Eur Food Res Technol (2003) 218:26–31 DOI 10.1007/s00217-003-0794-0

ORIGINAL PAPER

Medeni Maskan · Halil I˙. Bag˘cı

The recovery of used sunflower seed oil utilized in repeated deep-fat frying process Received: 6 May 2003 / Revised: 11 August 2003 / Published online: 1 October 2003
Springer-Verlag 2003

Abstract In this study, 50 consecutive deep-fat fryings were done by frying potato samples, each weighing 100 g, in sunflower seed oil at 170 C. Significant chemical and physical changes in sunflower seed oil were observed during frying. A number of official methods were used to evaluate its adsorption abilities including free fatty acids (FFA), peroxide value (PV), conjugated dienes at 232 nm, secondary oxidation products at 270 nm and specific heat value determination. These parameters were determined in oil samples taken after each of the ten fryings before and after adsorbent treatment. A mixture of 2% pekmez earth, 3% bentonite, and 3% magnesium silicate was used as the adsorbent mixture. The FFA content of oil increased from 0.17 to 0.29% during frying. The use of adsorbents reduced FFA content of the used oil to 0.13%, i.e., a value below the FFA content of fresh oil (0.17%). Peroxide values decreased during frying because of decomposition of peroxides at high temperatures. A significant reduction was obtained in peroxide and conjugated diene values (K232 value) due to the adsorbent treatment. However, the treatment increased the amount of secondary oxidation products (K270 value). The specific heats of untreated used oil were higher than specific heats of adsorbent treated used oil over the entire frying process. Keywords Frying oil · Adsorbents · FFA · PV · Specific heat

Introduction Deep-fat/oil frying is extensively used in food processing both industrially and at home, and fried potato products M. Maskan ()) · H. I˙. Bag˘cı Food Engineering Department, University of Gaziantep, 27310 Gaziantep, Turkey e-mail: [email protected] Tel.: +90-342-3601200/2309 Fax: +90-342-3601105

are one of its largest applications. It is important for the food industry and food regulatory agencies to have simple and objective methods for quality evaluation of used frying oil. In the case of the food industry, a significant economic advantage can be gained from the ability to determine the appropriate point at which a frying oil is no longer suitable for use [1]. In addition, such methods would provide information regarding the safety and nutritional quality of their final products [2]. The nature of the deep-fat frying process is such that a progressive degradation of certain qualities of the cooking oil occurs during continuing use [3, 4]. Which test is the best indicator of oil quality is still unclear. The snack food industry tends to use free fatty acid (FFA) content as a chemical marker for predicting oil stability of products as they pass through distribution to the consumer. However, the characteristics which are of most interest to the potato chip and similar industries are FFA content, viscosity and color change of the vegetable cooking oil as well as formation and decomposition of hydroperoxides and polymerization via complex free radical processes at elevated temperatures above 160 C [5]. The latter compounds are harmful to the human body and also lead to the rejection of used frying oil [2, 6]. A large proportion of fats and oils is used in frying processes. Used frying oils are generally discarded because oxidized lipid degrades the quality of fried foods, however, discarded oil still has a large portion of triglycerides. Economic considerations and the need to produce fried foods of desirable quality have stimulated an interest in the purification of used frying oil. Purification could be achieved by removing the undesirable oxidized low molecular weight material and polymers [7]. Current interest in the study of evaluation of used frying oil is apparent from the relatively large number of papers appearing in the last few years. Some researchers have concentrated on the utilization of used oil as a fuel oil alternative [8]: biodiesel production [9, 10. 11, 12] as alternative fuels for diesel engines. On the other hand, Jaswir et al. [13] studied the effect of the addition of

27

natural and synthetic antioxidants to frying oil in order to extend its frying life. Recently, a few researchers have focused their attention on the purification of used frying oil in order to make the oil edible by removing degradation products. Several different methods have been used for this purpose. The membrane processing technique used by Subramaninan et al. [7] to enhance the shelf-life of used frying oils by removing oil-soluble impurities significantly decreased the quantity of the impurities. Yoon et al. [14] investigated the separation of triglycerides from used frying oil by supercritical carbon dioxide extraction. However, these two systems have still not gained application industrially because of high investment and operating costs. The most popular method is the recovery of used frying oil by adsorbent treatments. It is simple and efficient. Most of the investigations have been done by Lin et al. [15, 16, 17], but in their studies there are no detailed studies on the chemical and thermal properties of used frying oil, during frying and after adsorbent treatment. Therefore, the aims of this study were to determine 1) FFA, PV, specific heat and specific extinctions, (E1% 1 cm) patterns of sunflower seed oil during frying and 2) the patterns of the same parameters of the used frying oil purified by an adsorbent mixture, which had been optimized previously [18]. It is believed that used frying oil can be upgraded to the level existent in freshly refined oils, thus, providing considerable savings to food processors.

Materials and methods Materials Fresh potatoes and sunflower seed oil were purchased from a local market in Gaziantep,Turkey. Chemicals (NaCl, NaOH, glacial acetic acid, chloroform, KI, Na2S2O3, n-hexane and ethanol) for analyses were purchased from Merck Company (Germany) and phenolphthalein and starch were from Riedel de Haen (Germany). Potato preparation for frying The potatoes were peeled and washed under tap water. The washed potatoes were then cut into slices (6.0€1.237.8€1.25 49.5€ 4.09 mm3) using a vegetable slicer. Uniform size is essential for uniform heat transfer between the potato slices and the frying oil. The sliced potatoes were weighed into 100-g portions and soaked in a 2.5% NaCl solution for 5 min. This reduces the oil absorption capacity and prevents surface darkening of the potato slices due to oxidation. It also positively affects the surface properties such as improving the rigidity of potato slices by complexing pectin. The complexes formed decrease the solubility of pectin and prevent the potato from disintegration during frying [13]. Following draining, the potato slices were blotted with a paper towel before frying. Frying and frying oil sampling An electrical deep-fat fryer (Mares de Luxe-2, France) with a frying basket and 3-L-oil capacity was used for frying the potato slices. The fryer was operated at a temperature of 170 C. Exactly 1.5 L of sunflower seed oil was heated to 170 C at the beginning and samples of potatoes weighing 100 g were fried for 6 min. The

frying period was taken from literature and preliminary studies. During the frying process, the lid of the fryer was closed and the basket was immersed into the hot oil. After 6 min, the fried slices were removed from the fryer. Then, the next sample was fried. The time between frying cycles was 1 min. The oil was subjected to 50 consecutive fryings. In order to determine the effect of frying on the oil properties, 140 mL of frying oil was withdrawn for the fresh oil and after each of the following frying cycles: 1st, 10th, 20th, 30th, 40th, 50th , cooled in a cool room at 13 C and filtered using a coarse cellulosic filter paper to eliminate any suspended matter. The volume of oil was not replenished to the original volume with fresh oil after any of the fryings. The oil samples were centrifuged, filtered and stored in a refrigerator at 4 C in glass-stoppered flasks until used. Adsorbents In order to remove fat soluble degradation products and insoluble particles, a mixture of 2% pekmez earth (CaCO3 containing special natural white soil), 3% bentonite, 3% magnesium silicate, and 92%, by weight, used sunflower seed oil was prepared for purification of the used oil . This mixture was found to be the optimum mixture as previously investigated [18]. Chemical analysis Analyses were carried out on frying oil samples before and during frying as well as after adsorbent treatment. FFA were assayed according to the methods of AOCS Ca 5a-40 [19] and the common iodometric method was used for PV determination. In order to eliminate the problems associated with this method [5], the corrected iodometric titration method in BS 684, Section 2.14 is recommended [20]. Oxidation products were evaluated by UV absorbance at 232 nm (initial products of oxidation, i.e., conjugated hydroperoxides) and at 270 nm (final oxidation products, i.e.,FFA, aldehydes, and ketones) by using a Beckman Model 24 spectrophotometer (Marca Reg, USA). They were measured as specific extinctions (K232 and K270), E1%1 cm, i.e., extinction of 1% solution (g/100 ml) of the oil in n-hexane, in a thickness of 1 cm at the specified wavelength [19]. Differential scanning calorimeter (DSC) analysis DSC measures the amount of energy (heat) absorbed or released by a sample as it is heated, cooled, or held at constant temperature. The heat flow of the adsorbent mixture in treated and untreated oil samples was determined using DSC (Perkin Elmer Instrument Pyris 6 DSC model, Norwalk, CT, USA) with a sensitivity of 20 mW at a temperature range from 0 to 80 C. It was equipped with a Haake thermostat DC50 and a cooler K20 (Karlsruhe, Germany) unit. The scan rate was 12 C/min and about 17 to 24.5 mg of oil sample was placed in the hermetically sealed aluminium pans. An empty aluminium pan and its lid was used as an inert reference. Nitrogen (99.999% purity) was used as the purge gas at a flow rate of approximately 20 mL/min. Three scans were run for each frying cycle. The manufacturer’s software (Pyris 6 Kinetics Software) program was used to analyse and plot the thermal data, and calculate specific heat values of the used oil. Statistical analysis Statistical analysis was conducted using the SigmaPlot (Scientific Graphing Software, version 4.00). The paired t-test was applied to the properties of treated and untreated used frying oils. Trends were considered significant when means of compared pairs differed at the P<0.05 level.

28

Fig. 1 Effect of frying and adsorbent treatment on free fatty acids (FFA) of used sunflower seed oil samples

Results and discussion Effect of adsorbent mixture on properties of used sunflower seed oil Effect on FFA The FFA content, resulting from the hydrolysis of triacylglycerol as well as further decomposition of hydroperoxides, is one of the most important indicators of frying oil deterioration. The released fatty acids are more susceptible to thermal oxidation under frying temperatures. The oxidized products of fatty acids give the off-flavors and odors (hydrolytic rancidity) to the frying medium and fried foods [15, 21]. Therefore, controlling the level of FFA within a reasonable range would prevent the breakdown of fats. Determination of FFA appears to be the method favored by many operations for quality control evaluation of used frying oil [22]. The FFA content of oil increased from 0.17 to 0.29% (i.e., about 65% increase) during repeated frying in this study. The adsorbent mixture showed a remarkable ability to remove FFA to below the initial FFA value of 0.17% as shown in Fig. 1. This excellent ability is probably owing to the magnesium silicate content of the mixture, which has the most active sites, largest surface area [16], and gives the mixture basic properties, enabling it to attract acids and other polar color compounds. The difference among adsorbent treated and untreated used frying oil was significant (P<0.05). Other researchers investigated mixtures of various compositions with the intention to improve the recovery ability of adsorbents. Lin et al. [16] reported about 74% reduction in FFA content by a treatment of oil with an adsorbent mixture (wt%) of 4.5% clay, 0.5% charcoal, 2.5% MgO, and 2.5% celite (the rest is oil). McNeill et al. [23] prepared mixtures of activated carbon and silica adsorbents with different compositions for FFA reduction purposes in

Fig. 2 Effect of frying and adsorbent treatment on peroxide value (PV) of used sunflower seed oil samples

used oil. They observed 28–59% reduction. In other investigations, Subramanian et al. [7], and Zhang and Addis [24] found no significant change in FFA content during membrane processing of used oil and paper filter, plus diatomaceous earth treatments of heated oils, respectively. These results clearly show that the prepared adsorbent mixture used in our study gives better results than mixtures reported in other studies. Effect on peroxides Peroxide formation is a major concern from the point of view of rancidity and toxicology.. Food lipid oxidation products such as peroxides, the free radicals involved in their formation and propagation, malonaldehyde, and several cholesterol oxidation products are reported to promote atherosclerosis and coronary heart disease [7].The change in PV (meq of peroxide per kg of sample) of the oil is shown in Fig. 2. The PV of the used oil was significantly higher (P<0.05) than the treated oil except in the 50th frying cycle. Peroxides declined from 12.7 to about 4 during continued use. It is the inherent thermal instability of the peroxy-bond at severe frying temperatures which results in the formation of aldehydes and ketones (rancid off-flavor compounds) followed by a significant increase in absorptivity at 270 nm (Fig. 3). This would be in agreement with the results of Thais et al. [25]. A decrease in PV during frying has been observed by several researchers [6, 24, 25]. On the other hand, others [13, 26] have reported an increase in PV of oil during heating and frying. The decline in PV during frying may be explained by the fact that the moisture content of the food being fried has a positive effect in that it creates as a steam-blanket effect over the surface of the oil, thereby reducing contact with air as well as helping to volatilize and remove peroxides, flavors, and odors which would otherwise

29

Fig. 3 Effect of frying and adsorbent treatment on extinction coefficient (K270) of used sunflower seed oil samples at 270 nm

Fig. 4 Effect of frying and adsorbent treatment on extinction coefficient (K232) of used sunflower seed oil samples at 232 nm

accumulate in the frying oil [22]. Considering all these points, food scientists [6, 25] have concluded that the PV is not a good index for the measure of oxidation because hydroperoxides are unstable under frying conditions. Moreover, the PV of oil samples at frying temperatures of 160 to 180 C are zero [27]. It is clear that (Fig. 2) most of the peroxides present in the used frying oil were eliminated significantly (P<0.05) during adsorbent treatment process, as indicated by the lower PV (about 4).

temperature. Therefore, absorptivity at 232 nm cannot be used as a substitute for the PV under these frying conditions. The used untreated frying oil showed a slight increase in K270 demonstrating formation of secondary oxidation products. The increase in K232 and K270 values are in agreement with the results of others [25, 29]. However, the K232 and K270 values estimated in this study are higher than the values published in some literature [25, 28, 30]. This maybe owing to the mild processing conditions applied in other studies and also due to the stability of the oils used. It was observed from the specific extinction values at 232 nm that some conjugated dienes and trienes were actually rejected to a small extent by application of adsorbents mixture to the used frying oil samples, i.e., the decomposition products produced during frying were reduced by the application of the adsorbents mixture. The reduction was significant (P<0.05) for primary oxidation products. However, the used treated frying oil showed significant (P<0.05) increase in K270 compared to untreated used oil, demonstrating formation of secondary oxidation products during sample cooling and adsorbent treatment. This problem may be the result of an inability of the adsorbents used to remove secondary oxidation products. The results show that the determination of these two parameters may be useful to track oxidative deterioration of oil during the frying process.

Effect on specific absorption Specific absorption determines the oxidation levels of the oil. The primary products of lipid oxidation are hydroperoxides known as peroxides. Similar to PV, absorbance at 232 nm also measures the degree of primary oxidation. Therefore, the amount of peroxide is directly correlated with the specific absorption at 232 nm as described by Jaswir et al. [13]. If oxidation continues secondary oxidation products are produced. These products are aldehydes, ketones, and alcohols and are monitored by specific absorption at 270 nm [25, 28]. Determination of absorptivity (extinction coefficient) in the UV spectrum is simple, does not depend on a chemical reaction or color development, and requires a small sample size. The absorptivity values, K270 and K232, against the number of fryings are shown in Figs. 3 and 4, respectively. The K232 values increased gradually with the increase in frying number. The variation of K232 (Fig. 4) presented a pattern different from that of PV as indicated in Fig. 2. It means that conjugated dienes formed during frying increased absorbance at 232 nm in addition to PV formed even at low concentrations. The conjugated dienes are formed during oxidation of polyunsaturated fatty acid resulting in a shift in one of the double bonds. In this study, the absorptivity values obtained were not related to the PV due to decomposition of peroxides at the frying

Effect on specific heat values Reliable estimates of thermal-physical properties of foods are essential in order to understand and describe various thermal processes (e.g., the rate of heat transfer during frying) as well as to optimize the design and operation of heating and cooling systems, and modeling of frying systems. Specific heat, thermal conductivity, and density

30

fresh (unused) oil. As the oil degrades, the heat capacity decreases and the flywheel “slows down” to a lower value.

Conclusions and recommendations The study of the refinement of used sunflower seed oil, utilized for frying, by adsorbent treatments revealed the following conclusions:

Fig. 5 Effect of frying and adsorbent treatment on specific heat values of used sunflower seed oil samples

are the heat transfer properties of interest [31, 32]. The specific heat measurements were based on the comparison of the heat flow rates into the sample and reference pan subjected to identical temperature programs. The following equation was used to calculate specific heats from heat flow versus temperature curves, obtained from DSC [31]: Cpapp ¼ H2
H1 ðT2
T1 Þ  m

ð1Þ

where, Cpapp is the apparent specific heat, (H2-H1) is the change in enthalpy over the temperature range (T2-T1) studied and m is the mass of oil sample. The apparent specific heat values of used oil before and after adsorbent treatment are presented in Fig. 5. Generally, the specific heats of untreated used oil were higher than the specific heats of treated used oil over the entire frying series except for the first frying (number 1). The average specific heat values were 4.11€0.25 and 3.86€0.24 j/g C for untreated and treated used frying oils, respectively. This difference was not significant statistically (P>0.05). Since specific heat strongly depends on composition [32, 33], the higher specific heat for untreated used frying oil may be associated with changes in molecular structure of the oil occurring during frying. The specific heats of compounds newly formed during frying may cause an apparent increase in the specific heat of used oil. Most of these compounds were eliminated by adsorbent treatment, hence, the composition of the recovered oil changed, resulting in lower specific heat values. Overall, specific heat decreased as the frying number increased. Similar results were obtained by Gloria and Aguilera [34] and, Tan and Man [26] for oils heated for 10 and 12 h, respectively. Also, Paul and Mittal [3} observed a significant decrease (14%) in specific heat of canola oil at higher levels of fat degradation (7 days of frying). This phenomena has been explained by Blumenthal [5]. He states that initially the oil has a high heat capacity which diminishes with use. This means that there is a “thermal flywheel”, or energy sink at a high level in a

The FFA content, K232 and K270 values increased during frying. However, the PV and specific heat values of frying oil decreased as frying number increased. The adsorbent mixture prepared improved all the parameters studied except secondary oxidation products measured at 270 nm. The increase in these products was attributed to the formation of aldehydes during the adsorbent treatment process. The adsorbent mixture may be used for refinement of used frying oils industrially. For future work, it would be complementary to the present study to investigate the effect of adsorbent treatments on the quantity of tocopherols, total polar components, polymers, and health implications of the refined used oil. Especially, the formation of polycyclic aromatic hydrocarbons must be investigated during frying as well as their elimination, since they may cause cancer [35]. Also, the flavor and oxidative stability of treated used oil should be studied. Acknowledgements The authors thank the University of Gaziantep, Turkey, for research support.

References 1. Aguilera JM, Gloria-Hernandez H (2000) J Food Sci 65:476– 479 2. Clark WL, Serbia GW (1991) Food Technol 45:84–89 3. Paul S, Mittal GS (1996) J Food Proc Eng 19:201–221 4. Paul S, Mittal GS (1996) J Food Eng 30:389–403 5. Blumenthal MM (1991) Food Technol 45:68–71, 94 6. White PJ (1991) Food Technol 45:75–80 7. Subramanian R, Nandini KE, Shella PM, Gopalakrihna AG, Raghavorao KSMS, Nakajima M, Kimura T, Maekawa T (2000) JAOCS 77:323–328 8. Karaosmanog˘lu F, Beker UG (1996) Energ Sour 18:637–644 9. Alcantara R, Amores J, Canoira L, Fidalgo E, Franco MJ, Navarro A (2000) Biomass Bioenerg 18:515–527 10. Bak YC, Choi JH, Kim SB, Kang DW (1996) Korean J Chem Eng 13:242–245 11. Cvengros J, Povazanec F (1996) Biosource Technol 55:145– 150 12. Neto P RC, Rossi LFS, Zagonel GF, Ramos LP (2000) Quim Nova 23:531–537 13. Jaswir I, Che Man YB, Kitts DD (2000) Food Res Int 33:501– 508 14. Yoon J, Han BS, Kang YC, Kim KH, Jung MY, Kwon YA (2000) Food Chem 71:2785–2790 15. Lin S, Akoh CC, Reynolds AE (1998) J Food Lipid 5:1–16 16. Lin S, Akoh CC, Reynolds AE (1999) JAOCS 76:739–744 17. Lin S, Akoh CC, Reynolds AE (2001) Food Res Int 34:159–166

31 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

Maskan M, Bag˘cı HI˙ (2003) Eur Food Res Technol 217:215– 218 Firestone D (1990) Official and tentative methods of the American Oil Chemists’ Society, 4th edn. American Oil Chemists’ Society, Champaign, Robards K, Kerr AF, Patsadiles E (1988) Analyst 113:213–224 Nawar WW (1996) In: Fennema OR (ed) Food Chemistry, 3rd edn. Marcel Dekker, New York, pp 225–319 Stevenson SG, Vaisey-Genser M, Eskin NAM (1984) JAOCS 61:1102–1108 McNeill J, Kakuda Y, Kamel B (1986) JAOCS 62:1564–1567 Zhang WB, Addis PB (1992) J Food Sci 57:651–654 Vieire TMFS, Regitano-d’Arce MAB (1999) J Agric Food Chem 47:2203–2206 Tan CP, Che Man YB (1999) Food Chem 67:177–184 Karsten N (1996)http://www.aarhus.com/aarhusolie/aarhusolie.nsf/links

28. Albi T, Guinda LA, Perez-Camino MC, Leon M (1997) J Agric Food Chem 45:3000–3003 29. Petukov I, Malcomson LJ, Przybylski R, Armstrong L (1999) JAOCS 76:627–632 30. Gomes T, Caponio F (1997) J Agric Food Chem 45:1381–1384 31. Karunabar B, Mishra SK, Bandyopadhyay S (1998) J Food Eng 37:345–351 32. Ngadi MO, Mallikarjunan P, Chinnan MS, Radhakrisnan S, Hung YC (2000) J Food Proc Eng 23:73–87 33. Hwang CY, Heldman DR, Chao RR, Taylor TA (1999) J Food Sci 64:141–144 34. Glooria H, Aguilera JM (1998) J Agric Food Chem 46:1363– 1368 35. Leon-Camacho M, Viera-Alcaide I, Ruiz-Mendez MV (2003) Eur J Lipid Sci Technol 105:9–16