**2. Esterification in solvent-free system**

After determining the optimal working parameters in the solvent-free system (molar ratio, temperature, enzyme concentration, initial water content) the reactions were completed in integrated system, where pervaporation – and effective and mild membrane separation process (Garcia, 1999) - was connected to the bioreactor.

#### **2.1 Experimental**

The catalyst used was Novozyme 435, a commercial *Candida antarctica* lipase (E.C. 3.1.1.3. Triacylglycerol acylhydrolase) immobilized on a macroporous acrylic resin with a water content of 1-2 % w/w. The enzyme was provided as a gift by Novo Nordisk A/S (Denmark). The nominal activity of the catalytic preparation was approximately 7000 Propyl Laurate Units (PLU/g). One propyl laurate unit (PLU) is defined as the number of mol of n-propyl laurate obtained in the standard test corresponding to the esterification of lauric acid with n-propyl alcohol, after 15 min at atmospheric pressure.

The fusel oil was provided as a gift by Distillery Győr (Hungary). All the other chemicals used in analysis were of analytical grade and purchased from Reanal Ltd. (Hungary) and Sigma Chemical Co. (USA).

Two different procedures were used for ester production. Firstly, synthesis of esters was carried out in shaking flasks (150 rpm) containing 25 ml solution of several alcohols and oleic acid mixture with different molar ratios, different temperatures and various amounts of enzyme by using New Brunswick Scientific (USA) shaking incubator to study the esterification kinetics. The starting time of the reaction was the addition of the enzyme.

In the other procedure a 200 ml round flask reactor was thermostated and connected with a pervaporation unit using hydrophilic membrane for continuous removal of water produced. The reaction mixture was circulated through the pervaporation unit by a peristaltic pump. The vacuum pump, manometer and the cooled traps were connected to the pervaporation unit.

The laboratory scale pervaporation unit was purchased from Carbone Lorraine (Germany) and it was jacketed later. The membrane surface area was 2.0\*10-2 m2. The membranes used for the pervaporation experiments (PERVAP 2201, PERVAP 2202, CMC-VP-43) were provided by GFT (Germany) and Celfa (Switzerland).

Aliquots of the reaction mixture were withdrawn periodically and the residual acid content was assayed by titrating against potassium hydroxide (0.1 M) using phenolphthalein as an indicator and ethanol as a quenching agent. The percentage esterification was calculated from the values obtained for the blank and the test samples. The fusel oil esters were confirmed by chromatographic analyses of the samples using a Hewlett Packard Model 5890 Series II GC equipped with a flame ionisation detector and a 30 m HP-FFAP capillary column. The percentage esterification calculated by both GC analysis (which showed product formation) and titrimetry (which showed acid consumption) were found to be in good agreement. The water content of the reaction mixture was measured by Mettler DL 35 automatic Karl-Fischer titrator

#### **2.2 Results**

122 Food Industrial Processes – Methods and Equipment

For example, esters produced from long-chain fatty acids (12–20 carbon atoms) and shortchain alcohols (3–8 carbon atoms) have been used increasingly in the food, detergent, cosmetic and pharmaceutical industries. Esters prepared from the reaction of long-chain acids with long-chain alcohols (12–20 carbon atoms) also have important applications as plasticizers and lubricants (Zaidi et al., 2002, Dossat et al., 2002). The direct effect of the ester group on the physical properties of a lubricant is to lower the volatility and raise the flash

Compared with conventional chemical synthesis from alcohols and carboxylic acids using mineral acids as a catalyst, the use of enzymes such as lipases to produce these high valueadded fatty acid esters in solvent-free media may offer many significant advantages (Yadav & Lathi, 2003). These include the use of any hydrophobic substrate, higher selectivity, milder processing conditions and the ease of product isolation and enzyme reuse. The ecological properties of oleochemical esters have been intensively studied within the last couple of years. In general, their aquatic toxicity is very low or almost negligible. For the aquatic compartment the fish, daphnia, algae and bacteria are the most relevant test organisms and standardized test

Esterification reactions by lipase in non-conventional media have been studied in our laboratory for long (Gubicza et al., 2003). Enzymatic esterification of fatty acids and ingredients of fusel oil was studied by (Gulati et al., 2003) using lipase from *Aspergillus tereus*. They found that in n-hexane solvent the alcohols were able to react with the fatty acids (miristic acid, palmitic acid, stearic acid), except oleic acid. Using other lipase preparations (*Candida antarctica, Candida rugosa, Rhizomucor miehei*, *Porcine pancreas*), however, made it possible the successful oleic acid esterification with similar low molecular weight alcohols. Description of the correct kinetics on the particular esterification reaction is

In our earlier work natural aroma esters were produced by enzymatic esterification in organic solvents and in solvent-free media (Gubiza, 2000). In this work the purpose was to find a utilisation of fusel oil, where bio-lubricants can be manufactured. The alcohol compounds of fusel oil were esterified with oleic acid using enzyme catalysis in nonconventional, solvent-free media (section 2) and in ionic liquid (section 4), moreover the

After determining the optimal working parameters in the solvent-free system (molar ratio, temperature, enzyme concentration, initial water content) the reactions were completed in integrated system, where pervaporation – and effective and mild membrane separation

The catalyst used was Novozyme 435, a commercial *Candida antarctica* lipase (E.C. 3.1.1.3. Triacylglycerol acylhydrolase) immobilized on a macroporous acrylic resin with a water content of 1-2 % w/w. The enzyme was provided as a gift by Novo Nordisk A/S (Denmark). The nominal activity of the catalytic preparation was approximately 7000 Propyl Laurate Units (PLU/g). One propyl laurate unit (PLU) is defined as the number of mol of

methods, such as laid down in the OECD methods 201–210 (Willing, 1999).

even more difficult due to the various possible inhibition effects.

kinetics of the reaction was described (section 3).

**2. Esterification in solvent-free system** 

**2.1 Experimental** 

process (Garcia, 1999) - was connected to the bioreactor.

point.

The esterification reaction of oleic acid with the fusel oil fraction occurs as follows:

oleic acid + fraction of fusel oil = oleate esters + water

In this reversible reaction, the molar ratio of reactants, temperature, enzyme and removal of water from the reaction mixture are the variables affecting the conversion and the reaction rate.

#### **2.2.1 Water content**

Water level is critical for enzymes and affects enzyme action in various ways: by influencing enzyme structure via noncovalent bonding and disruption of hydrogen bonds; by facilitating reagent diffusion; and by influencing the reaction equilibrium. Too low water content usually reduces enzyme activity. High water content can also decrease reaction rates by aggregating enzyme particles and causing diffusional limitations. The optimal amount of water is often within a narrow range.

Time [h] 01234567

Fig. 2. Influence of reaction temperature on the synthesis of isoamyl-oleate. Oleic acid, 1.84

It is well-known, that acid/alcohol molar ratio is one of the most important parameters in enzymatic esterifications. Since the reaction is reversible, an increase in the amount of one of the reactants will result higher ester yields and as expected, this will shifts the chemical equilibrium towards the product side. One way of shifting the reaction toward the synthesis is to increase the alcohol concentration. However, high alcohol concentration may slow down the reaction rates due to inhibition. Therefore, it is necessary to optimize the actual excess nucleophile concentration to be employed in a given reaction. In order to determine the effect of molar ratio, oleic acid was esterified at molar ratios of 1:1, 1:2, 1:5 and 2:1 oleic acid/iso-amyl alcohol under the following conditions: 40 °C temperature applying 0.5 %

> Time [h] 01234567

Fig. 3. Influence of the acid/alcohol molar ratio on the synthesis of isoamyl-oleate.

mol/l; i-amyl-alcohol, 3.78 mol/l; Novozym 435, 0.5 %; speed, 150 rpm.

1:5 1:2 1:1 2:1 30 °C 40 °C 50 °C 60 °C 119

Ester yield [%]

**2.2.3 Effect of the molar ratio** 

0

enzyme concentration, 0.5 % initial water content.

Ester yield [%]

0

Novozym 435, 0.5 %; temperature, 40 °C; speed, 150 rpm.

20

40

60

80

100

20

40

60

80

100

Optimal water content is not only important to preserve the catalytic activity of an enzyme, but also to achieve high reaction rates and yields, and stability of the enzyme. Water requirements for enzymes in organic media vary greatly; therefore each enzyme must be examined at various levels of hydration.

Fig. 1. Influence of water content on the synthesis of isoamyl oleate. Oleic acid, 1.84 mol/l; iamyl-alcohol, 3.78 mol/l; temperature, 40 °C; Novozym 435, 0.5 %; speed, 150 rpm

The role of the water content in the reaction mixture was studied in the range of 0.1 – 1.5 %. As it can be seen in Figure 1 the ester yield as a function of the initial water content has a minimum. In very low water content (0.1 %) the amount of ester produced is small: in this case water present in the reaction mixture is not enough for the enzyme's optimal work. Increasing the water content the yield is growing, and a maximal value is reached at about 0.5 % initial water content. Then at higher and higher water contents the ester yields obtained are gradually decreased, which can be explained by the fact that at high water content the opposite reaction, the hydrolysis is favoured. Based on the results of these measurements, 0.5 % initial water content was used in the further experiments.

#### **2.2.2 Effect of temperature**

The effect of temperature was studied in a series of experiment in the range of 30 - 60 °C under the same conditions (initial water content, molar ratio, enzyme concentration). No thermal deactivation was observed up to 60 °C, and – as it is shown in Figure 2 – the concentration of the oleates produces increased in higher temperatures. This curve is typical of enzyme with high thermostability and which thermal denaturation, during the time of the assay, is negligible. After 12 hours reaction time (not shown) 92 % conversion was obtained in each case.

The activation energy of the enzymatic reaction was determined based on the well-known Arrhenius-equation. The logarithm of the reaction rate data were plotted as a function of reciprocal temperature and the activation energy was calculated from the slope of the regression line. It's value was 16.2 kJ/mol, which is similar to the value reported by (Garcia, et al., 1999) (11.7 kJ/mol), determined for the esterification of isopropyl alcohol and palmitic acid in solvent-free system by the same Novozyme 435 lipase preparation.

Fig. 2. Influence of reaction temperature on the synthesis of isoamyl-oleate. Oleic acid, 1.84 mol/l; i-amyl-alcohol, 3.78 mol/l; Novozym 435, 0.5 %; speed, 150 rpm.

#### **2.2.3 Effect of the molar ratio**

124 Food Industrial Processes – Methods and Equipment

Optimal water content is not only important to preserve the catalytic activity of an enzyme, but also to achieve high reaction rates and yields, and stability of the enzyme. Water requirements for enzymes in organic media vary greatly; therefore each enzyme must be

> 0,5% 1,5 % 0,1 % 0,3 %

Time [h] 01234567

Fig. 1. Influence of water content on the synthesis of isoamyl oleate. Oleic acid, 1.84 mol/l; i-

The role of the water content in the reaction mixture was studied in the range of 0.1 – 1.5 %. As it can be seen in Figure 1 the ester yield as a function of the initial water content has a minimum. In very low water content (0.1 %) the amount of ester produced is small: in this case water present in the reaction mixture is not enough for the enzyme's optimal work. Increasing the water content the yield is growing, and a maximal value is reached at about 0.5 % initial water content. Then at higher and higher water contents the ester yields obtained are gradually decreased, which can be explained by the fact that at high water content the opposite reaction, the hydrolysis is favoured. Based on the results of these

The effect of temperature was studied in a series of experiment in the range of 30 - 60 °C under the same conditions (initial water content, molar ratio, enzyme concentration). No thermal deactivation was observed up to 60 °C, and – as it is shown in Figure 2 – the concentration of the oleates produces increased in higher temperatures. This curve is typical of enzyme with high thermostability and which thermal denaturation, during the time of the assay, is negligible. After 12 hours reaction time (not shown) 92 % conversion was

The activation energy of the enzymatic reaction was determined based on the well-known Arrhenius-equation. The logarithm of the reaction rate data were plotted as a function of reciprocal temperature and the activation energy was calculated from the slope of the regression line. It's value was 16.2 kJ/mol, which is similar to the value reported by (Garcia, et al., 1999) (11.7 kJ/mol), determined for the esterification of isopropyl alcohol and palmitic

amyl-alcohol, 3.78 mol/l; temperature, 40 °C; Novozym 435, 0.5 %; speed, 150 rpm

measurements, 0.5 % initial water content was used in the further experiments.

acid in solvent-free system by the same Novozyme 435 lipase preparation.

examined at various levels of hydration.

Ester yield [%]

**2.2.2 Effect of temperature** 

obtained in each case.

0

20

40

60

80

100

It is well-known, that acid/alcohol molar ratio is one of the most important parameters in enzymatic esterifications. Since the reaction is reversible, an increase in the amount of one of the reactants will result higher ester yields and as expected, this will shifts the chemical equilibrium towards the product side. One way of shifting the reaction toward the synthesis is to increase the alcohol concentration. However, high alcohol concentration may slow down the reaction rates due to inhibition. Therefore, it is necessary to optimize the actual excess nucleophile concentration to be employed in a given reaction. In order to determine the effect of molar ratio, oleic acid was esterified at molar ratios of 1:1, 1:2, 1:5 and 2:1 oleic acid/iso-amyl alcohol under the following conditions: 40 °C temperature applying 0.5 % enzyme concentration, 0.5 % initial water content.

Fig. 3. Influence of the acid/alcohol molar ratio on the synthesis of isoamyl-oleate. Novozym 435, 0.5 %; temperature, 40 °C; speed, 150 rpm.

integrated system was designed for the enzymatic reaction, where a thermostated bioreactor is combined with a pervaporation unit The membrane unit consists of two stainless steel plate, the membrane is located between them. The permeate passed through the membrane was condensed in traps cooled by dry ice – acetone mixture. The permeate side was kept under 0.8 kPa vacuum pressure. During the procedure the feed stream was recirculated in liquid phase on the primary side of the membrane module, while the permeate was obtained in vapour phase. The reaction mixture was recirculated through a glass filter to keep the immobilised enzyme in the reactor. In the integrated system the bioreactor was thermostated at 60 °C temperature, and the experiments were carried out under the optimal conditions determined earlier. Samples were taken from the reactor and the analysis has shown that the water content was managed to keep at 0.5 %. In the integrated system 99.8 %

121

Firstly hydrophilic pervaporation membranes were tested. It was turned out that the CMC-VP-43 type membrane has the highest flux (0.05 kg/m2h) and selectivity (58.94), so this

The tribological properties of the biolubricant manufactured were tested according to the IP and ASTM standards. The chemical, physical parameters of the product were given in Table 1. The thermal and oxidation stabilities were measured by the modofied IP 8.6 method using an oxidative and thermal treating. The results obtained were compared to a commercial,

synthetic, DB-32 dicarboxylic acid ester used as a reference lubricating oil (Table 2).

**Properties Biolubricant DB-32** 

Density g/cm3 0.864 0.925

As it can be seen from Table 1, the bioester produced is considered as a low flashpoint lubricant. Due to its low pour point, it can be applied also at low temperatures. Its viscosity

ASTM colour – 0 0 Flash point (Cleveland) °C 211 204 Pour point °C −27 n.d. Viscosity at 100 °C (KV100 °C) mm2/s 2.28 2.7 Viscosity at 40 °C (KV40 °C) mm2/s 6.39 9.03 Viscosity index (VIE) – 207 148 Acid value mg KOH/g 0.01 0.1 Iodine–bromine value g I/100 g 68.4 n.d.

Table 1. Properties of the biolubricant compared to the reference oil

conversion of the oleic acid was achieved.

n.d.: no data.

membrane was used for the further experiments.

**2.2.6 Tribological properties of the biolubricant** 

As expected, a higher ester conversion was obtained in a shorter period of time for 1:2 and 1:5 molar ratios compared to 1:1. The highest acid conversion was achieved in case of 1:2 molar ratio, as it can be seen from Figure 3, thus alcohol excess was used in the further experiments.

#### **2.2.4 The effect of chain length of alcohols in fusel oil on the esterification reaction**

The main alcohol compounds of fusel oil, two with linear chains (ethanol, propanol) and three with a branch chains (isopropyl alcohol, isobutyl alcohol and isoamyl alcohol) were applied in the next serial of experiments to determine the effect of chain length. In the measurements the conditions were as follows: 1:2 acid/alcohol molar ratio, 0.5 % enzyme and 60 °C temperature.

The results shown in Figure 4, that the esterification rates of oleic acid with ethanol are smaller than with isoamyl alcohol, indicating some effect of the length and structure of the alcohol molecule. The difference between the isoamyl oleate and ethyl oleate was 33 %. Based on this result, it was assumed that esterification of oleic acid with model solution of fusel oil will be similar to the ester synthesis with i-amyl alcohol, since it is the main compound of fusel oil.

#### **2.2.5 Removal of excess water produced by pervaporation**

The optimal parameters of the reaction was determined in the shaken flasks experiments: 0.5 % water concentration, 1:2 oleic acid-isoamyl alcohol molar ratio, 60 °C temperature. According to the analysis data without water removal, the reaction reaches equilibrium after 12 hours reaction time.

Fig. 4. Influence of chain length of alcohols. Acid/alcohol molar ratio, 1:2; Novozym 435, 0.5 %; temperature, 60 °C; speed, 150 rpm.

The esterification of oleic acid resulted in 92 % yield, while water produced as excess during the process and it accumulated as a second phase in the bottom of the flask. Since water excess has a strong inhibition effect on the enzyme, stops the further conversion of substrate, therefore it should be removed. Pervaporation seemed a suitable membrane separation technique for continuous water removal, helping to convert all the substrate. Thus an

As expected, a higher ester conversion was obtained in a shorter period of time for 1:2 and 1:5 molar ratios compared to 1:1. The highest acid conversion was achieved in case of 1:2 molar ratio, as it can be seen from Figure 3, thus alcohol excess was used in the further

**2.2.4 The effect of chain length of alcohols in fusel oil on the esterification reaction**  The main alcohol compounds of fusel oil, two with linear chains (ethanol, propanol) and three with a branch chains (isopropyl alcohol, isobutyl alcohol and isoamyl alcohol) were applied in the next serial of experiments to determine the effect of chain length. In the measurements the conditions were as follows: 1:2 acid/alcohol molar ratio, 0.5 % enzyme

The results shown in Figure 4, that the esterification rates of oleic acid with ethanol are smaller than with isoamyl alcohol, indicating some effect of the length and structure of the alcohol molecule. The difference between the isoamyl oleate and ethyl oleate was 33 %. Based on this result, it was assumed that esterification of oleic acid with model solution of fusel oil will be similar to the ester synthesis with i-amyl alcohol, since it is the main

The optimal parameters of the reaction was determined in the shaken flasks experiments: 0.5 % water concentration, 1:2 oleic acid-isoamyl alcohol molar ratio, 60 °C temperature. According to the analysis data without water removal, the reaction reaches equilibrium after

> Time [h] 01234567

Fig. 4. Influence of chain length of alcohols. Acid/alcohol molar ratio, 1:2; Novozym 435, 0.5

The esterification of oleic acid resulted in 92 % yield, while water produced as excess during the process and it accumulated as a second phase in the bottom of the flask. Since water excess has a strong inhibition effect on the enzyme, stops the further conversion of substrate, therefore it should be removed. Pervaporation seemed a suitable membrane separation technique for continuous water removal, helping to convert all the substrate. Thus an

ethyl-alcohol n-propanol i-butanol i-amyl-alcohol

**2.2.5 Removal of excess water produced by pervaporation** 

experiments.

and 60 °C temperature.

compound of fusel oil.

12 hours reaction time.

Ester yield [%]

%; temperature, 60 °C; speed, 150 rpm.

0

20

40

60

80

100

integrated system was designed for the enzymatic reaction, where a thermostated bioreactor is combined with a pervaporation unit The membrane unit consists of two stainless steel plate, the membrane is located between them. The permeate passed through the membrane was condensed in traps cooled by dry ice – acetone mixture. The permeate side was kept under 0.8 kPa vacuum pressure. During the procedure the feed stream was recirculated in liquid phase on the primary side of the membrane module, while the permeate was obtained in vapour phase. The reaction mixture was recirculated through a glass filter to keep the immobilised enzyme in the reactor. In the integrated system the bioreactor was thermostated at 60 °C temperature, and the experiments were carried out under the optimal conditions determined earlier. Samples were taken from the reactor and the analysis has shown that the water content was managed to keep at 0.5 %. In the integrated system 99.8 % conversion of the oleic acid was achieved.

Firstly hydrophilic pervaporation membranes were tested. It was turned out that the CMC-VP-43 type membrane has the highest flux (0.05 kg/m2h) and selectivity (58.94), so this membrane was used for the further experiments.

#### **2.2.6 Tribological properties of the biolubricant**

The tribological properties of the biolubricant manufactured were tested according to the IP and ASTM standards. The chemical, physical parameters of the product were given in Table 1. The thermal and oxidation stabilities were measured by the modofied IP 8.6 method using an oxidative and thermal treating. The results obtained were compared to a commercial, synthetic, DB-32 dicarboxylic acid ester used as a reference lubricating oil (Table 2).


n.d.: no data.

Table 1. Properties of the biolubricant compared to the reference oil

As it can be seen from Table 1, the bioester produced is considered as a low flashpoint lubricant. Due to its low pour point, it can be applied also at low temperatures. Its viscosity

using lipase from *Aspergillus tereus*. They found that in n-hexane solvent the alcohols were able to react with the fatty acids (miristic acid, palmitic acid, stearic acid), except oleic acid. Using other lipase preparations (*Candida antarctica, Candida rugosa, Rhizomucor miehei, porcine pancreas*), however, made it possible the successful oleic acid esterification with similar low molecular weight alcohols. In Table 3 results published on the kinetics of enzymatic esterification of oleic acid with short chain alcohols in organic solvents are

Garcia et al. (1996) studied the kinetics of i-propyl-oleate formation by *Candida antarctica* lipase. The model used was an ordered bi-bi type containing 13 kinetic parameters. Thus the model seems too complicated having high uncertainty. Esterification of butyl alcohol by *Candida rugosa* lipase was studied by (Zaidi et al., 2002), where ping-pong bi-bi mechanism was assumed in the kinetical model with 5 parameters. However, the range of substrate concentration measured was quite narrow (0.1 – 1 mol/L), and the error of the modelling

Alcohol Enzyme Model Inhibition Reference Note

both S and P competitive

both acid and alcohol (Garcia et al., 1996)

(Zaidi et al., 2002)

al., 2001)

al.,2000 )

al., 2002)

alcohol (Oliveira et

alcohol (Goddard et

alcohol (Hazirka et

too many parameters fitted to one measurement; high error of fitting

123

error is 28 %; narrow range of substrate concentration (0.1-1 mol/L)

Difference between the parameters is 15 order of magnitude; quasi one-substrate kinetics

Different kinetics for each alcohol concentration; pseudo onesubstrate kinetics

Narrow range of substrate concentration (0.3- 0.8 mol/L)

ordered bi-bi 13 parameters

ping-pong bi-bi 5 parameters

random bi-bi 4 parameters

Michaelis-Menten 2+1 parameters

ping-pong bibi 4 parameters

Immobilized *Rhizomucor miehei* lipase was applied for ethyl oleate synthesis by (Oliveira et al., 2001). In order to describe the kinetics, random bi-bio model was used, which contained 4 parameters. However, the difference between the parameter values determined was 15 order of magnitude, implying that the effect of certain parameter is nearly negligible comparing to the others. Esterification of ethyl alcohol and oleic acid by immobilized

summarized.

i-propyl alcohol

butyl alcohol

ethyl alcohol

ethyl alcohol

ethyl alcohol

was found very high (28 %).

Novozym 435 *Candida antarctica* 

immobilised *Candida rugosa*

immobilised *Rhizomucor miehei* 

immobilised *Rhizomucor mihei*

soluble, from porcine pancrease

Table 3. Kinetic studies on esterification of oleic acid

index is quite high, thus the viscosity of the ester is not highly influenced by temperature. The good conversion of oleic acid resulted in the low acid number, while the iodin-bromine number refers to the quite low unsaturation.

The data in Table 2 show that the stability of the measured ester based on thermal and oxidative treatment is higher than the reference oil, i.e. the increase in viscosity and acid number is lower, the change of its colour is acceptable. According to these data the biolubricant is suggested to use mainly at the high speed and low load regime of the tribological circumstances. In the mechanical industry it can be applied e.g. as a cooling lubricant compound for metalworking processes, moreover in particular lubrication processes where lubricant loss may occur, e.g. mist lubrication, chain lubrication, launch engine lubrication.


Table 2. The effect of thermal and oxidative treatments

#### **2.2.7 Study on toxicity of biolubricant by zebrafishes**

To study the effect of the particular biolubricant on the water environment, its acute toxicity was assessed with Acute Fish Toxicity Test on Zebrafish (*Brachydanio rerio*), over an exposure period of 96 hours in a static system. A limit test was conducted according to the OECD Guidelines for the Testing of Chemicals, Procedure No. 203 (1992). In this kind of test the limit in LC50 number is 100 mg/l. In the experiment two 20 l aquaria were used, in the control there was no biolubricant, while in the other aquarium its nominal concentration was 100 mg/l. Zebrafishes (7 fishes in both aquaria) were used in the test, which are extremely sensitive to the waste compounds in living waters. Analytical methods were applied to measure the parameters of the water (chemical properties, temperature, oxigen saturation, pH…etc.). The observation of fishes were carried out at regular intervals (3, 6, 24, 48, 72 and 96 hours) and mortality was determined.

As a result of the experiment no signs of reaction or mortality were detected. Thus the 24 h, 48 h, 72 h and 96 h LC50 number of the particular biolubricant is > 100 mg/l. So it can be concluded that the biolubricant is not toxic for the living water.

### **3. Kinetics approach**

Before developing a complete method for enzymatic manufacture of this biolubricant, a detailed kinetic analysis should be carried out on the reaction mechanism. Enzymatic esterification of fatty acids and ingredients of fusel oil was studied by (Gulati et al., 2003)

index is quite high, thus the viscosity of the ester is not highly influenced by temperature. The good conversion of oleic acid resulted in the low acid number, while the iodin-bromine

The data in Table 2 show that the stability of the measured ester based on thermal and oxidative treatment is higher than the reference oil, i.e. the increase in viscosity and acid number is lower, the change of its colour is acceptable. According to these data the biolubricant is suggested to use mainly at the high speed and low load regime of the tribological circumstances. In the mechanical industry it can be applied e.g. as a cooling lubricant compound for metalworking processes, moreover in particular lubrication processes where lubricant loss may occur, e.g. mist lubrication, chain lubrication, launch

> **Oxidative treatment (200 °C, 12 h, airflow 20 dm3/h)**

> > **VIE Colour Acid**

**value (mg KOH/g)** 

**VK40 °C (mm2/s)**

number refers to the quite low unsaturation.

**Lubricant Thermal treatment** 

**VK40 °C (mm2/s)**

Table 2. The effect of thermal and oxidative treatments

**2.2.7 Study on toxicity of biolubricant by zebrafishes** 

48, 72 and 96 hours) and mortality was determined.

**3. Kinetics approach** 

concluded that the biolubricant is not toxic for the living water.

**VK100 °C (mm2/s)** **(200 °C, 12 h)** 

**VIE Colour Acid** 

**value (mg KOH/g)**

DB-32 2.75 9.45 140 7 5.5 5.79 29.47 143 8 50.8 Biolubricant 2.29 6.41 209 0 0.1 5.57 25.91 162 8 10.6

To study the effect of the particular biolubricant on the water environment, its acute toxicity was assessed with Acute Fish Toxicity Test on Zebrafish (*Brachydanio rerio*), over an exposure period of 96 hours in a static system. A limit test was conducted according to the OECD Guidelines for the Testing of Chemicals, Procedure No. 203 (1992). In this kind of test the limit in LC50 number is 100 mg/l. In the experiment two 20 l aquaria were used, in the control there was no biolubricant, while in the other aquarium its nominal concentration was 100 mg/l. Zebrafishes (7 fishes in both aquaria) were used in the test, which are extremely sensitive to the waste compounds in living waters. Analytical methods were applied to measure the parameters of the water (chemical properties, temperature, oxigen saturation, pH…etc.). The observation of fishes were carried out at regular intervals (3, 6, 24,

As a result of the experiment no signs of reaction or mortality were detected. Thus the 24 h, 48 h, 72 h and 96 h LC50 number of the particular biolubricant is > 100 mg/l. So it can be

Before developing a complete method for enzymatic manufacture of this biolubricant, a detailed kinetic analysis should be carried out on the reaction mechanism. Enzymatic esterification of fatty acids and ingredients of fusel oil was studied by (Gulati et al., 2003)

**VK100 °C (mm2/s)**

engine lubrication.

using lipase from *Aspergillus tereus*. They found that in n-hexane solvent the alcohols were able to react with the fatty acids (miristic acid, palmitic acid, stearic acid), except oleic acid. Using other lipase preparations (*Candida antarctica, Candida rugosa, Rhizomucor miehei, porcine pancreas*), however, made it possible the successful oleic acid esterification with similar low molecular weight alcohols. In Table 3 results published on the kinetics of enzymatic esterification of oleic acid with short chain alcohols in organic solvents are summarized.

Garcia et al. (1996) studied the kinetics of i-propyl-oleate formation by *Candida antarctica* lipase. The model used was an ordered bi-bi type containing 13 kinetic parameters. Thus the model seems too complicated having high uncertainty. Esterification of butyl alcohol by *Candida rugosa* lipase was studied by (Zaidi et al., 2002), where ping-pong bi-bi mechanism was assumed in the kinetical model with 5 parameters. However, the range of substrate concentration measured was quite narrow (0.1 – 1 mol/L), and the error of the modelling was found very high (28 %).


Table 3. Kinetic studies on esterification of oleic acid

Immobilized *Rhizomucor miehei* lipase was applied for ethyl oleate synthesis by (Oliveira et al., 2001). In order to describe the kinetics, random bi-bio model was used, which contained 4 parameters. However, the difference between the parameter values determined was 15 order of magnitude, implying that the effect of certain parameter is nearly negligible comparing to the others. Esterification of ethyl alcohol and oleic acid by immobilized

the concentration of substrate A (first), the concentration of substrate B (second), enzyme concentration, the amount of products formed altogether. Initial reaction rate can be described by including only the first three factors, taken into

account the fact that product concentration is 0 at the beginning of the reaction:

*v*

'

be used:

where

parameter of the reaction.

*v*

73 73 *kk*

*v*

parameter equation can be obtained:

*kk Vm* )(

1

[ ] []

*A B*

[ ] []

*m A B iB*

*K BK AK B*

*V*

This model contains four parameters, which can be further supplemented by the product inhibition factors. In this way two more parameters are added into the equation (Eq. 5). In the equation K'AB parameter is the rate of the apparent product formation, the ping-pong

> \*[ ] [ ] [ ] 1 1\* <sup>1</sup> [ ] [ ]\*[ ] [ ]

> > )(

*<sup>k</sup> KiB* )(

In our case, water is one of the products in esterification. But it is not only a product, a small amount of water should be present initially in the reaction mixture to keep in an active formation of the enzyme structure. In the beginning of the reaction, however, the initial concentration of water does not change significantly, therefore water content (P) can be considered as constant (its effect is negligible). Thus the equation can be simplified and a 5-

[ ] 1 1

 

[ ] [ ]\*[ ] [ ]

*K K BK A AB K B*

*iB*

*m A AB B*

*V*

731 327 *kkk kkk KA*

> 6 5 *k*

 

*V*

*m A AB B*

*KK P B K P A AB K B K*

)(

)( '

7351 7642 *kkkk kkkk <sup>K</sup> AB* 

*iB iP*

)( 735 763 *kkk kkk KB* 

(3)

125

(4)

4 3 *k <sup>k</sup> KiP*

(6)

(5)

*m A B*

*V <sup>v</sup> K K*

If substrate inhibition is considered, as well, the three-parameter equation should be completed with another constant (Dezbaradica et al., 2006) and the following equation can

[ ] 1 1

 

*Rhizomucor miehei* lipase was studied by (Goddard et al., 2000) as well. Michaelis-Menten model was used for the description of the reaction, however, different kinetics was used in each alcohol concentrations, which is considered as a pseudo-one substrate model.

Soluble porcine pancreatic lipase was applied for the ethyl oleate synthesis by (Hazarika et al., 2002). They assumed ping-pong mechanism, as well, containing 4 parameters, however the range of substrate concentration studied was even narrower (0.3 – 0.8 mol/L) than in case of (Zaidi et al., 2002). Description of the correct kinetics on the particular esterification reaction is even more difficult due to the various possible inhibition effects. As it is shown in Table 3, ethyl alcohol as a substrate was in all cases considered as inhibitor, while in the esterifications with other alcohols, both substrates were regarded as inhibitors.

As a summary, it seems from Table 3 that the kinetic models/parameters published so far can not be considered as a proper, detailed kinetic description of the enzymatic process for oleic acid esterification with short chain alcohols, moreover there has been no data found on esterification with i-amyl alcohol. Therefore our aim was to elaborate a proper, sophisticated model for this particular reaction.

#### **3.1 Kinetics model**

Kinetics of enzymatic reactions can be described by the well-known Michaelis-Menten model. For reactions having 2 substrates and 2 products (bi-bi reactions), its application is quite complicated since various mechanisms can be considered according to the order and rate both substrates' binding and products' releasing to/of the enzyme active sites (random, ordered, ping-pong…etc.). Since majority of the kinetical studies suggested ping-pong bi-bi mechanism for the enzymatic esterification of oleic acid and short chain alcohols, we also considered it as an initial point for the description. According to Cleland (Cleland, 1979) the ping-pong bi-bi mechanism can be outlined as follows:

$$A + E \xleftarrow{k\_1} \xleftarrow{k\_1} \left(\frac{EA}{FP}\right) \xleftarrow{k\_3} P + F + B \xleftarrow{k\_5} \frac{k\_5}{k\_6} \left(\frac{FB}{EQ}\right) \xleftarrow{k\_7} Q + E \tag{1}$$

In the first step the enzyme reacts with substrate A forming an AE enzyme-substrate complex, which is transformed into FP modified complex – by an internal rearrangement. Product P comes off the complex, then the modified enzyme molecule F is able to react with substrate B, forming a new enzyme-substrate complex FB. It is transformed into an enzymeproduct complex EQ, then Q moves to the bulk solution. Finally, enzyme E becomes free and can react with another substrate A molecule. In the equation the reaction rate constants are shown (k1 – k8), among them k1, k3, k5 and k7 belong to the towards direction of the reaction, while the others to the backwards direction (with negative sign). Summarising the steps, the formation rate of product can be written as follows:

$$\nu = \frac{(k\_1 k\_2 k\_3 k\_4 AB - k\_2 k\_4 k\_4 k\_8 PQ)E\_0}{E + (EA + FP) + F + (FB + EQ)}\tag{2}$$

where E0 is the initial enzyme concentration and E is the actual enzyme concentration (the other capital letters mean concentration of the particular compound). This model, however, is too complicated to apply it in practice. To simplify the situation, the main parameters influencing the rate of product formation are selected, as follows (Janssen et al., 1996):

the concentration of substrate A (first), the concentration of substrate B (second), enzyme concentration, the amount of products formed altogether.

Initial reaction rate can be described by including only the first three factors, taken into account the fact that product concentration is 0 at the beginning of the reaction:

$$\upsilon = \frac{V\_m}{1 + \frac{K\_A}{[A]} + \frac{K\_B}{[B]}} \tag{3}$$

If substrate inhibition is considered, as well, the three-parameter equation should be completed with another constant (Dezbaradica et al., 2006) and the following equation can be used:

$$w = \frac{V\_m}{1 + \frac{K\_A}{[A]} \left(1 + \frac{[B]}{K\_{iB}}\right) + \frac{K\_B}{[B]}}\tag{4}$$

This model contains four parameters, which can be further supplemented by the product inhibition factors. In this way two more parameters are added into the equation (Eq. 5). In the equation K'AB parameter is the rate of the apparent product formation, the ping-pong parameter of the reaction.

$$\upsilon = \frac{V\_m}{1 + \left(\frac{K\_A}{[A]} + \frac{K\_{AB}^{'} \ast ^\ast [P]}{[A] \ast ^\ast [B]}\right) \left(1 + \frac{[B]}{K\_{iB}}\right) + \frac{K\_B}{[B]} \ast \left(1 + \frac{[P]}{K\_{iP}}\right)}\tag{5}$$

where

130 Food Industrial Processes – Methods and Equipment

*Rhizomucor miehei* lipase was studied by (Goddard et al., 2000) as well. Michaelis-Menten model was used for the description of the reaction, however, different kinetics was used in

Soluble porcine pancreatic lipase was applied for the ethyl oleate synthesis by (Hazarika et al., 2002). They assumed ping-pong mechanism, as well, containing 4 parameters, however the range of substrate concentration studied was even narrower (0.3 – 0.8 mol/L) than in case of (Zaidi et al., 2002). Description of the correct kinetics on the particular esterification reaction is even more difficult due to the various possible inhibition effects. As it is shown in Table 3, ethyl alcohol as a substrate was in all cases considered as inhibitor, while in the

As a summary, it seems from Table 3 that the kinetic models/parameters published so far can not be considered as a proper, detailed kinetic description of the enzymatic process for oleic acid esterification with short chain alcohols, moreover there has been no data found on esterification with i-amyl alcohol. Therefore our aim was to elaborate a proper, sophisticated

Kinetics of enzymatic reactions can be described by the well-known Michaelis-Menten model. For reactions having 2 substrates and 2 products (bi-bi reactions), its application is quite complicated since various mechanisms can be considered according to the order and rate both substrates' binding and products' releasing to/of the enzyme active sites (random, ordered, ping-pong…etc.). Since majority of the kinetical studies suggested ping-pong bi-bi mechanism for the enzymatic esterification of oleic acid and short chain alcohols, we also considered it as an initial point for the description. According to Cleland (Cleland, 1979) the

> 13 57 24 6 8 *kk kk EA FB A E PFB Q E kk kk FP EQ*

In the first step the enzyme reacts with substrate A forming an AE enzyme-substrate complex, which is transformed into FP modified complex – by an internal rearrangement. Product P comes off the complex, then the modified enzyme molecule F is able to react with substrate B, forming a new enzyme-substrate complex FB. It is transformed into an enzymeproduct complex EQ, then Q moves to the bulk solution. Finally, enzyme E becomes free and can react with another substrate A molecule. In the equation the reaction rate constants are shown (k1 – k8), among them k1, k3, k5 and k7 belong to the towards direction of the reaction, while the others to the backwards direction (with negative sign). Summarising the

> ( ) 1257 2468 0 ( )( ) *k k k k AB k k k k PQ E E EA FP F FB EQ*

where E0 is the initial enzyme concentration and E is the actual enzyme concentration (the other capital letters mean concentration of the particular compound). This model, however, is too complicated to apply it in practice. To simplify the situation, the main parameters influencing the rate of product formation are selected, as follows (Janssen et al., 1996):

(1)

(2)

each alcohol concentrations, which is considered as a pseudo-one substrate model.

esterifications with other alcohols, both substrates were regarded as inhibitors.

model for this particular reaction.

ping-pong bi-bi mechanism can be outlined as follows:

steps, the formation rate of product can be written as follows:

**3.1 Kinetics model** 

$$\begin{aligned} V\_m &= \frac{k\_3 k\_\gamma}{k\_3 + k\_\gamma} \quad K\_A = \frac{k\_\gamma (k\_2 + k\_3)}{k\_1 (k\_3 + k\_\gamma)} \quad K\_B = \frac{k\_3 (k\_6 + k\_\gamma)}{k\_5 (k\_3 + k\_\gamma)} \quad K\_{\mu} = \frac{k\_3}{k\_4} \\\ K\_{\iota B} &= \frac{k\_\iota}{k\_6} \quad K'\_{\iota B} = \frac{k\_2 k\_4 (k\_6 + k\_\gamma)}{k\_1 k\_5 (k\_3 + k\_\gamma)} \end{aligned}$$

In our case, water is one of the products in esterification. But it is not only a product, a small amount of water should be present initially in the reaction mixture to keep in an active formation of the enzyme structure. In the beginning of the reaction, however, the initial concentration of water does not change significantly, therefore water content (P) can be considered as constant (its effect is negligible). Thus the equation can be simplified and a 5 parameter equation can be obtained:

$$\upsilon = \frac{V\_m}{1 + \left(\frac{K\_A}{[A]} + \frac{K\_{AB}}{[A]^\ast [B]}\right) \left(1 + \frac{[B]}{K\_{iB}}\right) + \frac{K\_B}{[B]}} \tag{6}$$

comparison of the two relevant relaxation times. The ratio of the relaxation time of biocatalysis rate, tr and that of the diffusion rate, td shows which process should be

and 2 ( ) *<sup>d</sup>*

Oleic acid – having slower diffusivity – was chosen for the calculations and the highest reaction rate-substrate concentration value-pair was taken from Table 4. Thus tr was

> ( ) *<sup>r</sup> <sup>C</sup> t s*

Diffusion constant (D) of oleic acid in n-heptane was determined according to Shibel (Perry,

*<sup>T</sup> D k* 

Vc of oleic acid is 1152 cm3/mol, thus Vs is obtained as 460 cm3/g. In this way D diffusion

The mass transfer coefficient can be calculated – based on the Sherwood number – from the

The average diameter of Novozyme 435 immobilised lipase preparation is 0.06 cm, thus the

2 42 1,61.10 / 55,9 ( ) (5,3 \* 10 / ) *<sup>d</sup>*

Comparing the values of tr and td it can be concluded that diffusion rate is three order of magnitude higher than the reaction rate, thus the rates measured in the enzymatic process

In Fig. 5a the initial reaction rates are presented as a function of oleic acid (substrate 1) concentration, while in Fig. 5b the same data are shown as a function of i-amyl alcohol (substrate 2) concentration. It can be clearly seen that the i-amyl alcohol has a considerably

 

*D cm s t s k cm s*

5 2

0 5 0 2

> 1 <sup>3</sup> *b s*

*sl <sup>D</sup> <sup>t</sup>*

*<sup>k</sup>* (7)

127

*r C* (8)

*<sup>V</sup>* (9)

1,048 *V V S C* 0,285 (10)

*SL* 2 / *<sup>k</sup> k Dd* (11)

(12)

0 ( )0

*r <sup>C</sup> <sup>t</sup> r C*

VS molar volume density was estimated from its critical volume (Vc):

considered as the limiting step.

calculate as:

1969):

The relaxation times can be defined as follows :

coefficient is calculated as 1.61x 10-5 cm2/s.

value of the mass transfer coefficient is 5.3x 10-4 cm/s. The relaxation time for the diffusion is calculated as:

*SL*

and oleic acid a slight inhibition effect on the enzymatic reaction.

diffusion coefficient and the particle size:

can be considered as the real reaction rates.

**3.4 Kinetical analysis** 

It can be seen that the difference between the 4- and 5-parameter equations (Eqs. 4 and 6) is the KAB factor. Its influence is significant only in the case when concentrations of both substrates are very low. However, we do not plan to carry out measurements under these conditions, thus it is assumed that no significant difference will be experienced in the modelling results obtained by using the two systems.
