Subscriber access provided by University of Winnipeg Library
Characterization of Natural and Affected Environments
Patterns of personal exposure to urban pollutants using personal passive samplers and GC×GC/ToF-MS Carlos A Manzano, Nathan Gray Dodder, Eunha Hoh, and Raul G. E. Morales Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.8b06220 • Publication Date (Web): 21 Dec 2018 Downloaded from http://pubs.acs.org on December 27, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 30
Environmental Science & Technology
1
Patterns of personal exposure to urban pollutants using personal passive samplers and
2
GC×GC/ToF-MS
3 4
Carlos A. Manzano,1,2, * Nathan G. Dodder,2,3 Eunha Hoh,2 Raul Morales1
5
1Center
for Environmental Science, Faculty of Science, Universidad de Chile, Santiago, Chile
6
2School
of Public Health, San Diego State University, San Diego, CA, USA
7
3San
Diego State University Research Foundation, San Diego, CA, USA
8 9
Abstract
10
The performance of silicon wristband passive samplers (WB), combined with comprehensive two-
11
dimensional gas-chromatography/time-of-flight mass-spectrometry (GC×GC/ToF-MS), for the analysis
12
of urban derived pollutants in the personal environment was evaluated. Cumulative 5-day exposure
13
samples from 27 individuals in areas with different geographical/socioeconomic characteristics within the
14
Santiago Metropolitan Region (Chile) were collected during winter and summer (2016-2017). Samples
15
were extracted without cleanup/fractionation and analyzed using targeted and nontargeted methods. The
16
quantified semi-volatile organic compounds (SVOCs, n=33) (targeted analysis), and tentatively identified
17
features (n=595-1,011) (nontargeted analysis) were classified according to their use/source. Seasonal
18
differences were observed in the targeted analysis, while seasonal and spatial differences were observed
19
in the nontargeted analysis. Higher concentrations of combustion products were observed in winter, while
20
higher concentrations of consumer products were found in summer. Spatial differences were observed in
21
hierarchical clustering analysis of the nontargeted data, with distinct clusters corresponding to specific
22
sub-regions of the urban area. Results from this study provide spatial and seasonal distributions of urban
23
pollutants within an urban area and establish the utility of linking WB with nontargeted analysis as a tool
24
to identify and prioritize new exposures to urban contaminants at the local/community level.
25
Keywords. Wristbands, nontargeted, urban pollutants, fingerprint, GC×GC/ToF-MS
ACS Paragon Plus Environment
Environmental Science & Technology
26
1. Introduction
27
Urban air pollution, connected to rapidly rising urban populations, is a top environmental concern and
28
has been recognized as a global challenge in the 21st century.1-3 It is well-known that residents of urban
29
settlements are exposed to a complex mixture of pollutants of natural and synthetic origin.4-8 Pollutants
30
originating from different sources, such as dietary or smoking habits, consumer products, and
31
heating/cooking fuels and their combustion products may be part of this complex mixture. Although some
32
urban derived pollutants have been associated with adverse health effects,9-11 there is little or no
33
information about a potentially vast number of pollutants to which urban populations are exposed.
34
Therefore, there is a need to derive a comprehensive description of urban pollutants at the local level (i.e.,
35
within cities and communities).
36
Most environmental research and urban monitoring efforts have historically used stationary active
37
sampling methods to measure a limited number of pollutants believed to pose potential risk.12 Some
38
advantages of using active sampling methods include the ability to provide reliable quantitative
39
concentration data, identify high concentration peak episodes, and provide high temporal resolution.13-15
40
Government-operated air quality monitoring networks have been built worldwide based on this
41
approach.16-19 Active air sampling has been used in our study area (Santiago, Chile) since the 1990s.19
42
However, estimating personal exposure using stationary active samplers may have limitations due to the
43
proximity effect, in which contaminant sources near the subject could cause higher exposures that might
44
not be captured by monitoring stations.20 Additionally, active monitoring networks represent mostly
45
outdoor environment, thus indoor sources of pollutants (i.e., where people spend ~80-90% of their day21,
46
22)
47
tested to minimize these uncertainties and improve on exposure studies,23 but often they can be a burden
48
for study participants because of pump noise, weight and equipment malfunction, in addition to their
49
relatively high cost.24
50 51
tend to be underrepresented. Personal active samplers (e.g., air monitoring backpacks) have been
Passive sampling technology represents an attractive alternative to active sampling, does not require external energy, and is generally low-cost.25 However, some disadvantages of using passive sampling
ACS Paragon Plus Environment
Page 2 of 30
Page 3 of 30
Environmental Science & Technology
52
techniques include a low temporal resolution (due of their lower sampling rate), unsuitability for short-
53
term monitoring and difficulties with automation.13, 25 In passive sampling, chemicals are sequestered
54
from the environment through passive diffusion into a receiving phase until equilibrium is reached.
55
Passive samplers act over long periods of time to provide time-weighted average concentrations, they are
56
resistant to accidental or extreme variations of concentrations and are suitable for long-term monitoring of
57
complex mixtures.25-29 Recent studies have demonstrated the effectiveness of silicon wristbands (WB) as
58
personal passive sampler devices to capture atmospheric semi-volatile organic compounds (SVOCs) with
59
log octanol-air partition coefficient ranging from 2.1-13.7.30 Additionally, WB may capture pollutants
60
through contact with skin, which may include pollutants from air, water, and personal care products, and
61
thus could be more representative of the totality of the personal exposure to urban pollutants.31, 32 This is
62
relevant as it has been estimated that SVOC uptake by skin absorption can even exceed uptake by
63
inhalation.33 WB provide the potential for a comprehensive assessment of personal exposure by capturing
64
multiple exposure pathways over a time period, which include workplace, indoor and outdoor exposures
65
through inhalation and dermal absorption.34 WB have been used in studies ranging from 48 hour to one
66
month exposure periods to measure pesticides,35-38 chemicals used in personal care products,36, 39 flame
67
retardants,31, 40 and other organic pollutants in urban, rural, outdoor and indoor settings.41-44 Although
68
personal exposures are likely composed of multiple classes of contaminants, with a large variability
69
observed among subjects,45 most of these studies have focused on target analytes, which often includes
70
extensive sample cleanup and fractionation.
71
Comprehensive two-dimensional gas chromatography (GC×GC/ToF-MS) has gained wide
72
implementation in the analysis of complex environmental samples since pioneered nearly 26 years ago.46
73
The theoretical peak capacity of GC×GC has been determined as being an order of magnitude higher than
74
conventional GC methods,47 and has been effectively used for the simultaneous analysis of multiple
75
classes of chemicals in complex environmental matrices: snow samples, soil, sediments and biota.48-57
76
In this study, the performance of low-cost WB samplers combined with GC×GC/ToF-MS with targeted
77
and nontargeted approaches was evaluated for the measurement of urban derived pollutants in the
ACS Paragon Plus Environment
Environmental Science & Technology
78
personal environment during 5-days exposures. We tested two hypotheses: 1) WB passive samplers
79
combined with multiple chemical analysis based on GC×GC/ToF-MS can be used as a tool to identify
80
spatial and seasonal differences in exposure to urban contaminants in various sub-regions of the same
81
metropolitan area, 2) That nontargeted analysis can detect exposure patterns unobserved by a 47-
82
compound targeted-analysis alone. This work is a methodological advancement in the use of WB passive
83
samplers to assess urban contaminant exposure. The results from this study provided preliminary spatial
84
and seasonal distributions of urban contaminants in the Santiago Metropolitan Region and can be used for
85
future prioritization efforts to characterize urban contaminants at the local level.
86 87 88
2. Materials and Methods
89
Study design. The protocols for recruitment of volunteers, collection, processing and analysis of samples
90
were pre-approved by the Ethical Committee of the Faculty of Sciences of the University of Chile
91
(approved and signed on 12/21/2015). Volunteers were recruited from three high school institutions of
92
different residential areas of the Santiago Metropolitan Region located at >12 km from each other: Las
93
Condes (sampling area ‘LC’: 33°25’28” S; 70°33’16” W), Maipu (sampling area ‘M’: 33°29’57” S;
94
70°45’32” W) and Puente Alto (sampling area ‘PA’: 33°36’41” S; 70°34’14” W) (Figure 1). According to
95
the latest Chilean National Population Census, sampling area LC is mainly inhabited by upper-middle to
96
high-income families (projected 2018 population of 288,228 people), sampling area M is mainly
97
associated with middle to upper-middle income families (projected 2018 population of 563,449), whereas
98
sampling area PA is mostly inhabited by low-middle to middle income families (projected 2018
99
population of 633,021).58 The selected residential areas have been reported to be affected differently by
100
urban expansion over the last 30 years. Land use studies have demonstrated that agricultural areas have
101
experienced high losses mainly due to new urban areas (i.e., housing and roads). However, the spread of
102
these new urban areas was not evenly distributed, as it was limited by the Andes mountain range to the
103
east. Therefore, the southern areas of the city experienced generalized urban expansion (sampling area M
ACS Paragon Plus Environment
Page 4 of 30
Page 5 of 30
Environmental Science & Technology
104
and PA), compared to the northern areas (sampling area LC).59 These new developments have produced
105
segregated communities with poor connectivity and detrimental environmental and socio-economic
106
impacts.60
107
At least five volunteers were recruited from each sampling area at the start of the study. The criteria for
108
participation were that the volunteer must travel 100 based on peak abundance (Table S4). The results were further processed using the statistical
184
compare add-on within ChromaTOF. Statistical compare is a post-data processing software package used
185
for the analysis of complex GC×GC chromatograms. It facilitates data mining and performs analyte
186
alignment based on retention time and mass spectral matches to make comparisons between groups of
187
samples.61 The WB extracts were divided into 7 groups according to their sampling location (i.e.,
188
sampling areas M, PA and LC during winter and summer, and blanks), and statistical compare parameters
189
were set conservatively to search for features with S/N>100 and present in 100% of WB in each group
190
(i.e., sampling area/season), therefore reducing personal variability (Table S4). The peak areas of the
191
resulting features were normalized to that of an internal standard (d12-chrysene) spiked before extraction,
192
and those showing ratios below 50:1 with respect to the average normalized peak area of the blanks were
193
manually excluded. Tentative peak identifications were assigned based on mass spectral matching to the
194
NIST 2011 MS library v2.0, with a minimum required similarity match of 500 (50%). Features below
195
50% similarity were reported as ‘unknown’. Note that the main objective was to screen for features
196
unique to each group of samples using tentative identifications with ≥ 50% similarity scores; and not to
197
confirm chemical identities, in which case a higher similarity score threshold would have been
198
appropriate.
199
Hierarchical clustering analysis on the log-transformed full data set (i.e., log10([abundance]+1)) with
200
Ward variance minimization was used for data interpretation and to identify closely related samples
201
without forcing associations between samples or analytes. Heatmaps were generated using Python’s
202
seaborn package v0.9.0.62
203 204
3. Results and Discussion
ACS Paragon Plus Environment
Page 8 of 30
Page 9 of 30
Environmental Science & Technology
205
Thirty WBs (n=30, worn by 30 people) were initially deployed. However, not all participants completed
206
the exposure study. One WB was lost, and another participant withdrew from the study, both from
207
sampling area LC, leaving it with only 3 WBs in winter and summer. Conversely, one extra WB (6 total)
208
was deployed in sampling area PA during the summer at the request of the participants. The participants
209
were composed of 15 women (56%) and 12 men (44%), ranging from 15-17 years old. The participant’s
210
evaluation of the experiment (on a scale from 1 to 10, with 10 being “very comfortable”) was 8.5±1.5
211
(Figure S1), which shows the participant’s willingness to wear WB samplers. All participants in this study
212
were high school students, therefore reducing variation due to occupational settings, and spent ~33% of
213
their time in school (given that the average number of hours in the school day in Chile is 8 hours). The
214
average commute reported was 3.6±2 min (
