The [4+2]‐Cycloaddition of α‐Nitrosoalkenes with Thiochalcones as a Prototype of Periselective Hetero‐Diels–Alder Reactions—Experimental and Computational Studies

Abstract

The [4+2]‐cycloadditions of α‐nitrosoalkenes with thiochalcones occur with high selectivity at the thioketone moiety of the dienophile providing styryl‐substituted 4H‐1,5,2‐oxathiazines in moderate to good yields. Of the eight conceivable hetero‐Diels–Alder adducts only this isomer was observed, thus a prototype of a highly periselective and regioselective cycloaddition has been identified. Analysis of crude product mixtures revealed that the α‐nitrosoalkene also adds competitively to the thioketone moiety of the thiochalcone dimer affording bis‐heterocyclic [4+2]‐cycloadducts. The experiments are supported by high‐level DFT calculations that were also extended to related hetero‐Diels–Alder reactions of other nitroso compounds and thioketones. These calculations reveal that the title cycloadditions are kinetically controlled processes confirming the role of thioketones as superdienophiles. The computational study was also applied to the experimentally studied thiochalcone dimerization, and showed that the 1,2‐dithiin and 2H‐thiopyran isomers are in equilibrium with the monomer. Again, the DFT calculations indicate kinetic control of this process.

Superdienophiles select one of eight conceivable pathways! The hetero‐Diels–Alder reactions of α‐nitrosoalkenes with thiochalcones are highly periselective and regioselective processes, providing only one of eight possible constitutional isomers. The thiochalcones undergo a regioselective, but reversible dimerization. High‐level DFT calculations reveal that these processes are kinetically controlled.

Introduction

The employment of cycloaddition reactions1 belongs to the most important strategies for the preparation of functionalized carbocyclic and heterocyclic compounds.2 The many applications of the (hetero‐)Diels–Alder reaction3 and of the 1,3‐dipolar cycloaddition (Huisgen reaction)4 for the selective and efficient formation of six‐ and five‐membered ring systems prove the particular relevance of these cycloadditions. In general, they proceed as concerted reactions and are classified as periselective processes that can be treated by the Woodward–Hoffmann rules.5 Although countless examples of these cycloadditions have been reported, there are still puzzling selectivity issues and mechanistic problems.6, 7 One interesting selectivity challenge arises if two 4 π‐systems are involved in (hetero‐)‐ Diels–Alder reactions. In the case of the cycloadditions of a heterodiene A (one heteroatom a) with a second heterodiene of general structure B (two heteroatoms b and c) the hypothetical formation of eight constitutional isomers  C–J can be depicted (Scheme 1). Products C and D (and G and H) would be the result of [4 π_A+2 π_B] reactions, whereas the isomeric compounds E and F (and I and J) would arise from [2 π_A+4 π_B] processes. Given that different perimeters of the two π‐systems are involved, the term periselectivity as introduced by Houk8 is applicable for the relationship of C–F.9 For each of these cycloadducts the regioisomers G–J are conceivable. In addition to these hetero‐Diels–Alder reactions, [3+2]‐cycloadditions of the s‐trans‐conformer of heterodiene B with A can form four isomeric five‐membered heterocycles that incorporate a 1,3‐dipole moiety. Although all reactions are formally allowed by the Woodward–Hoffmann rules, it is evident that not all products are favored by kinetic and/or thermodynamic factors. To our best knowledge, this type of periselective and regioselective reactions has been rarely studied.10

Scheme 1.

Hypothetical products C–J of Diels–Alder reactions of heterodiene A with heterodiene B (for clarity lone pairs at the centers a, b, and c are omitted); the conceivable four [3+2]‐cycloadditions are not depicted.

In this report, we describe two types of hetero‐Diels–Alder reactions of 1,3‐diaryl‐substituted α,β‐unsaturated thioketones 1 (known as thiochalcones, a=S) that represent heterodiene A. Firstly, the reversible dimerization of these compounds is analyzed and secondly their hetero‐Diels–Alder reactions with α‐nitrosoalkenes 10 (b=O, c=N) as second heterodiene component B. Thiocarbonyl compounds and especially non‐enolizable thioketones are well known as versatile building blocks for the preparation of sulfur heterocycles with variable ring size through cycloaddition reactions.11 Based on kinetic studies, Huisgen et al. named these compounds superdipolarophiles12a–12c and superdienophiles12d, 12e to emphasize their high reactivity towards 1,3‐dipoles and dienes, respectively. In comparison with aryl‐, hetaryl‐, and ferrocenyl‐substituted thioketones, the related thiochalcones with the general structure 1 are much less explored in cycloaddition chemistry. Noteworthy, thiochalcones 1 exist in solutions as equilibrium mixtures of monomeric and dimeric forms, namely 3,4‐dihydro‐1,2‐dithiin 2 and 3,4‐dihydro‐2H‐thiopyran derivatives 4 (Scheme 2).13 The dimeric products are formed through thia‐Diels–Alder reactions with thiochalcones playing at the same time the role of the heterodiene and the C=S dienophile (→1,2‐dithiins 2 or 1,3‐dithiins 3) or C=C dienophile (→thiopyrans 4 or 5). Surprisingly, in contrast to several α,β‐unsaturated thioaldehydes such as thioacrolein, there is no indication that thiochalcones 1 provide the regioisomeric dimers 3 (1,3‐dithiins) or the regioisomeric 3,4‐dihydro‐2H‐thiopyran 5.13d

Scheme 2.

Monomeric thiochalcones 1 equilibrating with dimers 2 and 4, and trapping of 1 through [3+2]‐ and [4+2]‐cycloadditions by using fluorinated 1,3‐dipoles or acetylenic dienophiles, respectively, and periselective hetero‐Diels–Alder reactions of 1 with α‐nitrosoalkenes 10 leading to cycloadducts 7, 9, and 11.

In pioneering work, Lewis acid‐mediated hetero‐Diels–Alder reactions of several thiochalcones 1 with activated ethylenes including chiral fumarates and maleates have been described.14 In contrast, in our recent publication, organocatalytic asymmetric [4+2]‐cycloadditions of thiochalcones 1 as heterodienes with in situ generated enantiopure dienamines derived from O‐silylated l‐prolinols have also been demonstrated.15 In addition, we described thia‐Diels–Alder reactions of 1 acting as heterodienes with acetylenic dienophiles16a, 16b as well as with 1,4‐quinones.16c Notably, in the case of unsymmetrically activated acetylenes such as 8, the [4+2]‐cycloadditions occurred with complete regioselectivity, and the sulfur atom attacked always the β‐position of the Michael‐type acceptor to give 4H‐thiopyrans 9 (Scheme 2).16b In contrast to the well‐established behavior of chalcones as reactive C=C dipolarophiles,17 [3+2]‐cycloadditions of thiochalcones 1 are almost unknown.18 Only in a recent publication,18a it was demonstrated that the 1,3‐dipolar cycloadditions of in situ generated electron‐deficient fluorinated nitrile imines 6, derived from trifluoroacetonitrile, occur in a fully chemo‐ and regioselective fashion onto the C=S bond of 1. The thioketone moiety played the role of a most suitable heterodipolarophile and 2,3‐dihydro‐1,3,4‐thiadiazoles 7 were obtained as products of these reactions (Scheme 2).

α‐Nitrosoalkenes constitute an exceptional class of highly reactive heterodienes, and their Diels–Alder reactions with C=C and C≡C dienophiles have been studied by experiments19 and with computational methods.20 Confirming an early study on α‐nitrosoalkene/thioketone cycloadditions,21 we recently described regioselective hetero‐Diels–Alder reactions of aryl‐, hetaryl‐, and ferrocenyl‐substituted thioketones with in situ generated electron‐deficient α‐nitrosoalkenes 10. The corresponding 4H‐1,5,2‐oxathiazine derivatives were obtained as the only products in a regio‐ and chemoselective manner.22 For the first time, alkyl‐substituted thioketones were successfully reacted with α‐nitrosoalkenes 10 also giving the expected heterocycles in good yields. In a preliminary experiment with 1,3‐diphenylprop‐2‐ene‐1‐thione (1 a, Ar¹, Ar²=Ph) and 1‐nitroso‐1‐phenylethylene (10 a, R=Ph) it was found that in this case the C=S bond was involved in the [4+2]‐cycloaddition reaction, yielding the corresponding 4‐styryl‐substituted 4H‐1,5,2‐oxathiazine of type 11 as major product.

The goal of the present work was the systematic examination of hetero‐Diels–Alder reactions of selected in situ generated α‐nitrosoalkenes 10 with a series of electron‐rich aryl‐, hetaryl‐, and ferrocenyl‐substituted thiochalcones 1 to establish scope and limitations of this periselective process. The obtained experimental results should be rationalized by theoretical studies that were also applied to the thiochalcone dimerization and the cycloadditions of related nitroso compounds and thioketones.

Results and Discussion

Synthesis of thiochalcones and dimerization

A series of thiochalcones 1 a–l bearing diverse aryl, hetaryl, and ferrocenyl substituents was selected for the present study. All compounds were prepared according to a known procedure based on the treatment of the corresponding chalcones 12 with Lawesson's reagent (LR) (Scheme 3).16 The “thiochalcone fractions” containing predominantly mixtures of dimers 2 and 4 were isolated chromatographically and subsequently used for further studies without separation of the components. In analogy to earlier reports,13b, 13c compounds bearing phenyl (1 a–g) or ferrocenyl (1 h, i) groups located at the C=S group (Ar¹) provided complex mixtures of dimers 2 and 4, and depending on the type of substituents, the composition of the mixtures slightly differed in the studied cases. For example, the “thiochalcone fraction” of 1 a was isolated chromatographically as a solid (2 a:4 a in approx. 1:1 ratio). When treated with petroleum ether at room temperature, a colorless solid precipitated, which was separated by filtration, recrystallized from benzene/diethyl ether and was studied by spectroscopy. The ¹H NMR spectrum immediately recorded in CDCl₃ solution at room temperature revealed the presence of two diastereomeric 1,2‐dithiins, cis‐ and trans‐2 a, as major components, and the structures of these isomers were confirmed by 2D NMR methods (for details, see the Supporting Information). The calculated NMR data also confirm the structure of these dimers 2 a as well as their isomers 4 a.

Scheme 3.

Synthesis of thiochalcones 1  a–l through thionation of chalcones 12  a–l with Lawesson's reagent (LR), and the structure of isolated 3,4‐dihydro‐2H‐thiopyran derivatives 4  j–l.

However, during the storage of the solution of 1,2‐dithiins cis‐ and trans‐2 a at room temperature overnight, the color turned blue and the ¹H NMR spectrum evidenced that, again, an equilibrium mixture containing diastereomers 2 a and 3,4‐dihydro‐2H‐thiopyran derivative 4 a exists in a 1:1:5 ratio. In addition, another set of signals was found in this NMR spectrum which indicated the presence of trace amounts of another dimeric product, presumably one of the two possible diastereomeric 1,3‐dithiins 3 a.23 These observations fit well with the literature report that two isomers of 1,2‐dithiins of type 2 form the dominant fraction of the thiochalcone dimers bearing Ph (for Ar¹) and SPh (instead of Ar²) substituents. This constitution was confirmed by a desulfurization experiment with Raney nickel, which led to isolation of 1,4‐diphenylhexane as the major product.13b

A similar solvent‐dependent equilibrium shift was observed for other “thiochalcone fractions” of compounds bearing para‐substituted phenyl (1 b and 1 c) and naphthyl (1 d) groups. However, in the case of analogs bearing the ferrocenyl group (1 e, 1 h, and 1 i) or with a hetaryl moiety located at the β‐position (1 f, 1 g), the ¹H NMR spectra of the purified material revealed either a very complex pattern or significant broadening of the signals was observed. Therefore, a reliable interpretation of the composition of the resulting mixtures was not possible. In contrast, more electron‐rich analogs bearing 3,4‐methylenedioxyphenyl (1 j), 2‐furyl (1 k), and 2‐thienyl (1 l) substituents as Ar¹ groups, provided almost exclusively 3,4‐dihydro‐2H‐thiopyrans 4 j–l as single diastereomers. These compounds were isolated and characterized spectroscopically. For example, the ¹H NMR spectrum of 4 j revealed the presence of a single set of diagnostic signals at 4.04 (dd, J=4.2, 6.5 Hz, HC(4)), 5.02 (d, J=11.1 Hz, HC(2)), 5.17 (dd, J=4.2, 11.1 Hz, HC(3)), and 6.24 ppm (d, J=6.5 Hz, HC(5)), attributed to hydrogen atoms of the central thiopyran ring. Moreover, the coupling constants J=4.2–4.4 Hz between the axial HC(3) and the equatorial HC(4) observed for the series 4 j–l strongly support the structure of endo‐cycloadducts (Scheme 3).13 Notably, also in the case of the a–d series only endo‐dimers 4 could be identified in the mixtures.

Hetero‐Diels–Alder reactions of thiochalcones with α‐nitrosoalkenes

It was assumed that the [4+2]‐cycloaddition reactions with α‐nitrosoalkenes 10 occur with monomeric thiochalcones 1 that are present in mixture with their dimers 2/4 and therefore different rates depending on the monomer stationary concentration in the reaction solution can be expected. Employing an established method,24 α‐nitrosostyrene 10 a was generated in situ by treatment of α‐chlorooxime 13 a (two equivalents) with potassium carbonate in dry dichloromethane at room temperature. The heterogeneous conditions guarantee low concentrations of 10 a and hence minimize the oligomerization of this reactive species. Due to the anticipated limited stability of the heterocyclic products, the reactions of 13 a with the dimers of thiochalcones 1 a–l were carried out at room temperature (or below), analogously to the already reported experiments with simple thioketones (Scheme 4).22 The conversions were monitored by TLC until the spots of the starting thiochalcone dimers completely disappeared. The ¹H NMR analysis of the crude reaction mixtures confirmed the formation of the expected [4+2]‐cycloadducts of type 11, contaminated in most of the cases by side products 14 (in general less than 10 %). The fairly stable major products were isolated by standard column chromatography in moderate to good yields up to 55 %, and the structure of the hitherto unknown 4H‐1,5,2‐oxathiazine derivatives 11 b–i was confirmed by standard spectroscopic methods supplemented with mass spectrometry and combustion analysis. For example, in the case of the ferrocenyl‐substituted 4H‐1,5,2‐oxathiazine 11 e, characteristic resonances attributed to C(3), C(4), and C(6) were found in the ¹³C NMR spectrum at 151.7, 23.5, and 87.0 ppm, respectively. In the constitutional isomer 11 h, bearing the ferrocenyl group at C(6), the corresponding signals were found at similar regions (152.1, 23.8, and 88.6 ppm). The attempted synthesis of 4H‐1,5,2‐oxathiazine derivatives 11 j–l performed under analogous reaction conditions (starting with dimers 4 j–l bearing electron‐donating substituents) led to the expected product only in a very low yield (11 j) or failed completely (11 k, 11 l). Instead, the formation of [4+2]‐cycloadducts 14 k and 14 l as mixtures of diastereomers was observed (Scheme 4). These results can be rationalized by the fact that the respective precursor dimers 4 exhibit enhanced stability and do not release sufficient amounts of the monomeric thiochalcone 1 at room temperature. Therefore, the reaction of heterodiene 10 a occurs mainly or exclusively with the dimers 4 leading to the formation of cycloadducts 14.

Scheme 4.

Cycloadditions of in situ generated 1‐nitroso‐1‐phenylethene (10  a) with thiochalcones 1  a–l (in equilibria with their dimers 2 and/or 4) leading to 4H‐1,5,2‐oxathiazines 11  a–j; (except from 14  b, side products of type 14 deriving from thiochalcone dimers endo‐4 were not isolated). [a] Reaction performed at 0 °C; [b] Not isolated; yield estimated based on the ¹H NMR of crude reaction mixture.

To gain more insight into this relationship of hetero‐Diels–Alder reactions of monomer 1 and/or dimer 4, the reactions of the parent “thiochalcone fractions” 2/4 with 10 a were studied in more detail and the crude product mixtures were analyzed by ¹H NMR spectroscopy (Table 1). When the reaction of 2 a/4 a with 10 a was performed under standard conditions (CH₂Cl₂, RT), the products 11 a and 14 a were formed in an approximate 4:1 ratio in favor of (E)‐3,6‐diphenyl‐6‐styryl‐4H‐1,5,2‐oxathiazine (11 a), which was finally isolated in 39 % yield (Table 1, entry 1). To evaluate the influence of daylight, the same reaction was performed under light exclusion, but essentially the same mixture of 11 a and 14 a was obtained. Notably, decreasing the reaction temperature to 0 °C led to a similar mixture, in which the components 11 a and 14 a were found in an approximate 9:1 ratio, and the major product 11 a was isolated in slightly higher yield of 45 % after a remarkably longer reaction time of 20 h (Table 1, entry 2). Finally, a “thiochalcone fraction” enriched in dimer endo‐4 a was prepared by equilibration of the dimer sample in CH₂Cl₂ overnight. By reaction with 10 a, the respective products 11 a and 14 a were now formed in an approximate 1:2 ratio (Table 1, entry 3). Unfortunately, an attempted isolation of pure samples of diastereomers 14 a by chromatography techniques failed due to similar polarity and limited stability of these cycloadducts. Noteworthy, the ¹H NMR analysis of the fractions obtained by column chromatography revealed the presence of higher amounts of 11 a and the dimer endo‐4 a, which indicates a possible decomposition of 14 a by cycloreversion across the thiopyran ring.

Table 1.

Ratio of products 11 and 14 formed in the reactions of model nitrosoalkene 10 a with “thiochalcone fractions” 2/4 as determined by ¹H NMR spectroscopy of the crude product mixture^[a]

Entry	Thiochalcone fraction 2/4	Ratio of crude product mixture 11:14:14'
1	a	78:20:2
2	a [b]	89:10:1
3	a [c]	29:65:6
4	b	67:29:4
5	b [c]	31:59:10
6	c	48:35:17
7	d	70:29:1
8	e	100:0:0
9	f	87:11:2
10	g	65:28:7
11	h	100:0:0
12	i	100:0:0
13	j	22:53:25
14	k	0:68:32
15	l	0:79:21

[a] Reaction conditions: CH₂Cl₂, room temperature, K₂CO₃, 13 a (2.0 equiv), thiochalcone fraction 2/4. [b] Reaction performed at 0 °C. [c] Reaction carried out by using thiochalcone fraction 2/4 pre‐equilibrated in CH₂Cl₂ overnight at room temperature.

Similar results were noticed in the reaction of 10 a with p‐bromophenyl‐substituted derivative 2 b/4 b. The standard “thiochalcone fraction” mainly furnished 11 b and minor amounts of 14 b/14 b', whereas a “thiochalcone fraction” enriched in 4 b provided 14 b/14 b' as major components (Table 1, entries 4 and 5). In this case, the subsequent purification by preparative TLC enabled isolation of small samples of both diastereomeric cycloadducts, 14 b (9 %) and 14 b' (2 %), obtained as a glassy solid and a thick colorless oil, respectively. The spectroscopic analysis of both products confirmed the anticipated bis‐heterocyclic structure of 14 b and 14 b' formed through [4+2]‐cycloaddition of 10 a with the dimeric thiochalcone, endo‐4 b. For example, in the ¹H NMR spectrum of 14 b, the signals attributed to the hydrogen atoms of the thiopyran ring were found at 3.61 (dd, J=3.5, 9.1 Hz, HC(3′)), 4.40 (dd, J=3.5, 6.6 Hz, HC(4′)), 4.97 (d, J=9.1 Hz, HC(2′)), and 6.36 ppm (d, J=6.6 Hz, HC(5′)), whereas the diagnostic resonances of the CH₂ group of the 1,5,2‐oxathiazine unit were located at 2.58 and 3.25 ppm (AB system, J=16.8 Hz). In the case of 14 b', a different pattern of signals was observed: the HC(3′) and HC(4′) resonances appeared as broadened pseudo‐triplets (for details, see the Supporting Information), whereas the highfield signals of the CH₂ group were found at 2.97 and 3.17 ppm (AB system, J=17.3 Hz). Thus, based on the characteristic chemical shifts observed for the diastereomeric compounds of type 14/14' in the ¹H NMR spectra, all products formed as major diastereomers in the series are considered to have the analogous relative configurations at the newly generated stereogenic center at C(6) of the 1,5,2‐oxathiazine ring. Unfortunately, the relative configuration could not be determined by spectroscopic methods and the attempted isolation of pure crystalline compounds of type 14/14' suitable for X‐ray diffraction was also unsuccessful.

The observed results of these reactions collected in Table 1 clearly demonstrate that the proportions of cycloadducts 11 and 14 formed under standard conditions reflect to some extent the composition of the “thiochalcone fraction”. Introduction of an electron‐donating group at the Ar¹ substituent (j, k, and l) in thiochalcone 1 (entries 13–15) or previous equilibration of “thiochalcone fractions” in CH₂Cl₂ (entries 3 and 5) increases the content of the endo‐4 dimer in the mixture, and hence, favors the formation of bis‐heterocyclic products of type 14/14'. To the best of our knowledge, these are the first examples of [4+2]‐cycloadditions of an electron‐deficient heterodiene with dimeric thiochalcones of type 4 playing the role of the electron‐rich C=S heterodienophile. In contrast, thiochalcones bearing the bulky ferrocenyl group either as Ar¹ (h, i) or Ar² (e) substituents lead to the expected monocyclic products 11 exclusively (entries 8, 11, and 12), very likely due to higher stationary concentrations of the monomeric thiochalcones 1.

In extension of the study, two particular electron‐deficient α‐nitrosoalkenes, 10 b (R=CF₃) and 10 c (R=COOEt) derived from 3‐bromo‐1,1,1‐trifluoro‐2‐(hydroxyimino)propane (13 b) and ethyl 3‐bromo‐2‐(hydroxyimino)propionate (13 c), respectively, were also tested in cycloadditions with the parent system 2 a/4 a (Scheme 5). By reacting of in situ generated 10 b with 1 a, the attempted preparation of the trifluoromethyl‐substituted analog of 4H‐1,5,2‐oxathiazine 11 a gave a mixture of products identified as 11 m and bis‐heterocyclic adducts 14 m and 14 m' formed in an approximate 6:3:1 ratio. However, the attempted isolation of these products by column chromatography resulted in complete decomposition. This observation is consistent with the previously described results of the reactions of the fluorinated α‐nitrosoalkene 10 b with simple thioketones,22 which did not afford stable products in most of the performed hetero‐Diels–Alder reactions. Finally, reaction of 1 a with in situ generated α‐nitrosoalkene 10 c, bearing an ethoxycarbonyl group, led to a mixture containing 11 n, 14 n and 14 n' in an approximate 2:5:2 ratio. The subsequent chromatographic purification followed by fractional crystallization enabled the isolation of a pure sample of the 14 n in 26 % yield.

Scheme 5.

Reactions of very electron‐deficient α‐nitrosoalkenes 10  b and 10  c generated from oximes 13  b or 13  c, respectively, with the dimers 2  a/4  a and its monomer 1  a.

In summary, the presented preparative results show that the composition of the “thiochalcone fraction” considerably determines the type of products formed, and that the substituent present in the heterodiene is of importance. In the studied system of the two heterodienes 1 and 10, the former component plays exclusively the role of the heterodienophile, reacting periselectively and regioselectively at the C=S bond, whereas the α‐nitrosoalkene acts as heterodiene to give styryl‐substituted 4H‐1,5,2‐oxathiazines 11. The C=C bond of these cycloadducts offers many options for further functionalization and hence libraries of unique heterocycles should be available for examination of their properties, for example, as biologically active compounds.

Computational Study

Methods and procedure

To study the reaction mechanisms of the dimerization and the hetero‐Diels–Alder reactions in detail, high‐level DFT calculations (PBE1PBE/def2‐TZVP+PCM(dichloromethane)+GD3BJ dispersion correction26, 27, 28, 29 on the basis of preceding B3LYP/6‐31G(d)30+GD3BJ geometry optimizations) were performed. In the following part, we report relative Gibbs free enthalpies with respect to the sum of respective educts s‐trans‐ 1 a and 10 a (ΔG ₂₉₈, kcal mol⁻¹, see the Experimental Part and the Supporting Information for details). This study concentrates on classical [4+2]‐cycloadditions of closed‐shell species; open‐shell species (radicals, radical cations, for example, by influence of light) were not considered in this work. Possible catalytically active species like Brønsted or Lewis acids and bases present in the reaction mixtures were not considered in the calculations.10d NMR‐calculations for 2 a, endo‐4 a and for 11 a are in accord with the experimental findings and support the reported constitutions and configurations of the products. Only products derived from E‐thiochalcone 1 a were considered.

Dimerization

In principle, the thiochalcone 1 a may form the four isomeric products 2–5 of hetero‐Diels–Alder‐reactions (see Scheme 2 and Scheme 6) if only the E‐form of 1 a is considered, and for each of the four products two diastereomers are possible due to the exo‐ or endo‐approaches of the components. The four principal products may be distinguished according to the number of bonds between the two sulfur atoms (1–4 bonds are possible). For all of the eight possible isomers, total and relative enthalpies were calculated and Scheme 6 summarizes the values with decreasing stability of the products. Among all isomers considered, compound exo‐4 a with a 1,5‐S,S‐arrangement came out to be by far the thermodynamically most stable isomer (−16.2 kcal mol⁻¹) and second is dimer exo‐5 a (1,4‐S,S‐arrangement, −12.8 kcal mol⁻¹). The diastereomers endo‐4 a and endo‐5 a are both slightly less stable compared with the respective exo‐isomers (−12.5 and −12.2 kcal mol⁻¹). They are followed by dimer trans‐2 a with an S−S‐bond within the heterocycle (1,2‐S,S‐arrangement) and its diastereomer cis‐2 a (−10.6 and −9.8 kcal mol⁻¹). The least stable dimers among these isomers are trans‐3 a and cis‐3 a (1,3‐S,S‐arrangement). In all cases, the diastereomers with a trans‐arrangement of the most spacious groups were calculated to be lower in enthalpy than the respective cis‐arrangements. Experimentally, the isomer endo‐4 a was found together with cis/trans‐2 a (see framed compounds in Scheme 6).

Scheme 6.

Results of the DFT‐calculations for the dimerization reaction of E‐thiochalcone 1 a leading to (hypothetical) cycloadducts 2 a–5 a (relative Gibbs free enthalpies ΔG ₂₉₈ are given in kcal mol⁻¹). The experimentally observed dimers endo‐4 a and cis/trans‐2 a are depicted in frames.

Furthermore, the transition states for the formation of the products in the sense of a synchronous, but possibly asymmetric reaction pathway were elucidated. According to these calculations, kinetic stability is expected for the heterocycles exo‐4 a, exo‐5 a, endo‐4 a (for structure, see Scheme 7, left), endo‐5 a and cis‐3 a with barriers for a cycloreversion of 22 kcal mol⁻¹ or more. Slightly smaller barriers were obtained for the cycloreversions of dimers trans‐2 a, cis‐2 a, and trans‐3 a, which indicate that these species may be subject of equilibration under the reaction conditions (room temperature, dichloromethane). These values are in good agreement with the NMR‐spectroscopic evidence for the equilibration of isomers trans‐ and cis‐2 a and endo‐4 a (see above).

Scheme 7.

Calculated transition state structures of endo‐4  a‐TS (left) and of 11  a‐TS (right). In both examples the transition state relevant C⋅⋅S distances are significantly shorter compared to the distances between the C⋅⋅C respectively C⋅⋅O atoms. endo‐4  a‐TS (left; C⋅⋅S distance 2.307 Å; C⋅⋅C distance 2.726 Å; imaginary frequency −205.55 cm⁻¹); 11  a‐TS (right; C⋅⋅S distance 2.483 Å; O⋅⋅C distance 2.953 Å; imaginary frequency −125.03 cm⁻¹). Hydrogen: white, carbon: grey, nitrogen: green, oxygen: red, sulfur: blue.

For the dimerization, low barriers (<12 kcal mol⁻¹) were calculated for the isomers exo‐4 a, endo‐4 a (Scheme 7, left), trans‐2 a, cis‐2 a, and trans‐3 a, whereas exo‐5 a, endo‐5 a, and cis‐3 a have barriers higher than 14 kcal mol⁻¹. Among the thermodynamically most stable calculated isomers, the one with the lowest barrier of 10.1 kcal mol⁻¹ is compound endo‐4 a, which was experimentally found and unambiguously characterized by NMR spectroscopy and calculation of the NMR data. Thus, the dimerization of thiochalcone 1 a is considered to be an essentially kinetically controlled process, leading to isomers trans‐ and cis‐2 a and the third most stable isomer endo‐4 a, but not to the thermodynamically most stable form exo‐4 a. The fact, that dimers trans‐ and cis‐3 a are not observed (within the analytical limits) may be due to the reversibility of their formation.

Hetero‐Diels–Alder reactions with nitroso compounds

The cycloaddition of s‐trans‐E‐thiochalcone 1 a with nitrosobenzene 15 was first computed as model reaction for the following hetero‐Diels–Alder reactions (Scheme 8). From the calculations, two products 16 a and 17 a generated by the [4+2]‐cycloadditions have to be expected, both of fairly low thermodynamic and kinetic stability. The three contiguous heteroatoms within the six‐membered ring may be the reason for the low stability of these cycloadducts. We are not aware of an experimental investigation of this reaction, but the calculations indicate that an equilibrium mixture of the precursors 1 a and 15 and cycloadduct 16 a can be expected.

Scheme 8.

Results of the DFT‐calculations for the reaction of E ‐thiochalcone 1  a with nitrosobenzene 15 to heterocycles 16  a and 17  a (relative Gibbs free enthalpies ΔG ₂₉₈ are given in kcal mol⁻¹).

In contrast, the calculations for the cycloaddition of α‐nitrosostyrene 10 a with thiobenzophenone 18 reveal that in this case the reaction proceeds over a relatively small barrier (9.7 kcal mol⁻¹) to a rather stable six‐membered ring product 19 a (−26.0 kcal mol⁻¹). The formation of regioisomer 20 a requires a quite large activation barrier (Scheme 9). Remarkably, the formation of a five‐membered heterocycle 21 a by a [3+2]‐cycloaddition requiring the s‐trans‐conformer of 10 a seems also possible, since its barrier is calculated to be only slightly higher (10.9 kcal mol⁻¹) and it shows pronounced thermodynamic stability (−28.5 kcal mol⁻¹). The formation of the less stable regioisomer 22 a requires a substantially higher activation enthalpy. Within the experimental limits, the formation of nitrones of type 21 a was not observed.22, 31, 32 Again, these data reveal that kinetic control of the cycloaddition is responsible for the formation of the experimentally found cycloadduct 19 a.

Scheme 9.

Results of the DFT‐calculations for the reaction of α‐nitrosostyrene 10  a with thiobenzophenone 18 leading to [4+2]‐cycloadducts 19  a and 20  a or [3+2]‐cycloadducts 21  a and 22  a (relative Gibbs free enthalpies ΔG ₂₉₈ are given in kcal mol⁻¹).

Following the situation sketched in Scheme 1, the thiochalcone–nitrosoalkene cycloadditions of (E)‐1 a and 10 a can afford eight different isomers (two of them with exo/endo diastereomers). Scheme 10 arranges the products following their calculated stability, among these isomers, two of them show a S−O bond (27 a, 29 a; 1,2‐S,O‐arrangement), two of them 1,3‐S,O‐ (11 a, (Scheme 7, right), 28 a), 1,4‐S,O‐ (24 a, trans‐25 a, cis‐25 a), and 1,5‐S,O‐arrangements (23 a, trans‐26 a, cis‐26 a). The most stable calculated isomer is isomer 23 a (−31.2 kcal mol⁻¹; 1,5‐S,O‐distance), followed by isomers 24 a and the experimentally found 11 a (−23.5 kcal mol⁻¹, Scheme 7, right). As expected, isomers with three contiguous heteroatoms (27 a, 28 a, 29 a) are less stable compared to the others due to unfavorable lone pair interactions.

Scheme 10.

Results of the DFT‐calculations for the reaction of E ‐thiochalcone 1  a with α‐nitrosostyrene 10  a (relative Gibbs free enthalpies ΔG ₂₉₈ are given in kcal mol⁻¹).

High barriers (>16 kcal mol⁻¹) for the cycloaddition were found for the reactions leading to isomers 23 a, 24 a, trans‐25 a, cis‐25 a, 27 a, and 28 a, smaller barriers (<15 kcal mol⁻¹) for 11 a (for structure see Scheme 7, right), trans‐26 a, cis‐26 a, and 29 a. Low barriers (<25 kcal mol⁻¹) for cycloreversion were only obtained for isomers cis‐26 a, trans‐26 a, and 29 a. Among the three thermodynamically most stable isomers (11 a, 23 a, and 24 a) the experimentally found and characterized isomer 11 a shows by far the lowest barrier for the cycloaddition reaction. Consequently, we interpret its preferred formation as a kinetically controlled process, leading to the thermodynamically third‐best isomer and confirming the role of the thiocarbonyl moiety as superdienophile. Other isomers may possibly not be formed under the reaction conditions (0 °C) or are subject of rapid equilibration of less stable cycloadducts. The calculations also suggest that under thermodynamic control the formation of the most stable 1,2‐oxazine derivatives of type 23 should be possible. Experimentally, an equilibration of 11 and the precursors 1 and 10 could not be proved due its limited stability.

Conclusions

A series of thiochalcones 1 was smoothly prepared by treatment of the corresponding chalcones 12 with Lawesson's reagent. As shown by NMR spectroscopy, the generated monomeric products 1 are in equilibrium with their dimers 2 (3,4‐dihydro‐1,2‐dithiins) and 4 (3,4‐dihydro‐2H‐thiopyrans), whereas the conceivable regioisomeric dimers 3 and 5 are not observed. Monomers such as 1 j–l with electron‐rich aryl groups Ar¹ give mainly or exclusively the dimers endo‐4 j–l. High‐level DFT calculations with the parent compound 1 a indicate that the dimerization process is kinetically controlled. Low barriers allow for the formation of endo‐4 a and cis/trans‐2 a, whereas the most stable dimer exo‐4 a is not formed due to a slightly higher barrier.

The monomers 1 can be successfully trapped by hetero‐Diels–Alder reactions with in situ generated nitrosoalkenes such as α‐nitrosostyrene 10 a. In these [4+2]‐cycloadditions the thioketone moieties of thiochalcones 1 function as superdienophiles with exclusive formation of styryl‐substituted 4H‐1,5,2‐oxathiazines 11. The cycloaddition hence proceeds with high periselectivity and regioselectivity. Closer inspection of crude product mixtures of the reactions shows that, depending on the thiochalcone substitution pattern, small or considerable amounts of the bis‐heterocyclic [4+2]‐cycloadducts 14 are formed, which derive from endo‐4 dimers. In particular, with electron‐rich groups Ar¹, where the corresponding monomers 1 are less favored in the equilibria, compounds 14 are found to be the major products. Again, the hetero‐Diels–Alder reaction only involves the thioketone moiety of endo‐4.

The DFT calculations show that the cycloaddition of α‐nitrosostyrene 10 a to the C=C double bond of 1 a should lead to the most stable, but not formed products 23 a and 24 a, followed by experimentally observed 11 a, and those arising from the cycloaddition of thiochalcone 1 a and the C=C double bond of 10 a (25 a, 26 a). By far least stable are the cycloadducts resulting from 1 a and the N=O double bond of 10 a (28 a, 29 a). Like the dimerization reactions of thiochalcone 1 a, the hetero‐Diels–Alder reaction of 1 a with 10 a to 11 a is also a kinetically controlled process. The thermodynamically most stable [4+2]‐cycloadduct 23 a that would originate from a reaction of the thiochalcone C=C double bond with the nitrosoalkene is apparently not formed due the high barrier of this cycloaddition.

Overall, the so far unstudied thiochalcone/α‐nitrosoalkene cycloadditions are identified as unique prototypes of periselective and regioselective processes. Only one of the eight possible constitutional isomers is observed—if the conceivable four [3+2]‐cycloadditions are also considered the high periselectivity is even more intriguing.

Experimental Section

For general information, all experimental and analytical details and computational details see the Supporting Information.

General procedure for reactions of the in situ generated α‐nitrosoalkenes with “thiochalcone fractions”: To a solution of the corresponding α‐halooxime 13 (2.00 mmol) in dry dichloromethane (6 mL), an excess of solid potassium carbonate (2.76 g, 20.0 mmol) was added. To the resulting suspension, a freshly prepared solution of the corresponding “thiochalcone fraction” 2/4 (1.00 mmol) in dry dichloromethane (2 mL) was added dropwise at room temperature, and stirring was continued until the characteristic color of the starting thiocarbonyl precursors faded. After completion of the reaction (confirmed by TLC) the precipitated inorganic materials were filtered off, washed with dichloromethane (2×4 mL), and the solvents were removed under reduced pressure. In all experiments the mass recovery was high (>90 %), and the ratio of products 11 and 14 identified in the crude mixtures and collected in Table 1 was established based on the registered ¹H NMR spectra. The residue obtained thereafter was purified by column chromatography (SiO₂, petroleum ether/dichloromethane 7:3, gradient 1:1) and the product was recrystallized from a petroleum ether/dichloromethane mixture to give the corresponding 4H‐1,5,2‐oxathiazine derivative 11 and/or product 14 as crystalline materials.

(E)‐3,6‐Diphenyl‐6‐styryl‐4H‐1,5,2‐oxathiazine (11 a):22 the product was obtained in improved yield by a modified general protocol, running the reaction at 0 °C for 20 h; yield: 161 mg (45 %); colorless crystals, m.p. 139–140 °C (decomp.). ¹H NMR (CDCl₃, 600 MHz): δ=3.42, 3.62 (AB system, J=17.4 Hz, 2 H, H₂C(4)), 6.52, 6.75 (AB system, J=15.9 Hz, 2 H, 2=CH), 7.24–7.42, 7.60–7.63, 7.72–7.75 ppm (3 m, 11 H, 2 H, 2 H, Ph). For further characterization see Ref. 22.

(E)‐6‐(4‐Bromostyryl)‐3,6‐diphenyl‐4H‐1,5,2‐oxathiazine (11 b): reaction time: 90 min; yield: 145 mg (33 %); colorless crystals, m.p. 196–198 °C (EtOAc) (decomp.). ¹H NMR (CDCl₃, 600 MHz): δ=3.40, 3.60 (AB system, J=17.5 Hz, 2 H, H₂C(4)), 6.50, 6.67 (AB system, J=15.9 Hz, 2 H, 2=CH), 7.24–7.46, 7.59–7.62, 7.70–7.73 ppm (3 m, 10 H, 2 H, 2 H, Ph, Ar). ¹³C NMR (CDCl₃, 151 MHz): δ=23.3 (t, C(4)), 86.7 (s, C(6)), 122.3 (s, CBr), 125.7, 126.9, 128.4, 128.66, 128.71*, 129.8, 129.9, 131.6, 131.8 (9 d, Ph, Ar,=CH), 134.6, 135.9, 139.8 (3 s, Ph, Ar), 152.3 ppm (s, C(3)); * signal with higher intensity. IR: $\tilde{\nu }$ =3053m, 3026m, 2914m, 1524m, 1476m, 1444m, 1368m, 1248m, 1245m, 1220m, 1197m, 951s, 918s, 836m, 758s, 745s, 711s, 693vs cm⁻¹. MS (ESI): m/z (%)=437 (6, [M+H]⁺), 223 (48), 149 (100). Anal. Calcd for C₂₃H₁₈BrNOS (436.37): C 63.31, H 4.16, N 3.21, S 7.35; found: C 63.11, H 4.20, N 3.17, S 7.23.

(E)‐3,6‐Diphenyl‐6‐(4‐methoxystyryl)‐4H‐1,5,2‐oxathiazine (11 c): reaction time: 60 min; yield: 110 mg (28 %); pale orange crystals, m.p. 159–160 °C (decomp.). ¹H NMR (CDCl₃, 600 MHz): δ=3.40, 3.61 (AB system, J=17.5 Hz, 2 H, H₂C(4)), 3.80 (s, 3 H, OMe), 6.37, 6.66 (AB system, J=15.9 Hz, 2 H, 2=CH), 6.82–6.86, 7.32–7.42, 7.59–7.62, 7.70–7.73 ppm (4 m, 2 H, 8 H, 2 H, 2 H, Ph, Ar). ¹³C NMR (CDCl₃, 151 MHz): δ=23.5 (t, C(4)), 55.3 (q, OMe), 87.0 (s, C(6)), 114.0, 125.7, 126.6, 126.9, 128.2 (5 d, Ar), 128.3 (s, Ar), 128.56, 128.60*, 129.8, 132.5 (4 d, Ar,=CH), 136.0, 140.1 (2 s, Ar), 151.9 (s, C(3)), 159.8 ppm (s, COMe); *signal with higher intensity. IR: $\tilde{\nu }$ =3068m, 2931m, 2855m, 1605m, 1515s, 1455m, 1367s, 1193m, 1053m, 933s, 858s, 793vs, 698s cm⁻¹. HRMS (ESI‐TOF): m/z [M+H]⁺ calcd for C₂₄H₂₂NO₂S: 388.1371; found: 388.1373.

(E)‐3,6‐Diphenyl‐6‐[2‐(naphth‐2‐yl)vinyl]‐4H‐1,5,2‐oxathiazine (11 d): reaction time: 60 min; yield: 150 mg (37 %); colorless crystals, m.p. 171–172 °C (decomp.). ¹H NMR (CDCl₃, 600 MHz): δ=3.43, 3.65 (AB system, J=17.5 Hz, 2 H, H₂C(4)), 6.64, 6.90 (AB system, J=15.9 Hz, 2 H, 2=CH), 7.34–7.47, 7.59–7.64, 7.75–7.80 ppm (3 m, 8 H, 3 H, 6 H, Ar). ¹³C NMR (CDCl₃, 151 MHz): δ=23.4 (t, C(4)), 87.0 (s, C(6)), 123.6, 125.7, 126.3, 126.4, 127.0, 127.5, 127.7, 128.1, 128.3, 128.6, 128.67, 128.68, 129.3, 129.8, 133.0 (15 d, Ar, =CH), 133.1, 133.3, 133.4, 136.0, 140.0 (5 s, Ar), 152.2 ppm (s, C(3)). IR: $\tilde{\nu }$ =3054m, 2914m, 1582m, 1489m, 1443m, 1392m, 1222m, 1097m, 955s, 916s, 887m, 752s, 745s, 708s, 693vs, 689vs cm⁻¹. HRMS (ESI‐TOF): m/z [M+H]⁺ calcd for C₂₇H₂₂NOS: 408.1422; found: 408.1426.

(E)‐3,6‐Diphenyl‐6‐[2‐(ferrocenyl)vinyl]‐4H‐1,5,2‐oxathiazine (11 e): reaction time: 24 h; yield: 250 mg (54 %); pale orange crystals, m.p. 155–156 °C (decomp.). ¹H NMR (CDCl₃, 600 MHz): δ=3.44, 3.46 (AB system, J=17.4 Hz, 2 H, H₂C(4)), 4.05 (s, 5 H, Fc), 4.23–4.25 (m, 2 H, Fc), 4.34–4.37 (m, 2 H, Fc), 6.11, 6.50 (AB system, J=15.7 Hz, 2 H, 2=CH), 7.33–7.42, 7.64–7.66, 7.70–7.74 ppm (3 m, 6 H, 2 H, 2 H, Ph). ¹³C NMR (CDCl₃, 151 MHz): δ=23.5 (t, C(4)), 67.1, 67.5, 69.2*, 69.3* (4 d, Fc), 81.1 (s, Fc), 87.0 (s, C(6)), 125.6, 125.9, 126.9, 128.54, 128.55, 128.6, 129.8, 131.9 (8 d, Ph,=CH), 136.0, 140.3 (2 s, Ph), 151.7 ppm (s, C(3)); *signal with higher intensity. IR: $\tilde{\nu }$ =3086m, 2888m, 1654m, 1588m, 1446m, 1399m, 1224m, 969s, 887m, 754s, 693vs cm⁻¹. HRMS (ESI‐TOF): m/z [M+H]⁺ calcd for C₂₇H₂₄FeNOS: 466.0928; found: 466.0929.

(E)‐3,6‐Diphenyl‐6‐[2‐(furan‐2‐yl)vinyl]‐4H‐1,5,2‐oxathiazine (11 f): reaction time: 60 min; yield: 145 mg (42 %); pale yellow crystals, m.p. 131–133 °C (decomp.). ¹H NMR (CDCl₃, 600 MHz): δ=3.41, 3.62 (AB system, J=17.4 Hz, 2 H, H₂C(4)), 6.29 (d_br, J≈3.3 Hz, 1 H, Fur), 6.36 (dd, J=1.8, 3.3 Hz, 1 H, Fur), 6.47, 6.52 (AB system, J=15.8 Hz, 2 H,=CH), 7.32–7.41, 7.60–7.62, 7.70–7.73 ppm (3 m, 7 H, 2 H, 2 H, Ph, Fur). ¹³C NMR (CDCl₃, 151 MHz): δ=23.4 (t, C(4)), 86.6 (s, C(6)), 110.1, 111.5, 120.9, 125.7, 126.8, 127.3, 128.62, 128.63, 128.64, 129.8 (10 d, Ph, Fur,=CH), 136.0, 139.9 (2 s, Ph), 142.7 (d, Fur), 151.4 (s, C(3)), 152.1 ppm (s, Fur). IR: $\tilde{\nu }$ =2905m, 2887m, 1573m, 1485m, 1444m, 1388m, 1246m, 1228m, 1088m, 989s, 954s, 822m, 733s, 688vs, 679vs cm⁻¹. MS (ESI): m/z (%)=370 (18, [M+Na]⁺), 348 (100, [M+H]⁺). Anal. Calcd for C₂₁H₁₇NO₂S (347.43): C 72.60, H 4.93, N 4.03, S 9.23; found: C 72.67, H 5.04, N 4.06, S 9.30.

(E)‐3,6‐Diphenyl‐6‐[2‐(thien‐2‐yl)vinyl]‐4H‐1,5,2‐oxathiazine (11 g): reaction time: 30 min; yield: 145 mg (40 %); pale yellow crystals, m.p. 151–152 °C (decomp.). ¹H NMR (CDCl₃, 600 MHz): δ=3.41, 3.61 (AB system, J=17.4 Hz, 2 H, H₂C(4)), 6.35, 6.84 (AB system, J=15.7 Hz, 2 H,=CH), 6.95 (dd, J=3.6, 5.1 Hz, 1 H, Thie), 6.99 (d_br, J≈3.6 Hz, 1 H, Thie), 7.20 (d_br, J≈5.1 Hz, 1 H, Thie), 7.32–7.42, 7.60–7.62, 7.70–7.73 ppm (3 m, 6 H, 2 H, 2 H, Ph). ¹³C NMR (CDCl₃, 151 MHz): δ=23.4 (t, C(4)), 86.7 (s, C(6)), 125.4, 125.8, 126.3, 126.9, 127.3, 127.5, 128.3, 128.64, 128.66, 128.69, 129.9 (11 d, Ph, Thie,=CH), 136.0, 139.9, 140.6 (3 s, Ph, Thie), 152.3 ppm (s, C(3)). IR: $\tilde{\nu }$ =2955m, 2927m, 2901m, 2847m, 1591m, 1493m, 1485m, 1444s, 1389m, 1276m, 1223m, 1197m, 1078m, 954s, 946s, 923s, 873m, 850m, 753s, 709s, 698vs, 689vs cm⁻¹. MS (ESI): m/z (%)=386 (12, [M+Na]⁺), 364 (100, [M+H]⁺). Anal. Calcd for C₂₁H₁₇NOS₂ (363.49): C 69.39, H 4.71, N 3.85, S 17.64; found: C 69.41, H 4.73, N 3.86, S 17.53.

(E)‐6‐Ferrocenyl‐3‐phenyl‐6‐styryl‐4H‐1,5,2‐oxathiazine (11 h): reaction time: 24 h; yield: 240 mg (52 %); beige crystals, m.p. 81–82 °C. ¹H NMR (CDCl₃, 600 MHz): δ=3.60, 3.67 (AB system, J=17.3 Hz, 2 H, H₂C(4)), 4.27 (m_c, 2 H, Fc), 4.31 (s, 5 H, Fc), 4.40–4.42, 4.46–4.48 (2 m, 1 H each, Fc), 6.51, 6.76 (AB system, J=15.9 Hz, 2 H, 2=CH), 7.27–7.30, 7.34–7.42, 7.44–7.47, 7.64–7.68 ppm (4 m, 1 H, 5 H, 2 H, 2 H, Ph). ¹³C NMR (CDCl₃, 151 MHz): δ=23.8 (t, C(4)), 66.9, 67.3, 68.5, 68.8, 69.4* (5 d, Fc), 84.8 (s, Fc), 88.6 (s, C(6)), 125.8, 126.8, 128.2, 128.3, 128.6, 128.7, 129.7, 131.2 (8 d, Ph,=CH), 135.9, 136.0 (2 s, Ph), 152.1 ppm (s, C(3)); * signal with higher intensity. IR: $\tilde{\nu }$ =3058m, 3026m, 2926m, 1578m, 1456m, 1444m, 1313m, 1221m, 967s, 906s, 818m, 725s, 689vs cm⁻¹. HRMS (ESI‐TOF): m/z [M+H]⁺ calcd for C₂₇H₂₄FeNOS: 466.0928; found: 466.0929.

(E)‐6‐Ferrocenyl‐3‐phenyl‐6‐[2‐(thien‐2‐yl)vinyl]‐4H‐1,5,2‐oxathiazine (11 i): reaction time: 24 h; yield: 260 mg (55 %); beige crystals, m.p. 102–104 °C. ¹H NMR (CDCl₃, 600 MHz): δ=3.58, 3.68 (AB system, J=17.3 Hz, 2 H, H₂C(4)), 4.26 (m_c, 2 H, Fc), 4.30 (s, 5 H, Fc), 4.40, 4.44 (2 s_br, 1 H each, Fc), 6.37, 6.88 (AB system, J=15.6 Hz, 2 H,=CH), 6.98 (dd, J=3.5, 4.9 Hz, 1 H, Thie), 7.01 (d_br, J≈3.5 Hz, 1 H, Thie), 7.21 (d_br, J≈4.9 Hz, 1 H, Thie), 7.39–7.42, 7.65–7.68 ppm (2 m, 3 H, 2 H, Ph). ¹³C NMR (CDCl₃, 151 MHz): δ=23.8 (t, C(4)), 66.9, 67.3, 68.6, 68.8, 69.4* (5 d, Fc), 84.5 (s, Fc), 88.4 (s, C(6)), 124.5, 124.9, 125.8, 127.0, 127.6, 127.9, 128.6, 129.8 (8 d, Ph, Thie,=CH), 136.0, 140.8 (2 s, Ph, Thie), 152.3 ppm (s, C(3)); *signal with higher intensity. IR: $\tilde{\nu }$ =3039m, 2957m, 2894m, 1493m, 1442m, 1378w, 1226m, 998s, 956s, 823m, 733s, 695s, 689vs cm⁻¹. MS (ESI): m/z (%)=472 (100, [M+H]⁺). Anal. Calcd for C₂₅H₂₁FeNOS₂ (471.41): C 63.69, H 4.49, N 2.97, S 13.60; found: C 63.87, H 4.73, N 2.84, S 13.60.

Quantum chemical calculations: Quantum chemical calculations (PBE1PBE/def2‐TZVP+PCM(dichloromethane)+GD3BJ dispersion correction)26, 27, 28, 29 were performed on the basis of preceding B3LYP/6‐31G(d)30+GD3BJ‐geometry optimizations using the Gaussian 09, Revision D.0133 and the Gaussian 16, Revision B.0134, package of programs. To obtain a most reliable structural information, several conformers of each isomer were investigated, in many cases after MM2‐conformational analysis. The transition‐state localizations were started with reaction‐path calculations by stepwise, independent elongation of both relevant bonds beginning with the cycloadducts (“retro‐Diels–Alder”) with full optimization of all other parameters. Then transition‐state searches or QST2 calculations on the basis of the obtained 3D‐hyperfaces followed. The s‐cis‐ as well as the s‐trans‐conformers of the reacting ene‐components were considered in the transition‐state searches; the respective energy lower isomeric transition states are represented in Scheme 6 and Scheme 10 and in the Supporting Information. We cannot exclude that due to the steric complexity of the reacting systems further transition‐state conformations and configurations exist. The explicit localization of bifurcations on the potential‐energy surfaces was beyond the scope of this experimentally oriented work⋅35 In many cases, IRC‐calculations were subsequently performed in order to characterize the respective stationary points.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supplementary

G. Mlostoń, K. Urbaniak, M. Jasiński, E.-U. Würthwein, H. Heimgartner, R. Zimmer, H.-U. Reissig, Chem. Eur. J. 2020, 26, 237.