The Synthesis and Base-Induced Breakdown of Triaryl 1,4-Oxathiins—An Experimental and DFT Study

1,4-Oxathiins are valued for a breadth of bioactivities and are known commercial fungicides. This article explores a novel preparation of 2,3,6-trisubstituted 1,4-oxathiin-S,S-dioxides via the reaction of benzyl 1-alkynyl sulfones and aryl aldehydes under basic conditions. A total of 20 examples possessing exclusively a trans arrangement of the 2,3-diaryl substituents are exhibited; the products demonstrate a variation of functional groups on the aryl ring attached to the heterocyclic ring system. The preparation is hindered by the base sensitivity of the products, and a ring-opened by-product typically contaminates the reaction mixture. A DFT assessment of the overall system includes a lithium counterion and offers possible pathways for the incorporation of the aldehyde, the cyclization step and the requisite proton transfers. In addition, the DFT work reveals options for the ring opening chemistry. It appears the trans 2,3-diaryl selectivity is set during the cyclization stage of the reaction sequence. The practical work uncovers a new reaction pathway to create a family of novel 1,4-oxathiin-S,S-dioxides whereas the computational work offers an understanding of the structures and possible mechanisms involved.

Of the protocols noted, the ring expansion chemistry of oxathiolanes, by a variety of approaches [30,[36][37][38], appears to be preferred when pursuing substrates with fungicidal and related bioactivities.Those compounds generally hold a -CH2CH2-opposite the alkene of the 1,4-oxathiin.More recent preparations have focused on adding ring substitution to create carbon stereocenters and key contributions are noted here.In work from the Cossy group, a variety of cis-2,6-disubstituted-1,4-oxathiane 4,4-dioxides were prepared with >98:2 diastereoselectivity by the Fe (III) catalyzed cyclization of substituted bis(2hydroxyethyl) sulfones [28].Grainger and coworkers employed -allyloxy and -propar- In an earlier communication [40], we introduced a fundamentally new method to assemble the 1,4-oxathiiin S,S-dioxide ring system bearing aryl groups at the 2, 3 and 6-positions of the ring and bearing the 2,3-aryl groups in trans stereochemistry (Figure 1B).Thus, the treatment of selected benzyl alkynyl sulfones (6) with base and an aldehyde led to the isolation of 1,4-oxathiin S,S-dioxides 5, although in low-to-moderate yield.The experimentation in that preliminary communication served as a proof-of-principle for the heterocycle assembly.Herein, we report a more detailed development and improved yield for the preparation of a family of aryl-substituted 1,4-oxathiin S,S-dioxides.We also probe the breakdown of the ring system, as observed under the reaction conditions, and employ computational chemistry to provide an understanding of the important pathways and intermediates for both ring formation and destruction.

Preparation of Substituted 1,4-Oxathiins
The previous communication introduced the cyclization reaction as represented with a specific example in Figure 1C.The products were exclusively trans (±)-2,3-dihydro-2,3,6triaryl-1,4-oxathiin, 4,4-dioxides (5).The structure of the ring system was confirmed by X-ray crystal analysis of a 2,3-di-(2-iodophenyl) analog [40].Other features indicative of the assigned structure includes a 1 H-1 H NMR coupling constant in the range of 11.3 to 11. 7 Hz for the 2-and 3-hydrogen resonances, in keeping with a diaxial configuration of the hydrogens.Those 2-and 3-hydrogen resonances appear on the 1 H NMR spectrum in the ranges of 5.81 to 6.48 ppm and 4.67 to 5.63 ppm, respectively.Also, consistent with the assigned structure is the lone vinylic resonance, appearing as a singlet at 5.85-6.51ppm.Overall, the reaction requires a benzyl anion attack of the added aldehyde and oxyanion cyclization onto the triple bond followed by substrate protonation to form the product.A probable mechanism is displayed in Scheme 1.A mechanistic and computational assessment of the pathways in the scheme will be discussed later in the paper (vide infra).
perimentation in that preliminary communication served as a proof-of-principle for the heterocycle assembly.Herein, we report a more detailed development and improved yield for the preparation of a family of aryl-substituted 1,4-oxathiin S,S-dioxides.We also probe the breakdown of the ring system, as observed under the reaction conditions, and employ computational chemistry to provide an understanding of the important pathways and intermediates for both ring formation and destruction.

Preparation of Substituted 1,4-Oxathiins
The previous communication introduced the cyclization reaction as represented with a specific example in Figure 1C.The products were exclusively trans (±)-2,3-dihydro-2,3,6triaryl-1,4-oxathiin, 4,4-dioxides (5).The structure of the ring system was confirmed by Xray crystal analysis of a 2,3-di-(2-iodophenyl) analog [40].Other features indicative of the assigned structure include a 1 H-1 H NMR coupling constant in the range of 11.3 to 11.7 Hz for the 2-and 3-hydrogen resonances, in keeping with a diaxial configuration of the hydrogens.Those 2-and 3-hydrogen resonances appear on the 1 H NMR spectrum in the ranges of 5.81 to 6.48 ppm and 4.67 to 5.63 ppm, respectively.Also, consistent with the assigned structure is the lone vinylic resonance, appearing as a singlet at 5.85-6.51ppm.Overall, the reaction requires a benzyl anion attack of the added aldehyde and oxyanion cyclization onto the triple bond followed by substrate protonation to form the product.A probable mechanism is displayed in Scheme 1.A mechanistic and computational assessment of the pathways in the scheme will be discussed later in the paper (vide infra).Scheme 1. Probable mechanism for oxathiin formation and breakdown.
Initial optimization protocols which led to our communicated results [40] were based primarily on the variation of the reaction temperature and on identity and equivalents of base, which eliminated NaH and KtOBu as viable options.The yield of 5a, the triphenyl Scheme 1. Probable mechanism for oxathiin formation and breakdown.
Initial optimization protocols which led to our communicated results [40] were based primarily on the variation of the reaction temperature and on identity and equivalents of base, which eliminated NaH and KtOBu as viable options.The yield of 5a, the triphenyl derivative reached a maximum yield of 54% [40].As it turned out, the chemistry leading to 5a proved anomalous within the domain of the initial evaluation conditions, and other examples afforded lower yields with very challenging chromatographic separation procedures, which usually incurred a loss of product.One problematic reaction component had already been identified as unreacted starting material.Another reaction constituent was characterized as an acyclic breakdown product of the 1,4-oxathiin, a recognized fate that is frequently [15,16,41], but not always [22], observed when 1,4-oxathiins are treated with base.In this case, the structure of the ring-opened product was eventually assigned to be keto sulfone 7Z (vide infra).
Given the problematic reaction constituents, additional optimization experiments were undertaken to further consume starting material while minimizing the formation of the by-product (7Z).Reactions were performed at temperatures previously established [40] but for different reaction times.Those experiments suggested additional concerns, wherein extended reaction times led to increased formation of intractable material.A quick chromatography of those mixtures was performed to remove only the intractable reaction constituents.Mass recovery and 1 H NMR spectrum peak integration provided ratios of the starting material, product 5a and by-product 7Z (Supplementary Materials, Table S1).From this analysis, six hours of reaction time offered the largest yield of the oxathiin, and that time was adopted for future reactions.
Recognizing that the base would ideally act in a catalytic manner (Scheme 1), additional experiments probed the efficiency of the heterocycle preparation using various equivalents of nBuLi.In those experiments, all with 2 eq. of added aldehyde, it was learned that 0.5 equiv. of base gave the best yield.Finally, a reduction of the equivalents of aldehyde improved the yield slightly more, while simplifying the constituency of the reaction mixtures (Table S2).With improved optimization and viable reaction conditions, there remained the issue of purification and isolation of the product.In TLC experiments and under subsequent flash chromatography conditions, the three reaction constituents created a difficult separation exercise.The product would often crystallize on the column thereby incurring greater losses and this outcome prompted alternative perspectives on the purification issue.
Although the oxathiins demonstrated a propensity to crystallize, a wide variety of recrystallization protocols failed to provide satisfactory results.Those recrystallization trials led to the realization that the crude reaction mixture could be triturated with warm ether to dissolve the unwanted keto-sulfone by-product (7) and leftover reactants while the oxathiin product resisted dissolution.The 1,4-oxathiin comprised the solid left behind and was found to be analytically pure in many cases.The ether solution was monitored for the product in each case and occasionally, small amounts of excess product were recovered through crystallization from the triturant.Another purification method was also applied in some instances.A flash column chromatography purification was initiated with a polarity that permitted crystallization of the product on the column while flushing unwanted materials.The oxathiin product was then eventually recovered in pure form by increasing the polarity of the elution solvent.The established reaction conditions with one or both purification protocols were applied to several examples, with the isolated yields of oxathiins 5 indicated in Table 1.The reactions were run on a 1-4 mmol scale.Entries #1-3 of Table 1 suggest the chemistry can be taken to a gram scale with no concerns.The examples demonstrate a variety of substituent patterns are possible on the benzyl portion of the starting benzyl alkynyl sulfone, as cyclized products are obtained in moderate to good yields (Table 1 #1-10).An assessment of substituent effects on reaction efficiency is hindered by the non-quantitative purification modes and the competing decomposition of the product.Substrates 5q-t were contaminated by a minor (≤10%) rotational isomer, presumably due to the steric demands of the 2-iodophenyl group situated on C3 between the SO 2 and (hetero)aryl group at the C2-position.
A reasonable mechanism for the product formation begins with full consumption of the sub-stoichiometric strong base (nBuLi or LDA) by deprotonation at the benzylic position of the starting material (Scheme 1).The consequent anion reacts with added aldehyde, to afford an alkoxide well positioned to undergo a 6-endo-dig conjugate addition onto the triple bond, creating a vinylic anion α to the sulfone.The cyclization may be reversible, but the cyclized form leads to the product through bimolecular proton transfer from another source, such as unreacted starting material or the carbon-3 of the product.This latter fate is presumed to lead to the undesired breakdown of the product.Variation of the aldehyde reactant created increased variability in the isolated yields (Table 1 # [11][12][13][14][15][16][17].Heterocyclic furfural and thiophene-2-carboxaldehyde gave reduced yields (#17, 20); pyridine-2-carboxaldehyde did not yield any cyclized products.Aliphatic aldehydes with acidic -hydrogens did not participate in the chemistry, presumably because of unwanted proton transfer and not aldehyde attack.Similarly, cinnamaldehyde and benzophenone also proved unsuitable.Use of 1-propynyl or 3,3-dimethyl-1-propynyl group as the starting alkyne component afforded product in only fair yields.
A reasonable mechanism for the product formation begins with full consumption of the sub-stoichiometric strong base (nBuLi or LDA) by deprotonation at the benzylic position of the starting material (Scheme 1).The consequent anion reacts with added aldehyde, to afford an alkoxide well positioned to undergo a 6-endo-dig conjugate addition onto the triple bond, creating a vinylic anion  to the sulfone.The cyclization may be reversible, but the cyclized form leads to the product through bimolecular proton transfer from another source, such as unreacted starting material or the carbon-3 of the product.This latter fate is presumed to lead to the undesired breakdown of the product.
As indicated, a ring-opened by-product typically contaminated the reaction mixture.The byproduct was assumed to originate from the deprotonation of the product, perhaps as a consequence of bimolecular proton transfers, followed by ring opening.During the early stages of the cyclization studies, the structure was assigned to be either 7E or 7Z since the availability of only one isomer did not permit a specific assignment.To pin down  Variation of the aldehyde reactant created increased variability in the isolated yields (Table 1 # [11][12][13][14][15][16][17].Heterocyclic furfural and thiophene-2-carboxaldehyde gave reduced yields (#17, 20); pyridine-2-carboxaldehyde did not yield any cyclized products.Aliphatic aldehydes with acidic -hydrogens did not participate in the chemistry, presumably because of unwanted proton transfer and not aldehyde attack.Similarly, cinnamaldehyde and benzophenone also proved unsuitable.Use of 1-propynyl or 3,3-dimethyl-1-propynyl group as the starting alkyne component afforded product in only fair yields.
A reasonable mechanism for the product formation begins with full consumption of the sub-stoichiometric strong base (nBuLi or LDA) by deprotonation at the benzylic position of the starting material (Scheme 1).The consequent anion reacts with added aldehyde, to afford an alkoxide well positioned to undergo a 6-endo-dig conjugate addition onto the triple bond, creating a vinylic anion  to the sulfone.The cyclization may be reversible, but the cyclized form leads to the product through bimolecular proton transfer from another source, such as unreacted starting material or the carbon-3 of the product.This latter fate is presumed to lead to the undesired breakdown of the product.
As indicated, a ring-opened by-product typically contaminated the reaction mixture.The byproduct was assumed to originate from the deprotonation of the product, perhaps as a consequence of bimolecular proton transfers, followed by ring opening.During the early stages of the cyclization studies, the structure was assigned to be either 7E or 7Z since the availability of only one isomer did not permit a specific assignment.To pin down  Variation of the aldehyde reactant created increased variability in the isolated yields (Table 1 # [11][12][13][14][15][16][17].Heterocyclic furfural and thiophene-2-carboxaldehyde gave reduced yields (#17, 20); pyridine-2-carboxaldehyde did not yield any cyclized products.Aliphatic aldehydes with acidic -hydrogens did not participate in the chemistry, presumably because of unwanted proton transfer and not aldehyde attack.Similarly, cinnamaldehyde and benzophenone also proved unsuitable.Use of 1-propynyl or 3,3-dimethyl-1-propynyl group as the starting alkyne component afforded product in only fair yields.
A reasonable mechanism for the product formation begins with full consumption of the sub-stoichiometric strong base (nBuLi or LDA) by deprotonation at the benzylic position of the starting material (Scheme 1).The consequent anion reacts with added aldehyde, to afford an alkoxide well positioned to undergo a 6-endo-dig conjugate addition onto the triple bond, creating a vinylic anion  to the sulfone.The cyclization may be reversible, but the cyclized form leads to the product through bimolecular proton transfer from another source, such as unreacted starting material or the carbon-3 of the product.This latter fate is presumed to lead to the undesired breakdown of the product.
As indicated, a ring-opened by-product typically contaminated the reaction mixture.The byproduct was assumed to originate from the deprotonation of the product, perhaps as a consequence of bimolecular proton transfers, followed by ring opening.During the early stages of the cyclization studies, the structure was assigned to be either 7E or 7Z since the availability of only one isomer did not permit a specific assignment.To pin down  Variation of the aldehyde reactant created increased variability in the isolated yields (Table 1 # [11][12][13][14][15][16][17].Heterocyclic furfural and thiophene-2-carboxaldehyde gave reduced yields (#17, 20); pyridine-2-carboxaldehyde did not yield any cyclized products.Aliphatic aldehydes with acidic -hydrogens did not participate in the chemistry, presumably because of unwanted proton transfer and not aldehyde attack.Similarly, cinnamaldehyde and benzophenone also proved unsuitable.Use of 1-propynyl or 3,3-dimethyl-1-propynyl group as the starting alkyne component afforded product in only fair yields.
A reasonable mechanism for the product formation begins with full consumption of the sub-stoichiometric strong base (nBuLi or LDA) by deprotonation at the benzylic position of the starting material (Scheme 1).The consequent anion reacts with added aldehyde, to afford an alkoxide well positioned to undergo a 6-endo-dig conjugate addition onto the triple bond, creating a vinylic anion  to the sulfone.The cyclization may be reversible, but the cyclized form leads to the product through bimolecular proton transfer from another source, such as unreacted starting material or the carbon-3 of the product.This latter fate is presumed to lead to the undesired breakdown of the product.
As indicated, a ring-opened by-product typically contaminated the reaction mixture.The byproduct was assumed to originate from the deprotonation of the product, perhaps as a consequence of bimolecular proton transfers, followed by ring opening.During the early stages of the cyclization studies, the structure was assigned to be either 7E or 7Z since the availability of only one isomer did not permit a specific assignment.To pin down As indicated, a ring-opened by-product typically contaminated the reaction mixture.The byproduct was assumed to originate from the deprotonation of the product, perhaps as a consequence of bimolecular proton transfers, followed by ring opening.During the early stages of the cyclization studies, the structure was assigned to be either 7E or 7Z since the availability of only one isomer did not permit a specific assignment.To pin down the structure, experiments were undertaken to probe the ring opening outcomes through deliberate α-sulfonyl deprotonation of oxathiin 5a.
Table S3 lists several experiments in which oxathiin 5a was treated with different bases and reaction conditions ranging from −35 • C to room temperature.The outcomes were consistently 7-25% of one double bond isomer and 75-93% of the other and the pair of isomers (7E/Z) could not be separated by preparative chromatography.The major product demonstrated 1 H NMR resonances that matched those observed for the by-product of the 5a preparation.However, the geometries of both isomers were still not established.
During the evaluation of various bases, it was observed that the isomeric ratio of the product demonstrated the most variability when DBU was employed.Recognizing that this tertiary base could prompt isomerization through reversible conjugate addition on the ring-opened sulfonyl ketone, the effects of this base were specifically pursued.Heating an isomeric mixture of ring-opened products with 3 equiv.of DBU in THF delivered a conversion to a single isomer after <72 h (Table 2).With a pure isomer in hand, and with a ~85% pure sample of the other isomer from the ring-opening trials, these compounds were differentiated through NOESY spectroscopy experiments (Supplementary Materials) [42,43].The major product of ring opening possessed the Z-geometry, whereas the DBU experiments revealed the E-isomer as the thermodynamically more stable alkene.As indicated, the isomerization presumably occurs by reversible conjugate addition of DBU to Z-isomer, single bond rotation and base release.It must be mentioned that the isomerization process may occur on the enolate form of the substrate, in keeping with harsher conditions required for isomerization.The observations suggest the mode of byproduct formation during the oxathiin synthesis is a kinetically controlled ring opening and reaction conditions do not bring about any isomerization of the byproduct.this tertiary base could prompt isomerization through reversible conjugate addition on the ring-opened sulfonyl ketone, the effects of this base were specifically pursued.Heating an isomeric mixture of ring-opened products with 3 equiv.of DBU in THF delivered a conversion to a single isomer after <72 h (Table 2).With a pure isomer in hand, and with a ~85% pure sample of the other isomer from the ring-opening trials, these compounds were differentiated through NOESY spectroscopy experiments (Supp.Mat.) [42,43].The major product of ring opening possessed the Z-geometry, whereas the DBU experiments revealed the E-isomer as the thermodynamically more stable alkene.As indicated, the isomerization presumably occurs by reversible conjugate addition of DBU to Z-isomer, single bond rotation and base release.It must be mentioned that the isomerization process may occur on the enolate form of the substrate, in keeping with harsher conditions required for isomerization.The observations suggest the mode of byproduct formation during the oxathiin synthesis is a kinetically controlled ring opening and reaction conditions do not bring about any isomerization of the byproduct.

Computational Chemistry
Computational analyses of the synthetic and degradative pathways were performed with the Gaussian 16/Gaussview 6 software package [44] at 298 °C using the M06-2x density functional, employed in the past for the chemistry of -sulfonyl anions [45].Calculations were all performed on the Graham computer cluster located at the University of Waterloo as part of the Digital Research Alliance of Canada.The scratch file system where calculations were carried out is a 3.2PB parallel high-performance filesystem.The Graham cluster is 41,548 cores and 520 GPU devices.We adopted a split valence Pople basis set [6-311 + G(d,p)] with diffuse functions on non-hydrogenic atoms.The M06 functionals have been shown to provide more accurate energies and geometries than the common B3LYP functional with conformationally flexible anions [46].Optimization and frequency calculations were carried out on reactants, pre-complexes, transition states and products for both the synthetic and degradative pathway using a CPCM solvation model with THF as

Computational Chemistry
Computational analyses of the synthetic and degradative pathways were performed with the Gaussian 16/Gaussview 6 software package [44] at 298 • C using the M06-2x density functional, employed in the past for the chemistry of α-sulfonyl anions [45].Calculations were all performed on the Graham computer cluster located at the University of Waterloo as part of the Digital Research Alliance of Canada.The scratch file system where calculations were carried out is a 3.2PB parallel high-performance filesystem.The Graham cluster is 41,548 cores and 520 GPU devices.We adopted a split valence Pople basis set [6-311 + G(d,p)] with diffuse functions on non-hydrogenic atoms.The M06 functionals have been shown to provide more accurate energies and geometries than the common B3LYP functional with conformationally flexible anions [46].Optimization and frequency calculations were carried out on reactants, pre-complexes, transition states and products for both the synthetic and degradative pathway using a CPCM solvation model with THF as the solvent.To minimize computational cost, explicit solvation was excluded from the calculations due to the computational intensity of the optimizations induced by the size of the molecular system.Transition state structures possessed a single imaginary frequency and were further validated by intrinsic reaction coordinate (IRC) calculations and/or analysis of the displacement vectors [47].The vibrational mode for that imaginary frequency was fully consistent with the atomic motion anticipated for the reaction under analysis.

Oxathiin Ring Formation
Past modelling studies of α-sulfonyl anion chemistry typically account for the metal counterion [48][49][50][51][52][53], although there are exceptions [45,54,55].This assessment engages a lithium counterion with the α-sulfonyl anion, as part of the transition state for addition to the aldehyde and for the chemistry thereafter.It is noted that one report employing a constrained, cyclic α-sulfonyl anion found an energetically favorable transition state that did not include lithium stabilization of the oxyanion forming during attack of aldehyde [48].In addition to employing a lithium counterion, for computational ease it was decided to model the α-sulfonyl carbanion chemistry as a monomeric entity, which has been observed among the known speciation forms in solution [56].Three optimized structures of the lithiated sulfone (8 a-c ) were identified, differing in the position of the lithium counterion.
The lowest energy structure (8 c ) had lithium as a bridging ion between the sulfonyl oxygens.Structures 8 a and 8 b were 0.2 and 0.8 kcal/mol higher in energy, respectively, and are rotational isomers about the anionic C-S bond (Figure 2).The narrow energy range is consistent with past contributions, which also report a small energetic gap among the various monomeric lithiated sulfone species [48,56].
lithium counterion with the -sulfonyl anion, as part of the transition state for addition to the aldehyde and for the chemistry thereafter.It is noted that one report employing a constrained, cyclic -sulfonyl anion found an energetically favorable transition state that did not include lithium stabilization of the oxyanion forming during attack of aldehyde [48]..In addition to employing a lithium counterion, for computational ease it was decided to model the -sulfonyl carbanion chemistry as a monomeric entity, which has been observed among the known speciation forms in solution [56].Three optimized structures of the lithiated sulfone (8a-c) were identified, differing in the position of the lithium counterion.The lowest energy structure (8c) had lithium as a bridging ion between the sulfonyl oxygens.Structures 8a and 8b were 0.2 and 0.8 kcal/mol higher in energy, respectively, and are rotational isomers about the anionic C-S bond (Figure 2).The narrow energy range is consistent with past contributions, which also report a small energetic gap among the various monomeric lithiated sulfone species [48,56].It remained a challenge to determine a logical starting lithiated sulfone for the attack of benzaldehyde.Further complicating matters was the expectation of numerous precomplexes between the aldehyde oxygen and any of the -lithiated sulfonyl species.To solve this issue, reasonable transition states with Li bridging the sulfonyl and carbonyl oxygens were located and intrinsic reaction coordinate analysis was used to pinpoint reactive aldehyde-lithiated sulfone precomplexes.Structures were then fully set with a final frequency/optimization; all precomplexes identified involved a form most closely related to either 8a or 8b.
A total of eight precomplexes (PCa-PCh) result from this approach, with two exemplified in Figure 3.With Li complexed to one sulfonyl oxygen of 8a, an attack of aldehyde may occur on the si or re face.Two more structures result when the Li is complexed to the other sulfonyl oxygen of 8a.The same alignments on 8b create the other four isomers.Precomplexes PCa,c,e,g eventually lead to the formation of the cis oxathiin (5cis) and similarly, pre-complexes PCb,d,f,h precede the formation of the trans oxathiin (5trans).Images of the precomplexes (PCa-h), the transition states for C2-C3-bond formation (PCa-h ‡ ), and the products of the addition (9a-h) are depicted in Figures S1-S3.The relative free energies of these structures are compiled diagrammatically in a linear reaction style in Figure 4, whereas Figure S4 separates the same information into cis and trans PES manifolds.Figure 3 shows exemplary structures for the PCa to 9a and PCd to 9d pathways, the lowest energy options It remained a challenge to determine a logical starting lithiated sulfone for the attack of benzaldehyde.Further complicating matters was the expectation of numerous precomplexes between the aldehyde oxygen and any of the α-lithiated sulfonyl species.To solve this issue, reasonable transition states with Li bridging the sulfonyl and carbonyl oxygens were located and intrinsic reaction coordinate analysis was used to pinpoint reactive aldehyde-lithiated sulfone precomplexes.Structures were then fully set with a final frequency/optimization; all precomplexes identified involved a form most closely related to either 8 a or 8 b .
A total of eight precomplexes (PC a -PC h ) result from this approach, with two exemplified in Figure 3.With Li complexed to one sulfonyl oxygen of 8 a , an attack of aldehyde may occur on the si or re face.Two more structures result when the Li is complexed to the other sulfonyl oxygen of 8 a .The same alignments on 8 b create the other four isomers.Pre-complexes PC a,c,e,g eventually lead to the formation of the cis oxathiin (5 cis ) and similarly, pre-complexes PC b,d,f,h precede the formation of the trans oxathiin (5 trans ).Images of the precomplexes (PC a-h ), the transition states for C 2 -C 3 -bond formation (PC a-h ‡ ), and the products of the addition (9 a-h ) are depicted in Figures S1-S3.The relative free energies of these structures are compiled diagrammatically in a linear reaction style in Figure 4, whereas Figure S4 separates the same information into cis and trans PES manifolds.Figure 3 shows exemplary structures for the PC a to 9 a and PC d to 9 d pathways, the lowest energy options to 10 cis and 10 trans anions, respectively.At this point in the reaction, the Li counterion acts as a bridge between the oxygen of benzaldehyde and the sulfone creating a cyclic structure that permits the usage of cis and trans labels going forward along the reaction pathway.The cis/trans configurations are maintained through to the oxathiin products.
Bond lengths, atom-to-atom distances and dihedral angles for the precomplexes (PC a-h ), transition states (PC a-h ‡ ) and products of benzaldehyde addition can be found in Table S4 of the Supplementary Materials, alongside a discussion of key trends for these potential energy surfaces.Overall, the free energies of activation for the addition reaction are low (Figure 4 or Figure S4).Although other researchers did not probe precomplexes in their assessment of similar addition reactions [45,50], an intramolecular example of addition to a ketone displays an enthalpic barrier of 2.0 kcal/mol [45].The addition products range from 5.3 to 11.2 kcal/mol lower in energy from the precomplexed starting materials.The two lowest energy addition products 9 d and 9 f have the phenyl groups in a trans relationship.
to 10cis and 10trans anions, respectively.At this point in the reaction, the Li counterion acts as a bridge between the oxygen of benzaldehyde and the sulfone creating a cyclic structure that permits the usage of cis and trans labels going forward along the reaction pathway.The cis/trans configurations are maintained through to the oxathiin products.Bond lengths, atom-to-atom distances and dihedral angles for the precomplexes (PCa-h), transition states (PCa-h ‡ ) and products of benzaldehyde addition can be found in Table S4 of the Supplementary Materials, alongside a discussion of key trends for these potential energy surfaces.Overall, the free energies of activation for the addition reaction are low (Figure 4 or S4).Although other researchers did not probe precomplexes in their assessment of similar addition reactions [45,50], an intramolecular example of addition to a ketone displays an enthalpic barrier of 2.0 kcal/mol [45].The addition products range from 5.3 to 11.2 kcal/mol lower in energy from the precomplexed starting materials.The two lowest energy addition products 9d and 9f have the phenyl groups in a trans relationship.Formation of 10 (anionic cis or trans oxathiin) requires a conjugate addition of the oxyanion to the triple bond of the alkyne.However, only half of the benzaldehyde addition products (9 e -h ) have the alkyne in suitable proximity.The remaining four products (9 a -d ) require a change in coordination of Li to the opposite sulfone oxygen, with rotation of the C 3 -S bond to place the alkyne in a pseudo-axial position.For example, the structures of trans addition products 9 d and 9 f display Ph-C 2 -C 3 -Ph dihedral angles only 1.2 • in difference but differ structurally based on which sulfonyl oxygen is coordinated to the Li.However, in 9 d , migration of the Li the other sulfonyl oxygen and rotation around the C 3 -S is required to position the alkyne in an axial position, creating 9 f , a change which facilitates cyclization (Figure 5).The relative energies of these products only differ by 0.8 kcal/mol (Figures 4 and 5).Similar relationships can be identified between products 9 a /9 e , 9 b /9 h and 9 c /9 g .These pairs show similar dihedral angles and bond lengths, so migration of Li sulfonyl oxygen to the other, with concomitant C 3 -S bond rotation, would essentially achieve their interconversion.The difference in free energy between these pairs is 0.7, 0.2 and 2.5 kcal/mol, respectively (Figures 3 and 4).To confirm the feasibility of the requisite interconversion, a transition state was sought for the C 3 -S bond rotation converting 9 d to 9 f , but optimizations never achieved convergence.However, as proof of principle, many of these transition state attempts are optimized to structures that position the alkyne in an axial position making it accessible to the oxyanion for the conjugate addition.
With the assumption these conformational changes occur as a prerequisite for cyclization, four products of benzaldehyde addition remain: 9 e and 9 g as cis adducts; 9 f and 9 h as trans adducts.While possessing identical stereogenic carbons defining them as cis adducts, 9 e and 9 g are diastereomers with sulfur as an additional stereogenic center due to Li coordination to one oxygen.By analogy with the discussion in the paragraph above, structure 9 g is expected to succumb to C 3 -S bond rotation, lithium repositioning and C 2 -C 3 bond rotation to convert to 9 e for a free energy benefit of 3.8 kcal/mol.Trans adducts 9 f and 9 h have a similar relationship wherein structure 9 h would gain 2.4 kcal/mol of free energy becoming 9 f after C 3 -S bond rotation and lithium repositioning and a C 2 -C 3 bond rotation that brings the C 2 -C 3 phenyl groups from diaxial to diequatorial.
Compound 9 e was found to be the anionic precursor to cis-oxathiin formation and is renamed 9 cis .Entity 9 cis may also be accessible by direct access from other cis structures 9 a , 9 c and 9 g .Correspondingly, any of 9 b , 9 d , 9 f and 9 h could lead to 9 trans , the precursor to trans-oxathiin formation.The preceding discussion outlines reasonable structure changes culminating in 9 f which is nearly identical to 9 trans in terms of bond lengths and shows only minor variation in the dihedral angles based on the C 2 -C 3 bond (Table S5).Hence, a very minor but energetically favorable structural change converts 9 f to 9 trans .Formation of 10 (anionic cis or trans oxathiin) requires a conjugate addition of the oxyanion to the triple bond of the alkyne.However, only half of the benzaldehyde addition products (9e-h) have the alkyne in suitable proximity.The remaining four products of Li sulfonyl oxygen to the other, with concomitant C3-S bond rotation, would essentially achieve their interconversion.The difference in free energy between these pairs is 0.7, 0.2 and 2.5 kcal/mol, respectively (Figures 3 and 4).To confirm the feasibility of the requisite interconversion, a transition state was sought for the C3-S bond rotation converting 9d to 9f, but optimizations never achieved convergence.However, as proof of principle, many of these transition state attempts are optimized to structures that position the alkyne in an axial position making it accessible to the oxyanion for the conjugate addition.With the assumption these conformational changes occur as a prerequisite for cyclization, four products of benzaldehyde addition remain: 9e and 9g as cis adducts; 9f and 9h as trans adducts.While possessing identical stereogenic carbons defining them as cis adducts, 9e and 9g are diastereomers with sulfur as an additional stereogenic center due to Li coordination to one oxygen.By analogy with the discussion in the paragraph above, structure 9g is expected to succumb to C3-S bond rotation, lithium repositioning and C2-C3 bond rotation to convert to 9e for a free energy benefit of 3.8 kcal/mol.Trans adducts 9f and 9h have a similar relationship wherein structure 9h would gain 2.4 kcal/mol of free energy becoming 9f after C3-S bond rotation and lithium repositioning and a C2-C3 bond rotation that brings the C2-C3 phenyl groups from diaxial to diequatorial.
Compound 9e was found to be the anionic precursor to cis-oxathiin formation and is renamed 9cis.Entity 9cis may also be accessible by direct access from other cis structures 9a, 9c and 9g.Correspondingly, any of 9b, 9d, 9f and 9h could lead to 9trans, the precursor to transoxathiin formation.The preceding discussion outlines reasonable structure changes culminating in 9f which is nearly identical to 9trans in terms of bond lengths and shows only minor variation in the dihedral angles based on the C2-C3 bond (Table S5).Hence, a very minor but energetically favorable structural change converts 9f to 9trans.
Changes in bond lengths and dihedral angles for the cyclization processes forming 10cis and 10trans can be found in Table 3; optimized structures are depicted in Figure 6.Attack of the oxyanion on C6 of the alkyne results in a significant decrease in the distance between Ob and C6 with final bond lengths of 1.426 Å and 1.423 Å for the cis and trans products, respectively.This conjugate addition onto the alkyne is accompanied by a lengthening of the C5-C6 bond to a value of 1.344 Å, and a C2-Ob bond to 1.434 Å in both structures.Loss of charge on Ob results in decreased association with Li, wherein the Li-Ob distance increases to 2.152 Å in the trans product and 2.287 Å in the cis product.This leads to increased association with the oxygen of the sulfone with atom-to-atom distance dropping by approximately 0.7 Å.As a result of cyclization, a formal charge is passed to Changes in bond lengths and dihedral angles for the cyclization processes forming 10 cis and 10 trans can be found in Table 3; optimized structures are depicted in Figure 6.Attack of the oxyanion on C 6 of the alkyne results in a significant decrease in the distance between O b and C 6 with final bond lengths of 1.426 Å and 1.423 Å for the cis and trans products, respectively.This conjugate addition onto the alkyne is accompanied by a lengthening of the C    C6, contributing to a shift in the position of the Li counterion somewhat closer to that of the anionic carbon.Relative free energy changes for cyclization are generally more significant than the prior benzaldehyde addition.The pre-cyclized compounds 9cis and 9trans are separated by 1.7 kcal/mol with the trans structure lower in energy.Activation barriers for the cis and trans pathways are 17.4 and 16.4 kcal/mol, respectively, with a ΔΔG ‡ of 2.7 kcal/mol.The Relative free energy changes for cyclization are generally more significant than the prior benzaldehyde addition.The pre-cyclized compounds 9 cis and 9 trans are separated by 1.7 kcal/mol with the trans structure lower in energy.Activation barriers for the cis and trans pathways are 17.4 and 16.4 kcal/mol, respectively, with a ∆∆G ‡ of 2.7 kcal/mol.The final products of these pathways, 10 cis and 10 trans , also show a significant energetic difference wherein the trans compound sits 3.1 kcal/mol lower than its cis counterpart.Hence, for the cyclization step leading to compound 10, the trans pathway not only delivers a product of greater thermodynamic stability but does so with a lower free energy of activation.
Immediate cyclization products 10 cis and 10 trans possess the lithium still in close proximity to O b .It was expected that more stable forms would hold the Li nearer the sulfonyl oxygens and the vinylic carbon (C 6 ).Such structures (10 cis' and 10 trans' ) were located as local minima and were found to possess identical free energies, substantially lower than 10 cis and 10 trans .Transition states for the conversion of species 10 to 10 could not be located.A proton source can react with any of 10 or 10 to provide neutral oxathiin, although a reaction with 10 may be expected to be faster based on its higher energy (Figure 7).Starting material 6 is logically suggested to be that proton source, as presented in Figure 7, but other options are also possible.A more complete discussion is presented after an outline of the computed ring-opening chemistry.

Oxathiin Ring Opening
The deliberate ring openings of 2,3-dihydro-1,4-oxathiin-4,4-dioxides [41] and some aryl-substituted benzo derivatives [16] have been evaluated for their synthetic utility, as the reaction unveils a reactive vinylic sulfone.In contrast, a saturated form of the ring system survives the double deprotonation protocol necessary for a successful Ramburg-Backlund protocol delivering the 2,5-dihydrofuran core required for the successful synthesis of (+)-solamin [22].In the work presented herein, we have observed the ring opening at an undesirable juncture during oxathiin preparation and when deliberately pursuing it (vide supra).Our practical findings grant partial understanding of the E vs Z product preferences in the ring opening.Hence, the possible ring opening mechanisms were explored computationally to help understand the formation products 7.
Given that deprotonated/lithiated structures of 5 would produce the ring-opened forms, a comprehensive search revealed eight optimized geometries (11a-h) for proton loss from the -sulfonyl, benzylic position (C3) that gave valid PESs for ring opening (The a, b,

Oxathiin Ring Opening
The deliberate ring openings of 2,3-dihydro-1,4-oxathiin-4,4-dioxides [41] and some aryl-substituted benzo derivatives [16] have been evaluated for their synthetic utility, as the reaction unveils a reactive vinylic sulfone.In contrast, a saturated form of the ring system survives the double deprotonation protocol necessary for a successful Ramburg-Backlund protocol delivering the 2,5-dihydrofuran core required for the successful synthesis of (+)-solamin [22].In the work presented herein, we have observed the ring opening at an undesirable juncture during oxathiin preparation and when deliberately pursuing it (vide supra).Our practical findings grant partial understanding of the E vs Z product preferences in the ring opening.Hence, the possible ring opening mechanisms were explored computationally to help understand the formation of decomposition products 7.
Given that deprotonated/lithiated structures of 5 would produce the ring-opened forms, a comprehensive search revealed eight optimized geometries (11 a-h ) for proton loss from the α-sulfonyl, benzylic position (C 3 ) that gave valid PESs for ring opening (The a , b , c designation for intermediates 11 bears no correlation to earlier usage of these letters for other compounds.).See Figure S5 for all eight PES pathways; Figure S6 shows lithiated oxathiin structures and is accompanied by some comments.Table S6 offers bond lengths and angles of optimized geometries through the ring-opening process.Given that we only observed isomer 12Z in our work, structures producing 12Z are the principal focus of this discussion.Those pathways involve 11 a-d and are incorporated into Figure 7. Lithiated species 11 a,b,d were ostensibly derived from cis oxathiin, meaning the proton was replaced with Li + before structure optimization.Structure 11c originated from the trans oxathiin, but the structural evolution of all precursors for ring opening during optimization created anionic structures unrelated to initial lithiated starting materials.
Lithiated structures 11 a-d are energetically accessible via proton removal from 5 cis or 5 trans via an isodesmic pathway that engages lithiated benzyl anion 8 c and releases neutral starting material 6.This is presented as an option in Figure 7 as one possible reaction for proton loss from isomers 5. Compounds 11 a-d all ring open with free energy barriers of 4.8-6.8kcal/mol, before reaching lithiated forms of 12Z (12Z a-d ) in a highly exergonic process.These energies are consistent with facile ring-opening energies of multiple structures and there is little to distinguish the four pathways.Ring-opening forms 12Z c/d are lower in energy than 12Z a/b , presumably due to the position of the Li atom, which bridges carbonyl and sulfonyl oxygens in the 12Z c/d ; all ring-opened lithiated species bring stable 12Z upon reaction quenching.
The preceding discussion does not address the fate of anions 11 e-h since final product 12E is not observed.However, it should be mentioned that the formation of 12E e from 11 e may represent a valid PES.That pathway delivers a thermodynamically more stable product (12E e vs 12Z c /12Z d , all with bridging Li), but has a much higher barrier (Figure S5); for energetic reasons, this is a non-competitive pathway for ring opening.Finally, consistent with the computational findings of the reactive intermediates and practical outcomes, alkene 7E was computed to be 6.4 kcal/mol more stable than 7Z.

Proton Transfer Options and Overall Pathway
The collection of starting options as shown in Figure 4 suggests several similar and reversible pathways are available, culminating in the energetically preferred formation of 9 trans over 9 cis (1.7 kcal/mol).The cyclization pathway of these two entities also prefers the trans option (2.7 kcal/mol).These comparative PESs are consistent with the observed experimental selectivity of the formation of 5 trans .It is conceivable that additional ring opening options follow from 5 trans through 11 c to byproduct 12Z.If any 5 cis is formed, it was not observed and would have to be broken down to ring opening product.It should be noted that deprotonation of 5 cis with a compound such as 8c is an exergonic process.Reversibility within the all trans (or all cis) 10 /5/11 manifold is possible but there is no opportunity to interchange the trans-diaryl stereochemistry for the cis-diaryl configuration once the initial ring formation has occurred.
No effort was made to find transition states for proton delivery to anions 10/10 or removal from isomers 5.As stated, product 5 trans could be readily formed by protonation of vinylic anions 10 trans/ 10 trans' by starting material 6 or another available acidic carbon among the entities in solution.As starting material depletes, a second option for a proton may be product 5, creating one of the forms of 11, some of which will ring open to 12Z.Table S7 demonstrates the free energy change for most of the available proton transfer options.The most conspicuous data in Table S7 suggest that the reaction of 10 cis with the cis or trans forms of 5 are the most endergonic reaction depicted.The predicted free energies of reaction of 9.0-12.6kcal/mol suggest 5 cis may be more prone to loss of proton.Other entries in the table are consistent with Figure 7 in that anionic 10 trans/ 10 cis represent the most reactive deprotonating agents.If 10 trans reverts to 10 trans' , formation of product 5 trans is still slightly endergonic if proton transfer occurs with 5 cis (1.1-3.3 kcal/mol).Finally, proton transfers between products 5 and various forms of benzyl anions 8 are reasonable, based on free energies of reaction near 0 kcal/mol.

General Reaction Procedures
All anhydrous reactions were carried out in glassware that was under flame-dried vacuum (or dried overnight in a 110 • C oven).Thin layer chromatography (TLC) was achieved on glass-backed plates with Silica Gel 60 (250 µm) containing a fluorescent indicator.Compounds were visualized through UV and I 2 indicators.Flash column chromatography was carried out using silica gel supplied by Silicycle (particle size 30-63 (mesh 230-400)). 1H NMR and 13 C NMR spectra were recorded on Bruker Avance 300, Avance 400 or Avance 600 spectrometer.Chemical shift values (δ) are reported in ppm relative to CDCl 3 (δ 7.26 ppm) unless otherwise noted.The proton spectra are reported as follows: δ (multiplicity, coupling constant J, number of protons).Multiplicities are indicated by s (singlet), d (doublet), t (triplet), q (quartet) and m (multiplet).High-resolution mass spectrometry was carried out at Queen's University, ON, Canada.Starting sulfones 6 were prepared as described previously [40,57]; characterization data for new compounds 6 are found in the Supplementary Materials.Aldehydes were purified prior to use if required.

Preparation of Trisubstituted 1,4-Oxathiins
General Method for Oxathiin-S,S-Dioxides Under a dry, inert atmosphere, starting sulfone (1 to 4 mmol, 1.0 eq) was dissolved in THF (0.05 M) and cooled to −78 • C. nBuLi (1.6 M, 0.5 eq in hexanes) or LDA (0.88 M, 0.5 eq in THF) was added dropwise and stirred at −78 • C for 15 min.Aldehyde (1.5 eq) was added to the solution and the reaction mixture was warmed slowly to −35 • C.After 6 h of stirring, the reaction was quenched with NH 4 Cl (aq) and warmed to rt.The aqueous layer was extracted with EtOAc (3 × 15 mL).The combined organic layers were washed with brine, dried over MgSO 4 and the solvent was evaporated under reduced pressure.Purification Method A: The crude residue was triturated in warm anhydrous ethyl ether several time.The decanted mixture contained impurities and the pure product remained undissolved.TLC inspection of the collected ether washes revealed if either a flash chromatography column or a 2nd trituration of the concentrated triturant was merited to obtain more product.Purification Method B: The crude residue was subjected to flash chromatography using 1:149:150 MeOH:DCM:hexanes (v/v/v) solvent system.After recovery of unwanted compounds, a flush with 10→30% MeOH in DCM afforded pure oxathiin.

Conclusions
This paper introduces details of a conceptually new synthetic protocol to assemble the trans-(±)-2,3-dihydro-2,3,6-triaryl-1,4-oxathiin, S,S-dioxide ring system.The protocol requires deprotonation of the benzyl position of benzyl alkynyl sulfones employing substoichiometric amounts of strong base.The anion so formed undergoes nucleophilic attack of an added aryl aldehyde and the ensuing alkoxy anion performs an intramolecular conjugate addition on the proximal triple bond of the original starting material.The protocol delivers fair to good yields of 2,3,6 trisubstuted-1,4-oxathiins, many in analytically pure form.The reaction is hindered by breakdown of the products under the reaction conditions, affording a kinetically formed keto sulfone, the presence of which reduces the yield of oxathiin and may hinder purification of the oxathiin product.
The preferred 1,4-oxathiin preparation for applications in pesticide science seems to be ring expansion chemistry of oxathiolanes.More recent work has demonstrated alternative protocols that introduce stereochemistry on the saturated 2-and 3-carbons.Our 1,4-oxathiin preparation complements those existing protocols for oxathiin synthesis in that it represents the only method for the placement of aryl groups at positions 2, 3 and 6.Moreover the relative stereochemistry at positions 2 and 3 is reliably set as trans during the chemistry.The preparation unveils a new library of highly substituted 1,4-oxathiins, which may hold value as bioactive compounds in a number of applications.From a strategy point of view, the methodology represents a previously unknown assembly mode for oxathiins and opens the door for additional variations on this reaction model.
A computational assessment of both oxathiin formation and breakdown reveal eight modes for benzyl anion attack of aryl aldehyde which converge to cis and trans addition products possessing an active alkoxide.The oxyanion can attack the pendant alkyne to

Figure 3 .
Figure 3. Computed structures for the addition of lithiated benzyl sulfones 8a/b to benzaldehyde.'a' series represents one of four cis addition modes.'d' series represents one of four trans addition modes.

Figure 3 .
Figure 3. Computed structures for the addition of lithiated benzyl sulfones 8 a/b to benzaldehyde.'a' series represents one of four cis addition modes.'d' series represents one of four trans addition modes.

Molecules 2023 , 21 Figure 4 .
Figure 4. Energy profile for aldehyde addition and eventual cis and trans oxathiin synthesis.

Figure 4 .
Figure 4. Energy profile for aldehyde addition and eventual cis and trans oxathiin synthesis.

Figure 5 .
Figure 5. Interconversion of matched aldehyde addition products 9d and 9f through C3-S bond rotation and lithium repositioning.

Figure 5 .
Figure 5. Interconversion of matched aldehyde addition products 9 d and 9 f through C3-S bond rotation and lithium repositioning.
5 -C 6 bond to a value of 1.344 Å, and a C 2 -O b bond to 1.434 Å in both structures.Loss of charge on O b results in decreased association with Li, wherein the Li-O b distance increases to 2.152 Å in the trans product and 2.287 Å in the cis product.This leads to increased association with the oxygen of the sulfone with atom-to-atom distance dropping by approximately 0.7 Å.As a result of cyclization, a formal charge is passed to C 6 , contributing to a shift in the position of the Li counterion somewhat closer to that of the anionic carbon.

Figure 6 .
Figure 6.Optimized geometries of reactants, transition states and products of the cyclization process leading to cis and trans oxathiin formation.

Figure 6 .
Figure 6.Optimized geometries of reactants, transition states and products of the cyclization process leading to cis and trans oxathiin formation.

Figure 7 .
Figure 7. Potential Energy surfaces for proton transfer and ring opening pathways of cis-and transoxathiin anions.Energy entries for species 9 and 10 introduced in Figure 4 are repeated here for context.

Figure 7 .
Figure 7. Potential Energy surfaces for proton transfer and ring opening pathways of cis-and trans-oxathiin anions.Energy entries for species 9 and 10 introduced in Figure 4 are repeated here for context.
d LDA was employed as the base.e Purified by Method B.
d LDA was employed as the base.e Purified by Method B.
a Moles of starting sulfone substrate.b Purification was by method A unless otherwise noted.(see Experimental); c nBuLi was the base unless otherwise indicated.d LDA was employed as the base.e Purified by Method B.
a Moles of starting sulfone substrate.b Purification was by method A unless otherwise noted.(see Experimental); c nBuLi was the base unless otherwise indicated.d LDA was employed as the base.e Purified by Method B.
a As mentioned in the text, although compounds submitted to the conditions are neutral isomers 7Z/E, anionic forms 12 (Scheme 1) are mostly likely to succumb to base attack and isomerization.b Measured at each juncture in time.c 76:24 is the starting ratio.

Relative Ratio of Isomers b
a As mentioned in the text, although compounds submitted to the conditions are neutral isomers 7Z/E, anionic forms 12 (Scheme 1) are mostly likely to succumb to base attack and isomerization.b Measured at each juncture in time.c 76:24 is the starting ratio.

Table 3 .
Bond lengths and angles for cyclization leading to cis and trans oxathiin formation.

Table 3 .
Bond lengths and angles for cyclization leading to cis and trans oxathiin formation.