Time-resolved infrared (TRIR) study on the formation and reactivity of organometallic methane and ethane complexes in room temperature solution

Abstract

We have used fast time-resolved infrared spectroscopy to characterize a series of organometallic methane and ethane complexes in solution at room temperature: W(CO)₅(CH₄) and M(η⁵C₅R₅)(CO)₂(L) [where M = Mn or Re, R = H or CH₃ (Re only); and L = CH₄ or C₂H₆]. In all cases, the methane complexes are found to be short-lived and significantly more reactive than the analogous n-heptane complexes. Re(Cp)(CO)₂(CH₄) and Re(Cp*)(CO)₂(L) [Cp* = η⁵C₅(CH₃)₅ and L = CH₄, C₂H₆] were found to be in rapid equilibrium with the alkyl hydride complexes. In the presence of CO, both alkane and alkyl hydride complexes decay at the same rate. We have used picosecond time-resolved infrared spectroscopy to directly monitor the photolysis of Re(Cp*)(CO)₃ in scCH₄ and demonstrated that the initially generated Re(Cp*)(CO)₂(CH₄) forms an equilibrium mixture of Re(Cp*)(CO)₂(CH₄)/Re(Cp*)(CO)₂(CH₃)H within the first few nanoseconds (τ = 2 ns). The ratio of alkane to alkyl hydride complexes varies in the order Re(Cp)(CO)₂(C₂H₆):Re(Cp)(CO)₂(C₂H₅)H > Re(Cp*)(CO)₂(C₂H₆):Re(Cp*)(CO)₂(C₂H₅)H ≈ Re(Cp)(CO)₂(CH₄):Re(Cp)(CO)₂(CH₃)H > Re(Cp*)(CO)₂(CH₄):Re(Cp*)(CO)₂(CH₃)H. Activation parameters for the reactions of the organometallic methane and ethane complexes with CO have been measured, and the ΔH^‡ values represent lower limits for the CH₄ binding enthalpies to the metal center of WCH₄ (30 kJ·mol⁻¹), MnCH₄ (39 kJ·mol⁻¹), and ReCH₄ (51 kJ·mol⁻¹) bonds in W(CO)₅(CH₄), Mn(Cp)(CO)₂(CH₄), and Re(Cp)(CO)₂(CH₄), respectively.

There is considerable interest in sigma-bonded organometallic alkane complexes, particularly since they have been identified as key intermediates in the transition metal-mediated CH activation process (1–3). Although such complexes generally are very short-lived intermediates (4), they have been known for over 30 years. Early experiments involved the photolysis of complexes such as Cr(CO)₆ and Fe(CO)₅ to generate the unstable intermediates Cr(O)₅ or Fe(CO)₄ in low-temperature matrices, where coordination to cocondensed CH₄ results in the formation of Cr(CO)₅(CH₄) and Fe(CO)₅(CH₄) (5, 6).

Flash photolysis experiments have demonstrated that the photolysis of Cr(CO)₆ in cyclohexane solution at room temperature forms Cr(CO)₅(cyclohexane) (7). Subsequently, several examples of alkane complexes in solution have been reported, and studies on the mechanism of the CH activation process have clearly demonstrated the role of these complexes in oxidative addition reactions (1, 8, 9). Time-resolved infrared (TRIR) spectroscopy has proved to be a powerful tool for the study of metal carbonyl alkane complexes. Their reactivity decreases on going both across and down groups 5, 6, and 7 (10–12), and these observations led to the identification of a very long-lived alkane complex, Re(Cp)(CO)₂(n-heptane) (Cp = η⁵C₅H₅), which has a lifetime of ≈25 ms at room temperature (13). The relative stability of Re(Cp)(CO)₂(alkane) complexes allowed Re(Cp)(CO)²(C₅H₁₀) to be observed at 180 K by NMR spectroscopy (14), and subsequent NMR studies have been carried out to determine the binding modes of a series of related alkanes to the Re(Cp′)(CO)² moiety (15, 16).

The activation of methane is of particular interest because of the potential of using this abundant hydrocarbon as both an energy source and chemical feedstock (17). Organometallic methane complexes have been characterized in low-temperature matrix isolation experiments (5, 6, 18). In solution, the existence of methane complexes has been inferred by examining product ratios and from the rates of CH activation and reductive elimination reactions in isotopic labeling experiments (19).

The lifetime of the alkane complexes depends strongly on the nature of the alkane, with cyclic alkanes tending to form longer-lived complexes than the corresponding linear alkanes do (12). Metal-alkane binding enthalpies have been determined by photoacoustic calorimetry (PAC) and TRIR spectroscopy for a range of group 6 [M(CO)₅(heptane)] and group 7 [M(Cp)(CO)²(heptane)] complexes in solution, and these are found to range between 40 and 57 kJ·mol⁻¹ (10, 20–24). Gas-phase TRIR studies have demonstrated that the shorter n-alkane species have progressively lower binding energies to the metal, with CH₄ being the most weakly bound (25). The failure to identify W(CO)₅(CH₄) by TRIR in the gas phase led to the conclusion that the CH₄ binding energy to the W(CO)₅ fragment is <21 kJ·mol⁻¹ (25). TRIR measurements on Cr(CO)₆ in the gas phase demonstrated that photolysis produces Cr(CO)₅. In the presence of CH₄ (600 torr; 1 torr = 133 Pa), a 2-fold decrease in the rate of regeneration of Cr(CO)₆ was interpreted to indicate a weak interaction between the Cr(CO)₅ fragment and CH₄, with an estimated binding enthalpy for CrCH₄ of 33 ± 8 kJ·mol⁻¹ (26). We recently have used a combination of supercritical fluids and TRIR spectroscopy to characterize a series of organometallic noble gas complexes at room temperature (11), leading to the characterization of Re(ⁱPrCp)(CO)(PF₃)Xe by NMR (27). We have used a similar approach to characterize an organometallic methane complex in room temperature solution after the photolysis of Fe(CO)₅ in supercritical methane (scCH₄), where we observed ³Fe(CO)₄ decaying into ¹Fe(CO)₄(CH₄) with k_obs = 5 ± 1 × 10⁶ s⁻¹ (28).

A number of theoretical studies on the coordination and activation of methane at metal centers also have been reported (29–33). Two of these studies are of particular interest to this work: (i) a study on the binding of CH₄ to W(CO)₅, which predicted the WCH₄ binding enthalpy to be ≈18 kJ·mol⁻¹ (30) and (ii) a paper on the interaction of CH₄ with the Re(Tp)(CO)² [Tp = Tris(pyrazolyl)borate] and Re(Cp)(CO)² moieties, where the ReCH₄ binding enthalpy was calculated to be 36 kJ·mol⁻¹ for the formation of Re(Cp)(CO)²(CH₄) from Re(Cp)(CO)² and CH₄ but with a low barrier to CH activation (33).

In this article, we investigate the formation and reactivity of a series of organometallic methane and ethane complexes in solution at room temperature and provide a detailed study of an organometallic methane complex formed under these conditions. We have carried out a TRIR study of both group 6 and 7 transition metal carbonyls in scCH₄ and liquid ethane (liqC²H₆). A related paper by Ball et al. (34) in this issue reports an investigation on the photolysis of Re(Cp)(CO)²(PF₃) in higher alkanes by TRIR and NMR, which shows a delicate balance between the coordination and activation of the alkane. Here, we also observe a balance between the coordination and activation of CH₄ and C²H₆ after interaction with Re(η⁵C₅R₅)(CO)² fragments (R = H or CH₃).

Results and Discussion

TRIR Studies of W(CO)₆ Photolysis in scCH₄.

The TRIR difference spectrum obtained 1 μs after photolysis (355 nm) of W(CO)₆ (1) in scCH₄ (4,000 psi, 296 K) in the presence of CO (17 psi) is shown in Fig. 1b. It is clear that the parent ν(CO) band at 1,991 cm⁻¹ is bleached, and two transient bands are formed at 1,968 and 1,942 cm⁻¹. In the presence of CO, these transient bands decay at the same rate (k_obs = 4.4 ± 0.1 × 10⁵ s⁻¹) as the parent band reforms. By comparison to previous matrix isolation (5) and TRIR experiments (35), these transient bands can be assigned to W(CO)₅(CH₄) (2). The observed rate constant for the decay of 2 depends linearly on the CO concentration, affording the second-order rate constant for the reaction of 2 with CO in scCH₄ (k_CO = 7.8 ± 0.8 × 10⁶ dm³·mol⁻¹·s⁻¹, 298 K). W(CO)₅(CH₄) is significantly more reactive (3–15 times) than other previously reported W(CO)₅(alkane) complexes (12). However, the rate of the alkane-CO substitution still is significantly below the diffusion-controlled rate in scCH₄ (≈8 × 10¹¹ dm³·mol⁻¹·s⁻¹, estimated from the Stokes–Einstein equations). We have carried out experiments in scAr to provide proof of the coordination of methane to W(CO)₅ in solution. Photolysis of 1 in scAr (5,100 psi, 296 K) under a low concentration of CO (3 psi) generates two transient bands at 1,969 and 1,941 cm⁻¹, which rapidly decay (k_obs = 1.2 ± 0.6 × 10⁷ s⁻¹) (Fig. 1c). Organometallic Ar complexes have been identified in low-temperature matrices (5), and theoretical studies have predicted that Ar binds to W(CO)₅ (36). We tentatively assign these bands to W(CO)₅Ar; however, further work is required to substantiate the nature of this transient species. When the experiment is repeated in the presence CH₄ (200 psi), the transient bands are observed at 1,945 and 1,970 cm⁻¹, and they decay on a much longer time scale (k_obs = 9.0 ± 2.0 × 10⁵ s⁻¹) (Fig. 1d). The increased lifetime of the transient in scAr doped with CH₄ indicates that 2 is formed in scAr. This work provides a detailed study of the reactivity of an organometallic methane complex (2) in solution at room temperature.

Fig. 1.

IR spectra and TRIR decay traces of W(CO)₅(L) (L = Ar, CH₄) in supercritical fluid solutions. (a) FTIR of W(CO)₆ in scCH₄ (4,000 psi, 296 K) and CO (17 psi). (b) TRIR difference spectrum of the same solution 1 μs after photolysis (355 nm). (c and d) Decay traces of the transient (1,969 cm⁻¹) after photolysis of W(CO)₆ in scAr (5,100 psi) with CO (3 psi) (c) and scAr (4,900 psi) with CH₄ (200 psi) and CO (3 psi) (d).

We have investigated the coordination of CH₄ to W(CO)₅ further by determining the activation parameters for the reaction of W(CO)₅(CH₄) with CO in scCH₄ from variable-temperature TRIR experiments. The interpretation of activation parameters for the ligand substitution reaction of an alkane complex with CO requires great care, particularly because the mechanism could be considered to be associative, dissociative, or interchange in nature. The reaction of 2 with CO in scCH₄ is faster (3.4 times) than that of the analogous n-heptane complex. However, the value of ΔH^‡ for the reaction of 2 with CO is found to be greater that that of the corresponding reaction for the n-heptane complex (Table 1). The ΔH^‡ value (30 ± 2 kJ·mol⁻¹) for the reaction of 2 with CO can be used as a lower limit for the WCH₄ binding enthalpy in scCH₄. It is interesting to compare this result with previous calculations (30), which predict a WCH₄ interaction of 18 kJ·mol⁻¹, and with gas-phase TRIR experiments, which estimate a CrCH₄ bond-dissociation energy of 33 ± 8 kJ·mol⁻¹ (26). Clearly, there is a significant WCH₄ interaction in solution at room temperature. We previously have shown that Re(Cp)(CO)²(heptane) is ≈700 times longer-lived than W(CO)₅(heptane) is (13). In an attempt to produce long-lived complexes of the lighter alkanes, we have investigated the photochemistry of M(Cp)(CO)₃ (M = Mn or Re) in scCH₄ and liqC²H₆.

Table 1.

Rate constants and activation parameters for the reaction of a series of organometallic alkane complexes with CO in supercritical (for CH₄) and liquid (for other alkanes) solutions

Complex	kCO, dm3·mol−1·s−1 at 298 K	ΔH‡, kJ·mol−1
W(CO)5(n-C7H16)†	2.3 (±0.2) × 106	20 ± 2
W(CO)5(CH4) (2)	7.8 (±0.2) × 106	30 ± 2
CpRe(CO)2(n-C7H16)§	2.5 (±0.2) × 103	46 ± 2
CpRe(CO)2(C5H10)§	1.2 (±0.2) × 103	32 ± 2
CpMn(CO)2(n-C7H16)§	8.0 (±0.6) × 105	36 ± 2
CpMn(CO)2(C5H10)§	2.5 (±0.4) × 105	24 ± 2
CpMn(CO)2(CH4) (4)	1.6 (±0.3) × 107	39 ± 2
CpMn(CO)2(C2H6) (5)	2.6 (±0.2) × 106	36 ± 2
CpRe(CO)2(CH4) (7)	2.7 (±0.4) × 104	51 ± 5
CpRe(CO)2(C2H6)¶	6.9 (±0.5) × 103	43 ± 2
Cp*Re(CO)2(CH4) (10)	5.0 (±0.3) × 104	47 ± 2
Cp*Re(CO)2(C2H6) (11)	2.6 (±0.1) × 104	43 ± 2

^†Ref. 12.

^§Ref. 38.

^¶Ref. 41.

TRIR Studies of Mn(Cp)(CO)₃ in scCH₄ and liqC₂H₆.

The TRIR difference spectra after photolysis (355 nm) of Mn(Cp)(CO)₃ (3) in scCH₄ (4,000 psi, 298 K) and liqC₂H₆ (1,600 psi, 297 K) both containing CO (60 psi) are shown in Fig. 2. In both cases, the parent is bleached, and two new transient bands at lower energy are produced. These transients are assigned to Mn(Cp)(CO)₂(CH₄) (4) and Mn(Cp)(CO)₂(C₂H₆) (5) in scCH₄ or liqC₂H₆, respectively, by comparison to known Mn(Cp)(CO)₂(alkane) complexes (37, 38). These transients decay in the presence of CO to reform 3 (for 4, k_obs = 2.5 ± 0.2 × 10⁶ s⁻¹ and k_CO = 1.6 ± 0.3 × 10⁷ dm³·mol⁻¹·s⁻¹; for 5, k_obs = 4.9 ± 0.1 × 10⁵ s⁻¹ and k_CO = 2.6 ± 0.2 × 10⁶ dm³·mol⁻¹·s⁻¹, all at 298 K).^§ These ethane and methane complexes also are significantly more reactive (3–20 times) toward CO than the analogous longer-chain n-heptane complexes are (Table 1).

Fig. 2.

TRIR difference spectra of Mn(Cp)(CO)₃ photolysis (355 nm) in scCH₄ (4,000 psi) with CO (60 psi) at 298 K (a) and C₂H₆ (1,600 psi) with CO (60 psi) at 297 K (b). The difference spectra in a and b were recorded at 120 ns and 3 μs, respectively, after photolysis.

The activation parameters for the reaction of 4 and 5 with CO to reform 3 have been investigated. The ΔH^‡ value determined for 4 (39 ± 2 kJ·mol⁻¹) again provides an estimate of the lower limit of the MnCH₄ binding enthalpy. Mn(Cp)(CO)₂(C₂H₆) (5) is found to be significantly less reactive than 4 is to CO substitution, indicating an increased binding strength of C₂H₆ to manganese compared with CH₄. However, the smaller value of ΔH^‡ (36 ± 2 kJ·mol⁻¹) suggests that this value is a considerable underestimate of the MnC₂H₆ binding enthalpy in 5. In principle, the values of ΔS^‡ from our activation parameters would shed light on whether this underestimate is attributable to changes in the reaction mechanism that might be occurring. However, interpretation of such values is problematic because of the potentially large errors in determining the ΔS^‡ values and possible entropy/enthalpy compensation effects (40).

Photolysis of Re(Cp)(CO)₃ in scCH₄.

The fragment Re(Cp)(CO)₂ produces long-lived alkane complexes with the higher alkanes, and we previously have reported the formation of Re(Cp)(CO)²(C²H₆) in a room temperature solution (41). The TRIR spectrum of Re(Cp)(CO)₃ (6) in scCH₄ (4,000 psi, 294 K) in the presence of CO (90 psi) obtained 20 μs after the laser flash shows that the parent ν(CO) bands at 1,951 and 2,038 cm⁻¹ are bleached, and four new transient bands are produced at 2,012, 1,963, 1,942, and 1,902 cm⁻¹ (Fig. 3). These transient peaks cannot be assigned to a single mononuclear CO loss intermediate as such a species cannot have more than two ν(CO) bands. The production of four new transient bands may indicate that more than one primary photoproduct is produced. In the presence of CO, all four bands decay at the same rate as the parent bands reform (k_obs = 6.41 ± 0.08 × 10³ s⁻¹).^¶ In the absence of CO, Re(Cp)(CO)²(alkane) is known to react with Re(Cp)(CO)₃ to form Re²(Cp)²(CO)₅ (42, 43). The new ν(CO) bands observed in this study are not attributable to this dimeric species. The two bands at 1,902 and 1,963 cm⁻¹ can be assigned to the solvated dicarbonyl species Re(Cp)(CO)²(CH₄) (7) by comparison to previous low-temperature IR and room temperature TRIR experiments (38, 44). The bathochromic shift in the ν(CO) bands compared with the parent species is consistent with loss of the π-acceptor ligand. Furthermore, the two bands at 1,942 and 2,012 cm⁻¹ have undergone a hypsochromic shift in comparison with 7, which is consistent with loss of a CO ligand and oxidation of the metal centre. This finding leads us to the tentative assignment of the bands at 1,942 and 2,012 cm⁻¹ to the CH activated species Re(Cp)(CO)²(CH₃)H (8). The activation of CH bonds by the related fragment Re(Cp*)(CO)² previously has been observed in the study of fluorinated arenes (45). The assignment of 7 and 8 is supported by density functional theory (DFT) calculations. We have carried out similar calculations to those previously reported (33) and have computed the IR frequencies of the ν(CO) vibrations. We have compared these gas-phase calculated values to the experimentally obtained band positions Re(Cp)(CO)₃ (TRIR: 1,951 and 2,038 cm⁻¹; DFT: 2,023 and 2,101 cm⁻¹), Re(Cp)(CO)²(CH₃)H (TRIR: 1,942 and 2,012 cm⁻¹; DFT: 2,014 and 2,078 cm⁻¹), and Re(Cp)(CO)²(CH₄) (TRIR: 1,902 and 1,963 cm⁻¹; DFT: 1,979 and 2,034 cm⁻¹). Although the absolute values are not comparable, there is a very good agreement between the trends in the calculated values and those observed in the TRIR experiments. We have not observed the ν(MH) band of 8, which would be expected to be present in the region of study, and as expected our DFT calculations have indicated that the ν(MH) bands of 8 are much lower in intensity than the ν(CO) bands. The observation of both the transient species 7 and 8 is in contrast to the photolysis of 3 in scCH₄ which yielded only 4. It is interesting to compare these results with the study of the interaction of dihydrogen with the M(Cp)(CO)² moiety. A nonclassical dihydrogen complex, Mn(Cp)(CO)²(η²H²), was observed for manganese, and for rhenium, oxidative addition of dihydrogen led to formation of the dihydride complex (46). Similarly, the reaction of Re(Cp)(CO)² with triethylsilane (SiEt₃H) forms Re(Cp)(CO)²(SiEt₃)H, whereas reaction with the analogous manganese complex generates a η²SiH complex (47).

Fig. 3.

IR spectra of Re(Cp)(CO)₃ in scCH₄. (a) FTIR spectrum of Re(Cp)(CO)₃ in scCH₄ (4,000 psi) with CO (90 psi) at 294 K. (b) TRIR difference spectra of same solution 20 μs after photolysis (266 nm).

In our experiments, the fact that 7 and 8 decay at the same rate suggests that these two species are in rapid equilibrium. We have varied both the temperature and the concentration of CO in scCH₄ solution, and accordingly, under all measured conditions, both sets of bands decay at the same rate. The observed rate constant for the decay of 7 and 8 increased linearly with CO concentration (k_CO = 2.7 ± 0.4 × 10⁴ dm³·mol⁻¹·s⁻¹, 298 K). For a system at equilibrium, it may be possible to vary product distribution by varying the temperature. Variable temperature experiments over a small temperature range (293–333 K) showed no detectable difference in product distribution larger than the signal noise, which indicates that the energetic difference between 7 and 8 is relatively small, and it may only be possible to detect such a difference over a large temperature range. Attempts to increase the temperature range of the experiments were hampered either by the high pressures involved or at low temperature by poor solubility. The assignment of a fast equilibrium between 7 and 8 is consistent with previous theoretical work (33) in which it was calculated that there was a small barrier to CH activation of CH₄ in 7 and the difference in enthalpy between 7 and 8 was reported to be only ≈4 kJ·mol⁻¹. The presence of a rapid thermal equilibrium between a sigma-complex and an oxidative addition product previously has been observed for the addition of dihydrogen to the Nb(Cp)(CO)₃ moiety in which both Nb(Cp)(CO)₃(η²H²) and Nb(Cp)(CO)₃H² were observed on the nanosecond time scale (10).

Photolysis of Re(Cp*)(CO)₃ in scCH₄.

We have investigated the interesting equilibrium between methane and methyl hydride complexes further by repeating these measurements using Re(Cp*)(CO)₃ (9), where there is increased electron density on the rhenium metal center. Photolysis (266 nm) of 9 was studied in solutions of both methane (4,000 psi, 298 K) and ethane (1,660 psi, 294 K), both containing CO (90 and 60 psi, respectively). The TRIR difference spectrum in scCH₄ showed a ν(CO) band pattern similar to that seen for the photolysis of Re(Cp)(CO)₃ (6), indicating the presence of both Re(Cp*)(CO)²(CH₄) (10) and Re(Cp*)(CO)²(CH₃)H (11) (Fig. 4). The complex trans-Re(Cp)(CO)²(CH₃)H previously has been synthesized, allowing us to assign 11 to cis-Re(Cp)(CO)²(CH₃)H (48). These results can be compared directly with the photolysis of 6, and there is a noticeable increase in the proportion of the oxidative addition product 11 relative to the organometallic methane complex 10.

Fig. 4.

TRIR difference spectra after photolysis of indicated species. Experiments were carried out in the presence of CO with 1,660 psi ethane or 4,000 psi methane. Lorentzian fit lines of the experimental data are shown. Red shaded areas indicate ν(CO) bands of the alkane complexes, Re(Cp)(CO)₂(C_nH_2n+2), and blue areas indicate the CH activated complexes, Re(Cp)(CO)₂(C_nH_2n+1)H. The Re(Cp)(CO)₃ in ethane spectrum is adapted from ref. 41.

Photolysis of 6 in liqC²H₆ only generates detectable amounts of Re(Cp)(CO)²(C²H₆) (41), whereas photolysis of 9 in ethane solution led to the formation of both Re(Cp*)(CO)²(C²H₆) (12) and a low concentration of Re(Cp*)(CO)²(C²H₅)H (13). For 10/11 and 12/13, the four photoproduct bands for each solution decayed at indistinguishable rates, indicating the presence of rapid equilibria. The rates of reaction of CO with 10/11 and 12/13 were found to be significantly faster than those seen after the photolysis of 6 (10/11: k_CO = 5.0 ± 0.3 ×10⁴ dm³·mol⁻¹·s⁻¹; 12/13: k_CO = 2.7 ± 0.1 ×10⁴ dm³·mol⁻¹·s⁻¹, both at 298 K). The faster rate for the reaction of 10/11 with CO relative to the corresponding reaction of 7/8 is consistent with a previous study on the lifetime of Mn(η⁵C₅R₅)(CO)²(n-heptane) (R = H or CH₃) toward small molecules in which faster rates of ligand substitution were observed for R = CH₃ (49).

We also have monitored the photolysis of 9 in scCH₄ by using picosecond TRIR (ps-TRIR) to investigate the formation of the equilibrium between alkane and alkyl hydride complexes (Fig. 5). This system was chosen because of the increased concentration of the CH activated species, which allowed for the direct monitoring of the establishment of the equilibrium between 10 and 11. Initially after photolysis, the parent ν(CO) bands are bleached, and two strong transient bands at 1,943 and 1,883 cm⁻¹ corresponding to Re(Cp*)(CO)²(CH₄) (10) are apparent (Fig. 5b). Also present at early times are additional ν(CO) bands (2,010, 1,986, and 1,919 cm⁻¹), which decay rapidly (k_obs = 1.7 ± 0.3 × 10¹⁰ s⁻¹) as the parent species partially reforms. Previous studies on the photolysis of Re(Cp)(CO)₃ in silane solutions also have observed similar short-lived species, which were assigned to the decay of the vibrationally excited ground-state molecule (47). The ν(CO) bands of 10 then are observed to decay partially, concurrent with the rise of the bands of Re(Cp*)(CO)²(CH₃)H (11) at 1,998 and 1,923 cm⁻¹, yielding an approximate rate for the activation of CH₄ and establishment of the equilibrium between 10 and 11 (k_obs = 5 ± 2 × 10⁸ s⁻¹) (Fig. 6).

Fig. 5.

ps-TRIR spectra showing the formation of Re(Cp)(CO)₂(CH₃)H. (a) FTIR of Re(Cp*)(CO)₃ in scCH₄ (4,000 psi) with CO (90 psi) at ≈298 K. (b–d) TRIR difference spectra of the same solution at 50 ps (b), 500 ps (c), and 2,000 ps (d) after photolysis (267 nm).

Fig. 6.

Kinetic traces of ν(CO) bands of Re(Cp*)(CO)₂(CH₄) (1,943 cm⁻¹) (a) and Re(Cp*)(CO)₂(CH₃)H (1,998 cm⁻¹) (b) after photolysis (267 nm) of Re(Cp*)(CO)₃ in scCH₄.

Kinetic Isotopic Effects.

The activation of CH₄ by Re(Cp)(CO)² and Re(Cp*)(CO)² has been investigated by using CD₄. Unfortunately, experiments in scCD₄ were prohibitively expensive, and we therefore have performed experiments in scAr doped with either CH₄ or CD₄. Photolysis of 6 in pure scAr only generates two CO bands at lower energy than those of 6, which are extremely short-lived even in the presence of a low concentration of CO (1 psi, k_obs = 2.0 ± 0.2 × 10⁶ s⁻¹) (Fig. 7a). The observed rate constant for the decay of this transient depends linearly on the CO concentration (k_CO = 9.0 ± 0.9 × 10⁸ dm³·mol⁻¹·s⁻¹, 298 K). This rate is below the expected diffusion-controlled rate (5 × 10¹¹ dm³·mol⁻¹·s⁻¹, estimated from the Stokes–Einstein equations), supporting the tentative assignment of this species to the solvated Ar complex. The photolysis of 6 in scAr (5,000 psi) doped with CH₄ (200 psi) and CO (11 psi) also has been investigated (Fig. 7b). It is clear that the results in scAr doped with CH₄ are very similar to those obtained in pure scCH₄ and that the same four transient bands are observed, which are assigned to 7 and 8 in scAr. The rate of decay of 7 and 8 is significantly slower (k_obs = 2.20 ± 0.05 ×10⁴ s⁻¹, 302 K) than the observed rate of decay of the transient in pure scAr. The experiment was repeated by using CD₄ rather than CH₄. In this experiment, the ν(CO) bands of Re(Cp)(CO)²(CD₄) (14) and Re(Cp)(CO)²(CD₃)D (15) are identifiable (Fig. 7c). A small shift (0 to 5 cm⁻¹) in the bands of 14/15 relative to 7/8 in scAr is observed. The observed rate of decay of the equilibrium mixture of 14/15 in scAr is greater (k_obs = 3.49 ± 0.05 × 10⁴ s⁻¹, 302 K) than that of 7/8 in scAr. There is a significant inverse kinetic isotope effect (k_H/k_D = 0.6 ± 0.1). We also find an inverse kinetic isotope effect (k_H/k_D = 0.4 ± 0.1) for the decay of Re(Cp*)(CO)²(CD₄) (16) and Re(Cp*)(CO)²(CD₃)D (17). An observed inverse kinetic isotope for the reductive elimination of an alkane from an alkyl hydride complex often is taken as an indicator of the presence of an alkane sigma-complex intermediate in the reductive elimination pathway (19, 50), which is consistent with our experimental assignments of the alkane/alkyl hydride equilibrium.

Fig. 7.

TRIR difference spectra of Re(Cp)(CO)₃ photolysis (266 nm) in scAr (5,060 psi, 298 K) with CO (10 psi) (a), scAr (5,000 psi, 298 K) doped with CH₄ (200 psi) and CO (11 psi) (b), and scAr (5,000 psi, 298 K) doped with CD₄ (200 psi) and CO (11 psi) (c). Spectra were recorded 80 ns (a), 10 μs (b), and 10 μs (c) after photolysis.

Activation Parameters.

Examination of the rate constant for the reaction of the equilibrium mixture 7/8 or 10/11 with CO indicates once more that the methane complexes 7 and 10 are significantly more reactive than the analogous heptane and cyclopentane complexes are (Table 1). The values of ΔH^‡ obtained for the reactions of 7/8 (51 ± 5 kJ·mol⁻¹) and 10/11 (47 ± 2 kJ·mol⁻¹) with CO in scCH₄ are again a lower limit of the ReCH₄ binding enthalpy. Previously, the lower limit of the Reheptane binding enthalpy has been reported to be 57 kJ·mol⁻¹ (23). Other lower limits for the Realkane binding enthalpy have been in the range of 32–46 kJ·mol⁻¹ (38). The ΔH^‡ for the reaction of Re(Cp)(CO)²(C²H₆) with CO in liqC²H₆ (43 ± 2 kJ·mol⁻¹) (41) is the same as that of the corresponding reaction of 12/13 (43 ± 2 kJ·mol⁻¹). These ΔH^‡ values clearly are lower than those for the corresponding reactions of 7/8 and 10/11 with CO.

Conclusions

In this paper, the observation of methane CH bond activation in solution at room temperature has been described. Using TRIR spectroscopy, we have provided examples of organometallic methane and ethane alkane complexes under such conditions. The organometallic methane and ethane complexes are all significantly more reactive than the previously reported analogous longer-chain alkane complexes. Despite this finding, we do observe a significant interaction between methane and all of the metal complex fragments examined, and it has been shown that methane has a surprisingly strong bond to rhenium in solution, with a lower limit of the binding enthalpy found to be 51 ± 5 kJ·mol⁻¹ in Re(Cp)(CO)²(CH₄). There is clearly more to learn about organometallic methane complexes and it is likely that fast TRIR will prove useful for this purpose.

Materials and Methods

W(CO)₆ (99%; Strem Chemicals, Newburyport, MA), Mn(Cp)(CO)₃ (Sigma–Aldrich), Re(Cp)(CO)₃ (99%; Strem Chemicals), Re(Cp*)(CO)₃ (99%; Strem Chemicals), CH₄ (research grade 99.995%; BOC Special Gases, Whitby, ON, Canada), C²H₆ (technical grade 99.95%; Air Products, Allentown, PA), Ar (pure shield grade; BOC), CD₄ (99 atom % D; Sigma–Aldrich), and CO (premier grade; Air Products) all were used as received.

Two different types of instrumentation were used to obtain TRIR data. Details of the diode laser-based TRIR apparatus have been described elsewhere (51). Briefly, the IR source, a continuous wave IR diode laser, was used to monitor the change in IR transmission at one IR frequency after UV excitation of the sample. Spectra were built up in a “point-by-point” fashion. ps-TRIR studies were performed on the PIRATE apparatus at the Rutherford Appleton Laboratory (Oxfordshire, U.K.), which also has been described in detail elsewhere (52). Supercritical and high-pressure liquid solutions for the experiments were prepared in a manner similar to that described in a previous study (28).

DFT calculations were performed by using Gaussian 03 (53) with the B3LYP hybrid functional (54–56) and the 6–311(G)(d,p) basis functions for all atoms except Re, for which the LanL2DZ ECP basis set augmented with “f” functions by using the contraction coefficients of Ehlers et al. (57–60) was used.

Supporting Information.

Further results, details of the calculations, and experimental procedures can be found in supporting information (SI) Figs. 8–13, SI Tables 2 and 3, and SI Text.

Abbreviations

TRIR

time-resolved infrared

scCH_4,

supercritical methane

liqC₂H_6,

liquid ethane

DFT

density functional theory

ps-TRIR

picosecond TRIR.