Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Article
Arctic and subarctic natural soils emit chloroform and brominated analogues by alkaline hydrolysis of trihaloacetyl compounds Christian Nyrop Albers, Ole Stig Jacobsen, Erico Marlon Moraes Flores, and Anders Risbjerg Johnsen Environ. Sci. Technol., Just Accepted Manuscript • Publication Date (Web): 09 May 2017 Downloaded from http://pubs.acs.org on May 15, 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 23
Environmental Science & Technology
1
Arctic and subarctic natural soils emit chloroform and brominated analogues
2
by alkaline hydrolysis of trihaloacetyl compounds
3 4
Christian N. Albers1,2*, Ole S. Jacobsen1, Erico M. M. Flores3, Anders R. Johnsen1
5 6 7
1
8
1350 Copenhagen K, Denmark
9
2
Geological Survey of Denmark and Greenland (GEUS), Department of Geochemistry, Øster Voldgade 10, DK-
Center for Permafrost (CENPERM), Department of Geosciences and Natural Resource Management, University
10
of Copenhagen, Denmark
11
3
Departamento de Química, Universidade Federal de Santa Maria, 97105-900, Santa Maria, RS, Brazil
12 13
* Email:
[email protected], phone +45 91333557, address: Øster Voldgade 10, DK-1350 Copenhagen K
14
1 ACS Paragon Plus Environment
Environmental Science & Technology
Page 2 of 23
15
Abstract
16
There has been increasing recognition of the occurrence of natural, halogenated organic compounds in marine
17
and terrestrial environments. Chloroform is an example of a halogenated organic compound with natural
18
formation as its primary source. Chloroform emission from soil has been reported from diverse Arctic,
19
temperate and (sub)tropical ecosystems. The terrestrial environment is a significant source to the atmosphere,
20
but little is known about the formation pathway of chloroform in soil. Here, we present evidence that
21
chloroform is formed through the hydrolysis of trichloroacetyl compounds in natural, organic-rich soils. In-situ
22
emissions of chloroform from soil in nine Arctic and subarctic ecosystems were linked to soil trichloroacetyl
23
turnover. The residence time from formation of the trichloroacetyl compounds in soil to the release of
24
chloroform to the atmosphere varied between 1 and 116 active months in unfrozen topsoil, depending on soil
25
pH. Non-specific halogenation that leads to trihaloacetyl formation does not discriminate between chloride and
26
bromide, and brominated analogues were formed alongside chloroform. Soil may therefore be a previously
27
unrecognised, natural source of brominated haloforms. The formation pathway of haloforms through
28
trihaloacetyl compounds can most likely be extended to other ecosystems with organic topsoils.
29 30
Introduction
31
In this article, we report on the formation mechanism of chloroform and its brominated analogues in natural
32
Arctic-subarctic soil. Chloroform has received considerable attention as a frequently detected pollutant in
33
groundwater1. Increasing attention has also been focused on the effects in the troposphere of short-lived
34
organohalogens, such as chloroform and its brominated analogues2-4. In the last 20 years, it has been
35
recognized that natural formation is the primary source of chloroform5 and that substantial amounts of
36
chloroform may be emitted from soils of temperate, coniferous plantations6-9 as well as natural forest
37
floors10,11. Indeed, based on atmospheric measurements in Mace Head, Ireland, the main sources of chloroform
38
in Europe are the less populated areas of Scandinavia, Ireland and Scotland12. This is in contrast to other
39
volatile organohalogens that have their primary European sources in industrialized areas12.
40
Although most previous studies have been conducted in forest plantations and other non-natural areas, it
41
seems that the formation of chloroform is a common process in many natural ecosystems. However, there is
42
very limited understanding about the mechanism of chloroform formation in soil.
43
2 ACS Paragon Plus Environment
Page 3 of 23
Environmental Science & Technology
44
Soil organic matter from chloroform-emitting forest plantations has recently been found to contain chemical
45
structures that can release chloroform at high pH13. The only known organochlorine structure that shows high
46
stability at low pH and hydrolyses to release chloroform at high pH is the trichloroacetyl group14,15.
47
Trihaloacetyl compounds are known intermediates in the nonspecific halogenation reactions that lead to
48
haloform formation during chemical water chlorination and paper pulp bleaching14. We therefore hypothesized
49
that the natural formation of chloroform and brominated analogs occurs through nonspecific halogenation of
50
soil organic matter leading to the formation of trihaloacetyl compounds in soil. These compounds will
51
eventually undergo nucleophilic attack by hydroxide ions (alkaline hydrolysis) and release haloforms (equation
52
1). The rate of hydrolysis will depend on the concentration of the hydroxide ion16, which is determined by the
53
soil pH.
SOM-R 54
HOX
O SOM-C-CX3
OH- hydrolysis
O CHX3 + SOM-C-O-
Trihaloacetyl compounds
(1)
55
where SOM-R is soil organic matter with a reactive group, such as a phenol, quinone, ketone etc. X is a halogen
56
atom, such as chlorine, bromine and possibly iodine.
57 58
At present, known halogenation mechanisms in the terrestrial environment include enzymatic reactions with
59
flavin-dependent halogenases, α-ketoglutarate dependent halogenases, methyl halide transferases or heme-
60
and non-heme haloperoxidases17,18 and non-enzymatic Fenton-like reactions19,20. Of these, only extra-cellular
61
chloro- or bromoperoxidases and Fenton-like reactions have the capacity to form free reactive chlorine or
62
bromine species such as HOCl and HOBr, and hence in theory lead to the formation of trihaloacetyl-groups in
63
soil organic matter. Using carbon and chlorine stable isotopes analysis, it was recently demonstrated in a lab-
64
experiment, that soil humic acid can be chlorinated by both hypochlorite and a commercial chloroperoxidase,
65
which led to pH dependent formation of chloroform, most likely through hydrolysis of trichloroacetyl
66
intermediates15,21. Indeed, most previous research on this subject has been conducted in the laboratory using
67
chemical chlorination with hypochlorite, which is extremely artificial compared to natural processes.
68
Trichloroacetyl compounds were detected in a limited study of three temperate, coniferous plantations13, but
69
data on the general occurrence of trichloroacetyl compounds in natural environments as well as the natural
70
occurrence of brominated analogs has not been reported previously. Empirical support of the trihaloacetyl
71
hypothesis based on analyses of environmental samples is therefore strongly needed.
3 ACS Paragon Plus Environment
Environmental Science & Technology
Page 4 of 23
72 73
The trihaloacetyl hypothesis has two testable predictions:
74
1. Trihaloacetyl compounds will be formed in haloform-producing soils, so that there will be a quasi-steady
75
state between trihaloacetyl formation and hydrolysis to haloforms. Trihaloacetyl compounds will hydrolyze
76
faster at higher pH, and these compounds will therefore show a shorter residence time in soils of higher pH.
77
2. The trihaloacetyl formation mechanism will be nonspecific and will not discriminate between chloride and
78
bromide. Brominated analogues of chloroform will therefore be formed if bromide is present in the soil.
79 80
We carried out a study to test these predictions for nine Arctic and subarctic ecosystems, in which chloroform
81
emissions have recently been reported11 (Fig. S1).
82 83
Experimental
84
Study sites
85
Four Arctic and five subarctic sites with different vegetation and soil properties were chosen for the study (Fig.
86
S1, Table 1). At each site, a 12-m permanent transect was established, with a distance of three metres between
87
the individual 0.06 m2 flux chamber bases. A detailed description of the transects and the study sites has
88
previously been published11.
89 90
Haloform analyses
91
Emissions of chloroform and brominated analogues were determined as described previously11. Briefly, 20-L
92
flux chambers were fitted onto the chamber bases for 65 minutes. Haloforms in the chamber headspace air
93
were trapped and stored in custom-made thermal desorption tubes (89×6.35 mm, containing Carbotrap B (22
94
mm), Carboxen 1003 (16 mm) and Carboxen 1000 (16 mm), Supelco, USA) for subsequent GC-ECD/MS analysis.
95
Before entering the thermal desorption tubes, the air was dried by passage through a Nafion™ desiccant
96
membrane dryer (DM-110-24, Perma Pure, NJ, USA). Each haloform emission was calculated based on four 1-L
97
samples. The chloroform emissions in Table 1 are the mean of two (Kan and Nar sites) or four (Abi and Dis
98
sites) sampling campaigns during the active season from late May to early September.
99 100
On 29 August 2014, 35 additional emission measurements were performed at site Kan-C, to better establish
101
the observed correlation between chloroform and bromodichloromethane emissions. Mobile chamber bases
4 ACS Paragon Plus Environment
Page 5 of 23
Environmental Science & Technology
102
instead of permanently installed bases were used for these samples, to allow the high number of
103
measurements to be performed. Concentrations of CFC11, which is neither consumed nor produced in the soil
104
and therefore should decrease in the chamber after each sampling, were used as a check that the system was
105
not leaking11. In addition, only analyses showing emissions of more than 15 ng m-2 h-1 chloroform or 0.2 ng m-2 h-
106
1
107
uncertain. Out of 35 samples, 18 fulfilled these criteria.
bromodichloromethane were used for further data analysis, as emissions lower than these values were highly
108 109
To sample soil air, a custom made probe comprising a 17-mm brass filter (Propartner, Denmark) connected to a
110
4-mm inner diameter polyamide tube (MB-LONGLIFE™ PA-12, Propartner, Denmark) was pushed to a depth of
111
20 cm. Soil air was then drawn using a membrane pump at 100 mL min-1 (210-1003MTX Twin Port Pocket
112
Pump, SKC, UK). The first 200 mL air was discarded to clean the probe and the following 400 mL was sampled in
113
thermal desorption tubes. A 0.45 µm PTFE filter (Midisart 2000, Sartorius, Germany) and a Nafion™ desiccant
114
membrane dryer were placed between the probe and the pump to remove dust particles and water vapour.
115
The system for collecting soil air was tested in the laboratory and found to neither emit nor adsorb measurable
116
levels of haloforms.
117 118
Soil samples
119
Soil was sampled from the transects by removing living mosses and lichens and then hammering a steel core (Ø
120
= 6 cm) to a depth of 10 cm. Two sets of soil samples were used in this study. The first set was sampled at the
121
nine sites during 2012 before the emission measurements. The 2012 samples were sampled 0.8 m from the
122
center of the chamber bases and used to determine total organic halogen. The second set, used for all other
123
measurements, was sampled in 2014 at the center of each flux chamber after the emission measurements
124
were terminated. The soil for analysis of total halogen and trichloroacetyl compounds was frozen on the day of
125
sampling and stored frozen at -18 °C. Before analysis, the soil was freeze dried and ground to fine powder. In
126
2014 an additional soil core was taken within each chamber base and stored at 5 °C until homogenization and
127
shaking with water (1:2.5, soil:water) to determine pH and perform the analysis of chloride and bromide by ion
128
chromatography (LC50-CD50, Dionex, CA, USA). Soil organic matter was determined as loss on ignition (550 °C,
129
2 hours).
130 131
Analysis of trichloroacetyl compounds
5 ACS Paragon Plus Environment
Environmental Science & Technology
Page 6 of 23
132
The concentration of trichloroacetyl compounds in soil from the nine sites was quantified after alkaline
133
hydrolysis followed by gas chromatographic detection of the formed chloroform as previously described13.
134
Briefly, 0.5 g freeze-dried and homogenized soil was put into a 120-mL serum bottle and then 2.5 mL 0.6M
135
NaOH was added and the bottle closed with crimp-caps containing an alumina-coated septum (Mikrolab
136
Aarhus, Denmark). After overnight shaking, a 9-mL headspace sample was taken with an airtight syringe and
137
injected into thermal desorption tubes for subsequent GC-ECD/MS analysis11. The chloroform in the headspace
138
(>80% of total chloroform) was corrected for sorption to organic matter and dissolution into water based on
139
partitioning experiments with different amounts of water and soil organic matter (supplementary methods).
140 141
The concentration of trichloroacetyl compounds was calculated as chloroform equivalents. The weight-based
142
trichloroacetyl concentrations were recalculated to area-based trichloroacetyl concentrations at each sampling
143
point based on the weight of the 10 cm soil core. The area-based trichloroacetyl concentration was then
144
divided by the mean chloroform emission at the same sampling point to obtain the mean residence time of
145
trichloroacetyl compounds in units of “months with non-frozen topsoil” (active months) for that specific
146
sampling point. Trichloroacetyl mean residence times at site Nar-B could not be reasonably calculated due to
147
chloroform emissions being too low and hence uncertain.
148 149
To test the pH-stability of trihaloacetyl compounds 0.2 g freeze dried and powdered Abi-A soil was weighed
150
into 120-mL serum bottles and pH was adjusted with 5 mL of a 1M citrate buffer (pH 4.7-5.8) or phosphate
151
buffer (pH 6.2-7.9) or just added 5 mL water (pH 4.3) and shaken gently for 24 hours at 23 °C.
152 153
Analysis of total organic halogen in soil
154
Total organic Cl and Br were determined using microwave-induced combustion and subsequent halogen-
155
specific detection, as previously described22,23. Briefly, inorganic halogen was removed from the powdered
156
samples by sequential washing (3 x 0.02M HNO3/0.2M KNO3 and 1 x water). The low pH during washing
157
ensured minimal loss of soil organic matter and ensured that trihaloacetyl compounds were not hydrolyzed.
158
The samples were then combusted in closed quartz vessels using 100 mg of sample in analytical triplicate. For
159
soil samples containing less than ~20% organic matter, a pyrohydrolysis method was applied according to the
160
conditions previously described24. Briefly, 100 mg of sample were mixed with 500 mg vanadium pentoxide in a
161
quartz boat and heated (10 min at 1000 °C; water flow-rate 1.0 mL min-1; humidified air flow-rate 200 mL min-
162
1
). In both methods, halogens were absorbed in 6 mL 50mM NH4OH that was diluted to 25 mL with water prior
6 ACS Paragon Plus Environment
Page 7 of 23
Environmental Science & Technology
163
to analysis by ion chromatography (chlorine) and inductively coupled plasma mass spectrometry (bromine)23.
164
Accuracy was evaluated using certified reference material of coal (NIST 1632c) and spiked samples were used
165
to evaluate the recovery of halogens.
166
Results and Discussion
167 168
Organic halogen in Arctic and subarctic soils
169
First, we determined the total organic and inorganic chlorine and bromine concentrations (Table 2) in soil from
170
nine transects where chloroform emission measurements were performed. Organic chlorine was present in all
171
chloroform emitting soils. These are the first such data reported for Arctic soils and interestingly, the total
172
organic chlorine concentrations of 63-341 mg kg-1 (175-1457 mg kg SOM-1) were very similar to what has been
173
reported for soils in warmer climates25,26. However, it is clear that chloroform emissions did not correlate with
174
total organic chlorine in soil (Table 1 and 2, Fig. S2), and that total organic chlorine in soil therefore cannot be
175
used as a proxy for chloroform emission. These results suggest that several chlorination mechanisms may occur
176
simultaneously in soil of which only some (those that produce free reactive chlorine species) contribute to
177
chloroform formation. An alternative explanation would be unspecific chlorination of different types of organic
178
matter leading to chlorinated compounds with different residence times at the different study sites. Organic
179
bromine was also present at all locations, ranging from 2-32 mg kg-1 (Table 2), corresponding to 6-90 mg kg
180
SOM-1. Organic bromine data are very scarce in the scientific literature, but our data from Arctic-subarctic
181
ecosystems are in the same range as those for temperate peatlands and coniferous plantations21,27. The organic
182
chorine/bromine ratio showed a substantial variation between study sites, ranging from only 4 at site Kan-C to
183
31 at site Abi-B (Table 2).
184 185
Trichloroacetyl compounds and their hydrolysis to chloroform
186
We then measured trichloroacetyl compounds in the soils of the transects and found that they were present in
187
all of the 44 soil samples analyzed, but at very different concentrations (Table 1, Fig. 1). Between sites, the
188
mean trichloroacetyl concentration varied 190-fold based on dry weight and 160-fold when expressed per soil
189
area (Table 1). There was also considerable within-site variation, often up to tenfold between soils in the five
190
haloform flux-chambers at each site. High trichloroacetyl concentrations were found only in soils with a pH
191
below 5.8 (Fig. 2a), which was in line with the first prediction that trichloroacetyl compounds become
192
increasingly unstable at higher soil pH. Furthermore, soils with higher pH (pH 6 to 7) showed up to 84 times
7 ACS Paragon Plus Environment
Environmental Science & Technology
193
higher chloroform emission:trichloroacetyl ratios compared to soils with lower pH (pH between 4 and 6). This
194
means that a low steady state concentration of trichloroacetyl compounds may be found even when
195
chloroform emissions are high. Our results are the first to demonstrate a general and natural occurrence of
196
trichloroacetyl compounds, and hence to support that chloroform is formed through trichloroacetyl
197
compounds in completely natural ecosystems. Furthermore, the correlation with pH has not previously been
198
demonstrated for environmental samples of any kind.
Page 8 of 23
199 200
It can be speculated that transport processes in the soil column might modify the emission fluxes to an extent
201
that they become independent of the formation processes, this is however unlikely for our soils.
202
Physicochemical processes such as sorption/-desorption and varying porosity would not change haloform
203
fluxes in a quasi-steady-state system. Degradation processes, on the other hand, may weaken the link between
204
haloform production and emission. Haloform degradation via halorespiration, which may take place under
205
highly reduced conditions, was probably negligible for most of our study sites that were oxidized down to the
206
permafrost or at least to a depth of 70 cm. One notable exception was the Kan-C site where the soil was
207
saturated with water, which may have led to locally reduced conditions, though the water was not stagnant
208
and methane emissions were undetectable or negative. Alternatively, some trihalomethane may be degraded
209
via microbial oxidation in the topsoil9. We did not determine the aerobic degradation potential of our soils, but
210
this process may account for some of the unexplained variation in our dataset.
211 212
By combining the trichloroacetyl concentrations and chloroform emissions, we could estimate a mean
213
residence time in soil for the trichloroacetyl compounds. These estimates were based on the assumptions that:
214
1) most of the trichloroacetyl compounds that release chloroform are located in the topsoil and 2) emission to
215
the atmosphere is the main loss route of chloroform from Arctic soils. These assumptions are a simplification,
216
but they are supported by three facts: 1) almost all organic matter in the Arctic soils and in temperate
217
coniferous plantations is located in the topsoil, 2) previous studies conducted in temperate coniferous
218
plantations show that chloroform is primarily formed in the topsoil7,9, and 3) 90% of the net chloroform
219
formation is released to the atmosphere in temperate coniferous plantations9. In Arctic ecosystems, the topsoil
220
is frozen most of the year, and chloroform emissions are considered negligible during this period11. We
221
therefore estimated trichloroacetyl residence times based only on the active periods when the soils were not
222
frozen. The estimated mean residence times varied from three active months at site Kan-C to 59 active months
223
at site Abi-B (Fig. 2b) with a full range from 1 to 116 months for individual flux chambers (Fig. S3). The
8 ACS Paragon Plus Environment
Page 9 of 23
Environmental Science & Technology
224
residence time decreased with increasing soil pH (Fig. 2b). This pH-dependence strongly supports our
225
hypothesis that the soil chloroform emissions originated from alkaline hydrolysis of trichloroacetyl compounds.
226 227
The mean residence time often showed high within-site variation (Fig. 2b), which was partly caused by within-
228
site variation in soil pH (Fig. S3). At site Nar-C, the organic layer extended 10 cm that were used for determining
229
the trichloroacetyl concentrations. This means, that we may have underestimated the area-based
230
trichloroacetyl concentrations for this site if trichloroacetyl compounds were present in significant amounts
231
below 10 cm. In addition, the mean residence times may have been slightly overestimated if some of the
232
chloroform was oxidized in the topsoil or leached.
233 234
We further investigated the influence of pH on the stability of the natural trichloroacetyl compounds in
235
laboratory tests with soil from the site with the highest chloroform emissions (Abi-A). The chloroform-release,
236
i.e. the trichloroacetyl stability, showed strong pH dependence with increased hydrolysis at higher pH (Fig. 2c).
237
Such a strong pH response has not previously been shown for any environmental sample containing natural
238
trichloroacetyl compounds, but was similar to what has been reported for the hydrolysis of trichloroacetyl
239
compounds formed by non-natural, chemical chlorination14,15. It should be noted that the laboratory
240
experiment (Fig 2c) shows the pH dependence of trichloroacetyl hydrolysis under laboratory conditions, which
241
cannot be used for prediction of in situ residence time (Fig. 2b). This is exemplified by the fact that the pH of
242
the Abi-A soil on average was 4.2 (Table 1), which corresponds to 1% hydrolysis of the trichloroacetyl
243
compounds in 24 hours at pH 4.3 (Fig 2c). This is faster than suggested by the in-situ mean residence time of
244
38 months (Fig 2b), but these two numbers are not directly comparable. First, the hydrolysis was determined at
245
23°C, which is much warmer than average in situ soil temperatures during the active months, which was 6°C11.
246
Second, soil colloids probably act as strong adsorbers of the trichloroacetyl molecules. Since the negatively
247
charged colloids attract hydrogen ions, the difference in acidity between colloid surfaces and the bulk solution
248
may be up to two pH units. Adsorbed molecules therefore ”experience” much lower pH than molecules in the
249
bulk solution28,29. As the pH dependence (Fig 2c) was determined under laboratory conditions, where the soil
250
was diluted 25-fold in 1M buffer, sorption equilibrium would presumably be shifted and more of the
251
trichloroacetyl compounds would be present in the bulk liquid due to dilution and high ionic strength.
252
Consequently, the trichloroacetyl compounds would be more labile than in-situ even at the same bulk pH.
253 254
Co-emission of brominated chloroform analogues
9 ACS Paragon Plus Environment
Environmental Science & Technology
Page 10 of 23
255
The formation of trichloroacetyl compounds requires nonspecific halogenation by free, reactive chlorine
256
species such as HOCl16, equation (1). The formation of HOCl by either chloroperoxidases or the fenton reaction
257
is, however, unspecific and other halides than chloride may be oxidized by both reaction mechanisms. This
258
means that HOBr (reactive bromine) would also be formed, if bromide ions were present18,19,30. Furthermore,
259
HOCl may react spontaneously with bromide and convert into HOBr due to the higher standard oxidation
260
potential of bromide compared to chloride31. If the proposed reaction mechanisms were correct (equation 1),
261
then brominated chloroform analogues would be formed along with chloroform when bromide is present. The
262
concentration of free bromide was below the detection limit (