5766
Langmuir 1997, 13, 5766-5769
Linear Solvation Energy Relationships To Explain Interactions Responsible for Solute Adsorption onto a Polar Polymeric Sorbent Brian J. Brune, Gregory F. Payne,* and Mahesh V. Chaubal Department of Chemical and Biochemical Engineering and Center for Agricultural Biotechnology, University of Maryland Baltimore County, 5401 Wilkens Avenue, Baltimore, Maryland 21228 Received February 12, 1997. In Final Form: July 31, 1997
Introduction The industrial separations of low-valued bulk chemicals has relied on distillation for separating volatile compounds and crystallization for the separation of nonvolatiles. Unfortunately, these traditional separation operations are often unable to satisfy the emerging demands for reduced energy costs and/or improved product purities. In many cases, adsorptive separations provide an alternative separations approachsespecially when the sorbent offers discriminating capabilities that yield high separation factors. Initially, large-scale adsorptive separations employed molecular sieves to separate compounds based on differences in shape and size.1-3 More recently, sorbents with polar/electrostatic surface sites are being used to separate molecules due to differences in sorbate-sorbent interaction energies. Examples of these latter include sorbents that exploit quadrupolar interactions for the preferential adsorption of nitrogen from oxygen,4,5 complexation by calcium to separate fructose from glucose,6-8 and interactions between silver and π electrons for olefinparaffin separations.9-11 To fully exploit these polar adsorptive mechanisms, the nature and energetic contributions of these specific interactions must be quantitatively understood. Polar adsorptive interaction mechanisms have been difficult to quantitatively characterize because most sorbents (e.g., activated carbon and silica) have complex and often ill-defined surface chemistries that permit solutes to adsorb at a variety of surface sites through a range of mechanisms with a distribution of interaction energies. When multiple mechanisms are responsible for adsorption, it is not only difficult to assess the contributions of individual mechanisms12 but, the overall selectivity for binding can also be reduced. Further, despite recent advances, computational methods have not yet advanced to the point that they can reliably predict experiment.13 In summary, there is a growing need to exploit specific adsorptive interactions for inexpensive separations; however, the current understanding of adsorption mechanisms limits a priori prediction of adsorption equilibrium or the (1) Broughton, D. B. Sep. Sci. Technol. 1984-1985, 19 , 723-736. (2) Ruthven, D. M. Principles of adsorption and adsorption Processes; Wiley: New York, 1984. (3) Jasra, R. V.; Bhat, S. G. T. Sep. Sci. Technol. 1988, 23, 945-989. (4) Yang, R. T.; Chen, Y. D.; Peck, J. D.; Chen, N. Ind. Eng. Chem. Res. 1996, 35, 3093-3099. (5) Jasra, R. V.; Choudary, N. V.; Bhat, S. G. T. Ind. Eng. Chem. Res. 1996, 35, 4221-4229. (6) Goulding, R. W. J. Chromatogr. 1975, 103, 229-239. (7) Ho, C.; Ching, C. B.; Ruthven, D. M. Ind. Eng. Chem. Res. 1987, 26, 1407-1412. (8) Cheng, Y. L.; Lee, T. Y. Biotech. Bioeng. 1992, 40, 498-504. (9) Yang, R. T.; Kikkinides, E. S. AIChE J. 1995, 41, 509-517. (10) Chen, N.; Yang, R. T. Langmuir 1995, 11, 3450-3456. (11) Wu, Z.; Han, S.; Cho, S.; Kim, J.; Chue, K.; Yang, R. T. Ind. Eng. Chem. Res. 1997, 36, 2749-2756. (12) Barrer, R. M. J. Colloid Interface Sci. 1966, 21, 415-434. (13) Chen, N.; Yang, R. T. Ind. Eng. Chem. Res. 1996, 35, 40204027.
S0743-7463(97)00149-2 CCC: $14.00
extrapolation of single solute data to predict multisolute adsorption. We have been studying polymeric sorbents because the homogeneous surface chemistry of polymers offers the potential for limiting adsorption to a small number of mechanistic interactions. Also, we have focused on an acrylic ester sorbent because it should be possible to exploit its polar functionality to confer selectivity to adsorption. Practically, acrylic ester sorbents are inexpensive and offer surface areas and mechanical properties appropriate for large-volume, bulk chemical separations. Initial studies provided phenomenological support for Scheme 1, which shows that solutes with hydrogen-bond donating ability (i.e., hydrogen-bond donor solutes) adsorb from a nonpolar solvent (i.e., hexanes) onto the acrylic ester sorbent due to the formation of a sorbate-sorbent hydrogen-bond.14-16 The quantitative framework developed for this hydrogen-bonding scheme was able to be extrapolated to explain separation factors observed for more complex solute systems17 and to predict separation factors for simple solute mixtures in more complex solvent arrangements.18,19 The purpose of the present work was to examine the adsorption of dipolar solutes that lack hydrogen-bond donor acidity and therefore do not fit the model of Scheme 1. To suggest the interaction mechanisms responsible for solute adsorption from hexanes onto the sorbent surface, the experimentally-measured adsorption equilibria was correlated in terms of standard linear solvation energy relationships (LSERs). LSERs were originally developed by Kamlet, Taft, and co-workers and have been modified over the years.20-23 LSERs have the general form
XYZ ) (XYZ)o + mM + sπ*+ aR + bβ
(1)
where XYZ represents a physicochemical property, (XYZ)o is a constant, and a, b, m, and s are regression coefficients. M is a term that varies depending on whether the correlation is being used to describe solvent or solute behavior. When LSERs are used to describe solvent behavior, M in eq 1 is related to the Hildebrand solubility parameter. When describing a compound’s behavior as a solute, M is related to the size of the cavity that must be formed for a solute to be transferred to a solvent. The last three terms in eq 1 describe dipolar and hydrogenbonding interactions, and quantitative measures for these terms have been provided by UV/vis/near-IR spectroscopic measurements.23 The term π* describes dipole/dipole and dipole/induced dipole interactions and is directly related to the dipole moment and polarizability.24 The hydrogenbonding acidity, R, describes a compound’s hydrogen-bond donating (HBD) ability, while β is the corresponding (14) Payne, G. F.; Payne, N. N.; Ninomiya, Y.; Shuler, M. L. Sep. Sci. Technol. 1989, 24, 457-465. (15) Payne, G. F.; Ninomiya, N. Sep. Sci. Technol. 1990, 25, 11171129. (16) Maity, N.; Payne, G. F.; Chipchosky, J. L. Ind. Eng. Chem. Res. 1991, 30, 2456-2463. (17) Chaubal, M. V.; Payne, G. F. Biotech. Prog. 1995, 11, 468-470. (18) Maity, N.; Payne, G. F. Langmuir 1991, 7, 1247-1254. (19) Payne, G. F.; Maity, N. Ind. Eng. Chem. Res. 1992, 31, 20242033. (20) Abraham, M. H. Chem. Soc. Rev. 1993, 22, 73-83. (21) Kamlet, M. J. Prog. Phys. Org. Chem. 1993, 19, 295-317. (22) Marcus, Y. Chem. Soc. Rev. 1993, 22, 409-416. (23) Reichardt, C. Chem. Rev. 1994, 94, 2319-2358. (24) Li, J.; Zhang, Y.; Dalas, A. J.; Carr, P. W. J. Chromatogr. 1991, 550, 101-134.
© 1997 American Chemical Society
Notes
Langmuir, Vol. 13, No. 21, 1997 5767 Scheme 1
hydrogen-bond accepting (HBA) ability.22,25 LSERs have traditionally been used to correlate the solvent effects on reaction rates and equilibria, and the phase transfer characteristics of solutes (e.g., octanol-water partition coefficients and chromatographic capacity factors).24,26,27 More recently, LSERs have been applied to the description of other thermodynamic properties such as partial molar enthalpies,28 activity coefficients at infinite dilution,29 and interaction parameters for surface tension.30
Figure 1. Benzaldehyde adsorption from hexanes onto acrylic ester and styrenic sorbents. For comparison, the adsorption data are normalized in terms of the specific surface areas. The adsorption temperature was 25 °C.
Materials and Methods The sorbents used in this study were Rohm and Haas resins XAD-7 (acrylic ester) and XAD-16 (styrenic)31 with specific surface areas reported by the supplier (Sigma Chemical) to be approximately 450 and 800 m2/gm, respectively. All sorbents were washed extensively with methanol and water and dried prior to use. For all studies, hexanes (supplied as a mixture of isomers by Fisher Chemical) were used as the solvent. Adsorption studies were conducted by equilibrating known amounts of sorbent with hexane solutions containing a single solute. The adsorbed concentration of solute (q) was calculated from the difference between the initial (Co) and equilibrated (C) concentrations of the solute in hexanes by
q ) (Co - C)V/A
(2)
where V is the volume of hexanes and A is the mass of resin. Hexane phase solute concentrations were determined by spectrophotometry.
Results and Discussion Figure 1 compares the adsorption of benzaldehyde from hexanes onto the acrylic ester sorbent with benzaldehyde adsorption onto a styrenic sorbent. To account for differences in the specific surface areas of these sorbents, the adsorbed concentration is normalized per unit surface area. The difference in adsorption affinities shown in Figure 1 demonstrates the importance of the polar surface of the acrylic ester sorbent for adsorbing polar solutes from hexanes. Since benzaldehyde lacks significant hydrogen-bond acidity,20,21 it seems unlikely that the hydrogen-bonding mechanism of Scheme 1 is responsible for adsorption of this solute onto the acrylic ester sorbent. To further study the adsorption of non-HBD solutes, a series of monosubstituted benzenes was studied. The solid lines in Figure 2 show the best fit lines characterizing the adsorption equilibria for these non-HBD solutes. Al(25) Abraham, M. H.; Buist, G. J.; Grellier, P. L.; McGill, R. A.; Prior, D. V.; Oliver, S.; Turner, E.; Morris, J. J.; Taylor, P. J.; Nicolet, P.; Maria, P. C.; Gal, J. F.; Abboud, J. L. M.; Doherty, R. M.; Kamlet, M. J.; Shuely, W. J.; Taft, R. W. J. Phys. Org. Chem. 1989, 2, 540-552. (26) Kamlet, M. J.; Doherty, R. M.; Abraham, M. H.; Marcus, Y.; Taft, R. W. J. Phys. Chem. 1988, 92, 5244-5255. (27) Li, J.; Zhang, Y.; Carr, P. W. Anal. Chem. 1992, 64, 210-218. (28) Sherman, S. R.; Suleiman, D.; Hait, M. J.; Schiller, M.; Liotta, C. L.; Eckert, C. A.; Li, J.; Carr, P. W.; Poe, R. B.; Rutan, S. C. J. Phys. Chem. 1995, 99, 11239-11247. (29) Sherman, S. R.; Trampe, D. B.; Bush, D. M.; Schiller, M.; Eckert, C. A.; Dallas, A. J.; Li, J.; Carr, P. W. Ind. Eng. Chem. Res. 1996, 35, 1044-1058. (30) Lee, L. H. Langmuir 1996, 12, 1681-1687. (31) Maity, N.; Payne, G. F.; Ernest, M. V., Jr.; Albright, R. L. React. Polym. 1992, 17, 273-287.
Figure 2. Adsorption of various solutes from hexanes onto the acrylic ester sorbent. The adsorption temperature was 30 °C. Data for the adsorption of phenol and benzyl alcohol were adjusted to 30 °C using previously reported adsorption affinities and enthalpies.16 The concentration range for each solute was chosen to exploit the sensitivity of the spectrophotometer while minimizing errors associated with sample dilution.
though the concentration range was varied for each solute to facilitate accurate determination of concentrations, Figure 2 shows that data for each solute is linear and can be extrapolated to the origin. The adsorptions of HBD solutes that are capable of hydrogen-bonding to the sorbent (e.g., phenol and benzyl alcohol) are illustrated by the dotted lines in Figure 2.16 As can be seen from Figure 2, HBD solutes capable of hydrogen-bonding to the acrylic ester sorbent have a higher adsorptive affinity compared to non-HBD solutes. Nevertheless, Figure 2 shows that dipolar solutes that lack HBD acidity can still adsorb onto the acrylic ester sorbent and that the adsorption affinities for these solutes vary considerably. To suggest interaction mechanisms responsible for adsorption, the adsorption affinity (i.e., the slope of the isotherm in the linear region) for a series of solutes is plotted as a function of the LSER parameters, as shown in Figure 3. The solutes used in Figure 3 are either monosubstituted benzenes (e.g., phenol, toluene, chlorobenzene, and anisole) or heterocyclic compounds with a single polar region (e.g., pyridine and quinoline). Consistent with Scheme 1, Figure 3a shows that for solutes capable of hydrogen-bonding to the sorbent, the adsorption affinity correlates to the HBD acidity parameter, R. However, Figure 3a shows that there are a series of compounds that lack HBD acidity (i.e., R ) 0) but have substantial adsorptive affinities. Parts b-d of Figure 3 show little correlation between the adsorption affinity and the HBA basicity (β), the dipolarity/polarizability (π*), or the cavity-forming term (V1/100). Thus, Figure 3 supports the conclusion that for solutes with significant R values, hydrogen-bonding is a dominant adsorption mechanism.
5768 Langmuir, Vol. 13, No. 21, 1997
Notes
Figure 4. Correlation between affinities for adsorption from hexanes onto acrylic ester sorbent and the solute’s dipolarity/ polarizability for solutes that lack hydrogen-bond donor acidity (i.e., R ) 0). The adsorption temperature was 30 °C. The open symbols, closed symbols, and asterisks refer to parameter values reported by Abraham,20 Kamlet,21 and Laurence et al.,32 respectively. In parentheses after the solute name is the experimentally measured adsorption affinity.
For solutes that lack HBD acidity (i.e., R ) 0), Figure 4 shows that the adsorption affinity can be related to π*, the dipolarity/polarizability term. Included in Figure 4 are π* values recently reported by Laurence et al.32 The correlation in Figure 4 suggests that dipolar or inductive mechanisms are responsible for the adsorption of monosubstituted benzenes or heterocyclics that have a single polar region but lack HBD acidity. Similar conclusions have been reported by Ageev et al.33 and Belyakova et al.,34 who performed chromatographic studies with a polar copolymeric sorbent, a hexane eluent, and a series of monosubstituted, non-HBD benzene derivatives. The chromatographic capacity factors observed in their studies could be correlated to the solute’s dipole moment. Similarly, surfaces with substantial polarity (e.g., oxidized carbon, silica, and sodium-exchanged zeolites) have been observed to yield higher adsorption affinities and/or adsorption heats for gases with significant quadrupole moments (e.g., N2 and CO2), while gas adsorption onto relatively nonpolar surfaces (e.g., graphitized carbon and silicalite) is correlated to gas polarizability and independent of gas polarity.12,35-38 The potential for dipolar interactions to provide a significant mechanism for adsorption from hexanes may explain the differences in benzaldehyde adsorption onto styrenic versus acrylic ester sorbents (Figure 1). Also, dipolar interactions may explain the strong adsorption of ion pairs onto acrylic ester sorbents.39,40 Quaternary ammonium ion pairs with either benzoate or benzenesulfonate, which lack obvious hydrogen-bond-donating ability, were observed to adsorb from hexanes onto the acrylic ester sorbent with a greater affinity and more negative enthalpy than phenol, a HBD solute that forms
Figure 3. Correlation between affinities for adsorption from hexanes onto the acrylic ester sorbent and the solute’s LSER parameters for (a) hydrogen-bond donor acidity, (b) hydrogenbond acceptor basicity, (c) dipolarity/polarizability, and (d) cavity formation. Adsorption affinity is a pseudoequilibrium constant obtained from the slope of the adsorption isotherms at 30 °C. The open symbols refer to parameter values from Abraham,20 while the closed symbols refer to parameter values from Kamlet.21 The hydrogen-bond acidity of acetophenone is reported to be either 020 or 0.06.21
(32) Laurence, C.; Nicolet, P.; Dalanti, M. T.; Abbound, J. M.; Notario, R. J. Phys. Chem. 1994, 98, 5807-5816. (33) Ageev, A. N.; Aratskova, A. A.; Belyakova, L. D.; Gvozdovich, T. N.; Kiselev, A. V.; Yahin, Y. I.; Kalal, J.; Svec, F. Chromatographia 1983, 17, 545-548. (34) Belyakova, L. D.; Platonova, N. P.; Shevchenko, T. I.; Svec, F.; Hradil, J. Pure Appl. Chem. 1989, 61, 1889-1896. (35) Kiselev, A. V. J. Colloid Interface Sci. 1968, 28, 430-442. (36) Golden, T. C.; Sircar, S. J. Colloid Interface Sci. 1994, 162, 182188. (37) Dunne, J. A.; Mariwala, R.; Rao, M.; Sircar, S.; Gorte, R. J.; Myers, A. L. Langmuir 1996, 12, 5888-5895. (38) Dunne, J. A.; Rao, M.; Sircar, S.; Gorte, R. J.; Myers, A. L. Langmuir 1996, 12, 5896-5904. (39) Ramakrishnan, S.; Payne, G. F. Langmuir 1994, 10, 3827-3835. (40) Payne, G. F.; Ramakrishnan, S. Ind. Eng. Chem. Res. 1995, 34, 575-584.
Notes
Langmuir, Vol. 13, No. 21, 1997 5769
a relatively strong hydrogen bond with the acrylic ester sorbent. The ability of quaternary ammonium ion pairs to interact strongly through polar interactions is also consistent with results of Kersten et al., who studied such ion pairs as the stationary phase for GC packing.41 Conclusion In principle, adsorption could offer high selectivities for industrial separations if inexpensive sorbents were available that allowed adsorption to occur through a limited number of specific sorbate-sorbent interactions. The polyacrylic ester sorbent used here is commercially available and inexpensive and may be able to limit adsorption to specific interactions due to its well-defined (41) Kersten, B. R.; Poole, S. K.; Poole, C. F. J. Chromatogr. 1989, 468, 235-260.
surface chemistry and polar functionality. In this study we examined the adsorption of a series of neutral solutes from hexanes onto this sorbent. The correlation of our results using standard physical organic chemistry relationships (i.e., LSERs) suggests that hydrogen-bonding and polar interactions play a predominant role in solute adsorption from a nonpolar solvent. More importantly, these correlations provide a quantitative framework to explain the results, and hopefully to permit extrapolation to predict adsorption in more complex systems (e.g., in multicomponent systems). Acknowledgment. Financial support was provided by the National Science Foundation through grant CTS9531812 and a REU supplement to grant BES-9315449. LA970149X