J. Phys. Chem. 1988, 92, 6140-6144
6140
for the Di. (see footnotes to Tables IX and X), and 5% errors were assumed for the activity derivatives required for these calculation^.^ The results at all of our concentrations obey the ORR within these assumed errors except for ZT = 0.5 m ~ b d m with - ~ wI = 1/3. The resulting error limits are small at high concentrations but become Because there are relatively quite large at ZT = 0.5 ~nol-dm-~. fewer osmotic and activity data at lower concentrations, the parameters of the Scatchard equation are less well determined, and thus the activity derivatives are more uncertain for that region. Consequently, we intend to measure additional low-concentration
isopiestic data before reporting L, and Li: values and formally testing the ORR. This should then reduce the large uncertainty in these calculations for low concentrations. Acknowledgment. This research was performed under the auspices of the Office of Basic Energy Sciences (Geosciences) of the U S . Department of Energy by Lawrence Livermore National Laboratory under Contract No. W-7405-ENG-48. We thank Athena Kyler for the word processing of this paper. Registry No. NaC1, 7647- 14-5; SrCI2, 10476-85-4.
Association of I nciusion Compounds in the Systems of ,&Cyclodextrin-Aniline-Sodium 1-Pyrenesulfonate and -Pyrene Sanyo Hamai Department of Physics, Miyazaki Medical College, Kiyotake. Miyazaki, 889- 16, Japan (Received: December 1, 1987; In Final Form: April 4, 1988)
-
Sodium I-pyrenesulfonate (PS) in aqueous solution forms a 1: 1 inclusion compound with P-cyclodextrin (CDx). When aniline (A) is added below 1 X loe2mol dm-3, the absorption maxima of a PS-CDx solution are red-shifted accompanied by an appearance of isosbestic points. In the presence of CDx the fluorescence quenching of PS by A takes place at a low A concentration. The fluorescence quenching is caused by the formation of a ternary inclusion compound PAC which is formed by incorporation of both one PS molecule and one A molecule into one CDx cavity. Addition of more than 1 X mol dm-3 A to the PS-CDx solution results in the appearance of a longer wavelength fluorescencefrom an electron donor-acceptor complex of PS with A. Comparisons have been made between the observed intensity data of this charge-transfer (CT) fluorescence and the calculated concentration curve of candidate species against the A concentration. This procedure has revealed that the CT fluorescence is due to an inclusion compound PAC-AC which is formed by an association of two different kinds of inclusion compounds, PAC and AC (1:l inclusion compound of A with CDx). In a CDx-A-pyrene system, it has been found that pyrene shows a behavior similar to that of PS. We have determined the equilibrium constants for the formation of PAC and PAC-AC and those of Py inclusion compounds corresponding to PAC and PACmAC.
-
Introduction Previously, we have reported that the same or different kinds of two 1:l inclusion compounds of /3-cyclodextrin (CDx) associate together to form a dimeric or a ternary inclusion compound in aqueous solution.' In a CDx-naphthalene system, the observation of excimer fluorescence of naphthalene has provided evidence for the association of the two 1:l inclusion compounds of CDx with naphthalene. In the case of a CDx-2-methoxynaphthalene-1,2dicyanoknzene system, CDx forms 1:1 inclusion compounds with 2-methoxynaphthalene and 1,2-dicyanobenzene, respectively. The charge-transfer absorption and fluorescence have been observed for this system, indicating that the above two different kinds of the 1:1 inclusion compounds can associate together. Dimers of 1: 1 inclusion compounds have been reported for other CDx complexes: a CDx-2-anthracenesulfonate system by Tamaki and Kokubu from a fluorescence study,2 a CDx-N-substituted phenothiazine system by Otagiri et al. from an N M R study,3 and a CDx-cyclobuta [ 1,2-b:3,4-b1diquinoxalinesystem by Yamaguchi from an induced circular dichroism s t ~ d y .Miyajima ~ et al. have suggested that a- and y-cyclodextrins themselves dimerize in aqueous s o l ~ t i o n . ~ X-ray studies on inclusion compounds in the crystalline state show the existence of dimeric CDx complexes constructed from a 1:1 CDx-ethyl cinnamate, a 1 :1 CDx-p-iodophenol, and a 1:1 CDx-2,5-diiodobenzoic acid complex, et^.^-^ These X-ray Hamai, S. Bull. Chem. SOC.Jpn. 1982, 55, 2721. Tamaki, T.; Kokubu, T. J. Inclusion Phenom. 1984, 2, 815. Otagiri, M.;Uekama, K.; Ikeda,K. Chem. Pharm. Bull. 1975,23, 188. Yamaguchi, H. J. Inclusion Phenom. 1984, 2, 747. Miyajima, K.; Sawada, M.; Nakagaki, M. Bull. Chem. SOC.Jpn. 1983, 56, 3556. (6) Hursthouse, M. B.; Smith, C. Z.; Thornton-Pett, M.; Utley, J. H. P. J . Chem. SOC.,Chem. Commun. 1982, 881. (1) (2) (3) (4) (5)
0022-3654/88/2092-6140$01.50/0
analyses in the solid state seem to foresee a possibility that an association of two 1:l inclusion compounds takes place fairly widely in aqueous solutions as well as in crystals. In the present investigation, the systems of CDx-aniline-sodium 1-pyrenesulfonate and -pyrene were studied through electronic absorption and fluorescence measurements.
Experimental Section Sodium 1-pyrenesulfonate (PS) purchased from Molecular Probes Inc. was used as received. Pyrene (Tokyo Kasei) was purified by column chromatography. /3-Cyclodextrin (Nakarai) was recrystallized three times from water.' Aniline (Wako) was distilled under reduced pressure. Doubly distilled water containing purified pyrene crystals was allowed to stand in the dark for several days, and its supernatant part was used for sample solutions of pyrene. The concentration of pyrene was about (2-3) X lo-' mol dmd3. PS concentrations for absorption and fluorescence studies and 1 X mol dm-3, respectively. Aerated were 2.5 X sample solutions were employed throughout this work. All the measurements were carried out at 5 i 0.1 OC. Absorption spectra were recorded on a Shimadzu UV-260 spectrophotometer. For the absorption measurements of pyrene solutions, an accumulation of spectral data, which were obtained at a slow scan speed, was repeated 20 or 30 times in order to improve a signal-to-noise ratio of the absorption spectrum. Fluorescence spectra were taken on a Shimadzu RF-501 spectrofluorometer equipped with a cooled Hamamatsu R-943 pho(7) Hamilton, J. A.; Sabesan, M. N.; Steinrauf, L. K.; Geddes, A. Biochem. Biophys. Res. Commun. 1976, 73, 659. (8) Stezowski, J. J.; Jogun, K. H.; Eckle, E.; Bartels, K. Nature (London) 1978, 274, 617. (9) Harding, M. M.; Maclennan, J. M.; Paton, R. M. Nafure (London) 1978, 274, 621.
0 1988 American Chemical Society
Association of Inclusion Compounds
The Journal of Physical Chemistry, Vol. 92, No. 21, 1988 6141
1.0
,
h/nm
400 I
350 ,
I
28
30
32
26
t / 1 0 3 crn-‘
Figure 2. Absorption spectra of PS (2.5 X mol dm-’) in aqueous solution containing a fixed concentration of CDx (8.4 X lo-’ mol dm”) and varying concentrations of A. Concentration of A: (1) 0 mol dm-’; (2) 1.10 X lo-’ mol dm-’; (3) 2.20 X lo-’ mol dm-’; (4) 5.49 X lo-’ mol dm-’; (5) 1.10 X mol dm-’; (6) 4.39 X mol dm-’.
0
0.5
1.0
Figure 3. Continuous variation plot monitored a t 346 nm. The PS concentration is 2.5 X mol dm”. [CDxIo [AIo = 3.0 X lo-’ mol dm-’.
+
bands with a disappearance of the isosbestic points. The spectral changes shown in Figure 2 indicate the existence of at least two kinds of inclusion species of PS. To examine whether the inclusion compound in a low concentration range of A (below 1 X mol dm-3) includes only one PS molecule or not, the PS conto 2 X lo4 mol dm-3 keeping centration was varied from 1 X and 4.39 X mol the CDx and A concentrations at 8.4 X dm-3, respectively. The absorbance at any wavelength was proportional to the PS concentration, indicating that only one PS molecule is included in the inclusion compound. When the sum of the added concentrations of CDx and A is small, the concentrations of PC and free PS are little affected by the formation of the inclusion compound responsible for the spectral change in Figure 2. As a consequence, a continuous variation method, which is usually utilized for two-component systems, can be applied to our three-component system in order to decide a stoichiometry concerning CDx and A. Upon the addition of CDx to a PS solution without A, the absorbance varies except for that at isosbestic points. Thus, 346 nm, at which an isosbestic point appears, has been selected as a monitoring wavelength. Under the conditions of [CDx], [A], = 3 X mol dm”, the differences from the absorbance at [CDx], = 0 mol dm-3 are plotted as a function of the CDx molar fraction (Figure 3). Here, [CDx], and [A], are initial concentrations of CDx and A, respectively. The difference in the absorbance goes through a minimum value at a CDx molar fraction of 0.5, and this result provides clear evidence for the inclusion compound of which the molar ratio of CDx to A is unity. Because a bulky PS molecule does not fully
-
+
(10) Hamai, S. Bull. Chem. Soc. Jpn. 1987, 60, 3505. (11) Harada, A.; Nozakura, S . Polym. J . 1982, 8, 141. (12) Klein, U. K. A.; Miller, D. J.; Hauser, M. Spectrochim. Acto, Part A 1976, 32A, 319. (13) Kobashi, H.; Takahashi, M.; Muramatsu, Y.;Morita, T. Bull. Chem. Soc. Jpn. 1981, 54, 2815. (14) Hoshino, M.; Imamura, M.; Ikehara, K.; Harna, Y . J . Phys. Chem. 1981, 85, 1820.
Hamai
The Journal of Physical Chemistry, Vol. 92, No. 21, 1988
6142
-2
A weak fluorescence from an exciplex between PS and A has been observed in organic solvents. In highly polar solvents, it is well-known that no exciplex emission is, in general, observed owing to a direct electron transfer without exciplex formation.l%z’ The result that PAC does not emit an exciplex or a C T fluorescence can be interpreted in terms of the relatively polar environment around a PS molecule which is incorporated into the CDx cavity. This conclusion is compatible with a picture of PAC in which an A molecule locates in the immediate neighborhood of a PS molecule inside the CDx cavity and in which a part of PS molecule extrudes from the CDx cavity into a water medium. As already mentioned, the absorption maxima of a PS solution containing CDx are further red-shifted accompanied by a vanishing of the isosbestic points when the concentration of A is raised above -1 X mol dm-3 (Figure 2). This result shows that the other inclusion compound of PS appears at high concentrations of A. The concentration dependence of the PS absorbance was investigated to decide whether the second PS inclusion compound includes more than one PS molecule or not. The apparent absorption coefficient at any wavelength was independent of the PS concentration from to 2 X lo4 mol dm-3, indicating that the second inclusion compound also includes only one PS molecule. This finding shows that neither association between two PACs nor that between PAC and PC occurs. The spectral change above 1 X lo-* mol dm-3 of A shown in Figure 2 is, therefore, due to the interaction of a PS molecule included in PAC with a species that does not contain a PS molecule. As such a species there are three candidates, A, AC, and CDx which all exist in the CDxA-PS system. Because the CDx cavity in PAC is fully filled with a PS and an A molecule, it is quite unlikely that CDx accommodating both PS and A further includes an additional A molecule in its cavity. However, there is a possibility that an additional A molecule associates with a CDx molecule including PS and A from outside of CDx through hydrogen bonding. Because the hydrogen-bonding interaction between the A and CDx molecule does not induce the absorption spectral change of PS, the above possibility can be excluded. Consequently, either CDx or AC is the species which can associate with PAC. Such an aggregation is shown by
0.04
1
1”’-
1
0 ‘ 0
3
2
1
[A],/
5
4
ld3mol
I 6
dm‘3
Figure 4. Plot of 1IIAP.S) against [A],. [PSIo= 2.5 [CDx],, = 8.4 X lo-’ mol dm-’.
X
mol dm-3.
occupy the CDx cavity, an A molecule can successfully enter the CDx cavity which already accommodates one PS molecule. From the above results, the absorption change in the low concentration range of A shown in Figure 2 can be attributed to the formation of a ternary 1 :1:1 inclusion compound PAC which is composed of one CDx, one PS, and one A molecule. The equilibria are represented by
PC+A&PAC
PS
-
(3)
+ AC & PAC
(4) where K3 and K4 are equilibrium constants for the formation of PAC from PC and that from PS, respectively. On the basis of an investigation of quenching of the PS fluorescence by A, Kobashi et al. have pointed out a possibility that a 1:l:l complex is formed among CDx, PS, and A.I3 In addition, from detailed studies on quenching and lifetime of the PS or pyrene fluorescence, Kano et al. have reported that a three-component complex is formed in CDx-aliphatic amine-PS or -pyrene s y ~ t e m s . l ~ ,Our ’ ~ assignment that PAC is responsible for the spectral change in Figure 2 ([A], < 1 X mol is consistent with their results. At low A concentrations, the fluorescence of a PS solution containing CDx is strongly quenched without shifts of the fluorescence peaks, indicating that PAC is nonfluorescent. In such a case, one can evaluate a K3 value according to the equation which has been derived in the previous paper:I
-
~ 1 +0 Kz[CDXlo) ~ ~ 1 + 0 / ~ (1 + K,[CDxlO)/a[P~lo (5) where Zf(PS) is the fluorescence intensity of a PS solution containing both CDx and A, a is a constant, and [PSIois an initial concentration of PS. In Figure 4, the data of l/ZdPS) are plotted as a function of [AIo, and the initial slope and the intercept yield a value of 2000 mol-’ dm3 as 4. In the absence of CDx the Stern-Volmer constant for fluorescence quenching of PS by A is 210 mol-l dm3. Since a guest molecule in the CDx cavity is effectively protected by CDx from a q ~ e n c h e r , ’we , ~ neglect the dynamic quenching of PS inside the cavity. Dynamic quenching of free PS by A in aqueous phase is calculated to be about 15% of the total fluorescence quenching at an A concentration of 1 X l U 3 mol dm-3, and the residual quenching comes from the static quenching caused by the formation of PAC1’ Consequently, an upward curvature in Figure 4 can be explained in terms of the dynamic quenching of free PS by A.’* Thus the K3 value of 2000 mol-l dm3 obtained above may be slightly overestimated. From a relationship of KlK3 = K2K4, K4 = 2000 mol-l dm3 was evaluated.
l/If(PSl = ~
I
~
3
~
~
~
(15) Kano, K.; Takenoshita, I.; Ogawa, T. Chem. Lett. 1980, 1035. (16) Kano, K.; Takenoshita, I.; Ogawa, T. J . Phys. Chem. 1982,86, 1833. (!7) In the same A concentration range as that shown in Figure 4, there are isosbestic points in the absorption spectra in Figure 2. Consequently, the further aggregation of PAC can be ruled out. (18) From the fluorescence lifetime measurements in the A concentration range from 1 X IO” to 6 X mol dm”, Kobashi et al. (ref 13) have revealed that the dynamic quenching of PS occurs in a PS solution containing both CDx and A at 22 O C .
PAC
+ AC 2PAC-AC
(6)
PAC
+ CDx
(7)
or ~
~
~
1
0
~
~
PACZ
where PAC-AC and PAC-C stand for an inclusion compound formed between PAC and AC and that between PAC and CDx, respectively. When A higher than 1 X loW2mol dm-3 is added to a PS solution containing 8.4 X mol dm-3 CDx, a new broad emission with a maximum at 540 nm appears at longer wavelengths than the PC (PS) fluorescence. Because the excitation spectrum of the broad emission is red-shifted compared with the absorption band of PAC, the new emission does not arise from PAC. As noted above, the fluorescence from the PS-A exciplex has been observed in organic solvents. Combined with the absorption spectral change shown in Figure 2, we assign the broad emission to the CT (charge-transfer) fluorescence from the EDA (electron donor-acceptor) complex of PS with A. Similarly, a C T fluorescence has been observed in the system of CDx-2methoxynaphthalene-1,2-dicyanobenzene although an exciplex fluorescence rather than a C T fluorescence is usually observed in organic solvents.’ These results exhibit that in some cases an EDA complex can be easily formed inside the CDx cavity (cavities) even when an EDA complex is not formed in organic solvents. The observation of the C T fluorescence in the CDx-A-PS system
-
(19) Knibbe, H.; Rollig, K.; Schafer, F. P.; Weller, A. J . Chem. Phys. 1967, 47, 1184. (20) Mataga, N.; Okada, T.; Yamamoto, N. Chem. Phys. Left. 1967, 1 ,
..,.
119
(21) Hirata, Y.; Kanda, Y.; Mataga, N. J . Phys. Chern. 1983, 87, 1659.
Association of Inclusion Compounds I
The Journal of Physical Chemistry, Vol. 92, No. 21, 1988 6143 I
I
0
I
1
1
I
I
I
I
1
2
3
4
[A],/lO-*
mol dm-3
Figure 5. Comparison of calculated concentration curves of PAGAC with observed data of IACT). Concentration curves and IACT) data are normalized at [A], = 0.0439 mol dm-'. Assumed value of K5:(1) 200 mol-' dm'; (2)300 mol-I dm3;(3) 400 mol-' dm'; (4)500 mol-' dm'; ( 5 ) 600 mol-' dm3.
implies that both a PS and an A molecule in PAC-AC or P A C C are shielded from a very polar medium of water to a greater extent than those in PAC and that as a result the EDA complex of PS with A experiences a less polar environment. Under our experimental conditions, the C T fluorescence intensity (If(CT)) is proportional to the concentration of PAC-AC or P A C C By use of an assumed value of K5, the concentration of PAC.AC can be calculated from the equations [PAC-AC] = KiK3K,[PS]o[A] [AC] [CDx]/(l Ki[CDx] X (1 -k K3[A1(1 + K,[AC]))) (8)
+
[AI = [Alo- tAC1
(9)
(XZ - 4K22[A]o[CD~],)'/2)/(2KJ (10) [CDX] = [ C D X ]-~ [AC] (11)
[AC] = ( X -
where
X = 1 + Kz([A]o + [ C D X ] ~ )
(12)
Figure 5 depicts the dependence of the calculated concentration of PACeAC on [CDxIo together with the data of the observed IACT). The best fit of the calculated curve for [PACeAC] to the observed IACT) is seen in Figure 5 when K5 is equal to 400 mol-' dm3. On the other hand, the [PACC] curves (not shown), which have been calculated by use of assumed K6 values, never fit the data points over the K6 range from 1.0 to 10000 mol-' dm3. Therefore, it is concluded that PACeAC is the inclusion compound responsible for the C T fluorescence. From the dependence of the If(CT) on [CDxIo as well as that on [AIo, we could further estimate a K5 value by using a similar procedure. This curve-fitting procedure gave K5 = 300 mol-' dm3, which was in good agreement with the K5value (400 mol-' dm3) obtained from the dependence of If(CT) on [AIo. This provides additional evidence for the existence of PAC-AC. On the other hand, K6 determined from the dependence of If(CT) on [CDxIo was 200 mol-' dm3. From the inconsistent result for K6 obtained from the different two methods, it was further confirmed that the species responsible for the C T fluorescence is not P A C C In a CDx-trimethylamine (TMA)-naphthalene system, on the other hand, Kano et al. have suggested a possibility that 1:l:l complexes composed of CDx, TMA, and naphthalene associate to form aggregates or that a 2: 1:l complex of CDx-TMA-naphthalene is formed.22 To estimate an environmental polarity around the EDA complex of PS with A in PACsAC, the peak wavelength of the C T fluorescence can be employed as a measure of the surrounding microscopic polarity. The fluorescence maximum (A,) of the PS-A exciplex is at 588 nm (17 010 cm-') and 592 nm (16 890 cm-') in 1-octanol and 1-pentanol, respectively. Since anionic PS of PS in nonpolar is insoluble in nonpolar organic solvents, A, solvents has been estimated as follows. In 1-octanol and l-pen(22) Kano, K.; Hashimoto, S.; Imai, A,; Ogawa, T. J. Inclusion Phenom. 1984, 2, 737.
tanol, A, of an exciplex of pyrene with A has been determined to be 535 nm (18 690 cm-') and 540 nm (18 520 cm-I), respecbetween pyrene and PS is tively. The energy difference in A, 1680 and 1630 cm-' in 1-octanol and 1-pentanol, respectively, and their average value of 1660 cm-' was used as the energy difference in A,, between pyrene and PS. Thus, we could estimate A, of PS in a nonpolar solvent by subtracting 1660 cm-' from A, of pyrene in the same solvent. By use of this procedure, A-, (pyrene) = 484, 485, and 503 nm in 1,4-dioxane, diethyl ether, and ethyl acetate gave ,A, (PS) = 526, 527, and 549 nm in the respective solvents. The ,A, value of PS, 549 nm, thus estimated in ethyl acetate is close to the peak wavelength (540 nm) of the C T fluorescence. Consequently, the dielectric constant of the environment around the EDA complex in PAC-AC is nearly the same as that of ethyl acetate (e = 6.02). On the other hand, in the system of CDx-2-methoxynaphthalene-1,2-dicyanobenzene, the polarity around an EDA complex of 2-methoxynaphthalene with 1,2-dicyanobenzene closely resembles the polarity in neat 1,Cdioxane (e = 2.21).' Such a difference in the polarity is due most probably to the different extent of the covering the guest molecules by two CDx molecules. In addition, the coexistence of two A molecules in PAC-AC may contribute to an increase in the microscopic polarity around the EDA complex which is composed of one PS and one A molecule since A is a polar molecule. Fluorescence Lifetimes of PS Solutions with CDx and with Both CDx and A . In order to further characterize the properties of inclusion compounds of PS, we measured fluorescence lifetimes of PS in water, a PS solution containing CDx, and that containing both CDx and A. The fluorescence lifetime of PS in water was found to be 65 ns at 5 "C. Despite the different temperatures, this value is in good agreement with a reported value of 65 ns at mol dm"), our 25 OC.16 For a PS solution with CDx (8.4 X data of the fluorescence decay could be analyzed practically as a monoexponential decay curve with a lifetime of 72 ns probably because the difference in the lifetime between PS and PC is small. This finding for the fluorescence decay of the PS solution with CDx is identical with the result obtained by Kano et a1.I6 The C T fluorescence decay of PAC-AC was monoexponential with a lifetime of 8.8 ns, indicating that only one species emits the C T fluorescence. A species responsible for the C T fluorescence can be attributed to a bimolecular excited complex which is composed of one PS and one A molecule in PAC-AC. As an alternative candidate, there is a termolecular excited complex formed among one PS molecule and two A molecules. However, we favor the former complex because the space of the CDx cavity is not wide enough for A electrons of one PS and two A molecules to overlap each other and because the fluorescence from the PS-A exciplex of a 1:l stoichiometry has been observed in organic solvents. Inclusion Compounds in the System of CDx-A-Pyrene ( P y ) . Py as well as PS is known to form a 1:1 inclusion compound PyC with The absorption maxima of Py are slightly redshifted upon adding CDx. As revealed by Nakajima26and Kalyanasundaram and Thomas,27the intensity of the vibronic bands of the Py fluorescence is very sensitive to solvents. The first vibronic band at 373 nm is intensified on going from nonpolar to polar solvents, while the third band at 383 nm decreases in intensity. The fluorescence spectrum of a Py solution containing CDx is analogous to that in relatively nonpolar solvents, indicating that the environment surrounding a Py molecule in PyC is less polar compared with that in bulk water. The Py fluorescence observed in aqueous CDx solution is essentially the same as that reported by Nakajima.23 An equilibrium constant Kl(py) for the formation of PyC, which corresponds to K 1 of PS,was determined to be 56 mol-' dm3 from the integrated intensity change in the Py fluorescence. A change in the absorption spectra of Py similar (23) Nakajima, A. Spectrochim. Acta, Part A 1983, 3 9 4 913. (24) Hashimoto, S.; Thomas, J. K. J. Am. Chem. SOC.1985, 107,4655. (25) Kusumoto, Y. Chem. Phys. Lett. 1987, 136, 535. (26) Nakajima, A. Bull. Chem. Soc. Jpn. 1971, 44, 3272. (27) Kalyanasundaram K.; Thomas, J. K. J . Am. Chem. SOC.1977, 99, 2039.
6144
J . Phys. Chem. 1988, 92, 6144-6148
to that of PS is observed when A is added to Py aqueous solutions containing CDx. Therefore, it is concluded that the equilibria corresponding to those for PS exist in the Py solutions containing both CDx and A. In fact, at low concentrations of A, the Py fluorescence is quenched without a change in the fluorescence spectrum, indicating the existence of a ternary inclusion compound PyAC which is formed from PyC and A. On the other hand, at high A concentrations a broad structureless emission appears with a maximum at 480 nm. The broad emission can be assigned to the C T fluorescence from an inclusion compound PyAGAC which is formed by the association of two different kinds of inclusion compounds PyAC and A C 2 * The peak position of the fluorescence from an exciplex between Py and A has been found to locate at 484 nm in 1,Cdioxane. This maximum wavelength is very close to that of the C T fluorescence from PyAC-AC. This finding exhibits that the polarity around an emissive species in PyAC-AC is nearly identical with that in neat 1,Cdioxane. With respect to the polarity around a species responsible for a C T fluorescence, the same result has already been pointed out for the system of
CDx-2-methoxynaphthalene-1,2-dicyanobenzene.1 (28) The Py excimer fluoresces with almost the same peak maximum. As mentioned below in the text, however, a short lifetime, 29 ns, of the broad emission suggests that the broad emission band is not due to the Py excimer.
The CT fluorescence from PyAC-AC decays monoexponentially like that from PACsAC: A lifetime of 29 ns has been obtained for the C T fluorescence from PyAC-AC. As in the case of PAC-AC, the species that emits the CT fluorescence in PyAC-AC is attributed to a binary excited complex between Py and A. By use of the same methods as those employed for the case of PS, equilibrium constants for Py, K,(Py), K4(Py), and K5(Py), which respectively correspond to K3, K4, and K5 for PS, were determined to be 66 000, 54 000, and 20 000 mol-' dm3, respectively. These values are about 30-50 times larger than those for PS. Such large differences in the equilibrium constants between Py and PS are due mainly to the steric hindrance of a sulfonato group of PS to the formation of each inclusion compound. In addition, the hydrophilic sulfonato group of PS appears to diminish the relevant equilibrium constants. Although there are significant differences in the relevant equilibrium constants between Py and PS, it should be emphasized that the same type of association occurs for parent Py and PS which posesses the bulky sulfonato group. Acknowledgment. I thank Professor Fumio Hirayama for his valuable discussion. Registry No. PS, 59323-54-5; CDx, 7585-39-9; A, 62-53-3; pyrene, 129-00-0.
A New Equation for Calculating Partial Cohesion Parameters of Solid Substances from Solubilities D. Reuteler-Faoro, P. Ruelle, H6 Nam-Trdn, C. de Reyff,? M. Buchmann, J. C. Ntgre, and U. W. Kesselring* Institut d'Analyse Pharmaceutique, Ecole de Pharmacie, UniversitP de Lausanne, CH- I005 Lausanne, Switzerland, and Laboratoire de Chimie Technique, Ecole Polytechnique FPdSrale de Lausanne, CH- 101 5 Lausanne, Switzerland (Received: December 14, 1987; In Final Form: April 4, 1988)
In this paper, a new expression of the Gibbs free energy of solution is suggested; it stems from general thermodynamics of mixing as well as from the works of Scatchard and Prentiss and from those of Srebrenik and Cohen. Its partial differentiation versus the number of moles of solid solute permits deduction of the chemical potential of the solute and hence of the solid. Since the chemical potentials of the same solute saturating two different solvents are together identical and equal to the chemical potential of the solid, a new relationship yielding the partial cohesion parameters of the solids is readily obtained. Proper use of this thermodynamic model implies that solubility equilibrium has been attained, that concentration at saturation is low, that no complex formations occur in solution, and finally that the geometric mean law is applicable to specific interactions. The partial cohesion parameters [MPa112]of phenylbutazone determined with this new model (6, = 24.5; 6, = 9.2; Bh = 7.6; 6, = 25.2) have been compared to those obtained by gas-solid chromatography.
Introduction The cohesion parameter,l initially named the solubility parameter 6 [MPa1/2] by Hildebrand and Scott,2 is an intrinsic parameter reflecting the intermolecular interaction capacities of liquids 6 = [(Av,$P
- RT)/V']'/'
(1)
where A,,#' [J mol-'] is the molar enthalpy of vaporization of the pure liquid at temperature T , V' [cm3 mol-'] is its molar volume a t the same temperature, R [8.3143 J K-' mol-'] is the perfect gas constant, and T [K] is the absolute temperature. To account for the different types of interactions which exist in real solutions, the total cohesion parameter has been split by Hansen3 into three components (ad, 6,, 6h) and into five components (&, ,a ai, 6,, 6,) by Karger, Keller, Snyder, and Owing to the nonsymmetrical nature of donor-acceptor interactions, the *To whom correspondence should be addressed. Ecole Polytechnique Fcdtrale de Lausanne.
0022-3654/88/2092-6144$01 SO/O
use of two acid-base partial cohesion parameters (a, 6,) to describe both proton acceptor and proton donor properties instead of one single parameter (6,) is normally necessary for a reliable quantitative evaluation of the behavior of multicomponent systems. ) Unfortunately, Lewis acid (6,) and Lewis base ( ~ 3 ~cohesion parameter values only exist for very few solvents, and even when they do exist, they do not result from independent determinations. Therefore, Hansen's three-component splitting was adopted in this work:
(1) Barton, A. F. M. Pure Appl. Chem. 1985, 51, 905. (2) Hildebrand. J. H.; Prausnitz, J. M.; Scott, R. L.Regular and Related Solutions; Van Nostrand Rheinold: New York, 1970. (3) Hansen, C. M. I d . Eng. Chem. Prod. Res. Deu. 1969, 8, 2. (4) Karger, B. L.; Snyder, L. R.; Eon, C. J. Chromatogr. 1976, 125, 71. (5) Karger, B. L.; Snyder, L. R.; Eon, C. Anal. Chem. 1978, 50, 2126. (6) Kelier, R. A.; Karger, B. L.; Snyder, L. R. Cas Chromatogr. Proc. Int. Symp. (Eur.) 1970, 8, 125. (7) Keller, R. A,; Snyder, L. R. J . Chromatogr. Sci. 1971, 9, 346.
0 1988 American Chemical Society
