Tellurite‐Squarate Driven Assembly of a New Family of Nanoscale Clusters Based on (Mo2O2S2)2+

Abstract The preparation and characterization of a new family of four polyoxothiometalate (POTM) clusters are reported, with varying size and complexity, based upon the dimeric [Mo2O2S2(H2O)6]2+ cation with the general formula (NMe4)aKb[(Mo2O2S2)c(TeO4)d(C4O4)e(OH)f] where a,b,c,d,e,f={1,7,14,2,4,10}=1, {Mo28Te2}; {2,26,36,12,10,48}=2, {Mo72Te12}; {0,11,15,3,3,21}=3, {Mo30Te3}; {2,6,12,2,4,16}=4, {Mo24Te2}. The incorporation of tellurite anions allowed the fine tuning of the templating and bridging of the available building blocks, leading to new topologies of increased complexity. The structural diversity of this family of compounds ranges from the highly symmetrical cross‐shaped {Mo24Te2} to the stacked ring structure of {Mo72Te12}, which is the largest tellurium‐containing POTM cluster reported so far. Also a detailed experimental analysis revealed that the pH isolation window extends from acidic to basic values. ESI‐MS analyses not only confirmed the stability of this family in solution but also revealed the stability of the observed virtual building blocks.


Introduction
Polyoxometalates (POMs) are molecular metal-oxide clusters that attract the interest of research groups due to their nanoscale size, unique structures and wide range of chemicalp roperties. [1] The fact that almosta ny elementi nt he periodic table can be incorporated into aP OM-based framework [2] renders them as exceptional candidates with highly modulars tructures and functionality.O nthe other hand, polyoxothiometalates (POTMs) is an under-studied subset of POM chemistry which emergedb yt he incorporationo fc halcogens; [3] the combination of aw ide range of applicationsa vailable to metal chalcogenidess uch as electronics, [4] hydrogen evolution [5] and batteries, [6] with the structurald iversity of POMs renders POTMs au nique family of compounds, which offers the opportunity for further exploration and discoveries. The incorporation of chalcogen elements into POTMs, for example, oxygen bridges are replaced by sulfur,a lters their behaviour and chemistry which leads to structures and properties not observed in conventional POMs. [7] The most common precursor utilised for the construction of large clusters is the dinuclear cation [Mo 2 O 2 S 2 ] 2 + , [8] due to its reactivity and flexibility to co-ordinate to appropriate templates,a nd generate libraries of building blocks. [9] Earlier work by Cadot et al. demonstrated that the dimer co-ordinates easily to carboxylates and that organic moleculesw ith multiple carboxyl groups can be used to expand the ring-shaped molecules that the dimer usually forms. [10] Work carriedo ut in our group revealed that certaint emplates can increaset he nuclearity and complexity much further than had been previously observed, with squaric acid, C 4 O 4 H 2 ,o r selenite, SeO 3 2À ,b oth granting access to new generations of building block libraries by acting either as templates or linkers; leadingn ot only to the formation of new and unanticipated structures, but also hugely increasing the size and structural diversityo ft hese compounds. [11,12] In order to fully utilise the potential of these compounds, it is necessary to understand their formation mechanism. The assembly of POM-based clusters proceeds through the acid condensation of the parent mononuclear metal-oxide anion, which can be influencedb yanumber of variables such as concentration and pH. [13] This process has also been shown to occur in stages;r ather than the full molecules forming directly from the mononuclear startingm aterials, their molecular growth proceeds through the formationo fs maller building units. These species, some of which cannot be isolated, [14] manifest themselves as structuralm otifs that repeatedly occur in many different POMstructures and are referred to as the building blocks (BBs) that make up the structure. [15] This concept was very eloquently illustrated by Müller et al. when describing the structures of the partially reduced "MolybdenumB lue" and Keplerate structures. [16] Self-assembly processes enable the entire field of supramolecular chemistry,w here it has been well established that pre-organised building blocks can spontaneously assemble into much larger and more complexa rchitecturest hat straddle the line between discrete molecules and bulk materials. [17,18] As such, identifying and determiningt he properties of viable building blocks [19] is ak ey aspect of this kind of chemistry and forms acritical part of the work reported here.

Results and Discussion
All of the molecules that are reported herein are derived from ac ommon set of precursors,s uch as [Mo 2 O 2 S 2 ] 2 + dimer units, C 4 O 4 2À and Te O 3 2À anions ( Figure 1). Even though the tellurite and squarate anions can act theoretically as templates, as well as linkers, and generate buildingb lock libraries that have been observedb efore,t here are no structural correlationst ot he final products.T hus, the presence and identity of the tellurite anion is revealed to be ac ritical factor in the formation of the final structure. In this case the tellurite anions act as linkersb etween the squarate-templated building blocks.
The BB A is unique to the tellurite/squarate system ( Figure 1), which comprises four [Mo 2 O 2 S 2 ] 2 + dimer units connected through pairs of co-ordinated hydroxide( OH À )g roups, with MoÀOb ond lengths of around2 .1 ,a nd as ingle squarate ion in the centre, co-ordinated to seven of the eight present Mo atoms through all four Oa toms of the squarate, with the average MoÀOb ond distance falling at approximately 2.3 .T he complete ring is formed by at ellurite anion, with the MoÀOb ond length here being slightly shorter at 2.0 .I n contrast, building block B is one [Mo 2 O 2 S 2 ] 2 + dimer unit short-er,w hile the squarate anion is co-ordinated to Mo centres through all four Oa toms. BB C can be considered ad erivative of building block A, which contains two [Mo 2 O 2 S 2 ] 2 + dimer units. This leaves one of the four squarate oxygen atoms free and uncoordinated. An interesting variation in bond length has been observed between the three squarateOatoms that are co-ordinated to Mo;t he two outer atoms have notably longer interactions than the central one, 2.5 and 2.2 respectively. D is av ery rare building block that has only been observed in two previous instances to our knowledge. [12a, 20] It is ah exameric building block consistingo ft hree dimer units, arranged in at riangular formation, centred on as ingle m 3 -O atom (MoÀOb ond length approx.2 .2 ). In addition to three m-OH links, the dimer units are also connected through Mo-m 3 S bonds of around 2.6 (standard MoÀSb ond lengths in the dimer are typically, 2.3-2.4 ).
These building blocks have been observed to form an umber of different structures that have been successfully synthesised and characterised( Figure2). Cluster 1,{ Mo 28 Te 2 }, comprises two A units connected further to two B units through the tellurite units, oriented perpendicular to the A units to form ac ross-like structurew ith dimensions 15. 6 25.5 (Figure 3) that crystallises in the monoclinic system space group C2/c. Cluster 1 has a D 2h point group,i ndicating three 2-fold rotation axes passing through the centre of the molecule, one bisecting the B building blocks, one bisecting the A building blocks and the third through the open cavity. Each of these axes is also containedw ithin am irror plane.
Structure 2,{ Mo 72 Te 12 }, is the largesta nd most complex of this family.Itisconstructed by A and C building blocks;initially a C unit bridges two A BBs. Twoa dditional A units are connectedt hrough the tellurite anions, arrangedp arallel to the first two, creatingi ne ffect multiple layerso fB Bs. An explana-  tion for the size and complex architectureo f2 lies in how the cluster is assembled from the available building blocks.A st he buildingb locks combine, see Figure 4, the intermediates formed at each stage have an accessible side, resulting in very labile speciest hat must continue building in order to form as table topology.Anotable feature of this compound is the Te -bridge between the building blocks that effectively creates at rapped trimeric TeÀOc hain formation ( Figure 5). It consists of three Te -atoms connectedt hrough O-bridges, with Te (a) being connected to Te(b) through an O-bridge forminga n angle of 118.58 while the Te(a)ÀOa nd Te(b)ÀOb ond lengths were found to be 2.24 and 2.07 respectively.Inasimilarfashion, Te (b) is connected to Te (c) through an O-bridge forming an angle of 116.28,w ith Te (b)ÀOa nd Te (c)ÀOb ond lengths of 2.13 and 2.12 respectively.
The two half-structures are joined throught he tellurite units of the second set of A building blocks, connected through hydroxide linkers to the other half. In the final structure, the two halves are oriented at approximately a9 0 8 angle to each other, with the open ends of each half towards the centre of the molecule, to form al ayered-ring structure of dimensions 24.0 26.6 that crystallises in the monoclinic system C2/c space group (Figure 4). The point group of this molecule is S 4 ,m eaning it has fewer symmetry elements than the previous example;m ore specifically,i ti ncorporates a2 -fold rotationa xis and a4 -fold improper rotationa xis that bisect the C building block in both halves of the structure.
Compound 3,{ Mo 30 Te 3 }, is ap ropeller-shaped molecule with 3-fold symmetry,c entred on one equivalent of D units, with the "blades" of the propeller being three A buildingb locks coordinating through the tellurite group of A to each vertex of D'st riangular structure,w ith the length of the sides of the triangular topology being 20.5 ( Figure 6). 3 displays C 3 symmetry,m eaning that the only symmetry element presenti sa3fold rotation axis through the m 3 -O atom of the D building block since the space occupiedb yt he three A buildingb locks is not distributed equally on both sides of the central D unit, which would be ap re-requisite for reflection to be as ymmetry operation for this molecule.L ike the previoust wo examples, this compound also crystallises in the monoclinic system space group C2/c.
Compound 4,{Mo 24 Te 2 }, is ah ighly symmetrical and aesthetically pleasing cross-shaped molecule of 15.4 in diameter, with several similarities to 1.T he main difference between the two is that 1 is constructed by two different BBs in contrast to 4 which is composed of one;f our B-type buildingb locks are arranged in as imilarf ashion to that observed in 1.T his change in BBs resultsi naslightly smaller nuclearity in 4 than in 1 (24 and 28 Mo centres,respectively).
As seen in 2,t ellurite becomes ab uildingb lock in itso wn right in this molecule in order to link the other buildingb locks  . Schematic representation of the sequential build-upo f2 from the BBs, starting with two A connecting with a C unit to form the first layer,followed by two additional A units to form asecond layer,w here two of these formationsa ssemble orthogonally to form af our-layered structure. BB C is denoteding reen.   (Figure 7). This compound has the point group D 4h , making it the most highly symmetrical structure of the four, which is further reinforced by this compound crystallising in the tetragonal system space group I4cm. The molecular symmetry elements include a4 -fold rotationa xis passing through both Te -atoms,f our 2-fold rotationa xes (two bisecting the two opposing pairs of buildingb locks, the other two going between them) and the relevant mirror planesa ssociated with these axes.
Apoint of interest when comparing 1 and 4 is the difference in orientation of the building blocks, where two of the BBs have been rotatedb y9 0 8.I ti sp ossible that the size and shape of the cavity in the centre of both of these molecules is important in revealing why the building blocks are arranged in this manner. The central cavity in both of these moleculesi si dentical in both size and shape, with the distance between the two Te -atomsb eing 5.3 while the shortestd istance between carbon atoms on opposing squaratei ons is 6.6 in both cases. We speculate that it may be able to takeu ps maller ionic species;h oweverm ore work would be required to establish whether this is the case.
One of the most interesting aspects of these molecules is the process by which they form. For each structure, the reaction is carried out at room temperature and all reagents are added in the same order.T he same volumeo fw ater is added to each reaction as solvent and each reactioni sc arriedo ut for the same length of time. The factors that have been observed to have the greatest bearing on the outcomeo ft he reaction are the ratio of the starting materials and the pH, which is unsurprising considering that it is well established that the condensation reaction that causes POMs to form is triggered by changes in pH. During the course of this work, many reactions were performed over al arge pH range in an effort to investigate the whole parameter space.T hus, it is possible to map the areas that are mostf avourablef or each compound. Figure 8s hows the pH ranges where each compound has been observed, along with the average yield obtained. Compounds 1 and 2 occur across ag reater range of pH values than the others, with 1 being ubiquitous at lower pH values (1.0-5.5) and 2 appearing to be more common at higherp H values (4.5-7.7)t han 1.I ti sn oteworthy that 2 is the only compound in the set capableo ff orming at both acidic and basic pH values. The other compounds are more restricted by the pH, with 3 and 4 both forming with ap Hw indow of lesst han two units. 4 assembles between pH 2.2-4.1. Surprisingly,c ompound 3 can be formed exclusively at pH values > 7, appearing between pH 7.0 and 8.8. Am ore fundamentalu nderstanding of how this system works couldb eg ained from looking at the pH ranges that the individual BBs occur in,a ss hown in Figure 9. These pH regions resultedb yc ombining the pH ranges identifiedf or each of the compounds constructed by as pecific set of building blocks, for example, BB A appears in compounds 1, 2 and 3,s ot he pH window for A type buildingb locks is the total range in which these compounds are found. A type building blocks, which appear in three of the four reported compounds,h as, unsurprisingly,t he broadest range and it occurs across the entire range of pH values investigated. B type building blocks extend from the lowest investigated pH across almost the entire range of acidic pH values. C has an arrower range than B,a nd is shiftedh igheru pt he pH scale. Finally, D has the narrowest range of the four BBs and only occurs in one compounda tt he higher end of the pH range in question.   The IR spectra (Figures S1-4 in Supporting Information) of all of these compounds are very similars ince all are formed from the same set of buildingb locks with identical modes of bondinga nd interaction. In all four spectra,t here is ab road signal at 3400 cm À1 ,i ndicative of the stretching of the OÀH bonds in water,a rising from the solvent molecules in the crystal structure and atmosphere.A lso arising from the solvent are the signals that appear at 1600cm À1 ,w hich correspond to bending vibrations of the water molecules. Signals at 2350 cm À1 are assigned to atmospheric CO 2 .T he signals at 1520 cm À1 are assigned to the CÀObonds of the squarate template. Other signals of interest occur between 480 and5 00 (Mo-S-Mo bridges), 690 (Mo-O-Mo bridges) and9 50 cm À1 (Mo= Ob onds).
Thermogravimetric analysis( TGA, FiguresS5-8) experiments show that all four compounds lose around 10-15 %o ft heir mass by 150 8C. This mass drop is assigned to crystallographic water content. Three of the four compounds experience am ass drop between 200 and 400-450 8C, accountingf or 10-11 %o ft he total mass of the sample;i nc ompound 2,t his is manifestedo vert wo overlapping steps, accounting for around 14 %o ft he total mass of the sample. For all of thesec ompounds,t he weightl oss is associated with the removal of carbon and sulfur content in the form of CO 2 and SO 2 ,r espectively.
Mass spectrometry experiments [21] showedt hat not only are these compounds stable in solution, through our observation of the intact clusters, but also we were able to identify the virtual BBs during the fragmentation process ( Figure 10). The distributione nvelopes that have been tentatively assigned (Figures S9-12 and tables S1-4 in the Supporting Information) have almostu niversally been either one of the BBs or ac ollection of multiple BBs that form af ragment of the distinct clusters. The clarity with which the buildingb locks appear in the mass spectrometrye xperiments suggests as equential mechanism of assembly,with the BBs formingf irst and subsequently combining into the structures that crystallise.

Conclusion
In conclusion, we have investigated the interaction of [Mo 2 O 2 S 2 ] 2 + units with the tellurite anion for the first time and directedt heir assembly into an ew family of nanoscale structures. Te lluriteh as been shown to act as al inker species, rather than at ruet emplate, when the system also includes the squarate anion.F our new clusters have been successfully synthesised and characterised that represent the structural diversity in-herenttoPOM chemistry.Interesting structural features include the adoption of high levelso fs ymmetry,a ccessible cavities that could potentially be used to uptake small ions and degrees of complexity that could not have been anticipated. Carefull investigation of the systema llowed us to map the reaction co-ordinates and identify the formation areas of specific building blocks. While there are overlaping areas, it was possible to identify clear pH ranges that favour the formationo f each compound, and infer from this ap icture of conditions that are preferred by the building blocks that form the basis of each of these structures.

Experimental Section
All reagents were purchased from Sigma Aldrich, Fisher Scientific and Alfa Aesar,a nd were used as provided with no further purification required. The (Mo 2 O 2 S 2 ) 2 + dimeric unit was prepared according to the modified procedure published by Cadot et al. in 1998. [7a] X-ray crystallography:D ata were collected at 150(2) Ku sing aBruker AXS Apex II (l(Mo Ka ) = 0.71073 )equipped with agraphite monochromator.S tructures were solved and refined using Direct methods with SHELXS-2014 [22] and SHELXL-2014 [23] using WinGX routines. [24] Refinement was achieved by full-matrix least squares on F 2 through SHELXL. Corrections for incident and diffracted beam absorption effects were applied using analytical methods. [25] All data manipulation and presentation steps were performed using WinGX. Details of interest about the structure refinement are given in the tables in the Supporting Information. Fourier transform infra-red spectroscopy:S amples were prepared as KBr discs and FTIR spectra were collected in transmission mode using aS himadzu IR Affinity-1S Fourier Transform Infra-Red Spectrophotometer.W avenumbers (n)are given in cm À1 . Electrospray ionisation mass spectrometry (ESI-MS) was performed on aW aters Synapt-G2 HDMS spectrometer operating in ion mobility mode, equipped with aq uadrupole and time of flight (Q/ToF) module for MS analysis. All samples were prepared by dissolving in 1:10 H 2 O:MeCN (HPLC grade) to ac oncentration of ca. 1 10 À5 m and injected directly at af low rate of 5 mLmin À1 using aH arvard syringe pump. All spectra were collected in negative ion mode and analysed using the Waters MassLynx v4.1 software. For all measurements the following parameters were employed:c apillary voltage:2 .5 kV;s ample cone voltage:1 0.0 V; extraction cone voltage:4.0 V; source temperature:808C; desolvation temperature: 180 8C; cone gas flow: 15 Lh À1 (N 2 ); desolvation gas flow: 750 Lh À1 (N 2 ).
Thermogravimetric analysis:A nalysis was performed on aT AI nstruments Q500 Thermogravimetric Analyser under nitrogen flow with at ypical heating rate of 10 8Cmin À1 from room temperature up to 800 8C. Elemental analysis:M o, Sa nd Te content were determined by ICP-OES analysis in the following way:5 -10 mg sample material was digested by adding 1mLd eionised water and 2mLc onc. HNO 3 to the sample in ad igestion beaker.T he sample solution was warmed until clear before being allowed to cool and af urther 5mLd eionised water added. The resulting solution was transferred quantitatively with washings to an Ac lass 50 mL volumetric flask and made up to the mark with deionised water.Ab lank sample was also prepared simultaneously to account for any digestion interferences. The samples were transferred to 50 mL polypropylene centrifuge tubes and analysed on an Agilent SVDV 5100 ICP-OES using the SVDV mode and appropriate calibration standards. Carbon and nitrogen content was analysed by the University of Glasgow microanalysis service within the School of Chemistry.P otassium content was determined using aC orning 410 Flame Photometer using the same samples and calibration standards used in the ICP-OES analysis.
All TGA and elemental analysis experiments were run on dry samples and as such, some crystallographic water had already been lost, accounting for the discrepancy between the crystallographic water content as stated in the molecular formulae and the water content found in the TGA experiments. ] 2 + (6.5 mL, 0.878 mmol) was added to give ab lack colour to the solution. 1 m K 2 CO 3( aq) was used to bring the pH of the solution to 5.2 with the colour turning to clear orange. The reaction mixture was stirred at room temperature for one hour,d uring which time the pH rose to 7.6. The reaction mixture was then filtered and kept at 18 8C. Within 2w eeks red rod-shaped crystals formed that were suitable for crystallography studies. 230 mg material was collected (58.38 %y ield based on Mo). Elemental analysis calcd (%) for C 12  (NMe 4 ) 2 K 6 [(Mo 2 O 2 S 2 ) 12 (TeO 4 ) 2 (C 4 O 4 ) 4 (OH) 16 ]·55 H 2 O( 4):N a 2 Te O 3 (0.05 g, 0.225 mmol) and C 4 O 4 H 2 (0.1 g, 0.877 mmol) were dissolved together in 20 mL distilled water to form ac loudy white solution. Dimeric [Mo 2 O 2 S 2 ] 2 + (5 mL, 0.68 mmol) was added to give ab lack colour to the solution. 1 m K 2 CO 3( aq) was used to bring the pH of the solution to 3.5, with the colour turning to clear orange. The reaction mixture was stirred at room temperature for one hour,a fter which the pH was 2.98. The reaction mixture was then filtered and kept at 18 8C. Within 2w eeks orange rod-shaped crystals formed that were suitable for crystallography studies. 103.4 mg material was collected (32.41 %yield based on Mo). Elemental analysis calcd (%) for C 24 (1), 1538940 (2), 1538941 (3)a nd 1538942 (4)c ontain the supplementary crystallographic data for this paper.T hese data are provided free of charge by The Cambridge Crystallographic Data Centre.
pH studies:I no rder to map the pH ranges for each compound, many reactions were performed under av ariety of conditions. No significant variation in reaction outcome was observed when the ratios between reagents were changed. All reagents were combined together in the flask, base added and the pH was recorded. After one hour of stirring, the pH of each reaction was recorded once more, which are the values used to establish the optimal pH ranges of each compound. pH was recorded twice due to this value changing over the course of the reaction. As election of the reaction conditions used, detailing the highest and lowest yield, manually set pH values and final pH values are given in Ta bles S9-S12 in the Supporting Information. The synthetic procedures given above correspond to the conditions that resulted in the highest yield.