Polyoxometalate-supported metal carbonyl derivatives: from synthetic strategies to structural diversity and applications

Abstract

Polyoxometalates (POMs) have demonstrated strong potential in various fields, such as catalysis, magnetism, medicine, photochemistry and materials science, mainly because of their remarkable physical and chemical properties. In the broad area of POM chemistry, the study of metal carbonyl-based POMs, as an important type of organometallic POM, has been a rather attractive direction in recent years. POM-supported metal carbonyl derivatives (PMCDs), which utilize the advantages of both POMs and metal carbonyls, have emerged as a promising class of molecules, and much effort has been devoted to their preparation and relevant applications in the last decade. Thus far, substantial development of the synthetic chemistry of PMCDs by using one-pot synthesis has been reported mainly by us. This review focuses on different structural complexes, accompanied by Lindqvist-type, Keggin-type, Dawson-type, and nonclassical-type PMCDs, and is meant to provide fodder and guidance for further exploration and discovery of more intriguing PMCDs with innovative architectures and remarkable functionality. Herein, we highlight and discuss the structural features of PMCDs based on various structural types. Furthermore, synthetic strategies and relevant applications, especially in terms of photochemical properties and catalysis, are reviewed. Ideally, these concepts and strategies can be extended to other organometallic POMs.

---

Jingkun Lu

Jingkun Lu was born in Henan, China in 1995 and received her B.S. and M.S. degrees in chemistry from Henan University. In 2014, she began to conduct scientific research in the group of Prof. Jingyang Niu in Henan Key Laboratory of Polyoxometalate Chemistry. Her research interests focus on the preparation, characterization and relevant catalytic properties of polyoxometalate-based metal carbonyl complexes.

Peipei He

Peipei He received her B.S. degree in applied chemistry at Henan University Minsheng College University in 2017. And in 2018, she was admitted to Henan University to pursue a master's degree. In the first year of her graduate school, she conducted a series of experiments using polyoxometalates as catalysts at the Key Laboratory of Polyoxometalates of Henan University. Then she went to Dalian Institute of Chemical Physics to study the synthesis of boron-doped carbon material catalysts and their role in aldehyde amidation.

Jingyang Niu

Jingyang Niu is the Special Professor of Chemistry of Henan Province at Henan University. He obtained his BA, MA and PhD degrees in 1986, 1989 and 1996 from Henan University, Northeastern Normal University and Nanjing University, respectively. His research interests focus on the synthesis, crystal structure and properties of novel polyoxometalates. He has received several awards, including the Distinguished Young Scholars of Henan Province, National government special allowance recipients, and Thousands of intellectual plans of Central Plain. He is the leader of the Henan Province Key Laboratory of Polyoxometalate at Henan University.

Jingping Wang

Jingping Wang is the Special Professor of Chemistry of Henan University. She received her BA and MA from Northeastern Normal University at 1986 and 1989, respectively. From 1989 to 2002 she was an assistant professor, a lecturer and an associate professor and then she became a professor at Henan University. She has received several awards, including the May 1st Labour Medal of Henan Province and Excellent University Key Teacher of Henan Province. Her research expertise is the synthesis, crystal structure and properties of novel polyoxometalates and their potential application in different fields.

---

1. Introduction

Polyoxometalates (POMs), a diverse family of anionic metal oxides based on early transition metals (e.g., Nb, Ta, V, Mo, W), have come to the fore due to their unique physical and chemical properties.^1–7 These molecules are built from corner-, edge- and face-shared metal–oxygen {MO_x} (x = 5, 6) polyhedra and span a very large diversity of morphologies (size and shape), resulting in a rich variety of structures involved in classical fields such as catalysis, photochemistry, electrochemistry, materials science, and medicine.^8–17 The first synthesized POM, (NH₄)₃[PMo₁₂O₄₀]·nH₂O, was reported by Berzelius in 1826.¹⁸ One century later, Keggin reported the crystallographic study of “H₃[PMo₁₂O₄₀]·29H₂O” based on powder X-ray diffraction,¹⁹ which was one of the most significant milestones in the history of POM research. The characterization of POMs and the systematic study of their properties began in the early 20^th century with Rosenheim.²⁰ In recent decades, POM research has seen tremendous growth, and a number of reviews have appeared that address different aspects of POM science.^21–27 One of the most interesting aspects of POM chemistry is the underlying deceptively simple synthetic procedures, which require a small number of steps or even just one step (one-pot syntheses).^28–33 Traditional “one-pot” synthesis with varying pH, temperature, concentration/type of metal oxide anion, heteroatom type/concentration, etc., has always been the main method to search for new POM clusters. Accompanying the development of new synthetic technologies, such as hydrothermal, microwave, refluxing and ionic liquid synthetic methods, applied to POM chemistry, the POM family has been significantly expanded and enriched.^34–38

The structural analogy between POMs and extended oxides was first noted by Baker and later expanded by Pope, Klemperer, Finke and Müller.^39–48 In view of the extensive literature on POM chemistry and the rapid developments in the field of POM coordination chemistry, the development of functional POM-based materials has been rather slow because POMs are usually crystalline solids that are hard to prepare. Recently, the most common route to the integration of POMs into functional architectures and devices rests on inorganic/organic hybrids. The functionalization of POM species by organic or organic groups has been heavily investigated, resulting in a new class of molecular complexes, but also extended POM compounds. The organic–inorganic hybrids are classified into two types according to the nature of the interaction between the organic and inorganic components. The first class (class I) involving noncovalent interactions has been reviewed by Müller and Zubieta, such as host–guest systems, and intermolecular complexes between POMs and organic substrates.^47,48 In the second class (class II), the organic and inorganic moieties are linked via strong covalent or non-covalent bonds. A special thematic issue in Chemical Reviews organized by Dolbecq divided class II into two parts.⁴⁹ The first section gave an exhaustive list of hybrid POMs where organic groups are covalently linked to POM units via p-block elements. The second part described hybrid POMs where the organic ligand is bound to d- and f-block elements grafted to the surface or encapsulated in the vacancy of a POM. The incorporation of d-block elements into vacant POM matrices is one of the oldest and most studied reactions in POM chemistry. In recent years, numerous studies have been devoted to the directed grafting of organic substrates onto the nucleophilic oxygen atoms of the POM core and to the introduction of the organometallic fragments into vacant POM complexes.^49,50 There are a limited number of POMs containing sufficient charge density at their surface oxygen atoms to covalently bind organometallics. And surface activation can be achieved by replacing Mo^VI or W^VI centers by one or more lower valent metals. In addition, the chemistry of organometallic derivatives of POMs is an area of growing interest. The organometallic derivatives of POMs provide a novel strategy to synthesize POM-based organic–inorganic hybrid materials because they may serve as models of solid oxide-supported organometallic compounds and the combination of the hard oxo ligand and soft ligands may result in novel properties. Since the pioneering work published by the groups of Ho and Klemperer, Zonnevijlle and Pope and Knoth and Harlow, organic-functionalized POMs, i.e., species in which one or some oxo or {MoO_x}ⁿ⁺ groups have been replaced with organic functional groups, have been studied extensively and now form the largest subclass of POM derivatives. In addition, Finke has emphasized that the organometallic POMs have two types, and the difference between organometallics supported on POMs and those incorporated into a POM is whether the organic species are firmly attached to a k³-O site of the POM surface oxygens.⁵¹

The research of POM-based organometallic derivatives is of huge significance in POM chemistry not only providing new dimensionalities in structures but also important owing to their materialistic advantages.^52–55 In this context, immobilization of metal carbonyl units onto the POM surface is expected to obtain functional compounds and emerges as an important category. POM-supported metal carbonyl derivatives (PMCDs) possess the dual advantages of metal carbonyl groups and POMs. In particular, PMCDs have found immerse interest as the chromophoric metal carbonyl forms fascinating supramolecular arrays, structure peculiarities and various size modifications which expands them in various thrust areas of research like single molecular magnets, photoluminescence, and catalysis.^56–59 However, among the reports of POM-supported organometallic compounds, very few papers describe the functionalization of POMs by metal carbonyl complexes. Despite their anionic charges, complete POMs have a rather low charge surface density. The nucleophilicity of the oxo ligands can be increased by substituting metals (Mo, W, V, Nb, Ta), and the resulting mixed addenda polyanions show enhanced reactivity towards organometallic fragments, which is exemplified by the grafting of various d⁶-fac-{ML₃} units ({M(CO)₃⁺}) Herein, we will focus our emphasis on the POM-supported metal carbonyl derivatives, and the purpose of this review is to give an extended survey of these compounds and their applications.

2. Synthetic strategies

In 1998, a comprehensive review by Gouzerh and Proust covered the sizeable literature on the organic and organometallic derivatives of POMs.⁵⁰ Furthermore, as an important subclass of POM-based organometallic derivatives, PMCDs have been synthesized and characterized for nearly forty years. PMCDs possess the dual advantages of metal carbonyl groups and POMs. Since the report of the first metal carbonyl adducts, [Nb₂W₄O₁₉M(CO)₃]³⁻ (M = Mn and Re), in 1980,⁶⁰ the synthesis and characterization of POM-supported organometallic derivatives have received much attention, while tricarbonyl derivatives of the hexaniobate and hexatantalate anions have been recently reported by Pope.⁶¹ As a consequence, syntheses are very often relatively difficult (mainly because polyanions do not have sufficient charge density to combine with metal carbonyl groups, along with their poor solubility in the same solvent, their radiation- or thermoinstability and the high costs of metal carbonyl compounds), whereas those of POM-supported carbonyl metal derivatives are relatively less difficult. Moreover, the manganese/rhenium carbonyl derivatives of POMs, with rich behaviours displayed by the d⁶ metal carbonyl compounds and the POMs, are especially important. Synthetically, the routes to produce new PMCDs are often very simple synthetic manipulations, and most of these hybrid compounds have been obtained by one-pot aqueous/hydrothermal methods. The rational synthesis of PMCDs is still a great challenge because one or more of the reaction conditions, such as temperature, pH, stoichiometry, reaction time, medium, and template identity, can have a considerable influence on the reaction outcome. In most cases, the synthetic strategy has been used to react organic-soluble forms of the POMs, generally via tetrabutylammonium counterions, with metal carbonyl complexes in organic solvents such as CH₃CN, CH₂Cl₂ and CH₃COOH. To date, discrete integrated metal carbonyls, which are compounds where the metal carbonyl component is incorporated into the polyoxometal framework, have been obtained in several ways, including the incorporation of a metal carbonyl into a lacunary POM, the oxidation of carbonyl dimers [{Cp*M(CO)₂}₂](M = W or Mo) and various aggregation processes triggered by protons and/or Lewis acids in aqueous or non-aqueous media.^62,63 Since 2008, our group has been dedicated to synthesizing novel PMCD structures, and we have obtained many new structures in H₂O–CH₃OH or H₂O–CH₃CN mixed solvents using one-pot aqueous synthesis. In this new review, we highlight the structural variety of reported PMCDs, and our aim is to survey the recent significant achievements and provide key information for designing novel structures with improved properties directed by their structure–property relationships.

3. Representative structure types of PMCDs

3.1 Lindqvist-type PMCDs

As early as 1980, Klemperer et al. reported the synthesis and characterization of the first example of PMCDs, [(OC)₃M(Nb₂W₄O₁₉)]³⁻ (M = Mn, Re) (Fig. 1), containing metal tricarbonyl units bonded to a triangle of oxygen atoms on the surface of the [Nb₂W₄O₁₉]⁴⁻ anion.⁶⁰ The complexes were prepared by the reaction of [(n-C₄H₉)₄N]₄(Nb₂W₄O₁₉), prepared from aqueous Na₂[(CH₃)₄N]₂(Nb₂W₄O₁₉) by using a cation-exchange resin, with an equimolar amount of [(OC₃)M(NCCH₃)₃](PF₆) (M = Re/Mn) in acetonitrile and characterized by infrared (IR) and nuclear magnetic resonance (NMR) spectroscopy. It was not 1985 that Klemperer investigated these complexes in more detail by using single-crystal X-ray diffraction analysis.⁶⁴ Specifically, the X-ray crystallographic study of [(OC)₃M(Nb₂W₄O₁₉)]³⁻ (M = Mn, Re) indicated that both metal sites were assumed to be 1/3 Nb and 2/3 W and that M(CO)₃⁺ (M = Re/Mn) was bound to three bridging oxygens of [Nb₂W₄O₁₉]⁴⁻ as expected from the 18-electron rule. Subsequently, Klemperer et al. reported the first polyoxoanion-supported rhodium and iridium carbonyl complexes,⁶⁵ [{(OC)₂Rh}₅(Nb₂W₄O₁₉)₂][(n-C₄H₉)₄N]₃, [{(OC)₂Rh}₃(Nb₂W₄O₁₉)₂][(n-C₄H₉)₄N]₅, [{(OC)₂Ir}₂(Nb₂W₄O₁₉)₂][(n-C₄H₉)₄N]₅, and [(OC)₂Ir(P₃O₉)][{(C₆H₅)₃P}₂N]₂. In addition, [{(OC)₂Rh}₅(Nb₂W₄O₁₉)₂][(n-C₄H₉)₄N]₃ was prepared by the reaction of either [{(C₇H₈)Rh}₅(Nb₂W₄O₁₉)₂][(n-C₄H₉)₄N]₃ with CO in a nitromethane solution or Nb₂W₄O₁₉[(n-C₄H₉)₄N]₄ with [(OC)₂RhCl]₂ in a 1,2-dichloroethane solution. [{(OC)₂Rh}₃(Nb₂W₄O₁₉)₂][(n-C₄H₉)₄N]₅ was prepared by the reaction of Nb₂W₄O₁₉[(n-C₄H₉)₄N]₄ with [(OC)₂RhCl]₂ in chloroform. The structure of the [{(OC)₂Rh}₃(Nb₂W₄O₁₉)₂]⁵⁻ anion was obtained by removing two (OC)₂Rh⁺ units from the structure of the [{(OC)₂Rh}₅(Nb₂W₄O₁₉)₂]³⁻ anion. [{(OC)₂Ir}₂(Nb₂W₄O₁₉)₂][(n-C₄H₉)₄N]₅ was synthesized by bubbling CO into a CH₃CN solution of [{(C₈H₁₂)Ir}₂H(Nb₂W₄O₁₉)₂][(n-C₄H₉)₄N]₅, and in this structure, two octahedral Nb₂W₄O₁₉⁴⁻ anions are linked together edge-to-edge by a proton and two (OC)₂Ir⁺ cations.

Fig. 1 Ball-and-stick model of the [(OC)₃M(Nb₂W₄O₁₉)]³⁻ (M = Mn, Re). Colour code: O, red; C, black; Nb/W, green; Mn/Re, purple.

Similarly, [(OC)₂Ir(P₃O₉)][{(C₆H₅)₃P}₂N]₂ was prepared by substitution of cyclooctadiene ligands in [(C₈H₁₂)Ir(P₃O₉)][{(C₆H₅)₃P}₂N]₂ with CO ligands.

In 1981, Day and co-workers reported preliminary results of the first X-ray crystallographic structure determinations of polyoxoanion-supported organometallic anions,⁶⁶ [(η⁵-C₅H₅)Ti(Mo₅O₁₈)]³⁻ and [(η⁵-C₅H₅)Ti(Mo₅O₁₈)MoO₂Cl]²⁻ (Fig. 2a and b), and the synthesis and structure of the first bifunctional polyoxoanion-supported organometallic, [(η⁵-C₅H₅)Ti(Mo₅O₁₈)Mn(CO)₃]²⁻. [(η⁵-C₅H₅)Ti(Mo₅O₁₈)MoO₂Cl]²⁻ was prepared by the reaction of [(η⁵-C₅H₅)Ti(Mo₅O₁₈)]³⁻ and 2 equiv. of aqueous HCl in a CH₃CN solution. The X-ray diffraction results of the [(η⁵-C₅H₅)Ti(Mo₅O₁₈)MoO₂Cl]²⁻ polyoxoanion showed that a MoO₂Cl⁺ unit is bonded to a triangle of three bridging oxygens in this structure. However, [(η⁵-C₅H₅)Ti(Mo₅O₁₈)MoO₂Cl]²⁻ is unstable towards water and heat, decomposing to [Mo₆O₁₉]²⁻ in CH₃CN. By replacement of MoO₂Cl⁺ with a Mn(CO)₃⁺ unit, [(η⁵-C₅H₅)Ti(Mo₅O₁₈)Mn(CO)₃]²⁻ was prepared from [(η⁵-C₅H₅)Ti(Mo₅O₁₈)]³⁻ and [(OC)₃Mn(NCCH₃)₃][PF₆] in CH₃CN and was relatively stable towards both water and heat in CH₃CN. The existence of [(η⁵-C₅H₅)Ti(Mo₅O₁₈)Mn(CO)₃]²⁻ was confirmed by the carbonyl region in the IR spectrum and the NMR spectrum.

Fig. 2 Ball-and-stick model of the [(η⁵-C₅H₅)Ti(Mo₅O₁₈)]³⁻ (a) and [(η⁵-C₅H₅)Ti(Mo₅O₁₈)MoO₂Cl]²⁻ (b) polyoxoanions. Colour code: O, red; C, black; Ti, grey; Cl, light green.

In 1993, Klemperer et al. reported five new Ru^I–Ru^I tetracarbonyl complexes: [(P₃O₉)₂Ru₂(CO)₄]⁴⁻, [{CpTi(W₅O₁₈)}₂Ru₂(CO)₄]⁴⁻, [(CH₃CN)₆Ru₂(CO)₄]²⁺, [(PPh₃)₂(CH₃CN)₄Ru₂(CO)₄]²⁺, and [(C₅H₅N)₆Ru₂(CO)₄]²⁺.⁶⁷ Notably, only two of these complexes were structurally characterized using single-crystal diffraction techniques: [(P₃O₉)₂Ru₂(CO)₄][(n-C₄H₉)₄N]₄·2CH₃CN and [(PPh₃)₂(CH₃CN)₄Ru₂(CO)₄](PF₆)₂ (Fig. 3a and b). [{CpTi(W₅O₁₈)}₂Ru₂(CO)₄]⁴⁻ was obtained by the reaction of [(CH₃CN)₆Ru₂(CO)₄](PF₆)₂ with 2 equiv. of TBA₃[CpTiW₅O₁₈] in a CH₂Cl₂ solution.

Fig. 3 Ball-and-stick model of [(P₃O₉)₂Ru₂(CO)₄]⁴⁻ (a) and [(PPh₃)₂(CH₃CN)₄Ru₂(CO)₄]²⁺ (b). Colour code: O, red; Ru, orange; P, yellow; C, black; N, dark green.

In 2001, Pope and co-workers reported ten 1 : 1 and 2 : 1 complexes of [Mn(CO)₃]⁺ and [Re(CO)₃]⁺ with [Nb₆O₁₉]⁸⁻ and [Nb₆O₁₉]⁸⁻ as potassium salts.⁶⁸ The complexes contain M(CO)₃ groups attached to the surface bridging oxygen atoms of the hexametalate anions to yield structures with nominal C_3V (1 : 1), D_3d (trans 2 : 1), and C_2V (cis 2 : 1) symmetry (Fig. 4a, b, and c). The formation of 2 : 1 complexes and the existence of both cis and trans isomers were first deduced from the NMR spectra. K₇[Re(CO)₃Nb₆O₁₉] was first synthesized by using an aqueous hydrothermal reaction technique, while the other complexes were obtained conventionally in aqueous solution. The discovery of K₇[Re(CO)₃Nb₆O₁₉] not only enriches the diversity of the structural chemistry of PMCDs but also provides inspiring directions to design and prepare many more intriguing structures. The Re compounds are stable up to 400–450 °C in the solid state and can be formed under hydrothermal conditions at 130 °C and pH > 10, while the Mn compounds lose CO at temperatures above 200 °C. The high stability of Re compounds may be assumed to be associated with the loss of Re₂O₇ formed by oxidation or disproportionation.

Fig. 4 Polyhedral/ball-and-stick representations of [M₆O₁₉{M′(CO)₃}_n]⁽⁸⁻ⁿ⁾⁻ (M = Nb, Ta; M′ = Mn, Re; n = 1, 2). Colour code: O, red; C, black; Re/Mn, pink; MO₆, dark green.

In 2003, Gouzerh et al. reported three crystal structures of (nBu₄N)₂[Re(CO)₃(H₂O){Mo₅O₁₃(OMe)₄(NO)}], (nBu₄N)₃[Na{Mo₅O₁₃(OMe)₄(NO)}₂{Mn(CO)₃}₂], and (nBu₄N)₄[Mn(H₂O)₂{Mo₅O₁₆(OMe)₂}₂{Mn(CO)₃}₂], which were obtained by the reaction of neutral or cationic manganese carbonyl species with the oxo-nitrosyl complex [Na(MeOH){Mo₅O₁₃(OCH₃)₄(NO)}]²⁻ and characterized by single-crystal X-ray diffraction.⁶⁹ Stirring of a 1 : 1 mixture of [Na(MeOH){Mo₅O₁₃(OCH₃)₄(NO)}]²⁻ and {Mn(CO)₃}⁺ in CH₃OH at room temperature can lead to the isolation of (nBu₄N)₂[Re(CO)₃(H₂O){Mo₅O₁₃(OMe)₄(NO)}]. In this paper, although (nBu₄N)₂[Re(CO)₃(H₂O){Mo₅O₁₃(OMe)₄(NO)}] and (nBu₄N)₂[Mn(CO)₃(H₂O){Mo₅O₁₃(OMe)₄(NO)}] are isomorphous, the crystal structure of only the Re-containing compound was obtained and fully analysed. As shown in Fig. 5a and b, the organometallic {Re(CO)₃}⁺ fragment was bonded to two adjacent axial oxygen atoms of the {Mo₅} ligand and achieved an 18-electron configuration by coordination to a molecule of water. (nBu₄N)₃[Na{Mo₅O₁₃(OMe)₄(NO)}₂{Mn(CO)₃}₂] was prepared by refluxing a mixture of [Na(MeOH){Mo₅O₁₃(OCH₃)₄(NO)}]²⁻ with either {Mn(CO)₃}⁺ or [MnBr(CO)₅] in CH₃OH. Each {Mn(CO)₃}⁺ fragment was linked to a distinct {Mo₅} unit through the oxygen atoms of two adjacent methoxo ligands and one bridging oxo ligand (Fig. 5c and d). Unlike the usual rule for PMCDs, where the {Mn(CO)₃}⁺ fragment (M = Mn or Re) binds to a triangle of bridging oxygen atoms similar to d⁶-fac-ML₃ fragments in general, such a coordination mode was unknown for the {Mo₅} ligand until the characterization of (nBu₄N)₃[Na{Mo₅O₁₃(OMe)₄(NO)}₂{Mn(CO)₃}₂]. In detail, each [{Mo₅O₁₃(OMe)₄(NO)}{Mn(CO)₃}]²⁻ unit interacts with Na⁺ through its four axial oxo ligands in such a way that Na⁺ displays distorted square-antiprismatic coordination. (nBu₄N)₄[Mn(H₂O)₂{Mo₅O₁₆(OMe)₂}₂{Mn(CO)₃}₂] can be obtained by refluxing a mixture of (nBu₄N)₂[Mo₂O₇] and [MnBr(CO)₅] in non-deaerated methanol or by adding (nBu₄N)₂[Mo₂O₇] to a mixture of Mn(NO₃)₂·4H₂O and [MnBr(CO)₅] in boiling methanol. The structure of this complex can be viewed as consisting of two independent [Mo₅O₁₆(OMe)₂}{Mn(CO)₃}]³⁻ units connected by a {Mn(H₂O)₂}²⁺ linker. Each {Mn(CO)₃}⁺ fragment is linked to a distinct [Mo₅O₁₆(OMe)₂]⁴⁻ through the oxygen atoms of the methoxo ligands and one bridging oxo ligand (Fig. 5e and f). In addition, the [Mo₅O₁₆(OMe)₂]⁴⁻ ion is a new member of the family of monovacant Lindqvist-type POMs and has not been previously characterized because this complex has limited stability in an uncomplexed form. In addition to side-on coordination to a {Mn(CO)₃}⁺ cation, octahedral coordination is achieved when each [Mo₅O₁₆(OMe)₂]⁴⁻ acts as a bidentate ligand towards the Mn^II centre and two mutually trans molecules of water are also present.

Fig. 5 Ball-and-stick representations and polyhedral representations of [Re(CO)₃(H₂O){Mo₅O₁₃(OMe)₄(NO)}]²⁻ (a) and (b), [Na{Mo₅O₁₃(OMe)₄(NO)}₂{Mn(CO)₃}₂]³⁻ (c) and (d), and [Mn(H₂O)₂{Mo₅O₁₆(OMe)₂}₂{Mn(CO)₃}₂]⁴⁻ (e) and (f). Colour code: O, red; C, black; Mo, blue; Re, pink; Mn, lilac; Na, yellow; MoO₆, dark green.

With continuous research on PMCDs, our group has reported an isopentatungstate-supported rhenium carbonyl derivative, KH[(CH₃)₄N]₃{[Re(CO)₃]₄[(μ₂-OH)(μ₃-O)(W₅O₁₈)]}·6H₂O in 2016.⁵⁶ As shown in Fig. 6, the [{Re(CO)₃}₄(μ₂-OH)(μ₃-O)(W₅O₁₈)]⁵⁻ polyoxoanion can be regarded as a four-metal carbonyl unit attached to a [W₅O₁₈]⁶⁻ subunit, which is closely related to the Lindqvist hexatungstate structure with one {WO₆} octahedron removed from {W₆O₁₉}. More intriguingly, the trimetric metal model cluster {Re₃} is capped on the [W₅O₁₈]⁶⁻ subunit through four Re–O–W bonds, and the structure possesses the largest number of metal carbonyl groups among PMCDs. Furthermore, the compound demonstrated good capability of catalysing alkenes and was found to efficiently catalyse the epoxidation of cyclooctene with high conversion (98.9%) and excellent selectivity (99%). The high catalytic efficiency in epoxidation of alkenes may be attributed to the pentatungstate framework.

Fig. 6 (a) Polyhedral representation of [W₆O₁₉]²⁻ and [W₅O₁₈]⁶⁻; WO₆ octahedral, green; (b) ball-and-stick representation of [{Re(CO)₃}₄(μ₂-OH)(μ₃-O)(W₅O₁₈)]⁵⁻; (c) ball-and-stick representation of {[Re(CO)₃]₄[(μ₂-OH)(μ₃-O)]}⁺. Colour code: Re, purple; O, red; C, gray; WO₆, green.

3.2 Keggin-type PMCDs

In 1979, Knoth et al. reported the first heteropolyanions that contain metal–metal bonds,⁷⁰ including [CpFe(CO)₂SnW₁₁PO₃₉]⁴⁻ (Cp = π-C₅H₅), [(OC)₃Co(SnW₁₁SiO₃₉)₂]¹¹⁻, [π-C₃H₅Pd(SnW₁₁PO₃₉)₂]¹¹⁻, and a high-molecular-weight completely inorganic polymer, [(OC)₃CoGeW₁₁SiO₄₀⁵⁻]_n. These organic derivatives of Keggin-structure heteropolyanions are prepared by reacting the “unsaturated” Keggin fragments [SiW₁₁O₃₉]⁸⁻, [SiMo₁₁O₃₉]⁸⁻, and [PW₁₁O₃₉]⁷⁻, with transition metal complexes that contain trichlorostannane or trichlorogermane ligands. These polyoxoanions were isolated as trimethylammonium, tetramethylammonium, and trimethylsulfonium salts and further characterized by elemental analysis and IR and NMR spectroscopy. Four years later, Knoth et al. obtained a new disubstituted Keggin-type PMSC, [{CpFe(CO)₂Sn}₂W₁₀PO₃₈]⁵⁻, by reacting CpFe(CO)₂SnCl₃ with an aqueous solution of sodium dihydrogen phosphate at pH = 8.6. The structure of this complex was demonstrated by W NMR and P NMR spectra, but no crystallographic data were obtained.

In 1987, Siedle et al. isolated several novel Keggin ion derivatives, [(Ph₃P)₂Rh(CO)(CH₃CN)]_nXM₁₂O₄₀ (X = P/Si; M = W/Mo; n = 3, 4), that contain coordinatively unsaturated [(Ph₃P)₂Rh(CO)]⁺ units.⁷¹ These complexes were structurally characterized by IR, NMR, and X-ray absorption spectroscopy three years later, and their catalytic activity and selectivity for olefin isomerization and hydroformation were also carefully described for the first time.

In 2008, Sadakane et al. reported a novel carbonyl–ruthenium-substituted undecatungstosilicate, [α-SiW₁₁O₃₉Ru^II(CO)]⁶⁻, which has been fully characterized in solution and in the solid state.⁷² More importantly, it was the first example of a metal carbonyl moiety being fully incorporated into the POM framework by means of a hydrothermal method. The incorporated ruthenium(II) shows a well-defined reversible redox couple, and the one-electron-oxidized ruthenium(III) derivative [α-SiW₁₁O₃₉Ru^III(H₂O)]⁵⁻ is unexpectedly stable in aqueous solution compared to carbonyl–ruthenium coordination complexes. A CV study also indicated that the lacunary α-Keggin-type polyoxotungstate was able to stabilize the uncommon Ru^III(CO) moiety, which usually releases its CO molecule in Ru^III(CO)–organic derivatives too quickly to be detected.

In 2012, our group reported a series of compounds based on polyoxoanions and manganese carbonyl groups: [Mn(CO)₃(CH₃CN)₃]_n[α-XM₁₂O₄₀] [n = 3, X = P^V, M = Mo^VI, W^VI], [Mn(CO)₃(CH₃CN)₃]_n[α-XM₁₂O₄₀]·1.5H₂O [n = 4, X = Si^IV, M = Mo^VI], and [Mn(CO)₃(CH₃CN)₃]_nH[α-XM₁₂O₄₀] [n = 3, X = Ge^IV, M = Mo^VI].⁷³ Owing to the relatively high activity of POM precursors in aqueous solution and the stability of [Mn(CO)₅]Br in non-aqueous solvents, we chose the CH₃CN/H₂O mixed solvent to explore the reactions of POM precursors with [Mn(CO)₅]Br. The molecular structural unit of these compounds consists of [Mn(CO)₃(CH₃CN)_n]⁺ cations and a well-known α-Keggin-type [α-PMo₁₂O₄₀]³⁻ anion. Considering the stationary fac-configuration of the [Mn(CO)₃(CH₃CN)_n]⁺ cation, we attempted to obtain the isomer containing mer-[Mn(CO)₃(CH₃CN)_n]⁺ cations by changing the experimental conditions; however, our attempts failed.

In the same year, our group reported two novel trivacant Keggin-type POM-based manganese carbonyl derivatives: K₈[(OC)₃Mn(A-α-H₂GeW₉O₃₄)]·10H₂O and K₈[(OC)₃Mn(A-α-H₂SiW₉O₃₄)]·11H₂O.⁵⁷ The two compounds were synthesized by degradation of the metastable [γ-XW₁₀O₃₆]⁸⁻ (X = Ge^IV, Si^IV) in CH₃CN–H₂O solvent (1 : 2, vol.). X-ray diffraction analysis indicates that the two complexes are isomorphic and consist of a [(OC)₃Mn]⁺ group and a trivacant [A-α-H₂XW₉O₃₄]⁸⁻ (X = Si^IV, Ge^IV) fragment, representing the first examples of trivacant Keggin-type metal carbonyl derivatives (Fig. 7). In this case, the [γ-XW₁₀] (X = Si^IV, Ge^IV) unit as a starting material was transformed to the [A-α-XW₉] unit in a CH₃CN-H₂O mixed solvent at pH = 7.04 or 7.16 at 80 °C. Moreover, the [γ-XW₁₀] unit plays a vital role in the formation of the two compounds, and good crystals of both compounds can be obtained in the range of CH₃CN–H₂O = 1 : 2–1 : 5, where the optimum ratio is 1 : 2. It should be noted that the [Mn(CO)₃] unit prefers to coordinate at the less nucleophilic bridging oxygens on the cap of the Keggin unit rather than at the lacunary sites. The reasons are as follows: firstly, the [(OC)₃Mn(A-α-H₂GeW₉O₃₄)]⁸⁻ polyoxoanion has twelve terminal oxygen atoms at the lacunary sites, and all of which have W–O double bands in the range of 1.750–1.767 Å.^56,59 From the viewpoint of the electronic structure, it is difficult for these terminal oxygen atoms to coordinate with the [Mn(CO)₃] unit because of their poor bonding ability. Secondly, some documents have confirmed that several different two-electron donor ligands L can react with carbonyl metal units to afford kinetically stable d⁶ low-spin octahedral 18-electron moieties fac-[(OC)₃ML₃]⁺ (M = Mn, Re).^66–69 So it is most likely that trigonal adjacent surface oxygens in Keggin POMs as two-electron L ligands coordinate with [(OC)₃M]⁺ featuring the fac-[(OC)₃MO₃]⁺ moiety. In the trivacant [A-α-H₂GeW₉O₃₄]⁹⁻ polyoxoanion, there is only an intact W₃O₉ triad on the cap of the Keggin unit. From the viewpoint of coordination chemistry and steric configuration, in order to favour an 18-electron configuration, the [Mn(CO)₃] unit prefers to coordinate at the (notionally) less nucleophilic bridging oxygens on the cap of the Keggin unit.

Fig. 7 (a) Ball-and-stick representation of K₈[(OC)₃Mn(A-α-H₂GeW₉O₃₄)]·10H₂O. (b) Polyhedral/ball-and-stick representation of K₈[(OC)₃Mn(A-α-H₂GeW₉O₃₄)]·10H₂O. (c) and (d) The packing arrangements of K₈[(OC)₃Mn(A-α-H₂GeW₉O₃₄)]·10H₂O along the a-axis. The hydrogen atoms, K cations and lattice water molecules are omitted for clarity.

In 2013, Hill et al. prepared a new series of complexes that contain two electron-donating groups, {M(CO)₃}⁺ ions (M = Re/Mn), on one polytungstate electron-accepting group (Fig. 8).⁵⁸ By utilizing [X₂W₂₂O₇₄(OH)₂]¹²⁻ (X = Sb, Bi) as synthetic precursors, these authors used these POM ligands to coordinate fac-{M(CO)₃}⁺ (M = Re/Mn) fragments and obtained four discrete Krebs-type “slipped-sandwich” structures: Na₁₁H[Sb₂W₂₀O₇₀{Re(CO)₃}₂]·34H₂O, Na₁₁H[Bi₂W₂₀O₇₀{Re(CO)₃}₂]·33H₂O, K₉Na₃[Sb₂W₂₀O₇₀{Mn(CO)₃}₂]·32H₂O, and K₉Na₃[Bi₂W₂₀O₇₀{Mn(CO)₃}₂]·32H₂O. The four compounds have been structurally characterized by single-crystal X-ray diffraction, and the convenient preparation of these structures was conducted in a weakly acidic aqueous solution (pH = 5–6). All four metal-donor-POM-acceptor compounds have similar structures and contain two identical β-B-[XW₉O₃₃]⁹⁻ (X = Sb or Bi) units joined by two WO₆ octahedra and two fac-{M(CO)₃}⁺ moieties. The charge transfer (CT) dynamics, investigated by femtosecond transient absorption (TA) spectroscopy of Na₁₁H[Sb₂W₂₀O₇₀{Re(CO)₃}₂]·34H₂O and Na₁₁H[Bi₂W₂₀O₇₀{Re(CO)₃}₂]·33H₂O combined with the density functional theory (DFT) calculations, indicates that both complexes exhibit metal-to-POM charge transfer (MPCT) transitions from the Re centres to the POM ligands. The CT transition from the Mn centres to the POM ligands in K₉Na₃[Sb₂W₂₀O₇₀{Mn(CO)₃}₂]·32H₂O and K₉Na₃[Bi₂W₂₀O₇₀{Mn(CO)₃}₂]·32H₂O leads to decomposition of the starting compounds.

Fig. 8 Polyhedral/ball-and-stick representation of [X₂W₂₀O₇₀{M(CO)₃}₂]¹²⁻ (X = Sb/Bi; M = Mn/Re). Colour code: Re, pink; O, red; C, grey; W, jasper; MO₆, jasper.

In the following year, Hill et al. reported several “twisted-sandwich” PMCDs,⁷⁴ K₇Na₃P₂W₂₃O₈₀{Re(CO)₃}₂·38H₂O, (C₃H₁₀N)₈Na₂P₂W₂₃O₈₀{Re(CO)₃}₂·10H₂O, and (C₃H₁₀N)₆KNa₃P₂W₂₃O₈₀{Mn(CO)₃}₂·7H₂O, which were obtained by the reaction of [α-PW₁₁O₃₉]⁷⁻ with solvated [M(CO)₃]⁺ (M = Re, Mn) in acidic aqueous solutions (pH < 4). Upon decreasing the pH value, the solution colours become darker, indicating an ongoing dimerization process. During the self-assembly processes, the resulting di-coordinated intermediates are chiral building blocks, but they exist as a racemic enantiomeric pair in the unit cells. This study provides us with more convenient syntheses and a novel POM scaffold, named the “twisted-sandwich” (Fig. 9).

Fig. 9 (a) Scheme of proposed synthetic pathways of K₇Na₃P₂W₂₃O₈₀{Re(CO)₃}₂·38H₂O and (C₃H₁₀N)₆KNa₃P₂W₂₃O₈₀{Mn(CO)₃}₂·7H₂O. Color code: O, red; C, black; P and PO₄, violet; Re or Mn, yellow; W and WO₆, blue or green. The yellow rectangles represent the mirror planes in [α-PW₁₁O₃₉]⁷⁻. (b) Representative distances (unit: Å) of the four terminal oxygens in [α-PW₁₁O₃₉]⁷⁻. (c) The trigonal coordination sites (Å). (d) UV-vis absorption spectra of K₇Na₃P₂W₂₃O₈₀{Re(CO)₃}₂·38H₂O. The first spectrum was recorded after 0.5 h of reaction. Spectra thereafter were recorded upon titration.

In 2014, our group reported an undecatungstoarsenate-supported carbonyl rhenium derivative,⁷⁵ [(AsW₁₁O₃₉){Re(CO)₃}₃(μ₃-OH)(μ₂-OH)]⁶⁻, which consists of one monovacant Keggin structure unit and a typical trimeric carbonyl rhenium cluster. [(CH₃)₄N]₁₃[H₁₁(AsW₁₁O₃₉)₄{(Re(CO)₃)₃(μ₃-OH)(μ₂-OH)}₄]·23H₂O was synthesized by the reaction of Re(CO)₅Cl and [(CH₃)₄N]₈[HAsW₉O₃₄]·11H₂O in the CH₃CN-H₂O mixed solvent under mild conditions. The trivacant metastable polyanion [HAsW₉O₃₄]⁸⁻ could easily transform into the [AsW₁₁O₃₉]⁷⁻ configuration in a weakly acidic environment (pH > 5 nearly). To the best of our knowledge, this complex is the only compound with a multinuclear {M(CO)₃}⁺ unit incorporated into the Keggin-type POM (Fig. 10a). Additionally, this complex shows excellent catalytic activity for cycloaddition of epoxy chloropropane, and its potential applications are considered (Fig. 10b).

Fig. 10 (a) Polyhedral/ball-and-stick representation of the [(AsW₁₁O₃₉){Re(CO)₃}₃(μ₃-OH)(μ₂-OH)]⁶⁻ polyoxoanion. Colour code: O, red; C, black; Re, pink; WO₆, green. (b) Cyclic carbonate synthesis from epoxides and CO₂.

In the same year, a monovacant Keggin-type POM-supported trirhenium carbonyl derivative, [(CH₃)₄N]₅H₂₃[(PW₁₁O₃₉){Re(CO)₃}₃(μ₃-O)(μ₂-OH)]₄·24H₂O, was synthesized by our group.⁵⁹ Each lattice unit consisted of 4 [(PW₁₁O₃₉){Re(CO)₃}₃(μ₃-O)(μ₂-OH)]⁷⁻ units, and the crystal structure of the polyoxoanion revealed a “cap” model of the trirhenium carbonyl cluster that was combined with the [PW₁₁O₃₉]⁸⁻ fragment (Fig. 11a). The POM moiety can be considered an extremely active tetradentate ligand providing excellent coordination to the Re centres of the trirhenium carbonyl cluster. The ability of this compound to catalyse cyclic carbonate synthesis from CO₂ and epoxides with an ionic liquid (1-ethyl-1-methylpyrrolidinium bromide) as a cocatalyst was also studied. DFT calculations demonstrated that the Re₃–C₇ bond is highly activated and more prone to breakage, thus providing proof that the Re₃ centre becomes a strong Lewis acid mediator for activating the epoxide. As proposed, the epoxide is first activated by coordination to the Re₃ centre, followed by epoxide ring opening upon nucleophilic attack by Br⁻, and subsequent interaction of the nucleophilic alkoxide intermediate with the electrophilic CO₂ to form the cyclic carbonate. Both the Lewis acidic centre (Re₃) and the nucleophile (Br⁻) have the same importance in this mechanism, and the excellent synergistic effects greatly promote the reactions. The proposed mechanism is supported by both theoretical and experimental results and provides new insights into the design of more powerful catalyst systems for the cycloaddition reaction (Fig. 11b and c).

Fig. 11 (a) Polyhedral/ball-and-stick representation of the [(PW₁₁O₃₉){Re(CO)₃}₃(μ₃-O)(μ₂-OH)]⁷⁻ polyanion. Colour code: O, red; C, black; Re, purple; WO₆, green. (b) Proposed reaction mechanism for the cycloaddition of CO₂ to epoxides. (c) Several highest occupied (H = HOMO) and lowest unoccupied (L = LUMO) molecular orbitals of polyoxoanion, with the orbital energies given in eV.

In 2015, our group prepared a new trivacant POM-based manganese carbonyl derivative,⁷⁶ (NH₄)₃H₃[{Mn(CO)₃}(Mn(H₂O)₂)(Mn(H₂O)₃)(TeW₉O₃₃)]₂·31H₂O, which was isolated in the mixed solvent of acetonitrile and water at room temperature. Notably, the polyanion [{Mn(CO)₃}(Mn(H₂O)₂)(Mn(H₂O)₃)(TeW₉O₃₃)]⁶⁻ is composed of two [{Mn(CO)₃}(β-B-TeW₉O₃₃)]⁷⁻ building blocks linked via a [Mn₄(H₂O)₁₀]⁸⁺ cluster, resulting in a stable “hamburger” structure (Fig. 12). This compound was synthesized by using a conventional method, and the raw material MnCl₂ plays an important role in the synthetic process; the complex cannot be produced in the absence of MnCl₂. In particular, the ratio of CH₃CN to H₂O also affects the crystal growth, and the optimal ratio is 1 : 4 to 1 : 5 because the solubilities of metal carbonyls and POMs are very different; hence, adjusting the mixing ratio of the solvent to optimize dissolution of both complexes is necessary.

Fig. 12 Polyhedral/ball-and-stick representation of [{Mn(CO)₃}(Mn(H₂O)₂)(Mn(H₂O)₃)(TeW₉O₃₃)]⁶⁻. Colour code: O, red; C, black; Mn, light purple; Te, yellow; WO₆, green.

In the same year, our group reported a nonavacant Keggin-type rhenium tricarbonyl derivative, [(NH₄)₅]{[PMo₃O₁₆][Re(CO)₃]₄}·1.5H₂O, which showed good catalytic activity for the cycloaddition reaction with pyrrolidinium bromide as a cocatalyst.⁷⁷ The DFT calculations prove that the HOMOs are localized on the tricarbonyl rhenium fragment of [PMo₃O₁₆{Re(CO)₃}₄]⁵⁻; in contrast, the LUMOs localized entirely within the [PMo₃O₁₆]⁹⁻ moiety are principally W–O antibonding orbitals (Fig. 13a and b). In this case, the Re2 atom is proven to be the Lewis acid centre when it is in the excited state, and the halide ions in ionic liquids act as Lewis base centres that could substantially lower the activation energy barrier of the cycloaddition reaction.

Fig. 13 (a) Polyhedral/ball-and-stick representation of [PMo₃O₁₆{Re(CO)₃}₄]⁵⁻. Colour code: O, red; C, black; Re, pink; PO₄, orange; WO₆, green. (b) The cycloaddition of CO₂ with epoxides with pyrrolidinium bromide as a cocatalyst and several highest occupied (H = HOMO) and lowest unoccupied (L = LUMO) orbitals of 1a and their orbital energies in eV.

In 2016, our group reported four novel organic–inorganic hybrid Keggin-type PMCDs, [(M₄(H₂O)₁₀)(XW₉O₃₃)₂{Mn(CO)₃}₂]ⁿ⁻ (X = Sb/Bi; M = Mn/Mn_3.5W_0.5), which were prepared by the reaction between [XW₉O₃₃]⁹⁻ (X = Sb/Bi), Mn(CO)₅Br and Mn²⁺ or the reaction between [{Mn(H₂O)}₃(SbW₉O₃₃)₂]¹²⁻/[(Mn(H₂O)₃)₂(WO₂)₂(BiW₉O₃₃)₂]¹⁰⁻ and Mn(CO)₅Br.⁷⁸ Single-crystal X-ray diffraction crystallography indicated that these complexes are the first examples of structurally characterized transition metal-substituted sandwich-type tungsto-antimonates/bismutates incorporated with carbonyl manganese groups. The organic–inorganic hybrids are composed of two {Mn(CO)₃} groups attached to a dimeric heteropolytungstate {M₄(B-β-XW₉O₃₃)₂} unit via six Mn^I–O–W bonds. Furthermore, the {M₄(B-β-XW₉O₃₃)₂} unit consists of a central symmetric parallelogram-like tetra-Mn^II cluster sealed into two identical [B-β-XW₉O₃₃]⁹⁻ fragments through Mn^II–O–W bonds. The [(Mn₄(H₂O)₁₀)(XW₉O₃₃)₂{Mn(CO)₃}₂]⁸⁻ (X = Sb/Bi) proved to be efficient for the electrocatalytic reduction of NO₂⁻.

In 2019, our group prepared a monomeric tellurotungstate(IV)-supported rhenium carbonyl derivative, Na₂H₂[(CH₃)₄N]₆[Te₂W₂₀O₇₀{Re(CO)₃}₂]·20H₂O,⁷⁹ by the one-pot reaction of Re(CO)₅Cl, Na₂WO₄·2H₂O and Na₂TeO₃ (Fig. 14). Single-crystal X-ray diffraction analysis revealed that this polyoxoanion was similar to the [X₂W₂₀O₇₀{M(CO)₃}₂]¹²⁻ (X = Sb, Bi and M = Re, Mn) polyoxoanions reported by Hill et al. The crystal structure of this compound comprises two [TeW₁₀O₃₅{Re(CO)₃}]⁵⁻ fragments, and each rhenium carbonyl group fac-{Re(CO)₃}⁺ was stabilized by a [TeW₁₀O₃₅{Re(CO)₃}]⁵⁻ ligand in the “out-of-pocket” structural motif. In addition, this complex can be used as an excellent catalyst for the epoxidation of different alkenes using H₂O₂ as an oxidant in acetonitrile.

Fig. 14 The preparation process of the [Te₂W₂₀O₇₀{Re(CO)₃}₂]¹⁰⁻ polyoxoanion. Colour code: O, red; C, black; Re, purple; Te, yellow; WO₆, green.

3.3 Dawson-type PMCDs

The first pioneering studies based on Dawson-type PMCDs were reported in 1997.⁸⁰ Finke et al. reported two derivatives, [{Re(CO)₃}P₂W₁₅Nb₃O₆₂]⁸⁻ and [{Ir(CO)₃}P₂W₁₅Nb₃O₆₂]⁸⁻, but the structures of these complexes have not been verified by single-crystal X-ray diffraction analysis. The [{Ir(CO)₃}P₂W₁₅Nb₃O₆₂]⁸⁻ complex is stable in the absence of water but decomposes quickly in the presence of even 1 equiv. of water. These authors also attempted to synthesize the analogous [P₂W₁₅Nb₃O₆₂]⁹⁻-supported Rh(CO)₂⁺ complex; however, ³¹P NMR revealed that this compound was unstable in solution at room temperature. Importantly, in this study, the authors also provided the first direct evidence that the overall symmetry of organometallic cations supported on soluble basic oxides can be altered by the presence of cations such as Na⁺ cations.

There were no reports of Dawson-type PMCDs from 1997–2010, but in 2011, Hill et al. reported a new complex comprising a [Re(CO)₃]⁺ unit supported on the defect Wells–Dawson-type POM [α₂-P₂W₁₇O₆₁]¹⁰⁻, which exhibits good visible-light-induced photoredox activity.⁸¹ The resulting complex, [P₄W₃₅O₁₂₄{Re(CO)₃}₂]¹⁶⁻, was prepared by mixing equivalent amounts of K₁₀[α₂-P₂W₁₇O₆₁]·20H₂O and Re(CO)₃(CH₃CN)₃(BF₄) in an acidic aqueous solution. In this structure, each [Re(CO)₃]⁺ moiety is stabilized by an [α₂-P₂W₁₇O₆₁]¹⁰⁻ ligand in an “out-of-pocket” structural motif (Fig. 15a). Comprehensive computational and spectroscopic studies have shown that the intense visible absorption of this complex can be attributed to an MPCT transition involving CT from the Re(I) centre to the POM. In addition, the orbitals and transition in this case are distinct from those in the well-documented heterobimetallic systems because the acceptor orbitals are delocalized and multimetallic.

Fig. 15 (a) Polyhedral and ball-and-stick representation of the [P₄W₃₅O₁₂₄{Re(CO)₃}₂]¹⁶⁻ polyoxoanion. Colour code: Re, pink; O, red; C, black; WO₆, green. (b) UV-vis and (c) Fourier transform infrared (FTIR) spectra in CH₂Cl₂.

In 2012, we showed that the introduction of Dawson-type precursors can result in several compounds that consist of [Mn(CO)₃(CH₃CN)₃]⁺ cations and polyoxoanions.⁸² [Mn(CO)₃(CH₃CN)₃]₆[α-X₂M₁₈O₆₂]H₂O [X = P^V, M = Mo^VI, W^VI; X = As^V, M = Mo^VI, W^VI] complexes were isostructural and were obtained in CH₃CN/H₂O solvents by exploring different solvent systems, such as CH₃COCH₃/H₂O, C₄H₈O/H₂O, CH₃CH₂OH/H₂O, and CH₃OH/H₂O.

In 2013, a POM-supported trirhenium carbonyl cluster, [P₂W₁₇O₆₁{Re(CO)₃}₃{ORb(H₂O)}(μ₃-OH)]⁹⁻, was synthesized and characterized.⁸³ This complex was prepared by the reaction between a POM precursor [P₄W₃₅O₁₂₄{Re(CO)₃}₂]¹⁶⁻ and [Re(CO)₃]⁺ complexes in a slightly acidic (PH = 5–6) aqueous solution. The crystal structure of this complex reveals an “out-of-pocket” motif with a trirhenium carbonyl “cap” grafted on the defect site of [α₂-P₂W₁₇O₆₁]¹⁰⁻ (the “support”), and the three Re(I) centres reside in different coordination environments. Interestingly, this complex is the first example of an oxocentered multimetal electron-donor substructure, mimicking a metal oxide-supported interfacial dyadic structure.

A comprehensive investigation of the photophysical properties of this complex using computational and two different time-resolved spectroscopic methods clearly reveals the presence of MPCT. TA spectroscopy was used to characterize the CT dynamics occurring at the interfaces of this “double cluster”, and time-resolved infrared spectroscopy was applied to provide detailed information on the excited states of [Re(CO)₃]⁺ complexes (Fig. 16). Additionally, it was found that the lifetime of this charge-separated excited state is short, and thus capturing it to drive a chemical reaction is challenging.

Fig. 16 The scheme of intramolecular MPCT. Colour code: O, red; C, black; Re, pink; PO₄, orange; WO₆, blue.

In 2015, Sadakane et al. reported an α₂-isomer of a mono-Ru-substituted Dawson-type heteropolytungstate with a carbonyl (CO) ligand,⁸⁴ [α₂-P₂W₁₇O₆₁Ru^II(CO)]⁸⁻, which was obtained by the reaction of [α₂-P₂W₁₇O₆₁]¹⁰⁻ and Ru(acac)₃ at 170 °C for 4 days. Note that this is the first single-crystal structure analysis of mono-Ru(CO)-substituted heteropolytungstate (Fig. 17).

Fig. 17 Ball-and-stick representation of [α₂-P₂W₁₇O₆₁Ru(CO)]⁸⁻. Green, blue, pink, yellow, and red balls represent belt tungsten atoms, cap tungsten atoms, and ruthenium, phosphorus, carbon, and oxygen, respectively.

In 2017, our group obtained a POM-supported [Mn(CO)₃]⁺ complex containing the mixed-metal cubane [Na(H₂O)₅](NH₄)₇[P₂W₁₅O₅₆Co₃(H₂O)₃(OH)₃Mn(CO)₃]·19H₂O by the one-pot reaction of [α-P₂W₁₅O₅₆]¹²⁻ with Co(OAc)₂·4H₂O and Mn(CO)₅Br in a slightly acidic (pH = 6.0) solution.⁸⁵ Notably, this compound is the first example of a monomeric tricobalt(II)-substituted Dawson-type PMCD. This new polyanion can be described as grafting the metal carbonyl group {Mn(CO)₃} onto the tricobalt-substituted POM framework {P₂W₁₅Co₃O₆₂}, which is similar to that of the previously reported Dawson-type [P₂W₁₅Nb₃O₆₂]⁹⁻ polyoxoanion-supported Re(CO)₃⁺ complex, [P₂W₁₅Nb₃O₆₂Re(CO)₃]⁸⁻. The {P₂W₁₅Co₃O₆₂} subclass consists of a {Co₃O₆} cap embedded in the defect site of the [α-P₂W₁₅O₅₆]¹²⁻, forming a P₂W₁₈-like saturated Well–Dawson structure (Fig. 18a). As shown in Fig. 18b, this complex shows weaker ferromagnetic interactions at low temperature. Furthermore, it also shows high catalytic activity in the cycloaddition of CO₂ with epoxides under mild reaction conditions with pyrrolidinium bromide as a cocatalyst (Fig. 18c).

Fig. 18 Polyhedral and ball-and-stick representation of [P₂W₁₅O₅₆Co₃(H₂O)₃(OH)₃Mn(CO)₃]⁸⁻ Colour code: O, red; C, black; Mn, light purple; Co, light pink; P, yellow; WO₆, green. (b) Plots of χ_MT versus T for the compound. Inset: Field dependence of magnetization at different temperatures. (c) The cycloaddition of CO₂ to epoxides with pyrrolidinium bromide.

In 2018, our group prepared a monomeric trinickel(II)-substituted Dawson-type PMCD,⁸⁶ Na(NH₄)₅[P₂W₁₅O₅₆Ni₃(H₂O)₃(μ₃-OH)₃(Re(CO)₃)₃]·18H₂O, which was isolated in the one-pot reaction of [α-P₂W₁₅O₅₆]¹²⁻ with Ni(OAc)₂·4H₂O and [Re(CO)₅]⁺. The polyanion can be described as a hetero-metallic cluster {Ni₃O₆Re₃(CO)₉} stabilized by an [α-P₂W₁₅O₅₆]¹²⁻ ligand in the defect structural motif through six μ₃-O bridges from six WO₆ octahedra and one μ₃-O bridge from the central PO₄ tetrahedron (Fig. 19a). Notably, this polyanion is unprecedented and is the first example of a monomeric hexasubstituted Dawson structure decorated with low-valent transition metal Ni ions and metal carbonyl groups. A possible formation process of this polyanion is as follows: first, three scattered Ni(II) ions are captured by the [P₂W₁₅O₅₆]¹²⁻ polyanion to form a monomeric trisubstituted Dawson unit, β-Ni₃P₂W₁₅. As depicted in Fig. 19b, three [Re(CO)₃]⁺ fragments are attached to the β-Ni₃P₂W₁₅ unit to form a {Ni₃Re₃} cluster-substituted Dawson POM.

Fig. 19 (a) Polyhedral /ball-and-stick representation of [P₂W₁₅O₅₆Ni₃(H₂O)₃(μ₃-OH)₃{Re(CO)₃}₃]⁶⁻. (b) Representations of the {Ni₃Re₃} cluster. Colour code: O, red; C, black; Re, purple; P, yellow; NiO₆, orange; WO₆, green.

3.4 Nonclassical-type PMCDs

In parallel with most classical-type PMCDs possessing representative structural features, nonclassical-type compounds have also experienced rapid development in recent years. In 2000, Gouzerh et al. reported a series of POM-incorporated metal carbonyl complexes,⁸⁷ such as (nBu₄N)[Mo₂O₅(OMe)₅{M(CO)₃}₂] (M = Mn/Re) (Fig. 20a), [Mo₂O₄(OMe)₆{Mn(CO)₃}₂] (Fig. 20b), (nBu₄N)₂[Mo₂O₆(OMe)₄{Re(CO)₃}₂] (Fig. 20c), (nBu₄N)[Mo₂O₄{MeC(CH₂O)₃}₂{Mn(CO)₃}] (Fig. 20d), [Mo₂O₄{MeC(CH₂O)₃}₂{Mn(CO)₃}₂] (Fig. 20e), and (nBu₄N)₂[Mo₆O₁₆(OMe)₂{MeC(CH₂O)₃}₂{Mn(CO)₃}₂] (Fig. 20f), characterized by X-ray crystal diffraction analysis. These structures were obtained by the reaction of [MBr(CO)₅] or solvated {M(CO)₃}⁺ ions (M = Mn/Br) with (nBu₄N)₂[Mo₂O₇] in methanol, sometimes in the presence of triols of the type RC(CH₂OH)₃ (R = Me or CH₂OH). It should be noted that these clusters are the first examples of POMs incorporating {Mn(CO)₃}⁺ and {Re(CO)₃}⁺ units and constitute a family of organometallic complexes whose structures led to the recognition of topologically equivalent organometallic and oxo(alkoxo)metal units. Moreover, (nBu₄N)[Mo₂O₅(OMe)₅{M(CO)₃}₂] (M = Mn/Re) represent the first observation of discrete mono-cubane-type POM derivatives, as the frameworks of other complexes are based on tetranuclear units that display the common rhomb-like structure. Furthermore, these molecular structures have been determined by single-crystal diffraction analysis, and the structural relationships reveal some electronic connections between low-valent and high-valent complex fragments.

Fig. 20 Polyhedral and ball-and-stick representations of [Mo₂O₅(OMe)₅{M(CO)₃}₂]⁻ (M = Mn/Re) (a), [Mo₂O₄(OMe)₆{Mn(CO)₃}₂] (b), [Mo₂O₆(OMe)₄{Re(CO)₃}₂]²⁻ (c), [Mo₂O₄{MeC(CH₂O)₃}₂{Mn(CO)₃}]⁻ (d), [Mo₂O₄{MeC(CH₂O)₃}₂{Mn(CO)₃}₂] (e), and [Mo₆O₁₆(OMe)₂{MeC(CH₂O)₃}₂{Mn(CO)₃}₂]²⁻ (f). Colour code: O, red; C, black; Re/Mn, purple; MoO₆, green.

Then, our group made a breakthrough in the field of nonclassical-type PMCDs. In 2008, we prepared a polyoxoanion-incorporated {Mn(CO)₃⁺} complex,⁸⁸ (n-Bu₄N)₂[Mo₆O₁₆(OCH₃)₂{HOCH₂C(CH₂O)₃}₂{Mn(CO)₃}₂], by the reaction of (n-Bu₄N)₄[Mo₈O₂₆] with Mn(CO)₅Br in methanol in the presence of C(CH₂OH)₄. Furthermore, single-crystal structure analysis revealed that the polyoxoanion is located at a crystallographic inversion centre and consists of two tetranuclear [Mo₃O₈(OCH₃){HOCH₂C(OCH₂)₃}{Mn(CO)₃}]⁻ moieties (Fig. 21).

Fig. 21 Polyhedral and ball-and-stick representation of [Mo₆O₁₆(OCH₃)₂{HOCH₂C(CH₂O)₃}₂{Mn(CO)₃}₂]²⁻. Colour code: O, red; C, black; H, grey; Mn, purple; MnO₆, green.

In the same year, we also reported three novel octatungstate-supported tricarbonyl metal derivatives,⁸⁹ H₆[Na(H₂O)₅]₂{[H₂W₈O₃₀][Mn(CO)₃]₂}·13H₂O, H₂[Na(H₂O)₅]₂[Na(H₂O)₄]₂[Na(H₂O)₂]₂{[H₂W₈O₃₀][Re(CO)₃]₂}·13H₂O, and [Na(H₂O)₅]₂ [Na₂(μ₂-H₂O)₂(H₂O)₄]₂[Mn(H₂O)₂]{[H₂W₈O₃₀][Mn(CO)₃]₂}, which were prepared by the reaction between Na₂WO₄·2H₂O and M(CO)₅Br (M = Mn^I, Re^I). These structures not only are the first examples of isopolyoxotungstate-supported metal carbonyl compounds and enrich the structural diversity of POM-based organometallic compounds but also realize the merging of metal carbonyl chemistry and isopolyoxotungstate chemistry. X-ray diffraction analyses reveal that the first two structures are almost isomorphic and exhibit isolated topologies, whereas the third structure displays a 1D chain architecture. This grafting of a [Mn(CO)₃]⁺ pendant cap on a POM unit via three bridging oxygen atoms is similar to the reported compounds (Fig. 22).

Fig. 22 (a) Ball-and-stick representation of [H₂W₈O₃₀{Mn(CO)₃}₂]⁸⁻ with the selected labelling scheme. (b) Combined polyhedral/ball-and-stick representation of polyoxoanion with the selected labelling scheme. Cyan octahedra: {W₃MnO₄} cubanes.

Five years later, we successfully obtained the first two examples of octamolybdate-supported tricarbonyl metal derivatives,⁹⁰ (NH₄)₄[H₄{[H₂Mo₈O₃₀][Mn(CO)₃]₂}]·12H₂O, and (NH₄)₄[{H₆Mo₈O₃₀}{Re(CO)₃}₂]·14H₂O, which were prepared by the reaction of (NH₄)₆[Mo₇O₂₄]·4H₂O with Mn(CO)₅Br and Re(CO)₅Cl, respectively. Single-crystal X-ray diffraction analysis revealed that both polyanions are isostructural and contain a [H₂Mo₈O₃₀]¹⁰⁻ framework that is grafted by two tricarbonyl metal fragments. Notably, the [{H₂Mo₈O₃₀}{M(CO)₃}₂]⁸⁻ clusters have not been reported until now in the POM-based organometallic family. As a part of our continuous work, we also obtained another octatungstate-supported tricarbonyl metal polyanion, Na₈(H₂O)₁₈(CH₃COOH)₂[{H₂W₈O₃₀}{Mn(CO)₃}₂]·4H₂O, with four crystallographically independent Na⁺ ions, by reacting Na₂WO₄·2H₂O, ErCl₃·6H₂O, CH₃COOH and Mn(CO)₅Br. The preparations of these structures provide us with an effective and feasible method of designing novel isopolyoxomolybdate and isopolyoxotungstate metal carbonyl derivatives.

In 2018, our group reported two monomeric tellurium-containing heteropolymolybdate-supported metal carbonyl derivatives,⁹¹ (NH₄)₅H₃{[Te₂Mo₁₂(OH)O₄₄][Mn(CO)₃]}·18H₂O and (NH₄)₈{[Te₂Mo₁₂(OH)O₄₄][Re(CO)₃]}·13H₂O, which were obtained by the conventional mixed-solvent solution method. To the best of our knowledge, these two compounds are the first examples of metal carbonyl groups supported on a telluromolybdate framework. A detailed crystallographic analysis reveals that the two compounds are isostructural and that the polyoxoanion [Te₂Mo₁₂(OH)O₄₄]⁹⁻ represents a nonclassical anion with considerably different bonding modes (Fig. 23). The irregular [Te₂Mo₁₂(OH)O₄₄]⁹⁻ framework comprises one upper [TeMo₃(OH)O₁₂]³⁻ moiety and one lower [TeMo₇O₂₉]¹²⁻ moiety, which are aggregated together by two corner-sharing {MoO₆} octahedra, forming a like-Well–Dawson-type structure. Moreover, the complexes exhibit good electrocatalytic activity for the reduction of nitrite.

Fig. 23 (a) Formation of (NH₄)₅H₃{[Te₂Mo₁₂(OH)O₄₄][Mn(CO)₃]}·18H₂O and (NH₄)₈{[Te₂Mo₁₂(OH)O₄₄][Re(CO)₃]}·13H₂O. (b) Polyhedral and ball-and-stick representations of [Te₂Mo₁₂(OH)O₄₄]⁹⁻. Polyhedral representations of (c) P₂W₁₈; (d) Mo₇O₂₉ moiety in [Te₂Mo₁₂(OH)O₄₄][Mn(CO)₃]⁸⁻; (e) Mo₇O₂₄ unit in (NH₄)₆[Mo₇O₂₄]·4H₂O. Colour code: Te, yellow; Mo/MoO₆ octahedra, azure; Mn, blue; O, red; C, grey; WO₆, light blue; PO₄, pink.

Then, we prepared a novel isotetramolybdate-supported rhenium carbonyl derivative, [(CH₃)₄N]₄[{Re(CO)₃}₄(Mo₄O₁₆)]·H₂O, by the reaction of Na₂MoO₄·2H₂O with tetramethylammonium chloride and Re(CO)₅Cl in a mixed solution (CH₃CN : H₂O = 3 : 10) at 80 °C. Single-crystal X-ray diffraction analysis reveals that the polyoxoanion can be described as grafting the four {Re(CO)₃} units onto the {Mo₄O₄} cubane (Fig. 24).⁹² Moreover, the compound can be used as an efficient catalyst for the oxidation of thioanisole into the corresponding sulfoxide in the presence of hydrogen peroxide.

Fig. 24 (a) Ball-and-stick representation of [H₂W₈O₃₀{Mn(CO)₃}₂]⁸⁻ with the selected labelling scheme. (b) Combined polyhedral/ball-and-stick representation of polyoxoanion with the selected labelling scheme. Cyan octahedra: {W₃MnO₄} cubanes.

In the same year, we reported a novel tetracarbonyl metal selenotungstate derivative,⁹³ Na_1.5H_4.5[(CH₃)₄N]₂{[Mn(CO)₃]₄(Se₂W₁₁O₄₃)}·9H₂O, which was synthesized via one-pot synthesis by the reaction of Na₂WO₄, SeO₂, and Mn(CO)₅Br with a pH-controlled assembly (pH 4.8–5.6) in a mixed CH₃CN/H₂O solvent. In addition, the {Se₂W₁₁} fragment is composed of the novel {SeW₃} moiety and {SeW₈} unit via two μ₂-O atoms and is completely different from the reported {Se₂W₁₁} fragment composed of the unreported {SeW₄} and {SeW₇} species, which are linked in a Wells–Dawson-like assembly (Fig. 25). The successful preparation of this compound provides us with an enlightening synthetic method for covalent grafting of metal carbonyl groups onto POM clusters and further guides the design of high-nuclear PMCDs, though it is quite difficult to obtain these complexes due to the steric hindrance of the organic groups. Moreover, this complex showed good activity in catalysing cyclic carbonate synthesis from epoxides and CO₂ under mild reaction conditions in conjunction with 1-ethyl-1-methylpyrrolidinium bromide.

Fig. 25 Polyhedral and ball-and-stick representations of the {Se₂W₁₁} fragment and its basic building blocks. Colour code: O, red; C, black; Re, green; W, cyan; Se, rose; O, red.

4. Relative applications

4.1 Catalytic properties

As we know, POMs are a type of intriguing catalyst that can be applied for a wide range of applications owing to their reasonably high thermal stability, and reversible electron transfer ability under mild conditions. Moreover, metal carbonyl complexes can also exhibit interesting catalytic properties in many reactions. Consequently, relevant catalytic studies on PMCDs, possessing advantages of both POMs and metal carbonyl complexes, have been conducted. In 1987, Siedle et al. reported that [(Ph₃P)₂Rh(CO)(CH₃CN)]_8−nXM₁₂O₄₀ and [(Ph₃P)₂Rh(CO)]_8−nXM₁₂O₄₀ (X = P/Si; M = W/Mo; n = 3, 4) were bifunctional catalysts in hydrogenation, which are capable of activating carbon monoxide and hydrogen at the rhodium centers as well as oxygen at tungsten oxide surfaces.⁷¹ Recently, in 2014, our group reported a monovacant Keggin-type PMCD, [(AsW₁₁O₃₉){Re(CO)₃}₃(μ₃-OH)(μ₂-OH)]⁶⁻, and it can be used as a catalyst for the synthesis of cyclic carbonates from carbon dioxide and epoxides under mild reactions with co-catalyst pyrrolidinium bromide.⁷⁵ Using 0.1 mol% (relative to epoxide) of the catalyst, the yield achieved was 84.3% and the outstanding turnover frequency (TOF) of 843 h⁻¹ was superior to that of the tricarbonyl rhenium(I) complex reported by Wong et al. (≈122) h⁻¹.⁹⁴ The mechanism proposed is supported by both theoretical and experimental results which provides new insights into the design of more powerful catalyst systems for the cycloaddition reaction. Then, [(AsW₁₁O₃₉){Re(CO)₃}₃(μ₃-OH)(μ₂-OH)]⁶⁻, [{PMo₃O₁₆}{Re(CO)₃}₄]⁵⁻ and [{Mn(CO)₃}₄(Se₂W₁₁O₄3)]⁸⁻, reported by us in 2014, 2015, and 2018, respectively, could act as Lewis acid catalysts and promote the conversion of CO₂ to cyclic carbonate under mild reaction conditions.^75,77,93 In 2016, the use of [{Re(CO)₃}₄(μ₂-OH)(μ₃-O)(W₅O₁₈)]⁵⁻ in the epoxidation of olefin with aqueous H₂O₂ provides the corresponding epoxides with a 98.9% conversion and 99% selectivity.⁵⁶

4.2 Electrochemistry and electrocatalysis

As we all know, POMs can deliver up to 24 electrons to another species, making them powerful electron reservoirs for multielectron reduction processes. In 2008, Sadakane et al. addressed a new polyoxoanion [α-SiW₁₁O₃₉Ru^II(CO)]⁶⁻, the CV of which showed a well-defined reversible one-electron redox pair and two two-electron redox pairs.⁷² The controlled potential electrolysis confirmed that this well-defined reversible redox pair indeed corresponds to a single electron transfer. When the pH of the solution was raised, the Ru^III/II redox potential did not change, whereas the two-electron redox potential shifted in a more negative direction. The redox potential of Ru^III/II is independent of pH, indicating that no proton is involved in the redox reaction. In 2012, our group reported two novel trivacant Keggin-type PMCDs, and CV measurements of both were carried out in the pH 6.7 mixed solvent CH₃CN–Na₂SO₄ (0.4 mol L⁻¹).⁵⁷ The cathodic peak currents are proportional to the scan rate, demonstrating that the redox processes of both compounds are surface-controlled. The electrocatalytic properties of the two complexes for NO₂⁻ oxidation were investigated, and the catalytic currents increase dramatically with the oxidation peak of NO₂⁻ with stepwise addition of nitrite. In 2016, two novel Keggin-type PMCDs [{Mn₄(H₂O)₁₀(β-MW₉O₃₃)₂{Mn(CO)₃}₂}]⁸⁻ (M = Bi/Sb) proved to be efficient for the electrocatalytic reduction of NO₂⁻. With the addition of modest amounts of sodium nitrite, an irreversible oxidation peak appears in the positive range at E_pc = 0.85–0.90 V (this id expected for NO₂⁻), while the reduction peak current of the W^VI-based wave increase progressively.⁷⁸ In 2018, two monomeric non-classical tellurium-containing PMCDs, [{Te₂Mo₁₂(OH)O₄₄}{M(CO)₃}]⁸⁻ (M = Re/Mn), also exhibit well electrocatalytic activities toward the reduction of nitrite.⁹¹

5. Conclusion and outlook

In recent decades, a large number of PMCDs have been prepared, and there are strong motivations to study the excellent photoelectric and catalytic properties of these complexes. In this review, synthetic strategies, structural diversities and relevant applications of PMCDs are presented. Conventional solution synthesis has already been proved to be a powerful and efficient method for synthesizing PMCDs, while the hydrothermal synthesis approach is rarely used. Various structural PMCDs have been divided into four general families: Lindqvist-type, Keggin-type, and Dawson-type complexes and other nonclassical compounds. This article shows that a large variety of nonclassical-type PMCDs have been isolated by one-pot synthesis, but this probably represents only a small amount of what is possible, in view of the very large number of POM precursors available and the ability to covalently attach metal carbonyl groups to the POM. This approach can be extended to organic cations, and the solvent system can be extended to an aqueous/organic solvent mixture, commonly water/CH₃CN. Moreover, the synthetic variables of greatest importance in synthesizing such clusters are the species and stoichiometric ratios of the raw materials, solution pH, ionic strength, type of carbonyl metals and reaction temperature, and all of these factors play an indispensable role in the formation and crystallization of the product phases. The self-assembly process makes the preparation of novel PMCDs a challenging task in synthetic chemistry because of the ambiguous reaction mechanism. Moreover, the poor solubility of metal carbonyl and the harsh reaction conditions make the preparation of new PMCDs a challenging task. To obtain new PMCDs with desired features and properties, the use of precursors or raw materials is another priority in the preparation process. It should be noted that the preparation of PMCDs is almost in the acid aqueous solution, and the dark environment during the preparation process is also necessary. In particular, among the organometallic POMs, the [M(CO)_n]⁺ in PMCDs can act as face-directed cations to mediate the crystallization process, resulting in novel PMCDs with intriguing properties. And PMCDs can also be catalytically active and exhibit interesting visible-light-induced photo-redox activity. Thus, the development of PMCDs is a growing area of technical interest. Of course, there are challenges in the scientific community related to the further exploration of novel PMCDs and the improvement of many techniques. Therefore, we hope that a careful reading of this review will help to explore more intriguing PMCDs.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Grants 21571050, 21573056, 21771053, and 21771054).

Notes and references

1. A. Müller, F. Peters, M. T. Pope and D. Gatteschi, Chem. Rev., 1998, 98, 239.
2. A. Blazevic and A. Rompel, Coord. Chem. Rev., 2016, 307, 42.
3. M. T. Pope and A. Muller, Angew. Chem., Int. Ed. Engl., 1991, 30, 34.
4. J.-D. Compain, P. Mialane, A. Dolbecq, I. M. Mbomekalle, J. Marrot, F. Sécheresse, E. Rivière, G. Rogez and W. Wernsdorfer, Angew. Chem., Int. Ed., 2009, 48, 3077.
5. P. Ma, F. Hu, J. Wang and J. Niu, Coord. Chem. Rev., 2019, 378, 281.
6. D. Li, P. Ma, J. Niu and J. Wang, Coord. Chem. Rev., 2019, 392, 49.
7. J. Lu, Y. Wang, X. Ma, Y. Niu, V. Singh, P. Ma, C. Zhang, J. Niu and J. Wang, Dalton Trans., 2018, 47, 8070.
8. S. Omwoma, W. Chen, R. Tsunashima and Y.-F. Song, Coord. Chem. Rev., 2014, 258, 58.
9. S.-S. Wang and G.-Y. Yang, Chem. Rev., 2015, 115, 4893.
10. D.-Y. Du, L.-K. Yan, Z.-M. Su, S.-L. Li, Y.-Q. Lan and E.-B. Wang, Coord. Chem. Rev., 2013, 257, 702.
11. H. Lv, Y. V. Geletii, C. Zhao, J. W. Vickers, G. Zhu, Z. Luo, J. Song, T. Lian, D. G. Musaev and C. L. Hill, Chem. Soc. Rev., 2012, 41, 7572.
12. A. Proust, B. Matt, R. Villanneau, G. Guillemot, P. Gouzerh and G. Izzet, Chem. Soc. Rev., 2012, 41, 7605.
13. L. Cronin and A. Müller, Chem. Soc. Rev., 2012, 41, 7333.
14. H. N. Miras, J. Yan, D.-L. Long and L. Cronin, Chem. Soc. Rev., 2012, 41, 7403.
15. K. Kamata, K. Yonehara, Y. Nakagawa, K. Uehara and N. Mizuno, Nat. Chem., 2010, 2, 478.
16. N. Mizuno, K. Yamaguchi and K. Kamata, Coord. Chem. Rev., 2005, 249, 1944.
17. Z. M. Zhang, T. Zhang, C. Wang, Z. K. Lin, L. S. Long and W. B. Lin, J. Am. Chem. Soc., 2015, 137, 3197.
18. J. J. Berzelius, Ann. Phys., 1826, 82, 369.
19. J. F. Keggin, Proc. R. Soc. London, Ser. A, 1934, 144, 75.
20. A. Rosenheim and R. Bilecki, Chem. Ber., 1913, 4, 543.
21. A. Bijelic, M. Aureliano and A. Rompel, Chem. Commun., 2018, 54, 1153.
22. J. M. Clemente-Juan, E. Coronado and A. Gaita-Ariño, Chem. Soc. Rev., 2012, 41, 7464.
23. M. Lechner, R. Güttel and C. Streb, Dalton Trans., 2016, 45, 16716.
24. J. M. Cameron, D. J. Wales and G. N. Newton, Dalton Trans., 2018, 47, 5120.
25. N. Narkhede, S. Singh and A. Patel, Green Chem., 2015, 17, 89.
26. J. Zhang, F. Xiao, J. Hao and Y. Wei, Dalton Trans., 2012, 41, 3599.
27. F. Pu, E. Wang, H. Jiang and J. Ren, Mol. BioSyst., 2013, 9, 113.
28. L.-H. Bi, U. Kortz, S. Nellutla, A. C. Stowe, J. Tol, N. S. Dalal, B. Keita and L. Nadjo, Inorg. Chem., 2005, 44, 896.
29. W.-C. Chen, C. Qin, X.-L. Wang, Y.-G. Li, H.-Y. Zang, Y.-Q. Jiao, P. Huang, K.-Z. Shao, Z.-M. Su and E.-B. Wang, Chem. Commun., 2014, 50, 13265.
30. B. Botar, Y. V. Geletii, P. Kögerler, D. G. Musaev, K. Morokuma, I. A. Weinstock and C. L. Hill, J. Am. Chem. Soc., 2006, 128, 11268.
31. B. S. Bassil, M. H. Dickman and U. Kortz, Inorg. Chem., 2006, 45, 2394.
32. C. Pichon, A. Dolbecq, P. Mialane, J. Marrot, E. Riviere and F. Sécheresse, Dalton Trans., 2008, 71.
33. P. I. Molina, H. N. Miras, D.-L. Long and L. Cronin, Dalton Trans., 2014, 43, 5190.
34. J. P. Wang, J. Du and J. Y. Niu, CrystEngComm, 2008, 10, 972.
35. J. P. Wang, P. T. Ma, Y. Shen and J. Y. Niu, Cryst. Growth Des., 2008, 8, 3130.
36. J. W. Zhao, D. Y. Shi, L. J. Chen, P. T. Ma, J. P. Wang and J. Y. Niu, CrystEngComm, 2011, 13, 3462.
37. J.-W. Zhao, J. Zhang, S.-T. Zheng and G.-Y. Yang, Inorg. Chem., 2007, 46, 10944.
38. E. Ahmed and M. Ruck, Angew. Chem., Int. Ed., 2012, 51, 308.
39. T. L. Jorris, M. Kozik and L. C. W. Baker, Inorg. Chem., 1990, 29, 4584.
40. M. T. Pope and A. Müller, Angew. Chem., Int. Ed. Engl., 1991, 30, 34.
41. K. Wassermann, M. H. Dickman and M. T. Pope, Angew. Chem., Int. Ed. Engl., 1997, 36, 1445.
42. V. W. Day, T. A. Eberspacher, W. G. Klemperer and C. W. Park, J. Am. Chem. Soc., 1993, 115, 8469.
43. Y. Lin, T. J. R. Weakley, B. Rapko and R. G. Finke, Inorg. Chem., 1993, 32, 5095.
44. A. Müller, S. Sarkar, S. Q. N. Shah, H. Bögge, M. Schmidtmann, S. Sarkar, P. Kögerler, B. Hauptfleisch, A. X. Trautwein and V. Shnemann, Angew. Chem., Int. Ed., 1999, 38, 3238.
45. A. Proust, R. Thouvenot and P. Gouzerh, Chem. Commun., 2008, 1837.
46. A. Dolbecq, E. Dumas, C. R. Mayer and P. Mialane, Chem. Rev., 2010, 110, 6009.
47. A. Müller, H. Reuter and S. Dillinger, Angew. Chem., Int. Ed. Engl., 1995, 34, 2328–2361.
48. J. Zubieta, Comments Inorg. Chem., 1994, 16, 153.
49. A. Dolbecq, E. Dumas, C. R. Mayer and P. Mialane, Chem. Rev., 2010, 110, 6009.
50. P. Gouzerh and A. Proust, Chem. Rev., 1998, 98, 77.
51. L. S. Ott and R. G. Finke, Coord. Chem. Rev., 2007, 251, 1075.
52. M. Tountas, Y. Topal, A. Verykios, A. Soultati, A. Kaltzoglou, T. A. Papadopoulos, F. Auras, K. Seintis, M. Fakis, L. C. Palilis, D. Tsikritzis, S. Kennou, A. Fakharuddin, L. Schmidt-Mende, S. Gardelis, M. Kus, P. Falaras, D. Davazoglou, P. Argitis and M. Vasilopoulou, J. Mater. Chem. C, 2018, 6, 1459.
53. Q. Li and Y. Wei, Chem. Commun., 2018, 54, 1375.
54. Y. Luo, M. Wächtler, K. Barthelmes, A. Winter, U. S. Schubert and B. Dietzek, Chem. Commun., 2018, 54, 2970.
55. J. Luo, K. Chen, P. C. Yin, T. Li, G. Wan, J. Zhang, S. T. Ye, X. M. Bi, Y. P. Yang, G. Wei and T. B. Liu, Angew. Chem., Int. Ed., 2018, 57, 4067.
56. J. Li, J. Guo, J. Jia, P. Ma, D. Zhang, J. Wang and J. Niu, Dalton Trans., 2016, 45, 6726.
57. J. Zhao, J. Wang, J. Zhao, P. Ma, J. Wang and J. Niu, Dalton Trans., 2012, 41, 5832.
58. C. Zhao, W. R. Cordoba, A. L. Kaledin, Y. Yang, Y. V. Geletii, T. Lian, D. G. Musaev and C. L. Hill, Inorg. Chem., 2013, 52, 13490.
59. Z. Huo, J. Zhao, Z. Bu, P. Ma, Q. Liu, J. Niu and J. Wang, ChemCatChem, 2014, 6, 3096.
60. C. J. Besecker and W. G. Klemperer, J. Am. Chem. Soc., 1980, 102, 7600.
61. A. V. Besserguenev, M. H. Dickman and M. T. Pope, Inorg. Chem., 2001, 40, 2582.
62. J. R. Harper and A. L. Rheingold, J. Am. Chem. Soc., 1990, 112, 4037.
63. F. Bottomley and J. Chen, Organometallics, 1992, 11, 1707.
64. C. J. Besecker, V. W. Day, W. G. Klemperer and M. R. Thompson, Inorg. Chem., 1985, 24, 44.
65. W. G. Klemperer and D. J. Main, Inorg. Chem., 1990, 29, 2355.
66. V. W. Day, M. F. Fredrich, M. R. Thompson, W. G. Klemperer, R. S. Liu and W. Shum, J. Am. Chem. Soc., 1981, 103, 3597.
67. W. G. Klemperer’ and B. Zhong, Inorg. Chem., 1993, 32, 5821.
68. A. V. Besserguenev, M. H. Dickman and M. T. Pope, Inorg. Chem., 2001, 40, 2582.
69. R. Villanneau, A. Proust, F. Robert and P. Gouzerh, Chem. – Eur. J., 2003, 9, 1982.
70. W. H. Knoth, J. Am. Chem. Soc., 1979, 101, 759.
71. A. R. Siedle, C. G. Markell, P. A. Lyon, K. O. Hodgson and A. L. Roe, Inorg. Chem., 1987, 26, 219.
72. M. Sadakane, Y. Iimuro, D. Tsukuma, B. S. Bassil, M. H. Dickman, U. Kortz, Y. Zhang, S. Ye and W. Ueda, Dalton Trans., 2008, 6692.
73. J. Zhao, J. Zhao, P. Ma, J. Wang, J. Niu and J. Wang, J. Mol. Struct., 2012, 1019, 61.
74. C. Zhao, E. N. Glass, J. M. Sumliner, J. Bacsa, D. T. Kim, W. Guo and C. L. Hill, Dalton Trans., 2014, 43, 4040.
75. Y. Zhang, D. Zhang, Z. Huo, P. Ma, J. Niu and J. Wang, RSC Adv., 2014, 4, 28848.
76. Y. Liu, Y. Zhang, P. Ma, Y. Dong, J. Niu and J. Wang, Inorg. Chem. Commun., 2015, 56, 45.
77. Z. Huo, J. Guo, J. Lu, Q. Xu, P. Ma, J. Zhao, D. Zhang, J. Niu and J. Wang, RSC Adv., 2015, 5, 69006.
78. J. Jia, Y. Zhang, P. Zhang, P. Ma, D. Zhang, J. Wang and J. Niu, RSC Adv., 2016, 6, 108335.
79. J. Lu, X. Ma, P. Wang, J. Feng, P. Ma, J. Niu and J. Wang, Dalton Trans., 2019, 48, 628.
80. T. Nagata, M. Pohl, H. Weiner and R. G. Finke, Inorg. Chem., 1997, 36, 1366.
81. C. Zhao, Z. Huang, W. Rodríguez-Córdoba, C. S. Kambara, K. P. O'Halloran, K. I. Hardcastle, D. G. Musaev, T. Lian and C. L. Hill, J. Am. Chem. Soc., 2011, 133, 20134.
82. J. Zhao, J. Zhao, P. Ma, J. Wang, J. Niu and J. Wang, J. Mol. Struct., 2012, 1019, 61.
83. C. Zhao, C. S. Kambara, Y. Yang, A. L. Kaledin, D. G. Musaev, T. Lian and C. L. Hill, Inorg. Chem., 2013, 52, 671.
84. K. Nishiki, H. Ota, S. Ogo, T. Sano and M. Sadakane, Eur. J. Inorg. Chem., 2015, 2714.
85. V. Singh, Y. Zhang, L. Yang, P. Ma, D. Zhang, C. Zhang, L. Yu, J. Niu and J. Wang, Molecules, 2017, 22, 1351.
86. J. Jia, P. Ma, P. Zhang, D. Zhang, C. Zhang, J. Niu and J. Wang, Dalton Trans., 2018, 47, 6288.
87. R. Villanneau, R. Delmont, A. Proust and P. Gouzerh, Chem. – Eur. J., 2000, 6(7), 1184–1192.
88. J. Wang, J. Li and J. Niu, Chem. Res. Chin. Univ., 2008, 24, 675.
89. J. Niu, L. Yang, J. Zhao, P. Ma and J. Wang, Dalton Trans., 2011, 40, 8298.
90. D. Zhang, J. Zhao, Y. Zhang, X. Hu, L. Li, P. Ma, J. Wang and J. Niu, Dalton Trans., 2013, 42, 2696.
91. P. Zhang, V. Singh, J. Jia, D. Zhang, P. Ma, J. Wang and J. Niu, Dalton Trans., 2018, 47, 9317.
92. J. Lu, X. Ma, V. Singh, Y. Zhang, P. Ma, C. Zhang, J. Niu and J. Wang, Dalton Trans., 2018, 47, 5279.
93. J. Lu, X. Ma, V. Singh, Y. Zhang, P. Wang, J. Feng, P. Ma, J. Niu and J. Wang, Inorg. Chem., 2018, 57, 14632.
94. W. L. Wong, K. C. Cheung, P. H. Chan, Z. Y. Zhou, K. H. Lee and K. Y. Wong, Chem. Commun., 2007, 2175.

---