Jaime A.
Flores
a,
Vincent N.
Cavaliere
a,
Dominik
Buck
a,
Balázs
Pintér
a,
George
Chen
b,
Marco G.
Crestani
a,
Mu-Hyun
Baik
*a and
Daniel J.
Mindiola
*a
aDepartment of Chemistry, E Kirkwood Rd, Bloomington, Indiana 47405, USA. E-mail: mindiola@indiana.edu; Fax: +1 812 855 8300; Tel: +1 812 855 2399
bDivision of Chemistry and Chemical Engineering, California Institute of Technology, E California Blvd, Pasadena, California 91125, USA
First published on 2nd June 2011
We demonstrate that a titanium-carbon multiple bond, specifically an alkylidyne ligand in the transient complex, (PNP)TiCtBu (A) (PNP− = N[2-P(CHMe2)2-4-methylphenyl]2), can cleanly activate methane at room temperature with moderately elevated pressures to form (PNP)Ti
CHtBu(CH3). Isotopic labeling and theoretical studies suggest that the alkylidene and methyl hydrogens exchange, either viatautomerization invoking a methylidene complex, (PNP)Ti
CH2(CH2tBu), or by forming the methane adduct (PNP)Ti
CtBu(CH4). The thermal, fluxional and chemical behavior of (PNP)Ti
CHtBu(CH3) is also presented in this study.
In this study we demonstrate that the transient titanium alkylidyne, (PNP)TiCtBu (PNP− = N[2-P(CHMe2)2-4-methylphenyl]2),10 can activate methane at room temperature to form (PNP)Ti
CHtBu(CH3). Isotopic labelling studies revealed that the Ti-methyl hydrogens in (PNP)Ti
CHtBu(CH3) exchange with the alkylidene hydrogen thereby suggesting either a tautomerization and/or abstraction pathway to be operative. Theoretical studies also support either pathway with the abstraction route having a slightly lower barrier.
![]() | ||
Scheme 1 Methane activation by complex 1viaA to form 4 as well as other reactivity of 1 and thermal stability of 4 and 4-d3. |
Exposure of a solution of 1 to CH4 at a pressure of 310 psi at 31 °C afforded the C–H activation product (PNP)TiCHtBu(CH3) (4) in 54% conversion (based on 31P NMR spectra) after 12 h with kavg = 5.3(5) x 10−5s−1 over five independent runs (Fig. 1, Graph A). A similar reaction was observed at 725 psi of CH4 (k = 7.2 × 10−5s−1, 50–60% conversion). Interestingly, conversion of 1 to 4 in 95% yield over 12 h could be achieved by increasing the methane pressure to 1150 psi at 31 °C with k = 7.9 × 10−5s−1. Legzdins observed similar reactivity of the transient butadiene complex, Cp*W(NO)(η2-CH2
CHCH
CH2) with methane at moderately elevated pressures (1025 psi).14 Therefore, varying the pressure of methane does not inflict any significant change on the pseudo-first order rate of decay of 1 (Fig. 1, Graph B), but it affects the yield of 4. Deuteration studies with CD4 at 260 psi verified the formation of the isotopologue, (PNP)Ti
CDtBu(CD3) (4)-d4, reaffirming that C–H activation most likely proceeds via1,2-CH bond addition across the alkylidyne ligand of A. Moreover, when CD4 was used, the average rate did not notably deviate from that observed for the activation of CH4 under similar conditions (kH/kD = 1.2(1), 31 °C, Graph A in Fig. 1), thus reasserting that the slow step in the 1 → 4 transformation precedes methane activation.
![]() | ||
Fig. 1 Kinetic studies for the conversion of 1 to 4 under various pressures of methane. Graph A depicts the pseudo-first order decay of 1 under 310 psi of CH4 and 260 psi CD4 as well as decay of 1-d1 under 310 psi of CH4. Graph B shows the dependence of rate on pressure for the conversion of 1 to 4 at 310, 725, and 1150 psi of CH4. |
From our previous experience in C–H activation, we anticipated that deuterating all α-positions in 1, to form (PNP)TiCDtBu(CD2tBu) (1-d3), should give rise to a KIE of ∼4 at 40 °C since α-hydrogen abstraction is expected to be overall rate-determining.10,12,15 To study the C–H activation step in more detail we prepared the d1 mixture of isotopomers, (PNP)Ti
CHtBu(CHDtBu)/(PNP)Ti
CDtBu(CH2tBu) (1)-d1 by treating 3-d1 with LiCH2tBu,10 so as to slow down the rate of decay of the precursor complex and expedite data collection at moderate pressures of CH4. When 1-d1 was exposed to 310 psi of CH4 pressure, KIE = 2.3 at 31 °C was observed (Fig. 1, Graph A), confirming that the formation of A, and not the intermolecular C–H bond activation step of methane, is rate-determining.
![]() | ||
Fig. 2 Eyring plot for the conversion of 4 to 2-d6 in C6D6 along with a table of rate constants. |
To confirm our spectroscopic characterization of 4, we prepared it independently by salt metathesis of 3 with 0.5 equivalents of freshly prepared Mg(CH3)2(OEt2)x in diethylether, in 97% isolated yield. Diagnostic features of 4 include an alkylidene and methyl resonance at 8.30 and 0.81 ppm, respectively, established by the 1H NMR spectrum collected at rt. The methyl protons (0.81 ppm) exhibit a triplet (3JHP = 3 Hz) due to coupling with the transoid phosphine groups. The phosphine resonances in complex 4 are significantly shifted (32.00 and 21.69 ppm, 2JPP = 43.9 Hz) from those observed in 3 or the degenerate alkylidene-alkyl 1 allowing for convenient monitoring of the reaction mixture by 31P NMR spectroscopy. By 13C NMR spectroscopy the alkylidene carbon resonance in 4 is extremely broad at rt. Thus, we conducted variable temperature multinuclear NMR experiments to resolve the 13C NMR spectrum of 4. Accordingly, the 13C NMR spectrum of 4 at 50 °C displays a more intense signal at 286 ppm (387 Hz) while lowering the temperature of the solution to −80 °C fully resolved this resonance into two distinct peaks, suggesting the presence of two species.13 By applying a combination of HMQC, 45° DEPT, and 1H-coupled 13C NMR experiments, the alkylidene resonances observed in the 13C NMR spectrum collected at −80 °C (299.2 ppm, doublet, 1JCH = 96.3 Hz and 268.8 ppm, doublet, JCH = 76.2 Hz), could be readily correlated to the downfield 1H NMR resonances at 8.90 and 7.90 ppm, respectively. As a result, we propose that complex 4 exists as a mixture of alkylidene isomers (also referred to as rotamers), 4-anti and 4-syn (Scheme 2),17 which isomerize rapidly on the NMR time scale with a low barrier of 10.9 kcal mol−1 at −28.9 °C. Variable temperature 31P NMR spectroscopy also corroborate the formation of two isomers of 4. The 31P NMR spectrum of 4 at −28.9 °C revealed coalescence of the two doublets observed at rt, while further cooling of the NMR solution to −60 °C resolved the broad resonances into two AB couplets, consistent with two distinct C1 symmetric titanium complexes in approximately 3:
2 ratio. Increasing the temperature to rt regenerated the original “averaged” AB set, indicating rapid chemical Ti
C bond rotation with respect to the NMR time scale. Therefore, complex 4, which can be derived from methane activation, exist as a mixture of alkylidene rotamers at rt.
![]() | ||
Scheme 2 Isomers 4-syn and 4-anti detected at low temperature, tautomerization to the methane adduct A-CH44, or methylidene complex 5, and reactivity showing 4 and not 5 is the resting tautomeric form. |
We have found that complex 4 behaves like an alkylidyne synthon analogous to 1. The treatment of 4 with Al(CH3)3 generated the known neopentylidyne-zwitterion (PNP)Ti(μ2-CtBu)(μ2-CH3)(AlMe2)18 concurrent with CH4 elimination. Likewise, the neopentylidyne functionality can also be trapped with NCtBu to render the azametallacyclobutadiene complex (PNP)Ti(CtBuCtBuN)19 along with release of CH4 (Scheme 2). Both complexes were generated quantitatively as established by 31P NMR spectroscopy as well as isolated in yields exceeding 70% without evidence of formation of another species.
![]() | ||
Fig. 3
1H NMR spectra at low field region for the thermolysis of (PNP)Ti![]() |
To corroborate that exchange does take place in 4, the rate of decay of 4-d4 to 2-d6 in C6D6 was also measured. The KIE is higher than that observed for 4-d1 (KIE = 2.06, 80 °C). The difference in rate constants between 4-d1 and 4-d4 must arise from hydrogen exchange (vide infra). As a result of our studies, we propose the exchange phenomenon in complex 4 to occur via a tautomerization, meaning formation of a methylidene complex (PNP)TiCH2(CH2tBu) (5), or by abstraction/addition via the methane adduct (PNP)Ti
CtBu(CH4), A-CH44 (Scheme 2). Our previous work established that the α-hydrogens in (PNP)Ti
CHtBu(CD2tBu) can rapidly exchange at room temperature to a mixture of isotopomers 1-d2 (1
:
2 ratio), so the proposed exchange observed here is not completely unexpected.10,12
![]() | ||
Fig. 4 Computed reaction profile illustrating the competition between methane extrusion (path b) and tautomerization (path a) in complex 4. Values shown in brackets are computed solution-state free energies (ΔGsol) in kcal mol−1. |
At present, we propose part of the difference in energy between 4 and 5 to be imposed by an α-agostic interaction present in the neopentylidene (and absent in the methylidene). This type of interaction can account for 1–10 kcal mol−1, which could explain a difference in energy of 8 kcal mol−1. In fact, the computed Ti–C–H angles of 118.5° and 129.7° and H–C–H angle of 111.8° in 5 indicate an extremely weak α-agostic interaction of the methylidene with the Ti(IV) center. These structural alterations obtained for the optimized geometry computed for 5, are much less pronounced than those observed in stable tungsten and molybdenum methylidene complexes,21 in which more acute M–C–H angles are observed (105–107°). Thus, the data presented here suggests that the α-agostic bond in the titanium-methylidene complex 5 might be playing a minor role. Presently, we do not know what other factors could be accounting for this difference in energy. To better understand why 5 is energetically difficult to generate, we investigated 4-TSaa. Fig. 5 shows the computed structure of 4-TSaa, which adopts a trigonal-bipyramidal geometry with the migrating hydrogen H1 being located halfway between the donor carbon C1 and the acceptor carbon C2 with C–H distances of 1.51 and 1.50 Å, respectively. Interestingly, in 4-TSaa both C1–H3 and C2–H1 bonds display an α-agostic interaction with the titanium center (dTi–H3 = 2.21, ∠Ti–C1–H3 = 90.6°, dTi–H1 = 2.09 Å, ∠Ti–C1–H3 = 83.6°) lowering the energy of this transition state.22 This step can be described as a metal mediated α-hydrogen migration as indicated by the relatively short Ti–H2 distance of 1.72 Å, a Wiberg bond order of 0.18 and Natural Population Analysis (NPA) electron density of 0.71 at H2 in the transition state.
![]() | ||
Fig. 5 Computed structures of 4-TSaa (left) and intermediate A-CH44. Only the most important atoms are shown for clarity (PNP ligand peripherals are omitted). |
As shown in Fig. 4, a conceivable pathway for observed hydrogen exchange in 4 is by way of the alkane sigma-complex, A-CH44. In contrast to the relatively long-lived [(2,6-(tBu2PO)2C5H3N)Rh(CH4)]+ σ-methane complex,23B-CH44, which was recently characterized in solution by Brookhart and Goldberg, A-CH44 must be an extremely labile σ-adduct lying 26 and 10 kcal mol−1 higher in energy than complex 4 and the separated products, respectively.23 In A-CH44 methane occupies one of the equatorial positions of the approximately trigonal-bipyramidal geometry and binds in an η2-C–H fashion resulting in a slightly elongated C1–H1 bond distance of 1.11 Å (Fig. 5) compared to the C–H distance of ∼1.14 Å in B-CH44. The calculated Ti–H1 and Ti–C1 distances of 2.26 and 2.85 Å are significantly longer than those found in B-CH44 (dRh–H = 1.87 Å and dRh–C = 2.38 Å). These longer C–H bonds are in line with the thermodynamics which suggest that methane is very loosely bound in A-CH44. The calculated enthalpy difference of 0.8 kcal mol−1 between the separated products and A-CH44 indicates that the methane extrusion step is almost entirely entropy driven24 and the upper limit of the corresponding barrier can be estimated to be 3–4 kcal mol−1. Intuitively, the rapid, entropy-assisted loss of methane from A-CH44 to yield A25 should prevent the reactivation of the C–H bond. At present however, our isotopic labeling and kinetic data cannot distinguish either pathway for exchange.
The dehydrogenation of CH4 to methylidene is especially important because it could allow for the conversion of methane to an industrially important reagent such as ethylene, via intermolecular “CH2” coupling routes.27 However, at present, this conversion has been proposed for silica-supported Ta(III) alkane metathesis catalysts or metal clusters.4,7,8 The former reaction has been predicted to occur by a series of steps such as σ-bond metathesis of (surface)Ta(H) with methane followed by α-hydride elimination to yield (surface)TaCH2(H).26 However, our work suggests that 1,2-CH bond addition9–11 and tautomerization can also be plausible pathways for the dehydrogenation of methane, forming in the process, a terminal titanium methylidene28 reminiscent of Tebbe's reagent, without necessitating a two-electron redox process.29 Unfortunately, our labeling and kinetic studies do not provide any evidence for secondary KIE taking place in the exchange process.
The possibility of a σ-methane complex, A-CH44, as an intermediate for exchange cannot be discarded, but such a putative species must be relatively long-lived in order to allow for H-exchange rather than replacement with the medium such as benzene. Previous work in our group has established that A activates benzene, via formation of an adduct A-C66H66, which has a long enough lifetime to allow for the determination of the intramolecular equilibrium KIE (C6H6/1,3,5-C6H3D3, 1.33(3)) which is different from the intermolecular KIE (C6H6/C6D6, 1.03(7)).10a These results suggest that binding of the arene is the slowest step in the C–H bond breaking process (but not rate-determining overall).
We estimate the rate of exchange in 4 to be slower than, but comparable to the elimination of methane (10−6 s−1 at 60 °C). The reactivity of 4 with Al(CH3)3 and NCtBu suggests this species to be an alkylidyne synthon, A. Unfortunately, the latter process is encumbered by the irreversible α-hydrogen abstraction to eliminate methane. Current efforts are being devoted to the other volatile paraffins since β-hydrogens might play an essential role.
Footnote |
† Electronic supplementary information (ESI) available: Experimental procedures, spectroscopic data for all new compounds, kinetics data, and theoretical protocols. See DOI: 10.1039/c1sc00138h |
This journal is © The Royal Society of Chemistry 2011 |