3. Results & Discussion

3.1. Biodiesel Purification

3.1.1. Acid value

Even with the increase in the acidity index after the treatment of biodiesel with Magnesol®, all values of the acidity index are lower than that established by the ANP standard, where the document cites that only biodiesel with an acidity index less than 0.5 mg KOH/g sample can be marketed. Figure 3 shows the acid value of the unpurified biodiesels and biodiesels purified with virgin (purified I) and recovered Magnesol® (purified II), where occurs increase acids values due to the affinity of Magnesol®adsorbing basic compounds, such as soap and catalyst residue (KOH) (Faccini et al., 2011).

3.1.2. Glycerol amount

Figure 4 shows the amount of glycerol in the biodiesel before and after the use of Magnesol®, and it is observed that the used of Magnesol®, and the reuse do Magnesol purified (originally used in the purification of biodiesel from virgin soybean oil), was not effective to purification of biodiesel, because the amount of glycerol found in the biodiesel is within the norms (below 5%). But the Magnesol® used in the purification of biodiesel from the frying oil, it can be reused with good efficiency, since the amounts of glycerol found in biodiesel A2.1 (5.03%) and A2.2 (2.7%) are within the values recommended by the legislation, showing a decrease in the amount of glycerol when compared to unpurified biodiesel (A2 - 6.41%). Probably, this decrease in glycerol observed in biodiesel after purification with recovered Magnesol® is due to the characteristics of the remaining compounds in the THF-purified Magnesol, because it is demonstrated in the literature that residual compounds with polar and nonpolar characteristics found in residues of the biodiesel and remaining in the Magnesol have greater affinity with glycerol, aiding in the purification of the biodiesel (Panagiotopoulou and Tsimidou, 2002; Wretensjo and Karlberg, 2002).

3.1.3. Flash point

In three fractions of biodiesels synthesized from virgin oil (A1) and subsequently purified with virgin Magnesol® (A1.1) and recovered Magnesol® (A1.2), flash point analyses were performed and the results obtained were 169°C, 165°C and 167°C, respectively. The results indicate that all fractions of purified biodiesel have values accepted by ANP, demonstrating good reusability after passing through the purification process. The same flash point analyses were performed with three fractions of biodiesel synthesized from frying oil (A2) and purified with virgin Magnesol® (A2.1) and recovered Magnesol® (A2.2), yielding results of 159 °C, 158°C and 160 °C, respectively. These results are in accordance with ANP 07/2008 standards that establish a flash point of 100°C as the minimum value to be used (Lôbo et al., 2009).

3.2. Magnesol®Purification

In Table 1, processes 1 and 3 correspond to the solvents used to purify the Magnesol® used in the purification of biodiesel derived from virgin oil, while processes 2 and 4 correspond to the solvents used to purify the Magnesol® used in the purification of biodiesel derived from frying oil. Results of the elemental analysis corresponding to the related process showed the best results for the Magnesol® used in the purification of biodiesel from virgin oil. Consequently, after process 4, it was decided to optimize the adsorbent purification process only for the Magnesol® used in the purification of biodiesel derived from virgin oil, and later, to apply this to the Magnesol® used in the purification of biodiesel derived from frying oil (Table 1).
The elemental analysis of virgin Magnesol® did not show any percentage of carbon, thus all carbon found after the purification of the adsorbent is an impurity derived from biodiesels. In the Magnesol® recovery process, the parameters optimized were reaction time, solvent type, solvent ratio and temperature. Above is Table 1 with the 19 Magnesol®recoveries carried out in this work. As the goal of this analysis was to determine the amount of carbon, the reactions that were demonstrated to be more efficient were reactions 9, 10, 14, 15 and 16.
Magnesol® purification was performed using several solvents of different polarities in order to verify the influence of the dielectric constant in the purification of biodiesels, because it is known that the by-products of biodiesel synthesis have polar and nonpolar characteristics. Knowing the dielectric constant (µ) of the solvents, it was expected that the solvents with a higher degree of polarity, such as H2O (used to make the solution of NaOH), µ = 80; CH3CN, µ = 37 and ethanol, µ = 30, would be the least effective. The CH2Cl2, although showing µ = 9.1, very close to µ of THF (µ = 7.5), also does not provide satisfactory results, perhaps because it did not show hydrogen interactions with the by-products generated in biodiesel synthesis. Ethyl acetate (µ = 6.02) is observed to have purification rates very close to THF because ethyl acetate has a µ slightly smaller than THF and it has the possibility of having hydrogen interactions with compounds withdrawn from the process of biodiesel purification.
In relation to the effect of the temperature during the reaction, it is observed that an increase in temperature (25 to 50°C) in the purification process of Magnesol® does not contribute significantly to a better result. This can be seen in the reactions using hexane (reactions 5 and 6) and ethyl acetate (reactions 07 and 08) in which a small decrease in impurities (amount of C, 0.54% and 0.06%, respectively) occurred. In addition, when comparing the THF solvent at 25 and 50 °C a slight variation in the amount of carbon (4.31 and 4.44%, respectively) is seen.
Another parameter analyzed in the reaction optimization was the number of extractions. When comparing reactions 04 and 05, it is evident that when three extractions are used (reaction 05) instead of a single extraction (batch), the purification of Magnesol® is more effective, even using 25% less solvent. Subsequently, by comparing the amount of solvent, it was observed that when increasing the amount of THF in the proportions of solvent/adsorbent from 15:1 (reaction 09) to 30:1 (reaction 14), Magnesol® was obtained with a C quantity 0.42% lower, and when testing the purification of Magnesol® by increasing the proportion of solvent/adsorbent to 45:1 and 60:1 (reactions 15 and 16, respectively), a very small decrease in impurity was observed, 0.06 and 0.08%, respectively, showing that above the solvent/adsorbent ratio 30:1, there is no considerable gain in impurities reduction. And finally, when the contact time between the solvent and Magnesol® was decreased, a decrease in the purification efficiency of the solvent was observed, indicating that the solvent and the Magnesol® are required to remain in contact for 30 minutes at each step during the washing process.

3.3. Characterization of Purified and Unpurified Magnesol®

In order to justify the efficiency of the recovery of Magnesol® (purified), SEM-FEG (Scanning Electron Microscopy-Field Emission Gun) analyses were performed, in which it was possible to compare the morphology and particle diameter. In addition, through the FTIR analysis it was possible to compare which chemical groups the Magnesol® can retain in its structure as well as evaluate the efficiency of the recovery process. Also, thermogravimetric analysis was performed, in which the different degradation temperatures of the virgin and recovered Magnesol® can be compared.

3.3.1. SEM-FEG

Figure 5 shows the SEM-FEG of virgin Magnesol® and Magnesol® recovered from the purification of biodiesel derived from virgin and frying soybean oils. From the micrographs obtained by SEM-FEG, the particle diameters of (a andd ) Pure Magnesol®, (b ande ) Magnesol® used to purify biodiesel derived from frying oil and (c and f ) Magnesol® used to purify biodiesel derived from virgin oil were calculated. A predominant microporous morphology was observed in the analyzed materials as it was also highlighted by Facicini et al., 2011.
It is noted that the process of purification and recovery causes breaking of part of the Magnesol® structure, probably due to the agitation and temperature of these processes. This corroborates with previous results regarding the reuse of Magnesol® because the structure remains porous and the breaking of part of the spherical structures increases the contact surface, compensating for the probable decrease in Magnesol® efficiency that can occur with the small amount of contamination that remains after the recovery process.

3.3.2. FTIR

Comparing the first spectrum with the RM-1 and RM-2 spectra, not difference is noted, specifically, the recovered Magnesol® (RM-1 and RM-2 spectra) demonstrates compatibility with the virgin Magnesol®. It may be further noted that the M-1 and M-2 spectra show some characteristic bands having absorptions between 3000 and 2800 cm-1(C-H), compatible with the presence of carbon sp3(carbon with just sigma bonds (σ)). The bands between 1820 and 1630 cm-1 show carbonyl functions of C=O and the region 1599 to 1500 cm-1 represents sigma (σ) and pi (π) bonds, i.e., unsaturated bonds between C=C (Lopes and Fascio, 2004; Silverstein and Bassler, 1962). With this, it can be observed that the recovered Magnesol® spectra (RM-1 and RM-2) do not contain the infrared absorption bands found in the M-1 and M-2 spectra, and thus, resemble the virgin Magnesol® FTIR spectrum (Figure 6).

3.3.3. Thermal gravimetric analysis

From the thermograms in Figure 7, it is observed that the degradation of virgin Magnesol® (VM) occurs in a single step. However, the degradation of the recovered Magnesol®(RM-1 and RM-2) occurs in two steps, because contaminants are present in the purified biodiesel which the THF solvent was not capable of removing. This result was proven by the presence of carbon in the elemental analysis.
As shown in Figure 7, VM, RM-1 and RM-2 experience maximum mass loss with temperature at 83 °C, 76 °C and 74 °C, respectively. This result is expected due to the volatile components and water steam beginning to be released in this first range of temperature starting at 42 °C, 42 °C and 35 °C and finishing at 230 °C, 266 °C and 266 °C, respectively. In addition, a second maximum mass loss temperature was observed for RM-1 and RM-2 (408 °C and 446 °C, respectively); mass loss in this temperature range was expected due to the small amount of carbon found in the products.