Subscriber access provided by AUSTRALIAN NATIONAL UNIV
Article
Release and Gas-Particle Partitioning Behaviors of Short-Chain Chlorinated Paraffins (SCCPs) During the Thermal Treatment of Polyvinyl Chloride Flooring Faqiang Zhan, Haijun Zhang, Jing Wang, Jiazhi Xu, Heping Yuan, Yuan Gao, Fan Su, and Jiping Chen Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b01965 • Publication Date (Web): 11 Jul 2017 Downloaded from http://pubs.acs.org on July 11, 2017
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 free 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 accessible to all readers and 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.
Environmental Science & Technology 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 27
Environmental Science & Technology
1
Release and Gas-Particle Partitioning Behaviors of
2
Short-Chain Chlorinated Paraffins (SCCPs) During the
3
Thermal Treatment of Polyvinyl Chloride Flooring
4
Faqiang Zhan,†,‡ Haijun Zhang, *,† Jing Wang,§,║ Jiazhi Xu,†,‡ Heping Yuan,†
5
Yuan Gao,† Fan Su,† and Jiping Chen†
6
†
7
Chemical Physics, Chinese Academy of Sciences, Dalian, 116023, China
8
‡
9
§
10
║
11
CH-8600, Switzerland
CAS Key Laboratory of Separation Science for Analytical Chemistry, Dalian Institute of
University of Chinese Academy of Sciences, Beijing 100049, China Institute of Environmental Engineering, ETH Zurich, Zurich CH-8093, Switzerland Empa, Swiss Federal Laboratories for Materials Science and Technology, Dübendorf
12 13
TOC/Abstract Art
14 15 16 17 18 19 20
*Corresponding Authors
21
Phone: +86-411-8437-9972, fax: +86-411-8437-9562; e-mail: hjzhang@dicp.ac.cn.
ACS Paragon Plus Environment
Environmental Science & Technology
22
ABSTRACT
23
Chlorinated paraffin (CP) mixture is a common additive in polyvinyl chloride (PVC)
24
products as a plasticizer and flame retardant. During the PVC plastic life cycle, intentional
25
or incidental thermal processes inevitably cause an abrupt release of short-chain CPs
26
(SCCPs). In this study, the thermal processing of PVC plastics was simulated by heating
27
PVC flooring at 100–200 °C in a chamber. The 1-h thermal treatment caused the release of
28
1.9%–10.7% of the embedded SCCPs. A developed emission model indicated that SCCP
29
release was mainly controlled by material–gas partitioning at 100 °C. However, release
30
control tended to be subjected to material-phase diffusion above 150 °C, especially for
31
SCCP congeners with shorter carbon-chain lengths. A cascade impactor (NanoMoudi) was
32
used to collect particles of different sizes and gas-phase SCCPs. The elevated temperature
33
resulted in a higher partition of SCCPs from the gas-phase to particle-phase. SCCPs were
34
not strongly inclined to form aerosol particles by nucleation, and less present in the Aitken
35
mode particles. Junge-Pankow adsorption model well fitted the partitioning of SCCPs
36
between the gas-phase and accumulation mode particles. Inhalation exposure estimation
37
indicated that PVC processing and recycling workers could face a considerably high risk
38
for exposure to SCCPs.
39
INTRODUCTION
40
Polyvinyl chloride (PVC) is one of the highest-volume plastics produced, and it is
41
widely used in packaging, electronic devices, the automobile industry and building
42
materials.1 To enable PVC processing and improve its physical and mechanical properties,
43
PVC is always mixed with additives, such as stabilizers, plasticizers, lubricants, flame
44
retardants and other agents. Chlorinated paraffins (CPs) are common additives in PVC
45
plastics as both plasticizers and flame retardants.2 They are extremely complex mixtures
46
that consist of thousands of congeners and isomers, which can be subdivided into
47
short-chain (SCCPs, C10−13), medium-chain (MCCPs, C14−17) and long-chain CPs (LCCPs, ACS Paragon Plus Environment
Page 2 of 27
Page 3 of 27
Environmental Science & Technology
48
C>17) according to their carbon chain length. Among them, SCCPs has been proposed as a
49
group of new persistent organic pollutants (POPs) by the Stockholm Convention (UNEP,
50
2017),3 due to their environmental persistence, long-range transport potential, tendency to
51
bioaccumulate and high toxicity.4–9
52
SCCPs have been subjected to regulation in Europe, the U.S. and Canada due to their
53
physical-chemical properties and adverse effects.8 In the European Union, articles
54
containing SCCPs in concentrations equal to or greater than 0.15% by weight are
55
prohibited.10 However, the use and production of SCCPs in most developing countries is
56
not yet restricted. In China, a larger quantity of commercial CP mixtures were produced
57
without being classified by carbon chain length, and approximately 70% of produced CP
58
mixtures were consumed by the PVC processing industry.11 It was estimated that the
59
consumption of SCCPs in the Chinese PVC processing industry was as high as 58,700
60
tons in 2011, and could reach 75,500 tons in 2016.11 Therefore, serious concerns should be
61
raised for the release of SCCPs from the PVC life cycle to the ambient environment.
62
Thermal treatment is an essential process for the processing and recycling of PVC
63
plastics, during which organic additives can be abruptly and prominently released to the
64
ambient environment under thermal stress. PVC processing techniques, including
65
extrusion, calendaring and injection molding, usually operate at 100−210 °C. Meanwhile,
66
the abruptly elevated temperature of PVC plastics usually occurs during open burning of
67
municipal solid waste, melting for PVC recycling, building fire accidents and short circuit
68
current of PVC cables. A previous study found that the emission of di-(2-ethylhexyl)
69
phthalate (DEHP) from PVC plastics increased 100 folds when temperature increased
70
from 20 °C to 80 °C.12 The evaporation rate of DEHP would be independent of DEHP
71
concentration in PVC plastics at ≥ 80 °C, and the transport of DEHP from PVC plastics to
72
the surrounding air was controlled by diffusion at ≥ 120 °C.13 An emission inventory
73
prediction assumed that the emission of SCCPs from PVC processing to air was 4-fold ACS Paragon Plus Environment
Environmental Science & Technology
74
higher than from the routine use of PVC plastics in China,11 according to the default
75
emission factors for the release of additives from polymer processing to air.14 Thus, an
76
investigation of the actual release characteristics of SCCPs from thermally treated PVC
77
plastics should be conducted.
78
During the thermal treatment of PVC plastics, released SCCPs can be rapidly associated
79
with aerosol particles formed from volatile compounds through nucleation and
80
coagulation.15, 16 Gas–particle partitioning, on the one hand, can reduce the gas-phase
81
concentration of SCCPs and thus increase the release rate.17 On the other hand, it can
82
significantly affect the transport and environmental fate of the released SCCPs.18 In
83
addition, the distribution of SCCPs in particle fractions of different sizes will also
84
determine the occupational exposure levels for PVC processing workers or firefighters to
85
SCCPs through inhalation. The primary objective of this study was to investigate the
86
release characteristics of SCCPs from thermally treated PVC plastics and gas–particle
87
partitioning behaviors of the released SCCPs. The obtained results will provide missing
88
information on the source characteristics of SCCPs from PVC processing and other
89
heating processes, and help estimate the health risk of inhalation exposure to SCCPs
90
released from the thermal treatment of plastics.
91
EXPERIMENTAL SECTION
92
Experimental Setup and Sampling. A resilient type of PVC flooring, purchased
93
from a building material market in China, was adopted as the model PVC material. The
94
PVC flooring was composed of a wear layer (pure PVC, thickness: 0.1 mm), a patterned
95
layer (thickness: 0.65 mm) and a foamed layer (thickness: 0.85 mm). The patterned layer
96
and foamed layer contained 60–70% PVC and 30–40% additives (mainly CaO and TiO2)
97
by weight. CPs with chlorine content of 52% by mass (CP-52) were added, and the total
98
content of SCCPs in raw PVC was determined to be 2.36 mg/g. Elemental abundance in
99
PVC flooring was analyzed by X ray fluorescence, and the results are listed in Table S1 of ACS Paragon Plus Environment
Page 4 of 27
Page 5 of 27
Environmental Science & Technology
100
the Supporting Information (SI).
101
The thermal treatment experiments were performed in a stainless steel chamber (length
102
1200 mm × width 900 mm × height 1200 mm) constructed in our lab and equipped with a
103
vent hole on the top and an electric heating plate on the bottom (SI Figure S1). The surface
104
temperature of the electric heating plate was precisely adjusted using a digital temperature
105
controller, and further verified by an infrared radiation thermometer. PVC flooring was
106
chopped into small discs with the same size (diameter: 123 mm, mass: approximately 17 g)
107
for each run. The chopped PVC flooring was placed on the center of the electric heating
108
plate after desired temperature (100 °C, 150 °C and 200 °C) was reached.
109
For the SCCP release experiment, seven sampling time points were established for each
110
temperature condition. At each sampling time point, the thermally treated PVC flooring
111
disc was taken out from the box and immediately weighed, and one seventh of the test
112
PVC flooring piece was cut down for SCCPs analysis. The remaining part of the test PVC
113
flooring piece was immediately placed back in the center of electric heating plate. The
114
weighing and sampling operation was completed within 2 min, and the operation time was
115
not counted as heating time.
116
For the gas–particle partitioning study, a cascade impactor (Mode 122-NR NanoMoudi,
117
MSP Corporation) was applied to collect particles with different diameter at the flow rate
118
of 30 L/min. The sampling nozzle was fixed at the center of the chamber, and the distance
119
from the heating plate was 0.5 m. Sampling was conducted for 8 h at 100 °C and 4 h at
120
150 °C and 200 °C. The average relative humidity in the chamber was 48%, 49% and 47%
121
for the thermal treatment at 100 °C, 150 °C and 200 °C, respectively.
122
There are 13 stages in the cascade impactor, and the 50% cutoff diameters ranged from
123
10 nm to 18000 nm. Uncoated aluminum foil was used as the impaction substrate to
124
collect the particles on impactor plates, followed by a back-up filter (Quartz Microfiber
125
Filters, WhatmanTM, 90 mm) to collect uncaptured particles. Polyurethane foam (PUF) ACS Paragon Plus Environment
Environmental Science & Technology
126
was placed under the back-up filter to collect SCCPs in the gas phase. Aluminum foil was
127
baked at 500 °C for 12 h in a muffle oven to remove organic substances. PUF was washed
128
with n-hexane and dichloromethane (DCM) three times using accelerated solvent
129
extraction (ASE 350, Dionex, USA). The impaction substrate was taken from the cascade
130
impactor and kept in an incubator for 24 h to reach equilibrium after sampling. The mass
131
of aluminum foil at each stage was measured on a microbalance (Mettler Greifensee,
132
Switzerland; precision: ± 1 µg) before and after the experiment, repeated three times and
133
averaged.
134
Sample Extraction and Cleanup. Raw PVC flooring samples and thermally treated
135
PVC flooring samples were cut and ground into small pieces (through 24 mesh sieve) with
136
liquid nitrogen, and 0.3 g of PVC samples were ultrasonically extracted with 20 mL
137
n-hexane/DCM (1: 1, v/v) three times (15 min for each run). Extraction of particles on
138
aluminum foils followed the same procedure as PVC samples. PUF and quartz microfiber
139
filters were Soxhlet-extracted for 24 h with 250 mL n-hexane/DCM (1: 1, v/v). Prior to
140
extraction, 2 ng of
141
samples, as the extraction internal standard. For PVC flooring samples, 2 µg of chlorinated
142
2-methyl undecane (branched C12-CPs, chlorine content: 55.4% by mass) was spiked as
143
the extraction internal standard.
13
C6-δ-hexachlorocyclohexane was spiked into particle and PUF
144
A silica gel column was used to clean up samples. The column was packed with 5.0 g
145
anhydrous Na2SO4, 3.5 g silica gel (activated at 650 °C for 12 h) and 6.0 g anhydrous
146
Na2SO4 from bottom to top. The extract was transferred to the column after concentration
147
to approximately 1 mL. The fraction was received by eluting with 40 mL n-hexane/DCM
148
(2: 1, v/v). The target solution was concentrated to approximately 0.5 mL, transferred to a
149
vial, evaporated to near dryness with a gentle stream of N2, and re-dissolved in 40 µL of
150
injection internal standard solution for instrumental analysis. Chlorinated 2-methyl nonane
151
(branched C10-CPs) and
13
C10-trans-CD were adopted as injection internal standards for ACS Paragon Plus Environment
Page 6 of 27
Page 7 of 27
Environmental Science & Technology
152
extracts of PVC flooring samples and extracts of particle and PUF samples, respectively.
153
Detailed information about chemicals and reagents is shown in the SI.
154
Instrumental Analysis and Quantification. In this study, the release of SCCPs
155
from the heating of PVC flooring was calculated according to the concentration loss of
156
SCCPs in the flooring before and after thermal treatment. In the tested PVC flooring
157
material, medium-chain CPs (MCCPs) were detected at high concentrations (SI Figure S2).
158
To avoid strong interferences of MCCPs with SCCP analysis on mass spectrometry,19 the
159
carbon skeleton reaction gas chromatography (GC) method was adopted for the analysis
160
of SCCPs in PVC flooring. The analysis was conducted on a Hewlett-Packard GC (HP
161
5890 Series II) equipped with a capillary DB-5 column (30 m × 0.32 mm i.d. × 0.25 µm
162
film thickness, J&W Scientific, USA) and FID detector. The concentrations of n-alkanes
163
(C10−13), produced from the dechlorination reaction were calibrated using n-alkane
164
standards (purity: > 99.9%); thereafter the concentrations of SCCPs with different carbon
165
chain lengths were quantified with corrections of the internal standard recovery rate and
166
average chlorine content (52% by mass). Detailed information about carbon skeleton
167
reaction GC method is shown in the SI.
168
In particle and PUF samples, SCCPs were the predominant CP components due to less
169
volatilization of MCCPs from the heating of PVC flooring (SI Figure S3), and thus the
170
interference of MCCPs in SCCP analysis on mass spectrometry could be ignored. In this
171
study, the SCCPs in particle and PUF samples were analyzed using a high-resolution gas
172
chromatography/electron capture negative ion−low resolution mass spectrometry
173
(HRGC/ECNI−LRMS) method, in which information about both the carbon chain lengths
174
and chlorine distribution of SCCPs could be obtained. The analysis was conducted on a
175
GCMS-QP2010 (Shimadzu, Japan) coupled with a capillary DB-5 column (15 m × 0.25
176
mm i.d. × 0.25 µm film thickness, J&W Scientific, U.S.). The quantification of SCCPs
177
followed the procedure described by Reth et al.20 Detailed information about ACS Paragon Plus Environment
Environmental Science & Technology
178
HRGC/ECNI−LRMS method and SCCP quantification is shown in the SI.
179
Quality Assurance and Quality Control. During particle sampling, the measured
180
gas flow rates and pressure drops of the cascade impactor all met the specified values. The
181
relative deviation in particle mass at each stage, measured by microbalance three times,
182
was less than 10%. For the analysis of SCCPs, procedural blanks were first performed for
183
each batch of solvents and adsorbents. No SCCPs were detected in the procedural blanks.
184
The procedural blanks spiked with a small quantity of SCCP mixture standards were used
185
to determine the method detection limit (MDL) of HRGC/ECNI−LRMS analysis, which
186
was calculated as three times the standard deviation at 4 time measurements. The MDL
187
value for the total SCCPs (∑SCCPs) in particle and PUF samples was estimated to be 0.49
188
ng/m3. In the gas–particle partitioning study, the field blank samples were obtained by
189
sampling ambient air through the cascade impactor. The concentrations of SCCPs in field
190
blank samples of back-up filter and aluminum foil were below MDL. SCCP level in field
191
PUF blank was determined to be 79.3 ng, which was 3−4 orders of magnitude lower than
192
those (54.3−815.8 µg) of field samples.
193
For the HRGC/ECNI−LRMS analysis of SCCPs, the recoveries of internal standard
194
13C6-δ- hexachlorocyclohexane in all particle and PUF samples were in the range of
195
51.4%−135.8%. The reported concentrations of SCCPs in particle and PUF samples were
196
not corrected by the recovery rates of extraction internal standard and field blanks. For the
197
carbon skeleton reaction GC analysis of SCCPs, the internal standard chlorinated 2-methyl
198
undecane in all PVC flooring samples ranged from 70.5% to 140.3%.
199
RESULTS AND DISCUSSION
200
Release Kinetics of SCCPs. An abrupt and prominent release of SCCPs was
201
observed during the initial stage of thermal treatment of PVC flooring. Within the initial
202
1-h, the overall loss of ∑SCCPs from PVC flooring mass was 1.9%, 7.4% and 10.7% at
203
heating temperatures of 100 °C, 150 °C and 200 °C, respectively (SI Table S3). SCCP ACS Paragon Plus Environment
Page 8 of 27
Page 9 of 27
Environmental Science & Technology
204
release rates gradually decreased with increased heating time, and SCCP congener groups
205
with shorter carbon chain lengths exhibited higher release rates (Figure 1 and SI Figure
206
S4).
207
208 209
Figure 1. Comparison of fitted and measured accumulative release of SCCP congener groups
210
from thermal treatment of PVC flooring.
211 212
A simplified schematic representation for the release of SCCPs from thermally treated
213
PVC flooring in the stainless steel chamber is shown in SI Figure S5. The release of
214
SCCPs from plastic materials can generally be described with three physical processes: (1)
215
material-phase mass-transfer (internal diffusion), (2) material−gas partitioning (boundary
216
condition) and (3) gas-phase mass-transfer (external diffusion).17,
217
diffusion coefficient D, material−gas partition coefficient K and convective mass-transfer
218
coefficient hm are the corresponding parameters governing the three processes. Based on
219
emission model of VOCs from building materials,22−24 a model that integrates these three
220
physical processes governing release of SCCPs from the thermally treated PVC flooring
221
was developed in this study, and model details are shown in the SI. When the key
222
coefficients K, D and hm were known, the model could be numerically solved by using the
223
finite difference method.24 ACS Paragon Plus Environment
21
Material-phase
Environmental Science & Technology
224
In this emission model, the coefficient hm for the laminar flow of ∑SCCPs was
225
evaluated according to the relationship of Sherwood number (Sh) with Reynolds number
226
(Re) and Schmidt number (Sc).23, 25 The calculated hm values were 0.017 m/s, 0.019 m/s
227
and 0.021 m/s at heating temperatures of 100 °C, 150 °C and 200 °C, respectively. The
228
coefficients K and D were obtained by fitting the variation of measured residual
229
concentrations of SCCPs in PVC flooring with heating time, based on the
230
Levenberg–Marquardt algorithm (details in SI). Before optimizing, a careful check for the
231
relative sensitivity of coefficients D and K was conducted. The result showed that no
232
linear dependence existed between D and K values (SI Figure S6), and thus these two
233
coefficients could be independently estimated by the inverse method. The fitted values for
234
the accumulative release of various SCCP congener groups were all significantly (P < 0.01)
235
correlated with the measured values (Figure 1 and SI Table S4), validating the developed
236
model. As shown in SI Table S5, the obtained K and D values both depended strongly on
237
the heating temperature. In addition, SCCP congener groups with shorter carbon chain
238
lengths showed larger D values and smaller K values.
239
The time-dependent release rates of SCCPs were predicted by the developed model
240
with estimated parameters (Figure 2). During the initial 1-h of thermal treatments at
241
150 °C and 200 °C, SCCP release rates declined dramatically. The rapid decreases resulted
242
from fast surface volatilization during thermal stress, which was controlled by the
243
material–gas partition coefficient K and convective mass-transfer coefficient hm.
244
Subsequently, SCCP release rates fell, suggesting that the release control was gradually
245
subjected to the material-phase diffusion. SCCP congener groups with shorter carbon
246
chain lengths were more easily volatilized from the surface of thermally treated PVC
247
flooring, and thus their release was more inclined to be limited by material-phase diffusion.
248
At the heating temperature of 100 °C, SCCP release rate did not present an obviously
249
sharp drop during the initial 1-h of thermal treatment, indicating that the material–gas ACS Paragon Plus Environment
Page 10 of 27
Page 11 of 27
Environmental Science & Technology
250
partition decisively affected the release rate of SCCPs from thermal treatment of PVC
251
flooring at 100 °C.
252
253 254
Figure 2. Predicted release rate of SCCP congener groups from thermal treatment of PVC
255
flooring.
256 257
Mass Size Distribution of the Emitted Particles. During the 4-h thermal
258
treatment of PVC flooring, the total mass of particles captured by the NanoMoudi stages
259
was 85.3 mg at heating temperature of 200 °C, which was 4.4-fold higher than that at
260
150 °C. Only 4.1 mg of particles was captured during 8-h thermal treatment at 100 °C.
261
The emitted particles generally exhibited a unimodal mass size distribution (SI Figure S7).
262
More than 90% of particles were deposited at Stages 6−10 (50% cutoff diameter: 56−1000
263
nm) at 100 °C, Stages 5−8 (50% cutoff diameter: 180−1800 nm) at 150 °C and Stages 3−8
264
(50% cutoff diameter: 180−5600 nm) at 200 °C, respectively. A continuous mass size
265
distribution of emitted particles was reconstructed using the data inversion method
266
described by Dong et al.26 The lognormal functions fit the gravimetric NanoMoudi data
267
well (significant at 0.01 level) (SI Figure S8), and converged to the following parameters:
268
geometric mean diameter (Dg) was 358 nm, 919 nm and 1990 nm for thermal treatment at
269
100 °C,150 °C and 200 °C, respectively (SI Table S6). The elevated temperature gave rise
ACS Paragon Plus Environment
Environmental Science & Technology
270
to an increase in larger particle yield. Scanning electron microscope images showed that
271
these larger aerosol particles mainly formed from the coagulation of ultrafine mode
272
particles (UFPs, size: < 100 nm) (SI Figure S9). Meanwhile, energy dispersive X-ray
273
spectrometer indicated that the main elements in aerosol particles captured by NanoMoudi
274
Stages 5−9 (50% cutoff diameter: 100−1000 nm) were C, O, Cl, Ca, Si, Mg and Ti (SI
275
Table S7). This suggested that these larger aerosol particles were mainly composed of
276
chlorine-containing organic compounds, and plastic additives such as calcium oxide,
277
silicon compounds, magnesium oxide and titanium oxide.
278
279 280
Figure 3. Mass size distribution reconstructed from NanoMoudi impactor data of
281
particle-bound SCCPs produced from thermal treatment of PVC flooring. Unit dm is mass of
282
SCCPs captured at certain stage when sampling volume is 1 m3, and Dp is the aerodynamic
283
diameter of particles deposited at a certain stage.
284 285
Distribution of the Total SCCPs in Gas and Particle Phases. During the 8-h
286
thermal treatment of PVC flooring at 100 °C, NanoMoudi captured 909.2 µg of SCCPs,
287
89.7% of which was in the gas phase (adsorbed on PUF). However, most emitted SCCPs
288
partitioned into the particle phase during the 4-h thermal treatment at 150 °C and 200 °C.
ACS Paragon Plus Environment
Page 12 of 27
Page 13 of 27
Environmental Science & Technology
289
The total masses of particle-bound SCCPs (captured by different stages and back-up filter)
290
were 634.1µg and 1262.0 µg at 150 °C and 200 °C, accounting for 64.5% and 95.9% of
291
the total captured SCCPs, respectively. The mass size distribution of SCCPs in the
292
particle-phase also featured a unimodal distribution (SI Figure S10). According to the data
293
inversion method,26 a continuous mass size distribution of particle-bound SCCPs was
294
reconstructed. As shown in Figure 3, the log-normal functions provided a good fit
295
(significant at 0.01 level) for the NanoMoudi impactor data of particle-bound SCCPs. The
296
fitting results indicated that of SCCP mass peaks occurred at particle sizes of 359 nm, 894
297
nm and 1725 nm for thermal treatments at 100 °C, 150 °C and 200 °C, respectively (SI
298
Table S8).
299
The concentrations of ∑SCCPs in particles showed obvious variation with changes in
300
particle size and heating temperature (Figure 4). In the ultrafine particles (UFPs) deposited
301
on Stages 10−13 (50% cutoff diameter: 10−56 nm), the ∑SCCPs concentrations were in
302
the range of 0.5−13.7 mg/g. Due to the larger deviation in mass measurement, the
303
concentrations of ∑SCCPs in the smallest UFPs deposited on the back-up filter (50%
304
cutoff diameter: < 10 nm) could not be accurately determined. ∑SCCPs concentrations in
305
the accumulation mode particles, consisting of particles deposited on Stages 3−9 (50%
306
cutoff diameter: 100−3200 nm), were in the range of 6.9−46.9 mg/g, significantly higher
307
than those in UFPs. The accumulation mode particles emitted from thermal treatment at
308
150 °C showed higher ∑SCCPs concentrations compared to those at 100 °C and 200 °C.
309
However, thermal treatment at 200 °C led to a higher load of SCCPs in coarse mode
310
particles deposited on the inlet stage and Stages 1 and 2 (50% cutoff diameter: 5.6−18 µm).
311
The average concentration of ∑SCCPs in coarse mode particles emitted from thermal
312
treatment at 200 °C was 37.2 mg/g, approximately 49-fold and 20-fold higher than those
313
at 100 °C and 150 °C, respectively.
314
ACS Paragon Plus Environment
Environmental Science & Technology
315 316
Figure 4. SCCP concentrations in size separated particles (collected by NanoMoudi) from
317
thermal treatment of PVC.
318 319
The UFPs comprise two sub-modes, i.e. nucleation mode (< 10 nm) and Aitken mode
320
particles (10−100 nm).27 During thermal treatment of PVC flooring, nucleation mode
321
particles can be formed through nucleation in the ambient air after rapid cooling and
322
dilution of emissions when the saturation ratio of gaseous compounds of low volatility (i.e.
323
HCl and larger organic molecules) reached a maximum.28−30 Subsequently, Aitken mode
324
particles were formed through collision and condensational growth of nucleation mode
325
particles.31 In this study, the concentrations of ∑SCCPs in UFPs with a 50% cutoff
326
diameter of 10−32 nm were lower, implying that SCCPs were not strongly inclined to
327
form aerosol particles by nucleation.
328
During the thermal treatment of PVC flooring, accumulation mode particles could be
329
produced by coagulation and surface growth of UFPs. Because UFPs had a lower
330
concentration of ∑SCCPs, the coagulation of UFPs did not seem to result in a larger
331
increase in ∑SCCPs concentration in the formed accumulative mode particles. The
332
significantly elevated concentration of ∑SCCPs in accumulation mode particles should
333
mainly result from the sorption of gas-phase SCCPs during the surface growth of particles.
334
Due to the hydrophobicity, volatilized SCCPs should be partitioned into only the organic ACS Paragon Plus Environment
Page 14 of 27
Page 15 of 27
Environmental Science & Technology
335
phase of particles. Therefore, it could be speculated that the outer layer of accumulation
336
mode particles should have higher organic matter content due to the continuous surface
337
condensation and adsorption of organic vapors.
338
During thermal treatment of PVC flooring at 200 °C, coarse mode particles were
339
deposited on the inlet stage and Stages 1 and 2 (50% cutoff diameter: 5.6−18 µm) were
340
found to be yellow organic liquid residues, suggesting that these coarse mode particles
341
were in an organic liquid state. These organic liquid particles should be formed by the
342
rapid condensation of semi-volatile organic compounds that were abruptly emitted from
343
the PVC flooring under higher thermal stress. They inevitably exhibited a larger capacity
344
to absorb the released SCCPs, and thus resulted in the high accumulation of SCCPs.
345
Distribution of SCCP Congener Groups in Gas- and Particle-Phases. To
346
understand the dependence of the gas−particle partitioning behavior of SCCPs on their
347
physical-chemical properties, SCCP congeners were categorized by both carbon chain
348
length and chlorine atom number into 24 congener groups, and their relative abundances
349
were compared (SI Figure S11). Thermal treatments of PVC flooring at three different
350
heating temperatures (100 °C, 150 °C and 200 °C) produced similar congener group
351
profiles of gas-phase SCCPs, which were characterized by the higher relative abundances
352
of C10-CPs (average: 82.6%) groups. In the particle-phases, the proportion of C10-CPs
353
largely decreased, whereas the proportions of C11−13-CPs increased noticeably.
354
Principal component analysis (PCA) was applied with unit variance (UV) scaling for
355
data regarding the relative abundances of SCCP congener groups to further investigate
356
their similarities and differences with respect to gas−particle partitioning behavior. The
357
two dimensions represented the two main factors, which explained most of the original
358
variables with a cumulative contribution of 74.0% (Figure 5). The loading factors of the
359
SCCP congener groups exhibited an increasing tendency with the increase of log KOA
ACS Paragon Plus Environment
Environmental Science & Technology
360
values on loadings 1 (SI Figure S12), suggesting that the first factor (Component 1 in the
361
score plot) was mainly explained by variables induced by the KOA values.
362 363 364 365 366 367 368 369 370 371 372 373
Figure 5. PCA score plot for congener group abundances of gas-phase and particle-phase
374
SCCPs released from thermal treatments of PVC flooring at three heating temperatures
375
(100 °C, 150 °C and 200 °C). The 50% cutoff diameters of nucleation mode particles, Aitken
376
mode particles, accumulation mode particles and coarse mode particles, were < 10 nm,
377
10−100 nm, 100−3200 nm and > 3200 nm, respectively.
378 379
In the PCA score plot, gas-phase and nucleation mode particles (50% cutoff diameter: