The Suzuki reaction, a palladium-catalyzed reaction of arylboronic acids with aryl halides, is a fairly universal method for constructing a new carbon–carbon bond in a number of aromatic compounds, which is widely used in fine organic synthesis to obtain multifunctional biaryl compounds and their heterocyclic analogs [1]. Numerous examples of the using of Suzuki cross-coupling for the synthesis of new drug compounds, liquid crystal composites, conductive polymers, multifunctional materials, and organic intermediates are presented in reviews [2, 3]. In addition to the traditional triphenylphosphine ligand [4], electron-donating sterically hindered phosphines [5], phosphine oxides [6], and N-palladacycles [7] have been proposed as effective ligands or catalyst precursors.

Research in the field of catalysis of cross-coupling reactions is aimed at the development of more active palladium catalysts, a detailed study of the catalytic transformations mechanisms, modification of the conditions for their implementation, determination of the catalytically active particles structure, and expansion of the reagents range [812].

The creation of active reusable heterogeneous catalysts [13] can significantly reduce the cost of expensive palladium and reduce the amount of residual metal in cross-coupling products, which is especially important in the synthesis of pharmaceuticals. One of the promising solutions to this problem is the development of bi- and polymetallic Pd–M catalysts, since due to the synergistic effect caused by the transfer of electron density from an electropositive metal (iron, cobalt, nickel, etc.) to a less electropositive palladium, efficient catalysts can be obtained with a low content of expensive palladium [14]. A detailed analysis of the problems and achievements in the field of heterogeneous catalysis by polymetallic nanoparticles of transition metals is presented in reviews [15, 16].

In continuation of studies [1723] on the development of new catalytic systems for cross-coupling reactions, bimetallic composites Pd–Ni(Co,Cu,Fe) were coated on aluminum oxide modified with L-proline in order to obtain reusable catalysts.

The main idea behind the creation of new catalytic materials was the non-covalent modification of aluminum oxide with a suitable amino acid, which would be highly soluble in water and slightly soluble in alcohol solvents. Amino groups present in amino acids form complexes with transition metals; as a result, the metal-activator and palladium are evenly distributed over the support surface, creating optimal conditions for the formation of highly dispersed catalytic composites. As a support, a conventional aluminum oxide for chromatography with a grain size of 0.063–0.2 mm and a pore diameter of 9 nm [Merck, Al2O3 90, active basic (activity stage I), 135–162 m2/g] was chosen. By treatment aluminum oxide with a saturated aqueous L-proline (Pro) solution, a proline-modified carrier Pro/Al2O3 (~1.5 mmol proline/g) was obtained. Then, mixtures of NiCl2 (CoCl2, CuCl2, or FeCl3) crystalline hydrates and Na2PdCl4 from solutions in anhydrous methanol, in which proline is slightly soluble, were applied onto the modified aluminum oxide. Subsequent reduction of MCl2(3)‒PdCl2‒Pro/Al2O3 (M = Ni,Co,Cu,Fe) hybrid materials with an excess of sodium borohydride yielded polymetallic composites: Pd‒Ni‒Pro/Al2O3 (1), Pd‒Co‒Pro/Al2O3 (2), Pd–Fe–Pro/Al2O3 (3), and Pd–Cu–Pro/Al2O3 (4). Treatment of composite 3 containing corrosion-unstable iron with sodium tetrachloropalladate resulted in a composite Pd‒Fe‒Pro/Al2O3@Pd (5) with a protective palladium coating. The scheme for the synthesis of polymetallic composites applied on aluminum oxide with a protective palladium coating is presented using the example of obtaining composites 3 and 5 (Scheme 1).

Scheme
scheme 1

1.

According to the data of atomic absorption analysis, composites 14 contain ~0.1 mmol Pd/g and ~0.8 mmol metal-activators/g; composite (5) contains ~0.3 mmol Pd/g and ~0.6 mmol Fe/g. To compare the catalytic activity, a Pd‒Pro/Al2O3 composite 6 with the same amount of palladium (0.1 mmol/g), as in bimetallic composites was prepared in a similar way from Na2PdCl4 and proline. According to SEM microscopy and EDS data, composites 16 are characterized by a uniform distribution of palladium, iron, cobalt, nickel, and proline. Figure 1 shows micrographs of the Pd‒Fe‒Pro/Al2O3@Pd composite 5 with element distribution maps. It can be seen from the presented data that in additionally palladium sample 5, palladium is uniformly distributed over the surface of the polymetallic composite.

Fig. 1.
figure 1

SEM micrograph (scale 100 μm) of the Pd–Fe–Pro/Al2O3@Pd composite 5 (a) with maps of the elements distribution: aluminum (b), oxygen (c), nitrogen (d), iron (e), and palladium (f).

The catalytic activity test and verification of the possibility of regeneration of the obtained polymetallic composites were carried out on a model reaction of 4-methoxyphenylboronic acid with 3-bromobenzoic acid in aqueous methanol (1 : 1) and in water (Scheme 2). All experiments were performed in the presence of Pd–Ni(Co,Cu,Fe)–Pro/Al2O3 composites (0.1 mol % Pd) at the boiling point of solvents in air.

Scheme
scheme 2

2.

As can be seen from the results presented in Table 1, the activity of all catalysts in aqueous methanol and in water is very high, and the nature of the second activating metal (Co, Cu, Ni, Fe) has little effect on the yield of the cross-coupling product, 4′-methoxybiphenyl-3-carboxylic acid (exp. 1–10). It should be noted that in the presence of the Pd‒Fe‒Pro/Al2O3 catalyst 3, the reaction mixture gradually acquired a light orange color due to the formation of iron(III) hydroxide due to the corrosion instability of iron in an aqueous-basic medium. Analysis of the reaction mixtures by atomic absorption spectroscopy after the reactions completion and separation of the catalyst by centrifugation did not reveal the presence of palladium in the solution at the sensitivity level of the method (~1 ppm). The monometallic Pd–Pro/Al2O3 catalyst 6 obtained in the absence of a more electropositive metal exhibits noticeably lower activity: in water at 100°C for 20 min, the yield of the cross-coupling product was 61% (compare exp. 10 and 11). The activity of bimetallic palladium catalysts applied on a mesoporous oxide carrier noticeably exceeds the activity of previously obtained magnetic polymetallic Pd‒Fe‒Co‒Ni composites [24]. Catalysts are easily isolated from reaction mixtures by simple filtration or centrifugation (3000 rpm, 5 min).

Table 1. Catalysis by polymetallic Pd‒Ni(Co,Cu,Fe)‒Pro/Al2O3 composites of the reaction of 3-bromobenzoic acid with 4-methoxyphenylboronic acida

Catalysts 2, 4, and 5 showed high catalytic efficiency [TON (catalyst turnover number) up to 103, TOF (catalyst turnover frequency) up to 4 103 h–1] in 5 recycles without activity loss in aqueous methanol and in water (Table 2). These results can be explained by rather strong binding of palladium to the surface of the metal (Ni,Co,Cu,Fe)-oxide (Al2O3) support in the bimetallic composites obtained in the presence of proline; only a small part of the applied palladium, which, due to the synergistic effect, has a high reactivity in the oxidative addition reaction, takes part in catalysis due to a reversible transition into solution. As a result, the catalyst retains its composition and activity. SEM micrographs of composite 5 after 5 recycles are shown in Fig. 2. From a comparison of the data presented in Figs. 1 and 2, it can be concluded that the morphology of the catalyst and the dispersity of applied palladium are retained.

Table 2. Catalysts 2, 4, 5 activity after recycling (5 cycles)
Fig. 2.
figure 2

SEM micrograph (scale 100 µm) of the Pd‒Fe‒Pro/Al2O3@Pd composite 5 regenerated after 5 recycles (a) with maps of the elements distribution: aluminum (b), oxygen (c), nitrogen (d), iron (e), and palladium (f).

The very high activity of applied bimetallic palladium catalysts, practically quantitative yields, and the use of water or aqueous methanol as the reaction medium make it possible to completely exclude chromatographic methods for purifying the cross-coupling product, simplify the isolation procedure to the maximum, and make the cross-coupling process environmentally safe. To isolate analytically pure samples, after the reaction completion, the reaction mixture was diluted with water, filtered, then 10–15 vol % ethanol was added, heated almost to boiling, and slowly acidified with 5% HCl with stirring. After cooling, a finely crystalline, well filterable precipitate of biphenylcarboxylic acid was obtained. According to atomic absorption spectroscopy, the amount of residual palladium in the isolated samples of biphenylcarboxylic acid is less than 1 ppm.

The synthetic capabilities of the developed catalysts are demonstrated by the example of the synthesis of the drug substance diflunisal (Scheme 3). 4-Hydroxy(2′,4′-difluoro)biphenyl-3-carboxylic acid (diflunisal) is a non-steroidal anti-inflammatory drug with analgesic and antipyretic effects was obtained from 2,4-difluorophenylboronic acid and 5-iodosalicylic acid by catalysis with composite 5 (0.01 mol % Pd) in water at 100°C for 5 min with quantitative yield. Diflunisal is approximately 20 times more effective in a hypersensitivity test and 9 times more effective in the treatment of adjuvant arthritis than aspirin [25].

Scheme
scheme 3

3.

The known method for the preparation of this compound [26] is based on carrying out the reaction in water in a strictly inert atmosphere catalyzed by a complex of palladium (2 mol %) with the phosphine ligand t-Bu-Amphos—[2-(di-tert-butylphosphino)ethyl]trimethylammonium chloride ( 20°C, 8 h, 95% yield). The obtained bimetallic catalysts have such high activity and stability that, in some cases, cheaper aryl chlorides can be used instead of aryl bromides and iodides (Scheme 4).

Scheme
scheme 4

4.

The reaction of 2-chloropyridin-3-amine with furan-2-ylboronic acid catalyzed by composite 5 (0.1 mol % Pd) in a 20% aqueous solution of an ionic liquid (Bu4NOAc) proceeds in 30 min and leads to 2-(furan-2-yl)pyridine-3-amine in practically quantitative yield. The need to carry out the reaction in the presence of an ionic liquid is due to the low solubility of 2-chloropyridine-3-amine in water, in contrast to halobenzoic acids (Table 1). In anhydrous dioxane catalyzed by 3 mol % Pd[Р(t-Bu)3]2, to complete this reaction, the reaction mixture must be refluxed for 18 h with a twofold excess of furan-2-ylboronic acid; the yield of 2-(furan-2-yl)pyridin-3-amine was 88% [27].

Thus, for the first time, promising and very active in the catalysis of the Suzuki reaction, bimetallic (Pd–metal) catalysts supported on aluminum oxide modified with proline have been developed. In the presence of the obtained catalysts, the reactions proceed in practically quantitative yields, which simplify the isolation of cross-coupling reaction products to the maximum. The developed catalytic materials Pd–Ni(Co,Cu,Fe)–Pro/Al2O3 are easily regenerated from the reaction mixture and can be reused up to 5 times without activity loss.

EXPERIMENTAL

1H and 13C NMR spectra (400 and 100 MHz, respectively) were recorded on a Bruker Avance II 400 spectrometer in DMSO-d6 or CDCl3. Chemical shifts are determined relative to residual solvent signals. Mass spectra were recorded on an Agilent 6890N device equipped with an Agilent HP-5ms capillary column (30 m × 0.25 mm × 0.25 µm) and an Agilent 5975C inert MSD detector, electron impact ionization with an electron energy of 70 eV (evaporator temperature, 250°С). Elemental analysis was performed on a vario Micro cube elemental CHNS analyzer. The palladium amount in polymetallic composites and in cross-coupling products was determined by atomic absorption spectroscopy on an Akvilon MGA-915 spectrometer. SEM micrographs were taken on a Zeiss LEO EVO 50 XVP scanning electron microscope equipped with an Oxford Instruments EDX INCA Energy 350 analyzer. The progress of the reactions was monitored by TLC on Merck Silica gel 60 F254 plates. Melting points were determined on a Kofler apparatus.

Reagents and solvents (Aldrich, Acros Organics, and Merck) were used without additional purification.

Pd‒M(Ni,Co,Fe,Cu)‒Pro/Al2O3 composites. To 10 g of Al2O3 preliminarily dried at 150°C for 2 h, a solution of 15 mmol of L-proline in 10 ml of water was added. The resulting suspension was stirred for ~1 h at 80°С at atmospheric pressure on a rotary evaporator until complete evaporation of water and dried for 1 h at 90°С. Solutions of 0.8 mmol of NiCl2 (CoCl2, FeCl3, or CuCl2) crystalline hydrates and 0.1 mmol of Na2PdCl4 in 5 mL of anhydrous methanol were added to samples of proline-modified aluminum oxide (1 g). The resulting suspensions were stirred for ~1 h at 60°С at atmospheric pressure on a rotary evaporator until the methanol was completely evaporated and dried for 1 h at 60°С. The samples of green (nickel), pink (cobalt), orange (iron) and blue (copper) were obtained. To each sample in an argon atmosphere, 2 mL of water, 0.2 mL of methanol were added, and with stirring, a solution of 4 mmol of NaBH4 in 2 mL of water. All samples, when the reducing agent solution was added, almost instantly turned dark gray. After hydrogen evolution was completed (~30 min), the samples were washed with water, acetone, and diethyl ether and dried for 1 h at 80°C. Bimetallic Pd‒Ni‒Pro/Al2O3 (1), Pd‒Co‒Pro/Al2O3 (2), Pd‒Fe‒Pro/Al2O3 (3), and Pd‒Cu‒Pro/Al2O3 (4) composites were obtained.

To increase the stability, 0.1 g of sample 3 in 2 mL of water was treated with stirring with 0.2 mL of a 0.1 M Na2PdCl4 aqueous solution (0.02 mmol), stirring was continued until the Na2PdCl4 solution became completely colorless. The solution was decanted, the residue was washed successively with water, acetone, diethyl ether, and dried for 1 h at 80°С. A sample of the Pd–Fe–Pro/Al2O3@Pd composite (5) with an additional palladium coating was obtained. Similarly, a Pd‒Pro/Al2O3 composite 6 was obtained from Na2PdCl4 and proline with the same amount of palladium as in bimetallic composites.

Suzuki reaction catalyzed by PdM(Ni,Co,Fe,Cu)Pro/Al2O3 composites (general procedure). To a mixture of 1.20 mmol of arylboronic acid, 1.00 mmol of aryl bromide, and 0.35 g (2.50 mmol) of K2CO3 in 5 mL of H2O (or 50% aqueous methanol) was added a catalyst (10 mg 14, 6, 3.3 mg 5, 0.1 mol % Pd). The reactor was placed in a silicon bath preheated to 120°C (for reactions in aqueous methanol) or to 160°C (for reactions in water), the reaction mixture was intensively stirred for 15–20 min at reflux (yields of the cross-coupling product are given in Table 1). The reactions progress was monitored by TLC (eluent hexane–Et2O, 3 : 1) using calibration solutions of the corresponding biaryl and aryl bromide (at a molar ratio of 1 : 1 and 9 : 1). After the reaction completion, the catalyst was separated by centrifugation (3000 rpm) and, after washing with water and ethanol, was reused. The centrifugate was diluted with water, filtered, 10– 15 vol % ethanol was added, heated to ~50°C, and slowly acidified with 5% HCl to pH 2–3 with stirring. As a result, well filterable precipitates were formed, and analytically pure samples of the cross-coupling product were obtained without the use of chromatographic methods. Yields in exp. 1, 4, 10 (Table 1) were also determined by 1H NMR using tetrachloroethane (0.5 mmol) as an internal standard.

4′-Methoxybiphenyl-3-carboxylic acid. Yield 0.224 g (98%) (exp. 10), white crystalline powder, mp 203.6–204°C (mp 202–203°C [28]). 1H NMR spectrum (400 MHz, DMSO-d6), δ, ppm: 3.83 s (3H, MeO), 7.05 d. d (2H, H3′,5′, J 6.8, 2.1 Hz), 7.56 t (1HAr, J 7.7 Hz), 7.66 d. d (2H, H2′,6′, J 6.8, 2.1 Hz), 7.83–7.94 m (2HAr), 8.11 d. d (1H, H4, J 7.8, 2.0 Hz), 13.12 br. s (1H, COOH). 13С NMR spectrum (100 MHz, DMSO-d6), δС, ppm: 55.2 (MeO), 114.5 (C3′,5′), 126.8 (C5), 127.5 (C2), 127.9 (C2′,6′), 129.2 (C6), 130.6 (C4), 131.45 (C3), 131.57 (C1′), 140.2 (C1), 159.2 (C4′), 167.3 (COOH). Found, %: C 73.59; H 5.39. C14H12O3. Calculated, %: C 73.67; H 5.30.

4-Hydroxy-(2′,4′-difluoro)biphenyl-3-carboxylic acid. Yield 0.245 g (98%), white crystalline powder, mp 212–213°C (mp 210–211°C [25]). 1H NMR spectrum (400 MHz, DMSO-d6), δ, ppm: 7.06 d (1H, H5, J 8.8 Hz), 7.16 d. d (1H, H5′, J 8.1, 8.1, 2.2 Hz), 7.33 d. d. d (1H, H3′, J 9.9, 9.9, 2.2 Hz,), 7.56 d. d (1H, H6′, J 15.4, 8.8 Hz), 7.66 d (1H, H6, J 8.1 Hz), 7.91 s (1H, H2). 13C NMR spectrum (100 MHz, DMSO-d6), δС, ppm: 105.1 d. d (C3′, JCF 27.1, 25.8 Hz,), 112.1 d. d (C5′, JCF 20.8, 2.8 Hz); 113.2 (C3); 117.6 (C5); 123.7 d. d (C1′, JCF 12.5, 4.2 Hz); 125.1 (C1), 130.3 d (C2, JCF 2.8 Hz), 131.5 d. d (C6′, JCF 9.7, JCF 4.2 Hz), 135.8 d (C6, JCF 2.8 Hz), 158.9 d. d (C2′, JCF 226.1, 12.5 Hz), 160.7 (C4), 161.8 d. d (C4′, JCF 224.7, 12.5 Hz), 171.6 (COOH). Found, %: C 62.36; H 3.27. C13H8F2O3. Calculated, %: C 62.41; H 3.22. F 15.19; O 19.18.

2-(Furan-2-yl)pyridine-3-amine [14]. Yield 0.152 g (95%), light yellow oil. 1H NMR spectrum (400 MHz, CDCl3), δ, ppm: 4.52 br. s (2H, NH2), 6.56 d. d (1H, H4 furan, J 3.4, 1.8 Hz), 6.97 d (1H, Н3 furan, J 3.3 Hz), 7.02 d. d (1H, H4, J 9.4, J 1.9 Hz), 7.17 d. d (1H, H5, J 9.4, 4.2 Hz), 7.40 d (1H, Н5 furan, J 1.8 Hz), 7.91 d. d (1H, H6, J 4.2, 1.9 Hz). 13C NMR spectrum (100 MHz, CDCl3), δС, ppm: 107.8 (C3 furan), 110.5 (C4 furan), 122.3 (C4), 123.6 (C5), 138.5 (C2), 139.7 (C3), 141.9 (C5 furan), 142.7 (C6), 151.2 (C2 furan). Mass spectrum, m/z (Irel, %): 160 (100) [M]+, 131 (62), 104 (17). Found, %: C 67.40; H 5.12; N 17.41. C9H8N2O. Calculated, %: C 67.49; H 5.03; N 17.49.