Introduction

In the past several decades, the development and application of hypervalent iodine(III) reagents have received considerable attention from organic chemists for their excellent properties1,2,3,4,5,6, including thermodynamic stability, environmental friendliness, and versatile reactivity7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24. In contrast to the most well-developed functional-group-transferring transformations enabled by trifluoromethyl-, fluoro-, azido-, alkynyl-, alkenyl-containing hypervalent iodine(III) reagents7,8,9,10,11,12,13,14,15,16, the application of the nitrooxyl (O2NO)-containing I(III) reagents has remained a challenge for organic chemists, as the existing O2NO-I(III) reagents 1a–c had not found wide application in organic synthesis since their discovery several decades ago (Fig. 1a)25,26. It was not until 2020 that Katayev’s group realized the first application of O2NO-I(III) reagent 1c in the preparation of nitrooxylated β-keto esters, 1,3-diketones, malonates, and oxindoles in the absence of oxidants or bases (Fig. 1b)27. It is worth mentioning that the asymmetric version of this transformation between the reaction of O2NO-I(III) reagent 1c with β-keto esters and β-keto amides were further investigated by Deng28 and Feng’s groups (Fig. 1b)29, respectively. In 2023, Deng’s group further accomplished the nitrooxylation of diverse substrates including cyclopropyl silyl ethers, β-keto esters, β-keto amides, 1,3-diketones, and β-naphthol, by using noncyclic O2NO-I(III) compound 1a as nitrooxylating reagent (Fig. 1b)30. In addition, a catalyst-free intermolecular dearomatization reaction of β-naphthols with reagent 1c under mild conditions to access various nitrooxylated β-naphthalenones was uncovered by You’s group recently (Fig. 1b)31. Obviously, the existing hypervalent iodine(III) reagents bearing the ONO2 group are limited in types, and their applications primarily focused on the nitrooxylation reactions featuring a fully-exo fashion. In this regard, the development of hypervalent O2NO-containing iodine(III) reagents and searching for their other unique applications should be highly desirable.

Fig. 1: Existing applications of O2NO-I(III) compounds and known accesses to furazans.
figure 1

a Known O2NO-I(III) compounds 1a–c. b Reported works utilizing O2NO-I(III) reagents 1a and 1c in a fully-exo fashion. c Reported strategies for the preparation of furazan derivatives. d This work: a O2NO-I(III) compound 1d and its application for synthesis of furazans as a nitrating reagent.

Furazans (1,2,5-oxadiazoles)32,33,34,35,36 constitute an important class of heterocycles that have been applied as energetic materials37,38,39,40,41,42,43 and biologically active agents44,45,46,47,48. Accordingly, a great deal of effort has been devoted to the assemblage of this unique class of skeletons. However, the known strategies for accessing Furazans are relatively limited. Literature survey showed that the synthesis of furazans could be realized via deoxygenation of furoxans by tri-substituted phosphite (Fig. 1c, react. 1)49,50,51, cyclization of vinyl azides with NOBF4 (Fig. 1c, react. 2)52, and dehydrative cyclization of vicinal bisoximes (Fig. 1c, react. 3)34,35,36,53,54,55,56,57,58,59,60,61,62,63 mediated by alkalinous34,53 or acidic additives54,55,56,57,58,59,60,61, I2P462 or PPh3/DIAD63. It is worth noting that this last strategy is the most widely-used methodology as vicinal bisoximes, the precursor of furazans, could be readily obtained from hydroxylamination of ammonia with glyoxals, glyoxal monooximes, cyano oximes, or acyl cyanides (Fig. 1c, react. 4)34,64. Additionally, Kwong’s group recently developed an expeditious metal-free [2 + 2 + 1] radical tandem cyclization reaction of arylketimine, realizing the synthesis of a series of furazan-fused quinolines by employing tert-butyl nitrite (tBuONO) to incorporate NO moiety into furazan framework in a fully-endo pattern (Fig. 1c, react. 5)65. Although all the above approaches have their respective merits in obtaining the corresponding furazan derivatives, the development of novel synthetic routes to access this unique heterocycle should still be of important synthetic value.

Here, we reported that benziodazole-type O2NO-I(III) 1d, being a hypervalent O2NO-containing iodine(III) compound, could react with β-monosubstituted enamines to trigger a copper-catalyzed radical nitration/cyclization/dehydration cascade, providing an alternative protocol to access the exclusive furazan heterocycles (Fig. 1d). Differing from the previous nitrooxylation reactions enabled by the existing O2NO-I(III) reagents 1a and 1c, O2NO-I(III) compound 1d in this work was used as nitrating reagent to incorporate its NO moiety to furazan skeleton in a fully-endo pattern.

Results and discussion

In order to further enrich the type of hypervalent iodine(III) reagents66,67, we were interested in investigating the preparation of a benziodazole-bearing O2NO group. Following the general procedure66,67,68,69, a ligand exchange reaction of benziodazole-type Cl-I(III) compound 1e with silver nitrate (AgNO3) was conducted in dried chloroform under an N2 atmosphere. The reaction afforded the expected benziodazole-type O2NO-I(III) 1d feasibly in 93% yield as a light yellow solid, which is stable for several months when stored at 0 °C in the absence of light (Fig. 2). Thermogravimetry-differential thermal analysis (TG-DTS) showed that compound 1d decomposed at 180.7 °C (for details see Supplementary Data 3). Furthermore, a single crystal of 1d was grown in a mixed solvent of chloroform/n-hexane at room temperature, and it crystallized in the monoclinic space group P21/c with Z = 4. An X-ray crystal analysis of compound 1d (Fig. 2) revealed a distorted T-shape geometry like the common hypervalent λ3-iodanes with an O11–I10–N16 bond angle of 162.32(8)° and I–ONO2 bond length of 2.336(2) Å. The length of the observed I–ONO2 bond in compound 1d is longer than its analogous 1a and 1c, i.e., 2.311(3) Å30, and 2.283(2) Å27, respectively, suggesting reduced covalent character. Compound 1d also possesses a planar geometry, as indicated by the torsion angles O14–N13–O11–I10 (8.7(3)°), I10–C19–C18–C20 (2.0(3)°), and O17–C20–N16–C21 (8.0(5)°) (for details see Supplementary Data 4).

Fig. 2: Preparation of O2NO-containing benziodazole-type I(III) 1d.
figure 2

Conversion to reagent 1d from 1e via ligand exchange reaction and single-crystal X-ray structure of 1d.

Initially, our efforts were focused on studying the feasibility of the nitrooxylation reaction of O2NO-I(III) 1d with β-monosubstituted enamine 2a in the presence of 10 mol% CuI in acetonitrile at 50 °C under nitrogen atmosphere. Unexpectedly, it was not the nitrooxylating product but the heterocyclic furazan 3a that was produced and isolated in 72% yield (Table 1, entry 1). The results of a solvent screening revealed that the reaction in other solvents, including DCE and 1.4-dioxane led to inferior yields of 3a (Table 1, entries 2–3), while no desired product was observed when THF, DMF, or HFIP was used (Table 1, entries 4–6). The following catalysts screening showed that the reaction proceeded with significant efficiency when CuBr or CuSCN was applied (Table 1, entries 7–8). However, when other copper reagents including CuCl, Cu2O, CuBr2, Cu(OAc)2, or Cu(OTf)2 were used, product 2a was obtained in relatively lower yield in each case (Table 1, entries 9–13). Other metal additives including FeBr2, PdCl2, Mn(OAc)2, Ni(acac)2, Co(acac)2, and RhCl(PPh3)3 were also investigated. All of them were proved to be compatible with this reaction except Co(acac)2 (for details see Supplementary Table S1). The result of a control reaction conducted in the absence of a copper catalyst provided no desired product, indicating that the copper catalyst is indispensable for the reaction to occur (Table 1, entry 14). Temperature was proved to be another important factor for an efficient transformation, with reaction run at 60 °C afforded the best outcome (Table 1, entries 15–17). Furthermore, the screening on dosage of O2NO-I(III) 1d indicated that 1.5 equivalents of the hypervalent iodine(III) reagent were necessary for complete consumption of the starting enamine 2a (Table 1, entries 18–19).

Table 1 Optimization of the reaction conditionsa,bView full size image

With the optimized reaction conditions in hand (Table 1, entry 17), the substrate scope of this newly established method was evaluated (Fig. 3). A series of substituted enamines 2 were proved to be compatible with the protocol, with all of the reactions proceeded successfully to afford the corresponding substituted furazans 3a–w. As can be seen from Fig. 3, aryl enamines substituted with electron-neutral, -donating, and -withdrawing groups reacted favorably to afford furazans 3a–g in moderate to good yields. In addition, halogen-containing substrates were also conveniently converted to the desired products 3h–k in satisfactory yields. In addition, enamines equipping heterocyclic furyl and thienyl groups, or an aromatic naphthyl substituent, were also suitable for this transformation, yields. Notably, the structure of product 3n unambiguously provides the corresponding furazans 3l–n in acceptable to good confirmed by X-ray single-crystal diffraction analysis (for details see Supplementary Data 5). Furthermore, the reaction of substrates bearing other alkoxycarbonyl substituents, such as butoxycarbonyl and ethoxycarbonyl group, also proceeded well with good efficiency (3o–r). Strikingly, the method could also be well applied to enamines containing the analogous electron-withdrawing cyano or aroyl substituents (3s–w). The utility of this method was further demonstrated by the gram-scale synthesis of compound 3a in a yield of 67% when 10 mmol of 2a was used under optimized conditions.

Fig. 3: Substrate scope study for synthesis of furazans 3.
figure 3

[a] Reaction conditions: enamine 2 (0.3 mmol, 1.0 equiv), O2NO-I(III) 1d (0.45 mmol, 1.5 equiv), CuI (10% mol) in acetonitrile (4 mL) under N2 atmosphere at 60 °C for 6 h. [b] Isolated yield. [c] Gram-scale synthesis of 3a, 67% (10 mmol of 2a was used).

To our surprise, when the reaction of substrate 4a with a menaphthyl moiety was conducted under the above-optimized reaction conditions, it was not the expected menaphthyl furazan but the benzylic CH2-oxidized compounds, i.e., naphthoyl furazan 5a as well as its precursor 5a’ that were isolated in a yield of 23% and 55%, respectively (Fig. 4, entry 1). Further study revealed that reaction of menaphthyl-substituted enamine 4a with 1.0 equiv of O2NO-I(III) 1d in the presence of CuI catalyst under nitrogen atmosphere gave 89% naphthoyl-containing enamine 5a’ (Fig. 4, entry 2), which could be further converted to product 5a under standard reaction conditions (see SI for details). Considering above facts as well as the result of the control reaction (see SI for details) where trace yield of 5a’ was formed in the absence of CuI catalyst (with most of starting materiel 4b unconsumed), we tentatively presumed that enamine 4a was first oxidized to benzoyl enamine 5a’ by O2NO-I(III) 1d assisted by CuI catalyst and then the formed 5a’ was converted to product 5a by further reacting with O2NO-I(III) 1d. Thus, a larger amount of 1d was employed to facilitate the complete conversion of enamine 4a to furazan 5a. When the amount of O2NO-I(III) 1d was increased to 2.5 equivalents, a 75% yield of furazan 5a was attained (Fig. 4, entry 3). Under the most optimal conditions, other benzyl enamines were investigated, and they were all converted to the corresponding furazans 5b–e with acceptable yields (Fig. 4).

Fig. 4: Substrate scope study for synthesis of furazans 5.
figure 4

[a] Reaction conditions: enamine 4 (0.3 mmol, 1.0 equiv), O2NO-I(III) 1d (0.75 mmol, 2.5 equiv), CuI (10% mol) in acetonitrile (4 mL) under N2 atmosphere at 60 °C for 6 h. [b]Isolated yield.

Derivatization of the obtained furazan derivatives was carried out to prove the utility of this method (Fig. 5). To our delight, furazan 3b could be further transformed into compound 6 via the one-pot two-step amidation70. In addition, azide 7 could be achieved from furazan 3k through nucleophilic substitution reaction71,72. Both of the two transformations provided access to new derivatized furazan-containing molecules, demonstrating the stability of the exclusive furazan skeleton under the respective reaction conditions.

Fig. 5: Further derivatization of the obtained furazans.
figure 5

Conversion to furazans 6–7 from compounds 3b and 3k via amidation and substitution, respectively.

To understand the mechanism of this copper-catalyzed O2NO-I(III) 1d-mediated transformation, a series of control experiments were conducted (Fig. 6). First, the reaction of substrate 2a under standard conditions produced co-product N-acetyl-2-iodobenzamide S1 in high yield (based on starting material 2a) (Fig. 6a). Then radical scavenger was introduced to investigate whether the reaction adopts a radical pathway. Specifically, when 1.5 equivalents of TEMPO was employed under standard conditions, the transformation was almost completely inhibited (Fig. 6b). Next, a radical clock experiment was carried out by introducing 1.5 equivalents of compound 8 to the reaction of substrate 2a under standard reactions, and it was found that furazan 3a, nitrated compounds 9 and 10 were isolated in yield of 26%, 44%, 17%, respectively (Fig. 6b). The outcomes of the above experiments strongly indicate that reaction process might encompass a radical pathway, and the reactive NO2 species73,74,75,76 might be a crucial intermediate formed in situ from O2NO-I(III) 1d during the process. To corroborate whether NO2 was the intermediate, control experiment by replacing O2NO-I(III) 1d with exogenous brown NO2 gas, generated from the known reaction77,78 of copper powder and concentrated nitric acid, were conducted (Fig. 6c). To our delight, furazan 3a was obtained in 88% from the reaction of treating substrate 2a with NO2 gas in presence of CuBr2 catalyst in acetonitrile at 60 °C for 0.5 h, while no 3a was detected when no copper catalyst was used. The result of the above control experiment strongly supports our assumption that NO2 is the reactive species that enables the nitration reaction to occur.

Fig. 6: Mechanism investigation.
figure 6

a Generation of co-product S1 under standard conditions. b Radical-trapping experiments. c Investigation on reactive NO2 generated in situ.

Based on the above results as well as the previous reports73,74,75,76,79,80,81,82,83, a plausible radical pathway including two parts (the formation of NO2 and the following NO2-radical addition/cyclization/dehydration cascade) was proposed for this transformation (Fig. 7). Initially, homolysis79,80,81,82,83 of O2NO-I(III) 1d under heating gives O2NO radical and N-radical A1. Single electron oxidation of CuI by A1 affords Cu(II) species A2. Meanwhile, dimerization of the generated O2NO radical generates intermediate B, which is unstable and undergoes dissociation to release oxygen gas as well as NO2 molecule, a reactive radical species that can dimerize into N2O4. Then, the radical addition of NO2 to the C–C double bond of enamine 2a furnishes radical species C. Next, one H radical of intermediate C is captured by Cu(II) species A2, leading to the formation of imine D as well as Cu(III) species A3, which undergoes reductive elimination to form N-acetyl-2-iodobenzamide S1 and CuI.

Fig. 7: Plausible mechanism.
figure 7

The formation of NO2 and the following NO2-radical addition/cyclization/dehydration cascade.

Next, two pathways (Fig. 7, the path a and b) were postulated for the formation of furazan 3a from intermediate D. In path a, enamine E was formed from imine D via tautomerism first. Then nucleophilic attack of the nitrogen atom of enamino moiety in intermediate E to its oxygen center of nitrone gave the cyclized intermediate F. Subsequent tautomerization of F achieved via the system of S1/S2 and following dehydration of the resulting intermediate H gave product 3a. While in path b, the intramolecular attack of oxygen atom of nitrone in intermediate D to nitrogen center of its imino moiety, with the concomitant formation of C–C double bond and cleavage of C–N bond occurred first to give intermediate I. Then intramolecular cyclization of I provided the cyclized intermediate J, which underwent similar tautomerization and following dehydration of the resulting intermediate H to afford furazan 3a.

Finally, enaminone 11 was also examined to explore whether it is applicable to this radical nitration/cyclization/dehydration cascade reaction. The results showed that enamine 11 was equally applicable for this transformation under standard conditions, furnishing furazan 5e in 76% yield (Fig. 8). Interestingly, the two N atoms in product 5e originate from the amino moiety of enamine 11 and the nitrooxyl moiety of O2NO-I(III) 1d, respectively, but in a completely reversed pattern to the ones of 5e generated from 4e (Fig. 4).

Fig. 8: Reaction utilizing enaminone 11 as starting material.
figure 8

Compound 11 was also a suitable substrate for our nitration/cyclization/dehydration cascade reaction.

In conclusion, we prepared a benziodazole-type hypervalent O2NO-I(III) compound 1d and had it applied to the synthesis of a series of exclusively heterocyclic furazans from β-monosubstituted enamines via an unprecedented copper-catalyzed radical nitration/cyclization/dehydration cascade. Differing from the existing O2NO-I(III) reagents that have been uniformly used as nitrooxylating reagents for introducing O2NO moiety, the O2NO-I(III) 1d described in this work can be regarded as a nitrating reagent and incorporate its NO moiety to furazan skeleton in a fully-endo pattern. Furthermore, the current method also provides an alternative approach, which is in nature different from the existing strategies34,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65, to the biologically interesting furazan heterocycles.

Methods

Procedure for synthesis of O2NO-Iodine(III) 1d

To a 200 mL two-necked round-bottomed flask were added compound 1e (3.23 g, 10 mmol, 1.0 equiv), AgNO3 (3.4 g, 20 mmol, 2.0 equiv) and dried CHCl3 (70 mL) under N2 atmosphere. The reaction mixture was stirred at room temperature in the dark for 3.5 days. The mixture was then filtered through a pad of Celite and washed with CHCl3 (1000 mL). The solvents were concentrated in a vacuum to give compound 1d as a white solid.

General procedure for the synthesis of substituted furazans 3 and 5

To a 20 mL Schlenk tube equipped with a stirrer was added β-monosubstituted enamine 2 (0.3 mmol, 1.0 equiv), O2NO-I(III) 1d (0.45 mmol, 1.5 equiv) and CuI (0.03 mmol, 6 mg, 10 mol%) under N2 atmosphere, followed by addition of acetonitrile (4 mL). The tube was screw-capped and stirred at 60 °C. After stirring for 6 h, the reaction mixture was diluted with dichloromethane, filtered through a pad of Celite, and concentrated in a vacuum. The residue was purified with silica gel chromatography (PE/EtOAc) to afford furazans 3. (When enamine 4 and O2NO-I(III) 1d (0.75 mmol, 2.5 equiv) were employed as starting materials under the above conditions, substituted furazans 5 was obtained).