Focusing on the Catalysts of the Pd- and Ni-Catalyzed Hirao Reactions

The Hirao reaction involving the phosphinoylation or phosphonation of aryl halides by >P(O)H reagents is a P–C bond forming transformation belonging to the recently very hot topic of cross-couplings. The Pd- or Ni-catalyzed variations take place via the usual cycle including oxidative addition, ligand exchange, and reductive elimination. However, according to the literature, the nature of the transition metal catalysts is not unambiguous. In this feature article, the catalysts described for the Pd(OAc)2-promoted cases are summarized, and it is concluded that the “(HOY2P)2Pd(0)” species (Y = aryl, alkoxy) is the real catalyst. In our model, the excess of the >P(O)H reagent served as the P-ligand. During the less studied Ni(II)-catalyzed instances the “(HOY2P)(−OY2P)Ni(II)Cl−” form was found to enter the catalytic cycle. The newest conclusions involving the exact structure of the catalysts, and the mechanism for their formation explored by us were supported by our earlier experimental data and theoretical calculations.

Montchamp and co-workers applied solvent additives (e.g., ethylene glycol or dimethoxyethane) to facilitate the P-C coupling of H-phosphinates [15,52,53]. According to them, on the one hand, the co-solvent promotes the ligand exchange by taking part in the conversion of the pentavalent Y 2 P(O)H to the Y 2 POH tautomer, on the other hand, it stabilizes the Pd-catalyst. The rate of the last step is also influenced by the steric and electronic properties of the ligands involved in complex III. A cis geometry of the aryl group and the P-moiety is required for the reductive elimination to occur. Unlike other C-heteroatom cross-coupling reactions, the elimination of the P-C coupled product may be promoted by electron-donating substituents [54]. In case of bidentate ligands, the so-called "bite angle" may also affect the reaction rate: a larger angle induce faster elimination [55].
Kalek and Stawinski found that the addition of ionic additives, mainly acetates, had a positive effect on the course of the P-C couplings regarding reaction time [13,16,50]. According to the studies, acetate ions play role in all stages of the catalytic cycle: a more active catalyst complex may be formed (see later) [13], and the presence of ions accelerates the ligand exchange and the reductive elimination as well [50].

Nickel-Catalyzed P-C Coupling Reactions
Nickel catalysis is also often applied in P-C cross coupling reactions. On the one hand, reductive procedures involving Ni(II) salts together with Zn or Mg as the reductant were described [23,24,56], or Ni(0)(cod) 2 was applied as the catalyst precursor [25,26,[57][58][59], on the other hand, the Ni salts were used without reductive agents [60][61][62][63][64]. It is worth noting that in the above reactions, especially in the reductive variations, 2,2 -bipyridine was applied as the ligand. It was assumed that Ni(II) is reduced to Ni(0), and the bipyridine complex of the reduced form takes part in the oxidative addition of the aryl or other halide. This is followed by ligand exchange including the entry of the >P(O)H reagent. The last step is the reductive elimination [65]. Hence, regarding the redox background, a similar process was substantiated as for the Pd-case. However, no convincing evidence was presented for the Ni(0) → Ni(II) supposition at the start of the cycle.
Moreover, the way how the P-C couplings applying Ni(II) salts take place in the absence of reductants has not been investigated for long [60][61][62][63][64].
As it was already mentioned, the Keglevich group developed a MW-assisted version of the P-C coupling, and it was possible to apply NiCl 2 instead of Pd(OAc) 2 . In these cases, there was no need to apply the usual P-ligands, instead, the excess of the >P(O)H reagent served as the ligand [47].

General Considerations
Let us survey the possible Pd complexes that were considered as catalysts in the Hirao reaction ( Figure 2). All agreed on that regarding the palladium salt precursors, Pd(OAc) 2 is the best choice due to its lipophilicity [10,12]. Although this salt consists of trimers or even polymers, its structure is generally represented as Pd(OAc) 2 . Buono et al. assumed the formation of the [{H(OPh 2 P) 2 }PdOAc] 2 complex (1) in the interaction of Pd(OAc) 2 and diphenylphosphine oxide [66]. The investigations of Amatore and Jutand on C-C coupling reactions showed that the interaction of Pd(OAc) 2 and triphenylphosphine leads to (Ph 3 P) 2 Pd(OAc) 2 complex, that is then converted to species 2 by reduction and the departure of an AcOion [67]. Kalek and Stawinski also proposed (Ph 3 P) 2 PdAcOcomplex 2 as an active species in the P-C couplings involving Pd(OAc) 2 as the catalyst precursor and PPh 3 as the ligand [13]. At the same time, complex 3 was considered by them as an inactive species in the catalytic cycle [13], regardless that Ackermann et al. supposed a similar species (the deprotonated version of 4) as an active catalyst in another kind of reaction, in the intermolecular α-arylation of 2-acylamino-chloroarenes to provide different oxindoles [68]. We suggested the simple "PdP 2 " type complex (4) as the real catalyst in the Hirao reaction on the basis of our experiments and theoretical calculations [44].

Formation of the PdP 2 Catalyst-Theoretical and Experimental Results
We were the first who studied the formation of Pd complex 4 from Pd(OAc) 2 and the excess of the Y 2 P(O)H reagent. Our observation was that in the "P-ligand-free" Hirao reaction, 1.3 equivalents of the >P(O)H reagent should be used if the Pd(OAc) 2 catalyst is applied in a quantity of 10 mol% (three equivalents are needed for the catalyst, 1 equivalent as the reducing agent, two equivalents as the P-ligand) [42,43]. Complex [HO(EtO) 2 P] 2 Pd (4, Y = EtO) was prepared by us in a separate experiment reacting Pd(OAc) 2 with 1.3 equivalents of (EtO) 2 P(O)H at 120 • C for 30 min. 31 P NMR shift of the Pd complex (4, Y = EtO) was in agreement with that reported by Kalek et al. [13]. This complex was then tested in the reaction of (EtO) 2 P(O)H with PhBr carried out as elaborated by us [44]. Diethyl phenylphosphonate was obtained in a yield of 72% proving that the in situ formed catalyst is identical to species 4 (Y = EtO). It was also proved by us [44] that the interaction of Pd(OAc) 2 and diphenylphosphine oxide furnishes a complex [(HO)Ph 2 P] 2 Pd (4, Y = Ph). Pd(II) was reduced to Pd(0), while the oxidation is manifested in by-product 5. The formation of this species along with the energetics are shown in Scheme 2 [69]. The mechanism of the catalyst formation was evaluated by theoretical calculations [69]. According to the general consideration and our earlier findings [44], the active oxidation state of Pd was assumed to be zero. As it was mentioned above, diphenylphosphine oxide serves as the reductive agent. The most probable reaction mechanism is presented in Scheme 3. The reaction of Pd(OAc) 2 with two molecules of Ph 2 P(O)H affords to Pd(II) complex 6, the deprotonation of which leads to species 7. Then, the acetate anion of the "PdOAc" moiety is replaced by a third Ph 2 P(O)H molecule affording complex 8. Then, an intermolecular O-insertion into the P-Pd bond of intermediate 8 gives species 9. In the next step, an acetate anion is connected to the P atom of the "Ph 2 POPd" moiety of complex 9 to furnish species 10. A reductive elimination from intermediate 10 via TS 11 provides complex 12. A final stabilization of species 12 by the elimination of Ph 2 P(O)OH and AcO − , and by the simultaneous incorporation of the Ph 2 POH species formed by regeneration via hydrolysis of acyloxyphosphine by-product results in the formation of complex 4 (Y = Ph) that is the active catalyst [44].  The comparison of the reactivity of (EtO) 2 P(O)H and Ph 2 P(O)H is not easy, as there are three independent and consecutive reaction steps: reduction of the Pd(II) to Pd(0) (1), ligation of the Pd(0) so formed (2), and the intrinsic P-C coupling itself (3). As regards to the gross reactivity, (EtO) 2 P(O)H seems to be more reactive than Ph 2 P(O)H, if the reaction times required at 120 • C are compared. To be able to decide on the reactivity of the >P(O)H species in the P-C couplings, concurrent reactions comprising 0.5 equivalents of both Ph 2 P(O)H and (EtO) 2 P(O)H to one equivalent of bromobenzene were performed in the presence of 10 mol% of Pd(PPh 3 ) 4 at 120 • C, and interrupted before completion that suggested that Ph 2 P(O)H is slightly more reactive, than (EtO) 2 P(O)H in the P-C coupling reaction itself (Scheme 5) [44]. Returning to the idea of Buono et al. they claimed the formation of another kind of complex from the interaction of Pd(OAc) 2 with a series of secondary phosphine oxides applied in a two equivalents quantity. The outcome is shown on the example of Ph 2 P(O)H as the P-reagent (Scheme 6) [66]. However, the structure of complex 1 was not confirmed. We believe that in the latter case not complex 1, but the [(HO)Ph 2 P] 2 Pd complex (4, Y = Ph) was formed.
Comparative theoretical calculations suggested that while complex 15 is not a stable species (Scheme 7), the dimer-like complex with two chloride anions (16) may exist (Scheme 8).

A Theoretical Study on the Ligation of Pd(0)
In our calculations, the complexation of Pd(0) with diarylphosphine oxides with Ph-, 3,5-Me 2 -C 6 H 3 and 2-Me-C 6 H 4 substituents and triphenylphosphine was investigated (Scheme 9) [69]. It was found that the mono-(17) and bis-ligated Pd(0) complexes (18) may be formed in exothermic reactions (in STEP 1 and STEP 2, respectively) for all the cases investigated, as shown by delta enthalpy (∆H) values of −150.5-90.6 kJ mol −1 and −91.3-27.7 kJ mol −1 , respectively. The formation of tri-coordinated complex 19 from di-coordinated 18 is not as favorable with Ph 2 POH and (3,5-Me 2 -C 6 H 3 ) 2 POH (∆H = 4.8 and 18.1 kJ mol −1 , respectively) as the bis-ligation. And what is more, with (2-Me-C 6 H 4 ) 2 POH and PPh 3 , as a consequence of steric hindrance, the tris-ligation (STEP 3) becomes endothermic (∆H = 51.3 and 53.9 kJ mol −1 , respectively). In case of the (2-MePh) 2 POH and PPh 3 ligands, the optimum is at bis-ligation, while with Ph 2 POH and 3,5-Me 2 -C 6 H 3 , the tris-ligation is optimal. The situation regarding enthalpies is rather similar for the tetra-ligation (20) [69]. The complex forming ability of the (2-Me-C 6 H 4 ) 2 POH species is comparable with that of PPh 3 . It can be also said that while forms 18 and 19 are real complexes, 20 is dissociated variation. It is noted that the complexation status of a catalyst is a crucial in the reaction it is involved in. An "under-complexed" Pd is not active and not stabilized, hence, it may decompose and precipitate. At the same time, an overcrowded complex may not be active enough. From among the complexes discussed, 18 may be the best catalyst, 19 and 20 are hindered forms. It is noteworthy that Pd(PPh 3 ) 4 may also exist in the Pd(PPh 3 ) 3 + PPh 3 form [70].  The beneficial effect of the 2-Me or 3-Me group in the phenyl ring was concluded. To get further information, 0.15 equivalents of (2-Me-C 6 H 4 ) 2 P(O)H was applied as the P-ligand in the Hirao reaction of Ph 2 P(O)H with bromobenzene that was carried out otherwise as above. As compared to the case, where Ph 2 P(O)H held the role of both the reagent and the preligand (Scheme 10, entries 1 and 2), the P-C coupling using (2-Me-C 6 H 4 ) 2 P(O)H as the catalyst ligand was significantly faster, as a complete conversion required 30 min (Scheme 10, entry 7). The selectivity of the "mixed" Hirao reaction was surprisingly good, only 5% by-product (2-Me-C 6  substituent has no effect on the course of the reaction. The results can be interpreted by assuming that the beneficial effect of (2-Me-C 6 H 4 ) 2 P(O)H and (3,5-Me 2 -C 6 H 3 ) 2 P(O)H is to promote the formation of a more active catalyst incorporating only two Ar 2 POH ligands. This is the consequence of steric hindrance due to the 2-Me group and 3-Me substituent, respectively. The activity of the Pd complexes with the P-ligands under discussion shows the following trend in respect of the P-species:

Theoretical and Experimental Studies on the Ni Catalyst
As it was mentioned in the Introduction (Section 1.2) not much is known about the exact nature of the Ni catalyst involved in the Hirao reaction. We investigated the route of the MW-assisted P-C coupling reaction of PhBr with diphenylphosphine oxide or diethyl phosphite applying NiCl 2 as the catalyst precursor. In our case, the excess of the >P(O)H reagent served as the P-ligand. Preparative experiments suggested that two equivalents of the P-reagent to the NiCl 2 was the optimum quantity [47,71]. Then, theoretical calculations suggested the following mechanism for the formation of the active catalyst entering in the Hirao reaction (Scheme 11) [71].  (25) is developed after a consecutive deprotonation and the loss of a chloride anion, involving species 23 and 24, respectively. This complex (25) will then participate in the oxidative addition with the aryl halide. It follows that, as a result, Ni(II) gets in the Ni(IV) state. This was indeed a fully surprising experience, but quantum chemical calculations confirmed this [71]. It was another surprise that in the "reductive" variation, again a Ni(II) → Ni(IV) oxidation was substantiated by the preparative experiments and calculations [72]. However, surveying the literature, the Ni(II) → Ni(IV) transition is well-known during the interconversion of different Ni complexes [73][74][75]. Moreover, a Ni(II) → Ni(IV) conversion was observed in the Ni-catalyzed alkylation of benzamides with alkyl halides [76]. Scheme 11. Mechanism for the formation of the active catalyst from NiCl 2 and diphenylphosphine oxide or diethyl phosphite calculated by the B3LYP/6-31G(d,p)//PCM(MeCN) method.

A Theoretical Study on the Ligation of Ni(II)
The consecutive complexation of NiCl 2 with the trivalent tautomeric form (Y 2 POH) of diphenylphosphine oxide (Y = Ph) and diethyl phosphite (Y = EtO) was evaluated by quantum chemical calculations (Scheme 12). The endothermic tautomerization of Y 2 P(O)H to Y 2 POH (if Y = Ph, ∆H = 22.9 kJ mol −1 , while if Y = EtO, ∆H = 25.1 kJ mol −1 ) [77] was also taken into consideration.

Conclusions
Considering different variations, our experimental data and the theoretical calculations suggested that in the cases when Pd(OAc) 2 was used as the catalyst precursor, "P 2 Pd(0)" was the initial species entering the catalytic cycle. Applying the excess of the Y 2 P(O)H reagent as the ligand, the catalyst is HOY 2 P-Pd(0)-PY 2 OH. Hence, considering that Y 2 P(O)H also serves as the reducing agent, three equivalents of the P-reagent to the Pd(OAc) 2 precursor is needed. At the same time, when NiX 2 is the catalyst precursor, no matter if there is a reducing agent or not, HOY 2 P-Ni(II)Cl-PY 2 Ois the active catalyst. The involvement of Ni(II) instead of Ni(0) in the catalytic cycle is a brand new and surprising observation. Utilizing quantum chemical calculations mechanistic protocols were proposed for the formation of the "P 2 Pd(0)" and the "P 2 Ni(II)XccH" catalysts. Ligation studies also supported our findings.