Rational Design and Synthesis of Highly Active ... - ACS Publications

5432

Organometallics 2009, 28, 5432–5444 DOI: 10.1021/om900568f

Rational Design and Synthesis of Highly Active Pincer-Iridium Catalysts for Alkane Dehydrogenation Sabuj Kundu, Yuriy Choliy, Gao Zhuo, Ritu Ahuja, Thomas J. Emge, Ralf Warmuth, Maurice Brookhart, Karsten Krogh-Jespersen,* and Alan S. Goldman* Department of Chemistry and Chemical Biology, Rutgers, The State University of New Jersey, New Brunswick, New Jersey 08903 Received July 1, 2009

“PCP”-pincer-ligated iridium complexes have been found to be highly effective catalysts for the dehydrogenation of alkanes. We report a computational and experimental study of the effect on catalytic activity resulting from systematically varying steric crowding by the substitution of methyl groups for the phosphino tert-butyl groups of (R4PCP)Ir (R4PCP = κ3-C6H3-2,6-(CH2PR2)2; R = t Bu or Me). DFT calculations for (R4PCP)Ir species (R4 = tBu4 or tBu3Me) indicate that the ratedetermining step in the n-alkane/1-alkene transfer dehydrogenation cycle is β-H elimination by (R4PCP)Ir(n-alkyl)(H). It is calculated that the transition state for this step is ca. 10 kcal/mol lower for (tBu3MePCP)Ir than for (tBu4PCP)Ir (relative to the corresponding free (R4PCP)Ir). However, this catalytically favorable effect is calculated to be partially offset by the strong binding of 1-alkene to (tBu3MePCP)Ir in the resting state, so the overall barrier is thus lower by only ca. 4 kcal/mol. Further Me-for-tBu substitutions have a smaller effect on the transition states, and the calculated energy of the olefin-bound resting states is lowered by comparable amounts; therefore these additional substitutions are predicted to have little overall favorable effect on catalytic rates. (tBu3MePCP)IrH4 has been synthesized and isolated, and its catalytic activity has been investigated. It is indeed found to be a more active catalyst precursor than (tBu4PCP)IrH4 for alkane transfer dehydrogenation. (tBu2Me2PCP)IrH4 was also synthesized and as a catalyst precursor is found to afford somewhat lower activity than (tBu3MePCP)IrH4. However, synthetic precursors of (tBu2Me2PCP)IrH4 tended to yield dinuclear clusters, while complex mixtures were observed during catalysis that were not amenable to characterization. It is therefore not clear if the lesser catalytic activity of (tBu2Me2PCP)Ir vs (tBu3MePCP)Ir derivatives is due to the energetics of the actual catalytic cycle or due to deactivation of this catalyst via the facile formation of clusters.

Introduction Olefins are ubiquitous as reagents and intermediates in organic chemistry in the synthesis of petrochemicals, commodity chemicals, and fine chemicals. For this reason, the catalytic dehydrogenation of alkanes and, more broadly, alkyl groups is a reaction with tremendous potential value. The development of soluble transition metal complexes as catalysts for alkane dehydrogenation was pioneered in the early 1980s by the groups of Crabtree and Felkin.1 Several catalytic systems were developed, although turnover numbers were limited by catalyst decomposition. Various groups have significantly contributed to progress in this field subsequently. A key breakthrough was the report by Kaska and *Corresponding author. E-mail: [email protected]; kroghjes@ rutgers.edu. (1) (a) Burk, M. J.; Crabtree, R. H; Parnell, C. P.; Uriarte, R. J. Organometallics 1984, 3, 816–817, and references therein for stoichiometric dehydrogenations. (b) Burk, M. J.; Crabtree, R. H. J. Am. Chem. Soc. 1987, 109, 8025–8032. (c) Felkin, H.; Fillebeen-Khan, T.; Holmes-Smith, R.; Lin, Y. Tetrahedron Lett. 1985, 26, 1999–2000. (2) Gupta, M.; Hagen, C.; Flesher, R. J.; Kaska, W. C.; Jensen, C. M. Chem. Commun. 1996, 2083–2084. pubs.acs.org/Organometallics

Published on Web 08/24/2009

Jensen of the very robust and effective pincer-ligated iridium complex (tBu4PCP)IrH2.2,3 This complex was subsequently found to selectively dehydrogenate the terminal position of n-alkanes4 and to catalyze dehydrogenation without the need for a sacrificial hydrogen acceptor.5 In subsequent work the effects of varying substituents at the para position of the aryl group,6,7 substituting O-atoms for the CH2 linkers,8,9 and the replacement of phosphino-tert-butyl (3) Gupta, M.; Hagen, C.; Kaska, W. C.; Cramer, R. E.; Jensen, C. M. J. Am. Chem. Soc. 1997, 119, 840–841. (4) Liu, F.; Pak, E. B.; Singh, B.; Jensen, C. M.; Goldman, A. S. J. Am. Chem. Soc. 1999, 121, 4086–4087. (5) Xu, W.; Rosini, G. P.; Gupta, M.; Jensen, C. M.; Kaska, W. C.; Krogh-Jespersen, K.; Goldman, A. S. Chem. Commun. 1997, 2273–2274. (6) Krogh-Jespersen, K.; Czerw, M.; Zhu, K.; Singh, B.; Kanzelberger, M.; Darji, N.; Achord, P. D.; Renkema, K. B.; Goldman, A. S. J. Am. Chem. Soc. 2002, 124, 10797–10809. (7) Zhu, K.; Achord, P. D.; Zhang, X.; Krogh-Jespersen, K.; Goldman, A. S. J. Am. Chem. Soc. 2004, 126, 13044–13053. (8) (a) G€ ottker-Schnetmann, I.; White, P.; Brookhart, M. J. Am. Chem. Soc. 2004, 126, 1804–1811. (b) G€ottker-Schnetmann, I.; White, P. S.; Brookhart, M. Organometallics 2004, 23, 1766–1776. (c) G€ottker-Schnetmann, I.; Brookhart, M. J. Am. Chem. Soc. 2004, 126, 9330–9338. (9) Morales-Morales, D.; Redon, R.; Yung, C.; Jensen, C. M. Inorg. Chim. Acta 2004, 357, 2953–2956. r 2009 American Chemical Society

Article

groups with phosphino-isopropyl groups have been explored. These catalysts have found application in tandem alkane-dehydrogenation/olefin-metathesis systems, which effect alkane metathesis,10,11 and substrates other than alkanes have been dehydrogenated.12,13 Recently it has been reported that the catalysts can be tethered onto solid supports, particularly alumina, without sacrificing reactivity.14

The presence of sterically bulky, robust, phosphinoalkyl groups (e.g., tBu) could offer protection against cluster formation and bimolecular catalyst deactivation. However, it would seem likely that such groups also strongly contribute to the activation barriers to both C-H bond addition and the requisite β-H elimination of the resulting iridium alkyl intermediate. Thus, these bulky groups afford advantages and disadvantages. We considered, however, that not all four R groups in a fragment (R4PCP)Ir contribute equally to this equation. Both C-H addition and β-H elimination are distinctly unsymmetrical with respect to a (R4PCP)Ir unit, and different tBu groups of the (tBu4PCP)Ir fragment are expected to exert substantially different steric effects on the respective transition states. A simple schematic (Figure 1) would suggest that, in the case of (tBu4PCP)Ir, replacement of even just one tBu group with a sterically much less demanding unit, viz., Me, could substantially favor these reaction steps, while it seemed unlikely that such a substitution would strongly promote the undesirable formation of di- or multinuclear clusters.15,16 Conversely, however, potential resting states could also be stabilized by the decreased crowding resulting from even a single Me-for-tBu substitution. To explore these issues in detail, we have conducted a combined computational/experimental study on the effect of substituting tBu groups on the tBu4PCP ligand by Me groups.17 Although this work is focused on pincer-Ir catalysts for alkane dehydrogenation, we note that the impor(10) Goldman, A. S.; Roy, A. H.; Huang, Z.; Ahuja, R.; Schinski, W.; Brookhart, M. Science 2006, 312, 257–261. (11) Ahuja, R.; Kundu, S.; Goldman, A. S.; Brookhart, M.; Vicente, B. C.; Scott, S. L. Chem. Commun. 2008, 253–255. (12) Morales-Morales, D.; Redon, R.; Wang, Z.; Lee, D. W.; Yung, C.; Magnuson, K.; Jensen, C. M. Can. J. Chem. 2001, 79, 823–829. (13) Zhang, X.; Fried, A.; Knapp, S.; Goldman, A. S. Chem. Commun. 2003, 2060–2061. (14) Huang, Z.; Brookhart, M.; Goldman, A. S.; Kundu, S.; Ray, A.; Scott, S. L.; Vicente, B. C. Adv. Synth. Catal. 2009, 351, 188–206. (15) Pelczar, E. M.; Emge, T. J.; Goldman, A. S. Acta Crystallogr., C 2007, 63, m323–m326. (16) Zhang, X.; Emge, T. J.; Goldman, A. S. Inorg. Chim. Acta 2004, 357, 3014–3018. (17) For a good example of subtle steric modulations resulting in dramatic effects on catalytic activity see: (a) Pool, J. A.; Lobkovsky, E.; Chirik, P. J. Nature (London, U.K.) 2004, 427, 527–530. (b) Pun, D.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc. 2008, 130, 6047–6054. (18) (a) Albrecht, M.; van Koten, G. Angew. Chem., Intl. Ed. 2001, 40, 3750–3781. (b) van der Boom, M. E.; Milstein, D. Chem. Rev. 2003, 103, 1759–1792. (c) The Chemistry of Pincer Compounds; Morales-Morales, D.; Jensen, C., Eds.; Elsevier: Amsterdam, 2007. (d) Benito-Garagorri, D.; Kirchner, K. Acc. Chem. Res. 2008, 41, 201–213.

Organometallics, Vol. 28, No. 18, 2009

5433

Figure 1. Schematic illustration indicating that different tBu groups of the (tBu4PCP)Ir fragment are expected to exert substantially different steric effects on the transition states for C-H addition (left) and β-H elimination (center). The canting of the PCP aryl ring plane relative to the P-Ir-P axis and a general labeling scheme for the phosphine substituents are also illustrated (right).

tance of pincer ligands in catalysis has been recognized much more widely,18-20 and aspects of this study may accordingly be more broadly applicable.21-25

Results and Discussion Computational Studies. Reactions 2-4 in Scheme 1 comprise the established pathway for dehydrogenation of an alkane (specifically of an ethyl group) by (tBu4PCP)Ir to produce (tBu4PCP)IrH2 and alkene;26 the same reactions, but in the reverse order, describe the hydrogenation of a terminal alkene. The two reaction sequences eqs 2-4 and eqs 4-2 therefore constitute a catalytic cycle for transfer dehydrogenation of an alkane using an olefinic sacrificial hydrogen acceptor. Reaction 1 describes the entry into the catalytic cycle from an out-of-cycle alkene-bound resting state. Electronic structure calculations have been conducted for the reactions depicted in Scheme 1 with R0 = C2H5, i.e., for the degenerate transfer dehydrogenation of n-butane using 1-butene as the sacrificial hydrogen acceptor, and different combinations of phosphinoalkyl groups R1-4 (Figure 1). Our DFT27 calculations made use of the PBE combination of exchange and correlation functionals;28 effective core (19) (a) Gatard, S.; Chen, C.-H.; Foxman, B. M.; Ozerov, O. V. Organometallics 2008, 27, 6257–6263. (b) Ozerov, O. V.; Watson, L. A.; Pink, M.; Caulton, K. G. J. Am. Chem. Soc. 2007, 129, 6003–6016, and references therein. (20) (a) Verat, A. Y.; Pink, M.; Fan, H.; Tomaszewski, J.; Caulton, K. G. Organometallics 2008, 27, 166–168. (b) Verat, A. Y.; Fan, H.; Pink, M.; Chen, Y. S.; Caulton, K. G. Chem.-Eur. J. 2008, 14, 7680–7686, and references therein. (21) (a) Kuznetsov, V. F.; Gusev, D. G. Organometallics 2007, 26, 5661–5666. (b) Kuznetsov, V. F.; Lough, A. J.; Gusev, D. G. Inorg. Chim. Acta 2006, 359, 2806–2811. (22) (a) Jiang, Y.; Longmire, J. M.; Zhang, X. Tetrahedron Lett. 1999, 40, 1449–1450. (b) Longmire, J. M.; Zhang, X.; Shang, M. Organometallics 1998, 17, 4374–4379. (23) Morales-Morales, D.; Cramer, R. E.; Jensen, C. M. J. Organomet. Chem. 2002, 654, 44–50. (24) (a) Bedford, R. B.; Betham, M.; Charmant, J. P. H.; Haddow, M. F.; Orpen, A. G.; Pilarski, L. T.; Coles, S. J.; Hursthouse, M. B. Organometallics 2007, 26, 6346–6353. (b) Baber, R. A.; Bedford, R. B.; Betham, M.; Blake, M. E.; Coles, S. J.; Haddow, M. F.; Hursthouse, M. B.; Orpen, A. G.; Pilarski, L. T.; Pringle, P. G.; Wingad, R. L. Chem. Commun. 2006, 3880–3882. (25) Bonnet, S.; Li, J.; Siegler, M. A.; von Chrzanowski, L. S.; Spek, A. L.; van Koten, G.; Klein Gebbink, R. J. M. Chem.-Eur. J. 2009, 15, 3340–3343, and references therein. (26) Renkema, K. B.; Kissin, Y. V.; Goldman, A. S. J. Am. Chem. Soc. 2003, 125, 7770–7771. (27) Parr, R. G.; Yang, W. Density-Functional Theory of Atoms and Molecules; University Press: Oxford, 1989. (28) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865.

5434


Kundu et al.

Scheme 1. Reaction Pathway for Dehydrogenation of an Alkane CH3CH2R0 by (R4PCP)Ir

Table 1. Computed Relative Gibbs Free Energies (kcal/mol) for the Catalytic Cycle Defined in Scheme 1a speciesb

R1,2 = tBu; R1,3 = tBu; R1,4 = tBu; R1 = tBu; R1-4 = R1-4 = R1-3 = tBu; t i Bu Pr R4 = Me R3,4 = Me gem R2,4 = Me trans R2,3 = Me meso R2-4 = Me R1-4 = Me

Ir(butene) þ butane -2.2 Ir þ butene þ butane 0.0 TS: C-H addition þ butene 20.8 Ir(H)(butyl) þ butene 14.4 TS: β-H elimination þ butene 30.6 16.9 Ir(H)2(butene) þ butene Ir(H)(H) þ butene þ butene 3.2 32.7 ΔG = (G highest energy TS) - (G [Ir(butene)] (lowest energy state))c Ir þ H2 f Ir(H)2 -14.3 a

-8.9 0.0 16.7 10.9 23.3 10.5 5.1 32.2

-7.9 0.0 15.8 9.9 20.4 6.4 2.4 28.3

-12.3

-15.0

-10.0 0.0 15.4 8.5 18.7 6.2 3.0 28.7

-10.4 0.0 12.9 8.5 16.8 4.7 2.8 27.3

-11.6 0.0 15.2 8.4 14.6 5.3 3.0 26.8

-13.8 0.0 12.0 7.9 12.3 1.6 2.5 26.1

-15.8 0.0 10.1 5.4 9.1 -1.9 3.3 25.9

-14.4

-14.6

-14.4

-14.9

-14.1

R1-4

For each catalyst species, the sum of the free energies of ( PCP)Ir, n-butane, and 1-butene defines the reference energy of 0.0 kcal/mol. Assumed reaction conditions are T = 423 K (150 °C), [n-butane] = 10 M, and [1-butene] = 1 M. The numbering scheme used for the phosphine alkyl groups is b R1-4 c PCP)Ir. See text for full explanation. shown in Figure 1. Ir = (

potentials replaced the inner electrons of Ir and P;29 valence basis sets of at least split-valence plus polarization quality were assigned to all atoms intimately involved in the catalytic cycle,29,30 and C, H atoms of lesser importance received splitvalence basis sets.30 Reactant, transition state, and product geometries were fully optimized for each hydrocarbon-catalyst combination. Potential energies combined with normal-mode analysis of all stationary points located on the potential energy surfaces allow for the computation of idealized (gas phase) Gibbs free energies for the catalytic cycle (Scheme 1 and Table 1). The columns in Table 1 are ordered left to right according to decreasing Gibbs energy for β-H elimination (eq 3). This step is computed to be rate-determining in the dehydrogenation cycle (Scheme 1) for all investigated (PCP)Ir catalysts, except (meso-tBu2Me2PCP)Ir. With the same single exception of (meso-tBu2Me2PCP)Ir, Table 1 also shows (from left to right) decreasing Gibbs activation energy for the C-H activation step. Substitution of a tBu group for smaller alkyl groups (iPr or Me) only slightly diminishes substituent electron donation to P (electron-donating ability t Bu > iPr > Me) and renders the phosphine slightly more electron withdrawing from the central Ir atom (relative to P(CH2R)tBu2). Such electronic effects would be expected to induce very modest changes in the activation barriers for C-H activation and β-H elimination; moreover, they might be expected to generate small barrier increases as the degree of alkyl substitution is reduced (“less electron-rich” Ir metal center). The general decrease in activation energies observed with decreasing extent of phosphine alkylation may thus be (29) (a) Andrae, D.; Haeussermann, U.; Dolg, M.; Stoll, H.; Preuss, H. Theor. Chim. Acta 1990, 77, 123–141. (b) Bergner, A.; Dolg, M.; Kuechle, W.; Stoll, H.; Preuss, H. Mol. Phys. 1993, 80, 1431–1441. (c) Pietro, W. J.; Francl, M. M.; Hehre, W. J.; DeFrees, D. J.; Pople, J. A.; Binkley, J. S. J. Am. Chem. Soc. 1982, 104, 5039–5048. (30) (a) Ditchfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys. 1971, 54, 724–728. (b) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213–222. (c) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 1980, 72, 650–654. (d) Clark, T.; Chandrasekhar, J.; Spitznagel, G. W.; Schleyer, P. v. R. J. Comput. Chem. 1983, 4, 294–301.

safely attributed to diminishing steric interactions (from left to right, approximately, in Table 1). From the data presented in the first three columns of Table 1 (R1-4 = tBu; R1-4 = iPr; and R1-3 = tBu, R4 = Me) we note that not only are the calculated energetic effects on reactions 1-4 arising from a single Me-for-tBu substitution substantial (5-10 kcal/mol), but also that just a single Mefor-tBu substitution engenders energetic effects generally larger than those resulting from four iPr-for-tBu substitutions. In particular, the TS for β-H elimination is lower by 10 kcal/mol in the (tBu3MePCP)Ir reaction manifold than in the (tBu4PCP)Ir manifold; the computed relative stabilization is about 7 kcal/mol when (iPr4PCP)Ir is the catalyst. The differential energy lowering for the C-H activation step is about half these values: 5 kcal/mol in the case of (tBu3MePCP)Ir and 4 kcal/mol for (iPr4PCP)Ir. Considering the overall potential improvement to the catalytic dehydrogenation process, we find that the significant stabilization of the rate determining β-H transition state is unfortunately (but not unexpectedly) partially offset by the stronger affinity of 1-butene for (tBu3MePCP)Ir and (iPr4PCP)Ir; the calculations predict differential binding energy increases of 6-7 kcal/mol. The penultimate row in Table 1 gives the calculated overall free energy of activation for the respective catalytic cycles, which is equal to that of the highest energy TS (β-H elimination in most cases) in each case, minus the calculated energy of the resting state (the iridium butene complex in all cases). Whereas the predicted overall catalytic free energy of activation in the case of the (tBu3MePCP)Ir catalyst (28.3 kcal/mol) is 4.4 kcal/mol less than that of (tBu4PCP)Ir (32.7 kcal/mol), it is only 0.5 kcal/mol less in the case of (iPr4PCP)Ir (32.2 kcal/mol). It is reassuring, however, that experimental data already reported support an increased efficiency of (iPr4PCP)Ir (relative to (tBu4PCP)Ir) in alkane dehydrogenation.31 A priori there are actually two plausible paths for the approach of the alkane to (tBu3MePCP)Ir, partially illustrated (31) Liu, F.; Goldman, A. S. Chem. Commun. 1999, 655–656.

Article


5435

Figure 2. Illustration (in perspective and depth-fading) of the transition states for C-H activation and β-H migration in the dehydrogenation reaction involving butane and (tBu3MePCP)Ir. Path A has the alkyl chain directed toward the less sterically demanding Me group; path B has the alkyl chain directed toward the sterically much more demanding t-Bu groups and is energetically disfavored by ca. 5 kcal/mol. The primary H and C atoms involved in the reaction are color-coded blue and red, respectively. Hydrogen atoms on aryl and methylene units of tBu3MePCP have been omitted for clarity. Free energy values (ΔG) are relative to free (tBu3MePCP)Ir and n-butane.

by paths A and B in Figure 2. In all the (PCP)Ir fragments, the PCP ligand attains a pseudo-mer-type configuration, and the dehydrogenation reaction takes place in the “horizontal” plane perpendicular to the P-Ir-P pseudoaxis. In this plane, the bulky phosphines create a partially open site trans to the C(PCP) atom; this is where the initial attack (C-H activation) must take place. With the alkane terminus in this opening, the alkyl chain may be oriented in the direction of the group exerting the least steric hindrance (Me), path A, or toward the two tBu groups, path B. Path A represents the lower energy path for both the C-H activation and β-H migration steps, and it is the one for which data are presented in Table 1. Path B has the C-H activation barrier 3.0 kcal/ mol higher than path A, but the difference in β-H migration barriers is more substantial at 5.1 kcal/mol. In the transition states for C-H activation, the alkyl chain remains oriented effectively away from the phosphine alkyl groups, so the energetic preference for path A is not large (Figure 2). However, in the transition state for β-H migration, the β-C and hence the entire alkyl chain is necessarily brought into close vicinity of the Ir atom and the phosphine alkyl groups

(Figure 2); substrate-catalyst steric interactions increase and the orientational preference (path A vs B) is enhanced. Since both reaction paths have the same resting state ((tBu3MePCP)Ir-butene complex), the overall difference in barrier heights between paths A and B is 5.1 kcal/mol. A second methyl group can in principle be introduced into (tBu3MePCP)Ir at three different positions, giving rise to gem (R1,2 = tBu; R3,4 = Me), trans (R1,3 = tBu; R2,4 = Me), and meso (R1,4 = tBu; R2,3 = Me) isomeric structures. At least from the simple perspective suggested by Figure 1, incrementally smaller energetic effects might be expected from the second Me-for-tBu substitution. Indeed, although the computed TS barriers do decrease somewhat relative to (tBu3MePCP)Ir (Table 1), concomitantly the binding energy of the olefin to (tBu2Me2PCP)Ir increases and the effects on the overall catalytic barriers resulting from the second Mefor-tBu substitution are consequently computed to be quite small. We calculate the overall barrier height for the dehydrogenation (Scheme 1) diminished by 1.5 kcal/mol for the (meso-tBu2Me2PCP)Ir catalyst and by only 1.0 kcal/mol for (trans-tBu2Me2PCP)Ir, relative to (tBu3MePCP)Ir; a small

5436


Kundu et al.

Figure 3. Illustration (in perspective and depth-fading) of the transition states for C-H activation and β-H migration in the dehydrogenation reaction involving butane and (meso-tBu2Me2PCP)Ir. Path A has the alkyl chain directed toward the less sterically demanding Me groups; in path B, the alkyl chain is directed toward the sterically more demanding tBu groups, and this reaction path is energetically disfavored by about 6 kcal/mol. The primary H and C atoms involved in the reaction are color-coded blue and red, respectively. Hydrogen atoms on aryl and methylene units of meso-tBu2Me2PCP have been omitted for clarity.

increase of 0.4 kcal/mol in overall barrier height is actually computed in the case of (gem-tBu2Me2PCP)Ir. Considering the intrinsic accuracy of the computations, the magnitudes of the above energies are probably too small to be taken fully at face value. Thus, the calculations suggest that at best only a modest increase in catalytic efficiency should be anticipated from the use of a (meso-tBu2Me2PCP)Ir or (trans-tBu2Me2PCP)Ir catalyst instead of (tBu3MePCP)Ir. Interestingly, C-H activation (not β-H elimination) is predicted to be the rate-determining step for alkane dehydrogenation by (meso-tBu2Me2PCP)Ir, although the difference in TS energies is very small (0.6 kcal/mol, Table 1) and tests the limits of accuracy for the computations. For this catalyst, we have also investigated two possible approaches of the alkane (see Figure 3): path A, where the alkyl chain is directed toward the two Me groups, and path B, in which the alkane orientation is toward the two tBu groups. Not surprisingly, path A is the lower energy path in both the C-H activation and β-H migration steps (data presented in Table 1). The β-H migration step clearly encounters minimal steric hindrance in path A (see Figure 3), which translates into a substantial decrease of almost 8 kcal/mol in barrier height relative to (tBu3MePCP)Ir (Table 1). The computed

C-H activation barrier is 1.6 kcal/mol higher for path B than for path A, but the difference in β-H migration barriers is substantially greater at 7.0 kcal/mol. Since both paths have the same resting state, the overall barrier height difference between paths A and B would be 6.4 kcal/mol. Continuing our computational Me-for-tBu substitution experiments, we find that the overall barriers for n-butane dehydrogenation with putative (tBuMe3PCP)Ir and (Me4PCP)Ir catalysts are virtually identical at 26.1 and 25.9 kcal/mol, respectively, and not noticeably different from that computed for (meso-tBu2Me2PCP)Ir (26.8 kcal/mol). Considering the expected ease of dimer or cluster formation for such sterically unhindered, minimally alkylated (PCP)Ir species, (tBuMe3PCP)Ir and (Me4PCP)Ir do not appear to be useful potential targets for synthesis. In summary, our DFT calculations suggest significantly improved catalytic performance for the single Me-for-tBu substituted (tBu3MePCP)Ir species with perhaps modest additional enhancement arising from a second Me-for-tBu suband/or stitution leading to (meso-tBu2Me2PCP)Ir tBu2Me2 PCP)Ir. (transExperimental Studies. Synthesis and Characterization of (tBu3MePCP)IrH4. Racemic PtBuMeH was synthesized by

Article


5437

Scheme 2. Synthesis of (tBu3MePCP)IrH4 (1)

the reaction, in diethyl ether, of LiAlH4 with PtBuMeCl, prepared as reported by Kolodyazhnyi32 or by the reaction of PMeCl2 with tBuMgCl according to the method of Wolfsberger33 (see Experimental Section). The reaction of excess (5 equiv) dibromo-m-xylene in acetone with tBu2PH gave 1,3-C6H4(CH2PtBu2)(CH2Br) 3 HBr (Scheme 2). After purification, the reaction of this monophosphine salt with PtBuMeH, followed by addition of triethylamine, gave tBu3Me PCP-H. In analogy with the synthesis of the parent (tBu4PCP)IrH4,2,3,7,34 this pincer ligand precursor was then reacted with [Ir(cyclooctadiene)Cl]2 to give pure (tBu3MePCP)IrHCl (two isomers), as characterized by 1H, 31 P, and 13C NMR (see Experimental Section). We were unable to obtain X-ray quality crystals of (tBu3MePCP)IrHCl, perhaps because of the formation of two isomers (presumably diastereomers with the hydride cis and trans to the phosphinomethyl group, respectively). A reaction batch in which a slight excess of [Ir(cyclooctadiene)Cl]2 was (inadvertently) present, however, yielded crystals of (tBu3MePCP)Ir(H)(μ-Cl2)Ir(COD); a single-crystal X-ray diffraction structure of this compound was obtained (Figure 4), revealing the presence of a chloridebridged (tBu3MePCP)IrHCl unit. (tBu3MePCP)IrHCl was treated with LiBEt3H under H2 atmosphere to give (tBu3MePCP)IrH4 (1).2,3,7,34 31P, 1H, and 13 C NMR spectra, as well as elemental analysis, were consistent with the formulation of 1 as (tBu3MePCP)IrH4 (Experimental Section). We were unable to obtain X-ray quality crystals of complex 1. However, in analogy with the known reaction of (tBu4PCP)IrH4,6 addition of 1 atm of CO gave (tBu3MePCP)Ir(CO), eq 5; the single-crystal X-ray diffraction structure of this complex affords supporting evidence of the structure of the (tBu3MePCP)Ir unit (Figure 5). (The structure of the (tBu4PCP)Ir analogue has been previously reported.12)

Catalytic Activity of (tBu3MePCP)IrH4. In accord with the calculated results discussed above, complex 1 was found to (32) Kolodyazhnyi, O. I. Zh. Obshch. Khim. 1981, 51, 2466–2480. (33) Wolfsberger, W. Chem.-Zeitung 1986, 110, 449–450. (34) Goldman, A. S.; Ghosh, R. In Handbook of C-H Transformations-Applications in Organic Synthesis; Dyker, G., Ed.; Wiley-VCH: New York, 2005; pp 616-622, 649-651.

Figure 4. ORTEP diagram of (tBu3MePCP)Ir(H)(μ-Cl2)Ir(COD).

Figure 5. ORTEP diagram of (tBu3MePCP)Ir(CO).

be a significantly more effective catalyst for alkane transfer dehydrogenation than (tBu4PCP)IrH4 or even the presumably much less crowded (iPr4PCP)IrH4 (see Table 2). For example, after 5 h at 150 °C, a solution of 1 (1.0 mM) in n-octane containing 1.14 M norbornene (NBE) gave 979 mM total octenes (979 turnovers) as compared with 96 and 208 mM obtained with the use of (tBu4PCP)IrH4 and (iPr4PCP)IrH4, respectively (entries 8, 9, and 10 in Table 2). The same trend of catalyst activity, 1 > (iPr4PCP)IrH4 > (tBu4PCP)IrH4, was found when tert-butylethylene (TBE), rather than NBE, was used as a sacrificial hydrogen acceptor (Table 3). When 1-hexene was used as the hydrogen acceptor, rates were somewhat slower but the same order of activity among

5438


Kundu et al.

Table 2. Transfer Dehydrogenation of n-Octane Using NBE as the Acceptor (150 °C)a entry

catalyst (1 mM) (ca. initial [NBE]) tBu4

1

( PCP)IrH4 [NBE] = 0.2 M (iPr4PCP)IrH4 [NBE] = 0.2 M

2

(tBu3MePCP)IrH4 [NBE] = 0.2 M

3

(tBu2Me2PCP)IrH4 [NBE] = 0.2 M

4

(tBu4PCP)IrH4 [NBE] = 0.45 M

5

(iPr4PCP)IrH4 [NBE] = 0.45 M

6

(tBu3MePCP)IrH4 [NBE] = 0.45 M

7

(tBu4PCP)IrH4 [NBE] = 1.1 M

8

(iPr4PCP)IrH4 [NBE] = 1.1 M

9

tBu3Me

PCP)IrH4 ( [NBE] = 1.1 M

10

a

time (min)

[NBE]

0 10 20 30 0 10 20 30 0 10 20 0 10 20 30 0 15 30 60 0 15 30 60 0 15 30 60 0 60 300 0 60 300 0 60 300

197 174 168 159 188 120 77 30 191 84 0 164 115 37 0 441 414 401 393 445 366 309 203 457 309 219 32 1146 1099 1047 1128 1037 913 1142 735 153

loss of NBE

total octenes

1-octene

2-transoctene

2-cisoctene

other

23 29 38

23 27 39

11 7 6

10 15 22

2 5 7

0 0 4

68 111 158

68 110 156

19 24 17

33 54 87

13 25 39

3 7 13

107 191

20 13

55 92

22 37

10 48

39 127 164

107 190 0 46 128 146

27 40 48

23 37 46

16 21 21

5 12 18

2 4 5

0 0 2

82 134 242

78 135 238

38 44 46

29 55 109

16 32 64

0 4 19

148 238 425

142 234 424

43 39 20

67 113 176

24 42 77

8 40 151

47 99

47 96

29 45

12 35

6 11

0 5

91 215

83 208

45 52

22 89

16 53

0 14

407 989

408 979

66 46

192 338

73 154

77 441

[catalyst] = 1.0 mM. Product concentrations (mM) measured by GC.

the catalysts was obtained (Table 4). This combination of experimental substrates (n-octane/1-hexene) resembles the computational choice (n-butane/1-butene) most closely. (However, NBE and TBE are generally preferred as acceptors for studies of this type since isomerization of R-olefin acceptors introduces an additional complicating factor.) Pincer-ligated iridium complexes have also been shown to be effective for acceptorless as well as transfer dehydrogenation of alkanes. For the acceptorless dehydrogenation of cyclodecane or n-undecane, (tBu3MePCP)IrH4 was again found to be more effective than (tBu4PCP)IrH4, although the difference was less pronounced than was found for transfer dehydrogenation (Tables 5 and 6). Given the reversibility of this dehydrogenation reaction, we suspect that loss of H2 from solution plays an important role in the kinetics of the observed reaction; in that case the nature of the catalyst is accordingly less important as one approaches equilibrium conditions in solution. In addition, under conditions where the resting state is the dihydride and loss of H2 is ratedetermining,35 catalyst 1 would offer no steric advantage. All three catalysts gave kinetic selectivity for dehydrogenation of n-octane at the terminal position; in all cases an (35) (a) Krogh-Jespersen, K.; Czerw, M.; Summa, N.; Renkema, K. B.; Achord, P.; Goldman, A. S. J. Am. Chem. Soc. 2002, 124, 11404– 11416. (b) Krogh-Jespersen, K.; Czerw, M.; Goldman, A. S. J. Mol. Catal., A 2002, 189, 95–110. (c) Krogh-Jespersen, K.; Czerw, M.; Goldman, A. S. In Activation and Functionalization of C-H Bonds; Goldberg, K. I.; Goldman, A. S., Eds.; ACS Symposium Series 885; American Chemical Society: Washington, DC, 2004; pp 216-233.

initial buildup of 1-octene was followed by isomerization and internal alkenes were soon the major products. In spite of the significant differences in kinetic barriers indicated by the DFT calculations, differences in terminal/internal selectivity are not very great (indicating that these selectivity differences are not primarily due to steric factors). Competition experiments revealed that the kinetic preference for dehydrogenation of n-octane versus cyclooctane was 5.7:1 for (tBu4PCP)Ir versus 2.3:1 for complex 1; this preference exists in spite of the very large thermodynamic effect favoring the dehydrogenation of cyclooctane.36 Synthesis of (tBu2Me2PCP)IrH4. The synthesis and isolation of tBu2Me2PCP precursors and complexes (Scheme 4) is complicated by the fact that meso and (chiral) trans diastereomers are possible. Racemic PtBuMeH was reacted with dibromoxylene, in analogy with the synthesis of the parent tBu4PCP species, to give a mixture of meso- and rac-(tBu2Me2PCP-H), Figure 6. Reaction with [Ir(COD)Cl]2 gave a very complex mixture of products. Crystals obtained from this mixture were characterized by X-ray diffraction and determined to be complex 7; this complex may be viewed structurally as two (meso-tBu2Me2PCP)IrHCl units bridged by a chloride (Figure 7). (36) (a) Roth, W. R.; Lennartz, H. W. Chem. Ber. 1980, 113, 1806– 1817. (b) Doering, W. E.; Roth, W. R.; Bauer, F.; Breuckmann, R.; Ebbrecht, T.; Herbold, M.; Schmidt, R.; Lennartz, H.-W.; Lenoir, D.; Boese, R. Chem. Ber. 1989, 122, 1263–1266.

Article


5439

Table 3. Transfer Dehydrogenation of n-Octane Using TBE as the Acceptor (150 °C)a entry

catalyst (1 mM) (ca. initial [TBE] )

time (min)

tBu4

1

( PCP)IrH4 [TBE] = 0.2 M

0 10 20 30 0 10 20 30 0 10 20 0 10 20 30 0 15 30 60 0 15 30 60 0 15 30 60

(iPr4PCP)IrH4 [TBE] = 0.2 M

2

(tBu3MePCP)IrH4 [TBE] = 0.2 M

3

(tBu2Me2PCP)IrH4 [TBE] = 0.2 M

4

(tBu4PCP)IrH4 [TBE] = 0.45 M

5

(iPr4PCP)IrH4 [TBE] = 0.45 M

6

(tBu3MePCP)IrH4 [TBE] = 0.45 M

7

a

[TBE] 186 153 129 121 200 107 75 43 208 74 10 150 77 0 445 408 391 378 476 371 316 206 475 300 214 22

loss TBE

total octenes

2-transoctene

1-octene

2-cisoctene

other

33 57 65

31 53 60

11 18 15

15 22 28

5 9 13

0 0 4

93 125 157

86 123 152

28 25 19

36 57 76

22 33 42

0 8 15

134 198

126 195

14 6

62 79

26 33

24 77

73 150

76 140

47 64 77

42 56 68

21 22 18

14 23 32

7 11 15

0 0 3

105 160 270

106 156 265

42 38 31

38 66 126

26 45 74

0 6 34

175 261 453

171 252 446

26 21 10

87 121 137

35 51 59

23 59 240


Table 4. Transfer Dehydrogenation of n-Octane Using 1-Hexene as the Acceptor (150 °C)a entry 1

catalyst (1 mM) (ca. initial [1-hex].)

time (min)

[1-hex]

(tBu4PCP)IrH4 [1-hex] = 0.45 M

0 10 15 30 0 10 15 30

464 437 424 342 470 300 264 141

(tBu3MePCP)IrH4 [1-hex] = 0.45 M

2

a

2-trans-hexene

2-cis-hexene

hexane

total octenes

1-octene

2-transoctene

2-cisoctene

10 20 70

4 7 23

9 12 28

7 11 20

6 10 16

1 1 3

0 0 1

81 109 146

23 32 44

52 73 120

48 64 111

40 49 71

6 11 29

2 4 11


The hydride was located at the expected positions that complete the octahedral coordination of the Ir atom and was approximately 0.80(13) e-/A˚3 on the difference Fourier map. Moreover, the geometry of the non-H atoms is more consistent with six-coordinate than five-coordinate (PCP)Ir (all angles in the planes bisecting the P-Ir-P axis are 90° ( 2°). While the structure as shown in Figure 7 would imply a mixed-valence Ir(III)-Ir(IV) complex, the presence of an additional proton, although not located in the refinement process, would lead to assignment as a more typical Ir(III)-Ir(III) complex. The two iridium centers are clearly equivalent chemically, as the corresponding bond distances are all essentially equal. Regardless of the Ir oxidation states, the presence of only methyl groups on one “side” of each monomeric unit allows a side-by-side approach that would not be possible with phosphino-tBu groups (C-C distances are C9-C27 = 3.80 A˚ and C14-C32 = 3.72 A˚). Reduction of the complex mixture with LiBEt3H under H2 in pentane (Scheme 4) gave what appeared to be a single species by 31P NMR (δ33.6 ppm) and 1H NMR (δ -8.80 ppm, t, JPH = 10.2 Hz). The 31P chemical shift is 40 ppm downfield from the free ligand (-7 ppm), which is similar to the difference in chemical shift between

(tBu4PCP)IrH4 (73 ppm) and the respective free ligand (33 ppm). The 1H NMR signal in the hydride region is quite similar to that of (tBu4PCP)IrH4 (δ -9.1 ppm, C6D6, t, JPH = 9.8 Hz). Thus, we can assign this complex as (tBu2Me2PCP)IrH4 (possibly a mixture of meso and trans diastereomers, although it appears as a single species by our NMR methods). Attempts to obtain crystals of this species yielded crystals of 11. Complex 11 may be viewed as the product of (meso-tBu2Me2PCP)Ir and (meso-tBu2Me2PCP)IrH2. From that perspective, the former fragment has inserted into a C-H bond of a phosphinomethyl group of the latter (colored blue in the drawing of Figure 8); in addition, a hydride from the latter fragment bridges the two iridium centers, while a hydrogen of a methyl group of the “(meso-tBu2Me2PCP)Ir” fragment forms an agostic or sigma-complex bridge with the (meso-tBu2Me2PCP)IrH2 unit (Figure 8). The (tBu2Me2PCP)IrH4 synthesized as above catalyzed n-octane transfer dehydrogenation, using either NBE or TBE as an acceptor, with rates greater than that of (tBu4PCP)IrH4 but less than that of (tBu3MePCP)IrH4. Under catalytic conditions NMR spectroscopy reveals a complex mixture of products. The formation of complexes 7

5440


Kundu et al.

and 11 indicates that the reduced steric bulk of the (tBu2Me2PCP)Ir unit (relative to (tBu4PCP)Ir or even (tBu3MePCP)Ir) increases the tendency toward the formation of dinuclear species. Thus, the decreased activity of this catalyst as compared with (tBu3MePCP)Ir is likely due to the formation of such, presumably catalytically inactive, species under catalytic conditions. Given the greater difficulty of isolating pure precursors and the lower catalytic Table 5. Acceptorless Dehydrogenation of Cyclodecane (CDA, bp = 201 °C)a entry

catalyst (1 mM) (ca. initial [TBE]) tBu4

1

(

2

(iPr4PCP)IrH4

3

(tBu3MePCP)IrH4

PCP)IrH4

time (h)

cis-CDE

trans-CDE

DEC

total

1 2 4 9 1 2 4 9 1 2 4 9

49 104 209 325 153 226 301 410 388 550 679 807

49 61 82 72 31 44 62 86 65 111 149 163

3 4 7 11 12 26 37 96 36 75 101 119

101 169 298 408 196 296 400 592 489 736 929 1089

productivity of (tBu2Me2PCP)Ir, the (tBu3MePCP)Ir species seem more promising for alkane dehydrogenation catalysis. We have, however, made progress toward the synthesis of enantiopure (S,S)-(tBu2Me2PCP)IrH4, which will be investigated in the context of asymmetric hydrogenation and dehydrogenation21-25 and reported on independently. Use of the Catalysts in Alkane Metathesis. Brookhart and Goldman have recently reported that alkane dehydrogenation catalysts, acting in tandem with olefin metathesis catalysts, can catalyze the metathesis of n-alkanes.10,11 Catalysts

Figure 6. Three stereoisomers of tBu2Me2PCP.

a [catalyst] = 1.0 mM. Product concentrations (mM) measured by GC. Oil bath temp = 230 °C, DEC = diethylcyclohexane, total = concentration of cis-CDE þ trans-CDE þ DEC.

Table 6. Acceptorless Dehydrogenation of n-Undecane (bp = 196 °C)a entry catalyst (1 mM) 1

(tBu4PCP)IrH4

2

(iPr4PCP)IrH4

3

(tBu3MePCP)IrH4

a

time 2-trans2-cis(h) 1-undecene undecene undecene total 1 2 4 8 1 2 4 8 1 2 4 8

12 17 33 30 10 11 11 12 18 22 30 28

6 9 12 18 6 7 7 8 12 17 21 35

2 3 4 6 2 2 3 3 3 5 6 10

20 29 49 54 18 20 21 23 33 44 57 63

[catalyst] = 1.0 mM. Product concentrations (mM) measured by GC. Oil bath temp = 230 °C.

Figure 7. ORTEP diagram and molecular structure of [(tBu2Me2PCP)IrHCl](μ-Cl), 7.

Scheme 4. Synthesis of (tBu2Me2PCP)IrH4

Article


5441

the electronic structure calculations (Table 1), which predict that such complexes would form the resting state in the catalytic cycle. Complex 12 is observed when 1-alkene is added to solutions of 1 or when other hydrogen acceptors react in the presence of n-alkane, resulting in formation of 1-alkene, in accord with the proposed formulation as (tBu3MePCP)Ir(1alkene). Thus, the reaction of 1 with 1-octene in p-xylene-d10 or COA solvent gives the same species (by 31P and 1H NMR) as observed in n-alkane solvent, which is consistent with lack of direct participation of solvent and the proposed formulation as (tBu3MePCP)Ir(1-alkene). Conversely, the reaction of 1 with hydrogen acceptors that are not 1-alkenes, NBE or TBE, in the absence of n-alkane (in either COA or p-xylened10 solvent) gives species that are distinct from 12 (presumably these are the NBE and TBE adducts, respectively). A solution of 12 was formed under the catalytic conditions described above. Volatiles were removed, and the residue was dissolved in p-xylene-d10. 31P and 1H NMR showed that species 12 was still present. When CO (1 atm) was added to the solution, the formation of (tBu3MePCP)Ir(CO) (31P and 1 H NMR) and free 1-alkene (1H NMR) was observed. Conclusions and Summary. Substitution of methyl groups for one or two of the four tBu groups on the catalyst (tBu4PCP)Ir has been studied, both computationally and experimentally, for its effect on catalytic alkane dehydrogenation activity. DFT calculations predict the rate-determining step in the n-alkane/1-alkene transfer dehydro0 genation cycle to be β-H elimination by (R3R PCP)t t 0 Ir(n-alkyl)(H) (R3R = Bu4 or Bu3Me; modeled with n-butane/1-butene substrates). The electronic structure calculations predict that a single Me-for-tBu substitution has a large favorable energetic effect on this step; the transition state is calculated to be ca. 10 kcal/mol lower for (tBu3MePCP)Ir(n-butyl)(H) than for (tBu4PCP)Ir(n-butyl)(H) (relative to the corresponding free (PCP)Ir fragments). However, this stabilizing effect on the overall barrier is predicted to be offset and reduced to ca. 4 kcal/mol by the stronger binding of 1-butene to (tBu3MePCP)Ir. (tBu3MePCP)IrH4 has been synthesized and isolated, and its catalytic activity has been investigated. In accord with the DFT calculations, (tBu3MePCP)Ir is indeed found to be more active than (tBu4PCP)Ir for catalytic transfer dehydrogenation. Also in agreement with the calculations, the resting state0 in n-alkane/1-alkene transfer dehydrogenation by 0 (R3R PCP)Ir is found to be (R3R PCP)Ir(1-alkene) for both (tBu3MePCP)Ir and (tBu4PCP)Ir. The effects of a second Me-for-tBu substitution are calculated to be less dramatic than the first with respect to catalytic activity for transfer dehydrogenation. Experimentally, (tBu2Me2PCP)IrH4 is found to possess lower catalytic

1 and 10, each acting in tandem with the Schrock olefin metathesis catalyst Mo(NAr)(CHCMe2Ph)(ORF6)2 (Ar = 2,6-i-Pr2C6H3; ORF6 = OCMe(CF3)2), give significantly greater rates than obtained with (tBu4PCP)IrH4 (see Table 7). The initial rates obtained with 1 and 10 are similar to each other, but the productivity using catalyst 10 levels off more quickly, indicative of more rapid decomposition, which is likely related to the reduced steric protection in 10. The 1-Alkene Complex Resting State. The catalytic transfer dehydrogenation of n-octane by 1, using 1-hexene, NBE, or TBE as the hydrogen acceptor, was studied by 31P and 1H NMR under conditions comparable to those employed in the catalytic studies described above but with a higher concentration, 20 mM, of catalyst 1. The NMR spectrum reveals a single major resting state, which is apparently independent of the nature of the hydrogen acceptor. Although we have been unable to isolate the observed complex species, the observations described below strongly indicate that it is the π-bound 1-alkene complex (tBu3MePCP)Ir(1-alkene), complex 12 (either 1-hexene or 1-octene), in accord with the results of

Figure 8. ORTEP diagram and molecular structure of compound 11.

Table 7. Alkane Metathesis of n-Hexane (10 mM Ir, 6.4 mM Mo, 20 mM TBE in n-hexane (7.6 M) at 125 °C); Product Concentrations (mM) Measured by GC entry 1

Ir catalyst (10 mM) tBu4

(

PCP)IrH4

iPr4

2 3

( PCP)IrH4 (tBu3MePCP)IrH4

4

(tBu2Me2PCP)IrH4

time (h)

total

C2

C3

C4

C5

C7

C8

C9

C10

C11

C12

C13

C14

2 8 21 1 3 8 1 2 7

144 391 786 208 574 936 227 236 451

20 40 12 5 11 19 8 7 9

19 39 140 29 78 160 37 39 65

13 31 83 23 62 114 28 30 52

39 93 229 63 154 252 72 80 138

17 49 137 46 120 167 46 54 93

3 8 41 18 50 70 17 17 31

3 10 51 27 70 92 18 20 33

29 114 75 13 38 62 15 18 34

1 4 11 1 3 7 2 3 7

0 1 3 0 2 4 1 1 3

0 1 2 0 1 3 0 0 1

0 2 1 0 1 1 0 0 1

5442


activity than (tBu3MePCP)IrH4 but was also found to be less amenable to characterization. It is therefore not clear if the lower activity is due to the energetics of the catalytic cycle itself or due to facile deactivation of the catalyst through dimerization. Thus, computationally it is predicted that the initial Mefor-tBu substitution on (tBu4PCP)Ir should have a substantial favorable effect on catalytic dehydrogenation activity; this conclusion is supported by experiment. The DFT calculations predict subsequent substitutions to have a smaller effect. This seems to be supported by experiment and reinforced by an increased tendency to form clusters that are presumably inactive.

Experimental Section General Considerations. All reactions were carried out under argon atmosphere using standard Schlenk techniques or in an argon-filled drybox. Anhydrous acetone, hexane, and pentane were purchased from Aldrich and were deoxygenated by purging with argon gas. p-Xylene, p-xylene-d10, and C6D6 were dried using Na/K and collected by vacuum transfer. All solvents (COA, n-octane, n-hexane) were distilled under vacuum from Na/K alloy. NBE was purified by vacuum sublimation. TBE and 1hexene were dried under Na/K alloy and vacuum transferred under argon. Varian instruments (300, 400, or 500 MHz) were used for the 1H, 13C, and 31P NMR experiments. The residual peak of the deuterated solvent was used as a reference for 1H chemical shifts. 31P NMR chemical shifts were referenced to a PMe3 standard, which appears at -62.1 ppm. A PMe3 internal standard in 31P NMR was employed in determining the yield. (tBu4PCP)IrH22,34 and (iPr4PCP)IrH231 (PCP = κ3-2,6-(tBu2PCH2)2C6H3) were prepared as described in the literature. GC analyses were carried out with a Thermal Focus GC with a flame ionization detector (FID) on an Agilent HP-1 column (100% dimethylpolysiloxane, 30 m length 0.32 mm i.d. 0.25 μm film thickness) or Supelco Petrocol DH column (100 m length 0.25 mm i.d. 0.5 μm film thickness). Calibration curves were prepared using standard samples. Synthesis of PtBuMeCl. tert-Butylmethylchlorophosphine was synthesized as reported by Kolodyazhnyi32 (and used in the synthesis of tBu2Me2PCP-H), or by a slight modification of the procedure of Wolfsberger33 (for use in the synthesis of tBu3MePCP-H) as follows. Under argon atmosphere in a 500 mL air-free flask, 5.0 g (42.8 mmol) of dichloromethylphosphine (PMeCl2, Aldrich) was added to 100 mL of anhydrous THF. The solution was then cooled to -78 °C using a dry ice/acetone bath. A tert-Butylmagnesium chloride solution (31.44 mL, 1.36 M in THF; Aldrich) was added slowly for 30 min and stirred for 1 h under argon atmosphere. The cooling bath was then removed, and stirring was continued at room temperature for an additional 1 h. The resulting THF product solution was used for subsequent steps (below), and the 31 P NMR spectrum was taken without filtration or further isolation. 31P{1H} NMR (THF): δ 118.2 (PtBuMeCl) (60% yield) and δ 12.2 (PtBu2Me). Synthesis of PtBuMeH. tert-Butylmethylphosphine was synthesized as per the following scheme. THF

PMeCl2 þ t BuMgCl sf Pt BuMeCl -78 °C, 2 h ðþPt Bu2 MeÞ

LiAlH4 , THF

sf

RT, 12 h

HPt BuMeðþ Pt Bu2 MeÞ

distillation

sf HPt BuMe 70 °C

To 1.63 g (42.8 mmol) of LiAlH4 in 300 mL of anhydrous THF was added dropwise the mixture of PtBuMeCl

Kundu et al. and PtBu2Me (described above) under argon atmosphere. Twelve hours of room temperature stirring gave a mixture of HPtBuMe and PtBu2Me. Pure HPtBuMe was separated out of the mixture by distillation under argon atmosphere. 31 P{1H} NMR (THF, 122 MHz): δ -39.5. Caution: HPtBuMe is pyrophoric and should not be exposed to air. Synthesis of 1,3-C6H5(CH2Br)(CH2PtBu2) 3 HBr (3). To 25 g (85 mmol) of 1,3-bis(bromomethyl)benzene (Aldrich) dissolved in 250 mL of degassed acetone was added 3.2 mL (17 mmol) of HPtBu2 (Aldrich). The mixture was refluxed for 2 h. After the reaction was over, 2 formed a white precipitate and 3 stayed in acetone solution along with excess 1,3-bis(bromomethyl)benzene. This solution was concentrated by removing acetone under vacuum and was added slowly to rapidly stirring Et2O (300 mL). An oily solid formed at the bottom. The solution was stirred well and the supernatant was removed by filtration. The solid was redissolved in acetone and reprecipitated with excess Et2O and was filtered out again. After repeating this procedure five times all 1,3-bis(bromomethyl)benzene was removed in Et2O solution and compound 3 remained as a pure white solid in 85% yield (5.9 g). 31P{1H} NMR (acetone-d6, 122 MHz): δ 34.96 (t, JPD = 72.4 Hz, PtBu2 3 HBr). 1H NMR (acetone-d6, 300 MHz): δ 8.03 (s, 1H, Ar-H), 7.75 (d, JHH = 7.5 Hz, 2H, Ar-H), 7.39 (m, 1H, Ar-H), 4.71 (s, 2H, CH2Br), 4.23 (d, JPH = 13.8 Hz, 2H, CH2P), 1.56 (d, JPH = 16.5 Hz, 18H, PtBu2). Synthesis of tBu3MePCP-H 3 2HBr (4). HPtBuMe (100 mL, 12.5 mmol) in THF was added to 5 g (12.2 mmol) of 3, followed by addition of 100 mL of acetone (to ensure 3 dissolves entirely in the solution). This mixture was refluxed for 2 h under argon atmosphere, and the HBr salt of the ligand (4) precipitated as a white solid (5.5 g). 31P{1H} NMR (D2O, 122 MHz): δ 48.09 (t, JPD = 66.7 Hz, PtBu2, HBr), 27.59 (t, JPD = 73.9 Hz, PtBuMe, HBr). Synthesis of tBu3MePCP-H (5). To 4.0 g of 4 (7.79 mmol) were added 4.5 mL of Et3N (15.6 mmol) and 100 mL of anhydrous hexane. This mixture was stirred for 4 days at room temperature. Afterward the solution was filtered and hexane was removed under vacuum to give 1.22 g (yield: 44.6%) of the ligand precursor 5. 31P{1H} NMR (C6D6, 162 MHz): δ 33.37 (s, PtBu2), -7.27 (s, PtBuMe). 1H NMR (C6D6, 300 MHz): δ 7.46 (s, 1H, Ar-H), 7.3 (d, JHH = 7.5 Hz, 1H, Ar-H), 7.16 (t, JHH = 7.5 Hz, 1H, Ar-H), 7.06 (d, JHH = 7.8 Hz, 1H, Ar-H), 2.82 (dd, JHH = 12.9 Hz, JPH = 4.5 Hz, 1H, CHHPtBuMe) 2.40 (dd, JHH = 12.9 Hz, JPH = 4.5 Hz, 1H, CHHPtBuMe), 2.77 (d, JPH = 2.4 Hz, 2H, CH2PtBu2), 1.08 (d, JHH = 9.6 Hz, 18H, PtBu2), 0.962 (d, JPH = 11.4 Hz, 9H, PtBuMe), 0.785 (d, JPH = 3.9 Hz, 3H, PtBuMe). Synthesis of (tBu3MePCP)IrHCl (6). To 1.0 g (2.84 mmol) of 5 were added 70 mL of toluene and 0.953 g (2.84 mmol) of [Ir(1,5cyclooctadiene)Cl]2 (Strem). This mixture was refluxed for 1 day under hydrogen atmosphere, and the solvent was then removed under vacuum. Compound 6 was extracted from the mixture with pentane, to give 1.57 g of a purple-red solid after removal of pentane (95% yield). NMR indicates a mixture of two isomers in a ratio of ca. 1.1:1 (the species in higher concentration appeared to give slightly broader peaks in both the 1H and 31P NMR spectra). 31P{1H} NMR (p-xylene-d10, 202 MHz): δ 69.98 (d, JPP = 345 Hz, PtBu2), 69.27 (d, JPP = 345 Hz, PtBu2), 48.94 (d, JPP = 342 Hz, PtBuMe), 43.27 (d, JPP = 346 Hz, PtBuMe). 1H NMR (p-xylene-d10, 500 MHz): δ 7.16 (m, 1H, Ar-H), 7.08 (m, 1H, Ar-H), 7.04 (m, 2H, Ar-H), 6.91(m, 2H, ArH), 3.12 (m, 6H, from two isomers, CH2PtBu2 and CHHPtBuMe), 2.69 (m, 2H, from two isomers, CHHPtBuMe), 1.41 (d, JPH = 8 Hz, 6H, from two isomers, PMe), 1.32 (m, 36H, from two isomer, PtBu2), 1.22 (d, tBu, JPH = 14 Hz, 9H, minor isomer, PtBuMe), 1.09 (d, tBu, JPH = 14 Hz, 9H, major isomer, PtBuMe), -36.88 (br, 1H, major isomer, Ir-H), -41.54 (t, JPH = 14.25 Hz, minor isomer, Ir-H). 13C NMR (C6D6, 126 MHz) (isomer with sharper peaks): δ 152.1 (dd, 2JCP = 12.0 Hz,

Article 3

JCP = 4.7 Hz, Ar), 149.8 (dd, 2JCP = 12.4 Hz, 3JCP = 4.2 Hz, Ar), 148.8 (dd, 2JCP = 13.3 Hz, 3JCP = 4.8 Hz, Ar), 123.5 (s, Ar), 122.3 (d, JCP = 16.4 Hz, Ar), 121.1 (d, JCP = 15.6 Hz, Ar), 39.4 (d, JCP = 32.6 Hz, PCH2), 38.5 (dd, JCP = 16.4 Hz, 3 JCP = 3.5 Hz, PC(CH3)3), 35.9 (d, JCP = 29.6 Hz, PCH2), 35.1 (dd, JCP = 18.2 Hz, 3JCP = 4.0 Hz, PC(CH3)3), 32.9 (dd, JCP = 25.6 Hz, 3JCP = 4.2 Hz, PC(CH3)3), 30.2 (dd, 2JCP = 3.6 Hz, 4 JCP = 1.3 Hz, PC(CH3)3), 29.6 (dd, 2JPC = 3.6 Hz, 4JCP = 1.1 Hz, PC(CH3)3), 27.2 (dd, 2JCP = 3.7 Hz, 4JCP = 1.0 Hz, PC(CH3)3), 10.2 (dd, JCP = 26.5 Hz, 3JCP = 2.5 Hz, PCH3). For the other isomer, with slightly broader peaks, the 13 C NMR shows peaks close to those of the isomer with sharper peaks. Synthesis of (tBu3MePCP)IrH4 (1). 6 (0.7 g, 1.2 mmol) was dissolved in 200 mL of pentane. Then 1.2 mL of LiBEt3H (1 M in THF 1.2 mmol) was added slowly at room temperature under H2 atmosphere. The solution changed color from red to light brown, and some precipitate formed. The solution was stirred for 1 day at room temperature under H2 atmosphere, and the precipitate was removed by filtration. After removing the solvent under vacuum, 0.49 g (yield: 74%) of compound 1 was isolated as a brown solid. 31P{1H} NMR (p-xylene-d10, 202 MHz): δ 72.84 (d, JPP = 327 Hz, PtBu2) 33.58 (d, JPP = 327 Hz, PtBuMe). 1H NMR (p-xylene-d10, 400 MHz): δ 7.11 (m, 1H, Ar-H), 7.09 (m, 2H, Ar-H), 3.38 (m, 2H, CH2PtBu2), 3.31 (m, 1H, CHPtBuMe), 3.10 (m, 1H, CHPtBuMe), 1.52 (dd, JPH = 9.5 Hz, JPH = 3 Hz, 3H, PMe), 1.19 (d, JPH = 12.8 Hz, 9H, PC(CH3)3), 1.17 (d, JPH = 12.8 Hz, 9H, PC(CH3)3), 1.05 (d, CH3, JPH = 14 Hz, 9H, PtBuMe), -9.02 (t, JPH = 9.9 Hz, Ir(H)4). 13C NMR (C6D6, 126 MHz): δ 151.9 (m, Ar), 149.52 (dd, 2JCP = 10.8 Hz, 3JCP = 4.7 Hz, Ar), 147.3 (dd, 2JCP = 11.0 Hz, 3JCP = 4.7 Hz, Ar), 123.7 (s, Ar), 121.5 (d, JCP = 15.6 Hz, Ar), 121.2 (d, JCP = 15.5 Hz, Ar), 45.6 (d, JCP = 35.7 Hz, PCH2), 41.7 (d, JCP = 30.5 Hz, PCH2), 33.4 (dd, JCP = 18.7 Hz, 3JCP = 4.0 Hz, PC(CH3)3), 32.8 (dd, JCP = 19.6 Hz, 3JCP = 4.4 Hz, PC(CH3)3), 30.1 (d, JCP = 3.5 Hz, PC(CH3)3), 29.6 (d, JPC = 3.6 Hz, PC(CH3)3), 27.6 (dd, JCP = 28.9 Hz, 3JCP = 4.5 Hz, PC(CH3)3), 26.1 (d, JCP = 4.7 Hz, PC(CH3)3), 18.4 (dd, JCP = 28.1 Hz, 3JCP = 4.3 Hz, PCH3). Anal. Found (calcd): C: 46.22 (46.11), P: 11.30 (11.34), H: 7.61 (7.5). (tBu3MePCP)Ir(CO). To a p-xylene-d10 solution of 5 mg of (tBu3MePCP)IrH4 (1; 9 μmol) in a J-Young tube was added 1 atm of CO. An immediate color change from red to yellow was observed. The solvent was removed and crystals were obtained from the hexane solution. 31P{1H} NMR (p-xylene-d10, 202 MHz): δ 84.2 (d, JPP = 288 Hz, PtBu2), 51.3 (d, JPP = 288 Hz, PtBuMe). 1H NMR (p-xylene-d10, 500 MHz): δ 7.07 (m, 2H, Ar-H), 7.03 (m, 1H, Ar-H), 3.28 (m, 2H, CH2PtBu2), 3.01 (m, 1H, CHPtBuMe), 2.80 (m, 1H, CHPtBuMe), 1.31 (d, JPH = 9.5 Hz, 3H, PMe), 1.23 (d, JPH = 14 Hz, 9H, PtBuMe), 1.08 (t, CH3, JPH = 16.7 Hz, 18H, PtBu2). IR: νCO(C6H6 solution) 1920 cm-1 (cf. 1917 cm-1 measured for (tBu4PCP)Ir(CO) in C6H6 solution). (tBu3MePCP)Ir(1-hexene) (12). To a p-xylene-d10 solution of 5 mg of (tBu3MePCP)IrH4 (9.1 μmol) in a J-Young tube was added 1-hexene (4.5 μL, 36 μmol) at room temperature. After 30 min, solvent was removed under vacuum and the resulting complex was characterized by NMR. 31P{1H} NMR (p-xylene-d10, 202 MHz): δ 64.1 (d, JPP = 353 Hz, PtBu2), 40.8 (d, JPP = 353 Hz, PtBuMe). 1H NMR (p-xylene-d10, 500 MHz): δ 7.29 (d, JHH = 7 Hz, 1H, Ar-H), 7.20 (d, JHH = 7 Hz, 1H, ArH), 7.15 (d, JHH = 7 Hz, 1H, Ar-H), 4.65 (m, 1H, HA, Irhexene), 3.54 (m, 1H, HC, Ir-hexene), 3.24 (m, 4H, CH2PtBu2 and CH2PtBuMe), 2.96 (m, 2H, CH2, Ir-hexene), 2.77 (dd, JHH = 7.5 Hz, JHH = 1.5 Hz (geminal), 1H, HB, Ir-hexene), 1.71 (m, 2H, CH2, Ir-hexene), 1.61 (m, 2H, CH2, Ir-hexene), 1.50 (dd, JPH = 7.5 Hz, JPH = 2.5 Hz, 3H, PMe), 1.28 (d, JPH = 11 Hz, 9H, PtBu), 1.24 (d, JPH = 11 Hz, 9H, PtBu), 1.09 (t, JPH =


5443

7.2 Hz, 3H, CH3, Ir-hexene), 1.05 (d, JPH = 12.5 Hz, 9H, PtBuMe).

(tBu4PCP)Ir(1-hexene). To 0.5 mL of a p-xylene-d10 solution of (tBu4PCP)IrH4 (5 mg, 8.3 μmol) in a J-Young tube was added 1-hexene (4.1 μL, 33.2 μmol) at room temperature. After 30 min solvent was removed under vacuum and the compound was characterized by NMR. 31P{1H} NMR (p-xylene-d10, 202 MHz): δ 59.7 (s, br, PtBu2), 57.1 (s, br, PtBu2). 1H NMR (p-xylene-d10, 500 MHz): δ 7.28 (d, JHH = 8 Hz, 2H, Ar-H), 7.16 (t, JHH = 5.5 Hz, 1H, Ar-H), 4.52 (m, 1H, HA, Ir-hexene), 3.78 (dt, JHH = 12 Hz (cis), JHH = 4.5 Hz, 1H, HC, Ir-hexene), 3.33 (br, 4H, CH2PtBu2), 2.86 (d, JHH = 7.5 Hz, 1H, HB, Ir-hexene), 2.32 (m, 2H, CH2, Ir-hexene), 1.68 (m, 2H, CH2, Ir-hexene), 1.58 (m, 2H, CH2, Ir-hexene), 1.35 (br, 18H, PtBu2), 1.23 (br, 18H, PtBu2), 1.09 (t, JPH = 7.2 Hz, 3H, CH3, Ir-hexene). Synthesis of 1,3-Bis[(tert-butylmethylphosphino)methyl]benzene (tBu2Me2PCP-H) (9). HPtBuMe (1.1 g, 10 mmol) and 1,3-bis(bromomethyl)benzene (1.32 g, 5 mmol) were dissolved in acetone (50 mL). The solution was refluxed overnight, leading to the formation of a pasty solid. Acetone was filtered out by a cannula filter, and the solid was washed with acetone (2 10 mL). The solid was dried and dissolved in the degassed water. To this solution was added dropwise an excess of NaOAc (2.5 g, 30 mmol) solution in degassed water (15 mL). A white precipitate was formed immediately, and the solution was stirred for 30 min. n-Hexane (50 mL) was added to this solution, and the white precipitate dissolved in hexane. The hexane-water mixture was stirred for 10 min; the organic layer was removed by cannula and the aqueous layer was washed with n-hexane (2 20 mL). The combined hexane solution was dried with Na2SO4 and the solvent evaporated under vacuum. The product 9 was an oily colorless liquid (560 mg, 36% yield). 31P{1H} NMR (C6D6, 162 MHz): δ -7.0 (s). 1H NMR (C6D6, 400 MHz): δ 7.04-7.28 (4H, Ar-H), 2.80 (dd, JPH = 12.9 Hz, JHH = 3.9 Hz, 2H), 2.37 (dd, JPH = 12.9 Hz, JHH = 6.9 Hz, 2H), 0.96 (d, JPH = 11.1 Hz, 18H), 0.76 (d, JPH =3.6 Hz, 6H). (R,R)-tBu2Me2PCP and (S,S)-tBu2Me2PCP: 31 P{1H} NMR (C6D6, 162 MHz): δ -7.0 (s). 1H NMR (C6D6, 400 MHz): δ 7.04-7.28 (4H, Ar-H), 2.80 (dd, JPH = 12.9 Hz, JHH = 3.9 Hz, 2H), 2.37 (dd, JPH = 12.9 Hz, JHH = 6.9 Hz, 2H), 0.95 (d, JPH = 11.1 Hz, 18H), 0.77 (d, JPH = 3.6 Hz, 6H). Attempted Synthesis of (tBu2Me2PCP)IrHCl. Compound 9 (465 mg, 1.5 mmol) and [Ir(1,5-cyclooctadiene)Cl]2 (503 mg, 1.5 mmol) were dissolved in toluene (20 mL), and hydrogen was bubbled through this solution. The solution was refluxed under hydrogen atmosphere for 48 h. Toluene was then removed by vacuum, and the residue was extracted with pentane (3 100 mL). The pentane washings were collected and the solvent evaporated under vacuum. A red solid was obtained (661 mg, 82% yield). By NMR data (31P, 1H), the product could be a mixture of several compounds, which were not characterized, except for the dinuclear complex [(tBu2Me2PCP)IrHCl]2(μ-Cl) (7), which was crystallized from the mixture and characterized by X-ray diffraction. Synthesis of (tBu2Me2PCP)IrH4 (10). The crude product obtained from the attempted synthesis of (tBu2Me2PCP)IrHCl described above (537.5 mg or 1.0 mmol assuming a formulation of (tBu2Me2PCP)IrHCl) was dissolved in pentane (250 mL), and hydrogen was bubbled into the solution for 2 h. The solution

5444


turned from red to yellow. A 1 M solution of LiBEt3H in THF (1 mL, 1 mmol) was then added dropwise to the solution under hydrogen atmosphere. This reaction was periodically monitored by 31P NMR until the disappearance of a peak at δ 24 was observed. Then the solution was stirred for another 2 h and filtered with a cannula filter. The filtrate was collected and pentane was removed under vacuum. The product was an orange solid (437 mg, 87% yield). 31P{1H} NMR (C6D6, 162 MHz): δ 32.9 (s). 1H NMR (C6D6, 400 MHz): δ 6.8-7.3 (m, 4H), 3.28-3.36 (dt, JHH = 3.6 Hz, JPH = 16.5 Hz, 2H), 2.99-3.07 (dt, JHH = 4.5 Hz, JPH = 16.2 Hz, 2H), 1.47 (t, JPH = 3.0 Hz, 6H), 0.94 (t, JPH = 6.9 Hz, 18H), -8.80 (t, JPH = 10.2 Hz, 4H). Transfer Dehydrogenation. In an argon-filled glovebox, the iridium complexes (1 μmol, taken from a stock solution (1 M)) were dissolved in n-octane or COA (1 mL) in a flask with a Kontes high-vacuum stopcock and an Ace Glass “adjustable electrode Ace-Thred adapter”, which allows removal of small samples (0.5 μL) with a microliter syringe. 1-Alkene, TBE, or NBE was added to the solution as acceptor. The flask was sealed tightly with a Teflon plug under an argon atmosphere, and the solution was removed from the glovebox and stirred in an oil bath at the specified temperature. Periodically, the flask was removed from the bath and cooled in an ice bath. An aliquot was removed from the flask with a 1 μL GC syringe and analyzed by GC. Turnover numbers were calculated for each aliquot using mesitylene, which was added as a GC standard. Results are summarized in Tables 2-4. Acceptorless Dehydrogenation of Alkanes. A catalyst solution (1.5 mL, 1 mM) was added in an argon atmosphere glovebox to a reactor consisting of a 5 mL round-bottom cylindrical flask fused to a water-jacketed condenser (ca. 15 cm). The top of the condenser was fused to two Kontes high-vacuum valves and an Ace Glass “adjustable electrode Ace-Thred adapter”. The solution was refluxed in an oil bath held at 230 °C (CDA bp = 201 °C or n-undecane bp = 196 °C). Escape of H2 is facilitated by a continuous argon stream above the condenser. Turnover numbers were calculated for each aliquot using mesitylene as GC standard. Results are summarized in Tables 5 and 6. Alkane Metathesis. In the glovebox, Ir catalyst (0.021 mmol), Mo catalyst (10 mg, 0.013 mmol), and TBE (5.4 μL, 0.042 mmol)

Kundu et al. were added to n-hexane (2 mL, 15.3 mmol) containing mesitylene (0.034 M as an internal standard). Two aliquots of this solution (0.5 mL each) were transferred to NMR tubes containing capillaries of PMe3 in mesitylene-d12 for reference and locking. The contents were cooled under liquid nitrogen and sealed under vacuum. The tubes were heated (in parallel) in a preheated oven at 125 °C, and NMR spectra were recorded at regular intervals. 13C{1H} NMR spectroscopy permits the resolution of all n-alkanes in the range C1-C12, although quantification is significantly less precise than is obtained by GC analysis (but improved by the use of inverse gating). No significant differences between spectra of the two aliquots were observed. When NMR did not show any further change in the composition of n-alkanes (23 h), the reaction mixture was analyzed by GC. The seal of one of the tubes was broken, and the solution and headspace were analyzed by GC. Alkane Metathesis: Analysis of Headspace (methane, ethane, and propane). After heating, the contents of the tube were brought to room temperature and the tube was then cooled under liquid nitrogen and shaken repeatedly to equilibrate and dissolve the gaseous products. The seal was then broken and replaced with a septum, and the solution was brought to room temperature. Then 200 μL of the headspace was sampled using a gastight syringe and analyzed by GC. Authentic samples of methane, ethane, and propane were used for calibration in the gas phase (but most of the lighter hydrocarbons observed, including ethane, remained in solution).

Acknowledgment. A.S.G. and K.K.J. thank the Division of Chemical Sciences, Office of Basic Energy Sciences, Office of Energy Research, U.S. Department of Energy, for support of this research. A.S.G., K.K.J., and M.B. thank the NSF for support of this work under the auspices of the Center for Enabling New Technologies through Catalysis (CENTC). Supporting Information Available: Crystal structure data and cif files for (tBu3MePCP)Ir(H)(μ-Cl2)Ir(COD), (tBu3MePCP)Ir(CO), and complexes 7 and 11. This material is available free of charge via the Internet at http://pubs.acs.org.

Rational Design and Synthesis of Highly Active ... - ACS Publications

Recommend Documents