Breathing Some New Life into an Old Topic: Chalcogen-Nitrogen π-Heterocycles as Electron Acceptors †

Recent progress in the design, synthesis and characterization of chalcogen-nitrogen π-heterocycles, mostly 1,2,5-chalcogenadiazoles (chalcogen: S, Se and Te) and their fused derivatives, possessing positive electron affinity is discussed together with their use in preparation of charge-transfer complexes and radical-anion salts—candidate building blocks of molecule-based electrical and magnetic functional materials.


Introduction
Despite rapid progress in the design, synthesis and characterization of molecule-based functional materials for electronics and spintronics, particularly (photo-) conducting, superconducting and magnetic ones, there is a permanent demand for new building blocks in the field (the relevant literature is too abundant to be cited completely, for selected recent references see [1][2][3][4][5][6][7][8][9][10][11][12][13][14]). Especially interesting OPEN ACCESS possibilities are associated with chalcogen-nitrogen chemistry [15][16][17][18] best known in materials science with polymeric sulfur nitride (SN) x [19]-the first macromolecular metal and low-temperature superconductor [20,21]. Another significant impact of this chemistry in materials science is that a large number of candidate building blocks-the spin and/or charge carriers for molecular functional materials were found amongst chalcogen-nitrogen π-heterocyclic radicals, both neutral and positively charged (i.e., radical cations) .
The best studied systems are sulfur-nitrogen, or thiazyl, π-heterocyclic radicals and radical cations. Their common structural feature is the SN fragment, i.e., the same moiety that polymeric sulfur nitride (SN) x is composed of. The free radical SN • (congener of nitric oxide) has never been detected by EPR spectroscopy in its ground 2 Π 1/2 state but the EPR spectrum of the thermally accessible excited 2 Π 3/2 state has been recorded instead in a gas phase. In condensed phases, this radical seems to be undetectable by EPR spectroscopy due to instability and/or strong g-anisotropy [45][46][47].
The diversity of properties displayed by potential materials based on the open-shell thiazyl species covering magnetic ordering, magnetic bistability and electric conductivity is not observed for any other class of compounds. Particularly, crystalline samples of the neutral thiazyl radicals and their Se congeners revealed spin-canted antiferromagnetism, as well as electrical conductivity in combination with spin-canted antiferromagnetism or ferromagnetism ( [48][49][50] and references therein). In the solid state, some π-heterocyclic thiazyl radicals demonstrate spin-state crossover accompanied by magnetic hysteresis. The bistability arises from the coexistence over a temperature range of two-solid state phases, one based on paramagnetic radicals, and the other on their weakly bonded diamagnetic π-dimers. It is believed that magnetically bistable materials can find applications in molecular spintronic devices, i.e., in magneto-thermal switching and information storage devices [30,34,[51][52][53].
The Se congeners of π-heterocyclic thiazyl radicals and radical cations are much less studied whereas Te congeners are unknown [22,24,[48][49][50]54]. This is due to well-known experimental difficulties associated with (organo)tellurium chemistry (see, for example, [54] and references therein). Conceptually, variation of chalcogens in the chalcogen-nitrogen π-heterocycles does not involve any serious challenge: S, Se and Te substitute each other isostructurally and isoelectronically, and 2,1,3-benzochalcogenadiazoles give representative example [55]. On the other hand, heavier-chalcogen congeners of open-shell thiazyl compounds are of enhanced interest especially in the context of the heavy-atom effect on electrical conductivity and magnetic properties [25][26][27]48,49].
Recently, methods for isolation of genuine chalcogen-nitrogen π-heterocyclic RAs in the form of thermally stable salts have been reported [70][71][72][73][74][75][76][77][78][79][80]. Structurally, the RAs belonged mostly to the 1,2,5-chalcogenadiazole ring system (Figure 1) but not only. An important feature of 1,2,5-chalcogenadiazoles and many other related π-heterocycles is their positive electron affinity. As a consequence, RAs are more thermodynamically stable than neutral precursors. The RAs reveal no propensity to isomerization or monomolecular decay. These features provide a basis for their isolation in the form of salts despite the obvious kinetic activity of these species, particularly towards atmospheric moisture and oxygen. The aforementioned 1,2,5-chalcogenadiazolidyl salts are the first representatives of a novel class of paramagnetic chemical compounds. Potentially, this class is broad since derivatives of some other chalcogen-nitrogen π-heterocyclic systems can be used as precursors of stable RAs. Due to this, a distinctive feature of this class can be specified as chalcogen-nitrogen π-heterocyclic anion bearing unpaired π-electron, i.e., possessing a spin S = 1/2. No restrictions are imposed on the cation. Persistent RAs of acyclic analogues of 1,2,5-thiadiazoles, i.e., sulfur diimides R-N=S=N-R (R = Alk, Ar), and their complexes with metal (Cr, Mo, W) carbonyls have also been known for a long time from solution EPR experiments [81] but have not been isolated to date.
In this review, we outline recent progress in the design, synthesis and characterization of chalcogen-nitrogen π-heterocycles, mostly 1,2,5-chalcogenadiazoles (chalcogen: S, Se and Te), and their fused derivatives, possessing positive electron affinity and discuss their use in the preparation of RA salts and CT complexes as candidate building blocks for molecule-based magnetic and conducting functional materials. The review deals first of all with the fundamental chemistry and physics of the heterocycles, and not with materials science. All possible applications mentioned in the text await realization of their potential, and in this aspect the field is in its infancy.
In some important aspects, this new field conceptually overlaps with, and continues development of, the chemistry of polysulfur-nitrogen heterocyclic compounds, i.e., compounds with unusually high proportions of sulfur and nitrogen atoms with respect to carbon atoms, originated by Charles W. Rees in the early 1980s [82][83][84][85][86].

Electron Affinity
First electron affinity (EA) [87] and first ionization energy (IE) are fundamental physical properties of molecules which are very important not only for their chemical reactivity but also for many of their applications in the field of functional materials. According to the Koopmans' theorem, first vertical EA and IE are numerically equal to the energies of the LUMO and HOMO (the frontier MOs) taken with the opposite signs, respectively. Basic research of dependence of both EA and IE on molecular composition and structure is still of significant scientific interest. Whereas IE can be only positive, EA can be both positive and negative. Especially interesting is rather rare positive EA.
According to quantum chemical calculations, the EA of many chalcogen-nitrogen π-heterocycles, both known and unknown, is positive, i.e., RAs are thermodynamically more stable than neutral precursors, with the exact numerical results depending on the level of theory [88]. Test calculations of gas-phase adiabatic EA of selected chalcogen-nitrogen π-heterocycles and TCNE at a number of different levels of the theory revealed that resource-economic (U)B3LYP/6-31+G(d) approach performs reasonably well even in comparison with the G3B3 one which accurately reproduced the experimental EA of TCNE. Correction for ZPE hardly affects the results (Table 1) [88,89]. Due to this, compounds of interest, mostly derivatives of 1,2,5-thia-and selenadiazoles presented below, were calculated at the (U)B3LYP/6-31+G(d) level of theory without ZPE correction. For generality, calculations also covered simple 1,2,5-telluradiazoles and some other chalcogen-nitrogen π-heterocycles which do not belong to 1,2,5-chalcogenadiazole ring system. Particularly, ring systems isomeric to 1,2,5-chalcogenadiazole one (chalcogen: S, Se) were taken into account as well as [1,8-c,d][1,2,6]naphthothiadiazine (known to form stable RA [63,65]) and its derivatives (Tables 2-6) [88]. The archetypal 1,2,5-and 1,2,3-thiadiazoles and their Se congeners possess small positive EA and strong electron-withdrawing substituents, e.g., CN are necessary to enlarge it to ca. 2 eV. On the contrary, 1,2,4-and 1,3,4-thiadiazoles and the Se analogue of the latter have negative EA. Electron-withdrawing substituents CF 3 and CN change the sign of EA of these compounds and significantly enlarge its value. With the same substituents R (R = CF 3 , CN), derivatives of 1,2,3-, 1,2,4and 1,3,4-chalcogenadiazole systems are better electron acceptors than those of 1,2,5-chalcogenadiazole system (chalcogen: S, Se), and Se and Te derivatives are better acceptors than S ones (Table 2). Overall, the nature of the chalcogen, isomerism and substitution pattern are important in determining their EA. According to the calculations, an interesting peculiarity is associated with 1,2,3-selenadiazole where on formation of the RA the Se-N bond spontaneously dissociates. Benzo-fused derivatives of 1,2,5-chalcogenadiazoles (i.e., 2,1,3-benzochalcogenadiazoles) reveal EAs enlarged by ca. 0.9 eV as compared with their monocyclic prototypes. Electron-withdrawing F, NO 2 and CF 3 substituents in the carbocycles enlarge the discussed property further ( Table 3).
An interesting trend is that in the isostructural series of 1,2,5-chalcogenadiazoles and their benzo-fused derivatives (chalcogen: S, Se, Te) the molecular EA increases with atomic number of the chalcogen, i.e., from S to Te despite the fact that atomic EA and Allen electronegativity decreases in this sequence as 2.08 (S), 2.02 (Se) and 1.97 (Te) eV, and 2.59 (S), 2.42 (Se) and 2.16 (Te), respectively. As follows from Table 1, this result is not an artifact of the DFT approach.
Within the structural classes, the π-MOs of the discussed heterocycles are isolobal, i.e., their shapes are invariant to the nature of chalcogen atoms [73].
Overall, according to the calculations, broad structural variation of the archetypal chalcogennitrogen π-heterocycles, including substitution, achievable with current synthetic methods should allow control of the EA of compounds of interest.
As with the EA, it is seen ( Table 7) that 1) the −1/0 electrochemical potential of the heterocycles varies in a broad range depending on their structure and composition; and 2) in the isostructural series of 2,1,3-benzooxa-and -chalcogenadiazoles (chalcogen: S, Se) electron-acceptor ability grows with the atomic number of X (X = O, S, Se).
One of the most interesting findings is however stability of RAs of polyfluorinated 2,1,3benzothia(selena)diazoles and their derivatives [56]. In contrast to RAs of polyfluorinated (hetero)aromatics such as pentafluoropyridine and octafluoronaphthalene which are highly unstable in solution at ambient temperature and cannot be detected by conventional EPR, these RAs are long-lived under electrochemical conditions in MeCN, and especially in DMF, at 295 K [56]. This motivates further work towards their isolation in the form of stable salts.
As follows from Tables 2-6, some other chalcogen-nitrogen π-heterocycles are expected to be precursors of stable RAs. Amongst them, 1,2,6-thia(selena)diazines (known to form stable RAs [63,65]) belong to the most promising and are already involved in ongoing research by the authors. For this reason, current progress in their synthesis is also briefly discussed.
The alkynyl units were attached to benzothiadiazole ring by using the Sonogashira coupling of dibromo derivative 7 with alkynes and palladium catalyst [116,117] (Scheme 7).
The Stille coupling was the most frequently used procedure for preparation of mono-and bis-thienylsubstituted thiadiazoles from a number of dibrominated precursors and trialkylstannanes [118][119][120][121]; in the case of pyridothiadiazole 8 only one bromine atom was substituted (Scheme 8).
The Suzuki coupling was found to be useful approach for the synthesis of thienyl derivative 20 from 4-bromo-2,1,3-benzothiadiazole 10, in that case rarely employed 3-thienylboronic acid was involved in the reaction [122] (Scheme 9).

Scheme 9.
Reaction of 4-bromo-2,1,3-benzothiadiazole 10 with 3-thienylboronic acid. To facilitate the Stille coupling reaction, a new approach for iodination of mono-and di-fluoro substituted benzothiadiazoles 21 was proposed [123]. The synthesis was performed successfully by a single-step Barker-Waters procedure in the presence of AgSO 4 and iodine in concentrated sulfuric acid at high temperature, and the target products were isolated in 65% yields (Scheme 10). Reaction of vicinal nitroamines with sulfur monochloride in the presence of organic bases provides a short and convenient synthetic approach to fused 1,2,5-thiadiazoles and their 2-oxides (Scheme 12) [124]. Scheme 12. Synthesis of benzo-fused 1,2,5-thiadiazoles and their 2-oxides from ortho-nitroanilines.

1,2,5-Telluradiazoles
It should be emphasized that the synthetic chemistry of 1,2,5-telluradiazoles is much more difficult experimentally compared with the chemistry of their Se and S congeners ( [97,148] and references therein).

Charge-Transfer Complexes
One of the most well known organic electron donors is tetrathiafulvalene (TTF), IUPAC name: 2,2'-bis(1,3-dithiolydene) [175]. Its 0/+1 electrochemical potential is 0.33 V [18], and the HeI UPS first vertical IE is ~6.8 eV [176,177]. With electron acceptors, TTF, as well as its various derivatives, form numerous CT complexes and radical-ion salts. In the majority of cases, TCNE, TCNQ and their derivatives were used as electron acceptors. Many of compounds prepared display electrical conductivity and superconductivity as well as interesting magnetic properties [4,14,95,175,178].
Initially, the 1,2,5-chalcogenadiazoles (chalcogen: S, Se) and their O congener fused with TCNQ ( Figure 4) were used in the synthesis of CT complexes and radical-ion salts [95,96]. This was motivated by the assumption that the heterocyclic moieties will enforce intermolecular interactions suppressing the metal-insulator transition typical of many TCNQ-based complexes and salts. A series of TCNQ derivatives fused with one 1,2,5-oxa-and-thia(selena)diazole unit (64a-c) was synthesized and their reduction potentials (E p , Table 7) were found to be positive, although for 64b,c the E p values were somewhat lower than that for TCNQ. These heterocyclic TCNQ derivatives gave a series of CT complexes with TTF derivatives [95].
Recently, it was recognized that 1,2,5-chalcogenadiazoles themselves can serve as effective electron acceptors (Section 2). With TTF, it was discovered that [1,2,5]thiadiazolo [3,4-c][1,2,5]thiadiazole 22 and 3,4-dicyano-1,2,5-telluradiazole 35 form EPR-silent CT complexes TTF·22 and TTF·35 2 , respectively, in the latter case despite the fact that the initial molar ratio of components in reaction solution was 1:1 [89]. The structures of both complexes were confirmed by XRD (Figures 5 and 6). The B97-D functional taking into account long-range dispersion interactions was employed in the calculations of the formation of CT complexes TTF·22 and TTF·35 2 . This method is well suited for the description of weakly bonded complexes (e.g., D-A, van der Waals, and H-bonded complexes) [181]. The def2-TZVP basis set with diffuse basis functions and ECP were used for Te [182]. According to the calculations, electron transfer from TTF is 0.24e onto molecule of 22 and 0.39e onto two molecules of 35. For complex TTF·22, the QTAIM calculations identified five bond critical points (BCPs) for C…N, S…N, and S…S bonding between its components. All the BCPs were characterized by low electron density ( BCP ρ = 0.5-1.1 × 10 −2 a.u.) and positive values of the Laplacian ( BCP ρ 2 ∇ = 1.6-2.6 × 10 −2 a.u.), the latter being typical of the closed-shell D…A interactions [89]. Experiments with stronger electron donor TTT featuring 0/+1 electrochemical potential of 0.15 V [183] are in progress [184].
With elemental potassium and in THF, 2,1,3-benzothiadiazole 68 was transformed into RA salt  Bicyclic chalcogenadiazoles 22 and 26 were reduced with thiophenolate. This approach made it possible to obtain RA salts 70 and 71 with diamagnetic cations of early alkali metals (Li + , Na + , K + ) encapsulated into corresponding crown ethers as well as the salt with (Me 2 N) 3 S + cation (Scheme 35). The structures of all salts were confirmed by XRD ( Figure 8) and their paramagnetic character by solid-state and solution EPR in combination with DFT calculations [77,78].   The salts prepared are thermally stable but air-sensitive. Their decomposition, however, is not a fast process. Specially designed heterogeneous hydrolysis of crystalline salt 70b with saturated water vapor at ambient temperature showed that it proceeds quite slowly and results in an unexpected formation of trithionate salt (Scheme 36, Figure 9) [185].  The thiophenolate-based approach has, however, a limited scope since, for example, the anion does not react with 3,4-dicyano-1,2,5-thiadiazole 72 and 2,1,3-benzothiadiazole 68 [186]. On the other hand, the reaction of thiophenolate with 3,4-dicyano-1,2,5-selenadiazole 73 and 3,4-dicyano-1,2,5telluradiazole 35 gave products of its hypercoordination at the Se and Te centers of the heterocycles, respectively (Scheme 37, Figure 10) [156,186]. It should be noted that similar hypercoordination at the Te center of compound 35 was observed for halides X − (X = F, Cl, Br, I) [97,156]. This previously unknown type of reactivity may be rather general for 1,2,5-chalcogenadiazoles containing heavier chalcogens. According to the XRD data, in all studied cases the length of the hypercoordinate bond is ca. 0.3-0.5 Å longer than the sum of corresponding covalent radii but ca. 1-1.2 Å shorter than the sum of corresponding van der Waals (VdW) radii. For example, in the hypercoordinate product shown in Scheme 36 and Figure 10 (middle) the Se-S distance is 2.722 Å as compared with the sum of the covalent radii of these atoms of 2.25 Å and the sum of their VdW radii of 3.70 Å. For derivatives of Ph-Se-S-Ph with known XRD structures the Se-S bond length varies in the range 2.20-2.22 Å, for neutral derivatives of 3-coordinated Se atom in the range 2.50-2.58 Å, and for derivatives of the 4coordinated Se atom in the range 2.78-3.05 Å ( [186] and references therein). In the hypercoordinate product featuring the Te-S bond (Figure 10, right), its length is 2.688 Å whereas the sums of the corresponding covalent and VdW radii are 2.43 and 3.86 Å, respectively [156].
According to the NBO calculations, in all cases the hypercoordinate bond is formed via negative hyperconjugation of a lone-pair orbital of X − (X = PhS, Hal) with antibonding σ*-MO of the chalcogen-nitrogen bond of heterocycle. This description agrees with the Alcock model suggested previously for secondary bonding interactions between atoms of heavy p-block elements and atoms with electron lone pairs [97,156,186].
The dichotomy between reduction to RA and hypercoordination to heavier chalcogen atom should be taken into account in further work in the field.  Interaction of compound 22 with TDAE gave homospin RA salt [TDAE] [22] 2 (Scheme 38) which was EPR-active in solution but silent in the solid state. According to XRD data (Figure 11), in the crystal the RAs form centrosymmetric π-dimers featuring interplanar separation of 3.25 Å whereas the sum of the VdW radii of two S atoms is 3.60 Å. These diamagnetic dimers are stable only in the solid state and dissociate in both the solution (EPR spectroscopy) and gas phase (DFT calculations) [76].   The heterospin, S 1 = S 2 = 1/2 and S 1 = 3/2, S 2 = 1/2, RA salts of 1,2,5-chalcogenadiazoles 22 and 75 were prepared with reducing agents bis(toluene)chromium(0) (CrTol 2 ) and decamethylchromocene (CrCp* 2 ), respectively, as [CrTol 2 ] [ Figure 13). Whereas salts 76 and 77 are EPR-active in both the solid-state and solution, salt 78 is EPR-silent in these states of aggregation due to the substantial zero-field splitting and fast relaxation of the cation provoking fast relaxation of the RA [71,73]. In solution the salts exist most likely in the form of ion pairs which is typical of the discussed RAs [57][58][59][60][61][62][63][64].    (80), respectively, were isolated [99]. Analysis of the XRD structure of the solvate 79·3CH 2 Cl 2 confirmed the complete CT with formation of the RA of 66 (Scheme 41). The EPR spectrum of the latter was recorded for CH 2 Cl 2 solution of salt 79. . In the latter, the unpaired electron and the negative charge are delocalized equally across two molecules of 66 [99].
Very recently [80], a new type of sulfur-nitrogen π-heterocycles represented by (6H-1,2,3benzodithiazol-6-ylidene)malononitrile 81 was electrochemically and chemically with TDAE and CrTol 2 (Scheme 42) reduced into the RA. Electrochemically generated RA and its salts [TDAE] [81] 2 (82) and [CrTol 2 ] [81] (83) were characterized by solution EPR. According to the EPR data, salt 82 is homospin S = 1/2 paramagnet in both the solid state and MeCN solution. Solid-state paramagnetism of salt 82 indicates that RAs do not form diamagnetic π-dimers. Salt 83 is heterospin, S 1 = S 2 = 1/2, since both paramagnetic ions were detected in its THF solution [80]. Electronic structure and EPR spectra of RAs obtained by electrochemical and chemical reduction of chalcogen-nitrogen π-heterocycles were analyzed using results of the DFT, HF and, in some cases, post-HF methods [70][71][72][73][74][75][76][77][78][79][80]. For all RAs under study, the SOMO is a π-type MO. Figure 14 displays the SOMO of the typical representative, i.e., RA of compound 22. This π-SOMO is antibonding for bonds SN and nonbonding for bonds CN. Multi-configuration calculations by the CASSCF method revealed an unusual electronic structure of π-dimers of RAs for the salt [TDAE] [22] 2 (see Figure 11). It was found that the major contribution (~80%) to the wave function of the singlet ground state of the dimer comes from the electronic configuration with two electrons occupying the bonding MO composed of weakly interacting SOMOs of the RAs. However, the contribution of the biradical component to this state is rather large (~20%) [76].
The DFT-calculated spin density distribution in the RAs studied shows that the density on their VdW surfaces is mainly positive with only small islands of negative values in the vicinity of the C-C bond of 1,2,5-chalcogenadiazole ring (for example, see Figure 15) [56,73]. Therefore, contacts between the like spin density regions are most probable for neighboring RAs in the crystals of homospin salts. This should lead to antiferromagnetic (AF) exchange interactions between RAs. If two paramagnetic species in heterospin salts mainly contact each other in the regions of unlike spin density, ferromagnetic (FM) interactions are possible [187][188][189][190][191][192][193][194].

Charge-Transfer Complexes
The standard two-contact method was used to measure the single crystals resistivity of the CT complexes TTF·22 and TTF·35 2 . The temperature dependence of resistance was recorded in the temperature range 300-320 K ( Figure 16). As follows from Figure 16, both complexes revealed low-gap semiconductor behavior with activation energy ~0.34 eV for TTF·22 and 0.40 eV for TTF·35 2 . For both CT compounds, the absolute conductivity showed noticeable increase upon white-light irradiation (0.5-1.5 sun intensity) which is congruent with the D-A system's ability to efficiently separate the generated charges. Although the relative increase in conductivity was small, this was attributed to the comparatively large dark currents arising from these low-gap semiconductors [89].

Radical-Anion Salts
Molar magnetic susceptibilities (χ) of a series of RA salts under discussion were measured in a wide temperature range 2-300 K. For the homospin salt [TDAE] [22] 2 characterized by centrosymmetric π-dimers of the RAs in the XRD structure, the magnetic susceptibility is temperatureindependent and equal to zero after subtracting the diamagnetic contribution [76].
Experimental χ(T) dependences for RA salts 70b,c with alkali metal cations were analyzed using phenomenological analytical temperature dependences [78]. For both salts, AF exchange interactions were observed, with low Neel temperatures (Table 8). For salt 70c, the Bonner-Fisher uniform chain model [190] perfectly describes the χ(T) in the whole temperature range ( Figure 17A). For salt 70b, the alternating chain model [190] led to reasonable agreement with experiment. Experimentally determined parameters of the AF exchange interactions for these and some other homospin RA salts are presented in Table 8. To support application of the Bonner-Fisher uniform chain model in the case of salt 70c, calculations of the exchange interactions for the pairs of neighboring RAs were performed. Five unique pairs of first-nearest neighbors were found in the crystal structure of salt 70c [77]. Spinunrestricted broken-symmetry approach [191][192][193][194][195] at HF, MP2 and B3LYP levels of theory was employed for the calculations. The calculated J value for one pair of RAs of 22 was at least one order of magnitude larger than other J values. Moreover, calculations supported the assumption that in the salt 70c the exchanged-coupled RAs form uniform chains. In addition, the data of Table 8 demonstrate that the J value calculated at the UB3LYP/TZVP level (−1.62 cm −1 ) is in very good agreement with experimentally determined value −1.22 cm −1 (the accuracy is about 30%).
The case of salt 71 is rather complicated since its structure is disordered. There are three different mutual orientations of the RAs of 26 in the selected pairs; the J value strongly depends on the orientation and changes from −8.08 cm −1 for pair b to −1.69 cm −1 for pair c (Table 8, Figure 18). Taking into account statistics of contacts, the average value is predicted to be −3.87 cm −1 . The latter value is about 80% larger than J obtained for salt 70c. Calculations also suggested that exchangecoupled RAs of salt 71 form infinite chains. However, these chains are characterized by three statistically distributed J values. Nevertheless, the Bonner-Fisher uniform chain model very well describes the χ(T) of salt 71 in the whole temperature range ( Figure 17B). The experimentally obtained J value is about 30% larger than that for salt 70c. This observation agrees qualitatively with the difference in the calculated J values [77].  For RA salt 74 ([CoCp 2 ] [22]), the temperature dependence of its magnetic susceptibility has a maximum at 9.7 ± 0.5 K indicating AF ordering of the spin system ( Figure 19A). The magnetic motif of salt 74 was also analyzed in terms of the pair exchange integrals (J) calculated similar to the case of 70c and 71. The magnetic motif was found to be two-dimensional (2D) and close to the S = 1/2 square lattice AF Heisenberg model ( Figure 20). The analysis of the χ(T) dependence was performed using the low-and high-temperature series expansions available for this model [196][197][198][199]. The exchange interactions between RAs of 22 (with S…S distance r1, Figure 20) were estimated to be J = −4.2 ± 0.7 cm −1 . With the general form of the Van Vleck equation [187,195,200]    Amongst investigated homospin RA salts, most complicated magnetic motif was predicted for 69 ([K(THF)] [68]) [74,88]. Neglecting small J values (J < 0.2 cm −1 ) leads to the 2D magnetic motif with both FM and AF interactions (Table 8). A reasonable agreement between experimental and theoretical temperature dependences was achieved with J values which are one and a half times higher than calculated ones (Table 8; Figure 19, C and D).
For heterospin salt 78, analysis of the experimental χT temperature dependence ( Figure 21A) based on the CASSCF and DFT calculations revealed only AF interactions: significant between RAs of 22 (−40 ± 9 cm −1 ), weak between [CrCp* 2 ] + cations (−0.58 ± 0.03 cm −1 ), and very weak between the RAs and cations. Thus, magnetic motif of this salt is very simple and consists of the AF-coupled pairs of RAs and AF-coupled pairs of cations. Experimentally determined J values are in good agreement with calculations (Table 9). For salts 76, calculations predict rather simple 2D magnetic motif ( Figure 22) with two types of AF interactions: between RAs (J 1 ) and between RAs and cations (J 2 ). The theoretical modeling of the χT temperature dependence was based on this magnetic motif and led to very good agreement with experiment ( Figure 21B). Experimentally determined J values were in reasonable agreement with experiment (Table 9). The magnetic motif of the salt 77 is much more complex, mainly due to a structural disorder of the RAs of 75. This 3D motif can be characterized by at least seven J parameters of both signs (FM and AF interactions, Table 9). To simulate correctly the χT temperature dependence of this material is intractable problem. However, even oversimplified model composed of two types of exchange-coupled pairs (J 1 = 9.72 and J 2 = −7.96 cm −1 ) perfectly describes experiment ( Figure 21C). This is in qualitative agreement with results of calculations predicting both FM and AF interactions in the anion…cation pairs (Table 9).

Summary and Future Perspectives
Available synthetic methods allow preparation of various 1,2,5-chalcogenadiazoles (chalcogen: S, Se, Te) and their numerous functional derivatives, particularly areno-and hetareno-fused derivatives. Overall, molecular diversity is very broad in the field and can further be broadened by means of DFT calculations-aided molecular design.
Derivatives of the 1,2,5-chalcogenadiazole ring system possess positive EA which depends on molecular composition and structure and can be enlarged by electron-withdrawing substituents, i.e., controlled. With various reducing agents, the compounds can be easily transformed into persistent RAs and the latter isolated in the form of thermally stable crystalline salts. The salts, both homospin (where only anions were paramagnetic) and heterospin (where both ions were paramagnetic) revealed mainly AF interactions in their spin systems. It should be noted that currently systems with AF interactions are receiving increased attention because of the experimental observation of the spin-liquid state, as well as their prospects in creating nanoscale memory cells [201,202].
With the McConnell I model dealing with spin polarization [187] dominance of AF interactions is expected for the homospin salts. In these salts, the spin density on the VdW surfaces of their RAs is mostly positive, with only small islands of negative spin density. For neighboring RAs in the crystal lattice, contacts of like spin density are most probable to give rise to AF exchange interactions between them, whereas for FM interactions, contacts of unlike spin density are required (except contacts of unlike density on orthogonal MOs leading to AF interactions). In general, the heterospin salts with paramagnetic sandwich cations possessing peripheral negative spin density on ligands are better suited for FM interactions and, therefore, are worth of further investigations.
Especially promising are heterospin salts with S = 1/2 cations [MAr 2 ] + (M = Cr, Mo, W). The IE of their precursors MAr 2 (reducing agents in the target salts' preparations) can be varied in a rather broad range depending on the ring substituents. With the same Ar ligands, the IE is practically equal for M = Cr (3d), Mo (4d), and W (5d) allowing one to cover the whole d block in a single approach. The cations with Mo and W are of special interest because of the strong spin-orbit coupling inherent in these heavy atoms. The strength of the spin-orbit coupling increases sharply with the atomic number as Z 4 to be sufficient for atoms with Z > 30. In the heterospin salts containing Mo (Z = 42) or W (Z = 74) atoms in the cation and heavier chalcogen Se (Z = 34) or Te (Z = 52) atoms in the anion, the strong spin-orbit coupling can lead to spin canting to originate a FM ground state even under conditions of AF exchange interactions between paramagnetic centers (the Dzyaloshinsky−Moriya mechanism [187]). Successful experiments with MoAr 2 reducing agents are already in progress.
Additionally to 1,2,5-chalcogenadiazoles, there are many other chalcogen-nitrogen π-heterocycles partially represented in Tables 2-6 and expected to be precursors of persistent RAs which can be isolated in the form of thermally stable salts, or effective electron acceptors in synthesis of new CT complexes. Therefore, one may hope that the discussed field of the chalcogen-nitrogen chemistry is very far from being exhausted one.