University of Southern California Dissertations and Theses

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

(USC Thesis Other)

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

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Oxidative Potential of Urban Atmospheric Particles:
Spatiotemporal Trends and Associations with Source-specific
Chemical Components

by

Arian Saffari

A dissertation presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the Requirements for the Degree
DOCTOR OF PHILOSOPHY
(ENVIRONMENTAL ENGINEERING)

May 2017

Copyright 2017 Arian Saffari
i

Dedication

To my parents, for their unconditional love and endless dedication. You are the
reason I can be here today.

To my sister, for being my first and best teacher.

&

To PeiPei, for her persistent love and support throughout the most critical years of
my life.
ii

Acknowledgements

First and foremost, special and sincere thanks to my PhD advisor, Prof.
Constantinos Sioutas, for giving me an opportunity to be part of his group, and for
his invaluable guidance and unwavering support at each and every stage of my
research. His unparalleled scientific vision, combined with his exceptional
enthusiasm has inspired me throughout these years and has helped me develop my
intellectual maturity.

Thanks to Prof. Jamie Schauer and Dr. Martin Shafer at the University of
Wisconsin-Madison, for providing me with their insightful advice over the years.

Thanks to the members of my candidacy and defense committees at USC, Prof.
George Ban-Weiss, Prof. Caleb Finch, Prof. Kelly Sanders, Prof. Felipe De Barros
and Prof. Genevieve Giuliano, for taking the time to review my thesis proposal and
dissertation and providing me with their constructive feedback.

Thanks to my past and present group mates at the USC aerosols laboratory, Dr.
Payam Pakbin, Dr. Nancy Daher, Dr. Winnie Kam, Dr. Dongbin Wang, Dr. Sina
Hasheminassab, Farimah Shirmohammadi, Mohammad Sowlat, Chris Lovett and
Amirhosein Shams, for being such wonderful colleagues and friends.

Finally, thanks to the large group of collaborators at USC and elsewhere who I had
the privilege of working with and who contributed to this dissertation.

iii

Table of Contents
Dedication ............................................................................................................... i
Acknowledgements .............................................................................................. ii
List of Figures ....................................................................................................... vi
List of Tables ........................................................................................................ ix
Abstract ................................................................................................................. xi

Chapter 1. Introduction ........................................................................................... 1
1.1. Particulate Matter (PM): Properties and Health Effects .......................... 1
1.2. Oxidative Potential: A Metric for Toxicity Assessment............................. 3
1.3. Rationale of the Research ............................................................................. 5
1.4. Dissertation Layout ....................................................................................... 6

Chapter 2. Sources and solubility of quasi-ultrafine elements in Los Angeles .. 9
2.1. Introduction .................................................................................................... 9
2.2. Methods ......................................................................................................... 11
2.2.1. Sampling Location/Technique......................................................................................11
2.2.2. Elemental Analysis .......................................................................................................14
2.2.3. Principal Component Analysis (PCA)..........................................................................15
2.3. Results and Discussion ................................................................................. 15
2.3.1. Sources of Elements .....................................................................................................15
2.3.2. Water Solubility of Elements .......................................................................................24
2.3.3. Spatial and Temporal Variation of Elemental Groups .................................................26
2.4. Summary and Conclusions ......................................................................... 35
2.5. Acknowledgements ...................................................................................... 36

iv

Chapter 3. Seasonal and Spatial Variability of ROS Activity in Los Angeles 37
3.1. Introduction .................................................................................................. 37
3.2. Methods ......................................................................................................... 39
3.2.1. Sampling Sites ..............................................................................................................39
3.2.2. Sampling Protocol ........................................................................................................40
3.2.3. Chemical/Toxicological Analyses ................................................................................41
3.3. Results and Discussion ................................................................................. 42
3.3.1. Spatiotemporal Variation of ROS Activity ..................................................................42
3.3.2. Association of ROS activity and Chemical Components .............................................46
3.4. Summary and Conclusions ......................................................................... 53
3.5. Acknowledgements ...................................................................................... 53

Chapter 4. Global Perspective on PM-induced ROS Activity ........................... 54
4.1. Introduction .................................................................................................. 54
4.2. Sampling and ROS Analysis Methodology ............................................... 56
4.3. Results and Discussion ................................................................................. 62
4.3.1. Comparison of Oxidative Potential in different Urban Areas ......................................62
4.3.2. Association of Oxidative Potential with Chemical Species and Sources .....................63
4.3.3. Effect of Particle Size on PM Oxidative Potential .......................................................67
4.3.4. Seasonal Effects on PM Oxidative Potential ................................................................68
4.4. Summary and Conclusions ......................................................................... 71
4.5. Acknowledgements ...................................................................................... 71

Chapter 5. Impact of Atmospheric Aging on PM-induced ROS Activity ........ 73
5.1. Introduction .................................................................................................. 73
5.2. Methods ......................................................................................................... 75
5.2.1. Site Selection, Sampling Schedule and Meteorology...................................................75
5.2.2. Sampling Method .........................................................................................................82
5.2.3. Chemical and Toxicological Analyses .........................................................................82
v

5.2.4. Chemical Mass Balance (CMB) Model .......................................................................84
5.3. Results and Discussion ................................................................................. 85
5.3.1. Chemical Composition .................................................................................................85
5.3.2. Oxidative Potential Source Apportionment..................................................................92
5.4. Summary and Conclusions ....................................................................... 101
5.5. Acknowledgements .................................................................................... 101

Chapter 6. Concluding Remarks ........................................................................ 102
6.1. Major Findings of the Current Work ...................................................... 102
6.2. Broad Implications and Recommendations ............................................ 103

Bibliography ......................................................................................................... 105
vi

List of Figures
Figure 2.1. Location of the sampling sites in the Los Angeles basin. ............. 13
Figure 2.2. Concentration of water soluble fraction of PM0.25 metals
(yearly average). Error bars represent one standard deviation. .................... 25
Figure 2.3. Yearly average water solubility of PM0.25 metals. Error bars
represent one standard deviation. ..................................................................... 25
Figure 2.4 (a-d). Seasonal variation of selected elements for PC1 (road dust)
at (a) Long Beach, (b) Los Angeles, (c) Riverside and (d) Lancaster.. .......... 30
Figure 2.5 (a-d). Seasonal variation of selected elements for PC2 (vehicular
abrasion) at (a) Long Beach, (b) Los Angeles, (c) Riverside and (d)
Lancaster.............................................................................................................. 31
Figure 2.6 (a-d). Seasonal variation of selected elements for PC3 (residual
oil combustion) at (a) Long Beach, (b) Los Angeles, (c) Riverside and (d)
Lancaster.............................................................................................................. 32
Figure 2.7 (a-d). Seasonal variation of selected elements for PC4 (cadmium
sources) at (a) Long Beach, (b) Los Angeles, (c) Riverside and (d) Lancaster..
............................................................................................................................... 33
Figure 2.8 (a-d). Seasonal variation of selected elements for PC5 (metal
plating) at (a) Long Beach, (b) Los Angeles, (c) Riverside and (d) Lancaster..
............................................................................................................................... 34
Figure 3.1 (a-d). Volume-based ROS activity at different sampling sites
during: (a) spring, (b) summer, (c) fall and (d) winter. Error bars
correspond to one standard deviation. Dashed lines indicate the average of
all sampling sites.. ............................................................................................... 44
Figure 3.2 (a-d). Mass-based ROS Activity at different sampling sites during:
(a) spring, (b) summer, (c) fall and (d) winter. Error bars correspond to one
standard deviation. Dashed lines indicate the average of all sampling sites...
............................................................................................................................... 45
vii

Figure 3.3 (a-c). Linear regression between predicted ROS activity and
measured ROS activity at (a) Long Beach, (b) Los Angeles and (c)
Riverside.... .......................................................................................................... 52
Figure 4.1. ROS activity normalized by PM mass (µg Zymosan/mg PM) and
volume of air (µg Zymosan/m3 air) at different locations and for different
size ranges. Box plots correspond to minimum, 1st quartile, median, 3rd
quartile and maximum...... ................................................................................. 59
Figure 4. 2. Mass-normalized ROS activity associated with PM2.5 in Lahore
and Milan and PM0.25 in Riverside and Los Angeles during summer (i.e.
June, July and August) and winter (i.e. December, January and February).
Error bars correspond to one standard error...... ............................................ 71
Figure 5.1. Location of the three sampling sites in the Los Angeles basin......
............................................................................................................................... 77
Figure 5.2. Wind Speed/Direction (a) and Ozone concentration (ppm) (b),
during the sampling periods in each site. Error bars represent one standard
error of hourly data...... ...................................................................................... 78
Figure 5.3. Diurnal variation of temperature (a) and relative humidity (b) at
the three sampling sites. Error bars represent standard error..... ................. 79
Figure 5.4. 24-hour back-trajectories arriving at the sampling sites at 6:00
pm, for an elevation range of 0-1000 meters and GDAS 0.5-degree
meteorological data...... ....................................................................................... 81
Figure 5.5. Mass concentration (µg/m3) of quasi-ultrafine PM (PM0.25) as
well as carbonaceous species (Organic Carbon (OC), Water Soluble Organic
Carbon (WSOC) and Elemental Carbon (EC)) at the three study locations.
Error bars represent one standard deviation of 7 bi-weekly composited
samples. Pie charts represent the water soluble and water insoluble fractions
of total organic carbon at each site.................................................................... 87
Figure 5.6. Total concentration (ng/m3) of elements and metals (a) as well as
their water soluble fraction (b) at the three study locations. Error bars
represent one standard deviation of 7 bi-weekly composited samples...... .... 88
viii

Figure 5.7. Concentration of detected organic species at the three study
locations: (a) Polycyclic Aromatic Hydrocarbons (PAHs), (b) Hopanes and
Steranes, (c) n-Alkanes and (d) Organic Acids...... .......................................... 91
Figure 5.8. Oxidative potential of particles at the three study locations: (a)
Normalized by the total PM mass (µg Zymosan/mg PM) and (b) Normalized
by the air volume (µg Zymosan/m3). Error bars represent standard
deviation of 7 bi-weekly composited samples...... ............................................. 94
Figure 5.9. Source contributions to the quasi-UF OC (µg/m3) derived from
the CMB model..... .............................................................................................. 97
Figure 5.10. Relative contribution of mobile sources and secondary OC to
the ROS activity...... .......................................................................................... 100

ix

List of Tables
Table 2.1. Meteorological parameters.. ............................................................ 13
Table 2.2 Annual average and standard error (SE) of the mass
concentration (ng/m3) of quasi-UFP (dp<0.25 µm) and its elemental
constituents at Long Beach (HUD) , Los Angeles (GRD, LDS, CCL, USC,
HMS, FRE), Riverside (VBR, GRA) and Lancaster (LAN) ........................... 17
Table 2.3. Principal component loadings (VARIMAX normalized) of
selected elements in quasi-
combined (HUD, GRD, LDS, CCL, USC, HMS, FRE, VBR, GRA, LAN) ... 21
Table 2.4. Principal component loadings (VARIMAX normalized) of
selected elements in quasi-UFP (dp<0.25 um) for rural sites cluster (GRA,
VBR and LAN). ................................................................................................... 22
Table 2.5. Principal component loadings (VARIMAX normalized) of
selected elements in quasi-UFP (dp<0.25 um) for urban sites cluster (HUD,
GRD, LDS, CCL, USC, HMS, FRE).. ............................................................... 23
Table 3.1. Pearson correlation coefficients (R) between monthly ROS
activity data (µg Zymosan/m3 air) and monthly concentrations of chemical
species (carbonaceous material, water soluble metals and inorganic ions).
Underlined numbers indicate values with R>0.7 and p<0.05... ...................... 48
Table 3.2. Pearson correlation coefficients (R) between monthly ROS
activity data (µg Zymosan/m3 air) and monthly concentrations of the water
insoluble fraction of metals. Underlined numbers indicate values with R>0.6
and p<0.05.... ........................................................................................................ 49
Table 3.3. Statistical output of multivariate regression model..... .................. 51
Table 4.1. Summary of studies investigating the ROS activity of atmospheric
PM using the cell-based macrophage assay...... ............................................... 58
Table 4.2. Student’s two-tailed t-test (p-value) results corresponding to the
comparisons between (a) Mass-normalized ROS activity levels in figure 4.1
and (b) volume-normalized ROS activity levels in figure 4.1...... ................... 60
x

Table 4.3. Species and sources associated with the ROS activity (R>0.7 and
p<0.05) at each location and PM size range...... ............................................... 67
Table 5.1. Pearson correlation coefficient (R) between ROS activity and
chemical species including OC, WIOC, WSOC, EC, individual metals and
of summed concentration of metals at the three study locations. Underlined
values correspond to R>0.65..... ......................................................................... 95
Table 5.2. Pearson correlation coefficient (R) between ROS activity and
CMB-derived sources at the three study locations. Underlined values
correspond to R > 0.65...... .................................................................................. 99
Table 5.3. Linear regression analysis between Secondary OC and mobile
sources (µg/mg PM) as independent variables and ROS (µg Zymosan/mg
PM) as dependent variable, for all three sites combined..... ......................... 100

xi

Abstract
There is a large body of literature indicating associations between airborne particulate matter
(PM) and increased risk of a wide range of adverse health outcomes in humans. One of the major
PM properties that significantly contributes to its health effects is the oxidative potential, which
induces cellular oxidative stress in biological systems upon exposure and triggers both localized
and systemic inflammatory responses, leading to a myriad of health effects with endpoints in the
respiratory tract as well as the cardiovascular and nervous systems. Despite the compelling
evidence and documentation of PM-related health effects, the state of knowledge regarding the
exact PM causative agents is fairly immature. Accordingly, current PM regulations mainly target
PM mass concentration, which may not be a good representative for the PM-induced health
effects and toxicity. These knowledge gaps necessitate an improved identification and
characterization of PM chemical composition, emission sources and their association with PM
toxicity and oxidative potential.
The aim of this dissertation was to investigate, in great detail, the associations between the PM
chemical components and toxicity, and to provide guidelines on the specific PM sources that
drive toxicity in urban atmosphere. To this end, a series of case studies were designed and
executed in the Los Angeles metropolitan area, as an example of a complex urban environment
impacted by a variety of PM sources. In these case studies, size resolved PM samples were
collected and analyzed for their chemical composition and toxicity. Subsequently, statistical and
source apportionment techniques such as molecular-marker chemical mass balance (MM-CMB),
principal component analysis (PCA) and regression modeling were deployed to quantify the
sources of PM and its most important sub-classes (such as metal particles) as well as to identify
the associations between toxicity and different parameters. In addition to the case studies in the
Los Angeles basin, the PM toxicity and chemistry data from a number of recent studies were also
pooled together in a meta-analysis, to test the consistency of the findings in different urban
environments. Findings of this work advance our knowledge of PM toxicity and its relationship
with chemistry and emission/formation sources, and provide valuable insights for more targeted
and cost-effective environmental regulations.
1

Chapter 1: Introduction
Adverse health impacts associated with exposure to particulate matter necessitates
implementation of regulatory measures to control the emission of these pollutants into the
atmosphere. Complexity of the chemical composition of these particles, combined with the
potentially significant impact of the regulation on the industry, however, has urged the scientific
community to more accurately characterize these pollutants and to establish plans for more
efficient and targeted regulations.

1.1. Particulate Matter (PM): Properties and Health Effects
Atmospheric Particulate Matte (PM) is a general term referring to the solid particles and/or liquid
droplets suspended in the earth’s atmosphere. PM can be a complex mixture of different organic
and inorganic chemical species, most of which are potentially harmful to human health. PM is,
therefore, identified by the United States Environmental Protection Agency (US-EPA) as one of
the criteria air pollutants (along with 5 other species including ground-level ozone, carbon
monoxide, sulfur oxides, nitrogen oxides and lead). To protect the health of the citizens, US-EPA
regulates PM and other criteria pollutants under the Clean Air Act (CAA) and by the National
Ambient Air Quality Standards (NAAQS). Atmospheric PM can originate from two broad types
of sources: primary and secondary (Fine et al., 2008). Primary sources refer to the sources that
directly emit PM in the atmosphere, such as combustion processes (e.g. vehicle exhaust,
residential wood and biomass burning, ship emissions), industrial activities and constructions.
Secondary sources, on the other hand, represent the generation of PM in the atmosphere due to
the photochemical processing. Particles that originate from secondary sources usually have
distinctive properties compared to their precursor primary particles.
2

Numerous epidemiological as well as clinical investigations have linked PM exposure to a wide
range of adverse health effects, including but not limited to premature death,(Hoek et al., 2002)
cardiovascular diseases (Delfino et al., 2005a), respiratory problems (Donaldson et al., 2002) and
neurodevelopmental disorders (Campbell et al., 2005). The adverse health endpoints of PM
exposure are, for the most part, dependent on their physical and chemical properties. Among
different physical properties of PM, particle size is the most important parameter mediating the
health effects (Sioutas et al., 2005). PM are typically classified into different categories based on
their aerodynamic diameter, namely coarse PM (particles with diameters between 10 to 2.5 µm),
Fine PM (particles with diameters smaller than 2.5 µm) and Ultrafine PM (particles with
diameters smaller than 0.1 µm). Although current NAAQS standards set by the US-EPA only
regulate coarse and fine PM, ultrafine PM are believed to have higher toxicity compared to the
larger regulated size ranges (Delfino et al., 2005; Hughes et al., 1998; Solomon et al., 2012).
Ultrafine particles typically account for a small fraction of the total PM mass in urban
atmosphere, although they dominate the total particle number. Due to their smaller size, ultrafine
particles have higher surface area per mass and therefore they can carry significant amounts of
toxic agents. Moreover, due to their smaller size, ultrafine particles have higher pulmonary
deposition efficiency and therefore can potentially penetrate deeper into the respiratory system
compared to the larger size ranges (Sioutas et al., 2005). Chemical content of PM is another key
factor affecting the toxicity and adverse health effects of PM inhalation. Depending on their
emission/formation source, atmospheric particles in an urban area can be made of different
subgroups of chemical species, including but not limited to inorganic ions (such as sulfate,
nitrate and ammonium), metals (iron, zinc, aluminum, copper and potassium, to name a few),
various groups of organic compounds (such as alkanes, organic acids, hopanes, steranes and
3

polycyclic aromatic hydrocarbons) and black carbon. It is noteworthy that other than adverse
health effects in humans, atmospheric PM are also significant mediators in several other
important environmental issues such as climate change, regional visibility range, acid rains and
damage to material (Seinfeld and Pandis, 2012).
1.2. Oxidative Potential: A Metric for Toxicity Assessment
Although the adverse health effects associated with the exposure to PM is well-established,
knowledge regarding the exact pathophysiological mechanisms through which PM toxicity is
induced in human body is still incomplete and is currently under extensive research. Based on
the clinical knowledge, one of the most important pathways leading to PM toxicity upon
inhalation is the oxidative stress, originated from the formation of reactive oxygen species (ROS)
within the epithelial cells and macrophages in the respiratory system (Oberdörster et al., 2005;
Shi et al., 2003; Squadrito et al., 2001; Tsai et al., 1999). There is solid evidence that elevated
ROS levels in the cells (and the resulting oxidative stress) can alter the redox status of the cells
and consequently trigger a series of acute and chronic responses such as inflammation in
pulmonary tract as well as the cardiovascular system (Squadrito et al., 2001) and mitochondrial
damage (Li et al., 2003) that consequently result in a myriad of adverse health outcomes.
Oxidative potential is, therefore, currently considered as the most relevant and robust surrogate
for assessment of the PM-induced toxicity.
Several methodologies have, so far, been introduced for measuring the oxidative potential of the
atmospheric PM. PM inhalation is a complex physiochemical process, considering the significant
heterogeneity, size and complicated composition of the particles. In general, the closer the metric
of oxidative stress measurement is to the actual biological pathway of PM inhalation, the more
4

comprehensive and reliable is the toxicity assessment. The current metrics for assessment of the
oxidative stress can be divided into two subcategories, Chemical (i.e. acellular) assays and
biological (i.e. cellular) assays (Ayres et al., 2008). A brief description of these two categories is
provided below:
Chemical (Acellular) Assays: Chemical assays develop procedures that reflect the chemical
agents in the atmospheric particles that are known to be responsible for their ability to induce
oxidative potential. Some of the prominent chemical assays for measurement of oxidative
potential include PM-catalyzed dithiothreitol (DTT) consumption assay, PM-catalyzed Hydroxyl
benzoate generation assay and Ascorbic Acid depletion assay (Ayres et al., 2008). DTT assay,
for instance, is based on the ability of redox-active PM components to transfer electrons from
DTT to oxygen, a process that generates superoxide radical (a strong oxidant agent) (Cho et al.,
2005). The rate of this reaction in the presence of PM is monitored and expressed as oxidative
potential induced by the particles. Hydroxyl benzoate assay is based on the interaction between
PM’s redox active agents (particularly metals) with a substrate such as salicylic acid to form
hydroxyl radical (a strong oxidative agent) (Coudray and Favier, 2000). Ascorbic Acid depletion
assay quantifies the oxidative potential by measuring the rate through which atmospheric PM
deplete Ascorbic Acid (a strong antioxidant agent present in the long lining fluid), rather than by
quantification of directly generated oxidant agents (Campian et al., 2002). Although the chemical
assays can, at least in theory, provide insight regarding the oxidative potential of particles, there
are significant gaps between physiological relevance of these assays to the actual in-vivo PM
exposure, as these assays do not reflect the biological response caused by the interaction of the
particles with the human cells. Simplicity, rapidness and relatively lower analytical costs are,
however, the major advantages of chemical assays compared to the biological assays.
5

Biological (cellular) Assays: Biological assays employ an in-vitro exposure module for
quantification of the oxidative stress. Intracellular ROS generated from the oxidative stress
associated with the lysosomal action of the macrophages (i.e. digestion of the soluble oxidative
stressors such as redox active water soluble metals) is considered as the most important
biological pathway leading to inflammatory responses (Imrich, 1998; Imrich et al., 2007). This
increased abundance of inflammatory mediators within the cells is thought to be directly
associated with the lung malfunction and aggravation of respiratory diseases such as asthma and
chronic obstructive pulmonary disease (COPD) (Delfino et al., 2005; Hoek et al., 2002). A series
of cellular assays have been developed during the past few years to better simulate these cell-
ROS interactions upon inhalation of ambient as well as engineered nanoparticles (Nel et al.,
2006).
1.3. Rationale of the Research
Despite commendable progress in PM-related toxicological research during the past decade,
appropriate extrapolation of the clinical data to the actual real-world PM exposure conditions
has, so far, been hindered, largely due to the yet incomplete knowledge regarding the specific
chemical components and associated PM emission/formation sources that dominantly drive the
PM-induced oxidative potential in the atmosphere. The regulatory measures in urban areas have,
therefore, been focused on the PM mass as the sole surrogate of the PM-related health effects,
with minimal consideration of the chemical and toxicological properties associated with the
specific emission sources. In the current regulatory framework, same mass emissions from all
different sources are treated as similar, while the relative toxicological potency of various
sources can be significantly different. The main purpose of this dissertation is to fill some of the
knowledge gaps between the current clinical knowledge regarding the PM-induced toxicity and
6

the actual sources and chemical components that drive the PM toxicity in the urban atmosphere,
with the aim of promoting current regulations toward a more targeted and effective approach.
The major objectives can be summarized as below:
 Evaluation of the spatial and temporal variability of the oxidative potential associated
with the ultrafine PM in the Los Angeles air basin, as an example of an urban
environment.
 Determination of the chemical species and emission sources that dominantly drive the
PM-induced oxidative potential in the atmosphere.
 Evaluation of the major sources associated with the redox-active chemical agents present
in ultrafine PM, with a special focus on metals.
 Evaluating the potential effect of the atmospheric aging on the PM-induced oxidative
potential.
 Comparison of the above-mentioned results among different urban areas with the aim of
identifying the overarching findings and providing global quantitative comparative data.
1.4. Dissertation Layout
This dissertation is organized into different chapters as described below:
Chapter 2: this chapter provides a detailed assessment of trace element and metal content of
quasi-ultrafine particulate matter (PM0.25) in the Los Angeles area. A year-long sampling
campaign at ten distinct locations across the Los Angeles air basin and an extensive chemical
analysis on the collected samples enabled us to have an extensive discussion about the temporal
and spatial variation of the elemental content of quasi-ultrafine particles. The sources of trace
elements and metals were further characterized by applying principal component analysis to the
7

dataset. Characterizing metal sources in this chapter established the much needed foundation of
PM toxicity versus PM source comparisons presented in the following chapters.
Chapter 3: This chapter provides a comprehensive assessment of reactive oxygen species (ROS)
activity of quasi-ultrafine particulate matter in the Los Angeles metropolitan area. A year-long
sampling campaign at 10 distinct locations across the Los Angeles air basin and a detailed
chemical analysis on the collected samples enabled an extensive and quantitative discussion
about the temporal and spatial variation of PM0.25-induced ROS activity of quasi-ultrafine
particles and its association with chemical species. Moreover, the findings in this chapter provide
additional insight on the role of water-solubility of chemical content in the ROS activity, and the
potential emission sources that have the highest contribution to the ROS activity at each location.
Chapter 4: in this chapter, a meta-analysis was performed to provide a detailed and quantitative
assessment of the potential toxicity of atmospheric particulate matter (PM) in various urban
locations in the world and its relation to chemical composition and emission sources. The results
of several previous studies in different locations over the past decade are integrated and analyzed
in this chapter in order to provide global quantitative comparative data. Of particular note is the
analysis of the main emission sources pertaining to the observed toxic effects and the
implications for targeted air quality regulation with the aim of reducing adverse health effects of
exposure to toxic PM in urban areas. The effect of particle size as well as photochemical aging
on the PM-induced toxicity is also discussed briefly based on the real-time in-vitro toxicity data
in various urban locations.
Chapter 5: this chapter provides quantitative insights on the relative impacts of primary and
secondary organic sources on the oxidative potential of quasi-ultrafine particles (PM0.25) in the
8

Los Angeles basin. The assessment was performed by collecting PM0.25 in three different time
segments of the day at three contrasting locations, followed by quantification of chemical
composition (namely metals and organic compounds), source apportionment using molecular
marker chemical mass balance (MM-CMB) model and eventually performing regression analysis
between the sources and oxidative potential in an effort to evaluate the changing contribution of
primary and secondary sources on the oxidative potential and quantitatively investigate the
impact of photochemical aging on PM toxicity.
Chapter 6: this chapter summarizes the major findings of this dissertation and provides
directions for future research on this topic, as well as recommendations for implementing more
cost-effective and targeted PM regulations in urban areas.
9

Chapter 2:
Sources and solubility of quasi-ultrafine elements in Los Angeles
Year-long sampling campaign of quasi-ultrafine particles (PM
0.25
) was conducted at 10 distinct
locations across the Los Angeles south coast air basin and concentrations of trace elements and
metals were quantified at each site using high-resolution inductively coupled plasma sector field
mass spectrometry. In order to characterize sources of trace elements and metals, principal
component analysis (PCA) was applied to the dataset. The major sources were identified as road
dust (influenced by vehicular emissions as well as re-suspended soil), vehicular abrasion,
residual oil combustion, cadmium sources and metal plating. These sources altogether accounted
for approximately 85% of the total variance of quasi-ultrafine elemental content. The
concentrations of elements originating from source and urban locations generally displayed a
decline as we proceeded from the coast to the inland. Occasional concentration peaks in the rural
receptor sites were also observed, driven by the dominant westerly/southwesterly wind
transporting the particles to the receptor areas.

This chapter is based on the following publication:
Saffari, A.; Daher, N.; Shafer, M. M.; Schauer, J. J.; Sioutas, C. Seasonal and spatial variation of
trace elements and metals in quasi-ultrafine (PM0.25) particles in the Los Angeles metropolitan
area and characterization of their sources. Environ. Pollut. 2013, 181, 14–23.

2.1. Introduction
Exposure to atmospheric particulate matter (PM) has been linked to several adverse health
effects, including but not limited to, respiratory diseases (Penttinen et al., 2001), cardiovascular
issues (Delfino et al., 2005) and neurodevelopmental disorders (Campbell et al., 2005). Health
effects of PM exposure are primarily linked to chemical composition of PM rather than total PM
mass (Claiborn et al., 2002; Tsai et al., 1999; Verma et al., 2009). There is growing literature
supporting the hypothesis that one of the important pathways underlying these adverse health
endpoints is the oxidative stress (e.g. ROS generation) that derives from the interaction of PM
10

with cells (Donaldson et al., 2002). Elevated ROS levels can alter the redox status of the cell and
consequently trigger a series of acute and chronic responses such as pulmonary inflammation
(Squadrito et al., 2001) and mitochondrial damage (Li et al., 2003).
Among different PM chemical constituents, certain metals and trace elements are of utmost
importance from a toxicological viewpoint purportedly due to their capability to increase the
redox activity of ambient PM (Shi et al., 2003; Valavanidis et al., 2005). Trace metals and
elements can be found in different size ranges. Considering that pulmonary deposition for
airborne particles increases dramatically with decreasing particle size (Kreyling et al., 2006;
Montoya et al., 2004), determining the metal content of the quasi-UFP and their corresponding
sources is extremely important in the overall assessment of public risk due to exposure to these
species and vital in designing strategies for controlling their emissions. A variety of these
sources can be found in the megacity of Los Angeles, with different source contributions
generating different concentrations of trace elements depending on the location. While numerous
metals and elements are considered to contribute to the toxic properties of PM, only 11 metals
(As, Be, Cd, Co, Cr, Hg, Mg, Ni, Pb, Sb, Se) are currently regulated under US-EPA. Water-
soluble component of PM, particularly metals, are specifically proved to be strongly associated
with the biological ROS activity since their interaction with cells can enhance the generation of
hydroxyl radical, a strong oxidative agent (Prahalad et al., 1999; Valavanidis et al., 2005).
Atmospheric parameters, such as vertical mixing height and wind patterns, as well as emission
sources of transition metals and their strengths, all result in temporal and seasonal variability,
which would consequently cause variations in quasi-ultrafine chemical composition and ROS
activity.
11

Other than temporal and spatial variability, chemical composition (and hence redox activity) of
atmospheric PM is strongly affected by particle size as well. Smaller particles typically have
larger surface area (Hughes et al., 1998) and higher pulmonary deposition efficiency compared
to larger particles (Chalupa et al., 2004) and they are capable of carrying higher proportions of
redox active chemical species that induce inflammatory effects (Kleinman et al., 2003). Ultrafine
particles are typically defined as particles with aerodynamic diameter smaller than 0.1 µm
(Sioutas et al., 2005). This chapter primarily focuses on particles with aerodynamic diameter
smaller than 0.25 µm (which includes the ultrafine size range along with a small portion of
accumulation mode (typically defined as PM
2.5-0.1
) and would be referred to as quasi-ultrafine
henceforth.
For the studies presented in this chapter, quasi-ultrafine samples were collected and analyzed at
10 distinct locations in the Los Angeles air basin for a period of one year, in an effort to identify
the spatial and temporal variations of quasi-ultrafine elemental content. The samples were
chemically analyzed and the concentration of several metals and elements has been quantified
for each season and location. Principal Component Analysis (PCA) was also applied to the
dataset in order to better understand the potential sources of quasi-ultrafine metals and elements
in the area. Moreover, water solubility of these metals was investigated, in order to provide
further information regarding the bioavailability of these species in biological systems.
2.2. Methods
2.2.1. Sampling Location/Technique
Sampling was conducted at 10 locations in the Los Angeles Basin (LAB), including source,
urban and/or near freeway, rural receptor sites downwind (eastwards) of urban Los Angeles, and
12

remote desert-like regions. Fig. 2.1 depicts the location of these sampling sites (designated with
3-letter codes). Throughout this chapter, the sampling sites are clustered as Long Beach
(including HUD site), Los Angeles (including GRD, LDS, CCL, USC, HMS and FRE sites),
Riverside (including VBR and GRA sites) and Lancaster (including LAN site). Geographically
speaking, these clusters are ranging from coast to inland, in increasing order of distance from the
port of Long Beach.
HUD site in Long Beach is considered as a source location in the basin due to its proximity to
several industrial activities in the Long Beach area as well as the Terminal Island freeway and I-
710 freeway, which is majorly impacted by emissions from heavy duty diesel vehicles (Liacos et
al., 2012). Los Angeles site cluster spanning west, east and central Los Angeles, represents an
urban location affected primarily by several freeways replete with sources of vehicular emissions
as well as emissions from the port which are transported with the dominant southwesterly air
stream, particularly in spring and summer. Riverside is a typical semi-rural receptor area located
downwind and therefore influenced by advected particles emitted from source sites. Lancaster,
located in a remote area to the north of the basin represents a desert-like environment
presumably far from both emission sources and advected PM. Selected meteorological
parameters at these site clusters are also presented in Table 2.1. These parameters include
seasonal averages of temperature, precipitation and averages of wind speed/direction and are all
acquired from the online database of California Air Resource Board (CARB). Temperatures are
highest in summer and lowest in winter for all site clusters with Lancaster having the largest
temperature variation among seasons. The general dominant wind direction was from coast to
inland (westerly/southwesterly) with spring and summer showing the strongest wind speeds for
all site clusters.
13

Figure 2.1. Location of the sampling sites in the Los Angeles basin.

Table 2.1. Meteorological parameters.
Site Cluster Season

Temperature
(°C)
Precipitation
(mm)
Wind
Speed (m/s)
(calm%)
Predominant
direction
Long Beach
Spring 16.3±5 12.0 1.6(4.1) SW
Summer 21.3±3.8 0.7 1.9(3.3) SW
Fall 19.9±4.9 54.8 1.2(5.6) W
Winter 13.8±5 168.9 0.7(2.9) NW
Los Angeles
Spring 17.2±5.8 9.7 2.1(0.6) SW
Summer 22.8±4.8 0.0 3.6(1.2) SW
Fall 20.9±5.7 184.9 1.6(4.6) SW
Winter 14.1±5.6 182.6 1.5(0.6) NE
Riverside
Spring 17.4±6.7 1.8 2.6(5) NW
Summer 25.6±6.4 5.6 3.6(2.7) W
Fall 22.0±7.1 1.5 1.6(6.5) NW
Winter 13.8±6.1 78.5 1.3(3.9) N
Lancaster
Spring 15.2±6.6 NA 4.1(13.5) W
Summer 27.9±5.4 NA 4.3(11.6) W
Fall 18.8±7.3 NA 1.5(30.5) W
Winter 8.7±4.7 NA 1.5(33.6) W
14

Ambient quasi-UFP were concurrently collected at the sampling sites on a time-integrated basis
once a week during a weekday, starting at 12:00 A.M PST and ending at 11:59 P.M PST for a
period of one year, using 2 parallel Sioutas personal cascade impactor samplers (Sioutas PCIS,
SKC, Inc, Eighty Four, PA, USA) operating with a flow rate of 9 lpm (Misra et al., 2002; Singh
et al., 2003). One PCIS was loaded with 37 mm Teflon filter (Pall Life Sciences, Ann Arbor, MI)
and the other one with 37 mm quartz filter (Whatman International Ltd, Maidstone, England). In
order to measure the collected mass, the Teflon filters were stabilized under controlled
temperature and relative humidity (21 ⁰C ± 2 ⁰C and 30% ± 5%) then weighed before and after
the sampling using a microbalance (Model MT5, Mettler-Toledo, Inc., Highstone, NJ, USA).
2.2.2. Elemental Analysis
Filters were collected on a weekly basis, with 4-5 weekly samples composited together into one
monthly sample in order to perform chemical analyses. The total elemental composition of the
quasi-UFP samples was determined from these monthly filter composites using a high resolution
magnetic sector inductively coupled plasma sector field mass spectrometry (ICP-SFMS,
Thermo-Finnigan Element 2) (Herner et al., 2006). Sections of the filter membranes were placed
in Teflon digestion bombs with a mixture of 1 mL of 16 M nitric acid, 0.25 mL of 12 M
hydrochloric acid, and 0.10 mL of hydrofluoric acid and the samples solubilized using a
microwave-aided acid digestion system (Milestone ETHOS+). Digestates were diluted to 15 mL
with high purity water (18MΩ/cm
-1
), stored in pre-cleaned low-density polyethylene (LDPE)
bottles then analyzed by SF-ICP-MS. Forty-nine elements were quantified. Propagated analytical
uncertainties were estimated from the uncertainties (square root of the sum of squares method) of
the SF-ICP-MS instrumental analysis, method blanks and digestion recoveries (Herner et al.,
15

2006). The water-soluble portion of the metals and elements was also quantified with a same SF-
ICP-MS analytical method, but using 10 ml Milli-Q water (Millipore, Bedford, MA, USA) for
extraction.
2.2.3. Principal Component Analysis (PCA)
To identify potential sources of metals and elements in quasi-UFP, Principal Component
Analysis (PCA) was applied to the monthly concentrations of PM0.25 elemental species. PCA
was conducted using SPSS statistical software (version 16.0). A VARIMAX rotation was
employed for interpretation of the PCs (Schaug et al., 1990) and factors with eigenvalues greater
than unity were retained in the analysis. Species that exhibited a signal-to-noise ratio (S/N)
above unity in more than 80% of the samples were first selected as variables in the analysis. For
these considered elements, concentration values with a S/N ratio below unity were replaced by a
random value between zero and the detection limit of the given specie in order to avoid
misleading correlations between variables. The number of variables was then further reduced
based on statistical measures of their sampling adequacy and minimally adequate sample size
requirement. 12 monthly data points were included in the PCA analysis for each of the 10
sampling sites. Therefore we had a total of 120 data points (samples) with 27 metals as variables,
resulting in a sample/variable ratio consistent with recommended criteria for a robust PCA
analysis suggested in earlier publications.
2.3. Results and Discussion
2.3.1. Sources of Elements
Table 2.2 summarizes the mass concentrations of quasi-UFP and elements, averaged annually for
different site clusters across the basin. All of the concentrations in the table are arithmetic means
16

of twelve monthly concentrations with their standard errors. The annual average quasi-ultrafine
mass concentration varies from a minimum of 9.0 µg/m
3
at rural desert-like Lancaster to a
maximum of 11.5 µg/m
3
at Long Beach, which is heavily influenced by port-related activities,
and is also in the proximity of I-710 and Terminal island freeways where the contribution of
mobile sources promotes ultrafine levels as the result of increased primary traffic related
emissions (Westerdahl et al., 2005). These freeways are some of the most significant sources of
PM emissions in the PM
0.25
size range (Ntziachristos et al., 2007). Elemental concentrations
ranged from 0.1-10 (ng/m3) for the trace metals (Ag, La, As, Mo, Co, Cr, V, Ni, Rb, Sr, Cd, Sn,
Sb and Pb) to levels around 50-200 (ng/m
3
) for the more abundant base particle-forming
elements such as S, Al, Fe, Ca, K and Na. Sulfur was the most dominant element reaching a peak
of 378 ng/m
3
at Long Beach followed by Fe, Ca and Al with peak concentrations of respectively
158, 110 and 107 ng/m
3
all observed at Riverside. The highest quasi-UF elemental
concentrations were generally observed at either the source site of Long Beach, or at the receptor
Riverside sites, while the minimum elemental concentrations are most frequently associated with
remote Lancaster site.
To identify potential sources of metals and elements in quasi-UFP, Principal Component
Analysis (PCA) was applied to the monthly concentrations of PM
0.25
elemental species. PCA was
conducted using SPSS statistical software (SPSS Inc., version 16.0). A VARIMAX rotation was
employed for interpretation of the PCs (Schaug et al., 1990) and factors with eigenvalues greater
than unity were retained in the analysis. Species that exhibited a signal-to-noise ratio (S/N)
above unity in more than 80% of the samples were first selected as variables in the analysis. For
these considered elements, concentration values with a S/N ratio below unity were replaced by a
random value between zero and the detection limit of the given specie in order to avoid
17

misleading correlations between variables. The number of variables was then further reduced
based on statistical measures of their sampling adequacy and minimally adequate sample size
requirement. PCA was primarily applied to a combined dataset from all ten sites. Further PCA
runs were also performed on pooled data from two distinct site clusters in order to compare the
elemental sources at different locations. The first cluster represents urban areas and includes
Long Beach and Los Angeles sites (HUD, LDS, CCL, GRD, USC, FRE, and HMS). In contrast,
the other cluster represents rural receptor locations and consists of rural Riverside (GRA, VBR)
and Lancaster (LAN) sites.
Table 2.2 Annual average and standard error (SE) of the mass concentration (ng/m
3
) of
quasi-UFP (dp<0.25 µm) and its elemental constituents at Long Beach (HUD) , Los Angeles
(GRD, LDS, CCL, USC, HMS, FRE), Riverside (VBR, GRA) and Lancaster (LAN)
Long Beach Los Angeles Riverside Lancaster
Average SE* Average SE Average SE Average SE
Mass Conc.
(µg/m3) 11.5 1.13 9.15 0.33 11.1 0.52 8.97 1.31
Li 0.10 0.03 0.07 0.01 0.16 0.03 0.11 0.03
Na 64.7 10.6 41.4 2.67 66.5 5.24 75.1 16.3
Mg 21.7 6.01 11.8 1.22 30.8 3.61 26.3 5.12
Al 90.2 24.7 46.5 4.58 107 13.1 99.6 17.8
S 378 37.0 290 15.6 262 26.1 190 37.2
K 61.2 14.4 42.7 4.36 88.9 9.32 78.2 23.2
Ca 83.8 23.8 43.8 5.36 110 12.7 67.8 11.6
Ti 11.2 3.09 6.76 0.69 10.7 1.10 12.0 3.26
V 8.54 0.87 3.72 0.19 2.36 0.18 0.72 0.16
Cr 1.43 0.35 1.22 0.17 1.48 0.51 0.68 0.11
Mn 3.07 0.79 1.78 0.17 3.76 0.42 3.09 0.78
Fe 148 37.8 111 10.9 158 17.5 151 27.2
Co 0.14 0.02 0.05 0.00 0.07 0.01 0.06 0.02
Ni 2.45 0.27 1.38 0.11 1.23 0.23 0.38 0.08
Cu 5.84 1.71 9.02 1.20 8.93 1.15 4.01 0.57
Zn 17.0 3.09 8.05 0.89 10.0 1.07 5.95 1.48
18

As 0.24 0.04 0.22 0.01 0.27 0.02 0.13 0.01
Rb 0.14 0.04 0.09 0.01 0.22 0.02 0.18 0.05
Sr 1.08 0.26 0.83 0.08 1.16 0.11 0.91 0.15
Mo 0.68 0.18 0.44 0.03 0.45 0.06 0.14 0.02
Ag 0.11 0.04 0.08 0.01 0.17 0.04 0.06 0.02
Cd 0.06 0.01 0.06 0.01 0.09 0.03 0.02 0.00
Sn 0.87 0.24 2.06 0.74 0.70 0.09 8.36 8.77
Sb 0.94 0.25 1.17 0.12 1.08 0.12 0.51 0.07
Ba 6.44 1.81 6.22 0.66 5.44 0.58 4.89 0.58
La 0.17 0.02 0.12 0.01 0.11 0.01 0.06 0.01
Pb 2.01 0.46 1.79 0.13 2.52 0.32 0.98 0.13

The results of the PCA for the combination of all ten sampling sites are presented in Table 2.3.
PCA results corresponding top rural and urban site clusters are also presented in Tables 2.4 and
2.5, respectively. Five distinct principal components are observed for the considered dataset
(Table 2.3). The first principal component (PC1) consists of both mineral elements (such as Mg
and Ca) and transitional metals (such as Fe, Co, Ti, Mn), and likely represents road dust enriched
with vehicular-related elements originating from sources such as lubricating/motor oil
(Ntziachristos et al., 2007) or tire wear (Dahl et al., 2006). These tire wear particles may include
metals used in softening filler oils emitted from tire-pavement interaction (Gustafsson et al.,
2008). This principal component contributes to about 36% of the total variance, which
underscores the significance of road dust in the overall metal and element composition of
particulate matter in quasi-ultrafine mode. While road dust has been identified as a major source
of coarse particle emissions (Amato et al., 2009), other studies have consistently been reporting
road dust emissions originating from tire-pavement interactions to be present in the smaller size
fractions as well, with mass mean particle diameters as low as 15-50 nm (Dahl et al., 2006;
Gustafsson et al., 2008).
19

The second principal component (PC2) shows significant loadings (greater than 0.7) for metals
abundant in brake wear such as Ba, Sb and Cu. Sanders et al (Sanders et al., 2003). showed that
emissions associated with brake wear can be found in the ultrafine size range, suggesting that the
potential source of PC2 is strongly impacted by abrasive vehicular emissions. When PCA is
applied to the urban sites, these metals appear in the road dust component (Table 2.5). However,
when PCA is applied to the receptor sites, Ba and Cu form a separate component (Table 2.4).
This observation is likely due to the dominance of vehicular emissions at the urban sites, which
makes it difficult to distinguish other emission sources of metals. Compared to all sites
combined, road dust constitutes a larger fraction of total variance in the system when PCA is
applied only to the urban sites (62% for urban sites versus 36% for all sites combined) and
loading factors also indicate the increased contribution of transitional metals originating from
primary vehicular emissions to the variance, due to the location of these urban sites near high
traffic freeways. In the rural receptor sites, however, road dust has a potentially more natural
origin, which makes it easier to distinguish between vehicular abrasion and road dust in these
areas.
In the third principal component (PC3), high loadings (greater than 0.8) of V, La and S as tracers
of ship emissions (Arhami et al., 2009) are observed which suggest this component to be
influenced by emissions from port of Long beach and the burning of other residual oils such as
fuel oil. High Sulfur concentrations can also be influenced by photochemical activities leading to
sulfate generation in the atmosphere. This component disappeared in PCA applied to the rural
receptor sites (Table 2.4).
20

Interpretation of the fourth principal component (PC4) is somehow unclear since this component
only shows a strong loading for Cd (0.869). Other elements including Pb and Ag had moderate
loadings (0.6-0.7). It is also noteworthy that Cd, Pb and Ag are grouped with the first component
(road dust) for the PCA of the urban sites (including Long Beach and Los Angeles) while they
form a separate component when PCA is applied to the rural receptor site cluster. This trend
suggests that this PC may potentially represent a local, non-vehicular source, enriched in
emissions from a variety of metallurgical activities (Thomaidis et al., 2003).
In the fifth principal component (PC5), high loadings (>0.8) are observed for Cr and Ni, which is
most probably associated with emissions originating from various industrial sources in the
vicinity of our sampling sites and metal plating activities in particular. All of these five
components together explain about 84% of the total variance in the dataset.
21

Table 2.3. Principal component loadings (VARIMAX normalized) of selected elements in
quasi-UFP (dp<0.25 m) for all sampling sites combined (HUD, GRD, LDS, CCL, USC,
HMS, FRE, VBR, GRA, LAN)
Elements
Principal Component
PC1: Road
Dust
PC2:
Vehicular
Abrasion
PC3: Residual
Oil
Combustion
PC4:
Cadmium
Sources
PC5: Metal
Plating
Rb .949 .173 -.030 .135 .078
Mg .946 .187 .028 .135 .081
Al .919 .224 .059 .135 .109
K .907 .235 .075 .130 .041
Mn .886 .347 .024 .143 .109
Ca .856 .358 .048 .212 .117
Ti .850 .403 .025 .094 .081
Na .848 .075 .223 .097 .146
Li .808 .249 .014 .323 .086
Fe .768 .583 -.010 .136 .136
Sr .683 .586 .042 .184 .066
Co .627 .276 .461 .056 .297
Rh .282 .868 .017 .173 .136
Ba .422 .835 -.002 .142 .060
Sb .243 .793 -.016 .350 .026
Cu .221 .756 -.005 .144 .150
Mo .257 .714 .170 .191 .389
As .339 .511 .068 .455 .062
Zn .452 .478 .477 .327 .033
S .049 -.201 .874 -.047 .003
La .254 .170 .847 .049 -.017
V -.170 .031 .808 -.029 .232
Cd .137 .225 .105 .869 .007
Ag .193 .229 -.133 .687 .063
Pb .381 .538 .063 .624 .123
Cr .281 .218 -.022 .044 .871
Ni .128 .207 .492 .087 .810
Eigenvalue 14.68 2.95 2.62 1.35 0.98
% Variance 35.8 20.7 10.9 9.2 6.9
% Cumulative
variance
35.8 56.5 67.4 76.6 83.5
Note: Variables with loading factors above 0.7 are in bold font and variables with loading factors
between 0.5 and 0.7 are underlined.

22

Table 2.4. Principal component loadings (VARIMAX normalized) of selected elements in
quasi-UFP (dp<0.25 um) for rural sites cluster (GRA, VBR and LAN).
Elements
Principal Component
PC 1 PC 2 PC 3 PC 4
Ti .968 .067 .066 .002
K .944 .073 .036 .172
Fe .899 .351 .026 .192
Mn .884 .351 .065 .132
Co .878 .162 .126 .331
Mg .877 .255 -.042 .247
Al .855 .309 -.041 .214
Na .855 -.297 .091 .211
Sr .838 .390 .134 .278
Ca .701 .557 .024 .362
Zn .415 .360 .344 .381
Mo .126 .802 .320 .100
Ba .567 .687 -.029 .241
As .205 .664 .450 .318
Cu .187 .640 .570 .136
Cd -.083 .003 .886 .183
Pb .156 .278 .774 .384
Ag -.018 .240 .766 -.227
Cr .331 .165 .074 .871
Ni .373 .261 .198 .833
Eigenvalue 11.44 3.36 1.19 1.18
% Variance 42.403 16.587 14.103 12.862

23

Table 2.5. Principal component loadings (VARIMAX normalized) of selected elements in
quasi-UFP (dp<0.25 um) for urban sites cluster (HUD, GRD, LDS, CCL, USC, HMS,
FRE).
Elements
Component
PC 1 PC 2 PC 3
Ca .970 .085 .097
Mn .960 -.035 .208
Rb .949 .042 .136
Mg .945 .138 .139
Fe .936 -.137 .260
Li .926 .111 .064
Ti .921 -.055 .200
K .917 .159 .064
Al .916 .189 .156
Ba .915 -.186 .183
Pb .906 -.022 .056
Zn .872 .281 .026
Sr .861 -.074 .102
Sb .853 -.190 .140
Cd .848 .256 -.058
Co .815 .177 .382
Na .742 .331 .370
Ag .741 -.200 .116
As .697 -.003 -.014
Cu .689 -.174 .287
S -.081 .918 .113
V -.170 .836 -.028
La .410 .805 .111
Cr .178 -.106 .959
Ni .129 .363 .880
Eigenvalue 16.13 2.91 1.74
% Variance 62.005 11.563 9.604

24

2.3.2. Water Solubility of Elements
Increased solubility of trace elements and metals has been linked to increased bio-availability to
human cells in both in vivo (Roberts et al., 2004) and in vitro studies. Figure 2.2 shows the
concentration of the water-soluble fraction of quasi-UF trace elements and metals at different site
clusters. Sampling sites were clustered into Long Beach (HUD), Los Angeles (GRD, LDS, CCL,
USC, HMS and FRE) and Riverside (VBR and GRA, LAN). The concentrations have different
orders of magnitude, ranging from values as low as 0.001-0.1 ng/m
3
(e.g. La, Ti, Co and Cd) to
higher values in the range of 1-100 ng/m
3
for metals such as Zn, Mg, K, Fe, Cu and V. Metals
with dominant vehicular sources (such as Zn, Ni, Fe, Cr and Mo) are generally showing higher
concentrations at either source (Long Beach) or urban (Los Angeles) site clusters compared to
rural receptor locations represented by Riverside, due to the increased strength of their emission
sources at Long Beach and Los Angeles area. Metals that more dominantly originate from re-
suspended soil and dust, on the other hand, have generally comparable concentrations at all site
clusters (e.g. Al, Ti and K). The average water-solubility of selected elements, including many
redox active metals (such as Cr, Cu, Fe, Mn, Mo, Ni and V), calculated as the mass ratio of
water-soluble concentration to total concentration, is depicted in Figure 2.3. The metals with the
highest water soluble fraction were Zn and Cd with 92% and 87% average solubility,
respectively. Ni, V, Cu, Co, Ba, Mn, Mo, Pb and As, all metals with dominant anthropogenic
sources such as residual oil combustion, vehicular abrasion and industrial and tailpipe emissions
displayed moderate water-solubility, with averages ranging from 38% to 69%. The lowest water-
solubility was observed for Ti and Fe (0.5% and 3.6% respectively), which both originate from
road dust emissions, particularly re-suspended soil and dust.
25

Figure 2.2. Concentration of water soluble fraction of

PM
0.25
metals (yearly average).
Error bars represent one standard deviation.

Figure 2.3. Yearly average water solubility of PM
0.25
metals. Error bars represent one
standard deviation.
26

2.3.3. Spatial and Seasonal Variation of Elemental Groups
Temporal variability of the elemental sources identified in the PCA analysis is shown in Figures
2.4 to 2.8 (a-d), for the 5 sources and 4 spatial clusters. The variances shown in Table 2.3
indicate that road dust was the dominant component at all of the site clusters, explaining as much
as 36% of the total variance when PCA is applied to all sites combined. As noted earlier, for
urban site cluster (including Los Angeles sites and Long Beach), the contribution of road dust to
the total variance of trace elements and metals is 62%, which is considerably higher compared to
all sites combined and rural receptor sites (including rural Riverside sites and Lancaster). Road
dust at the inland sites only accounted for 42% of the total variance. The presence of mineral
elements such as Mg, Al, K, Na, Sr, Mn, Fe and Ti in this component identifies the contribution
of road dust re-suspension caused by tire-pavement interactions. While PCA is applied only to
urban sites, particular transitional elements such as Co, Cd, Zn and Pb indicate higher loadings
(between 0.815 and 0.906) in this component associated with road dust compared to the case
when all sites are combined. According to previous studies, biomass burning can be another
potential source of potassium emissions. Our PCA results, however, indicate a very high loading
of K (0.90) in the first principal component and no significant loadings in any other component,
which suggests that road dust is the dominant source of K in the basin. The seasonal variation of
selected elements associated with road dust emissions are plotted in Fig. 2.4 (a-d) for four site
clusters (Long Beach, Los Angeles, Riverside and Lancaster). The patterns of seasonal variation
are different among site clusters. In summer and spring, the rural Riverside and Lancaster sites
consistently show higher concentrations possibly due to the higher southwesterly wind speed,
which helps transporting these metals in the quasi-UFP mode to the rural receptor sites
(Riverside, Lancaster) located downwind. In winter, however, the highest concentrations are
27

observed at Long Beach and the lowest at Lancaster which could be attributed to the lower
mixing height and decreased advection due to stagnation and lower wind speeds, resulting in
accumulation of these metals at the source site.
The second principal component when all sites are combined is identified as abrasive vehicular
emissions due to the high (above 0.7) or moderately high (between 0.5 and 0.7) loading of metals
such as Ba, Cu, Sb and Fe. All of these metals are previously linked to brake lining emissions as
well as tire wear. This principal component explains almost 21% of the total variance in the data
set. Fig. 2.5 (a-d) presents the seasonal variation of selected metals (Ba, Sb, Cu) associated with
abrasive vehicular emissions for different site clusters. All metals in this component show similar
spatial trends across site clusters with the highest concentrations occurring at either Riverside or
Lancaster during summer and spring. This trend could be related to increased advection from
upwind urban sites during these seasons. In contrast, during the colder seasons (i.e. winter and
fall), the concentration of these metals is higher at the source and urban sites (Los Angeles and
Long Beach), which could be attributed to the decreased mixing height and wind speed, resulting
in the accumulation of these elements near the urban locations.
When PCA is applied to all sites combined, S, V and La show significant loading (greater than
0.8) in PC3. Sulfur and vanadium are known as typical tracers of ship emissions, refinery
operations and residual oil combustion in all size ranges including the quasi-ultrafine mode in
our case. According to Table 2.3, this component contributes to as much as 11% of the total
variance in the dataset for all sites combined and also the urban sites. For the receptor sites,
however, these elements do not appear as an independent component, most likely due to the
location of these sites away from the coast, where most of these particles originate from ship
emissions and oil refineries in the area. Monthly variation of S, V and La at each site cluster is
28

depicted in Fig. 2.6 (a-d). S, La and V display overall similar trends. The highest variability is
observed at Long Beach, which is closest to the source location and therefore more sensitive to
local ship plumes. Occasional local increase in the port activities in particular months can change
the concentration of these elements in the vicinity of source locations. The concentration of V
and La consistently peaks at Long Beach, then decreases quite dramatically as we go from the
coast to the inland sites. The concentration decrease from the source site (Long Beach) to the
inland receptor sites (Riverside, Lancaster) is also more substantial during colder months (i.e.
December, January February), most likely due to decreased regional mass transport in winter
compared to summer and spring. We should also note that sulfur displays a considerably higher
winter time (i.e. December, January and February) concentration at Long Beach (281 ng/m3)
compared to downwind site clusters (61.4, 123 and 60.7 ng/m3 for Los Angeles, Riverside and
Lancaster, respectively), which indicates the dominance of local emissions from port-related
activities during the colder periods. During warmer months, however, sulfur has overall a more
uniform concentration across the basin, indicating that the majority of this species in the quasi-
ultrafine range is related to ammonium sulfate produced by photochemical reactions that occur
fairly uniformly across the basin.
Figure 2.7 (a-d) shows the temporal variability of the elements in PC4. Cd is the only element
with significant loading factor (0.869) in component four followed by Ag and Pb with loading
factors of 0.687 and 0.624 respectively. Cadmium has a variety of emission sources such as
plastic production and industrial metallurgical activities. An earlier study comprehensively
identified more than ten different potential sources leading to cadmium emissions in European
cities, including, but not limited to coal, wood, gasoline and oil combustion and cement
production (Pacyna et al., 1984). Steel production industries as well as phosphate fertilizers
29

extensively used in rural and agricultural locations are also introduced as other potential sources
of cadmium emissions into the atmosphere. At urban site clusters, Cd, Pb and Ag appear with
high loading factors (respectively 0.848, 0.906 and 0.741) in the first component associated with
road dust at urban sites implying that vehicular emissions are the dominant source of these
elements. When rural receptor sites are clustered with urban sites, however, the contribution of
non-vehicular related emissions to the variance is more distinguishable from road dust and
therefore these metals (with Cd being the dominant element) show up in the PCA results as a
separate component, indicating the predominance of aforementioned sources.
Fig. 2.8 (a-d) presents monthly patterns of Ni and Cr concentrations (i.e. PC 5) at each site
cluster. The concentrations are generally higher in colder periods (i.e. from October to January)
likely due to the lower mixing height. Ni and Cr concentrations are highest during December and
January at Long Beach followed by Los Angeles, Riverside and Lancaster, which could be
related to emissions from power plants and refineries in the Long Beach area. Another
observation here is the effect of plume events causing occasional abrupt changes associated with
monthly concentrations. One example can be the unexpected peak for Ni and Cr at Riverside in
November (3.18 and 6.22 ng/m3 respectively), which could be related to an increase in a local
source strength. The other example is noticeable peaks for Cr at Riverside in March, at Lancaster
in April and at Los Angeles in July and August. The peak for Cr at Riverside in March could also
be related to the increased long-range PM transport for this metal. As much as 70-80% of Cr is
associated with sub-micrometer particles (dp< 1 µm), with over 40% of the mass concentration
therefore transport capability in the atmosphere.
30

Figure 2.4 (a-d). Seasonal variation of selected elements for PC1 (road dust) at (a) Long Beach, (b) Los Angeles, (c) Riverside and (d) Lancaster.

31

Figure 2.5 (a-d). Seasonal variation of selected elements for PC2 (vehicular abrasion) at (a) Long Beach, (b) Los Angeles, (c) Riverside and (d) Lancaster.

32

Figure 2.6 (a-d). Seasonal variation of selected elements for PC3 (residual oil combustion) at (a) Long Beach, (b) Los Angeles, (c) Riverside and (d)
Lancaster.

33

Figure 2.7 (a-d). Seasonal variation of selected elements for PC4 (cadmium sources) at (a) Long Beach, (b) Los Angeles, (c) Riverside and (d) Lancaster.

34

Figure 2.8 (a-d). Seasonal variation of selected elements for PC5 (metal plating) at (a) Long Beach, (b) Los Angeles, (c) Riverside and (d) Lancaster.

35

2.4. Summary and Conclusions
In this chapter, monthly data of quasi-UFP and their constituent elements were studied in a year-
long sampling campaign at ten locations across the LAB. PCA was applied to site-pooled
elemental data as well as to two distinct site clusters representing urban and rural receptor
locations. Five components were identified in the quasi-ultrafine mode, namely road dust,
vehicular abrasion, residual oil combustion, cadmium sources and metal plating. The elemental
species marking road dust and metal plating were found at all of the site clusters. On the other
hand, the components associated with vehicular abrasion and cadmium sources were observed as
separate groups when only rural receptor sites were included in the analysis while they were
combined with the dominant road dust component when urban sites were considered in the PCA.
The component related to residual oil combustion (strongly affected by ship emissions and
refineries) appeared in the urban site cluster, including HUD as the source site near the port of
Long Beach. The major constituents of each component are highlighted as Ca, Mg, Na, K, Fe, Ti
and Al for road dust; Cu, Ba, Ca, Sb, Fe and Mo for vehicular abrasion; S, V and La for residual
oil combustion; Cd for the fourth component designated as cadmium sources; and Cr and Ni for
metal plating. The seasonal and spatial variation of these elemental groups was subsequently
investigated to provide an additional insight on their sources and transport patterns. The
elemental groups linked to metal plating and residual oil combustion generally showed a
decreasing trend in the transect from the coast to the inland receptor sites while occasional peaks
for some elements were observed at Riverside during summer and fall period when the dominant
westerly/southwesterly winds more effectively transport quasi-UFP from upwind source and
urban sites to the receptor locations. Although airborne particles in the quasi-ultrafine size range
are believed to exhibit enhanced biological activity in comparison with larger-sized particulate
36

matter due to their deep penetration and deposition in the respiratory system, they are the least
regulated particles under US-EPA. The findings of this chapter further clarify the complicated
source profile and transport patterns of this significant size range and can thus help the
progressive regulations on PM emissions in the megacity of Los Angeles.
2.5. Acknowledgements
This study was funded by the South Coast Air Quality Management District (SCAQMD) (award
#11527). We also would like to thank the staff at the Wisconsin State Laboratory of Hygiene for
their assistance with the chemical analyses. We also wish to acknowledge the support of USC
Provost’s and Viterbi’s Ph.D. fellowships.
37

Chapter 3:
Seasonal and Spatial Variability of ROS Activity in Los Angeles
In this chapter, seasonal and spatial variation in redox activity of quasi-ultrafine particles
(PM0.25) and its association with chemical species was investigated at 9 distinct sampling sites
across the Los Angeles metropolitan area. Biologically reactive oxygen species (ROS) assay
(generation of ROS in rat alveolar macrophage cells) was employed in order to assess the redox
activity of PM0.25 samples. Seasonally, fall and summer displayed higher volume-based ROS
activity (i.e. ROS activity per unit volume of air) compared to spring and winter. ROS levels
were generally higher at near source and urban background sites compared to rural receptor
locations, except for summer when comparable ROS activity was observed at the rural receptor
sites. Univariate linear regression analysis indicated association (R>0.7) between ROS activity
and organic carbon (OC), water soluble organic carbon (WSOC) and water soluble transition
metals (including Fe, V, Cr, Cd, Ni, Zn, Mn, Pb and Cu). A multivariate regression method was
also used to obtain a model to predict the ROS activity of PM0.25, based on its components. The
most important species associated with ROS were Cu and La at the source site of Long Beach,
and Fe and V at urban Los Angeles sites. At Riverside, a rural receptor location, WSOC and Ni
(tracers of secondary organic aerosol and metal plating, respectively) were the dominant species
driving the ROS activity.

This chapter is based on the following publication:
Saffari, A.; Daher, N.; Shafer, M. M.; Schauer, J. J.; Sioutas, C. Seasonal and spatial variation in
reactive oxygen species activity of quasi-ultrafine particles (PM0.25) in the Los Angeles
metropolitan area and its association with chemical composition. Atmos. Environ. 2013, 79, 566–
575.

3.1. Introduction
There is growing literature supporting the hypothesis that one of the important pathways
underlying the adverse health endpoints of exposure to particulate matter is the oxidative stress
(e.g. ROS generation) that derives from the interaction of PM with cells (Donaldson et al., 2002).
38

Elevated ROS levels can alter the redox status of the cell and consequently trigger a series of
acute and chronic responses such as pulmonary inflammation(Squadrito et al., 2001) and
mitochondrial damage (Li et al., 2003). Previous studies have indicated that particle size is one
of the most important factors mediating the health effects of PM (Li et al., 2009). Smaller
particles typically have larger surface area (Hughes et al., 1998) and higher pulmonary
deposition efficiency compared to larger particles (Chalupa et al., 2004) and they are capable of
carrying higher proportions of redox active chemical species that induce inflammatory effects
(Kleinman et al., 2008).
Ultrafine particles are typically defined as particles with aerodynamic diameter smaller than 0.1-
0.2 µm (Sioutas et al., 2005). Similar to chapter 2, this chapter also focuses on particles with
aerodynamic diameter smaller than 0.25 µm and would be referred to as quasi-ultrafine
henceforth. This size range is particularly important from a public health perspective. In a series
of cohort studies recently conducted in the Los Angeles basin (LAB), positive associations were
observed between quasi-ultrafine particles and biomarkers of adverse health effects (systematic
inflammation, platelet activation and ambulatory ST-segment depression, to name a few), while
poor or no association was found for larger size ranges (i.e. PM0.25-2.5 and PM2.5-10) (Delfino
et al., 2011, 2010, 2009, 2008). Moreover, higher ROS activity was induced by quasi-ultrafine
particles compared to larger size ranges, which likely indicates the higher potency of these
particles to initiate the aforementioned negative health effects (Hu et al., 2008).
Previous studies have reported water-soluble component of PM, particularly metals, to be
strongly associated with the biological ROS activity since their interaction with cells can
enhance the generation of hydroxyl radical, a strong oxidative agent (Goldsmith et al., 1998;
Prophete et al., 2006). Previous studies conducted in the LAB have investigated the relation of
39

redox potential with chemical constituents of PM and found positive associations between redox
potential measured by ROS activity and transition metals such as Fe, V, Ni, Cr, Cd, Zn and Pb
(Verma et al., 2010). Atmospheric parameters, such as vertical mixing height and wind patterns,
as well as emission sources of transition metals and their strengths, all result in temporal and
seasonal variability, which would consequently cause variations in quasi-ultrafine chemical
composition and ROS activity.
In this chapter, quasi-ultrafine samples were collected and analyzed at 9 distinct locations in the
Los Angeles basin for a period of one year, in an effort to identify the spatial and temporal
variations of PM0.25-associated ROS activity, and more importantly to provide more
comprehensive insights on the dominant emission sources driving the ROS activity of PM0.25,
which is essential for establishing more targeted PM regulations in the basin. The chemical
composition of quasi-ultrafine particles collected at the same sites, as well as sources of trace
elements and metals have been previously identified and reported in Chapter 2. The current
Chapter now focuses on the seasonal and spatial variation of the ROS cellular activity associated
with the quasi-ultrafine PM samples discussed above. Moreover, univariate and multivariate
regression analysis have been used to better understand the association of chemical species (and
their corresponding sources) with ROS activity and to obtain a model for predicting quasi-
ultrafine ROS activity based on its water soluble chemical content.
3.2. Methods
3.2.1. Sampling Sites
Samples were collected from nine distinct sites, each designated with a three-letter code, across
the Los Angeles air basin, as described in chapter 2, Figure 2.1. HUD is located in the Long
Beach area and is considered as a source site, given its vicinity to PM0.25 emissions associated
40

with industrial and port activities in the Long Beach area as well as Terminal Island freeway and
the I-710 freeway (Daher et al., 2013). GRD, LDS, CCL, USC, HMS and FRE are all located in
urban locations spanning west, east and central Los Angeles and are considered as urban sites.
PM at these locations is dominated by vehicular emissions from nearby freeways or surface
streets as well as particles advected from the source site HUD. VBR and GRA, which are located
further inland in rural/semi-rural regions in Riverside, are considered as receptor sites located
along the prevailing air trajectory from coast to inland. Temperatures were within seasonal
norms (highest in summer and lowest in winter) and highest wind speeds were observed during
spring and summer with a dominant westerly/southwesterly direction, transporting the particles
from coast to inland sites. Further details regarding the coordinates of sampling sites and
meteorological parameters can be found in chapter 2, as well as previous publications (Pakbin et
al., 2011).
3.2.2. Sampling Protocol
Time-integrated 24-hour quasi-ultrafine (PM0.25) samples were collected once a week during a
weekday (i.e. Monday, Tuesday, Wednesday, Thursday or Friday), starting at 12:00 A.M and
ending at 11:59 P.M, from April 2008 to March 2009. Particles were collected by inertial
separation method, using two parallel Sioutas personal cascade impactor samplers (Sioutas PCIS,
SKC, Inc., Eighty Four, PA, USA), each operating at a flow rate of 9 LPM, as described in
Chapter 2. One PCIS was loaded with 37 mm Teflon filters (Pall Life Sciences, Ann Arbor, MI)
and the other one with 37 mm quartz filters (Whatman International Ltd, Maidstone, England).
Mass concentrations were determined by pre-weighing and post-weighing the filters using a
microbalance (Model MT5, Mettler-Toledo, Inc., Highstown, NJ, USA), after equilibration
41

under controlled laboratory conditions (Temperature of 19-23 ⁰C and Relative humidity of 25-
35%).
3.2.3. Chemical/Toxicological Analyses
To measure the total concentration of elements, sections of monthly composited filter
membranes were digested in an acid mixture (containing HNO3, HF and HCl) and solubilized in
a Teflon digestion bomb using a microwave-assisted digestion system (Milestone ETHOS+).
The digestates were then analyzed using a high resolution inductively coupled plasma sector
field mass spectrometry (ICP-SFMS, Themo-Finnigan Element 2). The water-soluble portion of
the metals and elements was also quantified with a same ICP-SFMS analytical method, but using
10 ml Milli-Q water (Millipore, Bedford, MA, USA) for extraction. Elemental and organic
carbon (EC and OC) were quantified by the NIOSH Thermal Optical Transmission method and
ion chromatography (IC) was used to measure concentration of water soluble inorganic ions. The
water-soluble organic carbon (WSOC) content of the samples was determined using a Sievers
900 Total Organic Carbon Analyzer, following water-extraction and filtration. Further details
about these analytical methods can be found elsewhere (Herner et al., 2006; Shafer et al., 2010).
To quantify the biological reactive oxygen species (ROS) activity of the PM samples, filters
were extracted with 1 ml Milli-Q water and extracts were filtered through 0.22 µm
polypropylene syringe filters. Rat alveolar macrophage cells (cell line NR8383) were then
exposed to the PM extracts mixed with 2',7'-dichlorodihydrofluorescein diacetate (DCFH-DA).
DCFH-DA is cell-wall permeable and de-acetylated in the cytoplasm to 2',7'-
dichlorodihydrofluorescein (DCFH). The non-fluorescent DCFH was converted to the highly
fluorescent 2,7, dichlorofluorescein (DCH) by the ROS species produced within the cell. The
increase in fluorescence intensity of the samples compared to positive controls (Zymosan, Urban
42

Dust extracts) and negative controls (method blank), was monitored by a fluorescent plate reader
and ROS activity was reported in units of Zymosan equivalents. Additional details of the
macrophage-ROS protocol and method performance can be found in (Landreman et al., 2008).
3.3. Results and Discussion
3.3.1. Spatiotemporal Variations of ROS activity
The ROS activity of quasi-ultrafine particles, normalized by the volume of air (expressed in µg
Zymosan per m
3
air), at each sampling site is shown in Figure 3.1 (a-d) for different seasons.
This volume-based ROS activity, which reflects the potential toxicity associated with exposure
to inhaled PM, is generally higher during summer and fall (averaging 69.1 and 73.8 µg
Zymosan/m
3
, respectively, across sites) compared to spring and winter (42.9 and 27.3 µg
Zymosan/m
3
, respectively). At receptor sites VBR and GRA, ROS activity is highest during
summer (63.1 and 92.3 µg Zymosan/m
3
respectively), likely due to increased secondary organic
aerosol generation during the transport/aging of particles from source and urban areas to the
receptor sites (Sardar et al., 2005). At GRD, LDS, CCL, USC, HMS and FRE, all representing
urban and/or near-freeway locations spanning west, central and eastern Los Angeles, the ROS
activity was consistently highest during either fall (GRD: 107 µg Zymosan/m
3
, USC: 101 µg
Zymosan/m
3
, HMS: 76.8 µg Zymosan/m
3
and FRE: 64.6 µg Zymosan/m
3
) or summer (LDS:
63.8 µg Zymosan/m
3
and CCL: 81.4 µg Zymosan/m
3
). Seasonal and spatial variations of mass-
based ROS activity (i.e. normalized by total quasi-ultrafine PM mass, expressed in µg Zymosan
per mg of PM), is shown in Figure 3.2. (a-d). Comparing Figures 3.1. and 3.2, it is evident that
mass-based ROS activity which reflects the intrinsic toxicity of particles, showed very similar
43

trends to volume-based ROS activity, indicating that PM composition, more than PM mass
concentration was driving ROS activity.
44

Figure 3.1 (a-d). Volume-based ROS activity at different sampling sites during: (a) spring, (b) summer, (c) fall and (d) winter. Error bars
correspond to one standard deviation. Dashed lines indicate the average of all sampling sites.

45

Figure 3.2 (a-d). Mass-based ROS Activity at different sampling sites during: (a) spring, (b) summer, (c) fall and (d) winter. Error bars
correspond to one standard deviation. Dashed lines indicate the average of all sampling sites.

46

3.3.2. Association of ROS Activity and Chemical Components
Regression analysis was carried out in order to investigate the association of ROS activity with
chemical species. As the first step, univariate regression was employed between volume-based
ROS and PM chemical components, in order to examine how individual species correlate with
ROS activity and to also obtain insight about important species to be included in the multivariate
regression analysis. Table 3.1 shows Pearson correlation coefficients (R) between monthly ROS
activity data and monthly concentrations of carbonaceous material, water soluble metals and
inorganic ions, at each sampling site. Fe and V generally show the highest correlations (R>0.7)
with ROS activity across the basin. At the source site HUD, La highly correlates with ROS
activity (R=0.86). This metal is one of the tracers of residual oil combustion, as found in Chapter
2 of this dissertation. At GRD, LDS, CCL, USC, HMS and FRE, all located in the urban area of
Los Angeles, ROS activity is correlated (R>0.7) with several transition metals such as Fe, V,
Mn, Ni and Cr. Strong association between these transition metals and ROS activity was also
reported in other studies conducted in Los Angeles (Verma et al., 2009) and other urban areas in
the world such as Milan, Italy (Daher et al., 2012) and Lahore, Pakistan (Shafer et al., 2010).
Some correlations are also observed between ROS activity and Al, Mg and Ti. These metals
primarily originate from re-suspended soil and road dust and their correlation with ROS activity
is likely due to their co-variation with redox active Fe, which shares the same source (road dust)
with them in the quasi-ultrafine size range, as discovered in Chapter 2. Water soluble organic
carbon (WSOC) has the highest correlations with ROS activity at the two Riverside sites
(R=0.68 and 0.72 at VBR and GRA , respectively). WSOC has been previously linked to the
formation of secondary organic aerosols (SOA), particularly at rural areas (Snyder et al., 2009;
Weber et al., 2007). Its correlation with ROS at these sites, located at rural receptor locations of
47

Riverside County, is likely associated with the redox potential of SOA, found in higher
proportions at these two sites. Organic carbon (OC), which includes both water soluble and
insoluble carbonaceous species, is correlated with ROS at USC and GRA (R=0.91 and 0.79,
respectively). Correlations between ROS activity and EC are generally higher at near freeway
urban sites (e.g. GRD, USC and FRE, R= 0.64, 0.69 and 0.67, respectively) compared to
receptor sites (VBR and GRA, R= -0.22 and 0.32, respectively). EC is not known as an
intrinsically redox active species. It is, however, a typical tracer of primary vehicular exhaust
(Schauer, 2003) and therefore shares the same source with some of the redox active transition
metals (such as Fe, Cu, Mn, and Cd). It is noteworthy that the correlation coefficients observed
for both OC and EC at different sites are generally lower compared to transition metals. Water-
soluble inorganic ions such as nitrate and sulfate are not showing any notable correlation with
ROS activity. While these ions do not have any functional group leading to the formation of
ROS, their acidity may contribute to the overall PM toxiciy. To investigate the role of metals’
water-solubility on ROS induction, Pearson correlation coefficients between ROS activity and
the water insoluble concentration of metals were determined, as presented in Table 3.2. As
expected, much lower (and often negative) correlations are observed between the water insoluble
fraction of most metals and trace elements and ROS activity compared to their water-soluble
fraction, which again accentuates the importance of water solubility in the formation of reactive
oxygen species.

48

Table 3.1. Pearson correlation coefficients (R) between monthly ROS activity data (µg
Zymosan/m
3
air) and monthly concentrations of chemical species (carbonaceous material,
water soluble metals and inorganic ions). Underlined numbers indicate values with R>0.7
and p<0.05.

HUD GRD LDS CCL USC HMS FRE VBR GRA
EC 0.08 0.64 0.01 -0.06 0.69 0.32 0.67 -0.22 0.32
OC -0.30 0.41 -0.03 -0.48 0.91 0.45 0.13 0.34 0.79
WSOC -0.26 0.40 0.57 -0.20 0.32 0.55 0.44 0.68 0.72
Mg 0.01 0.68 0.19 -0.40 0.56 0.34 0.39 0.10 0.71
Al 0.80 0.47 0.67 0.17 0.29 0.63 0.82 -0.41 0.56
K 0.21 0.78 0.31 -0.21 0.75 0.66 0.48 0.22 0.73
Ti 0.64 0.75 0.23 0.50 0.58 0.59 0.81 0.56 0.34
V 0.60 0.78 0.70 0.72 0.34 0.74 0.84 0.65 0.65
Cr 0.13 0.69 0.66 0.04 0.68 0.74 0.77 0.52 0.71
Mn -0.11 0.63 0.55 -0.33 0.71 0.73 0.45 0.00 0.58
Fe 0.87 0.89 0.67 0.63 0.65 0.91 0.93 -0.11 0.80
Co 0.36 0.60 -0.14 -0.28 0.51 0.07 0.61 0.03 0.64
Ni 0.16 0.69 0.67 0.46 0.43 0.69 0.71 0.57 0.77
Cu 0.73 0.44 0.30 -0.27 -0.35 0.23 0.24 -0.36 -0.03
Zn 0.59 0.34 0.59 -0.02 0.80 0.59 0.64 0.47 0.68
As -0.23 0.54 0.25 -0.11 0.14 0.41 0.67 -0.15 0.47
Mo -0.08 0.70 0.09 -0.35 0.52 0.67 0.43 0.05 0.42
Cd 0.14 0.69 0.31 -0.16 0.86 0.44 0.47 0.14 0.05
Ba -0.16 0.39 -0.05 -0.39 0.54 0.25 0.20 -0.16 0.46
La 0.86 0.71 0.90 0.52 0.23 0.56 0.86 0.63 0.47
Pb 0.69 0.72 0.40 -0.16 0.52 0.21 0.63 0.33 0.06
NO
3
-
-0.32 0.59 0.13 -0.22 0.46 -0.32 0.64 -0.35 -0.02
SO
4
2-
0.55 0.24 0.44 0.41 0.39 0.51 0.64 0.54 0.60
NH
4
+
0.42 0.18 0.09 0.40 0.11 0.43 0.56 0.55 0.41

49

Table 3.2. Pearson correlation coefficients (R) between monthly ROS activity data (µg
Zymosan/m
3
air) and monthly concentrations of the water insoluble fraction of metals.
Underlined numbers indicate values with R>0.6 and p<0.05.

HUD GRD LDS CCL USC HMS FRE VBR GRA
Mg -0.39 0.22 -0.21 0.04 0.62 0.51 -0.46 -0.30 -0.54
Al -0.30 0.68 0.35 -0.23 0.37 0.25 0.22 -0.21 0.03
K -0.34 0.57 -0.49 -0.39 0.18 0.16 -0.73 -0.54 -0.72
Ti -0.34 0.45 -0.23 -0.38 0.27 0.13 0.19 -0.28 -0.10
V -0.32 0.54 0.10 -0.21 0.50 0.07 -0.36 0.34 -0.45
Cr -0.16 -0.06 -0.25 -0.65 0.35 -0.24 -0.27 -0.19 -0.26
Mn -0.29 0.35 -0.31 -0.26 0.41 -0.26 -0.20 -0.40 -0.47
Fe -0.29 0.37 -0.31 -0.46 0.46 0.03 -0.03 -0.27 -0.10
Co 0.19 0.35 -0.05 -0.06 0.39 0.05 -0.14 -0.32 -0.33
Ni
0.09 0.46 -0.20 -0.04 0.39 0.34 -0.24 -0.05 -0.36
Cu
-0.13 0.14 -0.60 -0.31 0.14 -0.28 -0.36 -0.69 -0.52
Zn
-0.41 0.42 -0.11 0.93 0.22 -0.34 -0.44 0.02 0.62
As
-0.07 0.36 0.02 -0.27 0.14 -0.26 -0.52 -0.21 -0.51
Mo
-0.23 0.28 -0.38 -0.47 0.46 -0.29 -0.45 -0.58 -0.42
Cd
-0.26 0.51 -0.43 0.20 0.58 0.21 -0.05 -0.51 0.02
Ba
-0.28 0.35 -0.43 -0.27 0.28 0.08 -0.23 -0.37 -0.55
La
0.59 0.76 0.59 0.16 0.77 0.44 0.70 0.36 0.76
Pb
-0.28 0.58 -0.16 -0.58 0.42 -0.24 -0.02 -0.24 -0.25

50

Multivariate linear regression was employed in order to provide insight on source tracers that
have most significant contribution to the ROS activity. Three site clusters were considered,
namely Long Beach (including HUD), Los Angeles (including GRD, LDS, CCL, USC, HMS
and FRE) and Riverside (including VBR and GRA). The analysis was performed using SPSS
statistical software (SPSS Inc., version 16.0) and the water-soluble species which showed
individual correlation (R>0.7, p<0.05) with ROS activity using the univariate regression were
included in the analysis. The model yielded the following equations for the prediction of ROS
activity using the water-soluble concentration data:
1) Long Beach: ROS = -3.85 + 10.4 × Cu + 7056 × La
2) Los Angeles: ROS = 11.9 + 3.93 × Fe + 15.3 × V
3) Riverside: ROS = -29.0 + 36.9 × WSOC + 24.0 × Ni
The ROS activity in the above equations is expressed in µg Zymosan/m
3
, water-soluble metal
concentrations (i.e. Cu, La, Fe, V and Ni) are in ng/m
3
and WSOC is in µg/m
3
. The statistical
output of the model is summarized in Table 3.3. As shown in the table, all variables are
statistically significant (P<0.05) and there is no co-linearity between the selected variables
(VIF<4). The predicted ROS activity based on this model is also plotted as a function of
measured ROS activity in Figure 3.3 (a-c) for all three site clusters. For Long Beach, Cu and La
explained most of the variance in the ROS activity (R
2
=0.82, slope=0.82, intercept=8.43). The
appearance of La reflects the influence of residual oil combustion emissions on the ROS activity
at Long Beach, along with vehicular emissions represented by Cu. At the Los Angeles site
cluster, Fe and V, tracers of vehicular emissions and residual oil combustion are the variables
that best predict the ROS activity. The coefficient of determination (R
2
) for the model is 0.63
with a slope of 0.63 and an intercept of 20.4 (Figure 3.3-b). At Riverside, WSOC and Ni
51

concentrations drive the model, indicating the influence of secondary organic aerosols (SOA)
and industrial activities (metal plating in particular) on the ROS activity at the receptor location.
The correlation between measured and predicted ROS activity for this case is 0.68, with slope of
0.68 and intercept of 13.1 (Figure 3.3-c).
Table 3.3. Statistical output of multivariate regression model.
Site Cluster Species
Unstandardized
Coefficients
Units
Standardized
Coefficients
P-Values
Variance
Inflation
Factor
(VIF)
Overall
R
2
Long Beach
Constant -3.853
µg
Zymosan/m
3

0.82
Cu 10.498
µg
Zymosan/ng
metal
.674 .016 1.01
La 7056.197
µg
Zymosan/ng
metal
.556 .032 1.01
Los Angeles
Constant 11.950
µg
Zymosan/m
3

0.63
Fe 3.936
µg
Zymosan/ng
metal
.309 .002 1.51
V 15.323
µg
Zymosan/ng
metal
.574 .000 1.51
Riverside
Constant -29.0
µg
Zymosan/m
3

0.68
WSOC 36.9
µg
Zymosan/µg
WSOC
0.58 0.000 1.03
Ni 24.0
µg
Zymosan/ng
metal
0.48 0.001 1.03

52

Figure 3.3 (a-c). Linear regression between predicted ROS activity and measured ROS activity at (a) Long Beach, (b) Los Angeles and (c)
Riverside.

y = 0.82x + 8.43
R² = 0.82
0
20
40
60
80
100
120
0 20 40 60 80 100 120 140
Predicted ROS activity (µg Zymosan/m3)

Measured ROS activity (µg Zymosan/m3)

(a)
y = 0.63x + 20.39
R² = 0.63
0
20
40
60
80
100
120
140
0 50 100 150 200
Predicted ROS activity (µg Zymosan/m3)

Measured ROS activity (µg Zymosan/m3)
(b)
y = 0.68x + 13.12
R² = 0.68
0
20
40
60
80
100
120
140
0 20 40 60 80 100 120 140 160 180
Predicted ROS activity (µg Zymosan/m3
Measured ROS activity (µg Zymosan/m3
(c)
53

3.4. Summary and Conclusions
The results presented in this chapter demonstrated that PM
0.25
- bound water soluble transition
metals, along with organic carbon (OC) and water soluble organic carbon (WSOC) have the
highest association with ROS activity. Our results underscore the significance of water soluble
species of PM
0.25
in the formation of ROS, since water soluble metals were considerably more
redox active than their insoluble fractions. At Long Beach, multivariate regression analysis
indicated dominant contributions from residual oil combustion and vehicular emissions to ROS
activity. Tracers of these sources and road dust re-suspension, potentially originating from the
numerous freeways and surface streets in the region, were also observed to contribute to ROS
activity at urban areas of Los Angeles. At rural Riverside, tracers of metal plating activities as
well as secondary organic aerosol (SOA) are major drivers of ROS activity. The results of this
study provide important information on seasonal and spatial variations in ROS activity and
potential sources of chemical species contributing to ROS formation. Furthermore, combination
of our findings with exposure assessment models can provide additional insight on personal
exposure to toxic quasi-ultrafine airborne particles, and therefore assist the regulation of PM
emissions in the Los Angeles Metropolitan area.
3.5. Acknowledgments
This study was funded by the South Coast Air Quality Management District (SCAQMD) (award
#11527). We also would like to thank the staff at the Wisconsin State Laboratory of Hygiene for
their assistance with the chemical analyses. We also wish to acknowledge the support of USC
Provost’s and Viterbi’s Ph.D. fellowships.
54

Chapter 4:
Global Perspective on the PM-induced ROS Activity
Several air pollution studies are employing chemical (abiotic) assays to assess oxidative stress of
PM, which, in theory, can be used as a surrogate for its toxicity assessment. These chemical
assays, however, lack the physiological relevance to real-world PM exposure, as they do not
incorporate the biological responses of exposure to PM. To advance our understanding of
sources and chemical elements contributing to aerosol toxicity and provide quantitative global
comparative data, we report here on the biological oxidative potential associated with size-
segregated airborne PM in different urban areas of the world, measured by a biological cell-
based (macrophage) reactive oxygen species (ROS) assay. Our meta-analysis indicates a
generally greater intrinsic PM toxicity as well as higher levels of exposure to toxic PM in
developing areas of the world. Water soluble forms of transition metals (e.g. Fe, Ni, Cu and V)
and water soluble organic carbon (WSOC) are the chemical species showing highest correlations
with the oxidative potential of PM water extracts across diverse urban settings, indicating that
residual oil combustion, industrial activities, traffic emissions as well as secondary organic
aerosol formation are likely important sources contributing to PM-induced toxicity.

This chapter is based on the following publication:
Saffari, A.; Daher, N.; Shafer, M. M.; Schauer, J. J.; Sioutas, C. Global Perspective on the
Oxidative Potential of Airborne Particulate Matter: A Synthesis of Research Findings. Environ.
Sci. Technol. 2014, 48, 7576–7583.

4.1. Introduction
In recent years, there has been increasing attention paid to air pollution associated with airborne
particulate matter (PM) due to the complicated composition and source profile of PM compared
to other air pollutants as well as their capability to carry significant quantities of toxic chemicals
(Pope et al., 2002; Sioutas et al., 2005). Numerous studies have linked PM exposure to a wide
range of adverse health endpoints, including, but not limited to, cardiovascular diseases (Delfino
55

et al., 2005), respiratory problems (Penttinen et al., 2001) and adverse neuro-developmental
effects (Morgan et al., 2011). Many of the toxic effects of PM are thought to be mediated by
inflammatory responses, originating from PM-induced oxidative activity leading to the
generation of reactive oxygen species (ROS) upon the interaction of PM with epithelial cells and
macrophages (Ayres et al., 2008; Li et al., 2003; Tao et al., 2003). Several studies have
attempted to develop assays to quantify PM oxidative potential, using mostly abiotic, non-
cellular methods (Miljevic et al., 2010). Although these methods can, in theory, be used to
measure PM oxidative stress, they have limited physiological relevance to real-world PM
exposure, since they do not directly simulate the actual PM-cell interaction and its biological
inflammatory responses. In a recent cohort epidemiological study conducted in the Los Angeles
area, it was shown that among several abiotic and cellular assays most commonly used to
measure particle oxidative potential, only the cell-based macrophage assay used in our present
analysis demonstrated a robust association with systematic inflammation biomarkers that lead to
adverse health endpoints (Delfino et al., 2010). This assay employs a fluorescent probe (2´,7´-
dichlorodihydrofluorescein diacetate, DCFH-DA) to quantify PM oxidative potential, induced by
the generation of ROS upon interaction of redox-active particle components with macrophage rat
alveolar cells (Landreman et al., 2008). In this chapter we integrate and analyze the results of
several studies that we conducted over the last decade at numerous different locations across the
world, in order to provide insight on PM-induced oxidative potential and its relation to chemical
composition and sources of airborne PM, as well as to give a global perspective on PM toxicity
and its implications on public health.

56

4.2. Sampling and ROS Analysis Methodology
Table 4.1 summarizes relevant information about previous studies that are considered in this
meta-analysis. The analytical measurements conducted in each study are also listed in Table 4.1.
In brief, comprehensive elemental characterization of PM collected on Teflon filters was
performed using microwave-assisted mixed-acid digestion and high-resolution sector field
inductively coupled plasma mass spectrometry (ICP-MS) of the digestate, as described in Herner
et al., (2006). Water soluble organic carbon (WSOC) was measured by a Sievers 900 Total
Organic Carbon Analyzer, following water-extraction and filtration (0.22 µm) of the samples
(Sullivan et al., 2004). Organic carbon (OC) and elemental carbon (EC) concentrations were
quantified from PM collected on quartz filters by NIOSH thermal optical transmission method
(Schauer, 2003). Gas chromatography mass spectrometry (GC-MS) (as described in Stone et al.
(2008)) as well as Ion chromatography (IC) (Stone et al., 2009). were further employed in some
of the studies to speciate, in detail, organic compounds and inorganic ions, respectively. ROS
activity of the PM samples in all of the studies was quantified using an in-vitro exposure assay of
PM extracts to rat alveolar macrophage cells (cell line NR8383). Details of this macrophage
ROS assay are discussed in Landreman et al. (2008) A summary of the method is provided
below:
PM extraction/preparation: Teflon filter-collected PM samples were extracted with high-purity
Milli-Q (18 mΩ) water. The extraction was performed using 16 hours of continuous agitation at
room temperature in dark conditions, followed by centrifugation (6600 rpm for approximately 1
min) and then filtering through 0.22 µm polypropylene syringe filters. Just prior to the
macrophage exposure, 10× concentrated solutions of salts glucose medium (SGM) were added to
the PM extracts.
57

Preparation of rat alveolar cells: Rat alveolar cells (NR8383) were obtained from the American
Type Culture Collection (ATCC) and grown in Hams F12 medium (comprised of 2mM L-
glutamine, 1.176 g/L sodium bicarbonate and 15% heat inactivated fetal bovine serum). The cell
lines were maintained at 37 ºC and 5% CO
2.
atmosphere.
In-vitro exposure and ROS detection: Non-adherent macrophage cells were harvested from the
growth medium and centrifuged at 750 rpm for 5 min. Following the centrifugation, the cell
medium was removed and replaced with the SGM to obtain a cell suspension concentration of
1,000 cells/µL. To each well of a 96 well plate, 100 µL of this suspension (containing 100,000
cells) was added and incubated for 2 hours at 37 ºC and 6% CO
2
. 15 minutes before the end of
the incubation period, PM extracts were mixed with a small volume of 2´7´-
dichlorodihydrofluorescein diacetate (DCFH-DA) solution to achieve a final concentration of 45
µM DCFH-DA. After the incubation period, the SGM was removed from the plated cells in the
well plate and immediately replaced with 100 µL of the PM solution. The cell exposure was
carried-out for 2.5 hours under 37 ºC / 6% CO
2
conditions. During this exposure period, DCFH-
DA that passes the cell membrane is de-acetylated by cytoplasmic esterases to 2',7'-
dichlorodihydrofluorescein (DCFH). If oxidizing species are present (e.g. ROS species generated
intracellularly from exposure to potentially toxic PM components), the DCFH is converted to the
highly fluorescent 2,7, dichlorofluorescein (DCF). The fluorescence intensity of the samples
after the exposure was determined at 504 nm excitation and 529 nm emission using a m5e
microplate reader (Molecular Devices). The fluorescence was then compared to positive controls
(including Zymosan (a β-1,3 polysaccharide of D-glucose), and urban dust extracts) and negative
controls (method blanks) and the overall oxidative activity reported in units of Zymosan
equivalents.
58

Table 4.1. Summary of studies investigating the ROS activity of atmospheric PM using the cell-based macrophage assay.
Particle
size
Study Location Environment Sampling Period
Sampling
Duration
Chemical
Measurements
PM
0.25

(Saffari et al., 2013b)

Los Angeles, USA Urban Apr 2008-Mar 2009 24-hour
OC, EC, WSOC, Metals,
Ions, Organics*
(Saffari et al., 2013b)

Long Beach, USA Near-Port/Industrial Apr 2008-Mar 2009 24-hour
OC, EC, WSOC, Metals,
Ions, Organics
(Daher et al., 2014)

Beirut, Lebanon Urban Jul 2012 -Aug 2012 7 am-5 pm
OC, EC, WSOC, Metals,
Ions, Organics
(Saffari et al., 2013b)

Riverside, USA Rural/Semi-rural Apr 2008-Mar 2009 24-hour
OC, EC, WSOC, Metals,
Ions, Organics
PM
2.5

(Verma et al., 2009b)

Los Angeles, USA
(During Wildfire)
Urban Oct 2007 24 hour WSOC, Metals, Organics
(Daher et al., 2014) Beirut, Lebanon Urban Jul 2012 -Aug 2012 7 am-5 pm
OC, EC, WSOC, Metals,
Ions, Organics
(Daher et al., 2012b) Milan, Italy Urban Dec 2009-Nov 2010 24-hr
OC, EC, WSOC, Metals,
Ions, Organics
(Saffari et al., 2013a)
Thessaloniki,
Greece
Urban Jan 2013-Feb 2013 24-hr
OC, EC, Metals, Ions,
Organics
(Zhang et al., 2008) Denver, USA Urban Jan 2003-Dec 2003 24 hour OC, EC, WSOC, Metals
(Shafer et al., 2010) Lahore, Pakistan Urban Jan 2007-Jan 2008 24-hour
OC, EC, WSOC, Metals,
Ions, Organics
PM
10-2.5

(Cheung et al., 2012) Los Angeles, USA Urban Apr2008-Mar 2009 24-hour
OC, EC, WSOC, Metals,
Ions, Organics
(Shafer et al., 2010) Lahore, Pakistan Urban Jan 2007-Jan 2008 24-hour
OC, EC, WSOC, Metals,
Ions, Organics
(Daher et al., 2014) Beirut, Lebanon Urban Jul 2012 -Aug 2012 7 am-5 pm
OC, EC, WSOC, Metals,
Ions, Organics
(Cheung et al., 2012) Riverside, USA Rural/Semi-rural Apr2008-Mar 2009 24-hour
OC, EC, WSOC, Metals,
Ions, Organics
* OC: organic carbon, EC: elemental carbon, WSOC: water soluble organic carbon, Ions: SO
4
2-
, NH
4
+
,NO
3
-
, PO
4
3-
, Cl
-
, K
+
, Na
59

Figure 4.1. ROS activity normalized by PM mass (µg Zymosan/mg PM) and volume of air (µg
Zymosan/m
3
air) at different locations and for different size ranges. Box plots correspond to minimum,
1
st
quartile, median, 3
rd
quartile and maximum.

60

Table 4.2. Student’s two-tailed t-test (p-value) results corresponding to the comparisons between (a) Mass-normalized ROS activity levels in
figure 4.1 and (b) volume-normalized ROS activity levels in figure 4.1.

Beirut (PM0.25)

Los Angeles (PM0.25) <0.001

Long Beach (PM0.25) <0.001 0.158

4.2.
(a)

Riverside (PM0.25) <0.001 0.123 0.392

Beirut (PM2.5) 0.178 <0.001 <0.001 <0.001

Milan (PM2.5) 0.777 0.003 0.02 0.021 0.045

Lahore (PM2.5) 0.029 0.518 0.438 0.421 <0.001 0.018

Denver (PM2.5) <0.001 0.002 0.001 0.002 <0.001 <0.001 0.024

Los Angeles (PM2.5) <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.023

Thessaloniki (PM2.5) <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.009 0.465

Lahore (PM10-2.5) 0.023 0.446 0.367 0.237 <0.001 0.035 0.446 0.002 0.002 0.001

Beirut (PM10-2.5) <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.057 0.673 0.673 <0.001

Los Angeles (PM10-
2.5)
<0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.021 0.414 0.414 <0.001 0.813

Riverside(PM10-2.5) <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.011 0.563 0.563 <0.001 0.791 0.732

Beirut (PM0.25)
Los Angeles (PM0.25)
Long Beach (PM0.25)
Riverside (PM0.25)
Beirut (PM2.5)
Milan (PM2.5)
Lahore (PM2.5)
Denver (PM2.5)
Los Angeles (PM2.5)
Thessaloniki (PM2.5)
Lahore (PM10-2.5)
Beirut (PM10-2.5)
Los Angeles (PM10-2.5)
Riverside(PM10-2.5)

61

Beirut (PM0.25)

Los Angeles (PM0.25) <0.001

Long Beach (PM0.25) <0.001 0.435

4.2.
(b)

Riverside (PM0.25) <0.001 0.322 0.172

Beirut (PM2.5) 0.468 <0.001 <0.001 <0.001

Milan (PM2.5) 0.066 <0.001 <0.001 <0.001 0.273

Lahore (PM2.5) <0.001 <0.001 <0.001 <0.001 <0.001 <0.001

Denver (PM2.5) <0.001 0.003 <0.001 <0.001 <0.001 <0.001 <0.001

Los Angeles (PM2.5) <0.001 0.083 0.113 0.189 <0.001 <0.001 <0.001 0.258

Thessaloniki (PM2.5) <0.001 0.009 0.008 0.011 <0.001 <0.001 <0.001 0.677 0.871

Lahore (PM10-2.5) 0.003 <0.001 <0.001 <0.001 <0.001 0.036 0.182 <0.001 <0.001 <0.001

Beirut (PM10-2.5) <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.294 0.145 0.49 <0.001

Los Angeles (PM10-
2.5)
<0.001 <0.001 0.006 0.01 <0.001 <0.001 <0.001 0.455 0.231 0.093 <0.001 0.587

Riverside(PM10-2.5) <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.287 0.189 0.24 <0.001 0.664 0.784

Beirut (PM0.25)
Los Angeles (PM0.25)
Long Beach (PM0.25)
Riverside (PM0.25)
Beirut (PM2.5)
Milan (PM2.5)
Lahore (PM2.5)
Denver (PM2.5)
Los Angeles (PM2.5)
Thessaloniki (PM2.5)
Lahore (PM10-2.5)
Beirut (PM10-2.5)
Los Angeles (PM10-2.5)
Riverside(PM10-2.5)

62

4.3. Results and discussion
4.3.1. Comparison of oxidative potential in different urban areas
Figure 4.1 displays the ROS activity of size-segregated PM in different locations, normalized by
PM mass (in units of Zymosan per mg PM, lower plot) as well as by the volume of air (in units
of Zymosan per m
3
of air, upper plot). ROS activity normalized by PM mass represents the
intrinsic PM toxicity and is driven by the mix of ROS-active PM constituents. ROS activity
normalized by the volume of air sampled provides, in-contrast, a metric of the actual exposure to
toxic PM upon inhalation and is driven, not only by the intrinsic PM toxicity, but also by the
total PM concentration. The latter may be influenced by several other factors, including but not
limited to, local wind and temperature conditions, inversion layers and emission strength of
pollution sources. As seen in Figure 4.1, the mass-normalized ROS activity varies substantially
across the different locations, suggesting that the contrasting mix of sources and resulting PM
compositions in these specific urban atmospheres can potentially result in significant variation in
toxicity. The location-dependence (or source-dependence) of PM toxicity is mostly obvious for
the PM2.5 size range. For instance, the median intrinsic PM2.5 ROS activity is as low as 483 and
572 µg Zymosan/mg PM for Los Angeles and Thessaloniki (Greece), respectively, while the
values observed for Milan and Beirut are 25-50 times higher. Median PM10-2.5 ROS activity in
Beirut is within the same range as that of Los Angeles and Riverside, while almost 10 times
lower than Lahore; potentially due to the remarkably high mass fraction of redox active species
(such as Mn) in Lahore’s coarse PM compared to Beirut and Los Angeles (Cheung et al., 2012;
Daher et al., 2014; Shafer et al., 2010). In smaller size ranges, however, Beirut has a significantly
higher intrinsic PM toxicity compared to other locations (including Lahore, one of the most
polluted urban areas globally) (Shafer et al., 2010). These differences are generally consistent
63

with the concentration and mix of redox-active species, including water-soluble transition
metals, measured in the PM at each location (Daher et al., 2014; Saffari et al., 2013b; Shafer et
al., 2010).
For a given site, the air volume-based ROS activity (presented in Figure 4.1, upper plot),
represents the actual exposure to toxic PM upon inhalation (as opposed to the mass-normalized
ROS activity, representing intrinsic PM toxicity) and is, therefore, directly proportional to PM
mass concentration. It is also noteworthy that the air volume-based ROS activity is of particular
importance in locations where extremely high PM levels are observed and/or during high
pollution episodes when PM levels are affected by local meteorological factors (e.g. formation of
inversion layers) and strength of local primary emission sources. PM2.5 in Lahore, for instance,
has a lower level of intrinsic toxicity compared to Beirut and Milan, as seen in Figure 4.1. The
volume-normalized toxicity of Lahore’s airborne PM2.5, however, is 4 and 3 times higher than
Beirut and Milan, respectively, due to the high PM2.5 levels in that city (Shafer et al., 2010),
leading to a higher overall population exposure to potentially toxic aerosols despite its lower
intrinsic toxicity. Statistical significance of the comparisons between the ROS activities shown in
Figure 4.1 is documented in Tables 4.2.a and 4.2.b for reference.
4.3.2. Association of oxidative potential with chemical species and sources
Table 4.3 summarizes PM components having the highest association (R>0.7 and p<0.05) with
ROS activity at each location and in each size range, as well as their corresponding modeled
sources. Water-soluble transition metals are believed to catalyze the reduction of hydrogen
peroxide to the redox active hydroxyl radical through the Haber-Weiss reaction pathway (Fantel,
1996). These species (most notable Fe, Cu and Ni) were frequently and consistently associated
64

with the macrophage ROS activity, as presented in Table 4.3. The association of transition
metals with ROS activity implies the importance of their emission source reconciliation. The
effect of fuel oil combustion as well as industrial activities on PM-induced redox activity is
evident, as indicated by the association of residual oil combustion tracers (e.g. V and Ni)
(Saffari et al., 2013a) with ROS activity at multiple locations and size ranges. Nickel, for
instance, is associated with PM0.25 and/or PM2.5 ROS activity in Beirut, Los Angeles,
Riverside, Milan and Lahore. V is also correlated with the ROS activity in Los Angeles for all of
the size ranges (PM0.25, PM2.5 and PM10-2.5), which likely reflects the influence of fuel
oil/residual oil combustion products from ship emissions and other port-related industrial
activities. Association of La with PM0.25 at Long Beach also indicates the dominant influence of
industrial activities (such as petrochemical/refinery emissions) in that area. Moreover, the
influence of traffic sources, including their contribution to road dust (e.g. tailpipe exhaust, re-
condensed lubricating motor oil and tire-pavement interaction emissions) (Saffari et al., 2013c),
is observed, particularly in the PM2.5 and PM0.25 size ranges, as indicated by the strong
association of vehicular tracers at multiple locations (e.g. Fe in Los Angeles, Thessaloniki and
Denver and Cu in Long Beach, Beirut and Milan) with ROS activity. In the coarse mode, tracers
of vehicular abrasion (e.g. Cu) as well as re-suspended soil and dust (e.g. Co and Mn), are
associated with the ROS activity in Los Angeles and Beirut (Cheung et al., 2012; Daher et al.,
2014).
The water-soluble fraction of organic aerosols is another potential component contributing to the
induction of ROS activity. At Riverside, which is a semi-rural location downwind of Long Beach
and Los Angeles, water soluble organic carbon (WSOC) was associated with ROS activity which
suggests the effect of secondary organic aerosol (SOA) formation during the aging/transport of
65

particles in the atmosphere on PM oxidative potential (Saffari et al., 2013b; Snyder et al., 2009).
This effect of SOA on PM oxidative potential was also evident at Denver and Milan where
WSOC was strongly correlated with PM2.5 ROS activity (Daher et al., 2012b; Zhang et al.,
2008). It is noteworthy that specific local sources may also have a dominant contribution to the
ROS activity at certain locations. PM2.5-associated redox activity in Thessaloniki, for instance,
is affected by wood and biomass combustion emissions, as evident by the association of
levoglucosan and water soluble K with ROS activity in Thessaloniki (Saffari et al., 2013a). This
is also consistent with the fact that these samples were collected in the winter period in that city
and during severe pollution episodes.
One of the important observations in this meta-analysis is the identification of the influence of
air quality regulations on reducing PM intrinsic toxicity (i.e. mass-normalized ROS activity) as
well as exposure to potentially toxic PM (i.e. volume-normalized ROS activity). Beirut and Los
Angeles can be taken as examples. These two cities are both complex urban and industrialized
areas with generally similar climate and geographical morphology. Los Angeles has some of the
strictest air quality regulations globally, while those in Beirut are considerably more permissive
(typical of other developing areas of the world as well). We can clearly observe that PM in
Beirut is remarkably more toxic compared to Los Angeles PM in all of the size ranges. Tracers
of fuel oil combustion (i.e. Ni and V), for instance, are 5-6 times more enriched in the PM0.25
size range in Beirut compared to Los Angeles, which reflects the influence of stricter regulations
on fuel oil burning in Los Angeles compared to Beirut (Daher et al., 2014; Saffari et al., 2013b).
The same observation holds for the mass fraction of metallic tracers of vehicular emissions (e.g.
Cu and Fe) in Beirut compared to Los Angeles. Water soluble Cu, for instance, is almost 2 times
66

more enriched in PM0.25 emissions in Beirut compared to Los Angeles, likely due to the stricter
regulations on mobile sources in the Los Angeles basin (Daher et al., 2014; Saffari et al., 2013b).
67

Table 4.3. Species and sources associated with the ROS activity (R>0.7 and p<0.05) at each location and PM size range.

Size
Range
Location
Traffic Emissions
Fuel Oil
Combustion
Biomass Burning
Secondary
Organics
Fe Cu Cr Pb Co Mn Zn As Pd Ce Ni La V S K Levoglucosan WSOC
PM
0.25

Los Angeles,
USA
*

*

Long Beach,
USA
* *

*

Beirut,
Lebanon

* *

Riverside,
USA

*

*
PM
2.5

Los Angeles,
USA
*

* *

* * *

Beirut,
Lebanon

*

*

*

Milan, Italy

* *

* *

*
Thessaloniki,
Greece
*

* *

* * *

* * *

Denver, USA
*

*
Lahore,
Pakistan
§

* *

* * * *

PM
10-2.5

Los Angeles
and
Riverside,
USA

*

*

*

Beirut,
Lebanon

* *

§ Correlations for Lahore correspond to pooled PM
2.5
and PM
10
data.

68

4.3.3. Effect of particle size on PM oxidative potential
The particle-size dependence of mass-normalized (intrinsic) ROS-activity is evident in Figure
4.1. With the exception of Lahore, all of the studied locations exhibit lowest intrinsic ROS
activities in the coarse mode (compared to fine PM and PM0.25). The Lahore exception is
potentially driven by the significantly higher abundance of a large group of metals commonly
found in re-suspended soils (e.g. Ce, Mn) as well as metals from industrial metallurgical
activities (Cd and Pb) in the coarse mode, compared to PM2.5 (Shafer et al., 2010). In Los
Angeles and Riverside, considerably higher (i.e. 5-12-fold) median intrinsic toxicity was
measured for the PM0.25 size range compared to larger fractions; this was driven primarily by
the higher water solubility as well as higher water soluble mass fraction, of redox active metals
(e.g. Fe, V and Ni, originating from residual oil combustion and/or vehicular emissions) (Saffari
et al., 2013c). This trend could also be, at least in part, attributed to the higher abundance of
water-soluble organic species (mainly originating from SOA formation as discussed in the
previous section) in the smaller size fractions. A positive correlation is observed between WSOC
and ROS activity in PM0.25 at Riverside while no association is observed between ROS activity
and WSOC in the coarse mode (Table 4.3). Moreover, in Beirut, the median ROS activity of
PM0.25 and PM2.5 are within the same range and almost 20 times higher than the ROS activity
associated with coarse PM. These observations altogether indicate that smaller particles are
generally associated with higher redox activity.
It should be noted that in reality, the potential toxicity of smaller particles relative to larger
particles may be even greater than what is measured by any assay, including our macrophage
ROS assay. Compared to larger particles, smaller particles (i.e. ultrafine fraction) have a much
69

higher deposition efficiency in the respiratory system (Chalupa et al., 2004; Sioutas et al., 2005),
as well as a higher penetration in the lower parts of the respiratory tract (i.e.
pulmonary/tracheobronchial regions) (Yeh et al., 1996). Moreover, based on clinical studies, the
abundance of natural anti-oxidants in lung lining fluid (e.g. glutathione and ascorbic acid)
decreases remarkably as we go from upper to lower (i.e. deeper) respiratory tract, where smaller
particles are prone to deposit (Kelly et al., 2003). Anti-oxidants in the lung lining fluid are
believed to protect the respiratory system against the oxidative stress imposed by inhalable
material and therefore, the overall toxicity impact of smaller particles to humans compared to
larger particles can be potentially higher than the one predicted by in-vitro toxicity assays.
Determining the degree to which in-vivo oxidative stress of particles is different from in-vitro
measurements remains unclear and requires further and much-needed clinical investigations.
4.3.4. Seasonal effects on PM oxidative potential
Differences in the chemical and physical properties of particles from secondary and primary
sources can potentially lead to differences in their oxidative potential. The effect of aging on PM
oxidative potential, measured by chemical (abiotic) assays (e.g. DTT assay), has been
investigated in previous studies (using both ambient measurement techniques as well as smog
chamber studies), and increased abiotic ROS activity for photo-chemically-aged particles
compared to fresh particles was reported (McWhinney et al., 2013; Rattanavaraha et al., 2011;
Saffari et al., 2014). Comparisons between ambient measurements and dynamometer studies
imply the possible effect of aging on the in-vitro ROS activity as well. Verma et al (2010)
evaluated the PM-induced macrophage (cell-based) ROS activity of fresh vehicular emissions in
Los Angeles, with and without particle emission control technologies, using dynamometer
facilities. The reported mass-normalized ROS activity for PM from the non-retrofitted vehicles
70

in that study was almost 10 and 2 times lower than that of ambient PM0.25 and PM2.5 in Los
Angeles, respectively. Photochemical activity and SOA formation during the transport of
particles from emission source to downwind receptor locations is likely the most important factor
contributing to the elevated in-vitro ROS levels in the ambient conditions compared to the fresh
engine exhaust. This hypothesis is further supported by comparing the intrinsic toxicity levels
during summer and winter in different locations. Figure 4.2. shows mass-normalized ROS
activity of PM0.25 in Los Angeles and Riverside and PM2.5 in Lahore and Milan during
summer (i.e. June, July and August) and winter (December, January and February). As seen in
the plot, ROS activity in Los Angeles and Riverside during summer averages 7961 (±1032) and
6971 (±285) μg Zymosan/mg PM respectively, which is 3.8 and 2.5 times higher than the winter
levels at the same locations. Summer-time toxicity levels in Lahore and Milan are also 1.8 and
1.5 times higher than winter-time levels, further suggesting the contribution of SOA components
to ROS-activity. The effect of photochemical aging is also evident by examining dominant
sources driving the PM0.25-associated ROS activity in the Los Angeles basin. As seen in Table
4.3, in Long Beach and Los Angeles (representing urban source locations in the basin impacted
by mostly freshly emitted aerosols), the macrophage ROS activity is largely driven by primary
emission sources (traffic emissions and residual oil combustion emissions in particular). In
contrast, in Riverside, which represents a receptor location in the Los Angeles basin, in which
the majority of the aerosol in summertime is advected from upwind source sites of central Los
Angeles after several hours since their emission (Kim et al., 2002), the effect of SOA formation
on ROS activity is clearly demonstrated. This result implies that as particles are transported in
the basin from source to receptor sites, photochemical formation of organic species contributes to
the overall PM redox activity. Controlled experimental studies (such as smog chamber studies)
71

would be helpful to further clarify the detailed toxico-chemical pathways through which the in-
vitro ROS activity of particles is affected by photochemical aging.

Figure 4. 2. Mass-normalized ROS activity associated with PM2.5 in Lahore and Milan and
PM0.25 in Riverside and Los Angeles during summer (i.e. June, July and August) and
winter (i.e. December, January and February). Error bars correspond to one standard
error.

72

4.4. Summary and Conclusions
In summary, the major pollution sources contributing to PM-induced oxidative potential (e.g.
residual/fuel oil combustion due to industrial activities and vehicular emissions) are among those
that can be effectively controlled by air quality regulations. The dominating influence of specific
chemical components (and their corresponding sources) on PM-induced ROS activity at different
locations is of particular note in this meta-analysis, which necessitates (but also improves
chances of successful intervention) the need for more targeted PM regulations to efficiently
mitigate the health effects of exposure to PM. As an example, while redox active transition
metals generally account for a small fraction of total PM mass (less than 10%), they are among
the major drivers of PM-induced oxidative potential (Table 4.3). Fuel oil combustion and
vehicular sources (abrasion as well as tailpipe emissions) are the two major contributors to ROS
activity as indicated by the association of their metallic tracers with ROS activity at multiple
locations. The water-soluble fraction of organic aerosols is also another major contributor to
ROS activity, with a more dominant effect at the locations with higher SOA formation. Accurate
identification and characterization of these sources at different urban settings is, therefore,
necessary for improving the air quality and reducing PM-associated toxicity in urban areas.
4.5. Acknowledgements
Financial support for this study was provided by South Coast Air Quality Management District
(SCAQMD) (award # 11527), United States Environmental Protection Agency (US-EPA) under
the Science to Achieve Results program (EPA-G2006-STAR-Q1) and Southern California
Particle Center through the US-EPA grant RD-8324-1301-0 to the University of Southern
California.
73

Chapter 5:
Impact of Atmospheric Aging on PM-induced ROS Activity
There is growing literature supporting the hypothesis that the most important pathway
underlying the adverse health effects of exposure to particulate matter (PM) is the oxidative
stress derived from the interaction of PM with cells. Differences in the chemical and physical
properties of particles from secondary and primary sources can potentially lead to differences in
their oxidative potential. The effect of aging on PM oxidative potential, measured by chemical
(abiotic) assays (e.g. DTT assay), has been investigated in previous chamber studies and
increased ROS activity for photochemically-aged particles compared to fresh particles was
reported. Other than smog chamber measurements, previous dynamometer studies imply the
possible effect of aging on the ROS activity as well. This chapter aims to investigate the effect of
atmospheric aging on the oxidative potential using field measurement techniques, in order to
provide much-needed insight regarding the toxicity of airborne particles in the real-world urban
atmosphere rather than laboratory conditions.

This chapter is based on the following publication:
Saffari, A.; Hasheminassab, S.; Wang, D.; Shafer, M. M.; Schauer, J. J.; Sioutas, C. Impact of
primary and secondary organic sources on the oxidative potential of quasi-ultrafine particles
(PM0.25) at three contrasting locations in the Los Angeles Basin. Atmos. Environ. 2015, 120,
286–296.

5.1. Introduction
Quasi-UFPs may originate from primary emissions (including, but not limited to vehicle exhaust,
fuel oil combustion, biomass burning and different industries) as well as from secondary aerosol
formation processes, such as photochemical aging (Ning et al., 2007). Differences in the physical
and chemical properties of particles from secondary and primary sources likely drive contrasts in
74

their intrinsic oxidative potential. There is extensive data documenting that primary sources (i.e.
particulate combustion by-products) significantly contribute to the toxicity and redox activity of
airborne particles, primarily due to the presence of transition metals as well as organic oxidant
agents, such as quinones, in these emissions (Cheung et al., 2009; Cormier et al., 2006;
Dellinger et al., 2000; Lewtas, 2007). However, emerging evidence suggests that photochemical
aging and secondary sources may as well significantly impact the overall toxicity and redox
activity of the ambient aerosols. For instance, multiple studies have shown that photochemical
aging of the diesel exhaust and soot particles in controlled laboratory chamber conditions
(designed to imitate atmospheric aging processes) enhance the PM oxidative potential measured
by the non-cellular dithiothreitol (DTT) assay (Antiñolo et al., 2015; Li et al., 2009; McWhinney
et al., 2013, 2011; Rattanavaraha et al., 2011). Similarly, in another chamber study, it was shown
that naphthalene-driven secondary organic aerosol (SOA) has enhanced DTT-based ROS activity
compared with the primary naphthalene aerosol (McWhinney et al., 2013). Comparison of the
ROS activity of fresh vehicular emissions in dynamometer facilities with the ROS activity of
ambient PM may also suggest that photochemical processes in the atmosphere can change the
oxidative properties of ambient particles (Geller et al., 2006; Saffari et al., 2014). Along the same
lines, recent studies by (Verma et al., 2015, 2014) suggested that SOA has major contributions to
the ROS activity of ambient PM2.5 in southeastern United States, although these studies
identified biomass burning as the most dominant source contributing to the ROS activity in that
region. Moreover, in chapter 4, we demonstrated higher summer-time cell-based ROS activities
compared to winter in multiple urban areas of the world (Saffari et al., 2014), again suggesting
the potential enhancement of PM-induced redox activity and toxicity through secondary aerosol
formation. Despite the likely importance of characterizing the relative impact of primary and
75

secondary sources on the oxidative potential of PM and major implications for effective air
pollution control policy, there is yet limited data available on how the formation of secondary
organic aerosols (SOA) affect the oxidative potential and PM toxicity in actual urban settings.
The primary objective of this study was to quantitatively investigate the relative impact of
primary and secondary sources on the oxidative potential of atmospheric particles in a real-world
urban environment. To this end, airborne quasi-UFPs were collected at 3 selected locations,
temporally sequentially, along the dominant air movement trajectory in the Los Angeles (LA)
south coast air basin, representing three contrasting sets of aerosols with a substantially different
mixture of primary and secondary origins. The oxidative potential of PM0.25 at these sites was
quantified using a cell-based macrophage ROS assay. Furthermore, comprehensive chemical
analyses were conducted on the collected samples, followed by a source apportionment analysis
using the molecular marker-based chemical mass balance (MM-CMB) model, in order to
investigate and quantify source-toxicity relationships and the relative impact of different sources
at contrasting levels of photochemical age and secondary organic aerosol formation.
5.2. Methods
5.2.1. Site Selection, Sampling Schedule and Meteorology
Site selection and sampling schedule criteria were designed in a way to distinguish, to the highest
practical extent possible, different levels of SOA formation in the LA basin. Samples were
collected in the summer season, with three sampling sites located at Long Beach, central Los
Angeles and Upland. Figure 5.1 shows the location of the sampling sites in the basin. The Long
Beach site (coordinates: 34.1036111, -117.6291667; elevation: 9 m) is located about 2 km north
of the Los Angeles/ Long Beach harbor, about 180 meters east (downwind) of the terminal island
76

freeway. Therefore, this site is highly impacted by fresh primary emissions from port activities,
in addition to a mix of mostly primary emissions from the light duty and heavy duty vehicle fleet
given its proximity to freeways, in particular the I-710 freeway which has a large (greater than
10%) diesel truck contribution (Kam et al., 2012). The Los Angeles site (coordinates:
33.8025000 -118.2200000; elevation: 80 m) is located 5 kilometers south of the downtown Los
Angeles, in a typical urban area about 150 meters east (downwind) of the I-110 freeway. The
Upland site (coordinates: 34.0191667 -118.2772222; elevation: 377 m) is located about 60
kilometers northeast of central Los Angeles in a residential area nearly 1500 meters north of the
I-10 freeway, representing urban background, impacted by advected particles from the upwind
central Los Angeles region during summer when the westerly/southwesterly wind is dominant, in
addition to local emission sources.
Samples were collected at each site weekly, during three consecutive weekdays, between June
and early October 2014. Each week, the sampling was started at a different weekday (Monday,
Tuesday or Wednesday), in order to include all of the 5 weekdays in our sampling schedule and
thus account for the possible differences in traffic density among different weekdays. To enhance
our resolution of different stages of photochemical SOA formation, each day of sampling was
segmented into three 4-hour periods based on the prevailing wind direction and distances
between the sites. To illustrate the viability of this strategy for site selection and sampling
protocol, wind direction/speed data at the three study locations are presented in Figure 5.2-a.
Wind direction was southerly at Long Beach during the sampling times (with a median speed of
2.8 m/s), and gradually shifted towards southwesterly and westerly directions further inland
towards the LA and Upland regions in mid-day and afternoon hours (with median speeds of 4.5
and 7.0 m/s, respectively). Also, Figure 5.3 shows the diurnal variation of temperature and
77

relative humidity (RH) at the three study locations. Hourly averages of temperature and relative
humidity ranged between 17-35 ˚C and 32-80%, respectively, with Upland having the highest
recorded temperatures and lowest RH among the three locations, followed by LA and Long
Beach. These trends are consistent with historical summer-time data in the basin (Hasheminassab
et al., 2014) and were used as one of the criteria for selecting the sampling sites.

Figure 5.1. Location of the three sampling sites in the Los Angeles basin.
78

Figure 5.2. Wind Speed/Direction (a) and Ozone concentration (ppm) (b), during the sampling periods in each site. Error bars represent one
standard error of hourly data.
(a)

(b)

79

Figure 5.3. Diurnal variation of temperature (a) and relative humidity (b) at the three sampling sites. Error bars represent standard error.

(a)
(b)
80

Based on the aforementioned meteorological conditions, for each day, samples were first
collected at Long Beach from 6:00 am to10:00 am, i.e., the typical rush hour period during which
the Long Beach region is mostly impacted by fresh primary emissions and also with the lowest
temperatures compared to midday and afternoon periods, which favor the partitioning of
semivolatile primary organic compounds into the particle phase (Verma et al., 2011). Sampling
was then continued between 10:00 am-2:00 pm at Los Angeles, during which the quasi-UFPs are
in an intermediate stage of photochemical transformation, with impacts from advected (and
diluted) primary emissions in the upwind regions (i.e. Long Beach harbor), in addition to early
stage of SOA formation. Lastly, quasi-UFPs were sampled at Upland from 2:00 pm-6:00 pm, the
time period that reflects the photochemically-aged SOA.
To further evaluate the extent to which the study locations were impacted by photochemical
activity, ozone concentration data are displayed in Figure 5.2-b. For Upland and Long Beach, the
ozone concentrations correspond to the exact same sampling locations. For the Los Angeles site,
however, the ozone concentrations were obtained from the nearest air quality monitoring site of
the South Coast Air Quality Management District (SCAQMD), located about 3 kilometers north
of our sampling location in central Los Angeles. Ozone is predominantly formed through
photochemical reactions involving nitrogen oxides (NOx) and volatile organic compounds
(VOCs), and can therefore be used as a general marker of the extent of photochemical
secondary aerosol formation (Docherty et al., 2008; Fine et al., 2008). Increasing ozone levels
during the sampling times while moving from Long Beach towards LA and Upland imply that
Long Beach during the sampling hours (6-10 am) is dominantly impacted by primary emissions
from mobile sources (vehicles/ships), while the impact and dominance of SOA formation
increases consistently as we move towards the downwind locations (LA and Upland) in the mid-
81

day and afternoon periods. As an additional analysis, Hybrid Single-Particle Lagrangian
Integrated Trajectory model (HYSPLIT 4, NOAA air resources laboratory) was also employed in
order to examine the long-range trajectories in our study area. While accurate identification of
trajectories between our sampling sites is hindered by the low resolution of meteorological data
available for the HYSPLIT model, the overall pattern of large-scale air movement in our study
area seems to be consistent with the local patterns (i.e. wind roses). Figure 5.4 highlights the
back-trajectories arriving at the three sampling sites on selected sampling days.
Figure 5.4. 24-hour back-trajectories arriving at the sampling sites at 6:00 pm, for an
elevation range of 0-1000 meters and GDAS 0.5-degree meteorological data.

08/12/2014
08/26/2014 09/02/2014
07/15/2014
LB
USC UPL
82

5.2.2. Sampling Method
Five parallel Sioutas Personal Cascade Impactor Samplers (Sioutas™ PCIS, SKC Inc., Eighty
Four, PA, USA) (Misra et al., 2002), each operating at 9 lpm, were deployed at each site to
collect particles in three size ranges: <0.25 μm (quasi-UFP), 0.25–2.5 μm, and 2.5–10 μm. In this
paper we focus on the results for the quasi-UFP fraction, given the importance of this fraction
from a toxicological perspective. At each site, the sampler was located inside of an air
conditioned trailer, and the inlet was placed about 3 meters above the trailer’s rooftop. Of the
five PCIS at each site, three were loaded with polytetrafluoroethylene (PTFE) membrane filters
(3 μm pore-size, Pall Life Sciences, Ann Arbor, Michigan, USA) and the other two PCIS were
loaded with quartz microfiber filters (Whatman International Ltd, Maidstone, England). A total
of 7 sets of samples were collected at each site, each representing a bi-weekly composite of the
collected quasi-ultrafine particles.
5.2.3. Chemical and Toxicological Analyses
Particle mass concentration was determined by pre- and post- weighing the Teflon and Zeflour
filters using a microbalance (Mettler Toledo Inc., Columbus, OH, USA), following equilibration
under controlled temperature (22–24 °C) and relative humidity (of 40–50%) conditions.
Elemental and organic carbon (EC and OC, respectively) content of the filters were quantified
using the NIOSH Thermal Optical Transmission method (Birch and Cary, 1996). Water soluble
organic carbon (WSOC) was quantified using a Sievers 900 Total Organic Carbon Analyzer
(Stone et al., 2008). The total and water-soluble fraction of elements were measured using a high
resolution inductively coupled plasma sector field mass spectrometry (ICP-SFMS, Themo-
Finnigan Element 2) (Herner et al., 2006). To conduct the total ICP-MS analysis, the particles
83

collected on the PTFE membrane filters were solubilized in an acid mixture (containing nitric
acid, hydrochloric acid and hydrofluoric acid), to allow a complete digestion of all water soluble
as well as water insoluble PM components (Lough et al., 2005). To quantify the water-soluble
fraction of elements, PM0.25 were extracted in high purity Milli-Q water (by agitation and
sonication in acid-leached polypropylene tubes) and the extracts were filtered using the 0.22 µm
syringe filters prior to the ICP-MS analysis. ICP-MS analysis has a number of advantages over
other methods (such as X-ray fluorescence (XRF)) for the analysis of a broad range of elements
in atmospheric particulate matter, although a generally good agreement has been reported
between these methods (Okuda et al., 2014). Organic speciation was conducted on the bi-weekly
quartz filter composites using gas chromatography mass spectrometry (GC-6980, quadrupole
MS-5973, Agilent Technologies) (Stone et al., 2009). In order to supply sufficient mass to the
chemical analyses, filters from two consecutive weeks were composited and the chemical
analyses were carried out on the bi-weekly composites for all analyses, except for EC/OC which
was done on individual quartz filters.
Oxidative potential of the PM samples was quantified using a macrophage-based reactive oxygen
species (ROS) assay, similar to the previous chapter. In brief, to conduct this fluorogenic cell-
based method, Teflon filter composites were extracted using 1.00 ml sterilized Milli-Q water, by
16 hours of agitation in room temperature in the dark followed by 30 minutes of sonication, with
an average extraction efficiency of about 92% based on gravimetric measurements before and
after the extraction. To perform the PM-macrophage exposure, rat alveolar cells (NR8383) were
grown in a solution containing L-glutamine, sodium bicarbonate and inactivated fetal bovine
serum. The cells were then maintained at 37 ˚C, and prior to the exposure, a concentrated
solution of salt glucose medium (SGM) was added to the extracted PM. The ROS activity was
84

then quantified by in-vitro exposure of the incubated cells and extracted PM. 2′,7′-
dichlorodihydrofluorescein diacetate (DCFH-DA; a membrane permeable compound responsive
to a broad spectrum of reactive oxygen species including hydroxyl radical, superoxide radical,
peroxide and peroxynitrite (Schoonen et al., 2006)) was used as the fluorescent probe. DCFH-
DA is de-acetylated in the cell cytoplasm by cell esterases, yielding 2′,7′-
dichlorodihydrofluorescein (DCFH). Non-fluorescent DCFH is converted by ROS species
produced within the cell cytoplasm into the highly fluorescent 2,7-dichlorofluorescein (DCF),
which was monitored using a microplate reader. Final ROS responses were reported in units of
Zymosan equivalents, since among typical ROS positive controls, Zymosan was found to have
the highest stability, sensitivity and signal/noise for the specific ROS assay employed in our
study. Further details on this method can be found in Landreman et al. (2008).
5.2.4. Chemical Mass Balance (CMB) Model
A molecular-marker based chemical mass balance (MM-CMB) model was applied to apportion
the sources of PM0.25 OC in the collected samples. The model was mathematically solved with
the CMB software (EPA CMB v8.2), developed by the US Environmental Protection Agency,
which applies an effective-variance least-squares algorithm to the linear combination of the
product of the source contribution and its concentration (Watson et al., 1984).
Based on the observed primary tracers, the following source profiles were included in the model:
diesel and gasoline motor vehicles (Kam et al., 2012; Liacos et al., 2012), biomass burning (Fine
et al., 2004; Sheesley et al., 2007), vegetative detritus (Rogge et al., 1993a), natural gas
combustion (Rogge et al., 1993b), and ship emissions (Minguillón et al., 2008). Mobile source
profiles are based on PM0.25 measurements, while the rest are associated with PM2.5, but were
85

assumed to be the same for PM0.25, similar to the approach applied by Minguillón et al., (2008)
and Hasheminassab et al. (2013). To conduct the CMB analysis, a set of molecular marker
species representative of different primary emission sources were selected and used as indicator
species in the model. These species include: EC, benzo(b)fluoranthene, benzo(k)fluoranthene,
benzo(e)pyrene, 17α(H)-22,29,30- trisnorhopane, sitostane, levoglucosan, heptacosane,
triacontane, hentriacontane, dotriacontane, tritriacontane, vanadium, and nickel. Following the
CMB modeling, univariate and multivariate regression analysis was performed using SPSS
statistical software (v. 16), with the ROS activity as the dependent variable and the CMB-derived
sources as the independent variables.
5.3. Results and Discussion
5.3.1. Chemical composition
Average PM0.25 mass concentrations, as well as the average concentrations of carbonaceous
species (OC, EC and WSOC) are shown in Figure 5.5 for the three study locations. Minimum
mass concentration was observed at Long Beach (8.55 µg/m
3
), followed by Los Angeles and
Upland which showed comparable levels (11.96 µg/m
3
and 13.04 µg/m
3
, respectively). Increased
mass concentration levels from Long Beach towards the further downwind locations (Los
Angeles and Upland) can be, to a major extent, due to the additional contribution from secondary
aerosol formation in the receptor regions; it is also possible that other local sources, such as re-
suspended sub-micron dust (especially at Upland), may be contributing to the PM mass, as will
be discussed in the subsequent sections.
OC levels exhibit a similar trend as total PM0.25 mass concentration, with a maximum level of
5.5 µg/m
3
at Upland, followed by Los Angeles (5.1 µg/m
3
) and Long Beach (3.9 µg/m
3
).
86

Increased OC levels at Upland and Los Angeles are most likely due to the SOA formation, in
addition to a possible and rather small contribution from local primary biogenic sources at the
Upland site, as the source apportionment results (following sections) attest. Similar to total OC,
water soluble organic carbon, a tracer of secondary organic formation (Fine et al., 2008), also
shows enhanced levels at the receptor regions (Upland and Los Angeles, with 2.4 and 2.8 µg/m
3

respectively) compared to Long Beach (1.1 µg/m
3
). These levels indicate that at Long Beach,
only 28% of the organic carbon is water soluble, while the level of OC water solubility is
enhanced to 53% and 43% at Los Angeles and Upland, implying that SOA formation is a
contributor to the increased OC levels at the receptor regions during the warm seasons. EC
comprises a small fraction of the quasi-ultrafine mass (less than 10%), with highest concentration
at Long Beach, which is consistent with the impact of the nearby port activity and freeway
emissions with large fractions of heavy-duty vehicles-HDVs.
Average total concentration of major and trace elements and their water-soluble fraction at the
three study locations are shown in Figures 5.6.-a and 5.6-b, respectively. Concentration levels
spanned a wide range of values, with their spatial variations primarily driven by the dominant
local and regional sources (similar to previous studies in the basin (Saffari et al., 2013a)). For
instance, elements such as Ti and Fe, which primarily originate from sub-micron particle dust re-
suspension (Gustafsson et al., 2008; Saffari et al., 2013b) exhibit highest concentrations in the
Upland region.. On the other hand, metals such as Zn and Mo, associated with vehicular and
roadway emissions, as well as Ni, V and Cr, that are dominantly emitted from industrial
activities (e.g. refineries and metal plating) (Saffari et al., 2013b) peak at the Long Beach site,
due to its proximity to the multiple anthropogenic sources as well as the port emissions.
87

Figure 5.5. Mass concentration (µg/m
3
) of quasi-ultrafine PM (PM0.25) as well as carbonaceous species (Organic Carbon
(OC), Water Soluble Organic Carbon (WSOC) and Elemental Carbon (EC)) at the three study locations. Error bars represent
one standard deviation of 7 bi-weekly composited samples. Pie charts represent the water soluble and water insoluble fractions
of total organic carbon at each site.

88

Figure 5.6. Total concentration (ng/m
3
) of elements and metals (a) as well as their water soluble fraction (b) at the three study
locations. Error bars represent one standard deviation of 7 bi-weekly composited samples.
(a)

(b)

89

Concentrations of organic species (clustered into their corresponding organic groups, namely
polycyclic aromatic hydrocarbons (PAHs), hopanes and steranes, n-alkenes and organic acids) at
the three study locations are presented in Figure 5.7 (a-d). PAHs comprise a list of toxic and
carcinogenic compounds (Lewtas, 2007; Perera et al., 2006), which are primarily associated with
the incomplete combustion of fossil fuels from gasoline and diesel vehicles (Manchester-Neesvig
et al., 2003; Polidori et al., 2008). Due to their semi-volatile nature (Verma et al., 2011), particle-
phase PAHs are generally more abundant during the cold seasons. Considering that this study
was conducted in the summer season, the low molecular weight PAHs were mostly partitioned in
the gas-phase and therefore were not detected. Six PAH species with moderate to high molecular
weights (and therefore less volatility) were detected in our PM0.25 samples as displayed in
Figure 5.7-a. Cumulative PAHs concentration was highest at Long Beach (0.32 ng/m
3
), followed
by Los Angeles (0.15 ng/m
3
) and Upland (0.09 ng/m
3
). Higher PAHs concentrations in Long
Beach can be due to the combined effects of stronger PAHs emission sources in the Long Beach
region, along with lower morning time temperatures (and therefore higher PAHs partitioning in
the particle phase) at this site compared to midday and afternoon hours at LA and Upland as
shown in Figure 5.3-a..
Hopanes and Steranes are groups of organic compounds that originate almost exclusively from
primary vehicular emissions in the Los Angeles Basin (more specifically, from lubricating oil of
gasoline and diesel vehicles) (Zheng et al., 2002). Cumulative concentration of hopanes and
steranes at the three study locations are shown in Figure 5.7-b. Similar to PAHs, hopanes and
steranes peak at Long Beach (total concentration of 0.31 ng/m
3
), and their abundance decreases
as we move further towards the receptor regions (0.19 and 0.15 ng/m
3
at LA and Upland,
respectively), indicating decreased emission contributions from vehicular sources in the
90

downwind areas. Cumulative concentration of major detected n-alkanes (namely C25-C33) are
presented in Figure 5.7-c. Similar to PAHs, hopanes and steranes, n-alkanes were also most
abundant in Long Beach (cumulative concentration of 8 ng/m3), followed by Los Angeles and
Upland (6 and 4 ng/m
3
). Unlike PAHs, hopanes, steranes and n-alkanes (each dominated by
primary sources) organic acids, shown in Figure 5.7-d, are dominantly formed through secondary
formation processes (oxidation of their gas-phase precursors followed by condensation of the
less volatile products into the particle phase) (Carlton et al., 2006). For this reason, the
cumulative concentration of the detected organic acids in this study was highest at Upland (44
ng/m
3
) where photochemical activity was highest, followed by Los Angeles and Long Beach (37
and 24 ng/m
3
, respectively). Among the organic acids measured in this study, phthalic acid and
octadecanoic acid were the two most abundant species, as shown in Figure 5.7-d.
91

Figure 5.7. Concentration of detected organic species at the three study locations: (a) Polycyclic Aromatic Hydrocarbons
(PAHs), (b) Hopanes and Steranes, (c) n-Alkanes and (d) Organic Acids.

0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Long Beach Los Angeles Upland
Concentration (ng/m
3
)
(a)
Coronene
Benzo(g,h,i)perylene
Indeno(1,2,3-cd)pyrene
Benzo (e) pyrene
Benzo(k)fluoranthene
Benzo(b)fluoranthene
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Long Beach Los Angeles Upland
Concentration (ng/m
3
)
(b)
ABB-20R-C29-Sitostane
ABB-20R-C27-Cholestane
22R-Homohopane
22S-Homohopane
17A(H)-21B(H)-Hopane
17A(H)-21B(H)-30-Norhopane
17A(H)-22,29,30-Trisnorhopane
0
5
10
15
Long Beach Los Angeles Upland
Concentration (ng/m
3
)
(c) Tritriacontane
Dotriacontane
Hentriacontane
Triacontane
Nonacosane
n-Octacosane
n-Heptacosane
n-Hexacosane
n-Pentacosane
0
10
20
30
40
50
Long Beach Los Angeles Upland
Concentration (ng/m
3
)
(d)
Sebacic Acid
Azelaic Acid
Suberic Acid
Pimelic Acid
Adipic Acid
Glutaric Acid
Methylphthalic Acid
Terephthalic Acid
Isophthalic Acid
Phthalic Acid
Tetracosanoic Acid
Eicosanoic Acid
Pinonic Acid
Nonadecanoic Acid
Octadecanoic Acid
92

5.3.2. Oxidative Potential Source Apportionment
Oxidative potential of the quasi-UFPs, quantified by a macrophage reactive oxygen species
(ROS) assay, is demonstrated in Figure 5.8 for the three study locations. In Figure 5.8-a, the
ROS activity (expressed as units of Zymosan equivalents) is normalized by the total quasi-UFP
mass and therefore it represents the intrinsic toxicity of the particles. Mass-normalized ROS
activity is highest at Long Beach (10277±1669 µg Zymosan/mg PM), likely due to the proximity
of this site to the multiple sources of potential toxic emissions (specifically two major freeways
with mixed gasoline and diesel vehicle fleet, with added impact of refinery/petrochemical
industries). The intrinsic ROS activity in Los Angeles remains within the same range as Long
Beach (8213±1892 µg Zymosan/mg PM, with no statistically significant difference compared to
Long Beach), but slightly decreases as we move inland towards the Upland region (6677±1615
µg Zymosan/mg PM, p=0.091).
In figure 5.8-b, ROS activities normalized by the volume of air is shown for the three sites.
While the mass-normalized expression (Figure 5.8-a) provides insight regarding the intrinsic
toxicities of the particles, the volume-normalized values indicate the severity of exposure to
quasi-UFPs and the redox-active components therein. Due to the increased quasi-UPF mass
concentrations in the receptor regions, the overall volume-normalized ROS levels show little
spatial variability across the three sites, with no statistically significant difference between the
averages observed at any of the sites. As shown in Figure 5.8-b, average volume-normalized
ROS levels at Long Beach, Los Angeles and Upland are 95±15, 86±19 and 75±12 µg
Zymosan/m
3
, respectively.
93

Univariate regression was performed at the three study locations, between the intrinsic ROS
activity and the mass fraction of OC, WSOC, water insoluble organic carbon (WIOC, estimated
as the difference of OC and WSOC), EC, 11 redox active and/or potentially toxic metals (namely
V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Cd and Pb), as well as the sum of the aforementioned metals
(presented in Table 5.1). OC exhibited moderate to strong association with the ROS activity at
the three sites (R-values of 0.40, 0.74 and 0.69 at Long Beach, LA and Upland, respectively).
WSOC and WIOC have been respectively linked to the secondary and primary portions of the
organic aerosols in previous studies (Snyder et al., 2009; Weber et al., 2007). Strong and
statistically significant correlations (R>0.8 and p<0.05) between WIOC and ROS at Long Beach
as well as WSOC and ROS at Upland are consistent with the hypothesis that increased
photochemical SOA formation can alter the sources that dominantly drive ROS activity along
different locations in the basin. EC also exhibits strong associations with the ROS activity in
Long Beach and Upland, implying that PM-induced ROS activity in the receptor region (e.g.
Upland) can still be impacted by the local primary emissions and/or advected emissions in the
upwind source regions. Metals are another toxicologically important fraction of quasi-UFPs, the
redox-activity of which has been reported in numerous past studies (Prahalad et al., 1999;
Valavanidis et al., 2005). Consistent with previous investigations, our data documents the
associations of several metals and ROS activity. Specifically, Cr, Cu and As in Long Beach, Fe
in Los Angeles and V, Cr and As in Upland were correlated (R>0.65) with the ROS activity as
shown in Table 5.1, indicating that the effect of various metal sources (most notably vehicular
emissions) on the ROS activity can be observed throughout the basin. Following expectations,
this impact is overall stronger in the vicinity of source regions (e.g. Long Beach and Los
Angeles).
94

Figure 5.8. Oxidative potential of particles at the three study locations: (a) Normalized by
the total PM mass (µg Zymosan/mg PM) and (b) Normalized by the air volume (µg
Zymosan/m
3
). Error bars represent standard deviation of 7 bi-weekly composited samples.

95

Table 5.1. Pearson correlation coefficient (R) between ROS activity and chemical species
including OC, WIOC, WSOC, EC, individual metals and of summed concentration of
metals at the three study locations. Underlined values correspond to R>0.65.

Species Long Beach Los Angeles Upland
OC 0.40 0.74** 0.69**
WIOC 0.81* 0.26 0.49
WSOC 0.56 0.59 0.84*
EC 0.73** 0.13 0.88*
V 0.27 0.44 0.82*
Cr 0.72** 0.22 0.85*
Mn 0.59 0.55 0.49
Fe 0.34 0.70** 0.59
Co 0.12 0.21 0.29
Ni 0.16 0.01 0.30
Cu 0.71** 0.26 0.50
Zn 0.50 0.41 0.39
As 0.86* 0.33 0.77**
Cd 0.58 0.05 0.10
Pb 0.10 0.01 0.63**
Sum of Metals
0.67** 0.48 0.50
*Indicates p < 0.05 ** Indicates p < 0.10

96

PM0.25 OC source apportionment results are shown in Figure 5.9. In this Figure, only the
sources estimates with statistically significant levels (i.e. values greater than 2 times the
uncertainty) are included. Source contribution of primary sources were all derived from the CMB
model and the secondary OC was also estimated as the difference of the water soluble organic
carbon and 71% of the biomass burning source estimate (i.e. the fraction of wood smoke that is
estimated to be soluble based on Sannigrahi et al. (2006). Accordingly, ―other primary OC‖ was
propagated as the difference between summed source contribution estimates of all primary and
secondary sources and the total OC concentration, for each individual sample. At Long Beach,
mobile primary emissions (from light duty and heavy duty vehicles-denoted as LDV and HDV,
respectively) were the dominant sources, contributing to as much as 72% of the total OC,
substantially greater than that estimated for Los Angeles and Upland where the contribution from
mobile sources was about 50% and 39%, respectively. Secondary OC, in contrast, comprised a
relatively smaller fraction of OC in Long Beach (about 25%) but was found at much higher
levels at the receptor regions with nearly 50% contribution at Los Angeles and about 40% at
Upland, indicating the increased SOA formation in Upland and Los Angeles during the mid-day
and afternoon periods compared to morning hours in Long Beach. A relatively small portion of
OC (about 18%) in Upland was remained unexplained. This portion (represented as ―other
primary OC‖ in the Figure) indicates the primary OC emissions from sources that were not
included in the CMB model. Some of these possible sources could include emissions from food
cooking (especially meat charbroiling emissions from restaurants as reported in many other
studies (Nolte et al., 1999; Schauer and Cass, 2000)), as well as primary biogenic emissions in
Upland, given the residential location of this site and its proximity to these sources. Ship
emission comprised a small portion (0.001-0.002 µg/m
3
) of quasi-UFP mass concentration in this
97

study. These values indicate an almost 10-fold decrease in concentrations from ship emissions
compared to source apportionment studies of PM0.25 conducted in 2007 and 2008 in Los
Angeles (Hasheminassab et al., 2013; Minguillón et al., 2008), likely due to the implementation
of new ship emission control action plans at Long Beach and San Pedro Bay ports of California
in 2010
Figure 5.9. Source contributions to the quasi-UF OC (µg/m
3
) derived from the CMB model.

To evaluate the relative impact of each source on the ROS activity, regression analysis was
performed between the ROS activity and the source contribution estimates. Table 5.2 presents
the Pearson correlation coefficients between ROS activity and sources. At Long Beach, Mobile
emissions (sum of HDV and LDV) exhibited the highest association with the ROS activity
(R=0.63). In contrast, at Upland secondary OC was the source with highest association with the
98

ROS activity (R=0.72), indicating that increased photochemical aging may enhance the
contribution of secondary organic sources to the PM-induced oxidative potential. Correlation
coefficients in Los Angeles imply a moderate impact from both secondary OC and mobile
emissions (R=0.48 and 0.65, respectively), followed by ship emissions (R=0.45) likely
originating from the upwind port region. ―Other primary OC‖ and dust did not show any positive
correlations with the ROS activity at any of the study locations. It is noteworthy that the
oxidative potential of PM from different sources and formation mechanisms is highly dependent
on site morphology and characteristics, as evident from past studies in other urban regions. For
instance, (Hellack et al., 2015) and (Janssen et al., 2014) reported traffic emissions as the main
source contributing to the ROS activity and OH radical formation potency of PM2.5 and PM10
in urban locations in Germany and Netherlands, respectively; while in a study conducted in
Tartu, Estonia, biomass burning and industrial combustion processes were identified as the
strongest ROS-active sources in the region (Orru et al., 2010).
To provide additional insight on the overall relative impact of the two major PM0.25 sources
throughout the basin, regression analysis was also attempted for all of the three sites combined,
with secondary OC and mobile emissions as the independent variables (Table 5.3). Univariate
regression analysis indicated that mobile emissions and secondary OC could each explain 48%
and 15% of the variability across the three sites, whereas multivariate regression analysis with
both sources included as independent variables would explain 58% of the variability (p<0.001).
Relative contribution of each source to the measured ROS activity is also shown in Figure 5.10.,
which further indicates that the impact of atmospheric aging on the measured ROS activity
(represented by the SOA source) increases as the particles are advected from coast to inland
regions of the LA basin.
99

Table 5.2. Pearson correlation coefficient (R) between ROS activity and CMB-derived
sources at the three study locations. Underlined values correspond to R > 0.65.
Source Long Beach Los Angeles Upland
Mobile Sources
(HDV+LDV)
0.63** 0.65** 0.23
Secondary OC 0.55 0.48 0.72**
Vegetative Detritus -0.91 -0.37 0.06
Biomass Burning 0.13 -0.11 -0.13
Ship Emissions 0.58 0.45 0.00
Natural Gas Combustion 0.39 -0.12 0.36
Other Primary OC _ _ -0.81
*Indicates p < 0.05 ** Indicates p < 0.10

100

Table 5.3. Linear regression analysis between Secondary OC and mobile sources (µg/mg
PM) as independent variables and ROS (µg Zymosan/mg PM) as dependent variable, for
all three sites combined.
Variable(s) Regression Equation R
2

Model Significance
(p-value)
Mobile (HDV+LDV) ROS = 3031 + 21.3 Mobile 0.48 <0.001
Secondary OC ROS = 4446 + 22.8 SOC 0.15 0.07
Mobile (HDV+LDV)

Secondary OC
ROS = -119 + 18.3 SOC + 20.2 Mobile 0.58 <0.001

Figure 5.10. Relative contribution of mobile sources and secondary OC to the ROS activity.

101

5.4. Summary and Conclusions
The evaluation of the relative impacts of primary and secondary sources on the PM-induced
oxidative potential, as discussed in this chapter, has major implications in developing air
pollution control strategies. Based on our source apportionment and regression analyses, primary
vehicular emissions from gasoline and diesel vehicles, along with SOA are the most abundant
sources of PM0.25, and have the highest contribution to and association with the PM0.25 ROS
activity throughout the Los Angeles basin. The relative importance of these two formation
mechanisms, however, appears to exhibit substantial spatial and temporal contrasts. It can be
inferred from our regression analysis that, overall, primary emissions are still the dominant
contributors to PM oxidative activity and associated health effects in a populated urban setting
like Los Angeles. Light and heavy duty vehicles are the dominant ROS-active sources in the
morning period and especially in the proximity of major primary emissions (i.e. close to the port
and freeways in the Los Angeles basin). However, as particles from vehicular emissions of both
light and heavy-duty vehicles are transported away from the source region during the day
towards the downwind regions of the basin, the ROS activity associated with photo-oxidation of
these primary particles become important to the ROS activity at the downwind location.
Therefore, despite the overall dominance of primary vehicular sources at locations close to
primary vehicle emission, the atmospheric aging of these emissions add to the impact on health
risks at downwind locations and should be accounted for in the design and implementation of
protective regulatory measures in the future.
5.5. Acknowledgments
This study was funded by the South Coast Air Quality Management District (SCAQMD) (award
#11527). We also wish to acknowledge the support of USC Viterbi’s PhD fellowship.
102

Chapter 6:
Concluding Remarks

6.1. Major Findings of the Current Work
Findings documented in this dissertation contribute to the current state of knowledge on the
following aspects:
Ultrafine PM oxidative potential: current national ambient air quality standards (NAAQS)
primarily focus on the mass concentration of PM2.5 and PM10 size fractions of the ambient
aerosols. Findings of this work not only reinforce the importance of considering the ultrafine size
fraction for future regulatory measures in urban areas, but also provide an in-depth analysis of
ultrafine particles toxicity and source characteristics. Accordingly, chapter 2 of this dissertation
was one of the first attempts on quantifying the sources of ultrafine PM metals and investigating
its temporal and spatial variability. Moreover, chapter 3 provides a uniquely comprehensive
documentation of ultrafine PM oxidative potential in a major urban environment, and is the first
study on ultrafine particles that successfully quantifies year-long variability of ultrafine PM
oxidative potential and its chemical components at 9 different locations covering a large
geographical region.
Associations of ultrafine PM oxidative potential and emission/formation sources: rigorous
quantification of the seasonal/spatial variability of ultrafine PM oxidative potential also enabled
investigations on the source/toxicity associations. While a handful of previous studies have
strived to document these relationships, findings of this dissertation provide data that are
substantially more comprehensive in terms of depth and breadth. Moreover, the synthesis study
in chapter 4 provides global quantitative comparative data on size resolved PM toxicity.
103

Impact of atmospheric aging on ultrafine PM oxidative potential: One of the major and long-
term challenges facing the environmental regulatory authorities is to distinguish the relative
importance of primary PM emissions versus secondary PM formation processes in terms of PM-
related health effects. While this topic has been investigated in a number of laboratory smog-
chamber studies and epidemiological investigations, the results shown in chapter 5 of this
dissertation is one of the first attempts for addressing these challenges in an actual urban
atmosphere using field collection and offline analysis techniques. As described in chapter 5, the
special sampling design in that study enabled us to capture the impact of aging on the chemical
composition of ambient ultrafine PM and to link those changes to the oxidative potential.
6.2. Broad Implications and Recommendations
Most important implications of the findings of this dissertation can be summarized as follows:
Advancing future air quality regulations: current PM regulations in the United States aim at
controlling the total PM mass concentration. Over the past few decades since the enactment of
the clean air act, the major trend of changes in PM regulations have been to decrease the
NAAQS mass concentration standards to lower levels. While these progressive actions have
been, overall, successful in improvement of air quality in major urban areas, they might not be
the most efficient and cost-effective method of regulation, given that PM mass is just a surrogate
of other more important PM characteristics, especially the chemical and toxicological properties.
Findings of this dissertation can potentially facilitate the design and implementation of more
targeted and cost-effective regulatory measures that would provide public health protection while
possibly lifting some of the unnecessary burden from the industrial sector.
Improving PM-related epidemiological studies: epidemiological studies typically require a large
dataset covering a variety of different populations and conditions. Comprehensive
104

documentation of the temporal and spatial variability of ultrafine particle composition and
oxidative potential discussed in this dissertation can be used as an input for longitudinal cohort
investigations that aim to study the PM health effects over long periods of time and wide
geographical regions. This data become particularly important for cohorts that are designed for
chronic health effects of air pollution exposure (e.g. study of lung cancer or the Alzheimer’s
disease), given that such studies are in dire need of long-term and geographically resolved PM
pollution and characterization data.
Clinical studies using source-specific PM samples: last but not least, the conclusions of this
dissertation provide a framework for design and implementation of health studies (e.g. in-vitro
and in-vivo exposure assays) using source-specific PM samples. While the statistical association
between PM sources and toxicity as described in this work provide valuable information on
possible health-related consequences of exposure to these pollutants, the actual biological
pathways through which these health effects are caused is fairly unknown. Moreover, all of the
source-specific PM oxidative potential data presented here are based on offline cell-based assays
that may not be fully representative of the real-world PM exposure in human body. Design and
execution of in-vivo studies using particles originating from specific emission/formation sources
(e.g. the traffic-dominated PM versus the photochemically-aged PM, similar to what was done in
chapters 5), will shed light on the biological mechanisms of PM toxicity and subsequent health
effects.

105

Bibliography
Amato, F., Pandolfi, M., Escrig, A., Querol, X., Alastuey, A., Pey, J., Perez, N., Hopke, P.K.,
2009. Quantifying road dust resuspension in urban environment by Multilinear Engine: A
comparison with PMF2. Atmos. Environ. 43, 2770–2780.
Antiñolo, M., Willis, M.D., Zhou, S., Abbatt, J.P.D., 2015. Connecting the oxidation of soot to
its redox cycling abilities. Nat. Commun. 6.
Arhami, M., Sillanpää, M., Hu, S., Olson, M.R., Schauer, J.J., Sioutas, C., 2009. Size-Segregated
Inorganic and Organic Components of PM in the Communities of the Los Angeles
Harbor. Aerosol Sci. Technol. 43, 145–160.
Ayres, J.G., Borm, P., Cassee, F.R., Castranova, V., Donaldson, K., Ghio, A., Harrison, R.M.,
Hider, R., Kelly, F., Kooter, I.M., Marano, F., Maynard, R.L., Mudway, I., Nel, A.,
Sioutas, C., Smith, S., Baeza-Squiban, A., Cho, A., Duggan, S., Froines, J., 2008.
Evaluating the Toxicity of Airborne Particulate Matter and Nanoparticles by Measuring
Oxidative Stress Potential—A Workshop Report and Consensus Statement.
Birch, M.E., Cary, R.A., 1996. Elemental carbon-based method for occupational monitoring of
particulate diesel exhaust: methodology and exposure issues. Analyst 121, 1183–1190.
Campbell, A., Oldham, M., Becaria, A., Bondy, S.C., Meacher, D., Sioutas, C., Misra, C.,
Mendez, L.B., Kleinman, M., 2005. Particulate Matter in Polluted Air May Increase
Biomarkers of Inflammation in Mouse Brain. NeuroToxicology 26, 133–140.
Campian, E.C., Cai, J., Benz, F.W., 2002. Acrylonitrile irreversibly inactivates glyceraldehyde-
3-phosphate dehydrogenase by alkylating the catalytically active cysteine 149. Chem.
Biol. Interact. 140, 279–291.
Carlton, A.G., Turpin, B.J., Lim, H.-J., Altieri, K.E., Seitzinger, S., 2006. Link between isoprene
and secondary organic aerosol (SOA): Pyruvic acid oxidation yields low volatility
organic acids in clouds. Geophys. Res. Lett. 33, L06822.
Carroll-Ann W. Goldsmith Amy Imrich, H.D.Y.N., 1998. Analysis of Air Pollution Particulate-
Mediated Oxidant Stress in Alveolar Macrophages. J. Toxicol. Environ. Health A 54,
529–545.
Chalupa, D.C., Morrow, P.E., Oberdörster, G., Utell, M.J., Frampton, M.W., 2004. Ultrafine
particle deposition in subjects with asthma. Environ. Health Perspect. 112, 879–882.
Cheung, K., Shafer, M.M., Schauer, J.J., Sioutas, C., 2012. Diurnal Trends in Oxidative Potential
of Coarse Particulate Matter in the Los Angeles Basin and Their Relation to Sources and
Chemical Composition. Environ. Sci. Technol. 46, 3779–3787.
106

Cheung, K.L., Polidori, A., Ntziachristos, L., Tzamkiozis, T., Samaras, Z., Cassee, F.R., Gerlofs,
M., Sioutas, C., 2009. Chemical Characteristics and Oxidative Potential of Particulate
Matter Emissions from Gasoline, Diesel, and Biodiesel Cars. Environ. Sci. Technol. 43,
6334–6340.
Cho, A.K., Sioutas, C., Miguel, A.H., Kumagai, Y., Schmitz, D.A., Singh, M., Eiguren-
Fernandez, A., Froines, J.R., 2005. Redox activity of airborne particulate matter at
different sites in the Los Angeles Basin. Environ. Res. 99, 40–47.
Claiborn, C.S., Larson, T., Sheppard, L., 2002. Testing the metals hypothesis in Spokane,
Washington. Environ. Health Perspect. 110, 547–552.
Cormier, S.A., Lomnicki, S., Backes, W., Dellinger, B., 2006. Origin and Health Impacts of
Emissions of Toxic By-Products and Fine Particles from Combustion and Thermal
Treatment of Hazardous Wastes and Materials. Environ. Health Perspect. 114, 810–817.
Coudray, C., Favier, A., 2000. Determination of salicylate hydroxylation products as an in vivo
oxidative stress marker. Free Radic. Biol. Med. 29, 1064–1070.
Daher, N., Ruprecht, A., Invernizzi, G., De Marco, C., Miller-Schulze, J., Heo, J.B., Shafer,
M.M., Shelton, B.R., Schauer, J.J., Sioutas, C., 2012a. Characterization, sources and
redox activity of fine and coarse particulate matter in Milan, Italy. Atmos. Environ. 49,
130–141.
Daher, N., Saliba, N.A., Shihadeh, A.L., Jaafar, M., Baalbaki, R., Shafer, M.M., Schauer, J.J.,
Sioutas, C., 2014. Oxidative potential and chemical speciation of size-resolved particulate
matter (PM) at near-freeway and urban background sites in the greater Beirut area. Sci.
Total Environ. 470–471, 417–426.
Dahl, A., Gharibi, A., Swietlicki, E., Gudmundsson, A., Bohgard, M., Ljungman, A., Blomqvist,
G., Gustafsson, M., 2006. Traffic-generated emissions of ultrafine particles from
pavement–tire interface. Atmos. Environ. 40, 1314–1323.
Delfino, R.J., Gillen, D.L., Tjoa, T., Staimer, N., Polidori, A., Arhami, M., Sioutas, C.,
Longhurst, J., 2011. Electrocardiographic ST-segment depression and exposure to traffic-
related aerosols in elderly subjects with coronary artery disease. Environ. Health
Perspect. 119, 196–202.
Delfino, R.J., Sioutas, C., Malik, S., 2005. Potential Role of Ultrafine Particles in Associations
between Airborne Particle Mass and Cardiovascular Health. Environ. Health Perspect.
113, 934–946.
Delfino, R.J., Staimer, N., Tjoa, T., Arhami, M., Polidori, A., Gillen, D.L., Kleinman, M.T.,
Schauer, J.J., Sioutas, C., 2010. Association of Biomarkers of Systemic Inflammation
with Organic Components and Source Tracers in Quasi-Ultrafine Particles. Environ.
Health Perspect. 118, 756–762.
107

Delfino, R.J., Staimer, N., Tjoa, T., Gillen, D.L., Polidori, A., Arhami, M., Kleinman, M.T.,
Vaziri, N.D., Longhurst, J., Sioutas, C., 2009. Air pollution exposures and circulating
biomarkers of effect in a susceptible population: Clues to potential causal component
mixtures and mechanisms. Environ. Health Perspect. 117, 1232–1238.
Delfino, R.J., Staimer, N., Tjoa, T., Polidori, A., Arhami, M., Gillen, D.L., Kleinman, M.T.,
Vaziri, N.D., Longhurst, J., Zaldivar, F., Sioutas, C., 2008. Circulating biomarkers of
inflammation, antioxidant activity, and platelet activation are associated with primary
combustion aerosols in subjects with coronary artery disease. Environ. Health Perspect.
116, 898–906.
Dellinger, B., Pryor, W.A., Cueto, B., Squadrito, G.L., Deutsch, W.A., 2000. The role of
combustion-generated radicals in the toxicity of PM2.5. Proc. Combust. Inst. 28, 2675–
2681.
Docherty, K.S., Stone, E.A., Ulbrich, I.M., DeCarlo, P.F., Snyder, D.C., Schauer, J.J., Peltier,
R.E., Weber, R.J., Murphy, S.M., Seinfeld, J.H., Grover, B.D., Eatough, D.J., Jimenez,
J.L., 2008. Apportionment of Primary and Secondary Organic Aerosols in Southern
California during the 2005 Study of Organic Aerosols in Riverside (SOAR-1). Environ.
Sci. Technol. 42, 7655–7662.
Donaldson, K., Brown, D., Clouter, A., Duffin, R., MacNee, W., Renwick, L., Tran, L., Stone,
V., 2002. The Pulmonary Toxicology of Ultrafine Particles. J. Aerosol Med. 15, 213–
220.
Fantel, A.G., 1996. Reactive oxygen species in developmental toxicity: Review and hypothesis.
Teratology 53, 196–217.
Fine, P.M., Cass, G.R., Simoneit, B.R.T., 2004. Chemical Characterization of Fine Particle
Emissions from the Fireplace Combustion of Wood Types Grown in the Midwestern and
Western United States. Environ. Eng. Sci. 21, 387–409.
Fine, P.M., Sioutas, C., Solomon, P.A., 2008. Secondary Particulate Matter in the United States:
Insights from the Particulate Matter Supersites Program and Related Studies. J. Air Waste
Manag. Assoc. 58, 234–253.
Geller, M.D., Ntziachristos, L., Mamakos, A., Samaras, Z., Schmitz, D.A., Froines, J.R., Sioutas,
C., 2006. Physicochemical and redox characteristics of particulate matter (PM) emitted
from gasoline and diesel passenger cars. Atmos. Environ. 40, 6988–7004.
Goldsmith, C.-A.W., Imrich, A., Danaee, H., Ning, Y., Kobzik, L., 1998. Analysis of air
pollution particulate-mediated oxidant stress in alveolar macrophages. J. Toxicol.
Environ. Health - Part A 54, 529–545.
Gustafsson, M., Blomqvist, G., Gudmundsson, A., Dahl, A., Swietlicki, E., Bohgard, M.,
Lindbom, J., Ljungman, A., 2008. Properties and toxicological effects of particles from
the interaction between tyres, road pavement and winter traction material. Sci. Total
Environ. 393, 226–240.
108

Hasheminassab, S., Daher, N., Saffari, A., Wang, D., Ostro, B.D., Sioutas, C., 2014. Spatial and
temporal variability of sources of ambient fine particulate matter (PM2.5) in California.
Atmospheric Chem. Phys. 14, 12085–12097.
Hasheminassab, S., Daher, N., Schauer, J.J., Sioutas, C., 2013. Source apportionment and
organic compound characterization of ambient ultrafine particulate matter (PM) in the
Los Angeles Basin. Atmos. Environ. 79, 529–539.
Hellack, B., Quass, U., Beuck, H., Wick, G., Kuttler, W., Schins, R.P.F., Kuhlbusch, T.A.J.,
2015. Elemental composition and radical formation potency of PM10 at an urban
background station in Germany in relation to origin of air masses. Atmos. Environ. 105,
1–6.
Herner, J.D., Green, P.G., Kleeman, M.J., 2006. Measuring the Trace Elemental Composition of
Size-Resolved Airborne Particles. Environ. Sci. Technol. 40, 1925–1933.
Hoek, G., Brunekreef, B., Goldbohm, S., Fischer, P., van den Brandt, P.A., 2002. Association
between mortality and indicators of traffic-related air pollution in the Netherlands: a
cohort study. The Lancet 360, 1203–1209.
Hu, S., Polidori, A., Arhami, M., Shafer, M.M., Schauer, J.J., Cho, A., Sioutas, C., 2008. Redox
activity and chemical speciation of size fractioned PM in the communities of the Los
Angeles-Long Beach harbor. Atmos Chem Phys 8, 6439–6451. doi:10.5194/acp-8-6439-
2008
Hughes, L.S., Cass, G.R., Gone, J., Ames, M., Olmez, I., 1998. Physical and Chemical
Characterization of Atmospheric Ultrafine Particles in the Los Angeles Area. Environ.
Sci. Technol. 32, 1153–1161.
Imrich, A., Ning, Y., Lawrence, J., Coull, B., Gitin, E., Knutson, M., Kobzik, L., 2007. Alveolar
macrophage cytokine response to air pollution particles: Oxidant mechanisms. Toxicol.
Appl. Pharmacol. 218, 256–264.
Janssen, N.A.H., Yang, A., Strak, M., Steenhof, M., Hellack, B., Gerlofs-Nijland, M.E.,
Kuhlbusch, T., Kelly, F., Harrison, R., Brunekreef, B., Hoek, G., Cassee, F., 2014.
Oxidative potential of particulate matter collected at sites with different source
characteristics. Sci. Total Environ. 472, 572–581.
Kam, W., Liacos, J.W., Schauer, J.J., Delfino, R.J., Sioutas, C., 2012. Size-segregated
composition of particulate matter (PM) in major roadways and surface streets. Atmos.
Environ. 55, 90–97.
Kelly, F.J., 2003. Oxidative stress: its role in air pollution and adverse health effects. Occup.
Environ. Med. 60, 612–616.
Kelly, F.J., Dunster, C., Mudway, I., 2003. Air pollution and the elderly: oxidant/antioxidant
issues worth consideration. Eur. Respir. J. 21, 70s–75s.
109

Kim, S., Shen, S., Sioutas, C., Zhu, Y., Hinds, W.C., 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, 297–307.
Kleinman, M.., Sioutas, C., Chang, M.., Boere, A.J.., Cassee, F.., 2003. Ambient fine and coarse
particle suppression of alveolar macrophage functions. Toxicol. Lett. 137, 151–158.
Kleinman, M.T., Araujo, J.A., Nel, A., Sioutas, C., Campbell, A., Cong, P.Q., Li, H., Bondy,
S.C., 2008. Inhaled ultrafine particulate matter affects CNS inflammatory processes and
may act via MAP kinase signaling pathways. Toxicol. Lett. 178, 127–130.
Kreyling, W.G., Semmler-Behnke, M., Möller, W., 2006. Ultrafine Particle–Lung Interactions:
Does Size Matter? J. Aerosol Med. 19, 74–83.
Landreman, A.P., Shafer, M.M., Hemming, J.C., Hannigan, M.P., Schauer, J.J., 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, 946–957.
Lewtas, J., 2007. Air pollution combustion emissions: Characterization of causative agents and
mechanisms associated with cancer, reproductive, and cardiovascular effects. Mutat. Res.
Mutat. Res., The Sources and Potential Hazards of Mutagens in Complex Environmental
Matrices - Part II 636, 95–133.
Li, N., Sioutas, C., Cho, A., Schmitz, D., Misra, C., Sempf, J., Wang, M., Oberley, T., Froines,
J., Nel, A., 2003. Ultrafine particulate pollutants induce oxidative stress and
mitochondrial damage. Environ. Health Perspect. 111, 455–460.
Li, N., Wang, M., Bramble, L.A., Schmitz, D.A., Schauer, J.J., Sioutas, C., Harkema, J.R., Nel,
A.E., 2009. The adjuvant effect of ambient particulate matter is closely reflected by the
particulate oxidant potential. Environ. Health Perspect. 117, 1116–1123.
Li, Q., Wyatt, A., Kamens, R.M., 2009. Oxidant generation and toxicity enhancement of aged-
diesel exhaust. Atmos. Environ., Air Pollution Related to Transport 43, 1037–1042.
Liacos, J.W., Kam, W., Delfino, R.J., Schauer, J.J., Sioutas, C., 2012. Characterization of
organic, metal and trace element PM2.5 species and derivation of freeway-based
emission rates in Los Angeles, CA. Sci. Total Environ. 435–436, 159–166.
Lough, G.C., Schauer, J.J., Park, J.-S., Shafer, M.M., DeMinter, J.T., Weinstein, J.P., 2005.
Emissions of Metals Associated with Motor Vehicle Roadways. Environ. Sci. Technol.
39, 826–836.
Manchester-Neesvig, J.B., Schauer, J.J., Cass, G.R., 2003. The Distribution of Particle-Phase
Organic Compounds in the Atmosphere and Their Use for Source Apportionment during
the Southern California Children’s Health Study. J. Air Waste Manag. Assoc. 53, 1065–
1079.
110

McWhinney, R.D., Badali, K., Liggio, J., Li, S.-M., Abbatt, J.P.D., 2013. Filterable Redox
Cycling Activity: A Comparison between Diesel Exhaust Particles and Secondary
Organic Aerosol Constituents. Environ. Sci. Technol. 47, 3362–3369.
McWhinney, R.D., Gao, S.S., Zhou, S., Abbatt, J.P.D., 2011. Evaluation of the Effects of Ozone
Oxidation on Redox-Cycling Activity of Two-Stroke Engine Exhaust Particles. Environ.
Sci. Technol. 45, 2131–2136.
McWhinney, R.D., Zhou, S., Abbatt, J.P.D., 2013. Naphthalene SOA: redox activity and
naphthoquinone gas–particle partitioning. Atmos Chem Phys 13, 9731–9744.
Miljevic, B., Fairfull-Smith, K.E., Bottle, S.E., Ristovski, Z.D., 2010. The application of
profluorescent nitroxides to detect reactive oxygen species derived from combustion-
generated particulate matter: Cigarette smoke – A case study. Atmos. Environ. 44, 2224–
2230.
Minguillón, M.C., Arhami, M., Schauer, J.J., Sioutas, C., 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, 7317–7328.
Misra, C., Singh, M., Shen, S., Sioutas, C., Hall, P.M., 2002. Development and evaluation of a
personal cascade impactor sampler (PCIS). J. Aerosol Sci. 33, 1027–1047.
Montoya, L., Lawrence, J., Murthy, G.K., Sarnat, J., Godleski, J., Koutrakis, P., 2004.
Continuous Measurements of Ambient Particle Deposition in Human Subjects. Aerosol
Sci. Technol. 38, 980–990.
Morgan, T.E., Davis, D.A., Iwata, N., Tanner, J.A., Snyder, D., Ning, Z., Kam, W., Hsu, Y.-T.,
Winkler, J.W., Chen, J.-C., Petasis, N.A., Baudry, M., Sioutas, C., Finch, C.E., 2011.
Glutamatergic Neurons in Rodent Models Respond to Nanoscale Particulate Urban Air
Pollutants in Vivo and in Vitro. Environ. Health Perspect. 119, 1003–1009.
Nel, A., Xia, T., Mädler, L., Li, N., 2006. Toxic Potential of Materials at the Nanolevel. Science
311, 622–627.
Ning, Z., Geller, M.D., Moore, K.F., Sheesley, R., Schauer, J.J., Sioutas, C., 2007. Daily
Variation in Chemical Characteristics of Urban Ultrafine Aerosols and Inference of Their
Sources. Environ. Sci. Technol. 41, 6000–6006.
Nolte, C.G., Schauer, J.J., Cass, G.R., Simoneit, B.R.T., 1999. Highly Polar Organic Compounds
Present in Meat Smoke. Environ. Sci. Technol. 33, 3313–3316.
Ntziachristos, L., Ning, Z., Geller, M.D., Sheesley, R.J., Schauer, J.J., Sioutas, C., 2007. Fine,
ultrafine and nanoparticle trace element compositions near a major freeway with a high
heavy-duty diesel fraction. Atmos. Environ. 41, 5684–5696.
111

Oberdörster, G., Oberdörster, E., Oberdörster, J., 2005. Nanotoxicology: An Emerging
Discipline Evolving from Studies of Ultrafine Particles. Environ. Health Perspect. 113,
823–839.
Okuda, T., Schauer, J.J., Shafer, M.M., 2014. Improved methods for elemental analysis of
atmospheric aerosols for evaluating human health impacts of aerosols in East Asia.
Atmos. Environ. 97, 552–555.
Orru, H., Kimmel, V., Kikas, Ü., Soon, A., Künzli, N., Schins, R.P.F., Borm, P.J.A., Forsberg,
B., 2010. Elemental composition and oxidative properties of PM2.5 in Estonia in relation
to origin of air masses — results from the ECRHS II in Tartu. Sci. Total Environ. 408,
1515–1522.
Pacyna, J.M., Semb, A., Hanssen, J.E., 1984. Emission and long-range transport of trace
elements in Europe. Tellus B 36B, 163–178.
Pakbin, P., Ning, Z., Shafer, M.M., Schauer, J.J., Sioutas, C., 2011. Seasonal and spatial coarse
particle elemental concentrations in the Los Angeles area. Aerosol Sci Technol 45, 949–
963.
Penttinen, P., Timonen, K.L., Tiittanen, P., Mirme, A., Ruuskanen, J., Pekkanen, J., 2001.
Ultrafine particles in urban air and respiratory health among adult asthmatics. Eur.
Respir. J. 17, 428–435.
Perera, F.P., Rauh, V., Whyatt, R.M., Tsai, W.-Y., Tang, D., Diaz, D., Hoepner, L., Barr, D., Tu,
Y.-H., Camann, D., Kinney, P., 2006. Effect of Prenatal Exposure to Airborne Polycyclic
Aromatic Hydrocarbons on Neurodevelopment in the First 3 Years of Life among Inner-
City Children. Environ. Health Perspect. 114, 1287–1292.
Polidori, A., Hu, S., Biswas, S., Delfino, R.J., Sioutas, C., 2008. Real-time characterization of
particle-bound polycyclic aromatic hydrocarbons in ambient aerosols and from motor-
vehicle exhaust. Atmos Chem Phys 8, 1277–1291.
Pope, C, 2002. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine
particulate air pollution. JAMA 287, 1132–1141.
Prahalad, A.K., Soukup, J.M., Inmon, J., Willis, R., Ghio, A.J., Becker, S., Gallagher, J.E., 1999.
Ambient Air Particles: Effects on Cellular Oxidant Radical Generation in Relation to
Particulate Elemental Chemistry. Toxicol. Appl. Pharmacol. 158, 81–91.
Prophete, C., Maciejczyk, P., Salnikow, K., Gould, T., Larson, T., Koenig, J., Jaques, P., Sioutas,
C., Lippmann, M., Cohen, M., 2006. Effects of select PM-associated metals on alveolar
macrophage phosphorylated ERK1 and -2 and iNOS expression during ongoing alteration
in iron homeostasis. J. Toxicol. Environ. Health - Part Curr. Issues 69, 935–951.
Rattanavaraha, W., Rosen, E., Zhang, H., Li, Q., Pantong, K., Kamens, R.M., 2011. The reactive
oxidant potential of different types of aged atmospheric particles: An outdoor chamber
study. Atmos. Environ. 45, 3848–3855.
112

Roberts, J.R., Taylor, M.D., Castranova, V., Clarke, R.W., Antonini, J.M., 2004. Soluble Metals
Associated with Residual Oil Fly Ash Increase Morbidity and Lung Injury After Bacterial
Infection in Rats. J. Toxicol. Environ. Health A 67, 251–263.
Rogge, W.F., Hildemann, L.M., Mazurek, M.A., Cass, G.R., Simoneit, B.R.T., 1993. Sources of
fine organic aerosol. 4. Particulate abrasion products from leaf surfaces of urban plants.
Environ. Sci. Technol. 27, 2700–2711.
Saffari, A., Daher, N., Samara, C., Voutsa, D., Kouras, A., Manoli, E., Karagkiozidou, O.,
Vlachokostas, C., Moussiopoulos, N., Shafer, M.M., Schauer, J.J., Sioutas, C., 2013a.
Increased Biomass Burning Due to the Economic Crisis in Greece and Its Adverse
Impact on Wintertime Air Quality in Thessaloniki. Environ. Sci. Technol. 47, 13313–
13320.
Saffari, A., Daher, N., Shafer, M.M., Schauer, J.J., Sioutas, C., 2014a. Seasonal and spatial
variation in dithiothreitol (DTT) activity of quasi-ultrafine particles in the Los Angeles
Basin and its association with chemical species. J. Environ. Sci. Health Part A 49, 441–
451.
Saffari, A., Daher, N., Shafer, M.M., Schauer, J.J., Sioutas, C., 2014b. Global Perspective on the
Oxidative Potential of Airborne Particulate Matter: A Synthesis of Research Findings.
Environ. Sci. Technol. 48, 7576–7583.
Saffari, A., Daher, N., Shafer, M.M., Schauer, J.J., Sioutas, C., 2013b. Seasonal and spatial
variation in reactive oxygen species activity of quasi-ultrafine particles (PM0.25) in the
Los Angeles metropolitan area and its association with chemical composition. Atmos.
Environ. 79, 566–575.
Saffari, A., Daher, N., Shafer, M.M., Schauer, J.J., Sioutas, C., 2013c. Seasonal and spatial
variation of trace elements and metals in quasi-ultrafine (PM0.25) particles in the Los
Angeles metropolitan area and characterization of their sources. Environ. Pollut. 181, 14–
23.
Sanders, P.G., Xu, N., Dalka, T.M., Maricq, M.M., 2003. Airborne Brake Wear Debris: Size
Distributions, Composition, and a Comparison of Dynamometer and Vehicle Tests.
Environ. Sci. Technol. 37, 4060–4069.
Sannigrahi, P., Sullivan, A.P., Weber, R.J., Ingall, E.D., 2006. Characterization of Water-Soluble
Organic Carbon in Urban Atmospheric Aerosols Using Solid-State 13C NMR
Spectroscopy. Environ. Sci. Technol. 40, 666–672.
Sardar, S.B., Fine, P.M., Mayo, P.R., Sioutas, C., 2005. Size-Fractionated Measurements of
Ambient Ultrafine Particle Chemical Composition in Los Angeles Using the
NanoMOUDI. Environ. Sci. Technol. 39, 932–944.
Schauer, J.J., 2003. Evaluation of elemental carbon as a marker for diesel particulate matter. J.
Expo. Anal. Environ. Epidemiol. 13, 443–453.
113

Schauer, J.J., Cass, G.R., 2000. Source Apportionment of Wintertime Gas-Phase and Particle-
Phase Air Pollutants Using Organic Compounds as Tracers. Environ. Sci. Technol. 34,
1821–1832.
Schaug, J., Rambæk, J.P., Steinnes, E., Henry, R.C., 1990. Multivariate analysis of trace element
data from moss samples used to monitor atmospheric deposition. Atmospheric Environ.
Part Gen. Top. 24, 2625–2631.
Schoonen, M.A.A., Cohn, C.A., Roemer, E., Laffers, R., Simon, S.R., O’Riordan, T., 2006.
Mineral-Induced Formation of Reactive Oxygen Species. Rev. Mineral. Geochem. 64,
179–221.
Seinfeld, J.H., Pandis, S.N., 2012. Atmospheric Chemistry and Physics: From Air Pollution to
Climate Change. John Wiley & Sons.
Shafer, M.M., Perkins, D.A., Antkiewicz, D.S., Stone, E.A., Quraishi, T.A., Schauer, J.J., 2010.
Reactive oxygen species activity and chemical speciation of size-fractionated
atmospheric particulate matter from Lahore, Pakistan: An important role for transition
metals. J. Environ. Monit. 12, 704–715.
Sheesley, R.J., Schauer, J.J., Zheng, M., Wang, B., 2007. Sensitivity of molecular marker-based
CMB models to biomass burning source profiles. Atmos. Environ. 41, 9050–9063.
Shi, T., Schins, R.P.F., Knaapen, A.M., Kuhlbusch, T., Pitz, M., Heinrich, J., Borm, P.J.A.,
2003. Hydroxyl radical generation by electron paramagnetic resonance as a new method
to monitor ambient particulate matter composition. J. Environ. Monit. JEM 5, 550–556.
Singh, M., Misra, C., Sioutas, C., 2003. Field evaluation of a personal cascade impactor sampler
(PCIS). Atmos. Environ. 37, 4781–4793.
Sioutas, C., Delfino, R.J., Singh, M., 2005. Exposure assessment for atmospheric Ultrafine
Particles (UFPs) and implications in epidemiologic research. Environ. Health Perspect.
113, 947–955.
Snyder, D.C., Rutter, A.P., Collins, R., Worley, C., Schauer, J.J., 2009. Insights into the origin of
water soluble organic carbon in atmospheric fine particulate matter. Aerosol Sci.
Technol. 43, 1099–1107. doi:10.1080/02786820903188701
Solomon, P.A., Costantini, M., Grahame, T.J., Gerlofs-Nijland, M.E., Cassee, F.R., Russell,
A.G., Brook, J.R., Hopke, P.K., Hidy, G., Phalen, R.F., Saldiva, P., Sarnat, S.E., Balmes,
J.R., Tager, I.B., Özkaynak, H., Vedal, S., Wierman, S.S.G., Costa, D.L., 2012. Air
pollution and health: bridging the gap from sources to health outcomes: conference
summary. Air Qual. Atmosphere Health 5, 9–62.
Squadrito, G.L., Cueto, R., Dellinger, B., Pryor, W.A., 2001. Quinoid redox cycling as a
mechanism for sustained free radical generation by inhaled airborne particulate matter.
Free Radic. Biol. Med. 31, 1132–1138.
114

Stone, E.A., Hedman, C.J., Sheesley, R.J., Shafer, M.M., Schauer, J.J., 2009. Investigating the
chemical nature of humic-like substances (HULIS) in North American atmospheric
aerosols by liquid chromatography tandem mass spectrometry. Atmos. Environ. 43,
4205–4213.
Stone, E.A., Snyder, D.C., Sheesley, R.J., Sullivan, A.P., Weber, R.J., Schauer, J.J., 2008.
Source apportionment of fine organic aerosol in Mexico City during the MILAGRO
experiment 2006. Atmospheric Chem. Phys. 8, 1249–1259.
Sullivan, A.P., Weber, R.J., Clements, A.L., Turner, J.R., Bae, M.S., Schauer, J.J., 2004. A
method for on-line measurement of water-soluble organic carbon in ambient aerosol
particles: Results from an urban site. Geophys. Res. Lett. 31.
Tao, F., Gonzalez-Flecha, B., Kobzik, L., 2003. Reactive oxygen species in pulmonary
inflammation by ambient particulates. Free Radic. Biol. Med. 35, 327–340.
Thomaidis, N.S., Bakeas, E.B., Siskos, P.A., 2003. Characterization of lead, cadmium, arsenic
and nickel in PM2.5 particles in the Athens atmosphere, Greece. Chemosphere 52, 959–
966.
Tsai, F.C., Apte, M.J., Daisey, J.M., 1999. An Exploratory Analysis of the Relationship Between
Mortality and the Chemical Composition of Airborne Particulate Matter. Inhal. Toxicol.
12.
Valavanidis, A., Fiotakis, K., Bakeas, E., Vlahogianni, T., 2005. Electron paramagnetic
resonance study of the generation of reactive oxygen species catalysed by transition
metals and quinoid redox cycling by inhalable ambient particulate matter. Redox Rep. 10,
37–51.
Verma, V., Fang, T., Guo, H., King, L., Bates, J.T., Peltier, R.E., Edgerton, E., Russell, A.G.,
Weber, R.J., 2014. Reactive oxygen species associated with water-soluble PM2.5 in the
southeastern United States: spatiotemporal trends and source apportionment. Atmos
Chem Phys 14, 12915–12930.
Verma, V., Fang, T., Xu, L., Peltier, R.E., Russell, A.G., Ng, N.L., Weber, R.J., 2015. Organic
aerosols associated with the generation of reactive oxygen species (ROS) by water-
soluble PM2.5. Environ. Sci. Technol. 49, 4646–4656.
Verma, V., Ning, Z., Cho, A.K., Schauer, J.J., Shafer, M.M., Sioutas, C., 2009a. Redox activity
of urban quasi-ultrafine particles from primary and secondary sources. Atmos. Environ.
43, 6360–6368.
Verma, V., Pakbin, P., Cheung, K.L., Cho, A.K., Schauer, J.J., Shafer, M.M., Kleinman, M.T.,
Sioutas, C., 2011. Physicochemical and oxidative characteristics of semi-volatile
components of quasi-ultrafine particles in an urban atmosphere. Atmos. Environ. 45,
1025–1033.
115

Verma, V., Polidori, A., Schauer, J.J., Shafer, M.M., Cassee, F.R., Sioutas, C., 2009b.
Physicochemical and Toxicological Profiles of Particulate Matter in Los Angeles during
the October 2007 Southern California Wildfires. Environ. Sci. Technol. 43, 954–960.
Verma, V., Shafer, M.M., Schauer, J.J., Sioutas, C., 2010. Contribution of transition metals in
the reactive oxygen species activity of PM emissions from retrofitted heavy-duty
vehicles. Atmos. Environ. 44, 5165–5173.
Watson, J.G., Cooper, J.A., Huntzicker, J.J., 1984. The effective variance weighting for least
squares calculations applied to the mass balance receptor model. Atmospheric Environ.
1967 18, 1347–1355.
Weber, R.J., Sullivan, A.P., Peltier, R.E., Russell, A., Yan, B., Zheng, M., de Gouw, J.,
Warneke, C., Brock, C., Holloway, J.S., Atlas, E.L., Edgerton, E., 2007. A study of
secondary organic aerosol formation in the anthropogenic-influenced southeastern United
States. J. Geophys. Res. Atmospheres 112, D13302.
Westerdahl, D., Fruin, S., Sax, T., Fine, P.M., Sioutas, C., 2005. Mobile platform measurements
of ultrafine particles and associated pollutant concentrations on freeways and residential
streets in Los Angeles. Atmos. Environ. 39, 3597–3610.
Yeh, H.-C., Cuddihy, R.G., Phalen, R.F., Chang, I.-Y., 1996. Comparisons of Calculated
Respiratory Tract Deposition of Particles Based on the Proposed NCRP Model and the
New ICRP66 Model. Aerosol Sci. Technol. 25, 134–140.
Zhang, Y., Schauer, J.J., Shafer, M.M., Hannigan, M.P., Dutton, S.J., 2008. Source
apportionment of in vitro reactive oxygen species bioassay activity from atmospheric
particulate matter. Environ. Sci. Technol. 42, 7502–7509.
Zheng, M., Cass, G.R., Schauer, J.J., Edgerton, E.S., 2002. Source Apportionment of PM2.5 in
the Southeastern United States Using Solvent-Extractable Organic Compounds as
Tracers. Environ. Sci. Technol. 36, 2361–2371.

Abstract (if available)

Abstract There is a large body of literature indicating associations between airborne particulate matter (PM) and increased risk of a wide range of adverse health outcomes in humans. One of the major PM properties that significantly contributes to its health effects is the oxidative potential, which induces cellular oxidative stress in biological systems upon exposure and triggers both localized and systemic inflammatory responses, leading to a myriad of health effects with endpoints in the respiratory tract as well as the cardiovascular and nervous systems. Despite the compelling evidence and documentation of PM-related health effects, the state of knowledge regarding the exact PM causative agents is fairly immature. Accordingly, current PM regulations mainly target PM mass concentration, which may not be a good representative for the PM-induced health effects and toxicity. These knowledge gaps necessitate an improved identification and characterization of PM chemical composition, emission sources and their association with PM toxicity and oxidative potential. ❧ The aim of this dissertation was to investigate, in great detail, the associations between the PM chemical components and toxicity, and to provide guidelines on the specific PM sources that drive toxicity in urban atmosphere. To this end, a series of case studies were designed and executed in the Los Angeles metropolitan area, as an example of a complex urban environment impacted by a variety of PM sources. In these case studies, size resolved PM samples were collected and analyzed for their chemical composition and toxicity. Subsequently, statistical and source apportionment techniques such as molecular-marker chemical mass balance (MM-CMB), principal component analysis (PCA) and regression modeling were deployed to quantify the sources of PM and its most important sub-classes (such as metal particles) as well as to identify the associations between toxicity and different parameters. In addition to the case studies in the Los Angeles basin, the PM toxicity and chemistry data from a number of recent studies were also pooled together in a meta-analysis, to test the consistency of the findings in different urban environments. Findings of this work advance our knowledge of PM toxicity and its relationship with chemistry and emission/formation sources, and provide valuable insights for more targeted and cost-effective environmental regulations.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

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

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

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

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

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

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

PDF

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

PDF

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...

PDF

Chemical and toxicological characteristics of particulate matter in urban environments with a focus on its sources, associated health impacts and mitigation policies

PDF

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

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

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

PDF

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

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

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

PDF

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

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

PDF

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

PDF

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

PDF

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

PDF

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

Asset Metadata

Creator Saffari, Arian (author)

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

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Environmental Engineering

Publication Date 02/22/2017

Defense Date 11/09/2016

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

Tag Air pollution,OAI-PMH Harvest,oxidative potential,particulate matter

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Sioutas, Constantinos (committee chair), Ban-Weiss, George (committee member), Finch, Caleb (committee member)

Creator Email arian.saffari@gmail.com,asaffari@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-341641

Unique identifier UC11257626

Identifier etd-SaffariAri-5086.pdf (filename),usctheses-c40-341641 (legacy record id)

Legacy Identifier etd-SaffariAri-5086.pdf

Dmrecord 341641

Document Type Dissertation

Rights Saffari, Arian

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