Personal exposure to particulate matter with aerodynamic diameter <2.5 µm (PM2.5) is impacted by indoor, outdoor, and personal activity-related (i.e., behaviors) sources in various microenvironments, spaces, and neighborhoods that individuals typically encounter [1, 2]. Prenatal exposure to PM2.5 specifically is associated with adverse maternal and fetal health outcomes [3,4,5]. Exposure in the 3rd trimester of pregnancy, when most fetal weight gain occurs is thought to be associated with low birth weight and impaired growth [6,7,8]. The toxicity of PM2.5 and its subsequent impact on health is driven by its chemical composition and main sources contributing to it [8,9,10]. Identifying and quantifying the main sources of personal PM2.5 exposure can shed light on particular mixtures that might pose a greater health risk and might otherwise be missed by solely investigating total PM2.5 mass concentration as a whole. This is particularly important in environmental health disparities contexts and for specific vulnerable populations such as pregnant women for whom meaningful recommendations to reduce exposures and health risks are needed [11, 12].

Main sources of PM exposure are generally resolved using source- and receptor-oriented modeling approaches [13, 14]. Based on the mass balance principle [15], receptor-oriented approaches utilize speciated measurements at receptors (i.e., receiving) to identify and quantify major sources (or source groups) impacting that location or individual [16, 17]. One of the most commonly used receptor models is the Positive Matrix Factorization Model (PMF) which solves for sources and their chemical profiles without pre-assuming what they are [18,19,20].

Several studies have conducted source apportionment in outdoor and indoor environments and investigated their health impacts [3, 21, 22]. For example, vehicular emissions, wood smoke, natural gas combustion, ship emissions, secondary aerosols, fresh and aged sea salt, and soil/road dust sources were resolved in outdoor PM (various size ranges) in Los Angeles, CA communities [23, 24]. Hasheminassab et al. [25] found mobile sources were the major contributor to both indoor (39 ± 21%) and outdoor (46 ± 17%) PM2.5 mass in three retirement homes in Los Angeles, Habre et al. [26] found cooking, cleaning, candle/incense burning, and smoking contributed significantly to indoor PM2.5 concentrations in New York City residences.

However, personal exposure to PM2.5 and its sources can be more complex to discern for several reasons. First, while personal monitoring (sampling air in the personal breathing zone) is considered the gold standard approach in external individual-level exposure assessment [27, 28], the cost and burden of conducting these studies is still high especially during pregnancy [29, 30]. Second, individuals get exposed to PM2.5 in multiple microenvironments and locations, often in close proximity to indoor sources or while mobile, sometimes generating PM2.5 from their activities (e.g., burning candles), and are always impacted by outdoor air pollution that infiltrates indoors or into the personal breathing zone [27, 28, 31, 32]. As such, disentangling sources that contribute to personal exposure can be more challenging compared to outdoor or indoor studies. This challenge is reflected by the relatively small number of studies conducting source apportionment in personal samples, most of which are collected 12 to 48 hour integrated PM samples [2, 33,34,35].

Of these, Özkaynak et al. [35] found that personal PM10 exposure was much higher than outdoor and indoor PM10 concentrations in 178 nonsmoking residents in Riverside, CA, and that these explained 16% and 50% of the variation in personal exposures, respectively. Cooking and smoking were important sources of personal exposure. Minguillón et al. [34] found wide variation in personal PM2.5 exposures of 54 pregnant women and reported on limitations of questionnaire data (e.g., recall error, accuracy of time and location of travel and activities) in helping to resolve sources.

To the best of our knowledge, no studies to date have conducted source apportionment on personal PM2.5 exposure samples in the 3rd trimester of pregnancy. We aimed to chemically characterize the composition of personal PM2.5 and resolve its main sources in the MADRES (Maternal And Developmental Risks from Environmental and Social Stressors) cohort in Los Angeles, CA. To accomplish this goal, we analyzed filter-based data from a personal monitoring sub-study in MADRES using the United States Environmental Protection Agency (EPA) PMF model [19]. We leveraged questionnaire data, geolocation monitoring (GPS), and geospatial modeling of environmental exposures (in residential neighborhoods and activity spaces) to confirm predicted source identities and understand how their mass contributions vary in relation to personal behaviors, indoor/outdoor sources, and time-activity patterns.


Study design

A total of 213 women in their 3rd trimester enrolled in the larger MADRES cohort were recruited into this personal monitoring sub-study between October 2016 and March 2020. MADRES is an ongoing prospective pregnancy cohort focused on predominantly low-income, Hispanic women and their babies residing in Los Angeles, CA. MADRES aims to address critical gaps in understanding environmental health disparities and the impacts of air pollution and social stressors on maternal and child health. The details of eligibility, enrollment, and follow-up of participants are described elsewhere [36]. Briefly, eligible participants for this sub-study were in the 3rd trimester at the time of recruitment, ≥18 years of age, and could speak either English or Spanish fluently. In the initial design, people living in a household with an active smoker were excluded to reduce the impact from smoking on personal PM2.5 exposures. However, in order to encourage all participants to contribute to sub studies, the non-smoking household criterion was not applied consistently throughout the study and was eliminated by the end of 2018. Informed consent was obtained for each participant. The University of Southern California’s Institutional Review Board (IRB) approved the study protocol.

Data collection

The 48-hr integrated personal PM2.5 measurements were collected to characterize the composition of personal PM2.5 and identify its main sources in the MADRES cohort through source apportionment analysis. Several other data sources were used from MADRES questionnaires and measurements and from external data sources in a second follow-on bivariate analysis to confirm source identities and understand the personal drivers that affect the mass contribution of each source. These included the following: questionnaires (collected at trimester 1, 2, and 3, respectively), GPS-derived time-activity patterns and environmental exposures within activity spaces from the personal monitoring study, modeled residential environmental exposures, and outdoor EPA PM2.5 chemical speciation data from a central site.

Personal PM2.5 measurements

The personal PM2.5 sampling design and protocol in the 3rd trimester is described in detail in O’Sharkey et al. [37] and Xu et al. [38]. Briefly, personal, 48-h integrated PM2.5 measurements were collected using a Gilian Plus Datalogging Pump (Sensidyne, Inc.) operating on a 50% cycle at 1.8 lpm flow rate with the sampling inlet located in the breathing zone. The pump is connected to a PM2.5 Harvard Personal Environmental Monitor (PEM) size-selective impactor with a 37 mm Teflon filter (2 µm pore size; Pall, Inc.). Participants were asked to wear the sampling device for the entire data collection period with a few exceptions. These included when it is unsafe to do so (e.g., driving), showering, or sleeping, in which case they were instructed to place the device near them in an unobstructed location.

Filters were analyzed gravimetrically to determine PM2.5 mass using an MT5 microbalance (Mettler Toledo, Columbus, OH, USA) in a dedicated humidity- and temperature-controlled chamber at the USC Exposure Analytics Laboratory. Filters were then sent to RTI International (Research Triangle Park, NC) to determine elemental composition of the following 33 species using Energy Dispersive X-Ray Fluorescence (EDXRF): barium (Ba), calcium (Ca), chlorine (Cl), copper (Cu), iron (Fe), potassium (K), magnesium (Mg), manganese (Mn), sodium (Na), nickel (Ni), sulfur (S), silicon (Si), titanium (Ti), zinc (Zn), aluminum (Al), bromine (Br), cobalt (Co), phosphorus (P), lead (Pb), selenium (Se), strontium (Sr), vanadium (V), cesium (Cs), zirconium (Zr), chromium (Cr), rubidium (Rb), arsenic (As), indium (In), silver (Ag), antimony (Sb), tin (Sn), cerium (Ce), and cadmium (Cd). Filters were also analyzed for concentrations of black carbon (BC), brown carbon (BrC), and environmental tobacco smoke (ETS) using a seven-wavelength optical transmittance integrating sphere method [39, 40].


Participants completed interviewer-administered questionnaires in trimester-specific visits as part of the larger MADRES cohort and an exit survey after completing the 48-hr monitoring period as part of the personal monitoring sub-study (Table S1 and S2). Data obtained from the MADRES questionnaires include the following: demographics (e.g., age, race, education, employment, income), housing characteristics (e.g., type of dwelling, building age). In addition, data on the following were available from the exit survey: time-activity patterns (e.g., time spent indoors and outdoors, commuting), home ventilation (e.g., window open, air conditioner use), current tobacco smoke exposure (primary and secondhand), and presence of any significant indoor sources of PM2.5 such as cooking or candle burning [36]. Participants’ home addresses at the 3rd trimester study timepoint were geocoded for residential exposure assessment.

Residential environmental exposure assessment

Daily ambient concentrations of PM2.5, PM10, nitrogen dioxide (NO2), and ozone (O3) were interpolated at the residence using inverse distance squared weighted interpolation from US EPA Air Quality System data [36]. Daily local traffic-related nitrogen oxides (NOx) concentrations at the residence were estimated using the CALINE4 line source dispersion model by roadway class [41]. Daily meteorology (temperature, precipitation, specific humidity, relative humidity, downward shortwave radiance, wind direction and wind speed) was assigned based on a 4 km × 4 km gridded model developed by Abatzoglou [42]. Forty-eight-hour integrated averages were calculated from daily measurements to correspond to the personal monitoring dates. For wind direction, four categories were created based on the 48-hr mean (arithmetic mean of two vector averages) as follows: 0–90° as wind blowing from northeast (NE), 91–180° as southeast (SE), 181–270° as southwest (SW), and 271–360° as northwest (NW), where a direction of 0° is due North on a compass.

GPS-derived time-activity patterns and environmental exposures within activity spaces

Smartphones with the study-developed madresGPS Android app pre-installed and programmed were used to log participants’ geolocation (GPS and metadata) and motion sensor data continuously at 10-s intervals for the 48-hr monitoring period. Data were then analyzed to derive time activity patterns as minutes spent staying at home or non-home locations (assumed to be indoors) and minutes spent on the road (or in transit, travel mode unknown) and then converted to percentages out of the 48-hr period for use in the analysis. The methods used to derive time-activity patterns were based on [43, 44] and described in more detail in Xu et al. [38].

GPS data was also used to construct 48-hr activity spaces and calculate environmental exposures encountered within them [38]. Briefly, activity spaces are defined as the local areas that individuals interact with when they move around during their daily activities [45, 46]. We constructed activity spaces using Kernel Density Estimation (KDE) method for each individual and then calculated the following environmental exposures within them: walkability index score, Normalized Difference Vegetation Index (NDVI, greenness), parks and open spaces, traffic volume on primary roads, and road lengths (for primary and secondary roads combined and for minor streets). Data sources for these measures are listed in Table S3. Briefly, KDE integrates time and space to account for durations of time spent at certain locations and incorporates a distance decay kernel function to assign higher weight to environmental features closer to the locations where participants spent the most time in (compared to locations they passed through) using pre-defined bin (e.g., 25 m) and neighborhood sizes (e.g., 250 m) [45, 47]. Activity space calculations were conducted in ArcGIS Pro 2.5 (Esri, Redlands, CA).

Outdoor PM2.5 chemical speciation in study region

PM2.5 metals and carbonaceous components concentrations were available every third day from the Chemical Speciation Network (CSN) [48] at the Downtown Los Angeles monitoring station—the most proximal and central site in the study area—and were downloaded from the EPA Air Quality System. Then, the data was linked to personal monitoring 48- hr time periods based on the overlapping dates.

Data analysis

Descriptive statistics were first conducted to plot the distributions of all the variables (e.g., individual/residential environment data) for the participants. Then, using USEPA PMF5.0 model, source apportionment analysis was performed on personal PM2.5 measurements to identify main sources, along with their contributions to PM2.5 mass. Once sources were resolved, bivariate analysis was conducted between each PM2.5 source and several variables describing personal behaviors, indoor/outdoor sources, and time-activity patterns to confirm source identities and understand the personal drivers that affect PM2.5 exposures.

Descriptive analysis

Descriptive statistics were calculated in SAS 9.4 (SAS Institute Inc) to describe the distribution of personal PM2.5 mass concentration, elemental components and carbon fractions (used in final PMF analysis), questionnaire variables (home ventilation, time-activity patterns, indoor sources, etc.), and environmental exposures (residential or GPS-derived).

Source apportionment analysis

The USEPA PMF 5.0 model was used to resolve and identify major sources of PM2.5 and quantify their mass contributions using measured concentrations and sample-specific uncertainties as inputs. Briefly, the PMF model uses factor analysis to identify source contributions and profiles for a given number of sources through solving the following equation: [19, 20, 49]

$${X}_{{ij}}=\mathop{\sum }\limits_{k=1}^{n}{g}_{{ik}}{f}_{{kj}}+{e}_{{ij}}$$

where Xij represents the concentration of chemical component j in sample i, gik represents the mass contribution of factor k in sample i, fkj represents the loading of chemical component j on factor k, and eij is the residual error for sample i and species j.

The PMF model solves Eq. (1) by minimizing the sum of squares object function Q for a given number of factors k: [20, 49]

$$Q=\mathop{\sum }\limits_{i = 1}^{n}\mathop{\sum }\limits_{j =1 }^{m}\left[{\frac{{e}_{{ij}}}{{u}_{{ij}}}}\right]^{2}$$

where uij is the uncertainty of species j in sample i. The model decomposes the concentrations matrix into a contributions g matrix and profiles f matrix and constrains results to be positive (or not significantly negative) [20, 50]. Each observation is individually weighted by its uncertainty in Eq. (2); therefore, samples with higher analytical uncertainties will have less influence on the solution.

Based on the PMF-calculated signal-to-noise ratio (S/N) indicating the degree of analytical noise relative to the concentration of each species [19], we categorized species as “Bad” (S/N ≤ 0.2, excluded from analysis), “Weak” (0.2 < S/N < 1, downweighted in the analysis), and “Strong” (S/N ≥ 1, retained). Although Pb and V had S/N < 0.2, they were included in the analysis as potentially important tracers of traffic and fuel oil, respectively, and set to “Weak.”

Of the 36 elemental and carbon species, the following 16 were finally included as “Strong” in the PMF analysis: BC, BrC, Ba, Ca, Cl, Cu, Fe, K, Mg, Mn, Na, Ni, S, Si, Ti, and Zn. We also included 9 “Weak” species as follows: Al, Br, Co, ETS, P, Pb, Se, Sr, and V. PM2.5 mass was designated as the total variable. An extra 10% modeling uncertainty was added to account for sampling or modeling errors not captured in the sample-specific analytical uncertainties [19]. Missing concentrations were replaced by species’ medians. One out of 213 (0.5%) samples were excluded as outliers based on multiple species’ concentrations.

Solutions with five to seven factors were scanned first to decide upon a reasonable factor number in a final model with 100 runs. Q values were checked for no undue influence from outliers and no local minimum solutions. Based on chemical loading profiles and prior knowledge, the optimal number of sources was selected which provided the most physically interpretable solution [50]. The convergent solution with the lowest Qrobust value (goodness-of-fit parameter excluding points with uncertainty-scaled residuals greater than 4) was selected [19]. Fpeak rotations, where positive F peak values sharpen the F matrix and negative values sharpen the G matrix were performed next to refine the solution. The optimal Fpeak value for solution rotation was chosen based on the smallest change in Q [19]. Residuals and R2’s for each species were checked for normality and model fit, respectively. Finally, diagnostics analysis of Displacement (DISP), Bootstrap (BS) (100 bootstraps, 0.6 minimum correlation), and Bootstrap-Displacement (BS-DISP) were performed to estimate the variability in the PMF solution under different scenarios. DISP focuses on effects of rotational ambiguity in the profiles or loadings; BS identifies whether there are a small set of observations that can disproportionately influence the solution; and BS-DISP includes effects of random errors and rotational ambiguity [19].

Bivariate confirmatory analysis of source identities and drivers

Bivariate analyses were conducted to further confirm source identities and examine trends in their mass contributions. Scatterplots, boxplots, and nonparametric statistics (Spearman correlations for continuous variables and Kruskal-Wallis test for categorical ones) were used to describe and test the relationships of select factors (selected from the literature and prior knowledge) with each source’s predicted mass contributions. These included demographics, time-activity patterns, home characteristics, indoor air pollution sources, and outdoor residential and activity space based environmental exposures as described earlier. Categorical variables with unbalanced values (≥85% of the records have one value) or with high missingness (≥80%) were dropped from further analysis.


Descriptive statistics

Most of the participants (>98%) resided in central and east Los Angeles, CA. The majority were Hispanic (78%), working (48%) during the 3rd trimester, and completed grade 12 or less (55%) (Table S1). Mean age was 28 years at consent (range 18–45 years) and mean parity was 2 (range 1–6). Among participants who reported annual household income, most of them had annual household income <$30,000 (68%).

Personal PM2.5 mass and component concentrations are provided in Table 1. Mean (SD) PM2.5 concentration was 22.3 (16.6) μg/m3. The optical carbon fractions BC, BrC, and ETS combined constituted on average 17% (3.8) μg/m3 of the total PM2.5 mass.

Table 1 PM2.5 mass and chemical component concentrations (in units of ng/m3 unless otherwise noted).

The distributions of home characteristics, indoor PM sources, and selected time activities as reported in questionnaires or derived from GPS are presented in Table S2. Based on the exit survey, 60% of participants spent some time near traffic when outdoors, and 34% spent more than 2 hrs per day commuting during the 48-hr monitoring period. During sampling, 60% of participants opened windows more than half of the time, 26% used air conditioning and 37% used fans at home. In terms of indoor PM sources, 38% were close to cooking smoke, 24% were close to burning candles or incense, and 39% spent time close to people smoking (cigarette, cigar, hookah, or pipe smoke) nearby. In addition, 56% of participants lived in an apartment, 44% were part of a household with >3 persons, and 43% lived in a home built after the 1980s. Participants also spent an estimated mean (SD) of their time at home 78.5% (19.6%) and at non-home locations 15.2% (15.7%).

Source apportionment results

Fourteen observations had missing values and were replaced by species median, including PM mass (3 obs), BC (3), BrC (6), and ETS (2). A five-factor solution combined fuel oil and secondhand smoking, while seven factors resulted in a non-interpretable factor with single high loading of Zn. A six-factor solution was chosen as the optimal, physically interpretable one (Qrobust = 5845.4 and Qtrue = 6143.2). An Fpeak rotation of 0.1 was then applied with 100 bootstraps, which resulted in no unmapped factors (Table S4). Some species with higher S/N ratio, e.g., BC, Cl, K, S, Ca, and Zn, had non-normal residuals (Table S5).

The six sources together explained 48% of the variability in PM2.5 mass concentration and included the following: Traffic, Secondhand Smoking, Aged Sea Salt, Fresh Sea Salt, Fuel Oil, Crustal (Fig. 1). Each of these is presented in more detail below along with supporting relationships from confirmatory bivariate analyses.

Fig. 1: PMF-predicted source loading profiles (in % of species).
figure 1

Sources are color coded as shown: .


The first source identified was traffic with high loadings of BC, Zn, and Ba (Fig. 1). It contributed on average 2.6% of personal PM2.5 mass (Table 2). Traffic was moderately positively correlated with crustal and inversely correlated with fresh sea salt and fuel oil sources (Table 3). It was strongly correlated with outdoor residential ambient (NO2 and PM2.5) and traffic-related (total CALINE4 NOx) air pollutants. It was also negatively correlated with outdoor O3. In addition, Traffic source contributions increased with greater length of primary roads in the KDE activity space (Table 4).

Table 2 Predicted source contributions to personal PM2.5 mass concentrations.
Table 3 Spearman correlation matrix of PMF-predicted PM2.5 source contributions, colored from low (blue) to high (red).
Table 4 Spearman correlations between PMF-predicted PM2.5 source contributions and select continuous variables.

Secondhand Smoking

This second source had a high loading of BrC and ETS (Fig. 1). With an average mass contribution of 11.9 μg/m3, it contributed the majority of personal PM2.5 mass (65.3% on average) (Table 2). Participants living in apartments seemed to have slightly higher exposure to this source compared to those living in houses (13.0 vs. 10.5 μg/m3, respectively, not significant, Fig. 2a). This source was also slightly negatively correlated with greater window opening time. Secondhand smoking (SHS) was also slightly higher when participants were close to 2+ persons smoking nearby as compared to one (among ~33% of participants reporting being close to someone smoking in the 48-hr monitoring period) (13.2 vs. 11.0 μg/m3 respectively, not significant). To eliminate the possibility that this source is capturing an outdoor biomass burning signal, we checked its correlation with outdoor potassium (elemental, K), potassium ion (K+), and elemental and organic carbon concentrations at the Downtown CSN site, all of which showed insignificant weak correlations (Table 4). SHS was also not correlated with any questionnaire variables on indoor cooking (frequency, use, etc.).

Fig. 2: Relationship between source mass contributions (y-axis) and environmental or home characteristics.
figure 2

a Secondhand smoking and home type, b aged sea salt and window opening time in the 48-h monitoring period, c fresh sea salt and average wind direction in the 48-h monitoring period, and d crustal and number of household occupants (*The relationships in (b d) were significant with Kruskal-Wallis test p value  < 0.05) .

Aged sea salt

This third source had high loadings of Na, Mg, and S (Fig. 1). It contributed on average 4.3% of the personal PM2.5 mass (Table 2). Aged sea salt was negatively correlated with the SHS source (Table 3). It was strongly positively correlated with outdoor O3 concentration and temperature and negatively correlated with wind speed (Table 4). Aged sea salt was also significantly positively correlated with window opening time (Fig. 2b).

Fresh sea salt

This fourth source had high loadings of Cl, Na, and Mg (Fig. 1). It contributed on average 4.7% of the personal PM2.5 mass (Table 2). Fresh sea salt was negatively correlated with traffic and SHS sources (Table 3). The mass contribution of this source were significantly different across different wind directions, with the highest contribution on days when average wind direction originated from the west (NW followed by SW, Fig. 2c). It was also positively correlated with outdoor residential wind speed and relative humidity. To eliminate the possibility of this being an indoor source related to aerosolized minerals from domestic water use or salt used in cooking [32, 35, 51], we checked its relationships with humidifier usage and time close to smoke from cooking, respectively. Even though the sample size was unbalanced (30 out of 212 reported using a humidifier), average mass contributions were lower (not significant) when people used a humidifier compared to not (0.5 vs. 0.9 μg/m3, respectively). Similarly, mass contributions were lower when participants reported spending more time close to smoke from cooking. In addition, fresh sea salt was moderately positively correlated with ambient Cl and Cl as measured at the Downtown Los Angeles site (Table 4).

Fuel oil

This source had high loadings of Cu, Ni, and V (Fig. 1). It contributed the second largest share of personal PM2.5 mass (11.7%) with an average mass contribution of 2.1 μg/m3 (Table 2). Fuel oil was positively correlated with crustal and negatively correlated with traffic sources (Table 3). This source was slightly higher (non-significant) in participants living in homes built before 1980 compared to after (2.4 vs. 1.9 μg/m3). Fuel oil was also positively correlated with outdoor NO2 and negatively correlated with O3 (Table 4).


The sixth and final source had high loadings of crustal elements Ca, Si, Ti, and Al (Fig. 1). It contributed 11.5% (2.1 μg/m3) of personal PM2.5 mass on average (Table 2). Crustal was moderately positively correlated with traffic and fuel oil sources (Table 3). Living in a household with 4+ occupants was associated with greater contributions of this source compared to fewer occupants (2.5 vs. 1.2 μg/m3, p value = 0.04, Fig. 2d). It was also positively correlated with outdoor NO2 and PM10 concentrations and negatively correlated with relative humidity and precipitation (Table 4).


We identified six sources of personal PM2.5 exposure along with their mass contributions in the 3rd trimester of pregnancy, in a population of 212 low-income, predominantly Hispanic pregnant women living in Los Angeles, CA. Secondhand smoking (SHS) followed by fuel oil and crustal were the biggest contributors to personal PM2.5 exposures. Of the six sources, SHS and crustal appeared to be of indoor origin (or closely related to indoor activities), while traffic, aged and fresh sea salt, and fuel oil were of outdoor origin. Secondhand smoke was the single largest contributor to total personal PM2.5 mass concentrations. The combined indoor source contributions (14 ± 10 μg/m3) were more than triple those of outdoor sources (4.2 ± 2.9 μg/m3), highlighting the importance of the indoor environment in contributing to personal exposures during this critical window of time.

SHS is also a well-known indoor air contaminant [52, 53] associated with a suite of adverse health effects [54,55,56]. This source had high loadings of BrC and ETS (the lab-reported marker of SHS), and some loadings of Br and K which were related to tobacco smoke in previous studies [39, 57, 58]. In order to avoid overloading the samplers with particles from primary tobacco smoke which would also overshadow chemical fingerprints from other sources if present, by design, we excluded participants who reported smoking themselves (this did not occur in this population) or those with an active smoker permanently residing in their household (despite this latter criterion not being consistently applied throughout the study). Despite the low smoking prevalence and these design restrictions, secondhand smoking was still identified as the source with the largest contribution to personal PM2.5 exposures in our study. The mass contributions of this source (and measured ETS concentrations) did not show any clear changes over time as the study progressed, suggesting that recruitment decisions did not significantly influence our findings. This source has been found in other personal PM2.5 monitoring studies [34, 35, 59]. Minguillón et al. [34] identified cigarette sources in the personal, indoor and outdoor environments for the pregnant women living in Barcelona, Spain. Özkaynak et al. [35] found nicotine mainly in the personal sample and indoor environment for the nonsmoking residents in Riverside, California. Using a real-time PM2.5 monitor in personal and residential environment, Zhang et al. [59] found SHS distribution in children with and without self-reported SHS exposure in New York City, where children from smoking families had six times that of SHS exposure for children from non-smoking families. All these studies, including the current one, demonstrate that SHS source is a common finding in the personal monitoring studies. This suggests that secondhand (and possibly thirdhand) smoke exposure remains an issue of public health concern especially during pregnancy and in the context of our population, despite the success of large-scale bans on public smoking. We found that participants living in apartments tended to have marginally higher exposure to secondhand smoking than those living in detached houses. This could suggest greater potential of secondhand smoke infiltration from adjacent units in an apartment building or from visitors smoking [60,61,62]. Anecdotal evidence from the study (data not shown) also suggests secondhand smoke exposure could be occurring outside of the residence, during public transit or while commuting, and warrants further investigation.

We also found both fresh and aged sea salt—sources of outdoor origin—contributed to personal PM2.5 exposures in our study with high loading of Na, Mg, Cl (fresh only), and S (aged only). Previous work identified sea salt or marine aerosol sources with similar loading profiles [16, 23, 63]. Despite only having access to average wind direction data over the 48-hr monitoring period (not an ideal wind measure compared to most frequent wind direction for example), fresh sea salt mass contributions were higher with westerly winds and higher wind speeds, which provides greater potential for aerosolization and airborne transport of sea salt particles from the Pacific Ocean. Habre et al. [23] found sea salt mass contributions to PM2.5 mass in southern California to be highest in coastal communities. As fresh sea salt ages and undergoes photochemical reactions that also lead to secondary O3 formation, chlorine is lost and sulfates are formed [23, 64]. Thus, aged sea salt resembles fresh sea salt in its loading profiles, except with S substituting Cl. Lower wind speed, stagnant conditions, and warmer temperatures provide more chemical reaction time and contribute to the loss of Cl and formation of O3 as fresh sea salt ages [65, 66]. Given the mild, coastal southern California climate year-round, it is not surprising to see both outdoor sea salt sources contributing to personal PM2.5 exposures of pregnant women in our study.

As for the crustal source, crustal elements such as Al, Ca, and Ti are naturally present in the earth’s crust. They originate outdoors and can enter the indoor environment as windblown dust or as dust tracked indoors on residents’ shoes. Once indoors, crustal materials will typically settle and get resuspended as indoor sources (or emissions of indoor origin) when disturbed by human movement or other activities (i.e., vacuuming). Therefore, the presence of more occupants in a household provides greater opportunities for re-suspension of crustal dust, similar to earlier studies [26]. Despite crustal being positively correlated with the traffic source, it did not have any loadings of known traffic-related markers (BC, Ba, Zn, Cu) suggesting it was not reflecting a road dust signal [67]. As such, crustal can be thought of as a predominantly indoor origin source despite the possibility of our participants getting exposed to crustal dust outside of their homes, in daily commutes, and outdoor activities.

Finally, we found that fuel oil and traffic sources contributed to personal PM2.5 exposures as well. Similar to previous studies, the fuel oil source had high loadings of Ni and V which are known tracers of heavy residual fuel oil combustion in industrial applications, certain heavy-duty machinery, and in marine engine emissions [68,69,70,71,72]. BC serves as a traffic marker especially for diesel engines [16, 26], while Zn, Ba, and Fe are related to motor vehicle exhaust emission, brake and tire wear, and diesel additives [73, 74]. Traffic is an important source of air pollution in Los Angeles, CA [23, 25]. Our study shows that greater concentrations of traffic-related pollutants outdoors (CALINE NOx and ambient NO2) and spending more time near primary roads (in actual activity spaces) correlate with greater personal exposure to the PM2.5 traffic source. It is also important to note that the traffic source had high loadings of tailpipe combustion markers (BC) and non-exhaust brake and tire wear markers (Ba and Zn, respectively) potentially capturing both tailpipe and non-tailpipe traffic exposures.

The strengths of this study include the personal PM2.5 measurements and detailed chemical composition analysis that allowed us to apportion the major sources of personal PM2.5 exposure during a critical time window of pregnancy. By integrating concurrently collected questionnaire data and geospatially modeled environmental exposures in activity spaces (from GPS) and in the residential neighborhood, the results further corroborate these sources, their determinants, and origin (primarily indoor vs outdoor). However, the PMF model only explained a portion of the variability in personal PM2.5 mass concentrations (R2 0.48). One possible reason could be that we did not measure organic carbon (OC) species on the Teflon filters in this study which are known to contribute a large fraction of indoor PM2.5 mass concentrations [26, 75]. It was also possible that some of most volatile OC species and ammonium nitrate among others volatilized off the filters; we found higher temperature was associated with lower personal PM2.5 mass in bivariate analysis [38]. We also did not measure volatile organic compounds or other species that could potentially contribute significantly to personal air pollution exposures (in different phases, and not just in PM2.5). The sample size of the study, while considered large in personal monitoring settings, and the short monitoring period may limit the generalizability and representativeness of personal PM2.5 exposures beyond this study area and across the full pregnancy and postpartum periods.


This is one of the few studies to conduct a thorough characterization of sources impacting personal PM2.5 exposures of predominantly Hispanic and low-income women during pregnancy in an environmental health disparities context. Our findings show the diversity of sources that impact personal PM2.5 exposures in pregnancy in Los Angeles, CA. With ~ 77% of personal exposures contributed by indoor sources, the findings highlight the importance of the indoor environment contributions to total PM2.5 exposures during pregnancy and the potentially incomplete understanding of this population’s exposures by solely relying on outdoor air pollution measures. These results may inform source-specific health investigations and subsequent interventions to reduce exposure to potentially more harmful components of PM2.5.