Reactive Extraction of Alcohols from Apolar Hydrocarbons with

7572

Ind. Eng. Chem. Res. 2004, 43, 7572-7580

Reactive Extraction of Alcohols from Apolar Hydrocarbons with Aqueous Solutions Boris Kuzmanovic´ ,† Norbert J. M. Kuipers,† Andre´ B. de Haan,*,† and Gerard Kwant‡ Separation Technology Group, Faculty of Science and Technology, University of Twente, P.O. Box 217, 7500 AE Enschede, The Netherlands, and DSM Research, P.O. Box 18, 6160 MD Geleen, The Netherlands

Use of aqueous solutions as reactive extraction solvents for the recovery of monohydroxyl alcohols present in few-percent concentrations in apolar hydrocarbons is evaluated. Two approaches are considered. The first, in which an aqueous solution containing a reactive extractant is applied for the recovery of the alcohol, shows only limited potential to be used. Among several potential extractants, only hydroxypropyl cyclodextrins and silver nitrate were able to achieve a slight improvement of the distribution ratio of benzyl alcohol relative to pure water. In addition to providing moderate values of the distribution ratios of benzyl alcohol, these extractants are also found to increase the solubility of the apolar organic solvent in the aqueous phase and thereby cause a decrease in selectivity toward the alcohol. The second considered approach, in which the alcohol is chemically modified prior to the extraction into an easily extractable form, in this case a monoester, has much more potential. A benign solvent, such as an aqueous solution of sodium hydrogen carbonate, can provide a distribution ratio of benzyl alcohol of up to 200, leaving the solubility of the organic solvent in the aqueous solutions unchanged relative to that in pure water. The modification of aromatic, cycloaliphatic, and linear aliphatic alcohols can be done efficiently in an apolar organic solvent without need for a catalyst. It is found that the back-recovery of alcohol extracted in this way can be performed by back-extraction after the monoester has been hydrolyzed. The rate of hydrolysis can be controlled by temperature. Hence, an efficient and selective recovery of alcohols can be achieved by reactive extraction using aqueous solutions of a hydrogen carbonate salt by modifying the alcohol into a monoester prior to the extraction. 1. Introduction Aromatic and medium- and high-molecular-weight aliphatic oxygenates are often present in few-percent concentrations in apolar organic solvents from which they need to be recovered. Some examples of industrial importance are the oxidation products of toluene, cyclohexane, and cyclododecane, as well as those obtained from the Fischer-Tropsch process. In oxidation processes, the recovery of acids, aldehydes, ketones, and alcohols is based on distillation,1 even though the solvent is present in concentrations of 75-95% and has a lower boiling point than the oxygenates. This results in the necessity of spending large amounts of energy on solvent evaporation. In the case of the FischerTropsch process, product recovery from the organic condensate is done by liquid-liquid extraction.1 However, high solvent-to-feed ratios (up to 6:1) are applied, and it is necessary to separate significant amounts of solvent from the raffinate. Given these drawbacks, other options, more energy-efficient and less demanding, are often considered. In a previous paper,1 we discussed the advantages of liquid-liquid extraction with water for the recovery of oxygenates from an apolar organic solvent. To overcome the poor solubility of these oxygenates in water, we * To whom correspondence should be addressed. Tel.: +3153-4895410. Fax: +31-53-4894821. E-mail: a.b.dehaan@ utwente.nl. † University of Twente. ‡ DSM Research.

considered the introduction of environmentally and toxicologically benign reversibly reactive salts in water, i.e., the use of aqueous solutions as reactive extraction solvents. In this way, the equilibrium distribution ratios of various carbonyl compounds (aldehydes and ketones) could easily be increased by several orders of magnitude up to several hundreds. In this paper, the applicability of reactive extraction for the recovery of slightly and moderately water-soluble monohydroxyl alcohols from apolar organic solvents into an aqueous solution is evaluated. Although water itself can provide fair values of the equilibrium distribution ratio for many of the alcohols considered, a high solventto-feed ratio and/or a large number of stages would still be required to achieve a satisfactory recovery of the alcohol using pure water as the extraction solvent. For example, because of the low distribution ratio of benzyl alcohol between water and toluene of about 0.3 at 25 °C, more than 35 stages with a solvent-to-feed ratio of 3.5 would be necessary to achieve 99% recovery. The distribution of the alcohol between the aqueous and apolar organic phases would increase significantly if suitable solutes selectively reacted with the alcohol, modifying it into a highly water-soluble form. If this increase could be at least 1 order of magnitude and preferably even higher, the extraction using a waterbased solvent would become more attractive. Benzyl alcohol in toluene is used as a representative organic alcohol/solvent mixture, whereas the most promising reactive compounds observed experimentally for that case are also evaluated for the extraction of

10.1021/ie049620p CCC: $27.50 © 2004 American Chemical Society Published on Web 10/15/2004

Ind. Eng. Chem. Res., Vol. 43, No. 23, 2004 7573

cyclohexanol from cyclohexane and 1-hexanol from n-hexane. In this way, the applicability to cycloaliphatic and linear aliphatic alcohols, in addition to aromatic ones, is also explored. Because of the structural similarities of water and alcohols, it can be expected that potential extractants that can react with the hydroxyl group of the alcohol will also interact with water and, thus, will not significantly increase the distribution ratio of the alcohol. Therefore, in parallel with the above-mentioned approach, another option, in which the alcohol is first modified into a monoester and subsequently exposed to reactive extraction, is also evaluated.

Table 1. Reversible Chemical Reactions Considered for the Reactive Extraction of Unmodified Alcohol into Aqueous Solution

2. Reactive Extraction of Unmodified Alcohol Few applications of reactive extraction in the recovery of alcohols, from either the organic or the aqueous phase, are known. Although certain amines and phosphoryl compounds have been evaluated as extractants for the recovery of ethanol from an aqueous solution into an organic solvent,2,3 these attempts did not result in a significantly larger distribution ratio than already provided by conventional solvents. Some authors applied a reaction to enable the extraction of diols.4,5 However, this case is closer to physical than reactive extraction, given that it involves a reaction in which the diol is converted into a less polar component prior to and not during the extraction. Different reversible chemical reactions are considered in this paper for use in the reactive extraction of alcohols. To ensure no solubility in the organic phase, the essential criterion for selecting suitable reactions is that both the reactive solute and the reaction product are charged, or at least highly water-soluble, species. The options considered here are presented in Table 1. It is known that diols and polyols form reversible charged complexes with borate salts.6-10 On the other hand, for monohydroxyl alcohols, this complexation is not reported in the literature and hence requires evaluation. Sodium borate is used to represent the borate salts. Another option involving borates is a potential complexation of a monohydroxyl alcohol with a boratediol complex.11 The borate complex formed with 1,2propanediol is used for this purpose. Esterification, as one of the most characteristic reaction of alcohols, is often used for their recovery.13 To apply this reaction, a monosalt of dicarboxylic acid, which will yield an ester that is also a salt, is explored by employing monosodium maleate to the extraction of benzyl alcohol. By forming an inclusion complex with such cyclodextrins, low-water-soluble alcohols potentially can be extracted into the aqueous phase. On the basis of geometry, a molecule with the size of toluene or benzyl alcohol fits the best in the cavity of β-cyclodextrins. This can be confirmed by the highest values of the complexation constants with toluene for β- compared to R- and γ-cyclodextrins.14 This suggests that a β-cyclodextrin can be used for the purposes of this investigation. Concerning the R-cyclodextrins, although the size of the cavity does not enable inclusion complexation of toluene, it is still reported that a toluene-size molecule, but with a hydroxyl group (such as phenol), has a significantly higher complexation constant than toluene itself.15-17 Therefore, a hydroxypropyl-R-cyclodextrin with an average molar substitution (AMS) of 0.6 and two hydroxy-

propyl-β-cyclodextrins with different degrees of substitution (0.6 and 1.0) are experimentally evaluated. The transition metal ions are known for their reversible π-complexation with unsaturated and aromatic hydrocarbons.18,19 Silver nitrate is chosen to represent this type of extractant. The overall distribution ratio of the alcohol (DAlc) was calculated by dividing the total concentration of the alcohol in the aqueous phase by its concentration in the organic phase, at equilibrium

DAlc )

caAlc + rcaCEAlc coAlc

(1)

where the subscript Alc represents alcohol and CEAlc the chemically extracted alcohol. r is the number of alcohol molecules per molecule of chemically extracted product (r ) 1 in all reactions with the exception of the reaction with borate where r ) 2). The selectivity of a solvent toward alcohol (βAlc) is calculated as the ratio of equilibrium distribution ratios of alcohol and apolar organic solvent (AOS)

βAlc )

DAlc DAOS

(2)

3. Reactive Extraction of Modified Alcohol Only a few examples can be found in the literature concerning the conversion of an alcohol into another compound in order to enable or improve its separation. In one case, an alcohol was converted into a ketone to allow the distillation of a ketone originally present in the solution.20 In another, propylene glycol or 1,3-

7574

Ind. Eng. Chem. Res., Vol. 43, No. 23, 2004

Scheme 1. (a) Monoesterification of Benzyl Alcohol by a Cyclic Anhydride, (b) Dissociation of a Monoester of Benzyl Alcohol by a Basic Salt, and (c) Hydrolysis of a Deprotonated Benzyl Alcohol Monoester in an Aqueous Salt Solution

Table 2. Properties of the Selected Cyclic Anhydrides cyclic anhydride

melting point (°C)

pKa2 of dicarboxylic acid

solubilitya

phthalic hexahydrophthalic 3-methylglutaric

131 29 40

5.40 6.76 5.44

lowb lowb NAc

a Water solubility of dicarboxylic acid (hydrolyzed cyclic anhydride). b Only qualitative information available.46 c No literature reports on the solubility were found; it must be evaluated prior to an eventual application.

propanediol was converted into dioxanes and efficiently extracted from dilute aqueous solutions into apolar organic solvents.4,5 To apply this approach to the extraction of an alcohol from an apolar organic solvent, the following criteria have to be satisfied: (i) the modification has to result in a compound that can be easily (and selectively) extracted into the aqueous phase, (ii) the modification reaction has to be attainable in the apolar organic solvent and must be reversible so that the alcohol can be regenerated after extraction, (iii) the modifying agent should not contaminate nor have any undesirable effects on the other products in the organic phase, and (iv) the agent should be easily regenerated after separation so that it can be reused. To fulfill the first criterion, the fact that carboxylic acids can be efficiently extracted into alkali aqueous solutions initiated the consideration of an alcohol modification into a carboxylic acid prior to extraction. Partial esterification of a cyclic anhydride, so-called monoesterification, would convert an alcohol into a monoester that is, at the same time, a carboxylic acid (Scheme 1a). This is an equilibrium reaction in which the generation of the monoester is favored. Monoesterification of different aliphatic alcohols with phthalic anhydride,21-26 maleic anhydride,27 or hexahydrophthalic anhydride28 is observed to be reasonably fast without any catalyst. Specific examples for benzyl alcohol are also observed.29,30 The formed monoester, i.e., the modified alcohol, has one carboxylic group available that can dissociate in the presence of a basic salt, producing a charged and therefore more water-soluble carboxylate (Scheme 1b). The overall distribution ratio of benzyl alcohol DBAlc in such a case is defined as

DBAlc )

a a caME + cME - + cBAlc

coME + coBAlc

(3)

and it is not considered at equilibrium, but monitored as a function of time. The subscript ME represents protonated monoester, ME- is deprotonated monoester,

and Balc is benzyl alcohol. Superscripts o and a denote the organic and aqueous phases, respectively. A suitable extraction salt has to be selected on the basis of the dissociation constant of the monoester, which can be taken as equal to the second dissociation constant of the adequate dicarboxylic acid.31,32 If a salt such as hydrogen carbonate could be applied (BAnn- is HCO3- in Scheme 1b), just a mildly basic solution (pH ≈ 9) will be involved, and the dissociation product will be carbonic acid (H2CO3), which can be removed from the aqueous phase as carbon dioxide. Once extracted into the aqueous phase, the monoester, either protonated or deprotonated, starts to hydrolyze. During the hydrolysis, which occurs spontaneously in the aqueous solution, the alcohol is regenerated. Hence, after extraction and separation of the two liquid phases, the alcohol can be regenerated from the monoester by hydrolysis in the aqueous solution (Scheme 1c). To increase the rate of hydrolysis during the regeneration, a pH shift33,34 and/or temperature shift,35 but also alternative methods such as ultrasonic acceleration36 or biodegradation,37 can be employed. A temperature shift is preferable because a pH shift usually requires introduction of an additional chemical that needs to be removed afterward. The regenerated alcohol could be back-extracted into a high-molecularweight apolar organic solvent and then distilled or, if it is poorly soluble in water, simply separated from the aqueous phase by decantation of the newly formed alcohol phase. The other reaction product of the monoester hydrolysis is the deprotonated dicarboxylic acid (Scheme 1c). Moreover, if an excess of anhydride is used in the modification reaction, it will also be extracted into the aqueous phase as deprotonated dicarboxylic acid. To be reused in the process, the acid has to be protonated and then converted back into the anhydride. Use of carbon dioxide under pressure for this protonation is the preferential option. In that case, a benign chemical would be used, and only hydrogen carbonate, which is not a byproduct, but a desired product given that it is used for the extraction of the monoester, would be generated.12 Concerning the conversion into the anhydride, many methods for dehydration of a dicarboxylic acid exist (using acetic anhydride, trifluoroacetic anhydride, or ethoxyacetylene).38 However, the simplest approach is conversion by heating. This option is feasible only for cyclic anhydrides that contain five to seven atoms in the ring.39 Three cyclic anhydrides were chosen for experimental evaluation of the monoesterification of benzyl alcohol in toluene: phthalic, hexahydrophthalic, and 3-methylglutaric anhydride (see Table 2). Although phthalic anhydride is poorly soluble in water, it is still included in the evaluation as one of the most produced and cheapest cyclic anhydrides, but also because the benzyl


alcohol monoester of this anhydride is used as the model compound for evaluation of extraction of a modified alcohol. Hexahydrophthalic anhydride was selected as a representative of highly soluble anhydrides although its pKa2 is higher than that of carbonic acid, denying the possibility of extracting the monoester of this anhydride using a hydrogen carbonate salt. 4. Experimental Section 4.1. Chemicals. Benzyl alcohol, 1-hexanol, cyclohexanol, cyclohexane, toluene, n-hexane, sodium hydrogen carbonate, di-sodium tetraborate decahydrate, ethanol, silver nitrate, phthalic anhydride, and benzoic acid were supplied by Merck (Darmstadt, Germany); dibenzofuran by Acros Organics (Geel, Belgium); and 1,2-propanediol, monosodium maleate, monobenzyl phthalate, hexahydrophthalic anhydride, hydroxypropyl-R-cyclodextrin with a molar substitution of 0.6, hydroxypropyl-βcyclodextrin with a molar substitution of 0.6, and hydroxypropyl-β-cyclodextrin with molar substitution of 1.0 by Sigma-Aldrich (Zwijndrecht, The Netherlands). 1,2,4-Trimethylbenzene was obtained from Fluka (Zwijndrecht, The Netherlands), whereas N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) was from Alltech (Breda, The Netherlands). All chemicals were used as received. MiliQ water was used in all experiments. 4.2. Reactive Extraction of Unmodified Alcohol. For the reactive extraction of the unmodified alcohol, the set of jacketed glass cells already described in an earlier paper was used.40 Forty milliliters of each the organic and aqueous solution were introduced into each vessel. The alcohol mass fraction in the initial organic solution was 1.5 wt %, and the initial concentration of extractant in the aqueous phase was varied between zero and its solubility. After the introduction of both solutions, the lid was placed on the cell, and the mixture was stirred at 450 rpm. After agitation (the time of agitation required to reach equilibrium was determined prior to equilibrium experiments), the mixture was allowed to settle for 1 h. During the agitation and settling, the temperature of the fluid in the vessels was kept constant at the required value with an uncertainty of 0.1 °C. After the settling period, 1-2 mL samples of both the organic and aqueous layers were taken for quantitative analysis. 4.3. Reactive Extraction of Modified Alcohol. Monoesterification Reaction. The monoesterification reaction was analyzed on the automated workstation described in detail in an earlier paper.40 The concentrations of reactants and product were measured as a function of reaction time. The reaction was performed in 20-mL vials, into which 17 mL of the organic mixture containing the anhydride and the organic solvent (toluene, trimethylbenzene, cyclohexane, n-hexane) had been introduced. The vial was placed in the orbital shaker and heated to the required temperature (in the range of 60-150 °C) with an uncertainty of 1 °C. The shaking was done by applying a rotation speed of 750 rpm, with 30 s of rotation in one direction, a 5-s pause, and 30 s of rotation in the opposite direction. The reaction was initiated by introducing the alcohol into the vial. The quantity varied with experiment, but it was not higher than 1.5 wt %. At specified times, 60-µL samples were taken from each vial and dispensed into 2-mL vials that contained about 0.5 mL of a chemical derivatization agent. The sample was left for at least 30 min to allow complete derivatization of the monoester and the alcohol. After

that, the sample was injected into the gas chromatograph (GC) for quantitative analysis. The purpose of the derivatization was to enable chemical analysis of the monoester, but also to halt the reaction at the time the sample was taken by converting all benzyl alcohol and monoester present in the sample into the inert silyl derivatives. Dissociation Extraction. The dissociation extraction experiments were performed on the same automated workstation as the monoesterification experiments. The extraction was done in 2-mL vials into which 850 µL of each phase, at the extraction temperature, had been introduced. The organic phase was a 0.1 M solution of monobenzyl phthalate in toluene, representing the modified benzyl alcohol solution, whereas the aqueous phase contained sodium hydrogen carbonate at different concentrations. During extraction, no agitation was applied because the pH probe occupied a significant volume of the experimental vessel and did not allow for it. As compensation for a possible nonuniform concentration profile within the phase because of the absence of agitation, a significant amount of the organic phase was sampled. At predefined times, the robot arm took a sample of 100 µL of the upper, organic layer and transferred it into a 2-mL vial that contained about 0.5 mL of the derivatization agent. The derivatized sample was analyzed by gas chromatography. No samples of the aqueous phase were taken as the used chemical method did not allow analysis of the aqueous phase. 4.4. Chemical Analysis. The components were quantified either by a CP-3800 gas chromatograph (GC) system with an FID detector, by a Series 2500 highpressure liquid chromatograph (HPLC), or by a Prostar HPLC (all from Varian, Palo Alto, CA). The characteristics of the methods used for each specific case are given below. Reactive Extraction of Unmodified Alcohol. For almost all cases, except when samples of the equilibrium experiments using aqueous cyclodextrin solutions were analyzed, high-pressure liquid chromatography was used. Reverse-phase chromatography was applied with Kromasil (n-decyl C8 chains) (250 mm × 4.6 mm; 10µm packing) as the stationary phase. UV detection at a wavelength of 254 nm was used for detection of the analytes. A mixture of methanol and ammonium acetate at a flow rate of 1 mL/min was used as the mobile phase. The determination of the concentration of the components was based on previously established calibration curves. Gas chromatography was used for determination of the concentrations in the experiments involving cyclodextrins. An ECTM-wax column (30-m length, 0.32-mm diameter, 0.25-µm packing; Alltech) was applied. The column temperature was raised from 100 to 240 °C in increments of 40 °C/min. Helium was used as the carrier gas with an initial flow of 11.8 mL/min. The temperatures of the detector and injector were kept constant at 300 and 275 °C, respectively, with a pressure in the injector of 276 kPa. Quantification of the components in the sample was done by using an internal standard method in which a 0.25 M solution of dibenzofuran in ethanol was used as the internal standard. Reactive Extraction of Modified Alcohol. Because of poor stability, the monoesters had to be derivatized prior to chemical analysis. Silylation using N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) was

7576


Figure 1. Overall equilibrium distribution ratio of benzyl alcohol between aqueous solution and toluene as a function of the initial extractant concentration in the aqueous phase at 25 °C: ), silver nitrate; 0, hydroxypropyl-β-cyclodextrin (AMS ) 1.0); b, hydroxypropyl-β-cyclodextrin (AMS ) 0.6); 2, hydroxypropyl-R-cyclodextrin (AMS ) 0.6); O, sodium borate; 4, 1,2-propanediol-borate complex; 1, monosodium maleate. The initial mass fraction of benzyl alcohol in toluene is 1.5 wt % (0.1 mol/L). The phase volume ratio is 1. The points represent the experimental results, and the lines show their trend.

the derivatization method applied. This procedure was chosen because it enables a fast conversion of the monoesters, but also of the alcohols, into stable silyl derivatives.41-43 Gas chromatography with a CP-SIL column (25-m length, 0.15-mm diameter, 1.2-µm packing; Varian) was used for this analysis. The column temperature was raised from 50 to 300 °C at a rate of 20 °C/min and, at the end, maintained at 300 °C for 5 min. The temperatures of detector and injector were kept constant at 300 and 270 °C, respectively, with a pressure in the injector of 154 kPa. Hydrogen was used as the carrier gas. Quantification of the components was done by using the apolar organic solvent present in the sample as the internal standard. 5. Results and Discussion 5.1. Reactive Extraction of Unmodified Alcohol. The measured overall equilibrium distribution ratios of benzyl alcohol for all evaluated extractants are summarized in Figure 1. Using pure water as the extraction solvent, it is confirmed that the mass balance holds within 2%. Only hydroxypropyl cyclodextrins and silver nitrate came close to the desired value of the distribution ratio of at least 3 for benzyl alcohol, which is 1 order of magnitude higher than the physical distribution ratio. The enhancements achieved by the other evaluated extractants were insufficient, and therefore, those results will not be discussed further. In addition, reactive extraction of unmodified alcohol is not investigated for cyclohexanol and 1-hexanol because of the only moderate improvement of the distribution ratio of benzyl alcohol. Silver nitrate improves the distribution ratio of benzyl alcohol by a factor of 5 relative to that in pure water. On the other hand, the results on selectivity (Figure 2) show that such a silver nitrate solution has a lower selectivity toward benzyl alcohol than pure water,

Figure 2. Selectivity of aqueous silver nitrate and cyclodextrin solutions toward benzyl alcohol relative to toluene as a function of the initial extractant (AgNO3 or cyclodextrin) concentration in the aqueous phase at 25 °C: 0, silver nitrate; 9, hydroxypropylβ-cyclodextrin (AMS ) 1.0); b, hydroxypropyl-β-cyclodextrin (AMS o,0 ) 0.6); 2, hydroxypropyl-R-cyclodextrin (AMS ) 0.6). wBAlc ) 1.5 o,0 wt % (cBAlc ) 0.1 mol/L). The points represent the experimental results, and the lines show their trend.

confirming the interaction of silver with both aromatic compounds in the solution. The decreasing trend of the selectivity in Figure 2 suggests even that the silver ion interacts more with toluene than with benzyl alcohol. However, because the selectivity is still rather high, an aqueous solution of silver nitrate can be used for the recovery of benzyl alcohol. All three evaluated cyclodextrins cause an increase of the distribution ratio of benzyl alcohol. The maximum values reached are between 1.3 and 2.7, which makes them the most attractive among the evaluated extractants. β-Cyclodextrins show a higher distribution ratio than the R form, whereas the β-cyclodextrin with the lower degree of substitution performs somewhat better. Possibly, a further reduction in the degree of substitution could enhance distribution ratio even further. Obviously, complexation of benzyl alcohol also occurs within the smaller cavity of an R-cyclodextrin. The selectivity of these extractants toward benzyl alcohol is shown in Figure 2. A decrease in selectivity for β-cyclodextrin solutions relative to that of pure water can be seen, whereas for R-cyclodextrin, it is safe to say only that the selectivity is similar to that of pure water. With a distribution ratio of about 3 (at a solvent-to-feed ratio of 1) and a selectivity toward the alcohol of at least 100, aqueous solutions of hydroxypropyl cyclodextrins have sufficient potential to be used for the recovery of benzyl alcohol. 5.2. Reactive Extraction of Modified Alcohol. 5.2.1. Monoesterification Reaction. Figure 3 shows typical concentration profiles of both reactants (alcohol and anhydride) and the product (monoester) in the reaction of benzyl alcohol with hexahydrophthalic anhydride. It can be seen that the consumption rates of both alcohol and anhydride are equal to the generation rate of monoester. This observation suggests that the desired reaction shown in Scheme 1a is occurring. Hence, this also means that there are no side reactions, such as diesterification. For the reaction of benzyl alcohol with phthalic anhydride, it was confirmed by chemical analysis that the product is really the desired monoester.


Figure 3. Concentrations as a function of time of the reactants and product in the monoesterification of benzyl alcohol with hexahydrophthalic anhydride in toluene at 100 °C when anhydride is used in excess (open symbols) and when equimolar amounts of the two reactants are used (solid symbols): b, O hexahydrophthalic anhydride; 9, 0 benzyl alcohol; 2, 4 monobenzyl hexahydrophthalate. The points represent the experimental results, and the lines show their trend.

Figure 4. Conversion of benzyl alcohol (ξBAlc) as a function of time and temperature in the reaction with hexahydrophthalic anhydride (b, 60 °C; 9, 100 °C; 2, 150 °C), phthalic anhydride (0, 100 °C; O, 125 °C; 4, 150 °C), and 3-methylglutaric anhydride (x, 60 °C; !, 100 °C; ∆, 150 °C) in toluene. A 6-fold molar excess of hexahydrophthalic and phthalic anhydrides was used, whereas the phthalic anhydride concentration was equal to its solubility at each o,0 o,0 temperature. wBAlc ) 0.8 wt % (cBAlc ) 0.06 mol/L). The points represent the experimental results, and the lines show their trend.

The concentrations of alcohol, anhydride, and monoester in the noncatalyzed monoesterification reaction performed in a dilute organic solution were measured as a function of time at various temperatures. The benzyl alcohol conversions in those reactions are shown in Figure 4 for benzyl alcohol with various anhydrides. A conversion higher than 90% can be achieved using hexahydrophthalic anhydride. This conversion is realized in approximately 100 min at 150 °C, which can be considered as relatively fast compared to a typical residence time in an oxidation reactor of 240-1800 min.44 Because hexahydrophthalic anhydride exhibits good solubility in apolar solvents, the conversion and reaction rate can made be even higher by using a higher excess of anhydride or a higher temperature. Regarding the temperature effect, this reversible reaction seems to be exothermic. It should be emphasized that the

Figure 5. Conversions of cyclohexanol (solid symbols) and 1-hexanol (open symbols) as a function of time at various temperatures in the reaction with 3-methylglutaric anhydride in toluene: b, O o,0 60 °C; 9, 0 100 °C; 2, 4 150 °C. wAlc ) 0.8 wt %. A 6-fold molar excess of anhydride is used. The points represent the experimental results, and the lines show their trend.

measurements at 60 and 100 °C were performed in toluene, whereas those at 150 °C were actually obtained in 1,2,4-trimethylbenzene. This substitution is acceptable, given that, at 100 °C, both the rates of the reaction and the equilibrium conversions are the same in the two solvents. Both the equilibrium conversion and the rate of reaction of benzyl alcohol with phthalic anhydride are lower than in the reaction with hexahydrophthalic anhydride. This is due to the low concentration of phthalic anhydride, which is equal to its solubility limit. The highest equilibrium conversion of benzyl alcohol is achieved with 3-methylglutaric anhydride. At 150 °C, the alcohol conversion into the monoester is almost 100%, whereas at the lower temperatures, equilibrium is not reached within the monitored time, although the equilibrium conversion can be expected to be at least as high as that at 150 °C considering the exothermic nature of the monoesterification reactions observed for the other two anhydrides. This suggests that the excess of anhydride does not need to be as high as 6-fold to obtain a high conversion of benzyl alcohol. On the other hand, the reaction rate in this case seems to be much more sensitive to the temperature than is the case with either hexahydrophthalic or phthalic anhydrides. Hence, the highest equilibrium conversions at reasonable reaction rates make 3-methylglutaric anhydride the best modifier for benzyl alcohol. Therefore, this anhydride was also selected for the experimental evaluation of the monoesterification reaction with the other two alcohols (see Figure 5). At the same temperature, the reaction rate decreases in the order 1-hexanol > benzyl alcohol > cyclohexanol. The rate sensitivity to temperature, i.e., the activation energy, seems to be similar for benzyl alcohol and cyclohexanol, whereas it is lower for 1-hexanol. On the other hand, the lowest equilibrium conversion at 150 °C was found for 1-hexanol. It is only about 50%, compared with 90% for cyclohexanol and close to 100% for benzyl alcohol. The reasons for these differences will not be discussed here because they are beyond the scope of the current investigation. In any case, it is confirmed that different alcohols can be efficiently modified into monoesters.

7578


Table 3. Experimentally Determined Monoesterification Rate Constants (k1) and Equilibrium Constants (K) anhydride

k1 [L/(mol min)]

K (L/mol)

hexahydrophthalic 3-methylglutaric

Benzyl Alcohol 4.6 × 102 exp(-3.0 × 104/RT) 1.9 × 105 exp(-5.2 × 104/RT)

2.3 × 10-5 exp(5.1 × 104/RT) very high

Cyclohexanol 7.9 × 103 exp(-4.5 × 104/RT)

4.7 × 10-2 exp(2.2 × 104/RT)

5.9 × 102 exp(-3.3 × 104/RT)

very high

3-methylglutaric 1-hexanol 3-methylglutaric

To quantify the reaction rates and equilibria, the reaction equilibrium constants and the rate constants are determined from the experimental data. This was done for 3-methylglutaric and hexahydrophthalic anhydrides, whereas the reaction with phthalic anhydride was not evaluated as not all of the anhydride introduced dissolved. According to the literature, the monoesterification is first order with respect to both reactants.24 Therefore, we assume the rate equation to be a reversible reaction with first-order kinetics with respect to each of the reactants, but also with respect to the product k1

A + B {\ }C k 2

-rA ) k1cAcB - k2cC

(4)

where A represents alcohol, B anhydride, and C the monoester. ci is the concentration at time t, k1 and k2 are reaction rate constants, and rA is the reaction rate. The integrated rate equation for c0C ) 0 and c0A * c0B 45 is

k1t ) (c0A - ceA)

solute (monobenzyl phthalate) were present in equimolar amounts. As expected, in all graphs, the distribution ratio raises from zero toward a maximum, after which it decreases toward the asymptote determined by the physical distribution ratio of the alcohol (about 0.3 for benzyl alcohol at this temperature). However, the value of the maximum reached differs from case to case. Figure 6 shows that an aqueous sodium hydrogen carbonate solution can be used for the extraction of monobenzyl phthalate. A distribution ratio of about 200 is reached when NaHCO3 is used in excess. This is a 3-orders-of-magnitude increase of the distribution ratio relative to the physical distribution ratio of benzyl alcohol in pure water. When NaHCO3 is in excess, the pH of the aqueous phase during the extraction is almost constant at about 9. As can be seen from the monoester concentration

(c0A - ceA)[c0A c0B - (c0A - ceA) (c0A - cA)]

ln c0A c0B - (c0A - ceA)2

c0A c0B[(c0A - ceA) - (c0A - cA)] (5)

where c0A and c0B are the initial concentrations of reactants and ceA is the concentration of alcohol at equilibrium. The right-hand-side term of eq 5 for each reaction at each measured temperature was plotted as a function of time. It was concluded that the experimental results can be successfully fitted by linear curves, suggesting that the proposed reaction rate equation is indeed adequate. Hence, for the reactions that are far from equilibrium, the equilibrium concentration is estimated by finding the best linear fit. The obtained slope of each graph represents the reaction rate constant, whereas the reaction equilibrium constant was calculated from the measured or estimated concentrations of monoester, alcohol, and anhydride at equilibrium. The reaction rate and equilibrium constants for the same reaction were fitted as a function of temperature by an Arrheniustype of equation, yielding the activation energy and preexponential constant. The obtained correlations are reported in Table 3. 5.2.2. Dissociation Extraction. Figure 6 shows the concentration profiles of benzyl alcohol and monobenzyl phthalate in the organic phase, as well as the calculated distribution ratio, for the extraction by an aqueous solution of sodium hydrogen carbonate. The first graph gives the results when the salt was used in excess, whereas for the lower graph, the extractant (salt) and

Figure 6. Concentrations of the components in the toluene-rich phase (c0i , open symbols) and overall distribution ratio of benzyl alcohol (DBAlc, solid symbols) as a function of time in the extraction of the modified alcohol at 50 °C into a (a) 1 M and (b) 0.1 M NaHCO3 aqueous solution: 0, benzyl alcohol; O, monobenzyl phthalate. The phase volume ratio is 1. The points represent the experimental results, and the lines show their trend.


noticeable influence of temperature on the extractability of the monoester. In conclusion, the alcohol regeneration can indeed be controlled by adjusting the temperature of the extract. 6. Conclusions

Figure 7. Concentrations of the components in the organic phase (c0i ) as a function of time in the extraction of modified alcohol into a 1 M NaHCO3 aqueous solution at 50 °C (solid symbols) and 70 °C (open symbols): 0, 9 benzyl alcohol; O, b monobenzyl phthalate. The phase volume ratio is 1. The points represent the experimental results, and the lines show their trend.

profile in Figure 6a, this pH is sufficient to enable a complete extraction of the monoester. On the other hand, not more than several percent of the extracted monoester hydrolyzes in about 30 h. Hence, it can be concluded that the sodium hydrogen carbonate obviously has the desired characteristics for a good extractant, as it enables complete dissociation of the monoester but does not cause a very fast hydrolysis. A relatively slow hydrolysis enables the distribution ratio to be maintained at high values for a longer period of time. For equimolar amounts of monoester and extractant, a maximum distribution ratio of about 10 is achieved. During this extraction, the pH of the aqueous phase decreases from above 9 to about 6, whereas the largest decrease occurs in the initial stage of the extraction. From the alcohol concentration profile (see Figure 6b), it is clear that the monoester hydrolysis is very slow in this pH range. This results in an efficient extraction, where the distribution ratio is almost constant at its maximum of about 10 for a long period. Having a pH of 6 at the end of monitoring suggests that roughly all of the NaHCO3 is consumed in the extraction. However, the final concentration of monoester in the organic phase does not go lower than 0.01 mol/L. This indicates that not all of the introduced extractant is used for the dissociation of monoester and implies that NaHCO3 should be used in slight excess. Finally, it should be emphasized that, for the above case of modified alcohol extraction, no influence of the mass-transfer rate has been evaluated. However, although the results might be different under different hydrodynamic conditions, it is confirmed by this work that very high values of the distribution ratio can be achieved. 5.2.3. Regeneration of Alcohol by a Temperature Shift. To evaluate the feasibility of the regeneration of benzyl alcohol from the extracted monoester by a temperature shift, the extractions at 50 and 70 °C were compared. Figure 7 shows a nearly 10-times faster hydrolysis rate of benzyl alcohol resulting from the applied temperature increase. On the other hand, the monoester concentration profile at 70 °C is roughly the same as that at the lower temperature, showing no

Aqueous solutions containing reactive salts or other adequate reactive extractants might have great potential for use as reactive extraction solvents in the recovery of alcohols present in few-percent concentrations in apolar organic solvents. Two approaches for the recovery of slightly or moderately water-soluble monohydroxyl alcohols are evaluated. The first, in which an aqueous solution containing a reactive extractant is applied for extraction of unmodified alcohol, shows limited potential for application. The second considered approach, in which the alcohol is modified prior to extraction into an easily extractable monoester, shows much greater potential. For example, a benign solvent, such as an aqueous solution of sodium hydrogen carbonate, can provide a distribution ratio of benzyl alcohol of up to 200 with an unchanged solubility of toluene in the aqueous phase relative to that in pure water. Modification of various alcohols (benzyl alcohol, cyclohexanol, and 1-hexanol) into a monoester can be done efficiently in an apolar organic solvent, without the need for a contaminating catalyst. Monoesters, which contain a carboxylic group, are efficiently extracted by dissociation extraction. Furthermore, back-recovery of the extracted modified alcohol (monoester) can be performed by a spontaneous hydrolysis reaction whose rate can be controlled by the temperature. Acknowledgment We thank Mr. Henny Bevers, Ms. Annemarie Montanaro, Mr. Kimmo Heinamaki, and Ms. Tessa van Schoonhoven for their participation in the research. Furthermore, we are grateful for the financial support of DSM, NWO-CW (Netherlands Organization for Scientific ResearchsChemical Science), and Novem (Netherlands Agency for Energy and Environment). Literature Cited (1) Kuzmanovic´, B.; Kuipers, N. J. M.; Haan, A. B. de.; Kwant G. Reactive Extraction of Carbonyl Compounds from Apolar Hydrocarbons Using Aqueous Salt Solutions. Ind. Eng. Chem. Res. 2003, 42, 2885-2896. (2) Pai, R. A.; Doherty, M. F.; Malone, M. F. Design of Reactive Extraction Systems for Bioproduct Recovery. AIChE J. 2002, 48, 514-526. (3) King, C. J. Separation Processes Based on Reversible Chemical Complexation. In Handbook of Separation Process Technology; Rousseau, R. W., Ed.; John Wiley & Sons: New York, 1987. (4) Malinowski, J. J. Reactive Extraction for Downstream Separation of 1,3-Propanediol. Biotechnol. Prog. 2000, 16, 76-79. (5) Broekhuis, R. R.; Lynn, S.; King, C. J. Recovery of Propylene Glycol from Dilute Aqueous Solutions via Reversible Reaction with Aldehydes. Ind. Eng. Chem. Res. 1994, 33, 3230-3237. (6) Wise, E. T.; Weber, S. G. A Simple Partitioning Model for Reversibly Cross-Linked Polymers and Application to the Poly(vinyl alcohol)/Borate System (“Slime”). Macromolecules 1995, 28, 8321-8327. (7) Keita, G.; Ricard, A.; Audebert, R.; Pezron, E.; Leibler, L. The Poly(vinyl alcohol)-Borate System: Influence of Polyelectrolyte Effects on Phase Diagrams. Polymer 1995, 36, 49-54. (8) Chen, C. Y.; Yu, T. Dynamic Light Scattering of Poly(vinyl alcohol)-Borax Aqueous Solution Near Overlap Concentration. Polymer 1997, 38, 2019-2025.

7580


(9) Tsuchiya, H. High-Performance Liquid Chromatographic Analysis of Polyhydroxyflavones Using Solid-Phase BorateComplex Extraction. J. Chromatogr. B 1998, 720, 225-230. (10) Abe, Y.; Shoji, T.; Kobayashi, M.; Qing, W.; Asai, N.; Nishizawa, H. Enantioselective Distribution of Amino-Alcohols in a Liquid-Liquid Two-Phase System Containing Dialkyl L-Tartarate and Boric Acid. Chem. Pharm. Bull. 1995, 43, 262-265. (11) Shibayama, M.; Sato, M.; Kimura, Y.; Fujiwara, H.; Nomura, S. 11B nmr Study on the Reaction of Poly(vinyl alcohol) with Boric Acid. Polymer 1988, 29, 336-340. (12) Kuzmanovic´, B. Reactive Extraction of Oxygenates with Aqueous Salt Solutions. Ph.D. Thesis, Febodruk, Enschede, The Netherlands, 2003. (13) Burfeind, J.; Schu¨gerl, K. Reaktiv-Membran-Extraktion von Ethanol. Chem. Ing. Technik. 1999, 71, 93-96. (14) Sanemasa, I.; Akamine, Y. Association of Benzene and Alkylbenzenes with Cyclodextrins in Aqueous Medium. Bull. Chem. Soc. Jpn. 1987, 60, 2059-2066. (15) Szejtli, J.; Osa, T. Cyclodextrins. In Comprehensive Supramolecular Chemistry; Atwood, J. L., Davies, J. E. D., Macnicol, D. D., Vogtle, F., Eds.; Elsevier Science: Oxford, U.K., 1996; Vol. 3. (16) Rekharsky, M. V.; Inoue, Y. Complexation Thermodynamics of Cyclodextrins. Chem. Rev. 1998, 98, 1875-1917. (17) Ross, P. D.; Rekharsky, M. V. Thermodynamics of Hydrogen Bond and Hydrophobic Interactions in Cyclodextrin Complexes. Biophys. J. 1996, 71, 2144-2154. (18) Rege, S. U.; Padin, J.; Yang, R. T. Olefin/Paraffin Separations by Adsorption: π-Complexation vs Kinetic Separation. AIChE J. 1998, 44, 799-809. (19) Munakata, M.; Wu, L. P.; Ning, G. L. A New Type of Multilayer SystemsSilver(I) Complexes of Polycyclic Aromatic Compounds. Coord. Chem. Rev. 2000, 198, 171-203. (20) Bataafsche Petroleum. A Process for Separating Ketones from Mixtures Thereof with Other Compounds. GB Patent 730486, 1955. (21) Siggia, S.; Hanna, J. G. Quantitative Organic Analysis via Functional Groups; John Wiley & Sons: New York, 1979. (22) Maksimenko, E. G.; Kirilovich, V. I.; Lopatina, N. M.; Kutsenko, A. I.; Maksimova, G. V.; Zhuk, R. V.; Uskach, Ya. L.; Bogach, E. V.; Melnikov, M. D.; Kutyanin, L. I. Method of Obtaining Alkylbenzylphthalates. Russian Patent 1187420, 1994. (23) Ellington, J. B.; Brennecke, J. F. Pressure Effect on the Esterification of Phthalic Anhydride in Supercritical CO2. J. Chem. Soc., Chem. Commun. 1993, 1094-1095. (24) Schwegler, M. A.; van Bekkum, H.; de Munck, N. A. Heteropolyacids as Catalysts for the Production of Phthalate Diesters. Appl. Catal. 1991, 74, 191-204. (25) Thorat, T. S.; Yadav, V. M.; Yadav, G. D. Esterification of Phthalic Anhydride with 2-Ethylhexanol by Solid Superacidic Catalysts. Appl. Catal. A 1992, 90, 73-96. (26) Skrzypek, J.; Zbigniew Sadlowski, J.; Lachowska, M.; Nowak, P. Kinetics of the Esterification of Phthalic Anhydride with 2-Ethylhexanol; Part IV: Non-Catalytic Process; Final Kinetics Comparison of the Catalysts. Chem. Eng. Process. 1998, 37, 163167. (27) Grzesik, M.; Skrzypek, J.; Tuszynski, T. Esterification of Maleic Anhydride with 2-Ethylhexanol in Case When No Catalyst Is Used and In the Presence of Sulfuric Acid as a Catalyst. Inz. Chem. Procesowa 1999, 20, 351-362. (28) Uozumi, Y. Laboratory of Complex Catalysis, Institute for Molecular Science, Okazaki, Japan. Personal communication, 2001.

(29) Ejlertsson, J.; Meyerson, U.; Svensson, B. H. Anaerobic Degradation of Phthalic Acid Esters During Digestion of Muncipal Solid Waste Under Landfilling Conditions. Biodegradation 1996, 7, 345-352. (30) Bashkirov, A. N.; Grozhan, M. M.; Kamzolikin, V. V.; Lapitsky, J. A. Process for the Production of Aromatic Alcohols. GB Patent 1198233, 1970. (31) Uozumi, Y.; Yasoshima, K.; Miyachi, T.; Nagai, S. Enantioselective Desymmetrization of Meso-Cyclic Anhydrides Catalysed by Hexahydro-1H-Pyrrolo[1,2-c]imidazones. Tetrahedron Lett. 2001, 42, 411-414. (32) Suzuki, T.; Yaguchi, K.; Suzuki, S.; Suga, T. Monitoring of Phthalic Acid Monoesters in River Water by Solid-Phase Extraction and GC-MS Determination. Environ. Sci. Technol. 2001, 35, 3757-3763. (33) Okanagan University College, Programs/Departments, Science, Chemistry, Chemistry Faculty, Nigel Eggers, Chem 317. http://www.people.ouc.bc.ca/neggers/Chem%20317/Course%20Notes/ HYDROLYSIS.doc (last accessed July 2002). (34) Andres, G. O.; Granados, A. M.; de Rossi, R. H. Kinetic Study of the Hydrolysis of Phthalic Anhydride and Aryl Hydrogen Phthalates. J. Org. Chem. 2001, 66, 7653-7657. (35) Krammer, P.; Vogel, H. Hydrolysis of Esters in Subcritical and Supercritical Water. J. Supercrit. Fluids 2000, 16, 189-206. (36) Tuulmets, A.; Raik, P. Ultrasonic Acceleration of Ester Hydrolysis. Ultrason. Sonochem. 1999, 6, 85-87. (37) Cartwright, C. D.; Owen, S. A.; Thompson, I. P.; Burns, R. G. Biodegradation of Diethyl Phthalate in Soil by a Novel Pathway. FEMS Microbiol. Lett. 2000, 186, 27-34. (38) Cox, A. Dicarboxylic and Polycarboxylic Acids. In Comprehensive Organic Chemistry; Barton, S. D.; Ollis, W. D.; Sutherland, I. O., Eds.; Pergamon Press: Oxford, U.K., 1979; Vol. 2, pp 661-718, Nitrogen Compounds, Carboxylic Acids, Phosphorus Compounds. (39) Smith, M. B.; March, J. March’s Advanced Organic Chemistry, 5th ed.; John Wiley & Sons: New York, 2001; Chapter 10. (40) Kuzmanovic´, B.; Delden, M. L. van; Kuipers, N. J. M.; Haan, A. B. de. Fully Automated Workstation for Liquid-Liquid Equilibrium Measurements. J. Chem. Eng. Data 2003, 48, 12371244. (41) Jonsson, S.; Boren, H. Analysis of Mono- and Diesters of o-Phthalic Acid by Solid-Phase Extractions with PolystyreneDivinylbenzene-Based Polymers. J. Chromatogr. A 2002, 963, 393-400. (42) Regis Technologies, Chromatography, GC Derivatization Reagents, Derivatization Procedures. http://www.registech.com/ gc/procedures.pdf (last accessed June 2003). (43) Alltech Associates Inc., Applied Science Labs. GC Derivatization Guide; Alltech-Applied Science: Deerfield, CT, 1986. (44) Kirk, R. E.; Othmer, D. F. Kirk-Othmer Encyclopedia of Chemical Technology; Wiley-Interscience: New York, 2003. (45) Wallas, S. M. Chemical Reaction Engineering Handbook of Solved Problems; Gordon and Breach: Amsterdam, 1995. (46) Grasselli, J. O., Ed. Atlas of Spectral Data and Physical Constants for Organic Compounds; CRC Press: Boca Raton, FL, 1973.

Received for review May 8, 2004 Revised manuscript received August 12, 2004 Accepted August 26, 2004 IE049620P

Reactive Extraction of Alcohols from Apolar Hydrocarbons with

Recommend Documents