Ind. Eng. Chem. Res. 1990,29, 1333-1338 Yerger, E. A.; Barrow, G. M. Acid-Base Reactions in Non-dissociating Solvents. Acetic Acid and Diethylamine in Carbon Tetrachloride and Chloroform. J. Am. Chem. SOC.1955a, 77, 4474-4481. Yerger, E. A,; Barrow, G . M. Acid-Base Reactions in Non-dissociating Solvents n-Butylamine and Acetic Acid in Carbon Tetra-
chloride and Chloroform. 6206-6207.
1333 J. Am. Chem. SOC. 1955b, 77,
Received for review August 30, 1989 Revised manuscript received February 12, 1990 Accepted February 21, 1990
Extraction of Carboxylic Acids with Amine Extractants. 3. Effect of Temperature, Water Coextraction, and Process Considerations Janet A. Tamadat and C. Judson King* Department of Chemical Engineering and Lawrence Berkeley Laboratory, University of California, Berkeley, California 94720
Coextraction of water during extraction of succinic acid by Alamine 336 in different diluents has been measured. The amounts of coextracted water lie in the same order as the solubilities of water in the diluents without amine present. Water coextraction with different acids follows the order fumaric > malonic > maleic = succinic > lactic > acetic. The effects of temperature on extraction of succinic and lactic acids by Alamine 336 with chloroform and methyl isobutyl ketone (MIBK) diluents have been measured. Enthalpies and entropies of complex formation have been derived from the results and are interpreted in terms of the differences in interactions among the species involved. Two approaches for regeneration through back-extraction into an aqueous phase are considered. These involve changes in the equilibrium relationship through a swing of temperature and a swing of diluent composition, respectively. The factors underlying the utility of each are explored and contrasted. The two approaches may be used in combination. In the production of carboxylic acids, processes such as fermentations produce multicomponent, aqueous solutions with product acid concentrations typically 10% w/w at most, and usually substantially less. Subsequent separation, purification, and concentration of these acids is difficult because of the high affmities of the acids for water. Distillation of dilute, nonvolatile acids involves large energy consumption for the heat of vaporization of water, which must be taken overhead. Furthermore, distillation cannot fractionate among nonvolatile acids. The low aqueous activity of carboxylic acids results in low distribution coefficients of acids into conventional solvents. Thus, solvent extraction with conventional solvents would require very high solvent flow rates and result in substantial dilution of the acid. Long-chain tertiary amines are effective extractants for carboxylic acids (Kertes and King, 1986). The strong interaction between the acid and amine allows for formation of acid-amine complexes and thus provides for high equilibrium distribution ratios. Additionally, the high affinity of the organic base for the acid gives selectivity for the acid over other nonacidic components in the mixture. This reaction is reversible, enabling recovery of the acid and recycle of the solvent. There are two steps to a practical extractive separation and recovery process. The first is extraction of the acid to produce an acid-loaded extract and a relatively acid-free aqueous raffnate. The second step transfers the acid from the solvent into a product phase and regenerates the extractant mixture which is recycled back to the extractor. Regeneration is a crucial step for development of a practical operation.
* T o whom correspondence should be addressed. Department of Chemical Engineering, E25-342, Messachusetts Institute of Technology, Cambridge, MA 02139. t Current address:
0888-5885/90/2629-1333$02.50/0
Water coextraction, Le., water that enters the organic phase with the solute, may also affect process economics. For example, it may be necessary to recover pure acid from an aqueous solution produced from the extract during regeneration. This work examines coextraction of water with carboxylic acid-amine complexes and two means of regeneration using back-extraction-temperature swing and diluent swing. Some of the chemistry involved in the effect of temperature on extraction is also discussed. McCabe-Thiele Diagram. A McCabe-Thiele operating diagram for dilute solutions (Figure 1)outlines the general form of an extractionlback-extraction acid recovery process (King, 1980). During the extraction step, the equilibrium curve must lie above the operating line. A low solvent flow rate corresponds to a relatively large slope of the operating line. Thus, high equilibrium distribution of acid into the solvent phase is necessary in order to achieve good product recovery at reasonable solvent flow rates. During regeneration, the situation is reversed; the equilibrium curve must lie below the operating line. The slope of the regenerator operating line should be lower than that for extraction, corresponding to low water flow rates, if the product is to be concentrated overall with substantial solute recovery. To achieve this, the equilibrium distribution of the acid into the organic phase must be low. Therefore, it is important to effect a downward shift in the equilibrium line between the extraction and regeneration stages. Such processes a r e known as "swing" processes.
Experimental Section Batch experiments were performed for the extraction of succinic and lactic acids by Alamine 336 (Henkel Corp.) in chloroform and methyl isobutyl ketone (MIBK) as diluents at various temperatures as described in part 1 (see also Tamada and King (1989)). Extraction experiments without amine were also performed at various tempera0 1990 American Chemical Society
1334 Ind. Eng. Chem. Res., Vol. 29, No. 7, 1990
-I 2 06 E
I
F P R
= Feed
-
Product
= Raflinate
,
-+-
'1
- 0-c--
-
I
Nitrobenzene Chloroform -4. 1Ocranol W
,.'
D8chloromethane
.
MIBK
;
= Wale! w -
is = Loaded __ Solvent RS
= Regenerated
Solvent
w 0 4 ,
I
I
Acid in Organic Phase (mol/Ll
Figure 2. Coextraction of water with succinic acid and 0.29 mol/L Alamine 336 in various diluents. The water concentration in the organic phase has been adjusted by subtracting the amount of water soluble in the solvent alone (Tamada and King, 1989) from the total coextracted water. 0 $ 0 2 1
Regenerator
ooy-
0 00
I
0 04
I
0 08
0 12
1 I
0 16
Solute in Aqueous Phase
Figure 1. Extraction/regeneration process using back-extractive regeneration. (a, top) Process flow diagram. (b, bottom) McCabeThiele diagram.
tures with MIBK, for which extraction by the diluent alone is significant. The temperature of the shaker bath was maintained within f 2 "C of the setpoint. The product phases were analyzed for equilibrium acid concentrations as described in part 1. Batch experiments to determine the amount of water coextracted with the acid were performed for a variety of carboxylic acid-Alamine 336-diluent systems. The water content of the product organic phase was measured by Karl Fischer titration, except in the case of MIBK, where water analyses were made by gas chromatography (Varian Model 3700) with thermal conductivity detection, using a 1.0-m Porapak Q (Waters Assoc.) column. Organic-phase acid concentrations were corrected for extraction by the diluent alone a t the appropriate temperature, as described in part 1. Best-fit equilibrium constants for the complexes formed were determined as described in part 1. Water Coextraction. Chemistry of Water Coextraction. Figure 2 compares the amount of water coextracted with succinic acid and 0.29 mol/L Alamine 336 in the various diluents. Water coextraction decreases in the order 1-octanol > MIBK > nitrobenzene > methylene chloride > chloroform > heptane (not shown), which is the same order as the solubility of water in the diluent alone (Tamada and King, 1989). Apparently, forces that allow the diluent to solvate water molecules effectively also cause solvation of the water molecules surrounding or attached to a complex. Additionally, studies were performed to compare the coextraction of water, which accompanies the extraction of various other acids, by Alamine 336 in chloroform, MIBK, and various alcohols (Tamada and King, 1989). Water coextraction at low acid concentrations decreases in the acid order fumaric > malonic > maleic = succinic > lactic > acetic. Thus, it is seen that monocarboxylic acids cany less water with them than do dicarboxylic acids, which may reflect the tendency of coextracted water molecules to associate with the carboxylate group. Fumaric acid may then give more water coextraction than other
acids because of the lack of internal self-association (part 2). The effect of temperature on water coextraction was also studied (Tamada and King, 1989). For the extraction of lactic and succinic acids by Alamine 336 in chloroform, water coextraction increases with increasing temperature. For the extraction of succinic acid by Alamine 336 in MIBK, there does not appear to be an effect of temperature on water coextraction. In general, selectivity of the acid over water in extraction by amine extractants is high, relative to the results with conventional solvents. The water carried into the extract would be minimal compared to the amount of water used in an aqueous back-extraction and therefore has little effect upon process viability. Effect of Temperature on Extraction of Carboxylic Acids. Baniel et al. (1981) report the effect of temperature on the extraction of citric acid by tridecylamine in petroleum fractions with an alcohol modifier, in xylene, and in nitrobenzene. The distribution of acid into the solvent phase decreases sharply with increasing temperature, sufficiently to allow back-extraction of the acid from the solvent in a fresh aqueous phase without overall dilution of the acid. Wennersten (1983) studied the extraction of citric acid by Alamine 336 in a variety of diluents a t 25 and 60 "C. Detailed data were obtained for Isopar H (a paraffinic kerosene), a 1:l (v/v) mixture of Isopar H and MIBK, and n-butyl chloride a t 25 and 60 "C. For the corresponding 35 "C increase in temperature, distribution ratios decreased by as much as a factor of 6. Sato et al. (1985) examined the extraction of lactic, tartaric, succinic, and citric acids by trilaurylamine in xylene at 20, 30,40, and 50 "C. For the 30 "C temperature increase, distribution ratios decreased by factors between 2 and 10, depending upon the type of acid and its concentration. Apparent Enthalpy and Entropy of Complexation. The complexation reactions in the organic phase involve proton transfer or hydrogen-bond formation (part 2) and are expected to be exothermic. Formation of a complex makes the system more ordered and thus decreases the entropy. Therefore, as the temperature is increased, the amount of acid extracted decreases. If the enthalpy and entropy of reaction are assumed to be constant over the temperature range, the expression -AH lnK=RT
+ -AS R
indicates that a plot of In K vs 1/ T gives a straight line. The slope is proportional to the enthalpy of reaction, and the intercept is proportional to the entropy.
Ind. Eng. Chem. Res., Vol. 29, No. 7, 1990 1335 Table I. Effect on Temperature on Partition and Dimerization Constants for Lactic and Succinic Acids in MIBK
C A , =~P[A] ~ ~ + 2&P2[AI2 acid lactic
succinic
T, "C 0 25 50 75 0 25 50 75
P
Kd
0.09 0.11 0.13 0.14 0.28 0.19 0.16 0.15
0.75 0.56 0.43 0.45
Standard State. For a reaction carried out a t temperature T and 1 atm of pressure, the overall change of state may be divided into (a) cooling (or heating) of reactants (A and AP1B) in solution a t their equilibrium concentrations from temperature T to 25 "C, (b) "unmixing" reactants at 25 "C from their equilibrium concentrations in solution to pure components, (c) reaction a t 25 "C of the pure reactants to form pure product, A,B, (d) dilution a t 25 "C of pure product to equilibrium concentration in solution, and (e) heating (or cooling) products in solution from 25 "C to temperature T. If the standard state is taken as pure liquid components a t 25 "C and 1 atm of pressure, transition c is the standard-state reaction, and hence, the association constant for this reaction is the true thermodynamic constant, K Transitions b and d reflect deviations of Kpl from:;fI due to the nonidealities of mixing of the species in the aqueous and organic phases. These effects were discussed in part 2. Transitions a and e account for the effect of temperature on the activity coefficient, which will cause deviations of the apparent AH and AS from the standard-state values. This deviation can be theoretically calculated from d In T / d ( l / T ) = Ahmi, (2) (Smith and Van Ness, 1975), where Ahmix is the partial molar heat of mixing. Experimental Results Extraction by MIBK Alone. Carboxylic acids dimerize readily in organic solvents. Table I gives the monomer partition coefficients and dimerization constants for the extraction of succinic and lactic acids by MIBK without the presence of amine, determined by the best fit to experimental data (Tamada and King, 1989). No dimerization was apparent for succinic acid at the concentrations studied. The acids exhibit contrasting behavior; as the temperature increases, extraction of succinic acid decreases, while extraction for lactic acid increases. The dimerization constant of lactic acid decreases with increasing temperature. Extraction by Tertiary Amine in Diluent. Figure 3 shows the experimental results for the extraction of succinic acid by Alamine 336 in (a) MIBK (0-75 " C ) and (b) chloroform (0-55 "C). Similar data were obtained for the extraction of lactic acid (Tamada and King, 1989). Extraction decreases with increasing temperature. Table I1 summarizes the association constants calculated by the best-fit modeling analysis described in part 1. (Logarithms are to the base 10.) Figure 4, a and b, shows log K vs 1/T for succinic and lactic acids. Tables I11 and IV show values for the apparent AH and AS for stepwise complex formation of (l,l), (2,1), and (3,l) complexes, calculated from a linear leastsquares fit of the data in Table I1 and Figure 4. The
"
0 0
0 2
0 1
0 3
Acid in Aqueous Phase (mol/L)
- 0 3
2 -E I
g o 2
f
z
E 0 1
n
9
0 0
0 04
0 0
0 08
0 12
0 16
Acid in Aqueous Phase (mol/L]
Figure 3. Effect of temperature on the extraction of succinic acid by 0.29 mol/L Alamine 336. (a, top) MIBK diluent. (b, bottom) Chloroform diluent. Table 11. Effect of Temperature on the Stepwise Formation Constants for Complexes of Lactic and Succinic Acids with 0.29 mol/L Alamine 336 in Chloroform and MIBK system T, "C log Kll log Kzl log Kgl 0 1.68 0.07 -0.07 lactic acid Alamine 336 25 1.31 0.21 -0.35 MIBK 50 1.12 -0.04 -0.26 75
lactic acid Alamine 336 chloroform succinic acid Alamine 336 MIBK succinic acid Alamine 336 chloroform
0 25 55
0 25 50 75
0 25 40 55
0.67 3.37 2.57 1.74 2.06 1.35 0.85 0.46 3.60 2.45 1.85 1.40
0.08 -0.62 -0.46 -0.36 0.86 0.75 0.47 0.09
-0.15
Table 111. Apparent Enthalpies of Complex Formation with Alamine 336O AH, kcal/mol system (1.1) (2.1) (3.1) AH transfer Lactic Acid MIBK -5.6 f 0.4 -0.4 f 0.9 -0.4 f 1.1 1.1f 0.1 chloroform -12.1 f 0.2 +1.9 f 0.3 Succinic Acid MIBK chloroform af
-9.2 f 0.4 -16.5 f 0.4
-4.4 f 1.0
-1.7 f 0.3
means values are standard error of estimate.
apparent enthalpies of association are more exothermic for succinic acid than for lactic acid and more exothermic in chloroform than in MIBK. The apparent entropy decrease is greater for succinic acid than for lactic acid and greater in chloroform than in MIBK.
1336 Ind. Eng. Chem. Res., Vol. 29, No. 7 , 1990
28
3 0
3 2
3 4
3 8
38
l / T x Id (K)
, Ad-
0 1 :
1
,;
- -, - - - -1k -,- - - - - -
-1
28
3 0
3 2
3 4
3 6
3 6
l / T x 18 (%) Figure 4. Determination of the apparent heats and entropies of reaction for the extraction of (a, top) succinic acid and (b, bottom) lactic acid by Alamine 336 in MIBK and chloroform. Lines correspond to values of the apparent enthalpies and entropies of stepwise formation of (1,1),@,I), and (3,l)complexes given in Tables I11 and IV.
Table IV. Apparent Entropies of Complex Formation with Alamine 336" AS, cal/(mol K) (2,l) (3J) Lactic Acid MIBK -12.6 f 0.4 -0.9 f 0.5 -2.09 f 0.7 chloroform -29.0 f 0.1 +4.2 f 0.1
system
AS transfer
(1,U
Succinic Acid MIBK -24.6 f 0.2 -11.9 f 0.6 chloroform -44.2i 0.2
0.7 f 0.1
-8.7 i 0.2
f means values are the standard errors of estimate.
Comparison of (1,l) with (2,l)and (3,l) Complexation. For the systems studied, (1,1)complexationis much more exothermic and involves a much greater loss of entropy than the formation of (2,l) or (3,l) complexes. This is reasonable when related to the findings in part 2, in which it was concluded that (1,l)complexation involves the formation of an ion pair, but higher complexes involve hydrogen-bond formation. Effect of Diluent. The relatively large differences in enthalpy and entropy loss between the two diluents for (1,l) complexation are consistent with the conclusion (part 2) that interaction of chloroform with the complex is specific hydrogen bonding. The association of chloroform with the complex is exothermic and increases the order (decreases the entropy) of the system. A comparison of the thermodynamic quantities for (2,l) complex formation of lactic acid with the amine in chloroform and MIBK reveals a contrast to the (1,l)behavior. In chloroform, AHzlis a small, positive quantity, indicating an endothermic reaction, and ASzl is positive, indicating an increase in system entropy upon the addition of the second acid to the (1,l)complex. This is consistent with
the hypothesis that chloroform orders itself around the (1,l) complex. Addition of the second acid disrupts the chloroform-complex interaction, which requires energy and increases the overall randomness of the system. Effect of Acid. The results in Table I11 indicate that (1,l) complexation is more exothermic for succinic acid than for lactic acid. Likewise, the entropy loss is greater for succinic acid. Either the partial molar heat of mixing of the complex in the solvent, the partial molar heat of mixing of the acid in the aqueous phase, or the AH+- must differ between lactic and succinic acids. The partial molar heats of mixing of acids in the aqueous phase have been estimated to be -0.65 kcal/mol for dilute succinic acid solutions (Apelblat, 1986) and -0.94 kcal/mol for lactic acid (Saville and Gundry, 1959; Holten, 1971). Although these figures are approximate, such a small difference suggests that this factor is not a significant contribution to the variation in the heat of complexation with acid for the concentrations studied in this work. The effects of temperature on the partition coefficients, P, for extraction of the acids by the diluent alone may yield information on the organic-phase heat of mixing. Since P = cAaq/cAorg,it follows that d In P / d ( l / T ) = Hhq - HA^^^ = -AHtrms (3) where AHtr,, is the heat of transfer from the organic to the aqueous phase. Table I11 shows values for AHbw calculated from the distribution of succinic acid and lactic acid into MIBK alone. Values for the heats of transfer of the acid from the aqueous phase to a organic phase of MIBK for lactic and succinic acids are +1.1and -1.7 kcal/mol, respectively. The 2.8 kcal/mol difference is a substantial portion of the 3.6 kcal/mol difference in AHll for the two acids in MIBK. Likewise, the difference in the entropy of transfer from the aqueous to organic phases is 7 cal/ (mol K) (Tamada and King, 1989), a substantial fraction of the 12 cal/(mol K) difference in AS,,. Thus, it is possible that the heat of mixing of the acid in the diluent is related to the heat of mixing of the complex in the solvent. Analogous calculations cannot be performed in chloroform diluent because the distribution of acid into the solvent phase is too low to be measured accurately. Although attributing the larger enthalpy of complexation of succinic acid compared to lactic acid to a difference in the enthalpy of mixing of the complex in the organic phase is reasonable, it should be considered to be speculative. The differences between enthalpy and entropy changes in MIBK for the (2,l) succinic acid-amine and -lactic acid complexes are about the same as for the (1,l) complexes and may again primarily reflect differences in the partial molar heats of mixing of the complexes in the organic phase. Temperature-Swing Regeneration. In a temperature-swing extraction/regeneration scheme, the extraction is carried out at a relatively low temperature, producing an acid-loaded organic extract and an aqueous raffinate waste stream containing the unwanted feed components. During regeneration, the extract is contacted with a fresh aqueous stream at a higher temperature to produce an acid-laden aqueous product stream and an acid-free organic phase. The concentration of the acid achievable in this stream depends upon the amount of change in the extraction equilibrium between temperatures and can be higher than that in the original aqueous feed stream. MIBK has desirable properties for use as a diluent in temperature-swing regeneration of succinic acid. Distribution ratios for extraction are high enough for facile extraction, and at higher temperatures, the acid readily
Ind. Eng. Chem. Res., Vol. 29, No. 7, 1990 1337 0.3
loo vol%
/
F I, W,A
i
-
A-B,C.H
I
- _T _ _ _ ;m I
m
r,P W,A
-
A = Acid B = Base F wci
-
(active) H = Heptane (inen) W = Water I I Undesired
0 0
02
0.4
Acid in Aqueous Phase (mol/L)
Figure 5. Effect of the volume percent of active diluent (chloroform) in inert diluent (n-heptane) for the extraction of succinic acid by 0.29 mol/L Alamine 336.
distributes back into the aqueous phase. Although the temperature effect is sharper, chloroform may not be as good a diluent choice for temperature-swing regeneration of succinic acid. Because of the substantial curvature of the equilibrium relationship, there will be a tangent pinch midway along the operating line for regeneration with a simple back-extraction process. A temperature increase of 55 "C does not lower the equilibrium curve enough for practical back-extraction. An alternative is to mix chloroform (or some other active diluent) with an inert diluent to lower the magnitudes of the distribution ratios. This is a very flexible way to bring both the extraction and regeneration equilibria into appropriate distribution ranges. Another possibility is to use a split-flow configuration to move the operating curve around the pinch point (Tamada and King, 1989). In such a design, partially regenerated solvent is removed midway along the regeneration and added midway along the extractor to the partially loaded solvent. Because of the lower enthalpy change, the extraction of lactic acid by Alamine 336 in MIBK and chloroform does not show as large a temperature effect. Hence, temperature-swing regeneration would be less effective for lactic acid than for succinic acid. Diluent-SwingRegeneration. Baniel et al. (1981) also describe regeneration by back-extraction following a change in diluent composition. In a diluent-swingprocess the extraction is carried out in a solvent composed of the amine and a diluent that promotes distribution of the acid into the organic phase. The composition of the acid-laden organic phase leaving the extractor is then altered, by either removal of the diluent or addition of another diluent, to produce a solvent system that promotes distribution of the acid to the aqueous phase. This altered organic phase is contacted with a fresh aqueous stream in the regenerator to produce the acid-laden aqueous product and the aciddepleted solvent for recycle to the extractor. Adjustment of diluent composition can also occur before this solvent stream reenters the extractor. The insight gained in parts 1 and 2 enables rational selection of effective diluent-swing processes. Equilibrium Studies. Figure 5 shows experimental results for succinic acid extracted by 0.29 mol/L Alamine 336 in a diluent composed of varying ratios of chloroform (active) to n-heptane (inert) at 25 "C. Similar results were obtained for succinic acid extracted by 0.29 mol/L Alamine 336 in mixtures of methylene chloride and n-heptane (Tamada and King, 1989). For 100% heptane diluent, two organic phases are formed. Although it must be considered to be empirical without further substantiation, a reasonable fit to the experimental data is achieved with a model of (1,1,3)-one acid, one amine, three diluent-complex
Figure 6. Diluent-swing regeneration process flow diagram. Ideally, the extraction stage selectively removes the acid from the aqueous multicomponent feed mixture, and the regeneration stage concentrates the aqueous product acid solution compared to the aqueous feed solution.
formation with log Kl13= -0.63 and -0.32 ( m ~ l / L ) for -~ chloroform and methylene chloride diluents, respectively (Tamada and King, 1989). Process Considerations. Removal of the active diluent has a strong effect on the equilibrium distribution of the acid in Figure 5. The equilibrium curve at high heptane concentration is no longer concave downward, as is the case at high chloroform concentration. Thus, the tangent pinch that may occur in a back-extraction regeneration process is ameliorated. In one example scheme (Figure 6), the active diluent is distilled from the extract leaving the primary extractor and is recycled to join the regenerated solvent leaving the regenerator. Note that the distillation of the extract has the additional advantage of preconcentration of the acid in the extract before regeneration. Another possibility involves use of a relatively nonvolatile active diluent (e.g., a long-chain alcohol) in the extractor. A volatile inert diluent is added to the loaded extract prior to regeneration and then removed from the regenerated extract by distillation before reentering the extractor. This has the disadvantageof diluting the extract stream and requiring distillation of larger amounts of solvent (after the regeneration) to obtain the same shift in the active/inert diluent ratio. However, this configuration has a subtle compensating advantage. Recall that the loading of dicarboxylic acids in inert diluents decreases with decreasing amine concentration (part 1). Therefore, halving the amine concentration more than halves the equilibrium amount of acid extracted at a given aqueous acid concentration. A greater shift in equilibrium is achieved for the same shift in active/inert diluent ratio.
Discussion Diluent swing and temperature swing are compared by Tamada and King (1989). At low feed concentrations for the succinic acid-Alamine 336-chloroform-heptane case, diluent swing provides greater concentrating ability, whereas at high feed concentrations temperature swing is superior in this aspect. This results from the slightly concave downward shape of the temperature-swing regeneration equilibrium curve and the concave upward shape of the diluent-swing regeneration equilibrium curve. Equipment costs for regeneration by diluent swing would be greater than for temperature swing, because of the distillation step. The energy consumption for a temperature-swing process comes from the need to heat the streams entering the regenerator. The energy consumption for a diluent-swingprocess relates to the heat input to the reboiler of the distillation column. Temperature swing is favored by a large heat of the complexation reaction. For diluent swing, the solvent flow rate is the most important
I n d . Eng. Chem. Res. 1990,29, 1338-1345
1338
factor in determining energy costs; a high concentration of solute in the extract will allow for a more efficient process. Coupling of diluent and temperature swing may be advantageous. The additional energy cost of the distillation over simple temperature swing may be offset by the greater concentration of the product and the fact that diluent swing offers less pinched, more favorable equilibrium curves.
Acknowledgment We are grateful to the late A. Steven Kertes for helpful discussions. We thank Eric Chan, Roy Kamimura, and Maninderpal Grewal, who made significant contributions toward the experimental portion of this work. This work was supported by a National Science Foundation Graduate Fellowship and by the Assistant Secretary for Conservation and Renewable Energy, Office of Energy Systems Research, Energy conversion and Utilization Technologies (ECUT) Division, U S . Department of Energy, under Contract DE-AC03-76SF00098. Registry No. MIBK, 108-10-1;succinic acid, 110-15-6;lactic acid, 50-21-5; dichloromethane, 75-09-2;nitrobenzene, 98-95-3; chloroform, 67-66-3; 1-octanol, 111-87-5; water, 7732-18-5.
Literature Cited Apelblat, A. Enthalpy of Solution of Oxalic, Succinic, Adipic, Maleic, Malic, Tartaric, and Citric Acids, Oxalic Acid Dihydrate, and
Citric Acid Monohydrate in Water at 298.15 K. J.Chem. Thermodyn. 1986, 18, 351-357. Baniel, A. M.; Blumberg, R.; Hajdu, K. Recovery of Acids from Aqueous Solutions. U.S.Patent 4,275,234, 1981; Chem. Abstr. 1982, 97, 109557.
Holten, C. H. Lactic Acid: Properties and Chemistry of Lactic Acid and Derivatives; Verlag Chemie: Copenhagen, 1971; pp 20-58. Kertes, A. S.; King, C. J. Extraction Chemistry of Fermentation Product Carboxylic Acids. BiotechnoL Bioeng. 1986,28, 269-282. King, C. J. Binary Multistage Separations: General Graphical Approach. In Separation Processes, 2nd ed.; McGraw-Hill Book Co.: New York, 1980; pp 258-296. Sato, T.; Watanabe, H.; Nakamura, H. Extraction of Lactic, Tartaric, Succinic, and Citric Acids by Means of Trioctylamine. Bunseki Kagaku 1985,34, 559-563. Saville, G.; Gundry, H. A. The Heats of Combustion, Solution and Ionization of Lactic Acid. Trans. Faraday SOC.1959, 55, 2036-2038. Smith, J. M.; Van Ness, H. C. Phase Equilibria. In Introduction to Chemical Engineering Thermodynamics, 3rd ed.; McGraw-Hill Book Co.: New York, 1975; Chapter 8, pp 290-375. Tamada, J. A. PhD. Dissertation, University of California, Berkeley,
1989.
Tamada, J. A.; King, C. J. Extraction of Carboxylic Acids by Amine Extractants. Report LBL-25571; Lawrence Berkeley Laboratory: Berkeley, CA, Jan 1989. 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.
Received for review August 30, 1989 Revised manuscript received February 12, 1990 Accepted February 21, 1990
Aqueous-Phase Adsorption of Trichloroethene and Chloroform onto Polymeric Resins and Activated Carbon Thomas E. Browne and Yoram Cohen* Department of Chemical Engineering, University of California, Los Angeles, Los Angeles, California 90024-1592
The adsorption of the EPA priority pollutants trichloroethene (TCE) and chloroform (CHC13)from aqueous solutions onto activated carbon and macroporous polymeric resins was investigated over a wide concentration range. Over much of the concentration range studied, activated carbon adsorbed more TCE and CHC13 on a weight basis and on a surface area basis than did the polymer resins. A t concentrations greater than 1000 hg/L, the adsorption capacity, on a surface area basis, for the different resins and for activated carbon was similar. The adsorption isotherms were fitted by the empirical Freundlich isotherm and the semiempirical isotherm of Jossens et al. The affinity of the solutes for the different resins was qualitatively described by a simple Hildebrand solubility parameter correlation.
Introduction Over the past 2 decades, it has often been suggested that macroporous polymeric resins could be useful for the adsorptive removal of a variety of organic solutes from aqueous systems (Gustafson et al., 1968; Paleos, 1969; Chriswell et al., 1977; Chudyk et al., 1979; Fox, 1979 Neely, 1980; Van Vliet and Weber, 1981; Chanda et al., 1983; Cornel and Sontheimer, 1986). In general, it has been found that polymeric resins have a lower adsorption capacity than does activated carbon for most organic solutes (McGuire and Suffet, 1978; Van Vliet and Weber, 1981; No11 and Gounaris, 1988). Several studies, however, demonstrated that for the solutes p-chlorophenol and ptoluenesulfonate (Van Vliet and Weber, 1981), p-chloro-
* Author to whom correspondence should be addressed.
phenol and phenol (No11 and Gounaris, 1988), p-nitrophenol (Kim et al., 1976), benzaldehyde (Furusawa and Smith, 1974),and phenol and carbon tetrachloride (Weber and Van Vliet, 1981), the slopes of the adsorption isotherms for the polymeric resins were steeper than the slopes of the corresponding isotherms for activated carbon. The results of these latter studies suggested that at sufficiently high solute concentrations the polymeric resins may have a higher capacity than does activated carbon. It is well-known that the pore size distribution and surface area affect the resin’s capacity for solute adsorption (Paleos, 1969; Kunin, 1977; Feeney, 1979; Chudyk et al., 1979; Slejko, 1985; Cornel and Sontheimer, 1986). Yet most of the studies on water decontamination rarely report adsorption capacity based on the measured surface area (i.e., moles of solute/squared meter), and even fewer studies have considered the possible influence of the in-
0888-5885190J 2629-1338$02.50/0 0 1990 American Chemical Society
