Study of the Antioxidant Power of the Waste Oil from Palm Oil Bleaching Clay

: Palm oil is one of the most consumed oils, one of whose reﬁning steps is the removal of pigments and other substances using bleaching clay as adsorbent. Worldwide production of this oil was 70 million tons in 2017, requiring 1 t of clay to produce 1000 t of reﬁned oil. The residual bleaching clay, having an oil fraction (12.70%) rich in phenolics, carotenoids and tocopherols, was extracted in this study with ethanol to obtain an antioxidant-rich palm oil bleaching extract (POBE), with the aim of using it as a natural antioxidant source. The POBE antioxidant capacity determined by the DPPH method corresponded to a 20.29% inhibition of radical formation. The POBE was also tested for its potential to enhance oxidative stability of passion fruit, pracaxi and Brazil nut oils used as reference oils, and compared to common synthetic antioxidants (tert-butylhydroquinone and propyl gallate), either separately as controls or in mixtures with them. Besides the increased oxidative stability of these oils induced by the POBE, a positive synergistic e ﬀ ect between it and the synthetic antioxidants was observed. These results taken together suggest that the exploitation of the waste oil from bleaching clay as an additive to improve the oxidative stability of biofuels or lubricating oils is feasible.


Introduction
Palm (Elaeis guineensis) is grown on over 11 million hectares worldwide [1], and its oil is one of the most important branches of agribusiness in many countries. In 2017 alone, worldwide palm oil production reached around 70 million tons [2]. Because of its high content in fatty acids and substances with strong antioxidant activities, such as tocopherols, tocotrienols and carotenoids, palm oil is one of the most consumed in the world and serves a variety of sectors, among which are the food, cosmetic, biofuel and energy industries [2][3][4][5]. However, in order to meet such market demands, it is necessary to remove impurities, pigments and unpleasant odors [6] through a process known as vegetable oil

Clay Characterization
The clay crystalline structure was investigated before and after the bleaching process by Powder X-ray Diffraction (PXRD) using a X´PERT PRO MPD (PW 3040/60) diffractometer (PANalytical, São Paulo, SP, Brazil) operating with Cu radiation and a 1.54 Å wavelength. Thermal behavior was investigated by thermogravimetric and derivative analysis (TG/DTG) under synthetic air (21% oxygen, 78% nitrogen and 1% other trace gases) at a 10 • C/min heating rate, in the temperature range of 25-800 • C, using a thermogravimetric analyzer, model DTG-60H (Shimadzu, Kyoto, Japan). The oil content in bleaching clay was quantified by nuclear magnetic resonance (NMR) using a Minispec mq 7.5 analyzer (Bruker Optics, Ettlingen, Germany).

Obtaining and Characterization of Palm Oil Bleaching Extract and Vegetable Oils
Palm oil bleaching extract (POBE) was obtained according to Fernandes et al. [29] with some modifications. Briefly, clay after bleaching was ground, mixed with 400 mL of ethanol, left at rest for seven days, stirred for one minute each day and finally vacuum filtered. The solvent was removed from the POBE by rotary evaporation at 40 • C.

Antioxidant Activity of Palm Oil Bleaching Extract and Vegetable Oils
Antioxidant activities of vegetable oils and POBE were determined by the 2,2-diphenyl-1-picrylhydrazyl radical (DPPH) scavenging method, according to Brand-Williams et al. [31] with an Evolution Array UV-Vis spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA) at 515 nm wavelength. Results were expressed as percentage inhibition of radical formation (% I) in the sample compared to the initial DPPH concentration (% I), according to the equation % I = ((Ac − As)/Ac) × 100, where Ac is the absorbance of the control and As that of the sample. Results were expressed as percentage inhibition of radical formation (% I) in the sample compared to the initial DPPH concentration.
Total phenolic compounds were determined according to the Folin-Ciocalteu method described by Singleton et al. [32], using gallic acid as the standard. Quantitation was performed using the above-mentioned UV-Vis spectrophotometer at 760 nm wavelength and the equation TPC (mgGA/g) = Các.gallic × V volumetric balloon (mL) × F/m(g) × 103, where TPC = phenolic compound concentration; Các.gallic = concentration in gallic acid; V volumetric balloon (mL) = extract volume after extraction; F = dilution factor for reading; and m = sample mass. Results were expressed in milligrams of eqivalent gallic acid per gram of sample (mgGAE/g).
Total carotenoids were determined according to the methodology described by Rodriguez-Amaya [33] by wavelength scanning from 300 to 600 nm with the above UV-Vis spectrophotometer. The β-carotene scanning spectrum showed maximum absorption at 450 nm, which was used as a standard for quantifying carotene concentration in oils. Total carotenoid content (TC) was calculated using the β-carotene specific extinction coefficient in petroleum ether (ε = 2592), according to the equation TC (µg/g) = As. × 10 4 v/m × ε, where As. is the absorbance; v is the added petroleum ether volume in mL; and m = sample mass in g. Results were expressed as mg/g of total carotenoids.

Oxidative Stability of Vegetable Oils
The oxidative stability of vegetable oils was determined as the oxidative induction period according to the AOCS Cd 12b-92 method using an oxidative stability instrument, model Rancimat 873 (Metrohm, Herisau, Switzerland). Water conductivity was measured along the air bubbling, which was performed at a flow rate of 10 L/h in each oil (5.0 g) and a temperature of 110 • C. The oxidative induction period (OIP) was assumed as the time interval between the start of a measurement and the conductivity increase as a result of the oxidation products' solubilization. Samples of each oil were weighed and homogenized with a magnetic stirrer for 10 min before performing oxidative stability assays.

Synergistic Effect of Palm Oil Bleaching Extract with Synthetic Antioxidants
Binary mixtures were prepared using 500, 1000 and 2000 ppm POBE together with both the above synthetic antioxidants at a concentration of 50 or 100 ppm and analyzed for their effects on the vegetable oils' oxidative stability.

Clay Characterization before and after the Palm Oil Bleaching Process
Clay samples before and after the palm oil bleaching process were analyzed by powder X-ray Diffractometry (PXRD) to detect any changes in the crystalline phases induced by it:

•
The diffractograms shown in Figure  It is possible to infer the presence of iron oxide, which indicates that there is mica in the material, while the peak at 2θ = 27 • may correspond to impurities such as quartz, as suggested by Santos [35].
The clay diffractogram before bleaching ( Figure 1A) agrees with what is found in the literature for the same material used for commercial anthocyanin adsorption from red cabbage [36], while that after bleaching shows a decrease in peak intensity ( Figure 1B), indicating decreased material crystallinity. This result may have been due to partial dissolution of the clay crystalline lamellae after acid treatment in the bleaching process [37,38]. However, even after bleaching the clay structure remained practically the same, suggesting that it may be profitably reused after removal of adsorbed palm oil.
The content of clay-impregnated oil quantified by low field Nuclear Magnetic Resonance (NMR) was approximately 12.70%, i.e., a lower value than reported in the literature (20%-30%) [12,39], probably due to the higher efficiency of our bleaching process.
Energies 2020, 13, 804 5 of 13 phases both before and after bleaching: montmorillonite (Na0.3(AlMg)2Si4O10OH26H2O-JCPDS 12-0219), illite ((K,H3O)Al2Si3AlO10(OH)2-JCPDS 26-0911), kaolinite (Na0.3Al4Si6O15(OH)64H2O-JCPDS 29-1490), quartz (SiO2-JCPDS 05-490) and mica (K-Mg-Fe-Al-Si-O-H2O-JCDPS 02-0227). It is possible to infer the presence of iron oxide, which indicates that there is mica in the material, while the peak at 2θ = 27° may correspond to impurities such as quartz, as suggested by Santos [35]. The clay diffractogram before bleaching ( Figure 1A) agrees with what is found in the literature for the same material used for commercial anthocyanin adsorption from red cabbage [36], while that after bleaching shows a decrease in peak intensity ( Figure 1B), indicating decreased material crystallinity. This result may have been due to partial dissolution of the clay crystalline lamellae after acid treatment in the The thermogravimetric curves of Figure 2 show a substantial difference in clay mass loss before and after treatment: • Two peaks in the clay DTG curve can be observed before bleaching (Figure 2A) in the temperature range of 25-120 • C, pointing out successive mass losses likely associated with free water evaporation and structural water loss from some mineral. A third, almost imperceptible event occurring at 120-190 • C could be attributed to the decomposition of the hydroxyl groups of compounds constituting a clay structure, such as Al(OH) 3 and Fe(OH) 2 [38]. Another event between 400 and 540 • C that occurred with approximately 4.6% mass loss was probably the result of the decomposition of the hydroxyl groups (structural water), mainly from kaolinite [40]. The last 2.1% mass loss observed between 550 and 700 • C is also likely to have been caused by the loss of hydroxyls from other clay minerals. Similar results were found by Miranda et al. [38] for montmorillonite clays, with a loss of 8.9% in the first event, 5.1% in the second event and 1.76% corresponding to structural dehydration.
Even TG/DTG curves of clay after bleaching highlight a first event in the range of 25 to 160 • C with 4.9% mass loss ( Figure 2B), possibly due to evaporation of free water molecules, and at least two successive events in the temperature ranges of 170-270 • C and 270-340 • C. Considering that between 170 and 340 • C there was no clay mass loss before bleaching (Figure 2A), the last two events may be attributed to the loss of hydroxyl groups (structural water), ions or compounds adsorbed on clay lamella, probably originating from the bleaching process (7.1%), as well as the burning of oily organic matter adsorbed on clay (12.5%), respectively. The latter hypothesis appears to be confirmed by the similarity of clay thermal profile after bleaching between 270 and 340 • C to those found in DTG curves of vegetable oils [22,41,42]. As expected, the 12.5% loss attributed to organic material burning is very close to the content of oil embedded in clay determined by low-field NMR (12.7%). Finally, a last event, which occurred between 410 and 600 • C, indicates successive mass losses (6.1%) that can be attributed to clay hydroxyls, mainly kaolinite and other clay minerals [40]. suggesting that it may be profitably reused after removal of adsorbed palm oil.
The content of clay-impregnated oil quantified by low field Nuclear Magnetic Resonance (NMR) was approximately 12.70%, i.e., a lower value than reported in the literature (20%-30%) [12,39], probably due to the higher efficiency of our bleaching process.
The thermogravimetric curves of Figure 2 show a substantial difference in clay mass loss before and after treatment:

•
Two peaks in the clay DTG curve can be observed before bleaching (Figure 2A) in the temperature range of 25-120 °C, pointing out successive mass losses likely associated with free water evaporation and structural water loss from some mineral. A third, almost imperceptible event occurring at 120-190 °C could be attributed to the decomposition of the hydroxyl groups of compounds constituting a clay structure, such as Al(OH)3 and Fe(OH)2 [38]. Another event between 400 and 540 °C that occurred with approximately 4.6% mass loss was probably the result of the decomposition of the hydroxyl groups (structural water), mainly from kaolinite [40]. The last 2.1% mass loss observed between 550 and 700 °C is also likely to have been caused by the loss of hydroxyls from other clay minerals. Similar results were found by Miranda et al. [38] for montmorillonite clays, with a loss of 8.9% in the first event, 5.1% in the second event and 1.76% corresponding to structural dehydration.
Even TG/DTG curves of clay after bleaching highlight a first event in the range of 25 to 160 °C with 4.9% mass loss ( Figure 2B), possibly due to evaporation of free water molecules, and at least two successive events in the temperature ranges of 170-270 °C and 270-340 °C. Considering that between 170 and 340 °C there was no clay mass loss before bleaching (Figure 2A), the last two events may be attributed to the loss of hydroxyl groups (structural water), ions or compounds adsorbed on clay lamella, probably originating from the bleaching process (7.1%), as well as the burning of oily organic matter adsorbed on

Physicochemical Characterization of Palm Oil Bleaching Extract and Vegetable Oils
The fatty acid profiles of passion fruit (PFO), pracaxi (PO) and Brazil nut (BNO) oils presented in Table 1 show that they consist mainly of unsaturated fatty acids, which makes them particularly susceptible to oxidation. For this reason, they were chosen to test the POBE antioxidant power. PFO had linoleic (64.32%), oleic (18.49%) and palmitic (10.91%) acids as the major components. Since the first one is a polyunsaturated fatty acid, its larger content compared to other reference vegetable oils suggests higher susceptibility to lipid oxidation [43,44]. The same type of oxidative sensitivity, although less pronounced, would be expected for BNO, of which these acids constitute 33.86%, 39.22% and 14.37%, respectively, in agreement with the composition reported by Rodrigues et al. [45]. On the other hand, PO had a higher content of saturated fatty acids, namely behenic (14.79%), lignoceric Energies 2020, 13, 804 7 of 13 (10.88%) and stearic (5.77%) acids, and a lower content of linoleic acid (11.81%), from which higher oxidative stability would be expected.
Since palm oil refining did not significantly alter its fatty acid composition, its oxidative stability, which is generally attributed to the high concentration of saturated fatty acids [46], should not have been influenced. As expected, the residue removed from clay after bleaching had a fatty acid composition very similar to that of crude and refined palm oils, with higher percentages of oleic (44.32%), palmitic (39.82%), linoleic (9.13) and stearic (4.25%) acids.
Resolution RDC No. 270/2005 of ANVISA [47] establishes maximum limits of acidity and peroxide indices of 4 mgKOH/g and 15 meqO 2 /kg, respectively, for unrefined and cold-pressed oils and fats intended for human use. These values can be compared with the basic quality parameters of vegetable oils listed in Table 2, in order to clarify possible POBE effects on their oxidative stability. PFO and PO are in accordance with current Brazilian legislation, while BNO had a slightly higher acidity index, probably due to the presence of undesirable substances in seeds resulting from the selection or their inadequate storage [48]. High acidity indices, like that detected in BNO and to a lesser extent in PO, are indicative of a high content of free fatty acids resulting from the action of vegetable or microbial lipases [49]. Indeed, seeds of oil-rich species, such as pracaxi and Brazil nut, are especially susceptible to the action of these enzymes because of their exposition during harvesting that increases their moisture content and causes hydrolytic rancidity [50].
The oxidative induction periods (OIPs) of PO (1.93 h) and BNO (2.10 h) were significantly shorter than those of PFO (3.40 h), which indicates higher oxidative stability of the last oil, contrary to what was expected by the fatty acid profile. This means that not only the lipid composition, but also other quality parameters should be taken into account when assessing oxidative stability, such as the acidity and peroxide indices. Both do in fact indicate the oil state of preservation and quality, by evaluating the products resulting from lipid oxidation, i.e., the degree of deterioration. The relatively short OIP values of the three reference oils tested suggest the possibility of improving the oxidative stability of other oils and biofuels as well, by supplementing them with antioxidant-rich waste oils such as that proposed in this study.

Bioactive Compounds and Antioxidant Activity of Palm Oil Bleaching Extract and Vegetable Oils
The contents of bioactive compounds in the oils investigated in this work are listed in Table 3. The contents of total phenolic compounds (TPCs) in all oils tested indicate the presence in their chemical structures of hydroxyl groups and aromatic rings, both in simple and polymeric forms, which contribute to their antioxidant power. However, this content was particularly high in crude palm oil and POBE and significantly lower in the reference vegetable oils. In addition, as expected, the lowest content was found in refined palm oil, hence demonstrating the effectiveness of the bleaching process. Therefore, it is likely that a portion of the TPCs was degraded during the oil refining process, but another one was concentrated in the POBE and clay. It is also possible to observe that a large amount of total carotenoids (TCs) was lost during bleaching, because the POBE had a TC value (357.4 ppm) that was almost thrice that of refined palm oil, but less than half of that of crude oil. According to Rossi et al. [52], the use of clays and synthetic silica mixtures in the bleaching process is responsible for the removal of no less than 30% to 50% of crude palm oil carotenoids, suggesting that residual oily extract could be an effective antioxidant additive. However, contrary to what is often stressed in the literature, there is no direct relationship between the TPC content and the antioxidant activity of natural products, many other causes being involved in the latter [53]. The TPC content does not determine antioxidant activity, but rather the structure of phenolic compounds. These, depending on the number and position of the hydroxyl groups, as well as replacements in their aromatic rings, may or may not present antioxidant properties [5].
Crude palm oil showed an antioxidant activity expressed as percentage inhibition of DPPH . (87.1%) not so different from that reported in the literature (59.02-88.16%), whose broad variability may have depended on the different industrial processes used [54,55]. However, the refining process led to an almost 80% decrease of this activity, confirming the literature data [55]. Although much lower than palm oil antioxidant activity, the one of the POBE (20.3%) was about 5-10-fold those of PFO and PO, respectively, and very close to that of BNO (Table 3). Since the TPC and TC contents of PFO and BNO are close to each other and the OIP of the former much longer than that of the latter (Table 2), it is likely that the oxidation inhibition mechanisms of these bioactive compounds are distinct and the inhibition percentages not proportional within of the oil. For this same reason, although PO had the lowest bioactive compound contents (TPC = 84.5 mgGAE/g; TC = 10.9 mg/g) along with the lowest DPPH . scavenging capacity (2.2%) among the three reference oils, its OIP was not significantly shorter than that of BNO.

Effect of Palm Oil Bleaching Extract or Synthetic Antioxidants on Vegetable Oil Stability
POBE addition to the reference vegetable oils led to a significant OIP lenghtening at concentrations ≥ 200 ppm, although this effect was well below that observed using tert-butylhydroquinone (TBHQ) and propyl gallate (PG) as synthetic antioxidants (Figure 3), whose overall effectiveness decreased in the order PFO > PO > BNO; i.e., according to the worsening of oil quality parameters (Table 2). However, the POBE performance as an antioxidant additive in PO was, when used at 1000 ppm, even better than that of synthetic antioxidants. On the other hand, the worst performance in BNO may have been the result of a different action mechanism or of its higher oxidation degree, in that bioactive compounds may have reacted with oxidation products thus reducing the POBE's protective action.
In Tables 4 and 5 it is possible to observe the influence of the POBE, in blend with synthetic antioxidants, on the OIP of vegetable oils. In general, the use of POBE in combination with PG (Table 4) or TBHQ (Table 5) increased the vegetable oils' OIP, thus demonstrating the occurrence of a synergism between them, especially in the case of PG that showed the best results. This made it possible to increase the oxidation protection of these oils without exceeding the maximum limit allowed by law for the use of synthetic antioxidants.  In Tables 4 and 5 it is possible to observe the influence of the POBE, in blend with synthetic antioxidants, on the OIP of vegetable oils. In general, the use of POBE in combination with PG (Table 4) or TBHQ (Table 5) increased the vegetable oils' OIP, thus demonstrating the occurrence of a synergism between them, especially in the case of PG that showed the best results. This made it possible to increase  PFO, which displayed the best quality parameters ( Table 2), was the one among the vegetable oils tested that gave the best inhibition results, even quadrupling the OIP using the 100/2000 (ppm/ppm) PG/POBE blend. The OIP of BNO, which in contrast showed the worst quality parameters, using the 50/500 (ppm/ppm) PG/POBE blend, was almost the same as that obtained using the maximum PG concentration allowed by the legislation (100 ppm according to the Brazilian legislation (ANVISA RDC nº 33/2001)) [34]. There is no limit established for natural antioxidants, which is the case of the POBE. This allowed halving the dosage of this synthetic antioxidant, which, like all others, in large quantities can be harmful to human health. However, in the case of PO, a progressive increase in the TBHQ/POBE dosage led to a maximum OIP threshold of about 5.5 h, which was significantly exceeded only by using the maximum POBE dosage in the blend (Table 5).

Conclusions
The aim of this work was to recover the clay used as an adsorbent in the palm oil bleaching process and to use the residual oil extract (POBE) as an antioxidant additive in order to increase the oxidative stability of vegetable oils. For this purpose, the passion fruit, Brazil nut and pracaxi oils were tested as references for future studies on biofuels and lubricating oils. Thermogravimetric and NMR analyses of clay after bleaching indicated a 12% residual oil content, whose physicochemical characterization showed a high antioxidant potential related to its high contents of phenolic compounds and carotenoids.
The antioxidant capacity of POBE, tested at different concentrations in reference vegetable oils, was comparable to those of PG and TBHQ, which are commonly used to improve the stability of biofuels and lubricating oils. In general, starting from a POBE concentration of 200 ppm, an increase in the oxidative induction period was observed for the studied oils, which is expected to lead to an increase in their shelf life. PG showed higher efficiency than TBHQ as synthetic antioxidant, probably due to its larger number of active hydroxyl groups, which could have compensated the polarization effect of the aromatic ring, favoring the release of more protons.
In mixtures with these synthetic antioxidants, especially PG, the residual oil showed a positive synergistic effect on the oxidative stability of pracaxi, Brazil nut and passion fruit oils, whose oxidative induction periods, using 100/1000 (ppm/ppm) PG/POBE in oils, were lengthened to as much as 12.2, 11.1 and 11.1 h, respectively. Such a synergistic effect make way for future replacement of these synthetic antioxidants with natural alternatives such as POBE, to minimize their negative effects on human health and, at the same time, solve the environmental problem of waste oil disposal and bleaching clay exploitation.
These results point to the future substitution of synthetic antioxidants with natural alternatives, minimizing the negative effects of these substances on human health. POBE proved to be efficient in its antioxidant capacity, adding value to residual oils from the oil refinement processes and contributing to the reduction of the environmental problem caused by the exploration of bleaching clay. These results make way for future replacement of these synthetic antioxidants with natural alternatives such as POBE.