Synthesis, Structure and Nickel Carbonyl Complexes of Dialkylterphenyl Phosphines

FULL PAPER 1 Synthesis, Structure and Nickel Carbonyl Complexes of Dialkylterphenyl Phosphines Mario Marín,[b] Juan J. Moreno,[a] Carlos Navarro-Gilabert,[a] Eleuterio Álvarez,[a] Celia Maya,[a] Riccardo Peloso,[a] M. Carmen Nicasio,*[b] Ernesto Carmona*[a] Dedicated to Professor William D. Jones on the occasion of his 65th birthday Abstract: We describe the experimental and computational characterization of a series of dialkylterphenyl phosphines, PR2Ar’ (Figure 2). Molecules of composition PR2ArDtbp2, for R = Me, Et, iPr, c-C5H9 and c-C6H11, are accompanied by five PR2Ar’ phosphines containing the bulky alkyl groups iPr, c-C5H9 or c- C6H11, in combination with Ar’ = ArXyl2, ArXyl’2, or ArPh2 (L1-L10). Steric and electronic parameters have been determined computationally and from IR and X-ray data obtained for the phosphines and for some derivatives, including tricarbonyl and dicarbonyl nickel complexes, Ni(CO)3(PR2Ar’) and Ni(CO)2(PR2Ar’). In the solid state, molecules of PR2Ar’ adopt one of the three possible structures represented in Figure 3, formally related by rotation around the Cipso-P bond. Information on their relative energies and on the influence of the free phosphine structure on its coordination chemistry towards Ni(CO)n (n = 2, 3) fragments has been gained by experimental and computational methods. Introduction Tertiary phosphines, PR3, and related molecules containing a three-coordinate phosphorus atom, PX3 (X = OR, NR2 and others), form an invaluable family of ligands, with an influence in organometallic chemistry and homogeneous catalysis that cannot be overstated.1-3 By and large, during the second half of the past century, studies on transition metal complexes focused on alkyl and aryl phosphines such as PMe3, PiPr3, P(c-C6H11)3, PPh3 and others. Notwithstanding the versatility of these ligands, relatively little progress was made on the synthesis of monodentate phosphines until fairly recently.4 Some twenty years ago, Buchwald and coworkers demonstrated that when bound to palladium, dialkylbiaryl phosphines were able to catalyse with high proficiency C-C and C-N coupling reactions of aryl halides, including unactivated aryl chlorides.5 Lately, Buchwald’s phosphines,6 along with other bulky phosphines,7-9 have become indispensable ligands in organometallic chemistry and catalysis, and within the recent past many tertiary phosphines of assorted electronic and steric characteristics have been made available.10- 15 As of late, remarkable success in unveiling subtle phosphine ligand effects has been witnessed in nickel, palladium and gold catalysis.16 Besides, for tri(1-adamantyl)phosphine, a molecule for which steric and electronic properties beyond conventional limits were disclosed, unique catalytic features, seemingly influenced by van der Waals dispersion forces,17 were uncovered.12a Furthermore, widely used parameters such as Tolman’s cone angle 18 and the buried volume %Vbur19 may be envisaged as accounting, respectively, for remote and proximal steric effects, the former potentially giving rise to significant dispersion forces.16a The impact of dialkylbiaryl phosphines in organotransition metal chemistry and catalysis prompted us to explore the mterphenyl analogues, PR2Ar’, for Ar’ = terphenyl radical. The conspicuous features of sterically demanding terphenyl fragments, either as bulky organometallic aryl ligands20 or as substituents of various types of Lewis bases21,22 have been investigated. At the outset of our work, however, just a few terphenyl phosphine ligands, and a consequently small number of their complexes with late transition metals, had been described.23-25 Our first reports centred Rh, Ir, Pt and Au of PMe2Ar’ phosphines (Ar’ = C6H3-2,6- (C6H3-2,6-R2)2 for R = Me and iPr, in shorthand notation PMe2ArXyl2 and PMe2ArDipp2, respectively), and evinced their aptitude to stabilize low-coordinate structures, as well as their potential to adopt different coordination modes (Figure 1), where P-bonding is complemented by relatively weak M···Carene interactions with a flanking aryl ring of the terphenyl substituent.23- 28 Figure 1. Different coordination modes found for terphenyl phosphines.23-28 The objective of the present work was the description of a series of PR2Ar’ molecules selected in accordance with the [a] J. J. Moreno, C. Navarro-Gilabert, Dr. Eleuterio Álvarez, Dr. Celia Maya, Dr. R. Peloso, Prof. Dr. E. Carmona Instituto de Investigaciones Químicas (IIQ), Departamento de Química Inorgánica and Centro de Innovación en Química Avanzada (ORFEO-CINQA), Consejo Superior de Investigaciones Científicas (CSIC) and Universidad de Sevilla Avda. Américo Vespucio 49, 41092 Sevilla, Spain E-mail: gu[email protected] [b] M. Marín, Prof. Dr. M. C. Nicasio Departamento de Química Inorgánica, Universidad de Sevilla Aptdo 1203, 41071 Sevilla (Spain) E-mail: mnica[email protected]s Supporting information for this article is given via a link at the end of the document. FULL PAPER 2 Figure 2. Structural formulae, abbreviated names and numbering scheme for the terphenyl phosphines described in this work (upper part) and others already in the literature (bottom row). Crystallographic data for phosphines with underlined numbers available. [a] See reference 24a. following criteria. Firstly, besides some PMe2-containing phosphines,28,29 we set out to prepare PR2Ar’ phosphines of the branched and cyclic alkyl groups iPr and c-C6H11, respectively, which are phosphine substituents extensively employed in organometallic chemistry and catalysis. The new terphenyl phosphines find applications in efficient Pd-catalyzed amination reactions and Cu-promoted cycloaddition of azides and alkynes. These and other catalytic processes will be described separately. We also considered of interest the c-pentyl phosphine analogues,30-35 despite known drawbacks for P(c-C5H9)3 as, for instance, its facile oxidation and the higher reactivity of c-C5H9 rings relative to c-C6H11 ones, due mainly to steric strains.34 For the second group of phosphines, we chose a terphenyl moiety containing two 3,5-bis(t-butyl)phenyl substituents, 3,5-C6H3- (CMe3)2, at the 2- and 6-positions of the central aryl ring (in short ArDtbp2). As presented in Figure 2, a total of five PR2ArDtbp2 ligands have been prepared for R = Me, Et, iPr, c-C5H9 and c-C6H11, (L1- L5). Attempts to prepare their tert-butyl analogues, PtBu2Ar’, proved unsuccessful. Here, we focus attention on ligand synthesis and structural characterization by X-ray diffraction methods. We show that all dialkylterphenyl phosphines that have been authenticated by X- ray crystallography, exhibit a solid-state structure that corresponds to one of the three conformations depicted in Figure 3, for which different coordination properties towards unsaturated metal fragments, MLn, can be foreseen. Besides, we consider closely Ni(0)-CO-PR2Ar’ complexes, not only tricarbonylphosphine species, Ni(CO)3(PR2Ar’), 1·PR2Ar’, similar to those employed by Tolman to ascertain phosphine electronic properties,35 but also some Ni(CO)2(PR2Ar’) derivatives, 2·PR2Ar’, for which an uncommon oblique trigonal pyramidal geometry, with a relatively weak Ni-η2-Carene interaction has been unveiled. Although many Ni(0) tricoordinate complexes are familiar, analogous phosphine dicarbonyl Ni(0) compounds are unknown. Furthermore, we were surprised to learn that nearly fifty years after the generation by Tolman of Ni(CO)3(PR3) complexes in CH2Cl2 solutions,35 there is an astonishing dearth of crystallographic information on compounds of this type.36 Results and Discussion Synthesis and spectroscopic characterization of PR2Ar’ ligands Our recent synthesis of terphenyl phosphines, PR2Ar’,28 followed a slightly modified earlier procedure23a and consisted in the stepwise reaction of a terphenyl Grignard reagent, Mg(Ar’)X, with PCl3, followed by alkylation using, once more, the appropriate magnesium agent, Mg(R)Br, as shown in Scheme 1a. Whereas this method yields satisfactory results for linear hydrocarbyl R groups and it was thus applicable to the synthesis of the new phosphines L1 and L2, attempts to extend it to the branched or cyclic i-propyl, c-pentyl and c-hexyl phosphines led either to unreacted starting materials or to complex mixtures of products that could not be characterized (Scheme 1b). Emulating the synthesis of biaryl phophines,37 the alkylation of PX2Ar’ was FULL PAPER 3 effected in the presence of CuCl. After careful monitoring of reaction conditions (temperature, concentration of reagents and copper salt), best results were obtained performing the reaction at room temperature and using PX2Ar’:Mg(R)Br molar ratios of ca. 1:4, in the presence of overstoichiometric quantities of CuCl (approximately 1.5 equiv. relative to PX2Ar’). The formation of the target phosphines was always accompanied by small amounts of Cu(X)(PR2Ar’) complexes. Work-up of reaction mixtures was complicated by the generation of metallic copper and of sticky, insoluble materials that easily clogged filtration apparatus.37b The synthesis of the new phosphines is summarized in Scheme 2, particular details regarding their purification are presented in the Experimental Section and the Supporting Information. As stated earlier, all attemps to prepare PtBu2Ar’ phosphines were fruitless. Scheme 1. General synthesis of m-terphenyl dialkyl phosphines with linear hydrocarbyl groups. Scheme 2. Optimized synthesis of m-terphenyl dialkyl phosphines PR2Ar’ with branched or cyclic substituents. Table 1 collects relevant information on the new phosphines L1-L10. Corresponding data reported originally for L11-L16 are also included. Before discussing pertinent characterization data, trends in isolated yields deserve some brief comments. All phosphines but L16 and L9 can be produced on gram-scale runs in 50-60% yields, hence comparable to those given originally for dialkylbiaryl analogues37 and for the recently prepared PMe2Ar’ (Methyl-JohnPhos; ca. 52% three-step overall yield).29b With the exception of PEt2ArXyl2, L16, reported to form in unexpected low yields (about 40%),28 it can be noticed that sterically demanding R groups provoke a significant decrease in isolated yields, which remain in the 35-60% range for iPr, c-C5H9 and c-C6H11, in contrast with the 70-80% values of the PMe2Ar’ analogues. This may explain failure to isolate PtBu2Ar’ phosphines. It appears plausible that steric hindrance triggers side reactions involving, among others, formal hydride transfer from Mg(R)Br to PX(R)Ar’ intermediates, thereby giving rise to alkene and secondary phosphine P(H)R(Ar’) products.38 In support of this hypothesis, 31P NMR resonances around −40 ppm, that could be due to P(H)R(Ar’) molecules,39 were frequently detected in the reaction crudes. Furthermore, reactions aimed at the synthesis of PR2ArDipp2, for R = iPr, c-C5H9 and c-C6H11, did not afford the desired products, and for R = c-C5H9 the secondary phosphine PH(c-C5H9)ArDipp2, with  (31P) = −43.8 ppm, was isolated and characterized by X-ray crystallography (see the SI, Figure S9). The similar phosphine PH(iPr)ArXyl2, with (31P) = −37.7 ppm, was also isolated and characterized (see the SI). Table 1. Yields (based on P), 31P{1H} NMR chemical shifts and  C-P-C angles for the ligands listed in Figure 2. Ligand Yield (%) 31P{1H} NMR (ppm)  C-P-C (º) L1, PMe2ArDtbp2 71 −36.6 298.53 L2, PEt2ArDtbp2 65 −12.8 L3, PiPr2ArDtbp2 49 12.6 310,72 L4, P(c-C5H9)2ArDtbp2 37 0.9 309,44 L5, P(c-C6H11)2ArDtbp2 51 1.6 L6, PiPr2ArXyl2 57 16.2 315,84 L7, P(c-C5H9)2ArXyl2 57 4.6 313,84 L8, P(c-C6H11)2ArXyl2 56 10.1 316,59 L9, PiPr2ArXyl’2 53 15.3 L10, PiPr2ArPh2 58 16.3 L11, PMe2ArPh2 [a] 80 −35.8 298,63 L12, PMe2ArXyl2 [b] 60 −40.4 309,76 L13, PMe2ArMes2 [c] 65 −36.9 309,02 L14, PMe2ArDipp2 [d] 54 41.3 308,45 L15, PMe2ArTripp2 [a] 54 −40.7 L16, PEt2ArXyl2 [a] 40 −8.0 [a] See ref. 28. [b] See ref. 26a. [c] See ref. 23a. [d] See ref. 26b. The new terphenyl phosphines were obtained as analytically pure white solids that could be stored under air for extended periods of time. This is in agreement with the behaviour unveiled previously for dialkylbiaryl phosphines.29,40a Some of the PR2Ar’ phosphines exhibit solution dynamic behaviour, as discussed later for PiPr2ArDtbp2, L3. In general, however, the room temperature 1H and 13C{1H} NMR spectra recorded for L1-L10 are simple and indicative of a high degree of apparent symmetry introduced by rotation around the P-Cipso bond. 31P{1H} chemical shifts span a range of over 50 ppm (Table 1), from −36.6 (L1) to 16.3 ppm (L10). As expected, the major influence in the  value comes from the R groups.10,39 Given that, as already pointed out, the syntheses of bulky phosphines often give rise to kinetically FULL PAPER 4 competitive, undesirable side reactions, predicting the 31P NMR chemical shifts of target molecules can be helpful. With the aid of the 31P NMR contributions compiled for the R groups of L1-L10,10 the  values listed in Table 1 lead to group contributions for ArXyl2 and ArDtbp2 of approximately +4 and +1 ppm, respectively. A comparable value of +2 ppm can be estimated for ArDipp2, yet the latter is based on only two PR2ArDipp2 phosphines (Table 1). It should be remarked that for tertiary phosphines, deviations between estimated and experimental values are usually ≤3 ppm, exceeding rarely 6 ppm.10 Solution, X-ray and Gas–Phase Molecular Structure of Dialkylterphenyl Phosphines, PR2Ar’ The solid-state structures of the molecules of the newly reported terphenyl phosphines L1, L3, L4 and L6-L8 were determined by X-ray crystallography. To have on hand a sufficiently large number of structures, the molecular geometries of PMe2ArPh2 (L11), PMe2ArXyl2 (L12) and PMe2ArDipp2 (L14) were also ascertained.28 Adding these X-ray data to other already reported,24a,28 a structural database of more than a dozen X-ray structures becomes currently available, allowing to discriminate among three distinct phosphine conformations which formally interconvert by rotation around the Cipso-P bond (Figure 3). Supplementary to X-ray analysis, solution and computational studies on selected phosphines were developed. The three solid-state structural types are represented schematically in Figure 3 with A, B and C labels. Notice, however, that in Figure 3: (i) The R’ substitution pattern at the side rings is 3,5 in A but 2,6 in B. (ii) PMe2Ar’ phosphines exhibit structure of type A or B, while C is characteristic of the bulkier PR2Ar’ phosphines presented in Figure 2 (R = iPr, c-C5H9 or c-C6H11). (iii) To mitigate steric hindrance between the P-bound alkyl groups and the neighbouring flanking ring, the PR2 moiety in structures B and C bends away from the ring in close proximity, such that one of the P-Cipso-Cortho angles widens (>120°) at the expense of the other (<120°). On the basis of the experimental and computational results discussed in this work, we propose that a certain phosphine PR2Ar’ adopts in the solid state one of the three A, B or C structures, depending on steric repulsions among the phosphorus substituents and electron-electron repulsion between the phosphorus atom lone pair and the -system of the adjacent ring. On these grounds, structure A is the preferred geometry for the least sterically demanding phosphines studied, namely PMe2ArDtbp2 (L1) and PMe2ArPh2 (L11), the two featuring P-Me bonds and no ring substitution at positions 2 and 6. Besides L1 and L11, A is also the structural type found for P(H)2ArMes2 and PMe2ArXyl’2 (Mes = 2,4,6-C6H2Me3; Xyl’ = 3,5-C6H3Me2), reported by Wehmschulte and coworkers.24a Concerning PMe2Ar’ phosphines bearing 2,6-disubstituted flanking rings, the enlarged steric impediments brought in by these substituents results in an observable structural change from conformation A to B. Combination of a ca. 50° rotation of the – PMe2 half around the P-Cipso bond to make one of the P-Me bonds almost coplanar with the Ar’ central ring, and opening of the P- Cipso-Cortho bond angle of the PMe2 unit and the closer side ring, partially relieves steric tension making conformation B somewhat more favourable than A. Experimentally, this is the molecular geometry ascertained for PMe2ArXyl2 (L12) and PMe2ArDipp2 (L14), as well as for the previously reported24a,28 PMe2ArMes2 (L13), P(C≡CH)2ArMes2, P(CH2CH=CH2)2ArDipp2 and the Buchwald-type biaryl phosphine Methyl-JohnPhos.29a Figure 3. Different conformations adopted by dialkylterphenyl phosphines in the solid state. The third type of structure found in the solid-state for the PR2Ar’ investigated becomes the preferred one for the bulky, branched or cyclic alkyl groups iPr, c-C5H9 and c-C6H11 (L3, L4, L6-L8 in Figure 2). Now, steric hindrance among the R groups and the substituents at the flanking aryl rings forces the bulky P- bonded alkyl units to place themselves above and below the plane of the central aryl ring, i.e. frontwards and backwards in the perspective shown for conformation C in Figure 3. Likewise, and similar to structure B, to relieve steric tension between the R groups and the proximal ring, the PR2 moiety bends away from the ring, enlarging the corresponding P-Cipso-Cortho bond angle (see below for details). Interestingly, along with the bulky PR2Ar’ phosphines studied in this work, all the sterically demanding dialkylbiaryl phosphines analysed by the group of Buchwald,40a or by others,40b,c feature also a structure of type C. To gain insights into the solution behaviour of terphenyl phosphines, variable temperature NMR and complementary computational studies were additionally carried out. A comparative analysis of PMe2ArDtbp2 (L1, structure A) and PiPr2ArDtbp2 (L3, structure C), appears appropriate. Computational calculations revealed that for L1, rotation around the P-Cipso bond is a very facile process, with transition states between conformations of the order of 3-3.5 kcal·mol-1 relative to the ground state structure (Figure S1). This is in accordance with solution NMR data down to −80 ºC, where both flanking rings remain equivalent (Figure S3). In contrast, the 1H-NMR spectrum of L3 reveals inequivalence of the two rings at −80 ºC, as denoted, for instance, by the observation of two signals for the t-Bu groups with  1.29 and 1.28 ppm, or by the appearance of two distinct signals (7.17 and 7.04 ppm) for the ortho H atoms of the Dtbp rings (Figure S4). Upon warming, exchange of the flanking rings becomes faster, averaging the exchanging sites. Thus, at −50 ºC the two t-Bu resonances merge and originate a broad signal centred at 1.33 ppm. From this and other data of the interchanging H-positions, an energy barrier ΔG‡ ≈ 11.7 kcal·mol-1 can be FULL PAPER 5 estimated, using Eyring equation and the value of the rate constant determined at the coalescence temperature (−50 ºC). In reasonable agreement with these experimental results, the computational analysis of terphenyl rotation around the P-Cipso bond reveals that ring exchange in the ground state structure C needs surmounting an energy barrier of ca. 14 kcal·mol-1 (see SI, p. 16 for details). With this information in mind, some comments are pertinent regarding the coordination capabilities of the different terphenyl phosphine structural types. Regardless of the nature of Ar’, structure A is readily accessible for dimethylterphenyl phosphines PMe2Ar’. It is of note that for this rather symmetric conformation, the phosphorus lone-pair points to a void region of space with little steric interference from the terphenyl substituents. Classical P- coordination is therefore expected, with little or no front-strain (F- strain),34 depending upon the nature of the unsaturated transitionmetal Lewis acid centre. As conformations B and C can also be readily accessed by these phosphines, other coordination modes involving the -system of the proximate aryl ring, i.e. -P,n-Carene coordination (see Figure 1), can be foreseen, contingent on metal needs. Contrary to this situation, for bulky phosphines with structure C the electron density of the phosphorus lone pair faces the nearby aryl ring such that considerable F-strain might arise when bonding to crowded, or relatively crowded, transition-metal fragments. At the same time, this directionality of the lone pair facilitates the formation of complementary M···Carene bonds, that is, once again -P,n-Carene binding. As discussed later, these considerations are in excellent agreement with the reactivity found for PR2Ar’ phosphines and Ni(CO)4 as a source of unsaturated “Ni(CO)n” fragments (n = 2, 3). Figure 4 contains two views of the molecular structure of PMe2ArDtbp2, (L1), emphasizing the regular distribution of the −PMe2 half relative to the terphenyl group. Besides other metrics, this is nicely evinced by two almost identical P-Cipso-Cortho bond angles of 121.3(2) and 121.4(2)º. A conspicuous structural feature clearly perceivable in Figure 4a is the deviation of the flanking aryl ring ipso carbon atoms, C7 and C21, by ca. 0.32 and 0.33 Å from the plane of the central aryl ring, with the result of two Me groups of t-Bu substituents in opposite rings approaching to a distance of about 4.77 Å, only ca. 20% longer than twice the van der Waals radius of a Me group (2.0 Å).41 While it is tempting to attribute this distortion to London dispersion forces,17 dispersion-corrected DFT-D3 calculations do not support this assumption. Moreover, the unsubstituted terphenyl analogue, PMe2ArPh2, (L11), exhibits a similar deformation, with deviations of the Cipso atoms of about 0.28 Å (Figure S8). It must also be noted that due to the absence of substituents in the ortho positions of the lateral rings, in conformation A the latter are rotated around the Cortho-Cipso bond from the almost perpendicular arrangement with respect to the central ring observed in the other two conformations. This results in an angle between the planes containing the lateral rings and the plane of the central ring of 48° in the structure of PMe2ArPh2 (vs. an average of ca. 85° in PMe2ArXyl2). Figure 4. Two views of the molecular structure of PMe2ArDtbp2, L1. Selected bond lengths (Å) and angles (º): P-C(1), 1.864(3); P-C(35), 1.848(3); P-C(36) 1.851(3); C(36)-P-C(35), 95.2(2); C(35)-P-C(1), 101.4(1); C(36)-P-C(1), 101.9(1); P-C(1)-C(2), 121.3(2); P-C(1)-C(6), 121.4(2). The molecular structures of PMe2ArDipp2 (L14, structure of type B) and PiPr2ArXyl2 (L6, structure C) are illustrated in Figure 5. X- ray data for other terphenyl phosphines studied in this work that display structure B or C can be found in the Supporting Information (Figures S10-S11). For PMe2ArDipp2 (Figure 5), one of the P-Me bonds namely P-C(2), is near to coplanar with the central aryl ring, with a C(Me)−P−Cipso−Cortho torsion angle of 19.2º. In addition, to attenuate steric repulsions between the P-bonded methyl groups and the adjacent aryl ring, the pertinent P-Cipso- Cortho bond angle distends to 128.64(9)º with an accompanying decrease of the other to 113.30(9)º. Molecules of PiPr2ArXyl2 (Figure 5) and of other i-propyl-, c-pentyl- and c-hexyl-terphenyl phosphines investigated, present also two distinct P-Cipso-Cortho bond angles, the wider between ca. 127.5 and 130.6º, and the smaller in the interval 111.6-115.1º. But the prominent structural feature in these bulky phosphines is doubtless the placing of the two R groups in opposite regions of space relative to the plane of the terphenyl central aryl ring (above and below in the perspective shown in Figure 5). Some steric properties of the terphenyl phosphines in their Ni(0)-CO-PR2Ar’ complexes, in particular their angular symmetric deformation coordinate S4’ parameters,42 will be analysed in the coming section. Gas-phase, energy-minimized structures of selected terphenyl phosphines were obtained with the aid of DFT calculations. Fair agreement between experimental and calculated geometries was found. Relevant bond distances and angles are collected in Table S1. FULL PAPER 6 Figure 5. Molecular structures of PMe2ArDipp2 (L14) (above) and PiPr2ArXyl2 (L6) (below). Selected bond lengths (Å) and angles (º): L14: P-C(1), 1.858(1); P- C(31), 1.834(2); P-C(32) 1.839(2); C(31)-P-C(32), 99.29(9); C(31)-P-C(1), 109.54(6); C(32)-P-C(1), 99.61(6); P-C(1)-C(2), 128.64(9); P-C(1)-C(6), 113.30(9) . L6: P-C(6), 1.872(2); P-C(16), 1.882(2); P-C(19), 1.866(2); C(16)-P- C(19), 104.16(7); C(16)-P-C(6), 107.99(6); C(19)-P-C(6), 103.69(7); P-C(6)- C(5), 129.1(1); P-C(6)-C(7), 113.6(1) . Dialkylterphenyl Phosphine Nickel Carbonyl Complexes We considered of interest ascertaining the electronic properties of dialkylterphenyl phosphines analysing their coordination chemistry towards Ni(CO)4, and studying the resulting Ni(0) carbonyl-phosphine complexes, Ni(CO)n(PR2Ar’). The bulkiness of PR2Ar’, as well as the relatively small nickel covalent radius of 1.34 Å,43 anticipated coordination of only one phosphine per nickel atom. In his pioneering studies, Tolman generated Ni(CO)3(L) complexes for a variety of three-coordinate P-donor ligands by the room temperature reaction of Ni(CO)4 and L in CH2Cl2, and determined the widely employed Tolman electronic parameter for the ligand L,  or TEP, as the frequency of the symmetric A1 carbonyl stretching mode in the corresponding Ni(CO)3(L) species.18,35a Despite the marked, longstanding impact of this work in organometallic chemistry and catalysis, it is all the more surprising that, as briefly cited, a CSD search (Cambridge Structural Database) revealed36 an astonishing scarcity of X-ray data on Ni(CO)3(PR3) complexes of commonly utilized, commercially available alkyl and aryl phosphines. Thus, some ferrocenylphosphine44a and fluoroalkylphosphine44b Ni(CO)3(PR3) compounds have been characterized, among others, by X-ray crystallography. In addition, carbonylation of a binuclear NiI-NiI complex of the p-terphenyl diphosphine 1,4-bis(2-(diisopropylphosphino)phenyl) benzene, afforded a crystalline material shown by X-ray diffraction45 to be a mixture of nickel(0) carbonyls, in which 80% of the phosphines are bound to Ni(CO)2 and the remaining phosphines to Ni(CO)3. IR data were not given. Yet, Ni(CO)3(PR3) complexes of PiPr3, PPh3, P(c-C6H11)3 and other common phosphine derivatives are conspicuously absent, with the relevant exception of Ni(CO)3(PtBu3).44c Moreover, attempts to synthesize Ni(CO)3(P(c- C6H11)2Ar) complexes of Buchwald-type biaryl (Ar) phosphines proved unsuccessful, because the ligands appeared to be too sterically demanding to stabilize such complexes.46 Recently, the steric and electronic properties of another series of dialkylbiaryl phosphines, PR2ArPh, were investigated using IR and X-ray data collected for a diversity of Cr(0), Pd(0) and Pd(II) compounds.29a We found that the target Ni(0)-CO-PR2Ar’ complexes could be accessed with comparable reaction outcomes either by treatment of Ni(CO)4 with PR2Ar’ or by carbonylation of 1:1 mixtures of Ni(cod)2 and PR2Ar’ (cod = 1,5-cyclooctadiene). For convenience, the latter procedure was optimized and employed in all reactions investigated.47 The interaction of equimolar mixtures of Ni(CO)4 and PR2Ar’ in THF solution could be readily monitored by IR and 31P{1H} NMR spectroscopy, showing that depending upon the nature of the phosphine, two types of complexes could form, namely, tricarbonyls Ni(CO)3(PR2Ar’) (1) or dicarbonyl species Ni(CO)2(PR2Ar’) (2) (Scheme 3). The two kinds of compound give rise to the expected carbonyl stretching bands, that were registered in the vicinity of 2065 (A1) and 1980 cm-1 (E modes) for tricarbonyls 1, and around 1995 (sym) and 1920 (asym) for the dicarbonyl complexes 2. Similarly, though 31P NMR chemical shifts for PR2Ar’ depend markedly on the nature of the R group (Table 1), for compounds 1·PR2Ar’ a ∆ shift of ca. 30-40 ppm to higher frequencies relative to the free phosphine was recorded, and of about 40-50 ppm, i.e., an extra ~6-10 ppm shift, for the dicarbonyls Ni(CO)2(PR2Ar’). Comparison with NMR data already in the literature for other transition metal terphenyl phosphine complexes27,28 suggests classical P-coordination of the phosphine in 1·PR2Ar’ and bidentate binding in 2·PR2Ar’, resulting from additional Ni···Carene electronic interactions with the terphenyl substituent. This supposition is supported by X-ray studies and will be discussed later in sufficient detail. Scheme 3. Synthesis of complexes 1·PR2Ar’ and 2·PR2Ar’. As represented in Scheme 3a, only the dimethyl and diethyl terphenyl phosphines, PMe2Ar’ and PEt2Ar’, originated the sought Ni(CO)3(PR2Ar’) complexes, 1·PR2Ar’, in reactions that took place readily at room temperature and a CO pressure of 1 bar. Instead, ligands containing the more sterically demanding iPr, c- C5H9 and c-C6H11 alkyl groups afforded either the dicarbonyl derivatives 2·PR2Ar’ of Scheme 3b, or non-isolable complexes. In this instance, however, reaction mixtures had to be stirred under vacuum, in many cases at high temperatures (around FULL PAPER 7 80 °C). A fair number of Ni(CO)n(PR2Ar’) (n = 2, 3) complexes were fully characterized by microanalysis, IR and NMR spectroscopy, and some of them were additionally authenticated by X-ray crystallography. Others were, however, generated in solution for IR spectroscopy studies (Tables 2 and 3). The Tolman electronic parameter, TEP, of commercial P(c-C5H9)3 was additionally measured, and for the sake of completeness complex Ni(CO)3(PPh3)48 was also prepared and crystallographically characterized. Likewise, two dialkylbiaryl phosphines, specifically P(c-C6H11)2ArTripp (XPhos) and PtBu2ArTripp (tBuXPhos) were examined, and, whereas in accordance with previous studies the latter originated no isolable products,46 we were able to characterize a stable Ni(CO)3(XPhos) complex. Within the series of dicarbonyl derivatives (Scheme 3b), X-ray studies provided precise details of the unusual nickel coordination environment existing in 2·PiPr2ArDtbp2 and 2·P(c-C5H9)2ArXyl2. Table 2. IR Wavenumbers (cm-1) for the Carbonyl Stretching Vibrations in Ni(CO)3(PR3) complexes in CH2Cl2 solution. Ligand Complex CO(A1) CO(E) L12 Ni(CO)3(PMe2ArXyl2) 2063.8 1987 L13 Ni(CO)3(PMe2ArMes2) 2063 1987 L1 Ni(CO)3(PMe2ArDtbp2) 2063 1988 L14 Ni(CO)3(PMe2ArDipp2) 2062.9 1986 L15 Ni(CO)3(PMe2ArTripp2) 2062 1985 L2 Ni(CO)3(PEt2ArDtbp2) 2061 1985 L5 Ni(CO)3(P(c-C6H11)2ArDtbp2) 2060 1980[b] XPhos Ni(CO)3(P(c-C6H11)2ArTripp) 2059 1980 P(c-C5H9)3 Ni(CO)3(P(c-C5H9)3) 2059 1980 P(c-C6H11)3 Ni(CO)3(P(c-C6H11)3)[a] 2056.4 1973 PtBu3 Ni(CO)3(PtBu3)[a] 2056.1 1971 PPh3 Ni(CO)3(PPh3)[a] 2068.9 1990 [a] From ref. 18. [b] Partially obscured band. Tolman evinced that the symmetric A1 mode of the carbonyl ligands in Ni(CO)3(PR3) spanned the range 2056.1 (PtBu3) to 2011 cm-1 (PF3), while the degenerate E modes were recorded between 1971 (PtBu3) and 2016 cm-1 (PCl2Ph) (the figure for PF3 was not given).35a As represented in Table 2, the new Ni(CO)3(PR2Ar’) complexes featuring a coordinated dimethylterphenyl phosphine ligand are characterized by A1 stretching vibrations in the narrow range 2062 to 2064 cm-1, i.e. a few wavenumbers below PMe2Ph (2065.3 cm-1) and even slightly down the PMe3 value (2064.1 cm-1). These data confirm the electron-rich nature of PMe2Ar’ ligands, i.e. their overall electrondonor capacity. Similarly, for the Ni(CO)3(PR2Ar’) derivatives of PEt2ArDtbp2 (L2) and P(c-C6H11)2ArDtbp2 (L5), the A1 were registered at 2061 for L2, and 2060 for L5. The above figures for L2 should be compared with the 2063.7 and 1982 cm-1 stretchings found for PEt2Ph.18 Regarding P(c-C6H11)2Ar’ phosphines, it is noteworthy that the TEP obtained in our work for Ni(CO)3(P(c-C6H11)2ArTripp) of 2059 cm-1 compares well with that given above for L5, but it is somewhat higher than the 2054 cm-1 value that was estimated by Nolan et al.46 by linear correlation from ν(CO)average in IrCl(CO)2(XPhos).49 We note further that the TEPs obtained for P(c-C5H9)2ArDtbp2 (L5), and P(c-C6H11)2ArTripp (i.e. XPhos) of 2060 and 2059 cm-1, respectively, are only a few cm-1 higher than for P(c-C6H11)3 (2056.4 cm-1) and comparable to the 2060.6 cm-1 figure that can be calculated for P(c-C6H11)2Ph employing Tolman substituent contributions.18 The TEP measured in this work for P(c-C5H9)3 is 2059 cm-1. The symmetric and antisymmetric ν(CO) stretchings found for the formally three-coordinate dicarbonyl complexes 2·PR2Ar’ can be found in Table S1. Although non-carbonyl-containing, threecoordinate, 16 valence-electron Ni(0) complexes have long been known,50 analogous CO complexes are very rare and seem to be limited to a few examples encompassing derivatives incorporating strongly nucleophilic carbene ligands (see also reference 45). To the best of our knowledge, truly three-coordinate, 16 valenceelectron Ni(CO)2(PR3) compounds are unknown, and those reported herein constitute no exception, because X-ray studies to be described next unambiguously demonstrate the existence in compounds 2·PR2Ar’ of weak Ni···Carene interactions implicating one of the terphenyl flanking aromatic rings. This structural peculiarity limits comparison of IR (CO) data with four-coordinate Ni(CO)2(PR3)2 complexes, and also with the few known examples of three-coordinate Ni(CO)2(L) derivatives of carbene ligands.51 Notwithstanding these shortcomings, it is appropriate remarking that for the 2·PR2Ar’ complexes collected in Table S1, sym centres in the proximity of 1995 cm-1 and asym at about 1923 cm- 1, values comparable to the 1990 and 1926 wavenumbers characteristic of the four-coordinate, bis-PMe3 complex Ni(CO)2(PMe3)2. With due caution, given the different nature of the compounds, the IR properties of complexes 2·PR2Ar’ reflect the high metal basicity of their [Ni(-P,2-Carene-PR2Ar’)] metal fragment. We have studied the solution dynamic behaviour of 2·PiPr2ArDtbp2 by variable temperature NMR spectroscopy. Two intramolecular rearrangements can be envisioned. First, interchange of the two degenerate 2-structures, each involving the ipso and one of the ortho carbon atoms of the Dtbp ring engaged in nickel bonding, through a -P,1-Cipso transition state (see Scheme 4), is expected to be fast (only 1.7 kcal·mol-1 energy barrier according to DFT calculations), generating an effective plane of symmetry containing the terphenyl central ring and the Ni-P bond. Second, since a relatively weak Ni-2-Carene bonding interaction is foreseeable (vide infra), temporary cleavage of the Ni-2-Dtbp linkage to form a truly three-coordinate Ni(CO)2(PiPr2ArDtbp2) intermediate or transition state, followed by rotation around the Cipso-P bond would result in the shuffle of the two Dtbp units. In the room temperature 1H NMR spectrum of complex 2·PiPr2ArDtbp2 the two isopropyl groups of the phosphine ligand are equivalent and originate a multiplet centred at 1.94 ppm (2 H, doublet of septets, 2JHP = 13.3, 3JHH = 6.6 Hz) accompanied by another, also well-defined multiplet spanning across the 0.92-0.84 ppm interval (6 H + 6 H; 3JHP = 10.4, 9.4 Hz). In like manner, a doublet 13C NMR resonance arises at 198.9 ppm (2JCP = 10 Hz) due to two equivalent carbonyl ligands. It is therefore evident that the first of the aforementioned dynamic processes, namely, interconversion of the two equivalent -P,2-Carene structures, occurs in a swift manner under ambient conditions. In contrast, the tBu protons of the two lateral rings remain inequivalent at room temperature and are seen in the form of a broad resonance at FULL PAPER 8 Scheme 4. Solution dynamic behaviour of 2·PiPr2ArDtbp2 exchanging degenerate Ni-2-Carene structures. around 1.39 ppm. Nevertheless, this signal resolves in two singlets at 1.36 and 1.39 ppm (18 H, 18 H) on cooling at −30 °C. As this change is not accompanied by noticeable variations of the 1H NMR resonances of the –PiPr2 moiety, in all probability it can be proposed to be due to significantly slowed down commuting of the Dtbp rings, and hence to attainment at low temperatures of a static structure in which the phosphine coordination mode is akin to that found in the solid state. From the value of the rate constant for exchange determined at the coalescence temperature of ca. 15 °C, an energy barrier ∆G‡ ≈ 15 kcal·mol-1 can be estimated52 for the flipping of the terphenyl flanking rings. DFT calculations on this process are in accordance with this value (see SI). To attain unambiguous, definitive structural information on the nickel carbonyl terphenyl phosphine complexes, 1·PR2Ar’ and 2·PR2Ar’, single crystals suitable for X-ray studies were grown for some of these compounds. Concretely, the tricarbonyl derivatives of PMe2ArXyl2, PMe2ArDipp2 and PMe2ArDtbp2, as well as the dicarbonyls Ni(CO)2(PiPr2ArDtbp2) and Ni(CO)2(P(c-C5H9)2ArXyl2), were crystallographically characterized. For comparative purposes, the structure of the known Ni(CO)3(PPh3),48 was likewise determined (Figure S12). Figure 6. Molecular structure of tricarbonyl Ni(CO)3(PMe2ArDtbp2) (1·L1). Selected bond lengths (Å) and angles (º): Ni-P, 2.2401(7); Ni-C(21), 1.797(3); Ni-C(20), 1.782(3); P-C(1), 1.861(3); P-C(19), 1.832(2); C(19)-P-C(19), 97.2(2); C(19)-P-C(1), 105.38(9). As illustrated in Figures 6 and S12, the three tricarbonyls investigated feature the expected, somewhat distorted tetrahedral geometry around the metal centre. A conspicuous structural feature is the common conformation of type A (Figure 3) adopted by the phosphine in these complexes. As can be seen, not only the two P-Me bonds occupy the same region of space relative to the plane of the terphenyl central aryl ring, but moreover the – PMe2 moiety is symmetrically disposed with respect to this ring such that the two P-Cipso-Cortho angles are practically identical (ca. 121° in the three complexes). As free, non-coordinated molecules, only PMe2Ar’ adopts a conformation alike A, inasmuch as PMe2ArXyl2 and PMe2ArDipp2 prefer a structure of type B. In view of the meagre differences in energy between conformations A and B in the free phosphines, it seems plausible that to attenuate F- strain between the B phosphine conformation and the Ni(CO)3 fragment in the Ni(CO)3(PMe2Ar’) complex, conformation A becomes favoured. As mentioned earlier, in conformation A the phosphorus lone pair points towards an unoccupied region of space, whereas B-type complexation would place the Ni(CO)3 moiety relatively close to one of the flanking aryl rings of the terphenyl substituent. Figure 7. Molecular structures of dicarbonyls (Ni(CO)2(PiPr2ArDtbp2) (2·L3) and Ni(CO)2(P(c-C5H9)2ArXyl2) (2·L7). Selected bond lengths (Å) and angles (º): 2·L3: Ni-P, 2.2056(6); Ni-C(41), 1.763(3); Ni-C(42), 1.776(3); Ni-C(7), 2.449(2); Ni- C(8), 2.332(2); P-C(1), 1.849(2); P-C(35), 1.863(3); P-C(38), 1.848(3); C(35)-P- C(38), 106.1(2); C(35)-P-C(1), 104.0(1); C(38)-P-C(1), 104.5(1). 2·L7: Ni-P, 2.2037(8); Ni-C(33), 1.767(4); Ni-C(34), 1.770(4); Ni-C(7), 2.438(3); Ni-C(8), 2.414(3); P-C(1), 1.850(2); P-C(23), 1.845(3); P-C(28), 1.845(3); C(23)-P-C(28), 104.2(1); C(23)-P-C(1), 107.7(1); C(28)-P-C(1), 104.3(1). FULL PAPER 9 A comparison of selected bond distances and angles for compounds 1·PR2Ar’ and related Ni(CO)3(L) structures is given in Table 3. The Ni-P distances in 1·PMe2ArXyl2 and 1·PMe2ArDtbp2 are very similar (2.248(1) and 2.240(1) Å) and comparable to that Table 3. Selected Structural Data for Ni(CO)3(L) and Ni(CO)2(L) complexes of tertiary phosphine and nucleophilic carbene ligands. Ligand Complexes Ni‒P (Å) Ni‒CO (average, Å) ∑C‒P‒C (°) S4’ (°) L1 Ni(CO)3(PMe2ArDtbp2) 2.240 1.792 308.0 38.0 L12 Ni(CO)3(PMe2ArXyl2) 2.248 1.780 311.1 31.5 L14 Ni(CO)3(PMe2ArDipp2) 2.294 1.776 315.7 24.4 PPh3 Ni(CO)3(PPh3) 2.224 1.800 308.7 37.5 PtBu3[a] Ni(CO)3(PtBu3) 2.289 1.721 323.4 12.9 IPr[b] Ni(CO)3(IPr) 1.979 1.792 CAACMethyl[c] Ni(CO)3(CAACMethyl) 1.963 1.800 L3 Ni(CO)2(PiPr2ArDtbp2) 2.206 1.769 314.6 26.0 L7 Ni(CO)2(P(c-C5H9)iPr2ArXyl2) 2.204 1.770 316.2 23.4 IAd[a] Ni(CO)2(IAd) 1.96[d] 1.760 ItBu[b] Ni(CO)2(ItBu) 1.96[d] 1.751 C(PPh3)2[e] Ni(CO)2(C(PPh3)2) 1.99[d] 1.746 [a] Ref. 44c. [b] Ref. 51b. [cb] Ref. 51c. [d] d(Ni-CCarbene). [e] Ref. 51a in Ni(CO)3(PPh3), the latter being equal to 2.235(1) Å. The most sterically demanding of the three phosphines, PMe2ArDipp2, forms the longest Ni-P bond in this series at 2.294(2) Å. Referring to Ni- CO bond distances, which for the three complexes studied cluster around 1.78 Å, it is hardly surprising that they are shorter than in Ni(CO)4 (ca. 1.82 Å) and comparable to those in Ni(CO)3(L) complexes containing PPh3, N-heterocyclic carbenes (NHC) or cyclic (alkyl)(amino)carbenes (CAAC) ligands (Table 4).51 Probably, the length of the Ni-CO bonds in these complexes is more a reflection of the sterics than of the electron density at the Ni(0) centre. The C-O distances, with values of approximately 1.14 Å, are also unexceptional and of scant structural utility, because, as stated by Cotton and Wilkinson,53 in the 2-3 bond order range concerned, the CO bond length is relatively insensitive to bond order. For the three 1·PMe2Ar’ complexes investigated, the sum of the C-P-C bond angles is approximately 10° larger than for the free ligands and amounts to 315.7(3)° (PMe2ArDipp2), 311.1(3)° (PMe2ArXyl2) and 308.0(2)° (PMe2ArDtbp2). The values of the angular symmetric deformation coordinate S4’ for the phosphine ligands in complexes 1·PR2Ar’ are also included in Table 3.42 It is worth recalling that PMe3, PPh3 and PtBu3 possess mean S4’ parameters of 46.5, 27.6 and 2.6°, respectively, and that S4’ values for a given phosphine may span a considerable wide range.42 Consequently, it is best to compare series of related complexes. A recent study based on trans-PdCl2(PR3)2 complexes provided S4’ values for Buchwald-type dialkylbiaryl phosphines, PR2ArPh2, of ca. 28° (R = Me), 29.3° (Et), 29.8° (Ph) and 16.9 (c-C6H11).29a Corresponding parameters for PMe3, PMe2Ph and PtBu2Ph in these complexes were 35.3, 32.8 and 18.4°, respectively. As shown in Table 3, X-ray data for the Ni(CO)3(PR3) complexes discussed in this work lead to S4’ values of 38.0° (PMe2ArDtbp2), 31.5° (PMe2ArXyl2) and 24.4° (PMe2ArDipp2) in what refers to terphenyl phosphines, and of 37.5° for PPh3 and 13.0 for PtBu3. Clearly, the steric demands of our terphenyl phosphines are significantly smaller than for PtBu3, and in comparison with PPh3, PMe2ArDtbp2 is equivalent and PMe2ArXyl2 and PMe2ArDipp2 increasingly larger. As already noted, PR2Ar’ phosphines containing the sterically demanding iPr, c-C5H9 and c-C6H11 alkyl groups did not provide isolable Ni(CO)3(PR2Ar’) complexes, although for P(c- C6H11)2ArDtbp2 minor amounts of the tricarbonyl derivative accompanied the major dicarbonyl reaction product. Once again, it should be recalled that the very bulky PR2Ar’ phosphines adopt a structure of type C (Figure 3) with the phosphorus lone pair facing the  system of one of the side aromatic rings, as found earlier for dialkylbiaryl phosphines.40 Under these circumstances, it is reasonable to surmise that F-strain between the phosphine and Ni(CO)3 fragment destabilizes the purported Ni(CO)3(PR2Ar’) complexes and promotes formation of dicarbonyls 2·PR2Ar’ (see Supporting Information, p. 18). X-ray studies on complexes Ni(CO)2(PiPr2ArDtbp2) and Ni(CO)2(P(c-C5H9)2ArXyl2) demonstrate that their seemingly unsaturated, three-coordinate structure corresponds in fact to four-coordinated nickel centres, as in each complex there is a relatively weak Ni-η2-Carene interaction with one of the Dtbp or Xyl substituents (Figure 7). The molecules of these complexes have in the solid state a very uncommon oblique triangular pyramidal geometry in which one of the Cipso-Cortho bonds of the proximal lateral ring occupies the apex. The electronic interaction with the flanking ring is undoubtedly weak,54 as it is characterized by fairly long Ni···Carene distances. For example, in the Xyl-substituted Ni(CO)2(P(c-C5H9)2ArXyl2) complex, where the nickel atom interacts with one of the Cipso=C(Me) bonds of a lateral ring, the two Ni-C bond distances are equal to ca. 2.44 (to C7) and 2.41 Å (to C8), both standing well above the 1.97 Å value of the sum of the covalent radii of Csp2 (0.73 Å) and Ni (1.24 Å).43 Complex Ni(CO)2(PiPr2ArDtbp2) possesses no substituents at the ortho