A dinuclear nickel peroxycarbonate complex: CO₂ addition promotes H₂O₂ release

Abstract

Nickel coordination compounds featuring Ni–O bonds are key structural motifs in both bioinorganic and synthetic chemistries. They serve as precursors for organic substrate oxidation and are commonly invoked intermediates in water oxidation and oxygen reduction schemes. Herein, we disclose a series of well-defined dinuclear nickel complexes that, upon treatment with CO₂ and H₂O₂, afford the first nickel-bound peroxycarbonate. This unprecedented nickel–oxygen intermediate is stabilized by hydrogen bonding templated across the bimetallic core. Contrasting copper and iron analogues, the nickel peroxycarbonate reversibly dissociates H₂O₂, a process that is shown to be accelerated by exogenous CO₂.

---

The ubiquity of Cu, Fe, and Mn enzymes has led to extensive studies interrogating metal-oxygen intermediates of mono- and bimetallic complexes of these ions.¹ In contrast, Ni is generally employed by Nature for C₁ conversion chemistry,² but has recently received attention with respect to oxidation reactivity given its role in the active sites of superoxide dismutase (NiSOD; Fig. 1A)³ and dioxygenase enzymes.⁴ Ni(II) complexes are typically inert to dioxygen, however modifications to the supporting ligand(s) have been demonstrated to override this tendency.⁵ Even so, the vast majority of nickel–oxygen complexes are accessed via addition of strong oxidants, including H₂O₂, peroxyacids, and alkylperoxides.⁶ Continued preparation and study of new classes of nickel–oxygen intermediates is critical to elucidating the chemistry of nickel metalloenzymes, informing fundamental mechanisms in synthetic oxidation catalysis, and providing a basis for comparison to copper and iron homologs.^6c

Fig. 1 Proposed mechanism for bioinorganic superoxide reduction to H₂O₂ (A) and representative examples of the diverse nickel–oxygen intermediates accessed synthetically via both O₂ activation and H₂O₂/base addition (B).

Given the high electronegativity of Ni, its low oxophilicity, and O(π)/Ni(d) repulsions, nickel–oxygen species can demonstrate potent oxidative reactivity.⁷ Mononuclear nickel oxidants—generated from the treatment of Ni(II) precursors with peroxyacids—are capable of oxidizing unactivated aliphatic C–H bonds, yet the nature of the active oxidant remains unknown.⁸ Tyrosinase-like Ni^III-μ(O)₂-Ni^III diamond cores are competent for intramolecular aliphatic⁹ and aromatic¹⁰ C–H oxygenation, ultimately sourcing the oxidizing equivalents from hydrogen peroxide (H₂O₂). They are likewise proposed to be the active oxidants in catalytic aromatic hydroxylation protocols.^1a,11 Recently, this has been scrutinized;¹² more exotic nickel–oxygen species have been reported under similar reaction conditions and cannot be strictly ruled out as reaction-relevant intermediates (Fig. 1B).^9,13

Given that many mechanistic proposals rely on self-assembly of monometallic precursors to dinuclear nickel–oxygen species, our lab has targeted well-defined bimetallic model systems supported by a bicompartmental ligand scaffold. Herein, we describe the synthesis of a series of multinuclear nickel complexes that serve as precursors for base-free H₂O₂ activation chemistry. These Ni(II) compounds were found to readily react with carbon dioxide (CO₂), affording a dinickel complex bridged by κ²-bicarbonate linkages. The CO₂ in these ligands plays a key role in subsequent H₂O₂ reactivity, yielding the first structurally characterized bimetallic peroxycarbonate complex, [LNi₂(O₂CO₂)(H₂O)₂]⁺. Contrasting previously reported peroxycarbonates,¹⁴[LNi₂(O₂CO₂)(H₂O)₂]⁺ releases H₂O₂ when warmed to 0 °C. In addition to expanding the library of characterized nickel–oxygen intermediates (Fig. 1B), this work lends credence to proposed peroxyacid adducts that are relevant to nickel-mediated alkane and alkene oxidations.^8a,c,d

Treating lithium phenolate pro-ligand, LiL, with two equiv. of Ni(OTf)₂, generates a dinickel complex¹⁵ featuring both inner- (¹⁹F δ = 0.81 ppm) and outer-sphere (¹⁹F δ = –74.42 ppm) triflate anions. The nickel(II) centres afford sharp paramagentically shifted resonances by ¹H NMR spectroscopy, a fingerprint correlated to the bis(triflate) cation [LNi₂(OTf)₂]⁺ (Scheme 1), via complementary single crystal X-ray diffraction (SCXRD) analysis. μ-OH ligands are purported intermediates and/or precursors in H₂O₂-mediated oxidation chemistry;¹⁶ these groups were targeted via treatment of [LNi₂(OTf)₂]⁺ with NaOH. Under anaerobic conditions, addition of base results in generation of a tetranuclear complex with μ-OH ligands linking two distinct LNi₂ motifs, the formulation of which was corroborated as [L₂Ni₄(OH)₄]²⁺ by SCXRD. The four metal ions adopt a roughly tetrahedral arrangement (τ₄ = 0.97),¹⁷ with bridging hydroxide and aryloxide ligands capping each of the six edges. The Ni–Ni distances across the aryloxide bridges are slightly elongated (3.90 Å) relative to those spanned by the hydroxide ligands (3.75 Å), suggesting that the intermolecular μ-hydroxo is more accessible with the semi-rigid dinucleating scaffold. When the same reaction is performed open to air, a distinct tetranickel complex results, with two bridging carbonate and two aquo ligands, [L₂Ni₄(CO₃)₂(H₂O)₂]²⁺, via the uptake of CO₂ from ambient air (Scheme 1). All four Ni(II) centres remain six-coordinate with C₂ symmetry and μ-κ²O,O:κO carbonate binding. The solid-state structure of this species highlights expansion of the inter-nickel distances between the two bimetallic units (Ni⋯Ni_ave. = 4.96 Å), and distinct Ni–Ni contacts within the two bicompartmental chelates, with the κ²-carbanato moieties enforcing a more compact Ni₂ spacing (viz. 3.45 vs. 3.81 Å). The solid state asymmetry observed for [L₂Ni₄(CO₃)₂(H₂O)₂]²⁺ is affirmed by the solution spectroscopy, and electrospray ionization mass spectrometry (ESI-MS) shows parent ion peaks consistent with both tetranuclear units, [L₂Ni₄(OH)₄]²⁺ and [L₂Ni₄(CO₃)₂(H₂O)₂]²⁺.

Scheme 1 Synthesis of well-defined multinuclear nickel hydroxide complexes and subsequent reactivity with carbon dioxide.

Whereas solutions of [L₂Ni₄(OH)₄]²⁺ prove air-stable, they readily react with concentrated CO₂ (1 atm) to generate a new species, as indicated by a slight blue shift in the UV-visible spectrum (Fig. 2, left) and a distinct paramagnetically shifted signature in the ¹H NMR spectrum.¹⁸ The same spectral features are reproduced when CO₂ (1 atm) is added to [L₂Ni₄(CO₃)₂(H₂O)₂]²⁺. Vibrational spectroscopy shows a sharp carbonyl stretch at 1687 cm⁻¹ that is sensitive to ¹³CO₂ isotopic labelling (Fig. 2, right), most consistent with a terminal (bi)carbonate assignment.^14h Whereas this reaction product has eluded characterization in the solid state, ESI-MS of reactions stemming from the treatment of either precursor with ^12/13CO₂ corroborates generation of a bis(bicarbonate) cation, [LNi₂(CO₃H)₂]⁺, the dinuclear nature of which is further supported by DOSY NMR experiments (Scheme 1).

Fig. 2 Electronic absorption spectra evidencing conversion of [L₂Ni₄(OH)₄]²⁺ () and [L₂Ni₄(CO₃)₂(H₂O)₂]²⁺ () to [LNi₂(CO₃H)₂]⁺ ( & , respectively) in MeCN solution (left). Solution IR spectra (MeCN) corroborating isotopic sensitivity of the bicarbonate (¹²C ;¹³C ) carbonyl stretch (right).

As an entry point to oxidative chemistry, we pursued H₂O₂ addition to these multinuclear Brønsted-basic nickel species. Peroxycarbonate ligands are shown to form from both CO₂ addition to reactive metal peroxo species^14b,d,f–k and H₂O₂ insertion into metal carbonatos.^14a Treating [L₂Ni₄(CO₃)₂(H₂O)₂]²⁺ or [LNi₂(CO₃H)₂]⁺ with excess aqueous H₂O₂ (50 wt%) at –30 °C resulted in an immediate colour change to pale purple. Low temperature ¹H NMR spectroscopy supported the formation of a mixture of paramagnetic species; cleaner reactivity to a single major product was observed when the oxidant addition was carried out in the presence of exogenous CO₂. Crystalline purple plates were formed when Et₂O/MeCN mixtures of this new product were maintained at low temperature (−17 °C). SCXRD analysis revealed generation of a dinuclear peroxycarbonate complex, [LNi₂(O₂CO₂)(H₂O)₂]⁺ (Fig. 3); the first nickel peroxycarbonate and the first structurally characterized bimetallic example.

Fig. 3 Synthesis, solid-state structure,‡ and reactivity of a nickel peroxycarbonate complex. Bond metrics are reported in angstroms. Substrate oxidations were conducted at 0 °C for 4 hrs in either MeCN or MeCN/DCM.

Contrasting what was proposed for dicopper complexes on a similar phenolate-bridged ancillary ligand,^14h the present dinickel peroxycarbonate shows preferential binding to a single Ni centre (Fig. 3). The metrical parameters highlight that the O₂CO₂-bound Ni ion in [LNi₂(O₂CO₂)(H₂O)₂]⁺ adopts a distorted octahedral coordination environment, with peroxycarbonate bond lengths similar to Suzuki's related monometallic Fe analogue.^14e The O–O bond is slightly elongated in the present case (cf. 1.466(6) vs. 1.455(5) Å) and more closely matches that of the recently reported peroxybicarbonate anion (1.469(2) Å).¹⁹ The second nickel centre in [LNi₂(O₂CO₂)(H₂O)₂]⁺ is likewise pseudo-octahedral with two water molecules completing the coordination sphere. A striking feature of the solid-state structure is the orientation and proximity of the aquo ligands to the formally anionic oxygen atoms of the peroxycarbonate, which are well within the range of H-bonding contacts (OH₂⋯O_ave. = 2.682(5) Å). H-bonding is a proven strategy for stabilizing reactive metal-oxygen species;²⁰ we hypothesize that nickel templated intramolecular H-bonding imbues added stability to the peroxycarbonate moiety in the present system.

The reported reactivity of base-metal peroxycarbonate complexes varies, with documented examples of both O–O bond scission^14a,c and thermal disproportionation to liberate O₂.^14h Moreover, peroxycarbonate complexes have been shown to be competent oxidants in their own right, demonstrating O-atom^14a,b,e–h and H-atom transfer^14a,f reactivity. In light of this precedent, we were keen to explore the thermal (in)stability of [LNi₂(O₂CO₂)(H₂O)₂]⁺, which demonstrated relatively slow conversion to [L₂Ni₄(CO₃)₂(H₂O)₂]²⁺ in MeCN solution at 0 °C (t_1/2ca. 40 min.). This is markedly different from Karlin's dicopper analogue, which reacts rapidly via thermal disproportionation at temperatures as low as –50 °C,^14h a distinction attributed to the intramolecular H-bonds (vide supra). A balanced reaction for formation of [L₂Ni₄(CO₃)₂(H₂O)₂]²⁺ necessitates the loss of H₂O₂—efforts to spectroscopically identify/quantify H₂O₂ were unsuccessful, but reactivity probes with phosphine reagents corroborated nearly quantitative peroxide release (92%; Fig. S21, ESI†).§ In this way, [LNi₂(O₂CO₂)(H₂O)₂]⁺ demonstrates novel reactivity for this ligand motif, acting as a reservoir for H₂O₂.

We next sought to explore the potential of [LNi₂(O₂CO₂)(H₂O)₂]⁺ to serve as a bona fide oxidant via low-temperature reactions with various substrates. Kinetic assays at 0 °C, under pseudo first-order conditions, displayed non-integer rate laws (Fig. 4 and ) for two electronically differentiated phosphines. These results are more consistent with reversible H₂O₂ binding than the direct reaction of PR₃ with [LNi₂(O₂CO₂)(H₂O)₂]⁺. A kinetic regime in which H₂O₂ binding to form [LNi₂(O₂CO₂)(H₂O)₂]⁺ (Fig. 3, k_–1) is rate competitive with phosphine oxidation (k₂) rationalizes the observed reaction profiles.¶ To further bolster this hypothesis, the rate of carbonate formation from [LNi₂(O₂CO₂)(H₂O)₂]⁺ in the presence () and absence () of H₂O₂ was likewise investigated. The zero-order dependence on Ni observed for formation of [L₂Ni₄(CO₃)₂(H₂O)₂]²⁺ supports rate determining H₂O₂ release (rather than dimerization), and addition of exogenous H₂O₂ suppresses conversion (Fig. 4 and , respectively). A comparison of three additional substrates (Fig. 3, inset) demonstrated no significant difference in reactivity relative to hydrogen peroxide controls, further supporting [LNi₂(O₂CO₂)(H₂O)₂]⁺ acts as a storehouse for H₂O₂.

Fig. 4 Reaction kinetics tracking [LNi₂(O₂CO₂)(H₂O)₂]⁺ reactivity under conditions outlined in the accompanying legend. Ar = tris(4-trifluoromethylphenyl).

Reaction atmosphere—N₂vs. CO₂vs. air—has been shown to play a role in peroxycarbonate reactivity.^14a Under CO₂, the decomposition of [LNi₂(O₂CO₂)(H₂O)₂]⁺ is immediate, even at low temperature (thawing MeCN), quantitatively furnishing bicarbonate complex [LNi₂(CO₃H)₂]⁺. Again, this chemistry proceeds via the release of H₂O₂ with concomitant CO₂ uptake. This reversible sequestration of H₂O₂ at a base-metal-bound peroxycarbonate is unprecedented, differentiating the present Ni reactivity from that of both its Fe^14a,c and Cu^14g congeners. Subsequent chemical steps—dimerization to form [L₂Ni₄(CO₃)₂(H₂O)₂]²⁺, substrate oxidation, or CO₂ addition (generating [LNi₂(CO₃H)₂]⁺)—serve to drive the equilibrium toward exhaustive H₂O₂ release.

In summary, well-defined dinuclear nickel complexes have been accessed by employing a phenolate-bridged bicompartmental ligand scaffold. Hydroxide installation results in highly Lewis basic moieties that favour intermolecular bridging ([L₂Ni₄(OH)₄]²⁺) or CO₂ uptake ([L₂Ni₄(CO₃)₂(H₂O)₂]²⁺). Ensuing hydrogen peroxide chemistry affords the first reported example of a nickel peroxycarbonate complex, the stability of which is attributed to metal-templated intramolecular H-bonding interactions. Unlike previously reported metal peroxycarbonates,¹⁴[LNi₂(O₂CO₂)(H₂O)₂]⁺ reacts via reversible H₂O₂ release, as established by kinetics assays. H₂O₂ dissociation is shown to be driven by CO₂ addition, accelerating oxidant release. Compounding ambiguity regarding the exact nature of reactive nickel–oxygen intermediates accessed from H₂O₂ and base (cf.Fig. 1B), this work demonstrates that CO₂ may likewise play a non-innocent role in altering the nickel speciation. Further studies investigating the fundamental chemistry of multinuclear nickel precursors with myriad catalysis-relevant oxidants are underway in our laboratory.

This work was supported by the University of Michigan and the NSF (XRD Instrumentation–CHE-0840456). We thank Dr. Eugenio Alvarado for NMR spectroscopy expertise, Drs. Jeff Kampf and Fengrui Qu for assistance with SCXRD and Claire R. Patterson for XRD data of [LNi₂(OTf)₂]⁺. Roy Wentz aided in the design and fabrication of custom glassware and the Szymczak lab shared instrumentation that made this work possible.

Data availability

The data supporting this article have been included as part of the ESI.† Crystallographic data have been deposited in the CCDC (2348000–2348003)—https://www.ccdc.cam.ac.uk/structures.

Conflicts of interest

There are no conflicts to declare.

Notes and references

1. (a) L. Vicens, et al. , ACS Catal., 2020, 10, 8611; (b) L. Que and W. B. Tolman, Nature, 2008, 455, 333; (c) E. A. Lewis and W. B. Tolman, Chem. Rev., 2004, 104, 1047; (d) L. M. Mirica, et al. , Chem. Rev., 2004, 104, 1013.
2. S. W. Ragsdale, Crit. Rev. Biochem. Mol. Biol., 2004, 39, 165.
3. (a) D. P. Barondeau, et al. , Biochem., 2004, 43, 8038; (b) J. Wuerges, et al. , Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 8569.
4. (a) J.-H. Jeoung, et al. , Angew. Chem., Int. Ed., 2016, 55, 3281; (b) T. C. Pochapsky, et al. , Nat. Struct. Bio., 2002, 9, 966.
5. (a) D. M. Beagan, et al. , J. Am. Chem. Soc., 2024, 146, 12375; (b) A. J. McNeece, et al. , J. Am. Chem. Soc., 2020, 142, 10824; (c) N. Zhao, et al. , J. Am. Chem. Soc., 2020, 142, 21634; (d) P.-C. Duan, et al. , J. Am. Chem. Soc., 2018, 140, 4929.
6. (a) H. Noh and J. Cho, Coord. Chem. Rev., 2019, 382, 126; (b) T. Corona and A. Company, Chem. Eur. J., 2016, 22, 13422; (c) M. R. Halvagar, et al., 3.17-Small Molecule Models: Cu, Ni, Co, in Comprehensive Inorganic Chemistry II, ed. J. Reedijk and K. Poeppelmeier, 2nd edn, Elsevier, Amsterdam, 2013, p. 455.
7. J. R. Winkler and H. B. Gray, in Molecular Electronic Structures of Transition Metal Complexes I, ed. D. M. P. Mingos, P. Day and J. P. Dahl, Springer, Berlin, Heidelberg, 2012, 17.
8. (a) J. Nakazawa, et al. , J. Am. Chem. Soc., 2013, 135, 6010; (b) S. Hikichi, et al. , Dalton Trans., 2013, 42, 3346; (c) T. Nagataki, et al. , Dalton Trans., 2007, 1120; (d) T. Nagataki, et al. , Chem. Commun., 2006, 4016.
9. (a) J. Cho, et al. , Inorg. Chem., 2006, 45, 2873; (b) S. Itoh, et al. , J. Am. Chem. Soc., 2001, 123, 11168; (c) K. Shiren, et al. , J. Am. Chem. Soc., 2000, 122, 254.
10. (a) T. Tano, et al. , Bull. Chem. Soc. Jpn., 2010, 83, 530; (b) K. Honda, et al. , Angew. Chem., Int. Ed., 2009, 48, 3304; (c) A. Kunishita, et al. , Inorg. Chem., 2009, 48, 4997.
11. (a) S. Muthuramalingam, et al. , Catal. Sci. Tech., 2019, 9, 5991; (b) Y. Morimoto, et al. , J. Am. Chem. Soc., 2015, 137, 5867.
12. K. Farshadfar and K. Laasonen, Inorg. Chem., 2024, 63, 5509.
13. (a) M. T. Kieber-Emmons, et al. , J. Am. Chem. Soc., 2006, 128, 14230; (b) E. J. Brown, et al. , Angew. Chem., Int. Ed., 2005, 44, 1392; (c) S. Hikichi, et al. , J. Am. Chem. Soc., 1998, 120, 10567.
14. (a) T. Tsugawa, et al. , Chem. Lett., 2015, 44, 330; (b) G. Meier and T. Braun, Angew. Chem., Int. Ed., 2012, 51, 12564; (c) H. Furutachi, et al. , J. Am. Chem. Soc., 2005, 127, 4550; (d) M. Yamashita, et al. , J. Am. Chem. Soc., 2005, 127, 7294; (e) K. Hashimoto, et al. , Angew. Chem., Int. Ed., 2002, 41, 1202; (f) M. Aresta, et al. , Inorg. Chem., 1996, 35, 4254; (g) I. Sanyal, et al. , J. Am. Chem. Soc., 1993, 115, 11259; (h) P. P. Paul, et al. , J. Am. Chem. Soc., 1991, 113, 5322; (i) M. Schappacher and R. Weiss, Inorg. Chem., 1987, 26, 1189; (j) L. Dahlenburg and C. Prengel, J. Organomet. Chem., 1986, 308, 63; (k) P. J. Hayward, et al. , J. Chem. Soc. D, 1969, 987b.
15. (a) H. T. Zhang, et al. , Angew. Chem., Int. Ed., 2023, 62, e202218859; (b) S. S. Massoud, et al. , Dalton Trans., 2015, 44, 2110; (c) Y. Gultneh, et al. , Polyhedron, 1998, 17, 3351.
16. A. Rajeev, et al. , ACS Catal., 2022, 12, 9953.
17. L. Yang, et al. , Dalton Trans., 2007, 955.
18. D. Huang, et al. , Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 1222.
19. Z. Yan, et al. , J. Am. Chem. Soc., 2023, 145, 22213.
20. (a) E. W. Dahl, et al. , Chem. Commun., 2018, 54, 892; (b) E. W. Dahl, et al. , J. Am. Chem. Soc., 2018, 140, 10075; (c) C. L. Ford, et al. , Science, 2016, 354, 741; (d) A. S. Borovik, Acc. Chem. Res., 2005, 38, 54; (e) C. E. MacBeth, et al. , Science, 2000, 289, 938.

---

Footnotes

† Electronic supplementary information (ESI) available. CCDC 2348000–2348003. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4cc02241f

‡ Thermal anisotropic displacement ellipsoids are shown at the 50% probability level; non-aquo hydrogen atoms and the triflate counterion are omitted for clarity. The hydrocarbyl ligand backbone is depicted as a wireframe for simplicity.

§ Control reactions of [L₂Ni₄(CO₃)₂(H₂O)₂]²⁺ and phosphine under O₂ did not result in any detectable oxidation to phosphine oxide on the timescale of these experiments.

¶ The difference in phosphine oxidation rate is insufficient to rule out direct reaction of the phosphine with [LNi₂(O₂CO₂)(H₂O)₂]⁺; however, in totality, the data are inconsistent with this mechanism.

---