Synthesis of dimethyl carbonate in supercritical carbon dioxide

THERMODYNAMICS

Synthesis of dimethyl carbonate in supercritical carbon dioxide

D. Ballivet-TkatchenkoI, ^*; R. A. Ligabue^{I, II}; L. Plasseraud^I

^ILaboratoire de Synthèse et Electrosynthèse Organométalliques, UMR 5188 CNRS, Université de Bourgogne-UFR Sciences et Techniques, BP 47870, 21078 Dijon, France. Phone: + 33 3 80 39 37 70, Fax: + 33 3 80 39 37 72 E - mail: [e-mail]

^IIDepartamento de Química Pura, Pontifícia Universidade Católica do Rio Grande do Sul, Av. Ipiranga 6681, Prédio 12, CEP 90619-900, Porto Alegre - RS, Brazil

ABSTRACT

The reactivity of carbon dioxide with methanol to form dimethyl carbonate was studied in the presence of the n-butylmethoxytin compounds n-Bu₃SnOCH₃, n-Bu₂Sn(OCH₃)₂ , and [n-Bu₂(CH₃O)Sn]₂ O. The reaction occurred under solventless conditions at 423 K and was produced by an increase in CO₂ pressure. This beneficial effect is primarily attributed to phase behavior. The mass transfer under liquid-vapor biphasic conditions was not limiting when the system reached the supercritical state for a CO₂ pressure higher than 16 MPa. Under these conditions, CO₂ acted as a reactant and a solvent.

Keywords: Dimethyl carbonate; Supercritical carbon dioxide; Tin(IV) compounds.

INTRODUCTION

Switching to greener technologies can not only benefit the environment but also increase manufacturing efficiency and reduce waste. Green chemistry includes a number of concepts and principles (Anastas and Warner, 1998), such as waste minimization, solvent selection, atom utilization, catalysis, and alternative synthetic routes from sustainable resources. Therefore, the challenge in fundamental research is to provide tools and understanding for developing new reaction pathways and products according to these principles. For example, linear organic carbonates are interesting target since their conventional production involves the use of toxic phosgene (Shaikh and Sivaram, 1996). One of these, dimethyl carbonate (DMC), has numerous potential applications, avoiding carcinogenic risks of dimethyl sulfate and methyl halides (Tundo, 2001) and the toxicity of phosgene (Rivetti, 2000). Interestingly enough, DMC is considered an option as an oxygenate in reformulated gasoline, but its current price is too high for application (Pacheco and Marshall, 1997).

DMC preparation routes of increasing interest involve carbonate interchange, oxidative carbonylation of alcohols, and carbonation of alcohols (Ballivet-Tkatchenko and Sorokina, 2003). These reactions take place in the presence of catalysts for enhancement of productivity and selectivity. The carbonation reaction is of special interest because the co-reactants are carbon dioxide and methanol. Carbon dioxide is nontoxic, an advantage over phosgene and carbon monoxide, and easy to handle and to store, but much less reactive (Aresta and Quaranta, 1997). However, the implementation of safer technologies justifies the assessment of chemical reactions based on carbon dioxide as it can be viewed as a renewable raw material. Moreover, the use of supercritical carbon dioxide as a substitute solvent has received much attention because of its tunable properties by variation in pressure and temperature (Jessop and Leitner, 1999). Supercritical carbon dioxide may be a particularly advantageous reaction medium when it serves as both a reactant and a solvent (Ballivet-Tkatchenko et al., 2004). The improved rates for catalytic hydrogenation of carbon dioxide support this approach (Jessop et al., 1994). On the other hand, emerging strategies take advantage of multiphasic conditions under which the catalyst is insoluble in scCO₂-rich phase and the reaction products, soluble (Ballivet-Tkatchenko et al., 2003). Therefore, cost-effective operations are needed in associating reaction and separation steps in a single reactor. Moreover, recycling of the catalyst is in principle easier.

It has been reported that organometallic compounds based on Sn(IV) can produce reaction (1) with a beneficial effect of CO₂ pressure (Isaacs et al., 1999; Sakakura et al., 1999; Ballivet-Tkatchenko et al., 2000 and 2003).

However, the reaction mechanism is poorly understood, justifying further studies for improvement of catalyst design.

This paper describes the reactivity of CO₂ with n-butylmethoxytin(IV) derivatives and focuses on DMC formation in the presence of methanol under supercritical conditions in accordance with equation (1). Multinuclear magnetic resonance (NMR) and infrared (IR) spectroscopies, and volumetric experiments were used for characterization of the organometallics. The chemical reaction (1) was conducted in the absence of solvent. The effect of additives such as Lewis bases (imidazoles) on the DMC rate as well as the effect of the CO₂ pressure are discussed.

MATERIALS AND METHODS

All experiments were performed under dry oxygen-free argon using Schlenk tube techniques. The solvents were dried over appropriate dessicants and distilled under argon immediately prior to use. The butylstannanes n-Bu₃SnOCH₃ and n-Bu₂Sn(OCH₃)₂ were synthesized from n-Bu₃SnCl and n-Bu₂SnCl₂ (Aldrich), respectively (Ballivet-Tkatchenko et al., 2000). The butyldistannoxane [n- Bu₂(CH₃O)Sn]₂O was obtained by reaction between an equimolar mixture of n-Bu₂SnO (Aldrich) and n-Bu₂Sn(OCH₃)₂ in accordance with the literature (Ballivet-Tkatchenko et al., 2003). Methanol (Aldrich, 99.8+ %) was dried over magnesium methylate. CO₂ N45 TP was purchased from Air Liquide. NMR spectra were recorded at 295 K in deuterated chloroform (Aldrich 99.9 %) on a Bruker Avance 300 spectrometer (¹H = 300.131, ¹¹⁹ Sn = 111.910, and ¹³C = 75.475 MHz). Infrared (IR) spectra were recorded on a Bruker Vector 22 equipped with a Specac Golden Gate^TM ATR device.

In a typical experiment under CO₂ pressure, 20 cm³ of methanol were added to the tin compound (4 mmol based on tin) in a Schlenk tube. Then the solution was transferred to a 120 cm³ stainless steel batch autoclave. Finally, CO₂ was introduced at 4 MPa, and room temperature. The autoclave was heated up to 423 K (controlled by an internal thermocouple) and the pressure was adjusted to the desired value by a high-pressure CO₂ pump (Top Industrie S.A.). After 12 h of reaction under magnetic stirring, the autoclave was cooled down to 273 Kand depressurized and the condensed phase transferred to a Schlenk tube. A trap-to-trap distillation under vacuum at room temperature allowed separation of the organics, which were analyzed by GC (Fisons 8000, J&W Scientific, DB-WAX 30m capillary column, FID detector), and the tin residue was characterized by multinuclear NMR and IR spectroscopies.

RESULTS AND DISCUSSION

The insertion of CO₂ into Sn-OR bonds of the compounds n-Bu₃SnOR, n-Bu₂Sn(OR)₂, and [n-Bu₂(RO)Sn]₂O occurs readily at room temperature (R = CH₃, i-C₃H₇). The new species, characterized by multinuclear NMR, IR, volumetry, and elemental analysis, correspond to alkylcarbonate tin(IV) derivatives. The ¹³C NMR signature of the alkylcarbonato fragment (Sn-OCO₂R) is evidenced by a signal in the 159-156 ppm range. Moreover, a strong IR band due to the CO₃ group is also present at 1660-1600 cm^-1. The quantitative determination of carbon dioxide uptake corresponds to the transformation of only one alkoxy group, whatever the number of alkoxy groups originally found around tin. In the din-butyl series, this result is consistent with the ¹¹⁹Sn NMR spectra, which show resonances between –130 and –220 ppm corresponding to pentacoordinated tin atoms (Smith and Tupciauskas, 1978). Accordingly, dimeric structures are proposed as shown in Figure 1.

Recent determinations of (CH₃)₂Sn(OCH₃)₂ structure and its carbonated form by single-crystal X-ray diffraction analysis corroborate these assignments (Choi et al., 1999). Therefore, dimers can be viewed as stable structures, either in the solid state or in solution, at room temperature. A modeling study by DFT calculations also shows that the dimer is more stable than the monomer (Ballivet-Tkatchenko et al., 2002). Alkoxy groups ensure this stability due to their oxygen atoms acting as a Lewis base. Only one of the two alkoxy groups is involved in this bonding mode, leaving the other alkoxy moiety in the terminal position. As shown by the X-ray diffraction study, the insertion of CO₂ into the Sn-OCH₃ bond only takes place with the terminal alkoxy due to the higher basicity of the oxygen atom. The stannyl carbonate fragment Sn-OCO₂CH₃ thus formed is reversibly converted to the starting Sn-OR species under vacuum at room temperature. This property is of primary importance for the reactivity of the system for DMC formation.

The adducts with n-Bu₃SnOR and [n-Bu₂(RO)Sn]₂O are more stable; the quantitative de-insertion of CO₂ is now observed under vacuum treatment between 303 and 363 K, depending on the R group. The [n-Bu₂(RO)Sn]₂O compound also has a dimeric structure as well as its carbonated form (Figure 2).

In this distannoxane series, alkoxy and oxo groups ensure the stability of the dimeric form through four coordination bonds. Such structures are well documented for tetraorganodistannoxanes and related systems (Chandrasekhar et al., 2002).

When the compounds n-Bu₂Sn(OCH₃)₂ , [n-Bu₂(CH₃O)Sn]₂ O, and n-Bu₃SnOR were dissolved in methanol and then mixed with CO₂, and heated in a batch reactor for 12 h, DMC is selectively formed with a maximum yield at 423 K. The order of reactivity found, n-Bu₂Sn(OCH₃)₂ > [n-Bu₂(CH₃O)Sn]₂ O > n-Bu₃SnOR, indicates that the best DMC yield is obtained for the compound having two methoxy groups per tin (Table 1, entries 1-3). Concomitantly, the DMC:Sn molar ratio is close to one.

This result is in agreement with the carbonation of only one methoxy group (Figures 1, 2). CO₂ pressure effect was also found to have an effect. Increasing the pressure from 9 to 20 MPa led to a progressive increase in DMC yield from 0.4 to 0.9 (DMC:Sn molar ratio). A further increase up to 25 MPa had no effect. Several explanations can account for this observation. On the one hand, as the CO₂ adduct is reversible, the carbonated species should be favored at higher pressures and on the other hand, the phase diagram of the reaction mixture may be such that phase changes may occur in this pressure range.

As we had no opportunity to perform spectroscopic measurements at 423 K and 20 MPa, we decided to check the first hypothesis by changing the electronic properties of tin in adding Lewis bases such as imidazole (HIm) and 1-methylimidazole (1-MeIm) to the reaction mixture. Coordination of imidazoles is known for Sn(IV) compounds forming 1:1 and 2:1 adducts (Pettinari, 1999). If coordination occurs under the reaction conditions, the monomeric form n-Bu₂Sn(OCH₃)₂ will be stabilized and the two methoxy groups, carbonated. Therefore, the DMC:Sn ratio will reach a value of two. The results reported in Table 1 (entries 4-7) show no drastic alteration in the DMC yield with imidazoles as additives.

The study of the phase diagram was more informative. In this work, the pressures applied ranged from 9 to 25 MPa. They are beyond the critical pressures of CO₂ and methanol, 7.38 and 8.09 MPa, respectively, but the reaction temperature (423 K) is in between the critical temperatures of CO₂ (Tc = 304.2 K) and methanol (Tc = 512.6 K). Accordingly, the binary mixture may undergo phase changes in pressure and CO₂-methanol composition. The liquid-vapor (LV) case consists in a methanol-rich liquid phase in which the tin compounds are soluble. The chemical reaction will therefore take place in the liquid phase. The solubility of CO₂ in methanol has been reported to decrease with increasing temperature and to increase with pressure (Chang et al., 1997). Accordingly, in our study under isothermal conditions an increase in CO₂ pressure will shift the equilibria depicted in Figures 1 and 2 to the carbonated specie thus enhancing the rate of DMC formation, in agreement with the experimental results. However, the biphasic LV regime may no longer be valid for high pressures. One might expect to switch to monophasic conditions. Hence, this phase effect may be superimposed.

The CO₂-methanol phase behavior has been investigated by a number of researchers. Data available for the conditions defined in our study (423 K, 60 < CO₂ < 76 mol%) point to a corresponding critical pressure of around 16 MPa (Wells et al., 2003). In order to discover to what extent the addition of tin compounds affects the system, the phase behavior of the three-component mixture was visually examined by conducting the reaction in a 30 cm³ stainless steel autoclave equipped with sapphire windows (Top Industrie S. A.). We observed the changes at the liquid-vapor interface while heating the reactor from room temperature to 423 K with an initial CO₂ composition of 76% (mol), providing a final pressure of 23 MPa at 423 K. The autoclave was loaded with a methanolic solution of 0.47% (mol) n-Bu₂Sn(OCH₃)₂ at room temperature and then CO₂ was added. Upon addition, the liquid phase expanded due to CO₂ absorption. Heating led to the progressive shrinking of the liquid phase. The merging of the liquid and vapor phases was observed around 16 MPa and 410 K. Hence, on the basis of this observation, the increase in DMC rate with CO₂ pressure occurs with a phase change from a biphasic LV system to a monophasic supercritical one. The benefit arises from the reactants being in a single-fluid phase so that the chemical reaction occurs homogeneously, avoiding CO₂ mass transfer limitations. At the present level of our understanding, it is the most important contribution of the supercritical conditions. However, specific solvation properties of the supercritical fluid should not be underestimated (Brennecke and Chateauneuf, 1999) and further work in this direction is in progress.

CONCLUSIONS

The carbonation of methanol to form dimethyl carbonate in the presence of n-butylmethoxytin compounds was achieved under solventless conditions at 423 K. Mass transfer of gaseous CO₂ into the methanolic solution was limited under liquid-vapor biphasic conditions. The rate of DMC formation was lower than under monophasic conditions. The single-fluid phase was reached by increasing CO₂ pressure, resulting in a supercritical medium where CO₂ acted as a reactant and a solvent. This case study herein described highlights of optimization of a chemical reaction that can be achieved by controlling the phase behavior of the mixture. The benefit of increasing the rate and selectivity is regularly emphasized by experts in the domain of supercritical fluids. The main information still needed is the description of phase equilibria existing in the reactor as a function of pressure and temperature. This objective requires an interdisciplinary approach with complementary expertise in synthetic chemistry, analytical chemistry, and chemical engineering.

ACKNOWLEDGEMENTS

We are grateful for the financial support of this work received from the Centre National de la Recherche Scientifique and from the Ministère de la Recherche as a postdoctoral grant (RL).

NOMENCLATURE

n-Bu

normal butyl chain

¹³C NMR carbon-13 nuclear magnetic resonance spectroscopy CH₃ methyl i-C₃H₇ isopropyl CO₂ carbon dioxide DFT density functional theory DMC

dimethyl carbonate

HIm imidazole IR infrared spectroscopy LV liquid-vapor 1-MeIm 1-methylimidazole mol mole NMR

nuclear magnetic resonance spectroscopy

CH₃O methoxy group P pressure (MPa) Pc critical pressure (MPa) ppm NMR chemical shift in part per million R

alkyl group

sc supercritical ¹¹⁹Sn tin-119 nuclear magnetic resonance NMR spectroscopy T temperature (K) Tc

critical temperature (K)

t time (h)

Received: October 20, 2004

Accepted: October 21, 2005

• Anastas, P. T. and Warner, J. C., Green Chemistry: Theory and Practice. Oxford University Press, New York (1998).
• Aresta, M. and Quaranta, E., Carbon Dioxide: A Substitute for Phosgene, CHEMTECH 27, 32 (1997).
• Ballivet-Tkatchenko, D., Douteau, O. and Stutzmann, S., Reactivity of Carbon Dioxide with n-Butyl(phenoxy)-, (alkoxy-), and (oxo)stannanes: Insight into Dimethyl Carbonate Synthesis, Organometallics, 19, 4563 (2000).
• Ballivet-Tkatchenko, D., Jerphagnon, T. and Chermette, H., CO₂ as a C1-Building Block for Dialkyl Carbonates. Environmental Challenges and Greenhouse Gas Control for Fossil Fuel Utilization in the 21st Century, Kluwer Academic/Plenum Publishers, 371 (2002).
• Ballivet-Tkatchenko, D., Picquet, M., Solinas, M., Francio, G., Wasserscheid, P. and Leitner, W., Acrylate Dimerisation under Ionic Liquid-Supercritical Carbon Dioxide Conditions, Green Chemistry, 5, 232 (2003).
• Ballivet-Tkatchenko, D. and Sorokina, S., Linear Organic Carbonates. Recovery and Utilization of Carbon Dioxide, Kluwer Acad. Pub., Dordrecht, pp. 261-277 (2003).
• Ballivet-Tkatchenko, D., Jerphagnon, T., Ligabue, R., Plasseraud, L. and Poinsot, D., The Role of Distannoxanes in the Synthesis of Dimethyl Carbonate from Carbon Dioxide, Appl. Catal. A: General, 255, 93 (2003).
• Ballivet-Tkatchenko, D., Camy, S. and Condoret, J.-S., Carbon Dioxide, a Solvent and Synthon for Green Chemistry. Environmental Chemistry, Springer, Berlin, chap. 49 (2004).
• Brennecke, J. F. and Chateauneuf, J. E., Homogeneous Organic Reactions as Mechanistic Probes in Supercritical Fluids, Chem. Rev., 99, 433 (1999).
• Chandrasekhar, V., Nagendran, S. and Baskar, V., Organotin Assemblies Containing Sn-O Bonds, Coord. Chem. Rev., 235, 1 (2002).
• Chang, C. J., Day, C-Y., Ko, C-M. and Chiu, K-L., Density and P-x-y Diagrams for Carbon Dioxide Dissolution in Methanol, Ethanol, and Acetone Mixtures, Fluid Phase Equilibria, 131, 243 (1997).
• Choi, J-C., Sakakura, T. and Sako, T., Reaction of Dialkyltin Methoxide with Carbon Dioxide Relevant to the Mechanism of Catalytic Carbonate Synthesis, J. Am. Chem. Soc., 121, 3793 (1999).
• Isaacs, N. S., O'Sullivan, B. and Verhaelen, C., High Pressure Routes to Dimethyl Carbonate from Supercritical Carbon Dioxide, Tetrahedron, 55, 11949 (1999).
• Jessop, P. G., Ikarlya, T. and Noyori, R., Homogeneous Catalytic Hydrogenation of Supercritical Carbon Dioxide, Nature (London, United Kingdom), 368, 231 (1994).
• Jessop, P. G. and Leitner, W, Chemical Synthesis Using Supercritical Fluids. Wiley-VCH, Weinheim (1999).
• Pacheco, M. A. and Marshall, C. L., Review of Dimethyl Carbonate Manufacture and its Characteristics as a Fuel Additive, Energy & Fuels, 11, 2 (1997).
• Pettinari, C., Organotin(IV) Derivatives of Imidazoles, Pyrazoles and Related Pyrazolyl and Imidazolyl Ligands, Main Group Met. Chem., 22, 661 (1999).
• Rivetti, F., The Role of Dimethyl Carbonate in the Replacement of Hazardous Chemicals, C. R. Acad. Sci. Paris Série IIC Chimie, 3, 497 (2000).
• Sakakura, T., Choi, J-C., Saito, Y., Masuda, T., Sako, T. and Oriyama, T., Metal-Catalyzed Dimethyl Carbonate Synthesis from Carbon Dioxide and Acetals, J. Org. Chem., 64, 4506 (1999).
• Shaikh, A-A.G. and Sivaram, S., Organic Carbonates, Chem. Rev., 96, 951 (1996).
• Smith, P. J. and Tupciauskas, A. P., Chemical Shifts of ¹¹⁹Sn Nuclei in Organotin Compounds, Ann. Reports on NMR Spectroscopy, 8, 291 (1978).
• Tundo, P., New Developments in Dimethyl Carbonate Chemistry, Pure Appl. Chem., 73, 1117 (2001).
• Wells, P. S., Zhou, S. and Parcher, J. F., Unified Chromatography with CO₂-based Binary Mobile Phases, Anal. Chem., 75, 18A (2003).

*

To whom correspondence should be adressed

Publication Dates

• Publication in this collection
  07 July 2006
• Date of issue
  Mar 2006

History

• Received
  20 Oct 2004
• Accepted
  21 Oct 2005