Current Issues in Molecular Catalysis Illustrated by Iron Porphyrins as

Article pubs.acs.org/accounts

Current Issues in Molecular Catalysis Illustrated by Iron Porphyrins as Catalysts of the CO2‑to-CO Electrochemical Conversion Cyrille Costentin,* Marc Robert,* and Jean-Michel Savéant* Université Paris Diderot, Sorbonne Paris Cité, Laboratoire d’Electrochimie Moléculaire, Unité Mixte de Recherche Université - CNRS No. 7591, Bâtiment Lavoisier, 15 rue Jean de Baïf, 75205 Cedex 13 Paris, France CONSPECTUS: Recent attention aroused by the reduction of carbon dioxide has as main objective the production of useful products, the “solar fuels”, in which solar energy would be stored. One route to this goal is the design of photochemical schemes that would operate this conversion using directly sun light energy. An indirect approach consists in first converting sunlight energy into electricity then using it to reduce CO2 electrochemically. Conversion of carbon dioxide into carbon monoxide is thus a key step through the classical dihydrogen-reductive Fischer− Tropsch chemistry. Direct and catalytic electrochemical CO2 reduction already aroused active interest during the 1980−1990 period. The new wave of interest for these matters that has been growing since 2012 is in direct conjunction with modern energy issues. Among molecular catalysts, electrogenerated Fe(0) porphyrins have proved to be particularly efficient and robust. Recent progress in this field has closely associated the search of more and more efficient catalysts in the iron porphyrin family with an unprecedentedly rigorous deciphering of mechanisms. Accordingly, the coupling of proton transfer with electron transfer and breaking of one of the two C−O bonds of CO2 have been the subjects of relentless scrutiny and mechanistic analysis with systematic investigation of the degree of concertedness of these three events. Catalysis of the electrochemical CO2-to-CO conversion has thus been a good testing ground for the mechanism diagnostic strategies and the all concerted reactivity model proposed then. The role of added Brönsted acids, both as H-bond providers and proton donors, has been elucidated. These efforts have been a preliminary to the inclusion of the acid functionalities within the catalyst molecule, giving rise to considerable increase of the catalytic efficiency. The design of more and more efficient catalysts made it necessary to propose “catalytic Tafel plots” relating the turnover frequency to the overpotential as a rational way of benchmarking the catalysts within iron porphyrins and among all available molecular catalysts, independently of the characteristics of the electrolytic cell in use. To be reliable, such assignments of the intrinsic characteristics of catalysts are grounded in the accurate elucidation of mechanisms. Without forgetting the importance of large scale electrolysis, not only mobilization of all resources of nondestructive techniques such as cyclic voltammetry was necessary to achieve this challenge, but also new approaches, such as foot-of-the-wave analysis combined with raising of scan rate, had to be applied. The latest improvement in catalyst design was to render it water-soluble while preserving, or even augmenting, its catalytic efficiency. The replacement of the nonaqueous solvents so far used by water makes the CO2-to-CO half-cell reaction much more attractive for applications, allowing its association with a water-oxidation anode through a proton-exchange membrane. Manipulation of pH and buffering then allow CO2-to-CO conversions from those involving complete CO-selectivity to ones with prescribed CO−H2 mixtures. Overall, it appears that not only are iron porphyrins the most efficient catalysts of the CO2-to-CO electrochemical conversion but also they can serve to illustrate general issues concerning the field of molecular catalysis as a whole, including other reductive or oxidative processes.

1. INTRODUCTION One of the most important targets of contemporary energy and environmental research is the reductive conversion of carbon dioxide into fuels by means of solar energy.1,2 Two approaches may then be devised. One is to set up photochemical schemes that would operate this conversion using sun light energy directly. Another route toward the same ultimate goal is to first convert solar energy to electricity, which will then be utilized to reduce CO2 electrochemically. Direct electrochemical injection of an electron into the CO2 molecule, forming the corresponding anion radical, CO2•−, © 2015 American Chemical Society

requires very high energy. Indeed, in N,N′ dimethylformamide (DMF), the standard potential of the CO2/CO2•− couple is as negative as −1.97 V vs SHE (standard hydrogen electrode).3 Electrochemical conversion of CO2 to any reaction product thus requires catalytic schemes that preferably avoid this intermediate. It should also be borne in mind that reduction of CO2 to any product entails the association of proton transfer with electron transfer, if only on thermodynamic grounds, as appears in Figure Received: May 23, 2015 Published: November 12, 2015 2996

DOI: 10.1021/acs.accounts.5b00262 Acc. Chem. Res. 2015, 48, 2996−3006

Article

Accounts of Chemical Research

oxalate.10 It accompanied the first attempts to use transition metal complexes as catalysts of the CO2-to-CO conversion.11 A new wave of interest has grown since 2012 in conjunction with contemporary energy issues. These are the recent advances that are discussed in this Account with two connected objectives in mind. One is the quest for more efficient catalysts of the CO2-toCO electrochemical conversion. The other is the systematic deciphering of mechanisms using nondestructive techniques such as cyclic voltammetry, without forgetting the importance of preparative-scale electrolysis as the ultimate goal of catalyst design. The importance of coupling proton transfer with electron transfer appeared since the very first studies of electrochemical CO2-to-CO conversion catalysis by iron porphyrins, but recent studies have extended the number of usable acids and provided a detailed picture of the reaction mechanism. This could be established by a systematic analysis of the CV current−potential responses as a function of the reactants’ concentrations. One obstacle to this endeavor was the appearance of “secondary phenomena” that make the current−potential responses deviate from the canonical behavior expected from simple catalytic reaction schemes. The strategies proposed to eliminate the interference of this undesired phenomena, combining “foot-ofthe-wave analysis” and raising the scan rate in cyclic voltammetry, apply to the particular catalytic reaction under examination but have a much broader scope. They are the aim of section 2. With these tools in hand, we will discuss, in the section 3, the catalysis mechanism in the case of the simplest iron porphyrin of the family, namely, FeTPP (Chart 1). Once it was observed that a Brönsted acid such as phenol considerably enhances catalysis, the idea rapidly came to mind that installing the phenol moieties inside the molecule (CAT and FCAT, see Chart 1) should further boost catalysis if only because of the large local acid concentration thus available. Catalysis gets indeed more efficient. However, the mechanism changes compared with that for FeTPP, as well as the degree of concertedness between electron transfer, proton transfer, and breaking of one of the C−O bonds of CO2. These findings are the object of section 4. At this stage, it became apparent that a reliable way of comparing the performance of the catalysts one to the other in terms of overpotential and efficiency is needed, inside and outside the iron porphyrin family. Section 5 addresses this question, showing that rational benchmarking of the catalysts should be based on the comparison between “catalytic Tafel plots” relating the turnover frequency to the overpotential. Comparison between CAT and FCAT is an example showing that manipulation of substituents may not be a zero-sum game as discussed in section 6. For a long time, nonaqueous polar solvents (mostly DMF and acetonitrile) have been used to investigate the properties of the various catalysts of the electrochemical CO2-to-CO conversion, even if some water was sometimes added to speed up catalysis. The results thus obtained in nonaqueous or weakly aqueous media enabled the discovery of remarkably efficient and selective catalysts of the CO2-to-CO conversion. They were also the occasion of notable advances in the field of mechanisms and theory of concerted bond-breaking/proton−electron transfer. One should however recognize that, from the point of view of practical applications, the use of nonaqueous solvents is not the most exciting aspect of these results. One would rather like to use water as the solvent. We have found that the porphyrin WSCAT (Chart 1), which bears trimethylammonium substituents in para-

1.4 Efficient catalytic schemes must therefore include protoncoupled electron transfer associated with heavy atom bond

Figure 1. Variation of the apparent standard potential with pH (Pourbaix diagrams) for the reductive conversion of CO2 into various products.4

breaking. This even applies to the CO2-to-CO conversion despite that the target product does not contain hydrogen atoms. Proton transfer from Brönsted acids present (AH) does accompany the cleavage of one of the C−O bonds of CO2 with production of one water molecule (Scheme 1). Scheme 1

Carbon monoxide may be an interesting step toward the desired fuels since it can be used as feedstock for the synthesis of alkanes through the classical Fischer−Tropsch process. A number of molecular catalysts for the homogeneous electrochemical CO2-to-CO conversion have been proposed. They mainly derive from transition metal complexes.5−8 The initial step of the catalytic process (Scheme 2) is the electrochemical generation of an appropriately reduced state of Scheme 2a

M, metal complex; N, N − 1, formal oxidation degrees of the metal in the complex a

the catalyst, which reacts with the substrate, CO2, leading to the product, CO, with regeneration of the starting form of the catalyst. This reaction entails the breaking of one of the two C−O bonds of CO2, the transfer of an additional electron, and the transfer of two protons. Establishing the sequencing of these events and their degree of concertedness is not an easy task. It requires the rigorous application of all resources of electrochemical mechanistic analysis to reach reliable conclusions. One of the most thoroughly investigated families of such catalysts is that of iron porphyrins brought electrochemically to the oxidation degree 0. There has been two periods in the investigation of this field. The first one in the years 1980−19909 followed studies of the direct electrochemical reduction of CO2 and of its reduction by means of organic mediators in nonaqueous solvents, leading competitively to formate and 2997


Article

Accounts of Chemical Research Chart 1. Iron-Porphyrin Catalysts for CO2-to-CO Electrochemical Conversion

Figure 2. (a) Red curve, standard one-electron Nernstian diffusioncontrolled wave; blue curve, “canonical” Nernstian pure kinetic catalytic wave (see text). (a′) Linear analysis of the canonical catalytic wave. (b, b′) cyclic voltammetry of FeTPP (1 mM) in DMF + 0.1 M n-Bu4NBF4, in the presence of 0.23 M CO2 and 3 M PhOH on a Hg electrode at 21 °C. Variations with the scan rate (V/s): 0.1 (green), 1 (red), 10 (yellow), 50 (blue): (b) cyclic voltammetric responses; (b′) foot-of-thewave analyses. (c) Catalytic rate constant derived from the foot-of-thewave analyses in panel b′.

position of each of the phenyl groups of TPP, fulfills this requirement while being an efficient catalyst of the electrochemical CO2-to-CO conversion. Section 7 is devoted to this catalyst and to the possibility of obtaining selectivity for CO or CO + H2 (syngas) mixtures in prescribed proportions. It would appear that iron porphyrins not only are efficient catalysts of the CO2-to-CO electrochemical conversion but also can serve as illustrating examples of general issues concerning the field of molecular catalysis as a whole. These aspects will be addressed in our concluding remarks.

2a). The return trace is exactly superimposable to the forward trace and obeys the following equation:12 i=

0 FSCcat Dcat 2kcatCA0 F 0 ⎤ 1 + exp⎡⎣ RT (E − Ecat )⎦

(1)

E0cat

(E = electrode potential, = standard potential of the catalyst couple, Dcat = diffusion coefficient of the catalyst, kcat = rate constant of the catalytic reaction, C0cat = total concentration of catalyst, and S = electrode surface area) when the electrode electron transfer is fast enough for the Nernst law to apply. The current−potential curve may be analyzed as depicted in Figure 2a′, giving rise to a characteristic straight line. Although Figure 2a,a′ and eq 1 form the base of the analysis of current−potential responses, “secondary phenomena” may interfere distorting the S shape, making peaks appear as well as hysteresis between the forward and backward traces. These “secondary phenomena” may involve substrate consumption, inhibition by products,

2. FIGHTING “SECONDARY PHENOMENA”: FOOT-OF-THE-WAVE ANALYSIS AND RAISING THE SCAN RATE Considering a two-electron process as depicted in Scheme 2 with the second electron transfer taking place in solution, if fast catalytic reaction is involved and the substrate is negligibly consumed during the experiment, cyclic voltammetric current− potential responses take a very characteristic S-shaped form independent from scan rate (blue curve in Figure 2a),12,13 quite different from the diffusion-controlled peaks (red curve in Figure 2998


Article


Figure 3. That carbon monoxide is the main product of the catalytic reduction of CO2 may be guessed from the backward trace of the catalytic current. It is indeed well-known that iron(II) porphyrins strongly bind CO. The small wave around −1.0 V thus corresponds to the oxidation of FeITPP into the strongly stabilized OC/FeIITPP complex resulting in a substantial shift of the FeITPP oxidation potential (bottom of Figure 3). This is confirmed by preparative scale electrolyses in the presence of each of the four different acids investigated (Table 1).

deactivation of the catalyst, etc., all phenomena that grow up together with catalysis.14a An example, pertaining to FeTPP, is given in Figure 2b.14b,c Analyses according to the procedure defined in Figure 2a′ are, as expected, linear at the foot of the wave, where catalysis is not too strong (Figure 2b′), thus allowing the determination of the catalytic rate constant, kcat. This example illustrates the importance of combining foot-of-the-wave analysis (FOWA) and raising of scan rate: FOWA for the slowest scan rate has a very short linear portion from which one can only derive a very approximate value of kcat, whereas accuracy is definitely better for the three other scan rates. Another strategy could have been attempting to identify these secondary phenomena and take them into account in the interpretation and treatment of the data. In most cases, this would be a hopeless approach to the question since these phenomena are not known independently, may not be contained in the list given above, and may combine with one another. It seems more effective to minimize their interference in the absence of precise data that would allow their simulation.

Table 1. Catalysis of the CO2-to-CO Electrochemical Conversion by FeTPPa acid

concentration (M)

Faradaic yield (%)

CF3CH2OH

0.1 0.5 1 0.1 0.5 1 0.1

97 98 100 100 98 94 31b

PhOH

3. CATALYSIS OF THE CO2/CO ELECTROCHEMICAL CONVERSION BY SIMPLE IRON(0) PORPHYRINS: COUPLING ELECTRON TRANSFER WITH PROTON TRANSFER14B Four different Brönsted acids were shown to boost catalysis, namely, water, trifluoroethanol, phenol, and acetic acid. Stronger acids were avoided since the iron(0) porphyrin is then a good catalyst of proton reduction as shown in the case of protonated triethylamine.15 Figure 3 shows a typical example of the catalytic reduction of CO2 by Fe0TPP in the presence of an acid, AH. Starting with

AcOH a

Preparative-scale electrolyses in DMF + 0.1 M nBu4NPF6 at a mercury pool working electrode, with [FeTPP] = 1 mM, the solution being saturated in CO2 (0.23 M), with various concentrations of CF3CH2OH, PhOH, and AcOH. Electrolysis potential: −1.46 V vs SHE. bDegradation of the catalyst occurs.

Cyclic voltammograms were systematically recorded as a function of the concentrations of CO2 and each of the abovementioned acids and treated as depicted in the preceding section in order to eliminate the effect of secondary phenomena, thus leading to the determination of an apparent catalytic rate constant, kap. The reaction order in CO2 was observed to be 1, and the variations of the apparent rate constant with acid concentration were as summarized in Figure 4. As transpires from this figure, particular attention was devoted to the role of water in the presence of the other acids to free data analysis of the possible role of residual water. Careful analysis of the reaction orders derived from the data reported in Figure 4 allows establishment of the mechanism shown in Scheme 3 and of the related kinetic constants (Table 2). From these observations and after discarding one by one the other possible pathways, reaction 3 appears to follow a mechanism in which proton transfer and C−O bond breaking are concerted with an intramolecular electron transfer (red pathway in Scheme 4). The two carbon−oxygen bonds do not play a symmetrical role in the reaction since one C−O bond is broken during the proton-coupled intramolecular electron transfer process while the other one is preserved in CO. It follows that the forward reaction goes through the Fe I asymmetrical adduct rather than the FeII symmetrical adduct. The reverse reaction conversely goes through the FeII form rather than through the FeI form of the FeIICO adduct. It is worth noting that the representation of the initial Fe0−CO2 adduct by − Fe(I)CO2•− as the predominant resonance form is in accordance with the results of DFT calculations.16 There is a linear correlation between the log of the kinetic constant and the pK of the acid (Figure 5), the variation of which amounts to a variation of the driving force of the reaction. The fact that the symmetry factor characterizing this correlation, 0.35, is significantly smaller than 0.5 falls in line with the concerted character of the rate-determining step.17

Figure 3. Cyclic voltammetry of FeTPP (1 mM) in DMF + 0.1 M nBu4NPF6, in the absence (blue) and presence of 0.23 M CO2 and of 10 mM PhOH (red). Sketch of the electrochemical reactions.

FeIITPP, two reversible reduction waves are observed in the absence of CO2 and AH corresponding successively to the generation of FeITPP and Fe0TPP. Addition of CO2 and AH does not change the first wave but produces a drastic increase and change of shape of the second wave. The second wave is the result of the catalytic reaction sketched in the upper inset of 2999


Article

Accounts of Chemical Research Table 2. Catalysis of the CO2-to-CO Electrochemical Conversion by FeTPP: Kinetic constants AH H2O CF3CH2OH PhOH AcOH a

pK 31.5 24.0 18.8 13.3

K1k2 (M−2 s−1)

K1k2k3 (M−3 s−1)

K1k2k3 KIEa

2

4 × 104 6 × 105

10 8 × 104 8 × 106 1.2 × 108

1.8 2.5 1.5

Kinetic H/D isotope effect.

Scheme 4a

Figure 4. Variations of the apparent catalytic rate constant with the concentration of acids in the presence of variable amounts of water. Concentrations in M, kap in M−1 s−1. Solid lines: magenta, lowconcentration first-order; green, low-concentration second-order; red, high-concentration first-order; blue, global fittings (see text).

a

Stepwise pathways in blue; concerted pathway in red.

Scheme 3

4. INSTALLING THE ACID FUNCTIONALITIES IN THE CATALYST MOLECULE16,18 Starting from the boosting effect of Brönsted acids, the idea of installing such acid functionalities in the catalyst molecule was based initially on the expectation of a simple concentration effect assuming that the catalytic mechanism would remain the same as

Figure 5. Variations of the kinetic constant K1K2k3 with pK of the added acid.

with FeTPP. Local concentrations on the order of 100 M may indeed be conceived. Such concentrations cannot be achieved in solution without involving the direct reduction of the acid, which 3000


Article

Accounts of Chemical Research would muddle any effect on the catalysis of the CO2-to-CO conversion. That catalysis is indeed boosted by the presence of the phenol functionalities in ortho and ortho′ positions of the TPP phenyl rings manifests itself by a huge increase of the catalytic current (Figure 6).

Figure 8. CAT reaction mechanism at the prewave (in blue) and the catalytic wave (in red). The illustrating CV curves are taken from Figure 7 ([CO2] = 0.23 M, [PhOH] = 1 M). The scan rates are the same as in Figure 7.

Figure 6. Cyclic voltammetry of 1 mM TPP (left) and CAT (right) in DMF + 0.1 M n-Bu4NPF6 in the presence of 0.23 M CO2 and 1 M H2O at 0.1 V/s.

Figure 9. DFT structure of the Fe0−CO2 adduct in CAT. Reproduced with permission from ref 16. Copyright 2014 American Chemical Society.

Peaks being then observed instead of plateaus, it is also clear that secondary phenomena are at work. Preparative-scale electrolyses pointed to the quasi-exclusive formation of CO as with FeTPP. Careful inspection of the CV current−potential responses, after getting rid of secondary phenomena by raising the scan rate (see section 2) as shown in Figure 7, revealed that the mechanism has changed. Albeit not easy to detect, there is a prewave in front of the catalytic wave (Figures 7 and 8) at which the FeI porphyrin not only is reduced to the Fe0 complex that binds CO2 (see Figure 9), as it does in the case of TPP, but also gets protonated before

being reduced a second time giving rise at a slightly more negative potential to the catalytic process (Figure 8). This is a consequence of the stabilization of the initial Fe0−CO2 adduct by the H-bonds provided by the pendant OH groups installed in the ortho positions of the phenyl rings (see Figure 9), which is the main cause of the strong acceleration of catalysis. Catalysis then entails the reduction of the protonated Fe0−CO2 adduct according to a mechanism in which electron transfer, proton transfer, and C−O bond breaking are all concerted. Stabilization of the Fe0−CO2 adduct by the OH groups installed in the

Figure 7. Cyclic voltammetry of 1 mM CAT in DMF + 0.1 M n-Bu4NPF6 in the presence of 0.046 (top) and 0.23 (bottom) M CO2 and from left to right 0.3, 0.5, 1, 2, and 3 M PhOH as a function of the scan rate (V/s): 0.1 (blue), 0.2 (light green), 0.5 (red), 1 (yellow), 5 (magenta), 10 (orange), 20 (cyan), 30 (dark green). In black, simulation of the catalytic current. 3001


Article

Accounts of Chemical Research bulk dCproduct

molecule is also indicated by the DFT results: as compared with TPP, the gain is ca. 0.2 eV, although, in absolute value, the adduct formation is predicted to remain uphill in both cases, contrary to experimental evidence.16

dt

=

0 TOFmax Dcat S Ccat 0 ⎤ V 2 1 + exp⎡⎣ F (E − Ecat )⎦ RT

Examples of catalytic Tafel plots are displayed in Figure 11, as derived from cyclic voltammetry or preparative-scale data. Good catalysts stand at the upper left corner of the log TOF versus η plot and vice versa. It is also interesting to note that the catalytic Tafel plot for a given catalyst provides a guide to its use, allowing one to compromise between energy saving (overpotential) and fastness (TOF). Catalytic Tafel plots allow benchmarking of the catalysts independently of the particular electrolytic cell in use. However, when a particular catalyst has been selected and the balance between overpotential and TOF decided, the rate at which the product is generated in an electrolysis experiment depends on the cell characteristics, taking again the Nernstian case as example,

5. CATALYST BENCHMARKING So far, comparison between the catalyst efficiency has been simply done by comparing the current obtained in cyclic voltammetry at the same scan rate. A more rigorous comparison is needed for the catalysts we are mentioning here but also with other molecular catalysts of the same reaction that have been reported in the literature. Two parameters have to be considered in this connection, namely, the overpotential and the turnover frequency (TOF) or, more exactly, the relationship between the two, that is, “the catalytic Tafel plots” characterizing each catalyst.14 The overpotential, η, is a simple thermodynamic notion, namely, the difference between the standard potential of the reaction to be catalyzed and the electrode potential: η = E0,ap CO2/CO − E. The turnover frequency is defined as TOF = Nproduct/Nactive‑cat., where Nproduct is the number of moles of product generated per unit of time and Nactive‑cat. is the maximal number of moles of active catalyst. By reference to homogeneous catalytic reactions, it is tempting to consider Nactive‑cat. as the number of moles of catalyst contained in the whole electrolysis compartment. In fact, in the framework of a fast catalytic electrochemical process, not all catalyst molecules participate in catalysis but only those that are contained within a thin reactiondiffusion layer adjacent to the electrode, as illustrated in Figure 10 in the case of a reductive process. The reaction-diffusion layer is thinner the faster the catalytic reaction. Then, the catalytic Tafel plots obey the following equation, in the case where the catalyst redox couple is fast (complies with the Nernst law):

bulk dCproduct

dt

=

TOFmax Dcat S 0 ⎤ V 1 + exp⎡⎣ F (E − Ecat )⎦ RT

indicating that the faster the production the larger the ratio between the working electrode surface area (S) and the volume of the solution (V), as it is the case for any preparative scale electrolysis. Another important factor of merit is the stability of the catalyst. The above analysis applies during the electrolysis period of time where the catalyst may be considered as stable, in which case, TOF is constant and the turnover number, TON, increases continuously with time at a pace defined by TOF. In practice, progressive deactivation of the catalyst calls for another definition of the turnover number that may be named limiting turnover number, TONlim, which expresses the number of catalytic cycles performed by the catalyst before disappearance. Recent work26 showed that TONlim is simply equal to the ratio of the rate constants of catalysis and catalyst deactivation. This essential factor of merit of the catalyst can be determined at the end of the electrolysis when the catalyst entirely disappeared. It may however be more conveniently obtained before complete inactivation of the catalyst by observing of the variations with time of the charge passed, of the product formed, and of the decay of the catalyst concentration in solution. In practical terms, the most interesting situation is when the deactivation over catalysis rate ratio is small, even though the catalyst is not unconditionally stable. Then, the catalytic Tafel plot can be derived from the observation of the turnover frequency during the early stages of electrolysis. It should also be borne in mind that this turnover frequency vs overpotential relationship can be derived from less tedious cyclic voltammetric experiments in which catalyst inactivation plays a marginal role as discussed earlier in this section.

6. EFFECT OF ELECTRON-WITHDRAWING SUBSTITUENTS A first example is provided by the FCAT catalyst (Chart 1). Compared with CAT, the same OH functionalities are present, favoring H-bonding of the Fe0−CO2 adduct and protonation en route to catalysis. The inductive effect of the five fluorine atoms on each phenyl group is expected to be favorable to catalysis in that it moves the standard potential of the catalyst in the positive direction. However, the same inductive effect is also expected to

Figure 10. Catalysis of a reductive process. Concentrations profiles at the plateau of the oxidized (dotted lines) and reduced (full lines) forms of the catalyst in the absence (blue lines) and presence (red lines) of the substrate. 3002


Article


Figure 11. Benchmarking of the CO2-to-CO conversion catalysts by means of catalytic Tafel plots as derived from the data reported in refs 13, 16, and 18−25.

8).16,20 The ensuing thermodynamic and kinetic characteristics then allowed the determination of the catalytic Tafel plots shown in Figure 11, thus confirming that FCAT is definitely a better catalyst than CAT. Another example is provided by the iron porphyrin noted as WSCAT in Chart 1, in which four trimethylammonium substituents have been introduced in para-positions of the phenyl groups of TPP. This molecule was initially designed to catalyze the CO2-to-CO electrochemical conversion in water (see next section). It may also be investigated in DMF and compared with the other catalysts under the same conditions. It thus appears as a quite efficient catalyst, even more efficient than CAT and FCAT, as revealed by the corresponding catalytic Tafel plot shown in Figure 11. It is easy to conceive that the strong inductive effect of the four trimethylammonium substituents is able to very significantly move the catalytic wave toward positive potentials as indeed observed in Figure 12. The reasons that this very same inductive effect does not decrease the TOFmax are not obvious, calling for a more detailed analysis of the mechanism and for a more quantitative estimation of the key thermodynamic and kinetic parameters of the reaction than presently available.

decrease the electron density on the iron in the iron(0) complex and thus to be unfavorable to the formation of the Fe0−CO2 adduct. As indicated by the CV responses of the two complexes in the presence of CO2 (Figure 12), the balance of the two effects appears to be globally favorable to catalysis.16,19 A more rigorous mechanism analysis is however required to substantiate the comparison. This was carried out in the same way as for CAT, giving rise to CV responses that are very similar to those displayed in Figure 7 leading to the same mechanism (Figure

7. THE CO2/CO ELECTROCHEMICAL CONVERSION IN WATER The work reported in the preceding sections, carried out in nonaqueous or partially aqueous media, enabled the discovery and progressive development of remarkably efficient and selective catalysts of the CO2-to-CO electrochemical conversion. It was also successful in establishing mechanisms and rational bases for catalyst benchmarking. It however must be recognized that, from the standpoint of practical applications, one would rather like to use water as the solvent. This would ease the optimization of CO2-to-CO half-cell reaction as well as its

Figure 12. Cyclic voltammetry of 1 mM WSCAT (red), FCAT (blue), and CAT (green) in DMF + 0.1 M n-Bu4NPF6 at 0.1 V s−1 in the presence of 0.23 M CO2 and of 1 M PhOH; i0p, the peak current of the reversible FeII/FeI wave at the same scan rate represents a one-electron stoichiometry. 3003


Article


efficiency with no degradation of the selectivity. The last touch of this venture was the design of a water-soluble porphyrin obtained by inclusion of four trimethylammonio substituents in the parapositions of the TPP phenyl rings. In pH 7 carbonate-buffered water, it catalyzes efficiently the formation of CO with very little hydrogen formation. Manipulation of pH and addition of buffers are expected to allow the generation CO−H2 mixtures (syngas) in prescribed proportions. These findings also open the route to the design and optimization of a full electrolysis cell in which the WSCAT porphyrin would be entrapped in a polymeric film deposited on the electrode surface in association with a wateroxidation anode by means of a proton-exchange membrane. Obtaining these results was the fruit of the development of concepts and experimental strategies that can be applied to molecular catalysis in general. One may also view the ironporphyrin venture as an experimental validation of these concepts and strategies. It also allowed uncovering of unexpected problems, the solution of which can be transposed to other catalytic systems. These transposable knowledge items may be summarized as follows. • Without forgetting that preparative scale electrolysis is the ultimate goal, application of a nondestructive technique such as cyclic voltammetry is the main tool of mechanism analysis. It has to be systematic and rigorous. In the quest for reaction mechanisms, variations of the CV responses with all available parameters (concentrations and scan rate) should be scrutinized in detail. A few “digital simulations” do not suffice. All possible pathways but one should be explored and ruled out successively. As known from the very beginning of cyclic voltammetry,12a fast catalytic reactions entail “pure kinetic” conditions, which are expected to give rise to S-shaped CV responses. Observation of scan rate dependent peaks under such circumstances points to the interference of secondary phenomena such as substrate consumption, inhibition by products, deactivation of the catalyst, and any other phenomenon that increases with the charge passed. Rather than trying to simulate these phenomena, the nature and characteristics of which are not precisely known, a better option is to minimize their interference by restricting the analysis to the foot of the catalytic wave. Another approach is to increase the scan rate so as to reduce the charge passed. One can also combine profitably foot-of-the waveanalysis and raising of scan rate. Ignoring these effects and treating the raw peak data with equations designed for plateaus leads to unreliable rate “constants” that vary with the scan rate.27 • Since turnover frequency and overpotential are mutually dependent quantities, related by catalytic Tafel plots, a rational benchmarking of the catalysts ought to be based on the comparison of these plots established for each catalyst as illustrated by Figure 11. An important peculiarity of catalytic electrochemistry, namely, the fact that the catalytic reaction takes place within a thin reaction-diffusion layer, should be taken into account. As a consequence, the TOF may then be defined so as to obtain a catalytic Tafel plot that reflects the intrinsic properties of the catalyst regardless of the characteristics of the electrochemical cell. Another interest of the catalytic Tafel plot of a given catalyst is that it allows the optimization of its use by trading between energy saving (low overpotential) and rapidity (high TOF). Concerning

association with a water-oxidation anode by means of a protonexchange membrane. Rather than electrode coating with an insoluble catalyst, complete solubility in water allows easy manipulation of pH and buffering of the system. In future attempts to entrap the catalyst in a polymeric film deposited on the electrode surface, full solubility in water should also minimize adverse self-interactions between catalyst molecules and guarantee full access of the bathing solution. Foreseeable difficulties in these attempts were that CO2 is poorly soluble in water and that it is partially converted into carbonic acid, making it possible that the CO2-to-CO conversion might be challenged by H2 evolution from reduction of carbonic acid or hydrated protons. Despite these challenging circumstances, the WSCAT iron tetraphenylporphyrin made water-soluble by the four trimethylammonio para-substituents of the phenyl rings (Chart 1) does catalyzes the CO2-to-CO electrochemical conversion provided that the pH is close to neutrality.20 This can be observed in cyclic voltammetry where a very large current is seen in the FeI/0 potential region when CO2 is introduced into the solution (Figure 13a).

Figure 13. Cyclic voltammetry of 0.5 mM WSCAT at 0.1 V/s in water + 0.1 M KCl brought to pH = 6.7 by addition of KOH (a) under 1 atm CO2 or (b) in the absence of 1 atm CO2.

CO was found to be largely predominant (typically 90%) with the formation of only a very small amount of hydrogen (typically 10%) upon preparative-scale electrolyses around −1 V vs SHE in a carbonate buffered solution of pH close to 7. In a formate buffered solution of pH 3.7, electrolysis resulted in the exclusive formation of hydrogen, whereas in 0.1 M phosphate buffer at pH 7, a 50−50 CO−H2 mixture was obtained. These findings open the way to the possibility of adjusting the pH and the buffer content of the solution so as to generate CO− H2 mixtures (syngas) in prescribed proportions.

8. CONCLUSIONS AND PERSPECTIVES After the first wave of interest aroused by iron-porphyrins as catalysts of the CO2-to-CO electrochemical conversion during the late 1980s−mid-1990s period, the topic has received a booming renewed attention in recent times accompanying the daring contemporary energy and environmental challenges. They now appear as robust, efficient, and selective catalysts of the CO2-to-CO electrochemical conversion. The role of Brönsted acids, both as H-bond and proton donors, has been confirmed and explained in detail at the mechanistic level. Modification of the initial tetraphenylporphyrin framework by inclusion of pendant acid groups or electron-withdrawing substituents on the phenyl groups of TPP has allowed improvement of the catalytic 3004


Article

Accounts of Chemical Research catalyst stability as an additional factor of merit, leading to the notion of limiting turnover number, all that has been said at the end of section 5 is completely general and need not being repeated here. • The role of acids as cocatalyst in the push−pull catalytic process9c,28 (electrons from the reduced complex are pushed into the substrate and pulled out by the acid) discussed here is general and has a symmetrical counterpart in oxidations where bases play a similar role. Whatever the exact mechanism of their action, it is not surprising, even only for a local concentration effect, that installing such groups right in the catalyst molecule has beneficial effects. It is however worth analyzing the mechanisms that produce these effects as done here, where it has been possible to distinguish between the contributions of Hbonding and proton transfer. It is also important to bear in mind that reprotonation of the intramolecular proton transfer centers is a part of the process that may partake in rate limitations (as well as, symmetrically, the refreshment of the pendant bases in oxidative processes). • The favorable or unfavorable effect of substituents on catalysis, within a family of molecules, is a topic where important starting observations have been made, giving rise to first elements of rationalization. For example, in the comparison between CAT and FCAT (Figure 11), one could have easily guessed that the presence of the electronwithdrawing fluorine substituents would have a favorable effect on catalysis in the sense that it should push the catalyst standard potential toward positive values. But the same inductive effect was expected to have a series of unfavorable consequences: (i) decrease of the electron density on the iron at the 0 oxidation state entailing a lesser tendency to form an adduct with CO2, (ii) decrease of the electron density on the oxygen atoms of CO2 in the adduct resulting in less stabilization of the adduct by the local phenol groups, (iii) more difficult follow-up protonation by the same group, and (iv) taking also into account concerted coupling with the breaking of one of the C−O bonds. Despite this, FCAT appears to be a better catalyst than CAT in terms of catalytic Tafel plots, meaning that, grossly speaking, the effect of these unfavorable factors on the TOFmax is much less important than the positive shift of the standard potential of the catalyst. Clearly in this case, as well as in other families of catalysts (including the effect of electron-donating groups in catalysis of oxidations), more experimental material should be made available, mechanistically analyzed, and theoretically organized in order to obtain a reasonably comprehensive picture of this complex reactivity−structure issue. • The design of water-soluble catalysts of the CO2-to-CO conversion is obviously an important step toward practical applications in which the molecular catalyst would be immobilized onto the electrode surface by means of a negatively charged polymer and associated with a wateroxidation anode through a proton-exchange membrane. Manipulation of pH and buffer content of the solution so as to generate CO−H2 mixtures (syngas) in prescribed proportions is greatly facilitated by the opportunity to operate the catalysis in water. The same approach may be applied to other catalysts of CO2 electrochemical reduction. It may also be transposed to molecular catalysis of other reductive or oxidative processes in which electron

■

transfer and bond breaking (forming) events are in close association with proton transfer.

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest. Biographies Cyrille Costentin received his undergraduate education at Ecole Normale Supérieure in Cachan. He is, since 2007, Professor at the University Paris Diderot. His interests include mechanisms and reactivity in electron transfer chemistry with particular recent emphasis on electrochemical and theoretical approaches to proton-coupled electron transfer processes. Marc Robert was educated at the Ecole Normale Supérieure in Cachan. He is, since 2004, Professor at the University Paris Diderot. His interests include electrochemical, photochemical, and theoretical approaches of electron transfer reactions, as well as proton-coupled electron transfer processes in both organic chemistry and biochemistry. Jean-Michel Savéant received his education in the Ecole Normale Supérieure in Paris. He is, since 1985, Directeur de Recherche au Centre National de la Recherche Scientifique at the University Paris Diderot. His current research interests involve all aspects of molecular and biomolecular electrochemistry as well as mechanisms and reactivity in electron transfer chemistry and biochemistry.

■

ACKNOWLEDGMENTS Samuel Drouet, Guillaume Passard, and Arnaud Tatin are thanked for their participation to the work described in this Account.

■

REFERENCES

(1) (a) Lewis, N. S.; Nocera, D. G. Powering the planet: Chemical challenges in solar energy utilization. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 15729−15735; 2007, 104, 20142. (b) Gray, H. B. Solar fuel II. The quest for catalysts. Eng. Sci. 2009, 2, 27−31. (2) Aresta, M.; Dibenedetto, A.; Angelini, A. Catalysis for the valorization of exhaust carbon: From CO2 to chemicals, materials, and fuels. Technological use of CO2. Chem. Rev. 2014, 114, 1709−1742. (3) Lamy, E.; Nadjo, L.; Savéant, J. M. Standard potential and kineticparameters of electrochemical reduction of carbon-dioxide in dimethylformamide. J. Electroanal. Chem. Interfacial Electrochem. 1977, 78, 403−407. (4) Frese, K. W., Jr. In Electrochemical and Electrocatalytic Reactions of Carbon Dioxide, Sullivan, B. P.; Krist, K.; Guard, H. E., Eds.; Elsevier, New York, 1993; Chapter 6, pp 145−216. (5) Windle, C. D.; Perutz, R. N. Advances in molecular photocatalytic and electrocatalytic CO2 reduction. Coord. Chem. Rev. 2012, 256, 2562− 2570. (6) (a) Savéant, J.-M. Molecular catalysis of electrochemical reactions. Mechanistic aspects. Chem. Rev. 2008, 108, 2348−2378. (b) Costentin, C.; Robert, M.; Savéant, J.-M. Catalysis of the electrochemical reduction of carbon dioxide. Chem. Soc. Rev. 2013, 42, 2423−2436. (7) Schneider, J.; Jia, H.; Muckerman, J. T.; Fujita, E. Thermodynamics and kinetics of CO2, CO, and H+ binding to the metal centre of CO2 reduction catalysts. Chem. Soc. Rev. 2012, 41, 2036−2051. (8) Berardi, S.; Drouet, S.; Francas, L.; Gimbert-Surinach, C.; Guttentag, M.; Richmond, C.; Stoll, T.; Llobet, A. Molecular artificial photosynthesis. Chem. Soc. Rev. 2014, 43, 7501−7519. 3005


Article


illustrated by proton-assisted reductive cleavage of an O−O bond. Proc. Natl. Acad. Sci. U. S. A. 2011, 108, 8559−8564. (b) Costentin, C.; Robert, M.; Savéant, J.-M.; Tard, C. Breaking bonds with electrons and protons. Models and examples. Acc. Chem. Res. 2014, 47, 271−280. (18) Costentin, C.; Passard, G.; Robert, M.; Savéant, J.-M. Ultraefficient homogeneous catalyst for the CO2-to-CO electrochemical conversion. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 14990−14994. (19) Costentin, C.; Drouet, S.; Robert, M.; Savéant, J.-M. A local proton source enhances CO2 electroreduction to CO by a molecular Fe catalyst. Science 2012, 338, 90−94. (20) Costentin, C.; Robert, M.; Savéant, J.-M.; Tatin, A. An Efficient and Selective Molecular Catalyst for the CO2/CO Electrochemical Conversion in Water. Proc. Natl. Acad. Sci. U. S. A. 2015, 112, 6882− 6886. (21) Raebiger, J. W.; Turner, J. W.; Noll, B. C.; Curtis, C. J.; Miedaner, A.; Cox, B.; DuBois, D. L. Electrochemical reduction of CO2 to CO catalyzed by a bimetallic palladium complex. Organometallics 2006, 25, 3345−3351. (22) Wong, K.-Y.; Chung, W.-H.; Lau, C.-P. The effect of weak Bronstëd acids on the electrocatalytic reduction of carbon dioxide by a rhenium tricarbonyl bipyridyl complex. J. Electroanal. Chem. 1998, 453, 161−169. (23) Smieja, J. M.; Sampson, M. D.; Grice, K. A.; Benson, E. E.; Froehlich, J. D.; Kubiak, C. P. Manganese as a substitute for rhenium in CO2 reduction catalysts: The importance of acids. Inorg. Chem. 2013, 52, 2484−2491. (24) Chen, Z.; Chen, C.; Weinberg, D. R.; Kang, P.; Concepcion, J. J.; Harrison, D. P.; Brookhart, M. S.; Meyer, T. J. Electrocatalytic reduction of CO2 to CO by polypyridyl ruthenium complexes. Chem. Commun. 2011, 47, 12607−12609. (25) Sampson, M. D.; Nguyen, A. D.; Grice, K. A.; Moore, C. E.; Rheingold, A. L.; Kubiak, C. P. Manganese catalysts with bulky bipyridine ligands for the electrocatalytic reduction of carbon dioxide: Eliminating dimerization and altering catalysis. J. Am. Chem. Soc. 2014, 136, 5460−5471. (26) Costentin, C.; Passard, G.; Savéant, J.-M. Benchmarking of homogeneous electrocatalysts. Overpotential, turnover frequency, limiting turnover number. J. Am. Chem. Soc. 2015, 137, 5461. (27) Steffey, B. D.; Curtis, C. J.; DuBois, D. L. Electrochemical reduction of CO2 catalyzed by a dinuclear palladium complex containing a bridging hexaphosphine ligand: Evidence for cooperativity. Organometallics 1995, 14, 4937−4943. (28) Electrons from the Fe0 complex are pushed in CO2 and pulled out by the acid with breaking of one C−O bond.9c

(9) (a) Hammouche, M.; Lexa, D.; Savéant, J.-M.; Momenteau, M. Catalysis of the electrochemical reduction of carbon dioxide by iron(″0″) porphyrins. J. Electroanal. Chem. Interfacial Electrochem. 1988, 249, 347−351. (b) Bhugun, I.; Lexa, D.; Savéant, J.-M. Ultrae6fficient selective homogeneous catalysis of the electrochemical reduction of carbon dioxide by an iron(0) porphyrin associated with a weak broensted acid cocatalyst. J. Am. Chem. Soc. 1994, 116, 5015− 5016. (c) Bhugun, I.; Lexa, D.; Savéant, J. M. Catalysis of the electrochemical reduction of carbon dioxide by iron(o) porphyrins: Synergystic effect of weak bronsted acids. J. Am. Chem. Soc. 1996, 118, 1769−1776. Lewis acids play a similar role: (d) Hammouche, M.; Lexa, D.; Momenteau, M.; Savéant, J.-M. Chemical catalysis of electrochemical reactions. Homogeneous catalysis of the electrochemical reduction of carbon dioxide by iron (″0″) porphyrins. Role of the addition of magnesium cations. J. Am. Chem. Soc. 1991, 113, 8455− 8466. (e) Bhugun, I.; Lexa, D.; Savéant, J. M. Catalysis of the electrochemical reduction of carbon dioxide by iron(0) porphyrins. Synergistic effect of Lewis acid cations. J. Phys. Chem. 1996, 100, 19981− 19985. (10) (a) Gressin, J. C.; Michelet, D.; Nadjo, L.; Savéant, J. M. Electrochemical reduction of carbon-dioxide in low proton media. New J. Chem. 1979, 3, 545−554. (b) Gennaro, A.; Isse, A. A.; Severin, M.-G.; Vianello, E.; Bhugun, I.; Savéant, J.-M. Mechanism of the electrochemical reduction of carbon dioxide at inert electrodes in media of low proton availability. J. Chem. Soc., Faraday Trans. 1996, 92, 3963−3968. (c) Gennaro, A.; Isse, A. A.; Savéant, J. M.; Severin, M. G.; Vianello, E. Homogeneous electron transfer catalysis of the electrochemical reduction of carbon dioxide. Do aromatic anion radicals react in an outer-sphere manner? J. Am. Chem. Soc. 1996, 118, 7190−7196. (11) (a) Hawecker, J.; Lehn, J. M.; Ziessel, R. Electrocatalytic reduction of carbon-dioxide mediated by re(bipy) (co)3cl(bipy = 2,2′-bipyridine). J. Chem. Soc., Chem. Commun. 1984, 328−330. (b) Beley, M.; Collin, J. P.; Ruppert, R.; Sauvage, J. P. Nickel(ii)-cyclam: An extremely selective electrocatalyst for reduction of carbon dioxide in water. J. Chem. Soc., Chem. Commun. 1984, 1315−16. (c) Beley, M.; Collin, J. P.; Ruppert, R.; Sauvage, J. P. Electrocatalytic reduction of carbon dioxide by nickel cyclam2+ in water: Study of the factors affecting the efficiency and the selectivity of the process. J. Am. Chem. Soc. 1986, 108, 7461−7467. (12) (a) Savéant, J.-M.; Vianello, E. In Recherches sur les courants catalytiques en polarographie oscillographique à balayage linéaire de tension. Étude théorique; Longmuir, I. S., Ed.; Advances in Polarography; Pergamon Press: Cambridge, U.K., 1959; pp 367−374. (b) Savéant, J.M.; Vianello, E. Potential-sweep chronoamperometry: Kinetic currents for first-order chemical reaction parallel to electron-transfer process (catalytic currents). Electrochim. Acta 1965, 10, 905−920. (13) Costentin, C.; Savéant, J.-M. Multielectron, multistep molecular catalysis of electrochemical reactions: Benchmarking of homogeneous catalysts. ChemElectroChem 2014, 1, 1226−1236. (14) (a) Costentin, C.; Drouet, S.; Robert, M.; Savéant, J.-M. Turnover numbers, turnover frequencies and overpotential in molecular catalysis of electrochemical reactions. Cyclic voltammetry and preparative-scale electrolysis. J. Am. Chem. Soc. 2012, 134, 11235−11242. (b) Costentin, C.; Drouet, S.; Passard, G.; Robert, M.; Savéant, J.-M. Proton-coupled electron transfer cleavage of heavy-atom bonds in electrocatalytic processes. Cleavage of a C-O bond in the catalyzed electrochemical reduction of CO2. J. Am. Chem. Soc. 2013, 135, 9023−9031. (c) Revisiting the analysis in ref 14b, it appears that the correct value of E0cat is −1.453 instead of −1.428 V vs SHE then used, thus leading to the FOW analyses in Figure 2b′ and the rate data in Figure 2c. (15) Bhugun, I.; Lexa, D.; Savéant, J.-M. Homogeneous catalysis of electrochemical hydrogen evolution by iron(0) porphyrins. J. Am. Chem. Soc. 1996, 118, 3982−3983. (16) Costentin, C.; Passard, G.; Robert, M.; Savéant, J.-M. Pendant acid-base groups in molecular catalysts: H-bond promoters or proton relays? Mechanisms of the conversion of CO2 to CO by electrogenerated iron(0) porphyrins bearing prepositioned phenol functionalities. J. Am. Chem. Soc. 2014, 136, 11821−11829. (17) (a) Costentin, C.; Hajj, V.; Robert, M.; Savéant, J.-M.; Tard, C. Concerted heavy-atom bond cleavage and proton and electron transfers 3006


Current Issues in Molecular Catalysis Illustrated by Iron Porphyrins as

Recommend Documents