at SciVerse ScienceDirect

Energy 58 (2013) 296e304

Contents lists available

Energy

journal homepage: www.elsevier .com/locate/energy

A comparative evaluation of physical and chemical propertiesof biodiesel synthesized from edible and non-edible oils and studyon the effect of biodiesel blending

A.E. Atabani a,*, T.M.I. Mahlia b, H.H. Masjuki a, Irfan Anjum Badruddin a,Hafizuddin Wan Yussof c, W.T. Chong a, Keat Teong Lee d

aDepartment of Mechanical Engineering, University of Malaya, 50603 Kuala Lumpur, MalaysiabDepartment of Mechanical Engineering, Universiti Tenaga Nasional, 43009 Kajang, Selangor, Malaysiac Faculty of Chemical and Natural Resources Engineering, University Malaysia Pahang Lebuhraya Tun Razak, 26300 Kuantan, Pahang, Malaysiad School of Chemical Engineering, Universiti Sains Malaysia, Engineering Campus, Seri Ampangan, 14300 Nibong Tebal, Pulau Pinang, Malaysia

a r t i c l e i n f o

Article history:Received 14 December 2012Received in revised form25 April 2013Accepted 1 May 2013Available online 1 July 2013

Keywords:Crude oil characteristicsBiodiesel productionPhysical and chemical propertiesBlending effect

* Corresponding author. Tel.: þ60 122314659.E-mail address: a_atabani2@msn.com (A.E. Ataban

0360-5442/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.energy.2013.05.040

a b s t r a c t

Traditionally, biodiesel has been produced from edible oils due to their low free fatty acids. However,their use has elevated some issues such as food versus fuel and many other problems that have nega-tively affected their economic viability. Therefore, exploration of non-edible oils may significantly reducethe cost of biodiesel especially in poor countries which can barely afford the high cost of edible oils. Thispaper aims to produce biodiesel from several edible and non-edible oils that are readily available in theSouth East Asian market. These oils include; Jatropha curcas, Calophyllum inophyllum, Sterculia foetida,Moringa oleifera, Croton megalocarpus, Patchouli, Elaeis guineensis (palm), Cocos nucifera (coconut),Brassica napus (canola) and Glycine Max (soybean) oils. This was followed by an investigation of physico-chemical properties of the produced biodiesel. This paper also discusses the concept of biodieselblending to improve some of the properties of these feedstocks. For instance, blending of SFME andCoME improves the viscosity of SFME from 6.3717 mm2/s to 5.3349 mm2/s (3:1), 4.4912 mm2/s (1:1) and3.879 mm2/s (1:3). The properties of other biodiesel blends were estimated using the polynomial curvefitting method.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Currently, more than 80% of theworld’s energy needs depend onfossil fuels. However, the declining reserve of fossil fuel and thegrowing carbon dioxide emissions are driving the world attentionto reduce dependence on fossil fuel. In turn, this has promoted theinterest in bioenergy, including biodiesel, as one of the primerenewable energy sources. Biodiesel is gaining worldwide atten-tion. Edible oils are considered as the first generation of biodieselfeedstock. Biodiesel has been produced in the US and Europe usingedible oils because they have surplus of them, can achieve highbiodiesel yield and easy processing due to their low free fatty acids.However, their use has raised many concerns such as food versusfuel problem and some environmental problems such as seriousdestruction of vital soil resources, deforestation and usage of much

i).

All rights reserved.

of the available arable land as can be seen in many countriesespecially highly populated countries such as China and India. All ofthese factors negatively affected the economic viability of biodieselproduction from edible oils. It is known that the cost of feedstockalone represents 75% of the cost of biodiesel. Therefore, explorationof new low-cost agricultural non-edible crops and the utilization ofby-products in the biodiesel production may significantly reducethe cost of biodiesel especially in poor countries which can hardlyafford the high cost of edible oils [1e14].

The availability of any feedstock for biodiesel production de-pends on the regional climate, geographical locations, local soilconditions and agricultural practices of any country [15]. Forinstance, canola oil is mainly used in Europe, palm and coconut oilin South East Asia and soybean oil and animal fats are used in theUnited States. However, the collective supply of these fats and oils isnot adequate to displace fossil fuels. For example, if all U.S. soybeanproduction were dedicated to biodiesel, an estimated 6% of dieseldemand would be satisfied. Thus, alternative feedstocks for bio-diesel production such as Jatropha curcas, Sterculia foetida, Moringa

Nomenclature

CIME Calophyllum inophyllum methyl esterCME Canola methyl esterCMME Croton megalocarpus methyl esterCoME coconut methyl esterJCME Jatropha curcas methyl esterMOME Moringa oleifera methyl esterPaME Patchouli methyl esterPME palm oil methyl esterSFME Sterculia foetida methyl esterSME soybean methyl ester

A.E. Atabani et al. / Energy 58 (2013) 296e304 297

oleifera, Croton megalocarpus, and Calophyllum inophyllum haveattracted considerable attention. These feedstocks are consideredmore sustainable as they can be grown on land that is unsuitable forother crops [2,3,16e18].

Therefore, it can be understood that the combined supply ofedible and non-edible oils is currently one of the available solutionsto meet the world energy demand and reduce the dependency onthe edible oils.

1.1. Objectives of this study

The main objective of this work is to produce biodiesel fromseveral edible and non-edible oils that are readily available in theSouth East Asian market. These oils include; J. curcas, C. inophyllum,S. foetida, M. oleifera L., C. megalocarpus, Patchouli, Elaeis guineensis(palm), Cocos nucifera (coconut), Brassica napus L. (canola) andGlycine Max (soybean) oils (Fig. 1). Moreover, a detailed investiga-tion of physical and chemical properties of the produced biodieselsfrom these feedstocks has been fully presented in this study. Thisis very important to assess the quality of these feedstocks. More-over, this paper also suggests the concept of biodieselebiodieselblending from these feedstocks to improve some of the mainproperties such as viscosity, cloud, pour and cold filter pluggingpoint. In this paper, the polynomial curve fitting method was sug-gested to predict the properties of biodiesel blends.

2. Materials and methods

2.1. Materials and chemicals

The crude oils of Palm, Coconut, Canola and Soybean were ob-tained from local market. M. oleifera and C. megalocarpus oils weresupplied from University Sains Malaysia (USM). J. curcas, C. ino-phyllum and S. foetida oils were purchased fromMinistry of Forestryof the Republic of Indonesia (Bogor, Indonesia). Patchouli oil waspurchased from Banda Aceh (Indonesia). Other chemicals such asmethanol, H2SO4, KOH and Na2SO4 were obtained from SigmaAldrich (Malaysia). Qualitative filter paper (filtres Fioroni) of150 mm size was supplied from (Metta Karuna Enterprise,Malaysia). The production of biodiesel was conducted in the labo-ratory scale using 500 ml batch reactor.

2.2. Structure of the study

This paper aims to produce biodiesel from various edible andnon-edible oils that are readily available in the South East Asianmarket. Therefore, the following figure (Fig. 2) gives a summary ofthe implemented flow chart of this paper.

2.3. Biodiesel production procedures

2.3.1. Production of biodiesel from palm, coconut, canola andsoybean oils

In this process, crude palm, coconut, canola and soybean oilswere reacted with 25% (v/v oil) of methanol and 1% (m/m oil) ofpotassium hydroxide (KOH) and maintained at 60 �C for 2 h and400 rpm stirring speed. After completion of the reaction, the pro-duced biodiesels were deposited in a separation funnel for 12 h toseparate glycerol from biodiesel. The lower layer which containedimpurities and glycerol was drawn off.

2.3.2. Production of biodiesel from J. curcas, C. inophyllum, S.foetida, M. oleifera, C. megalocarpus and Patchouli oils2.3.2.1. Esterification process. This process was employed primarilyto reduce the acid value of these feedstocks prior to the secondtransesterification step. In this process, the molar ratio of methanolto refined J. curcas, C. inophyllum, S. foetida, M. oleifera, C. mega-locarpus and Patchouli oils was maintained at 12:1 (50% v/v oil). 1%(v/v oil) of sulphuric acid (H2SO4) was added to the pre-heated oilsat 60 �C for 3 h under 400 rpm stirring speed in a glass reactor. Oncompletion of this reaction, the products were poured into aseparating funnel to separate the excess alcohol, sulphuric acid andimpurities presented in the upper layer. The lower layers wereseparated and entered into a rotary evaporator and heated at 95 �Cunder vacuum conditions for 1 h to remove methanol and waterfrom the esterified oils. Due to the high acid value of C. inophyllum,this process has been conducted twice to reduce the acid value ofthe oil to less than 4 mgKOH/g oil.

2.3.2.2. Transesterification process. In this process, the esterifiedoils of J. curcas, C. inophyllum, S. foetida, M. oleifera, C. megalocarpusand Patchouli were reacted with 25% (v/v oil) of methanol (6:1molar ratio) and 1% (m/m oil) of potassium hydroxide (KOH) andmaintained at 60 �C for 2 h and 400 rpm stirring speed. Aftercompletion of the reaction, the produced biodiesels were depositedin a separation funnel for 12 h to separate glycerol from biodiesels.The lower layers which contained impurities and glycerol werediscarded.

2.3.3. Post-treatment processThe methyl esters formed in the upper layer from the previous

processes were washed with distilled water to remove theentrained impurities and glycerol. In this process, 50% (v/v oil) ofdistilled water at 60 �C was sprayed over the surface of the estersand stirred gently. This process was repeated several times until thepH of distilled water became neutral. The lower layers were thendiscarded and upper layers were poured entered into a rotaryevaporator to remove the excess water andmethanol. Methyl esterswere further dried using Na2SO4. Finally, the final product wasfiltered using qualitative filter paper (filtres Fioroni).

2.4. Measurement of physicalechemical properties of crude oils andbiodiesels

In this study, the physical and chemical properties of the crudeoils and their respective biodiesels were tested according to ASTMD6751. These properties include; viscosity, viscosity index, density,flash point, CP (cloud point), PP (pour point), CFPP (cold filterplugging point), CCR, calorific value, oxidation stability, copperstrip corrosion and total sulphur beside some other non-ASTMproperties such as transmission, absorbance and refractive index.Table 1 shows the equipment used in this study to analyse theseproperties. Moreover, the details of their manufacturers and theASTM D6751 methods were also included in the table1.

Fig. 1. Feedstock pictures.

A.E. Atabani et al. / Energy 58 (2013) 296e304298

2.5. Blending of biodiesel

The test fuels were blended by using a homogenizer device at aspeed of 2000 rpm. The homogenizer was fixed on a clamp on avertical stand, which allows changing of the homogenizer’s height.To mix the fuels by using the homogenizer, the plug is turned onand the appropriate speed is selected by using the selector which islocated on top of the homogenizer.

In this study the effect of biodiesel blending ratios of 1:1 (v/v)and 1:3 (v/v) on some physical and chemical properties has beenstudied and presented. These include viscosity, cloud point, pourpoint and cold filter plugging point. In this paper, polynomial curvefitting method was used to estimate the properties of other bio-diesel blends. This method is an attempt to describe the relation-ship between variable X as a function of available data and aresponse Y, which seeks to find a smooth curve that best fits thedata. Mathematically, a polynomial of order k in X is expressed inthe following form:

Y ¼ Co þ C1X þ C2X2 þ/þ CkX

k (1)

3. Results and discussion

3.1. Characterization of crude oils

The properties of crude oils were presented in Table 2. The mainfindings from this table show that S. foetida oil possesses thehighest kinematic viscosity of 75.826 mm2/s at 40 �C. C. mega-locarpus oil has the best viscosity index of 224.2. Patchouli oilpossesses the lowest CFPP and highest refractive index and calorificvalue of 1 �C, 1.5069 and 42,986 kJ/kg respectively. Canola has thehighest flash point of 290.5 �C. M. oleifera oil was found to have thebest oxidation stability of 41.75 h. Coconut oil possesses the highesttransmission of 91.2% (T) and lowest Absorbance of 0.04 (abs).

3.2. Characterization of biodiesel

The physical and chemical properties of the produced biodieselsfrom different edible and non-edible oils are shown in Table 3. Thefollowing section will discuss some of the obtained results andcompare them with each other.

Fig. 1. (continued).

A.E. Atabani et al. / Energy 58 (2013) 296e304 299

3.2.1. Kinematic viscosityViscosity is the most important property of any fuel as it in-

dicates the resistance of a material to shear or flow. It thereforeaffects the operation of the fuel injection equipment and sprayatomization, particularly at low temperatures when the increase inviscosity affects the fluidity of the fuel [14,15,18]. Table 3 shows theobtained results of kinematic viscosity of CIME, JCME, SFME, PME,CoME, CME, SME, CMME, PaME and MOME. From this table, it canbe seen that CoME possesses the lowest kinematic viscosity of3.1435 mm2/s followed by CMME (4.0707 mm2/s), SME(4.3745 mm2/s), CME (4.5281 mm2/s), PME (4.6889 mm2/s), JCME(4.9476 mm2/s), MOME (5.0735 mm2/s), CIME (5.5377 mm2/s),PaME (6.0567 mm2/s) and finally SFME (6.3717 mm2/s). The

kinematic viscosities of biodiesel are all lower that those presentedby their respective oils as can be seen in Table 2. This is an expectedfinding since biodiesel molecules are single, long chain fatty esterswith higher mobility than the bigger and bulkier triglyceridemolecules [19]. It can be observed that all results are in agreementwith the standard specified by ASTM D6751of (1.9e6 mm2/s)except for SFME and PaME which both have slightly higher vis-cosity than the limit prescribed by ASTM D445. A comparison ofsome results in the present study and those in literature was alsoconducted. It was found that CME possesses slightly higher vis-cosity (4.5281mm2/s) than that of [13] with 4.50mm2/s and [19,20]with 4.439mm2/s. CoME possesses higher viscosity (3.1435mm2/s)than that of [19,20] with 2.726 mm2/s. JCME possesses higher

Characterization of crude oils

Biodiesel production

Physical and chemical properties of biodiesel

Investigation of biodiesel blending opportunities

Fig. 2. Flow chart of the present study.

A.E. Atabani et al. / Energy 58 (2013) 296e304300

viscosity (4.9476 mm2/s) than that of [19] with 4.253 mm2/s, [21]with 4.42 mm2/s, [22] with 4.84 mm2/s and [13,20] with 4.80mm2/s but lower than that of [22] with 5.38 mm2/s. MOME pos-sesses higher viscosity (5.0735 mm2/s) than that of [19] with 4.008mm2/s, [20,23] with 4.83 mm2/s, [24] with 4.91 mm2/s and [25]with 4.80 mm2/s. PME possesses higher viscosity (4.6889 mm2/s)than that of [19] with 4.570mm2/s, [21] with 4.52 mm2/s, [22] with4.5 mm2/s and of [13] with 4.42 mm2/s but lower than that of [20]with 5.07 mm2/s. SME possesses higher viscosity (4.3745 mm2/s)than that of [19,20] with 4.039 mm2/s and [13,22] with 4.08 mm2/s.SFME possesses slightly higher viscosity (6.3717 mm2/s) than that

Table 1Equipment list.

Property Equipment

1 Kinematic viscosity SVM 3000 e automatic2 Flash Point Pensky-martens flash point e automatic

NPM 4403 Oxidation stability 873 Rancimat e automatic4 Cloud and Pour point Cloud and Pour point tester e automatic

NTE 4505 CCR Micro-Carbone Conradson

Residue Tester e automatic NMC 4406 Total sulphur Multi EA 5000 e automatic

7 CFPP Cold filter pluggingpoint e automatic NTL 450

8 Density SVM 3000 e automatic9 Copper strip corrosion Seta copper corrosion bath 11300-010 Dynamic viscosity SVM 3000 e automatic11 Viscosity Index (VI) SVM 3000 e automatic12 Caloric value C2000 basic calorimeter e automatic13 Refractive Index RM 40 Refractometer e automatic14 Transmission Spekol 150015 Absorbance Spekol 1500

N/S ¼ not specified in ASTM D6751 test method.

of [26] with 6.0 mm2/s. While, CMME possesses lower viscosity(4.0707 mm2/s) than that of [27] with 4.8 mm2/s and [28] of4.56 mm2/s. Moreover, CIME possesses lower viscosity(5.5377 mm2/s) than that of [29] with 5.724 mm2/s but higher thanthat of [30,31] with 4 mm2/s and 4.92 mm2/s respectively.

3.2.2. Oxidation stabilityOxidation stability is an important parameter to investigate; it is

an indication of the degree of oxidation, potential reactivity withair, and can determine the need for antioxidants [14,19]. Table 3shows the obtained results of oxidation stability. It can be seenthat PME possesses the highest oxidation stability of 23.56 h fol-lowed by MOME (12.64 h), CoME (8.01 h), CME (7.08 h), CIME(6.12 h), JCME (4.84 h), SME (4.08 h), SFME (1.46 h), CMME (0.71 h)and PaME (0.022 h). It can be seen that PME, MOME, CoME, JCME,SME, CME and CIME possess oxidation stabilities that agree withthe range specified by ASTM D6751standard of min 3 h. However,the oxidation stability of SFME, CMME and PaME are less than theprescribed ASTM standard. This indicates that the use of antioxi-dants is necessary to meet standards specifications. The compari-son with other results in literature revealed that the oxidationstability of CME (7.08 h) is comparable to that of [19,20] with 7.6 h.CoME possesses lower oxidation stability (8.01 mm2/s) than that of[19,20] with 35.5 h. However, the result in the present study satisfyboth ASTM d6751 and EN 14214 standards of minimum 3 h and 6 hrespectively. While, JCME possesses higher oxidation stability(4.84 h) than that of [19,20] with only 2.3 h. Moreover, MOMEpossesses higher oxidation stability (12.64 h) than that of [19,20]with only 2.3 h, [23] of 3.61 h and [25] of 3.52 h. PME possesseshigher oxidation stability (23.56 h) than that of [19] with only 0.2 hand [20] with 4 h. SME possesses higher oxidation stability (4.08 h)than that of [19,20] with only 2.1 h.

3.2.3. Cloud, pour and cold filter plugging pointThe cloud point (CP) is the temperature at which wax crystals

first becomes visible when the fuel is cooled. Pour point (PP) is thetemperature at which the amount of wax out of solution is suffi-cient to gel the fuel, thus it is the lowest temperature at which thefuel can flow. Cold filter plugging point (CFPP) refers to the tem-perature at which the test filter starts to plug due to fuel compo-nents that have started to gel or crystallize. It is commonly used asindicator of low temperature operability of fuels and reflects their

Manufacturer ASTM D6751 ASTM D6751 limit

(Anton Paar, UK) D 445 1.9e6.0(Normalab, France) D 93 130 min

(Metrohm, Switzerland) D 675 3 h min(Normalab, France) D 2500 and D 97 Report

(Normalab, France) D 4530 0.050% m max

(Analytical Jena, Germany) D 5453 15 Max (S15)500 Max (S500)

(Normalab, France) D 6371 N/S

(Anton Paar, UK) D 1298 N/S(Stanhope-Seta, UK) D 130 No.3 max(Anton Paar, UK) N/S N/S(Anton Paar, UK) N/S N/S(IKA, UK) N/S N/S(Mettler Toledo, Switzerland) N/S N/S(Analytical Jena, Germany) N/S N/S(Analytical Jena, Germany) N/S N/S

Table 2Physical and chemical properties of crude oils.

Property Jatropha Sterculia Calophyllum Coconut Palm Canola Soybean Croton Moringa Patchouli

1 Kinematic viscosity (mm2/s) at 40 �C 48.095 75.826 55.478 27.640 41.932 35.706 31.7390 29.8440 43.4680 9.81752 Kinematic viscosity (mm2/s) at 100 �C 9.1039 13.608 9.5608 5.9404 8.496 8.5180 7.6295 7.2891 9.0256 2.21513 Dynamic viscosity (mpa s) at 40 �C 43.543 69.408 51.311 25.123 37.731 32.286 28.796 27.1570 38.9970 9.29334 Viscosity index (VI) 174.1 184.8 165.4 168.5 185.0 213.5 223.2 224.20 195.20 �21.605 Flash point (�C) 258.5 246.5 236.5 264.5 254.5 290.5 280.5 235 263 146.56 CFPP (�C) 21 29 26 22 23 15 13 10 18 17 Density (kg/m3) at 40 �C 0.9054 0.9153 0.9249 0.9089 0.8998 0.9042 0.9073 0.9100 0.8971 0.94668 High caloric value (kJ/kg) 38,961 39,793 38,511 37,806 39,867 39,751 39,579 39,331 39,762 42,9869 Copper strip corrosion (3 h at 50 �C) 1a 1a 1a 1a 1a 1a 1a 1a 1a 1a10 Refractive index 1.4652 1.4651 1.4784 1.4545 1.4642 1.471 1.4725 1.4741 1.4661 1.506911 Transmission (%T) 61.8 26.6 34.7 91.2 63.2 62.9 65.2 87.5 69.2 71.412 Absorbance (Abs) 0.209 0.574 0.46 0.04 0.199 0.202 0.186 0.058 0.16 0.14613 Oxidation stability (h at 110 �C) 0.32 0.15 0.23 6.93 0.08 5.64 6.09 0.14 41.75 0.13

A.E. Atabani et al. / Energy 58 (2013) 296e304 301

cold weather performance [15,18]. Table 3 shows the obtained re-sults of CP, PP and CFPP. It can be observed that CIME, JCME, MOMEand PME have relatively higher CP, PP and CFPP than SME, PaME,CME, CoME, CMME and SFME. The comparison of some results inthe present study and those in literature shows that CME possess acomparable CP and CFPP of�3 �C and�10 �C compared to�3.3 and�13 of Refs. [19,20]. CME possesses similar CP and PP of �3 �C and�9 �C compared to �4 �C and �9 �C of Ref. [13]. Same observationwas drawn for SMEwith CP and CFPP of 1 �C and�3 �C compared to0.9 and �4 of Ref. [19,20]. While CoME possesses slightly higher CPand CFPP of 1 �C and �1 �C compared to 0 �C and �4 �C ofRef. [19,20]. However, JCME possesses very high CP and CFPP of10 �C compared to 2.7 �C and 0 �C of Refs. [19,20]. MOME possesseshigher CP and CFPP of 21 �C and 18 �C compared to 13.3 �C and 13 �Cof Refs. [19,20] and 18 �C and 17 �C of Ref. [25]. MOME possessesslightly higher PP of 19 �C compared to 17 �C of Ref. [25]. The resultsof CP and CFPP for PME were similar to Refs. [19,20] of 13 �C and12 �C respectively. CIME possesses lower CP 12 �C compared to13.2 �C but higher PP of and 13 �C compared to 4.3 �C of Ref. [32].CMME possesses higher CP and PP of �3 �C and �2 �C compared to�4 �C and �9 �C of Ref. [28]. SFME possesses slightly higher pourpoint of 2 �C compared to 1 �C of Ref. [26].

3.2.4. Flash pointFlash point of a fuel is the temperature at which it will ignite

when exposed to a flame or a spark [14,15,18]. Table 3 shows theobtained results of flash point. It can be seen that CIME, JCME, SME,CME, SFME, PME, MOME and CMME have flash points of 162.5 �C,

Table 3Physical and chemical properties of the produced biodiesel.

CIME JCME SME

1 Viscosity at 40 �C (mm2/s) 5.5377 4.9476 4.37452 Viscosity at 100 �C (mm2/s) 1.998 1.8557 1.7643 Viscosity at 40 �C (mpa s) 4.8599 4.2758 3.80144 Density at 40 �C (kg/cm3) 0.8776 0.8642 0.8695 Oxidation stability (h at 110 �C) 6.12 4.84 4.086 CFPP (�C) 11 10 �37 Cloud point (�C) 12 10 18 Pour point (�C) 13 10 19 Flash point (�C) 162.5 186.5 202.510 Copper strip corrosion (3 h at 50 �C) 1a 1a 1a11 High caloric value (kJ/kg) 39,513 39,738 39,97612 CCR (m/m%) 0.4069 0.0440 0.020413 Total sulfur (mg/kg) 4.11 3.84 0.8614 Absorbance (abs) at WL 656.1 0.057 0.045 0.03715 Transmission (%) at WL 656.1 87.7 90.3 9216 Refractive index (RI) at 25 �C 1.4574 1.4513 1.455317 Viscosity index 183.2 194.6 257.8

N/D ¼ not determined.

186.5 �C, 202.5 �C, 186.5 �C, 130.5 �C, 214.5 �C, 176 �C and 164 �Crespectively. These results agree with the specification of flashpoint in ASTM D6751 of minimum 130 �C except for CoME andPaME which have a flash point of 118.8 �C. The comparison withliterature shows that the flash points of JCME, CME and PME ob-tained in this study (186.5 �C, 186.5 �C and 214.5 �C) are higher thanthose presented in Ref. [13] of 135, 170 and 182 �C respectively.Moreover, the results of CIME presented in this study of 162.5 �C arehigher than those presented in Refs. [29e31] of 151 �C, 140 �C and140 �C respectively. The flash point of MOME (176 �C) is higher thanthat presented in Ref. [25] of 162 �C. The flash point of SME(202.5 �C) is higher than that presented in Refs. [20,26] of 178 �C.The flash point of CoME 118.5 �C is slightly higher than that pre-sented in Ref. [20] of 110 �C. For CMME the result obtained in thisstudy (164 �C) is lower than that obtained in Ref. [28] of 189 �C.Same observation was drawn for SFME which has flash point of130.5 �C compared to 162 �C in Ref. [26].

3.2.5. High caloric valueCaloric value is an important parameter in the selection of a fuel.

The caloric value of biodiesel is generally lower than of dieselbecause of its higher oxygen content [14,15,18]. Table 3 shows theobtained results of calorific value of the produced biodiesel. It canbe observed that PaME possesses the highest calorific value of44,180 kJ/kg followed by CME (40,195 kJ/kg), MOME (40,115 kJ/kg),PME (40,009 kJ/kg), SFME (40,001 kJ/kg), SME (39,976 kJ/kg),CMME (39,786 kJ/kg), JCME (39,738 kJ/kg), CIME (39,513 kJ/kg) andfinally CoME (38,300 kJ/kg).

CME SFME PME CoME MOME CMME PaME

4.5281 6.3717 4.6889 3.1435 5.0735 4.0707 6.05671.7864 2.1954 1.7921 1.3116 1.9108 1.6781 1.82233.9212 5.5916 4.0284 2.705 4.3618 3.453 5.58480.866 0.8776 0.8591 0.8605 0.8597 0.8704 0.92217.08 1.46 23.56 8.01 12.64 0.71 0.022�10 2 12 �1 18 �4 �17�3 1 13 1 21 �3 <�33�9 2 15 �4 19 �2 <�33186.5 130.5 214.5 118.5 176 164 118.51a 1a 1a 1a 1a 1a 1a40,195 40,001 40,009 38,300 40,115 39,786 44,1800.0291 0.2911 0.0118 0.0114 0.022 0.028 0.3850.83 7.02 1.81 0.94 N/D N/D 77.10.041 0.057 0.05 0.035 0.046 0.041 391.1 87.9 89.1 92.3 90 91.1 01.4544 1.4557 1.4468 1.4357 1.4494 1.4569 1.5032236.9 174.4 203.6 230.8 206.7 276.3 61.8

7

12y = -6.8571x - 15.543x + 11.943

R² = 0.984310

15

A.E. Atabani et al. / Energy 58 (2013) 296e304302

3.3. Effect of biodiesel blending on some physical and chemicalproperties

3.3.1. Effect of CoME, PME and SFME blends on kinematic viscosityThe effect of blending CoME and PME with SFME on kinematic

viscosity has been studied. Blending of PME with SFME improvedthe viscosity of SFME from 6.3717 mm2/s to 6.0482 mm2/s (3:1),5.5995 mm2/s (1:1) and 5.3254 (1:3) respectively as can be seen inFig. 3a. Moreover, the following equation has been developed fromthis figure to predict the viscosity at any PME-SFME blends asfollow:

ViscosityðSFME�PMEÞ ¼ �0:5159x2�1:1195xþ6:35990� x� 100ðxh%PMEÞ�R2 ¼ 0:9908

�(2)

Blending of CoME with SFME improved the viscosity of SFMEfrom 6.3717 mm2/s to 5.3349 mm2/s (3:1), 4.4912 mm2/s (1:1) and3.879 (1:3) respectively as can be seen in Fig. 3b. Moreover, thefollowing equation has been developed from this figure to predictthe viscosity at any CoME-SFME blends as follow:

ViscosityðSFME�CoMEÞ ¼0:9533x2�4:1182xþ6:34570�x�100ðxh%CoMEÞ�R2 ¼ 0:9981

�(3)

3.3.2. Effect of CME, PME, JCME and CIME blends on cloud, pour andcold filter plugging point

The effect of blending of CME with PME, JCME and CIME on thecloud, pour and cold filter plugging point has been studied. As canbe seen in Figs. 4e6, blending of CME with PME, JCME and CIMEimproved the cold flow properties of PME, CIME and JCMErespectively. The developed equation in these figures can be used to

(a)

(b)

4.68895.3254

5.59956.0482

6.3717

y = -0.5159x - 1.1195x + 6.3599R² = 0.9908

0.0000

1.0000

2.0000

3.0000

4.0000

5.0000

6.0000

7.0000

8.0000

0% 25% 50% 75% 100%

Vis

cosi

ty a

t 40

˚C (

mm

²/s)

% of PME in the blends

3.1435

3.879

4.4912

5.3349

6.3717

y = 0.9533x - 4.1182x + 6.3457R² = 0.9981

0

1

2

3

4

5

6

7

0% 25% 50% 75% 100%

Vis

cosi

ty a

t 40

˚C (

mm

²/s)

% of CoMe in the blends

Fig. 3. Prediction of kinematic viscosity (a) PME and SFME blends (b) CoME and SFMEblends.

predict the CFPP, CP and PP of these blends at any blending ratiobased on CME% in the blends.

3.3.2.1. Prediction of CP, PP and CFPP of PME-CME blends.

CP ¼ 3:4286x2 � 20:629xþ 13:429 0 � x � 100

R2 ¼ 0:9704(4)

PP ¼ �2:2857x2 � 20:114xþ 14:114 0 � x � 100

R2 ¼ 0:9784(5)

CFPP ¼ �6:8571x2 � 15:543xþ 11:943 0 � x � 100

R2 ¼ 0:9843 (6)

3.3.2.2. Prediction of CP, PP and CFPP of JCME-CME blends.

CP ¼ �1:1429x2 � 12:857xþ 10:457 0 � x � 100

R2 ¼ 0:979(7)

PP ¼ �13:714x2 � 6:2857xþ 10:286 0 � x � 100

R2 ¼ 0:9785(8)

CFPP ¼ �6:8571x2 � 14:743xþ 10:543 0 � x � 100

R2 ¼ 0:9639 (9)

(a)

(b)

(c)

-10

-5

4

-15

-10

-5

0

5

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

CF

PP

(˚C

)˚C

)˚C

)

% of CME in the blends

-10-7

3

710

y = -6.8571x - 14.743x + 10.543R² = 0.9639

-15

-10

-5

0

5

10

15

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

CF

PP

(

% of CME in the blends

-10

-5

2

811

y = -5.7143x - 16.286x + 11.486R² = 0.9918

-15

-10

-5

0

5

10

15

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

CF

PP

(

% of CME in the blends

Fig. 4. Prediction of CFPP (a) CME and PME blends (b) CME and JCME blends (c) CMEand CIME blends.

(a)

(b)

(c)

-9

-1

47

15

y = -2.2857x - 20.114x + 14.114R² = 0.9784

-15

-10

-5

0

5

10

15

20

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

PP

(˚C

)˚C

)

% of CME in the blends

-9

-4

5

810y = -13.714x - 6.2857x + 10.286

R² = 0.9785

-15

-10

-5

0

5

10

15

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%PP

(

% of CME in the blends

-9

-1

5

1113

y = -13.714x - 8.6857x + 13.286R² = 0.9972

-15

-10

-5

0

5

10

15

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%PP

(˚C

)

% of CME in the blends

Fig. 6. Prediction of PP (a) CME and PME blends (b) CME and JCME blends (c) CME andCIME blends.

(a)

(b)

(c)

-3-2

5

9

13

y = 3.4286x - 20.629x + 13.429R² = 0.9704

-6-4-202468

10121416

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

CP

(˚C

)˚C

)˚C

)

% of CME in the blends

-3-1

4

8

10y = -1.1429x - 12.857x + 10.457

R² = 0.979

-6

-4

-2

0

2

4

6

8

10

12

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

CP

(

% of CME in the blends

-3

0

6

9

12

y = -3.4286x - 12.171x + 12.171R² = 0.9867

-6

-4

-2

0

2

4

6

8

10

12

14

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

CP

(

% of CME in the blends

Fig. 5. Prediction of CP (a) CME and PME blends (b) CME and JCME blends (c) CME andCIME blends.

A.E. Atabani et al. / Energy 58 (2013) 296e304 303

3.3.2.3. Prediction of CP, PP and CFPP of CIME-CME blends.

CP ¼ �3:4286x2 � 12:171xþ 12:171 0 � x � 100

R2 ¼ 0:9867 (10)

PP ¼ �13:714x2 � 8:6857xþ 13:286 0 � x � 100

R2 ¼ 0:9972 (11)

CFPP ¼ �5:7143x2 � 16:286xþ 11:486 0 � x � 100

R2 ¼ 0:9918 (12)

3.3.3. Sample calculationFrom Fig. 3 and based on Eq. (2) in Section 3.3.1, it can be seen

that the viscosity of PME and SFME blends can be predicted usingthe following equation:

Viscosity�SFME� PME

�¼ � 0:5159x2 � 1:1195xþ 6:3599

0 � x � 100�xh%PME

�

Therefore, the viscosity of PME-SFME biodiesel blend can bepredicted based on the above equation as follow:

Viscosityð25%PMEÞ ¼ � 0:5159*ð25%Þ^2� 1:1195*�25%

�

þ 6:3599 ¼ 6:0478 mm2=s

4. Conclusion

Production of biodiesel from non-edible oils may significantlyreduce the cost of biodiesel. In this study, production of biodieselwas performed using several edible and non-edible oils that areavailable in the South East Asian market. It has been reportedthat the yield of biodiesel from all these feedstocks was highenough to produce biodiesel in a practical way. Moreover, thephysical and chemical properties of these feedstocks were eval-uated. It was found that most of the properties of biodiesels arefollowing the standard specified by ASTM D 6751. In addition,this paper also suggests the concept of biodiesel blending inorder to improve their properties such as viscosity, flash, cloud,pour and cold filter plugging point. For instance, blending ofSFME and CoME improve the kinematic viscosity of pure SFMEfrom 6.3717 to 3.879 mm2/s. It was also found that blending hasimproved the cold flow properties of PME, CIME and JCMErespectively.

Acknowledgement

The authors would like to acknowledge the Ministry of HigherEducation of Malaysia and The University of Malaya, Kuala Lumpur,Malaysia for the financial support under UM.C/HIR/MOHE/ENG/06(D000006-16001). The authors would also like to thank the tech-nical staff at Faculty of Chemical Engineering & Natural Resources(FKKSA), University Malaysia Pahang for their valuable help andsupport.

A.E. Atabani et al. / Energy 58 (2013) 296e304304

References

[1] Chung K-H, Kim J Lee, Ki-Young. Biodiesel production by transesterification ofduck tallow with methanol on alkali catalysts. Biomass and Bioenergy2009;33(1):155e8.

[2] da Silva JPV, Serra TM, Gossmann M, Wolf CR, Meneghetti MR,Meneghetti SMP. Moringa oleifera oil: studies of characterization and bio-diesel production. Biomass and Bioenergy 2010;34(10):1527e30.

[3] Deka DC, Basumatary S. High quality biodiesel from yellow oleander (Thevetiaperuviana) seed oil. Biomass and Bioenergy 2011;35(5):1797e803.

[4] Rashid U, Anwar F, Knothe G. Biodiesel fromMilo (Thespesia populnea L.) seedoil. Biomass and Bioenergy 2011;35(9):4034e9.

[5] Teixeira da Silva de La Salles KM, Meneghetti SMP, Ferreira de La Salles W,Meneghetti MR, dos Santos ICF, da Silva JPV, de Carvalho SHV, et al. Charac-terization of Syagrus coronata (Mart.) Becc. oil and properties of methyl estersfor use as biodiesel. Industrial Crops and Products 2010;32(3):518e21.

[6] Amani MA, Davoudi MS, Tahvildari K, Nabavi SM, Davoudi MS. Biodieselproduction from Phoenix dactylifera as a new feedstock. Industrial Crops andProducts 2013;43(0):40e3.

[7] Fröhlich A, Rice B. Evaluation of Camelina sativa oil as a feedstock for biodieselproduction. Industrial Crops and Products 2005;21(1):25e31.

[8] Liu Y, Xin Hong-ling, Yan Yun-jun. Physicochemical properties of stillingia oil:feasibility for biodiesel production by enzyme transesterification. IndustrialCrops and Products 2009;30(3):431e6.

[9] Yang F-X, Su Yin-Quan, Li Xiu-Hong, Zhang Q, Sun Run-Cang. Preparation ofbiodiesel from Idesia polycarpa var. vestita fruit oil. Industrial Crops andProducts 2009;29(2e3):622e8.

[10] Kansedo J, Lee KT, Bhatia S. Cerbera odollam (sea mango) oil as a promisingnon-edible feedstock for biodiesel production. Fuel 2009;88(6):1148e50.

[11] Silitonga AS, Atabani AE, Mahlia TMI, Masjuki HH, Irfan Anjum Badruddin,Mekhilef S. A review on prospect of Jatropha curcas for biodiesel in Indonesia.Renewable and Sustainable Energy Reviews 2011;15:3733e56.

[12] Wang R, Zhou Wan-Wei, Hanna MA, Zhang Yu-Ping, Bhadury PS, Wang Y,et al. Biodiesel preparation, optimization, and fuel properties from non-ediblefeedstock, Datura stramonium L. Fuel 2012;91(1):182e6.

[13] Gui MM, Lee KT, Bhatia S. Feasibility of edible oil vs. non-edible oil vs. wasteedible oil as biodiesel feedstock. Energy 2008;33(11):1464e653.

[14] Atabani AE, Anjum Badruddin Irfan, Mahlia TMI, Masjuki HH, Chong WT,Keat Teong Lee. Investigation of physical and chemical properties ofpotential edible and non-edible feedstocks for biodiesel production, acomparative analysis. Renewable and Sustainable Energy Reviews2013;21(May):749e55.

[15] Atabani AE, Silitonga AS, Anjum Badruddin Irfan, Mahlia TMI, Masjuki HH,Mekhilef S. A comprehensive review on biodiesel as an alternative energyresource and its characteristics. Renewable and Sustainable Energy Reviews2012;16(4):2070e93.

[16] Moser BR, Vaughn SF. Coriander seed oil methyl esters as biodiesel fuel:unique fatty acid composition and excellent oxidative stability. Biomass andBioenergy 2010;34(4):550e8.

[17] Moser BR, Vaughn SF. Efficacy of fatty acid profile as a tool for screening feed-stocks for biodiesel production. Biomass and Bioenergy 2012;37(0):31e41.

[18] Atabani AE, Silitonga AS, Ong HC, Mahlia TMI, Masjuki HH, AnjumBadruddina Irfan, et al. Non-edible vegetable oils: a critical evaluation of oilextraction, fatty acid compositions, biodiesel production, characteristics, en-gine performance and emissions production. Renewable and SustainableEnergy Reviews 2013;18(February):211e45.

[19] Sanford SD, White JM, Shah PS, Wee C, Valverde MA, Meier GR. Feedstock andbiodiesel characteristics. Available from: http://www.regfuel.com/pdfs/Feedstock%20and%20Biodiesel%20Characteristics%20Report.pdf; 2011 [cited04.11.12].

[20] Karmakar A, Karmakar S, Mukherjee S. Properties of various plants andanimals feedstocks for biodiesel production. Bioresource Technology2010;101(19):7201e10.

[21] Oghenejoboh KM, Umukoro PO. Comparative analysis of fuel characteristics ofbio-diesel produced from selected oil-bearing seeds in Nigeria. EuropeanJournal of Scientific Research 2011;58(2):238e46.

[22] Balat M. Potential alternatives to edible oils for biodiesel productione a reviewof current work. Energy Conversion and Management 2011;52(2):1479e92.

[23] Rashid U, Anwar F, Moser BR, Knothe G. Moringa oleifera oil: a possible sourceof biodiesel. Bioresource Technology 2008;99(17):8175e9.

[24] Kafuku G, Lam MK, Kansedo J, Lee KT, Mbarawa M. Heterogeneous catalyzedbiodiesel production from Moringa oleifera oil. Fuel Processing Technology2010;91(11):1525e9.

[25] RashidU, Anwar F, AshrafM, SaleemM, Yusup S. Application of response surfacemethodology for optimizing transesterification ofMoringa oleifera oil: biodieselproduction. Energy Conversion and Management 2011;52(8e9):3034e42.

[26] Devan PK, Mahalakshmi NV. Study of the performance, emission andcombustion characteristics of a diesel engine using poon oil-based fuels. FuelProcessing Technology 2009;90(4):513e9.

[27] Aliyu B, Shitanda D, Walker S, Agnew B, Masheiti S, Atan R. Performance andexhaust emissions of a diesel engine fuelled with Croton megalocarpus(musine) methyl ester. Applied Thermal Engineering 2011;31(1):36e41.

[28] Kafuku G, Mbarawa M. Biodiesel production from Croton megalocarpus oiland its process optimization. Fuel 2010;89(9):2556e60.

[29] Sudradjat R. Calophyllum inophyllum L. a potential plant for biodiesel.Available from: http://www.iges.or.jp/en/bf/pdf/activity20090204/session2/Sudradjat.pdf; 2011 [cited 19.11.12].

[30] SathyaSelvabala V, Selvaraj DK, Kalimuthu J, Periyaraman PM, Subramanian S.Two-step biodiesel production from Calophyllum inophyllum oil: optimiza-tion of modified b-zeolite catalyzed pre-treatment. Bioresource Technology2011;102(2):1066e72.

[31] Sahoo PK, Das LM, Babu MKG, Naik SN. Biodiesel development from high acidvalue polanga seed oil and performance evaluation in a CI engine. Fuel2007;86(3):448e54.

[32] Sahoo PK, Das LM, Babu MKG, Arora P, Singh VP, Kumar NR, et al. Comparativeevaluation of performance and emission characteristics of jatropha, karanjaand polanga based biodiesel as fuel in a tractor engine. Fuel 2009;88(9):1698e707.