Abhishek V. Dubey and
A. Vijay Kumar*
Department of Chemistry, Institute of Chemical Technology, Matunga, Mumbai, Maharashtra, India-400019. E-mail: vijayakki@gmail.com
First published on 4th May 2016
Mussel-inspired magnetically recoverable and reusable Pd nanoparticles immobilised on polydopamine-coated nano-Fe3O4 were prepared and characterized by techniques such as TEM, SEM, XPS, XRD, FT-IR and DSC-TGA. The catalyst demonstrated remarkable activity for Suzuki–Miyaura cross-coupling reactions of aryl halides and aryldiazonium salts with arylboronic acids in a benign solvent system. A wide range of substrates were converted, including the challenging relatively poorly reactive aryl bromides. The yields of the products were good to excellent. Moreover, the protocol has several advantages such as tolerance of sensitive functional groups, magnetic recovery and reusability of the catalyst up to five cycles. Mechanistic studies showed that the palladium nanoparticles followed a dissolution and redeposition pathway during the catalytic cycle. The developed methodology was successfully utilized for the synthesis of a precursor of the sartan drugs.
In this context, nanomaterials based on iron oxides have recently burgeoned as recoverable catalysts owing to their abundance, nontoxic nature and magnetic recoverability. Many versions of metal nanoparticles, complexes, organocatalysts, etc. immobilized on iron oxides (coated with silica, dopamine, carbons and polymers) for various cross-coupling reactions and other organic transformations have been developed.6 Usually, the procedures used for preparing these capped iron oxides demand the use of expensive, toxic precursors under inert conditions. In addition, these procedures are carried out in harsh reaction conditions with toxic solvents at elevated temperatures, which limit their utility. In addition, the need for specialised conditions/equipment for their recovery make them less suitable for scale-up. Besides these shortcomings, the leaching of supported metal nanoparticles and iron has been observed for uncapped bare ferrite catalysts.7 Therefore, a robust catalyst that can be easily prepared from readily available precursors and capable of operating efficiently in non-toxic solvents is highly desirable.
In view of the above limitations, our attention was drawn to biomimicry, particularly the marine mussels, which have a special ability to latch onto surfaces. They do this by secreting 3,4-dihydroxyphenylalanine (DOPA)-containing adhesive proteins, which become hard byssal threads when released at saline pH. These threads function as anchors and thus prevent the mussels from getting washed away with the tide.8
Polydopamine is a polymer whose behavior is similar to that of the adhesive proteins of mussels. It in particular displays unique properties such as excellent biocompatibility and affinity for surfaces, and can be easily prepared by self-polymerization of dopamine (the well-known neurotransmitter) at ambient conditions. For these reasons, polydopamine has emerged as a promising biomimetic material with a wide range of applications in energy devices, water treatment, sensing, nano drug delivery, catalysis, etc.9 There has been a long debate about its structure, and is presumed to consist of many functional groups such as indolylamines, quinones, catechols and imines similar to the natural melanins.10 These moieties are able to chelate metal ions and also have an inclination to bind metal oxides, and hence serve dual purposes of surface adhesion and stabilization of metal nanoparticles. The catechol and quinone domains tend to enhance organocatalysis, photochemical oxidation of water, and photodegradation.11
Over the past few years, biomimetic palladium catalysis has come into the limelight on account of its versatility and significant catalytic activity under milder conditions than those used for the above-described processes.12 Therefore considering the above observations, and to continue developing the methodology for using magnetically recoverable catalysts for organic synthesis,13 we present herein mussel-inspired polydopamine (PDA)-coated nanoferrites14 supported by palladium nanoparticles (Pd/Fe3O4@PDA) for use in the Suzuki reaction under mild conditions to address the issues of recoverability, benign support and reusability (Fig. 1).
To confer success, the synthesised material was characterized by techniques such as TEM, SEM-EDX, XPS, XRD, DSC-TGA, FT-IR and ICP. The EDX analysis confirmed the presence of Pd, along with Fe, C and O. The TEM images showed a uniform PDA coating loaded with evenly distributed palladium nanoparticles with dimensions of 10–70 nm. The amount of metal loaded was revealed by ICP-AES analysis to be 3.88 wt%. X-ray photoelectron spectroscopy (XPS) is the most effective technique to identify the type of Pd species on a heterogeneous support.15 The XPS analysis of the native catalysts showed Pd 3d core level lines, and the peaks at 333.13 eV and 335.2 eV were assignable to the Pd(0) species at different environments on the support. The peaks at 337.8 eV and 342.7 eV can be attributed to the Pd(II) species. The intensities of these peaks in the spectra clearly indicated that the major species present was Pd(0) with the rest being Pd(II), which can be PdO, in agreement with literature values16 (Fig. 3a). The spectra also showed the peaks corresponding to Fe–O, CO, C–C and C–H bonds17 (Fig. 3b–f).† Furthermore, we have analysed the catalyst by thermogravimetric analysis (TGA) to determine its thermal stability; analysis showed it to be stable upto 675 °C preceded by mass loss at 100 °C attributable to absorbed water (see ESI).† All these characterization techniques corroborated the immobilization of palladium on the support. Thus, having succeeded with the immobilization of palladium, we tested the catalyst for the cross-coupling reaction of 4-methoxyiodobenzene and phenylboronic acid.
Different conditions, made by varying parameters such as solvents, bases and temperature, were tested to optimize product yield. The best yields resulted from a water–ethanol (1:
1) solvent mixture with (0.46 mol% Pd) of catalyst, K2CO3 as base at 80 °C. Under these conditions, the substrates underwent complete conversion in 6 h to afford the cross-coupling product 4-methoxy-1,1-biphenyl (1a) in 98% yield (entry 7, Table 1). The 1H and 13C-NMR spectroscopic data confirmed the formation of the cross-coupling product (refer to ESI†).
Entry | Solvent | Base | Time (h) | Temp. (°C) | Yieldb (%) |
---|---|---|---|---|---|
a Reaction conditions: phenylboronic acid (0.55 mmol), 4-methoxyiodobenzene (0.5 mmol), base (2 equiv.), Pd/Fe3O4@PDA (0.46 mol% Pd).b Isolated yield.c No improvement of yields even after 24 h.d Pd/Fe3O4@PDA (0.23 mol% Pd).e K2CO3 (1 equiv.). | |||||
1 | H2O | K2CO3 | 6 | 80 °C | 57c |
2 | EtOH | K2CO3 | 6 | 80 °C | 58c |
3 | EtOH![]() ![]() ![]() ![]() |
Cs2CO3 | 6 | 80 °C | 98 |
4 | EtOH![]() ![]() ![]() ![]() |
Na2CO3 | 6 | 80 °C | 75 |
5 | EtOH![]() ![]() ![]() ![]() |
K2CO3 | 6 | 80 °C | 68d |
6 | EtOH![]() ![]() ![]() ![]() |
K2CO3 | 6 | 80 °C | 78e |
7 | EtOH![]() ![]() ![]() ![]() |
K2CO3 | 6 | 80 °C | 98 |
8 | EtOH![]() ![]() ![]() ![]() |
K2CO3 | 5 | 80 °C | 88 |
9 | EtOH![]() ![]() ![]() ![]() |
K2CO3 | 24 | 70 °C | 80 |
10 | EtOH![]() ![]() ![]() ![]() |
K2CO3 | 24 | 60 °C | 71 |
With the optimised conditions in hand, we then proceeded to screen a variety of substrates for their conversions. The reaction of phenylboronic acid with substituted iodobenzenes was carried out. Iodobenzenes bearing electron-donating and electron-withdrawing groups such as methoxy, methyl and cyano groups afforded the products (1b–m) in the range of 98–77% yield in six hours. The polyaromatic aryl products viz. 1-naphthyl (1k), 4-biphenyl (1l) and heteroaryls viz. 2-pyridyl (1i), 2-thiophene (1j) were obtained in good to excellent yields (79–97%) (1i–l, Table 2). The product (1f) of the reaction of 4-methoxy iodobenzene with 4-trifluoromethyl phenylboronic acid was also obtained in 89% yield. The 2,6-disubstituted halide substrate showed moderate conversion due to steric hindrance.
We further extended the scope of the methodology by testing various aryl bromides for their reactivity with arylboronic acids. Different aryl bromides bearing electron-donating and electron-withdrawing groups such as 4-methoxy, 3-methyl, 4-methyl, 4-nitro, 4-cyano, 2-pyridyl, 3-pyridyl 2-acetyl and 4-formyl groups were reacted with phenylboronic acid, 2-methoxyphenylboronic acid and 4-trifluomethoxyphenylboronic acid. The reactants underwent smooth conversions to afford the products (1n–1aa) in excellent to good yields ranging from 95–75% with an extended reaction times of 8–12 hours when compared with the iodobenzenes. The reactions of 2-methoxy phenylboronic acid with bromobenzene and 4-trifluoromethoxy phenylboronic acid with 4-methyl bromobenzene have also successfully afforded the products (1w, 1y) in 95% and 89% yields (Table 3). The compound 2,6-methylbromobenzene showed only 72% conversion due to steric hindrance. Chlorobenzene, 2-chloropyridine and 2-chlorothiophene did not react with phenylboronic acids even at relatively high temperatures.
Generally, in the Suzuki cross-coupling reaction, the aryl bromides are less reactive and require higher temperatures to achieve good conversions, when compared with the iodobenzenes. The reaction proceeds with difficulty when electron-donating groups are present on the aryl bromides. In contrast, with our catalytic system, we were able to obtain good to excellent yields even in the case of challenging substrates such as bromobenzenes, especially those with electron-donating groups. Additionally, the products/reactants bearing cyano and carbonyl groups, which are prone to undergo hydrolysis and oxidation, did not show any formation of hydrolysed (or) oxidized products (entries 1t and 1r, Table 3). All of these observations, i.e., the preservation of the sensitive functional groups and the enhanced reactivity of substrates, showed the mild conditions and efficiency of the developed protocol.
To explore further the versatility of the catalyst, we tested another cross-coupling reaction, one that has been less exploited with supported palladium catalysts. We chose the Suzuki cross-coupling reaction of aryldiazonium salts with arylboronic acids.18 Initially, we tested a reaction of 4-nitrobenzenediazonium tetrafluoroborate with phenylboronic acid by varying different conditions such as solvent and temperature (ESI† for the optimisation condition). The reaction with methanol as solvent at room temperature after three hours gave the maximum 85% yield of the biaryl product (1ab). The other solvents gave inferior yields when compared with methanol. Thus, we continued further with methanol as solvent at room temperature and tested various aryldiazonium tetrafluoroborate salts for their reactions with different arylboronic acids. We were able to obtain good to moderate yields of products in most cases. The reaction of 4-methoxybenzenediazonium tetrafluoroborate with 4-nitrophenylboronic acid gave a very low yield since these compounds did not have the functional groups favouring cross-coupling, i.e., an electron-withdrawing group on the arylboronic acid and an electron-donating group on the aryldiazonium salt (Table 4).
After completing the study of the substrates, we proceeded to test the recyclability of the catalyst. This test was done by recovering the particles after completion of the reaction and reusing them for the next cycle. The recovery of the catalyst was facile due to the magnetic properties of the ferrite nanoparticles. The particles were easily recovered with the help of an external laboratory magnet precluding any specialized techniques such as centrifugation and filtration. We were able to recycle the catalyst for up to five cycles with no loss of activity and with only a slight deviation of product yields (Scheme 2).
We next investigated the leaching of palladium from the support, since it is a major concern in heterogeneous catalysis. This issue was probed by separating the catalyst from the reaction mixture after the fifth cycle and subjecting the solution to ICP analysis. The Pd leaching was found to be negligible (6.1 ppm). In addition, surface analyses of the recycled catalyst were carried out by using TEM and XPS to check for the intactness of the particles. The TEM analysis showed very little nanoparticle aggregation, in particular when compared with the native catalyst and no change in polydopamine morphology (Fig. 2). The catalyst after being reused for five cycles showed XPS peaks at 338 eV and 343.4 eV, which we assigned to Pd(IV) species. The peaks at 334.9 eV and 340.4 eV were assigned to the Pd(0).15,16 We presumed that Pd(IV), a higher oxidation-state species of Pd, was formed under the aqueous conditions, as reported earlier in the case of supported palladium catalysts.16 It was quite difficult to accurately assign the Pd 3d peak at 346–347 eV. The peak could not be attributed to a metal plasmon or palladium satellite S2, as it was too intense. We cannot offer any explanation at this moment and intend to further investigate this issue in near future. In the XPS spectra, peaks accounting to potassium might have been due to remnants of the base used in the reaction, and the increase in the percentage of oxygen was presumably due to the presence of water molecules.
To gain more insights into how palladium was operating in the reaction, we performed a hot filtration experiment. Under the optimised reaction conditions, the cross-coupling between 4-iodoanisole and phenylboronic acid was carried out for two hours (∼60% conversion according to GCMS), and then the catalyst was removed by centrifugation (3000 rpm), and the obtained filtrate was placed in a separate reaction flask and the reaction continued at 80 °C for 24 hours. This reaction showed a 62% conversion according to GCMS. Another hot filtration test was carried out at the same condition (reaction conversion after two hours ∼61%) but with the catalyst filtered without centrifugation to prevent the back deposition of leached Pd species. This reaction showed ∼69% conversion according to GCMS (see ESI for experimental details†). It can be concluded that the leached species in the non-centrifuged experiment catalysed the reaction. From these observations, we speculated that the palladium nanoparticles dissolved during the oxidative addition step and redeposited on the surface after the reductive elimination step, similar to the mechanisms described in the previous reports.19–21
Having completed these tests, we then turned our focus to the synthesis of a biphenyl-based drug molecule (or) precursor. Our attention was drawn towards the synthesis of the important lifesaving sartan series of drugs, namely losartan, valsartan, irbesartan, olmesartan and candesartan, which are known hypertension angiotensin AT1 antagonists.22 These molecules are synthesized industrially from a common biphenyl precursor, 4′-methyl-[1,1′-biphenyl]-2-carbonitrile, by a Suzuki protocol. We therefore tried the cross-coupling of 4-methylphenylboronic acid and 2-bromobenzonitrile with our Pd/Fe3O4@PDA nanocatalyst. The reaction was completed in 8 h, and upon work up and purification gave the product 4′-methyl-[1,1′-biphenyl]-2-carbonitrile (1ak) in an 82% yield (Scheme 3). Thus, herein, we successfully demonstrated the synthesis of the common starting material for the industrial syntheses of seven sartan drugs.
![]() | ||
Scheme 3 Synthesis of the biphenyl precursor for the sartan series of drugs by using Pd/Fe3O4@PDA nanoparticles. |
Footnote |
† Electronic supplementary information (ESI) available: TEM, SEM, XPS, EDX, FT-IR and DSC-TGA analysis of the catalyst. Details of experimental procedures, 1H-NMR and 13C-NMR for all compounds are available. See DOI: 10.1039/c6ra03395d |
This journal is © The Royal Society of Chemistry 2016 |