J. Am. Chem. SOC.1993, 115, 701-707
701
Probing Solute-Entrainer Interactions in Matrix-Modified Supercritical C 0 2 JOAM Zagrobelny and Frank V. Bright* Contributionfrom the Department of Chemistry, Acheson Hall, State University of New York at Buffalo, Buffalo, New York 14214. Received June 25, 1992
Abstract: Binary supercritical fluids composed of C 0 2 and small amounts of acetonitrile or methanol (entrainers) are studied as a function of fluid density, using pyrene as the solute probe. These experiments show how preferential entrainer clustering effects pyrene excimer formation. To this end, steady-state and time-resolved fluorescence spectroscopy are used in concert to probe the ground and excited states of pyrene. Results show that entrainers enhance solutesolvent clustering and slow the, traditionally diffusion-controlled,excimer formation reaction. Clustering also helps shield the excimer from non-radiative deactivation processes as evidenced by a longer decay time in the near-critical region. All data over our concentration range are consistent with a homogeneous pyrene ground state. There is no evidence for solutesolute (pyrene-pyrene) interactions prior to excitation. However, once in the excited state, excimer formation is facilitated due to the decreased fluid viscosity compared to liquids.
Introduction
The characteristic critical point for any chemical system is defined by its critical temperature (T,)and pressure (P,). Immediately below these points exists an equilibrium between the liquid and gaseous phases. Above the critical p i n t the two phases coalesce into a supercritical fluid.’ One unique feature of supercritical fluids is that the solvent characteristics are variable over a wide range.2 For example, in the rear-critical region one can use temperature and pressure to adjust density from gas-like to liquid-like. However, the transport properties (Le., diffusivity and viscosity) remain more on the order of a Thus, supercritical solvents possess enhanced solvating power compared to gases and improved mass transport compared to liquids.’-3 During this past decade, the unique characteristics of supercritical fluids have been used widely to effect separations and More recent work has shown that supercritical fluids can be used as novel reaction media.’b-8-’4 On the basis of the well-know fact that solvents can affect reaction kinetics,15 the tunability of supercritical fluids provides, in principle, a convenient way to control chemical reactions without changing the molecular structure of the solvent. However, to fully realize the potential of supercritical fluids, it becomes necessary to understand the molecular interactions and the system phase behavior. To this end, chemists and engineers have investigated various solutesupercritical fluid systems. ErlichI6 first observed negative partial molar volumes for polar solutes in supercritical fluids. This anomalous result lead to more detailed thermodynamic (e.g., solubility and partial molar properties at infinite dilution) in supercritical fluids. For example, Eckert et al.18*19 recovered partial molar volumes which were large and negative (approximately 100 times the bulk solvent molar volume) in the near-critical region. These experiments suggested an augmentation of solvent molecules about the solute a t infinite dilution. This eventually lead to the idea that the fluid composition in the cybotactic region was enriched relative to the bulk-the so called clustering phenomena. Following from these experiments several groups have used molecular dynamics simulations and m ~ d e l i n g ~and & ~~~p e c t r o s c o p y ~ ’ -to ~ ~probe ~~”~ this clustering phenomena. Debenedetti et al.24925qualitatively described these solutefluid interactions for three limiting cases: attractive, weakly attractive, or repulsive. Each case was characterized by solutesolvent partial molar volume, enthalpy, and correlation length fluctuations. Solute-fluid clustering was apparent only in the attractive case.24J5 More recent results26 predict a dynamical solute-fluid cluster. Cochran et al.28*29calculated pair correlation functions and fluctuation integrals for supercritical systems using integral equation theories. Near the critical point, these authors observed Author to whom all correspondence should be addressed.
0002-7863/93/1515-701$04.00/0
the growth of the solutesolvent fluctuation integral resulting from an increase in the correlation length of the density fluctuations.
( I ) (a) Reid, R. C.; Prausnitz, J. M.; Poling, B. E. The Properties ofcases and Liauids. 4th ed.:McGraw-Hill: New York. 1987. Ib) Suoercritical Fluid Techndlo&Reviews in Modern Theory and Applications; Bruno,T. J., Ely, J. F., Eds.; CRC Press: Boca Raton, FL, 1991. (2) McHugh, M. A.; Krukonis, V. J. Supercritical Fluid ExtractionPrinciples and Practice: Butterworths: Boston, MA, 1986. (3) Klesper, E. Angew. Chem., Int. Ed. Engl. 1978, 17, 138. (4) Novotny, M. V.; Springston, S.R.; Peaden, P. A,; Fjeldstad, J. C.; Lee, M. L. Anal. Chem. 1981, 53, 407A. (5) Supercritical Fluid Chromatography; Smith, R. M., Ed.;Royal Society of Chemistry Monograph: London, 1988. (6) Berger, T. A.; Deye, J. F. J . Chromatogr. Sci. 1991, 29, 141. (7) Brennecke, J. F.; Eckert, C. A. AIChE J . 1989, 35, 1409. (8) Supercritical Fluid Technologv-Theoretical and Applied Appmches in Analytical Chemistry; Bright, F. V., McNally, M. E., Eds.; ACS Symp. Ser. 488; American Chemical Society: Washington, DC, 1992; Chapter 1. (9) Fulton, J. L.; Yee, G. G.; Smith, R. D. J . Am. Chem. SOC.1991, 113, 8327. (10) Howdle, S. M.; Healy, M. A.; Poliakoff, M. J . Am. Chem. Soc. 1990, 1 12.4804. (11) Narayan, R.; Antal, M. J. J . Am. Chem. Soc. 1990, 112. 1927. (12) Erickson, J. C.; Schyns, P.; Cooney, C. L. AIChE J. 1990,36,299. (13) Nakamura, K.; Fujii, H.; Chi, Y. M.; Yano, T. Ann. N.Y. Acad. Sci. 1990, 613, 319. (14) Nakamura, K. Trends Biotechnol. 1990, 8, 288. (1 5) March, J. Advanced Organic Chemistry, 3rd ed.;Wiley-Interscience: New York, 1985. (16) Ehrlich, P.; Fariss, R. J . Phys. Chem. 1969, 73, 1164. (17) Wu, P. C.; Ehrlich, P. AlChE J . 1973, 19, 533. (18) Eckert, C. A.; Ziger, D. H.; Johnston, K. P.; Ellison, T. K. Fluid Phase Equilib. 1983, 14, 167. (19) Eckert, C. A.; Ziger, D. H.; Johnston, K. P.; Kim, S. J. Phys. Chem. 1986, 90, 2738. (20) Debenedetti, P. G. Chem. Eng. Sci. 1987, 42, 2203. (21) Kim, S.; Johnston, K. P. Ind. Eng. Chem. Res. 1987, 26, 1206. (22) Kim, S.; Johnston, K. P. AlChE J . 1987, 33, 1603. (23) Debenedetti, P. G.; Kumar, S. K. AIChE J . 1988, 34.645. (24) Debenedetti, P. G.; Mohamed, R. S.J . Chem. Phys. 1990,90,4528. (25) Debenedetti, P. G.; Petsche, I. B.; Mohamcd, R. S. Fluid Phase Equilib. 1989, 52, 347. (26) Petsche, 1. 8.; Debenedetti, P. G. J. Phys. Chem. 1991, 95, 386. (27) Petsche, I. B.; Debenedetti, P. G. J. Chem. Phys. 1989, 91, 7075. (28) Cochran, H. D.; Lee, L. L. In Supercritical Fluid Science and Technology; Johnston, K. P., Penninger, J. M. L., Eds.; ACS Symp. Ser. 406; American Chemical Society: Washington, DC, 1989; Chapter 3. (29) Cochran, H. D.; Lee, L. L.; Pfund, D. M. Proc. Int. Symp. Supercritical Fluids 1988, 245. (30) Johnston, K. P.; Kim, S.;Combs, J . In Supercritical Fluid Science ond Technology;Johnston, K. P., Penninger, J. M. L., Eds.; ACS Symp. Ser. 406; American Chemical Society: Washington, DC, 1989; Chapter 5 .
0 1993 American Chemical Society
702 J. Am. Chem. SOC.,Vol. 115, No. 2, 1993
Zagrobehy and Bright
These r e s ~ l t support s ~ ~ ~clustering ~~ in attractive solute-solvent pairs and cavitation in repulsive systems. This work also suggests that solutesolvent clustering may enhance solutesolvent interactions. Spectroscopic measurements provide a molecular-level view of solute-fluid interaction and support cluster f o r m a t i ~ n . ~ ' q ~ ~ , ~ & ~ ~ For example, steady-state and time-resolved fluorescence studies of the solute 6-propionyl-2-(dimethylamino)naphthalene (PRODAN) in CF3H show clear evidence for the formation of solute-fluid clusters.33 Additional s t ~ d i e s ~show ~ J ~that in the Figure 1. Energy-level diagram for pyrene excimer formation: hv. abnear-critical region the local density of N 2 0 about PRODAN is sorbed photon; kM, de-excitation rate from monomer species; k D M , bfapproximately 2.5 times greater than the bulk. The average molecular rate coefficient for formation of pyrene excimer; k M D , uninumber of solvent molecules in the cluster has been determined molecular rate coefficient for dissociation of the pyrene excimer; and kD, e ~ p e r i m e n t a l l y ,and ~ ~ , is ~ ~in good agreement with many of the de-excitation rate from the excimer species. theoretical calculations (vide supra).2b22 Additional fluoresccncebased on the twisted intramolecular charge transfer emission at pyrene concentrations in the low micromolar range. (TICT) of (N,ZVdimethylamino)bnzonitrile(DMABN) and ethyl On the basis of these results, it was proposed that the excimer p(N,N-dimethylamino)benzmte(DMAEB) in supercritical fluids was a manifestation of increased solute-solute (pyrenepyrene) has shown that solute-fluid clustering can occur even at a reduced interactions. This concept has been further supported47by integral density ( p , = pcxp/pc)of O S . equation calculations based on molecular distribution functions Nonpolar polyaromatic hydrocarbons have also been used to using the correlation function for the Lennard-Jones mixture probe solute-fluid and solute-solute interactions in supercritical simulating both C02-naphthalene and C02-pyrene. These simsolvents. Brennecke et al.37.38reported on the steady-state ulations feature increased height of the solute-solute pair disfluorescence of pyrene in supercritical C02, C2H4, and CF3H. tributions near the critical point and predict increased short-range Pyrene emission exhibits several vibronic bands and is very useful solute concentration about a solute molecule. for probing solutesolvent interactions. Specifically, the intensity Our most recent w0rlP83~~ has centered on using the photophysics of the 0-0 transition (I,)is extremely solvent d e p e ~ ~ d e and n t ~ ~ ~ of ~ pyrene in pure supercritical fluids to probe the kinetics of increases with increasing solutesolvent interactions. In contrast, solute-fluid and solutesolute interactions. From steady-state and the third vibronic band intensity (I3)shows little variance between time-resolved fluorescence spectroscopy we concluded that the ~ o l v e n t s . Thus, ~ ~ - ~by ~ following Z1/13 it is possible to probe sopyrene excimer is observed a t micromolar concentrations in sulutesolvent interactions?f44 Brennecke et al. found that this ratio, percritical C02and C2H4 because of (1) increased diffusivity and as expected, increased with fluid d e n ~ i t y ; ~however, ~ > ~ * it is un(2) fluid-excimer interactions which stabilize the excimer excited usually large in the region near the critical point, indicating state. Thus, solute-fluid clusters do not affect the mechanism enhanced solutefluid interactions, i.e., clustering. More recent of excimer formation in COz and C2H4; however, once formed results by Knutson et al.4s compared the experimental clustering in the excited state, they affect the excimer decay rate.48*49In data to molecular dynamics simulations involving the first, second, contrast, pyrene excimer formation occurs in CF3H because the and third solvation shells. These results showed that the observed apparent excited-state equilibrium constant favors excimer fordensity augmentation arose only from the first solvation shell. m a t i ~ n . Here ~ ~ the recovered bimolecular rates for excimer The effects of pyrene concentration on its static emission have production are slowed appreciably (compared to that predicted also been investigated in supercritical f l ~ i d s . The ~ ~ , most ~ ~ inby diffusion control) as a consequence of enhanced solute-fluid triguing aspect of these results is the appearance of an excimer-like interaction^.^^ The bulk of the work in supercritical fluid science and technology has used COz as the solvent.la2 This is a result of its (31) Johnston, K. P.; Kim, S.; Wong, J. M. Fluid Phase Equilib. 1987, nontoxicity, low cost, nonflammability, and mild critical conditions 38. - - , 39. -( T , = 31.04 OC,Pc = 72.85 atm, pc = 0.468 g / ~ m ~ Unfor).~ (32) Hrnjez, B. J.; Yazdi, P. T.; Fox, M. A.; Johnston, K. P. J. Am. Chem. tunately, COz alone does not dissolve polar solutes well, and even SOC.1989, 111, 1915. nonpolar solutes have limited ~olubility.~ However, addition of (33) Betts, T. A.; Zagrobelny, J.; Bright, F. V. J . Supercrit. Fluids 1992, 5. 48. small amounts (1-5 mol %) of relatively volatile entrainers (co(34) Betts, T. A.; Zagrobelny, J.; Bright, F. V. J . Am. Chem. SOC.1992, solvents) increases the solvent strength of C02' but does not 114, 8163. significantly alter the critical properties 0: density compared to (35) Betts, T. A. Ph.D. Thesis, State University of New York at Buffalo, a pure Thus, entrainers can improve the selectivity (Le., 1992. the ability to extract solute in preference to another) of super(36) Sun, Y.-P.; Fox, M. A.; Johnston, K. P. J . Am. Chem. Soc. 1992,114, 1187. critical solvents. Spectroscopic methods have been used to show (37) Brennecke, J. F.; Eckert, C. A. In Supercritical Fluid Science and that these improvements can be described in terms of specific Technologv; Johnston, K. P., Penninger, J. M. L., as. ACS ; Sypm. Ser. 406; soluteentrainer interaction^.^^^^ American Chemical Society: Washington, DC, 1989; Chapter 2. In an effort to improve our understanding of solutefluid (38) Brennecke, J. F.; Tomasko, D. L.; Peshkin, J.; Eckert, C. A. Ind. Eng. Chem. Res. 1990, 29, 1682. clustering in entrainer-modified supercritical fluids, we have in(39) Brennecke, J. F.; Tomasko, D. L.; Eckert, C. A. J . Phys. Chem. 1990, vestigated their effect on pyrene excimer formation. Here we 94, 1692. report on the density and temperature dependence of pyrene (40) Ben-Amotz, D.; LaPlant, F.; Shea, D.; Gardecki, J.; List, D. In Suphotophysics in binary supercritical fluids composed of C02and percritical Fluid Technology-Theoretical and Applied Approaches in AMlytiCd Chemistry; Bright, F. V., McNally, M. E., Eds.; ACS Symp. Ser. acetonitrile or methanol. These results provide new information 488; American Chemical Society: Washington, DC, 1992; Chapter 2. on entrainers in supercritical fluid clustering. Acetonitrile and (41) Blitz, J. P.; Yonker, C. R.; Smith, R. D. J . Phys. Chem. 1989, 93, 6661. (42) Zerda, T. W.; Song, X . ; Jonas, J. Appl. Spectrosc. 1986, 40, 1194. (43) Dong, D. C.; Winnik, M. A. Can. J . Chem. 1984, 62, 2560. (44) Kalyanasundaram, K.; Thomas, J. K. J . Am. Chem. SOC.1977, 99, 2039. (45) Knutson, B. L.; Tomasko, D. L.;Eckert, C. A.; Debenedetti, P. G., Chialvo. A. A. In Supercritical Fluid Technology-Theoretical and Applied Approaches in Analytical Chemistry; Bright, F. V., McNally, M. E., Eds.; ACS Symp. Ser. 488; American Chemical Society: Washington, DC, 1992; Chanter 5. (46) Birks, J. B. Photophysics of Aromatic Molecules; Wiley-Interscience: New York, 1970.
(47) Wu, R.-S.; Lee, L. L.; Cochran, H. D. Ind. Eng. Chem. Res. 1990. 29, 971. (48) Zagrobelny, J.; Betts, T. A.: Bright, F. V. J . Am. Chem. Soc. 1992, 114, 5249. (49) Zagrobclny, J.; Bright, F. V. J . Am. Chem. SOC.1992, 114, 7821. (50) Dobbs, J. M.; Wong, J. M.; Johnston, K. P. J . Chem. Eng. Data 1986, 31, 303. (51) Dobbs, J. M.; Wong, J. M.; Lahiere, R. J.; Johnston, K. P. Ind. Eng. Chem. Res. 1987, 26, 56. (52) Dobbs, J. M.; Johnston, K. P. Ind. Eng. Chem. Res. 1987,26, 1476. (53) Lemert, R.M.; Johnston, K. P. Ind. Eng. Chem. Res. 1991,30, 1222.
Study of Matrix- Modified Supercritical COz
J. Am. Chem. Soc., Vol. 115, No. 2, 1993 103
methanol also provide a convenient means to compare aprotic and protic entrainers.
-
0.8 ..
-1
In normal liquids, the kinetic model shown in Figure 1 describes the pyrene monomer and excimer emission process.46 The species M, M*, and D*denote the ground-state monomer, excited-state monomer, and excited-state dimer (excimer), respectively. The terms kM,kD, k M D , and k D M represents the emissive rate coefficients for the monomer and excimer, the non-radiative unimolecular D* M* M (reverse) rate, and the non-radiative bimolecular M* M D* (forward) rate, respectively. The hv symbolism denotes an absorption process that populates M* only; no D exists in the ground state.46 The individual rate terms given in Figure $1are recovered by simultaneous, global analysis of multiple fluorescence decay experiment^.^^-^' That is, we analyze simultaneously the intensity decay traces from multiple wavelength and pyrene concentration experiments and recover the rate terms directly. The goodness-of-fit between the experimental data and the assumed model is judged by the reduced x2 (x2,) residuals and autocorrelation function.54 The theoretical treatment used to define the global algorithm can be found in ref 54. From Einstein-Smoluchowski diffusion theory,58 the bimolecular rate coefficient for a diffusion-controlled reaction is given by
- ++-
kDM
= 8000RT/0.37
>E
0.6--
-" cn
0.4
-.
0.2
--
v
C
Q
n
4
0.0
20
120
170
220
Pressure (bar) Figure 2. Isotherms for CO,. The area contained within the dashed box indicates the region over which experiments were carried out. 1.2
MeOH
(1)
where R is the gas constant (J mol-' K-I), T i s the temperature in Kelvin, and 7 is the fluid viscosity in poise. For pyrene excimer emission in normal liquids, the bimolecular rate ( k D M ; M* M D*)generally follows this simple diffusion model.46
-
70
+
1.2
MeCN
Experimental Section Detailed information on the high-pressure apparatus and time-resolved instrumentation can be found elsewhere.48 All steady-state fluorescence experiments were carried out using a SLM 48000 M H F spectrofluorometer (SLM Instruments) modified to accommodate the high-pressure optical In the current configuration, the temperature and pressure imprecision are 0.1 OC and 0.2 bar, respectively. At 450 W Xe-arc lamp serves as the excitation source, and monochromators are used for excitation and emission wavelength selection. All time-resolved experiments were performed using an in-house constructed N2 laser/boxcar-based system interfaced to a personal computera4* The control/acquisition BASIC software was developed and written in house. The time-resolved intensity decay data are analyzed using a commercially available software package (Globals Unlimited). Sample Preparation. SFC grade C 0 2 (
