Non-Innocent Base Properties of 3- and 4-Pyridyl-Dithia- and Diselenadiazolyl Radicals; the Effect of N-Methylation

Condensation of persilylated nicotinimideamide and isonicotinimideamide with sulfur monochloride affords double salts of the 3-, 4-pyridyl-substituted 1,2,3,5-dithiadiazolylium DTDA cations of the general formula [3-, 4-pyDTDA][Cl][HCl] in which the pyridyl nitrogen serves as a noninnocent base. Reduction of these salts with triphenylantimony followed by deprotonation of the intermediate-protonated radical affords the free base radicals [3-, 4-pyDTDA], the crystal structures of which, along with those of their diselenadiazolyl analogues [3-, 4-pyDSDA], have been characterized by powder or single-crystal X-ray diffraction. The crystal structures consist of "pancake" π-dimers linked head-to-tail into ribbonlike arrays by η2-S2---N(py) intermolecular secondary bonding interactions. Methylation of the persilylated (iso)nicotinimide-amides prior to condensation with sulfur monochloride leads to N-methylated double chloride salts Me[3-, 4-pyDTDA][Cl]2, which can be converted by metathesis into the corresponding triflates Me[3-, 4-pyDTDA][OTf]2 and then reduced to the N-methylated radical triflates Me[3-, 4-pyDTDA][OTf]. The crystal structures of both the N-methylated double triflate and radical triflate salts have been determined by single-crystal X-ray diffraction. The latter consist of trans-cofacial π-dimers strongly ion-paired with triflate anions. Variable temperature magnetic susceptibility measurements on both the neutral and radical ion dimers indicate that they are diamagnetic over the temperature range 2-300 K.


Introduction
1,2,3,5-Dithiadiazolyl (DTDA) radicals have been widely studied for many years, 1 initially with a view to their utilization as building blocks for molecular magnetic materials, 2 molecular semiconductors and conductive charge transfer salts [DTDA] + [X] -. 3 More recently they have also been explored as ligands to dand f-block metals, 4 fluorophores in solution and polymer composites, 5 external field-driven NLO switches, 6 and paramagnetic guests in porous 7 and nonporous host frameworks. 8 In the absence of steric protection, DTDA radicals associate in the solid state, forming nominally diamagnetic dimers which display a partial or complete loss of their S = ½ magnetic signature. Selenium-based (DSDA) analogues have also been investigated, and in these systems the tendency to dimerize is even stronger, although, as with DTDAs, 9 steric effects can impede dimerization. 10 Pairwise association of the radicals by overlap of their SOMOs can occur in a variety of ways, some of which are shown in Chart 1. Of these, by far the most common is the cis-cofacial mode, which involves the direct superposition of two radicals separated by a pair of 4-center 2-electron 11 S---S (or Se---Se) 12 contacts. The nature of these interactions, now commonly referred to as "pancake" -bonds, 13 has been the subject of much debate, the focus being on whether the pairing of the singly occupied molecular orbitals (SOMOs) of two radicals affords an open-or closed-shell singlet state.

Chart 1
Considerable synthetic effort has been directed towards the control of the dimerization mode and packing pattern of the resulting pairs, partly with the intent of encouraging -stacked dimer arrays, which historically were viewed as the most effective way to introduce a pathway for charge transport. 1c,3a,14 One approach that has served well has been the use of intermolecular secondary bonding interactions (SBIs) between the disulfide (diselenide) unit of the DTDA (DSDA) ring and a lone-pair carrying heteroatom attached to the 4-substituent. The effect was first demonstrated in the 2-, 3-and 4-cyanophenyl DTDA (DSDA) dimers, 15 and later extended to cyanofuryl 16 and cyanothienyl 17 derivatives. In all these materials the dimers are connected into ribbon-like arrays linked by η 2 -S2---NC SBIs ( Figure 1). The resulting ribbons then pack into layers which, in some cases, afford the desired superimposed -stack arrays, with the DSDA derivatives displaying small bandgap semiconductive properties. 14 From a magnetic perspective a vital modification of this strategy, developed by Rawson and coworkers in the mid-1900's, involved attachment of a cyano (or nitro) group to an otherwise fully fluorinated phenyl substituent on a DTDA radical. 18 The impact was two-fold: (i) the η 2 -S2---NC (or η 2 -S2---O2N) SBIs locked the radicals into ribbons, and (ii) the steric protection afforded by perfluorination of the phenyl group suppressed dimerization. The discovery of these undimerized DTDA radicals and their sometimes startling magnetic properties 19 represents one of the major triumphs in the development of radical-based molecular materials. Exploration of ligand design and the resulting effect on solid state architecture continues, and the range of structure-making SBIs has been expanded. The packing of 2-chlorophenyl substituted DTDA dimers, for example, are dominated by short η 2 -S2---Cl intermolecular contacts which, like η 2 -S2---NC SBIs, are effective in generating layered structures. 20 Likewise the 2-pyridyl-substituted radicals 2-pyDTDA and 2-pyDSDA and related pyrimidyl derivatives developed by Preuss and coworkers serve as chelating (Chart 2) ligands to paramagnetic metal ions. The resulting coordination complexes display a rich array of magnetic properties, ranging from single molecule magnets to magnetically ordered materials and magnetothermal switches. 21 By contrast, the related 3-, 4-pyDTDA radicals 22 have received relatively little attention. A number of metal complexes have been reported, 23,24 but information on the radicals themselves is sparse, and the corresponding 3,4-pyDSDA derivatives are, to date, unknown. The relative dearth of information on these particular radicals may stem in part from the presence of a basic nitrogen on the ligand, which has caused some difficulties in their isolation. To address this issue we provide here full structural characterization of the dimers of the 3-, 4-pyDTDA and 3, 4-pyDSDA derivatives as well as modifications to the synthetic procedures made necessary by the "noninnocent" base properties of the pyridyl ligand. In addition, as a first step in taking advantage of 6 the coordination chemistry of the ligand, we also report the preparation and structural characterization of the N-methyl-pyridinium substituted radical salts Me [OTf] and Me [OTf].

Results and Discussion
Synthesis. Early synthetic routes to the DTDA framework involving the addition reactions of thiazyl halides 25 or sulfur monochloride/ammonium chloride mixtures 26 to nitriles have been largely replaced by methods based on the condensation of an amidine, more particularly a persilylated amidine, 27 with sulfur monochloride or dichloride, to afford an oxidized dithiadiazolylium chloride salt. Reduction of the latter, typically with triphenylantimony, liberates the neutral radical, which may be purified by vacuum sublimation. This methodology, which is easily extended to the preparation of selenium-based analogues, 28 provides the basis for the synthesis of 3-, 4-pyDTDA and 3-, 4-pyDSDA described here.
The necessary persilylated amidines 3-and 4-pyADS were generated by the addition of lithium bis(trimethylsilylamide) to the appropriate nitrile (3-NCpy, 4-NCpy), followed by quenching the intermediate lithiated amidinate with chlorotrimethylsilane (Scheme 1A). Typically, such reactions employ diethyl ether or toluene/ether mixtures as the reaction solvent, but for these pyridyl-substituted nitriles the use of the more strongly basic solvent tetrahydrofuran is critical in order to suppress solvation of Li + ions with the pyridyl nitrogen and to keep the otherwise insoluble lithiated intermediate in solution. Consistently, the subsequent silylation step can only 7 be effected in tetrahydrofuran at reflux. The resulting persilylated amidines may then be purified by fractional distillation in vacuo.
Most silylated amidines undergo a smooth condensation with excess sulfur mono-or dichloride to afford directly the appropriate dithiadiazolylium chloride salt [DTDA] [Cl] as an insoluble yellow-orange solid. However, for both 3-and 4-pyDTDA this general procedure requires modifications arising from the fact that the initial condensation reaction affords not the simple salt but its highly moisture-sensitive hydrochloride double salt, that is, [3, 4-pyDTDA] [Cl][HCl], in which the pyridyl nitrogen is protonated (Scheme 1B). By contrast, protonated intermediates are not observed in the preparation of the analogous perfluorinated derivative 4-(F4-py)-DTDA, as a result of the weaker base strength of the perfluoropyridyl moiety. 29 Reduction of these double salts with triphenylantimony in MeCN then affords the corresponding hydrochloride salts of the radical, that is, [3-, 4-pyDTDA] [HCl]. Finally, rinsing these insoluble hydrochloride salts with a solution of excess triethylamine in MeCN liberates the free-base radicals 3-, 4-pyDTDA, which may be purified by fractional sublimation in vacuo. The sequence of events may be conveniently tracked by infrared spectroscopy, as illustrated in Figure 2 (see also Figure S1) which shows the spectral changes observed during the preparation of 4-pyDTDA.

Scheme 1
To explore more fully the stepwise nature of the synthesis of 3-, 4-DTDA, we pursued the Nmethylated analogues of the protonated intermediates described above, with a view to obtaining structural information. To this end we prepared the N-methylated amidinium triflate salts Me Preparation of the selenium-based radicals 3-, 4-pyDSDA was more straightforward than that described above for 3-, 4-pyDTDA, possibly because the condensation reaction employed stoichiometric amounts of SeCl2 generated by comproportionation of Se and SeCl4, 28 as a result of which the presence of adventitious HCl (present in S2Cl2) was all but eliminated. Accordingly, the simple chlorides [3-, 4-pyDSDA] [Cl] were generated directly during the condensation (Scheme 2), and reduction of these salts with triphenylantimony then released the corresponding radical dimers, which could be purified by sublimation in vacuo.

Scheme 2
Crystallography on Radical Dimers. Analytically pure samples of the dimers of the four radicals 3-, 4-pyDTDA and 3-, 4-pyDSDA were generated by fractional sublimation in vacuo.
Crystals of 3-pyDTDA and 3-, 4-pyDSDA so obtained were suitable for structural analysis by single crystal X-ray diffraction methods. Crystal data are summarized in Table S1 and ORTEP drawings of the dimer units are shown in Figure 3. The almost spherical, powdery nodules of 4-pyDTDA obtained by sublimation did not diffract as single crystals, but did generate a reproducible powder X-ray diffraction (PXRD) pattern consisting of a few sharp peaks and some broad "humps" (Figure 4), an appearance characteristic of a disordered nanocrystalline phase. 30 Material obtained by recrystallization from odichlorobenzene afforded a similar pattern. Changes in sample preparation, such as sample grinding, did not lead to a marked change in linewidth or resolution of the diffractogram.

11
Attempts to index the PXRD data collected on both laboratory and synchrotron X-ray sources with conventional indexing programs (DICVOL, McMaille, Topas) were unsuccessful, mainly because of the paucity and poor resolution of some of the diffraction peaks, which hindered assignment of a space group and specification of a unit cell. Eventually, however, by comparing the observed diffraction pattern with that predicted for the selenium analogue 4-pyDSDA we were able to index the data manually to the C2/c space group and identify a unit cell similar to that observed for both 4-pyDSDA and its perfluorinated variant 4-(F4-py)DTDA. 29,31 Figure 4. Observed and calculated PXRD pattern for cis-cofacial 4-pyDTDA (λ = 0.825925 Å).
The PXRD data was subsequently modelled in DASH using synthetic annealing methods based on a molecular unit adapted from the known molecular coordinates of 4-cyanophenyl-DTDA. 15 This procedure afforded a plausible solution based on cis-cofacial dimers ( Figure 3c) linked head-to-tail into antiparallel chains in a fashion very similar to that found for 4-pyDSDA. This solution was refined by Rietveld methods using a rigid-body constraint in which only the unit cell parameters were optimized; the resulting crystal data are provided in Table S2 and the observed/calculated PXRD patterns are presented in Figure 4. Two other solutions with the same cell parameters but based on trans-cofacial and trans-antarafacial dimers were similarly refined.
The resulting (optimized) crystal data are provided in Table S2 and

Crystals of the N-methylated double triflates Me[3-, 4-pyDTDA][OTf]2 and radical ion triflates
Me [OTf] obtained by crystallization from MeCN were all suitable for structural characterization by single crystal X-ray diffraction. Crystal data for these materials are summarized in Table S3, and ORTEP drawings with intramolecular metrics are provided in    . Data were collected in cooling mode using a static field H of 0.1 T over the range shown 2-300 K. As may be seen in Figure 11, the resulting plots of  (corrected for diamagnetic contributions) versus T plots are remarkably similar. Both the neutral and charged radicals show essentially diamagnetic behavior, with a concentration of free spin defect impurities around the 1% level, which is as expected for DTDA dimers. There is no evidence for thermally induced dissociation up to 300 K, the limit of the experiment.

Summary and Conclusions
The chemistry and structural properties of DTDA and DSDA radicals have been extensively studied for many years. A wide range of 4-substituents have been explored, with increasing interest in the introduction of ligands that can enhance or complement the coordination chemistry of the radical itself. 21,23,24 In principle a 3-or 4-pyridyl ligand provides an effective secondary binding site for metal ions, but the basic nature of these ligands also introduces subtleties in the preparation of the radicals to which they are bound. Here we have provided a detailed account of the preparation of 3-, 4-pyDTDA and their Se-based analogues, using procedures that reflect the intermediacy of protonated radicals. Treatment of the protonated intermediates with base allows for isolation and structural characterization of the native radicals. In the solid state 3, 4-pyDTDA and 3, 4-pyDSDA crystallize as diamagnetic dimers linked head-to-tail into ribbon-like arrays by short η 2 -S2---N(py) and η 2 -Se2---N(py) intermolecular secondary bonding interactions. In the case of 4-pyDTDA disorder in the packing of these molecular ribbons leads to a nanocrystalline morphology.
As a first step in exploiting the basic properties of the 3-, 4-pyridyl ligands we have prepared the N-methylated radical triflates Me [OTf], the crystal structures of which consist of diamagnetic trans-cofacial dimers strongly ion-paired with triflate anions. Development of the charge transfer chemistry of these latter salts opens the door to the development of new functional molecular materials with potentially interesting magnetic and/or charge transport properties.

Experimental Section
General Methods and Procedures. The starting materials 3-and 4-cyanopyridine, benzyltriethylammonium chloride, lithium bis(trimethylsilylamide), chlorotrimethylsilane, methyl trifluoromethanesulfonate (methyl triflate), silver trifluoromethanesulfonate (silver triflate), sulfur monochloride, potassium iodide, and triphenylantimony were obtained commercially; all were used as received, save for KI, which was dried in vacuo. Selenium tetrachloride was prepared following the literature procedure. 33 Reagent grade acetonitrile (MeCN) was dried by distillation from P2O5 and CaH2, anhydrous grade tetrahydrofuran (THF) was dried by distillation from sodium/benzophenone and diethyl ether was dried over 4 Å molecular sieves. Unless otherwise specified, all reactions and synthetic procedures were carried out under an atmosphere of nitrogen or argon. Fractional sublimations of all radical dimers were performed in an ATS series 3210 three-zone tube furnace, mounted horizontally and linked to a series 1400 temperature control system. Infrared spectra (Nujol mulls, KBr optics) were recorded on a Nicolet Avatar FTIR spectrometer or Bruker Alpha with Platinum ATR module with a diamond IR transmitting crystal at 2 cm -1 resolution. NMR spectra were recorded on a Bruker Avance III HD 300 MHz spectrometer using anhydrous deuterated solvents. Elemental analyses 20 were performed in-house on an Elementar Vario EL III elemental analyzer or by MHW Laboratories, Phoenix, AZ 85018.  35;H,9.25;N,12.44. Found: C,53.40;H,9.28;N,12.22. -3-yl)-1,2,3,5-dithiadiazolyl, 3-pyDTDA. Following the procedure described above for 4-pyDTDA, sulfur monochloride (13.5 g, 0.100 mol) was added to a solution of 3-pyADS (1.70 g, 5.04 mmol) in 50 mL MeCN. The resulting slurry was heated at reflux for 2 h, then cooled to room temperature, and the orange precipitate of the crude double salt solid [3-pyDTDA] [Cl][HCl] was filtered off, washed with 3  20 mL MeCN and dried in vacuo, yield 1.21 g (4.74 mmol). This crude material was added to 50 mL MeCN, along with solid Ph3Sb  39.54;H,2.21;N,23.06. Found: C,39.36;H,2.32;N,23.17.
The reaction was allowed to warm to room temperature and stirred for 16 h to afford a yellow solution. The solvent was flash distilled to leave a white solid, which was washed with 20 mL toluene and the fine white solid collected by filtration and dried in vacuo; yield 5.66 g (11.28 mmol, 91 %). 1 1.42;N,8.48. Found: C,22.81;H,1.66;N,8.70.  21.82;H,1.42;N,8.48. Found: C,22.13;H,1.52;N,8.77.  26.11;H,1.46;N,15.22. Found: C,26.24;H,1.50;N,15.20. EPR Spectroscopy. X-Band EPR spectra for 3-and 4-pyDTDA were recorded at ambient temperature using a   Agilent SupraNova diffractometer equipped with multilayer optics monochromated dual source (Cu and Mo) and Atlas detector, using CuKα ( = 1.54184 Å) radiation; data acquisitions, reductions and analytical face-index based absorption corrections were made using the program CrysAlisPRO. 39 For 3-, 4-PyDSDA X-ray data were collected at 296 K using ω scans with a Bruker APEX II CCD detector on a D8 three-circle goniometer and Mo Kα (λ = 0.71073 Å) radiation. In all cases, the structures were solved using the ShelXS 40 program and refined on F 2 by full matrix least squares techniques with the ShelXL8 program using the interface in the Olex2 (v.1.2) program package. 41 Powder Crystallography. Powdered samples of sublimed 4-pyDTDA were loaded into borosilicate glass capillaries and sealed. Data were collected at 293 K using synchrotron radiation (λ = 0.825925 Å) available on beamline I11 at the Diamond Light Source. As a result of the paucity and poor resolution of some of the observed reflections attempts to index the data using DICVOL, 42 McMaille 43 and Topas 44 were unsuccessful. However, starting from the cell settings and space group (C2/c) of 4-pyDSDA manual indexing of the data was eventually achieved. A structure search was then performed in DASH 45 using simulated annealing methods and a molecular unit based on the atomic coordinates of 4-cyanophenyl-DTDA. 15 A rigid body constraint was initially employed, followed by release of the ring-to-ring (py-DTDA) torsion angle. A solution was found which closely matched the cis-cofacial mode of dimerization and packing observed in 4-pyDSDA (Figures 3 and 6 Table S2. The deposited CIF file for 4-pyDTDA corresponds to the cis-cofacial dimer;

Preparation of Me
for completenesss the refined crystal coordinates for the other two models are available in Tables   S4 and S5.
Magnetic Susceptibility Measurements. DC magnetic susceptibility measurements were performed at a field of 1000 Oe over the temperature range 2-300 K on a Quantum Design MPMS SQUID magnetometer. Diamagnetic corrections were made using Pascal's constants. 48