Molecular catalysis of h2 Evolution: Diagnosing heterolytic versus homolytic pathways

Article abstract of DOI:10.1021/ja505845t

Molecular catalysis of H₂ production from the electrochemical reduction of acids by transition-metal complexes is one of the key issues of modern energy challenges. The question of whether it proceeds heterolytically (through reaction of an initially formed metal hydride with the acid) or homolytically (through symmetrical coupling of two molecules of hydride) has to date received inconclusive answers for a quite simple reason: the theoretical bases for criteria allowing the distinction between homolytic and heterolytic pathways in nondestructive methods such as cyclic voltammetry have been lacking heretofore. They are provided here, allowing the distinction between the two pathways. The theoretical predictions and the diagnosing strategy are illustrated by catalysis of the reduction of phenol, acetic acid, and protonated triethylamine by electrogenerated iron(0) tetraphenylporphyrin. Rather than being an intrinsic property of the catalytic system, the occurrence of either a heterolytic or a homolytic pathway results from their competition as a function of the concentrations of acid and catalyst and the rate constants for hydride formation and H₂ evolution by hydride protonation or dimerization.

Full text of DOI:10.1021/ja505845t

Article
pubs.acs.org/JACS
Molecular Catalysis of H2 Evolution: Diagnosing Heterolytic versus
Homolytic Pathways
Cyrille Costentin,* Hachem Dridi, and Jean-Michel Saveant*
́
́ ́ ́
Universite Paris Diderot, Sorbonne Paris Cite, Laboratoire d’Electrochimie Moleculaire, UMR 7591 CNRS, 15 rue Jean-Antoine de
Baïf, F-75205 Paris Cedex 13, France
S
* Supporting Information
ABSTRACT: Molecular catalysis of H2 production from the
electrochemical reduction of acids by transition-metal com-
plexes is one of the key issues of modern energy challenges.
The question of whether it proceeds heterolytically (through
reaction of an initially formed metal hydride with the acid) or
homolytically (through symmetrical coupling of two molecules
of hydride) has to date received inconclusive answers for a
quite simple reason: the theoretical bases for criteria allowing
the distinction between homolytic and heterolytic pathways in nondestructive methods such as cyclic voltammetry have been
lacking heretofore. They are provided here, allowing the distinction between the two pathways. The theoretical predictions and
the diagnosing strategy are illustrated by catalysis of the reduction of phenol, acetic acid, and protonated triethylamine by
electrogenerated iron(0) tetraphenylporphyrin. Rather than being an intrinsic property of the catalytic system, the occurrence of
either a heterolytic or a homolytic pathway results from their competition as a function of the concentrations of acid and catalyst
and the rate constants for hydride formation and H2 evolution by hydride protonation or dimerization.
be prepared by an analysis of the diffusion−reaction problem
leading to the identification of the minimal number of
governing dimensionless parameters.
INTRODUCTION
■
In addition to its fundamental interest, the ongoing attention
attracted by catalysis of the production of hydrogen upon
reduction of acids has a double motivation. One aims at a better
understanding of hydrogenases.1−6 The other is of a synthetic
nature and relates to modern energy challenges.7−12 Whether
the reaction follows a homolytic or heterolytic reaction
pathway, as illustrated by Scheme 1,13,14 is a question that
has been addressed for a long time15 and continues to be a
debated issue.16−20 Diverging conclusions have often been
reached even with the same catalyst or closely related catalysts.
This is the case with the popular cobaloxime H2 evolution
catalysts. They were first considered to follow a homolytic
pathway,8b,16a,19 whereas more recent studies favor a heterolytic
pathway.20 Several quantum-chemical calculations of thermo-
dynamic quantities that may help suggest which is the preferred
reaction pathway have been carried out, also favoring the
heterolytic pathway.17,18,20c
Cyclic voltammetry (CV) has been widely used to investigate
the reaction mechanism. However, reliable mechanistic
conclusions could not be reached for the simple reason that
theoretical relationships and the ensuing diagnostic criteria are
lacking for reaction schemes corresponding to homolytic
pathways and to their competition with heterolytic pathways.
One may resort to “digital simulation”21 of experimental cyclic
voltammograms. However, the number of parameters is so
large that many different schemes with ad hoc parameter values
can match the data, thus preventing any effective determination
of the reaction mechanism. Digital simulations may be quite
useful to finalize the mechanism determination, but they should
The aim of the present contribution is to provide these
lacking relationships and the ensuing diagnostic criteria. On the
basis of these criteria, strategies allowing the distinction
between the two types of reaction pathways and the
identification of the parameters that govern their competition
will be proposed. These strategies will be illustrated by an
experimental example, namely, catalysis of the reduction of
phenol, acetic acid, and protonated triethylamine in N,N′
dimethylformamide (DMF) by electrogenerated iron(0)
tetraphenylporphyrin (Fe0TPP).
RESULTS AND DISCUSSION
■
1. Reminder of the Characteristics of a Simple One-
Electron−One-Step Catalytic Reaction Scheme. For a
simple catalytic reaction scheme, the catalytic CV response may
exhibit diverse shapes that are summarized in Figure 1. Since
we are interested in efficient catalysts, the global catalytic rate
constant is implied to be large, and therefore, “pure kinetic
conditions” are achieved, corresponding to the upper three
zones in Figure 1. Under the assumption that enough acid is
present to render its consumption negligible in the course of
the CV run, the kinetic situation belongs to the upper right-
hand zone in Figure 1. An S-shaped CV response is thus
Received: June 10, 2014
Published: September 4, 2014
© 2014 American Chemical Society
13727
dx.doi.org/10.1021/ja505845t | J. Am. Chem. Soc. 2014, 136, 13727−13734

Journal of the American Chemical Society
Article
a b
,
CV responses depend upon three dimensionless parameters
[see the Supporting Information (SI)]:
Scheme 1. EEPDimP Pathways
k2
k1
kdCc0at
k1CA0H
keCc0at
k1CA0H
,
,
and
where C0cat and CA0H are the concentrations of catalyst and acid,
respectively, and the rate constants k are defined in Scheme 1.
The first two parameters measure the rapidity of the second
protonation step and the dimerization step, respectively,
relative to the first protonation step. They are indicative of
the propensity of the key intermediate MN(H−) ↔ MN−1(H•)
to be or not be at steady state. In each of these cases, the ratio
of the two first parameters provides a measure of the
competition between the second protonation step and
dimerization, i.e., the competition between the heterolytic
and homolytic pathways. The third parameter is related to the
reduction of the MN complex, which closes the catalytic loop. It
may take place at the electrode surface or in the solution by
reaction with the MN−2 complex. If k1 is large, the MN complex
(like other intermediates) is formed close to the electrode
surface and is reduced before it has time to meet and react with
the MN−2 complex (ECE-type situation), and vice versa for
large values of keC0cat/k1CA0H [the solution electron transfer
(SET) situation].
a
MJ (J = N, N − 1, N − 2) denotes a complex with a metal at the
oxidation degree J. The vertical double-headed arrow indicates
resonance between two forms of the metal complex. AH denotes
the acid and A− its conjugated base. The k’s are the corresponding rate
We first examine the SET situation (keC0cat/k1CA0H → ∞).
The zone diagram in Figure 2 summarizes the characteristics of
the S-shaped CV response as a function of the two parameters
k2/k1 and kdC0cat/k1CA0H. The kinetic zones are obtained upon
giving extreme values to these parameters (0 and ∞,
respectively) so as to reach simplified behaviors that depend
on a lesser number of parameters (one and zero, respectively).
Because of experimental uncertainties in measuring plateau
currents (∼5%) and half-wave potentials (∼5 mV), these
simpler behaviors are reached for finite small and large values of
the dimensionless parameters, respectively. In the determi-
nation of the boundaries between zones, the simplest zone is
privileged (zero-parameter zone over one-parameter zone, one-
parameter zone over two-parameter zone, and so forth). The
area of the two-parameter zone G (where G = “general”) is thus
minimized at the benefit of the one- and zero-parameter zones.
The meaning of the various zones and the corresponding
expressions for the plateau currents and half-wave potentials are
detailed in the caption of Figure 2. Derivations of these
expressions are provided in the SI.
b
constants. The acronym EEPDimP describes the nature and
sequencing of the successive steps. E denotes electrode electron
transfer, P denotes protonation, and Dim denotes the H2-evolving
dimerization step. This scheme was adapted from a more general
formalism22a to make it more specific to the present case.
The ECE situation (keC0cat/k1CA0H → 0), gives rise to another
zone diagram (Figure 3). It is not very different from the SET
zone diagram (Figure 2), and the expressions for the plateau
currents and half-wave potentials are similar albeit not identical
in the two cases.
Mechanism diagnosis derives from the combined observation
of the plateau current and half-wave potential and their
variation or lack of variation with the acid and catalyst
concentrations according to the equations in the ellipsoidal
inserts that characterize the zero-parameter zones. The working
curves in the rectangular insets indicate the variations to be
expected when passing from one zero-parameter zone to the
other upon variation of the catalyst and acid concentrations. An
estimate of the three rate constants ensues. The reliability of
the diagnostic is then checked by means of digital simulation,
and the values of the rate constants are simultaneously adjusted
for an improved fit of the experimental data.
Figure 1. Zone diagram showing the different shapes of the CV
response as a function of the dimensionless rate (ordinate) and excess
(abscissa) parameters.22b,c
expected, being characterized by a plateau current ipl and a half-
wave potential E1/2
2. Competition between the Heterolytic and Homo-
lytic Pathways in the Framework of Scheme 1. The same
assumptions as in the preceding section are made in the
following discussion of the heterolytic versus homolytic
competition of Scheme 1. Within this framework, the S-shaped
.
13728
dx.doi.org/10.1021/ja505845t | J. Am. Chem. Soc. 2014, 136, 13727−13734

Journal of the American Chemical Society
Article
Figure 2. Competition between the heterolytic and homolytic pathways in the SET regime (keC0cat/k1CA0H → ∞). The zone diagram summarizes the
various kinetic behaviors as a function of the two dimensionless parameters k2/k1 and kdC0cat/k1CA0H. The zones are labeled using the following
abbreviations: HE, heterolytic; HO, homolytic; S, steady-state; N, non-steady-state; G, general. The effects of the experimental parameters contained
in the dimensionless parameters are summarized, as logarithms, by the set of arrows on top of the diagram, which allows traveling through the
various zones. The blue (HES) and green (HEN) zones correspond to heterolytic pathways in which the key intermediate MN(H−) ↔ MN−1(H•) is
or is not in steady state, respectively. Likewise, the red (HOS) and yellow (HON) zones corresponds to homolytic pathways in which the key
intermediate MN(H−) ↔ MN−1(H•) is or is not in steady state, respectively. In these four cases, closed-form expressions for the current and half-
wave potential apply. They are given in the oval inset of the same color as the corresponding zone. Within the heterolytic case, the passage from the
steady-state regime to the non-steady-state regime occurs through zone HEG. The variations of the plateau current (thick black line) and half-wave
potential (dotted line) with the appropriate parameter are shown as working curves in the square inset of the same color (cyan). The thin line
corresponds to the variation of the plateau current in the corresponding limiting zone. Vertical dashed lines indicate the boundaries between the
zones. The same applies, within the homolytic case, to the passage from the steady-state regime to the non-steady-state regime through zone HOG
(orange). The magenta zone (GS) depicts the passage from the heterolytic pathway to the homolytic pathway in the steady-state regime. Here the
variations of the current and half-wave potential are given against a dimensionless parameter that directly represents the competition between the
two pathways. The same applies to the light-green zone (GN), which depicts the passage from the heterolytic pathway to the homolytic pathway in
the non-steady-state regime, although a different dimensionless competition parameter is used. The boundaries between the zones correspond to a
10% uncertainty of the plateau current and a 5 mV uncertainty in the half-wave potential. The potentials are given in volts. S is the electrode surface
area, and D is the catalyst diffusion coefficient.
The easiest situation is when the expressions for the plateau
current and half-wave potential are different when passing from
one zero-parameter zone to the other, allowing the distinction
between the two zones by means of the variation or lack of
variation of the plateau current and half-wave potential as the
acid and catalyst concentrations are varied. In the converse case,
where these formal expressions for the plateau current and half-
wave potential are the same, as for example in the case of zones
HES and HOS, one may have the impression that making a
distinction between the two zones is impossible. In fact, the
following strategy may be applied. The expressions for the
plateau current differ by a factor of 2 or √2 upon crossing the
transition zone GS. Consequently, if the transition zone is
reached upon changing the acid and/or catalyst concentration,
the distinction between the HES and HOS zones becomes
possible. If no transition zone is observed, then a guess can be
made that the system is in either the HES zone or the HOS
zone, and k1 may be estimated in each case. In addition, the
13729
dx.doi.org/10.1021/ja505845t | J. Am. Chem. Soc. 2014, 136, 13727−13734

Journal of the American Chemical Society
Article
Figure 3. Competition between the heterolytic and the homolytic pathways in the ECE regime (keC0cat/k1CA0H → 0). This zone diagram uses the
same definitions and characteristics as in Figure 2. The plateau currents in the HES and HEN zones (and in the corresponding transition zones) are
modified in comparison with the SET regime.
absence of a transition zone implies that 2kdC0cat/k1CA0H is either
<100 (HES) or >100 (HOS). The ensuing values of kd, being
smaller or not smaller than the diffusion limit, finally allow the
effective distinction between the two zones.
The next section is dedicated to an experimental illustration
of these procedures.
3. Catalysis of the Reduction of Protonated Triethyl-
amine, Phenol, and Acetic Acid in DMF by Electro-
generated Fe0TPP: Raw CV Data. Figure 4 summarizes the
CV data obtained with the three acids, using the concentrations
of acid and catalyst as operational parameters to move the point
representing the system in the zone diagram (Figures 2 and 3).
In the absence of acid (Figure 4, top left), the CV response
exhibits two reversible waves corresponding to the first two
electron transfer steps in Scheme 1, with standard potentials
(E0) equal to −1.34 and −2.04 V vs Fc+/Fc, respectively. One
consequence of this potential sequencing is that the SET
reaction at the end of Scheme 1 is strongly downhill and
consequently endowed with a rate constant ke that is most
likely close to the diffusion limit in all three cases. We know
from previous work23 that the iron(II) hydride formed upon
electron transfer to the iron(0) complex is more difficult to
reduce than the iron(I) complex. This observation validates the
reaction sequence represented in Scheme 1 rather than other
sequences of electron transfers, proton transfers, and
dimerization steps, which are briefly discussed in section 6.
Of the three acids, Et3NH+ is the most acidic (pK = 9.2 in
DMF24). The first protonation step is thus expected to be the
most rapid in the series. This is the reason that we used scan
rates of 10, 15, and 20 V s−1 (Figure 4, top right, where only the
10 V s−1 data are shown), which are relatively high in order to
minimize side phenomena starting from acid consumption25
but still low enough for pure kinetic conditions to remain
fulfilled (see Figure 1). With the weaker acids AcOH and
PhOH (pK = 13.2 and 18.8, respectively24), catalysis is less
vigorous (Figure 4, middle and bottom), and a scan rate of 2 V
s−1 offers a good compromise for minimizing the side effects
while preserving the pure kinetic conditions.
4. Catalysis of the Reduction of Protonated Triethyl-
amine, Phenol, and Acetic Acid in DMF by Electro-
generated Fe0TPP: Treatment of the CV Data. Because
Et3NH+ is the strongest acid of the series and gives rise to
13730
dx.doi.org/10.1021/ja505845t | J. Am. Chem. Soc. 2014, 136, 13727−13734

Journal of the American Chemical Society
Article
Figure 5. Catalysis of the reduction of Et3NH+ by electrogenerated
Fe0TPP. Shown are plots of the normalized plateau current (left) and
the half-wave potential referred to the catalyst standard potential
(right) against the square root of the acid concentration. [FeTPP] = 1
mM. Blue dots: experimental data from the top-right panel of Figure 4.
Open triangles: simulation.
unambiguously in the HEN zone, i.e., following a heterolytic
pathway with the hydride key intermediate not being at steady
state. Applying the relationships given in Figure 3 for the HEN
zone leads to the values of k1 and k2 listed in Table 1.26 Taking
Table 1. Rate Constants (M−1 s−1)
acid
k1
k2
kd
Et3NH+
AcOH
PhOH
2 × 108
106
5 × 103
105
−
6.5 × 102
3 × 104
25
104
ke as equal to 2 × 108 M−1 s−1 leads to values of the parameter
keCc0at/k1CA0H smaller than 1 over the whole range, thus
justifying the assumption of an ECE regime. Simulation27 of
the variations of the plateau current and of the half-wave
potential using the values of the rate constants thus estimated
confirms these conclusions (Figure 5).
We come now to PhOH, which gives rise to the smallest
catalytic currents in the series and is therefore likely to fall
within the SET regime. The data (Figure 4 bottom) were
treated accordingly, giving rise to the variations of the plateau
current and the half-wave potential versus the square root of
the acid concentration shown in Figure 6. There is a clear
deviation of the plateau current from proportionality to the
square root of the acid concentration upon decreasing the acid
Figure 4. Cyclic voltammetry of FeIITPP in DMF + 0.1 M n-Bu4NPF6.
Top left: 1 mM FeIITPP in the absence of acid at 0.1 V s−1. Top right:
1 mM FeIITPP at 10 V s−1 in the presence of Et3NH+(M): 0.0061
(blue), 0.0134 (green), 0.0219 (red), 0.0301 (yellow), 0.0421
(magenta), 0.0664 (cyan). Middle left: 1 mM FeIITPP at 2 V s−1 in
the presence of AcOH (M): 0.029 (blue), 0.054 (green), 0.084 (red),
0.124 (yellow), 0.174 (magenta), 0.224 (cyan), 0.274 (orange).
Middle right: 5 mM FeIITPP at 2 V s−1 in the presence of AcOH (M):
0.01 (blue), 0.03 (green), 0.055 (red), 0.085 (yellow), 0.125
(magenta), 0.175 (cyan), 0.225 (orange), 0.275 (light blue), 0.375
(purple), 0.5 (gray), 0.7 (light red), 0.9 (black). Bottom left: 1 mM
FeIITPP at 2 V s−1 in the presence of PhOH (M): 0.18 (blue), 0.355
(green), 0.68 (red), 1 (yellow), 1.24 (magenta), 1.72 (cyan), 2
(orange), 2.3 (light blue). Bottom right: 5 mM FeIITPP at 2 V s−1 in
the presence of PhOH (M): 0.18 (blue), 0.355 (green), 0.68 (red),
1.24 (yellow), 1.72 (magenta), 2.3 (cyan). i0p is the peak current of the
one-electron Nernstian FeI/0 wave at the same scan rate in the absence
of acid [i0p = 0.446FSCc0atD1/2(Fv/RT)1/2].
efficient catalysis, we may infer that k1 is large and therefore
that the kinetic regime is likely to be of the ECE type. We may
thus analyze the data in this framework, i.e., using the zone
diagram of Figure 3 and the accompanying expressions and
working curves. The experimental values of the normalized
plateau current ipl/i0p (where i0p is the peak current of the one-
electron Nernstian FeI/0 wave at the same scan rate in the
absence of acid) and the half-wave potential E1/2 (referred to
the catalyst standard potential E0) are plotted against the square
root of the acid concentration in Figure 5. The proportionality
of the plateau current to the square root of the acid
concentration and the constancy of the half-wave potential at
a positive value relative to the standard potential of the catalyst
are two diagnostic criteria that point to the system being
Figure 6. Catalysis of the reduction of PhOH by electrogenerated
Fe0TPP. Shown are plots of the normalized plateau current (left) and
the half-wave potential referred to the catalyst standard potential
(right) against the square root of the acid concentration. [FeTPP] = 1
mM (blue symbols) or 5 mM (red symbols). Dots: experimental data
from the bottom panels of Figure 4. Open triangles: simulation.
13731
dx.doi.org/10.1021/ja505845t | J. Am. Chem. Soc. 2014, 136, 13727−13734

Journal of the American Chemical Society
Article
concentration and increasing the catalyst concentration,
indicating that the system is traveling across zone GN in the
zone diagram in Figure 2 (also see the expanded view of zones
HEN, GN, and HON in Figure 8). Determination of the rate
constants was done as follows. The data obtained at high
PhOH concentrations are assumed to correspond to the HEN
zone. The plateau current and half-wave potential obtained
under these conditions thus allowed the determination of k1
and k2 (Table 1). On the basis of these values, simulation27 of
the voltammograms over the whole PhOH concentration range
led to the value of kd (Table 1). This procedure allows taking
into account deviations of the cyclic voltammograms from pure
kinetic conditions (see Figure 1), which are assumed in the
zone diagram but are no longer fulfilled when the phenol
concentration is decreased because of the small catalytic rate
constants. We may note that the variation of the half-wave
potential with the concentrations is of no practical use because
of the large uncertainties deriving from the proximity of a
second catalytic wave that became very high as the acid
concentration was raised and thus artificially pushed the first
catalytic wave toward more positive values (Figure 4 bottom).
Taking again a value of 2 × 108 M−1 s−1 for ke leads to values of
the parameter keCc0at/k1CA0H that are larger than 10 over the
whole range of the two concentrations, thus justifying the
assumption of an SET regime.
PhOH. We note a slight deviation of the plateau current plot
from proportionality to the square root of the acid
concentration at low acid concentrations. This observation
points to the interference of a homolytic pathway in this
concentration range. The values of k1 and k2 may be
determined from the large linear portion of the i/i0p plot at
high acid concentrations and from the value of E1/2 in the same
concentration range, leading to k1 = 106 M−1 s−1 and k2 = 103
M−1 s−1. The SET versus ECE parameter varies as 0.2 ≤ keCc0at/
k1CA0H ≤ 30 within the range of acid and catalyst
concentrations. It follows that the system is not exactly under
the SET regime, calling for a slight correction of k1 and k2
obtained by means of simulation and finally leading to the
values reported in Table 1. The variations of the half-wave
potential at 1 mM catalyst concentration (Figure 7, top right)
are not precise enough to corroborate the interference of a
homolytic pathway as noted earlier. However, this is clearly
confirmed by the E1/2 data obtained at a catalyst concentration
of 5 mM (Figure 7, bottom right). Simulations27 allowed
adjustment of the kd value to reproduce the experimental data.
5. Comparison of the Three Acids. On the basis of the
data summarized in Table 1, an expanded view of the bottom-
left corner of the zone diagram (Figure 8) shows that the key
The variations of the plateau current and the half-wave
potential versus the square root of the acid concentration
observed with AcOH are shown in Figure 7 for two values of
the catalyst concentration (1 and 5 mM). The data were
analyzed in the framework of the SET regime, as suggested by
the fact that the catalysis is not very vigorous as it is with
Figure 8. Expanded view of the zone diagram showing how the system
travels through the heterolytic-to-homolytic transition zone as a
function of the nature of the acid, its concentration, and the
concentration of catalyst.
intermediate MN(H−) ↔ MN−1(H•) is not at steady state for all
three systems. The reduction of the strongest acid, Et3NH+,
follows an entirely heterolytic reaction pathway, whereas the
two other concurrently take a heterolytic and a homolytic path.
It also interesting to note that the two protonation rate
constants are linearly correlated with the pK of the acid (Figure
9).
6. Alternative Reaction Pathways. Besides the EEPDimP
pathways (Scheme 1) that we have investigated so far, one may
alternatively consider two other sequences of the E, P, dim, and
P steps within the reaction pathway, as shown in Schemes 2
and 3. The variations of the plateau current and half-wave
Figure 7. Catalysis of the reduction of AcOH by electrogenerated
Fe0TPP. Shown are plots of the normalized plateau current (left) and
the half-wave potential referred to the catalyst standard potential
(right) against the square root of the acid concentration. Top:
[FeTPP] = 1 mM. Bottom: [FeTPP] = 5 mM. Blue dots: experimental
data from the middle panels of Figure 4. Open triangles: simulations
for heterolytic + homolytic pathways (blue) and the heterolytic
pathway alone (magenta).
13732
dx.doi.org/10.1021/ja505845t | J. Am. Chem. Soc. 2014, 136, 13727−13734

Journal of the American Chemical Society
Article
Scheme 3. EPDimEP Pathways
Figure 9. Correlations of the two protonation rate constants k1 (red
dots) and k2 (blue dots) with the pK of the acid.
Scheme 2. EPEDimP Pathways
dominant reaction steps as a function of the experimental
parameters is particularly helpful.
(3) The analysis was focused on fast catalytic systems under
pure kinetic conditions in which side phenomena (e.g.,
substrate consumption, inhibition by products, and deactivation
of the catalysts) have been minimized so as to give rise to
canonical S-shaped CV responses independent of the scan rate
(Figure 1, top right). It is noted en passant that for the same
conditions the same analyses apply to other techniques and to
preparative-scale electrolyses as well. The assumption of
negligible consumption of the substrate could be lifted, with
systems tending toward the conditions of total catalysis (Figure
1, top left), but this would come at the price of an increase in
the number of independent parameters of the system.
(4) For the sake of simplicity, we limited ourselves to
EEPDimP mechanisms (Scheme 1), but the analysis could
easily be repeated for other mechanisms such as EPEDimP
(Scheme 2) and EPDimPE (Scheme 3).
potential for these pathways should then be analyzed in a quite
similar manner by means of zone diagrams accompanied by
adapted expressions and working curves that can be used to
unravel the heterolytic/homolytic competition.
(5) The illustration using the reduction of Et3NH+, AcOH,
and PhOH catalyzed by electrogenerated Fe0TPP demon-
strated the feasibility of the diagnosing strategy and may serve
as a guide for future studies of other catalysts. While the first of
these acids exclusively follows a heterolytic pathway, the other
two provide good examples of heterolytic versus homolytic
competition.
CONCLUDING REMARKS
■
(1) The outcome of this work is that we now have theoretical
tools that allow diagnosis of the occurrence of heterolytic
versus homolytic pathways in the catalysis of H2 evolution from
reduction of acids.
(2) Rather than being an intrinsic property of the catalytic
system, the occurrence of either a heterolytic or a homolytic
pathway results from their competition as a function of the
concentrations of acid and catalyst and of the rate constants for
hydride formation and H2 evolution by hydride protonation or
dimerization. The use of a kinetic zone diagram based on a
minimal number of dimensionless parameters to trace out the
ASSOCIATED CONTENT
* Supporting Information
Experimental details and derivations of the equations allowing
the establishment of the kinetic zone diagram and the various
expressions and working curves depicting the variations of the
plateau current and half-wave potential with the appropriate
■
S
13733
dx.doi.org/10.1021/ja505845t | J. Am. Chem. Soc. 2014, 136, 13727−13734

Journal of the American Chemical Society
Article
(17) (a) Solis, B. H.; Hammes-Schiffer, S. Inorg. Chem. 2011, 50,
11252. (b) Solis, B. H.; Hammes-Schiffer, S. J. Am. Chem. Soc. 2011,
133, 19036.
parameters. This material is available free of charge via the
Internet at http://pubs.acs.org.
(18) Muckerman, J. T.; Fujita, E. Chem. Commun. 2011, 47, 12456.
(19) Hu, X.; Brunschwig, B. S.; Peters, J. C. J. Am. Chem. Soc. 2007,
129, 8988.
(20) (a) Razavet, M.; Artero, V.; Fontecave, M. Inorg. Chem. 2005,
44, 4786. (b) Baffert, C.; Artero, V.; Fontecave, M. Inorg. Chem. 2007,
46, 1817. (c) Bhattacharjee, A.; Andreiadis, E. S.; Chavarot-Kerlidou,
M.; Fontecave, M.; Field, M. J.; Artero, V. Chem.Eur. J. 2013, 19,
15166.
AUTHOR INFORMATION
Corresponding Authors
cyrille.costentin@univ-paris-diderot.fr
saveant@univ-paris-diderot.fr
Notes
■
The authors declare no competing financial interest.
́
(21) Saveant, J.-M. Elements of Molecular and Biomolecular Electro-
REFERENCES
■
chemistry: An Electrochemical Approach to Electron Transfer Chemistry;
John Wiley & Sons: Hoboken, NJ, 2006; pp 121−125.
(1) Tard, C.; Pickett, C. J. Chem. Rev. 2009, 109, 2245.
(2) Felton, G. A. N.; Mebi, C. A.; Petro, B. J.; Vannucci, A. K.; Evans,
D. H.; Glass, R. S.; Lichtenberger, D. L. J. Organomet. Chem. 2009,
694, 2681.
(3) (a) Vincent, K. A.; Parkin, A.; Armstrong, F. A. Chem. Rev. 2007,
107, 4366. (b) Armstrong, F. A.; Belsey, N. A.; Cracknell, J. A.; Goldet,
G.; Parkin, A.; Reisner, E.; Vincent, K. A.; Wait, A. F. Chem. Soc. Rev.
2009, 38, 36.
(4) Gloaguen, F.; Rauchfuss, T. B. Chem. Soc. Rev. 2009, 38, 100.
(5) (a) Liebgott, P.-P.; Leroux, F.; Burlat, B.; Dementin, S.; Baffert,
C.; Lautier, T.; Fourmond, V.; Ceccaldi, P.; Cavazza, C.; Meynial-
Salles, I.; Soucaille, P.; Fontecilla-Camps, J. C.; Guigliarelli, B.;
Bertrand, P.; Rousset, M.; Leg
(b) Abou Hamdan, A.; Liebgott, P.-P.; Fourmond, V.; Gutier
O.; De Lacey, A. L.; Infossi, P.; Rousset, M.; Dementin, S.; Leg
(22) (a) Costentin, C.; Savea
1226. (b) Savea
(c) See p 109 in ref 21.
́
nt, J.-M. ChemElectroChem 2014, 1,
́
nt, J.-M.; Su, K.-B. J. Electroanal. Chem. 1984, 171, 341.
(23) Bhugun, I.; Lexa, D.; Savea
3982.
́
nt, J.-M. J. Am. Chem. Soc. 1996, 118,
(24) Izutsu, K. Electrochemistry in Nonaqueous Solvents; Wiley-VCH:
Weinheim, Germany, 2002; Chapter 3.
(25) (a) Besides substrate consumption, the main side phenomena
are inhibition by products and deactivation of the catalyst. These
effects increase with the charge passed during the CV experiment and
are therefore minimized by raising the scan rate.25b (b) Costentin, C.;
́
er, C. Nat. Chem. Biol. 2010, 6, 63.
rez-Sanz,
er, C.
Drouet, S.; Robert, M.; Savea
11235.
́
nt, J.-M. J. Am. Chem. Soc. 2012, 134,
́
́
(26) These values are in good agreement with those previously
determined under total catalysis conditions.23
(27) These simulations were performed using the DigiElch software.
See: Rudolph, M. J. Electroanal. Chem. 2003, 543, 23.
Proc. Natl. Acad. Sci. U.S.A. 2012, 109, 19916. (c) Fourmond, V.;
Baffert, C.; Sybirna, K.; Lautier, T.; Abou Hamdan, A.; Dementin, S.;
Soucaille, P.; Meynial-Salles, I.; Bottin, H.; Leg
2013, 135, 3926.
́
er, C. J. Am. Chem. Soc.
(6) Thoi, V. S.; Sun, Y.; Long, J. R.; Chang, C. J. Chem. Soc. Rev.
2013, 42, 2388.
(7) (a) Rakowski Dubois, M.; Dubois, D. L. Acc. Chem. Res. 2009, 42,
1974. (b) Rakowski DuBois, M.; DuBois, D. L. Chem. Soc. Rev. 2009,
38, 62.
(8) (a) Gray, H. B. Nat. Chem. 2009, 1, 7. (b) Dempsey, J. L.;
Brunschwig, B. S.; Winkler, J. R.; Gray, H. B. Acc. Chem. Res. 2009, 42,
1995.
(9) Gust, D.; Moore, T. A.; Moore, A. L. Acc. Chem. Res. 2009, 42,
1890.
(10) Koper, M. T. M.; Bouwman, E. Angew. Chem., Int. Ed. 2010, 49,
3723.
(11) (a) Artero, V.; Fontecave, M. Coord. Chem. Rev. 2005, 249,
1518. (b) Tran, P. D.; Artero, V.; Fontecave, M. Energy Environ. Sci.
2010, 3, 727. (c) Artero, V.; Chavarot-Kerlidou, M.; Fontecave, M.
Angew. Chem., Int. Ed. 2011, 50, 7238. (d) Artero, V.; Fontecave, M.
Chem. Soc. Rev. 2013, 42, 2338.
(12) (a) Lee, C. H.; Dogutan, D. K.; Nocera, D. G. J. Am. Chem. Soc.
2011, 133, 8775. (b) Roubelakis, M. M.; Bediako, D. K.; Dogutan, D.
K.; Nocera, D. G. Energy Environ. Sci. 2012, 5, 7737.
(13) (a) This catalytic reaction scheme is the exact homogeneous
molecular counterpart of the reduction of Brønsted acids on platinum,
with the first steps being equivalent to the Volmer step, the homolytic
step of the Volmer−Tafel step, and the heterolytic step of the
Volmer−Heyrovsky step.13b (b) Bockris, J. O.; Reddy, A. K. N. In
Modern Electrochemistry; Plenum Press: New York, 1970; Vol. 2,
Chapter 10, pp 1231−1250.
(14) Other competitive heterolytic versus homolytic mechanisms
with different sequencing of the electron transfer, proton transfer, and
dimerization steps are possible. They will be envisaged later on.
(15) (a) Collman, J. P.; Wagenknecht, P. S.; Lewis, N. S. J. Am. Chem.
Soc. 1992, 114, 5665. (b) Collman, J. P.; Ha, Y.; Wagenknecht, P. S.;
Lopez, M. A.; Guilard, R. J. Am. Chem. Soc. 1993, 115, 9080.
(16) (a) Dempsey, J. L.; Winkler, J. R.; Gray, H. B. J. Am. Chem. Soc.
2010, 132, 1060. (b) Dempsey, J. L.; Winkler, J. R.; Gray, H. B. J. Am.
Chem. Soc. 2010, 132, 16774. (c) Valdez, C. N.; Dempsey, J. L.;
Brunschwig, B. S.; Winkler, J. R.; Gray, H. B. Proc. Natl. Acad. Sci.
U.S.A. 2012, 109, 15589.
13734
dx.doi.org/10.1021/ja505845t | J. Am. Chem. Soc. 2014, 136, 13727−13734