A mechanistic study into the effect of acetic acid on methanol synthesis

Abstract

The effect of the presence of higher oxygenates on the synthesis of methanol over an unmodified commercial Cu/ZnO/Al₂O₃ catalyst operating under realistic industrial conditions (523 K, 50 barg) was investigated using acetic acid as a suitable probe molecule. On addition of 1 mol% acetic acid the methanol yield decreased by 36%. The acetic acid was totally converted to a range of products namely, ethanol, methyl acetate, carbon monoxide and carbon dioxide. Use of d₄-acetic acid revealed no kinetic isotope effect in the formation of any product. However a range of CX₃CH₂OH species were formed including CD₃CH₂OH, CDH₂CH₂OH and CH₃CH₂OH revealing H/D exchange in the methyl fragment. Using [1-¹³C]-acetic acid and [2-¹³C]-acetic acid it could be shown that the production of CO/CO₂ from the adsorbed acetate was through decarboxylation. The residual adsorbed methyl fragment decomposed to give carbon, which was re-oxidised with adsorbed oxygen.

---

1. Introduction

The ability to develop highly active, selective and poison-resistant catalysts for the production of C₂₊ oxygenates is becoming increasingly important as the role of these oxygenates expand within the fuel and chemical industries.^1,2 C₂₊ oxygenates are oxygenated hydrocarbons such as acetaldehyde, ethanol, acetic acid and their higher carbon number homologues, and the ability to synthesise them has become increasingly important due to their high value and versatile applications (in comparison with hydrocarbon products such as paraffin).^1,2 As a fuel additive or alternative, C₂₊ oxygenates offer many advantages including lower hydrocarbon, CO or NO_x emissions in the exhaust gas due to more complete combustion to CO₂ and H₂O.³ In the chemical industries, C₂₊ oxygenates can be utilized in many ways, e.g. as precursors for synthesising more complex organic chemicals, in the alkylization and solubilisation of coal and as solvents in their own right.^3,4

The hydrogenation of carbon monoxide to give C₂₊ oxygenates has been of interest to researchers since as far back as 1913, when BASF discovered that cobalt and osmium catalysts could produce mixtures of alcohols, esters and acids at pressures of 20 MPa and temperatures of up to 673 K.⁴ Since that time, much work has focused on characterizing and developing the performance of catalysts suitable for the production of C₂₊ oxygenates.^1,5–9 Although several technologies have been employed commercially using heterogeneous catalysts,⁴ the hydrogenation of carbon monoxide with high activity and selectivity to a single C₂₊ oxygenate remains elusive.^8,10

Much of the recent focus in the production of C₂₊ oxygenates by hydrogenation of carbon monoxide has centred on the use of modified methanol synthesis catalysts.^6,11–14 Methanol synthesis catalysts of the formulation Cu/ZnO/Al₂O₃ have been employed in industry since their introduction in 1966 by ICI, and have proved to be both selective and active for the production of methanol from a synthesis gas feed of CO/CO₂/H₂. These catalysts have been well characterised and the mechanism of methanol synthesis over them well understood.

Modification of Cu/ZnO/Al₂O₃ catalysts to produce C₂₊ oxygenates usually involves doping of the catalyst with various additives, such as alkali or alkaline earth compounds.⁶ These additives have been shown to increase the CO dissociation capability whilst suppressing hydrogenation. This yields an increase in the olefin to paraffin ratio and an increase in selectivity to longer chain oxygenates. However as a general trend, as the selectivity to higher alcohols increases, conversion of carbon monoxide decreases.

In the synthesis of C₂₊ oxygenates over methanol synthesis type catalysts it would be of interest to determine how higher oxygenate species interact on the surface of the catalyst under working conditions. In the current study it was decided to investigate the effects of the presence of higher oxygenates on the synthesis of methanol over an unmodified commercial catalyst operating under realistic industrial conditions. Acetic acid was selected as a suitable probe molecule.

2. Experimental

For the study an industrial methanol synthesis catalyst of the formulation CuO/ZnO/Al₂O₃ was used with a typical Cu surface area of 31 m² g⁻¹. Catalytic tests were carried out using a 300 ml Berty reactor fitted with a 075-mkΙΙ magnedrive impeller for agitation. Gaseous feeds of CO, CO₂, H₂ and N₂ (all BOC, >99.99%) were monitored and controlled via the use of mass flow controllers. Nitrogen was fed as an internal standard to allow the calculation of conversion. A pressure of 50 barg was achieved by use of a gas booster unit and was set using a back pressure regulator. Acetic acid was fed into the system prior to the reactor via an HPLC pump. The liquid feed was vaporised at a temperature of 423 K and mixed with the gas feeds. All tubing pre- and post-reactor was heated to 463 K to prevent condensation of reactants and products. All valves were also heated to the same temperature in a purpose built oven. Aliquots of the effluent stream were analysed using a Thermofinnigan Trace GC fitted with two columns and TCD detectors. Helium (BOC, >99.99%) was employed as the carrier gas. Columns used were; a 50 m molsieve 5A for CH₄, CO, N₂ and C₂H₆ separation and a 30 m HP PLOT Q column for all other species. For reactions where labelled acetic acid was used, a quadrupole mass spectrometer (QMS) was connected to the effluent line post GC analysis.

All catalytic testing took place under realistic methanol synthesis conditions (523 K, 50 barg and a GHSV of 11 000). Prior to reaction, catalyst pellets (1.36 ml) were placed in the reaction basket and reduced in situ at 523 K for 3 h under a flow of 50 ml min⁻¹ H₂ at atmospheric pressure. The catalyst was then introduced to the reaction feed gas mix of N₂/CO/CO₂/H₂ in a 0.4 : 1 : 1 : 8 molar ratio. The reactor was pressurised and once all the reaction parameters had been attained analysis began. All acetic acid additions were at 1 mol% except d₄-acetic acid which was added at 1.5 mol%.

The following materials were used without further purification: acetic acid (DWR, 100%), acetic acid-d₄ (Sigma Aldrich, 99.5atom%), acetic acid-1-¹³C and acetic acid-2-¹³C (both Sigma Aldrich, 99 atom%).

3. Results

3.1 Unlabelled acetic acid addition

As a baseline experiment and to allow easy determination of mass spectra for the labelled additions, an experiment feeding in pure, unlabelled acetic acid was followed. The results are shown in Fig. 1 and 2.

Fig. 1 Reaction profile from 1.0 mol% unlabelled acetic acid addition into CO/CO₂/H₂ system over commercial methanol synthesis catalyst (dotted lines indicate addition and removal of acid).

Fig. 2 Reaction profile from 1.0 mol% acetic acid addition into CO/CO₂/H₂ system over commercial methanol synthesis catalyst (dotted line indicates addition and removal of acid).

Prior to the addition of acetic acid, the methanol synthesis system was allowed to reach steady state. A methanol yield of approximately 9% was typical, with no significant side-products detected. A reasonable concentration of water was present due to water gas shift activity. Steady levels of all components were observed after only a few hours on-line, indicating that equilibrium had been achieved.

After eight hours steady state had been attained and unlabelled acetic acid was fed into the system via the HPLC pump at low concentration (1.0 mol%, 0.307 μmole min⁻¹). Initially when the acid was introduced there was a brief increase in methanol levels, which was followed by a significant decay in methanol production: at its lowest point methanol production decreased by approximately 36%.

As well as the observed decrease in methanol synthesis activity, several other effects were noted upon addition of the acid. Ethanol and methyl acetate were detected and increased in yield throughout the addition, plateauing just before the acetic acid feed was discontinued (27 h, time on stream). An increase in water production was also observed.

The addition of the acid into the system also resulted in an increase in both CO and CO₂ concentrations. Although the decrease in conversion to methanol would account for an increase in carbon oxides, only ∼30% of the observed increase could be attributed to this route. As the only other carbon containing species present in the system was acetic acid, this indicated that it was acting as a source of CO/CO₂ upon its introduction.

Finally upon cessation of the acid feed, the system returned to steady state conditions with no apparent decrease in methanol synthesis activity. This indicated that the acid only had a transient effect on the catalyst surface, and did not alter the catalyst permanently.

3.2 Labelling studies

For the ¹³C labelling experiments the labelled acid was fed in to the system at a rate of 1.0 mol% (3.07 × 10⁻⁷ moles min⁻¹). In the deutero acid experiment, addition was made at the 1.5 mol% level (4.60 × 10⁻⁷ moles min⁻¹).

3.2.1 [1-¹³C] and [2-¹³C] labelled acetic acid additions. Using fresh catalyst the methanol synthesis reaction was again allowed to attain steady state prior to introducing the acid feed. Upon addition of the acid the system behaved exactly as previously shown, with a decrease in methanol production and increases in CO, CO₂ and H₂O concentrations. Production of ethanol and methyl acetate were also observed. Mass spectra data for both [1-¹³C]- and [2-¹³C]-acetic acid labelling experiments were found to be identical and therefore, only data for the [2-¹³C]-acetic acid run is shown.

By comparison with the data obtained from the unlabelled acetic acid experiment it was possible to follow the migration of the labelled carbon from the acid into various species. Firstly, incorporation of the labelled carbon into CO₂ and ethanol were observed by careful comparison of both the GC and MS traces, as illustrated in Fig. 3 and 4.

Fig. 3 GC data from [2-¹³C] labelled addition.

Fig. 4 Mass spectrometry data from [2-¹³C] labelled addition.

This confirmed that a route for the conversion of acid to CO₂ existed. As the mass spectra for experiments using [1-¹³C]- and [2-¹³C]-acetic acid were identical, this indicated that both carbon atoms of the acid could be converted to CO₂ requiring a breakdown of the “carbon backbone” of the acid. The ¹³C-label was unable to be observed in carbon monoxide (due to overlap with methanol 29 fragment), however we would expect it to be present via the water gas shift reaction. An increase in unlabelled CO₂ was also observed and was attributed to the decrease in methanol production as well as conversion of the unlabelled carbon of the acid.

As expected the carbon label was also observed in ethanol and methyl acetate confirming direct production of these species from the acid by hydrogenation and esterification.

Finally, no label was detected in the methanol fragments confirming that the conversion of acid to methanol did not take place.

3.2.2 d₄ labelled acetic acid addition. Upon introduction of fully deuterated acetic acid into the methanol synthesis system, no kinetic isotope effect was observed in the formation of any product. Deuterium was found to be incorporated into hydrogen, water, ethanol, methyl acetate and methanol. Incorporation of deuterium into ethanol (Fig. 5) and methyl acetate was observed to varying degrees and up to a maximum of four deuterium atoms (D) were found to be present. Exchange of deuterium from the acid with surface hydrogen was shown to proceed, as significant formation of HD took place.

Fig. 5 Mass spectra data for the addition of 1.5 mol% d₄-labelled acetic acid into a methanol synthesis system.

Incorporation of the deuterium was observed in the methanol. This may be due to deuterium on the surface of the catalyst being utilised during the hydrogenation of CO₂ or simple H/D exchange processes. The direct production of methanol from acetic acid was shown not to occur from the ¹³C labelling experiments. Post reaction TPO of the catalyst confirmed the presence of deuterium on the catalyst surface with the formation of HDO.

4. Discussion

The results from the study revealed that upon addition of acetic acid into a methanol synthesis system, a number of effects could be observed. Firstly, the level of methanol production briefly increased by ∼6% and then dropped dramatically by ∼36%, secondly ethanol and methyl acetate were formed and thirdly an increase of both CO and CO₂ concentrations was observed.

The increase in methanol production on introduction of the acetic acid followed by the drop in production can be explained by competitive adsorption of the acetic acid. Acetic acid has been shown to be adsorbed strongly onto copper surfaces in a bidentate orientation to form acetate species.^15–17 As the interaction is relatively strong, the adsorption of the acid would displace other more weakly bound methanol precursor species, such as methoxy groups, leading to an apparent increase in methanol production. The subsequent drop in production is then expected as the adsorbed acetic acid reduced the number of surface sites where methanol synthesis could take place. In keeping with this analysis of the run data revealed a molar ratio of 1.5 : 1 between acid molecules fed in and the decrease in methanol production.

As well as decreasing methanol production, the adsorbed acetate species were found to be highly reactive, and resulted in the production of two new species—ethanol and methyl acetate. Carbon labelling of the acetic acid revealed that both carbons from the acid were present in the two species, indicating that the carbon framework of the acid was left intact during their formation and no labelled species was observed in the methyl component of methyl acetate. Production of ethanol from the addition of acetic acid over copper catalysts has been observed previously in the literature by Cresseley et al.,¹⁸ and was revealed to be due to hydrogenation of the formed acetate species on the surface of the catalyst:

CH₃CO₂H + 2H₂ → CH₃CH₂OH + H₂O

Cressely studied the hydrogenation of acetic acid over Cu/SiO₂ catalysts and the results showed ethanol in high yield, along with other products such as acetaldehyde and ethyl acetate.¹⁸ Although neither acetaldehyde nor ethyl acetate was observed in the current study, methyl acetate was detected in a small concentration. In Cressely's work, the production of ethyl acetate was linked to ethanol reacting with acetic acid to form the ester.¹⁸ However in the reaction shown here, methanol was the most abundant alcohol (8 : 1 ratio of methanol : ethanol). The production of methyl acetate was therefore attributed to be the reaction of the methanol in the system with the acetic acid:

CH₃CO₂H + CH₃OH ↔ CH₃CO₂CH₃ + H₂O

Hydrogenation of the deuterated acid should produce CD₃CH₂OH (49 amu) as the principal ethanol isotopomer and indeed no evidence is obtained for D-incorporation into the –CH₂– unit. However a range of CX₃CH₂OH species are formed including CD₃CH₂OH (49 amu), CDH₂CH₂OH (47 amu) and CH₃CH₂OH (46 amu) revealing H/D exchange in the methyl fragment (Fig. 5).

Both the hydrogenation of the acid and the esterification reaction involved the production of water as a by product. The addition of deutero-labelled acid revealed the formation of HDO as expected

CD₃CO₂D + 2H₂ → CD₃CH₂OH + DHO

but also low levels of D₂O. The amount of D₂O formed is more than would be expected from random isotope exchange given the vast excess of H₂, suggesting that the concentration of D on the surface was higher than would be anticipated by the gas phase ratios.

As well as the production of ethanol and methyl acetate, the introduction of the acid into the system also resulted in an increase of both CO and CO₂ concentrations. Although, the decrease in methanol production would account for some of the carbon oxide increase, it would not account for it all suggesting that the acetic acid was contributing to the enhancement in carbon dioxide. This is supported by the labelling studies, where [¹³C]-acetic acid gave rise to [¹³C]CO₂. Studies performed on copper single crystals^17,19–22 have shown that the decomposition of surface acetates resulted in evolutions of CO₂, CH₃CO₂H, CH₄, H₂C₂O (ketene) and the deposition of carbon fragments on the surface. The desorption rates for the fragments were identical, indicating a common intermediate.¹⁷ The common intermediate was postulated to be acetic anhydride, formed on the surface of the catalyst from two acetate groups reacting together.¹⁷ However, the decomposition profile for acetate has been shown to vary markedly depending on the metal, the metal particle morphology and the presence of different adsorbates. For example the presence of oxygen on the metal was shown to have a pronounced effect on the decomposition of acetates.^19–21 Adsorbed oxygen has the effect of ordering the acetate, generating small islands whereby the attractive forces between the molecules stabilise them.^21,23,24 When acetate does decompose it happens auto-catalytically, to yield CO₂, H₂ and adsorbed carbon.^19–21,24

Overall, the most common route for acetate decomposition on metal surfaces was via decarboxylation, to yield CO₂ and CH₄:^{17,19–21,25}

CH₃CO₂H ↔ CH₄ + CO₂

However in our tests methane was not detected by GC or MS analysis. On closer inspection of the decomposition mechanism, the acetate group is proposed to “fragment” on the metal surface to yield gas phase CO₂ and an adsorbed methyl group and adsorbed H on the catalyst surface:^17,19–22

CH₃CO₂H ↔ CO₂ + CH₃(ads) + H(ads)

The methyl group then further decomposes to yield surface carbon and gas phase hydrogen.^{19–21,23,24} Density Functional Theory (DFT) has shown that this occurs by the tilting of the acetate group such that the methyl group can interact with the surface leading to C–H bond cleavage,^26,27 followed by C–C bond scission. Interaction of the acid hydrogen with the catalyst surface was revealed to take place during the addition of the deuterated acid, as HD was readily formed. However the carbon balance of the acid fed into the systems revealed that there was no surface carbon deposited during the period where the acid was introduced to the system (Table 1).

Table 1 Acid balance from unlabelled and labelled acetic acid additions

	1.0 mol% unlabelled acid addition	1.0 mol% [2-^13C] labeled acid addition	1.5 mol% d4 acid addition
Acid to EtOH (%)	14	14	21
Acid to Meth Ac (%)	1	1	1
Acid to CO/CO2 (%)	84	89	84
Total (%)	99	104	106

---

If decomposition of the acetate was taking place on the surface, and decarboxylation was the route, then re-oxidation of the surface carbon would need to take place on the catalyst surface.

To consider how oxidation of the surface carbon might proceed in the methanol synthesis system, the surface of the catalyst has to be considered. Under working conditions the surface is in a dynamic state, with a small percentage of the surface covered with adsorbed atomic oxygen.^28–30 Adsorbed oxygen is derived from the hydrogenation of CO₂ to yield methanol:

CO₂ + 2H₂ ↔ CH₃OH + O(ads)

Over the Cu/ZnO/alumina catalyst this adsorbed oxygen species has been shown, using ¹²C¹⁸O₂/¹³C¹⁶O isotope tracers to react with carbon monoxide and carbon dioxide.^31,32 Under normal conditions, adsorbed oxygen reacts with either CO or H₂ to form CO₂ or H₂O. However, if adsorbed oxygen was to react with surface carbon instead of CO or H₂ then no carbon laydown would occur. If this process were to proceed, decomposition of the acetic acid would yield CO and/or CO₂ and 2H₂, while interconversion between CO₂ and CO would take place via the WGS reaction. The proposed model explains both the CO and the CO₂ increases observed in the system. However, oxidation of surface carbon could, in principle, have taken place via a reverse Boudouard mechanism, (C(ads) + CO₂ ↔ 2CO) but under the reaction conditions the thermodynamics are not favourable (ΔG₅₂₃ = +80 kJ mol⁻¹). The same is true for the reaction of surface carbon with steam (C(ads) + H₂O ↔ CO + H₂, ΔG₅₂₃ = +60 kJ mol⁻¹).

Although no carbon was calculated to be on the surface from the acetic acid addition some carbon was known to have been deposited during the methanol synthesis reaction. Hence a temperature programmed oxidation (TPO) of the catalyst was performed as it was possible that there had been isotope exchange between the surface carbons prior to oxidation. The TPO from the experiment when [2-¹³C]CH₃COOH was added is shown in Fig. 6. Clearly there is a [¹³C]CO₂ response for the low temperature evolution. However when the same experiment is repeated with [1-¹³C]CH₃COOH no [¹³C]CO₂ is evolved. This confirms the decarboxylation decomposition mechanism with the carbon on the surface derived from the CH₃ group.

Fig. 6 TPO of a catalyst that had been subjected to a period of [2-¹³C]CH₃COOH addition during methanol synthesis.

5. Conclusion

The studies revealed that the addition of acetic acid into a working methanol synthesis system had a number of effects.

• Methanol production decreased dramatically and was attributed to a competitive adsorption effect between the adsorbed acid (acetate) and one or more of the methanol precursor species on the catalyst surface

• The acetate species was found to be highly reactive and went on to form ethanol, methyl acetate and CO/CO₂. Ethanol and methyl acetate were formed through hydrogenation of the acid and esterification with methanol respectively. The production of CO/CO₂ from the adsorbed acetate was through the decomposition on the catalyst surface, via decarboxylation. Using carbon-13 isotopic labelling it was shown that the residual adsorbed methyl fragment decomposed to give carbon and was re-oxidised with adsorbed oxygen. The use of CD₃COOD revealed no kinetic isotope effect in the hydrogenation, esterification or decomposition. Hydrogen exchange was observed in the methyl group of ethanol.

References

1. S. C. Chuang, J. G. Goodwin and I. Wender, J. Catal., 1985, 95, 435–446.
2. S. S. C. Chuang, R. W. Stevens and R. Khatri, Top. Catal., 2005, 32, 225–232.
3. S. A. Hedrick, S. S. C. Chuang, A. Pant and A. G. Dastidar, Catal. Today, 2000, 55, 247–257.
4. H. D. Schindler, Coal Liquefaction—A Research and Development Needs Assessment, US Department of Energy, Mclean, VA, 1989, vol. 2.
5. T. Koerts and R. A. van Santen, J. Catal., 1992, 134, 13–23.
6. J. C. Slaa, J. G. van Ommen and J. R. H. Ross, Catal. Today, 1992, 15, 129–148.
7. A. M. Efstathiou, T. Chafik, D. Bianchi and C. O. Bennett, J. Catal., 1994, 148, 224–239.
8. F. G. A. van den Berg, J. H. E. Glezer and W. M. H. Sachtler, J. Catal., 1985, 93, 340–352.
9. T. P. Wilson, P. H. Kasai and P. C. Ellgen, J. Catal., 1981, 69, 193–201.
10. V. Ponec, Catal. Today, 1992, 12, 227–254.
11. N. Zhao, R. Xu, W. Wei and Y. Sun, React. Kinet. Catal. Lett., 2002, 75, 297–304.
12. E. Tronconi, L. Lietti, P. Forzatti and I. Pasquon, Appl. Catal., 1989, 47, 317–333.
13. J. G. Nunan, C. E. Bogdan, K. Klier, K. J. Smith, C.-W. Young and R. G. Herman, J. Catal., 1989, 116, 195–221.
14. R. Xu, W. Wei, W.-H. Li, T.-D. Hu and Y.-H. Sun, J. Mol. Catal. A: Chem., 2005, 234, 75–83.
15. P. Baumann, H. P. Bonzel, G. Pirug and J. Werner, Chem. Phys. Lett., 1996, 260, 215–222.
16. B. A. Sexton, Chem. Phys. Lett., 1979, 65, 469–471.
17. M. Bowker and R. J. Madix, Appl. Surf. Sci., 1981, 8, 299–317.
18. J. Cressely, D. Farkhani, A. Deluzarche and A. Kiennemann, Mater. Chem. Phys., 1984, 11, 413–431.
19. M. Bowker, Catal. Today, 1992, 15, 77–100.
20. M. Bowker and Y. X. Li, Catal. Lett., 1991, 10, 249–258.
21. N. Aas and M. Bowker, J. Chem. Soc., Faraday Trans., 1993, 89, 1249–1255.
22. Z. Li, F. Calaza, F. Gao and W. T. Tysoe, Surf. Sci., 2007, 601, 1351–1357.
23. R. J. Madix, J. L. Falconer and A. M. Suszko, Surf. Sci., 1976, 54, 6–20.
24. J. L. Falconer, and R.J. Madix, Surf. Sci., 1974, 46, 473–504.
25. W. Rachmady and M. A. Vannice, J. Catal., 2000, 192, 322–334.
26. M. Neurock, J. Catal., 2003, 216, 73–88.
27. E. Hansen and M. Neurock, J. Phys. Chem. B, 2001, 105, 9218–9229.
28. M. Muhler, E. Törnqvist, L. P. Nielsen, B. S. Clausen and H. Topsøe, Catal. Lett., 1994, 25, 1–10.
29. C. T. Campbell, Appl. Catal., 1987, 32, 367–369.
30. I. Chorkendorff, P. B. Rasmussen, H. Christoffersen and P. A. Taylor, Surf. Sci., 1993, 287, 208–211.
31. G. Lui, D. Willcox, M. Garland and H. H. Kung, J. Catal., 1985, 96, 251.
32. S. D. Jackson, J. Catal., 1989, 115, 247–249.

---