The effect of different filter aid materials and winterization periods on the oxidative stability of sunflower and corn oils

Abstract

The effects of different filter aid materials at different ratios (0.3% or 0.6%, w/w) and three different (6 h, 12 h, 24 h) winterization times, on free fatty acid (FFA), peroxide, iodine, p-anisidine and conjugated diene values and copper (Cu), iron (Fe), nickel (Ni) levels of sunflower and corn oils were studied. FFA amounts of both sunflower and corn oils showed no significant changes during winterization process. However, peroxide, p-anisidine and conjugated diene values increased during winterization due to the holding time in winterized tank and atmospheric conditions. Iodine values of oils increased as unsaturation rises. In respect of trace metals, the increase in copper and iron levels was found higher than the increase of nickel level.

Para esta investigación se analizaron los efectos que diferentes materiales filtrantes auxiliares en distintos ratios (0,3% y 0,6% w/w) y tres distintos periodos de invernización (6 h, 12 h, 24 h), tienen sobre el ácido graso libre (AGL), los valores de peróxido, de yodo, de p-anisidina y de dieno conjugado, así como en los niveles de cobre (Cu), hierro (Fe) y níquel (Ni), en aceites de girasol y de maíz. Se constató que, durante el proceso de invernización, las cantidades de ácido graso libre no mostraron cambios significativos ni en el aceite de girasol ni en el aceite de maíz. Sin embargo, los valores de peróxido, de p-anisidina y de dieno conjugado, aumentaron a lo largo de este proceso debido al tiempo de retención en el tanque previsto para realizarlo y a las condiciones atmosféricas. Asimismo, a medida que se elevó la insaturación se registraron aumentos en los índices de yodo de los aceites. En términos de los metales traza, el incremento verificado en los niveles de cobre y de hierro fue superior al observado en los niveles de níquel.

Keywords:

• free fatty acid
• filter aid materials
• oxidative stability
• trace metals
• vegetable oils
• winterization

Palabra claves:

• ácido graso libre
• materiales filtrantes auxiliares
• estabilidad oxidativa
• metales traza
• aceites vegetales
• invernización

Introduction

It is impossible to consume crude vegetable oils. To bring in edible oil properties a series of refining process is applied to oils. Crude vegetable oils contain some undesirable compounds such as free fatty acids (FFAs), phosphatides, peroxides, aldehydes, pigments, trace metals, polymers, waxes, mono and diglycerides. The amount and nature of these compounds are related to growing conditions of seeds, storage conditions and processing technology (Kayahan, 2003). Those compounds must be removed from the oil without destroying triglyceride structure by refining. Winterization is a step in refining vegetable oils to remove compounds which crystallize at low temperatures and cause turbidity in oils (Coutinho et al., 2009; Turkulov, Dimić, Karlović, & Vukša, 1986). Trace metals present in oils and fats may be of natural origin or present due to processing procedures such as bleaching (Fe), hardening (Ni, Cu) as well as corrosion of the processing equipment (Fe, Ni). They have catalytic effect on the mechanism of autoxidation (Reyes & Campos, 2006). Normally, trace metals can be removed from the oils during degumming step applied before winterization (Wesdorp, 1996).

During winterization step, some quality characteristics (acidity, peroxide values, spectrophotometric constants, etc.) of oils which are the accelerating effect of oxidation can be increased, and in the deodorization step applied after winterization, it is difficult to lower them. During deodorization, the FFAs, aldehydes, ketones and other short chain compounds resulting from oxidation are removed (Hamm & Hamilton, 2000).

The final product should have acceptable flavour, odour, colour and oxidative stability. One of the key quality points during processing is avoidance of exposure of the oil to air. This becomes particularly important in the winterization step (Hamm & Hamilton, 2000).

Since it is cheaper than the others, perlite is widely used as filter aid material for the winterization process of vegetable oils in Turkey, whereas in European countries and United States, kieselguhr, which has a porous structure, is generally used for this purpose. It is thought that the present study is important for the vegetable oil industry because up to our knowledge the present study is the first research planned to determine the effects of using these two filter aid materials at different concentrations on the oxidative stabilities of both sunflower and corn oils during winterization. For this purpose, some quality characteristics were determined such as FFA, peroxide value, iodine value, p-anisidine and conjugated diene values. Additionally, trace metal (Fe, Cu and Ni) levels were also detected. At the same time, the effects of these filter aid materials on the oxidative stability of the oil samples were also compared with each other. The objective of this study was to evaluate the oxidative stability of vegetable oils during winterization processes by using different filter aid materials in different ratios.

Materials and methods

Materials

Degummed, neutralized and bleached sunflower and corn oils were obtained from commercial vegetable oil refineries in high-density 30-L polyamide tanks. All chemicals and reagents used were analytical grade of E. Merck, Germany.

Filter aid materials were obtained from Altınyağ Extractors Inc in İzmir. Water contents of filter aid materials were 0.5% and 0.06% for kieselguhr and perlite, respectively.

Methods

Winterization process

Perlite and kieselguhr were used as filter aid materials in concentrations of 0.3 and 0.6%, respectively. The winterization process was carried out in three different periods of time (6 h, 12 h and 24 h) and in two replications. Also, a control winterization was performed without adding any filter aid materials. The pilot winterization unit having a capacity of 50 L was used in the research. The unit was composed of refrigeration and crystallization tanks, a digital temperature-control panel, a compressor with Freon12 refrigerant, holding and balance tanks, a plate filter, a feed tank, a mixer, a centrifugal pump and junction pipes. The winterization process includes three-step cooling, crystallization and filtration. In pre-cooling, the temperature was reduced from ambient temperature to 10°C by mixing for 2 h with a plate-type heat exchanger. For better crystallization and maturation, the mixture was cooled to 4 ± 1°C. Then, the filter aid material was added by weighing. At this temperature, the mixer speed was set at 30 rpm/min to obtain slow mixing for 6 h, 12 h and 24 h. Thus, wax crystals formed during this period gets matured. After maturation, the mixture was heated to 8–10°C to reduce the viscosity and facilitate filtration. The crystalline phase was removed from the mixture by using filter plates under about 1 bar pressure.

Sampling

After filtering the winterized oil at 6 h, 12 h and 24 h, 500 mL of sample from each batch was taken for analysis. The sampling design is given in Figure 1. All analyses were repeated two times.

Figure 1. Factorial experiment plan for winterization sampling process.

Figura 1. Plan de experimentación factorial para el proceso de muestreo de invernización.

Chemical analysis

FFA contents (as percentage oleic acid), average iodine values and peroxide values of the oil samples were determined as defined in AOCS (1993), TS 4961 ISO 3961 (1997) and TS 4964 ISO 3960 (2004). Conjugated diene values were determined according to the American Oil Chemist’s Society Method Ti 1a-64 (AOCS, 1997a). p-Anisidine value analysis was done according to the procedure described by the American Oil Chemist’s Society Method Cd 18–90 (AOCS, 1997b). The absorbance of solutions for conjugated diene and p-anisidine values was measured in a UV–VIS spectrophotometer (Shimadzu UV-1601, Kyoto, Japan). All analyses were also performed for the sample CS (control sample) which was not winterized.

Determination of copper, nickel and iron levels

Standard solutions

The working standard solutions were prepared by diluting the stock solutions of 1000 mg/L of copper, nickel and iron. The ratio of standard solutions prepared for each element was different.

Sample preparation

Microwave acid digestion method was used for sample preparation (Sarojam, 2009). An amount of 0.2 g of samples was taken into digestion tubes and 5 mL of HNO₃ (65%), 1 mL of HCl and 3 ml of H₂O₂ (30%) were added. The samples were digested in a Mars5 microwave digestion unit (CEM Corporation, Matthews, USA) according to the programme given in Table 1. After digestion, the samples were diluted up to 25 mL with 2 mL/L HNO₃. Duplicated analysis was performed on the samples. Blank digestion was also carried out in the same way.

Table 1. Microwave acid digestion programme (Sarojam, 2009).

Tabla 1. Programa de digestión de ácido de microondas (Sarojam, 2009).

Step	Ramp time, min	Hold time, min	Cooling period, min	Pressure (MPa)	Temperature (°C)
1	10	10	5	0.758	110
2	10	10	5	1.034	150
3	20	10	5	1.310	190

Atomic absorption spectrophotometer apparatus and analysis conditions

The trace metal levels were determined by PerkinElmer AAnalyst 800 model atomic absorption spectrometer (PerkinElmer Instruments, Shelton, CT, USA) equipped with transversely heated graphite atomization (THGA), and a longitudinal Zeeman effect background corrector was also used. Lumina lamp, electrodeless discharge lamp (EDL) and hollow cathode lamps were employed as radiation sources. The operating parameters for working elements by graphite furnace were set as recommended by the application note of Sarojam (2009) for edible oils. Atomic absorption spectrophotometer (AAS) graphite furnace operating parameters and temperature programme are given in Tables 2 and 3, respectively. Samples were injected into the graphite tubes using a PerkinElmer AS-800 Autosampler. The atomic absorption signal was determined in peak height mode against a calibration curve.

Table 2. AAS graphite furnace operating parameters (Sarojam, 2009).

Tabla 2. Parámetros de operación de horno de grafito AAS (Sarojam, 2009).

Parameters	Fe	Cu	Ni
Current (mA)	30	15	25
Wavelength (nm)	248.3	324.8	232
Slit (nm)	0.2	0.7	0.2
Mode	AA-BG	AA-BG	AA-BG
Sample volume (µL)	20	20	20
Matrix modifier	0.015 mg MgNO3	0.005 mg Pd and 0.003 mg MgNO3	–
Modifier volume (µL)	5	5	0

Table 3. AAS graphite furnace temperature programme (Sarojam, 2009).

Tabla 3. Programa de temperatura de horno de grafito AAS (Sarojam, 2009).

Steps	Operation	Temperature, °C	Ramp time, s	Hold time, s	Internal flow, mL/min
1	First drying	110	1	30	250
2	Second drying	130	15	30	250
3	Pyrolysis	1400 (Fe), 1200 (Cu), 1100 (Ni)	10	20	250
4	Atomization	2100 (Fe), 2000 (Cu), 2300 (Ni)	0	5	0
5	Clean-out	2450	1	3	250

Statistical analysis

Data were statistically analysed by analysis of variance (ANOVA) using the SAS software package. The PROC MIXED procedure was applied followed by Fisher’s LSD when the differences between applications were significant (p < 0.05). Factorial design was used as experimental design. Winterization applications and time were two factors in the experimental design. Filter aid materials are discussed together with the amounts used and evaluated as a single factor (SAS, 2001).

Results and discussion

Changes in FFA levels of the oil samples

The effects of using different filter aid materials and different winterization treatments on FFA contents of sunflower and corn oils are given in Figure 2. It can be seen that the FFA levels did not change significantly during winterization in both oils (p > 0.05). But the initial average FFA level (0.10%) of sunflower oil was increased to 0.14% by using kieselguhr as filter aid material at a concentration of 0.3%, whereas initial average FFA level (0.12%) of corn oil was increased to 0.16% by using kieselguhr as filter aid material at a concentration of 0.3% and 0.6%. Generally, slight increases were observed in winterization by using kieselguhr than in winterization by using perlite. When sunflower and corn oils were compared to each other, it was seen that an increase in free fatty acidity of corn oil was higher than an increase in that of the sunflower oil. It is thought that this is because of the higher linolenic acid content of corn oil when compared to sunflower oil. According to TS 888 (2003) and TS 886 (2005), the highest FFA content of refined sunflower and corn oils must be lower than 0.3%.

Figure 2. Changes in free fatty acid levels of sunflower and corn oils during different winterization processes.

Figura 2. Cambios en los niveles de ácido graso libre (AGL) de los aceites de girasol y de maíz durante distintos procesos de invernización.

Notes: KW: Winterization in which kieselguhr was used; PW: winterization in which perlite was used; CW: Control Winterization in which no filter aid material was used; CS: Control sample.

Nota: KW: Invernización en la que se usó kieselgur; PW: invernización en la que se usó perlita; CW: invernización de control en la que no se usó ningún material filtrante auxiliar; CS: muestra de control.

According to the results of the statistical analyses, it was determined that the interaction of winterization time and type of filter aid material did not affect the changes on the FFA level of the sunflower and corn oils (p > 0.05).

Changes in peroxide values of the oil samples

The peroxide value of sunflower oil increased from 3.3 meq/kg to 6.7 meq/kg during the winterization processes by using kieselguhr (0.6%) at the end of 12 h (Figure 3). This increase was also higher for the corn oil during winterization process in which kieselguhr was used as filter aid material. The peroxide value of control corn oil sample was 5 meq/kg in average. After winterization processes, this value was determined as 10.8 meq/kg (Figure 3). It is thought that this higher increase was because of the unsaturation degree of corn oil, temperature and presence of O₂. In the meantime, existence of trace metals shows pro-oxidant effect to the increasing peroxide value (Pehlivan, Arslan, Gode, Altun, & Özcan, 2008). Peroxide value is related to the formation of aldehydes which are the primary oxidation products so it must not exceed 10 meq/kg in refined vegetable oils (TS 886, 2005; TS 888, 2003). The increase in peroxide value of the samples during winterization in which perlite was used was found lower than in the peroxide value obtained for the samples during winterization in which kieselguhr was used. It is thought that this situation is related to the higher water content of kieselguhr when compared to perlite.

Figure 3. Changes in peroxide values of sunflower and corn oils during different winterization processes.

Figura 3. Cambios en los valores de peróxido en los aceites de girasol y de maíz durante distintos procesos de invernización.

Notes: KW: Winterization in which kieselguhr was used; PW: Winterization in which perlite was used; CW: Control winterization in which no filter aid material was used; CS: Control sample.

Nota: KW: Invernización en la que se usó kieselgur; PW: invernización en la que se usó perlita; CW: invernización de control en la que no se usó ningún material filtrante auxiliar; CS: muestra de control.

The effect of the concentration filter aid materials was found statistically important on the changes in peroxide value changes of sunflower and corn oils (p < 0.05), whereas winterization period showed no effect on the changes in peroxide values of the samples (p > 0.05)

Changes in p-anisidine values of the oil samples

p-Anisidine value is a criterion for the determination of the oxidation level. Average p-anisidine values of the samples of both corn and sunflower oils showed both increases and decreases during winterization process. As seen from Figure 4, average p-anisidine value of control sunflower oil was 1.55, whereas this value was 2.55 at the 24 h of winterization of sunflower oil for which 0.3% kieselguhr was used as filter aid material. Similar results were obtained for corn oil. At 24 h of the winterization of corn oil by using 0.6% kieselguhr, the average p-anisidine value became 2.50. The initial p-anisidine value of this sample was 1.44. It is seen that the effect of kieselguhr on the changes in p-anisidine values of the samples was more when compared to perlite. There is not a standard value for p-anisidine present. It is related to the formation of aldehydes which are the secondary oxidation products. p-Anisidine value analysis was carried out in order to control the oxidation and increase as the oxidation progress. The presence of the ambient light is one of the factors that increase the value of p-anisidine (Raza et al., 2009). Therefore, it is thought that p-anisidine value of the oil samples could be increased due to the exposure to light, moisture content of filter aid materials and holding period in the tank.

Figure 4. Changes in p-anisidine values of sunflower and corn oils during different winterization processes.

Figura 4. Cambios en los valores de p-anisidina en los aceites de girasol y de maíz durante distintos procesos de invernización.

Notes: KW: Winterization in which kieselguhr was used; PW: Winterization in which perlite was used; CW: Control winterization in which no filter aid material was used; CS: Control sample.

Nota: KW: Invernización en la que se usó kieselgur; PW: invernización en la que se usó perlita; CW: invernización de control en la que no se usó ningún material filtrante auxiliar; CS: muestra de control.

According to Ryan, Mestrallet, Nepote, Conci, and Grosso (2007), the p-anisidine value of refined sunflower oil was 1.88, whereas it was 3.14 in refined corn oil. These findings are similar to our results. Raza et al. (2009) investigated the changes in p-anisidine values of sunflower oil samples which were exposed to auto-oxidation and photo-oxidation. As they reported, average p-anisidine value of sunflower oil which was exposed to auto-oxidation increased to 10.03 from 1.51, whereas this value was 17.16 for the sample which was exposed to photo-oxidation.

According to the statistical analyses results, it was determined that the interaction of winterization period and the type of filter aid material showed no effect on the change in p-anisidine values of sunflower oil samples (p > 0.05). On the other hand, the interaction of winterization period and the type of filter aid material was found statistically important on the changes in average p-anisidine values of corn oil samples (p < 0.05).

Changes in conjugated diene values of the oil samples

Conjugated diene values were determined to control the oxidation degree of vegetable oils. They generally increase during different winterization treatments in both types of oils (Figure 5). Initial average conjugated diene value of sunflower oil was 0.11%, and the maximum increase was observed in the winterization for the sample for which kieselguhr was used as filter aid material at a concentration of 0.6%. At the end of 24 h, this value was 0.25%. On the other hand, at the beginning of the winterization, average conjugated diene value of corn oil was 0.13%. It was observed that this value became 0.22% at 24 h of the winterization process in which perlite was used at a concentration of 0.6%. Raza et al. (2009) have identified the conjugated diene value of sunflower oil as 0.09% before oxidation. Conjugated diene value of the samples starts to increase during the first stages of the oxidation and continues to increase during oxidation. There is no standard value present for the oils.

Figure 5. Changes in conjugated diene values of sunflower and corn oils during different winterization processes.

Figura 5. Cambios en los valores de dieno conjugado en los aceites de girasol y de maíz durante distintos procesos de invernización.

Notes: KW: Winterization in which kieselguhr was used; PW: Winterization in which perlite was used; CW: Control winterization in which no filter aid material was used; CS: Control sample.

Nota: KW: Invernización en la que se usó kieselgur; PW: invernización en la que se usó perlita; CW: invernización de control en la que no se usó ningún material filtrante auxiliar; CS: muestra de control.

According to the statistical analyses results, only the effect of using different filter aid materials was found important on the changes of conjugated diene values of sunflower oil samples (p < 0.05). And, it was found that winterization period had significant importance on the changes in conjugated diene values of corn oil samples (p < 0.05). It is thought that it was related to the higher linolenic acid content of corn oil than of sunflower oil.

Changes in iodine values of the oil samples

The effects of using different filter aid materials and different winterization treatments on FFA contents of sunflower and corn oils are given in Figure 6. Depending on the winterization time, the iodine values of sunflower and corn oils increased. The iodine value of sunflower oil reached to 132.5 g/100 g from 123.2 g/100 g in winterization by using kieselguhr at a concentration of 0.6% (24 h), whereas it reached to 146.3 g/100 g from 123.5 g/100 g in corn oil (24 h). This situation could be explained by removing the waxes and improving unsaturation (O'Brien, Farr, & Wan, 2000). In different studies, the average iodine value of sunflower oil was found to be 139 g/100 g (Anjum, Anwar, Jamil, & Iqbal, 2006; Raza et al., 2009). Naz, Sheikh, Siddiqi, and Sayeed (2004) determined in their study that the iodine value of corn oil was 125 g/100 g.

Figure 6. Changes in iodine values of sunflower and corn oils during different winterization processes.

Figura 6. Cambios en los valores de yodo en los aceites de girasol y de maíz durante distintos procesos de invernización.

Notes: KW: Winterization in which kieselguhr was used; PW: Winterization in which perlite was used; CW: Control winterization in which no filter aid material was used; CS: Control sample.

Nota: KW: Invernización en la que se usó kieselgur; PW: invernización en la que se usó perlita; CW: invernización de control en la que no se usó ningún material filtrante auxiliar; CS: muestra de control.

The effect of interaction of the type of filter aid materials and the time of winterization did not significantly affect the changes in iodine values of sunflower oil (p > 0.05). The effect of winterization time significantly affected the changes in iodine values of corn oil samples.

Changes in trace metal levels of the oil samples

Fe, Cu and Ni, which are the most important trace metals in oils, were investigated during different winterization treatments and the changes are given in Table 4. The Fe and Cu contents increased continuously in both oils due to the winterization time and filter aid materials. The increase of Ni content was not as high as in Fe and Cu. It is thought that the material of pilot winterization unit or filter aid materials used may help the increase in metals in samples. The Fe, Cu and Ni levels determined in perlite and kieselguhr are 6.73–7.11 mg/kg, 0.46–0.56 mg/kg and 0.55–0.62 mg/kg, respectively. Fe is the predominant metal among the filter aid materials. Together with this, the initial Fe and Cu contents of sunflower and corn oil samples were almost high, that is, 1.04–0.69 mg/kg and 0.12–0.12 mg/kg, respectively. This situation could be based on the climate and environmental conditions of the area where oil seeds are growing. At the end of the winterization process in which 0.3% perlite was used, average Fe content of sunflower oil was determined as 1.62 mg/kg, whereas this value was 1.54 mg/kg for the corn oil sample which was winterized by using 0.3% kieselguhr. On the other hand, Cu content of sunflower oil which was winterized by using 0.6% kieselguhr was 0.51 mg/kg, whereas it was found as 0.26 mg/kg for corn oil at the end of the winterization for which perlite was used (Table 4).

Table 4. Changes at trace metal levels of sunflower and corn oils during different winterization processes (mg/kg ± SD).

Tabla 4. Cambios en los niveles de metales traza en los aceites de girasol y de maíz durante distintos procesos de invernización (mg/kg ± DE).

			Fe		Cu		Ni
			Sunflower	Corn	Sunflower	Corn	Sunflower	Corn
Time (h)	Control sample		1.04 ± 0.06	0.69 ± 0.18	0.12 ± 0.06	0.12 ± 0.04	0.42 ± 0.01	0.18 ± 0.007
6	Control winterization		1.30 ± 0.00	0.71 ± 0.15	0.21 ± 0.02	0.16 ± 0.02	0.44 ± 0.02	0.20 ± 0.01
12			1.21 ± 0.28	0.72 ± 0.17	0.21 ± 0.08	0.20 ± 0.06	0.53 ± 0.11	0.17 ± 0.03
24			1.48 ± 0.34	0.85 ± 0.13	0.24 ± 0.03	0.23 ± 0.06	0.50 ± 0.05	0.18 ± 0.04
6	perlite	0.3%	1.37 ± 0.26	1.10 ± 0.08	0.25 ± 0.04	0.21 ± 0.05	0.49 ± 0.16	0.19 ± 0.01
12			1.24 ± 0.06	1.20 ± 0.08	0.33 ± 0.06	0.25 ± 0.04	0.51 ± 0.01	0.20 ± 0.00
24			1.62 ± 0.3	1.32 ± 0.06	0.33 ± 0.08	0.26 ± 0.04	0.48 ± 0.00	0.22 ± 0.00
6		0.6%	1.15 ± 0.03	1.42 ± 0.08	0.37 ± 0.15	0.20 ± 0.01	0.48 ± 0.08	0.18 ± 0.02
12			1.38 ± 0.24	1.46 ± 0.03	0.29 ± 0.05	0.24 ± 0.01	0.52 ± 0.01	0.18 ± 0.01
24			1.55 ± 0.33	1.51 ± 0.03	0.27 ± 0.01	0.26 ± 0.01	0.53 ± 0.03	0.23 ± 0.007
6	kieselguhr	0.3%	1.12 ± 0.13	1.43 ± 0.06	0.19 ± 0.05	0.16 ± 0.01	0.54 ± 0.05	0.22 ± 0.007
12			1.34 ± 0.05	1.43 ± 0.04	0.23 ± 0.06	0.19 ± 0.04	0.53 ± 0.04	0.20 ± 0.00
24			1.57 ± 0.08	1.54 ± 0.30	0.34 ± 0.11	0.23 ± 0.00	0.50 ± 0.06	0.21 ± 0.03
6		0.6%	1.19 ± 0.007	1.25 ± 0.01	0.51 ± 0.00	0.20 ± 0.04	0.48 ± 0.00	0.20 ± 0.007
12			1.57 ± 0.08	1.25 ± 0.03	0.32 ± 0.09	0.22 ± 0.03	0.43 ± 0.01	0.18 ± 0.00
24			1.59 ± 0.09	1.23 ± 0.03	0.34 ± 0.00	0.26 ± 0.01	0.45 ± 0.02	0.21 ± 0.007

As seen from the statistical analyses of the samples, the effect of using different filter aid materials did not significantly affect the changes in Fe content of sunflower oil (p > 0.05), whereas this fact affected the changes on Fe content of corn oil (p < 0.05). But the effect of different winterization times on the Fe changes of both oils was found significantly important (p < 0.05). It is estimated that metal transition occurs from filter aid material to oils. The effect of winterization time was found statistically important on the Cu changes of sunflower and corn oils (p < 0.05), whereas using different filter aid materials significantly affected the changes in Cu contents of sunflower oil (p < 0.05).

In terms of the stability of oils, Fe and Cu contents must be below 0.1 mg/kg and 0.02 mg/kg, respectively (Smouse, 1995). According to the Pakistan National Standard Institute, the limiting value of Ni must be below 0.50 mg/kg (Anon., 1990). Ni content found in sunflower oil was rather close to limiting value (maximum 0.54 mg/kg). For corn oil, it has been found quite low according to the limiting value and reached maximum to 0.23 mg/kg in the winterization by using perlite at a concentration of 0.6% in 24 h. The effect of kieselghur and perlite was found to be statistically important on the changes of Ni levels of sunflower oil (p < 0.05), whereas different winterization times caused significant effects for corn oil (p < 0.05).

Reyes and Campos (2006) used microwave acid digestion method in their study. They determined Ni and Cu levels as 4.18–4.30 mg/kg and 2.66–3.65 mg/kg in refined corn oil, respectively. Brevedan, Carelli, and Crapiste (2000) found Fe and Cu contents of sunflower oil as 3.6–15.1 mg/kg and 1.5–1.8 mg/kg, respectively. Generally, the results about trace metals differ in vegetable oils. It can be seen that our findings are not as high as compared with literatures. According to TS 888 (2003) and TS 886 (2005), Cu and Fe contents of refined sunflower and corn oil must be mostly 0.1–1.5 mg/kg. As Codex Alimentarius Standard (2009) found, Fe and Cu contents in refined and raw oils must be mostly 1.5–5.0 mg/kg and 0.1–0.4 mg/kg, respectively. The Fe and Cu results obtained (Table 4) are between maximum values proposed by Codex Alimentarius Standard (2009) for refined and crude oils.

Conclusions

Using different filter aid materials and applying different winterization times affected the oxidative stability of sunflower and corn oils. Within this framework, the peroxide value, conjugated diene and p-anisidine values and trace metal levels which are determined for the purpose of oxidation control increases during winterization in both oils. It is estimated that this increase is due to the holding time of oils in winterized tank, presence of atmospheric O₂ and ambient temperature during the analysis done. The slight increase in FFA levels of oil was found insignificant.

A significant positive correlation between the iodine values and conjugated diene values of the sunflower (r = 0.67) and corn oils (r = 0.70), was observed. During winterization processes, conjugated diene value was also increased due to the holding time. Also, positive and low correlations were found between p-anisidine values and Fe and Cu levels of sunflower (r₁ = 0.38, r₂ = 0.52) and corn oils (r₁ = 0.56, r₂ = 0.32).

When the filter aid materials were compared to each other, it was determined that perlite was more effective than kieselguhr on the oxidative stability of oils. Highest increase in free fatty acidity, peroxide value, conjugated diene value, p-anisidine and trace element content was determined for the winterization process in which kieselguhr was used. Smallest increases in these values were determined for the process in which no filter aid material was used. Higher levels of conjugated diene, p-anisidine, FFAs and trace metals obtained due to the slight oxidation could be declined or removed during deodorization process which will be applied after winterization step.