Scholarly article on topic 'Kinetics of the pre-treatment of used cooking oil using Novozyme 435 for biodiesel production'

Kinetics of the pre-treatment of used cooking oil using Novozyme 435 for biodiesel production Academic research paper on "Chemical sciences"

CC BY
0
0
Share paper
OECD Field of science
Keywords
{Biodiesel / "Liquid chromatography" / "Used cooking oil" / "Novozyme 435" / "Kinetic model" / "Fatty acid methyl esters (FAMEs)" / Esterification / "Ping Pong Bi Bi mechanism"}

Abstract of research paper on Chemical sciences, author of scientific article — Kathleen F. Haigh, Goran T. Vladisavljević, James C. Reynolds, Zoltan Nagy, Basudeb Saha

Abstract The pre-treatment of used cooking oil (UCO) for the preparation of biodiesel has been investigated using Novozyme 435, Candida antarctica Lipase B immobilised on acrylic resin, as the catalyst. The reactions in UCO were carried out using a jacketed batch reactor with a reflux condenser. The liquid chromatography–mass spectrometry (LC–MS) method was developed to monitor the mono-, di- and triglyceride concentrations and it was found that the method was sensitive enough to separate isomers, including diglyceride isomers. It was found that the 1,3 diglyceride isomer reacted more readily than the 1,2 isomer indicating stereoselectivity of the catalyst. This work showed that Novozyme 435 will catalyse the esterification of free fatty acids (FFAs) and the transesterification of mono- and diglycerides typically found in UCO when Novozyme 435 is used to catalyse the pre-treatment of UCO for the formation of biodiesel. A kinetic model was used to investigate the mechanism and indicated that the reaction progressed with the sequential hydrolysis esterification reactions in parallel with transesterification.

Academic research paper on topic "Kinetics of the pre-treatment of used cooking oil using Novozyme 435 for biodiesel production"

Contents lists available at ScienceDirect

Chemical Engineering Research and Design | ^ |

journal homepage www.elsevier.com/locate/cherd

Kinetics of the pre-treatment of used cooking oil ^ using Novozyme 435 for biodiesel production

CrossMark

Kathleen F. Haigha, Goran T. VladisavljeviCa, James C. Reynoldsb) Zoltan Nagya, Basudeb Sahac'*

a Department of Chemical Engineering, Loughborough University, Loughborough, Leicestershire LE11 3TU, United Kingdom

b Department of Chemistry, Loughborough University, Loughborough, Leicestershire LE11 3TU, United Kingdom c Centre for Green Process Engineering, Department of Applied Sciences, Faculty of Engineering, Science and the Built Environment, London South Bank University, London SE1 0AA, United Kingdom

abstract

The pre-treatment of used cooking oil (UCO) for the preparation of biodiesel has been investigated using Novozyme 435, Candida antarctica Lipase B immobilised on acrylic resin, as the catalyst. The reactions in UCO were carried out using a jacketed batch reactor with a reflux condenser. The liquid chromatography-mass spectrometry (LC-MS) method was developed to monitor the mono-, di- and triglyceride concentrations and it was found that the method was sensitive enough to separate isomers, including diglyceride isomers. It was found that the 1,3 diglyceride isomer reacted more readily than the 1,2 isomer indicating stereoselectivity of the catalyst. This work showed that Novozyme 435 will catalyse the esterification of free fatty acids (FFAs) and the transesterification of mono- and diglycerides typically found in UCO when Novozyme 435 is used to catalyse the pre-treatment of UCO for the formation of biodiesel. A kinetic model was used to investigate the mechanism and indicated that the reaction progressed with the sequential hydrolysis esterification reactions in parallel with transesterification.

© 2014 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Keyuiords.Biodiesel; Liquid chromatography; Used cooking oil; Novozyme 435; Kinetic model; Fatty acid methyl esters (FAMEs); Esterification; Ping Pong Bi Bi mechanism

1. Introduction

Biodiesel is defined as mono-alkyl esters of long chain fatty acids derived from renewable sources, particularly vegetable oil and animal fats (Knothe, 2010; Zainal-Abidin-Murad et al., 2012). The most common commercial process is to convert vegetable oil to biodiesel by means of a transesterification reaction, with methanol in the presence of a basic catalyst, such as sodium hydroxide, resulting in the formation of fatty acid methyl esters (FAME), as shown in Fig. 1. This reaction proceeds in a stepwise manner with the triglycerides (TG) being converted to diglycerides (DG) and the DGs subsequently converted to monoglycerides (MGs).

Vegetable oil is an expensive raw material and there are ethical concerns with using a potential food source for fuel. As a result alternative raw materials such as Jatropha curcas

(Patil et al., 2009), by-products from oil refining such as palm fatty acid distillate (Talukder et al., 2009), animal fats, algal oil (Semwal et al., 2011) and used cooking oil (UCO) (Akoh et al., 2007; Enweremadu and Mbarawa, 2009) are currently being investigated to mitigate these issues. The advantage of UCO is that a waste material is diverted from landfill, simultaneously eliminating the competition with food and reducing the cost. The cooking process causes the vegetable oil, TGs, to breakdown to form, DGs, MGs, and free fatty acids (FFAs). The FFAs react with the basic catalysts used during transesterification in a saponification side reaction which can form an emulsion and reduce product yield. In addition, most biodiesel specifications impose an upper limit on the FFAs content as FFAs can cause deposits and engine damage (Knothe, 2005). Ester-ification can be used to convert the FFAs to biodiesel and a schematic of the esterification process is shown in Fig. 2.

* Corresponding author. Tel.: +44 02078157190; fax: +44 02078157699. E-mail address b.saha@lsbu.ac.uk (B. Saha).

Received 26 July 2013; Received in revised form 2 December 2013; Accepted 4 January 2014 0263-8762/$ - see front matter © 2014 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.10167j.cherd.2014.01.006

Alcohol Glycerol

Fig. 1 - A schematic representation of the transesterification reaction.

Catalyst

Free Fatty Acid Alcohol

Fig. 2 - A schematic representation of the esterification reaction.

Fatty Acid Alkyl Water Ester (Biodiesel)

The heterogeneous catalysts used to investigate esterification for biodiesel production can be classified as inorganic acid catalysts (Cordeiro et al., 2008; Park et al., 2010), ionexchange resin catalysts (Ozbay et al., 2008; Russbueldt and Hoelderich, 2009) and immobilised enzymes (Talukder et al., 2009; Souza et al., 2009). Ion-exchange resins and immobilised enzymes are more suitable for an environmentally benign process because relatively high conversions are possible at relatively benign conditions (Enweremadu and Mbarawa, 2009; Zainal-Abidin-Murad et al., 2012)

Novozyme 435, Candida antarctica Lipase B immobilised on acrylic resin, was investigated for the esterification pre-treatment of biodiesel from UCO and the catalytic performance was compared to an ion-exchange resin catalyst, Purolite D5082 (Haigh et al., 2013). It was found a faster conversion was possible at more benign conditions when Novozyme 435 was used compared to Purolite D5082. In addition, it was found that at the higher reaction temperatures (50-60 °C) the amount of FAME formed was greater than the amount of FFAs consumed based on the esterification reaction as shown in Fig. 2. Lipases including Novozyme 435 have also been investigated for transesterification and hydrolysis (Ganesan et al., 2009; Lam et al., 2010; Tongboriboon et al., 2010) although it has been found that the reaction rate is slow compared to esterification. In order to determine if Novozyme 435 can simultaneously catalyse esterification and transesterification reactions in UCO, it is necessary to monitor the MG, DG and TG concentrations in addition to FAME and FFAs. Gas chromatography (GC) and liquid chromatography (LC) are the most common methods for investigating the production of biodiesel (Li et al., 2008). Substances with high molecular weights, high boiling points and low volatility are not easily vapourised and separated by GC, although this problem can be overcome by means of a silylation reaction and the use of internal standards (Holcapek et al., 1999; Li et al., 2008). LC is a versatile analytical technique and most samples do not require derivatisation and as a result numerous methods have been investigated for the separation and quantification of biodiesel components (Holcapek et al., 1999; Li et al., 2008; Santori et al., 2009). Ultra violet (UV) and mass spectrometry (MS) detectors have been investigated for the detection of biodiesel components (Holcapek et al., 1999; Turkan and Kalay, 2006; Santori et al., 2009). The disadvantage of UV detectors for complex

mixtures is that the response is based on the concentration and number of double bonds. MS detectors are useful because they can also provide structural information.

Kinetic modelling has been used to investigate the esterification of FFAs to form biodiesel. In most cases a model system, with a single fatty acid is used and it has been found that the reaction follows a Ping Pong Bi Bi kinetic model (Al-Zuhair et al., 2006; Mahmud et al., 2010). A schematic of the Ping Pong Bi Bi mechanism for esterification is shown in Fig. 3 with an FFA first reacting with the enzyme to form a complex (E-FFA) leading to the formation of water. Methanol (MeOH) can then react with the surface to form the E-FAME and the FAME is subsequently released.

Three mechanisms were assessed for the transesterification of palm oil, using immobilised Pseudomonas sp. and ethanol based on the variation of the TG, MG, DG, biodiesel and alcohol concentrations (Cheirsilp et al., 2008). Mechanism 1 assumed hydrolysis followed by esterification with the overall reactions following a Ping PongBi Bi mechanism. Mechanism 2 assumed that the complexes bound the enzyme surface react fast and was a simplified version of Mechanism 1. Mechanism 3 assumed that transesterification (direct alcoholysis) of the triglycerides occurred in parallel with the hydrolysis esterifi-cation reaction sequence. The mechanism which best fitted the experimental data was Mechanism 3.

This work has investigated the use of LC-MS to determine the concentration of TG, DG and MGs in the reaction mixture during the esterification pre-treatment of UCO for the production of biodiesel. This data was subsequently analysed based on the model proposed by Cheirsilp et al. (2008) for simultaneous esterification and transesterification, i.e., Mechanism 3. The aim of this work was to determine the components and reactions that contribute to the additional FAME formation during the esterification pre-treatment of UCO for the

Fig. 3 - Schematic representation of the Ping Pong Bi Bi mechanism for esterification of UCO.

production of biodiesel using Novozyme 435 and develop a novel model to describe the reactions.

2. Experimental

2.1. Materials

Methyl ester, mono-, di- and triglyceride standards were purchased from Aldrich, UK. Novozyme 435 was donated by Novozyme UK Ltd. and used as supplied. The UCO was supplied by GreenFuel Oil Co., Ltd., UK and had an average molecular weight of 278gmol—1 and FFAs content of approximately 6.4wt%. Analytical grade methanol, toluene and 2-propanol were purchased from Fisher Scientific UK Ltd. HPLC grade 2-propanol and acetonitrile were used for the HPLC work and purchased from Fisher Scientific UK Ltd. All solvents were used as supplied.

2.2. Esterification of UCO

The esterification reactions were carried out using a jacketed batch reactor with a reflux condenser. The stirrer motor was a Eurostar Digital IKA-Werke. The temperature was monitored by means of a Digitron, 2751-K thermocouple and this information was used to set the temperature on the Techne, TE-10D Tempette water bath. The UCO and methanol were added to the reactor and heated to the required temperature, after sampling, the catalyst was added to initiate the esterifica-tion reaction. The sample tube was fitted with metal gauze to prevent withdrawal of catalyst when taking samples and the samples were withdrawn by means of a syringe. All samples were analysed for FFAs, FAME, MG, DG and TG concentrations.

2.3. FAME and FFA analysis

FAME concentration was determined using gas chromatography-mass spectrometry (GC-MS) (Hewlett Packard HP-6890), equipped with a DB-WAX (J & W Scientific) capillary column (30 m x 0.25 mm) packed with polyethylene glycol (0.25 ^m film thickness); helium gas at a flow rate of 1.1mLmin—1 was used as the carrier gas. The amount of sample injected was 2 ^L. The temperature of the injector and detector was 250 ° C. The initial oven temperature of 70°C was held for 2min, then increased at 40°Cmin—1 up to 210 °C, then increased at 7 °Cmin—1 up to 230 °C and the final temperature was held for 11 min. Methyl heptadecanoate was used as the internal standard.

The %FFAs in all samples was determined by titration using the ASTM D974 method. 2 g of sample was dissolved in 100 mL of a solution of toluene:2-propanol:water (volume ratio of 100:99:1) and titrated using p-naphtholbenzein as an indicator.

2.4. Details of the Ultra performance liquid chromatography-mass spectrometry (UPLC-MS) method

The LC-MS was carried out using a Waters Acquity ultra performance liquid chromatography (UPLC) system interfaced to a Waters Synapt HDMS quadrupole time-of-flight (TOF) mass spectrometer, using an electrospray ionisation source operating in the positive ion mode. A Phenomenex Kinetix C18 UPLC column (150 mm x 2.1mm x 2.1 ^m) was used for the separation. The chromatography used a binary method with acetonitrile as solvent A and 2-propanol as solvent B. The separation was carried out using a binary gradient with a flow

Fig. 4 - Conceptual scheme of the reaction mechanism investigated.

rate of 0.15mLmin—1 starting with 90% acetonitrile and 10% 2-propanol changing to 30% acetonitrile in 20 min.

3. Development of the model

The model used in this work was based on the model proposed by Cheirsilp et al. (2008), i.e., Mechanism 3. In order to simplify the model it was assumed that the components can be grouped together. It was also assumed that the inhibition of the enzyme by alcohol followed the competitive inhibition model. It has been shown that for the reaction conditions investigated, there are no mass transfer limitations (Haigh et al., 2013) and as a result this model applies to the intrinsic reaction kinetics.

It was found that the water (W) concentration increased when using UCO and Novozyme 435 as the catalyst. This is because the FFAs initially present in UCO lead to the esterification reaction and water formation. In addition, it is not necessary to add water to the reaction mixture when Novozyme 435 is used as the catalyst. As a result, the model proposed by Cheirsilp et al. (2008) was amended to allow for water formation and the resulting conceptual scheme for the reaction is shown in Fig. 4. According to this reaction scheme, the free enzyme (E) reacts with alcohol (Al), triglycerides (T), diglycerides (D), monoglycerides (M) and FFAs (F) to form the associated complexes E-Al, E-T, E-D, E-M and E-F, respectively. The hydrolysis and esterification reactions are then expected to occur in parallel with the E-T, E-D and E-M complexes reacting with W to form F or with Al to form esters (Es) eventually leading to the formation of glycerol (G).

The rate expressions have been developed by assuming that the hydrolysis and esterification reactions are rate limiting and the resulting rate equations are:

dTi = — (Vmx[W| + VeT[Al])[T][E1

dp = ((VmT[W] + VeT[Al])[T] — (VmD [W] + Veo[Al])[D])|E*]

dM = ((VmD [W] + VeD[Al])[D] — (VmM [W] + VeM[Al])[M])[E*]

= (VmM[W] +VeM[Al])[E*j

d[Fj _ -d[Wj dt = dt

= ((VmT[Tj + VmD[Dj + VmM[M])[W] - VeEs[Fj[Alj)[E*j

d[Esj -d[Al]

= (VeT[Tj + VeD[Dj + VeM [Mj + VeEs[Fj)[Alj[E*j

where the [E*j is: [E*j =

1 + KmT [Tj + KmD[Dj + KmM[Mj + KmF[F] + (Al/Ki)

VmT, VmD and VmM are the hydrolysis rate constants and are defined as: VmT = kskj^, VmD = k7k5/ks and VmM = kj^/k^. VeEs is the esterification rate constant and is defined as VeEs = ki5ki3/ki4. VeT, VeD and VeM are the rate constants for transesterification and are defined as: VeT = k4k1/k2, VeD = k8k5/k6 and VeM = k12k9/k10. The equilibrium constants KmT, KmD, KmM, KmF and the inhibition constant KI are defined as: KmT = k1/k2, KmD = k5/ks, KmM = k9/kw, KmF = k13/k:4 and Ki= k!7/k16.

The unknown parameters were determined by fitting the model equations to the batch experimental data using MAT-LAB.

4. Results and discussion

4.1. Results of the LCMS analysis

A typical chromatogram generated by LC-MS for UCO is shown in Fig. 5 using the base peak ion (BPI) setting. UCO is known to consist of a mixture of various MG, DG and TGs, and as a result the chromatogram for UCO has a large number of peaks. The analysis was carried out using a mass spectrometer and

Fig. 5 - A typical UPLC chromatogram of used cooking oil (UCO) using the base peak ion (BPI) setting. The inset shows the extracted ion chromatogram for the 1,2 and 1,3 isomers of dioleoyl-glycerol which gives a sodiated molecular ion [M+Na]+ at m/z 643.5.

1 Sample 9-240 min reaction time 1

Jill Sample 3-20 min reaction time

Standard diolein predominantly 1,3 isomer A

0 5 10 15 20

Time (min)

Fig. 6 - Selected ion LC-MS chromatograms showing depletion of 1,3 positional isomer of dioleoyl-glycerol [M+Na]+ ion at m/z 643.5 and comparison with the dioleoyl-glycerol standard containing predominantly 1,3 positional isomers.

therefore an extracted ion chromatogram for each species can be generated based on the mass-to-charge (m/z) ratio. Each component has a positive charge due to the incorporation of an environmental Na+ ion. The extracted ion chromatogram for dioleoyl-glycerol is shown as an inset in Fig. 5.

UCO has a complex composition and in order to simplify the method it has been assumed that each component of a given species gives a similar mass spectrometric response. TVo calibration standards were used for each of the MG, DG and TG species. It can be seen from the inset in Fig. 5 that many of the components form multiple peaks due to isomers and as a result the total area under all peaks for the relevant component was used as the peak area.

It has been found that individual diglycerides can be separated into two peaks corresponding to the 1,2 and 1,3 positional isomers of the diglyceride. During the reaction these peaks disappear at different rates, an example of this is shown in Fig. 6. The peaks at two reaction times for dioleoyl-glycerol have been compared to the 1,3 diolein standard which is composed predominantly of the 1,3 isomer. These data show that the second peak corresponds to the 1,2 isomer, and that this isomer is transesterified preferentially over the 1,3 isomer by the Novozyme 435 enzyme catalyst. A comparison of the concentration of the two palmitoyl-oleoyl-glycerol isomers during the reaction is shown in Fig. 7. The concentration data has been calculated for this species because there is better separation of the isomers than dioleoyl-glycerol and it has been

Fig. 7 - A comparison of the rate of disappearance of diglyceride peaks (palmitoyl-oleoyl-glycerol at m/z 617.5).

Table 1 - Details of the composition of UCO.

Property Value

Fatty acid composition (wt%)

Linoleic acid 43

Oleic acid 36

Palmitic acid 13

Stearic acid 3.8

Linolenic acid 3.6

Molar mass, average TGs (gmol—1) 867.4

FFA (wt/wt%) 8.42

Acid value (mg KOH g—1) 16.6

TG concentration (%) 84

DG concentration (%) 7.0

MG concentration (%) 0.3

assumed that the isomers elute in the same order. An enzyme catalyst, Novozyme 435, is being used to catalyse the reaction and this data shows that this catalyst has some stereoselectivity. Enzymes have been investigated for the manufacture of biodiesel, but this data indicates that information on the fatty acid composition of a particular oil may not be sufficient to determine how well the enzyme will catalyse a particular reaction. The isomer composition can also determine the rate and extent of reaction.

4.2. Results of the batch experiments

A summary of the composition of the UCO used for these experiments is given in Table 1. A comparison of the FAME formation compared to FFAs consumption at 50 °C using Novozyme 435 as the esterification catalyst is shown in Fig. 8. A schematic of the esterification of FFAs to FAME for the pre-treatment of UCO is shown in Fig. 2. From this data it can be seen that the FAME formation exceeds the FFAs consumption by a factor of approximately 2.

It has been proposed that the additional FAME formation is due to transesterification side reactions and an overall schematic for the transesterification reaction is shown in Fig. 1. This reaction proceeds in a stepwise manner with the TGs converted to DGs and the DGs converted to MGs. A comparison of the change in concentration of MG, DG, TG, FAME and FFAs is shown in Fig. 9. The MG concentration increases slightly and then decreases, which is typical of a reaction intermediate. The DG concentration decreases steadily during the course of the reaction. At the end of the 6 h reaction time most of the MG and DGs have been consumed and this

Fig. 8 - A Comparison of the FFAs consumption and FAME formation using Novozyme 435 as the esterification catalyst at 50 °C.

Fig. 9 - Change in the concentration of triglycerides (TG), diglycerides (DG), monoglycerides (MG), fatty acid methyl esters (FAME/biodiesel) and fatty acids (FFAs) during the pretreatment of UCO for biodiesel production. This experiment was carried out using Novozyme 435 as the catalyst at 50 ◦ C.

would account for the additional FAME formation. In contrast the change in TG concentration is very small. Overall, there is a slight decrease in the concentration of triglycerides. Thus while the TGs will react in the transesterification reaction, the rate is much slower when compared to the transesterification of MGs and DGs. This is consistent with previous experiments where it was found that when Novozyme 435 was used as a transesterification catalyst, the transesterification of triglycerides was the rate limiting step with no accumulation of MG or DGs during the reaction (Turkan and Kalay, 2006).

4.3. Kinetic modelling results

The differential equations (1)-(6) were solved by means of the MATLAB "ode23s" function. This was combined with the MAT-LAB "fminsearch" function to minimise the residual of the sum of squares, i.e., the difference between the experimental and calculated values. A comparison of the experimentally determined concentrations compared to those predicted by the model is shown in Fig. 10 for the MG, DG, TG and FAME concentrations. From Fig. 10 it can be seen that some of the concentration data is scattered particularly the TG and FAME concentrations. Overall the data followed a reasonable trend and as a result the scatter has been attributed to experimental error. These experiments have been repeated and the overall results are reproducible. In the case of the UPLC work a quality control sample was injected periodically during the run and this confirmed that the instrument performance remained consistent. There is a reasonable fit between the model and the experimental data.

In the case of DG concentrations it can be seen that the model predicts a faster initial decrease in the concentration than the experimental data. One of the assumptions made as part of the model development was that all the DG species could be treated as a single component. However, it can be seen from Figs. 6 and 7 that the 1,2 and 1,3 positional isomers react at different rates and the DG trend is the sum of these two trends. The model could be improved by allowing for both DG isomers, however, this would increase the complexity of the model and substantially increase the complexity of the component analysis.

The parameters predicted by the model for this system are shown in Table 2. These values have been compared to values

Fig. 10 - A comparison of the concentrations from the experiments compared to those predicted by the model with (a) the MG, DG and TG concentrations and (b) showing the FAME concentrations.

found in the literature (Cheirsilp et al., 2008). From this data it can be seen that there are substantial differences between these values, as expected, because the systems investigated are different. Novozyme 435 was used instead of immobilised Pseudomonas sp., methanol was used in place of ethanol, water was not added to the reaction mixture and the model has been modified to allow for water formation during the esterification reaction. Novozyme 435 is known to result in slow reactions with TGs (Tongboriboon et al., 2010; Ganesan et al., 2009) and the parameters for the TG reactions (VmT and VeT) are lower

Table 2 - The model parameters.

Parameters Predicted values Literature valuesa

(mol-1 min-1) (mol-1 min-1)

VmT 0.0130 1.27

VmD 0.447 1.35

VmM 78.6 3.25

VeEs 48.4 23.1

VeT 0.00430 45.9

VeD 25.5 19.6

VeM 18.5 16.1

KmT 0.00770 0.482

KmD 0.00290 0.387

KmM 3.92 0.329

KmF 0.00410 0.187

Ki 764 14.7

Key experimental parameters

Catalyst type Novozyme 435 Pseudomonas sp.

Reagent Methanol Ethanol

Water addition No Yes

a Cheirsilp et al. (2008).

than the literature values. From this data it can be seen that the DGs and MGs react more readily than the TGs. As expected the parameter of the esterification of FFAs (Ves) is greater than the literature value because Novozyme 435 has been shown to favour the esterification reaction (Tongboriboon et al., 2010). Overall the data indicates that Novozyme 435 can catalyse the transesterification reaction in parallel with the hydrolysis esterification reaction sequence. The value for the alcohol inhibition constant (KI) is greater than the literature value and this can be accounted for because methanol was used instead of ethanol. The two catalysts exhibit different levels of alcohol tolerance.

5. Conclusions

The pre-treatment of UCO, with Novozyme 435 as a catalyst, for the preparation of biodiesel has been investigated using LC-MS to monitor the mono-, di- and triglyceride concentrations with kinetic modelling to assess the mechanism. It has been shown that the LC-MS method has sufficient sensitivity to monitor the progress of specific isomers with the 1,3 DG isomers being consumed faster than the 1,2 isomer.

This work was carried out because earlier experiments focused on the esterification pre-treatment of UCO to form biodiesel showed that there was an excessive formation of FAME at high temperatures. From the MG, DG and TG data it can be seen that the MGs and DGs are consumed during the reaction while the TG concentrations decreases slightly. Cheirsilp et al. (2008) proposed that enzyme catalysed reactions proceed with transesterification in parallel with the hydrolysis esterification sequence of reactions. The model was amended to allow for water formation and it was found that there was a good fit between the model and the experimental results. These results show Novozyme 435 can simultaneously catalyse the hydrolysis, esterification and transesterification reactions in UCO.

Novozyme 435 has been investigated as an esterification pre-treatment catalyst for the production of biodiesel. However, these results indicate that alternative uses of Novozyme 435 duringbiodiesel production maybe more suitable. Enzyme catalysts are less sensitive to FFAs compared to their chemical counterparts. As a result, enzymatic transesterification could be carried out first followed by a simultaneous esteri-fication and transesterification reaction using Novozyme 435. This process could be used to ensure that all FFAs as well as unreacted MGs and DGs are removed from the product stream.

Acknowledgements

We would like to thank EPSRC for the PhD scholarship to KH. We would also like to thank GreenFuel Oil Co., Ltd., UK for supplying the UCO and Novozymes UK Ltd. (Dr. David Cowan) for supplying the enzyme catalyst and his help and advice with using Novozyme 435 for this project.

References

Akoh, C.C., Chang, S.-W., Lee, G.-C., Shaw, J.-F., 2007. Enzymatic approach to biodiesel production. J. Agric. Food Chem. 55, 8995-9005.

Al-Zuhair, S., Jayaraman, K.V., Krishnan, S., Chan, W.-H., 2006. The effect of fatty acid concentration and water content on the production of biodiesel by lipase. Biochem. Eng. J. 30, 212-217.

Cheirsilp, B., H-Kittikun, A., Limkatanyu, S., 2008. Impact of transesterification mechanisms on the kinetic modeling of biodiesel production by immobilized lipase. Biochem. Eng. J. 42, 261-269.

Cordeiro, C.S., et al., 2008. A new zinc hydroxide nitrate

heterogeneous catalyst for the esterification of free fatty acids and the transesterification of vegetable oils. Catal. Commun. 9, 2140-2143.

Enweremadu, C.C., Mbarawa, M.M., 2009. Technical aspects of production and analysis of biodiesel from used cooking oil—a review. Biochem. Eng. J. 13, 2205-2224.

Ganesan, D., Rajendran, A., Thangavelu, V., 2009. An overview on the recent advances in the transesterification of vegetable oils for biodiesel production using chemical and biocatalysts. Rev. Environ. Sci. Biotechnol. 8, 367-394.

Haigh, K.F., Abidin, S.Z., Vladisavljevic, G.T., Saha, B., 2013. Comparison of Novozyme 435 and Purolite D5081 as heterogeneous catalysts for the pretreatment of used cooking oil for biodiesel production. Fuel 111, 186-193.

Holcapek, M., Jandera, P., Fischer, J., Prokes, B., 1999. Analytical monitoring of the production of biodiesel by high-performance liquid chromatography with various detection methods. J. Chromatogr. A 858 (1), 13-31.

Knothe, J.G., 2010. Biodiesel and renewable diesel: a comparison. Prog. Energy Combust. Sci. 36, 364-373.

Knothe, G., 2005. Fuel properties. In: Knothe, G. (Ed.), The

Biodiesel Handbook. AOCS, Champaign, Illinois, pp. 76-164.

Lam, M.K., Lee, K.T., Mohamed, A.R., 2010. Homogeneous,

heterogeneous and enzymatic catalysis for transesterification of high free fatty acid oil (waste cooking oil) to biodiesel: a review. Biotechnol. Adv. 28, 500-518.

Li, Y., Bao, G., Wang, H., 2008. Determiniation of 11 fatty acids and fatty acid methyl esters in biodiesel using ultra performance liquid chromatography. Chin. J. Chromatogr. 26, 494-498.

Mahmud, M.S., Safinski, T., Nelson, M.I., Sidhu, H.S., Adesina, A.A., 2010. Kinetic analysis of oleic acid esterification using lipase as catalyst in a microaqueous environment. Ind. Eng. Chem. Res. 49, 1071-1078.

Ozbay, N., Oktar, N., Tapan, N., 2008. Esterification of free fatty acids in waste cooking oils (WCO): role of ion-exchange resins. Fuel 87, 1789-1798.

Park, Y.-M., et al., 2010. Esterification of used vegetable oils using the heterogeneous WO3/ZrO2 catalyst for production of biodiesel. Bioresour. Technol. 101 (Suppl.), S59-S61.

Patil, P.D., Gude, V.G., Deng, S., 2009. Biodiesel production from Jatropha curcas, waste cooking, and camelina sativa oils. Ind. Eng. Chem. Res. 48, 10850-10856.

Russbueldt, B.M.E., Hoelderich, W.F., 2009. New sulfonic acid ion-exchange resins for the preesterification of different oils and fats with high content of free fatty acids. Appl. Catal. A: Gen. 362, 47-57.

Santori, G., Arteconi, A., Di Nicola, G., Moglie, M., Stryjek, R., 2009. Quantitation of compounds in biodiesel mixtures with reversed-phase liquid chromatography. Energy Fuels 23, 3783-3789.

Semwal, S., Arora, A.K., Badoni, R.P., Tuli, D.K., 2011. Biodiesel production using heterogeneous catalysts. Bioresour. Technol. 102, 2151-2161.

Souza, M.S., et al., 2009. Biodiesel synthesis via esterification of feedstock with high content of free fatty acids. Appl. Biochem. Biotechnol. 154, 74-88.

Talukder, M.M.R., Wu, J.C., Lau, S.K., Cui, L.C., Shimin, G., Lim, A., 2009. Comparison of Novozym 435 and Amberlyst 15 as heterogeneous catalyst for production of biodiesel from palm fatty acid distillate. Energy Fuels 23, 1-4.

Tongboriboon, K., Cheirsilp, B., H-Kittikun, A., 2010. Mixed lipases for efficient enzymatic synthesis of biodiesel from used palm oil and ethanol in a solvent-free system. J. Mol. Catal. B: Enzym. 67, 52-59.

Türkan, A., Kalay, S., 2006. Monitoring lipase-catalyzed methanolysis of sunflower oil by reversed-phase high-performance liquid chromatography: elucidation of the mechanisms of lipases. J. Chromatogr. A 1127, 34-44.

Zainal-Abidin-Murad, S., Haigh, K., Saha, B., 2012. Esterification of free fatty acids in used cooking oil using ion exchange resins as catalysts: an efficient pre-treatment method for biodiesel feedstock. Ind. Eng. Chem. Res. 51, 14653-14664.