University of Southern California Dissertations and Theses

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...

(USC Thesis Other)

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

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 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 of PM-bound metals for application
in source apportionment studies

By

Mohammad Hossein Sowlat

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

May 2019

Copyright 2019 Mohammad Hossein Sowlat
ii
Dedication

To my parents who love me unconditionally and made me who I am today.

&

To my beloved wife, my best friend, and my life partner, Saeedeh, who supported me all
the way through the toughest years of my life.
iii
Acknowledgements
First and foremost, I would like to sincerely thank my PhD advisor, Professor
Constantinos Sioutas, whose guidance, mentorship, and vision made this work possible. I
am extremely grateful for his continued support, sometimes from thousands of kilometers
away, during the course of my PhD. I believe there is nothing more valuable for a PhD
student than constructive feedback, comments, and criticism, which I was fortunate
enough to continuously receive from Professor Sioutas in all steps of my doctoral work at
USC. I am eternally indebted to him for all of his support.
I would also like to thank Professor James Schauer and Dr. Martin Shafer of University
of Wisconsin-Madison for their continued help and support in many of the projects that I
undertook during my doctoral studies at USC. Their comments and suggestions were
always helpful and resulted in the delivery of higher quality work.
Special thanks to Professor George Ban-Weiss, with whom I had the pleasure and
privilege of working in multiple occasions and on different projects. Your unconditional
support is highly appreciated.
My sincere thanks to my PhD candidacy and defense committee members, Professor
George Ban-Weiss, Professor Rob McConnell, Professor Kelly Sanders, and Professor
Felipe De Barros for providing me with their constructive comments.
I am also grateful to world-famous researchers at the Institute of Atmospheric Sciences
and Climate of the National Research Council of Italy, Dr. Stefano Decesari, Dr. Silvia
Sandrini, Dr. Stefania Gilardoni, Dr. Maria Cristina Facchini, and Dr. Sandro Fuzzi for
their help and support in the collaborative study on fog toxicity in Po Valley, Italy.
Last but not least, I would also like to extend my gratitude to my former and current
colleagues and group-mates at USC's Aerosol lab for their sincere help and support in the
iv
research projects that I have been involved in during my doctoral studies at USC. Special
thanks to Dr. Sina Hasheminassab and Dr. Dongbin Wang for being amazing and very
supportive mentors and friends. I learned a lot from them and will be eternally grateful to
them. I would also like to thank my amazing colleagues Amirhosein Mousavi, Sina
Taghvaee, and Giulia Simonetti for making my life so much easier with their amazing
work and being a constant source of energy and enthusiasm. My thanks also go to
Farimah Shirmohammadi, Christopher Lovett, Milad Pirhadi, and Ehsan Soleimanian; it
was certainly a pleasure to work with all of them.
This work was partly supported by the USC Viterbi School of Engineering’s PhD
Fellowship award. We would also like to thank the financial support from the United
States National Institute of Allergy and Infectious Diseases (award number:
5R01AI065617-15), the National Institutes of Health (grant number: 5R01ES024936-02),
and the Regione Emilia Romagna as part of the Supersito project (Decreto Regionale
428/10) and co-funded by the European Union’s Seventh Framework Programme
(FP7/2007-797 2013) under grant agreement no 603445 (BACCHUS).

v
Table of Contents
List of Figures vii
List of Tables ix

Chapter 1: Introduction 1
1.1. Background 1
1.2. Overview 3

Chapter 2: Source apportionment of ambient particle number concentrations
in central Los Angeles using positive matrix factorization (PMF)
7
2.1. Introduction 7
2.2. Methodology 11
2.2.1. Sampling site 11
2.2.2. Sampling time, method, and instrumentation 12
2.2.3. Auxiliary variables 14
2.2.4. Meteorology in central Los Angeles 15
2.2.5. PMF model 18
2.2.6. Input matrices 21
2.3. Results and discussion 21
2.3.1. Overview of the data 21
2.3.2. Number of Factors 24
2.3.3. Factor identification 36
2.4. Summary and conclusions 45

Chapter 3: Enhanced toxicity of aerosol in fog conditions in the Po Valley,
Italy
47
3.1. Introduction 47
3.2. Methods 50
3.2.1 Measurement site 50
3.2.2 Aerosol and fog-water sampling 51
3.2.3 Chemical analysis of water-soluble components 52
vi
3.2.4 Total elemental analysis of aerosol and fog samples 52
3.2.5 ROS analysis 53
3.2.6 Statistical analyses 54
3.3. Results 55
3.3.1. Partitioning of PM mass and chemical components between fog
droplets and interstitial aerosol
55
3.3.2. ROS activity of fog water and aerosol samples 60
3.3.3 Association of ROS activity to PM and fog chemical components 66
3.4. Discussion and conclusions 75

Chapter 4: Development and field evaluation of an online monitor for near-
continuous measurement of iron, manganese and chromium in coarse
airborne particulate matter (PM)
78
4.1. Introduction 78
4.2. Materials and methods 82
4.2.1. Reagents and standards 82
4.2.2. System configuration 83
4.2.3. Field evaluation tests and continuous long-term operation 86
4.2.4. SF-ICPMS analysis of filter samples 98
4.3. Results and Discussion 89
4.3.1. Calibration 89
4.3.2. Comparison between online-measured data with filter samples data 90
4.3.3. Diurnal trends of Fe, Mn, and Cr in coarse PM concentrations and
relationship to meteorological parameters in central Los Angeles
92
4.4. Summary and conclusions 103

Chapter 5: Conclusions and recommendations 105

Bibliography

107

vii
List of Figures

Figure 2.1: Diurnal variations of important meteorological parameters in the cold
and warm phases. Error bars correspond to one standard error
17
Figure 2.2: Average number and volume size distributions of all the input
samples to the PMF model in the cold and warm phases (the graphs represent
geometric means ± SE)
23
Figure 2.3: The number size distributions as well as the auxiliary variables
profiles for each of the factors resolved by the PMF model
26
Figure 2.4: Volume size distributions along with the explained variation (%) of
each factor profile resolved by the PMF model
27
Figure 2.5: Relative contribution of each factor to the total number
concentrations: a) overall phases; b) cold phase; and c) warm phase
29
Figure 2.6: Contribution (particles/cm
3
) of each of the PMF-resolved factors to
the total number concentrations in the cold and warm phases
30
Figure 2.7: Diurnal variations (geometric means) of number concentrations
(particles /cm
3
) from each factor resolved by the PMF model in the cold and
warm phases. Error bars correspond to one standard error
31
Figure 2.8: Weekday/weekend analysis of each of the factors resolved by the
PMF model (values are geometric means). Error bars correspond to one standard
error
33
Figure 2.9: Correlation between the measured vs. PMF-predicted total number
concentrations (particles/cm
3
) for the entire sampling period
36
Figure 3.1: Mass concentrations (µg/m
3
) of the Po Valley fog water and aerosol
during the study period. Bars represent geometric means and error bars
correspond to one standard error (SD)
56
Figure 3.2: Concentrations (a), and mass fractions (b), of the water-soluble
components in the fog water and aerosols. Bars represent geometric means and
error bars correspond to one standard error (SE)
58
Figure 3.3: Airborne concentrations (ng/µg) (a) and mass fractions (ng/m
3
) (b) of
metals/elements in fog water and aerosol samples collected in the Po Valley area
in Fall 2015. Bars represent geometric means and error bars correspond to one
standard error (SD)
59
Figure 3.4: Per-volume (µg/m
3
) (a) and per-mass (µg/µg) (b) concentration of
WSOC in daytime/interstitial and fog water samples in the Po Valley in Fall
2015. Bars represent geometric means and error bars correspond to one standard
error (SD)
62
Figure 3.5: Per-volume (MicrogramZymosan/m
3
air) (a) and per-mass
(MicrogramZymosan/mg PM) (b) ROS activity of the fog water and aerosol
samples collected in the Po Valley in fall 2015. The results pertain to parallel
aerosol and fog samples. The values for aerosol samples are based on the
63
viii
unfiltered ROS analysis protocol. Bars represent geometric means and error bars
correspond to one standard error (SD)
Figure 3.6: Per-volume (MicrogramZymosan/m
3
air) (a) and per-mass
(MicrogramZymosan/mg PM) (b) ROS activity of the aerosol samples collected
in the Po Valley in fall 2015. The values for aerosol samples are based on both
filtered (representing the redox activity of WS components) and unfiltered
(representing the redox activity of both WS and WI components) ROS analysis
protocols. Bars represent geometric means and error bars correspond to one
standard error (SD)
65
Figure 4.1: System schematic of the coarse PM metal monitor 84
Figure 4.2: Correlation between the metal concentrations measured online with
off-line concurrent measurements obtained using filter samplers. Error bars
represent one standard deviation of multiple online (n = 12) and offline (n=3)
measurements for each data point
92
Figure 4.3(a-c): The diurnal variation of meteorological parameters in the study
location averaged over the study period; a) wind speed; b) temperature; and c)
relative humidity. Error bars represent one standard error (SE) of the mean
94
Figure 4.4(a-d): Prevailing wind direction and speed (a) during the study period
and its relationship with concentrations of Fe (b), Mn (c), and Cr (d)
97
Figure 4.5(a-d): Diurnal variations of the coarse PM mass concentrations and the
online-measured metal concentrations averaged over the study period; a) coarse
PM; b) Fe; c) Mn; and d) Cr. For panel a, the error bars represent one standard
error (SE) of the mean. The box plots represent the interquartile range (the
bottom and top lines of the box representing the first and the third quartiles,
respectively). The line inside the box represents the median, while the whiskers
above and below the box represent the 95
th
and 5
th
percentiles
101
Figure 4.6(a-c): Diurnal variations of the mass fractions of the online-measured
metal concentrations averaged over the study period; a) Fe; b) Mn; c) Cr. Error
bars represent one standard error (SE) of the mean
102
Figure 4.7: Weekday-weekend analysis of the metallic species concentrations
during the study period. The P-values shown on the graph correspond to
independent samples T-test
103

ix
List of Tables

Table 2.1: Summary of the input parameters to the PMF model in this study 15
Table 2.2: Summary statistics for the parameters included in the PMF model 22
Table 2.3: Spearman correlation coefficient matrix indicating the association
between the auxiliary variables and the factors resolved by the PMF model. R
values above 0.5 are bolded
35
Table 3.1: Spearman rank correlation coefficients between per-volume
concentrations of water-soluble species as well as metals/elements in the fog
water samples and the corresponding ROS levels. Correlation coefficients which
were statistically significant (at P<0.05) are highlighted in bold
67
Table 3.2: Spearman rank correlation coefficients between concentrations of the
metals/elements, markers of organic aerosol, and water-soluble (WS)
components in the aerosol samples and the corresponding ROS levels.
Statistically significant (P<0.05) correlation coefficients are highlighted in bold
68
Table 3.3: Output of multiple linear regression (MLR) analysis using ROS
activity as the dependent variable and ambient concentrations of the measured
chemical species as independent variables
70
Table 3.4: Spearman rank correlation coefficients between WSOC and inorganic
species in the fog water samples. Correlation coefficients which were
statistically significant (at P<0.05) are highlighted in bold
70
Table 3.5: Results of the Principal Component Analysis (PCA) performed on the
aerosol samples
71
Table 4.1: Results of the system calibration for each of the individual metals
measured Table
90
4.2: Summary statistics of the parameters collected over the entire study period 93

1
Chapter 1:
Introduction

1.1. Background
The rapid pace of urbanization and industrialization has created a wide range of environmental
challenges in metropolitan areas around the globe, and air pollution is among the most
challenging ones (Javid et al., 2016; Lovett et al., 2018b; Naddafi et al., 2012; Sowlat et al.,
2011). Ambient particulate matter (PM) is of particular importance, due to its distinct
characteristics and diverse health impacts. For example, ambient PM can come in vastly different
sizes; in fact, there is a 4-order-of-magnitude difference in the diameter of the smallest and the
largest particles that can remain suspended in the air for significant periods of time (from a few
nm to tens of μm). Ambient PM can also come from diverse sources, including, but not limited
to, vehicles (exhaust and non-exhaust), industries, biomass burning, wildfires, and resuspension
of road dust (Mousavi et al., 2019b; Sowlat et al., 2016a; Sowlat et al., 2013; Sowlat et al., 2012;
Taghvaee et al., 2018b; Wang et al., 2016a; Xue et al., 2019). They can also be formed
secondarily due to chemical/photochemical reactions in the atmosphere. And, finally, ambient
PM can posses a wide range of chemical components, including carbonaceous materials,
elements and metals, water-soluble components, and organic species (Hassanvand et al., 2014;
Mousavi et al., 2018a; Mousavi et al., 2018b; Shahsavani et al., 2012b; Shahsavani et al., 2017;
Shirmohammadi et al., 2018; Shirmohammadi et al., 2017a). All of these characteristics have
made ambient PM a distinct air pollutant which has become the target of thousands of scientific
research studies around the globe.
2
The health impacts of exposure to ambient particulate matter (PM) have been the subject of
numerous epidemiological and toxicological studies. These studies have found statistically
significant associations between ambient PM exposure and increased risk of cardiovascular
diseases, respiratory diseases, neurodegenerative effects, lung caner, and even premature death
(Brunekreef and Forsberg, 2005; Davis et al., 2013c; Decesari et al., 2017; Dockery and Stone,
2007; Gauderman et al., 2015; Miller et al., 2007; Pope et al., 2004b; Pope Iii et al., 2002).
According to an earlier global burden of disease (GBD) study, exposure to ambient PM is
responsible for more than 3 million premature deaths all over the world every year (Lim et al.,
2013). However, a more recent GBD study indicated that exposure to air pollution is now
standing above AIDS, malaria, and malnutrition in terms of the disease burden, and coming
closely second to cigarette smoking, causing an estimated 6.5 million premature deaths in 2015
around the world (Wang et al., 2016b), and of this figure, 4.5 million death are attributable to
ambient PM exposure.
It is noteworthy that the vast majority of the epidemiological and toxicological studies that have
evaluated the health impacts of exposure to ambient PM have used PM mass concentrations as
the sole metric of PM exposure, while PM has other characteristics that can play significant roles
in terms of toxicity and, in turn, human health effects (Sowlat et al., 2016a). For instance, it has
been shown in more recent studies that PM size, number concentration, chemical composition,
and even surface area are also playing major roles in driving health impacts associated with
ambient PM (Brook et al., 2010; Chen et al., 1991; Davis et al., 2013b; Delfino et al., 2010b;
Dreher et al., 1997; Kasumba et al., 2009; Lighty et al., 2000; Oberdörster et al., 1994; Peters et
al., 1997; Sowlat et al., 2016a). In addition, rather than the total mass of ambient PM, particular
chemical components, including water-soluble organic carbon, black carbon (BC), redox-active
3
metals, and polycyclic aromatic hydrocarbons (PAHs) have been found to cause toxicity and
adverse human health effects as well (Decesari et al., 2017; Delfino et al., 2005; Donaldson et
al., 2003; Hassanvand et al., 2015; Li et al., 2009; Lovett et al., 2018a; Mousavi et al., 2019a;
Peters et al., 2006; Shahsavani et al., 2012a; Shirmohammadi et al., 2018; Taghvaee et al., 2019;
Taghvaee et al., 2018a; Tao et al., 2003), while other chemical components, such as secondary
ammonium nitrate and sulfate are considered as innocuous PM species (Argyropoulos et al.,
2016; Decesari et al., 2017; Mousavi et al., 2019a), although the contribute to a major fraction of
ambient PM, especially in the PM
2.5
size fraction (due to their secondary nature) (Mousavi et al.,
2018c; Sowlat et al., 2016a).
This makes the study of the physical, chemical, and toxicological characteristics of ambient PM
quite critical, as it provides significant insight into the contributing sources which are deriving
the toxicity of ambient PM and, therefore, most likely the health impacts. This information is
crucial for designing air pollution control schemes to effectively mitigate air pollution problem
in an area. The research presented herein provides valuable information on the physical and
chemical characteristics of ambient PM number and mass under different atmospheric
conditions, provides insight into their sources, and finally, offers a novel approach for high time-
resolution measurement of the chemical components of ambient PM that can be used in source
apportionment studies.

1.2. Overview
Our work begins by the study of physical characteristics and sources of ambient PM number
concentrations in central Los Angeles, in a location significantly influenced by traffic-related
emissions (Sowlat et al., 2016a). In this study, particle number size distributions were measured
4
over a year-long sampling campaign in central Los Angeles, CA, covering the full size spectrum
from 13 nm to 10 µm. We also collected data/performed measurements for many other
parameters, namely elemental and organic carbon (EC/OC), black carbon (BC), PM mass
concentrations, gaseous pollutants, meteorological variables, and counts of traffic, to assist us in
our source apportionment attempt. All these variables were then used as inputs to the USEPA's
Positive Matrix Factorization (PMF) model (version 5.0) to identify the sources of particle
number concentrations in central Los Angeles. The contribution of each source was quantified
and the temporal (seasonal and diurnal) variations for each sources were compared and
discussed.
In the next work, we present the chemical composition and toxicity of ambient PM and fog water
in a rural location in the Po Valley, Italy, an area that experiences severe fog episodes during the
winter time, explore the sources that contribute to ambient PM, fog water, and their associated
toxicity, and determine how fog partitioning impacts the chemical composition and toxicity of
ambient PM (Decesari et al., 2017). For this purpose, we collected fog water samples using a fog
water collector, while daytime and nighttime particles were also collected using a high-volume
sampler. The nighttime particles collected in this study corresponded to interstitial particles,
which means that these are particles that are not scavenged by fog. Samples were analyzed for
their carbonaceous, ionic, elemental and metallic components. We also analyzed the fog water as
well as PM samples for oxidative potential, using the alveolar macrophage (AM) assay. We then
explored how fog formation might impact the chemical composition and, in turn, toxicity of
ambient PM. The results from study have important implications for other areas, such as Los
Angeles, that experience fog episodes, because, as described in detail in the paper, this
5
phenomenon can critically increase the toxicity of ambient PM through aqueous-phase
chemistry.
And, finally, in the third and last study, we developed an online metal monitor for time-resolved
measurement of (a time resolution of 2-hr) three important redox-active metals, namely Iron
(Fe), Manganese (Mn), and Chromium (Cr), that are bound to ambient coarse particulate matter
(PM) (i.e., PM
10-2.5
) (Sowlat et al., 2016b). Due to the great importance of these redox-active
metals in driving the toxicity of ambient PM, developing new techniques for their time-resolved
measurements would greatly benefit studies on the identification of their source and formation
mechanisms in the atmosphere, especially given that traditional sampling methods have quite
low time resolutions, usually 24 hr. To develop the monitor, we used virtual impactors (VIs) to
first enrich coarse PM concentration into the target flow and then capture them in water slurry
using a Biosampler. We then added specific chemical reagents (pertinent to each metal) to the
slurry samples, which lead to the formation of colored complexes, the intensity of which was
then measured using a Micro Volume Flow Cell (MVFC) coupled with UV/VIS
spectrophotometry. The intensity of the color (measured by light absorption) was used as a
metric for the concentration of the target metal inside the slurry, which is proportional to the
ambient concentration of that metal, calculated through calibration experiments on dilution series
of known concentrations. The monitor was then deployed in the field, and its performance was
measured and compared with that of a standard method (i.e., inductively coupled plasma-mass
spectrometry (ICP-MS)) over a four-month period from January through April 2016. The data
collected by the monitor were then used to obtain time-resolved time-series and diurnal
variations plots of metal concentrations, which were used in combination of wind rose plots in
infer information about the sources. The metal monitor showed excellent promise in obtaining
6
time-resolved concentration data for the target metals (i.e., Fe, Mn, Cr), which can be potentially
used to increase our knowledge of the sources and formation mechanisms of these metals by
performing source apportionment studies on the collected data.
7
Chapter 2:
Source apportionment of ambient particle number concentrations in central
Los Angeles using positive matrix factorization (PMF)
In this study, the positive matrix factorization (PMF) receptor model (version 5.0) was used to
identify and quantify major sources contributing to particulate matter (PM) number concentrations,
using PM number size distributions in the range of 13 nm to 10 μm combined with several auxiliary
variables, including black carbon (BC), elemental and organic carbon (EC/OC), PM mass
concentrations, gaseous pollutants, meteorological, and traffic counts data, collected for about 9
months between August 2014 and 2015 in central Los Angeles, CA. Several parameters, including
particle number and volume size distribution profiles, profiles of auxiliary variables, contributions of
different factors in different seasons to the total number concentrations, diurnal variations of each of
the resolved factors in the cold and warm phases, weekday/weekend analysis for each of the resolved
factors, and correlation between auxiliary variables and the relative contribution of each of the
resolved factors, were used to identify PM sources. A six-factor solution was identified as the
optimum for the aforementioned input data. The resolved factors comprised nucleation, traffic 1,
traffic 2 (with a larger mode diameter than traffic 1 factor), urban background aerosol, secondary
aerosol, and soil/road dust. Traffic sources (1 and 2) were the major contributor to PM number
concentrations, collectively making up to above 60% (60.8–68.4%) of the total number concentrations
during the study period. Their contribution was also significantly higher in the cold phase compared
to the warm phase. Nucleation was another major factor significantly contributing to the total number
concentrations (an overall contribution of 17%, ranging from 11.7 to 24%), with a larger contribution
during the warm phase than in the cold phase. The other identified factors were urban background
aerosol, secondary aerosol, and soil/road dust, with relative contributions of approximately 12% (7.4–
17.1%), 2.1% (1.5–2.5%), and 1.1% (0.2–6.3%), respectively, overall accounting for about 15%
(15.2–19.8%) of PM number concentrations. As expected, PM number concentrations were
dominated by factors with smaller mode diameters, such as traffic and nucleation. On the other hand,
PM volume and mass concentrations in the study area were mostly affected by sources with larger
mode diameters, including secondary aerosols and soil/road dust. Results from the present study can
be used as input parameters in future epidemiological studies to link PM sources to adverse health
effects as well as by policymakers to set targeted and more protective emission standards for PM.

This chapter is based on the following publication:

Sowlat, M.H., Hasheminassab, S., Sioutas, C., 2016. Source apportionment of ambient particle
number concentrations in central Los Angeles using positive matrix factorization (PMF). Atmospheric
Chemistry and Physics 16, 4849-4866.

2.1. Introduction
Numerous epidemiological studies have provided compelling evidence linking exposure to
ambient particulate matter (PM) with increased risk of respiratory and cardiovascular diseases,
hospitalization, and premature mortality (Brunekreef and Forsberg, 2005; Dockery and Stone,
8
2007; Gauderman et al., 2015; Miller et al., 2007; Pope et al., 2004b; Pope Iii et al., 2002).
According to the most recent global burden of disease study, over 3 million premature deaths
annually occur all around the globe due to exposure to ambient PM (Lim et al., 2013). It should,
however, be noted that most of these epidemiological studies have related the aforementioned
health outcomes with solely the mass concentrations of PM and, therefore, do not adequately
represent submicron particles (Ogulei et al., 2007), mainly because this PM fraction contributes
negligibly to total ambient PM mass (Delfino et al., 2005; Vu et al., 2015). More recently,
studies have associated human health effects with particles characteristics other than mass
concentration, including size, number concentration, chemical composition, and even surface
area (Brook et al., 2010; Chen et al., 1991; Davis et al., 2013b; Delfino et al., 2010b; Dreher et
al., 1997; Kasumba et al., 2009; Lighty et al., 2000; Oberdörster et al., 1994; Peters et al., 1997).
Even though our knowledge of which particle characteristic (mass, size, surface area, etc.) can be
considered as the best predictor for human health outcomes is limited, there is growing evidence
highlighting the critical role of particle size and number concentrations from a human health
effect perspective (Vu et al., 2015). For example, studies have indicated that ultrafine particles
(UFP, i.e. particles with an aerodynamic diameter of <100 nm) have higher toxicity per unit mass
(Donaldson et al., 1998; Li et al., 2003; Nel et al., 2006; Oberdörster et al., 2002), have higher
deposition efficiencies in the lung (Venkataraman, 1999), and penetrate deeper into the alveolar
regions of lungs (Sioutas et al., 2005b). Additionally, several studies have also found that PM
number concentrations (mostly UFPs) can be associated with adverse effects on human health,
particularly for cardiovascular diseases (Delfino et al., 2005; Peters et al., 1997; Wichmann et al.,
2000).
9
Regulations on PM number concentrations have already been implemented on motor vehicle
emissions in a few countries. For instance, the Euro 5b and 6 have set a limit to particle number
emission factors, in addition to particle mass emission limits, for heavy-duty and gasoline
vehicles (http://www.dieselnet.com/standards/eu/ld.php, accessed 20 October 2015). It is also
expected that this approach be gradually adopted in other parts of the world (Friend et al., 2013),
based mainly on the critical health implications of the PM number concentrations, especially in
smaller fractions like UFP. This emphasizes the necessity of identification and quantification of
PM sources based on number as well as mass (Friend et al., 2013). This allows for source-
specific assessment of health effects of exposure to PM to provide us with the knowledge
required to develop efficient control strategies for PM emissions from major sources to minimize
those health effects (Yue et al., 2008).
Positive Matrix Factorization (PMF) is one of the most widely used receptor models that have
been successfully applied to identify and quantify sources of atmospheric particles. The vast
majority of previous efforts has been devoted to the identification of sources that contribute to
the mass of particles using PMF on chemically-speciated PM mass data in different parts of the
world (Alleman et al., 2010; Dutton et al., 2010; Lim et al., 2010; Sofowote et al., 2015; Sowlat
et al., 2013; Sowlat et al., 2012). Recently, attempts have been made to characterize sources that
contribute to particle number, rather than mass, using PMF applied to particle number size
distribution data. These studies have adopted different approaches in their source apportionment,
including: (1) using particle number size distribution together with gaseous pollutants, chemical
composition, meteorological, or traffic data in the PMF analysis (Beddows et al., 2015; Harrison
et al., 2011; Kasumba et al., 2009; Ogulei et al., 2007; Ogulei et al., 2006b; Thimmaiah et al.,
2009; Zhou et al., 2005), (2) using particle number size distribution and chemical composition
10
data in separate and/or combined PMF runs (Beddows et al., 2015; Gu et al., 2011), (3)
comparing PMF results with actual events during the study period (Ogulei et al., 2007), and (4)
simply correlating the PMF results with gaseous pollutants data (Friend et al., 2013; Friend et al.,
2012; Kim et al., 2004). It is noteworthy that the major factors resolved by these studies have
been nucleation, traffic, secondary aerosol, urban background, and wood burning.
Numerous studies have been performed in Los Angeles evaluating PM number concentrations as
well as size distributions, with a focus on vehicular emissions as a major source of particle
number in urban areas (Singh et al., 2006; Zhang et al., 2005). Source apportionment of
atmospheric particles has also been extensively studied in Los Angeles, but almost all of the
studies have focused on the contribution of different sources to PM mass rather than PM number
concentration (Ham and Kleeman, 2011; Hasheminassab et al., 2013; Hwang and Hopke, 2006;
Kim and Hopke, 2007; Kim et al., 2010; Schauer and Cass, 2000). To the best of our knowledge,
no source apportionment studies have ever been performed in Los Angeles on particle number
size distributions using PMF. The only study providing a source apportionment of particle
number concentrations in Los Angeles is that of Brines et al. (2015), in which major sources
contributing to particle number concentrations were identified in five high-insolation cities
around the world (Barcelona, Madrid, Roma, Los Angeles, and Brisbane) using the k-means
clustering method. It should, however, be noted that, in case of Los Angeles, the Brines et al.
(2015) study used particle number size distribution data for a rather limited time period (i.e., 3
months); moreover, the studied size distributions ranged from 13 nm to 400 nm, thus excluding
potentially important PM sources contributing to the larger size fractions of PM
2.5
and/or PM
10
.
In the present work, we collected high-resolution (5-min measurements), wide-spectrum particle
number size distribution data (i.e., 13 nm to 10 µm, covering the nucleation, Aitken,
11
accumulation, and coarse PM modes) over a long period of time (i.e., 9 months, covering both
warm and cold seasons) in a location near downtown of Los Angeles, California, to identify and
quantify sources contributing to particle number concentrations using the most recent version of
the PMF model (version 5.0). We also included gaseous pollutants (i.e., CO, NO, NO
2
, O
3
),
particle mass (PM
10-2.5
and PM
2.5
), meteorological (temperature, relative humidity (RH), and
wind speed), black carbon (BC), elemental carbon (EC) as well as primary (POC) and secondary
organic carbon (SOC), and traffic (counts of light-duty vehicles (LDVs) and heavy-duty vehicles
(HDVs)) data as inputs to help identify the factors resolved by the model. Results from the
present study can be used as a platform for future health effect studies to estimate the source-
specific impact of exposure to PM from a number concentration perspective, which is critical for
development and establishment of abatement strategies and standards in order to minimize the
most relevant health outcomes.

2.2. Methodology
2.2.1. Sampling site
Continuous measurements were carried out at the particle instrumentation unit (PIU) located on
the University of Southern California's (USC) park campus, approximately 3 km south of
downtown Los Angeles, CA. The PIU is located within approximately 150 m downwind of a
routinely congested interstate freeway, i.e. I-110, and is also in close proximity to parking and
construction facilities. Previous studies conducted by this research group have indicated that the
PIU is a mixed urban site that is also heavily impacted by vehicular emissions (Geller et al.,
2004; Hasheminassab et al., 2014b; Moore et al., 2007).

12
2.2.2. Sampling time, method, and instrumentation
Continuous measurements were conducted at the PIU from August 2014 through March 2015 as
well as in August 2015. To obtain number size distribution of atmospheric particles in the size
range of 14 -760 nm (mobility diameter), a Scanning Mobility Particle Sizer (SMPS
TM
, TSI
Model 3081) was used, which was connected to a Condensation Particle Counter (CPC, model
3020, TSI Inc., USA). Particles in the size range of 0.3-10 µm (optical diameter) were measured
using an Optical Particle Sizer (OPS
TM
, Model 3330, TSI Inc., USA). The time resolution for
these two instruments was 5 min. The OPS instrument was calibrated by the manufacturer using
Polystyrene Latex (PSL) particles, which have a dynamic shape factor of 1 (i.e., spherical
particles) and a refractive index of 1.59. It should also be noted that the measurements provided
by the OPS instrument depend primarily on the refractive index and the dynamic shape factor
(Hasheminassab et al., 2014b). Numerous studies have indicated that for spherical particles, the
size selection offered by optical particle counters, such as the OPS instrument, is quite similar to
the actual physical diameter of the particle being measured (Chen et al., 2011; Hasheminassab et
al., 2014b; Hering and McMurry, 1991; Reid et al., 1994). That said, there is compelling
evidence in the literature supporting the fact that the refractive index and the dynamic shape
factor for ambient aerosols in urban areas (such as Los Angeles) are quite similar to those of PSL
particles (Covert et al., 1990; Ebert et al., 2004; Hänel, 1968; Kent et al., 1983; Stolzenburg et
al., 1998; Strawa et al., 2006; Watson et al., 2002). To further evaluate this assumption, we used
the Multi-Instrument Manager (MIM
TM
) software, developed by TSI Inc., USA, which estimates
the refractive index and dynamic shape factor of aerosols from parallel SMPS and OPS scans.
The output from this software indicated that the average real part of the refractive index for the
aerosols collected in this study was 1.59±0.01 and their dynamic shape factor was 0.99±0.02.
13
This finding is also in concert with the results of Hasheminassab et al. (2014b), which reported
an average shape factor of near unity at the same sampling site, using the apparent and material
density of aerosols. Hence, further adjustment of OPS sizing was deemed unnecessary and the
OPS size distribution, with the original size selection, was merged with the SMPS size spectra.
Therefore, size bins covering the range of 13.6-514 nm from SMPS (without subsequent
combination into larger size fractions) were merged with the OPS channels from 0.522 to 9.01
µm as the input data to the PMF model. More detailed information on the sensitivity of the OPS
sizing to the refractive index and the dynamic shape factor of aerosols can be found in
(Hasheminassab et al., 2014b).
Black carbon (BC) concentrations, with a time resolution of 15 min, were measured using a
portable Aethalometer (Magee scientific, model AE-42). Hourly concentrations of elemental
carbon (EC) and organic carbon (OC) were measured using a semi-continuous EC/OC carbon
aerosol analyzer (Model 4, Sunset Laboratory Inc., USA), using the thermal/optical
transmittance measurement protocol of the National Institute of Occupational Safety and Health
(NIOSH 5040). By applying the “EC tracer method”, Saffari et al. (2016) estimated the primary
organic carbon (POC) and secondary organic carbon (SOC) concentrations from total OC at the
same location. These two parameters (i.e., POC and SOC) were also used as input parameters in
the PMF model, as they can provide valuable input regarding the detection of primary and
secondary sources of PM. The EC tracer method has been discussed in detail elsewhere (Day et
al., 2015;Saffari et al., 2016). Briefly, the main assumption in this method is that EC and POC
are released from similar sources; therefore, this approach is most applicable where combustion
is the main source of ambient POC (Day et al., 2015). It is noteworthy that, in the present study,
the sampling site was located in close proximity to a major freeway, thereby making the EC
14
tracer suitable for the data collected in this location, as it has also been used in previous studies
in the same sampling site (Polidori et al., 2007;Saffari et al. 2016) as well as similar locations in
the Los Angeles basin (Na et al., 2004; Strader et al., 1999). In this method, the following
equations can be used after determining the ratio of POC to EC to estimate the concentration of
SOC:
POC = [OC/EC]
p
x EC + b (1)
SOC = OC – POC (2)
where, [OC/EC]
p
is the POC to EC ratio; b is the intercept of the linear regression between POC
and EC, which is considered to be the portion of POC associated with non-combustion
emissions. Using equation (1), the slope and the intercept of the regression line were found to be
1.55 (±0.07) and 0.45 (±0.24), respectively. More detailed information on the results obtained
using the EC tracer method can found elsewhere (Saffari et al., 2016). It is also noteworthy that
we used the “high EC edge method” to determine observations with a high probability of
dominant POC contribution, which is believed to be a more accurate method for the
identification of the [OC/EC]
p
ratio compared to the traditional approach, as discussed by Day et
al. (2015), and has also been successfully applied in a number of previous studies (Harrison and
Yin, 2008; Lim and Turpin, 2002; Na et al. 2004).

2.2.3. Auxiliary variables
To help better identify the factors resolved by the PMF model, additional parameters, including
gaseous pollutants (i.e., CO, NO, NO
2
, and O
3
) and particulate matter mass concentrations in two
size fractions (i.e., PM
10-2.5
and PM
2.5
), meteorological parameters (i.e., temperature, relative
humidity, and wind speed), and traffic flow data (counts of LDVs and HDVs), were also
15
included in the model as auxiliary variables. Hourly concentrations of particulate mass and
gaseous pollutants together with hourly measurements of meteorological parameters were
acquired from the online data base of California Air Resources Board (CARB), for the sampling
site located in Downtown Los Angeles (North Main St.), approximately 3 km to the northeast of
the PIU. The hourly traffic flow data were acquired from the nearest vehicle detection station
(VDS) to our sampling site on the Freeway I-110, operated by the freeway performance
measurement system (PeMS), under the California Department of Transportation (CalTrans).
Table 1 provides a summary statistics of the input parameters to the PMF model in this study. To
achieve the same time resolution across all variables, we calculated hourly-averaged data points
for all variables.

Table 2.1: Summary of the input parameters to the PMF model in this study.
Parameter Source of data Time resolution in original data set
EC, OC Sunset monitor 1 hr
Size Distribution (14-760 nm) SMPS 5 min
Size Distribution (0.3-10 µm) OPS 5 min
BC Aethalometer 15 min
PM mass concentration data (PM
10-2.5
, PM
2.5
) CARB 1 hr
Gaseous Pollutants (NO, NO
2
, CO, O
3
, SO
2
) CARB 1 hr
Meteorological data (T, RH, WS) CARB 1 hr
Traffic data (counts of LDV and HDV) PeMS 1 hr

2.2.4. Meteorology in central Los Angeles
To evaluate the impact of meteorological conditions on factor contributions as well as to better
identify the resolved factors based on their expected seasonal trends, the study period was
16
partitioned into two phases, i.e., colder phase (from November to February) and warmer phase
(from August to October as well as March), and the model outputs, except for factor profiles, are
presented for each phase accordingly. Figure 1 illustrates the diurnal variation of important
meteorological parameters, namely, temperature, RH, wind speed, and solar radiation, in the cold
and warm phases. As can be seen from the figure, on average, temperature was 5-7 °C higher in
the warm phase than in the cold phase, although the trends were similar in both phases.
Minimum temperatures were observed in the early morning (coinciding with morning rush
hours), while maximum temperatures were seen at around noon. Conversely, RH peaked at night
and exhibited a minimum in the early afternoon. RH was also slightly higher in the warm phase
than in the cold phase. As expected, wind speed peaked in the early afternoon during the warm
phase and slightly shifted to the evening in the cold phase, while the slowest winds were blown
during nighttime. The wind speeds were also higher in the warm phase compared to the cold
phase. Solar radiation had a consistent trend in both phases, peaking at noon, with the levels
being higher in the warm phase than in the cold phase, as one would expect. Similar trends and
levels were also observed by Hasheminassab et al. (2014b) in central LA, indicating the
occurrence of stable atmospheric conditions during nighttime until morning rush hours,
especially in colder months of the year.

17
Figure 2.1: Diurnal variations of important meteorological parameters in the cold and warm
phases. Error bars correspond to one standard error.
10
12
14
16
18
20
22
24
26
28
30
1 3 5 7 9 11 131517 192123
Hour of the Day
Temperature (°C)
Cold Phase Warm Phase

0
10
20
30
40
50
60
70
80
90
100
1 3579 11 13 15 17 19 21 23
Hour of the Day
RH (%)
Cold Phase Warm Phase

0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
1 3 5 7 9 11 131517192123
Hour of the Day
Wind Speed (m/s)
Cold Phase Warm Phase

0
100
200
300
400
500
600
700
800
900
1000
1 3 5 7 9 111315 17 19 2123
Hour of the Day
Solar Radiation (W/m2)
Cold Phase Warm Phase

18
2.2.5. PMF model
PMF, first developed by Paatero and Tapper (1993), is a multivariate statistical model used for
identifying and quantifying the contribution of different sources to a set of samples using the
fingerprints of those sources. This multivariate factor analysis tool decomposes a matrix of
speciated data into two sub-matrices, i.e., factor profiles and source contributions, as shown
below (Krecl et al., 2008):
X=G.F+E (3)
where, X is the matrix of samples (here, particle number size distribution together with auxiliary
variables data); G is the matrix containing source contributions; F is the matrix containing factor
profiles; and E is the residual matrix.
The above equation can also be expressed mathematically, as the following (Norris et al., 2014):


 
p
k
ij kj ik ij
e f g x
1
(4)
where, x
ij
is the PM number concentration (or concentration of another auxiliary species) for the
ith sample and the jth size bin (or species); p is the number of factors that contribute to the PM
number concentrations; g
ik
is the relative contribution of kth factor to ith sample; f
ik
is the PM
number concentration of jth size bin in the kth factor; and e
ij
is the residual (observed‒estimated)
value for the ith sample and jth size bin.
With the constraint that no sample can have a significantly negative contribution and using a
least-square method, the PMF then resolves factor profiles and contributions by attempting to
minimize the Q value, as shown below (Paatero, 1997; Paatero and Tapper, 1994):
2
1
1 1
.















 

  ij
p
k
kj ik ij m
j
n
i
u
f g x
Q (5)
19
where, u
ij
is the uncertainty associated with the sample x
ij
.
One of the advantages of the PMF model is weighting every single value in the input data matrix
using user-provided uncertainties, enabling the model to allow for measurement confidence in
resolving the factor profiles and contributions (Norris et al., 2014). In the present work, since no
measurement uncertainties were available for the input parameters, we applied the method
suggested by Ogulei et al. (2006a; 2006b) and Zhou et al. (2014) to calculate the uncertainties for
individual data points inserted into the model. For this purpose, measurement errors were first
estimated for each data point using the following equation:
) (
1
j
j i ij
N N C    (6)
where, σ
ij
is the estimated measurement error for the ith sample and jth size bin (or concentration
of auxiliary variables); C
1
is an empirical constant usually between 0.01 and 0.05; N
ij
is the
observed number concentration for the ith sample and jth size bin (or concentration of auxiliary
variables); and j N is the arithmetic mean of the PM number concentrations for the jth size bin
(or concentration of auxiliary variables).
The value of the measurement method obtained from the above equation is then used to calculate
the measurement uncertainty, according to the following equation:
) , max(
2 ij ij ij ij
y x C S    (7)
where, S
ij
is the calculated uncertainty associated with the ith sample and jth size bin; C
2
is an
empirical constant usually between 0.1 and 0.5; and Y
ij
is the value calculated by the model for
x
ij
. In the present work, C
1
and C
2
values of 0.05 and 0.1 were chosen to obtain the most
physically interpretable solution using a trial and error approach.
In the present study, the most recent version of the PMF model, version 5.0, newly released by
the United States Environmental Protection Agency (PMF guide), was used. Uncertainties
20
associated with the resolved factor profiles were estimated using three error estimation methods,
namely, Displacement (DISP) analysis, Bootstraps (BS) method, and a combination of DISP and
BS methods (BS-DISP). For the DISP analysis, a solution was considered valid if the observed
drop in the Q value was below 0.1% and there was no factor swaps for the smallest dQ
max
(i.e.,
4). For the BS method, 100 runs were selected and a solution was considered valid if all of the
factors had a mapping of above 90%. For the BS-DISP analysis, a solution was considered valid
if the observed drop in the Q value was below 0.5% (Brown et al., 2015; Norris et al., 2014;
Paatero et al., 2014).
The PMF model was run in the robust mode, which down-weights the effect of values with high
uncertainties (i.e., values set as “weak” in the model) on the final solution resolved by the model
(Brown et al., 2015). Missing values were replaced by interpolating the previous and the next
data points in the matrix; however, to decrease the effect of these replaced values on the final
solution, their uncertainty was set as three times the mean uncertainty for that species (that is
practically what the model does to set a species as “weak”). Based on the recommendations
presented by Brown et al. (2015), genuine zero values were included in the input matrix. Particle
number concentration (PNC) was selected as the “total variable”, and the PMF model
automatically turned it into a weak species by increasing its uncertainty by a factor of 3. An extra
model uncertainty of 5% was also set to account for errors that are not covered in the input
uncertainty values (Reff et al., 2007), since the uncertainty matrix only includes the effect of
random as well as experimental errors.

21
2.2.6. Input matrices
The model was run in two different scenarios, one with EC/OC data, which included 1053
samples of 131 species, and one without EC/OC data, which included 2976 samples of 129
species. This was due mainly to the fact that the EC/OC data were being collected in parallel for
a different study that coincided with the current work in a span of time shorter than the entire
study period. Therefore, in order to keep the large number of samples from the main study (i.e.,
2976) as well as to use the critical advantage of having EC/OC data in the factor identification
process, it was decided to run the PMF model in two different scenarios, one including EC/OC
data and one without these data. It should also be noted that although the latter matrix contained
BC data, this variable was excluded from the former matrix to avoid double counting, as EC was
already included in the dataset.

2.3. Results and discussion
2.3.1. Overview of the data
Table 2 presents the statistical characteristics of the species included in the PMF model. In this
table, signal-to-noise (S/N) ratio is a parameter that indicates if the variability in the
measurements is real or within the data noise. In the current version of the model, i.e., PMF 5.0,
the method used for calculating the S/N ratio has been updated compared to the previous
versions, resolving the disadvantages associated with the previous method of S/N calculation (for
a more detailed discussion on the S/N calculation methods, see SI). In the current method, if the
resulting S/N ratio is above 1, it can be concluded that the species has a reliable signal. As
reported in Table 2, all the species in the input matrix had S/N ratios well above 1, indicating
very strong signals for all the variables.
22
Table 2.2: Summary statistics for the parameters included in the PMF model.
Species Geometric Mean Standard Error Min Max S/N ratio
Total number
concentration (#/cm
-3
)
6860.00 94.10 524.00 32400.00 7.00
PM
10-2.5
(µg/m
3
) 15.90 0.19 2.00 77.00 7.00
PM
2.5
(µg/m
3
) 14.50 0.23 1.00 101.00 6.90
CO (ppm) 0.58 0.01 0.10 2.19 7.10
NO (ppb) 8.46 0.57 1.00 212.00 5.80
NO
2
(ppb) 22.50 0.23 1.90 75.00 7.10
O
3
(ppb) 17.40 0.33 2.00 105.00 6.80
BC (µg/m
3
) 1.14 0.02 0.124 9.13 6.90
POC*(µg/m
3
) 2.20 0.08 0.10 19.20 6.80
SOC* (µg/m
3
) 2.13 0.05 0.04 16.30 7.10
EC* (µg/m
3
) 1.01 0.03 0.01 7.34 8.80
RH (%) 50.40 0.40 6.00 99.00 7.10
Temperature (°C) 18.80 0.13 3.89 38.33 7.30
Wind speed (m/s) 4.03 0.04 1.00 14.00 6.80
LDV (#/h) 3790 34 691 7620 7.10
HDV (#/h) 153 3 5 920 6.80
*Values are pertaining to the runs including EC/OC data.

Figure 2 depicts the average number and volume size distributions of all the input data to the
PMF model by phase, which were collected during the entire study period. As shown in the
figure, the vast majority of the particles were smaller than 100 nm, and the number concentration
had a mode diameter at around 40 nm. Additionally, a significantly higher number concentration
was observed in the cold phase compared to the warm phase, which is consistent with the results
from the previous studies conducted in Los Angeles (Hudda et al., 2010; Singh et al., 2006).
23
Regarding volume concentrations, we observed one minor volume mode at the size range of 300-
500 nm and a major mode at around 4-6 µm. In this case, the volume concentration was higher in
the cold phase than in the warm phase for the minor mode diameter (at 300-500 nm), while a
sharper peak was observed for the major mode diameter (at around 4-6 µm) in the warm phase
compared to the cold phase. This PM volume size distribution is typical of urban areas (Vu et al.,
2015), and is also consistent with the findings of a previous study conducted recently by this
research group at the same sampling location (Hasheminassab et al., 2014b).

Figure 2.2: Average number and volume size distributions of all the input samples to the PMF
model in the cold and warm phases (the graphs represent geometric means ± SE).
0
2000
4000
6000
8000
10000
12000
10 100 1000 10000
Diameter (nm)
dN/dLogDp (Particle/cm3)
Cold Warm
0
5
10
15
20
25
30
10 100 1000 10000
Diameter (nm)
dV/dlogDp (µm3/ cm3)
COLD WARM

24
2.3.2. Number of Factors
In the present work, the PMF model was run several times using different number of factors,
input uncertainty matrices (as noted in the methods section), and extra modeling uncertainties to
obtain the best and most physically applicable solution. Additionally, we used several criteria to
determine the best solution resolved by the model, including: 1) particle number size distribution
profiles for different factors; 2) volume size distribution profiles for the resolved factors; 3)
profiles of auxiliary variables for different factors; 4) contribution of each factor in different
seasons to the total number concentrations; 5) diurnal variations of each of the resolved factors in
the cold and warm phases; 6) diurnal variations of each of the resolved factors in weekdays
versus weekends; and 7) correlation between auxiliary variables and the relative contribution of
each of the resolved factors. The six-factor solution was found to present the most physically
explainable one, and was, therefore, chosen as the final solution. When the model was run with
one less factor (i.e., 5-factor solution), the model could not distinguish the two traffic factors, and
Traffic 1 and Traffic 2 factors were merged together. On the other hand, when the model was run
with one more factor (i.e., 7-factor solution), a new factor was resolved by the model , having a
mode diameter between that of “urban background aerosol” and “secondary aerosol”, but
without having any distinct diurnal, seasonal, or weekday/weekend trends or auxiliary variables
profile. Therefore, this factor could not be meaningfully interpreted and identified, prompting us
to choose the 6-factor as the optimal solution.
Figure 3 illustrates the number size distributions as well as the auxiliary variables profiles for
each of the factors resolved by the PMF. Figure 4 indicates volume size distribution of each
factor. In Figures 3, 4 the black solid lines represent absolute concentrations (number or volume)
of each size bin and should be read from the left Y axis, while the grey triangles represent the
25
explained variation of each size bin and should be read from the right Y axis. The relative
contributions (overall, and by cold or warm phases) of each factor to the total number
concentrations are shown in Figure 5. Figure 6 illustrates the contribution (particles/cm
3
) of each
of the PMF-resolved factors to the total number concentrations in the cold and warm phases
within box and whisker plot. The diurnal variations and the weekday/weekend trends (geometric
means) for each of the factors are illustrated in Figures 7 and 8, respectively. The spearman
correlation coefficient matrix indicating the association between the auxiliary variables and the
factors resolved by the PMF model is also presented in Table 3. Figure 9 also illustrates the
correlation between the measured and PMF-predicted total number concentrations for the entire
sampling period. As can be seen in the figure, the high correlation between measured and
predicted values (R
2
=0.99) and very close to 1 slope of the regression line indicate that the PMF
model has been successful in modeling the input data and apportioning the total PM number
concentrations to the resolved factors.

26
Figure 2.3: The number size distributions as well as the auxiliary variables profiles for each of
the factors resolved by the PMF model.
Nucleation
0
500
1000
1500
2000
2500
3000
3500
4000
10 100 1000 10000
Diameter (nm)
dN/dLogDp
(Particle/cm3)
0
20
40
60
80
100
Explained Variation
(%)
Concentration Percentage
Nucleation
0
10
20
30
40
50
60
70
80
PM10-2.5
PM2.5
CO
NO
NO2
O3
BC
RH
Temp
WS
LDV
HDV
Species
Normalized Concentration

Traffic 1
0
1000
2000
3000
4000
5000
6000
7000
8000
10 100 1000 10000
Diameter (nm)
dN/dLogDp
(Particle/cm3)
0
10
20
30
40
50
60
70
80
Explained Variation (%)
Concentration Percentage
Traffic 1
0
10
20
30
40
50
60
70
80
PM10-2.5
PM2.5
CO
NO
NO2
O3
BC
RH
Temp
WS
LDV
HDV
Species
Normalized Concentration

Traffic 2
0
10
20
30
40
50
60
70
80
PM10-2.5
PM2.5
CO
NO
NO2
O3
BC
RH
Temp
WS
LDV
HDV
Species
Normalized Concentration
Traffic 2
0
1000
2000
3000
4000
5000
6000
10 100 1000 10000
Diameter (nm)
dN/dLogDp
(Particle/cm3)
0
10
20
30
40
50
60
Explained Variation (%)
Concentration Percentage

Urban Background Aerosol
0
500
1000
1500
2000
2500
3000
10 100 1000 10000
Diameter (nm)
dN/dLogDp
(Particle/cm3)
0
10
20
30
40
50
60
70
80
Explained Variation (%)
Concentration Percentage
Urban Background Aerosol
0
10
20
30
40
50
60
70
80
PM10-2.5
PM2.5
CO
NO
NO2
O3
BC
RH
Temp
WS
LDV
HDV
Species
Normalized Concentration

27
Secondary Aerosol
0
500
1000
1500
2000
2500
3000
3500
4000
10 100 1000 10000
Diameter (nm)
dN/dLogDp
(Particle/cm3)
0
10
20
30
40
50
60
70
80
90
Explained Variation (%)
Concentration Percentage
Secondary Aerosol
0
10
20
30
40
50
60
70
80
PM10-2.5
PM2.5
CO
NO
NO2
O3
BC
RH
Temp
WS
LDV
HDV
Species
Normalized Concentration

Soil/Road Dust
0
500
1000
1500
2000
2500
3000
3500
4000
10 100 1000 10000
Diameter (nm)
dN/dLogDp
(Particle/cm3)
0
20
40
60
80
100
Explained Variation (%)
Concentration Percentage
Soil/Road Dust
0
10
20
30
40
50
60
70
80
PM10-2.5
PM2.5
CO
NO
NO2
O3
BC
RH
Temp
WS
LDV
HDV
Species
Normalized Concentration

Figure 2.4: Volume size distributions along with the explained variation (%) of each factor
profile resolved by the PMF model.
Nucleation
0
5
10
15
20
25
10 100 1000 10000
Diameter (nm)
dV/dLogDp (µm3/cm3)
0
10
20
30
40
50
60
70
80
90
100
Explained variation (%)
Volume concentration (dV/dLogDp) Explained variation (%)

Traffic 1
0
5
10
15
20
25
10 100 1000 10000
Diameter (nm)
dV/dLogDp (µm3/cm3)
0
10
20
30
40
50
60
70
80
90
100
Explained variation (%)
Volume concentration (dV/dLogDp) Explained variation (%)

28
Traffic 2
0
5
10
15
20
25
10 100 1000 10000
Diameter (nm)
dV/dLogDp (µm3/cm3)
0
10
20
30
40
50
60
70
80
90
100
Explained variation (%)
Volume concentration (dV/dLogDp) Explained variation (%)

Urban Background Aerosol
0
5
10
15
20
25
10 100 1000 10000
Diameter (nm)
dV/dLogDp (µm3/cm3)
0
10
20
30
40
50
60
70
80
90
100
Explained variation (%)
Volume concentration (dV/dLogDp) Explained variation (%)

Secondary Aerosol
0
5
10
15
20
25
10 100 1000 10000
Diameter (nm)
dV/dLogDp (µm3/cm3)
0
10
20
30
40
50
60
70
80
90
100
Explained variation (%)
Volume concentration (dV/dLogDp) Explained variation (%)

Soil/Road Dust
0
5
10
15
20
25
10 100 1000 10000
Diameter (nm)
dV/dLogDp (µm3/cm3)
0
10
20
30
40
50
60
70
80
90
100
Explained variation (%)
Volume concentration (dV/dLogDp) Explained variation (%)

29
Figure 2.5: Relative contribution of each factor to the total number concentrations: a) overall
phases; b) cold phase; and c) warm phase.
Overall
Traffic2
27.5%
Nucleation
17.3%
Traffic1
39.9%
Secondary
Aerosol
2.1%
Soil/Road
Dust
1.1%
Urban
Background
12.2%

Cold Phase
Traffic2
25.0%
Nucleation
11.7%
Traffic1
43.4%
Secondary
Aerosol
2.5%
Soil/Road
Dust
0.2%
Urban
Background
17.1%

Warm Phase
Traffic2
27.6%
Nucleation
24.0%
Traffic1
33.2%
Urban
Background
7.4%
Soil/Road
Dust
6.3%
Secondary
Aerosol
1.5%

30
Figure 2.6: Contribution (particles/cm
3
) of each of the PMF-resolved factors to the total number
concentrations in the cold and warm phases.

31
Figure 2.7: Diurnal variations (geometric means) of number concentrations (particles /cm
3
) from
each factor resolved by the PMF model in the cold and warm phases. Error bars correspond to
one standard error.
Nucleation
0
200
400
600
800
1000
1200
1400
1600
1800
2000
0 1 2 3 4 5 6 7 8 9 1011121314151617181920212223
Hour of the Day
Number Concentration
(Particle/cm3)
Cold Phase Warm Phase

Traffic 1
0
1000
2000
3000
4000
5000
6000
7000
0 1 2 3 4 5 6 7 8 9 1011121314151617181920212223
Hour of the Day
Number Concentration
(Particle/cm3)
Cold Phase Warm Phase

Traffic 2
0
500
1000
1500
2000
2500
3000
0 1 2 3 4 5 6 7 8 9 1011121314151617181920212223
Hour of the Day
Number Concentration
(Particle/cm3)
Cold Phase Warm Phase

32
Urban Background Aerosol
0
500
1000
1500
2000
0 1 2 3 4 5 6 7 8 9 1011121314151617181920212223
Hour of the Day
Number Concentration
(Particle/cm3)
Cold Phase Warm Phase

Secondary Aerosol
0
50
100
150
200
250
300
0 1 2 3 4 5 6 7 8 9 10 1112 131415 161718 1920 212223
Hour of the Day
Number Concentration
(Particle/cm3)
Cold Phase Warm Phase

Soil/Road Dust
0
50
100
150
200
250
300
350
0 1 2 3 4 5 6 7 8 9 10 1112 131415 161718 1920 212223
Hour of the Day
Number Concentration
(Particle/cm3)
Cold Phase Warm Phase

33
Figure 2.8: Weekday/weekend analysis of each of the factors resolved by the PMF model (values
are geometric means). Error bars correspond to one standard error.
Nucleation
0
200
400
600
800
1000
1200
1400
1600
1800
2000
01234567 891011121314151617181920212223
Hour of the Day
Number Concentration
(Particle/cm3)
Weekday Weekend

Traffic 1
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
01234567 891011121314151617181920212223
Hour of the Day
Number Concentration
(Particle/cm3)
Weekday Weekend

Traffic 2
0
500
1000
1500
2000
2500
3000
01234567 891011121314151617181920212223
Hour of the Day
Number Concentration
(Particle/cm3)
Weekday Weekend

34
Urban Background Aerosol
0
200
400
600
800
1000
1200
1400
1600
01234567 891011121314151617181920212223
Hour of the Day
Number Concentration
(Particle/cm3)
Weekday Weekend

Secondary Aerosol
0
50
100
150
200
250
0 1 2 3 4 5 6 7 8 9 1011 121314 151617 181920 212223
Hour of the Day
Number Concentration
(Particle/cm3)
Weekday Weekend

Soil/Road Dust
0
20
40
60
80
100
120
140
0 1 2 3 4 5 6 7 8 9 1011 121314 151617 181920 212223
Hour of the Day
Number Concentration
(Particle/cm3)
Weekday Weekend

35

Table 2.3: Spearman correlation coefficient matrix indicating the association between the
auxiliary variables and the factors resolved by the PMF model. R values above 0.5 are bolded.

Species Nucleation Traffic 1 Traffic 2 Urban
Background
Aerosol
Secondary
Aerosol
Soil/Road
Dust
PM
10-2.5

0.17* 0.21* 0.35* 0.27* 0.09* 0.39*
PM
2.5
-0.24* -0.09* 0.05* 0.33* 0.69* 0.23*
CO 0.04 0.41* 0.58* 0.64* 0.17* 0.28*
NO -0.01 0.48* 0.59* 0.52* 0.27* 0.24*
NO
2
0.08* 0.50* 0.60* 0.57* 0.33* 0.14*
O
3
0.57* 0.34* 0.40* -0.35* 0.46* 0.19*
BC 0.01 0.53* 0.70* 0.71* 0.13* 0.22*
POC 0.09* 0.62* 0.28* 0.30* 0.24* 0.29*
SOC 0.46* 0.12* 0.43* 0.58* 0.46* 0.20*
EC 0.17* 0.47* 0.56* 0.60* 0.20* 0.17*
RH -0.26* -0.32* -0.30* -0.05* 0.43* 0.33*
Temp 0.52* -0.23* -0.18* -0.39* 0.34* 0.47*
WS 0.57* 0.00 0.07* -0.04* -0.25* 0.62*
LDV 0.22* 0.70* 0.42* 0.05* 0.01 0.02
HDV 0.23* 0.52 0.43* -0.08* -0.12* -0.01
* Indicates R values that are statistically significant (P<0.05).

36
Figure 2.9: Correlation between the measured vs. PMF-predicted total number concentrations
(particles/cm
3
) for the entire sampling period.
y = 0.99(±0.0007)x - 21.14(±6.99)
R
2
= 0.99
0
10000
20000
30000
40000
0 10000 20000 30000 40000
Measured
Predicted

2.3.3. Factor identification
Factor 1: Factor 1 has a number mode at <20 nm, a volume mode at <20 nm, and contributes
17.3% (11.7-24%) to the total number concentrations (Figures 3, 4, and 5). This factor has strong
positive (except for RH, with which this factor has negative correlation) associations with
temperature, wind speed, SOC, and O
3
(Table 3), which are also statistically significant (p<0.05).
These associations are also apparent from high loadings of temperature, RH, wind speed, and O
3

in the auxiliary variables profile (Figure 3). The contribution of this factor to the total number
concentration was also higher in the warm phase than in the cold phase, when higher
37
temperatures, wind speeds, and solar radiation are observed (Figure 1); this was the case both in
terms of percent contribution (24% in the warm phase vs. 11.7% in the cold phase) and number
concentration (589±25 particles/cm
3
in the cold phase vs. 1153±28 particles/cm
3
in the warm
phase) (Figures 5 and 6). The diurnal variations for this factor also revealed a sharp peak in the
afternoon (2-6 PM) (Figure 7), which coincides with very high temperatures, wind speeds, and
solar radiation as well as with minimum RH (Figure 1). A minor peak was also observed during
morning rush hours (6-8 am), which suggests the partial influence from traffic sources, as also
observed by loadings of HDV and LDV in this factor (Figure 3). However, there was no
significant distinction in the diurnal variation patterns of this factor in weekdays compared to
weekends (Figure 8).
The above characteristics are all typical of a "nucleation" factor, during which new particles are
formed via photochemical events under high temperatures, high wind speeds, and low RH
(Beddows et al., 2015; Brines et al., 2015; Dall'Osto et al., 2012; Vu et al., 2015). The minor
peak in the early morning can also be explained by the cooling, following dilution, of vehicular
exhaust emissions, which leads to the partitioning of semi-volatile exhaust gases into the particle
phase; this process is further enhanced by to the lower temperatures during that time of day
(Harrison et al. 2011; Ntziachristos et al. 2007; Janhall et al. 2004; Charron and Harrison, 2003).
Our findings are most specifically consistent with those of the study of Brines et al. (2015), in
which the authors had reported nucleation as one of the major sources of UFPs in five high-
insolation cities, including Los Angeles using the data obtained from the same sampling location.
They observed very similar diurnal variation for nucleation, with a minor peak in the early
morning and a major peak in early afternoon at the same sampling location in Los Angeles.

38
Factor 2: Factor 2 is mostly represented by particles at 20-40 nm and contributes about 40%
(33.2-43.4%) to the total number concentration (Figures 3 and 5). It also has a volume
concentration peak at around 30-40 nm (Figure 4). Judging by the loadings presented in the
auxiliary variables profile (Figures 3) and correlation coefficients presented in Table 3, this
factor has clear associations with gaseous pollutants (e.g., CO, NO, and NO
2
), BC, EC, and POC
(from the scenario containing EC/OC data, which themselves are indicators of vehicular
emissions (Gu et al., 2011; Ogulei et al., 2006b). In addition, high species loadings (Figure 3)
and correlation coefficients (Table 3) of LDV and HDV counts can also be observed for this
factor, indicating the influence of nearby passing traffic on this factor. The contribution of this
factor to the total number concentration was also much higher in the cold phase than in the warm
phase, when lower temperatures, wind speeds, and solar radiation (Figure 1) lead to increased
atmospheric stability and lower mixing height (Hasheminassab et al., 2014a); this was the case
both in terms of percent contribution (43.4% in the cold phase vs. 33.2% in the warm phase) and
number concentration (3166±66 particles/cm
3
in the cold phase vs. 1201±61 particles/cm
3
in the
warm phase) (Figures 5 and 6). The diurnal variations also revealed a distinctive pattern peaking
in the morning rush hours (around 7-8 AM) (Fig. 7). The weekday/weekend analysis also
indicated that this factor had higher contributions during the weekdays compared to the
weekends (Figure 8). Therefore, this factor can be attributed to traffic tailpipe emissions.
Previous source apportionment studies on number size distributions have also associated such
characteristics with fresh vehicular emissions (Beddows et al., 2015; Dall'Osto et al., 2012; Vu et
al., 2015). This factor is denoted as “traffic 1”, given that another factor attributed to traffic
emissions was resolved, which will be discussed in the following section. The characteristics of
39
this traffic factor are in agreement with what Brines et al. (2015) reported for five high-insolation
cities, including Los Angeles.

Factor 3: This factor has a major peak in the Aitken mode (60-100 nm) and contributes 27.5%
(25-27.6%) to the total number concentration (Figures 3 and 5). It also exhibited a volume
concentration peak at around 100 nm (Figure 4). Judging by the loadings presented in the
auxiliary variables profile (Figure 3) and correlation coefficients presented in Table 3, significant
associations can be observed between this factor and gaseous pollutants (e.g., CO, NO, and
NO
2
), as well as with BC (and EC from the scenario containing EC/OC data). Although weaker
than those of Factor 2, there are significant positive associations between this factor and LDV
and HDV counts (Figure 3 and Table 3), suggesting the likely influence of nearby passing traffic.
This factor also had a significantly higher contribution to the total number concentrations in the
cold phase than in the warm phase (an average of 1755±56 particles/cm
3
in the cold phase vs.
1059±43 particles/cm
3
in the warm phase (Figure 6)), in spite of the fact that its percent
contribution to the total PM number concentrations was comparable in both phases, and slightly
higher in the warm phase (25% in the cold phase vs. 27.6% in the warm phase (Figure 5)). This
is due mainly to the fact that the contribution of the "traffic 1" factor is so large in the cold phase
that has significantly obscured the percent contribution of other factors in this phase, even
though their absolute contributions in terms of total number concentrations were higher in the
cold phase.
The diurnal variations for this factor also indicated clear peaks during the morning rush hours (6-
8 am) in both phases, along with another peak at late night during the cold phase, most likely due
to the stagnant atmospheric conditions during this time of the year, which traps the emissions in
40
lower altitudes (Figure 7). This diurnal profile suggests a major contribution from semi-volatile
compounds in the atmosphere, particularly in the cold phase, as reflected in the substantial
increase at nighttime. The weekday/weekend analysis for this factor revealed larger contributions
in weekdays than in weekends, especially during daytime hours. The slightly higher nighttime
contribution of this factor in the weekends compared to the weekdays can be attributed to the
larger number of within-city travels being made on holiday nights. These levels and trends are,
overall, suggestive of emissions from vehicular sources. However, the larger size range of this
factor compared to factor 2, combined with the involvement of EC and SOC (as observed from
the scenario containing EC/OC data) as well as BC, suggest that although this factor also
originates from “traffic”, the particles are “older” (i.e., more aged) than those observed in factor
2 and are mostly in the Aitken and Accumulation modes; therefore, it was labeled as "traffic 2".
This finding is also consistent with those of the previous studies (e.g., Brines et al. (2015)), in
which the authors detected distinct traffic factors (with a collective relative contribution of
approximately 60% in Los Angeles at the same sampling site) using a different source
apportionment method, named k-means cluster analysis. It should be noted that it is quite
common in source apportionment studies performed on size-segregated PM number
concentrations to detect more than one traffic factors, due primarily to the fact that particle sizes
may change, as particles undergo processes including agglomeration as well as evaporation or
condensation of semi-volatile species from- or onto their surface following their release in the
atmosphere (Harrison et al., 2016; Kim et al., 2004; Zhou et al., 2005).
It is also noteworthy that the traffic 2 factor has a slightly higher HDV loading than traffic 1
factor (Figure 3). It also has a somewhat stronger positive correlation with HDV (R=0.43) than
with LDV (R=0.41), while the traffic 1 factor has a stronger correlation with LDV (R=0.69) than
41
with HDV (R=0.52). Additionally, the stronger correlation of traffic 2 factor with EC and BC
compared to traffic 1 leads us to the hypothesis that HDV might be contributing more to this
factor than LDV is. Vu et al. (2015) have also suggested that observing a number concentration
mode at the size range of 60-100 nm can be a result of incomplete combustion of diesel fuel,
consisting of pyrolytic EC and OC. Other studies have also found two particle modes, or factors,
for traffic-related emissions. Although the emissions in both of these two modes are believed to
come from the same fleet of vehicles, they have different formation mechanisms and chemistry,
with particles associated with the second mode (i.e., soot mode) assumed to have an elemental
carbon core. This is consistent with the findings of the present study, judging by the mode
diameter and high loading of BC, EC, and OC in the traffic 2 factor (Figure 3) and the strong
correlation of this factor with BC, EC, and OC (Table 3). Additionally, studies have indicated
that a fraction of diesel PM emissions, which is generally in the range of 50-200 nm, comprises
particles that have an elemental core, with low-vapor-pressure hydrocarbons and sulfur
compounds being adsorbed on their surface (Burtscher, 2005). Therefore, it might be likely that
this factor is representing a higher contribution of HDV emissions, although stronger evidence is
required to confirm this hypothesis.

Factor 4: Factor 4, which contributes 12.2% (7.4-17.1%) to the total number concentration, is
represented by a number mode at around 220 nm and a volume mode at around 250 nm (Figures
3, 4, and 5). The profile for the auxiliary variables also indicates high loadings for gaseous
pollutants (e.g., CO, NO, and NO
2
) and BC (Figure 3) as well as for EC and SOC (when the
PMF model was run with the EC/OC data). The large correlation coefficients of this factor with
the aforementioned species also confirm its strong association with these parameters (Table 3).
42
The lower-than-unity NO/NO
2
ratio for this factor also suggests that these particles are aged
compared to the newly formed particles (Liu et al., 2014). This is also supported by the stronger
positive correlation of this factor with SOC than with POC, suggesting the fact that this factor is
not coming from direct emissions and has most likely undergone processes and reactions in the
atmosphere. As can be inferred from Figures 5 and 6, the contribution of this factor is
significantly higher in the cold phase than in the warm phase, both in terms of percent
contribution (17.1% in the cold phase vs. 7.4% in the warm phase) and the absolute contribution
to the total number concentration (1200±41 particles/cm
3
in the cold phase vs. 284±23
particles/cm
3
in the warm phase). As seen in Figure 7, the diurnal variations for this factor also
exhibit a clear peak during morning hours, which indicates higher concentrations when
atmosphere is more stable and wind speeds are low, especially in the cold phase when these
conditions are even more intense (Figure 1). The weekday/weekend analysis also revealed a
slightly elevated contribution of this factor to the total number concentrations during morning
rush hours, especially during the weekdays, suggesting the small influence of traffic emissions
on this factor. Previous studies have indicated that these are characteristics of the “urban
background aerosol”, as suggested by (Beddows et al., 2015; Dall'Osto et al., 2012).

Factor 5: Factor 5 has a number and volume mode at around 500 nm and a minor number mode
at 50 nm (looking at the black dots, representing the explained variations) (Figures 3 and 4). This
factor contributes 2.1% (1.5-2.5%) to the total number concentration (Figure 5). It is also
associated with high loadings of PM
2.5
mass concentration (i.e., major contributor to PM
2.5

mass), NO, NO
2
, Temperature, RH (Figure 3), and SOC (as observed from the scenario
containing EC/OC data). This is also supported by the results of the correlation analysis
43
presented in Table 3, indicating that this factor has strong positive correlations with PM
2.5
, NO,
NO
2
, Temperature, RH, and SOC. The overall small contribution of this factor to the total
number concentration was slightly higher in the cold phase than in the warm phase; this was the
case both in terms of percent contribution (2.5% in the cold phase vs. 1.5% in the warm phase)
and number concentration (111±11 particles/cm
3
in the cold phase vs. 100±5 particles/cm
3
in the
warm phase) (Figures 5 and 6). The diurnal variation for this factor also reveals a significant
increase during nighttime, especially during the cold phase (Figure 7). However, the
weekday/weekend analysis did not reveal any distinctive trend pertaining to the day of the week
for this factor (Figure 8). These pieces of evidence point to “secondary aerosols” as the most
appropriate title for this factor, which is consistent with the results of previous PMF studies both
on number size distributions and chemical speciation data (Beddows et al., 2015; Hasheminassab
et al., 2014a). Table 3 indicates a much higher correlation of this factor with SOC than POC (R
values of 0.5 and 0.2, respectively). The association of this factor with RH and temperature,
along with its higher contribution to particle number during the cold phase, particularly at night,
support the hypothesis that this factor likely represents the fraction of aerosols produced by
secondary reactions on a regional scale, including ammonium nitrate (whose partitioning in the
PM phase increases with decreasing temperature and increased RH), but also secondary organic
aerosols from nighttime and/or aqueous phase reactions, as indicated in earlier studies in this
area (Hersey et al., 2011; Venkatachari et al., 2005). In a previous source apportionment study on
PM
2.5
chemical speciation data in downtown Los Angeles, we also found a similar factor profile,
representing a mixture of secondary components (dominated by secondary nitrate and SOC) with
higher contribution during the cold season (Hasheminassab et al., 2014a). Moreover, previous
studies have shown that secondary organic aerosol formed at nighttime together with ammonium
44
nitrate are major contributors to the mass concentrations of PM
2.5
, which was also observed in
the present work from the high loading of PM
2.5
mass concentration in this profile (Figure 3)
(Arhami et al., 2010; Hasheminassab et al., 2014a; Saffari et al., 2016).

Factor 6: Factor 6 is dominated by particles at around 1 µm and above (Figure 3). This factor
also had a volume mode at > 1 µm (Figure 4). Although this factor contributes only 1.1% (0.2-
6.3%) to the total number concentration (Figure 5), it is associated with high loadings of coarse
PM and PM
2.5
(great contributor to mass) (Figure 3). In addition, high loadings of temperature
and wind speed were observed for this factor (Figure 3). Table 3 also indicates strong correlation
of this factor with coarse PM, PM
2.5
, temperature, and wind speed. The contribution of this factor
to the total number concentration was also higher in the warm phase than in the cold phase, both
in terms of percent contribution (0.2% in the cold phase vs. 6.3% in the warm phase) and number
concentration (14±1 particles/cm
3
in the cold phase vs. 243±3 particles/cm
3
in the warm phase)
(Figures 5 and 6). The diurnal variations for this factor exhibited significantly higher
contributions during daytime, especially in the warm phase (Figure 7), when atmosphere is
unstable, wind speed is high, and the mixing height is at its maximum (Figure 1). However, the
weekday/weekend analysis did not reveal any distinctive trend pertaining to the day of the week
for this factor (Figure 8). Based on all of the abovementioned characteristics, this factor was
named “soil/road dust” (Gietl et al., 2010; Harrison and Booker, 2001; Harrison et al., 2012).
This is also quite consistent with the findings of the study of Hasheminassab et al. (2014a), in
which the authors apportioned the sources of ambient fine particulate matter across the state of
California. In that study, the authors observed a lower contribution of the soil factor to particle
mass concentrations in the northern regions of the state of California, mainly because of higher
45
RH and increased precipitation that inhibit the re-suspension of soil due to strong winds
(Harrison and Booker, 2001). In the present study, similarly, the contribution of this factor was
higher in the warm phase, when higher temperatures and wind speeds facilitate the re-suspension
of soil and dust (Figure 1).

2.4. Summary and conclusions
The present study was the first attempt to characterize major sources of PM number
concentrations and quantify their contributions using the PMF receptor model applied on PM
number size distributions in the range of 13 nm to 10 µm combined with several auxiliary
variables, including BC, EC/OC, PM mass, gaseous pollutants, meteorological, and traffic flow
data, in central Los Angeles. The six-factor solution was found to be the most physically
applicable solution for the input data: nucleation, traffic 1, traffic 2, urban background aerosol,
secondary aerosol, and soil. Traffic sources (1 and 2) were the major contributor to PM number
concentrations, making up to above 60% of the total number concentrations combined, with
larger contributions in the cold phase compared to the warm phase, when lower temperatures,
wind speeds, and solar radiation lead to increased atmospheric stability and lower mixing height.
The contribution of traffic factors was largest during morning and afternoon rush hours; it was
also higher in the weekdays compared to the weekends, as expected. In agreement with the
findings of previous studies in Los Angeles, nucleation was another major factor contributing to
the total number concentrations (17%), having a larger contribution in the warm phase than in
the cold phase. The diurnal variations for this factor also revealed a sharp peak in the afternoon
(2-6 PM), which coincides with high temperatures, wind speeds, and solar radiation as well as
with minimum RH, providing ideal conditions for the occurrence of photochemical nucleation
46
processes, especially during warmer seasons. Urban background aerosol, secondary aerosol, and
soil, with relative contributions of approximately 12%, 2.1%, and 1.1%, respectively, overall
accounted for approximately 15% of PM number concentrations. However, these factors
dominated the PM volume and mass concentrations, due mainly to their larger mode diameters.

47
Chapter 3:
Enhanced toxicity of aerosol in fog conditions in the Po Valley, Italy
While numerous studies have demonstrated the association between outdoor exposure to atmospheric
particulate matter (PM) and adverse health effects, the actual chemical species responsible for PM
toxicological properties remain a subject of investigation. We provide here reactive oxygen species
(ROS) activity data for PM samples collected at a rural site in the Po Valley, Italy, during the fog
season (i.e., November-March). We show that the intrinsic ROS activity of Po Valley PM, which is
mainly composed of biomass burning and secondary aerosols, is comparable to that of traffic-related
particles in urban areas. The airborne concentration of PM components responsible for the ROS
activity decreases in fog conditions, when water-soluble species are scavenged within the droplets.
Thanks to this partitioning effect of fog, the measured ROS activity of fog water was contributed
mainly by water-soluble organic carbon (WSOC) and secondary inorganic ions rather than by
transition metals. We found that the intrinsic ROS activity of fog droplets is even greater (> 2.5 times)
than that of the PM on which droplets are formed, indicating that redox-active compounds are not
only scavenged from the particulate phase, but are also produced within the droplets. Therefore, even
if fog formation exerts a scavenging effect on PM mass and redox-active compounds, the aqueous-
phase formation of reactive secondary organic compounds can eventually enhance ROS activity of
PM when fog evaporates. These findings, based on a case study during a field campaign in November
2015, indicate that a significant portion of airborne toxicity in the Po Valley is largely produced by
environmental conditions (fog formation and fog processing) and not simply by the emission and
transport of pollutants.

This chapter is based on the following publication:

Decesari, S., Sowlat, M.H., Hasheminassab, S., Sandrini, S., Gilardoni, S., Facchini, M.C., Fuzzi, S.,
Sioutas, C., 2017. Enhanced toxicity of aerosol in fog conditions in the Po Valley, Italy. Atmospheric
Chemistry and Physics 17, 7721-7731.

3.1. Introduction
There is a rapidly growing body of epidemiological evidence identifying major health impacts
associated with population exposure to airborne particulate matter (PM), including, but not
limited to, respiratory and cardiovascular diseases, as well as neurodegenerative effects (Pope et
al., 2002); (Pope et al., 2004a); (Davis et al., 2013c; Dockery and Stone, 2007; Gauderman et al.,
2015). Even if the international air quality standards for atmospheric PM are based on mass
concentrations (PM10 and PM2.5 for particles of diameter below 10 or 2.5 µm, respectively), the
WHO acknowledges that it is likely that not every PM component is equally important in
48
causing these health effects (WHO, 2007, 2013). Research on traffic-related PM has provided a
first epidemiological evidence of the links between adverse health effects and PM chemical
composition (Bates et al., 2015; Janssen et al., 2011), in line with the results of several
toxicological studies (Nel, 2005); (Cassee et al., 2013); (Fang et al., 2015b), while outside urban
areas, the characterization of PM and associated health effects is sparse. During transport in the
atmosphere, particles emitted by traffic are progressively diluted and eventually transformed by
chemical processes that enrich them in secondary components (Crippa et al., 2014; Zhang et al.,
2007). Moreover, the oxidation of reactive volatile organic compounds (VOCs) supplies a range
of water-soluble compounds (e.g., formaldehyde, glyoxal), partitioning into the particulate phase
at high relative humidities (deliquescent aerosols and cloud/fog droplets), where they can be
further oxidized into new secondary organic compounds (so called “aqueous SOA”) (Ervens et
al., 2011). The current understanding of the health impacts of PM secondary components is
certainly much more limited than for traffic-related aerosols. Although recent chamber studies
have indicated increased oxidative potential of PM emitted from a variety of sources (e.g.,
combustion or biomass burning) after undergoing oxidation and/or secondary formation
(Antinolo et al., 2015; McWhinney et al., 2013; McWhinney et al., 2011), there is a lack of direct
field toxicological observations, which is partly caused by the difficulty of disentangling the
secondary PM fraction from other oxidized organic material of primary nature in field
conditions. In this study, we investigate the toxicological properties of PM at a polluted rural
site, where the presence of fog is responsible for aerosol scavenging and deposition but also for
secondary aerosol formation through aqueous-phase processes. Fog and low-level clouds are
transient phenomena in the atmosphere but their occurrence can be high in wintertime in certain
areas of the globe, including many highly-populated sites enclosed in orographic basins.
49
(Cermak et al., 2009) showed that several pollution hotspots in Europe, including Benelux, the
Ruhr district, the basins of Paris and London and the Po Valley, experience low-level clouds and
fogs for 35% to 60% of the days in winter months. The fraction of fog days in fall/winter in the
Californian Central Valley (6.5 millions inhabitants) is ca. 20% according to (Baldocchi and
Waller, 2014). Fog frequencies of ca. 10% in winter are also characteristic of the Yangtze River
corridor (Niu et al., 2010), and even greater values (20% to more than 35%) are typical of the
Indo-Gangetic plain (Saraf et al., 2011). All these regions of the globe commonly experience PM
pollution peaks in winter months, during the fog season. In this season of the year, the same
stable weather conditions favor the accumulation of air pollutants and fog formation. Therefore,
fog-processing is potentially a major driver for secondary aerosol formation in wintertime at all
these sites. In (Gilardoni et al., 2016), we provided a first estimate of secondary organic aerosols
produced by aqueous-phase processing of smoke particles in Europe: 0.1 to 0.5 Tg of organic
carbon per year, corresponding to 4 – 20% of total primary organic aerosol emissions in the
region.
In the present study, aerosol chemical and toxicological measurements were carried out in the Po
Valley, Italy, where radiation fogs frequently occur during the cold season (up to 25% of the
time in fall-winter months in rural areas, according to recent studies (Giulianelli et al., 2014).
With approximately 20 million inhabitants (~30% of the Italian population), the Po Valley area
has the highest population density across the country, and ranks among the top European
“pollution hotspots” in terms of mortality attributable to PM exposure (Kiesewetter and Amann,
2014). We provide here the first measurements of the redox activity (as a proxy for toxicological
potential) of fog water and interstitial (i.e., unscavenged) aerosols. The toxicological assay used
in this study is capable of quantifying the oxidative potential associated with ambient PM,
50
initiated via generation of reactive oxygen species (ROS) due to the interaction of target cells
with redox-active components of ambient PM (Landreman et al., 2008); (Daher et al., 2012;
Verma et al., 2012b). We show that ROS, in the rural Po Valley PM, occur in concentrations
comparable to that of a PM in a megacity, and that the ROS levels are even amplified in fog
water with respect to the PM fraction scavenged within fog droplets. Results from the present
study provide a basis for prospective epidemiological programs to evaluate how fog
scavenging/processing of PM impacts on human health.

3.2. Methods
3.2.1 Measurement site
Fog samples were collected from 30 November to 30 December 2015 at the meteorological
station Giorgio Fea in San Pietro Capofiume (44°39′15″ latitude, 11°37′29″ longitude), a rural
site located 30 km northeast of Bologna (Italy) in the eastern part of the Po Valley (northern
Italy). From 30 November to 4 December, an intensive observation period was scheduled, with
the concurrent sampling of fog and aerosol samples and the deployment of a HR-ToF-AMS
(Aerodyne Research) for online aerosol measurements. During the sampling campaign, a total of
6 aerosol samples and 16 fog samples were collected. Additionally, fog samples collected after
the intensive observation period (i.e., after 4 December) were pooled in groups of two or three
for the analysis of metals and oxidative potential.

51
3.2.2 Aerosol and fog-water sampling
Aerosol samples were collected by a PM1 Tecora Echo High Volume sampler equipped with a
PM1 Digitel sampling head, operating at 500 L min
-1
flow. One daytime and one nighttime
sample were collected every day from 9:00 to 18:00 LT and from 18:00 to 9:00 LT respectively,
on prewashed and prebaked Pall quartz fiber filters. Samples were wrapped with aluminum foil,
zipped in plastic bags, and stored in freezer at −20°C until analysis. Fog samples were collected
using an automated, computer driven active string collector described in (Fuzzi et al., 1997). A
Particulate Volume Monitor PVM-100 (Gerber, 1991), used to determine fog liquid water
content (LWC) at 1 min time resolution, was used to activate the string collector. The collector
and its strings were extensively cleaned at the beginning of the fog season, two weeks before the
samples object of this study were collected. In particular, the stainless-steel strings were carried
to the CNR-ISAC chemical laboratory in Bologna where they were gently brushed to remove
any stuck residues from fog sampling in the previous winter and washed with milliQ water in an
ultrasonic bath. Finally they were brought back to the field and mounted on the fog collector
before first the automatic sampling system was switched on around mid-November. A LWC
threshold of 0.08 g m
‐3
was chosen for the activation as an indicator of fog presence, which
roughly corresponds to a 200 m visibility. Concentrations of analytes in fog samples, expressed
in µg mL
-1
, were converted into µg m
-3
by multiplying with the fog liquid volume (mL m
-3
). The
latter was not estimated by the mass of sampled fog water multiplied by the flux of the fog
collector, because the collection efficiency is typically much smaller than 1 (about 40%,
according to Fuzzi et al. 1997). We used instead the liquid water content (LWC) measured by the
PVM-100 and averaged over the sampling time of the fog collector to multiply by the
52
concentrations of chemicals in fog water to get air equivalent concentrations (µg m
-3
) of the fog
components.

3.2.3 Chemical analysis of water-soluble components
Prior to chemical analysis a quarter of each aerosol quartz-fibre filter was extracted with 10 mL
of 18-MΩ Milli-Q water by sonication for 30 minutes. Liquid extracts were filtered to remove
quartz residues, then analyzed for Water Soluble Organic Carbon (WSOC) with a total organic
carbon analyzer (Shimadzu TOC-5000A). Fog samples were filtered on 47 mm quartz-fibre
filters (Whatman, QMA grade) within few hours after collection to remove suspended
particulate. Conductivity and pH measurements were carried out immediately, then samples were
stored frozen until further analysis. WSOC was determined in the same way as for aerosol filter
extracts, in addition the ionic composition of fog was determined by two Thermofisher
ICS_2000 ion chromatographs equipped with an IonPac AS11 2x250 mm separation column for
anions and IonPac CS16 3x250 mm Dionex separation column, self-regenerating suppressors
and KOH and MSA as eluents for anions and cations respectively. In both cases a gradient
elution allowed the separation and detection of both inorganic and organic anions and cations.

3.2.4 Total elemental analysis of aerosol and fog samples
The magnetic sector inductively coupled plasma mass spectrometry (SF-ICP-MS; Thermo-
Ginnigan Element 2) unit was used for the total elemental analysis of the aerosol and fog
samples. Detailed information about the procedure can be found elsewhere (Zhang et al., 2008)
(Okuda et al., 2014). Briefly, the samples are first digested in an automated microwave-aided
53
(oven-aided in case of fog samples) digestion system (Milestone ETHOS+) using a mixture of
ultra-high-purity acids (1.0 mL of 16 M nitric acid and 0.25 mL of 12 M hydrochloric acid).
Then, the digestates are analyzed by SF-ICPMS after being diluted to 15 mL with high-purity
water (18 mΩ). A combination of the SF-ICPMS instrumental analysis, method blanks, and
digestion recoveries were used to estimate the analytical uncertainties. A total of 49 elements
were determined using this method. Further details of the analytical method can be found in
(Zhang et al., 2008) and (Okuda et al., 2014). The full chemical analyses (WSOC, ion
chromatography, metals) was performed on 6 aerosol samples and 7 fog samples.

3.2.5 ROS analysis
The ROS assay, an in vitro exposure assay of PM extracts to rat alveolar macrophage cells (cell
line NR8383)(Landreman et al., 2008), was used as a measure of PM toxicity of the samples. In
this method, samples (only filters in this step) were first extracted using an initial sonication
period for 15 min with high-purity 10 Milli-Q (18 mΩ) water. The samples are then continuously
agitated in the dark at room temperature for 16 h, followed by another 15-min sonication and 1-
min agitation by a vortex mixer. A portion of the sample suspension was passed through 0.22 µm
polypropylene syringe filters (referred to as filtered samples, representing WS components),
while the other portion of the sample suspension underwent the ROS assay without filtration
(referred to as the unfiltered samples)(Shafer et al., 2016). ROS activity analysis was carried out
on 20 samples (= 6 aerosol + 7 unfiltered fog + 7 filtered fog samples).
In the in vitro exposure and ROS detection step, the membrane-permeable 2′,7′-dichlorodihydro-
fluorescein diacetate (DCFH-DA) probe gets deacetylated by cytoplasmic esterases to 2′,7′-
dichlorodihydro- fluorescein (DCFH) upon entering the macrophage cell. In the presence of
54
oxidizing species (for example, the ROS species generated due to exposure to toxic PM
components), DCFH is converted to its fluorescent form, which is DCF. In the next step, an M5e
microplate reader (Molecular Devices, CA, USA) is used to determine the fluorescence intensity
of each cell after the exposure at 488 nm excitation and 530 nm emission. The raw fluorescence
data were corrected using blanks and normalized using Zymosan (ZYM) positive controls (i.e.,
β-1,3 polysaccharide of D-glucose from Sigma Aldrich, MO, USA) (Shafer et al., 2016). Filter
blanks were used to account for the potential impact of the biologically active quartz filters on
the ROS results. The ROS activity of filter blanks (N = 2) resulted to be of the order of 280  13
fluorescence units (FU) for unfiltered extracts and of 27  23 FU for filtered extracts. For
comparison, the ROS activity of PM1 filter samples was as high as 3896  350 FU and 1518 
118 FU for unfiltered and filtered extracts, respectively. Following subtraction of the
contribution of blanks, the overall per-mass oxidative potential (representing the inherent
toxicity) of samples was reported in units of Zymosan equivalents. The per-volume ROS activity
(i.e., normalized by the volume of air samples) of samples was also calculated via multiplying
the mass-based ROS by the PM mass concentration of a given sample. This provides a measure
of the actual toxicity of the air inhaled by individuals, which is a very important parameter in
exposure assessment studies (Wittkopp et al., 2013) (Zhang et al., 2007) (Delfino et al., 2010a).

3.2.6 Statistical analyses
Spearman rank correlation was used to explore the association of individual components of the
aerosol and fog with the ROS activity. We also applied the principal component analysis (PCA)
to the ambient concentrations of the chemical components of the aerosols in order to identify the
55
source factors that contribute to PM levels and ROS activity. In this analysis, the VARIMAX-
normalized rotation approach was employed to identify the uncorrelated source factors (Henry,
1987). Additionally, we also applied the multi-linear regression (MLR) analysis to identify the
source factors (as represented by the relevant species) that mostly contribute to ROS activity of
aerosols and fog water. In this analysis, several combinations of species with high loadings in the
PCA were regressed against the ROS levels, and the ones leading to the highest R2 value were
kept in the final solution. The statistical significance was evaluated at both P values of <0.1 and
<0.05.

3.3. Results
3.3.1. Partitioning of PM mass and chemical components between fog droplets and interstitial
aerosol
Figure 1 indicates the mass concentrations of fog water and daytime/interstitial aerosols in the Po
Valley area during the study period (Nov-Dec 2015). The results pertain to concurrent aerosol
and fog samples. As can be seen in the figure, daytime (i.e., out-of-fog) aerosol exhibited the
highest mass concentrations (35.9±17.3 μg/m3), whereas the average mass concentration of the
interstitial aerosol was 15.7±6.3 μg/m3. The average mass concentration of fog water (soluble
and insoluble) components was also slightly higher than that of interstitial aerosols (i.e., ~17
μg/m3). These results are in agreement with those previously observed in the area: for example,
Gilardoni et al. (2014) reported an average PM1 concentration of 32 μg/m3 under clear
conditions (i.e., during the day), while the nighttime aerosol concentrations were as low as 10
μg/m3, indicating that the average interstitial aerosol concentration during fog episodes was 66%
lower than that of daytime aerosols as a result of fog scavenging. Under the assumption that PM
56
sources and mixing height are constant, the fog “partitioning” acts by simply transferring mass
(and selectively, chemical compounds) to and from the aerosol and the fog phase, and this is
supported by the results shown in Figure 1, in which the sum of mass concentrations of fog water
and interstitial aerosols is almost equal to the mass concentrations of daytime aerosols.

Figure 3.1: Mass concentrations (µg/m
3
) of the Po Valley fog water and aerosol during the study
period. Bars represent geometric means and error bars correspond to one standard error (SD).
0
5
10
15
20
25
30
35
40
45
PM Mass concentration (µg/m
3
)
Daytime aerosol Nighttime aerosol Fog

Figure 2(a,b) illustrates the concentrations and mass fractions of WS components (inorganic and
organic ionic species) in fog water and aerosol samples. As shown in the figure, the
concentrations of WS components were significantly higher in daytime and fog water samples
than in interstitial (or nighttime) aerosols, confirming that WS components from atmospheric
aerosols partitioned to a great extent into fog water during fog episodes, reaching scavenging
57
rates of 80% in the case of nitrate, chloride, acetate and methanesulfonate (MSA). In addition,
the excess mass concentrations of certain ionic components (chloride, ammonia and especially
acetate and formate) in fog samples compared to daytime PM1 (Figure 2 (a)) cannot be explained
by a simple partitioning process, and implies the enrichment of such compounds in fog water
from the gas-phase followed by aqueous-phase reactions (hydration, acid-base reactions,
oxidation). Finally, as shown in the mass fraction chart (Figure 2(b)), the levels of WS
components are highest in the fog water, in certain cases (e.g., acetate, formate, and ammonia)
much higher than those of daytime aerosols.
Figure 3(a, b) illustrate the airborne concentrations (ng/m
3
) and mass fractions (ng/μg) of the
metals/elements in fog water and aerosol samples. The results pertain to parallel aerosol and fog
samples. Our results are comparable with those of previous studies in the Po Valley area in terms
of the concentrations of metals/elements in the aerosol samples (Mancinelli et al., 2005;
Canepari et al., 2014; Perrino et al., 2014). For species with usually high WS fractions (including
Ca and Mg), mass fractions were higher in fog water than within interstitial aerosol. However,
mass fractions of combustion-related species (including V, Mn, Fe, Cu, and Pb) were greater in
the interstitial aerosol with respect to fog water. Their scavenging rates (daytime minus nighttime
concentrations) are lower than 60% and much smaller than those of WS ionic compounds
(Figure 2). Therefore, combustion-related metals are enriched in aerosol particles that are poor
condensation nuclei and, therefore, remain in the particle phase during fog episodes, while the
particles carrying the WS components are efficiently scavenged into fog water (Gilardoni et al.,
2014).

58
Figure 3.2: Concentrations (a), and mass fractions (b), of the water-soluble components in the
fog water and aerosols. Bars represent geometric means and error bars correspond to one
standard error (SE).
a)
0.001
0.01
0.1
1
10
Concentration (µg/m
3
)
Daytime aerosol Nighttime aerosol Fog

b)
0.0001
0.001
0.01
0.1
1
Mass fraction (µg/µg)

59
Figure 3.3: Airborne concentrations (ng/µg) (a) and mass fractions (ng/m
3
) (b) of
metals/elements in fog water and aerosol samples collected in the Po Valley area in Fall 2015.
Bars represent geometric means and error bars correspond to one standard error (SD).
a)
0.0001
0.001
0.01
0.1
1
10
100
1000
Na
Mg
Al
Ca
Ti
V
Cr
Mn
Fe
Cu
Zn
As
Cd
Ba
Pb
Th
Metal concentration (ng/m
3
)
Daytime aerosol Nighttime aerosol Fog

b)
0.00001
0.0001
0.001
0.01
0.1
1
10
100
Na
Mg
Al
Ca
Ti
V
Cr
Mn
Fe
Cu
Zn
As
Cd
Ba
Pb
Th
Metal mass fraction (ng/µg)

60
3.3.2. ROS activity of fog water and aerosol samples
When fog forms, aerosol particles are selectively scavenged into fog droplets, with the smaller
and more hydrophobic particles left unscavenged as interstitial aerosol. This process tends to be
selective with respect to specific chemical components (“partitioning”) and represents a useful
way to study the toxicological properties of externally-mixed PM components under field
conditions. Past studies in the Po Valley (Facchini et al., 1999; Gilardoni et al., 2014) highlighted
that fog scavenging efficiency for different chemical components of atmospheric particles is
related to their water solubility, with higher scavenging efficiencies for water-soluble (WS)
species (e.g., inorganic ions and water-soluble organic carbon, WSOC) and lower scavenging
efficiencies for hydrophobic compounds (e.g., elemental carbon and several metals) (Gilardoni et
al., 2014). Therefore, during fog episodes, fog droplets are enriched in WS components, while
water-insoluble (WI) species dominantly partition into interstitial aerosols (Hallberg et al., 1992)
(Fuzzi et al., 1988). Results from the present study are in good agreement with the published
literature on the effect of fog scavenging on the aerosol properties: Figures 2 and 3 indicated that
organic as well as inorganic WS components (e.g. MSA, oxalate, nitrate, and sulfate) and metals
with high WS fractions (including Ca, Na, and Mg) have the highest scavenging rate, whereas
other combustion-related species with lower solubility (i.e. Fe, Cr, and Zn) are scavenged much
less efficiently. Metals can be also scavenged by fog (Mancinelli et al., 2005) but not to the same
extent as secondary organic and inorganic species, resulting in higher metal enrichment of the
interstitial aerosol. The results on partitioning of WSOC between interstitial aerosols and fog
water are presented in Figure 4. Contrary to the chemical species discussed above, WSOC is a
complex mixture of chemical species originating from a variety of sources. According to
previous studies in the region (Gilardoni et al., 2014; 2016), oxidized particulate organic
61
compounds include biomass burning products as well as secondary organic species, which also
account for the products of transformation of biomass burning compounds upon fog processing.
As can be seen in Figure 4(b), the mass fraction of WSOC in PM1 was comparable between
daytime and interstitial aerosols, analogously to what is observed for some of the WS ionic
species, like sulfate and oxalate, and contrary to others (e.g., nitrate and ammonium) which are
considerably depleted in interstitial particles with respect to daytime aerosol. However, the
airborne concentration of WSOC in daytime aerosol was approximately two times higher than
that of interstitial aerosols (Figure 1(a)). Additionally, the concentration of WSOC was much
higher in the fog water than in daytime and nighttime aerosols, both on a per-mass and per-
volume basis.
We present here for the first time results on fog scavenging effects on ROS activity of the
aerosol. The average ROS activity of daytime (i.e, without fog) and nighttime (in-fog as well as
interstitial) aerosol bulk extracts and for fog water is reported in Figure 5. On a per m
3
of air
volume basis, the ROS activity of fog water and daytime aerosol are quite comparable and about
four times greater than that of interstitial aerosol. However, per PM mass ROS activities of the
daytime and interstitial aerosols are very similar (1877±849 and 1870±1229 microgram
Zymosan/mg of PM, respectively), while the ROS activity of fog water was considerably higher
(5194±610 MicrogramZymosan/mg of PM) than that of both types of aerosols. The variability of
ROS activity of PM in the Po Valley in the fog season is therefore complex. The extrinsic (on a
per-volume basis) ROS activity in PM decreases upon fog scavenging similarly to total PM1
mass concentrations (-70% and -66%, respectively) causing the toxicity of interstitial particles
being comparable to that of the original aerosol population. In contrast, the ROS activity of fog
water is clearly higher than the scavenged fraction of ROS activity of PM (calculated as the
62
difference between daytime and interstitial PM activity). This trend reflects very closely that of
WSOC levels (Figure 4). The excess of WSOC mass in fog water with respect to daytime aerosol
indicates that organic solutes in fog droplets do not originate simply from aerosol scavenging,
and that additional sources from aqueous-phase processing of absorbed volatile organic
compounds must be taken into account. Secondary organic aerosol (SOA) sources in fog water
could therefore explain the “amplification” of ROS activity with respect to the parent aerosol
population.

Figure 3.4: Per-volume (µg/m
3
) (a) and per-mass (µg/µg) (b) concentration of WSOC in
daytime/interstitial and fog water samples in the Po Valley in Fall 2015. Bars represent
geometric means and error bars correspond to one standard error (SD).
a)
0
2
4
6
8
10
12
WSOC (µg/m
3
)
Daytime aerosol Nighttime aerosol Fog

b)
0.00
0.05
0.10
0.15
0.20
0.25
0.30
WSOC (µg/µg)

63

Figure 3.5: Per-volume (MicrogramZymosan/m
3
air) (a) and per-mass (MicrogramZymosan/mg
PM) (b) ROS activity of the fog water and aerosol samples collected in the Po Valley in fall
2015. The results pertain to parallel aerosol and fog samples. The values for aerosol samples are
based on the unfiltered ROS analysis protocol. Bars represent geometric means and error bars
correspond to one standard error (SD).
a)
0
20
40
60
80
100
120
140
Per‐volume ROS
(MicrogramZymosan/m
3
)
Daytime aerosol Nighttime aerosol Fog

b)
0
1000
2000
3000
4000
5000
6000
7000
Per‐mass ROS
(MicrogramZymosan/mg PM)

To further explore the difference in the toxicity of daytime and interstitial aerosols, the ROS
activity of daytime and nighttime aerosols was also evaluated based on the filtered ROS analysis
protocol, representing the toxicity of the WS components (Figure 6). As already mentioned,
volume-based ROS activity of daytime aerosols was 3-5 times higher than that of interstitial
64
aerosols, but this difference becomes even greater in the filtered extracts (a factor of 7, see
Figure 6(a)). It should be noted that during the daytime, the per PM-mass filtered ROS activity of
aerosols (which relates to the redox activity of only the WS fraction) is almost half of the
unfiltered ROS levels (representing the redox activity of both WS and WI fractions). However,
during nighttime, the per PM mass ROS activity of the WS fraction contributes approximately
only 25% to the overall per-mass ROS aerosol activity (Figure 6(b)). These results imply that
during daytime, WS and WI PM fractions contribute roughly equally to the ROS activity of the
aerosol, whereas during nighttime, the WI components (comprising mostly elements, metals and
insoluble carbonaceous material) contribute to as much as 75% of the aerosols ROS activity.
Therefore, even if daytime and interstitial aerosols exhibit the same intrinsic toxicity (Figure 5),
this is driven by different chemical composition in the two aerosol populations. The most critical
difference between daytime and interstitial (i.e., unscavenged) aerosols is that the latter are
depleted of WS components, including WSOC (Figure 1) and inorganic aerosols (Figures 2 and
3) which are efficiently scavenged by fog. Therefore, as discussed above, the higher ROS
activity of daytime aerosols compared to that of nighttime/interstitial aerosols must be attributed
to WS components, while, at nighttime, the aerosol ROS activity is mainly driven by WI
components. It is also noteworthy that the highest ROS activity was observed in the fog water,
which is enriched in WS components compared to the daytime and interstitial aerosols.

65
Figure 3.6: Per-volume (MicrogramZymosan/m
3
air) (a) and per-mass (MicrogramZymosan/mg
PM) (b) ROS activity of the aerosol samples collected in the Po Valley in fall 2015. The values
for aerosol samples are based on both filtered (representing the redox activity of WS
components) and unfiltered (representing the redox activity of both WS and WI components)
ROS analysis protocols. Bars represent geometric means and error bars correspond to one
standard error (SD).

a)
0
20
40
60
80
100
Daytime aerosol Nighttime aerosol
Per‐volume ROS
(MicrogramZymosan/m
3
)
Unfiltered Filtered

b)
0
500
1000
1500
2000
2500
3000
Daytime aerosol Nighttime aerosol
Per‐mass ROS
(MicrogramZymosan/mg PM)

66
3.3.3 Association of ROS activity to PM and fog chemical components
To investigate the main species driving the ROS activity in fog water, we performed Spearman
rank correlation as well as Multilinear regression (MLR) analysis between per-volume
concentrations of all species and per-volume ROS activity of fog water, the results of which are
presented in Tables 1, 2, and in Table 3. As can be seen in Table 1, most of the WS species were
strongly correlated with the ROS activity of fog. However, this does not necessarily mean that all
of these species, including inorganic ions, are responsible for the aerosol redox activity. Previous
studies have indicated that some of these associations with toxicologically innocuous species
(e.g., inorganic ions) are observed because of the co-linearity of inorganic ions with important
redox active species, including WSOC, and not because of the toxicity of these components per
se (Ntziachristos et al., 2007b) (Cho et al., 2005) (Verma et al., 2012a). For instance, (Verma et
al., 2012a) found strong positive correlations between the redox activity of quasi-ultrafine PM
(PM0.18) and NO
3
-
, SO
4
2-
, NH
4
+
, and WSOC; however, the authors concluded that the strong
correlations observed between these inorganic ions and redox activity is because of their co-
linearity with WSOC (as also shown in Table 4) rather than their own toxicity, as these
components are not mechanistically active in these assays (Cho et al., 2005). It should be noted,
however, that inorganic acids can indirectly affect the ROS activity of fogs and wet aerosols by
solubilizing redox-active metals (Fang et al., 2017; Giulianelli et al., 2014; Oakes et al., 2012).
Such mechanisms can be important for areas where SO
2
levels are high, while they are less
plausible for environments such as the Po Valley where aerosols and fogs are largely neutralized
by ammonia, as demonstrated by the typical fog pH of about ~6 , according to Giulianelli et al.
(2014).

67
Table 3.1: Spearman rank correlation coefficients between per-volume concentrations of water-
soluble species as well as metals/elements in the fog water samples and the corresponding ROS
levels. Correlation coefficients which were statistically significant (at P<0.05) are highlighted in
bold.
Species
Correlation
Coefficient
Species
Correlation
Coefficient
Water-soluble components
Acetate 0.72* Cl
-
0.86
Dimethylamine 0.58 K
+
0.93
Ethylamine 0.90 Mg
2+
0.90
Formate 0.75* Na
+
0.89
Methansulfonate 0.89 NH
4
+
0.82
Methylamine 0.68* NO
2
-
-0.75
Oxalate 0.52 NO
3
2-
0.97
Trimethylamine 0.89 SO
4
2-
0.61
Ca
2+
0.86 WSOC 0.93

Metals/Elements
Na 0.26 Ni 0.26
Mg 0.60 Cu 0.49
Al 0.49 Zn 0.60
Ca 0.49 As 0.26
Ti 0.49 Cd 0.49
68
V 0.26 Ba 0.49
Cr 0.49 Pb 0.60
Mn 0.60 Th 0.49
Fe 0.49
*Denotes statistical significance at P<0.1.

The attribution of ROS activity of fog to dissolved organic substances was also confirmed by the
results of the MLR analysis (Table 3), demonstrating that WSOC alone explains 98% of the
variability in the fog water ROS levels. We observed much smaller correlation coefficients
between the ROS activity of fog water samples and elemental/metallic components, which have
been shown to correlate with particle toxicity in earlier studies (Ntziachristos et al., 2007b) (Cho
et al., 2005) (Hu et al., 2010) (Verma et al., 2012a). This is due to the lower concentration of
metals/elements in the fog water compared to daytime and interstitial aerosols, as shown in
Figure 3.

Table 3.2: Spearman rank correlation coefficients between concentrations of the metals/elements,
markers of organic aerosol, and water-soluble (WS) components in the aerosol samples and the
corresponding ROS levels. Statistically significant (P<0.05) correlation coefficients are
highlighted in bold.
Species
Correlation
Coefficient
Species
Correlation
Coefficient
Markers of organic aerosol
m/z 44 (marker of 0.43 m/z 60 (marker of 0.31
69
OOA
1
) BBOA
2
)
Water-soluble components
Organic matter 0.43 K
+
0.60
WSOC 0.66 Mg
2+
0.62
Acetate 0.66 Na
+
0.60
Formate 0.66 NH
4
+
0.31
Methansulfonate 0.83 NO
3
-
0.66
Oxalate 0.83 SO
4
2-
0.66
Cl
-
0.66
Elements/metals
Na 0.54 Ni 0.77*
Mg 0.03 Cu 0.77*
Al -0.25 Zn 0.60
Ca 0.33 As 0.66
Ti -0.52 Cd 0.31
V 0.74 Ba -0.15
Cr 0.26 Pb 0.66
Mn 0.66 Th 0.14
Fe 0.77*
*Denotes statistical significance at P<0.1.
1 OOA: Oxygenated organic aerosol
2 BBOA: Biomass burning organic aerosol

70
Table 3.3: Output of multiple linear regression (MLR) analysis using ROS activity as the
dependent variable and ambient concentrations of the measured chemical species as independent
variables.
Category Species Standardized
coefficients
Units Partial R R P-value R
2

Fog Samples WSOC 0.97 µg Zymosan/ m
3
air 0.98 0.98 <0.0001 0.96
MSA 0.79 µg Zymosan/ m
3
air 0.94 0.016
Aerosols
Fe 0.30 µg Zymosan/ m
3
air 0.73
0.97
0.161
0.94

Table 3.4: Spearman rank correlation coefficients between WSOC and inorganic species in the
fog water samples. Correlation coefficients which were statistically significant (at P<0.05) are
highlighted in bold.
Species
Correlation
Coefficient
Species
Correlation
Coefficient
Water-soluble components
Acetate 0.86 Cl
-
0.78
Dimethylamine 0.81 K
+
0.99
Ethylamine 0.81 Mg
2+
0.89
Formate 0.89 Na
+
0.75*
Methansulfonate 0.96 NH
4
+
0.96
Methylamine 0.88 NO
2
-
-0.53
Oxalate 0.68* NO
3
2-
0.96
71
Trimethylamine 0.96 SO
4
2-
0.79
Ca
2+
0.96
*Denotes statistical significance at P<0.1.

Table 3.5: Results of the Principal Component Analysis (PCA) performed on the aerosol
samples.
Source factors Species
Transition metals (tracers
of combustion-related
sources)
Water-soluble components
(tracers of secondary
aerosol)
Fe 0.98 0.19
Mn 0.96 0.24
V 0.93 0.35
Ni 0.94 -0.08
Cu 0.96 0.23
WSOC 0.03 0.98
MSA 0.49 0.85
Oxalate 0.46 0.79
NO
3
-
-0.17 0.99
SO
4
2-
0.22 0.96

72
To investigate the major PM chemical species that contribute to the ROS activity of daytime and
interstitial aerosols, we performed a principal component analysis (PCA) followed by a linear
correlation and a multilinear regression (MLR) analysis on the data pertaining to aerosols
samples. It should be noted that, due to the limited number of aerosol samples analyzed for ROS
activity (a total of 6), the data for both daytime and nighttime (i.e., interstitial) aerosols were
combined together, and the following analyses were performed on the pooled data. Table 5
presents the results of the PCA conducted on the chemical components of daytime and interstitial
aerosols. As can be seen in the table, two source factors were resolved, together explaining 95%
of the variance in the data. The first source factor comprises species that most likely are of
combustion origin (e.g., traffic, power plants), even though some of them (e.g., Fe, Mn, and Cu)
may also come from non-exhaust traffic sources in other areas (Sanderson et al., 2014). The
metals in this group, including V, Fe, Cu, Mn, and Ni, are considered toxic and known as redox-
active metals (Argyropoulos et al., 2016) (Sowlat et al., 2016a; Wang et al., 2016a). The second
source factor consists of water-soluble components, including WSOC as well as the ionic
components (methane-sulfonate, oxalate, sulfate, and nitrate) which are tracers of secondary
aerosols (Hu et al., 2010) (Hasheminassab et al., 2013). We should point out that the PCA
grouping of the chemical species reflects both the possible day-to-day variations in the
contributions of the specific sources to PM mass as well as the diurnal cycles in concentrations
governed by the specific fog scavenging rates (Figures 2, 3). Table 2 shows the Spearman rank
correlation coefficients between the WS components and elemental/metallic species of the
aerosols and the corresponding ROS levels. Very strong positive correlations were observed
between the ROS activity of aerosol samples and many of the WS components, which are also
known tracers of secondary organic aerosols (SOA); for example, the Spearman rank correlation
73
coefficients between the ROS activity and the concentrations of MSA and oxalate were 0.83 and
0.83, respectively. Similar as in the case of fog water, the observed association of inorganic ions
and low-molecular weight organic acids with the aerosol ROS activity is probably because of the
co-linearity of the ionic species and (unspeciated) redox-active organic compounds.
Interestingly, correlation coefficients are greater for the SOA tracers (oxalate, MSA) than for the
bulk WSOC mixture (R = 0.66) and for the whole organic mass measured by the AMS (R =
0.43), suggesting that the ROS activity is prevalently contributed by the secondary fraction of
particulate organic matter. However, Table 2 also shows strong positive and statistically
significant correlations for several metal/elemental species, such as Fe, Ni, and Cu, which are
primary components of the aerosol. The correlation results were also corroborated by the
outcome of the MLR analysis (Table 3), indicating that MSA (representative of the secondary
organic aerosols source factor) and Ni (representative of the primary combustion-related source
factor) were highly correlated with the ROS activity in the aerosol samples. Judging by the
standardized coefficients (0.79 and 0.30 MicrogramZymosan/m
3
air for MSA and Fe,
respectively) and partial correlation values (0.94 and 0.73 for MSA and Fe respectively), the
secondary organic aerosol fraction (represented by MSA in the MLR analysis) overall has a
higher contribution to the aerosol ROS activity compared to combustion-related metals
(represented by Fe in the MLR analysis). The results discussed above imply that during daytime,
the aerosol ROS activity is driven by a combination of WS organics and redox active metals.
However, during nighttime, when fog scavenging partitions the WS components into the water
phase and enriches the nighttime (or interstitial) aerosols with metals, the aerosol ROS activity is
mainly driven by elemental/metallic components, while the ROS activity of fog is mostly driven
by WS organic compounds.
74
WSOC is the main driver of toxicity in fog water and also a likely contributor of toxicity of PM1
during daytime when the contribution of WS components to toxicity is significant (Fig. 2).
Submicron organic aerosols in rural Po Valley originate mainly from combustion processes
(biomass burning) but a significant fraction of mass is the product of chemical transformations
occurring in wet aerosols and fog droplets (“aqueous SOA”) (Gilardoni et al., 2016). Biomass
burning organic tracers and their products of oligomerization formed in the aqueous phase
contain hydroquinone and catechol moieties that can participate to redox reactions leading to
ROS formation. The formation of such redox active organic compounds occurs in the aqueous
phase, which explains a correlation with ionic tracers such as oxalate. Another source of reactive
WSOC dissolved in deliquesced aerosols and especially in fog water is the uptake of water-
soluble reactive gases. It is well known that organic vapours, like low-molecular weight organic
acids and carbonyls, can account for a significant fraction of WSOC in fog (up almost 50%
according to (Collett et al., 2008)). Some, like formic acid, exhibit ROS activity (Du et al.,
2008). In this study, the contribution of low-molecular weight organic acids to WSOC was on
average 8% ± 4%, which represents a lower limit for the contribution of WS VOC to fog solutes
(as formaldehyde and other carbonyls were not measured). The uptake of reactive gases could
explain the excess of WSOC in fog water compared to the scavenged fraction of daytime PM1
(Figure 4) as well as the higher intrinsic toxicity of fog components, as shown in Figure 5.

75
3.4 Discussion and Conclusions
This study reports ROS activity data for PM1 and fog water samples at a continental rural area
virtually free of local traffic emissions and where aerosol mass is prevalently contributed by
biomass burning and secondary organic and inorganic components (Gilardoni et al., 2014). The
aerosol chemical composition differed substantially from that of urban PM for which the induced
oxidative potential has been documented by several studies (Saffari et al., 2014) and can be put
in relation to the redox-active activity of traffic-related aerosols (Ning and Sioutas, 2010).
Despite the very limited in situ traffic emissions at the Po Valley site, the ROS levels recorded
for PM (1879±851 and 1873±1199 MicroZymosan/mg during day and night respectively) and
fog samples (4142±1347 MicroZymosan/mg) in our study are comparable with those reported by
(Saffari et al., 2014), who reported ROS values of approximately 800 and 6000
MicrogramZymosan/mg PM in Los Angeles (LA) for PM2.5 and quasi-ultrafine PM (PM0.25),
respectively. Additionally, in a more recent on-road study conducted in 2015 inside two major
freeways of Los Angeles, the average (±SD) per-mass ROS activity of PM2.5 particles was
found to be 3660 (±1743) and 3439 (±3058) MicrogramZymosan/mg PM for I-110 and I-710,
respectively (Shirmohammadi et al., 2017b) , indicating that the intrinsic toxicity of the Po
Valley fog water is higher than that of PM2.5 particles collected on the freeways of Los Angeles.
Although direct comparison of the ROS activity of the fog and aerosol samples collected in this
study with those of ambient aerosol previously reported by other studies is associated with some
uncertainties (due mainly to the different size range of the particles collected and the fact that the
ROS activity appears to be strongly particle size-dependent (Sioutas et al., 2005a) (Saffari et al.,
2014), nonetheless these results point to possible health effects associated with PM exposure
during fog episodes in the Po Valley, the toxicity of which are comparable or, in many cases,
76
higher than that of the highly toxic traffic-related PM. Our results show that the toxicity of
aerosol particles accumulating in orographic depressions at the mid-latitudes during the cold
season, which is normally peak concentration season for PM in many continental areas, can be
further amplified by the formation of fog, whose intrinsic toxicity is even greater than that of the
original aerosol particles scavenged into the droplets. Moreover, the redox potential of fog
solutes is mainly driven by oxidized organic compounds, which also explains the excess of ROS
activity in fog water with respect to the scavenged fraction of the aerosol. The effect of
secondary organic species on ROS activity in fog and aerosols in the Po Valley is clearly
demonstrated by the fact that fog water exhibits a higher intrinsic toxicity with respect to PM1,
despite its depletion of redox active metals as a consequence of the systematically different
scavenging rates between WS and WI aerosol species. The contribution of WS secondary species
to ROS activity is supported also by: a) the results of MLR analysis, b) the scavenging rate of
total redox-active compounds (71% of ROS activity) which is higher than that for most
combustion-related metals (typically < 60%), corroborating the contribution of hygroscopic
particles to ROS activity; and c) the results obtained from the filtered extracts indicating that ca.
50% of ROS activity is attributable to WS species in daytime (i.e., out-of-fog) conditions. The
origin of secondary organic compounds responsible for the ROS activity of aerosol WSOC in the
Po Valley cannot fully be elucidated based on this set of data. However, previous observations at
the same site in the fog season showed that SOA are produced in large amounts by aqueous
reactions in fog droplets and deliquesced aerosols starting from organic compounds emitted from
biomass burning (Gilardoni et al., 2016). Production of biomass burning SOA is accompanied by
the formation of redox active organic compounds, including hydroquinones. We can therefore
speculate that the enhancement of ROS activity in fog water is at least partly irreversible, as
77
evaporating fog droplets become enriched on newly-formed redox-active secondary species. The
relevance of secondary sources of toxicity in fog and fog-processed aerosols calls for more
stringent controls on possible precursor emissions, which should be pursued by policy makers
including international authorities (as secondary PM components concentrations are often
triggered by transboundary pollution) (Kiesewetter and Amann, 2014). Finally, the health effects
of the exposure to fog water toxics depends on fog frequency, and in turn on climate conditions
and climate change. Substantial reductions (-50%) of fog occurrence in the last 30 years have
been documented for the Po Valley and many other European locations, and these dramatic
changes have been attributed to global warming and changes in cloud condensation nuclei
concentrations (Vautard et al., 2009) (Giulianelli et al., 2014). The enhanced toxicity of fog
droplets observed in this study suggests that the historical reduction of fog frequency may result
in an unintended improvement of air quality in many continental areas, overlapping also with the
deliberate reduction of PM emissions put into practice since the early 90’s in many developed
countries.
78
Chapter 4:
Development and field evaluation of an online monitor for near-continuous
measurement of iron, manganese and chromium in coarse airborne
particulate matter (PM)
A novel air sampling monitor was developed for near-continuous (i.e., 2-hr time resolution)
measurement of Iron (Fe), Manganese (Mn), and Chromium (Cr) concentrations in ambient coarse
particulate matter (PM) (i.e., PM
10-2.5
). The developed monitor consists of two modules: 1) the coarse
PM collection module, utilizing two virtual impactors (VIs) connected to a modified BioSampler to
collect ambient coarse PM into aqueous slurry samples; 2) the metal concentration measurement
module, which quantifies the light absorption of colored complexes formed through the reactions
between the soluble and solubilized target metals and pertinent analytical reagents in the collected
slurries using a Micro Volume Flow Cell (MVFC) coupled with UV/VIS spectrophotometry. The
developed monitor was deployed in the field for continuous ambient PM collection and measurements
from January to April 2016 to evaluate its performance and reliability. Overall, the developed monitor
could achieve accurate and reliable measurements of the trace metals Fe, Mn, and Cr over long
sampling periods, based on the agreement between the metal concentrations measured via this online
monitor and off-line parallel measurements obtained using filter samplers. Based on our results, it can
be concluded that the developed monitor is a promising technology for near-continuous measurements
of metal concentrations in ambient coarse PM. Moreover, this monitor can be readily configured to
measure the speciation (i.e., water-soluble portion as well as specific oxidation states) of these metal
species. These unique abilities are essential tools in investigations of sources and atmospheric
processes influencing the concentrations of these redox-active metals in coarse PM.

This chapter is based on the following publication:

Sowlat, M.H., Wang, D., Simonetti, G., Shafer, M.M., Schauer, J.J., Sioutas, C., 2016. Development
and field evaluation of an online monitor for near-continuous measurement of iron, manganese, and
chromium in coarse airborne particulate matter (PM). Aerosol Science and Technology 50, 1306-
1319.

4.1. Introduction
The current body of epidemiological studies provides unequivocal evidence for the association
between exposure to particulate matter (PM) and increased risk of cardiovascular, neurological
as well as respiratory diseases, hospitalization, and premature death (Brunekreef and Forsberg,
2005; Davis et al., 2013a; Delfino et al., 2005; Delfino et al., 2010b; Dockery and Stone, 2007;
Gauderman et al., 2015; Perez et al., 2008; Pope et al., 2004b; Pope Iii et al., 2002). It should be
79
noted that most of these studies have linked the health end-points with PM mass concentrations
of airborne PM, however, there is growing evidence highlighting the important role of the
chemical composition of ambient PM, rather than simply the mass in driving the health outcomes
(Claiborn et al., 2002; Verma et al., 2009).

Redox-active transition metals, including iron (Fe), manganese (Mn), and chromium (Cr), are
among the few chemical components of ambient PM that studies have documented an
association with human health impacts, likely due to their ability to induce oxidative stress (e.g.,
via the generation of reactive oxygen species (ROS)), which usually occurs as a cascade of
events that significantly increases the ROS concentration in the target cells and causes
inflammatory responses in cells and tissues (Delfino et al., 2005; Donaldson et al., 2003; Li et
al., 2009; Peters et al., 2006; Tao et al., 2003). Therefore, a focus of many recent studies has
been on the evaluation of sources, transport, and chemical speciation of these redox-active trace
metals (Chester and Stoner, 1974; Fang et al., 2015a; Harrison et al., 2012; Moreno et al., 2006;
Pérez et al., 2008; Putaud et al., 2004; Putaud et al., 2010; Viana et al., 2008). Several studies
have also examined the particle-size dependency of oxidative potential and relationship to
chemical composition (Ayres et al., 2008; Charrier and Anastasio, 2012; Miljevic et al., 2010;
Ntziachristos et al., 2007a; Shafer et al., 2010; Tao et al., 2003; Valavanidis et al., 2005).
Nonetheless, the impact of individual chemical components of PM on the health end-points is
still relatively poorly understood; therefore, the development of innovative PM measurement
techniques capable of providing high-resolution of chemically-speciated data and their use in
concurrent toxicological studies can drastically improve our current understanding of the main
PM species driving these health effects (Karlsson et al., 1997; Wang et al., 2015; Weber et al.,
80
2001). These novel techniques will also help improve our understanding of the dynamics of
atmospheric PM emission, dispersion, and fate through the provision of data with higher time
resolution. In particular, the development of techniques/methods capable of measuring different
forms of redox-active metals (i.e., different oxidation states as well as solubility characteristics)
would be very important in studying aerosol toxicity. This is because the toxicity of PM-bound
metals is significantly impacted by their solubility, with the water-soluble fraction of the redox-
active metals better reflecting in many cases the bioavailable pool (Heal et al., 2005; Shi et al.,
2003). Additionally, in most cases, the toxicity of these metals, per se, is dependent upon their
oxidation state (Khlystov and Ma, 2006; Majestic et al., 2007; Majestic et al., 2006).
Recently, a novel online monitor was developed and successfully applied for the near-continuous
measurement of PM
2.5
-associated Fe, Mn, and Cr (Wang et al., 2016a). The system comprised of
two modules: 1) the PM collection module employing an aerosol-into-liquid collector; and 2) the
chemical analysis module employing Micro Volume Flow Cell (MVFC) coupled with
spectrophotometry. One of the critical advantages of the developed system over previously
available technologies (e.g., aerosol time-of-flight mass spectrometry (ATOFMS) or X-ray-
fluorescence (XRF) based technologies) is that it can measure these metals in both total or water-
soluble forms, as well as in different oxidation states with a relatively high time resolution (i.e., 2
hr), which can provide crucial information on the biological pathways underlying the adverse
health effects, which are mostly speciation-driven (Wang et al., 2016a).
There has been a variety of novel technologies developed to measure the on-line concentrations
metals in the fine PM size fraction, including the particle-into-liquid sampler (PILS) (Weber et
al. 2001); the Cooper Xact
TM
625 Monitoring System; the AMMS-100 Focused Photonics
(Hangzhou), Inc., China; and the semi-continuous elements in aerosol system (SEAS) (Kidwell
81
and Ondov 2001, 2004). Despite the recent development of these high time-resolution
measurement methods for PM
2.5
-bound elements, to the best of our knowledge, similar systems
have not yet been developed for the measurement of redox-active metals in the coarse PM
fraction (PM
10-2.5
, i.e., particles with an aerodynamic diameter of between 2.5 µm and 10 µm)
and, therefore, are highly needed. One of the factors hindering the development of such systems
for coarse PM is the low solubility of these species in the coarse size range (Birmili et al., 2006;
Wang et al., 2013), which limits the ability to accurately measure the total PM-bound
concentration of the target metals. A large fraction of many of these metals and trace elements is
partitioned in the coarse PM fraction (Allen et al., 2001), which further highlights the need for
developing techniques capable of measuring coarse PM-bound redox-active metals with a high
time-resolution. Such tools will be crucial in investigations of the health impacts induced by the
coarse fraction of PM, as recent studies have indicated significant association between these
metals and the redox activity of coarse PM (Becker et al., 2005; Becker and Soukup, 2003;
Cheung et al., 2011; Cheung et al., 2012; Shafer et al., 2010; Wang et al., 2015). To address
these needs, in the present study, we developed a new technique for near-continuous (i.e., a time
resolution of 2 hr) measurement of three coarse PM-bound redox-active metals, namely Fe, Mn,
and Cr. Following laboratory characterization, the system was operated in the field from January
to April 2016 to assess its performance and reliability with minimum supervision over a
relatively long period of time. Additionally, in order to check the validity of the results, the
online-measured data were compared with those obtained by means of inductively coupled
plasma mass spectrometry (ICP-MS) analysis conducted on filter samples collecting
concurrently coarse PM through a parallel sampling system. Finally, in this paper, we present the
82
results of four months of system deployment in the field, enabling us to explore the diurnal
concentration profiles of these redox-active metals in ambient coarse PM.

4.2. Materials and methods
4.2.1. Reagents and standards
The water used for the collection of ambient coarse PM and system rinsing was produced by a
Millipore A-10 water purification system (EMD Millipore, Billerica, MA). All the acids were
trace metal grade (VWR). All the chemicals and reagents were prepared in polypropylene
laboratory equipment, which were acid-washed (with either 4 M hydrochloric acid (HCl) or 4 M
nitric acid (HNO
3
) and water-rinsed before use. For Fe, a 1000 ppm stock solution was first
prepared gravimetrically from (NH
4
)
2
Fe(SO
4
)
2
(ACS) salt, and then diluted to obtained standard
solutions in the range of 0-200 ppm. The Fe analytical reagent (i.e., Ferrozine) was prepared by
addition of 133 mg of ferrozine (Sigma) to 50 mL of water that contained 65 µL of 4 M HCl
(Rastogi et al., 2009). The Fe reducing agent (HA) was prepared by dissolving 19.3 mg of
hydroxylamine hydrochloride (HA) solid (Sigma), with a purity of 99.9999%, in 50 mL of water
(Majestic et al., 2006). One molar sodium hydroxide (NaOH) solution was prepared from reagent
grade NaOH solid (AMRESCO) and was used for pH adjustment.
For Mn, a 100 ppm stock solution was first prepared gravimetrically from MnCl
2
(ACS) salt, and
then diluted to obtained standard solutions in the range of 0-100 ppm. The manganese analytical
reagent (formaldoxime; FAD) was prepared by dissolving 20 g of HA in 450 ml of water,
followed by the addition of 10 ml of 37% formaldehyde solution, and the solution was then made
83
up to 500 ml with water. As with Fe, 1 M NaOH solution was used for pH adjustment. Both the
Fe and Mn stock solutions were acidified to pHs below 1 using a 4 M HCl solution.
For Cr, a 100 ppm stock solution was first prepared gravimetrically from K
2
CrO
7
(ACS) salt, and
then diluted to obtained standard solutions in the range of 0-100 ppb. In the case of Cr, the stock
solution was acidified using 4 M HNO
3
. Additionally, in order to prepare the chromium
analytical reagent (i.e., diphenylcarbazide, DPC), 167 mg of the reagent was dissolved into 100
ml of acetone. Afterwards, it was mixed with 1.67% H
2
SO
4
solution at 1:1 volume ratio
(Khlystov and Ma, 2006). Finally, 0.143 ml of H
2
O
2
was diluted to 100 ml of 0.1 M NaOH
solution to prepare the 0.1% H
2
O
2
solution.

4.2.2. System configuration
The system configuration is presented in Figure 1. The developed system comprises two
modules: 1) the coarse PM collection module; and 2) the metal concentration measurement
module. The coarse PM collection module utilizes two virtual impactors (VIs) connected in
parallel, coupled with a modified BioSampler (i.e., liquid impinger) (BioSampler, SKC West,
Inc., Fullerton, CA) technology, for which extensive details can be found elsewhere (Wang et al.,
2015). Briefly, the air is drawn into two round nozzle VIs, each having a major flow rate of 100
L/min and a minor flow rate of 5 L/min. This total flow rate corresponds to a theoretical 50%
cutpoint of 1.5 µm in aerodynamic diameter, concentrating airborne coarse PM into the VI’s
minor flow. As discussed in our earlier studies (Wang et al., 2015), although the theoretical cut
point of the VIs is slightly lower than the traditional definition of the coarse PM size range (i.e.,
2.5 µm), this cut point was selected to ensure that the entire size range of ambient coarse PM is
concentrated by the maximum enrichment factor of about 20. The two minor flows are combined
84
into a total flow rate of 10 L/min entering the modified BioSampler, in which particles are
captured in 20 ml of ultrapure water to form a particle-liquid suspension. The collection
efficiency of the BioSampler at that flow rate is near 100% for particles above 1.5 µm (Kim et
al., 2001).
Figure 4.1: System schematic of the coarse PM metal monitor

Ambient air
PM
10
inlet
100 lpm
VI
VI: Virtual impactor
100 lpm
5 lpm
Biosampler
10 lpm
Ultrapure water
Peristaltic pump
Level detector
5 lpm
Acid
NaOH
Syringe pump for
injecting reagents
pH meter
Reaction
bottle
pH probe
Serpentine
reactor and
reaction coils
Spectrophotometry
unit (UV light,
MVFC, and
spectrophotometer)
Data collection
Chemical analysis module PM collection module
Peristaltic pump
85
The second module, for which in-depth details can be found in (Wang et al., 2016a), starts with
the transferring of the aqueous sample into a capped bottle for chemical reactions. In this step,
concentrated acids (i.e., 4 M HCl for Fe and Mn, and 4 M HNO
3
combined with 1% H
2
O
2
for
Cr) are added to the sample at a volume ratio of 1:20 (acid : sample). The addition of acids
results in very low pH levels (i.e., around 0.5) to ensure that nearly all metal species are
solubilized during the 10 min residence time of this step. The pH is then adjusted to 5-7 for Fe-
ferrozine and 7-8.5 for Mn-FAD reactions by adding sufficient amounts of 1 M NaOH. It should
be noted that no NaOH is added for Cr measurements, since the Cr-DPC reaction is optimal at
pH=1. In the next step, oxidation/reduction agents (i.e., HA) are added to the sample solution to
convert all different oxidation states of the target metals to a uniform state. Additionally, for Mn
measurements, the ethylenediaminetetraacetic acid (EDTA, 0.08 N) solution was added to the
sample at a volume ratio of 1% to eliminate any possible interference from Fe existing in the
ambient coarse PM sample (Majestic et al., 2007). For Cr, however, no oxidation agent was
added to the sample solution, mainly because all the Cr(III) was already converted to Cr(VI) due
to the addition of H
2
O
2
. After a total residence time of 10 min in the serpentine reactor and then
in the reaction coil, the sample is transferred to a 10-cm path-length, optical flow cell (60 µL
internal volume) MVFC (FIA-ZSMA-ML-100-TEF, Ocean Optics, Inc., Dunedin, FL) for the
spectrophotometry detection of the metals. The concentration of each metal was determined at
the wavelength with the highest level of absorption for the corresponding analytical reagent (i.e.,
ferrozine at 562 nm for Fe, FAD at 450 nm for Mn, and DPC at 540 nm for Cr), while a
wavelength of 700 nm, at which there is effectively zero absorptivity for the analytical reagents,
was selected to determine the background sample absorption. It should be noted that the three
target metals were not simultaneously measured, due mainly to limitations in the available
86
equipment. Nonetheless, concurrent measurements can be easily implemented by splitting the
concentrated slurry flow and using multiple spectrophotometry units in parallel. Additionally, in
this study we focused on measuring the total concentrations of the three metals, however, as
indicated in our previous work by Wang et al. (2016), the system configuration can be easily
modified to measure water-soluble and different oxidation states of the target metals. The former
can be achieved by removing the acid digestion step and using inserting an in-line liquid
filtration unit (e.g., a 0.20 μm polypropylene syringe filter) to remove the water-insoluble
fraction before the spectrophotometric measurement. The latter can be achieved by performing
measurements with and without the addition of the oxidation and reduction reagents. For
instance, the water-soluble Fe can be measured by omitting the HCl addition in the measurement
sequence and filtration before the sample passes through the MVFC. Additionally, in order to
measure Fe(II), the addition of the HA solution would be omitted from the procedure to prevent
reducing Fe(III) to Fe(II). Furthermore, Fe(II) and total Fe can be measured using two distinct
measurement lines, and the Fe(III) concentration can be easily calculated by subtracting Fe(II)
concentration from that of total Fe. As illustrated by Wang et al. (2016), the same procedure can
also be applied to Mn and Cr to measure water-soluble fraction and/or different oxidation states.

4.2.3. Field evaluation tests and continuous long-term operation
The developed system for the near-continuous measurement of Fe, Mn, and Cr in coarse PM was
deployed at the Particle Instrumentation Unit (PIU), located on the University of Southern
California's (USC) park campus, for field evaluation tests and continuous long-term run (i.e.,
approximately four months, from January to April 2016). As shown by previous studies
performed at this location, this is a mixed urban site where ambient PM is largely impacted by
87
vehicular emissions (Hasheminassab et al., 2014b; Sowlat et al., 2016a), as it is located 150 m
downwind of a major freeway, i.e., I-110. It is also located approximately 3 km directly to the
south of downtown Los Angeles, CA, and less than 2 km to the southwest of another major
freeway, i.e., the I-10.
Ambient coarse PM was collected in 2-hr time intervals, resulting in 12 data points per day, and
the concentrations of the target metals (i.e., Fe, Mn, and Cr) were subsequently measured using
the system described above. After each day of sampling, the system was manually switched to
measure the next metal species, therefore each metal was sampled every third day. Coarse PM
sampling was done 6 days a week, starting on Tuesday and ending on the next Monday to
incorporate weekend measurements as well. Every week, calibration and system maintenance
was also performed on the day the system was stopped. Each metal was measured in two days
per week. Additionally, the sequence of measuring these metals was changed every week to
make sure that we capture the variations of metals concentrations during the week as well as
during the weekends.
In addition, 5 time-integrated 24-hr filter samples were collected for each metal in parallel to the
online measurements in order to compare the online concentrations to off-line measurements
conducted on filters. For this purpose, another virtual impactor, with the same major and minor
flow rates (i.e., 100 and 5 L/m, respectively) sampled in parallel with the on-line sampler. The
air drawn by the minor flow of this virtual impactor was passed through 37-mm Teflon filters
(Teflo, Pall Corp., Life Sciences, 1-µm pore, Ann Arbor, MI), which were held in a 37 mm air
sampling cassette (Zefon International Inc., Ocala, FL). The filter samples were collected over a
24-hr interval to ensure that sufficient coarse PM mass was collected for the subsequent
elemental analysis using an magnetic sector inductively coupled plasma mass spectrometry (SF-
88
ICPMS) instrument. To directly compare the 2-hr online-measured data with those obtained from
24-hr filter samples, the online data were averaged over 24 hr.
Concurrent to the measurements of the target metals using the developed technique, ambient
coarse PM mass concentrations were continuously measured using a coarse particle mass
monitor previously developed, for which extensive details can be found in (Misra et al., 2001).
Briefly, the continuous coarse particle mass monitor (CCPM) consists of three major
components, including a PM
10
inlet, a round nozzle VI with a 2.5-µm cut point, and a standard
tapered element oscillating microbalance (TEOM 1400A, Thermo Fisher Scientific, MA, USA)
instrument. In this system, particles are drawn through the VI with a total flow rate of 50 L/min.
Coarse particles are then concentrated into the minor flow of 2 L/min and drawn through the
TEOM for continuous measurement of PM mass concentrations, while PM
2.5
were drawn into
the major flow (48 L/min).

4.2.4. SF-ICPMS analysis of filter samples
Following acid digestion of the PM, the total elemental composition of the particulate matter
(PM) collected on the 37mm Teflon filters was determined using magnetic sector inductively
coupled plasma mass spectrometry (SF-ICPMS; Thermo-Finnigan Element 2) (Okuda et al.
2014). Filter membranes were placed in micro Teflon PFA digestion vessels and PM solubilized
using a mixture of ultra-high-purity acids (1.0 mL of 16 M nitric acid, 0.25 mL of 12 M
hydrochloric acid and 0.10 mL of hydrofluoric acid) in an automated microwave-aided digestion
system (Milestone ETHOS+). This protocol effects a complete solubilization of all particle
phases and element species. Digestates were diluted to 15 mL with high purity water
(18 MΩ/cm
−1
) in pre-cleaned low-density polyethylene (LDPE) bottles and then analyzed by SF-
89
ICPMS. Forty-nine elements were quantified. Propagated analytical uncertainties were estimated
from the uncertainties (square root of the sum of squares method) of the SF-ICPMS instrumental
analysis, method blanks, and digestion recoveries. The latter correspond to the standard deviation
of replicate analyses of National Institute of Standards and Technology (NIST) Standard
Reference Materials (SRM). Six solid samples of three SRMs were digested and analyzed with
every analytical batch of samples. Further details of these protocols can be found in previous
publications (Zhang et al. 2008; Saffari et al. 2013; Okuda et al. 2014).

4.3. Results and Discussion
4.3.1. Calibration
For each metal, the measurement module (i.e., the spectrophotometry unit) of the developed
system was calibrated using the dilution series prepared from the stock solution. Table 1 presents
the results of the system calibration for each of the metals. As presented in the table, the
calibration results indicated robust linear associations between metal concentrations and
absorbance level at the respective analytical wavelength - the R-square of the calibration curve
was above 0.99 for all of the metals. Regression line slopes of 0.0042 (±0.0002) (Abosorbance
Unit (AU)/ppb), 0.0016 (±0.0001) (AU/ppb), and 0.0056 (±0.0004) (AU/ppb) for Fe, Mn, and
Cr, respectively, confirm the high sensitivity of the long optical path spectrophotometric
methods. It should be noted that system calibration was repeated multiple times during the
sampling campaign as an additional quality assurance step to corroborate the validity of the on-
line measurements. The limit of detection (LOD) for each metal was determined using the field
method blanks prepared according to the following procedure: ultrapure water was first injected
into the system and the relevant chemicals and reagents were added to the water as if this was an
90
actual aerosol sample. Then, the blank sample was passed through the MVFC and the absorbance
was measured. The LOD was estimated as 3 times the standard deviation of the field method
blanks (Khlystov and Ma, 2006; Majestic et al., 2007; Rastogi et al., 2009; Wang et al., 2016a).
According to the measurements and calculations mentioned above, the LODs were estimated to
be 0.3 ppb (i.e., 0.25 ng/m
3
), 0.2 ppb (i.e., 0.16 ng/m
3
), and 0.2 ppb (i.e., 0.16 ng/m
3
) for Fe, Mn,
and Cr measurements (based on a flow rate of 200 L/min and a collection time of 2-hr),
respectively. Our system has also high measurement sensitivity, demonstrated by the signal-to-
noise (S/N) ratios of 231, 83, and 38 for Fe, Mn, and Cr, respectively, based on the average metal
concentration levels observed in the current study.

Table 4.1: Results of the system calibration for each of the individual metals measured
Element Range Calibration curve (units
of slope: AU
a
/ppb)
R
2
Limit of Detection
b

(LOD) (ppb)
Limit of Detection
c

(LOD) (ng/m
3
)
Fe 0-200 ppb Y =0.0042x -0.0166 0.99 0.3 0.25
Mn 0-100 ppb Y=0.0016x+0.0075 0.99 0.2 0.17
Cr 0-100 ppb Y=0.0056x-0.0008 0.99 0.2 0.17
a
AU is absorbance unit
b
The LOD was calculated as 3 times the standard deviation of the field method blanks (i.e., ultrapure water as a
blank sample plus the pertinent reagents)
c
Estimated based on a sampling flow rate of 200 L/min over a 2-hr collection time

4.3.2. Comparison between online-measured data with filter samples data
As noted earlier, time-integrated 24-hr filter samples were collected for each metal in parallel to
the online measurements to verify the integrity of the readings by the developed sampler. Figure
2 presents the linear regression and correlation between the total (i.e., soluble plus insoluble)
91
concentrations of the metallic species measured by the developed monitor and those obtained
from ICP-MS analysis of the off-line filter samples collected in parallel. For these samples, the
ranges of total concentrations were 30-180 ng/m
3
for Fe, 5-27 ng/m
3
for Mn, and 2-14 ng/m
3
for
Cr. As shown in the figure, very good agreement was observed between the total concentrations
of the metallic species measured by the developed monitor and those obtained from the ICP-MS
analysis performed on the collected filters, based on the slope of regression line (i.e., 0.86±0.05)
and the R-square value (i.e., 0.96). This agreement based on total concentrations indicates that
the ambient coarse PM has been effectively digested. Additionally, the average (±SD)
online/filter concentration ratios were 0.83(±0.18), 0.93(±0.20), and 0.85(±0.11), respectively for
Fe, Mn, and Cr. The approximately 10% lower on-line versus off-line filter-based concentrations
could be attributed to the loss of particles (mostly the insoluble fraction) on the walls and bottom
of the Biosampler, particle loss in the tubing and in the reaction vessel, and the likely resistance
of a small fraction of the insoluble part of PM-bound metals to the acid digestion/solubilization
procedure.

92
Figure 4.2: Correlation between the metal concentrations measured online with off-line
concurrent measurements obtained using filter samplers. Error bars represent one standard
deviation of multiple online (n = 12) and offline (n=3) measurements for each data point.
Filter concentrations (ng/m
3
)
110 100
Online measurements (ng/m
3
)
1
10
100
Y=(0.86±0.05)X
R
2
=0.96

Previous studies have indicated that these metal species in the coarse PM fraction are largely
(between 60% and >90%) insoluble (Allen et al., 2001). Therefore, even though we did not
directly measure the insoluble fraction of these metals, the very good agreement between the
online-measured data and offline filter samples corroborates the method’s capability to
efficiently digest/solubilize the insoluble fraction and accurately measure the total concentration
of Fe, Mn, and Cr.

4.3.3. Diurnal trends of Fe, Mn, and Cr in coarse PM concentrations and relationship to
meteorological parameters in central Los Angeles
The developed monitor for online measurement of Fe, Mn, and Cr concentrations was deployed
at the PIU site for an approximate four month period, from January to April 2016, to assess the
long-term performance of the system. Table 2 presents the summary statistics of the online-
Fe data
Mn data
Cr data
93
measured concentrations of coarse PM-bound Fe, Mn, and Cr during the study period. This table
also provides the mass concentration of coarse PM measured using the CCPM running
concurrently with the metal monitor at the PIU. As shown in the table, the average coarse PM
mass concentration was 11.6 (±9.0) µg/m
3
over the study period. The coarse PM Fe
concentrations averaged 57.8 (±43.0) ng/m
3
(ranging from 14.4-180.0 ng/m
3
) and were higher
than those of Mn (i.e., 15.0 (±10.3) ng/m
3
,

ranging from 2.2-37.2 ng/m
3
) and Cr (i.e., 6.9(±4.1)
ng/m
3
, ranging from 1.5-17.2 ng/m
3
). In addition, The levels observed in the present study are in
concert with those previously reported in central LA at the same sampling site (i.e., the PIU),
both for particle mass concentrations and coarse PM-bound metals (Cheung et al., 2011; Cheung
et al., 2012).

Table 4.2: Summary statistics of the parameters collected over the entire study period
Variables
Geometric
Mean
Standard
deviation
Minimum Maximum
Fe (ng/m
3
) 57.8 42.9 14.4 180.0
Mn (ng/m
3
) 14.9 10.3 2.3 37.2
Cr (ng/m
3
) 6.9 4.1 1.5 17.1
Wind speed (m/s) 1.9 1.2 0.9 8.5
Temperature (°C) 15.2 4.8 6.1 32.8
RH (%) 47.9 23.9 6.0 99.0
Coarse PM (µg/m
3
) 11.6 8.9 1.20 55.1

94
Data for meteorological parameters, including wind speed and direction, temperature, and
relative humidity (RH) (Table 2) were acquired from the California Air Resources Board's
(CARB) online database for the sampling site that is located in North Main St., downtown Los
Angeles, which is located approximately 3 km northeast of our sampling site (i.e., the PIU).
Figure 3(a-c) illustrates the average diurnal variation of meteorological parameters (i.e.,
temperature, wind speed, and RH) in the study area during our sampling campaign. As shown in
the figure, maximum temperatures (around 20 °C) were observed in the middle of the day,
whereas RH (Figure 3(b)) peaked at night, reaching values up to 60%. The diurnal variations for
wind speed (Figure 3(c)) exhibited a sharp peak (around 3 m/s) in the afternoon. These results
are consistent with those reported previously in central Los Angeles (Cheung et al., 2011;
Cheung et al., 2012).

Figure 4.3(a-c): The diurnal variation of meteorological parameters in the study location
averaged over the study period; a) wind speed; b) temperature; and c) relative humidity. Error
bars represent one standard error (SE) of the mean.
a)
0
1
1
2
2
3
3
4
01 2345 678 9 1011121314151617181920212223
Hour of the Day
Windspeed (m/s)

95
b)
0
5
10
15
20
25
012 3456 7 8 9 1011121314151617181920212223
Hour of the Day
Temperature (°C)

c)
10
20
30
40
50
60
70
80
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour of the Day
RH (%)

Figure 4(a) presents the wind rose in the sampling site, indicating the direction and speed of the
prevailing wind in the study area during our campaign. As shown in the figure, the wind was
most commonly blowing from the northeast (approximately 27% of the time), with the highest
speeds ranging from 4 to 5 m/s. The second major wind direction during the sampling period was
westerly and southwesterly winds, exhibiting even higher wind speeds (above 5 m/s).
Figures 4(b-d) present the concentrations of the metals in association with wind direction during
the entire study period, presented separately for daytime and nighttime. As can be seen from the
plots, for Fe and Mn, the highest concentrations (i.e., 150-200 ng/m
3
for Fe and 30-40 ng/m
3
for
Mn) were observed during two distinct periods, i.e., westerly/southwesterly winds during the
daytime, and northeasterly winds during the nighttime. Our sampling site is located
approximately 150 m to the east of the I-110 freeway; it is also in close proximity (less than 2
km to the southwest) to the I-10 freeway. Therefore, the observed high Fe and Mn concentrations
during daytime westerly winds and nighttime northeasterly winds can be associated with traffic
96
emissions (i.e., road dust/resuspension soil) brought to the site by prevailing winds from these
two major freeways. It is also noteworthy that, for both Fe and Mn, although high concentrations
were observed during nighttime northeasterly winds, the average daytime concentrations during
westerly and southwesterly winds (i.e., 89.5 (±30.0) ng/m
3
and 19.4 (±11.6) ng/m
3
for Fe and
Mn, respectively) were higher than those during nighttime (i.e., 64.5 (±39.9) ng/m
3
and 13.6
(±8.7) ng/m
3
for Fe and Mn, respectively). In contrast to the Fe and Mn trends, as can be
observed in Figure 4(d), the coarse PM Cr concentrations follow distinctly different trends that
are suggestive of contributions of other potential local sources, including metal plating facilities
using Cr-containing paints (Ospital et al., 2008; Propper et al., 2015), in addition to the road
dust/resuspension soil noted earlier.
Figure 5(a) illustrates the diurnal variations of the coarse PM mass concentrations in the
sampling location during the study period. As can be seen in the figure, the PM mass
concentration starts to sharply increase at 8 am and reaches a maximum at 11-am-noon,
decreasing to slightly lower ranges until 6 pm, and then sharply decreases down to the lowest
values of 8-10 µg/m
3
at nighttime. This trend clearly indicates the impact of meteorological
conditions and traffic volume on CPM mass variations. As shown in Figures 3(a-c), during the
midday period, we have the lowest RH and highest wind speed, both of which favor the
resuspension of soil and road dust. This is also in agreement with the results from previous
studies conducted in central LA at the same sampling site (Pakbin et al., 2010; Pakbin et al.,
2011).

97
Figure 4.4(a-d): Prevailing wind direction and speed (a) during the study period and its
relationship with concentrations of Fe (b), Mn (c), and Cr (d).

a) b)

c) d)

The average diurnal variations of Fe concentrations are shown in Figure 5(b). The box represents
the interquartile range (the bottom and top lines of the box representing the first and the third
98
quartiles, respectively). The line inside the box represents the median, while the whiskers above
and below the box represent 95
th
and 5
th
percentiles. As can be seen from the figure, two peaks
were observed, one major peak during morning rush hours (8-10 am) and one minor peak during
late afternoon/early evening hours (4-8 pm), bringing the average Fe concentration up to around
90 ng/m
3
, compared to the lower values of around 40 ng/m
3
during other hours of the day. These
two peaks illustrate the impact of traffic-related emissions (most likely resuspended dust, given
the size range of the PM being measured in this study) on CPM levels at the sampling site. This
trend was also observed in Figure 4(b), indicating high Fe concentrations during westerly and
southwesterly winds, bringing atmospheric particles from the major freeway (i.e., I-110) located
approximately 150 m to the west of our sampling site. There is also a smaller third peak in
concentration observed in the middle of the day (1-3 pm), which corresponds to maximum
temperatures and wind speeds and minimum RHs, as shown in Figures 3(a-c). This peak can be
attributed to southeasterly winds, which are most frequent in the middle of the day (Wang et al.,
2015), particularly in the cold season (Dec – March, as in this sampling campaign). This finding
is also consistent with those reported by Sowlat et al. (2016a), observing very high contributions
from the soil/road dust during high temperatures and wind speeds and very low RHs.
Figure 5(c) shows the diurnal variations of Mn concentrations at the sampling location during the
study period. As can be seen from the figure, the Mn concentrations peaked in the early morning
(around 8-10 am), with an average concentration of 25 ng/m
3
. As in the case of Fe
concentrations, this peak can be attributed to the impact of traffic-related emissions (most likely
resuspended dust) on CPM levels at the sampling site, which was also observed in Figure 3(c),
indicating the highest Mn concentrations during westerly and southwesterly winds, which carried
road dust particles from the major freeway (i.e., I-110) to our sampling site. In contrast to the Fe
99
diurnal variations, we did not observe a second peak in late afternoon/early evening hours;
however, the Mn concentrations remained relatively high during the daytime hours (slightly less
than 20 ng/m
3
), which is the time period corresponding to maximum temperatures and wind
speeds and minimum RHs, as shown in Figures 3(a-c), indicating the probable influence of the
soil dust factor. This finding is in concert with previous studies performed in central Los
Angeles, showing the major influence of soil dust on coarse PM-bound Mn concentrations
(Pakbin et al. 2010; Pakbin et al. 2011; Cheung et al. 2012; Shirmohammadi et al. 2015). Mn is
therefore not dominated by mobile sources like Fe, and at least a major fraction of it in this size
range probably originates from resuspension of soil dust. This is also confirmed by the results of
the weekday/weekend analysis (Figure 7), which did not reveal any significant differences
between Mn concentrations on weekdays compared to weekends.
Figure 5(d) indicates the diurnal variations of measured concentrations of Cr during the study
period. Two distinct peaks were observed, one during morning rush hours (around 9 am) and one
during afternoon hours (3-5 pm), bringing the average Cr concentration up to around 12 ng/m
3
,
compared to the lower values of around 4-6 ng/m
3
during other hours of the day. Although these
two peaks can be partly attributed to the impact of re-suspended road dust from 110 freeway,
coarse PM-bound Cr may also be affected by other smaller local sources supported also by the
association between Cr concentrations and wind direction data (Figure 4(d). Figures 6(a-c)
illustrate the diurnal variations of Fe, Mn, and Cr mass fractions averaged over the entire study
period. As shown in the figures, the diurnal trends for the mass fractions were almost similar to
those of airborne concentrations for all three of the metals, exhibiting major peaks in early
morning for all the metals and a second peak in late afternoon/early evening hours for Fe.
However, it is apparent from the diurnal trends of the mass fractions of these metals that the first
100
peak occurs earlier in the morning (4-6 am versus 8-10 am), which is anti-correlated with coarse
PM mass concentrations, implying the dilution of these metals by a less metal-enriched phase
during morning rush hours.
Figure 7 presents a weekday/weekend analysis for Fe, Mn, and Cr concentrations at the sampling
site over the entire study period. As can be seen from the figure, the concentration of Fe was
significantly higher during the weekdays (about 75 ng/m
3
) compared to the weekends (around 33
ng/m
3
) (P<0.001), suggesting the major role of traffic in elevating Fe concentrations, which is
consistent with the interpretation of the data in the metal rose plots as well as in the diurnal
variation plots (Figures 5 and 6). However, the Mn and Cr concentrations were not statistically
different (P>0.05) between the weekdays and weekends (averaging about 15 ng/m
3
for Mn and
about 7 ng/m
3
for Cr).

101
Figure 4.5(a-d): Diurnal variations of the coarse PM mass concentrations and the online-
measured metal concentrations averaged over the study period; a) coarse PM; b) Fe; c) Mn; and
d) Cr. For panel a, the error bars represent one standard error (SE) of the mean. The box plots
represent the interquartile range (the bottom and top lines of the box representing the first and the
third quartiles, respectively). The line inside the box represents the median, while the whiskers
above and below the box represent the 95
th
and 5
th
percentiles.

a) b)

c) d)

0
2
4
6
8
10
12
14
16
18
20
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour of the Day
Coarse PM mass concentration
(µg/m
3
)
102
Figure 4.6(a-c): Diurnal variations of the mass fractions of the online-measured metal
concentrations averaged over the study period; a) Fe; b) Mn; c) Cr. Error bars represent one
standard error (SE) of the mean.
a)
0
1
2
3
4
5
6
7
8
2 4 6 8 10 12 14 16 18 20 22 24
Hour of the Day
Fe mass fraction (ng/µg)

b)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2 4 6 8 10 12 14 16 18 20 22 24
Hour of the Day
Mn mass fraction (ng/µg)

c)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
2 4 6 8 10 12 14 16 18 20 22 24
Hour of the Day
Cr mass fraction (ng/µg)

103
Figure 4.7: Weekday-weekend analysis of the metallic species concentrations during the study
period. The P-values shown on the graph correspond to independent samples T-test.
0
10
20
30
40
50
60
70
80
90
Fe Mn Cr
Ambient concentraion (ng/m
3
)
Weekday Weekend
P<0.001
P=0.12
P=0.18

4.4. Summary and conclusions
In this study, a novel monitor was developed for near-continuous measurement of Fe, Mn, and
Cr concentrations in ambient coarse PM (i.e., PM
10-2.5
) with a time resolution of 2 hr. This
monitor consisted of two modules, the first one utilizing two VIs connected to a modified
BioSampler (i.e., liquid impinger) to draw and concentrate coarse PM into slurry samples, and
the second one utilizing a Micro Volume Flow Cell (MVFC) coupled with spectrophotometry to
measure metal concentrations. Our results indicate that all major species of the target metals can
be efficiently digested by directly collecting ambient coarse PM into aqueous slurries. The
validity of the online measurements was investigated using parallel time integrated filter-
collected data obtained from ICP-MS analysis, which indicated very good agreement between
the online measurements and parallel filters. The average concentrations of the three metals
104
during the study period were 57.8 ng/m
3
, 15.0 ng/m
3
, and 6.9 ng/m
3
for Fe, Mn, and Cr,
respectively, consistent with published data from the study site. Diurnal variations of these
metals generally followed that of coarse PM mass concentrations, suggesting the influence of
meteorological conditions and traffic sources on these metals in coarse PM. Results from the
present study indicate that the developed monitor is capable of achieving measurements with
high accuracy and reliability over long sampling periods with minimum supervision. These
features, combined with the unique abilities of the system to measure water-soluble and different
oxidation states of these metals make it a promising technology to achieve near-continuous
measurements of metal concentrations in ambient coarse PM, enabling a better understanding of
the atmospheric processes and sources involved in formation and transport of these redox-active
metals in the coarse PM size range.
105
Chapter 5:
Conclusions and recommendations
The results from these studies provide insight into the physico-chemical characteristics and
sources of particle number and mass concentrations and how atmospheric conditions can impact
the chemical characterization and toxicity of ambient PM, and how we can improve our
understanding of sources contributing to PM levels in an area by developing techniques for
measurement of ambient PM components with fine time resolutions.
Results from the first study indicated that in central Los Angeles, traffic sources are the major
contributor to ambient PM number concentrations, with a total contribution of more than 60% to
the total number concentrations. Nucleation was the second major contributor (17%) to ambient
PM number concentrations in central Los Angeles. This underscores the major impact of traffic
sources in an urban environment, like Los Angeles, because almost all of the precursors of the
nucleation particles also come from traffic sources. This also shows that although stringent
regulations in the State of California have been able to significantly reduce the emissions from
vehicles in terms of PM mass, traffic still plays the most important role in driving the
concentrations of particle number. Given that particle number are a good representative of
ultrafine (UFP) particles, and that UFPs have the ability to penetrate deep into the lung and cause
significant health impacts, it might be prudent to devise additional standards to control vehicular
emissions in terms of particle number as well, something that is already in place in some parts of
Europe. In terms of future research, it would be very interesting to further evaluate the
relationship between sources of particle number concentrations and individual health points to be
able to evaluate source-specific health outcomes due to exposure to ambient PM number.
Another idea to explore would be repeating the same study near major sources of PM emissions,
106
for instance near the LAX airport or ports of Los Angeles and Long Beach to explore how the
size distributions and source contributions would vary near these major sources, especially in
case of the airport since airplane emissions are quite rich in UFPs.
In the second study, for the first time, we indicated how formation of fog in the Po Valley can
impact the chemical characteristics and toxicity of ambient PM. It would be worthwhile to look
further into this topic by repeating similar studies in other locations around the world, such as
Los Angeles, that experience severe fog episodes during colder months of the year. This would
be quite important in terms of the public health impacts, since we showed in this study that the
occurrence of fog episodes can enhance the toxicity of ambient PM, and leaving more toxic PM
in the atmosphere even when the fog water is evaporated.
And, finally, research presented in chapter 4 (third study) indicated how we can improve our
understanding of the diurnal variations of PM-bound redox-active metals by developing and
deploying a novel monitor capable of time-resolved measurements. This is because particles are
emitted and formed in the atmosphere at time resolutions ranging from minutes to hours, and we
cannot fully understand the source contributions and formation mechanisms using conventional
filter-based, time-integrated measurements of chemical components having time resolutions of
24 hours and longer. These time-resolved data have a great potential to be used in source
apportionment studies to more precisely determine the sources of these metals in a target area.
Therefore, they deserve a follow-up source apportionment study using data collected using this
monitor.

107
Bibliography
Alleman, L.Y., Lamaison, L., Perdrix, E., Robache, A., Galloo, J.-C., 2010. PM10 metal
concentrations and source identification using positive matrix factorization and wind
sectoring in a French industrial zone. Atmospheric Research 96, 612-625.
Allen, A.G., Nemitz, E., Shi, J.P., Harrison, R.M., Greenwood, J.C., 2001. Size distributions of
trace metals in atmospheric aerosols in the United Kingdom. Atmospheric Environment 35,
4581-4591.
Antinolo, M., Willis, M.D., Zhou, S.M., Abbatt, J.P.D., 2015. Connecting the oxidation of soot
to its redox cycling abilities. Nature Communications 6.
Argyropoulos, G., Besis, A., Voutsa, D., Samara, C., Sowlat, M.H., Hasheminassab, S., Sioutas,
C., 2016. Source apportionment of the redox activity of urban quasi-ultrafine particles
(PM0.49) in Thessaloniki following the increased biomass burning due to the economic crisis
in Greece. Science of the Total Environment 568, 124-136.
Arhami, M., Minguillón, M.C., Polidori, A., Schauer, J.J., Delfino, R.J., Sioutas, C., 2010.
Organic compound characterization and source apportionment of indoor and outdoor
quasi ‐ultrafine particulate matter in retirement homes of the Los Angeles Basin. Indoor Air
20, 17-30.
Ayres, J.G., Borm, P., Cassee, F.R., Castranova, V., Donaldson, K., Ghio, A., Harrison, R.M.,
Hider, R., Kelly, F., Kooter, I.M., 2008. Evaluating the toxicity of airborne particulate matter
and nanoparticles by measuring oxidative stress potential—a workshop report and consensus
statement. Inhalation Toxicology 20, 75-99.
Baldocchi, D., Waller, E., 2014. Winter fog is decreasing in the fruit growing region of the
Central Valley of California. Geophysical Research Letters 41, 3251-3256.
Bates, J.T., Weber, R.J., Abrams, J., Verma, V., Fang, T., Klein, M., Strickland, M.J., Sarnat,
S.E., Chang, H.H., Mulholland, J.A., Tolbert, P.E., Russell, A.G., 2015. Reactive Oxygen
Species Generation Linked to Sources of Atmospheric Particulate Matter and
Cardiorespiratory Effects. Environmental Science & Technology 49, 13605-13612.
Becker, S., Mundandhara, S., Devlin, R.B., Madden, M., 2005. Regulation of cytokine
production in human alveolar macrophages and airway epithelial cells in response to ambient
air pollution particles: further mechanistic studies. Toxicology and applied pharmacology
207, 269-275.
Becker, S., Soukup, J., 2003. Coarse (PM2.5-10), Fine (PM2.5), and ultrafine air pollution
particles induce/increase immune costimulatory receptors on human blood-derived
monocytes but not on alveolar macrophages. Journal of Toxicology and Environmental
Health Part A 66, 847-859.
Beddows, D.C.S., Harrison, R.M., Green, D.C., Fuller, G.W., 2015. Receptor modelling of both
particle composition and size distribution from a background site in London, UK.
Atmospheric Chemistry and Physics Discussions 15, 10123-10162.
108
Birmili, W., Allen, A.G., Bary, F., Harrison, R.M., 2006. Trace metal concentrations and water
solubility in size-fractionated atmospheric particles and influence of road traffic.
Environmental Science & Technology 40, 1144-1153.
Brines, M., Dall'Osto, M., Beddows, D.C.S., Harrison, R.M., Gómez-Moreno, F., Núñez, L.,
Artíñano, B., Costabile, F., Gobbi, G.P., Salimi, F., 2015. Traffic and nucleation events as
main sources of ultrafine particles in high-insolation developed world cities. Atmospheric
Chemistry and Physics 15, 5929-5945.
Brook, R.D., Rajagopalan, S., Pope, C.A., Brook, J.R., Bhatnagar, A., Diez-Roux, A.V.,
Holguin, F., Hong, Y., Luepker, R.V., Mittleman, M.A., 2010. Particulate matter air
pollution and cardiovascular disease an update to the scientific statement from the American
Heart Association. Circulation 121, 2331-2378.
Brown, S.G., Eberly, S., Paatero, P., Norris, G.A., 2015. Methods for estimating uncertainty in
PMF solutions: Examples with ambient air and water quality data and guidance on reporting
PMF results. Science of the Total Environment 518, 626-635.
Brunekreef, B., Forsberg, B., 2005. Epidemiological evidence of effects of coarse airborne
particles on health. European Respiratory Journal 26, 309-318.
Cassee, F.R., Heroux, M.E., Gerlofs-Nijland, M.E., Kelly, F.J., 2013. Particulate matter beyond
mass: recent health evidence on the role of fractions, chemical constituents and sources of
emission. Inhalation Toxicology 25, 802-812.
Cermak, J., Eastman, R.M., Bendix, J., Warren, S.G., 2009. European climatology of fog and
low stratus based on geostationary satellite observations. Quarterly Journal of the Royal
Meteorological Society 135, 2125-2130.
Charrier, J.G., Anastasio, C., 2012. On dithiothreitol (DTT) as a measure of oxidative potential
for ambient particles: evidence for the importance of soluble transition metals. Atmospheric
Chemistry and Physics Discussions 12, 11317-11350.
Chen, G., Ziemba, L.D., Chu, D.A., Thornhill, K.L., Schuster, G.L., Winstead, E.L., Diskin,
G.S., Ferrare, R.A., Burton, S.P., Ismail, S., 2011. Observations of Saharan dust
microphysical and optical properties from the Eastern Atlantic during NAMMA airborne
field campaign. Atmospheric Chemistry and Physics 11, 723-740.
Chen, L.C., Peoples, S.M., Amdur, M.O., 1991. Pulmonary effects of sulfur oxides on the
surface of copper oxide aerosol. The American Industrial Hygiene Association Journal 52,
187-191.
Chester, R., Stoner, J.H., 1974. The distribution of Mn, Fe, Cu, Ni, Co, Ga, Cr, V, Ba, Sr, Sn, Zn,
and Pb, in some soil-sized particulates from the lower troposphere over the world ocean.
Marine Chemistry 2, 157-188.
Cheung, K., Daher, N., Shafer, M.M., Ning, Z., Schauer, J.J., Sioutas, C., 2011. Diurnal trends in
coarse particulate matter composition in the Los Angeles Basin. Journal of Environmental
Monitoring 13, 3277-3287.
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. Environmental Science & Technology 46, 3779-3787.
109
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. Environmental Research 99, 40-47.
Claiborn, C.S., Larson, T., Sheppard, L., 2002. Testing the metals hypothesis in Spokane,
Washington. Environmental Health Perspectives 110, 547.
Collett, J.L., Herckes, P., Youngster, S., Lee, T., 2008. Processing of atmospheric organic matter
by California radiation fogs. Atmospheric Research 87, 232-241.
Covert, D.S., Heintzenberg, J., Hansson, H.-C., 1990. Electro-optical detection of external
mixtures in aerosols. Aerosol Science and Technology 12, 446-456.
Crippa, M., Canonaco, F., Lanz, V.A., Aijala, M., Allan, J.D., Carbone, S., Capes, G., Ceburnis,
D., Dall'Osto, M., Day, D.A., DeCarlo, P.F., Ehn, M., Eriksson, A., Freney, E., Hildebrandt
Ruiz, L., Hillamo, R., Jimenez, J.L., Junninen, H., Kiendler-Scharr, A., Kortelainen, A.M.,
Kulmala, M., Laaksonen, A., Mensah, A., Mohr, C., Nemitz, E., O'Dowd, C., Ovadnevaite,
J., Pandis, S.N., Petaja, T., Poulain, L., Saarikoski, S., Sellegri, K., Swietlicki, E., Tiitta, P.,
Worsnop, D.R., Baltensperger, U., Prevot, A.S.H., 2014. Organic aerosol components
derived from 25 AMS data sets across Europe using a consistent ME-2 based source
apportionment approach. Atmospheric Chemistry and Physics 14, 6159-6176.
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., 2012. Characterization, sources and redox
activity of fine and coarse particulate matter in Milan, Italy. Atmospheric Environment 49,
130-141.
Dall'Osto, M., Beddows, D.C.S., Pey, J., Rodriguez, S., Alastuey, A., Harrison, R.M., Querol,
X., 2012. Urban aerosol size distributions over the Mediterranean city of Barcelona, NE
Spain. Atmospheric Chemistry and Physics 12, 10693-10707.
Davis, D.A., Bortolato, M., Godar, S.C., Sander, T.K., Iwata, N., Pakbin, P., Shih, J.C., Berhane,
K., McConnell, R., Sioutas, C., 2013a. Prenatal exposure to urban air nanoparticles in mice
causes altered neuronal differentiation and depression-like responses. PLoS One 8, e64128.
Davis, D.A., Bortolato, M., Godar, S.C., Sander, T.K., Iwata, N., Pakbin, P., Shih, J.C., Berhane,
K., McConnell, R., Sioutas, C., 2013b. Prenatal exposure to urban air nanoparticles in mice
causes altered neuronal differentiation and depression-like responses.
Davis, D.A., Bortolato, M., Godar, S.C., Sander, T.K., Iwata, N., Pakbin, P., Shih, J.C., Berhane,
K., McConnell, R., Sioutas, C., Finch, C.E., Morgan, T.E., 2013c. Prenatal Exposure to
Urban Air Nanoparticles in Mice Causes Altered Neuronal Differentiation and Depression-
Like Responses. Plos One 8.
Decesari, S., Sowlat, M.H., Hasheminassab, S., Sandrini, S., Gilardoni, S., Facchini, M.C.,
Fuzzi, S., Sioutas, C., 2017. Enhanced toxicity of aerosol in fog conditions in the Po Valley,
Italy. Atmospheric Chemistry and Physics 17, 7721-7731.
Delfino, R.J., Sioutas, C., Malik, S., 2005. Potential role of ultrafine particles in associations
between airborne particle mass and cardiovascular health. Environmental Health
Perspectives, 934-946.
110
Delfino, R.J., Staimer, N., Tjoa, T., Arhami, M., Polidori, A., Gillen, D.L., George, S.C., Shafer,
M.M., Schauer, J.J., Sioutas, C., 2010a. Associations of Primary and Secondary Organic
Aerosols With Airway and Systemic Inflammation in an Elderly Panel Cohort. Epidemiology
21, 892-902.
Delfino, R.J., Tjoa, T., Gillen, D.L., Staimer, N., Polidori, A., Arhami, M., Jamner, L., Sioutas,
C., Longhurst, J., 2010b. Traffic-related Air Pollution and Blood Pressure in Elderly Subjects
With Coronary Artery Disease. Epidemiology (Cambridge, Mass.) 21,
10.1097/EDE.1090b1013e3181d1095e1019b.
Dockery, D.W., Stone, P.H., 2007. Cardiovascular risks from fine particulate air pollution. New
England Journal of Medicine 356, 511-513.
Donaldson, K., Li, X.Y., MacNee, W., 1998. Ultrafine (nanometre) particle mediated lung
injury. Journal of Aerosol Science 29, 553-560.
Donaldson, K., Stone, V., Borm, P.J.A., Jimenez, L.A., Gilmour, P.S., Schins, R.P.F., Knaapen,
A.M., Rahman, I., Faux, S.P., Brown, D.M., 2003. Oxidative stress and calcium signaling in
the adverse effects of environmental particles (PM10). Free Radical Biology and Medicine
34, 1369-1382.
Dreher, K.L., Jaskot, R.H., Lehmann, J.R., Richards, J.H., Ghio, J.K.M.A.J., Costa, D.L., 1997.
Soluble transition metals mediate residual oil fly ash induced acute lung injury. Journal of
Toxicology and Environmental Health Part A 50, 285-305.
Du, L., Su, Y.Y., Sun, D.B., Zhu, W.H., Wang, J.Y., Zhuang, X.H., Zhou, S.N., Lu, Y.J., 2008.
Formic acid induces Yca1p-independent apoptosis-like cell death in the yeast Saccharomyces
cerevisiae. Fems Yeast Research 8, 531-539.
Dutton, S.J., Vedal, S., Piedrahita, R., Milford, J.B., Miller, S.L., Hannigan, M.P., 2010. Source
apportionment using positive matrix factorization on daily measurements of inorganic and
organic speciated PM2.5. Atmospheric Environment 44, 2731-2741.
Ebert, M., Weinbruch, S., Hoffmann, P., Ortner, H.M., 2004. The chemical composition and
complex refractive index of rural and urban influenced aerosols determined by individual
particle analysis. Atmospheric Environment 38, 6531-6545.
Ervens, B., Turpin, B.J., Weber, R.J., 2011. Secondary organic aerosol formation in cloud
droplets and aqueous particles (aqSOA): a review of laboratory, field and model studies.
Atmospheric Chemistry and Physics 11, 11069-11102.
Facchini, M.C., Fuzzi, S., Zappoli, S., Andracchio, A., Gelencser, A., Kiss, G., Krivacsy, Z.,
Meszaros, E., Hansson, H.C., Alsberg, T., Zebuhr, Y., 1999. Partitioning of the organic
aerosol component between fog droplets and interstitial air. Journal of Geophysical
Research-Atmospheres 104, 26821-26832.
Fang, T., Guo, H.Y., Zeng, L.H., Verma, V., Nenes, A., Weber, R.J., 2017. Highly Acidic
Ambient Particles, Soluble Metals, and Oxidative Potential: A Link between Sulfate and
Aerosol Toxicity. Environmental Science & Technology 51, 2611-2620.
Fang, T., Verma, V., Bates, J.T., Abrams, J., Klein, M., Strickland, M.J., Sarnat, S.E., Chang,
H.H., Mulholland, J.A., Tolbert, P.E., 2015a. Oxidative potential of ambient water-soluble
PM2.5 measured by Dithiothreitol (DTT) and Ascorbic Acid (AA) assays in the southeastern
111
United States: contrasts in sources and health associations. Atmospheric Chemistry and
Physics Discussions 15, 30609-30644.
Fang, T., Verma, V., Guo, H., King, L.E., Edgerton, E.S., Weber, R.J., 2015b. A semi-automated
system for quantifying the oxidative potential of ambient particles in aqueous extracts using
the dithiothreitol (DTT) assay: results from the Southeastern Center for Air Pollution and
Epidemiology (SCAPE). Atmospheric Measurement Techniques 8, 471-482.
Friend, A.J., Ayoko, G.A., Jager, D., Wust, M., Jayaratne, E.R., Jamriska, M., Morawska, L.,
2013. Sources of ultrafine particles and chemical species along a traffic corridor: comparison
of the results from two receptor models. Environmental Chemistry 10, 54-63.
Friend, A.J., Ayoko, G.A., Jayaratne, E.R., Jamriska, M., Hopke, P.K., Morawska, L., 2012.
Source apportionment of ultrafine and fine particle concentrations in Brisbane, Australia.
Environmental Science and Pollution Research 19, 2942-2950.
Fuzzi, S., Orsi, G., Bonforte, G., Zardini, B., Franchini, P.L., 1997. An automated fog water
collector suitable for deposition networks: Design, operation and field tests. Water Air and
Soil Pollution 93, 383-394.
Fuzzi, S., Orsi, G., Nardini, G., Facchini, M.C., McLaren, S., McLaren, E., Mariotti, M., 1988.
HETEROGENEOUS PROCESSES IN THE PO VALLEY RADIATION FOG. Journal of
Geophysical Research-Atmospheres 93, 11141-11151.
Gauderman, W.J., Urman, R., Avol, E., Berhane, K., McConnell, R., Rappaport, E., Chang, R.,
Lurmann, F., Gilliland, F., 2015. Association of improved air quality with lung development
in children. New England Journal of Medicine 372, 905-913.
Geller, M.D., Fine, P.M., Sioutas, C., 2004. The relationship between real-time and time-
integrated coarse (2.5–10 μm), intermodal (1–2.5 μm), and fine (< 2.5 μm) particulate matter
in the Los Angeles Basin. Journal of the Air & Waste Management Association 54, 1029-
1039.
Gerber, H., 1991. Direct measurement of suspended particulate volume concentration and far-
infrared extinction coefficient with a laser-diffraction instrument. Applied Optics 30, 4824-
4831.
Gietl, J.K., Lawrence, R., Thorpe, A.J., Harrison, R.M., 2010. Identification of brake wear
particles and derivation of a quantitative tracer for brake dust at a major road. Atmospheric
Environment 44, 141-146.
Gilardoni, S., Massoli, P., Giulianelli, L., Rinaldi, M., Paglione, M., Pollini, F., Lanconelli, C.,
Poluzzi, V., Carbone, S., Hillamo, R., Russell, L.M., Facchini, M.C., Fuzzi, S., 2014. Fog
scavenging of organic and inorganic aerosol in the Po Valley. Atmospheric Chemistry and
Physics 14, 6967-6981.
Gilardoni, S., Massoli, P., Paglione, M., Giulianelli, L., Carbone, C., Rinaldi, M., Decesari, S.,
Sandrini, S., Costabile, F., Gobbi, G.P., Pietrogrande, M.C., Visentin, M., Scotto, F., Fuzzi,
S., Facchini, M.C., 2016. Direct observation of aqueous secondary organic aerosol from
biomass-burning emissions. Proceedings of the National Academy of Sciences of the United
States of America 113, 10013-10018.
112
Giulianelli, L., Gilardoni, S., Tarozzi, L., Rinaldi, M., Decesari, S., Carbone, C., Facchini, M.C.,
Fuzzi, S., 2014. Fog occurrence and chemical composition in the Po valley over the last
twenty years. Atmospheric Environment 98, 394-401.
Gu, J., Pitz, M., Schnelle-Kreis, J., Diemer, J., Reller, A., Zimmermann, R., Soentgen, J.,
Stoelzel, M., Wichmann, H.E., Peters, A., 2011. Source apportionment of ambient particles:
comparison of positive matrix factorization analysis applied to particle size distribution and
chemical composition data. Atmospheric Environment 45, 1849-1857.
Hallberg, A., Ogren, J.A., Noone, K.J., Heintzenberg, J., Berner, A., Solly, I., Kruisz, C.,
Reischl, G., Fuzzi, S., Facchini, M.C., Hansson, H.C., Wiedensohler, A., Svenningsson, I.B.,
1992. Phase partitioning for different aerosol species in fog. Tellus Series B-Chemical and
Physical Meteorology 44, 545-555.
Ham, W.A., Kleeman, M.J., 2011. Size-resolved source apportionment of carbonaceous
particulate matter in urban and rural sites in central California. Atmospheric Environment 45,
3988-3995.
Hänel, G., 1968. The real part of the mean complex refractive index and the mean density of
samples of atmospheric aerosol particles. Tellus 20, 371-379.
Harrison, R.M., Beddows, D.C.S., Dall’Osto, M., 2011. PMF analysis of wide-range particle size
spectra collected on a major highway. Environmental Science & Technology 45, 5522-5528.
Harrison, R.M., Jones, A.M., Beddows, D.C.S., Dall'Osto, M., Nikolova, I., 2016. Evaporation
of traffic-generated nanoparticles during advection from source. Atmospheric Environment
125, Part A, 1-7.
Harrison, R.M., Jones, A.M., Gietl, J., Yin, J., Green, D.C., 2012. Estimation of the contributions
of brake dust, tire wear, and resuspension to nonexhaust traffic particles derived from
atmospheric measurements. Environmental Science & Technology 46, 6523-6529.
Harrison, R.M., Yin, J., Mark, D., Stedman, J., Appleby, R. S.,, Booker, J., and Moorcroft, S.,
2001. Studies of the coarse particle (2.5–10 μm) component in UK urban atmospheres.
Atmospheric Environment 35, 3667–3679.
Hasheminassab, S., Daher, N., Saffari, A., Wang, D., Ostro, B.D., Sioutas, C., 2014a. Spatial and
temporal variability of sources of ambient fine particulate matter (PM2.5) in California.
Atmospheric Chemistry andPhysics 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. Atmospheric Environment 79, 529-539.
Hasheminassab, S., Pakbin, P., Delfino, R.J., Schauer, J.J., Sioutas, C., 2014b. Diurnal and
seasonal trends in the apparent density of ambient fine and coarse particles in Los Angeles.
Environmental Pollution 187, 1-9.
Hassanvand, M.S., Naddafi, K., Faridi, S., Arhami, M., Nabizadeh, R., Sowlat, M.H., Pourpak,
Z., Rastkari, N., Momeniha, F., Kashani, H., 2014. Indoor/outdoor relationships of PM10,
PM2.5, and PM1 mass concentrations and their water-soluble ions in a retirement home and a
school dormitory. Atmospheric Environment 82, 375-382.
113
Hassanvand, M.S., Naddafi, K., Faridi, S., Nabizadeh, R., Sowlat, M.H., Momeniha, F.,
Gholampour, A., Arhami, M., Kashani, H., Zare, A., 2015. Characterization of PAHs and
metals in indoor/outdoor PM10/PM2. 5/PM1 in a retirement home and a school dormitory.
Science of the Total Environment 527, 100-110.
Heal, M.R., Hibbs, L.R., Agius, R.M., Beverland, I.J., 2005. Total and water-soluble trace metal
content of urban background PM10, PM2.5 and black smoke in Edinburgh, UK. Atmospheric
Environment 39, 1417-1430.
Henry, R.C., 1987. Current factor-analysis receptor models are ill-posed. Atmospheric
Environment 21, 1815-1820.
Hering, S.V., McMurry, P.H., 1991. Optical counter response to monodisperse atmospheric
aerosols. Atmospheric Environment. Part A. General Topics 25, 463-468.
Hersey, S.P., Craven, J.S., Schilling, K.A., Metcalf, A.R., Sorooshian, A., Chan, M.N., Flagan,
R.C., Seinfeld, J.H., 2011. The Pasadena Aerosol Characterization Observatory (PACO):
chemical and physical analysis of the Western Los Angeles basin aerosol. Atmospheric
Chemistry and Physics 11, 7417-7443.
Hu, D., Bian, Q.J., Lau, A.K.H., Yu, J.Z., 2010. Source apportioning of primary and secondary
organic carbon in summer PM2.5 in Hong Kong using positive matrix factorization of
secondary and primary organic tracer data. Journal of Geophysical Research-Atmospheres
115.
Hudda, N., Cheung, K., Moore, K.F., Sioutas, C., 2010. Inter-community variability in total
particle number concentrations in the eastern Los Angeles air basin. Atmospheric Chemistry
and Physics 10, 11385-11399.
Hwang, I., Hopke, P.K., 2006. Comparison of source apportionments of fine particulate matter at
two San Jose Speciation Trends Network sites. Journal of the Air & Waste Management
Association 56, 1287-1300.
Janssen, N.A.H., Hoek, G., Simic-Lawson, M., Fischer, P., van Bree, L., ten Brink, H., Keuken,
M., Atkinson, R.W., Anderson, H.R., Brunekreef, B., Cassee, F.R., 2011. Black Carbon as an
Additional Indicator of the Adverse Health Effects of Airborne Particles Compared with
PM10 and PM2.5. Environmental Health Perspectives 119, 1691-1699.
Javid, A., Hamedian, A.A., Gharibi, H. and Sowlat, M.H., 2016. Towards the Application of
Fuzzy Logic for Developing a Novel Indoor Air Quality Index (FIAQI). Iranian Journal of
Public Health 45(2), 203–213.
Karlsson, A., Irgum, K., Hansson, H.-C., 1997. Single-stage flowing liquid film impactor for
continuous on-line particle analysis. Journal of aerosol science 28, 1539-1551.
Kasumba, J., Hopke, P.K., Chalupa, D.C., Utell, M.J., 2009. Comparison of sources of
submicron particle number concentrations measured at two sites in Rochester, NY. Science
of the Total Environment 407, 5071-5084.
Kent, G.S., Yue, G.K., Farrukh, U.O., Deepak, A., 1983. Modeling atmospheric aerosol
backscatter at CO 2 laser wavelengths. 1: Aerosol properties, modeling techniques, and
associated problems. Applied Optics 22, 1655-1665.
114
Khlystov, A., Ma, Y., 2006. An on-line instrument for mobile measurements of the spatial
variability of hexavalent and trivalent chromium in urban air. Atmospheric Environment 40,
8088-8093.
Kiesewetter, G., Amann, M., 2014. Urban PM2.5 levels under the EU Clean Air Policy Package.
TSAP Report N.12. International Institute for Applied Systems Analysis (IIASA),
Laxenburg, Austria.
Kim, E., Hopke, P.K., 2007. Source characterization of ambient fine particles in the Los Angeles
basin. Journal of Environmental Engineering and Science 6, 343-353.
Kim, E., Hopke, P.K., Larson, T.V., Covert, D.S., 2004. Analysis of ambient particle size
distributions using unmix and positive matrix factorization. Environmental Science &
Technology 38, 202-209.
Kim, E., Turkiewicz, K., Zulawnick, S.A., Magliano, K.L., 2010. Sources of fine particles in the
South Coast area, California. Atmospheric Environment 44, 3095-3100.
Kim, S., Jaques, P.A., Chang, M., Froines, J.R., Sioutas, C., 2001. Versatile aerosol
concentration enrichment system (VACES) for simultaneous in vivo and in vitro evaluation
of toxic effects of ultrafine, fine and coarse ambient particles Part I: Development and
laboratory characterization. Journal of aerosol science 32, 1281-1297.
Krecl, P., Hedberg Larsson, E., Ström, J., Johansson, C., 2008. Contribution of residential wood
combustion and other sources to hourly winter aerosol in Northern Sweden determined by
positive matrix factorization. Atmospheric Chemistry and Physics 8, 3639-3653.
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 Science and Technology 42, 946-957.
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. Environmental Health Perspectives 111, 455.
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. Environmental Health Perspectives 117, 1116-1123.
Lighty, J.S., Veranth, J.M., Sarofim, A.F., 2000. Combustion aerosols: factors governing their
size and composition and implications to human health. Journal of the Air & Waste
Management Association 50, 1565-1618.
Lim, J.-M., Lee, J.-H., Moon, J.-H., Chung, Y.-S., Kim, K.-H., 2010. Source apportionment of
PM10 at a small industrial area using Positive Matrix Factorization. Atmospheric Research
95, 88-100.
Lim, S.S., Vos, T., Flaxman, A.D., Danaei, G., Shibuya, K., Adair-Rohani, H., AlMazroa, M.A.,
Amann, M., Anderson, H.R., Andrews, K.G., 2013. A comparative risk assessment of burden
of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions,
1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. The lancet
380, 2224-2260.
115
Liu, Z.R., Hu, B., Liu, Q., Sun, Y., Wang, Y.S., 2014. Source apportionment of urban fine
particle number concentration during summertime in Beijing. Atmospheric Environment 96,
359-369.
Lovett, C., Shirmohammadi, F., Sowlat, M.H., Sioutas, C., 2018a. Commuting in Los Angeles:
cancer and non-cancer health risks of roadway, light-rail and subway transit routes. Aerosol
and Air Quality Research 18, 2363-2374.
Lovett, C., Sowlat, M.H., Saliba, N.A., Shihadeh, A.L., Sioutas, C., 2018b. Oxidative potential
of ambient particulate matter in Beirut during Saharan and Arabian dust events. Atmospheric
Environment 188, 34-42.
Majestic, B.J., Schauer, J.J., Shafer, M.M., 2007. Development of a manganese speciation
method for atmospheric aerosols in biologically and environmentally relevant fluids. Aerosol
Science and Technology 41, 925-933.
Majestic, B.J., Schauer, J.J., Shafer, M.M., Turner, J.R., Fine, P.M., Singh, M., Sioutas, C., 2006.
Development of a wet-chemical method for the speciation of iron in atmospheric aerosols.
Environmental Science & Technology 40, 2346-2351.
Mancinelli, V., Decesari, S., Facchini, M.C., Fuzzi, S., Mangani, F., 2005. Partitioning of metals
between the aqueous phase and suspended insoluble material in fog droplets. Annali Di
Chimica 95, 275-290.
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. Environmental Science & Technology 47, 3362-3369.
McWhinney, R.D., Gao, S.S., Zhou, S.M., Abbatt, J.P.D., 2011. Evaluation of the Effects of
Ozone Oxidation on Redox-Cycling Activity of Two-Stroke Engine Exhaust Particles.
Environmental Science & Technology 45, 2131-2136.
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. Atmospheric Environment 44,
2224-2230.
Miller, K.A., Siscovick, D.S., Sheppard, L., Shepherd, K., Sullivan, J.H., Anderson, G.L.,
Kaufman, J.D., 2007. Long-term exposure to air pollution and incidence of cardiovascular
events in women. New England Journal of Medicine 356, 447-458.
Misra, C., Geller, M.D., Shah, P., Sioutas, C., Solomon, P.A., 2001. Development and
Evaluation of a Continuous Coarse (PM10–PM25) Particle Monitor. Journal of the Air &
Waste Management Association 51, 1309-1317.
Moore, K.F., Ning, Z., Ntziachristos, L., Schauer, J.J., Sioutas, C., 2007. Daily variation in the
properties of urban ultrafine aerosol—Part I: Physical characterization and volatility.
Atmospheric Environment 41, 8633-8646.
Moreno, T., Querol, X., Alastuey, A., Viana, M., Salvador, P., De la Campa, A.S., Artiñano, B.,
De la Rosa, J., Gibbons, W., 2006. Variations in atmospheric PM trace metal content in
Spanish towns: illustrating the chemical complexity of the inorganic urban aerosol cocktail.
Atmospheric Environment 40, 6791-6803.
116
Mousavi, A., Sowlat, M.H., Hasheminassab, S., Pikelnaya, O., Polidori, A., Ban-Weiss, G.,
Sioutas, C., 2018a. Impact of particulate matter (PM) emissions from ships, locomotives, and
freeways in the communities near the ports of Los Angeles (POLA) and Long Beach (POLB)
on the air quality in the Los Angeles county. Atmospheric Environment 195, 159-169.
Mousavi, A., Sowlat, M.H., Hasheminassab, S., Polidori, A., Shafer, M.M., Schauer, J.J.,
Sioutas, C., 2019a. Impact of emissions from the Ports of Los Angeles and Long Beach on
the oxidative potential of ambient PM0.25 measured across the Los Angeles County. Science
of the Total Environment 651, 638-647.
Mousavi, A., Sowlat, M.H., Lovett, C., Rauber, M., Szidat, S., Boffi, R., Borgini, A., De Marco,
C., Ruprecht, A.A. and Sioutas, C., 2019b. Source apportionment of black carbon (BC) from
fossil fuel and biomass burning in metropolitan Milan, Italy. Atmospheric Environment 203,
252-261.
Mousavi, A., Sowlat, M.H., Hasheminassab, S., Polidori, A., Sioutas, C., 2018b. Spatio-temporal
trends and source apportionment of fossil fuel and biomass burning black carbon (BC) in the
Los Angeles Basin. Science of the Total Environment 640, 1231-1240.
Mousavi, A., Sowlat, M.H., Sioutas, C., 2018c. Diurnal and seasonal trends and source
apportionment of redox-active metals in Los Angeles using a novel online metal monitor and
Positive Matrix Factorization (PMF). Atmospheric Environment 174, 15-24.
Naddafi, K., Sowlat, M., Safari, M., 2012. Integrated assessment of air pollution in Tehran, over
the period from September 2008 to September 2009. Iranian Journal of Public Health 41, 77-
86.
Nel, A., 2005. Air pollution-related illness: Effects of particles. Science 308, 804-806.
Nel, A., Xia, T., Mädler, L., Li, N., 2006. Toxic potential of materials at the nanolevel. Science
311, 622-627.
Ning, Z., Sioutas, C., 2010. Atmospheric Processes Influencing Aerosols Generated by
Combustion and the Inference of Their Impact on Public Exposure: A Review. Aerosol and
Air Quality Research 10, 43-58.
Niu, F., Li, Z.Q., Li, C., Lee, K.H., Wang, M.Y., 2010. Increase of wintertime fog in China:
Potential impacts of weakening of the Eastern Asian monsoon circulation and increasing
aerosol loading. Journal of Geophysical Research-Atmospheres 115.
Norris, G., Duvall, R., Brown, S., Bai, S., 2014. EPA Positive Matrix Factorization (PMF) 5.0
Fundamentals and User Guide, U.S. Environmental Protection Agency, Office of Research
and Development, Washington, DC 20460.
Ntziachristos, L., Froines, J.R., Cho, A.K., Sioutas, C., 2007a. Relationship between redox
activity and chemical speciation of size-fractionated particulate matter. Particle and Fibre
Toxicology 4, 1.
Ntziachristos, L., Froines, J.R., Cho, A.K., Sioutas, C., 2007b. Relationship between redox
activity and chemical speciation of size-fractionated particulate matter. Particle and Fibre
Toxicology 4, 5-5.
117
Oakes, M., Ingall, E.D., Lai, B., Shafer, M.M., Hays, M.D., Liu, Z.G., Russell, A.G., Weber,
R.J., 2012. Iron Solubility Related to Particle Sulfur Content in Source Emission and
Ambient Fine Particles. Environmental Science and Technology 46, 6637-6644.
Oberdörster, G., Ferin, J., Lehnert, B.E., 1994. Correlation between particle size, in vivo particle
persistence, and lung injury. Environmental Health Perspectives 102, 173.
Oberdörster, G., Sharp, Z., Atudorei, V., Elder, A., Gelein, R., Lunts, A., Kreyling, W., Cox, C.,
2002. Extrapulmonary translocation of ultrafine carbon particles following whole-body
inhalation exposure of rats. Journal of Toxicology and Environmental Health Part A 65,
1531-1543.
Ogulei, D., Hopke, P.K., Chalupa, D.C., Utell, M.J., 2007. Modeling source contributions to
submicron particle number concentrations measured in Rochester, New York. Aerosol
Science and Technology 41, 179-201.
Ogulei, D., Hopke, P.K., Wallace, L.A., 2006a. Analysis of indoor particle size distributions in
an occupied townhouse using positive matrix factorization. Indoor Air 16, 204-215.
Ogulei, D., Hopke, P.K., Zhou, L., Pancras, J.P., Nair, N., Ondov, J.M., 2006b. Source
apportionment of Baltimore aerosol from combined size distribution and chemical
composition data. Atmospheric Environment 40, 396-410.
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.
Atmospheric Environment 97, 552-555.
Ospital, J., Cassmassi, J., Chico, T., 2008. Multiple air toxics exposure study in the south coast
air basin. MATES III Draft Final Report, South Coast Air Quality Management District,
Diamond Bar, CA (USA).
Paatero, P., 1997. Least squares formulation of robust non-negative factor analysis.
Chemometrics and Intelligent Laboratory Systems 37, 23-35.
Paatero, P., Eberly, S., Brown, S.G., Norris, G.A., 2014. Methods for estimating uncertainty in
factor analytic solutions. Atmospheric Measurement Techniques 7, 781-797.
Paatero, P., Tapper, U., 1993. Analysis of different modes of factor analysis as least squares fit
problems. Chemometrics and Intelligent Laboratory Systems 18, 183-194.
Paatero, P., Tapper, U., 1994. Positive matrix factorization: A non ‐negative factor model with
optimal utilization of error estimates of data values. Environmetrics 5, 111-126.
Pakbin, P., Hudda, N., Cheung, K.L., Moore, K.F., Sioutas, C., 2010. Spatial and temporal
variability of coarse (PM10−2.5) particulate matter concentrations in the Los Angeles area.
Aerosol Science and Technology 44, 514-525.
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 Science and Technology
45, 949-963.
Perez, L., Tobias, A., Querol, X., Künzli, N., Pey, J., Alastuey, A., Viana, M., Valero, N.,
González-Cabré, M., Sunyer, J., 2008. Coarse particles from Saharan dust and daily
mortality. Epidemiology 19, 800-807.
118
Pérez, N., Pey, J., Querol, X., Alastuey, A., López, J.M., Viana, M., 2008. Partitioning of major
and trace components in PM10–PM2.5–PM1 at an urban site in Southern Europe.
Atmospheric Environment 42, 1677-1691.
Peters, A., Veronesi, B., Calderón-Garcidueñas, L., Gehr, P., Chen, L.C., Geiser, M., Reed, W.,
Rothen-Rutishauser, B., Schürch, S., Schulz, H., 2006. Translocation and potential
neurological effects of fine and ultrafine particles a critical update. Particle andFfibre
Toxicology 3, 1.
Peters, A., Wichmann, H.E., Tuch, T., Heinrich, J., Heyder, J., 1997. Respiratory effects are
associated with the number of ultrafine particles. American Journal of Respiratory and
Critical Care Medicine 155, 1376-1383.
Pope, C.A., Burnett, R.T., Thun, M.J., Calle, E.E., Krewski, D., Ito, K., Thurston, G.D., 2002.
Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air
pollution. Jama-Journal of the American Medical Association 287, 1132-1141.
Pope, C.A., Burnett, R.T., Thurston, G.D., Thun, M.J., Calle, E.E., Krewski, D., Godleski, J.J.,
2004a. Cardiovascular mortality and long-term exposure to particulate air pollution -
Epidemiological evidence of general pathophysiological pathways of disease. Circulation
109, 71-77.
Pope, C.A., Burnett, R.T., Thurston, G.D., Thun, M.J., Calle, E.E., Krewski, D., Godleski, J.J.,
2004b. Cardiovascular mortality and long-term exposure to particulate air pollution
epidemiological evidence of general pathophysiological pathways of disease. Circulation
109, 71-77.
Pope Iii, C.A., Burnett, R.T., Thun, M.J., Calle, E.E., Krewski, D., Ito, K., Thurston, G.D., 2002.
Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air
pollution. JAMA 287, 1132-1141.
Propper, R., Wong, P., Bui, S., Austin, J., Vance, W., Alvarado, Á., Croes, B., Luo, D., 2015.
Ambient and Emission Trends of Toxic Air Contaminants in California. Environmental
Science & Technology 49, 11329-11339.
Putaud, J.-P., Raes, F., Van Dingenen, R., Brüggemann, E., Facchini, M.C., Decesari, S., Fuzzi,
S., Gehrig, R., Hüglin, C., Laj, P., 2004. A European aerosol phenomenology—2: chemical
characteristics of particulate matter at kerbside, urban, rural and background sites in Europe.
Atmospheric Environment 38, 2579-2595.
Putaud, J.P., Van Dingenen, R., Alastuey, A., Bauer, H., Birmili, W., Cyrys, J., Flentje, H.,
Fuzzi, S., Gehrig, R., Hansson, H.-C., 2010. A European aerosol phenomenology–3: Physical
and chemical characteristics of particulate matter from 60 rural, urban, and kerbside sites
across Europe. Atmospheric Environment 44, 1308-1320.
Rastogi, N., Oakes, M.M., Schauer, J.J., Shafer, M.M., Majestic, B.J., Weber, R.J., 2009. New
technique for online measurement of water-soluble Fe (II) in atmospheric aerosols.
Environmental Science & Technology 43, 2425-2430.
Reff, A., Eberly, S.I., Bhave, P.V., 2007. Receptor modeling of ambient particulate matter data
using positive matrix factorization: review of existing methods. Journal of the Air & Waste
Management Association 57, 146-154.
119
Reid, J.S., Cahill, T.A., Wakabayashi, P.H., Dunlap, M.R., 1994. Geometric/aerodynamic
equivalent diameter ratios of ash aggregate aerosols collected in burning Kuwaiti well fields.
Atmospheric Environment 28, 2227-2234.
Saffari, A., Daher, N., Shafer, M.M., Schauer, J.J., Sioutas, C., 2014. Seasonal and spatial
variation in dithiothreitol (DTT) activity of quasi-ultrafine particles in the Los Angeles Basin
and its association with chemical species. Journal of Environmental Science and Health Part
a-Toxic/Hazardous Substances & Environmental Engineering 49, 441-451.
Saffari, A., Hasheminassab, S., Shafer, M.M., Schauer, J.J., Chatila, T.A., Sioutas, C., 2016.
Nighttime Formation of Aqueous-Phase Secondary Organic Aerosols in Los Angeles and its
Implication for Fine Particulate Matter Composition and Oxidative Potential. Atmospheric
Environment 133, 112-122
Sanderson, P., Delgado-Saborit, J.M., Harrison, R.M., 2014. A review of chemical and physical
characterisation of atmospheric metallic nanoparticles. Atmospheric Environment 94, 353-
365.
Saraf, A.K., Bora, A.K., Das, J., Rawat, V., Sharma, K., Jain, S.K., 2011. Winter fog over the
Indo-Gangetic Plains: mapping and modelling using remote sensing and GIS. Natural
Hazards 58, 199-220.
Schauer, J.J., Cass, G.R., 2000. Source apportionment of wintertime gas-phase and particle-
phase air pollutants using organic compounds as tracers. Environmental Science &
Technology 34, 1821-1832.
Shafer, M.M., Hemming, J.D.C., Antkiewicz, D.S., Schauer, J.J., 2016. Oxidative potential of
size-fractionated atmospheric aerosol in urban and rural sites across Europe. Faraday
Discussions 189, 381-405.
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. Journal of
Environmental Monitoring 12, 704-715.
Shahsavani, A., Naddafi, K., Haghighifard, N.J., Mesdaghinia, A., Yunesian, M., Nabizadeh, R.,
Arahami, M., Sowlat, M., Yarahmadi, M., Saki, H., 2012a. The evaluation of PM10, PM2.5,
and PM1 concentrations during the Middle Eastern Dust (MED) events in Ahvaz, Iran, from
April through September 2010. Journal of Arid Environments 77, 72-83.
Shahsavani, A., Naddafi, K., Haghighifard, N.J., Mesdaghinia, A., Yunesian, M., Nabizadeh, R.,
Arhami, M., Yarahmadi, M., Sowlat, M.H., Ghani, M., 2012b. Characterization of ionic
composition of TSP and PM10 during the Middle Eastern Dust (MED) storms in Ahvaz,
Iran. Environmental Monitoring and Assessment 184, 6683-6692.
Shahsavani, A., Yarahmadi, M., Hadei, M., Sowlat, M.H., Naddafi, K., 2017. Elemental and
carbonaceous characterization of TSP and PM10 during Middle Eastern dust (MED) storms
in Ahvaz, Southwestern Iran. Environmental Monitoring and Assessment 189, 462.
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. Journal of Environmental Monitoring 5,
550-556.
120
Shirmohammadi, F., Lovett, C., Sowlat, M.H., Mousavi, A., Verma, V., Shafer, M.M., Schauer,
J.J., Sioutas, C., 2018. Chemical composition and redox activity of PM0.25 near Los Angeles
International Airport and comparisons to an urban traffic site. Science of the Total
Environment 610, 1336-1346.
Shirmohammadi, F., Sowlat, M.H., Hasheminassab, S., Saffari, A., Ban-Weiss, G., Sioutas, C.,
2017a. Emission rates of particle number, mass and black carbon by the Los Angeles
International Airport (LAX) and its impact on air quality in Los Angeles. Atmospheric
Environment 151, 82-93.
Shirmohammadi, F., Wang, D.B., Hasheminassab, S., Verma, V., Schauer, J.J., Shafer, M.M.,
Sioutas, C., 2017b. Oxidative potential of on-road fine particulate matter (PM2.5) measured
on major freeways of Los Angeles, CA, and a 10-year comparison with earlier roadside
studies. Atmospheric Environment 148, 102-114.
Singh, M., Phuleria, H.C., Bowers, K., Sioutas, C., 2006. Seasonal and spatial trends in particle
number concentrations and size distributions at the children's health study sites in Southern
California. Journal of Exposure Science and Environmental Epidemiology 16, 3-18.
Sioutas, C., Delfino, R.J., Singh, M., 2005a. Exposure assessment for atmospheric ultrafine
particles (UFPs) and implications in epidemiologic research. Environmental Health
Perspectives 113, 947-955.
Sioutas, C., Delfino, R.J., Singh, M., 2005b. Exposure assessment for atmospheric ultrafine
particles (UFPs) and implications in epidemiologic research. Environmental Health
Perspectives, 947-955.
Sofowote, U.M., Su, Y., Dabek-Zlotorzynska, E., Rastogi, A.K., Brook, J., Hopke, P.K., 2015.
Sources and temporal variations of constrained PMF factors obtained from multiple-year
receptor modeling of ambient PM2.5 data from five speciation sites in Ontario, Canada.
Atmospheric Environment 108, 140-150.
Sowlat, M.H., Gharibi, H., Yunesian, M., Mahmoudi, M.T., Lotfi, S., 2011. A novel, fuzzy-
based air quality index (FAQI) for air quality assessment. Atmospheric Environment 45,
2050-2059.
Sowlat, M.H., Hasheminassab, S., Sioutas, C., 2016a. Source apportionment of ambient particle
number concentrations in central Los Angeles using positive matrix factorization (PMF).
Atmospheric Chemistry and Physics 16, 4849-4866.
Sowlat, M.H., Naddafi, K., Yunesian, M., Jackson, P.L., Lotfi, S., Shahsavani, A., 2013. PM10
source apportionment in Ahvaz, Iran, using positive matrix factorization. CLEAN–Soil, Air,
Water 41, 1143-1151.
Sowlat, M.H., Naddafi, K., Yunesian, M., Jackson, P.L., Shahsavani, A., 2012. Source
apportionment of total suspended particulates in an arid area in southwestern Iran using
positive matrix factorization. Bulletin of Environmental Contamination and Toxicology 88,
735-740.
Sowlat, M.H., Wang, D., Simonetti, G., Shafer, M.M., Schauer, J.J., Sioutas, C., 2016b.
Development and field evaluation of an online monitor for near-continuous measurement of
iron, manganese, and chromium in coarse airborne particulate matter (PM). Aerosol Science
and Technology 50, 1306-1319.
121
Stolzenburg, M., Kreisberg, N., Hering, S., 1998. Atmospheric size distributions measured by
differential mobility optical particle size spectrometry. Aerosol Science and Technology 29,
402-418.
Strawa, A.W., Elleman, R., Hallar, A.G., Covert, D., Ricci, K., Provencal, R., Owano, T.W.,
Jonsson, H.H., Schmid, B., Luu, A.P., 2006. Comparison of in situ aerosol extinction and
scattering coefficient measurements made during the Aerosol Intensive Operating Period.
Journal of Geophysical Research: Atmospheres (1984–2012) 111.
Taghvaee, S., Sowlat, M.H., Diapouli, E., Manousakas, M.I., Vasilatou, V., Eleftheriadis, K.,
Sioutas, C., 2019. Source apportionment of the oxidative potential of fine ambient particulate
matter (PM2.5) in Athens, Greece. Science of the Total Environment 653, 1407-1416.
Taghvaee, S., Sowlat, M.H., Hassanvand, M.S., Yunesian, M., Naddafi, K., Sioutas, C., 2018a.
Source-specific lung cancer risk assessment of ambient PM2.5-bound polycyclic aromatic
hydrocarbons (PAHs) in central Tehran. Environment International 120, 321-332.
Taghvaee, S., Sowlat, M.H., Mousavi, A., Hassanvand, M.S., Yunesian, M., Naddafi, K.,
Sioutas, C., 2018b. Source apportionment of ambient PM2.5 in two locations in central
Tehran using the Positive Matrix Factorization (PMF) model. Science of the total
environment 628, 672-686.
Tao, F., Gonzalez-Flecha, B., Kobzik, L., 2003. Reactive oxygen species in pulmonary
inflammation by ambient particulates. Free Radical Biology and Medicine 35, 327-340.
Thimmaiah, D., Hovorka, J., Hopke, P.K., 2009. Source apportionment of winter submicron
Prague aerosols from combined particle number size distribution and gaseous composition
data. Aerosol and Air Quality Resarch 9, 209-236.
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 Report 10, 37-51.
Vautard, R., Yiou, P., van Oldenborgh, G.J., 2009. Decline of fog, mist and haze in Europe over
the past 30 years. Nature Geoscience 2, 115-119.
Venkatachari, P., Hopke, P.K., Grover, B.D., Eatough, D.J., 2005. Measurement of particle-
bound reactive oxygen species in Rubidoux aerosols. Journal of Atmospheric Chemistry 50,
49-58.
Venkataraman, C., 1999. Comparison of particle lung doses from the fine and coarse fractions of
urban PM-10 aerosols. Inhalation Toxicology 11, 151-169.
Verma, V., Polidori, A., Schauer, J.J., Shafer, M.M., Cassee, F.R., Sioutas, C., 2009.
Physicochemical and toxicological profiles of particulate matter in Los Angeles during the
October 2007 southern California wildfires. Environmental Science & Technology 43, 954-
960.
Verma, V., Rico-Martinez, R., Kotra, N., King, L., Liu, J., Snell, T.W., Weber, R.J., 2012a.
Contribution of Water-Soluble and Insoluble Components and Their
Hydrophobic/Hydrophilic Subfractions to the Reactive Oxygen Species-Generating Potential
of Fine Ambient Aerosols. Environmental Science & Technology 46, 11384-11392.
122
Verma, V., Rico-Martinez, R., Kotra, N., King, L., Liu, J.M., Snell, T.W., Weber, R.J., 2012b.
Contribution of Water-Soluble and Insoluble Components and Their
Hydrophobic/Hydrophilic Subfractions to the Reactive Oxygen Species-Generating Potential
of Fine Ambient Aerosols. Environmental Science & Technology 46, 11384-11392.
Viana, M., Kuhlbusch, T.A.J., Querol, X., Alastuey, A., Harrison, R.M., Hopke, P.K.,
Winiwarter, W., Vallius, M., Szidat, S., Prévôt, A.S.H., 2008. Source apportionment of
particulate matter in Europe: a review of methods and results. Journal of Aerosol Science 39,
827-849.
Vu, T.V., Delgado-Saborit, J.M., Harrison, R.M., 2015. Review: Particle number size
distributions from seven major sources and implications for source apportionment studies.
Atmospheric Environment 122, 114-132.
Wang, D., Pakbin, P., Shafer, M.M., Antkiewicz, D., Schauer, J.J., Sioutas, C., 2013.
Macrophage reactive oxygen species activity of water-soluble and water-insoluble fractions
of ambient coarse, PM2.5 and ultrafine particulate matter (PM) in Los Angeles. Atmospheric
Environment 77, 301-310.
Wang, D., Shafer, M.M., Schauer, J.J., Sioutas, C., 2015. A new technique for online
measurement of total and water-soluble copper (Cu) in coarse particulate matter (PM).
Environmental Pollution 199, 227-234.
Wang, D., Sowlat, M.H., Shafer, M.M., Schauer, J.J., Sioutas, C., 2016a. Development and
evaluation of a novel monitor for online measurement of iron, manganese, and chromium in
ambient particulate matter (PM). Science of The Total Environment 565, 123-131.
Wang, H., Naghavi, M., Allen, C., Barber, R., Carter, A., Casey, D., Charlson, F., Chen, A.,
Coates, M., Coggeshall, M., 2016b. Global, regional, and national life expectancy, all-cause
mortality, and cause-specific mortality for 249 causes of death, 1980â 2015: a systematic
analysis for the Global Burden of Disease Study 2015. The Lancet 388, 1459-1544.
Watson, J.G., Chow, J.C., Lowenthal, D.H., Stolzenburg, M.R., Kreisberg, N.M., Hering, S.V.,
2002. Particle size relationships at the Fresno supersite. Journal of the Air & Waste
Management Association 52, 822-827.
Weber, R.J., Orsini, D., Daun, Y., Lee, Y.N., Klotz, P.J., Brechtel, F., 2001. A particle-into-
liquid collector for rapid measurement of aerosol bulk chemical composition. Aerosol
Science & Technology 35, 718-727.
WHO, 2007. The world health report 2007 - A safer future: global public health security in the
21st century, Geneva.
WHO, 2013. Research for universal health coverage: World health report 2013, Geneva
Wichmann, H.-E., Spix, C., Tuch, T., Wölke, G., Peters, A., Heinrich, J., Kreyling, W.G.,
Heyder, J., 2000. Daily mortality and fine and ultrafine particles in Erfurt, Germany part I:
role of particle number and particle mass. Research report (Health Effects Institute), 5-86.
Wittkopp, S., Staimer, N., Tjoa, T., Gillen, D., Daher, N., Shafer, M., Schauer, J.J., Sioutas, C.,
Delfino, R.J., 2013. Mitochondrial Genetic Background Modifies the Relationship between
Traffic-Related Air Pollution Exposure and Systemic Biomarkers of Inflammation. Plos One
8.
123
Xue, J., Xue, W., Sowlat, M.H., Sioutas, C., Lolinco, A., Hasson, A., Kleeman, M.J., 2019.
Seasonal and Annual Source Appointment of Carbonaceous Ultrafine Particulate Matter
(PM0.1) in Polluted California Cities. Environmental Science & Technology 53, 39-49.
Yue, W., Stölzel, M., Cyrys, J., Pitz, M., Heinrich, J., Kreyling, W.G., Wichmann, H.E., Peters,
A., Wang, S., Hopke, P.K., 2008. Source apportionment of ambient fine particle size
distribution using positive matrix factorization in Erfurt, Germany. Science of the Total
Environment 398, 133-144.
Zhang, K.M., Wexler, A.S., Niemeier, D.A., Zhu, Y.F., Hinds, W.C., Sioutas, C., 2005.
Evolution of particle number distribution near roadways. Part III: Traffic analysis and on-
road size resolved particulate emission factors. Atmospheric Environment 39, 4155-4166.
Zhang, Q., Jimenez, J.L., Canagaratna, M.R., Allan, J.D., Coe, H., Ulbrich, I., Alfarra, M.R.,
Takami, A., Middlebrook, A.M., Sun, Y.L., Dzepina, K., Dunlea, E., Docherty, K., DeCarlo,
P.F., Salcedo, D., Onasch, T., Jayne, J.T., Miyoshi, T., Shimono, A., Hatakeyama, S.,
Takegawa, N., Kondo, Y., Schneider, J., Drewnick, F., Borrmann, S., Weimer, S.,
Demerjian, K., Williams, P., Bower, K., Bahreini, R., Cottrell, L., Griffin, R.J., Rautiainen,
J., Sun, J.Y., Zhang, Y.M., Worsnop, D.R., 2007. Ubiquity and dominance of oxygenated
species in organic aerosols in anthropogenically-influenced Northern Hemisphere
midlatitudes. Geophysical Research Letters 34.
Zhang, Y.X., 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. Environmental Science & Technology 42, 7502-7509.
Zhou, L., Hopke, P.K., Stanier, C.O., Pandis, S.N., Ondov, J.M., Pancras, J.P., 2005.
Investigation of the relationship between chemical composition and size distribution of
airborne particles by partial least squares and positive matrix factorization. Journal of
Geophysical Research: Atmospheres (1984–2012) 110.

Abstract (if available)

Abstract The research presented herein provides valuable information on the physical and chemical characteristics of ambient PM number and mass under different atmospheric conditions, provides insight into their sources, and finally, offers a novel approach for high time-resolution measurement of the chemical components of ambient PM that can be used in source apportionment studies. Our work begins by the study of physical characteristics and sources of ambient PM number concentrations in central Los Angeles, in a location significantly influenced by traffic-related emissions. In this study, particle number size distributions were measured over a year-long sampling campaign in central Los Angeles, CA, covering the full size spectrum from 13 nm to 10 µm. Particle number size distributions were then used as inputs to the USEPA's Positive Matrix Factorization (PMF) model (version 5.0) to identify the sources of particle number concentrations in central Los Angeles. The contribution of each source was quantified and the temporal (seasonal and diurnal) variations for each sources were compared and discussed. In the next work, we present the chemical composition and toxicity of ambient PM and fog water in a rural location in the Po Valley, Italy, an area that experiences severe fog episodes during the winter time, explore the sources that contribute to ambient PM, fog water, and their associated toxicity, and determine how fog partitioning impacts the chemical composition and toxicity of ambient PM. For this purpose, we collected fog water samples using a fog water collector, while daytime and nighttime particles were also collected using a high-volume sampler. Samples were analyzed for their carbonaceous, ionic, elemental and metallic components. We also analyzed the samples for oxidative potential, using the alveolar macrophage (AM) assay. We then explored how fog formation might impact the chemical composition and, in turn, toxicity of ambient PM. The results from study have important implications for other areas, such as Los Angeles, that experience fog episodes, because, as described in detail in the paper, this phenomenon can critically increase the toxicity of ambient PM through aqueous-phase chemistry. And, finally, in the third and last study, we developed an online metal monitor for time-resolved measurement of (a time resolution of 2-hr) three important redox-active metals, namely Iron (Fe), Manganese (Mn), and Chromium (Cr), that are bound to ambient coarse particulate matter (PM) (i.e., 10₁₀₋₂.₅). To this end, we used virtual impactors (VIs) to first enrich coarse PM concentration into the target flow and then capture them in water slurry using a Biosampler. We then added specific chemical reagents (pertinent to each metal) to the slurry samples, which lead to the formation of colored complexes, the intensity of which (which is proportional to metal concentrations) was then measured using a Micro Volume Flow Cell (MVFC) coupled with UV/VIS spectrophotometry. The monitor was then deployed in the field, and its performance was measured and compared with that of a standard method (i.e., inductively coupled plasma-mass spectrometry (ICP-MS)) over a four-month period from January through April 2016. The data collected by the monitor were then used to obtain time-resolved time-series and diurnal variations plots of metal concentrations, which were used in combination of wind rose plots in infer information about the sources.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

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

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

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

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

PDF

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

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

PDF

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

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

PDF

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

PDF

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

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

PDF

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

PDF

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

PDF

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

PDF

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

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

PDF

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

PDF

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

PDF

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

PDF

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

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

Fabrication, deposition, and characterization of size-selected metal nanoclusters with a magnetron sputtering gas aggregation source

PDF

Personal exposure to particulate matter PM2.5 sources during pregnancy and birthweight

Asset Metadata

Creator Sowlat, Mohammad Hossein (author)

Core Title 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...

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Engineering (Environmental Engineering)

Publication Date 03/22/2019

Defense Date 02/15/2019

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

Tag aerosol,OAI-PMH Harvest,particulate matter,physico-chemical characterization,source apportionment,toxicity

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Sioutas, Constantinos (committee chair), Ban-Weiss, George (committee member), McConnell, Rob (committee member)

Creator Email hsowlat@gmail.com,sowlat@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-135491

Unique identifier UC11675682

Identifier etd-SowlatMoha-7174.pdf (filename),usctheses-c89-135491 (legacy record id)

Legacy Identifier etd-SowlatMoha-7174.pdf

Dmrecord 135491

Document Type Dissertation

Format application/pdf (imt)

Rights Sowlat, Mohammad Hossein

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