J. Phys. Chem. 1992, 96, 931-936 hancement. This indicates that the plane of ellipticine bound to DNA must be perpendicular to the silver surface. Thus, since the complexation of ellipticine with DNA takes place through no orientational rearrangement, it is definitely concluded that the interaction of the ellipticine plane, either free or bound to DNA, takes place perpendicularly to the silver surface. Moreover, it may be assumed that its long side makes a noticeable angle relatively to the metal surface.
Conclusion Surface-enhanced Raman spectra of ellipticine and 2-Nmethylellipticinium, as well as their complexes with DNA, were easily obtained in silver colloid solutions. Indeed, once adsorbed onto the silver surface, these dyes exhibit efficient quenching of their fluorescent emission and strong enhancement of Raman scattering. The analysis of these spectra were performed on the basis of model chromophores (isoquinoline, indole, and naphthalene) and from the comparison with near-IR FT Raman spectra of solid samples. From this study the following main conclusions can be drawn. 1. The pH effect on the intensity of skeletal modes shows that the ring vibrations are very sensitive to the protonation state of
931
the molecule, thus expressing the large electronic perturbation induced by protonation at the N(2) nitrogen atom. 2. While the adsorption of ellipticine onto the silver surface gives rise to a huge enhancement of its Raman spectra, this process does not affect its interaction with DNA. 3. The orientation of ellipticine molecule relative to the silver surface is the same either for the free or for the intercalated molecule, the plane of the molecule being normal to the silver surface. Moreover, it is likely that the long side of the molecule is significantly tilted with respect to the surface. Previous studies have shown that ellipticine (and derivatives) are useful probes to analyze drug-DNA interactions. This study proved that besides the described techniques used, SERS constitutes one of the most powerful tools to detect these interactions at submicromolar drug concentration levels. Moreover, when other spectroscopic methods as sensitive as SERS (like spectrofluorimetry) are unserviceable, SERS still remains of great interest. For these reasons, SERS investigations of nonfluorescent drugs, which interact with DNA according to a nonintercalative process, are currently underway. Registry No. Ellipticine, 519-23-3; 2-methylellipticine,69467-91-0; silver, 7440-22-4.
Relaxation Studies on Model Compounds of Cyclohexyl-Based Polyacrylates Ricardo Diaz-Calleja; Evaristo Riande,**tand Julio San Romint Luboratorio de Termodinbmica, E TSII, yniversidad PolitBcnica de Valencia. 46071 Valencia, Spain, and Instituto de Ciencia y Tecnologia de Polimeros (CSIC), 28006 Madrid, Spain (Received: June 26, 1991; In Final Form: September 11, 1991)
The equilibrium and dynamic dielectric properties of cyclohexyl isobutyrate (CI) and cyclohexyl2,4-dimethylglutarate(CG), model compounds of the repeating unit and the dimer of poly(cyclohexy1acrylate) (PCA), respectively, are determined. The results show that the dipole moment at 30 OC of cyclohexyl2,4-dimethylglutarate(2.84 D) is significantly higher than the value of this quantity (2.09 D) for its aromatic counterpart, phenyl 2,4-dimethylglutarate (PG), as a consequence of the strong preference for the high polarity t*,t* conformation in the racemic diad of the former compound. The dielectric relaxation spectrum of CI presents a relatively strong secondary relaxation, in comparison with that of the compounds with higher conformational versatility CG. Other esters such as PG and m-chlorophenyl 2,4-dimethylglutarate(MCPG)also exhibit weak secondary relaxations in their dielectric spectra. Attempts are made to explain at the molecular level the weak subglass absorptions of the esters based on 2,4-dimethylglutaricacid. The a relaxations in the model compounds are interpreted in terms of the coupling model.
Introduction Whereas it is a well-recognized fact that the most prominent relaxation in amorphous polymers, the glass-rubber or a relaxation, is caused by long-range generalized motions involving rotations about skeletal bonds, the exact nature of other secondary relaxations taking place at temperatures well below the glass transition temperature, where long-range motions are frozen, still is a matter of conjecture.I4 On one hand, it has been assumed that the subglass relaxations or B processes are related to reorientations of flexible side groups in polymer chains.5,6 However, the fact that subglass relaxations also are detected in symmetrical polymers without flexible side groups seems to suggest that these relaxations are produ'd by motions of the side groups presumably coupled with motions of a few skeletal bonds of the main chain. On the other hand, the presence of B processes in small molecules such as bromobenzene, lchloronaphthalene, etc. has given support to models that assume that intermolecular interactions play an important role in the development of secondary relaxation^.^ t Universidad PolitCnica de Valqcia. f
Instituto de Ciencia y Tecnologia de Polimeras.
0022-3654/92/2096-931$03.00/0
In a recent work, the dielectric relaxation spectra of phenyl and chlorophenyl esters of poly(acry1ic acid) have been reported.' An inspection of the spectra permits to detect a subglass relaxation whose strength is weak for polymers with polar symmetric side groups such as poly(pheny1 acrylate) (PPA) and poly@-chlorophenyl acrylate) (PPCPA); however, polymers with polar asymmetric side groups, such as poly(m-chlorophenyl acrylate) (PMCPA) and poly(o-chlorophenylacrylate) (POCPA), exhibit a subglass absorption whose strength is comparable to that of the glass-rubber relaxation. These results therefore suggest that only a small part of the apparent mean-square dipole moment &*of the symmetric side groups in the molecular chains relaxes mainly (1) McCrum, N. G.; Read, B. E.; Williams, G. Anelastic and Dielectric Effects in Polymeric Solids; Wiley-Interscience: New York, 1967. (2) Williams, G . Adu. Polym. Sci. 1979, 33, 59. (3) The Glass Transition and the Nature of the Glassy State; Goldstein, M., Simha, R., Eds.; Ann. N . Y.Acad. Sci. 1976, 279, 1-246. (4) Boyd, R. H. Polymer 1985, 26, 1123. (5) Buerger, D. E.; Boyd, R. H.Macromolecules 1989, 22, 2694. (6) Bgerger, D. E.; Boyd, R. H. Macromolecules 1989, 22, 2699. (7) Diaz-Calleja, R.; Riande, E.; San RomBn, J. Macromolecules 1991, 24, 264.
0 1992 American Chemical Society
932
The Journal of Physical Chemistry, Vol. 96, No. 2, 1992
Diaz-Calleja et al.
TABLE I: Values of the Glass Transition Temperature T,for Cyclohexyl Isobutyrate (CI), Cyclohexyl 2,4-Mmethylglutarate (CC), Phenyl 2,4-Dimethylglutarate (PG), and m-Chlorophenyl 2,4-Dimethylglutarate (MCPG) compound T., "C compound T., O C CI t r g r > tO,gO, although the changes in polarity are much lower than in the other cases. The relaxation strength of the secondary relaxation may be expressed in terms of the familiar relationship2'
bonds in Figure 7a.
to be treated macroscopically, in which the molecule i is located, can be expressed as
p = pi
+ &k
Smith and Boyd's approach*22the term N be written as
(7) may
~ @ f i ~ 2
Nfispo2 = N(pi(O.P*)
(8) fkd.
Where p* is the Statistical mechanical average OfP With The second factor of the second member of this equation can be written
+ z(bi({)*bk*) k
(pi(S")'P*) = (pi({)')
(9)
with pk* being the dipole moment of pk assuming pi({) is fixed. If the intermolecular interactions are negligible, pk* will be independent of pi({) and eq 9 becomes (pi"'*) =
(p?)
+ ( h i ( { ) ) k' c ( p k * )
(10)
At temperatures above Tgr( p i ( { ) ) averages to zero, but this may not be the situation in the subglass region where overall molecular tumbling does not occur. Therefore if it is further assumed that the macroscopic region associated with P is not permanently polarized, one obtains
(p) = ( p i ( { ) ) + c ( p k ( { ) ) = 0 k
(1 1)
and eq 1 1 becomes22 where bo2is the uncorrelated mean-square dipole moment and gs a factor that accounts for the intramolecular and intermolecular correlations in the /3 process. In general, the total dielectric relaxation strength e,, - e,, is related to the reorientation of N p dipoles per unit volume over all spatial orientations in addition to the intramolecular reorientations. Consequently, the term Npgslloz in eq 6 depends on the dipolar reorientation of N identical molecules per unit of volume, in which each individual i has an instantaneous conformation tiassociated with a dipole moment p i ( { ) . The dipole moment P of a spherical region large enough (21) FrBhlic, H. Theory of Dielectrics, 2nd ed.; Oxford University Press: Oxford, U.K., 1959.
N&,#02 = N[(Pk({)')- (cLk({))'(bk({))]
(12)
The value of ( p ( { ) ) is obtained as the sum of the weighted components along the x, y , z coordinate axes corresponding to the dipole moments of all the conformations. Calculations of this kind carried out on CG gives for ( P k ( { ) 2 ) and (&({))'(&({)) in eq 12 the values of 8.2 and 5.5 D2, respectively, if the first set of statistical weights given before are used. Since N = 2N, the term gspo2= 1.35 D2, that is, about 36% of po2shoufd relax through the process if intermolecular interactions were negligible. However, the very low strength of the /3 process in CG (Figure 3) suggests that intermolecular interactions must play an important (22) Smith, G. D.; Boyd, R. H. Mucromolecules 1991, 24, 2731.
936 The Journal of Physical Chemistry, Vol. 96, No. 2, 1992
role in the development of Secondary relaxations in this compound. Williams and ceworkers2have suggested that a rigid molecular dipole may exist in a variety of environments and in a given environment may be partially relaxed by local motions in that environment, giving rise to the secondary relaxations. The fact that secondary relaxations can also be detected in simple molecules such as chloronaphthalene, chlorobenzene, etc. where intramolecular contributions to the dielectric strength of these compounds should apparently not be important, led J ~ h a rto i ~propose ~ that a number of low-density regions,statistically distributed, may exist in the glass in which dipoles are located in many possible orientations. The secondary relaxations would arise from noncooperative but highly hindered motions of some of these molecules in these local zones which are encaged by large regions where requirements of cooperative motions make relatively inmobile the arrangement of the majority of dipoles. According to this latter model, transitions of opposite polarity presumably occur in small islands of molecules of CG that produce nearly nil dielectric activity. The same explanation may be valid for FG ' and MCFG. On the other hand, the relatively low intensity exhibited by the p relaxation in poly(cyclohexy1 acrylate) probably involves rotations about C-C* bonds in such a way that only a small change in the distribution of rotational angles about C-C* bonds takes place in the process.
The u Relaxation Process The complex dielectric permittivity is related to the normalized decay function $ ( t ) of the a process by the familiar expression24
Diaz-Calleja et al. TABLE V hrameters of the Cwplisg Model for the Clrss-Rubber Relrution of Cyclobexyl Isobutyrate (CI), Cyclohexyl and 2,4Dimethylglntuate (CC), Phenyl 2,4Diwthylglutarate (E), m-Chloropbeayl2,4DimetbylgIutuate (MCPG)
T,, OC
compound
CI CG
PG MCFG 1.0
-
0.0
-
0.8
-
0.7
-
m I
_
-
;?.e iii
0.5
-
0.4
-
0.3
-
0.2
I
I -1
I
I
log
where + ( t ) is commonly expressed by the Kohlrausch-Williams-Watts (KWW) e q u a t i ~ n ~ ~ , ~ ~ $ ( t ) = exp(-t/T)T
(14)
Recent studies carried out by Ngai and co-~orkers*'.~* suggest that coupling between the relaxing species provides a time-dependent effect that slows down the relaxation already in progress due to the effect of the interactions between these species and the heat bath. The model developed by these authors gives for $(t) an expression similar to the KWW equation
+(r) = exp(-t/r*)'-" where
T*,
(15)
named the effective relaxation time, is given by T*
= [(I - n ) w ; ~ ~ ] ' / ( ' - ~ )
(16)
In this expression, w, is a characteristic time dependent on the complexity of the system and n is related to the coupling between the relaxing species; n lies in the interval 0 < n < 1, and its value is higher the larger is the coupling between the relaxing species. The primitive relaxation time 7 0 is given by a Doolittle-type equation 70 a
exp[Bo/(T- Tm)1
(17)
with T , being the temperature at which a primitive relaxation process would not take place. From eqs 15, 16, and 17, the effective relaxation time can be written as T * a exp[Bo/(l - n ) l / ( T - T,) (18) (23) Johari, G. P. Ann. N . Y.Acad. Sci. 1976, 279, 117. (24) Williams, G.; Watts, D. C.; Dev, S.B.; North, A. M.Tram. Faraday SOC.1971,67., 1223. (25) Kohirausch, R. Ann. Phys. Chem. 1847, 12 (3), 3931. (26) Williams, G.; Watts, D. C. Trans. Faraday SOC.1970.66, 80. (27) Ngai, K. L.; Rajagopal, A. K.; Teitler, S.J . Chem. Phys. 1988, 88, 5086. (28) Ngai, K. L.; Mashimo, S.; Fytas, G . Macromolecules 1988,21,3030.
& A 1 - n), K 2750 2140 3114 3050
n 0.48 0.36 0.44 0.48
-160 -100 -9 1 -9 1
%i5Jm:,
I
2
F m e 9. Master curves for the glass-rubber relaxation of cyclohexyl isobutyrate (CI) (filled symbols) and cyclohexyl2,4-dimethylglutarate (open symbols). Dotted and continuous lines were calculated from the KWW equation assuming 9 = 0.52 and 0.64, respectively.
Values of T* were obtained from the maxima of the u absorptions and their natural logarithms plotted apinst l/(T- T,), the values of T, beiig considered those for which straight lines are obtained. Values of the slopes m = Bo/(1 - n) and T , are given in Table V. Master curves for the dielectric loss of the a process were obtained by horizontal shifting of the isotherms of the normalized losses measured in the vicinity of the glass transition temperature. It can be seen in Figure 9 that the distribution of the relaxation times associated with the glass-rubber transition seems to be wider in CI than in CG, in spite of the higher conformational versatility of the latter compound. The curves were found to fit eq 15 for the values of n indicated in Table V. The parameter n decreases from 0.48 for CI to 0.36 for CG, suggesting that the complexity of the former compound in the relaxation is higher than that of the latter. It should be pointed out that, with the exception of CG, n lies in the range 0.40-0.48 as occurs in polymers, and therefore, this parameter does not seem to be very sensitive to molecular weight. It has been suggestedB that Bo is independent on the probe and that the value of n for the dipolar relaxation (ndJ is lower than or equal to that for shear compliance (n,) and this in tum is lower than or equal to the value of this quantity for density fluctuation (nb). However, the results obtained in this work suggest that the invariance of Bo does not seem to hold for the same probe in different systems. Actually, the value of this quantity rises from ca. 1370 K for CG to 2080 K for FG. However, the remarkable closeness of the values of Bo for both CI and CG suggests that the invariance of this parameter presumably holds for homologous series of molecular compounds.
Acknowledgment. This work was supported by the CICYT through Grants PB-89-0069 and MAT 88-0555. Registry No. CI, 1129-47-1; CG, 137768-94-6; PCA (homopolymer), 27458-65-1.