Acid-Catalyzed Etherification of Glycerol with Tert-Butanol: Reaction Monitoring through a Complete Identification of the Produced Alkyl Ethers

Abstract

Higher tert-Butyl glycerol ethers (tBGEs) are interesting glycerol derivatives that can be produced from tert-butyl alcohol (TBA) and glycerol using an acid catalyst. Glycerol tert-butylation is a complex reaction that leads to the formation of five tBGEs (two monoethers, two diethers, and one triether). In order to gain insight into the reaction progress, the present work reports on the monitoring of glycerol etherification with TBA and p-toluensulfonic acid (PTSA) as homogeneous catalysts. Two analytical techniques were used: gas chromatography (GC), which constitutes the benchmark method, and ¹H nuclear magnetic resonance (¹H NMR), whose use for this purpose has not been reported to date. A method for the quantitative analysis of tBGEs and glycerol based on ¹H NMR is presented that greatly reduced the analysis time and relative error compared with GC-based methods. The combined use of both techniques allowed for a complete quantitative and qualitative description of the glycerol tert-butylation progress. The set of experimental results collected showed the influence of the catalyst concentration and TBA/glycerol ratio on the etherification reaction and evidenced the intrinsic difficulties of this process to achieve high selectivities and yields to the triether.

1. Introduction

Glycerol is currently produced in large amounts as a byproduct of the biodiesel industry. According to ChemAnalyst the global glycerol market was about 1 million tons in 2021, and it is expected to grow at a compound annual rate of 4.5% until 2030. The personal care, cosmetic, and pharmaceutical industry sectors dominate this demand [1]. On the other hand, the production of glycerol associated to biodiesel is much higher. Indeed, according to the International Energy Agency [2], 45,712 million liters of biodiesel were produced in 2021, which allows estimating the biodiesel production at about 40.2 million tons (assuming a mean biodiesel density of 0.88 g/cm³) and that of glycerol in crude (non-refined) form at 4.4 million tons. Therefore, there is great interest in developing new uses capable of absorbing the surplus in order to improve the economic balance of the biodiesel production processes and introducing that sustainable resource in the value chain, thus contributing to the circular economy.

Since the large-scale emergence of biodiesel as an alternative fuel, some 25 years ago, many review papers have appeared reporting on the progress made in the methodss for transforming glycerol into value-added products. Referring to some of the recent studies, Morais Lima et al. [3] described the production of propylene glycol, acrolein, epichlorohydrin, dioxalane, dioxane, and glycerol carbonate through chemical routes and that of 1,3-propanediol, n-butanol, citric acid, ethanol, butanol, propionic acid, mono-, di-, and triacylglycerols, cynamoil esters, glycerol acetate, and benzoic acid by means of biochemical processes, mainly enzymatic. Checa et al. [4] discussed the rational formulation of the catalysts required depending on the chemistry of the transformation route according to reforming (steam and aqueous phase), hydrogenolysis, reduction, selective oxidation, and acetalization reactions. Direct uses of crude glycerol and recent valorization approaches such as the production of alkyl-aromatics and activated carbon were also highlighted. Other important conversion processes were dehydration, pyrolysis, gasification, selective transesterification, etherification, fermentation, oligomerization, and polymerization [5]. Kaur et al. [6] emphasized the environmental advantages of the biological conversion of crude glycerol and included among the valuable products polyglycerols, polyhydroxyalkanoates, solketal, trehalose, and various organic acids (lactic, glyceric, succinic, docosahexaenoic, and eicosapentaenoic). A recently proposed and very promising route of glycerol valorization is its catalytic deoxygenation for (bio)olefin (e.g., propylene) production [7]. The plethora of possible products that can be obtained from glycerol illustrates its frequent designation as a platform chemical.

There is also big interest in the applications of glycerol as a fuel (through combustion) and as a source of fuels (e.g., hydrogen, biogas, syngas, and ethanol) and fuel additives [5,8,9]. Fuel additives are commonly used in order to improve thermal engines performance, reduce their pollutant emissions, and modify specific physicochemical properties of commercial gasoline, diesel, and biodiesel. Oxygenated derivatives of glycerol such as ethers, acetals, and esters (acetates) have been reported as fuel additives [10,11]. The tert-butyl glycerol ethers (tBGEs) resulting from the reaction between isobutylene or tert-butanol and glycerol are precisely of particular interest as concerns the present work. Depending on the number of hydroxyl groups of the glycerol molecule that become alkylated, this reaction, also known as glycerol tert-butylation (see Figure 1), leads to the formation of two monoether isomers (tB1GE-a and tB1GE-b), two diether isomers (tB2GE-a and tB2GE-b), and a triether (tB3GE). Due to the limited solubility of the monoethers in the most common fuels, the products preferred as additives are the diethers and, especially, the triether. The tBGEs have been found to increase the octane number of gasoline and have been claimed as substitutes for methyl and ethyl tert-butyl ethers (MTBE and ETBE, respectively). When used as diesel and biodiesel additives, the main positive effects of tBGEs are the reduction in particulate matter and soot emissions [11]. On the other hand, alkylated glycerol monoethers have interesting surfactant and biological properties and find application as components of cosmetics and personal care and pharmaceutical products [12,13,14].

Glycerol etherification reactions have been thoroughly reviewed by Palanychamy et al. [15]. Glycerol tert-butylation involves three consecutive and reversible steps, leading to the successive formation of tert-butyl glycerol mono-, di-, and triethers (see Figure 1). The reaction is typically carried out in the presence of strong acid catalysts, and when tert-butanol (tert-butyl alcohol, TBA) is used as the alkylating agent, each step is accompanied by the liberation of a water molecule. Much of the early work on this field has been performed reacting glycerol and isobutylene (isobutene, 2-methylpropene, IB); a commercial process was developed based on this synthetic route [16]. IB requires operating the reactor under pressure (around 20 atm) in order to keep the olefin in the liquid state, although it is immiscible with glycerol, thus leading to a heterogeneous reaction system characterized by mass transport limitations [16,17]. These features, and the possibility of IB oligomerization to form diisobutylenes as side reaction, have been considered disadvantages that have encouraged the use of TBA instead of IB in more recent works. Nevertheless, the presence of coproduced water has been found to negatively affect the acid catalysts and introduce thermodynamic limitations that introduce thermodynamic limitations that make more complex reaching high yields of the higher (di-, and specially, tri-) tBGEs with TBA than with IB [18]. As for the catalysts required, homogeneous acids, i.e., those that are soluble in glycerol, such as p-toluensulfonic acid or the heteropoly acid H₃PW₁₂O₄₀, are much more active than the heterogeneous ones and allow obtaining significantly higher yields of di- and tri-tBGEs [16]. The interest in avoiding the use of the homogeneous acids due to corrosion, safety, and environmental issues has fostered the search for solid acid catalysts among which cation exchange resins with highly crosslinked structure, large pore zeolites, sulfonated mesostructured silicas and carbons, and supported tungstophosphoric acid have provided the best results with both IB [19,20,21,22,23,24,25] and TBA [19,26,27,28,29,30,31,32]. These materials require a fine-tuning of their acid and textural properties in order to develop suitable catalytic activity and selectivity toward higher tBGEs; at the same time, they are also very sensitive to water, which solvates the hydrophilic active sites, rendering them poorly active.

In the vast majority of glycerol tert-butylation reports available, it is customary to lump the isomers as monoethers and diethers, and even the diethers and triether are sometimes lumped as higher ethers. In the present work, procedures for the identification and analysis of the different tert-butyl glycerol ethers are presented. The five tBGEs were obtained in our laboratories, isolated, and completely characterized by HRMS-ESI+, ATR-FT-IR, and NMR. A straightforward methodology is presented that allows for fast and reliable monitoring of the reaction between glycerol and tert-butanol (TBA) catalyzed with p-toluensulfonic acid (PTSA) combining ¹H NMR and conventional GC-FID analyses. It is expected this way to contribute to a complete characterization of the reacting system, as well as providing an improved description of the steps involved in the tert-butylation reaction.

Gas chromatography (GC) is the benchmark technique for the quantitative analysis of tBGEs; however, except made of tB1GE-a, the tBGEs are not easily available, which complicates the equipment calibration. Melero et al. [20,33] proposed to extrapolate the response factor obtained for tB1GE-a to the higher ethers. Other authors determined the response factors for all the individual ethers after column chromatography separation and purification from the reaction mixture [23,24,34], which is quite laborious.

Nuclear magnetic resonance (NMR) has become in recent years a high-throughput analytical technique for the characterization of complex mixtures. This is the case, for example, of the monitoring and/or quantitative analysis of reaction mixtures from the digestion of woody biomass [35,36,37,38], lignin depolymerization [39], or transesterification reaction for biodiesel production [40,41]. However, the purification of tertiary mono and di tert-butyl glycerol ethers is elusive, and to the best of our knowledge, their characterization has not been reported to date. As for the secondary mono and di-tert-butyl glycerol ethers and tri-tert-butyl glycerol ether, their NMR spectra have been reported [42,43,44]. Nevertheless, descriptions are, in some cases, imprecise when providing the chemical shifts for the ¹H NMR [44] or simplistic when explaining the spin systems of etherification products [42]. Indeed, the chemical shifts reported by Jamróz et al. [44], which describe the spin system, and those reported by González et al. [42] corresponding either to the glycerol skeleton hydrogen atoms or to the methyl groups in the tert-butyl moieties did not match at all between them.

2. Results and Discussion

2.1. Characterization of Etherification Products

The five tert-butyl ethers of glycerol were synthetized as indicated in the Materials and Methods Section 3.3. After their isolation and purification, the corresponding chemical structure and expected formula were confirmed by NMR and ESI⁺ (see Figure A1). Glycerol and di- and tri-tert-butyl ethers presented the expected [M + Na]⁺ as the major peak (m/z 227.1613 for tB2GE-a, m/z 227.1626 for tB2GE-b, and m/z 283.2267 for tB3GE). This peak was accompanied by 2M + Na⁺ in the case of tB2GE-a. More reactive monoethers presented the peak corresponding to [M + Na]⁺ (m/z 171.0996). Condensation of the primary hydroxyl groups was observed under the ESI⁺ analysis conditions for tB1GE. Thus, in the case of tB1GE-a, m/z 301.1988 was detected after the condensation of two molecules under the analysis conditions, producing [2M-H₂O + Na]⁺. tB1GE-b presented an additional primary hydroxyl group, and in addition to [M + Na]⁺ (m/z 171.0996), the major peak appeared at m/z 413.2647, which corresponds to [3M-3H₂O + Na⁺] as a result of the higher reactivity of these primary hydroxyl groups, due to their lower steric hindrance, to become a crown-ether like structure under the analysis conditions.

ATR-FT-IR spectra of the isolated compounds showed the gradual disappearance of the hydroxyl O–H stretching band at ca. 3400 cm⁻¹ as the degree of etherification increased accompanied by the intensification of the aliphatic bands between 2974 and 2834 cm⁻¹ that correspond to the C–H stretching mode (see Figure A2). As expected, no major differences were observed between tB2GE-a and tB2GE-b with this technique.

As for the NMR spectra,

Table 1. Chemical shifts and coupling constant for the hydrogen atoms on the glycerol skeleton of the tBGEs as identified in Figure 2.

Ether	H1a (J1a,2)	H1b (J1b,2)	(J1a,1b)	H2	H3a (J3a,2)	H3b (J3b,2)	(J3a,3b)	C1–OC [CH3]3	C2–OC [CH3]3	C3–OC [CH3]3
tB1GE-a	3.43 (5.88)	3.49 (3.92)	(9.10)	3.78	3.65 (4.95)	3.70 (3.91)	(11.4)	1.20	-	-
tB1GE-b	3.65	-	-	3.71	3.65 a	-	-	1.24	-	-
tB2GE-a	3.37 (5.92)	3.42 (5.06)	(8.97)	3.78	3.37 a (5.92) a	3.42 b (5.06) b	(8.98)	1.19	-	1.19
tB2GE-b	3.34 (8.58)	3.41 (4.30)	(8.62)	3.65	3.61	3.61	-	1.21	1.19	-
tB3GE	3.27 (5.32)	3.37 (5.92)	(9.22)	3.60	3.27 a (5.32) a	3.37 b (5.92) b	(9.22) a,b	1.17	1.20	1.17

^a Symmetry H-1a and H-3a. ^b Symmetry H-1b and H-3b.

gathers the chemical shifts and coupling constant for the hydrogen atoms on the glycerol skeleton that are identified in Figure 2. Given the symmetry of tB1GE-b, tB2GE-a, and tB3GE, the ¹H NMR signals were easier to assign (see Figure A5, Figure A7 and Figure A11). In contrast, the secondary C₂ carbon of tB1GE-a and tB2GE-b was asymmetric, and therefore, C₁ was diasterotopic, which complicated the interpretation of their ¹H NMR spectra (see Figure A3 and Figure A9). The hydrogen atoms on the glycerol skeleton of tB1GE-a appeared as a set of three ¹H NMR signals in the range from 3.42 ppm to 3.81 ppm (see Figure 3a), whereas both hydroxyl hydrogen atoms appeared as broad shoulders at 2.40 ppm. As expected, the tert-butyl moiety appeared as a singlet at 1.19 ppm. The assignation of H and C signals was performed using ¹³C APT and HMBC correlation. The ¹H signal at 1.19 ppm from –C–CH₃ presented long-range correlation with the quaternary C atom at 73.66 ppm (see Figure 3b). These signals also presented long-range correlation with the signal of H₁ centered at 3.43 ppm, which also correlated with C₃ at 63.92 ppm in the HMBC spectrum (see Figure 3b). C₂ is a chiral center; therefore, both H₁ and H₃ are diasterotopic. Hence, they appeared as a set of two signals each (H_1a and H_1b; H_3a and H_3b) with large geminal coupling constants, JH_1a-H_1b = 9.1 Hz and JH_3a-H_3b = 11.4 Hz. The coupling constants with H₂ were in the 3.9–5.8 Hz range. Because of the coupling with the non-equivalent H₁ and H₃ atoms, the ¹H NMR signals for H₂ were shown as multiplet centered at 3.78 ppm. The ¹H NMR spectra of tB1GE-a could be satisfactorily simulated using WINDNMR [45].

Analysis of the spectra for symmetric tB2GE-a and tB3GE was much simpler. In the case of tB2GE-a, the ¹H NMR spectrum showed the signal of the hydrogen atom on the tert-butyl group as a singlet at 1.19 ppm, whereas the signal corresponding to the hydroxyl group appeared as a doublet at 2.54 ppm (J = 4.48 Hz). The signal for H₂ appeared as a hexuplet centered at 3.78 ppm showing an apparent coupling constant of 5.27 Hz. Because of the symmetry of the molecule, H₁ and H₃ were equivalent and turned out a unique signal at 3.39 ppm. However, as in the case of tB1GE-a, H_a and H_b presented slightly different chemical shifts because of their diasterotopic character. Indeed, nuclei H_1a,3a and H_1b,3b showed slightly different chemical shifts of 3.37 ppm and 3.42 ppm, respectively, with a large geminal coupling constant of J_ab = 8.97 Hz being J_a2 = 5.92 Hz and J_b2 = 5.06 Hz. Therefore, the signal for H₂ centered at 3.78 ppm actually corresponded to a triple triplet with J_a2 = 5.92 Hz and J_b2 = 5.06 Hz but could not be accurately resolved. The ¹³C APT spectrum (Figure 3c) allowed the easy assignation of carbon atom signals, as indicated in

Table 2. Chemical shifts for the carbon atoms of the tBGEs.

Ether	C1	C1–O–C–(CH3)3 a	C1–O–C–(CH3)3 a	C2	C2–O–C–(CH3)3 a	C2–O–C–(CH3)3 a	C3
tB1GE-a	63.92	27.59	73.69	70.80	-	-	64.69
tB1GE-b	63.83	-	-	71.12	28.68	74.75	63.83
tB2GE-a	63.09	27.70	73.17	70.38	-	-	63.09
tB2GE-b	64.19	28.31	74.24	69.82	27.37	73.32	65.49
tB3GE	63.53	27.72	72.75	71.34	28.57	73.87	63.53

^a Chemical shifts correspond to the carbon atom written in italics and underlined.

.

The ¹H NMR spectrum of tB3GE presented a similar pattern to that of tB2GE-a. H_1a and H_3a appeared at 3.27 ppm, showing large coupling constants with H_1b and H_3b, J_ab = 9.22 Hz, and J_a2 = 5.32 Hz, whereas H_1b and H_3b appeared at 3.37 ppm with J_b2 = 5.92 Hz. The similarity for the J₂ coupling constants suggested that H₂ had the appearance of a well-defined quintuplet whose apparent coupling constant (J = 5.57 Hz) averaged J_a2 and J_b2. The ¹³C APT spectra for tB2GE-a and tB3GE were far simpler and allowed easier identification of the corresponding signals (see Appendix B). In the case of tB2GE-a, the signal at 27.52 ppm was attributed to the primary methyl groups on the tert-butyl moieties, the signal at 72.99 ppm to the quaternary carbon on the tert-butyl moieties, and that at 62.9 ppm to the secondary C₁ and C₃. Similarly, for tB3GE, the HMBC spectrum showed long-range correlations between the hydrogen signal at 1.17 ppm and the quaternary C atom at 72.56 ppm, as well as the hydrogen signal at 1.20 ppm and the quaternary C atom at 73.68 ppm, allowing the assignation of these quaternary C atoms.

Elusive tB1GE-b and tB2GE-b have been recently identified by GC-MS [43], although their NMR characterization has not been reported. tB1GE-b, as in the case of tB2GE-a and tB3GE, shows a symmetry plane, so that a simple spectrum could be expected. Nevertheless, chemical shifts of H₂, H_a, and H_b were so close that signals corresponding to H_a and H_b were broad and appeared in the 3.60–3.68 ppm range, and the coupling constants could not be accurately determined. In the case of H₂, it appeared as an apparent quintuplet (J = 4.67 Hz) centered at 3.71 ppm. Concerning the tB1GE-b ¹³C spectrum, it was recorded using an APT sequence that allowed the fast assignation of the signal at 28.68 ppm to the primary methyl carbon and the ones at 63.83 ppm and 71.12 ppm to the C₁ and C₃ secondary carbons and the C₂ tertiary carbon, respectively (see Appendix B). The long-range correlation of the signal at 1.24 ppm that corresponds to the CH₃ groups allowed the identification of a small signal at 74.75 ppm assigned to the quaternary carbon atom on the tert-butyl moiety.

Finally, the ¹H NMR spectrum for tB2GE-b showed four groups of signals in the glycerol skeleton and two singlets corresponding to the tert-butyl groups at 1.21 ppm and 1.19 ppm. The hydroxyl group appeared as a double doublet at 2.51 ppm (J = 3.78 Hz, J = 7.94 Hz) due to coupling with the diasterotopic H₃. This hydroxyl signal showed strong HSQC-TOCSY correlation with the carbon atom at 65.65 ppm that corresponded then to C₃ (Figure 4a). Once C₃ was assigned, the heteronuclear ¹H–¹³C experiment combined with ¹³C APT allowed easy assignation of C₁, C₂, H₁, and H₂ (Figure 4b). The hydrogen atom on the hydroxyl group presented a clear NOE effect with the hydrogen atoms on the tert-butyl group at 1.19 ppm (Figure 4c) that were hence assigned to the tert-butyl group on C₂. The long-range correlation in the HMBC spectrum (Figure 4d) allowed the identification of the quaternary carbon atoms.

The ¹H signals for H₂ and H₃ signals overlapped, making the resolution of the system difficult. As for H₁, two ¹H NMR signals were observed, the first centered at 3.41 ppm and the second at 3.34 ppm (Figure 5). The geminal coupling constant for H_1a and H_1b was 8.60 Hz, and they both coupled with H₂ with J = 4.30 Hz and J = 8.58 Hz, respectively, causing the triplet aspect of the signal at 3.34 ppm that indeed corresponded to a double doublet. The chemical shifts for H₂ and H_3a and H_3b were determined using the HSQC-TOCSY cross-signals with C₂ and C₃, respectively. The coupling constants for H₂ and H₃ needed to be determined using the spectrum simulation module in Topspin 3.6.2 and are gathered in Table 1.

2.2. Glycerol Tert-Butylation Monitoring through GC Analyses

Gas chromatography (GC) is the benchmark analytical technique for monitoring etherification reactions between glycerol and tert-butyl alcohol (TBA), although it presents several drawbacks. Polyols like glycerol have low vapor pressures that makes necessary the use of relatively harsh analysis conditions together with long analysis times that do not ensure the precise quantification of glycerol and, consequently, accurate mass balances. In addition, the complete derivatization by silylation of the reaction mixture is difficult and laborious and requires high quantities of specific reagents. Nevertheless, as shown in Figure 6, it allows the appropriate separation of the five tBGEs, glycerol, and the internal standard in ca. 20 min, although unreacted TBA cannot be quantified as it elutes with the solvent used to dilute the sample taken from the reaction mixture.

Figure 7 shows the tBGE content of the reaction mixture expressed as molar fractions as a function of the glycerol conversion achieved after 24 h of tert-butylation reaction at temperatures between 70 and 110 °C, 8 wt.% PTSA, referred to the initial glycerol content and TBA/glycerol molar ratios within the 4:1–16:1 range. It can be observed that tB1GE-a and tB2GE-a were the main tert-butyl glycerol ethers produced. In addition, the tB1GE-a content was much higher than that of tB2GE-a, which only reached significant concentrations once the monoether was sufficiently abundant, in accordance to a reaction scheme in series. The formation of 1- and 1,3-ethers from the condensation of primary hydroxyl groups of glycerol to produce tB1GE-a and tB2GE-a was more probable against the formation of 2- and 1,2-ethers, tB1GE-b, and tB2GE-b [26]. Indeed, glycerol was a triol having double the number of primary than secondary hydroxyl groups. In addition, primary hydroxyls were preferred for tert-butylation due to steric effects because the tert-butyl group was a voluminous moiety. This explained in part the very low concentrations of the triether achieved, which was present in detectable amounts when the glycerol conversions reached values above ca. 0.75. Another reason was the thermodynamic limitations that appeared when TBA was used as the alkylating agent [18]. The tB1GE-b and tB2GE-b contents were also very low. The evolution of the molar fractions suggested that tB1GE-b disappeared to form tB2GE-b and that this diether reacted to form the triether tB3GE, as indicated by arrows in Figure 7.

2.3. Glycerol Tert-Butylation Monitoring through ¹H NMR Analyses

Given that the ¹H NMR signal was directly proportional to the amount of hydrogen atoms present in the sample, in principle, no calibration was required for the quantification of samples whose analysis required only 90 s. Initially, the ¹H NMR spectra of the isolated glycerol and the tB1GE-a, tB2GE-a, and tB3GE. tBGEs were superimposed (Figure 8a) in an attempt to find out a relation between the results of the integration of the different spectra regions and the molar fraction of each compound. However, the spectra of the reaction mixtures were slightly different from those corresponding to the isolated product superposition (Figure 8b). This was due to the differences in the solvent dielectric properties due to the presence of high amounts of TBA and glycerol in the reaction mixture. Hence, the spectra of the samples from the etherification reactions were much easier to interpret than those corresponding to the component superposition, which allowed the simplification of the quantitative analysis defining four integration regions (denoted as R_j, j = 1, 2, 3, 4) and assuming negligible the contributions of tB1GE-b and tB2GE-b. As for the rest of the compounds, their contributions to the several integration regions are gathered in

Table 3. Contribution of the compounds indicated to the integration of the NMR spectra regions.

Region, Rj	δ (ppm)	Glycerol	tB1GE-a	tB2GE-a	tB3GE
1	3.850–3.604	5	3	1	-
2	3.524–3.429	-	2	-	-
3	3.429–3.333	-	-	4	2
4	3.333–3.225	-	-	-	2

. An obvious drawback of this procedure was that no distinction could be made between both monoethers and diethers.

Accordingly, the molar fractions of glycerol and the tBGEs were calculated from the values ($n_{\mathrm{i}}$) given by Equations (1)–(6), which were proportional to the number of moles of each compound present in the sample. In these equations, ${IA}_{\mathrm{R}\mathrm{j}}$ (j = 1,2,3,4) are integration values corresponding to the region j according to Table 3. The conversion of glycerol ($X_{\mathrm{G}\mathrm{l}\mathrm{y}\mathrm{c}}$) and the tBGE selectivities ($S_{\mathrm{i}}$) can be calculated according to Equations (5) and (6), respectively, considering that no products other than tBGEs and unreacted glycerol were present in the reaction mixture. Due to the abovementioned limitations of NMR analyses, in what follows, selectivities are reported for tB1GE and tB2GE that lump both monoethers and both diethers, respectively.

$$n_{t\mathrm{B}1\mathrm{G}\mathrm{E}} = \frac{{IA}_{\mathrm{R}2}}{2}$$

(1)

$$n_{t\mathrm{B}2\mathrm{G}\mathrm{E}} = \frac{({IA}_{\mathrm{R}3}-{ IA}_{\mathrm{R}2})}{4}$$

(2)

$$n_{t\mathrm{B}3\mathrm{G}\mathrm{E}}={IA}_{\mathrm{R}4}$$

(3)

$$n_{\mathrm{G}\mathrm{l}\mathrm{y}\mathrm{c}}=\frac{{IA}_{\mathrm{R}1}-3·n_{t\mathrm{B}1\mathrm{G}\mathrm{E}}-n_{t\mathrm{B}2\mathrm{G}\mathrm{E}}}{5}=\frac{(4·{IA}_{\mathrm{R}1}-6·{IA}_{\mathrm{R}2}-{IA}_{\mathrm{R}3}-{IA}_{\mathrm{R}4})}{20}$$

(4)

$$X_{\mathrm{G}\mathrm{l}\mathrm{y}\mathrm{c}}=\frac{n_{t{\mathrm{B}}_1\mathrm{G}\mathrm{E}}+{ n}_{t{\mathrm{B}}_2\mathrm{G}\mathrm{E}}+n_{t{\mathrm{B}}_3\mathrm{G}\mathrm{E}}}{n_{\mathrm{G}\mathrm{l}\mathrm{y}\mathrm{c}}+n_{t{\mathrm{B}}_1\mathrm{G}\mathrm{E}}+{ n}_{t{\mathrm{B}}_2\mathrm{G}\mathrm{E} }+n_{t{\mathrm{B}}_3\mathrm{G}\mathrm{E}}}$$

(5)

$$S_{\mathrm{i}}=\frac{n_{\mathrm{i}}}{n_{t{\mathrm{B}}_1\mathrm{G}\mathrm{E}}+n_{t{\mathrm{B}}_2\mathrm{G}\mathrm{E}}+n_{t{\mathrm{B}}_3\mathrm{G}\mathrm{E}}} i=n_{t\mathrm{B}1\mathrm{G}\mathrm{E}},n_{t\mathrm{B}2\mathrm{G}\mathrm{E}},n_{t\mathrm{B}3\mathrm{G}\mathrm{E}}$$

(6)

Figure 9 shows the relation between the selectivities to tert-butyl mono-, di-, and triethers of glycerol obtained, calculated from the results of the analyses performed by GC and ¹H NMR of the reaction samples. In general, a good agreement is observed; however, some samples led to larger discrepancies. The quantification of glycerol was identified as the main source of error, which reached ca. 5% and 7% for the NMR and GC analyses, respectively. In the case of the tBGEs, the errors were reduced to ca. 3% with both techniques. Higher errors could be associated to homogenization difficulties, particularly in samples with very low or very high glycerol conversions. In the first case, the high polarity and viscosity of glycerol complicated the sample manipulation. In the second case, the large difference in polarity between the reaction products, especially the di- and triethers, and that of the reactants led to the formation of micro-emulsions through phase segregation.

Monitoring of the glycerol tert-butylation reactions through ¹H NMR has allowed illustrating the effects of some of the reaction conditions. In this regard, Figure 10 shows the influence on the glycerol conversion of the catalyst (PTSA) concentration after 24 h of reactions conducted at 70 °C and TBA/glycerol molar ratios of 4:1, 8:1, and 16:1. It can be seen that as expected, at a given TBA/glycerol ratio, the glycerol conversion increases at increasing PTSA concentration. For example, at the TBA/glycerol molar ratio of 4:1, the conversion increased from ca. 70% to 95% when the catalyst concentration passed from 8 wt.% to 32 wt.%. However, at a given catalyst content, the glycerol conversion decreased as the TBA/glycerol molar ratio increased. This was explained by the fact that the catalyst concentration was referred to the initial glycerol content of the reaction mixture. Therefore, the catalyst concentration over the total reaction volume decreased as the TBA/glycerol ratio increased due to the dilution caused by increasing amounts of TBA. For instance, when the catalyst concentration was fixed at 32 wt.% referred to the glycerol amount, the overall catalyst concentration decreased from 6.4 wt.% to 1.9 wt.% and finally 1.0% when the TBA/glycerol ratio increased from 4:1 to 8:1 and 16:1, respectively.

Figure 11 shows the tBGE selectivities for the conversion points included in Figure 7. It can be seen that the selectivities were dictated by their own reaction progress, that is, the glycerol conversion. The catalyst content and TBA/glycerol ratio affected the conversion that could be achieved in a given reaction time, in this case, 24 h, but did not seem to influence the selectivity. In other words, the highest glycerol conversions attained corresponded to the reactions performed at the lowest TBA/glycerol ratio (4:1) and the highest PTSA concentration (32 wt.% referred to the initial glycerol amount) considered. High glycerol conversions were necessary to obtain the highest possible di- and triether selectivities. The first ones reached values of ca. 35% at their highest, whereas in the conditions of the present study, maximum tB3GE selectivities of ca. 8% were obtained. As concerns the monoethers, maximum conversions were obtained at the lowest glycerol conversion. In accordance with the in-series scheme that followed the tert-butylation reaction, the first products were proportionally more abundant at short reaction times (in batch processes), when they had little opportunity of being converted into higher ethers. As for the temperature, an effect similar to the rest of reaction variables was found, having a positive influence on the glycerol conversion but not affecting the tBGE selectivities.

2.4. Etherification of the Tert-Butyl Glycerol Monoether

With the purpose of increasing the yield of the higher tBGEs, the tert-butylation reaction was carried out starting from tB1GE instead of glycerol as indicated in Section 3.2. Figure 12 shows the evolution with reaction time of the tBGE molar fractions monitored through ¹H NMR (Figure 12a) and GC (Figure 12b). It was clear that the monoether converted into the diether without having a significant impact on the triether concentration. After the fifth day of reaction, the monoether conversion reached 38%; at that time, a new charge of TBA and catalyst was performed, aiming at further converting the monoether. The conversion increased up to 56% after two additional days of reaction. The GC analyses allowed distinguishing between both diethers, and the obtained results (Figure 12b) suggested that the triether was mainly formed from tB2GE-b. This seemed reasonable in view of the much stronger steric hindrance that would entail its formation from tB2GE-a. However, tB2GE-b was much less abundant than tB2GE-a, that is, the diether with both glycerol primary hydroxyl groups etherified. This showed that there were intrinsic difficulties in obtaining high selectivities to the tert-butyl glycerol triether through the homogeneously acid-catalyzed tert-butylation of glycerol with TBA.

3. Materials and Methods

3.1. Materials and Analytical Techniques

tert-Butanol (TBA), anhydrous glycerol (99.5%), and 1,3,5-trimethoxybenzene (as the internal standard) were purchased from Acros Organics (Fairlawn, NJ, USA) p-Toluenesulfonic acid (PTSA) used as the homogeneous catalyst was purchased from Panreac S.L. (Darmstadt, Germany) CDCl₃ was purchased from Carlo-Erba (Val de Reuill, France) and used as received.

The attenuated total reflectance (ATR) infrared (IR) spectra were recorded on an Avatar 360 FT-IR spectrometer (ThermoFisher Scientific, Walthman, MA, USA). The gas chromatography (GC) analyses were performed on a Shimadzu gas chromatograph equipped (Kyoto, Tapan) with a flame ionization detector (FID) and a DB-23 (30 m, 0.32 mm ID, 0.25 µm) column. During the analyses, the oven temperature was kept for 10 min at 90 °C, then it was raised from 90 to 150 °C at a rate of 25 °C/min, and finally, it was maintained for 8 min at 150 °C. The samples for the GC analysis were prepared from 0.040 g of the reaction mixture that were diluted with 2.5 mL of a 2 g/L solution of 1,3,5-trimetoxybenzene in acetonitrile.

The nuclear magnetic resonance (NMR) analyses were performed on a Bruker Ascend 400 spectrometer (Rheinstetten, Germany) operated at 400 MHz and equipped with a PA BBO 5 mm probe. All ¹H and ¹³C chemical shifts were reported using the δ scale and were referenced to the residual signal of CHCl₃ at 7.26 ppm and that of CDCl₃ at 77.16. The pulse programs were the previously installed zg30 for ¹H with 16 scans. CDCl₃ was the solvent of choice after discarding DMSO-d₆ and CD₃OD due to the overlapping of residual signals from CD₂HOD and H₂O, respectively, from those of the reaction products.

3.2. Tert-Butylation Reactions

Glycerol etherification reactions were performed in a 100 mL stainless steel stirred autoclave at 30 bar, TBA/glycerol molar ratios ranging between 4:1 and 16:1, catalyst concentrations of 8–32 wt.% PTSA referred to the glycerol loaded into the reactor, and temperatures within 70–90 °C. The samples were withdrawn from the reactors during the course of the reaction by means of the appropriate recirculating valves to maintain the pressure and agitation conditions.

An experiment was carried out using 1.5 g of the monoethers (tB1GEs), 10 g TBA and 0.24 g glycerol (TBA/glycerol molar ratio of 13.5:1), 70 °C, and 16 wt.% of PTSA catalyst referred to the tB1GEs. After five days of reactions, a new charge into the reactor of 10 g TBA and 0.24 g glycerol was carried out.

3.3. Isolation of the Tert-Butyl Ethers

In order to obtain the tBGEs for the NMR identification, a glycerol etherification reaction was conducted on the stainless steel autoclave at 90 °C and 30 bar with a TBA/glycerol molar ratio of 4:1 and 8 wt.% PTSA referred to the glycerol loaded into the reactor. After 24 h of reaction, the resulting mixture was concentrated in a rotary evaporator to remove the unreacted alcohol. Afterward, ca. 17 g of the tBGE/glycerol mixture were charged into a chromatography column using M60 silica as stationary phase. tB3GE, tB2GE-a, and tB2GE-b (see Figure 1) were separated using hexane/ethyl acetate (9:1) as the mobile phase, as reported by González et al. [42]. On the other hand, the tB1GE-a and tB1GE-b monoethers were eluted using a hexane/ethyl acetate (1:9) mixture.

4. Conclusions

Glycerol tert-butylation is a complex reaction leading to the formation of five glycerol tert-butyl ethers (tBGEs). All of them have practical interest: the monoethers as surfactants and components of cosmetics and pharmaceuticals and the di- and triethers as fuel additives. In order to suitably monitor the progress of the reaction between glycerol and tert-butanol (TBA), a method based on ¹H NMR analyses was developed in the present work that allowed for the quantification of unreacted glycerol and the tBGEs in only 90 s without the need for equipment calibration. These features are clear advantages compared with conventional GC analyses when fast, almost real-time, monitoring of the reaction is required. In contrast, the method was not able to distinguish between both monoether and both diether isomers, which were lumped into two groups of reaction products. For that reason, it was necessary to combine ¹H NMR and GC analyses to obtain a complete characterization of the reaction mixture.

The set of results available for the development of the analytical methods provided information of interest as concerns the formation of higher ethers. Glycerol tert-butylation is a consecutive reaction in which primarily formed monoethers lead to diethers that are finally converted into the triether. According to our results, the triether seemed to be formed from tB2GE-b instead of tB2GE-a due to easier access of the third tert-butyl moiety to a primary carbon atom than to a secondary one. However, the fact that tB2GE-a was much easier to form than tB2GE-b due to the higher reactivity of primary hydroxyls compared with the secondary ones, the double number of primary as compared with secondary hydroxyls present in glycerol, and steric effects explained the difficulties in forming the tert-butyl glycerol triether through this synthetic route.