Hydrogen Atom or Proton Coupled Electron Transfer? - American

organic framework, was investigated by multireference wave function theory. ..... for oxygen, grey for carbon, white for hydrogen, blue for cobalt, li...
0 downloads 0 Views 1MB Size
Subscriber access provided by Nottingham Trent University

Article

Hydrogen Atom or Proton Coupled Electron Transfer? C-H Bond Activation by Transition Metal Oxides Carlo Alberto Gaggioli, Joachim Sauer, and Laura Gagliardi J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.9b04006 • Publication Date (Web): 21 Aug 2019 Downloaded from pubs.acs.org on August 21, 2019

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 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 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.

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 30 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

Hydrogen Atom or Proton Coupled Electron Transfer? C-H Bond Activation by Transition Metal Oxides Carlo Alberto Gaggioli a, Joachim Sauer* b and Laura Gagliardi* a a

Department of Chemistry, Chemical Theory Center and Supercomputing Institute, University of

Minnesota–Twin Cities, 207 Pleasant Street SE, Minneapolis, Minnesota 55455, United States b

Institut für Chemie, Humboldt-Universität zu Berlin, Unter den Linden 6, 10099 Berlin, Germany

Abstract The C-H bond activation in oxidative dehydrogenation of propane by hetero-bimetallic oxide clusters (first row transition metals), deposited on the zirconium oxide node of the NU-1000 metal organic framework, was investigated by multireference wave function theory. The redox-active part of the systems studied has composition (CoO)(MO)(OH)2 with M= Ti, Mn, Fe, Co, Ni, Cu, Zn. In this series, the energy of H transfer from propane to the metal oxide (E) varies from -26 kcal/mol for M=Cu, Zn to 85 kcal/mol for M=Ti. This is accompanied by a change in the mechanism from hydrogen atom transfer, M2+(dn) O−  M2+(dn) OH− , for M=Cu, Zn to proton coupled electron transfer, Mm+(dn) O2−  M(m-1)+(dn+1) OH−, for M=Ni, Co, Fe, Mn, Ti. Whereas for M=Ni (E = -13 kcal/mol) Ni+III is reduced to Ni+II, for M=Co, Fe, Mn (E = 1, 10, 6 kcal/mol, respectively) it is Co+III that is reduced to Co+II, For M=Ti, Ti maintains its +IV oxidation state and Co+II is reduced to Co+I.

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 30

1. Introduction As a central question in chemistry, the activation of C–H bonds by metal oxides in general, and by different types of transition metal oxides in particular, has been an active area of research over decades1–8 and continues to do so till today, see, e.g., refs.9–15 The types of metal oxides range from solid catalysts1,3,4,9,10,12 over enzymes5,6,14,15 and molecular compounds2,6,13 to gas phase species.7,8,11,16 Among the reactions considered, there are industrially relevant processes3,4 as the oxidative dehydrogenation of alcohols to aldehydes10 and of propane to propene, as well as the spectacular selective conversion of n-butane to maleic acid anhydride.9 Among the small alkanes, methane is in the focus not only because the C–H bond is least reactive, but also because it is the principal component of natural gas and as such it becomes increasingly important as energy source and raw material in chemical industry. Although enzymes like methane mono-oxygenase can functionalize methane at ambient conditions,14 a corresponding viable industrial process has not yet been developed, neither have improved catalysts for oxidative coupling of methane resulted in processes with higher yield of C2 products.17 Therefore, atomistic understanding with the aim to design model compounds and improved catalysts is still a priority, and much insight is being gained from discussing H atom abstraction from C–H bonds in terms of proton-coupled electron transfer (PCET).9,10,11,13 A special case of the latter is hydrogen atom transfer (HAT) from a C–H bond to an oxyl radical: the electron and proton together attach to the O atom and form a new bond. C–H + [O−,Mm+] C–H + O–M



 C + [H–O−,Mm+]

(1a)

 C + H–O–M

(1b)

The two equations above are the ionic (1a) and covalent (1b) representations of the HAT process.

-2-

ACS Paragon Plus Environment

Page 3 of 30 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

HAT is a facile radical recombination reaction. It occurs with gas phase cluster cations generated by ionization of simple metal oxides or fully oxidized transition metals with empty d-shells.7,8 Experiments have shown that species like Al8O12 and V4O10 abstract hydrogen from CH4 in agreement with quantum chemical calculations which predict exothermic reactions (of the order of 100 kJ/mol) with no barriers.18,19 Similarly, e.g., Sc4O4 and Ce4O9 have been found to abstract H atoms from n-butane with calculated reaction energies of -18 and -17 kcal/mol, respectively.16

For surface oxyl species, generated by Li-doping of solid MgO as postulated by

Lunsford,1 barriers as low as 3 ±2 kcal/mol have been calculated.20 Typical for transition metal oxides is a different type of PCET with the proton attaching to the O atom of the metal-oxo bond and the electron populating available d-states at the metal site C–H + [O2−,Mm+(dn)]  C + [HO−,M(m-1)+(dn+1)] C–H +

O=M(dn)  C +

HO–M(dn+1)

(2a) (2b)

In this article we reserve the name PCET to this type of reaction, to distinguish it from HAT (eq. 1a and 1b). The barriers for PCET reactions are much higher. For example, values between 50 – 120 kJ/mol have been reported for the activation of n-butane on vanadium phosphate oxide (50100 kJ/mol)9 and of propane on oxide-supported vanadia (80-120 kJ/mol),21 or on cobalt supported on ZrO-nodes of the NU-1000 metal-organic framework (MOF) NU-1000 (ca. 80 kJ/mol).12 The higher barriers for PCET become understandable if this reaction (eq. 2) is decomposed into two hypothetical steps, formation of an oxyl species by ligand to metal charge transfer (LMCT), [O2−,Mm+(dn)]  [O−,M(m-1)+(dn+1)] O=M(dn)





O–M(dn+1)

(3a)



(3b)

-3-

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

and subsequent H atom transfer according to eqs. (1a/1b). Once the LMCT excitation energy, ELMCT, has been spent, eq. (3a/b), the facile, exoenergetic radical recombination (1a/1b) will occur, EPCET(2a/b)  ELMCT(3a/3b) + EHAT(1a/1b) This explains that the activity of supported transition metal oxide catalysts increases with decreasing UV-visible absorption edge energy, e.g., for vanadia catalysts a rate increase of two orders of magnitude has been observed when the edge energy drops from 3.0 to 2.2 eV.22 It also explains that electronically excited states of transition metal oxo species are more reactive. For example, silica-supported vanadia, as other transition metal oxides, is a viable photocatalysts.23 Similarly, in their study of alkane hydroxylation by an oxo-iron(IV) species Ye and Neese found that on the reaction path an oxyl iron (III) species is formed according to eqs. 3a/3b that abstracts the H-atom from the CH bond.15

Scheme 1. Two different electronic states of a transition metal – oxygen species that lead to different H abstraction mechanisms, hydrogen atom transfer (HAT) and proton-coupled electron transfer (PCET).

Here, we are interested in transition metal ions with partially occupied d-shells which show valence isomerism24,25 of the type describe by eq. (3a/3b) and, hence, can switch between the less reactive -4-

ACS Paragon Plus Environment

Page 4 of 30

Page 5 of 30 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

metal oxo state and the more reactive metal oxyl state, i.e., between PCET and HAT depending on the relative stability of the higher, Mm+(dn), or lower, M(m-1)+(dn+1), oxidation state. For example, a [CoO]+ species could exist as [Co2+(d7)O−]+

or

[Co3+(d6)(O2−)]+

(4)

see Scheme 1. In the former case we would expect high reactivity (HAT, eq. 1a) and in the latter case lower reactivity (PCET, eq. 2a). Theoretically predicting which of the two electronic structure types is prevailing in a given system and which of the two mechanisms, HAT or PCET, will be operating requires quantum-mechanical multi-reference (MR)26,27 calculations. We present here such calculations on the binuclear cobalt oxide species (CoO)2(OH)2 and its transition metal doped analogues, (CoO)(MO)(OH)2 with M= Ti, Mn, Fe, Ni, Cu, Zn. These species form the redox-active part of compounds with the composition [(CoO)(MO)(OH)2] (ZrO2)4 Zr2(CH3COO)8 which had been used in ref.28 to model cobalt oxo-hydroxo species deposited on the ZrO2 nodes of the metal-organic framework NU1000.12 Density functional theory (DFT) had been applied to examine the effect of doping on the activation enthalpy for H abstraction from propane.28 Here we present MR calculations on the enthalpy for this reaction which show a change in mechanism from PCET to HAT with the variation of the dopant, leading to a decrease of the predicted reaction energy from + 85 to - 26 kcal/mol. According to the Brønsted-Evans-Polanyi principle29 we expect a proportional variation of the activation enthalpies.

2. Computational Models and Methods 2.1. Models and DFT Structure Optimization

-5-

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

We depart from the cluster model used in ref.28 for DFT calculations which was formed by extracting a zirconium oxide node from a NU-1000 metal-organic (MOF) framework structure optimized under periodic boundary conditions. The linker molecules were replaced with acetate (CH3COO-) groups and the positions of the atoms in the methyl group of the acetate were kept frozen. The composition of the node model was (ZrO2)4(Zr)2(CH3COO)8. The model was then completed by attaching the Co2O2(OH)2 cobalt-oxo-hydroxo species to one face of the zirconium oxide node yielding the [Co2O2(OH)2](ZrO2)4Zr2(CH3COO)8 model which we call “large” (Zr6). It is shown in the Supporting Information, Figure S1. For efficiency of the computationally more demanding MR calculations, we adopt truncated “small” (Zr2) models of composition [(CoO)(MO)(OH)2]ZrO2Zr(OH)4(H2O)4 for our study of the first H atom abstraction from propane. The reactant structure consists of propane interacting with the “small” (Zr2) model through van der Waals interactions, while the product consists of a propyl radical and a hydrogen atom (previously bonded to propane) bonded to an oxygen atom that is part of the Co/M/Zr-oxide node. In the following, we call the former vdW complex, and the latter Pr complex, see Figure 1. These “small” models were obtained from the “large” acetate-saturated cluster used in ref.28 by including only the first coordination sphere of the Co and M atoms and two proximal Zr atoms with all the oxygen atoms bonded to them. Twelve terminal hydrogen atoms were added to keep charge neutrality which resulted in OH and H2O ligands that complete the coordination sphere of the Zr4+ ions. Their positions were fixed to be in the same direction as the removed Zr atoms and in the same direction as the removed carbon atom of the acetate. The oxygen-hydrogen bond distances were optimized using the same level of theory (M06-L) as for the larger cluster. In this

-6-

ACS Paragon Plus Environment

Page 6 of 30

Page 7 of 30 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

way we minimized artificial H-bond interactions arising from the hydrogen capping of the node. For comparison of the “large” and “small” models see Supporting Information, Figure S1.

vdW

Pr

Figure 1: Structure drawings (top row) and ball-and-stick representation (bottom row) of models for the vdW complex with propane, vdW, and the product complex with the propyl radical, Pr. Color code: red for oxygen, grey for carbon, white for hydrogen, blue for cobalt, light green for zirconium. For the “small” models we performed a further, partial structure optimization using the M06 hybrid functional30 in which only the positions of the Co and M atoms, the two bridging O atoms, and the H atom that is transferred from propane in the vdW complex to a bridging O atom in the product

-7-

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

were allowed to move, while the position of the other atoms were kept fixed. The latter DFT structure optimizations were performed with Gaussian 09.31 For M=Ni, Co, and Fe, we also localized transition structures (TS) adopting a “small” model for the TS following the same procedure as described before for the DFT geometry optimization, namely optimizing the “small” model (capped with OH- and H2O ligands) with M06-L. Subsequently, we computed the transition structures using the M06 functional by optimizing only the positions of the Co and M atoms, the two bridging O atoms, and the H atom that is transferred from propane in the vdW complex to a bridging O atom, keeping the position of the other atoms fixed. For M=Zn and Cu we were able to locate the TS complexes by following the same procedure described above, and by optimizing the atoms constituting the propane molecule as well. 2.2. Multireference calculations We performed single point MR calculations at the M06 structures using the MOLCAS-8.2 package.27 Since previous MR calculations28 have shown that different spin states are close in energy (within 1-3 kcal/mol), we performed the calculations only for the highest spin state of a given compound, which can accommodate both the [O2−,Mm+(dn)] and [O−,M(m-1)+(dn+1)] electronic valence states. We employed the complete active space self-consistent field (CASSCF)32 and restricted active space self-consistent field (RASSCF)33,34 methods, followed by second order perturbation theory (CASPT235,36/RASPT234,37) to include dynamical correlation. We made use of Hirshfeld spin populations,38 computed using the MultiWfn package,39 as indicative of the valence state, i.e. the electron configuration on the transition metal ion and the oxygen species. In the CASSCF/CASPT2 calculations, an active space (AS) of n electrons in m orbitals (n,m) was chosen, and all configurations that arise from all possible excitations of the n electrons in the m

-8-

ACS Paragon Plus Environment

Page 8 of 30

Page 9 of 30 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

orbitals are included in the wave function. As an example, for the complex having Co and Zn as metals (in the sextet spin state), we included in the active space all the unpaired electrons, namely five electrons in five orbitals, CASSCF(5,5). The general choice of active space for CASSCF was to include all the d orbitals of the metal atoms that contain unpaired electrons. In some cases, one of the d orbitals is exchanged with one p orbital of the oxygen, depending on the stability of different metal oxidation states (vide infra). We subsequently explored larger active spaces using the restricted active space SCF (RASSCF) formalism. In the RASSCF model, the active orbitals are divided into three distinct subspaces, RAS1, RAS2 and RAS3. In RAS1, the orbitals are doubly occupied and at most double excitations from RAS1 are allowed. In RAS2, n electrons in m orbitals are included and a full configuration interaction is performed. In RAS3, virtual orbitals are included, and at most double excitations to RAS3 are allowed. The following notation will be used: RASSCF(nae in nao)/(nae2 in nao2)/n, where n indicates the maximum number of electrons excited from RAS1 or into RAS3 (n=2 in this case), (nae in nao) describes the global RAS(1-3) active space, and (nae2 in nao2) the RAS2 space. For example for Co_Zn, in RASSCF(9,14)/(5,5)/2 there are 5 electrons and 5 orbitals in RAS2 (chosen as the d orbitals of the metal having an unpaired electron), 2 doubly occupied orbitals in RAS1 (chosen as two p orbitals localized on bridging oxygen atoms) and 7 empty orbitals in RAS3 (chosen as all the correlating orbitals, defined as the unoccupied orbitals with a higher quantum number with respect to the occupied orbitals). These orbitals are shown in Figures 2 and 3 for the Co_Zn vdW and Pr, respectively. As in the CASSCF case, described above, in some cases one p orbital of oxygen and one d orbital of the metal are exchanged, namely the former will be in RAS2 and the latter in RAS1 (see next section for a case by case description). We explored the convergence of the active space, by including more p orbitals of the bridging oxygen atoms and corresponding correlating orbitals in the active space, for the

-9-

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

Co/Zn complex and inspecting spin densities and reaction energies and found that the chosen RASSCF(9,14)/(5,5)/2 active space is accurate enough, see Supporting Information, Table S3. The chosen active spaces and spin states for all the Co/M systems studied are given in Table S4.

Figure 2: RASSCF(9,14)/(5,5)/2 active orbitals with occupation numbers in brackets for M=Zn for the vdW complex.

- 10 -

ACS Paragon Plus Environment

Page 10 of 30

Page 11 of 30 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

Figure 3: RASSCF(9,14)/(5,5)/2 active orbitals with occupation numbers in brackets for M=Zn for the Pr complex. 3. Results and Discussion 3.1. DFT structures First, we compare the small and large model (for M=Zn, Co, Fe) reaction energies and electronic structures, to see whether there is a significant difference between the two. The results are shown in Table S1. The differences in the reaction energy are in the range of 3 (for M=Co) to 8 (for M=Zn) kcal/mol. Nevertheless, the Hirshfeld spin populations deviates by only 0.1 on average (maximum deviation is 0.17 for spin on O1 for Co_Zn vdW), and therefore we conclude that the electronic structure of the active site is qualitatively correctly described by the small model. Selected bond distances of the partially optimized M06 structures are given in Tables S1a and S1b of the Supporting Information. Figure 4 shows the bond distances of the CoO2M core for M=Co and Zn and their change on adding an H atom. In the Co/Co system there are two O2− ions and on PCET one of them is converted into an OH− ion. In agreement with the reduced ionic bond strength

- 11 -

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

of the latter, both Co-O distances get longer. With M=Zn (or Cu), there is a O− ion in the Co/M vdW complex. Forming a Co(III)-oxyl species, it stays close to the Co3+ ion whereas the distance to the Zn2+ (or Cu2+) ion is much longer, 3.06 (or 3.02) Å. On H atom transfer, this bond is shortened in the product complex to 2.38 (or 2.37) Å. The distance between OH− and Zn2+ is 0.36 Å longer than the distance between OH− and Co3+, again in agreement with the smaller electrostatic valence of Zn2+ compared Co3+.

Figure 4. Bond distances (Å, M06 functional) for Co/Co and Co/Zn vdW (left) and Pr complexes (right). The change on product formation is given in italics.

In general, the structure type of the Co/M/Zr-oxide cluster shown in Figure 1 does not change between the reactant (vdW) and the product (Pr). Only for the Co/Mn system we found a different structure isomer for the product complex which we call “restructured” (RES-Pr). Figure 3 shows it compared to the usual “isomorphous” (Pr) substitution structure. In the vdW complex of this system one OH group is simultaneously linked to Zr and Mn (with bond distances of 2.54 and 2.07 Å respectively, see figure 5), making the coordination of Mn equal to 5 (see tables S2a and S2b). This coordination number is known to be stable for Mn.40 In the “restructured” product complex,

- 12 -

ACS Paragon Plus Environment

Page 12 of 30

Page 13 of 30 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

RES-Pr, one of the bridging oxygen atoms is far apart from Co and Mn, with distances of 4.14 and 3.30 Å, respectively, getting instead very close to Zr, at 1.91 Å.

vdW

Pr

RES-Pr

Figure 5. Structure of the “small” (CoO)(MnO)(OH)2 Zr2O2(OH)4(H2O)4 model of the Co/Mn vdW complex (left), product complex Pr (middle) and comparison to lower energy product “restructured” (RES-Pr) isomer (right). Color code: see Fig. 1, purple for manganese.

3.2. Electronic structures of the vdW and product complexes

- 13 -

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

Table 1 shows the RASSCF Hirshfeld spin populations on the Co, M and bridging O atoms of the CoO2M core for the vdW complexes. For the Co/Co complex they suggest a [Co3+(d6),O2−]+ valence state which is maintained also for Co in the doped Co/M complexes with M=Zn, Cu, Ni, Fe, and Mn. The high stability of the Zn2+(d10) and Cu2+(d9) configurations explains the presence of O− radical species in the Co/Zn and Co/Cu systems, whereas in the Co/Ni, Co/Fe, and Co/Mn systems the [M3+(dn) O2−] valence state is more stable than the [M2+(dn+1) O−] one. For the Co/Ni vdW complex, we managed to find the [M2+(dn+1) O−] solution, but it is 22.4 kcal/mol less stable than the [M3+(dn) O2−] solution, see Table S5. Attempts to find additional solutions with different oxidation states for other metals always converged to the configuration reported in Table 1, see Supporting Information, section S3. In the Ti species, since Ti prefers to be in the Ti4+(d0) state, Co assumes the Co2+(d7) configuration and both O ions are O2−. These results suggest that H abstraction from propane will occur as HAT with low if any barrier for the Zn and Cu doped systems, whereas PCET will dominate for the Co/Co system and the other dopants (Ni, Co, Fe, Mn and Ti). In the product complex, an H atom has been transferred from the propane molecule to the metal oxide. For the Co/Zn and Co/Cu vdW complexes featuring an O− radical species HAT results in an OH− group. For all the other systems there is PCET. The electron is accommodated in the dshell of the metal ion whereas the proton is added to an O2− ion forming an OH− group. For Co/Co one of the Co3+(d6) ions is reduced to Co2+(d7). This is also the case for Co/Fe and Co/Mn because Co3+ is more easily reduced than Fe3+ and Mn3+, respectively. In contrast, in Co/Ni Co stays Co3+(d6) because the reduction of Ni3+ to Ni2+ is more stable than the reduction of Co3+ to Co2+. Ti doping is a special case. Ti is so stable as Ti4+ ion that Co2+ in [Co2+,O2−/Ti4+,O2−] is reduced to Co+ in [Co+,OH−/Ti4+,O2−].

- 14 -

ACS Paragon Plus Environment

Page 14 of 30

Page 15 of 30 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

Table 1: RASSCF Hirshfeld spin populations on the Co, M and bridging O atoms of the CoO2M core for the vdW complex. The last column shows the electron configuration derived from the former. M=

O1

O2

Co

M

Electron Configuration

Zn

0.94

0.12

3.61

0.01

Co3+(d6) O 2−/Zn2+(d10) O−

Cu

0.95

0.13

3.62

0.94

Co3+(d6) O 2−/Cu2+(d9) O−

Ni

0.24

0.22

3.63

2.50

Co3+(d6) O 2−/Ni3+(d7) O 2−

Co

0.24

0.14

3.61

3.58

Co3+(d6) O 2−/Co3+(d6) O 2−

Fe

0.20

0.19

3.56

4.62

Co3+(d6) O 2−/Fe3+(d5) O 2−

Mn

0.21

0.09

3.60

3.73

Co3+(d6) O 2−/Mn3+(d4) O 2−

Ti

0.03

0.01

2.83

0.01

Co2+(d7) O 2− /Ti4+(d0) O 2−

Table 2: RASSCF Hirshfeld spin populations on the Co, M and bridging O atoms of the CoO2M core for the product complex. The spin populations on the propyl radical (sum of all atoms) and on the C atom are also given. The last column shows the electron configuration derived from the spin populations. M=

O1

O2

Co

M

Propylb

C

Electron Configuration

Zn

0.06

0.15

3.58

0.01

0.96

0.64

Co3+(d6) O2−/Zn2+(d10) OH−

Cu

0.06

0.15

3.60

0.94

0.95

0.66

Co3+(d6) O2−/Cu2+(d9) OH−

Ni

0.07

0.12

3.64

1.90

0.95

0.63

Co3+(d6) O2−/Ni2+(d8) OH−

Co

0.09

0.12

2.85

3.64

0.94

0.64

Co3+(d6) O2−/Co2+(d7) OH−

Fe

0.08

0.17

2.85

4.57

0.94

0.64

Co2+(d7) OH−/Fe3+(d5) O2−

Mna

0.00

0.07

2.88

3.82

0.96

0.66

Co2+(d7) OH−/Mn3+(d4) O2−

Ti

0.03

0.02

1.88

0.01

0.95

0.66

Co+(d8) OH− /Ti4+(d0) O2−

a“

Restructured” product complex (RES-Pr, see Fig. 5).

b

Sum of spin populations on all atoms of the propyl radical.

- 15 -

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

3.3. H Abstraction Mechanism and Reaction Energies Table 3 provides information about the change of the spin populations on the atoms of the CoO/MO fragment of the Co/M complexes on attachment of the H atom which is abstracted from the C-H bond of propane. Figure 6 shows the corresponding RASPT2 reaction energies (numbers in Table 3). The sum of the spin population changes on the CoO/MO atoms is between -0.86 and -0.95 for all systems, but for different Co/M cases the dominating contribution is on different atoms. Four different cases can be distinguished: (i) For the Co/Zn and Co/Cu systems the substitution of Co by Zn and Cu, respectively, has resulted in an O− species. On H attachment its spin is quenched, i.e. there is HAT converting an O− species into OH− which is strongly exoenergetic, about 26 kcal/mol, with almost no dependence on the dopant. This fits to reaction energies of about – 24 and -29 kcal/mol predicted by DFT for the multinuclear Al8O12+ and V4O10+ gas phase clusters, respectively, that have been experimentally shown to abstract hydrogen from CH4,18,19 and also to the energies of -17 to -18 kcal/mol for the observed reactions of Sc4O4- and Ce4O9+ with n-butane.16 In all other cases listed in Table 3 PCET occurs with the proton attaching to one of the bridging O atoms and the electron reducing a transition metal ion. In the following we describe PCET case by case, highlighting which metal ion is reduced during PCET. (ii) In the Co/Ni system Ni+III is reduced to Ni+II which is easier than reducing Co+III to Co+II. Since Ni prefers the formal (+II) rather than (+III) oxidation state, the H atom attachment is also exoenergetic, but only by about 13 kcal/mol. Table 3: Hirshfeld spin density differences between Pr and vdW complexes, Δspin, obtained with RASSCF for the bridging O atoms as well as for the Co and M (dopant) atoms. The numbers in bold are used for the atoms where the largest difference occurs. The RASPT2 energy difference

- 16 -

ACS Paragon Plus Environment

Page 16 of 30

Page 17 of 30 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

between Pr and vdW complexes, ΔE, in kcal/mol, is also reported, together with the type of redox process. The rows labeled TS show the corresponding spin density differences and energy differences (energy barriers) between the TS structures and vdW complexes. Δspin M=

O1

O2

Co

M

ΔE

Redox process

Zn

-0.88

0.03

-0.03

0.00

-25.5

[Zn2+(d10)] O−  OH−

-0.65

0.04

-0.03

0.01

-12.1

[Zn2+(d10)] O−  OH− b

-0.89

0.02

-0.02

0.00

-26.1

[Cu2+ (d9)] O−  OH−

-0.64

0.03

-0.04

0.01

-11.5

[Cu2+(d9)] O−  OH− b

-0.17

-0.10

0.01

-0.60

-12.9

Ni3+(d7) O2−  Ni2+(d8) OH− c

0.44

-0.01

0.05

-0.58

16.0

Ni3+(d7) O2−  Ni2+ (d8)] O− c

-0.15

-0.02

-0.76

0.06

1.1

Co3+(d6) O2−  Co2+(d7) OH− d

0.12

-0.04

-0.77

0.08

23.3

Co3+(d6) O2−  Co2+(d7) OH− d

-0.12

-0.02

-0.71

-0.05

9.6

Co3+(d6) O2−  Co2+(d7) OH− d

0.08

-0.01

-0.71

-0.05

26.0

Co3+(d6) O2−  Co2+(d7) OH− d

Mna

-0.21

-0.02

-0.72

0.09

6.3

Co3+(d6) O2−  Co2+(d7) OH− d

Ti

0.00

0.01

-0.95

0.00

85.1

Co2+(d7) O2−  Co+ (d8) OH− c

TS Cu TS Ni TS Co TS Fe TS

b

b

a“

Restructured” product complex (RES-Pr, see Fig. 5). H atom transfer (HAT). c Proton coupled electron transfer (PCET) with M (=Ni) reduction. d PCET with Co reduction from +III to +II. e PCET with Co reduction from +II to +I. b

(iii) For Co/Co, Co/Fe and Co/Mn, regardless of the dopant, Co+III is reduced to Co+II. This explains that the positive reaction energies are similar, varying between 1.1 kcal/mol and 9.6 kcal/mol only. Reaction energies of the same order of magnitude, 2.4 kcal/mol and 12.0 kcal/mol,

- 17 -

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

depending on the product spin-state, have been calculated before for the H abstraction from propane by V3O7+ (CH3)2C H[O=VV(d0)V2O6]+ 

(CH3)2C[HOVIV(d1)V2O6]+

and found to be compatible with the experimental results.41,42 (iv) On substitution of Ti for Co in the Co/Ti system, Ti assumes a +IV formal oxidation state and forces Co into the (+2) state, see Table 1. It turns out that Ti(+IV) is so stable that on H attachment it is rather Co+II that is further reduced to Co+I rather than Ti+IV being reduced to Ti+III. The reaction is very unfavorable, 85 kcal/mol endoenergetic. The change in mechanism sets also limits to attempts to find a simple reactivity descriptor such as the spin population or the charge on the H accepting atom.28 Only in case of HAT, (eqs. 1a, 1b) we may expect a correlation with the spin population on O1, see also the discussion in ref. 43. For the two examples of type (i) here, Co/Zn and Co/Cu, the small difference in the spin population is indeed reflected in a corresponding small difference in the reaction energy.

- 18 -

ACS Paragon Plus Environment

Page 18 of 30

Page 19 of 30 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

Figure 6. Reaction energies computed with RASPT2 (diamonds) and M06 (green squares). The labels indicate the doping metal, M, in the Co/M system. Color code for diamonds: red for HAT, blue for PCET with M reduction (from +III to +II), black for PCET with Co reduction (from +III to +II), purple for PCET with Co reduction (from +II to +I).

For PCET we may rather think of two parameters, the basicity of the O atom that accepts the proton and the energy required to reduce the transition metal ion. When comparing Co/Ni – type (ii) with Co/Co it is Ni+III that is more easily reduced to the +II oxidation state than Co+III, and for Co/Ti – type (iv) - it is very difficult to reduce Co+II to Co+I which explains that the reaction considered is 75 kcal/mol more endoenergetic than for any other Co/M system. With respect to making predictions without having to localize transition structures or product structures, the energy of hydrogenation is recommended.7,44 It just doubles the effort one is spending on the calculations for the reactant model that are required to get descriptors like charges or spin densities. For M=Zn, Cu, Ni, Co and Fe we calculated energy barriers and we analyzed the electronic structure. The spin rearrangement from vdW to TS is reported in Table 3. For the spin populations of the TS structures see Table S6. For the systems that react through a HAT mechanism (namely Co_Zn and Co_Cu), in the Pr there is a spin population change of -0.88 and -0.89 at O1, respectively. In the TS the spin population change is -0.65 and -0.64 at O1, pointing to a HAT mechanism (with oxygen reduction) in which more than half of the spin rearrangement has occurred. The energy barriers are -12.1 kcal/mol for Co_Zn and -11.5 kcal/mol for Co_Cu, respectively. This means that the reaction is barrierless at the RASPT2 level of theory (using the DFT optimized TS geometry). The energy barriers follow the trend of the reaction energies (-25.5 and -26.1 kcal/mol, see Table 3).

- 19 -

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

For the systems that react through a PCET mechanism, in the product state of Co_Ni, Co_Co, and Co_Fe there is a spin population change of -0.60 (at Ni), -0.76 (at Co) and -0.71 (at Co), respectively, indicating a reduction to Ni+II, Co+II and Co+II, respectively, see Table 3. In the transition state, the spin population changes are only marginally different, -0.58, -0.77 and -0.71, respectively, indicating that the transfer of the electron into the metal d-states is already completed. Correspondingly, the energy barriers for Co/Ni, Co/Co and Co/Fe, 16.0, 23.3, and 26.0 kcal/mol, respectively, follow the trend of the reaction energies (Table 3), -12.9, 1.1, and 9.6 kcal/mol, respectively. The relationship between the energy barriers and reaction energies is shown in Figure S2, follows the trend corresponding to the Brønsted-Evans-Polanyi principle29. There is a peculiarity with the Co/Ni system, however. For all three systems, in the product state the transfer of an electron in the metal d-states is accompanied with transfer of the proton to the oxygen atom and creation of the propyl radical site (PCET), see eq. 2a. In the transition state, transfer of the electron is almost completed as the spin population changes on the metal sites indicate. While for Co/Co and Co/Fe the corresponding formation of the radical site at the propyl species is almost completed (spin populations of 0.66 and 0.67, respectively, compared to 0.94 for both systems in the product state), for the Co/Ni transition structure the propyl spin population is 0.29 only. Moreover, for this system a substantial change (0.44) of the spin population on the (proton accepting) O atom is found (Table 3). Hence, for the Co/Ni system the transition state resembles the intermediate of the two-step process mentioned in the introduction, first formation of an oxyl species by LMCT, eq. (3a), and subsequent H atom transfer (eq. 1a): C–H + [O2−,Ni3+(d7)]  {C–H[O−,Ni2+(d8)]}TS  C + [H–O−,Ni2+(d8)] This confirms that the Co/Ni system is an intermediate case, while it shares with Co/Co and Co/Fe the absence of an oxyl species in the initial (vdW) state, [O2−,Ni3+(d7)], its transition state

- 20 -

ACS Paragon Plus Environment

Page 20 of 30

Page 21 of 30 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

resembles the LMCT state [O−,Ni2+(d8)] with an oxyl species which, for type (i) species, is present as the initial state. As for the reaction energy, the explanation is the low stability of Ni+III species compared to Ni+II, and consequently not only is the reaction energy 14 kcal/mol more exothermic for Co/Ni than for Co/Co, but also the barrier is 6.7 kcal/mol lower. A similar case is the C-H bond activation by oxo-Fe(IV) species with an [Fe4+(d4)O2−]2+ core. Ye and Neese showed that part of the reaction path is the formation of an [Fe3+(d5)O−]2+ species by LMCT that then performs the actual H-atom abstraction.45

3. 4. Comparison with DFT results Figure 6 shows the DFT reaction energies (M06 functional) compared to the MR results. The trend of increasing reaction energies from Co/Zn to Co/Ti is the same as found with MR (RASPT2) calculations, but the M06 results show much less variation, between -5 and +43 kcal/mol only (see also Table S9), instead of -26 to +85 kcal/mol found with RASPT2 (Table 3). Similarly, the reaction energies computed with M06-L (with structures also optimized with M06-L) vary between 4.0 and 61.9 kcal/mol (see Table S12). The Hirshfeld spin densities for the reactants (vdW) and products (Pr) as well as their change are shown in Tables S7-S9 of the SI. Unlike the MR results, the spin density is not localized on the CoO2M part, but delocalizes over other oxygen and metal atoms of the whole Co/M/Zr-oxide cluster. For M06, the sum over the four atoms of the CoO2M core of the spin densities differences between Pr and vdW complexes is between -0.65 (Co/Co) and -0.80 (Co/Zn), see Table S9, whereas the MR calculations predict values between -0.95 and -0.86 for all systems (Table 3). The Co/Ti system is an extreme case. Due to the strong preference of Ti to stay in the +IV oxidation state, our MR calculations predict that on PCET, Co goes from Co2+(d7) to Co+(d8), while the

- 21 -

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 22 of 30

DFT(M06) calculations show no change on the CoO2Ti fragment but rather delocalize the electron over the ZrO2Zr(OH)4(H2O)4 part of the model with spin density changes of -0.29 on O, -0.10 on Zr and -0.30 on the H atoms. This delocalization makes it difficult to assign valence states as we did for the MR calculations. For M=Zn, Cu the reaction may still be described as HAT (m=2), [Mm+(dn+1)] O−  [Mm+(dn+1)] OH−,

(5a)

although only about 70% of the total spin change is on O1. For M=Ni, Co, Fe and Mn, the (delocalized) spin changes on the two O atoms and on the two metal atoms are of similar magnitude which means that M06 describes the reactions of these systems as a mixture of both valence states, eq. (5a) and [Mm+(dn)] O2−  [M(m-1)+(dn+1)] OH−,

(5b)

and we cannot distinguish between HAT and PCET anymore. Regarding the transition state calculations obtained with DFT (M06 functional, data in Tables S10S11), the spin density changes are less pronounced than what is found with MR calculations. For Co_Zn and Co_Cu, the change in spin density on O1 is -0.38 and -0.37, respectively. For Co_Ni, the change in spin density on Ni is only -0.20, and also for Co_Co and Co_Fe, the change on Co is only -0.24 and -0.25, respectively. The energy barriers using DFT (12.4, 18.0, 19.0, 26.0 and 29.0 kcal/mol for Co_Zn, Co_Cu, Co_Ni, Co_Co and Co_Fe, respectively) follow the trend of the reaction energies (-5.1, -0.9, -1.9, 13.5, and 14.8 kcal/mol, respectively). We can conclude that DFT is not fully capturing the mechanism predicted by MR calculations. The DFT energy barriers are close to the RASPT2 ones for the systems that react through PCET (within 3-4 kcal/mol), but differ significantly for the systems that react through HAT (differences of 25-30 kcal/mol). The

- 22 -

ACS Paragon Plus Environment

Page 23 of 30 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

relationship between the energy barriers and the reaction energies by using both MR and DFT is shown in Figure S2, revealing a trend.

4. Summary and conclusion For Co2O2(OH)2 cobalt oxo hydroxo species deposited on Zr2O2(OH)4(H2O)4 models of the [Zr6O8]8+ nodes of the metal-organic framework NU-1000 MR calculations predict that they abstract H atoms from propane with a reaction energy of just 1 kcal/mol. The mechanism is proton coupled electron transfer (PCET) according to eqs. (2a/2b). On replacing one of the Co atoms with Fe or Mn, the mechanism remains the same; Co is reduced from Co+III to Co+II and the reaction energy increases a little, to 6 and 10 kcal/mol for Mn and Fe, respectively. In contrast, on doping with Ti the reaction energy increases substantially, to 85 kcal/mol. The mechanism is still PCET, but Ti is so stable in its +IV oxidation state that, in the (CoO)(TiO)(OH)2 species, it forces Co into the +II oxidation state. The reduction is now occurring from Co+II to Co+I, which is a very unfavorable process: H + Co2+(d7) O2−  Co+(d8) OH− Since Ni is more readily reduced from +III to +II than Co, on substitution of one of the Co atoms with Ni the energy for the PCET reaction becomes negative (-13 kcal/mol): H + Ni3+(d7) O2−  Ni2+(d8) OH− The most active (CoO)(MO)(OH)2 species are obtained when doping with the low-valent Zn and Cu metals, i.e., M=Zn, Cu. The latter prefer the M(m-1)+(dn)O− valence state to the Mm+(dn-1) O2− and the reaction proceeds as direct H atom transfer (HAT), cf. eqs. 1a/1b, with a reaction energy of -26 kcal/mol for both species.

- 23 -

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

The different mechanisms suggest different reactivity descriptors. While the spin population may be useful for HAT reactions, for PCET reactions both the charge on the proton receiving oxygen atom (basicity) and the reduction energy of the transition metal ion may play a role. The recommended descriptor for computational screening is the energy of H attachment. In general, we expect that, on replacing transition metal oxide species, e.g., [Co+IIIO-II]+ with lowvalent transition metals, e.g., [Zn+IIO-I]+, C-H bond activation via PCET switch to HAT and, hence, becomes more facile. We note, however, that exceedingly active species may not be good catalysts because they get easily poisoned and it takes some effort to regenerate them in the catalytic cycle. Moreover, selectivity will decrease if activity increases. Based on the insight gained here, we can discuss the reactivity of other dopants. We found that whether the mechanism is HAT or PCET depends on the oxidation state of the metal. If the metal is in high oxidation state, e.g. +III, and there is little spin density on the oxygen atom, the preferred route will be PCET. Which of the two metal ions is reduced in the PCET depends on the relative reduction potential. For example, a C-H activation using Cr+III as a dopant should undergo a PCET with Co+III reduced to Co+II, since Cr+III is a highly stable oxidation state. Another example would be Co_Ce: since Ce+IV is easily reduced to Ce+III, the reactant state will be Co+III/Ce+III (rather than Co+II/Ce+IV). On H addition, the mechanism will proceed through a PCET with Co reduction to Co+II (Ce+III will not be further reduced). Another example would be the doping with Ag, having a Co_Ag complex. In this case Ag is stable in a +I oxidation state, and therefore the spin will localize at oxygen forming a O− species. This should yield a HAT mechanism for the C-H activation reaction, with a very exothermic reaction energy (and therefore low energy barrier). Experimental verification of our prediction may not be easy for metal oxide species supported on MOF nodes because it will be difficult to synthesize them with precise control of nuclearity and

- 24 -

ACS Paragon Plus Environment

Page 24 of 30

Page 25 of 30 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

composition. However, reactivity studies on mass-selected mixed-metal gas phase clusters in combination with structure characterization by infra-red spectroscopy will be able to provide experimental answers11,16,6,7 to the questions addressed in the present computational study. Specifically, experiments by Asmis and co-workers for, e.g., [(CeO2)(VO2)]+ cations provided evidence for the larger stability of the Ce+III/V+V oxidation states compared to Ce+IV/V+IV,46 and such experiments are underway for gas phase clusters featuring the (CoO)+(MO)+ motif. ASSOCIATED CONTENT

The Supporting Information is available free of charge on XXXX. Description of computational models, additional DFT and multireference calculations, RASSCF molecular orbitals, cartesian coordinates and absolute electronic energies (both DFT and multireference) are reported.

AUTHOR INFORMATION Corresponding Authors

*J. S. E-mail: [email protected] *L. G. E-mail: [email protected] ORCID Carlo Alberto Gaggioli: 0000-0001-9105-8731 Joachim Sauer: 0000-0001-6798-6212 Laura Gagliardi: 0000-0001-5227-1396

Notes The authors declare no competing financial interest. ACKNOWLEDGEMENTS This work was supported by the Inorganometallic Catalyst Design Center, an EFRC funded by the DOE, Office of Basic Energy Sciences (DE-SC0012702), and by German Research Foundation

- 25 -

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

(DFG). The authors acknowledge the Minnesota Supercomputing Institute (MSI) at the University of Minnesota for providing computational resources. J.S. has been supported by a Reinhart Koselleck grant of German Research Foundation (DFG) and by the “Fonds der Chemischen Industrie”.

REFERENCES (1) Lunsford, J. H. The Catalytic Oxidative Coupling of Methane. Angew. Chem., Int. Ed. 1995, 34, 970-980. (2) Mayer, J. M. Hydrogen Atom Abstraction by Metal - Oxo Complexes : Understanding the Analogy with Organic Radical Reactions. Acc. Chem. Res. 1998, 31, 441–450. (3) Haber, J. Oxidation of Hydrocarbons. In Handbook of Heterogeneous Catalysis; Ertl, G.; Knözinger, H.; Weitkamp, J., Eds. VCH: Weinheim, 1997; Vol. 5, pp 2253-2274. (4) Muhler, M. Oxidative Dehydrogenation of Alcohols to Aldehydes, Ketones, and Carboxylic Acids. In Handbook of Heterogeneous Catalysis, Ertl, G.; Knözinger, H.; Weitkamp, J., Eds. VCH: Weinheim, 1997; Vol. 5, pp 2274-2284. (5) Shaik, S.; De Visser, S. P.; Ogliaro, F.; Schwarz, H.; Schröder, D. Two-State Reactivity Mechanisms of Hydroxylation and Epoxidation by Cytochrome P-450 Revealed by Theory. Curr. Opin. Chem. Biol. 2002, 6, 556–567. (6) Limberg, C. Oxygenation The Role of Radicals in Metal-Assisted Oxygenation Reactions. Angew. Chem. Int. Ed. 2003, 42, 5932-5954. (7) Sauer, J. C-H Bond Activation by Transition Metal Oxides. In Computational Modeling for Homogeneous and Enzymatic Catalysis: A Knowledge‐ Base for Designing Efficient Catalysts; Morokuma K.; Musaev D. G., Eds. Wiley-VCH: Weinheim, 2008; pp 231–244. (8) Schwarz, H. Thermal Hydrogen-Atom Transfer from Methane : A Mechanistic Exercise. Chem. Phys. Lett. 2015, 629, 91–101. (9) Cheng, M.; Goddard, W. A. The Critical Role of Phosphate in Vanadium Phosphate Oxide for the Catalytic Activation and Functionalization of n‑ Butane to Maleic Anhydride. J. Am. Chem. Soc. 2013, 135, 4600-4603. (10) Kropp, T.; Paier, J.; Sauer, J. Support Effect in Oxide Catalysis: Methanol Oxidation on Vanadia/Ceria. J. Am. Chem. Soc. 2014, 136, 14616–14625. (11) Li, J.; Zhou, S.; Zhang, J.; Schlangen, M.; Weiske, T.; Usharani, D.; Shaik, S.; Schwarz, H. Electronic Origins of the Variable Efficiency of Room-Temperature Methane Activation by Homo- and Heteronuclear Cluster Oxide Cations [XYO2]+(X, Y = Al, Si, Mg): Competition between Proton-Coupled Electron Transfer and Hydrogen-Atom Transfer. J. Am. Chem. Soc. 2016, 138, 7973–7981. (12) Li, Z.; Peters, A. W.; Bernales, V.; Ortuño, M. A.; Schweitzer, N. M.; Destefano, M. R.; Gallington, L. C.; Platero-Prats, A. E.; Chapman, K. W.; Cramer, C. J.; Gagliardi L.; Hupp J. T.;

- 26 -

ACS Paragon Plus Environment

Page 26 of 30

Page 27 of 30 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

Farha O. K. Metal-Organic Framework Supported Cobalt Catalysts for the Oxidative Dehydrogenation of Propane at Low Temperature. ACS Cent. Sci. 2017, 3, 31–38. (13) Markle, T. F.; Darcy, J. W.; Mayer, J. M. A New Strategy to Efficiently Cleave and Form C–H Bonds Using Proton-Coupled Electron Transfer. Sci. Adv. 2018, 4, eaat5776. (14) Wang, V. C.; Maji, S.; Chen, P. P.; Lee, H. K.; Yu, S. S.; Chan, S. I. Alkane Oxidation : Methane Monooxygenases , Related Enzymes , and Their Biomimetics. Chem. Rev. 2017, 117, 8574-8621. (15) Ye, S.; Neese, F. Nonheme Oxo-Iron (IV) Intermediates Form an Oxyl Radical upon Approaching the C – H Bond Activation Transition State. Proc. Natl. Acad. Sci. 2011 108, 12281233. (16) Ding, X.; Wu, X.; Zhao, Y.; He, S.-G., C–H Bond Activation by Oxygen-Centered Radicals over Atomic Clusters. Acc. Chem. Res. 2012, 45, 382-390. (17) Kondratenko, E. V.; Baerns, M. Oxidative Coupling of Methane. In Handbook of Heterogeneous Catalysis, 2nd ed.; G. Ertl, H. K., F. Schüth, J. Weitkamp, Ed. WileyVCH: 2008; Vol. 6, pp 3010-3023. (18) Feyel, S.; Döbler, J.; Höckendorf, R.; Beyer, M. K.; Sauer, J.; Schwarz, H., Activation of Methane by Oligomeric (Al2O3)x+ (x=3,4,5): The Role of Oxygen-Centered Radicals in Thermal Hydrogen-Atom Abstraction. Angew. Chem., Int. Ed. 2008, 47, 1946-1950. (19) Feyel, S.; Döbler, J.; Schröder, D.; Sauer, J.; Schwarz, H. Thermal Activation of Methane by Tetranuclear [V4O10]+. Angew. Chem. Int. Ed. 2006, 45, 4681-4685. (20) Kwapien, K.; Paier, J.; Sauer, J.; Geske, M.; Zavyalova, U.; Horn, R.; Schwach, P.; Trunschke, A.; Schlögl, R. Sites for Methane Activation on Lithium-Doped Magnesium Oxide Surfaces. Angew. Chem., Int. Ed. 2014, 53, 8774–8778. (21) Beck, B.; Harth, M.; Hamilton, N. G.; Carrero, C.; Uhlrich, J. J.; Trunschke, A.; Shaikhutdinov, S.; Schubert, H.; Freund, H. J.; Schlögl, R.; Sauer, J.; Schomäcker, R. Partial Oxidation of Ethanol on Vanadia Catalysts on Supporting Oxides with Different Redox Properties Compared to Propane. J. Catal. 2012, 296, 120–131. (22) Chen, K.; Bell, A. T.; Iglesia, E. The Relationship between the Electronic and Redox Properties of Dispersed Metal Oxides and Their Turnover Rates in Oxidative Dehydrogenation Reactions. J. Catal. 2002, 209, 35-42. (23) Fumiaki, A.; Tsunehiro, T. Propylene Oxide Synthesis and Selective Oxidation over Supported Metal Oxide Photocatalysts with Molecular Oxygen. Chem. Lett. 2006, 35, 468–473. (24) Sato, D.; Shiota, Y.; Juha, G.; Yoshizawa, K. Theoretical Study of the Mechanism of Valence Tautomerism in Cobalt Complexes. J. Phys. Chem. A 2010, 114, 12928-12935. (25) Tezgerevska, T.; Alley, K. G.; Boskovic, C. Valence Tautomerism in Metal Complexes : Stimulated and Reversible Intramolecular Electron Transfer between Metal Centers and Organic Ligands. Coord. Chem. Rev. 2014, 268, 23–40. (26) Phung, Q. M.; Feldt, M.; Harvey, J. N.; Pierloot, K. Toward Highly Accurate Spin State Energetics in First-Row Transition Metal Complexes: A Combined CASPT2/CC Approach. J. Chem. Theory Comput. 2018, 14, 2446–2455.

- 27 -

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

(27) Aquilante, F.; Autschbach, J.; Carlson, R. K.; Chibotaru, L. F.; Delcey, M. G.; De Vico, L.; Fdez. Galván, I.; Ferré, N.; Frutos, L. M.; Gagliardi, L.; Garavelli, M.; Giussani, A.; Hoyer, C. E.; Li Manni, G.; Lischka, H.; Ma, D.; Malmqvist, P. Å.; Müller, T.; Nenov, A.; Olivucci, M.; Pedersen, T. B.; Peng, D.; Plasser, F.; Pritchard, B.; Reiher, M.; Rivalta, I.; Schapiro, I.; SegarraMartí, J.; Stenrup, M.; Truhlar, D. G.; Ungur, L.; Valentini, A.; Vancoillie, S.; Veryazov, V.; Vysotskiy, V. P.; Weingart, O.; Zapata, F.; Lindh, R. Molcas 8: New Capabilities for Multiconfigurational Quantum Chemical Calculations across the Periodic Table. J. Comput. Chem. 2016, 37, 506-541. (28) Simons, M. C.; Ortuno, M. A.; Bernales, V.; Gaggioli, C. A.; Cramer, C. J.; Bhan, A.; Gagliardi, L. C−H Bond Activation on Bimetallic Two-Atom Co‑ M Oxide Clusters Deposited on Zr-Based MOF Nodes: Effects of Doping at the Molecular Level. ACS Catalysis 2018, 8, 2864 2869. (29) Evans, M. G.; Polanyi, M. Inertia and Driving Force of Chemical Reactions. Trans. Faraday Soc. 1938, 34, 11–24. (30) Zhao, Y.; Truhlar, D. G. The M06 Suite of Density Functionals for Main Group Thermochemistry , Thermochemical Kinetics , Noncovalent Interactions , Excited States , and Transition Elements : Two New Functionals and Systematic Testing of Four M06-Class Functionals and 12 Other Functionals. Theor Chem Acc. 2008, 215–241. (31) Frisch, M. J. .; Trucks, G. W. .; Schlegel, H. B. .; Scuseria, G. E. .; Robb, M. A. .; Cheeseman, J. R. .; Scalmani, G. .; Barone, V. .; Mennucci, B. .; Petersson, G. A. .; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr., J.A; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian, 09; Gaussian, Inc.: Wallingford CT, 2009. (32) Roos, B. O.; Taylor, P. R.; Siegbahn, P. E. M. A Complete Active Space SCF Method (CASSCF) Using a Density-Matrix Formulated Super-CI Approach. Chem. Phys. 1980, 48, 157−173. (33) Malmqvist, P. Å.; Rendell, A.; Roos, B. O. The Restricted Active Space Self-ConsistentField Method, Implemented with a Split Graph Unitary Group Approach. J. Phys. Chem. 1990, 94, 5477–5482. (34) Malmqvist, P. Å.; Pierloot, K.; Shahi, A. R. M.; Cramer, C. J.; Gagliardi, L. The Restricted Active Space Followed by Second-Order Perturbation Theory Method: Theory and Application to the Study of CuO2 and Cu2O2 Systems. J. Chem. Phys. 2008, 128, 204109.

- 28 -

ACS Paragon Plus Environment

Page 28 of 30

Page 29 of 30 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

(35) Andersson, K.; Malmqvist, P.-Å.; Roos, B. O.; Sadlej, A. J.; Wolinski, K. Second-Order Perturbation Theory with a CASSCF Reference Function. J. Phys. Chem. 1990, 94, 5483–5488. (36) Andersson, K.; Malmqvist, P. Å.; Roos, B. O. Second-Order Perturbation Theory with a Complete Active Space Self-Consistent Field Reference Function. J. Chem. Phys. 1992, 96, 1218– 1226. (37) Sauri, V.; Serrano-Andrés, L.; Shahi, A. R. M.; Gagliardi, L.; Vancoillie, S.; Pierloot, K. Multiconfigurational Second-Order Perturbation Theory Restricted Active Space (RASPT2) Method for Electronic Excited States: A Benchmark Study. J. Chem. Theory Comput. 2011, 7, 153-168. (38) Hirshfeld, F. L. Bonded-Atom Fragments for Describing Molecular Charge Densities. Theor. Chim. Acta 1977, 44, 129–138. (39) Lu, T.; Chen, F. Multiwfn: A Multifunctional Wavefunction Analyzer. J. Comput. Chem. 2012, 33, 580–592. (40) Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry, Vol. 6; New York, Wiley, 1988. (41) Feyel, S.; Schröder, D.; Rozanska, X.; Sauer, J.; Schwarz, H. Gas-Phase Oxidation of Propane and 1-Butene with [V3O7]+ : Experiment and Theory in Concert. Angew. Chem. Int. Ed. 2006, 45, 4677-4681. (42) Rozanska, X.; Sauer, J. Oxidative Dehydrogenation of Hydrocarbons by V3O7+ Compared to Other Vanadium Oxide Species. J. Phys. Chem. A 2009, 113, 11586-11594. (43) Saouma, C. T.; Mayer, J. M. Do Spin State and Spin Density Affect Hydrogen Atom Transfer Reactivity? Chem. Sci. 2014, 5, 21–31. (44) Kwapien, K.; Sierka, M.; Döbler, J.; Sauer, J. Reactions of H2, CH4, C2H6, and C3H8 with [(MgO)n]+ Clusters Studied by Density Functional Theory. ChemCatChem 2010, 2, 819-826. (45) Ye, S.; Neese, F. Nonheme oxo-iron (IV) intermediates form an oxyl radical upon approaching the C–H bond activation transition state. PNAS 2011, 108, 1228-1233. (46) Jiang, L.; Wende, T.; Claes, P.; Bhattacharyy, S.; Sierka, M.; Meijer, G.; Lievens, P.; Sauer, J.; Asmis, K. R. Electron Distribution in Partially Reduced Mixed Metal Oxide Systems: Infrared Spectroscopy of CemVnOo+ Gas-Phase Clusters. J. Phys. Chem. A 2011, 115, 11187–92. (47) Mafuné, F.; Masuzaki, D.; Nagata, T. Reduction Site in CenVmOk+ Revealed by Gas Phase Thermal Desorption Spectrometry. Top. Catal. 2018, 61, 42–48. For Table of Contents Only

- 29 -

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

- 30 -

ACS Paragon Plus Environment

Page 30 of 30