Aug. 20, 1964
CHARGE-TRANSFER COMPLEXES IN SOLUTION
only 2-5Yc isomer conversion resulted. Poassium ferrioxalate actinometry was employed.82 Tubes containing solutions of potassium ferrioxalate (0.006 F ) in 0.1 '%1 sulfuric acid were irradiated prior t o , during, and a t the end of each experiment. The production of ferrous ion was determined by measuring the optical density of its complex with 1,lO-phpanthroline. The extinction coefficient of the complex a t 5100 A , , (1.132 f 0.003) X l o 4 1. mole-' cm.-', was determined by the recommended procedures2 using a standardized solution of ferrous +fate. The quantum yield for ferrous ion production a t 3660 A . was taken as 1.21.82,86 Excitation Energies.--A detailed study of ,phosphorescence spectra from sensitizers used in this work will be published shortly.89 The values of triplet excitation energies (used) refer t o 0-0 band maxima in phosphorescence spectra of sensitizer solutions in glasses consisting of methylcyclohexane and isopentane ( 5 :1 b y volume) a t i 7 " K . The emission spectra of biacetyl and benzil were also measured at room temperature in various solvents. The phosphoroscopes used in such experiments have been d e s ~ r i b e d .In ~ ~some ~ ~ ~cases phosphorescence could not be detected and the excitation energies refer t o 0-0 bands in So+ TI (85) T h e higher value (1.7) reported b y Lee a n d SeligergB is probably in error, T h e ferrioxalate actinometer has been checked against uranyl oxalate actinometrya2 9' and against benzophenone-benzhydrol actinometry88 a n d in all cases the values of Hatchard a n d Parker appear t o be correct. (86) J. Lee and H . H . Seliger, paper presented a t Rochester Photochemistry Symposium, M a r c h , 1963. (87) J. H . Baxendale and N. K . Bridge, J . Phys. Chem., 5 9 , 783 (1955). (88) D. Cowan, W. H a r d h a m , a n d G. S. H a m m o n d , unpublished results. (89) W. G . Herkstroeter, A. A. Lamola, and G . S . Hammond, unpublished results. (90) P. A. Leermakers, G . W. Byers, A. A. Lamola, and G. S . Hammond, J . Am. Chem. SOC.,8 6 , 2670 (1963).
[CONTRIBUTION FROM
THE
3217
absorption spectra measured in ethyl iodide, carbon disulfide, or in the presence of high pressures of oxygen. Such values were taken from the literature .49,91-g4 The triplet-state energy of duroquinone was estimated by assuming it t o be identical with the excitation energies of 2,3dimethyl-l,4-benzoquinoneand 2,5-dimethyl-l,4-benzoquinone. 95 T h e phosphorescence spectra of 4-acetylbiphenyl and 2-acetylfluorene were recorded and excitation energies of 67.5 and 64.1 kcal./mole were indicated for these two compounds, respectively. These values were not used because the presence of emitting impurities was suspected. I n view of the stilbene stationary states established in the presence of these sensitizers and the fact t h a t the 0-0 band of the phosphorescence spectrum of 4-benzoylbiphenyl is reported to be a t 60.6 k c a l . / m ~ l e ,we ~ ~ tentatively assumed t h a t the S O phosphorescence bands for 4-acetylbiphenyl and 2-acetylfluorene are close t o 60.6 k c a l . / m ~ l e . ~ ~ The 0-0 band of the phosphorescence spectrum of benzanthrone was estimated t o lie a t about 46.0 k ~ a l . / m o l e . ~ * (91) D. S . McClure, J . Chem. Phys., 17, 905 (1949). (92) D. P . Craig and I. G. Ross, J. Chem. Soc., 1589 (1954). (93) S . P . McGlynn, M. R. Padhye, and M. Kasha, J. Chem. P h y s . , 1 3 , 593 (1955). (94) M. R. Padhye, S . P. McGlynn, and M. Kasha, ibid., 14, 588 (1956)' J a p a n , 36, 295 (1962). (95) A. Kuboyama, Buli. Chem. SOC. (96) V. Ermolaev and A . Terenin, J . chim. phys., 66, 699 (1958). (57) Kote t h a t the energies of 2-acetonaphthone and %naphthyl phenyl ketone are essentially identical. (98) T h e phosphorescence spectrum of 3-bromo-7H-benz [d,e]anthracen-7one (3-hromobenzanthrone) in heptane a t 77'K. shows a 0-0 band a t 45.7 kcal./mole.~Q (99) D . S . Shigorin, S . A. Shcheglova, and N. S. Dokunikhin, Doki. Akad. Nauk S S S R , 137, 1416 (1561); Proc. Acad. Sci. Phys. Chem. Sect., 137, 371 (1961).
UNIVERSITY OF MISSISSIPPI, USIVERSITY,MISSISSIPPI]
Charge-Transfer Complexes in Solution. I. Spectrophotometric Studies of Aromatic Hydrocarbon-Aromatic Nitro Compounds Dissolved in Carbon Tetrachloride BY N. B.
JURINSKI A N D
P. A. D.
DE
MAINE'
RECEIVED F E B R U A R26, Y 1964 Here there is reported an ultraviolet absorption spectral study of the charge-transfer complexes formed between certain methylbenzenes and selected aromatic nitro compounds in carbon tetrachloride a t 20 and 45O, Toluene, m-xylene, mesitylene, durene, pentamethylbenzene, and hexamethylbenzene were the donors used. Acceptors were nitrobenzene, m-dinitrobenzene, p-dinitrobenzene, sym-trinitrobenzene, and sym-trinitrotoluene. The new computer method2 designed t o eliminate ambiguities from the literature was used to process all experimental information. For a given donor, the absorptivity of the complex ( a c ) increases a s the strength of the acceptor is increased by increasing the number of nitro groups. This fact is predicted by the charge-transfer theory. T h e relative order of QC for each donor is nitrobenzene, m-dinitrobenzene, p-dinitrobenzene, symtrinitrotoluene, and sym-trinitrobenzene. This contrasts with the well-known contradictions of Mulliken's charge-transfer hypotheses reported for complexes of a given acceptor with d o n x s of increasing strength. Formation constants, computed with the assumption t h a t only a 1: 1 complex is farmed, vary with the wave length, and the absorptivities of the complexes vary with the temperature. These facts are explained by the simultaneous formation of isomeric 1 : 1 and higher order complexes. Average formation constants for each of the twenty-two systems studied a t 20 and 45" are given.
Introduction nitro systems. From these works it has been deter. mined8 t h a t steric effects play an important role It has been well established that aromatic nitro in complex formation. Further, it has been shown compounds are capable of forming complexes of the charge-transfer type as described by M ~ l l i k e n ~ - ~ t h a t an increase in the number of nitro groups in the acceptor molecule produces an increase in the degree of and more recently by Dewar and Lepley.6 Bier' association as shown by the magnitude of association has investigated sym-trinitrobenzene complexed with constants with a number of amines.1° Thompson and aromatic amines and hydrocarbons in chloroform. de Maine2 have studied the effect of solvent on the Foster and co-workerss-l0 have studied a number of sym-trinitrobenzene-naphthalene system. A change (1) Author to whom inquiries should be addressed a t the Chemistry of solvent was found to be able to vary the magnitude Department, University of California, S a n t a Barbara, Calif. of the association constants, K , by a factor of a t ( 2 ) C. C. Thompson, Jr., a n d P. A. D . de Maine, J . Am. Chem. Soc., 8 6 , 3090 (1963). least five. Also, the solvent has been shown to change (3) R. S . Mulliken, ibid,, 7 1 , 600 (1950). the absorptivity of the co-nplex. (4) R . S . Mulliken, ibid., 74, 811 (1952). (5) R . S. Mulliken, J. Phys. Chem., 66, 801 (1952). (6) M . J. S . Dewar and A. R. Lepley, J. Am. Ckem. SOC.,83, 4560 (1961). ( 7 ) A. Bier. Rec. frau. ckim., 76, 866 (1956).
(8) R. Foster, J . C h r m SOC.,1075 (1960). (9) R . Foster and T . J . Thomson, Trans. F a l a d a y SOC.,59, 296 (1903). (10) B. Dale, K . Foster. and I). I.. Hammick, J . C h ~ mSo< , :398H (1954)
N . E . JURINSKI
3218
ASD
The formation of a 1: 1 complex has long been postulated to explain certain spectral data with varying degrees of success. Benesi and Hildebrand" first suggested a method of calculating the association constant, K ,and the molar absorptivity, U C , of a complex
+
K
for the system: X D e C . The Scott12aand the Ketelaar 12h equations are two more recent treatments. Recently de Maine and Seawright13 have developed an iterative method of analysis of spectroscopic data which was employed in this study and which is discussed more fully below. The validity of a spectral analysis performed in any of these procedures has often been questioned. Variations in the value of K have shown (at times) a dependence upon wave length contrary to what is predicted. Scott12a has attributed variations to the neglect of activity coefficients. Ketelaar, et have tried to correct for absorption due to an undercutting band of the free acceptor to obtain constant results. Corkill, et ~ l . , 'have ~ used a competitive system of donors to study the decrease in intensity of a well known band of one complex by the competing equilibrium for the common acceptor by the second donor. de Maine and Seawright's methodI3 has adequately allowed for all corrections of this nature and thus any variations noticed cannot be attributed to absorption by some of the uncomplcxed species. =1 further consideration is the possibility that more than one type of complex is formed, thus invalidating the assumption of 1: 1 interaction producing a single product. Higher order interactions forming 2 : 1 and other complexes have been considered likely due to the excess of one reagent normally required as an experimental limitation imposed by a mathematical approximation in most analyses. Using carefully selected systems Landauer and McConnell, l5 Foster, et al.,l f i and Hayman" have studied this possibility and have found, respectively, evidence for a 2 : 1 complex, no evidence for a 2 : 1 complex, and support for a nonspectroscopic method indicating higher order complexes. Jurinskilg has shown t h a t an exact solution for the formation constants ( K 1and Kz) and the molar absorptivities ( U C , and a& for 1: 1 and 2 : 1 simultaneous complexes may be determined assuming Beer's law is valid for all species. It is seen in this treatment that the association "constant" determined assuming only 1: 1 complex formation is not a true constant and thus could be expected to vary with wave length. Orgel and M ~ l l i k e nhave ~ ~ postulated t h a t isomeric and contact charge-transfer complexes may play an important role but that their presence should lead to a wave length independent value of K . Several workers have reported variations in the calculated (11) H A Renesi and J . H. Hildebrand, J . A m Chem. Soc., 71, 2703 (1949). ( 1 2 ) (a) R I,, Scott, Rer. i ~ ~chim., v . 75, 787 (1956). (h) J A . A. Ketelaar, C. van d e Stolpe, A. Goudsmit, a n d W. Dzcubas, ibid., 71, 1104 ( 1 9 5 2 ) . (13) P A . I). de Maine a n d R . 1). Seawright. "Digital Computer Programs for Physical Chemistry," Vol 1, >lacmillan C o . , New York, N. Y 1963. (11) J . > f , Corkill, R. Foster, and I). I* Hammick, J . Chem. Soc., 1202 (1955) (1.5) J. I.andauer and H hlcConnel1, J . A n i . Chcm. Soc., 74, 1221 (1962). (IC,) R . Foster. I ) . 1, Hammick, and P J Placito. J . Chem. S o c . , 3881 f 19 56) (17) H. J . H a y m a n , J i'hprn. P h y s . . 37, 2290 (19R2). (18) N B. Jurinski, J . Miss. Acad. S c i . . 10, 74 (1964). (19) 1,. E. Orgel and I
