Polymer-Supported Dibutylstannyl Azide: An Efficient and Recoverable Reagent in the One-Pot Synthesis of Aryl Azides and 5-Aryl 1 H -tetrazoles

Herein, a very stable and highly reactive insoluble polymer-supported organotin azide was prepared from a Merrifield resin and its azide loading was determined by elemental analysis. This immobilized azide was employed in one-pot diazotization-azidodediazoniation of aromatic amines to provide a wide range of aryl azides in very good yields under mild conditions. On the other hand, this supported reagent was successfully applied in the 1,3-dipolar cycloaddition reaction with aromatic nitriles to provide 5-aryl 1 H -tetrazoles in good yields. In addition, the recyclability of tin azide from the supported dibutyltin sulfonate recovered in the synthesis processes was reported. The developed methods allow the recycling and reuse of the supported tin azide, without significant loss of reactivity after four cycles in our reaction scale and less than 20 ppm of residual tin concentration in final products without additional purification.


Introduction
Organotin compounds play a major role in decisive steps for the development of many organic synthesis methodologies involved in the production of natural products, 1 catalysts, 2 radiopharmaceuticals, 3 and molecules of biological interest. 4 Indeed, some organotin complexes are fungicides, 5 antifoulants, 6 pesticides, 7 bactericides, 8 and antitumor agents. 9 However, the valuable potential of this chemical tool is currently confronted with the controversy surrounding the toxicity of their byproducts, the stoichiometric proportion in which they are generated and the difficulties associated with their complete elimination in the final products. 10 Over the last years, different procedures that facilitate the separation of tin residues from the desired products have been developed. Among these, a few of the implemented strategies are: conversion of tin byproducts to insoluble polymeric tin fluorides, methodologies involving a catalytic amount of an organotin reagent with in situ recycling by means of a metal hydride reagent and reactions carried out in ionic liquids or polar solvents. 11 The most convenient approaches to reduce contamination and facilitate the purification of products have been the design of organotin reagents anchored on different matrices. 11 The high increase in the number of papers describing the use of polymeric supports in organic chemistry, over the past decade, is a demonstration of its impact on scientific community. There are several main factors which contribute to the popularity of the technique, such as the elimination or simplification of purification steps, the easy recovery and regeneration of the reactants or catalysts supported without considerable loss of reactivity, the tin contents in the reaction crude reduced to levels below 50 ppm, among others. 11 Considering these remarkable advantages, we have recently reported a preliminary study focused on the synthesis of aryl azides, via diazotizationazidodediazoniation of anilines in a one-pot stepwise procedure by using an immobilized organotin azide. 12 Herein, we report the expansion of the substrate scope to several aniline derivatives.
It is well known that tetrazoles are important and versatile heterocycles with a wide range of chemical applications. 13 In medicinal and pharmaceutical chemistry, this interest is due to its ability as a carboxylic acid bioisostere for the design of active pharmaceutical targets. 14 These compounds together with their derivatives act as analgesics, 15 antivirals, 16 antibacterials, 17 antifungals, 18 anti-HIV drug candidates, 19 and anticancer agents. 20 Several procedures have been reported for their synthesis by reaction between organic nitriles and different azides such as silyl, tin and sodium azide under the influence of different Lewis acid catalysts. 21 Among them, the use of tributyltin azide is the most preferable and practicable one in terms of its availability, safety, stability and solubility in organic solvents, as it is exemplified on the first synthesis of valsartan, 22 and in the numerous efforts for its improvement. 23 Recently, Chrétien et al., 24 developed a method for synthesizing 5-substituted 1H-tetrazoles via polymer-supported organotin alkoxide which is used for the in situ formation of tributylstannyl azide by reaction with trimethylsilyl azide. This heterogeneous synthesis afforded a residual tin concentration in the products compatible to pharmaceutical applications. Furthermore, the use in a catalytic amount of the organotin reagent allows an alternative type of recyclability. However, this protocol used an excess of TMSN 3 , a hazardous reagent, 25 not reusable or recoverable. Furthermore, it does not allow the subsequent functionalization of the stannylated tetrazole restricting the scope of this procedure. Notably, we have not found precedents for the use of immobilized organotin azides. Therefore, in connection with our continuing effort to study the synthetic potential of organotin compounds, 26 and considering the relevance of tetrazoles in organic synthesis, we were encouraged to explore the 1,3-dipolar cycloaddition reaction between a polymer-supported organotin azide and various nitriles with a special focus on the recovery and reuse of this reagent.

Results and Discussion
To further expand the substrate scope of our previously reported azidodeamination method, 12 we carried out the reactions of a representative series of anilines and polymer-supported organotin azide 3b prepared from the resin-bound dibutyltin chloride 3a by means of two consecutive treatments with tributylstannyl azide (TBSnN 3 ) in dimethylformamide (DMF) at room temperature for 30 h each (Scheme 1). The azide loading on 3b (0.84 mmol N 3 g -1 ) was determined by elemental analysis and afterwards, the excess of TBSnN 3 were recovered from the filtrates and the washing solutions by adding an equivalent amount of NaN 3 with respect to the experimental azide loading of 3b.
The optimal reaction conditions, established with azides 2a-2c, for in situ generation of the aryldiazonium tosylates, followed by its azidodediazoniation by means of the addition of 3b swollen in MeCN, were applied to a series of aromatic amines (Scheme 2). Finally, good isolated yields of aryl azides 2 were obtained in the presence of both electron-withdrawing and electrondonating groups with a high tolerance to various common functionalities, owing to the mild conditions employed. These results confirmed that the employment of the resinbound tin azide meant a great improvement as regards to isolation of aryl azides and removal of tin byproducts from the crude reaction. The reusability of 3b, obtained from recovered 3c by treatment with TBSnN 3 , 12 was evaluated in four-run recycling tests for the azidodeamination of 2-amino-4,5-dimethoxybenzoic acid (1d). As shown in Table 1, the successively yields obtained from the corresponding aryl azide (2d) showed that no significant loss in reactivity took place. It is important to highlight that in our test conditions the number of runs was limited by the available mass of 3b after the workup of each reaction.
Afterwards, we set out to explore the synthetic potential of the resin-bound tin azide 3b as dipolar partner in the synthesis of 5-aryl 1H-tetrazoles through the Huisgen 1,3-dipolar cycloaddition with nitriles. We started by testing the reported conditions for reactions of tributyltin azide in solution phase, 27 employing benzonitrile (4a) as model substrate, in an equimolar ratio with respect to 3b. After heating the mixture of 3b and benzonitrile in xylenes at reflux during 24 h, subsequent filtration and washing, we carried out the solvolysis of the polymer-supported stannyl tetrazole with a solution of concentrated hydrochloric acid in MeOH [HCl aq /MeOH 10% v/v]. After 30 min, an impure 5-phenyl 1H-tetrazole (5a) was obtained from the crude filtrates as confirmed by 1 H and 13 C nuclear magnetic resonance (NMR). Despite we were not able to determine the identity of the additional signals in the NMR spectra, we could observe, by thin layer chromatography (TLC) analysis, the presence of tin residues in the sowing point. Considering that such residues could be form due to the C-O bonds breaking off the spacer to the Merrifield resin [Sn-(CH 2 ) 3 -O-], promoted by the applied solvolysis conditions, we decided to investigate the proton-mediated cleavage step in order to avoid this drawback.
At first, we set out to explore the solvolysis on a previously swollen resin. After the 24 h cycloaddition step, we swelled the resin-bound stannyl tetrazole in dichloromethane (DCM) and carried out the solvolysis with HCl aq /MeOH 10% v/v. In these conditions, 5a was obtained in 52% yield but we found evidence of tin residues by TLC analysis. Taking these results into account, we decided to evaluate the effect of the concentration from the hydrochloric acid solutions at different reaction times. We carried out the respective solvolysis steps by using HCl aq /MeOH 5 and 2% v/v at 10 and 30 min each. After performing the corresponding washes, by TLC analysis of the collected filtrates, we were able to determine that the use of HCl aq /MeOH, regardless of its concentration or reaction time, promotes the release of tin residues from the solid support. Based on these results, we carried out a new reaction using TsOH • H 2 O as an alternative source of proton to perform the cleavage of the Sn-tetrazole bond. The new solvolysis procedure was carried out by swelling the polymeric material in DCM and stirring with a 0.15 M solution of TsOH . H 2 O in MeOH during 30 min at room temperature. Thereby, the corresponding spectroscopically pure 5a could be obtained by simple filtration and washing in 72% yield.
Taking into account that the efficiency in the solvolysis step depends on the nature of the Merrifield resin spacer, we carried out the synthesis of 5-(2-methylphenyl)-1H-tetrazole (5b) using TsOH • H 2 O ( Table 2, entry 1). Despite that a cycloaddition step of 24 h provided a 66% yield and taking into account a certain degree of steric hindrance of the substrate, we decided to evaluate whether a longer reaction time would improve this result or not. In order to accomplish this, we performed the synthesis of 5b, keeping the heating mixture of 2-methylbenzonitrile (4b) with 3b at reflux for 48 h ( Table 2, entry 2). After usual workup, this product was obtained in 60% yield. By TLC analysis of the supernatant before the solvolysis step, we could observe the presence of the corresponding tetrazole together with tin residues, which had not been detected in the previous reaction. We realized that the longer reaction time caused the protodestannylation of the supported stannyl tetrazole.
One of the many advantages to employ a solidsupported reactant is the possibility to work with an excess of reagents that allow driving reactions to completion and the easy recovery of the unused reagents. Consequently, we decided to carry out the same reaction using 2 equivalents of benzonitrile 4b. In this opportunity, spectroscopically pure 5b was obtained without significant yield improvement ( Table 2, entry 3).  On the other hand, we envisioned that the accelerating effect of microwaves could potentially reduce the reaction time. In such context, during the last years, several procedures have been developed to obtained tetrazoles from nitriles and silyl or sodium azide promoted by microwaves. 28 To analyze the effect of microwaves in this cycloaddition reaction, we carried out the synthesis of 5b under microwave heating, using a power of 250 W and a temperature of 150 °C, for two cycles of 20 min ( Table 2, entry 4). Unfortunately, it was only obtained in 35% yield and, by TLC analysis, we once again observed the corresponding product in the supernatant before doing the solvolysis step. Next, we decided to carry out the cycloaddition by using a lower power (150 W) and a temperature of 150 °C for two cycles of 20 min ( Table 2, entry 5). Once more, we only obtained 5b in 38% yield. After these discouraging results, we tried to perform the synthesis of tetrazoles within 24 h under thermal conventional heating. This time, we obtained various 5-aryl 1H-tetrazoles in good yields (Scheme 3). Moreover, after work-up and isolation of 5a and 5b, without further purification, several analyses by inductively coupled plasma (ICP) were carried out and exhibited a very low tin contamination (under 20 ppm).
At this point, we decided to examine the recyclability of 3b. To accomplish this, the polymer-supported dibutyltin sulfonate 3c, recovered by filtration at the end of the reactions, was treated with TBSnN 3 according to our protocol previously reported. The recycled polymersupported dibutyltin azide 3b was reused in the synthesis of 5a, as a model reaction. The corresponding tetrazole 5a was obtained in three consecutive runs of 67, 60, 48% yield, respectively. In this case, the number of times 3b was reused is not only limited by its available mass after the workup of the reaction. Furthermore, a high degradation of the resin is evident due to the temperature used in the cycloaddition step, as we detected when the reaction was carried out for 48 h.

Conclusions
In this report, it was demonstrated the efficiency of a polymer-supported dibutyltin azide in the one-pot synthesis of aryl azides via diazotization of anilines under mild conditions. Furthermore, to the best of our knowledge, this is the first report concerning the cycloaddition reaction with nitriles using a supported organotin azide, in order to provide 5-aryl 1H-tetrazoles in good yields without need of further purification. The developed methods are very advantageous in terms of yields, easier and efficient workup, and low tin pollution of products in comparison with conventional protocols. The most important advantage lies on the recyclability and reuse of the supported tin azide being reused up to four times in our reaction scale. However, on a larger scale it can be reused multiple times. Furthermore, the supported tin reagent was found to be very stable and can be stored at 2-4 °C, for an extended period of time, without losing reactivity. Finally, it is noteworthy that tin contamination of the products was limited to less than 20 ppm even using equimolar amounts of supported tin azide.

Materials and/or methods
Unless otherwise stated, analytical grade reagents and solvents were purchased from Sigma-Aldrich (St. Louis, United States) and Anedra (Buenos Aires, Argentina) and used as received. TBSnN 3 was obtained from TBSnCl as described in our previous report. 12 Resin-bound dibutyltin chloride (3a) was prepared according to the known literature procedure, 29   . Gel-phase 13 C NMR (CDCl 3 ) were recorded with optimized set parameters, 30 and gel-phase 119 Sn NMR (CDCl 3 ) were performed as the routine experiments. The acquisition of mass spectra and analytical determinations were performed using a gas chromatography-mass spectrometry (GC-MS) instrument (HP5-MS capillary column, 30 m × 0.25 mm × 0.25 μm) equipped with a HP-5972 selective mass detector operating at 70 eV in electron-ionization (EI) mode (Agilent Technologies, Santa Clara, USA). Program: 50 °C for 2 min with increase 10 °C min -1 to 280 °C; injection port temperature: 200 °C. Melting points were determined on a Reichert-Kofler (Vienna, Austria) hot-stage microscope and were uncorrected. Microanalytical data were obtained using an Exeter Analytical CE-440 CHN/O instrument (Coventry, England). The tin content was determined by inductively coupled plasma-atomic emission spectroscopy (ICP-AES) analysis at LANAQUI laboratories (CERZOS-CONICET-UNS, Bahía Blanca, Argentina).
Because azides are potentially explosive compounds, all azidation reactions and subsequent workups should be operated carefully and conducted in a fume hood with the sash positioned as low as possible. For safety instructions on lab-scale synthesis of azido compounds see the literature. 31 Aryl azides 2d-2n and tetrazoles 5a-5e are known compounds whose physical and spectroscopic properties are either in agreement with those previously reported.
General procedure for azidodeamination of aryl amines by using 3b, and its later recovery In a 25 mL round bottom flask, a solution of aryl amine 1 (0.5 mmol) and TsOH•H 2 O (0.114 g, 0.6 mmol) in MeCN (10 mL) was cooled to 0 °C in an ice bath and stirred for 15 min, tert-butyl nitrite (0.061 g, 70 μL, 0.6 mmol) was added dropwise and the solution was kept under stirring for further 15 min at the same temperature. After the addition of the resin 3b (0.73 g, 0.6 mmol, 0.82 mmol N 3 g -1 ) swollen in MeCN the suspension was allowed to attain room temperature and was gently stirred for 24 h. Then, the polymeric material was removed by vacuum filtration, washed successively with MeCN and Et 2 O (3 × 5 mL each). Thereafter, the combined filtrates and washing solutions were concentrated in vacuum to give the corresponding pure aryl azide 2. For the aryl azides 2g and 2h the combined organic layers were subjected to quantitative GC-MS analysis prior to vacuum concentration. The collected resin 3c was further washed successively with EtOH, DCM and Et 2 O (3 × 5 mL each), dried under reduced pressure to constant weight, and stored at 2-4 °C, opportunely, after several experiments, it was reacted with a 5-fold excess of TBSnN 3 in accordance with its theoretical loading (1.04 mmol g -1 ). After the first reaction process and subsequent workup, applying the same above-described procedures for its synthesis, 3b was recovered as a yellowish resin (in around 60% yield) and was found to contain 0.76 mmol of N 3 g -1 . elemental analysis: C, 67. 19;H,6.93;N, Beige solid; 80% yield; 1 H NMR (300 MHz, CDCl 3 ) d 7.59 (s, 1H), 6.65 (s, 1H), 3.97 (s, 3H), 3.91 (s, 3H) ( Figure S3, SI section); 13 C NMR (75 MHz, CDCl 3 ) d 167. 5, 154.2, 146.5, 134.1, 114.7, 112.7, 102.7, 56.5, 56.4 ( Figure S4, SI section).

General procedure for 1,3-dipolar cycloaddition between nitriles and 3b
In a 25 mL Schlenk tube under nitrogen atmosphere, nitrile 4 (0.5 mmol), 3b (0.60 g, 0.50 mmol, 0.84 mmol N 3 g -1 ) and xylenes (6 mL) were heated under reflux for 24 h. Then, the polymeric material was removed by vacuum filtration, and washed successively with xylenes and acetone (3 × 5 mL each). Next, it was carried out the cleavage of Sn-N bond by stirring the polymeric material in DCM (4 mL) with a 0.15 M solution of TsOH•H 2 O in MeOH (4 mL) at room temperature for 30 min. The resultant resin was removed by vacuum filtration and, washed successively with DCM and Et 2 O (3 × 5 mL each). The organic layer was washed with water, dried over MgSO 4 and concentrated under reduced pressure giving the corresponding pure tetrazole 5.