Main Group Redox Catalysis of Organopnictogens: Vertical Periodic Trends and Emerging Opportunities in Group 15

Abstract

A growing number of organopnictogen redox catalytic methods have emerged–especially within the past ten years–that leverage the plentiful reversible two-electron redox chemistry within group 15. The goal of this Perspective is to provide the context to understand the dramatic developments in organopnictogen catalysis over the past decade with an eye towards future development. An exposition of the fundamental differences in the atomic structure and bonding of the pnictogens, and thus the molecular electronic structure of organopnictogen compounds, is presented to establish the backdrop against which organopnictogen redox reactivity–and ultimately catalysis–is framed. A deep appreciation of these underlying periodic principles informs an understanding of the differing modes of organopnictogen redox catalysis and evokes the key challenges to the field moving forward. We close by addressing forward-looking directions likely to animate this area in the years to come. What new catalytic manifolds can be developed through creative catalyst and reaction design that take advantage of the intrinsic redox reactivity of the pnictogens to drive new discoveries in catalysis?

Graphical Abstract

1. Introduction

Chemistry is patterned by elemental properties arising from the quantum structure of atoms.¹ As systematized in the periodic table, an element’s periodic position corresponds with approximate expectations about its properties.² Accordingly, the redox reactivity of the elements is usefully (even if somewhat over-simplistically) abstracted according to their periodic ‘block’. For elements in the s- and f-blocks, single oxidation states³ (+I or +II for s elements,^4-6 +III for f elements^7,8) tend to prevail; by contrast, numerous stable oxidation states separated by modest reduction potentials proliferate among the transition metals of the d-block.⁹ Especially for the late transition metals of the second (4d) and third row (5d), the prevalence of accessible two-electron redox processes provides the thermodynamic and mechanistic basis¹⁰ upon which innumerable groundbreaking discoveries in catalytic synthesis are built.^11-14

The elements of the p-block—especially the ‘heavier’ entrants of principle quantum number n≥3—are more akin to their neighbors in the d-block than they are to either the s- or f-blocks in terms of breadth of accessible oxidation states. Representatively, compounds of the group 15 elements (collectively known as the pnictogens,^15,16 abbreviated Pn) express a rich redox reactivity,^17-20 where the valence electronic ns²np³ configuration gives rise to compounds that span −III to +V oxidation states.^21-25

Correspondingly, discrete chemical reactions involving redox events at pnictogen centers have been described since at least the early 19^th century.^26,27 Since that time, many developments in the synthetic chemistry of organopnictogen-based two-electron redox are intimately connected to pioneering achievements of 20^th century organic chemistry. Staudinger’s reduction of organic azides by P(III) reagents to give P(V) iminophosphoranes is a bedrock reaction in organic synthesis^28-30 that continues to find new applications in catalysis^31-33 and chemical biology.^34-39 Wittig’s olefin synthesis,^40-42 which leverages the driving force P^III→P^V=O, ushered in a new era in industrial preparation of carotenoids, such as vitamin A.⁴³ Further down group 15, unique aryl transfer reagents were introduced by Barton based on the conversion Bi^V→Bi^III,^44-46 a forerunner to ongoing oxidative organopnictogen method development. In short, the impact of organopnictogen-based redox methods on synthesis is both long and celebrated.

New developments that merge elementary organopnictogen redox reactions into catalytic cycles involving formal two-electron redox cycling have been gathering pace, especially within the past decade. These developments, proceeding in parallel with ongoing synthetic redox method developments elsewhere in the p-block in Groups 13,^47-50 14,^51,52 16,^53-58 and 17,^59-68 represent the vanguard of a new class of redox catalysts composed of main group elements that evoke an analogy with well-established activation modes of the late d-block elements.^69-74

Along with ample fundamental science motivations, the attractiveness of redox catalysts derived from the heavier group 15 elements is buoyed in a practical sense by the relative abundance and low cost of these pnictogens.⁷⁵ Phosphorus is abundant both in the earth’s crust (1300 ppm) and in the biosphere, being the only member of the pnictogen family other than nitrogen that is essential to life. While the heavier pnictogens are comparatively more scarce (As, 5.7 ppm; Sb, 0.75 ppm; Bi, 0.23 ppm), all are produced on >20,000 ton scale annually.⁷⁶ And though bismuth is only roughly as abundant terrestrially as palladium (0.52 ppm) and platinum (0.5 ppm), it is 10³–10⁴ times less expensive on a per kilogram basis (cf. $7.50/kg for Bi, ~$75000/kg for Pd, ~$33000/kg for Pt). Indeed, established non-redox activation modes in organopnictogen catalysis (i.e. Lewis acid,^77-79 Lewis base,^80-82 and frustrated Lewis pair^83-86 catalysis), along with the long history of Group 15 compounds as supporting ligands in organometallic chemistry,^87-94 serve as a validation of the viability of organopnictogens as constituents of practical catalysts.

In this Perspective, we wish to highlight exciting recent advances in the burgeoning field of organopnictogen redox catalysis. Our major goals are: (1) to identify the pivotal contributions defining the current state of the art and (2) to articulate future directions that are likely to define the forefront of research moving forward. Toward these goals, we first trace the fundamental periodic properties of the group 15 elements and then illustrate how these periodic trends are expressed in the diversity of reactions driven by group 15 redox catalysis. In this way, we hope to convey not only an appreciation of the new synthetic capabilities revealed by group 15 redox catalysis, but also a context for understanding of the relationships—both similarities and distinctions—between the congeneric elements in terms of their catalytic chemistry. By conceptualizing group 15 redox catalysis in this way as a worthy catalytic modality, we hope that this Perspective will knit together the broad cross-section of synthetic inorganic and organic chemists active in the organopnictogen area and serve to nucleate new efforts in this productive and promising area of research.

2. Periodicity and Vertical Trends in Group 15

Given that an informed understanding of the periodic trends and the related structural, bonding, and electronic features of organopnictogens establishes the guiding principles for further development of this field of catalysis, the purpose of this section is to provide a targeted evaluation of key features of the elements themselves and organic molecules containing them that drive the redox catalytic reactivity endemic to each pnictogen. Interested readers can find further elaboration of many of these themes in prior monographs and reviews.^95-100

2.1. Trends in Atomic Electronic Structure.

The importance of atomic electronic structure in chemical bonding and reactivity is an essential feature of molecular orbital theory. As expanded below, the relative importance of s and p valence atomic functions in organopnictogen bonding and molecular structure—and thus reactivity— varies intrinsically with spatial and energetic atomic orbital disposition.

2.1.1. Valence orbital size.

A graph of the radial probability maxima for the valence s and p orbitals of the group 15 elements is given in Figure 1A.¹⁰¹ As expected for the increasing principal quantum number, the radial extension of the valence AOs increases down the group, but three subtleties of the periodic atomic electronic structure are noteworthy. First, the increase in size—though monotonic—is not smooth. Instead, a ‘sawtooth’ shape is evident, such that the van der Waals radii of P (1.80 Å) and As (1.85 Å) are clustered, as are Sb (2.05 Å) and Bi (2.07 Å). This effect has been attributed to a ‘secondary periodicity’^102,103 arising from incomplete screening of nuclear charge owing to the intervention of the d- and f-elements on period 4 (As) and 6 (Bi), respectively (i.e. the ‘scandide’ and ‘lanthanide’ contractions).¹⁰⁴ Second, the increase in radial extension does not affect s and p orbitals equivalently.¹⁰⁵ For valence 2s and 2p orbitals of nitrogen, the probability maximum in the radial distribution function is nearly identical (0.54 and 0.52 Å, respectively), but for the 3s and 3p orbitals of phosphorus it differs by ca. 15%. The radial differences between ns and np are even more pronounced for As, Sb, and Bi. This phenomenon arises because the 2p orbital lacks a core shell of the same angular momentum (l=1) and thus does not have a radial node, whereas radial nodes are requisite for all p orbitals of higher principal quantum number (n>2) to satisfy quantum orthogonality. In effect, the first-filled p orbital shell exerts an outward effect on all higher p shells through ‘primogenic repulsion,’ as coined by Pyykkö.^106,107 Kaupp has further emphasized the importance of radial nodes in main group bonding and reactivity.^108,109 Third, spin-orbit coupling and relativistic effects take on significant importance for bismuth.^110-113 The 6p_1/2 and 6p_3/2 spinors diverge markedly in radial extension, and the 6s orbital experiences a significant contraction compared to a notional ‘nonrelativistic bismuth.’ The importance of these orbital effects, especially the latter, has very profound consequences for the chemical and redox reactivity of bismuth (vide infra).

Figure 1.

(A) Atomic orbital radial probability function of Group 15 elements. (B) Valence atomic orbital 1-electron ionization energies of Group 15 elements.

2.1.2. Valence orbital ionization energies.

A plot of the valence atomic orbital one-electron ionization energies is shown in Figure 1B.¹⁰¹ As seen especially for the heavier pnictogens (P─Bi), valence p orbital energy increases uniformly down group 15. By contrast, the s orbital ionization energy does not exhibit such a monotonic trend. Instead, the ‘sawtooth’ profile is again seen; note for instance that the magnitude of the one-electron binding energy of the As 4s orbital is slightly larger than that of the P 3s orbital and that the Bi 6s orbital ionization energy is substantially larger than that of the Sb 5s orbital. These effects can be traced back to the d- and f-block contractions,¹⁰⁴ which is augmented in the latter case by the relativistic stabilization of the Bi 6s orbital and spin-orbit splitting of the p½ and p³⁄₂ orbital energies.¹¹⁴

2.2. Trends in Molecular and Electronic Structure.

2.2.1. Bonding and Hybridization.

The interplay of AO radial sizes and energies has significant effects on the bonding of the heavier pnictogens. Kutzelnigg has explained that the decreased spatial overlap of the s and p orbitals down group 15 results in less s/p mixing and a lifting of the orthogonality for s/p hybrid orbitals.^115,116 As illustrated by Kaupp for the series H₃Pn (Pn=P─Bi), valence s-character accumulates in the non-bonding lone-pair orbital down the group, and the Pn─H bonds tend to be made increasingly from essentially unhybridized p-orbitals.¹⁰⁹ This ‘hybridization defect’ arising from the increasingly disparate s and p orbital sizes generally leads to weakening of σ bond energies down group 15. Thus, for the series H₃Pn (Pn=P─Bi), a consistent decrease in the Pn─H bond dissociation enthalpy is observed down the group (P: 81.4, As: 74.6, Sb: 63.3, Bi: 51.8 kcal/mol).^117,118

2.2.2. Tricoordination.

Data for the triphenylpnictogen(III) compounds (Ph₃Pn) in the Cambridge Structural Database¹¹⁹ exemplify the periodic trend in molecular structure that trace the molecular-electronic structure nexus (Fig. 2). In accord with the trend in atomic size (Sect 2.1.1), a sawtooth-like increase in Pn─C bond lengths in the PnPh₃ series – PPh₃ (CSD-1238522),¹²⁰ AsPh₃ (CSD-1318411),¹²¹ SbPh₃ (CSD-1318403),¹²² BiPh₃ (CSD-1468789)¹²³ – is observed, where P─C (1.93 Å) and As─C (1.96 Å) are shorter bond lengths than Sb─C (2.15 Å) and Bi─C (2.25 Å). Relatedly, the average bond angle ∠C-Pn-C decreases down the group: ∠C-P-C 102.7°, ∠C-As-C 100.4°, ∠C-Sb-C 96.6°, and ∠C-Bi-C 93.7°. Two mutually reinforcing effects drive this trend: (1) the longer bond lengths of the heavier pnictogens ease steric crowding between the aryl substituents and thus permit narrower bond angles, and (2) the s/p hybridization defect leads to increasingly directional bonding down the group (i.e. higher p-orbital contribution to Pn─C bonding and greater accumulation of s-character in the nonbonding lone pair). The longer bond lengths and greater pyramidalization of the heavier pnictogens are common features of trigonal tricoordinate group 15. As a corollary, the barrier to pyramidal inversion of trivalent organopnictogens via the ‘umbrella coordinate’ increases down the group.^124,125 By transit from a pyramidal C_3v to a planar D_3h geometry, the HOMO nonbonding lone pair (2a₁) correlates with the atomic p orbital oriented along the rotational axis. The energetic penalty to planarization thus imposed, which is accentuated in the case of bismuth by the relativistic stabilization of the 6s orbital relative to the 6p set,¹²⁶ has been correlated with the electronegativity of the central pnictogen within the context of a second-order Jahn-Teller effect.¹²⁷

Figure 2.

(top) Solid-state structures for Ph₃Pn (Pn = P, As, Sb, Bi) viewed orthonormal to one of the equivalent C_α-Pn-C_α’ planes. Periodic variation in bond angles and pyramidalization are thereby best visualized. (bottom) Tabulated structural data for Ph₃Pn, and computed inversion barriers of PnH₃.

As will be detailed in subsequent sections, many of the organopnictogen compounds that exhibit catalytic redox properties are nontrigonal (i.e. no local threefold symmetry).⁷⁴ The interrelation of molecular geometry and electronic structure of nontrigonal compounds can be approached by reference to the frontier correlation diagram in Figure 3. Descent from local C_3v symmetry by progression along the bending (e symmetry) normal mode gives C_s-symmetric structures. Electronically, the consequence of this symmetry-lowering distortion is a lifting of the degeneracy of the unfilled orbitals resulting in a decrease in the HOMO-LUMO energy gap. Computational and experimental validation for this electronic picture has been established for nontrigonal chelates of pnictogen(III) triamide compounds.^128-131 The ability to construct pnictogen compounds of diverse molecular shapes by appropriate constraint allows for electronic structure tailoring with profound consequences for the future of catalysis in this area.

Figure 3.

Qualitative correlation diagram for frontier orbitals in C_3v symmetry (center) upon descent to C_s symmetry (left and right). Orbital projections are viewed down the σ plane.

2.2.3. Pentacoordination.

In parallel to the foregoing discussion of tricoordinate pnictogen(III) compounds, the pentaphenylpnictogen(V) compounds (Ph₅Pn) first prepared by Wittig^132-134 illustrate relevant periodic trends for molecular compounds in pentacoordination (Fig. 4). Solid state structures for Ph₅P (CSD-1232414)¹³⁵ and Ph₅As (CSD-1230863)¹³⁶ are well-described as trigonal bipyramidal (τ = 0.90 and 0.98, respectively). By contrast, the heavier congeners Ph₅Sb (CSD-1232410)¹³⁷ and Ph₅Bi (CSD-1254431)¹³⁸ crystallize as distorted square pyramidal structures (τ = 0.25 and 0.22, respectively).^139,140 These static structures provide snapshots spanning the Berry pseudorotation coordinate,^141,142 and spectroscopic evidence supports that they persist in solution.¹⁴³ Intriguingly, whereas Ph₅P, Ph₅As, and Ph₅Sb are all colorless solids, Ph₅Bi is violet.^144,145 Seppelt and Pyykkö have provided evidence that a ligand-to-metal charge transfer excitation in the visible region results from Bi-based LUMO composed of the relativistically-stabilized 6s orbital.^146,147 Without relativistic considerations, the HOMO-LUMO gap is predicted to be 27% larger, such that “‘nonrelativistic’ pentaphenylbismuth would not be violet.” The connection between the observed low-energy optical transition and the propensity for Ph₅Bi to react as an electrophilic aryl transfer reagent has been noted.¹⁴⁸

Figure 4.

Solid-state structures and structural data of Ph₅Pn.

2.3. Trends in Dative and Redox Reactivity.

2.3.1. Measures of donor reactivity.

The dissociation enthalpy for Lewis adducts with group 13 Lewis acids provides a measure of donor ability of trivalent organopnictogens.¹⁴⁹ On the basis of gas phase experiments with AlH₃, acid-base adduct formation is most favorable for P and least favorable for Bi (Fig, 5, left). These findings correlate with qualitative observations regarding nucleophilic reactivity; triphenylphosphine and triphenylarsine readily undergo alkylation with methyl iodide, but triphenylstibine requires the more reactive trimethyloxonium electrophile (Me₃O)BF₄ to undergo quaternization, while triphenylbismuth is not quaternized even with (Me₃O)BF₄.^150,151 However, steric effects often are entangled with this underlying trend. Specifically, the relatively small atomic radii of phosphines and arsines relative to stibines and bismuthines give rise to a substantial repulsive interaction with sterically encumbered Lewis acids (ⁱPr₃Pn─Al^tBu₃ series, Fig. 5, right), resulting in accordingly diminished energetic stabilization of the Lewis adduct. In effect, the lighter pnictogens are more sensitive to steric influences than their heavier congeners.^152-154

Figure 5.

Gas-phase Pn–Al distances and dissociation enthalpies (D_e) of Lewis adducts H₃Al–PnⁱPr₃ and ^tBu₃Al–PnⁱPr₃.

2.3.2. Aqueous reduction potentials.

The standard electrode potentials for the group 15 ions in aqueous solution establish an important trend governing the redox reactivity of these elements.¹⁵⁵ As shown in the Frost diagram in Figure 6,¹⁵⁶ phosphorus is the only element for which the Pn(III) and Pn(V) oxyacids are more stable than the elemental form. These positive oxidation states become increasingly unstable down the group; high valent Bi(V) is the least stable among Pn(V) congeners. This increasing preference for the lower valent state among the heavier group 15 elements can be viewed as a manifestation of the ‘inert pair effect,^{98-100,157-159} which in turn may be related to the hybridization defects within the high-valent compounds.¹¹⁶ Computed Pn(V)=O bond energies from the reaction H₃Pn + 0.5O₂ → H₃Pn=O series (MP2/DZ+d level) display a similar effect, wherein phosphorus forms the most stable oxide while bismuthine oxide is energetically uphill.¹⁶⁰ As such, as a general rule the Pn^III/Pn^V redox couples can be summarized as follow: P^III/P^V is strongly reducing, As^III/As^V and Sb^III/Sb^V are mildly oxidizing, and Bi^III/Bi^V is strongly oxidizing.

Figure 6.

Frost oxidation state diagram for heavier pnictogens under acidic conditions.

2.3.3. Thermodynamics of reductive elimination from 5-coordinate pnictoranes

Similarly to the relative stabilities of the pnictide oxides described above, evaluation of the relative thermodynamic stabilities PnX₅ compounds with respect to PnX₃ illuminates a periodic trend (Fig. 7). Among the PnF₅ congeners, BiF_5, which is known to fluorinate hydrocarbons,^161,162 is at least 45 kcal/mol less stable than the lighter congeners,^163,164 such as the stable, Lewis acidic PF₅. Similarly, the PnH₅ series¹⁶⁵ displays an irregular thermodynamic trend for the liberation of H₂ and PnH₃, in which decomposition of BiH₅ is at least 20 kcal/mol more favorable than any of the lighter congeners, owing again to the substantially more oxidizing nature of Bi(V). As will be shown, these general characterizations manifest in markedly differing reactivity of organopnictogens, and thus provide a framework for appreciating the divergences in application in redox catalysis.

Figure 7.

Gas-phase energies (kcal/mol) for the reactions PnF₅→PnF₃ + F₂ and PnH₅→PnH₃ + H₂.

The foregoing atomic properties and molecular reactivity trends are the fundamental backdrop against which the varied organopnictogen reactivity described in this Perspective is brought into relief. As described below, these periodic trends govern much of the divergent structure, bonding, and electronic nature of the recently uncovered examples of organopnictogen redox catalysis.

3. State-of-the-Art Developments in Organopnictogen Redox Catalysis

Although an initial report can be traced to 1981,¹⁶⁶ the overwhelming majority of demonstrations in the field of organopnictogen redox catalysis have come in the past decade. In this section, the key developments will be discussed, organized first by pnictogen element and then by reaction type. The purpose of this section is to develop a systematic perspective on the state of the field, with an eye toward understanding prevailing themes and, accordingly, gaps in current knowledge which might present avenues for further research.

3.1. Organophosphorus Redox Catalysis.

As described in the preceding section, the redox chemistry of organophosphorus molecules is primarily driven by the reducing nature of the P(III) state and the relative stability of the P(V) oxidation state, especially those compounds possessing P(V)=O moieties.^167-169 As such, the oxidation of P(III) compounds to stable P(V) species can be accomplished with a variety of oxidants of even modest oxidizing power.^170-177 Conversely, the reduction of P(V) to P(III) is often contrathermodynamic, thus requiring relatively forcing conditions or bespoke molecular design;^178-182 this presents the primary challenge in achieving organophosphorus redox cycling.

In practice, the relatively strong P(V)=O bond typically requires strong reductants¹⁸³ such as metal hydrides to generate the P(III) species via the intermediacy of a hydridophosphorane.¹⁸⁴ However, the barrier to reduction is lower for constrained cyclic phosphine oxides relative to unstrained cyclic¹⁸⁵ or acyclic^186,187 congeners, providing a rationale for catalyst design operating in the P^III/P^V=O couple. Reductive ligand coupling represents another available avenue for the reduction of P(V) species, as described in a seminal report by Mann in 1948^188-192 and modernized into a programmable process by McNally.^193-195 Nevertheless, the poor driving force inherent to P(V) to P(III) reduction presents the key challenge in achieving self-contained organophosphorus redox cycling.

3.1.1. Catalytic Wittig Reaction.

The Wittig reaction is a cornerstone of organophosphorus chemistry, and efforts to render it catalytic in phosphine require a strategy for mild and swift reduction of the phosphine oxide byproduct to enable P^III/P^V redox cycling. The past 12 years have seen the successful application of novel organophosphorus molecules to achieve such a feat.^196-199 In 2009, O’Brien reported the first example of organophosphorus redox catalysis using a five-membered phospholane oxide (3-methyl-1-phenylphospholane 1-oxide) P1•[O] operating in the P^III/P^V=O couple in the context of a Wittig reaction (Fig. 8A).²⁰⁰ This strategy uses a mild hydrosilane reductant, Ph₂SiH₂, to reduce the phospholane oxide precatalyst to the active P(III) species, which can then undergo quaternization, deprotonation, and Wittig reaction to obtain the desired product olefin and regenerate the phospholane oxide pre-catalyst. In 2013, O’Brien significantly lowered the reaction temperature for catalytic Wittig reactions to ambient temperature via the use of a Bronsted acidic additive, 4-nitrobenzoic acid, which enhances the rate of reduction of phosphine oxide P2•[O] (Fig. 8B).²⁰¹ O’Brien further developed a series of electron-deficient phospholane oxide precatalysts, including P3•[O], to enable the use of non-stabilized ylides in the catalytic Wittig reaction (Fig. 8C).²⁰²

Figure 8.

Phospholane-catalyzed Wittig reaction with (A) stabilized ylides and mechanism, (B) stabilized ylides at ambient temperature through inclusion of Brønsted acid additive, and (C) unstabilized ylides through development of electron-deficient phospholane catalyst.

Other organophosphorus catalyst scaffolds have proven adept at achieving catalytic Wittig reactions. In 2019, Werner demonstrated the utility of phosphetane²⁰³ oxide P4•[O]²⁰⁴ to enable catalytic Wittig reaction at 1 mol% catalyst loading at ambient temperature in the absence of any acidic additive (Fig. 9).²⁰⁵ Simple phosphine oxide precatalysts, such as Ph₃PO, Oct₃PO, or Bu₃PO, have been explored for catalytic Wittig reactions, but to this point have required the assistance of microwave heating or Bronsted acid additive at high temperature.^206,207

Figure 9.

Phosphetane-catalyzed Wittig reaction with stabilized ylides at ambient temperature.

In 2014, Werner demonstrated the first enantioselective catalytic Wittig reaction operating in a P^III/P^V=O couple, highlighting some challenges in realizing such a method (Fig. 10). In this work, a variety of chiral phosphine catalysts are applied for desymmetrization of prochiral haloketone 11 to give enantioenriched diketone 12. The most promising result utilizes (S,S)-Me-DuPhos (P5), a C₂-symmetric bisphospholane,²⁰⁸ with phenylsilane as the terminal reductant in dioxane via microwave heating at 150 °C, which gives 39% yield and 62% ee.^209,210

Figure 10.

Chiral phospholane-catalyzed asymmetric Wittig cyclization.

The formation of the phosphorus ylide can also be achieved in the absence of base through conjugate addition to activated olefins and proton transfer, as exemplified by Werner and Lin (acrylates),^211-214 Vouturiez (ynoates),^215,216 Kwon (allenes),²¹⁷ and Lin (enones)²¹⁸ using a selection of catalysts previously described (Fig. 11A). Of particular note is an enantioselective variant enabling the synthesis of (trifluoromethyl)cyclobutenes (Fig. 11B)²¹⁹ developed by Voituriez in 2018 with Kwon’s bicyclic chiral phosphine oxide HypPhos P7•[O].

Figure 11.

A) Catalytic Wittig reactions from unsaturated ylide precursors. B) Asymmetric organophosphorus-catalyzed (trifluoromethyl)cyclobutene formation via a conjugate addition/Wittig olefination reaction.

3.1.2. Catalytic Staudinger Reactions and Aza-Wittig.

In 2012, van Delft and Rutjes reported the first catalytic Staudinger reaction with a dibenzophosphole catalyst P10 and PhSiH₃ as reductant (Fig. 12A).²²⁰ In contrast to iminophosphorane hydrolysis employed in the stoichiometric reaction, the catalytic reaction involves direct reduction of the P(V) iminophosphorane with PhSiH₃ for the formation of the amine product and regeneration of the phosphine catalyst.²²¹ PPh₃ (P11) could also be used in place of the dibenzophosphole under identical conditions, albeit with significantly prolonged reaction times. Mecinović later demonstrated an ambient temperature protocol by employing an optimized hydrosilane reductant.²²² Catalytically formed iminophosphoranes from PPh₃ (P11) can also be used for Staudinger amidation reactions (Fig. 12B),²²³ although the precise mechanism of the redox cycle is unclear.^224-226

Figure 12.

A) Dibenzophosphole-catalyzed Staudinger reduction. B) PPh₃-catalyzed Staudinger ligation.

Other applications of iminophosphorane intermediates in the context of P^III/P^V=O cycling include catalytic aza-Wittig^227-230 and diaza-Wittig reactions (Fig. 13A).²³¹ In 2018, Kwon demonstrated the first catalytic asymmetric Staudinger-aza-Wittig reaction^232,233 with high levels of stereoinduction via desymmetrization of diketones using HypPhos catalyst P12 with the assistance of a Brønsted acid additive (Fig. 13B).²³⁴

Figure 13.

A) Catalytic aza-Wittig reactions using benzo[b]phosphindole. B) Catalytic enantioselective aza-Wittig synthesis of chiral heterocycles catalyzed by HypPhos. Bz = benzoyl; Cy = cyclohexyl; Ts = tosyl.

3.1.3. Catalytic Appel and Mitsunobu Reactions.

Organophosphorus catalyzed oxidation-reduction condensation reactions,^235,236 such as the Appel and Mitsunobu reactions, face challenges of reagent compatibility (between halenium/azo oxidant and hydrosilane reductant) and product stability. In 2011, Rutjes and van Delft achieved a P^III/P^V=O catalyzed Appel bromination (Fig. 14A).¹⁸⁵ In this transformation, diethyl bromomalonate (DEBM) is an ideal bromenium donor, showing good compatibility with hydrosilane reductants. Further, the dibenzophosphole catalyst P10 is exclusively reactive toward the bromenium source, thus selectively generating the electrophilic bromophosponium ion, but unreactive towards the brominated products.^237,238 Recently, Werner further extended the scope to chlorination of alcohols with benzotrichloride as oxidant and trioctylphosphine (P13) as the catalyst (Fig. 14B).²³⁹ Catalytic Appel conditions with PPh₃ (P11) can also be used to drive amide couplings between carboxylic acids and amines, as demonstrated by Mecinović in 2014 (Fig. 14C).²⁴⁰ Alternatively, Denton has extensively developed redox-neutral P^V-mediated dehydrative halogenation reactions using Ph₃PO as catalyst with oxalyl chloride as dehydrative reagent to enable phosphine oxide/phosphonium cycling.^241-248

Figure 14.

Organophosphorus-catalyzed Appel A) bromination, B) chlorination, and C) amidation. Bn = benzyl.

Recently, an annulation of amines and carboxylic acids was described via organophosphorus-driven recursive dehydration using phosphetane catalyst P4•[O], DEBM, and PhSiH₃ or Ph₂SiH₂ (Fig. 15).²⁴⁹ In this tandem catalytic reaction, the catalytically-generated bromophosphonium first induces amide coupling and then cyclodehydration in a second catalytic turnover. To facilitate the coupling of alkyl amines, fully-substituted diethyl (methyl)bromomalonate (DEMBM) is required to suppress N-alkylation. These conditions enable the coupling of pharmaceuticals, such as ibuprofen, without racemization at adjacent stereocenters, as well as the synthesis of dihydroisoquinoline natural products such as dihydropapaverine. Interestingly, the use of diethyl chloromalonate as the oxidant, and thus a chlorophosphonium intermediate as the dehydrating species, results in only amide bond formation.

Figure 15.

Phosphetane-catalyzed tandem annulation of amines and carboxylic acids by sequential C–N and C–C bond formation. p-tol = para-tolyl; ⁱBu = iso-butyl.

In 2010, O’Brien again successfully applied precatalyst P1•[O] in a catalytic Mitsunobu-type reaction (Fig. 16).²⁵⁰ Later, Aldrich disclosed some initial efforts into recycling both phosphine oxide and the azocarboxylate reagent, by an iron-phthalocyanine catalyzed process in the presence of oxygen.²⁵¹ However, a detailed study from Taniguchi reported difficulty in reproducing both yield and enantiomeric ratio for some examples, as well as successful product formation in the absence of hydrazine catalyst. These results indicate this reaction might not undergo a true Mitsunobu process, and further study appears to be necessary.^252,253 Recently, Denton has used creative catalyst design to enable redox-neutral P^V-based catalysis operating in a phosphine oxide/phosphonium cycle to achieve a highly successful catalytic Mitsunobu reaction.²⁵⁴

Figure 16.

Phospholane-catalyzed Mitsunobu-type reaction.

3.1.4. Catalytic Reductive O-Atom Transfer.

Owing to the strongly reducing nature of trivalent P(III) compounds, phosphines are excellent O-atom acceptors from a variety of oxygenated substrates. In 2010, Woerpel described the first P^III/P^V=O catalyzed reductive O-atom transfer by selective reduction of alkyl silyl peroxides to silyl ether products.²⁵⁵ The overall reaction is initiated by concerted insertion of triphenylphosphine into the O─O bond. Labeling and crossover studies demonstrate that a concerted elimination/silyl transfer step is operative in generating the silyl ether products and a phosphine oxide, which could in turn be selectively reduced by a titanium(III) hydride generated in situ.

To expand P^III/P^V=O catalyzed O-atom transfer to less-oxidizing oxygenated substrates, the catalytic chemistry of a biphilic²⁵⁶ phosphetane catalyst scaffold has been developed. In 2015, a phosphetane-catalyzed deoxygenative condensation reaction of α-keto esters and carboxylic acids via formal carbene insertion into the protic O─H bond of the acid was described (Fig. 17).²⁵⁷ The reaction initiates by Kukhtin-Ramirez addition²⁵⁸ of the P(III) phosphetane P14 to the keto ester substrate 48. Proton transfer from the benzoic acid followed by Arbuzov-like²⁵⁹ displacement of phosphine oxide P14•[O] from intermediate P14b results in formation of α-acyloxy ester product 50. The catalytic cycle is closed by reduction of phosphetane oxide P14•[O] to P14 by the hydrosilane reductant.

Figure 17.

P^III/P^V=O catalyzed deoxygenative condensation of α-keto esters with carboxylic acids.

The phosphetane scaffold is also effective for engaging nitro groups in O-atom transfer. Building on seminal stoichiometric work by Cadogan,^260-263 in 2017 a catalytic synthesis of indazoles and benzotriazoles from nitroimine and -azo starting materials, respectively, using P4•[O] as precatalyst under comparatively mild conditions was described (Fig. 18).²⁶⁴

Figure 18.

Biphilic phosphetane-catalyzed N–N bond-forming Cadogan heterocyclization via P^III/P^V=O redox cycling.

In this transformation, DFT models implicate a [3+1] cycloaddition of P(III) species to the nitro group as the turnover-limiting step. In accord with empirical observations, the barrier to this step with a phosphetane is significantly lower in energy than with an acyclic trialkylphosphine. Distortion-interaction analysis²⁶⁵ of the relevant transition structures (Fig. 19) shows that the differential barrier arises from an enhanced stabilizing interaction energy for the phosphetane rather than a diminished distortion penalty.^266-269 In effect, the contracted endocyclic C-P-C bond angle results in a low-lying LUMO, thus imbuing the phosphorus center with increased biphilic character relative to acyclic and larger phosphacyclic compounds. For comparison, a similar catalytic Cadogan transformation described by Nazaré using a larger-ring phospholene oxide precatalyst requires higher catalyst loadings and significantly longer reaction times.²⁷⁰

Figure 19.

Transition structures and distortion/interaction analyses for (3+1) transition states (M06-2X/6-311++g(d,p)): (A) phosphetane TS and (B) Me₃P TS. Phosphine distortion energy (ΔE_dP^‡) in green, nitromethane distortion energy (ΔE_dN^‡) in blue, fragment interaction energy (ΔE_i^‡) in red, activation energy (ΔE^‡) in black. All energies in kcal/mol without zero-point correction.

This approach to catalytic nitro deoxygenation has been similarly applied to C─N bond-forming reactions for the synthesis of carbazoles and indoles, as shown in Figure 20.²⁷¹ Here, oxazaphosphirane intermediate 55 was observed at low temperature as the immediate precursor to carbazole formation. DFT calculations suggest an oxazaphosphirane as the pivotal intermediate, which thermally dissociates phosphine oxide P4•[O] to reveal a free nitrene capable of evolving to the carbazole product via C─H amination.²⁷²

Figure 20.

Biphilic organophosphorus-catalyzed intramolecular C_sp²-H amination and identification of oxazaphosphirane intermediate.

Given that such an oxazaphosphirane intermediate might be targeted to further reaction development via heterolytic ring opening with a Lewis acid, introduction of an arylboronic acid partner to the P^III/P^V=O catalyzed nitro deoxygenation manifold resulted in a new reductive C─N cross coupling of nitroarenes and boronic acids (Fig. 21).²⁷³ The scope was subsequently expanded to allow the reductive coupling of nitromethane with both boronic acids and esters, providing an efficient strategy for installation of the MeHN− fragment with inexpensive and easy-to-handle nitromethane as the methylamine surrogate.²⁷⁴ By virtue of the nonmetal main group-catalyzed conditions for this C─N coupling, useful chemoselectivities are observed, establishing the method as a complement to existing transition metal-catalyzed techniques. Mechanistic investigations support a pathway involving formation of the oxazaphosphirane intermediate P4b, followed by engagement with the boronic acid 57 to make betaine P4c, leading to product formation via 1,2-metallate shift. This pathway is predicted to outcompete evolution of the oxazaphosphirane to a free nitrene 60, accounting for the excellent selectivity for intermolecular cross-coupling.^275,276 The C─N coupling event can be telescoped with subsequent ring closing events to allow for the synthesis of N-aryl heterocycles (58) by a cross-coupling/condensation cascade, as depicted in Figure 22.²⁷⁷

Figure 21.

P^III/P^V=O catalyzed intermolecular reductive C–N cross coupling of nitroarenes and boronic acids. Boc = tert-butyloxycarbonyl.

Figure 22.

P^III/P^V=O-catalyzed cascade synthesis of N-functionalized azaheterocycles.

Phosphetane oxide P4•[O] also efficiently catalyzes deoxygenative processing of sulfonyl chlorides (including trifluoromethyl- and heteroarylsulfonyl derivatives) by O-atom transfer (Fig. 23).²⁷⁸ This approach has been applied to an electrophilic sulfenylation of indoles via fleeting sulfenyl(ium) electrophilic equivalents.

Figure 23.

Phosphetane-catalyzed (fluoroalkyl)sulfenylation via deoxygenation of sulfonyl chlorides. Pin = pinacol.

3.1.5. Catalytic Hydride and Hydrogen Transfer.

Phosphetane-based catalysts have also been shown to drive regioselective transpositive reduction of allylic bromides through the intermediacy of P(V) hydrides (Fig. 24).²⁷⁹ The reaction benefits from the colocalized donor and acceptor properties of the phosphetane to achieve the necessary changes in both oxidation state and coordination number. Specifically, the reaction starts with quaternization of phosphetane P15 by the allylic bromide. In the presence of the stoichiometric reductant LiAlH(O-tBu)₃, hydride is delivered directly to the phosphorus center of allylic phosphonium cation P15a to give a hydridophosphorane P15b that is observable by low temperature ³¹P NMR spectroscopy. VT-NMR kinetics experiments and DFT calculations indicate that decomposition of pentacoordinate hydridophosphorane P15b to the reduction products occurs regiospecifically via a concerted 5-membered, 6-electron transition state (P15c). This pericylic γ-reductive elimination illustrates the unique merger of conventional organic and organometallic reactivities in catalytic chemistry of the p-block compounds.

Figure 24.

Organophosphorus-catalyzed regioselective reductive transposition of allylic bromides.

In a conceptually complementary hydride transfer reaction, an unusual transfer hydrogenation of azobenzene with ammonia borane catalyzed by P^III/P^V cycling was developed (Fig. 25). In this work, planar compound P16, introduced by Arduengo,^280-282 reacts with H₃N•BH₃ to give dihydridophosphorane P16a.²⁸³ Dihydride P16a in turn serves as a reactive hydrogen donor, transferring an H₂ equivalent to a variety of electrophilic organic acceptors. The combined reactivities of P16 as hydrogen acceptor from ammonia-borane and P16a as hydrogen donor to an organic substrate permit the use of this phosphorus platform as a catalyst for transfer hydrogenation. Although alternative pathways have been suggested via DFT studies,^284-286 experimental mechanistic investigations lead to the assertion that hydrogen transfer catalysis in this case involves P16⇌P16a cycling. These results establish precedent for ‘dihydride’ transfer hydrogenation with a p-block catalyst.

Figure 25.

P^III/P^V-catalyzed transfer hydrogenation of azobenzene.

3.2. Organoarsenic Redox Catalysis.

The redox reactivity of organoarsenic compounds is similar, albeit less well developed, when compared to organophosphorus congeners, as might be expected by the similar valence orbital IEs of P and As (see Fig. 2B). For instance, As(III) molecules similarly undergo oxidation to As(V) with mild oxidants,^287,288 and arsonium ylides can be generated from arsonium salts^289,290 or carbene transfer^291,292 for use in Wittig-type olefination reactions. In contrast, the As(V) oxidation state is less thermodynamically stable than P(V) (see Fig. 6), such that pentacoordinate arsoranes are known to undergo reductive elimination via ligand coupling, ²⁹³ and O-atom transfer of R₃As=O + PR₃ → R₃As + O=PR₃ is both kinetically and thermodynamically accessible.^294,295

3.2.1. Catalytic Wittig Reactions.

Taking advantage of the favorable deoxygenation of arsine oxides by P(III) reagents, the first report of organoarsenic redox catalysis was published in 1989 by Shi and Huang who described a tributylarsine-catalyzed Wittig olefination of aldehydes with activated bromoalkanes (Fig. 26).²⁹⁶ Triphenylphosphite, itself not competent to drive the direct olefination reaction, serves as a terminal O-atom acceptor by deoxygenation of the arsine oxide formed by Wittig olefination. Recently, Imoto and Naka have demonstrated the ability of an arsolane to efficiently catalyze similar transformations by As^III/As^V=O cycling with a hydrosilane reductant at 100 °C.²⁹⁷

Figure 26.

Organoarsine-catalyzed Wittig reaction employing triphenylphosphite as stoichiometric O-atom acceptor.

A second approach to arsine-catalyzed Wittig reactions involves Fe-porphyrin-catalyzed carbenoid transfer to generate the requisite arsenic ylide, as demonstrated by Tang (Fig. 27).^298,299 In an initial report from 2007, triphenylarsine (As2) catalyzes the olefination of aldehydes with ethyl diazoacetate in the presence of an Fe-porphyrin catalyst, where sodium dithionite is the terminal reductant enabling turnover at As. In a follow-up study in 2012, the arsine catalyst is immobilized on a polymer support to enable olefination of aldehydes and ketones with use of a soluble hydrosilane reductant, PMHS, at 110 °C to enable redox cycling of the arsine catalyst. Taken together, these reports demonstrate the utility of organoarsenic compounds in the catalytic generation of arsonium ylides for olefination and the propensity for reduction of the catalytic arsine oxides intermediates. However, concerns about toxicity and stability of the organoarsenic compounds have limited the utility of such transformations, especially as new strategies for facile turnover of phosphine oxides have emerged (see Sect 3.1.1). It remains to be seen whether there are any transformations unique to organoarsenic redox catalysis that would overcome the perceived barriers to use of As in synthesis.

Figure 27.

Organoarsine-catalyzed Wittig-type olefination of aldehydes with diazo compounds, with Fe-porphyrin co-catalyst to facilitate carbene transfer. TCP = tetra(para-chlorophenyl)porphyrinate.

3.3. Organoantimony Redox Catalysis.

As compared to both P and As, the chemistry of organoantimony compounds is distinguished by the less reducing nature of the Sb(III) oxidation state and more oxidizing nature of the Sb(V) oxidation state.³⁰⁰ As such, whereas oxidation of Sb(III) species can be accomplished by reaction with strong oxidants such as bromine, peroxides, o-quinones, and iodoso compounds, stibines do not typically undergo quaternization with alkyl halides or Michael acceptors.³⁰¹ Conversely, the lower stability of the Sb(V) compounds results in enhanced oxidizing power in relation to the lighter pnictogens, as depicted in Fig. 9. Consequently, oxidative transformations of substrates, such as alcohol oxidation, have been described using Sb(V) compounds.³⁰² These stoichiometric reactions have been translated to a limited set of organoantimony-catalyzed methods.

3.3.1. Catalytic Oxidation Reactions.

Organoantimony redox catalysis is characterized by a conspicuous opportunity for further development. At present, only two publications have appeared in this area, each of which describes an identical overall transformation under slightly modified conditions, depicted in Figure 28. In 1982, Akiba translated a stoichiometric triphenylantimony dibromide-mediated oxidation of α-hydroxyketones to α-diketones into a catalytic protocol, employing as little as 10 mol% of the Sb(V) catalyst.³⁰³ Upon single turnover, the resultant reducing Ph₃Sb (Sb2) can be oxidized by the exogenous bromine surrogate 2,3-dibromo-3-phenylpropionate to regenerate the oxidizing Sb(V) dibromide (Sb1), turning over the cycle. 20 years later, Kurita described a more practical implementation, in which 10 mol% triphenylstibine (Sb2) is used directly as catalyst under aerobic oxidation conditions to effect the same transformation in nearly quantitative yield.³⁰⁴

Figure 28.

Benzoin oxidation via organoantimony redox catalysis.

In contrast to this mild, efficient reaction with SbPh₃, the use of stoichiometric PPh₃ or BiPh₃ both provide no benzil product (81), owing to chemical inertness of the P(V) and Bi(III) states, respectively. In fact, this catalytic oxidation represents the microscopic reverse of well-established P(III)-mediated 1,2-dicarbonyl reduction by Kukhtin-Ramirez addition.^305-308 Further, reaction employing AsPh₃ (As2) under air is sluggish and poorly efficient, demonstrating the varied reactivity of congeneric organopnictogens, which are each best suited to particular applications. However, this approach to catalytic alcohol oxidation via organoantimony catalysis has never been extended beyond these activated α-hydroxyketones.

3.4. Organobismuth Redox Catalysis.

The redox chemistry of the Bi^III/Bi^V couple is dominated by the manifestation of the inert pair effect.^{98-100,157,158} Owing to the poor spatial and energetic overlap of Bi valence s and p orbitals,^{108-109,115-116} with drastic relativistic effects of the heavy atom nucleus, only very strong oxidants can convert a Bi(III) center to Bi(V); accordingly BiCl₃ does not yield BiCl₅ upon exposure to chlorine.³⁰⁹ However, Bi(V) species, such as Ph₃Bi(OAc)₂, are accessible through oxidation with peroxides, for example, and have been used extensively as strong oxidants, such as in alcohol oxidation, olefin oxidation, and oxidative cleavage of diols.^310,311 Further, the strongly oxidizing nature of Bi(V) centers has resulted in the development of ligand coupling reactions utilizing triaryl Bi(V) reagents, e.g. in the arylation of phenols.^44-46 Recently, these principles have been applied by Ball to programmed, stoichiometric o-arylation of phenols by arylboronic acids via the intermediacy of triaryl Bi(V) species.³¹²

The Bi^I/Bi^III couple has been much less studied in the context of organopnictogen chemistry, as only recently have discrete redox events in this manifold been explored. Of particular note is the seminal work of Dostál, who has demonstrated that Lewis base-stabilized aryl-Bi(III) dihydrides undergo facile release of H₂ to generate the corresponding aryl-Bi(I) compounds,^313,314 which are then amenable to oxidative addition to deliver Bi(III) species.^315-317 Bismuth(III) alkoxides also undergo Bi─O homolysis in certain cases,^318-319 a potentially relevant step in the SOHIO ammoxidation process for the synthesis of acrylonitrile from propylene.^320-322 These rare examples represent the early stages of accessing low-valent organobismuth centers to enable redox events and have begun to find application in catalysis.

3.4.1. Catalytic Oxidation Reactions.

Much of the pioneering synthetic method development using organobismuth molecules can be attributed to Barton and coworkers. Indeed, the very first demonstration of any organopnictogen exhibiting redox catalysis was reported by Barton and Motherwell in 1981 (Fig. 29),¹⁶⁶ in which triphenylbismuth (Bi1) catalyzes oxidative cleavage of α-glycols using a stoichiometric oxidant such as tert-butyl hydrogen peroxide (TBHP) or N-bromosuccinimide (NBS). This discovery was predicated upon the observation that, in the stoichiometric variant using triphenylbismuth carbonate as the oxidant, quantitative conversion to triphenylbismuth (Bi1) is observed. As such, simply by slow addition of an exogenous oxidant to regenerate a Bi(V) species, catalytic turnover can be achieved with catalyst loadings as low as 1%. Similar reactivity of both cis- and trans-decalin-9,10-diols suggests an open intermediate enabling the oxidative cleavage, as opposed to a cyclic intermediate as has been invoked for Criegee, Malaprade, and related oxidations.³²³ Here, it is relevant to note the difference in reactivity as exhibited in the SbPh₃-catalyzed oxidation of benzoins as described in section 3.3.1, which is limited to more activated substrates.³⁰⁴

Figure 29.

BiPh₃-catalyzed α-glycol cleavage via Bi(V) oxidation. NBS = N-bromosuccinimide.

Postel and Duñach later described a series of oxidative cleavage reactions catalyzed by Bi(III) mandelate, under molecular oxygen in DMSO.^324,325 Here, epoxides can be oxidized in situ to α-diketones, which are further oxidatively converted to two equivalents of carboxylic acid. Related reactions point to a dual Lewis acidic and redox role for Bi(III) in these reactions.^326-329 Other Bi(III)-catalyzed oxidation reactions, including benzylic and allylic hydroxylation with TBHP, have been reported; however, mechanistic evidence is not supportive of a Bi-redox cycle.^330-332

3.4.2. One-electron redox via open shell intermediates.

The first demonstration of radical-mediated organobismuth catalytic reactivity was described by Coles in the context of oxidative coupling of TEMPO with phenylsilane with release of H₂ (Fig. 30).³³³ In this reaction, the isolable Bi(II) radical catalyst Bi2 can reversibly bond to TEMPO (85) to generate metastable Bi(III)-TEMPOxide Bi2a, which is proposed to undergo metathesis with a Si─H bond, generating the TEMPO─Si bonded product and Bi(III)-hydride Bi2b. This species was previously shown stoichiometrically to undergo oxidative loss of hydrogen, thus regenerating Bi(III) catalyst Bi2.^318,319 Similar catalytic reactivity was recently demonstrated by Lichtenberg using a diaryl(bismuth)thiolate catalyst under UV irradiation.³³⁴

Figure 30.

Bi^II/Bi^III-catalyzed dehydrogenative O–Si bond formation. Dipp = 2,6-di-iso-propylphenyl.

3.4.2. Catalytic Cross-Coupling.

While Bi(III) and Bi(V) reagents have been used as organometallic nucleophiles and electrophiles, respectively, in transition metal-catalyzed cross couplings for more than 20 years,³³⁵ Bi-catalyzed redox cross-coupling reactions have only recently been reported. A transition metal-like cross-coupling reaction catalyzed by two electron-processes at a Bi center was described by Cornella in 2020 (Fig. 31).³³⁶ In this work, tethered Lewis base-supported Bi(III)-bismuthane catalyst Bi3 undergoes transmetalation with an aryl boronic ester to generate triarylbismuthane Bi3a. Then, oxidation by strongly oxidizing fluoropyridinium reagent 91 yields Bi(V) species Bi3b, which is stabilized by the pendant Lewis basic sulfoximine. Finally, reductive elimination forges the new C─F bond of product 88 and regenerate Bi(III) species Bi3, turning over the cycle. This chemistry takes advantage of a tethered biaryl sulfoximine ligand framework on Bi to both stabilize highly oxidizing Bi(V) intermediates with the pendant Lewis base and yield selective ligand coupling of the exocyclic aryl ligand with the apical fluoride substituent. As described in a follow-up report, perfluoroalkyl sulfonate salts are successfully coupled using bis-CF₃ bismuthane Bi4 bearing a sulfone tether to provide aryl triflate and nonaflate products.³³⁷ In this catalytic platform, rational ligand design to optimize geometric and electronic properties at the central pnictogen atom serve to unveil novel, transition metal-like reactivity.

Figure 31.

Bi-catalyzed fluorination and triflation of aryl boronic esters and acids, respectively. Proposed mechanism of fluorination. Tf = triflyl.

3.4.3. Catalytic Reductive Deoxygenation.

Cornella has also explored the Bi^I/Bi^III couple for catalysis in the context of transfer hydrogenation of azoarenes and nitroarenes (Fig. 32).³³⁸ Using an NCN-chelated bismuthinidene (Bi5) first described by Dostál,^313,314 an unstable Bi(III)-dihydride (Bi5a) is putatively formed by reaction with ammonia borane, in reverse analogy to the loss of H₂ from a Bi(III)-dihydride (Bi5a) originally described by Dostál. In this catalytic reaction, the putative Bi(III)-dihydride (Bi5a) intermediate delivers H₂ across either N─N or N─O π-bonds, accomplishing a transfer hydrogenation with good functional group tolerance. Mechanistic studies support the intermediacy of the Bi(III)-dihydride (Bi5a), as its protonated cation (Bi5b) can be detected by HRMS in both stoichiometric and catalytic reactions. Here, the Bi^I/Bi^III cycle is exploited to first receive an equivalent of H₂ from ammonia borane and then deliver it to an activated π substrate, similarly to earlier work carried out in the P^III/P^V couple.²⁸³ This reaction is the first demonstration of organopnictogen catalysis operating in the Pn^I/Pn^III couple, thus paving the way for low-valent organopnictogen chemistry in catalysis.

Figure 32.

Bismuthinidene-catalyzed transfer hydrogenation of azoarenes and nitroarenes to hydrazines and hydroxylamines, respectively, with ammonia borane.

Cornella further demonstrated the catalytic utility of the Bi^I/Bi^III couple of bismuthinidenes such as Bi5-Bi7 for the reductive deoxygenation of N₂O, through a distinct mechanistic pathway (Fig. 33).³³⁹ In this study, rapid deoxygenation of N₂O by Bi5 liberates N₂ and produces a dimeric [Bi₂O₂] species as detected by ESI-HRMS. Through careful tuning of the pendant imine ligands, aldimine-supported Bi6 and ketimine-supported Bi7 yield dimeric [Bi₂O₂] and monomeric [Bi─OH] scaffolds, respectively, upon exposure to N₂O, with the structures unambiguously identified by single crystal X-ray crystallography. These results seemingly indicate an unstable, basic Bi^III-oxide intermediate derived from O-atom transfer from N₂O. Both aforementioned oxide-derived adducts yield parent compounds Bi6 and Bi7, along with HOBpin and O(Bpin)₂, upon exposure to HBpin at ambient temperature. Accordingly, catalytic reduction of N₂O is feasible using Bi5, Bi6, and Bi7, with Bi5 delivering the most rapid and efficient conversion, even at catalyst loadings as low as 0.1 mol%. This demonstration of Bi^I/Bi^III catalysis combines the reducing nature of the Bi(I) state with a facile reduction of a Bi(III) oxide equivalent to enable redox cycling at ambient temperature, evocative of the body of work in P^III/P^V=O redox catalysis.

Figure 33.

Bismuthinidene-catalyzed reductive deoxygenation of N₂O via the intermediacy of Bi(III)-oxide equivalents with pinacolborane. m-Tp = meta-terphenyl.

4. Outlook

The quickening pace of progress in organopnictogen redox catalysis within the past fifteen years assures the continued vibrancy of this exciting area of research in the years to come. Looking ahead, we anticipate significant opportunities for ongoing discovery across a broad scientific front, including:

• Designing Catalysts with Improved Redox Leveling. A greater mastery over precision redox tuning will be needed to enable catalysis with greater speed (turnover frequency) and greater durability (turnover number). An appreciation for the connection between catalyst composition/structure and redox driving force of elementary reaction steps will be a necessary initial step in this quest, but a further attentiveness to round-trip thermodynamics will also be needed for catalysis. Detailed mechanistic and thermochemical studies that identify kinetic bottlenecks and parasitic branching points will be essential to inform new catalyst designs that enable faster turnover at milder conditions with lower catalyst loading. In the limit, such a high level of redox mastery would enable the reversible use of a given Pnⁿ/Pnⁿ⁺² couple specified only by the reaction thermodynamics of the stoichiometric inputs.

• Taming Underexplored Two-Electron Redox Couples for Catalysis. Although periodic trends shape the innate driving forces for two-electron redox events at pnictogens (Sect. 2), novel design of organopnictogen compounds might open space for catalytic cycles operating within ‘atypical’ redox couples. As exemplified in Sect. 3, the Pn^III/Pn^V couple has been widely employed in catalysis; by contrast, the lower-valent Pn^I/Pn^III couple is still comparatively underdeveloped. Seminal work on the chemistry of Pn(I) centers have demonstrated their ability to achieve challenging bond activation reactions.^340,341 Although the generation of low-valent Pn(I) species under mild conditions poses the most immediate barrier to expansion of organopnictogen redox catalysis to the Pn^I/Pn^III couple, pioneering work from Cornella in Bi^I/Bi^III catalysis ^338,339 establishes feasibility and points to further opportunities. By expanding the accessibility of diverse redox states—presumably through ligand design and substituent effects—new channels of reactivity might be made available.

• Controlling Stereochemistry in Organopnictogen Redox Catalysis. The pioneering achievements of Werner,^209,210 Voituriez,²¹⁹ and Kwon²³⁴ (Sect. 3.1) establish the viability of stereochemical control within organopnictogen redox catalysis; however, new chiral organopnictogen catalysts will be needed to advance beyond these initial discoveries. For instance, it remains to be seen whether deliberate incorporation of ‘secondary sphere’ interactions can be leveraged to effect stereochemical discrimination.^33,342-346 The opportunities and/or complexities associated with stereogenic pnictogen chirality centers and their stereochemical fluxionality—especially in pentacoordination (i.e. polytopal isomerism)^141,142—have not yet been explored in a systematic fashion. Indeed, given challenges presented by the varying coordination numbers, geometries and valence electron counts encountered in organopnictogen redox catalysis, the emergence of new heuristics of asymmetric design may be needed.

• Merging Organopnictogen Redox with Established Catalytic Modes. The merger of organopnictogen redox catalysis with other enabling modes of catalysis (organocatalysis, transition metal catalysis, Bronsted acid/base catalysis, inter alia) could lead to the development of further powerful classes of reactions. Such catalytic cycles could be envisioned to work in tandem, cascade,^347-351 or synergistic modes,³⁵² owing to the mutual compatibility of each catalytic mode of molecular transformation. Such mergers could make use of the distinct reactivities inherent to the aforementioned platforms and create opportunities for unveiling novel transformations.

• Embracing One-electron Open-shell Reactivity. Stable covalent bonds are (mostly) two-electron constructs, but their catalytic synthesis by stepwise one-electron processes presents potentially enabling reaction channels.³⁵³-³⁵⁹ Open-shell reactivity within organopnictogen catalysis is therefore ripe for development. Organopnictogen radicals and radical ions are well-known entities,^360,361 whose reactivity can be triggered by photochemistry^362-367 or electrochemistry.^368,369 Among possible scenarios, single-electron oxidation or reduction of catalytic intermediates^370-377 could unveil new reaction pathways, including selective bond activation or challenging atom transfer processes.^378-382 Alternatively, single-electron pathways could be accessed to facilitate otherwise sluggish catalytic turnover, such as electrocatalytic reduction of phosphine oxides.^383-386

• Beyond Homogenous Organic Reaction Media. The development of catalytic systems that can operate in nonorganic media will be a necessity to realize a broader potential of group 15 redox catalysis in contexts beyond organic synthesis. Noting the prevalence of organopnictogen redox chemistry in chemical biology in the form of the Staudinger ligation,^34-39 the development of water-compatible reaction systems presents an appealing challenge to the growth of the field of organopnictogen redox catalysis.³⁸⁷ Indeed, recent work utilizing P(V) chemistry to selectively label serine residues³⁸⁸ and Bi(V) chemistry to arylate phenols³¹² point to the potential of Pn(V) to enable selective bond-forming reactions. Alternatively, an adaptation of the design principles for homogeneous group 15 redox catalysis to heterogeneous catalyst development similarly presents untold prospects for discovery.

5. Concluding Remarks

To close, we return to the question posed at the end of the Abstract: “What new catalytic manifolds can be developed through creative catalyst and reaction design that take advantage of the intrinsic redox reactivity of the pnictogens to drive new discoveries in catalysis?”³⁸⁹ This is a critical question, and though a detailed answer may not be knowable except in hindsight, the contours of a reply surely can be traced in outline. Organopnictogen redox catalysis is a relatively young entrant to the science of catalysis presently populated by numerous highly successful catalytic modalities, each a towering achievement. Within this crowded context, organopnictogen redox catalysis must aspire to more than mimicry of existing techniques; it must express something authentic and inimitable. On this front, it seems likely that the most compelling opportunities presented by this emerging field—those that will maximize scientific and practical impact—will be realized through the discovery of new bond (dis)connections or functional group interconversions that are truly native to organopnictogen redox catalysis. We assert that the periodic trends—both within Group 15, and between Group 15 and others in the p-block—impart the pnictogens generally, and each of the pnictogen elements individually, with distinctive properties, providing a varied palette of components for catalyst design and reaction development. The diversity of characteristics in Group 15 position organopnictogen redox catalysis to achieve unique reaction classes that are without direct precedent or complement in the armory of catalytic synthesis. Along this trajectory, the progress achieved thus far in organopnictogen redox catalysis is but a tantalizing preamble to a future of ongoing discovery.