A Qualitative Comparison of the Reactivities of 3,4,4,5-Tetrachloro-4H-1,2,6-thiadiazine and 4,5-Dichloro-1,2,3-dithiazolium Chloride

The high yielding transformations of 3,4,4,5-tetrachloro-4H-1,2,6-thiadiazine into 3,5-dichloro-4H-1,2,6-thiadiazin-4-one (up to 85%) and 2-(3,5-dichloro-4H-1,2,6-thiadiazin-4-ylidene)malononitrile (up to 83%) have been investigated and compared to the analogous transformations of the closely-related 4,5-dichloro-1,2,3-dithiazolium chloride (Appel’s salt) into 4-chloro-5H-1,2,3-dithiazol-5-one and 2-(4-chloro-5H-1,2,3-dithiazol-5-ylidene)malononitrile. Furthermore, cyclocondensation of 3,4,4,5-tetrachloro-4H-1,2,6-thiadiazine with 2-aminophenol and 1,2-benzenediamines gave fused 4H-1,2,6-thiadiazines in 68%–85% yields.

Appel's salt 2, discovered in 1985, is readily prepared from chloroacetonitrile and disulfur dichloride [17] and has since found numerous uses as a scaffold for the synthesis of 1,2,3-dithiazole derivatives [18][19][20][21]. Dithiazolium 2 exists as a salt and not in the covalent form 2′, and it is a planar and 6π aromatic system, although worthy of note is the 5,5-difluoro analogue, which is a non-ionic covalent bound molecule that can be distilled and isolated as an oil [17]. The chemistry of the dithiazolium 2 is governed by the electrophilicity of the C-5 carbon, and it readily reacts with nucleophiles to give neutral 5H-1,2,3-dithiazoles.

Preparation of 3,4,4,5-Tetrachloro
Two methods exist in the literature for the preparation of the tetrachlorothiadiazine 1: the reaction of dichloromalononitrile (6) and SCl2 [15] and the reaction of N-2,2-trichloro-2-cyanoacetimidoyl chloride (7) [24] with elemental sulfur. In our hands, both reactions worked well; however, the second method was preferred, as it gave better yields and avoided the use of the toxic and hard to access SCl2 (Scheme 3). Compound 7, even though it looks very reactive, was easy to prepare and isolate and can be stored for up to six months at 0 °C. Moreover, the two reports on the isolation of tetrachlorothiadiazine 1 gave conflicting distillation data for the product (100 °C, 8 mbar [25] vs. 90 °C, 4 mbar [26] 3,5-Dichloro-4H-1,2,6-thiadiazin-4-one (3) can be prepared from the reaction of glacial formic acid with 3,4,4,5-tetrachloro-4H-1,2,6-thiadiazine (1) [15]. In our hands, this transformation was sensitive to the quality of the formic acid used as both a solvent and reagent. Interestingly, no other chemistry has been reported for tetrachlorothiadiazine 1 apart from this reaction and its degradation in moist air to give 2-chloromalonamide (8) (Scheme 4) [26]. As such, we reinvestigated the transformation of tetrachlorothiadiazine 1 into the thiadiazinone 3 (Section 2.2. Comparing the syntheses of tetrachlorothiadiazine 1 with dithiazolium 2, it was clear that both required similarly hazardous reagents for their preparations. The synthesis of dithiazolium 2, however, was clearly easier, as it required only commercially-available chloroacetonitrile and disulfur dichloride and can be isolated as a salt by simple filtration. On the other hand, the synthesis of the thiadiazine 1 required the use of chlorine gas to prepare the less readily-available dichloromalononitrile and SCl2 reagents, and the product 1 was isolated by vacuum distillation.

Preparation of the Thiadiazinone 3 and Dithiazolone 9
Both the covalent thiadiazine 1 and the ionic 1,2,3-dithiazolium 2 are hydrolytically unstable, reacting with water or even in moist air. However, while the degradation of dithiazolium 2 in moist air gives a brown mass from which the dithiazolone 9 can be isolated by sublimation [17], the analogous degradation of tetrachlorothiadiazine 1 leads to ring cleavage to afford 2-chloromalonamide (8) (Scheme 4) [26].
Several improved methods are known for the transformation of Appel's salt 2 to dithiazolone 9 [17,27], but only details of the formic acid-induced reaction have been reported for the hydrolysis of the tetrachlorothiadiazine 1 [1,2,15]. As such, we investigated other routes to synthesize the two ketones 3 and 9.
The hydrolytic conditions investigated included treatment with formic acid, acetic acid, sodium and silver nitrates and DMSO ( Table 1). The highest yielding (85%) and cleanest reaction conditions for the hydrolysis of the tetrachlorothiadiazine 1 into the thiadiazinone 3 involved the use of AgNO3 in MeCN (Table 1, Entry 6), while for the hydrolysis of Appel's salt 2 to dithiazolone 9, the cleanest conditions were using neat glacial formic acid ( Table 1, Entry 1; 89%). While most procedures worked equally well for both transformations, significant differences in behavior were also observed.
The reaction of tetrachlorothiadiazine 1 with glacial formic acid (98% purity) gave the thiadiazinone 3 in a reasonable 75% yield, but the use of technical-grade formic acid (85% purity), containing ca. 15% water, led to degradation of the starting material, affording only a trace of 2-chloromalonamide (<1%). In contrast, Appel's salt 2 was less sensitive to the presence of water in the formic acid and with the respective reactions gave dithiazolone 9 in 89% and 71% yields ( Table 1, Entries 1 and 2).
Moreover, the classical reaction of Appel's salt 2 with NaNO3 (1 equiv) in dichloromethane (DCM), at ca. 20 °C, to give dithiazolone 9 in a 72% yield [17] failed to give any product with the tetrachlorothiadiazine 1; however, when DCM was replaced by MeCN, the desired thiadiazinone 3 was obtained in a 71%-72% yield ( Table 1, Entries 4 and 5). Presumably, the more polar MeCN encouraged the equilibrium between the covalent and ionic forms of thiadiazine 1 to shift favorably towards the latter. This study showed that the tetrachlorothiadiazine 1 and Appel's salt 2 have similar reactivity to many of the investigated reagents, indicating that both compounds, despite having a different form (covalent vs. ionic) are similarly electrophilic. However, thiadiazine 1, which is non-aromatic, is more sensitive to the presence of water in the reaction conditions.
Despite the fact that the simple pyridine-mediated condensation of malononitrile with Appel's salt 2 gives only a low yield of ylidenemalononitrile 11 (40%) [32], when the tetrachlorothiadiazinone 1 was treated with malononitrile (1.1 equiv) and 2,6-lutidine (2 equiv) in dry DCM, at ca. 20 °C, the starting material was consumed quickly (TLC, 10 min) to give the ylidenemalononitrile 10 in a 78% yield ( Table 2, Entry 1). Attempts to improve the yield of this reaction involved screening the base [1,8- and solvents (MeCN, THF, PhMe), but these led to either degradation or to lower yields. Nevertheless, in DCM at ca. 20 °C using lutidine (2 equiv) as the base, increasing the equivalents of malononitrile from 1.1-1.5 equiv afforded the ylidenemalononitrile 10 in a slightly better yield (82%), while a further increase to 2 equiv of malononitrile gave a similar 83% yield ( Table 2, Entries 2 and 3). The reaction could be scaled up to 4 mmol, and while this led to a drop in yield (73%) ( Table 2, Entry 5), it also enabled a chromatography free work-up that involved passing the reaction mixture through a short plug of SiO2, washing the organic phase with 2M HCl and H2O and precipitating the product from THF/pentane to give the ylidene 10 in a preparatively useful 64% yield (Table 2, Entry 6).  Attempts to develop a base-free protocol were less effective, as heating a PhMe solution of the tetrachlorothiadiazine 1 with malononitrile (2 equiv) at reflux (110 °C) led to complete consumption of the starting material only after 48 h and isolation of the desired ylidene 10 in a low 27% yield. Moreover, the condensation reactions of the thiadiazinone 3 with benzene-1,2-diamine or sodium 2-aminophenoxide to give fused systems are known (Scheme 9) [16]. Even though the thiadiazinone 3 was inert to reactions with primary amines at the C-4 position, after an initial nucleophilic addition in the C-3 position, intramolecular cyclizations readily occur with bisnucleophiles to afford tricyclic systems in excellent yields, e.g., 4-chloro-10H- [1,2,6]thiadiazino [3,4-b]quinoxaline (18) and 4-chlorobenzo [5,6] [1,4]oxazino[2,3-c][1,2,6]thiadiazine (19) (Scheme 9). Worthy of note was that while an excellent yield was reported for oxazine 19, this required an initial base activation (deprotonation) of the hydroxy group of 2-aminophenol to direct the reaction. The dichloromethylene of the tetrachlorothiadiazine 1 was expected to be more electrophilic and therefore more reactive than the C-3, C-4 and C-5 positions of the thiadiazinone 3. As such, bisnucleophiles were anticipated to initially attack the geminal dichloromethylene. This reactivity mimics that of Appel's salt 2 (Scheme 8) with bisnucleophiles where the first nucleophilic displacement occurs at the more electrophilic C-5 position.

Cyclisation
Treatment of the tetrachlorothiadiazine 1 with 2-aminophenol or benzene-1,2-diamine in MeCN at 20 °C for 1 h afforded the fused heterocycles 18 (68%) and 19 (68%), respectively (Scheme 10). Tentatively, the moderate yields (68%) can be attributed to the high reactivity of the tetrachlorothiadiazine 1 that can presumably suffer from both halophilic and thiophilic attack, leading to its degradation. A proposed mechanism for this transformation involves the initial formation of imine 20, which we know from previous studies readily forms by reaction of thiadiazine 1 with arylamines [22]. Subsequently, two routes are proposed, either a 6-exo cyclisation to give the final products (Route a) or a 5-endo cyclisation occurs to form the spirocyclic compounds 21 that then ring-opens, assisted by electron release from both the nitrogen's lone pair and possibly the ring sulfur, subsequently cyclizing at the C-3 position (route b). Interestingly, when using N′-phenylbenzene-1,2-diamine, an unsymmetrical diamine, the less reactive secondary amine ends up cyclizing on the thiadiazine C-3 position to give 4-chloro-10-phenyl-10H-[1,2,6][3,4-b]quinoxaline (22) in an 85% yield. While this route to tricyclic systems 18, 19 and 22 was non-quantitative, it offered two distinct advantages: firstly, it avoided the need to access the thiadiazinone 3, thereby reducing the number of steps to the final products, and secondly, it offered an alternative regioselectivity that avoided the need to base activate bisnucleophiles, such as the 2-aminophenol. [1,2,6]thiadiazino [3,4-b]quinoxaline (19) Interestingly, during the synthesis and isolation of 4-chloro-10H- [1,2,6]thiadiazino [3,4-b]quinoxaline (19), we observed its decomposition to 3-aminoquinoxaline-2-carbonitrile (23). Traces of this product were initially observed during chromatography of quinoxaline 19, so we decided to investigate the reactivity of this compound. Quinoxaline 19 was stable under basic conditions, as it was recovered unchanged after 48 h stirring in neat Et3N, while it was unstable in acid, as heating a solution in glacial AcOH at reflux for 15 min led to complete consumption of the starting material and isolation of 3-aminoquinoxaline-2-carbonitrile (23) as the only product in a 38% yield. Alternatively, heating a solution of 19 in aqueous HCl/THF at 80 °C gave the quinoxaline 23 in an 80% yield (Scheme 11). Furthermore, quinoxaline 19 was also unstable in oxidizing and reducing conditions. Namely, stirring a DCM solution of 19 with MnO2 (10 equiv) at ca. 20 °C led to complete consumption of the starting material after 22 h and isolation of aminoquinazoline 23 in an 85% yield (Scheme 11). The dissolving metal reduction with Zn (4 equiv) in AcOH led to degradation of 19 and isolation of 3-aminoquinoxaline-2-carbonitrile (23) in only a low 6% yield.

General Methods and Materials
All chemicals were commercially available, except those whose synthesis is described. Anhydrous Na2SO4 was used for drying organic extracts, and all volatiles were removed under reduced pressure. All reaction mixtures and column eluents were monitored by TLC using commercial glass-backed thin layer chromatography (TLC) plates (Merck Kieselgel 60 F254) [39]. The plates were observed under UV light at 254 and 365 nm. The technique of dry flash chromatography was used throughout for all non-TLC-scale chromatographic separations using Merck Silica Gel 60 (less than 0.063 mm, Merck KGaA, Darmstadt, Germany). Melting points were determined using a PolyTherm-A, Wagner & Munz, Kofler-Hotstage Microscope apparatus (Wagner & Munz, Munich, Germany) or were determined using a TA Instruments differential scanning calorimeter (DSC) Q1000 with samples hermetically sealed in aluminum pans under an argon atmosphere; using heating rates of 5 °C/min (DSC melting points listed by onset and peak values). Solvents used for recrystallization are indicated after the melting point. UV spectra were obtained using a Perkin-Elmer Lambda-25 UV-VIS spectrophotometer (Perkin-Elmer, Waltham, MA, USA), and inflections are identified by the abbreviation "inf". IR spectra were recorded on a Shimadzu FTIR-NIR Prestige-21 spectrometer (Shimadzu, Kyoto, Japan) with the Pike Miracle Ge ATR accessory (Pike Miracle, Madison, WI, USA), and strong, medium and weak peaks are represented by s, m and w, respectively. 1 H-and 13 C-NMR spectra were recorded on a Bruker Avance 300 (at 300 and 75 MHz, respectively, Bruker, Billerica, MA, USA) or a 500 machine (at 500 and 125 MHz, respectively). Deuterated solvents were used for the homonuclear lock, and the signals are referenced to the deuterated solvent peaks. Attached-proton-test (APT) NMR studies identified quaternary and tertiary carbons, which are indicated by (s) and (d) notations, respectively. MALDI-TOF mass spectra were recorded on a Bruker Autoflex III Smartbeam instrument. Low resolution (EI) mass spectra were recorded on a Shimadzu Q2010 GC-MS with a direct inlet probe. Dichloromalononitrile (6) [24], N-2,2trichloro-2-cyanoacetimidoyl chloride (7) [24] and dimethylsulfonium dicyanomethylide (14) [40] were prepared according to the reported procedures. Tetrachloro-4H-1,2,6-thiadiazine (1)