Selectivity between Lactic Acid and Glucose during Recovery of Lactic

Mar 15, 1996 - broths, methods of recovery that utilize separating agents, such as .... with the C8 chain predominating about 2 to 1 (Henkel. Corp., 1...
2 downloads 0 Views 572KB Size
Download PDF
Ind. Eng. Chem. Res. 1996, 35, 1215-1224

1215

Selectivity between Lactic Acid and Glucose during Recovery of Lactic Acid with Basic Extractants and Polymeric Sorbents Youyuan Dai† and C. Judson King* Lawrence Berkeley Laboratory and Department of Chemical Engineering, University of California, Berkeley, California 94720

During recovery of product carboxylic acids from fermentation broths, it is important to maximize the selectivity for the desired acid, as opposed to substrate sugars. In this work uptakes of glucose and competitive uptakes of lactic acid and glucose have been measured for the extractant Alamine 336 in various diluents and three commercially available basic solid polymeric sorbents. The results show that swelling is the main factor governing the selectivity between lactic acid and glucose for the polymeric sorbents. Because of a high uptake capacity and relatively low swelling, Dowex MWA-1 gives a higher selectivity in the pH 5-6 range than do Amberlite IRA35 and Reillex 425. Extraction with Alamine 336 provides a much higher selectivity, but a lower capacity, than the polymeric sorbents. The extent of water coextraction depends strongly on the diluent used, and larger amounts of water coextracted correspond to larger uptakes of glucose. Introduction Carboxylic acids are readily made by fermentation and are among the most attractive substances that can be manufactured from biomass (Ng et al., 1983; Jain et al., 1989). Recovery of carboxylic acids from fermentation broths presents a challenging separation problem. Because of the dilute, complex nature of fermentation broths, methods of recovery that utilize separating agents, such as extractants and solid sorbents, that are selective for carboxylic acids are attractive (King, 1987; Tung and King, 1994). [The terms “sorbent” and “sorption” are used to connote situations where a polymeric solid can take up solute by either or both surface attachment (adsorption) or bulk infusion (absorption).] Important characteristics of extractants and solid sorbents are a high capacity for the acid, a high selectivity for the acid as opposed to water and substrate (e.g., glucose), regenerability, and, depending upon the process configuration, biocompatibility with microorganisms. Many fermentations, such as that for lactic acid production, are subject to end-product inhibition (Yabannavar and Wang, 1991; Davison and Thompson, 1992). If a solid sorbent can be used in situ or in an external recycle loop, higher overall yields can be achieved. Many fermentations to produce carboxylic acids operate most effectively at pH above pKa1 of the acid product. For example, lactic acid (pKa ) 3.86), is typically produced at pH 5-6 (Vickroy, 1985; Jain et al., 1989). One approach for recovering carboxylic acids from such solutions is to use agents that are sufficiently basic to retain a substantial capacity several pH units above the pKa1 of the carboxylic acid. Tung and King (1994) investigated extraction and sorption of lactic and succinic acids using different basic extractants and polymeric sorbents. The results showed that the uptake in the pH 5-6 range varied substantially from one agent to another and was strongly dependent upon the basicity and capacity of the agent. Similar results have been * To whom correspondence should be addressed. † Present address: Department of Chemical Engineering, Tsinghua University, Beijing 100084, China.

0888-5885/96/2635-1215$12.00/0

reported by other researchers (e.g., Yabannavar and Wang, 1987, 1991; Yang et al., 1991; Evangelista et al., 1994). Agents to be used in fermentation processes should provide high selectivity between the product carboxylic acid and substrate sugars. In addition to conservation of substrate, one reason why very high selectivities are often sought is the tendency for small amounts of sugars in a product such as lactic acid to cause discoloration. Reported measurements of uptake capacities for sugars and for selectivity achieved between carboxylic acids and sugars with extractants and/or solid sorbents are few and fragmentary. Kaufman et al. (1994) screened a series of solid sorbents preliminarily for possible utilization in biparticle fluidized-bed fermentation to produce lactic acid from immobilized Lactobacillus delbreuckii, measuring selectivity between acid and sugar as well as other properties. Kaufman et al. (1995) presented selectivities between lactic acid and glucose obtained with a fixed bed of the sorbent Amberlite IRA-35 under a particular set of operating conditions. Evangelista et al. (1994) reported breakthrough curves for lactic acid and glucose during fixed-bed adsorption with actual fermentation broths. Wennersten (1983) investigated extraction of citric acid from fermentation broths by the extractant Alamine 336 in different diluents and observed the color of the extract phase and the phase separation rate. The best selectivity and phase separation were obtained using a nonpolar diluent, e.g., n-hexane. Wang et al. (1991) reported that the partition coefficient of glucose between an aqueous solution and trioctylphosphine oxide (TOPO) in kerosene was 0.002, implying a high selectivity for carboxylic acids as opposed to glucose. San-Martin et al. (1992) reported results for extraction of lactic acid by Alamine 336 in various diluents and noted no effect of the simultaneous presence of lactose on the uptake capacity for lactic acid. Scho¨ller et al. (1993) measured uptakes of lactic acid with a liquid-membrane system and reported that the presence of glucose affected the swelling of the liquidmembrane material but did not affect the uptake of lactic acid. In this work, equilibria for uptake of glucose alone and competitive uptakes of lactic acid and glucose from aqueous solution were examined over a wide pH range © 1996 American Chemical Society

1216

Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996

Table 1. Properties of Basic Sorbents Useda sorbent manufacturer matrix functional group max temp (°C) apparent pKa N2 BET surface area (m2/g) a

Dowex MWA-1 Dow Chemical Co. poly(styrene-divinylbenzene) 3° amine 100 6.70b 23

Reillex 425 Reilly Tar & Chemical Co. poly(4-vinylpyridine) pyridine 100 3.92b 90

Amberlite IRA-35 Rohm & Haas Co. acrylic 3° amine 60 8.32b 20-30

Sources: Dow Chemical Co., 1983; Reilly Industries, 1989; Rohm & Haas Co., 1981. b Gustafson et al., 1970.

and at different temperatures for three basic polymeric sorbents as well as for Alamine 336 in various diluents. Experiments were carried out with synthetic aqueous solutions rather than actual fermentation media so as to control conditions and limit the presence of other substances that could affect the equilibria.

(Waters Model 401) detector. A mobile phase of 0.01 N H2SO4 was used. The decrease of glucose concentration in aqueous solution was used to compute the composite uptake of glucose, i.e., the degree of surface or site enrichment, by eq 1 (Kipling, 1965). Any water and/or glucose

Qcg ) W0(Cg,i - Cg,f)/m

Materials and Methods Materials. The solid sorbents used and their manufacturers and properties are listed in Table 1. All three sorbents were washed successively with water, methanol, aqueous NaOH, and aqueous HCl and then were dried to constant weight in a vacuum oven at 60 °C and 60 kPa. Reillex 425 was prewetted by stirring in methanol and then displacing the methanol with water. The other two sorbents were initially in the dry form before use. Alamine 336 (Henkel Corp.), an aliphatic, straightchain tertiary amine, was used as received from the manufacturer. The alkyl groups are mainly C8 and C10, with the C8 chain predominating about 2 to 1 (Henkel Corp., 1984). Alamine 336 is a pale yellow liquid, with a specific gravity of 0.81 g/cm3 at 25 °C. A molecular weight of 406 (Tamada et al., 1990) was used for Alamine 336 in this work. Alamine 336 was used at a concentration of 0.300 mol/L (about 15% v/v) in a diluent composed of varying proportions of 1-octanol and dodecane. Dodecane (Aldrich) and 1-octanol (Aldrich) were used as received. D-(+)-Glucose monohydrate, USP (Fluka Chemie AG, Buchs), was used as received. Lactic acid (Aldrich, 85% in water, ACS reagent grade) was diluted with water to approximately 15% (w/w) and boiled under total reflux for 12 h to hydrolyze dimers or multimers present in the concentrated solution. All aqueous solutions were prepared from distilled water that had been further purified by passage through a Milli-Q water filtration system (Millipore Corp.). Glucose Sorption Isotherms. Sorption isotherms for glucose onto various sorbents were generated by contacting known weights of sorbents (typically 1.0 g, dry basis) and glucose solutions of known concentration (typically 10 g) in 20 mL scintillation vials sealed with foil-lined caps. The vials were placed in a temperaturecontrolled shaker bath at an oscillation rate of 120 min-1 for 24 h, which preliminary tests demonstrated to be more than sufficient for equilibration. Equilibration temperatures were 25, 40, and 60 °C for Dowex MWA-1 and Reillex 425 and 25 and 40 °C for Amberlite IRA35, for which the highest allowable operating temperature was found to be less than 60 °C. The pH of the solution was not adjusted, and buffering agents were not used. Initial and final concentrations of glucose were determined by high-performance liquid chromatography, HPLC (Perkin-Elmer Series 10), using a Bio-Rad fast acid analysis column with a differential refractometer

(1)

initially adhering to or absorbed into the sorbent must be included in W0 and Cg,i. Since depletions of glucose in the bulk solution are small, it was important to gain very high precision in the analysis by HPLC. Typical reproducibilities of analysis were within 1%. Competitive Sorption of Lactic Acid and Glucose vs pH. A total of 10 g of aqueous solution containing 4.0% (w/w) lactic acid and 1.0% glucose was contacted with 1 g of sorbent (dry basis) in a sealed vial, as in the glucose sorption experiments. The aqueous solutions were adjusted to various pH values with aqueous NaOH. After equilibration, the pH of the solution was measured with an Orion Model 601A pHmeter equipped with an Orion Ross pH electrode. The acid and glucose concentrations of the aqueous phase were determined by HPLC. The composite uptake of glucose, Qcg, is calculated from eq 1 and that of lactic acid, Qca, by an analogous expression:

Qca ) W0(Ca,i - Ca,f)/m

(2)

Individual Uptake. The total uptakes of solutes and water cannot be determined by simply measuring the changes of solute concentration in the supernatant solution during equilibration. The individual uptakes of lactic acid (Qia), glucose (Qig), and water (Qiw) are defined, following Kipling (1965), as:

Qiw )

Qia )

W0Ca,i - WfCa,f m

(3)

Qig )

W0Cg,i - WfCg,f m

(4)

W0(1 - Ca,i - Cg,i) - Wf(1 - Ca,f - Cg,f) (5) m

The individual uptakes give the total amounts of solute or water taken up by the sorbent, including selective uptake at the surface and/or on reactive sites and nonselective uptake due to swelling and pore filling. Calculation of the individual isotherms requires an assumption about the boundary between sorbed and bulk solution. In this work, the sorbed solution is defined as that solution which is retained by the sorbent after centrifugation for 10 min at 2000 rpm in a IEC HN-S2 (Damon/IEC Division) centrifuge. Equations 6 and 7 describe the relationships between the composite and individual uptakes. The individual

Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996 1217

W0(Ca,i - Ca,f)/m ) Qia - (Qia + Qig + Qiw)Ca,f

(6)

W0(Cg,i - Cg,f)/m ) Qig - (Qia + Qig + Qiw)Cg,f

(7)

uptake must always be greater than the composite uptake, and the composite uptake approaches the individual uptake at very low solute concentration and/ or when there is very little water uptake. In this work, composite uptakes were computed directly from measurements of solution concentrations before and after sorption. After centrifugation, the total solution uptake Qig + Qiw or Qia + Qig + Qiw was measured, and individual uptakes were then computed from eqs 6 and 7. Glucose Extraction Equilibria. Aqueous and organic solutions of known glucose concentration were contacted in the same way as described for glucose sorption experiments. Experiments were performed at 25, 40, and 60 °C, using 10 mL each of the aqueousand organic-phase solutions. The aqueous and organic phases were separated by decantation following centrifugation for 8 min at 2000 rpm in an IEC HN-S2 (Damon/IEC Division) centrifuge. Initial and final glucose concentrations in the aqueous phase were determined by HPLC, as described previously. Equilibrium glucose concentrations in the organic phase were calculated by mass balance. Competitive Extraction of Lactic Acid and Glucose vs pH. Ten milliliter aqueous solutions of 0.45 mol/L lactic acid and 0.055 mol/L glucose that were adjusted to various pH values using aqueous NaOH were contacted with an equal volume of 0.30 mol/L Alamine 336 in the appropriate diluent in 40 mL vials. After equilibration, the phases were separated by centrifugation. The aqueous-phase pH was measured as described previously. Lactic acid and glucose concentrations in aqueous solutions were determined simultaneously by the HPLC method previously described. Organic-phase acid concentrations were determined by two-phase titration with aqueous NaOH, using phenolphthalein as an indicator. Mass balances for lactic acid closed within 2.0%. Water Contents of Organic Phases. Organicphase water concentrations were determined with a Quintel Model MS-1 automatic Karl Fischer titrator. Changes in phase volumes due to water transfer to the organic phase were taken into account in the calculation of equilibrium distribution ratios. Description of Extraction Equilibrium. Protondonating diluents such as 1-octanol may form a hydrogen bond with the available oxygen on the carboxylic group of the acid-amine complex and thus hamper binding of a second acid onto this complex (Kertes and King, 1986; Tamada et al., 1990). For the range of concentrations in which the experimental work was carried out, it was assumed that only a (1, 1) complex formed for lactic acid with the amine extractant. Transfer of the acid from water into the amine-free 1-octanol diluent was also taken into account. The distribution ratio D of lactic acid between solvent and solution was therefore related to the complexation constant as follows:

D) where

B0K11 + φP 1 + K11C

(8)

Figure 1. Composite sorption isotherms for glucose onto Dowex MWA-1 at different temperatures.

Figure 2. Composite sorption isotherms for glucose onto Reillex 425 at different temperatures.

C ) Ca,w/(1 + 10pH-pKa) Equation 8 ignores simple partitioning of the acid into the amine and dodecane. A least-squares-fitting method was used to determine the apparent equilibrium constant, K11. The minimization parameter used was the sum of the absolute values of (Ca,w,pred - Ca,w,expt)/Ca,w,expt, where the subscript “pred” denotes the value predicted by eq 1 and the subscript “expt” denotes the experimentally determined value. Results and Discussion Glucose Sorption Isotherms. Figures 1-3 show composite isotherms for adsorption of glucose onto the three sorbents at 25, 40, and, for two of the sorbents, 60 °C. Full data underlying these and subsequent sorption results are presented by Dai and King (1995b). Glucose exhibits negative deviations from ideality in aqueous solution, i.e., activity coefficients less than 1 (Miyajima et al., 1983). Correspondingly, addition of

1218

Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996

Figure 3. Composite sorption isotherms for glucose onto Amberlite IRA-35 at different temperatures.

Figure 5. Effect of pH on composite uptakes of glucose and lactic acid by Reillex 425. Ca,i ) 4.0%; Cg,i ) 1.0%; W0/m ) 10.

Figure 4. Effect of pH on composite uptakes of glucose and lactic acid by Dowex MWA-1. Ca,i ) 4.0%; Cg,i ) 1.0%; W0/m ) 10.

Figure 6. Effect of pH on composite uptakes of glucose and lactic acid by Amberlite IRA-35. Ca,i ) 4.0%; Cg,i ) 1.0%; W0/m ) 10.

glucose serves to increase the surface tension of water. Hence, there should be little or negative surface enrichment of glucose if there is no preferential chemical interaction of the sugar with the polymeric sorbent surface. This prediction is borne out by the weak interactions and low capacities exhibited by all three sorbents. Glucose capacities decrease only slightly with increasing temperature, corresponding to a low enthalpy of adsorption for glucose. Competitive Sorption of Lactic Acid and Glucose vs pH. Figures 4-6 show the measured surface excesses of lactic acid and glucose on the three different polymeric sorbents as a function of the pH of the aqueous solution, for a constant feed concentration of the solutes and the constant initial phase ratio (10 g of solution/g of sorbent). The observed effect of pH on adsorption of the weakly ionizable acid is a well-known phenomenon. The results correspond closely to previous results (Getzen et al., 1969; Ward and Getzen, 1970; Mu¨ller et al., 1980; Kuo et al., 1987; Tung and King, 1994), indicating that there is little or no effect of glucose on the sorption of lactic acid.

For typical lactic acid fermentations, the sorbent should demonstrate substantial uptake in the pH 5-6 range (Tung and King, 1994; Dai and King, 1995a). The results in Figures 4-6 show that the uptake in this range is strongly dependent upon the basicity of the sorbent. Reillex 425 gives virtually no selective uptake of lactic acid in the pH 5-6 range. In the same pH range, uptakes are 0.17-0.28 g of lactic acid/g of sorbent for Dowex MWA-1 and 0.29-0.35 g of lactic acid/g of sorbent for Amberlite IRA-35. The uptake data for lactic acid with all three sorbents are well described by 1:1 complexation (Dai and King, 1995b), which gives a functional form equivalent to the Langmuir equation.

Qca KCa,f ) Qm 1 + KCa,f

(9)

The curves in Figures 4-6 are calculated by means of eq 9. It is interesting to note that all three sorbents give small negative surface enrichments of glucose at all

Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996 1219

Figure 8. Individual sorption of water by polymeric sorbents at 25 °C. Ca,i ) 4.0%; Cg,i ) 1.0%; W0/m ) 10.

Figure 7. Individual uptakes of water from aqueous glucose solution by Dowex MWA-1, Reillex 425, and Amberlite IRA-35. Table 2. Surface Depletion of Glucose during Simultaneous Sorption of Lactic Acid and Glucose (25 and 40 °C) temp (°C) 25

40

Dowex MWA-1 pH Qcg (mg/g)

Reillex 425 pH Qcg (mg/g)

Amberlite IRA-35 pH Qcg (mg/g)

2.78 4.88 6.07 7.00 7.47 8.05 8.69 8.79 2.75 3.71 4.64 5.36 6.19 6.73 7.26 8.01

2.15 2.62 3.16 3.76 4.25 5.09 6.04 6.48 2.09 2.79 3.31 3.99 4.68 5.29 5.86 6.14

4.38 4.98 6.08 6.48 7.22 7.65 8.09 8.29 3.49 4.78 5.44 5.80 6.33 6.68 7.29 8.20

-8.1 -8.1 -7.4 -5.3 -4.5 -4.0 -3.6 -3.4 -7.4 -7.2 -8.1 -6.7 -5.5 -4.9 -4.4 -4.3

-3.6 -3.2 -2.8 -2.2 -2.1 -1.4 -1.4 -1.4 -3.7 -3.3 -3.3 -3.1 -2.4 -2.8 -3.4 -2.7

-13.1 -12.9 -13.1 -12.5 -12.4 -11.8 -11.4 -10.3 -15.4 -14.8 -14.3 -14.7 -11.9 -11.0 -9.3 -9.3

values of pH, as shown in Figures 4-6 and reported quantitatively in Table 2. Comparison with Figures 1-3 shows that lactic acid appears to displace glucose that is sorbed in the absence of lactic acid. As a result of the nonenrichment of glucose by adsorptive mechanisms, the uptake of glucose is determined by the amount of uptake of bulk solution, e.g., by swelling and/ or interstitial entrainment. Water Uptake and Swelling. Figure 7 shows measured individual uptakes of water from glucose solutions by the three polymeric sorbents. The sorbents all swell appreciably. The water uptake by Amberlite IRA-35 is substantially higher, and the water uptake of Reillex 425 is somewhat higher than that of Dowex MWA-1. The water capacities of the three sorbents decrease slightly as the concentration of glucose in the

Figure 9. Relationship of glucose and water uptakes by the three polymeric sorbents at 25 and 40 °C. Ca,i ) 4.0%; Cg,i ) 1.0%; W0/m ) 10.

equilibrium solution increases. This is probably the result of competitive uptake between glucose and water and the decrease in the activity of water as the glucose concentration increases. Water uptake increases with increasing operating temperature, reflecting greater relaxation of the polymer chains. Figure 8 presents water uptakes from the solutions of lactic acid and glucose at 25 °C. The increase in water uptake with increasing lactic acid concentration results from increased swelling, attributable to solvation of the amine-carboxylic acid complex. Results for water uptakes at 40 °C (Dai and King, 1995b) are essentially the same as those at 25 °C. As shown in Figure 9 there is a direct correlation between uptakes of glucose and water for all three polymeric sorbents. It thereby appears that water brings glucose into the sorbent proportionately. Selectivity of Competitive Sorption. Figure 10 shows selectivities of total uptakes obtained between lactic acid and glucose at 25 °C. Figure 11 shows selectivities between lactic acid and water at 25 °C,

1220

Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996

Figure 10. Selectivity (proportion of lactic acid to glucose in sorbate) at 25 °C. Ca,i ) 4.0%; Cg,i ) 1.0%; W0/m ) 10.

Figure 12. Extraction equilibria of glucose for 15% (v/v) Alamine 336 with 85% (v/v) 1-octanol.

Figure 11. Inverse selectivity (proportion of water to lactic acid in sorbate) at 25 °C. Ca,i ) 4.0%; Cg,i ) 1.0%; W0/m ) 10.

Figure 13. Extraction equilibria of glucose for 15% (v/v) Alamine 336 with 42.5% (v/v) 1-octanol and 42.5% (v/v) dodecane.

reported as “inverse selectivities”sthe weight ratio of water to lactic acid taken up. In Figures 10 and 11 the abscissa is the pH of the equilibrium aqueous phase. Results at 40 °C (Dai and King, 1995b) are essentially the same as those at 25 °C. It is clear that, among the basic sorbents studied, Dowex MWA-1 gives the best selectivities for uptake of lactic acid, as opposed to water and glucose, in the pH 5-6 range. Although Amberlite IRA-35 gives higher uptakes of lactic acid in the middle pH range, the marked swelling of that sorbent leads to substantial uptakes of water and glucose and decreases the selectivities for lactic acid over water and glucose. The uptake and selectivity for lactic acid with Reillex 425 are much lower than for Dowex MWA-1 or Amberlite IRA-35 in the pH 5-6 range. Glucose Extraction Equilibria. Figures 12-14 report equilibria for the water-glucose-Alamine 336/ diluent system, at 25, 40, and 60 °C with different diluent compositions. Nearly linear equilibrium relationships and low distribution ratios are obtained. As the content of active diluent (1-octanol) and the tem-

perature increase, the extraction of glucose increases. Full data underlying these and subsequent extraction results are reported by Dai and King (1995c). Competitive Extraction of Lactic Acid and Glucose vs pH. Capacities for extraction of lactic acid by Alamine 336 in diluents of three different compositions are presented in Figure 15 as functions of equilibrium aqueous-phase pH. The results are reported in terms of the stoichiometric loading (mol acid/mol extractant). Since the extraction equilibria involve the concentration of undissociated acid in the aqueous phase, the extraction capacity for a carboxylic acid decreases with increasing pH. As has been shown previously (Kertes and King, 1986; Tamada et al., 1990), a greater proportion of 1-octanol in the diluent increases the distribution of lactic acid to the solvent phase, because of the ability of the protic 1-octanol to associate with the acid-amine complex. As has also been found previously (Wennersten, 1983; Tamada and King, 1990), the loadings of Alamine 336 with lactic acid in all three diluents decrease with increasing operating temperature. Using the experimental data in this work and appropriate models (Tung, 1993), the apparent equilibrium

Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996 1221 Table 3. Apparent Complexation Equilibrium Constant K11 and Measured Partition Coefficient P for Lactic Acid K11 (L/mol) solvent

25 °C

15% Alamine 336 with 85% 1-octanol 189; 15% Alamine 336 with 42.5% 1-octanol 29.5 and 42.5% dodecane 15% Alamine 336 with 8.5% 1-octanol 1.96 and 76.5% dodecane

191a

40 °C 60 °C 99 21.1 1.54

44.5 14.9 1.09

P

a

solvent

25 °C

40 °C

60 °C

1-octanol

0.205

0.215

0.228

Fitting results from Tung (1993).

Figure 14. Extraction equilibria of glucose for 15% (v/v) Alamine 336 with 8.5% (v/v) 1-octanol and 76.5% (v/v) dodecane.

Figure 16. Effect of pH on extraction of glucose by 15% (v/v) Alamine 336 with 85% (v/v) 1-octanol. Ca,i ) 0.45 M; Cg,i ) 0.055 M; phase ratio ) 1.0.

Figure 15. Effect of pH on extraction of lactic acid by 15% (v/v) Alamine 336 with three different diluent compositions. Ca,i ) 0.45 M; Cg,i ) 0.055 M; phase ratio ) 1.0.

constants, K11, were obtained. The results are listed in Table 3 and are similar to those obtained by Tung and King (1994) for single-component extractions at the same operating conditions. The partition coefficients of lactic acid from water into otherwise pure 1-octanol at different temperatures were also measured and are shown in Table 3. Measured extraction capacities for glucose with the same solvent mixtures are shown in Figures 16-18, where it is seen that the uptakes of glucose are essentially independent of pH. All three solvent mixtures

Figure 17. Effect of pH on extraction of glucose by 15% (v/v) Alamine 336 with 42.5% (v/v) 1-octanol and 42.5% (v/v) dodecane. Ca,i ) 0.45 M; Cg,i ) 0.055 M; phase ratio ) 1.0.

provide very low uptake capacities of glucose. Distribution ratios of glucose are somewhat lower than those for extraction of glucose alone. This may again reflect competition with lactic acid. Water Coextraction. Measurements of water co-

1222

Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996

Figure 18. Effect of pH on extraction of glucose by 15% (v/v) Alamine 336 with 8.5% (v/v) 1-octanol and 76.5% (v/v) dodecane. Ca,i ) 0.45 M; Cg,i ) 0.055 M; phase ratio ) 1.0.

Figure 20. Relationship of uptake for glucose to uptake for water into three different solvent mixtures at 25 and 60 °C. Ca,i ) 0.45 M; Cg,i ) 0.055 M; phase ratio ) 1.0.

Figure 19. Water coextraction for extraction of lactic acid by 15% (v/v) Alamine 336 in various diluents at 40 °C. Ca,i ) 0.45 M; Cg,i ) 0.055 M; phase ratio ) 1.0.

Figure 21. Separation factor between lactic acid and glucose for 15% (v/v) Alamine 336 in various diluents at 40 °C. Ca,i ) 0.055 M; phase ratio ) 1.0.

extraction (Dai and King, 1995c) showed that the coextraction of water was not affected by the presence of glucose and increased slightly with increasing temperature. Figure 19 shows the water coextraction measured in the experiments examining competitive extraction between lactic acid and glucose at 40 °C. Increasing uptake of lactic acid increases the amount of water coextracted. This result suggests that water interacts with lactic acid, or more likely the acid-amine complex, in the organic phase and thereby accompanies the extracted acid into the organic phase. The amount of water coextracted with lactic acid into the solvent phases increases with an increase in the proportion of 1-octanol in the diluent. Over the temperature range 25-60 °C, the amount of water coextracted with lactic acid was insensitive to temperature (Dai and King, 1995c). As is shown in Figure 20, the extent of water coextraction depends strongly upon the diluent used and the operating temperature. Coextracted water becomes greater at higher operating temperature and for dilu-

ents having large proportions of 1-octanol, and the coextracted water facilitates accommodation of glucose into the organic phase. But for a given diluent, additional water coextraction caused by the increase in uptake of lactic acid at lower pH does not significantly increase the uptake of glucose (Dai and King, 1995c). Separation Factor for Competitive Extraction. The separation factor between lactic acid and glucose, Ra,g, is expressed as

Ra,g )

Ca,o Cg,w Ca,w Cg,o

(10)

Figure 21, based upon data reported in earlier figures, shows the separation factor between lactic acid and glucose for extraction by Alamine 336 in various diluents at 40 °C. The uptake capacity of lactic acid has the most important influence on the separation factor between lactic acid and glucose. The separation factor,

Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996 1223 Table 4. Comparison of Extraction and Sorption for Selective Separation between Lactic Acid and Glucose (40 °C) uptake (mg/g of fresh solvent) solvent

pH

lactic acid

glucose

water

15% Alamine 336 in 1-octanol

2.97 3.88 4.76 5.34 5.95 6.44 6.67 7.35 2.75 3.71 4.64 5.36 6.19 6.73 7.26 8.01

38.94 35.09 27.81 19.12 8.59 3.43 2.12 0.47 359 351 318 269 185 117 62.4 42.3

0.054 0.052 0.051 0.053 0.053 0.052 0.053 0.053 12.0 11.9 10.4 10.3 8.6 6.5 5.0 4.0

41.54 40.92 36.32 29.47 22.89 20.20 18.68 18.42 1440 1421 1391 1316 1141 959 826 746

Dowex MWA-1

therefore, decreases with increasing pH and with decreasing ratio of 1-octanol to dodecane in the diluent mixture. Comparison of Extraction and Sorption. The performances of extraction and sorption for selective separation between lactic acid and glucose are compared in Table 4. The initial concentrations of lactic acid and glucose for these data were 4.0% and 1.0% (w/w), respectively. The weight ratio of aqueous solution to fresh solvent or dry sorbent was equal to 10. The operating temperature was 40 °C. The solvent used was 15% (v/v) Alamine 336 with 85% (v/v) 1-octanol. The sorbent used was Dowex MWA-1. The uptake of lactic acid for extraction was calculated from the fitting parameters. The densities of organic and aqueous phases are assumed to be constant, and the effect of phase ratio on uptakes of glucose and water has been ignored. The results show that the uptake of lactic acid for sorption by Dowex MWA-1 is much higher than that for extraction by Alamine 336. This advantage is more significant in the higher pH range. On the other hand, the uptake capacity for water is much higher for sorption than for extraction, and this results in a substantially higher uptake of glucose onto the sorbent. Marked swelling of the sorbent in the solution reduces the selectivities for acid over water and glucose. Acknowledgment This research was supported by the Biological and Chemical Technology Research Program, Advanced Industrial Concepts Division, Office of Industrial Technologies, U.S. Department of Energy. Y.D. received a financially supported study leave from Tsinghua University, Beijing, China. Nomenclature B0 ) initial amine concentration (mol/L) Ca,f ) weight fraction of lactic acid in the final solution Ca,i ) concentration of lactic acid in the initial solution (weight fraction for sorption; mol/L for extraction) Ca,o ) lactic acid concentration in the organic phase (mol/ L) Ca,w ) lactic acid concentration in the aqueous phase (mol/ L) Cg,f ) weight fraction of glucose acid in the final solution

Cg,i ) concentration of glucose in the initial solution (weight fraction for sorption; mol/L for extraction) Cg,o ) glucose concentration in the organic phase (mol/L) Cg,w ) glucose concentration in the aqueous phase (mol/L) Cw,o ) water concentration in the organic phase (mol/L) Cw,w ) water concentration in the aqueous phase (mol/L) D ) distribution ratio of lactic acid K ) complexation constant for sorption (g of resin/g of lactic acid) Ka ) dissociation constant of lactic acid (mol/L) K11 ) apparent equilibrium constant for the (1,1) complex (L/mol) m ) mass of dry sorbent (g) P ) partition coefficient Qca ) composite uptake of lactic acid (g of lactic acid/g of sorbent) Qcg ) composite uptake of glucose (g of glucose/g of sorbent) Qia ) individual uptake of lactic acid (g of lactic acid/g of sorbent) Qig ) individual uptake of glucose (g of glucose acid/g of sorbent) Qiw ) individual uptake of water (g of water/g of sorbent) Qm ) maximum value of sorbed lactic acid at high concentration (g of lactic acid/g of sorbent) W0 ) initial mass of the bulk solution (g) Wf ) final mass of the bulk solution (g) Greek Letters Ra,g ) separation factor between lactic acid and glucose φ ) volume fraction of 1-octanol in the solvent mixture

Literature Cited Dai, Y.; King, C. J. Modeling of Fermentation with Continuous Lactic Acid Removal by Extraction Utilizing Reversible Chemical Complexation. Report No. LBL-37485; Lawrence Berkeley Laboratory: Berkeley, CA, July 1995a. Dai, Y.; King, C. J. Selective Separation of Lactic Acid and Glucose. I. Sorption by Basic Polymer Sorbents. Report No. LBL-37837; Lawrence Berkeley Laboratory: Berkeley, CA, Sept 1995b. Dai, Y.; King, C. J. Selective Separation of Lactic Acid and Glucose. II. Extraction by Alamine 336 in Various Diluents. Report No. LBL-37838; Lawrence Berkeley Laboratory: Berkeley, CA, Sept 1995c. Davison, B. H.; Thompson, J. E. Simultaneous Fermentation and Separation of Lactic Acid in a Biparticle Fluidized-Bed Bioreactor. Appl. Biochem. Biotechnol. 1992, 34/35, 431-439. Dow Chemical Co. DOWEX Ion Exchange Resins Engineers’ Handbook; Midland, MI, 1983. Evangelista, R. L.; Mangold, A. J.; Nikolov, Z. L. Recovery of Lactic Acid by Sorption. Appl. Biochem. Biotechnol. 1994, 45/46, 131144. Getzen, F. W.; Ward, T. M. A Model for the Adsorption of Weak Electrolytes on Solids as a Function of pH. 1. Carboxylic AcidCharcoal Systems. J. Colloid Interface Sci. 1969, 31, 441-453. Gustafson, R. L.; Filius, H. F.; Kunin, R. Basicities of Weak Base Ion Exchange Resins. Ind. Eng. Chem. Fundam. 1970, 9, 221229. Henkel Corp. Alamine 336; Red Line Technical Bulletin; Henkel Technical Center: Minneapolis, MN, 1984. Jain, M. K.; Datta, R.; Zeikus, J. G. High-Value Organic Acids Fermentation-Emerging Processes and Products. In Bioprocess Engineering: The First Generation; Ghose, T. K., Ed.; Ellis Horwood Limited: Chichester, U.K., 1989; pp 366-389. Kaufman, E. N.; Cooper, S. P.; Davison, B. H. Screening of Resins for Use in a Biparticle Fluidized-Bed Bioreactor for the Continuous Fermentation and Separation of Lactic Acid. Appl. Biochem. Biotechnol. 1994, 45/46, 545-554. Kaufman, E. N.; Cooper, S. P.; Clement, S. L.; Little, M. H. Use of a Biparticle Fluidized-Bed Bioreactor for the Continuous and Simultaneous Fermentation and Purification of Lactic Acid. Appl. Biochem. Biotechnol. 1995, 51/52, 605-620. Kertes, A. S.; King, C. J. Extraction Chemistry of Fermentation Product Carboxylic Acids. Biotechnol. Bioeng. 1986, 28, 269282.

1224

Ind. Eng. Chem. Res., Vol. 35, No. 4, 1996

King, C. J. Separation Processes Based on Reversible Chemical Complexation. In Handbook of Separation Process Technology; Rousseau, R. W., Ed.; Wiley: New York, 1987; pp 760-764. Kipling, J. J. Adsorption from Solutions on Non-electrolytes; Academic Press: New York, 1965; pp 28-42. Kuo, Y.; Munson, C. L.; Rixey, W. G.; Garcia, A. A.; Frierman, M.; King, C. J. Use of Adsorbents for Recovery of Acetic Acid from Aqueous Solution. Part 1. Factors Governing Capacity. Sep. Purif. Methods 1987, 16 (1), 31-64. Miyajima, K.; Sawada, M.; Nakagaki, M. Studies on Aqueous Solutions of Saccharides. I. Activity Coefficients of Monosaccharides in Aqueous Solutions at 25 °C. Bull. Chem. Soc. Jpn. 1983, 56, 1620-1623. Mu¨ller, G.; Radke, C. J.; Prausnitz, J. M. Adsorption of Weak Organic Electrolytes from Aqueous Solution on Activated Carbon. Effect of pH. J. Phys. Chem. 1980, 84, 369-376. Ng, T. K.; Busche, R. M.; McDonald, K. C.; Hardy, R. W. F. Production of Feedstock Chemicals. Science 1983, 219, 733740. Reilly Industries, Inc. Reillex: A New Family of Crosslinked Polyvinylpyridines from Reilly, Indianapolis, IN, 1989. Rohm & Haas Co. Amberlite/Duolite Ion Exchange Resins: Technical Notes; Philadelphia, PA, 1981. San-Martin, M.; Pazos, C.; Coca, J. Reactive Extraction of Lactic Acid with Alamine 336 in the Presence of Salts and Lactose. J. Chem. Technol. Biotechnol. 1992, 54, 1-6. Scho¨ller, C.; Chaudhuri, J. B.; Pyle, D. L. Emulsion Liquid Membrane Extraction of Lactic Acid from Aqueous Solutions and Fermentation Broth. Biotechnol. Bioeng. 1993, 42, 50-58. Tamada, J. A.; King C. J. Extraction of Carboxylic Acids with Amine Extractants. 3. Effect of Temperature, Water Coextraction, and Process Considerations. Ind. Eng. Chem. Res. 1990, 29, 1333-1338. Tamada, J. A.; Kertes, A. S.; King C. J. Extraction of Carboxylic Acids with Amine Extractants. 1. Equilibria and Law of Mass Action Modeling. Ind. Eng. Chem. Res. 1990, 29, 1319-1326.

Tung, L. A. Recovery of Carboxylic Acids at pH greater than pKa. Report No. LBL-34669; Lawrence Berkeley Laboratory: Berkeley, CA, Aug 1993. Tung, L. A.; King, C. J. Sorption and Extraction of Lactic and Succinic Acids at pH > pKa1. I. Factors Governing Equilibria. Ind. Eng. Chem. Res. 1994, 33, 3217-3223. Vickroy, T. B. Lactic Acid. In Comprehensive Biotechnology; Blanch, H. W., Drew, S., Wang, D. I. C., Eds.; Pergamon Press: New York, 1985; Vol. 3, pp 761-776. Wang, C. J.; Bajpai, R. K.; Iannotti, E. L. Nondispersive Extraction for Recovery Lactic Acid. Appl. Biochem. Biotechnol. 1991, 28/ 29, 589-603. Ward, T. M.; Getzen, F. W. Influence of Aromatic Acids on Activated Carbon. Environ. Sci. Technol. 1970, 4, 64-67. Wennersten, R. The Extraction of Citric Acid from Fermentation Broth Using a Solution of a Tertiary Amine. J. Chem. Technol. Biotechnol. 1983, 33B, 85-94. Yabannavar, V. M.; Wang, D. I. C. Bioreactor System with Solvent Extraction for Organic Acid Production. Ann. N.Y. Acad. Sci. 1987, 506, 523-535. Yabannavar, V. M.; Wang, D. I. C. Extractive Fermentation for Lactic Acid Production. Biotechnol. Bioeng. 1991, 37, 10951100. Yang, S. T.; White, S. A.; Haso, S. T. Extraction of Carboxylic Acids with Tertiary and Quaternary Amines: Effect of pH. Ind. Eng. Chem. Res. 1991, 30, 1335-1342.

Received for review October 13, 1995 Revised manuscript received February 27, 1996 Accepted February 27, 1996X IE9506274

X Abstract published in Advance ACS Abstracts, March 15, 1996.