Iridium complexes bearing 2,6-bisphosphinomethyl aryl (PCP motif) and many related pincer ligands, including those with PNP, PPP, and bis (NHC)aryl (CCC) motifs, have been explored and developed extensively in the past 25 years. In particular such complexes have seen great success in C -H bond activation including catalytic alkane dehydrogenation and tandem reactions based upon dehydrogenation [1][2][3][4][5][6][7][8][9][10][11], as well as reactions involving cleavage and formation of C -O [12,13], N -H [14][15][16], and other strong bonds [17]. Iridium complexes of pincer ligands with terminal N-coordinating groups (e.g. NCN-type, such as 2,6bis-oxazolinephenyl, i.e. Phebox) have also seen development in catalysis and strong-bond activation. This chemistry, however, has typically not been analogous to that of the aforementioned ligands with phosphino-or carbene-coordinating "arms" (terminal groups) which largely operates via Ir(I) complexes. Instead, the chemistry of such NCN-iridium complexes has largely focused on Ir(III) carboxylate complexes that are believed to operate via concerted metalation-deprotonation (CMD) or other mechanisms involving high-oxidation complexes [18][19][20][21][22][23][24][25].
N-coordinating groups potentially offer significant advantages over P-coordinating and other "soft" groups, including ease of ligand synthesis, cost, and resistance to oxidation. From a fundamental perspective, it is of interest to understand the effect of such variations on catalytic or stoichiometric reactivity. However, whereas PXP-type pincer ligands (typically X = C or N) largely incorporate sterically demanding phosphino groups (e.g. P t Bu 2 or P i Pr 2 ), the NXN pincer ligands explored in this context have largely made use of groups such as oxazolines, which are much less bulky, and in which the limited steric bulk is positioned very differently than in the PCP complexes. These factors obfuscate any meaningful comparisons.
In this context we wished to explore the chemistry of an NCN complex with a relatively bulky N-coordinating group, and particularly one in which the steric bulk is not positioned only near the coordination site trans to the coordinating aryl carbon as in the case of Phebox. Toward this end we have synthesized adducts, 1-L, of the iridium dibromide complex of the phenyldiimine ligand, 2,6-bis(1-((2,6-dimethylphenyl) imino)ethyl)phenyl ( Xy PhDI). The complexes isolated represent somewhat unusual examples of late-metal pincer complexes bearing two electron-withdrawing low-field ancillary ligands (bromides). We have investigated the thermodynamics of the binding of the various ligands, L. DFT calculations are found to be in very good agreement with the experimental results. Encouraged by this agreement, we compare these results with binding thermodynamics calculated for the corresponding hydrides, and for the Ir(I) complexes of Xy PhDI and other pincer ligands. In an effort to synthesize the dihydride, which is of particular interest as a precursor of the corresponding Ir(I) fragment, we treated the dibromide precursor 1-MeCN with KO t Bu under H 2 atmosphere. This resulted in the formation of a bimetallic species, the structure of which was determined crystallographically. Remarkably the molecular structure revealed that one equivalent CH 3 CN had been hydrogenated to give ethylamine, and a second equivalent had undergone C-C bond cleavage to give an iridium center with a methyl and C-bound cyanide bridged to the second metal center. [27]. Crystals of 1-MeCN were grown by diffusion of pentane into a THF solution at room temperature, and the molecular structure was determined by scXRD (Fig. 1).
Acetonitrile was displaced from 1-MeCN by reaction with 1.2 equiv of pyridine, PMe 3 , N-ethylamine, or P(OMe) 3 , or CO (1 atm), to give 1py, 1-PMe 3 , 1-NH 2 Et, 1-P(OMe) 3 and 1-CO respectively. Their molecular structures are shown in Fig. 2a-e.
Bubbling ethylene through a toluene solution of 1-MeCN to dryness yielded a solid that was redissolved in benzene under argon atmosphere. Crystals were obtained by slow evaporation and the molecular structure of the product, 1-C 2 H 4 , was determined by scXRD (Fig. 2f).
Addition of ethylene atmosphere to a toluene solution of 1-MeCN (without bubbling to dryness) led to a mixture of 1-MeCN and 1-C 2 H 4 . The equilibrium of Scheme 2 was established with K eq = 0.181 (ΔG • = 1.0 kcal/mol).
Bubbling solutions of 1-C 2 H 4 or 1-MeCN to dryness with N 2 gas resulted in no substitution, nor did the analogous approach lead to substitution with H 2 gas. It appears that these species bind much more weakly than C 2 H 4 , or acetonitrile, if at all. Likewise addition of 1-hexene to a solution of 1-C 2 H 4 did not result in any observable substitution.
Addition of CO (1.9 atm) to a benzene-d 6 solution of 1-PMe 3 did not result in any substitution or other reaction. Conversely, however, addition of 1.2 equiv PMe 3 to a solution of 1-CO resulted in complete conversion to 1-PMe 3 . Given that CO typically binds very strongly to iridium complexes, including Ir(III) complexes, we found this result to be somewhat intriguing [28,29].
Equilibrium binding constants were determined for various pairs of ligands. In all cases the equilibrium was reached from both directions, by starting with a given complex 1-L, adding the complementary ligand L', waiting until equilibrium was apparently reached, and then adding an additional quantity of ligand L. K eq ([ML' eq ][L eq ]/[ML eq ][L' eq ] for each ligand pair L/L') was determined for the following ligand pairs at 25 • C: MeCN/C 2 H 4 , K eq = 0.181 (Scheme 2); MeCN/ t BuNH 2 , K eq = 32.0; t BuNH 2 / i PrNH 2 , K eq = 48.2; i PrNH 2 /py, K eq = 3.71; py/EtNH 2 , K eq = 2.63; py/CO, K eq = 13.8; EtNH 2 /CO, K eq = 4.95; EtNH 2 /PPh 2 Me, K eq = 2.15. For PPh 2 Me/P(OMe) 3 , K eq = 29.2 was determined at 80 • C. The corresponding relative free energies of binding are given in Table 1.
Electronic structure (DFT) calculations were then conducted, initially for those complexes that were studied experimentally. Geometries were optimized in the gas phase using the M06 functional [30] and split valence basis set 6-31G(d,p) for C, H, N, O and P [31][32][33][34][35]. For Ir, the Stuttgart-Dresden effective core potential was used for the 60 core electrons; the associated basis set was used for the 17 valence electrons (SDD) [36]. All calculations were done at standard conditions of temperature (298.15 K) and pressure (1 atm); full details are given in the SI. Ir-L binding energies were calculated and are given in Table 2 BDFEs)(see SI for electronic energy, enthalpy and entropy of binding). The calculated relative free energies of binding are in excellent agreement with experimentally obtained values (Table 1). For those ligands for which relative binding free energies were obtained experimentally, the root-mean-square deviation of all calculated relative values compared with all relative values extrapolated from the experimental measurements is 0.8 kcal/mol [37]; we consider this to be very satisfactory agreement.
With potential ligands H 2 , N 2 and 1-hexene we were unable to observe any displacement of C 2 H 4 ; this is in agreement with their calculated low energies of binding. For H 2 and N 2 in particular, the failure to observe substitution or loss of 1-C 2 H 4 , even after bubbling solutions of 1-C 2 H 4 to dryness with the respective gas (and thereby providing a very strong entropic driving force for substitution) indicates that these molecules bind particularly weakly to 1 (if at all). The very unfavorable calculated free binding energies are consistent with this result. At the other extreme, none of the ligands used in this study were able to displace PMe 3 to any observable extent, in accord with its calculated very high relative free binding energy (13.7 kcal/mol greater than ethylene, Table 1).
Having established the ability of the computational method to reliably calculate relative energies of binding to 1, we used such calculation to study binding of the same ligands to related Ir fragments, specifically the corresponding dihydride (2), and the 14-electron Ir(I) fragment, ( Xy PhDI)Ir (3; no ancillary ligands) (Fig. 3). We calculate that for amines and phosphines, the variability of the energy of binding to these three fragments was fairly small, with bond dissociation free energies (BDFEs) for each ligand found to be within a range spanning 6 kcal/mol. In contrast, the BDFE of CO increased strongly over the series 1 < 2 < with the free energy of CO binding to Ir(I) fragment 3 (37.4 kcal/mol) being 19.3 kcal/mol greater (ΔG 1 -ΔG 3 ) than binding to Ir(III) fragment 1 (18.1 kcal/mol). Presumably this large variability reflects the degree of increasing π-donating ability among the various fragments, and commensurately increased metal-ligand π-backbonding.
Perhaps more surprising, in our view, was the magnitude of the variation calculated for binding of N 2 . N 2 generally binds much more weakly than CO and the complexes calculated in this study suggest no exception to that rule (e.g. the calculated free energy of N 2 binding to Scheme 2. Equilibrium between 1-MeCN and 1-C 2 H 4 .
Free energies of binding to 1 relative to 1-C 2 H 4.
Experimental Calculated Δ(Calc -Exptl)
.01 -0.82 0.19 t-BuNH 2 -3.05 -4.01 -0.96 i-PrNH 2 -5.34 -7.19 -1.85 pyridine -6.11 -6.75 -0.64 EtNH 2 -6.68 -7.58 -0.90 PPh 2 Me -7.13 -8.69 -1.56 CO -7.64 -8.52 -0.88 P(OMe) 3 -9.49 -12.11 -2.62 PMe 3 ≪ -9.5 -13.72 -is only 2.0 kcal/mol). But although N 2 is also generally considered to be a much weaker π-acceptor than CO[39] the variation of N 2 binding energies among the complexes is nearly as great as that found for CO; N 2 is calculated to bind 14.1 kcal/more strongly to fragment 3 than to fragment 1 (ΔG 1 -ΔG 3 ). NBO analysis of the CO and N 2 adducts of 1, 2 and 3 indicates, as expected, that CO is a better π-acceptor than N 2 (see SI). However, as the electron-donating ability of the fragment increases (1 < 2 < 3), the electronic occupancy of the ligand π * orbitals is calculated to increase approximately as much (actually slightly more) for N 2 than for CO [40]
. This seems very consistent with the calculated increase in the Ir-N 2 BDE being comparable to the increase calculated for the Ir-CO BDE for this series of complexes. Ethylene and 1-hexene show sensitivity to the nature of the fragment that is approximately equal to that of N 2 (ΔG 1 -ΔG 3 = 13.0 kcal/mol and 13.5 kcal/mol respectively) while trimethyl phosphite is comparably sensitive (ΔG 1 -ΔG 3 = 11.4 kcal/mol). Acetonitrile (ΔG 1 -ΔG 3 = 7.6 kcal/mol) shows variability somewhere in between that of the amine ligands and those ligands that are apparently very sensitive to π-donating ability (e.g. CO, N 2 , olefins).
H 2 is calculated to add to the three ( Xy PhDI)Ir fragments investigated to give a dihydrogen complex with a relatively short H-H distance varying only from 0.81 Å (1-H 2 ) to 0.85 Å (3-H 2 ). Thermodynamically, the binding energy shows moderate sensitivity to the nature of the iridium fragment (ΔG 1 -ΔG 3 = 9.5 kcal/mol).
Generally speaking, our calculations predict that π-acceptor ligands such as CO, but also N 2 and alkenes, bind much more strongly to the ( Xy PhDI)Ir species with greater π-electron-donating ability (( Xy PhDI) IrH 2 and ( Xy PhDI)Ir), with a sensitivity to π-electron-donating ability that is much greater than calculated for ligands such as amines; this is consistent with the most fundamental organometallic precepts and classical organometallic bonding descriptions. In such descriptions, CO is presented as a strong π-acceptor while amines are pure sigma donors or are even considered to be π-donors as well as σ-donors. From that perspective, however, it is noteworthy that the binding energy of Nethylamine for example is even slightly greater toward the more electron-rich iridium fragments; the model of simple donation seems inconsistent with these data. Moreover, the variation found for phosphines is very nearly equal to that for amines; this seems inconsistent with the former being considered even modest π-acceptors [41,42]. A full analysis of the origin of these effects is beyond the scope of this paper, but we plan to address these questions in future work.
We have also calculated, for comparison with ( Xy PhDI)Ir, the energies of binding of monodentate ligands to the (pincer)Ir(I) fragments (Phebox)Ir, ( iPr PCP)Ir, and ( tBu PCP)Ir (Table 2 and Fig. 3). Of these, the Phebox pincer ligand is of course the most closely related of these to Xy PhDI, having the same diaminoaryl NCN motif. van Koten and co-workers have shown that nickel complexes of these two pincer ligands have fairly similar redox properties as might be expected [26].
The sterically undemanding ligands H 2 , N 2 , and CO bind more strongly to the (Phebox)Ir fragment than to ( Xy PhDI)Ir, while acetonitrile binds slightly less strongly. This could suggest that (Phebox)Ir is a more π-electron-donating fragment but a slightly poorer σ-acceptor.
Larger differences between Ir-L BDFEs of (Phebox)IrL versus ( Xy PhDI)IrL are seen in the case of larger ligands, with the binding to (Phebox)Ir being weaker in all cases. The respective complexes are four-coordinate d 8 , therefore approximately square planar, and therefore this indicates specifically that the binding site trans to the Ir-bound aryl carbon is more crowded in (Phebox)Ir than in ( Xy PhDI)Ir. Inspection of the calculated structures of the Xy PhDI complexes supports this conclusion. The N-xylyl groups are oriented so that this trans coordination site of the Xy PhDI complexes is significantly more open than that of (Phebox)Ir. For example, in the respective PMe 3 complexes there are numerous close contacts (d H-H < 2.4 Å) between the Phebox methyl groups and the PMe 3 ligand (Fig. 4a), but no close contacts between coordinated PMe 3 and the Xy PhDI ligand (Fig. 4b).
Notably, although the trans site of the ( Xy PhDI)Ir fragment is much more open than that of (Phebox)Ir, the sites cis to the Ir-bound carbon are fairly crowded in the case of ( Xy PhDI)Ir while extremely open in the case of (Phebox)Ir. This is illustrated in Fig. 5 with space filling models of the respective (pincer)Ir(PMe 3 ) complexes and the Buried Volume maps of the (pincer)Ir fragments of Fig. 6 [43,44]. Qualitatively at least, the distribution of steric bulk in the ( Xy PhDI)Ir fragment resembles that of the (PCP)Ir fragments more closely than that of (Phebox)Ir, in that the (PCP)Ir fragments are also more crowded at the coordination sites cis to the Ir-bound carbon than at the trans sites. The symmetry of the ( tBu PCP) Ir fragment (Fig. 6c) makes the qualitative resemblance to ( Xy PhDI)Ir (Fig. 6a) more apparent than for the ( iPr PCP)Ir (Fig. 6d) fragment because the two i-propyl groups on each P atom are generally oriented in opposite directions (one tertiary C -H bond toward the Ir center and the other pointed away). Quantitatively however the Buried Volume calculations indicate that the ( Xy PhDI)Ir fragment (%V BUR = 68.2 %) is much more similar to ( iPr PCP)Ir (67.3 %) than to ( tBu PCP)Ir (78.4 %).
To assess the magnitude of the effect of steric crowding at the coordination sites cis to the Ir-bound carbon in the ( Xy PhDI)Ir unit, we calculated the thermodynamics of the (hypothetical) reaction of ( Ar PhDI)Ir(CO) with Br 2 to give trans-( Ar PhDI)IrBr 2 (CO), where Ar = xylyl (the present system) and Ar = phenyl (i.e. a model PhDI ligand lacking the xylyl methyl groups) (Fig. 7a). The addition of Br 2 is found to be 16 kcal/mol less favorable for the more crowded Xy PhDI complex. This effect is presumably entirely due to steric crowding, since any electronic effect is expected to be very small and to favor oxidative addition to the Xy PhDI complex. We also calculate that trans-addition of Br 2 to (phebox)Ir(CO) has a free energy approximately equal to that for ( Ar PhDI)Ir(CO), 71.2 kcal/mol, consistent with these complexes both
Calculated free energies (ΔG • , kcal/mol) of binding of various ligands to trans-( Xy PhDI)IrBr 2 (1), trans-( Xy PhDI)IrH 2 (2) a , ( Xy PhDI)Ir ( 3), (Phebox)Ir ( 4), ( iPr PCP)Ir ( 5) and ( tBu PCP)Ir (6).
L ( Xy PhDI)IrBr 2 ( Xy PhDI)IrH 2 a ( Xy PhDI)Ir (Phebox)Ir ( iPr PCP)Ir ( tBu PCP)Ir H 2 2.7 -5.5 -6.7 -9.2 -13.8 -9.5 N 2 -2.0 -10.3 -16.1 -17.2 -20.1 -19.2 1-hexene -6.6 -12.9 -20.1 -15.1 -21.4 -8.0 C 2 H 4 -9.6 -14.0 -22.6 -16.4 -23.7 -16.7 MeCN -10.4 -14.0 -18.0 -16.4 -18.4 -17.1 Fig. 3. Calculated free energies of binding of various ligands to fragments 1 -6. S. Das et al.
having negligible steric crowding at the cis coordination sites as would be expected (see Fig. 5). The crowding at the cis positions of trans-( Xy PhDI)IrBr 2 (CO) is clearly manifest in its geometric structure (Fig. 7b). The coordination sphere is distorted and the xylyl ring is canted so that one bromide ligand sits between the xylyl methyl groups and the CO ligand, while the other bromide is positioned between the methyl groups and the PhDI backbone. The corresponding C ipso -Ir-Br angles are 94. All ligands that we have studied are calculated to bind much more strongly to ( iPr PCP)Ir than to (Phebox)Ir (Fig. 3). Compared with ( Xy PhDI)Ir, however, the amines bind less strongly to ( iPr PCP)Ir while the phosphines bind more strongly to ( iPr PCP)Ir. The small π-accepting ligands bind more strongly to ( iPr PCP)Ir than to ( Xy PhDI)Ir, with CO in particular binding 12 kcal/more strongly, while P(OMe) 3 also binds much more strongly to ( iPr PCP)Ir (by 12.8 kcal/mol). P(OMe) 3 is known to be a good π-acceptor, but presumably much less π-accepting than CO.
The greater energy of binding of P(OMe) 3 to ( iPr PCP)Ir versus ( Xy PhDI)Ir may therefore be a combination of greater π-donating ability of the ( iPr PCP)Ir fragment combined with a greater tendency to bind to the Pdonating ligands generally, perhaps related to the "softness" of these ligands in contrast with the "hard" N-donors.
At room temperature, the 1 HNMR spectrum of 1-C 2 H 4 in the presence of free C 2 H 4 indicates rapid exchange of free and bound ethylene. Decoalescence of the respective 1 H NMR signals was observed at slightly reduced temperature, and the individual signals, attributable to free and bound C 2 H 4 respectively, were sharp at 255 K. Dynamic NMR allowed determination of the rate constants for exchange, over the temperature range 255 K -298 K (Table 3), by simulation using the dNMR feature in the program Topspin [45]. An Eyring plot (Fig. 8) of the rates thus obtained yielded activation parameters ΔH ‡ = 25.0 kcal/mol and ΔS ‡ = 36.0 ± 1.0 cal/mol•deg. The positive activation entropy indicates that the reaction proceeds via dissociation of ethylene, as would be expected of an 18-electron ethylene complex. The activation enthalpy is very close to the calculated thermodynamic value of the enthalpy of dissociation, ΔH • = 23.7 kcal/mol. Taken at face value, this would and d) to illustrate the greater crowding of (Phebox)Ir at the site trans to the Ir-bond C atom, in contrast with the greater steric hindrance of ( Xy PhDI)Ir(PMe 3 ) at the sites cis to the Ir-bond C atom. PMe 3 methyl groups in green and non-coordinating atoms of the pincer backbones in light grey. imply, as might also be expected, that addition of ethylene to the 16electron dissociation product, 1, has a near-zero enthalpic barrier of approximately ΔH ‡ = 1.3 kcal/mol. These results offer experimental support for the DFT-calculated thermodynamic values for ethylene (albeit approximate). Accordingly, they also support the validity of the absolute values of the DFT-calculated binding free energies of those ligands for which the relative (to ethylene) binding free energies are in agreement with experimental values.
In an attempt to reduce 1-MeCN to the corresponding dihydride or Ir (I) complex, the complex was treated with KO t Bu (3 equiv) in benzene under H 2 atmosphere and was left to stir at room temperature overnight under the hydrogen atmosphere. Benzene was evaporated and the solution was extracted with pentane to remove excess base. Crystals were grown under inert atmosphere at room temperature by diffusion of pentane into a concentrated benzene solution. Unexpectedly, scXRD revealed the product to be a binuclear bridging cyanide complex with the molecular structure as shown in Fig. 9. Although hydrides were not located unambiguously by crystallography, the 1 H NMR spectrum of the crystals, after dissolving in benzene-d 6 , indicated the presence of two equivalent hydride ligands (-26. a reaction of much interest [49][50][51][52][53][54][55][56]. The examples that are perhaps most closely related to the present work have been reported by Garcia and Jones [57][58][59][60][61][62].
We investigated the reaction of the ( Xy PhDI)Ir fragment with CH 3 CN computationally and located a transition state for direct oxidative addition of the acetonitrile C -C bond. However the barrier of this reaction was calculated to be too high, at 33.1 kcal/mol, to account for the formation of 7 at room temperature. Lewis acids, however, have been reported to accelerate C-CN bond cleavage [58,63] or the reverse, C-CN reductive elimination [64,65]. (Notably, Garcia and Jones have reported an example where Lewis acid binding actually inhibited C-C cleavage [57]). Therefore, given that the nitrile group of 7 is bridging, we considered that the second ( Xy PhDI)Ir unit played a role in promoting the C -C cleavage reaction. In accord with this hypothesis, the free energy barrier to C -C cleavage was calculated to be only 10.4 kcal/mol when the acetonitrile N atom was coordinated to a ( Me PhDI)IrH 2 fragment (the N-coordinated (PhDI)Ir moiety was truncated in the transition state, with xylyl groups replaced by methyl groups for computational simplicity; 3*-H 2 ; Fig. 10). Although the free energy of this binding is positive, ΔG • = 7.7 kcal/mol, due to a large unfavorable entropy term (TΔS = -16.7 kcal/mol at 298 K), the overall calculated barrier of C -C cleavage, ΔG ‡ = 18.1 kcal/mol, is still dramatically lowered by assistance from the second (PhDI)Ir center. The initial C -C cleavage product and the final, ethylamine-coordinated, product were calculated to be significantly exergonic relative to 3-MeCN, 3-H 2 , and EtNH 2 .
The calculations indicate, remarkably, that in the transition states for C-CN cleavage, both metal assisted and non-assisted, the Ir-CN bond is significantly shorter in the transition states (1.94 Å and 1.96 Å for Irbridged and non-bridged respectively) than in the products in which the Ir-CN bond is fully formed (2.10 Å and 2.07 Å) (Fig. 11). Note, however, that as would generally be expected for an oxidative addition reaction, the Ir-CH 3 distance is significantly shorter in the products than in the transition states. Acetonitrile C-C oxidative addition by (dippe) Ni(0) [59] and Cp*(PMe 3 )Rh(I) [60] (not assisted by a second metal) has been investigated computationally in detail by Jones. While our results are generally in agreement with those, in the Rh(I) case the M-CN bond in the TS (2.00 Å) was slightly longer [60] than in the C-C cleavage product (1.97 Å). In the case of Ni(0), however, as in the present systems the M-C bond was shorter in the TS (1.82 Å) than in the product (1.88 Å) although the difference was not as pronounced.
Alkyl cyanide elimination/C-C bond formation has been compared with alkyl migration to CO (i.e. CO insertion into M-alkyl bonds [65][66][67]. This perspective might help to rationalize this unusual example of a TS for C -C cleavage with a M-C distance shorter than that of the product with the fully formed M-C bond. For example, we have computationally studied alkyl migration to CO of Mn(CO) 5 (CH 2 Ar), and found that the M-CO bond in the migration transition state is shorter (1.82 Å) than in the carbonyl reactant (1.86 Å) or in the acyl product of migration (1.88 Å) [68]. Note also that in the transition states calculated in this work, as well as in the systems studied by Jones [59,60], there is a significant agostic interaction with the acetonitrile methyl C -H bond. In our studies of alkyl migration to CO (or alkyl migration from acyl ligand to metal) it was shown that formation of an analogous agostic interaction played a significant role in the energy of the transition state [68]; these shared feature would seem to further support the proposed relationship between alkyl migration and alkyl-CN cleavage/ elimination. H 2 IrL, while there is little difference for ligands that are less π-accepting.
We also compared ( Xy PhDI)Ir-L BDFEs with BDFEs of (pincer)Ir-L complexes for pincers Phebox, iPr PCP, and tBu PCP. The (Phebox)Ir fragment, compared with ( Xy PhDI)Ir, forms slightly stronger bonds with the smallest π -accepting ligands, but (Phebox)Ir-L bonds are significantly weaker with bulky ligands, particularly those that are not significantly π-accepting. Although Buried Volume calculations indicate that the (Phebox)Ir fragment has overall greater "unburied" volume than ( Xy PhDI)Ir, the coordination site occupied by L in the fourcoordinate d 8 (pincer)IrL complexes (i.e. the site trans to Ir-bound aryl C) is significantly more crowded in (Phebox)Ir. The motif of ( Xy PhDI)Ir, with greater steric crowding at the cis sites and a more open site trans to the aryl C is also found for both ( iPr PCP)Ir and ( tBu PCP)Ir pincer ligands.
The ( R PCP)Ir fragments appear to be more π-donating than the Phebox or PhDI fragments, as illustrated by stronger binding to CO or N 2 , but with respect to sterically demanding ligands, the BDFEs of ( Xy PhDI)IrL are somewhere between those of the very crowded ( tBu PCP)Ir and the much less crowded ( iPr PCP)Ir.