Corn germs (5.9% moisture, 42.6% oil by dry basis) as clean and dry material were sampled (2 batches of samples, 5 kg each) for this study from the oil production line of a commercial facility (Cargill, Bursa, Turkey) and were gifted to us. From the same production line, refined winterized corn oil samples (2 samplings, 1 kg each) were also gifted to us. All chemicals used were analytical or chromatographic grade and purchased from Merck Co. (Darmstadt, Germany) and Sigma Chem. Co. (St. Louis, USA). The internal standards used in the chromatographic analyses were purchased from Supelco (Bellefonte, PA, USA), Nu-Check (Elysian, MN, USA) and Sigma Chem. Co. (St. Louis, USA).

The moisture content of the corn germs was adjusted to 12% by water conditioning (Aydeniz et al., 2014; Yılmaz et al., 2015a) before cold pressing, since that moisture level was determined as the best point for maximum oil yield and ease of the operation by the pre-experiments. The prepared corn germs were cold pressed with a laboratory scale cold press machine (12 kg seed/h capacity, single head, 2 hp, 1.5 kw power, Koçmaksan ESM 3710, İzmir, Turkey) in two separate batches for the 2 replicates of the study. The cold pressing machine was operated at 30 rpm screw rotation speed, 10 mm exit die and max 40 ºC oil exit temperature. The cold pressed oils were first filtered through a filter bag which captures micron size particles (Miroil RB22Fs, Allentown, USA) to separate suspended materials and then centrifuged in a refrigerated centrifuge (Sigma 2-16 K, Postfach, Germany) at 6797xg for 10 min at room temperature. The obtained clear oil was weighed and placed into amber colored glasses, flushed with nitrogen and stored in a fridge during the analyses.

The refractive indices of the oil samples were measured with Abbe 5 (Bellingham and Stanley, UK) refractometer at 25 ºC. Apparent viscosity values were assessed by a Brookfield Viscometer (model DV II+Pro with Rheocalc software, Brookfield Eng. Lab., Inc., MA, USA) with LV-SC4-18 spindle and 30 rpm at 25 ºC. The specific gravity and turbidity (25 ºC) values of the oil samples were measured with the AOCS method Cc 10c-95 (AOCS Methods, 1997) and by a Hach 2100 AN Turbidimeter (USA), respectively. CIE color values (L, a* and b*) of the oils were determined using a Minolta Colorimeter CR-400 (Minolta Camera Co., Osaka, Japan). The sediment contents of the oils were measured gravimetrically according to Dündar Emir et al., (2015).

The free fatty acid contents and acid values of the samples were determined according to the AOCS method Ca 5a-40 and Cd 3d-63 (AOCS Methods, 1997). The peroxide values and iodine numbers were measured by the AOCS methods Cd 8-53 and Cd 1-25 (AOCS Methods, 1997), respectively. The p-anisidine value was determined by AOCS method Cd 18-90 (AOCS Methods, 1997). The saponification numbers and amounts of unsaponifiable matter were evaluated by following method Tl 1a-64 (AOCS Methods, 1997) and by the ISO 3596 method (ISO Method 3596, 2000). Total phenolic content by Folin-Ciocalteu reagent and antioxidant capacity by the Trolox equivalent antioxidant capacity (TEAC) were analyzed according to a previously described technique (Aydeniz and Yılmaz, 2012). The carotenoid contents expressed as total carotenoid and zeaxanthin in the cold pressed and refined corn oils were measured using an UV-Visible Spectrophotometer (Agilent 8453, Waldbronn, Germany) at 445 nm and petroleum ether: acetone (1:1, v/v) as solvent by a method based on Franke et al., (2010). The chlorophyll content expressed as mg pheophytin-a per kg oil was measured in accordance with method 13i-96 (AOCS Methods, 1997).

The fatty acid methyl esters (FAMEs) were prepared according to method Ce 2-66 (AOCS Methods, 1997). The identification of the fatty acid methyl esters was performed by Gas Chromatograph (Agilent Technologies 7890B, Palo Alto, CA, USA) equipped with a HP 88 capillary column (100m × 0.25mm ID × 0.2μm, J&W Scientific Co, CA, USA) and a flame ionization detector (FID) (Agilent Technologies, Palo Alto, CA, USA). The GC experimental conditions were as follows; oven temperature: 120 °C for 1 min, 175 °C (10 °C/min) for 10 min, 210 °C (5 °C/min) for 5 min and 230 °C (5 °C/min) for 5 min; injection volume 1 μL, injector split ratio 1:50, flow rate 2 mL/min, carrier gas hydrogen, injector and the detector temperatures were 250 and 280 °C, respectively. Fatty acid methyl esters were quantified by co-chromatography with a FAME mixture (37-components, C4-C24, Supelco, Bellefonte, PA, USA).

The sterol compositions of the oil samples were evaluated by the ISO 12228 method (ISO Method 12228, 1999). Sterol fractions separated on TLC after saponification of the oil samples were analyzed with Gas Chromatography (Agilent Technologies 7890B, Palo Alto, CA, USA) equipped with a flame ionization detector (FID) (Agilent Technologies, Palo Alto, CA, USA) and DB5 capillary column (30m × 0.25mm ID × 0.1 μm, J &W Scientific Co, CA, USA). The analysis was performed under the following GC conditions: oven temperature: 60 °C for 2 min, 60-220 °C (40 °C/min) for 1 min, 220-310 °C (5 °C/min) for 30 min, injection volume 1 μL, injector split ratio 1:100, flow rate 0.8 ml/min carrier gas hydrogen (30 ml/min), injector and the detector temperatures were 290 and 300 °C, respectively. Phytosterol identification was achieved by comparing the relative retention times from samples with those of commercially available standards (cholesterol, brassicasterol, stigmasterol, β-sitosterol) under the same operating conditions.

The α-Tocopherol contents of corn oil samples were determined by reverse-phase HPLC using an Shimadzu HPLC system (Shimadzu Corporation, Kyoto, Japan) fitted with a LC-20AT HPLC pump, a DGU-20A5R degasser, a CTQ-10ASVP column oven, a RF-20A diode array detector according to the procedure of Grilo Câmara et al., (2014) with minor modifications. 20 μL samples (0.15 g oil in 3 mL dichloromethane) were injected by an SIL-20AHT auto sampler into the Inertsil ODS-3 column (250mm × 4.6mm ID × 5μm, GL Sciences Inc., Japan). Isocratic elution was used with a flow rate of 1.6 mL/min. The mobile phase consisted of methanol: water (98:2 v/v). Detector wavelengths were 290 nm for excitation and 330 nm for emission. Tocopherol quantification was performed using α-tocopherol standard (Merck Co., Darmstadt, Germany).

The phenolic compounds were analyzed with a RP-HPLC system coupled to a SPD-M20A diode array detector (Shimadzu Corporation, Kyoto, Japan) according to Moulehi et al., (2012) with minor modifications. The extraction of phenolic compounds from all oil samples was performed by Garcia-Villalba et al., (2010) and all methanolic extracts were filtered through a 0.45 μm membrane filter before injection into the HPLC system. The separation was carried out on a Zorbax Eclipse Plus C18 column (250mm × 4.6 mm ID × 5μm, Agilent Technologies, USA) at 25 °C. The mobile phase consisted of acetonitrile (solvent A) and water with 0.2 % sulphuric acid (solvent B) and the flow rate was kept at 0.5 mL/min. The gradient program was as follows: 80% A/20% B 0–18 min, 70% A/30% B 18-24 min, 60% A/40% B 24–30 min, 50% A/50% B 30-36 min, 40% A/60% B 36–40 min, 45% A/55% B 40-45 min, 35% A/65% B 45-50 min, 20% A/80% B 50-52 min, 10% A/90% B 52–54 min, 100% B 54-70 min, 45% A/55% B 70-72 min, 80% A/20% B 72-74 min. The injection volume was 20 μl, and obtained peaks were monitored at 280 nm. Phenolic compounds were identified according to their retention times against those of commercially available standards.

Thermal parameters of the oil samples were determined with a Perkin-Elmer 4000 Series Differential Scanning Calorimeter (DSC) (Groningen, The Netherlands). The instrument was previously calibrated with Indium and Zinc standards and purged with nitrogen at 50 mL/min flow rate. First, 5-7 mg oil sample were weighed and sealed hermetically in aluminium pans. Then the temperature program was applied by heating samples from room temperature to 110 °C by 10 °C/min rate, then cooling to -70 °C by 10 °C/min rate and holding at that temperature for 3 min, and finally heating again to 50 °C by 5 °C/min rate. Pyris 1 Manager Software was used to calculate the thermal parameters, melting temperature (T_m), melting enthalpy (ΔH_m), crystallization temperature (T_c) and crystallization enthalpy (ΔH_c) (Yılmaz et al., 2015c). Oxidative induction time (OIT) was also measured with the DSC by heating the sample from 30 °C to 130 °C at 20 °C/min heating rate, while flushing nitrogen (50 mL/min), and then by isothermal temperature programming at 130 °C with 50 mL/min purified oxygen (99.8%) application on the samples. The thermograms of the samples were calculated by using 1 Manager Software to calculate the OIT value (Yılmaz et al., 2015c).

The volatile compounds in the corn oils were analyzed according to the technique of Krist et al., (2006) with minor modifications. Volatile compounds were collected from the oil samples with the headspace solid phase microextraction (SPME) (2 cm to 50 / 30 μm DVB / Carboxen / PDMS, Supelco, Bellafonte) technique (Pawliszyn, 2012). First, 2 g oil sample were weighed into a 40 mL amber SPME vial (Supelco, Bellafonte, USA) and 1 g NaCl and 10 μL internal standards-IS (0.1 μL of 2-methyl-3-heptanone and 6 μL of 2-methylvaleric acid dissolved in 1 mL methanol) were added. The closed vial was vortexed for 1 min and put in a water bath (GFL, Germany) at 40 °C for 10 min to collect the volatiles in the headspace. Then, a SPME (2 cm to 50 / 30 μm DVB / Carboxen / PDMS, Supelco, Bellafonte) needle was inserted into the vial and SPME fiber was exposed to a depth of 2 cm in the headspace for 10 min at 40 °C in a water bath. Finally, the volatile compounds isolated on the fiber were injected into the GC/MS (Agilent 6890N/Agilent 5875C mass spectrometer, Agilent Technologies, Wilmington, DE, USA) equipped with a non polar HP5 MS column (30m × 0.25 mm ID × 0.25 μm, J&W Scientific, Folsom, CA). The GC conditions were described in a previous study (Yılmaz et al., 2015a). Identification of the volatiles was based on the comparison of the mass spectra of unknown compounds with those in the National Institute of Standards and Technology (NIST, 2008) and Wiley Registry of Mass Spectral Data, databases (McLafferty, 2005) and Retention (Kovats) index. Volatile compound quantification was based on the relative abundances calculated positively by the Equation given below.

Sensory attributes (appearance, color, odor, taste/flavor) of each oil sample were assessed by 110 individual panelists (61 females - 49 males aged 20-45 years) using a 5-point hedonic scale (1 = dislike extremely to 5 = like extremely). All the oil samples were coded with three digit numbers and presented in clear glass cups with metal lids. Panelists first evaluated the appearance and color of each sample, and then tasted and swallowed it. Two samples of each treatment group were tested twice by each of the panelists at different sessions, randomly. The panelists were provided with warm water and unsalted crackers to cleanse the palate between sensory evaluations. All evaluations were at room temperature and under day light. The mean scores of the sensory attributes collected by the hedonic scales were calculated.

Cold pressing of the corn germs was duplicated and 2 batches of refined winterized corn oils were sampled to be the duplicate samples. For each of these duplicate samples, all analyses were performed in at least duplicate, if not triplicate. All values in the tables provided average of these determinations ± standard error. Comparison of the measured properties for corn oil samples and statistically significant differences between groups were determined by one-way ANOVA and Tukey’s tests. Sensory evaluation data was analyzed with non-parametric Kruskal-Wallis and Dunn’s test. Minitab ver. 16.1.1 (Minitab, 2010) and SPSS package (SPSS, 1994) programs were used for all statistical evaluations. There was a minimum 95% level of confidence for the statistics.

The measured physical properties of the cold pressed corn oils (CPCO) and refined winterized corn oils (RWCO) are presented in Table 1. The two oil samples can be visualized in Figure 1 as well. Most of the differences in appearance can easily be observed from Figure 1 and Table 1.

Although there is no significant difference between samples for refractive indices, statistically significant difference exists for the viscosity values. A significant difference exists between the two oil samples for the turbidity value. It is quite expected that refined winterized oils are truly clear since during the refining procedures all particles causing certain turbidities were removed. The cold pressed oil was only centrifuged to remove water and suspended materials, but some fine suspended materials and high molecular weight fractions still remain inside the oil. Supporting this assumption, the sediment content of CPCO was much higher than RWCO, indicating some possible reasons for high turbidity. Statistically significant differences also exist for the color parameters (Table 1). The L value (0-black/100-white) or luminosity of RWCO is much higher, indicating a correlation with the low turbidity measured. The a* value (+red/-green) indicates that the CPCO is more reddish than the other sample. Actually RWCO is rather greenish in color as shown by the a* value. The other color parameter, b* value (+yellow/-blue) showed that RWCO is more yellow than the CPCO sample (Table 1 and Figure 1). Strecker et al., (1996) suggested 1.4753 as a refractive index, and 30.80 and 18.15 cP as viscosities at 40 and 60 ºC for corn oils. Generally, our results concur with these values.

The main chemical characteristics of the corn oil samples are also summarized in Table 1. Free fatty acidity (FFA) of CPCO was 1.16 %, and higher than that of the RWCO (0.23 %), but this can be considered quite low compared to other cold pressed oils (Yılmaz et al., 2015a; Yılmaz et al., 2015b). The measured acid values are in accordance with the FFA values as well. According to the codex standard for vegetable oils (Codex Stan, 2015), the upper limit of acid value is 4.0 mg KOH/g oil. Hence, the CPCO is quite acceptable in terms of acidity. The peroxide value (PV) of CPCO (9.29 meq O₂/kg) was significantly higher than that of the RWCO (6.67 meq O₂/kg). According to the Codex Standard for Named Vegetable Oils (Codex Stan, 2015), the PV must be lower than 15 meq O₂/kg for cold pressed oils, and CPCO can be acceptable by this limit as well. Similarly, p-anisidine (p-AV) values agree with the PVs measured. The oxidation level of oil heavily depends on its production, processing, and storage conditions; hence, only acceptable upper limits could be meaningful for suitability of consumption. Each oil might have different oxidation stability, depending on its chemical composition, purity, and storage conditions (O’Brien, 2004). The PV and p-AV reported in this study indicate that CPCO does not pose any immediate health hazard based on oil oxidation, and is consumable as it is pressed. The iodine numbers indicate significant difference between the two samples (Table 1). This is also expected, since in the RWCO sample, the waxes and higher melting point triglyceride fractions must have been removed by the winterization process, whereas those higher melting components are likely to remain in the CPCO sample after pressing and centrifugation. In the literature, iodine values for corn oils are given between 118 and 133 (Strecker et al., 1996; O’Brien, 2004). The sediment content, color parameters and turbidity values have indicated that an adsorbent treatment and/or full industrial filtration of CPCO after pressing may significantly improve these properties. The saponification number of CPCO was higher than that of the RWCO (Table 1). Both Strecker et al., (1996) and O’Brien, (2004) reported saponification numbers of corn oils to be 187 to 193. The number measured in the CPCO (205.43 mg KOH/g oil) sample is higher than those reported in the literature, which may be due to the high molecular weight component present in the unrefined cold pressed oil. The unsaponifiable matter content of the samples was 1.44 and 1.13% for the CPCO and RWCO, respectively. The unsaponifiable number for corn oils is given from 1.3 to 2.3% (O’Brien, 2004). Hence, there is an agreement between our findings and the literature data.

Some other chemical properties related to the nutritional value of the oils were also reported (Table 1). Total phenolic content, antioxidant capacity, total carotenoid and zeaxanthin contents of the CPCO sample were significantly higher than those measured in the RWCO sample. For both samples, there are no quantifiable levels of chlorophyll. These results clearly prove that most phenolics and pigments are lost during the refining processes. These compounds constitute antioxidant capacity, and in fact no measured antioxidant capacity was observed in the RWCO sample. This suggests an important nutritional loss as well. In one study (Daugan et al., 2011), the β-carotene content of commercial corn oil was reported as 0.91 ppm. Another study (Rafalowski et al., 2008) reported the β-carotene content of cold pressed corn oil as 3.56 mg/100 g oil, and indicated its absence in the refined oil. Tuberoso et al., (2007) reported the chlorophylls content as 4.9 mg/kg, β-carotene content as 0.9 mg/kg, and TEAC activity to be 2.30 mmol/L for corn oil. The total phenolic content as caffeic acid equivalent for corn oil was reported as 1.26 mg CAE/100 g oil, and DPPH radical scavenging activity and EC50 values were reported as 11.1% and 14.8 μg, respectively (Siger et al., 2008). Our results and literature usually concur, but some differences exist depending on the material source and differences in analytical methods.

Five fatty acids were quantified in the corn oil samples (Table 2). The fatty acid composition of the CPCO and RWCO samples were very similar. Both included around 53% linoleic and 30% oleic acid as the major fatty acids.

In general, the fatty acid compositions of our samples concur with the literature (Strecker et al., 1996; O’Brien, 2004; Codex Stan, 2015; Rafalowski et al., 2008). There were statistically significant differences between the two samples for the sterol composition (Table 2). 24-Methylene cholesterol, campesterol, Δ-7-campesterol, Δ-5,23-stigmasterol, cholesterol and Δ-5,24-stigmasterol were absent in the RWCO, while quantified in small amounts in the CPCO sample. Among the fifteen different sterols, the majority was composed of β-sitosterol (60-62%), campesterol (19-22%), stigmasterol (6-7%) and sitosterol (2-4%). These findings are in agreement with those ranges given in the standard for vegetable oils (Codex Stan, 2015). Strecker et al., (1996) indicated β-sitosterol, campesterol and stigmasterol as the major sterols of corn oils. The only tocopherol quantified in these samples was α-tocopherol (Table 2). Amount of α-tocopherol in the CPCO (248.83 mg/kg) was significantly higher than that found in the RWCO (159.60 mg/kg) sample. Clearly during refining, significant losses of α-tocopherol have occurred. Rafalowski et al., (2008) have determined the tocopherol composition of both refined and cold pressed corn oil among ten other oils. They have quantified 27.22, 91.76, and 5.93 mg/100 g oil of α-tocopherol, γ-tocopherol and δ-tocopherol in the cold pressed sample, while the same tocopherols were quantified in refined oils as 26.11, 44.31, and 0.0 mg/100 g oil, respectively. Our study concurs in this finding that during refining most tocopherols are lost. Strecker et al., (1996) has given the tocopherol composition of corn oil as 19.1, 94.2, and 4.2 mg/100 g for the α-, γ- and δ-tocopherols. Clearly, we measured only α-tocopherol, and our finding is in agreement with the literature.

Some flavonoids and phenolic acids were measured and reported in Table 3 for the CPCO sample.

No phenolics were quantified in the refined corn oil samples. It can be said that during refining, all phenolics were degraded or lost for the refined corn oil. Among the analyzed 8 flavonoids, only hesperidin (23.81 mg/kg), and rutin (14.17 mg/kg) were quantified. Except for o-coumaric acid, gallic, syringic, trans-ferulic and rosmaniric acids were quantified as the phenolic acids (Table 3). In one study (Pandey et al., 2013), different maize samples (baby corn, raw, boiled, baked, fried, ripe corns, popcorn, stylar) were analyzed for phenolic acid composition. Raw maize contained 4.6, 3.2, 0.99 and 0.57 μg/g fresh weight of gallic, caffeic, ferulic and salicylic acids, respectively. Another study (Tuberoso et al., 2007) reported the phenolic compositions of nine different oils including maize oil. In maize oil, 2.8 ppm vanillin, 0.5 ppm ferulic acid, and 0.9 ppm trans-cinnamic acid were quantified. Siger et al., (2008) have measured the phenolic acid content of cold pressed corn oil. They have quantified 1.7, 1.9, 5.8, and 0.6 μg/100 g oil of p-hydroxybenzoic, p-coumaric, ferulic, and sinapic acids,respectively. All these literature data indicate that cold pressed corn oils include some important phenolic compounds, and these molecules are bioactive components for some health benefits, including antioxidant, anti-bacterial, anti-inflammatory, anti-allergic, anti-thrombotic, anti-carcinogenic activities, hepatoprotective and vasodilatory actions (Pandey et al., 2013).

Crystallization and melting temperatures and enthalpies of the two corn oil samples are reported in Table 4. Since a thermal cycling procedure was applied during the thermal analysis, it was possible to observe the whole behavior of the corn oils during heating (up to 50 ºC) and cooling (up to – 70 ºC) stages.

There were some statistically significant differences between the samples. Crystallization onset and peak temperatures were lower for the cold pressed sample than the refined-winterized sample. Likewise, melting peak temperature and enthalpy were lower for the cold pressed sample. These differences are expected, and might be caused by the waxes and/or high melting fractions present in the CPCO sample. Those fractions were removed as expected from the RWCO sample by the winterization process applied previously. It was stated that fats and oils exhibit a range of crystallization and melting temperatures, rather than definite points since they are heterogeneous mixtures of different triglycerides. In addition, a shift between crystallization and melting onset temperatures are acknowledged as a common phenomenon for unsaturated oils (Co and Marangoni, 2012). We have not found thermal data for corn oils in the literature, and this study may provide this information for interested researchers.

The oxidation induction times (OIT) of the samples determined by DSC are also presented in Table 4. Clearly, the cold pressed sample was significantly more stable against oxidation due to its prolonged induction time (65.74 min). Since cold pressed oils retain most of their minor components including phenolics and tocopherols compared to refined oils, they might have higher oxidation stability. This was also the case for these samples as evidenced from the measured total phenolic contents and antioxidant activity (Table 1) and phenolic compositions (Table 3).

The volatile aromatic compounds determined in the CPCO and RWCO samples are presented in Table 5. There are 22 volatiles detected in the cold pressed, and 7 volatiles and hexane detected in the refined sample. Clearly, most volatiles are lost in the refined sample, possibly during the alkali refining and deodorization processes.

The most abundant volatiles present in the CPCO sample were hexanal, acetic acid, 3-methyl-butanal, hexanoic acid and butanoic acid while only minor amounts of hexanal, heptanal, hexanoic acid, limonene, 2-ethyl hexanol, 3-carene, and nonanal remained in the RWCO sample. It is well known that in any food product, the characteristic aroma is provided by the volatiles present all together, but not in an order of higher concentrations, rather in the order of aroma threshold value. The aroma threshold value is defined as the minimum concentration of any aromatic volatile which can be perceived by humans. It is well known that the aroma threshold of some volatiles is much lower, and they can contribute much more to the perceived aroma at their very low concentrations than the other volatiles present in the same food at very high amounts (Meilgaard et al., 1991). The green, fresh, herbal, and sour attributes are the most frequent descriptors for the measured aromatic volatiles in the samples (Table 5). There is a fairly limited number of studies about volatile compounds of corn oils. In one study (Tuan, 2011), 29 volatile compounds were identified in the corn oils. There were some similarities between this and that study, and octane, 2,6-dimethyl-3-heptanone, 2-heptenal, 2-pentylfuran, nonanal, hexedecanoic acid, 9-octadecenoic acid ethyl ester were determined as the same or very similar compounds. The other volatiles were usually different isomers of the same base compounds, but acetic acid, α-pinene, benzaldehyde, limonene, and 3-carene were identified only in our study. These volatiles distinctly provide almond, citrus, pine, buttery and terpenic aroma notes. It is quite possible that cold pressing of the corn germ yielded the advantage of the presence of these various volatiles for the oil. Solid Phase Microextraction (SPME) with different resins was studied to identify the volatiles in corn and soybean oils (Steenson et al., 2002). In corn oil, 13 different volatiles were identified with 2-heptenal, trans-2-octenal, pentane and 2,4-heptedienal as the most abundant ones. These and others listed are similar to those determined in this study (Table 5). In another study (Goicoechea and Guillén, 2014), the volatile compounds generated in corn oils during storage at room temperature for various times (12-103 months) were measured. Usually oxidized samples included many secondary oxidation compounds including alkanals, 2-alkenals, 2,4-alkadienals, acids, alcohols, ketones, alkylfurans, lactones and α,β-unsaturated aldehydes. As storage time extended, the kinds and amounts of volatiles were enhanced. Even in this study, at earlier storage times, some similarities between the detected volatiles (hexanoic and butanoic acid, hexenal, nonanal, limonene, and some furans etc.) of our samples were observed.

The presence of hexane (22.93 μg/kg oil) in the RWCO sample is an important issue to address here. Since it is absent in the CPCO sample, it can easily be argued that hexane is originated from the solvent extraction process applied previously to the RWCO sample. The commercial hexane used in oil extraction is a mixture of low molecular weight linear, branched and cyclic saturated alkanes. It is acknowledged that hexane is a toxic substance, and acute inhalation of humans to high doses creates dizziness, giddiness, slight nausea, and headache; chronic exposure is characterized with polyneuropathy in humans, with numbness in the extremities, muscular weakness, blurred vision, headache, and fatigue (TGK, 2002; Mirghani and Che Man, 2003). It is stated in the Turkish codex standard for the determination of the maximum limits of definite contaminants in foods (TGK, 2002) that the maximum residual limit of hexane in fats/oils is 5 mg/kg, whereas the United States Environmental Protection Agency has published some health data for hexane but it was incomplete (Mirghani and Che Man, 2003). It was also stated that hexane residues in refined foods would be less than 2 ppm, since if 80 g fat per day is consumed by a 70 kg weight person, such a residual level would be equivalent to 2.29 μg/kg/day, which is a toxicologically effective level (Hautfenne et al., 1987). Although the residual hexane level in the RWCO sample is much lower than the permitted maximum level, CPCO still provides one of the major advantages of being free of any solvents or chemicals.

Hedonic scores of appearance, color, odor and taste/flavor for the two corn oil samples with 110 volunteer consumers are shown in Figure 2.

All sensory scores for the CPCO sample were significantly lower than the RWCO sample. Except for the taste/flavor score of the CPCO sample, all attribute scores for both samples were higher than the neutrality point of 3.0 (neither like nor dislike); hence, it could be claimed that consumers accepted both samples. Improvements especially in the taste/flavor and odor of the CPCO sample were determined as the main requirements based on the consumer test. A simple adsorption and/or combined adsorption/winterization process would improve sensory characteristics as well as turbidity and sediment values of the CPCO sample. This process may not cause significant losses in bioactive compounds. A deodorization process might greatly improve the odor characters of CPCO, but deodorization usually causes heavy nutrient losses (O’Brien, 2004) and we would not suggest it for cold pressed oils. Similar consumer responses to cold pressed capia pepper seed oil (Yılmaz et al., 2015a) and tomato seed oil (Yılmaz et al., 2015b) were observed in our laboratory.