Impurity ion complexation enhances carbon dioxide reduction catalysis

Article abstract of DOI:10.1021/acscatal.5b00808

Herein, we show that group 11 CO₂ reduction catalysts are rapidly poisoned by progressive deposition of trace metal ion impurities present in high purity electrolytes. Metal impurity deposition was characterized by XPS and in situ stripping voltammetry and is coincident with loss of catalytic activity and selectivity for CO₂ reduction, favoring hydrogen evolution on poisoned surfaces. Metal deposition can be suppressed by complexing trace metal ion impurities with ethylenediaminetetraacetic acid or solid-supported iminodiacetate resins. Metal ion complexation allows for reproducible, sustained catalytic activity and selectivity for CO₂ reduction on Au, Ag, and Cu electrodes. Together, this study establishes the principal mode by which group 11 CO₂ reduction catalysts are poisoned and lays out a general approach for extending the lifetime of electrocatalysts subject to impurity metal deposition.

Full text of DOI:10.1021/acscatal.5b00808

Research Article
pubs.acs.org/acscatalysis
Impurity Ion Complexation Enhances Carbon Dioxide Reduction
Catalysis
Anna Wuttig and Yogesh Surendranath*
Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States
S
* Supporting Information
ABSTRACT: Herein, we show that group 11 CO2 reduction
catalysts are rapidly poisoned by progressive deposition of
trace metal ion impurities present in high purity electrolytes.
Metal impurity deposition was characterized by XPS and in
situ stripping voltammetry and is coincident with loss of
catalytic activity and selectivity for CO2 reduction, favoring
hydrogen evolution on poisoned surfaces. Metal deposition
can be suppressed by complexing trace metal ion impurities
with ethylenediaminetetraacetic acid or solid-supported
iminodiacetate resins. Metal ion complexation allows for
reproducible, sustained catalytic activity and selectivity for CO2
reduction on Au, Ag, and Cu electrodes. Together, this study establishes the principal mode by which group 11 CO2 reduction
catalysts are poisoned and lays out a general approach for extending the lifetime of electrocatalysts subject to impurity metal
deposition.
KEYWORDS: electrocatalysis, CO2 reduction, surface poisoning, metal impurities, energy storage
control over product selectivity, which can be easily
compromised by impurities that interact with or irreversibly
alter the surface.
1. INTRODUCTION
Catalytic reactions are highly sensitive to the presence of
impurities. Trace constituents in the reaction medium can have
an outsized effect on reaction efficiency and selectivity by
interfering with the active species or mediating turnover by
themselves. For example, metal and halide contaminants both
promote and poison activity and selectivity in a myriad array of
heterogeneous reactions,1−4 have led to serendipitous catalyst
discovery in organic synthesis,5,6 and can dramatically alter
enzymatic activity.7−9
Trace impurities also play a dominant role in electrochemical
energy catalysis. For example, trace Co2+ impurities have led to
false positives in the development of molecular oxygen
evolution reaction (OER) catalysts,10 and trace iron11,12 and
platinum13 impurities have been shown to enhance the OER on
heterogeneous catalysts. Likewise, Pt surfaces are readily
poisoned for the oxygen reduction reaction (ORR) by various
impurities,14 whereas “metal-free” ORR activity has been
ascribed to trace metal impurities.15 Trace metal impurities
have also been shown to convolute the facet-dependence of
hydrogen evolution reaction (HER) activity on Au surfaces.16
The electrochemical reduction of CO2 to fuels is of growing
interest because it provides an attractive platform for the
storage of intermittent renewable energy in energy-dense
chemical bonds.17−19 Relative to other energy conversion
reactions such as ORR, OER, and HER, CO2 reduction (CDR)
is particularly sensitive to surface structure and composition
because of the myriad array of CDR products thermodynami-
cally accessible over a narrow potential range.20−22 The
development of practical CDR catalysts requires unparalleled
Group 11 metal surfaces are regarded as the most promising
heterogeneous catalysts for this reaction because they display
low to moderate overpotentials for CO production and, in the
case of copper, generate higher order products including
methane and ethylene.23−31 However, planar group 11 metal
surfaces are known to lose their catalytic activity and selectivity
for CO2 reduction over the time scale of minutes to hours
under steady-state electrolysis.22,32−37 For example, copper
surfaces lose one-half of their catalytic activity for methane
production within 20 min of polarization,33,35 and CO
production selectivity on Au32 and Ag34 decreases within
minutes of electrolysis. Despite posing a clear obstacle to the
practical implementation of CO2 reduction technologies, the
mechanistic basis for this activity loss remains poorly
understood.
Researchers have posited that the deactivation is due to the
deposition of trace metal ion22,35 or organic impurities35 in the
electrolyte, while others have suggested that this activity loss is
unavoidable, resulting from the accumulation of catalytic
intermediates that poison the surface over time.22,33,38−41 On
the basis of these hypotheses, strategies have been proposed to
prolong catalyst lifetimes, including periodic oxidative pulsing
of the electrode to remove adsorbed organics32−34 and long-
Received: April 18, 2015
Revised: May 21, 2015
Published: June 24, 2015
© 2015 American Chemical Society
4479
DOI: 10.1021/acscatal.5b00808
ACS Catal. 2015, 5, 4479−4484

ACS Catalysis
Research Article
Figure 1. Surface chemistry following CDR catalysis on group 11 metal surfaces. Narrow scan X-ray photoelectron spectra of Cu (A), Ag (B), and
Au (C) rotating electrodes before (black) catalysis and after 45 min of CDR catalysis in native Ci electrolyte (green) and EDTA-containing
electrolyte (red). Cyclic voltammetry traces of rotating electrodes at 2500 rpm in native Ci electrolyte (green) or EDTA-containing Ci electrolyte
(red) after (D) 120 min of CDR catalysis at −1.00 V on Cu; (E) 45 min of CDR catalysis at −0.90 V on Ag; and (F) 45 min of CDR catalysis at
−0.70 V on Au. In all cases, CV scans were recorded at 50 mV s−1 scan rate with a positive direction of scan.
term (>9 h) pre-electrolysis using a sacrificial electrode to
scavenge trace metal ion impurities in the electrolyte.35 The
former is impractical because it progressively alters the catalyst
surface structure,42 and pre-electrolysis has been shown, in
many cases, to be ineffective,43,44 irreproducible,33,34 and is
both energy- and time-intensive. We note that high surface area
electrodes37,38,44,45 should exhibit reduced sensitivity to trace
impurities, but these systems are difficult to probe mechanis-
tically due to their complex morphology and structure. Thus,
the systematic development of new CDR catalysts would
benefit from a clear understanding of the principal CDR
deactivation mechanism on these surfaces and the development
of simple strategies for sustaining catalyst activity and selectivity
over time.
Herein, we demonstrate, unambiguously, that metal ion
deposition is the principal mode of catalyst deactivation for Cu,
Ag, and Au metal surfaces and show, for the first time, that
catalyst deactivation can be entirely eliminated by irreversibly
coordinating trace metal ions in situ with ethylenediaminetetra-
acetic acid (EDTA) or ex situ with a solid-supported
iminodiacetate resin. The high binding affinity46 and rapid
complexation kinetics (k1 ≈ 1010 M−1 s−1)47−50 of EDTA and
solid-phase analogues make this a general strategy for
maintaining high CDR activity over time, regardless of structure
of the catalyst.
typical for selective CO2 reduction (−1.0 V, all potentials are
quoted versus the reversible hydrogen electrode, RHE) reveals
the buildup of zinc and lead impurities (Figures 1A and S1).
Similarly, prolonged electrolyses of silver (−0.9 V) and gold
(−0.7 V) rotating cone electrodes reveal the accumulation of
zinc and copper impurities (Figures 1B, C and S1). Previous
studies conducted at similar electrolysis potentials were unable
to detect trace metal impurity signals via XPS.25 Our use of a
rotating electrode serves to accelerate the rate of diffusion-
limited metal deposition, increasing the surface impurity
population (see SI calculation). These XPS results indicate
that all group 11 metal surfaces are subject to contamination via
impurity deposition even in cases where high purity electrolytes
are employed. These results are in line with the relatively high
thermodynamic M2+/0 redox potentials for Cu2+ (0.75 V) and
Pb2+ (0.28 V) and Zn2+ (−0.35 V)51 under CDR conditions
and the detection of μM concentrations of Zn2+ by inductively
coupled plasma mass spectrometry (ICP-MS). Given that the
thermodynamic potentials for metal deposition are close to or
positive of the thermodynamic potentials for CDR catalysis, we
expect that impurity deposition will also occur on high surface
area group 11 catalysts that operate at lower CDR over-
potentials, albeit at slower rates.
To gain further insight into changes in surface composition
during CDR catalysis, we characterized electrodes via cyclic
voltammetry (CV) immediately following (∼1 s time delay)
reductive polarization. The first CV sweeps recorded after short
(15 min) and prolonged (>45 min) CDR catalysis in native Ci
electrolyte reveal a progressive rise, Figure S2A, in broad
oxidative features at −0.22 and 0.26 V for Cu, Figure 1D
(green). Likewise, for Ag electrodes, we observe a continuous
rise, Figure S2B, in broad oxidative features at −0.25 and 0.50
V, Figure 1E (green). A similar experiment on Au reveals a rise,
Figure S2C, of features at 0.11, 0.50, and 0.85 V with the
appearance of distinct shoulders upon longer electrolysis,
Figure 1F (green). These features are not observed prior to
CDR catalysis (Figure S2A−C, black) or upon subsequent
cycling of the electrode after CDR catalysis. Together, the
2. RESULTS AND DISCUSSION
Group 11 Surfaces Accumulate Impurity Metal
Poisons During CO2 Reduction Catalysis. To characterize
the purity of group 11 surfaces following CDR catalysis, we
analyzed electrodes by X-ray photoelectron spectroscopy
(XPS) following electrolysis in unpurified CO2-saturated 0.1
M NaHCO3 (Ci) electrolyte (full experimental details,
including solution, electrode, and electrochemical cell prepara-
tion, are available in the SI). Despite using 18 MΩ cm water
and electrolyte salts of the highest purity (99.9999%, Aldrich
TraceSELECT) available commercially, XPS of copper rotating
disk electrodes following 45 min of electrolysis at potentials
4480
DOI: 10.1021/acscatal.5b00808
ACS Catal. 2015, 5, 4479−4484

ACS Catalysis
Research Article
observations suggest that these waves originate from irrever-
sible oxidative stripping (M0→ Mn+ + ne−) of impurity metals
electrodeposited on Cu, Ag, and Au surfaces during CDR
catalysis.
To assign these stripping waves to metal ion impurities, we
collected CV traces following CDR in Ci electrolyte, containing
50 μM of various M2+ salts of metals detected via XPS. This
intentional introduction of an impurity metal ion allows us to
probe the stripping behavior of a single M2+ candidate under
experimental conditions relevant to CDR catalysis. For Cu
electrodes polarized at CDR potentials in Zn2+-containing
electrolytes, the first CV scan following CDR catalysis reveals a
broad Zn stripping feature at −0.30 V, Figure S3A (red), in line
with one oxidative wave, at −0.22 V, Figure 1D (green),
observed for CDR performed in native Ci electrolyte. An
analogous experiment performed with Pb2+-containing electro-
lyte reveals oxidative features at 0.15 and 0.26 V (Figure S3A,
blue), attributed to stripping of bulk and underpotential
deposited (UPD) lead on Cu, as previously characterized in
acidic electrolytes.52 Consistent with the low concentrations of
Pb in native Ci electrolyte (Table S1), the stripping feature
observed in Figures 1D (green) matches the UPD stripping
potential of 0.26 V in Figure S3A (blue). Together, these
results indicate that Cu is susceptible to progressive fouling by
Zn and Pb deposition during CDR catalysis.
Similar experiments conducted with Ag and Au working
electrodes in Zn2+-containing electrolytes reveal Zn-stripping
features at −0.35 and −0.23 V for Ag (Figure S3B, red), and at
−0.10 and 0.12 V for Au (Figure S3C, red). In both cases, the
more positive peak is attributed to Zn UPD stripping
features,53−55 which match the stripping waves at −0.25 V on
Ag (Figure 1E, green) and 0.11 V on Au (Figure 1F, green)
observed following CDR in native Ci electrolyte. Analogous
experiments performed with Ag and Au electrodes in Cu2+-
containing electrolytes reveal broad stripping features at 0.50
and 0.60 V for Ag (Figure S3B, green) and 0.65 and 0.85 V for
Au (Figure S3C, green). These features are close to the
stripping waves observed at 0.50 V on Ag (Figure 1E, green),
and 0.56 and 0.85 V on Au (Figure 1F, green) following CDR
in native Ci electrolyte. Slight differences in potentials may be
due to alloying of codeposited Cu and Zn from native Ci
electrolytes,56 as opposed to the deposition of Cu alone from
Cu2+-enriched electrolytes. Stripping features for other
common first row transition metals, such as Fe and Ni,
would be expected to occur at values less positive than that of
Cu,51,57leading us to assign the aforementioned peaks to Cu
stripping. Together with XPS data, analysis of stripping
voltammograms demonstrates that group 11 electrodes are
susceptible to fouling by Pb, Zn, and Cu deposition during
CDR catalysis.
Figure 2. Faradaic efficiencies for CDR and HER product formation
on gold foil in electrolytes of varying purity. Activity of Au for CO (A)
and H2 (B) formation at −0.70 V in native Ci electrolyte (black
squares), Ci electrolyte containing 3.4 μM EDTA (red circles), and
Chelex-treated Ci electrolyte (blue triangles).
for CO production is accompanied by a rise, from ∼23% to
∼77%, in Faradaic efficiency for the hydrogen evolution
reaction (HER). The observed CO and H2 account for
∼90% of the steady state current, and no other gas phase
products (see SI for experimental details) were detected,
suggesting that solution-phase products, such as formate,23 may
account for the balance. Notwithstanding, CDR selectivity is
progressively eroded over time for electrolysis performed in
native Ci electrolyte. In order to suppress the observed
deactivation, we employed established pre-electrolysis meth-
ods35 to clean the electrolyte. In our hands, this method proved
ineffective, leading to similar deactivation in CO production
Faradaic efficiency from ∼40% to ∼16% (Figure S4).
Seeking an alternative to pre-electrolysis methods, we
envisioned that introducing EDTA to Ci electrolytes would
prevent metal deposition, thereby enhancing long-term
catalysis. CV scans of a Au electrode recorded immediately
following 45 min of CDR catalysis in Ci electrolyte containing
3.4 μM EDTA (Figure 1F, red) do not exhibit the Zn and Cu
stripping features observed for the same experiment conducted
in the absence of EDTA (Figure 1F, green). We also do not
observe Zn and Cu features by XPS after 45 min of CDR
catalysis in EDTA-containing electrolyte (Figure 1C, red),
indicating an impurity-free surface. A broad oxidative feature is
observed beginning at 0.60 V (Figure 1F, red), which we
attribute to the oxidation of surface-adsorbed CO formed
during electrolysis.58 We speculate that in untreated electro-
lytes, this weak feature is masked by the copper stripping wave.
In line with the absence of metal fouling observed by CV, we
find that ex situ and in situ metal ion chelation significantly
attenuates catalyst deactivation. Au electrodes display sustained
CDR activity when operated in Ci electrolyte containing 3.4
μM EDTA (Figure 2A, red circles). The concentration of
EDTA was chosen on the basis of the M2+ impurity
concentration measured by ICP-MS of the native Ci electrolyte.
The initial Faradaic efficiency for CO production is ∼80% with
only a slight decay to ∼70% after 2 h of steady-state electrolysis.
Further enhancement in catalytic activity is afforded by
purification of Ci electrolyte with solid-supported iminodiace-
tate resin (Chelex) prior to electrolysis; CO production on Au
is sustained at ∼80% over 2 h (Figure 2A, blue triangles). In
addition, the use of EDTA-containing or Chelex-treated Ci
electrolyte enables greater reproducibility in product selectivity
and catalyst activity, as evidenced by tighter error bars
compared to the data obtained from native Ci electrolyte.
Furthermore, the sustained CDR selectivities for Au are
reflected in sustained partial current densities for CO (Figures
S5A) and H2 production (Figure S5B), indicating that impurity
Chelation Inhibits Impurity Deposition and Enables
Sustained CO2 Reduction Catalysis. Impurity metal
deposition leads to dramatic changes in CDR selectivity on
group 11 catalysts. At −0.70 V, CO production on polycrystal-
line Au foil electrodes operated in native Ci electrolytes
commences with ∼60% Faradaic efficiency (FE) but declines to
∼10% over the course of 2 h (Figure 2A, black squares, error
bars shown in Figures 2−4 are the standard deviation of three
independent measurements), which is consistent with literature
reports.32,34 The large error bars observed for electrolyses
performed in native Ci electrolyte reveal a high degree of run-
to-run variability, consistent with trace metal impurities
strongly influencing catalytic efficiency. Catalyst deactivation
4481
DOI: 10.1021/acscatal.5b00808
ACS Catal. 2015, 5, 4479−4484

ACS Catalysis
Research Article
chelation provides for sustained intrinsic rates of product
formation. Upon treatment of the electrolyte with either EDTA
or Chelex, the current densities for CO production remains
high (1.2 mA cm−2) relative to HER (0.1 mA cm−2). Taken
together with the CV and XPS data highlighted above, these
results suggest that impurity metal deposition over the course
of CDR catalysis is the principal source of electrode
deactivation on Au.
Silver electrodes display similar deactivation profiles when
operated in native Ci electrolyte. At −0.90 V, CO production
on Ag in native Ci electrolyte commences with ∼50% Faradaic
efficiency but declines over the course of 2 h to ∼33% (Figure
3A, black squares). This deactivation is accompanied by a
Figure 4. Faradaic efficiencies for CDR and HER product formation
on copper foil in electrolytes of varying purity. Activity of Cu for CH4
(A), CO (B), C2H4 (C) and H2 (D) formation at −1.00 V in native Ci
electrolyte (black squares), Ci electrolyte containing 3.4 μM EDTA
(red circles), and Chelex-treated Ci electrolyte (blue triangles).
Figure 3. Faradaic efficiencies for CDR and HER product formation
on silver foil in electrolytes of varying purity. Activity of Ag for CO (A)
and H2 (B) formation at −0.90 V in native Ci electrolyte (black
squares), Ci electrolyte containing 3.4 μM EDTA (red circles), and
Chelex-treated Ci electrolyte (blue triangles).
similar potentials.35 Over the same time period, the minority
production of CO, ∼5%, remains constant within error over the
2 h electrolysis period, Figure 4B. We only examine gas-phase
products in this study, and we note that copper also mediates
the production of liquid products such as ethanol, propanol,
and formate23 under these conditions.
corresponding rise from ∼38 to ∼52% in current going to the
HER over the same period. Similar to the results observed on
Au, following 45 min of CDR catalysis in Ci electrolyte
containing 3.4 μM EDTA, stripping waves are not observed by
CV (Figure 1E, red), and XPS spectra show no Zn or Cu peaks
(Figure 1B, red), indicating a metal-impurity-free surface. The
trailing cathodic feature ending at −0.15 V is attributed to
residual catalytic current and is not observed in subsequent
scans.
Consistent with the CV data, EDTA enables sustained and
improved CO production, commencing at ∼58% and declining
only slightly to 52% after 2 h of electrolysis (Figure 3B, red
circles). As for Au, further enhancements in long-term catalytic
activity are observed for Ci electrolytes purified by treatment
with Chelex prior to electrolysis: CO production on Ag is
sustained at ∼60% over the entire 2 h period (Figure 3A, blue
triangles). Consistent with the retention in FE, the total
currents for product formation are preserved at 0.65 mA cm−2
for CO (Figure S6A) and at 0.4 mA cm−2 for H2 formation
(Figure S6B). The error bars for the Ag data are large possibly
due to variance in the sulfuric acid etching treatment prior to
each run. As for Au electrodes, these results suggest that
impurity metal deposition during CDR catalysts is the principal
source of deactivation of Ag electrodes.
Copper electrodes display complex deactivation profiles
when operated at −1.00 V in native Ci electrolyte. At early
times, the principal gas-phase CDR product is C2H4 with ∼12%
Faradaic efficiency, Figure 4C. However, after 2 h of
electrolysis, this C2 product is not observed at all. Similarly,
CH4 production commences at ∼7% Faradaic efficiency but is
not observed after 2 h of electrolysis, Figure 4A. The decline in
selectivity for higher-order CDR products is accompanied by a
rise in H2 production yield from ∼65% to ∼85%, Figure 4D.
These results are consistent with previous reports of Cu at
The data obtained in the case of Cu differ from those
collected when using Ag and Au; we do not observe enhanced
long-term CDR activity on Cu electrodes operated in EDTA-
containing Ci electrolyte. The deactivation profiles for C2H4
and CH4 production as well as the rise in H2 production are
similar to that observed for native Ci electrolyte (Figure
4A,C,D, red circles). Surprisingly, the introduction of EDTA to
the electrolyte promotes a decline in CO production Faradaic
efficiency from ∼5% to ∼0% over the course of 2 h, Figure 4B.
Although these observations seem to imply that EDTA is
ineffective at preventing metal deposition on Cu electrode
surfaces, voltammetry scans recorded immediately following 2 h
of electrolysis in the presence of EDTA do not display any
stripping waves (Figure 1D, red), and XPS spectra recorded
following electrolysis show no Zn or Pb peaks (Figure 1A, red),
suggesting a metal-impurity-free surface. The trailing cathodic
feature ending at 0.00 V is attributed to residual catalytic
current and is not observed in subsequent scans. On the basis
of these results, we propose that the observed inefficacy of
EDTA in preventing CDR deactivation on Cu may result from
direct interaction of the chelator with the surface and/or
chelator-induced surface restructuring. Interestingly, the rise in
CO production observed for Cu operated in native Ci
electrolyte, which is not seen in EDTA-containing Ci
electrolyte, suggests that Pb or Zn metal deposition, observed
by XPS (Figure 1A) and CV (Figure 1D), may promote release
of CO intermediates from Cu surfaces.
The observed deactivation of Cu electrodes can, however, be
eliminated by pretreatment of Ci electrolyte with Chelex. The
principal CDR product is CH4 with ∼20% Faradaic efficiency
sustained over 2 h (Figure 4A, blue triangles). Similarly, C2H4
and CO production is sustained at ∼5% (Figure 4C, blue
4482
DOI: 10.1021/acscatal.5b00808
ACS Catal. 2015, 5, 4479−4484

ACS Catalysis
Research Article
(3) Koel, B. E.; Kim, J. In Handbook of Heterogeneous Catalysis; Ertl,
triangles) and ∼2% (Figure 4B, blue triangles), respectively,
over the same time period. It appears that removal of metal ion
impurities from the electrolyte alters product selectivity even at
early reaction times, as shown by the substantial enhancement
in methane yield and corresponding decrease in C2H4 Faradaic
efficiency.
G., H. Knozinger, F. Schuth, J. W, Eds.; Wiley-VCH: Weinheim, 2008;
̈
̈
Vol. 5, pp 1593−1624.
(4) Butt, J.; Peterson, E. E. Activation, Deactivation, and Poisoning of
Catalysts; Academic Press: San Diego, 1988.
(5) Buchwald, S. L.; Bolm, C. Angew. Chem., Int. Ed. 2009, 48, 5586−
5587.
Importantly, the sustained CDR selectivities described above
for Cu are reflected in sustained partial current densities for
CO2 reduction products, including CH4 (Figure S7A), CO
(Figure S7B), C2H4 (Figure S7C), and H2 (Figure S7D). These
results indicate that impurity chelation provides for sustained
intrinsic rates of CDR product formation. Thus, the methods
described here may prove particularly valuable for detailed
studies of CDR that require temporal fidelity of the catalyst
surface.
(6) Crabtree, R. H. Chem. Rev. 2012, 112, 1536−1554.
(7) Klug, D.; Rabani, J.; Fridovich, I. J. Biol. Chem. 1972, 247, 4839−
4842.
(8) Peracchi, A. Trends Biochem. Sci. 2001, 26, 497−503.
(9) Lee, S. H.; Ha, S. H.; Lee, S. B.; Koo, Y.-M. Biotechnol. Lett. 2006,
28, 1335−1339.
(10) Ullman, A. M.; Liu, Y.; Huynh, M.; Bediako, D. K.; Wang, H.;
Anderson, B. L.; Powers, D. C.; Breen, J. J.; Abruna, H. D.; Nocera, D.
̃
G. J. Am. Chem. Soc. 2014, 136, 17681−17688.
(11) Corrigan, D. A. J. Electrochem. Soc. 1987, 134, 377−384.
(12) Trotochaud, L.; Young, S. L.; Ranney, J. K.; Boettcher, S. W. J.
Am. Chem. Soc. 2014, 136, 6744−6753.
3. CONCLUSIONS
In summary, we have shown that the commonly observed
activity loss of group 11 catalysts for CDR arises principally
from impurity metal deposition. Additionally, we have shown
that impurity metal deposition can be reduced or eliminated by
chelation with EDTA or electrolyte purification via treatment
with a solid-supported metal-chelating resin. The results
highlight the extremely high sensitivity of CDR to metal ion
impurities and the value of post-electrolysis CVs in diagnosing
impurity-assisted fouling. The simple approach described here
establishes an accessible and general protocol for evaluating
CO2 reduction and other fuel-forming electrocatalysts under
conditions that preserve the fidelity of the native surface.
(13) Bockris, J. O.; McHardy, J. J. Electrochem. Soc. 1973, 120, 61−66.
(14) Paulus, U. A.; Schmidt, T. J.; Gasteiger, H. A. Handbook of Fuel
Cells; John Wiley & Sons: Chichester, 2010.
(15) Wang, L.; Ambrosi, A.; Pumera, M. Angew. Chem., Int. Ed. 2013,
52, 13818−13821.
(16) Perez, J.; Gonzalez, E. R.; Villullas, H. M. J. Phys. Chem. B 1998,
102, 10931−10935.
́
(17) Kondratenko, E. V.; Mul, G.; Baltrusaitis, J.; Larrazabal, G. O.;
́
Perez-Ramírez, J. Energy Environ. Sci. 2013, 6, 3112−3135.
(18) Olah, G. A.; Prakash, G. K. S.; Goeppert, A. J. Am. Chem. Soc.
2011, 133, 12881−12898.
(19) Whipple, D. T.; Kenis, P. J. A. J. Phys. Chem. Lett. 2010, 1,
3451−3458.
(20) Rosenthal, J. In Progress in Inorganic Chemistry; Karlin, K. D.,
Ed.; John Wiley & Sons: Hoboken, 2014; pp 299−338.
(21) Benson, E. E.; Kubiak, C. P.; Sathrum, A. J.; Smieja, J. M. Chem.
Soc. Rev. 2009, 38, 89−99.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.5b00808.
■
S
(22) Hori, Y. In Modern Aspects of Electrochemistry; Vayenas, C.,
White, R., Gamboa-Aldeco, M., Eds.; Springer, New York, 2008; Vol.
42, pp 89−189.
Full experimental details, survey XPS spectra, stripping
voltammetry experiments in the presence of various Mn+
salts, and intrinsic rates of CO2 reduction product
formation in electrolytes of different purities (PDF)
(23) Hori, Y.; Wakebe, H.; Tsukamoto, T.; Koga, O. Electrochim. Acta
1994, 39, 1833−1839.
(24) Kuhl, K. P.; Cave, E. R.; Abram, D. N.; Jaramillo, T. F. Energy
Environ. Sci. 2012, 5, 7050−7059.
(25) Hatsukade, T.; Kuhl, K. P.; Cave, E. R.; Abram, D. N.; Jaramillo,
T. F. Phys. Chem. Chem. Phys. 2014, 16, 13814−13819.
(26) Sreekanth, N.; Phani, K. L. Chem. Commun. 2014, 50, 11143−
11146.
(27) Peterson, A. A.; Nørskov, J. K. J. Phys. Chem. Lett. 2012, 3, 251−
258.
AUTHOR INFORMATION
Corresponding Author
*E-mail: yogi@mit.edu.
Notes
■
(28) Calle-Vallejo, F.; Koper, M. T. M. Angew. Chem., Int. Ed. 2013,
52, 7282−7285.
The authors declare no competing financial interest.
(29) Montoya, J. H.; Peterson, A. A.; Nørskov, J. K. ChemCatChem
2013, 5, 737−742.
ACKNOWLEDGMENTS
■
The authors gratefully acknowledge Melissa Zastrow for
assistance with purification protocols. This research was
supported through a Research Agreement with Saudi Aramco,
a Founding Member of the MIT Energy Initiative, by the Air
Force Office of Scientific Research under award FA9550-15-1-
0135, and by the MIT Department of Chemistry through junior
faculty funds for Y.S. A.W. is supported by a predoctoral
fellowship from the National Science Foundation. This work
made use of the MRSEC Shared Experimental Facilities at
MIT, which is supported in part by the NSF under award
DMR-0819762.
(30) Nie, X.; Esopi, M. R.; Janik, M. J.; Asthagiri, A. Angew. Chem.,
Int. Ed. 2013, 52, 2459−2462.
(31) Peterson, A. A.; Abild-Pedersen, F.; Studt, F.; Rossmeisl, J.;
Norskov, J. K. Energy Environ. Sci. 2010, 3, 1311−1315.
(32) Kedzierzawski, P.; Augustynski, J. J. Electrochem. Soc. 1994, 141,
L58−L60.
(33) Jermann, B.; Augustynski, J. Electrochim. Acta 1994, 39, 1891−
1896.
(34) Kostecki, R.; Augustynski, J. Berichte der Bunsengesellschaft fur
̈
Phys. Chemie 1994, 98, 1510−1515.
(35) Hori, Y.; Konishi, H.; Futamura, T.; Murata, A.; Koga, O.;
Sakurai, H.; Oguma, K. Electrochim. Acta 2005, 50, 5354−5369.
(36) Yano, H.; Shirai, F.; Nakayama, M.; Ogura, K. J. Electroanal.
Chem. 2002, 533, 113−118.
REFERENCES
■
(37) Chen, Y.; Li, C. W.; Kanan, M. W. J. Am. Chem. Soc. 2012, 134,
19969−19972.
(1) Twigg, M. V.; Spencer, M. S. Top. Catal. 2003, 22, 191−203.
(2) Besson, M.; Gallezot, P. Catal. Today 2003, 81, 547−559.
4483
DOI: 10.1021/acscatal.5b00808
ACS Catal. 2015, 5, 4479−4484

ACS Catalysis
Research Article
(38) Kas, R.; Kortlever, R.; Yılmaz, H.; Koper, M. T. M.; Mul, G.
ChemElectroChem 2015, 2, 354−358.
(39) DeWulf, D. W.; Jin, T.; Bard, A. J. J. Electrochem. Soc. 1989, 136,
1686−1691.
(40) Wasmus, S.; Cattaneo, E.; Vielstich, W. Electrochim. Acta 1990,
35, 771−775.
(41) Kyriacou, G.; Anagnostopoulos, A. J. Electroanal. Chem. 1992,
322, 233−246.
(42) Hoogvliet, J. C.; Dijksma, M.; Kamp, B.; van Bennekom, W. P.
Anal. Chem. 2000, 72, 2016−2021.
(43) Batista, E. A.; Temperini, M. L. A. J. Electroanal. Chem. 2009,
629, 158−163.
(44) Lu, Q.; Rosen, J.; Zhou, Y.; Hutchings, G. S.; Kimmel, Y. C.;
Chen, J. G.; Jiao, F. Nat. Commun. 2014, 5, Article no. 3242.
(45) Li, C. W.; Kanan, M. W. J. Am. Chem. Soc. 2012, 134, 7231−
7234.
(46) Furia, T. E. In CRC Handbook of Food Additives; Furia, T. E.,
Ed.; CRC Press: Boca Raton, 1972; pp 271−294.
(47) Eigen, M.; Wilkins, R. G. Mechanisms of Inorganic Reactions;
Kleinberg, J., Murmann, R. K., Fraser, R. T. M., Bauman, J., Eds.;
Advances in Chemistry; American Chemical Society: Washington,
DC., 1965; Vol. 49, pp 55−80.
(48) Bril, K.; Bril, S.; Krunhotz, P. J. Phys. Chem. 1956, 60, 251−252.
(49) Bril, K.; Bril, S.; Krumholz, P. J. Phys. Chem. 1955, 59, 596−600.
(50) Tanaka, N.; Osawa, H.; Kamada, M. Bull. Chem. Soc. Jpn. 1963,
36, 530−534.
(51) Vanysek, P. CRC Handbook of Chemistry and Physics, 95th ed.;
́
Haynes, W. M., Ed.; CRC Press: Boca Raton, 2014.
́
(52) Brisard, G. M.; Zenati, E.; Gasteiger, H. A.; Markovic, N. M.;
Ross, P. N. In Solid-Liquid Electrochemical Interfaces; Jerkiewicz, G.,
Soriaga, M. P., Uosaki, K., Wieckowski, A., Eds.; ACS Symposium
Series; American Chemical Society: Washington, DC., 1997; Vol. 656,
pp 142−155.
(53) Takahashi, S.; Hasebe, K.; Aramata, A. Electrochem. Commun.
1999, 1, 301−304.
(54) Adzic, G. J. Electrochem. Soc. 1981, 128, 1691−1697.
(55) Chu, M. G. J. Electrochem. Soc. 1981, 128, 2281−2286.
(56) Fonticelli, M. H.; Posadas, D.; Tucceri, R. I. J. Electroanal. Chem.
2004, 565, 359−366.
(57) Gerischer, H.; Kolb, D. M.; Przasnyski, M. Surf. Sci. 1974, 43,
662−666.
(58) Kita, H.; Nakajima, H.; Hayashi, K. J. Electroanal. Chem.
Interfacial Electrochem. 1985, 190, 141−156.
4484
DOI: 10.1021/acscatal.5b00808
ACS Catal. 2015, 5, 4479−4484