Absorption and Oxidation of Nitrogen Oxide in Ionic Liquids

Article abstract of DOI:10.1002/chem.201601166

A new strategy for capturing nitrogen oxide, NO, from the gas phase is presented. Dilute NO gas is removed from the gas phase by ionic liquids under ambient conditions. The nitrate anion of the ionic liquid catalyzes the oxidation of NO to nitric acid by atmospheric oxygen in the presence of water. The nitric acid is absorbed in the ionic liquid up to approximately one mole HNO₃per mole of the ionic liquid due to the formation of hydrogen bonds. The nitric acid can be desorbed by heating, thereby regenerating the ionic liquid with excellent reproducibility. Here, time-resolved in-situ spectroscopic investigations of the reaction and products are presented. The procedure reveals a new vision for removing the pollutant NO by absorption into a non-volatile liquid and converting it into a useful bulk chemical, that is, HNO₃.

Full text of DOI:10.1002/chem.201601166

DOI: 10.1002/chem.201601166
Full Paper
&
Ionic Liquids
Absorption and Oxidation of Nitrogen Oxide in Ionic Liquids
Andreas J. Kunov-Kruse, Peter L. Thomassen, Anders Riisager, Susanne Mossin,* and
[a]
Rasmus Fehrmann*
Abstract: A new strategy for capturing nitrogen oxide, NO,
from the gas phase is presented. Dilute NO gas is removed
from the gas phase by ionic liquids under ambient condi-
tions. The nitrate anion of the ionic liquid catalyzes the oxi-
dation of NO to nitric acid by atmospheric oxygen in the
presence of water. The nitric acid is absorbed in the ionic
nitric acid can be desorbed by heating, thereby regenerating
the ionic liquid with excellent reproducibility. Here, time-re-
solved in-situ spectroscopic investigations of the reaction
and products are presented. The procedure reveals a new
vision for removing the pollutant NO by absorption into
a non-volatile liquid and converting it into a useful bulk
chemical, that is, HNO3.
liquid up to approximately one mole HNO per mole of the
3
ionic liquid due to the formation of hydrogen bonds. The
Introduction
depends on the association between the CO molecule and
2
[8]
the anion. The bonds formed are often relatively weak and
therefore, the gas can be easily expelled from the IL in a regen-
eration step. Water is inevitably present in a flue gas and be-
The man-made emission of NO (N O, NO, NO ) is of environ-
x
2
2
mental concern because NO participates in detrimental photo-
x
[12]
chemical reactions in both the troposphere and the strato-
cause water also interacts with the anion of the IL, water is
a likely inhibitor in both cases. Oxygen from the flue gas may
[
1]
sphere. NO reacts with hydrocarbon-containing pollutants to
x
[13]
form health-threatening smog in densely populated areas, as
well as being active in depleting the ozone layer. The end
also interfere with the absorption. Here, a different strategy
to remove NO and NO from a gas stream is proposed, where
2
product, that is, NO , contributes to acid rain. To remove the
the water and oxygen present are actively utilized to chemical-
ly convert the NO into nitric acid, HNO , which then accumu-
2
NO from flue gases, ammonia or hydrocarbons are generally
x
x
3
used as reductants in the excellently performing selective cata-
lytic reduction (SCR) reaction. This technology is, however, not
lates in the IL due to the formation of strong hydrogen
[14]
bonds.
attractive for specialized applications, for example, for NO re-
The current work presents in-situ attenuated total reflec-
tance Fourier transform infrared (ATR-FTIR) spectroscopic re-
sults of the interaction between a thin IL film and a combined
gas stream containing dilute NO gas, atmospheric air, and
water. From time-resolved experiments mechanistic evidence
for the oxidation and absorption reactions as well as the stoi-
chiometry of the absorption have been extracted. The accumu-
x
moval on ships or in biomass fired and co-fired power plants,
where an end-of-pipe cleaning technology is preferred.
The unique properties of ionic liquids (ILs) include high ther-
mal stability, practically no vapor pressure, and a large electro-
[
2–5]
chemical window.
Recent research has focused on the use
of ILs for selective absorption of acidic gases such as CO and
2
[
6–9]
SO2.
In addition a recent comprehensive review on gas solu-
lated HNO can be desorbed by heating the IL above 1208C.
3
bilities in ILs concerns not only CO and SO but also reports
The IL is completely regenerated with no trace of decomposi-
tion products observed.
2
2
10]
[
on all other gases known to solubilize in ILs. It is remarkable
that no data on NO solubilities in ILs are reported so far. SO2
capture by ILs is suggested to depend on the association of
Results and Discussion
[
7]
SO with the anion of the IL. This has later been confirmed
2
by X-ray investigations of crystalline [TMGH]Cl·SO (TMGH=
A thin film of an IL was monitored with ATR-FTIR spectroscopy
during exposure to a controlled gas flow containing NO. A
schematic of the setup is shown in Figure S1 in the Supporting
Information. The influence of the cations and anions of the IL
on the reactivity was also investigated. The employed cations
are depicted in Scheme 1.
2
[
11]
1
,1,3,3-tetramethylguanidinium). CO capture by ILs similarly
2
[a] Dr. A. J. Kunov-Kruse, MSc. P. L. Thomassen, Assoc. Prof. A. Riisager,
Assoc. Prof. S. Mossin, Prof. R. Fehrmann
Centre for Catalysis and Sustainable Chemistry, Department of Chemistry,
Technical University of Denmark, DK-2800 Kgs. Lyngby (Denmark)
E-mail: slmo@kemi.dtu.dk
The characteristic IR bands observed and monitored during
ꢀ
ꢀ1
the reactions were: nitrate (NO ): n˜ =1340 cm (vs), nitric
rf@kemi.dtu.dk
3
ꢀ
1
acid (HNO ): n˜ =948 (s), 1660 cm (s); water (H O): n˜ =3400 (s),
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201601166.
3
2
ꢀ
1
1640 cm (s); dinitrogen tetroxide (N O ): n˜ =1858 (w), 1750
2
4
Chem. Eur. J. 2016, 22, 1 – 12
1
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Scheme 1. IL cations used in this work. BMIM=1-butyl-3-methylimidazolium,
EMIM=1-ethyl-3-methylimidazolium, BMMIM=1-butyl-2,3-dimethylimidazo-
lium, cholinium (Ch)=N-(2-hydroxyethyl)-N,N,N-trimethylammonium.
ꢀ
1
ꢀ
ꢀ1
(
vs), 750 cm (w); acetate (CH COO ): n˜ =1550 cm (vs); and
3
ꢀ
1
[15,16]
acetic acid (CH COOH): n˜ =1710 cm (vs).
The bands cor-
3
Figure 1. ATR-FTIR spectra of [BMIM][NO
ing NO (2% v/v), O
corresponds to the spectrum of the pure IL exposed to air saturated with
water and the black line to the steady-state spectrum obtained after 15 min
gas exposure. Intermediate spectra are indicated with gray lines. The most
intense identifiable bands are indicated.
3
] exposed to a gas stream contain-
(16% v/v), and water (ꢁ1% v/v) at RT. The blue line
responding to the IL cations do not change significantly
[
17]
2
during the experiments.
According to calculation, the nitric acid band at n˜ =
ꢀ
1
9
48 cm corresponds to the symmetric O-N-O bend. The ex-
periments are in agreement with this because the band does
not shift when the water in the gas stream is exchanged with
ꢀ
1
D O. The other characteristic peak for HNO at n˜ =1660 cm
become available (see Scheme 2). Additionally, the nitrate
band splits as less symmetric hydrogen-bonded species are
generated, giving bands at n˜ =1426, 1330, 1295, and
2
3
ꢀ
1
shifts to n˜ =1610 cm
with deuterium substitution
(Scheme 2).
ꢀ
1
1
250 cm , see Figure 1. Similarly, within twenty seconds of
gas exposure the characteristic bands of the solvated HNO
3
ꢀ
1
molecule at n˜ =940 and 1660 cm increase in intensity. An iso-
ꢀ
1
sbestic point is clearly observed at n˜ =1400 cm . Another iso-
ꢀ
1
sbestic point at n˜ =1300 cm is less well defined due to weak
ꢀ
1
intensity bands from HONO at n˜ =1260 and 1290 cm . The
presence of isosbestic points confirms that essentially only two
forms of the nitrate ion exist in the solution, that is, free as in
pure [BMIM][NO ] and in a strong hydrogen bond with HNO .
3
3
A reference experiment was performed in order to assess
whether the reaction proceeds without O . No reaction was
2
observed when [BMIM][NO ] was first left to equilibrate with
3
the atmosphere and then exposed to NO (10% v/v) for 80 min
3ꢀ
3
3ꢀ
Scheme 2. Top) Assigned IR modes of NO , HNO
, and their complexes. As-
(
Figure S2 in the Supporting Information). Therefore, the dis-
ꢀ
signment based on calculated frequencies of NO , NO
The analogous modes of DNO were assigned by observation of the red
shift of lines in the experimental spectra. Calculations were performed by
···HNO , and HNO .
3 3 3
proportionation of NO and the presence of a NO impurity in
2
3
the gas stream are both dismissed as the reason for the ob-
[
18]
using the B3LYP functional and the 6-31G(d,p) basis set in Gaussian 09.
served activity. A control experiment with dry NO and O on
2
a dry IL film was also performed and we could determine that
the reaction was indeed slower when water was the limiting
factor. The only source for the protons in the observed HNO3
was water present either in the IL or in the gas phase and no
other type of reactivity was observed in experiments with NO
gas. If NO, O , and H O are all present, the characteristic HNO
3
In-situ ATR-FTIR experiments
[BMIM][NO3]
A thin film of [BMIM][NO ] was applied to the diamond surface
3
2
2
of the IR cell and exposed to a gas stream containing different
bands appear within minutes of exposure of the IL film to the
concentrations of NO (1–6% v/v) in combination with O (8–
gas mixture. Thus it was concluded that HNO is only formed
2
3
1
8% v/v), water (ꢁ1% v/v), and balance N at different tem-
when all three components are present simultaneously.
2
peratures. When applied, the IL forms a thin film on the dia-
mond, thereby minimizing mass transfer limitations. Thus, the
experiment is envisioned as a model for the interaction be-
tween the gas phase and a supported ionic liquid phase (SILP)
[
BMIM][CH COO]
3
The experiment performed using [BMIM][CH COO] is shown in
3
[
19]
ꢀ1
material.
Figure 2. Initially, a prominent water band at n˜ =3400 cm
The IR spectra obtained for the 2% v/v NO experiments at
room temperature are shown in Figure 1. The characteristic
(not shown) was present because [BMIM][CH COO] is hygro-
3
scopic and has a significant content of water when in equilibri-
um with the atmosphere. It was eroded quickly as the reaction
progressed. The prominent acetate bands at n˜ =1570 and
ꢀ
1
very intense band at approximately n˜ =1340 cm of the ni-
trate ion immediately decreases in intensity as acidic protons
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Figure 2. ATR-FTIR spectra of [BMIM][CH
3
COO] exposed to a gas stream con-
(8% v/v), and water (ꢁ1% v/v) at RT. The blue line
Figure 3. ATR-FTIR spectra of [BMIM]Cl exposed to a gas stream containing
taining NO (6% v/v), O
2
NO (6% v/v), O
2
(8% v/v), and water (ꢁ1% v/v) at RT. The blue line corre-
corresponds to the spectrum of the pure IL exposed to air saturated with
water, whereas the gray and black lines correspond to the spectra recorded
after 7 and 12 min gas exposure, respectively. The green line corresponds to
the spectrum recorded after heating to 1208C and subsequent cooling to
RT.
sponds to the spectrum of the pure IL exposed to air saturated with water,
the gray lines correspond to the spectra recorded after 2, 3, and 4 min of
gas exposure. The black line corresponds to the spectrum at steady state
after 6 min of gas exposure.
ꢀ
1
1
380 cm disappeared within a few minutes, whereas a band
ꢀ
1
from acetic acid concurrently appeared at n˜ =1710 cm . After
longer time of exposure, the HNO bands appeared and bands
3
corresponding to both acids co-existed. After gas exposure for
more than five minutes (or after gentle heating to 808C) the
acetic acid was lost, as indicated by the disappearance of the
bands assignable to both acetate and acetic acid. Hence, only
the bands from nitric acid and hydrogen-bonded nitrate re-
mained. When heated to 1208C the nitric acid bands were fi-
nally also completely lost, and the resulting spectrum features
ꢀ
1
the characteristic nitrate band at n˜ =1340 cm corresponding
to the spectrum of pure [BMIM][NO ]. The minor differences
3
were assigned to the different water content of the ILs. De-
sorption is gradual and most of the nitric acid is desorbed
before 1208C. Interestingly, the temperature of complete de-
sorption corresponds to the boiling temperature of the azeo-
Figure 4. ATR-FTIR spectra of [EMIM][CF
taining NO (6% v/v), O
corresponds to the spectrum of the pure IL, the black line corresponds to
the spectrum of the saturated IL at steady state (17 min), and the gray lines
correspond to intermediate spectra. The green line corresponds to the spec-
trum after heating to 1208C and subsequent cooling to RT.
3 3
SO ] exposed to a gas stream con-
(8% v/v), and water (ꢁ1% v/v) at RT. The blue line
2
[
20]
trope of nitric acid and water (b.p. 1218C), whereas pure
nitric acid boils at 838C. On the other hand, water and acetic
acid (b.p. 1188C) do not form an azeotrope and the complete
loss of acetic acid was accordingly observed at significantly
lower temperatures.
Notably, the strong bands from nitrate were not observed
after desorption, whereas the strong bands originating from
the triflate ion were observed throughout the experiment. The
latter bands only changed moderately after accumulation of
HNO in the IL, suggesting that hydrogen bonding of HNO to
[
BMIM]Cl
A similar experiment for [BMIM]Cl is shown in Figure 3. Here,
the bands from the cation are more readily identified. The final
spectrum obtained after exposure to NO, O , and H O was
2
2
3
3
identical to the spectrum recorded in the [BMIM][NO ] experi-
the anion was present but less pronounced than observed in
the experiments with the nitrate IL. Hence, the strong triflate
3
ments. The absence of isosbestic points, however, indicates
ꢀ
1
the formation of an intermediate. The strong band at n˜ =
band centered at n˜ =1270 cm (asymmetric S=O stretch) lost
ꢀ
1
1
360 cm identifies the intermediate as a non-hydrogen-
intensity and became broadened when HNO was present but
3
bonding nitrate anion (see below).
regained intensity after desorption.
[
EMIM][CF SO ]
[BMMIM][NO3]
3
3
[
EMIM][CF SO ] was investigated in the same setup (Figure 4).
The C2-position of the IL imidazolium ring can be deprotonat-
ed when the anion of the IL is a strong base, and the formed
3
3
The oxidation reaction forming HNO3 was observed, but at
a slower rate compared to the nitrate- and chloride-based ILs.
[21]
carbene has been shown to be reactive towards CO2.
In
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
3
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
order to discard an analogous reaction path in the current
study, IR experiments were also performed with [BMMIM][NO3]
having a methyl group at the C2-position of the imidazolium
ring. The obtained spectral results (Figure S3 in the Supporting
producible. Importantly, all conclusions are equally valid from
ꢀ
1
quantification of the n˜ =1660 cm band. The IR spectrum in
the vicinity of the characteristic HNO vibrations at n˜ =
3
ꢀ
1
948 cm were deconvoluted by Gaussian functions, by using
[23]
Information) were identical to the results for [BMIM][NO ], vali-
the method described previously. The area under the Gaussi-
3
ꢀ
1
dating that the carbene reaction path is not relevant. Hence,
an function fitted to the n˜ =948 cm band in the experimen-
+
all experimental results suggested that the [BMIM] ion does
tal spectrum can be correlated to the HNO content by the use
3
not partake in the reaction. The non-involvement of the cation
was further confirmed in experiments performed with acetate-
and nitrate-based ILs containing the cheap, biodegradable,
and non-toxic cholinium cation (Scheme 1). The results are
shown in Figures S4 and S5 in the Supporting Information and
are analogous to the ones for imidazolium ILs with the same
anion.
of the standard curve. The uncertainty of the total HNO con-
3
tent determined by this method is estimated to be 10–20%
based on the quality of the reference spectra and the fitting
procedure.
The fitting procedure was applied to all time-resolved spec-
tra at RT of [BMIM][NO ], [EMIM][CF SO ], [BMIM]Cl, [Ch]
3
3
3
[CH COO] as well as [BMMIM][NO ] when exposed to NO (6%
3
3
v/v) with O and water in excess. The results are shown in
2
Figure 6. For the cholinium-based ILs, IR bands from the cation
[
BMIM][NO ] and NO
3 2
By using dry NO2 gas instead of NO the experiment with
BMIM][NO ] was repeated (Figure 5). NO2 was willingly ab-
[
3
sorbed in the IL and dimerized in the condensed phase, giving
characteristic bands from N O at n˜ =1858 (w), 1750 (vs), and
2
4
ꢀ
1
[22]
7
50 cm (w). Bands from HONO and HNO were also promi-
3
nent after absorption, indicating disproportionation of N O in
2
4
the presence of water in the IL. When the gas feed was
changed to dinitrogen saturated with water, the bands specific
for N O decreased immediately and the characteristic HNO
3
2
4
bands increased.
ꢀ
1
Figure 6. Integrated area of the n˜ =948 cm HNO
3
band as a function of
(8% v/v), and
the exposure time to a gas stream containing NO (6% v/v), O
2
water (ꢁ1% v/v) at RT. The area is converted to mole ratios by using the
standard curve obtained from Figure S6 in the Supporting Information. The
3
experiment with [BMIM][NO ] was performed with the gas above the IL
changing immediately to the target composition. All the other experiments
were performed with a more gradual change of the gas flow.
are overlapping with the quantifiable HNO bands. By careful
3
use of the peak deconvolution method, reliable data were ob-
tained. The steady-state saturation point approximately corre-
sponds to a molar ratio of 1.1:1 (HNO /IL) for the acetate, chlo-
3
ride, and nitrate ILs, whereas the triflate IL converges at a ratio
of approximately 0.6. The apparent difference in the absorp-
tion rate observed between [BMIM][NO ] and [BMMIM][NO ]
3
Figure 5. ATR-FTIR spectra of [BMIM][NO ] exposed to a gas stream contain-
ing NO
posed to air saturated with water, the gray line corresponds to the spectrum
recorded after NO exposure for 3 min, and the black line belongs to the
2
at RT. The blue line corresponds to the spectrum of the pure IL ex-
3
3
was due to a difference in the absorption setup and is not con-
sidered to be significant. The time-resolved spectra of [BMIM]
2
spectrum recorded after subsequent 30 s of exposure to N
water.
2
saturated with
[
NO ] were measured with an improved setup allowing the IL
3
to be exposed to the NO/O /H O mixture exactly at t=0,
2
2
whereas the switch for the other ILs was dependent on re-
placement of a non-negligible volume of gas before the ATR
cell, resulting in a concentration gradient during the first mi-
nutes of the experiment. Accordingly, the time line has been
shifted 20 s to partially correct for this. All experiments per-
formed under identical conditions support the conclusion that
the cation does not influence the rate of absorption or the ob-
Quantification of HNO3
A standard curve for the determination of the HNO concentra-
3
tion in the IL was generated by measuring IR spectra of solu-
tions prepared by dissolving known amounts of concentrated
nitric acid in [BMIM][NO ] at RT (Figure S6 in the Supporting In-
3
ꢀ
1
formation). Either of the HNO bands at n˜ =1660 and 948 cm
3
can be used to quantify the amount of HNO present in the IL,
tained molar ratio of HNO to IL.
3
3
but because a water band overlaps with the former band, the
results obtained by using the latter were more reliable and re-
The IL anion, on the other hand, is decisive for the evolution
of the spectra. The acetate IL shows conversion of NO to HNO3
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
4
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
as fast as the nitrate IL, whereas the triflate IL shows a slower
conversion. The slow rate of accumulation of HNO in [BMIM]
3
[
CF SO ] is assigned to triflate being less nucleophilic towards
3 3
activation of NO. Evaluation of the IR spectra of the IL suggests
that HNO3 forms hydrogen bonds to other HNO molecules,
3
weak hydrogen bonds to the oxygen atoms in the triflate ion,
and no hydrogen bonds to the fluorine atoms, which are very
poor hydrogen-bond acceptors. Triflate is the ion of a stronger
acid than HNO and is not protonated during the reaction.
3
Notably, the chloride IL has a significantly different overall
profile with an initial induction period followed by first a slow
and then a fast HNO3 formation rate, before converging to
almost the same molar HNO /IL ratio as the nitrate ILs. The dif-
3
ference is assigned to chloride being a stronger base than ni-
[
24]
trate in ILs, making strong bonds to protons. The water con-
tent of [BMIM]Cl in equilibrium with the atmosphere was de-
termined by Karl-Fischer titration to be 5% w/w, which corre-
ꢀ
sponds to a molar ratio of about 1:2 for H O to Cl . This is
2
3
Figure 7. Top) Double logarithmic plot of maximal linear rates of HNO for-
much higher than for any of the other ILs used (ꢁ1% w/w
mation (r) in [BMIM][NO
partial pressure of NO in the gas phase. Bottom) Relative rate of HNO3 for-
mation in [BMIM][NO ] with 2% v/v NO with excess water and oxygen at dif-
3
] at RT with excess water and oxygen as function of
H O for [BMIM][NO ]).
2
3
After the first minutes of gas exposure some water was
3
ferent temperatures. A dotted line is added as a guide to the eye. Rates are
determined as the maximal linear rate of the change in intensity of the
likely driven out of the liquid and the reaction also consumed
ꢀ
1
water resulting in the n˜ =3400 cm water band decreasing
ꢀ1
n˜ =948 cm band in the in-situ ATR-FTIR spectra.
fast. All protons generated in the reaction became bound to
chloride leaving no protons to generate HNO . Accordingly,
3
a relatively narrow band corresponding to non-hydrogen-
bonded nitrate appears in Figure 3 in the time interval corre-
sponding to 50–150 s in Figure 6 and the apparent rate of the
HNO generation is lower. After 150 s the rate of HNO forma-
crease with temperature. Due to the two competing effects it
was not possible to generate an Arrhenius plot for the total re-
action based on the experiments described here. The overall
maximum rate was observed around 308C. We assume that
the rate of the NO oxidation reaction is limiting the overall
rate below this temperature and that the solubility of the reac-
tants in the IL determines the overall rate above this tempera-
ture.
3
3
tion in [BMIM]Cl increases sharply to the same value as for
BMIM][CH COO], [BMIM][NO ], [Ch][NO ], and [BMMIM][NO ]
[
3
3
3
3
and the band of the non-hydrogen-bonded nitrate disappears.
The rate of accumulation of nitric acid in the IL is very de-
pendent on the NO concentration as well as the temperature.
[
BMIM][NO ] was saturated with HNO at room temperature
Bulk experiments
3
3
after 50 s when exposed to 6% v/v NO, excess O , and water.
2
Absorption followed by gravimetric analysis
Absorption of NO in [BMIM][CH COO] and [BMIM][NO ] was
With 1% v/v NO it took 3000 s. In the in-situ ATR-IR experi-
ments the maximal linear rate of increase in intensity of the
3
3
HNO band was determined. The experiment was performed
performed by passing the gas slowly through bulk ILs (ꢁ1 g)
in two types of experiments. The gas uptake was followed
gravimetrically. Series 1 was performed with 10% v/v NO in N2
3
multiple times with different NO concentrations at RT and the
measured rates are plotted in Figure 7, top. From these experi-
ments an apparent order of the reaction with respect to NO
was determined to be 2.0 (ꢂ0.3).
bubbling through [BMIM][CH COO] at 1008C for about ten
3
days to obtain steady state. The resulting solution was allowed
to cool to RT under continuous gas exposure. t=0 was chosen
as the point when the temperature of the solution was at RT.
The solution was left with time to equilibrate. After several
days a new steady state had been reached (Figure 8). Oxygen
and water was not deliberately added in this experiment, be-
cause it was originally intended as a pure NO absorption ex-
periment. Series 2 is a reproduction of series 1 with a different
Once the IL is saturated with HNO , the reaction cannot be
3
followed further with IR spectroscopy performed on the IL
film, but the reaction between NO and HNO is likely to contin-
3
ue although at a significantly lower rate. The IL provides
a sticky surface where NO from the gas phase can react with
HNO giving NO and HNO , which are liberated to the gas
3
2
2
stream if the IL is locally saturated. However, the liberated spe-
cies might be absorbed further downstream.
batch of the IL. In series 3, pre-dried [BMIM][NO ] was bubbled
3
The absorption of NO by the ILs is very temperature-depen-
dent giving rise to a decrease in the absorption capacities with
increasing temperature (Figure 7, bottom). The solubility of
through with a gas stream of 5% v/v NO containing water and
oxygen in excess (1 and 10% v/v, respectively). The weight
was followed closely until steady state. Despite the difference
in the conditions for series 1 and 2 compared to series 3 an
identical weight increase compared to time t=0 was found at
steady state.
[25]
gases in ILs generally decreases with increasing temperature.
We expect the same to be the case for NO, NO , and O . On
2
2
the other hand, the rate of NO oxidation is expected to in-
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
were compared to the NMR spectra of pure [BMIM][NO ] (Fig-
3
ures S8 and S9 in the Supporting Information) and were found
to be almost identical. The spectra revealed no indication of
proton or carbon signals attributed to acetic acid [these would
1
13
be found at d( H)=2.1 (3H), 11.4 ppm (1H); d( C)=20.8 (CH ),
3
1
78.1 ppm (COOH)] in the sample after desorption. On the
1
other hand, some well-defined signals at d( H) (CDCl )=13.8
3
1
and d( H) ([D ]DMSO)=5–6 ppm (br) were observed indicating
6
that desorption of HNO was not complete. No decomposition
3
products were observed in the NMR spectra after this absorp-
tion–desorption cycle, and the chemical shifts of the imidazoli-
um ion signals remained essentially unchanged compared to
Figure 8. Increase in the mass of the bulk ILs as a function of the time by
passing a gas through the IL through a sintered glass frit at RT. The gain of
mass is assumed to be due to HNO
a mole fraction of HNO relative to the IL. In series 1 and 2 [BMIM][CH
was exposed to a gas flow of 10% v/v NO in N at 1008C for several days
before cooling to RT (at t=0) under a continued gas flow. In series 3, pre-
dried [BMIM][NO ] was exposed to a gas mixture of 5% v/v NO with excess
and water at RT.
3
only and has been recalculated to
the pure [BMIM][NO ], except for a slight upfield shift (0.05–
3
3
3
COO]
0.20 ppm) of the aromatic protons caused by the acidic envi-
2
ronment.
3
O
2
Comparison between the IL and aqueous nitrate solution
It is relevant to compare the reaction rates of the uptake and
the oxidation of NO in IL media to the one in an analogous
aqueous nitrate solution in order to prove that the IL is in fact
enhancing the reaction rate and not just providing a non-vola-
tile reaction medium. Unfortunately, it is not possible to design
a comparable experiment by using neither the in-situ FTIR
setup nor the gravimetric analysis of absorption in a bulk
sample. The water in the aqueous nitrate solution will evapo-
rate fast leaving just the nitrate salt on the IR diamond and, in
the bulk experiment, evaporation of water will change the
weight of the solution. Instead the reactions were compared
by quantifying the developed acidic protons by volumetric
The observations are explained in the following way: In the
initial stages of series 1 and 2 some water is present in the
BMIM][CH COO] and the precautions to keep out oxygen and
[
3
water over an extended period were not sufficient. Thus, NO is
being oxidized to HNO , which is a stronger acid than acetic
3
acid and [BMIM][CH COO]+HNO3 will give [BMIM]
3
[
NO ]+CH COOH. The generated acetic acid is not retained in
3 3
the liquid at 1008C and is lost to the gas phase with time. At
steady state the conversion is complete. This is in correspond-
ence with the observation of the viscosity of the IL changing
drastically because [BMIM][NO ] has a much lower viscosity
3
than [BMIM][CH COO] due to a weaker hydrogen-bonding pat-
analysis as follows. In series 4, a solution of NaNO in water
3
3
[
20,21]
tern between the cations and anions.
After the tempera-
having the same molar concentration of nitrate as [BMIM][NO3]
(ꢁ6m) was prepared. NO (5% v/v) and air saturated with
water was bubbled through 3.0 mL of the solution with a rate
ture is lowered to RT the starting point in series 1, 2, and 3 are
now analogous. Series 3 reaches steady state faster because
the access to oxygen and water is not limited by diffusion
through tubing or the presence as impurity in the gases or
setup. The end point after waiting to steady state is the same
for series 1, 2 and 3. Thus, we conclude that the reactivity ob-
served for series 1 and 2 is the same as for series 3 despite the
initial difference in the IL and the reaction conditions.
ꢀ
1
of 10 mLmin in a series of experiments. After 2, 4, 6, 8, and
10 h, respectively, the gas flow was stopped and the solution
was diluted with degassed water. Thymol blue was added as
acid/base indicator and titration was performed with a 0.10m
aqueous solution of NaOH. In series 5 the same experiments
were performed on [BMIM][NO ]. Both series 4 and 5 showed
3
an increasing amount of acidic protons being accumulated in
the liquids, as a function of the exposure time. The higher vis-
cosity of the IL should result in a hindered mass transfer across
Thermogravimetric analysis
Thermogravimetric analysis (TGA) performed on the pure
the gas/liquid interface compared to an aqueous NaNO solu-
3
[
BMIM][CH COO] compared to the product at steady state after
tion. Despite this effect favoring the conversion in the aqueous
solution it was observed that the IL had accumulated approxi-
mately three times as many acidic protons as the aqueous ni-
trate solution during the first couple of hours (see Figure S10
in the Supporting Information). After longer experimental
times, the IL approaches saturation and reacts slower, whereas
the aqueous solution continues to have a linear increase.
Water evaporating from the aqueous solution during the ex-
periment was partially avoided by using a low flow rate and
the air in the flow being saturated with water. Nevertheless,
some water loss could not be avoided judged by the weight
change of the solution. The aqueous solution produces some
3
reaction with NO + O + H O confirms the transformation of
2
2
the IL (see Figure S7 in the Supporting Information). [BMIM]
NO ] decomposes at temperatures well above 2408C, which is
[
3
6
08C higher than the decomposition temperature of [BMIM]
[CH COO].
3
NMR spectroscopy
1
13
H and C NMR analyses (in CDCl and [D ]DMSO) were per-
3
6
formed on the bulk [BMIM][CH COO] sample after absorption
3
to steady state with NO at room temperature and subsequent
desorption for 3 h at 100–1108C (without gas flow). The data
precipitate of NaNO with time because the solubility is ex-
3
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
ceeded locally. This is not a problem for the IL where no pre-
cipitates are observed. The use of a frit section at the end of
the gas tube inserted in the aqueous solution provided irrepro-
show that the ILs have excellent chemical stability towards
H N O species. This is rationalized with the imidazolium ring
x
y
z
+
being well protected from electrophilic attack by NO2 or simi-
lar species due to its positive charge. As seen from the pro-
posed reaction pathway below, radicals originating from dioxy-
gen are not believed to be part of the turnover of the reaction.
The fact that the organic cations are unaffected by the harsh
treatment is also a strong indication that the formation of un-
controllable reactive radical species is controlled in the IL
ducible results as NaNO precipitates in the frit. Therefore, an
3
open-end (Ø=1mm) glass tube was used for both experi-
ments.
Comparison between inert gas, NO, CO , and SO
2
2
media by the more stable radicals NO and NO . Uncharged or-
2
In a realistic flue gas CO and SO are also present and can in-
2
2
ganic molecules or other impurities present in the IL are likely
less protected and therefore oxidized and decomposed com-
teract with the ionic liquid. In order to separately assess the in-
teractions of these gases with the ionic liquid the following
control experiments were performed: five different gas mix-
tures were bubbled through approximately 1 g of initially dry
pletely to CO and water giving a self-cleaning process of the
2
IL during heating to the regeneration temperature.
ꢀ
1
NO is always present to some extent in pressurized NO gas
2
[
BMIM][NO ] with a rate of 40 mLmin (see Table 1 and Fig-
3
bottles due to the slow, but irreversible disproportion reaction
ure S11 in the Supporting Information).
[26]
of NO to form N O and NO . However, because no HNO for-
2
2
3
mation was observed in the absence of dioxygen in the feed
gas (Figure S2 in the Supporting Information), the dispropor-
tion reaction cannot account for the observed generation of
3
Table 1. Absorption of gases in [BMIM][NO ] determined gravimetric and
by titration with NaOH(aq).
HNO . Similarly, the level of autoxidation of NO by dioxygen in
3
Experiment
1
2
3
4
5
the gas phase before contact with the IL was monitored with
gas-phase UV/Vis spectroscopy and was found to be very low
and thus not responsible for the observed reactivity. Further-
CO
SO
NO [%]
2
[%]
[%]
–
–
–
10
1.4
24
50
–
–
10
1.4
24
0.11
0
–
1
–
10
1.4
24
–
–
5
10
1.4
8
–
–
0.1
10
1.4
36
2
[27]
O
2
[%]
more, preliminary
and ongoing investigations of the gas
H
2
O [%]
conversion reaction with supported ionic liquid phase (SILP)
absorbers reveal much higher conversion dynamics over the IL
phase compared to bypass experiments. So far, several absorp-
tion (at RT)–desorption (at 1308C) cycles performed by SILP ab-
sorbers revealed a completely reversible and unchanged ca-
pacity of the absorber.
after absorption for x h
weight increase [wtwt IL]
mol H / mol IL
ꢀ
1
0.10
0
0.09
0.004
0.53
1.10
0.11
0.10
+
In the first three experiments the weight increase is only
due to water because acid–base titration reveals that the
amount of acidic protons in the solutions after absorption is
close to zero. For experiment 4 with 5% NO, the IL is almost
saturated after 6 h and the increase in both weight and acidity
of the solution is very high. For experiment 5 with 1000 ppm
NO, the results show that the reaction does take place but the
saturation is slow due to the low concentration of NO. The
weight increase is not significantly higher than for experi-
ments 1–3, but acidic protons corresponding to 10% of the
molar amount of the IL were found after 36 h. This corre-
sponds to approximately 15% of the NO in the total gas
volume that was bubbled through 1 g of the IL for 36 h.
Proposed reaction cycle
The overall reaction (1) is thermodynamically favored with
ꢀ
1
D G=ꢀ210 kJmol , but it is very slow under ambient condi-
r
tions.
4
NOðgÞþ3 O ðgÞþ2 H OðlÞ ! 4 HNO ðlÞ
ð1Þ
2
2
3
ILs have ionic strengths, which are of course much higher
than those of molecular liquids and higher than those of brine
solutions as well. For ILs consisting of large organic cations
and small inorganic anions, the base strength and the nucleo-
philicity of the anion is much higher than if the anion was sol-
Discussion
[24]
The cations of the ILs do not decompose even when heated to
vated in water. Normally, NO is kinetically inert toward many
common reactants but ILs provide a condensed medium con-
taining strong nucleophiles. The absence of a full hydration
shell decreases the activation barrier for nucleophilic attack.
Based on the experimental evidence given here, a mechanism
for the oxidation reaction of NO in IL is suggested.
1
208C in concentrated nitric acid; the original IR spectra of the
nitrate and triflate ILs are completely recovered after cooling
to RT even after many cycles. If bulk [BMIM][NO ] saturated
3
with HNO is kept completely dry and heated to above 1208C
3
for longer periods, some yellow coloration is observed, but no
irreversible decomposition products have been observed.
When exposing [BMIM]Cl to NO, species corresponding to
both hydrochloric acid and nitric acid are present in the solu-
tion, but the IL does not decompose despite being saturated
with this very potent oxidation mixture. These observations
The oxidation of NO is reported in the literature to have the
same rate law as given in Equation (2) in the gas phase and in
aqueous solution despite the total reactions being differ-
[28,29]
ent.
In this work, a second-order dependence on p(NO)
over the IL film was also found (see Figure 7), and we expect
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
the same rate law is relevant for the rate-determining step in
this IL solution.
N O ðILÞþH OðILÞ ! HNO ðILÞþHNO ðILÞ
ð10Þ
ð11Þ
ð12Þ
2
4
2
3
2
ꢀ
þ
HNO ðILÞ ! NO2 ðILÞþH ðILÞ
2
d½NOꢃ
2
ꢀ
¼ k½NOꢃ ½O ꢃ
ð2Þ
2
HNO ðILÞþNO2 ðILÞ ! HN O4ꢀðILÞ
ꢀ
dt
2
2
Gas phase:
ꢀ
ꢀ
The protons generated will prefer NO2 over NO3 because
ꢀ
nitrite is the stronger base. The dimeric species HN O (IL) and
2
4
2
4
NOðgÞþO ðgÞ ! 2 NO ðgÞ
ð3Þ
ð4Þ
2
2
the nitrous acid dimer have a strong hydrogen bond keeping
the two nitrogen(III) species close together. The nitrogen(III)
species have thus been activated for oxidation by dioxygen in
one step. The product is the analogous nitric acid nitrate
dimer:
Aqueous solution:
NOðgÞþO ðgÞþ2 H OðlÞ ! 4 H ðaqÞþ4 NO ꢀðaqÞ
þ
2
2
2
In ILs an adduct can form between the IL anion and the NO
HN O4 ðILÞþO ðILÞ ! HN O ꢀðILÞ
ꢀ
ð13Þ
molecule capturing and activating the NO for further reactions,
because the anion is a powerful base and nucleophile. In the
case of a nitrate IL, the attack on NO can involve oxidation of
the NO nitrogen atom, which is thermodynamically favored.
The product has one negative charge and is likely to be
caught in the IL phase and to participate in further reactions.
The initiation step in the condensed IL phase is thus proposed
to be:
2
2
2
6
This corresponds well to the observed molar ratio of approx-
imately 1:1 between HNO3 and the IL after gas saturation.
When the IL is saturated with HNO there are just enough pro-
3
tons to allow every proton to form a strong hydrogen bond
between two nitrate anions. No free basic nitrate anions are
then available for nucleophilic attack on NO, only allowing
HNO to accumulate very slowly in the IL after this point (see
3
NOðgÞþNO3ꢀðILÞ ! N O4ꢀðILÞ
ð5Þ
Figure 7). The nitrate–nitric acid dimer is clearly observable in
the IR spectrum and is the final product until the nitric acid is
desorbed by raising the temperature:
2
ꢀ
The unstable adduct N O4 is then suggested to decompose
2
into two species with the intermediate oxidation states of +III
and +IV as given in reaction (6). Reactions (5) and (6) are anal-
HN
O
ꢀðILÞ ! HNO
ðgÞþNO
ꢀðILÞ
ð14Þ
2
6
3
3
ogous to the reaction between NO and HNO in the atmos-
3
[30–33]
phere or on surfaces.
An analogue reaction to reaction (13) has previously been
suggested to explain why the rate of the oxidation of NO by
x
N O ðILÞ ! NO ðILÞþNO2ꢀðILÞ
ꢀ
ð6Þ
2
4
2
dioxygen was observed to be several orders of magnitude
[35,36]
faster just around 08C in water droplets.
The freeze-con-
It was shown (see Figure 5) that NO dimerizes to N O in
2
2
4
centration effect leads the HNO2 concentration to increase
drastically in the remaining liquid phase by a factor of 240–
the IL when insufficient water is present (reaction (7)).
[37]
1
1000. The formed HNO dimers are activated towards oxi-
2
2
NO ðILÞ ! N O ðILÞ
ð7Þ
2
2
4
dation by O2:
ðHONOÞ ðaqÞþO ðaqÞ ! 2 HNO ðaqÞ
We expect N O to be able to disproportionate to ions in an
2
4
ð15Þ
2
2
3
ionic media as given in reaction (8). This corresponds to the
anhydrous form of the known disproportionation reaction oc-
[
34]
In the IL phase a single proton assures that the nitrogen(III)
species are held together, activating them towards oxidation
curring on surfaces.
N O ðILÞ ! NO þðILÞþNO ꢀðILÞ
ð8Þ
2
by O in an analogous manner.
2
4
2
2
Reaction (13) is suggested to be the rate-determining step
after a series of pre-equilibria, that is, reactions (5)–(7) and re-
actions (10)–(12). This is in accordance with the overall reaction
order of two with respect to NO. The complete reaction path-
way consisting of reactions (5)–(7) and reactions (10)–(14) is
shown in the catalytic cycle in Scheme 3, having the stoichiom-
etry of the total reaction given in Equation (1). The individual
reactions with coefficients are listed in Table S1 in the Support-
ing Information. The appearance of the cycle is complicated by
the demands of stoichiometry and the suggestion of elemen-
tary reactions but can be considered in a simple way as nitrate
Two nitrite anions from reactions (6) or (8) could hypotheti-
cally be oxidized by O in one step, as given in reaction (9):
2
ꢀ
2
NO2 ðILÞþO ðgÞ ! 2 NO3ꢀðILÞ
ð9Þ
2
However, reaction (9) is probably not very favored even in
an IL phase because two anionic species have to be close to-
gether in order to allow reaction with one O molecule. Under
2
dry conditions, the final products after complete oxidation are
+
ꢀ
the ionic species NO2 and NO3 , as in the solid state of N O .
2
5
In the majority of the experiments enough water was pres-
ent allowing reactions (10)–(12) to take place and several of
the intermediates are observed in Figure 5.
oxidizing NO to HNO and thereby being reduced itself to ni-
3
trite. O then re-oxidizes nitrite to nitrate, thereby regenerating
2
the catalyst.
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
8
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Conclusion
Nitrate is a catalyst for the oxidation of NO by O to nitric acid.
2
Nitrate is reduced to nitrite by NO and is re-oxidized by O2,
thereby restoring the catalyst. Aqueous nitrate solutions have
modest activity towards this reaction. By using a nitrate-based
ionic liquid, the catalyst is an integral part of the reaction
medium, the reaction is at least three times faster than in
aqueous solution with a comparable nitrate concentration,
and, most importantly, the reaction can be run in a continuous
flow without the reaction media being lost by evaporation.
Under exactly the same conditions, the IL does not absorb
either CO or SO to any measurable extent.
Scheme 3. Stoichiometric reaction cycle for the oxidation of NO to HNO
catalyzed at RT by nitrate based ILs (reactants: red, products: blue, inter-
mediates: black).
3
2
2
Interconversion between the different nitrogen species in
the condensed IL phase is likely dynamic and fast and the con-
centrations are the equilibrium values. The most reactive spe-
cies towards oxidation by dioxygen is expected to be a nitrite–
nitrous acid adduct or a nitrous acid dimer. Both are poised to
deliver four electrons to a dioxygen molecule and make the
corresponding dimeric nitrogen(V) species without oxygen rad-
ical intermediates. The hydrogen bonds formed serve to retain
the nitrogen species in the IL phase removing them from the
gas phase. This method is therefore a tool to absorb exhaust
It should be noted that the chemistry is similar to what
would be expected in aqueous solution. The difference ob-
served is due to an enhanced activity of the nitrate anions in
IL solution because they are not solvated by water molecules.
Also the interaction between typically large cations of ILs with
the anions is significantly weaker compared to the interaction
+
by the typically smaller inorganic cations (e.g., Na ) with the
NO species from off gases and storing them. The absorption
x
[
38]
III
anions. Furthermore, the generation of dimeric units of N
ability can be restored in a desorption step at higher tempera-
tures. No deterioration of the absorption properties of the IL is
observed even after many absorption–desorption cycles. After
species coupled by hydrogen bonding has a favorable impact
on the rate of oxidation with O , which is the rate-determining
2
step.
saturation, all nitrate ions participate in strong hydrogen
ꢀ
In ILs with other anions than nitrate, a different initiation re-
action for the catalytic cycle must take place: Acetate is not an
oxidant but is still a good nucleophile that can capture NO in
the condensed phase. The reaction mechanism for this initia-
tion is not clear, but it is possible that a small percentage of
bonds and the molar ratio between NO3 and HNO is approxi-
3
mately 1:1. Thus, the resulting mixture after saturation with
HNO3 can be formulated as a new IL, that is, [BMIM]
[NO ·HNO ]. The product after desorption is a concentrated
3
3
HNO stream that can be condensed with water as nitric acid
3
NO present in the gas can disproportionate and start the pro-
of commercial grade.
2
cess or that O could participate in radical reactions for the ini-
The cation of the IL has no apparent influence on the ab-
sorption reaction and in most cases the anion is converted to
nitrate after one absorption–desorption cycle. Thus, ILs like the
non-toxic cholinium chloride used as a vitamin additive for
livestock, are expected to be just as effective as traditional imi-
dazolium-based ILs and can be obtained at a much lower cost.
In order to overcome the mass-transfer limitations for the in-
teraction of bulk ILs with the gas phase, solid absorbers with
ILs impregnated on porous silica or other porous supports can
2
tiation reaction. After a while, nitrate accumulates and the re-
action proceeds like in nitrate-based ILs. The ILs with anions,
which are stronger bases than nitrate, will be converted to the
corresponding nitrate IL after one cycle. Acetate is a very pow-
[
24]
erful Lewis and Brønsted base in a pure IL and FTIR measure-
ments performed on [BMIM][CH COO] (Figure 2) and [Ch]
3
[
CH COO] (Figure S4 in the Supporting Information) show that
3
ꢀ
1
IR features assignable to acetate at n˜ =1550 cm decrease
very quickly as the nitrate and acetic acid bands increases. As
NO is absorbed and oxidized, nitric acid is formed but the pro-
tons are removed immediately by acetate. Judged from the
[9,27]
be made. The resulting SILP materials
are very promising
for the application as reversible NO absorbers in end-of-pipe
x
installations of industrial units emitting NO like, for example,
x
ꢀ
1
decreasing intensity of the band at n˜ =1715 cm with time,
power plants, ships, and cement factories. Investigations with
such SILP absorbers are ongoing.
the acetic acid is slowly removed from the liquid even at RT. In
the bulk experiments with [BMIM][CH COO], acetic acid was
3
absent according to NMR analysis after one absorption–de-
sorption cycle. When starting from a more basic anion than ni-
trate, it is suggested that the generated, unprotonated nitrate Experimental Section
anion is involved in the further absorption and oxidation of
[
BMIM][NO ] (>98%, IoLiTec), [BMIM][CH COO] (>98%, Sigma–Al-
3
3
NO. Furthermore, as acids accumulate in the IL, the nitric acid
will be retained more strongly than the other acids due to its
formation of a high-boiling azeotrope with water having a pat-
tern of strong hydrogen bonds.
drich), [BMIM]Cl (95%, Sigma–Aldrich), [BMMIM][NO ] (>98%, IoLi-
Tec), [Ch][CH COO] (95%, Sigma–Aldrich), and [EMIM][CF SO ]
(>99%, IoLiTec) were obtained from commercial sources and used
as received. [Ch][NO ] was produced in situ by exposing [Ch]
3
3
3
3
3
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
9
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[
CH COO] to a gas mixture of NO/O /water in nitrogen followed by
10 h, respectively, by adding a total of 50 mL degassed water in
several portions and transferring the resulting solution to an Erlen-
meyer flask. Thymol blue was added as acid/base indicator and the
amount of acidic protons was determined by titration with NaOH
3
2
heating to 1208C in a flow of nitrogen.
The gas streams were mixed from 10% v/v NO in N , N , CO , 2%
v/v SO in N and air. All gases were delivered by AGA. The gas
flow was controlled by mass flow controllers (Bronkhorst). To
2
2
2
2
2
(0.100m). For series 4 and 5, the experiments were reproduced
with an open-end glass tube. This had little influence for the ex-
periments with IL but problems with clogging of the frit with
NaNO3 in the aqueous solution bulk experiments were avoided
giving more reproducible results.
obtain a wet gas, the air or N stream was led through water or
2
D O (99.9%, Sigma–Aldrich) in a bubble flask fitted with a glass frit
2
and joined with the NO gas just before the reactor. NO and O2
were led through stainless steel tubing (Swagelok) close to the re-
actor chamber in order to minimize the autoxidation of NO in the
NMR spectra were recorded at RT on a Bruker 300 MHz NMR instru-
ment by using CDCl solutions/inhomogeneous mixtures of pure
gas phase, which is facilitated by sunlight. NO was generated by
3
2
[
BMIM][NO ] and [BMIM][CH COO] as well as of the ILs after satura-
treating copper metal with HNO3 (65%, Sigma–Aldrich). Prior to
subjecting the IL sample to the gas, the gas was dried through
a column with activated molecular sieves (4 ꢁ). ATR-FTIR experi-
ments were performed on a Nicolet iS5 spectrometer equipped
with a thermo-regulated (up to 3008C) Specac golden gate high-
temperature diamond ATR cell. This was fitted with an add-on cap,
allowing the controlled gas mixture to flow over the diamond of
the instrument at atmospheric pressure (Figure S1 in the Support-
ing Information). The experiments were performed with 1–6.5% v/
3
3
tion with NO gas in the bulk at RT as described above for series 3
and subsequent desorption without flow at 1108C for 3 h.
1
[BMIM][NO ]: H NMR (300 MHz, CDCl ): d=0.86 (3H; butyl), 1.28
(2H; butyl), 1.78 (2H; butyl), 3.93 (3H; methyl), 4.16 (2H; butyl),
7.42 (1H; C5), 7.48 (1H; C4), 9.62 ppm (1H; C2); C NMR (300 MHz,
CDCl ): d=13.5, 19.5, 32.2, 36.4, 49.9, 122.5, 124.0, 137.7 ppm (im-
3
3
13
3
purities: 54.7, 104.6 ppm).
1
[
BMIM][CH COO] after reaction: H NMR (300 MHz, CDCl ): d=0.92
3
3
v NO, 8–20% v/v O , 1–1.5% v/v water with a flow rate of
2
(3H; butyl), 1.33 (2H; butyl), 1.84 (2H; butyl), 3.96 (3H; methyl),
ꢀ
1
5
0 mLmin . The IL was applied as a thin film in the center of the
4
1
3
.20 (2H; butyl), 7.37 (1H; C5), 7.41 (1H; C4), 9.42 (1H; C2),
diamond before securing the add-on cap. For the IL screening ex-
periments, the gas exchange was performed with a valve outside
the add-on cap and a reproducible delay before the reaction was
13
3.52 ppm (0.5H; HNO3); C NMR (300 MHz, CDCl ): d=13.5, 19.6,
3
2.2, 36.5, 50.1, 122.5, 123.9, 137.5 ppm (impurity: 54.7 ppm).
TGA was performed on a Mettler Toledo TGA/DSC 1 STARe system
in a flow of nitrogen of 70 mLmin . Samples were heated from RT
to 6008C with a heating rate of 108Cmin .
present. The time-resolved experiments with [BMIM][NO ] were
ꢀ1
3
performed with the atmosphere in the cap being equilibrated
before quickly adding the cap over the IL film on the ATR diamond.
The reaction and absorption were followed in real time by record-
ing spectra continuously. For each spectrum, 4–16 scans with
ꢀ1
Determination of water contents in the ILs were performed by
Karl-Fischer titration on a Metrohm 888 Titrando by analyzing
three samples of 0.1 mL.
ꢀ1
4
cm resolution (approximately 5–20 s measuring time) were per-
formed. The spectra were background- and ATR-corrected in
OMNIC 8.2 assuming a refractive index of 1.5 and a reflectance
[39]
angle of 908. Gaussian deconvolution by using PeakFit was per- Acknowledgements
ꢀ
1
formed individually in the area of the n˜ =948 and 1660 cm
bands of HNO . To allow for batch deconvolution, a method was
3
This work is financially supported by the Energinet.dk through
the PSO project 10521 and the Copenhagen Cleantech Cluster
[34]
developed for Gnuplot 4.1. Further details on the deconvolution
[23]
method can be found in the literature. The shown full spectra in
Figures 1–6 were all obtained at RT. Spectra (not shown) obtained
at temperatures up to 608C were the basis for the results shown in
Figure 7, bottom.
(
GAP project no. 91050-ko). S.M thanks the Danish Independ-
ent Research Council DFF (project no. 09-070250). The FP7
COST Action CM1306 EXIL is acknowledged for financial sup-
port for dissemination. Dr. Johannes Due Hansen, Thorey Gret-
tarsdottir, and Anita Godiksen, Centre for Catalysis and Sustain-
able Chemistry, DTU Chemistry, Technical University of Den-
mark are acknowledged for performing preliminary investiga-
tions. LAB S.A, France and DONG Energy, Denmark are also
thanked for financial support and valuable discussions.
The bulk experiments were performed in five series in a 15 mL
custom-made cylindrical Schlenk flask fitted with a tube (with or
without a glass frit) reaching almost to the bottom and valves
both on the inlet and outlet. Enough liquid is added to cover the
opening of the insert so the gas is passing through the liquid.
After closing the valves, the entire flask can be disconnected from
the gas line and weighed.
For series 1 and 2, [BMIM][CH COO] (1 mL) was used. For series 3
3
Keywords: gas absorption · ionic liquids · IR spectroscopy ·
nitric acid · NOx
and 5, [BMIM][NO ] (1 mL and 0.9 mL, respectively) was used. For
3
series 4, NaNO (2 mL, 6m) in degassed water was used. The gases
3
ꢀ
1
were pure 10% v/v NO at 10 mLmin for series 1 and 2; 5% v/v
NO, 10% v/v O , and approximately 1% v/v water at 30 mLmin
ꢀ1
2
[1] Health Aspects of Air Pollution with Particulate Matter, Ozone and Nitro-
gen Dioxide: Report on a WHO Working Group, World Health Organiza-
tion, Regional Office for Europe, 2003.
for series 3; and 5% v/v NO, 10% v/v O , and approximately
2
ꢀ
1
1
% v/v water at 10 mLmin for series 4 and 5. All five series were
followed gravimetrically by weighing several times. In series 1 and
, the reaction vial was submerged in an oil bath at 1008C while
[
[
2] N. Meine, F. Benedito, R. Rinaldi, Green Chem. 2010, 12, 1711–1714.
3] C. P. Fredlake, J. M. Crosthwaite, D. G. Hert, S. N. V. K. Aki, J. F. Brennecke,
J. Chem. Eng. Data 2004, 49, 954–964.
2
under the gas flow for several days to obtain steady state. Then
the oil bath was removed and the flask was cooled down to RT.
Still under flow, a new steady state was obtained at room tempera-
ture after several days. Series 3 was performed at RT and followed
by weighing multiple times. Series 4 and 5 consists of several indi-
vidual experiments at RT, which were stopped after 2, 4, 6, 8, and
[
[
4] A. Noda, M. A. B. H. Susan, K. Kudo, S. Mitsushima, K. Hayamizu, M. Wa-
tanabe, J. Phys. Chem. B 2003, 107, 4024–4033.
5] H. Tokuda, K. Hayamizu, K. Ishii, M. A. B. H. Susan, M. Watanabe, J. Phys.
Chem. B 2004, 108, 16593–16600.
[6] J. Huang, A. Riisager, P. Wasserscheid, R. Fehrmann, Chem. Commun.
2006, 4027–4029.
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
10
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
[
7] C. Wang, G. Cui, X. Luo, Y. Xu, H. Li, S. Dai, J. Am. Chem. Soc. 2011, 133,
[21] H. A. Duong, T. N. Tekavec, A. M. Arif, J. Louie, Chem. Commun. 2004,
112 – 113.
11916–11919.
[
[
8] J. F. Brennecke, B. E. Gurkan, J. Phys. Chem. Lett. 2010, 1, 3459–3464.
9] P. Thomassen, A. Riisager, A. J. Kunov-Kruse, S. Mossin, H. Kolding, S.
Kegnæs, R. Fehrmann, ECS Trans. 2013, 50, 433–442.
[22] R. N. Wiener, E. R. Nixon, J. Chem. Phys. 1957, 26, 906–908.
[23] A. J. Kunov-Kruse, A. Riisager, S. Saravanamurugan, R. W. Berg, S. B. Kris-
tensen, R. Fehrmann, Green Chem. 2013, 15, 2843–2848.
[24] Z. Guo, B.-M. Lue, K. Thomasen, A. S. Meyer, X. Xu, Green Chem. 2007, 9,
1362–1373.
[
[
10] Z. Lei, C. Dai, B. Chen, Chem. Rev. 2014, 114, 1289–1326.
11] R. W. Berg, P. Harris, A. Riisager, R. Fehrmann, J. Phys. Chem. A 2013, 117,
1
1364–11373.
[25] J. L. Anthony, J. L. Anderson, E. J. Maginn, J. F. Brennecke, J. Phys. Chem.
B 2005, 109, 6366–6374.
[26] H. Tsukahara, T. Ishida, Y. Todoroki, M. Hiraoka, M. Mayumi, Free Radical
Res. 2003, 37, 171–177.
[
[
[
12] L. Cammarata, S. G. Kazarian, P. A. Salter, T. Welton, Phys. Chem. Chem.
Phys. 2001, 3, 5192–5200.
13] L. J. Murphy, A. M. McPherson, K. N. Robertson, J. A. C. Clyburne, Chem.
Commun. 2012, 48, 1227–1229.
14] A. Riisager, A. J. Kunov-Kruse, S. Mossin, R. Fehrmann, Absorption and
Oxidation of NO in Ionic Liquids, 2013, WO2013079597.
15] F. Mꢂlen, M. Herman, J. Phys. Chem. Ref. Data 1992, 21, 831–881.
16] D. C. Harris, M. D. Bertolucci, Symmetry and Spectroscopy, Oxford Univer-
sity Press, New York, 1978.
[27] H. Kolding, P. Thomassen, S. Mossin, S. Kegnæs, A. Riisager, J. Rogez, G.
Mikaelian, R. Fehrmann, ECS Trans. 2014, 64, 97–108.
[28] P. C. Ford, D. A. Wink, D. M. Stanbury, FEBS Lett. 1993, 326, 1–3.
[29] S. Goldstein, G. Czapski, J. Am. Chem. Soc. 1995, 117, 12078–12084.
[30] B. J. Finlayson-Pitts, N. A. Saliba, L. M. Wingen, W. S. Barney, M. Mochida,
H. Yang, Rep. Calif. Air Ressources Board Calif. Environ. Prot. Agency 2001,
1–166.
[31] M. Mochida, B. J. Finlayson-Pitts, J. Phys. Chem. A 2000, 104, 9705–9711.
[32] J. Kleffmann, T. Benter, P. Wiesen, J. Phys. Chem. A 2004, 108, 5793–
5799.
[33] R. Svensson, E. Ljungstrçm, Int. J. Chem. Kinet. 1988, 20, 857–866.
[34] T. Kinugawa, S. Enami, A. Yabushita, M. Kawasaki, M. R. Hoffmann, A. J.
Colussi, Phys. Chem. Chem. Phys. 2011, 13, 5144–5149.
[35] D. E. Damschen, L. R. Martin, Atmos. Environ. 1983, 17, 2005–2011.
[36] P. K. Mudgal, S. P. Bansal, K. S. Gupta, Atmos. Environ. 2007, 41, 4097–
4105.
[
[
[
17] K. Noack, P. S. Schulz, N. Paape, J. Kiefer, P. Wasserscheid, A. Leipertz,
Phys. Chem. Chem. Phys. 2010, 12, 14153–14161.
[
18] Gaussian 09, Revision A.02, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Men-
nucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian,
A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara,
K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro,
M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Ko-
bayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyen-
gar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B.
Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,
O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg,
S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cio-
slowski, D. J. Fox, Gaussian Inc., Wallingford CT, 2009.
[37] N. Takenaka, A. Ueda, T. Daimon, H. Bandow, T. Dohmaru, Y. Maeda, J.
Phys. Chem. 1996, 100, 13874–13884.
[38] Q. Yang, H. Xing, Z. Bao, B. Su, Z. Zhang, Y. Yang, S. Dai, Q. Ren, J. Phys.
Chem. B 2014, 118, 3682–3688.
[39] PeakFIT v. 4.11, Systat Software Inc, 2009.
[
[
19] H.-P. Steinrꢃck, P. Wasserscheid, Catal. Lett. 2015, 145, 380–397.
20] M. Thiemann, E. Scheibler, K. W. Wiegand in, Ullmann’s Encyclopedia of
Industrial Chemistry, Vol. 24, Wiley-VCH, Weinheim, 2012, pp. 177–223.
Received: March 11, 2016
Published online on && &&, 0000
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
11
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
FULL PAPER
The answer is NO! A new vision for re-
moving NO from exhaust gases is pre-
sented. Nitrate-based ionic liquids cata-
&
Ionic Liquids
A. J. Kunov-Kruse, P. L. Thomassen,
A. Riisager, S. Mossin,* R. Fehrmann*
lyze the fast oxidation of NO by O to
2
nitric acid (see figure), which is accumu-
lated in the non-volatile ionic liquid in
a ratio of 1:1 at room temperature. The
reaction is followed in situ in a thin film
by ATR-FTIR spectroscopy. A mechanism
for the catalytic reaction is suggested.
&
& – &&
Absorption and Oxidation of Nitrogen
Oxide in Ionic Liquids
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
12
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!