1074
NOTES
Figure 2. Spectral changes resulting from the exchange of sepiolite with DIO: (a) sepiolite after 30-min evacuation a t 350'; (b) after 30-min exchange with DnO and 30-min evacuation, both a t 350"; (c) after 24-hr exchange with D20 and 30-min evacuation, both a t 350'.
phenomenon for dioptase.) If this is correct, it would indicate that the adsorbed molecule of water is linked through its oxygen atom to the surface. The most likely site is an electrophilic species, probably a magnesium ion. This spatial arrangement allows the hydrogens freedom to produce two sharp bands in the normal hydroxyl stretch region, similar to the spectrum observed from a dilute solution of H20 in carbon tetrachloride. Dehydration above 400" is predominantly irreversible and tkte regeneration of bands C and D lost in this treatment does not occur on the addition of water vapor a t 25" to the calcined specimen. The hydroxyl stretch bands A and B, one of which, A, rapidly exchanges with D 2 0 and interacts with adsorbed ethylene and pyridine, must originate from hydroxyl groups located in different environments. Angell and Schaff er6 observed hydroxyl bands at 3745, 3688, 3643, and 3540 cm-' for a magnesium-exchanged Y-type zeolite and assigned them, respectively, to SiOH groups, possibly AlOH, a proton which has combined with surface oxygen to give SiOH adjacent to 3-coordinated aluminium, and H-bonded OH. It seems likely that the band detected at 3740 cm-l (A) in this investigation arises from free hydroxyls located on silicon atoms, since all forms of dehydrated silica display absorption in the range 3740-3750 cm-l. The sharp band at 3674 cm-' is assigned either to a hydroxyl on a magnesium ion, corresponding to the 3688-cm-' band of Angell, or to hydroxyl produced by the reaction of The Journal of Physical Chemistry
dehydrated protons with the neighboring oxygen lattice. As the band is observed after room-temperature evacuation, the latter explanation is less plausible. Adsorbed pyridine did not exhibit spectra typical of a pyridinium specie^,'^^ but, as adsorption affected only one of the four bands, this does not constitute a very comprehensive test. Bands a t 1695 and 1425 em-' have not been identified. (6) C.L.Angell and P. C. Schaffer, J . Phys. Chem., 69,3463 (1965). (7) E.P.Parry, J . Catal., 2 , 371 (1963). (8) M.R. Basila, T. R. Kantner, and K. H. Rhee, J . Phys. Chem., 68, 3197 (1964).
Aliphatic Semidiones.
VIII.
Hyperfine Splitting by Oxygen and Carbon Atoms in Semidiones'
by Glen A. Russell and Graham R. Underwood Department of Chemistry, Iowa State University, Ames, Iowa 60010 (Received September 6,1967)
Hyperfine splitting by 170has been observed in the esr spectra of nitroxides,2 p-benxosemiquinone,s nitrobenzene radical anions,' and substituted phenoxy radi~als.~
1075
NOTES We have measured u0 for the radical anions produced by reduction of structures 1-3. It is recognized that structure 3 yields a trans-semidione (RC(0 -)=C(O-)-
Table I : Hyperfine Splitting Constants (G) in Dimethoxyethane Solution a t 25" Substrate
R).6
0
il
lb 2
(CHs)sC C(CHs)s 1
3
bH1 0.12(18) 2 82 (2), 0.29 (2)' 0.34' I
-11.1 -9.95 -10.5
52.7(1),7.7(6) 6.15 (2)d 3 . 8 (2),d 2 . 4 (6)
a All a0 values assumed to be negative. Line widths increased with increasing magnetic field: E. de Boer and E. L. Mackor, J. Chem. Phys., 38, 1450 (1963). * Previously reported (N. Hirota and S. I. Weissman, J. Amer. Chern. Soc., 82, 4424 (1960)) uH = 0.12 and UC = 7.6 (6) and 49.6 G in tetrahydrofuran solution. Previously reported: G. A. Russell, G. W. Holland, and K.-Y. Chang, ibid., 89,6629 (1967). Total intensity was approximately 2% of spectral intensity. ' Reference 6. Also reported as 0.31 G (G. A. Russell and E. T. Strom, ibid., 86, 744 (1964)) in dimethyl sulfoxide (80y0)-t-butyl alcohol (20%).
n
4
0
n
In other cases, the reduction of the a-diketone could conceivably occur with decarbonylation to yield the ketyl RCOCOR
+e
---t
RzCO--
[C6H6COCOCeH6]* -
+ Qcoopc
I ,
GAUSS
A
I
+ co
constant might serve as a definitive test for the ketyl (1) or semidione (3) structure. The introduction of 170into structures 2 and 3 occurred readily by treatment of the diketone with "0-enriched (-10%) water (YEDA Research and Development Co., Ltd., Rehovoth, Israel) in dimethoxyethane solution. For structure 1, the ketone and water in the presence of 1.5 M hydrogen chloride in dimethoxyethaine solution were heated a t 200" (sealed tube) for 3 days. Upon cooling, structures 1 and 3 separated from an aqueous layer. For structure 2, the solvent was removed under reduced pressure. The substrates were reduced in dimethoxyethane at 25" by sodium-potassium alloy. Typical spectra are given in Figures 1 and 2 and the coupling constants are listed in Table I. On the basis of the general theory of Karplus and Fraenkel,* the value of u0 is expected to be given by eq 1 for the radical anions derived from structures 1-3.6 Values of Qo' = -40 and &coo = -17 G have been u0 = Q o O p o
20
10
1
+ CO
Processes of this type occur readily for triketones: for example, in the electrolytic reduction of l,&diphenyl1,2,3-pr0panetrione.~ The investigation of u0 in structures 1-3 was undertaken to establish if this splitting
C&,COCOCO~&,H6
0
I
(1)
~uggested.~These values lead to a predicted value of -9.95 G for structure 2 and -12.9 G for structure 3 (estimating PO = 0.11 and pc = 0.35 for the cis-semi-
l
l
C
? Figure 1. First-derivative esr spectrum of 170-enriched potassium ketyl in dimethoxyethane solution a t 25". Two 13Csplittings (natural abundance) are also seen. The I7O hyperfine splittings show line width broadening as the magnetic field increases. At the concentration of ketyl employed, UH = 0.12 G (18 hydrogens) is not seen.
dione (2), and PO = 0.20 and pc sernidione (3)).
=
0.29 for the truns-
(1) This work was supported by grants from the National Institutes of Health (GM 13,000-02)and the National Science Foundation (GP 6402X). (2) J. C.Baird, J . Chem. Phya., 37, 1879 (1962). (3) W. M.Gulick, Jr., and D. H. Geske, J. Amer. Chem. Soc., 88, 4119 (1966). (4)W. M.Gulick, Jr., and D H. Geske, ibid., 87, 4049 (1965). (5) (a) K.Dimroth, F. BLr, and A. Berndt, Angew. Chem., 77, 217 (1965); (b) K.Dimroth, A. Berndt, F. BLr, R. Volland, and A. Schweig, ibid., 79, 69 (1967). (6) G.R. Luckhurst and L. E. Orgel, Mol. Phys., 7, 297 (1963). (7) G.A. Russell and S. A. Weiner, J . Amer. Chem. Soc., 89, 6623 (1967). (8) M.Karplus and G. K. Fraenkel, J. Chem. Phys., 35,1312 (1961). (9) G. A. Russell, E. T. Strom, E. R. Talaty, and S. A. Weiner, J . Arne?. Chem. Soc., 88, 1998 (1966); G. A. Russell and R. S. Stephens, J . Phys. Chem., 70, 1320 (1966). Volume 78,Number 9 March 1968
1076
SOTES trans-biacetyl (5.6 G) radical anions in dimethyl sulfoxide solution. The absence of a third a' for structure 1 is surprising. Perhaps the 7.7-G a' is due to eight magnetically equivalent carbon atoms (the six methyl carbons and the two quaternary carbon atoms). The absence of detectable splitting due to UCO' in structures 2 and 3 is a puzzling feature. For the transbiacetyl radical anion very sharp lines have been reported, but only a single a' is seen (4.5 G). Here, any other value of a' would have to be less than 2 G and probably less than 1 G.14
Figure 2. First-derivative esr spectrum of potassium salt of l70-enriched semidione in dimethoxyethane solution at 25". The central three lines of the spectrum due to uH = 2.82 G ( 2 hydrogens) are off scale, as are a pair of satellites that coincide with two of the I7O splittings. The 1% satellites (a0= 6.15 G ) are easily seen in the absence of I7O splittings. At lower concentrations (better resolution), a second triplet, uH = 0.29 G, is seen.
McLachlan spin densities,l0 calculated to give the best fit with the available yielded values of &coo = -20 and Qo0 = -32 G. Applying these constants yielded predicted values of a0 of -10.5 (cis) and -12.2 (trans) G. However, it is obvious from the magnitude of &coo and Qo0 preferred by others3 that no reasonable spin distributions in the di-t-butyl ketyl could possibly account for an oxygen hfsc of - 11.1 G. It does appear that using &COO = 0 and Qo0 = -40 & 4jb G actually provides a reasonable fit with all data if po is assumed to be about 0.25 for the radical anions derived from structures 1-3. The values of U C H and ~ ~ UGH: for the radical anions derived from structures 1 and 3 are surprising. The ketyl with the highest carbonyl carbon spin density gives the smallest hydrogen coupling and the largest 13CH3coupling. Perhaps there is a sign reversal in the hydrogen coupling. In any event, the mechanism of this coupling is not clear." The l3C hyperfine splittings assigned to two identical carbon atoms in the semidiones derived from structures 2 and 3 could be due to the carbonyl carbon atoms12 or the a-carbon atoms. In view of the assignment of ~CH: in ethyl radical as 13.57 G,13 the splitting by the a-carbon atoms would be expected to be pc(Qc!cc). From Fessenden's data, we estimate QCQ' = 13.57/ [l - (3(26.9)/508)] = 16 G, and using pc 0.25, a value of aaC 4 G is predicted. The values found are 6.15 G for structure 2 and 3.8 G for structure 3. The difference in values of a' found for structures 2 and 3 is in accord with the conclusions that the cis-semidiones possess a higher spin density on the carbonyl carbon atom than the trans isomer.g This is reflected in the values of U C H in ~ ~ the cis-biacetyl (7.0 G) and
-
The Journal of Physical Chemistry
-
(10) A. D. McLachlan, Mol. Phys., 3, 233 (1960). (11) K. H. Hausser, H. Brunner, and J. C. Jochims, ibid., 10, 253 (1966). (12) E. T. Strom and G. A. Russell, J . Chem. Phys., 41, 1514 (1964). (13) R. W. Fessenden, J . Phys. Chem., 71, 74 (1967). (14) Diisopropylsemidione ((CH,)zCHC(O.)=C(O-)CC(CHs)~I-I)has been prepared by Mr. H. Mallcus with 13C a t the carbonyl and methine positions. The observed values of ac are 0.8 and 4.0 G, re-
spectively. Biacetyl radical anion (CHsC(O.)CHs) has been labeled with 13C at the methyl position by Mr. D. Lawson. The values of ac observed are 5.2 =t0.1 and 4.5 G for the cis and trans structurea, respectively. The carbonyl carbon atoms in biacetyl radical anion 2 G. have also been labeled, ac = 0.50 (trans) and ~ 1 . (cis)
The Argon- Sensitized Radiolysis of Methane and Ethane in the Liquid Phase' by Norman V. Klassen2 Radiation Research Laboratories, Mellon Institute, Pittsburgh, Pennsylvania (Received September 22, 1967)
A study of the argon-sensitized radiolysis of liquid ethylene was reported p r e v i o ~ s l y . ~Energy transfer from argon to ethylene in the liquid phase was found to be very efficient. The G values (based on the total energy absorbed by the solutions) of acetylene, ethane, and n-butane were similar for the radiolysis of an argon solution containing only 0.05 mol % ethylene and for the radiolysis of pure ethylene. It seemed significant that the G values of all measured products were constant when the ethylene concentration of the irradiated solutions was varied between 1 and 0.05%. The present report of the radiolysis of argon-methane and argon-ethane solutions in the liquid phase shows a similar pattern of product formation with change in hydrocarbon concentration.
Experimental Section The methane and ethane were purified by gas chromatography to reduce hydrocarbon impurities to levels (1) Supported in part by the U. S. Atomic Energy Commission. (2) Division of Applied Physics, National Research Council of Canada,
Ottawa, Ontario. (3) N. V. Klassen, J . Phys. Chem., 71, 2409 (1967).
