C–H Bond Activation and C–C Coupling of Methane on a Single Cationic Platinum Center: A Spectroscopic and Theoretical Study

We spectroscopically investigated the activation products resulting from reacting one and multiple methane molecules with Pt+ ions. Pt+ ions were formed by laser ablation of a metal target and were cooled to the electronic ground state in a supersonic expansion. The ions were then transferred to a room temperature ion trap, where they were reacted with methane at various partial pressures in an argon buffer gas. Product masses observed were [Pt,C,2H]+, [Pt,2C,4H]+, [Pt,4C,8H]+, and [Pt,2C,O,6H]+, which were mass-isolated and characterized using infrared multiple-photon dissociation (IRMPD) spectroscopy employing the free electron laser for intra-cavity experiments (FELICE). The spectra for [Pt,2C,4H]+ and [Pt,4C,8H]+ have several well-defined bands and, when compared to density functional theory-calculated spectra for several possible product structures, lead to unambiguous assignments to species with ethene ligands, proving Pt+-mediated C–C coupling involving up to four methane molecules. These findings contrast with earlier experiments where Pt+ ions were reacted in a flow-tube type reaction channel at significantly higher pressures of helium buffer gas, resulting in the formation of a Pt(CH3)2+ product. Our DFT calculations show a reaction barrier of +0.16 eV relative to the PtCH2+ + CH4 reactants that are required for C–C coupling. The different outcomes in the two experiments suggest that the higher pressure in the earlier work could kinetically trap the dimethyl product, whereas the lower pressure and longer residence times in the ion trap permit the reaction to proceed, resulting in ethene formation and dihydrogen elimination.


■ INTRODUCTION
Large amounts of methane are present in natural gas, but its great stability hinders widespread utilization by the chemical industry. Methane must be activated before it can be converted to a higher-value chemical such as methanol or ethene. Currently, methane is first thermally cracked, forming synthesis gas, before it can be converted into, for example, synthetic fuels via the Fischer−Tropsch process employing Fe-or Co-based catalysts. 1 These processes require high temperatures and pressures, which cause the overall process to be energy inefficient. Indeed, Crabtree has pointed out, even though organometallic CH bond activation dates to 1962, there remains no "general, selective, efficient catalytic functionalization reactions of unactivated sp 3 CH bonds". 2 In the gas phase, studies of alkane activation by metal cations date to the seminal work of Ridge in 1979, who observed the facile C−H (and C−C) bond activation of butanes by Fe + . 3 In recent years, there have been many investigations of how transition metals (TMs) interact with methane. 4−6 Studying these systems under well-defined conditions with highly sensitive mass-spectrometric methods can offer detailed information on reaction mechanisms and the electronic structure needed on the metal center that allows facile C−H bond activation. In such studies, it was found that several third-row TM cations are capable of dehydrogenating methane at room temperature. 7−18 The structures of several [TM,C,2H] + products were recently identified using infrared multiplephoton dissociation (IRMPD) action spectroscopy. Comparisons to theory identified TM + −CH 2 carbene structures for Ta + , W + (both with agostic distortions), and Pt + (having the classic C 2v structure), and H−TM + −CH hydrido carbyne structures for Os + and Ir + , where reaction with the latter metal ion also yielded a minority carbene product. 16,19−21 In 1990, Irikura and Beauchamp observed that several of these thirdrow TM cations could also dehydrogenate multiple methane molecules, presumably by oligomerization. 7,22,23 For Pt + , they reported dehydrogenation of up to five methane molecules but only characterized the first step, i.e., Pt + + CH 4 → PtCH 2 + + H 2 as facile. The structures of the products resulting from multiple methane dehydrogenations by any metal cation were never determined directly. Wheeler et al. studied the reaction products of Pt + and Ir + with multiple methane molecules formed in a helium buffer gas under relatively high helium pressure conditions using IRMPD spectroscopy. 24,25 For the reaction with Pt + , they identified the [Pt,3C,10H] + and [Pt,4C,14H] + products formed as Pt(CH 3 ) 2 (CH 4 ) + and Pt(CH 3 ) 2 (CH 4 ) 2 + , where the spectroscopy was enabled by the IR-induced loss of methane molecules. 24 On the basis of quantum chemical calculations, they also concluded that Pt(CH 3 ) 2 + was not the global minimum structure for [Pt,2C,6H] + , with species such as (H 2 )Pt(C 2 H 4 ) + being significantly more stable. Therefore, the reaction of PtCH 2 + with methane could form both Pt(CH 3 ) 2 + and (H 2 )Pt(C 2 H 4 ) + exothermically, but the barrier leading to the first species is below the energy of the PtCH 2 + + CH 4 reactants, whereas that for the second product is above it, preventing its formation. In an earlier theoretical study, Diefenbach et al. already concluded that PtCH 2 + should react with methane to form a Pt(C 2 H 4 ) + product rather than the bis-carbene Pt(CH 2 ) 2 + . 26 In the current study, we reinvestigate the potential of gasphase Pt + ions to activate multiple methane molecules. In contrast to previous work, the reactions take place in a radiofrequency (RF) ion trap, allowing more control over the methane partial pressure and reaction times and, consequently, over the product distribution formed. The lower pressure reaction conditions in our RF ion trap lead to the formation of other product species than those observed by Wheeler et al. in the molecular beam environment. 24 In our experiment, the reaction conditions are a closer match to the single-collision conditions of the experiments by Irikura and Beauchamp. 7 After a fixed reaction time in our RF ion trap, all ions were transferred to the Fourier-transform ion cyclotron resonance mass spectrometer (FTICR-MS) coupled to the Free Electron Laser for IntraCavity experiments (FELICE). 27 24 Finally, use of the FTICR-MS instrument allows products of interest to be mass-isolated, foregoing the need to rely on messenger tagging and enhancing the experimental sensitivity. Comparison of the experimental IRMPD spectra with theoretical calculations then permits the unambiguous identification of [Pt,2C,4H] + and [Pt,4C,8H] + as species containing one and two ethene ligands, thereby documenting the dehydrogenation and C−C coupling of multiple methane molecules on the single platinum center. ■ METHODS Experimental Section. The experiments were carried out in an FTICR-MS instrument coupled to a FELICE beamline 27,28 with a laser vaporization source described earlier. 29,30 Pt + ions were produced by irradiating a rotating Pt disk by a frequency-doubled Nd:YAG laser operating at 30 Hz and producing ∼3 mJ pulses. To cool the ions formed and to entrain them in a supersonic expansion when exiting the source region, a pulse of He carrier gas was injected through a piezoelectric valve prior to ablation. Ions were then transferred by ion optics and a quadrupole mass filter operated in guiding mode to a sectioned RF linear quadrupole ion trap with rectangular electrodes and two DC-biased circular end electrodes with a hole to admit or expel ions. The ion trap is enclosed by a room temperature cylinder that separates it from the background vacuum. Ions were thermalized by collisions with argon, introduced through a leak valve into the cylinder at a pressure of approximately 8 × 10 −4 mbar measured with a Baratron pressure gauge directly connected to the cylinder. Methane was let in via a second leak valve. A methane partial pressure ranging from 2 × 10 −5 up to 8 × 10 −4 mbar was used to maximize the intensity of the ionic species of interest.
After reacting over a period of approximately 200 ms, all ions formed were expelled from the trap by reducing the voltage on one of the end electrodes. Ions were subsequently deflected by 90°using a DC quadrupole bender and transferred to one of four ICR trapping cells coinciding with the center of the magnetic field produced by a 7 T superconducting magnet. Here, unwanted masses were ejected via a combination of single frequency, chirped, and stored-waveform inverse Fourier transform (SWIFT) pulses. 31 After mass isolation, the ions of interest were irradiated by a single macropulse of the FELICE free-electron laser. 27 Upon resonant vibrational irradiation, photofragmentation was induced, after which all ions present were detected. The molecular response to the IR laser is represented by the fragmentation yield Y F , defined as where I P (ν) and I F (ν) represent the mass spectral intensities of precursor and fragment species as a function of IR frequency ν. Experimental spectra were acquired by plotting the fragmentation yield as a function of ν. IR spectra were normalized on the macropulse energy inferred from outcoupling a fraction of the intracavity light, which was also used for wavelength calibration. FELICE macropulses had a pulse energy of approximately 0.3 J when investigating [Pt,2C,4H] + over the whole spectral range and 0.7 J for [Pt,4C,8H] + . Spectral bandwidths range from 0.9% of the central frequency in the low-frequency region to 0.33% in the high-frequency region. ICR cell 1 coincides with the focus of FELICE and thus has the highest photon fluence. ICR cells 2, 3, and 4 are 100, 200, and 300 mm away from the focus of FELICE, respectively. The FELICE Rayleigh length of 82 mm leads to an IR fluence in cell 4 that is 14 times lower than in cell 1. 27,29 Computational Section. To interpret the experimental IR spectra and to rationalize product formation pathways, density functional theory (DFT) calculations were carried out. Calculations were performed using the Gaussian16 software package 32 at the UB3LYP level of theory 33,34 using the def2-TZVPPD basis set, which had previously been proven accurate for describing the systems of interest. 24 To benchmark our methods, we calculated the binding energies of Pt + −H, Pt + −CH 2 , and Pt + −CH 3 to be 2.95, 5.04, and 2.85 eV, respectively, compared to experimentally determined values of 2.81 ± 0.05, 4.80 ± 0.03, and 2.67 ± 0.08 eV, respectively. 17,35 For platinum, this basis set uses an effective core of [Kr]4d 10 4f 14 and explicitly treats the 5s, 5p, 6s, and 5d valence electrons. Both doublet and quartet spin states were investigated. For all structures, harmonic frequencies were calculated to ascertain that they are either a true minimum or a first-order transition state. Intrinsic reaction coordinate calculations were performed to make sure that the transition states connect the desired intermediates. All frequencies used were scaled by a factor of 0.97 to account for anharmonicity and the redshift associated with the IRMPD process. This value is in between 0.939 and 0.983, as found earlier for similar systems. 16,24 The calculated spectra were convoluted with a Gaussian line shape function with a FWHM of 20 cm −1 for comparison to experimental spectra. To assess potential broadening resulting from the rotational substructure of vibrational transitions for selected species, we simulated rotational profiles for each vibrational band by assuming pure a-, b-, or c-type transitions and identical rotational constants (retrieved from the DFT calculations) for ground and excited vibrational states. To do so, we used Prof. L. Meerts' homebuilt software for diagonalizing the rotational Hamiltonian and convoluted the individual transitions, weighted by a single Boltzmann temperature factor, with a 0.9% FWHM Gaussian line shape function. Finally, all calculated energies reported in this work were zero-point energy-corrected using unscaled vibrational frequencies.
■ RESULTS AND DISCUSSION Products Formed in the Reaction of Pt + with Methane. Figure 1 shows a composite mass spectrum of ions observed when reacting Pt + ions with methane in the ion trap, created by adding several mass spectra with varying methane partial pressures in the ion trap together. Pt + ions dominate in absence of methane, showing its characteristic isotopic distribution of masses 194, 195, 196, and 198 amu (33, 34, 25, and 7% natural abundance, respectively). As the methane partial pressure is increased, [Pt,nC,2nH] + (n = 1, 2, 4) product ions gradually appear, with [Pt,4C,8H] + being evident when both argon and methane have a partial pressure of approximately 8 × 10 −4 mbar in the ion trap. Interestingly, no clear sign of a [Pt,3C,6H] + product was observed. Instead, a set of minor products at m/z = 240, 241, and 242 was observed, 46 Da higher in mass compared to the three primary Pt isotopes. This product could be identified mass-spectrometrically as [Pt,2C,O,6H] + , and for reasons explained below, this appears to be a water adduct of the [Pt,2C,4H] + product, presumably formed by a trace water contamination in either the argon or methane lines. When spectroscopically investigating a single ionic species using IR light, we ensured that all other species were ejected from the ICR cell. Only the 194 Pt and 195 Pt isotopes were used to obtain IRMPD spectra of the molecular ions examined in order to avoid ambiguities associated with products that had not dehydrogenated.
In stark contrast to the experiments carried out by Wheeler et al. in a molecular beam environment, 24 no [Pt,nC,(2n + 2)H] + (n = 1−4) products were observed. Our initial interpretation for this difference is that this is a result of the significantly longer reaction times (200 ms versus ∼100 μs) and the lower total pressures (10 −4 to 10 −3 mbar versus an estimated pressure exceeding 10 mbar). Such differences are in accordance with results reported for similar experimental designs to the one used by Wheeler et al. 36−38 The reaction in the higher pressure flow tube may thus trap intermediate reaction products. A final difference between the two experiments is that in the expansion into vacuum to form a molecular beam, products may be complexed with extra methane molecules. IR Spectroscopy of [Pt,C,2H] + . As a first test to check whether the products formed in the current experimental instrument are the same as those formed in the higher-pressure flow tube of Lapoutre et al. 16 we compare the IRMPD spectrum for the [Pt,C,2H] + product with the spectrum of PtCH 2 + reported earlier. Figure 2a shows the IRMPD spectrum recorded in ICR cell 3 for [Pt,C,2H] + , with H 2 loss as the fragmentation channel. It is compared to the IRMPD spectrum recorded using a molecular beam experiment as shown in Figure 2b. 16 Both spectra are quite similar, dominated by an intense band at 749 cm −1 and a weaker band at 977 cm −1 . Overall, the bands in the current spectrum, especially the band at 749 cm −1 , appear a bit broader, and the peak at 977 cm −1 seems to have more pronounced shoulders on both sides. The band at 665 cm −1 observed in the earlier spectrum is not obvious in the present spectrum, although it could be buried in the tail of the 749 cm −1 band. The band at 749 cm −1 is assigned to two overlapping modes: the Pt−C stretch calculated at 747 cm −1 and the in-plane CH 2 rocking mode calculated at 726 cm −1 , with the latter having an intensity three times smaller. The 977 cm −1 experimental band corresponds to the out-of-plane CH 2 wagging mode calculated at 1016 cm −1 . Rovibrational simulations at 77 K (gray) and room temperature (blue) are shown in Figure 2c. Here, it can be seen that the absence of the 665 cm −1 band, the appearance of a broader shoulder structure for the 977 cm −1 band, and the   16 (c) Room temperature (blue) and 77 K (gray) simulated spectrum of the PtCH 2 + ( 2 A 1 ) structure, including rovibrational band shapes and the harmonic frequencies in black. The asterisk (*) denotes a frequency that was shifted down 40 cm −1 before scaling by 0.97.

Inorganic Chemistry
pubs.acs.org/IC Article broadening of the 749 cm −1 band all can be rationalized by the higher temperature at which the current spectrum was recorded. Although both reactions have taken place at room temperature, the spectrum in Figure 2b was taken after the reactive mixture was expanded into a vacuum, thereby undergoing significant cooling of the internal degrees of freedom, whereas the ions formed in the current experiment were transferred without such cooling. We conclude that the room-temperature simulation for this experiment is accurate enough to explain most features. It must be noted that as in a previous publication, the b-type transition associated with the in-plane C−H wagging vibration (denoted with an asterisk in Figure 2c) was calculated at 748 cm −1 (unscaled) but has been shifted down by 40 cm −1 before scaling by 0.97 to better match the observed spectrum. 18 IR Spectroscopy of [Pt,2C,4H] + . Figure 3a shows the IRMPD spectrum of [Pt,2C,4H] + (m/z = 222, 223). It was recorded using FTICR cell 3 between 250 and 780 cm −1 and using FTICR cell 4 between 735 and 1745 cm −1 via H 2 loss into the m/z = 220, 221 mass channels. At higher IR intensities, we observe a second loss channel, namely, into the bare Pt + mass channel as shown in Figure S1. Two strong IR bands are present at 391 and 1030 cm −1 , and a medium-strong band is found at 1404 cm −1 with a tail toward the blue. Weaker bands are found at 634 and 1186 cm −1 .
The experimental spectrum is compared to calculated spectra of several trial structures. We found seven structures for the [Pt,2C,4H] + species, five of which are shown in Figure  3b−f, including simulated rovibrational envelopes. The other two, one being an isomer with very high energy and the other being a rotational isomer of the structure shown in Figure 3d, are shown in Figure S2. The lowest-energy structure is the PtC 2 H 4 + complex (ethene ligand), a structure with C 2v symmetry where Pt lies 2.04 Å above the center of the C−C bond. It is located 0.53 eV below the PtCH 2 + + CH 4 reactants when the elimination of H 2 is assumed. A PtCHCH 3 + complex (ethylidene ligand) is found to be 0.24 eV higher in energy than the ethene complex. This structure can be seen as a carbene structure where one of the hydrogen atoms is replaced by a methyl group. Transfer of a hydrogen from the methyl group to the platinum results in a HPtC 2 H 3 + species lying 0.55 eV above the PtC 2 H 4 + structure and slightly above the energy of the reactants. A similar structure in which the H ligand lies trans to the vinyl ligand (instead of the cis orientation shown in Figure 3d) was found to lie 0.01 eV higher in energy and has a very similar spectrum ( Figure S2) . It is the only structure offering predictions for all bands observed. Three of the bands predicted are readily attributed to the IRactive vibrations of ethene, which are found at 826, 949, and 1444 cm −1 in the free molecule. 39 The most intense band in the experimental spectrum (1030 cm −1 ) can be assigned to the ethene-asymmetric CH 2 wagging mode (in the Pt−C−C plane, 949 cm −1 in free ethene) predicted at 1027 cm −1 for the complex. Its symmetric counterpart, symmetry-forbidden in ethene (but Raman-active at 943 cm −1 ), carries IR intensity because of the presence of Pt and is predicted at 1037 cm −1 with an intensity eight times lower than the IR-active band. Both frequencies are slightly blueshifted upon binding to Pt + , as explained below. The asymmetric wagging mode is a b-type transition that, when its rovibrational envelope is simulated (see Figure 3b), even explains the double-peaked maximum observed. The experimental band at 1404 cm −1 is assigned to the concerted (out-of-phase) ethene CH 2 scissoring mode (1444 cm −1 in free ethene) predicted at 1420 cm −1 . This band has a shoulder on the high-energy side that can be assigned to the C−C stretch vibration predicted at 1497 cm −1 . Because the C−C stretch vibration is found by Raman spectroscopy at 1623 cm −1 in free ethene, this mode is affected by the complexation as discussed further below. 39 The out-of-phase in-plane CH 2 rocking mode, weak in free ethene at 826 cm −1 , does not gain any intensity and is predicted at 803 cm −1 ; from the inset of Figure 3a, it can be seen that no sign for this band is observed experimentally. However, when irradiating at higher intensities ( Figure S1b), a very weak band is observed at this frequency.
The intense experimental band at 391 cm −1 is not due to an ethene vibration. The calculations predict that this band is a combination of the symmetric and asymmetric Pt−ethene stretches calculated at 360 and 373 cm −1 . The former is the most intense and, being an a-type transition, explains the band's relative sharpness. It rivals the asymmetric CH 2 wagging mode at 1027 cm −1 in intensity despite having a significantly lower integrated IR intensity. Two bands remain: the weak experimental band at 634 cm −1 , visible in the inset in Figure  Inorganic Chemistry pubs.acs.org/IC Article 3a, matches a band calculated at 629 cm −1 , associated with the rocking motion of the whole ethene molecule with respect to Pt. The last experimental band at 1186 cm −1 matches a mode possessing both in-phase scissoring and C−C stretch character, here predicted at 1211 cm −1 (only Raman-active in free ethene, 1342 cm −1 ). All experimental band frequencies, calculated frequencies, and intensities, including a mode description, are gathered in Table 1.
Upon ethene binding to Pt + , the wagging vibrational frequencies are found to increase, whereas a frequency decrease was observed for the C−C stretch and scissoring modes. The lower frequency and thus smaller force constant of the C−C stretch vibration of PtC 2 H 4 + compared to free ethene can be explained by the weakening of the C−C bond by electron donation to Pt + . Our calculations show that the C−C bond length increases from 1.325 Å in free ethene to 1.402 Å in PtC 2 H 4 + , while the C−H bond length remains virtually the same (1.083 Å in free ethene, 1.085 Å when bound to Pt + ). Further, upon binding to Pt + , the ethene molecule is no longer planar, with all hydrogen atoms lying slightly below the C−C axis and oriented away from the Pt atom. The carbon atoms thus go from sp 2 hybridization in free ethene to incorporate more sp 3 -like hybridization in PtC 2 H 4 + . The transfer of electron density from ethene to the Pt + ion is also explored by examining the calculated local charges on the atoms as shown in Table 2. A Mulliken population analysis, which is based on electron density present in orbitals of a nucleus, indicates that Pt + has a charge of +0.74e, while ethene has a total charge of +0.26e. This charge on ethene is not equally shared: both carbon atoms sum to −0.52e, and the four hydrogen atoms sum to +0.78e. An alternative metric, the atomic polar tensor (APT) charges, indicates that the net positive charge on Pt in the PtC 2 H 4 + complex equals +0.49e, while the ethene molecule has +0.51e. Here, all carbon and hydrogen atoms have a charge of approximately +0.08e.
We also observed [Pt,2C,O,6H] + at m/z = 240, 241, and irradiation of this ion led to a primary fragment ion at m/z = 222, 223 (loss of 18 Da) and a secondary fragment ion at m/z = 220, 221 (loss of 20 Da) as identified in Figure S6. As discussed in the Supporting Information, this ion can be identified as (H 2 O)Pt + (C 2 H 4 ), i.e., the platinum ethene product complexed with adventitious water in the vacuum system. This result further validates the observation of C−C coupling in the present system.
Reaction Pathway for the Formation of PtC 2 H 4 + . To rationalize the formation of ethene when reacting Pt + with two methane molecules, we computationally investigated a potential reaction pathway starting from the platinum carbene product, PtCH 2 + , which is formed exothermically by reacting Pt + with one methane molecule as described earlier. 14,15,17 PtCH 2 + can react with another methane molecule to form the PtC 2 H 4 + product via the pathway indicated by the black trace in Figure 4, with detailed energetics of the intermediates in Table 3. All species shown are on the doublet spin surface, except for methane and dihydrogen, which are singlets. The reaction starts via physisorption of the methane molecule onto PtCH 2 + , forming intermediate 1. The adsorption energy associated with this step is large enough (0.77 eV) to allow transfer of a hydrogen atom from the CH 4 group to the platinum atom via transition state TS1, forming intermediate 2, HPt(CH 2 )(CH 3 ) + . Now, the carbon atoms from the carbene and methyl groups can couple together via TS2. This step has the highest energy barrier along the path toward PtC 2 H 4 + . TS2 is higher in energy than the PtCH 2 + + CH 4 reactants by 0.16 eV; however, it still lies 0.21 eV below the energy of the Pt + + 2 CH 4 reactants. Thus, without dissipation, the energy in the system is sufficient to overcome this barrier. However, if the PtCH 2 + product formed from Pt + and CH 4 is fully equilibrated with its surroundings, one would expect an equilibrium with at least part of the population trapped before this barrier. Carbon−carbon-coupling via TS2 leads to  40 Can this potential energy surface explain why Wheeler et al. found the Pt(CH 3 ) 2 + product by reacting Pt + with two methane molecules, while the current study finds PtC 2 H 4 + instead? 24 The reaction pathway toward Pt(CH 3 ) 2 + is shown in red in Figure 4. Starting from intermediate 2, it is possible to transfer the hydrogen atom on platinum to the carbene group  92 eV, which we found to be associated to an excited state of 2 B 1 symmetry, whereas we found a 2 A 1 ground state at −1.31 eV. The ground state has a doubly occupied π* orbital and a singly occupied sd hybrid on Pt, whereas the excited state has a singly occupied π* orbital and a doubly occupied sd hybrid.
From this reaction pathway, we conclude that the higherpressure environment of the flow-tube reaction cell in the molecular beam experiment used by Wheeler et al. thermalizes the product ions more efficiently. These thermalizing collisions lead to the PtCH 2 + reactant possessing less excess energy such that when it reacts with methane, it cannot easily dehydrogenate. Instead, it gets kinetically trapped at the lowest-energy [Pt,2C,6H] + species, Pt(CH 3 ) 2 + , that can be formed before TS2. In the room-temperature ion trap used here, there are not only fewer thermalizing collisions, but the long storage time and low endothermicity of the +0.16 eV barrier associated with TS2 also allow the reaction to proceed, and no kinetic trapping of intermediates is observed. The same mild endothermicity may also explain why this reaction was characterized as not facile by Irikura and Beauchamp, 7 certainly when comparing this barrier to the lack of a barrier for dehydrogenation of the first methane by Pt + . The dehydrogenation of CH 4 by Pt + yields PtCH 2 + + H 2 and has been experimentally determined to be exothermic by 0.08 ± 0.03 eV. 4 Finally, we address the fragmentation observed under IR irradiation. Loss of H 2 from PtC 2 H 4 + to form PtC 2 H 2 + is calculated to require 1.49 eV, not considering any possible barriers, whereas the bond energy between Pt + and C 2 H 4 is calculated to be 2.94 eV. Thermodynamically, H 2 loss from PtC 2 H 4 + is thus favored over ethene loss. All values are included in Table 4 together with the energetics of other ions characterized using IRMPD spectroscopy.

IR Spectroscopy of [Pt,4C,8H] + .
When we increased the methane concentration in the ion trap to 8 × 10 −4 mbar, we observed a [Pt,4C,8H] + species. In contrast to the case for [Pt,2C,4H] + , resonant IR irradiation under moderate intensities led to multiple loss channels: H 2 loss into the m/z = 248, 249 mass channels, C 2 H 4 loss into the m/z = 222, 223 mass channels, and C 2 H 4 + H 2 loss into the m/z = 220, 221 mass channels. The branching ratio intensities of precursor and fragments as a function of IR frequency are shown in Figure  S3. Figure 5a shows the IRMPD spectrum of [Pt,4C,8H] + , which contains two strong bands at 998 and 1434 cm −1 and two bands of medium intensity at 1315 and 1544 cm −1 .
The IRMPD spectrum is compared to the calculated spectra of several trial structures. We found a total of 12 stable isomers    Figure 5b−e, and the other eight are shown in Figure S4. The calculated isomers shown in Figure 5 were selected based on their relatively low energies and the match of the calculated IR spectrum with the experimental IRMPD spectrum. The structures located have different binding motifs to the Pt + ion as we found σ bonds, π bonds, and long-range interactions for physisorbed methane molecules. Examples of calculated isomers shown in Figure S4 include methane propyne, methane propadiene, methyl allyl, 1-and 2-butene, and hydrido cyclobutyl. All twelve isomers are on the doublet spin surface. The lowest-energy quartet structure found is a PtCH 2 C(CH 3 ) 2 + species (isobutene ligand), 2.59 eV higher in energy than the putative global minimum of Pt(C 2 H 4 ) 2 + . When comparing the obtained IR spectrum of [Pt,4C,8H] + with the calculated spectra of the isomers selected for Figure 5, we observe a good match with the Pt(C 2 H 4 ) 2 + species, the lowest-energy structure. Other species, for instance, the doublet PtCH 2 C(CH 3 ) 2 + at 0.74 eV relative to Pt(C 2 H 4 ) 2 + , exhibit bands at similar frequencies with reasonable intensities, but they do not provide any more satisfactory agreement. The broad structure encompassing the 1315−1434−1544 cm −1 triad is tempting to assign as a mixture of species, but we rather suspect that the broad structures are a result of power broadening given the higher pulse energy used here. Rotational envelopes play less of a role for these species as all have rotational constants at most half those of PtC 2 H 4 + . We thus identify the [Pt,4C,8H] + species as a complex containing two ethene ligands that are bound on opposite sides of the Pt, with the C−C bonds oriented at an angle of 90°. The experimental band at 998 cm −1 is assigned to be a combination of the symmetric and asymmetric wagging motion of the ethene molecules in Pt(C 2 H 4 ) 2 + (see Table 5 From a catalytic standpoint, the observation that ethene loss is the dominant decomposition pathway for Pt(C 2 H 4 ) 2 + is an interesting finding. We can safely assume that H 2 elimination is not barrierless and that the loss of C 2 H 4 is the kinetically preferred pathway. This implies that PtC 2 H 4 + itself can be seen as a catalyst for C−H bond activation and C−C coupling in the formation of ethene from two methane molecules. Indeed, the formation of Pt(C 2 H 4 ) 2 + + 2 H 2 from PtC 2 H 4 + + 2 CH 4 is calculated here to be exothermic by 0.07 eV, although we did not calculate the full potential energy surface. Combined with the 2.11 eV required to remove the second ethene ligand and recover the PtC 2 H 4 + catalyst, this cycle requires a significant energy input of 2.03 eV, but this energy is certainly lower than the original route of cracking followed by Fischer−Tropsch synthesis. It can thus be speculated that similar reaction pathways could be facilitated by supported single-atom catalysts. As such, we expect that the details of the reaction path presented in this work, which appear robust as they can rationalize experimental findings in both low-and highpressure regimes, can help guide the development of such catalytic systems.
One example of a detail along the reaction path in Figure 4 is the rate-limiting step for C−C coupling, i.e., TS2. In the precursor to this transition state, intermediate 2, the platinum center has three ligands (H, CH 2 , and CH 3 ), each of which is bound to the platinum cation by a single covalent bond. This fills the 5d and 6s orbitals on Pt, while the CH 2 ligand is left