University of Southern California Dissertations and Theses

Particulate matter (PM) exposure for commuters in Los Angeles: chemical characterization and implications to public health

(USC Thesis Other)

Particulate matter (PM) exposure for commuters in Los Angeles: chemical characterization and implications to public health

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

PARTICULATE MATTER (PM) EXPOSURE FOR
COMMUTERS IN LOS ANGELES: CHEMICAL
CHARACTERIZATION AND IMPLICATIONS TO
PUBLIC HEALTH

by

Winnie Kam

A dissertation presented to the

FACULTY OF THE USC GRADUATE SCHOOL

In partial fulfillment of the requirements for the degree of

Doctor of Philosophy

in

Environmental Engineering

May 2013

Copyright 2013 Winnie Kam

Dedication

To my parents, Kay and Paul,
for their unconditional love and support.

i
ii

Acknowledgements
The last four years have been an amazing journey for me both personally and
professionally. Since joining the aerosol lab at University of Southern California in 2009, first as
a master’s student then as a PhD student, I have grown immensely as an academic scientist. This
thesis would not have been possible without the help of many people for whom I would like to
pay my sincerest gratitude.
First and foremost, I would like to thank my advisor Professor Constantinos Sioutas for
his confidence in me and for his guidance in my doctoral research work. His deep knowledge of
aerosol research and scientific intuition has been invaluable towards my progress as a PhD
student. I am forever indebted to him for providing this priceless opportunity for me to be a part
of his premiere research group that continues to be at the forefront of aerosol research.
I would also like to thank all members of my guidance committee, Professor Ronald C
Henry, Professor Jiu-Chiuan Chen, Professor Scott Fruin, and Professor James Moffett for their
thoughtful suggestions and continual support of my research work. Special thanks to Professor
James J Schauer and Dr. Martin M Shafer of University of Wisconsin-Madison for their
insightful comments, which have been a substantive contribution to the development of my
dissertation.
Since research in the aerosol lab is largely a collective effort, my thesis would not be
possible without the help of a number of former and current colleagues at USC: Dr. Zhi Ning, Dr.
Kalam Cheung, Dr. Payam Pakbin, Dr. Vishal Verma, Dr. Neelakshi Hudda, Nancy Daher,
James Liacos, Dongbin Wang, Niloofar Hajibeiklou, and Shruthi Balasubramanian. I would like
to especially acknowledge Dr. Zhi Ning, Assistant Professor at City University of Hong Kong,
iii

for his mentorship during the early stages of my study and for his patience and friendship. It has
been a great pleasure to work with everyone that has been a part of the aerosol lab. In addition, I
would also like to thank Dr. Evangelia Kostenidou, Genevieve McSpaden, Ivy Zheng, and Xu
Zhen. Although their stay with the aerosol lab was brief, their involvement and commitment to
various research projects will be remembered.
Lastly, I would like to thank my fiancé, Kenneth C Leung, for his unending support and
trust in me during the course of my study.

iv

Table of Contents
Dedication ........................................................................................................................................ i
Acknowledgements ......................................................................................................................... ii
List of Figures ............................................................................................................................... vii
List of Tables .................................................................................................................................. x
Abstract ......................................................................................................................................... xii
Chapter 1 Introduction .................................................................................................................... 1
1.1 Background ...................................................................................................................... 1
1.2 Characteristics of airborne particles ................................................................................. 2
1.3 Health effects associated with PM ................................................................................... 3
1.4 Rationale for research ...................................................................................................... 4
1.4.1 Exposure assessment for passengers on the L.A. Metro subway and light-rail ........ 6
1.4.2 Exposure assessment for passengers in private vehicles in L.A. .............................. 7
1.5 Major research questions ................................................................................................. 8
Chapter 2 Particulate matter (PM) concentrations in underground and ground-level rail systems
of the Los Angeles Metro ............................................................................................................. 11
2.1 Sampling methodology for L.A. Metro subway and light-rail study ............................. 11
2.1.1 Monitoring instruments and sampling campaigns .................................................. 12
2.1.2 Quality assurance ................................................................................................... 14
2.2 Results and discussion ................................................................................................... 15
2.2.1 DustTrak correction factor ...................................................................................... 15
2.2.2 Overview of personal exposure concentrations of Metro red and gold line ........... 16
2.2.3 PM concentrations at station platforms and inside trains ....................................... 17
2.2.4 Comparison of red and gold line PM levels to nearby air monitoring sites ........... 21
2.2.5 Inter-correlations of PM concentrations to investigate sources .............................. 22
2.2.6 Comparison to worldwide rail systems ................................................................... 24
2.3 Conclusion ..................................................................................................................... 26

v

Chapter 3 Chemical characterization and redox potential of coarse and fine particulate matter
(PM) in underground and ground-level rail systems of the Los Angeles Metro .......................... 27
3.1 Sample analysis .............................................................................................................. 27
3.2 Results and discussion ................................................................................................... 28
3.2.1 Mass balance ........................................................................................................... 28
3.2.2 Crustal species ........................................................................................................ 30
3.2.3 Non-crustal species ................................................................................................. 31
3.2.4 Reactive oxygen species (ROS) activity ................................................................. 35
Chapter 4 Size-segregated composition of particulate matter (PM) in major roadways and surface
streets ............................................................................................................................................ 39
4.1 Introduction .................................................................................................................... 39
4.2 Experimental methodology ............................................................................................ 39
4.2.1 Sampling instrumentation ....................................................................................... 41
4.2.2 Sample analysis ....................................................................................................... 43
4.3 Results and discussion ................................................................................................... 44
4.3.1 Overview of campaign ............................................................................................ 44
4.3.2 Mass balance ........................................................................................................... 46
4.3.3 Inorganic ions.......................................................................................................... 50
4.3.4 EC and OC .............................................................................................................. 53
4.3.5 Comparison to previous studies in Los Angeles ..................................................... 55
4.4 Conclusion ..................................................................................................................... 57
Chapter 5 On-road emission factors of PM pollutants for light-duty vehicles (LDVs) based on
real-world urban street driving conditions .................................................................................... 58
5.1 Introduction .................................................................................................................... 58
5.2 Experimental methodology ............................................................................................ 59
5.2.1 Sampling route ........................................................................................................ 59
5.2.2 Sampling methodology and analysis ...................................................................... 60
5.2.3 Emission factors calculation ................................................................................... 61
5.3 Results and discussion ................................................................................................... 62
5.3.1 Emission factors for major PM components and species ....................................... 62
5.3.2 Comparison of PM
2.5
emission factors to previous LDV studies ........................... 67
5.3.3 n-Alkanes and calculation of carbon preference index (CPI) ................................. 75
vi

5.4 Conclusion ..................................................................................................................... 77
Chapter 6 A comparative assessment of PM
2.5
exposures in light-rail, subway, freeway, and
surface street environments in Los Angeles and estimated lung cancer risk ................................ 79
6.1 Introduction .................................................................................................................... 79
6.2 Experimental methodology ............................................................................................ 79
6.2.1 Sampling methodology ........................................................................................... 80
6.2.2 Route description .................................................................................................... 82
6.2.3 Sample analysis ....................................................................................................... 83
6.3 Results and Discussion .................................................................................................. 83
6.3.1 Comparability of the two campaigns ...................................................................... 83
6.3.2 Major PM components ............................................................................................ 87
6.3.3 Metals and trace elements ....................................................................................... 89
6.3.4 Water solubility of metals and trace elements ........................................................ 94
6.3.5 Polycyclic aromatic hydrocarbons (PAHs) ............................................................. 95
6.3.6 Lung cancer risk for commuters ............................................................................. 98
6.4 Conclusion ................................................................................................................... 100
Chapter 7 Conclusions and recommendations for future research ............................................. 102
7.1 METRO study conclusion............................................................................................ 102
7.2 On-road study conclusion ............................................................................................ 103
7.3 Integration of the METRO and on-road study conclusion........................................... 105
7.4 Recommendations for future research ......................................................................... 106
7.4.1 Limitations of the current studies ......................................................................... 106
7.4.2 Recommendations for future research .................................................................. 107
7.4.3 Recommendations for regulatory control ............................................................. 108
Publications from this thesis ....................................................................................................... 111
Bibliography ............................................................................................................................... 112

vii

List of Figures
Figure 1.1 Time spent traveling to work for the L.A. population (%). ....................................... 5
Figure 2.1 Map of sampling routes, the Metro red line (subway) and gold line (light-rail),
two nearest air quality monitoring stations, and the University of Southern
California (USC) urban ambient site ....................................................................... 11
Figure 2.2 (a) Carry-on suitcase setup for station/train intensive campaign, two DustTraks
with PM
2.5
and PM
10
inlet and TSI Q-Trak. (b) Setup for the personal exposure
campaign, three personal cascade impactor samplers (PCIS) with battery-
powered pumps and DustTrak ................................................................................. 13
Figure 2.3 Mass concentrations of PM
2.5
for the subway line (a) and the light-rail line (c).
Fine fraction (PM
2.5
/PM
10
) and coarse fraction (PM
10-2.5
/PM
10
) data from
DustTrak are presented for the subway line (b) and light-rail line (d) with PCIS
mass concentrations. ................................................................................................ 18
Figure 2.4 Continuous PM
2.5
, PM
10
, and CO
2
data for approximately one hour of sampling
for the subway line (a) and the light-rail line (b). Shaded areas represent times
when the subject was riding inside the train. ........................................................... 21
Figure 2.5 Comparison of PM
2.5
concentrations from personal exposure campaign with
ambient levels from two stations (Downtown L.A. and Burbank). Each point
represents a daily 3.5-hour average. (a) Red line (subway) vs Downtown L.A., (b)
red line vs Burbank, (c) gold line (light-rail) vs Downtown L.A., and (d) gold
line vs Burbank. ....................................................................................................... 22
Figure 2.6 Correlation of PM
2.5
and coarse PM data at all stations and inside train for
subway line (a) and light-rail line (b). Each data point represents an average of
the 7 days of sampling from the station/train intensive campaign. ......................... 23
Figure 3.1 Mass reconstruction of the 7 identified categories (a) coarse and (b) fine PM.
Gravimetric mass concentrations are also presented. .............................................. 29
Figure 3.2 Upper Continental Crustal (UCC) enrichment factors (EFs) for (a) coarse and (b)
fine mode using total elemental concentrations. ...................................................... 34
Figure 3.3 Reactive oxygen species (ROS) activity (a) per air volume (m
3
) and (b) per
gravimetric PM mass (mg). ..................................................................................... 38
viii

Figure 4.1 Map of the three sampling routes: 1) I-110 (blue), 2) I-710 (red), and
Wilshire/Sunset (purple). The USC background site is denoted by the red star. .... 40
Figure 4.2 Sampling schematic of the inlet into the vehicle and the instrumental set up. ........ 43
Figure 4.3 Size-fractionated mass summary the three roadways and USC background site.
S1 and S2 represent the two sets of samples collected, with a sampling duration
of approximately 50 hours for each set. Sampling dates are shown in Table 4.1. ... 45
Figure 4.4 Mass balance constructed based on five identified categories for a) PM
10-2.5
, b)
PM
2.5-0.25
, and c) PM
0.25
. Error bars represent one positive standard deviation. ...... 46
Figure 4.5 Inorganic ions concentrations for a) PM
10-2.5
, b) PM
2.5-0.25
, and c) PM
0.25
. Error
bars represent one standard deviation. ..................................................................... 51
Figure 4.6 Size-segregated concentrations of a) total carbon (TC), b) elemental carbon (EC),
c) organic carbon (OC), and d) water-soluble OC (WSOC).................................... 54
Figure 4.7 Comparison of PM
2.5
concentrations of a) mass and b) OC and EC to previous
studies conducted at fixed sites in the vicinity of the I-110 and I-710. ................... 56
Figure 5.1 Map of sampling route (Wilshire/Sunset Boulevards) and USC background site. . 60
Figure 5.2 Comparison of the velocity profile of the current study (Wilshire/Sunset) and the
two test cycles (FTP and UDC) of the dynamometer studies. The current study
shows a typical sampling hour with a 30s resolution based on GPS data. The
FTP and UDC driving schedules use a 1s resolution and can be found at
www.epa.gov. .......................................................................................................... 68
Figure 5.3 PM
2.5
emission factors for PM components for the current study (bars) and
previous LDV studies (markers) for PM mass, OC and EC. ................................... 69
Figure 5.4 PM
2.5
emission factors for metals and elemental species for the current study
(bars) and previous LDV studies (markers). Fe is shown separately due to its
difference in magnitude. .......................................................................................... 72
Figure 5.5 Comparison of PM
2.5
emission factors (μg/kg of fuel) between (a) PAHs and (b)
hopanes and steranes. ............................................................................................... 73
Figure 5.6 n-alkane concentrations (C19-C40) for (a) PM
10-2.5
, (b) PM
2.5-0.25
, and (c) PM
0.25
. 76
Figure 6.1 Map of five commute environments: 110 (green), 710 (blue), Wilshire/Sunset
(purple), METRO red line (red), and METRO gold line (yellow). The USC
ix

reference site is denoted by the star and the South Coast Air Quality
Management District (SCAQMD) monitoring site is denoted by the triangle. ....... 81
Figure 6.2 Comparison of major PM components at the USC reference site for the two
campaigns to assess comparability of data. All bars presented in this study
represent upper and lower data points (N=2). ......................................................... 84
Figure 6.3 Comparison of major PM components (OC, WSOC, EC, and TC) for the five
commute microenvironments. EC appears on a separate axis to highlight
differences. ............................................................................................................... 88
Figure 6.4 Comparison of concentrations of total metals and trace elements for the five
commute environments. ........................................................................................... 90
Figure 6.5 Comparison of water-solubility (%) of metals and trace elements for the five
microenvironments separated into high and low solubility species. ....................... 95
Figure 6.6 a) Concentrations of 11 PAHs and b) sum of PAHs concentrations and ΣBaPeq
for the five commute environments. ........................................................................ 96

x

List of Tables
Table 2.1 Summary of the personal exposure campaign. Average PCIS and DustTrak PM
concentrations. DustTrak correction factors are calculated based on PM
2.5

concentrations from the PCIS. ................................................................................. 16
Table 2.2 Average PM
10
, PM
2.5
, and coarse PM concentrations of the station/train intensive
sampling campaign. ................................................................................................. 17
Table 2.3 Correlation coefficients between PM
2.5
and coarse PM for the light-rail
line. N=7 .................................................................................................................. 24
Table 2.4 A comparison of PM
10
and PM
2.5
average concentrations for worldwide subway
systems. Fine fractions (PM
2.5
/PM
10
) are also presented. ........................................ 25
Table 3.1 Average total mass concentrations (ng/m
3
) and ratios relative to USC of crustal
and non-crustal metals. ............................................................................................ 30
Table 3.2 Coefficients of determination (R
2
) of total crustal and non-crustal species.
Correlation includes gold and red line data only for coarse and fine PM (N=7). .... 33
Table 3.3 ROS coefficients of determination (R
2
) with WS crustal and non-crustal species
and other PM components. Data includes coarse and fine data for red line, gold
line, and USC ambient site (N=9). ........................................................................... 36
Table 4.1 Summary of meteorological parameters from nearby air quality monitoring sites
(South Coast Air Quality Management District (SCAQMD)). Traffic data is from
the CalTrans database. ............................................................................................. 41
Table 4.2 Mass concentrations of metals (ng/m
3
). For PM
10-2.5
, N=2 with standard
deviations; for PM
2.5-0.25
and PM
0.25
, N=1. .............................................................. 47
Table 5.1 Fuel-based emission factors (mass of pollutant emitted per kg of fuel) of PM
components and metals and trace elements for three PM size fractions. Pollutants
with concentrations close to or less than USC background levels have been
omitted. .................................................................................................................... 63
Table 5.2 Fuel-based emission factors (mass of pollutant emitted per kg of fuel) of
polycyclic aromatic hydrocarbons (PAHs) and hopanes and steranes for three
PM size fractions. Pollutants with concentrations close to or less than USC
background levels have been omitted. ..................................................................... 66
xi

Table 5.3 Description of previous light-duty vehicle (LDV) studies used for comparison. .... 67
Table 5.4 Sum of the concentrations of n-alkanes (C19-C40), Cmax and its corresponding
concentrations, and CPI values. Values shown include uncertainty of one
standard deviation. ................................................................................................... 77
Table 6.1 Summary of sampling dates and times and meteorological parameters for the
METRO and on-road studies. Meteorological parameters are based on South
Coast Air Quality Management District (SCAQMD) monitoring site. ................... 80
Table 6.2 Average concentrations of metals and trace elements and PAHs at USC site
during the two campaigns. (N=2) ............................................................................ 86
Table 6.3 Meteorological parameters and gaseous pollutant measurements at South Coast
Air Quality Management District (SCAQMD) monitoring site in downtown Los
Angeles. ................................................................................................................... 87
Table 6.4 Mass concentrations of major PM components, metals and trace elements, and
PAHs at the five microenvironments (N=2). Only one sample for PAHs was
analyzed for the METRO gold line.......................................................................... 91
Table 6.5 Lung cancer risk calculations based on a commuter lifetime of 45 years, 2
hours/day, and 5 days/week. Unit risk factors for rodent and epidemiology are
1.1E-4 and 2.1E-3 (μg/m
3
)
-1
, respectively. .............................................................. 99

xii

Abstract
According to the U.S. Census Bureau, 570,000+ commuters in Los Angeles travel for
over 60 minutes to work. Studies have shown that a substantial portion of particulate matter (PM)
exposure can occur during this commute depending on the mode of transport. This thesis focuses
on the PM exposure for commuters of four microenvironments in Los Angeles including subway,
light-rail, freeways, and surface streets.
The first part of the thesis focuses on the subway and light-rail commute environments.
Elevated concentrations of PM have been found in a number of worldwide underground transit
systems, with major implications regarding exposure of commuters to PM and its associated
health effects. An extensive sampling campaign was to measure PM concentrations in two lines
of the Los Angeles Metro system – an underground subway line (Metro red line) and a ground-
level light-rail line (Metro gold line). Considering that a commuter typically spent 75% of time
inside the train and 25% of time waiting at a station, subway commuters were exposed on
average to PM
10
and PM
2.5
concentrations that were 1.9 and 1.8 times greater than the light-rail
commuters. The average PM
10
concentrations for the subway at station platforms and inside the
train were 78.0 μg/m
3
and 31.5 μg/m
3
, respectively; for the light-rail line, corresponding PM
10

concentrations were 38.2 μg/m
3
and 16.2 μg/m
3
. Regression analysis demonstrated that personal
exposure concentrations for the light-rail line are strongly associated with ambient PM levels
(R
2
=0.61), while PM concentrations for the subway line are less influenced by ambient
conditions (R
2
=0.38) and have a relatively stable background level of about 21 μg/m
3
. Mass
balance showed that in coarse PM (PM
10-2.5
), iron makes up 27%, 6%, and 2% of gravimetric
mass for red line, gold line, and USC, respectively; in fine PM, iron makes up 32%, 3%, and 1%.
Non-crustal metals, particularly Cr, Mn, Co, Ni, Mo, Cd, and Eu were elevated for the red line
xiii

and, to a lesser degree, the gold line. Bivariate analysis showed that reactive oxygen species
(ROS) activity is strongly correlated with water-soluble Fe (R
2
=0.77), Ni (R
2
=0.95), and OC
(R
2
=0.92). A multiple linear regression model (R
2
=0.94, p<0.001) using water-soluble Fe and
OC as predictor variables was developed to explain the variance in ROS.
The second part of the thesis focuses on PM exposure for private commuters of freeways
and surface streets. An on-road sampling campaign was conducted to assess on-road PM
composition for three size fractions (PM
10-2.5
,PM
2.5-0.25
, and PM
0.25
) on three representative
roadways in Los Angeles: 1) the I-110, a high-traffic freeway composed mostly of light-duty
vehicles (LDVs), 2) the I-710, a major freeway for heavy-duty vehicles (HDVs) travelling to and
from the Ports of Los Angeles and Long Beach, and 3) Wilshire/Sunset Blvd, two major surface
streets. Results showed that the PM
0.25
fraction is heavily influenced by vehicular emissions, as
indicated by average roadway PM concentrations that were 48.0±9.4% higher than those
observed at USC (p<0.05), while the PM
10-2.5
fraction is mostly influenced by resuspension of
road dust and the PM
2.5-0.25
fraction is mainly composed of secondary species. With very low EC
levels in PM
10-2.5
, the most notable difference among the three roadway environments was the
PM
2.5
EC levels observed on the I-710, which are 2.0±0.2 μg/m
3
and 4.1 times greater than USC.
Next, fuel-based emission factors (mass of pollutant per kg of fuel) were calculated to
assess the emissions profile of a light-duty vehicle (LDV) traffic fleet characterized by stop-and-
go driving conditions that are reflective of urban street driving. Emission factors for metals and
trace elements were highest in PM
10-2.5
while emission factors for polycyclic aromatic
hydrocarbons (PAHs) and hopanes and steranes were highest in PM
0.25
. PM
2.5
emission factors
were also compared to previous freeway, roadway tunnel, and dynamometer studies based on an
LDV fleet to determine how various environments and driving conditions may influence
xiv

concentrations of PM components. The on-road sampling methodology deployed in the current
study captured substantially higher levels of metals and trace elements associated with vehicular
abrasion (Fe, Ca, Cu, and Ba) and crustal origins (Mg and Al) than previous LDV studies. The
semi-volatile nature of PAHs resulted in higher levels of PAHs in the particulate phase for LDV
tunnel studies (Phuleria et al. 2006) and lower levels of PAHs in the particulate phase for
freeway studies (Ning et al. 2008). With the exception of a few high molecular weight PAHs, the
current study’s emission factors were in between the LDV tunnel and LDV freeway studies. In
contrast, hopane and sterane emission factors were generally comparable between the current
study, the LDV tunnel, and LDV freeway, as expected given the greater atmospheric stability of
these organic compounds.
Lastly, PM exposures for all commute environments were compared using mass per
volume of air as the metric of comparison. Metals associated with stainless steel, notably Fe, Cr,
and Mn, were elevated for the red line (subway), most likely from abrasion processes between
the rail and brakes; elements associated with tire and brake wear and oil additives (Ca, Ti, Sn, Sb,
and Pb) were elevated on roadways. Elemental concentrations on the gold line (light-rail) were
the lowest. Overall, the 710 exhibited high levels of PAHs (3.0 ng/m
3
), most likely due to its
high volume of HDVs, while the red and gold lines exhibited low PAH concentrations (0.6 and
0.8 ng/m
3
for red and gold lines, respectively). Lastly, lung cancer risk due to inhalation of PAHs
was calculated based on a commuter lifetime (45 years for 2 hours per workday). Results showed
that lung cancer risk for the 710 is 3.8 and 4.5 times higher than the light-rail (gold line) and
subway (red line), respectively. With low levels of both metals and PAH pollutants, our results
indicate that commuting on the light-rail (gold line) may have potential health benefits when
compared to driving on freeways and busy roadways.

Chapter 1 Introduction
1.1 Background
An aerosol is defined as a solid or liquid suspended in a gas medium. It is a two-phase
system and describes various forms of microscopic particles that remain in the air such as
resuspended soil dust, particles generated from vehicular combustion, photochemically formed
particles, and sea salt from the ocean. Particulate matter (PM) refers to the particles or liquid
droplets in the aerosol and is responsible for environmental effects such as visibility and climate
as well as numerous adverse health effects. In addition, PM is a major component of
photochemical smog and influences surface albedo by decreasing the amount of heat reaching
the surface (Seinfeld and Pandis 2006). The composition of PM is highly complex and varies
depending on local sources, source strength, and atmospheric processes such regional transport
and gas-to-particle partitioning. Therefore, investigating the physico-chemical and toxicological
characteristics of PM is crucial in understanding its environmental and health effects for both
policymakers and for the general public.
Particulate matter is made up of a number of chemical constituents including inorganic
ions (nitrate and sulfate), crustal metals and trace elements, elemental carbon (EC), and organic
species. The chemical components are derived from both natural and anthropogenic sources.
Natural sources include resuspended crustal elements, sea spray, and windborne biological
materials; anthropogenic sources include vehicular emissions, burning of fossil fuels and
biomass, and emissions from industrial activity. Particles that are emitted directly into the
atmosphere are known as primary pollutants, while particles formed in the atmosphere are
known as secondary pollutants, i.e. photochemical reactions with gaseous precursors (i.e.
1
2

nitrogen oxides). Once emitted, particles may undergo various physical and chemical processes
that may alter particle size and chemical composition.
1.2 Characteristics of airborne particles
In the context of PM, the most important parameter is particle size, which is usually
expressed as aerodynamic diameter or dp. Because particles exist in various shapes, aerodynamic
diameter is defined as the diameter of a unit density sphere that has the same settling velocity as
the particle. Airborne particles can range from the submicron (<1um) mode to tens of microns in
size. There are three major PM size fractions: coarse mode (or PM
10-2.5
) contains particles in the
range of 2.5 to 10μm, accumulation mode contains particles in the range of 100nm to 2.5μm, and
ultrafine PM are particles less than 100nm. Fine PM (or PM
2.5
) refers to particles less than 2.5μm
and PM
10
refers to particles less than 10μm. Particles greater than 10μm are typically of less
interest because these particles are characterized by low atmospheric residence times and
respiratory deposition in the nasal region, while PM
10
can enter the thoracic region and is of
great interest to air pollution studies and for regulatory purposes.
The PM size ranges exhibit differences in respiratory deposition, atmospheric formation
and deposition mechanisms, particle composition, and optical properties. The coarse fraction is
formed mainly from mechanical processes such as grinding, erosion, and wind resuspension, and
due to its relatively high settling velocity, its primary deposition mechanism is gravitational
settling (Hinds 1999). The accumulation mode is formed mainly through physical atmospheric
aging processes such as coagulation of smaller particles and growth of existing particles by
condensation. This mode tends to remain in the atmosphere for longer because its removal
mechanism is neither dominated by gravitational settling or coagulation processes. Ultrafine
particles are formed through incomplete combustion and gas-to-particle nucleation processes and
3

are primarily removed through coagulation with other particles into a larger size mode. Although
they have negligible mass, they dominate in particle number concentration and are efficiently
deposited by diffusional mechanisms into all regions of the respiratory tract, including the
alveolar region. In addition, its greater surface area per mass compared with larger particles
renders ultrafine particles to be more biologically active (Brown et al. 2001; Oberdorster et al.
2005).
Currently, PM
10
and PM
2.5
are regulated in the National Ambient Air Quality Standards
(NAAQS) under the Environmental Protection Agency (EPA), which uses mass concentration
(μg/m
3
) as the metric for regulation. The law sets two standards: the primary standard is designed
to protect public health (i.e. sensitive populations such as children, elderly, and those with
respiratory illnesses) and the secondary standard is designed to protect public welfare (i.e.
visibility, damage to buildings and crops).
1.3 Health effects associated with PM
Adverse health effects associated with PM remains the one of the main motivations for
current aerosol research. Numerous studies have found a link between respiratory, pulmonary,
cardiovascular effects and long-term exposure to atmospheric PM (Samet et al. 2000; Schwartz
et al. 2002; Pope and Dockery 2006; Li et al. 2009). Recent in-vivo and in-vitro studies have
shown that ultrafine particles may trigger a proinflammatory response in the mouse brain that
can contribute to neurodegenerative diseases (Campbell et al. 2005; Morgan et al. 2011).
Although the biological mechanisms responsible for toxicity of PM are still uncertain, numerous
studies have found a positive correlation with PM toxicity and its chemical components,
including organic carbon (OC) and elemental carbon (EC) (Mar et al. 2000; Metzger et al. 2004),
4

trace metals (Saldiva et al. 2002), and quinones and polycyclic aromatic hydrocarbons (PAHs)
(Xia et al. 2004).
It is also postulated that PM components have oxidative properties and the potential to
generate reactive oxygen species (ROS) which contribute to oxidative stress in the human body.
While ROS is a natural byproduct of aerobic metabolism, an imbalance in ROS levels may affect
tissue oxygen homeostasis. An increase in ROS concentrations has been shown to play a direct
role in pulmonary inflammation (Tao et al. 2003), which may lead to decreased lung function
and exacerbation of respiratory diseases such as asthma and chronic obstructive pulmonary
disease (COPD). The magnitude of ROS generation has also been hypothesized to be driven by
redox reactions of soluble PM constituents such as transition metals (Verma et al. 2010) and
organic compounds (Cho et al. 2005) as well as the PM size fraction (Hu et al. 2008). Because of
the complex chemical nature of PM and the spatial and temporal variation of local PM sources,
further research is needed to understand the physical processes and the chemical components that
contribute to ambient PM.
1.4 Rationale for research
In major metropolitan areas, vehicular emissions are the primary source of ambient PM
(Schauer et al. 1996), and are of particular importance to populations in the vicinity of trafficked
areas or downwind of major freeways. Studies conducted near freeways and major roadways
have found that PM levels were substantially elevated relative to areas that are farther from the
traffic source (Zhu et al. 2002; Ning et al. 2010). Populations in the proximity of trafficked
roadways are most susceptible to PM health effects (Tonne et al. 2007). However, the most
sensitive demographics are developing children near roadways (Brunekreef et al. 1997; Dales et
al. 2009) and the elderly (Liao et al. 1999; Creason et al. 2001).
5

Although the U.S. EPA estimates people spend an average of approximately 90% of their
time indoors (U.S. EPA “The Inside Story: A Guide to Indoor Air Quality”), a large portion of
their total daily PM exposure may be due to time spent outdoors or during their commute on
roadways. The U.S. Census Bureau estimates a total of 4.5M workers that are age 16 and over in
Los Angeles, who have a mean travel time to work of 30.7 minutes, resulting in a round trip
commute of approximately 1 hour. However, Figure 1.1 shows that the travel time to work is not
normally distributed and approximately 13.8% travel for 60 minutes or more each way to work
(2011 American Community Survey), accounting for over 621,000 people in Los Angeles.

Figure 1.1 Time spent traveling to work for the L.A. population (%).
Source: 2011 American Community Survey (ACS)

As one of the largest and most populated cities in the world, a number of transport modes
are available for the residents of L.A. including the subway, light-rail, train, bus, and private
vehicles. Due to the sprawled urban nature of L.A., private vehicles dominates as the major
mode of transportation and accounts for over 80% of the working population (2011 ACS),
contributing to the notorious traffic congestion of its major highway systems. Depending on the
0%
5%
10%
15%
20%
<10 10-14 15-19 20-24 25-29 30-34 35-44 45-59 >60
% of working L.A.
population
Commute time to work (minutes)
6

mode of commute, the passenger may be exposed to PM of various species and concentration
levels. Thus, understanding the chemical composition of PM for different commute
microenvironments is essential in assessing passenger exposure and potential health endpoints
associated with PM inhalation. In particular, my current research will focus on the exposure
assessment for public and private commuters in five different commute environments: light-rail,
subway, freeways, and surface streets.
1.4.1 Exposure assessment for passengers on the L.A. Metro subway and light-rail
Metro systems are an important transportation mode in megacities across the world that
commuters take on a daily basis. However, recent measurements in cities across the world
indicate that subway systems may present a unique microenvironment with particulate matter
(PM) concentrations subject to different influences than ground-level sources. Earlier studies
have documented elevated PM levels in major subway systems across the world. Mean exposure
levels in the London Underground rail system were 3-8 times higher than street-level
transportation modes (Adams et al. 2001); average daytime PM
2.5
and PM
10
levels in a Paris
railway station were approximately 5-30 times higher than levels on Paris streets (Raut et al.
2009). In addition, elevated concentrations of elements especially Fe, Mn, Cu, Ni, Cr have been
observed in numerous subway systems relative to ambient urban concentrations. A personal
exposure assessment of a passenger on the Helsinki subway system determined an increase of 3%
for total fine PM exposure levels, but nearly 200% increase for Fe, 60% increase for Mn, and 40%
increase for Cu (Aarnio et al. 2005). Another study estimated that commuters in London
spending 2h in the subway per day would increase their personal daily exposure by 17 μg/m
3
(Seaton et al. 2005). High Mn, Cr and Fe concentrations of 160-350 times greater than the
7

median for outdoors residential areas were observed for teenage commuters in the New York
City subway system (Chillrud et al. 2004).
Although passengers spend a relatively short amount of time in subway systems,
exposures to high concentrations of PM with enriched levels of certain elements may have
significant health implications. Few toxicological studies, primarily in-vitro, have been
conducted on the health impacts of subway particles. Stockholm subway particles were found to
be 8 times more genotoxic than ambient particles and up to 4 times more like to cause oxidative
stress to cells (Karlsson et al. 2005). Furthermore, the Stockholm subway particles caused more
DNA damage than particles produced from wood combustion (Karlsson et al. 2006). On the
other hand, studies have reported PM levels to be lower for the Hong Kong (Chan et al. 2002)
and Guangzhou (Chan et al. 2002) subway systems than compared to other transport modes.
Therefore, differences between subway ventilation methods, braking systems, wheel type, air
conditioning, system age, and train motive source make it impossible to directly extrapolate
results from previous studies to other subway systems.
1.4.2 Exposure assessment for passengers in private vehicles in L.A.
In major metropolitan areas, the major source of PM is from vehicular combustion
(Schauer et al. 1996; Querol et al. 2001), and populations in the proximity of trafficked roadways
are most susceptible to PM health effects (Tonne et al. 2007). However, the most sensitive
demographics are developing children near roadways (Brunekreef et al. 1997; Dales et al. 2009)
and the elderly (Liao et al. 1999; Creason et al. 2001). Although the underlying physiological
pathways for PM toxicity are uncertain, several studies have postulated that certain PM
components may play a role, including elemental carbon (EC), organic carbon (EC), and trace
metals (Metzger et al. 2004; Ostro et al. 2008; Verma et al. 2010).
8

As a result, it is important to characterize PM at major roadway environments, where
elevated levels of PM have been observed (Zhu et al. 2002). Various methodologies have been
used to assess the impact of vehicular emissions on ambient air. Chassis dynamometer studies
can measure the emissions of a target vehicle in a controlled environment (Schauer et al. 1999;
Yanowitz et al. 1999), but cannot account for particle aging and atmospheric dilution effects.
Another method is roadside sampling, where continuous and time-integrated instruments are
used to measure both physical and chemical components of PM at designated sites downwind of
roadways (Kuhn et al. 2005b; Ning et al. 2010). Several recent studies have conducted on-road
sampling, in which a mobile laboratory (typically a hybrid or electric vehicle) equipped with
various continuous instruments have been used to measure black carbon (BC) (Fruin et al. 2004),
particle size distribution (Gouriou et al. 2004; Westerdahl et al. 2005; Weimer et al. 2009), PM
2.5
,
particle-bound PAH, and gaseous pollutants including NOx, CO, ozone, and hydrocarbons
(Bukowiecki et al. 2002; Ning and Chan 2007; Weiss et al. 2011). However, there are no
previous on-road studies that have reported time-integrated chemical speciation data, which
would substantially enhance the current knowledge base of on-roadway PM as well as provide a
more comprehensive exposure assessment for commuters.
1.5 Major research questions
The objective of my thesis is to investigate a major research problem and address a
number of relevant research questions. The overarching theme of my proposal is the PM
exposure of commuters using different transport modes and how the different
microenvironments contribute to chemical composition of the PM. Although the results
presented will mostly focus on the mass concentrations and chemical composition of PM
samples collected in the various commute microenvironments, toxicity results will also be
9

discussed with implications to public health assessment. The major research questions I am
proposing to investigate in my thesis are:
1) In terms of PM composition, what are passengers exposed to on the L.A. Metro subway
and light-rail lines and to what degree is it elevated or not elevated relative to urban
ambient air?
2) How does the PM exposure in the two aforementioned L.A. Metro environments compare
to passengers commuting in private vehicles on major roadways and surface streets in
L.A.?
3) Of the 5 commute microenvironments (subway, light-rail, freeway with mostly light-duty
vehicles, freeway with a higher percentage of heavy-duty vehicles, and major arterial
roads) presented, which type of passengers are exposed to the greatest PM toxicity and
what are the main factors contributing to this health effect?

As air pollution is a regional problem, the results from my current work and my future
investigation will be specific to the areas of Los Angeles. Chapter 1 has provided an introduction
to air pollution and the major health problems associated with particulate matter in the context of
major metropolitan areas. It has also provided the rationale as to why assessing PM exposure for
various commute microenvironments is essential in understanding how differences in PM
composition can contribute to adverse health effects. Chapter 2 and 3 discusses the results based
on a METRO study in Los Angeles investigating a light-rail (METRO gold line) and a subway
(METRO red line). Major results for PM mass and chemical speciation data and toxicity results
are presented. Chapters 4 and 5 will present the results based on a major on-road study for three
differential private commute environments including an HDV freeway (710), an LDV freeway
10

(110), and major surface streets. Chapter 4 focuses on a mass balance analysis and comparison of
major PM components, while chapter 5 focuses on LDV emission factors based on
measurements at the surface street environment. Chapter 6 integrates the METRO and on-road
campaign and focuses comparing PM components and species that are associated with adverse
health effects between the five commute microenvironments. At the end, estimates of lung
cancer risk are computed for all environments and are compared. Chapter 7 provides a
conclusion of the thesis and recommendations for future work and regulation.

11

Chapter 2 Particulate matter (PM) concentrations in underground
and ground-level rail systems of the Los Angeles Metro

2.1 Sampling methodology for L.A. Metro subway and light-rail study
Two lines of the L.A. Metro system were sampled in this study – the red and gold lines.
The red line is a subway line that spans approximately 17 km connecting downtown L.A. to
North Hollywood. Weekly ridership for the red line is estimated to be 150,000, which is the
highest of the Metro rail lines and accounts for almost 50% of system wide ridership. The gold
line, which began operation in 2003, is a ground-level light-rail line that connects Pasadena to
downtown L.A. Weekly ridership for the gold line is estimated to be 35,000 as of August 2010.
For both lines, trains pass every 8-10 minutes during rush hours and 10-12 minutes during
normal hours. A one-way trip to and from Union Station is approximately 30 minutes
(www.metro.net). Figure 2.1 shows a map of where the Metro red and gold lines run, two nearby
air quality monitoring sites (Downtown L.A. and Burbank), and the University of Southern

Figure 2.1 Map of sampling routes, the Metro red line (subway) and gold line (light-rail), two nearest air quality monitoring
stations, and the University of Southern California (USC) urban ambient site
12

California (USC) urban ambient site, which was used to represent urban background
concentrations.
2.1.1 Monitoring instruments and sampling campaigns
The Q-Trak Indoor Air Quality Monitor Model 7565 (TSI Inc., Shoreview, MN) was
used to determine CO
2
concentrations, temperature, and relative humidity at a logging interval of
15 seconds. The DustTrak Aerosol Monitor Model 8520 (TSI Inc., Shoreview, MN) was used to
measure continuous PM
2.5
and PM
10
concentrations at a logging interval of 30 seconds. Previous
studies have shown that light-scattering aerosol measuring devices are subject to error when
relative humidity is greater than 60% (Lowenthal et al. 1995; Sioutas et al. 2000; Chakrabarti et
al. 2004). Q-Trak reported relative humidity levels that are within the operating range of the
DustTrak. Airborne PM was collected with the Sioutas™ Personal Cascade Impactor Sampler
(SKC Inc., Eighty-Four, PA), also referred to as PCIS (Misra et al. 2002; Singh et al. 2003),
which was operated with a Leland Legacy Pump (SKC Inc., Eighty-Four, PA) at a flow rate of 9
liters per minute (lpm) (Brinkman et al. 2008). The pumps were calibrated with Gilian
Gilibrator-2 Air Flow Calibrator (Sensidyne Inc., Clearwater, FL) before and after sampling.
During the sampling, the pump flows were checked regularly with flow meters. The PCIS was
prepared using one impaction stage with a cutpoint of 2.5 μm and an after filter stage, collecting
coarse and fine PM, respectively. For the purpose of chemical analysis, the PCIS were loaded
with two types of filters. One was loaded with PTFE (Teflon) filters, with a 25mm Zefluor
supported PTFE filter (Pall Life Sciences, Ann Arbor, MI) as the impaction substrate and a
37mm PTFE membrane filter with PMP ring (Pall Life Sciences, Ann Arbor, MI) as the after
filter. The other unit was loaded with quartz microfiber filters (Whatman International Ltd,
Maidstone, England). The Teflon filters were gravimetrically analyzed using a MT5
13

Microbalance (Mettler-Toledo Inc., Columbus OH), which has a detection limit of 10 μg. A total
of 9 PCIS were deployed in this study, with 3 PCIS for each line and 3 PCIS at the fixed site.
Figure 2.2b shows the setup of the PCIS inside the carry-on suitcases that were used for PM
collection on the subway.

Figure 2.2 (a) Carry-on suitcase setup for station/train intensive campaign, two DustTraks with PM
2.5
and PM
10
inlet and TSI Q-
Trak. (b) Setup for the personal exposure campaign, three personal cascade impactor samplers (PCIS) with battery-powered
pumps and DustTrak

The sampling campaign took place from May 3 – August 13, 2010. Two sub-campaigns
were undertaken during this period: 1) the station/train intensive campaign, which focuses on
measuring real-time PM
2.5
, PM
10
, and CO
2
concentrations simultaneously at each station and
inside the train, and 2) the personal exposure campaign, focusing on airborne PM exposures for
Metro commuters by sampling concurrently on the red line, gold line, and at the USC ambient
fixed site. The station/train intensive sampling occurred on a weekly basis and alternated
between the two lines each week (i.e. each line is sampled every other week), accruing 7 days of
sampling for each line. The subject carried a suitcase equipped with two DustTraks, one with a
PM
2.5
inlet and one with a PM
10
inlet, and a Q-Trak (Figure 2.2a). For the personal exposure
campaign, the red line, gold line, and USC fixed ambient site were sampled concurrently for 3.5
14

hours (9:30 am to 1:00 pm) for 4 out of the 5 weekdays. The campaign was designed to
determine the personal PM exposure of riders on both lines and compare with each other and
with an urban ambient site (USC). The two subjects were directed to spend approximately 75%
of time in the train and 25% of time at stations, which represents a typical commute of a
passenger. The two subjects each carried a suitcase with 3 PCIS (two with PTFE filters and one
with quartz filters) and 3 pumps inside and one DustTrak with a PM
2.5
inlet (Figure 2.2b). The
PCIS are attached with pre-tested PM
10
inlets (Arhami et al. 2009) so that the impaction
substrates are only collecting coarse particles (PM
10-2.5
). The USC urban ambient site was also
equipped with an identical suitcase and inlets with 3 PCIS (2 Teflon, 1 quartz) inside. For each
site, two samples were collected.
2.1.2 Quality assurance
To determine the comparability of the two DustTraks deployed for the sampling
campaign, the DustTraks were tested by collocated sampling before, in the middle, and at the end
of the campaign. A correlation of the PM readings for the two DustTraks shows that they are
within 10% of each other (y = 1.10x + 0.001) and have an R
2
of 0.99. During the campaign, the
DustTrak was maintained at its working flow rate of 1.7 lpm. In addition, the DustTraks were re-
zeroed and their impaction plates were cleaned on a daily basis. The Q-Trak was calibrated by
zero checking and re-zeroing, if necessary, before and during the campaign. The 9 PCIS used for
the campaign were also tested by collocated sampling, and gravimetric analysis revealed that the
PCIS agreed within 10-15% in mass concentrations. After sampling each day, the PCIS with
filter substrates were sealed with parafilm and stored in a -4°C freezer.
15

2.2 Results and discussion
2.2.1 DustTrak correction factor
The PM measurements obtained from the DustTrak monitor form the basis of the results
discussed in this study. The DustTrak monitor is a direct-reading photometer, operated based on
aerosol light scattering. Thus, physical properties of the sampled aerosols such as density, index
of refraction and size distribution strongly influence DustTrak measurements (Gorner et al.
1995). The DustTrak monitors were factory-calibrated against a gravimetric measurement of the
International Organization for Standardization (ISO) 12103-1, A1 dust (Arizona Test Dust).
When the physical characteristics of the sampled aerosols are significantly different from the test
aerosol, a correction factor (CF) is needed to adjust the DustTrak readings to actual PM
concentrations. Depending on the sampling environment and aerosol characteristics, DustTrak
monitors typically read higher than the reference methods by a factor of 2-4 (Chung et al. 2001;
Yanosky et al. 2002; Kim et al. 2004). In this study, impaction-based (PCIS) PM measurements
and DustTrak PM
2.5
readings were collected simultaneously during the personal exposure
campaign. The DustTrak CF was calculated based on the mass concentrations derived from the
PCIS, which is shown in Table 2.1. The CFs are 1.43 and 1.86 for sampling periods 1 and 2 for
the light-rail (gold) line, respectively, similar to CFs found in other environments (Heal et al.
2000; Kingham et al. 2006). Contrary to CF factors reported in literature, the CFs are somewhat
low (0.85 and 1.11 for periods 1 and 2, respectively) for the subway line. Due to the number of
DustTraks available for this campaign, a PM
10
CF could not be calculated. All DustTrak data
reported in this study have been corrected, in which PM
10
data was adjusted with its
corresponding PM
2.5
CF.
16

Table 2.1 Summary of the personal exposure campaign. Average PCIS and DustTrak PM concentrations. DustTrak
correction factors are calculated based on PM
2.5
concentrations from the PCIS.

2.2.2 Overview of personal exposure concentrations of Metro red and gold line
Table 2.1 presents the results from the personal exposure campaign, which was designed
to measure the PM exposure of a typical commute of a rider (75% of time spent inside train and
25% of time spent waiting at a station platform). The average PM
10
, PM
2.5
, and PM
10-2.5
, or
coarse PM, mass concentrations obtained from the PCIS, and the PM
2.5
mass concentrations
obtained from the DustTrak are presented. Coarse PM is a subset of PM
10
and is calculated as the
difference of the adjusted PM
10
and PM
2.5
values. The personal exposure PM
10
and PM
2.5

concentrations on the subway line are 1.9 and 1.8 times greater than the corresponding
concentrations for the light-rail line, indicating that subway commuters are exposed to almost
double the PM concentrations of light-rail commuters. In comparison to the urban ambient site
(USC), PM
10
and PM
2.5
exposure concentrations for the subway line are on average 1.4 and 1.7
times higher, respectively; for the light-rail line, the concentrations are 0.76 and 0.94 times that
of corresponding USC ambient concentrations. Using a paired t-test, the personal exposure PM
2.5

levels for the subway and light-rail line reported by the DustTrak are statistically different
(p<0.001). Interestingly, personal exposure to coarse PM levels for the subway line are almost
Period Dates of sampling
PM
10
(μg m
-3
)
PM
2.5
(μg m
-3
)
Coarse PM
(μg m
-3
)
PM
2.5
(μg m
-3
)
Correction
factor
Subway system 1 May 3 - Jun 11, 2010 45.8 33.6 12.1 28.6 0.85
(red line) 2 Jun 14 - Aug 13, 2010 41.6 32.0 9.6 35.5 1.11
1 May 3 - Jun 11, 2010 22.7 16.6 6.1 23.9 1.43
2 Jun 14 - Aug 13, 2010 23.3 20.1
3.2
37.3 1.86
1 May 3 - Jun 11, 2010 31.0 19.0 12.1 - -
2 Jun 14 - Aug 13, 2010 29.5 20.0 9.5 - -
Ground-level light-rail
system (gold line)
Urban ambient site
(USC)
PCIS Dust Trak
17

equivalent to the urban ambient site, while light-rail levels are on average 0.43 times those of
ambient levels. The lower coarse PM exposure is most likely due to the subject spending 75% of
the time inside the train. This factor is further investigated in the next section.
2.2.3 PM concentrations at station platforms and inside trains
The results from the station/train intensive campaign for the subway line, light-rail line,
and urban ambient site are summarized in Table 2.2. In general, the subway’s platforms and train
have PM
10
and PM
2.5
concentrations that are approximately double those of the light-rail’s
platforms and train levels; however, coarse PM levels for the subway platforms and trains are 2.4
and 2.9 times greater than the light-rail platforms and train levels, likely a result of the enclosed
tunnel environment of the subway line. The light-rail platform PM concentrations are
comparable to the USC fixed site concentrations, while the subway stations have PM
10
, PM
2.5
,
and coarse PM levels that are 2.5, 2.8, and 2.0 times greater than those at USC.

Table 2.2 Average PM
10
, PM
2.5
, and coarse PM concentrations of the station/train intensive sampling campaign.

Stations (all) 78.0 ± 16.5 56.7 ± 11.3 21.3 ± 5.6
Train 31.5 ± 10.8 24.2 ± 6.9 7.3 ± 6.4
Stations (all) 38.2 ± 4.1 29.4 ± 4.2 8.8 ± 1.4
Train 16.2 ± 6.8 13.7 ± 5.3 2.5 ± 2.4
PM
2.5
(μg m
-3
)
PM
10
(μg m
-3
)
19.9 30.7
Coarse PM
(μg m
-3
)
10.8
Light-rail line (gold)
Subway line (red)
Urban ambient site (USC)
18

PM ratio
0.2
0.4
0.6
0.8
1.0
PCIS ratio
Fine PM
Coarse PM
c)
Union Station
Chinatown
Lincoln/Cypress
Heritage Square
Southwest Museum
Highland Park
Mission
Fillmore
Del Mar
Memorial Park
Lake
Allen
Sierra Madre
TRAIN
PM concentration ( g m
-3
)
10
20
30
40
50
60
PM
2.5
(Personal Exposure)
PM
2.5
(Station/Train)
PM
10
(Station/Train)
d)

b)
PM ratio
0.2
0.4
0.6
0.8
1.0
PCIS ratio
Fine ratio
Coarse ratio
a)
Union Station
Civic Center
Pershing Square
7th St/Metro Center
Westlake/MacArthur
Wilshire/Vermont
Vermont/Beverly
Vermont/Santa Monica
Vermont/Sunset
Hollywood/Western
Hollywood/Vine
Hollywood/Highland
Universal City
North Hollywood
TRAIN
PM concentration ( g m
-3
)
20
40
60
80
100
120
PM2.5 (Personal exposure)
PM2.5 (Station/train)
PM10 (Station/train)

Figure 2.3 Mass concentrations of PM
2.5
for the subway line (a) and the light-rail line (c). Fine fraction (PM
2.5
/PM
10
)
and coarse fraction (PM
10-2.5
/PM
10
) data from DustTrak are presented for the subway line (b) and light-rail line (d)
with PCIS mass concentrations.
Figures 2.3a and 2.3c show the average PM
10
and PM
2.5
concentrations from the
station/train intensive campaign and the average PM
2.5
concentrations from the personal
exposure campaign for the subway line and light-rail line, respectively, with error bars of one
19

standard deviation. Note that the values presented for the station/train intensive campaign are the
average of bi-weekly concentrations. For the underground subway stations, the concentrations
vary (i.e. PM
10
values range from 50 to 100 μg/m
3
) while the ground-level, light-rail station
PM
10
concentrations are distributed in a narrower range (31 to 48 μg/m
3
). It is possible that the
ventilation system installed at some stations along the subway line may be more efficient at
removing PM than at other stations. Also, the bi-weekly variation of the light-rail line is much
greater than the subway line variation; one standard deviation of the average PM values is 50%
and 25% for the light-rail and subway line means, respectively. This indicates a greater temporal
variation of PM levels in the light-rail line than the subway line. To further confirm this
observation, a paired t-test was performed for each line between the mean PM
2.5
concentrations
of the personal exposure and station/train intensive campaigns. At 95% confidence interval, the
light-rail line mean data are significantly different (p=0.001), while the subway line mean data
are not significantly different (p=0.64). This suggests that the light-rail line PM concentrations
may vary according to seasonal and meteorological conditions, while the subway line PM
concentrations are less influenced by temporal changes.
The fine fraction (PM
2.5
/PM
10
) and coarse fraction (PM
10-2.5
/PM
10
) of PM are shown for
all stations and inside the train with corresponding error bars in Figures 2.3b and 2.3d for the
subway line and light-rail line, respectively. The calculated PCIS fine and coarse fractions are
also shown in the figures to demonstrate the agreement between the PCIS and DustTrak data. For
the subway line, the station platforms and train have overall fine PM fraction averages of 0.73
and 0.79, respectively; for the light-rail line, the corresponding fine fraction averages are 0.78
and 0.86. In general, commuters are exposed to a somewhat lower fine PM fraction and thus a
greater coarse fraction while they are waiting at the stations than while riding inside the train.
20

This is consistent with a subway study conducted in Taiwan, which also found PM
2.5
/PM
10
to be
higher inside the train (0.75-0.78) than at station platforms (0.67-0.75) (Cheng et al. 2008). A
possible reason for the lower coarse fraction inside the trains is that the air conditioning system
of the train may be more efficient at removing larger coarse mode particles than smaller particles
in the fine mode. Another subway study in Hong Kong found that a non-air-conditioned
transport system had a significantly lower fine fraction (0.63-0.68) than an air-conditioned
system (0.71-0.78) (Chan et al. 2002).
Figures 2.4a and 2.4b show real-time PM
2.5
, PM
10
, and CO
2
concentrations collected
simultaneously for an hour of sampling for the subway line and light-rail line, respectively. In
general, PM
2.5
and PM
10
levels follow each other consistently. A noticeable build-up of CO
2

occurs inside the train, whereas CO
2
levels drop rapidly as the commuter steps out of the train.
The CO
2
measured inside the train is primarily from the exhaled breath of the riders. It is also
important to note that when commuters stand right next to the train door, they are exposed to an
immediate flux of particles and a reduction of CO
2
when the train door opens, which can be seen
by the simultaneous peaks and dips in Figure 2.4a. Inside the train, CO
2
level reaches up to 1200
ppm, which is 3 to 4 times higher than the level of ambient CO
2
concentrations, but still not at a
level of concern for commuters. A study in the Seoul subway system reported CO
2
levels ranging
from 1153 to 3377 ppm (Park and Ha 2008). In Figure 2.4a, as the train departs from Union
Station, CO
2
concentrations stay level until the train reaches 7
th
St/Metro Center, where a large
number of passengers enter the train, thus creating a surge in CO
2
levels. Although the number of
passengers in the train car was not recorded, the accumulation of CO
2
depends strongly on this
factor.
21

Figure 2.4 Continuous PM
2.5
, PM
10
, and CO
2
data for approximately one hour of sampling for the subway line (a)
and the light-rail line (b). Shaded areas represent times when the subject was riding inside the train.

2.2.4 Comparison of red and gold line PM levels to nearby air monitoring sites
To investigate the influence of ambient PM concentrations on the two Metro lines,
figures 2.5a-d show the association of personal exposure PM
2.5
concentrations of the subway line
(a-b) and the light-rail line (c-d) with the PM
2.5
levels recorded at the two nearest monitoring
sites in Downtown L.A. and Burbank (Figure 2.1). The sites are maintained by the South Coast
Air Quality Management District (SCAQMD). Each data point represents a 3.5-hour average
concentration from each day of sampling (N=54). The linear regression analysis reveals a
moderately strong relationship between light-rail line personal exposure concentrations with
ambient levels (R
2
=0.62 and 0.59 for the Downtown L.A. and Burbank site, respectively) and a
weaker relationship between subway line personal exposure concentrations with ambient levels
(R
2
=0.38 and 0.38). This suggests that the light-rail line is more influenced by ambient PM
levels than the subway line. Also, given the relatively small y-intercept of the regression lines
22

(2.25 and 1.45 µg/m
3
in figures 2.5c and 2.5d, respectively), local emissions are presumably the
main source of pollutants for the light-rail line. On the other hand, the subway line appears to
have an almost constant background concentration of approximately 21 μg/m
3
. The linear
regression analysis implies that on a day with high episodic ambient PM concentrations, light-
rail commuters may be subjected to comparable or higher personal exposure concentrations than
subway commuters.
y = 0.40x + 21.7
R
2
= 0.38
N = 49
Ambient PM
2.5
concentration ( g m
-3
) from air quality
monitoring site in downtown Los Angeles
10 20 30 40 50 60
Metro red line (subway) PM
2.5

concentration ( g m
-3
)
10
20
30
40
50
60
a)
y = 0.49x + 21.0
R
2
= 0.38
N = 49
Ambient PM
2.5
concentration ( g m
-3
) from air quality
monitoring site in Burbank
10 20 30 40 50
Metro red line (subway) PM
2.5
concentration ( g m
-3
)
10
20
30
40
50
60
b)

c)
Ambient PM
2.5
concentration ( g m
-3
) from air quality
monitoring site in downtown Los Angeles
10 20 30 40 50 60
Metro gold line (light-rail) PM
2.5
concentration ( g m
-3
)
10
20
30
40
50
y = 0.61x + 2.25
R
2
= 0.62
N = 52
d)
Ambient PM
2.5
concentration ( g m
-3
) from air quality
monitoring site in Burbank
10 20 30 40 50
Metro gold line (light-rail) PM
2.5
concentration ( g m
-3
)
10
20
30
40
50
y = 0.74x + 1.45
R
2
= 0.59
N = 52

Figure 2.5 Comparison of PM
2.5
concentrations from personal exposure campaign with ambient levels from two
stations (Downtown L.A. and Burbank). Each point represents a daily 3.5-hour average. (a) Red line (subway) vs
Downtown L.A., (b) red line vs Burbank, (c) gold line (light-rail) vs Downtown L.A., and (d) gold line vs Burbank.

2.2.5 Inter-correlations of PM concentrations to investigate sources
To investigate the sources of PM in the two microenvironments, station platforms and
trains, correlation analysis between PM
2.5
and coarse PM was done. In Figure 2.6a and 2.6b, the
23

data points represent the station/train intensive campaign average of the 7 days of sampling for
each station and inside the train for the subway line and light-rail line, respectively. Since it was
previously established that the two microenvironments have a common source of PM, a linear
regression was performed for both the station and train data points. The high correlation
(R
2
=0.89) for the subway line scatter plot (Figure 2.6a) indicates that PM
2.5
and coarse PM have
a common origin. Although this common PM source cannot be determined based on the data
presented in this manuscript, previous studies have attributed metallic components of PM to
originate from the friction of the wheels on the steel rails, the vaporization of metals due to
sparking, wear of brakes (Pfeifer et al. 1999; Sitzmann et al. 1999), and particulate resuspension
and dispersion from train and passenger movement (Chan et al. 2002; Raut et al. 2009). The
upcoming chemical analysis of fine and coarse PM will help determine the degree to which the
aforementioned sources may contribute to PM exposure in the underground environment.
a)
Coarse PM ( g m
-3
)
10 20 30
PM
2.5
( g m
-3
)
20
40
60
80
Stations
Train
y = 2.0x + 13.6
R
2
= 0.89
b)
Coarse PM ( g m
-3
)
2 4 6 8 10 12 14
PM
2.5
( g m
-3
)
10
20
30
40
Stations
Train
y = 1.26x + 17.7
R
2
= 0.21

Figure 2.6 Correlation of PM
2.5
and coarse PM data at all stations and inside train for subway line (a) and light-rail
line (b). Each data point represents an average of the 7 days of sampling from the station/train intensive campaign.

Figure 2.6b shows a weak correlation between PM
2.5
and coarse PM for the light-rail
stations and train (R
2
=0.21), suggesting that they do not share a common source. However, given
the temporal variation of ambient air and its strong influence on the personal PM exposure levels
of the light-rail line (Figure 2.5c and 2.5d), the data points presented in the figure, which
24

represent the average of the 7 days of sampling, do not account for this significant factor. To
account for the day-to-day variation of light-rail line PM concentrations, a linear regression of
the daily fine and coarse PM levels was conducted for each station and the train (N=7). The
correlation coefficients are presented in Table 2.3, which range from 0.52 to 0.92. This moderate
to strong correlation suggests that fine and coarse PM for the light-rail line may indeed have a
common source. It is reasonable to hypothesize that the primary source of particulate pollution
for the light-rail line is from local emissions (vehicular traffic, road dust, photochemical
reactions, etc). The daily operations of the light-rail trains (i.e. movement of the train) may also
affect PM levels, but their impact is expected to be considerably smaller, given its exposed
environment.

Table 2.3 Correlation coefficients between PM
2.5
and coarse PM for the light-rail line. N=7

2.2.6 Comparison to worldwide rail systems
The Los Angeles Metro system, which began operation in 1993, is a relatively new rail
system compared to other systems around the world. Table 2.4 displays the average and range of
PM
10
and PM
2.5
concentrations and its fine fraction (PM
2.5
/PM
10
) for different rail systems. In
R
Union Station 0.81
Chinatown 0.80
Lincoln/Cypress 0.88
Heritage Square 0.77
Southwest Museum 0.82
Highland Park 0.85
Mission 0.74
Fillmore 0.73
Del Mar 0.68
Memorial Park 0.52
Lake 0.75
Allen 0.84
Sierra Madre 0.67
TRAIN 0.92
Gold line
25

comparison to the PM levels of the subway systems presented, the particulate levels measured at
the two rail lines of the Los Angeles Metro system fall on the relatively ‘cleaner’ side. On
average, the underground stations of the Metro red line have PM
10
and PM
2.5
levels that are 2.5
and 2.9 times greater than the USC urban ambient site, while an underground station in the
Stockholm subway system had levels that are 4.8 and 11.2 times greater than its corresponding
ambient site (Johansson and Johansson 2003). A study on the Seoul subway system reported
levels of an underground station to be only 2.3 and 1.3 times greater than ambient
concentrations, but the ambient PM
10
and PM
2.5
concentrations were 155 μg/m
3
and 102 μg/m
3

respectively (Kim et al. 2008). The Seoul subway train PM levels are also exceptionally high
because of the lack of a mechanical ventilation system inside the train (Park and Ha 2008). A
study on the London subway system, the oldest rail system in the world, reported underground

Table 2.4 A comparison of PM
10
and PM
2.5
average concentrations for worldwide subway systems. Fine fractions
(PM
2.5
/PM
10
) are also presented.
City (study year) Measurement location
average range (min-max) average range (min-max)
Los Angeles (2010) in train (gold line - ground level) 14 3-38 16 6-53 0.88 Current study
in train (red line - underground) 24 11-62 31 14-107 0.77
ground level station platforms (all stations) 29 4-77 38 8-184 0.76
underground station platforms (all stations) 57 9-130 78 14-197 0.73
urban ambient site 20 - 31 - 0.65
Taipei (2007) in train (underground) 31 19-51 40 22-71 0.78 Cheng et al. (2008)
underground station platform 44 22-91 66 29-130 0.67
ground level station platform 33 7-94 44 11-131 0.75
Paris (2006) underground station platform (rush hours) 93 - 320 - 0.29 Raut et al. (2009)
underground station platform (normal hours) 61 - 200 - 0.31
Helsinki (2004) in train (underground) 21 17-45 - - - Aarnio et al. (2005)
underground station platform 50 37-87 - - -
Seoul (2004) underground station platform 129 82-176 359 238-480 0.36 Kim et al. (2008)
in train (underground) 126 115-136 312 29-356 0.40
urban ambient site 102 41-174 155 79-254 0.66
Stockholm (2000) underground station platform 258 105-388 469 212-722 0.55 Johansson and Johansson (2003)
urban ambient site 23 3-89 98 6-454 0.23
New York City (1999) integration of 5h at station platform and 3h in train 62 - - - - Chillrud (2004)
Hong Kong (1999) in train (mostly underground) 33 21-48 44 23-85 0.75 Chan et al. (2002)
in train (mostly ground level) 46 29-68 60 41-89 0.77
London (1999) in train (underground line) 247 105-371 - - - Adams et al. (2001)
in train (above ground line) 29 12-42 - - -
PM
2.5
( μg m
-3
) PM
10
(μg m
-3
)
Fine fraction
(PM
2.5
/PM
10
)
Reference
26

train PM
2.5
levels around 250 μg/m
3
(Adams et al. 2001). A study on the New York City subway
system measured PM
2.5
levels on average of 62.5 μg/m
3
(Chillrud et al. 2004). The PM levels of
the L.A. Metro system are comparable to the Taipei, Helsinki, and Hong Kong subway systems,
which are generally newer systems and equipped with more efficient ventilation systems and
advanced braking technologies. Higher fine fractions (PM
2.5
/PM
10
) are also observed for the
newer rail systems, while the older systems (London, Paris, Stockholm) exhibit lower fine
fractions, consistent with the fact that older ventilation systems may be less efficient at filtering
larger particles.
2.3 Conclusion
An intensive particulate sampling campaign was conducted in spring and summer of
2010 to compare two types of rail systems on the L.A. Metro, an underground subway system
(Metro red line) and a ground-level light-rail system (Metro gold line). In general, commuters of
the subway line are exposed to greater PM concentrations than commuters of the light-rail line
by almost two-fold. Regression analysis showed that the light-rail line is heavily influenced by
ambient PM levels and its particulate pollutants originate from local sources, such as vehicular
emissions and road dust. The subway line is less influenced by ambient PM levels and has an
additional source of airborne particulate pollution that is generated from the daily operation of
trains. Strong correlations of PM
2.5
and PM
10
between train and stations reveal that PM from
stations is the main source of PM inside trains. PM
2.5
and coarse PM are also highly correlated,
suggesting they are also derived from the same source. The next chapter will provide a
comprehensive chemical analysis that will be essential in determining the sources of PM in the
two rail lines as well as their toxicological potential.

27

Chapter 3 Chemical characterization and redox potential of coarse
and fine particulate matter (PM) in underground and
ground-level rail systems of the Los Angeles Metro

3.1 Sample analysis
The Teflon filters were equilibrated for 24h and then weighed before and after sampling
to determine gravimetric mass concentrations using a MT5 Microbalance (Mettler-Toledo Inc.,
Columbus, OH; uncertainty of ±5μg) in a temperature and relative humidity-controlled room.
The filters were subsequently cut into 3 equal sections. The first section was analyzed by means
of magnetic-sector Inductively Coupled Plasma Mass Spectroscopy (SF-ICPMS) to determine
total elemental composition using an acid extraction (Zhang et al. 2008). The second section was
extracted using Milli-Q water and aliquots were dispensed for SF-ICPMS analysis to determine
water-soluble elemental composition and for ion chromatography (IC) analysis to determine the
PM concentrations of inorganic ions (Kerr et al. 2004). For the third section, an alveolar
macrophage assay was used to determine the reactive oxygen species (ROS) activity of aqueous
suspensions of the collected PM. The location of the alveolar macrophage on the inner epithelial
surface of the lung renders this assay an appropriate model of pulmonary inflammation in
response to PM exposure. In addition, the fluorescent probe (DCFH-DA) used in the assay is
sensitive toward a number of ROS. Details of the assay, extraction protocol, and detection
methodology are explained in Landreman et al. (2008) (Landreman et al. 2008). The quartz
substrates were prebaked at 550°C for 12h and stored in baked aluminum foil prior to sampling.
Elemental and organic carbon (EC/OC) was determined using the Thermal Evolution/Optical
Transmittance analysis (Birch and Cary 1996) and organic compounds were determined using
28

GC/MS (Schauer et al. 1999). Due to limitations in PM mass collection, only fine PM was
analyzed for GC/MS.
3.2 Results and discussion
3.2.1 Mass balance
To reconstruct total PM mass concentration for the red line, gold line, and USC ambient
site, chemical species were grouped into seven categories: water-soluble organic carbon (WSOC)
and water-insoluble organic carbon (WISOC), elemental carbon (EC), inorganic ions, crustal
metals less Fe (CM), elemental Fe, and trace metals. WSOC and WISOC were calculated using a
multiplier of 1.8 to account for the oxygen, nitrogen, and hydrogen associated with organic
carbon (Turpin and Lim 2001). Inorganic ions are the sum of Cl
-
, NO
3
-
, SO
4
2-
, PO
4
3-
, Na
+
, K
+
,
and NH
4
+
. The CM category represents the sum of Al, K, Ca, Mg, Ti, and Si, each of which were
multiplied by appropriate factors to convert to oxide mass (Cheung et al. 2011). Because silicon
data was not acquired, Si was estimated by multiplying Al with a factor of 3.41 (Hueglin et al.
2005). For the purpose of this study, all Fe data is presented as total elemental Fe. Coarse and
fine PM mass reconstructions were calculated in the same manner under the assumption that
chemical species in both modes originate from the same source.
Figure 3.1 shows the mass reconstruction based on the seven identified categories along
with total gravimetric mass concentration for the gold line, red line, and USC ambient site for (a)
coarse and (b) fine PM. Results are based on an average of the two periods sampled, except for
the gold line coarse PM mode, which is a composite of the two periods. Error bars in this study
represent one standard deviation. In coarse PM, the gravimetric mass concentration of the gold
line is approximately 40% of the USC ambient site, while mass concentration for the red line is
29

almost equivalent to USC ambient site. Even though the gold line runs outdoors, the effect of
being inside the train significantly reduces personal coarse PM exposure; however, this effect is
not apparent for the red line, in which a previous study on the L.A. Metro has demonstrated an
additional source of PM in the underground environment (Chapter 2). In fine PM, the red line
gravimetric mass concentration is approximately 70% greater than both the gold line and USC
ambient site concentrations, while concentrations for gold line are only 5% less than USC
ambient site. This suggests that the additional source of PM for the red line has a greater
influence in the fine mode than in the coarse mode, and that being inside the train for the gold
line does not substantially reduce fine PM exposure.
Gold Line Red Line USC
mass concentration ( g m
-3
)
2
4
6
8
10
12
14
Gold Line Red Line USC
mass concentration ( g m
-3
)
10
20
30
40
WSOC
WISOC
EC
Inorganic ions
CM
Fe
Trace metals
Gravimetric mass

Figure 3.1 Mass reconstruction of the 7 identified categories (a) coarse and (b) fine PM. Gravimetric mass
concentrations are also presented.

The most significant difference between the three sites is the abundance of Fe in the
subway environment in both PM modes. In coarse PM, the gravimetric mass of red line, gold
line, and USC ambient site contains 27%, 6%, and 2% Fe; fine PM in the corresponding sites
a b
30

contains 32%, 3%, and 1% Fe. This significant presence of Fe in the subway air has major
implications in terms of personal exposure of subway passengers. The mass reconstruction for
USC ambient site in coarse mode shows that about 48% of the gravimetric mass is unidentified.
This is consistent with a coarse PM study in the L.A. basin that found 20-50% unidentified mass
in urban sites and attributed the fraction to the uncertainty of the OC multiplier coupled with
conversion factors for crustal oxides (Cheung et al. 2011).
3.2.2 Crustal species
Crustal species are elements that are derived from soil origins. They account for a
significant portion of urban aerosols, especially in the coarse mode. Table 3.1 shows the average
concentration of crustal species in coarse and fine PM for the gold line, red line, and USC
ambient site. Except for fine mode Al and Ca, the concentrations of the latter crustal species (Mg,
K, and Ti) in both modes for the red and gold lines are remarkably similar. For the red line

Table 3.1 Average total mass concentrations (ng/m
3
) and ratios relative to USC of crustal and non-crustal metals.

31

sample, crustal species in the coarse mode also exhibit the same patterns as the inorganic ions in
which the concentrations follow each other in the two sequential sampling periods (not shown),
indicating the influence of ambient crustal aerosols. However, crustal species in the fine mode
did not exhibit this trend. It is important to note that the average red line concentrations of fine
mode Al and Ca are greater than corresponding USC concentrations, suggesting these two
species may have an additional non-crustal source, which will be discussed in greater detail in
the following section.
3.2.3 Non-crustal species
Selected non-crustal species concentrations (ng/m
3
) and gold and red line to USC
ambient site concentration ratios are also shown in Table 3.1. The species were selected based on
its elevated concentrations relative to USC concentrations and results from other subway systems
(Salma et al. 2007; Murruni et al. 2009). Numerous studies have determined Fe to be ubiquitous
in subway environments, and are present in elevated concentrations relative to street levels by up
to 50 times (Nieuwenhuijsen et al. 2007). The current study has also observed a number of non-
crustal species, particularly transition metals, to have significantly higher concentrations in the
subway environment than USC ambient levels. It is important to note that the enrichment ratios
of the red line relative to the gold line and USC are greater in the fine mode than in the coarse
mode. For the red line, Fe concentrations (3.0 and 10.6 µg/m
3
in the coarse and fine mode,
respectively) are 12 and 45 times greater than the corresponding USC concentrations and 11 and
22 times greater than those for the gold line. A study in Budapest found Fe concentrations at
station platforms to be 33.5 and 15.5 µg/m
3
for PM
10-2.0
and PM
2.0
, respectively, accounting for
40% and 46% of their corresponding total PM mass (Salma et al. 2007). Consistent with other
worldwide subway studies, Mn, Cr, Co, Ni, Cu, and Ba were observed on the red line to have
32

concentrations that are at least 2 times higher than corresponding USC ambient levels. In
comparison to the gold line personal exposure, passengers on the red line are exposed to
substantially higher levels of most of these trace elements. In addition, Mo, Cd, and Eu have also
been identified to be significantly enriched in the subway environment, especially in the fine
mode. Mo exhibits the greatest enrichment ratios in both modes and both lines; for red line, Mo
concentrations are 113 and 146 times greater than USC levels in coarse and fine mode,
respectively; for the gold line, Mo concentrations are 5 and 6 times greater than USC levels. The
enriched levels of these non-crustal species observed for the red line can be attributed to its
enclosed environment and a significant underground source that has resulted in the accumulation
and subsequent resuspension of PM dust.
Subway dust is primarily generated by the frictional processes of the wheels, rails, and
brakes of the system as well as by the mechanical wearing of these parts. Particles can also be
formed by the condensation of gaseous Fe species from the sparking between the third-rail and
the train (Kang et al. 2008). Stainless steel, which is used for the rail tracks and the main body of
the train for both lines, is an iron-based alloy mixed with chromium and other metallic elements
to enhance its properties. However, the composition of the stainless steel employed by the red
and gold line could not be found as it may be proprietary information of the manufacturer.
Linear regression analysis was conducted for the crustal and non-crustal species to
determine the inter-correlation in coarse and fine PM. Table 3.2 shows the coefficients of
determination (R
2
) of these species. The regression analysis only includes total metals data from
the gold and red line samples (N=7) based on the assumption that the selected non-crustal
species from the gold and red line environment are derived from a different source than
corresponding species from USC ambient site. The number of data points includes 3 coarse PM
33

data points (1 from gold line, 2 from red line) and 4 fine PM data points (2 from gold line, 2
from red line). While Table 3.1 established the elevated concentrations of Al and Ca for the red
line relative to USC ambient site, bivariate linear regression analysis reveals that both species are
strongly correlated with the majority of non-crustal species, suggesting that Al and Ca may have
non-crustal sources for the gold and red line environments in addition to soil-derived sources.
The strong correlation between Al, Ca, Cr, Mn, Fe, Co, Ni, Cu, Mo, Cd, Ba, and Eu suggests that
these elements may share a common source, and may be components of stainless steel used by
the subway and light-rail systems in this study.

Table 3.2 Coefficients of determination (R
2
) of total crustal and non-crustal species. Correlation includes gold and
red line data only for coarse and fine PM (N=7).

Cu exhibits lower R
2
values with the other stainless steel elements, but is still well
correlated and also appears to be clustered with Zn, Ti, K, and Ca. Although Ba is strongly
correlated with the other non-crustal elements (R
2
>0.96), it is not typically used as an alloy in
stainless steel, but has been identified with the wear of brakes (Furuya et al. 2001). Zn is the only
Mg Al K Ca Ti Cr Mn Fe Co Ni Cu Zn Mo Cd Ba Eu
Mg 1
Al 0.94 1
K 0.43 0.42 1
Ca 0.90 0.96 0.40 1
Ti 0.68 0.68 0.92 0.64 1
Cr 0.58 0.74 0.27 0.85 0.44 1
Mn 0.54 0.71 0.24 0.82 0.40 1.00 1
Fe 0.54 0.70 0.22 0.82 0.38 0.99 1.00 1
Co 0.55 0.70 0.21 0.83 0.38 0.99 0.99 1.00 1
Ni 0.59 0.74 0.30 0.86 0.47 1.00 0.99 0.99 0.98 1
Cu 0.52 0.61 0.75 0.67 0.79 0.73 0.70 0.67 0.66 0.76 1
Zn 0.35 0.42 0.92 0.42 0.86 0.40 0.38 0.34 0.33 0.43 0.86 1
Mo 0.52 0.67 0.15 0.80 0.31 0.96 0.97 0.98 0.99 0.95 0.56 0.25 1
Cd 0.54 0.71 0.26 0.82 0.43 0.99 0.99 0.99 0.99 0.99 0.72 0.41 0.96 1
Ba 0.57 0.73 0.26 0.84 0.43 1.00 1.00 1.00 0.99 0.99 0.72 0.40 0.96 0.99 1
Eu 0.56 0.73 0.26 0.83 0.43 1.00 1.00 0.99 0.99 0.99 0.72 0.40 0.95 0.99 1.00 1
Non-crustal species
Crustal
species
Non-crustal species
Crustal species
34

non-crustal species not strongly correlated with other non-crustal species, but instead exhibits
strong correlation with K and Ti. Zn is typically used as a coating on steel for corrosion
protection (Marder 2000), but our regression analysis suggests its elevated concentrations may
be from another source. A subway study in Buenos Aires found that the main source of Zn was
from street-level vehicular emissions (Murruni et al. 2009).
Figure 3.2 shows the crustal enrichment factors (EF) for 22 elements for the gold line, red
line, and USC ambient site in the (a) coarse and (b) fine mode. The crustal ratios are calculated
based on Upper Continental Crust (UCC) values from Taylor and McLennan (1985) (Taylor and
McLennan 1985). Total elemental concentrations are first normalized by Al and then divided by

Figure 3.2 Upper Continental Crustal (UCC) enrichment factors (EFs) for (a) coarse and (b) fine mode using total
elemental concentrations.
35

the relative abundance of the corresponding UCC ratio. A much higher crustal EF indicates
anthropogenic origin, while an EF approaching 1 indicates crustal origin. The elements are
sorted based on the decreasing order of the USC ambient site crustal EFs. The pattern in fine and
coarse mode crustal EFs are remarkably similar to each other. Crustal EFs for Na, La, K, and Mg
in both modes for the gold and red lines are lower than USC ambient site, suggesting the source
is ambient air. In both modes, Mo has the highest crustal EF for the gold and red lines, followed
by Fe, Mn, Ba, Cr, and Ni. It is evident that the source of these enriched elements is substantially
greater on the red line than on the gold line. Table 3.1 showed that concentrations of Ni, Cr, and
Ba for the gold line are lower or similar to corresponding concentrations at the USC ambient site,
but the crustal EF analysis reveals that the EFs are actually greater than USC ambient site EFs by
approximately 2-3 times. This suggests that these elements have indeed been influenced by
additional sources (i.e. rail abrasion, brake wear, etc.), which are not affecting the vicinity of the
USC ambient site.
3.2.4 Reactive oxygen species (ROS) activity
A number of transition metals that have been identified in this study to be present in
elevated concentrations are known to contribute to the generation of reactive oxygen species
(ROS), resulting in oxidative stress (Verma et al. 2010). Table 3.3 shows the ROS coefficients of
determination (R
2
) with WS crustal species, WS non-crustal species, OC, WSOC, EC, and ions.
The analysis includes both PM modes for the red line, gold line, and USC (N=9). Non-crustal
species, namely Fe (R
2
=0.77), Ni (R
2
=0.95), and Cr (R
2
=0.84), show strong correlations with
ROS activity, while most of the crustal species show poor correlations. Although Cd also shows
a strong correlation with ROS activity, Cd is not redox active and its concentrations are below
any toxicity threshold. Of the other PM components, OC exhibited the highest correlation with
36

ROS (R
2
=0.92). A multiple linear regression (MLR) analysis was conducted to further
investigate the contribution of PM components to ROS activity. We elected WS Fe and OC as
our two independent variables to be predictors of ROS activity because they each represent
distinct pollution sources and have also been shown to be redox active based on the ROS assay
in previous studies (Hu et al. 2008; Verma et al. 2010). A fitted least-squares model was
obtained using SigmaPlot resulting in the following equation:
ROS = -16.624 + 0.663 × Fe + 0.0318 × OC
The overall model is statistically significant (p<0.001) and has an adjusted R
2
=0.94; the
constant was reported not to be statistically significant (p>0.05). In addition, the correlation
between the measured and predicted ROS values is excellent (R
2
=0.96; y[predicted
ROS]=0.96x[measured ROS]+3.41). The results suggest that WS Fe, a PM species present in
elevated concentrations in rail environments, and OC, an indicator of ambient vehicular traffic,
can explain 94% of the variance of the measured ROS activity.

Table 3.3 ROS coefficients of determination (R
2
) with WS crustal and non-crustal species and other PM
components. Data includes coarse and fine data for red line, gold line, and USC ambient site (N=9).

37

Figure 3.3 shows the ROS activity of particles for the gold line, red line, and USC
ambient site in ng of Zymosan units (a) per volume (m
3
) and (b) per mass (mg). The per volume
basis is relevant for the personal exposure assessment of passengers, while the per mass basis is a
measure of the intrinsic properties of the particles collected. On a per volume basis, fine PM
accounts for 90-98% of total ROS activity. In addition, ROS activity observed on the red line is
greater than USC ambient site and gold line activity by 65% and 55%, respectively. Even though
total concentrations of ROS-active metals (Fe, Ni, and Cr) in both modes are 4-44 times greater
on the red line than at USC ambient site (Table 3.1), ROS activity differs by less than 2 times.
The opposite is observed when comparing the two PM modes, in which ROS activity differs by
7-29 times and total concentrations of Fe, Ni, and Cr only differ by 2-4 times. Based on these
observations, it is clear that the soluble fraction of the metals plays a dominant role in ROS
activity. On a per mg basis, gold line ROS activity in fine mode is 13% greater than red line and
USC ambient site activity, while red line and USC ambient site ROS activity are comparable.
Our results suggest that one unit of PM mass on the gold line may be as intrinsically toxic as one
unit of PM mass from the red line, however, from a personal exposure perspective, PM
originating from the red line generates greater ROS activity per m
3
of air than PM from the gold
line and at USC ambient site.
38

Figure 3.3 Reactive oxygen species (ROS) activity (a) per air volume (m
3
) and (b) per gravimetric PM mass (mg).

Although our results suggest that subway and light-rail air may not be intrinsically more
toxic than USC ambient air, it should be noted that a more appropriate personal exposure
assessment of transport microenvironments in L.A. should be a comparison with the predominant
mode of commute—private vehicles. Passengers of the L.A. Metro may actually be subjected to
lower levels of PM and toxic co-pollutants such as EC and transition metals. To date, freeway
commuter exposure assessments of on-road PM chemical composition measurements are limited.
My future investigation will provide a comprehensive exposure assessment of “on-road” PM
physical and chemical characterization of trafficked freeways and arterial roads of urban L.A.
This will offer a greater understanding of the health impacts of different transport modes to the
public health.

39

Chapter 4 Size-segregated composition of particulate matter (PM)
in major roadways and surface streets

4.1 Introduction
The novelty of this study lies in the sampling instrumentation used to collect on-road PM.
The use of portable, lightweight battery-operated pumps, which operate at relatively high flow
rates coupled with cascade impactors made it possible to collect size-fractionated PM. A quasi-
isokinetic inlet was deployed to collect time-integrated samples along three routes in Los
Angeles: a major roadway dominated by light-duty vehicles (LDVs), a major roadway with a
higher fraction of heavy-duty vehicles (HDVs), and two major surface streets. Concurrent
samples were collected at a fixed site at the University of Southern California (USC),
representing an urban background site. Other studies have provided measures of roadway
ambient air pollutants which may be representative of a busy traffic area, but the current study
focuses on assessing private commuter exposure by selecting three distinct commute
environments which encompass various traffic volumes, traffic composition, and driving
conditions. Chemical analysis has been performed for three size fractions of PM, including PM
10-
2.5
, PM
2.5-0.25
, and PM
0.25
. The objectives of this study are: 1) provide a chemical comparison of
the three roadway environments and USC background site, 2) discuss factors that may contribute
to the differences in chemical composition of the roadway environments, and 3) compare results
to previous studies that have been conducted at fixed sites near the same roadways.
4.2 Experimental methodology
The in-vehicle sampling campaign was undertaken in Los Angeles in March-April 2011.
Sampling was conducted for 11 hours per day on Monday - Friday, from 6:00 a.m. to 5:00 p.m.
Figure 4.1 shows the three sampling routes that were selected to each represent a distinct
40

roadway environment. The I-110 is a high-traffic freeway that runs 51-km from the Port of Los
Angeles through downtown Los Angeles to Pasadena and is composed mostly of light-duty
vehicles (LDVs); the I-710 is a 43-km freeway and has a higher heavy-duty vehicle (HDV)
composition than the I-110 because it serves as the main corridor for HDVs traveling to and from
the Ports of Los Angeles and Long Beach; Wilshire/Sunset Boulevards is a 48-km major surface
street route that traverses downtown Los Angeles, Koreatown, Miracle Mile, Beverly Hills, and
Hollywood.

Figure 4.1 Map of the three sampling routes: 1) I-110 (blue), 2) I-710 (red), and Wilshire/Sunset (purple). The USC
background site is denoted by the red star.
A summary of the traffic data is shown in Table 4.1, which is taken from the CalTrans
Performance Measurement System (PeMS) database. The total vehicle flow for the I-110 (6378
vehicles/hr) is approximately 1.5 times higher than total flow for the I-710 (4247 vehicles/hr).
However, the I-110 truck flow (243 trucks/hr) is approximately half of the I-710 truck flow (470
trucks/hr), yielding a truck composition of 3.9% and 11.3% for the I-110 and I-710, respectively.
41

In addition, the northern section of the I-110, which is approximately 12-km, is limited to LDVs
only. It is important to note that even though traffic flows may differ for various parts of the
freeway, the traffic sites were carefully selected to be located in the center of the freeway to
characterize the entire freeway. The Wilshire/Sunset route had a total flow of 1,839 vehicles/hr
with a negligible truck flow. Concurrent sampling was conducted at University of Southern
California (USC) and served as our background site (Moore et al. 2007; Ning et al. 2007). Two
samples were collected for the three routes and for the USC background site.

Table 4.1 Summary of meteorological parameters from nearby air quality monitoring sites (South Coast Air Quality
Management District (SCAQMD)). Traffic data is from the CalTrans database.

4.2.1 Sampling instrumentation
The sampling vehicle (Honda Insight Hybrid 2011) was equipped with six Personal
Cascade Impactor Samplers (PCIS) (Misra et al. 2002; Singh et al. 2003), each operated
individually with a battery-powered Leland Legacy pump (SKC Inc., Eighty-Four, PA) at a flow
rate of 9 liters per minute (lpm). The pumps were calibrated with a Gilian Gilibrator-2 Air Flow
Calibrator (Sensidyne Inc., Clearwater, FL) before and after sampling, and pump flows were
checked regularly with flow meters throughout the sampling campaign. Two PCIS units were
operated concurrently at the USC background site. The inlet for roadway air was designed to
have a 90° bend that is characterized by a 50% cutpoint of 10 μm (Peters and Leith 2004). Each
PCIS had two impaction stages – the first with a cut point of 2.5 μm to collect PM
10-2.5
particles
Route
Dates of
sampling
Temp
(deg C)
± RH (%) ±
Prevailing wind
direction
Wind speed
(m/s)
±
Total/truck flow
(vehicle/hour)
Truck
composition (%)
110 S1 3/1-3/8/11 17.8 3.8 54.2 17.0 W 3.2 0.9
110 S2 4/11-4/18/11 18.7 2.5 60.1 13.1 SW 3.1 0.8
710 S1 3/17-3/25/11 15.8 3.2 55.3 16.2 W 3.6 0.9
710 S2 4/19-4/25/11 18.7 1.7 64.5 8.3 W 3.4 0.9
WS S1 3/9-3/16/11 21.7 3.8 48.3 16.0 SW 2.9 1.2
WS S2 4/26-5/2/11 25.8 4.2 35.6 13.4 SW 3.7 1.0
6378/243
1839/NA
4247/470
3.9
11.3
NA
42

and the second with a cut point of 0.25 μm to collect PM
2.5-0.25
particles– and an after-filter stage
to collect PM
0.25
. For the purpose of conducting comprehensive chemical analysis, three PCIS
were loaded with PTFE (Teflon) filters and three were loaded with quartz filters. In the Teflon
units, 25-mm Zefluor-supported PTFE filters (Pall Life Sciences, Ann Arbor, MI) were used as
the impaction substrates and 37-mm PTFE membrane filter with PMP ring (Pall Life Sciences,
Ann Arbor, MI) were used as after-filters; in the quartz units, quartz microfiber filters (Whatman
International Ltd, Maidstone, England) were used as both the impaction substrates and after-
filters. Two DustTrak (Model 8520, TSI Inc., Shoreview, MN) were also deployed to measure
continuous PM
2.5
and PM
10
mass concentrations. However, DustTrak results will not be
presented in this paper.
Roadway air was supplied to the PCIS units and DustTrak monitors by a custom-
designed, stainless steel inlet with an inner diameter of 0.95 cm (3/8 inch). The curved PM
10

inlet directed roadway air into the vehicle window to a manifold with six branches (one for each
PCIS) with the two DustTrak connected to the end (Figure 4.2). Total flow rate through the inlet
was 57.4 lpm (6 PCIS * 9 lpm + 2 DustTrak * 1.7 lpm), yielding a velocity of 13.4 m/s (or 30
mph) at the inlet. Hence, isokinetic sampling was achieved when the vehicle was moving at 13.4
m/s. Anisokinetic sampling may result in the overestimation or underestimation of relatively
larger particles (i.e. PM
10-2.5
) and can be calculated based on the following equation (Belyaev and
Levin 1974):

43

where C is the estimated PM concentration, C
o
is the actual PM concentration, U
o
is the free
stream velocity, or vehicle speed, U is the inlet velocity, and St is Stokes number. When the
vehicle speed is above 13.4 m/s or U
o
> U, sub-isokinetic sampling occurs; conversely, when
vehicle speed is below 13.4 m/s or U
o
< U, super-isokinetic sampling occurs. Considering the
two most extreme cases for d
p
=10 μm, when the vehicle speed is 22.3 m/s (or 50 mph),
C/C
o
=1.34, and when the vehicle speed is 4.5 m/s (or 10 mph), C/C
o
=0.58. For particles smaller
than 5 μm, the resulting over- or under- estimation of concentration in these two extreme cases of
vehicle speed is less than 20%, and decreases rapidly with decreasing particle size. PM
2.5
is
virtually unaffected by anisokinetic sampling.

Figure 4.2 Sampling schematic of the inlet into the vehicle and the instrumental set up.

4.2.2 Sample analysis
After completion of sampling, Teflon filters were analyzed gravimetrically and
chemically while quartz filters were only analyzed chemically. The Teflon filters were
equilibrated for 24 hours, then pre- and post-weighed in a temperature and relative humidity-
44

controlled room using a MT5 Microbalance (Mettler-Toledo Inc., Columbus, OH) to determine
gravimetric mass concentrations. The Teflon filters were analyzed by magnetic-sector
Inductively Coupled Plasma Mass Spectroscopy (SF-ICPMS) using acid extraction to determine
total elemental composition (Zhang et al. 2008) and ion chromatography (IC) to determine the
PM concentrations of inorganic ions (Kerr et al. 2004). The quartz substrates were prebaked at
550°C for 12 hours and stored in baked aluminum foil prior to sampling. Elemental and organic
carbon (EC/OC) content was determined using the Thermal Evolution/Optical Transmittance
analysis (Birch and Cary 1996). Water extracts of OC yielded WSOC. Organic speciation
analysis was conducted with gas chromatography mass spectroscopy (GC/MS) (Schauer et al.
1999). However, organics and metals speciation results are not discussed here and will be the
topic of upcoming manuscripts.
4.3 Results and discussion
4.3.1 Overview of campaign
Table 4.1 provides the dates of sampling, a summary of the meteorological conditions,
and the traffic flows on the sampled roadways. Meteorological data are obtained from the nearby
Downtown Los Angeles and Long Beach air quality monitoring sites that are maintained by the
South Coast Air Quality Management District (SCAQMD). Data presented are for the
corresponding dates and times of sampling. Variation in both temperatures (°C) and relative
humidities (%) were expected considering the daily sampling duration (6 a.m. to 5 p.m.). The
prevailing wind direction during sampling hours was consistent across all samples with an
onshore breeze from W or SW. Average wind speed (m/s) was also relatively consistent.
45

Figure 4.3 shows the size-fractionated PM mass concentrations of the two samples
collected for each route and for USC background site. In terms of PM size composition, no PM
mode is dominant and PM
10-2.5
, PM
2.5-0.25
, and PM
0.25
account for an average of 37.2±4.1%,
28.3±3.9%, and 34.6±3.6% of total PM mass, respectively, across all sites. Although there is
some variation in total mass concentrations between the two sets of samples, the USC
background site has the lowest levels of 24.1±2.8 μg/m
3
while the three sampled roadways
exhibited comparable levels of 32.2±3.32 μg/m
3
.

Figure 4.3 Size-fractionated mass summary the three roadways and USC background site. S1 and S2 represent the
two sets of samples collected, with a sampling duration of approximately 50 hours for each set. Sampling dates are
shown in Table 4.1.

46

4.3.2 Mass balance
Figure 4.4a-c shows the mass balance for each site (I-110, I-710, Wilshire/Sunset, and
USC) and each size fraction. Gravimetric mass concentrations are shown with error bars
representing one standard deviation. The mass balance was constructed based on 5 categories:
inorganic ions, organic matter (OM), elemental carbon (EC), crustal metals, and trace metals.
Inorganic ions consist of chloride, nitrate, phosphate, sulfate, sodium, ammonium, and potassium.
Organic matter is calculated by multiplying OC by a correction factor of 1.8 (Turpin and Lim
2001). Crustal metals include Mg, Al, K, Ca, Ti, and Fe multiplied by a corresponding factor to
account for oxide forms, with the exception of Si which is derived from Al (Chow et al. 1994;
Marcazzan et al. 2001; Hueglin et al. 2005). Trace metals include the remaining elements.
Individual metal mass concentrations (ng/m
3
) for the three roadway environments and for USC
background site are shown in Table 4.2, and will be discussed in more detail in Chapters 4 and 5.

a. PM
10-2.5
b. PM
2.5-0.25
c. PM
0.25

Figure 4.4 Mass balance constructed based on five identified categories for a) PM
10-2.5
, b) PM
2.5-0.25
, and c) PM
0.25
.
Error bars represent one positive standard deviation.
47

Table 4.2 Mass concentrations of metals (ng/m
3
). For PM
10-2.5
, N=2 with standard deviations; for PM
2.5-0.25
and
PM
0.25
, N=1.

4.3.2a PM
10-2.5

In PM
10-2.5
, USC exhibited the lowest gravimetric and reconstructed mass concentrations
while Wilshire/Sunset exhibited the highest in both (Figure 4.4a). In spite of the PM mass
differences, the composition of the five designated PM components are relatively consistent
across all routes and USC, with crustal metals accounting for 40-50% of reconstructed mass, ions
I-110 ± I-710 ±
Wilshire/
Sunset
± USC ± I-110 I-710
Wilshire/
Sunset
USC I-110 I-710
Wilshire/
Sunset
USC
Li 0.11 0.03 0.09 0.01 0.13 0.03 0.12 0.03 0.29 0.21 0.39 0.32 0.13 0.08 0.15 0.11
B 0.31 0.04 0.39 0.05 0.25 0.03 0.17 0.03 0.47 0.31 0.39 0.40 1.92 0.94 2.49 1.11
Mg 83.51 31.29 81.39 1.61 135.22 36.23 131.22 16.03 70.71 67.38 78.18 46.00 28.22 19.04 25.29 21.46
Al 151.81 49.86 147.60 29.13 224.66 50.56 149.13 76.88 97.24 113.60 80.06 39.25 100.19 68.56 90.20 101.64
P 5.20 0.63 4.72 1.05 3.80 2.24 1.83 1.45 10.64 8.76 6.90 7.28 9.03 7.88 7.04 4.74
S 114.07 26.20 109.38 19.00 124.97 9.60 109.59 1.71 264.42 191.58 210.31 312.35 336.90 231.42 221.81 286.22
K 82.56 24.10 89.01 11.78 99.20 10.74 99.98 14.35 56.27 54.40 49.19 49.18 61.19 34.82 52.06 38.43
Ca 206.40 37.27 203.28 8.89 272.08 24.61 192.02 64.12 113.93 131.81 97.32 56.82 106.66 77.71 97.37 68.60
Ti 31.99 0.48 26.28 3.13 35.30 2.06 21.13 0.44 23.66 20.97 12.35 7.82 22.84 13.45 15.72 8.53
V 0.54 0.10 0.74 0.03 0.58 0.04 0.49 0.05 0.56 0.73 0.45 0.60 0.86 1.19 1.01 1.04
Cr 2.70 0.10 2.00 0.14 3.68 0.18 1.04 0.31 2.02 1.92 1.48 0.66 2.10 1.52 1.79 2.18
Mn 6.93 0.45 6.40 0.11 7.73 0.05 3.78 0.36 5.65 5.74 3.72 2.49 4.32 3.42 3.73 2.34
Fe 765.05 139.84 651.78 31.05 893.98 40.27 318.14 15.33 598.85 598.40 338.31 136.67 356.75 267.31 328.50 134.49
Co 0.12 0.00 0.17 0.07 0.14 0.00 0.19 0.10 0.09 0.10 0.08 0.06 0.08 0.07 0.08 0.10
Ni 1.01 0.41 1.13 0.15 0.96 0.16 0.55 0.03 0.59 0.60 0.35 0.34 1.15 1.01 0.56 1.24
Cu 47.86 9.00 28.82 0.05 53.54 4.46 12.22 2.57 36.48 23.17 21.97 5.97 23.92 13.63 21.37 7.98
Zn 11.12 0.95 10.77 1.45 12.45 0.66 6.34 0.72 11.48 11.25 9.82 6.01 18.26 12.97 10.72 8.33
As 0.14 0.00 0.12 0.07 0.15 0.03 0.10 0.02 0.31 0.17 0.16 0.19 0.35 0.23 0.33 0.27
Rb 0.24 0.07 0.24 0.04 0.26 0.03 0.25 0.02 0.15 0.16 0.09 0.11 0.16 0.09 0.13 0.11
Sr 3.88 0.65 3.60 0.72 3.81 0.12 2.40 0.53 3.03 2.66 1.70 0.93 4.42 1.53 1.65 0.87
Mo 1.65 0.33 1.03 0.02 2.38 0.15 0.48 0.12 1.50 1.12 1.11 0.34 1.15 1.06 1.23 0.46
Ag 0.05 0.02 0.05 0.01 0.03 0.01 0.09 0.10 0.05 0.12 0.04 0.02 0.10 0.03 0.13 0.10
Cd 0.05 0.00 0.05 0.00 0.07 0.00 0.03 0.02 0.07 0.06 0.06 0.04 0.09 0.06 0.07 0.05
Sn 2.70 0.84 0.59 0.02 6.04 0.09 0.00 0.00 4.24 2.30 3.55 1.65 3.68 1.83 3.93 2.06
Sb 7.33 1.42 3.93 0.00 7.35 0.61 1.77 0.40 6.39 4.07 3.35 1.12 4.12 2.45 3.74 1.25
Ba 55.25 16.70 45.00 3.72 61.76 0.48 17.95 5.93 53.37 57.07 24.91 7.84 28.36 23.46 21.52 7.05
La 0.15 0.01 0.15 0.01 0.18 0.02 0.18 0.12 0.17 0.22 0.12 0.13 0.10 0.09 0.09 0.15
Ce 0.50 0.11 0.42 0.02 0.41 0.02 0.26 0.11 0.40 0.45 0.19 0.14 0.29 0.22 0.19 0.14
Pr 0.03 0.00 0.02 0.00 0.03 0.00 0.02 0.01 0.02 0.02 0.01 0.01 0.01 0.01 0.01 0.01
Nd 0.08 0.01 0.08 0.00 0.09 0.01 0.06 0.02 0.06 0.07 0.03 0.02 0.05 0.03 0.04 0.03
Eu 0.07 0.02 0.06 0.00 0.08 0.00 0.03 0.01 0.07 0.08 0.03 0.01 0.04 0.03 0.03 0.01
W 0.06 0.02 0.14 0.06 0.08 0.01 0.07 0.03 0.09 0.15 0.07 0.07 0.11 0.14 0.12 0.09
Pb 1.14 0.03 1.88 0.53 1.55 0.24 1.19 0.26 2.12 2.11 2.37 2.10 2.61 1.87 2.55 2.15
PM
2.5-0.25
PM
0.25
PM
10-2.5
48

for 30-40%, OM for 15-25%, trace metals for 3%, and EC for 1%. This consistency
demonstrates that the differential microenvironment of each roadway does not play a substantial
role in the composition of PM
10-2.5
. This can also be explained by the fact that the major sources
of PM
10-2.5
are not from traffic sources (i.e. vehicular combustion), but rather from resuspension
of road dust, large sea salt particles, mechanically generated particles, and biogenic particles (i.e.
pollen, fungal spores).
The unidentified mass in this mode is the largest of the three PM size fractions, with I-
110 having the lowest of 26% unidentified mass and USC site having the highest of 41%. This
underestimation of reconstructed mass calculations can be attributed to the uncertainty in the OC
multiplication factor and other mineral compounds that may exist in oxide form. Previous studies
of this size fraction in urban areas of Los Angeles have also found that crustal metals and
inorganic ions account for the majority of the mass balance (Sardar et al. 2005; Arhami et al.
2009) and that PM
10-2.5
may be characterized by a relatively high percentage of unidentified mass
(Cheung et al. 2011).
4.3.2b PM
2.5-0.25

Figure 4.4b shows the PM composition for PM
2.5-0.25
for the 3 roadways and USC
background site. Contrary to the PM
10-2.5
composition, there is more variability in some of the
PM components, notably the inorganic ions, which account for 35% of reconstructed mass at I-
710 and 69% at USC background site. Crustal metals also exhibited variability across the sites,
with 11% for USC and 32% for I-710, and OM ranged from 14-26% with the lowest fraction at
USC. This variability in composition suggests that this mode may be more influenced by sources
in the immediate environment (i.e. vehicular resuspension, tire and brake wear). Inorganic ions
49

are primarily formed in the atmosphere in the presence of gaseous precursors (SO
2
, NO
2
), and
thus secondary in nature (Seinfeld and Pandis 2006). Previous studies have also found that PM in
PM
2.5-0.25
is dominated by inorganic ions (primarily sulfate and nitrate) in various regions of Los
Angeles (Sardar et al. 2005; Arhami et al. 2009). Although there appears to be a greater
difference in inorganic ions between the various roadways and USC background site, statistical
analysis showed that the differences are not statistically significant (p=0.17). EC and trace metals
are low contributors in this mode, accounting for 1-2% and 4-5%, respectively. Finally, 15-25%
of gravimetric mass concentrations were unidentified for the roadway sites while USC exhibited
an overestimation of 10% of mass.
4.3.2c PM
0.25

In PM
0.25
, USC exhibits the lowest total PM mass concentrations (Figure 4.4c). On
average, the three roadway environments have gravimetric PM concentrations that are
48.0±9.4% higher than levels observed at USC (p<0.05), which strongly suggests that vehicular
emissions from both LDVs and HDVs are major contributors to PM in this mode. OM is the
dominant component and its contribution is consistent across all sites, accounting for 50-60% of
reconstructed mass. A previous study in Los Angeles also determined OM to be the dominant
contributor of PM
0.25
(Arhami et al. 2009). Contrary to the larger size fractions, EC exhibited
substantial variation across the three roadway environments, accounting for 15% for the I-710,
10% for the I-110, 6% for Wilshire/Sunset and 6% for USC. This observation is consistent with
the application of using EC as a surrogate for primary emissions and as a tracer of diesel fuel
combustion (Schauer 2003). The significantly higher contribution of EC in the two freeway
environments relative to Wilshire/Sunset and USC (p<0.05) can be attributed to the fleet fraction
of HDVs, of which the I-710 has the highest of 11.3%, the I-110 at 3.9%, and Wilshire/Sunset
50

with a negligible HDV fraction. Of the three size fractions, PM
0.25
appears to be most heavily
influenced by traffic emissions. Inorganic ions’ contribution exhibited slight variation, with the I-
710 at a low of 11% and USC at a high of 21%; crustal metals’ contribution ranged from 12% at
I-710 and 19% at USC; trace metals’ contribution were consistently low at 3-5% at all sites.
PM
0.25
is characterized by the lowest fraction of unidentified mass, with 6% for the I-110 and
19% for both the I-710 and Wilshire/Sunset.
The authors acknowledge that quartz filters are subject to both positive (adsorption) and
negative (volatility) artifacts of OC, especially for particles collected through filtration (i.e.
PM
0.25
) (Turpin et al. 2000). However, significant organic adsorption and volatility artifacts are
not expected in this study due to the very high PM mass loadings (Kim et al. 2001) and the low
pressure drop of the PCIS (Misra et al. 2002). Particles in PM
10-2.5
and PM
2.5-0.25
are collected by
impaction and are thus less susceptible to sampling artifacts (Zhang and McMurry 1987).
4.3.3 Inorganic ions
Figure 4.5 shows the inorganic ions concentrations for five inorganic ions (nitrate, sulfate,
sodium, chloride, and ammonium). Ion balance calculations showed that particles in PM
2.5-0.25

and PM
0.25
were almost fully neutralized; however, the calculation was not possible for PM
10-2.5
without information for calcium and carbonate.
4.3.3a PM
10-2.5

In PM
10-2.5
, the average of total inorganic ion concentrations of the four sites is 2.5±0.2
μg/m
3
(Figure 4.5a). Nitrate is the dominant component and its contribution to total ions ranges
from 31.9% on the I-710 to 50.1% for the I-110. A previous study in Los Angeles in this size
51

Figure 4.5 Inorganic ions concentrations for a) PM
10-2.5
, b) PM
2.5-0.25
, and c) PM
0.25
. Error bars represent one standard
deviation.

fraction also found nitrate to be the major component of inorganic ions (Cheung et al. 2011).
Given the southwesterly sea breeze from the Pacific Ocean, sea salt is considered to be a
contributor to PM
10-2.5
and its primary components, sodium and chloride, have concentrations of
0.9-1.2 μg/m
3
and make up 37-49% of total inorganic ions. On average, sea salt levels are
a
c
b
52

27.7±11.0% higher on roadways relative to USC levels, however, the difference is not
statistically significant (p=0.43). Sulfate concentrations are relatively low (0.27-0.29 μg/m
3
) and
constitute 10-12% of total ions across all sites.
4.3.3b PM
2.5-0.25

It is clear from the mass balance that inorganic ions dominate the PM
2.5-0.25
mode (Figure
4.4b) and that its concentrations are highest in this mode (p<0.05) (Figure 4.5b). The I-110, I-710,
Wilshire/Sunset, and USC sites have total inorganic ions concentrations of 3.9±0.5 μg/m
3
,
2.4±1.6 μg/m
3
, 3.3±0.1 μg/m
3
, and 5.0±1.4 μg/m
3
, respectively. However, the differences in
inorganic ions concentrations are not statistically significant (p=0.17). With the exception of
chloride, the concentrations of inorganic ions are comparable to or lower than USC levels.
Although the total concentrations vary, the individual inorganic ion contributions are relatively
consistent across all sites, with nitrate dominating and accounting for 44-52%, followed by
sulfate (17-29%) and sodium (10-18%).
4.3.3c PM
0.25

Of the three size fractions, the PM
0.25
has the lowest total inorganic ion concentrations,
with an average of 1.29±0.14 μg/m
3
across the four sites (Figure 4.5c). Overall, there is not a
substantial amount of variation in concentration and composition among the three roadway
environments. Sulfate is the dominant component and contributes 52-58% of total inorganic
compounds across all sites, followed by ammonium (18-29%) and nitrate (8-19%). Previous
studies in the Los Angeles basin also found sulfate to be the major inorganic component in
PM
0.25
(Minguillon et al. 2008; Arhami et al. 2009). Similar to PM
10-2.5
and PM
2.5-0.25
, the
53

inorganic ions concentrations in PM
0.25
at the roadway environments are generally comparable or
less than USC background levels.
4.3.4 EC and OC
Figure 4.6 shows the average total carbon (TC), EC, OC, and water-soluble OC (WSOC)
concentrations for the four sites. TC is the sum of EC and OC (Figure 4.6a). PM
0.25
, PM
2.5-0.25
,
and PM
10-2.5
constitute 67-71%, 16-19%, and 11-15% of total TC, respectively. The I-710
exhibits the greatest TC concentration of 9.3±2.2 μg/m
3
and USC has the lowest of 4.1±0.2
μg/m
3
. In terms of enrichment ratios relative to USC, the I-710, I-110, and Wilshire/Sunset are
2.2±0.4, 1.7±0.2, and 1.7±0.1, respectively. Of TC, OC accounts for 78-91% and EC accounts
for 9-22%.
EC concentrations are significantly higher in PM
0.25
relative to the other two size
fractions (p<0.05), and 80-95% of total EC is accounted for in this mode for all four sites (Figure
4.6b). The most substantial difference among the three roadway environments is the EC levels
observed on the I-710, of which EC concentrations are 2.1±0.2 μg/m
3
and 4.1 times greater than
USC, while levels on the I-110 and Wilshire/Sunset are 1.1±0.2 μg/m
3
and 0.6±0.02 μg/m
3
and
2.1 and 1.3 times greater, respectively. Given that EC is a primary pollutant and can be used as a
tracer for diesel emissions at least in the Los Angeles Basin, the differences in EC concentrations
may be explained by the truck density in each roadway. The I-710 and I-110 had truck flows of
470/hr and 243/hr, respectively, for the periods of sampling (CalTrans). Although the I-110 has a
greater total vehicle flow than the I-710, the near two-fold difference in truck flows highlights
the significant contribution of HDVs to EC. The Wilshire/Sunset traffic flow consists mainly of
passenger cars and few compressed natural gas (CNG) buses, which emit lower levels of PM and
54

gaseous pollutants than conventional diesel buses (Hesterberg et al. 2008). The negligible HDV
flow is evident in the comparable levels of EC on Wilshire/Sunset to the USC background site.

Figure 4.6 Size-segregated concentrations of a) total carbon (TC), b) elemental carbon (EC), c) organic carbon (OC),
and d) water-soluble OC (WSOC).

Similar to EC, OC is significantly higher in PM
0.25
(p<0.05), of which 62-67% of total
OC is accounted for in all 4 sites, while in PM
2.5-0.25
and PM
10-2.5
, OC accounts for 17-21% and
13-17%, respectively (Figure 4.6c). Previous studies of ultrafine PM, PM
0.18
, in Los Angeles also
found OC to be the dominant PM component (Hughes et al. 1998; Arhami et al. 2009). Total OC
concentrations for the I-110, I-710, Wilshire/Sunset, and USC are 6.1±0.8 μg/m
3
, 7.3±2.0 μg/m
3
,
6.2±0.1 μg/m
3
, and 3.6±0.04 μg/m
3
, respectively; enrichment ratios for the roadways relative to
a
b
c d
55

USC are 1.7±0.2, 2.0±0.5, and 1.7±0.1, respectively. Organic mass originates from a mix of both
primary (vehicular emissions) and secondary sources (photochemical reactions) (Kim et al.
2002). The more than two-fold difference in OC concentrations observed on the three roadway
environments shows the significant contribution of vehicular traffic to OC.
Figure 4.6d shows WSOC concentrations, which are also significantly higher in PM
0.25

(p<0.05). Total WSOC concentrations for the I-110, I-710, Wilshire/Sunset, and USC are
1.8±0.02 μg/m
3
, 1.5±0.1 μg/m
3
, 2.2±0.2 μg/m
3
, and 1.1±0.03 μg/m
3
, and accounts for 29.4±3.8%,
21.9±5.6%, 36.3±3.1%, and 29.8±0.8% of total OC. The relatively low WSOC fractions
observed are consistent with a previous study that showed OC in urban settings is less oxidized
than OC in rural settings, and thus less water-soluble (Salma et al. 2007).
4.3.5 Comparison to previous studies in Los Angeles
Figure 4.7 shows a comparison of PM
2.5
mass, OC, and EC concentrations from the
current study to previous studies conducted at fixed sites near the I-710 and I-110 (Kuhn et al.
2005a; Kuhn et al. 2005b; Phuleria et al. 2007). In Phuleria et al. (2007), the I-710 study was
conducted directly adjacent to the roadway, at approximately 10 m from the centerline of the
freeway and sampling took place in February and March 2006 from 11 a.m. to 7 p.m. Kuhn et al
(2005a) was conducted in winter 2005 (January) and Kuhn et al. (2005b) was conducted in
summer 2004 (May). However, the sampling site for these two studies was in the northern
“gasoline-only” portion of the I-110 where HDVs were not allowed, while results from the
current study are based on a 3.9% HDV fraction. OC concentrations in the ultrafine fraction were
omitted due to positive adsorption artifacts from the Kuhn et al. (2005a and 2005b) studies.
Differences in PM mass and OC concentrations are not substantial for either roadway
56

comparisons. However, average EC concentrations seem to have decreased by approximately
50% for both roadways in the last 5 years.

*OC in ultrafine fraction for Kuhn 2005 studies are omitted due to positive adsorption artifacts
Figure 4.7 Comparison of PM
2.5
concentrations of a) mass and b) OC and EC to previous studies conducted at fixed
sites in the vicinity of the I-110 and I-710.

This improvement may be partially explained by differences in sampling conditions (i.e.
time of year), but is most likely the result of the Port of Los Angeles Clean Truck Program,
which calls for a progressive ban on older polluting drayage trucks (San Pedro Bay Ports Clean
Air Action Plan (CAAP)). In October 2008, all pre-1989 drayage trucks were banned from
entering the Port and in January 2010, all 1989-1993 drayage trucks were banned in addition to
1994-2003 drayage trucks that were not retrofitted. Finally, in January 2012, drayage trucks that
do not meet the 2007 Federal Clean Truck Emission Standards will be banned. The impact of the
program has been regional, and its effectiveness is illustrated in the decrease in EC levels of the
current I-110 study (with 3.9% HDVs) in comparison to the previous I-110 study (no HDVs).
This is consistent with a study that observed a reduction in BC emission factors after the
a b
57

implementation of diesel particle filter (DPF) retrofit and truck replacement program in 2010 at
the Port of Oakland (Dallmann et al. 2011).
4.4 Conclusion
An in-vehicle sampling campaign was conducted to assess the on-road chemical
composition of PM in three roadway environments (I-110, I-710, and Wilshire/Sunset) for three
size fractions (PM
10-2.5
, PM
2.5-0.25
, and PM
0.25
). Based on average PM concentrations, PM
0.25
is
heavily influenced by vehicular emissions, which is evident in its substantial contribution to TC,
including both EC and OC components, while PM
2.5-0.25
and PM
10-2.5
are less impacted by on-
road sources. Inorganic ions compositions (%) were found to be relatively consistent across the
three roadways. Although concentrations in PM
2.5-0.25
and PM
0.25
at USC were higher than
corresponding levels at roadways, the differences were not significant (p=0.17). The most
notable finding from this study is the elevated levels of EC on the I-710, which were 4.0 times
greater than the USC ambient site, while the I-110 and Wilshire/Sunset roadways were
approximately 2.1 and 1.2 times greater, respectively. In comparison to previous studies, EC
levels appear to have decreased substantially on both the I-110 and I-710, which may be
explained by the effectiveness of the Port of Los Angeles Clean Truck Program that began in
2008.
58

Chapter 5 On-road emission factors of PM pollutants for light-duty
vehicles (LDVs) based on real-world urban street driving
conditions

5.1 Introduction
In this study, an on-road sampling campaign was conducted on two major surface streets
in Los Angeles to assess the effect on PM emissions of urban street driving conditions, which are
essentially intermittent stop-and-go driving. The major advantage of the on-road sampling
methodology deployed in the current study is that it captures PM mass using a driving scheme
that is much closer to urban driving conditions than the aforementioned studies. The
dynamometer studies are based on standardized test cycles of individual vehicles and roadside
samples can only cover localized driving conditions. The frequent acceleration and deceleration
of a fleet of vehicles is expected to emit higher levels of PM than cruise conditions (Yanowitz et
al. 1999). For the surface streets selected, the vehicle fleet is almost entirely composed of LDVs
(passenger cars). Fuel-based emission factors of PM components species are determined and
then compared to previous LDV studies. Lastly, n-alkane concentrations are discussed and
carbon preference indices (CPIs) are calculated to assess the anthropogenic source contribution
to n-alkanes. Results from this study are of great significance as they represent driving
conditions on heavily trafficked surface streets typical of many urban locations throughout the
United States.

59

5.2 Experimental methodology
5.2.1 Sampling route
The on-road sampling campaign was conducted on two major surface streets in Los
Angeles, CA – Wilshire and Sunset Boulevards. The route was selected because it represents two
important streets that are heavily trafficked on a daily basis and traverses major parts of Los
Angeles including Hollywood, Miracle Mile, Beverly Hills and Downtown Los Angeles. The
sampling route is approximately 47 km (29 mi), with each loop requiring 2-5 hours of sampling
depending on traffic congestion. Concurrent sampling was conducted at a site located at the
University of Southern California (USC), representing an urban background site which has been
well documented in previous studies (Moore et al. 2007; Ning et al. 2007). Figure 5.1 shows the
sampling route and the USC background site. Due to the time- and resource-intensive nature of
this campaign, only two samples were collected, of which each sample represents over 50 hours
of on-road PM sampling. Sampling was conducted on weekdays from 6am to 5pm; sample 1 was
collected March 9-16, 2011 and sample 2 was collected on April 26-May 2, 2011. While the
authors acknowledge that the sample size is small which may not provide an accurate estimate of
the variability and uncertainty of on-road PM species presented in this study, the sampling period
is very long and each sample corresponds to a fairly wide range of meteorological and traffic
conditions of Los Angeles.
60

Figure 5.1 Map of sampling route (Wilshire/Sunset Boulevards) and USC background site.

5.2.2 Sampling methodology and analysis
A summary of the sampling methodology is described in this text. For a complete
description, please refer to section 4.3. The mobile laboratory used in this study was a Honda
Insight Hybrid 2011 vehicle, modified with a window insert for the aerosol inlet. PM mass was
collected onto filter substrates for three size fractions: PM
10-2.5
(10 > d
p
> 2.5 μm), PM
2.5-0.25
(2.5
> d
p
> 0.25 μm), and PM
0.25
(d
p
< 0.25 μm). Roadway PM was collected onto six Personal
Cascade Impactor Samplers (PCIS) (SKC Inc., Eighty-Four, PA) (Misra et al. 2002) loaded with
two impaction stages and an after filter through a PM
10
(d
p
< 10 μm) inlet into the vehicle. Each
PCIS was operated with a battery-powered Leland Legacy pump (SKC Inc., Eighty-Four, PA) at
a flow rate of 9 liters per minute. The resulting aerosol velocity at that flow rate through the inlet
is 13.4 m/s (30 mph). Hence, isokinetic sampling was achieved when the vehicle was moving at
13.4 m/s. Anisokinetic sampling may result in the overestimation or underestimation of relatively
larger particles (i.e. PM
10-2.5
) due to sub-isokinetic flows at the aerosol inlet. For d
p
= 10 μm, the
estimated PM concentration is calculated to be 1.2 times greater than the actual concentration
61

when vehicle speed is 17.9 m/s (40 mph) whereas the estimated PM concentration is 0.6 times
the actual concentration when vehicle speed is 4.5 m/s (10 mph) (Belyaev and Levin 1974). For
particles smaller than 5 μm, the expected over- or under- estimation of concentration in these two
cases of vehicle speed is less than 20%, and decreases with decreasing particle size. PM
2.5
is
virtually unaffected by anisokinetic sampling. When the vehicle is stopped, there are no
anisokinetic effects as the aerosol stream (roadway air) velocity is 0 m/s.
For the purpose of chemical analysis, two types of impaction substrates were used –
PTFE (Pall Life Sciences, Ann Arbor, MI) and quartz (Whatman International Ltd, Maidstone,
England) filters. The filters were stored overnight in a refrigerator until they were to be used the
next sampling day. To determine total metal and trace elemental composition, the PTFE filters
were extracted with acid to be analyzed by magnetic-sector inductively coupled plasma mass
spectrometry (SF-ICPMS) (Zhang et al. 2008). Elemental and organic carbon (EC and OC,
respectively) was determined from the quartz substrates using thermal evolution/optical
transmittance analysis (Birch and Cary 1996). Water extracts of OC yielded WSOC. Organic
speciation was conducted with gas chromatography mass spectroscopy (GC/MS) (Schauer et al.
1999).
5.2.3 Emission factors calculation
The emission factors presented in this study are based on fuel consumption. They are
defined as the mass of pollutant emitted per mass of fuel burned and are calculated based on the
following equation (Geller et al. 2005; Phuleria et al. 2006; Ning et al. 2008):
(1)
62

where E
P
is the emission factor of pollutant P expressed in mg/kg of fuel; [P]
st
and [P]
bg
are
pollutant concentrations in μg/m
3
on streets and at the background site, respectively; [CO
2
] is the
concentration of CO
2
in μg of C/m
3
; w
c
is the carbon weight fraction of the considered fuel,
which is 0.85 for gasoline. This equation assumes that the carbon mass from vehicular exhaust is
mostly in the form of CO
2
, and that the mass of other carbon-containing compounds (i.e. CO,
black carbon, hydrocarbons) are negligible relative to the total emitted carbon mass (Yli-Tuomi
et al. 2005). A similar approach has been used in numerous studies to determine emission factors
of various traffic environments including tunnels and freeways (Phuleria et al. 2006). It should
be noted that the emission factors for species not associated with roadway emissions may not be
accurately calculated from Eq. (1), the results of which should be interpreted with caution. In
addition, PM species with very similar levels to corresponding background levels are not an
important part of the roadway emissions profile.
5.3 Results and discussion
5.3.1 Emission factors for major PM components and species
Table 5.1 shows the emission factors of major PM components (mass, OC, EC, and
WSOC) and metals and trace elements for three size fractions (PM
10-2.5
, PM
2.5-0.25
, and PM
0.25
).
The results are presented as the average of the 2 samples and its uncertainty. Emission factors for
species with concentrations similar to or lower than corresponding background levels are not
shown in the table. Overall, the PM
0.25
fraction has the greatest emission factors in all of the
major PM components. The most substantial difference between the three size fractions is the
OC emission factors, of which the PM
0.25
emission factors are 3.9 and 4.8 times greater than
PM
10-2.5
and PM
2.5-0.25
emission factors, respectively. WSOC accounts for 55% of OC for PM
0.25
,
63

while WSOC only accounts for 28% and 25% of OC, respectively. The relatively low emission
factors for EC, which can be used as a surrogate for diesel emissions at least in urban settings
(Schauer 2003) are expected, given that the traffic composition on the Wilshire/Sunset route
consists primarily of LDVs. Chapter 4 reported the EC concentrations on the current route are
approximately 3 times lower than the I-710, a freeway with 12% HDVs, and approximately 2
times lower than the I-110, a freeway with 4% HDVs.

Table 5.1 Fuel-based emission factors (mass of pollutant emitted per kg of fuel) of PM components and metals and
trace elements for three PM size fractions. Pollutants with concentrations close to or less than USC background
levels have been omitted.
Wilshire/Sunset ± Wilshire/Sunset ± Wilshire/Sunset ±
PM components (mg/kg of fuel)
mass 75.8 18.8 62.7 35.8 99.4 98.9
OC 12.6 2.43 10.3 2.59 49.1 5.58
WSOC 3.51 0.82 2.59 0.57 26.8 5.20
EC - - - - 5.29 4.09
metals and trace elements ( μ g/kg of fuel)
Mg 582 824 909 180 108 68.6
Al 2360 3330 1150 234 - -
S 436 227 - - - -
K 24.9 35.2 - - 383 296
Ca 2270 2530 1140 267 808 280
Ti 401 49.4 128 33.3 202 40.7
Cr 74.6 4.55 23.2 3.12 - -
Mn 111 9.73 34.6 7.41 38.9 7.25
Fe 16300 554 5700 641 5450 680
Ni 11.4 5.26 - - - -
Cu 1170 42.6 452 32.7 376 32.7
Zn 173 3.14 108 35.8 67.3 43.4
Sr 40.1 11.8 21.8 4.99 21.9 4.7
Mo 53.9 0.38 21.8 3.08 21.4 3.33
Sb 158 4.56 63.1 5.43 70.0 6.26
Ba 1240 165 482 54.7 406 47.5
Eu 1.65 0.324 0.615 0.073 0.551 0.063
W 0.46 0.65 0.083 0.118 0.827 0.09
Pb 10.4 14.3 7.53 6.95 11.3 7.3
PM
10-2.5
PM
0.25
PM
2.5-0.25
64

For metals and trace elements, the emission factors are highest in the PM
10-2.5
fraction,
followed by PM
2.5-0.25
and lowest for the PM
0.25
fraction, consistent with previous studies in the
Los Angeles basin that have found metals and trace elements to contribute most to larger size
fractions of PM
10
(Sardar et al. 2005; Krudysz et al. 2008). In general, emission factors for
metals and trace elements in PM
10-2.5
are 2-5 times higher than corresponding emission factors in
PM
2.5-0.25
and PM
0.25
. The most notable observations are the emission factors of Fe, which are
substantially higher than all other metals and trace elements in the three size fractions. Al, Ca,
Cu, and Ba also have relatively high emission factors (>1000 μg/kg of fuel), followed by Mg, S,
Ti, Mn, Zn, and Sb (>100 μg/kg of fuel).
In addition to road dust resuspension, numerous studies have associated metals with
brake lining (Cu, Ba, Sb), automotive lubricant oil (Ca), and vehicular wear debris (Fe, Zn, Pb, S)
(Sternbeck et al. 2002; Harrison et al. 2003; Lough et al. 2005). Catalytic converters can also be
a source of precious metals (Pd, Pt, Rh) in the roadway environment (Prichard and Fisher 2012).
In addition, a number of metals and trace elements may have both natural (crustal materials and
soil) and anthropogenic sources. A previous study on PM
10-2.5
in the Los Angeles area reported
that mineral and crustal metals and elements, most notably Fe, Ca, Al, Mg, K, Ti, and Mn
contributed to 33% of total variance in mass, while abrasive vehicular markers such as Cu, Ba,
and Sb, account for 16% of variance (Pakbin et al. 2011). Although crustal sources contribute to
metals and trace elements, the on-road emission factors presented indicate that metal and trace
elemental concentrations are substantially higher than our urban background levels and that
vehicular abrasion is a major source of metals and trace elements in heavily-trafficked surface
streets. Previous studies which have investigated the size distribution profile of metals in
roadways environments have also identified vehicular combustion and lubricant oil to be a
65

source of metals and trace elements (i.e. Cu, Pb, Ca) in the sub-micrometer size range (Lin et al.
2005; Lough et al. 2005; Schauer et al. 2006). Lough et al. (2005) observed a local peak for Cu
and Pb in the 0.1μm size range in a tunnel study and Lin et al. (2005) reported a bimodal
distribution for K, Ca, Zn, and Pb in a roadway study. However, the only species that show an
apparent peak in PM
0.25
is Pb and Sb. Given the 3 PM size fractions studied, it cannot be
determined if there are local peaks in the sub-micrometer size range of the other species.
In roadway environments, while the greatest fraction of metals and trace elements are
derived from vehicular wear products, organic compounds including PAHs, hopanes, steranes,
and n-alkanes are mostly derived from lubricant oil and the incomplete combustion of fuel
(Rogge et al. 1993a; Fine et al. 2004), and thus can be used as tracers for primary traffic
emissions. Contrary to what was observed for metals and trace elements, Table 5.2 shows
emission factors for PAHs are highest in PM
0.25
(1.4-6.5 μg/kg of fuel) while emission factors in
PM
10-2.5
and PM
2.5-0.25
are comparably low (0.1-0.8 μg/kg of fuel). On average, emission factors
in PM
0.25
are 18.9 and 19.7 times higher than corresponding PAHs in PM
10-2.5
and PM
2.5-0.25
,
respectively. This is consistent with observations in a study for a tunnel bore reserved for LDVs
only that found gasoline engine-derived PAHs were found almost entirely in PM
0.12
(Miguel et al.
1998). Except for benzo(ghi)fluoranthene, a low molecular weight (MW) PAH, the emission
factors for high MW PAHs (benzo(ghi)perylene and coronene) are greater than most of the
lighter MW PAHs. Previous studies have also found that HDVs, which were a negligible part of
the fleet composition of Wilshire/Sunset, are the major source of low MW PAHs, whereas both
LDVs and HDVs contribute to higher MW PAHs (Schauer et al. 2002; Ning et al. 2008). The
emission factors of benzo(a)pyrene have been omitted as the concentrations were below
detection limit.
66

Table 5.2 Fuel-based emission factors (mass of pollutant emitted per kg of fuel) of polycyclic aromatic
hydrocarbons (PAHs) and hopanes and steranes for three PM size fractions. Pollutants with concentrations close to
or less than USC background levels have been omitted.

Lastly, emission factors of hopanes and steranes are shown in Table 5.2. These organic
compounds have been used as reliable tracers of vehicular emissions as they are primarily
derived from automotive lubricant oil (Schauer et al. 1996; Cass 1998). Similar to PAHs, the
emission factors in PM
10-2.5
and PM
2.5-0.25
are generally low and are on average 0.8±0.5 and
0.3±0.1 μg/kg of fuel, respectively, while the average for PM
0.25
is 2.9±1.6 μg/kg of fuel. In
comparison to the previously discussed PM components, hopanes and steranes have relatively
low overall emission factors. A tunnel study which apportioned emission factors between LDVs
and HDVs found that hopane and sterane emission factors for HDVs were approximately 8-14
times greater than LDVs for ultrafine PM (d
p
< 0.18 μm) and 10-20 times greater for
Wilshire/Sunset ± Wilshire/Sunset ± Wilshire/Sunset ±
PAHs ( μ g/kg of fuel)
Pyrene 0.27 0.16 0.07 0.06 2.4 1.2
Benzo(ghi )fluoranthene - - 0.34 0.13 6.5 6.6
Benz(a )anthracene 0.13 0.19 0.10 0.15 1.4 0.46
Chrysene 0.24 0.04 0.32 0.08 2.2 0.32
Benzo(b +k )fluoranthene 0.08 0.11 0.79 0.62 4.2 1.0
Benzo(e )pyrene 0.19 0.07 0.08 0.12 2.9 0.37
Benzo(a )pyrene - - - - - -
Indeno(1,2,3-cd)pyrene 0.32 0.45 0.32 0.33 1.5 2.1
Benzo(ghi )perylene 0.25 0.13 0.30 0.27 4.3 0.93
Coronene 0.13 0.11 0.11 0.05 4.3 0.58
Hopanes and steranes (μg/kg of fuel)
17 α(H)-22,29,30-Trisnorhopane 0.38 0.32 0.11 0.16 - -
17 α(H)-21 β(H)-30-Norhopane 0.81 1.0 0.14 0.19 2.5 2.0
17 α(H)-21 β(H)-Hopane 1.8 1.4 - - 5.3 2.4
22S-Homohopane 1.3 0.83 0.16 0.22 2.3 3.2
22R-Homohopane 0.65 0.92 - - 1.7 2.4
22S-Bishomohopane 0.86 1.2 0.56 0.80 - -
22R-Bishomohopane 0.90 1.3 0.26 0.36 - -
α β β-20R-C27-Cholestane 0.38 0.54 0.26 0.37 - -
α β β-20S-C27-Cholestane 0.33 0.46 0.22 0.32 - -
ααα-20S-C27-Cholestane 0.54 0.74 0.31 0.44 - -
PM
10-2.5
PM
2.5-0.25
PM
0.25
67

accumulation PM (2.5 > d
p
> 0.18 μm) (Phuleria et al. 2006). The following section will
compare emission factors from the current study to previous LDV studies conducted in freeway,
tunnel, and laboratory dynamometer environments.
5.3.2 Comparison of PM
2.5
emission factors to previous LDV studies
Table 5.3 shows the sampling/testing dates, location, and relevant results for previous
PM
2.5
LDV emission factor studies to be compared to the current study. Ning et al. (2008)
presents a study that was conducted at a fixed site downwind of the I-110 in Los Angeles at the
northern portion of the freeway that allows LDVs only. Studies by Phuleria et al. (2006) and
Geller et al. (2005) were conducted at the Caldecott Tunnel in Berkeley, CA in Bore 2, which is
reserved for LDVs only. The dynamometer studies are based on a composite of LDV gasoline
vehicles of various makes, models, and odometer values (Schauer et al. 2002; Fujita et al. 2007).
Schauer et al. (2002) focused on 9 catalyst-equipped LDVs (model years 1981-1994) driven
through the cold-start Federal Test Procedure (FTP) urban driving cycle; the Fujita et al. (2007)
study is based on 14 LDVs (model years 1984-1997) driven through a modified unified driving
cycle (UDC) from warm-start. Figure 5.2 shows a comparison of the driving cycles between the
two dynamometer studies and the velocity profile of the current study. The FTP and UDC cycles
have an average velocity of 9.5±7.1 m/s (21.2±15.9 mph) and 11.0±8.8 m/s (24.6±19.7 mph)
(www.epa.gov) while the current study has an average of 7.1±4.7 m/s (16.0±10.5 mph). The

Table 5.3 Description of previous light-duty vehicle (LDV) studies used for comparison.
Sampling location/test cycle Sample/test period Relevant results for comparison Reference
Wilshire/Sunset Blvds March-May 2011 - current study
LDV freeway (northern portion of I-110) May-Jun 2004, Jan 2005 mass, EC, OC, metals, PAHs, hopanes and steranes Ning et al. (2008)
LDV tunnel (Bore 2 of Caldecott Tunnel) Aug-Sept 2004 PM mass, EC, OC, metals Geller et al. (2005)
LDV tunnel (Bore 2 of Caldecott Tunnel) Aug-Sept 2004 PAHs, hopanes and steranes Phuleria et al. (2006)
LDV dynamometer study (warm-start UDC) Summer 2001 mass, EC, OC, metals, PAHs, hopanes and steranes Fujita et al. (2007)
LDV dynamometer study (cold-start FTP) - PM mass, PAHs, hopanes Schauer et al. (2002)
68

time (current study)
06:30 06:40 06:50 07:00 07:10 07:20 07:30
Velocity (mph)
0
20
40
60
80
Test time (s)
0 500 1000 1500 2000
Velocity (m/s)
0
5
10
15
20
25
30
35
Current study (Wilshire/Sunset Boulevards)
Federal Test Procedure (FTP)
Unified Driving Cycle (UDC)

Figure 5.2 Comparison of the velocity profile of the current study (Wilshire/Sunset) and the two test cycles (FTP
and UDC) of the dynamometer studies. The current study shows a typical sampling hour with a 30s resolution based
on GPS data. The FTP and UDC driving schedules use a 1s resolution and can be found at www.epa.gov.

emission factors from the two dynamometer studies are distanced-based (mass of pollutant
emitted per distance traveled), thus for the purpose of comparison, these factors are converted to
fuel-based values. LDV gasoline properties used for conversion are fuel consumption
(12L/100km) and density (740g/L) (Kirchstetter et al. 1999). All data shown are PM
2.5
emission
factors, and the current data are calculated as the sum of PM
0.25
and PM
2.5-0.25
.
Figure 5.3 shows the average emission factors of PM mass, OC and EC for the current
study (shown as bars) and previous LDV studies (shown as markers). All positive errors bars for
the current study represent the upper range of the data (N=2), negative error bars are not shown
for visual clarity. On average, the PM mass emission factor for the current study is higher than
earlier studies, but there is considerable variation in the values reported. The higher emissions of
PM mass observed on the surface streets may be explained by the current on-road sampling
methodology employed, which may capture road dust and vehicular abrasion emissions more
69

effectively than fixed roadside sampling and dynamometer testing, coupled with the frequent
braking encountered on street driving conditions.
mass OC* EC
Emission factor (mg/kg of fuel)
0
50
100
150
200
250
Wilshire/Sunset
LDV freeway (Ning et al. 2008)
LDV tunnel (Geller et al. 2005)
LDV dyno (Fujita et al. 2007)
LDV dyno (Schauer et al. 2002)

*Due to positive OC adsorption artifacts, Ning et al. (2008) reports a possible overestimation of OC and Geller et al.
(2005) omits the ultrafine fraction of OC.
Figure 5.3 PM
2.5
emission factors for PM components for the current study (bars) and previous LDV studies
(markers) for PM mass, OC and EC.

For OC, the emission factor for the current study lies in between the other studies. Due to
positive organic vapor adsorption artifacts from the use of quartz substrates, Ning et al. (2008)
reported an overestimation of the OC emission factors as they discuss in their analysis, whereas
Geller et al. (2005) omitted the OC emission factors in the ultrafine fraction due to similar
concerns about the effect of positive artifacts due to gas phase adsorption, which would be far
more pronounced in the confined environments of a roadway tunnel. As in any study using
quartz filters to capture PM for EC/OC analysis, we acknowledge the possibility of OC
adsorption artifacts affecting our collections. These OC artifacts are only relevant for the after
filter stage (PM
0.25
), as the other size ranges (PM
2.5-0.25
and PM
10-2.5
) are collected by impaction
and thus are not susceptible to adsorption artifacts (Turpin et al. 2000). Two earlier studies in
Los Angeles (Kim et al. 2001; Sardar et al. 2005) examined the effect of PM loading on positive
70

artifacts due to adsorption of vapor phase OC on quartz filters. Sardar et al. (2005), investigated
the OC adsorption artifacts in ultrafine PM using micro-orifice uniform-deposit impactor
(MOUDI) at 30 lpm, and showed that the adsorption process reaches saturation at a
concentration of roughly 1.5 μg/m
3
for 24 hr sampling, corresponding to approximately 65 μg of
OC. Similarly, Kim et al. (2001), who investigated artifacts at a rural site using the PM
2.5

Federal Reference Method (FRM) at 16.7 lpm, found that positive OC artifacts reach a saturation
of approximately 2 μg/m
3
regardless of PM mass concentration, corresponding to a loading of 48
μg of OC. Thus, as the PM mass loadings increase, the OC adsorption artifacts tend to decrease.
The PM
2.5
OC levels in this study are 5.2±0.3 μg/m
3
and correspond to over 400 μg of OC,
which suggests that positive organic artifacts may be less of a concern in the current results given
the high OC loadings. In addition, total PM
2.5
loadings were on the order of 1.5-2 mg. It is
important to note that the other PM constituents (i.e. EC, metals) are not subject these artifacts.
Lastly, because the samples were immediately stored in a refrigerator when not in use, we do not
believe the filters were subject to additional negative artifacts due to volatilization.
For EC, the emission factors for all studies are relatively low (< 30 mg/kg of fuel). This is
expected considering the negligible influence of HDVs, which are the major emitters of EC in an
urban traffic environment (Schauer 2003). Geller et al. (2005), who also apportioned emission
factors between LDVs and HDVs, found that HDVs emit 709±54 mg/kg of fuel compared to
29.4±3.3 mg/kg of fuel for LDVs.
Figure 5.4 shows a comparison of emission factors of metal and trace elements. The
emission factors of Fe are shown on a separate axis due to their difference in magnitude
compared to other studies. These metals and trace elements were selected on the basis they are
associated with vehicular emissions or crustal materials based on the earlier discussion and have
71

been presented in the previous LDV studies as an essential part of the roadway emissions profile.
The most important observation is that all of the emission factors from the current study are
substantially higher than corresponding species of the previous studies. In comparison to the
LDV freeway and tunnel studies, the current study’s emission factors are approximately 12 times
greater for Mg and Al, and 2-5 times greater for Fe and the remaining species. As discussed in
the previous section, a number of the metals and trace elements presented may be derived from
vehicular (tire and brake wear, lubricant oil, catalytic converters, etc.) and crustal sources. The
higher emission factors of Mg and Al observed on the surface street are most likely from road
dust resuspension, as these species are primarily from crustal sources (Grieshop et al. 2006). The
differences in emission factors between the current study and the previous LDV tunnel and
freeway studies can be explained by the on-road sampling methodology employed in the current
study, which may capture substantially higher levels of metals and trace elements associated with
vehicular abrasion and particulate resuspension. Measurements from Ning et al. (2008) were
conducted at 3m downwind of the freeway while measurements from Geller et al. (2005) were
conducted in the air ducts just above the tunnel ceiling at 50m from the tunnel exit, both of
which are reasonable locations for measuring site-specific traffic emissions, but neither method
captures the additional on-road metal and elemental concentrations that are observed on-road
while driving with other vehicles. Another factor that may contribute to the elevated emission
factors is that surface street driving is characterized by driving conditions involving frequent
acceleration and decelerations, while the freeway and tunnel studies are measuring emission
factors from vehicles moving primarily under cruise conditions (Ning et al. 2008; Geller et al.
2005). In comparison to the dynamometer study (Fujita et al. 2007), the emission factors of the
current study are 80 times higher for Fe and 12-15 times higher for Mg, Ca, and Ba. This was not
72

surprising considering that, by design, dynamometer testing captures primarily tailpipe emissions
and fails to capture any road dust and products of vehicular abrasion.
Mg Al K Ca Ti Cu Ba
Emission factors (mg/kg of fuel)
0.0
0.5
1.0
1.5
2.0
2.5
Wilshire/Sunset
LDV freeway (Ning et al. 2008)
LDV tunnel (Geller et al. 2005)
LDV dyno (Fujita et al. 2007)
Fe
0
2
4
6
8
10
12
14

Figure 5.4 PM
2.5
emission factors for metals and elemental species for the current study (bars) and previous LDV
studies (markers). Fe is shown separately due to its difference in magnitude.

Figure 5.5a shows the comparison of emission factors between PAHs. Overall, the
emission factors from the current study are lower than EFs from the LDV tunnel (Phuleria et al.
2006) and higher than the fixed site downwind of the LDV freeway (Ning et al. 2008). The
average ratio of emission factors of the current study to the LDV tunnel is 0.7±0.4; with the
exception of the high MW PAHs (indeno(1,2,3-cd)pyrene, benzo(ghi)perylene, and coronene),
the ratio to the LDV freeway levels is 3.8±1.6. Due to the semi-volatile nature of PAHs, this
group of compounds can partition between gaseous and particulate phase depending on the vapor
pressure of a particular PAH (Naumova et al. 2003). Since smaller PAHs (less than 5 aromatic
rings) have higher vapor pressures than larger PAHs, they can be found in either phase
depending on the environment. The enclosed nature and the lower, constant temperatures of the
73

tunnel environment may favor the partitioning of semi-volatile PAHs into the particulate phase.
In contrast, the higher temperatures coupled with greater dilution conditions at the fixed site
downwind of the freeway may shift PAHs into the gaseous phase, resulting in less PAHs in the
particulate phase and lower PAH emission factors. Although the current study is also subject to
a
Pyrene
Benzo(ghi)fluoranthene
Benz(a)anthracene
Chrysene
Benzo(b+k)fluoranthene
Benzo(e)pyrene
Benzo(a)pyrene
Perylene
Indeno(1,2,3-cd)pyrene
Benzo(ghi)perylene
Coronene
Emission factors ( g/kg of fuel)
0
2
4
6
8
10
12
14
16
Wilshire/Sunset
LDV freeway (Ning et al. 2008)
LDV tunnel bore (Phuleria et al. 2006)
LDV dyno (Fujita et al. 1007)
LDV dyno (Schauer et al. 2007)

b
17 α (H )-22, 29 , 30 -Tri s no rho pa ne
17 α (H )-21β (H )-30-Norho pa ne
17 α (H )-21β (H )-H op an e
22S-Homohopane
22R-Homohopane
22S-Bishomohopane
22R-Bishomohopane
α β β -20 (R +S )-C 27 -C ho l es t an e
α α α -20 S -C 27 -C ho l es t an e
Emission factors ( g/kg of fuel)
0
2
4
6
8
10
Wilshire/Sunset
LDV freeway (Ning et al. 2008)
LDV tunnel (Phuleria et al. 2006)
LDV dyno (Fujita et al. 2007)
LDV dyno (Schauer et al. 2002)

Figure 5.5 Comparison of PM
2.5
emission factors (μg/kg of fuel) between (a) PAHs and (b) hopanes and steranes.
74

open atmospheric conditions, the results are based on on-road sampling of major surface streets
characterized by intermittent acceleration and deceleration of vehicles, as noted earlier. In
comparison to the two dynamometer studies and with the exception of a few PAHs, the emission
factors of the current study are generally higher. This may be explained by differences in driving
conditions, composition of the LDV fleet as well as the sheer traffic volume on Wilshire/Sunset
Boulevards, while results from the dynamometer studies represent only a composite of select
vehicles.
Figure 5.5b shows the comparison of emission factors between hopanes and steranes.
With the exclusion of 17 α(H)-22,29,30-trisnorhopane and ααα-20S-C27-cholestane, the ratios of
emission factors of the current study to the LDV tunnel and LDV freeway are 1.1±0.5 and
0.8±0.4, respectively. Contrary to observations for PAHs, emission factors of hopanes and
steranes for the current study are substantially closer to corresponding emission factors for the
LDV tunnel despite differences in sampling location and conditions. This is consistent with a
previous study that found hopanes and steranes to be less sensitive to variations in temperature
than PAHs (Zielinska et al. 2004). As mentioned earlier, the major source of hopanes and
steranes is from automotive lubricating oil. In addition, this finding highlights the use of hopanes
and steranes as reliable biomarkers of vehicular emissions for various traffic environments and
driving conditions. Since these compounds are found to be generally stable, it is unlikely they
will react fast enough in the atmosphere to compromise their use as a primary tracer on a
regional time scale (i.e. a few hours) (Rogge et al. 1993a). However, in comparison to the two
dynamometer studies, the current study’s emission factors are generally higher, similar to PAHs.
As previously mentioned, this observation further shows how results from dynamometer tests
75

represent idealized driving conditions and may be considerably different from actual traffic
environments.
5.3.3 n-Alkanes and calculation of carbon preference index (CPI)
Particulate n-alkane homologues (C19-C40) were quantified in this study to further
assess anthropogenic influence of organic PM composition for the two major surface streets in
Los Angeles. Figure 5.6 shows the mass concentrations (ng/m
3
) of n-alkanes in PM
10-2.5
, PM
2.5-
0.25
, and PM
0.25
for Wilshire/Sunset Boulevards and USC background site. In PM
10-2.5
, overall n-
alkane concentrations are comparable between Wilshire/Sunset and USC background site;
however, n-alkane concentrations in PM
2.5-0.25
and PM
0.25
between the surface street and USC are
significantly different (p<0.05). It is interesting to note that peak concentrations in all size
fractions are observed at C29 and C31 n-alkanes for both Wilshire/Sunset and USC. Biogenic
debris, namely leaf abrasion products, is characterized by high odd-carbon number
predominance of C29-C33 n-alkanes while fine particulate n-alkanes emitted from combustion
processes show no carbon number preference but rather high concentrations of n-alkanes in C19-
C25 (Rogge et al. 1993b). The same study also found that n-alkanes derived from tire wear
debris have no carbon number preference, but observed greater concentrations for high
molecular weight n-alkanes (>C30) with peak concentrations at C37. This suggests that, in
PM
0.25
, n-alkanes observed on Wilshire/Sunset may be influenced by an additional source of
biogenic and vehicular combustion products than n-alkanes at the USC background site; in
PM
2.5-0.25
, n-alkanes on Wilshire/Sunset may be influenced by biogenic sources and tire wear
debris and less by combustion sources. These results are not surprising considering that the on-
road sampling methodology deployed in this study can capture substantially greater amounts of
76

PM derived from combustion and tire wear as well as biogenic products that have been
pulverized and subsequently resuspended in the form of road dust from vehicular turbulence.

a
n-alkane carbon number
19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
concentration (ng/m
3
)
0
1
2
3
4
Wilshire/Sunset
USC
b
n-alkane carbon number
19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
concentration (ng/m
3
)
0
1
2
3
4
c
n-alkane carbon number
19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
concentration (ng/m
3
)
0
1
2
3
4

Figure 5.6 n-alkane concentrations (C19-C40) for (a) PM
10-2.5
, (b) PM
2.5-0.25
, and (c) PM
0.25
.

77

The carbon preference index, CPI, is an indicator that can be used to determine the
contribution of biogenic organic matter and anthropogenic sources to organic aerosol. The alkane
CPI is defined as the sum of the concentrations of odd carbon number alkanes divided by the
sum of the concentrations of even carbon number alkanes (Simoneit 1989; Brown et al. 2002;
Arhami et al. 2009). Biogenic products have an odd carbon preference and have a high CPI (up
to 7), while organic matter with a high anthropogenic contribution will not have a carbon number
preference and have CPIs ranging from 1 to 2. C
max
is defined as the carbon number of the alkane
which has the highest concentration of the n-alkanes. Table 5.4 shows the sum of the
concentrations of n-alkanes (C19-C40), C
max
and its corresponding concentration, and CPI values
for Wilshire/Sunset and USC in the three PM size fractions. Overall, all CPIs are in the range of
1-2, indicating a substantial anthropogenic influence for both environments and all size ranges.
For Wilshire/Sunset, although the concentration sums of n-alkanes in PM
2.5-0.25
and PM
0.25
are
approximately 2 times higher than PM
10-2.5
, the CPIs are comparable.

Table 5.4 Sum of the concentrations of n-alkanes (C19-C40), Cmax and its corresponding concentrations, and CPI
values. Values shown include uncertainty of one standard deviation.

5.4 Conclusion
To investigate on-road emission factors of PM pollutants for two major surface streets in
Los Angeles (Wilshire and Sunset Boulevards), a sampling campaign was conducted and size-
segregated PM was collected. The current study is the first on-road study to report
comprehensive chemical speciation data including metals and trace elements, PAHs, hopanes
Wilshire/Sunset USC Wilshire/Sunset USC Wilshire/Sunset USC
Σn-alkanes (ng/m
3
) 12.9 ± 4.0 12.0 ± 2.7 25.4 ± 9.1 11.4 ± 1.2 25.1 ± 5.0 13.8 ± 0.7
Cmax (ng/m
3
) C29 (2.1 ± 1.2) C31 (1.8 ± 0.6) C31 (3.3 ± 0.3) C31 (1.5 ± 0.04) C31 (2.8 ± 0.2) C29 (1.7 ± 0.4)
CPI 1.7 ± 1.1 1.0 ± 0.1 1.0 ± 0.2 1.1 ± 0.3 1.1 ± 0.0 1.1 ± 0.1
PM
10-2.5
PM
2.5-0.25
PM
0.25
78

and steranes, and n-alkanes based on time-integrated data. The fuel-based PM emission factors
essentially represent emissions from a LDV fleet with frequent acceleration and deceleration
driving conditions. Previous studies have only reported chemical speciation data from PM
collected at fixed sites in the proximity of a LDV freeway (Ning et al. 2008) and a LDV tunnel
bore (Phuleria et al. 2006; Geller et al. 2005). Results from the current study revealed that the on-
road sampling methodology deployed can capture higher levels of PM compared to earlier
studies, in particular, higher levels of metals and trace elements associated with vehicular
abrasion (Fe, Ca, Cu, and Ba) as well as with crustal origins (Mg and Al). Overall, the PM
2.5

emission factors of PAHs from the current study are lower than the LDV tunnel and higher than
the LDV freeway while hopane and sterane emission factors are comparable between the studies.
Finally, alkane CPIs were calculated and were indicative of substantial anthropogenic source
contribution for surface streets in Los Angeles.
79

Chapter 6 A comparative assessment of PM
2.5
exposures in light-rail,
subway, freeway, and surface street environments in Los
Angeles and estimated lung cancer risk

6.1 Introduction
The focus of this study is to assess and compare PM
2.5
exposures between five different
transport environments in the Los Angeles area (subway, light-rail, surface street, two major
freeways with a low and high truck fraction). Key species for comparison are EC, OC, total and
water-soluble metals and trace elements, and PAHs. They are presented as mass per m
3
of air.
Sources of PM components and species will be discussed in detail as well as its contribution
from the commute environment. Lastly, estimates of lung cancer risk based on concentrations of
PAHs are calculated to determine risk associated with each commute environment. The novelty
of the current study lies in its focus on the exposure risk assessment for commuters of various
microenvironments of Los Angeles. The PM species presented and discussed in this study have
either been identified as carcinogens or as hazardous to human health. In addition, results from
this study are of primary interest not only for commuters of these specific commute
environments, but also for residents and pedestrians who are in the vicinity of major roadways
that are sources of these pollutants.
6.2 Experimental methodology
This study represents the integration of two major campaigns that were conducted
independently but with common measurement methods. The two campaigns will be referred to
as the METRO study and the on-road study. Table 6.1 shows a summary of the sampling dates
and times, routes, and meteorological parameters including average temperature, relative
80

humidity (RH), prevailing wind direction, and wind speed. Meteorological data is based on
South Coast Air Quality Management District (SCAQMD) monitoring site. Average
temperatures varied from 17 – 24 °C for the two campaigns due to different sampling periods.
Note that the differences in the variation of relative humidity were due to the longer sampling
time period for the on-road campaign (6:00AM-5:00PM) as compared to the METRO campaign
(9:30AM-1:00PM). Overall, the wind direction and wind speed are comparable in the two
campaigns. The University of Southern California (USC) site, which is centrally located near
downtown Los Angeles, was sampled concurrently during both campaigns as a reference site.
Previous studies have also been conducted at this site (Moore et al. 2007; Ning et al. 2007).
Figure 6.1 shows a map of the five commute environments, USC reference site, and SCAQMD
monitoring site.

Table 6.1 Summary of sampling dates and times and meteorological parameters for the METRO and on-road
studies. Meteorological parameters are based on South Coast Air Quality Management District (SCAQMD)
monitoring site.

6.2.1 Sampling methodology
Both campaigns used the same instruments to collect PM for the purpose of comparing
results. PM
2.5
was collected using the compact Personal Cascade Impactor Sampler (PCIS) (SKC
Inc., Eighty-Four, PA) (Misra et al. 2002; Singh et al. 2003), which was operated with portable
Leland Legacy pumps (SKC Inc., Eighty-Four, PA). The PCIS was prepared using an impaction
Route Dates of sampling Sampling times
Temperature
(°C)
RH (%)
Prevailing wind
direction
Wind speed
(m/s)
METRO study
METRO lines 5/3-8/13/10 9:30am - 1:00pm 24.0 ± 3.5 55 ± 9.7 SW 3.2 ± 0.9
on-road study
110 3/1-3/8/11, 4/11-4/18/11 6:00am - 5:00pm 18.3 ± 4.5 57.2 ± 21.5 W 3.2 ± 1.2
710 3/17-3/25/11, 4/19-4/25/11 6:00am - 5:00pm 17.3 ± 3.6 59.9 ± 18.2 W 3.5 ± 1.3
Wilshire/Sunset 3/9-3/16/11, 4/26-5/2/11 6:00am - 5:00pm 23.8 ± 5.7 42.0 ± 20.9 SW 3.3 ± 1.6
81

Figure 6.1 Map of five commute environments: 110 (green), 710 (blue), Wilshire/Sunset (purple), METRO red line
(red), and METRO gold line (yellow). The USC reference site is denoted by the star and the South Coast Air Quality
Management District (SCAQMD) monitoring site is denoted by the triangle.

stage with a cutpoint of 2.5 μm and an after filter stage, which collected PM
2.5
. For the purpose
of comprehensive chemical analysis, two types of filters were used—PTFE (Teflon) filters (Pall
Life Sciences, Ann Arbor, MI) and quartz microfiber filters (Whatman International Ltd,
Maidstone, England). Two sets of samples (N=2) were collected for both campaigns.
For the METRO campaign, subjects used suitcases to carry the instruments and spent
25% of their time waiting at stations and 75% of their time inside the train to simulate a typical
commuter. The sampling duration for the METRO campaign was on weekdays from May to
August 2010 from 9:30AM to 1:00PM. For the on-road campaign, the sampling vehicle was a
Honda Insight Hybrid 2011 equipped with a curved inlet for roadway air entry to the sampling
instruments. Since anisokinetic effects, which may result in the overestimation or

82

underestimation of relatively larger particles (i.e. PM
10-2.5
) (Hinds 1999) need to be considered
for on-road PM measurements, it was determined that PM
2.5
, the PM fraction of interest in this
study, is largely unaffected by anisokinetic sampling (refer to Chapter 4). The sampling duration
for the on-road campaign was on weekdays from March to April 2011 from 6:00AM to 5:00PM.
6.2.2 Route description
A detailed description of the methodology of the METRO study has been described in
section 2.2, so only a brief description follows. In the METRO study, two lines (the red line and
gold line) were sampled concurrently using identical instruments. The red line is a subway line
powered by electric third-rail that connects Downtown Los Angeles to North Hollywood and has
the highest ridership of the METRO system; the gold line is a ground-level light rail line
powered by overhead electric lines that connects Downtown Los Angeles to Pasadena. Trains
pass every 8-12 minutes depending on the hour.
The on-road study has been described in detail in Chapters 4 and 5, so a summary of the
routes and methodology follows. On-road sampling was conducted for three roadways during
different time periods (Table 6.1). The three roadways each represent differential private
commute environments. The 110 is a high-traffic freeway that runs from the Port of Los Angeles
through Downtown Los Angeles to Pasadena, and is composed mostly of light-duty vehicles
(LDVs); the 710 is less trafficked but serves as the main corridor for heavy-duty vehicles (HDVs)
traveling to and from the Port of Los Angeles; Wilshire/Sunset Boulevards are major surface
street routes that are composed primarily of LDVs and negligible HDVs and characterized by
“stop-and-go” (frequent acceleration and deceleration) driving conditions. In the Los Angeles
Basin, the 110 and 710 represent the lowest HDV (3.9%) and highest HDV (11.3%) traffic
composition of the freeways in the area, respectively. Traffic data from the CalTrans PeMS 2011
83

database showed that total traffic flows for the 110 and 710 in one direction are 6,378 and 4,247
vehicles/hr, respectively, and truck flows are 243 and 470 trucks/hr. Wilshire/Sunset has a total
flow of 1,839 vehicles/hr and negligible truck flows.
6.2.3 Sample analysis
Both campaigns were analyzed using the same analytical methodology. The Teflon filters
were gravimetrically analyzed to determine mass concentration using a MT5 Microbalance
(Mettler-Toledo Inc., Columbus, OH). For chemical analysis, Teflon filters were extracted by
acid and subsequently analyzed by magnetic-sector Inductively Coupled Plasma Mass
Spectroscopy (SF-ICPMS) to determine metals and trace elements (Zhang et al. 2008). The
quartz filters were analyzed using the Thermal Evolution/Optical Transmittance analysis to
measure elemental and organic carbon (EC/OC) (Birch and Cary 1996). Water extracts of the
samples determined WSOC and water-soluble metals and trace elements. Quartz filters were also
analyzed for polycyclic aromatic hydrocarbons (PAHs) concentrations by gas chromatography
mass spectroscopy (GC/MS) using an established solvent extraction and molecular
quantification analysis protocol as initially developed by Mazurek et al. (1987) and subsequently
advanced by other studies (Schauer et al. 1999; Schauer et al. 2002).
6.3 Results and Discussion
6.3.1 Comparability of the two campaigns
The interpretation of results of this study largely depends on the comparability of PM
components and chemical speciation data across the two sampling campaigns. Chemical
speciation results from PM collected at the USC site (N=2) were compared for the two
campaigns since sampling there occurred concurrently as a reference site, while using the same
84

sampling instrumentation and analyzed with the same methodology. To distinguish the two data
sets, ‘USC on-road’ refers to reference results from the on-road campaign and ‘USC METRO’
refers to the reference results from the METRO campaign. Figure 6.2 shows the average
concentrations (μg/m
3
) of major PM components and the range, unless otherwise noted. Organic
carbon (OC) is a major component of PM
2.5
in the Los Angeles Basin, and can constitute
approximately 30-40% of PM
2.5
in the Basin (Sardar et al. 2005). According to mass balance
analysis from the previously published manuscripts at the USC site, organic matter (OM), which
is OC multiplied by a correction factor of 1.6±0.2 for an urban aerosol (Turpin and Lim 2001),
constitutes the second largest component of PM
2.5
after inorganic ions (section 4.4.2). The
OC WSOC* EC TC
mass concentration ( g/m
3
)
0
1
2
3
4
5
USC on-road
USC METRO

*N=1 for USC METRO data
Figure 6.2 Comparison of major PM components at the USC reference site for the two campaigns to assess
comparability of data. All bars presented in this study represent upper and lower data points (N=2).

85

current study presents OC with no correction factor. The OC levels at USC are 3.5 and 3.1 μg/m
3

for the METRO and on-road campaigns, respectively, varying by 11%; and water-soluble OC
(WSOC) constitutes 51% and 32% of OC, respectively. Elemental carbon (EC) for the USC
METRO is approximately 2 times higher than the USC on-road campaign, which could be the
result of reductions in EC due to the implementation of the recent Port of Los Angeles Clean
Truck program that started in 2008. Total carbon (TC), which is the sum of EC and OC, is 4.5
and 3.6 μg/m
3
for the USC METRO and USC on-road campaigns, respectively, thus varying by
20%.
A table of comparison of metals and trace element and PAH concentrations at USC
during the two campaigns is shown Table 6.2. The major difference observed between the two
campaigns are the sulfur levels which are 2.4 times higher during the METRO campaign than
during the on-road campaign, consistent with the higher and more variable levels of S observed
at the light-rail and subway lines. This can be explained by the fact that the METRO campaign
was conducted in the summertime, the photochemical period in Los Angeles. In Los Angeles, S
in the particulate phase is mostly in the form of ammonium sulfate, which is primarily formed
from gaseous precursors of SO
2
in the presence of sunlight and reaches its highest levels in
summertime. Another important element, Na, which is influenced by sea salt, differs by 30%
between the two campaigns. Otherwise, most of the other metals and trace elements differ by
less than 20%. For PAHs, chrysene exhibits the highest concentration during the USC METRO
campaign and is 3.6 times greater than during the on-road study. All other PAHs are generally
less than 0.1 ng/m
3
.
86

Table 6.2 Average concentrations of metals and trace elements and PAHs at USC site during the two campaigns.
(N=2)

In addition to PM species, meteorological parameters (temperature, relative humidity,
wind direction and speed) and gaseous pollutants (NOx, CO, and O
3
) were compared during the
two campaigns based on measurements at the SCAQMD monitoring site (Table 6.3). Average
temperatures are warmer during the METRO campaign since it is during the summertime while
the on-road campaign is in the spring; average relative humidity (%) and wind speed is similar.
NOx and CO concentrations are greater by about 20-25% on average during the on-road
campaign because these species are emitted by traffic sources and are typically higher in the
morning hours (6:00-9:00AM) and decline throughout the day. On the other hand, O
3
, a
average ± average ± average ± average ±
Metals and trace elements (ng/m
3
) PAHs* (ng/m
3
)
Na 780.5 289.6 591.5 46.2 Pyrene 0.05 - 0.04 0.01
Mg 76.3 6.6 67.5 3.6 Benzo(ghi )fluoranthene 0.04 - bdl bdl
Al 133.3 14.3 140.9 9.7 Benz(a )anthracene 0.03 - 0.01 0.00
S 1412.4 188.9 598.6 38.7 Chrysene 0.18 - 0.05 0.005
K 95.3 10.4 87.6 9.3 Benzo(b )fluoranthene 0.11 - 0.09 0.08
Ca 144.6 30.8 125.4 7.4 Benzo(k )fluoranthene 0.05 - 0.02 0.01
Ti 18.3 1.3 16.3 1.0 Benzo(e )pyrene 0.08 - 0.05 0.06
Cr 3.1 2.2 2.8 0.1 Benzo(a )pyrene bdl - bdl bdl
Mn 5.2 0.8 4.8 0.2 Indeno(1,2,3-cd )pyrene 0.05 - 0.03 0.002
Fe 236.8 8.1 271.2 12.0 Benzo(ghi )perylene 0.11 - 0.14 0.05
Co 0.12 0.0 0.16 0.0 Coronene 0.06 - 0.02 0.01
Ni 2.9 1.5 1.6 0.2
Cu 14.6 0.9 13.9 0.5
Zn 16.3 3.62 14.3 1.16
Mo 1.06 0.1 0.80 0.1
Cd 0.09 0.0 0.09 0.0
Sn 3.5 0.6 3.7 0.1
Sb 2.4 0.55 2.4 0.09
Ba 13.5 0.9 14.9 0.8
Eu 0.011 0.001 0.021 0.001
Pb 2.9 0.0 4.3 0.2
#N/A #N/A #N/A #N/A
bdl denotes below detection limit
*only one sample analyzed for USC METRO for PAHs
USC METRO USC on-road USC METRO USC on-road
87

secondary pollutant, is 12% higher during the METRO campaign because O
3
levels are typically
lower in the morning hours (6:00-9:00AM). Overall, there is more variability in the
meteorological parameters and gaseous pollutants due to the longer sampling period during the
on-road campaign, but corresponding p-values demonstrate that none of the parameters are
statistically significant between the two campaigns (p>0.05). Overall, considering the differential
sampling dates and times of the two campaigns and less a few PM species, the PM
2.5
components
and species do not vary substantially. As a result, concentrations (mass per m
3
of air) will be
presented in this manuscript as the metric of comparison among the five commute
microenvironments.

Table 6.3 Meteorological parameters and gaseous pollutant measurements at South Coast Air Quality Management
District (SCAQMD) monitoring site in downtown Los Angeles.

6.3.2 Major PM components
Figure 6.3 shows the concentrations of major PM components (OC, WSOC, TC, and EC)
for the five microenvironments. The bars represent the upper and lower data points. Overall, the
710 exhibits the highest concentrations of OC and EC, and thus TC, while the two METRO lines
have the lowest concentrations of OC and TC. The relatively higher levels observed on the 710
can be attributed to its higher volume of heavy-duty vehicles (HDVs), which are the greatest
emitters of EC (Schauer 2003) and OC (Geller et al. 2005; Ntziachristos et al. 2007; Phuleria et
al. 2007) in a traffic environment. For OC, the concentrations of the 110 and Wilshire/Sunset are
comparable and are approximately 20% higher than the METRO lines even though reference OC
levels at USC were 11% higher during the METRO campaign. It is also important to note that
Dates of sampling Sampling times
Temperature
(°C)
RH (%)
Prevailing wind
direction
Wind speed
(m/s)
NOx (ppm) CO (ppm)
METRO campaign 5/3-8/13/10 9:30am - 1:00pm 24.0 ± 3.5 55 ± 9.7 SW 3.2 ± 0.9 0.035 ± 0.016 0.47 ± 0.14
on-road campaign 3/1-5/2/11 6:00am - 5:00pm 19.8 ± 8.1 53.0 ± 35.1 W 3.3 ± 2.4 0.043 ± 0.043 0.58 ± 0.32
p-value (α=0.05) - - 0.25 0.82 - 0.22 0.61 0.28
88

EC is largely from primary emissions while OC has a primary and secondary component
(Schauer 2003), and thus less influenced by primary emissions. As discussed in Chapter 3, the
source of OC in the METRO red line is primarily from the entrance of ambient air through the
ventilation system. For WSOC, the concentrations of the microenvironments range from 1.2-2.0
μg/m
3
, with the METRO red line exhibiting the lowest concentration and Wilshire/Sunset
exhibiting the highest. For EC, the 710 concentrations are 2.0-3.3 times higher than the other
commute microenvironments. Similar to OC, the main source of EC for the METRO lines is the
influence of ambient air. Although Wilshire/Sunset is a highly-trafficked roadway environment,
its low levels of EC can be explained by its negligible HDV volumes, designating it as a
primarily LDV roadway. Moreover, Geller et al. (2005) showed that the fuel-based emission
factors (mg of pollutant emitted per kg of fuel burned) for HDVs and LDVs for PM
2.5
is 710 and
29 mg/kg of fuel, respectively, which means HDVs emit almost 25 times more EC than LDVs
per kg of fuel burned.
OC WSOC TC
mass concentration ( g/m
3
)
2
4
6
8
10
12
110
710
Wilshire/Sunset
Gold Line (light-rail)
Red Line (subway)
EC
0.5
1.0
1.5
2.0
2.5

Figure 6.3 Comparison of major PM components (OC, WSOC, EC, and TC) for the five commute
microenvironments. EC appears on a separate axis to highlight differences.
89

6.3.3 Metals and trace elements
A number of metals and trace elements that have been quantified in this study are
identified as hazardous air pollutants under the U.S. EPA Clean Air Act Amendments of 1990,
which includes Sb, Cd, Cr, Co, Pb, Mn, and Ni compounds. The U.S. EPA states that these
airborne pollutants are known to cause or possibly cause cancer or other serious health effects,
including birth defects, further emphasizing the importance of understanding the exposure of
metals and trace elements for public and private commuters as well as residents who are in the
vicinity of major roadways. Figure 6.4 and Table 6.4 shows the concentrations (ng/m
3
) of total
metals and trace elements for the commute microenvironments, and are presented based on
concentration levels. Overall, the METRO red line (subway) exhibits the highest concentrations
for Fe, Mn, Mo, Ba, Cr, Co, Ni, and Cd. For Fe, the red line is approximately 10 times higher
than the 110 and 710, 15 times higher than Wilshire/Sunset, and 20 times higher than the gold
line (light-rail). The sources of these enriched species have been discussed extensively in section
3.3, thus only a brief summary follows. In the enclosed underground environment, subway dust
is generated mainly from the abrasion processes between the rail, wheels, and brakes as well as
by the mechanical wearing of the parts. The rail tracks and the main body of the train are
composed of stainless steel, an Fe-based alloy mixed with other metallic elements to enhance its
properties. The same study also demonstrated that Fe is strongly correlated with Cr, Mn, Co, Ni,
Mo, Cd, and Eu (R
2
> 0.9, p < 0.05), indicating that these species may be components of
stainless steel. In addition, PM may be resuspended due to train and passenger movement.
Although the gold line (light-rail) exhibits relatively low concentrations of these steel-associated
elements, the study determined that based on crustal enrichment factors (Taylor and McLennan
1985), of which values close to 1 indicates crustal origins and greater values indicate
90

Fe
mass concentration (ng/m
3
)
2000
4000
6000
8000
10000
12000
14000
Na Mg Al S K Ca Mn Mo Ba
200
400
600
800
1000
1200
110
710
Wilshire/Sunset
Gold line (light-rail)
Red line (subway)
Ti Cr Ni Cu Zn Sn Sb Pb
mass concentration (ng/m
3
)
20
40
60
80
Co Cd Eu
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6

Figure 6.4 Comparison of concentrations of total metals and trace elements for the five commute environments.

anthropogenic sources, these elements have enrichment factors that are 2-3 times greater than the
USC reference site. This suggests that these elements have indeed been influenced by additional
sources (i.e. steel abrasion, wear of rail parts, etc.) that do not influence USC. Regardless of its
additional source, the concentrations of metals and trace elements on the gold line (light-rail) are
the lowest of the five commute environments, meaning commuters of this light-rail line are
exposed to the lowest amounts of these airborne toxics.

91

Table 6.4 Mass concentrations of major PM components, metals and trace elements, and PAHs at the five
microenvironments (N=2). Only one sample for PAHs was analyzed for the METRO gold line.

The differential microenvironments play a major role in the contribution of various
sources of metals and trace elements presented in this study. For example, Na may be part of soil
dust, but because Los Angeles is largely affected by the southwesterly onshore breeze from the
Pacific Ocean, sea salt is a substantial source of Na in the Basin. The higher Na concentrations
observed on the three roadways relative to the METRO lines are expected in light of its major
110 ± 710 ±
Wilshire/
Sunset
±
METRO
Gold line
±
METRO
Red line
±
PM components ( μg/m3)
OC 5.0 0.6 6.3 1.3 5.2 0.3 4.0 0.1 4.4 0.2
WSOC 1.6 0.01 1.4 0.1 2.0 0.2 1.8 - 1.2 0.1
EC 1.0 0.2 2.0 0.2 0.6 0.01 1.0 0.2 0.8 0.01
TC 6.1 0.7 8.3 1.3 5.8 0.3 5.1 0.2 5.2 0.2
metals and trace elements (ng/m3)
Fe 955.6 43.6 865.7 42.0 666.8 30.8 490.5 195.1 10599.2 1723.7
Na 480.9 37.2 400.7 31.2 660.1 50.7 212.5 3.2 288.0 130.8
Mg 98.9 5.3 86.4 5.0 103.5 5.8 30.6 20.7 63.7 29.5
Al 197.4 12.6 182.2 11.7 170.3 11.1 61.7 54.6 150.8 47.5
S 601.3 38.3 423.0 27.3 432.1 28.0 601.2 359.8 802.1 123.1
K 117.5 12.5 89.2 9.4 101.3 11.8 57.6 10.0 62.1 12.0
Ca 220.6 12.9 209.5 12.8 194.7 11.6 74.2 39.4 189.5 61.2
Mn 10.0 0.4 9.2 0.4 7.4 0.3 5.8 2.2 84.9 13.1
Mo 2.6 0.2 2.2 0.1 2.3 0.1 6.3 2.1 155.6 26.9
Ba 81.7 4.5 80.5 4.6 46.4 2.4 18.4 11.5 215.6 33.3
Ti 46.5 2.5 34.4 1.9 28.1 1.6 9.4 3.0 11.8 2.8
Cr 4.1 0.2 3.4 0.2 3.3 0.1 2.1 0.9 23.1 4.7
Co 0.2 0.0 0.2 0.01 0.2 0.01 0.1 0.03 1.2 0.2
Ni 1.7 0.2 1.6 0.2 0.9 0.2 1.4 0.5 11.9 2.6
Cu 60.4 2.2 36.8 1.4 43.3 1.5 37.5 2.5 64.8 11.3
Zn 29.7 2.4 24.2 1.9 20.5 1.6 23.9 5.5 29.7 2.6
Cd 0.2 0.0 0.1 0.0 0.1 0.01 0.1 0.1 1.0 0.1
Sn 7.9 0.3 4.1 0.2 7.5 0.3 3.3 1.4 4.0 0.4
Sb 10.5 0.4 6.5 0.3 7.1 0.3 3.7 1.6 1.8 0.7
Eu 0.1 0.01 0.1 0.01 0.1 0.003 0.01 0.01 0.1 0.02
Pb 4.7 0.3 4.0 0.2 4.9 0.3 2.4 1.1 2.9 0.4
PAHs (ng/m3)
Pyrene 0.21 0.02 0.83 0.10 0.13 0.05 0.06 - 0.08 0.01
Benzo(GHI)fluoranthene 0.22 0.04 0.32 0.09 0.24 0.24 bdl - 0.06 0.00
Benz(a)anthracene 0.10 0.03 0.16 0.06 0.06 0.02 bdl - 0.02 0.00
Chrysene 0.23 0.06 0.25 0.07 0.14 0.01 0.14 - 0.19 0.02
Benzo(b)fluoranthene 0.31 0.05 0.30 0.09 0.23 0.07 0.10 - 0.12 0.01
Benzo(k)fluoranthene 0.10 0.01 0.10 0.07 0.05 0.02 0.05 - 0.04 0.00
Benzo(e)pyrene 0.21 0.01 0.24 0.09 0.15 0.05 0.04 - 0.08 0.05
Benzo(a)pyrene bdl - 0.09 0.05 bdl - bdl - bdl -
Indeno(1,2,3-cd)pyrene 0.09 0.06 0.04 0.01 0.09 0.09 0.06 - 0.05 0.00
Benzo(GHI)perylene 0.45 0.10 0.51 0.16 0.30 0.10 0.12 - 0.10 0.00
Coronene 0.10 0.10 0.17 0.14 0.17 0.03 0.06 - 0.06 0.01
92

sea salt influence. Moreover, the higher levels of Na observed at Wilshire/Sunset relative to the
110 and 710 (average of 33% higher), can be explained by the proximity of the Wilshire/Sunset
route to the coast, while the 110 and 710 routes are located more inland, and thus less influenced
by sea salt. The average concentrations of Mg, Al, and K for the three roadways are 96.3±8.8,
183.3±13.6, 102.6±14.2 ng/m
3
, respectively, and are all higher than average corresponding
concentrations on both METRO lines (47.2±23.4, 106.2±63.0, and 59.9±3.2 ng/m
3
). This is
consistent with previous studies which determined that these species are primarily of crustal
origin and derived from resuspension rather than from vehicular sources (Sternbeck et al. 2002).
Studies have shown that Fe, Al, Ca, Mg, K, Na are abundant in crustal materials, but road
dust may be enriched with some of these elements, indicating anthropogenic sources (Harrison et
al. 2003; Lough et al. 2005). While the focus of this study is not to quantify the enrichment of
these species relative to a reference site, our data is consistent with earlier roadside studies for
elements which have a contribution from traffic sources (i.e. Cu, Ba, Pb, Fe, Ca, Sb, etc.). The
most obvious observations are the elements that are associated with vehicular traffic but not rail
abrasion or wear, which are Ca, Ti, Sn, Sb, and Pb. These elements are primarily derived from
vehicular wear processes such as brake and tire wear (Ti, Sb, and Pb) and motor oil additives (Ca)
(Lough et al. 2005; Grieshop et al. 2006). The average concentrations for the three roadways are
208.3±13.0 ng/m
3
for Ca, 36.3±9.4 ng/m
3
for Ti, 6.51±2.1 ng/m
3
for Sn, 8.0±2.2 ng/m
3
for Sb,
and 4.6±0.5 ng/m
3
for Pb, while corresponding averages for the two METRO lines are
131.9±81.6, 10.6±1.7, 3.7±0.5, 2.7±1.4, and 2.6±0.4 ng/m
3
. HDVs are known to be greater
emitters of elements associated with brake wear due to the greater mechanical force required to
decelerate relative to LDVs (Sternbeck et al. 2002). However, our results found no statistically
93

significant differences for Ca, Ti, Sn, Sb, and Pb between the 110, which is composed of 3.9%
HDVs, and 710, which is composed of 11.3% HDVs (p=0.42).
Elements associated with both steel and vehicular sources include Fe, Ba, Cu, and Zn.
Traffic sources for these elements include engine wear (Fe), brake wear (Ba and Cu), and tire
wear (Zn) (Sternbeck et al. 2002; Sanders et al. 2003). BaSO
4
is known to be commonly used in
brake linings (Garg et al. 2000). The overall trend for these four elements is substantially higher
concentrations in the red line (subway) relative to the other environments except for Zn, which
exhibit comparable levels for the five environments. For Fe, Ba, and Cu, concentrations on the
three roadways vary within 50% of each other (with averages of 829.4±147.8 ng/m
3
for Fe,
69.6±20.0 ng/m
3
for Ba, and 46.8±12.2 ng/m
3
for Cu), and are generally higher than the gold line
(490.5 ng/m
3
for Fe, 18.4 ng/m
3
for Ba, and 37.5 ng/m
3
for Cu).
Sulfur, which is the predominant element in all environments, shows unexpectedly high
concentrations for the METRO lines relative to the three roadways. As discussed earlier, since S
is in the form of ammonium sulfate in the Los Angeles Basin (Hughes et al. 2000), this is most
likely explained by the warmer temperatures observed during the METRO campaign, which
would lead to greater formation of sulfate from the increase in photochemical reactions. A
previous study in the basin also found that sulfate levels were higher in the summer than in the
winter (Sardar et al. 2005). Another source of S is from fuel, which is known to be emitted at
higher rates from HDVs due to higher S content in diesel fuel (Lowenthal et al. 1994), or S can
be emitted from fuel, motor oil, and additives such as zinc dithiophosphate (Lough et al. 2005).
94

6.3.4 Water solubility of metals and trace elements
In regards to PM exposure, water solubility of particle-bound elements is an important
property that contributes to its bioavailability to human cells. In particular, soluble transition
metals have been shown to mediate cardiopulmonary injury (Costa and Dreher 1997). Numerous
studies have also shown that soluble transition metals (i.e. Fe, Ni, Cr) may generate reactive
oxygen species (ROS) through Fenton-like reactions, resulting in oxidative stress (Loft et al.
2005; Verma et al. 2010). Figure 6.5 shows the water solubility (%) of selected metals and trace
elements in this study grouped by high and low solubility classes. Solubility of elements for the
METRO campaign has been discussed in section 3.3, so only a brief discussion follows. Overall,
elements in the METRO red line are the least soluble but have the highest total elemental
concentrations, while the METRO gold line has the highest solubility of the five environments
but the lowest total elemental concentrations. A number of these metals and trace elements
quantified in this study are typically in the form of oxides or other compounds in the urban
environment (Chow et al. 1994), and thus have varying solubility depending on their specific
form. In addition, various isotopes have different solubility as well. The most notable
observation is the differential solubility of Fe, which exhibited total Fe concentrations of 10.6
μg/m
3
for the red line (subway) but is only 0.8% soluble, while the gold line (light-rail) exhibited
total Fe concentrations of 0.5 μg/m
3
but is 11.5% soluble. This yields water-soluble
concentrations of 79.8 and 57.9 ng/m
3
for the red line and gold line, respectively, which differs
only by 1.4 times as opposed to the 20 times difference for total concentrations. For elements
that are typically salts (i.e. Na, Mg, S in the form of sulfate) generally have higher solubilities, as
expected. Except for a few elements, solubility of elements for the three roadway environments
95

are comparable and lie in between the solubilities of the gold and red line for a number of
transition metals including Cr, Mn, Fe, Co, Ni, Cu, Zn, and Cd.
Na Mg S K Ca Mn Co Ni Cu Zn Mo Cd Sb Ba Eu
water solubility (%)
0
20
40
60
80
100
110
710
Wilshire/Sunset
Gold line (light-rail)
Red line (subway)
Al Ti Cr Fe Sn Pb
0
5
10
15
20

Figure 6.5 Comparison of water-solubility (%) of metals and trace elements for the five microenvironments
separated into high and low solubility species.

6.3.5 Polycyclic aromatic hydrocarbons (PAHs)
The U.S. EPA classifies 16 PAHs as priority pollutants based on their carcinogenicity
and mutagenicity. Many PAHs and their derivatives are identified as probable (Group 2A) or
possible (Group 2B) carcinogens as defined by the International Agency for Research on Cancer
(IARC). Therefore, it is essential to identify and quantify the concentrations of PAHs of which
public and private commuters are exposed to on a daily basis. Figure 6.6a and 6.6b and Table 6.4
shows the average and range of concentrations of 11 PAHs (ng/m
3
) and total PAH concentrations
for the five commute environments. It is important to note that the temporal and seasonal
differences in sampling times may affect PAH concentrations to a certain degree. Total PAH
concentrations are substantially higher on the roadway environments than on the two METRO
lines. Specifically, the 710, 110, and Wilshire/Sunset roadways are 4.2, 2.8, and 2.2 times higher
than the average of the two METRO lines, respectively. For most of the individual PAH species,
96

the 710 levels are generally 2-3 times higher than the other two roadway environments, except
for pyrene, where levels are 4-8 times higher. Although the 710 has the highest total PAH
concentrations, statistical analysis using a t-test showed that the concentrations of all the PAHs
(N=11) on the 710 were not significantly different with the 110 (p=0.37) or Wilshire/Sunset
(p=0.18). PAH concentrations for the two METRO lines are consistently lower than the
roadways, which is expected considering the main source of PAHs in the subway and light-rail
environment is most likely entrainment from ambient air. Pairwise multiple comparison tests
(Tukey test) was performed to further investigate statistical significance, which determined that
only two pairs (710 and red line, 710 and gold line) were significantly different (p<0.05).
a
Pyrene
Benzo(ghi)fluoranthene
Benz(a)anthracene
Chrysene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(e)pyrene
Benzo(a)pyrene
Indeno(1,2,3-cd)pyrene
Benzo(ghi)perylene
Coronene
mass concentration (ng/m
3
)
0.2
0.4
0.6
0.8
1.0
110
710
Wilshire/Sunset
Gold line (light-rail)
Red line (subway)
b
110
710
Wilshire/Sunset
Gold line (light-rail)
Red line (subway)
Total PAHs concentration (ng/m
3
)
1
2
3
4
BaPeq concentration (ng/m
3
)
5
10
15
20
25
Total PAHs
BaPeq

Figure 6.6 a) Concentrations of 11 PAHs and b) sum of PAHs concentrations and ΣBaPeq for the five commute
environments.
In urban environments, the main source of PAHs is from fuel and combustion processes
(Miguel et al. 1998). Both LDVs and HDVs emit PAHs, but HDVs emit PAHs at substantially
higher amounts than LDVs (Phuleria et al. 2006; Ning et al. 2008). Numerous studies which
have apportioned PAH emission factors (μg of pollutant/kg of fuel burned) for HDVs and LDVs
97

through tunnel studies (Phuleria et al. 2006) and dynamometer studies (Schauer et al. 1999;
Schauer et al. 2002) have found that HDVs can emit up to 50 times more PAH levels than LDVs.
The same studies also found that low molecular weight (MW) (MW ≤ 228) PAHs (i.e. pyrene)
are primarily emitted by HDVs and high MW (MW ≥ 276) PAHs (i.e. benzo(ghi)perylene,
indeno(1,2,3-cd)pyrene) are emitted by both HDVs and LDVs, which is consistent with the
results of our current study. Although the 110 has high total traffic flows but low HDV flows
(6378 veh/hr and 243 trucks/hr) and the 710 has a lower total traffic flow but higher truck flow
(4247 veh/hr and 470 trucks/hr), the near 2-fold difference in truck volumes present on the 710 is
most likely responsible for the higher concentrations of light MW PAHs.
An important property of PAHs is their semi-volatile nature. PAHs can be found in the
urban environment in the gaseous phase or adsorbed onto particles in the solid phase based on its
vapor pressure. Ambient temperatures can also play a role in the presence of particle-bound
PAHs (Eiguren-Fernandez and Miguel 2012). Generally, high MW PAHs have lower vapor
pressures than low MW PAHs. Thus, pyrene (MW = 202), can partition between the gas and
particle phase depending on ambient temperatures, while indeno(1,2,3-cd)pyrene (MW = 276)
and coronene (MW = 300) are found almost entirely in the particle phase. This is consistent with
a previous study which found that emission factors for high MW PAH based on roadside
sampling at the 110 and 710 were comparable to the reconstructed LDV and HDV emission
factors based on tunnel sampling in spite of the different temperatures and dilution conditions,
while low MW PAH emission factors differed by over two times (Ning et al. 2008). For the
current study, the ambient temperatures for the two campaigns varied from 17 to 24 °C (Table
6.1). Of the 3 roadway environments, Wilshire/Sunset exhibited the lowest PAH levels,
consistent with the higher observed temperatures. Although the METRO campaign also had
98

higher average ambient temperatures, the PAH levels for the red and gold lines are also
influenced by the particle removal efficiency of the subway and train ventilation systems and
thus particle penetration into the train.
6.3.6 Lung cancer risk for commuters
As mentioned earlier, PAHs are a major public health concern due to its carcinogenicity,
and more specifically, to lung cancer risk. A number of PAHs identified in this study are
classified as priority pollutants under the U.S. EPA. According to the IARC, benzo(a)pyrene, or
BaP, is classified as a probable carcinogen (Group 2A). Since BaP has been studied and its
cancer potency values have been well established, it is commonly used as the index compound
for which potency activity of other PAH compounds and its derivatives are compared to.
Relative potency values are referred to as potency equivalent factors (PEFs), for which BaP has a
PEF of 1. Benz(a)anthracene has a PEF of 0.1, meaning it has 1/10
th
the potency of BaP. PEFs
for other PAH compounds are from OEHHA. The calculation of lung cancer risk follows the
method in Sauvain et al. 2003, and a brief summary follows. The PEFs are multiplied by the
corresponding PAH concentration to determine the BaP equivalent concentrations (BaPeq). The
sum of the individual BaPeq (ΣBaPeq) is subsequently multiplied by its unit risk factor (μg/m
3
)
-1
,
which has been determined based on rodent or epidemiology studies (Collins et al. 1998;
Bostrom et al. 2002; Sauvain et al. 2003). Unit risk factors based on rodent and epidemiology
studies are 1.1E-4 and 2.1E-3 (μg/m
3
)
-1
, respectively. Figure 6.6b shows ΣBaPeq as a marker for
each of the 5 microenvironments. Since the unit risk factors in Sauvain et al. (2003) are based on
occupational continuous exposures (45 years for 8 hr/day), the current unit risk factors are
determined by a multiplication factor of 0.18 to account for the lower risk for a commuter’s
lifetime on the road, which is considered to be 45 years for 5 days/week for 2 hours/day.
99

Table 6.5 shows the concentrations of ΣBaPeq and corresponding unit risk factors and lung
cancer risk based on earlier rodent and epidemiology studies. Note that the unit risk factors
between the rodent and epidemiology studies differ by approximately 19 times, yielding cancer
risk values that differ by the same factor. Overall, the 710 exhibits the highest cancer risk. The
710 is greater than the two METRO lines by an average of 4.2 times, and is greater than the 110
and Wilshire/Sunset by 1.9 and 2.7 times, respectively. Although the results offer meaningful
insight into the cancer risk that various commuters face on a daily basis, the authors
acknowledge that there are relatively large uncertainties associated with the results as seen with
the differences in unit risk factors between the rodent and epidemiology studies. In addition,
concentrations of PAHs may exhibit temporal variation. For example, a previous study in
Wilmington, CA, which is located between the 110 and 710 and in the proximity of the Ports of
Los Angeles and Long Beach found that lung cancer risk is highest during rush hour traffic
(around 8:00AM) and lowest in the late afternoon (around 5:00PM) (Polidori et al. 2008),
whereas the current study represents results based on time-integrated samples from 6:00AM
to5:00PM. Nonetheless, results from this study are substantive in assessing the lung cancer risk
for commuters on the light-rail, subway, freeway, and surface street environments in Los
Angeles.

Table 6.5 Lung cancer risk calculations based on a commuter lifetime of 45 years, 2 hours/day, and 5 days/week.
Unit risk factors for rodent and epidemiology are 1.1E-4 and 2.1E-3 (μg/m
3
)
-1
, respectively.
ΣBaPeq
(ng/m
3
)
rodent epidemiology
110 12.7 1.4 27.1
710 23.3 2.7 49.7
Wilshire/Sunset 8.6 1.0 18.4
Gold line (light-rail) 6.3 0.7 13.5
Red line (subway) 5.1 0.6 11.0
lung cancer risk
(x10
-6
)
100

6.4 Conclusion
This study compares the major PM components (EC, OC, WSOC), metals and trace
elements, and PAHs in PM
2.5
for public and private commuters in five differential environments:
subway (METRO red line), light-rail (METRO gold line), surface street (Wilshire/Sunset), and
two freeways which represent the highest (710) and lowest (110) truck compositions in Los
Angeles. The 710 exhibited the highest EC and OC levels most likely due to its higher volume of
HDVs, while the two METRO lines had the lowest EC and OC levels. Metals and trace elements
quantified in this study are derived from a variety of sources depending on the commute
environment. Substantially high levels of Fe and other steel-associated elements (Mn, Mo, Ba,
Cr, Co, Ni, and Cd) were observed on the red line (subway) and substantially low levels were
observed on the gold line (light-rail). Major sources in the rail environment are steel abrasion and
wear of parts. Another group of elements (Ca, Ti, Sn, Sb, and Pb) was identified to be associated
with urban traffic sources only, which are generated from vehicular wear processes and emitted
from motor oil additives. Additionally, a number of the elements are primarily of crustal (i.e. Mg,
Al) or sea salt (Na) origins, and not influenced by rail or traffic sources. In the roadway
environment, PAHs are primarily derived from vehicular emissions and total PAHs were found
to be substantially higher on the 710, consistent with earlier studies which found HDVs to be
significantly greater emitters of PAHs than LDVs (Phuleria et al. 2006; Ning et al. 2008). Lastly,
lung cancer risk was estimated and the 710 was determined to have the greatest cancer risk,
while the two METRO lines had the lowest risk. Since the gold line (light-rail) was observed to
have low concentrations of both PAHs and metals and trace elements, this suggests that
commuting on a light-rail may have potential health benefits due to lower PM exposure levels as
opposed to driving on freeways and major roadways. Results from this study are especially
101

important for understanding PM
2.5
exposure not only for daily commuters but also for residents
and pedestrians who are in the proximity of roadways and can be subjected to these pollutants
and its inherent health risks.

102

Chapter 7 Conclusions and recommendations for future research
This thesis focuses on the size-fractionated PM exposure for five different commute
environments in Los Angeles (light-rail, subway, freeway with highest HDV fraction, freeway
with lowest HDV fraction, and major surface streets). Individual routes were selected to
represent unique microenvironments within a local region of Los Angeles and to highlight
various sources that contribute to the emissions profile of PM depending on commute
environment. A comprehensive chemical analysis was conducted to assess the chemical
composition of PM collected in the various microenvironments, including OC, EC, WSOC,
inorganic ions, total and water-soluble metals, and speciated organics. Implications to health
were also discussed including toxicological results and estimates to lung cancer risk based on
particle-bound PAHs.
7.1 METRO study conclusion
Chapters 2 and 3 presented the results from the METRANS campaign. Chapter 2 focused
on the mass concentrations of PM
2.5
and PM
10
based on continuous measurements on the light-
rail (METRO gold line) and subway (METRO red line). Results showed that subway commuters
are exposed to greater PM concentrations than light-rail commuters, and the light-rail
environment is heavily influenced by ambient PM while the subway environment has an
additional source of airborne PM from the daily operation of trains. Correlation analysis was also
conducted to show that PM
2.5
and coarse PM are highly correlated, and thus from the same
source. Due to the enclosed nature of the subway line, PM concentrations are inherently higher
Chapter 3 focused on the chemical composition of PM collected on the subway and light-rail
lines. A mass balance was conducted and showed that iron constitutes approximately 30% of
total mass for the subway line for both coarse and fine PM fractions. Bivariate regression
103

analysis investigating ROS activity showed that ROS is strongly correlated with water-soluble Fe,
Ni, and OC. A multiple linear regression model using water-soluble Fe and OC to be predictors
of ROS was developed and explained 94% of the variance in ROS. In addition, on a per volume
basis, commuters on the red line exhibit 55% more ROS activity than riders on the gold line.
Results also showed that a number of metals, including Cr, Mn, Co, Ni, Mo, Cd, and Eu were
elevated for the red line and, to a lesser degree, the gold line. This campaign is the first to make
time-integrated and continuous measurements of PM for the Los Angeles METRO.
7.2 On-road study conclusion
Chapters 4-6 present the results from the on-road study. This study’s novelty lies in the
sampling methodology deployed to collect size-fractionated on-road PM. While numerous on-
road studies have been conducted in various roadway environments, none of the previous have
collected time-integrated PM for the purpose of a comprehensive chemical analysis. Three
roadways (HDV freeway, LDV freeway, and major surface streets) were selected to represent
different commute environments that are typical of Los Angeles commuters. The mobile
laboratory was a hybrid vehicle, and was selected for its low emission. The aerosol inlet was
specially designed to account for anisokinetic effects, which may underestimate or underestimate
PM collection depending on size range. Calculations in section 4.3.1 showed that coarse particles
less than 5 μm in diameter are affected by anisokinetic effects by less than 20%, while PM
2.5
is
virtually unaffected.
Chapter 4 presents the first of the results from this major on-road sampling campaign.
This chapter focuses on the mass balance of three size fractions (PM
10-2.5
, PM
2.5-0.25
, and PM
0.25
)
of the three roadway environments. It is clear that each size fraction is influenced by different
source contributions varying from vehicle emissions, road dust resuspension, and natural sources
104

such as sea salt. For all roadways, PM
0.25
is heavily influenced by vehicular emissions, as
expected since particles from combustion are typically in the ultrafine size fraction. The other
two size fractions are less influenced by roadway emissions. PM
2.5-0.25
is dominated by inorganic
ions which are secondary pollutants formed from gaseous precursors (SO
2
and NOx) in the
presence of sunlight. In an urban environment, the gaseous precursors are typically emitted from
anthropogenic sources such as tailpipe emissions. PM
10-2.5
is dominated by crustal metals, most
likely from the resuspension of road dust generated from the turbulence of on-road traffic. Major
PM components (EC, OC, and WSOC) were investigated and the most notable observation are
the elevated EC concentrations on the HDV freeway (710). Although elevated, EC
concentrations are lower in comparison to previous studies, which may be due to the
implementation of the Clean Truck Program that started in 2008.
Chapter 5 presents the second part of the results from the on-road campaign, and focuses
on the surface street environment (Wilshire and Sunset Boulevards). On-road emission factors
are calculated for LDVs since the surface streets selected are predominantly trafficked by LDVs.
The surface street environments differ from the other two roadway environments is that it is
characterized by stop-and-go driving conditions due to frequent acceleration and deceleration of
vehicles. Emission factors are based on a urban reference site (USC), where concurrent sampling
was conducted, and are presented for metals and trace elements, PAHs, and hopanes and steranes.
Results are compared to earlier an earlier LDV tunnel and LDV freeway study for PM
2.5
only.
Overall, results show that the on-road sampling methodology deployed may capture higher levels
of PM compared to earlier roadside studies. It is also observed that the overall emission factors
of PAHs are lower than the LDV tunnel study and higher than the LDV freeway study. This is
most likely due to the semi-volatile nature of PAHs which has the ability to partition between the
105

gaseous and particulate phase depending on ambient conditions. Due to the cooler temperatures
and enclosed nature of the LDV tunnel environment, PAHs will most likely to partition to the
particulate phase, thus contribute to higher levels of PAHs, while the warmer temperatures and
open atmospheric conditions of the LDV freeway will like shift PAHs to its gaseous phase.
While the two surface streets are also subject to open atmospheric conditions, as noted earlier,
the LDV fleet on surface streets is subject to frequent stop-and-go conditions while the LDV
fleet on freeways is mostly under cruise conditions. Emission factors for hopanes and steranes
were comparable between the current study and the two earlier LDV studies, as expected since
hopanes and steranes are relatively stable regardless of environment.
7.3 Integration of the METRO and on-road study conclusion
The last part of the thesis presents an integration of the results of the METRO study and
the on-road study and essentially ties the entire thesis together. In addition to comparing the
chemical composition of the five commute microenvironments, this chapter of the thesis focuses
more on the health implications of commuter exposure to PM. Since PM
2.5
is the most relevant in
terms of health and risk assessment, results and discussions are limited to this size fraction. Since
a number of metals and trace elements are listed as hazardous air pollutants and a number of
PAHs are classified as priority pollutants under the U.S. EPA, these two major groups of species
become the focus of this integrative chapter. Due to the various commute environments studied,
three major sources contribute to the levels of metals and trace elements depending on the
environment. These three sources include vehicular emissions (tire and brake wear), rail abrasion,
and natural (crustal or sea salt sources). For example, Fe, Mn, Mo, Ba, Cr, Co, Ni, and Cd were
observed on the subway line to be substantially higher, indicating these elements may be
associated with stainless steel and become airborne due to the steel abrasion and wear of train
106

and rail parts. Another group of elements including Ca, Ti, Sn, Sb, and Pb was identified to be
associated with traffic sources only and mostly arise from vehicle wear processes and from
engine oil, for which they are additives. Elements including Mg, Al, and Na originate from
crustal or sea salt, otherwise natural, origins and not influenced by either rail or traffic sources.
PAH concentrations were highest in the HDV freeway environment, as expected since HDVs are
substantially greater emitters of PAHs relative to LDVs (Phuleria et al. 2006; Ning et al. 2008).
Since PAHs are classified as probable or possible carcinogens under the IARC, lung cancer risk
was estimated based on speciated PAHs. Results show that the HDV freeway exhibits the
greatest cancer risk while the two METRO lines exhibit the lowest risk since they have lower
levels of PAHs. Based on the results from this study, commuting on the light-rail line may have
potential health benefits since it has low concentrations of both metals and trace elements and
PAHs, and thus lower risk of lung cancer.
7.4 Recommendations for future research
7.4.1 Limitations of the current studies
Size-fractionated PM was collected in five differential commute environments in Los
Angeles and a comprehensive chemical analysis was conducted to compare major PM
components, metals and trace elements, inorganic ions, and speciated organics. Two separate
campaigns were conducted at different time periods and dates. The METRO campaign was
conducted during the summer of 2010 from 9:30AM to 1:00PM while the on-road campaign was
conducted in the spring of 2011 from 6:00AM to 5:00PM. Due to the time-, resource-, and labor-
intensive nature of the two campaigns, it was not possible to conduct concurrent sampling at all
five commute microenvironments. This is inherently one of the limitations of the current study,
for which there may be temporal or seasonal variations that may contribute to the differences in
107

PM components and species that were compared in Chapter 6. It was shown that at the USC
reference site, which was sampled during both campaigns, differences in PM components, metals
and trace elements, organic species, atmospheric conditions less a few parameters (i.e.
temperature), and gaseous pollutants were not significantly different during the two campaigns
(p>0.05). While measures were taken to ensure results would be consistent as possible (i.e. using
same sampling instruments and chemical analysis), given the limitations, concurrent sampling of
all microenvironments would have been the ideal method of conducting this massive campaign.
Numerous studies have shown that while ambient PM is dynamic and may vary on a
daily basis, PM trends have been observed on a seasonal, diurnal, and temporal basis (Sardar et
al. 2005; Hughes et al. 2000). The studies are conducted for a few months out of the entire year
and are only representative of the time period sampled. Although PM exposure in other similar
commute environments or during another time period can be extrapolated based on the results of
this study, PM exposure will vary depending on regional and temporal differences. Regardless of
this limitation, this study represents an extensive emissions profile that may be typical of many
urban commute environments that are dominated by a mix of anthropogenic and natural sources.
7.4.2 Recommendations for future research
The research presented in this manuscript has valuable implications to public risk
assessment as well as developing a comprehensive emissions profile for various common
commute microenvironments. Although the results are limited to a certain time period and to the
localized region of Los Angeles, researchers looking to investigate PM chemical composition
and exposure for commuters can use the current study as a foundation for developing similar
future research projects. Further studies can be designed to investigate a more intensive
campaign that focuses on on-road diurnal trends (i.e. morning vs evening commute time periods).
108

Since the current study represents the integration of 11-hour time intervals on weekdays, the
results cannot take into account the variability that occurs in the course of the day. Studies can
also be conducted to focus on the seasonal differences throughout a year where an on-road can
be conducted during preferential time periods that represent the extreme weather conditions (i.e.
summer and winter). This would be a great extension of the current on-road research that has
been conducted and it would be especially interesting to investigate the diurnal on-road variation
of PM exposure for commuters during their morning and evening commutes.
Similar future studies can also be conducted on other light-rail and subway systems. It
would be interesting to investigate time-integrated PM composition inside the train and while
waiting at stations to differentiate between the two locations. The current study represents and
integration of waiting at the stations (25% of time) and sitting inside the train (75%) of the time
to represent a typical commuter environment. It may also be worthwhile to investigate a line that
is partially underground and partially ground-level or a line that is entire above ground. The PM
composition may be slightly different from ground-level PM due to elevation differences.
In addition to the future research proposals, a greater sample size (N) would enhance the
analysis of the data set in providing a greater understanding of the variability that may occur in
different time periods. A greater sample size would also benefit epidemiologists investigating
exposure risk for commuters in various microenvironments.
7.4.3 Recommendations for regulatory control
It was determined in the METRO study that a number of metals were elevated especially
in the subway system, and to a lesser degree, in the light-rail system. Levels were substantially
higher in the subway system due to its underground environment and enclosed nature, thus
109

particles generated may accumulate in the underground stations. Certain stations exhibited higher
concentrations than other stations (Chapter 2), which could be due to the age of the station where
some stations may be more recent and thus have improved ventilation systems built in. In regards
to controlling PM pollutants, namely Fe and other stainless steel associated elements, ventilation
systems should be upgraded to enhance airflow of subway air to ambient air and to increase
entrainment of ambient air into the subway station. In-train exposure can also be improved by
upgrading the ventilation system or installing an air conditioning system. Since Chapter 2
showed that PM concentrations inside the train are correlated with PM concentrations at the
station, this suggests that PM at the stations is the main source of PM inside the train. Thus, to
ultimately improve rider exposure to PM inside the train, where commuters spend most of their
time, measures should be taken to improve air quality at the stations.
In regards to controlling on-road PM pollutants, the major source of these pollutants must
be taken into account as well as the PM size fraction of interest. In the traffic environment, it is
clear that the major sources of elevated PM species are from vehicular wear, tailpipe emissions,
and resuspended road dust. The smallest PM size fraction, PM
0.25
, is dominated by OC and is
most likely a result of incomplete combustion byproducts; the larger PM size fraction, PM
10-2.5
,
is dominated by crustal materials, which is most likely from natural sources. While there may not
be any control technology to reduce resuspension of road dust (nor may be necessary to), it is
essential to target effective controls towards the reduction of tailpipe emissions, specifically from
older model vehicles that are less efficient and heavy polluters. A more effective control method
would be to target HDVs, which are substantially greater emitters of EC, OC, metals, and
organic species (Phuleria et al. 2007; Geller et al. 2005).
110

The Port of Los Angeles, which is responsible for the majority of HDVs that travel along
the 710 (HDV freeway), has already implemented the Clean Truck Program in 2008. The
objective of the program is to establish a progressive ban of older model HDVs
(www.portoflosangeles.org). In October 2008, all pre-1989 trucks were banned from entering the
Port; in January 2010, all 1989-1993 trucks were banned, in addition to 1994-2003 trucks that
had not been retrofitted; in January 2012, all trucks that did not meet the 2007 Federal Clean
Truck Emissions Standards were banned from the Port. The program has been extremely
effective in reducing emissions and improving overall air quality in the region. A recent study
showed that health costs from PM emitted from HDVs exceeded 440M dollars in 2005, but
decreased by 36%, 90%, and 96% after meeting the 2008, 2010, and 2012 program deadlines
(Lee et al. 2012). The study shows that the Clean Truck Program has been effective in that it
exceeded its target goal, with substantive reductions in the cost of health outcomes.
In addition to reductions in health costs, emissions have also decreased on the roadways.
Based on results of the current studies, EC levels (Figure 4.7b), which can be used as a tracer of
HDVs at least in an urban environment (Schauer 2003), has decreased nearly 50% in the last 5
years based on comparison of the current HDV freeway study that was conducted in 2011 in
comparison to an earlier study that was conducted in 2006 (Phuleria et al. 2007). The previous
study and current study represents two years before and three years after the implementation of
the Clean Truck Program, respectively. Section 4.4.5 also showed that conditions on the 110
freeway, which has a lower truck fraction, may also have improved based on EC levels. Overall,
there have been obvious improvements in the overall air quality for roadways in Los Angeles in
the past five years. After the implementation of the last stage of the Clean Truck Program in
2012, it would be worthwhile to further investigate PM emissions on the 710 roadway.
111

Publications from this thesis
Chapter 2:
Kam W., Cheung K., Daher N., and Sioutas C. (2011) “Particulate matter (PM) concentrations in
underground and ground-level rail systems of the Los Angeles Metro.” Atmospheric Environment, 45(8):
1517-1524.
Chapter 3:
Kam W., Ning Z., Shafer M. M, Schauer J. J., and Sioutas C. (2011) “Chemical characterization of coarse
and fine particulate matter (PM) in underground and ground-level rail systems of the Los Angeles Metro.”
Environmental Science and Technology, 45(16): 6769–6776.
Chapter 4:
Kam W., Liacos J., Schauer J. J., Delfino R. J., and Sioutas C. (2012) “Size-segregated composition of
particulate matter (PM) in major roadways and surface streets.” Atmospheric Environment, 55: 90-97.
Chapter 5:
Kam W., Liacos J., Schauer J. J., Delfino R. J., Sioutas C. (2012) “On-road emission factors of PM
pollutants for light-duty vehicles (LDVs) based on urban street driving conditions.” Atmospheric
Environment, 61: 378-386.
Chapter 6:
Kam W., Delfino R.J., Schauer J.J., Sioutas C. (2013) “A comparative assessment of PM2.5 exposures in
light-rail, subway, freeway, and surface street environments in Los Angeles and estimated lung cancer
risk.” Environmental Sciences: Processes and Impacts, doi: 10.1039/c2em30495c.

112

Bibliography
Aarnio, P., T. Yli-Tuomi, et al. (2005). "The concentrations and composition of and exposure to
fine particles (PM2.5) in the Helsinki subway system." Atmos. Environ. 39(28): 5059-
5066.
Adams, H. S., M. J. Nieuwenhuijsen, et al. (2001). "Fine particle (PM2.5) personal exposure
levels in transport microenvironments, London, UK." Sci. Total Environ. 279(1-3): 29-44.
Arhami, M., M. Sillanpaa, et al. (2009). "Size-Segregated Inorganic and Organic Components of
PM in the Communities of the Los Angeles Harbor." Aerosol Sci. Technol. 43(2): 145-
160.
Belyaev, S. P. and L. M. Levin (1974). "Techniques for collection of representative aerosol
samples." J. Aerosol Sci. 5: 325-338.
Birch, M. E. and R. A. Cary (1996). "Elemental carbon-based method for monitoring
occupational exposures to particulate diesel exhaust." Aerosol Sci. Technol. 25(3): 221-
241.
Bostrom, C. E., P. Gerde, et al. (2002). "Cancer risk assessment, indicators, and guidelines for
polycyclic aromatic hydrocarbons in the ambient air." Environ. Health Persp. 110: 451-
488.
Brinkman, M., K. Cowen, et al. (2008). "Sioutas PCIS with Leland Legacy Pump:
Environmental Technology Verification Report." Mickey Leland National Urban Air
Toxics Research Center (NUATRC). Number 12.
Brown, D. M., M. R. Wilson, et al. (2001). "Size-dependent proinflammatory effects of ultrafine
polystyrene particles: A role for surface area and oxidative stress in the enhanced activity
of ultrafines." Toxicol. Appl. Pharmacol. 175(3): 191-199.
Brown, S. G., P. Herckes, et al. (2002). "Characterization of organic aerosol in Big Bend
National Park, Texas." Atmos. Environ. 36(38): 5807-5818.
Brunekreef, B., N. A. H. Janssen, et al. (1997). "Air pollution from truck traffic and lung
function in children living near motorways." Epidemiology 8(3): 298-303.
Bukowiecki, N., J. Dommen, et al. (2002). "A mobile pollutant measurement laboratory-
measuring gas phase and aerosol ambient concentrations with high spatial and temporal
resolution." Atmos. Environ. 36(36-37): 5569-5579.
Campbell, A., M. Oldham, et al. (2005). "Particulate matter in polluted air may increase
biomarkers of inflammation in mouse brain." Neurotoxicology 26(1): 133-140.
Cass, G. R. (1998). "Organic molecular tracers for particulate air pollution sources." Trac-Trends
in Analytical Chemistry 17(6): 356-366.
113

Chakrabarti, B., P. M. Fine, et al. (2004). "Performance evaluation of the active-flow personal
DataRAM PM2.5 mass monitor (Thermo Anderson pDR-1200) designed for continuous
personal exposure measurements." Atmos. Environ. 38(20): 3329-3340.
Chan, L. Y., W. L. Lau, et al. (2002). "Commuter exposure to particulate matter in public
transportation modes in Hong Kong." Atmos. Environ. 36(21): 3363-3373.
Chan, L. Y., W. L. Lau, et al. (2002). "Exposure level of carbon monoxide and respirable
suspended particulate in public transportation modes while commuting in urban, area of
Guangzhou, China." Atmos. Environ. 36(38): 5831-5840.
Cheng, Y. H., Y. L. Lin, et al. (2008). "Levels of PM10 and PM2.5 in Taipei Rapid Transit
System." Atmos. Environ. 42(31): 7242-7249.
Cheung, K., N. Daher, et al. (2011). "Spatial and temporal variation of chemical composition and
mass closure of ambient coarse particulate matter (PM10-2.5) in the Los Angeles area."
Atmos. Environ. 45(16): 2651-2662.
Chillrud, S. N., D. Epstein, et al. (2004). "Elevated airborne exposures of teenagers to
manganese, chromium, and iron from steel dust and New York City's subway system."
Environ. Sci. Technol. 38(3): 732-737.
Cho, A. K., C. Sioutas, et al. (2005). "Redox activity of airborne particulate matter at different
sites in the Los Angeles Basin." Environ. Res. 99(1): 40-47.
Chow, J. C., J. G. Watson, et al. (1994). "Temporal and Spatial Variations of Pm(2.5) and Pm(10)
Aerosol in the Southern California Air-Quality Study." Atmos. Environ. 28(12): 2061-
2080.
Chung, A., D. P. Y. Chang, et al. (2001). "Comparison of real-time instruments used to monitor
airborne particulate matter." J. Air Waste Manag. Assoc. 51(1): 109-120.
Collins, J. F., J. P. Brown, et al. (1998). "Potency equivalency factors for some polycyclic
aromatic hydrocarbons and polycyclic aromatic hydrocarbon derivatives." Regul. Toxicol.
Pharm. 28(1): 45-54.
Costa, D. L. and K. L. Dreher (1997). "Bioavailable transition metals in particulate matter
mediate cardiopulmonary injury in healthy and compromised animal models." Environ.
Health Persp. 105: 1053-1060.
Creason, J., L. Neas, et al. (2001). "Particulate matter and heart rate variability among elderly
retirees: the Baltimore 1998 PM study." J. Expo. Anal. Environ. Epidemiol. 11(2): 116-
122.
Dales, R., A. J. Wheeler, et al. (2009). "The Influence of Neighborhood Roadways on
Respiratory Symptoms Among Elementary Schoolchildren." J. Occup. Environ. Med.
51(6): 654-660.
114

Dallmann, T. R., R. A. Harley, et al. (2011). "Effects of diesel particulate filter retrofits and
accelerated fleet turnover on drayage truck emissions at the Port of Oakland." Environ.
Sci. Technol. In press.
Eiguren-Fernandez, A. and A. H. Miguel (2012). "Size-Resolved Polycyclic Aromatic
Hydrocarbon Emission Factors from On-Road Gasoline and Diesel Vehicles:
Temperature Effect on the Nuclei-Mode." Environ. Sci. Technol. 46(5): 2607-2615.
Fine, P. M., B. Chakrabarti, et al. (2004). "Diurnal variations of individual organic compound
constituents of ultrafine and accumulation mode particulate matter in the Los Angeles
basin." Environ. Sci. Technol. 38(5): 1296-1304.
Fruin, S. A., A. M. Winer, et al. (2004). "Black carbon concentrations in California vehicles and
estimation of in-vehicle diesel exhaust particulate matter exposures." Atmos. Environ.
38(25): 4123-4133.
Fujita, E. M., B. Zielinska, et al. (2007). "Variations in speciated emissions from spark-ignition
and compression-ignition motor vehicles in California's south coast air basin." J. Air
Waste Manag. Assoc. 57(6): 705-720.
Furuya, K., Y. Kudo, et al. (2001). "Seasonal variation and their characterization of suspended
particulate matter in the air of subway stations." J. Trace Microprobe Tech. 19(4): 469-
485.
Garg, B. D., S. H. Cadle, et al. (2000). "Brake wear particulate matter emissions." Environ. Sci.
Technol. 34(21): 4463-4469.
Geller, V. D., S. B. Sardar, et al. (2005). "Measurements of particle number and mass
concentrations and size distributions in a tunnel environment." Environ. Sci. Technol.
39(22): 8653-8663.
Gorner, P., D. Bemer, et al. (1995). "Photometer measurement of polydisperse aerosols." J.
Aerosol Sci. 26(8): 1281-1302.
Gouriou, F., J. P. Morin, et al. (2004). "On-road measurements of particle number concentrations
and size distributions in urban and tunnel environments." Atmos. Environ. 38(18): 2831-
2840.
Grieshop, A. P., E. M. Lipsky, et al. (2006). "Fine particle emission factors from vehicles in a
highway tunnel: Effects of fleet composition and season." Atmos. Environ. 40: S287-
S298.
Harrison, R. M., R. Tilling, et al. (2003). "A study of trace metals and polycyclic aromatic
hydrocarbons in the roadside environment." Atmos. Environ. 37(17): 2391-2402.
Heal, M. R., I. J. Beverland, et al. (2000). "Intercomparison of five PM10 monitoring devices
and the implications for exposure measurement in epidemiological research." Journal of
Environmental Monitoring 2(5): 455-461.
115

Hesterberg, T. W., C. A. Lapin, et al. (2008). "A comparison of emissions from vehicles fueled
with diesel or compressed natural gas." Environ. Sci. Technol. 42(17): 6437-6445.
Hinds, W. C. (1999). Aerosol Technology: Properties, Behavior, and Measurement of Airborne
Particles. New York, NY, Wiley-Interscience.
Hu, S., A. Polidori, et al. (2008). "Redox activity and chemical speciation of size fractioned PM
in the communities of the Los Angeles-Long Beach harbor." Atmospheric Chemistry and
Physics 8(21): 6439-6451.
Hueglin, C., R. Gehrig, et al. (2005). "Chemical characterisation of PM2.5, PM10 and coarse
particles at urban, near-city and rural sites in Switzerland." Atmos. Environ. 39(4): 637-
651.
Hughes, L. S., J. O. Allen, et al. (2000). "Evolution of atmospheric particles along trajectories
crossing the Los Angeles basin." Environ. Sci. Technol. 34(15): 3058-3068.
Hughes, L. S., G. R. Cass, et al. (1998). "Physical and chemical characterization of atmospheric
ultrafine particles in the Los Angeles area." Environ. Sci. Technol. 32(9): 1153-1161.
Johansson, C. and P. A. Johansson (2003). "Particulate matter in the underground of Stockholm."
Atmos. Environ. 37(1): 3-9.
Kang, S., H. Hwang, et al. (2008). "Chemical Compositions of Subway Particles in Seoul, Korea
Determined by a Quantitative Single Particle Analysis." Environ. Sci. Technol. 42(24):
9051-9057.
Karlsson, H. L., A. G. Ljungman, et al. (2006). "Comparison of genotoxic and inflammatory
effects of particles generated by wood combustion, a road simulator and collected from
street and subway." Toxicol. Lett. 165(3): 203-211.
Karlsson, H. L., L. Nilsson, et al. (2005). "Subway particles are more genotoxic than street
particles and induce oxidative stress in cultured human lung cells." Chem. Res. Toxicol.
18(1): 19-23.
Kerr, S. C., J. J. Schauer, et al. (2004). "Regional haze in Wisconsin: sources and the spatial
distribution." J. Environ. Eng. Sci. 3(3): 213-222.
Kim, B. M., J. Cassmassi, et al. (2001). "Positive organic carbon artifacts on filter medium
during PM2.5 sampling in the South Coast Air Basin." Aerosol Sci. Technol. 34(1): 35-41.
Kim, J. Y., S. R. Magari, et al. (2004). "Comparison of fine particle measurements from a direct-
reading instrument and a gravimetric sampling method." Journal of Occupational and
Environmental Hygiene 1(11): 707-715.
Kim, K. Y., Y. S. Kim, et al. (2008). "Spatial distribution of particulate matter (PM10 and PM2.5)
in Seoul Metropolitan Subway stations." J. Hazard. Mater. 154(1-3): 440-443.
116

Kim, S., S. Shen, et al. (2002). "Size distribution and diurnal and seasonal trends of ultrafine
particles in source and receptor sites of the Los Angeles basin." J. Air Waste Manag.
Assoc. 52(3): 297-307.
Kingham, S., M. Durand, et al. (2006). "Winter comparison of TEOM, MiniVol and DustTrak
PM10 monitors in a woodsmoke environment." Atmos. Environ. 40(2): 338-347.
Kirchstetter, T. W., R. A. Harley, et al. (1999). "On-road measurement of fine particle and
nitrogen oxide emissions from light- and heavy-duty motor vehicles." Atmos. Environ.
33(18): 2955-2968.
Krudysz, M. A., J. R. Froines, et al. (2008). "Intra-community spatial variation of size-
fractionated PM mass, OC, EC, and trace elements in the Long Beach, CA area." Atmos.
Environ. 42(21): 5374-5389.
Kuhn, T., S. Biswas, et al. (2005b). "Physical and chemical characteristics and volatility of PM
in the proximity of a light-duty vehicle freeway." Aerosol Sci. Technol. 39(4): 347-357.
Kuhn, T., S. Biswas, et al. (2005a). "Diurnal and seasonal characteristics of particle volatility and
chemical composition in the vicinity of a light-duty vehicle freeway." Atmos. Environ.
39(37): 7154-7166.
Landreman, A. P., M. M. Shafer, et al. (2008). "A macrophage-based method for the assessment
of the reactive oxygen species (ROS) activity of atmospheric particulate matter (PM) and
application to routine (daily-24 h) aerosol monitoring studies." Aerosol Sci. Technol.
42(11): 946-957.
Lee, G., S. You, et al. (2012). "Assessing air quality and health benefits of the Clean Truck
Program in the Alameda corridor, CA." Transportation Research Part a-Policy and
Practice 46(8): 1177-1193.
Li, N., M. Y. Wang, et al. (2009). "The Adjuvant Effect of Ambient Particulate Matter Is Closely
Reflected by the Particulate Oxidant Potential." Environ. Health Perspect. 117(7): 1116-
1123.
Liao, D. P., J. Creason, et al. (1999). "Daily variation of particulate air pollution and poor cardiac
autonomic control in the elderly." Environ. Health Perspect. 107(7): 521-525.
Lin, C. C., S. J. Chen, et al. (2005). "Characteristics of metals in nano/ultrafine/fine/coarse
particles collected beside a heavily trafficked road." Environ. Sci. Technol. 39(21): 8113-
8122.
Loft, S., C. Stevns-Hansen, et al. (2005). "Oxidative stress and inflammation as mediators of
health effects or air pollution." Free Radical Res. 39: S26-S26.
Lough, G. C., J. J. Schauer, et al. (2005). "Emissions of metals associated with motor vehicle
roadways." Environ. Sci. Technol. 39(3): 826-836.
117

Lowenthal, D. H., C. F. Rodgers, et al. (1995). "Sensitivity of Estimated Light Extinction
Coefficients to Model Assumptions and Measurement Errors." Atmos. Environ. 29(7):
751-766.
Lowenthal, D. H., B. Zielinska, et al. (1994). "Characterization of Heavy-Duty Diesel Vehicle
Emissions." Atmos. Environ. 28(4): 731-743.
Mar, T. F., G. A. Norris, et al. (2000). "Associations between air pollution and mortality in
Phoenix, 1995-1997." Environ. Health Perspect. 108(4): 347-353.
Marcazzan, G. M., S. Vaccaro, et al. (2001). "Characterisation of PM10 and PM2.5 particulate
matter in the ambient air of Milan (Italy)." Atmos. Environ. 35(27): 4639-4650.
Marder, A. R. (2000). "The metallurgy of zinc-coated steel." Prog. Mater Sci. 45(3): 191-271.
Metzger, K. B., P. E. Tolbert, et al. (2004). "Ambient air pollution and cardiovascular emergency
department visits." Epidemiology 15(1): 46-56.
Miguel, A. H., T. W. Kirchstetter, et al. (1998). "On-road emissions of particulate polycyclic
aromatic hydrocarbons and black carbon from gasoline and diesel vehicles." Environ. Sci.
Technol. 32(4): 450-455.
Minguillon, M. C., M. Arhami, et al. (2008). "Seasonal and spatial variations of sources of fine
and quasi-ultrafine particulate matter in neighborhoods near the Los Angeles-Long Beach
harbor." Atmos. Environ. 42(32): 7317-7328.
Misra, C., M. Singh, et al. (2002). "Development and evaluation of a personal cascade impactor
sampler (PCIS)." J. Aerosol Sci. 33(7): 1027-1047.
Moore, K. F., Z. Ning, et al. (2007). "Daily variation in the properties of urban ultrafine aerosol -
Part I: Physical characterization and volatility." Atmos. Environ. 41(38): 8633-8646.
Morgan, T. E., D. A. Davis, et al. (2011). "Glutamatergic Neurons in Rodent Models Respond to
Nanoscale Particulate Urban Air Pollutants in Vivo and in Vitro." Environ. Health
Perspect. 119(7): 1003-1009.
Murruni, L. G., V. Solanes, et al. (2009). "Concentrations and elemental composition of
particulate matter in the Buenos Aires underground system." Atmos. Environ. 43(30):
4577-4583.
Naumova, Y. Y., J. H. Offenberg, et al. (2003). "Gas/particle distribution of polycyclic aromatic
hydrocarbons in coupled outdoor/indoor atmospheres." Atmos. Environ. 37(5): 703-719.
Nieuwenhuijsen, M. J., J. E. Gomez-Perales, et al. (2007). "Levels of particulate air pollution, its
elemental composition, determinants and health effects in metro systems." Atmos.
Environ. 41(37): 7995-8006.
118

Ning, Z. and T. L. Chan (2007). "On-road remote sensing of liquefied petroleum gas (LPG)
vehicle emissions measurement and emission factors estimation." Atmos. Environ. 41(39):
9099-9110.
Ning, Z., M. D. Geller, et al. (2007). "Daily variation in chemical characteristics of urban
ultrafine aerosols and inference of their sources." Environ. Sci. Technol. 41(17): 6000-
6006.
Ning, Z., N. Hudda, et al. (2010). "Impact of roadside noise barriers on particle size distributions
and pollutants concentrations near freeways." Atmos. Environ. 44(26): 3118-3127.
Ning, Z., A. Polidori, et al. (2008). "Emission factors of PM species based on freeway
measurements and comparison with tunnel and dynamometer studies." Atmos. Environ.
42(13): 3099-3114.
Ntziachristos, L., Z. Ning, et al. (2007). "Fine, ultrafine and nanoparticle trace element
compositions near a major freeway with a high heavy-duty diesel fraction." Atmos.
Environ. 41(27): 5684-5696.
Oberdorster, G., E. Oberdorster, et al. (2005). "Nanotoxicology: An emerging discipline
evolving from studies of ultrafine particles." Environ. Health Perspect. 113(7): 823-839.
Ostro, B. D., W. Y. Feng, et al. (2008). "The impact of components of fine particulate matter on
cardiovascular mortality in susceptible subpopulations." Occup. Environ. Med. 65(11):
750-756.
Pakbin, P., Z. Ning, et al. (2011). "Seasonal and Spatial Coarse Particle Elemental
Concentrations in the Los Angeles Area." Aerosol Sci. Technol. 45(8): 949-U156.
Park, D. U. and K. C. Ha (2008). "Characteristics of PM10, PM2.5, CO2 and CO monitored in
interiors and platforms of subway train in Seoul, Korea." Environment International
34(5): 629-634.
Peters, T. M. and D. Leith (2004). "Particle deposition in industrial duct bends." Ann. Occup.
Hyg. 48(5): 483-490.
Pfeifer, G. D., R. M. Harrison, et al. (1999). "Personal exposures to airborne metals in London
taxi drivers and office workers in 1995 and 1996." Sci. Total Environ. 235(1-3): 253-260.
Phuleria, H. C., M. D. Geller, et al. (2006). "Size-resolved emissions of organic tracers from
light-and heavy-duty vehicles measured in a California roadway tunnel." Environ. Sci.
Technol. 40(13): 4109-4118.
Phuleria, H. C., R. J. Sheesley, et al. (2007). "Roadside measurements of size-segregated
particulate organic compounds near gasoline and diesel-dominated freeways in Los
Angeles, CA." Atmos. Environ. 41(22): 4653-4671.
119

Polidori, A., S. Hu, et al. (2008). "Real-time characterization of particle-bound polycyclic
aromatic hydrocarbons in ambient aerosols and from motor-vehicle exhaust." Atmos.
Chem. Phys. 8(5): 1277-1291.
Pope, C. A. and D. W. Dockery (2006). "Health effects of fine particulate air pollution: Lines
that connect." J. Air Waste Manag. Assoc. 56(6): 709-742.
Prichard, H. M. and P. C. Fisher (2012). "Identification of Platinum and Palladium Particles
Emitted from Vehicles and Dispersed into the Surface Environment." Environ. Sci.
Technol. 46(6): 3149-3154.
Querol, X., A. Alastuey, et al. (2001). "PM 10 and PM2.5 source apportionment in the Barcelona
Metropolitan area, Catalonia, Spain." Atmos. Environ. 35(36): 6407-6419.
Raut, J. C., P. Chazette, et al. (2009). "Link between aerosol optical, microphysical and chemical
measurements in an underground railway station in Paris." Atmos. Environ. 43(4): 860-
868.
Rogge, W. F., L. M. Hildemann, et al. (1993a). "Sources of Fine Organic Aerosol .2. Noncatalyst
and Catalyst-Equipped Automobiles and Heavy-Duty Diesel Trucks." Environ. Sci.
Technol. 27(4): 636-651.
Rogge, W. F., L. M. Hildemann, et al. (1993b). "Sources of Fine Organic Aerosol .3. Road Dust,
Tire Debris, and Organometallic Brake Lining Dust - Roads as Sources and Sinks."
Environ. Sci. Technol. 27(9): 1892-1904.
Saldiva, P. H. N., R. W. Clarke, et al. (2002). "Lung inflammation induced by concentrated
ambient air particles is related to particle composition." Am. J. Respir. Crit. Care Med.
165(12): 1610-1617.
Salma, I., R. Ocskay, et al. (2007). "Sampling artefacts, concentration and chemical composition
of fine water-soluble organic carbon and humic-like substances in a continental urban
atmospheric environment." Atmos. Environ. 41(19): 4106-4118.
Salma, I., T. Weidinger, et al. (2007). "Time-resolved mass concentration, composition and
sources of aerosol particles in a metropolitan underground railway station." Atmos.
Environ. 41(37): 8391-8405.
Samet, J. M., F. Dominici, et al. (2000). "Fine particulate air pollution and mortality in 20 US
Cities, 1987-1994." N. Engl. J. Med. 343(24): 1742-1749.
Sanders, P. G., N. Xu, et al. (2003). "Airborne brake wear debris: Size distributions, composition,
and a comparison of dynamometer and vehicle tests." Environ. Sci. Technol. 37(18):
4060-4069.
Sardar, S. B., P. M. Fine, et al. (2005). "Seasonal and spatial variability of the size-resolved
chemical composition of particulate matter (PM10) in the Los Angeles Basin." Journal of
Geophysical Research-Atmospheres 110(D7).
120

Sauvain, J. J., T. V. Duc, et al. (2003). "Exposure to carcinogenic polycyclic aromatic
compounds and health risk assessment for diesel-exhaust exposed workers." Int. Arch.
Occ. Env. Hea. 76(6): 443-455.
Schauer, J. J. (2003). "Evaluation of elemental carbon as a marker for diesel particulate matter."
J. Expo. Anal. Environ. Epidemiol. 13(6): 443-453.
Schauer, J. J., M. J. Kleeman, et al. (1999). "Measurement of emissions from air pollution
sources. 1. C-1 through C-29 organic compounds from meat charbroiling." Environ. Sci.
Technol. 33(10): 1566-1577.
Schauer, J. J., M. J. Kleeman, et al. (1999). "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.
Schauer, J. J., M. J. Kleeman, et al. (2002). "Measurement of emissions from air pollution
sources. 5. C-1-C-32 organic compounds from gasoline-powered motor vehicles."
Environ. Sci. Technol. 36(6): 1169-1180.
Schauer, J. J., G. C. Lough, et al. (2006). Characterization of Metals Emitted from Motor
Vehicles, Health Effects Institute. Report number 133.
Schauer, J. J., W. F. Rogge, et al. (1996). "Source apportionment of airborne particulate matter
using organic compounds as tracers." Atmos. Environ. 30(22): 3837-3855.
Schwartz, J., F. Laden, et al. (2002). "The concentration-response relation between PM2.5 and
daily deaths." Environ. Health Perspect. 110(10): 1025-1029.
Seaton, A., J. Cherrie, et al. (2005). "The London Underground: dust and hazards to health."
Occup. Environ. Med. 62(6): 355-362.
Seinfeld, J. H. and S. N. Pandis (2006). Atmospheric Chemistry and Physics: From Air Pollution
to Climate Change. Hoboken, N.J., John Wiley & Sons, Inc.
Simoneit, B. R. T. (1989). "Organic-Matter of the Troposphere .5. Application of Molecular
Marker Analysis to Biogenic Emissions into the Troposphere for Source
Reconciliations." Journal of Atmospheric Chemistry 8(3): 251-275.
Singh, M., C. Misra, et al. (2003). "Field evaluation of a personal cascade impactor sampler
(PCIS)." Atmos. Environ. 37(34): 4781-4793.
Sioutas, C., S. Kim, et al. (2000). "Field evaluation of a modified DataRAM MIE scattering
monitor for real-time PM2.5 mass concentration measurements." Atmos. Environ. 34(28):
4829-4838.
Sitzmann, B., M. Kendall, et al. (1999). "Characterisation of airborne particles in London by
computer-controlled scanning electron microscopy." Sci. Total Environ. 241(1-3): 63-73.
121

Sternbeck, J., A. Sjodin, et al. (2002). "Metal emissions from road traffic and the influence of
resuspension - results from two tunnel studies." Atmos. Environ. 36(30): 4735-4744.
Tao, F., B. Gonzalez-Flecha, et al. (2003). "Reactive oxygen species in pulmonary inflammation
by ambient particulates." Free Radic. Biol. Med. 35(4): 327-340.
Taylor, S. R. and S. M. McLennan (1985). The Continental Crust: Its Composition and
Evolution. Oxford, Boston, Palo Alto, Victoria, Blackwell Scientific Publications.
Tonne, C., S. Melly, et al. (2007). "A case-control analysis of exposure to traffic and acute
myocardial infarction." Environ. Health Perspect. 115(1): 53-57.
Turpin, B. J. and H. J. Lim (2001). "Species contributions to PM2.5 mass concentrations:
Revisiting common assumptions for estimating organic mass." Aerosol Sci. Technol.
35(1): 602-610.
Turpin, B. J., P. Saxena, et al. (2000). "Measuring and simulating particulate organics in the
atmosphere: problems and prospects." Atmos. Environ. 34(18): 2983-3013.
Verma, V., M. M. Shafer, et al. (2010). "Contribution of transition metals in the reactive oxygen
species activity of PM emissions from retrofitted heavy-duty vehicles." Atmos. Environ.
44(39): 5165-5173.
Weimer, S., C. Mohr, et al. (2009). "Mobile measurements of aerosol number and volume size
distributions in an Alpine valley: Influence of traffic versus wood burning." Atmos.
Environ. 43(3): 624-630.
Weiss, M., P. Bonnel, et al. (2011). "On-Road Emissions of Light-Duty Vehicles in Europe."
Environ. Sci. Technol. 45(19): 8575-8581.
Westerdahl, D., S. Fruin, et al. (2005). "Mobile platform measurements of ultrafine particles and
associated pollutant concentrations on freeways and residential streets in Los Angeles."
Atmos. Environ. 39(20): 3597-3610.
Xia, T., P. Korge, et al. (2004). "Quinones and aromatic chemical compounds in particulate
matter induce mitochondrial dysfunction: Implications for ultrafine particle toxicity."
Environ. Health Perspect. 112(14): 1347-1358.
Yanosky, J. D., P. L. Williams, et al. (2002). "A comparison of two direct-reading aerosol
monitors with the federal reference method for PM2.5 in indoor air." Atmos. Environ.
36(1): 107-113.
Yanowitz, J., M. S. Graboski, et al. (1999). "Chassis dynamometer study of emissions from 21
in-use heavy duty diesel vehicles." Environ. Sci. Technol. 33(2): 209-216.
Yli-Tuomi, T., P. Aarnio, et al. (2005). "Emissions of fine particles, NOx, and CO from on-road
vehicles in Finland." Atmos. Environ. 39(35): 6696-6706.
122

Zhang, X. Q. and P. H. McMurry (1987). "Theoretical-Analysis of Evaporative Losses from
Impactor and Filter Deposits." Atmos. Environ. 21(8): 1779-1789.
Zhang, Y. X., J. J. Schauer, et al. (2008). "Source apportionment of in vitro reactive oxygen
species bioassay activity from atmospheric particulate matter." Environ. Sci. Technol.
42(19): 7502-7509.
Zhu, Y. F., W. C. Hinds, et al. (2002). "Concentration and size distribution of ultrafine particles
near a major highway." J. Air Waste Manag. Assoc. 52(9): 1032-1042.
Zielinska, B., J. Sagebiel, et al. (2004). "Emission rates and comparative chemical composition
from selected in-use diesel and gasoline-fueled vehicles." J. Air Waste Manag. Assoc.
54(9): 1138-1150.

Abstract (if available)

Abstract According to the U.S. Census Bureau, 570,000+ commuters in Los Angeles travel for over 60 minutes to work. Studies have shown that a substantial portion of particulate matter (PM) exposure can occur during this commute depending on the mode of transport. This thesis focuses on the PM exposure for commuters of four microenvironments in Los Angeles including subway, light-rail, freeways, and surface streets. ❧ The first part of the thesis focuses on the subway and light-rail commute environments. Elevated concentrations of PM have been found in a number of worldwide underground transit systems, with major implications regarding exposure of commuters to PM and its associated health effects. An extensive sampling campaign was to measure PM concentrations in two lines of the Los Angeles Metro system – an underground subway line (Metro red line) and a ground-level light-rail line (Metro gold line). Considering that a commuter typically spent 75% of time inside the train and 25% of time waiting at a station, subway commuters were exposed on average to PM₁₀ and PM₂.₅ concentrations that were 1.9 and 1.8 times greater than the light-rail commuters. The average PM₁₀ concentrations for the subway at station platforms and inside the train were 78.0 μg/m3 and 31.5 μg/m3, respectively

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Temporal, spatial and toxicological characteristics of coarse particulate matter in an urban area and relation to sources and regulations

PDF

Size-resolved particulate matter (PM) in urban areas: toxico-chemical characteristics, sources, trends and health implications

$Exposure assessment and source apportionment of size fractions of airborne particulate matter$

PDF

Exposure assessment and source apportionment of size fractions of airborne particulate matter

PDF

Ultrafine particle concentration inside vehicles: models for exposure assessment

PDF

Investigation of physico-chemical characteristics of particulate matter from vehicular sources

PDF

Development of novel techniques for evaluating physical, chemical and toxicological properties of particulate matter in ambient air

PDF

Characterization of black carbon: from source to evolution of physical and optical properties in the atmosphere

$Physico-chemical properties and source apportionment of size-fractionated airborne particulate matter in urban areas with implications for public health$

PDF

Physico-chemical properties and source apportionment of size-fractionated airborne particulate matter in urban areas with implications for public health

PDF

Toxicity of urban particulate matter: long-term health risks, influences of surrounding geography, and diurnal variation in chemical composition and the cellular oxidative stress response

PDF

Influence of atmospheric processes on the physico-chemical properties of primary particles and their impact on public exposure

PDF

Measurement and methods of assessing the impact of prevalent particulate matter sources on air quality in southern California

PDF

Investigating the temporal trends, sources, and toxicity of ambient particulate matter (PM) in different metropolitan environments, and development of a novel aerosol generation setup for inhalat...

PDF

Physico-chemical characteristics and sources of ambient PM mass and number concentrations and their associated toxicity, and development of novel techniques for high time-resolution measurement o...

$Chemical and toxicological characteristics and historical trends of size-fractioned particulate matter from traffic-related emissions in Los Angeles$

PDF

Chemical and toxicological characteristics and historical trends of size-fractioned particulate matter from traffic-related emissions in Los Angeles

PDF

Toxicological characteristics of particulate matter in an urban environment and their linkage to the source-specific constituents

PDF

Analysis of sources and profiles of organic carbon in ambient particulate matter across fine and coarse sizes and introducing an optical technique for real-time urban dust measurement

PDF

Assessing indoor and outdoor air quality by characterizing the physicochemical and toxicological properties of particulate matter using…

PDF

Oxidative potential of urban atmospheric particles: spatiotemporal trends and associations with source-specific chemical components

PDF

Impact of urban source emissions on ambient particulate matter (PM) composition and associated toxicity in local and regional scales using source appointment models

PDF

Investigating the role of urban emission sources on redox-active PM compounds and the chemical analysis of the standardized diesel exhaust particles

Asset Metadata

Creator Kam, Winnie (author)

Core Title Particulate matter (PM) exposure for commuters in Los Angeles: chemical characterization and implications to public health

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Environmental Engineering

Publication Date 01/11/2013

Defense Date 11/16/2012

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag commuter exposure,Los Angeles,OAI-PMH Harvest,particulate matter,roadways,surface streets,transportation

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Sioutas, Constantinos (committee chair), Chen, Jiu-Chiuan (committee member), Fruin, Scott (committee member), Henry, Ronald C. (committee member), Moffett, James W. (committee member)

Creator Email miss.winniekam@gmail.com,wkam@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-126320

Unique identifier UC11291363

Identifier usctheses-c3-126320 (legacy record id)

Legacy Identifier etd-KamWinnie-1391.pdf

Dmrecord 126320

Document Type Dissertation

Rights Kam, Winnie

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA