Polyoxometalates (POMs) are a class of metal-oxygen cluster compounds composed of pre-transition metals, such as Mo^Ⅴ, Mo^Ⅵ, W^Ⅵ, V^Ⅴ, Nb^Ⅴ and Ta^Ⅴ, by shared oxygen atoms [1, 2]. Due to the ample of redox electron pairs, pure POMs are sensitive to stimulation [3, 4]. In recent decade, a great number of metal-substituted POMs [5], pure organic-ligands sheeted POMs [6], hybrid organic-inorganic POMs compounds [7], and POM-based metal-organic frameworks [8, 9] have been widely investigated to regulate the physical and chemical performances. Among them, the direct modification of organic ligands on POMs can not only improve the stability under specific conditions but also tune the electronic structures while the strategy is limited by a finite number of atoms that can be utilized to coordinate [6]. The introduction of metal ions for coordination to bridge POMs and organic ligands has been proved to be a feasible way for functionalized POMs [5, 10].

Thiacalixarene, a cyclic phenol tetramer bridged by sulfur atoms, is an ideal multidentate ligand for building nanoclusters [11-13]. Thiacalix[4]arene generally bond metals to form a shuttlecock-like polynuclear secondary building unit (PSBU), which can be served for modular construction for various nanoclusters [14-16]. Thiacalix[4]arene-based nanoclusters have received exceptional attention due to their magnetic [17], catalytic [18], electrochemical properties [19], and molecular recognition [20]. As a result of the π-rich cavity and strong inclusion properties, calixarene and their derivatives can be used as hosts to include guests through intermolecular interactions to form host-guest complexes [21, 22]. The modification POMs with calixarenes is beneficial for fabricating functional hybrid materials with combined merits of the two components [23, 24]. Very recently, we successfully combined a sulfur-modified Mo₈ with a Co₄-thiacalix[4] arene unit to produce a bimetallic Co₄Mo₈ electrocatalyst, which showed Co-Mo synergy effect and durability for electrocatalytic water oxidation [25]. However, the research on thiacalix[4]arene-based molybdenum containing nanoclusters is still in the infancy and it is also challenging for synthesizing such complexes [24-28].

As our continuous on thiacalix[4]arene supported functional nanoclusters, here we report two nickel-thiacalix[4]arene modified POM-based bimetallic coordination clusters: [Ni₄(TC4A)]₂[(Mo₅^ⅤMo₃^ⅥO₂₄)(PO₄)] (+Solvent) (Ni₈PMo₈) and [Ni₄(TC4A)]₂[(Mo₅^ⅤMo₃^ⅥO₂₄)(OH)(CO₃)] (+Solvent) (Ni₈Mo₈) (H₄TC4A = p-tert-butylthiacalix[4]arene). The specific details of the syntheses, characterizations, and structure analyses for two clusters and the electrochemical detection and electrocatalytic oxidation for glucose have been provided (Some detailed information of this work was supplied in Supporting information.).

Single–crystal X–ray diffraction demonstrated that both Ni₈PMo₈ and Ni₈Mo₈ crystallize in the monoclinic system with the space group P2₁/m (Table S1 in Supporting information). As shown in Fig. 1, Ni₈PMo₈ cluster can be regarded as a sandwich structure composed of two tail-to-tail shuttlecock-like Ni₄-TC4A secondary building units (SBUs) in-between one {Mo₈O₂₄}(Mo₈) unit. The Mo₈ unit in Ni₈PMo₈ can be viewed as a ring of eight Mo atoms from the decomposion and re-selfassembly of H₃PMo₁₂O₄₀ (PMo₁₂). The inner space of Ni₈PMo₈ located disordered {PO₄} units which is also from PMo₁₂ source (Fig. 1a). Since PMo₁₂ can be decomposed and utilized as Mo sources, we are curious what structure we will get by directly using H₂MoO₄. Under equivalent synthesis conditions, the replacement of PMo₁₂ with H₂MoO₄ resulted a Ni₈Mo₈ cluster with the similar framework to that of Ni₈PMo₈. However, in inner cavity of Ni₈Mo₈ coordinated with OH⁻ and CO₃²⁻ anions. The OH⁻ bonded to the bottom of one Ni₄-TC4A SBU while CO₃²⁻ is close to another SBU (Fig. 1b). CO₃²⁻ may be derived from CO₂ in air, since prolonged exposure of the starting mixture for Ni₈Mo₈ under ambient prior to solvothermal reaction can significantly increase crystal yield. All the Ni centers in two nanoclusters are six-coordinated with two μ₂-O oxygen atoms from phenol, two μ₃-O atoms sharing with Mo, one sulfur atom from TC4A and one oxygen from the inner units. The Mo coordination enviroment is complicated in both two nanoclusters. By splitting disordered modes for {PO₄} and CO₃²⁻ units based on symmetry operation, one can find the Mo centers for two nanoclusters is also six-coordinated in general. The five coordiates (four bridging μ₃-O and one terminal O) are in normal environment as that in popular POM [3] and the six one is from the disordered inner units. The {PO₄} units bridge eight Mo centers via two pairs (each for three) of disordered μ₃-O atoms while CO₃²⁻ anion connected eight Mo by two pairs (each for two) of disordered μ₂-O. The splited μ₃-O atoms from {PO₄} in Ni₈PMo₈ bond with two Mo and one Ni centers. The μ₄-O with normal location in {PO₄} and CO₃²⁻ units coordinate with four Ni in a Ni₄-TC4A SBU simultaneously, respectively. The average nearest Ni–Mo distances bridged by two μ₃-O on the surface of Mo₈ units is 3.025 Å and 3.096 Å for Ni₈PMo₈ and Ni₈Mo₈, respectively. The combination of the bond valance calculations, the bonding modes analysis, and the overall charge banlance consideration indicate the bivalence of Ni cations, mixture valence for Mo (five for V and three for Ⅵ) in two nanoclusters and pentavalence for P in Ni₈PMo₈ (Table S2 in Supporting information). The partially reduction of Mo from Ⅵ in feeding materials to V in the products might caused by thiacalixarene ligand and methanol solvent under solvothermal condition, which is also observed in literatures [10, 25, 29]. The disordered solvent molecules filled in the crystal interstices have been subtracted by SQUEEZE during structure refinement (Fig. S1 in Supporting information). The two clusters have also been characteried by PXRD, FT-IR and TGA. PXRD indicate the purity of the bulk samples (Fig. S2 in Supporting information) and TGA showed both frameworks of two clusters are stable up to ca. 180 ℃ in N₂ media (Fig. S3 in Supporting information). FT-IR spectra showed characteristic C−H vibration p-tert-butyl groups and phenolic groups of calixarene (Fig S4 in Supporting information).

In POM chemistry, some classical examples has shown that the inner units encapsulated by poly-molybdenum can cause significant difference catalytic performances [5, 29]. Both Ni₈PMo₈ and Ni₈Mo₈ clusters contain a Mo₈ ring, but the components wrapped in the center are completely different. The difference prompted us to study the electrochemical behavior of two metal clusters. Moreover, recent reports have also suggested the possibility of POM-based materials for glucose detection [30, 31]. Furthermore, we also found that metal-thiacalixarene cluster might be a good catalyst as the abundant M-S bonds that can facilitate electron transfer in catalytic reactions [25, 32].

Based on above considerations, the electrochemical performances and electrocatalytic activities for glucose oxidation (G-oxidation) of two clusters were investigated, respectively. Working electrodes were prepared by dissolving the two clusters in CH₂Cl₂ and then directly coating on carbon paper (CP), denoted as Ni₈PMo₈-CP and Ni₈Mo₈-CP. The cyclic voltammetry (CV) of two clusters in 0.1 mol/L KOH at different scan rates is depicted in Fig. 2. A pair of redox peaks located at 0.519 V (Ⅰ), 0.423 V (Ⅰ') for Ni₈PMo₈-CP and 0.532 V (Ⅱ), 0.436 V (II') for Ni₈Mo₈-CP (scan rate: 10 mV/s; potentials vs. HgO/Hg reference electrode, the same afterwards) were observed, respectively, which can be assigned to the redox process of Ni^Ⅱ/Ni^Ⅲ [33, 34]. It is noteworthy that although the two working electrodes presented linear surface-controlled plots and similar peak differences (< 5 mV), both the anodic peak (i_pa) and cathodic peak (i_pc) intensity of Ni₈Mo₈-CP is obviously higher than that of Ni₈PMo₈-CP under a same scan rate. The observations indicated the inner different components of the clusters, PO₄^3- for Ni₈PMo₈ and OH⁻ and CO₃²⁻ for Ni₈Mo₈, are responsible for the different electrochemical performances.

In the case of low glucose concentration, compared with that of without glucose, the i_pa in CVs linearly increased and moved in the direction of high potential for both electrodes (Fig. 3). The significantly elevated i_pa intensity and less variation of i_pc indicate the electrochemical G-oxidation by Ni^Ⅲ [33, 34]. G-concentration vs. current density (j) were well fitted (correlation coefficient: 0.99) as linear plots with slopes being 1.28 and 0.64 in the 0.1–2 µmol/L and 0.2–4 µmol/L ranges for Ni₈PMo₈-CP and Ni₈Mo₈-CP, respectively. The smaller slope and the wide detecting G-concentration ranges for Ni₈Mo₈-CP indicated the superior oxidation capacity than that of Ni₈PMo₈-CP. The sensitivity and limit of detection are 1.28 mA L µmol⁻¹ cm⁻² and 0.04 µmol/L, 0.64 mA L µmol⁻¹ cm⁻² and 0.03 µmol/L for Ni₈PMo₈-CP and Ni₈Mo₈-CP, respectively. The glucose detection performances two clusters are comparable or better than some Ni-based electrochemical sensor materials (Table S3 in Supporting information).

Linear sweep voltammetry (LSV, scan rate: 5 mV/s) also was applied to investigate the electrocatalytic activities for two electrodes for G-oxidation (Fig. S5 in Supporting information). After the glucose detection was completed, the current intensity in LSV curves gradually increased with the continuous addition of glucose. The total glucose consumption of Ni₈Mo₈-CP was ca. 8 µmol/L at 0.65 V before the current density remained unchanged after two successive addition operations, which is about 1.5 times than Ni₈PMo₈-CP (5.4 µmol/L). The G-concentration vs. Δj (comparing with in the absence of glucose) were fitted as linear dependent plots with slopes being 1.32 and 1.91 for Ni₈Mo₈-CP and Ni₈Mo₈-CP, respectively, which indicates the better capacity of the former electrode for G-oxidation.

Oxygen evolution reaction (OER) is the main competing reaction for G-oxidation [35]. The LSV profiles of the G-oxidation and OER for Ni₈PMo₈-CP and Ni₈Mo₈-CP electrodes in 0.1 mol/L KOH solution with 20 mmol/L glucose are presented in Fig. 4a. The addition of glucose significantly lowered the electrode initial overpotential from ca. 0.7 V (OER) to 0.3 V (G-oxidation) for electrocatalytic oxidation reaction. The Tafel slopes for G-oxidation estimated be 181.71 and 209.53 mV/dec (0.3–0.4 V) for Ni₈PMo₈-CP and Ni₈Mo₈-CP, respectively. There is almost no current density for OER at 0.65 V. Comparingly, Ni₈Mo₈-CP presented a current density of 6.89 mA/cm² at same potential for G-oxidation, which is higher than that of Ni₈PMo₈-CP (5.12 mA/cm²). The long-term G-oxidation were carried out at a constant potential of 0.65 V by the chronoamperometric measurements (Fig. 4b). The current density decreased dramatically for Ni₈Mo₈-CP and reached to a relative constant j volume in about 3 h. In the same electrocatalytic G-oxidation time, the current density of Ni₈PMo₈-CP decreases slowly compared with that of Ni₈Mo₈-CP. Additionally, the smaller charge-transfer resistance (R_ct) of the Ni₈Mo₈-CP electrode in contrast to Ni₈PMo₈-CP also revealed the favourable electron transport rate and catalytic kinetics (Fig. S6 in Supporting information) [35].

To figure out how the inner components influence the electrocatalytic performances for G-oxidation, the rinsed and re-polished (renewed) electrodes were further investigated by CVs. Both of the renewed electrodes showed that the anode reaction (Ni^Ⅱ→Ni^Ⅲ) current density increased and the anodic peaks moved to high potential comparing to the initial electrodes at a scan rate 50 mV/s (Fig. S7 in Supporting information). Obvious redox pairs with equal peak difference of 84 mV besides to Ni^Ⅱ-Ni^Ⅲ were observed for both renewed electrodes at a low scan rate of 1 mV/s, which can be assign to the Mo^Ⅴ→Mo^Ⅵ (Fig. S8 in Supporting information). The corresponding peak differences for Ni^Ⅱ/Ni^Ⅲ redox pairs is 104 mV and 120 mV for renewed Ni₈Mo₈-CP and Ni₈PMo₈-CP electrodes, respectively. It is reasonable that the presence of Mo^Ⅴ/Mo^Ⅵ significantly lowered the peak difference of the active Ni^Ⅱ→Ni^Ⅲ in Ni₈Mo₈-CP than that for Ni₈PMo₈-CP. The smaller peak difference indicated the faster and reversible redox processes for Ni₈Mo₈-CP than that for Ni₈PMo₈-CP, which might afford the former better electrocatalytic performances. Further LSV experiments on renewed electrodes indicated the recyclability for electrodes except for some changes that accordance with the CV investigations discussed above (Fig. S9 in Supporting information).

X-ray photoelectron spectrum (XPS) was used to probe the chemical environments of the as-synthesized crystals and renewed samples. Ni₈PMo₈ and Ni₈Mo₈ crystals showed almost identical chemical environments for Ni 2p and Mo 3d (Fig. S10 in Supporting information). The well fitted Ni_2p high-resolution spectra with peak positions at 855.59 and 873.23 eV indicated the presence of Ni^Ⅱ for both crystal clusters [34, 36]. The high-resolution of spectra Mo 3d for crystal samples divided into two pairs fitted peaks at 232.06 eV and 234.8 eV, 232.6 eV and 235.7 eV, which could be attributed to Mo^Ⅴ and Mo^Ⅵ, respectively [25]. The calculated Mo^Ⅵ and Mo^Ⅴ ratios are all close to 3:5, which is accordance to that found in the crystal structure refinement (Table S4 in Supporting information). Generally, one metal center response for active site to facilitate the electron transfer of electrolytes while another one can tune the intrinsic catalytic performances in a bimetallic electrocatalyst [25]. Both Ni^Ⅲ signal and increasement of Mo^Ⅵ ratio were observed for the renewed samples (Fig. 5). The Ni^Ⅲ/Ni^Ⅱ and Mo^Ⅵ/Mo^Ⅴ ratio estimated to be 0.33 and 0.70 for Ni₈PMo₈ and 0.61 and 0.90 for Ni₈Mo₈, respectively, which indicated the interacted synergistic effect between Ni and Mo. The presence of both larger Ni^Ⅲ/Ni^Ⅱ and Mo^Ⅵ/Mo^Ⅴ ratios for Ni₈Mo₈ indicated the superior oxidation capacity than that of Ni₈PMo₈. The above discussions indicate that the electrocatalytic glucose oxidation mainly originated from Ni^Ⅱ/Ni^Ⅲ redox active sites and tuned by Mo^Ⅴ/Mo^Ⅵ sites.

Additionally, the high-resolution spectra for C 1s, O 1s and S 2p kept almost unchanged before and after glucose oxidation (Fig. S11 in Supporting information). Furthermore, no Ni or Mo element signal was detected in the electrolytes after long-term oxidation by ICP-AES analysis. Moreover, the EDS mapping for the samples after oxidation experiments show the well overlapped elements in the same area (Figs. S12 and S13 in Supporting information). All these results proved the stability of the cluster frameworks after reaction.

In summary, we have successfully synthesized and characterized two bimetallic coordination clusters Ni₈PMo₈ and Ni₈Mo₈. The framework of the two clusters is built from a poly-molybdenum Mo₈ sandwiched by two Ni₄-thiacalxiarene units. The inner cavity of two clusters incorporated different components, PO₄³⁻ for Ni₈PMo₈ and OH⁻ and CO₃²⁻ for Ni₈Mo₈, which resulted from the different Mo feed sources. The same exterior structure while different interior "chips" of the two clusters afforded the prepared binder free and active carbon free electrodes with distinguishable electrochemical sensing and electrocatalytic performances for glucose. The synergistic effect of Ni-Mo is identified experimentally and the superior effect for Ni₈Mo₈ than that for Ni₈PMo₈ can be attributed to the weak coordination of OH⁻ and CO₃²⁻ in the former core, and the strong coordination of PO₄³⁻ in the latter. Our work combines polyoxometalate and calixarene merits to prepare two bimetallic clusters, providing an example for studying the structure-activity relationship.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

This work was supported by the National Natural Science Foundation of China (No. 91961110).

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2022.03.088.