Direct and Scalable Electroreduction of Triphenylphosphine Oxide to Triphenylphosphine

Abstract

The direct and scalable electroreduction of triphenylphosphine oxide (TPPO)the stoichiometric byproduct of some of the most common synthetic organic reactionsto triphenylphosphine (TPP) remains an unmet challenge that would dramatically reduce the cost and waste associated with performing desirable reactions that are mediated by TPP on a large scale. This report details an electrochemical methodology for the single-step reduction of TPPO to TPP using an aluminum anode in combination with a supporting electrolyte that continuously regenerates a Lewis acid from the products of anodic oxidation. The resulting Lewis acid activates TPPO for reduction at mild potentials and promotes P−O over P−C bond cleavage to selectively form TPP over other byproducts. Finally, this robust methodology is applied to (i) the reduction of synthetically useful classes of phosphine oxides, (ii) the one-pot recycling of TPPO generated from a Wittig reaction, and (iii) the gram-scale reduction of TPPO at high concentration (1 M) with continuous product extraction and in flow at high current density.

A Systems Approach to a One‐Pot Electrochemical Wittig Olefination Avoiding the Use of Chemical Reductant or Sacrificial Electrode

“…The activation of the P=O bond in TPPO via formation of 1 further leads to a drastic decrease of the redox potential as evident from the cyclic voltammograms recorded in acetonitrile solutions with 0.15 m tetrabutylammonium hexafluorophosphate (TBAPF 6 ). A pseudo‐reversible one‐electron‐reduction of TPPO was observed at 2.92 V (vs. Fc/Fc + ) (Figure b, black curve). The methylation of P=O to P−OMe + causes an anodic shift of 870 mV and irreversible reduction of 1 (Figure b, red curve) which, in contrast, has not been previously achieved with TPPO→LA (LA: B(OAr) 3 and AlCl 3 ) adducts, apparently due to weaker association of TPPO and LA .…”

“…A pseudo‐reversible one‐electron‐reduction of TPPO was observed at 2.92 V (vs. Fc/Fc + ) (Figure b, black curve). The methylation of P=O to P−OMe + causes an anodic shift of 870 mV and irreversible reduction of 1 (Figure b, red curve) which, in contrast, has not been previously achieved with TPPO→LA (LA: B(OAr) 3 and AlCl 3 ) adducts, apparently due to weaker association of TPPO and LA . However, the estimated free energy change (Δ G ) determined by DFT calculations for the one‐electron‐reduction of [Ph 3 P(OMe)] + (Figure S5) is about 63.7 kcal mol −1 preferable in comparison to TPPO, while that of the TPPO→AlCl 3 adduct is stabilized by only 9–10 kcal mol −1 , as reported recently .…”

“…The methylation of P=O to P−OMe + causes an anodic shift of 870 mV and irreversible reduction of 1 (Figure b, red curve) which, in contrast, has not been previously achieved with TPPO→LA (LA: B(OAr) 3 and AlCl 3 ) adducts, apparently due to weaker association of TPPO and LA . However, the estimated free energy change (Δ G ) determined by DFT calculations for the one‐electron‐reduction of [Ph 3 P(OMe)] + (Figure S5) is about 63.7 kcal mol −1 preferable in comparison to TPPO, while that of the TPPO→AlCl 3 adduct is stabilized by only 9–10 kcal mol −1 , as reported recently . Additional four different phosphine oxides (R 3 PO) were selected bearing phenyl and/or alkyl substituents R attached to the P center and using the same methodology of phosphoryl P=O bond via alkylation, corresponding methoxy‐ and/or ethoxy‐phosphonium salts [R 3 P(OR)](OTf) were synthesized and isolated (see Supporting Information, Figures S6–S17).…”

“…The displacement of Ph groups in 1 by alkyl substituents has also a significant influence on the WOR efficiency: using [Ph 2 MeP(OMe)](OTf), MeOTf and PhCHO lowers the yields of styrene to 31 % (entry 5 in Table , Figures S37 and S38), while monoethyl substitution on the P atom ([Ph 2 EtP(OMe)](OTf)) leads merely to unwanted (P)O−C bond scission and formation of Ph 2 EtP=O (entry 6 in, Table , and Figures S39–S41). Notably, due to the higher reduction potentials of fully alkyl‐substituted phosphine oxides ( n Bu 3 PO, n Oct 3 PO), electro‐chemical and/or chemical reduction remain unfavorable under the here applied conditions . Although DFT calculations performed with [ n Bu 3 P(OMe)] + revealed that the one‐electron reduction of the phosphonium center is energetically favorable, no minima was obtained in the potential energy surface for the two‐electron‐reduction to n Bu 3 P (Scheme S2).…”

“…In the earlier reports of electroreduction of TPPO to TPP, the LAs such as B(OAr) 3 , AlCl 3 and Me 3 SiCl act as sinks to trap O 2− , producing the very stable [{(OAr) 3 B} 2 O] 2− diborate, Al 2 O 3 , and (Me 3 Si) 2 O disiloxane, respectively. In contrast, the electro‐reductive P−O bond cleavage of 1 and 2 furnishes the respective alcohols CH 3 OH (from 1 in 68 % yields) and C 2 H 5 OH (from 2 in 75 % yields), respectively (see Figures S52–S54).…”