Electrochemical degradation of butyltrimethylammonium bis(trifluoromethylsulfonyl)imide for lithium battery applications

Article abstract of DOI:10.1039/c4nj00355a

Ionic liquids (ILs) are being considered as electrolytes for lithium ion batteries due to their low volatility, high thermal stability, and wide electrochemical windows which are stable at the strongly reducing potentials present in Li/Li⁺ batteries. Lithium metal deposition occurs under strongly reducing conditions and the effect that Li metal and any overpotential has on the stability of ILs is important in furthering the application of ILs in lithium based batteries. Here, N-butyl-N-trimethylammonium bis(trifluoromethylsulfonyl)imide was exposed to various potential differences in order to collect and characterize the volatile products. The IL produced more volatile products when exposed to strong reducing potentials which included reactive products such as hydrogen, alkanes, and amines. Water is a known contributor to hydrogen production in reducing environments, but the IL is also a source of hydrogen. If Li⁺ was present, the preferred pathway of reduction was plating of the lithium onto the working electrode, thus decreasing the reaction rate of degraded ILs.

Full text of DOI:10.1039/c4nj00355a

NJC
View Article Online
View Journal | View Issue
PAPER
Electrochemical degradation of butyltrimethyl-
ammonium bis(trifluoromethylsulfonyl)imide for
lithium battery applications
Cite this: New J. Chem., 2014,
38, 3879
Christopher L. Klug,ab Nicholas J. Bridges,*a Ann E. Visser,a Stephen L. Crumpa and
Eliel Villa-Alemana
Ionic liquids (ILs) are being considered as electrolytes for lithium ion batteries due to their low volatility, high
thermal stability, and wide electrochemical windows which are stable at the strongly reducing potentials
present in Li/Li+ batteries. Lithium metal deposition occurs under strongly reducing conditions and the effect
that Li metal and any overpotential has on the stability of ILs is important in furthering the application of
ILs in lithium based batteries. Here, N-butyl-N-trimethylammonium bis(trifluoromethylsulfonyl)imide was
exposed to various potential differences in order to collect and characterize the volatile products. The IL
produced more volatile products when exposed to strong reducing potentials which included reactive
products such as hydrogen, alkanes, and amines. Water is a known contributor to hydrogen production
in reducing environments, but the IL is also a source of hydrogen. If Li+ was present, the preferred
pathway of reduction was plating of the lithium onto the working electrode, thus decreasing the
reaction rate of degraded ILs.
Received (in Montpellier, France)
10th March 2014,
Accepted 27th May 2014
DOI: 10.1039/c4nj00355a
www.rsc.org/njc
products. With ILs, it is assumed that the observed electrochemical
cathodic and anodic potential limits are associated with the
Introduction
In electrochemical applications, the stability of the electrochemical reduction of cations and oxidation of anions, respectively.9,10
medium and the width of the electrochemical window determine However, some results show the IL anion bis(trifluoromethyl-
the utility and application. For example, in lithium ion batteries, sulfonyl)imide (Tf2Nꢀ) undergoes reductive decomposition
the electrolyte must remain inert to electrochemical potentials before the IL cation.11 Based on quantum mechanical calcula-
sufficient to reduce lithium ions to metal. Alkyl carbonates were tions for cathodic decomposition products of 1,1-butylmethyl-
some of the first solvents to be used in lithium ion batteries since pyrrolidinium ([C4,1pyrr][Tf2N]), Kroon et al. propose a range of
they have high dielectric constants and low viscosities.1 Other decomposition products; methylpyrrolidine and a butyl radical,
electrolyte systems examined for this application include polymer dibutylmethylamine radical via ring opening, and butylpyrrolidine
and gel electrolytes.2 A particular emphasis in the development of and a methyl radical.12 After [C4,1pyrr][Tf2N] electrochemical
advanced electrolytes for lithium ion batteries is reducing the degradation experiments, 1-methylpyrrolidone, octanes, octenes,
volatility and flammability associated with alkyl carbonates.
isobutanol, dibutylmethylamine, and 1-butylpyrrolidine were identi-
ILs are attractive media for lithium ion batteries since fied.12 The reduction reactions for quaternary ammonium-based ILs
they offer high conductivity, high thermal stability, and wide have been investigated computationally and there are two possible
electrochemical windows. Recent reports show a variety of ways one-electron reduction reactions, one of which includes loss of a
in which ILs have been used in energy storage applications.3–6 propyl group.13 The gas phase calculations show that the C–N bond
To overcome the relatively high viscosity of ILs, ILs can be breakage (loss of a propyl group) is energetically favorable over the
diluted with traditional solvents with little impact on the reduction reaction where the IL cation remains intact.
conductivity.7,8
Previous literature using ILs do not always use term ‘‘Ionic
The electrochemical redox reactions of ILs have also been Liquid’’, but should not be ignored.14 In this study the use of
experimentally investigated to determine the IL decomposition tetraethylammonium tertafluoroborate ([N2,2,2,2][BF4]) was used
as an electrolyte solvent in acetonitrile diluent, commercially
available under the trade name Digirena E. Digirena E is a
traditional organic solvent with the addition of an IL as the
electrolyte, thus the bulk of the properties are of the acetonitrile
solvent. The experiments focused on real life conditions of
a Savannah River National Laboratory, Aiken, SC 29808, USA.
E-mail: Nicholas.Bridges@SRNL.doe.gov; Tel: +1 803-725-7279
b Georgia Regents University, Department of Chemistry and Physics, 1120 15th St.,
Augusta, 30912, USA
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 3879--3884 | 3879

View Article Online
Paper
NJC
temperature between 70–100 1C, and long exposures to very high
potential (2.5–30 h of exposure to electrochemical potentials as
high as 6 V, no reference point provided). Extensive analysis of
the by-products produced includes several multi-step reactions
with actively formed intermediates.
The N-butyl-N-trimethylammonium bis(trifluoromethylsulfonyl)-
imide ([N4,1,1,1][Tf2N]) IL has been a particularly promising IL for
battery applications,15,16 thus our investigation on the effects of
charge and current density on the degradation of [N4,1,1,1][Tf2N]
with the addition of a co-solvent, acetonitrile. Headspace gas
samples were collected over each half-cell independently. The
non-condensable gases were measured by mass spectrometry (MS),
and product identities were confirmed with gas chromatography-
mass spectrometry (GC-MS) and Fourier transform-infrared
spectroscopy (FT-IR). The presence of lithium ion in the system
did not affect the reduction reaction, just the rate of production
due to the electrochemical deposition of the lithium metal.
Fig. 1 Effect of variable charge on the reductive degradation of [N4,1,1,1][Tf2N]
with ꢀ5 mA for 8 hours (red), compared to the blank (black). All data was
normalized to the pressure of each sample.
Results and discussion
Direct injection mass spectrometry (DI-MS)
An earlier report had shown electrochemical windows for a
variety of ILs and a lower cathodic limit of ꢀ2.6 V for
[N4,1,1,1][Tf2N] in comparison to ꢀ2.4 V for [C4,1pyrr][Tf2N], vs.
ferrocene couple.17 Other researchers have reported on the
stability of quaternary ammonium cations, especially in the
presence of lithium salts, such that the cathodic limit allows for
battery research to be pursued using [N4,1,1,1][Tf2N].13,16,18 At
the potentials used for plating and stripping Li, currents
Fig. 2 Effect of charge (ꢀ5 mA, variable time) on the concentration of
degradation products.
observed in the experimental setup can approach or slightly database spectra for fragmentation of butene or butadiene
exceed ꢁ5 mA. Since many batteries can take a few hours to isomers. Mass 71.2 could result from the degradation of the
fully recharge, any materials used in the battery should be able quaternary ammonium cation to form a buteneamine, or it
to tolerate currents of several milliamps for several hours.
could be from fragmentation of butan-1-amine.20 Increasing
To examine the effect of variable charge on the degradation the degradation time (total charge) resulted in an exponential
of [N4,1,1,1][Tf2N], ꢁ5 mA was run for 1, 2, 4, 6, or 8 hours increase in the concentration of degradation products as shown
through [N4,1,1,1][Tf2N] samples diluted with acetonitrile. While in Fig. 2. (The DI-MS was not calibrated for quantification of all
there are a number of peaks present in all of the spectra at an the gaseous species measured. The rate of increase of the
m/z less than 45 amu, hydrogen was the only one of these peaks response from the DI-MS is not guaranteed to be linear to the
which is due to the degradation of the IL, the other peaks were increase in the quantity of the species.)
inert environmental background. Results from the reduction of
IL for 8 hours are shown in Fig. 1. Products clearly related to the would lead to the formation of the CF3 radical, which could
Cleavage of the terminal trifluoromethyl group from [Tf2N]ꢀ
ꢂ
total charge passed through the system are found at m/z of 2.01, abstract hydrogen producing a volatile compound. Within the
+
55.1, 56.1, 57.1, 58.1, 59.1, 69.1, 70.1, 71.2, 72.2, and 101.4 amu. MS, trifluoromethane was observed as CF3 and CF2H+ at m/z
Other studies using radiolysis indicate that the formation 69.1 and 51.1 amu, respectively.9,20 The formation of trifluoro-
of alkyl moieties from quaternary ammonium cations with methane is the only evidence of the anion’s degradation and no
and without amine groups is possible.19 Computationally, the evidence was observed for the formation of sulfur containing
electrochemical reduction of quaternary ammonium cations to gases or gases that contained nitrogen originating from electro-
an amine and an alkyl radical has also been found to be the chemical degradation of the [Tf2N]ꢀ in this study.
most likely reaction. The alkyl radicals could be quenched by
The [N4,1,1,1][Tf2N] diluted with an equal volume of aceto-
the abstraction of hydrogen from neighboring cations. The nitrile was reduced by constant charge (72 C) over variable
alkyl and amine products from [N4,1,1,1][Tf2N] would be volatile times to study the effect of current density on degradation. The
and ranging from m/z 16 to as high as 185 amu. The products at normalized integrated peak areas for hydrogen, butane, butene,
m/z 55.1–59.1, 71.2, 72.2, and 101.4 amu are consistent with and trifluoromethane are shown in Fig. 3, illustrating the
alkyl, alkenyl, and amine products as described in literature.19 constant production of degradation gases given constant
The fragmentation pattern observed in the mass spectra resembles charge across six different current densities. Within the tested
3880 | New J. Chem., 2014, 38, 3879--3884
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014

View Article Online
NJC
Paper
Fig. 5 Effect of lithium ions on the degradation of [N4,1,1,1][Tf2N] for
8 hours at +5 mA (green, solid) and ꢀ5 mA (red solid) with 0.5 M Li; and
without Li (black, dashed[+] and dotted [ꢀ]). The product spectra from
reduction are offset by a factor of 10. Spectra are normalized to individual
sample pressures.
Fig. 3 Effect of current on hydrogen (2), butane (58), butene (56) and
trifluoromethane (69) production from reduction of [N4,1,1,1][Tf2N]. All data
was normalized to the pressure of each sample, and the maximum
measured concentration. Since the product of current and time was set
to a constant 20 mA h, and the same foil area was used in each
experiment, the current applied is proportional to the current density.
hence decreasing the reduction efficiency of the [N4,1,1,1]+ into
alkyl and amine products. The trifluoromethane product results
from degradation through direct oxidation of the [Tf2N]ꢀ or from
indirect degradation of the [Tf2N]ꢀ during reduction of the IL.
In the samples from oxidative degradation with Li+ in solution
(Fig. 5) there are common peaks at m/z of 51.1 and 69.1 amu, both
of which have been attributed to trifluoromethane.
GC-MS analysis
The [N4,1,1,1][Tf2N] was degraded with ꢀ5 mA for 8 hours, and
the chromatogram of the gas sample is shown in Fig. 6 with
significant peaks numbered. The first peak (peak 1) appears to
be broad, and begins as a shoulder of the most intense peak in
the chromatogram and through mass spectrum was confirmed
to be trifluoromethane.
The fragmentation patterns for peaks 2, 3, and 5 are identical
and match those for cis- and trans-2-butene and 2-methylpropene.
With column separations being roughly based on the boiling
point, peak 2 would be 2-methylpropene (b.p. = ꢀ6.9 1C), peak 3
trans-2-butene (b.p. = 0.8 1C), peak 5 is cis-2-butene (b.p. = 4 1C),
and peak 6 is butane. Peak 4 appearing at 3 minutes was assigned
to 2-methylprop-1-en-1-amine, as shown in Fig. 7. This assign-
ment was made based on the molecular ion at m/z = 71, and
the base peak at m/z = 55 corresponding to a loss of NH2.
Fig. 4 Effect of variable charge on the oxidative degradation of [N4,1,1,1][Tf2N]
with +5 mA for 8 hours (green), compared to the blank (black). All data was
normalized to the pressure of each sample.
range, the production of gases did not depend on current
density but only on the total charge passed, thus no threshold
for degradation dependent upon current density was observed.
The oxidized samples at +5 mA for varied time produced
hydrogen and trifluoromethane, Fig. 4 represents only eight
hours of oxidation. If there was any production of butane/
butene from the oxidation of the IL, it was at the analytical
detection limit. The majority of butane/butene produced would
be from the reduction half-cell. In contrast to the reductive
degradations, there was no relationship observed between
hydrogen production and the amount of oxidizing charge.
Trifluoromethane was present at a much lower yield from
oxidation than from reduction, but was still clearly related to
the total charge passed.
Samples of [N4,1,1,1][Tf2N] with 0.5 M lithium bis(trifluoro-
methylsulfonyl)imide (Li[Tf2N]) diluted with equal volumes of
acetonitrile were degraded to compare gaseous products with
and without Li+ present. The mass spectra from the degrada-
tion samples with Li+ present are similar to those without Li+,
as shown in Fig. 5. The same gas peaks are present in
the samples from reductive degradation at m/z of 50.1 through
59.1 amu, 69.1 amu, and 101.4 amu. In further comparison of
the systems with Li to those without, it was observed that the
addition of lithium ion decreased the production of all of the
observed products formed from reduction, except for trifluoro-
methane. The Li ion is reduced to the metal in this system,
Fig. 6 Total ion count versus time for [N4,1,1,1][Tf2N] degraded with ꢀ5 mA
for 8 hours.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 3879--3884 | 3881

View Article Online
Paper
NJC
Fig. 9 FT-IR spectra of non-condensable gas degradation products
formed from ꢀ5 mA over 8 hours in [N4,1,1,1][Tf2N] in a 10 cm path length
cell. Identified products include: ammonia, methane, butane, trimethyl-
amine, trifluoromethane, and acetonitrile used to dilute the ionic liquid and
increase conductivity.
Fig. 7 Mass spectra for peak 4 at 3 minutes (top) identified as 2-
methylprop-1-en-1-amine, and a peak at 8.6 minutes (bottom) identified
as butan-1-amine. Background has been subtracted.
931, 908, and 868 cmꢀ1 were assigned to ammonia. Acetonitrile
(diluent) was observed at several peaks including 2988, 2622,
1554, 1514, 1421, 1377, 1337, 1059, and 716 cmꢀ1 with the
neighbouring broad peaks. Overlapping with one of the broad
acetonitrile peaks near 716 cmꢀ1 is a peak at 700 cmꢀ1, which
along with peaks at 1152, 1208, and 3035 cmꢀ1 are attributed to
trifluoromethane. Trimethylamine is also assigned to the peaks
at 2962, 2823, 2774, 1457, 1271, 1179 (as a shoulder of the
1152 cmꢀ1 peak), and 1050 cmꢀ1. Some low and broad absorbance
peaks remain at 1692, 1635, 1380, 895, and 789 cmꢀ1 which at this
point have not been identified.
Another amine, eluting at 8.6 minutes, is shown and identified
in Fig. 7 as butan-1-amine. This product is identified based on
the molecular ion at m/z = 73, the loss of NH2 leading to the
peak at m/z = 57, and the typical butyl fragmentation pattern.
Eluting at 14.2 minutes was a product predicted to be N-ethyl-
N-methylbutan-1-amine, as shown in Fig. 8. This assignment was
made based on the molecular ion at m/z = 114, and the base peak
at m/z = 57 corresponding to a butyl fragment. The fragment at
m/z = 72 is again a butan-1-amine (–H), and the peak at m/z = 85
corresponds to a loss of an ethyl group from the N-ethyl-N-
Hydrogen production as a function of water content
methylbutan-1-amine. This product is thought to result from a
The hydrogen observed under reducing conditions could result
from radical reactions within the ionic liquids, or from the
reduction of water. All ILs will maintain some hydrated water in
them. In order to better understand the source of the hydrogen,
the same IL system was ‘doped’ with water and held under
similar reducing conditions. The dried IL had been measured
for water content by Karl Fischer titration, and was found to
contain 52.5 ꢁ 2.1 ppm of water. Analysis in the literature has
shown the water content of saturated [N4,1,1,1][Tf2N] to be
14 000 ppm.22 As seen in Fig. 10, there does seem to be a
maximum rate of hydrogen production. Other researchers have
observed a plateau effect at higher water concentrations
for hydrogen production from water electrolysis in a different
ILs, which they attributed to the solubility limit of water.23
+
rearrangement reaction of the [N4,1,1,1
]
within the electro-
chemical cell yielding the charge neutral amine.
FTIR analysis
The absorbance spectrum of a gas sample from [N4,1,1,1][Tf2N]
degraded with ꢀ5 mA for 8 hours was measured in a 10 cm path
length cell, equilibrated to atmospheric pressure. The resulting
spectrum, shown in Fig. 9, was analysed using a database
from Infrared Analysis, Inc. and the Aldrich Library of FT-IR
Spectra.21 Methane was identified by the peaks at 3113, 3017,
2927, and 1306 cmꢀ1 and the associated rotational bands. The
sensitivity of methane is greater in the FTIR measurements
than the mass-spectrometry based analysis. Butane was
assigned to the 2968, 2877, and 1464 cmꢀ1 peaks. The peaks
at 3334 (and rotational bands), 1671, 1625, 1122, 1046, 992, 966,
Fig. 10 Hydrogen production from dried [N4,1,1,1][Tf2N], water saturated
[N4,1,1,1][Tf2N], and intermediate concentrations, with the integrated peak
area of hydrogen (m/z = 2.01) plotted verses water content.
Fig. 8 Mass spectrum for a product eluting at 14.2 minutes, identified as
N-ethyl-N-methylbutan-1-amine.
3882 | New J. Chem., 2014, 38, 3879--3884
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014

View Article Online
NJC
Paper
A different study in acidic IL attributed lower rates of hydrogen Experimental
production to the slow Volmer reaction, being the simultaneous
Electrochemical degradation
electron transfer to a proton and its adsorption to the electrode
surface.24 Because the samples studied for hydrogen production
were below the saturation limit of water, the plateau effect is
most likely due to rate limitations of the charge transfer reaction,
but further study is needed for conclusive evidence. However, the
important findings are that in practical applications, even the
driest of ILs will still retain some water leading to additional
hydrogen production, and that even an ionic liquid with no
water would still lead to the reductive formation of hydrogen gas.
An H-cell was designed and made by the SRNL Glass Shop such
that gas samples could be drawn into a pre-evacuated bottle
from one side of the apparatus. The [N4,1,1,1][Tf2N] purchased
from IoLiTec was diluted with an equal volume of acetonitrile and
decolorized with activated carbon equal to a tenth of the IL mass.
The suspension was filtered, and the cleaned [N4,1,1,1][Tf2N] was
dried at 90 1C under argon flow, then the pressure was reduced
until a constant volume was achieved.
Equal volume ratios of dry acetonitrile from Sigma-Aldrich
with [N4,1,1,1][Tf2N] were used to dilute the IL. A platinum foil
working electrode of approximately 4 mm by 16 mm by 0.2 mm
was lowered halfway into the IL, and a larger piece of platinum
foil was used as the counter electrode. At least 8 hours were
allowed for complete wetting of the Vycor frit before the
experiment was started. A PAR273A potentiostat was used with
CorrWare software.
At the end of the degradation run the pre-evacuated bottle
was opened collecting the non-condensable gases from the
headspace at sub-atmospheric pressures. Blank samples were
also collected, where all conditions were the same except no
current was passed through the system. For the lithium loaded
systems, dry Li[Tf2N] salt from Sigma-Aldrich was used to
prepare 0.5 M solutions of Li in [N4,1,1,1][Tf2N].
Conclusions
The electrochemical degradation of [N4,1,1,1][Tf2N] in acetonitrile
produces flammable gases, greenhouse gases, and noxious and toxic
gases within short exposure times not too far out of standard Li/Li+
potentials. With the exception of trifluoromethane, all gaseous
by-products originate from the cation, [N4,1,1,1]+, summarized
in Table 1. The source of all nitrogen-based compounds was
the quaternary ammonium cation, as there was no observed
evidence of sulphur-based compounds supporting further
degradation of the anions.
The presence of dissolved water increases the production of
hydrogen via electrolysis, but ILs will also produce hydrogen due
to internal reduction. In reducing environments, the presence of
lithium ion decreases the electrochemical degradation of the IL
due to lithium metal deposition on the working electrode. In the
cases of overpotentials or depletion of Li+, the rate of gaseous
production from degradation of the IL will increase.
Gas analysis (DI-MS)
An OmniStar GSD 320 O2C mass spectrometer with an yttriated
iridium filament was used for gas analysis. The mass spectro-
meter was run using Quadera software in scan mode from 1 to
199 amu (m/z) at 200 ms amuꢀ1 and a resolution of 5. Back-
ground scans were taken with nitrogen purge gas flowing for
three scans before analyzing each sample over three scans. The
sample pressures above the mass spectrometer isolation valve
were at least 2 psia, and did not vary during analysis by more than
0.01 psia. The sample signal was calculated as the difference
between the average of the background scans taken just prior to
the samples and the average of the sample scans. Spectra were
analyzed by normalizing the sample signal to the gas pressure.
With proper system design and controls, ILs could remain
viable for commercial use in Li/Li+ systems. Where ILs would be
used as an electrolyte in a Li/Li+ system, the highly conductive
nature of the ILs, with no measureable vapour pressure still
provides them as a viable option for commercial use.
Table 1 Gaseous degradation products from dried [N4,1,1,1][Tf2N]. Masses
in parentheses correspond to products only observed by FT-IR. Masses in
brackets were only observed by GC-MS. All other products were observed
using DI-MS in combination with data from GC-MS, FTIR, and literature
information
Gas analysis (GC-MS)
Product mass
on DI-MS (m/z) degradation product
Reductive
Secondary method
of confirmation
The headspace samples were analyzed by purge and trap Gas
Chromatography-Mass Spectroscopy (GC-MS) by injecting the
gases into the purge system during the purge cycle. Analytical
separations were carried out on a Hewlett Packard 6890 gas
chromatograph, equipped with a 20 m DB-624 column with
0.18 mm diameter and 1.0 mm film thickness. Gas samples were
injected directly into a water purge system, so water soluble
products such as ammonia, light amines, and methane would
not be carried through onto the GC column. Species with m/z o
50 amu were also excluded from the collected data due to
experimental background. Quantitation was performed using a
Hewlett Packard 5973 mass selective detector. Initial chromato-
grams were obtained by injecting 1 mL gas samples without
splitting the injection. The injection port was held at 200 1C and
2.01
Hydrogen
Methane
Ammonia
N/A
FTIR
FTIR
GC-MS, FTIR
GC-MS
(16.0)
(17.0)
51.1
55.1
56.1
57.1
58.1
59.1
69.1
70.1
71.2
72.2
[73.0]
101.4
[114.0]
Trifluoromethane
2-Butene, 2-methylpropene
2-Butene, 2-methylpropene
Butane
Butane, trimethylamine
Trimethylamine
Trifluoromethane
2-Methylprop-1-en-1-amine
2-Methylprop-1-en-1-amine
Butan-1-amine
GC-MS
GC-MS, FTIR
GC-MS, FTIR
FTIR
GC-MS, FTIR
GC-MS
GC-MS
GC-MS
Butan-1-amine
GC-MS
GC-MS
N,N0-Dimethylbutylamine
N-Ethyl-N-methylbutan-1-amine GC-MS
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 3879--3884 | 3883

View Article Online
Paper
NJC
the column flow was 0.5 mL minꢀ1. The oven temperature was held
at 27 1C for eight min at injection then raised at 20 1C minꢀ1 up to
200 1C. The mass spectrometer electron multiplier was set for 1200
volts for the initial chromatographic measurements.
4 S. Pohlman, B. Lobato, T. A. Centeno and A. Balducci, Phys.
Chem. Chem. Phys., 2013, 15, 17827.
5 A. Brandt, C. Ramirez-Castro, M. Anouti and A. Balducci,
J. Mater. Chem. A, 2013, 1, 12669.
6 D. R. MacFarlane, N. Tachikawa, M. Forsyth, J. M. Pringle,
P. C. Howlett, G. D. Elliott, J. H. Davis, M. Wantanabe,
P. Simon and C. A. Angell, Energy Environ. Sci., 2014, 7, 232.
7 E. T. Fox, E. Paillard, O. Borodin and W. A. Henderson,
J. Phys. Chem. C, 2013, 117, 78.
8 S. Menne, R.-S. Ku¨hnel and A. Balducci, Electrochim. Acta,
2013, 90, 641.
9 M. Buzzeo, C. Hardacre and R. Compton, ChemPhysChem,
2006, 7, 176.
10 E. I. Rogers, B. Sljukic, C. Hardacre and R. Compton,
J. Chem. Eng. Data, 2009, 54, 2049.
11 P. C. Howlett, E. I. Izgorodina, M. Forsyth and
D. R. MacFarlane, Z. Phys. Chem., 2006, 220, 1483.
12 M. C. Kroon, W. Buijs, C. J. Peters and G.-J. Witkamp, Green
Chem., 2006, 8, 241.
Gas analysis (FT-IR)
A 10 cm path length glass cell was purged with nitrogen gas,
and sealed with rubber gaskets. The 10 cm cell was placed in
the sampling chamber of a Nicolet 6700, and a background
spectrum was collected using 1 cmꢀ1 resolution, triangular
apodization, and 512 scans. An empty syringe was inserted to
partially evacuate the glass cell, removed, and the sample
syringe was inserted and sample was backfilled into the cell.
A rubber gasket was briefly opened to equilibrate the sample
pressure to atmospheric pressure. All data was collected at
room temperature and atmospheric pressure. Automatic CO2
and H2O subtraction was performed by the OMNIC software on
the instrument. During data analysis spectral subtraction was
performed with the Essential FT-IR program as components
were identified, if matched standard spectra were available.
13 Y.-H. Tian, G. S. Goff, W. H. Runde and E. R. Batista, J. Phys.
Chem. B, 2012, 116, 11943.
14 P. Kurzweil and M. Chwistek, J. Power Sources, 2008, 176, 555.
15 J. A. Vega, J. Zhou and P. A. Kohl, J. Electrochem. Soc., 2009,
156, A253.
16 R. Wibowo, S. E. W. Jones and R. G. Compton, J. Chem. Eng.
Data, 2010, 55, 1374.
17 N. J. Bridges, A. E. Visser, M. J. Williamson, J. I. Mickalonis
and T. M. Adams, Radiochim. Acta, 2010, 98, 243.
18 J. Sun, M. Forsyth and D. R. MacFarlance, J. Phys. Chem. B,
1998, 102, 8858.
19 D. Behar, P. Neta and C. Schultheisz, J. Phys. Chem. A, 2002,
106, 3139.
Acknowledgements
Savannah River National Laboratory is operated by Savannah
River Nuclear Solutions. This manuscript has been authored by
Savannah River Nuclear Solutions, LLC under Contract No.
DEAC09-08SR22470 with the U.S. Department of Energy. The
United States Government retains and the publisher, by accepting
this article for publication, acknowledges that the United States
Government retains a non-exclusive, paid-up, irrevocable, world-
wide license to publish or reproduce the published form of this
work, or allow others to do so, for United States Government
purposes.
20 I. A. Shkrob, S. D. Chemerisov and J. F. Wishart, J. Phys.
Chem. B, 2007, 111, 11786.
21 C. J. Pouchert, The Aldrich Library of FT-IR Spectra, Aldrich
Chemical Company, Milwaukee, 1st edn, 1989, vol. 3.
22 J. Jacquemin, P. Husson, A. A. H. Padua and V. Maier, Green
Chem., 2006, 8, 172.
23 R. F. de Souza, J. C. Padilha, R. S. Gonçalves and J. Tault-
Berthelot, Electrochem. Commun., 2006, 8, 211.
References
1 Y. Ein-Eli, S. F. McDevitt and R. Laura, J. Electrochem. Soc.,
1998, 145, L1.
2 W. H. Meyer, Adv. Mater., 1998, 10, 439.
3 E. B. Fox, A. E. Visser, N. J. Bridges and J. W. Amoroso, 24 Y. Meng, L. Aldous, S. R. Belding and R. G. Compton, Phys.
Energy Fuels, 2013, 27, 3385.
Chem. Chem. Phys., 2012, 14, 5222.
3884 | New J. Chem., 2014, 38, 3879--3884
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014