**2. Results**

#### *2.1. Structural Analysis*

The structure of [β-Mo8O26] <sup>4</sup><sup>−</sup> is preorganized for the coordination of different metal cations due to the presence of two trans-located lacunes (Scheme 1). During the study of complexation in the Ag+/[β-Mo8O26] <sup>4</sup>−/L (L = auxiliary ligand) systems [50,51,58], we found a large number of equilibria that can be easily shifted by the addition of different ligands. In the present case, we tested 4-aminopyridine (py-NH2) as an auxiliary ligand in order to produce a 1D {-Mo8-Ag-py-NH2-Ag-Mo8-} coordination polymer. Instead of this, the reaction gives (Bu4N)3[Ag(py-NH2)Mo8O26] as the main product (phase purity was confirmed by XRPD, see Supplementary Materials, Figure S1). In the crystal structure (SCXRD details are collected in Supplementary Materials Table S1) Ag+, [β-Mo8O26] <sup>4</sup><sup>−</sup> and py-NH2 combine into another type of 1D coordination polymer when [Ag(py-NH2)Mo8O26] 3− anions stack together via py-NH2 . . . O=Mo interactions (Figure 1).

**Figure 1.** py-NH2 . . . O=Mo interactions in the crystal structure of **1**.

Three typical bonding distances surround Ag+: d(Ag1-N1) = 2.27(3) Å, d(Ag1-O6) = 2.372(7) Å, d(Ag1-O2) = 2.544(6) Å and two longer contacts d(Ag1-O9) = 2.639(7) and d(Ag1-O13) = 2.689(7) Å indicate CN = 3+2 for Ag+. These distances are in agreement with the previously published pyridinium complexes of this type [50]. The distances for py-NH2 . . . O=Mo interactions fill the interval between 2.974 and 3.285 Å. The shortest N . . . O contacts 2.974 and 3.021 Å are depicted in blue in Figure 1.

The formation of this coordination polymer via NH2 ... POM interactions is very interesting. We did the structural search for bonding between the oxoligands of the [β-Mo8O26] <sup>4</sup><sup>−</sup> lacunes and H-atoms, and collected 11 hits (BURBOH, CASNIU, COPFIW, EWILIG, GEBYER, GISHEW, HIJSUR, MAXPUZ, MEPNIH, VEHTAF, YAGNOJ) from CCDC (ConQuest Version 2020.2.0). We will use the corresponding refcodes of deposited crystal structures as references in the description below.

The interactions between [β-Mo8O26] <sup>4</sup><sup>−</sup> and Me2NH(R), Me2NH2 <sup>+</sup> and NH4 <sup>+</sup> in the crystal structures of YAGNOJ (a); BURBOH (b); HIJSUR (c); GISHEW (d) are shown in Figure 2.

**Figure 2.** *Cont*.

**Figure 2.** H ... O=Mo interactions in the crystal structure of: (**a**) YAGNOJ; (**b**) BURBOH; (**c**) HIJSUR; (**d**) GISHEW.

According to the structural analysis, R3NH+, R2NH2 <sup>+</sup> and NH4 <sup>+</sup> interact with terminal O=Mo groups of [β-Mo8O26] <sup>4</sup><sup>−</sup> lacunes. Moreover, even Me4N+ can interact with the lacune (VEHTAF). In the crystal structure of GEBYER, the [β-Mo8O26] <sup>4</sup><sup>−</sup> lacunes interact with two H2O molecules. In the case of **1**, we detected interaction between the neutral NH2-group protons with the O=Mo groups of polyoxomolybdate. This illustrates strong attraction between the lacune terminal oxoligands and H-atoms possessing some acidity (chiefly N–H, but also C-H in Me4N+). Considering this, we can suggest direct proton transfer exactly to these oxoligands-producing terminal Mo-OH group, which is highly reactive (M–O π-bonding breaking) and initiates further rearrangement of octamolybdate into hexamolybdate. The detailed mechanistic studies of this transformation are still absent. In this research, we used this channel to initiate the reaction between [β-Mo8O26] 4− and different protonated polyoxometalates serving as proton source. Such direct reactions between two different polyoxometalates are poorly studied. The HPLC-ACP-AES technique was used to control the products.

#### *2.2. Reactivity of [β-Mo8O26] 4*−

The first candidate for this type of reaction was easily prepared (Bu4N)4H2[V10O28]. The HPLC-ICP-AES chromatogram of pure (Bu4N)4[β-Mo8O26] in acetonitrile shows a major molybdenum peak (tR = 3.6 min), corresponding to the octamolybdate anion [β-Mo8O26] <sup>4</sup>−, and a minor peak (tR = 4.8 min), which can be assigned as a hexamolybdate anion [Mo6O19] <sup>2</sup><sup>−</sup> (Figure 3a) [59]. The profile of the major peak is asymmetric due to the presence of [α-Mo8O26] <sup>4</sup>−, according to the previous ESI-MS data, demonstrating the absence of any other molybdates in the solution [50]. The HPLC-ICP-AES chromatogram of a freshly prepared solution of (Bu4N)4H2[V10O28] shows a single peak containing vanadium (tR = 3.0 min), which confirms the presence of individual vanadate anion [V10O28] <sup>6</sup><sup>−</sup> in the solution (Figure 3b). Moreover, the addition of 2 eq of Bu4NOH to the solution of (Bu4N)4H2[V10O28] does not reflect any POM transformation.

**Figure 3.** HPLC-ICP-AES chromatograms of (**a**) [β-Mo8O26] <sup>4</sup><sup>−</sup> and (**b**) [V10O28] <sup>6</sup><sup>−</sup> in acetonitrile.

The reaction between (Bu4N)4[β-Mo8O26] and (Bu4N)4H2[V10O28] in DMSO does not proceed at room temperature, according to 51V NMR data, which is the fastest way to check the reaction progress. The reaction mixture must be heated over 50 ◦C to activate the polyoxometalates' transformation. The HPLC-ICP-AES technique was used to investigate the reaction products at different molar ratios of the reagents. The reaction time was 10 min.

For molar ratio 5/1 (Mo:V = 5/1) at Co of (Bu4N)4[β-Mo8O26] = 6 mM, we observed one peak with the atomic ratio Mo:V = 2.2 (tR = 4.3 min) (Figure 4a) and a second V-free peak (tR = 4.8 min), which may be ascribed to unreacted octamolybdate (Figure 4a). With an increase in the vanadate concentration (Mo/V = 5:2 molar ratio), the same major peak with atomic ratio Mo:V = 2.5 was observed, the intensity of which doubled (Figure 4b). In addition, a chromatogram shows a minor Mo-free peak (tR = 3.0 min), which indicates an excess of the decavanadate anion in this case (Figure 3b).

**Figure 4.** HPLC-ICP-AES chromatograms of the POM mixture at (**a**) the initial Mo/V ratio = 5/1 and (**b**) Mo/V ratio = 5/2.

No significant changes in the chromatograms were observed with a further increase in the concentration of vanadate. The atomic ratio Mo:V = 2.2 indicates the formation of [V2Mo4O19] <sup>4</sup><sup>−</sup> Lindqvist type anions as the reaction product. According to 51V NMR, the total intensity of the other V peaks is ca. 1% of the intensity of the signal from the major product (See NMR part).

The next candidate to study the proton transfer controlled reaction with (Bu4N)4[β-Mo8O26] was [H4SiW12O40]·14H2O. Preliminary experiments showed that the reaction between (Bu4N)4[β-Mo8O26] and [H4SiW12O40]·14H2O in acetonitrile proceeded slowly and led to the formation of a number of products in comparable amounts. Therefore, CH3CN was replaced with dimethyl sulfoxide (DMSO). The HPLC-ICP-AES chromatogram of a freshly prepared solution of silicotungstic acid in DMSO shows a single peak containing tungsten (tR = 6.2 min), which confirms the presence of individual silicotungstate anion [SiW12O40] <sup>4</sup><sup>−</sup> in the solution (Figure 5a) (The intensities of Si lines are significantly lower than W or Mo and cannot be adequately estimated).

**Figure 5.** HPLC-ICP-AES chromatograms of (**a**) [SiW12O40] <sup>4</sup><sup>−</sup> and (**b**) [β-Mo8O26] <sup>4</sup><sup>−</sup> in dimethyl sulfoxide.

The HPLC-ICP-AES chromatogram of (Bu4N)4[β-Mo8O26] in DMSO (Figure 5b) is similar to the chromatogram in acetonitrile (Figure 3a) and shows a major molybdenum peak (tR = 4.5 min), corresponding to the octamolybdate anion [β-Mo8O26] <sup>4</sup>−, and a minor peak of [Mo6O19] <sup>2</sup><sup>−</sup> [59]. Since the viscosity of DMSO is 5 times that of acetonitrile, we were forced to reduce the concentration of the ion-pair reagent in the HPLC eluent to prevent column overpressure. Therefore, the peak retention times in DMSO increased. The HPLC-ICP-AES technique was used to investigate the reaction products between (Bu4N)4[β-Mo8O26] and [H4SiW12O40]·14H2O at different molar ratios. For the Mo/W = 10/1 molar ratio at Co of (Bu4N)4[β-Mo8O26] = 3 mM, we observed four peaks (Figure 6a): (i) unreacted octamolybdate (tR = 4.5 min), (ii) poorly separated peak with atomic ratio Mo:W = 2.3 (tR = 4.7 min), (iii) hexamolybdate (tR = 5.7 min) and (iv) Mo-free peak (tR = 6.2 min) from unreacted silicotungstic acid. With an increase in the tungstate concentration (Mo/W = 10/2 molar ratio), the same major peak with atomic ratio Mo:W = 2.3 was observed (Figure 6b). In addition, the chromatogram shows minor W-free peaks (tR = 4.5 min, tR = 5.6 min) and a single peak containing tungsten (tR = 6.2 min), which may indicate an excess of the tungstate anion. Further increase in the concentration of tungstate (Mo/W = 10/4 molar ratio) leads to the disappearance of the first molybdenum peak ([β-Mo8O26] <sup>4</sup>−, tR = 4.5 min) and an increase in the intensity of the peak of unreacted tungstate.

**Figure 6.** HPLC-ICP-AES chromatograms of the POM mixture at (**a**) the initial Mo/W ratio = 10/1, (**b**) Mo/W ratio = 10/2 and (**c**) Mo/W ratio = 10/4.

Thus, according to the HPLC-ICP-AES results, the proton transfer between the silicotungstic acid and [β-Mo8O26] <sup>4</sup><sup>−</sup> triggers metal redistribution with the formation of Lindqvist type [W2Mo4O19] <sup>2</sup><sup>−</sup> anion as the reorganization product of [β-Mo8O26] <sup>4</sup>−. Curiously, [α-Mo8O26] <sup>4</sup><sup>−</sup> does not react in this case. The Keggin anion almost completely converts into the mixed Lindqvist at Mo/W ratio = 10/1 (Figure 6a). We reported a similar process earlier, when direct reaction of [H3PW12O40] with [NbO(C2O4)2] <sup>−</sup> yielded [PW11NbO40] <sup>4</sup><sup>−</sup> [60].

The reaction between (Bu4N)4[β-Mo8O26] and acetic acid was investigated with the HPLC technique. The reaction was run in DMSO (Co of = 3 mM) by the addition of various concentrations of acetic acid (Figure 7).

The HPLC chromatogram of a freshly prepared solution of (Bu4N)4[β-Mo8O26] shows the peaks from octamolybdate [β-Mo8O26] <sup>4</sup><sup>−</sup> (Figure 5, peak no. 3, tR = 4.5 min) and hexamolybdate [Mo6O19] <sup>2</sup><sup>−</sup> (Figure 7, peak no. 5, tR = 5.6 min) in the ratio of 95:5. Addition of 0.001 M acetic acid decreases the octamolybdate peak intensity, while causing an increase in the hexamolybdate peak intensity and the appearance of a new peak (peak no. 4, tR = 5.1 min). Further increase in the acetic acid concentration continues to reduce the intensity of the octamolybdate peak and leads to an increase in the intensity of peak no. 4, as well as the appearance of two minor peaks (peak no. 1,2) of smaller molybdates. At an acetic acid concentration of 0.008 M, the intensity ratio of the peaks corresponding to octamolybdate (peak no. 3), the new product (peak no. 4), and hexamolydate (peak no. 5), is 1.8:3.4:1, respectively. No further changes in the ratio of species in solution was observed

with an increase in the concentration of acetic acid from 0.008 M to 0.01 M; however, the intensity of all peaks decreases by 1.5, and further acidification leads to the formation of a white precipitate, which makes the HPLC analysis unapplicable.

**Figure 7.** The HPLC-UV chromatograms of (Bu4N)4[β-Mo8O26] in dimethyl sulfoxide with the addition of acetic acid.

From this observation it follows that the transformation of [β-Mo8O26] <sup>4</sup><sup>−</sup> into [Mo6O19] 2− can be explained as direct dimolybdate ([Mo2O7] <sup>2</sup>−) elimination, as was proposed earlier. There are two simple molybdate anions in the reaction mixture. In the literature there is a structure of K[MoO2(OAc)3]·HOAc [61] complex, showing the possibility of [MoO2(OAc)3] <sup>−</sup> existence in the solution. The new peak (Figure 7, peak no. 4) can be assigned as [Mo8O24(OAc)2] <sup>4</sup>−, with the same structure as reported for the malonate derivative ((NH4)4[Mo8O24(C3H2O2)2]·4H2O) [62].

#### *2.3. NMR*

NMR spectroscopy was anticipated to be an informative tool to study the reaction between (Bu4N)4[β-Mo8O26] (**Mo8**) and (Bu4N)4H2[V10O28] (**V10**) due to the presence of both 51V and 95Mo NMR active isotopes. We measured 95Mo NMR spectra for the following solutions to study the effects of acidification of **Mo8** by Hpts (Hpts = *p*-toluenesulfonic acid) (Figure 8).

**Figure 8.** 95Mo NMR data for **Mo8** acidification: (**a**) 70 mg **Mo8** + 2 mg of Hpts in 600 μL DMSO-*d*6; (**b**) 70 mg **Mo8** + 5 mg of Hpts in 600 μL DMSO-*d*6; (**c**) 70 mg **Mo8** + 10 mg of Hpts in 600 μL DMSO-*d*6.

As can be seen, addition of Hpts as a non-coordinating organic acid to the solution of (Bu4N)4[β-Mo8O26] leads to the disappearance of the [α-Mo8O26] <sup>4</sup><sup>−</sup> isomer and an increase in the amount of [Mo6O19] <sup>2</sup>−. Simple (mononuclear or binuclear) Mo-containing complexes were not detected, most likely due to the fast exchange. The signals from such species should appear in a negative region, in comparison with the literature [63].

The reaction between **Mo8** and **V10** was studied using both 95Mo and 51V NMR (Figure 9).

**Figure 9.** 95Mo NMR data describing **Mo8** and **V10** reaction: (**a**) 70 mg **Mo8** (300 μL DMSO-d6) + 55 mg **V10** (300 μL DMSO-*d*6) room temperature; (**b**) 70 mg **Mo8** (300 μL DMSO-d6) + 22 mg **V10** (300 μL DMSO-*d*6), after 10 min at 60 ◦C; (**c**) 70 mg (Bu4N)2[Mo6O19] in 600 μL DMSO-*d*6; (**d**) 70 mg **Mo8** in 600 μL DMSO-*d*6; (**e**) 70 mg of (Bu4N)3Na[V2Mo4O19] in 600 μL DMSO-*d*<sup>6</sup> \* Chemical shift (CS) = 137.4301 ppm, \*\* CS = 123.3993 ppm.

The 51V NMR spectra (Figures S2–S4) show exclusive formation of [V2Mo4O19] <sup>4</sup>−, to the detriment of other mixed metal Lindqvist molybdovanadates, meaning that such reactions can offer a straightforward way to this anion. Two signals in the 95Mo NMR spectrum of (Bu4N)4[β-Mo8O26] indicate an equilibrium between α and β isomers, as described in the literature [64]. In the case of spectrum *b* (Figure 9), the baseline correction was not as accurate, and the peaks from **Mo8** have slightly negative intensities. Moreover, due to this problem, the profile of the main signal is also not as correct. Nevertheless, we can postulate the presence of [V2Mo4O19] <sup>4</sup><sup>−</sup> and [Mo6O19] <sup>2</sup><sup>−</sup> in the reaction mixture.

#### **3. Materials and Methods**

#### *3.1. Physical Methods*

(Bu4N)4[β-Mo8O26], (Bu4N)2[Mo6O19], (Bu4N)2[Mo2O7], (Bu4N)4H2[V10O28] and (Bu4N)3Na[V2Mo4O19] were prepared according to the literature data (Inorg. Synth. Vol. 27). DMSO was distilled in vacuo over NaOH. [H4SiW12O40]·14H2O was manufactured by "The Red Chemist" (Saint Petersburg, USSR) and checked with FT-IR and TGA prior to use. Other reagents were of commercial quality (Sigma Aldrich) and were used as purchased. IR spectra were recorded on a Bruker Vertex 60 FT-IR spectrometer. Elemental analysis was carried out on a MICRO Cube CHN analyzer.

**Synthesis of (Bu4N)3[Ag(py-NH2)Mo8O26] (1)**: (Bu4N)4[β-Mo8O26] (200 mg, mmol) was dissolved in 2 mL of DMF under sonication, and afterward, 20 mg (mmol) of py-NH2 was added to the observed clear solution. Solid AgNO3 (32 mg, mmol) was added to the reaction mixture under sonication. The resulting mixture was placed in Et2O atmosphere at 4 ◦C to obtain crystalline material. The crop of large colorless crystals was isolated after 48 h. Yield was 175 mg (90% based on initial octamolybdate).

Elemental analysis. Calcd C, H, N (%) for **1**: 30.1, 5.4, 3.3; found C, H, N (%): 30.0, 5.3, 3.2. IR (KBr, cm<sup>−</sup>1): 3485 (w), 3336 (m), 3222 (w), 2959 (vs), 2929 (s), 2872 (s), 1632 (vs), 1611 (s), 1554 (w), 1522 (m), 1480 (vs), 1456 (s), 1375 (m), 1355 (w), 1333 (w), 1278 (w), 1214 (m), 1148 (w), 1100 (w), 1060 (w), 1028 (w), 1014 (m), 970 (s), 947 (vs), 928 (vs), 905 (vs), 865 (vs), 847 (vs), 825 (s), 811 (m), 700 (vs), 657 (vs), 567 (m), 551 (s), 520 (s), 472 (m), 444 (w), 407 (s).

#### *3.2. NMR*

51V and 95Mo NMR spectra were recorded on a Bruker Avance III 500 spectrometer (BBI detector), using NaVO3 and Na2MoO4 as internal standards. Spectra were measured in DMSO-*d*<sup>6</sup> at room temperature using standard 5 mm NMR tubes.

#### *3.3. X-ray Diffraction on Single Crystals*

Crystallographic data and refinement details are given in Table S1 (Supplementary Materials). The diffraction data for **1** were collected on a Bruker D8 Venture diffractometer with a CMOS PHOTON III detector and IμS 3.0 source (Mo Kα radiation, λ = 0.71073 Å) at 150 K. The ϕ- and ω-scan techniques were employed. Absorption correction was applied by SADABS (Bruker Apex3 software suite: Apex3, SADABS-2016/2 and SAINT, version 2018.7-2; Bruker AXS Inc.: Madison, WI, USA, 2017). Structures were solved by SHELXT [65] and refined by full-matrix least-squares treatment against |F|2 in anisotropic approximation with SHELX 2014/7 [66] in the ShelXle program [67]. H-atoms were refined in geometrically calculated positions.

CCDC 2215910 contains the supplementary crystallographic data. These data can be obtained free of charge via http://www.ccdc.cam.ac.uk/conts/retrieving.html, or from the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: (+44) 1223-336-033; or e-mail: deposit@ccdc.cam.ac.uk.
