A dual catalytic strategy for carbon–phosphorus cross-coupling via gold and photoredox catalysis

1During the past decade, homogeneous gold reactions based on Au(i) or Au(iii) catalysis have emerged as an extraordinary tool to create molecular complexity. In these reactions, gold most-commonly acts as a redox-neutral and carbophilic π-acid that activates carbon–carbon multiple bonds towards nucleophilic attack. Alternatively, gold-catalyzed transformations employing a stoichiometric external oxidant, such as Selectfluor, have allowed entry into pathways involving Au(i)/Au(iii); however, the majority of these reactions still involve intermediates generated from activation of a carbon–carbon π-bond. The requirement for stoichiometric amounts of strong oxidizing reagents has generally limited the chemistry to π-bonds and aromatic compounds. Recently, stepwise oxidation of gold(i) complexes by photoredox-generated radical species has emerged as an alternative strategy for accessing Au(i)/Au(iii) coupling reactions.

2Organophosphorus compounds have drawn increasing attention due to their broad applications in biological, pharmaceutical, and material sciences. These compounds are commonly accessed through transition metal-catalyzed coupling processes. More recently, the desired coupling has been achieved through the reaction of phosphonate esters or phosphine oxides with highly electrophilic arylcopper(iii) intermediates generated from oxidation of copper(i) with diaryliodionium(iii) salts [eqn (1)]. While the reported combined photoredox/gold-catalyzed reactions have relied on the intervention of carbon–carbon π-bonds [eqn (2)], we hypothesized that the gold(iii) intermediates generated in this manner might also engage in coupling reactions with other nucleophilic species [eqn (3)].

3To this end, we explored the dual photoredox/gold-catalyzed coupling reaction of p-tolyldiazonium with diethyl phosphite (Table 1). The initial screening of solvents found that the desired product was formed in 37% yield when acetonitrile was employed as solvent (Table 1, entry 1). Other solvents commonly employed in photocatalysis, such as DMF and EtOH, were also tested and afforded the product in 50% and 65%, respectively (Table 1, entries 2 and 3). In order to exploit the better solubility of diazonium salts and tautomerization of H-phosphonates in polar solvents, we explored whether a solvent mixture with ethanol might improve the reaction outcome. The co-solvent consisting of MeCN/EtOH (4: 1) gave the best result, affording the arylphosphonate in 82% yield (Table 1, entries 4–7). Lower yields were obtained when Ru(bpy)3Cl2 or Ir(ppy)3 were used as the photocatalyst (Table 1, entries 8 and 9). No product was observed with IPrAuCl as the gold catalyst (Table 1, entry 10), and reducing the amount of photocatalyst or gold catalyst both gave lower yields (Table 1, entries 11 and 12). Moreover, no product was detected when the reaction was performed in the absence of the gold catalyst (Table 1, entry 13) and significantly reduced yields of 3a were observed in the absence of the Ru(bpy)3(PF6)2 and/or light (Table 1, entries 14 and 15). We also examined replacing the gold catalyst with those derived from palladium, silver or copper salts; however, only moderate yield was obtained when Pd(OAc)2 was used (Table 1, entry 16), and no product was detected with other catalysts even when the reaction time was prolonged to 16 h (Table 1, entries 17–22).

4With the optimized conditions in hand, we investigated the scope of the diazonium substrates. Aryldiazoniums salts bearing electron-donating groups at their para-positions, such as methyl, phenyl and methoxy, were coupled with diethyl phosphite affording the corresponding products in good to excellent yields (Table 2, compounds 3a–3c). However, the reactivity was dramatically decreased when aryldiazoniums salts containing strong withdrawing groups such as –CF3 and –NO2 were used (Table 2, compounds 3k and 3l). As expected, the P-arylation using aryl diazoniums with halogen in their para and meta positions proceeded efficiently, and yields of 71–86% were obtained (Table 2, compounds 3e–3g). However, 2-bromophenyl diazonium was less reactive comparatively and 37% yield of the product was obtained (Table 2, compound 3h). The naphthyl phosphonate was isolated in 72% yield under the standard reaction conditions (Table 2, compound 3i).

5We next turned our attention to an evaluation of the scope and limitations of our reaction with different types of P(O)H compounds. As seen in Table 3, aryl diazonium salts with electron-donating, electron-withdrawing and halogen substituents reacted with H-phosphonate diesters bearing different alkyl groups efficiently, and yields of 74–90% were obtained (Table 3, entries 1–5). The reaction also proceeded smoothly with dibenzyl and ethyl phenylphosphinate as coupling partners (Table 3, entries 6 and 8). The more challenging coupling of H-phosphonate diphenylester also occurred under the gold-catalyzed reaction conditions, albeit it in diminished yield (Table 3, entry 7).

6Interestingly, phenyl phosphinic acid was also a competent nucleophile for this reaction yielding products of a three-component coupling between the diazonium salt, arylphosphinic acid and the alkyl alcohol solvent [eqn (4)].

7Additionally, the intermediate aryl diazonium salt can be generated without purification from the corresponding aniline. For example, 3e was obtained in 69% through one-pot, two-step procedure for the diazotization and P-arylation of 4-fluoroaniline, compared with 86% under standard conditions [eqn (5)].

8In conclusion, we have developed the first gold-catalyzed oxidative P-arylation of H-phosphonates promoted by visible light photoredox catalysis. The reaction proceeds under mild reaction conditions (room temperature, no base) and shows excellent substrate scope, including the use of phosphinic acids as coupling partners. More broadly, the use of photoredox catalysis to achieve the oxidation event required for cross-coupling,, avoids the need for strong oxidants associated with known gold-catalyzed coupling reactions. This feature putatively allows for increased functional group compatibility, as clearly demonstrated by the gold-catalyzed formation of alkyne-substituted phosphinate ester 3v, in which the potentially reactive carbon–carbon π-bond is left intact, and can subsequently be engaged in a copper-catalyzed alkyne-azide click reaction [eqn (6)]. The development of this strategy for cross-coupling and detailed mechanistic studies is ongoing in our group.