Scholarly article on topic 'Optimization of the methanolysis of lard oil in the production of biodiesel with response surface methodology'

Optimization of the methanolysis of lard oil in the production of biodiesel with response surface methodology Academic research paper on "Chemical sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Egyptian Journal of Petroleum
OECD Field of science
Keywords
{Biodiesel / Methanolysis / "Lard oil" / "Central composite design" / Optimization / "Analysis of variance"}

Abstract of research paper on Chemical sciences, author of scientific article — Chinyere B. Ezekannagha, Callistus N. Ude, Okechukwu D. Onukwuli

Abstract Methanolysis of lard oil to biodiesel was optimized using central composite design (CCD) of response surface methodology to delineate the effects of five levels, four factorson the yield of biodiesel. A total of 30 individual experiments were conducted and designed to study these process variables. A statistical model predicted that the highest conversion yield of lard biodiesel would be 96.2% at the following optimized reaction conditions: reaction temperature of 65°C, catalyst amount of 1.25%, time of 40min, methanol to oil molar ratio of 6:1 at 250rpm. Experiments performed at the predicted optimum conditions yielded 96% which was in good agreement with the predicted value. This study shows that lard oil as a low cost feedstock is a good source of raw material for biodiesel production and a sustainable biodiesel production could be achieved with proper optimization of the process variables.

Academic research paper on topic "Optimization of the methanolysis of lard oil in the production of biodiesel with response surface methodology"

Egyptian Journal of Petroleum xxx (2017) xxx-xxx

HOSTED BY

Contents lists available at ScienceDirect

Egyptian Journal of Petroleum

journal homepage: www.sciencedirect.com

Full Length Article

Optimization of the methanolysis of lard oil in the production of biodiesel with response surface methodology

Chinyere B. Ezekannagha a, Callistus N. Ude b, Okechukwu D. Onukwulia

a Department of Chemical Engineering, NnamdiAzikiwe University, Awka P.M.B. 5025, Anambra State, Nigeria b Projects Development Institute (PRODA), P.M.B. 01609 Emene-Enugu, Nigeria

ARTICLE INFO ABSTRACT

Methanolysis of lard oil to biodiesel was optimized using central composite design (CCD) of response surface methodology to delineate the effects of five levels, four factorson the yield of biodiesel. A total of 30 individual experiments were conducted and designed to study these process variables. A statistical model predicted that the highest conversion yield of lard biodiesel would be 96.2% at the following optimized reaction conditions: reaction temperature of 65 °C, catalyst amount of 1.25%, time of 40 min, methanol to oil molar ratio of 6:1 at 250 rpm. Experiments performed at the predicted optimum conditions yielded 96% which was in good agreement with the predicted value. This study shows that lard oil as a low cost feedstock is a good source of raw material for biodiesel production and a sustainable biodiesel production could be achieved with proper optimization of the process variables.

© 2016 Egyptian Petroleum Research Institute. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Article history: Received 27 August 2016 Revised 4 November 2016 Accepted 4 December 2016 Available online xxxx

Keywords: Biodiesel Methanolysis Lard oil

Central composite design Optimization Analysis of variance

1. Introduction

Petroleum derived fuels have been the major source of energy globally. This source is finite and at the current rate of consumption emanating from the global population explosion, it will get depleted in the near future. In addition, the soaring prices due to uncertain political situation in some oil-producing countries, scarcity, and environmental concerns have necessitated the research for an alternative and renewable energy sources.

Our society is highly dependent on petroleum for its activities and Nigeria is not left out. This has led to numerous challenges and untold hardship to the citizens which would have been mitigated if there were sustainable alternative fuels. A viable alternative is biodiesel. An alternative fuel must be technically feasible, economically acceptable and readily available [1].

Biodiesel consists of the simple alkyl esters of fatty acids derived from a renewable lipid feedstock such as vegetable oil or animal fat. It is oxygenated, sulphur free, biodegradable, non-toxic and environmentally friendly alternative automotive fuel [2]. Its use does not require any major modifications in the existing diesel engine. The advantage of this bio-fuel over the conventional diesel fuel also includes high cetane number, higher heating value,

low smoke and particulates, low carbon monoxide and hydrocarbon emissions.

Biodiesel fuels are attracting increasing attention worldwide as a blending component or direct replacement for diesel fuel in vehicle engines. The major constraint in wide spread use of biodiesel is the production cost which includes the costs of raw materials and the process operation. The cost of raw materials represents approximately 60-75% of the total cost of biodiesel production [3-5]. As a future prospective, biodiesel has to compete economically with petroleum diesel fuels. One way of reducing the biodiesel production costs is to use the less expensive/low cost feed stock containing fatty acids such as animal fats, inedible oils, restaurant waste oil, frying oil, products of the refining vegetable oil instead of from the edible vegetable oil which could lead to food crisis [6-8]. These low cost feed stocks are more challenging to process because they contain high amount of free fatty acids (FFA) but could be overcome by improving on the production process by the use of two stage processes (esterification and transesterification)and using the optimum reaction conditions for maximum biodiesel yield.

Several processes have previously been developed for the production of biodiesel via acid, alkali, enzyme catalyzed and non-catalyzed processes. The process of transesterification with an alkali catalyst and short chain alcohol as most often conducted tends to yield the highest productivity in the shortest time. Most of the studies show best properties of biodiesel was obtained by

Peer review under responsibility of Egyptian Petroleum Research Institute. http://dx.doi.org/10.1016/j.ejpe.2016.12.004

1110-0621/® 2016 Egyptian Petroleum Research Institute. Production and hosting by Elsevier B.V.

This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

C.B. Ezekannagha et al./Egyptian Journal of Petroleum xxx (2017) xxx-xxx

using KOH as catalyst. Methanol is commonly and widely used in biodiesel production due to its low cost and availability. Other alcohols such as ethanol, isopropanol, and butanol may also be used. Ethanol is derived from renewable biomass, thus relatively cheap. However, ethanol forms an azeotrope with water so it is expensive to purify the ethanol during recovery. Also, the yield of fatty acid ethyl esters is less compared to methyl esters as well as separation of glycerol is the main constrains in the process of ethanolysis. A key quality factor for the primary alcohol is the water content which interferes with the transesterification reactions and can result in poor yields and high level of soap, free fatty acids (FFA) and triacylglycerols (TAG) in the final fuel [9,10]. Unfortunately, all the lower alcohols are hygroscopic and are capable of absorbing water from the air.

According to Enweremadu et al. [11], transesterification reaction involves some critical parameters which strongly influence the final yield. These parameters are peculiar to transesterification of triglycerides in general. From the review of the transesterifica-tion methods, the most relevant variables are namely; free fatty acid and water content in the oil, reaction temperature, molar ratio of alcohol to oil, type of catalyst, type/chemical structure of alcohol, amount/concentration of catalyst, reaction time, intensity of mixing (rpm), use of co solvents. However, Leung et al. [12] reported that, there are four primary/main factors affecting the yield of biodiesel namely; alcohol/oil quantity, Reaction time, Reaction temperature, and catalyst concentration.

The traditional one factor at a time method of analysis is time consuming and does not take into consideration the interaction effects between the factors hence optimization method with respect to design of experiment such as central composite design of response surface methodology to establish best conditions for biodiesel production, in addition to studying the main and interaction effects of the process factors thereby providing the best approach in establishing a model correlating the response variable and the independent variables. Response surface methodology (RSM) is a widely used statistical tool which has been applied in research into complex variable processes and has great advantage in the optimization of reaction conditions and the study of interactions among experimental variables within the range studied, allowing a better understanding of the process while reducing the experimental time and costs [13]. The central composite design (CCD) of response surface methodology was suitable approach for simultaneous study of effects of process variables on the biodiesel production and has previously been successfully applied in the study and optimization of biodiesel production with rapeseed oil, soybean oil, cotton seed oil, et cetera and in several biotechnolog-ical and chemical processes [14-16].

Aiming to achieve increased knowledge on previously described study specifically on biodiesel production from animal fat as raw material, the present work focused on assessing the effects of five levels, four factors and the correlation among the process variables within the range of temperature (50-70 °C), catalyst concentration (0.5-1.5 wt% of fat), time (20-100 min) and molar ratio of methanol to oil (3:1-15:1) and evaluating optimum parameters for the biodiesel yield by using Response surface methodology (RSM).

2. Materials and methods

2.1. Materials

method by subjecting it to heating in a pan without the presence of water at 110 °C for 1 h (under atmospheric pressure to avoid any degradation) to remove water, the waxy, and other suspended and residual matters. Melted fat was then filtered to remove the insoluble materials (such as meat and bone particles) known as cracklings. The processed pork fat was stored in air tight opaque plastic jars to prevent oxidation. The lard oil was characterized to determine the acid value, specific gravity, viscosity, water content, saponification and iodine values so as to ascertain the appropriate pre-treatment method to be used for the oil before used for the reaction.

2.2. Experimental methods

2.2.1. Transesterification procedure

A batch reactor of 500 ml capacity equipped with a reflux condenser and magnetic stirrer was charged with the desired amount of oil (50 ml) heated to 65 °C in a water bath with agitation. A measured amount of catalyst (potassium hydroxide) was then thoroughly mixed in known quantity of methanol till it dissolved completely to give potassium methoxide. The potassium methox-ide was added to the reactor and the reaction timed immediately after the addition of the potassium methoxide. It was transferred into separating funnel and allowed to settle for an hour. Two distinct layers were observed; a thick brown layer (glycerol) at the bottom and a yellowish colour layer constituting the upper layer (biodiesel). FAME layer (Biodiesel) was washed with distilled water to remove unreacted catalyst, methanol and residual glycerol and heated slightly to remove any residual water in it.

The percentage yield was taken.

% Yield =

Weight of Fatty Acid Methyl Ester Weight of Oil Used :

The transesterification was carried out at optimum rotation speed of 250 rpm based on literature data to achieve maximum conversion. The reaction parameters were chosen as follows: temperature ranged from 50 °C to 70 °C, mass ratio of catalyst to oil from 0.5% to 1.5%, time from 20 min to 100 min, molar ratio of methanol to oil from 3:1 to 15:1. The procedures above were used for each experiment executed at different parameters using the experimental design matrix in Table 2.

2.2.2. Biodiesel characterization

Standard procedures were used to characterize the properties of lard biodiesel; acid value (A.V), viscosity (i), specific gravity (S.G), saponification value (S.V), iodine value (I.V), cetane number (C.N), higher heating value (HHV), flash point, cloud point and pour point of the biodiesel. The determined fuel properties were compared with the ASTM standards for fuel. Most of the properties analyzed determine the efficiency of a fuel for diesel engines. There are other aspects or characteristics which do not have a direct bearing on the performance, but are important for reasons such as environment impact etc.

2.2.3. FTIR spectroscopy analysis

The various functional groups present in the raw oil and biodiesel sample were determined with Fourier Transform infrared (FTIR).

Methanol (CH3OH, 99.8% purity) and potassium hydroxide were bought from Cornraws Company Ltd., Enugu and of analytical grade, unless otherwise stated. Mixed pork lard was obtained from new market in Enugu and was rendered according to the method of Dias et al. [17]. The pork lard was rendered using dry-rendering

2.3. Design of experiment

The experimental design selected for this study is Central Composite Design (CCD) and the response measured which is the dependent variable is the yield of biodiesel.

C.B. Ezekannagha et al./Egyptian Journal of Petroleum xxx (2017) xxx-xxx

Table 1

Independent variables and their levels for central composite design of the lard oil transesterification.

Independent variable

Symbol

Levels

Reaction temperature Catalyst amount Time

Methanol/Oil

Wt% Mins

Xi X2 X3 X4

50 0.5 20 3

60 1.00 60 9

65 1.25 80 12

70 1.5 100 15

Table 2

Central composite design arrangement (experimental design matrix).

Std Run Coded factors X4 Actual factors Time (Mins) Methanol/ Oil (-)

X1 X2 X3 Reaction Temp. (°C) Catalyst Amount (wt%)

1 22 -1 -1 -1 -1 55 0.75 40 6

2 18 1 -1 -1 -1 65 0.75 40 6

3 17 -1 1 -1 -1 55 1.25 40 6

4 11 1 1 -1 -1 65 1.25 40 6

5 23 -1 -1 1 -1 55 0.75 80 6

6 3 1 -1 1 -1 65 0.75 80 6

7 30 -1 1 1 -1 55 1.25 80 6

8 4 1 1 1 -1 65 1.25 80 6

9 28 -1 -1 -1 1 55 0.75 40 12

10 5 1 -1 -1 1 65 0.75 40 12

11 9 -1 1 -1 1 55 1.25 40 12

12 10 1 1 -1 1 65 1.25 40 12

13 25 -1 -1 1 1 55 0.75 80 12

14 20 1 -1 1 1 65 0.75 80 12

15 27 -1 1 1 1 55 1.25 80 12

16 12 1 1 1 1 65 1.25 80 12

17 2 -2 0 0 0 50 1 60 9

18 1 2 0 0 0 70 1 60 9

19 16 0 -2 0 0 60 0.5 60 9

20 7 0 2 0 0 60 1.5 60 9

21 21 0 0 -2 0 60 1 20 9

22 26 0 0 2 0 60 1 100 9

23 14 0 0 0 -2 60 1 60 3

24 19 0 0 0 2 60 1 60 15

25 15 0 0 0 0 60 1 60 9

26 29 0 0 0 0 60 1 60 9

27 13 0 0 0 0 60 1 60 9

28 24 0 0 0 0 60 1 60 9

29 8 0 0 0 0 60 1 60 9

30 6 0 0 0 0 60 1 60 9

In order to optimize the process variables for biodiesel production from lard oil transesterification, examine the combined effect of the four different independent variables; reaction temperature, catalyst amount, reaction time, methanol to oil ratio on yield and derive a model, five levels, four factors central composite factorial design (CCD) which includes 24 = 16 factorial points plus 6 central points and 2 x 4 = 8 star points leading to a total of 30 experiments was adopted in this study. Table 1 shows the independent factors and their corresponding actual values while Table 2 shows the design matrix. The matrix for the four variables was varied at these five levels (-a, -1, 0, +1, and +a). The factor levels of the variables investigated were chosen by considering the preliminary tests on effect of individual variables on biodiesel yield and operating limits of the biodiesel production process conditions.

2.4. Model fitting

Each response variable of the experimental planning (Table 2) was fitted to a second order polynomial equation generated by Design-expert 8 trial version software ©2013by Stat-Ease Inc., USA and presented in Eq. (2) aiming to correlate the response variable with the independent variables. The quadratic response

surface model presented in Eq. (2), which considered the linear, quadratic and the interaction effects of the variables, was used

for the preliminary regression fits.

4 4 34

y = bo + £bx + £b«*2 + £ £ bm (2)

¡=1 ¡=1 ¡=1 j=i+i

where Y is the response factor (fatty acid methyl ester contents), xi = the Ith term of independent factor, po = intercept, pi = linear model coefficient, pii = quadratic coefficient for the factor i, and pij = linear model coefficient for the interaction between factors i and j.

Table 2 shows the complete experimental design matrix of central composite design for the factorial design. The order in which the runs were made was randomized to avoid systematic errors.

2.5. Statistical analysis for biodiesel production using response surface methodology

Experimental data shown in Table 3 were analyzed via response surface methodology using Design-Expert version8 software, ©2013 by Stat-Ease, Inc., USA and then interpreted. The analytical

4 C.B. Ezekannagha et al./Egyptian Journal of Petroleum xxx (2017) xxx-xxx

Table 3

Central composite design, experimental and predicted values of biodiesel yield.

Std Run Reaction Catalyst Time Methanol Experimental Predicted Residual

order order Temp. Amount to oil ratio Biodiesel Biodiesel

(°C) (wt%) (mins) (-) Yield (%) Yield (%)

1 22 SS 0.7S 40 6 80 78.666667 1.333333

2 18 6S 0.7S 40 6 87 8S.416667 1.S83333

3 17 SS 1.2S 40 6 84 82.916667 1.083333

4 11 6S 1.2S 40 6 96 96.166667 -0.16667

S 23 SS 0.7S 80 6 87 8S.416667 1.S83333

6 3 6S 0.7S 80 6 89 88.666667 0.333333

7 30 SS 1.2S 80 6 76 77.166667 -1.16667

8 4 6S 1.2S 80 6 88 86.916667 1.083333

9 28 SS 0.7S 40 12 82 80.916667 1.083333

10 S 6S 0.7S 40 12 78 77.166667 0.833333

11 9 SS 1.2S 40 12 86 86.666667 -0.66667

12 10 6S 1.2S 40 12 90 89.416667 0.S83333

13 2S SS 0.7S 80 12 86 86.166667 -0.16667

14 20 6S 0.7S 80 12 80 78.916667 1.083333

1S 27 SS 1.2S 80 12 80 79.416667 0.S83333

16 12 6S 1.2S 80 12 77 78.666667 -1.66667

17 2 S0 1 60 9 72 72.916667 -0.91667

18 1 70 1 60 9 78 78.916667 -0.91667

19 16 60 0.S 60 9 81 83.916667 -2.91667

20 7 60 l.S 60 9 89 87.916667 1.083333

21 21 60 1 20 9 92 93.916667 -1.91667

22 26 60 1 100 9 90 89.916667 0.083333

23 14 60 1 60 3 82 83.916667 -1.91667

24 19 60 1 60 1S 78 77.916667 0.083333

2S 1S 60 1 60 9 89 87.S l.S

26 29 60 1 60 9 88 87.S 0.S

27 13 60 1 60 9 87 87.S -0.S

28 24 60 1 60 9 88 87.S 0.S

29 8 60 1 60 9 87 87.S -0.S

30 6 60 1 60 9 86 87.S -l.S

steps used include: analysis of variance (ANOVA), regression analysis, and response surface plots of the interaction effects of the factors to evaluate optimum conditions for the yield of Biodiesel. The linear, quadratic and linear interactive effects of the process variables on the yield of biodiesel were calculated and their respective significance evaluated by ANOVA test. The p-value was used as the basis for measuring the significance of the regression coefficients, values of p less than 0.05 signified that the coefficient is significant otherwise insignificant. The response of the transesterification process was used to develop a mathematical model depicted in Eq. (2) that correlates the yield of FAME to the transesterification process variables studied. The adequacy of the model was tested by the coefficient of determination (R2) value as compared to the adjusted R2 value. The central composite design conditions and responses, are given in Tables 3. The actual yields of biodiesel produced at different process parameters were calculated and are also contained in Table 3.

3. Results and discussion

3.1. Characterization

Experimental values of various properties of lard oil and lard oil biodiesel according to American Society for Testing and Materials standards (ASTM) methods using standard apparatus for the measurement are shown in Table 4.

The Characterization results in Table 4 show that lard oil presented a low acid and iodine value of 0.84 mg KOH/g oil and 54.57 respectively which deviated from the results recorded by some researchers like Gutierrez et al. [18] who recorded raw lard oil acid value of 1.58 mg KOH/g. This could probably result from the feed of the pig and in agreement with the report of Ockerman, [19], who stated that pigs that have been fed different diets will have lard oil with significantly different fatty acid content and

iodine value. However, comparable results of acid value were recorded by Jeong et al. [4] and Dias et al. [17] who obtained the acid values of 0.67 and 0.71 mg KOH/g respectively though with varying iodine values of 84 and 67 respectively. The acid value of the lard biodiesel in this study was found to be 0.28 mg KOH/g with FFA of 0.14% after transesterification reaction which is lower than that of raw lard (0.84 mg KOH/g). The result of which indicates that acid value of the biodiesel decreased significantly after transesterification reaction.

The most significantly affected property of the lard oil was its viscosity, as shown in Table 4. Results of the analysis showed that the viscosity of the pork lard decreased on transesterification.

The determined physiochemical properties are in compliance to the specifications of ASTM.D6751-07; hence the lard methyl ester produced has characteristics suitable for use as fuel for diesel engines.

3.2. Fourier Transform infrared (FTIR) spectroscopy analysis

The FTIR spectra of the pork lard (raw oil) and transesterified pork lard (biodiesel) were carried out in the range from 500 to 4000 cm-1 to study the effect of transesterification on the pork lard. The FTIR spectra results of the raw oil and biodiesel are shown in Figs. 1 and 2 respectively. The changes in the (%) transmission and functional groups indicated the modifications that occurred during the transesterification process. The FTIR spectra of the raw oil and biodiesel have similar pattern because of few similarities that exist among triglycerides and methyl ester in their functional groups. However, small difference is observed in these three regions (C=O ester, CH3 and C-O ester) because biodiesel had a different compound/functional group (CH3) bonded compared to the lard oil. In Fig 2, the strong ester peak at 1750 cm-1 (C=O ester) and at 1170-1200 cm-1 (C-O ester) are clearly present in the spectra. This is in agreement with the study of Meyer et al. (2000) who

C.B. Ezekannagha et al./Egyptian Journal of Petroleum xxx (2017) xxx-xxx

Table 4

Fuel properties of lard methyl ester using ASTM methods.

Properties

Lard biodiesel

ASTM limits

Acid value Free fatty acid Specific gravity@30 "C Viscosity@40 "C Saponification value Iodine value Water content Peroxide value Cetane number Higher heating value Flash point Cloud point Pour point

Mg KOH/g oil

Mg KOH/g oil

gb/100goil

MEq/kg

MJ/kg °C °C °C

0.84 0.42 0.9088 25.26 225.8 54.57 TRACE 80.0-

0.8732

130 minimum

0.50 max

0.86-0.90 1.9-6.0

3 minimum

47 minimum

-3 to 12 -15 to 10

Wavenumber (i Fig. 1. FT-IR Spectra of raw oil (lard).

stated that the wave frequency (cm-1) for (C=O ester) functional group is 1750-1725 cm-1. Outside these two regions, another characteristic peak that indicates the presence of CH3 group in the mixtures of methyl ester can be observed at 1445 cm-1. Similar result was reported by [20,21].

3.3. Evaluation of regression model for the lard biodiesel yield

The correlation between the experimental process variables and biodiesel yield was evaluated using the central composite design (CCD) modeling technique. Second order polynomial regression equation fitted between the response yield of FAME (Y) and the process variables: Reaction temperature(X1), Catalyst amount (x2), Time(X3), Methanol/oil molar ratio(X4). From Table 5, the ANOVA results showed that the quadratic model is suitable to analyze the experimental data. The model for percentage of FAME content (Y) in terms of the coded factors of the process variables is given by:

Yield of FAME% = 87.50 + 1.50X1 + 1.00X2 - 1.00X3

- 1.50X4 + 1.63X1X2 - 0.87X1X3

- 2.62X1X4 - 3.12X2X3 + 0.38X2X4

- 0.37X3X4 - 2.90X2 - 0.40X2 + 1 10X3

- 1.65X:

To develop a statistically significant regression model, the significance of the regression coefficients was evaluated based on the p-values. The coefficient terms with p-values more than 0.05 are insignificant and are removed from the regression model. The analysis in Table 5 shows that linear terms of temperature, catalyst, time, methanol/oil molar ratio, quadratic terms of temperature, time, and methanol and interactive terms of temperature and catalyst, temperature and methanol, catalyst and time that is

4000 3500 ЗООО 2500 2000 1500 10ОО 500

Wavenumber (cm-1)

Fig. 2. FT-IR spectra of lard biodiesel.

6 C.B. Ezekannagha et al./Egyptian Journal of Petroleum xxx (2017) xxx-xxx

Table 5

Analysis of variance (ANOVA) for the regression model equation, values of regression coefficients and their significant tests and effects.

Source Coeff. SS Df MS F Std. Err P-value Prob > F

Model 87.50 833.53 14 59.54 21.35 0.68 <0.0001 Significant

X1Temperatu 1.50 54.00 1 54.00 19.36 0.34 0.0005

X2CatalystAmt 1.00 24.00 1 24.00 8.61 0.34 0.0103

X3-Time -1.00 24.00 1 24.00 8.61 0.34 0.0103

X4Methanol/Oil -1.50 54.00 1 54.00 19.36 0.34 0.0005

X1X2 1.63 42.25 1 42.25 15.15 0.42 0.0014

X1X3 -0.87 12.25 1 12.25 4.39 0.42 0.0535

X1X4 -2.62 110.25 1 110.25 39.53 0.420.42 <0.0001

X2X3 -3.12 156.25 1 156.25 56.03 0.42 <0.0001

X2X4 0.38 2.25 1 2.25 0.81 0.42 0.3833

X3X4 -0.37 2.25 1 2.25 0.81 0.42 0.3833

X? -2.90 230.01 1 230.01 82.47 0.32 <0.0001

X22 -0.40 4.30 1 4.30 1.54 0.32 0.2335

X23 1.10 33.44 1 33.44 11.99 0.32 0.0035

X24 -1.65 74.30 1 74.30 26.64 0.32 0.0001

Residual 41.83 15 2.79

Lack of Fit 36.33 10 3.63 3.30 0.0997

Pure Error 5.50 5 1.10 Insignificant

Cor Total 875.37 29

R2 = 0.9522 Adj.R2 = 0.9076.

Predicted vs. Actual

70.00 75.00 80.00 85.00 90.00 95.00 100.00

Actual

Fig. 3. Plot of the predicted versus the actual values of yield of FAME.

X1, X2, X3, X4, X1,X3,X2,X1X2, X1X4, X2X3 are significant model terms. The model reduces to Eq. (4) after eliminating the insignificant coefficients.

Yield of FAME% = 87.50 + 1.50X1 + 1.OOX2 - 1.OOX3

- 1.50X4 + 1.63X1X2 - 2.62X1X4

- 3.12X2X3 - 2.90X2 + 1.10X3 - 1.65X2 (4)

3.4. Model adequacy check

The analysis of variance indicated that the quadratic polynomial model was significant and adequate to represent the actual relationship between yield of FAME and the significant model variable as depicted by very small p-value of 0.0001.

The significance and adequacy of the established model was also elaborated by high value of coefficient of determination (R2)

90 85 80 75

B: Catalyst Amount

5700 A: Temperature

0.75 55.00

Fig. 4. Surface plot of the Interaction effect of temperature and catalyst on yield of lard biodiesel at constant time and methanol to oil molar ratio.

C.B. Ezekannagha et al./Egyptian Journal of Petroleum xxx (2017) xxx-xxx

value of 0.9522 and adj.R2 value of 0.9076 for the developed correlation. It implies that the model explains 95.22% of the total variation in the yield of FAME which is attributed to the experimental variables.

A line of unit slope, i.e. line of perfect fit with points corresponding to zero error between the predicted values and actual values was shown in Fig. 3 indicating that the model represents a relatively good description of the experimental data regarding the yield of FAME. It could be seen from the graph that all the points are very close to the line of perfect fit hence, there is adequate correlation between the predicted values and the experimental values of the independent variable which further elaborated the adequacy of the model.

The ''Lack of Fit F-value" of 3.30 implies the Lack of Fit is not significant relative to the pure error. Non-significant lack of fit is good because we want the model to fit.

3.5. Response surface plots for lard biodiesel production

The interactive effects of the process variables on the percent yield of FAME were studied by plotting three dimensional surface curves against any two independent variables, while keeping other variables at their central (0) level. The 3D curves of the response (yield of methyl ester) from the interactions between the variables are shown in Figs. 4-9. The process variables were found to have significant interaction effects.

The interactive effect of temperature and catalyst on the yield of methyl ester is positive as shown in Table 5 that is, increasing both variables increases the yield of biodiesel and this could be attributed to the fact that at higher reaction temperature there is higher rate of reaction as the reacting species collide more frequently and also simultaneous increase in catalyst concentration will increase the availability of catalysts for the reacting species

40.00 55.00

Fig. 5. Surface plot of the interaction effect of temperature and time on yield of lard biodiesel at constant catalyst amount and methanol to oil molar ratio.

Fig. 6. Surface plot of the interaction effect of temperature and methanol/oil molar ratio on yield of lard biodiesel at constant catalyst amount and time.

C.B. Ezekannagha et al./Egyptian Journal of Petroleum xxx (2017) xxx-xxx

80.00 1.25

C: Time 48.m\^^-o85 b: Catalyst Amount

40.00 0.75

Fig. 7. Surface plot of the interaction effect of catalyst amount and time on yield of lard biodiesel at constant temperature and methanol to oil molar ratio.

thereby pushing the reaction forward in favor of increased yield of methyl ester. The response surface plot of the interactive effect of temperature and catalyst is shown in Fig. 4, the yield of FAME increased with simultaneous increase in temperature and catalyst to about 65 °C and 1.25 wt% respectively beyond which the value of yield declined. The same trend was observed for the interactive effects of catalyst and methanol and the response surface plot is shown in Fig. 8.

The interactive effects of time and methanol on the yield of FAME are negative as shown in Table 5, that is increasing both variables resulted in decrease in the yield of FAME and this could be attributed to the fact that the optimum molar ratio for maximum biodiesel yield from animal fat to be in between 4.8:1 and 6.5:1 [22]. The response surface plot of the interactive effects of time and methanol are shown in Fig. 9, the yield of FAME declined beyond the molar ratio (6:1) and 40 min respectively because excess methanol deactivated the catalyst activity thereby reducing

its effectiveness and after 40 min there is declination in yield as all the reaction has been completed and the process reached equilibrium. Prolonged heating made the produced biodiesel to revert back to its original state and thus reducing overall biodiesel yield. The same trend was observed in the interactive effects of temperature and time, temperature and methanol, catalyst and time, and the response surface plots of the interactive effects are shown in Figs. 5-7 respectively. Generally increase in temperature resulted in corresponding increase in yield of FAME as was evidenced in Figs. 4-6.

Reaction temperature and methanol to oil molar ratio are the most significant process variable that affect the yield of the FAME as indicated by their highest F values in the ANOVA (Table 5).

The optimum conditions are: reaction temperature 65 °C; catalyst amount 1.25 wt%; time 40 min; methanol to oil molar ratio 6:1 and optimum FAME yield at these optimum conditions was predicted to be 96.2%. In order to validate the predicted

95 90 85 80 75

D: Methanol/Oil 7.00^^^^ 0 85 B: Catalyst Amount

6.00 0.75

Fig. 8. Surface plot of the interaction effect of catalyst amount and methanol to oil molar ratio on yield of lard biodiesel at constant temperature and time.

C.B. Ezekannagha et al./Egyptian Journal of Petroleum xxx (2017) xxx-xxx

90 85 80 75 70

D: Methanol/Oil 7.00^^^^ te.oo C: Time

6.00 40.00

Fig. 9. Surface plot of the interaction effect of time and methanol to oil molar ratio on yield of lard biodiesel at constant temperature and catalyst amount.

optimum values; experiments were carried out at these optimum conditions. The experimental value of 96% agreed closely with that obtained from the regression model. The high experimental value could be accounted for by the low fatty acid and water content determined and the optimum reaction conditions used for the experiment. This is comparable to the work of Jeong et al. [4] with experimental value of lard biodiesel yield 97.8%.

4. Conclusion

Lard oil as low cost feedstock is suitable for methyl ester (biodiesel) production as biodiesel was successfully produced from it via transesterification. The parameters used for quality of biodiesel determination tested in this work under optimal conditions satisfied the standards for biodiesel quality. The experiments conducted in this study using response surface methodology to determine the optimal reaction conditions for the production of biodiesel from lard gave the optimal values of the variables as follows: Reaction temperature of 65 °C, catalyst amount of 1.25% (w/ w), time of 40 min, and methanol to oil ratio of 6:1. At this predicted optimum condition, the predicted fatty acid methyl ester (FAME) content was 96.2%. The experimental value of 96% was well within the estimated value of the model. This demonstrated that response surface methodology with appropriate experimental design can be effectively applied to the optimization of the process parameters in methyl ester production. Through appropriate optimization of process parameters, economically viable biodiesel could be produced from low cost feedstock-lard oil which could substitute or a blending component with petroleum-based diesel in existing diesel engine without modification to meet the ever increasing demand of fuel oil.

References

[1] L.C. Meher, V.D. Sagar, S.N. Naik, Technical aspects of biodiesel production by transesterification, Renew. Sustain. Energy Rev. 10 (2006) 248-268.

[2] J.M. Encinar, J.F. Gonzalez, A. Rodriguez-Reinaeres, Biodiesel from used frying oil. Variables affecting the yields and characteristics of the biodiesel, Ind. Eng. Chem. Res. 44 (2005) 5491-5499.

[3] T. Krawczyk, Biodiesel Inform. 7 (1996) 801-822.

[4] G. Jeong, H. Yang, D. Park, Optimization of transesterification of animal fat acid methyl ester using response surface methodology, Bioresour. Technol. 100 (1) (2009) 25-30.

[5] F. Ma, M.A. Hanna, Biodiesel production: a review, Bioresour. Technol. 70 (1999) 1-15.

[6] V.B. Veljkovic, S.H. Lakicevic, O.S. Stamenkovic, Z.B. Todorovic, K.L. Lazic, Biodiesel production from tobacco (Nicotianatabacum L.) seed oil with a high content of free fatty acids, Fuel 85 (2006) 2671-2675.

[7] M. Canakci, J. Van Gerpen, Biodiesel production from oils and fats with high free fatty acids, Trans. ASAE 44 (6) (2001) 1429-1436.

[8] M. Mittelbach, B. Porkits, Silberholz, 1992. Production and fuel properties of fatty acidmethyl esters from used frying oil. Pro. Altern. Energy.

[9] A. Demirbas, Biofuels: Securing the Planet's Future Energy Needs, Springer, London (England), 2009. P. 336, e-ISBN 978-1-81882-011-1.

10] M.K. Lam, K.T. Lee, A.R. Mohammed, Homogeneous, heterogenous and enzymatic catalysis for transesterification of high free fatty acid oil (waste cooking oil) to biodiesel: a review, Biotechnol. Adv. 28 (4) (2010) 500-518.

11] C.C. Enweremadu, M.M. Mbarawa, Technical aspects of production and analysis of biodiesel from used cooking oil. A review, Renew. Sustain. Energy Rev. 13 (2009) 2205-2224.

12] D.Y.C. Leung, X. Wu, M.K.H. Leung, A review on biodiesel production using catalyzed transesterification, Appl. Energy 87 (2010) 1083-1095.

13] B.H. Hameed, L.F. Lai, L.H. Chin, Production of biodiesel from palm oil (Elaeis guineensis) using heterogenous catalyst: an optimized process, Fuel Process. Technol. 90 (2009) 606-610.

14] C.J. Shieh, H.F. Liao, C.C. Lee, Optimization of lipase catalyzed biodiesel by response surface methodology, Bioresour. Technol. 88 (2003) 103-106.

15] E. Uosukainen, M. Lamsa, Y. Linko, P. Linko, M. Leisola, Optimization of enzymatic transesterification of rapeseed oil ester using response surface and principal component methodology, Enzyme Microb. Technol 25 (1999) 236243.

16] D.H. Jeong, Park, Response surface methodological approach for optimization of enzymatic synthesis of sorbitan methacrylate, Enzyme Microbiol. Technol. 39 (3) (2006) 381-386.

17] J.M. Dias, M.C. Alvim-Ferraz, M.F. Almeida, Production of biodiesel from acid waste lard, Bioresour. Technol. 100 (2009) 6355-6361.

18] M.C. Gutierrez, M. Berrios, A.F. Chica, A. Martin, Residual lard fat: a good alternative as biodiesel raw material. Pre-processing of manure and organic waste for energy production. RAMIRAN-Research network on recycling of agricultural municipal and industrial residues in agriculture, 2010.

19] H.W. Ockerman, L. Basu, Edible rendering-rendered products for humanuse, in: D.L. Meeker (ed)., Essential Rendering: All About Animal By-products Industry. National Renderers Association, Arlington, VA, 2006, pp. 95-110.

20] H.T. Yun, F.A. Nurul, B. Mahiran, Biodiesel Production via transesterification of palm oil using NaOH/Al2O3 catalysts, Sains Malaysiana 40 (6) (2011) 587-594.

21] N.G. Siatis, A.C. Kimbaris, C.S. Pappas, P.A. Tarantilis, M.G. Polissiou, Improvement of biodiesel production based on the application of ultrasound: monitoring of the procedure by FTIR spectroscopy, J. Am. Oil Chem. Soc. 83 (1) (2006) 53-57.

[22] B. Freedman, E.H. Pryde, T.L. Mounts, Variables affecting the yield of fatty esters from transesterified vegetable oils, J. Am Oil Chem. Soc. 61 (10) (1984) 1638-1643.