Eco-efficient electrocatalytic C–P bond formation

Electrosynthesis of tertiary phosphines catalyzed by nickel complexes

Phosphines are an important class of organophosphorus compounds. In recent years, research has been focused on catalytic synthesis of phosphines and it is still under full development [72], [73], [74]. The transition-metal-catalyzed cross-coupling reactions for C–P bond formation receive more and more attention among other suitable methods.

Nickel-catalyzed coupling between chlorophosphines and organic (aryl, hetaryl or perfluoroalkyl) halides

Electrochemical approach can be used to design a versatile technique for preparation of various types of tertiary phosphines on the basis of cross-coupling reactions of diphenylchloro- and phenyldichlorophosphines with aryl halides or C-halogen derivatives of sulfur- or nitrogen-containing heterocycles catalyzed by electrochemically generated Ni⁰bpy complexes [6k, 8] (from NiBr₂bipy as the catalyst precursor). The reactions were carried out in an undivided cell equipped with a magnesium or zinc anode in the absence of specially added supporting electrolyte. The use of soluble anodes in some cases makes it possible not only to simplify the electrochemical process significantly, but also to control the reaction direction. The method includes electroreduction of organic halide in the presence of mono- or dichlorophenyl phosphine added gradually and utilizes a Ni-bpy catalytic system [71], [75], [76], [77]. The electrolysis was performed at a constant current density and room temperature, until the concentration of the desired product in solution became constant. The unsymmetrical tertiary phosphines were obtained as a result of the cross-coupling reaction under mild conditions and with rather high yields. The yields varied from 45 to 70% (Fig. 1), they depend on the nature of substituents in the aromatic ring and anode material.

Fig. 1:

Cross-coupling of chlorophosphines with aryl or hetaryl halides.

One should note that magnesium anode should be used for the synthesis of tertiary phosphines with donor substituents, and zinc anode is recommended for the preparation of phosphines with acceptor substituents.

The method is efficient for both aromatic halides with acceptor and donor substituents in the ring and heteroaromatic species (pyridine, thiophene, pyrimidine and pyrazole halides). Its advantages include single-stage character and mild conditions (room temperature) of the process.

We suppose [76], [77] that the first stage of a catalytic cycle is an electrochemical reduction of Ni(II) complex to Ni(0) complex (Fig. 2). The latter is able to react with aryl bromide forming the organonickel compounds [Ni(Ar)Br(bpy)] (Ar=see Fig. 1), which reacts with chlorophosphine forming the finished product Ph₂PAr and regenerating the Ni(II) complex.

Fig. 2:

Key stages of cross-coupling.

We have concluded that the phosphines are generated through two different pathways. The Ni(0) species appear and undergo oxidative addition with aryl halides at the lower potential values (Figs. 3 and 5, cycle A). Their reaction with chlorophosphines results in the product ArPPh₂.

Fig. 3:

Electrochemically induced arylphosphine formation.

Fig. 4:

Electrosynthesis of arylphosphine at two different electrode potentials.

Fig. 5:

Catalytic cycle of arylphosphine synthesis.

Products are formed much faster at higher potential values (Figs. 4 and 5, cycle B), presumably through the following sequence of reactions:

The arylnickel(II) complexes [Ni(Ar)X(bpy)] are reduced and cleave X⁻ (as already discussed for the electrocatalytic C–C coupling). The aryl phosphide complexes [Ni(Ar)Cl(PPh₂)(bpy)] formed by oxidative addition of the chlorophosphine are reduced under the applied conditions and undergo reductive elimination resulting in the target compounds (Fig. 5, cycle B). As the catalyst consumption proceeds, the reactions of the catalytic cycle B gradually become dominant [76], [77].

Perfluoroalkyl substituted phosphines draw much attention because of their potential applications as ligands in organometallic chemistry [78], [79], [80], particularly in fluorous biphase catalysis [81], [82]. A recently published comprehensive review of fluoroalkyl containing phosphines has only a short column devoted to the direct synthesis of fluorinated phosphines from R_FI and elemental phosphorus, and the majority of the mentioned methods have obvious drawbacks [83]. In order to fill this information gap, we provide a versatile and convenient single-step electrochemical method of obtaining fluoroalkyl phosphines and phosphine oxides [84]. Aiming to expand the scope of the proposed method, we realized the cross-coupling of diphenylchlorophosphine and organofluorine halides with good yields (up to 73%) under the influence of electrochemically generated complexes of nickel in low oxidation state as it is demonstrated in Fig. 6 [84].

Fig. 6:

Perfluoroalkylation of diphenylchlorophosphine.

The structure of the products was unequivocally determined by 2D NMR methods and confirmed by the good correlation between calculated (GIAO B3LYP/6-31G(d)//B3LYP/6-31G(d) and experimental data on ¹⁹F spectra. Some compounds were previously obtained by Rossi et al. through reaction of Ph₂P⁻ ions and R_FI in HMPA, DMPU or liquid ammonia under photochemical irradiation [85]. However, the yields were low and authors did not provide ³¹P NMR data for the phosphines. The electrochemically prepared products have high purity.

The selective monoarylation of phosphorus trichloride is one of the key aspects in preparation of the new organophosphorus ligands for catalytic systems, since dichloroarylphosphines are useful precursors of various organophosphorus compounds. The traditional synthesis using Grignard or organolithium reagents usually does not give high yields and requires severe conditions and flammable media, complicating the industrial synthesis of compounds of this class and requiring serious safety measures [12], [13], [86].

The interaction of a σ-complex MesNiBrbpy prepared by the oxidative addition of MesBr to an electrochemically generated complex Ni⁰bpy with phosphorus trichloride is followed by monoarylation of the latter (Fig. 7) [87]. Gradual adding of a solution of MesNiBrbpy in Et₂O to a solution containing 10-fold excess of PCl₃ in the same solvent at −70°C enables a selective monoarylation of phosphorus trichloride with formation of dichloromesitylphosphine as the only phosphorus-containing product. The yield of the isolated MesPCl₂ was 68% [87].

Fig. 7:

Dichloromesitylphosphine preparation.

Metal-catalyzed functionalization of the white phosphorus

White phosphorus (P₄) is a key material for chlorine-free technologies of phosphorus derivatives preparation [68], [71], [88]. The problem of selective tetrahedron opening in white phosphorus and its direct functionalization gains increasing significance due to the search for new ecologically safe synthetic routes to obtain organophosphorus compounds. The electrocatalytic functionalization of white phosphorus was also accomplished using transition metal complexes as catalysts [68], [89], [90], [91], [92], [93]. The activation of P₄ takes place under mild conditions promoted by the electrogeneration of σ-organometal complexes in the presence of P₄. The reaction was documented for Ni(0) species electrogenerated from Ni(II) complexes and stabilized by 2,2′-bipyridine (bpy). The electrolysis was carried out using soluble anode (Al, Mg, Zn) in an undivided cell with a phosphorus emulsion in DMF or acetonitrile containing an organic halide and a Ni(II) complex, [Ni(bpy)₃](BF₄)₂, as catalyst. Under these experimental conditions, white phosphorus may be efficiently converted to phosphines and phosphine oxides [89], [90], [91], [92], [93].

Mechanistic studies were carried out for a specific case of electrocatalytic arylation of the white phosphorus. These investigations allowed to determine that the Ni(0) complex, initially obtained via reduction of Ni(II), oxidatively adds the organic halide to give [NiX(Ar)(bpy)] species that mediate the catalytic formation of P–C bonds according to Fig. 8:

Fig. 8:

Electrocatalytic arylation of the white phosphorus.

Nature of the soluble anode drastically influences the final product, although such behavior does not have a simple interpretation (Fig. 9). The use of zinc anode leads to a complete conversion of P₄ to soluble OPC (organophosphorus compounds), mainly, tertiary phosphines, while aluminum anode promotes phosphine oxide formation. Cyclic polyphosphorus compounds, such as (PhP)₅, are produced, when magnesium anode is used [89], [90], [91], [92], [93]:

Fig. 9:

Synthesis of OPC with different soluble anode.

These processes are intriguing, because they form the basis for the potential of metal electrocatalysis that allows alkylation and arylation of the white phosphorus under mild conditions. Remarkably, these processes combine the high efficiency in OPC formation with the total control of the product selectivity depending on the careful choice of the metal anode.

A new approach to triphenyl- [93], [94] (Fig. 10) or perfluoroalkylphosphine [84] (Fig. 11) preparation was developed using the white phosphorus activation under the influence of zinc compounds in undivided electrolyzer:

Fig. 10:

Undivided electrolysis of P₄ and organic halide using Zn anode.

Fig. 11:

Perfluoroalkylation of white phosphorus.

Cyclic regeneration of the Zn-catalyst takes place at the cathode, the reactive σ-complex forms in the bulk of the solution. The σ-complex attacks the P₄ molecule resulting in tertiary phosphines as the target product in good yields (80%) [93], [94].

A novel electrochemical synthetic pathway to fluoroalkyl phosphines and phosphine oxides directly from the white phosphorus and organofluorine halides under the influence of electrochemically generated complexes of nickel in low oxidation state or organofluorine nickel σ-complexes obtained on their basis was reported [84]. For this purpose, electrochemical reduction of solutions of fluoroalkyl iodides in DMF in the presence of tetraphosphorus emulsion and NiBr₂bpy complex as a catalyst was performed (Fig. 11).

The reaction proceeds in undivided electrochemical cell equipped with Pt-cathode and sacrificial Zn-anode without any supporting electrolyte. The catalyst is zero-valent nickel or zinc generated. The metal sigma-complex generated in solution volume acts as an initiator in the electrocatalytic cycle with soluble anodes. The high selectivity and rate of electrosynthesis allow carrying out this process with high level of conversion of the white phosphorus to target organophosphorous products.

Metal-catalyzed ligand-directed phosphorylation of C(sp2)–H bonds

Transition metals catalyzed functionalization of widely abundant, but usually inert C–H bonds into carbon-carbon (C–C) and carbon-heteroatom (including C–P) bonds is very promising. Unlike traditional methods used for formation of these bonds, direct C–H-substitution excludes the necessity of preliminary functionalization (for example, halogenation or borylation) of the initial substrate, which decreases the number of steps and diminishes the amounts of undesirable products in the multistep processes. Atom-economy and shorter synthetic routes with less number of steps can thus be attained [95], [96]. For this reason, the functionalization of С–Н bonds catalyzed by transition metals attracts considerable attention [97], [98], [99], [100]. One of the main problems of new synthetic methods utilizing palladium (or other transition metals) catalysts in high oxidation states is the need stoichiometric amounts of co-oxidants. For example, transition metal salts or organic oxidants, such as silver salts, organic peroxides, and electrophilic fluorine-containing reagents, which are often expensive or poorly separable from reaction mixtures, can be used as co-oxidants [97], [98], [99], [100]. In addition, broad screening is needed to choose an optimum oxidant for a new ligand or substrate or when changing the reaction conditions.

A cheaper, “green”, and more controlled electrochemical alternative to the C–H substitution reactions under mild electrosynthesis conditions at controlled anodic potential is proposed. Various known synthetic approaches for oxidative C–H functionalization always include one or more oxidation steps. We assumed that the oxidation of Pd^II complexes can be efficiently conducted at an electrode through replacing of a chemical oxidant by electrochemical oxidation. This allowed us to obtain the assumed bimetallic Pd^III and/or Pd^IV complex via electrosynthesis and to use it as a model intermediate to study the routes of catalytic C–H-substitution. One of our tasks was to perform the whole process of C–H functionalization electrochemically, under generation of the oxidized form of the palladium catalyst at the electrode, which has been never accomplished earlier [69], [70], [101], [102], [103].

Although various reactions of direct functionalization of C–H bonds including those catalyzed by transition metals are already known, examples of carbon-phosphorus bond formation are rather restricted, most likely, because of the strong coordinating properties of the phosphorus reagents [24], [25], [36], [39], [104], [105], [106], [107], [108], [109]. There are rather few electrochemical methods for hydrogen atom substitution by phosphorus-containing groups and the last successful examples were described for the functionalization of nitroaromatic substrates only [110], [111], [112]. The formation of the carbon-phosphorus bond under the influence of catalytic transition metals is considered to be an important technique for the synthesis of various types of phosphorus compounds, such as phosphonates, phosphinates, phosphine oxides, phosphines, and others. However, the first catalytic method to introduce phosphorus groups into various structures has been developed only recently [66], [113], [114] although the importance of such processes is especially significant because of the green chemistry requirements.

In spite of success of some researchers (see, e.g. Refs 66, 113, 114) in the ligand-directed C–H-phosphorylation of aromatic substrates, many complicated factors, namely, high temperatures of the processes, prolonged reaction duration, the necessity to use an excess of expensive silver acetate AgOAc as an oxidant, an additional base and reagents for the facilitation of the reductive elimination step, and not always satisfactory yields, indicate that the search for new, more efficient solutions is extremely important.

We used the accumulated experience of electrochemical C–H-acetoxylation and perfluoroalkylation [69], [70], [101], [102] to develop a method of ligand-directed C–H-phosphorylation by electrochemical oxidation under mild conditions involving Pd(OAc)₂ as a catalyst and diethyl phosphite (DEP) as an accessible phosphorus precursor [115], [116], [117]. The use of the electrochemical method allowed to conduct the reaction under mild conditions without addition of an oxidant (AgOAc or other) with a controlled potential at room temperature and to analyze particular the reaction steps.

The combined electrochemical oxidation of 2-phenylpyridine, DEP (ratio 1:1.1, dropwise addition of DEP during the whole electrolysis), and palladium acetate Pd(OAc)₂ (10 mol.%) gives the product of ortho-phosphorylation of 2-phenylpyridine with 35–68% yield depending on the process conditions (Fig. 12).

Fig. 12:

Pyridine-directed phosphonation of C(sp²)-H bond.

The best result (68%) was obtained for electrolysis in the presence of sodium acetate as a base and benzoquinone (reagent that facilitates the reductive elimination step).

Intermediate six-membered palladium complex [(PhPy)Pd(EtO)₂P(O)]₂ was synthesized in order to determine the mechanism of the process. For this purpose, the reaction of acetate dimer of phenylpyridylpalladium (1) with DEP (ratio 1:2) in acetonitrile (Fig. 13) was performed.

Fig. 13:

Substitution of acetate bridges to phosphonate bridges.

The reaction occurs readily at 20°C, and the complex [(PhPy)Pd(EtO)₂P(O)]₂ precipitates from the reaction mixture in a day with 90% yield [38], [39]. Earlier, to synthesize dibutylpalladium phosphonate complex [(PhPy)Pd(BuO)₂P(O)]₂ from α-hydroxybutyl phosphonate and Pd₂(OAc)₂(PhPy)₂ (1), the reaction mixture was heated in dioxane at 120°C in the presence of 2.2 eq. K₂HPO₄[114]. Evidently, the ligand exchange occurs much more easily when DEP is used.

Then, complex [(PhPy)Pd(EtO)₂P(O)]₂ was oxidized at a Pt electrode in MeCN at 20°C under oxygen-free conditions at the potential of its first oxidation peak with the complete conversion to the single product of ortho-phosphorylation of 2-phenylpyridine (Fig. 14) with an almost quantitative yield [115], [116].

Fig. 14:

Electrochemical oxidation of dipalladacycle [(PhPy)Pd(EtO)₂P(O)]₂.

Electrochemical characteristics of the complex [(PhPy)Pd(EtO)₂P(O)]₂ were studied by cyclic voltammetry. The CV curve of this complex exhibits two irreversible oxidation peaks, their potentials lie in a more positive region (E_p¹=1.18 V) compared to the potentials of the complex [(PhPy)Pd(OAc)]₂ (1) (E_p¹=0.62 V).

It has been shown [102] that complex 1 is oxidized in two steps with transfer of two electrons in each step. The first oxidation peak is reversible and characterized by the metal-centered electron transfer and transformation of each Pd^II into Pd^III. Two electrons are also transferred during the second step, but the oxidation of Pd^III to Pd^IV in MeCN is irreversible [116]. The formation of intermediate Pd^III was detected for the oxidation of the complex 1, whereas no paramagnetic intermediates were observed in case of the complex [(PhPy)Pd(EtO)₂P(O)]₂. However, it is most likely that the oxidation of this complex proceeds through the same two steps (also two peaks, but at more positive potentials) [115], [116].

Thus, the substitution of the hydrogen atom in the ortho-position of 2-phenylpyridine by the phosphonate group occurs selectively under electrooxidative conditions and requires no additional reagents that promote reductive elimination step, which is always used in all studies published earlier [113], [114] (N-methylmaleimide (NMMI), benzoquinone, PPh₃, and others). We reproduced the reductive elimination step by heating the complex [(PhPy)Pd(EtO)₂P(O)]₂ in MeCN (at 120°C) in the presence of NMMI (2 eq.) for 4 h. However, only half of the initial complex decomposed giving the product of 2-phenylpyridine ortho-phosphorylation. Thus, the electrochemical oxidation favors the step of reductive elimination of intermediate [(PhPy)Pd(EtO)₂P(O)]₂.

According to the oxidation potentials of the key intermediates [(PhPy)Pd(μ-OAc)]₂ (1) and [(PhPy)Pd(EtO)₂P(O)]₂, the oxidation of phosphorus complex [(PhPy)Pd(EtO)₂P(O)]₂ is more difficult than that of the acetate complex 1. The comparison of the chemical (oxidized by AgOAc) and electrochemical oxidations shows that the acetate complex is oxidized first. As a result, it leads to an undesirable side reaction of ortho-acetoxylation of 2-phenylpyridine [102]. To obtain the target product with a good yield, it is necessary to provide conditions for the predominant oxidation of the phosphorus complex, i.e. to perform electrolysis at the oxidation potential of complex [(PhPy)Pd(EtO)₂P(O)]₂. The presumable mechanism of electrochemical C–Hphosphorylation of 2-phenylpyridine is presented in Fig. 15.

Fig. 15:

Catalytic cycle of 2-phenylpyridine C-H phosphorylation.

The electrolysis at the oxidation potential of phosphorus complex [(PhPy)Pd(EtO)₂P(O)]₂ favors catalytic process and increases the target product yield. Thus, the new approach to introduction of the phosphonate group into arylpyridines was developed using 2-phenylpyridine as an example. The approach is based on the electrochemical oxidation of a mixture of 2phenylpyridine and diethyl phosphite at room temperature in the presence of palladium acetate (Fig. 15) [115], [116].

The intermediate six-membered binuclear palladium cycle bound to the phosphonate ligand through oxygen and phosphorus atoms [(PhPy)Pd(EtO)₂P(O)]₂ was isolated. The preparative oxidation of the complex selectively gives the target product of C–H bond phosphorylation with a quantitative yield [115], [116].

In order to empower the ligand -directed phosphonation of C–H aromatic bonds, a series of diphosphonate-bridged dipalladacycles [(phpy)Pd(EtO)₂P(O)]₂, [(bhq)Pd(EtO)₂P(O)]₂, [(phpz)Pd(EtO)₂P(O)]₂ (phpy=2-phenylpyridine, bhq=benzo[h]quinoline, phpz=1phenylpyrazole) was prepared and characterized by NMR spectroscopy and cyclic voltammetry in acetonitrile solutions and in carbon paste electrode [117]. Comparative quantitative studies of the redox properties of structurally related diphosphonate-bridged dipalladacycles under the same conditions, their stability in high oxidation states, the influence of aromatic ligand nature on the potentials of Pd(III)/Pd(II) and Pd(IV)/Pd(III) transfers and electrochemical synthesis of arylphosphonates from these complexes were performed [117].

Diphosphonate-bridged dipalladacycles 4–6 with 2-phenylpiridide (phpy), benzo[h]quinoline (bhq), 1-phenylpyrazole (phpz) ligands were synthesized through treatment of corresponding acetate complexes [LPd(OAc)]₂1–3 with diethyl H-phosphonate (EtO)₂P(O)H (Fig. 16):

Fig. 16:

Diphosphonate-bridged dipalladacycles synthesis.

One should note that the geometry of the bridged dinuclear Pd complexes depends on the nature of the bridging ligands and directly affects their redox reactivity (vide intra). For example, while diphosphonate-bridged palladacycles are essentially planar [114], [116] and exhibit no metal-metal interactions (Pd-Pd distance is 4.345 Å for 4), the acetato-bridged complexes adopt an open “clamshell” geometry with two Pd centers in close proximity (Pd-Pd distance change from 2.862A˚ for complex [(phpy)Pd(OAc)]₂ to 2.842 Å for bhq and phpz analogs) and exhibit weak d⁸–d⁸ interactions between two Pd centers and π-stacking interactions between two phenylpyridine or related ligands [118], [119]. The oxidation potentials (E_p¹) of the acetate complexes 1–3 in CH₃CN solutions increase in the row phpy≤bhq<phpz. The order does not change in case of CPE: phpy<bhq<phpz, but reversibility of the first peak E_p¹ almost disappears. For the phosphonate complexes 4–6 in CH₃CN, E_p¹ increases in the row bhq<phpy (phpz complex is insoluble) while CPE exhibits opposite trend: phpz<phpy<bhq.

The dipalladium phosphonate complexes 4–6 exhibit two-electron irreversible oxidations both in CH₃CN solutions and in CPE. The irreversibility can be assigned to a fast chemical reaction after the electron transfer. That is likely a reductive elimination reaction producing arene (phpy, bhq, phpz) phosphonation products. The proposed oxidation scheme (Fig. 17) is following:

Fig. 17:

Oxidation of diphosphonate-bridged dipalladacycle.

Complexes of palladium in low oxidation state formed during reductive elimination step should be oxidized at these potentials to Pd(II) or even Pd(III). These assumptions were confirmed by preparative electrolysis.

Electrochemical oxidation of diphosphonate-bridged palladacycles 4 and 5 was investigated in a divided electrolyzer in CH₃CN solution and monitored by ³¹P NMR spectroscopy (Fig. 18). Initial [(phpy)Pd(P(O)(OEt)₂)]₂4 has δ=97.2 ppm in the ³¹P spectrum. After one electron is passed per each Pd atom, the signal completely disappears, whereas two new signals of 16.2 and 69.5 ppm are observed in the reaction mixture. The signal of δ=16.2 relates to arylphosphonate while δ=69.5 probably refers to a mononuclear palladium complex. Further passing of electricity leads to a gradual consumption of the palladacycle (δ=69.5 decreased) and formation of the arylphosphonate product (δ=16.2 increased).

Fig. 18:

³¹P spectra evolution during preparative oxidation of 4.

Monomerization of [(phpy)Pd(OEt)₂P(O)]₂4 into (phpy)Pd(CH₃CN)[P(O)(OEt)₂] (the most probable structure) induced by an electron transfer reaction is unusual. The assumption of the monomeric structure of the intermediate with δ=69.5 ppm was proved by the proximity of this value to the chemical shifts of other mononuclear palladium phosphonate complexes, such as Stockland’ arylpalladium complexes bu₂bpyPd(Ar)(P(O)(OR)₂) (δP=70.8–78.8 ppm) [120], [121]. Eight electrons per molecule of 4 are required to convert dimeric complex into the product (Fig. 18). The number of required electrons is greater than described in Fig. 17, where the maximum consumed electron number equals to six per dimer molecule. However, this scheme disregards monomerization of 4 and participation of Pd(IV) complexes in the process.

Monomeric palladium complexes are known to be oxidized at more positive potentials than dimeric ones with the same ligand environment [102]. We assume that the same trend is valid for the phosphonate complexes. The oxidation potential shifts to more positive region from 1.1 to 1.6 V during the preparative electrolysis (the working electrode potential is constantly monitored in galvanostatic mode) and attains the value of Pd³⁺/Pd⁴⁺ transfer. Scheme of the dipalladium complex oxidation, where the electron number is relevant to the electrolysis data, can be described as follows (Fig. 19).

Fig. 19:

Preparative oxidation of dipalladacycle 4.

Similar results of preparative oxidation were obtained for [(bhq)Pd(OEt)₂P(O)]₂5 (δ=95.2 ppm), which quantitatively converts into benzo[h]quinoline phosphonate (δ=23.2 ppm), when 8F electricity per mol is passed. The complex [(phpz)Pd(EtO)₂P(O)]₂6 is insoluble in pure CH₃CN. We have found that it is possible to achieve solubility of this complex by adding a small amount of amine such as pyridine or triethylamine. The removal of amine in vacuo returns complex 6 to the original insoluble white solid state with the same melting point corresponding to 6. This method allowed us to analyze 6 using spectral methods, to confirm its structure, and to carry out the electrochemical oxidation in preparative conditions. Thus, electrochemical oxidation of 6 (δ=81.5 ppm) quantitively yields diethyl-2-(1H-pyrazol-1-yl) phenylphosphonate (δ=23.4 ppm).

Thus, phosphonate palladacycles are oxidized irreversibly at more positive potentials, as distinguished from the acetate palladacycles, and quantitatively afford corresponding arylphosphonates in preparative oxidations carried out under mild conditions without any added oxidants. Therefore, this oxidation way may be applied for the synthesis of new arylphosphonates from different arenes in ligand-directed aromatic C–H phosphonation reactions.

Metal-induced oxidative phosphorylation of arenes

Synthesis of arylphosphonates through direct phosphorylation of C–H bonds in aromatic substrates remains one of the most important approaches, since it generally meets the accepted criteria of “green chemistry”. Thus, a search for the catalytic conditions is especially important, since there are only few known catalytic phosphorylation reactions of aromatic compounds. Perhaps, benzene is one of the most problematic objects in this regard, since its structure lacks factors facilitating C–H substitution, i.e. a functional group, which activates bonds or promotes direct functionalization. As a rule, if conditions for phosphorylation of benzene are selected properly, its derivatives give even higher yields of the target products [104], [122], [123].

It was suggested that phosphorylation of benzene at room temperature in one step was feasible using advantages of electrochemical metal complex catalysis [67], [124]. The catalytic systems for electrochemical phosphorylation of benzene were represented by metal complexes and salts (M) in the oxidation state II, which can undergo electrochemical oxidation to M^III: PdCl₂bipy (bipy=2,2′-bipyridine), manganese citrate Mn₃(C₆H₅O₇)₂, Mn₃(C₆H₅O₇)₂/CoCl₂dmphen (dmphen=1,10-dimethylphenanthroline), MnCl₂, MnCl₂/CoCl₂dmphen, MnCl₂/CoCl₂bipy, MnSO₄, MnSO₄/Ni(BF₄)₂bipy, MnCl₂bipy, Ni(BF4)2dmphen, and CoCl₂dmphen [125]. The process was carried out with the equimolar ratio benzene: dialkyl H-phosphonate (1:1) at room temperature. Such conditions have never before given good yields of arylphosphonate (Fig. 20). The catalytic systems were tested using commercially available, relatively stable phosphorylating agent diethyl H-phosphonate.

Fig. 20:

Benzene phosphorylation.

The most efficient catalytic systems are MnSO₄/CoCl₂dmphen and MnCl₂/CoCl₂bipy (yield to 90%) [125]. The phosphorylation virtually does not occur in the absence of catalysts. It is important to point out that 100% conversion of diethyl phosphite in these syntheses is reached after passing 2F of electricity. In contrast to the results published earlier [104] [according to which there is no phosphorylation at all in the absence of Mn(OAc)₂], the phosphonate can be also obtained in the presence of a mono-component catalyst (Mn^IIor Co^II) under electrochemical conditions, though in this case the yield of the target product is considerably lower. To find out the factors influencing efficiency of electrocatalytic phosphorylation, a CVA method was used [125]. Generalized scheme for the catalytic cycle was suggested (Fig. 21):

Fig. 21:

Catalytic cycle of phosphorylation using Co-Mn system.

Catalytic scheme involving a mono-component catalyst simplifies (Fig. 22).

Fig. 22:

Catalytic cycle of phosphorylation using M²⁺L precursor.

The suggested method for the electrosynthesis of diethylphenylphosphonate proceeds under mild conditions (room temperature, normal pressure) with 1:1 ratio of the reagents and gives the product with a high yield (up to 90%) and 100% conversion [125] in the presence of bimetallic catalytic systems Mn^II/Co^IIL (MnSO₄/CoCl₂dmphen or MnCl₂/CoCl₂bipy). Mechanism of the process requires further studies, but we suppose that the first step of the cycle is probably formation of the metal (or metals) phosphonate undergoing oxidation at low anodic potentials with elimination of the phosphonate radical, which reacts with benzene.

This new approach was upgraded and extended to the phosphorylation of benzenes bearing both electron withdrawing and electron donating substituents in the ring and some coumarins (coumarin, 6-methylcoumarin, 7-methylcoumarin) under the action of dialkyl H-phosphonate (RO)₂P(O)H (R=Et, n-Bu, i-Pr). These processes are based on catalytic oxidation of arene and H-phosphonate (1:1) mixture in electrochemically mild conditions (room temperature, normal pressure) using a new bimetallic catalyst system Mn^IIbpy/Ni^IIbpy (1%). This method produces dialkyl aryl phosphates with high yields (up to 70%) and 100% conversion of Hphosphonate (Fig. 23) [126]:

Fig. 23:

Electrochemically induced phosphorylation of arenes.

It was assumed the catalytic cycle involves the following steps (Fig. 24):

Fig. 24:

Catalytic cycle of phosphorylation using Ni-Mn system.

To sum it up, a single-stage method of direct electrochemical phosphorylation of aromatic C–H bonds resulting in dialkyl aryl phosphonates (aryl=benzene, coumarins) is provided. The process takes place under mild conditions (room temperature, normal pressure) with 1:1 reagent ratio, thus providing products with high yields (70%) and 100% conversion, when a bimetallic catalyst system Mn^IIbpy/Ni^IIbpy is used. The success was achieved due to fast catalytic cycle and rapid regeneration of metal catalysts on the electrode [125], [126].

Electrosynthesis of primary phosphines from alkenes and white phosphorus

Primary phosphines are valuable starting materials for chemical reactions. Because of their importance as precursors or active intermediates in numerous reactions, there is a growing interest to a new preparation method from available raw materials such as elemental phosphorus. The syntheses of primary phosphines usually involve expensive, multi-step, and long-term procedures, such as: reduction of phosphonous dihalides with LiAlH₄, preparation from metal phosphides and alkyl or aryl halides, hydrolysis of alkyl and arylphosphonous dihalides, reaction under Friedel-Crafts conditions, pyrolysis of biphosphines and triphosphines, etc. [12], [127]. Selective method of primary phosphine preparation directly from white phosphorus without traditional chlorination stages is highly advantageous.

Reactions of olefins with phosphorylating agents are usually not selective without a catalyst. It is known [128] that styrene reacts with PH₃ at 70°C and the pressure of 28–30 atmospheres under the action of radical initiators. This reaction results in formation of a mixture of primary, secondary and tertiary phosphines with yields of 6–36%. Ways of the primary phosphines preparation from styrene or α-methylstyrene and PH₃ in superbasic medium (KOH/DMSO) with yield of 20–30% are also well known [129].

However, the attempts to elaborate an approach to phosphine derivatives with one or two P–H bonds from P₄, e.g. primary or secondary phosphines R₂PH, RPH₂, phosphorus acids H₃PO₃, H₃PO₂ and other important precursors in phosphorus chemistry were not successful for the long time. The main problem of all known reactions concerns either low yield of a product due to the formation of nonreactive polyphosphides and consequently low phosphorus conversion, or the use of expensive reagents, such as rhodium complexes [130], [131], [132]. Our attention was attracted by rather old publications on phosphine electrolytic production, which was carried out in the 1960s. Thus, it was demonstrated that the cathodic reduction of white phosphorus in aqueous solutions on metals with high hydrogen overvoltage results in the formation of PH₃ with yield up to 95%. These results were patented in Germany, USA and Great Britain [133], [134], [135], [136], [137], [138], [139]. Some technological refinement for this process was recently suggested by Japanese researchers, who patented a reaction cycle for generating PH₃ while stirring [140]. In 20^th century several attempts were made to use electrochemically generated PH₃ for subsequent synthesis on the base of its reactions in situ, however, the obtained results were not encouraging, as they resulted in complex mixtures of products (e.g. with styrene) [141], [142], [143] and/or low yields.

It is worth noting that the topic of white phosphorus conversion into phosphine is well supported by patents. The C–P bond formation requires conditions that would allow instant (as formed) conversion of PH₃ and other phosphine intermediates into organic phosphines. This will allow avoiding the accumulation of intermediate toxic and dangerously explosive phosphine, converting it into undetectable “conventional intermediate” (Fig. 25):

Fig. 25:

Electrolysis of P₄ in aqueous solution.

Joint electrolysis of white phosphorus emulsion and alkene in the aqueous acetic buffer solution results in the formation of only primary phosphine under these conditions (Fig. 26) [93]:

Fig. 26:

Primary phosphine formation from P₄ in aqueous acetic buffer solution.

The mechanism of phosphine with P–C bond formation is not quite clear. We can assume following: phosphorus centered radicals and radical-anions are generated from white phosphorus under conditions of electrochemical reduction [144]. Radicals and radical-anions can join with weakly electrophilic alkenes (e.g. styrene) through nucleophilic mechanism and with nucleophilic alkenes (alkyl ethene, phenyl ethers) – through radical mechanism.

Electrochemical reduction of P₄ molecule in protogenic conditions is known to take place at the cathode with high hydrogen overvoltage, for example, at lead cathode [145], [146], [147]. The electrochemical rupture of the P–P bonds resulting in the formation of phosphine is provided by the presence of active proton donors through the protonation of intermediates (Fig. 27):

Fig. 27:

Electrochemical reduction of P₄ molecule.

But on the other hand, phosphine does not react with alkene without a catalyst. Phosphine can be added to С=С bonds by both ionic and radical mechanisms in the presence of initiators. Alkenes of different structure react with phosphine under rather rigid conditions (60–90°C, 30–47 atm, acid catalyst; or in the superbasic medium such as КОН/DMF) [148]. Apparently, under electrolysis conditions the investigated reaction is initiated by reduction of phosphine to ⁻PH₂ at the electrode or proceeds through the intermediate phosphides or radical and radical-anions formed at white phosphorus reduction. It is generally impossible to exclude proceeding of several competitive reactions of phosphide-anion formation, both its protonation in solution and reduction of P–H bond at the electrode, and target addition of phosphide-anion to alkene.

The formation of secondary or tertiary phosphines was not observed even at significant alkene excess in the initial mixture. According to NMR ³¹P spectrum, the only by-products discovered in the reaction mixture in aqueous part of electrolyte were inorganic acids of phosphorus (Fig. 28), hypophosphorous acid. The proposed method is characterized by the following advantages: mild conditions (room temperature) of the process and the one-step process of primary phosphine preparation directly from white phosphorus.

Fig. 28:

Products of electrolysis of white phosphorus in water – phosphines and their derivatives (organic and inorganic).

It should be noted that the use of some alkenes does not promote formation of the desired phosphines [93].

Results obtained in the course of electrosynthesis are explained on the basis of voltammetric data of the substrates. Reduction potentials of alkene successfully reacting with P₄ in joint electrolysis should be more negative, than the potential of white phosphorus reduction (Е=−2.2 V). Therefore, the desired products are obtained through P₄ reduction.

Inorganic hypophosphorus acid appeared to be a by-product. The choice of alkenes is defined by their low electrochemical activity, so they are not reduced in the accessible potential region at the glassy-carbon electrode, while they are reduced at less negative potentials at the lead electrode.

Ferrocene phosphonation

Given the importance of ferrocene derivatives with phosphorous containing substituents, the search of a new, more convenient, single-stage and selective approaches to their synthesis remains a relevant objective of the modern science. We proposed a new approach to ferrocene phosphorylation through a “masked” phosphorylating agent, namely α-hydroxylalkylphosphonate, based on electrochemical reduction of a mixture of ferrocene and (Me)₂C(OH)P(O)(OC₂H₅)₂ at −50°C [149] (Fig. 29). Such method enables obtaining of the product, diethyl ferrocenyl phosphonate, with high yields (up to 90%) and 100% conversion of the initial phosphonate within single stage. It is evidenced with experiments that the process of ferrocene reduction is carried out with preservation of the iron charge in the ferrocene fragment and with the formation of a cyclopentadienyl ligand radical anion at −3.3 V ref. Ag/AgCl (at −50°C).

Fig. 29:

Ferrocene phosphonation.

The synthesis of phosphorylated ferrocene in one stage using phosphorous acid as the phosphorylating agent with the yield of 36% was realized. These reactions are not catalytic, but this is a new achievement for the electrochemical formation of compounds with the C–P bond, which is difficult to approach by other methods.