2339
Enthalpies of Ion Pairing for Substituted Nitrobenzene
Equilibrium Studies by Electron Spin Resonance. 1V. Enthalpies of Ion Pairing for Substituted Nitrobenzene Anion Radicals Gerald R. Stevenson* and Luis Echegoyen Chemistry Department, University of Puerto Rico, Rio Piedras, Puerto Rico 00931 (Received April 11, 7973) Publication costs assisted b y the University of Puerto RiCO
The anion radicals of several substituted nitrobenzenes have been prepared by alkali metal reduction in hexamethylphosphoramide. For the systems in which the u+ value of the para Substituent is less than 0.7, the esr spectra for the “free” ion and ion pair were observed simultaneously. A F for the reaction between the ion pair and “free” ion was determined from temperature-dependent studies. AW was found to correlate linearly with the u+ values of the para substituents. The negative value of p indicates that the enthalpy of the dissociation of the ion pair to form the “free” ion increases with the electron-withdrawing ability of the para substituent. This is due to the fact that there is more ordering of the solvent by weaker ion pairs.
metal in HMPA. The coupling constants for the anion radicals observed are given in Table I. Only for the compounds I-VIII, which have a + values between 0.7 and -0.3, was it possible to observe the “free” ion and the ion pair simultaneous!y, Figure 1. For the para-substituted PhN02-HMPA-Li systems where the para substituent has a u+ value between 0.3 and -0.3 only the “free” ion could be observed at -lo”, and only the ion pair could be observed at high temperatures, about 80”. The nitrogen hyperfine coupling constants, A N , for the “free” anion radicals are insensitive to temperature changes and have the values given in Table I, while those for the ion pairs always increase with increasing temperature. At intermediate temperatures, relative esr line intensities are taken from the line height multiplied by the extrema to extrema line width squared. Plots of In (CY)/(@) us. 1/RT for all of these systems yield straight lines with slopes that are independent of the concentrations of anion radical, Figure 2. These simple modified van’t Hoff plots yield slopes of --ALP for the ion /3 CY M+ where /3 = ion pair (1) pair dissociation reaction, eq 1. The enthalpy of equilibrimuch larger than the concentration of the “free” ion (a), um 1 varies greatly with the para substituent. Subjecting a plot of simply In (CY)/@) us. 1/RT yields a straight line these data to a Hammett type correlation, we have obwith a slope of - ANO of dissociation of the ion pair.3 tained a reasonably linear relationship between u+ values Esr has previousiy been used to determine the Hamfor the para substituent and A T of the dissociation reacmett-Streitwieser correlation constants relating the coution, Figure 3. The slope of the line resulting from a plot pling constants for substituted PhN02 anion radicals with of AH” us. a+ is taken to be -RTp. A t 25” p has a value of u values.5 Krygowski, et a1.,6 have found that a linear cor-23 f 4. The error represents the standard deviation relation is obtained between the half-wave potential for taken from a computer analysis of the best slope. A more the polarographic reduction of a series of nitroaromatic accurate determination of p would necessitate the use of compounds and ur (the Streitwieser position constant) .7 more para-substituted nitrobenzenes. However, the choice However, there are no reports of a Hammett correlation of para substituents is severely limited in that the anion with ion pairiing equilibria. This correlation would be imradical of the compound must be thermally stable, the portant to provide information as to the variation of the anion radical must allow simultaneous observation of the structure of ion pairs with changes in the charge density ion pair and the “free” ion, and the two superimposed for the ion pair. Here we wish to report the enthalpies of spectra must be simple enough for analysis. ion pair dissociation for some para-substituted PhNQz Compounds IX-XI gave only the “free” ion upon reducanion radicals and a positive correlation of these enthaltion with Li in HMPA. For the case of p-dinitrobenzene pies to Brown’s U + values. the two nitrogen coupling constants are identical with those of the protons yielding an esr spectrum consisting of Results nine equally spaced lines. The intensities, 0.95.1: Para-substituted nitrobenzenes (I-XII) have been re15:30: 36:30: 15.5:5.1:0.9, compared to the theoretical duced to their respective anion radicals with lithium values, 1:6:17:30:36:30:17:6: 1,give excellent agreement.
To date there have been two reports of thermodynamic parameters controlling equilibria between a simultaneously observed anion radical ion pair and free ion.f,2 Allendoerfer and Papezl have determined the enthalpy and other thermodynamic parameters for the equilibrium between the ion pair and free ion of alkali metal durosemiquinone solutions. More recently we have reported the simultaneous observation of the “free” ion and ion pair of nitrobenzene2 ( P h N 0 2 ) and the enthalpies of ion pairing for the PhNO2-hexamethylphosphoramide (HMPA) metal systeme.3 PhN02 reduced by lithium metal in HMPA yields a solution that not only contains the PhNO2 “free” lon and .~ ion pair, but it also contains the PhNOz d i a n i ~ n This fact makes it difficult to obtain the metal concentration for the metal that is not involved in ion pairing. This metal concentration is necessary in order to obtain the true thermodynamic equilibrium constant for the ion pair dissociation, eq 1. However, since the M+ concentration is
+
The Journal of Physical Chemistry, Voi 77, No i 9 , 1973
Gerald R. Stevenson and Luis Echegoyen
2340
0
..
ii
-10
-1s
Figure 1. Low-field half of the esr spectrum for t h e system pchloronitrobenzene-HMPA-Li at 25”. The .arrows mark the first three lines of the ion pair. At lower temperatures only t h e “free” ion is observed, and at high temperatures only the ion pair is
Figure 3. A plot of A H ” of dissociation of the ion pair to form the “free” ion vs. the d value of the substituents I-VII. p taken from this plot is -23.
observed.
I
1 C’IRI
1.9
i
/
Figure 4. Esr spectra of the PhNQ2-HMPA-Na system as a function of temperature: (A) -IO”, “free” ion; (8)20”, “free” ion and ion pair simultaneously; (C) 70°, ion pair. Figure 2. A plot of In ( C Y ) / ( & pylnitrobenzene-H M PA-Li.
vs. 1 / R T for the system p-isopro-
For the system XII-HMPA-Li only one anion radical could be observed between -10 and 80”. The coupling constant for the nitrogen is 10.82 G, Table I. This value is too large to be attributed to the “free” ion even for the electron-pushing methoxy group. Further, A N increases The Journal of Physical Chemistry, Vol. 77,
No. 19, 1973
with increasing temperature, and all of the “free” ions show AN’S that are independent of temperature. This conclusion is confirmed by the fact that sodium and lithium reductions exhibit pronounced g tensor anisotropy (this is characteristic of ion pairs) .8 For the system PhNOz_HlilPA-Na, only the “free” anion radical is observed a t -IO”, while above 80” only the ion pair is apparent, Figure 4. Unlike the Li reductions,
Enthalpies of Ion Pairing for Substituted Nitrobenzene
2341
TABLE I: Coupling Constants in Gauss, Enthalpies of Equilibrium 1, and &r Values for the Para Substituents for the Systems Para-Substituted PhN02-HMPA-Li at 25” “Free” Compd
I II Ill IV
V VI VI I
Substituent
H
i-Pro t-BU CI CN
COOCH3 CON;
Vlll IX
Et CHO
X XI XI I
COPh
NO2
OCH3
AN
8.48a 8.93 8.85 7.77 4.90 5.18 6.53 8.71 3.54
A0
ion An
3.34 1.01 3.44 1.00 3.43 1.23 3.43 1.08 2.89 0.51 2.76 0.40 3.09 0.80 3.36 1.02 0.12 0.12 1.92 1.70 1.16 1.16 1.16 3.92 2.30 No “free” ion was observed
ton pair Asubst
4.22 (H) 1.89 (H)
0.86 ( I N ) 0.40 (3H) 3.09 (2H) 2.61 (H) 1.16 (1N)
AN
A0
Am
Asubst
4.10 1.92
3.40 1.09 10.85 3.43 1.11 11.39 3.44 1.18 11.25 3.58 1.24 10.27 0.5 9.81 2.9 2.8 0.4 7.63 3.09 1.14 9.97 11.17 3.36 1.02 No ion Dair was observed
0.9 0.4 3.09
No ion pair was observed No ion pair was observed 10.82 3.46 1.13
db 0
-0.28 -0.25 0.1 1 0.66 0.48 0.4c -0.30
- A H ” ,kcal/mol 9.0 f 0.8 12.5 f 1.2 13.3 I0.8 10.8 f 0.4 0 1.1 I 0.2 7.0 f 0.6
0.79 0.26
-0.78
*
a The uncertainty in the coupling constants is 0.05 G for ail cases where three significant figures are given. u+ values are taken from C. D. Ritchie and W. F. Sager, Progr. Phys. Org. Chem., 2. 334 (1964). u+ for this group has not been reported. The value given is estimated from that far CONH.
metal splitting from the Na nuclei is observed a t all temperatures for the ion pair. Owing to the large amount of overlap between the “free” ion and ion pair spectra, the ratio of ( a ) to (/I) could only be determined accurately over a narrow temperature range. The plot of In ( a ) / ( / I ) us. 1IRT yields an enthalpy of -7.2 kcallmol for dissociation of the ion pair. Addition of hexane to solutions containing the two ions always increases the concentration of the ion pair a t the expense of the “free” ion. Discussion It has been established that the two ions simultaneously observed by esr are the “free” ion and ion ~ a i r . ~ %Here 4,9 we have observed that the enthalpy for the ion pair dissociation (eq 1) is a function of the para substituent, but all of the systems for which AH” could be determined yield a negative A F . The negative values of AH” indicate that there is more solvent ordering due to the “free” ion plus the cation than there is for the ion pair. This is in agreement with the work of Hirota.lo For the para-substituted PhN02-HMPA-Li systems the large negative enthalpies of ion pair dissociation are a result of the strong solvation of the unassociated cation by the HMPA. Tighter ion pairs are generated by stronger electron-pushing groups in the para position, which increases the charge density in the NO2 group, thus requiring less solvation of the cation by the HMPA. From this it is obvious that the ion pair dissociation involving tighter ion pairs will yield more negative enthalpies. The “free” ion plus the solvated lithium cation, however, have essentially the same solvation shield for all of the systems studied. This would require that the enthalpy for reaction 2 would grow more negative with the electron-releasing character of the para substituent. This is in accord with our observed negative p value. We should note here that AGO changes in the opposite direction with the para substituent. Of course, this is due to AS”, which is negative for reaction 1. For the cases of strong electron-withdrawing groups, i.e., NOz, CN, CHO, etc., very little if any ion pairing is observed. This is the expected result, since the charge density on the NO2 group is relatively small, as evidenced by the lower AN’Sfor these systems. The p-methoxy group
has just the opposite effect. The increased charge density on the NO2 group due to the electron-pushing nature of this substituent ( a + = -0.78) favors the ion pair, accounting for the fact that only the ion pair is observed in solution. Increasing the temperature always results is an increase in A N for the ion pair due to the fact that the dielectric constant of the solvent decreases with increasing temperature allowing the formation of tighter ion pairs. This same effect was observed when hexane was added to the HMPA solutions. The fact that ion pairing increases a t higher temperatures is confirmed by the observation that AN^ increases with increasing temperature for the PhNQ2HMPA-Na system. Decreasing the dielectric constant of the solvent by the addition of hexane also has the same effect upon the ratio of ( a ) to ( p ) as does increasing the temperature. The sign of p , the large negative values of AFT, and the variation of A N and ANa for the ion pairs all indicate stronger solvent interactions with the unassociated cation and “free” ion than with the ion pair. The large negative value for p (-23) is explained in part by the fact that HMPA is the most powerful cation solvator known,ll and small changes in the structure of the ion pair result in large changes in the heat of solution of the ion pair. Experimental Section The esr spectrometer system and the method of formation of the anion radicals were exactly the same as previously described.3 p-Ethyl- and p-isopropylnitrobenzene were prepared by nitration of the cprresponding alkylbenzene. The products were purified by vacuum distillation and preparative gasliquid chromatography. The N-methyl-N-tert-butyl-p-nitrobenzamide12 and p-nitromethylbenzoate13 were prepared as described in the literature. All of the remaining compounds were purchased from Aldrich Chemical Co. and recrystallized before use.
Acknowledgment. We are very grateful to Research Corporation for support of this work. We also wish to thank Dr. G. M. Rubottom for helpful discussion. The Journal of Physical Chemistry, Vol. 77, No. 19, 1973
2342
Fraser P. Price and Joachim W. Wendorff
References a n d Notes (1) R. D. Allendoerfer and R. J. Papez, J. Phys. Chem., 76, 1012 (1972). (2) G. R. Stevenson, L. Echegoyen, and L. R. Lizardi. J. Phys. Chem., 76, 2058 (1972), (3) G. R. Stevenson, L. Echegoyen, and L. R. Lizardi, J. Phys. Chem., 76, 1439 (1972). (4) G. R. Stevenson and L. Echegoyen, J . Phys. Chem., submitted for publication. (5) W. C. Danen, C. T. West, T. T. Kensler, and T. J. Tipton, J, Amer. Chem. SOC., 94, 4830 (1972).
(6) T. M: Krygowski, M. Stencel, and Z. Galus, J. Electroanal. Chem., 39, (1972). (7) A. Streitwieser, Jr., "Molecular Orbital Theory for Organic Chemists," Wiley, New York, N. Y., 1962, p 326. (8) G. R. Stevenson and L. Echegoyen, unpublished results. (9) G. R. Stevenson and H. Hidalgo, J . Phys. Chem., 77, 1027 (1973). (10) N. Hirota, J. Phys. Chem., 71, 127 (1967). (11) H. NormanLAngew. Chem., Int. Ed. Engl., 6(12), 1046 (1967). (12) G. M. Rubottom, Tetrahedron Lett.. 44, 3887 (1969). (13) R. E. Ireland, D. A. Evans, D. Glover, G. M. Rubottom, and H. Young, J. Org. Chem., 34, 3717 (1969).
Transitions in Mesophase Forming Systems. V. Kinetics of Transformation and roperties of Cholesteryl Stearate' Fraser P. Price* and Joachim H. Wendorff Polymer Science and Engineering. University of Massachusetts, Amherst. Massachusetts 01002 (Received March 15, 1973) Publication costs assisted by the National lnstitutes of Health
The equilibrium density-temperature behavior and the interphase transformation kinetics have been studied for carefully purified cholesteryl stearate employing tthe techniques of precision dilatometry. In the crystalline solid the density at a given temperature depends upon the thermal path followed in attaining that temperature. The effect is not large, amounting to a t most 5% in the density. However, it is reproducible and does not seem to be due to voids in the sample. There is an approximately 5% volume change at the solid-isotropic transition temperature of 81.5". The isotropic-cholesteric volume change is 0.17% and occurs a t 77.5". The transformation kinetics into the solid state indicate the homogeneous nucleation of spheres at temperatures below 74.7". The isotropic-cholesteric transformation in the range 75.4-77.0' is characterized by the growth of homogeneously nucleated rods which develop from disk-like nucleii. This seems to be the general case for transformations in which the developing phase is a liquid crystal.
Introduction In this laboratory we are engaged in studies of the transformation kinetics in mesophase forming systems and in the precise determination of the temperature behavior of the densities of these systems. Previously, we have studied cholesteryl acetate,Z cholesteryl nonanoate,3 cholesteryl myristate,4 as well as p-azoxyanisole (PAA).5 All three esters of cholesteryl exhibit a cholesteric phase. In the myristate and the nonanoate this mesophase is enantiotropic while in the acetate it is monotropic. The smectic state does not occur in the acetate, is enantiotropic in the myristate, and is monotropic in the nonanoate. In PAA the nematic mesophase is enantiotropic.6 All these mesophase formers exhibit density-temperature behavior indicative of marked pretransition effects only on the low-temperature side of the transition. In the cholesteryl esters the rate of transformation from a given state to one stable at a lower temperature usually is sufficiently slow for the kinetics of the transformation to be observed. The transformations into the solid states are particularly slow and apparently are governed by the kinetics of the nucleation process.7 In PAA, however, the isotropicnematic transformation is too rapid to follow with our techniques and the transformation into the solid state is also nucleation controlled but it occurs at such high suThe Journal of Physical Chemistry, Vol. 77. No. 19, 1973
percoolings that, when it takes place, it occurs essentially instantaneously. In cholesteryl stearate, as in cholesteryl acetate, all the mesophases are monotropic. The solidisotropic transformation temperature of the stearate is variously given as 81.8-85. lo, the isotropic-cholesteric transition as 71.4-71.0", and the cholesteric-smectic transition is 69.9O.8 We are interested in delineating quantitatively the effects of chemical structure on mesophase behavior of the esters of cholesterol and this paper deals with the properties of the various transitions of cholesteryl stearate. We are concerned with the equilibrium values of the densities as well as with the kinetics of transformation. Experimental Section Muteriul. The mercury used to fill the dilatometers was obtained from the Sargent Welch Co. (Reagent Grade ACS). Samples of cholesteryl stearate were obtained from the Eastman Kodak Co., Rochester, N. Y. The cholesteryl stearate was purified by recrystallization from 1-pentanol, washed several times with a water-ethanol mixture, and dried under vacuum at various temperatures, up to temperatures above the melting point. Sometimes this recrystallization and washing procedure, without the drying, was repeated several times.
