Ce-Mediated Molecular Tailoring on an Unprecedented Gigantic Polyoxometalate {Mo132} into Half-Closed Product {Ce11Mo96} for Achieving High Proton Conduction

doi:10.21203/rs.3.rs-2614076/v1

Download PDF

Article

Ce-Mediated Molecular Tailoring on an Unprecedented Gigantic Polyoxometalate {Mo₁₃₂} into Half-Closed Product {Ce₁₁Mo₉₆} for Achieving High Proton Conduction

https://doi.org/10.21203/rs.3.rs-2614076/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 18 Aug, 2023

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

Precise synthesis of polyoxometalates (POMs) is important for the fundamental understanding of the relationship between the structure and function of each building motif. However, it is a great challenge to realize the atomic-level tailoring of specific sites in POMs without altering the major framework. Herein, we report the first case of Ce-mediated molecular tailoring on the unprecedented gigantic {Mo₁₃₂}, which has a closed structural motif involving a never seen {Mo₁₁₀} decamer. Such capped wheel {Mo₁₃₂} undergoes a rare quasi-isomerism with known {Mo₁₃₂} ball displaying different optical behaviors. Experiencing an ‘Inner-On-Outer’ binding process with the substituent of {Mo₂} reactive sites in {Mo₁₃₂}, the site-specific Ce ions drive the dissociation of {Mo₂^*} clipping sites and finally give rise to a predictable half-closed product {Ce₁₁Mo₉₆}. By virtue of the tailor-made open cavity, the {Ce₁₁Mo₉₆} achieves high proton conduction, nearly two orders of magnitude than that of {Mo₁₃₂}. This work offers a significant step toward the controllable assembly of POM clusters through a Ce-mediated molecular tailoring process for desirable properties.

Physical sciences/Chemistry/Inorganic chemistry/Solid-state chemistry

Physical sciences/Chemistry/Supramolecular chemistry/Crystal engineering

Polyoxometalates (POMs), a unique subset of molecular clusters, possess diverse structural types and potential applications from biomedicine to catalysis and materials science.^[1–3] Recently, targeted design and synthesis of POM clusters have aroused much concern to explore the basic principles of structure and bonding as well as to expand the functionalities of POM-based materials and devices.^[4–5] Advances have been conducted in terms of diverse nuclearities, sizes, and shapes, for instance, the general building block strategy with ‘bottom-up’ molecule design^[6] has enabled the assemblies of POM architectures, especially giant POM entities that bridge the gaps between molecular oxo anions and infinite oxides.^[7] However, the controllable preparation of giant POM clusters including precise atomic-level tailoring and structural functionality still remains a great challenge, which is highly desirable for engineering functional molecular metal oxides and fundamental understanding of how different motifs contribute to its overall properties.^{[7, 8]} The bottleneck mainly relies on the limited type of giant POM clusters that can be used as rational tailoring precursor: (1) it should have available reactive or clipping building block sites that are favorable for the site-specific tailoring to target POM assemblies without changing the whole structure; (2) wheel- or cage-shaped giant POMs may be ideal models as they could provide confined and sizable environments for molecular container (e.g. water vapor adsorption) and proton transport purposes.^{[7, 9, 10]} Therefore, the exploration of eligible giant parent POM clusters and tailor-made derivative for unusual functionality is a striking task.

Polyoxomolybdate ‘big wheels’ are a well-researched iso-POM family^[10] constructed from fundamental repeating {Mo₁₁} building blocks, namely mix-valenced {Mo₈}-linked by {Mo₂}- and {Mo₁}-type units, as presented in the archetypal tetradecameric {Mo₁₅₄}^[11] and hexadecameric {Mo₁₇₆}^[12]. In principle, other oligomeric assemblies (e.g. decamer, 10 sets of {Mo₁₁} units) along with inherent open cavity of Mo wheels were possible to be formed by grouping a number of common building blocks. Such Mo wheels could be further modified by Mo-based capped fragments to afford cage-like architectures, the merely reported example is {Mo₂₄₈} that undergoes a molecular growth process with {Mo₁₇₆} as the parent scaffold^[12]. Of particular important are the bimetallic {Mo₂} units in Mo wheels, it is relatively reactive and can be easily eliminated, coordinated by carboxylic acids, or replaced by electrophiles, to produce Mo wheel species with ‘defect’ sites^[13], amino acid-^[14] or lanthanide (Ln)-functionalization^[15], respectively. At this level, it is reasonable to speculate that the doping of Ln centers into Mo wheel is favorable to regulate the ring shapes and sizes if both the coordination modes and positions of Ln centers on the wheels could be manipulated in a controlled manner. Therefore, we considered the cage-like Mo wheels involving enough {Mo₂} sites to be suitable giant POM precursor model for exploring the possible Ln-mediated precise atomic-level tailoring process. What’s more, giant POM isomers or quasi-isomers, which remain little reports, are ideal models for investigation of structure − property relationships in metal oxides materials.^[16] Inspired by this, we report an unprecedented {Mo₁₃₂} (1) with a closed structural motif arising from the symmetrical growth of two {Mo₁₁} caps located on both sides of a highly expected but never seen {Mo₁₁}₁₀ = {Mo₁₁₀} decamer. The discovery of capped wheel 1 along with the classical {Mo₁₃₂} ball^[17] constitutes a unique pair of structural quasi-isomers in POMs that present different optical behaviors. Especially, a Ce-mediated molecular tailoring process was demonstrated through the substituent of edge-sharing {Mo₂} reactive sites by Ce centers featuring overall ‘Inner-On-Outer’ binding modes, which drives the resection of one {Mo₁₁} cap of 1 through the site-specific dissociation of {Mo₂^*} clipping sites and finally gives rise to a predictable half-closed {Ce₁₁Mo₉₆} (2) product (Scheme 1). The enhanced proton conductivity between both clusters arising from Ce-mediated tailoring process has been fully examined and discussed.

The decameric ‘pure’ molybdenum framework 1 was prepared by reducing an acidified aqueous mixture of Na₂MoO₄·2H₂O, Na₂S₂O₄ and CH₃COONa·3H₂O under hydrothermal condition, while the tailoring product 2 was obtained through the reaction of an aqueous solution of 1 with cerium chloride. Both compounds were characterized crystallographically and their formula assignments (as below) are fully verified through an array of techniques as confirmed in giant wheel- and ball-shaped polyoxomolybdates^[10], such as single crystal X-ray structure analysis, elemental analyses, redox titration, UV/Vis spectroscopy as well as bond valence sum (BVS) calculations (see support information for detailed discussion, Tables S1-S6 and Figures S1-S3):

Na₂₂H₃₀{Mo₁₃₂O₃₇₂(OH)₁₀(H₂O)₁₂(SO₄)₅(CH₃COO)₂₀}·122H₂O ≡ Na₂₂H₃₀{1a}·92H₂O (1)

Na₃H₆{Ce₁₁Mo₉₆O₂₈₆(H₂O)₁₀₁(SO₄)₈}·94H₂O ≡ Na₃H₆{2a}·94H₂O (2).

Structural Analysis of Mo₁₃₂

The single-crystal X-ray structural analysis (Table S7) displays that 1 crystalizes in tetragonal space group P4₂/ncm. It not only possesses a classical wheel topology with a D_5d symmetry, but also has two caps grafted to the both rims of the wheel, thus affords a closed {Mo₁₃₂} capped wheel (Fig. 1). In this sense, polyoxoanion 1a could be divided into two parts, one is the main {Mo₁₁₀} wheel (Fig. 1a) composed of {Mo₈}, {Mo₂}, and {Mo₁} building blocks, each occurring ten times, and the others remain two {Mo₁₁} caps (Fig. 1c). Remarkably, {Mo₁₁₀} wheel contains the fundamental {Mo₁₁} building blocks in Mo wheel family (Scheme S1) including theoretical ‘{Mo₁₃₂}’ model, archetypal {Mo₁₅₄} and {Mo₁₇₆} discovered by Müller et al.,^[10] which constructs the first pure decameric Mo wheel reported to date. It should be noted that {Mo₂} units between two neighboring {Mo₆} pentagons are rather different in the known Mo wheels: the edge-sharing {Mo^V₂} units in 1a have a shorter Mo···Mo distance (ca. 2.554 Å) compared to corner-sharing {Mo₂} units in other oligomers (Fig. 1g). Each Mo atom of the edge-sharing {Mo₂} makes use of both oxygen atoms from equatorial plane and axial positions for binding to relative outer rim defined by adjacent pentagons of the wheel framework (Fig. 1g). The edge-sharing {Mo₂} units are the common building blocks for Mo ball^[17] and highly reduced Mo POMs^[18–19]. Such major structural difference originated from {Mo₂} units gives rise to a tight {Mo₁₁₀} Mo wheel (ring diameter: ca. 26 Å, Figure S4a) than expected. For this reason, this decameric pure molybdenum framework reported herein is the smallest available Mo wheel of this type as proposed by Cronin et al.^[20] As a matter of principle, the hypothetical dodecameric ‘{Mo₁₁}₁₂’ member with a mixture of both edge-sharing and corner-sharing {Mo₂} units may also be found experimentally in the near future (Scheme S1). The Mo wheel family may serve as an ideal model to understand new phenomena in confinement situations in terms of their selectable internal nanospaces.

On both sides of the {Mo₁₁₀} wheel are capped by two pentagonal {(Mo)Mo₅}-containing {Mo₁₁} units (Fig. 1h). It is interesting to observe the in-plane growth of two {Mo₁₁} caps occurs along the direction of inner rim defined by {Mo^V₁} groups (denoted as {Mo₁^*}, Figs. 1e and S5, belong to the {Mo₈} units in {Mo₁₁₀} wheel) between the adjacent {Mo₂} units in {Mo₁₁₀} wheel. Both capping moieties are accordingly turned at an angle of 36° relative to one another (Fig. 1b), in line with the corresponding relative positions of {Mo₁^*} units. Such {Mo₁}-driven in-plane growth process in 1a is quite different from the corner-sharing {Mo₂}-mediated in-plane growth ({Mo₂₄₈})^[12] or longitudinal growth process ({Mo₁₈₀})^[13] (Figure S6). In this respect, Mo wheels are favored to trap specific molybdenum fragments/intermediates (e.g. {Mo₁₇}^[14], {Mo₃₀}^[13], {Mo₃₆}^[12]) that can’t be survived in separate form in ring rims or centers for closed cage-like or host-guest architectures, respectively. The main interception driving force, such as hydrogen bonding or coordination bonds, comes from the simple Mo-based species themselves and their organic derivatives on the inner rims of the wheels. In 1a, this covalently bonded Mo-based cap unit with the composition {(Mo^VO)₅(Mo^VI)Mo^VI₅O₂₁(H₂O)₆} is not unique since it has been used for essential groups in building Keplerate {Mo₁₃₂} ball.^[17] Thus, 1a represents the fundamental building blocks of both Keplerate Mo sphere and Mo wheels simultaneously. Furthermore, another kind of edge-sharing {Mo^V₂} units (denoted as {Mo₂^*}, Figs. 1e and S5) have been realized during the combination of wheel and corresponding caps. One Mo atom within cap and above mentioned {Mo₁^*} in wheel contribute to this {Mo₂^*} unit with its four O atoms in the equatorial planes for connectivity (Fig. 1e). Therefore, the slightly bent {Mo₂^*} facilitates an asymmetrical curved growth along the equatorial plane (Fig. 1e) to afford ‘oblate spheroid’ topological structure of 1a that achieves a new type of cavity with the maximum internal diameter ca. 1.0 nm (Fig. 1d). We find that the assembly modes of edge-sharing {Mo₂^*} units reported here point the way towards a fresh type of combination between two {Mo₆} pentagons compared to that in spherical Keplerate {Mo₁₃₂} cluster^[17], whose {Mo₂} mainly boosts the symmetrical growth as a function of linkers. It follows that two types of edge-sharing {Mo₂} units with unequal connectivity during the assembly lay foundation of the formation of 1a (Fig. 1f). In general, small ligands (e.g. sulfates, phosphate, or carboxylate-containing species) are prone to replace coordinated water molecules on the inner surface of {Mo₂} in Keplerate {Mo₁₃₂} capsule^[21] and Mo wheels^[22] for effective stabilization or functionalization. This behavior is also presented in cage-like 1a but with anisotropic ligand distributions (Fig. 1i). The sulfate internal ligands each with half site-occupancy disorder (5 {SO₄²⁻} in total) are found to be attached to 10 well-constructed {Mo₂^*} units (Figure S7a) while acetate external ligands (20 {CH₃COO⁻} in total) are grafted onto both sides of {Mo₁₁₀} wheel with the 20 positions consisting of two Mo adjacent atoms came from {Mo₂} units and {Mo₆} pentagons (Figure S7b). The current ligand self-sorting of rigid cage-like entity may give 1a to mimic biological behavior such as small molecular recognition or responsive sensing. After the successful substitution of ligated water, the bidentate carboxylate here not only exerts a forceful constraint on the Mo atoms for pushing them to be much closer to each other, but also offers the possibility of external groups modification that has not been shown in polyoxomolybdates, which is expected to consider 1a as a network synthon for preparing unprecedented POMs-based high-dimensional supramolecular frameworks.^[23]

Isomerism in 132-nuclearity POMs: Mo Brown vs Green

Exploring the molecular isomerism phenomenon is a significant task which allows for understanding possible assembly mechanism and overall properties^[24–25], but such molecular isomerism in POMs is still rare.^{[16, 26]} Herein, the development of the Mo-based building blocks assembly strategy brings the isomerism in 132-Mo-atom clusters in the case of their identical reduction degree (45%, Mo^V₆₀Mo^VI₇₂) and characteristic {Mo(Mo)₅} motifs with similar connection rules. 1a contains 12 {Mo₆} pentagons as well as 10 edge-shared {Mo^V₂} linked by other simple Mo species that give rise to a capped wheel 1a while the spherical {Mo₁₃₂} is built up of 12 {Mo₆} pentagons and 30 edge-shared {Mo^V₂} units, hence, they represent quasi-isomers. This molecular quasi-isomerism offers a model for knowing how different motifs within the 132-Mo-atom POMs contribute to their overall properties, for example, UV/Vis absorption spectra in solution and the solid state.

Figure 2 shows 1a displays an intense dark green color with an absorption peak in the long wavelength range (ca. 720 nm), which remains different from classical {Mo₁₅₄} wheel featuring blue color with the characteristic absorption peak located ca. 750 nm. In addition, the Keplerate-type {Mo₁₃₂} reveals an brown color and possesses an absorption peak in the short wavelength range (ca. 450 nm) in the UV/Vis spectra in solution.^[18] The reason for different UV/Vis absorption in solution is the delocalization of the reducing electrons. It is well-known that {Mo₁₅₄} wheel (Mo Blue) contains the electrons from reduction delocalized over the structure, in the central belt spanning the ring. As for the {Mo₁₃₂} ball (Mo Brown), the electrons are localized in reduced and isolated {Mo^V}₂ Mo–Mo bonds. When the reducing electrons are delocalized not only over the ring but also the isolated Mo–Mo bonds in {Mo^V₂} units, this contributes to the capped wheel 1a defined as a brand new ‘Mo Green’ family. Moreover, the solid UV/Vis absorption spectra of both {Mo₁₃₂} clusters also show different absorption band (Figure S8a). 1 exhibits a much wider absorption range including UV and visible. On this basis, the diffuse reflectance spectrum shows that the corresponding well-defined optical band-gap energies of 1 can be assessed at about 1.13 eV (Figure S8b). Meanwhile, the relevant conduction band of 1 was also tested by Mott–Schottky measurements at different frequencies, which features a smaller E_g (Figures S8c) and higher conduction band (Figure S9 and S10) compared to {Mo₁₃₂} ball. In total, this pair of ‘molybdenum framework isomer’ exhibits disparate optical behaviors, which may act as cluster models that bridge the gap between molecular isomerism and phase isomerism of molybdenum oxide for the purpose of better understanding both structures and photochemical performance of phase-dependent nanomaterials.^[27]

Ce-Mediated Tailoring Product {Ce₁₁Mo₉₆}

The stability of giant capped wheel 1 in solution is necessary to be confirmed as the precision during tailoring process can’t be separated from the whole integrity of this parent. Both UV/Vis and Raman spectra before and after tetrabutylammonium (TBA) salt exchange of 1 are conducted for this purpose. The dark green precipitation of TBA-1 was obtained by the mixture of a diluted aqueous solution containing 1 as well as an aqueous solution of TBABr (Figure S11). UV/Vis absorption spectrum of 1 in aqueous solution remains almost coincident for 24h (Figure S12). Moreover, the unanimous absorptions and profiles of UV/Vis absorption spectra of 1 in water and TBA-1 in acetonitrile solution imply the good stability of 1 (Figure S12). As for the Raman spectrum of 1, two characteristic peaks at 992 and 818 cm^− 1 could be attributed to the vibrations of terminal Mo = O bands and Mo − O−Mo bridges (Figure S13), respectively^{[20, 28]}. In addition, all scattering peaks had been maintained in the Raman spectrum of TBA-1, which shows that the structure of 1 was kept during the cation exchange process. The stability of 1 encourages us to employ this cage-like giant POM cluster featuring abundant edge-shared {Mo₂} units as a suitable precursor for exploring the rational molecular tailoring in the presence of cerium centers. It is possible to drive the tailoring process at the building-block level when Ce species were used as strong electrophiles to replace {Mo₂} units on the parent {Mo₁₁₀} wheel. Fortunately, we were able to get the half-closed Ce-containing product 2 after a careful optimization of the reaction conditions. Single-crystal X-ray structure analysis shows that 2a crystallizes in Pnma space group and displays poor aqueous solubility. All {Mo₂} units on the {Mo₁₁₀} wheel had been replaced by cerium ions that give rise to such half-closed motif with one {Mo₁₁} cap missing (Figs. 3a and 3b). Meanwhile, the substitution of eleven Ce centers greatly reduces the symmetry of 2a (C_2v point group) as compared to tailoring parent 1a (D_5d point group). The half-closed motif based on Ln-Mo POMs is rather rare since the only one remains the {Mo₁₃₀Ce₆} suffers from a molecular growth of the hypothetical ‘{Mo₁₂₀Ce₆}’.^[13] By contrast, the half-closed 2a here is derived from a well-defined 1a with a predictable Ce-mediated molecule tailoring process.

From the crystal structure, as shown in Fig. 3, this unprecedented Ce-mediated molecule tailoring arises from the Ce centers adopting the ‘Inner-On-Outer’ binding modes related to different coordination environments (Fig. 4). (1) Ce ‘On’ binding modes (Fig. 4a). In total there are five Ce centers (Table S8, Ce-O bands, 2.474(10)-2.673(14) Å) located just on the position of {Mo₂} reactive sites of the wheel. Each features the same 4-connected rod as presented in {Mo₂} units linked with neighboring {Mo₆} pentagons. From the respective distances between relevant {Mo₈} fragments corners across the {Mo₂} or Ce bonding sites for the tailorable 1a and half-closed 2a, we can find that the overall distances between equivalent O atoms pairs are reduced slightly (from 5.6 to 4.6 Å for the inner rim and from 3.2 to 2.7 Å for the belt polyhedra) since the monatomic {Ce(H₂O)₃} and {Ce(H₂O)₅} moieties are a litter smaller than the bimetallic {Mo₂} unit (Figs. 3c and 3e). It is therefore reasonable to prove that the five 4f centers all adopting Ce ‘On’ binding modes cause a certain shrinkage of the rim of the wheel and hence the missing of the {Mo₁₁} cap on one side. This key point has been well confirmed by the neutral Ln-substituted Mo ring with no caps as a result of the same Ln ‘On’ binding manner.^[18] (2) Ce ‘Inner’ (Fig. 4b), ‘Outer’ (Fig. 4d) as well as ‘Inner-Outer’ (Fig. 4c) binding modes (Table S8, Ce-O bands, 2.469(13)-2.68(2) Å). The other 4f substitution sites are responsible for why the {Mo₁₁} cap on the other side that are kept retained, which produces a asymmetric molecule tailoring process that has not been established in other POMs. {Ce(H₂O)₄} moiety with Ce ‘Inner’ binding mode and {Ce(H₂O)₇} moieties with Ce ‘Outer’ binding modes attach to the inner and outer surfaces of the half-closed oblate spheroid so as to replace {Mo₂} units, respectively. Furthermore, Ce ‘Inner-Outer’ binding modes are defined as a set of two Ce centers take the place of {Mo₂} units both inner and outer surfaces of major molybdenum framework. All outer {Ce(H₂O)₇} moiety act as 2-connected rods while all inner {Ce(H₂O)₄} moiety act as 3-connected rods (Figure S14). To the best of our knowledge, Ce ‘Inner’ and ‘Inner-Outer’ binding modes had not been reported in 4f-doped gigantic Mo wheels.^{[13, 15]} Specifically, the aforementioned distances between equivalent O atoms pairs are not changed that lead to the retainability of {Mo₁₁} cap in 2a. In term of the small ligands in 2a, 3 {µ₂-η¹:η¹:η⁰:η⁰-SO₄^2-} groups are decorated onto Ce centers with Ce ‘On’ binding modes for further stabilization of 4f ions, the remaining {SO₄^2-} groups (Table S9, S-O bands, 1.438(11)-1.613(18) Å) are still adhered to {Mo₂^*} units as presented in 1a while all grafted {CH₃COO^-} groups disappear on account of Ce substitution. In fact, the inside sulfates bring great influence during the assembly of Ce centers featuring Ce ‘Inner’ binding modes. One Ce ion seems to be ‘trapped’ into half-closed 2a by the linkage of two identical {SO₄^2-} groups (Fig. 3d) through two Ce-O-S bridges (Ce4-O244, 2.458(10) Å; S1-O244, 1.461(11) Å), and the other ‘trapped’ Ce ions disordered over two positions (half site-occupancy disorder in crystallography) possess the alternate linkage (Fig. 3f) guided by three {SO₄^2-} groups (the central one adopts a novel µ₄-η¹:η¹:η¹:η¹ bridging mode) through four Ce-O-S bridges (Ce3-O99, 2.450(15) Å; Ce3-O181, 2.546(14) Å; Ce3-O99, 2.450(15) Å; S4-O181, 1.448(13) Å; S5-O99, 1.520(15) Å). Based on this, both linkages contribute to the successful encapsulation of [{CeO₅(H₂O)₄}(SO₄)₂] and [{CeO₄(H₂O)₄}₂(SO₄)₃] guest fragments inside half-closed 2a (Figs. 3d and 3f).

All eleven Ce^III ions are nine-coordinated with the + 3 oxidation state testified by X-ray photoelectron spectroscopy (XPS) analysis (see support information for all Mo^V/VI, S^VI and Ce^III XPS in 1 and 2, Figure S15-S19) that could be described as distorted monocapped square antiprism.^[15] It can be seen that the coordination modes and positions of Ce ions had been effectively controlled, the expected tailoring product, namely half-closed 2a, was obtained. In this regard, it appears that 1a remains a rather suitable parent in directing the formation of tailoring product 2a owing to its two types of edge-sharing {Mo₂} units. The {Mo₂} units of {Mo₁₁₀} wheel are prone to be replaced by Ce centers with the Ce ‘On’ binding modes, which provides the driving force for the occurrence of molecule tailoring process, while the {Mo₂^*} units consisting of two Mo atoms from the wheel and cap could be separated play the part of clipping sites to access the site-specific resection of {Mo₁₁} cap (Scheme 1 and Fig. 1f). The molecule tailoring on 1a not only brings the losing of {Mo₁₁} cap but also prompts the missing of five Mo atoms of the wheel without altering the main molybdenum framework (Figure S20). Also, the BVS calculations^{[20, 29]} of Mo-based double cubanes and UV/Vis spectrum (characteristic band is centered around 730 nm) suggesting 20 reducing electrons are delocalized throughout the Mo wheel structure as expected.

Proton Conductivity

Recently, POMs have emerged as advanced solid proton conductors in light of their structural diversity with oxygen-rich surface and chemical stability, etc.^[30–32]1 and 2 possess abundant H₂O molecules, rich H⁺ counter cations, terminal oxygen atoms and high stability towards hydrous condition, at this level, we choose proton conductivity tests to evaluate the functionality of 1 and 2 related to Ce-mediated precise tailoring strategy. Both compacted pellets of crystalline powder samples were used for alternating current (AC) impedance measurements. Various parameters, including relative humidity (RH, 60% ~ 98%) and temperature (30°C ~ 80°C), had been fully explored and the resulting performance implies the molecular tailoring product 2 displays a rather better proton conductivity than parent 1 (Fig. 5). At 30°C and 60% RH, the conductivity value of 2 was measured to be 4.02 × 10^− 6 S cm^-1 (Fig. 5a). The conductivity has rapidly grown to 2.50 × 10^− 2 S cm^-1 upon increasing RH to 98% (Fig. 5b). Similar situation also occurs in 1 (Figs. 5b and S21) with conductivity ranging from 6.19 × 10^− 7 S cm^-1 (60% RH) to 2.46 × 10^− 4 S cm^-1 (98% RH). In general, such highly humidity dependent proton conductivity has often been discovered in POMs-based solid proton conductors as the water molecules play a major role during proton movement process.^[33–34] This is in line with the water vapor adsorption of 1 and 2 that increase with the rising humidity as demonstrated by the room-temperature H₂O vapor absorption and desorption isotherms (Fig. 5c). The difference in humidity dependent proton conduction in both compounds may be identified as half-closed motif 2 possesses an outstanding water uptake capacity (312.52 cm³ g^-1) compared to closed motif 1 (160.59 cm³ g^-1). As the temperature increase from 30 to 80°C, the proton conductivity increases with the increasing temperature. This may be attributed to possible promotion of forming hydronium ion based on H₂O and H⁺ at elevated temperatures and thus accelerate the proton transition.^[35] The conductivity of 1 under 98% RH gradually raises and reached a maximum of 1.02 × 10^− 3 S cm^-1 (Figs. 5e and S22). Interestingly, the bulk conductivity of 2 increase very slowly from 2.50 × 10^− 2 S cm^-1 to 9.10 × 10^− 2 S cm^-1 with the temperature elevated from 30 to 80°C at 98% RH (Figs. 5d and 5e), suggesting less temperature dependent proton conductivity, which is a rare example of metal-oxo clusters-based proton conductors that keep high and stable proton conductivity over a relatively wide temperature region.^[36–37] The maximum proton conductivity of 2 close to the order of 10^− 1 S cm^-1 features one of the representative POMs-based crystalline proton conductors^[38–41] (Table S10) and is comparable to those of well-known crystalline metal-organic (e.g. MOF-808-4SA-150^[42], BUT-8-(Cr)A^[43], MOF-808C^[44], and covalent − organic frameworks (e.g. H₃PO₄@NKCOF-1^[45], iCOFMs-SO₃H^[46]). The activation energy (E_a) at 98% RH for the proton transfers in 1 and 2 is calculated to be 0.26 eV and 0.35 eV, respectively (Fig. 5e), based on the Arrhenius equation, indicating the Grotthuss mechanism (typically E_a < 0.4 eV) by non-diffusive routes, that is a proton hops alone between a proton donor and an acceptor without a molten carrier.^[47] Fast proton conduction in the best-known Nafions under low-temperature and high RH could also be recognized as this mechanism.^[48] The electron conductivity (direct-current experiment) and ionic conductivity (deuterated water experiment, Figure S23) measurements confirmed that the conduction is based on the transport of protons.^{[49, 50]} Moreover, the conductivity value of 2 up to 6.92 × 10^− 2 S cm^-1 remains nearly two orders of magnitude than that of 1 under the same conditions (70°C, 98% RH). Therefore, we can reach the conclusion that the enhanced proton conductivity as well as lower activation energy of 2 had been realized by this powerful Ce-mediated precise tailoring strategy from precursor 1.

The chemical environments of inter- and intra-molecular proton conduction for both compounds had been presented in order to further gain the possible mechanism of enhanced proton conductivity at the molecular level.^{[38, 51]} Each {Mo₁₃₂} in 1 is connected to eight adjacent {Mo₁₃₂} by sixteen intermolecular atomic C···O distances around 3.0 Å (Figure S24) while each {Ce₁₁Mo₉₆} in 2 is connected to seven adjacent {Ce₁₁Mo₉₆} by twelve intermolecular atomic O···O distances from 2.8 to 3.1 Å (Figure S25). The obvious intermolecular interactions result both compounds in 3D continuous array, whose frameworks featuring condensed packing modes are full of crystalline water molecular forming the hydrogen-bonding networks that are beneficial for effective proton hopping and transfer. Besides, the overall energy barrier for proton transport is also determined by the intramolecular proton-exchange process in POMs^[51], this theory has been well-established here. In contrast to closed 1, the half-closed 2 remains better proton conductivity may be due to the following three tips in terms of its intramolecular microenvironment: (1) Open Cavity. It is the most critical factor related to the enhanced proton conductivity. The sealed cavity volume of 1 is ca. 905.265 Å³ (Figure S26) and this void has still been retained over 50% with an open motif (ca. 487.109 Å³, Figure S27) in 2 after the Ce-mediated molecular tailoring process. (2) Inside {SO₄^2-} Groups. The un-coordinated O atoms of sulfate groups {µ₂-η¹:η¹:η⁰:η⁰ (bidentate)/µ₃-η¹:η¹:η¹:η⁰ (tridentate)} with stronger Lewis acidity may be served as better proton hopping sites than the terminal coordinated O sites of the metal ions. (3) Inside {Ce(H₂O)_x} (x = 3 to 5) Moieties. The utilization of 9-coordinate Ce³⁺ could stabilize the half-closed motif and its abundant aqua ligands may also participate in building hydrogen-bonding network. Based on the analysis above, the open and hydrophilic cavity provided by {SO₄^2-} and {Ce(H₂O)_x} groups is large enough to trap guest and adsorbed water molecular within 2, which is confirmed by its higher water uptake capacity in the water vapor adsorption isotherm, realizing dense and extended hydrogen-bonding network (Figure S28, short O···O distances, 2.6–2.9 Å) for promoting proton transport. Thus, under high RH conditions, 2 possesses low E_a and high proton conductivity over a relatively wide temperature range. The proton conductivity stability and durability of both compounds had been fully explored. The cooling and heating conductivity curves remain almost unchanged (Figures S29 and S30), indicating their stability under harsh conditions. Furthermore, their proton conductivities were monitored for over a monitoring period of 15h (Fig. 5f), and little changes were observed. At this level, no obvious changes could be identified about sample morphology of 2 with high proton conductivity through SEM results (Figures S31). In addition, the stability assessments are further testified by experimental data from IR (Figures S32-S33), Raman (Figures S34), Mo^V/VI and Ce^III XPS (Figures S35-S37) of 1 and 2 before and after proton conduction tests. All data suggest that the structural integrity of 1 and 2 are still retained during the tests.

In summary, we have demonstrated an unprecedented Ce-mediated molecular tailoring strategy from parent 1 to specific product 2 in giant POM assemblies. Both compounds were fully characterized by exhaustive physicochemical techniques in solution and the solid. The tailorable parent features a capped wheel structural motif involving a never seen {Mo₁₁₀} decameric wheel, which also undergoes a quasi-isomerism in 132-nuclearity POMs, namely Mo Brown (spherical Mo₁₃₂) and fresh Mo Green (1, capped wheel Mo₁₃₂). To our knowledge, such molecular isomerism phenomenon has not been explored in POMs. 1 possesses high stability in solution and thus a Ce-mediated tailoring process was produced by the substituent of edge-shared {Mo₂} reactive sites by Ce centers featuring overall ‘Inner-On-Outer’ binding modes, which drives the resection of one {Mo₁₁} cap of 1 by the dissociation of {Mo₂^*} clipping sites and finally gives rise to a predictable half-closed product 2. The proton conductivity results showcase that 2 remains one of the well-known POMs-based crystalline proton conductors that keeps high and stable proton conductivity over a relatively wide temperature region. This work constitutes a major step toward the controllable assembly of POM clusters through a precise tailoring process to optimize its whole properties.

Materials and characterization. All reagents and solvents were purchased from commercial sources and used as received. IR spectra were recorded on an Alpha Centauri FTIR spectrophotometer on pressed KBr pellets in the range 400 ~ 4000 cm^− 1. Elemental analyses were analyzed on a Plasma-Spec-II ICP atomic emission spectrometer. Water contents were determined by TG analyses on a PerkinElmer TGA7 instrument in flowing N₂ with a heating rate of 10°C min^− 1. UV − Vis spectra were obtained by using a 752 PC UV/Vis spectrophotometer. XPS was performed on a ESCALAB 250 spectrometer. The vacuum inside the analysis chamber was maintained at 6.2×10^− 6 Pa during the analysis. The morphology of crystal was characterized with SEM (FESEM; XL30, FEG, FEI Company). Raman spectra were conducted by a Jobin Yvon confocal laser Raman spectroscopy. The Mott–Schottky spots were carried out at ambient environment using the electrochemical workstation (CHI 760E) in a standard three-electrode system at frequencies of 800, 1000 and 1200 Hz. The water adsorption − desorption isotherms at 298 K were measured using a Micrometrics ASAP-2020 Mautomatic specific surface area.

Alternating-Current (AC) impedance measurements were performed on AUTOLAB electrochemical workstation at a frequency range from 0.1 Hz to 10 MHz. The relative humidity and temperature values were set using a BPS-50CL constant temperature and humidity chamber. The proton conductivity was calculated using the following equation:

\(\sigma {\text{=}}\frac{L}{{SR}}\)

where σ is the proton conductivity (S·cm^− 1), L is the thickness of the sample (cm), S is the area of the sample (cm²), and R is the resistance (Ω). The activation energy (E_a) was calculated according to the following equation:

\(\sigma T{\text{=}}{\sigma _0}{{\text{e}}^{\frac{{ - {E_a}}}{{{k_B}T}}}}\)

where T is the temperature (K), σ₀ is the prefactor, and k_B is the Boltzmann constant.

Materials synthesis. Synthetic procedures for both compounds are given in the Supplementary Information.

Single-crystal X-ray diffraction. Single-crystal X-ray diffraction data for 1 and 2 were recorded on a Bruker Apex CCD II area-detector diffractometer with graphite-monochromated Mo_Kα radiation (λ = 0.71073 Å) at 173(2) K. Absorption corrections were applied using multi-scan technique and performed by using the SADABS program. The structures were solved by direct methods and refined on F² by full-matrix least squares methods by using the SHELXTL minimization on Olex 2.0 software package. Detailed formula determination is given in the Supplementary Information.

Data availability

The X-ray crystallographic data for structures reported in this article have been deposited at the Cambridge Crystallographic Data Centre (CCDC), under deposition number CCDC 2225462 (1) and 2225463 (2). Both data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif. All relevant data supporting the findings of this study are available from the corresponding authors on request.

Acknowledgement

This work was financially supported by the NSFC of China (no. 21801038), Natural Science Foundation of Jilin Province – Free Exploration General Project (YDZJ202201ZYTS331), and the Science and Technology Research Foundation of Jilin Educational Committee (JJKH20221158KJ), the Fundamental Research Funds for the Central Universities.

Author contributions

W.-C.C and X.-L.W. conceived and designed the experiment. X.-X.L. and C.-H.L. conducted synthesis and characterization. M.-J.H., C.-Y.S. and C.Q. performed the proton conductivity tests. W.-C.C, K.-Z.S. and X.-L.W. solved the crystal structures. W.-C.C., Y.Y., X.-L.W. and Z.-M.S. co-wrote the manuscript. All the authors contributed to the analysis of the results.

Additional information

The authors declare no competing financial interests. Supplementary information accompanies this paper at www.nature.com/naturechemistry. Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/. Correspondence and requests for materials should be addressed to W.C.C and X.L.W

Competing ﬁnancial interests

The authors declare no competing financial interests.

Cameron, J. M. Supramolecular assemblies of organo-functionalised hybrid polyoxometalates: from functional building blocks to hierarchical nanomaterials. Chem. Soc. Rev. 51, 293–328(2022).
Yang, P. & Kortz, U. Discovery and evolution of polyoxopalladates. Acc. Chem. Res. 51, 1599−1608(2018).
Gao, Y. et al. Polyoxometalates as chemically and structurally versatile components in self-assembled materials. Chem. Sci. 13, 2510–2527(2022).
Lockey, D. et al. Investigating the autocatalytically driven formation of Keggin-based polyoxometalate clusters. Matter 5, 302–313(2022).
Zhang, G. et al. Complexed semiconductor cores activate hexaniobate ligands as nucleophilic sites for solar-light reduction of CO₂ by water. Angew. Chem. Int. Ed. 61, e202213162(2022).
Li, C. et al. Self-Assembly of anionic polyoxometalate–organic architectures based on lacunary phosphomolybdates and pyridyl Ligands. J. Am. Chem. Soc. 141, 7687−7692(2019).
Virovets, A. V. Peresypkina, E. & Scheer, M. Chem. Rev. 121, 14485–14554(2021).
Gao, M. Y. et al. Precisely tailoring heterometallic polyoxotitanium clusters for the efficient and selective photocatalytic oxidation of hydrocarbons. Angew. Chem. Int. Ed. 61, e202215540(2022).
Colliard, I. et al. Snapshots of Ce₇₀ toroid assembly from solids and solution. J. Am. Chem. Soc. 143, 9612–9621(2021).
Cronin, L. & Müller, A. From serendipity to design of polyoxometalates at the nanoscale, aesthetic beauty and applications. Chem. Soc. Rev. 41, 7333–7334(2012).
Liu, T. et al. Self-assembly in aqueous solution of wheel-shaped Mo₁₅₄ oxide clusters into vesicles. Nature 426, 59–62(2003).
Müller, A. et al. Molecular growth from a Mo₁₇₆ to a Mo₂₄₈ cluster. Nature 397, 48–50(1999).
Xuan, W. et al. Exploring the molecular growth of two gigantic half-closed polyoxometalate clusters {Mo₁₈₀} and {Mo₁₃₀Ce₆}. Angew. Chem. Int. Ed. 56, 9727–9731(2017).
Xuan, W. et al. Ligand-directed template assembly for the construction of gigantic molybdenum blue wheels. Angew. Chem. Int. Ed. 58, 10867–10872(2019).
Xuan, W. et al. Stereoselective assembly of gigantic chiral molybdenum blue wheels using lanthanide ions and amino acids. J. Am. Chem. Soc. 141, 1242−1250(2019).
Mahnke, L. K. et al. Configurational isomerism in polyoxovanadates. Angew. Chem. Int. Ed. 57, 2972–2975(2018).
Haviv, E. et al. Guest transition metals in host inorganic nanocapsules: single sites, discrete electron transfer, and atomic scale structure. J. Am. Chem. Soc. 142, 14504–14512(2020).
Ribó E. G. et al. Engineering highly reduced molybdenum polyoxometalates via the incorporation of d and f block metal ions. Angew. Chem. Int. Ed. 61, e202201672(2022).
Li, X. X. et al. An unprecedented fully reduced {Mo^V₆₀} polyoxometalate: from an all-inorganic molecular light-absorber model to improved photoelectronic performance. Chem. Sci. 13, 4573–4580(2022).
Ribo, E. G. et al. Synthesis, assembly, and sizing of neutral, lanthanide substituted molybdenum blue wheels {Mo₉₀Ln₁₀}. J. Am. Chem. Soc. 142, 17508–17514(2020).
Pow, R. W. et al. Embedding alkenes within an icosahedral inorganic fullerene {(NH₄)₄₂[Mo₁₃₂O₃₇₂(L)₃₀(H₂O)₇₂]} for trapping volatile organics. Chem. Sci. 11, 2388–2393(2020).
She, S. et al. Peptide sequence mediated self-assembly of molybdenum blue nanowheel superstructures. Chem. Sci. 12, 2427–2432(2021).
Boyd, T. et al. POMzites: a family of zeolitic polyoxometalate frameworks from a minimal building block library. J. Am. Chem. Soc. 139, 5930–5938(2017).
Kang, X. & Zhu, M. Structural isomerism in atomically precise nanoclusters. Chem. Mater. 33, 39–62(2021).
Tian, S. et al. Structural isomerism in gold nanoparticles revealed by X-ray crystallography. Nat. Commun. 6, 8667–8672(2015).
Winter, R. S. et al. Nanoscale control of polyoxometalate assembly: a {Mn₈W₄} cluster within a {W₃₆Si₄Mn₁₀} cluster showing a new type of isomerism. Chem. Eur. J. 19, 2976–2981(2013).
Kuwahara, Y. et al. Mild Deoxygenation of sulfoxides over plasmonic molybdenum oxide hybrid with dramatic activity enhancement under visible light. J. Am. Chem. Soc. 140, 9203–9210(2018).
Schäffer, C. et al. Unprecedented and differently applicable pentagonal units in a dynamic library: a Keplerate of the type {(W)W₅}₁₂{Mo₂}₃₀. Angew. Chem. Int. Ed. 48, 149–153(2009).
Brown, I. D. & Altermatt D. Bond-valence parameters obtained from a systematic analysis of the inorganic crystal structure database. Acta Crystallogr. Sect. B: Struct. Sci. 41, 244–247(1985).
Lai, R. D. et al. Assemblies of increasingly large ln-containing polyoxoniobates and intermolecular aggregation–disaggregation interconversions. J. Am. Chem. Soc. 144, 19603–19610(2022).
Ogiwara, N. et al. Proton conduction in ionic crystals based on polyoxometalates. Coord. Chem. Rev. 462, 214524(2022).
Sahoo, R., Pal S. C., & Das, M. C. Solid-state proton conduction driven by coordinated water molecules in metal–organic frameworks and coordination polymers. ACS Energy Lett. 7, 4490–4500(2022).
Lin, J. et al. Self-assembly of giant Mo₂₄₀hollow opening dodecahedra. J. Am. Chem. Soc. 142, 13982−13988(2020).
Meng, X. et al. Proton-conducting crystalline porous materials. Chem. Soc. Rev. 46, 464−480(2017).
Chand, S. et al. Metal–organic frameworks and other crystalline materials for ultrahigh superprotonic conductivities of 10⁻² S cm⁻¹ or Higher. Chem. Eur. J. 25, 6259(2019).
Cheng, D. et al. Recent advances on polyoxometalate-based ion-conducting electrolytes for energy-related devices. Energy Environ. Mater. DOI: 10.1002/eem2.12341(2022)
Xie, W.-L. et al. Keeping superprotonic conductivity over a wide temperature region via sulfate hopping sites-decorated zirconium-oxo clusters. Small. 18, 2205444(2022).
Yang, P. et al. Lewis acid guests in a {P₈W₄₈} archetypal polyoxotungstate host: enhanced proton conductivity via metal-oxo cluster within cluster assemblies. Angew. Chem. Int. Ed. 57, 13046–13051(2018).
Li, S. et al. High proton-Conductivity in covalently linked polyoxometalate-organoboronic acid-polymers. Angew. Chem. Int. Ed. 60, 16953–16957(2021).
Zhu, M. et al. Macrocyclic polyoxometalates: selective polyanion binding and ultrahigh proton conduction. Angew. Chem. Int. Ed. 61, e202200666(2022).
Liu, J.-C. et al. Aggregation of giant cerium–bismuth tungstate clusters into a 3D porous framework with high proton conductivity. Angew. Chem. Int. Ed. 57, 8416–8420(2018).
Sharma, A. et al. Superprotonic conductivity of MOF-808 achieved by controlling the binding mode of grafted sulfamate. Angew. Chem. Int. Ed. 60, 14334–14338(2021).
Yang, F. et al. A flexible metal–organic framework with a high density of sulfonic acid sites for proton conduction. Nat. Energy. 2, 877–883(2017).
Basu, O. et al. Defect engineering in a metal–organic framework system to achieve super-protonic conductivity. Chem. Mater. 34, 6734–6743(2022).
Yang, Y. et al. Combined intrinsic and extrinsic proton conduction in robust covalent organic frameworks for hydrogen fuel cell applications. Angew. Chem. Int. Ed. 59, 3678–3684(2020).
Shi, B. et al. Short hydrogen-bond network confined on COF surfaces enables ultrahigh proton conductivity. Nat. Commun. 13, 6666(2022).
Kim, S. et al. Achieving superprotonic conduction in metal–organic frameworks through iterative design advances. J. Am. Chem. Soc. 140, 1077–1082(2018).
Sone, Y. Ekdunge, P. & Simonsson, D. Proton conductivity of nafion 117 as measured by a four‐electrode AC impedance method. J. Electrochem. Soc. 143, 1254−1259(1996).
Xu, H. Tao S. & Jiang, D. Proton conduction in crystalline and porous covalent organic frameworks. Nat. Mater. 15, 722–726(2016).
Liu B. et al. A bismuth oxide/graphene oxide nanocomposite membrane showing super proton conductivity and low methanol permeability. Chem. Sci. 10, 556–563(2019)
Zhou, Y. et al. Insight into proton transfer in phosphotungstic acid functionalized mesoporous silica-based proton exchange membrane fuel cells. J. Am. Chem. Soc. 136, 4954–4964(2014).

Scheme 1 is available in the Supplemental Files section.

There is NO Competing Interest.

1.cif
cif for 1
2.cif
cif for 2
checkcif.pdf
checkcif for 1 and 2
SupportingInformation.doc
Scheme1.tif
Scheme 1. Schematic illustration of Ce-mediated molecular tailoring on {Mo₁₃₂} polyoxomolybdate. Ce-mediated molecular tailoring from an unprecedented gigantic {Mo₁₃₂} capped wheel POM featuring structural quasi-isomerism with known {Mo₁₃₂} ball to a predictable half-closed {Ce₁₁Mo₉₆} product.

Download PDF

Journal Publication

published 18 Aug, 2023

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

Ce-Mediated Molecular Tailoring on an Unprecedented Gigantic Polyoxometalate {Mo₁₃₂} into Half-Closed Product {Ce₁₁Mo₉₆} for Achieving High Proton Conduction

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Discussion

Methods

Declarations

References

Scheme 1

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1