Subscriber access provided by University of Florida | Smathers Libraries
Article
Heterocyclic aromatics in petroleum coke, snow, lake sediments and air samples from the Athabasca oil sands region Carlos A Manzano, Christopher H. Marvin, Derek C.G. Muir, Tom Harner, Jonathan W. Martin, and Yifeng Zhang Environ. Sci. Technol., Just Accepted Manuscript • Publication Date (Web): 28 Apr 2017 Downloaded from http://pubs.acs.org on April 30, 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 24
Environmental Science & Technology
Heterocyclic aromatics in petroleum coke, snow, lake sediments and air samples from the
1
Athabasca oil sands region
2 3
Carlos A. Manzano1, *, Chris Marvin1, Derek Muir1, Tom Harner2, Jonathan Martin3, Yifeng Zhang3
4 5 6
1
Aquatic Contaminants Research Division, Environment & Climate Change Canada, Burlington, ON
7
2
Air Quality Processes Research Section, Environment & Climate Change Canada, Toronto ON
8
3
Division of Analytical and Environmental Toxicology, University of Alberta, Edmonton, AB
9 10 11
Abstract The aromatic fractions of snow, lake sediment, and air samples collected during 2011-2014 in the Athabasca Oil
12
Sands Region were analyzed using two-dimensional gas chromatography following a non-targeted approach.
13
Commonly monitored aromatics (parent and alkylated-polycyclic aromatic hydrocarbons, dibenzothiophenes) were
14
excluded from the analysis, focusing mainly on other heterocyclic aromatics. The unknowns detected were classified
15
into isomeric groups, and tentatively identified using mass spectral libraries. Relative concentrations of heterocyclic
16
aromatics were estimated, and were found to decrease with distance from a reference site near the center of the
17
developments, and with increasing depth of sediments. The same heterocyclic aromatics identified in snow, lake
18
sediments and air were observed in extracts of delayed petroleum coke, with similar distributions. This suggests that
19
petroleum coke particles are a potential source of heterocyclic aromatics to the local environment, but other oil
20
sands sources must also be considered. Although the signals of these heterocyclic aromatics diminished with
21
distance, some were detected at large distances (>100 km) in snow and surface lake sediments, suggesting that the
22
impact of industry can extend >50 km. The list of heterocyclic aromatics and the mass spectral library generated in
23
this study can be used for future source apportionment studies.
24
ACS Paragon Plus Environment
Environmental Science & Technology
25
Introduction
26
The Oil Sands deposits of Alberta are considered the third world’s largest oil reserve,1 with proven reserves of
27
~1.7 trillion barrels of bitumen distributed in the Athabasca, Cold Lake, and Peace River deposits.2 In the Athabasca
28
oil sands region (AOSR), bitumen is extracted using open pit mining and in-situ extraction technology, and it
29
undergoes on-site upgrading to produce a synthetic light crude oil and diluent for transport of bitumen by pipeline.3
30
The upgrading process, which includes coking and catalytic hydrocracking, results in large volumes of residual
31
petroleum coke (petcoke).4 Petcoke is a granular carbonaceous residual product, it has high sulfur content and high
32
heavy aromatics-to-aliphatic ratios,5 and it is currently being stockpiled in the AOSR as a potential future source of
33
energy.4
34
The local oil sands industry in the AOSR is known to release polycyclic aromatic compounds (PACs), volatile
35
organics, and priority pollutant elements to the atmosphere.6-19 PACs are ubiquitous organic contaminants that are
36
present as complex mixtures of a wide range of homologues and congeners having different molecular weights and
37
structures. Mixtures of PACs can include compounds with only carbon and hydrogen atoms in their structure, such
38
as unsubstituted polycyclic aromatic hydrocarbons (unPAHs) and alkylated PAHs (aPAHs), and also heterocyclic
39
compounds in which one or more carbon atoms are replaced by nitrogen (aza-arenes) or sulfur atoms (thia-arenes).20
40
The former (unPAHs, and C1-C4-substituted aPAHs) and some selected heterocyclic aromatics (C0-C4-
41
dibenzothiophenes (DBTs)) have been extensively studied in the AOSR and are included in ongoing monitoring
42
programs.21 Other heterocyclic aromatics apart from DBTs have generally not been included. Their inclusion could
43
provide a great deal of complementary information for source identification and improve assessments of risks of
44
exposure and health effects.22-24
45
There are natural and anthropogenic sources of PACs in the AOSR: stack emissions from bitumen upgrading,
46
diesel exhaust from mining and transportation equipment, haul roads, airborne particulate matter from the mines and
47
land disturbances, forest fires and naturally exposed bitumen. It has been suggested that atmospheric partitioning of
48
PACs from tailing ponds can also be a potentially significant source, although there is uncertainty in this
49
hypothesis.25-28 More recent studies have suggested that atmospheric deposition of petcoke particles can be a major
50
source of PAHs within ~30 km from the upgraders,29 and there is evidence for their presence in lake sediments as far
51
as 150 km away.30
ACS Paragon Plus Environment
Page 2 of 24
Page 3 of 24
Environmental Science & Technology
52
Airborne petcoke has the potential to exacerbate pre-existing lung ailments and may have additive or synergistic
53
effects with other environmental toxins, with incidental ingestion of fugitive dust and fibers as the primary pathway
54
for human exposure in occupational studies.31 Little is known about the effects of petcoke as fine particulate matter
55
on exposed populations and ecosystems, nor about petcoke ecotoxicity. Therefore, improved chemical
56
characterization of petcoke particles and fibers, and a more rigorous quantification of fugitive dust emissions from
57
storage piles, are required. Petcoke derived from the AOSR has a high sulfur content,4 and therefore potentially high
58
abundance of heterocyclic aromatics. Heterocyclic aromatics have been described in petcoke from other sources,32, 33
59
however to our knowledge petcoke from the AOSR has never been characterized for its heterocyclic aromatics
60
content.
61
The chromatographic analysis of heterocyclic aromatics (i.e. thia-arenes, aza-arenes) has shown to be
62
problematic for some matrices containing also unPAHs and aPAHs.34-37 Therefore, any chromatographic method
63
used for the analysis of complex mixtures must show enough selectivity to separate these groups. In this context,
64
two-dimensional gas chromatography coupled with time-of-flight mass spectrometry (GC×GC/ToF-MS), and using
65
non-traditional GC columns, was recently utilized to separate thia-arenes and aza-arenes from other PACs in
66
samples collected in the AOSR.38
67
This study focused on the analysis of novel heterocyclic aromatics in the aromatic fraction of snow, lake
68
sediment and passive air samples collected from the AOSR between 2011-2014 using GC×GC/ToF-MS. The
69
relative abundance and percent distribution of 259 heterocyclic aromatics were evaluated and compared to those
70
found in petcoke extracts obtained from Zhang et al.29 The results from this study provide preliminary source
71
identification using patterns of heterocyclic aromatics, and include a mass spectral library that can be used as
72
reference for future research in the AOSR and more detailed source apportionment studies.
73 74
Materials and Methods
75
Snow Samples: Snow samples for this study were collected in late winter 2012, as part of a larger study under the
76
Joint Canada-Alberta Monitoring (JOSM) program,21 focusing on the deposition of mercury and PACs in snow.13,18
77
Samples were collected during late February and early March 2012 in the Regional Municipality of Wood Buffalo in
78
northern Alberta, at varying distances from a reference point located close to the main development area (57.018 N;
79
-111.485 W) (Figure 1, Table S1) on the Athabasca River and near the Suncor upgrader. Samples were kept frozen
ACS Paragon Plus Environment
Environmental Science & Technology
80
until processing at the Environment & Climate Change Canada (ECCC) Centre for Inland Waters (Burlington,
81
Ontario). Melted snow was filtered by pumping through a GF/F filter connected to a 30 cm Teflon column (2 cm
82
i.d.) packed with 50 g of pre-cleaned XAD-2 resin. Filters were then extracted using a 1:1 hexane-acetone mixture
83
followed by 100% dichloromethane (DCM) using a DionexTM ASETM 350 pressurized liquid extraction system
84
(PLE) (Thermo Fisher Scientific, Waltham, MA, USA). XAD-2 resins were extracted using an elution column with
85
acetone and DCM. Both DCM fractions were back-extracted using a 3% sodium chloride solution, dried using
86
sodium sulfate, combined and concentrated. The extracts were further processed at the ECCC Air Quality Research
87
Division laboratory (Ottawa, Ontario), where they were fractionated on silica gel solid phase columns with 100%
88
hexane followed by 100% benzene (aromatic fraction). The aromatic fractions were analyzed by GC×GC/ToF-MS.
89
Further details on pre- and post-deployment sample cleanup and processing can be found in Manzano et al,18, 38 and
90
in the supporting information.
91
Lake sediment samples: Sediment cores were obtained from the center of small lakes with undisturbed catchments
92
and water depths deeper than 1.5 m, located varying distances from the reference site near the center of the
93
development area.19, 39 Four lakes were used in this study, and were located at 10, 16, 35, and 100 km from the
94
reference site (Figure 1, Table S1). Sediment cores were originally sampled contiguously at 0.5 cm resolution for the
95
upper 20 cm, and at 1 cm resolution below 20 cm. Age of each sediment section was determined using 210Pb, 137Cs
96
and 226Ra radio isotopic techniques and the constant rate of supply model,19 and was completed by Flett Research,
97
Ltd. (Winnipeg, Manitoba, Canada). Samples were processed by AXYS Analytical Services (Sidney, BC, Canada)
98
using their method MLA-021, which is based on US Environmental Protection Agency (EPA) methods 1625B and
99
8270C/D.19 Briefly, samples were defrosted and homogenized manually, and Soxhlet extracted with DCM for 18 h.
100
The extracts were then fractionated on a silica column with 100% pentane followed by 100% DCM. The DCM
101
fractions were then subjected to alumina cleanup using hexane and DCM (aromatic fraction). The aromatic fractions
102
were analyzed by GC×GC/ToF-MS. Only 3 sections of each lake core were used. Based on previous publications,
103
sections that were selected corresponded to different stages of the industrial development in the area: (a) prior to or
104
during the very early stages of large scale industrial development of the AOSR (i.e. ‘deep sediments’, dated pre-
105
1970), (b) low to intermediate growth (i.e. ‘mid sediments’, dated ~1970-2000); and (c) recent rapid expansion (i.e.
106
‘surface sediments’, dated post ~2000).40 Further details on sediment sampling protocols and sample processing and
107
cleanup can be found in Kurek et al.19
ACS Paragon Plus Environment
Page 4 of 24
Page 5 of 24
Environmental Science & Technology
108
Air samples: Air samples were collected at the AOSR in 2014, as part of a larger monitoring program under
109
JOSM.41, 42 Polyurethane foam (PUF) passive air samplers were used due to logistical challenges associated with
110
monitoring in this largely underdeveloped and remote region. PUF disks were deployed in double-domed sampling
111
chambers, identical to those used in the Global Atmospheric Passive Sampling network (GAPS).43 Passive samplers
112
were exposed for consecutive two-month periods at 17 sites and a subset of 4 sites were selected for this study based
113
on their proximity to the reference site, corresponding to April and May 2014 (Figure 1, Table S1). All pre-
114
deployment and post-deployment cleanup have been described previously.41, 42 Briefly, PUF disks were cleaned
115
using PLE with acetone, petroleum ether (PE) and acetonitrile prior to deployment, and dried under nitrogen. The
116
PUF disks retrieved were extracted using PE/acetone (75/25, v/v; 2 cycles). All extracts were then fractionated on a
117
silica gel column with PE, PE/acetone (50/50, v/v) (aromatic fraction), and methanol. The aromatic fractions were
118
analyzed by GC×GC/ToF-MS.
119
Petcoke extracts: Authentic petcoke samples from two different coking technologies (i.e. fluid and delayed) used in
120
bitumen upgraders in the AOSR were obtained and processed by the Division of Analytical and Environmental
121
Toxicology at the University of Alberta in Edmonton. The same samples were used in Zhang et al.29 Approximately
122
10 g of each sample were extracted using 250 mL of DCM twice, followed by 250 mL of methanol. The extracts
123
were then filtered using a GB-140 Glass Fiber Membrane filter with 90 mm diameter and pore size 0.4 µm. The
124
extracts were then concentrated using a rotary evaporation, and transferred to small GC vials. A procedural blank
125
was carried through the same processing. All the above pre-treatment for petcoke was performed in an ultraclean
126
organics lab at the University of Alberta.
127
GC×GC/ToF-MS analysis: Snow, lake sediments, passive air samples, and petcoke extracts were analyzed using a
128
GC×GC/ToF-MS Pegasus 4D (Leco, St Joseph, MI, USA). The instrument consisted of an Agilent 7890B gas
129
chromatograph (Palo Alto, CA, USA) equipped with a secondary oven, split/splitless injector and a consumable-free
130
modulator operated at -80 ºC. The first-dimension column was a liquid crystalline LC-50 (10 m × 0.18 mm × 0.10
131
µm) (J&K Scientific, Edwardsville, Nova Scotia, Canada) followed by a nano-stationary phase NSP-35 (1 m × 0.15
132
mm × 0.10 µm) (J&K Scientific) in the second dimension. The two GC columns were connected using an Agilent
133
CPM union (part No. G3182-61580). The data processing and analysis was performed using ChromaToF v.4.50.8,
134
and further statistical analysis was completed using SigmaPlot v.12.5. Further details on method development and
135
column configuration can be found in previous publications,38, 44 and in the supporting information (Table S2).
ACS Paragon Plus Environment
Environmental Science & Technology
A standard solution containing 13 PACs, purchased from ChemServices (West Chester, PA, USA), at 100 pg µL-
136 137
1
138
QA/QC purposes and to test for carryover throughout the analysis. Additionally, d10-fluoranthene and d12-chrysene
139
were used as internal standards. The standard solution was injected every 7 samples (for snow extracts), and every 3
140
samples (for air and lake sediment extracts), and showed consistent relative standard deviations (RSD) for total peak
141
areas (average 41%, range 32-48%), peak full width at half height (average 5%, range 2-11%) and peak height
142
(average 39%, range 30-46%). Blank toluene samples, and field blanks did not show detectable signals for
143
heterocyclic PACs. The internal standards also showed consistent peak areas, with RSDs of 23% for all snow
144
samples (n=25), 22% for all lake sediment samples (n=12), and 48% for all passive air samples (n=4).
in toluene, as well as field blank samples collected with snow samples, and blank toluene solutions were used for
145
Concentrations of heterocyclic aromatics could not be precisely determined due to the lack of authentic
146
standards. However, their relative abundance was determined using the ratio of the total peak area of the target
147
compound to the total peak area of an internal standard (i.e.: d10-fluoranthene for snow and lake sediments, d12-
148
chrysene for air samples). Total peak area was determined using the three major modulated peaks, following a
149
partial summation method.45, 46 Additionally, first dimension retention indexes (RIs) for all PACs identified were
150
determined using fluorene (RIstd = 200), phenanthrene (RIstd = 300), chrysene (RIstd = 400), benzo[a]pyrene (RIstd =
151
450), and benzo[g,h,i]perylene (RIstd = 500) as bracketing compounds with their respective RI constant values
152
(RIstd).
153
A snow sample, collected at the reference site (snow site S1) and previously known to have elevated
154
concentrations of PACs,6, 18 was analyzed first as a reference matrix. ChromaToF was arbitrarily set to identify
155
3,000 peaks with signal-to-noise ratios (S/N) higher than 50. Peaks coming from column bleed and solvent residues
156
were manually excluded from the results, along with commonly monitored unPAHs, aPAHs and DBTs.
157
The remaining peaks were classified into potential isomeric groups based on their mass spectra (MS), their
158
retention times, and observed elution patterns. Because molecular similarity grouping is not evident when using LC-
159
50×NSP-35,38 the groups were identified using specific m/z fragments rather than the total ion chromatogram
160
(Figure S1). The groups were completed by manually adding peaks with similar MS and S/N>50 that ChromaToF
161
could have missed during data processing. A total of 259 peaks were classified into 21 isomeric groups, with one
162
group divided into three subgroups (i.e. Unknowns (U)-276: G1, G2, and G3). The molecular and confirmation ions
ACS Paragon Plus Environment
Page 6 of 24
Page 7 of 24
Environmental Science & Technology
163
for all groups identified can be found in the supporting information (Table S3). The data processing and analysis of
164
all samples was completed by a single GC×GC operator to reduce potential errors due to selection criteria.
165 166 167
Results and Discussion Each of the 21 groups shared a common MS pattern among its member compounds. Therefore, the combined or
168
‘group MS’ was determined to reduce potential errors due to matrix interferences in individual compounds. These
169
combined MS were obtained by averaging the relative abundances for all ion fragments in individual MS for each
170
member of the group, and considered only those fragments present in more than one isomer with relative
171
abundances higher than 1%. The combined MS for the 21 groups can be found in the supporting information (Figure
172
S2). Peak name assignment was based on a 70% threshold compared to the NIST MS library 2011 (Figure 2, Figure
173
S2), otherwise the group was reported as ‘Unknown’ (U).
174 175 176
Relative abundance of heterocyclic aromatics: The 259 heterocyclic aromatics identified in snow at site S1 were used to build a reference method for the
177
analysis of other snow, lake sediments, air samples and petcoke extracts. This time, peak location in the 2D-
178
chromatogram, RIs and MS similarities >50% to the reference sample were used for positive peak identification.
179
RSDs for all RIs determined in all samples ranged from 0-0.6%, with the largest RSDs observed in late eluting
180
compounds (e.g. U-318, -324, -330 and -332) (Table S4). The RSDs determined were highest in the near-field snow
181
samples, which was likely a consequence of a strong matrix effect. This idea was reinforced by the background
182
noise level observed, calculated based on the default method in ChromaToF and using the standard deviation of the
183
baseline, which was also higher in near-field snow (range 7-21), compared to delayed petcoke (range 4-6), near-field
184
surface lake sediment (range 3-5), and near-field air (range 3-4) (Table S4). However, and despite the differences
185
observed in noise, background level and RIs, the spatial distribution of compounds in the 2D-chromatogram was
186
similar among all samples and extracts (Figure 3). Although differences in individual relative abundances were
187
observed and not all compounds were present in all media.
188
Petcoke: Fluid and delayed petcoke are both derived from bitumen with high sulfur content (up to 8%),47 which can
189
be an indicator for the presence of heterocyclic aromatics. However, delayed petcoke is produced at lower
190
temperatures than fluid petcoke, and therefore it is known to produce larger quantities of volatiles and show higher
ACS Paragon Plus Environment
Environmental Science & Technology
191
concentrations of PAC than fluid petcoke.29, 47 In this study, the 259 heterocyclic aromatics identified in snow at site
192
S1 were also found in delayed petcoke. Fluid petcoke and the procedural blank showed low relative abundance of
193
the 259 heterocyclic aromatics (S/N> L3 > L4). This is consistent with local open pit mining of bitumen in the area near the
217
Muskeg River (where site L2 is located) compared to deposits near the McKay River (site L1) (Figure 1). Relative
218
abundances for all heterocyclic aromatics identified in lake sediments can be found in the supporting information
ACS Paragon Plus Environment
Page 8 of 24
Page 9 of 24
Environmental Science & Technology
219
(Figure S5). These findings are similar to what has been reported for lakes located 6-90 km away from the reference
220
site, in which unPAH concentrations and fluxes were found to increase for recently deposited sediments, with a
221
clearer trend observed in lakes located at less than 25 km away.19 The petrogenic signature observed was attributed
222
to the deposition of bitumen on dust particles associated with wind erosion from open pit mines.19, 48 Additionally,
223
an increase in unPAH concentrations and fluxes over the past 30 years have been observed as far as lakes in
224
northern Saskatchewan, east of the Athabasca oil sands.40 However, wildfires were identified as the principal source
225
of unPAHs to those lakes, based on retene concentrations and on compound specific stable isotope analysis of
226
recently deposited sediments.40 The analysis of the 259 heterocyclic aromatics presented here could be beneficial for
227
future source apportionment studies, given that they have less numerous natural sources than those for the
228
homocyclic counterparts.49
229
Passive Air Samples: The relative abundance for the heterocyclic aromatics was higher at site A1 (
