Journal of The Electrochemical Society, 2020 167 086507
4
3
DEA (pK
role of the amine proton.
a
= 18.5 in MeCN). This allows one to directly probe the
or carbamic acid is responsible for fouling. As such, what follows is
meant to be qualitative, as such trends are reproducible.
Sparging an MeCN solution of TEA with CO2 results in no
noticeable equilibrium, as probed by IR spectroscopy. The CV of
CVs were taken between 10 and 100 mM of amine, the range
being chosen to ensure no visible precipitation, with exception of
TEA again shows no reduction under an N atmosphere (Fig. 7).
TMG (Figs. S24–S29). Given that the concentration of CO in
2
2
51
However, under CO
2
an irreversible reduction does occur, with an
MeCN is 0.28 M, the equilibrium should lie towards the carbamic
acid and not the carbamate. As the concentration of aniline is
increased (Figure S24), the current response changes from little
depletion being observed to a peak current that is shifted anodically;
the CVs of 75 and 100 mM aniline are superimposable. A similar
trend is observed for the other amines, the effect being more
pronounced for some.
onset potential of −1.96 V and Einf of −2.64 V (Fig. 7). Addition of
equimolar TMG gives a current response that appears to be less than
the superposition of TMG and the zwitterion. This decrease may be
due to electrode fouling, which we observed with all amines.
Addition of water has little effect on the resulting CV.
Triethanolamine (TEOA) is a commonly used sacrificial reduc-
5
2
tant for photocatalytic reductions, including those with CO2.
Moreover, Ishitani has shown that CO can insert into a Re-TEOA
bond and hence concentrate the CO for photochemical and
The CV of the six amines under CO are shown in Fig. 9a. Only
2
2
TMG shows a true peak, whilst all other amines do not plateau.
Notably, the onset potential for morpoline is significantly shifted
anodically, and the maximum current density is also much larger. A
direct comparison is further complicated by the equilibrium of
carbamate and carbamic acid, which is present under these condi-
tions.
In the presence of TMG, the CVs all show cathodic shifts
consistent with the formation of the carbamate (Fig. 9b). Here, the
amine identity has an effect on the shape of the peak as well as the
current density. From our IR experiments, under these conditions
only the carbamate is formed (carbonate for TEOA), allowing for
more ready comparison. With TEA, no carbamate formation should
2
1
9,29
electrochemical reactions.
CO does not readily react with the tertiary amine, as described
above, or at the alcohol functionality to give a carbonic acid.
Indeed, sparging a MeCN solution of TEOA with CO does not
show an appreciable equilibria as determined by IR spectroscopy
Fig. S32). Given that TEOA is more basic than aniline (pK = 15.93
This reactivity is unusual in that
5
2
2
(
a
43
in MeCN), this likely is due to the inability to form a carbamic acid
from the zwitterion. Addition of TMG to a solution of TEOA under
−
1
2
CO gives a species with an IR stretch at 1644 cm , which we
assign as the carbonate.
The CVs of TEOA under N shows no significant reduction up to
occur with TMG, given that the only interaction with CO is that of a
2
2
−
2.7 V (Fig. 8). When the solution is sparged with CO
2
, a modest
zwitterion. This is observed in Fig. 9b, whereby there is little change
in the CV response. DEA does not show much current enhancement
with the addition of TMG. As IR analysis indicates that now the
equilibrium lies completely towards the carbamate, we attribute this
lack of enhancement to systematic fouling which may minimize the
potential application of this amine in combination with an electro-
catalyst. TEOA, morpholine, and aniline all showed significant
current enhancement upon addition of TMG. Their onset potential
and maximum current densities both follow the same trend, with
TEOA being most anodically shifted and morpholine the least
response is observed, with Einf at −2.60 V. While we did not observe
any reaction by IR spectroscopy, we speculate that a small
concentration of either the zwitterion or the carbonate is formed in
solution, allowing for the reduction to occur. Upon addition of
equimolar TMG, the CV changes significantly. Now, there is a sharp
increase in current density at −1.94 V and Einf is anodically shifted
2
relative to that of TEOA under CO . This suggests that addition of
TMG aids in stabilizing a species, likely the carbonate. This is the
1
9
same proposed species in Ishitani’s system. This shift suggests that
while TEOA itself may not react with CO , limiting its extent in the
2
a
shifted. This trend cannot be rationalized on pK , but is consistent
photo- and electrocatalytic reductions of CO under dilute condi-
with the IR stretches of the resulting carbamate. This strongly
suggests that what is being reduced is a carbamate species and not
free CO . This also indicates that the resulting carbamate IR stretch
2
2
tions, the addition of a base allows for a higher concentration of CO2
to be present in solution. Addition of equimolar water bears no effect
on the onset potential.
is a good indicator of onset potential.
Bulk electrolysis.—To determine if the current observed is due to
Effect of water.—Water is a weak Brønsted acid, commonly
reduction of CO , bulk electrolysis was performed, with conditions
employed as a proton source for the electrochemical reduction of
2
54
outlined in Table II. Formate was quantified from the electrolyzed
solution as described before, and the headspace gas was injected into
carbon dioxide. If carbamates form, then in the presence of water a
5
speciation change to bicarbonate should ensue. Equimolar water has
a GC to quantify CO and H
instance do we observe good Faradaic efficiencies. This may be due
to short-circuiting our setup, or formation of a solution species that
2
we could not observe. We also did not observe any CO or H . Given
our lower detection limits (see SI), we surmise that at most this can
account for ∼0.034% FE. In all instances, we do observe formate,
with some enhancement observed with addition of TMG. Notably,
formate production on gold electrodes has been found to occur from
2
. No other gases were observed. In no
little effect on the CV of TEA with or without TMG, consistent with
the lack of a carbamate species being accessible; the discrepancies
observed are attributed to electrode fouling. Equimolar water only
affects TEOA in the presence of TMG, shifting the onset potential
less negative. The CVs of the secondary amines and aniline were all
impacted with addition of water, the effect more pronounced when
the equilibrium is shifted towards the carbamate (presence of TMG).
The similarity in the peak shape and onset potential suggests a single
species, bicarbonate, which is now being reduced.
4
2
2
bicarbonate reduction whilst CO from CO reduction. Consistent
with this notion, formate is likely produced from direct reduction of
the carbamate.
Bulk electrolysis.—Product analysis from bulk electrolysis
showed no formation of CO or H
2
. This is surprising given that
Discussion
Zhang predominately observes formation of H for bulk electrolysis
of DMA + CO at different electrodes. However, they acknowl-
2
edge that their system contains upwards of 0.3 M water, while we
are working under strictly anhydrous conditions (see SI). With no
2
24
Comparison of the electrochemical performances of 1°, 2° and
° amines.—Electrode fouling does occur, as evident from irrepro-
ducibility in subsequent CV scans using the same electrode.
circumvent this, a freshly polished electrode was used for each scan
with the exception of scan-rate dependences, see SI) and the CVs
presented in the text and SI have been replicated. We cannot rule out
that the electrodes are fouled to various extents with certain amines,
making it difficult to draw conclusions. The fouling was more
pronounced in systems that lacked TMG, suggesting that the amine
3
36,53
To
2
protons from water, less H is expected to form. Zhang and
coworkers also observed a selectivity for formate over CO at certain
electrodes, though CO production always accompanied formate
(
24
production. That formate is selectively formed under the condi-
tions described here is beneficial for application to electrocatalytic
systems whereby CO is the sole product.