PE Soxhlet extraction separated the substances of interest from insoluble inorganic (sand) and organic material (wood and cellulose). The FT-IR spectrum shows that the sludge contains mainly free fatty acids. Figure 1 (top) shows the main absorption signal of C = O stretching at 1697.5 cm− 1 for carboxylic acids, where no ester band of glycerides is observed. Then, methanol-based esterification was performed according to Nata et al. [1] Due to the difficulty in performing a liquid-liquid extraction in a separatory funnel, a continuous liquid-liquid extractor with EA was used. Then, the ester interchange reaction was carried out with continuous heating for 24 h [6], as corroborated by MS spectra of the less polar fraction obtained after CC. The FT-IR spectrum in Fig. 1 (bottom) shows the typical C = O stretching band for esters at 1737.5 cm− 1.
GC-MS analysis of the isolated fraction (column chromatography, 35 % mass yield, Rf = 0.33 SiO2, PE/EA 90:1) was carried out to identify the main compounds in that sample. Analyte MS spectra comparison using a mass spectral library search [7] corroborates the formation of ethyl esters of fatty acids instead of the expected methyl esters. Figure 2 presents the resulting gas chromatogram; the two main compounds in the fraction (91.7 %) were identified as ethyl palmitate (C18H36O2, 53.6 %) and ethyl oleate (C20H38O2, 38.1 %) with a higher match factor correspondence for its (E)-stereoisomer, in agreement with Nata et al. [1]. The match factor is the measured value of the direct match of peak m/z values and relative intensities, while the reverse match factor ignores all peaks that are in the sample spectrum but not in the library spectrum [8]. Clearly, a transesterification reaction occurred during the liquid-liquid extraction with EA and turned the methyl esters into their ethyl derivatives, as the MS spectra data comparison demonstrates. The MS comparison is shown in Fig. 3 (top) for ethyl palmitate and in Fig. 3 (bottom) for ethyl oleate.
Table 1
Match factors from the comparison of the mass spectra for the main analytes obtained in the CG-MS spectrum of the ester mixture and compounds in the NIST library.
Peak Name | tR (min) | Area (%) | Scan No. | Molecular ion mass (g/mol) | Assigned formula | RDBE* | Assigned compound | Match Factor |
1 | 9.128 | 0.87 | 444 | 228.2 | C14H28O2 | 1.0 | Ethyl laurate [Ethyl dodecanoate] | 870 |
Methyl tridecanoate | 584 |
2 | 11.874 | 1.09 | 924 | 256.2 | C16H32O2 | 1.0 | Ethyl myristate [Ethyl tetradecanoate] | 910 |
Methyl pentadecanoate | 632 |
3 | 13.717 | 1.59 | 1246 | 270.3 | C17H34O2 | 1.0 | Methyl palmitate [Methyl hexadecanoate] | 947 |
Ethyl pentadecanoate | 605 |
4 | 14.758 | 53.54 | 1428 | 284.2 | C18H36O2 | 1.0 | Ethyl palmitate [Ethyl hexadecanoate] | 903 |
Methyl heptadecanoate | 625 |
5 | 16.171 | 0.93 | 1675 | 296.3 | C19H36O2 | 2.0 | Methyl oleate [Methyl (9Z)- Octadec-9-enoate] | 926 |
Methyl (9E)- Octadec-9-enoate | 922 |
6 | 17.316 | 38.13 | 1868 | 310.3 | C20H38O2 | 2.0 | Ethyl (9E)-octadec-9-enoate | 917 |
Ethyl oleate [Ethyl (9Z)-octadec-9-enoate] | 892 |
Methyl (10Z)-nonadec-10-enoate | 726 |
7 | 17.562 | 3.84 | 1918 | 312.3 | C20H40O2 | 1.0 | Ethyl stearate [Octadecanoic acid ethyl ester] | 782 |
Methyl nonadecanoate | 671 |
* Ring and double bond equivalent |
NIST MS Search 2.3 was used to compare the mass spectra of methyl and ethyl ester derivatives with the same molecular ion mass (same molecular formula). Considering the absence of glycerides (as ester signals) in the sludge IR spectra (Fig. 1, black spectrum), the POS contained predominantly carboxylic acids similar to compositions described in other reports [9], i.e., palmitic acid and oleic acid as the main components. Then, the POS sample underwent Fischer esterification (heating under reflux of the carboxylic acid mixture in methanol), and the resulting reaction mixture was heated in basic aqueous media with ethyl acetate; under these conditions, the only expected derivatives of our POS were methyl and ethyl ester carboxylates. Based on that assumption, the molecular formula and ring and double bond equivalent (RDBE) assignments of the main signals in the chromatogram were calculated with MS Interpreter version Beta 3.1a considering only CxHyO2 formulas. NIST MS Search 2.3 was used for the mass spectra comparison, and the results for the match factors are summarized in Table 1.
As reported by Dubé et al. [10], for our similar biphase system, ester interchange should occur in the interphase, as shown in Fig. 4. The use of a weak base solution allowed the removal of the FFAs into the aqueous layer [11] as carboxylates, avoiding any interference in the reaction. Then, alkaline transesterification occurred in a similar way to the normal alkaline-catalyzed transesterification of vegetable oils to produce FAME and glycerol [11], considering that a small concentration of methoxide anion formed in the basic media through deprotonation of the residual methanol (used in the previous Fischer esterification) (Fig. 4, orange arrow). With that assumption, our hypothesis is that the methoxide anion initially transesterified the ethyl acetate used as extraction solvent, turning it into methyl acetate and liberating ethoxide anion into the interphase where conditions allowed the ester to interchange from FAME to FAEE. The contact of the warm solvent with the aqueous layer provided the energy needed to drive the reaction to form FAEE products, and the equilibrium was shifted by excess ethyl acetate and the extended duration of the process (24 h).
Although very low amounts of ester derivatives of linoleic acid were expected, they were not observed. Work is in progress to increase the yield of the first separated fraction analyzed in this report. This research covered several aspects of green chemistry, and the main compounds obtained are expected to serve as surfactants in diverse oil/water systems.