Catalyst-controlled selectivity in the C–H borylation of methane and ethane

See allHide authors and affiliations

Science  25 Mar 2016:
Vol. 351, Issue 6280, pp. 1421-1424
DOI: 10.1126/science.aad9289

Methane borylation in a cyclohexane sea

Although methane combusts readily at high temperatures, it is generally the hardest hydrocarbon to transform under gentler conditions, owing to its particularly strong C-H bonds. Cook et al. now show that soluble rhodium, iridium, and ruthenium catalysts can slice through these C-H bonds to add boron substituents to methane at 150°C. Smith et al. report the iridium-catalyzed reaction using phosphine ligands to enhance activity. Both studies were performed in cyclohexane solvent, revealing a remarkable selective preference for the methane reaction over functionalization of the cyclic hydrocarbon.

Science, this issue pp. 1421 and 1424


The C–H bonds of methane are generally more kinetically inert than those of other hydrocarbons, reaction solvents, and methane functionalization products. Thus, developing strategies to achieve selective functionalization of CH4 remains a major challenge. Here, we report transition metal–catalyzed C–H borylation of methane with bis-pinacolborane (B2pin2) in cyclohexane solvent at 150°C under 2800 to 3500 kilopascals of methane pressure. Iridium, rhodium, and ruthenium complexes all catalyze the reaction. Formation of mono- versus diborylated methane is tunable as a function of catalyst, with the ruthenium complex providing the highest ratio of CH3Bpin to CH2(Bpin)2. Despite the high relative concentration of cyclohexane, minimal quantities of borylated cyclohexane products are observed. Furthermore, all three metal complexes catalyze borylation of methane with >3.5:1 selectivity over ethane.

Over the past 50 years, numerous homogeneous transition-metal catalysts have been developed for the C–H functionalization of liquid alkanes [for example, via dehydrogenation (1), oxygenation (2), carbonylation (3), borylation (47), and C-, N-, and O-atom insertion (8, 9)]. However, relatively few of these catalysts have been translated to analogous reactions of methane (1014). This is largely due to the particular challenges associated with methane C–H functionalization. First, the C–H bonds of methane are stronger than those of most liquid alkanes [the C–H bond dissociation energies (BDEs) of methane, n-hexane (1°C–H), and cyclohexane are 105, 101, and 99.5 kcal/mol, respectively (15, 16)]. As such, methane C–H bond cleavage is prohibitively slow with many catalysts. Second, homogeneous alkane functionalization reactions are typically conducted by using neat alkane as the solvent (4, 5, 14), so the use of methane gas as a substrate poses challenges with respect to identifying a compatible reaction solvent (12, 17). Last, the reaction solvent and the CH3X products of methane functionalization typically contain more reactive C–H bonds than those of CH4. As such, developing strategies to achieve selective functionalization of CH4 in the presence of solvent and CH3X remains a challenging problem (1014).

We sought to identify a methane C–H functionalization process in which selectivity (both for CH4 versus CH3X functionalization and for CH4 versus solvent C–H functionalization) could be tuned through modification of the homogeneous transition-metal catalyst. To accomplish this goal, we focused on the catalytic C–H borylation of methane with B2pin2 (Fig. 1A). Over the past 15 years, there has been tremendous progress in the development of transition-metal catalysts for the C–H borylation of liquid alkane substrates. Catalysts based on iridium (Ir) (18, 19), rhodium (Rh) (2022), rhenium (Re) (23), and ruthenium (Ru) (24) have been reported for liquid alkane C–H borylation, typically by using the alkane substrate as the solvent and B2pin2 as the borylating reagent (19, 21, 2325). With the vast majority of liquid alkane substrates, the selectivity of C–H borylation is dominated by steric factors, with terminal (primary) C(sp3)–H bonds undergoing selective functionalization over secondary or tertiary C–H sites (25, 26). This selectivity has been reported to be largely independent of the nature of the transition-metal catalyst. For example, the C–H borylation of n-alkanes (n-CnH2n+2) with B2pin2 affords 1-Bpin-CnH2n+1 as the sole detectable product with Ir-, Rh-, Re-, and Ru-based catalysts (18, 20, 23, 24).

Fig. 1 Reactivity and selectivity challenges in the C–H borylation of methane.

(A) Proposed methane C–H borylation reaction. (B) Challenges with selectivity in methane C–H borylation.

In certain contexts, the introduction of a Bpin substituent has been shown to electronically activate adjacent C–H bonds toward further C–H borylation by rendering them more acidic (27, 28). This electronic activation has been best studied in the context of benzylic substrates, in which the C–H borylation of 1°-benzylic C–H bonds is often slower than that of the 2° α-boryl benzylic C–H bonds of the products (29, 30). However, the interplay between these steric and electronic effects has not been extensively explored in the C–H borylation literature, especially as a function of catalyst metal identity. As discussed below, these issues are expected to be particularly salient in the context of methane C–H borylation (Fig. 1B).

In 2005, Hall and co-workers reported density functional theory (DFT) calculations that suggest that Cp*Rh complexes (Cp*, pentamethylcyclopentadienyl) should be capable of catalyzing the C–H borylation of CH4 (22). Despite these encouraging computational results, there have been no subsequent experimental studies establishing the feasibility and/or exploring the selectivity of methane C–H borylation with these or any other catalysts. In a methane C–H borylation reaction, three major C–H bond–containing molecules will be present in solution: methane, CH3Bpin, and solvent (cyclohexane) (Fig. 1B). Among these three molecules, methane has the most sterically accessible C–H bonds, CH3Bpin has the most electronically activated (acidic) C–H bonds, and the reaction solvent, cyclohexane, is statistically favored because of its high concentration. Our studies sought to (i) experimentally establish the feasibility of metal-catalyzed methane C–H borylation; (ii) determine which factor (or factors) dominate selectivity in this transformation (sterics, electronics, or statistics); and (iii) probe whether different catalysts can be used to tune the selectivity of the reaction.

We selected Rh complex 1 for our initial investigations of methane C–H borylation on the basis of Hall and co-workers’ DFT calculations, which predicted a relatively low barrier for CH4 activation with this complex (22). The initial reactions were conducted in a Parr high-pressure batch reactor (45 mL volume) at 150°C, using 1.5 mole percent (mol %) of 1, 3500 kPa of methane, and 0.89 mmol of B2pin2 as the limiting reagent (31). As discussed above, the choice of solvent is particularly critical because any C–H bonds in the solvent must be less reactive with 1 than those of CH4. Thus, we first examined solvents without C–H bonds [perfluoromethylcyclohexane (PFMCH) and perfluorohexane (PFH)]. However, modest yields of methane C–H borylation products were obtained (Table 1, entries 1 and 2), likely because of the low solubility of the Rh catalyst in these media. We next examined cyclopentane (c-C5H10) and cyclohexane (c-C6H12) as solvents (Table 1, entries 3 and 4). These cycloalkanes are both known to be poor substrates for Rh-catalyzed C–H borylation (6, 20, 21) because the 2°C–H bonds are relatively sterically congested and weakly acidic (32). Cyclohexane proved to be optimal, affording CH3Bpin in 74% yield with only traces (~2%) of the solvent C–H borylation product c-C6H11Bpin (Table 1, entry 4). Under these conditions, high selectivity was also observed for the mono-borylation of methane [ratio of CH3Bpin to bis-borylated CH2(Bpin)2 was 10:1]. Increasing the loading of catalyst 1 to 3 mol % resulted in 99% yield of CH3Bpin, while maintaining excellent selectivity for CH3Bpin over c-C6H11Bpin and CH2(Bpin)2 (entry 6). Lowering the catalyst loading to 0.75 mol % resulted in decreased yield of CH3Bpin (51%) but increased turnover number (68 turnovers) (entry 5) relative to the standard conditions.

Table 1 Methane C–H borylation catalyzed by 1.

Embedded Image

View this table:

We next examined Ir and Ru complexes 2/3 and 4 as potential catalysts for methane C–H borylation. These complexes were selected on the basis of their known catalytic activity for the C–H borylation of liquid alkanes (18, 19, 24, 33, 34). Under the optimal conditions for catalyst 1, the combination of Ir complex 2 and ligand 3 (18) afforded moderate yield (45%) of CH3Bpin, whereas Ru complex 4 provided 67% yield (Table 2, entries 1 and 2). To more quantitatively compare these three catalysts, reaction progress was monitored as a function of time, and the results are shown in Fig. 2. These studies show that all of the reactions achieve a maximum yield within 10 hours. However, the initial reaction rate with Rh catalyst 1 is approximately four times faster than that with 2/3. Furthermore, 4 displays a lengthy induction period (~ 2 hours), suggesting that it serves as a precatalyst for this transformation (24, 35).

Table 2 Impact of catalyst on the yield and selectivity of methane C–H borylation.

Embedded Image

View this table:
Fig. 2 Time studies showing formation of CH3Bpin using 2800 kPa of CH4.

Rh complex 1 is indicated with red squares, Ir complex 2/3 with blue circles, and Ru complex 4 with green triangles.

In the Table 2 data, the choice of catalyst has a major impact on the selectivity of C–H borylation, both for methane versus cyclohexane and for methane versus CH3Bpin. In particular, Rh and Ru catalysts 1 and 4 exhibit much higher selectivity for CH4 than does the Ir catalyst system 2/3. This effect is observed even when the reactions are stopped at similar yield of CH3Bpin (~50% yield; Table 2, entries 1, 3, and 4 for comparison). The ratio of CH3Bpin to c-C6H11Bpin ranged from 82:1 (with catalyst 4) to 3:1 (with catalyst 2/3). Similarly, the CH3Bpin to CH2(Bpin)2 ratio ranged from 31:1 (with catalyst 4) to 4:1 (with catalyst 2/3). These results indicate that tuning of the catalyst structure can be used to control this undesired over-functionalization reaction.

To more quantitatively evaluate selectivity as a function of catalyst, we conducted competition experiments between CH4 (3500 kPa, 1.1 M) (36) and CH3Bpin (0.13 M, 1 equivalent relative to B2Pin2) with each of the catalysts 1, 2/3, and 4. The time course of each reaction is shown in Fig. 3. The yield of CH3Bpin (Fig. 3, blue circles, right y axes) represents the additional CH3Bpin formed from the C–H borylation of CH4 (measured in excess of 100%, given the CH3Bpin equivalent added at the outset), whereas the yield of CH2(Bpin)2 (Fig. 3, red squares, left y axes) represents the product of C–H borylation of CH3Bpin.

Fig. 3 Evaluation of the selectivity in CH4/CH3Bpin borylation.

(A to C) Reaction time profiles (top) for (A) Ir catalyst 2/3, (B) Rh catalyst 1, and (C) Ru catalyst 4. Red squares (left y axes) represent formation of CH2(Bpin)2, and blue circles (right y axes) represent formation of CH3Bpin. CH3Bpin:CH2(Bpin)2 ratios at early time points, relative rates, and ΔΔG values for the three catalyst systems are given below their respective time profile graphs.

With Ir catalyst 2/3 (Fig. 3A), the quantity of diborylated product present exceeds that of CH3Bpin at all time points. In contrast, the concentration of CH3Bpin is much greater than that of CH2(Bpin)2 throughout the reactions catalyzed by Rh complex 1 and Ru complex 4 (Fig. 3, B and C, respectively). Using the ratio of CH3Bpin:CH2(Bpin)2 obtained at early time points and the concentrations of CH4 and CH3Bpin added at the onset, we can estimate kCH4/kCH3Bpin (and thus approximate ΔΔG) for the C–H borylation of CH4 versus CH3Bpin for each catalyst (complete details of these calculations are provided in the supplementary materials). As shown in the bottom of Fig. 3, positive ΔΔG values are observed for catalysts 1 and 2/3, reflecting faster C–H borylation of CH3Bpin versus CH4. The values of ΔΔG are estimated as 0.53 and 2.48 kcal/mol for 1 and 2/3, respectively. In contrast, Ru catalyst 4 shows a reversal in selectivity, exhibiting a preference for CH4 over CH3Bpin, with an estimated ΔΔG of –0.5 kcal/mol.

The relative reactivity of methane and ethane is another important issue (given that ethane is the secondmost abundant component of natural gas) but is rarely addressed in alkane C–H functionalization reactions. In the few reported systems in which this has been studied, ethane is usually found to be much more reactive (17, 3739). As shown in Fig. 4A, catalysts 1, 2/3, and 4 all catalyze the C–H borylation of ethane at 150°C in cyclohexane. Again, ethane borylation occurs in preference to cyclohexane borylation and shows a similar dependence on metal catalyst as with methane, with selectivities ranging from 5:1 (with 2/3) to >100:1 (with 4).

Fig. 4 Comparison of catalysts for ethane borylation.

(A) Batch ethane borylation results for catalysts 2/3, 1, and 4. (B) Methane and ethane one-pot competition. Selectivity factor is the preference for methane over ethane borylation corrected for statistics and solubility.

To probe catalyst selectivity for methane versus ethane, known molar quantities of each gas were added to the high-pressure reactor. The reactions were run to complete conversion of B2pin2, and the ratio of CH3Bpin to CH3CH2Bpin was determined with each catalyst. These ratios were then corrected for the number of C–H bonds in each substrate as well as the relative solubilities of the two gases (36). As shown in Fig. 4B, all three catalysts exhibit a >3.5:1 preference for the C–H borylation of methane relative to ethane, which is consistent with sterically controlled selectivity. Additionally, the level of selectivity varies with the catalyst. The Ir catalyst 2/3 and Ru catalyst 4 both react approximately fourfold faster with methane C–H bonds, whereas 1 is more selective for methane (approximately sixfold faster). These results further highlight the impact of catalyst on both reactivity and selectivity in the C–H borylation of light alkanes.

Overall, we have demonstrated that catalyst structure has a major impact on reaction rates and selectivities in the C–H borylation of methane. Over-functionalization of the initial product, CH3Bpin, can be limited through the appropriate selection of catalyst. These results open up exciting possibilities for catalyst design (to further modulate reactivity and selectivity in methane C–H borylation) as well as the application of the concepts delineated here for other light alkane C–H functionalization reactions.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S19

Tables S1 to S14

References (4057)

References and Notes

  1. These initial conditions were selected on the basis of (i) published conditions for alkane borylation reactions (20, 21, 24, 26) as well as (ii) the fixed volume of our Parr high-pressure reactor.
  2. The 1-catalyzed background reaction of c-C5H10 and c-C6H12 with B2pin2 in the absence of methane afforded 20.3 and 4.2% yield of the solvent C–H borylation products, respectively.
  3. All three of these reactions passed the Hg drop test (supplementary materials), which is consistent with homogeneous catalysis.
  4. The concentrations of methane and ethane under the reaction conditions were determined by using Raman spectroscopic analysis of solutions of CH4 or CH3CH3 in C6D12 (supplementary materials).
  5. Equation of state was derived from (56) as implemented in the National Institute of Standards and Technology (NIST) Webbook fluid properties calculator:
  6. Similar observations of methane in propylene carbonate are available in (57).
Acknowledgments: The work conducted by A.K.C. (primarily involving evaluation of catalysts 1 and 2/3) was supported by NSF under the Centers for Chemical Innovation (CCI) Center for Enabling New Technologies through Catalysis (CENTC) Phase II Renewal, CHE-1205189. The work conducted by S.D.S. (primarily involving evaluation of catalyst 4 and gas solubility measurements) was supported by the U.S. Department of Energy Office of Basic Energy Sciences (contract DE-FG02-08ER 15997). The Parr reactors used in this work were purchased with funds from the NSF, under the CCI CENTC Phase II Renewal, CHE-1205189. We gratefully acknowledge D. Samblanet for assistance with the gas solubility measurements. The University of Michigan has filed for a provisional patent on this work.

Stay Connected to Science

Navigate This Article