NOTES
3363
that the initial thermoelectric power found on the first week after the preparation of electrodes was reduced by about 20 pv/deg in 3 weeks to the final steady-state value of about 130 pvldeg. The u value of trimethylbenzylammonium chloride may be estimated roughly to be -9.7 X deg-’, assuming B = 0.94.
Instability Constants of Silver-Amine Complexes in Isopropyl Alcohol’ 9
b
E
W
N
by J. L. Pauley and H. H. Hau2 Department of Chmistry, Kansas State College of Pittsburg, Pittsburg, Kansas (Received April 18, 1966)
1
0
I Tcmp.’C dT
20 -5
25 0
30 +5
Figure 1. Emf of combined thermocells of type 1 a t 0.01 M around the mean temperature 2 5 O , where the temperature on two terminal electrodes was kept a t 25” constantly.
of tetraalkylarnmonium chlorides studied here are found to be closely parallel with those of tetraalkylammonium bromides except tetra-n-butylammonium bromides. The value of -13.27 X lom3deg-’ found here for tetra-n-butylammonium chloride is far lower than the value of -22.94 X deg-’ given previously by one of the present writers for tetra-n-butylammonium bromide. However, the latter value seems to be somewhat too high, presumably because the observed value of (dE/dT)o for a combined thermocell system, LiBrBu4NBr, studied in the previous work, might have been somewhat too high (perhaps higher by several tens in pv/deg), when one remembers that in the previous work the measurements were carried out without allowing the cells to stand for a sufficiently long time as taken in the present work, because it was found in the present case of tetra-n-butylammonium chloride
Complex stability constants in aqueous systems have been investigated quite extensively, but comparatively little has been published regarding nonaqueous systems. Mead, Maricle, and Streuli3 have determined formation constants of a number of silver-amine complexes in acetone; Jonassen, Fagley, Rolland, and Yates4 have reported stability constant data for silverethylamine complexes in ethyl and isopropyl alcohols. Peard and Pflaum6have studied a number of complexes of heterocyclic amine-silver complexes in ethanol and acetonitrile, and several investigator^^-^ have contributed to various aspects of the effects of solvent on stability constants. Because of changes in dielectric constant, solvating ability, and Lewis acidity or basicity of the solvent as the solvent system is varied, complex formation constants should be significantly affected by changes in solvent. Dielectric constant effects might also be ~
~
(1) Presented before the Division of Physical Chemistry of the
American Chemical Society at the First Regional Meeting, Kansas City, Mo., Nov 1965. (2) Taken in part from the dissertation of H. H. Hau to the graduate school of Kansas State College of Pittsburg in partial fulfillment of the requirements for the Master of Science degree. (3) K. K. Mead, D. L. Maricle, and C. A. Streuli, AnaZ. Chem., 37, 237 (1965). (4) H. B. Jonassen, T. F. Fagley, C. C. Rolland, and P. C. Yates, J . Phys. Chem., 58, 286 (1954). (5) W. J. Peard and R. T. Pflaum, J. Am. Chem. SOC.,80, 1593 (1958). (6) L. G . Van Uitert, W. C. Fernelius, and B. E. Douglas, ibid., 7 5 , 2739 (1953). (7) L. G. Van Uitert and C. G. Haas, ibid., 75, 451 (1953). (8) D. C. Luehrs, R. T. Iwamoto, and J. Kleinberg, I m r g . Chem., 5 , 201 (1966). (9) L. G. Sillen and A. E. Martell, “Stability Constants of MetalIon Complexes,” The Chemical Society, London, 1964.
Volume 70, Number 10
October 1066
NOTES
3364
expected to be significant as reported in other ~ t u d i e s . ~ ~ ~ Table I: Effect of Background Electrolyte Concentration on Since measurements of complex stability are really pK1 Values for n-Butylamine-Silver( I) Complex a measure of the competition of solvent and ligand for ligancy sites, a decrease in the polarity of the solvent, Concn of LiClOa, M 0.8 0.5 0.2 0.0 as reflected by a decreased dielectric constant, should 9.1 9.2 9.4 10.4 pK i favor increased complex stability.
Experimental Section Reagents. The amines used were reagent grade and were freshly distilled before use except for the din-propyl-, diisobutyl-, and tributylamines. These latter were used as obtained. The isopropyl alcohol used was dried over activated synthetic zeolite before use. The silver perchlorate and lithium perchlorate were dried under vacuum and stored over desiccant until used. All additions and transfers of reagents were made in a drybox and solutions were stirred with dry nitrogen during potential measurements. Procedure. Silver ion activities were determined by a modification of the method of Bodlanderlo and Koch.” The reference half-cell consisted of silver-plated platinum electrodes12 in anhydrous isopropyl alcohol solutions 0.025 M in AgC104 and 0.5 M in LiC104. The variable half-cell consisted of a silver-plated platinum electrode in 0.5 M LiC104 solution in anhydrous isopropyl alcohol containing varying, but always large excess, amounts of amine and varying amounts of AgC104. For each complex studied, measurements were made on three sets of solutions each having a different amine concentration. In each set there were four solutions in which the amine concentration was keDt constant while silver ion concentrations were varied. The two cells were electrolytically connected by an aqueous agar bridge containing KNO,. The time of contact of the bridges with the alcoholic solutions was kept as small as possible to minimize diffusion effects. KN03 solutions in isopropyl alcohol were tried as bridge solutions but were found to be unsatisf a ~ t o r y . In ~ two different sets of measurements the KY03 concentration of the bridge solution was varied from 2 to 3 m. No significant change in the cell emf was noted, suggesting that junction potentials were small or at least independent of the concentration of the bridge solution and also of the concentration of the silver ion in the test and reference solutions. In one set of measurements, the concentration of the LiC104 background electrolyte was varied to determine the effects of ionic strength - on instability constant, Ki- (Table I). PKi values increased in a somewhat . regular way with decreasing electrolyte concentration, but no simple relationship between Ki and concentration or ionic strength could be observed. Measurements were made using an L and N student I
potentiometer and extended range galvanometer. Two pairs of essentially identical electrodes were compared for each emf determination in each solution to eliminate spurious electrode effects. Coordination numbers and instability constants were . calculated from observed potentials by the method of Jonassen, et aL12except that concentrations were used rather than activities in the absence of reliable data for activity coefficients. The relationship derived by these investigators is given by n
where pKi is the negative log of the instability constant, AgR+ is the silver ion concentration in the reference half-cell, E is the observed potential, n is the coordination number of the silver ion, LT is the total added ligand concentration, and AgT+ is the total silver ion concentration in the variable half-cell. A plot of -E/0.05916 vs. -%(log LT - nAgT+) log nAgT+ should give a straight line with a slope of 1.0 for n = 2 if the silver ion shows its normal coordination number and if the system is reversible.
+
Results ~
Average values of the slopes obtained on the basis of eq 1 are shown in Table 11. Values represent the average of three sets of experiments. For each run, silver ion concentrations varied from 0.0007 to 0.02 M . All results are the average of values for two independent pairs of electrodes. As can be seen, the slopes do not vary appreciably from 1 except perhaps for the diisobutyl, trilz-propyl, and tri-n-butylamine complexes. It is possible that in these complexes the silver ion may have a coordination number other than 2. One might predict, on the basis of steric effects, that the coordination number of the silver ion might be decreased for these amines and the lower value for the diisobutyl complex would be
_
The Journal of Physical Chemistry
(10) G. Bodlander as cited in S. Glasstone, “Textbook of Physical Chemistry,” 2nd ed, D. Van Nostrand Co., Inc., New York, N. Y., 1946.DD - _ 972-974. (11) F. K. V. Koch, J . Chem. SOC.,2053 (1930). (12) A. s. Brown, J . Am. Chem. SOC..56, 646 (1934).
3365
NOTES
Table 11: Slope of Plot of -E/0.0591 us. -n Log (LT- nAgT+) Run A
c
Amine concn, M
Amine
n-Propyl n-Butyl(0.2 M LiClOd) n-Butyl n-Butyl(0.8 M LiC104) Isobutyl t-Butyl Di-n-propyl Di-n-butyl Diisobutyl Triethyl Tri-n-propyl Tri-n-butyl 0
+ Log (AgT+) for Different Amine Complexes“ Run B
>
Slope
1.0 f O . 1
Amine concn, M
Run C
M
Amine concn, M
Slope
1.94 1.62
1.0 f O . 1
2.43 2.02
0.98f 0.1 0.98f 0.1
2.92 2.42
0.96 f 0.1 0.99=t0.1
1.62 1.62
1.0 2cO.l 0.95 0.3
2.02 2.02
1.0 *O.l 1.0 *0.1
2.42 2.42
0.99f 0.1 1.0 f O . l
1.60 1.52 1.17 1.78 1.73 1.58 1.59
1.0 f O . 1 1.0 f O . l 0.89 f 0.6 0.98f 0.1 0.92 f 0.3 1.08 f 0.1 1.48 f 0.4
2.00 1.90 1.46 2.02 1.96 1.87 1.80 1.85
0.98f 0.2 1.0 f O . 1 0.91f 0.3 0.99i0.3 0.86f 0.5 0.98=t0.3 1.46AI 0.3 1.22f 0.5
2.40 2.28 1.75 2.26 2.19 2.16 ...
0.97f 0.1 0.97f 0.2 0.92f 0.3 0.96i0.3 0.76f 0.3 0.96f 0.2 ...
...
...
...
Background electrolyte is 0.5 M LiC104, except as noted.
3.1
2.81 2.7
tri-n-bu tyl
‘t
0
di-n-butyl
Figure 1.
consistent with this. It does not seem reasonable, however, to postulate an increase in coordination number of the silver ion with the tri-n-butyl- and trin-propylamines in concentration ranges where less hindered amines show no such increase. It is thus concluded that the deviations noted for these three complexes represent failure of the equations to represent adequately the system rather than any real change in the coordination number of the silver ion. The independence of the slope upon the concentraobserved for the nOf t’he background butylamine complex suggests that the background electrolyte does not compete successfully with the amine for ligancy sites of the silver ion, and also that the 1iganCy O f the silver ion is not significantly affected by the electrolyte concentration. Various workers have reported a general relationship
between pKi and ~ K Bthe , negative log of the base l 3 l l 4 Experimental ionization constant of the values of pKi obtained in this investigation, plotted against ~ K values B for the various amines in water, values in water were are shown in Figure 1. used in the absence of such data in isopropyl alcohol and on the assumption that the relative values for the various amines would not be greatly changed in going from water to isopropyl alcohol. The amines fall clearly into three groups, primary, secondary, and tertiary. I n each of these groups a reasonably linear proportionality exists between pKi and ~ K B .The silver complexes formed from primary amines are about one to two orders of magnitude more stable than those with secondary amines, which are in turn two to three orders of magnitude more stable t,han complexes with tertiary amines. This is the order found by Brown1’ for the basic strength of ethyl-, diethyl-, and triethylamines with the highly hindered reference acid, tri-t-butylboron rather than the order when HC1 is used as the reference acid. This suggests that for the silver-amine complexes, steric effects are more significant than inductive effects in determining changes in stability with substitution. This does not
(13) R. J. Bruehlman and F. H. Verhoek, J. Am. Chem. soc., 7 0 , 1401 (1948). (14) J. Bjerrum, Chem. Rev., 46, 381 (1950). (15) H.K.Hall, Jr., J . Am. Chem. SOC.,79, 5441 (1957). (16)N. A.Lange, Ed., “Handbook of Chemistry,” 9th ed, Handbook Publishers, Inc., Sandusky, Ohio, 1956. (17) A. R. Day, “Electronic Mechanisms of Organic Reactions,” American Book CO., New York, N. Y.,1950,pp 220-221.
Volume 70, Number 10 October 1966
NOTES
3366
appear to be unique to the isopropyl alcohol solvent system since the same order is observed in water and acetone as shown in Table 111.
Table I11 : pKi of Silver-Amine Complexes in Different Solvents
Solvent system
Div p K i of silver-amine complexelectric nIsoDi-n- Diiso- Tri-nconstant Butyl butyl butyl butyl butyl
Isopropyl alcohol, 18.3 9.17 8.75 8.13 6.84 5.70 0 . 5 M LiClOa Acetone, 0 . 5 M 20.7 10.29 9.68 10.05 8.83 . . . LiClOa Water 78.54 7.48 7.24 4.10 . . . 3.82
Acknowledgments. This work was financed in part by an Academic Year Extension of the Research Participation Program at Oklahoma State University sponsored by the National Science Foundation. The assistance and suggestions of Dr. Paul Arthur of Oklahoma State University are gratefully acknowledged.
Kinetics of the Gas Phase Pyrolysis
of Tetranitromethane' by J. M. Sullivan and A. E. Axworthy Rocketdyne, A Division of North Ameriean Aviation, Canoga Park, California (Received May 10, 1966)
Jonassen4 and others have suggested that the pKi of amine complexes should increase as the dielectric constant of the solvent decreases and in the case of ethylamine in the solvents water, ethanol, and isopropyl alcohol such a correlation appears to exist, perhaps because of the very similar nature of the solvents. That this is not the only factor involved can be seen by comparing the values of pKi in isopropyl alcohol, acetone, and water as shown in Table 111. Although it is often difficult to compare results of different investigations, the comparison shown in Table I11 would suggest that specific solvent effects must play a significant factor in complex stabilities although dielectric constant effects almost certainly are of importance. Further studies involving different solvent systems may be useful in determining the relative importance of these factors.
Conclusions Complexes of silver with amines fall into three groups, those with primary, secondary, and tertiary amines. Within these groups, the stability of the complexes correlates with the base strength of the amine. It does not, however, correspond to the order of the base strength of the amines in going from primary to secondary to tertiary, suggesting that steric hindrance is a significant factor. The relative stabilities of the amine complexes do not appear to change with changes in solvent although the solvent does affect the absolute values of instability constants. The dielectric constant may have a significant bearing on solvent changes on stability of complexes but specific solvent effects are also significant and may, in the case of solvents of quite different nature, be more important than the effects of dielectric constant. The Journal of Physical Chemistry
Reported herein are the results of our study of the pyrolysis of tetranitromethane (TNM).
Experimental Section The 90.6-ml monel electrically heated, stirred flow reactor was identical with that described by Sullivan and TNR4 vapor of 99% purity was mixed with helium and passed through the reactor at 1 atm total pressure and a TNM initial partial pressure of 1.09 mm (runs 1-23). The storage tank was repressurized with helium prior to run 24 giving an initial TIVM partial pressure of 0.48 mm in the final eight experiments. The reactor was by-passed to permit analysis of the unreacted TNM. A sample of the gas stream from the reactor (or the by-pass) was periodically analyzed for TNM by introducing it into the helium carrier gas stream of a gas chromatograph by means of an unlubricated Beckman gas sampling valve. The reduction in TNRI concentration with residence time in the reactor was followed chromatographically using a 3-ft column of 5% S.E.30 silicone oil on Celite, 60-100 mesh, at room temperature and a thermal conductivity detector. The flow rates through the reactor were measured with a soap bubble flow meter connected to the exit stream. The measured flow rates and reactor volume were corrected to reactor temperature. lSiIass spectrometric analysis of the decomposition products showed the presence of the species NOz, NO, N20, and COz. The presence of NO2 and COZ (1) This work was supported by the U. S. Air Force and the Advanced Research Projects Agency under Contract No. AF04(611)9380 and ARPA Order No. 24. (2) J. M. Sullivan and T. J. Houser, Chem. I d . (London), 1057 (1965).
