ULTRASOUND-ASSISTED EXTRACTION OF SUNFLOWER OIL FROM THE CAKE AFTER SUNFLOWER SEED PRESSING

The influence of ultrasound application on the extraction (UE) of the cake after sunflower seed pressing was studied. Three different solvents were used for extraction in the Soxhlet apparatus: n-hexane, petroleum ether and extraction petrol. Petrol for extraction showed the highest yield. It has been shown that the using of ultrasound improved the extraction kinetics in the initial period – washing, and in the second period – the diffusion of oil from the mass of the cake. The parameters of the non-stationary diffusion model with and without the use of ultrasound were k' = 0.039 min, k' = 0.026 min and b' = 0.713, i.e. b' = 0.589, respectively. Oil yield was also slightly higher in the ultrasonically supported extraction.


Introduction
The production of sunflower oil is one of the most characteristic examples of the application of extraction on an industrial level.Therefore, this process is well developed and is the subject of research in many studies (Baümler et al., 2016;Rai et al., 2016).The production of sunflower oil involves several stages: separating the grain from impurities, moistening the grain in the water vapor for easier separation of the shell, pressing, extracting the residual oil from the cake, and final treatment of the oil (dearomatization and decolorization).The cake, after extraction, contains a certain amount of residual oil and as such is used for the animal food production.Preferably, the oil content of the cake after extraction at an acceptable time interval is as low as possible.
The results presented in this study are preliminary research results obtained within the framework of the project for improving the efficiency of the technological process of extracting sunflower oil in the oil company "Plima" in Kruševac.The goal of the project was to find out the rationality of the adaptation of the existing equipment.

Materials
As an extraction material in this study, a cakeof the mechanically pressed sunflower seeds was used.The crushing of the cake was done in an electric grinder with rotary blades (15 000 rpm.).The resulting particles were then separated through standard sieves.For the extraction, a fraction from 0.4 mm to 0.5 mm was used.The content of moisture in the samples is given in Table 1.The n-hexane, petrol ether and high-purity petrol were used for the extraction.Extraction in a Soxhlet apparatus (SA) In this study, two extraction techniques were applied.The ultrasound-assisted extraction in a modified Soxhlet apparatus (SA) (Luque-Garcıa and Luque de Castro, 2004), as shown in Figure 1, then a classical batch reactor for extraction with stirring, and an ultrasonic generator (BE) were used in this research.
Ultrasound-assisted extraction of sunflower oil from the cake after sunflower seed pressing 197 The (SA) technique with and without the use of ultrasound was used to determine the maximum yield.The technique (BE) with and without ultrasound was used to monitor the extraction kinetics.For the (SA) technique, a sample of 10 g and 100 mL of the appropriate solvent was used.The extraction time was 6 hours (24 cycles).After the final extraction, the solvent was eliminated in a vacuum evaporator at 50°C.

Extraction in the batch reactor
The extraction kinetics in this study was investigated in a classical batch reactor with stirring and the ultrasonically assisted extraction in the batch reactor.In both cases, high-purity petrol was used as a solvent.A 1:10 hydromodule was applied, and a sample of 5 g of the sunflower cake was placed in an Erlenmeyer of 100 mL with 50 mL of the solvent.The set-up consisted of an ultrasonic cleaner (EI, Niš, Serbia; total nominal power: 2 x 50 W; and internal dimensions: 300 x 220 x 155 mm), operating at 40 kHz frequency.The Erlenmeyer was then immersed in an ultrasonic bath filled with water up to 1/3 height.The operating temperature was constant at 25 o C (±0.2 o C).The extraction times were 2.5 min, 5 min, 10 min, 20 min, 40 min and 60 min.After completion of the extraction, the solvent was removed in a vacuum evaporator at 50°C.

Results and Discussion
Extraction in the Soxhlet apparatus (SA) The results of the sunflower oil extraction from the pressed sunflower seeds are shown in Table 2 and Figure 2.There was a positive contribution to the use of ultrasound.Compared to n-hexane and petrol ether, petrol showed the best yield, both without and with ultrasound.The contribution of ultrasound to all samples ranged from 1% to 1.5%.A large number of models can be found in the literature describing solid-liquid extraction.However, the kinetics of the extraction process of substances from raw materials has most often been modeled using the unsteady diffusion through plant material (Stanković et al., 1994), the film theory (Pekić et al., 1988) and the empirical equation of Ponomaryov (1976).Some of them are shown in Table 4.   Kitanović et al. (2008) In this paper, a model based on the unsteady diffusion (non-stationary diffusion model) is used, which has the following form: where: b' is the washing coefficient, 1; k' is the slow diffusion coefficient, min -1 , q is the oil content in the seeds during the extraction, g/100 g; q o is the oil Ultrasound-assisted extraction of sunflower oil from the cake after sunflower seed pressing 201 content initially present in the seeds, g/100 g, and τ is time in minutes.The b'and k' were determined experimentally from the slope and segment by plotting ln(q/q o ) against τ.The linear regression analysis, in this work, was done using Microsoft Office Excel 2007.
The model is based on the two-stage extraction mechanism and two parameters included.The first parameter (b') characterizes the washing stage (the so-called washing coefficient) and the second parameter (k') characterizes the slow extraction (the so-called slow extraction coefficient).A graphic presentation of a change of the ln(1-q/q o ) during the time is shown in Figure 4. Significant deviations from the model equation could only be observed in the initial period.The values of both kinetic parameters are presented in Table 3.
On the basis of linear forms of the kinetic curves (shown in Figure 4) and the linear correlation coefficient, it can be concluded that the kinetic models applied fitted well the experimental data.The coefficients of linear correlation were higher than 0.995 based on values in Table 3.In the second stage of extraction, a high level of coefficients of linear correlation (R) illustrates a very good agreement between the theory model and the experiment.
Based on the values of the rapid (b') and slow (k') parameters of the extractions shown in Table 3, the positive effect of ultrasound is seen.The coefficient of slow extraction without ultrasonic application was k' = 0.026 min -1 , whereas with the use of ultrasound k' = 0.039 min -1 .The coefficient of slow extraction was found to be 50% higher with the use of ultrasound.The effect of ultrasound is probably due to cavitation.The cavitation phenomenon is likely to partially destroy the structure of the solid phase (Toma et al., 2001).In addition, the cavitation phenomenon is likely to partially increase the temperature of oil in the solid phase.This affects the viscosity of the oil as well.In this way, the diffusion of the oil through the solid phase is accelerated.
The coefficient of washing without the application of ultrasound was b' = 0.589, whereas with the use of ultrasound b' = 0.713.The use of ultrasound increased the rate of extraction in the initial period (washing) by 21%.Obviously, a greater contribution of the ultrasound to the extraction of the oil was recorded in a slower phase, in the diffusion, through the solid material.

Conclusion
The obtained results show that the ultrasound had a positive effect on extraction in the Soxlet apparatus.For the same number of cycles in the Soxlet apparatus, the yield with ultrasound was about 1.5% higher.In relation to n-hexane and petroleum, the extraction with petrol showed the best results (yield) in extraction, both without and with ultrasound.
A two-parameter mathematic model was used for the extraction kinetics modeling; a model based on unsteady diffusion through solid material.In the second stage, the applied mathematical model best followed experimental results.The linear correlation coefficient was R = 0.995 without ultrasound and R = 0.997 by using of ultrasound.A positive effect of ultrasound, both on the extraction rate and on the oil yield, was confirmed by the applied model (b', k').As is known, in addition to ultrasound, many other parameters (grain size, temperature, hydrodynamic conditions, etc.) also affect the extraction efficiency.Therefore, the obtained results provide a good base for further research.

Figure 2 .
Figure 2. A graphic presentation of ultrasound application in sunflower oil extraction of the mechanically pressed sunflower seeds in various solvents.

Table 1 .
The moisture content in the samples, % m/m.

Table 2 .
The sunflower oil residue in samples after pressing in operating conditions, determined without and with ultrasonically assisted extraction, % m/m.

Table 3 .
Parameters of the unsteady diffusion model without and with ultrasuond.

Table 4 .
Some mathematical models describing solid-liquid extraction.