Boron Lewis Acid Catalysis Enables the Direct Cyanation of Benzyl Alcohols by Means of Isonitrile as Cyanide Source

The development of an efficient and straightforward method for cyanation of alcohols is of great value. However, the cyanation of alcohols always requires toxic cyanide sources. Herein, an unprecedented synthetic application of an isonitrile as a safer cyanide source in B(C6F5)3-catalyzed direct cyanation of alcohols is reported. With this approach, a wide range of valuable α-aryl nitriles was synthesized in good to excellent yields (up to 98%). The reaction can be scaled up and the practicability of this approach is further manifested in the synthesis of an anti-inflammatory drug, naproxen. Moreover, experimental studies were performed to illustrate the reaction mechanism.


Introduction
The need for the development of new reactions that are based on applying the atomeconomy concept [1] and avoiding the use of toxic reagents has become a consensus. Alcohols are highly attractive starting materials for synthesis because they are stable, have low toxicity, and are available. Direct nucleophilic substitution of an alcohol is attractive since water is, in principle, the only by-product [2][3][4][5]. However, this reaction is difficult because hydroxide is such a poor leaving group and therefore alcohols are classically derivatized to halides or pseudohalides prior to substitution, which results in the formation of vast amounts of waste. Thus, the development of new catalytic methodologies for dehydrative substitutions of alcohols was considered a central issue, as demonstrated by the inclusion of the "direct substitution of alcohols" in the ACS Green Chemistry Institute ® Pharmaceutical Roundtable's 2018 update on key green chemistry research areas [6]. In this context, the deoxygenative cyanation of readily available benzyl alcohols represent one of the most powerful methods for preparing α-aryl nitriles [7][8][9][10][11][12][13][14][15][16][17][18][19][20], an important class of core structures found in bioactive moleculars [21] and functional materials [22], and precursors that have applications in the synthesis of well-known drugs such as verapamil [23], naproxen [24], and cytenamide [25] as shown in Figure 1.
However, these methods suffer from major disadvantages such as the presence of hazardous and toxic cyanide sources and the use of stoichiometric activating reagents.
Recently, catalytic synthesis of α-aryl nitrile from benzyl alcohols was successfully developed. Ding's group performed pioneering work on the direct cyanation of α-aryl alcohols with trimethylsilyl cyanide (TMSCN) by indium halide catalysis [31]. Later, other Lewis acids, such as FeCl 3 ·6H 2 O [32], Zn(OTf) 2 [33], Brønsted acid montmorillonite The need for the development of new reactions that are based on applying the atomeconomy concept [1] and avoiding the use of toxic reagents has become a consensus. Alcohols are highly attractive starting materials for synthesis because they are stable, have low toxicity, and are available. Direct nucleophilic substitution of an alcohol is attractive since water is, in principle, the only by-product [2][3][4][5]. However, this reaction is difficult because hydroxide is such a poor leaving group and therefore alcohols are classically derivatized to halides or pseudohalides prior to substitution, which results in the formation of vast amounts of waste. Thus, the development of new catalytic methodologies for dehydrative substitutions of alcohols was considered a central issue, as demonstrated by the inclusion of the "direct substitution of alcohols" in the ACS Green Chemistry Institute ® Pharmaceutical Roundtable's 2018 update on key green chemistry research areas [6]. In this context, the deoxygenative cyanation of readily available benzyl alcohols represent one of the most powerful methods for preparing α-aryl nitriles [7][8][9][10][11][12][13][14][15][16][17][18][19][20], an important class of core structures found in bioactive moleculars [21] and functional materials [22], and precursors that have applications in the synthesis of well-known drugs such as verapamil [23], naproxen [24], and cytenamide [25] as shown in Figure 1.  As early as 1967, the one-pot method for the conversion of alcohols into cyanides based on the concept of the Mitsunobu reaction using NaCN as the cyanide source has been described [26]. Subsequently, there are a few reports on one-pot transformations of alcohols to α-aryl nitrile using Me3SiCl/NaI/NaCN [27], PPh3/nBu4NCN/DDQ [28], N-(ptoluenesulfonyl)imidazole (TsIm)/NaCN [29], and PPh3/DEAD/acetone cyanohydrin [30].

MeO
However, these methods suffer from major disadvantages such as the presence of hazardous and toxic cyanide sources and the use of stoichiometric activating reagents.
In recent years, B(C 6 F 5 ) 3 as a non-metallic Lewis acid has received widespread attention because of its strong Lewis acidity, commercial availability, and environmental friendliness [62][63][64][65][66][67][68]. Although still limited in its success, it mainly involves the B(C 6 F 5 ) 3 -catalyzed activation of hydroxyl groups, as reported by Meng, Zhao and Chan [69,70], Marek [71], Tang [72], Maji [73], Gevorgyan [74], and Moran [75]. Inspired by these reports and building on our ongoing interest in the developing atom-economic reactions [76][77][78][79], we questioned if the direct cyanation of alcohols with isocyanides in the presence of B(C 6 F 5 ) 3 could be realized to meet the requirements of atom economy and green chemistry (Scheme 1C). However, this hypothesis may face considerable challenges, such as the following: (a) the catalyst should be stable in wet and Lewis basic conditions, and (b) tert-butylisocyanide/B(C 6 F 5 ) 3 and nitrile/B(C 6 F 5 ) 3 adducts can be easily formed, as reported by Berke and Erker and co-workers [80]. It is unknown whether B(C 6 F 5 ) 3 can maintain its catalytic activity during the current cyanation reaction, and (c) the catalyst should be able to dissociate from nitrile products. Finally, (d) another challenge is to suppress the B(C 6 F 5 ) 3 -catalyzed homoetherification of alcohols reported by Chan and co-workers [70].
With the optimized conditions identified, we then proceeded to explore the scope of isocyanides (Scheme 2). Besides tBu-NC (2a), other tertiary amine-derived isocyanides, such as 2b and 2c, can also be used as a novel cyano source in the current direct cyanation of α-aryl alcohols. In contrast, when secondary amine-and aniline-derived isocyanides were used as substrates, only the corresponding ether products were obtained (not shown). Of note, treatment of 1a and 2c with the standard conditions afforded 3a in 81% NMR yield together with internal alkene 5 in 53% NMR yield and terminal alkene 6 in 38% NMR yield (Scheme 2b), indicating a tertiary carbon cation might be an intermediate.
Next, we turned to explore the generality of this cyanidation reaction with a variety of alcohols with tBu-NC (2a). As shown in Scheme 3, a wide range of benzylic alcohols can smoothly react with 2a under the optimized conditions, giving the corresponding α-aryl nitriles in good to excellent yields. Diarylsubstituted alcohols (R 2 = aryl) underwent reaction with tBu-NC to furnish the corresponding products (3a-3k) in 19−98% yields.  Diarylsubstituted alcohols bearing electron-donating (methoxy, alkyl) groups at the para-, ortho-or meta-positions of the benzene rings reacted smoothly (1a−1f). A variety of electron-withdrawing functional groups at the para-positions of the benzene rings such as −Br, −CF3, and -CO2Me were tolerated, affording the desired products in moderate to good yields (1g−1i). However, a low yield of 3j was isolated from 1j having an ortho-nitro group on the aromatic substituent. The naphthyl-containing alcohol 1k and heteroarylcontaining alcohol 1l also afforded the corresponding products in good to excellent yields. In addition, alkyl-substituted alcohols (R 2 = methyl, ethyl, and tert-butyl) were also welltolerated to afford the desired products 3m−3o. Of note, a competitive side reaction encountered in the reaction with 1m or 1n is the formation of styrene derivatives via the dehydration reactions of alcohols. Besides R 1 = methoxy (1p−1q), the substituent R 1 can be benzamide (1r). Benzhydrol 1s was tested but afforded the corresponding product 3s in low yield together with (oxybis(methanetriyl))tetrabenzene in 64% NMR yield. However,  Diarylsubstituted alcohols bearing electron-donating (methoxy, alkyl) groups at the para-, ortho-or meta-positions of the benzene rings reacted smoothly (1a−1f). A variety of electron-withdrawing functional groups at the para-positions of the benzene rings such as −Br, −CF3, and -CO2Me were tolerated, affording the desired products in moderate to good yields (1g−1i). However, a low yield of 3j was isolated from 1j having an ortho-nitro group on the aromatic substituent. The naphthyl-containing alcohol 1k and heteroarylcontaining alcohol 1l also afforded the corresponding products in good to excellent yields. In addition, alkyl-substituted alcohols (R 2 = methyl, ethyl, and tert-butyl) were also welltolerated to afford the desired products 3m−3o. Of note, a competitive side reaction encountered in the reaction with 1m or 1n is the formation of styrene derivatives via the dehydration reactions of alcohols. Besides R 1 = methoxy (1p−1q), the substituent R 1 can be benzamide (1r). Benzhydrol 1s was tested but afforded the corresponding product 3s in low yield together with (oxybis(methanetriyl))tetrabenzene in 64% NMR yield. However, Diarylsubstituted alcohols bearing electron-donating (methoxy, alkyl) groups at the para-, orthoor meta-positions of the benzene rings reacted smoothly (1a-1f). A variety of electron-withdrawing functional groups at the para-positions of the benzene rings such as -Br, -CF 3 , and -CO 2 Me were tolerated, affording the desired products in moderate to good yields (1g-1i). However, a low yield of 3j was isolated from 1j having an ortho-nitro group on the aromatic substituent. The naphthyl-containing alcohol 1k and heteroarylcontaining alcohol 1l also afforded the corresponding products in good to excellent yields. In addition, alkyl-substituted alcohols (R 2 = methyl, ethyl, and tert-butyl) were also welltolerated to afford the desired products 3m-3o. Of note, a competitive side reaction encountered in the reaction with 1m or 1n is the formation of styrene derivatives via the dehydration reactions of alcohols. Besides R 1 = methoxy (1p-1q), the substituent R 1 can be benzamide (1r). Benzhydrol 1s was tested but afforded the corresponding product 3s in low yield together with (oxybis(methanetriyl))tetrabenzene in 64% NMR yield. However, 1,2,3,4-tetrahydronaphthalen-1-ol (1t) and allylic alcohol (1u) underwent smooth cyanation. It is worth noting that the reported indium halide-catalyzed protocol with TMSCN as the cyanide source is not amenable to primary alcohol 1v for cyanation reaction [31]. Our system, however, gives reasonable yield for the same substrate. To our delight, the precursors for naproxen and cytenamide, respectively, can also be efficiently obtained in high yields by this protocol (3w and 3x).
Our system, however, gives reasonable yield for the same substrate. To our delight, the precursors for naproxen and cytenamide, respectively, can also be efficiently obtained in high yields by this protocol (3w and 3x). To disclose the synthetic practicability of the developed method, we studied a gramscale reaction, and 66% yield of 3w was obtained, which might provide potential value in To disclose the synthetic practicability of the developed method, we studied a gramscale reaction, and 66% yield of 3w was obtained, which might provide potential value in synthetic chemistry. Having established a protocol for the efficient synthesis of α-aryl nitrile 3w, (±)-naproxen, a nonsteroidal anti-inflammatory drug [24], was prepared in three steps from commercially available materials (Scheme 4).
Molecules 2023, 28, x FOR PEER REVIEW 6 of 15 synthetic chemistry. Having established a protocol for the efficient synthesis of α-aryl nitrile 3w, (±)-naproxen, a nonsteroidal anti-inflammatory drug [24], was prepared in three steps from commercially available materials (Scheme 4).

Scheme 4. Scale-up synthesis and synthetic transformations.
To gain insight into the reaction mechanism, several control experiments were conducted. When an enantiomerically pure sample of (R)-1y was subjected to standard conditions, the resulting nitrile 3y was obtained in racemic form (Scheme 5a). Moreover, (R)-1y was also employed to perform the etherification reaction under the standard condition. The 1 H NMR spectrum showed the ether 4y was obtained in 50% NMR yield with a 48/52 ratio of the two diastereomers (Scheme 5b). These control experiments support an SN1 pathway and rule out a concerted SN2 mechanism. The tert-Butylisocyanide-B(C6F5)3 adduct (8) was easily prepared [80] and used as a catalyst in the current reaction, affording the corresponding product 3a in 99% NMR yield (Scheme 5c). As shown in Table 1, entry 10, the reaction did form 3a in 20% NMR yield along with ether 4a in 30% NMR yield at 80 °C. Treatment of 4a with the standard setup then gave the desired 3a in 70% NMR yield (Scheme 5d).

Scheme 5. Control experiments.
Based on the results above, a plausible mechanism is proposed in Scheme 6. First, tert-butylisocyanide 2a forms a reversibly Lewis adduct 8 with B(C6F5)3 [80]. The homoetherification of alcohol in the presence of the B(C6F5)3 quickly delivers the ether 4 [70], To gain insight into the reaction mechanism, several control experiments were conducted. When an enantiomerically pure sample of (R)-1y was subjected to standard conditions, the resulting nitrile 3y was obtained in racemic form (Scheme 5a). Moreover, (R)-1y was also employed to perform the etherification reaction under the standard condition. The 1 H NMR spectrum showed the ether 4y was obtained in 50% NMR yield with a 48/52 ratio of the two diastereomers (Scheme 5b). These control experiments support an S N 1 pathway and rule out a concerted S N 2 mechanism. The tert-Butylisocyanide-B(C 6 F 5 ) 3 adduct (8) was easily prepared [80] and used as a catalyst in the current reaction, affording the corresponding product 3a in 99% NMR yield (Scheme 5c). As shown in Table 1, entry 10, the reaction did form 3a in 20% NMR yield along with ether 4a in 30% NMR yield at 80 • C. Treatment of 4a with the standard setup then gave the desired 3a in 70% NMR yield (Scheme 5d).
Molecules 2023, 28, x FOR PEER REVIEW 6 of 15 synthetic chemistry. Having established a protocol for the efficient synthesis of α-aryl nitrile 3w, (±)-naproxen, a nonsteroidal anti-inflammatory drug [24], was prepared in three steps from commercially available materials (Scheme 4). To gain insight into the reaction mechanism, several control experiments were conducted. When an enantiomerically pure sample of (R)-1y was subjected to standard conditions, the resulting nitrile 3y was obtained in racemic form (Scheme 5a). Moreover, (R)-1y was also employed to perform the etherification reaction under the standard condition. The 1 H NMR spectrum showed the ether 4y was obtained in 50% NMR yield with a 48/52 ratio of the two diastereomers (Scheme 5b). These control experiments support an SN1 pathway and rule out a concerted SN2 mechanism. The tert-Butylisocyanide-B(C6F5)3 adduct (8) was easily prepared [80] and used as a catalyst in the current reaction, affording the corresponding product 3a in 99% NMR yield (Scheme 5c). As shown in Table 1, entry 10, the reaction did form 3a in 20% NMR yield along with ether 4a in 30% NMR yield at 80 °C. Treatment of 4a with the standard setup then gave the desired 3a in 70% NMR yield (Scheme 5d).

Scheme 5. Control experiments.
Based on the results above, a plausible mechanism is proposed in Scheme 6. First, tert-butylisocyanide 2a forms a reversibly Lewis adduct 8 with B(C6F5)3 [80]. The homoetherification of alcohol in the presence of the B(C6F5)3 quickly delivers the ether 4 [70],

Scheme 5. Control experiments (a-d).
Based on the results above, a plausible mechanism is proposed in Scheme 6. First, tert-butylisocyanide 2a forms a reversibly Lewis adduct 8 with B(C 6 F 5 ) 3 [80]. The homoetherification of alcohol in the presence of the B(C 6 F 5 ) 3 quickly delivers the ether 4 [70], which could furnish an adduct 9 with B(C 6 F 5 ) 3 through the oxygen center. Subsequently, the adduct 9 could break into an intermediate 10 and carbocation 11. However, an alternative reaction path for the formation of the carbocation 11 directly from alcohol in the presence of the in situ-generated strong Brønsted acid B(C 6 F 5 ) 3 ·nH 2 O or boron Lewis acid B(C 6 F 5 ) 3 (not shown, see Supporting Information for details) cannot be ruled out [69,70]. The carbocation 11 could then be intercepted instantaneously by the tBu-NC (2a) to afford an intermediate 12 with a borate anion 10 as the counteranion. The stability of the tertiary carbon cation is the driving force to break the C-N bond in 12, leading to the α-aryl nitrile 3 and 2-methylpropene by proton elimination via a tert-butyl carbon cation intermediate (supported by Scheme 2b) [81]. The borate anion 10, on the other hand, transforms into alcohol 1 with the regeneration of the B(C 6 F 5 ) 3 catalyst.
Molecules 2023, 28, x FOR PEER REVIEW 7 of 15 which could furnish an adduct 9 with B(C6F5)3 through the oxygen center. Subsequently, the adduct 9 could break into an intermediate 10 and carbocation 11. However, an alternative reaction path for the formation of the carbocation 11 directly from alcohol in the presence of the in situ-generated strong Brønsted acid B(C6F5)3 nH2O or boron Lewis acid B(C6F5)3 (not shown, see Supporting Information for details) cannot be ruled out [69,70]. The carbocation 11 could then be intercepted instantaneously by the tBu-NC (2a) to afford an intermediate 12 with a borate anion 10 as the counteranion. The stability of the tertiary carbon cation is the driving force to break the C-N bond in 12, leading to the α-aryl nitrile 3 and 2-methylpropene by proton elimination via a tert-butyl carbon cation intermediate (supported by Scheme 2b) [81]. The borate anion 10, on the other hand, transforms into alcohol 1 with the regeneration of the B(C6F5)3 catalyst. Scheme 6. Proposed mechanism.

General Information
All reactions were performed in flame-dried glassware using conventional Schlenk techniques under a static pressure of nitrogen unless otherwise stated. Liquids and solutions were transferred with syringes. The known alcohols 1 [31,35] and tBu-NC-B(C5F5)3 adduct 8 [80] were prepared according to reported procedures. (R)-1y was prepared in 80% yield according to the known procedure [82] (95% ee of (R)-1y was determined by HPLC: OJ-H Column, 5/95 iPrOH/hexane, 0.5 mL/min, 254 nm, 35 °C; retention time = 75.36 min (minor), 81.66 min (major)). Tris(pentafluorophenyl)borane (B(C5F5)3, 98%, Energy Chemical) and tert-butyl isocyanate (97%, Energy Chemical) were purchased from commercial suppliers and used as received. Other commercially available reagents were purchased from Sigma-Adrich, Leyan and Bide Chemical Company. All solvents (tetrahydrofuran, toluene, and 1,2-dichloroethane etc.) were dried and purified following standard procedures. Technical grade solvents for extraction or chromatography (petroleum ether, CH2Cl2, and ethyl acetate) were distilled prior to use. Analytical thin layer chromatography (TLC) was performed on silica gel 60 F254 glass plates by Merck. Flash column chromatography was performed on silica gel 60 (40-63 μm, 230-400 mesh, ASTM) by Grace using the indicated solvents. 1 H, 13 C NMR spectra (Supplementary Materials) were recorded in CDCl3 on Bruker AV400 instruments. Chemical shifts are reported in parts per million (ppm) and are referenced to the residual solvent resonance as the internal standard (CHCl3: δ = 7.26 ppm for 1 H NMR and CDCl3: δ = 77.0 ppm for 13 C NMR). Data are reported as follows: chemical shift, multiplicity (s = singlet, d = doublet, t = triplett, q = quartet, m = multiplet), coupling constants (Hz), and integration. Mass spectra were recorded on a THERMO FINNIGAN LTQ-XL. The MS inlet capillary temp was always maintained at 275 °C and capillary voltage at 5 kV. No other source gases were used when Scheme 6. Proposed mechanism.

General Information
All reactions were performed in flame-dried glassware using conventional Schlenk techniques under a static pressure of nitrogen unless otherwise stated. Liquids and solutions were transferred with syringes. The known alcohols 1 [31,35] and tBu-NC-B(C 5 F 5 ) 3 adduct 8 [80] were prepared according to reported procedures. (R)-1y was prepared in 80% yield according to the known procedure [82] (95% ee of (R)-1y was determined by HPLC: OJ-H Column, 5/95 iPrOH/hexane, 0.5 mL/min, 254 nm, 35 • C; retention time = 75.36 min (minor), 81.66 min (major)). Tris(pentafluorophenyl)borane (B(C 5 F 5 ) 3 , 98%, Energy Chemical) and tert-butyl isocyanate (97%, Energy Chemical) were purchased from commercial suppliers and used as received. Other commercially available reagents were purchased from Sigma-Adrich, Leyan and Bide Chemical Company. All solvents (tetrahydrofuran, toluene, and 1,2-dichloroethane etc.) were dried and purified following standard procedures. Technical grade solvents for extraction or chromatography (petroleum ether, CH 2 Cl 2 , and ethyl acetate) were distilled prior to use. Analytical thin layer chromatography (TLC) was performed on silica gel 60 F254 glass plates by Merck. Flash column chromatography was performed on silica gel 60 (40-63 µm, 230-400 mesh, ASTM) by Grace using the indicated solvents. 1 H, 13 C NMR spectra (Supplementary Materials) were recorded in CDCl 3 on Bruker AV400 instruments. Chemical shifts are reported in parts per million (ppm) and are referenced to the residual solvent resonance as the internal standard (CHCl 3 : δ = 7.26 ppm for 1 H NMR and CDCl 3 : δ = 77.0 ppm for 13 C NMR). Data are reported as follows: chemical shift, multiplicity (s = singlet, d = doublet, t = triplett, q = quartet, m = multiplet), coupling constants (Hz), and integration. Mass spectra were recorded on a THERMO FINNIGAN LTQ-XL. The MS inlet capillary temp was always maintained at 275 • C and capillary voltage at 5 kV. No other source gases were used when digestion was performed in microdroplets. The samples were dissolved in 1:1 methanol:water.

Typical Procedure for Direct Cyanation of Alcohols
In a glove box, alcohol 1 (0.2 mmol), isocyanide 2 (0.3 mmol, 1.5 equiv), B(C 6 F 5 ) 3 (10.2 mg, 20 µmol, 10 mol%), and toluene (2.0 mL) were added to an oven-dried 10 mL pressure vial. The vial was sealed and removed from the glove box. The reaction mixture Molecules 2023, 28, 2174 8 of 15 was stirred at 100 • C (oil bath) for 2-18 h. After the reaction was completed, the reaction mixture was purified by silica gel column chromatography by using petroleum ether/ethyl acetate mixture to obtain the desired nitrile 3.

Procedure for the Preparation of Naproxen
To a solution of 6-methoxy-2-naphthaldehyde (1.86 g, 10.0 mmol, 1.0 equiv) in THF (20 mL, 0.5 M), methylmagnesium bromide (4.0 mL, 12 mmol, 3.0 M, 1.2 equiv) was added. When the reaction was judged to have reached completion (as determined by TLC), sat. NH 4 Cl was added slowly at 0 • C, and the mixture was extracted with EtOAc. The combined organic layers were washed with brine, dried over MgSO 4 , and purified by column chromatography on silica gel to obtain 1w (1.60 g, 80% yield).

Conclusions
In conclusion, by taking advantage of isonitriles as low-toxic CN surrogates in the metal-free cyanation of alcohols, an efficient and green method for the direct catalytic synthesis of α-aryl nitriles was developed (up to 98% yield). To the best of our knowledge, this is the first B(C 6 F 5 ) 3 -catalyzed transformation of isonitriles. Control experiments support an S N 1 pathway and rule out a concerted S N 2 mechanism. The in situ-generated ether 4 can be converted to the desired α-aryl nitriles under the current catalytic system via cleavage of the C-O bond. The use of readily available starting materials, low catalyst loading, a broad substrate scope, ease of scale-up, and application in the synthesis of the precursors for naproxen and cytenamide make this approach very practical and attractive. With these advantages, we expect that this method will find wide applications in organic synthesis.