Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Article
Gas-Phase Mechanisms of the Reactions of Reduced Organic Nitrogen Compounds with OH Radicals Nadine Borduas, Jonathan P.D. Abbatt, Jennifer Grace Murphy, Sui So, and Gabriel da Silva Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.6b03797 • Publication Date (Web): 03 Oct 2016 Downloaded from http://pubs.acs.org on October 7, 2016
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 37
Environmental Science & Technology
1
Gas-Phase Mechanisms of the Reactions of Reduced
2
Organic Nitrogen Compounds with OH Radicals
3
Nadine Borduas1*, Jonathan P. D. Abbatt1, Jennifer G. Murphy1, Sui So2, Gabriel da Silva2
4
1
Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, ON, M5S 3H6
5
2
Chemical and Biomolecular Engineering, University of Melbourne, Victoria 3010, Australia
6
ABSTRACT: Research on the fate of reduced organic nitrogen compounds in the atmosphere
7
has gained momentum since the identification of their crucial role in particle nucleation and the
8
scale up of carbon capture and storage technology which employs amine-based solvents.
9
Reduced organic nitrogen compounds have strikingly different lifetimes against OH radicals,
10
from hours for amines to days for amides to years for isocyanates, highlighting unique functional
11
group reactivity. In this work, we use ab initio methods to investigate the gas-phase mechanisms
12
governing the reactions of amines, amides, isocyanates and carbamates with OH radicals. We
13
determine that N−H abstraction is only a viable mechanistic pathway for amines and we identify
14
a reactive pathway in amides, the formyl C−H abstraction, not currently considered in structure-
15
activity relationship (SAR) models. We then use our acquired mechanistic knowledge and
16
tabulated literature experimental rate coefficients to calculate SAR factors for reduced organic
17
nitrogen compounds. These proposed SAR factors are an improvement over existing SAR
18
models because they predict the experimental rate coefficients of amines, amides, isocyanates,
19
isothiocyanates, carbamates and thiocarbamates with OH radicals within a factor of two, but
ACS Paragon Plus Environment
1
Environmental Science & Technology
20
more importantly because they are based on a sound fundamental mechanistic understanding of
21
their reactivity.
22
TOC GRAPHIC
Page 2 of 37
23 24
INTRODUCTION
25
Reduced organic nitrogen compounds, characterized by a C−N bond, are important for the study
26
of air quality and climate.1-3 Prominent examples of reduced organic nitrogen compounds in the
27
atmosphere include amines, amides, isocyanates, carbamates, cyanates and their sulfated
28
homologues, thiocarbamates and isothiocyanates. The C−N bond, where the N atom is in its -3
29
oxidation state, cannot typically be formed in the atmosphere. Thus, reduced organic nitrogen
30
compounds are emitted as such, and their fate in the atmosphere is govern by oxidation reactions,
31
particle formation and deposition.
32
These molecules have a wide range of biogenic and anthropogenic sources to the atmosphere and
33
have been identified and quantified in ambient air.4 Mixing ratios of reduced organic nitrogen
34
compounds vary from below pptv up to ppmv levels depending on the molecule, location, time
35
of day and meteorology. Amines are the most commonly measured organic nitrogen molecules
36
in the atmosphere. For example, gas phase trimethylamine concentrations were reported up to 40
37
pptv in rural areas and up to 6 pptv up in urban areas, suggesting an agricultural source.5-7 The
ACS Paragon Plus Environment
2
Page 3 of 37
Environmental Science & Technology
38
significant role of amines in aerosol formation and growth, as well as their recent large scale use
39
as solvents in carbon capture and storage (CCS) technologies, has validated focused attention on
40
understanding their fate in the environment.1,8-11 Indeed, at a CCS plant in Norway, peaks of up
41
to 10 ppbv of monoethanolamine (MEA), 300 ppbv of pyrazine and 800 ppbv of nitromethane
42
were reported at the top of a stripper column.12 It is known that the current benchmark solvent
43
for CCS, MEA, has a short atmospheric lifetime of approximately 2 h, governed by its reactivity
44
towards the OH radical.13-15 Other sources of reduced organic nitrogen molecules to the
45
atmosphere include direct emissions from industrial solvents, biomass burning, cigarettes, and
46
animal husbandry as well as oxidative chemistry of amines.1,4,8,16,17 Indeed, the gas-phase
47
oxidation of amines produces amides as well as isocyanates.13,15,18,19
48
The major atmospheric sinks for reduced organic nitrogen compounds are recognized as being
49
oxidation by OH radicals, and to a lesser extent by NO3 radicals and ozone. Because most
50
amines, amides and isocyanates do not absorb photons of wavelengths in the actinic window,
51
photolysis is not generally competitive.20,21 Loss to aerosol particles is another important sink for
52
some species. For example, amines may act as bases and help nucleate particles and/or contribute
53
to particle growth, which can impact climate directly by scattering light and indirectly by acting
54
as cloud condensation nuclei.20 Reduced organic nitrogen compounds are also thought to be
55
responsible for some of the colouring in brown carbon aerosols, again impacting climate through
56
their light-absorbing properties.22 Some reduced organic nitrogen compounds are also toxic. In
57
particular, methyl isocyanate and isocyanic acid may pose serious health effects if inhaled in
58
mixing ratios above 1 ppbv.23-25 In addition, nitrosamines and nitramines, oxidation products of
59
aminyl radicals reacting with NOx in urban areas for example, are carcinogens.21 Generally,
60
reduced organic nitrogen compounds are not important radiative forcing agents due to their small
ACS Paragon Plus Environment
3
Environmental Science & Technology
Page 4 of 37
61
gas-phase concentrations. However, the highest global warming potential of any compound
62
detected in the atmosphere is currently perfluorotributylamine.26
63
When organic nitrogen is quantified in both the gas and particle phases, the majority is typically
64
found in the particle phase and often in the water soluble fraction of particles.1,10 For example,
65
the concentration of aliphatic amines contributing to dissolved organic nitrogen in rainwater was
66
estimated to be < 1-14 nmol N m -3.3 Wet and dry deposition of organic nitrogen compounds are
67
estimated to be ~ 25% of the atmospheric global nitrogen deposition flux.2
68
A number of laboratory experiments, theoretical calculations and field studies have aimed to
69
better understand the fate of reduced organic nitrogen compounds and were recently reviewed in
70
the context of their use and/or production in CCS plants.4,8,20,21 Their chemical mechanisms have
71
been largely developed by experimental product studies, and more recently are being supported
72
by computational chemistry studies on amines,27-32 as well as recent studies on amide oxidation
73
mechanisms.17,33,34 In this study, we evaluate through computational chemistry, the mechanisms
74
involved in the gas-phase oxidation of reduced organic nitrogen compounds with OH radicals to
75
better understand their overall atmospheric fate. We then compare their mechanisms to highlight
76
the impact adjacent functionalities have on the nitrogen atom. Particular attention is given to
77
amines, amides and isocyanates, common functionalities found in the gas and particle phases.
78
Their lifetimes are also strikingly different, ranging from hours to years depending on the
79
functional group.8,21 We opted to study the simplest molecule of each class, i.e. methylamine,
80
formamide and isocyanic acid for relevance and calculation simplicity. We find that when
81
exposed to OH radicals, oxidation generally occurs on the adjacent C atom of N-containing
82
functionalities. In other words, amines are oxidized to amides and amides are oxidized to
83
isocyanates. We also evaluate a model carbamate molecule, N-methyl methylcarbamate since
ACS Paragon Plus Environment
4
Page 5 of 37
Environmental Science & Technology
84
structure-activity relationship (SAR) analyses for many reduced organic nitrogen compounds
85
originate from the evaluation of carbamate rate coefficients with OH radicals.35 Our mechanistic
86
approach is comprehensive as we consider all possible reaction sites on each functionality
87
including C−H and N−H abstractions, OH additions to carbonyl and OH additions to N atoms,
88
and subsequently identify probable reaction pathways. We then use this insight alongside a
89
compiled database of experimental rate coefficients to build SAR factors that rely on a
90
fundamental understanding of the reactivity of N-containing molecules.
91
COMPUTATIONAL METHODS
92
Computational ab initio methods were employed using the Gaussian 09 code.36 Structures of the
93
reactants were first optimized using the M06-2X density functional, with the 6-31G(2df,p) basis
94
set, and subsequently evaluated using the G3X-K composite theoretical method which combines
95
a series of Hartree-Fock, Møller-Plesset perturbation (MP4/6-31G(2df,p) and MP4/6-31+G(d))
96
and coupled cluster theory calculations (CCSD/6-31G(d)).37,38 A sample input file for the
97
execution of the G3X-K method in Gaussion 09 as well as optimized geometries are presented in
98
the Supplementary Information. G3X-K theory is used in this study because it was specifically
99
designed for thermochemical kinetics and reproduces barrier heights in the DBH24/08 database
100
with an average accuracy of 0.6 kcal mol-1.38 In addition, we did not use diffuse functions (+) in
101
the basis set throughout our calculations, simply to optimize computation time as the energies are
102
similar with and without their incorporation (see Table S1). The energies reported are for 0 K
103
and were calculated from the sum of the electronic and the zero point energies. The accuracy of
104
the energies stated is expected to be within 1 kcal mol-1.38 Bond dissociation energies (BDE) are
105
the difference between ground state energies of the products and the reactants of the bond
106
dissociated reaction.
ACS Paragon Plus Environment
5
Environmental Science & Technology
Page 6 of 37
107
Throughout this work, we describe transition state energy, used interchangeably with barrier
108
height energy, as the energy difference between the transition state and the sum of the reactants’
109
energies. Although many systems exhibit weak pre-complex formation between the reduced
110
organic nitrogen compound and the OH radical, these pre-complexes do not necessarily lie along
111
the reaction coordinate and/or may not be collisionally stabilized. Thus, we assume that their
112
energies do not dominate the overall flux through the transition state.39 While we approach the
113
mechanistic analysis by assuming that the reactivity will largely scale with the height of the
114
transition state, we recognize that entropic factors and tunneling may also play a role,
115
particularly for cases with transition states that are submerged relative to the reactant energy. In
116
addition, we estimate that a 1 kcal mol-1 difference in activation energy translates to
117
approximately a factor of 5 difference in rate coefficients using the Arrhenius equation for room
118
temperature rate coefficients that are on the order of 10-12 cm3 molec-1 s-1. Nonetheless, we
119
emphasize here that the theoretical calculations serve solely to identify reactive pathways and not
120
to calculate rate coefficients.
121
CALCULATION RESULTS
122
1. Bond dissociation energy (BDE)
123
From the energy diagrams of methylamine, formamide, N,N-dimethylformamide, N,N-
124
dimethylacetamide, N,N-dimethylpropanamide, N-methylpropanamide, isocyanic acid and N-
125
methyl methylcarbamate with OH radicals presented in this section (Schemes 1-4 and S1-S4), we
126
calculate bond dissociation energies (BDE) of C−H and N−H bonds in reduced organic nitrogen
127
compounds (Table 1). In general, our values compare well with previously reported
128
literature.40,41 We are aware of only one experimentally determined C−H bond BDE for the
ACS Paragon Plus Environment
6
Page 7 of 37
Environmental Science & Technology
129
reduced organic nitrogen compounds we investigated (methylamine). However, computed BDEs
130
exist and all compare well with our calculated BDEs for C−H bonds.40,42-44 For carbamates for
131
example, the methylamine C−H bond and the methoxy C−H bond were computed at 96 kcal mol-
132
1
133
the experimental BDE value for the N−H bond in formamide is 100.8 kcal mol-1, whereas we
134
calculate a value of 114.6 kcal mol-1.41 Yet, three previous ab initio studies have calculated
135
similar BDEs at 114.5 kcal mol-1, 113.6 kcal mol-1 and 113.2 kcal mol-1.42,43,45 Moreover, the
136
BDE of the N−H bond in amines (like methylamine) is expected to be lower than in amides (like
137
formamide) due to the delocalisation of the nitrogen’s lone pair in the amide functionality as
138
predicted by molecular orbital theory.
139
Table 1: Calculated and experimental bond dissociation energies for reduced organic nitrogen
140
compounds
and 100 kcal mol-1 respectively by Berry et al. and are comparable to our values.44 In addition,
Organic nitrogen
Formyl C−H bond (kcal mol-1)
alkyl C−H bond (kcal mol-1)
N−H bond (kcal mol-1)
Calculated
Calculated
Experimental
Calculated
Experimental
methylamine
NA
91.9
92.746,47
98.8
10042,48,49
formamide
93.3a
114.6
100.841
N,N-dimethylformamide
93.7
N,N-dimethylacetamide
NA
N,N-dimethylpropanamide
NA
N-methylpropanamide
NA
NA 92.5a
?
90.6a 97.9b 90.6a 93.6b 93.3a 95.1b
? ? ? ? ? ?
NA NA NA 105.5
~98.741
ACS Paragon Plus Environment
7
Environmental Science & Technology
Isocyanic acid
141 142 143 144 145
NA
NA
Page 8 of 37
109.8
109.750
NA 93.2a ? N-methyl 106.8 ~10551 c methylcarbamate NA 98.5 ? a b c C−H bond on the N-alkyl side; C−H bond on the amide side C−H bond on the methoxy side; ref 49 is for N-methyl ethylcarbamate51 NA = not applicable; ? = yet to be measured/reported
2. Methylamine + OH
146
Scheme 1 shows the three possible mechanisms for the OH radical to react with methylamine,
147
the simplest of amines, in the gas phase. Both the C−H and the N−H abstraction mechanisms are
148
competitive for this alkylamine, with barriers to reaction close to the entrance level energies of
149
the reactants and within 1 kcal mol-1 of each other. Both reactions are exothermic, as expected
150
due to the formation of water as a by-product. The third mechanism investigated is OH addition
151
to the N atom with a concerted C−N bond cleavage, which is highly endothermic with a clearly
152
inaccessible transition state energy.
153
Methylamine’s gas phase reaction was experimentally investigated by Atkinson et al., by Carl et
154
al. and by Onel et al.27,52,53 These authors find a room temperature rate coefficient of around 2 ×
155
10-11 cm3 molec-1 s-1 (lifetime of ~ 7 h); this fast rate coefficient is consistent with a low transition
156
state energy. Onel et al. also investigated this reaction by ab initio methods and found transition
157
states for both H-abstraction mechanisms slightly below the entrance level energies of the
158
reactants.27 Our results are consistent, albeit slightly higher in energy, a difference in part
159
attributable to the different model chemistries employed and our own theoretical method
160
uncertainty of 1 kcal mol-1.
ACS Paragon Plus Environment
8
Page 9 of 37
Environmental Science & Technology
161 162
Scheme 1. Theoretical energy diagram for methylamine + OH. Energies are 0 K enthalpies in
163
kcal mol−1, at the G3X-K level of theory.
164
3. Amides + OH
165
Amides are known to be products of amine oxidation but our understanding of their fate in the
166
atmosphere is limited.8,15,21,54 Their functionality differs from amines in that amides have a
167
carbonyl moiety adjacent to the nitrogen. The nitrogen’s lone pair is largely delocalized in the
168
carbonyl’s -system, as evidenced by higher BDEs for the N−H bond compared to amines
169
(Table 1). This effect translates to poor nucleophilicity and thus poor reactivity with electrophilic
170
OH radicals, and explains amides’ longer atmospheric lifetimes.17,54 Scheme 2 depicts the energy
171
diagram of formamide’s four possible mechanisms of reaction with OH radicals. The lowest
ACS Paragon Plus Environment
9
Environmental Science & Technology
Page 10 of 37
172
energy transition state is the formyl C−H abstraction, 6 kcal mol-1 lower than the next lowest
173
transition state energy. Thus, this pathway governs the majority of formamide’s reactivity,
174
analogous with the formyl C−H bond in aldehydes.55,56 This mechanism is also quite exothermic
175
compared to the N−H abstraction pathway. In contrast to amines, amides have the potential to
176
react with OH radicals by addition to the carbonyl moiety. We find that this pathway is 9.9 kcal
177
mol-1 above the entrance level energy of the reactants but is exothermic and so is a viable yet an
178
unlikely mechanism. As with amines, the OH addition to the nitrogen in amides is high in energy
179
and endothermic.
180
In previous studies, the room temperature rate coefficient of formamide and OH radicals was
181
measured to be 4.4 × 10-12 cm3 molec-1 s-1 and was supported by ab initio calculations, yet solely
182
focused on the H-abstraction mechanisms.17,34 The formyl C−H abstraction produces a C-
183
centered formyl radical which can go on to react with O2 and form isocyanic acid.17
184
Interestingly, the formyl C−H abstraction is also competitive for N-alkylated amides (see
185
Scheme S1). To complement the previous ab initio work on the reactivity of amides with OH
186
radicals, we also investigated the mechanistic pathways of N,N-dimethylacetamide, N,N-
187
dimethylpropanamide and N-methylpropanamide in Schemes S2, S3 and S4 respectively. Similar
188
mechanistic observations can be made when comparing all these amides: an easy C−H
189
abstraction from methylated amides and an unlikely N−H bond abstraction from amides (in
190
contract to amines).
ACS Paragon Plus Environment
10
Page 11 of 37
Environmental Science & Technology
191 192
Scheme 2. Theoretical energy diagram for formamide + OH. Energies are 0 K enthalpies in kcal
193
mol−1, at the G3X-K level of theory.
194
4. Isocyanic acid + OH
195
Isocyanates differ from the amide functionality by being hybridized sp2 at the nitrogen and sp at
196
the adjacent carbon. In isocyanic acid, the nitrogen’s lone pair is perpendicular to the -system
197
and is therefore not delocalised.57 Isocyanic acid’s OH-reaction energy diagram is presented in
198
Scheme 3. Three mechanisms are plausible: N−H abstraction and OH additions to the carbonyl
199
or to the nitrogen. All three mechanisms are high in energy and so isocyanic acid’s rate
ACS Paragon Plus Environment
11
Environmental Science & Technology
Page 12 of 37
200
coefficient is expected to be very slow at room temperature (Scheme 3). Indeed, the experimental
201
rate coefficient for the reaction of isocyanic acid with OH radicals was measured only at high
202
temperatures and if extrapolated to room temperature is approximately 10-15 cm3 molec-1 s-1,
203
translating to a lifetime of decades.24,58 The OH-addition to the carbonyl (barrier height of + 9.1
204
kcal mol-1) and the N−H abstraction (barrier height of + 19.4 kcal mol-1) pathways are
205
exothermic compared to the OH-addition to the nitrogen which we have now shown is
206
consistently endothermic in amines, amides and isocyanates. No pre-complex could be isolated
207
in our ab initio calculations between isocyanic acid and OH, consistent with all the mechanisms
208
having transition state energies significantly above that of the reactants. Thus, we do not expect
209
gas phase OH radicals to be a sink for isocyanic acid. Rather, it will likely partition to the
210
aqueous phase and undergo further reactions including hydrolysis.59
ACS Paragon Plus Environment
12
Page 13 of 37
Environmental Science & Technology
211 212
Scheme 3. Theoretical energy diagram for isocyanic acid + OH. Energies are 0 K enthalpies in
213
kcal mol−1, at the G3X-K level of theory.
214
5. N-Methyl methylcarbamate + OH
215
We opted to use N-methyl methylcarbamate as our carbamate substrate rather than the simpler
216
methylcarbamate as the latter does not yet have a reported rate coefficient with OH radicals in
217
the literature. Scheme 4 shows the theoretical energy diagram of the five mechanistic pathways
218
relevant to the reaction of N-methyl methylcarbamate with OH radicals. The difference between
219
carbamates and amides is the presence of an oxygen atom next to the amide functionality. This
220
extra oxygen atom provides additional electron density into the carbonyl, decreases the degree of
ACS Paragon Plus Environment
13
Environmental Science & Technology
Page 14 of 37
221
lone pair delocalization at the nitrogen and enhances the reactivity of the molecule towards the
222
electrophilic OH radicals. This claim is supported by the smaller bond dissociation energy (BDE)
223
for the N−H bond in N-methyl methylcarbamate (106.8 kcal mol−1) than in formamide (+ 114.6
224
kcal mol−1) for example (Table 1). Nonetheless, the C−H abstraction mechanisms are expected to
225
dominate the reactivity of carbamates since the two C−H abstraction transition state energies for
226
N-methyl methylcarbamate are below the energies of the reactants. Both OH addition
227
mechanisms, i.e. to the carbonyl or to the nitrogen, require high energy to proceed and are
228
endothermic.
229
We note here that N-methyl methylcarbamate’s experimental rate coefficient (4.3 × 10-12 cm3
230
molec-1 s-1) is slower than that of methylamine (2 × 10-11 cm3 molec-1 s-1) despite the former
231
having calculated lower transition state energies (see Table 2).27,35 Transition state energies are
232
relative to the entrance level energy of the reactants and are therefore not directly comparable
233
between molecules. On the other hand, BDEs are absolute values and we see then that the C−H
234
bond in methylamine has a lower BDE than N-methyl methylcarbamate (see Table 1), consistent
235
with a faster rate coefficient.
ACS Paragon Plus Environment
14
Page 15 of 37
Environmental Science & Technology
236 237
Scheme 4. Theoretical energy diagram for N-methyl methylcarbamate + OH. Energies are 0 K
238
enthalpies in kcal mol−1, at the G3X-K level of theory.
239
MECHANISTIC DISCUSSION
240
We gain insight into the mechanisms governing the reactivity of the four subclasses of reduced
241
organic nitrogen compounds investigated, namely amines (Scheme 1), amides (Scheme 2, S1,
242
S2, S3 and S4), isocyanates (Scheme 3) and carbamates (Scheme 4), by comparing their
243
calculated transition state energies (Table 2). We note here that we use the calculated transition
244
state energy solely to inform us on likely mechanisms and reactive sites.
ACS Paragon Plus Environment
15
Environmental Science & Technology
Page 16 of 37
245
Table 2: Tabulated transition state values from Schemes 1-4 and S1-S4 (with a 1 kcal mol-1
246
accuracy)
Transition state energies (kcal mol-1)
Organic nitrogen + OH
248
N−H abstraction
OH addition to carbonyl
OH addition to nitrogen
Formyl C−H
Nmethylated C−H
methylamine
NA
+ 1.2
+ 0.2
NA
+ 37.4
formamide
+ 0.2
NA
+ 6.3
+ 9.9
+ 41.7
- 0.9
- 1.2
NA
-
-
NA
- 2.4
NA
-
-
NA
- 2.4
NA
-
-
N-methylpropanamide
NA
- 2.2
+ 2.1
-
-
Isocyanic acid
NA
NA
+ 19.4
+ 9.1
+ 32.7
- 1.6 / - 0.1
+ 2.5
+ 14.1
+ 43.6
N,Ndimethylformamide N,Ndimethylacetamide N,Ndimethylpropanamide
247
C−H abstraction
N-methyl NA methylcarbamate NA = not applicable, - = not computed 1. C−H abstraction
249
The C−H abstraction mechanism dominates the reactivity of reduced organic nitrogen
250
compounds when present. This observation is expected as the C−H bond is less polarized than
251
heteroatom X−H bonds, making it more prone to break homolytically and react with the
252
electrophilic OH radical. C−H bonds in organic nitrogen compounds have consistently lower
253
bond dissociation energies than N−H bonds as seen in Table 1. The preference of OH radicals to
ACS Paragon Plus Environment
16
Page 17 of 37
Environmental Science & Technology
254
react with C−H bonds over N−H bonds leads to the oxidation of carbon while the nitrogen
255
remains in the -3 oxidation state throughout oxidative conversion from amines to amides to
256
isocyanates. Building upon this logic, we do not expect oxidation of reduced organic nitrogen
257
compounds to substantially contribute to the production of nitrogen oxides in ambient air. In
258
other words, the N atom in reduced organic nitrogen molecules does not end up as the N atom in
259
nitrogen oxide molecules.
260
2. N−H abstraction
261
The N−H abstraction mechanism which leads to aminyl radicals after the initial OH attack is
262
solely competitive for amines. The increasing trend observed for transition state energies of the
263
N−H abstraction mechanism is as follows: amine < carbamate < amide < isocyanate (see Table
264
2). Indeed, the N−H abstraction in methylamine is the only calculated barrier height which is
265
close to the entrance level energy of the reactants (at + 0.2 kcal mol-1), indicating a rapid and
266
favorable reaction. Importantly, aminyl radicals may go on to react with nitrogen oxides to form
267
nitramines and nitrosamines, which are known carcinogens.60-66 It is noteworthy that amides,
268
isocyanates and carbamates are not expected to form nitramines and nitrosamines directly from
269
OH radicals abstracting N−H bonds, based on the high energy barrier that yields their
270
corresponding aminyl radical. Nonetheless, as Bunkan et al. explicitly discuss, further reactions
271
of amide C-centered radicals may eventually lead to fragmentation and formation of aminyl
272
radicals from amides, yielding nitramines.33
273
3. OH addition to carbonyls
274
The addition of OH radicals to unsaturated bonds is a common mechanistic pathway for alkenes
275
and alkynes. For completeness, we also explored this mechanism for carbonyls, as they exist in
ACS Paragon Plus Environment
17
Environmental Science & Technology
Page 18 of 37
276
reduced organic nitrogen compounds. Although this mechanism is exothermic in formamide and
277
isocyanic acid, it is slightly endothermic for N-methyl methylcarbamate (Schemes 2, 3 and 4).
278
Nonetheless, it possesses a high energy barrier to reaction and is not expected to be competitive
279
with C−H or N−H abstraction pathways. In the case of isocyanic acid, this mechanism dominates
280
its reactivity, but this mechanism is still sufficiently slow at ambient temperatures to be
281
negligible in the context of isocyanic acid’s atmospheric fate.
282
4. OH addition to nitrogen
283
The OH addition to the N atom pathway was also investigated for methylamine, formamide,
284
isocyanic acid and N-methyl methylcarbamate’s reactions with OH radicals. This mechanism
285
was proposed to occur for sulfides and was originally thought to extend to amines.67 We show in
286
Schemes 1 to 4 that this mechanism is consistently high in energy and endothermic for all classes
287
of N-containing compounds studied. This observation remains true even for electron rich amines
288
like trimethylamine, where we calculate at the G3X-K level an activation energy of + 42.3
289
kcal/mol. Therefore, OH additions to N may lead to weak complexes,68 but will not lead to a
290
productive reaction pathway and would simply dissociate.
291
SAR MODELLING DISCUSSION
292
We now use our mechanistic insight to develop structure-activity relationship (SAR) group rate
293
constants and substituent factors to predict the atmospheric fate of reduced organic nitrogen
294
compounds. Our study is constructed on a fundamental mechanistic understanding of reactivity
295
with OH radicals to improve the existing empirical SAR model. We would like to emphasize at
296
this point that we do not calculate rate coefficients with our computed potential energy surfaces.
297
Rather, we use these energy diagrams to identify reactive and non-reactive pathways to lead us in
ACS Paragon Plus Environment
18
Page 19 of 37
Environmental Science & Technology
298
our development of the SAR factors described in this part of the study. We center our discussion
299
on aliphatic reduced organic nitrogen while appreciating that related aromatic compounds are
300
also of environmental importance.
301
1. Description of SAR models
302
Structure-activity relationships (SAR) models were developed as a predictive tool for estimating
303
the room temperature rate coefficients and hence the atmospheric lifetime of organic compounds
304
against the OH radical and later against other atmospheric oxidants.56,68-71 The SAR model’s
305
general approach consists of predicting an overall rate coefficient by summing the individual rate
306
constants of every reactive site on a molecule of interest. Relevant mechanisms for volatile
307
organic compound reactions with OH radicals include H-abstraction, addition to unsaturated
308
carbon-carbon bonds and addition to heteroatoms.68 The SAR method includes two types of
309
factors relevant to our reduced organic nitrogen analysis. First, there are the SAR group rate
310
constants for H-abstractions from any atom (C, N, O, etc.), denoted by a k value with a subscript
311
denoting the site (or functional group such as k-NH-, k-OH) of abstraction. There are three types of
312
SAR group rate constants specifically for C−H abstractions, one for each reaction occurring at a
313
primary (kprim), secondary (ksec) or tertiary (ktert) carbon center. Second, this C−H abstraction
314
SAR group rate constant (kprim, ksec, or ktert) is then multiplied by a substituent factor denoted as
315
F(X) to obtain an overall rate constant specific to the functional group’s reactivity. Sample
316
equations for C−H abstractions are given below and are based on Atkinson et al.68
317
𝑘𝑜𝑣𝑒𝑟𝑎𝑙𝑙 (𝐶𝐻3 − 𝑋) = 𝑘𝑝𝑟𝑖𝑚 𝐹(𝑋)
318
𝑘𝑜𝑣𝑒𝑟𝑎𝑙𝑙 (𝑌 − 𝐶𝐻2 − 𝑋) = 𝑘𝑠𝑒𝑐 𝐹(𝑋)𝐹(𝑌)
ACS Paragon Plus Environment
19
Environmental Science & Technology
319
Page 20 of 37
𝑘𝑜𝑣𝑒𝑟𝑎𝑙𝑙 ( 𝑌𝑍>𝐶𝐻 − 𝑋) = 𝑘𝑡𝑒𝑟𝑡 𝐹(𝑋)𝐹(𝑌)𝐹(𝑍)
320
There currently exist published empirical SAR group rate constants and substituent factors for
321
amines, carbamates or thiocarbamates functionalities, derived from limited experimental rate
322
coefficients.8,35,68,69,72 There are no SAR group rate constants specifically for amides,
323
isocyanates or isothiocyanates. Guided by our ab initio results, we examine the mechanisms
324
governing reduced organic nitrogen molecules’ reactivity in order to build upon past SAR
325
analyses. We tabulate existing experimental rate coefficients for reduced organic nitrogen
326
molecules with OH radicals and formulate SAR model equations defining koverall for each
327
compound (see Table S1 and model equations in the Supplementary Information). We use only
328
experimental rate coefficients in our statistical analysis and omit theoretically calculated rate
329
coefficients although the latter are becoming more reliable (but still few exist for N-containing
330
compounds).71,73 We then solve this overdetermined linear system of equations using a built-in
331
function in the program IGOR Pro (see Supplementary Information). Revised SAR factors are
332
proposed to better capture the differences in reactivity based on substituents on the N atom and
333
to disambiguate the existing factors for N-containing molecules (Figure 1 and Table 3). The SAR
334
factors are reported with two significant figures and without uncertainties as they represent
335
solutions to the overdetermined linear equations, remaining consistent with the SAR
336
literature.68,69,74
337
ACS Paragon Plus Environment
20
Page 21 of 37
Environmental Science & Technology
338 339
Figure 1: The SAR group rate constants represent reactivity at the highlighted H atom in
340
yellow and the SAR substituent factors represent reactivity imparted by the functionality
341
highlighted in blue to the H atom in bold. R1 and R2 can be either H or any alkyl group.
342
Table 3: The proposed SAR group rate constants kX-H and substituent factors F(X) for
343
aliphatic reduced organic nitrogen compounds Functional group Methyla
Group rate constant (cm3 molec-1 s-1) k-CH3 = 0.136 × 10-12 k-CH2- = 0.934 × 10-12 k-CH< = 1.94 × 10-12
Substituent factor (unitless) F(-CH3) = 1.00 F(-CH2) = 1.23 F(-CH