Divergent decomposition pathways of DMSO mediated by solvents and additives

KEYWORDS DMSO (dimethyl sulfoxide) plays an increasingly significant role in various synthetic processes by generating diverse active intermediates in situ, which actively participate in reactions. It is crucial to control the formation of these active intermediates to prevent their mutual interference during utilization. Our previous research identified methyl methanethiosulfonate (MMTS) as a major decomposition product of DMSO when catalyzed by small amounts of (COCl)2 in CH3CN. In the current study, we investigated how different solvents and additives can mediate the formation of MMTS. Complete avoidance of MMTS formation was achieved in 1,4-dioxane, while only trace amounts were observed in toluene, MTHF, or CHCl3. Moreover, the decomposition pathway of DMSO in these solvents was effectively mediated through the addition of strong acids (HX, where X = TfO, ClO4, I, Br, or Cl) or in the presence of LiI, CH3I, or Br2. The effects of solvents and additives on the decomposition of DMSO were explored. The possible mechanisms for the decomposition of DMSO under different conditions were proposed and discussed. GRAPHICAL ABSTRACT


Introduction
Swern oxidation is the most well-known application of DMSO as a reagent in synthesis [1].In recent years, an increasing number of synthetic methods have been developed by using synthons derived from DMSO, highlighting its remarkable versatility as a valuable reagent rather than merely a solvent in synthetic processes [2].DMSO serves as a source for diverse synthons, including oxygen [3], methyl [4], methylene [5], methylthio [6], methylthiomethyl [7], methyl sulfoxide [8], or methyl sulfoxide methyl group [9].
Our research has been primarily focused on the utilization of DMSO in conjunction with (COCl) 2 in synthesis.It has been discovered that this combination can generate several active intermediates, depending on the specific reaction conditions.For example, it generates CH 3 SCl as an electrophilic reagent that undergoes electrophilic addition with C = C bonds [10], and produces chlorodimethylsulfonium salt acting as a dehydrating reagent to convert amides or aldoximes into nitriles [11] or acting as a Cl + source in chlorination of 1,3-dicarbonyls [12], and afford HCHO as a methylene synthon in the preparation of benzoxazines [13] or methylenebisamides [14].In addition, one straightforward method for the preparation of methyl methanethiosulfonate (MMTS) from DMSO has been established in our previous work [15], which is significant since MMTS is a crucial methanesulfenylating reagent.
Given the diverse applications of formaldehyde (HCHO) in synthesis, our research efforts have been focused on developing methods for utilizing HCHO generated in situ from DMSO.However, a significant drawback was observed in the form of methyl methanethiosulfonate (MMTS), which was co-produced with HCHO and posed challenges during the utilization of HCHO.The interference caused by MMTS complicated the purification process, particularly through column chromatography, due to its similar polarity to the target molecules.To address this issue, our aim was to prevent the formation of MMTS during the generation of HCHO from DMSO.It was discovered that the choice of solvent polarity had a substantial impact on the decomposition pathway of DMSO.Furthermore, certain additives, including strong acids, LiI, CH 3 I, or Br 2 , were found to facilitate and mediate the decomposition of DMSO (Scheme 1).In this report, we present our findings regarding the factors influencing the decomposition pathway of DMSO.These results provide valuable insights for the application of DMSO as a reagent in synthesis by effectively mediating its decomposition pathway.

Results and discussion
The effect of solvents on the decomposition of DMSO.The influence of various solvents on the decomposition of DMSO was systematically investigated, and the corresponding results are summarized in Table 1.All the decomposition products, except for formaldehyde (HCHO), derived from DMSO were quantified by 1 H NMR using CH 2 Br 2 as an internal standard.The yields presented in Table 1 were calculated based on the quantities of products obtained per DMSO molecule.The reaction conditions for entry 1 in Table 1 were based on the similar conditions reported in our previous work [15].The reaction was monitored by 1 H NMR and the peak of DMSO disappeared after refluxing for 2 h.The identified decomposition products in the reaction mixture, as determined by 1 H NMR, included MMTS, dimethyl disulfide (DMDS), and dimethyl sulfide (DMS).Among Scheme 1. Decomposition of DMSO under different conditions.three sulfur-containing products, DMS exhibited the highest yield (45%), whereas the yield of MMTS (17%) closely matched the isolated yield reported in our previous work [15] (entry 1, Table 1).Surprisingly, when the reaction was conducted in 1,4-dioxane under reflux conditions, complete disappearance of DMSO required 4 h.Notably, neither MMTS nor DMS was detected by 1 H NMR spectroscopy.Instead, DMDS became the predominant product with a yield of 25%, accompanied by small amounts of chloromethylmethylsulfide (CMMS) at 8% yield (entry 2, Table 1).Lowering the reaction temperature to approximately 85°C (close to the boiling point of CH 3 CN) significantly decelerated the reaction, extending the reaction time to 13 h for completion (entry 3, Table 1).Compared to the reaction in entry 2, the yields of DMDS and CMMS remained nearly unchanged, while DMS was identified as a minor product with a yield of 13%.When toluene was used instead, the reaction reached completion after refluxing for 5 h.DMDS remained the major decomposition product (20% yield) with small amounts of MMTS, DMS, and CMMS observed (entry 4, Table 1).When 2-methyltetrahydrofuran (2-MTHF) was employed as the solvent, DMDS was obtained as the major product with a yield of 24% after refluxing for 12 h.Trace amounts of MMTS, DMS, and CMMS were also found, with DMS and CMMS yielding 5% and 8%, respectively (entry 5, Table 1).When the reaction was carried out in CHCl 3 , the decomposition of DMSO became very sluggish.About 40% of DMSO remained unreacted after refluxing for 12 h.And DMDS remained the primary decomposition product (13% yield) with trace amounts of MMTS and a few amounts of DMS and CMMS (11% and 5% yields) (entry 6, Table 1).
Apart from the sulfur-containing decomposition products identified through 1 H NMR, it is important to note that HCHO was also expected to be among the decomposition products.This is supported by the observation of white solid paraformaldehyde adhering to the bottom of the reflux condenser.In our previous work, we successfully utilized formaldehyde generated from the decomposition of DMSO in CH 3 CN as a methylene synthon for the synthesis of benzoxazines [13].
These results suggested that there might be two different pathways for the decomposition of DMSO in these solvents.Based on our previous work [15], we proposed a possible pathway for the formation of MMTS from DMSO in the presence of catalytic amounts of (COCl) 2 in CH 3 CN (Scheme 2).This pathway provides a comprehensive explanation for the composition of the mixture resulting from the decomposition of DMSO in CH 3 CN.In contrast, the decomposition of DMSO in the other four solvents (1,4-dioxane, toluene, 2-methyltetrahydrofuran (2-MTHF), and CHCl 3 ) displayed distinct characteristics compared to CH 3 CN.The first major difference was that MMTS, the marker component in CH 3 CN, was not generated at all in dioxane, and only trace amounts were observed in toluene, 2-MTHF and CHCl 3 .In addition, the most main component, DMS, in CH 3 CN did not predominate in 1,4-dioxane, toluene, 2-MTHF or CHCl 3 , while DMDS, a minor component in CH 3 CN, became the predominant component in these four kinds of solvents.These observations clearly suggest an alternative decomposition pathway for DMSO in these solvents.Based on the components derived from the decomposition of DMSO, a possible pathway was proposed as follows (Scheme 3).CMMS was observed in 1,4-dioxane, 2-MTHF and CHCl 3 , which was inferred to be formed by the elimination of chlorodimethylsulfonium salt followed by an addition of Cl − .This process is recognized as the Pummerer rearrangement [16].During the formation of CMMS, HCl was released at the same time.DMSO was protonated twice in the presence of HCl to generate the intermediate I, which lost a molecule of water to produce chlorodimethylsulfonium salt.These sequential reactions are the most commonly acknowledged steps in the widely accepted mechanism for deoxygenating sulfoxides using hydrogen halides [17].
The chlorodimethylsulfonium salt underwent an elimination to produce the intermediate II and release HCl.The nucleophilic attack of H 2 O on the intermediate II generates the intermediate III, which decomposed after a proton transfer to afford methanethiol and protonated formaldehyde.The coupling of methanethiol by autoxidation produced DMDS.The protonated formaldehyde deprotonated to release HCl, which entered the next catalytic cycle.Such a catalytic cycle can explain why DMDS was the predominant component in these solvents.Only trace amounts of CMMS were detected in toluene, which might be due to hydrolysis of CMMS at high temperature under reflux conditions.And the presence of small amounts of MMTS and DMS in toluene, 2-MTHF, and CHCl 3 implied that Path 1 shown in Scheme 2 was the minor pathway for the decomposition of DMSO in these solvents.Based on Path 1 shown in Scheme 2, the formation of MMTS was accompanied by DMS.However, for the reaction of entry 3 in Table 1, DMS was detected but without the formation of MMTS.In this case, it was inferred that the formation of DMS in dioxane might occur through the decomposition of methoxydimethylsulfonium salt shown at the bottom of Scheme 3.
These results indicated that the polarity of solvent has significant impact on the decomposition of DMSO initiated by (COCl) 2 .Polar solvent, such as CH 3 CN, favored Path 1 shown in Scheme 2, while Path 2 shown in Scheme 3 predominated in less polar solvents, such as 1,4-dioxane, toluene, 2-MTHF, or CHCl 3 .The sluggishness of the reaction in CHCl 3 might result from its lower reaction temperature under reflux owing to its relatively lower boiling point.
The effect of strong acids on the decomposition of DMSO.The reaction in 1,4dioxane, as indicated by entry 2 in Table 1, was observed to complete after 4 h of reflux.However, by reducing the reaction temperature to 85°C, the reaction time significantly increased to 13 h (entry 3, Table 1).This observation led us to consider strategies for accelerating the decomposition of DMSO in 1,4-dioxane.Given that the catalytic cycle of Path 2 in Scheme 3 involves the protonation of DMSO twice, it was hypothesized that the addition of an additional acid might facilitate the reaction.
To further investigate the influence of acids on the decomposition of DMSO, a range of acids were evaluated, namely TfOH, HClO 4 , HI, HBr, HCl, and TFA.HClO 4 , HI, HBr, and HCl were employed as 70%, 56%, 40%, and 36% aqueous solutions, respectively, while TfOH and TFA were used as analytically pure reagents.Remarkably, the reaction described in entry 3 of Table 1 exhibited significant acceleration in the presence of strong acids (0.1 equiv.),including TfOH, HClO 4 , HI, HBr, and HCl (entries 1-5, Table 2).Among these acids, the reaction proceeded most rapidly with HBr, achieving complete conversion in approximately 5 min (entry 4, Table 2).Moreover, the reaction proceeded relatively slowly with HCl, requiring approximately 3 h for completion (entry 5, Table 2).In contrast, TFA, which is considered a relatively weak acid, had a negligible effect on the reaction rate (entry 6, Table 2).
Notably, in the reactions where TfOH, HClO 4 , or HI were added, the formation of MMTS was observed once again, similar to the reaction in CH 3 CN.Additionally, DMS became the predominant component in these reactions.These findings strongly suggested that the decomposition of DMSO in 1,4-dioxane, in the presence of these acids, followed Path 1 as illustrated in Scheme 2. The influence of different amounts of acids on the reactions was investigated.Remarkably, the reactions were significantly enhanced when 0.01 equivalents of acids were added, except in the case of HCl (entries 7-11, Table 2).Notably, the addition of 0.01 equivalents of concentrated HCl had a negligible effect on the reaction rate (entry 11, Table 2).Intriguingly, when 0.01 equivalents of TfOH, HClO 4 , or HBr were used, only trace amounts of MMTS were observed, and DMDS once again became the predominant product (entries 7, 8, and 10, Table 2).This observation suggested that under these conditions, the decomposition of DMSO followed Path 2 as depicted in Scheme 3. DMS was observed as one of the major products in these reactions (entries 7, 8, and 10, Table 2), which supports the assumption that it originated from the decomposition of an intermediate methoxydimethylsulfonium salt shown at the bottom of Scheme 3.
We investigated whether these strong acids could individually promote the decomposition of DMSO in the absence of (COCl) 2 .The results revealed that the presence of 0.1 equiv. of TfOH, HClO 4 , or HI did not lead to the decomposition of DMSO (entries 12-14, Table 2).However, 0.1 equivalents of HBr effectively catalyzed the decomposition of DMSO (entry 15, Table 2).In contrast, when 0.1 equiv. of concentrated HCl were present, the reaction proceeded very slowly, with 75% of DMSO remaining after heating at 85°C for 3 h (entry 16, Table 2).Interestingly, when the reaction was conducted in CH 3 CN under reflux, the presence of 0.1 equiv. of HBr did not induce the decomposition of DMSO (entry 17, Table 2).
Notably, the decomposition of DMSO in the presence of 0.1 equiv. of HBr was particular, which produced DMS as the major product with almost no MMTS observed (entries 4 and 15, Table 2).As aforementioned, the formation of DMS was accompanied by MMTS based on Path 1 shown in Scheme 2. Therefore, it was inferred that DMSO might decompose through a different pathway in the presence of 0.1 equiv. of HBr, which was proposed as follows (Scheme 4).First, DMSO reacted with HBr to generate an intermediate bromodimethylsulfonium salt, which further decomposed to release DMS and Br 2 .Subsequently, DMSO reacted with Br 2 to produce the intermediate bromodimethylsulfonium salt that reenters the catalytic cycle, while simultaneously generating HOBr.HOBr exhibited instability and underwent decomposition, generating H 2 O, O 2 , and Br 2 .The generated Br 2 further reacted with H 2 O to form HBr, which played a crucial role in the catalytic cycle.The formation of HCHO was postulated to occur through the reaction of bromodimethylsulfonium with DMSO, leading to the production of an intermediate methoxydimethylsulfonium salt.Subsequently, this intermediate decomposed to yield DMS and HCHO.
In order to validate proposed Path 3 in Scheme 4, Br 2 (0.1 equiv.) was used instead of HBr in the reaction of entry 4 in Table 2 (entry 18, Table 2).The reaction proceeded at a similar rate as that of entry 4, and the composition of the decomposition products closely resembled that of entry 4 as well.The main product obtained was DMS, with trace amounts of MMTS, DMDS, and CMMS.Interestingly, the decomposition of DMSO also occurred in the presence of Br 2 alone (entry 19, Table 2), and the resulting decomposition products were similar to those observed in the reaction of entry 18.These findings suggested that Br 2 , like HBr, could facilitate the decomposition of DMSO.Previous reports have indicated that Br 2 can be generated from the reaction of DMSO with HBr, where it serves as an oxidant for thiols [18] or as a reagent for bromination reactions of alkenes, alkynes, or ketones [19].Therefore, it is reasonable to consider Path 3 in Scheme 4, which involves Br 2 as a key intermediate in the catalytic cycle.
Furthermore, an earlier study conducted by Oae et al. also described the conversion of DMSO into DMS facilitated by the presence of HBr or Br 2 , all in the absence of any solvent [20].Their research proposed two distinct catalytic pathways, both involving the reaction of DMSO with HBr, yielding DMS, Br 2 , and H 2 O as products.In one of these pathways, the bromination of DMSO resulted in the formation of α-bromodimethyl sulfoxide, which then underwent a Kornblum reaction with DMSO, ultimately forming an intermediate, CH 3 S( = O)CHO.This intermediate subsequently underwent hydrolysis to generate methanesulfinic acid, which, in turn, was oxidized by DMSO, yielding methanesulfonic acid as the final product.In contrast, the other pathway involved the bromination of DMS to produce α-bromodimethyl sulfide, which, similar to the previous pathway, underwent a Kornblum reaction to produce an intermediate, CH 3 SCHO.This intermediate experienced hydrolysis, resulting in methanesulfenic acid, which was then oxidized by DMSO to produce methanesulfonic acid.Notably, both pathways yielded the same product composition with a consistent ratio, namely 75% DMS, 25% CH 2 O, and 25% CH 3 SO 3 H.The authors also noted that the addition of 5% DMS could accelerate the reaction.We attempted the reaction with added DMS, but observed no acceleration in our system.Moreover, only trace amounts of CH 3 SO 3 H were detected in our experiment.Consequently, we believe that the decomposition of DMSO in dioxane, catalyzed by HBr or Br 2 in our work, differs from the decomposition of pure DMSO catalyzed by HBr or Br 2 in the study conducted by Oae et al.
In prior studies focusing on the deoxygenation of sulfoxides with hydrogen halides, hydriodic acid was regarded as the most effective acid catalyst [21].However, in our current experimental setup, we observed that no reaction occurred when only a catalytic quantity of HI was present, whereas the reaction proceeded in the presence of a catalytic amount of HBr.This discrepancy might be attributed to the influence of the reaction environment on the outcomes.
It was noteworthy that no white solid of paraformaldehyde was observed at the bottom of reflux condenser when the reactions were carried out at 85°C in the presence of 0.1 equiv. of HI, HBr or HCl (entries 3-5, Table 2).In order to detect whether HCHO was formed in these reactions, 1 equiv. of ethylene glycol was added to these reaction mixtures after DMSO (1 equiv.) was consumed completely.After stirring the mixture at 85°C for 1 h, 1,3dioxolane was detected by 1 H NMR, which implied that HCHO was also produced in these reactions (Scheme 5).
The experimental results demonstrated that the presence of strong acids could enhance the decomposition of DMSO initiated by (COCl) 2 in 1,4-dioxane.Specifically, the addition of HI, as well as 0.1 equiv. of TfOH and HClO 4 , influenced the decomposition pathway  of DMSO in 1,4-dioxane, leading to the formation of MMTS as one of the major products.Interestingly, among all the strong acids that accelerated the reactions, only HBr was capable of promoting the reactions independently without the need for (COCl) 2 .Furthermore, it was observed that a relatively weak acid, such as TFA, had a negligible effect on the decomposition rate of DMSO.The effect of LiI or CH 3 I on the decomposition of DMSO.Moreover, a noteworthy discovery was the significant promotion of DMSO decomposition in 1,4-dioxane in the presence of catalytic amounts of LiI or CH 3 I.The addition of 0.01 equiv. of LiI reduced the reaction time from 4 h to 1 h for the reaction described in entry 2 of Table 1 (entry 1, Table 3).However, this change in reaction conditions also resulted in a different composition of decomposition products.The products now included MMTS, DMDS, DMS, and CMMS.Notably, compared to the reaction in CH 3 CN (entry 2, Table 1), the quantities of MMTS and DMS decreased, while the level of DMDS slightly increased.Additionally, CMMS was observed in the reaction mixture.
A proposed mechanism for the promotion of the reaction by LiI is as follows (Scheme 6).As mentioned previously, CMMS formation occurred through the reaction of DMSO with Scheme 6.The mechanism of decomposition of DMSO promoted by LiI.
(COCl) 2 , involving subsequent elimination and addition steps with concurrent release of HCl.Protonation of DMSO by the in situ generated HCl took place, and LiI acted as a nucleophile, attacking the protonated DMSO to form CH 3 I and CH 3 SOH.The reaction between DMSO and CH 3 I, driven by nucleophilic substitution, produced a dimethylmethoxysulfonium salt, which then decomposed, liberating HCHO, DMS, and HCl.The released HCl participated in the subsequent catalytic cycle.The unstable CH 3 SOH was further transformed into MMTS and DMDS.This proposed catalytic cycle provides a plausible explanation for the observed formation of all decomposition products.
According to the proposed catalytic cycle in Scheme 6, CH 3 I was identified as a crucial intermediate, suggesting its role in catalyzing the decomposition of DMSO.To validate this hypothesis, the reaction was carried out using CH 3 I instead of LiI, while keeping the other reaction conditions identical to those described in entry 1 of Table 3. Remarkably, the results closely resembled those obtained in the presence of LiI (entry 2, Table 3) in terms of both the reaction time and the composition of decomposition products.This finding provided compelling evidence that CH 3 I could indeed promote the decomposition of DMSO, confirming the validity of the proposed mechanism presented in Scheme 6.
Furthermore, the impact of the amount of LiI on the reaction was investigated.It was observed that the reaction time was reduced to 2.5 h when 0.1 equiv. of LiI was added (entry 3, Table 3).However, when 0.5 equiv. of LiI was introduced, the reaction was completely inhibited (entry 4, Table 3).This observation led to the inference that the reaction involving CH 3 I and DMSO in the proposed catalytic cycle was reversible.Consequently, a high concentration of I − was found to be unfavorable for the forward conversion of the reaction (Scheme 6).
The addition 0.1 and 0.01 equiv. of CH 3 I made no difference on the reaction rate (entry 5, Table 3).Furthermore, 0.5 equiv. of CH 3 I, similar to LiI, also prevented DMSO from decomposition (entry 6, Table 3).
The reactions were also carried out in the presence of LiI or CH 3 I only without (COCl) 2 .However, no reaction occurred at all (entries 7 and 8, Table 3), which indicated that HCl generated by (COCl) 2 was indispensable for the catalytic cycle in Scheme 6.
For the reactions in CH 3 CN, the addition of 1% LiI or CH 3 I did not impose significant effect on the decomposition of DMSO.The results were nearly identical to reactions carried out in the absence of them (entries 9 and 10, Table 3).However, it was interesting to note that the addition of 0.1 equiv. of LiI or CH 3 I to the reaction of entry 1 in Table 1 inhibited the reaction significantly (entries 11 and 12, Table 3), which can be accounted for by the following reasons.In the case of LiI, it is likely that LiI intercepts the hydroxydimethylsulfonium salt in the catalytic cycle, resulting in the production of CH 3 I and CH 3 SOH.This interception prevents DMSO from entering the catalytic cycle described in Scheme 2. On the other hand, CH 3 I may consume Cl − , which is necessary for the intermediate dimethylmethoxysulfonium salt to undergo elimination and release HCl, HCHO, and DMS.These reactions disrupt the catalytic cycle.Therefore, it can be inferred that catalytic amounts of LiI or CH 3 I can modulate the decomposition pathway of DMSO initiated by (COCl) 2 in 1,4-dioxane by participating in a distinct catalytic cycle.This newly identified pathway differs from the pathways illustrated in Schemes 2-4.In this pathway, the decomposition of DMSO in 1,4-dioxane can generate MMTS and DMS as the main products.
While the deoxygenation of DMSO by HI has been extensively studied [21,22], our observations deviate from the findings in the existing literature.For instance, we noticed that CH 3 I can fulfill the same role as LiI.Additionally, it's worth noting that increasing the amounts of lithium iodide or methyl iodide in the system doesn't necessarily lead to more favorable results.In CH 3 CN, even a 10% concentration of lithium iodide can entirely suppress the reaction.These unique phenomena represent novel findings that haven't been previously documented in the literature.

Conclusion
DMSO undergoes different decomposition pathways depending on the solvent used.In less polar solvents like 1,4-dioxane, toluene, 2-MTHF, or CHCl 3 , the main decomposition products are DMDS and HCHO.This pathway is distinct from the one observed in CH 3 CN, where the major products are DMS, HCHO, and MMTS.
Furthermore, the addition of catalytic amounts of strong acids, LiI, or CH 3 I can mediate the decomposition of DMSO in 1,4-dioxane, leading to the formation of MMTS as a significant decomposition product.This process involves a catalytic cycle in the presence of LiI or CH 3 I, with CH 3 I playing a crucial role as an intermediate.Among the strong acids tested, HBr stands out as the only one capable of promoting the decomposition of DMSO independently, without the need for (COCl) 2 initiation.DMS serves as a key indicator of the decomposition process.In the presence of HBr, the decomposition of DMSO proceeds through a catalytic cycle involving Br 2 as a critical intermediate.These findings highlight that DMSO is not an entirely inert solvent and can undergo decomposition under various conditions.Caution should be taken utilizing DMSO as a solvent.The utilization of HCHO and MMTS derived from DMSO is currently underway in our lab.

Experimental section
General Information.NMR spectra were obtained on a Bruker AV 300 spectrometer ( 1 H NMR at 300 MHz, 13 C{ 1 H} NMR at 75 MHz) in CDCl 3 .Reagents and solvents are commercial grade and were used as supplied.All the reagents were purchased from Innochem.
General Procedure for the decomposition of DMSO in different solvents with a catalytic amount of (COCl) 2 .To a solution of DMSO (30 mmol, 1 equivalent) in solvent (12 mL) was added dropwise a solution of oxalyl chloride (3 mmol, 0.1 equivalents) in solvent (8 mL) at room temperature.The mixture was heated to reflux in an oil bath after addition of (COCl) 2 was completed.The reaction was monitored by 1 H NMR until DMSO disappeared completely.Then the mixture was cooled to room temperature and a defined amount of CH 2 Br 2 as an internal standard was added.The amounts of the decomposition products were quantified by an 1 H NMR internal standard method.The 1 H NMR spectra were attached in Supporting Information.
General Procedure for the decomposition of DMSO in 1,4-dioxane with a catalytic amount of (COCl) 2 in the presence of an additive.To a solution of DMSO (30 mmol, 1 equivalent) in 1,4-dioxane (12 mL) was added dropwise a solution of oxalyl chloride (3 mmol, 0.1 equivalents) in 1,4-dioxane (4 mL) and a solution of additive (0.1 or 0.01 equivalents) in 1,4-dioxane (4 mL) at room temperature successively.The mixture was heated to 85°C in an oil bath after addition of (COCl) 2 and the additive was completed.The reaction was monitored by 1 H NMR until DMSO disappeared completely.Then the mixture was cooled to room temperature and a defined amount of CH 2 Br 2 as an internal standard was added.The amounts of the decomposition products were quantified by an 1 H NMR internal standard method.The 1 H NMR spectra were attached in Supporting Information.
General Procedure for the decomposition of DMSO in 1,4-dioxane in the presence of a strong acid, or LiI, or CH 3 I without (COCl) 2 .To a solution of DMSO (30 mmol, 1 equivalent) in 1,4-dioxane (12 mL) was added dropwise a solution of strong acid, or LiI, or CH 3 I (3 mmol, 0.1 equivalents) in 1,4-dioxane (8 mL) at room temperature.The mixture was heated to 85°C or reflux in an oil bath after addition.The reaction was monitored by 1 H NMR. The amounts of the decomposition products were quantified by an 1 H NMR internal standard method.The 1 H NMR spectra were attached in Supporting Information.

Scheme 4 .
Scheme 4. Possible pathway of decomposition of DMSO in the presence of HBr.

Scheme 5 .
Scheme 5.The capture of HCHO by ethylene glycol.

Table 1 .
Decomposition of DMSO in different solvents.

Table 2 .
Decomposition of DMSO in different solvents in the presence of extra acids.
a Generally, the reactions were carried out with the 10:1 ratio of DMSO/(COCl) 2 in the presence of a strong acid at 85°C in 1,4-dioxane, except the control reactions in the absence of (COCl) 2 in entries 12-17, and 19.The reaction in entry 17 was performed in CH 3 CN under reflux.b The yields determined by 1 H NMR. c Not detected by 1 H NMR. d Not quantified by 1 H NMR. e The formation of HCHO was confirmed by capture with ethylene glycol.f No reaction occurred.g 75% of DMSO remained.

Table 3 .
Decomposition of DMSO in different solvents in the presence of LiI or CH 3 I.Generally, the reactions were carried out with the 10:1 ratio of DMSO/(COCl) 2 in the presence of LiI or CH3I under reflux in 1,4-dioxane or CH 3 CN, except the control reactions in the absence of (COCl) 2 in entries 7 and 8.
bThe yields determined by 1 H NMR.cNot quantified by 1 H NMR. d Almost no reaction occurred.e Not detected by 1 H NMR.