Heterogeneously Catalyzed Aerobic Oxidation of Methane to a Methyl Derivative

Abstract A promising strategy to break through the selectivity‐conversion limit of direct methane conversion to achieve high yields is the protection of methanol via esterification to a more stable methyl ester. We present an aerobic methane‐to‐methyl‐ester approach that utilizes a highly dispersed, cobalt‐containing solid catalyst, along with significantly more favorable reaction conditions compared to existing homogeneously‐catalyzed approaches (e.g. diluted acid, O2 oxidant, moderate temperature and pressure). The trifluoroacetic acid medium is diluted (<25 wt %) with an inert fluorous co‐solvent that can be recovered after the separation of the methyl trifluoroacetate via liquid–liquid extraction at ambient conditions. Silica‐supported cobalt catalysts are highly active in this system, with competitive yields and turnovers in comparison to known aerobic transition metal‐based catalytic systems.


Introduction
In 2018, it is estimated that nearly 145 billion cubic meters of unused natural gas,amethane-rich by-product of oil extraction, was flared owing to the lack of commercial technologies and incentives to bring it to market. [1] Methane derived from natural gas is ahighly abundant resource that is widely used in the production of commodity chemicals and liquid energy carriers,n otably methanol. [2] Although wellestablished industrial routes for methane utilization exist, these processes generally proceed through an indirect, twostep pathway,w ith syngas as an intermediate product.
Consequently,t hese energy-and capital-intensive processes are rendered economically unviable for methane valorization at small-and mid-scale facilities, [2a,3] such as remote and decentralized shale oil production sites. [4] Thec hallenge of developing am ore scale-flexible direct methane conversion route has therefore motivated both the academic and industrial communities over the previous years. [5] Many approaches have evolved to address this challenge, including the selective functionalization of the methane CÀH bond to form primary oxygenates,such as methanol or other derivatives,o ver heterogeneous catalysts. [2b, 4c, 5d, 6] However, existing approaches to partial oxidation are constrained by an umber of limitations,m ost notably,t he selectivity-conversion paradigm arising from the vulnerability of methanol to over-oxidation. [5d,7] Comparisons between diverse solid catalyst systems reveal ac lear trend of decreasing selectivity to methanol with increasing conversion of methane,alimit that severely restricts achievable yields and necessitates operating conditions that are highly impractical for commercialization. [5d, 7b,c,8] In order to achieve industrially relevant methane conversion without compromising high selectivity for methanol, it is imperative that methanol is stabilized to prevent over-oxidation. [7b, c, 9] An interesting "product protection" strategy,d emonstrated primarily in homogeneous catalytic systems,i st he protection of methanol via esterification to methyl bisulfate [10] and methyl trifluoroacetate. [11,18,23] These esters are less prone to further oxidation under typical reaction conditions [7b] and, therefore,o ffer the possibility to circumvent the selectivity-conversion limit during methane conversion. In later steps,t hese esters can be hydrolyzed back to methanol, thereby widening the array of potential commercial applications for this chemistry.T he translation of this protection strategy to use with aheterogeneous catalyst has been largely unexplored, with the notable exception of Schüth and co-workers who synthesized ahighly active solid catalyst based on the homogeneous Periana platinum bipyrimidine complex for the conversion of methane to methyl bisulfate in oleum using aSO 3 oxidant. [6e] Thea ttractive yields of current methane-to-methyl ester processes notwithstanding,t hese processes still fall short of potential commercial application. Rather than focusing only on obtaining high product yields,amore holistic consideration of process conditions,cost and handling of reagents,and necessary separations and recycle streams is essential. These factors,w hich have been rigorously defined by industrial experts in this field, [12] constitute af urther set of criteria for commercialization. In particular,s everal key factors have precluded commercialization of these processes,w hich include:1)the homogeneous nature of the catalysis,translating into challenges in product and catalyst recovery;2)the use of strong corrosive acids in an undiluted form, which leads to greater operational hazards and costly equipment; [13] 3) the use of economically unacceptable oxidants,such as potassium persulfate and hydrogen peroxide; [13] and 4) the difficulty in hydrolyzing the ester to methanol due to the highly exother-mic interaction of water with the reaction mixture,w hich potentially results in the undesired evaporation of methanol. [14] Aprocess that demonstrates the breakthrough performance of methane-to-methyl-esters ystems while simultaneously overcoming these conventional challenges has not, to our knowledge,been previously demonstrated.
To address the pressing limitations of syngas-free methane conversion, we propose an approach that combines the heterogeneously catalyzed partial oxidation of methane with subsequent esterification of the product in areaction medium of trifluoroacetic acid (TFA) diluted in an inert perfluoroalkane co-solvent. Perfluoroalkanes,s uch as perfluorohexane, are inert, stable at elevated temperatures, [15] and can exhibit high solubilities for gases. [16] By diluting TFAt om anageable concentrations of below 25 wt %i nan on-corrosive and oxidation-resistant perfluoroalkane,anumber of improvements are attained, namely:astrongly reduced corrosivity of the reaction medium;i mproved stability of ah eterogeneous catalyst through operation in am ilder environment;e nhanced recovery of the methyl ester via asimple liquid-liquid extraction with an on-fluorous polar solvent;a nd improved hydrolysis conditions.

Results and Discussion
Previous studies have shown that an umber of homogeneous transition metal-based catalysts display activity for methane to methyl trifluoroacetate conversion, including those based on copper, [11c] manganese, [11b] and cobalt. [23] As apreliminary step,therefore,avariety of transition metals on solid supports were synthesized and screened for activity in ab atch reactor system ( Figure 1). Based on the initial screening results presented in the Supporting Information, cobalt-containing silica catalysts synthesized via an incipient wetness impregnation (Co/SiO 2 -IWI) with an aqueous cobalt nitrate solution showed the most promising activity.
Them ethyl ester is recovered from the fluorous reaction medium through afacile liquid-liquid extraction with apolar solvent at room temperature.A cetonitrile-d3 is an aprotic polar solvent used as the extractant to preserve the product as the methyl ester and directly measure the product concentrations using 1 HNMR. Only residual amounts of the methyl ester remain in the fluorous phase immediately after contact with the deuterated acetonitrile phase as determined through 1 HNMR (Supporting Information, Figures S1 and S2), and therefore the total product yield can be determined from the deuterated acetonitrile phase after the extraction. Substituting water as the non-fluorous extracting solvent could provide opportunities for product separation with enhanced ester hydrolysis conditions compared to processes that rely on undiluted reaction mediums.I mportantly,t his simple and highly effective product separation method provides as traightforward pathway for recycling the fluorous cosolvent, thereby greatly reducing the overall usage of this component. This unique advantage in product separation and solvent recycle has not been previously described in published methane-to-methyl-ester systems.
Herein, catalysts with cobalt loadings of 0.1 wt %, 0.5 wt %, 1.5 wt %, 5wt%,a nd 10 wt %w ere synthesized via an incipient wetness impregnation method (Co/SiO 2 -IWI) and studied for the catalytic partial oxidation of methane. Figure 2a illustrates the dependence of the methyl ester productivity on the cobalt loading of the Co/SiO 2 -IWI catalysts under the typical reaction conditions.Amethyl ester productivity of approximately 250 mmol g cat À1 h À1 is obtained with the 0.5 wt %Co/SiO 2 -IWI catalysts.Increasing the cobalt content of the catalysts up to 5wt% results in only minor changes in the productivity of the methyl ester ( Figure 2a,b lue markers). Cobalt utilization is substantially higher for the low-loaded catalysts,with amaximum of nearly 175 mmol g Co À1 h À1 achieved using the 0.1 wt %C o/SiO 2 -IWI material, which then decreases rapidly with increasing cobalt loading ( Figure 2a,g reen markers). Thep erformance of the 10 wt %C o/SiO 2 -IWI catalyst is substantially inferior to the catalysts with lower cobalt content in terms of both productivity and cobalt utilization. In combination with the average particle sizes from the corresponding transmission electron microscopy (TEM) images (see Supporting Information), these results suggest that the more active catalysts are those with the higher cobalt dispersions.C onsequently,t he most efficient utilization of cobalt is realized in the lowest cobaltloaded cases.
Theh eterogeneity of the reaction for the Co/SiO 2 -IWI materials was confirmed via ahot filtration test and ICP-OES of the reaction medium (Supporting Information, Figure S4). Additional tests demonstrating the necessity of cobalt and molecular oxygen oxidant are summarized in Table S3 of the Supporting Information. To further assess the stability of the Co/SiO 2 -IWI catalysts,t he catalyst material was recovered after an initial catalytic test and recycledback into the reactor for asecond run conducted under the same conditions.Before recycling, some samples underwent at hermal treatment consisting of ac alcination to at least 250 8 8Ci ns tatic air and are denoted as "reactivated" samples.S amples that did not receive any treatment before the second run are referred to as "spent" samples.The catalytic productivity obtained with the spent 0.5 wt %C o/SiO 2 -IWI catalyst after reactivation is comparable to that obtained with the freshly synthesized catalyst (Figure 2b). Without the reactivation step,t his catalyst demonstrates decreased activity.
On inspecting the fresh catalysts by TEM, we find large cobalt-containing particles in those with high loadings,and Xray diffraction (XRD) of these catalysts confirms the presence of Co 3 O 4 (see Supporting Information). X-ray absorption spectroscopy (XAS) further corroborates the dominance of as pinel Co 3 O 4 at each of the weight loadings in the fresh state and suggests the presence of aminority Co IIlike fraction, which becomes less significant as the cobalt loading increases (Supporting Information, Figure S9). Figure 2c shows the Co K-edge X-ray absorption near edge structure (XANES) of the 0.5 wt %C o/SiO 2 -IWI catalyst in the fresh, spent, and reactivated state.T he lower binding energy feature around 7725 eV associated with aC o II component is noticeably more prominent in the spent and reactivated catalysts as compared to the fresh catalysts. Linear combination analysis (LCA) fits to the XANES of these materials (see Supporting Information) suggests that the reactivated catalyst regains more of the Co III /Co II character indicative of the initial spinel structure,b ut still retains asignificant, loading-dependent amount of the second Co(OH) 2 -like character (Figure 2c,i nset). Fort he higherloaded samples,such as the 5wt% Co/SiO 2 -IWI catalyst, the cobalt structure appears to be less perturbed by reaction in the spent catalysts and more closely resembles the initial structure following reactivation (Supporting Information, Figure S11).
In the case of the 0.1 wt %Co/SiO 2 -IWI catalysts,both the spent and reactivated catalysts show equivalent ester productivities that are approximately 50 %ofthat obtained with the fresh catalyst (Figure 2e). Despite the decrease in activity observed after the initial reaction, the spent and reactivated samples still show better cobalt utilization than the higherloaded catalysts,w ith cobalt-based ester productivities around 80 mmol g Co À1 h À1 (corresponding to approx. 240 mmol g cat À1 in Figure 2e). From an activity standpoint, the aerobic reactivation protocol does not have as ignificant effect on the 0.1 wt %catalyst. TheXAS of both the spent and reactivated materials that reveals as trong resemblance between these materials (Figure 2f)s upports this observation. Theeffects of catalysis manifest principally through the development of as houlder in the XANES,a ssociable with Co II @7 725 eV in the higher-loaded samples (Figure 2c). There is an explicit linear relationship between the fraction of cobalt associated with the 7725 eV Co II XANES feature and the obtained methyl ester yield over the studied range of cobalt loadings (Figure 2d). The1 0wt% Co/SiO 2 -IWI catalyst noticeably deviates from this trend, which likely results from the greatly increased domain size and reduced dispersion due to the formation of very large cobalt particles and agglomerates,asevidenced by TEM and XRD (see Supporting Information). In the case of the 0.1 wt %m aterial, the XANES feature @7725 eV Co II is transformed into as trong peak rather than as houlder (Figure 2f), and the overall XANES envelopes of 0.1 wt %Co/SiO 2 -IWI in the spent and reactivated state are markedly different from those of the higher weight-loaded samples.
Thep re-edge feature at ca. 7709 eV is indicative of afraction of the cobalt that initially exists with the tetrahedral (T d )s ymmetry expected from the Co II component of the spinel structure of Co 3 O 4 (Figure 2f,i nset). [17] This feature is removed in the spent catalyst, suggesting the removal of T d symmetry from the 0.1 wt %sample post-reaction. Moreover, the overall shift of the Co K-edge edge position to lower energy further suggests that the octahedral (O h )C o III component of the Co 3 O 4 ,h as also been consumed. Analysis of the EXAFS further indicates that ac omplete structural transformation of the cobalt species present on the 0.1 wt % catalyst from the Co 3 O 4 starting phase into ahighly dispersed O h Co II species has occurred (see Supporting Information). Thecombination of catalytic data with the XAS suggests that the new O h Co II species can still be associated with activity for methane oxidation. Themost active cobalt species across the series of catalysts,t herefore,m ay not be the Co 3 O 4 nanoparticles that are predominantly present in the fresh catalysts across the range of loadings,b ut rather an ewly formed species that is best observed in the low-loaded, highly dispersed catalyst materials.
By increasing the liquid hold-up in the reactor and reducing the partial pressure of oxygen, the highest oxygenbased yield of 12 %ofthe theoretical maximum was reached with 1.5 wt %C o/SiO 2 catalyst (Supporting Information, Table S5). Figure 3a ranks the performance of the cobaltcatalyzed process reported herein in comparison to other methane oxidation approaches that solely use dioxygen as the oxidant based on two important parameters:productivity and oxygen conversion to the desired product. This comparison reveals the considerable performance-gap that exists between aerobic heterogeneous systems for methane-oxidation. The latter operate aerobically with aN ADH electron carrier and are characterized by ahighly selective conversion of methane to methanol. [26] As high-temperature heterogeneous catalytic approaches using different transition metals typically result in am uch poorer selectivity even at lower conversion, these systems appear on the very left in Figure 3a.T he performance of the Co/SiO 2 -IWI catalyst is distinct from these other approaches and advances toward the more efficient bio-enzymatic systems.F urthermore,m ethanol productivity achieved per gram of the cobalt catalyst is of the same order as that achieved per gram of dry/ wet cells in the biological conversion of methane.W ith the Co/SiO 2 -IWI catalyst, ap roductivity up to 0.03 kg methanol kg cat À1 h À1 was attained at low cobalt loadings under low-pressure conditions and ap ronounced enhancement is expected on increasing the feed pressure further. Thev olumetric productivity,o rs pacetime-yield (STY), is an additional metric useful for comparing processes.A lthough optimization of this parameter was not af ocus of this work, volumetric productivity in this case would likely be greatly enhanced through increasing the amount of catalyst in the reactor,i ncreasing methane partial pressures,a nd targeting the retention of efficient cobalt utilization at higher weight loadings.
Apart from excellent oxygen-based product yields,working at lower partial pressure of methane showed improved selective methane conversion to the methyl ester.R eactions performed with a2 .5 %m ethane feed resulted in methanebased ester yields of up to 17 %( Supporting Information, Table S4). While the transition metal-based catalysts depicted in Figure 3b reach 2-3 %m ethane-based yields at best, biological systems consistently report yields close to 30 %. [26] Theh igh activity of MMO enzymes coupled with the high selectivity in methane conversion and oxygen utilization has been proven difficult to be replicated in heterogeneous systems.T he high activity of the cobalt-based catalyst along with the use of esterification as a" product protection" strategy enables aperformance that is comparable to the bioenzymatic systems. Table 1l ists the activity of the Co/SiO 2 -IWI catalyst with known homogeneous catalytic systems that oxidize methane to methyl trifluoroacetate.T he performance of these systems is compared on the basis of the reported turnovers (TO) over the reaction period and the productivity.W en ote that the turnovers reported typically represent al ower bound on the amount of product formed per mole of catalyst, since in most cases it appears that the catalyst may not be fully deactivated at the end of the reaction period. TheC o/SiO 2 -IWI catalyst achieved turnovers of up to 31 and ac orresponding productivity of 10.3 mol ester mol Co À1 h À1 .T his performance is competitive to even homogeneous systems that employ stronger oxidants (K 2 S 2 O 8 ,H 2 O 2 ), higher partial pressures of methane, and/or more complex redox cycling schemes.T aken together, Figure 3a nd Table 1i llustrate the outstanding catalytic performance of the system that is achieved with an economical oxidant and under reaction conditions that are more favorable than those generally employed in state-of-the-art methane-to-methyl-ester systems.

Conclusion
An improved methane-to-methyl ester process with competitive performance for the selective conversion of methane was designed by systematically addressing the limitations of conventional systems.T he use of af luorous co-solvent as an acid diluent brought several advantages,n otably am ilder reaction environment that allowed working in ah eterogeneous mode and af acile product and solvent recovery method through ah ighly effective extraction with an on-fluorous solvent. Theperformance of this cobalt-catalyzed process far exceeds other comparable heterogeneous transition metalbased high-temperature catalytic processes,w hich are generally restricted to much lower methane conversion and product selectivity.T hrough future work to elucidate the reaction mechanisms,t arget improvement of the cobalt utilization at high metal loadings,and optimization of process conditions,this novel catalytic approach holds the promise of significant future improvement.