Electrochemical, Electrocatalytic, and Magnetic Properties of Vanadium-Containing Polyoxometalates

: Mono-substituted vanadium-containing Dawson-type polyoxometalates having the general formula α 1 -[V IV W 17 X 2 O 62 ] 8 − and α 2 -[V IV W 17 X 2 O 62 ] 8 − , with X = As or P, were synthesised and subject to a comprehensive electrochemical study comprising the pH dependency. These POMs exhibit an electrocatalytic behaviour towards the oxidation of thiols (namely cysteine), rendering them interesting species for mimicking oxidative stress situations, at physiological pH values. The efﬁciency of the electro-oxidation was assessed with thiols of different nature, and the substrate that responded best was used to compare the electrocatalytic capabilities of the POM series. The magnetic behaviour of these POMs was also evaluated and compared to their analogues, α 1 - and α 2 -[V V W 17 X 2 O 62 ] 7 − (X = As or P), at low temperatures and showed, as expected, a paramagnetic behaviour of V IV based compounds.


Introduction
The polycondensation of oxoanions in aqueous or nonaqueous solution leads to the formation of large molecular structures referred to as polyoxometalates (POMs) [1,2]. A classic POM contains several oxygen atoms, and often hydrogen atoms and at least two other different elements in positive oxidation states. Its structure resembles a fit of discrete oxide fragments of defined sizes and shapes. Its formula is [X x M m O y ] q− , where M is generally tungsten, molybdenum, or vanadium, or other elements such as niobium, zinc, or tantalum. M can also be a mixture of these elements, which are then in their highest oxidation states. The element X, when it exists, is referred to as a hetero-element, and is usually an atom of phosphorus, arsenic, silicon, or any other transition or nonmetallic element [3].
POMs are characterised by the ratio M/X, which is the number of atoms of the condensed element, M, to the number of atoms of the hetero-element, X. Many compounds are obtained either by varying the M/X ratio (12, 11, 9, 18, 17...) or by changing X (Si IV , Ge IV , P V , As V ...) or M (W VI , Mo VI , V V ...) [2]. Different models have been proposed to describe the arrangement of M atoms condensed around X hetero-atoms. These models all have in common the basic unit MO 6 , which is an octahedron whose apexes are each occupied by an oxygen atom, and the centre by the metal atom M. In this study, we were interested in the M/X = 18/2 model, known as the Dawson structure [4].
In the unreduced state, the two compounds [X2W18O62] 6-(where X = As or P), which have all their tungsten atoms in the +VI oxidation state, are diamagnetic. It appears interesting to see how their electronic properties will evolve as a function of the nature, oxidation state, and position of substituents on the tungstic skeleton of these compounds.
In the unreduced state, the two compounds [X 2 W 18 O 62 ] 6-(where X = As or P), which have all their tungsten atoms in the +VI oxidation state, are diamagnetic. It appears interesting to see how their electronic properties will evolve as a function of the nature, oxidation state, and position of substituents on the tungstic skeleton of these compounds.
In this study, we were interested in α 1 -K 8 2 O substituted mixed compounds (with X = As or P) [41]. These compounds were synthesised from the two POMs α-K 6 [X 2 W 18 O 62 ].14H 2 O (where X = As or P) [42,43], then characterised in solution by electrochemistry. The description of the syntheses of the different compounds used in this study is given in the ESI. Cyclic voltammetry allowed us to show the influence of pH on the redox properties of the four compounds. Magnetic studies carried out on the four compounds containing V 4+ centres, as well as on the homologues containing V 5+ centres, confirmed that the former were paramagnetic and the latter diamagnetic.

Materials and Methods
All the compounds used in this study and their precursors were synthesised according to published methods, and their purity attested by electrochemistry [41][42][43][44]. Indeed, cyclic voltammetry and controlled potential coulometry are two electrochemical techniques that allow the unambiguous characterisation of POMs and the certification of their purity [45][46][47].

Magnetic Properties
Magnetic measurements were carried out with a MPMS-5 Quantum Design SQUID magnetometer with applied fields of 1 T for the M vs. T measurements and from 0 to 5 T for the M vs. H measurements at 6, 4, and 2 K. Sample powders were pressed in pellets and held in a straw with sewing cotton to avoid sample holder corrections and the orientation of the crystallites along the magnetic field. The susceptibility data were corrected from diamagnetic contribution by using Pascal's tables.

Electrochemistry
Pure water was obtained with a Milli-Q Integral 5 purification set. All reagents were of high-purity grade and were used as purchased without further purification: CH 3  Electrochemical data were obtained using an EG & G 273 A potentiostat driven by a PC with the M270 software. A divided cell with a standard three-electrode configuration was used for cyclic voltammetry experiments. The reference electrode was a saturated calomel electrode (SCE) and the counter electrode a platinum gauze of large surface area; both electrodes were separated from the bulk electrolyte solution via fritted compartments filled with the same electrolyte. The working electrode was a 3 mm outer diameter glassy carbon (GC) from Mersen, France. The pretreatment of the electrode before each experiment is adapted from a method described elsewhere [48]. Prior to each experiment, solutions were thoroughly deaerated for at least 30 min with pure argon. A positive pressure of this gas was maintained during subsequent work. The potentials are quoted against SCE unless otherwise stated, and all experiments were performed in solution at room temperature. Results were very reproducible from one experiment to the other, and slight variations observed over successive runs were rather attributed to the uncertainty associated with the detection limit of our equipment (potentiostat, hardware, and software) and not to the working electrode pretreatment, nor to possible fluctuations in temperature (polyanion concentrations: 0.5 mM).

Cyclic Voltammetry Characterisation at pH 3
The four compounds, α 1 -K 8 [41,49,50]. Indeed, in the cyclic voltammograms (CVs) in Figure 2a below and Figure S1a in the SI, we observed that the CVs of the phosphorus-containing species were slightly shifted to the left (towards the negative potentials) compared to the CVs of the arsenic-containing species. The W(VI) centres of the α 1 -As and α 2 -As species were more easily reduced than those of their phosphorus-containing counterparts (α 1 -P and α 2 -P). This difference in the reduction peak potential, ∆E c , between the different tungsten waves varied between 38 and 64 mV for the α 1 -As and between 34 and 72 mV for the α 2 -As (see Table 1). Indeed, previous studies have shown that Dawson-type POMs containing As as a hetero-element were easier to reduce than their P-containing counterparts [46,47]. This outcome, which has also been observed and studied in the case of Keggin-type POMs, was the consequence of differences in the atomic radius of the heteroatoms and in their electronegativities [51,52]. electronegativities [51,52].
On the other hand, and as a logical consequence, the V IV centre of the arsenic-containing species was more difficult to oxidise than that of the phosphorus-containing ones. For isomers α1, there was a potential difference of 78 in favour of α1-P, and this difference was lower for isomers α2, with a value of 62 mV in favour of α2-P.
The comparison of the CVs of the two isomers was also in line with the results of previous studies, and allowed us to distinguish them very clearly, especially at the level of their third reduction wave. Indeed, the third reduction step appeared as a two-electron wave with a single reduction peak located at −0.882 V and −0.922 V vs. SCE for α1-As and α1-P, respectively; whereas in the case of the α2 isomers, this ultimate reduction occurred in two separate one-electron steps, each with reduction peaks located at −0.728 V and −0.826 V vs. SCE for α2-As and at −0.774 V and −0.876 V vs. SCE for α2-P, respectively (Figures 2b and S1b; Table 1).  Figure 2. CVs of (a) α1-As (black) and α1-P (red); (b) α1-As (black) and α2-As (red) in 0.2 M Na2SO4 + H2SO4/pH3. Working electrode: GC; counter electrode: Pt; reference electrode: SCE. Scan rate: 10 mV s −1 .  Figure 2. CVs of (a) α 1 -As (black) and α 1 -P (red); (b) α 1 -As (black) and α 2 -As (red) in 0.2 M Na 2 SO 4 + H 2 SO 4 /pH3. Working electrode: GC; counter electrode: Pt; reference electrode: SCE. Scan rate: 10 mV s −1 . On the other hand, and as a logical consequence, the V IV centre of the arseniccontaining species was more difficult to oxidise than that of the phosphorus-containing ones. For isomers α 1 , there was a potential difference of 78 in favour of α 1 -P, and this difference was lower for isomers α 2 , with a value of 62 mV in favour of α 2 -P.
The comparison of the CVs of the two isomers was also in line with the results of previous studies, and allowed us to distinguish them very clearly, especially at the level of their third reduction wave. Indeed, the third reduction step appeared as a two-electron wave with a single reduction peak located at −0.882 V and −0.922 V vs. SCE for α 1 -As and α 1 -P, respectively; whereas in the case of the α 2 isomers, this ultimate reduction occurred in two separate one-electron steps, each with reduction peaks located at −0.728 V and −0.826 V vs. SCE for α 2 -As and at −0.774 V and −0.876 V vs. SCE for α 2 -P, respectively (Figure 2b and Figure S1b; Table 1).

Influence of the pH
It is known that, in general, in an aqueous medium, the exchange of electrons involving POMs on the interface between the solution and the surface of the working electrode is accompanied by an exchange of protons. It is therefore interesting to study the variation of the redox properties of these compounds as a function of the pH, in a buffered medium. The four compounds were characterised in two other media, 0.4 M CH 3 CO 2 Na + CH 3 CO 2 H/pH 5.0 and 0.2 M Na 2 HPO 4 + H 3 PO 4 /pH 7.0.
When we went from pH 3.0 to pH 5.0 and then pH 7.0, the redox potentials attributed to the tungstic centres became increasingly negative, whereas those of the vanadic centres remained invariant (Tables S1 and S2). It should be noted that in certain cases, this shifting of the redox waves towards more negative redox potentials was accompanied by a more or less marked evolution of their shapes (Figure 3). At pH 3.0, the first reduction step of the four compounds was bielectronic, and a well-defined wave with a single reduction peak was observed on the CVs. At pH 5.0, on the CVs of the α 1 isomers, we observed a slight splitting of this wave with the appearance of a second distinct peak in the case of compound α 1 -P, whereas in the case of compound α 1 -As, it was just a shoulder. At pH 7.0, the separation was clear regardless of the CV of the compound considered. This observation showed that the capture of two electrons during the first reduction step was facilitated by the relatively high proton concentration of the pH 3.0 medium. Indeed, the ratio of proton to POM concentration, [H+]/[POM], was 2 at pH 3.0. When we moved to pH 5.0 and pH 7.0, it was divided by 100 and by 10,000, respectively, which made electron transfer more difficult.

Influence of the pH
It is known that, in general, in an aqueous medium, the exchange of electrons involving POMs on the interface between the solution and the surface of the working electrode is accompanied by an exchange of protons. It is therefore interesting to study the variation of the redox properties of these compounds as a function of the pH, in a buffered medium. The four compounds were characterised in two other media, 0.4 M CH3CO2Na + CH3CO2H/pH 5.0 and 0.2 M Na2HPO4 + H3PO4/pH 7.0.
When we went from pH 3.0 to pH 5.0 and then pH 7.0, the redox potentials attributed to the tungstic centres became increasingly negative, whereas those of the vanadic centres remained invariant (Tables S1 and S2). It should be noted that in certain cases, this shifting of the redox waves towards more negative redox potentials was accompanied by a more or less marked evolution of their shapes (Figure 3). At pH 3.0, the first reduction step of the four compounds was bielectronic, and a well-defined wave with a single reduction peak was observed on the CVs. At pH 5.0, on the CVs of the α1 isomers, we observed a slight splitting of this wave with the appearance of a second distinct peak in the case of compound α1-P, whereas in the case of compound α1-As, it was just a shoulder. At pH 7.0, the separation was clear regardless of the CV of the compound considered. This observation showed that the capture of two electrons during the first reduction step was facilitated by the relatively high proton concentration of the pH 3.0 medium. Indeed, the ratio of proton to POM concentration, [H+]/[POM], was 2 at pH 3.0. When we moved to pH 5.0 and pH 7.0, it was divided by 100 and by 10,000, respectively, which made electron transfer more difficult. A closer comparison of the CVs of the α 1 isomers and those of the α 2 isomers led us to make an observation that may have escaped us at first sight: as early as pH 3.0, the first reduction step of α 1 -X (X = As or P) was already more or less composite ( Figure S2). Indeed, this first reduction wave presented a flattened cone shape compared to the pointed cone shape of the α 2 isomer. The other observation worth noting was that only this first reduction step split into two monoelectronic steps when the pH of the medium increased. The other reduction steps were just shifted to negative potentials, without splitting. Moreover, the third reduction step of the α 2 isomer merged with the fourth reduction step into a single, bielectronic step, as from pH 5.0 ( Figure S3).
In order to be able to study all these phenomena, which seemed to be the consequence of the variation of the concentration of protons in the medium, and to eliminate any disturbance that would come from the nature of the electrolyte, we then chose to work with an electrolyte solution that had a buffering capacity in a wide pH window, the Britton-Robinson buffer (S3). For this study, only the phosphorus-based compounds, α 1 -P and α 2 -P, were analysed in a pH range of 1.0 to 8.0, a range in which these vanadic POMs have been shown to be stable [44]. A comparative study between the As-and the P-containing compounds did not appear to be of any interest, as the only difference observed between the two families of compounds was the potential shift mentioned above.
When looking at the redox behaviour of the vanadic centre, it can be seen that for both isomers α 1 and α 2 , the value of the apparent potential, E 0 = (E a + E c )/2, of the V V/IV couple remained invariant in this interval ( Figure 4). We concluded that the oxidation of the V IV centre was not accompanied by an exchange of protons, but was just an exchange of electrons, as shown in the chemical equation (1) below: Magnetochemistry 2021, 7, x FOR PEER REVIEW 6 of 16 Figure 3. CVs of (a) α1-As; (b) α1-P; (c) α2-As and α2-P in 0.2 M Na2SO4 + H2SO4/pH3 (black), 0.4 M CH3CO2Na + CH3CO2H/pH 5.0 (red) and 0.2 M Na2HPO4 + H3PO4/pH 7.0 (blue). Working electrode: GC; counter electrode: Pt; reference electrode: SCE. Scan rate: 10 mV s −1 .
A closer comparison of the CVs of the α1 isomers and those of the α2 isomers led us to make an observation that may have escaped us at first sight: as early as pH 3.0, the first reduction step of α1-X (X = As or P) was already more or less composite ( Figure S2). Indeed, this first reduction wave presented a flattened cone shape compared to the pointed cone shape of the α2 isomer. The other observation worth noting was that only this first reduction step split into two monoelectronic steps when the pH of the medium increased. The other reduction steps were just shifted to negative potentials, without splitting. Moreover, the third reduction step of the α2 isomer merged with the fourth reduction step into a single, bielectronic step, as from pH 5.0 ( Figure S3).
In order to be able to study all these phenomena, which seemed to be the consequence of the variation of the concentration of protons in the medium, and to eliminate any disturbance that would come from the nature of the electrolyte, we then chose to work with an electrolyte solution that had a buffering capacity in a wide pH window, the Britton-Robinson buffer (S3). For this study, only the phosphorus-based compounds, α1-P and α2-P, were analysed in a pH range of 1.0 to 8.0, a range in which these vanadic POMs have been shown to be stable [44]. A comparative study between the As-and the P-containing compounds did not appear to be of any interest, as the only difference observed between the two families of compounds was the potential shift mentioned above.
When looking at the redox behaviour of the vanadic centre, it can be seen that for both isomers α1 and α2, the value of the apparent potential, E 0′ = (Ea + Ec)/2, of the V V/IV couple remained invariant in this interval ( Figure 4). We concluded that the oxidation of the V IV centre was not accompanied by an exchange of protons, but was just an exchange of electrons, as shown in the chemical equation (1) below: In the case of the redox processes attributed to the tungstic backbone, the redox behaviours of the two isomers were not similar, as was the case for the vanadic centre. In the case of the α1-P isomer, for example, one would have expected to see the first reduction wave lose its composite character completely when the pH of the electrolyte became well below 3. On the contrary, at pH 1.0, the composite character of this first reduction wave was strengthened, and the CV recorded in this medium had a shoulder at −0.22 V vs. SCE which stood out very clearly from the main peak located at −0.32 V vs. SCE. The shape of the other two reduction waves that followed, although bielectronic, corresponded to a single reduction peak located at −0.49 V and −0.77 V vs. SCE, respectively ( Figure 5a). Taking the pH 1.0 solution as a starting point, and increasing the pH very slowly, we In the case of the redox processes attributed to the tungstic backbone, the redox behaviours of the two isomers were not similar, as was the case for the vanadic centre. In the case of the α 1 -P isomer, for example, one would have expected to see the first reduction wave lose its composite character completely when the pH of the electrolyte became well below 3. On the contrary, at pH 1.0, the composite character of this first reduction wave was strengthened, and the CV recorded in this medium had a shoulder at −0.22 V vs. SCE which stood out very clearly from the main peak located at −0.32 V vs. SCE. The shape of the other two reduction waves that followed, although bielectronic, corresponded to a single reduction peak located at −0.49 V and −0.77 V vs. SCE, respectively (Figure 5a). Taking the pH 1.0 solution as a starting point, and increasing the pH very slowly, we observed the progressive disappearance of the shoulder of the first wave. At pH 2.3, this shoulder had completely disappeared, and we had a slightly broad wave with a single reduction peak at −0.39 V vs. SCE (Figure 5a). The CV of α 1 -P then consisted of three two-electron waves that shifted, all three towards negative potential values, as the pH of the solution increased. This pattern of three two-electron waves having a single reduction peak each would remain the same until pH values around 3.5. Figure 5b). However, there would still be a final separation of this first reduction step, as from pH 6.3.

(
Of the three reduction steps observable in this wide pH range, only the first wave exhibited this behaviour. As for the next two, the only observation made was the very expected decrease in the values of the reduction peak potentials when the pH of the electrolyte increased. However, it should be noted that, for the second reduction wave, a plateau was reached at pH values between 3.64 and 5.38. This behaviour already was noted in one of our previous studies, in which an inversion of the slope of the evolution of the reduction peak potential values was observed between pH 4 and 5 [48]. The redox behaviour of the α2-P isomer as a function of the pH was more classical. No wave splitting was observed for pH values below 5.4 ( Figure 6a). All three bielectronic waves shifted towards the negative potential values, with all three keeping their starting shapes almost unchanged. When we compared the first reduction of the α2-P isomer with that of the α1-P isomer, we realised that their behaviours, which seemed to diverge between pH 1.0 and pH 5.4, began to converge, as from pH 6.3 ( Figure S4). At pH 7.5, two almost superimposable monoelectronic waves were obtained (Figure 6b). From pH 3.8 onwards, the progressive separation of the first reduction wave into two more or less distinct steps was again observed. At pH 4.7, there were two distinct reduction peaks at −0.42 V and at −0.49 V vs. SCE. These two peaks would gradually disappear to give way to a wave with a single reduction peak, as shown on the CV recorded at pH 5.5 ( Figure 5b). However, there would still be a final separation of this first reduction step, as from pH 6.3.
Of the three reduction steps observable in this wide pH range, only the first wave exhibited this behaviour. As for the next two, the only observation made was the very expected decrease in the values of the reduction peak potentials when the pH of the electrolyte increased. However, it should be noted that, for the second reduction wave, a plateau was reached at pH values between 3.64 and 5.38. This behaviour already was noted in one of our previous studies, in which an inversion of the slope of the evolution of the reduction peak potential values was observed between pH 4 and 5 [48].
The redox behaviour of the α 2 -P isomer as a function of the pH was more classical. No wave splitting was observed for pH values below 5.4 ( Figure 6a). All three bielectronic waves shifted towards the negative potential values, with all three keeping their starting shapes almost unchanged. When we compared the first reduction of the α 2 -P isomer with that of the α 1 -P isomer, we realised that their behaviours, which seemed to diverge between pH 1.0 and pH 5.4, began to converge, as from pH 6.3 ( Figure S4). At pH 7.5, two almost superimposable monoelectronic waves were obtained (Figure 6b).
observed the progressive disappearance of the shoulder of the first wave. At pH 2.3, this shoulder had completely disappeared, and we had a slightly broad wave with a single reduction peak at −0.39 V vs. SCE (Figure 5a). The CV of α1-P then consisted of three twoelectron waves that shifted, all three towards negative potential values, as the pH of the solution increased. This pattern of three two-electron waves having a single reduction peak each would remain the same until pH values around 3.5.
From pH 3.8 onwards, the progressive separation of the first reduction wave into two more or less distinct steps was again observed. At pH 4.7, there were two distinct reduction peaks at −0.42 V and at −0.49 V vs. SCE. These two peaks would gradually disappear to give way to a wave with a single reduction peak, as shown on the CV recorded at pH 5.5 (Figure 5b). However, there would still be a final separation of this first reduction step, as from pH 6.3.
Of the three reduction steps observable in this wide pH range, only the first wave exhibited this behaviour. As for the next two, the only observation made was the very expected decrease in the values of the reduction peak potentials when the pH of the electrolyte increased. However, it should be noted that, for the second reduction wave, a plateau was reached at pH values between 3.64 and 5.38. This behaviour already was noted in one of our previous studies, in which an inversion of the slope of the evolution of the reduction peak potential values was observed between pH 4 and 5 [48]. The redox behaviour of the α2-P isomer as a function of the pH was more classical. No wave splitting was observed for pH values below 5.4 (Figure 6a). All three bielectronic waves shifted towards the negative potential values, with all three keeping their starting shapes almost unchanged. When we compared the first reduction of the α2-P isomer with that of the α1-P isomer, we realised that their behaviours, which seemed to diverge between pH 1.0 and pH 5.4, began to converge, as from pH 6.3 ( Figure S4). At pH 7.5, two almost superimposable monoelectronic waves were obtained (Figure 6b). Let us now turn to the last reduction step of the α 2 -P isomer. Indeed, in the pH range between 2 and 4, this wave split to give two very distinct reduction peaks. Outside these two limits, as shown on the CVs in Figure S5, there was a single-peak reduction wave. This singular behaviour was only observed in the case of this isomer. Indeed, beyond the first reduction step, no wave splitting was observed in the case of the α 2 isomer.
In summary, when the pH of the electrolyte was varied, the redox behaviour of the two isomers could be distinguished, as from the first reduction wave. The latter, which was globally bielectronic, could split into two monoelectronic steps or merge into a single step, depending on the pH. The pH ranges in which the wave separation or fusion were observed were not identical for the two isomers (see Scheme 1). In fact, for the α 2 -isomer, the reduction took place in a single bielectronic step up to pH 5.5, and above this pH value, a process with two very distinct monoelectronic steps was observed. In the case of the α 1 -isomer, we observed an alternation of bielectronic and monoelectronic processes. Up to pH 1.9, two single-electron steps were observed more or less distinctly, which merged from this pH value onwards into a single double-electron wave. A new separation was observed at pH 3.7, from which two new monoelectronic steps would coexist until pH 5, at which point they merged again. A final splitting took place, as from pH 6.2. Figure 6. (a) CVs of α2-P in Britton-Robinson buffer at pH 1.0 (black) and pH 5.4 (red); (b) CVs of α1-P (black) and α2-P (red) in Britton-Robinson buffer at pH 7.5. Working electrode: GC; counter electrode: Pt; reference electrode: SCE. Scan rate: 100 mV s −1 .
Let us now turn to the last reduction step of the α2-P isomer. Indeed, in the pH range between 2 and 4, this wave split to give two very distinct reduction peaks. Outside these two limits, as shown on the CVs in Figure S5, there was a single-peak reduction wave. This singular behaviour was only observed in the case of this isomer. Indeed, beyond the first reduction step, no wave splitting was observed in the case of the α2 isomer.
In summary, when the pH of the electrolyte was varied, the redox behaviour of the two isomers could be distinguished, as from the first reduction wave. The latter, which was globally bielectronic, could split into two monoelectronic steps or merge into a single step, depending on the pH. The pH ranges in which the wave separation or fusion were observed were not identical for the two isomers (see Scheme 1). In fact, for the α2-isomer, the reduction took place in a single bielectronic step up to pH 5.5, and above this pH value, a process with two very distinct monoelectronic steps was observed. In the case of the α1isomer, we observed an alternation of bielectronic and monoelectronic processes. Up to pH 1.9, two single-electron steps were observed more or less distinctly, which merged from this pH value onwards into a single double-electron wave. A new separation was observed at pH 3.7, from which two new monoelectronic steps would coexist until pH 5, at which point they merged again. A final splitting took place, as from pH 6.2. The redox behaviour of the α1-P isomer was also singular at the second reduction wave (Figure 7a below). While the evolution of the reduction peak potentials as a function of the pH was continuous and uniform for the α2-P isomer (for pH < 6.3), a plateau was observed between pH 3.6 and 4.6 for the α1-P isomer. On the other hand, the two isomers showed an almost identical behaviour as far as the third reduction wave was concerned. The evolution of the values of the reduction peak potentials as a function of the pH of the electrolyte followed a similar linear trend for both isomers (see SI). This corresponded broadly to the involvement of two protons concomitant with the exchange of two electrons. The redox behaviour of the α 1 -P isomer was also singular at the second reduction wave (Figure 7a below). While the evolution of the reduction peak potentials as a function of the pH was continuous and uniform for the α 2 -P isomer (for pH < 6.3), a plateau was observed between pH 3.6 and 4.6 for the α 1 -P isomer. On the other hand, the two isomers showed an almost identical behaviour as far as the third reduction wave was concerned. The evolution of the values of the reduction peak potentials as a function of the pH of the electrolyte followed a similar linear trend for both isomers (see SI). This corresponded broadly to the involvement of two protons concomitant with the exchange of two electrons.

Electrocatalytic Properties
The 0.2 M Na2HPO4 + H3PO4/pH 7.0 medium was selected to evaluate and compare

Electrocatalytic Properties
The 0.2 M Na 2 HPO 4 + H 3 PO 4 /pH 7.0 medium was selected to evaluate and compare the efficiency of the four compounds with respect to the electro-oxidation of a thiol (3mercaptopropanol). For this purpose, we will focus on the V V/IV redox couple. We observed for each of the compounds a reversible and monoelectronic oxidation wave when they were the sole electroactive species in solution, as shown in Figures S6b and S7a. In the presence of the thiol, we noticed the exponential increase of the oxidation current, as well as the loss of reversibility of the said waves. These two phenomena clearly indicated the existence of an electrocatalytic process at the level of the V V/IV wave (Figure 8a, Figures  S6a and S7b). Two parameters allowed us to quantify and compare the efficiency of these four compounds with respect to the electrocatalytic oxidation of the thiol: (1) the catalytic efficiency, CAT, which was calculated using the formula CAT = [(I cat − I 0 )/I 0 ] × 100; and (2) the onset potential of the catalytic reaction, ONSET.

Electrocatalytic Properties
The 0.2 M Na2HPO4 + H3PO4/pH 7.0 medium was selected to evaluate and compare the efficiency of the four compounds with respect to the electro-oxidation of a thiol (3mercaptopropanol). For this purpose, we will focus on the V V/IV redox couple. We observed for each of the compounds a reversible and monoelectronic oxidation wave when they were the sole electroactive species in solution, as shown in Figures S6b and S7a. In the presence of the thiol, we noticed the exponential increase of the oxidation current, as well as the loss of reversibility of the said waves. These two phenomena clearly indicated the existence of an electrocatalytic process at the level of the V V/IV wave (Figures 8a, S6a, and S7b). Two parameters allowed us to quantify and compare the efficiency of these four compounds with respect to the electrocatalytic oxidation of the thiol: (1) the catalytic efficiency, CAT, which was calculated using the formula CAT = [(Icat − I0)/I0] × 100; and (2) the onset potential of the catalytic reaction, ONSET.
As expected, the electrocatalytic process began earlier in the case of the α2-isomers, for which the oxidation of the vanadium centre took place at more favourable potentials. However, when comparing the CAT values calculated at the same potential (E = 0.75 V vs. SCE), it became clear that the α1-isomers showed the highest catalytic efficiency ( Table  2). This was in agreement with previous experimental and theoretical studies, which showed that the electron density was higher at the equatorial region of the Dawson structure (position α1-) than at its apical region (position α2-) [53,54]. As expected, the electrocatalytic process began earlier in the case of the α 2 -isomers, for which the oxidation of the vanadium centre took place at more favourable potentials. However, when comparing the CAT values calculated at the same potential (E = 0.75 V vs. SCE), it became clear that the α 1 -isomers showed the highest catalytic efficiency (Table 2). This was in agreement with previous experimental and theoretical studies, which showed that the electron density was higher at the equatorial region of the Dawson structure (position α 1 -) than at its apical region (position α 2 -) [53,54]. Table 2. Oxidation peak currents in the absence (I 0 ) and in the presence (I CAT ) of thiol. CAT and ONSET data were calculated or measured from values taken from CVs recorded in 0.2 M Na 2 HPO 4 + H 3 PO 4 /pH 7.0 (see above for the conditions).

Magnetic Properties
The magnetic behaviours of the four V 4+ compounds (α 1 -As V 4+ , α 2 -As V 4+ , α 1 -P V 4+ , and α 2 -P V 4+ ) and two V 5+ compounds (α 2 -AsV 5+ and α 2 -PV 5+ ) were investigated. V 5+ compounds were assumed to be diamagnetic, but their characterisation could allow us to evaluate the diamagnetic correction that should be applied for the calculation of the magnetization of V 4+ compounds. Indeed, this evaluation by Pascal's tables is always delicate in the case of POMs, as is the case for V 4+ POMs, for which we had only a spin 1/2 for a molar mass of the order of 4900 g.mol −1 and an important number of water molecules. M vs. B at 2 K plots are shown in Figure 9, and χ m T vs. T curves (where χ m is the molar susceptibility) in the range of 250 to 2 K (B = 1 T) in Figure 10. Both compounds presented a weak paramagnetic behaviour that may have been due to V 4+ impurities. The amount of impurities could be deduced from the M vs. B curves at 2 K. For an ion S = 1/2 with g = 2, the value of the magnetization expected at 5 T was 0.93 µB (black curve in Figure 9). Here, for α 2 -AsV 5+ the magnetization was 0.0115 µB and 0.0313 for α 2 -PV 5+ , representing about 1.2% and 3.3% of impurities, respectively.

Magnetic Properties
The magnetic behaviours of the four V 4+ compounds (α1-As V 4+ , α2-As V 4+ , α1-P V 4+ , and α2-P V 4+ ) and two V 5+ compounds (α2-AsV 5+ and α2-PV 5+ ) were investigated. V 5+ compounds were assumed to be diamagnetic, but their characterisation could allow us to evaluate the diamagnetic correction that should be applied for the calculation of the magnetization of V 4+ compounds. Indeed, this evaluation by Pascal's tables is always delicate in the case of POMs, as is the case for V 4+ POMs, for which we had only a spin 1/2 for a molar mass of the order of 4900 g.mol −1 and an important number of water molecules. M vs. B at 2 K plots are shown in Figure 9, and χmT vs. T curves (where χm is the molar susceptibility) in the range of 250 to 2 K (B = 1 T) in Figure 10. Both compounds presented a weak paramagnetic behaviour that may have been due to V 4+ impurities. The amount of impurities could be deduced from the M vs. B curves at 2 K. For an ion S = 1/2 with g = 2, the value of the magnetization expected at 5 T was 0.93 μB (black curve in Figure 9). Here, for α2-AsV 5+ the magnetization was 0.0115 μB and 0.0313 for α2-PV 5+ , representing about 1.2% and 3.3% of impurities, respectively. For α2-AsV 5+ and α2-PV 5+ , the χmT decreased linearly from 0.035 cm 3 .mol −1 K (0.03 cm 3 .mol −1 .K, respectively) at 250 K to about 0.009 cm 3 mol −1 K (0.015) at 20 K. If it is expected that the impurity obeys Curie's law, the slope of this straight line can be due to the presence of temperature-independent paramagnetism (TIP) relative to the POM. Curie's law can be written: χMT = xiC + χTIP.T, where C stands for the Curie constant, xi the fraction of impurity, and χTIP the TIP. The value obtained for the TIP was 115 × 10 −6 cm 3 .mol −1 for α2-AsV 5+ and 16 × 10 −6 cm 3 mol −1 for α2-PV 5+ . TIP with similar values could be expected for the V 4+ compounds.

Magnetic Properties
The magnetic behaviours of the four V 4+ compounds (α1-As V 4+ , α2-As V 4+ , α1-P V 4+ , and α2-P V 4+ ) and two V 5+ compounds (α2-AsV 5+ and α2-PV 5+ ) were investigated. V 5+ compounds were assumed to be diamagnetic, but their characterisation could allow us to evaluate the diamagnetic correction that should be applied for the calculation of the magnetization of V 4+ compounds. Indeed, this evaluation by Pascal's tables is always delicate in the case of POMs, as is the case for V 4+ POMs, for which we had only a spin 1/2 for a molar mass of the order of 4900 g.mol −1 and an important number of water molecules. M vs. B at 2 K plots are shown in Figure 9, and χmT vs. T curves (where χm is the molar susceptibility) in the range of 250 to 2 K (B = 1 T) in Figure 10. Both compounds presented a weak paramagnetic behaviour that may have been due to V 4+ impurities. The amount of impurities could be deduced from the M vs. B curves at 2 K. For an ion S = 1/2 with g = 2, the value of the magnetization expected at 5 T was 0.93 μB (black curve in Figure 9). Here, for α2-AsV 5+ the magnetization was 0.0115 μB and 0.0313 for α2-PV 5+ , representing about 1.2% and 3.3% of impurities, respectively. For α2-AsV 5+ and α2-PV 5+ , the χmT decreased linearly from 0.035 cm 3 .mol −1 K (0.03 cm 3 .mol −1 .K, respectively) at 250 K to about 0.009 cm 3 mol −1 K (0.015) at 20 K. If it is expected that the impurity obeys Curie's law, the slope of this straight line can be due to the presence of temperature-independent paramagnetism (TIP) relative to the POM. Curie's law can be written: χMT = xiC + χTIP.T, where C stands for the Curie constant, xi the fraction of impurity, and χTIP the TIP. The value obtained for the TIP was 115 × 10 −6 cm 3 .mol −1 for α2-AsV 5+ and 16 × 10 −6 cm 3 mol −1 for α2-PV 5+ . TIP with similar values could be expected for the V 4+ compounds.

Magnetic Properties
The magnetic behaviours of the four V 4+ compounds (α1-As V 4+ , α2-As V 4+ , α1-P V 4+ , and α2-P V 4+ ) and two V 5+ compounds (α2-AsV 5+ and α2-PV 5+ ) were investigated. V 5+ compounds were assumed to be diamagnetic, but their characterisation could allow us to evaluate the diamagnetic correction that should be applied for the calculation of the magnetization of V 4+ compounds. Indeed, this evaluation by Pascal's tables is always delicate in the case of POMs, as is the case for V 4+ POMs, for which we had only a spin 1/2 for a molar mass of the order of 4900 g.mol −1 and an important number of water molecules. M vs. B at 2 K plots are shown in Figure 9, and χmT vs. T curves (where χm is the molar susceptibility) in the range of 250 to 2 K (B = 1 T) in Figure 10. Both compounds presented a weak paramagnetic behaviour that may have been due to V 4+ impurities. The amount of impurities could be deduced from the M vs. B curves at 2 K. For an ion S = 1/2 with g = 2, the value of the magnetization expected at 5 T was 0.93 μB (black curve in Figure 9). Here, for α2-AsV 5+ the magnetization was 0.0115 μB and 0.0313 for α2-PV 5+ , representing about 1.2% and 3.3% of impurities, respectively. For α2-AsV 5+ and α2-PV 5+ , the χmT decreased linearly from 0.035 cm 3 .mol −1 K (0.03 cm 3 .mol −1 .K, respectively) at 250 K to about 0.009 cm 3 mol −1 K (0.015) at 20 K. If it is expected that the impurity obeys Curie's law, the slope of this straight line can be due to the presence of temperature-independent paramagnetism (TIP) relative to the POM. Curie's law can be written: χMT = xiC + χTIP.T, where C stands for the Curie constant, xi the fraction of impurity, and χTIP the TIP. The value obtained for the TIP was 115 × 10 −6 cm 3 .mol −1 for α2-AsV 5+ and 16 × 10 −6 cm 3 mol −1 for α2-PV 5+ . TIP with similar values could be expected for the V 4+ compounds.
) at 2 K compared to the theoretical curve for S=1/2. Regarding to the V 4+ compounds, the results are shown in Figure 11 in the form of χmT vs. T plots.
For the As compounds, χmT decreased linearly when lowering the temperature from 250 K to 20 K. As seen previously, this may have been due to presence of TIP, which could be estimated at 60 × 10 −6 and 29 × 10 −6 cm 3 .mol −1 for α1-AsV 4+ and α2-AsV 4+ , respectively. The Curie constants and the corrections deducted from these measurements are summarised in Table 3. Below 20 K, the decrease of χT was faster to reach 0.332 and 0.338 cm 3 .mol −1 K. This decrease was expected at low temperatures under a magnetic field of 1 T (see S8).

Magnetic Properties
The magnetic behaviours of the four V 4+ compounds (α1-As V 4+ , α2-As V 4+ , α1-P V 4+ , and α2-P V 4+ ) and two V 5+ compounds (α2-AsV 5+ and α2-PV 5+ ) were investigated. V 5+ compounds were assumed to be diamagnetic, but their characterisation could allow us to evaluate the diamagnetic correction that should be applied for the calculation of the magnetization of V 4+ compounds. Indeed, this evaluation by Pascal's tables is always delicate in the case of POMs, as is the case for V 4+ POMs, for which we had only a spin 1/2 for a molar mass of the order of 4900 g.mol −1 and an important number of water molecules. M vs. B at 2 K plots are shown in Figure 9, and χmT vs. T curves (where χm is the molar susceptibility) in the range of 250 to 2 K (B = 1 T) in Figure 10. Both compounds presented a weak paramagnetic behaviour that may have been due to V 4+ impurities. The amount of impurities could be deduced from the M vs. B curves at 2 K. For an ion S = 1/2 with g = 2, the value of the magnetization expected at 5 T was 0.93 μB (black curve in Figure 9). Here, for α2-AsV 5+ the magnetization was 0.0115 μB and 0.0313 for α2-PV 5+ , representing about 1.2% and 3.3% of impurities, respectively. For α2-AsV 5+ and α2-PV 5+ , the χmT decreased linearly from 0.035 cm 3 .mol −1 K (0.03 cm 3 .mol −1 .K, respectively) at 250 K to about 0.009 cm 3 mol −1 K (0.015) at 20 K. If it is expected that the impurity obeys Curie's law, the slope of this straight line can be due to the presence of temperature-independent paramagnetism (TIP) relative to the POM. Curie's law can be written: χMT = xiC + χTIP.T, where C stands for the Curie constant, xi the fraction of impurity, and χTIP the TIP. The value obtained for the TIP was 115 × 10 −6 cm 3 .mol −1 for α2-AsV 5+ and 16 × 10 −6 cm 3 mol −1 for α2-PV 5+ . TIP with similar values could be expected for the V 4+ compounds.

Magnetic Properties
The magnetic behaviours of the four V 4+ compounds (α1-As V 4+ , α2-As V 4+ , α1-P V 4+ , and α2-P V 4+ ) and two V 5+ compounds (α2-AsV 5+ and α2-PV 5+ ) were investigated. V 5+ compounds were assumed to be diamagnetic, but their characterisation could allow us to evaluate the diamagnetic correction that should be applied for the calculation of the magnetization of V 4+ compounds. Indeed, this evaluation by Pascal's tables is always delicate in the case of POMs, as is the case for V 4+ POMs, for which we had only a spin 1/2 for a molar mass of the order of 4900 g.mol −1 and an important number of water molecules. M vs. B at 2 K plots are shown in Figure 9, and χmT vs. T curves (where χm is the molar susceptibility) in the range of 250 to 2 K (B = 1 T) in Figure 10. Both compounds presented a weak paramagnetic behaviour that may have been due to V 4+ impurities. The amount of impurities could be deduced from the M vs. B curves at 2 K. For an ion S = 1/2 with g = 2, the value of the magnetization expected at 5 T was 0.93 μB (black curve in Figure 9). Here, for α2-AsV 5+ the magnetization was 0.0115 μB and 0.0313 for α2-PV 5+ , representing about 1.2% and 3.3% of impurities, respectively. For α2-AsV 5+ and α2-PV 5+ , the χmT decreased linearly from 0.035 cm 3 .mol −1 K (0.03 cm 3 .mol −1 .K, respectively) at 250 K to about 0.009 cm 3 mol −1 K (0.015) at 20 K. If it is expected that the impurity obeys Curie's law, the slope of this straight line can be due to the presence of temperature-independent paramagnetism (TIP) relative to the POM. Curie's law can be written: χMT = xiC + χTIP.T, where C stands for the Curie constant, xi the fraction of impurity, and χTIP the TIP. The value obtained for the TIP was 115 × 10 −6 cm 3 .mol −1 for α2-AsV 5+ and 16 × 10 −6 cm 3 mol −1 for α2-PV 5+ . TIP with similar values could be expected for the V 4+ compounds.
) under a magnetic field of 1 T.
For α 2 -AsV 5+ and α 2 -PV 5+ , the χ m T decreased linearly from 0.035 cm 3 .mol −1 K (0.03 cm 3 .mol −1 .K, respectively) at 250 K to about 0.009 cm 3 mol −1 K (0.015) at 20 K. If it is expected that the impurity obeys Curie's law, the slope of this straight line can be due to the presence of temperature-independent paramagnetism (TIP) relative to the POM. Curie's law can be written: χ M T = x i C + χ TIP .T, where C stands for the Curie constant, x i the fraction of impurity, and χ TIP the TIP. The value obtained for the TIP was 115 × 10 −6 cm 3 .mol −1 for α 2 -AsV 5+ and 16 × 10 −6 cm 3 mol −1 for α 2 -PV 5+ . TIP with similar values could be expected for the V 4+ compounds.
Regarding to the V 4+ compounds, the results are shown in Figure 11 in the form of χ m T vs. T plots.
For the P compounds, the χmT increased linearly when lowering the temperature. This may have indicated that the diamagnetism was not correctly taken into account. An additional correction (χdia) of −112 × 10 −6 and −25 × 10 −6 cm 3 .mol −1 for α1-PV 4+ and α2-PV 4+ , respectively, was added so that Curie's law was verified (χmT = C + χdiaT). Nevertheless, to exclude other possibilities of such behaviour (antiferromagnetic coupling for the As-containing POMs or ferromagnetic coupling for the P-containing POMs), magnetization M vs. field B measurements were carried out at low temperatures (6, 4, and 2 K). The results are depicted in Figure 12. These curves, in the form of M vs. B/T plots, could be simulated with a Brillouin function [55] for a S = 1/2 ion, thus showing that the For the P compounds, the χmT increased linearly when lowering the temperature. This may have indicated that the diamagnetism was not correctly taken into account. An additional correction (χdia) of −112 × 10 −6 and −25 × 10 −6 cm 3 .mol −1 for α1-PV 4+ and α2-PV 4+ , respectively, was added so that Curie's law was verified (χmT = C + χdiaT). Nevertheless, to exclude other possibilities of such behaviour (antiferromagnetic coupling for the As-containing POMs or ferromagnetic coupling for the P-containing POMs), magnetization M vs. field B measurements were carried out at low temperatures (6, 4, and 2 K). The results are depicted in Figure 12. These curves, in the form of M vs. B/T plots, could be simulated with a Brillouin function [55] for a S = 1/2 ion, thus showing that the ), α 2 -As V 4+ ( T (see S8).
For the P compounds, the χmT increased linearly when lowering the temperature. This may have indicated that the diamagnetism was not correctly taken into account. An additional correction (χdia) of −112 × 10 −6 and −25 × 10 −6 cm 3 .mol −1 for α1-PV 4+ and α2-PV 4+ , respectively, was added so that Curie's law was verified (χmT = C + χdiaT). Nevertheless, to exclude other possibilities of such behaviour (antiferromagnetic coupling for the As-containing POMs or ferromagnetic coupling for the P-containing POMs), magnetization M vs. field B measurements were carried out at low temperatures (6, 4, and 2 K). The results are depicted in Figure 12. These curves, in the form of M vs. B/T plots, could be simulated with a Brillouin function [55] for a S = 1/2 ion, thus showing that the ), α 1 -PV 4+ ( T (see S8).
For the P compounds, the χmT increased linearly when lowering the temperature. This may have indicated that the diamagnetism was not correctly taken into account. An additional correction (χdia) of −112 × 10 −6 and −25 × 10 −6 cm 3 .mol −1 for α1-PV 4+ and α2-PV 4+ , respectively, was added so that Curie's law was verified (χmT = C + χdiaT). Nevertheless, to exclude other possibilities of such behaviour (antiferromagnetic coupling for the As-containing POMs or ferromagnetic coupling for the P-containing POMs), magnetization M vs. field B measurements were carried out at low temperatures (6, 4, and 2 K). The results are depicted in Figure 12. These curves, in the form of M vs. B/T plots, could be simulated with a Brillouin function [55] for a S = 1/2 ion, thus showing that the ) and α 2 -PV 4+ ( T (see S8).
For the P compounds, the χmT increased linearly when lowering the temperature. This may have indicated that the diamagnetism was not correctly taken into account. An additional correction (χdia) of −112 × 10 −6 and −25 × 10 −6 cm 3 .mol −1 for α1-PV 4+ and α2-PV 4+ , respectively, was added so that Curie's law was verified (χmT = C + χdiaT). Nevertheless, to exclude other possibilities of such behaviour (antiferromagnetic coupling for the As-containing POMs or ferromagnetic coupling for the P-containing POMs), magnetization M vs. field B measurements were carried out at low temperatures (6, 4, and 2 K). The results are depicted in Figure 12. These curves, in the form of M vs. B/T plots, could be simulated with a Brillouin function [55] for a S = 1/2 ion, thus showing that the ).
For the As compounds, χ m T decreased linearly when lowering the temperature from 250 K to 20 K. As seen previously, this may have been due to presence of TIP, which could be estimated at 60 × 10 −6 and 29 × 10 −6 cm 3 .mol −1 for α 1 -AsV 4+ and α 2 -AsV 4+ , respectively. The Curie constants and the corrections deducted from these measurements are summarised in Table 3. Below 20 K, the decrease of χT was faster to reach 0.332 and 0.338 cm 3 .mol −1 K. This decrease was expected at low temperatures under a magnetic field of 1 T (see S8). For the P compounds, the χ m T increased linearly when lowering the temperature. This may have indicated that the diamagnetism was not correctly taken into account. An additional correction (χ dia ) of −112 × 10 −6 and −25 × 10 −6 cm 3 .mol −1 for α 1 -PV 4+ and α 2 -PV 4+ , respectively, was added so that Curie's law was verified (χ m T = C + χ dia T).
Nevertheless, to exclude other possibilities of such behaviour (antiferromagnetic coupling for the As-containing POMs or ferromagnetic coupling for the P-containing POMs), magnetization M vs. field B measurements were carried out at low temperatures (6, 4, and 2 K). The results are depicted in Figure 12. These curves, in the form of M vs. B/T plots, could be simulated with a Brillouin function [55] for a S = 1/2 ion, thus showing that the V 4+ ions were isolated, and that the decrease or the increase observed in χ m T in the range 20-250 K was not the result of a magnetic coupling.
Magnetochemistry 2021, 7, x FOR PEER REVIEW 12 of 16 V 4+ ions were isolated, and that the decrease or the increase observed in χmT in the range 20-250 K was not the result of a magnetic coupling.  Table 3.
For the AsV 4+ compound, the g values obtained, even if weak, were in agreement with the values found in the literature [56,57], but those obtained for the PV 4+ compound were very weak, confirming the possible underestimation of the molar mass for these compounds (Table 3).
Despite the experimental difficulties related to the weak magnetization of these POMs (a spin of 1/2 in a large diamagnetic matrix), it was confirmed that the V 4+ -containing compounds behaved as isolated S = 1/2 monomers, while the V 5+ -containing ones were globally diamagnetic.

Magnetic Properties
The magnetic behaviours of the four V 4+ compounds (α1-As V 4+ and α2-P V 4+ ) and two V 5+ compounds (α2-AsV 5+ and α2-PV 5+ ) were in V 5+ compounds were assumed to be diamagnetic, but their chara low us to evaluate the diamagnetic correction that should be applied the magnetization of V 4+ compounds. Indeed, this evaluation by Pas delicate in the case of POMs, as is the case for V 4+ POMs, for which w for a molar mass of the order of 4900 g.mol −1 and an important numbe M vs. B at 2 K plots are shown in Figure 9, and χmT vs. T curves (w susceptibility) in the range of 250 to 2 K (B = 1 T) in Figure 10. Both co a weak paramagnetic behaviour that may have been due to V 4+ impu impurities could be deduced from the M vs. B curves at 2 K. For an i the value of the magnetization expected at 5 T was 0.93 μB (black cur for α2-AsV 5+ the magnetization was 0.0115 μB and 0.0313 for α2-PV 5+ 1.2% and 3.3% of impurities, respectively. For α2-AsV 5+ and α2-PV 5+ , the χmT decreased linearly from 0.0 cm 3 .mol −1 .K, respectively) at 250 K to about 0.009 cm 3 mol −1 K (0.015) at that the impurity obeys Curie's law, the slope of this straight line ca ence of temperature-independent paramagnetism (TIP) relative to th can be written: χMT = xiC + χTIP.T, where C stands for the Curie const impurity, and χTIP the TIP. The value obtained for the TIP was 115 × AsV 5+ and 16 × 10 −6 cm 3 mol −1 for α2-PV 5+ . TIP with similar values co the V 4+ compounds.

Magnetic Properties
The magnetic behaviours of the four V 4+ compounds (α1-As V 4+ , α2-As and α2-P V 4+ ) and two V 5+ compounds (α2-AsV 5+ and α2-PV 5+ ) were investiga V 5+ compounds were assumed to be diamagnetic, but their characterisa low us to evaluate the diamagnetic correction that should be applied for the the magnetization of V 4+ compounds. Indeed, this evaluation by Pascal's tab delicate in the case of POMs, as is the case for V 4+ POMs, for which we had o for a molar mass of the order of 4900 g.mol −1 and an important number of wat M vs. B at 2 K plots are shown in Figure 9, and χmT vs. T curves (where χ susceptibility) in the range of 250 to 2 K (B = 1 T) in Figure 10. Both compoun a weak paramagnetic behaviour that may have been due to V 4+ impurities. T impurities could be deduced from the M vs. B curves at 2 K. For an ion S = 1 the value of the magnetization expected at 5 T was 0.93 μB (black curve in Fi for α2-AsV 5+ the magnetization was 0.0115 μB and 0.0313 for α2-PV 5+ , repre 1.2% and 3.3% of impurities, respectively. For α2-AsV 5+ and α2-PV 5+ , the χmT decreased linearly from 0.035 cm 3 . cm 3 .mol −1 .K, respectively) at 250 K to about 0.009 cm 3 mol −1 K (0.015) at 20 K. If that the impurity obeys Curie's law, the slope of this straight line can be du ence of temperature-independent paramagnetism (TIP) relative to the POM can be written: χMT = xiC + χTIP.T, where C stands for the Curie constant, xi t impurity, and χTIP the TIP. The value obtained for the TIP was 115 × 10 −6 cm AsV 5+ and 16 × 10 −6 cm 3 mol −1 for α2-PV 5+ . TIP with similar values could be the V 4+ compounds. ), and 2 K ( Magnetochemistry 2021, 7, x FOR PEER REVIEW 12 of 16 V 4+ ions were isolated, and that the decrease or the increase observed in χmT in the range 20-250 K was not the result of a magnetic coupling.  Table 3. For the AsV 4+ compound, the g values obtained, even if weak, were in agreement with the values found in the literature [56,57], but those obtained for the PV 4+ compound were very weak, confirming the possible underestimation of the molar mass for these compounds (Table 3).
Despite the experimental difficulties related to the weak magnetization of these POMs (a spin of 1/2 in a large diamagnetic matrix), it was confirmed that the V 4+ -containing compounds behaved as isolated S = 1/2 monomers, while the V 5+ -containing ones were globally diamagnetic.
); Brillouin function (-) for a S = 1/2 ion. The g values are given in Table 3.
For the AsV 4+ compound, the g values obtained, even if weak, were in agreement with the values found in the literature [56,57], but those obtained for the PV 4+ compound were very weak, confirming the possible underestimation of the molar mass for these compounds (Table 3).
Despite the experimental difficulties related to the weak magnetization of these POMs (a spin of 1/2 in a large diamagnetic matrix), it was confirmed that the V 4+ -containing compounds behaved as isolated S = 1/2 monomers, while the V 5+ -containing ones were globally diamagnetic.

Conclusions
The redox behaviour of four POMs from the Dawson family was studied in detail using cyclic voltammetry. New observations were made regarding the influence of the pH of the electrolyte on the nature of electron transfer and the evolution of the associated redox potentials. The position of the V substituent within the POM structure, α 1 or α 2 , had a very marked influence on the behaviour of the redox waves associated with the tungstic skeleton. The low-temperature magnetic behaviour of these four compounds containing a V IV centre was compared to that of their counterparts containing a V V centre, which confirmed a paramagnetic behaviour for the former and a diamagnetic behaviour for the latter.