Fluorescence Spectroscopy Applied to Monitoring Biodiesel Degradation: Correlation with Acid Value and UV Absorption Analyses

The techniques used to monitor the quality of the biodiesel are intensely discussed in the literature, partly because of the different oil sources and their intrinsic physicochemical characteristics. This study aimed to monitor the thermal degradation of the fatty acid methyl esters of Sesamum indicum L. and Raphanus sativus L. biodiesels (SILB and RSLB, resp.). The results showed that both biodiesels present a high content of unsaturated fatty acids, ∼84% (SILB) and ∼90% (RSLB). The SILB had a high content of polyunsaturated linoleic fatty acid (18  :  2), about 49%, and the oleic monounsaturated (18  :  1), ∼34%. On the other hand, RSLB presented a considerable content of linolenic fatty acid (18  :  3), ∼11%. The biodiesel samples were thermal degraded at 110°C for 48 hours, and acid value, UV absorption, and fluorescence spectroscopy analysis were carried out. The results revealed that both absorption and fluorescence presented a correlation with acid value as a function of degradation time by monitoring absorptions at 232 and 270 nm as well as the emission at 424 nm. Although the obtained correlation is not completely linear, a direct correlation was observed in both cases, revealing that both properties can be potentially used for monitoring the biodiesel degradation.


Introduction
e human interest in using the biomass available on the planet has been increased due to the global population growth, accelerated industrial development, reduced global petroleum reserves, and concerns about environmental impacts [1]. In this context, the search for alternative sources of renewable energy has growing interest. Besides renewable, the alternative energy sources should also be ecologically correct, socially sustainable, and economically viable. Clearly, biofuels compose part of the alternative ecologically friendly energy because they are produced from renewable sources and also more friendly to the environment than fossil-derived fuels [2].
Typically, oleaginous species are used as the main source of raw materials to produce biofuel [3]. Biodiesels can be produced by di erent oleaginous sources depending on the raw material available in each country and/or region. Consequently, it have been reported that biodiesels can be synthesized from soybean, corn, canola, palm, animal fats, recycled greases, and others [4,5]. For instance, in 2016/2017 soybean oil was the main feedstock for biodiesel production worldwide (341 million tons), for a total of 558 million tons of total oilseeds [6]. e desirable characteristics of the raw materials for biodiesel production include (i) adaptability to local growing conditions; (ii) regional availability; (iii) high oil content; (iv) favorable fatty acid composition; (v) compatibility with existing agricultural practices; (vi) agricultural inputs; (vii) by-product markets; (viii) land compatibility; and (ix) rotational adaptability with other cultures [7].
Nevertheless, depending on the region, di erent vegetable oils are available to biodiesel production according to the agricultural, social, and commercial factors as previously mentioned. As consequence, di erent biodiesels are produced with particular chemical compositions, such as secondary compounds (polyphenols, carotenoids, and chlorophylls) and triacylglycerols (TAGs) with di erent lengths and unsaturation contents of the fatty acid carbon chain. e secondary compounds and TAG's compositions are directly associated with the susceptibility of biodiesel degradation: (i) secondary compounds (antioxidants) can promote biodiesel stability, and (ii) unsaturated carbon chain content makes biodiesel more susceptible to the oxidation [8].
e biodiesel commercialization is dependent of several quality parameters [9]. Among them, acid value (AV) is an important parameter to evaluate the biodiesel characteristics. is parameter indicates the formation of organic acids from TAG's oxidation by molecular oxygen [10]. High AV indicates that biodiesel is very degraded, and its commercialization is impaired [11].
In the course of biodiesel degradation process, several small molecules are generated, such as peroxides, aldehydes, ketones, and carboxylic acids, among others [9]. ese molecules are formed by reaction between carbon chains of biodiesel with singlet oxygen, forming oxidized compounds [12]. Recent studies have demonstrated the potential of the spectroscopic technique in determining some important transformations in the molecular structure of the alkyl esters from biodiesel during degradation process [13,14]. Despite the fact that the formed degradation, speci cally carboxylic acids, be satisfactory quanti ed by the titration method (AV parameter), it remains unclear the relation of generating these compounds with biodiesel light absorption/emission.
Our previous works have shown that the acidi cation of vegetable oils re ects on the changes in the absorbances at 232 and 270 nm with direct impact on the emission pro le [15,16]. However, the optical behavior is still not completely understood for biodiesel, especially for unconventional vegetable sources studied here (from Raphanus sativus L. and Sesamum indicum L. seeds). In this sense, the present study aimed to apply uorescence spectroscopy for monitoring biodiesel degradation and determine its correlation with acid value and UV absorbance changes.

Oil Extraction. Raphanus sativus L. (RSL) and
Sesamum indicum L. (SIL) seeds were commercially obtained from local companies. e seeds were dehydrated in laboratory oven under temperature of 60°C for 14 hours. en, the seeds were submitted to hydraulic press extraction obtaining a yield of extracted oil of 32% (w/w) and 45% (w/w) for RSL and SIL oils, respectively. e extracted oils were kept in a dark environment (refrigerator) at −4°C until methyl esters synthesis. Unre ned crude oils were used in all analyses.

Biodiesel Production.
Primary, the acid values for RSL and SIL oils were titrated, obtaining the values of 0.35 and 0.82 mg KOH/g, respectively. e biodiesel was obtained by the transesteri cation of the extracted crude oils, using the potassium methoxide as a catalyst. Initially, the potassium methoxide was obtained by reacting 40 mL of methanol with 1.8 g of KOH under stirring at 60°C for 30 min. After that 90 g of oil was added in the solution and remained under stirring at 60°C during approximately 2 h for the transesteri cation reaction. is reaction was monitored by thin layer chromatography as described in our recent study [17,18]. After the transesteri cation process, the mixture was decanted to separate the biodiesel from the by-product glycerol. e phase containing the esters of interest was washed with distilled water and then dried with anhydrous MgSO 4 , ltered and concentrated in a rotary evaporator at 55°C for approximately 3 h [15].

Sample Degradation.
e degradation was carried out as follows: (i) the biodiesel samples were weighed in amber asks, approximately 3.5 g of sample (in each ask) in which the samples were prepared in triplicate; (ii) all samples were subjected to heating in an air circulation laboratory oven at 110°C for 48 hours; (iii) initially, asks were removed and stored (n � 3) every 2 hours up to the 24 hour period, and nally, asks were sampled in 36 and 48 hours. e samples were stored at approximately −4°C until analyses. e biodiesel oxidation was governed by the biodiesel-atmospheric air interaction without any additional air ow rate. e biodiesel-air contact was de ned simply by the size of the ask aperture, a radius of approximately 0.8 cm with 2.0 cm 2 of area with constant air renovation. All analyses were performed in triplicates considering three independent asks containing degraded biodiesel samples.

Fatty Acid Composition.
e analyses of the fatty acid composition were performed by using a gas chromatography system according to the method Ce 2-66 [19]. e fused silica SP-2560 column (100 m and 0.25 mm) was used in the separation process. e isothermal temperature was programmed at 140°C for 5 min followed by heating at 4°C min −1 up to 240°C, remaining at this level for 30 min. e temperature of the vaporizer was 250°C and detector was 260°C, using helium as carrier gas.

Oxidative Stability Study.
e oxidative stability, expressed by the induction period (IP) in hours, was determined in duplicate by a Rancimat apparatus according to the EN 14112 method. Samples weighing 3.0 g (±0.1) were added into a sealed glass vessel reaction, heated at constant temperature of 110°C, and analyzed under a constant air ow rate (10 L·h −1 ) passing through the samples and then into measuring vessel containing 50 mL ultrapure water [20].

Acid Value Analysis.
e variation of total acid value during the degradation period was evaluated by using the classical titrations. In three Erlenmeyers were added 2.0 g of sample, 25 mL of the solution ethyl ether : ethyl alcohol (2 : 1), and two drops of the phenolphthalein indicator. e samples were then titrated with 0.1 mol·L −1 potassium hydroxide, duly standardized, until the appearance of the pink coloration observed for at least 30 seconds [21].

UV Absorption and Fluorescence Spectroscopy
Analysis. All samples were diluted in a hexane HPLC grade. e concentration was 0.025% (w/v) and 0.030% (w/v) for the biodiesels produced from S. indicum L. (SILB) and R. sativus L. (RSLB), respectively. Molecular absorption and emission analyses were performed at room temperature using a 10 mm thick quartz cell.

Statistical Analyses.
e statistical analyses were performed by two-way ANOVA followed by the Tukey's (P < 0.05) test using GraphPad Prism (GraphPad 164 Software, Version 6.0, San Diego, CA). Table 1 shows the composition of fatty acids present in each oilseed. e results show that both feedstocks have a high content of unsaturated fatty acids, ∼84% for S. indicum L. and ∼90% for R. sativus L. As can be clearly seen, SIL has a higher content of linoleic polyunsaturated fatty acid (18 : 2) and oleic monounsaturated fatty acid (18 : 1), which were approximately 49% and 34%, respectively. On the other hand, RSL presented a considerable content of linolenic fatty acid (18 : 3), ∼11%. It was also observed that RSL presented two monounsaturated fatty acids, which were not found in SIL (eicosenoic acid (20 : 1) and erucic acid (22 : 1)). Moreover, signi cant quantities of oleic (18 : 1), >32%, for both feedstocks were quanti ed. e fatty acid composition in the feedstock is very important because the biodiesel degradation can be estimated by knowing the type and quantity of fatty acids.

Fatty Acid Composition.
is estimation is possible considering the unsaturations present in the carbon chains from fatty acids because these regions are susceptible to atmospheric oxygen attack. e oxygen reaction leads to the breaking of these chains to form unwanted oxidized compounds that change the initial physicochemical properties of the samples [22]. However, similar quantities of unsaturated fatty acids (∼84% for SIL and ∼90% for RSL) may not re ect an identical pro le of degradation due to the presence of secondary compounds (intrinsic antioxidants), depending on its chemical feature and content in the oilseeds [23,24].

Acid Value Determination.
e monitoring of the degradation process was carried out by the quanti cation of the acidic compounds ( Figure 1). In general, a similar pro le for the formation of acidic compounds in both biodiesels was observed. ere is an increase in the formation of these compounds in the rst 24 hours of thermodegradation,   tending to a plateau with a AV of ∼0.35 and ∼0.32 mg KOH/g for SILB and RSLB, respectively, until 48 hours. It is interesting to note that there is a change in the acidity pro le between these two biodiesels in approximately 9 hours, presenting a relative inversion. is means that the RSLB presents lower susceptibility to oxidation and formation of acidic compounds than SILB after 9 h of induced degradation. e di erence in the acid values from SILB to RSLB may be related to the major content of linoleic acid (18 : 2), approximately 49% present in the SIL fatty acid composition. is trend makes the biodiesel derived from SIL more susceptible to degrade than biodiesel from RSL. In this process, the small organic acids can be formed from radical reactions; in which, the oxygen reacts with the unsaturations forming organic hydroperoxides, further the formation of aldehydes and nally oxidized to a stable product such as carboxylic acid [22,25]. In summary, Figure 1 shows that although the degradation pro les are not equal, a similar response of acid value analyses was observed for both SILB and RSLB.

UV Absorption Measurement.
e carbonic chains of the methyl esters undergo oxidation exhibit absorption of light at two characteristic wavelengths, at 232 nm and 270 nm.   ese wavelengths refer to modi cations in the structures of the carbon chains and/or the formation of new molecules (degradation products). At the initial stage of degradation, there is a more prominent increase in absorption at 232 nm (primary oxidation), which occurs due to the formation of conjugated dienes. At the end of the degradation process, there is a more prominent increase in absorption at 270 nm (secondary oxidation), associated with aldehydes, ketones, and carboxylates [26,27].
In Figure 2(a), it is possible to see the formation of the primary compounds in the biodiesels. ere was a higher formation of these compounds in SILB in the rst 10 hours of induced degradation. On the other hand, RSLB presents lower production of these primary compounds in the rst 15 hours, being formed more intensely after 24 hours. It is interesting to observe a similar trend for the response of the measured absorbance at 270 nm (Figure 2(b)), comparing with the absorbance at 232 nm. is means that the primary and secondary compounds are generated proportionally in both biodiesels [28]. Moreover, the higher absorption at 232 and 270 nm for SILB can be attributed to its composition in terms of linoleic acids.
Previous studies have reported that an estimate of the level of the biodiesel oxidation can be obtained by analyzing the absorbance ratio of 232 to 270 nm [26,27,29]. Figures 2(c) and 2(d) show that both biodiesels form large amounts of more oxidized compounds after ∼10 hours during the thermal degradation. is response can be understood by the similarity in the unsaturated fatty acids for both biodiesels (∼84% for SIL and ∼90% for RSL).
In addition, the correlation behavior between the quanti ed acid values and the absorbance responses of the primary and secondary compounds generated in the oxidative process was performed. Figure 3 shows that, in general, considering at both wavelengths, there is a direct, but not linear correlation in the values for both biodiesels. Moreover, an opposite behavior was observed for each biodiesel. e inclination of the correlation curve is greater for SILB than for RSLB, suggesting a higher sensitivity of absorbance measurements comparatively to acid value titration for the SILB. In this sense, the results suggest that the oxidative stage of the SILB can be better studied by monitoring the alterations in the UV absorbances. Di erently, RSLB degradation can be evaluated with similar sensitivity applying AV analysis or UV absorption at 232 or 270 nm.

Fluorescence Pro les of Biodiesels.
Alternatively, the excitation/emission pro le was investigated through the contour maps presented in Figures 4 and 5 for SILB and RSLB, respectively. Figure 4 shows changes in uorescence intensities from 280 to 720 nm with excitation of 220 to 400 nm. Figures 4(a) and 4(d) reveal that the oxidation process induced a strong reduction in the emission intensity at ∼320 nm (excitations at ∼235 and ∼285 nm). Similarly, a decrease in the emission intensity at ∼630 nm, when excited at 235 and 285 nm, was also observed as presented in Figures 4(c) and 4(f).
is last emission is attributed to chlorophylls [27,30], clearly degraded under thermal conditions. Additionally, Figures 4(b) and 4(e) show that an increase in the emission intensity in the 415-430 nm range was determined, with excitation from 345 to 380 nm, revealing the production of uorescent compounds during the degradation process.
is behavior is similar to that observed by other studies and corroborates with our previous studies which demonstrated that conjugated tetraenes are produced during the biodiesel degradation [31,32]. Figure 5 shows changes in uorescence intensities from 310 to 500 nm, with excitation of 220 to 400 nm for RSLB. Figures 5(a) and 5(d) show the reduction in the emission intensities at ∼320 nm. According to the literature, the emission at 322 nm region is related to the natural antioxidant tocopherols [33]. e results also revealed a decrease in the emission intensity from ∼480 to ∼500 nm as presented Journal of Analytical Methods in Chemistry 5 in Figures 5(c) and 5(f). is emission region was recently reported to be associated with the carotenoids as reported by Silva et al. [30]. Similarly to that observed for SILB, it was possible to observe the appearance of emissions at ∼425 nm, directly associated with the degradation compounds. As previously discussed, this emission is associated with the conjugated tetraenes which are formed during the degradation of the methyl esters [32,34]. From the emission/excitation contour maps, some candidate wavelengths were selected to monitor the oxidative process of SILB and RSLB, as presented in Figure 6. e uorescence pro le of SILB shows that the emission intensity at 322 nm decays rapidly to ∼300 a.u. in the rst 10 hours (Figure 6(a)). For RSLB, it was observed that the intensity decrease is less abrupt (Figure 6(b)), but equally signi cant. In addition, it is interesting to note the convergence of uorescence intensity to speci c values. Moreover, these behaviors are very similar to that observed for AV analysis and UV absorbances pro les, with change on the response at approximately 10 hours of degradation time. Similar decrease of emission may be observed at wavelengths close to 630 nm for SILB limiting to the initial 15 hours of the degradation process ( Figure 6(e)). For RSLB, the decrease in emission at 488 nm occurs similarly to that observed for emission at 321 nm with satisfactory response in the initial 24 hours of the oxidative process (Figure 6(f)). On the other hand, the oxidation process may be monitored by increase of the emission at 424 nm, being a very promising parameter to monitor the oxidative process of both biodiesels (Figures 6(c) and 6(d)).

Fluorescence, Absorbance, and Acid Value
Correlations. e emission pro le at 424 nm may be associated with classical technique acid value and more recently with absorptions at 232 and 270 nm [16]. Figure 7(a) shows the correlation of uorescence intensity at 424 nm with acid values for RSLB and SILB, respectively. ese results show that there is a direct correlation in both cases, indicating that the emission at 424 nm is a potential tool for monitoring the oxidative stability for both SILB and RSLB. ese correlations will be better addressed in Section 3.6. Figures 7(b) and 7(c) show the obtained correlations of emission at 424 nm with absorbance at 232 and 270 nm, respectively. e results present a direct but not linear correlation. Furthermore, interesting to note di erent correlation plotted for each biodiesel.
ese results can be associated with speci c functions, such previously demonstrated [16]; however, it is clear that there is a nonlinear correlation.

AV Titration versus Emission at 424 nm.
To evaluate the potential of molecular uorescence spectroscopy for monitoring the degree of oxidation of the biodiesel samples, a close analysis of the correlation between uorescence intensity at 424 nm and acid value was performed as shown in Figure 8. Clearly, uorescence emission at 424 nm can be used to estimate the acid values using linear equations (1) and (2), respectively, which were obtained from the data tting presented in Figure 8: Em 424 � 2.1 + 17.1 * AV.
3.7. Consideration about Oxidative Stability of Biodiesels. e induction periods determined in this work for biodiesels from SILB and RSLB were (0.07 ± 0.01) h and (1.60 ± 0.10) h, respectively. ese results show that RSLB is more stable than SILB. In fact, in Figure 2, it was possible to see that the formation of primary and secondary compounds of degradation is more rapidly formed for SILB than RSLB; a possible explanation may be centered on the higher content of polyunsaturated fatty acids present in the SILB, making SILB more susceptible to degradation. e classical method for evaluating the stability of the biodiesels is the Rancimat method [35]. However, several alternatives methods were proposed to increase the knowledge about a particular sample, adding with information obtained from more classical analyses such as titration of peroxide, iodine, and acid values [36]. Other methods, such as infrared spectroscopy, PetroOXY, ultrasonic-accelerated oxidation, di erential scanning calorimetry (DSC), and thermogravimetric analysis (TGA) [8,37], have been also recently applied to monitor the biodiesel quality. Each of these techniques can provide speci c sample diagnostic which they should be crossed and interpreted to obtain a detailed picture of the biodiesel quality condition. In this context, the present study shows that due to the close correlation between biodiesel emission and acid values, uorescence spectroscopy can be potentially applied to monitoring the biodiesel degradation.

Conclusion
is work showed that the composition of fatty acids was feedstock dependent, with high content of unsaturated fatty acids, principally polyunsaturated in SILB. e high content of unsaturated fatty acids re ects in the transformations undergone by RSLB and SILB carbon chains accused by acid value analysis, UV absorption, and uorescence analysis. e oxidation stages of the biodiesels were monitored by the formation of acidic compounds through the classic determination of the acid value, obtaining a similar change for both biodiesels. It was also found that the absorption pro les at 232 and 270 nm were more sensible to discriminate the biodiesels from di erent feedstocks under degradation process, evidenced by the correlation between the absorbance and acid value analysis. Alternatively, our results showed that molecular uorescence spectroscopy can be used for monitoring the degradation stages of the RSLB and SILB according to the linear correlation between the emission at 424 nm and AV. Finally, it was demonstrated that the acid value for both biodiesels can be predicted by analyzing the emission at 424 nm. We summarize that the present ndings revealed that uorescence spectroscopy can be potentially used to monitor the biodiesel degradation.

Conflicts of Interest
e authors declare that there are no con icts of interest regarding the publication of this paper. Acknowledgments e authors would like to thank Conselho Nacional de Desenvolvimento Cientí co e Tecnológico (CNPq),   Journal of Analytical Methods in Chemistry 9