Natural gas (NG), which is primarily composed of methane (typically 85%-90%), is one of the most abundant hydrocarbon-based feedstocks. NG is considerably more efficient and cleaner than other fossil fuels, because of its higher H/C ratio and British thermal unit value. However, the economic value of NG is greatly underplayed, partly because its gaseous nature hinders its transportation and use. Methane conversion to high-value liquid chemicals and fuel products has long been regarded as the “Holy Grail” by chemists, because this is key to use NG as an important chemical feedstock. Recently, the increase in global NG reserves and the successful extraction of unconventional shale gas have greatly stimulated the development of gas-to-liquid (GTL) technologies.
The Shilov system, first reported in 1972, has had significant impacts on the development of organometallic GTL C1 chemistry [1, 2, 3]. The Shilov process converts alkanes, including CH4, to other hydrocarbon chemicals such as CH3OH and CH3Cl via three key steps.
(1) A C-H activation step - formation of a methylplatinum(II) intermediate from the reaction of CH4 with Pt(II) (i.e., [PtCl6]2-).
(2) A functionalization step - oxidation of the methylplati-num(II) species to form a methylplatinum(IV) complex.
(3) A catalyst regeneration step - nucleophilic addition at the methyl carbon to yield an organic product, with reduction of the Pt(IV) complex (i.e., [PtCl4]2-) back to Pt(II).
The net reaction is
R-H + [PtCl6]2- + H2O → R-OH + [PtCl4]2- + 2H+ + 2Cl- (1)
One of the most important features of the original Shilov system is that it can activate the stable C-H bond in CH4 at temperatures as low as 80 °C. However, further increasing the reaction temperature (> 120 °C) often results in irreversible disproportionation and combination of Pt(II) complexes to form Pt(IV) and Pt(0), which destabilizes the Pt(II) catalyst as a result of metallic Pt(0) precipitation from the aqueous solution:
2 [PtCl4]2- → Pt(0) + [PtCl6]2- + 2Cl- (2)
Consequently, only low reaction temperatures can be used, resulting in low conversion rates.
Significant efforts have been devoted to improving the stabilities and reactivities of Pt(II)-based catalytic systems using different organic ligand molecules, but commercially practical solutions have not been found [4, 5, 6, 7, 8, 9]. For instance, one of the best-known systems was developed by Periana at Catalytica in the 1990s; the catalyst in the system is (bpym)PtCl2 [bpym = η2-(2,2'-bispyrimidyl)] [10, 11, 12]. This system shows excellent efficiency and selectivity in the conversion of CH4 to CH3SO3H, which can be further hydrolyzed to CH3OH. In contrast to the original Shilov system, which is performed in aqueous solution, Catalytica’s catalyst is deactivated by water and the CH3OH produced; therefore, the conversion reaction can only be performed in highly concentrated (> 96%) sulfuric acid, which limits its commercial feasibility [13].
Ionic liquids (ILs) are superior solvents for a wide variety of organic and inorganic compounds, and have many advantages over conventional volatile organic solvents. ILs are therefore ideal coordination ligands for organometallic catalysts and are good solvents for many catalytic systems [14, 15]. Our recent studies have identified several N-heterocyclic ILs that are both thermally and chemically stable under extreme C-H bond activation reaction conditions, which involve powerful Pt-based catalysts at ~200 °C [16]. We anticipated that addition of selected ILs to the Shilov system would reverse the Pt(0) precipitation reaction by maintaining a significant amount of the Pt catalyst in solution, therefore a higher reaction temperature could be used, and the CH4 conversion rate and catalytic effi-ciency would be improved.
In this study, we focused on developing mechanistic and kinetic understandings of high-temperature Shilov systems, to improve the stability and reactivity. Pt(II)-catalyzed CH4 conversion reactions were performed using a specially designed sealed gold-tube mini-reactor, to enable high-pressure (25.5 MPa) conditions to be used, to improve the CH4 gas solubility in aqueous solution. The solubility of methane in aqueous solution under ambient conditions is low (0.0012 g/kg of water at 60 °C), and is enhanced 1000-fold under high pressure (0.1202 g/kg of water at 20.0 MPa, and 0.1531 g/kg of water at 30.0 MPa, at 60 °C) [17]. H/D exchange experiments were performed to determine the reactivities and selectivities of the reactions. Possible mechanisms for stabilizing the Pt catalyst in aqueous solution are discussed. The stability at high temperatures is attributed to the effect of concentrated Cl-. A comprehensive reaction network using a series of first-order parallel sequen-tial substitution reactions, described by the Arrhenius equation, was set up to determine the detailed reaction kinetics of the high-temperature Shilov process.
K2PtCl4 and H2PtCl6 were purchased from Sigma-Aldrich. The ILs 1-methylimidazolium bisulfate ([1mim][HSO4]) and 1-methylimidazolium chloride ([1mim][Cl]) were purchased from Fluka, and pyrazinium bisulfate ([pyrz][HSO4]) was synthesized by protonation of pyrazine with sulfuric acid. For the H/D exchange experiments, only K2PtCl4 was used as the catalyst, and D-substituted chemicals, i.e., water (D2O), CH3COOH (CD3COOD), sulfuric acid (D2SO4), and hydrogen chloride (DCl), were purchased from Sigma-Aldrich and used as additives. For example, 30% CD3COOD solution was obtained by mixing CD3COOD with pure D2O. In the reactions, the only hydrogen source was therefore activated methane.
A mini gold-tube (length 10 cm, inner diameter 1.09 cm) was used as the mini-reactor. We chose gold as the material for the mini-reactor because of its chemical inertness, and its flexibility, which allows volume expansion and contraction by external control of the confining pressure. It ensured 100% mass balance, and uniform temperature control was provided by a large box furnace. Efficient contact between the gas-phase methane and liquid-phase reaction media containing the catalysts and liquid stabilizers was achieved by applying external hydraulic pressure.
The Pt catalyst was dissolved in various reaction media at desired concentrations. The prepared reaction liquid (0.3 mL) was loaded into the gold-tube reactor, which was presealed at one end. CH4 gas (0.34 MPa, with 3% Ar as an internal standard) was also loaded into the reactor, and the other end was sealed by arc welding. The sealed gold-tube was placed in a stainless-steel autoclave and an external hydraulic pressure of up to 25.5 MPa was applied. The autoclave with the gold-tube was placed in a box furnace to maintain the required reaction temperature through hot-air circulation. After reaction, the sealed gold tube was inserted into a vacuum line and all gas compounds were analyzed using gas chromatography (GC). The remaining liquid and solid residues (if any) were collected and analyzed. The total weights of the sealed gold-tube before and after the reaction were recorded and compared. Typical weight differences were within 0.1 mg, indicating no leakages during the reaction.
A Hewlett-Packard 6890 GC with a J&W GS-GASPRO column, thermal conductivity detector, and mass selective detector was used to determine the gas composition and for D-isotopic analyses. The amount of methane gas was deduced from the internal standard Ar integration, and the methane isomers were identified from their characteristic ion masses, i.e., 20 for CD4, 19 for CHD3, 18 for CH2D2, and 17 for CH3D. All standard methane isomer samples (16-CH4, 17-CHD3, 18-CH2D2, 19-CH3D, and 20-CD4) were previously analyzed using the same GC-MS instrument to determine their fragmentation patterns, and stored for use as internal references. The methane isomer compositions were derived from the sample fragmentation patterns using the deconvolution method.
Pt(0) precipitation in the Shilov system and the effects of different stabilizers at 140 and 185 °C were investigated, with a focus on examining the effects of Cl- concentration on Pt stabilization. Table 1 summarizes the experimental results obtained under various conditions. The total H/D exchange ratio reflects the degree of CH4 conversion, and the relative H/D exchange ratio for CH3D and CH4-xDx is used as an indication of product selectivity. The turnover number (TON) was determined from the ratio of the molar quantity of activated CH4 to that of catalyst. Inductively coupled plasma-atomic emission spectroscopy (ICPAES; Leeman Labs Prodigy) was used to confirm that the precipitated deposit was metallic Pt(0). A piece of the opened gold tube with precipitates was immersed in 2% HNO3 solution in an ultrasonic washer overnight, cleaned with deionized water, and dried under vacuum. The total weight of the solid sam-ple was recorded. The sample was then dissolved in chloroa-cetic acid and subjected to ICP-AES to determine the concen-trations of the elements in the sample solution. The results showed that the only metallic elements in the solid residue from the precipitates were Pt (0.15 wt%, 906.1 μg/g) and Au (99.85 wt%).
Shilov suggested that addition of CH3COOH might signifi-cantly increase methane conversion (~30 times) in low-tem-perature, low-conversion reactions. He attributed this to che-late formation between carboxylic anions and the Pt catalyst, which would prevent Pt precipitation. Without the addition of a chelating agent, the H/D exchange ratio was quite small in pure D2O, even when we applied a high external pressure (3700 psi) to increase methane solubility in the aqueous phase (Table 1, entry 1). Addition of 30% CD3COOD signifi-cantly promoted H/D exchange (Table 1, entry 2), even for small amounts of catalyst; this is in general agreement with Shilov’s observations. However, at 150 °C, extensive Pt(0) precipitation still occurred. This was supported by the finding that performing the reaction under the same conditions for a longer time (12 h, Table 1, entry 3) did not significantly improve the TON, suggesting that the catalyst was deactivated within 4 h, and the coating Pt(0) layer did not catalyze C-H activation. However, CF3COOD enhanced the reactivity to a lesser extent, even at a higher catalyst concentration and longer reaction time (Table 1, entry 4, 150 °C, 5 h). This is probably because the stronger electrophilicity of F atoms makes the chelating ability of the F3COO group lower than that of the D3COO group. Changing the catalytic system from K2PtCl4 to H2PtCl6 or their mixture (Table 1, entries 5 and 6) did not improve the Pt stability. These results suggest that the addition of CH3COOH is not enough to prevent Pt(0) precipitation at high temperatures. The substantial increases in the conversion could arise from the enhanced CH4 solubility.
The addition of excess DCl and D2SO4 at 140 °C can also effectively prevent Pt(0) precipitation (Table 1, entries 7 and 8). Even at a higher temperature (185 °C) and a longer reaction time (24 h), the addition of DCl stabilizes the Pt catalyst in aqueous media, but D2SO4 does not (Table 1, entries 9 and 10). The stabilizing effect of DCl is independent of the addition of carboxylic acid (Table 1, entries 11 and 12), but the Cl- concentration has a significant effect. These results suggest that Cl- strongly affects Pt stabilization, as a result of the equilibrium reaction:
Pt(0) + [PtCl6]2- + 2 Cl- → [PtCl4]2- (3)
This is the reverse of the Pt(0) precipitation reaction (reaction (2)). More importantly, the proton concentration and solution acidity have less effect on Pt(0) stabilization, because sulfuric acid did not prevent Pt(0) precipitation at 185 °C. This suggests the possibility of replacing high concentrations of H+ with more eco-friendly cations, while maintaining a high Cl-concentration, using ILs such as [1mim][Cl].
The addition of various ILs (Table 1, entries 13-18) resulted in virtually no Pt precipitation at 140 °C, indicating that selected ILs stabilized the Pt catalyst. However, the overall H/D exchange rates also decreased. Possible explanations are that the added ILs do not improve the CH4 solubility and/or formation of Pt-IL complexes decreases the catalytic activity (as in the case of Catalytica’s system, which required a temperature of ~220 °C). The selected ILs did not stabilize the Pt catalyst at 185 °C (Table 1, entries 19 and 20).
The maximum Cl- effect was determined from a series of H/D exchange experiments at various DCl concentrations (Table 2, entries 21-27, and Fig. 2). All reactions were con-ducted at 200 °C and an external hydraulic pressure of 25.5 MPa for 4 h in 30% CD3COOD with 20 mmol/L K2PtCl4 as the catalyst. The results clearly show that the amount of silver coating decreased with the increasing Cl- concentration, and the conversion also decreased. These results suggest that a highly concentrated Cl- solution effectively prevented Pt(0) precipitation at high temperatures, but the activity of the Pt catalyst was also reduced.
The addition of highly concentrated Cl- and use of a high-pressure reactor enabled high CH4 conversions to be achieved. Under these conditions, multiply substituted methane isotopologues such as CH3D, CH2D2, CHD3, and CD4 were observed, indicating that multiple C-H bond activations occurred. The high-temperature Shilov system therefore involves a complex network of multiple and sequential reactions.
A series of isothermal experiments were conducted at 200 °C (25.5 MPa external pressure). The reaction system consisted of K2PtCl4 (5 mmol/L) as the catalyst, and 30% CD3COOD and DCl (9 mol/L in D2O solution) as the reaction medium. A number of gold-tube mini-reactors loaded with the same reactants were extracted from the box furnace at reaction times of 4, 8, 20, 24, 32, and 100 h. When the gold tube was opened after the reactions, no silver coating was observed, indicating that severe catalyst deactivation had not occurred, because of the addition of excess Cl-. Individual D-substituted methane compounds were identified using GC-MS; the data are summarized in Table 3.
The kinetics of the high-temperature Shilov system, which involves multiple substitution reactions, was investigated using a reaction network involving a number of parallel reactions:
CH4 → a11CH3D + a12CH2D2 + a13CHD3 + a14CD4 (4)
CH3D → a21CH2D2 + a22CHD3 + a23CD4 (5)
CH2D2 → a31CHD3 + a32CD4 (6)
CHD3 → CD4 (7)
If the reactions are assumed to be first order, the rate constant k of each reaction can be determined from the Arrhenius equation:
k = Afexp(-Ea/RT) (8)
We developed a precursor kinetic scheme to optimize the kinetic parameters and stoichiometric coefficients. For exam-ple, the reaction kinetics for production of each of the D-substituted isotopologues (CH3D, CH2D2, CHD3, and CD4) from CH4, represented by equation (4), will be different. Four precursors, p-CH3D, p-CH2D2, p-CHD3, and p-CD4 are introduced:
CH4 → a11p-CH3D + a12p-CH2D2 + a13p-CHD3 + a14p-CD4 (4a)
p-CH3D → CH3D (4b)
p-CH2D2 → CH2D2 (4c)
p-CHD3 → CHD3 (4d)
p-CD4 → CHD4 (4e)
Reaction (4a) is considered to be fast (low activation), and CH4 is quickly transformed to the four precursors, which follow different reaction kinetics, (4b), (4c), (4d), and (4e), to generate four isotopologues. Similarly, for reactions (5) and (6), three-precursor and two-precursor schemes, respectively, were used. There are therefore 13 reactions in the overall reaction network, i.e., five for isotopic substitutions of CH4 (one for precursor generation, four for reactions of individual components), four for isotopic substitutions of CH3D (one for precursor generation, three for reactions of individual components), three for isotopic substitutions of CH2D2 (one for precursor generation, two for reactions of individual components), and one for isotopic substitution of CHD3 to form CD4. It is worth noting that we assume that the kinetic processes have not achieved final equilibrium, therefore all reactions are kinetically controlled, i.e., there are no recombination reactions involved.
Optimization and parameter fitting were performed itera-tively. The concentrations of all methane isotopologues were computed for a given parameter set (frequency factor, activation energies, and stoichiometric coefficients). An error function summing the square of errors between the calculated and measured concentrations was minimized using the simplex optimization method. In theory, the frequency factor Af depends on the reaction conditions such as temperature and pressure, and reflects the degree of molecular collisions. All the reactions considered have similar mechanisms, therefore we assume a constant frequency factory, Af = 109, for all reactions. There are a total of nine stoichiometric coefficients and 10 activation energies that need to be optimized (we assume that the three precursor generation reactions are fast, and their activation energies are small, i.e., 20 kcal/mol). The commercially available Kinetics05 and OptimRx1.35.0 software packages were used for the optimization procedure.
The optimized kinetic parameters are listed in Table 4, and the theoretical predictions made based on optimized kinetic fitting with the experimental data are shown in Fig. 3.
Table 4 Reactions in methane H/D exchange experiments for kinetic fitting, and preliminary fitting results.
Reaction
Ea (kcal/mol)
CH4→a11CH3D-p1+a12CH2D2-p1+a13CHD3-p1+a14CD4-p1
a11 = 0.561, a12 = 0.437, a13 = 0.00207, a14 = 0.0000142
E04 = 20
(4)
CH3D-p1→CH3D
E4a= 29.9
(4a)
CH2D2-p1→CH2D2
E4b = 29.9
(4b)
CHD3-p1→CHD3
E4c = 39.7
(4c)
CD4-p1→CD4
E4d = 40.0
(4d)
CH3D→a21CH2D2-p2+a22CHD3-p2+a23CD4-p2
a21 = 0.396, a22 = 0.591, a23 = 0.013
E05 = 20
(5)
CH2D2-p2→CH2D2
E5a = 29.8
(5a)
CHD3-p2→CHD3
E5b = 29.5
(5b)
CD4-p2→CD4
E5c = 31.1
(5c)
CH2D2→a31CHD3-p3+a32CD4-p3
a31 = 0.682, a32 = 0.318
E06 = 20
(6)
CHD3-p3→CHD3
E6a =29.9
(6a)
CD4-p3→CD4
E6b = 28.9
(6b)
CHD3→CD4
E07 = 30.4
(7)
First, the activation energy of CH4 → CH2D2 (29.9 kcal/mol) is nearly equal to that of CH4 → CH3D (29.9 kcal/mol), but much less than those of CH4 → CHD3 (39.7 kcal/mol) and CH4 → CD4 (40.0 kcal/mol), indicating that the primary generation reactions from CH4 to CH3D and CH2D2 occur simultaneously. Secondly, the activation energies of CH3D→CH2D2 (29.8 kcal/mol), CH3D→CHD3 (29.5 kcal/mol), and CH3D→CD4 (31.1 kcal/mol), CH2D4→CHD3 (29.9 kcal/mol) and CH2D2→CD4 (28.9 kcal/mol), and CHD3→CD4 (30.4 kcal/mol) are all very close to or less than those of the first generations of CH2D2 and CHD3. This suggests that because all four C-H bonds are equivalent, the activation conditions for the second, third, and fourth C-H bonds are the same as those for the first C-H bond. Third ly, the stoichiometric coefficients of CH4→CHD3 and CD4 are much smaller (< 0.1%) than those of CH3D and CH2D2. CH3D and CH2D2 will therefore both be initially generated from CH4. After more than 20 h, further isotopic replacements occur such that CHD3 and CD4 are generated from CH3D and CH2D2. After 100 h, all multi-substituted compounds also become reactive, and the overall process is no longer kinetically controlled. The thermodynamic equilibrium effect becomes predominant; our results indicate that CH3D and CD4 are more energetically stable than CH2D2 and CHD3.
The original Pt-catalyzed Shilov system for CH4 activation can be stabilized at high temperature (~200 °C) by the inclusion of excess Cl- in the aqueous solution. The CH4 conversion rate can be further improved by enhancing the gas solubility using a high-pressure reactor. The mechanism of Pt stabilization with concentrated Cl- is attributed to an equilibrium shift for the reaction Pt(0) + PtCl62- + 2 Cl- ↔ 2 PtCl42-; this inhibits Pt(0) precipitation. Selected chloride-based ILs also stabilize the Pt catalyst at a lower reaction temperature (~140 °C). The carboxylic group from CH3COOH or CF3COOH can chelate with the Pt catalyst, but the stabilization effects are smaller than that of Cl-.
CH4 is a highly symmetric (tetrahedral symmetry, Td) molecule in which all four C-H bonds are equivalent. A catalytic system with reaction conditions under which a C-H bond can be activated will probably also activate the other three C-H bonds, and such sequential reactions often lead to poor product selectivity. We used a high-pressure reactor and excess Cl- to maintain the catalytic activity at higher reaction temperatures and achieved nearly 100% CH4 conversion in a Shilov-type reaction. We were therefore able to study the detailed mechanism and kinetics of the reaction, including C-H over-activation reactions. A reaction network consisting of a series of parallel chemical reactions was established, and the kinetic parameters (activation energies and stoichiometric coefficients) were optimized by fitting the experimental data.
The following conclusions were drawn.
1. The H atoms in CH4 are substituted with D atoms to give CH3D or CH2D2 via a first-order substitution reaction. The theoretical fitting results gave similar activation energies (~29.9 kcal/mol), reflecting the experimental observation of early generation of both CH3D and CH2D2.
2. Direct tri-or tetra-substitutions of CH4 to form CHD3 and CD4 require much higher energies (39.7 kcal/mol and 40.0 kcal/mol) than mono-and di-substitutions. Direct conversions from CH4 to CHD3 and CD4, i.e., simultaneous activations of three or four C-H bonds, are therefore more difficult.
3. The activation energies for the conversions of CH3D and CH2D2 fall within the same range as their generation from CH4 (~29.9 kcal/mol). This suggests that mono-or di-substituted methane isotopologues are subject to further substitutions, because if a reaction system can activate one C-H bond of CH4, it is probable that other C-H bonds will also be activated.
4. In our system, where a sufficient supply of D atoms is available, the final products are CHD3 and CD4 in a molar ratio of 80:20, suggesting that multiply D-substituted CH4 molecules are thermally more stable than mono-substituted ones.
5. For CH4 conversion to CH3Cl and other CH4 chlorination reactions, excess CH4 is typically present, therefore the final product distribution will be determined by the availability of Cl atoms. If CHCl3 or CCl4 is the target end product, a sig-nificantly large amount of Cl- is needed.
6. It is difficult to achieve high selectivity for singly substi-tuted methane (i.e., CH3Cl), not only because of possible higher substitution reactions, but also because the formation of CH2Cl2 is possible. Even if higher substitution reactions can be prevented, the best achievable CH3Cl selectivity is proba-bly 50%.
2015, Vol. 36 Issue (10): 1777-1784
PDF (521 KB)
2015, Vol. 36 Issue (10): 1777-1784
PDF (521 KB)