17) 39 38 22 7.5 209 163 127 86 n-Dodecylcyclohexane (C18) 61 44 30 13 150 129 117 82 n-Tridecylcyclohexane (C19) 12 29 27 15 138 116 98 63 n-Tetradecylcyclohexane (C20) 9.8 18 15 9.6 78 79 64 42 n-Pentadecylcyclohexane (C21) 6.8 14 11 5.5 54 45 38 28

191 Compound Natural dilution method High volume dilution method 0 kW 25 kW 50 kW 75 kW 0 kW 25 kW 0 kW 75 kW n-Hexadecylcyclohexane (C22) 4.3 9.3 8.2 4.1 45 40 30 17 a n-Heptadecylcyclohexane (C23) ND 7.9 6.6 2.8 31 22 22 14 n-Octadecylcyclohexane (C24) ND 5.6 4.2 2.0 26 18 14 10 n-Nonadecylcyclohexane (C25) ND 5.3 3.5 1.9 13 8.1 5.7 7.1 Sum of saturated cycloalkanes 242 221 154 71 1079 876 735 500

PAHs Naphthalene (Nap) 4.1 1.6 1.3 3.2 35 22 12 7.4 Acenaphthylene (Acy) 2.0 0.8 0.7 1.6 17 11 7 4.3 Acenaphthene (Ace) 3.0 1.6 0.7 3.7 33 25 11 11 Fluorene (Flu) 7.1 3.2 2.0 5.3 32 21 16 11 Phenanthrene (Phe) 77 42 28.8 30 224 143 91 70 Anthracene (Ant) 4.1 2.4 1.3 3.3 17 12 9 5.8 Fluoranthene (Flt) 25 29 33.3 20 78 51 52 42 Pyrene (Pyr) 67 62 66.0 35 146 88 77 56 Benzo(a)anthracene (Baa) 5.1 11 13.1 8.3 21 15 16 12 Chrysene (Chy) 5.1 13 15.0 11 29 19 17 16 Benzo(b)fluoranthene (Bbf) 3.0 11 10.5 3.0 5.9 7.20 7.35 7.5 Benzo(k)fluoranthene (Bkf) 2.0 4.8 3.3 1.5 5.0 6.64 6.74 6.2 Benzo(a)pyrene (Bap) 4.1 10 8.5 5.3 11.1 12.14 11.71 10 Indeno[1,2,3-cd]pyrene (Ind) 6.6 6.3 3.9 3.7 8.2 9.64 7.63 7.4 Dibenz(a,h)anthracene (Dba) 1.7 0.0 0.0 1.8 2.9 3.01 2.45 2.5 Benzo(ghi)perylene (Bgp) 3.0 3.2 1.3 2.9 6.4 7.53 8.24 6.1 Sum of PAHs 220 202 190 139 673 454 353 275

Alkylated PAHs 1-MN 30 4.5 5.5 4.8 112 51 26 16 2-MN 86 18 18 14 228 118 54 35 1,2-DMN 31 14 5.4 2.2 145 78 21 15 1,3-DMN 235 128 69 39 562 434 210 113 1,4-DMN ND ND 53 19 274 152 101 68 1,5-DMN 134 203 12 18 326 248 84 68 1,6-DMN 61 41 21 12 327 161 83 45 1,7-DMN 25 21 ND ND 101 75 33 27 1,8-DMN ND ND ND ND 0 0 0 0 2,3-DMN ND ND ND ND 24 16 12 5 2,6-DMN 54 27 13 9 171 102 48 44 2,7-DMN 49 24 ND ND 147 81 27 11 TMN 2914 1591 256 123 5625 3230 543 422 1-MPh 1770 747 50 41 3157 1528 164 118

192 Compound Natural dilution method High volume dilution method 0 kW 25 kW 50 kW 75 kW 0 kW 25 kW 0 kW 75 kW 2-MPh 2168 949 161 35 3978 2019 289 156 DMPh 3420 1821 236 95 6268 3199 577 362 Sum of alkylated PAHs 10977 5590 899 412 21445 11490 2271 1503

Alkylbenzenes

Toluene (C1-B) 11 14 24 27 41 46 175 198

C2-Benzenes (C2-B) 21 22 37 50 72 69 322 358

C3-Benzenes (C3-B) 15 17 48 79 43 54 346 464

C4-Benzenes (C4-B) 16 15 106 148 47 43 644 826

C5-Benzenes (C5-B) 6.2 22 88 105 23 65 550 554

C6-Benzenes (C6-B) 3.2 13 42 55 11 35 227 278 Sum of alkylbenzenes 71 103 346 464 236 312 2264 2678 n-Alkanoic acids

Hexanoic acid (C6) ND ND 9.0 23 10 15 23 45

Heptanoic acid (C7) 18 23 56 62 27 52 100 134

Octanoic acid (C8) 40 60 78 94 57 127 174 168

Nonanoic acid (C9) 88 104 123 140 125 208 271 256

Decanoic acid (C10) 289 256 208 198 398 487 443 318

Undecanoic acid (C11) 354 259 179 133 478 467 294 184

Dodecanoic acid (C12) 591 406 231 183 786 691 343 217

Tridecanoic acid (C13) 444 256 130 54 582 410 172 69

Tetradecanoic acid (C14) 1955 856 361 125 2503 1285 481 170

Pentadecanoic acid (C15) 1238 421 190 50 1559 590 236 62

Hexadecanoic acid (C16) 1991 547 216 62 1899 712 295 84

Heptadecanoic acid (C17) ND 1.3 0.0 2.5 5.2 8.4 4.8 3.1

Octadecanoic acid (C18) 4.1 8.2 5.3 5.2 10 11 8.2 5.3 Sum of n-alkanoic acids 7012 3201 1785 1133 8441 5063 2845 1716

Aromatic acids Benzoic acid 226 289 376 430 305 426 469 677 a ND: not detected, i.e. the concentration is lower than detection limit or the compound is not present in the sample.

193 Table C.5 Compound concentration (ng/m3) in particulate and gas phase diesel emissions collected with high volume dilution method (LSDF). Compound Particulate phase Gas phase abbreviationa 0 kW 25 kW 50 kW 75 kW 0 kW 25 kW 0 kW 75 kW n-Alkanes

C10 414 515 352 309 136976 174113 122843 112078

C11 584 732 644 589 155926 203587 180209 166199

C12 791 1066 823 701 182801 259092 220656 194255

C13 1201 1629 1022 996 199363 294799 263693 263940

C14 1780 2173 1388 1473 195823 293366 315076 312276

C15 3863 4666 2171 2567 206106 251963 280072 310565

C16 7854 8428 3825 3891 183518 203853 231871 276279

C17 11354 13638 5455 5572 153544 183653 217305 250734

C18 13711 15651 7466 6526 86420 109513 140640 151958

C19 12161 14960 7893 7075 29933 51851 76285 101382

C20 9859 13674 6965 6678 6651 17459 33575 50164

C21 6840 9406 5308 5103 2741 3083 16703 21240

C22 3708 5661 2952 2846 627 1004 2068 2268

C23 2045 2506 2097 1852 146 846 1028 1394

C24 1498 1911 1169 1431 68 412 592 669

C25 1155 1270 1375 1617 37 272 322 372 Sum of n-alkanes 78817 97887 50906 49227 1503570 2172268 2251197 2458172

Branched alkanes

C14 549 812 672 777 108678 171370 160596 180255

C15 958 1745 1368 1145 140850 277479 250412 254205

C18 3791 6485 7295 8026 132686 237981 375362 488322

C19 4287 7136 7828 8813 63275 147633 254892 367380

C20 5177 7852 8884 9606 29501 89386 174902 243950 Other branched 81514 83223 23514 29021 1059676 1414790 752453 1102803 alkanes Sum of branched 96276 107253 49562 57389 1534665 2338639 1968617 2636915 alkanes

Cycloalkanes

C13 76 86 113 110 8263 11722 17242 17427

C14 163 216 265 291 9760 14688 21494 24177

C15 358 490 600 634 9287 15271 21184 21684

C16 599 729 847 883 7776 11260 15390 18178

C17 742 973 1058 1085 6252 9812 13635 15449

C18 533 770 974 1034 2327 4898 7440 8666

C19 491 693 814 803 1013 2235 4682 5982

194 Compound Particulate phase Gas phase abbreviationa 0 kW 25 kW 50 kW 75 kW 0 kW 25 kW 0 kW 75 kW

C20 276 473 531 528 295 909 2134 2733

C21 192 269 314 359 108 232 458 689

C22 159 238 252 212 17 157 233 324

C23 111 132 184 173 7.5 67 137 185

C24 94 106 120 121 3.2 49 86 117 b C25 46 49 47 90 ND 16 35 60 Sum of saturated 3841 5224 6120 6323 45109 71316 104150 115671 cycloalkanes

PAHs Nap 126 132 101 94 9600 19138 23086 26520 Acy 59 68 62 55 2100 3812 5160 5745 Ace 119 147 92 134 3320 6412 9955 12590 Flu 114 128 134 143 2440 4184 6386 7593 Phe 797 854 760 887 3388 7756 11798 12901 Ant 62 69 72 74 355 738 841 1234 Flt 279 307 434 532 215 382 571 816 Pyr 519 524 644 706 139 540 718 884 Baa 76 92 129 152 14 34 57 79 Chy 103 115 145 198 11 23 65 67 Bbf 21 43 61 95 ND ND 1.6 2.6 Bkf 18 40 56 78 ND ND ND ND Bap 40 72 98 127 ND ND 2.0 2.0 Ind 29 57 64 93 ND ND 2.0 1.0 Dba 10 18 20 31 ND ND 0.73 1.7 Bgp 23 45 69 77 ND ND ND 1.7 Sum of PAHs 2395 2709 2940 3475 21741 43157 58825 68677

Alkylated PAHs 1-MN 399 303 217 205 25520 27543 32574 34988 2-MN 812 704 450 438 52831 63142 65441 75169 1,2-DMN 518 465 175 190 12144 15611 18189 18485 1,3-DMN 2003 2585 1749 1428 30632 37206 43556 57594 1,4-DMN 975 903 837 854 15723 32760 39958 52836 1,5-DMN 1159 1479 700 855 25664 41285 46329 57014 1,6-DMN 1163 957 687 571 16177 25902 33338 35498 1,7-DMN 359 447 275 341 9126 13485 17665 21550 1,8-DMN ND ND ND ND ND ND ND ND 2,3-DMN 86 95 99 58 1995 2692 3470 3660 2,6-DMN 608 608 400 550 10175 13082 15849 21420

195 Compound Particulate phase Gas phase abbreviationa 0 kW 25 kW 50 kW 75 kW 0 kW 25 kW 0 kW 75 kW 2,7-DMN 523 483 229 141 11484 7212 4656 3864 TMN 20027 19259 4523 5336 27801 39899 60421 72386 1-MPh 11242 9111 1366 1492 940 1307 2176 2848 2-MPh 14162 12038 2407 1973 2444 2513 2894 3517 DMPh 22318 19073 4804 4578 1053 1526 1682 1978 Sum of alkylated 76353 68508 18917 19010 243709 325165 388199 462805 PAHs

Alkylbenzenes

C1-B 144 274 1459 2510 8333 15719 69533 95941

C2-B 255 411 2680 4533 7659 13285 120220 219847

C3-B 152 321 2885 5866 3709 8494 112611 259674

C4-B 167 257 5365 10439 2175 4978 132269 293494

C5-B 82 387 4578 6999 770 4897 62162 149636

C6-B 39 208 1894 3520 331 2316 27113 61804 Sum of 840 1858 18862 33867 22978 49689 523908 1080395 alkylbenzenes n-Alkanoic acids

C6 36 89 191 573 2189 4465 9572 30768

C7 95 308 836 1690 2803 7824 29095 49411

C8 204 757 1447 2126 4699 18172 37616 63959

C9 444 1243 2255 3239 6156 20300 46378 84557

C10 1419 2906 3694 4016 11326 29057 55405 68273

C11 1703 2785 2446 2332 8733 17002 19640 20254

C12 2800 4118 2854 2741 5916 13274 12230 12961

C13 2073 2447 1429 876 2707 6320 5554 3997

C14 8911 7660 4009 2152 5585 8037 8650 5384

C15 5552 3515 1970 788 1336 1707 2208 1148

C16 6763 4242 2456 1056 384 424 1409 743

C17 18 50 40 39 1.6 6.9 7.5 8.7

C18 36 65 68 66 1.4 3.9 7.6 8.6 Sum of n-alkanoic 30054 30185 23695 21696 51838 126592 227770 341472 acids

Aromatic acids Benzoic acid 1085 2540 3907 8561 a The abbreviation used in this table refers to the corresponding compound listed in Table C.4. b ND: not detected, i.e. the concentration is lower than detection limit or the compound is not present in the sample.

196 Table C.6 Normalized emission rates of compounds in gas and particulate phase diesel emissions collected with high volume dilution method (LSDF). Compounds mg emission/kg fuel used mg emission/g compound in fuel 0kW 25kW 50kW 75kW 0kW 25kW 50kW 75kW n-Alkanes n-Decane (C10) 9.03 6.24 3.50 2.75 0.75 0.52 0.29 0.23 n-Undecane (C11) 10.29 7.30 5.14 4.08 0.91 0.65 0.46 0.36 n-Dodecane (C12) 12.07 9.30 6.29 4.77 0.70 0.54 0.37 0.28 n-Tridecane (C13) 13.19 10.60 7.52 6.49 0.46 0.37 0.26 0.22 n-Tetradecane (C14) 12.99 10.57 8.99 7.68 0.51 0.41 0.35 0.30 n-Pentadecane (C15) 13.81 9.17 8.02 7.67 0.50 0.33 0.29 0.28 n-Hexadecane (C16) 12.58 7.59 6.70 6.86 0.53 0.32 0.28 0.29 n-Heptadecane (C17) 10.84 7.05 6.33 6.28 0.42 0.27 0.24 0.24 n-Octadecane (C18) 6.58 4.47 4.21 3.88 0.75 0.51 0.48 0.44 n-Nonadecane (C19) 2.77 2.39 2.39 2.66 0.55 0.48 0.48 0.53 n-Eicosane (C20) 1.09 1.11 1.15 1.39 0.49 0.51 0.53 0.63 n-Heneicosane (C21) 0.63 0.54 0.63 0.65 0.58 0.41 0.57 0.59 n-Docosane (C22) 0.29 0.24 0.14 0.13 0.38 0.32 0.19 0.17 n-Tricosane (C23) 0.14 0.12 0.09 0.08 0.65 0.54 0.40 0.36 n-Tetracosane (C24) 0.10 0.08 0.05 0.05 0.96 0.78 0.47 0.48 a n-Pentacosane (C25) 0.08 0.06 0.05 0.05 NA NA NA NA Total n-alkanes 104 81 65 61 0.55 0.43 0.34 0.32

Branched alkanes

Norfarnesane (C14) 7.18 6.16 4.58 4.43 0.63 0.54 0.40 0.39

Farnesane (C15) 9.32 9.98 7.15 6.25 0.96 1.03 0.74 0.64

Norpristane (C18) 8.97 8.74 10.87 12.15 1.12 1.09 1.36 1.52

Pristine (C19) 4.44 5.53 7.47 9.21 0.76 0.94 1.27 1.57

Phytane (C20) 2.28 3.48 5.22 6.21 0.48 0.73 1.09 1.30 Other branched alkanes 75 54 22 28 0.23 0.16 0.07 0.08 Total branched alkanes 107 87 57 66 0.29 0.24 0.16 0.18

Cycloalkanes n-Heptylcyclohexane (C13) 0.55 0.42 0.49 0.43 0.04 0.03 0.04 0.03 n-Octylcyclohexane (C14) 0.65 0.53 0.62 0.60 0.06 0.05 0.05 0.05 n-Nonylcyclohexane (C15) 0.63 0.56 0.62 0.55 0.06 0.05 0.06 0.05 n-Decylcyclohexane (C16) 0.55 0.43 0.46 0.47 0.06 0.05 0.05 0.05 n-Undecylcyclohexane (C17) 0.46 0.39 0.42 0.40 0.07 0.06 0.07 0.07 n-Dodecylcyclohexane (C18) 0.19 0.20 0.24 0.24 0.09 0.10 0.12 0.11 n-Tridecylcyclohexane (C19) 0.10 0.10 0.16 0.17 0.60 0.63 0.95 1.01 n-Tetradecylcyclohexane (C20) 0.04 0.05 0.08 0.08 1.50 1.98 3.03 3.19 n-Pentadecylcyclohexane (C21) 0.02 0.02 0.02 0.03 NA NA NA NA

197 Compounds mg emission/kg fuel used mg emission/g compound in fuel 0kW 25kW 50kW 75kW 0kW 25kW 50kW 75kW n-Hexadecylcyclohexane (C22) 0.01 0.01 0.01 0.01 NA NA NA NA n-Heptadecylcyclohexane (C23) 0.008 0.007 0.009 0.009 NA NA NA NA n-Octadecylcyclohexane (C24) 0.006 0.006 0.006 0.006 NA NA NA NA n-Nonadecylcyclohexane (C25) 0.003 0.002 0.002 0.004 NA NA NA NA Total saturated cycloalkanes 3.22 2.74 3.13 2.99 0.06 0.05 0.06 0.06

PAHs Naphthalene (Nap) 0.64 0.69 0.66 0.65 0.85 0.91 0.88 0.87 Acenaphthylene (Acy) 0.14 0.14 0.15 0.14 0.89 0.87 0.93 0.89 Acenaphthene (Ace) 0.23 0.23 0.29 0.31 2.66 2.76 3.36 3.67 Fluorene (Flu) 0.17 0.15 0.19 0.19 1.68 1.54 1.85 1.89 Phenanthrene (Phe) 0.28 0.31 0.36 0.34 1.11 1.25 1.44 1.37 Anthracene (Ant) 0.03 0.03 0.03 0.03 3.65 3.84 3.46 4.27 Fluoranthene (Flt) 0.03 0.02 0.03 0.03 NA NA NA NA Pyrene (Pyr) 0.04 0.04 0.04 0.04 8.66 7.60 7.74 7.79 Benzo(a)anthracene (Baa) 0.006 0.005 0.005 0.006 NA NA NA NA Chrysene (Chy) 0.008 0.005 0.006 0.006 NA NA NA NA Benzo(b)fluoranthene (Bbf) 0.001 0.002 0.002 0.002 NA NA NA NA Benzo(k)fluoranthene (Bkf) 0.001 0.001 0.002 0.002 NA NA NA NA Benzo(a)pyrene (Bap) 0.003 0.003 0.003 0.003 NA NA NA NA Indeno[1,2,3-cd]pyrene (Ind) 0.002 0.002 0.002 0.002 NA NA NA NA Dibenz(a,h)anthracene (Dba) 0.001 0.001 0.001 0.001 NA NA NA NA Benzo(ghi)perylene (Bgp) 0.001 0.002 0.002 0.002 NA NA NA NA Total PAHs 1.59 1.64 1.76 1.77 1.17 1.21 1.29 1.30

Alkylated PAHs Methylnaphthalenes (MN)b 5.23 3.28 2.80 2.71 1.82 1.14 0.98 0.94 Dimethylnaphthalenes (DMN)c 9.24 7.05 6.48 6.78 0.99 0.76 0.69 0.73 Trimethylnaphthalenes (TMN)d 3.14 2.11 1.85 1.90 0.26 0.17 0.15 0.15 Methylphenanthrenes (MPh)e 1.89 0.89 0.25 0.24 2.46 1.16 0.33 0.31 Dimethylphenanthrenes (DMPh)f 1.54 0.74 0.18 0.16 NA NA NA NA Total alkylated PAHs 21 14 12 12 0.83 0.56 0.46 0.47

Alkylbenzenes

Toluene (C1-B) 0.56 0.57 2.02 2.41 0.40 0.42 1.47 1.75

C2-Benzenes (C2-B) 0.52 0.49 3.49 5.49 0.04 0.04 0.27 0.42

C3-Benzenes (C3-B) 0.25 0.32 3.28 6.50 0.03 0.03 0.33 0.65

C4-Benzenes (C4-B) 0.15 0.19 3.91 7.44 0.02 0.02 0.40 0.77

C5-Benzenes (C5-B) 0.06 0.19 1.90 3.84 0.01 0.03 0.34 0.69

C6-Benzenes (C6-B) 0.02 0.09 0.82 1.60 0.00 0.02 0.16 0.31

198 Compounds mg emission/kg fuel used mg emission/g compound in fuel 0kW 25kW 50kW 75kW 0kW 25kW 50kW 75kW Total alkylbenzenes 1.57 1.84 15 27 0.03 0.04 0.34 0.61

Total unburned hydrocarbons 239 189 155 171 a Not available since these compounds are not present in diesel fuels. b Methylnaphthalenes include 1- and 2-methylnaphthalene. c Dimethylnaphthalenes include all isomer that have been identified as dimethyl- or ethyl- naphthalene based on NIST library search or standard compound. d Trimethylnaphthalenes include all isomer that have been identified as trimethyl- or ethylmethyl- naphthalene based on NIST library search or standard compound. e Methylphenanthrenes include all isomer that have been identified as methylphenanthrene or methylanthracene based on NIST library search or standard compound. f Dimethylphenanthrenes include all isomer that have been identified as dimethyl- or ethyl- phenanthrene or anthracene based on NIST library search or standard compound.

199 Table C.7 Compound concentration (µg/g DPM) in DPM collected with high volume dilution method (HSDF). Compound 0kW 25kW 50kW 75kW Compound 0kW 25kW 50kW 75kW abbreviation abbreviation n-Alkanes

C10 6.9 7.7 6.1 12 C18 4635 4619 2460 1511

C11 28 17 13 12 C19 4229 4953 3423 1681

C12 54 18 10 8 C20 4557 5457 3014 1483

C13 234 157 54 29 C21 3516 4239 3028 1411

C14 633 442 215 126 C22 1660 2311 1213 882

C15 1194 1089 624 429 C23 1014 1476 1002 565

C16 3193 2160 2006 1268 C24 725 997 666 351

C17 4173 3233 1415 892 C25 492 1305 568 390 Sum of n-alkanes 30345 32481 19718 11048

Branched alkanes

C14 75 131 76 119 C19 2321 2232 1402 1049

C15 133 172 61 38 C20 2733 3091 1587 1073

C18 1720 1993 1134 863 Other branched 37576 20506 11950 4410 alkanes Sum of branched 44558 28125 16209 7553 alkanes

Cycloalkanes

C13 71 33 26 16 C20 783 1357 1326 988

C14 109 99 68 48 C21 879 1010 336 207

C15 237 289 133 112 C22 445 453 210 148

C16 630 694 323 161 C23 466 514 201 127

C17 965 1191 587 329 C24 323 432 135 76

C18 1396 1589 949 593 C25 272 319 104 68

C19 1392 1303 572 421 Sum of saturated 7967 9283 4969 3294 cycloalkanes

PAHs Nap 35 33 24 28 Baa 184 86 27 21 Acy 39 11 6.4 4.3 Chy 130 87 30 19 Ace 82 26 10 5.4 Bbf 60 30 27 11 Flu 90 79 25 4.0 Bkf 43 26 10 8.2 Phe 553 549 154 129 Bap 24 18 7.4 2.6 Ant 46 38 22 5.2 Ind 28 26 18 10 Flt 225 178 71 20 Dba 10 14 3.6 3.4

200 Compound 0kW 25kW 50kW 75kW Compound 0kW 25kW 50kW 75kW abbreviation abbreviation Pyr 428 308 87 59 Bgp 58 29 8.2 8.2 Sum of PAHs 2034 1538 532 339

Alkylated PAHs 1-MN 89 29 16 19 1,8-DMN 0 0 0 0 2-MN 255 100 51 48 2,3-DMN 21 20 11 4 1,2-DMN 151 97 19 12 2,6-DMN 155 104 38 36 1,3-DMN 548 579 185 141 2,7-DMN 160 84 22 11 1,4-DMN 146 178 86 50 TMN 8385 5557 695 585 1,5-DMN 347 713 78 99 1-MPh 4481 2369 202 133 1,6-DMN 165 212 70 67 2-MPh 5789 2999 360 202 1,7-DMN 111 95 30 24 DMPh 11606 4738 666 423 Sum of alkylated 32410 17875 2528 1854 PAHs

Alkylbenzenes

C1-B 80 107 82 70 C4-B 83 94 321 342

C2-B 137 159 116 166 C5-B 90 119 255 284

C3-B 72 117 131 196 C6-B 47 74 107 103 Sum of 509 671 1011 1161 alkylbenzenes n-Alkanoic acids

C6 12 20 29 60 C13 857 633 293 120

C7 36 66 136 216 C14 3754 1970 826 301

C8 77 167 255 306 C15 2527 866 377 111

C9 188 276 439 388 C16 2969 1225 433 149

C10 494 620 643 471 C17 7.7 15 9.2 6.0

C11 692 647 482 320 C18 15 19 14 7.8

C12 1322 1098 645 382 Sum of n-alkanoic 12950 7623 4583 2838 acids

Aromatic acids Benzoic acid 448 592 680 896 a The abbreviation used in this table refers to the corresponding compound listed in Table C.4. b ND: not detected, i.e. the concentration is lower than detection limit or the compound is not present in the sample.

201 C.2 Conversion of Compound Emissions between Different Units

The concentration of compounds in DPM is expressed in terms of µg/g DPM and ng/m3.

The conversion between the two units is as follows:

Compound concentration (ng/m3) = compound concentration (µg/g DPM) × mass of

DPM collected (mg) / volume of stack gas collected (m3) (C.1)

The following data are used to convert the concentration of compounds in DPM between the two units.

Table C.7 Sampling record for DPM collected with high volume dilution method (LSDF). Engine load (kW) Mass of DPM collected (mg) Volume of stack gas sampled (m3) 0 53.82 15.12 25 72.80 12.21 50 103.84 12.47 75 124.53 9.85

For the emission rate of compounds in total diesel emissions (including gas and particulate phase), two expressions are used: mg compound emission/kg of fuel combusted and mg compound emission/g compound in diesel fuel. The following two equations are used to calculate the two normalized emission rates.

CS × QS ER1 = (C.2) RDF × ρ DF

ER1 ER2 = (C.3) CDF

Where:

ER1 = mg compound emission/kg of fuel combusted, mg/kg DF

3 CS = compound concentration in diesel emissions (gas + particulate phase), µg/m

3 QS = stack gas flow rate, m /hr, listed in Table C.1

202 RDF = fuel usage rate, L/hr, listed in Table C.1

ρDF = diesel fuel density, 800 g/L

ER2 = mg compound emission/g compound in diesel fuel, mg/g

CDF = compound concentration in diesel fuel, g/kg DF

203 Appendix D

PASH DATA FOR DIESEL FUEL AND DIESEL EMISSIONS

Table D.1 PASHs in low and high sulfur diesel fuels (LSDF and HSDF). Compounda Formula Sulfur contentb (µg S/g DF) ± std (δ; n = 3) LSDF HSDF Total organic sulfurc S 433±9 2284±12 T to 2-PT (7 compounds)d NDe ND

BT C8H6S 1.34±0.05 8.21±0.17

2-MBT C9H8S 0.11±0.02 7.68±0.14 f 3-MBT C9H8S 3.21±0.09 8.39±0.12 g 5-MBT C9H8S 2.86±0.13 15.29±0.45

3,5-DMBT C10H10S 3.85±0.16 18.71±0.42

2,3,5-TMBT C11H12S 8.28±0.24 63.56±2.46

2,3,7-TMBT C11H12S 10.72±0.19 75.91±2.78

2,3,4,7-TTMBT C12H14S 4.32±0.11 24.36±0.56

DBT C12H8S 15.23±0.29 83.99±1.92

4-MDBT C13H10S 21.22±0.52 66.78±1.64

4,6-DMDBT C14H12S 10.95±0.31 20.81±0.35

2,4,6-TMDBT C15H14S 4.34±0.15 6.63±0.24 h Ph45T C14H8S 2.08±0.13 3.03±0.12 i BN12T to DiAT (6 compounds) C16H10S ND ND a The compound names for the abbreviations are as listed in Glossary of acronyms and symbols. b For both fuels (LSDF and HSDF), three sets of samples were injected. c Total sulfur was calculated by integrating the entire sulfur response as one peak. Integration began where the sulfur emission line increased from baseline or at the beginning of the first peak (7 min for diesel fuel). The final integration point was designated after the sulfur emission returned to baseline and remained stable (25 min for diesel fuel). The standard temperature program indicated in Table 6-1 was used. d The 7 compounds include T, 2-MT, 3-MT, 2,3-DMT, 2,5-DMT, 2-ET, and 2-PT. e ND (not detected) indicates result below the minimum detectable level (MDL). MDLs are based on results for a test mixture (Agilent part 8500-5067). The specification for sulfur 181 is 2 pg/sec. The MDL for our laboratory was typically about 0.5 pg/sec, or 1 pg S for a peak having a 2-sec width (at half height). f The coelution between 3- and 4-MBT may occur. g The coelution between 5- and 6-MBT may occur. h This peak is assigned as Ph45T based on the retention time of a standard compound. It is possible that this peak also contains a C3-DBT because C3-

DBTs can also elute in this region. Currently, we do not have a C3-DBT standard to examine this possibility. i The 6 compounds include BN12T, BN21T, BN23T, Ph34T, BPh9,10T, and DiAT.

204 Table D.2 PASHs in gas and particle phase diesel emissions (0 kW through 75 kW) when burning low and high sulfur diesel fuels. Compounda Formula Sulfur content (µg S/g DPM) ± std (δ; n = 3)b Gas phase emission at LSDF L0-Gc L25-G L50-G L75-G Total organic sulfurd S 775±35 786±27 450±26 400±30 e T C4H4S ND ND ND ND

2-MT C5H6S 5.56±0.62 10.15±0.89 3.42±0.33 5.41±0.39

3-MT C5H6S 9.90±0.78 19.01±1.25 6.10±0.52 10.37±0.89

2,3-DMT C6H8S 3.86±0.45 5.32±0.63 2.09±0.29 4.03±0.44

2,5-DMT C6H8S 4.17±0.54 6.68±0.56 2.45±0.28 4.59±0.38

2-ET C6H8S 4.03±0.49 6.82±0.61 2.72±0.32 4.13±0.26

2-PT C7H10S 1.73±0.28 4.35±0.52 1.54±0.27 2.76±0.37

BT C8H6S 9.56±0.82 7.76±0.67 3.96±0.36 4.18±0.43

2-MBT C9H8S 9.52±0.71 5.11±0.43 3.67±0.41 4.41±0.51 f 3-MBT C9H8S 12.20±0.88 9.22±0.79 5.88±0.51 5.87±0.55 g 5-MBT C9H8S 4.90±0.52 2.24±0.33 2.06±0.28 2.14±0.30

3,5-DMBT C10H10S 9.11±0.63 9.34±0.86 6.49±0.56 5.25±0.49

2,3,5-TMBT C11H12S 23.31±1.57 17.87±1.49 12.59±0.97 10.92±0.93

2,3,7-TMBT C11H12S 31.36±2.05 23.86±1.84 16.96±1.27 14.92±1.22

2,3,4,7-TTMBT C12H14S 2.28±0.33 1.85±0.26 1.38±0.18 0.85±0.17

DBT C12H8S 22.82±1.55 13.00±1.06 10.76±0.93 7.68±0.68

4-MDBT C13H10S 9.86±0.81 5.29±0.55 5.91±0.56 5.31±0.47

4,6-DMDBT C14H12S 3.59±0.42 ND ND ND 8 compoundsi ND ND ND ND

Particle phase emission at LSDF L0-P L25-P L50-P L75-P Total organic sulfur S 679±32 734±27 819±25 631±35 14 compoundsj ND ND ND ND

2,3,4,7-TTMBT C12H14S 1.48±0.15 1.96±0.28 1.53±0.25 1.63±0.22

DBT C12H8S 2.84±0.23 2.68±0.32 1.77±1.89 1.51±1.75

4-MDBT C13H10S 9.07±0.68 7.33±0.88 4.71±0.55 3.55±0.41

4,6-DMDBT C14H12S 10.30±1.08 12.40±1.03 11.57±1.21 11.34±1.09

2,4,6-TMDBT C15H14S 13.37±1.22 34.50±2.94 27.69±2.49 22.38±2.23 h Ph45T C14H8S 6.42±0.75 22.39±2.19 14.86±1.78 11.68±0.94

BN12T C16H10S 0.99±0.37 2.37±0.36 4.09±0.83 5.58±0.63

BN21T C16H10S 1.13±0.25 2.87±0.43 3.83±0.75 5.11±0.82

BN23T C16H10S 0.95±0.22 1.74±0.31 2.23±0.41 4.69±0.76

Ph34T C16H10S ND ND ND 2.25±0.50

BPh9,10T C20H12S ND ND ND ND

DiAT C24H12S ND ND ND ND

205 Table D.2 (Continued) Compound Formula Sulfur content (µg S/g DPM) ± std (δ; n = 3) Particle phase emission at HSDF H0-P H25-P H50-P H75-P Total organic sulfur S 1585 1599 1349 923 12 compoundsk ND ND ND ND

2,3,5-TMBT C11H12S 2.81 1.93 0.98 1.85

2,3,7-TMBT C11H12S 3.44 3.40 2.00 1.25

2,3,4,7-TTMBT C12H14S 14.28 12.67 8.52 2.49

DBT C12H8S 30.29 25.18 11.21 4.41

4-MDBT C13H10S 32.39 26.86 19.81 11.39

4,6-DMDBT C14H12S 22.91 29.28 15.31 8.25

2,4,6-TMDBT C15H14S 28.08 42.40 27.20 18.56

Ph45T C14H8S 19.05 34.78 17.80 9.55

BN12T C16H10S 1.89 3.22 5.54 7.68

BN21T C16H10S 1.69 3.06 4.61 6.86

BN23T C16H10S 1.50 2.98 4.33 6.49

Ph34T C16H10S ND 0.47 2.48 0.48

BPh9,10T C20H12S ND ND ND 0.81

DiAT C24H12S ND ND ND ND

Total organic sulfur in gas phase at HSDF 1768 1654 1282 1077 a The compound names for the abbreviations are as listed in Glossary of acronyms and symbols. b For DPM at LSDF, three sets of samples were collected and injected, but for DPM at HSDF only one set of samples was collected. c L or H indicates low or high sulfur diesel fuel was used; 0 through 75 is the engine load (kW); G or P indicates gas phase or particle phase. d Total sulfur was calculated by integrating the entire sulfur response as one peak. Integration began where the sulfur emission line increased from baseline or at the beginning of the first peak (7 min for DPM). The final integration point was designated after the sulfur emission returned to baseline and remained stable (32 min for DPM). Notes e-h are the same as those for Table D.1. i The 8 compounds include 2,4,6-TMDBT, Ph45T, BN12T, BN21T, BN23T, Ph34T, BPh9,10T, and DiAT. j The 14 compounds include T, 2-MT, 3-MT, 2,3-DMT, 2,5-DMT, 2-ET, 2-PT, BT, 2-MBT, 3-MBT, 5-MBT, 3,5-DMBT, 2,3,5-TMBT, and 2,3,7-TMBT. k The 12 compounds include T, 2-MT, 3-MT, 2,3-DMT, 2,5-DMT, 2-ET, 2-PT, BT, 2-MBT, 3-MBT, 5-MBT, and 3,5-DMBT.

206 Appendix E

DATA FOR PHENANTHRENE-ENRICHED FUEL

Table E.1 Composition of DPM collected when burning phenanthrene-enriched diesel fuela (high volume dilution sampling, 0 kW). Compound µg/g ng/m3 Compound µg/g ng/m3 DPM DPM n-Alkanes n-Decane (C10) 112 331 n-Octadecane (C18) 4740 13988 n-Undecane (C11) 195 574 n-Nonadecane (C19) 4344 12819 n-Dodecane (C12) 226 667 n-Eicosane (C20) 4293 12670 n-Tridecane (C13) 423 1248 n-Heneicosane (C21) 2826 8339 n-Tetradecane (C14) 554 1635 n-Docosane (C22) 1289 3803 n-Pentadecane (C15) 1006 2969 n-Tricosane (C23) 668 1971 n-Hexadecane (C16) 2680 7909 n-Tetracosane (C24) 432 1274 n-Heptadecane (C17) 3639 10739 n-Pentacosane (C25) 363 1070 Sum of n-alkanes 27788 82005

Branched alkanes

Norfarnesane (C14) 172 507 Pristine (C19) 1394 4114

Farnesane (C15) 301 889 Phytane (C20) 1771 5228

Norpristane (C18) 1195 3527 Other branched alkanes 28487 84070 Sum of branched alkanes 33321 98334

Cycloalkanes n-Heptylcyclohexane (C13) 21 61 n-Tetradecylcyclohexane (C20) 100 294 n-Octylcyclohexane (C14) 50 149 n-Pentadecylcyclohexane (C21) 69 203 n-Nonylcyclohexane (C15) 133 392 n-Hexadecylcyclohexane (C22) 49 145 n-Decylcyclohexane (C16) 236 696 n-Heptadecylcyclohexane (C23) 36 105 n-Undecylcyclohexane (C17) 214 633 n-Octadecylcyclohexane (C24) 27 79 n-Dodecylcyclohexane (C18) 217 642 n-Nonadecylcyclohexane (C25) 16 49 n-Tridecylcyclohexane (C19) 193 569 Sum of saturated cycloalkanes 1361 4016

PAHs Naphthalene (Nap) 88 260 Benzo(a)anthracene (Baa) 37 110 Acenaphthylene (Acy) 47 140 Chrysene (Chy) 53 157

207 Compound µg/g ng/m3 Compound µg/g ng/m3 DPM DPM Acenaphthene (Ace) 93 274 Benzo(b)fluoranthene (Bbf) 11 31 Fluorene (Flu) 84 248 Benzo(k)fluoranthene (Bkf) 9 26 Phenanthrene (Phe) 7527 22213 Benzo(a)pyrene (Bap) 17 50 Anthracene (Ant) 44 131 Indeno[1,2,3-cd]pyrene (Ind) 14 43 Fluoranthene (Flt) 155 458 Dibenz(a,h)anthracene (Dba) 4.8 14 Pyrene (Pyr) 294 868 Benzo(ghi)perylene (Bgp) 10 28 Sum of PAHs 8488 25049

Alkylated PAHs 1-MN 315 929 1,8-DMN ND 0 2-MN 665 1963 2,3-DMN 52 155 1,2-DMN 397 1171 2,6-DMN 415 1226 1,3-DMN 1339 3951 2,7-DMN 323 953 1,4-DMN 662 1955 TMN 10759 31751 1,5-DMN 814 2402 1-MPh 12058 35584 1,6-DMN 821 2424 2-MPh 15694 46315 1,7-DMN 245 724 DMPh 23249 68612 Sum of alkylated PAHs 67810 200114

Alkylbenzenes

Toluene (C1-B) 90 265 C4-Benzenes (C4-B) 97 285

C2-Benzenes (C2-B) 166 491 C5-Benzenes (C5-B) 58 171

C3-Benzenes (C3-B) 88 260 C6-Benzenes (C6-B) 26 76 Sum of alkylbenzenes 524 1547 n-Alkanoic acids

Hexanoic acid (C6) 36 106 Tridecanoic acid (C13) 715 2111

Heptanoic acid (C7) 73 217 Tetradecanoic acid (C14) 2837 8373

Octanoic acid (C8) 157 464 Pentadecanoic acid (C15) 1959 5782

Nonanoic acid (C9) 508 1500 Hexadecanoic acid (C16) 2262 6676

Decanoic acid (C10) 533 1574 Heptadecanoic acid (C17) 7.1 21

Undecanoic acid (C11) 968 2857 Octadecanoic acid (C18) 13 37

Dodecanoic acid (C12) 36 106 Sum of n-alkanoic acids 10070 29718

Aromatic acids Benzoic acid 363 1071 a Low sulfur diesel fuel with 198 mg/L phenanthrene was used as the normal fuel. After phenanthrene was added, its concentration in Ph-enriched fuel was 10000 mg/L. b ND: not detected, i.e. the concentration is lower than detection limit or the compound is not present in the sample.

208