Alastair R.
McInroy
a,
John M.
Winfield
a,
Chris C.
Dudman
b,
Peter
Jones
b and
David
Lennon
*a
aSchool of Chemistry, Joseph Black Building, University of Glasgow, Glasgow, G12 8QQ, UK. E-mail: David.Lennon@Glasgow.ac.uk; Tel: +44-(0)-141-330-4372
bInovyn, South Parade, Runcorn, Cheshire WA7 4JE, UK
First published on 9th December 2015
In previous work by the authors, aspects of the surface chemistry connected with methyl chloride synthesis over an η-alumina catalyst have been examined. This communication considers a role for Group 1 metal salts to modify the catalytic performance of the well characterised η-alumina catalyst. Firstly, based on a previously postulated mechanism for the reaction of methanol on η-alumina, a mechanism for methyl chloride synthesis over the η-alumina catalyst is proposed. Secondly, the validity of the new mechanism is tested by observing how the (i) type and (ii) loading of the Group 1 metal salt may perturb methyl chloride selectivity. The outcomes of these measurements are rationalised with reference to the postulated mechanism. Overall, this study represents an example of how a proposed reaction mechanism has been used to inform and guide a catalyst development strategy for a large-scale industrial process.
CH3OH + HCl → CH3Cl + H2O | (1) |
2CH3OH → CH3OCH3 + H2O | (2) |
In 2000 ICI (UK) filed a patent demonstrating improved yields of CH3Cl when an η-alumina catalyst was doped with a Group 1 metal salt.7 In order to improve the understanding of why a doped η-alumina catalyst should convey favourable catalytic performance when applied to synthesis, we examined aspects of synthesis-related surface chemistry over an un-doped commercial grade η-alumina catalyst.8 Lundie et al. probed the surface acidity of η-alumina, identifying and characterising a distribution of active sites. Four different sites were identified: (i) strong, (ii) medium-strong, (iii) medium-weak and (iv) weak Lewis acid sites.9 That work paved the way for investigations of the adsorption of the reagents (CH3OH and HCl) on this surface. The interaction of methanol on η-alumina was probed using inelastic neutron scattering (INS) in order to determine the nature and form of the adsorption complex.10 Further studies of methanol on η-alumina using a combination of infrared spectroscopy and mass spectrometry led to the proposal of a mechanism for the formation of the dimethyl ether by-product.11 Crucially, the mechanism made a case for site-selective chemistry. Next, from studies of the adsorption of HCl over η-alumina, it was possible to discern how dissociative adsorption of HCl connected with hydroxyl group formation and hydroxyl exchange reactions.12
Progressing to the bimolecular reaction over the un-doped η-alumina catalyst, temperature-programmed reaction measurements showed how modest CH3OCH3 production tracks the dominant formation of methyl chloride. A coincidence of reaction profiles was attributed to both product and by-product formation processes involving a common intermediate, namely methoxy groups residing in adsorption sites of similar Lewis acidity.13 Two reviews of these investigations of η-alumina catalysts have recently appeared: Parker and Lennon have described how INS has been used to probe aspects of methyl chloride synthesis related chemistry over η-alumina,14 and Lennon and Winfield have reviewed adsorption and reactions over η-alumina as part of a wider examination of heterogeneous catalysts used for large-scale syntheses of selected chlorohydrocarbons and fluorohydrocarbons.15
Other work carried out elsewhere has produced a series of papers describing various aspects of related chemistry and chemical engineering. Representative studies include the use of micro-reactors16,17 and the use of ZnCl2 modified zeolite and alumina based catalysts.18,19 The latter work illustrates how the addition of a zinc salt can modify the acid properties of an γ-alumina catalyst to beneficially influence product yields.
As stated above, we have previously postulated a mechanism for the reaction of methanol over an η-alumina catalyst including site selectivity.11 Linking this postulate with catalyst characterisation studies,9 the authors deduce that it should be possible to extend this model to account for actual CH3Cl production over this substrate. Such a scheme has direct relevance to the industrial operation. Thus, in this paper, we propose a mechanism for the site-selective formation of methyl chloride over an un-doped η-alumina catalyst, Fig. 1.
![]() | ||
Fig. 1 A schematic representation of the site-selective formation of methyl chloride over η-alumina. The red circles represent a strong/medium-strong Lewis acid site; the blue circles represent a medium-weak Lewis acid site; the red/blue shaded circles represent a strong/medium-strong/medium-weak Lewis acid site. The definition and form of these sites is described in ref. 9. |
Following our observations reported in the previous work on η-alumina,9–13Fig. 1 shows methyl chloride to be formed from the reaction of adsorbed methoxy species and adsorbed chloride. Importantly for methyl chloride production, the chlorine species must be adsorbed on the medium-weak site, as previous temperature-programmed measurements indicate that chloride adsorbed on strong and medium-strong sites is too strongly adsorbed to engage in methyl chloride production.12,13 In contrast, the methoxy species may be adsorbed on any of the three strongest Lewis acid sites. Following deductions on the principal step for dimethyl ether production over this catalyst,11 the key step for methyl chloride formation is believed to be concerted C–Cl formation and C–O breakage. The mechanism is consistent with reports on the process of the reverse reaction by Yates and co-workers.5
In recognition of the earlier patent report on the benefits of Group 1 metal dopants for η-alumina,7 the validity of the site-selective chemistry indicated in Fig. 1 is tested by examining how certain additives affect catalytic performance. Given that Li and Na cations may intercalate into the alumina lattice, we have examined (i) the effect of modifying a commercial grade η-alumina with K and Cs salts and (ii) the effectiveness of these new materials for methyl chloride synthesis.
The data from Fig. 2 in the range 487–541 K can be used to generate Arrhenius plots that correspond to constrained methanol conversions of ∼20%, Fig. 3. Linearity for both methyl chloride and dimethyl ether production is maintained over this temperature range and yields apparent activation energies of 88.3 (±0.9) and 94.8 (±1.1) kJ mol−1 respectively, with the error representing the standard deviation from three replicate runs. The Ea value for dimethyl ether production is comparable with that reported by de Vore et al. (98 kJ mol−1) for its formation over a γ-alumina catalyst.21 The similarity in Ea values for product and by-product production over the present catalyst suggests that they may be linked by a step common to both processes.
![]() | ||
Fig. 3 Arrhenius plots for methyl chloride (black squares) and dimethyl ether (green squares) production over the un-doped η-alumina catalyst. |
The activation energy information is helpful in understanding the different temperature profiles, Fig. 2, which are attributed to a kinetic effect. As the Ea value for CH3OCH3 is marginally higher than that for CH3Cl, the reaction rate for the former will rise faster with increasing temperature. Thus, at elevated temperatures, the surface methoxy concentration will be depleted, thereby reducing the probability of nearby/adjacent adsorbed methoxy species; a condition necessary for the formation of dimethyl ether.
Fig. 6 shows how dimethyl ether production varies with respect to temperature for two doping levels of KCl modifier; Fig. 7 shows the equivalent plot for CsCl. Here the effect is more dramatic than that observed for the formation of CH3Cl (Fig. 4 and 5), where a Group 1 metal salt loading of 1.0 mmol g(cat)−1 dramatically reduces the formation of CH3OCH3; particularly so with the Cs salt (Fig. 7). Specifically, at 560 K and with reference to dimethyl ether production over the plain η-alumina catalyst, Fig. 6 shows the dimethyl ether concentration to be reduced by 92.1% at a loading of 1.0 mmol KCl g(cat)−1, whilst Fig. 7 shows the dimethyl ether concentration to be reduced by 97.2% at a loading of 1.0 mmol CsCl g(cat)−1.
Methyl chloride selectivity, S(MC), is defined according to eqn (3),
![]() | (3) |
Collectively, Fig. 4–7 indicate that the modified alumina catalysts exhibit an improved selectivity for methyl chloride. Although the addition of the chemical modifier reduces the concentration of product and by-product, this decline in activity is significantly more marked for dimethyl ether compared with methyl chloride production. For example, at 573 K the un-doped η-alumina catalyst displays a methyl chloride selectivity of 95% (Fig. 2). However, for the case of the 1.0 mmol CsCl g(cat)−1 modified catalyst at the same temperature methyl chloride selectivity increases to >99.5% (Fig. 5 and 7). This improvement in the selectivity of the process is highly relevant when extrapolated to the large-scale unit operation.
The comparable trend for methyl chloride production as a function of Group 1 metal salt loading is shown in Fig. 9 and shows methyl chloride activity to be abruptly retarded up to a metal salt loading of ∼0.3 mmol g(cat)−1; thereafter increasing the level of the Group 1 metal salt is less influential. The reduced activity is attributed to a reduction in the surface concentration of adsorbed methoxy species due to the ion pair removing the stronger Lewis acid sites.
Fig. 9 shows that Group 1 metal salt loadings ≥0.3 mmol g(cat)−1 lead to approximately equal methyl chloride production rates, with the KCl modified alumina seemingly displaying a slightly enhanced activity at a loading of 1.0 mmol KCl g(cat)−1. However, Fig. 8 shows that, at comparable loadings, the CsCl modified catalyst exhibits the greater reduction in dimethyl ether production. Table 1 presents the methyl chloride selectivity values for the base alumina plus the modified catalysts at high loadings (reaction temperature = 563 K).
Catalyst | Methyl chloride selectivity/% |
---|---|
η-Al2O3 | 94.5 |
η-Al2O3 + 1.0 mmol KCl g(cat)−1 | 99.0 |
η-Al2O3 + 1.0 mmol CsCl g(cat)−1 | 99.6 |
Formation of dimethyl ether during methyl chloride production is problematical in the industrial operation, as the by-product can react with chlorine elsewhere in the chloromethanes production process to produce undesirable chlorinated hydrocarbon derivatives. This means that, from the industrial perspective, enhanced methyl chloride selectivity is the primary objective. Therefore, with reference to Table 1, the CsCl doped catalyst exhibits the more desirable performance: S(MC) = 99.6%. Often in heterogeneous catalysis enhanced selectivity for a specific product is achieved at the cost of activity. That principle is indeed evident here; Fig. 5 shows that doping of the base alumina catalyst with CsCl at a loading of 1.0 mmol CsCl g(cat)−1 reduces the methyl chloride production rate at 560 K by 57%. However, this residual activity is economically acceptable and the 5.1% improvement in methyl chloride selectivity indicated in Table 1 minimises significantly purification issues at the end of the commercial process.
In order to understand better the site-selective chemistry connected with methyl chloride production, it is informative to review the mechanism proposed for CH3OCH3 production.11 One way to eliminate essentially all dimethyl ether formation would be for the Group 1 metal salt to “cap” the strong and medium-strong Lewis acid sites. Adopting this assumption, the profile observed in Fig. 8 is interpreted to indicate that loadings of ≥0.3 mmol g(cat)−1 lead to the strong and medium-strong Lewis acid sites being deactivated by association with a Group 1 metal salt ion pair. Fig. 10 presents a variant of the original η-alumina/methanol mechanism (i.e. no HCl involvement) proposed in ref. 11 but, crucially, the area of the schematic diagram bounded by the solid black line signifies the “capping” of the strong and medium-strong Lewis acid sites. The unavailability of these sites to the incident methanol molecules prevents formation of CH3OCH3. This proposal is entirely consistent with the observations presented in Fig. 6–8. It is also noted that the capping of the strong Lewis acid sites will prevent the formation of surface formate species via the evolution of CH4(g) and H2(g), as previously reported for the un-doped η-alumina catalyst.11
![]() | ||
Fig. 10 Proposed methanol transformations during temperature programming of an η-alumina catalyst doped with 1.0 mmol g(cat)−1 of a Group 1 metal salt. The areas bounded by the solid black line represent chemistry no longer accessible due to the Group 1 metal “capping” the strong and medium-strong Lewis acid sites. This mechanism is a variation on that proposed for the reaction of methanol on η-alumina,11 that is reproduced with permission from Lundie et al., J. Phys. Chem. B, 2005, 109, 11592. Copyright 2005 American Chemical Society. |
The logic applied above for by-product formation can be readily extended to the industrially relevant formation of methyl chloride over η-alumina catalysts and is represented schematically in Fig. 11, which is a variant of Fig. 1. The shaded area of Fig. 11 indicates the inaccessibility of the strong and medium-strong Lewis acid sites that occurs at a Group 1 metal loading of 1.0 mmol g(cat)−1. This leaves the medium-weak Lewis acid site as the active site: both methoxy and Cl chemisorb at this class of site and combine to produce CH3Cl selectively (S(MC) = 99.6%). This proposal is consistent with the profiles depicted in Fig. 5 and 7. Importantly, the last step shown in Fig. 11, which results in the evolution of water, constitutes an active site regeneration step that allows the catalytic cycle to continue; this is a pre-requisite for steady-state operation.
• Temperature-programmed reaction measurements for methyl chloride and dimethyl ether over the un-doped η-alumina catalyst indicate activation energies of 88.3 (±0.9) and 94.8 (±1.1) kJ mol−1 respectively.
• Although doping of the η-alumina catalyst leads to a substantial reduction in the rate of CH3Cl formation (Fig. 5, 1.0 mmol CsCl g(cat)−1 reduces the production rate for CH3Cl at 560 K by 57%), it significantly improves the selectivity with respect to CH3Cl (Fig. 7/Table 1: 1.0 mmol CsCl g(alumina)−1 increases the methyl chloride selectivity at 563 K from 94.5% to 99.6%).
• On the basis of methyl chloride activity and selectivity, modification of the η-alumina catalyst with CsCl at a loading of ≥0.3 mmol g(cat)−1 is considered to present the best catalyst formulation for sustained CH3Cl production with minimal formation of CH3OCH3 where kinetic control of the product composition is possible.
• With reference to the modified reaction mechanism, this improvement in product selectivity is attributed to Group 1 salt ion pairs “capping” the strong and medium-strong Lewis acid sites of the η-alumina catalyst,9 with methyl chloride production occurring solely on medium-weak Lewis acid sites.
This journal is © The Royal Society of Chemistry 2016 |