NOTES
459
molecule. Thus the total yield of NH2 radicals determined by product analysis corresponds to (P(NH2) = 0.64. This value again is only a lower limit since reaction 4 is not included and S z formation very probably involves more than two NHp radicals per KZ molecule formed. In fact a maximum of 6 NI-I2 radicals may be used up per N2 molecule formed (2
+ 4x)NH, +Sp + (2 - 2x)Hz + 42”s (x being 0 < < 1) ~t:
(5)
Equation 5 holds only if no radicals other than NHp are involved in I T 2 formation. Contributions from reactions of ethyl radicals will be of minor importance since the C2Hs yield is well accounted for by disproportionation of CzHB radicals.” Only a tentative evaluation of the primary quantum yield of ammonia decomposition can be given since the ratio of the rate constants K 3 / K 4and the mechanism of nitrogen formation are not known. On the basis of the discussion given above this study leads to a value for ethyl rndicals corresponding to an H-atom yield of 0.93 and to an NH2yield of 0.72-0.96. Both results are lower than the expected quantum yield of unity for (1). They are, however, in good agreement with (P(HD) = 0.94 observed in the photolysis of ND3 at 2062 in the presence of 620 torr of p r ~ p a n e . ~ In a series of runs the light intensity was reduced to l / 4 9 of the normal value. The results are given in Table 11. The yields of alkanes remain constant whereas the quantum yield of ethylamine is reduced. No hexane formation was observed although it would have been 3 X detected if formed with a quantum yield Thus the attack of C2H5on CzH4can be neglected in our system. However, an attack of NHZ on CzH4 is postulated. This is supported by the reduced yield of ethylamine and hydrazine and the higher yield of butylamine and ethylenediamine among the products.
>
+ CzH4 +n”zCzH4 NHzC2H4 + CzH5 +IYH2C4Hg SHzCzH4 + NHz +NH2CzHdNHz SH2
(6)
(7) (8)
The relative rate of ethylamine to butylamine for-
Table I1 : Product Formation at Reduced Intensity (8.83 X 10’8 quanta/sec) Products
Butane Ethane Ethylamine Butylamine Hydrazine Not identified
Quantum yields
CzHa consumed
0.20 0.045 0.28
0.40 0.09 0.28 0.06
0.06
0.025
...
mation is 23 for the high-intensity runs compared with 4.5 in the photolysis at low intensity. Furthermore, a qualitative agreement exists between the change in the rate of hydrazine formation and the production of ethylamine and butylamine, respectively. Steadystate calculations indicate that NZH4 should decrease by a factor of approximately 0.3 in irradiations at low intensity, which is in agreement with our findings.
Acknowledgment. The authors wish to thank Professor Groth for his stimulating interest and helpful discussions.
w.
(11) J. A. Kerr and F. A. Trotman-Diokenson, Progr. Reaction Kinetics, 1 , 105 (1961).
The Dielectric Properties of Pyridine Complexes with Dichloro- and Trichloroacetic Acids by S. R. Gough and A. H. Price] Edward Davies Chemical Laboratory, University College of Wales, Aberystwyth, Great Britain (Recehed July $9, 1968)
Dielectric investigations of 1: 1 acid-base complexes in the microwave region of the spectrum yield information on the stability of the complex and its dipole moment (hence ionic character) and may provide information on the proton-jumping process in the equilibrium between the hydrogen-bonded and the ionic forms
Complexes involving acetic acid and monochloroacetic acid with pyridine exist predominantly in the hydrogenbonded form, while the trichloroacetic acid-pyridine complex is ionic.2 Dielectric studies up to 1.8 GHz show that these complexes behave as rigid structures and gave no evidence of any proton-jumping process. The dichloroacetic acid complex presents a possible intermediate case where the proton jumping may occur in the microwave region. The results reported here are an investigation of the dielectric properties (in the solid state and in benzene solution) of the dichloroacetic acid-pyridine complex at frequencies up to 8.5 GHz and an extension of the data2 on the trichloroacetic acid complex up to 36 GHz. Experimental Section The dichloroacetic acid was purified by fractional distillation under reduced pressure. The fraction boiling
0.01 __ 0.84
(1) All correspondence should be sent to this author. (2) M. Davies and L. Sobozyk,J . Chem. Soo., 3000 (1962). Volume 78, Number 2 February 1960
NOTES
460
a t 88-89' (10-11 mm of pressure) was collected. Analar grade trichloroacetic acetic acid was crystallized from chloroform and was stored over phosphorus pentoxide. Both the benzene and pyridine were dried and fractionally distilled. The specific conductance of the benzene used ranged from 10-l2 to 10-13 ohm-1 cm-l. Benzene was used as a solvent and measurements were made at 20'. Solutions were prepared in a drybox and care was taken to exclude moisture from the measuring cells. I n all cases l : l molar ratio of the acid to base was maintained. Solid complexes were prepared by precipitation from concentrated benzene solution containing a 1 : 1 molar ratio of the acid and base. The dichloroacetic acidpyridine complex was hygroscopic. Suitably shaped solid samples were prepared for insertion into the dielectric cells by pressing the powdered solid in specially manufactured dies. Permittivity (e') and dielectric loss factor (e") were measured in the frequency range 5 Hz-320 kHz using a Thompson bridge3with calibrated dielectric cells. Over the frequency range 250 MHz-8.5 GHz, standing wave techniques were used €or measurement4 with the sample contained within the wave guide. At 36 GHz dielectric loss factors were determined by direct measurement of the attenuation produced by the introduction of a known length of dielectric into the appropriate waveguide.6 Table I : Permittivity ( E ' ) of the Solid Complex Complex
Frequency range
Dichloracetic acid
250 MHz-8. 5 GHz
Trichloracetic acid
1 GHz-8.5 GHz
Temp, O C
20
-180 20
-180
aocf
CIA-B
* 0
2.0 0.0
ClpA-B
/ /
CISA-B
This work Lit. dolo (ref 2)
I
0.5
1.0 -Log
1.5
2.0
c,
Figure 1. Log &-log C plots for substituted acetic acid-pyridine complexes in benzene solution (20').
tional to the frequency of measurement and is entirely due to ionic conduction. The equivalent conductance 10K '
A=-----
c
-
E'lf
1.80
x
10BC
(where K is the specific conductance, C is the molar concentration, and f is the measuring frequency, in hertz) of these solutions is shown in Figure 1, together with the results of Davies and Sobczyk for the monoand trichloroacetic acid complexes. The results for the dichloroacetic acid complex fall within the regular pattern observed for the other substituted acetic acid complexes.2 The ionic conductance arises from, at most, the dissociation of about lo-' part of the solute. The dipole moment of the complex is calculated from the concentration variation of the solution static permittivity ( E O ) and refractive index (n) using the Guggenheim equation.' The results are shown in Figure 2 where A = (€0 - n2) - (€1 - n12) (the subscript 1 refers to the solvent) is plotted against the molar con-
f'
4.4h0.2 3.8A0.2
./'
3.7k0.2 3.4f0.2
Results and Discussion Solid Complexes. The permittivity of solid samples of the pyridine complexes with dichloro- and with trichloroacetic acids are shown in Table I. Assuming additivity of the electronic polarizability of the components of the complex a square of the refractive index of 2.2 is calculated for both complexes. This is much lower than the permittivity of the solid complexes and indicates a large atomic polarization probably associated with proton transfer between the covalent and ionic forms. Such a high atomic polarization would not be expected if complex formation resulted in only a stretching of the carboxylic 0-H bond.6 Low-Frequency (6 Hx-310 kHz) Results on Solutions. The dielectric loss factor for the complex a t any one concentration with the range 0.01-0.40 M is proporThe Journal of Physical Chemistry
t
1.20
a
0.80
0.40 0.00 0.00
0.10
0,20 C(solute),M.
0,30
Figure 2. Guggenheim plot for dichloroacetic acid-pyridine complex in benzene solution.
(3) A. M. Thompson, Proc. Inst. Eleo. Engrs. (London), B103, 1
(1956). (4) 8. Roberts and A, yon IIippel. J. Appl. Phys., 17, 611 (1948); G.Williams, J . Phys. Chem., 63, 534 (1969). (6) H.Kramer, 2.Physik, 157, 134 (1969). (6) J. W. Smith, J. Chim. Phys., 61, 125 (1964). (7) E. A, Guggenheim, Trans. Faraday SOC.,45, 714 (1949).
NOTES 0.40
46 1
1
0
0.175 M solution
0.30
Zu,
0.20
0.10
0.00
8.50
9.50
9.00 Log
10.00
t:
Figure 3. Variation of the dielectric loss factor ( E ” ) with frequency for dichloroacetic acid-pyridine complex in benzene solution. The solid line follows the Debye equation with = 0.36 and 7 = 83 x 10-12 see for the 0.175 M solution see for and E”,,, = 0.31 and 7 = 83 X the 0.162 M solution.
centration of the complex. The slope of the graph (strictly as C --c 0) is proportional to the square of the solute dipole moment and gives a value of 6.00 D for the dichloroacetic-pyridine complex. High-Frequency (260 MHx-36 GHx) Results on Solutions. I n liquids the dispersion of the orientation polarization occurs a t microwave frequencies and the variation of the associated dielectric loss factor follows the Debye equation 6’’
=
(€0
- Em)----1
UT
+
d T 2
where E,, is the permittivity at low (or zero) frequency, ,E is the permittivity at frequencies just beyond where the orientation polarization ceases to contribute, a( = 2 ~ j is) the angular frequency, and T is the relaxation time of the polar species. The absorption maximum (e”, = (€0 - ,E )/2) is proportional to the
O.I5
0.12
ir
0
o
Lit. data (ref 2)
0.09 B *
0.06
0.03 0.00 L
‘
8.5
9.0
9.5
10.0
10.5
11.0
Log f.
Figure 4. Dielectric absorption of trichloroacetio acid-pyridine complex in benzene solution. The solid line follows the Debye equation with e”m = 0.108 and I = 83 X 10-12 Bee.
square of the dipole moment of the relaxing species. The experimental results for the dichloroacetic acidpyridine complex (complex concentration, 0.175 and 0.162 M) are shown in Figure 3. The experimental results are adequately described by a single relaxation time. There is no evidence of a second relaxation process which would be expected if proton jumping occurred at microwave frequencies. The observed relaxation time (83 X 10-l2 sec) is a reasonable value for the relaxation time of a rigid species but is very much sec higher than the relaxation time of about 55 X expected by interpolation of the results of Davies and Sobc.zyk.2 It is also larger than the relaxation time of 65 X sec reported for the trichloroacetic acid-pyridine complex. Consequently dielectric loss factor measurements on the trichloroacetic acid complex were repeated and extended over a much wider frequency range. These results are shown in Figure 4 together with the original datae2 The revised relaxation time of 83 X 10-l2 sec is identical with the relaxation time of the dichloroacetic acid complex. The increased volume expected on replacing the dichloroacetic acid by the larger trichloroacetic acid is probably offset by a closer approach of the components in the more ionic trichloroacetic acid complex. No satisfactory explanation is available for the large difference between the relaxation time of the dichloroacetic acid complex and the value of sec reported for the monochloroacetic acid 46 X complex.2 The dipole moment may be calculated from the maximum dielectric loss factor (E’’, = (eo - ,E )/2) using the Onsager equation as applied to dilute solutionss
where the subscripts 1 and 2 refer to the pure solvent and to the pure solute, respectively; Nz is the number of solute molecules per cubic centimeter; and em2 is the pure solute permittivity at frequencies where only the atomic and electronic polarizations are significant. The permittivity of the solute in the pure state (Table I) gives an adequate representation of em2. The dipole moment of the dichloroacetic acid complex is thus 5.45 k 0.10 D. This is lower than the 6.00 D calculated from static permittivity measurements. The difference between these dipole moments arises from the neglect of the atomic polarization contribution in the calculation from the static permittivity data. A dipole moment of 4.17 D is calculated for the dichloroacetic acid complex using the bond moments and configuration shown in Figure 5. Free rotation is assumed about the Cr-C2 bond ( a similar model assuming free rotation yields a calculated dipole moment (8) L. Onsager, J . Amer. Chem. Soc., 58, 1486 (1938). Volume 75,Number 8 FeBruarv 1968
462
COMMUNICATIONS TO THE EDITOR
/
Figure 5. Bond moments and assumed configuration for the dichloroacetic acid-pyridine complex.
of 2.4 D for ethyl dichloroacetate; the lit’erature value is 2.63 Dg). Complex formation leads to increased polarity in the 0-H----N bond. The dipole moment of
the complex (5.45 D) is reproduced if the moment of the (0-H----N) group is increased from 3.7 to 5.1 I). With an OH bond length of 1.02 A in the complex,2this leads to an effective charge of 2.8 X 10-’0 esu at, the 0 and H centers. I n the free acid the effective charge at the 0 and H centers (separation, 0.96 A) is 1.57 X lO-’O esu. Thus complex formation nearly doubles the ‘‘ionicity’’ of the OH bond. This result is not unreasonable in view of the approximations involved in the group moment calculations and the neglect of polarizability effects. They suffice, however, to establish the qualitative conclusion that the proton remains predominantly attached to the oxygen atom rather than to the nitrogen atom. (9) A. L. hfcClellan, “Tables of Experimental Dipole Moments,” W. H. Freeman and Co., San Francisco, Calif., 1963.
COMMUNICATIONS T O T H E EDITOR Charge Distribution in Some Alkanes and Their Mass Spectra
Si?.: I n the past 15 years several papers’-5 have appeared discussing the possibility of predicting the fragmentation pattern of organic ion-molecules on the basis of a proposal by Lennard-Jones and Hall,’ who supposed the probability of fragmentation to be larger for those bonds where the density of the net positive charge is larger. Some authors3-j have pointed out that such a proposal appears to be inconsistent with the experimental results, on the basis of a comparison of the mass spectra of n-alkanes (which show a pronounced maximum at Ca, sometimes C4, fragments) with their calculated net positive charge distributions (which have a maximum at the center of the molecule). Surprisingly enough, these authors do not seem to take into account what seems to us a primary question, Le., to what extent does the distribution of the ions collected by the mass spectrometer reproduce that originating from the first fragmentation of the parent ion, thus accepting a sort of unjustified identification between mass spectrum and fragmentation pattern. Xore recently, Lorquet5 makes the still more surprising statement that the Lennard-Jones and Hall hypothesis (which, he says, does not hold for n-alkanes, due to a “too evenly distributed charge”) is experimentally verified in the mass spectra of branched alkanes. He supports his statement by the mass spectra of five hydrocarbons. The Journal of Phgsical Chemislry
It should be pointed out, however, that all the alkanes considered by this last author have such a formula as to give C3 or Cq fragments on fission of the most highly charged bond, i.e., those fragments which are in any case the most abundant. I n fact, an inspection of the mass spectra of two branched alkanes of larger sizeeand of their net positive charge distributions (Figure l),calculated by us for comparison (by the same method as Lorquet’s), shows the following facts: (1) the only feature which clearly distinguishes these spectra from those of the n-alkanes with the same number of C atoms is the higher abundance of the ions c n - 6 and C n-8; (2) fragments which cannot originate by a single fragmentation (ie., those ranging from C7 to Clz for the first compound and from C9 to CIS for the second one) are roughly as abundant as in the spectra of the normal alkanes with the same number of C atoms. From these observations we must conclude that the shape of the spectra is largely determined by refragmentation phenomena; as a consequence, the well-known “weakness” of the tertiary bonds is observable only through
(1) J. Lennard-Jones and G. G. Hall, Trans. Faraday Soc., 48, 581 (1952). (2) R. Thompson, Conference on Applied Mass Spectrometry, Institute of Petroleum, London, 1953, p 154. (3) N. D. Coggeshall, J. Chem. Phgs., 30, 593 (1959). (4) K.Fueki and K. Hirota, N i p p o n Kagalcu Zasshi, 81, 212 (1960). (5) J. C. Lorquet, Mol. Phys., 9, 101 (1965). (6) The spectra were obtained from those of American Petroleum Institute Project 44, Carnegie Institute of Technology, Pittsburgh, Pa., by normaliring to 100 the sum of the abundances of all the ions.