of 9
1
Dramatic
HER Suppression on Ag Electrodes via Molecular Films
for Highly Selective CO
2
to CO Reduction
Arnaud Thevenon
,
Alonso Rosas
-
Hernández,
Alex M. Fontani H
erreros, Theodor Agapie,
*
and Jonas
C. Peters*
Joint Center for Artificial
Photosynthesis (JCAP) and Division of Chemistry and Chemical Engineering, California Institute
of Technology (Caltech), Pasadena, California 91125, USA
.
KEYWORDS:
electrocatalysis,
CO
2
RR
,
HER
,
solar fuels
,
modified electrodes, silver electrodes
.
ABSTRACT
:
The carbon dioxide reduction reaction (CO
2
RR) in aqueous electrolytes suffers from efficiency loss due to the com-
petitive hydrogen evolution reaction (HER).
Developing efficient methods to suppress HER is a crucial step toward CO
2
utilization
.
Herein, we report the selective conversion of CO
2
to CO on planar silver electrodes with Faradaic efficiencies >99 % using simple
pyridinium
-
based additives. Similar to our previous studies on copper electrodes, the additives form an organic film which alt
ers
CO
2
RR selectivity. We report electrochemical kinetic and other mechanistic data to shed light on the role of these organic layer
s in
suppressing HER. These data suggest that hydrogen production is selectively inhibited by the growth of a hydrophobic or
ganic layer
on the silver surface that limits proton but not CO
2
mass transport at certain applied potentials. The data also point to the involvement
of a proton
-
transfer within the rate determining step of the catalysis, instead of the commonly observed e
lectron
-
transfer step for the
case of planar Ag electrodes.
INTRODUCTION
While anthropogenic CO
2
emissions present a daunting societal
challenge with respect to climate change, concentrated CO
2
sources also afford an attractive CO
2
-
recycling opportunity.
1
The electrochemical carbon dioxide reduction reaction
(CO
2
RR) is a catalytic process by which CO
2
-
recycling can be
driven via renewably derived electricity, hence affording a path-
way toward zero
-
or low
-
carbon chemical/fuel feedstocks
.
2
4
For such an approach to have practical utility a host of technical
challenges need
s
to be met, with high chemical selectivity for
desired C
-
containing products being prominent among
them
.
The
use of metallic electrodes
as catalytic materials
is
an im-
portant strategy to facilitate the transformation of CO
2
into a
range of such desirable p
roducts
.
5
9
Nevertheless
, the mecha-
nistic landscape of CO
2
RR is complex and competing proton
-
coupled electron transfer (PCET) pathways can be operative;
control of product selectivity remains a central issue.
A
p-
proaches
to rationally tune the
selectiv
ity o
f
catalytic surfaces
are needed.
Nanostructuring of metallic surfaces, to increase the number of
exposed active sites, has been widely used as a strategy to pre-
pare electrodes with improved reaction kinetics for CO
2
RR.
8,10
14
Detailed analyses of the catal
ytic activity as a function of the
electrochemical surface area (ECSA) have suggested that
nanostructuring does not alter the intrinsic activity of the active
sites in any significant manner.
15
This observation is in agree-
ment with the fact that the outcom
e of electrochemical CO
2
RR
is determined by the adsorption energies of reaction intermedi-
ates,
16
which are typically correlated through thermodynamic
scaling relations.
17
Strategies to break scaling relations and con-
trol reaction pathways to favor a single
product are therefore
desired.
18
20
In this context, molecular modifications of electrode surfaces
provide an attractive approach for the electrosynthesis of de-
sired products in CO
2
RR.
21
29
Our research team has recently
focused on studying the interacti
on of molecular films with cop-
per electrodes to control the selectivity of CO
2
RR.
30
32
We have
disclosed that water
-
soluble
N
-
substituted pyridinium
-
type ad-
ditives undergo electrochemically induced reductive dimeriza-
tion
in
situ
, leading to the deposition
of an organic film onto the
surface of a polycrystalline Cu electrode. As a consequence,
methane generation is highly suppressed and faradaic efficien-
cies (FE) for C
2+
products can be substantially increased, up to
80%.
30
Further, when such molecular films
are electrodeposited
onto copper
-
based gas diffusion electrodes, it is suggested that
selective stabilization of surface intermediate *CO
atop
enhances
the electrosynthesis of ethylene, with reaction rates up to 230
mA cm
-
2
.
31
These electrodeposited molecular films have also
demonstrated utility in the stabilization of nanostructured cop-
per surfaces for the selective production of ethylene.
32
These results with copper notwithstanding, it remains to now
determine whether other
metal electrodes can be similarly mod-
ified, and if so, how such additives, once deposited, may tailor
the electrocatalytic profiles of these metals in the context of
CO
2
RR (or other reductive transformations). The simplicity of
the approach offers attracti
ve opportunities here. In addition,
understanding the mechanistic basis of how such molecular
films alter the CO
2
RR selectivity profile may aid in the design
of new organic
-
metal interfaces with tailored selectivities.
29,31
Studies aimed at building such u
nderstanding are complicated
by the rich product profile of CO
2
RR on copper electrodes.
By contrast, metallic silver surfaces catalyze primarily CO
2
-
to
-
CO conversion in aqueous electrolytes, with concomitant pro-
duction of H
2
and a small amount of HCOOH dep
ending on the
2
potential applied.
9,33
In this regard, comparing modified and un-
modified silver electrodes, the former being prepared via a sim-
ilar additive approach to that which we have explored with cop-
per, provides an appealing strategy for mechanistic s
tudies ow-
ing to the simplicity of the product profile. Herein, we explore
the mechanistic basis for pyridinium
-
based additives to attenu-
ate HER at organic
-
silver interfaces. We show that certain
N
-
substituted pyridinium additives alter the CO
2
RR product pr
o-
file of Ag foils by selectively inhibiting proton (HER) but not
CO
2
reduction within a certain potential window, thereby pro-
ducing CO with extremely high selectivity. A mechanistic hy-
pothesis for this effect is proposed based on a combination of
electroca
talytic, kinetic, and surface characterization studies.
RESULTS AND DISCUSION
Selective Electroconversion of CO
2
to CO
. Bulk electrolysis
experiments were performed on a polycrystalline silver elec-
trode with CO
2
-
saturated 0.1 M KHCO
3
electrolyte (pH = 6.
8)
using a recently reported custom flow cell.
34
Potentials were
measured versus a leakless Ag/AgCl electrode
and converted to
the RHE scale according to the equation: E
(RHE)
= E
Ag/AgCl
+
E
o
Ag/AgCl
+
0.059 pH; where
E
Ag/AgCl
is the measured potential,
E
o
Ag/AgCl
= 0.1976 V at 25
o
C, and pH is the
pH
value of the bulk
.
In the absence of a molecular additive, bare silver electrodes
produced exclusively CO, H
2
, and HCOOH within the tested
potential range (from −0.80 V to −1.20 V,
Fig
ure
1a
-
b
,
Fig
ure
S1, Table S1). Trace CH
4
was detected at higher potential val-
ues, consistent with previous reports.
9,32
The highest FE for CO
was obtained at −1.10 V (FE
CO
= 74%) with concomitant pro-
duction of H
2
(FE
H2
= 19%) and HCOOH (FE
HCOOH
= 3%).
At
bias more reducing than −1.20 V and less negative than −0.90
V, H
2
was the major product and the production of HCOOH
remained constant (FE
HCOOH
< 5%).
Figure 1
. (a) Faradaic efficiencies; (b) total and partial current den-
sities obtained during CO
2
RR at −0.99 V in a CO
2
-
saturated 0.1 M
KHCO
3
electrolyte without or with 10 mM of either
1
-
Br
2
or
2
-
Cl
.
Chemical structure of
1
-
Br
2
(c),
2
-
Cl
(d) and their reduction to
4,4’
-
(1
-
Br)
2
,
2H
-
(1
-
Br)
,
4H
-
1
,
4,4’
-
(2)
2
and
2H
-
2
, respectively.
The electrocatalytic p
erformance of Ag foils was then evaluated
in the presence of molecular additives using a 10 mM solution
of either
1
-
Br
2
or
2
-
Cl
(Fig
ure
1, Fig
ure
S2 and S3, Table S2
and S3). Similar to our previous studies with copper foils,
30,31
a
noticeable organic film
was electrodeposited
in situ
during ca-
talysis onto the surface of the silver electrodes. Most notably,
HER was completely suppressed within the range from −1.00
V to −1.10 V (Fig
ure
1a). In the presence of
1
-
Br
2
, CO
2
was
selectively converted to CO with q
uantitative FE. Similar selec-
tivities were observed when a silver electrode, previously func-
tionalized with the organic film (by the same above
-
mentioned
protocol), was used in bulk electrolysis experiments together
with an additive
-
free electrolyte, confi
rming the role of the or-
ganic film for suppressing hydrogen production rather than so-
lution based pyridinium (Table S4). In addition, the selectivity
was maintained over 24 h with no detectable H
2
. The suppres-
sion of H
2
was also observed under acidic condi
tions at pH val-
ues as low as 4.2, in a phosphate buffer at −1.1 V, highlighting
the robustness of these organic films (Table S5).
This hybrid catalytic system represents the first example of a
silver electrode that is completely selective (up to our detect
ion
limit) for CO
2
-
to
-
CO electroconversion in pH
-
neutral aqueous
solution.
35
39
With respect to demonstrating high current densi-
ties (up to 300 mA/cm
−2
), silver
-
containing gas diffusion elec-
trodes, derived from a Ag coordination polymer, were
recently
shown to produce CO at FEs >96%.
40
In the presence of molecular additive
1
-
Br
2
, the partial current
density for CO (
j
CO
) increased by 20% compared to bare silver
(Fig
ure
1b), indicating that the molecular film promotes the pro-
duction of CO. This o
bservation is notable, as it points to a role
of the film in tuning selectivity beyond inhibition of H
2
evolu-
tion. Additionally, the maximum efficiency for CO is observed
at a potential that is anodically shifted by 100 mV relative to
that observed for bar
e Ag. At potentials less negative than −0.85
V and more reducing than −1.30 V, H
2
is the predominant prod-
uct, even though the partial current densities were 30−60%
lower than those measured for bare silver at the same potentials.
The post
-
catalysis surface
compositions of Ag electrodes with
and without the additives present in the electrolyte were ana-
lyzed
ex situ
by X
-
ray photoelectron spectroscopy (XPS, Fig
ure
S4). For convenience, the post
-
catalysis electrodes in the pres-
ence of
1
-
Br
2
or
2
-
Cl
are herein
abbreviated as
Ag
-
1
and
Ag
-
2
.
Consistent with previous reports,
41
43
ex situ
post
-
catalysis XPS
spectra of bare Ag and
Ag
-
1
feature two peaks separated by 6
eV at 368.4 eV and 374.4 eV. These peaks correspond to Ag
3d
5/2
and Ag 3d
3/2
, respectively, indicat
ing that Ag
0
is predomi-
nately present on the surface of both electrodes. By contrast, the
Ag 3d peaks could not be observed on
Ag
-
2
, likely due to the
thick layer of film electrodeposited (note: the organic film pro-
duced from additive
2
-
Cl
is significantly
thicker than the one
formed from
1
-
Br
2
,
vide infra
). As expected, the XPS spectra
of both
Ag
-
1
and
Ag
-
2
feature N 1s peaks due to the presence
of the organic layer. The spectrum of
Ag
-
1
consists of one broad
N 1s peak at 400.81 eV which was deconvoluted i
nto two peaks
centered at 401.71 eV and 400.71 eV. These peaks are assigned
to quaternary and tertiary amines, respectively.
44,45
The ratio of
tertiary/quaternary amine is approximatively 3:2, which differs
from the ratio we have previously observed on an
additive
-
mod-
ified Cu electrode (tertiary/quaternary amine ratio 1:1)
32
and
suggests a different film composition. Surprisingly, the XPS
spectrum of
Ag
-
2
consists of one N 1s peak at 403.8 eV, ca. 4
eV higher in energy compared to our previous observation f
or
3
a similarly modified copper electrode.
32
The C 1s and O 1s re-
gions of the XPS spectrum of
Ag
-
2
are also shifted to higher
energy. We therefore suspect that the observed shift is due to a
charging effect of the surface of the modified silver electrode
du
e to its thickly insulating organic layer.
46
The organic films could be extracted from post
-
catalysis Ag
electrodes using deuterated DMSO (
Ag
-
1
) and dichloro-
methane (
Ag
-
2
) to investigate their chemical composition by
1
H
NMR spectroscopy (Fig
ure
S5
-
S10). Th
e
1
H NMR spectrum of
the extracted organic film from
Ag
-
2
revealed the presence of
only two species. As previously observed with Cu electrodes,
30
the major species corresponds to the 4,4’
-
coupled dimer [Fig
ure
1d,
4,4’
-
(2)
2
, 63%]. Interestingly, a minor constituent was un-
ambiguously identified as tolyl
-
4
-
dihydro
-
pyridine (Fig
ure
1d,
2H
-
2
, 36%). This was further confirmed by independent syn-
thesis of
2H
-
2
from the reduction of
2
-
Cl
by Na
2
S
2
O
4
in basic
water and comparison with its published spectrum.
47
The organic film on
Ag
-
1
proved more complex to analyze. The
1
H NMR spectrum of the d
6
-
DMSO
-
extracted film showed the
presence of three primary species
(integrated intensity ~95% of
total), along with resonances associated with other trace spe-
cies. Similar to
Ag
-
2
, the characteristic resonances of the previ-
ously reported 4,4’
-
coupled dimer [Fig
ure
1c,
4,4’
-
(1
-
Br)
2
]
were observed (28% of total).
32
Surpris
ingly, the corresponding
2,2’
-
(1
-
Br)
2
dimeric isomer (not shown in Fig
ure
1) was not
detected, contrasting with the findings on copper electrodes.
32
A previously unobserved set of resonances, similar to, but dis-
tinct from
4,4’
-
(1
-
Br)
2
, were observed. This
species has been
tentatively assigned as a mono
-
reduced
2H
-
(1
-
Br)
(Fig
ure
1c,
34%, Fi
gure
S11). The last set of resonances suggested the pres-
ence of a symmetric molecule with an aromatic:alkenyl protons
ratio of 1:2. Accordingly, these resonances have been
attributed
to a doubly
-
reduced species
4H
-
1
(Fig
ure
1c, 38%). Discussion
of these assignments is provided in the Supporting Information.
Altogether, the ratio of tertiary/quaternary amine species deter-
mined by
1
H NMR spectroscopy is 3:2, in agreement with
the
aforementioned XPS data.
The surface of
Ag
-
1
and
Ag
-
2
was also analyzed
ex situ
by scan-
ning electron microscopy (SEM, Fig
ure
2, Fig
ure
S12
-
14) and
atomic force microscopy (AFM,
Fig
ure
S15) after 65 min of
bulk electrolysis at −0.99 V. The SEM and AFM images of a
Ag electrode in the absence of additive are relatively similar to
a freshly polished electrode with a slight increase in the surface
roughness (from 132 nm to 154 nm), indicati
ng only little sur-
face reconstruction during catalysis.
33
Interestingly, the surface
of
Ag
-
1
is partially covered by spherical particles of different
sizes (Fig
ure
2a). Our recent study on Cu electrodes demon-
strated that
1
-
Br
2
promotes nanostructuring of C
u foils by a cor-
rosion mechanism.
32
However, in the present case, energy dis-
persive X
-
ray spectroscopy (EDS) data showed that the ob-
served particles on
Ag
-
1
are composed of carbon and nitrogen
and not Ag (Fig
ure
2c). In addition, the particles were easily
washed away by rinsing the electrode with DMSO, confirming
that they correspond to the organic film (Fig
ure
S13). The
height of the particles is estimated to be up to 200 nm by a cross
-
sectional SEM image of a physical vapor deposited (PVD) Ag
electrode, w
here the organic film was electrodeposited under
the aforementioned conditions. In contrast, the SEM images of
Ag
-
2
clearly show a full coverage of the electrode with a thick
organic layer (Fig
ure
2b and 2d). The cross
-
sectional SEM im-
age of
Ag
-
2
shows a p
orous organic film with a thickness of
around 9
μ
m. Despite differences in surface coverage and film
thickness, both
Ag
-
1
and
Ag
-
2
efficiently suppressed HER. To
get a better understanding of how HER is suppressed in the
presence of the organic film, mecha
nistic studies were per-
formed.
Figure 2
. Top view SEM images of (a)
Ag
-
1
and (b)
Ag
-
2
. Cross
-
sectional SEM images and inserted EDS spectrum of (c)
Ag
-
1
and
(d)
Ag
-
2
.
The Organic Film Alters the Rate Determining Step
. Cur-
rent
-
voltage (Tafel) plots were probed over the potential range
from −0.60 V to −1.30 V (i.e. 0.50 > η > 1.20), for bare Ag,
Ag
-
1
and
Ag
-
2
. The Tafel analysis was performed within the linear
regime as shown in Fig
ure
3a (see also the S
upporting
I
nfor-
m
ation and
Fig
ure
S16). For bare Ag, the obtained Tafel slope
(157 mV/dec) is comparable to previously reported values
(130−150 mV/dec), which have been assigned to a rate deter-
mining step (RDS) involving an electron transfer (ET) to
CO
2
.
36,38,41,48
55
In c
ontrast, the Tafel slope obtained with
Ag
-
1
and
Ag
-
2
is significantly smaller (91 mV/dec and 107 mV/dec,
respectively), indicating improved kinetics for the ET step to
CO
2
. Based on the obtained values for the Tafel slope, it is dif-
ficult to unambiguously
assign the RDS with
Ag
-
1
and
Ag
-
2
.
The theoretical value of the Tafel slope for the first ET or PCET
as the RDS is 118 mV/dec, while the involvement of a proton
transfer (PT) as the RDS results in a value of 59 mV/dec.
50,56
In
the non
-
linear region of the
Tafel plot at high applied potential
(Fig
ure
S16), it can be seen that the CO
2
RR activity decreases
with increasing potential on Ag,
Ag
-
1
and
Ag
-
2
which likely
correlates with a depletion of the *CO
2
•−
intermediate.
57
This
could be attributed to an intrins
ic issue for Ag electrodes, where
the CO
2
concentration is too low at a high applied potential,
likely due to a high CO
2
RR reaction rate, to maintain high
CO
2
RR selectivity.
4
Figure 3
.
(a) Linear regime of the Tafel plot recorded in CO
2
-
saturated 0.1 M KHCO
3
and (b) [HCO
3
] dependence of the CO
current density recorded at different concentrations of CO
2
-
saturated KHCO
3
electrolyte (0.1 M, 0.33 M, 0.66 M and 1 M) at
−0.90
V
NHE
for Ag (
),
Ag
-
1
(●) and
Ag
-
2
(
). Each data point
was recorded at least two times to ensure reproducibility. Also see
Fig
ure
S16 for non
-
linear behavior at higher applied bias.
The electrodeposit
ion of the molecular film onto the surface of
the silver electrode may impose additional mass transport limi-
tation effects, complicating the analysis of electrokinetic stud-
ies. To gain further evidence that the partial current density for
CO
2
RR (
j
CO
) is ki
netically controlled within the voltage range
of the linear region of the Tafel plot (Figure 3a), we compared
the intrinsic activity of bare silver with
Ag
-
2
as a function of
the electrodeposition time of the additive. Within a kinetically
controlled regim
e, we expect that the intrinsic activities for
j
CO
and
j
H2
should be independent of the electrodeposition time. To
test this hypothesis,
Ag
-
2
was prepared by electrodeposition of
2
-
Cl
during different time intervals (from 10 ms to 60 s) at
0.70 V. This po
tential was chosen to ensure that all of the con-
sumed current was used to electrodeposit the molecular addi-
tive, and not for HER. The electrochemically active surface area
(ECSA) was measured before and after the electrodeposition of
the film. The CO and H
2
partial current densities were then
measured at −0.9 V, a potential value within the linear range of
the Tafel plot (see the S
upporting Information
for methodology,
Fig
ure
S24
-
27, Table S5 and S6). These studies were performed
with
Ag
-
2
but not
Ag
-
1
as in the latter case we were unable to
reliably control the amount of film electrodeposited on the elec-
trode surface. The instability of the one
-
electron reduced spe-
cies
(2
-
Cl)
, which increases its propensity for dimerization
compared to
(1
-
Br)
, like
ly contributes to the former’s efficacy
in controlled film formation.
As the electrodeposition times increase, the ECSA exponen-
tially decreases (
vide infra
for more discussion). We also no-
ticed that
j
H2
and
j
CO
decrease linearly with the ECSA, showing
that
the intrinsic activities for
CO
and
H
2
production are con-
stant for any electrodeposition time investigated, and are similar
to bare Ag. This behavior indicates that the film does not affect
the diffusion rates of proton carriers and CO
2
at low overpoten-
ti
als, thus keeping constant the local concentration of proton
carriers and CO
2
at the electrode surface, as in the case of a ki-
netically
-
controlled regime.
To shed more light on a possible involvement of a proton carrier
in the RDS, the influence of the pro
ton donor environment on
the CO
2
RR activity was studied.
58
It has been previously
demonstrated that the source of protons could come from
HCO
3
.
14,35,50,56
Following previous electrokinetic studies per-
formed on mesoporous electrodes,
36,41,48
50,52,56
we ra
n bulk elec-
trolysis at −0.90
V
NHE
, within the kinetic regime, while varying
the concentration of HCO
3
. Acknowledging that the local con-
centration of HCO
3
at the electrode interface could differ from
the bulk due to the diffusion limitations through the m
olecular
film, we have assumed that the local concentration of HCO
3
is
linearly correlated with the bulk concentration.
59
The plot of
log(
j
CO
) against log[HCO
3
] exhibits a slope of ca. 1.0 for
Ag
-
1
and
Ag
-
2
, and 0 for bare Ag, indicating an
approximate first
order in [HCO
3
] for
Ag
-
1
and
Ag
-
2
and a zeroth order for Ag
(Fig
ure
3b). The data available point to a rate determining step
involving a proton transfer from HCO
3
to
Ag
-
1
or
Ag
-
2
, dis-
tinct from bare Ag.
Site Poisoning and the Diffusion
Layer Suppress HER
. To
better understand the mechanism of HER suppression in the
presence of the organic film, linear sweep voltammetry (LSV)
measurements were recorded in an N
2
-
saturated 0.1 M KHCO
3
electrolyte (Fig
ure
S20 and S21). Results obtained on ba
re Ag,
Ag
-
1
and
Ag
-
2
show that the onset potential for HER is shifted
up to 400 mV cathodically in the presence of the organic film.
Ag
-
1
displays the lowest HER activity, with current densities
15 times lower than Ag at −1.10 V. The possibility of mass
tr
ansfer limitations induced by the organic film were studied us-
ing a rotating disk electrode (RDE). On the bare Ag electrode,
the HER current density increased from −2.2 mA/cm
2
to −5.6
mA/cm
2
upon increasing the rotation speed from 500 rpm to
5000 rpm, impl
ying convection mass transport limitation (Fig-
ure
S22 and S23). In contrast, the HER current density remained
constant for
Ag
-
1
and
Ag
-
2
, at any rotation speed in this same
range, ruling out any convection mass transfer limitation. How-
ever, the HER activit
ies for
Ag
-
1
and
Ag
-
2
are 10 to 25 times
lower than bare Ag, which could reflect a diffusion mass
transport limitation, likely a consequence of the presence of the
hydrophobic organic film.
The influence of the partial pressure of CO
2
was also studied to
c
ompare bare Ag with
Ag
-
1
, and
Ag
-
2
, to determine whether a
mass transport limitation due to the organic film might be oper-
ative (Fig
ure
S24 and S25). The partial pressure of CO
2
(
p
CO
2
)
was varied between 0.2 atm and 1.0 atm, keeping a constant
flow of 5 sc
cm by using N
2
as a balance gas. At −0.9
V
NHE
and
−1.1
V
NHE
,
the plot of
j
CO
as a function of
p
CO
2
gave identical
traces for Ag and
Ag
-
1
, suggesting that the organic film on
Ag
-
1
does not limit the concentration of CO
2
close to the electrode
surface. In contrast, the trace obtained for
Ag
-
2
is significantly
lower than for bare Ag and
Ag
-
1
, indicating that CO
2
diffusion
5
is limited in the presence of a thicker film. These results explain
the lower activity of
Ag
-
2
for CO
production in bulk electroly-
sis experiments compared to
Ag
-
1
. It is also worth noting that
even at
p
CO
2
as low as 0.2 atm no traces of H
2
were detected
with
Ag
-
1
and
Ag
-
2
, highlighting the ability of these organics
films for suppressing HER.
Related to th
e aforementioned experiments performed to
demonstrate that CO
2
RR is under a kinetic regime at −0.9 V,
the correlation between HER current densities and ECSA for
different electrodeposition times of
2
-
Cl
were measured at −1.1
V, i.e. under the optimum condi
tions for CO
2
RR activity. The
results show that the ECSA on the Ag electrode decreases ex-
ponentially with the electrodeposition time, reaching a plateau
where only 40% of the area remains available (Fig
ure
S29a).
Accordingly, the HER current density also d
ecreases exponen-
tially with the electrodeposition time (Fig
ure
S30). The correla-
tion between the ESCA and the HER current density is shown
in Fig
ure
4 (black trace, ■, Table S10), where, in contrast to the
plot shown in Figure S19, three regimes can be obs
erved. First,
the HER current density decreases linearly with the ECSA, until
roughly 40% of the electrode is covered by the organic film.
The current density then decreases more sharply, by ~ 50%,
over a very small change in the ECSA. Finally, the value o
f the
HER current density changes very slowly, even as the ECSA is
further depleted to ca. 60% of the total.
Figure 4
. Normalized ECSA against normalized partial current
density for HER (■) and CO (●) plot recorded at −1.1 V. For each
experiment, the fil
m was electrodeposited over diff
erent time peri-
ods (10 ms
to 1 min
) at −0.70 V from a 10 mM solution of
2
-
Cl
dissolved in 0.1 M KHCO
3
.
HER experiments were performed un-
der
an
N
2
atmosphere (see the Supporting Info for more details).
The dotted lines serve only as a visual guide. Each data point was
recorded at least two times to ensure reproducibility.
Similarly, the correlation between
j
CO
and the ECSA was inves-
tigated with
Ag
-
2
, using similar reaction conditions as for HER
studies but under a CO
2
atmosphere (Fig
ure
4, blue trace, ●,
Table S11, Fig
ure
S31 and S32). As expected, the ECSA de-
creases exponentially with the electrodeposition time, indicat-
ing that the organic layer is
produced independently whether N
2
or CO
2
is used. The plot of
j
CO
as a function of the ECSA indi-
cates a linear decrease of CO production as a function of the
active area, until reaching a plateau at an activity that is ~20%
less active than that of bare A
g. Prolonged electrodeposition
times reduce only slightly the ECSA without altering the
CO
2
RR activity.
The data shown here demonstrate that the film selectively sup-
presses HER over CO
2
RR at potentials from −1.0 to −1.2 V.
The data are consistent with a s
cenario in which the rate of dif-
fusion of protons through a hydrophobic film is dramatically
attenuated by comparison to CO
2
.
58
It is also possible that HER
and CO
2
RR occur at distinct active sites
60
on the silver elec-
trode, with the hydrophobic film preferentially electrodeposit-
ing onto and inhibiting HER sites. Our available data do not
unambiguously distinguish between such scenarios.
Overall, in the absence of a molecular additive, bare silver ele
c-
trodes produce a mixture of primarily CO and H
2
(75% CO vs
20% H
2
; ~ 5% HCOOH at −0.99 V) during CO
2
RR. The Tafel
analysis, and the zeroth and first order dependence in HCO
3
and CO
2
, respectively, are consistent with the first ET step to
CO
2
being the RD
S, as previously observed
.
36,38,41,48
55
Fig
ure
5
shows a working mechanistic model for inhibition of HER on
the molecularly coated Ag electrodes studied here. At potentials
less negative than −1.0 V, i.e. in the kinetic regime, Tafel anal-
ysis suggests imp
roved kinetics for CO
2
RR with
Ag
-
1
and
Ag
-
2
, indicating an enhanced rate for the first ET step that activates
CO
2
. This could be a consequence of interactions (whether di-
rect, or from secondary sphere interactions, for example via wa-
ter molecules) of the f
ilm with CO
2
as it is activated and reduced
on the Ag surface, analogous to previously proposed activation
of CO at a Cu
-
film interfac
e.
31
Electrokinetic data point to the
involvement of a proton transfer from HCO
3
within the RDS
(Fig
ure
5b). At potential
s between −1.0 V to −1.2 V, the hydro-
phobic layer formed from the electrodeposition of the additive
limits the diffusion of proton carriers to the electrode surface,
but not CO
2
. Significantly, the production of H
2
is completely
suppressed, with CO being s
electively generated during
CO
2
RR. At potentials more reducing than −1.2 V, the diffusion
of CO
2
becomes limited, creating a depletion of the *CO
2
•−
in-
termediates (Fig
ure
5c). Consequently, the production of H
2
be-
comes more favorable. This working hypothes
is warrants test-
ing via future operando spectroscopic measurements that will
aid in further characterization of the organic
-
silver interface.
CONCLUSIONS
We have reported on a simple and efficient method to fully sup-
press HER and thereby facilitate CO
2
RR
on Ag electrodes using
organic films. The films are conveniently generated at the sur-
face of the silver electrode during bulk electrolysis by
in situ
reduction of pyridinium
-
based organic additives. Electrokinetic
studies demonstrate the impact of the film on the mechanism of
CO
2
reduction: a proton transfer from HCO
3
is involved in the
RDS, which is in contrast to the more typically observed elec-
tron
transfer as RDS on bare Ag surfaces. Moreover, slow dif-
fusion of proton carriers through the hydrophobic layer is re-
sponsible for a dramatic decrease in HER. Consequently, CO
2
can be selectively reduced to CO, in an aqueous electrolyte,
with FEs >99%. Str
uctural changes in the additive used result
in different film morphologies with distinct consequences on
electrocatalysis. For example, the thick film derived from
2
-
Cl
efficiently suppresses HER but also significantly reduces
CO
2
RR. This detrimental lower
ing in partial current density for
CO
2
RR from
2
-
Cl
is eliminated when the partially covered elec-
trode derived from
1
-
Br
2
is used instead. The differences ob-
served between
2
-
Cl
and
1
-
Br
2
highlight the potential for ra-
tionally tuning electrocatalysis by func
tionalizing the electrode
with tailored organic additives.
6
Figure
5.
Our working hypothesis to rationalize CO and H
2
evolution on a polycrystalline silver electrode with the electrodeposited organic
film. The Ag e
lectrode and organic film are depicted in gray and orange, respectively. The rate constants
k
H+
and k
CO2
correspond to the rate
of diffusion of proton carriers and CO
2
, respectively, through the organic film to the electrode interface. The rate constants
k
HER
and k
CO2RR
correspond to the rate of HER and CO
2
RR respectively.
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is
available free of charge on the ACS
Publications website.
Experimental details for bulk electrolysis experiments, surface
characterization, and additional electrokinetic data
(PDF)
AUTHOR INFORMATION
Corresponding Author
s
*
Email: agapie@caltech.edu
*
Em
ail:
jpeters@caltech.edu
Author Contributions
These authors contributed equally
to this work
.
Notes
The authors declare no competing financial interest
.
ACKNOWLEDGMENT
S
NMR, AFM, and XPS, SEM and EDX measurements were col-
lected at the NMR
Facility (Division of Chemistry and Chemical
Engineering), the Molecular Materials Research Center (Beckman
Institute) and the Analytic Facilities (Division of Geological and
Planetary Sciences) of the California Institute of Technology, re-
spectively. This
material is based upon work performed by the Joint
Center for Artificial Photosynthesis, a DOE Energy Innovation
Hub, supported through the Office of Science of the U.S. Depart-
ment of Energy under Award Number DE
-
SC0004993. A.T.
acknowledges Marie Skłodow
ska
-
Curie Fellowship H2020
-
MSCA
-
IF
-
2017 (793471). J.C.P also acknowledges additional sup-
port from the Resnick Sustainability Institute at Caltech.
REFERENCES
(1)
Bushuyev, O. S.; De Luna, P.; Dinh, C. T.; Tao, L.; Saur, G.;
van de Lagemaat, J.; Kelley, S
. O.; Sargent, E. H. What Should
We Make with CO
2
and How Can We Make It?
Joule
2018
,
2
,
825
832.
(2)
Chu, S.; Cui, Y.; Liu, N. The Path towards Sustainable Energy.
Nat. Mater.
2017
,
16
, 16
22.
(3)
Martín, A. J.; Larrazábal, G. O.; Pérez
-
Ramírez, J. Towa
rds
Sustainable Fuels and Chemicals through the Electrochemical
Reduction of CO
2
: Lessons from Water Electrolysis.
Green
Chem.
2015
,
17
, 5114
5130.
(4)
Appel, A. M.; Bercaw, J. E.; Bocarsly, A. B.; Dobbek, H.; Du-
Bois, D. L.; Dupuis, M.; Ferry, J. G.; Fuji
ta, E.; Hille, R.; Kenis,
P. J. A.; Kerfeld, C. A.; Morris, R. H.; Peden, C. H. F.; Portis,
A. R.; Ragsdale, S. W.; Rauchfuss, T. B.; Reek, J. N. H.; See-
feldt, L. C.; Thauer, R. K.; Waldrop, G. L. Frontiers, Opportu-
nities, and Challenges in Biochemical and
Chemical Catalysis
of CO
2
Fixation.
Chem. Rev.
2013
,
113
, 6621
6658.
(5)
Nitopi, S.; Bertheussen, E.; Scott, S. B.; Liu, X.; Engstfeld, A.
K.; Horch, S.; Seger, B.; Stephens, I. E. L.; Chan, K.; Hahn, C.;
Nørskov, J. K.; Jaramillo, T. F.; Chorkendorff, I
. Progress and
Perspectives of Electrochemical CO
2
Reduction on Copper in
Aqueous Electrolyte.
Chem. Rev.
2019
,
119
, 7610
7672.
(6)
Dinh, C.
-
T.; Burdyny, T.; Kibria, M. G.; Seifitokaldani, A.;
Gabardo, C. M.; García de Arquer, F. P.; Kiani, A.; Edwards, J
.
P.; De Luna, P.; Bushuyev, O. S.; Zou, C.; Quintero
-
Bermudez,
R.; Pang, Y.; Sinton, D.; Sargent, E. H. CO
2
Electroreduction to
Ethylene via Hydroxide
-
Mediated Copper Catalysis at an Ab-
rupt Interface.
Science
2018
,
360
, 783
787.
(7)
Yang, K. D.; Lee,
C. W.; Jin, K.; Im, S. W.; Nam, K. T. Current
Status and Bioinspired Perspective of Electrochemical Conver-
sion of CO
2
to a Long
-
Chain Hydrocarbon.
J. Phys. Chem. Lett.
2017
,
8
, 538
545.
(8)
Zhang, L.; Zhao, Z.
-
J.; Gong, J. Nanostructured Materials for
Het
erogeneous Electrocatalytic CO
2
Reduction and Their Re-
lated Reaction Mechanisms.
Angew. Chem. Int. Ed.
2017
,
56
,
11326
11353.
(9)
Hori, Y. Electrochemical CO
2
Reduction on Metal Electrodes.
In
Modern Aspects of Electrochemistry
; Vayenas, C. G.,
White,
R. E., Gamboa
-
Aldeco, M. E., Eds.; Springer: New York, 2008;
Vol. 42, pp 89
189.
(10)
Seh, Z. W.; Kibsgaard, J.; Dickens, C. F.; Chorkendorff, I.; Nør-
skov, J. K.; Jaramillo, T. F. Combining Theory and Experiment
in Electrocatalysis: Insights into M
aterials Design.
Science
2017
,
355
, eaad4998.
(11)
Mistry, H.; Varela, A. S.; Kühl, S.; Strasser, P.; Cuenya, B. R.
Nanostructured Electrocatalysts with Tunable Activity and Se-
lectivity.
Nat. Rev. Mater.
2016
,
1
, 16009.
H
CO
2
H
+
Carrier
2
H
CO
2
CO
k
CO2RR
k
H
Voltage < -1.0 V
Kinetic regime:
-
k
H
,
k
CO2
>>
k
HER
,
k
CO2RR
-
RDS: PT or PCET
Diffusion limitation of protons
-
k
CO2
>>
k
H
>
k
CO2RR
>>
k
HER
-
Higher CO
2
RR rate, HER suppressed
2e
-
k
HER
H
2
2e
-
H
2
O
Voltage > -1.2 V
Depletion of *CO
2
:
-
k
CO2RR
,
k
HER
>>
k
CO2
,
k
H
- HER predominant
H
CO
2
H
+
Carrier
2
H
CO
2
CO
k
CO2RR
k
CO2
k
H
2e
-
k
HER
2
H
H
2
2e
-
H
2
O
H
CO
2
H
+
Carrier
2
H
CO
2
CO
k
CO2RR
k
CO2
k
H
2e
-
H
2
O
k
CO2
2
H
(a)
(b)
(c)
-1.0 V < Voltage < -1.2 V
Kinetic regime (RDS: PT or PCET)
-
k
H
+
,
k
CO2
>>
k
HER
,
k
CO2RR
-
HER predominan
t
Diffusion limitation of protons
-
k
CO2
>>
k
H
+
>
k
CO2RR
>>
k
HE
R
-
Higher CO
2
RR rate, HER suppressed
Depletion of *CO2
•−
-
k
CO2RR
,
k
HE
R
>>
k
CO2
,
k
H
+
-
HER predominant