152
K. K. SRIVASTAVA
Dielectric Relaxation Study of Some Pure Liquids by K. K. Srivastava Department of Physics, Panjab Undversi4y, Chandigarh, India
(Received April 50,1060)
The dielectric constants and losses of liquid diphenyl ether, dibutyl ether, N,N’-dimethylaniline,and N,”diethylaniline have been measured at two frequencies (9.72 and 31.82 GHz) in the microwave range and at three different temperatures. The calculated values of thermodynamicand distribution parameters indicate the possibilities of (i) more than one relaxation mechanism and (ii) bond formation in the cases of liquid dimethyland diethylanilines. The relaxation times for these anilines showed frequency dependence as well.
Introduction The present work essentially deals with the dielectric relaxation study of diphenyl ether, dibutyl ether, N,N’-dimethylaniline, and N,N’-diethylaniline as pure liquids. Extensive data are already available in the literature for the diphenyl ether molecule1-’ and, as such, this work would hardly add anything to the existing situation. The purpose for the study of this molecule is therefore mainly to test the reliability of experimental technique adopted by the author and see if the results obtained tally with those already reported. The other molecules have, in general, been studied in a number of nonpolar solvents but not many data seem to be available for their study in liquid state. It was therefore thought that this work may provide some more information for determining their molecular behavior. Experimental Section Chemicals. All the liquids were imported from Dr. Theodor Schuchardt, Miinchen, West Germany, with the specifications : diphenyl ether, pure grade; dibutyl AR ether, pure grade; N,N’-dimethylaniline, Grade, free from monomethylaniline; and N,N’-diethylaniline, pure grade. These chemicals were used without any further purification except N,N’-diethylaniline, which was distilled under vacuum and dried before use. Their physical constants, if compared with the literature values shown in Table I, indicate the extent of their purity. Static Dielectric Constant, Density, Refractive Index, and Viscosity Measurements. The static dielectric constant values were determined by the help of a dielectrometer,8 designed and fabricated by the author, working on the principle of heterodyne beat method. The density was measured with a pycnometer (capacity -12 ml) which could be dipped inside a thermostatic bath having an accuracy of *0.03”. The values of refractive indices (for sodium D-line) were measured at different temperatures by a HilgerAbbey refractometer having an arrangement for the circulation of temperature-controlled water through a jacket enclosing the experimental liquid. The Journal of Physical Chemistry
The viscosity values were measured by the Ubbelohde viscometer, duly calibrated with conductivity water. The entire viscometer was kept inside a thermostatic bath with a glass window through which the time of fall was noted. The instrument had an accuracy of 0.01%. The viscosity of the unknown was found by the formula 77 =
dt 770-
doto
where 1is the viscosity of the unknown and qo is that of water at the given temperature, while d and do are their densities; t and to are the efflux times for the unknown and water, respectively, at the same temperature. The experimentally observed values of the parameters mentioned above are shown in Table I along with their literature values. Dielectric Constant and Loss Measwements. These measurements were made at two frequencies, 9.72 and 31.82 GHz. Figure 1 gives the block diagram of the arrangement used. The entire assembly for the X-band (9.72 GHz), manufactured by Microwave Instruments Ltd., consisted of a reflex klystron as energy source, a wavemeter for the measurement of frequency, wave guides for transmitting the energy and a traveling probe (S.W.D.) for detecting the energy. The output from the standing wave detector was fed to the galvanometer (sensitivity: 22 X A) provided with a lamp and scale arrangement. A (1) W. E. Vaughan and C. P. Smyth, J . Phys. Chem., 65, 98 (1961). (2) K. Higasi, “Dipole, Molecule and Chemistry,” a monograph series of the Research Institute of Applied Electricity, No. 13, 1965, p 21.
(3) K. Higasi, “Dielectric Relaxation and Molecular Structure,” a monograph series of the Research Institute of Applied Electricity, No. 9 1961, p 21. (4) D. M. Roberti, 0. F. Kalman, and C. P. Smyth, J . Amer. Chem. Soc., 82, 3523 (1960). (5) Reference 3, p 9. (6) F. Hufnagel, G . Klages, and P. Knobloch, 2.Naturforsch., 17a, 96 (1962). (7) J. E. Anderson and C. P. Smyth, J. Chem. Phus., 42,473 (1965). (8) P. K. Sharma and K. K. Srivastava, Oyo Denki Kenkyusho Iho, 16,53 (1964).
DIELECTRIC RELAXATION STUDY OF SOME PURELIQUIDS
153
Table I T,OC
-Static Obsd
dielectric constantLit.
____-Refractive index----
g/mlLit.
--Density, Obsd
Obsd
---Viscositv, Obsd
Lit.
- . cP-Lit.
Diphenyl Ether 30 40 50
3.660 3.593 3.517
3 6lQ a t 40"
1.0669 1,0584 1.0499
1 .0661d at30'
1.5760 1.5711 1.5670
1.5762d a t 30"
2.92 2.39 1.95
1.78c a t 60"
1.3980 1.3900 1.3867
1.39906d at 15" (for He yellow line)
0.58 0.50 0.45
0 . 602d a t 30"
1.5523d a t 32.5'
1.13 0.96 0.83
1.16d a t 30"
1.61 1.30 1.09
0 , 783d
Dibutyl Ether 30 40 50
3.116 3.029 2.952
Not available
0. 75976d a t 30°C
0.7612 0.7528 0 7439 I
N,N'-Dimethylaniline 30 40 50
4.998 4.864 4.722
4.91b a t 20"
0.9495 0.9415 0.9329
0. 94804d
a t 30"
1.5530 1.5481 1.5431
N,N '-Diethylanilhe 30 40 50
5.037 4.877 4.729
Not available
0.9263 0.9181 0.9102
0 . 9348d a t 20"
1.5362 1.5311 1.5267
1 .54105d a t 22.3"
a t 75.5"
a J. H. Calderwood and C. P. Smyth, J . Amer. Chem. Soc., 78, 1295 (1956). A. A. Maryott and E. R. Smith, "Table of Dielectric Constants of Pure Liquids," National Bureau of Standards Circular 514, U. S. Government Printing Office, 1951. e W. E. Vaughan, W. S. Lovell, and C. P. Smyth, J. Chem. Phys., 36,535 (1962). J. Timmermans, "Physicochemical Constants of Pure Organic Compounds," Elsevier Press, New York, N. Y., 1950.
MICROMETER DRIVE MECHANISM
t
U
-
KLYSTRON
POWER SUPPLY
i - T KLYSTRON 0 S C ILL ATOn
Figure 1. Block diagram for microwave measurements.
With the liquid in the cell and the plunger at the cell window, the traveling probe is set to a minimum power position. The shorting plunger is then turned upward, increasing the depth of the liquid, and the output of the standing wave detector (probe position fixed) is noted. The distance between successive minima is equal to Xd/2. For low loss materials, the dielectric constant is given by e ' = ($)2
+
mica window was used to separate the top part of the wave guide, forming the dielectric cell, into which was put the experimental liquid. This cell was enclosed by an outer jacket through which temperaturecontrolled water was circulated. The measurements in the J band (31.82 GHz) were carried oute by using a klystron (EMl/R9546) and a frequency meter, Type PRD 538. The micrometer-driven plunger arrangements for both the frequency ranges were fabricated a t the workshop of Allahabad University. The accuracy of measurement was i 2 % for E' and =k5% for E".
where A, is the cut-off wavelength, XO is the free space wavelength, and Ad is the wavelength in the dielectric filled guide of the cell. The dielectric loss is determined by bringing the plunger down to the mica window so that the liquid length is zero. The field pattern in the standing wave detector is then recorded by moving the probe, and the width at twice minimum power points Azo is determined. The inverse voltage standing wave ratio po is calculated by relation
Theory of Measurements and Results The values of dielectric constants and losses were determined by the method developed by Heston, Franklin, Hennelly, and Smyth'O for low loss liquids. The experimental procedure of this method, for a shortcircuit case, is as follows.
where X, is the wavelength in the airfilled wave guide. The value of por obtained from eq 2, is corrected for (9) These measurements were carried out in the laboratory of Prof. Krishna Ji of Allahabad University, India. (10) W. M. Heston, A. D. Franklin, E. J. Hennelly, and C. P. Smyth,
J . Amer. Chem. Soc., 72,3443 (1950). Volume 74, Number 1
January 8, 1070
K. K. SRIVASTAVA
154 the approximation involved in deriving the re1ation.l' Next, the plunger is kept at different positions in steps of hd/2 and the standing wave pattern is recorded by moving the probe of the standing wave detector. The width at twice minimum power points is determined in each case. The values of pn for each plunger position are found by applying the relation (2) alongwith necessary corrections, as stated above. The slope dp,/dn is then determined by plotting a p n vs. n curve, showing a linear relationship. The dielectric loss factor e" is finally determined by the relation
0.5
i'
c
c
m
2.4 \2-6
2.8
3.0
3.4
3.2
3.6
(3) The values of E' and e", for the two frequencies and different temperatures, have been tabulated in Table 11.
0.5
Table I1 7 - 9 . 7 2 GHz-7
T,OC
Figure 2. Cole-Cole arc plot for diphenyl ether a t 40'. Open circles, this work; closed circles, ref 1.
f'
,---31,82 E'
E''
t
0-30°C.
x----so"c.
GHze"
Diphenyl Ether 30 40 50
3.35 3.32 3.32
0.38 0.33 0.30
2.8 I
3.07
0.44
2.49 2.49 2.49
0.36 0.37 0.34
3.2 I
3. I0
____+
3'4
3.6
E'
Dibutyl Ether 30 40 50
3.03 2.78 2.76
0.22 0.26 0.25
Figure 3. Cole-Cole arc plot for diphenyl ether a t 30 and 50".
N,N '-Dimethylaninline 30 40 50
3.39 3.67 4.08
0.42 0.47 0.51
2.93 2.93 3.03
0.27 0.31 0.60
N,N '-Diethyhiline 30 40 50
2.78 2.74 2.83
0.31 0.23 0.16
2.66 2.69 2.60
0.24 0.27 0.12
Distribution Parameter and Relaxation Time. The values of distribution parameter (a)and macroscopic relaxation time (T) are obtained by the help of a ColeFigure 4. Cole-Cole arc plot for dibutyl Cole')' plot of E" against e' for different frequencies ether a t 30, 40, and 50'. and temperatures. Figures 2-6 illustrate that these plots for all the experimental liquids represent an arc Knowing a, the macroscopic relaxation time r is of a circle intersecting the abscissa axis at the values of e, and eo, and having their centers lying below the found by the relation abscissa axis. eo is the static dielectric constant and 21 - =: the value of em is obtained by squaring the refractive (4) 2c index value. The diameter drawn through the center from the (11) A. R. Von Hippel, "Dielectric Materials and Applications," e, makes an angle an/2 with the e' axis. Tan a ~ / 2 The Technology Press of M.I.T., 1954, Chapter 11, p 86. is determined from the plots and a is calculated. (12) K. 8.Cole and R. H. Cole, J . Chem. Phya., 9,341 (1941). The Journal of Physical Chemistry
155
DIELECTRIC RELAXATION STUDY OF SOMEPURE LIQUIDS 0
x
I
-30°C
Table I11
-----4dC
13 -.-.-58
c
Macroscopic relaxation time, T X 1012 sea
Distribution parameter a
T,OC
Microscopic relaxation time, T O X lo**sec
Diphenyl Ether 30 40 50
0.34 0.35 0.33
5.0 4.1 2.7
4.4 3.6 2.4
Dibutyl Ether 30 40 50
0.30 0.25 0.25
5.7 4.7 4.2
5.0 4.2 3.6
N,N'-Dimethylaniline 30 40 50
0.59 0.55 0.34
46.1 14.0 4.7
38.2 11.6 3.9
N,N '-Diethylanilhe Figure 5. Cole-Cole arc plot for dimethylaniline a t 30, 40, and 50".
30
0.52 0.54 0.71
40 50
354.0 475.8 1231.7
290.3 390.1 1022.3
Table IV Free energy of activation kcal/mol T,'C
AFs
Entropy of activation cal/mol
A&
AFq
A&
Enthalpy of activation kcal/mol AH Axe
Diphenyl Ether 30 40 50
2.07 2.04 1.87
5.64 5.70 5.76
10.75 10.52 10.73
-5.76 -5.47 -5.48
5.33
3.99
2.41
2.48
21.53
2.86
...
3.81
Dibutyl Ether 30
40 50
2.16 2.13 2.14
4.71 4.78 4.87
-7.37 -7.34 -7.41
N,N'-Dimethylaniline
\. \.
30 40 50
\. \.
3.41 2.81 2.22
4.93 5.00 5.07
30 40 50
Figure 6. Cole-Cole arc plot for diethyladine a t 30, 40, and 50".
where v is the distance on the Cole-Cole plot between eo and the experimental point, u is the distance between that point and e,, and w is the angular frequency. Further, the Powles internal field correction is applied and the molecular relaxation time 70 is calculated from the macroscopic relaxation time 7 by the relationla =
2eo
59.78 59.81 59.76
-6.82 -6.83 -6.85
N,N '-Diethylaniline
b
70
0.81 0.85 0.82
+ 3eo
em 7
(5)
...
...
...
5.23 5.33 5.39
... ... . ..
-4.69 -4.64 -4.91
These values of a,7 , and T~ for the experimental liquids at different temperatures have been shown in Table 111. Thermodgnamic Parameters. These parameters have been calculated by the help of Eyring's equationsl4-l6 (13) J. G.Powles, J. Chem. Phys., 21,633 (1953). (14) H.Eyring, ibid., 4, 283 (1936). (15) A. E. Stearn and H. Eyring, ibid., 5,113 (1937). (16) S.Glasstone, K.J. Laidler, and H. Eyring, "The Theory of Rate Processes," MoGraw-Hill, New York, N. Y., 1941,pp 544-551. Volume 74,Number 1 January 8, 1970
K. E(. SRIVASTAVA
156
'
7 s -
eAFe/RT
kT and
where the notations all have their usual significance. AF, and AF, are the free energies of activation for dipole relaxation and viscous flow. These values are related to the enthalpy of activation and entropy of activation by the relations
AH,
- TAS,
AF, = AH,
- TAS,
AF,
=
(8)
and
(I/Tl
(9)
Figure 8. Log 7 and log
(T
X
XIO'
T )us. (1/T)for dibutyl ether.
Equations 6 and 7 indicate that the plots of log (T 2') vs. 1 / T and log 17 vs. 1/T should give approximately a linear relationship (see Figures 7-11) with slopes equal to A H , / R and A H J R , respectively, from which AH, and AH, are calculated. Also, from eq 6 and 7 , one gets
(3
AF,
=
2.303 R T log -
AF,
=
2.303 RT log
and
(Zd
where V is taken to be the molar volume and is found from the density measurements. Knowing T , 17, and T , one can evaluate AF, and AF, and subsequently AS, and AS, from eq 8 and 9. Table IV lists the values of all these parameters.
Discussion Diphenyl Ether. Figure 2 shows the Cole-Cole plot for this molecule a t 40". For comparison's sake, the
-5.30
2-40-
-'.'x,
F V
-ii*ro 9
-k o I
The Journal of Physical Chemistry
.b-90
( I/T ) x
Figure 9. Log 7 and log for dimethylaniline.
(T
X
lo3
T )us. (1/T)
literature values of E' and E''for diphenyl ether at 40", observed by Vaughan and Smyth' have also been plotted on the same curve. Open circles correspond to the values determined by the author whereas the closed circles refer to the values observed by Vaughan and Smyth. All the points fall fairly well on the Cole-Cole plot which indicates, if nothing else, a t least the reliability of the present experimental data. Similar plots for the same molecule a t temperatures 30 and 50" have been shown in Figure 3. Although it has been shown that relaxation of diphenyl ether fit a involves two distinct r n e c h a n i s m ~ , ~the ~ ~ ~data '
DIELECTRIC RELAXATION STUDYOF SOMEPURELIQUIDS
€or dimethylaniline, where Figure 10. Log (T X 2') vs. (1/T) the relaxation time ( 7 ) corresponds to 31.82 GHz.
(I/T)
Figure 11. Log 7 and log
(T
x103
X 2') vs. ( 1 / T ) for diethylaniline.
These values are 0.34/30", 0.35/40", and 0.33/50° as against 0.21/40°, reported by Vaughan and Smyth.' Slightly higher values obtained by the author may be due to a different method adopted for the determination of a. Maybe a plot of In jv/u 121s. In w would bring the a values down but it could not be checked as the measurements for e' and e" were carried out only a t two frequencies. The relaxation time for liquid diphenyl ether, as reported by Smyth, et aL12is 5.1 psec whereas the author found a value of 5.0 psec at 30".
157 The literature value3 for free energy of activation (AF,) is 2.20 kcal/mol a t 40" whereas the present work at 40" gives a value of 2.04 kcal/mol. Roberti, Kalman, and Smyth4 found AH, and AH, to be 1.9 kcal/mol and 3.8 kcal/mol for diphenyl ether in Nujol. These values for pure liquid have been found to be 5.33 and 3.99 kcal/mol, respectively. Even by orthodox methods16 the value for AHe comes out to be 5.21 kcal/mol a t 40". Dibutyl Ether. No data on relaxation study seem to be available in the literature for this molecule. The e' and E" values for the three temperatures represent the Cole-Cole plot fairly well (See Figure 4). The point corresponding to 9.72 GHz/30" falls far beyond the curve, which may be an experimental error. The values of distribution parameter a also indicate the usual trend with temperature. The macroscopic relaxation time r calculated for both the frequencies (9.72 GHz and 31.82 GHz) showed a difference of only 5%. Their average values, therefore, have been inserted in Table 111. These values are, 5.7 psec/30", 4.7 psec/40", and 4.2 psec/50". The relaxation times for the same molecule in different solvents, e.g., benzene, cyclohexane, carbon tetrachloride, heptane, decalin, and Nujol, have also been determined in the same laboratoryO17 This value in benzene has been found to be 6.5 psec a t 25" and shows an almost regular variation with the viscosity of the medium giving r = 14.2 psec for Nujol, having the largest viscosity coefficient. The viscosity of dibutyl ether at 30" is 0.58 cP, which is almost the same as that of benzene (0.57 cP). Obviously, the small difference in the two relaxation times hardly indicates any observable difference involved between the two systems. Also, considering the molecular volumes of dibutyl ether, diphenyl ether, and benzophenone, one finds that they have almost similar values of 170.9, 159.5, and 170 cc, respectively, but the relaxation times of dibutyl ether and benzophenone in benzene are 6.5 psec and 18.7 psec,'* respectively. This is possibly due to the fact that the butyl group sweeps a lesser area of solvent molecules while relaxing as compared to the phenyl group. N,N'-Dimethylaniline. In this case also, a perusal of Figure 5 reveals that the experimentally determined values of e' and e" at different temperatures represent the Cole-Cole arc fairly well. The values of distribution parameter a are found to be quite large, thus indicating the possibility of more than one relaxation mechanism. These values are 0.59/30°, 0.55/40", and 0.34/50", which show the usual trend with temperature. Whereas the molecular relaxation times for these temperatures are 38.2, 11.6, and 3.9 (17) This work has been done in this laboratory in collaboration with J. K. Vij and is shortly to be communicated for publication. (18) Reference 2, p 21 Volume 74, Number 1 January 8, 1970
K. K. SRIVASTAVA
158 psec, Garg and Smyth,l9 however, found a value of 28 psec for pure liquid. Other details of their work are not available since their findings have not been published. Grubb and Smyth, in their paper,20 reported the measurements performed by Dr. K. Fish which do not agree with those reported here. It appears that the main source of discrepancy is the method adopted for calculations. The author feels that for low-loss substances, the method of Heston, et al.,1° (adopted in this paper) gives more reliable results than that of Surber21(adopted by the authors20) which gives accurate results for medium- and high-loss substances. Also, as low a value for a (0.02) as reported in the paper20 should not account for the molecule having a relaxation mechanism suggested by Smyth.'S ChitokuZ2also reported the values of the average relaxation time for this molecule in benzene and dioxane which are 14.5 and 30.4 psec at 20" but the distribution parameters, obtained by him for these solvents, are zero for both the cases. On comparing the viscositites of pure N,N'-dimethylaniline and dioxane, one finds that these are almost of the same order, having values equal to 1.13 CP at 30" and 1.2SZ3CPat 20", respectively. Higasi, et a1.,22already suggested the formation of bonds between aminohydrogen atoms in aniline and an oxygen atom of dioxane molecule which leads to a high relaxation time in dioxane. Another factor noticed for this molecule is about its large variation in the relaxation time found at experimental frequencies. The a0 values, mentioned above and noted in Table 111, correspond to the frequency 9.72 GHz, whereas the values calculated for 31.82 GHz, using eq 4 and 5, are 84.6 psec/30°, 59.8 psec/4O0, and 11.5 psec/SO". On comparing Figures 9 and 10, one finds that a log (7 T) vs. 1/T plot is a perfect straight line if 7 values correspond to 9.72 GHz but it is not true if the values of 7 ,obtained for 31.82 GHz, are used (Figure 10). No adequate explanation can be given at this stage with the limited data available but it appears that in this case, 7rn seems to be dependent on the frequency of observation, whereas this quantity is assumed to be constant for the validity of Cole-Cole's rule. 7rn is the most probable relaxation time corresponding to the frequency that gives the greatest value of E". N,N'-Diethylaniliiae. The study of liquid N,N'diethylaniline indicates its unusual behavior in many ways. For example, the distribution parameter and the relaxation time increase with increasing temperature (see Table I11 and Figure 11). The molecular relaxation times for the liquid have been found to be 290.3 psec/30", 390.1 psec/40°, and 1022.3 psec/SO". The literaturez2values for the same substance in benzene and dioxane are 33.2 psec and 47.4 psec at 20". Like that of N,N'-dimethylaniline, the relaxation time for this substance also shows frequency dependence. The ,Journal of Physical Chemistry
However, the concentration of almost all the experimentally observed points into a small region of a ColeCole plot leads one to suspect the experimental inaccuracy involved to distinguish one point from the other in this range of frequency (see Figure 6). It is therefore felt that measurements a t still lower frequencies are very essential in this case. Thermodynamic Parameters. These parameters have been calculated for all the substances under study at different temperatures and the results are shown in Table IV. The values of AFc, AH,, and AS, have not been calculated for N,N'-diethylaniline because of the uncertainty involved, but it is expected that AS, for this molecule would be almost of the same magnitude as that of N,N'-dimethylaniline. Table IV, however, shows that AF,, the free energy of activation for the relaxation process, is always less than AF,,, the free energy of activation for viscous flow, as is always expected. The values of AS,, for all the molecules are negative whereas AS, values have been found to be positive. It is therefore apparent that there must be a considerable difference between the molecular processes involved and that of any concept of "internal viscosity." Furthermore, when the values of this entropy are as large as they often are here (especially in case of aniline), we know that activation must involve more than merely a single molecule. In order to explain the large positive values of AS,, especially in cases of dimethyl and diethyl anilines, the author would like to state the following. Suppose we attempt to go from one stable arrangement of a small region in a dielectric to another in the most rapid manner possible, utilizing only the thermal motions of molecules. Therefore, the surrounding of a given molecule must move simultaneously with the molecule. This might occur in two ways, i e . , (a) the various molecules involved may cooperate in their movements and rotate together in much the
Table V : Dipole Moment (in Debye units) at 30' PHigasi NOnsctger
Substance
(pure liquid)
Diphenyl ether Dibutyl ether N,N '-Dimethylaniline N ,N '-Diethylanilhe
1.11 1.29 1.48 1.67
(soln. in benzene)
1.19 1.18 1.62 1.79
(19) 5. K. Garg and C. P. Smyth, unpublished work; this value has been quoted in the Chemical Society Publication, "Molecular Relaxntion Processes," No. 20, Academic Press, London, 1966, p 8. (20) E. L. Grubb and C. P. Smyth, J . Amer. Chem. SOC.,83, 4879 (1961). (21) W. H. Surber, Jr., J . Appl. Phys., 19,514 (1948). (22) K. Chitoku and K. Higasi, Bull. Chem. Soc., Jap. 39, 2160 (1966). (23) K. Chitoku and K. Higasi, ibid., 39,2166 (1966).
159
CALCULATION OF 3.7RTm FOR IONIC LIQUID same manner as a set of interlocked gears or (b) the molecules may momentarily completely disengage one another and resume their stable configuration with their net dipole moment oriented in a new direction. Process (a) will presumably require little energy, while (b) will require considerable energy. The possibility of process (a) seems to be less probable if one looks at Table V which gives the values of dipole monents for pure liquids, calculated by Onsager’s equation p2
+
9kT M ( 2 ~ 0
= --
4TN d
€,)(EO
360(6m
+ 2)
+ 2)
and the dipole moments of these substances in benzene as solvent, calculated by Higasi’s equation24 4,/
= B(& -
aD)’’%
Since these values do not show any appreciable difference, the possibility of mechanism (a) is less probable as other Wise P(0nsager) would be different than P(Higasi) because the latter case deals with the sohtions in a
very low concentration range where the interlocking effect is absent, while with mechanism (b), on the other hand, there will be a great increase of freedom in the activated complex, so giving a positive entropy of activation whose magnitude will be a measure of the extent of the momentary violation of the requirements for stability. Incidently, the hydrogen bonding of aniline has already been r e p ~ r t e d . ~Maybe ~,~~ therefore, this is the reason for very large observed relaxation times in the cases of dimethyl- and diethylanilines.
Acknowledgment. The writer wishes to express his gratitude to the World University Service for providing UNESCO gift coupons for the purchase of chemicals. He is also grateful to Professor Krishna J i for providing facilities to work on the J-band assembly. (24) K. Higasi, Oyo Denki Kenkyusho Iho, 4,231 (1952). (25) J. C.Dearden and W. F. Forbes, Can. J . Chem., 38,896 (1960). (26) B. N,R ~ Oet, al., ibid., 40,963 (1962).
Semiempirical Calculation of 3.7RTmTerm in the Heat of Activation for Viscous Flow of Ionic Liquid by T. Emi and J. O’M. Bockris Electrochemistry Laboratory, U,niversity of Pennsylvania, Philadelphia, Pennsylvania
19104 (Received January 24, 1969)
The heat of activation in transport in ionic liquids is rededuced in terms of the work of hole formation in the liquid. Theoretical values calculated for the simple molten salts for which data are available agree reasonably with those determined experimentally.
1. Previous Position It has been found experimentally‘ that the Arrhenius type activation energies for viscous flow, E,, self-diffusion of cations, ED,,and anions, ED-,all reveal practically the same value in a molten salt, namely as E,, E E D ,E ED-E 3.7RTm
(1)
where R is the gas constant and T , is the melting point of the molten salt. Equation 1 is also applicable to simple organic liquids and to some liquid metals (Figure 1). Several attempts have been made in our laboratory to interpret the above relationship by extending the arguments of the theory of holes proposed by Furth,2
who considered the liquid state to be a mixture of holes and the matter outside them. The matter is a continuum with the normal surface tension, Q, of the liquid. The distribution of hole sizes is given by the formula
1.
W(r) = .lCe-El”dp,dp,dp,dp, (2) where C, the normalizing factor, is obtained by integrating W ( r ) from zero to infinity with respect to r; p,, p,, p , are the momenta corresponding to the coordinates x, y, and x , and p r is the momentum corre(1) L. Nanis and J. O’M. Bockris, J . Phys. Chem., 67, 2865 (1963); J. O’M. Bockris and 8. R. Richards, ibid., 69,671 (1965). (2) R. F W h , Proc. Cambridge Phil. Soc., 37, 252 (1941); F. H. Stillinger in “Molten Salt Chemistry,” Ed., M. Blander, Interscience Publishers, New York, N. Y.,1964,p 1. Volume 74, Number 1 January 8 , 1970