Catalysis in an Ionic Fluid, Supercritical Fluid Two Phase System
The invention relates to a process to carry out catalytic reactions involving a permanent gas as a reactant using an ionic liquid and a supercritical fluid.
Supercritical fluids are fluids heated above their critical temperatures (usually compressed gas) . Such fluids can move between different states of density without any phase transitions. Since the supercritical fluid can change density continuously, slight changes of temperature or pressure manipulates the thermodynamic and transport properties of the fluid.
Catalytic reactions may be carried out in ionic liquids, in supercritical fluids or in two-phase mixtures where one phase is an organic solvent and the other is either water or a fluorinated solvent. It is also known that supercritical fluids can be used to extract organic materials from ionic
liquids. It is further known that catalytic reactions of two or more liquid or solid reagents can be carried out in ionic liquid/supercritical fluid biphasic mixtures.
However catalytic reactions carried out in ionic liquids which involve permanent gases are often not very successful and a second (organic) solvent may be added to improve the process.
It has unexpectedly been found that the use of a supercritical fluid in catalytic reactions which involve the use of a permanent gas reactant allows unexpected improvement of the catalytic reaction in terms of chemoselectivity to the desired product, regioselectivity to the desired isomer of the product, product recovery and/or separation of the product from the catalyst.
The present invention provides a biphasic reaction system, said system comprising an ionic liquid as a first phase and a supercritical fluid as a second phase, said system further comprising a permanent gas as a reactant.
The present invention further provides a process for carrying out catalytic reactions in a biphasic system comprising an ionic liquid as a first phase and a supercritical fluid as a second phase, said process comprising providing a permanent gas as a reactant.
The reaction mixture will generally comprise at least one reactant and at least one catalyst in addition to the permanent gas and the biphasic system of supercritical fluid and ionic liquid. Advantageously such reactions can be carried out in a continuous flow mode.
The ionic liquid may be any salt A+B that has a melting point below the temperature at which the reaction is carried out. Some non-exclusive examples include salts where A+ has the structure (I) shown below.
R, R1, R2, R3 and R4 are all organic groups which might be independently chosen in the group consisting of H; aryl groups; straight or branched alkyl groups, preferably having a chain length of 1-28 carbon atoms, optionally branched with alkyl or aryl groups as described above, but in which one or more of the CH2 groups is substituted by 0, S, Se, NH, NR5, PH, PR6, SiH2 or SiR7 2 (where R5, R6 and R7 are all defined as R to R4 above) .
Alternatively, A+ may be a quaternary phosphonium cation of the formula PR8R9R10R11+; a quaternary ammonium cation of the formula NR12R13R14R15+ wherein R8 to R15 are all organic groups which might be
independently chosen in the group consisting of H; aryl groups; straight or branched alkyl groups, preferably having a chain length of 1-28 carbon atoms, optionally branched with alkyl or aryl groups as described above, but in which one or more of the CH2 groups is substituted by 0, S, Se, NH, NR5, PH, PR6, SiH2 or SiR7 2 (where RΞ, R6 and R7 are all defined as R to R4 above) . A+ may also be a cationic form of an heterocycle such as a pyridine, pyrimidine, pyrrole or other nitrogen or phosphorus containing heterocycles, which may optionally be substituted on the ring, including on the N or P atom, with alkyl or aryl groups similar to the radicals R to R4 described above.
B" may be any anion which in combination with the cation A+ affords the appropriate melting point. Non exclusive examples are: halides (like F", Cl", Br~ or I~) ; EX6 ~, where E represents P, As or Sb and X represents F, Cl, Br or I; MX or M2X7 ~ where M represents B, Al, Ga I or Tl and X is an halide as described above.
The supercritical fluid is chosen so that its critical point is below that of the reaction conditions. For most of the catalytic reactions carbon dioxide is a suitable fluid. However other gases like toluene, ethane, ethene, propane, propene, other hydrocarbons with 4-10 carbons atoms, ammonia and S02 may also be suitable.
The gaseous reactants involved in the catalytic reaction are compounds which behave as permanent gases in the conditions of the reaction. Non- exclusive examples include: hydrogen, carbon monoxide, nitrogen, methane, ethane, ethene .
The catalytic reaction carried out according to the process of the invention may be any catalytic reaction in which at least one of the reactants is a permanent gas. Non-exclusive examples include catalytic hydrogenation, carbonylation, hydrocarbonylation and multistep reactions in which one step involves a permanent gas. Hydroformylation is particularly preferred.
The secondary reactants which can be used for these reactions include any compound or mixture of compounds capable of undergoing catalytic reactions with one or more permanent gases. Non-exclusive examples include compounds containing double or triple bonds between carbon atoms, between carbon and nitrogen, oxygen or phosphorus, between nitrogen and oxygen or between phosphorus and oxygen. Suitable compounds may also comprise a hydro-carbon chain which is optionally substituted at any point with groups defined as for R to R4. Such reactant compounds can also comprise one or more functional moiety like alkenes, dienes, compounds containing several double bonds, imines, alkynes, cyanides, nitro compounds, alcohols, aldehydes, carboxylic acids, carbon monoxide, carbon dioxide.
Preferably, reactant (s) and product (s) of the reaction show some solubility in the supercritical fluid at the reaction temperature and pressure. This allows a continuous flow process to be carried out, and to simplify the recovery of the product (s) and/or reactant (s) .
The catalyst is any compound or combination of compounds that can catalyse the required reaction. Preferably the catalyst is soluble in the ionic liquid and insoluble in the supercritical fluid at the reaction temperature and pressure.
Non-exclusive examples include metal complexes having ligands bound to the metal through one or more electron donor atoms like H, C, Si, Ge, Sn, Pb, N, P, As, Bi, 0, S, Se, Te, F, Cl, Br or I. The ligand may contain more than one of these donor atoms and may form a ring with the central metal. Other examples of suitable catalysts include Bronstead or Lewis acids, a Lewis base or even a component of the ionic liquid or a product of its reaction with an additive such as water.
The process can be carried out in batch mode with in si tu decompression of the supercritical fluid followed by recovery of the products from the ionic liquid, with which they may or may not be miscible.
In one embodiment the process is carried out in batch mode as described above but after the reaction has taken place a stream of fresh
supercritical fluid is passed through the reactor and the superpercritical fluid containing the product (s) of the reaction is flushed from the reactor. This can be carried out either at the reaction temperature and pressure or at some other suitable temperature and/or pressure. The conditions of the reaction together with the reactants and products are advantageously chosen so that the supercritical fluid will drive the reaction products, the permanent gas and any unreacted reactants out of the reactor where they can be easily recovered and separated.
According to a preferred aspect of the invention the process is carried out in continuous flow mode. In this case, the reactants are mixed with the supercritical fluid and continuously fed into the reactor to contact and react with the ionic liquid and the catalyst. The stream which exits from the reactor may carry (in addition to the supercritical fluid) the products of the reaction, unreacted permanent gas(es) and any remaining reactants. Of course the conditions of the reaction together with the reactants and product (s) are advantageously chosen so that they show some solubility and can be mixed with the supercritical fluid.
In either of the last two cases the products obtained by the process may be recovered by decompression (to remove the permanent gases and the supercritical fluid) and purified by fractional distillation or crystallisation. The supercritical
fluid with or without the permanent gas(es) and any unreacted reactants can be then recompressed and re-introduced into the reactor.
The catalytic reaction may be carried out in a single reactor, or in a series of interconnected reactors using a multipass system. In such a system the stream of gases at the reactor outlet is continuously fed back into the reactor while a proportion of the product of the reaction is removed from the exit stream and replaced with fresh reactants, permanent gas(es) and supercritical fluid.
The reaction temperature is determined as a function of the particular catalytic reaction which is to be carried out and should be selected to allow the catalytic reaction to take place. Such temperatures usually range between -196 and 500°C.
The partial pressure of the permanent gas(es) used may be any pressure which allows the reaction to take place, but preferably ranges between 10~7 and 1000 bar. The overall pressure, which is made up of the partial pressure of the permanent gases, the supercritical fluid and the vapour pressures of any other volatile components may be any pressure at which the reaction will occur, but is preferably chosen so that the phase excluding the ionic liquid is a single supercritical phase or liquid in contact with a gas. The overall pressure is preferably less than 1000 bar.
Another aspect of the invention is a reactor designed to carry out a catalytic reaction according to the process of the invention in a continuous flow mode.
There are several techniques that can be employed in support of the development and scale-up of a supercritical (SCF) process. One such method is analytical supercritical fluid extraction (SFE) , which can be used to study the effects of matrix composition, sample preparation and extraction variables before scale-up. The variables, which affect the efficiency of an extraction or the ability to perform selective extractions, are pressure, temperature, flow rate, nature of the matrix and composition of extracting fluid. Figure 1 shows a generic diagram of a typical analytical SFE system. In principle the operation of analytical SFE is identical to its process analogue although in general the recycling of solvent is not practised.
The gaseous solvent is delivered from a cylinder and at a constant rate by a compressor or pump. The fluid is preheated and may be premixed with a modifier prior to passage through an extraction cell. A restriction unit is then used to control pressure levels upstream. Alternatively, system pressure can be controlled by the pump and flow rates controlled through a series of expansion valves situated after the extraction vessel. Analysis of the extract in this "off-line" set up
is then performed by conventional means . Alternatively an analytical instrument can be incorporated into the system for on-line anaylsis of the extracting fluid phase. The fluid phase is often analysed by UV-visible spectroscopy, infrared spectroscopy or chromatography. Whether performing analysis on-line or off-line, analytical SFE can be used to study the effects of the aforementioned factors on a reduced scale. The only significant difference between analytical and process SFE is that of fluid purity. Analytical studies require high purity solvents whereas cost effectiveness is of greater importance in industrial applications. Fluid purity is therefore compromised, with lower purity grades being used, to make a large-scale operation economically viable.
For continuous flow reactions, in which a supercritical fluid is used to transport reagents into and products from a reaction medium, the design of a laboratory scale continous flow reactor is similar to that of the SFE unit described, with the exception that a reaction takes place in the extraction vessel. Like an SFE system, pressure can either be controlled upstream, with gas flow rates set through a series of expansion valves, or downstream, with the gas being metered into the system at a constant rate. In principle, both of these approaches are equally effective on a laboratory scale although there are differences in the cost-effectiveness of their construction and commissioning.
A laboratory-scale, continuous flow reactor, has been developed which enables the transport of both liquid and gaseous reagents into and produces from an ionic liquid/catalyst phase using scC0 as the transport medium. The reactor has been developed to study homogeneous catalysis in supercritical fluid-ionic liquid biphasic systems and is depicted in Figure 3. The feasibility of the process for the hydroformylation of long-chain alkenes has been demonstrated and the operation of the flow reaction to this effect is as follow: SFC/SFE grade C02 is fed from a cylinder (1) , via a dip tube, into a refrigerated reservoir which supplies liquid C0 to ■ an air driven liquid pump (2) . The C0 pump is specifically designed for the pressurisation of liquefied gases and delivers the gas to a constant pressure. The pressure of the C02 supply is simply controlled by regulating the air supply to the pump, which works on a compression ratio of 115:1. The pressurised C0 then passes through a high- pressure regulator (3) that controls system pressure down stream up to the point of decompression (11) . The C02 feed then passes through a non-return valve (4) and a T-piece where the CO/H2 is also introduced into the system. The synthesis gas feed is delivered from a second cylinder (5) and is passed through an air driven gas booster (6) capable of increasing the cylinder supply pressure up to 750 Bar. The synthesis gas fee is then metered into the system by a dosimeter (7).
The CO/H2/C02 gas stream passes through a second non-return valve (8) , situated directly before the autoclave (9) , where the gas stream is introduced into the ionic liquid phase via a dip tube . The liquid reactant is introduced separately, into the ionic liquid/catalyst phase, via a second dip tube and is fed continuously from a HPLC pump (10) (the liquid reactant can also be dissolved in the supercritical phase prior to its introduction into the autoclave with only minor modifications to the reactor) . This second dip tube is also used to vent the ionic liquid from the reactor under pressure before shut down. The ionic liquid and catalyst are container within a hastelloy autoclave (ca 50 ml) which is fitted with a magnetically driven stirrer and an internal thermocouple for feedback to a temperature control unit.
The gas stream leaves the reactor via a third port and is decompressed in two stages using pressure regulators (11 & 12) . The first expansion valve (11) is used to decompress the gas stream to pressures typically in the range 2-10 Bar and is heated to avoid freezing. The second expansion valve (12) reduces the gas stream to near atmospheric pressure and a third micro-metering valve (13) governs the accurate flow of gas through the system. The gas stream leaving the first expansion valve is water-cooled up to the point of the second expansion valve (12) where the gas stream then passes through a heat exchanger (14) . Collection vessels (15) are situated after the head
exchanger and are cooled in a refrigerated recirculating bath (16) , which also provides the cooling for the heat exchange coil. The gas stream then passes through a second collection vessel (17) , the micro-metering valve and finally a flow meter (18) .
With the set-up described above, both liquid and gas reactants are metered into the system at a constant rate whilst the flow rate of C02 is controlled downstream. It is also possible, however, to control the flow rate of C02 into the system upstream and regulate pressure downstream as previously discussed. We have the capability of controlling system pressure and C0 flow rates by each of these methods.
During each catalytic run the contents of the collection vessels are generally analysed by GC- MS/GC-FID, NMR and AA. However, it is also possible to measure the composition of the gas stream leaving the reaction by on-line FTIR (19) .
The gas stream is currently not recycled in the methodology described above, although it is possible to completely recycle the gas stream through liquification and recompression.
The reactor also contains an additional liquid injection loop (20) for the purpose of cleaning the reactor at the decompression stage.
A pressure transducer (P) , which measures system pressure, is linked through a trip switch unit .that provides the power supply to the temperature control unit, dosimeter and HPLC pump. When a pre- set pressure limit is exceeded, the power to these units is cut enabling the system to be operated safely in the absence of an operator.
The collection vessels were originally cooled in acetone/card ice baths. However, the trapping of residual alkene proved to be difficult with only 50% of the volume of injected liquid being collected. The trapping of liquid, following decompression of the gas stream, has proven to be the major problem associated with the reactor. We have since introduced the heat exchange coil to improve the efficiency of liquid collection. In principle, the liquid mixture being extracted from the reactor could be fractionated by sequential pressure drops. This would require multiple decompression stages with collection vessels situated after each pressure drop. The reactor can easily be modified to incorporate this additional separation process.
The invention will now be illustrated by examples (which do not restrict the scope of the invention) and drawings, in which:
Fig. 1 is a schematic representation of a generic supercritical fluid extraction unit.
Fig. 2 is a schematic representation of a reactor device to be used in a preferred embodiment of the invention to carry out the process in a continuous flow mode.
Fig. 3 is a schematic representation of the high- pressure, continuous flow reactor to be used in a preferred embodiment of the invention to carry out the process in a continuous flow mode.
Fig. 4 is a graph showing the change of catalyst turnover against time for continuous flow hydroformylation of l-Octene using a SCF-1L biphasic system as described in Example 38.
Example 1: Hydroformylation of 1-hexene
A 50 cm3 hastelloy autoclave fitted with a mechanical stirrer, thermocouple and pressure sensor was degassed with C0/H2 and charged with: - 4.0 cm3 of N-methyl-N' -butyl imidazolium hexafluorophosphate (ionic liquid) ; - 2 cm3 of 1-hexene (reactant) ; - 0.022 mmol of [Rh2(0Ac) ] and 0.64 mrrtl of P(0Ph)3 (catalyst).
The autoclave was then sealed and pressurised with C0/H2(1:1, 70 bar). 18 cm3 of liquid C02 was pumped in using a hplc pump. The autoclave was sealed and heated to 100°C until the total pressure reached 230 bar. The autoclave contents were then stirred
at 100°C for 1 hour, cooled and depressurised. The organic phase was separated from the ionic liquid using a syringe and analysed by gas chromatography (GC) .
40% conversion of the 1-hexene had occurred with 83.5% selectivity to aldehydes (linear :branch or 1 :b ratio=6.1) .
Example 2 (Comparative Example) : Hydroformylation of 1-hexene without the use of supercritical fluid.
The reaction set forth in Example 1 was repeated but the supercritical fluid (C02) was omitted. A conversion rate of 1-hexene of over 99% was observed but with only 15.7% selectivity to aldehydes (l:b=2.4), the remaining products being aldol condensation products . By comparison the selectivity obtained in Example 1 is more than 6- fold superior to that obtained when no supercritical fluid is used.
Example 3 (Comparative Example) : Hydroformylation of 1-hexene using toluene in place of the supercritical fluid.
The reaction set forth in Example 1 was repeated but the supercritical fluid (C02) was replaced with toluene (18 cm3) . A conversion of 1-hexene of over 99% was observed with a selectivity of 83.9% to the aldehyde (l:b=2.45). By comparison, the linear to branch ratio is less than half that of Example 1.
Examples 4 to 8 :
A 50 cm3 hastelloy autoclave fitted with a mechanical stirrer, thermocouple and pressure sensor was degassed with CO/H and charged with : 4.0 cm3 of N-methyl- N' -butyl imidazolium hexafluorophosphate (ionic liquid) ; - 2 cm3 of 1-hexene (reactant) ; and - 0.039 mmol of [Rh2(OAc) ] and 0.26 mmol of P(m-C6H4S03Na)3 (catalyst).
The autoclave was then sealed and pressurised with C0/H2 (1:1, 40 bar). Liquid C02 (18 cm3) was pumped in using an hplc pump. The autoclave was sealed and heated to 100°C, until the total pressure was 200 bar. The autoclave was stirred at 100°C for two hours and the stir was stoppped.
A stream of supercritical C02 (scC02) at a temperature of 100°C and a pressure of 200 bar was passed into the bottom of the reactor using a dip tube. At the reactor outlet the supercritical phase containing the products was transferred through a pressure control valve into a second autoclave at 10 bar and -50°C.
Once all the product had been transferred from the reactor, the transfer pipe was closed and the supply of scC02 terminated.
The reactor autoclave, still containing the ionic liquid and the catalyst was cooled and depressurised before being charged with more 1- hexene (2 cm3), C0/H2 (1:1, 40 bar) and C02 (18 cm3) and heated. This procedure was repeated so that five experiments (Examples 4 to 8) could be carried out using the same charge of catalyst and ionic liquid. The main reactor containing the products was depressurised and opened after each run. The products were analysed by GC and the second autoclave cleaned before use for subsequent runs.
The results are summarised in Table 1.
Table 1 - Product analysis for Examples 4-8
Examples 9-11
The reactions described in Experiments 4-8 were repeated but using 0.22 mmol of [Rh2(OAc)4_, 0.7 mmol of P(OPh)3 in place of P (rn-C6H S03Na) 3 , and 1- nonene (2 cm3) in place of 1-hexene. The reaction time was 1 hour. The results are summarised in Table 2.
Table 2 - Product analysis for Examples 9-11
Examples 12-15
The reactions described in Experiments 9-11 were repeated but using higher grade (low water content) C02. The total pressure in these reactions was 180-184 bar. The results are summarised in Table 3.
Table 3 - Product analysis for Examples 12-15
a Additional P(0Ph)3 (0.7 mmol) was added at the start of this run.
Examples 16-20
The reactions described in Experiments 12-15 were repeated but using N-methyl-N1butyl-2-methyl imidazolium hexafluorophosphate. The total pressure in these reactions was 173-190 bar. The results are summarised in Table 4.
Table 4 - Product analysis for Examples 16-19
Examples 21-24
The reactions described in Experiments 9-11 were repeated but using 0.8 mmol of P (0- -C H4C9Hιg) 3 in place of P(0Ph)3. The results are summarised in
Table 5.
Table 5 - Product analysis for Examples 21-24
Examples 25-36
The reactions described in Experiments 12-15 were repeated but using [Ph2P (.m-C6H4S03) ] [N-butyl, N- methylimidazolium] (1.5 mmol) in place of P(0Ph)3.
The total pressure in these reactions was 170-186 bar. The results are summarised in Table 6.
Table 6 - Product analysis for Examples 25-36
Example 37
The apparatus shown in Figure 2 was constructed and the autoclave charged with :
6 cm3 of N-methyl-N' -butyl imidazolium hexafluorophosphate (ionic liquid) ; - 0.020 g of [Rh2(0Ac)4] and 0.25 g of P(0Ph)3 (catalyst) and the stirrer started.
Liquid reactant 1-hexene was fed at 0.03 cm3 min-1. Permanent gas C0/H2 was fed at 6.45 nL min-1 (C0:1- hexene = 5:1) from the compressor and C02 was fed
through the pump and pressure controller at an overall pressure at 200 bar. The autoclave temperature was set at 100°C. The overall exit flow was less than 15 nL min-1. Samples were collected over hourly periods for a five hour period and after a further two hour purge . They were analysed by GC and the results are shown in
Table 7. In all cases, the collected fractions were colourless, confirming that the rhodium catalyst had not leached. GC analysis did not show the presence of phenol from hydrolysis of the ligand.
Table 7 - Products from the continuous flow hydroformylation of 1-hexene
There was no sign of isomerised alkene, nor of aldol condensation products .
Example 38 i Continuous Flow Hydroformylation of 1- Octene using a SCF-IL Biphasic System
The operating conditions and results of a catalytic run are outlined below:
Phosphine : [tppds] [PMI] 2 (1.88g, 2.8mmol) Ionic liquid : [BMIM] PF6 (4ml) Rhodium precursor : Rh2(OAc) (40mg, 0. lδmmol) System pressure : 200 Bar System temperature : 1002C Stirrer speed : 1800 rpm 1-Octene flow rate : 0.03 mLmin"1 (0.19mmol min-1) Permanent gas : 1:1 CO:H mixture Permanent gas flow rate : 6.62 mmol min-1 Expansion valve temperature : 35aC Pressure at 2nd expansion valve : ca 5 Bar C0 flow rate : 0.5 nLrnin"1 Total volume injected : 130.5 ml Total volume collected : 63.09 ml
The apparatus shown in Figure 3 was constructed and the autoclave charged with a solution of ionic phosphine ligand ([tppds] [PMI]2) and a rhodium precursor (Rh2(OAc)4) dissolved in an ionic liquid ( [BMIM] PF6) . The reactor was then purged with a low pressure of C02 to remove air from the system. 40 Bar of permanent gas was introduced into the autoclave. The autoclave was heated to the system temperature of 1002C and its contents rapidly stirred for around 1 hour to enable foundation of the catalyst. The pressure was then increased to operating pressure by the addition of C02. The pressure of the permanent gas supplied to the dosimeter is set to a level higher than the system pressure, namely 350 Bar. This "overpressure" was used to calculate the dosimeter switching rate for
a given mass of synthesis gas onto the system. Permanent gas and C0 were then flowed through the system at the pre-determined rates given above for several millimetres, prior to the addition of 1- octene which is also fed continuously and at a constant rate of 0.03 mLmin"1 (0.19 mmol min-1). The system pressure was reduced to 5-10 Bar by the first expansion valve and to atmospheric pressure by the second. The system was allowed to run continuously with reactants and products being transported into and from the stationary ionic phase by the mobile C02 phase. Liquid extracted from the reactor is trapped in the collection vessels and its composition is analysed sequentially by NMR, AA and GC .
Collection vessels were cooled in dry ice/acetone baths and the system was operated continuously for three days with a total of 12 fractions collected and analysed during this period. The percentages of aldehydes were determined on the basis that all unrecovered material is octene or isomerised alkenes and therefore represent minimum possible conversion. Conversions were determined on the basis that rz-octane is present in the starting material at the same concentration observed in recovered fractions (as determined by GC analysis) . The results of this run are tabulated below (see Table 8) and represented graphically (see Figure 4) as catalyst turnover number against time. The linearity of the graph shown in Figure 4 over the 72.5 hour period shows that the catalyst is stable
at least over this period of time. The l:b ratio of the product aldehydes is constant (3.8) throughout the run, showing the ligand oxidation is not occurring. Rhodium analysis (by atomic absorption) of the recovered products shows <lppm of Rh in any of the samples which amounts to less than 0.06% of the initial rhodium loading. GCMS and NMR analysis of liquid recovered from the reactor show the presence of only the two product aldehydes and unreacted reactant. Analysis therefore shows that the continuous flow supercritical fluid-ionic liquid biphasic system provides a method for continuous flow homogeneous catalysis with built in separation of the products from both the catalyst and the reaction solvent even for relatively involatile products. With further modifications to the current reactor it will also be possible to (i) fractionate product aldehydes and unreacted starting material (ii) recycle the gas stream leaving the reactor. These modi ications will provide a process in which all materials are recycled with facile separation of products from catalyst, unreacted reactant and reaction solvent.
These experiments show that the process of the invention allows good conversion together with a high selectivity of the catalytic reaction and that the process can be operated in continuous flow mode .
Table 8 - Product Analysis for Example 3i
-^1