Continuous reactions in supercritical fluids; a cleaner, more selective synthesis of thymol in supercritical CO₂

Abstract

Continuous fixed-bed catalytic Friedel–Crafts alkylation of m-cresol with different alkylating agents, isopropanol (IPA) and propylene, has been carried out using supercritical CO₂, scCO₂, as an alternative and more environmentally friendly reaction medium, for the synthesis of the fine chemical thymol. Both a solid Lewis acid catalyst (γ-Al₂O₃) and a solid Brønsted acid catalyst (Nafion^® SAC-13) have been investigated over a range of reaction conditions to optimise yield and selectivity for thymol. The reaction product distribution was found to be related to the type of catalyst employed. This is likely to have been due to the different reaction pathways through which the reaction occurred, a direct Friedel–Crafts alkylation in the case of Brønsted type acids and a Fries rearrangement when employing the Lewis catalyst. The new technique of 2DCOR-GC analysis was employed to establish the order of formation of the different species generated in the reaction over the two catalysts in scCO₂.

---

Rodrigo Amandi

Rodrigo Amandi studied chemistry in San-Pablo CEU University in Madrid, Spain, where he obtained his degree in 1999. Later on in the year he joined the research group of Dr Flaviano Garcia-Alvarado where he investigated the synthesis and characterisation of new cathodic materials for rechargeable lithium batteries. In September 2000, he started his PhD in the Clean Technology Group, at the University of Nottingham, UK, under the supervision of Prof. Martyn Poliakoff. His research was developed in collaboration with the fine chemical manufacturer Thomas Swan & Co. and was focused on the application of supercritical fluids as replacement solvents for the synthesis of fine chemicals. After completion of his PhD, in late 2003, he started a post-doc within the Clean Technology Group funded by the chemical manufacturer Schenectady Pratteln. He is currently employed as a post-doc, in the Institute of Process Engineering at the ETH in Zurich, Switzerland, under the supervision of Prof. Philipp Rudolf von Rohr.

Introduction

Friedel–Crafts alkylation of aromatic compounds is among the most fundamental and useful reactions for carbon–carbon bond formation in aromatic rings.¹ However it is one of the more wasteful in terms of by-product formation, atom efficiency and catalyst usage. Minimisation, or preferably, elimination of waste in chemicals manufacture, is a major objective of Green Chemistry. Hence, the development of economically viable and more environmentally acceptable Friedel–Crafts alkylation processes is highly desirable. One possibility is the use of reaction media such as supercritical fluids, particularly supercritical carbon dioxide (scCO₂), as a replacement for conventional organic solvents. The use of scCO₂ as the reaction medium in the above reactions might offer the following advantages: enhancement of the reaction rate;^2,3 improved heat and mass transfer at the surface of the catalyst^4,5 and there is some evidence for an increase in the catalyst lifetime due to the extraction of coke precursors.^6,7

When combined with solid acid catalysts, scCO₂ forms the basis of a very efficient process. Friedel–Crafts alkylation of aromatics has been successfully conducted in scCO₂ in batch reactors by a number of groups.^8,9 Batch reactions in scCO₂ are economically difficult to scale-up because of the cost of large high pressure vessels. However, the use of continuous processes reduces the size of the reactor required to generate a given amount of product and increases safety. This type of process¹⁰ has been demonstrated with the opening of the world's first industrial-scale, multi-purpose supercritical flow reactor in July 2002. Continuous fixed-bed catalytic flow reactors have been used in a wide range of supercritical reactions including hydrogenation,¹¹ hydroformylation,¹² etherification¹³ and indeed Friedel–Crafts type alkylation of simple aromatics.¹⁴

In this work, we describe the successful application of continuous Friedel–Crafts alkylation in the synthesis of the fine chemical thymol (6-isopropyl-3-methyl-phenol) 1. Thymol possesses some important physiological properties, and it is also widely used in perfumery as an important intermediate in the manufacture of menthol, the major component of fragrances with a peppermint odour.¹⁵ Industrially, thymol is obtained by the liquid phase isopropylation of m-cresol 2 with propene using activated alumina as catalyst.¹⁶ Wimmer et al.¹⁷ have described a process for the synthesis of thymol by reacting m-cresol with propene, but using wide and medium pore size (pore diameter of 5–7 Å) zeolites.

Scheme 1 Products formed in the alkylation of 2 over solid acid catalysts in scCO₂.

In the last few years, other types of solid acid catalysts have been employed in the synthesis of thymol, including calcined Mg–Al hydrotalcites,¹⁸ zinc aluminate spinel (ZnAl₂O₄)¹⁵ and mesoporous Al-MCM-41 molecular sieves of different Al ∶ Si ratios.¹⁹

The most important reaction parameter to be controlled in the synthesis of thymol is selectivity towards the product of interest. This is because the boiling points of all the possible isomers generated during the reaction, 2-isopropyl-3-methyl-phenol3, 4-isopropyl-3-methyl-phenol4 and 5-isopropyl-3-methyl-phenol5 are very similar, making the distillation step needed to separate all the components very difficult indeed; in particular, separation of thymol from 3. For this reason, generation of unwanted isomers is an issue that must be minimised by careful control of the reaction parameters. Also, an excess of the aromatic species is generally used to decrease the amount of dialkylated and polyalkylated derivatives formed during the reaction.

Results and discussion

Effect of catalyst

The effect of the different reaction parameters, e.g. temperature, pressure, concentration of organic material within scCO₂ and molar ratio of reactants, has been investigated in depth for both solid catalysts investigated, with the aim of optimising the yield and selectivity for 1. Table 1 compares the results obtained on the alkylation of 2 with isopropanol (IPA) at the optimum reaction temperature found for both catalysts.

Table 1 % Yield of products and selectivity towards 1 for the Friedel–Crafts alkylation of 2 with IPA over solid acid catalysts in scCO₂^a

Catalyst	% Yield^d						% Select.^e1
	1	3	4	5	6	Dialkyl.
a 200 bar, 3 ∶ 1 molar ratio (2 ∶ IPA) and 10%w/w of organic material in scCO2. b 175 °C. c 275 °C. d The remaining material was unreacted 2. e Selectivity for 1 was calculated by dividing the number of moles of 1 generated by the sum of the number of moles of 1, 3, 4, 5, 6 and dialkylated species formed in the reaction.
Nafion^® SAC-13^b	52.0	2.5	21.3	4.4	0	6.6	61.2
γ-Al2O3^c	54.0	9.7	1.0	0.4	0.8	2.0	78.6

---

It is clear from Table 1 that the major by-product formed during the reaction, when employing Nafion^® SAC-13 is 4. Further changes in the reaction pressure (100–400 bar), molar ratio of reactants from 3 ∶ 1 to 10 ∶ 1 (2 ∶ IPA) and concentration of organic material in scCO₂ (10–50%w/w) did not have a significant influence on the reaction outcome in terms of yield and selectivity of products. In contrast, when using γ-Al₂O₃ as a solid acid catalyst, the major by-product observed was 3. Also, a small amount of isopropyl-3-methyl-benzylether 6 was observed when using γ-Al₂O₃. Increasing the reaction temperature above 275 °C, caused a decrease in the yield and selectivity towards thymol as well as an increase in the yield of 4 and 5. There was a substantial increase in the by-products 3 and 6 when the concentration of organic substrate within scCO₂ (see Table 2) was increased with a corresponding decrease in the thymol yield.

Table 2 Effect of the concentration of organic material in scCO₂ for the alkylation of 2 with IPA over γ-Al₂O₃^a

%w/w of organic	% Yield^b						% Select.^b1
	1	3	4	5	6	Dialkyl.
a 200 bar, 275 °C and 3 ∶ 1 molar ratio (2 ∶ IPA). b See Table 1 for definition.
5	72.0	1.4	2.4	0.6	0.2	1.8	92.2
10	54.0	9.7	1.0	0.4	0.8	2.0	78.6
20	37.1	15.4	0.9	0.1	1.0	1.4	66.1
50	20.0	13.0	0.6	0	10.0	0.4	44.9

---

The amount of organic substrate pumped into the supercritical flow system per unit of time, is related to the residence time of reactants across the catalyst bed. As seen from Table 2, the amounts of 3 and 6 generated during the reaction decreased with increasing residence time, but also during this same increase in residence time, an increase in the yield of 1 and 4 were observed. This behaviour suggests that 6 is the first species formed in the reaction, and that 1, 3 and 4 are formed through a Fries rearrangement process in which 3 (kinetic product) is formed before 1 and 4 (thermodynamic products) according to Scheme 2. The yield of 5 also increased, albeit moderately, with increasing residence time. Formation of 5 is believed to occur via an isomerisation process from the more stable species generated in the reaction (1, 3 and 4).²⁰

Scheme 2 The proposed mechanism for the alkylation of 2 with IPA using γ-Al₂O₃ as the catalyst in scCO₂.

The alkylation of anisole (methyl-phenylether) using IPA and n-propanol as the alkylating agents over γ-Al₂O₃, was also attempted. No alkylated derivatives whatsoever were observed across the range of temperature from 100 °C to 250 °C, at 200 bar, 3 ∶ 1 molar ratio of reactants (aromatic : alkylating agent) and a concentration of organic substrate in scCO₂ of 10%w/w. The only products detected, were phenol, o-cresol, m-cresol and p-cresol. Formation of these cresols would come from the Fries rearrangement of the starting material, anisole, under the reaction conditions investigated. The fact that a non-hydroxy aromatic cannot be alkylated by IPA using γ-Al₂O₃ as catalyst, is consistent with the Fries rearrangement pathway postulated.

2DCOR-GC analysis was carried out in order to investigate further the proposed mechanism for the formation of thymol. 2DCOR-GC is a new technique developed at Nottingham,²¹ which is particularly applicable to unselective reactions monitored by GC. The technique has previously been described in detail.²¹ Briefly, 2DCOR-GC uses a set of chromatograms from a system which has been perturbed, for example by changing residence time in a reactor, and uses these chromatograms to generate a 2D map which is rich in information. One of the advantages of 2DCOR-GC is an increase in resolution obtained by stretching chromatograms in a second dimension, which is demonstrated²¹ by the resolution between the two isomers of thymol, 1 and 3, on this particular chromatographic column. In our experiments the reaction residence time was used as the perturbation in order to distinguish in which order the reaction products are produced. Only the synchronous relationships are shown in Fig. 1, because these are sufficient to show that the mechanisms of the two catalysts are clearly very different. One advantage of 2DCOR-GC is that, even at this level of analysis, it is possible to see significant differences between a series of apparently similar chromatograms. Fig. 1a shows the synchronous map produced by the reaction performed with Nafion^® SAC-13. A detailed analysis of this correlation map can be found elsewhere.²¹ However, it is clear that many products are produced from this reaction. The correlations and their strengths indicate that the reaction is most likely a direct alkylation. A more detailed analysis shows that 6 is formed first, with 3 being produced more rapidly than 1 and 4. 5 is produced at the slowest rate. By comparison Fig. 1b is the synchronous correlation map obtained from the reaction performed over γ-Al₂O₃ catalyst. It can be seen that this is a cleaner reaction producing fewer products and intermediates. The relationships between 3 and 1 and 4 indicate that all three products most probably originate from 6; this is consistent with the postulated Fries rearrangement. Again, 3 is produced more rapidly than 1 but in smaller quantities. Since the reaction residence time is the same during the perturbation for both experiments, the lack of 5 production over the γ-Al₂O₃ catalyst indicates that this reaction is not a direct alkylation, and that transalkylation reactions do not occur within the residence time in the reactor. The 2DCOR-GC analysis presented here supports the proposed routes for production of 1 over the two catalysts.

Fig. 1 2DCOR-GC synchronous map produced from the chromatograms recorded using the residence-time perturbed reaction of 2 with IPA over (a) Nafion^® SAC-13 catalyst and (b) γ-Al₂O₃ catalyst. 2DCOR-GC can produce two types of correlation maps, synchronous and asynchronous. Only the synchronous maps are shown here. If the mechanisms were related, then, under the same perturbation and level of contouring, the maps would appear identical. Clearly, the maps are different so the mechanisms cannot be identical. The maps are symmetrical about the diagonal axis. Signals along this diagonal, referred to as ‘autopeaks’, reflect the change in a particular GC signal during the course of the perturbation; a strong autopeak reflects a strong change in signal. The autopeaks are labelled 1, 2, etc according to Scheme 1 (note that the autopeaks of 1 and 3 almost overlap). Off-diagonal peaks indicate correlations between two signals centred about ν₁ and ν₂ positions. Thus, the point X indicates that the correlation between 2 and 6 is weaker in (a) than in (b), because the contour plot for X is larger in (b). In addition there are more peaks to the right of X along the bottom of (a) than in equivalent positions in (b). This indicates that 2 is correlated to more products in (a) than in (b), precisely reflecting the greater selectivity of the Lewis acid catalyst in (b). It can be seen that in both cases, 1, 4 and 3 are synchronously related, but only in (a) is there an autopeak for 5. In both (a) and (b) it can be seen that 1 and 3 are formed, and can be correlated to 6, as indicated by the direction of the cross-correlation peaks; this indicates that both species have a common relationship. Both the synchronous and asynchronous maps are available in colour as ESI‡ and a full analysis of the correlation map (a) is given in ref. 18.

Further studies focused on increasing the yield and selectivity for 1 using γ-Al₂O₃ as heterogeneous catalyst. To maximise the yield and selectivity for 1, the residence time across the catalyst bed had to be maximised by increasing the volume of the reactor tube used in the experiments, by increasing its length, as well as by decreasing the bulk flow of CO₂ through the system. Alternatively, a further increase in pressure would have increased the residence time, but as stated before, pressure did not have a major impact on the reaction outcome. Table 3 and Table 4 show the effects of increasing the reactor volume to double its original value and of decreasing the bulk flow of scCO₂ respectively.

Table 3 Influence of the reactor size on the alkylation of 2 with IPA over γ-Al₂O₃ in scCO₂^a

Reactor volume/mL	% Yield						% Select. 1
	1	3	4	5	6	Dialkyl.
a 200 bar, 275 °C, 5%w/w of organic in scCO2 and 1 ∶ 0.8 molar ratio (2 ∶ IPA).
10	55.0	1.5	1.1	0.3	0.3	2.9	90.5
20	69.5	1.2	1.0	0.3	0.3	4.2	91.0

---

Table 4 Effect of the reaction bulk flow in the synthesis of thymol via alkylation of 2 with IPA over γ-Al₂O₃ in scCO₂^a

Bulk flow/L min^−1	% Yield						% Select. 1
	1	3	4	5	6	Dialkyl.
a 200 bar, 275 °C, 10%w/w of organic in scCO2 and 1 ∶ 0.8 molar ratio (2 ∶ IPA).
0.65	48.5	5.8	0.7	0.2	0.5	3.5	82.3
0.32	46.5	6.5	0.7	0.2	0.5	3.7	80.6

---

As expected, by increasing the reactor volume to 20 mL, the yield and selectivity for 1 was increased while decreasing the amount of 3 generated. In contrast, when the flow of CO₂ through the system was decreased to half the initial rate, the reaction outcome was almost unaffected. This was unexpected, because, as explained before, a decrease in the bulk flow of CO₂ should have increased the yield of 1. Phase measurements (Fig. 2) were carried out to determine if a phase transition was occurring upon decreasing the amount of organic within scCO₂.

Fig. 2 p vs. T phase diagram for the alkylation of 2 with IPA at 1 ∶ 0.8 molar ratio (2 ∶ IPA) at 5%w/w (■) and 10%w/w (◆). The phase measurements were conducted using a fibre optic probe developed at the University of Nottingham.²²

Under the conditions studied (200 bar, 275 °C), the reaction proceeded in a single phase for both of the concentrations of organic substrate investigated. Thus, a phase transition cannot explain why the reaction outcome did not vary when the bulk flow of CO₂ was decreased. Nevertheless, a possible explanation could be that the water generated when the reaction was conducted at 5%w/w is sufficiently removed by the scCO₂ from the catalyst surface. In contrast, at 10%w/w, the water generated in the reaction is too much, and consequently the water removal from the catalyst surface is not sufficient, therefore decreasing the catalyst activity to some extent. To investigate the effect of water on the catalyst performance, water was deliberately added to the mixture of reactants in a 1 ∶ 0.8 ∶ 0.8 molar ratio (2 ∶ IPA ∶ H₂O). The results are shown in Table 5.

Table 5 Effect of water on the alkylation of 2 with IPA over γ-Al₂O₃ in scCO₂^a

	% Yield						% Select. 1
	1	3	4	5	6	Dialkyl.
a 200 bar, 275 °C, 5%w/w of organic in scCO2 and 1 ∶ 0.8 molar ratio (2 ∶ IPA).
Added H2O	46.1	4.4	0.7	0.2	0.4	2.4	85.2
No H2O	55.0	1.5	1.1	0.3	0.3	2.9	90.5

---

Table 5 clearly shows that water has an effect on the performance of the catalyst by decreasing the yield and selectivity for 1. The water generated in the reaction, in addition to the water added to the reactants, was not sufficiently removed by the scCO₂, hence a deactivation of the catalyst occurred.

Effect of alkylating agent

The effect of different alkylating agents on the alkylation of m-cresol to form thymol was also investigated. To this end, propylene was employed in the alkylation of 2 over γ-Al₂O₃, and the results obtained were compared to those achieved when using IPA. The use of propylene as the alkylating agent avoids the formation of water as a reaction by-product, therefore eliminating the catalyst deactivation previously observed. The effect of temperature was first investigated. The results are displayed in Table 6.

Table 6 Effect of the reaction temperature on the alkylation of 2 with propylene over γ-Al₂O₃ in scCO₂^a

T/°C	% Yield						% Select. 1
	1	3	4	5	6	Dialkyl.
a 200 bar, 1 ∶ 0.8 molar ratio (2 ∶ propylene) and 10%w/w of organic material in scCO2.
200	32.0	8.2	0.3	0.1	0.5	3.6	71.4
250	47.0	1.7	1.1	0.5	0.2	7.3	80.1
275	50.6	1.2	1.6	0.6	0.2	5.2	84.5
300	44.1	1.5	3.2	1.4	0.1	6.4	76.8

---

Maximum yield in conjunction with very high selectivity for 1 was achieved at 275 °C. The amount of 3 generated decreased upon increasing the reaction temperature. The yield of 3 at 275 °C, was rather low compared to the amount generated using IPA as the alkylating agent as it was shown in Table 4. In contrast, the yield of polyalkylated species increased by a large amount and the yield of 4 was also significantly higher.

Further investigations focused on whether an increase in the concentration of organic substrate, 2 and propylene, within scCO₂, had the same effect on the catalyst performance as when employing IPA as the alkylating agent (Table 7). However, yield and selectivity for thymol were maintained to similar values even upon increasing the concentration of organic substrate within scCO₂ to 40%w/w. This was thought to be related to the higher reactivity of olefins compared to alcohols. Also, the “dry” reaction conditions when employing propylene as the alkylating agent in scCO₂, clearly increased the catalytic performance of γ-Al₂O₃, therefore maximising the reaction throughput.

Table 7 Effect of the concentration of organic material in scCO₂ for the alkylation of 2 with propylene over γ-Al₂O₃^a

%w/w of organic	% Yield						% Select. 1
	1	3	4	5	6	Dialkyl.
a 200 bar, 275 °C and 1 ∶ 0.8 molar ratio (2 ∶ propylene) in scCO2.
10	50.6	1.2	1.6	0.6	0.2	5.2	84.5
20	50.2	1.6	1.3	0.9	0.2	6.5	83.2
40	49.7	1.3	1.3	0.8	0.2	7.0	82.8

---

Conclusions

The alkylation of 2 using both IPA and propylene as the alkylating agent, can successfully be carried out in scCO₂ as a continuous process to form thymol highly selectively. The choice of catalyst has been shown to be crucial for the selective alkylation of 2 with IPA in scCO₂. The use of the Lewis acid γ-Al₂O₃ limited the formation of unwanted products to a great extent compared to the solid Brønsted acid catalyst investigated (Nafion^® SAC-13). This was thought to be related to the different reaction mechanisms that were taking place: a direct Friedel–Crafts alkylation reaction when employing Nafion^® SAC-13 and a Fries rearrangement process when using γ-Al₂O₃. 2DCOR-GC has proved to be a powerful and useful tool that could be used to elucidate the order of formation of the different species generated during a reaction, and consequently understand the mechanism of the chemical processes that are occurring.

It appears that scCO₂ can effectively remove the small quantities of water that are formed in the reaction. However, increasing the amount of organic substrate pumped into the system led to a decrease in catalyst performance as a result of the larger amount of water generated. Employing propylene as the alkylating agent eliminates the water formed as a by-product, and therefore increases the catalytic performance of γ-Al₂O₃ significantly.

Ideally, the alkylation reaction could be performed using IPA as alkylating agent instead of propene, which is used in current industrial processes, thus, eliminating the handling hazards associated with the use of propene. A final question relates to the precise role of the CO₂ in this reaction. We have performed the reaction of IPA + 2 over the γ-Al₂O₃ catalyst in the absence of CO₂ (1 ∶ 0.8 ratio of 2 ∶ IPA, 0.12 mL min⁻¹ flow rate, reactor volume was 10 mL, 200 bar, 272–276 °C). Surprisingly, the overall selectivity (48.2%) for 1 was almost identical to that observed with CO₂ (48.5%). Crucially however, the yield of the two isomers 3 and 4 were doubled. This is of great importance because of the problems of separating 3 and 4 from 2. Thus, the effect of CO₂ appears to be significant in suppressing the formation of these minor products but the precise mechanism remains unclear.

Continuous flow scCO₂ systems have shown a real advantage for the alkylation of aromatic compounds, to produce species of great industrial interest, such as thymol. Further experiments, including catalyst lifetime, need to be carried out to complete full implementation of this process in the chemical industry.

Experimental

A schematic diagram of the supercritical fluid continuous flow apparatus used to carry out all the experiments described in this paper, is shown in Fig. 3. The CO₂ used as the solvent in our reactions was stored as a liquid, in a cylinder next to the apparatus. The liquid CO₂ was fed into a refrigerated reciprocating pump (D) and compressed to reaction pressure via pressure multiplication by a compressed air supply. The liquid organic substrate (A) was pumped into the system using a HPLC pump. Alternatively, liquified gaseous reactants were pumped using a JASCO PU-1580-CO2 pump. The scCO₂ and the organic substrate were mixed mechanically in a static mixer (B, unheated NPT crosspiece filled with glass beads of 1.5–2 mm diameter) before the mixture reached the reactor (C). Thermocouples placed inside the catalyst bed, in the product(s) stream leaving the reactor and in the heating block were used to monitor the reaction temperature. The reactor used in the majority of our experiments, consisted of a 12 mm (OD) 316-stainless steel tubing with an internal volume of 10 mL. Reactors with larger internal volume were also employed in some of the experiments as indicated. The reaction pressure was dropped stepwise after the reactor, to separate the product(s) from the fluid. A flow meter was connected to the vent line to measure the flow rate of the exhaust gases. The bulk flow of the gaseous CO₂ was set to 0.65 L min⁻¹ at 1 bar and 20 °C, corresponding to 1.06 g of CO₂ min⁻¹. Analysis of the samples collected was carried out using GC (Shimadzu GC-17a with an AOC 20i autosampler. The carrier gas used was helium and the detector was a FID. The GC column used was BETA DEX 110 (30 m length, 0.25 mm ID and 0.25 µm of film thickness)). The results reported in the tables reflect the general trends of yield and selectivity across all reactor fractions collected during an experiment.

Fig. 3 Schematic of the supercritical fluid continuous flow apparatus employed in this research, where D represents the SCF pump and B the static mixer where the organic substrate (A) and the CO₂ mix before entering the reactor C.

Two solid acid catalysts were employed in our experiments, Nafion^® SAC-13 and γ-Al₂O₃. Nafion^® SAC-13 (supplied by Aldrich) is a highly porous silica network wherein nanometre sized Nafion^® resin particles are entrapped. The composite catalyst SAC-13 contained 13 wt% nanosized Nafion^® particles in the porous silica matrix. The surface area of this catalyst is 200 m² g⁻¹ and its exchange capacity²³ is 0.15 mequiv. g⁻¹. γ-Al₂O₃ had a surface area ranging from 140 to 160 m² g⁻¹. Its composition was as follows: Al₂O₃: 95%w/w, SiO₂: 0.035%w/w, Fe₂O₃: 0.025%w/w and Na₂O: 0.008%w/w.

Acknowledgements

We thank Thomas Swan & Co. Ltd., Schenectady Pratteln, EPSRC, EUFP5 project “CPFCO2” and the Marie Curie Programme for funding. We also thank Maia Sokolova for the phase behaviour measurements, Dr E. Garcia-Verdugo for his help and Dr A. Biland for his helpful comments.

References

1. Friedel–Crafts and Related Reactions, ed. G. A. Olah, Wiley, New York, 1963–64, vol. I–IV.
2. R. S. Oakes, A. A. Clifford and C. M. Rayner, J. Chem. Soc., Perkin Trans. 1, 2001, 917.
3. E. J. Beckman, J. Supercrit. Fluids, 2004, 28, 121.
4. P. G. Jessop and W. Leitner, Chemical Synthesis Using Supercritical Fluids, Wiley-VCH, Weinheim, 1999.
5. A. Baiker, Chem. Rev., 1999, 99, 453.
6. B. Subramaniam, Appl. Catal., A: Gen., 2001, 212, 199.
7. B. Subramaniam, C. J. Lyon and V. Arunajatesan, Appl. Catal., B: Environ., 2002, 37, 279.
8. M. A. Harmer and K. W. Hutchenson, Chem. Commun., 2002, 18.
9. J. E. Chateauneuf, J. Foote and K. Nie, Abstr. Pap. Am. Chem. Soc., 2000, 220, 74.
10. P. Licence, J. Ke, M. Sokolova, S. K. Ross and M. Poliakoff, Green Chem., 2003, 5, 99.
11. S. van den Hark, M. Harrod and P. Moller, J. Am. Oil Chem. Soc., 1999, 76, 1363.
12. N. J. Meehan, A. J. Sandee, J. N. H. Reek, P. C. J. Kamer, P. van Leeuwen and M. Poliakoff, Chem. Commun., 2000, 1497.
13. W. K. Gray, F. R. Smail, M. G. Hitzler, S. K. Ross and M. Poliakoff, J. Am. Chem. Soc., 1999, 121, 10711.
14. T. W. Kuo and C. S. Tan, Ind. Eng. Chem. Res., 2001, 40, 4724.
15. H. Grabowska, W. Mista, J. Trawczynski, J. Wrzyszcz and M. Zawadzki, Appl. Catal., A: Gen., 2001, 220, 207.
16. W. Biederman, H. Koller and K. Wedemeyer, US Pat., 4 086 283, 1978.
17. P. Wimmer, H. J. Buysch and L. Puppe, US Pat., 5 030 770, 1991.
18. S. Velu and S. Sivasanker, Res. Chem. Intermed., 1998, 24, 657.
19. V. Umamaheswari, M. Palanichamy and V. Murugesan, J. Catal., 2002, 210, 367.
20. A. Klein and K. Wedemeyer, US Pat., 3 968 173, 1976.
21. J. R. Hyde, R. A. Bourne, I. Noda, P. Stephenson and M. Poliakoff, Anal. Chem., 2004, 76, 6197.
22. M. V. Avdeev, A. N. Konovalov, V. N. Bagratashvili, V. K. Popov, S. I. Tsypina, M. Sokolova, J. Ke and M. Poliakoff, Phys. Chem. Chem. Phys., 2004, 1258.
23. M. A. Harmer and Q. Sun, Appl. Catal., A: Gen., 2001, 221, 45.

---

Footnotes

† This work was presented in part at the Green Solvents for Synthesis Meeting, held in Bruchsal, Germany, 3–6 October, 2004.

‡ Electronic supplementary information (ESI) available: Synchronous and asynchronous correlation maps for the 2DCOR-GC data presented in Fig. 1. See http://www.rsc.org/suppdata/gc/b4/b418983c/

---