Mechanochemical Dimerization of Aldoximes to Furoxans

Solvent-free mechanical milling is a new, environmentally friendly and cost-effective technology that is now widely used in the field of organic synthesis. The mechanochemical solvent-free synthesis of furoxans from aldoximes was achieved through dimerization of the in situ generated nitrile oxides in the presence of sodium chloride, Oxone and a base. A variety of furoxans was obtained with up to a 92% yield. The present protocol has the advantages of high reaction efficiency and mild reaction conditions.


Introduction
The furoxans (1,2,5-oxadiazole 2-oxides) are an important class of heterocyclic compounds with a long history [1]. Because of their ability to release NO [2,3], they play an important role in biochemistry, pharmaceuticals and other fields [4][5][6][7][8]. Over the past few decades, extensive work has been devoted to their synthesis. Among them, the commonly used approaches are oxidation of aldoximes, dehydration of nitrobenzenes and pyrolysis of o-nitroazidobenzenes [9][10][11][12][13][14][15][16][17][18]. In general, most of these preparation methods are very useful, but they often suffer from some drawbacks, such as the use of complicated starting materials, special oxidants, toxic organic solvents and so on. Therefore, it is necessary to develop an efficient and environmentally friendly method to synthesize furoxans from readily available starting materials.
Aldoximes have drawn increasing attention in organic synthesis due to their easy availability, better selectivity and tolerance to various functional groups. Oxone (2KHSO 5 ·KHSO 4 ·K 2 SO 4 ) is a stable and non-toxic inorganic oxidant, as demonstrated in reactions involving aldoximes [39][40][41][42]. Oxone has also been employed in mechanochemical reactions [42][43][44][45]. In 2019, the Tong group reported a protocol to generate nitrile oxides through NaCl/Oxone oxidation of aldoximes [41]. We previously reported the formation of N-acyloxyimidoyl chlorides from the mechanochemical solvent-free reaction of aldoximes with NaCl and Oxone in the presence of Na 2 CO 3 (Scheme 1a) [42]. It was found that product distribution and product yield were sensitive to the molar ratio of the reagents as well as the employed base. Interestingly, certain amounts of furoxans could be generated from aldoximes under the modified conditions. To further study this new reaction, we decided to optimize the

Results and Discussion
Our initial investigation was started by using (E)-4-methylbenzaldehyde oxime (1a) as a representative substrate. A mixture of 1a (0.2 mmol), NaCl (1.0 equiv.), Oxone (1.0 equiv.) and Na2CO3 (1.0 equiv.) together with four stainless-steel balls (5 mm in diameter) was introduced into a stainless-steel jar (5 mL) and milled (30 Hz) in a Retsch MM400 mixer mill (Retsch GmbH, Haan, Germany) at room temperature for 30 min. After separation, an 8% yield of 2a was obtained ( Table 1, entry 1). The product distribution was significantly affected by the choice of the used base. When Na2CO3 was replaced by other inorganic bases, including NaO t Bu, NaOAc and NaHCO3, only a trace amount of 2a was obtained (Table 1, entries 2-4). Satisfyingly, the desired product 2a was isolated in 36% yield when K2CO3 was employed as the base (  [7][8][9]. To our delight, when triethylamine (NEt3) was employed, 2a could be obtained in a 79% yield (Table 1, entry 10). Based on this result, the influence of the amount of NEt3 on the product yield was investigated. The product yield was decreased to 69% and 53% when the amount of NEt3 was increased to 1.25 equiv. and 1.5 equiv., respectively (Table 1, entries 11 and 12), showing a detrimental effect of excess NEt3. The exact reason is unknown so far. On the other hand, the yield was reduced to 62% and 38% when the equivalent of NEt3 was less than the required stoichiometric amount (0.75 equiv. and 0.5 equiv., respectively) ( Table 1, entries 13 and 14). When a mixture of 1a (0.2 mmol), NaCl (1.0 equiv.), Oxone (1.0 equiv.) and NEt3 (1.0 equiv.) was magnetically stirred at room temperature for 2 h, only a trace amount of 2a could be obtained (Table 1,  entry 15). This result demonstrated the great advantage of the current reaction by ball milling over magnetic stirring. Then, the reaction time was investigated. When the reaction time was shortened from 30 min to 15 min, the desired product 2a was obtained in only a 47% yield (Table 1, entry 16). When prolonging the reaction time to 40 min, the yield had no further improvement (Table 1, entry 17). When the amounts of NaCl and Oxone were increased, the yield essentially remained the same (Table 1, entries 18 and  19). The LAG protocol has been shown to improve reaction efficiency in mechanochemical reactions [31,32,[46][47][48][49][50]. Accordingly, several liquids were added to the reaction mixture as LAG agents. However, ethyl alcohol (EtOH), dichloromethane (DCM), ethyl acetate (EtOAc) and acetonitrile (CH3CN) were detrimental to the reaction, and 2a were obtained in 50-71% yields (Table 1, entries 20-23). Therefore, optimized reaction condi-Scheme 1. Comparison of different pathways in our previous and current work.

Results and Discussion
Our initial investigation was started by using (E)-4-methylbenzaldehyde oxime (1a) as a representative substrate. A mixture of 1a (0.2 mmol), NaCl (1.0 equiv.), Oxone (1.0 equiv.) and Na 2 CO 3 (1.0 equiv.) together with four stainless-steel balls (5 mm in diameter) was introduced into a stainless-steel jar (5 mL) and milled (30 Hz) in a Retsch MM400 mixer mill (Retsch GmbH, Haan, Germany) at room temperature for 30 min. After separation, an 8% yield of 2a was obtained (Table 1, entry 1). The product distribution was significantly affected by the choice of the used base. When Na 2 CO 3 was replaced by other inorganic bases, including NaO t Bu, NaOAc and NaHCO 3 , only a trace amount of 2a was obtained (Table 1, entries 2-4). Satisfyingly, the desired product 2a was isolated in 36% yield when K 2 CO 3 was employed as the base (  10). Based on this result, the influence of the amount of NEt 3 on the product yield was investigated. The product yield was decreased to 69% and 53% when the amount of NEt 3 was increased to 1.25 equiv. and 1.5 equiv., respectively ( Table 1, entries  11 and 12), showing a detrimental effect of excess NEt 3 . The exact reason is unknown so far. On the other hand, the yield was reduced to 62% and 38% when the equivalent of NEt 3 was less than the required stoichiometric amount (0.75 equiv. and 0.5 equiv., respectively) ( Table 1, entries 13 and 14). When a mixture of 1a (0.2 mmol), NaCl (1.0 equiv.), Oxone (1.0 equiv.) and NEt 3 (1.0 equiv.) was magnetically stirred at room temperature for 2 h, only a trace amount of 2a could be obtained (Table 1, entry 15). This result demonstrated the great advantage of the current reaction by ball milling over magnetic stirring. Then, the reaction time was investigated. When the reaction time was shortened from 30 min to 15 min, the desired product 2a was obtained in only a 47% yield (Table 1, entry 16). When prolonging the reaction time to 40 min, the yield had no further improvement (Table 1, entry 17). When the amounts of NaCl and Oxone were increased, the yield essentially remained the same (Table 1, entries 18 and 19). The LAG protocol has been shown to improve reaction efficiency in mechanochemical reactions [31,32,[46][47][48][49][50]. Accordingly, several liquids were added to the reaction mixture as LAG agents. However, ethyl alcohol (EtOH), dichloromethane (DCM), ethyl acetate (EtOAc) and acetonitrile (CH 3 CN) were detrimental to the reaction, and 2a were obtained in 50-71% yields (Table 1, entries 20-23). Therefore, optimized reaction conditions were established as follows: 0.2 mmol of 1a, 1.0 equiv. of NaCl, 1.0 equiv. of Oxone and 1.0 equiv. of NEt 3 at 30 Hz for 30 min (Table 1, entry 10).   With the optimized reaction conditions in hand, the scope and generality of this reaction were then examined, and the results are shown in Scheme 2. At first, a variety of aromatic aldoximes (1a-o) were explored and found to be compatible under the optimized reaction conditions. The substrate 1b with no substituent on the phenyl ring gave the corresponding product 2b in a 52% yield. For the aldoxime 1c containing the strong electron-rich para-OMe group, very low efficiency was observed with NEt3 as a base. To our delight, product 2c could be obtained in a 43% yield when NEt3 was replaced by Na2CO3. As for the para-halogen-substituted (E)-benzaldehyde oximes 1d-f, the desired products 2d-f were synthesized in 69-78% yields. For the substrate 1g with the para-substituted CO2Me group, the corresponding product 2g was isolated in a 62% yield after prolonging the reaction time to 60 min. When the aldoxime 1h bearing the strong electron-deficient para-NO2 was investigated, the desired product 2h was obtained in only a 30% yield. However, NaO t Bu could replace NEt3 to achieve a high yield of 91% for product 2h. As for the meta-substituted substrates 1i-l bearing Me, F, Cl and Br, the  With the optimized reaction conditions in hand, the scope and generality of this reaction were then examined, and the results are shown in Scheme 2. At first, a variety of aromatic aldoximes (1a-o) were explored and found to be compatible under the optimized reaction conditions. The substrate 1b with no substituent on the phenyl ring gave the corresponding product 2b in a 52% yield. For the aldoxime 1c containing the strong electron-rich para-OMe group, very low efficiency was observed with NEt 3 as a base. To our delight, product 2c could be obtained in a 43% yield when NEt 3 was replaced by Na 2 CO 3 . As for the para-halogen-substituted (E)-benzaldehyde oximes 1d-f, the desired products 2d-f were synthesized in 69-78% yields. For the substrate 1g with the para-substituted CO 2 Me group, the corresponding product 2g was isolated in a 62% yield after prolonging the reaction time to 60 min. When the aldoxime 1h bearing the strong electron-deficient para-NO 2 was investigated, the desired product 2h was obtained in only a 30% yield. However, NaO t Bu could replace NEt 3 to achieve a high yield of 91% for product 2h. As for the meta-substituted substrates 1i-l bearing Me, F, Cl and Br, the desired products 2i-l were isolated in 69-91% yields. The disubstituted substrates 1m-o were also compatible under the standard reaction conditions, affording products 2m-o in good yields of 85-92%. Unfortunately, it was found that heteroaromatic aldoximes, including (E)-nicotinaldehyde oxime, (E)-thiophene-2-carbaldehyde oxime and (E)-benzofuran-2-carbaldehyde oxime, were not suitable substrates for the current reaction. To further illustrate the substrate scope of this reaction, the substrates were extended from the aromatic aldoximes to aliphatic aldoximes with Na 2 CO 3 as the base. Gratifyingly, the (E)-2-phenylacetaldehyde oxime 1p gave the corresponding product 2p in a 70% yield. When (E)-3-phenylpropanal oxime 1q was employed under the newly modified reaction conditions, the desired 2q was obtained in a 78% yield. Another aliphatic aldoxime 1r formed from hexanal was also applicable to our reaction and provided 2r in a 50% yield. The structures of products were unambiguously confirmed by single-crystal X-ray diffraction analysis with 2c as an example (see the Supplementary Materials for details). During the course of our studies, we tried to use the ortho-substituted (E)-2-methylbenzaldehyde oxime (1s) as the substrate. Intriguingly, the nitrile oxide 3s rather than its dimer was obtained in an 86% yield, probably due to the steric hindrance caused by the ortho-substituent. Similarly, the 1,3-dipole 3t was isolated in a 90% yield when (E)-2,4,6-trimethylbenzaldehyde oxime (1t) was employed (Scheme 3). It is noteworthy that NEt 3 was used as the base for the efficient formation of furoxans in most cases. For the strong electron-deficient aldoxime 1h bearing 4-NO 2 Ph group, a stronger base NaO t Bu could dramatically increase the product yield. In contrast, for the electron-rich substrates including aldoxime 1c containing the 4-OMePh group and aliphatic aldoximes 1p-r, a weaker base Na 2 CO 3 was required. The exact reasons for these phenomena are not yet clear but it is likely that the different basicity of the employed three bases matches the formation of the corresponding 1,3-dipolar nitrile oxides and the subsequent dimerization.
During the course of our studies, we tried to use the ortho-substituted (E)-2-methylbenzaldehyde oxime (1s) as the substrate. Intriguingly, the nitrile oxide 3s rather than its dimer was obtained in an 86% yield, probably due to the steric hindrance caused by the ortho-substituent. Similarly, the 1,3-dipole 3t was isolated in a 90% yield when (E)-2,4,6-trimethylbenzaldehyde oxime (1t) was employed (Scheme 3). To gain insight into the reaction mechanism of this transformation, control experiments were performed (Scheme 4). The reaction of 1a (0.2 mmol), NaCl (1 equiv.) and Oxone (1 equiv.) afforded hydroximoyl chloride 4a in an 87% yield under our solvent-free ball-milling conditions. Then, 4a was allowed to react with NEt3 (1 equiv.) under the ball-milling conditions and produced 2a in a 77% yield. The effect of LAG on these two reactions was also examined. It was found that CH3CN as LAG seemed to retard the formation of 4a to a certain degree and showed nearly no effect on the subsequent dimerization process. Thus, these control experiments demonstrated that 4a should be the key intermediate for the transformation to 2a and explained why a slightly overall lower yield for the formation of 2a was observed with CH3CN as the LAG agent (71% vs. 79%, Table 1, entry 23 vs. entry 10). To gain insight into the reaction mechanism of this transformation, control experiments were performed (Scheme 4). The reaction of 1a (0.2 mmol), NaCl (1 equiv.) and Oxone (1 equiv.) afforded hydroximoyl chloride 4a in an 87% yield under our solvent-free ballmilling conditions. Then, 4a was allowed to react with NEt 3 (1 equiv.) under the ball-milling conditions and produced 2a in a 77% yield. The effect of LAG on these two reactions was also examined. It was found that CH 3 CN as LAG seemed to retard the formation of 4a to a certain degree and showed nearly no effect on the subsequent dimerization process. Thus, these control experiments demonstrated that 4a should be the key intermediate for the transformation to 2a and explained why a slightly overall lower yield for the formation of 2a was observed with CH 3 CN as the LAG agent (71% vs. 79%, Table 1, entry 23 vs. entry 10). To gain insight into the reaction mechanism of this transformation, control experiments were performed (Scheme 4). The reaction of 1a (0.2 mmol), NaCl (1 equiv.) and Oxone (1 equiv.) afforded hydroximoyl chloride 4a in an 87% yield under our solvent-free ball-milling conditions. Then, 4a was allowed to react with NEt3 (1 equiv.) under the ball-milling conditions and produced 2a in a 77% yield. The effect of LAG on these two reactions was also examined. It was found that CH3CN as LAG seemed to retard the formation of 4a to a certain degree and showed nearly no effect on the subsequent dimerization process. Thus, these control experiments demonstrated that 4a should be the key intermediate for the transformation to 2a and explained why a slightly overall lower yield for the formation of 2a was observed with CH3CN as the LAG agent (71% vs. 79%, Table 1, entry 23 vs. entry 10).

Scheme 4. Control experiments.
On the basis of the above experimental results and previous literature [41,42], a plausible mechanism is proposed (Scheme 5). First, NaCl is oxidized by Oxone to generate the chlorinating species I. Then, aldoxime 1 undergoes a chlorination reaction with I to provide the hydroximoyl chloride 4 via the possible intermediate II or III. Subse-

Scheme 4. Control experiments.
On the basis of the above experimental results and previous literature [41,42], a plausible mechanism is proposed (Scheme 5). First, NaCl is oxidized by Oxone to generate the chlorinating species I. Then, aldoxime 1 undergoes a chlorination reaction with I to provide the hydroximoyl chloride 4 via the possible intermediate II or III. Subsequently, 4 eliminates HCl with the aid of base, and the resulting nitrile oxide 3 undergoes 1,3-dipolar addition to the C≡N bond of another 3 to give the dimerization product 2. The aldoximes used in the above experiments were prepared according to the reported procedure [41] and were determined as the E-isomers [51]. The E-isomers of aldoximes could be converted into the corresponding Z-isomers under acidic conditions [51]. When a mixture of (Z)-4-methylbenzaldehyde oxime (1a') and the E-isomer 1a in a molar ratio of 7:1 was employed to replace the single isomer 1a, 2a was isolated in a 75% yield (Scheme 6), indicating that both E-and Z-isomers of aldoximes could provide furoxans in essentially the same yields.  The aldoximes used in the above experiments were prepared according to the reported procedure [41] and were determined as the E-isomers [51]. The E-isomers of aldoximes could be converted into the corresponding Z-isomers under acidic conditions [51]. When a mixture of (Z)-4-methylbenzaldehyde oxime (1a') and the E-isomer 1a in a molar ratio of 7:1 was employed to replace the single isomer 1a, 2a was isolated in a 75% yield (Scheme 6), indicating that both Eand Z-isomers of aldoximes could provide furoxans in essentially the same yields. The aldoximes used in the above experiments were prepared according to the reported procedure [41] and were determined as the E-isomers [51]. The E-isomers of aldoximes could be converted into the corresponding Z-isomers under acidic conditions [51]. When a mixture of (Z)-4-methylbenzaldehyde oxime (1a') and the E-isomer 1a in a molar ratio of 7:1 was employed to replace the single isomer 1a, 2a was isolated in a 75% yield (Scheme 6), indicating that both E-and Z-isomers of aldoximes could provide furoxans in essentially the same yields. To demonstrate the utility of the obtained furoxans, 2a could be deoxygenated by triethyl phosphite (P(OEt) 3 ) to provide 1,2,5-oxadiazole 5a in 91% yield at 165 • C for 12 h under an argon atmosphere (Scheme 7) [52]. The aldoximes used in the above experiments were prepared according to the reported procedure [41] and were determined as the E-isomers [51]. The E-isomers of aldoximes could be converted into the corresponding Z-isomers under acidic conditions [51]. When a mixture of (Z)-4-methylbenzaldehyde oxime (1a') and the E-isomer 1a in a molar ratio of 7:1 was employed to replace the single isomer 1a, 2a was isolated in a 75% yield (Scheme 6), indicating that both E-and Z-isomers of aldoximes could provide furoxans in essentially the same yields. For the purpose of comparing the present solvent-free reaction with its liquid-phase counterpart, we performed the reaction of 1a (0.2 mmol) with NaCl (1.0 equiv.), Oxone (1.0 equiv.) and NEt 3 (1.0 equiv.) in several organic solvents including dimethyl sulfoxide (DMSO), N,N-dimethylformamide (DMF), acetonitrile (MeCN) and 1,2-dichloroethane (DCE) at room temperature for 2 h. The results showed that MeCN was the best solvent, and the yield of product 2a was 47%. It is obvious that the present mechanochemical protocol has the higher yield (79% vs. 47%) and shorter reaction time (30 min vs. 120 min) under solvent-free conditions compared to the liquid-phase counterpart reaction. The possible reason is that the possibility of close contact of 1,3-dipoles for dimerization under the solvent-free mechanical milling conditions is much higher than that in the liquid phase.
Green chemistry metrics, such as complete and simple environmental factors (cEF and sEF), atom economy (AE) and reaction mass efficiency (RME), for the mechanosynthesis of 2a were quantified, showing advantages in greenness compared with those of its liquidphase counterpart (see the Supplementary Materials for details).

General Information
All reagents were obtained from commercial sources and used without further purification. NMR spectra were recorded on a Bruker Advance III HD 400 NMR spectrometer (Bruker BioSpin AG, Fällanden, Switzerland; 400 MHz for 1 H NMR; 101 MHz for 13 C NMR; 376 MHz for 19 F NMR) and a Bruker Advance III HD 500 NMR spectrometer (Bruker BioSpin AG, Fällanden, Switzerland; 500 MHz for 1 H NMR; 126 MHz for 13 C NMR; 471 MHz for 19 F NMR). 1 H NMR chemical shifts were determined relative to TMS at 0.00 ppm, CDCl 3 at δ 7.26 ppm or DMSO-d 6 at δ 2.50 ppm. 13 C NMR chemical shifts were determined relative to TMS at 0.00 ppm, CDCl 3 at δ 77.16 ppm or DMSO-d 6 at δ 39.52 ppm. Data for 1 H NMR and 13 C NMR are reported as follows: chemical shift (δ, ppm), multiplicity (s = singlet, d = doublet, t = triplet, m = multiplet). High-resolution mass spectra (HRMS) were taken on a Waters Acquity UPLC-Xevo G2 QTof mass spectrometer (Waters, Milford, MA, USA) with FTMS-ESI in positive mode. Ball-milling reactions were performed in a MM400 mixer mill (Retsch GmbH, Haan, Germany), using a 5 mL stainlesssteel jar with four 5 mm diameter stainless-steel balls and were milled at a frequency of 1800 rounds per minute (30 Hz) at room temperature. E-Aldoximes 1 were prepared according to the reported protocol [41]. A mixture of Z-isomer 1a' and E-isomer 1a was obtained in molar ratio of 7:1 by stirring E-1a in trifluoroacetic acid (TFA)-CHCl 3 at 0 • C for 20 min followed by removal of the volatiles in vacuo [51]. Single crystals of 2c were grown from dichloromethane/n-hexane at 4 • C.

Mechanochemical Synthesis and Characterization of Products 2a-r, 3s and 3t
A mixture of aldoximes 1a-t (0.2 mmol), NaCl (0.2 mmol), Oxone (0.2 mmol) and base (0.2 mmol) together with four stainless-steel balls (5 mm in diameter) was introduced into a stainless-steel jar (5 mL). The reaction vessel along with another identical vessel was closed and fixed on the vibration arms of a Retsch MM400 mixer mill, and was vibrated at a rate of 1800 rounds per minute (30 Hz) at room temperature for 30 min. After completion of the reaction, the resulting mixtures from the two runs were combined and extracted with dichloromethane and water. The organic layer was decanted, and the aqueous layer was extracted by dichloromethane (2 × 20 mL). The combined organic extracts were evaporated to remove the solvent in vacuo. The residue was separated by flash column chromatography on silica gel with ethyl acetate/petroleum ether as the eluent to afford products 2a-r, 3s and 3t.

Mechanochemial Synthesis of 4a from 1a and 2a from 4a
A mixture of 1a (27.3 mg, 0.2 mmol), Oxone (124.4 mg, 0.2 mmol) and NaCl (12.3 mg, 0.2 mmol) together with four stainless-steel balls (5 mm in diameter) was introduced into a stainless-steel jar (5 mL). The reaction vessel and another same vessel were closed and fixed on the vibration arms of a Retsch MM400 mixer mill and were vibrated at a rate of 1800 rounds per minute (30 Hz) at room temperature for 30 min. After completion of the reaction, the resulting mixtures from the two runs were combined and extracted with dichloromethane and water. The organic layer was decanted, and the aqueous layer was extracted by dichloromethane (2 × 20 mL). The combined organic extracts were evaporated to remove the solvent in vacuo. The residue was separated by flash column chromatography on silica gel with ethyl acetate/petroleum ether as the eluent to afford 4a (59.8 mg, 87% yield). CH 3 CN, the best LAG in the overall dimerization reaction, was also examined for the synthesis of 4a. By following the above procedure, the reaction of 1a (56.8 mg, 0.4 mmol) with NaCl (23.8 mg, 0.4 mmol) and Oxone (245.8 mg, 0.4 mmol) in the presence of CH 3 CN (44 µL) afforded 4a (49.0 mg, 69% yield).
A mixture of 4a (34.2 mg, 0.2 mmol) and NEt 3 (28 µL, 0.2 mmol) together with four stainless-steel balls (5 mm in diameter) was introduced into a stainless-steel jar (5 mL). The reaction vessel and another same vessel were closed and fixed on the vibration arms of a Retsch MM400 mixer mill and were vibrated at a rate of 1800 rounds per minute (30 Hz) at room temperature for 30 min. After completion of the reaction, the reaction vessels were washed with ethyl acetate three times (3 × 5 mL) and the combined solution was evaporated to remove the solvent in vacuo. The residue was separated by flash column chromatography on silica gel with ethyl acetate/petroleum ether as the eluent to afford 2a (41.2 mg, 77% yield). CH 3 CN, the best LAG in the overall dimerization reaction, was also examined for the synthesis of 2a from 4a. By following the above procedure, the reaction of 4a (68.2 mg, 0.4 mmol) with NEt 3 (56 µL, 0.4 mmol) in the presence of CH 3 CN (12 µL) afforded 2a (39.6 mg, 74% yield).

Deoxygenation Reaction of 2a
A mixture of 2a (27.0 mg, 0.1 mmol) and triethyl phosphite (0.5 mL) was heated at 165 • C under an argon atmosphere for 12 h. After completion of the reaction, the reaction vessels were washed with ethyl acetate three times (3 × 5 mL), and the combined solution was evaporated to remove the solvent in vacuo. The residue was separated by flash column chromatography on silica gel with ethyl acetate/petroleum ether as the eluent to afford 5a (23.1 mg, 91% yield). 1

Synthesis of 2a in Liquid Phase
To a stirred solution of 1a (27.5 mg, 0.2 mmol) in CH 3 CN (2 mL) were added NaCl (12.1 mg, 0.2 mmol), Oxone (123.4 mg, 0.2 mmol) and NEt 3 (28 µL, 0.2 mmol). The reaction mixture was allowed to stir at room temperature for 2 h. Then, the reaction mixture was filtered through a silica gel plug with ethyl acetate as the eluent and, subsequently, the solvent was removed under reduced pressure. The residue was purified by flash chromatography on silica gel with petroleum ether/ethyl acetate as eluent to give product 2a (12.7 mg, 47% yield).

Conclusions
In conclusion, we have successfully developed a solvent-free dimerization reaction of aldoximes to obtain furoxans in the presence of sodium chloride, Oxone and a base under solvent-free ball-milling conditions. The starting materials are easily available, and various aromatic and aliphatic aldoximes can be employed as the substrates. This protocol has advantages of ambient reaction conditions, high yields, solvent-free and catalyst-free conditions. Finally, a plausible reaction mechanism is proposed to explain the formation of furoxans.
Author Contributions: G.-W.W. supervised the project, analyzed data, discussed with R.-K.F. and wrote the manuscript; R.-K.F. and K.C. did experiments and provided a draft; C.N. characterized the X-ray structure of 2c. All authors contributed to the revision. All authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by National Natural Science Foundation of China, grant number 21372211.