A PROCESS FOR PREPARING AN UNREFINED EXTRACT OF VEGETABLE OIL OR ANIMAL FAT

Hie present invention relates to a process for preparing an unrefined extract of vegetable oil or animal fat and to a method for preparing a refined extract.

As a response to tax incentives and legislative measures, biodiesel consumption in the UK, mainland Europe and the USA is rising. The major legislative measure in Europe is EU directive 2003/30/EC which directs States to set indicative targets for biofuel production in 2005 and 2010. It recommends that transport fuels comprise 2% biofuel (in practice either biodiesel or bioethanol) by 2005 and 5.75% by 2010, Although these percentages are low, their implementation will represent a huge increase in biodiesel consumption in Europe.

Biodiesel is a sustainable alternative to crude-oil derived diesel ("petrodiescl") and may be used in a pure form in newer engines without damage or be combined with standard diesel in ratios varying from 2 to 20% biodiesel. Rapeseed oil is used in the manufacture of biodiesel for powering motor vehicles. Worldwide production of rapeseed oil (including carxola) rose to 46.4 million metric tons in 2005 which is the highest recorded total (source; FAO), Rapeseed oil is the preferred oil stock for biodiesel production in most of Europe, partly because rapeseed produces more oil per unit of land area than other oil sources such as say beans. Due to this increased demand, current prices of rapeseed oil are very high. Due to the cost of current processes for crushing and refining, rapeseed-derived biodiesel costs more to produce than standard diesel.

In Asia and Africa especially,, Jatropha is seen as the most promising source of feedstock for biodiesel production, Jatropha can produce around 1,590 litres per hectare compared with 1000 litres per hectare for rapeseed,

Biodiesel is currently produced predominantly by reacting vegetable oils with methanol in the presence of an alkaline liquid catalyst (usually sodium or potassium methoxide) to produce fatty acid methyl esters (biodiesel) and glycerol (by-product) in a reactor (such as a stirred tank). The reaction is a transestexificatkαi reaction of triglycerides which axe the main constituent of vegetable oil with, methanol. The process of biodiesel production usually involves two separate serial steps. In the first Step, oil is extracted from oilseed by crushing and solvent extraction. In the second step, oil is reacted with methanol to form biodiesel. Typically the process takes place in three separate locations; oilseed growth (farm), crushing and solvent extraction (vegetable oil production facility) and conversion of oil to biodiesel (biodiesel plant).

A disadvantage of this process is the formation of soaps from the reaction of the alkaline catalyst with free fatty acids or the saponification of the triglycerides and biodiesel. These reactions consume the catalyst and hinder phase separation of biodiesel from glycerol and the subsequent yield is reduced. The glycerol contains salts from neutralisation of the catalyst and requires a more costly step to reach high grade purity.

Almost all biodiesel plants worldwide use homogeneous catalysts. The disadvantages of using such catalysts are that they are continuously consumed and require a certain amount of effort to remove.

The Estørfip-H process produces 160000 te/year of biodiesel using a heterogeneous spinel-based catalyst but operates at higher temperatures (230 ^β C) than conventional processes (US-B-6878837). The savings from the reduction in downstream separation steps must be more than equal to the extra cast of the higher temperature process. Hie promise of a cheaper pxoee$s has resulted in a large body of research into heterogeneous catalysts including: (Peterson, G- and Scairah, W. Rapeseed oil transesterification by heterogeneous catalysts. Journal of the American Oil Chemists' Society 61(1984) 1593-1597; Xiβ, W., Peng, H. and Chen, L. Transestertfication of soybean oil catalyzed by potassium haded on alumina as a solid base catalyst. Applied Catalysis A: General 300 (200$) 67-74; Watkins, R., Ue, A. and Wilson, K. Li-CaO catalysed triglyceride transesterification for biodieseϊ applications. Green Chemistry 6 (2004) 33S-340; and Suppes, G. J., Daaari, M.A., Doskocil, EJ., Mfunkidy, PJ. and Goff MJ. Transestέrification of soybean ail with zeolite and metal catalysts. Applied Catalysis A: General 257 (2004) 213-223). Various catalysts have been evaluated with varying degrees of success. Solid bases have generally been observed to be more active than metal compounds (Suppes [supra]). Solid acids tend to require more extreme reaction conditions than solid bases (Lotero, E., Liu, Y,, Lopez, D-j. Suwannkarn, K., Bruce, D. and Goodwin, J. Synthesis of biodiesel via acid catalysis. Industrial and Engineering Chemistry _^ Research 44 (2005) 5353-5365). Solid bases are the most well-studied class of catalysts. Most studies have concluded that activity increased with base strength and pore size.

In order to produce biodiesel in bulk, an industrially practical catalyst must be easy to prepare on a large scale. Alkali metal-doped metal oxides have shown promising activity for the transestβπfication of vegetable oils to biodiesel. Li doped onco CaO has been shown to give greater than 90% conversion after 10 minutes for the traosesterification of tributyrate (a short chain triglyceride used as a model substrate for vegetable oil (Watkins [supra]). Alkali metal salts loaded on alumina also catalysed the txansesterification of vegetable oil with methanol (Xie [supra); Ebiura, T., Echiiea, T., Ishikiwa, A,, Murai, K. and Tosbide, B. Selective transesterification of triolein with methanol to methyl oleate and glycerol using alumina loaded with alkali metal salt as a solid-base catalyst. Applied Catalysis A: General 283 (2005) 111-116) end alkali metal salts on magnesium oxide showed some activity for the transesterification of vegetable oil (Peterson, [supra]). '

There are examples of the use of a solid heterogeneous catalyst in biodiesel production. However, the conditions for their υae are extreme and the catalysts have not been used in a reactive extraction process, "Die use of catalysts in high temperature processes (greater than 100 ⁰ C) is disclosed in WO2006050925 and US2004034244 which require pressurised systems (the ambient pressure boiling point of methanol is 65 *C). More specifically, WO2006050925 describes a process for producing esters of fatty acids and glycerin using heterogeneous catalysts, in particular for producing biodiesel. The process comprises: reacting vegetable oils or animal fats with an aliphatic 10 monoalcohol at high temperatures in the presence of a catalyst which comprises magnesium oxide or mixed oxides of magnesium and aluminum (obtained by calcination of hydrotalcite-like compounds which contain Al and Mg with an atomic ratio of Mg/Al > 1) to form esters of fatty acids and glycerins separation of the υnreacted monoalcohol; and separation of the fatty acid ester* and of the glycerin. US2004034244 describes alkyi esters of fatty acids and high purity - ^•

glycerin produced by a process comprising a series of transestβtification reactions between a vegetable or animal oil and an aliphatic monoalcohol employing a heterogeneous catalyst (for example based on zinc aluminatc). The water content in the reaction medium is controlled to a value that is below a given limiting value.

Very few of the published studies describe catalysts which are industrially practical: either the conditions employed are extreme (Xie et al [supra] and Suppβs et al [supra]), the feedstock is not vegetable oil (Watkins et al [supra]) or the reaction is too slow (Kim et al, Transesterification of vegetable oil to biodiesel using heterogeneous base catalyst. Catalysis Today 93-95 (2004) 315-320).

LiMgO has however been described for use in oxidative coupling of methane (Ito et aL Oxidative dimerization of methane over a lithium promoted magnesium oxide catalyst. J. Am. Chcm. Soc. 107 (1985) 5062-506S) and aldol condensation of citral with acetonύ (Diez et al, Aldol condenstaion of citral with acetone en MgO and alkali promoted MgO catalysts, journal of Catalysis, 240 (2006) 235-244).

Poulloux et al, "Solid base catalyst for the synthesis ofphotosterol esters'" 2003 describe the use of LiMgO in a transesterification reaction from methyl ester and bulky alcohol (sitosterol) to methanol and long chained bulky ester (phytosterol). This is not a biodiesel reaction. The reaction was carried cut at 24O ⁰ C and LiMgO was found to be a lesser performing catalyst than others but favoured a side reaction.

Haas et αL "In situ Alkaline Transesterification: An Effective Method for the Production of Fatty Add Esters from Vegetable Oils", JAOCS, 8ϊ, no. 1, S3 - 89, (2004) report a laboratory-scale process for producing biodiesel in situ from soy using methanol but this has the disadvantage that it requires avast excess of methanol rendering the process uneconomical for commercial purposes,

Ozgul Yucel et a "Variables Affecting the Yields Qf Methyl esters Derived from in situ Esteriβcation of Rice Bran OtT, JAOCS, 79, no.6, (S 11 - 614 (2002) and "FA Manaalkylestersfram Rice Bran Oil by in situ Esterificatiori", JAOCS, 80, no. 1.81 - 84 (2003) report a process for extracting and reacting rice bran oil using methanol or efhanol with limited success.

Kondo et a "On-line Extraction-Reaction ofCanoϊa Oil with ethanol by immobilised Lipase in scCOT End. Eng, Chem. Res., 41, 5770-5774 (2002) report extraction and reaction in supercritical carbon dioxide using an immobilised enzyme. However, this process is difficult to scale up and is relatively slow. The capital Costs are high and die process is impractical and unlikely to be economically feasible for smaller scale units in particular.

The present invention is based on the recognition that a mixture of alcohols or of an alcohol and an organic solvent optionally together with a catalyst enables concurrent reactive-extraction of a useful unrefined extract from a natural source of vegetable oil or animal fat.

Thus viewed from a first aspect the present invention provides a process for preparing an unrefined extract of vegetable oil or animal fat comprising:

adding to a source of the vegetable oil or the animal fat a transesterifyiπg alcohol and an organic solvent, optionally together with a catalyst to produce ihe unrefined extract, wherein the unrefined extract comprises one or more monoalkyl esters of one Of more long chain fatty acids.

In a preferred, embodiment, the one or more moαoalkyl esters of one or more Hong chain fatty acids represent a major proportion of the unrefined extract The unrefined extract may constitute a biodiesel. The biodiesel may conform to a legislative definition of the term "biodiesel".

The unrefined extract may b«s refined into a refined extract by conventional steps. The refined extract may constitute a Modiesel. The biodiesel may conform to a legislative definition of the term "biodiesel".

The process of the invention integrates extraction and reaction into a single step and results in a reduction in the total number of unit operations and increased yield of a useful unrefined extract (eg biodiesel). For example, the process of the invention may be used in large-scale commercial or small-scale biodiesel production and -will permit the development of integrated technology for biodiesel production combining extraction and reaction in a one-step reactive extraction process. This will facilitate distributed production of biodiesel (eg by oilseed farmers) whose use of the technology would enable production on-site. Due to the high yield of useful unrefined extract (eg biodiesel) that may be achieved straightforwardly without extreme temperatures oz pressures, complex plants or costly reacτants, the process of the invention is economically viable on a large or small scale. A typical yield of biodiasel is at least as high as that achieved by conventional methods in which the extraction and reaction steps are separate.

The process of the invention has a number of additional benefits. Firstly farmers may move further along the value chain of their product Secondly a local market for biodiesel may be stimulated. Thirdly the economics of using currently agriculturally unproductive (eg "set-aside") land for fuel CPOpa may be improved. The process of the invention may also be particularly suitable for use in developing countries where it can compensate for poor supply of fossil fuel-derived transport fuels, particularly in remote communities. Another benefit is that the anticipated "distributed production" model has the significant environmental benefit that it decreases the number of tauter journeys to and from large centralised production facilities thereby decreasing the life-cycle COa emissions associated with the product

The unrefined extract may comprise one or more monoalkyl esters of one or more Ci i. ₂₄ fatty acids. The (or each) monoalkyl ester of the (or each) fatty acid may be independently of formula:

R ¹ O(OCR) wherein R is a long chain (eg Ca- ₂ 4-) alkyl and R ¹ is an optionally substituted C1-5- alkyl, particularly preferably a C|. ₃ -alkyl. Specifically preferred groups R ¹ are methyl and ethyl, especially preferably methyl.

The vegetable oil or animal fat may comprise one or more acylglycerols (glyeerideg), The vegetable oil or animal fat may comprise one Oϊ more (typically all) of the group selected from monσglycerides, diglycerides and triglycerides. The predominant acylglycerols are generally triglycerides. The monoglycerides, diglycerides and triglycerides may be long chain (eg Cπ- ₂ 4) fatty acid esters of glycerol.

The (or each) triglyceride may be independently of formula: RCOO-CH ₂ CH(-OOCR ^I )CH ₂ -OOCS." wherein each of R, R ¹ , and R" which may ^" be the same or different ia a long chain (eg CtI-M-) alkyl.

The process of the invention may achieve a conversion to the one or more monoalkyl esters of one or more long chain fatty acids of S0wt% or more of the total amount of acylglycerols. Ih a preferred embodiment, the conversion of the total amount of acylglycerols is in excess of 9Qwt%, preferably in excess of 95wt%, particularly preferably in excess of 9Swt%, more preferably in excess of 99wt ^p /o and may be about 100wt%.

The process of the invention may achieve a conversion to the one or more monoalkyl esters of one or more long chain fatly acids OfSU ¹ Ko or more of the total amount of triglycerides. In a preferred embodiment, the conversion of the total amount of triglycerides is in excess of 90%, preferably in excess of 95%, particularly preferably in excess of 98%, more preferably in excess of 99.5% and may be about 100%.

Preferably the ratio of transesteπfying alcohol to organic solvent is in the range 10:90 to 90;10, particularly preferably 30:70 to 70:30, especially preferably 40:60 to 60:40, more especially preferably 45:55 to 55:45, yet more especially preferably about 50:50.

The Transesterifying alcohol and organic solvent (optionally together with the catalyst) may be added to the source of vegetable oil or animal fat separately and concurrently or in a mixture. Preferably the transesterifying alcohol and organic solvent are ittiscible.

The transesterifying alcohol may be an acyclic (eg linear or branched), mono- or polycystic (eg aromatic), saturated or unsaturated alcohol. The transesterifying alcohol Ia preferably an optionally substituted d^-alcohol, particularly preferably a Ci_3-alcohol. Specifically preferred are methanol and ethanol, especially preferably methanol.

In a first preferred embodiment, the organic solvent is an alcohol solvent The alcohol solvent may be an acyclic (eg linear or branched), mono- or polycyclic (eg aromatic), saturated or unsaturated alcohol. The alcohol solvent is preferably an optionally substituted C^-alcohoL particularly preferably a C ₂ _«-alcohol, more preferably a C ₂₄ - alcohol. Specifically preferred are apropanol (eg n-propanol) or ethanol, especially preferably ethanol (preferably bioethanol).

In a particularly preferred embodiment, the transesterifying alcohol is methanol and the organic solvent is ethanol. Particularly preferably, the process comprises: adding to the source of vegetable oil or animal fat methanol and ethaaol, together with a catalyst

Ethanol and methanol give a high yield of useful unrefined extract (eg biodiesel). The yield is typically higher than (for example) is obtained by an alcohol/hexane mixture, The absence of a non-polar solvent (eg hexane) makes solvent recovery straightforward and high, temperatures and/or pressures are not required. Moreover the use of bϊoethanol is environmentally preferable to the use of a non-polar solvent such as hexane.

Preferably the ratio of methanol to ethanol is in the range 40:60 to 60:40, particularly preferably 45:55 to 55:45, especially preferably about 50:50.

In a second preferred embodiment, tihe organic solvent is a non-polar solvent. A preferred organic solvent is & Ci.i2-a!kane _s preferably a Ci-s-alkane. Specifically preferred is hexane.

In a particularly preferred embodiment, the transesterifying alcohol is methanol and the organic solvent is hexarte. Preferably the volume of methanol is in excess of 50% of the total volume of methanol and hexane.

In a particularly preferred embodiment,, the transesterifying alcohol is ethanol and the organic solvent is hexane. Preferably the volume of efb.an.ol is in excess of 10% of the total volume of ethanol and hexane.

Ih a preferred embodiment, the process comprises: adding to the source of vegetable oil or animal firt a transesterifying alcohol and an organic solvent, together with a catalyst

In a preferred embodiment, the process comprises; adding to the source of vegetable oil or animal fat the transesterifying alcohol and the organic solvent, together with a first portion of a catalyst to form an intermediate unrefined extract; and adding to the intermediate unrefined extract a second portion of a catalyst to produce the unrefined extract

Adding the catalyst portion-wise in two stages in accordance with this embodiment has been found advantageously to increase the conversion of the total amount of aeylglyeerols and/or of the triglycerides to the one or more monoalkyl esters of one or more long chain fatty acids. The second portion of catalyst may be added to the intermediate unrefined extract ten or more, preferably twenty or more minutes after the first portion of a catalyst has been added to the Source of vegetable oil or animal føt

The catalyst may be a liquid or solid catalyst. A solid catalyst exhibits advantages in terms of recovering the liquid extract

The catalyst may be dissolved, suspended or dispersed in the traπsesterifying aJcohol and/or the organic solvent. The catalyst may be an alkali metal or alkaline earth metal λiethoxidc or hydroxide such as calcium methtøride, sodium methoxide of sodium hydroxide, For example, the catalyst may be sodium hydroxide in solution in the transesterifyiπg alcohol and/or the organic solvent (eg at a molality in the range 0.05 to OJ).

Preferably the catalyst is heterogeneous. The use of a heterogeneous catalyst advantageously leaves no neutralisation salts in the glycerol and as the catalyst is not continuously added aftd disposed of, the levels of input and waste are reduced. The catalyst may also be retained in the reactor by simple filtration and does not need to be neutralised to quench the reaction so that the number of separation steps downstream may be mk-iraised.

Preferably the catalyst is a solid heterogeneous catalyst.

Preferably the catalyst has a base strength (pK _B H ₊ ) in excess of 8 _» particularly preferably in excess of \ O ₅ more preferably has a base strength (pKam) in the range U to IS.

Particularly preferably the catalyst is a metal oxide doped with an alkali metal. The alkali metal is preferably selected from the group consisting of Li, Na and K. Typically the alkali metal is doped on the metal oxide in an amount of 1 Owt% or less (typically about 5wt%).

The metal oxide is preferably an oxide of one or mote of the gjoup consisting of alkaline earth metals, group HIA metals, transition metals, post-transition metals and Janthanides. Preferred are oxides of alkaline earth metals and group UIA metals, particularly preferably oxides of Mg ₁ Ca or Al, especially preferably MgO, CaO ^' and AIJOJ, most preferably MgO and CaO.

The metal oxide may be subjected to an elevated temperture. In a preferred embodiment, the metal oxide is calcined.

Specifically preferred alkali metal doped-metal oxides are Li-CaO, Na-CaO, K-CaO and calcined Li-MgO ₃ especially preferably calcined Li-MgO. Calcined Li-MgO exhibits a rapid rate of transesterification (at the boiling point of methanol) and hag a low solubility rate In methanol.

The vegetable oil may he virgin oil feedstock or waste vegetable oil. The vegetable oil may be an oil of any of the group consisting of rapcseed, soybean, field peπnyciess, Jatroptøa, mustard, flax, sunflower, caαola, palm, hemp or algae _*

The animal fats may be tallow, lard, yellow grease, chicken iat or by-products of the production of Oraega-3 fatty acids from fish oil.

Preferably Uie oil is rapeseβd oil, soy or Jaftopha, particularly preferably rapeseβd oil,

The source of vegetable oil is preferably oilseed. Typically the integrity αf the skin of the oilseed is broken before use in the process of the invention in order to expose the

ύil-containing inner layers. The source of vegetable oil used in accordance with the process of the invention may be crushed, cracked, flaked or ground oilseed.

The process of the invention may further comprise : crushing, cracking, flaking or grinding oilseed.

IQ a typical embodiment an oilseed, such as lapeseed, soy or Jatropha is cracked or flaked using conventional cracking or flaking equipment to expose the inner surface area of the seed.

The transesterifying alcohol and organic solvent may be added to the oilseed at ambient or elevated temperature and/or pressure. The oilseed may be left exposed to the transesterifying alcohol and organic solvent for a specified period of time (eg one hour or more). The oilseed may be left exposed to the tørøesterifyiπg alcohol and organic solvent at an elevated temperature. The elevated tempβrture is generally Ie$$ than the boiling point of the organic solvent and/or the transesterifying alcohol. The oilseed may be left exposed to the transesterifying alcohol and organic solvent under agitation. For example, oilseed may be left exposed to the transesterifying alcohol and organic solvent in a stirred* heated vessel until reactive extraction is complete.

The process of the invention may be carried out in a t eaetor such as (for example) a stirred tank, oscillatory baffled reactor or decanter-extractor. The liquid part of the unrefined ^β xtr&ct may be isolated from the solid part (residue). This step may be carried out downstream by a conventional apparatus such as a decanter ox by filtration. The solid residue may constitute meal which may be used for animal feed or energy production. The liquid part may be removed from the reactor.

In a preferred embodhnent _} fte process further comprises: isolating a liquid part of the unrefined extract from a solid part

The unrefined extract may contain amounts (eg trace amounts) of soap, glycerol, fatty acids, monoglycerides, diglycerides and triglycerides. Typically the amount of monoglycsridφ, diglyceride and triglyceride is less than 5wt%, preferably less than 3wt%, more preferably less than lwt%,

The process of the invention may further comprise downstream processing steps. The or each processing step may be a conventional downstream processing step. Each of these steps contributes to refining the unrefined extract.

Viewed from a further aspect the present invention provides a method for preparing a refined extract comprising: preparing an. unrefined extract by a process as hereiribefote defined; and refining the unrefined extract to produce a refined extract.

The refined extract may constitute a biodiesel. The biodiesel may confortn to a legislative definition of the term "biodiesel".

For example, the refining step may comprise a washing step (eg with water). For example, the refining step may comprise a separating step (eg- using a conventional separating apparatus Such as a centrifuge or wash column) for example to

remove methanol, soap or catalyst For example, the refining step may comprise a drying step (eg using a drying bed). For example, the refining step may comprise an ion exchange step (eg using an ion exchange bed).

Typically the unrefined extract comprises glycerol. The unrefined extract may comprise a minor proportion of glycerol. In an embodiment of the method of the invention, the step of refining Hie unrefined extract may comprise: recovering or eliminating glycerol from the unrefined extract.

The recovery or elimination, of glycerol may be carried out downstream by conventional apparatus such as a settler. Glycerol is widely used commercially in the cosmetics industry or for energy production.

The unrefined extract may comprise a proportion of a. biolubricant. In an embodiment of the method of the invention, the step of refining the unrefined extract may comprise: recovering or eliminating biolubricaπt from the extract

The recovery or elimination of biolubπcant may be carried out downstream by conventional apparatus.

The extract may comprise a proportion of vitamin E. ϊn an embodiment of the method Of the invention, the step of refining the unrefined extract may comprise: recovering or eliminating vitamin E from the extract.

The recovery or elimination, of vitamin B may be carried out downstream by conventional apparatus.

The present invention will now be described in a non-limitative sense with reference to the accompanying Figures in which:

Figure 1 - Schematic illustration of a reactor for carrying out an embodiment of the process or method of the invention;

Figure 2 - Comparison of hexane and n-alcohols as solvents for extracting t apeseed oil;

Figure 3 - Thin layer chromatographs of an extract produced using Soxhlet apparatus; Figure 4 - Yield of biodiesel produced by an embodiment of the process of the invention as a function of percentage alcohol in hexane; Figure S - Mass of biodiesel produced by an embodiment of the process of the invention;

Figwe 6 - Thin layer chxomatographs of an extract produced from rapeseed after contact with a solvent + NaOH in accordance with an embodiment of the process of the invention;

Figure 7 - Comparison of the amount of extract produced from, rapeseed with and without a reaction according to an embodiment Of the process of the invention; Figure 8 - A conventional flowsheet for biodiesel production is shown on the left and a modified flowsheet for a heterøgeneously catalysed process is shown on the right; Figure 9 - CaO catalysed reaction^ 60*C, 6:1 molar ratio oil to methanol, 2g catalyst Triglycerides a, diglycerides «, monoglycerides □• (Error shown due to sampling errors, error in analysis too small to show);

Fϊguxe 10 - Li-CaO catalysed reaction, 60 ⁰ C ₅ 6:1 molar ratio oil to methanol, 2g catalyst. Triglycerides a, diglycerides •, monoglycerides □;

Figure 11 - Na-CaO catalysed reaction, 60'C, 6:1 molar ratio oil to methanol, 2g catalyst Triglycerides ■* diglycerides •, monoglycerides α;

Figure 12 - K-CaO catalysed reaction, 60 ^D C, 6:1 molar ratio oil to methanol, 2g catalyst. Triglycerides ■, diglycerides •, monoglyoerides □;

Figure 13 - K-CaO calcined catalysed τeaction, 60 ⁰ C, 6:1 molar ratio oil to methanol,

2 g catalyst. Triglycerides ■, diglycerides •, monoglycerides πj

Figore 14 - Li-MgO catalysed reaction, 60"C, 6:1 molar ratio oil to methanol, 2g catalyst. Triglycerides ■, diglyύβrides •, monoglyceiides □;

Figure IS - A schematic illustration of the transestarifieation of triglycerides to biodiasel; and

Figure 16 - K-CaO (calcined) catalysed two-stage reaction; fresh catalyst added after

30 minutes, 6O ⁰ C, 6:1 molar ratio oil to methanol, 2g catalyst Triglycerides ■, diglycerides *, monoglycerides □ (one-stage process solid lines, two-stage process dotted lines),

Example 1 - Solvents

Experimentation

A first experiment assessed the effectiveness of hexaάe, methanol, ethanol and propanol as solvents in extracting oil from rapeseed. A second experiment in accordance with an embodiment of the process of the invention measured the amount of bϊodiesel produced with mixtures of certain solvents, reactants and a catalyst (sodium hydroxide).

(1) Extractions using hexane, methanol, ethanol, propanol and a 50-50 methanol- ethanol mixture as solvents were performed using Soxhlet apparatus. In this apparatus, solvent vapour is generated by evaporation from a pool containing the extract and solvent. Hie solvent vapour condenses and fills a sample compartment containing the ground oil seeds in a paper cup. When the solvent just covers the seeds, it overflows creating a siphon that draws all ftø solvent and oil out of the sample compartment. In this way* the seeds are continuously extracted using pure solvent

25 g of seeds were used in each test. Hexane was tested because it is used as the solvent in commercial vegetable oil extraction processes. 200 cm ³ of the solvent were used for each extraction in the Soxhlet apparatus. Experiments lasted for sfot hours to ensure that extraction was complete.

The extract was transferred from the Soxhlet apparatus to a rotary evaporator where the solvent was removed under vacuum at temperatures in the range 60-90'C (depending on the solvent under teat), After evaporation the residue was weighed and a small sample placed on a thiα layer chromatography slide.

Thin layer chromatography (TLC) was used to analyse the residue and provide a quick and simple analysis of the biodiesel, mono-, di- and triglycerides. The TLC plates were 0,25mm silica gel. The mobile phase used on the TLC plates was 85 vol% petroleum ether, 13.5 vol% diethyl ether and 1.S vol% acetic acid. The TLC plates were prepared by placing them in a TLC chamber with 20 ml of the mobile phase and allowing them to be wetted entirely then dried for 5 minutes. 1 ml of extract was diluted with 5 ml of petroleum ether and a 10 μl sample taken from the diluted mixture and applied to the plate 1 cm from the base. The TLC plate was then put back

in the chamber and the mobile phase allowed to travel the length of the plate. After drying, the plates wej-e developed in a second chamber containing iodine crystals. Typical results are shown in Figure 3, It was possible to use the images produced from the TLC both qualitatively and quantitatively. lift general, the less polar a compound, the further it travelled up the plate, For example, in the hexane TLC, the spot at the start line represents monoglycerides, the diffuse spot just above the start line represents the diglycerides, the semicircular spot represents fee fatty acids and the large spot represents the triglycerides.

Each extraction was carried out three times to ensure reproducibility. Figure 2 shows the mass of extract pfcoduced as a result of extraction using the Soxhlet apparatus. The results have been normalised to the mass of extract produced using hexane. It can be clearly seen that as the length of the carbon chain of the alcohol increases, the amount of extract increases. It was observed that the colour of the extract ranged from green/yellow with hexane through to dark brown with methanol. This indicated flαat the composition of the extract was changing with the solvent. The TLC slides show that vegetable oil (triglycerides) was extracted when hexane, piopanol or ethanol was used as a solvent but very little triglyceride was extracted with methanol.

(2) The protocol for tfie reaction-extraction experiments according to the process of the invention was as follows. Alcohol-hydroxide solutions were prepared by dissolving 0.1 or 0.05 moles of sodium hydroxide in.1 kg of methanol, ethanol or propanol. 50 g of seeds were ground for 1 minute in an electric coffee grinder to ensure that the oil-bearing materia] inside the seed was exposed. The seeds were then transferred to a 250 ml flask and 200 ml of the solvent/NaOH mixture under test was poured into the flask. The flask was submerged in a water bath at 60'C and attached to a. shaker arm. λ water-cooled condenser was connected to the top of the flask to ensure that no solvent was lost by evaporation. The flask was shaken in the water bath for 60 minutes to allow the system to come tα both solid-liquid and reaction equilibrium. After 60 minutes the seeds and extract were separated using vacuum filtration. The extract was transferred to a rotary evaporator where the solvent was removed under vacuum at temperatures in the range 60-90" C depending on the solvent under test The mass of extract was recorded and a sample placed on a TLC slide to quantify the amount of biodiesel present (see Figure 6). This is possible because the size of a spot oα a TLC plate is proportional to the amount of the component in the sample. The method was calibrated using test mixtures that contained varying masses of biodiesel. mono-, di- and triglycerides. By measuring the relative sizes of spots for biodiesel and tri-glycerides, it was possible to determine the mass of biodiesel in a sample.

Figure 7 compares the mass of material extracted rron. the seeds by solvent extraction (k in the first experiment) with the mass of material extracted by the reaction- extracttøn in accordance with the process of the invention (/e in the second experiment). ϊπ, each case, the results have been normalised to the mass of extract produced using hexane. From Figure 7, it can be seen that in all cases the mass of material extracted increased in the presence of the catalyst. This indicates that trafljSesterifLcation consumes triglyceride as it is extracted. The greatest enhancement was observed for the ethaπol-methanol system.

Further experiments were carried out in accordance with an embodiment of the process of the invention using various proportions of methanol in hexane. The results are shown in Figure 4 where the yield of biodiesel is plotted as a function of the

percentage alcohol (containing sodium hydroxide) in hexane. When methanol is used, biodiesel is produced when the solvent mixture i$ more than 50% methanol. The amount produced increases rapidly to a maximum for 0.1 mαlal NaOH in methanol solutions and less rapidly for 0.05 molal solutions. The use of stronger methanol and sodium hydroxide solutions caused, saponification due to the presence of water formed when the sodium hydroxide dissolved in the alcohol. This may be overcome if necessaiy by using pure sodium methoxide as the catalyst With ethatiol, biodiesel was produced when the hexane and ethanol mixture contained more than 10% ethanol.

Further experiments were carried out in accordance with m. embodiment of the process of the invention using a 0.1 molal NaOH solution of 50/50 metbanol-ethanoi (that acted as both a solvent and. a reactant) and of methanol and ethanol (with and without hexane) and propanoL The mass of biodiesel produced by these extraction- reactions are shown in Figure S. Figure S shows that for methanol, ethanol and propanol, the yield of biodiesel decreases as the length, of the carbon chain of the alcohol increases. The mixture of ethanol and methanol was interesting in that the yield of biodiesel was larger than for either methanol or ethanol alone oτ any other variant (even those including hexane).

E-caπrole 2 - Catalyst

A series of alkali metal-doped metal oxides were prepared to identify catalysts øctivc in the transesterification of rapeseed oil to biodiesel. The effect of the alkali, rαetal dopant (Li, Na and K), the support (CaO, MgO and Al ₂ O ₃ ) and preparation method (calcined at 6Q0 ^β C or heated to 100*C) on catalytic activity was investigated. Once active catalysts 'were identified, other reaction parameters (molar excess of methanol and amount of catalyst) were varied in order to define the reaction conditions necessary for a hetetogeneously catalysed reaction to meet the European standards (ENl 4214) for a biodiesel to be sold as a fuel.

Figure 8 shows a conventional flowsheet for biodiesel production (on the left) and a modified flowsheet for a heterogeneously catalysed process (on the right).

Materials and Methods

(a) Catalyst preparation

Catalysts were prepared by the incipient wetness method. Rjeagent grade CaO (Fisher), MgO (VWR) and γ-AbOj (Merck) were used as supports and reagent grade LiNO ₃ , NaNOj and KNO ₃ (all Fisher) as the source of alkali metals. The alkali metal salts were dissolved in distilled water to give an alkali metal loading of S wt% on the support. The solution was mixed with the support and the resulting paste dried in an oven at 11O ⁰ C for 5 hours and then calcined at 60O ⁰ C for 5 hours.

(b) Characterisation

Catalyst surface area was measured using the BET isotherm method in a Micromeritics Pulse Chemisorb 2700. Samples were degassed at IQO ⁰ C under flowing argon and the adsorption of >fe at 77 K measured. The base strength was assessed

using Hammett iπdicatora dissolved in methanol following the method of Watkins [supra].

(c) Catalyst screening

Food grade rapeseβd oil (Henry Colbeck) was used as the source of triglyceride and reacted with reagent grade methanol (Fisher). As an initial test of activity, small scale reactions were carried out in parallel. 0.2 g of catalyst was added to a 6;1 molar ratio of methanol to oil (volumetrically 1:3.3) in a 2 ml sealed vial which was shaken in a IKA mini shaker at 600 ipm inside an incubator kept at 60 ⁰ C. After 3 hours, the reaction products were analysed by g&B chromatography using the method of Plank and Lorbeer (Flank, Cw, and Lorbeer, E. Simultaneous determination qf glycerol, and mono-, di- and tri- gfycerides in vegetable oil methyl esters by cøpftlary gas chromatography. Journal of Chromatography A 697 (1995) 461-468). 100 rag of the biodiesel sample was silylatβd with MSTFA and internal standards of butane- 1,2,4- ttfol and tricaprin added A Unicam Fro GC gas chromatograph was used with a J&W Scientific DB-IHT column (15m length, 0.32 mm internal diameter).

(d) Batch kinetics

50 ml of nαeihanol and 165 ml vegetable oil (6:1 molar ratio) were mixed with a magnetic stirrer at lOOrpm in a 400 ml sealed glass jacketed vessel at 6Q°C. A ~Sml sample was taken every 20 minutes, filtered and analysed by gas chromatography.

(e) Reusability

To determine whether any of the catalytic activity was attributable to homogeneous catalysis by the alkali metal leaching from the catalyst, the stage 1 screening was firstly repeated using the same catalyst seven times with no significant loss in activity. Additionally, the biodiesel produced was analysed for alkali metal content using flame photometry and atomic absorption spectroscopy.

Results and discussion (a) Initial Screening

The surface area of the CaO supported catalysts and MgO supported catalysts was low and less than the surface area of the Al ₂ Oa supported catalyst (see Table 1). Calcining the CaO and MgO based catalysts significantly reduced their surface area. However, the triglyceride molecule is large compared with the pore sizes so there is negligible internal diffusion of triglyceride into the catalyst particle and instead the reaction occurs at the catalyst particle surface. Increasing the internal surface area would therefore have negligible benefit.

Shown in Table 1 are the base strengths of the catalysts measured using Hammett indicators. The alkali metal-doped CaO catalysts (calcined and uncalcined) show higher base strength than CaO alone (8.2< pK. _B H+<)LQ.l), This means that doping the CaO with an alkali metal enhances the base strength of the catalyst and calcination is not required. When MgO is used as a support, the base strength is lower. After calcination, the base strength of Li on MgO increased but thews was no change for Na

and K on MgO. This implies that with Li as a dopant, there is a change in the structure of the MgO based catalyst. None of the alumina supported catalysts achieved a high level of base strength but after calcination the base strength of the alumina supported catalysts increased. The lower base strengths of alumina supported catalysts can be attributed to the non-basic nature of alumina.

Catalysts with enhanced base strength are active for the trausestcrification of vegetable oil, namely alkali metal doped-CaO catalysts and Li doped MgO after calctø&tion. After reaction for three hours, conversions in excess of 90% were achieved. Every catalyst with a base strength of higher thai), pK _BH+ >11 was active which shows that base strength is an important factor. There is no observed effect wh ^β λ changing the alkali metal dopant on CaO supported catalysts. None of the alumina-supported catalysts are active. This is due to the low loading of alkali metal (5%) compared with the 15% - 45% used in previous studies [Xie [supra]].

A homogeneous base catalysed reaction proceeds via the formation, of a methoxide ion which attacks the carbonyl group on the triglyceride to create a tetrahedral intermediate to form the biodiesel ester and lose the glycerol backbone (Sohuchardt, TJ., Sercheli, R. and Vargas SLM. Transesteriβcation of vegetable oils: a review. Journal of the Brazilian Chemical Society 9 (1997) 199-210). It is likely that the heterogeneous base catalysed reaction will proceed through a similar route and so the base strength, must be high in order that a proton can be abstracted from the methanol to attack the carbonyl group. Other studies on solid catalysts have noted an improvement in activity with increasing base strength (Suppes et al [supra] and Kim, H,, Kang, B., Kim, M., Park, Y.M., Kim, D., Lee, J. and Lee, K. Transesterification of vegetable oil to btodiesel using heterogeneous base catalyst. Catalysis today 93-95 (2004) 315-320).

Table 1: Properties of catalysts under test and conversion achieved in initial screening (60 ?C, 3 hours, 6:1 molar ratio methanol to oil, 5% catalyst).

Na,- Al ₂ Q ₃ (calcined) 10.1< PKBHí<11 6%

K- AIiOa (calcined) io.κpκ _B Hí<π 5%

G>) Reusability

As 1he reaction is normally catalysed by homogeneous alkali metal hydroxides or methoxides, it is important to check that these species have not leached from the solid catalyst and are not responsible for the observed activity of the catalysts. Na-OO catalyst was filtered and reused seven times with no loss in activity so it is unlikely that species are leached quickly from the catalyst into the solution.

Table 2 shows the iesidual alkali metal concentration in the bϋodiesel product. The uncalcined Na-CaO and K-CaO catalysts lost less than: 1% of the mass of alkali metal loaded onto the catalyst so the rate of loss is very small. However, when the catalyst had bean calcined, there was no measurable concentration of K in the biodiesal. This means that there was no leaching from the catalyst into the biodiesel product and catalysis is definitely heterogeneous. The catalyst will therefore not lose activity over time. The difference in behaviour between the calcined and non-calcined catalysts implied that there is a structural change due to calcination. As the K. species do not leach from the catalyst, it is likely they are incorporated into the lattice.

As the K-CaO catalyst was the most heterogeneous, it was this catalyst which was studied in most depth.

Table 2: Residual alkali metal levels in biodiesel as measured by flame photometry.

(c) Bench scale tests

Reaction profiles are shown for CaO m Figure 9, for Li-CaO in Figure 10, for Na- CaO in Figure 11, for KCaO in Figure 12, foi KCaO calcined in Figure 13 and for Li- MgO calcined h\ Figure 14. The conversion obtained after ISO minutes vrith CaO as a catalyst is expectedly poor as it does not have a high base strength. However, 70% of the triglycerides are converted to lesser glycerldes and biodiesel attributed to the formation of calcium methoxide on the surface of CaO (Giyglewioz, S. Rapeseødot! methyl esters preparation using heterogeneous catalysts. Bioresource Technology 70 (1999) 249-253).

The reactions catalysed by Li-CaO, Na-CaO and K-CaO have similar conversions after ISO minutes. This shows that changing the alkali metal on the catalyst docs not significantly affect activity. When fc-CaO was calcined, the reaction rate increased. The possible stnictnral change on calcination also increases the activity of the catalyst The Li-MgO catalysed reaction had the highest iftte (less than 2% triglyceride rematøed after just 40 minutes).

The alkali-metal doped-CaO catalysed reactions have similar shaped conversion versus time profiles (Figures 10-14). A high initial rate of reaction and then a plateau is observed with a simultaneous increase in the concentration of mono- and di- glycerides. Finally after -the plateau, the concentration of triglycerides decreases and the concentration of partial glycerides decreases more slowly. This plateau was somewhat reduced when the K-CaO caialyst was calcined. The Li-MgO catalyst did not exhibit the same profile but was more as would be expected as the concentrations of partial glycerides peaked early on in the reaction and were consumed as the reaction progressed.

The conversion versus time profile of the alkali metal doped-CaO catalysed feaction has a different shape to that of a homogeneously catalysed reaction where the partial glycerides have a maximum concentration just after the start of the reaction and are consumed as the reaction progresses and their concentration decreases to a level close to zero wtøre it remains (Vicente, G-, Martinez, M., Aracil J. and Estβbaπ, A. Kinetics of sunflower oil methmolysis. Industrial and Engineering Chemistry Research 44 (2005) 5447-S4S4). Similarly in heterogeneoυsly catalysed short-chain triglyceride traiϊ≤esterfjoation reactions tihe concentrations of the partial glycerides peaks at the beginning of the reaction period (Watkάns et a\ [supra] and Lopez, P., Goodwin, J., Bruce, D. and Lotero, B, Transesterification oftriacetin with methanol on solid add and base catalysts. Applied Catalysis A: General 295 (2005) 97-105). The reaction between soybean oil and methanol at 100 °C catalysed by a zeolite also shows a peak in partial glyceridc concentration at the start of the reaction (Suppβs ef al [supra]). Since we would expect the mechanism and reaction scheme to be similar to that in these previous studies, it seems unlikely that the plateau is attributable to a chemical effect It U much more probable that the plateau is due to a mass transfer affect.

The transesterification of triglycerides to biodiesel consists of three sequential steps (figure IS). The triglyceride reacts with methanol to create a diglyceride and loses a biodiesel ester. The digiyceride then reacts with methanol to create a monoglyceride and loses a second biodiesel ester. The monoglyceride then loses its final fatty acid chain as a biodiesel ester to leave glycerol. Glycerol is a very viscous and sticky substance and could coat the catalyst particles creating a mass transfer film. With reference to figure 9, it can be seen that the reaction rate slows dramatically after 40 minutes. At this time, approximately 70% of the triglycerides have reacted to biodiesel and glycerol and there are negligible concentrations of partial glycerides. Glycerol is insoluble in biodiesel and forms a separate phase and (unlike biodiesel) will not be washed off the catalyst particle and could create a film.

The presence of this film has two effects, The first effect is to block access to the active sites and the second effect is to create a local environment to the catalyst that is rich in glycerol and relatively poor in methanol and triglycerides. TMs proposed phenomenon leads to the two effects seen on the reaction profiles - the plateau and the relatively high concentration of partial glycerides.

Firstly, the film creates a mass transfer barrier for the methanol and triglycerides and they must difftise through the glycerol layer to reach the catalyst. This slows the rate of reaction creating the plateau region.

Secondly, the environment immediately surrounding the catalyst is rich in glycerol. This pushes the equilibrium of the reaction as shown in figure 3 back towards the reaclants and so favours the formation of the intermediate products. Therefore the mono- and di-glycerides are formed but not consumed because in the area surrounding the catalyst there is too much, glycerol.

The % mass of the glycerides remaining fox each catalyst is shown in table 3, Comparing the remaining glycerides with the level set by EU standard ENl 4214 (also shown in table 3), it can be seen that the biodiesel does not meet the standard and would have to be purified before being sold, The Li-MgO catalysed reaction meets the standard sat for triglyceride level but not for mono- and di-glyeeride.

To minimise costs, the amount of purification required would be minimised and ideally the reaction would reach as near as possible to 100% completion to avoid the fleβd for purification steps. To achieve this, the conditions of the reaction were changed to favour the formation of products. The glycerides remaining after 1, 2 and 3 hours are shown in table 4. Increasing the amount of catalyst front 2 g to 4 g did not increase the rate of the reaction - after 60 minutes there was Si-Jl 20% mass triglyceride which is comparable with the reference reaction. However the final concentrations of mono-, di- and triglycerides were reduced from the reference levels. Increasing the molar ratio of methanol increased the rate of the reaction such that only 5% mass triglyceride remained after 1 hour and the final concentrations of glycerides were also slightly reduced. However, neither process brought the biodiesel produced in line with the standard.

As the formation of a glycerol film slows the reaction by creating a mass transfer barrier, a two stage process was also tested. The products were removed front the reactor after 30 minutes, the liquid separated from the catalyst and replaced into the reactor and the reaction, continued with, the addition of new catalyst λ comparison of the glyeeride concentrations in the 1 and 2 stage processes is shown in table 5 and figure IS. The biggest difference is that after 40 minutes in the two stage process, there is only 6 % mas$ triglyceride but in the one stage process there is 54% mass triglyceride. There does appear to be a dear advantage in the two stage process in that fresh catalyst without a film of glycerol allows the reaction tσ proceed unhindered. This result offers some proof that there is a mass transfer limitation. The advantage of the two stage process decreases at longer times and after 80 minutes the glyceride concentration in the two stage process is only slightly less than for the one stage process. In the two Stage process ihe plateau region is removed, but the reaction does not proceed fully to completion because it is limited by the equilibrium between glycerol and partial glycerides,

Conclusions

Four active catalysts for the transesterification of rapeseβd oil to biodiesel have been identified, namely Li-CaO, Na-CaO, K-CaO and Li-MgO (only after calcination). There is a clear correlation between base strength and activity. The time for the reaction to progress to a reasonable conversion is aot prohibitively long, especially in the case of the most active catalyst Li-MgO -where after 40 minutes less than 5 % triglycerides remain. The tiniescales required for these catalysed reactions to reach,

near completion are feasible on a large scale and the preparation of these catalysts is not prohibitively difficult ox costly.

The Na-CaO catalyst was reused seven times without loss in activity and the calcined K-CaO catalyst lost no K through leaching into the biodiesøl produced. The catalysts will therefore have a reasonable lifetime in an industrial setting. However, there is a mass transfer limitation, possibly caused by the formation of a glycerol film limiting access to the active sites of the catalyst This can be partially overcome by employing a two-stage process, ^*

Table 3: remaining gfycerides after 180 minutes reaction, compared to EU standard.

Table 4: Glycerides remaining with 4g K-CaO calcined catalyst (2x increase) and 9: 1 molar ratio methanol to oil (1.5x increase) at 1, 2 and 3 hours.

Table 5: Comparison of one- and two-stage process where fresh catalyst added after 30 minutes