Open Access

Phosphines are highly versatile ligands for transition metal catalysts because of wide tuning abilites of their stereoelectronic properties. Bulky and basic phosphines, to a smaller extend also π-acidic phosphites were intensively studied whereas dicoordinated trivalent phosphorus compounds were comparatively little investigated in this respect. In part this may go back to the limited stability of many P=C compounds, in the case of the stable benzazaphosphole to low stabilityof complexes with non-zero-valent transition metals. With the availability of suitable chelate complexes this problems may be overcome. Because biaryl phosphines proved particularly useful as chelate ligands this work is focused on the development of convenient syntheses of new biaryl-type N-heterocyclic or functionally aryl substituted 1,3-benzazaphosphole P,N- P,P- and P,O-chelate ligands and the characterization of their structures. The pivotal point was to find an applicable synthetic route to the title ligands. Because currently transition metal catalyzed cross-coupling reactions are a hot field in catalytic research, the initial target of my work was the investigation of the applicability of suitable biaryl coupling reactions on 1,3-benzazaphospholes. There are several types of transition metal catalyzed biaryl couplings. One reaction, which is currently in the main focus by use of non-toxic and air stable coupling partners, often allowing water as environmental friendly solvent, is the Pd-catalyzed Suzuki-Miyaura coupling of an aryl halide with an arylboronic acid. To apply the Suzuki coupling to the synthesis of biaryl-type benzazaphospholes, the synthesis of either benzazaphosphol-2-boronic acids or reactive 2-halogen-benzazaphospholes have to be performed. Because of the successful introduction of functional groups in position 2 of benzazaphospholes via lithiation and reaction with electrophiles, the 2-lithiation of suitably available N-substituted benzazaphospholes and introduction of boryl groups or halogen by reaction with boronic acid esters or with a halogenating reagent like dibromoethane appeared as a realistic route and was chosen for closer study. N-Neopentyl-benzazaphosphole was selected by its relatively easy access and N-mesityl-benzazaphosphole as a N-aryl representative. From the two principal methods developed to synthesize 1,3-benzazaphospholes, only the synthesis and reduction of o-aniline phosphonic acid esters to o-phosphinoanilines and subsequent [4+1] cyclocondensation is promising to access N-substituted 2-CH benzazaphospholes. My first investigations targeted to improve the synthesis of the benzazaphosphole precursors. The invention of a Cu- instead of the earlier used Pd-catalyzed P-C coupling allows a more economical access to anilinophosphonates which were then transformed to 2H-1,3-benzazaphospholes by the established orthoformamide cyclocondensation. Several attempts of the coupling with careful control of dryness of all reagents and solvents were made in order to obtain pure 1,3-benzazaphosphole-2-boronic acid ester and, after mild hydrolysis, to isolate 1,3-benzazaphosphole-2-boronic acid. The coupling worked with N-mesityl-1,3-benzazaphosphole 13e, but the benzazaphosphol-2-boronic acid could not be obtained in pure form because of easy B-C bond cleavage during crystallization, certainly by the two ‘OH groups. For attempts with a reverted methodology, the synthesis of a 2-bromo-substituted benzazaphosphole was studied, which should be coupled with (hetero)arylboronic acids via Suzuki-Mijaura reaction. However, the 2-bromo-benzazaphosphole also could not be obtained in pure form, and a coupling experiment with phenyl boronic acid and catalysis with ligand free Pd/C failed. Therefore, other routes to biaryl-type benzazaphospholes were envisaged. Direct C-H functionalization has emerged over the past few years as an attractive strategy to enhance molecular complexity. This holds also for π-excess-type heterocycles like indoles, benzoxazoles or purines which allow direct CH-arylation in 2-position. These reactions generally involve palladium based catalysts and in some cases rhodium catalysts. In a series of experiments the catalytic arylation, heteroarylation and later also alkylation were studied with 1,3-benzazaphospholes 13a-e as precursors. The initial studies were carried out with iodobenzene, keeping similar reaction conditions as for 2-CH arylation of indoles. Then transition metal catalysts, bases and conditions were varied. The necessity and influence of a catalyst was established by blind experiments without transition metal catalyst which led to strong decrease of the reactivity. However, the transitional metal catalyzed reactions of N-substituted-1,3-benzazaphosphole with aryl- and heteroaryl halides did not give the desired 2-aryl-substituted 1,3-benzazaphosphole biaryl ligands but revealed a novel oxidative addition at the P=C double bond. In the presence of moisture benzazaphospholine-P-oxides are formed. Further exploration of the scope of this reaction showed that it is applicable to several functionally substituted aryl halides and heteroaryl halides. As besides PdX2 (X = Cl, OAc) also Pd(0)(PPh3)4 was found active as catalyst, it can be assumed, that the reaction occurs via a Pd(0) species and oxidative addition of the aryl halide at Pd(0). Because Pd(0) will coordinate stronger to the π-acidic benzazaphosphole than Pd(II) it is assumed that in the first step small equilibrium amounts of a Pd(0)benzazaphosphole complex will be formed which undergo the oxidative addition and then react to benzazaphospholium salt and furnish back a Pd(0) complex with 1,3-benzazaphosphole ligand. The benzazaphospholium salts are highly sensitive to moisture and react with traces of water to form benzazaphospholine-P-oxides 20 and acid, neutralized by the base. A cyclic species RR’P(OH)=CHR”, where the halogen is replaced by OH, may be assumed as intermediate which undergoes a rearrangement to the more stable RR’P(=O)-CH2R” tautomer, driven by the high P=O bond energy. After various investigations of the optimum conditions for the reaction, a number of new functionally substituted P-aryl or P-heteroaryl benzazaphospholine P-oxides and 1,3-dineopentyl-benzazaphospholine-3-oxide were isolated and characterized by 1H, 31P, 13C and HRMS data and two by crystallography. The biaryl-type 2-phenyl-1,3-benzazaphosphole is known since the earliest reports of these heterocycles, synthesized by cyclocondensation of 2-phosphinoaniline with benziminoester hydrochloride or in low yield with benzaldehyde. The latter method was further developed because of the compatibility of the aldehyde group with various donor functions. 2-Phosphinoaniline (12a) and 2-phosphino-4-methylaniline (12b) were heated with pyridine-2-carboxaldehyde under varied conditions, and a crucial role of acid catalyst was observed in the investigation. The results showed that the dehydrogenating cyclocondensation, if catalyzed by a suitable type and amount of acid catalyst, works well for primary phosphinoanilines 12a,b and a variety of reactive aldehydes, including N-heterocyclic and o- or m-functionally substituted arylaldehydes. In an equimolar ratio, on heating usually hydrogen is eliminated, at least formally, to furnish the aromatically stabilized 1H-1,3-benzazaphosphole ring systems of 35 whereas in other cases reductive side reactions occur, e.g. the N-CH2R substitution to 36 in reactions with two equivalents of aldehyde. Thus the synthesis of 1,5-dimethyl-1,3-benzazaphosphole (36a) was achieved by double cyclocondensation of 12b and formaldehyde in a 1:2 molar ratio. This provides the so far shortest way to synthesize N-substituted 1,3-benzazaphospholes and suggests, that the reaction is generally applicable in reactions with two equivalents of monoaldehyde. This puts the question if N-secondary o-phosphinoanilines such as N-neopentyl-2-phosphinoaniline (12d) can be cyclocondensed with aldehydes to benzazaphospholes or if a primary amino group is required. The successful experiment shows that cyclocondensation of N-secondary o-phosphinoanilines with suitable aldehydes is possible. N-Neopentyl-2-pyrido-1,3-benzazaphosphole was obtained in high yield. An interesting extension of the above reaction are cyclocondensations with compounds bearing two aldehyde groups. Double condensation of 12b with o-phthaldialdehyde was performed. It proceeded fast and gave tetracyclic-1,3-benzazaphosphole in high yield. Based on the NMR monitored primary formation of organoammonium phosphino glycolates from amines, phosphines and glyoxylic acid, followed by conversion to phosphinoglycines, it is assumed that the reaction proceeds by initial attack of the primary phosphino group of 12b at the carbonyl carbon atom of R-CHO, polarized with the help of the acid catalyst. The resulting P-C bonded secondary phosphine, containing an α-hydroxy group, may release water after transfer of a proton to oxygen in equilibrium, followed by attack of amine. This leads to formation of the dihydro-intermediate 34, observed by NMR reaction monitoring in several cases. Possible ways are releasing of H2 during reflux, directly giving 2-substituted NH-1,3-benzazaphospholes 35, or hydrogen transfer, connected e.g. with N-substitution leading to 1,2-disubstituted 1,3-benzazaphospholes 36. The second path is observed mainly when excess or double molar quantities of aldehydes are used at the start of the reaction. The two hydrogen atoms at P and C2 are consumed during the second condensation and formation of the NCH2R group and generate the P=C double bond. Finally, cyclocondensation of o-phosphinoanilines with aldehydes has proven as a useful method for the synthesis of biaryl type benzazaphosphole ligands. After thorough investigations, N-primary and secondary phosphino anilines were found cyclisable with various heteroaryl aldehydes upon refluxing in toluene in the presence of a suitable acid catalyst, and 11 new compounds were synthesized following this procedure and characterized by 1H, 31P, 13C NMR and HRMS data. For two compounds crystal structures were also obtained. First attempts to synthesize chelate complexes with the 2-(hetero)aryl-1,3-benzazaphospholes were started. A soluble 2-(o-diphenylphosphinophenyl)-1,3-benzazaphoasphole-Cr(CO)4 chelate complex was detected by NMR spectroscopy, whereas most products of the new ligands with Rh(COD) or NiCp complexes were insoluble in usual NMR solvents and require further efforts for synthesis and full analytical and structural characterization.