3909
ACIDDISSOCIATION IN ACETONE-WATER MIXTURES was sufficiently large to account for all the Nz produced during the methylamine photolysis. Therefore, we feel that the N2 we observe comes from the polymer photolysis and not from decomposition of ammonia or other nitrogen-containing compounds as suggested in the work a t longer wavelengthsea We made a determined effort to detect any HCN which might have been produced in this system. At the wavelength employed it is energetically possible that excited species such as CD2NHz*or CDBNH*produced in the initial step could further decompose to hydrogen and HCN. We employed both gas chromatography and mass spectrometry in searching for HCN but we observed only insignificantly small amounts of this product.
Conclusion The photolysis of methylamine proceeds primarily by an atomic mechanism with molecular elimination processes representing relatively minor contributions to the overall photodecomposition. Surprisingly, we find that D atom formation in CD3NH2photolysis is more important than H atom formation. This is contrary to the methylamine photochemistry a t longer wavelengths. The deposition of copious quantities of polymer on the walls and window of the reaction cell severely complicates this system and limits the quantitative data to products noncondensable at 77°K. The molecular yields we report are to be considered as maximum amounts and this is particularly true in the case of HD formation.
Acid Dissociation in Acetone-Water Mixtures.
An Anomalous Medium
Effect When London Dispersion Forces Are Large
by Dodd-Wing Fong and Ernest Grunwald Department of Chemistry, Brandeis University, Waltham, Massachusetts 06164 (Received April 29, 1869)
+
2,4,6,2’,4’,6’-Hexanitrodiphenylamine(HNDA) undergoes acid dissociation according to HNDA ~t H+ NDA-. At 25O, ~ K =A2.75 in water, 2.13 in 12 wt % acetone-water, and 1.26 in 24 wt % acetone-water; 0.1 144 p-cyanophenol and 0.1 M p-methoxyphenol likewise raise the dissociation constant. The direction of these changes is contrary to prediction based on the charge-in-dielectric-continuummodel, but is consistent with the expected large increase in solvation energy of NDA-, when acetone is added to water, as a result of London
dispersion forces.
For acid dissociation of the charge type HA = H+ the usual medium effect when acetone is added to an aqueous solution is for KA to decrease.’ This effect is attributed primarily to the lowering of the di\ / OZN NO2 electric constant, which opposes the formation of the (HNDA) ions.2 The change in solvent basicity is thought to be small, as long as water remains in large molar e x c e ~ s . ~ The change in solvation energy by van der WaalsLondon dispersion forces is likewise thought to be small, or at least not to be dominant. We find, on the contrary, that for acid dissociation of 0,N 2,4,6,2’,4’,6’-hexanitrodiphenylamine(HNDA),4J K A (NDA-) not only fails to decrease but actually increases by more than SO-fold as acetone, up to 24 wt %, is added to an (I) J. F. J. Dippy, 8. R. C. Hughes, and A. Rosanski, J . Chem. SOC.,1441 (1969), and references cited therein. aqueous solution. Our data are listed in Table I, and (2) W. F. K. WynneJones, Proc. Roy. SOC.,A140, 440 (1933). ~ K values A covering the entire acetone-water system Acetone is a much weaker base than water. See, for example, (3) are cited in Table 11. We find, moreover, that p-methN. C. Den0 and M. J. Wisotsky, J. Amer. Chem. SOC.,85, 1736 oxyphenol and p-cyanophenol also raise the KA of (1963).
+ A-,
Volume 73, Number I1 November 1969
3910
DODD-WINO FONO AND ERNEST GRUNWALD
Table I : ~ K ofA2,4,6,2',4',6'-Hexanitrodiphenylamine a t 25" in Acetone-Water Mixtures
0.0
2.701 3.000 3.230 3.421
11.75 f 0.43 15.34 3: 0.17 18.09 f 0.21 19.49 f 0.32
0.96 f 0.07 1.78 f 0.06 3.07 f 0.15 4.35 f 0.41
2.03
2.002 2.223 2.399 2.700 3.002
11.3Sd 13.73d 16.21 f 0.04 20.76 =t 0.12 23.08 f 0.06
0.80 1.13 1.68f 0.01 4.06 f.0.12 8.24 =k 0.20
4.06
1.524 1.700 1.855 1.922 2.002
34.64f 0.21 38.47 f 0.09 39.55 f 0.72 42.16 f 0.21 43.98 f 0.26
2.02 f 0.04 2.90 f 0.03 3.24 =t 0.27 4.40 f 0.12 5.66 f 0.23
2.72 f 0.03 2.75 f 0.01 2.74 f 0.02 2.78 f 0.04 (Av) 2.75 f 0.02 2.11 2.17 2.17 f 0.01 2.09 f 0.01 2.11 f 0.01 (Av) 2.13 f 0.04 1.22 f 0.01 1.24f 0.01 1.34f 0.03 1.28f 0.01 1.25=t 0.01 (Av) 1.26 f 0.04
'
poH = - log [H+], where [H+] = molar concentration of hydrogen ion at a given solvent composition. Mean and standard deviation, based on several absorbancy measurements between 4400 and 5000 A. ' KA = [H+][A-] /[HA], in molar concentrations. No attempt is made to estimate molar activity coefficients. Initial measurement.
Table 11: Effect of p-Methoxy- and p-Cyanophenol and of Acetone on ~ K ofA2,4,6,2',4',6'-Hexanitrodiphenylamine at 25" Acetone, wt %
0
12b
Phenol, 0.1 M
...
...
p-OCHa p-CN
24"
...
p-OCH3 p-CN
42 76.5 100
... ... ...
PKA~
2.75 f 0.02 2.13 f 0.04 1.83 f 0.03 1.73 zk 0.01 1.26 f 0.04 1.12 f 0.01 1.15f 0.05 2.5 f O . l d 2.9 f O.ld 3.3 f 0.3d
a Unless otherwise specified, KA is the equilibrium constant in terms of molar concentrations at the given solvent composition. 2.03M ; Table I. 4.06 M ; Table I. Results extrapolated to infinite dilution. D. J. Glover, ref 5.
'
HNDA in water-rich media (Table II), so that this unusual effect is not characteristic of acetone alone. I n fact, on a molar basis the effects produced by the phenols are even greater than those produced by acetone. In 12 wt % acetone, 6 log K A / ~isc3 for p-methoxyphenol, 4 for p-cyanophenol, and 0.37 for acetone, where c denotes the molar concentration of added phenol or acetone. The fact that the two phenols produce similar effects, in spite of the different electron donor-acceptor properties of their substituents, seems to rule out possible explanations that involve the formation of specific charge-transfer complexes. We believe that the medium effect on K A is in this case dominated by London dispersion forces. Some years ago, Grunwald and Price applied London's The Journal of Physical Chemistry
theory for complex moleculesa to the prediction of medium effects on relative acid strength and described certain conditions under which the dispersion forces should be especially large.' The following conditions apply here. The acid, HNDA, absorbs in the ultraviolet while the conjugate base also absorbs strongly in the visible. Thus the chromophoric groups (the amino-nitrogen atom and the six nitro groups) become very strong centers of van der Waals attraction in the NDA- ions,' and we must expect a relatively large change in solvation energy by London dispersion forces on acid dissociation. Any change in solvent that improves the van der Waals attraction of solvent molecules for solute molecules will tend to promote acid dissociation. Our data suggest the following interpretation. Water is a poor solvating agent by the London dispersion mechanism? When acetone or phenol is added to water, solvation by the London dispersion mechanism improves and the gain in interaction energy is greater for the NDA- ion than for HNDA. Up to 24 wt % acetone, this effect is large enough to outweigh the opposing effect of the lowered dielectric constant, and our data show that KA increases. At still higher concentrations of acetone, the balance of opposing effects changes, and Glover's datae show that K A decreases above 40% acetone (Table 11). Because of the dramatic increase in K Aup to 24 wt % acetone disclosed by our data, we believe that there is (4) R. Stewart and J. P. O'Donnell, J . Amer. Chem. SOC., 84, 493 (1962); Can. J. Chem., 42,1694 (1964). (5) D. J. Glover, J.Amer. Chem. SOC.,87,5275 (1965). (6) F. London, J . Phys. Chem., 46,305 (1942). (7) E. Grunwald and E. Price, J . Amer. Chem. SOC.,86, 4517 (1964).
ACIDDISSOCIATION IN ACETONE-WATER MIXTURES an enrichment of acetone in the solvent shells around the NDA- ions. However, the NDA- ions are fairly large and can accommodate quite a few acetone.molecules on adjacent sites. The displacement of water molecules by acetone molecules seems to take place gradually, and its effect on KA does not become saturated until the acetone concentration is in excess of 24 wt %. That saturation occurs eventually is shown by the much smaller effect of added phenol in 24 wt % acetone than in 12 wt % acetone (Table 11).
Experimental Part
Materials. Fisher certified acetone was distilled once before use. Eastman White Label 2,4,6,2',4',6'hexanitrodiphenylamine (HNDA) and p-methoxyphenol were recrystallized from acetone and dried in vacuo. p-Cyanophenol from Aldrich Chemical Co. was recrystallized from methanol-water and dried in vacuo. Chemically pure, deionized water was used without further purification. KA Measurements. To obtain KA, we measured [NDA-] spectrophotometrically in solutions of known pcH and known total [HNDA] concentration. (Quantities in brackets denote molar concentrations. p,H = -log [H+].) Optical densities were recorded between 4400 and 5000 A at 25", using a Beckman model DU-2 spectrophotometer and matched l-cm cells. Absorption of light in this range is due entirely to the NDAions. Stock solutions of the separate reactants were brought to 25" before mixing, so that after mixing, it required only a few minutes to reach thermal equilibrium in the thermostated cell compartment of the spectrophotometer. Optical densities were measured 10-15 min after mixing. The pcH of the solutions was measured with a calibrated Beckman research pH meter. At the higher HC1 concentrations (>0.02 P ) , it could be assumed that [H+] is equal to the molar concentration of added HC1 because lHNDAl is relatively small. Solutions with [H+] = 0.03 M served as pH standards to make the pH meter direct reading in pcH for measurements in the more dilute HC1 solutions at the given solvent composition. Experience accumulated in this laboratory has shown that pcH measured in this manner is equal
3911 to -log [H+] at the given solvent composition within close limits. The concentrations of un-ionized HNDA and of NDA- were less than 60 p M in all measurements. The pH ranged from 1.5 to 3.5. The data in Table I are typical. Molar extinction coefficients in water were measured for 38 pM [NDA-] in the presence of 0.05 F NaOH. Our results were in good agreement with Glover's.5 Although alkaline solutions of NDA- in water have optical densities that are stable for days, solutions of [NDA-] in 0.05 F NaOH in 12 and 24 wt % acetone change color soon after mixing of the components, and the optical density is highly unstable. However, stable optical densities could be obtained in both mixed solvents when NaOH was added so as to be only slightly in excess of the HNDA, which was on the order of 50 pM. In 24 wt % acetone, molar extinction coefficients obtained under such near-neutral conditions were identical with extinction coefficients obtained for solutions containing only 50 p M HNDA. It should be noted that acid dissociation is expected to be virtually complete for an acid of ~ K 1.26 A at a formal concentration of 50
PM.
The molar extinction coefficients in 12 and 24 wt % acetone were only slightly different from those in water, in agreement with G l o ~ e r showing ,~ that except for solvation effects, the light-absorbing species is the same. Because solution of solid HNDA in the acetonewater mixtures was inconveniently slow, the solid was dissolved in pure acetone and the resulting solution was diluted with water. Some of the aqueous acetone solutions were found to be metastable in the presence of HC1, because HNDA crystallized out overnight. However, crystallization was so slow as to be imperceptible during the time required to measure the optical densities, except in 12 wt % acetone below pH 2.3, where the change was less than 10% in 15 min. Here KA was calculated from the initial readings.
Acknowledgment. It is a pleasure to thank Professor Allan K. Colter for helpful discussions. We gratefully acknowledge the support of this work by the Petroleum Research Fund of the American Chemical Society and by the National Science Foundation.
vo'olume78,Number 11 November 1060
