This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Industrial Crops and Products 34 (2011) 1010– 1016

Contents lists available at ScienceDirect

Industrial Crops and Products

jo ur nal homep age: www.elsev ier .com/ locate / indcrop

Optimizing mechanical oil extraction of Jatropha curcas L. seeds with respect topress capacity, oil recovery and energy efficiency

S. Karaj ∗, J. MüllerUniversität Hohenheim, Institute of Agricultural Engineering, Tropics and Subtropics Group, Garbenstr 9, 70599 Stuttgart, Germany

a r t i c l e i n f o

Article history:Received 18 January 2011Received in revised form 3 March 2011Accepted 4 March 2011Available online 27 April 2011

Keywords:Physic nutOilseedBiofuelScrew pressVegetable oil

a b s t r a c t

The objective of this study was to optimize the mechanical oil extraction of Jatropha curcas seeds byincreasing the efficiency of oil recovery and decreasing oil residues in press cake. The experiments werecarried out with mechanical screw press type – Komet D85-1G. Four setups were created by parametercombination of two different screws (16 and 21.5 mm choke ring size), with two different press cylinders(1 and 1.5 mm bore size), three different nozzles (8, 10 and 12 mm nozzle diameter) and three rotationalspeeds (low, medium and high). Oil recovery reduced when rotational speed increases for all setups;highest oil was 89.4% (m/m). The oil recovery was increasing when energy input increased and decreas-ing when seed material throughput increased. The relations between energy input and seed materialthroughput followed a strict pattern, which correlated with oil recovery. This correlation can be used fordetermining the optimal operation parameters.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Jatropha curcas L., belonging to the family Euphorbiaceae is ashrub or tree that is propagated by cuttings and is widely plantedas a hedge to protect fields from browsing animals (Heller, 1996;Sirisomboon et al., 2007). The plant produces toxic seeds and istherefore non-edible to animals and humans. Toxicity of seeds isdue to the presence of curcine and diterpenes, mainly phorbol ester(Heller, 1996; Makkar and Becker, 1997; Makkar et al., 1998, 2008;Jongschaap et al., 2007).

J. curcas can be utilized for different purposes: erosion con-trol, living fence, ornamental plant or source of fire wood. Thebark is rich in tannin and produces a dark blue dye. Leaves havebeen used for rearing of silkworm, in dyeing and in medicine asanti-inflammatory substance (Openshaw, 2000; Basha et al., 2009).Seeds have been used as insecticide, soap and varnish production.Seed cakes have been used as fertilizer, solid fuel or in biogas pro-duction. Non toxic varieties or detoxified press cake has been usedas feed for animal (Openshaw, 2000; Sirisomboon et al., 2007).Due to depletion of fossil fuels and the green-house effect, theapplication as biofuel is probably the most interesting from botheconomical and ecological points of view (Beerens, 2007). Lessawareness was raised to possible human health hazard by usingbiogenic fuels until it was found that strong mutagenic compoundscan be released from biofuel combustion (Krahl et al., 2009).

∗ Corresponding author. Tel.: +49 711 459 22840; fax: +49 711 459 23298.E-mail address: shkelqim karaj@uni-hohenheim.de (S. Karaj).

The physical, mechanical and chemical properties as well aspotential use of extracted oil from J. curcas as transesterified oil,or as a blend with diesel has been widely studied (Augustus et al.,2002; Pramanik, 2003; Narayana and Ramesh, 2006; Sirisomboonet al., 2007; Karaj and Müller, 2010; Kratzeisen and Müller, 2010).The calorific value and cetane number of J. curcas oil are similar todiesel, but the density and viscosity are much higher (Namasivayamet al., 2007).

Various methods for recovering oil from seeds have been inves-tigated (Shah et al., 2005; Lim et al., 2010; Qian et al., 2010).Chemical extraction (solvent and enzymatic method) of J. cur-cas seeds has been studied and the optimal parameters for oilextraction such as solvent type, temperature range, solvent to solidratio, processing time and particle size have also been considered(Winkler, 1997; Shah et al., 2004, 2005; Hawash et al., 2008; Sayyaret al., 2009).

The mechanical oil extraction of J. curcas was reported as sub-optimal due to lack of knowledge about best operation parameters(Shah et al., 2005). Openshaw (2000) has reported the use of sun-flower seed mechanical screw presses for extracting J. curcas seedoil as unsuitable due to technical problems and low oil recovery.Beerens (2007) has reported results from using two mechanicalscrew presses: a mechanical cylinder press (BT50) and a strainerpress (Sayari). Maximum oil recovery was reported to be 79% forBT50 screw press and 87% for Sayari strainer press after dual pass-ing. Concerning the optimization of oil extraction efficiency, neitherthe influence of different settings of the screw press nor the result-ing dependent factors were reported in other studies. Therefore,this study was aiming to analyse different designed variable such

0926-6690/$ – see front matter © 2011 Elsevier B.V. All rights reserved.doi:10.1016/j.indcrop.2011.03.009

Author's personal copy

S. Karaj, J. Müller / Industrial Crops and Products 34 (2011) 1010– 1016 1011

as screw press, press cylinder, nozzle size and rotational speed tooptimize the mechanical oil extraction of J. curcas seeds by increas-ing the efficiency of oil recovery with a mechanical cylinder screwpress (Komet D85-1G).

2. Materials and methods

2.1. J. curcas seed material

For this study, dried J. curcas seeds were imported fromJaipur, Rajasthan (India). The place of origin was characterized byannual precipitation of 1000–1200 mm and a temperature rangeof 15–35 ◦C. The plantation was 11–12 years old and its annualyield was calculated to be around 3500 kg/ha. Seeds were har-vested manually during November–December 2007 and stored injute bags in a warehouse facility at temperatures between 14 ◦Cand 30 ◦C. After transport to Germany, J. curcas seeds were cleanedwith a pneumatic separator to remove impurities and then storedat room temperature at a range of 18–22 ◦C, attaining a moisturecontent of 8.3% (wet base) and oil content of about 36%. Moisturecontent of seeds was not altered and unshelled seeds were used foroil extraction optimization.

2.2. Mechanical screw press Komet D85-1G

The experiments were carried out with mechanical screwpress Komet D85-1G (IBG Monforts Oekotec, Mönchengladbach,Germany) powered by a 3.0 kW electric motor. Maximal capacityin terms of seed throughput was 25 kg/h. Different sensors wereinstalled for measuring direct and indirect parameters such as, tem-peratures, pressure, torque and rotational speed of the screw andoil recovery (Fig. 1).

Oil was collected and weighed after each experiment at oilcollection point (XI) into light-proof plastic containers. The tem-perature sensors were installed in five different positions (T1–T5).Body temperature of the mechanical press was measured alongthe screw shaft via thermocouples type K at point T1, T2 and T3.Temperature of extracted oil was measured by thermocouple justafter passing out of the press cylinder holes at point T4. Press caketemperature was measured 35 mm behind the nozzle outlet withan infrared sensor (Raytek TX, Berlin, Germany) at point T5, andthe whole system was monitored by an infrared thermal camera(Thermal Imager Fluke Ti20, Eindhoven, The Netherlands).

A mold cavity pressure sensor (p) (6157BB, Kistler Instruments,Winterthur, Switzerland) was installed close to the compres-sion zone (X) where the highest temperature and pressure wasexpected during operation. Screw speed was adjusted between0 and 600 rpm by a variable speed alternator (XIV), and a torquetransducer sensor (E) (BRBK-200-n, ETH Messtechnik, Gschwend,Germany) was installed for measuring rotational speed (�) andtorque (�) between motor and screw press cylinder via a couplingdevice (XII).

2.3. Independent and dependent variables

A number of design parameters were considered in this study foroptimizing oil extraction using the mechanical screw press (Fig. 2).These included: (i) two different screws with choke worm shaft ringsize 16 and 21.5 mm, labelled R8 and R11, (ii) two different presscylinders with mesh size 1 and 1.5 mm, labelled P1 and P1.5, and(iii) three different nozzles with restriction size 8, 10 and 12 mm,labelled N8, N10 and N12.

The rotational speed of the screw was adjusted on three lev-els; low, medium and high speed, namely 220, 290 and 355 rpmfor screw R8 and 115, 180 and 255 rpm for screw R11. The speed

limit range for different screws was determined through prelimi-nary tests. Screw R11 allowed for the application of lower speedsthan screw R8 without jamming during pressing. Therefore, thespeed levels for screw press R11 were set lower than for screwpress R8. The dependent variables such as oil content, oil recov-ery, system temperatures and time required for pressing as well astorque and pressure in the pressing chamber were monitored viaweight measurements and sensor control (Table 1). Data was col-lected by a data acquisition switch unit (34970A Agilent, HewlettPackard, Palo Alto, CA) and transferred via BenchLink Dataloggersoftware to a laboratory computer.

Oil recovery efficiency was the most important dependent vari-able considered in this study. The effects of independent variableson oil recovery optimization could be explained by consideringthe dependent variables. For increasing the extraction efficiencyincreased temperature is beneficial. Since friction inside the presscylinders generates heat which is passed in the oil, reducing the vis-cosity of the crushed seed material and facilitating oil brake point,which is the applied pressure for deliberating oil from crushedseed material (Sukumaran and Singh, 1989). For optimization of oilrecovery, pressure is the most interesting variable to monitor. Pres-sure is expected to alter when independent variables are changingand a general hypothesis is that higher pressure will lead to highertemperature generation and higher oil recovery efficiency (Willemset al., 2008, 2009).

Oil recovery efficiency was calculated by comparing the oil con-tent in the press cake to the initial oil content in the seeds extractedby Soxhlet apparatus, three replicates respectively, and using thefollowing formula (Beerens, 2007):

O� = 1 −[

OC/(1 − OC )OS/(1 − OS)

]· 100 (1)

with O� as oil recovery efficiency in % (m/m), OC as oil content inpress cake and OS as oil content in seed, in g (m/m).

Throughput was calculated by dividing the seed sample weightby the time required for pressing the sample as follows:

TP = S

t(2)

with TP throughput in kg/h, S sample size in kg, t time in h.In order to estimate the energy input of each extraction process,

power was calculated as follows:

P = � · ω (3)

with P as power in W, � as torque in Nm and ω as angular velocity.The specific energy input per kg seed processed (ES) in kWh kg−1

was calculated as ratio of total energy (Et) in kWh to sample size(S) in kg:

ES = Et

S(4)

where Et = P·t, time (t) in hour.

2.4. Experimental procedure

Four different experiment setups were carried out with combi-nations of different design parameters in order to determine theeffect of independent variables on dependent variables (Table 2).

In Setup 1, screw and press cylinder sizes (R8 P1) were kept con-stant, whereas nozzle was varied in combination with speed. Theconstant parameters for the other setups were R8 P1.5, R11 P1 andR11 P1.5 respectively for Setups 2, 3 and 4, where the same varia-tion of nozzle was applied through all setups. Speed was altered onthree levels from 220, 290 and 355 rpm on first and second setup,whereas on third and fourth setup speed levels were 115, 180 and255 rpm, respectively.

Author's personal copy

1012 S. Karaj, J. Müller / Industrial Crops and Products 34 (2011) 1010– 1016

II III IV V VI VII

T1T2 T3

T4

T5

VIII

IXXXI

XIII

XIV

p

E

I

XII

Fig. 1. Mechanical screw press for oil extraction and installed sensors, (I) feeding container, (II) feeding hopper, (III) housing, (IV) screw, (V) press cylinder with oil outlet holes,(VI) heating, (VII) nozzle, (VIII) press cake, (IX) press head, (X) compression zone, (XI) oil collector, (XII) coupling, (XIII) motor, (XIV) speed alternator, (T1–T5) temperaturesensors, (p) pressure sensor, (E) torque transducer sensor.

Fig. 2. Independent variables of mechanical screw press Komet D85-1G.

A batch of 10 ± 0.01 kg of unshelled J. curcas seeds with average(X, Y, Z) dimension size of 18. 4, 11.2 and 9.1 mm, length, widthand breadth, respectively was used for each experiment. Before thepressing process was initiated, the press head (the casing aroundthe compression zone and the nozzle, Fig. 1) was heated with anelectrical device to 90 ◦C. This was done to prevent the press fromjamming at the beginning of the operation (Ferchau, 2000; Beerens,2007). During preparatory heating, 5 kg of seeds were fed into thehopper and all sensors were installed (as illustrated in Fig. 1). Afterthe temperature reached 90 ◦C on the press head, the mechanical

screw press was operated with the adjusted speed (according tothe setup) and the electrical heater was switched off. The 5 kg ofseeds were pressed until all sensors were stabilized at given designparameters (such as temperatures, pressure, torque and rotationalspeed) and then the precise sample of 10 kg was fed into the hopper.The oil collection point was manually cleaned before the pressingof the 10 kg sample took place. The oil was collected in light-proofplastic containers and the press cake was collected in a plastic crate.

The raw oil and press cake were weighed after each experimentand taken to the laboratory for additional analyses. The data col-

Table 1Dependent variables for optimization of oil extraction.

Variable Symbol Unit Method Manufacturer

Oil content in seeds OS % Soxhlet (DGF-Einheitsmethoden, 2006)Temperature 1 T1 ◦C Thermocouple Type K Greisinger Electronic GmbHTemperature 2 T2 ◦C Thermocouple Type K Greisinger Electronic GmbHTemperature 3 T3 ◦C Thermocouple Type K Greisinger Electronic GmbHTemperature 4 T4 ◦C Thermocouple Type K Greisinger Electronic GmbHTemperature 5 T5 ◦C Infrared temperature sensor Raytek TXTime t h Computer timeTorque � Nm Torque transducer, BRBK-200-n ETH messtechnik GmbHPressure p bar Mold pressure sensor, 6157BB Kistler Instrument, AG

Author's personal copy

S. Karaj, J. Müller / Industrial Crops and Products 34 (2011) 1010– 1016 1013

Table 2Experimental setup: R8 and R11 screw press with 16 and 21.5 mm choke ring; P1and P1.5 press cylinders with 1 and 1.5 mm hole size; N8, N10 and N12 nozzles with8, 10 and 12 mm restriction size.

Setup 1 Setup 2 Setup 3 Setup 4R8 P1 R8 P1.5 R11 P1 R11 P1.5

N8 �220 N8 �220 N8 �115 N8 �115N8 �290 N8 �290 N8 �180 N8 �180N8 �355 N8 �355 N8 �255 N8 �255N10 �220 N10 �220 N10 �115 N10 �115N10 �290 N10 �290 N10 �180 N10 �180N10 �355 N10 �355 N10 �255 N10 �255N12 �220 N12 �220 N12 �115 N12 �115N12 �290 N12 �290 N12 �180 N12 �180N12 �355 N12 �355 N12 �255 N12 �255

� – rotational speed in rpm.

lected were statistically analysed using standard statistics such asmean values, median, and interquartile and regression analyses byOrigin Pro-8 software.

3. Results and discussion

3.1. Temperature generation during pressing process

Fig. 3 shows a typical temperature pattern on the mechanicalscrew press during pressing of J. curcas seeds. The screw press bodytemperature rises from point T1 (35 ◦C) to T3 (up to 100 ◦C).

Since T3 was located at the compression zone where the resis-tance of seed material against the press head was the highest,temperature recorded at this point was also the highest. The outsidesurface temperatures of the press head was detected up to 70 ◦C,whereas in bore holes temperatures up to 100 ◦C were detected bythermography.

The temperatures generated at all measuring points when usingdifferent screw presses such as R8 or R11 are depicted in Fig. 4.

The highest temperature was generated by using screw pressR8 rather than R11 in both graphics; this was due to the differ-ence in screw design. Screw press R8 had a smaller worm shaft sizeresulting in higher resistance of seeds against the screw press andgenerating more heat than screw press R11. The graphics confirmthat highest temperature was generated at compression zone (T3)for all analysed setups.

Oil temperature (T4) and press cake temperature (T5) dependedon temperature created at the compression zone (T3) and relateddependency was found as depicted in Fig. 5.

Fig. 5 shows that when temperature at the compression zone(T3) rises the temperature of oil (T4) rises as well. The statisticalanalyses for oil temperature (T4) reveal a good correlation to tem-perature at the compression zone (T3) that can be expressed as an

Fig. 3. Temperature pattern of the screw press (Thermal Imager Fluke Ti20).

Fig. 4. Temperature at different points (T1–T5) on four setups when using twodifferent screws (R8, R11) and two different press cylinders (P1, P1.5); error barsindicate min and max values of each temperature measuring point, and boxes depictmedian, mean values and interquartile range from 25-th to 75-th percentiles.

exponential function (R2 = 0.651, p ≤ 0.01):

T4 = 0.011 · e(T3/14.7) + 64.39 (5)

Also temperature of press cake (T5) increases simultaneouslywith the temperature at the compression zone (T3). However, herethe dependency follows a linear correlation (R2 = 0.801, p ≤ 0.01):

T5 = 1.13 · T3 − 40.11 (6)

Since the temperature of oil and press cake was difficult tobe measured directly via thermocouples, due to external environ-mental factors (conventions and irradiation) a reference measuringpoint has to be introduced, where the external environmental fac-tors are minimal. The reference measuring point should estimateoil temperature and press cake temperatures based on temper-ature values measured. It was observed that temperature at thebore holes of press head (Fig. 3) was not affected from wind con-vection and irradiation meaning constant temperature through allpressing conditions. These points might be suitable for indirect esti-mation of oil and press cake temperatures based on model 5 and 6.In order to prevent oil deterioration (acid value, phosphorous, cal-cium, magnesium, oxidation stability and water content) and presscake protein denaturation by overheating, a temperature controlsystem of oil and press cake temperature should be implementedon the mechanical screw press.

3.2. Oil recovery optimization

The dependent variables such as oil recovery efficiency, oilresidue in press cake, seed material throughput, specific energyinput, torque and pressure generated are given in Table 3. It was

Author's personal copy

1014 S. Karaj, J. Müller / Industrial Crops and Products 34 (2011) 1010– 1016

Table 3Oil recovery efficiency, oil residues in press cake, seed material throughput, specific energy, torque and pressure generated during the process of oil extraction.

Setup name Oil recoverya

O� , % (m/m)Oil residuesa

OC , % (m/m)Throughput TP, kg/h Energya ES , kWh/kg Torque �, Nm Pressure p, bar

S. 1 (R8 P1)b

N8 �220 84.2 8.1 3.79 0.377 58.9 ± 4.8 78.0 ± 7.8N8 �290 83.9 8.3 4.72 0.393 59.0 ± 4.6 62.9 ± 8.0N8 �355 81.2 9.6 5.56 0.425 60.6 ± 5.6 47.5 ± 11.3N10 �220 83.9 8.3 3.78 0.405 61.7 ± 4.9 67.6 ± 5.9N10 �290 82.7 8.8 4.96 0.360 58.3 ± 4.9 46.1 ± 4.8N10 �355 81.4 9.4 6.06 0.351 56.9 ± 5.2 43.7 ± 4.3N12 �220 84.8 7.9 3.85 0.416 70.0 ± 5.1 49.5 ± 6.8N12 �290 83.6 8.0 4.63 0.431 72.7 ± 5.5 58.6 ± 18.4N12 �355 81.8 9.3 5.99 0.412 64.0 ± 6.0 46.8 ± 19.2S. 2 (R8 P1.5)b

N8 �220 89.3 5.6 4.00 0.462 79.4 ± 4.4 71.0 ± 19.5N8 �290 86.8 6.9 5.19 0.445 76.4 ± 4.9 64.4 ± 17.0N8 �355 84.8 7.8 6.27 0.446 75.6 ± 4.7 58.4 ± 9.5N10 �220 85.7 7.4 3.94 0.460 77.0 ± 4.8 84.7 ± 14.9N10 �290 84.9 7.8 5.13 0.432 72.6 ± 4.7 75.5 ± 13.2N10 �355 83.5 8.5 6.28 0.443 74.8 ± 4.6 61.8 ± 8.2N12 �220 84.6 8.0 4.08 0.445 74.9 ± 5.0 57.3 ± 7.6N12 �290 83.8 8.3 5.18 0.423 71.1 ± 4.9 59.1 ± 11.3N12 �355 82.5 8.9 6.35 0.409 68.3 ± 4.7 36.6 ± 5.4S. 3 (R11 P1)b

N8 �115 84.5 8.0 5.36 0.293 131.5 ± 15.3 85.1 ± 21.5N8 �180 76.3 11.8 8.55 0.284 126.0 ± 4.9 79.7 ± 15.8N8 �255 69.8 14.5 11.65 0.271 119.1 ± 5.7 79.2 ± 17.6N10 �115 81.3 9.5 5.80 0.255 119.9 ± 7.5 42.4 ± 14.1N10 �180 73.7 12.9 8.94 0.262 114.1 ± 6.2 38.8 ± 13.6N10 �255 82.1 9.1 11.61 0.247 105.5 ± 7.7 54.7 ± 7.4N12 �115 76.6 11.6 5.99 0.265 131.6 ± 5.3 67.3 ± 18.0N12 �180 63.2 17.1 10.19 0.251 122.4 ± 4.9 67.9 ± 12.4N12 �255 51.5 21.4 12.96 0.222 105.4 ± 7.3 34.5 ± 8.0S. 4 (R11 P1.5)b

N8 �115 83.9 8.3 6.11 0.278 141.3 ± 4.8 75.2 ± 22.1N8 �180 78.4 10.8 9.56 0.261 131.1 ± 7.0 51.6 ± 13.6N8 �255 69.8 14.5 13.39 0.255 127.6 ± 3.9 33.9 ± 11.5N10 �115 76.7 11.6 6.77 0.245 138.3 ± 5.3 40.9 ± 9.9N10 �180 66.6 15.8 11.88 0.237 131.1 ± 6.0 37.3 ± 12.6N10 �255 32.2 27.6 14.33 0.201 108.9 ± 9.3 35.5 ± 7.0N12 �115 65.8 16.1 7.03 0.228 133.8 ± 5.5 44.4 ± 9.7N12 �180 55.8 19.9 11.56 0.202 116.2 ± 6.8 37.3 ± 6.0N12 �255 40.0 25.2 14.49 0.158 84.6 ± 8.2 20.4 ± 4.4

a O� , OC and ES were smaller than ±5% STD.b R – screw, P – press cylinder.

observed that screw R8 (Setups 1 and 2) shows higher oil recoverythan screw R11 (Setups 3 and 4) even though R11 was operatedwith lower rotational speed than R8. The highest oil recovery of89.4% (m/m) was achieved by using Setup 2 (R8 P1.5 N8 �220) andthe lowest oil recovery efficiency of about 32.2% (m/m) by usingSetup 4 (R11 P1.5 N10 �255). Oil residue in press cake was higherin Setups 3 and 4 than in 1 and 2.

Table 3 shows that screw press R8 (Setups 1 and 2) requires morespecific energy than screw press R11. Lower energy input meanslower oil recovery efficiency, higher oil residue in press cake andhigher seed material throughput.

Torque and pressure showed similar trends within the setups,mainly that when screw speed increased torque, pressuredecreased. Torque generated during oil extraction was higher forscrew press R11 (Setups 3 and 4) than screw press R8 (Setups1 and 2). Torque is related with resistance of seeds on screwpress against press cylinder. Since more materials were pro-cessed with screw press R11 (due to choke ring dimensions)than screw press R8 the resistance created and resulting torquewas expected to be higher. It was observed that torque wasdecreasing when rotational speed increases within setups, but thistrend was not proven to be the same for Setup 1. This mightbe explained with the fact that higher speed means higher tem-perature and lower viscosity results in lower resistance of seedmaterial.

The radial pressure generated during pressing reduced with theincrease of rotational speed and with increase of nozzle size. Theo-retically, higher screw speed means more seed material throughputand higher oil content residual in press cake since less time is avail-able for the oil to drain from the solids. At higher speed the viscositythus remains lower resulting in less pressure build-up and more oilcontent in press cake (Beerens, 2007; Willems et al., 2008, 2009).

Fig. 6 shows the correlation of specific energy input and oilrecovery efficiency versus material throughput. The statistical anal-yses for specific energy input (ES) and oil recovery efficiency (O�)revealed a high correlation to seed material throughput (TP) thatcan be expressed as linear function (R2 = 0.729) and exponentialfunction (R2 = 0.778), respectively:

ES = 0.5124 − 0.02362 · TP (7)

O� = 85.834 − 0.602 · e(TP/3.345) (8)

For each added unit of material throughput the specific energyinput and oil recovery efficiency would decrease according to thelinear function (7) and exponential function (8). The model could beused only for the given throughput data (3–16 kg/h) and mechan-ical screw press (Komet D85-1G). The highest oil recovery wasachieved with the lowest material throughput, for instance at 4 kg/hthroughput the oil recovery efficiency was around 90%. On theother hand lower throughput leads to highest specific energy input,

Author's personal copy

S. Karaj, J. Müller / Industrial Crops and Products 34 (2011) 1010– 1016 1015

Fig. 5. Correlation between oil temperature (T4) and press cake temperature (T5)vs. press head temperature (T3) on all analysed setups.

Fig. 6. Specific energy input and oil recovery efficiency vs. seed material throughput.

0 2 4 6 8 10 12 14 160.0

0.5

1.0

1.5

2.0

2.5

3.0

Oil production

Oil

Pro

duct

ion

OP,

kgoi

l/h

Throughput TP, kgseed /h

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Spe

cific

Ene

rgy

Eη

, kW

h kg

oil-1

Energy Efficiency

Fig. 7. Oil production capacity and specific energy vs. throughput of seeds.

where 4 kg/h throughput required around 0.40 kWh/kg specificenergy.

The optimal operation point can either be based on maximizingoil production capacity or specific pressing energy. Fig. 7 shows theoil production per unit of time based on oil recovery efficiency (Eq.(8)) and specific energy (Eq. (7)) versus seed material throughput.

The oil production capacity increases with increase of seedmaterial throughput and the maximal oil production capacity of2.76 kgoil/h was achieved at a throughput of 11 kgseed/h. It wasobserved that additional increase of throughput leads to decreaseof oil production capacity. This was because oil recovery efficiency(Eq. (8), Fig. 7) is rapidly decreasing when throughput furtherincreases.

The specific pressing energy which is the ratio of oil produc-tion and energy input needed for pressing reached the maximumreturn of 1.4 kgoil/kWh electrical energy also at a throughput of11 kgseed/h. The same trend was observed for energy efficiency,where additional increase of seed material throughput leads todecrease of energy efficiency.

4. Conclusions

The temperature generated on mechanical screw press was cre-ated due to resistance of seed material against the press head. Oiltemperature depended on press cylinder and press head tempera-tures; whereas press cake temperature was only influenced by heatgenerated from the press head. Temperatures up to 140 ◦C wererecorded and highest temperature was generated by using screwR8 rather than R11.

Oil recovery decreased and oil content in press cake and seedmaterial throughput was increased when rotational speed of screwpress was higher. Oil extraction with screw press Komet D85-1Gwas most efficient by using a screw with 16 mm choke ring size, apress cylinder with 1.5 mm bore size and a nozzle with 8 mm diam-eter in contrast to other combinations. Specific energy input and oilrecovery are correlated to seed material throughput. Throughputcan be adjusted by rotational speed of the screw to maximise eitheroil production or energy efficiency, in which fortunately both max-ima where close together. The optimal operation point for screwpress Komet D85-1G was at 11 kg/h Jatropha seed throughput,resulting in a press capacity of 2.67 kg/h in terms of recovered oil.The oil extraction efficiency was reaching study objectives. Futurestudies should focus on analysing different varieties, cultivationorigins and production practices as well as temperature influenceon oil and press cake quality indentifying threshold temperaturewhere oil and protein deterioration does not occur.

Author's personal copy

1016 S. Karaj, J. Müller / Industrial Crops and Products 34 (2011) 1010– 1016

Acknowledgements

The authors are grateful to the Bundesministerium für Bildungund Forschung, Berlin (Germany) for financial support of Project0330799A, Optimization of de-shelling and oil extraction of Jat-ropha curcas L. for direct use in plant oil stoves (01.06.2007 to30.06.2010).

References

Augustus, G.D.P.S., Jayabalan, M., Seiler, G.J., 2002. Evaluation and bioinduction ofenergy components of Jatropha curcas. Biomass Bioenerg. 23, 161–164.

Basha, S.D., Francis, G., Makkar, H.P.S., Becker, K., Sujatha, M., 2009. A comparativestudy of biochemical traits and molecular markers for assessment of geneticrelationships between Jatropha curcas L. germplasm from different countries.Plant Sci. 176, 812–823.

Beerens, P., 2007. Screw-pressing of Jatropha seeds for fuelling purposes in lessdeveloped countries. M.Sc. Thesis, Department of Sustainable Energy Technol-ogy, Eindhoven University of Technology, Eindhoven, p. 87.

DGF-Einheitsmethoden, 2006. Deutsche Einheitsmethoden zur Untersuchung vonFetten, Fettprodukten, Tensiden und verwandten Stoffen, Stand: 2.

Ferchau, E., 2000. Equipment for Decentralised Cold Pressing of Oil Seeds. Folkecentrfor Renewable Energy, Denmark.

Hawash, S., Kamal, N., El Diwani, G., 2008. Egyptian Jatropha oil extraction forbiodiesel production. Afinidad 65, 470–474.

Heller, J., 1996. Physic nut. Jatropha curcas L. Promoting the conservation anduse of underutilized and neglected crops. Institute of Plant Genetics and CropPlant Research, Gatersleben/International Plant Genetic Resources Institute,Rome.

Jongschaap, R.E.E., Corré, W., Bindraban, P., Brandenburg, W., 2007. Claims and Factson Jatropha curcas L. Plant Research International Wageningen, Wageningen, TheNetherlands, p. 42.

Karaj, S., Müller, J., 2010. Determination of physical, mechanical and chemicalproperties of seeds and kernels of Jatropha curcas L. Ind. Crops Prod. 32,129–138.

Krahl, J., Knothe, G., Munack, A., Ruschel, Y., Schröder, O., Hallier, E., Westphal, G.,Bünger, J., 2009. Comparison of exhaust emissions and their mutagenicity fromthe combustion of biodiesel, vegetable oil, gas-to-liquid and petrodiesel fuels.Fuel 88, 1064–1069.

Kratzeisen, M., Müller, J., 2010. Prediction of deposit formation during combustionof Jatropha oil from standard quality parameters. Fuel 89, 2769–2774.

Lim, S., Hoong, S.S., Teong, L.K., Bhatia, S., 2010. Supercritical fluid reactive extractionof Jatropha curcas L. seeds with methanol: a novel biodiesel production method.Bioresour. Technol. 102, 7169–7172.

Makkar, H.P., Becker, K., 1997. Potential of J. curcas seed meal as protein supplementto livestock feed, constraints to its utilisation and possible strategies to overcomeconstraints. In: Jatropha 97, February 23–27, 1997 , Managua, Nicaragua.

Makkar, H.P., Becker, K., Schmook, B., 1998. Edible provenances of Jatropha curcasfrom Quintana Roo state of Mexico and effect of roasting on antinutrient andtoxic factors in seeds. Plant Foods Hum. Nutr. 52, 31–36.

Makkar, H.P., Francis, G., Becker, K., 2008. Protein concentrate from Jatropha cur-cas screw-pressed seed cake and toxic and antinutritional factors in proteinconcentrate. J. Sci. Food Agric. 88, 1542–1548.

Namasivayam, C., Sangeetha, D., Gunasekaran, R., 2007. Removal of anions, heavymetals, organics and dyes from water by adsorption onto a new activated carbonfrom Jatropha husk, an agro-industrial solid waste. Process Saf. Environ. Prot. 85,181–184.

Narayana, J., Ramesh, A., 2006. Parametric studies for improving the performance of aJatropha oil-fuelled compression ignition engine. Renew. Energ. 31, 1994–2016.

Openshaw, K., 2000. A review of Jatropha curcas: an oil plant of unfulfilled promise.Biomass Bioenerg. 19, 1–15.

Pramanik, K., 2003. Properties and use of Jatropha curcas oil and diesel fuel blendsin compression ignition engine. Renew. Energ. 28, 239–248.

Qian, J., Shi, H., Yun, Z., 2010. Preparation of biodiesel from Jatropha curcas L. oilproduced by two-phase solvent extraction. Bioresour. Technol. 101, 7025–7031.

Sayyar, S., Abidin, Z.Z., Yunus, R., Muhammad, A., 2009. Extraction of oil from Jat-ropha seeds-optimization and kinetics. Am. J. Appl. Sci. 6, 1390–1395.

Shah, S., Sharma, A., Gupta, M.N., 2004. Extraction of oil from Jatropha curcas L.seed kernels by enzyme assisted three phase partitioning. Ind. Crops Prod. 20,275–279.

Shah, S., Sharma, A., Gupta, M.N., 2005. Extraction of oil from Jatropha curcas L. seedkernels by combination of ultrasonication and aqueous enzymatic oil extraction.Bioresour. Technol. 96, 121–123.

Sirisomboon, P., Kitchaiya, P., Pholpho, T., Mahuttanyavanitch, W., 2007. Physicaland mechanical properties of Jatropha curcas L. fruits, nuts and kernels. Biosyst.Eng. 97, 201–207.

Sukumaran, C.R., Singh, B.P.N., 1989. Compression of a bed of rapeseeds: the oil-point. J. Agric. Eng. Res. 42, 77–84.

Willems, P., Kuipers, N.J.M., De Haan, A.B., 2008. Hydraulic pressing of oilseeds:experimental determination and modeling of yield and pressing rates. J. FoodEng. 89, 8–16.

Willems, P., Kuipers, N.J.M, de Haan, A.B., 2009. A consolidation based extrudermodel to explore GAME process configurations. J. Food Eng. 90, 238–245.

Winkler, E., 1997. Enzyme-supported oil extraction from Jatropha curcas seeds. Appl.Biochem. Biotechnol. 63–65, 449–456.