Different Performance of Two Isomeric Phosphinobiphenyl Amidosulfonates in Pd-Catalyzed Cyanation of Aryl Bromides

A hydrophilic phosphinobiphenyl amidosulfonate, 2′-(dicyclohexylphosphino)-2{[(sulfonatomethyl)amino]carbonyl}[1,1′-biphenyl], triethylammonium salt (L2), was prepared and, together with its isomer bearing the polar amido-sulfonate tag in the position 4 of the biphenyl scaffold (compound L1), evaluated as a supporting ligand in Pd-catalyzed cyanation of aryl bromides using K4[Fe(CN)6] as the non-toxic cyanide source. The less sterically demanding ligand L1 was found to form more active catalysts than the newly prepared compound L2. A catalyst formed in situ from palladium(II) acetate and L1 efficiently mediated cyanation of aryl bromides bearing electron-donating substituents but failed in the analogous reactions with electron-poor substrates.


Introduction
Palladium-catalyzed cross-coupling reactions [1][2][3] have become an indispensable part of the synthetic chemist toolbox in research laboratories as well as in industry.In addition to reactions proceeding with the formation of new C-C bonds between two organic fragments that are utilized mainly for assembling the molecular skeleton, there have been developed analogous processes allowing for the introduction of various functional groups.The latter reactions include metal-mediated cyanation of organic halides (mostly aromatic) which affords synthetically valued nitriles [4,5].Some years ago, it was shown that this particular reaction can be performed in aqueous reaction media and with K 4 [Fe(CN) 6 ] as a practically non-toxic cyanide source [6].However, these benefits are paid for by the need for a careful choice of a catalytic system suitable for use in aqueous reaction media [7][8][9][10][11][12] and optimization of the reaction conditions mainly aimed at suppressing unwanted hydrolysis of the nitriles (the primary products) to the corresponding amides.
We have demonstrated that the Pd-catalyzed cyanation of aryl bromides with K 4 [Fe(CN) 6 ] can be performed very well with simple catalysts formed in situ from palladium(II) acetate and ferrocene-based phosphino-amidosulfonate [13] and phosphino-urea [14] donors.Recently, we have also reported the synthesis of a water-soluble ligand L1 (Scheme 1), inspired by the highly efficient phosphinobiphenyl donors introduced by Buchwald et al. [15], and utilized this compound as a supporting ligand in Pd-catalyzed Suzuki-Miyaura biaryl coupling performed in water [16].Considering the good catalytic results achieved with the L1-Pd catalyst in pure water, we decided to extend our reaction tests toward biphasic reaction media.As the testing reaction, we chose the mentioned Pd-catalyzed cyanation of aryl bromides to benzonitriles and compared the properties of donor L1 with its sterically more encumbered isomer, 2-(dicyclohexylphosphino)-2 -{[(sulfonatomethyl)amino]carbonyl} [1,1 -biphenyl], triethylammonium salt (compound L2 in Scheme 1), the synthesis of which is also reported in this contribution.
Scheme 1. Ligands evaluated in this study.
In the same vein, ligand L2 was synthesized (Scheme 2) in five steps from commercially available 2,2′-dibromo[1,1′-biphenyl] (1).This compound was firstly lithiated and phosphinylated with ClPCy2 (Cy = cyclohexyl) to afford an intermediate phosphine-bromide, which was isolated in the form of the air-stable borane adduct 2. In the following step, the carboxyl function was introduced by lithiation and carboxylation with carbon dioxide.Phosphinocarboxylic acid 3 was then converted to active pentafluorophenyl ester 4 and the ester was in turn reacted with aminomethylsulfonic acid in the presence of a catalytic amount of 4-(dimethylamino)pyridine to afford the desired ligand in Pprotected form, compound 5. Subsequent deprotection with 1,4-diazabicyclo[2.2.2]octane (dabco) [20] afforded the target phosphinobiphenyl amidosulfonate, which was isolated as a white microcrystalline solid containing solvating ethyl acetate in nearly 30% yield with respect to the starting dibromide 1.
In its electrospray ionization (ESI) mass spectrum, L2 shows the peak due to the anion [L2 − HNEt3] − at m/z 486.The IR spectrum contains bands, which can be assigned to ν(NH) (3404 and 3309 cm −1 ), amide vibrations (1663 and 1533 cm −1 ) and ν(C=O) of the residual ethyl acetate (1735 cm −1 ) [21].IR bands attributable to the terminal sulfonate groups are observed at 1177 and 1032 cm −1 [22].The presence of the carboxamide unit is manifested in the 13 C NMR spectrum, showing a signal at δC 168.18.Together with the 31 P NMR spectrum (δP −12.6), the 13 C NMR spectrum further corroborates the presence of an intact phosphine moiety.Finally, the 1 H NMR spectrum of L2 contains complicated multiplets due to the phenyl and cyclohexyl groups, signals of the triethylammonium cation and a pair of double doublets due to the diastereotopic NHCH2 protons (δH 4.00 and 4.20).

Scheme 1.
Ligands evaluated in this study.
In the same vein, ligand L2 was synthesized (Scheme 2) in five steps from commercially available 2,2 -dibromo[1,1 -biphenyl] (1).This compound was firstly lithiated and phosphinylated with ClPCy 2 (Cy = cyclohexyl) to afford an intermediate phosphine-bromide, which was isolated in the form of the air-stable borane adduct 2. In the following step, the carboxyl function was introduced by lithiation and carboxylation with carbon dioxide.Phosphinocarboxylic acid 3 was then converted to active pentafluorophenyl ester 4 and the ester was in turn reacted with aminomethylsulfonic acid in the presence of a catalytic amount of 4-(dimethylamino)pyridine to afford the desired ligand in P-protected form, compound 5. Subsequent deprotection with 1,4-diazabicyclo[2.2.2]octane (dabco) [20] afforded the target phosphinobiphenyl amidosulfonate, which was isolated as a white microcrystalline solid containing solvating ethyl acetate in nearly 30% yield with respect to the starting dibromide 1.
In its electrospray ionization (ESI) mass spectrum, L2 shows the peak due to the anion [L2 − HNEt 3 ] − at m/z 486.The IR spectrum contains bands, which can be assigned to ν(NH) (3404 and 3309 cm −1 ), amide vibrations (1663 and 1533 cm −1 ) and ν(C=O) of the residual ethyl acetate (1735 cm −1 ) [21].IR bands attributable to the terminal sulfonate groups are observed at 1177 and 1032 cm −1 [22].The presence of the carboxamide unit is manifested in the 13 C NMR spectrum, showing a signal at δ C 168.18.Together with the 31 P NMR spectrum (δ P −12.6), the 13 C NMR spectrum further corroborates the presence of an intact phosphine moiety.Finally, the 1 H NMR spectrum of L2 contains complicated multiplets due to the phenyl and cyclohexyl groups, signals of the triethylammonium cation and a pair of double doublets due to the diastereotopic NHCH 2 protons (δ H 4.00 and 4.20).

Pd-Catalyzed Cyanation of Aryl Bromides
For the initial screening experiments aimed at finding the best ligand and palladium source (Scheme 3), we employed the cyanation of 4-bromoanisole (7e) to give 4-methoxybenzonitrile (8e; Scheme 4), which can be easily monitored by 1 H NMR spectroscopy.The reactions were performed in the presence of 1 mol % of a Pd-catalyst and potassium carbonate as the base at 100 °C in a 1,4dioxane-water (1:1) mixture.The catalysts were generated before the reaction by mixing the respective ligand with a palladium precursor in dichloromethane and evaporation of the resultant solution because some of the palladium precursors were not sufficiently soluble in water.

Pd-Catalyzed Cyanation of Aryl Bromides
For the initial screening experiments aimed at finding the best ligand and palladium source (Scheme 3), we employed the cyanation of 4-bromoanisole (7e) to give 4-methoxybenzonitrile (8e; Scheme 4), which can be easily monitored by 1 H NMR spectroscopy.The reactions were performed in the presence of 1 mol % of a Pd-catalyst and potassium carbonate as the base at 100 • C in a 1,4-dioxane-water (1:1) mixture.The catalysts were generated before the reaction by mixing the respective ligand with a palladium precursor in dichloromethane and evaporation of the resultant solution because some of the palladium precursors were not sufficiently soluble in water.

Pd-Catalyzed Cyanation of Aryl Bromides
For the initial screening experiments aimed at finding the best ligand and palladium source (Scheme 3), we employed the cyanation of 4-bromoanisole (7e) to give 4-methoxybenzonitrile (8e; Scheme 4), which can be easily monitored by 1 H NMR spectroscopy.The reactions were performed in the presence of 1 mol % of a Pd-catalyst and potassium carbonate as the base at 100 °C in a 1,4dioxane-water (1:1) mixture.The catalysts were generated before the reaction by mixing the respective ligand with a palladium precursor in dichloromethane and evaporation of the resultant solution because some of the palladium precursors were not sufficiently soluble in water.The results collected in Table 1 indicate a generally superior performance of catalysts based on L1 but do not allow formulating any clear-cut clear conclusions for the series of analogous catalysts based on L1 and L2.The yields of the coupling product varied greatly upon changing the Pd precursor as well as the ligand-to-Pd ratio.With ligand L1, which was used during the subsequent reaction scope tests, the catalysts resulting from palladium(II) acetate and palladium(II) chloride afforded better yields than those formed from [PdCl(η 3 -C3H5)]2 and even from Pd(II) precursors containing C,N-chelating ligands, [PdCl(L NC )]2 (L NC = 2-[(dimethylamino-κN)methyl]phenyl-C 1 ) and its recently introduced analogues 6a and 6b [23,24] (see Scheme 3) which are readily activated by elimination of the chelating hydrocarbyl ligand.Notably, when Pd(OAc)2 (1 mol %) was utilized as the pre-catalyst without any added supporting ligand, the model cyanation reaction did not proceed to any appreciable extent (0% conversion under the standard reaction conditions).We have also noted that the in situ formed pre-catalysts are highly sensitive to air.Their preparation as well as subsequent catalytic use required a careful protection of the reaction mixture from the air to achieve good and reproducible results.This can be demonstrated by a massive decrease of the yield of the coupling product 8e when the model cyanation reaction was carried out in the air (see entry 2 in Table 1).Catalyst decomposition was confirmed by 31 P NMR spectroscopy, which revealed a whole set of signals for samples exposed to air.It is also noteworthy that a "successful" catalytic run could be easily recognized visually because the reaction mixture turned black due to formation of Pd(0), while in the case of low conversions, it remained yellow-orange.
The following reaction scope tests were performed with the pre-catalyst generated from Pd(OAc)2 and L1 (2 equiv.), which achieved nearly the best results during the screening experiments and makes use of the relatively cheap and easy-to-handle Pd precursor [25].The results obtained with different aryl bromides are presented in Table 2. Apparently, the tested catalyst efficiently mediated cyanation of electron-rich substrates while affording practically negligible yields in the case of aryl bromides bearing electron-withdrawing substituents.The conversions of the latter substrates were not improved even upon extending the reaction time from 2 h to one day with the sole exception of 4-chloro-1-bromobenzene (7g), for which the conversion to 8g increased from less than 5% to 10% (note: the corresponding amides, possibly resulting from hydrolysis of the nitriles, were detected in only negligible amounts).This striking difference noted for the two classes of substrates may reflect The results collected in Table 1 indicate a generally superior performance of catalysts based on L1 but do not allow formulating any clear-cut clear conclusions for the series of analogous catalysts based on L1 and L2.The yields of the coupling product varied greatly upon changing the Pd precursor as well as the ligand-to-Pd ratio.With ligand L1, which was used during the subsequent reaction scope tests, the catalysts resulting from palladium(II) acetate and palladium(II) chloride afforded better yields than those formed from [PdCl(η 3 -C 3 H 5 )] 2 and even from Pd(II) precursors containing C,N-chelating ligands, [PdCl(L NC )] 2 (L NC = 2-[(dimethylamino-κN)methyl]phenyl-C 1 ) and its recently introduced analogues 6a and 6b [23,24] (see Scheme 3) which are readily activated by elimination of the chelating hydrocarbyl ligand.Notably, when Pd(OAc) 2 (1 mol %) was utilized as the pre-catalyst without any added supporting ligand, the model cyanation reaction did not proceed to any appreciable extent (0% conversion under the standard reaction conditions).We have also noted that the in situ formed pre-catalysts are highly sensitive to air.Their preparation as well as subsequent catalytic use required a careful protection of the reaction mixture from the air to achieve good and reproducible results.This can be demonstrated by a massive decrease of the yield of the coupling product 8e when the model cyanation reaction was carried out in the air (see entry 2 in Table 1).Catalyst decomposition was confirmed by 31 P NMR spectroscopy, which revealed a whole set of signals for samples exposed to air.It is also noteworthy that a "successful" catalytic run could be easily recognized visually because the reaction mixture turned black due to formation of Pd(0), while in the case of low conversions, it remained yellow-orange.
The following reaction scope tests were performed with the pre-catalyst generated from Pd(OAc) 2 and L1 (2 equiv.), which achieved nearly the best results during the screening experiments and makes use of the relatively cheap and easy-to-handle Pd precursor [25].The results obtained with different aryl bromides are presented in Table 2. Apparently, the tested catalyst efficiently mediated cyanation of electron-rich substrates while affording practically negligible yields in the case of aryl bromides bearing electron-withdrawing substituents.The conversions of the latter substrates were not improved even upon extending the reaction time from 2 h to one day with the sole exception of 4-chloro-1-bromobenzene (7g), for which the conversion to 8g increased from less than 5% to 10% (note: the corresponding amides, possibly resulting from hydrolysis of the nitriles, were detected in only negligible amounts).This striking difference noted for the two classes of substrates may reflect changes in the reaction energetics (reductive elimination of ArCN is slower for electron-poor aryl groups [26]) and also a relatively low stability of the formed catalyst.changes in the reaction energetics (reductive elimination of ArCN is slower for electron-poor aryl groups [26]) and also a relatively low stability of the formed catalyst.In summary, the newly synthesized hydrophilic phosphinobiphenyl ligand L2 afforded less stable and less efficient catalysts for the Pd-catalyzed cyanation of aryl bromides with K4[Fe(CN)6] in 1,4-dioxane-water mixtures than its less sterically congested isomer L1.Catalysts formed in situ from Pd(OAc)2 and L1 efficiently mediated cyanation of aryl bromides bearing electron-donating substituents, being more active than the analogous catalysts obtained from Pd(OAc)2 and hybrid phosphinoferrocene ligands reported earlier [5,6].However, the latter catalytic systems appear to be more robust, allowing cyanation of less reactive electron-poor substrates for which the Pd(OAc)2/L1 catalyst failed.

Materials and Methods
Commercial K4[Fe(CN)6]•3H2O was dehydrated under vacuum prior to the use (ca.30 °C/10 Torr; Note: the reaction proceeds equally well with hydrated K4[Fe(CN)6] but the anhydrous salt was utilized to avoid possible variations in the reaction stoichiometry resulting from the use of partly dehydrated materials).Anhydrous tetrahydrofuran, dichloromethane and methanol were dried using a PureSolv MD5 Solvent Purification System (Innovative Technology Inc., Amesbury, MA, USA).Anhydrous N,N-dimethylformamide (Sigma-Aldrich, St. Louis, MO, USA) and solvents (Lach-Ner, Neratovice, Czech Republic) used for aqueous workup, column chromatography and crystallizations were used as received.Compound L1 [8]   13 C NMR), to external 85% aqueous H3PO4 ( 31 P NMR) or to external neat CFCl3 ( 19 F NMR). Mass spectra were recorded on an Esquire 3000 (Bruker) or an LTQ Orbitrap XL instrument (Thermo Fisher Scientific, Waltham, MA, USA; high resolution analyses).Infrared spectra were collected in Nujol mulls with an FTIR Thermo Nicolet 760 instrument (Thermo Fisher Scientific, Waltham, MA, USA) in the range 400-4000 cm −1 .

Materials and Methods
Commercial K 4 [Fe(CN) 6 ]•3H 2 O was dehydrated under vacuum prior to the use (ca.30 • C/10 Torr; Note: the reaction proceeds equally well with hydrated K 4 [Fe(CN) 6 ] but the anhydrous salt was utilized to avoid possible variations in the reaction stoichiometry resulting from the use of partly dehydrated materials).Anhydrous tetrahydrofuran, dichloromethane and methanol were dried using a PureSolv MD5 Solvent Purification System (Innovative Technology Inc., Amesbury, MA, USA).Anhydrous N,N-dimethylformamide (Sigma-Aldrich, St. Louis, MO, USA) and solvents (Lach-Ner, Neratovice, Czech Republic) used for aqueous workup, column chromatography and crystallizations were used as received.Compound L1 [8]   13 C NMR), to external 85% aqueous H 3 PO 4 ( 31 P NMR) or to external neat CFCl 3 ( 19 F NMR). Mass spectra were recorded on an Esquire 3000 (Bruker) or an LTQ Orbitrap XL instrument (Thermo Fisher Scientific, Waltham, MA, USA; high resolution analyses).Infrared spectra were collected in Nujol mulls with an FTIR Thermo Nicolet 760 instrument (Thermo Fisher Scientific, Waltham, MA, USA) in the range 400-4000 cm −1 .

Preparation of 2 -(Dicyclohexylphosphino)-2-bromo[1,1 -biphenyl]-borane (1:1) (Compound 2)
A two-necked flask was charged with 2,2 -dibromo[1,1 -biphenyl] (1; 4.68 g; 15.0 mmol), purged with argon and sealed with a rubber septum.Anhydrous tetrahydrofuran (60 mL) was introduced and the reaction flask was placed in a dry ice/ethanol bath.A solution of n-butyllithium in hexanes (6.0 mL of 2.5 M, 15.0 mmol) was introduced dropwise with continuous cooling and stirring, yielding a pale yellow solution.After 30 min, chloro-dicyclohexylphosphine (3.6 mL, 16.5 mmol) was added in one portion, causing the reaction mixture to turn purple.The reaction mixture was kept at −78 • C for another 30 min, and then warmed to room temperature whereupon its color changed to colorless.After stirring overnight, neat borane-dimethyl sulfide complex (1.6 mL, 16.5 mmol) was added via a syringe.The reaction mixture was stirred for another 30 min, quenched with methanol (5 mL) and evaporated to dryness.The solid residue was taken up with ethyl acetate (100 mL) and the extract was washed with saturated aqueous NaHCO 3 (50 mL), brine (50 mL) and dried over magnesium sulfate.The solvents were removed under vacuum and the solid residue was purified by column chromatography (silica gel, ethyl acetate/hexanes, 1:30 v/v).The second band was collected and evaporated to afford compound 2 as a colorless, slowly crystallizing oil.Yield 4.86 g (73%).Under argon, bromide 2 (4.43 g, 10.0 mmol) was dissolved in anhydrous tetrahydrofuran (60 mL) in a three-necked reaction flask equipped with a stirring bar.The solution was cooled in a dry ice/ethanol bath and then treated by n-butyllithium (4.4 mL of 2.5 M solution in hexanes, 11.0 mmol).The resulting purple solution was stirred at −78 • C for another 30 min and then poured onto crushed dry ice, whereupon the mixture turned colorless.After warming to room temperature, the solution was acidified with 1.25 M methanolic HCl (15 mL) and concentrated under vacuum.The solid residue was dissolved in ethyl acetate (50 mL), washed with brine (3 × 50 mL) and dried over magnesium sulfate.The solvents were removed under reduced pressure, leaving a crude product, which was purified by flash chromatography (silica gel, ethyl acetate/hexanes, 1:3 v/v).The first band was collected and evaporated to give 3 as a white solid.Yield 3.00 g (74%).

Pd-Catalyzed Cyanation of Aryl Bromides. General Procedure for Preparative Experiments
A reaction vessel was charged (in this order) with palladium acetate (2.2 mg, 1 mol % with respect to the aryl bromide), ligand L1 (11.8 mg, 2 mol % with respect to the aryl bromide), potassium hexacyanoferrate(II) (184 mg, 0.5 mmol), potassium carbonate (138 mg, 1.0 mmol) and the respective aryl bromide (1.0 mmol).The flask was equipped with a magnetic stirring bar, flushed with argon and sealed with septum.The solvent (1,4-dioxane/water, 1:1; 4 mL) was added, the septum was changed for glass stopper and the flask was transferred into an oil bath kept at 100 • C.After stirring for 2 h, the reaction mixture was cooled and diluted with water (2 mL) and ethyl acetate (5 mL).The organic layer was separated and washed with brine (10 mL).The aqueous layer was extracted with ethyl acetate (3 × 5 mL).The organic extracts were combined, dried over anhydrous magnesium sulfate, and evaporated under reduced pressure to afford pure nitrile, which was analyzed by 1 H and 13 C NMR spectroscopy.Analytical data of the coupling products matched those reported in the literature [5,6].

Scheme 3 .
Scheme 3. Structures of Pd precursors with N,C-supporting ligands.

Scheme 3 .
Scheme 3. Structures of Pd precursors with N,C-supporting ligands.Scheme 3. Structures of Pd precursors with N,C-supporting ligands.

Scheme 4 .
Scheme 4. The model cyanation reaction used during the screening experiments.

Scheme 4 .
Scheme 4. The model cyanation reaction used during the screening experiments.

a
Conditions: aryl bromide (1.0 mmol), K4[Fe(CN)6]•3H2O (0.5 mmol) and K2CO3 (1.0 mmol) were reacted in dioxane-water (2 mL each) at 100 °C for 2 h in the presence of 1 mol % of pre-catalyst formed in situ from Pd(OAc)2 and L1 (2 equiv.).Isolated yields are an average of two independent runs.For details, see Experimental Section.b Reaction in the presence of 2 mmol of K2CO3.

a
Conditions: aryl bromide (1.0 mmol), K 4 [Fe(CN) 6 ]•3H 2 O (0.5 mmol) and K 2 CO 3 (1.0 mmol) were reacted in dioxane-water (2 mL each) at 100 • C for 2 h in the presence of 1 mol % of pre-catalyst formed in situ from Pd(OAc) 2 and L1 (2 equiv.).Isolated yields are an average of two independent runs.For details, see Experimental Section; b Reaction in the presence of 2 mmol of K 2 CO 3 .

Table 1 .
Results of the screening experiments with various Pd precursors and donors L1 and L2 a .

Table 1 .
Results of the screening experiments with various Pd precursors and donors L1 and L2 a .

Table 2 .
Summary of the reaction scope tests.