Mechanistic Insight and Overpotential Tuning in a Series of

Subscriber access provided by Thompson Rivers University | Library

Article

Electrocatalytic H2O Reduction with f-Elements: Mechanistic Insight and Overpotential Tuning in a Series of Lanthanide Complexes Dominik Pascal Halter, Chad T Palumbo, Joseph W. Ziller, Milan Gembicky, Arnold L. Rheingold, William J Evans, and Karsten Meyer J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b11532 • Publication Date (Web): 29 Jan 2018 Downloaded from http://pubs.acs.org on January 29, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of the American Chemical Society is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

Electrocatalytic H2O Reduction with f-Elements: Mechanistic Insight and Overpotential Tuning in a Series of Lanthanide Complexes Dominik P. Halter,1 Chad T. Palumbo,2 Joseph W. Ziller,2 Milan Gembicky,3 Arnold L. Rheingold,3 William J. Evans,2* and Karsten Meyer1* 1

Department of Chemistry and Pharmacy, Inorganic Chemistry, University Erlangen-Nürnberg, Egerlandstraße 1, D-91058 Erlangen, Germany 2 Department of Chemistry, University of California, Irvine, California 92697-2025, United States 3 Department of Chemistry and Biochemistry, University of California, San Diego, 9500 Gilman Drive, MC 0332, La Jolla, California 92093, United States ABSTRACT: Electrocatalytic energy conversion with molecular f-element catalysts is still in an early phase of its development. We here report detailed electrochemical investigations on the recently reported trivalent lanthanide coordination complexes [((Ad,MeArO)3mes)Ln] (1–Ln), with Ln = La, Ce, Pr, Nd, Sm, Gd, Dy, Er, Yb that were now found to perform as active electrocatalysts for the reduction of water to dihydrogen. Reactivity studies involving complexes 1–Ln and the Ln(II) analogues [K(2.2.2-crypt)][((Ad,MeArO)3mes)Ln] (2–Ln) suggest a reaction mechanism that differs significantly from the reaction pathway found for the corresponding uranium catalyst [((Ad,MeArO)3mes)U] (1–U). While complexes 1–Ln activate water after electrochemical 1 e– reduction to yield the reduced 2–Ln species via a radical pathway, the 5f analogue 1–U directly reduces H2O via a 2 e– pathway. The electrocatalytic H2O reduction by complexes 1–Ln is initiated by the respective Ln(III)/Ln(II) redox couples, which gradually turn to more positive values across the Ln series. This correlation has been exploited to tune the catalytic overpotential of water reduction by choice of the lanthanide ion. Kinetic studies of the 1–Ln series were performed to elucidate correlations between overpotential and turnover frequencies of the 4f-based electrocatalysts.

Introduction Electrocatalysis is a powerful tool to enhance the efficiency of electrochemical processes, such as methanol oxidation in fuel cells, or H2O to dihydrogen reduction for energy storage.1–6 Numerous examples of homogeneous transition metal catalysts for electrochemical hydrogen production have been reported, which all contributed to a detailed understanding of reaction mechanisms and technical processes.7–12 The first and thus far only example of actinide electrocatalysis involves the trivalent uranium complex [((Ad,MeArO)3mes)U] (1–U), with ((Ad,MeArO)3mes)3– = 6,6',6''-((2,4,6-trimethylbenzene-1,3,5-triyl)tris(methylene))tris(2-(adamantan-1-yl)-4-methylphenolate) that mediates H2O reduction.6 This molecular U(III) catalyst provided valuable and unprecedented insight into f-element reactivity and homogenous catalysis, and will help to pursue further potential applications in f-element catalysis. Spectroscopic and synthetic analyses of several intermediates involved in the catalytic cycle suggest a mechanism with two key steps; namely, the chemical reduction of H2O to H2 (Scheme 1, reaction 1), and the successive electrochemical one-electron reduction to regenerate the U(III) catalyst (Scheme 1, reaction 2).6,13 However, some of the species proposed in the catalytic cycle, such as the initially formed U(III)–aquo complex, or the product of the subsequent oxidative addition, U(V)(OH)(H), could not be isolated due to the high reactivity and fleeting nature of these intermediates. In order to further advance our understanding of the electrocatalytic H2O reduction by f-elements in the mesitylene-anchored tris(aryloxide) ligand environment ((Ad,MeArO)3mes)3–,14 we have examined lanthanide-mediated electrocatalytic dihydrogen production from water.

Scheme 1. The two key steps of electrocatalytic H2O reduction by the U(III) catalyst [((Ad,MeArO)3mes)U], namely 1) chemical oxidation of U(III) and 2) electrochemical reduction of U(IV).

For a number of applications, lanthanides provide certain advantages over the actinides. With the exception of Pm, the lanthanides are not radioactive and are known for a consistent coordination chemistry, which allows for convenient handling.15 Utilizing the large range of available lanthanide metals allows the synthesis of an entire series of complexes with an identical, consistent ligand environment. With such a series in hand, the variable redox potentials of the individual lanthanides can be investigated in detail in order to explore and tune electrocatalytic performance. Therefore, the overpotential for catalytic H2 production from water is adjustable by choice of the lanthanide metal center. In order to minimize the overpotential, the Ln(III)/Ln(II) couple should be close to the thermodynamic potential of H2O reduction. In this context, the correlation of overpotential and chemical reactivity, which can aid the design of new catalytically active f-element complexes, is of particular interest. In addition, fleeting intermediates, such as the aforementioned U–OH2 complex involved in the catalytic cycle, are more likely to be isolated with the less reducing lanthanides; thus, providing further insight ACS Paragon Plus Environment

Journal of the American Chemical Society 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 9

into the reaction mechanism. Moreover, a comparison of the lanthanide reactivity reported herein, together with data of the previously described actinide catalyst 1–U, provides deeper insight into f-element electrocatalysis in general, and may open opportunities to develop new systems in the future. Results and discussion Reactivity studies on the structurally characterized trivalent lanthanide complexes of the general formula [((Ad,MeArO)3mes)Ln] (1–Ln), with Ln = La, Ce, Pr, Nd, Sm, Gd, Dy, Er, and Yb,16,17 represent the basis of this survey on electrocatalytic H2O reduction. Analogous to the uranium electrocatalyst 1–U, the reduced lanthanide species [K(2.2.2-cryptand)][(Ad,MeArO)3mes)Ln] (2–Ln) are accessible,16,17 principally allowing for well-defined lanthanide-mediated redox catalysis. This application of lanthanide redox chemistry is of high interest given that lanthanides are predominantly known for Lewis-acid-based catalysis.18–20 Complexes 1–Ln are prepared in a protonolysis reaction of the trivalent lanthanide complexes [Ln(N(SiMe3)2)3] (N(SiMe3)2 = 1,1,1,3,3,3-hexamethyldisilazane) with the protonated ligand (Ad,MeArOH)3mes.16,17 The respective divalent lanthanide compounds 2–Ln were obtained by reduction of 1–Ln with KC8 in the presence of 2.2.2-cryptand. While the divalent lanthanide complexes 2–Ln can reduce H2O to form 0.5 H2 and trivalent hydroxo complexes, as shown with [K(2.2.2-crypt)][(Ad,MeArO)3mes)Nd(OH)] (3–Nd), the parent Ln(III) complexes 1–Ln can accommodate water as a neutral ligand in the equatorial plane, as confirmed by the Nd(III) aquo complex [(Ad,MeArO)3mes)Nd(H2O)], 1–Nd–H2O. The identity and connectivities of these species were confirmed by XRD analysis. Given the steric hindrance of the adamantyl-derivatized tris(aryloxide) ligand system, the coordination of an additional ligand in the equatorial plane is unusual, but not unprecedented. Uranium complexes of the ((Ad,MeArO)3mes)3– ligand system tend to coordinate THF in the equatorial plane of the [((Ad,MeArO)3mes)U(X)(THF)] complexes (X = F, Cl, Br, I, and OH).6,21 Based on spectroscopic results and DFT analyses, a comparable equatorial coordination of H2O to the uranium catalyst [((Ad,MeArO)3mes)U] was previously proposed as the initial step of water reduction.6,13 The molecular structure of the aquo complex 1–Nd–H2O, shown in Figure 1, now provides further evidence for this proposed substrate binding step. Structural Analysis. In the course of our studies, the two trivalent neodymium complexes 1–Nd–H2O and 3–Nd were discovered and structurally characterized. Interestingly, and likely due to the labile nature of the aquo ligand, complex 1–Nd–H2O cocrystallizes with two equivalents of its desolvated form 1–Nd. A direct comparison of 1–Nd to the 5f homologue 1–U reveals significantly longer metal–arenecentr. distances for 1–Nd (2.489 Å) than for 1–U (2.353 Å). This is due to a notably lower degree of covalency in 4f vs. 5f metal-ligand bonding.16,22 Upon one-electron reduction of complex 1–Nd, to form the reactive Nd(II) species 2–Nd, the average Nd–OAr bond is elongated by 0.05 Å to 2.237(4) Å, and the Nd–arenecentr. distance is shortened to 2.366 Å, indicative of a more electron-rich metal center. Both XRD analyses and DFT calculations are in agreement with a metal-centered reduction to form a genuine Nd(II) ion.16

Figure 1. Top: Molecular structure of the previously reported Nd(III) precursor 1–Nd for visual comparison.16 Middle: Molecular structure of the new aquo complex 1–Nd–H2O in crystals of 1–Nd–H2O • 2 1–Nd; selected bond lengths of 1–Nd–H2O: Nd– OH2: 2.479(9) Å, average Nd–OAr: 2.20(2) Å, Nd–mescentr.: 2.489 Å, Ndoop –0.265 Å. Bottom: Molecular structure of the new hydroxo complex 3–Nd in crystals of 3–Nd; selected bond lengths of 3–Nd: Nd–OH: 2.156(5) Å, average Nd–OAr: 2.243(3) Å, Nd– mescentr.: 2.754 Å, Ndoop –0.012 Å. Hydrogen atoms, except for H2O and OH protons, and in case of 1–Nd–H2O co-crystallized 1–Nd, are omitted for clarity; all ellipsoids are presented at the 50% probability level.

ACS Paragon Plus Environment

Page 3 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society Structural analysis of the newly synthesized equatorial aquo complex 1–Nd–H2O, reveals that the coordination of H2O barely influences the coordination environment of the Nd(III) center. Upon water coordination (Nd–OH2O = 2.479(9) Å), the average U– OAr bond length of 2.20(2) Å, the NdOOP value of –0.265 Å, and the Nd–arenecentr. distance of 2.489 Å, remain unchanged compared to complex 1–Nd.16 In contrast, the axial coordination of OH– in 3–Nd introduces severe changes to the molecular structure. The average Nd–OAr bond length in 3–Nd is increased to 2.243(3) Å and the Nd ion is located in the plane of the three aryloxide oxygens, resulting in a negligible out of plane shift NdOOP of –0.012 Å and a significant elongation of the Nd–arenecentr. distance to 2.754 Å. These structural differences between the 1–Nd–H2O and 3–Nd are caused by the affinity of hard lanthanide ions towards hard donors, such as OH–. A structural comparison of the three Nd(III) complexes 1– Nd, 1–Nd–H2O, and 3–Nd is summarized in Table 1. Table 1. Selected bond lengths (Å) of complexes 1–Nd–H2O, and 3–Nd and complexes 1–Nd. Nd–Ox refers to the bond distance to the aquo or the hydroxo ligand, respectively.

1–Ua 1–Ndb 2–Ndb 1–Nd–H2O 3–Nd

Nd– OAr(ave)

Nd–Ox

Nd–arenecentr. NdOOP

2.169(3) 2.19(1) 2.237(4) 2.20(2) 2.243(3)

– – – 2.479(9) 2.156(5)

2.353 2.489 2.366 2.489 2.754

redox couples, to simplify all further discussions in this work. Figure 2 clearly shows that the reduction potentials of complexes 1–Ln shift gradually to more positive potentials from 1–La to 1– Yb. For complexes 1–Ln with lanthanides heavier than Sm, however, slight inconsistencies of this trend were observed (see Table 2). In contrast to the reversible reductions, the observed oxidations ranging from 0.12 to 0.59 V are irreversible for all investigated lanthanide complexes 1–Ln. With the exception of 1–Ce, the potential of this irreversible oxidation shifts to more positive potentials across the period from La to Er, which is in line with increasing Lewis acidity of the metals. An oxidation wave is not observed for 1–Yb, likely because it is shifted beyond the electrochemical solvent window of THF (see Figure 2 and Table 2). Given that irreversible oxidation was also observed for the 4f 0 La(III) complex, 1–La, which has no valence electrons, the oxidation event is likely ligand-based for all complexes 1–Ln. Transition metal complexes in similar phenolate ligand environments, such as [((tBu,tBuArO)tacn)M(OAc)2] (with M = Ga, Fe, Co), have been reported by Wieghardt and coworkers. These complexes are oxidized to form the corresponding phenoxyl radicals at potentials that compare well to those determined herein for 1–Ln.25 In analogy, the ligand-centered oxidation of complexes 1–Ln was tentatively assigned to the generation of phenoxyl radicals.

–0.475 –0.268 –0.530 –0.265 –0.012

a original data reported in reference [14]. b original data reported in reference [16]. Electrochemical Studies. In view of the recently reported electrocatalytic H2O reduction, mediated by uranium complex 1–U, the electrochemical behavior of complexes 1–Ln was also investigated (Figure 2). The results provide a valuable extension to previous literature reports of Ln(III)/Ln(II) redox potentials that are mainly based on calculations and molten salt data of simple halide salts. Particularly, for Sm and Yb, the literature-reported reduction potentials (–1.50 and –1.18 V, vs. NHE respectively)23,24 deviate significantly from those determined for the arene-anchored tris(aryloxide) complexes presented here. The cyclic voltammetry experiments with complexes 1–Ln revealed two characteristic signals; namely, a reversible reduction event, ranging from –3.08 V for 1–La to –2.12 V for 1–Yb, and an irreversible oxidation with peak potentials ranging from 0.01 V for 1–Ce to 0.59 V for 1–Er (all potentials are reported vs. Fc+/Fc). Apparent from the symmetry of the redox wave, the reduction becomes more reversible when the metal center is substituted across the lanthanide row from 1–La to 1–Nd, and is fully reversible for all remaining Ln complexes of the series (SI). The limited reversibility of the redox wave observed for complexes 1–La, 1–Ce, and 1–Pr is in accordance with the limited stability of the isolated reduced complexes 2–Ln. These findings complement a parallel synthetic study, revealing that the chemical reduction of the Ln(III) complexes 1–La, 1–Ce, and 1–Pr is ligand centered,17 whereas complexes 1–Nd to 1–Yb form genuine and stable Ln(II) species.16 Despite this situation, the reduction waves of complexes 1–Ln will only be referred to as Ln(III)/Ln(II)

Figure 2. Cyclic voltammetry analysis of selected complexes 1– Ln, with trace for solvent blank, illustrating the increasing reversibility and shifting reduction potential across the lanthanides series. The inset shows a scan rate-dependent analysis of the Nd(III)/Nd(II) couple of complex 1–Nd. Data were recorded with 0.1 M TBAPF6 in THF with a glassy carbon working electrode.



Page 4 of 9

Table 2. Redox potentials (V) of the Ln(III)/Ln(II) redox couples of all lanthanides (except La, Ce, Pr, for which the reductions are ligand-based), and peak potentials (V) of the irreversible ligand oxidations of compounds 1–La to 1–Yb; measured with a glassy carbon working electrode in 0.1 M TBAPF6 solutions (THF) and referenced vs. Fc+/Fc. Ln

Reduction [V]

Ligand Oxidation [V]

La Ce Pr Nd Sm Gd Dy Er Yb

−3.08 −2.93 −2.96 −2.93 −2.60 −2.90 −2.86 −2.87 –2.12

0.12 0.01 0.20 0.23 0.35 0.54 0.55 0.59 –

Having characterized the electrochemical properties of all complexes 1–Ln, their catalytic performance was investigated. Remarkably, upon H2O addition to the electrolyte (0.22 M), electrocatalytic H2O reduction with concomitant production of H2 (determined by GC-TCD) was observed for all complexes 1–Ln. The catalytic overpotential of the different electrocatalysts 1–Ln is dependent on the redox potential of the respective Ln(III)/Ln(II) couples (Figure 3, and SI). The observed catalytic onset potentials, herein defined as the potentials at which the catalytic current for H2 production reaches three times the background current, are roughly 0.3 V more positive than the respective half-wave potentials of 1–Ln reduction without substrate. The individual catalytic onset potentials are listed in Table 3. Clearly, the correlation of overpotentials and reduction potentials illustrates that the catalysis is facilitated by the unimpaired complexes 1–Ln. Control experiments to exclude catalytic effects of potential decomposition products were performed with the free aryloxide ligand as well as the trivalent lanthanide salt SmI3. These measurements, carried out under analogous conditions to the catalytic experiments, revealed that pure samples of the free ligand are electrochemically silent, whereas SmI3 features an irreversible Sm(III)/Sm(II) reduction at –2.81 V, but no catalytic activity. Mechanistic Studies. The reaction mechanism of the lanthanidemediated electrocatalytic H2O reduction was studied by synthetic, spectroscopic, and electrochemical methods. Based on the combined results, the reaction sequence presented in Scheme 2 is proposed to facilitate the catalytic H2O reduction. The electrochemical reduction of 1–Ln is the first step, forming reactive Ln(II) species that chemically reduce H2O to form the Ln(III) hydroxo complexes [((Ad,MeArO)3mes)Ln(OH)]– and 0.5 equiv. H2 in the second step. Subsequently, the complex anion [((Ad,MeArO)3mes)Ln(OH)]– is protonated by free H2O to form the charge-neutral Ln(III) aquo complexes [((Ad,MeArO)3mes)Ln(H2O)] (1–Ln– H2O). In the last step, the H2O ligand dissociates from the 1–Ln– H2O complex to regenerate the catalyst, 1–Ln. The reaction cascade of the lanthanide catalyzed H2O

Figure 3. Electrocatalytic reduction of H2O (0.22 M in THF) representatively shown for the lanthanide catalysts 1–Nd (black & grey) and 1–Sm (red & orange), illustrating a correlation between the catalytic overpotential and reduction potential of complexes 1–Ln. Data recorded in 0.1 M TBAPF6 electrolyte solutions (THF) with a glassy carbon working electrode. reduction, as presented in Scheme 2, deviates significantly from the two-step sequence reported for the analogous uranium catalyst (Scheme 1). In contrast to the trivalent uranium catalyst, the trivalent lanthanides do not reduce H2O directly, but initially form aquo complexes 1–Ln–H2O. Consequently, under catalytic conditions with excess H2O, the resting state of complexes 1–Ln remains Ln(III), whereas the resting state of catalyst 1–U is U(IV). Accordingly, in the lanthanide case, the electrochemical reduction is the required step to generate the active Ln(II) catalyst [((Ad,MeArO)3mes)Ln]– (2–Ln) (Scheme 3). For further insight into the reactivity of the catalytically active Ln(II) species, complexes 2–Ln were independently synthesized and characterized. Control experiments studying the reaction of complexes 2–Ln with 1 equiv. of H2O resulted in the expected formation of the Ln(III) hydroxo complexes [K(2.2.2-crypt)][((Ad,MeArO)3mes)Ln(OH)] (3–Ln) and dihydrogen (Scheme 3, reaction c).

Scheme 2. Proposed steps of electrocatalytic H2O reduction with the Ln(III) catalysts [((Ad,MeArO)3mes)Ln] (1–Ln). In case of 3–Nd, formation of the trivalent Ln–OH species was confirmed by single-crystal XRD analysis, bulk elemental analysis, and 1H NMR spectroscopy (SI). In contrast to the reported two-electron oxidative addition of H2O to [((Ad,MeArO)3mes)U], the proposed one-electron reduction mechanism of the lanthanides implies a radical pathway for H2O reduction, ultimately generating H• radicals as the source of the detected H2. Indeed, in situ EPR experiments following the stoichiometric H2O reduction by 2–Nd confirmed the formation of organic radicals due to


Page 5 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society H• scavenging in the stoichiometric hydrogen production process (SI). While the stoichiometric experimental data clearly identify the hydrogen-producing step of the reaction between Nd(II) complex 2–Nd and H2O as a one-electron process, the possibilities for the actual H2 formation mechanism are diverse. Given the steric bulk of the ligand, a dimeric pathway – with two metal centers simultaneously providing one electron – seems unlikely. Instead, a simple charge transfer to solvent (CTTS) mechanism may apply.26 Analogous to literature examples of anionic transition metal complexes, such as [IrCl6]3–, a free proton may bind to the complex anion of 2–Nd, possibly in proximity to the aryloxide oxygens, subsequently forming 1–Nd and H•.27 The coordination of a proton to tris-aryloxide ligands was actually identified for a related Co(II) complex, namely [((tBu,tBuArO)2(tBu,tBuArOH)N)CoII(NCCH3)] (SI). The oxidative addition of the proton to the metal site in 2–Ln is unlikely, since it requires the formation of Ln(IV) species. This would imply a deviating reactivity of complex 1–Ce, given that cerium – unlike the other lanthanides – has an accessible Ln(IV) oxidation state. However, our findings show a periodic trend with no exception for Ce. Another possibility is a simple outer-sphere electron transfer to solvated protons known for H2O reduction by alkali metals.28 Under electrocatalytic conditions, however, yet another, even different mechanism including a reaction cascade of ECCE or ECEC type may apply.29 The mechanistic studies reported herein focused on elucidating the regeneration of catalysts 1–Ln from 3–Ln. While the uranium catalyst 1–U is regenerated by reductively-induced OH– elimination (Scheme 1, eq. 2),6 another mechanism must be operative in case of the lanthanide analogues. Representative of all Ln(III) hydroxo species, complexes 3–Nd and 3-Gd were reduced with KC8 to yield lightly colored THF solutions, strongly deviating from those of the intensely-colored 2–Ln species.

Scheme 3. Synthetic steps towards the mechanistically relevant species formed during H2O reduction with the lanthanide

catalysts 1–Ln. Synthesis of a) the trivalent complexes 1–Ln with (Ad,MeArOH)3mes, b) the reactive, divalent complexes 2–Ln, and c) the trivalent hydroxo complexes 3–Ln with concomitant formation of H2. d) Protonation of hydroxo complexes 3–Ln to form the trivalent aquo species 1–Ln–H2O. e) Dynamic equilibrium of 1–Ln and the corresponding trivalent aquo complexes 1– Ln–H2O in solution. Since divalent Gd is EPR active at room temperature, 2–Gd was expected to yield an isotropic EPR signal centered at g = 1.98.16 The in situ EPR spectroscopic analysis of 3–Gd reduction with KC8, however, did not yield the expected signal of 2–Gd; thus, rendering the reductively-induced elimination of OH– unlikely. Instead, in the presence of free H2O, the protonation of the apical hydroxo ligand of the 3–Ln species was identified by 1H NMR spectroscopy. The protonation was demonstrated by titrating THF-d8 solutions of 3–Nd with D2O, resulting in gradual disappearance of the characteristic OH resonance at d = 5.42 ppm. The spectroscopically characterized but labile aquo species 1–Ln– H2O could not be isolated in bulk, likely due to the weak interaction of the H2O ligand with the trivalent lanthanide ions. This weak coordination is beneficial to regenerate the lanthanide catalysts via H2O dissociation; thereby, closing the catalytic cycle. Electrocatalytic reduction of H2O was also observed when 3–Ln species were employed as catalysts instead of 1–Ln, strongly suggesting that the hydroxo complexes 3–Ln are participating in the catalytic cycle. The chemical synthesis and interconversion of all relevant species is summarized in Scheme 3. Catalyst Performance. In addition to mechanistic and structural insights, the performance of the different lanthanide catalysts was investigated and compared. Figure 3 illustrates that the catalytic onset potential is influenced by the Ln(III)/Ln(II) couple (Table 2). Accordingly, the catalytic overpotential (h) can be adjusted by choice of the lanthanide ion. Obviously, the redox potential of the Ln(III)/Ln(II) couple, defining the catalytic overpotential, is also a measure of the driving force of the chemical H2O reduction step (Scheme 2, eq. 2). Consequently, a decreased overpotential of the electrochemical reaction step correlates with a reduced reaction rate for the consecutive chemical reaction step.30 While experimentally determined half-wave potentials of electrocatalysts typically directly translate to catalytic overpotentials, analysis of the 1–Ln series suggests that this correlation does not strictly apply for f-element catalysts presented herein. The extensive series of complexes 1–Ln provides therefore a rare opportunity to characterize the unique behavior of f-elements in catalysis. In the context of this work, we revisited our previous studies on the determination of the thermodynamic H2O reduction potential in THF. In a series of open circuit potential measurements (SI), using literature reported procedures and setups,31,32 we now determined the thermodynamic H2O reduction potential at –0.989 V vs Fc+/Fc in THF. Hence, based on the experimentally determined reduction potentials of the complexes reported herein (Table 2), the catalyst with the lowest predicted overpotential for H2O reduction should be 1–Yb with a theoretical value of h = 1.13 V. However, 1–Yb unexpectedly does not catalyze H2O reduction in the region of this theoretically calculated overpotential. Instead, the catalyst with the lowest experimentally confirmed overpotential in the series is 1–Sm, for which an onset potential of 2.30 V vs. Fc+/Fc was determined (corresponding to 1.31 V on the overpotential scale). Reasons for the decreased catalytic



activity of complex 1–Yb are diverse and will be discussed in the following.

Figure 4. Overlay CV plot of the two Nd(III) complexes [((Ad,MeArO)3mes)Nd] (1–Nd, black) and [K(2.2.2-cryptand)][((Ad,MeArO)3mes)Nd(OH)] (3–Nd, red). The CV was recorded with a glassy carbon working electrode in 0.1 M TBAPF6 solutions (THF) and referenced vs. Fc+/Fc . One important aspect is the increasing Lewis acidity across the period from 1–La to 1–Yb that stabilizes the catalytically relevant Lewis base adducts,33,34 such as complexes 1–Ln–H2O and 3–Ln, thereby decreasing the observed reaction rates for the late lanthanides. Additionally, the coordination of Lewis base ligands, such as OH–, shifts the Ln(III)/Ln(II) reduction potentials of complexes 1–Ln to more negative values. A comparison of the cyclic voltammograms obtained for the Nd(III) complexes 1– Nd and 3–Nd (Figure 4) shows that in case of hydroxo complex 3–Nd, the Ln(III)/L(II) reduction is no longer observed, because it is negatively shifted beyond the solvent window of THF. Under catalytic conditions, however, the Nd(III)–OH complex 3–Nd performs at an overpotential and activity identical to 1–Nd (SI), strongly suggesting that in the presence of free H2O, complexes 3–Ln is converted to 1–Ln (in equilibrium with 1–Ln–H2O) as illustrated in Scheme 3. A significant effect on the Ln(III)/Ln(II) reduction potential of complexes 1–Ln was also observed upon coordination of H2O, forming complexes 1–Ln–H2O. Due to the catalytic activity of complexes 1–Ln in the presence of H2O, this potential shift cannot be directly observed, but can be estimated from catalysis data. It is well known that the onset of electrocatalytic substrate reduction is generally expected at potentials slightly less negative than the correlated E1/2 potentials of the catalyst without substrate.35 The correlation of experimentally determined E1/2 values of complexes 1–Ln without substrate and their respective catalytic onset potentials, however, clearly reveals that upon traversing the series, the catalytic onset potential gradually shifts towards more and more negative values with respect to E1/2 of isolated complexes 1–Ln (SI). A suitable explanation for this observation is the increasing Lewis acidity and concomitant stability of 1–Ln– H2O species for the late lanthanides. Accordingly, the 1–Ln and 1–Ln–H2O equilibrium gradually shifts towards the aquo complex 1–Ln–H2O, which becomes the predominant species for the late lanthanides. The fact that onset potentials generally precede the E1/2 values of the catalyst’s resting state, implies that the

Page 6 of 9

reduction potentials of complexes 1–Ln–H2O (Table 3) are more negative compared to 1–Ln. Likewise, the stability of redox inactive hydroxo complexes 3–Ln, formed during the catalytic cycle, increases with increasing Lewis-acidity of the catalysts, inhibiting their activity. Due to the long-lived and inactive resting state, the in operando concentration of active catalyst is expected to be considerably lower than the initially provided catalyst loading of complexes 1–Ln. As a result, the Lewis acidity of the different lanthanides directly influences the experimentally observed overpotential and activity for H2O reduction with catalysts 1–Ln. Relative reaction rates kobs of electrocatalytic H2O reduction with 1–Ln were estimated by foot-of-the-wave analyses (FOWA, see SI),35–37 and are in accordance with a decreasing activity upon traversing the lanthanide series from 1–La to 1–Yb. Overall, the obtained rates were one order of magnitude greater for the early lanthanides than for the late members of the series. Given that the standard potentials of the catalysts 1–Ln in the presence of substrate cannot be determined precisely, approximations and simplifications were made. Accordingly, the obtained reaction rates satisfactorily allow the comparison of the relative catalyst performance within the series 1–Ln, but cannot serve as literature standards. Table 3. Tabulated values of Ln(III)/Ln(II) reduction potentials (V) estimated in the presence of substrate, and experimentally observed onset potentials of H2O reduction by 1–Ln. Half-wave potentials were determined from the current response under catalytic conditions as described in the SI. The reported catalytic onset potentials were determined as the potentials at which the catalytic current reaches three times the background current. Ln La Pr Ce Er Dy Gd Nd Sm Yb

E½ in presence of substrate [V] –3.21 –3.11 –2.99 –2.99 –2.98 –2.95 –2.94 –2.87 –––

catalytic onset potential [V] –2.87 –2.66 –2.82 –2.64 –2.59 –2.67 –2.57 –2.30 –––

Altogether, the presented results demonstrate that the development of effective catalysts cannot merely focus on lowering the overpotential alone. Instead, it must be considered that lowering the overpotential may also influence reaction mechanisms and chemical reaction rates. All three parameters must be well balanced for the successful design of future catalysts. Summary In summary, the electrochemical properties and chemical reactivity of the trivalent lanthanide series [((Ad,MeArO)3mes)Ln] (1–Ln) were investigated. Complexes 1–Ln were successfully employed for electrocatalytic H2 production from H2O. The overpotential for catalytic H2O reduction is dependent on the respective Ln(III)/Ln(II) reduction waves, which allows for reactivity tuning


Page 7 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society by choice of the lanthanide ion. While the less reducing lanthanides form catalysts with low-overpotential, they exhibit lower driving force for the chemical substrate conversion, ultimately leading to a kinetic termination of catalytic activity as seen in the case of 1–Yb. Mechanistic studies were performed on the Nd complexes, representative for the entire series. The Nd(II) complex [K(2.2.2-cryptand)][((Ad,MeArO)3mes)Nd] (2–Nd) was treated with H2O to form the Nd(III) hydroxo complex [K(2.2.2cryptand)][((Ad,MeArO)3mes)Nd(OH)] (3–Nd). This reaction was identified as the H2 production step of the catalytic cycle. In order to demonstrate the stability of complexes 1–Ln in the presence of water, the trivalent aquo complex [((Ad,MeArO)3mes)Nd(H2O)] (1–Nd–H2O) was crystallized and structurally characterized. This aquo complex also provided insight in the initial substrate binding step of electrocatalytic H2O reduction, which also supports previous proposals of the uranium aquo complex [((Ad,MeArO)3mes)U(H2O)] as an intermediate in the catalytic cycle of 1–U. Thus, the combined results presented here provide a more complete picture of f-element facilitated electrocatalytic H2O reduction.

ASSOCIATED CONTENT

Supporting Information. General considerations, synthetic procedures, experimental conditions, spectroscopic data, and mathematical equations. This material is available free of charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION Corresponding Authors * E-mail: [email protected] and [email protected]

Notes

The authors declare no competing financial interests. Author Contributions The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

ACKNOWLEDGMENT D.P.H acknowledges the Graduate School Molecular Science (GSMS) of FAU Erlangen-Nürnberg for the generous support. The Bundesministerium für Bildung und Forschung (BMBF, support codes 02NUK012C and 02NUK020C), the FAU Erlangen-Nürnberg, and COST Action CM1006 are acknowledged for funding. W.J.E. and C.T.P. acknowledge the U.S. National Science Foundation for support (CHE-1565776 to W.J.E.; DGE-1321846 to C.T.P.). Dr. Frank W. Heinemann (FAU) and Prof. Susanne Mossin (DTU) are acknowledged for their work on the Co(II) tris-aryloxide complex, [((tBu,tBuArO)2(tBu,tBuArOH)N)CoII(NCCH3)], presented in the SI.

REFERENCES (1) Huang, W.; Wang, H.; Zhou, J.; Wang, J.; Duchesne, P. N.; Muir, D.; Zhang, P.; Han, N.; Zhao, F.; Zeng, M. et al. Nat. Commun. 2015, 6, 10035. (2) Zhang, R.; Zhang, W.; Gao, L.; Zhang, J.; Li, P.; Wang, W.; Li, R. Applied Catalysis A: General 2013, 466, 264–271.

(3) Vilchez, F.; Gutierrez-Granados, S.; Alatorre Ordaz, A.; Galicia, L.; Herrasti, P. Journal of Electroanalytical Chemistry 2008, 614, 8–14. (4) Helm, M. L.; Stewart, M. P.; Bullock, R. M.; DuBois, M. R.; DuBois, D. L. Science 2011, 333, 863–866. (5) Thoi, V. S.; Sun, Y.; Long, J. R.; Chang, C. J. Chem. Soc. Rev. 2013, 42, 2388–2400. (6) Halter, D. P.; Heinemann, F. W.; Bachmann, J.; Meyer, K. Nature 2016, 530, 317–321. (7) Karunadasa, H. I.; Chang, C. J.; Long, J. R. Nature 2010, 464, 1329–1333. (8) Singh, W. M.; Baine, T.; Kudo, S.; Tian, S.; Ma, Xuan Anh N.; Zhou, H.; DeYonker, N. J.; Pham, T. C.; Bollinger, J. C.; Baker, D. L. et al. Angew. Chem. Int. Ed. 2012, 51, 5941– 5944. (9) Ridley, F.; Ghosh, S.; Hogarth, G.; Hollingsworth, N.; Holt, K. B.; Unwin, D. G. Journal of Electroanalytical Chemistry 2013, 703, 14–22. (10) Letko, C. S.; Panetier, J. A.; Head-Gordon, M.; Tilley, T. D. J. Am. Chem. Soc. 2014, 136, 9364–9376. (11) Wiedner, E. S.; Bullock, R. M. J. Am. Chem. Soc. 2016, 138, 8309–8318. (12) Garrett, B. R.; Click, K. A.; Durr, C. B.; Hadad, C. M.; Wu, Y. J. Am. Chem. Soc. 2016, 138, 13726–13731. (13) Halter, D. P.; Heinemann, F. W.; Maron, L.; Meyer, K. Nat. Chem. 2017, DOI: 10.1038/NCHEM.2899. (14) La Pierre, H. S.; Kameo, H.; Halter, D. P.; Heinemann, F. W.; Meyer, K. Angew. Chem. Int. Ed. 2014, 53, 7154–7157. (15) Johnson, D. A. J. Chem. Soc., Dalton Trans. 1974, 1671–1675. (16) Fieser, M. E.; Palumbo, C. T.; La Pierre, H. S.; Halter, D. P.; Voora, V. K.; Ziller, J. W.; Furche, F.; Meyer, K.; Evans, W. J. Chem. Sci. 2017, 8, 7424–7433. (17) Palumbo, C. T.; Halter, D. P.; Voora, V. K.; Guo, C. P.; Chan, A. K.; Fieser, M. E.; Ziller, J. W.; Hieringer, W.; Furche, F.; Meyer, K. et al. Inorg. Chem. 2017, submitted. (18) Edelmann, F. T. Lanthanide metallocenes in homogeneous catalysis. In Topics in Current Chemistry: Organolanthoid Chemistry-Synthesis, Structure, Catalysis; Hermann, W. A., Anwander, R., Edelmann, G., Eds. 179; Springer: New York, 1996; pp 247–276. (19) Shibasaki, M.; Yoshikawa, N. Chem. Rev. 2002, 102, 2187–2210. (20) Edelmann, F. T. Chem. Soc. Rev. 2012, 41, 7657–7672. (21) Halter, D. P.; La Pierre, H. S.; Heinemann, F. W.; Meyer, K. Inorg. Chem. 2014, 53, 8418–8424. (22) Neidig, M. L.; Clark, D. L.; Martin, R. L. Coordination Chemistry Reviews 2013, 257, 394–406. (23) Mikheev, N. B. Inorganica Chimica Acta 1984, 94, 241– 248. (24) Mikheev, N. B.; Kamenskaya, A. N. Coordination Chemistry Reviews 1991, 109, 1–59. (25) Kimura, S.; Bill, E.; Bothe, E.; Weyhermüller, T.; Wieghardt, K. J. Am. Chem. Soc. 2001, 123, 6025–6039. (26) Glebov, E. M.; Pozdnyakov, I. P.; Plyusnin, V. F.; Khmelinskii, I. Journal of Photochemistry and Photobiology C: Photochemistry Reviews 2015, 24, 1–15. (27) Eidem, P. K.; Maverick, A. W.; Gray, H. B. Inorganica Chimica Acta 1981, 50, 59–64. (28) Mason, P. E.; Uhlig, F.; Vaněk, V.; Buttersack, T.; Bauerecker, S.; Jungwirth, P. Nat. Chem. 2015, 7, 250–254. (29) Costentin, C.; Savéant, J.-M. CHEMELECTROCHEM 2014, 1, 1226–1236.



(30) Rafiee, M.; Miles, K. C.; Stahl, S. S. J. Am. Chem. Soc. 2015, 137, 14751–14757. (31) Pegis, M. L.; Roberts, J. A. S.; Wasylenko, D. J.; Mader, E. A.; Appel, A. M.; Mayer, J. M. Inorganic chemistry 2015, 54, 11883–11888. (32) Roberts, J. A. S.; Bullock, R. M. Inorg. Chem. 2013, 52, 3823–3835. (33) Brown, I. D.; Skowron, A. J. Am. Chem. Soc. 1990, 112, 3401–3403. (34) Xue, D.; Zuo, S.; Ratajczak, H. Physica B: Condensed Matter 2004, 352, 99–104.

(35) Lee, K. J.; Elgrishi, N.; Kandemir, B.; Dempsey, J. L. Nat. rev. chem. 2017, 1, 39. (36) Rountree, E. S.; McCarthy, B. D.; Eisenhart, T. T.; Dempsey, J. L. Inorg. Chem. 2014, 53, 9983–10002. (37) Costentin, C.; Drouet, S.; Robert, M.; Savéant, J.-M. J. Am. Chem. Soc. 2012, 134, 11235–11242.


Page 8 of 9

Page 9 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of the American Chemical Society

Insert Table of Contents artwork here


9

Mechanistic Insight and Overpotential Tuning in a Series of

Recommend Documents