A combination of trimethylsilyl chloride and hydrous natural montmorillonite clay: an efficient solid acid catalyst for the azidation of benzylic and allylic alcohols with trimethylsilyl azide

Abstract

We present a new procedure to convert in situ natural montmorillonite clay into a partially acidified one using a catalytic amount of trimethylsilyl chloride and trace water. We demonstrate that the acidic montmorillonite can catalyze the direct azidation of various benzylic and allylic alcohols with trimethylsilyl azide.

---

Montmorillonite (Mont) is one of the abundant naturally-occurring clays. It is composed of stacked, negatively charged, two-dimensional aluminosilicate layers with exchangeable cationic species, mostly sodium ions, between the layers. Substitution of the sodium ions by protons or multivalent metal ions turns the natural montmorillonite (Na-Mont) into an acidic clay which can be employed as a catalyst for various organic transformations.¹ For example, proton-exchanged montmorillonite (H-Mont) was utilized for the alkylation of typical nucleophiles, such as amine and allyltrimethylsilane, with alcohols.² We also demonstrated the efficient catalysis of tin(IV)-montmorillonite (Sn-Mont)³ for the reactions of various benzylic and allylic alcohols with typical nucleophiles, such as allytrimethylsilane,⁴ malonates,⁵ cyanotrimethylsilane⁶ and triethylsilane.⁷ These reactions are attributed to the strong acidity and the unique structural character of Sn-Mont; the alcohols first underwent protonation, followed by dehydration with Sn-Mont to form benzylic carbenium intermediates, which were trapped between the montmorillonite silicate layers. It is considered that the low nucleophilic montmorillonite silicate anions catalyzed the reaction not only by stabilizing the carbenium ions, but also by accelerating the reactions between the carbenium ions and the nucleophiles.

In contrast to conventional acid catalysts, such as BF₃·Et₂O or AlCl₃, the solid acid montmorillonites tend to tolerate the presence of trace water in the reaction system, which in most cases, helps simplify the synthetic procedures for various acid-catalyzed reactions as well as has a significant advantage regarding their reuse.

Previously, we reported the chlorination of various benzylic and allylic alcohols using trimethylsilyl chloride (TMSCl) and hydrous natural Na-Mont,⁸ which was comparable to the results obtained using the intrinsically acidic Sn-Mont or H-Mont. During this chlorination, TMSCl and trace water included in the Na-Mont generates hydrogen chloride, which is then trapped by the Na-Mont and promotes the generation of the corresponding carbenium ions from the alcohols through protonation and dehydration.^9,10 The subsequent reaction between the carbenium ions and the chloride ions then forms the chlorinated products. Since chloride ions have a relatively poor nucleophilicity, we postulate that if a more powerful nucleophile than a chloride ion coexists in the system, the carbenium ions will preferentially react with the more reactive nucleophile to afford the corresponding substituted products. If this is correct, the alcohols bearing poorly-leaving hydroxy groups can be directly substituted with preferred nucleophiles simply by mixing the natural Na-Mont, TMSCl and a nucleophile in a flask.

To confirm our hypothesis, we used benzhydrol 1a as a model benzylic alcohol and trimethylsilyl azide (TMSN₃) as the nucleophile. TMSN₃ has previously been used with catalytic amounts of metal catalysts, such as Fe(OTf)₃,¹¹ magnetic nano Fe₃O₄,¹² and InBr₃,¹³ as well as with BF₃·Et₂O,¹⁴ for the direct azidation of alcohols. Our results are summarized in Table 1.

Table 1 Optimization of reaction conditions

Entry	Catalyst	Solvent	Reaction time	4a^a	5a^a
a The ^1H-NMR yield based on mesitylene as the internal standard.b 8 mol% of BF3·Et2O was employed.c Refer to ESI for details on the preparation of hydrous Na2SO4.d 50 mg of NaY was employed as the catalyst.e 1 equiv. of Me3SiN3 and 1 equiv. of TMSCl were employed. Benzhydryl chloride was obtained as a side product in 20% yield.f 0.06 equiv. of AcCl was used instead of TMSCl.g 0.003 equiv. of TMSCl was employed.h No TMSCl was used.i 1.5 equiv. of Me3SiN3 was employed.j A trace amount of 1a remained.
1	—	CH2Cl2	40 min	0%	0%
2^b	BF3·Et2O	CH2Cl2	40 min	0%	0%
3^c	Hydrous Na2SO4	CH2Cl2	40 min	0%	0%
4^d	Na-Y	CH2Cl2	40 min	0%	0%
5^e	Na-Mont	CH2Cl2	40 min	79%	0%
6	Na-Mont	CH2Cl2	40 min	99%	0%
7^f	Na-Mont	CH2Cl2	40 min	98%	1%
8	H-Mont	CH2Cl2	40 min	99%	0%
9	Sn-Mont	CH2Cl2	40 min	99%	0%
10^g	Na-Mont	CH2Cl2	40 min	3%	0%
11^h	Na-Mont	CH2Cl2	40 min	0%	0%
12^i	Na-Mont	CH2Cl2	40 min	94%	2%
13	Na-Mont	Hexane	40 min	98%	≪1%
14	Na-Mont	Et2O	40 min	54%	5%
15	Na-Mont	Acetone	40 min	0%	18%
16^j	Na-Mont	CH2Cl2	10 min	93%	2%

---

As we expected, when a mixture of 1a, Na-Mont, 1 equiv. of TMSN₃, and 1 equiv. of TMSCl was stirred at room temperature (RT) for 40 min, the corresponding azide product 4a was obtained in 79% yield (entry 5 of Table 1) together with benzhydryl chloride in 20% yield, as a result of the simultaneous chlorination and azidation of the intermediary carbenium ions. In the hope of suppressing the chlorination, the reaction was performed again under the same conditions except using a catalytic amount (6 mol%) of TMSCl. After 40 min, we were surprised to find that 4a was formed in 99% yield together with benzhydryl chloride in less than 0.1% yield (entry 6). When the reaction was performed with a catalytic amount (6 mol%) of acetyl chloride (AcCl), which also deliberated HCl upon contact with water, 4a was formed in 98% yield together with 5a in 1% yield (entry 7). No reaction proceeded when a combination of Na-Mont and TMSN₃ or that of TMSCl and TMSN₃ was applied (entries 1 and 11), concluding that the combination of Na-Mont, TMSCl, and TMSN₃ are essential to the successful azidation. When the amount of TMSCl was reduced from 6 mol% to 0.3 mol%, 4a was only obtained in very low yield (entry 10), suggesting that the efficiency of the azidation significantly depends on the amount of TMSCl employed.

Surprisingly, when 8 mol% BF₃·Et₂O, a representative homogenous acid catalyst, was employed as a control catalyst, the formation of 4a was not observed.¹⁵ We also performed the reaction using typical acidic montmorillonites, such as H-Mont or Sn-Mont, as the acid catalysts. In each case, azide product 4a was obtained in almost quantitative yield (entries 8 and 9). A survey of the solvent effects disclosed that undried CH₂Cl₂ (water content: <0.2%) was the solvent of choice (entries 6, 13–15).

By use of the optimized conditions shown in entry 6, the scope of benzylic and other alcohols as well as benzylic ethers was next investigated. The results are summarized in Fig. 1 and 2.

Fig. 1 Scope of substrates. (a) The ¹H-NMR yield based on mesitylene as the internal standard. Reaction conditions: 1 mmol of 1a, 2 mmol of Me₃SiN₃, 0.06 mmol of TMSCl, 50 mg of Na-Mont and 5 mL of CH₂Cl₂. (b) In parentheses are the yields of products obtained after column chromatography. (c) The yield of the product obtained when 0.06 equiv. of AcCl was used instead of TMSCl. (d) Dehydration product was formed instead.

In general, the azidation method is applicable to various benzylic alcohols: benzhydrol 1a underwent the azidation in almost quantitative yield. Benzylic alcohol 1b bearing an electron-withdrawing chlorine, or those with electron-donating groups 1d and 1e gave the corresponding azides in almost quantitative yields. The azidation of benzylic alcohol 1c having a fluorine atom was also successful. 1-Arylethanols, such as 1f and 1g, also underwent the azidation in excellent yields. This method is also applicable for the azidation of allylic alcohols, such as 1h and 1i, yielding azides 4h and 4i in 90% and 96% isolated yields, respectively. It should be noted that for azide 4h, the azide ion is attached to the original position of the hydroxyl of 1h, not at the benzylic site. We compared the difference in the Gibbs free energy (G) between 4h and its possible regioisomer, 1-azido-1-phenyl-2-butene (4h′), calculated by Gaussian 03W at the B3LYP/6-31(d) level. According to the calculation, 4h is more stable by 5.1 kcal mol⁻¹ than 4h′, rationalizing the preferred formation of 4h to 4h′. The same site-selective substitution of 1h with a cyanide ion was also observed.⁶ Although simple benzyl alcohol 1j did not react at all, activated primary benzylic alcohol 1k underwent the azidation in 26% yield with TMSCl, but in 13% yield with AcCl, proving that using TMSCl is preferable to using AcCl as a source of HCl for the azidation reaction. An addition of a small amount of water to the reaction made no improvements of the poor yield, only showing the same yield of 26%.

Attempts to apply the method for the azidation of a secondary aliphatic alcohol 1l and a tertiary aliphatic alcohol 1m were not successful. On the other hand, tertiary benzylic alcohol 1n gave the corresponding azide in almost quantitative yield.

When the trimethylsilyl ether of benzhydrol 5b was subjected to the azidation conditions, the corresponding azide product was obtained in almost quantitative yield (Fig. 2).

Fig. 2 Azidation of dibenzhydril ether (5a) and benzhydril trimethylsilyl ether (5b). (a) The ¹H-NMR yield based on mesitylene as the internal standard. Reaction conditions: 0.5 mmol of 5a or 1 mmol of 5b, 2 mmol of Me₃SiN₃, 0.06 mmol of TMSCI, 50 mg of Na-Mont and 5 mL of CH₂Cl₂. (b) Na-Mont was activated at 135 °C in vacuum for 1 h prior to the reaction, CH₂Cl₂ was distilled from CaH₂ and used immediately after the distillation.

During the optimization of the reaction conditions, we observed that when a mixture of 1a, Na-Mont, TMSCl, and TMSN₃ was stirred for 10 min at RT, 4a was afforded in 93% yield together with the symmetrical ether of benzhydrol 5a in 2% yield (Table 1, entry 16). When the stirring period was prolonged, the yield of 4a increased along with the disappearance of 5a, indicating that 5a seemed to be one of the intermediates in the azidation.

To further elucidate the mechanism of the azidation, 5a was subjected to the standard azidation conditions. After a 40 min reaction, 4a was formed in 49% yield together with 49% of the intact 5a. In contrast, when the reaction was performed using activated Na-Mont and anhydrous CH₂Cl₂, the yield of 4a dropped to 9%, suggesting the important role of trace water in the azidation (Fig. 2). Based on these results together with the other observations on the azidation, we proposed a plausible mechanism for this azidation method as shown in Fig. 3.

Fig. 3 Plausible mechanism.

The reaction is most likely initiated by the hydrolysis of TMSCl by water included in the Na-Mont or the solvent to generate an equimolar amount of HCl, which is then trapped by the Na-Mont. The HCl inside the Na-Mont promotes the generation of a carbenium ion from an alcohol through protonation and subsequent dehydration. The involvement of carbenium ion-like species during the azidation reaction was confirmed by the azidation of (R)-1-phenylethanol (98% ee) which yielded a racemic mixture of the corresponding azides in 37% yield.¹⁶ The carbenium ion then reacts with TMSN₃ to generate the corresponding azide product as well as trimethylsilanol which is produced by the reaction with water. The regeneration of a proton in the last step of the reaction explains why the subsequent azidation can be repeated without any need to supply more TMSCl.¹⁷

One of the most important applications of azides in organic synthesis is the Cu(I)-catalysed Huisgen cycloadditions of azides to alkynes to produce 1,2,3-triazoles.¹⁸ We also investigated the possibility to perform a one-pot synthesis of 1,2,3-triazoles from benzylic alcohols via the azidation and the successive click reaction.¹⁹

After a screening of the reaction conditions, we found that the combination of CuI catalyst with Et₃N and propanoic acid was suitable to perform the one-pot synthesis of 1,2,3-triazole derivatives. A 0.5 equiv. of CuI was employed in the one-pot synthesis since cycloaddition involving bulky azides, such as 4a, tends to take a longer time to reach completion.²⁰ Benzylic azides 4a, 4f, and 4i underwent the one-pot cycloaddition to give the corresponding triazole derivatives in 99%, 76%, and 93% yields, respectively (Fig. 4). However, the attempt to synthesize the 1,2,3-triazole from a tertiary benzylic azide 4n failed, which is probably due to steric effects of such a bulky azide.²¹

Fig. 4 One-pot synthesis of 1,2,3-triazoles from alcohols. (a) The isolated yield from column chromatography with using hexane/ethyl acetate as eluents.(b) The reaction mixture was stirred for 4.5 h.

In conclusion, we developed a new and practical method to convert natural montmorillonite into a solid acid catalyst using a catalytic amount of TMSCl. The acidic montmorillonite effectively catalysed the azidation of various benzylic alcohols at ambient temperature. The reaction system allows the one-pot synthesis of 1,2,3-triazole derivatives in combination with the CuI catalysis.

Acknowledgements

We thank Prof. Shuichi Hiraoka of The University of Tokyo for the HRMS analysis, and Prof. Tohru Yamada and Dr Tatsuyuki Tsubo of Keio University for the chiral HPLC analysis.

Notes and references

1. Solid Supports and Catalysts in Organic Synthesis, ed. K. Smith, Ellis Horwood and Prentice Hall, 1992, pp. 100–129.
2. H-Mont: K. Motokura, N. Nakagiri, T. Mizugaki, K. Ebitani and K. Kaneda, J. Org. Chem., 2007, 72, 6006–6015; K. Motokura, N. Fujita, K. Mori, T. Mizugaki, K. Ebitani and K. Kaneda, Angew. Chem., Int. Ed., 2006, 45, 2605–2609.
3. Sn-Mont: Y. Masui, J. Wang, T. Teramura, T. Kogure, T. Tanaka and M. Onaka, Microporous Mesoporous Mater., 2014, 198, 129–138.
4. J. Wang, Y. Masui and M. Onaka, Tetrahedron Lett., 2010, 51, 3300–3303.
5. J. Wang, Y. Masui and M. Onaka, Synlett, 2010, 2493–2497.
6. J. Wang, Y. Masui and M. Onaka, ACS Catal., 2011, 1, 446–454.
7. M. A. Tandiary, Y. Masui and M. Onaka, Tetrahedron Lett., 2014, 55, 4160–4162.
8. M. A. Tandiary, Y. Masui and M. Onaka, Synlett, 2014, 25, 2639–2643.
9. Hydrous Na-Mont was obtained from Kunimine Industries, Japan, and contained 14 wt% water.
10. X.-R. Song, B. Song, Y.-F. Qiu, Y.-P. Han, Z.-H. Qiu, X.-H. Hao, X.-Y. Liu and Y.-M. Liang, J. Org. Chem., 2014, 79, 7616–7625.
11. L. Y. Chan, S. Kim, W. T. Chung, C. Long and S. Kim, Synlett, 2011, 415–419.
12. N. A. N. Aslam, S. A. Babu, D. K. Singh and A. Rana, Synlett, 2014, 25, 2201–2207.
13. A. Kumar, R. K. Sharma, T. V. Singh and P. Venugopalan, Tetrahedron, 2013, 69, 10724–10732.
14. (a) A. Koziara and A. Zwierzak, Tetrahedron Lett., 1987, 28, 6513–6516; (b) G. Srinu and P. Srihari, Tetrahedron Lett., 2013, 54, 2382–2385.
15. Examples of the azidation by the catalytic use of BF₃·Et₂O have been reported only for the aryl vinyl carbinols. See ref. 14b for details.
16. See ESI for details.†.
17. See ESI for details.†.
18. For reviews on the synthesis of 1,2,3-triazoles based on the cycloaddition of azides with alkynes see: (a) V. D. Bock, H. Hiemstra and J. H. van Maarseveen, Eur. J. Org. Chem., 2006, 51–68; (b) J. E. Hein and V. V. Fokin, Chem. Soc. Rev., 2010, 39, 1302–1315.
19. For examples of the one-pot synthesis of 1,2,3-triazoles with alcohols as the starting material see: (a) B. Sreedhar, P. S. Reddy and V. R. Krishna, Tetrahedron Lett., 2007, 48, 5831–5834; (b) N. A. Naveen Aslam, S. A. Babu, D. K. Singh and A. Rana, Synlett, 2014, 25, 2201–2207.
20. (a) B. Sreedhar, P. S. Reddy and V. R. Krishna, Tetrahedron Lett., 2007, 48, 5831–5834; (b) J. Doiron, A. H. Soultan, R. Richard, M. M. Toure, N. Picot, R. Richard, M. Cuperlovic-Culf, G. A. Robichaud and M. Touaibia, Eur. J. Med. Chem., 2011, 46, 4010–4024.
21. There have been no reports on the synthesis of the 1,2,3-triazole derivative from azide 4n to the best of our knowledge.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra13238f

---