C. C. THOMPSON, JR.,AND P. A. D. DE MAINE
2766
Solvent Effects on Charge-Transfer Complexes. 11. Complexes of 1,3,5=Trinitrobenzenewith Benzene, Mesitylene, Durene, Pentamethylbenzene, or Hexamethylbenzene’
by C. C. Thompson, Jr., Department of Soils and Crops, Rutgers, The State University, New Brunswiek, New Jersey
and P. A. D. de Maine Department of Chemistry, University of California, Santa Barbara, California
(Received March 23, 1966)
Formation constants ( K ) , heats of formation (AH), and absorptivities (e) of 1:1 chargetransfer complexes of 1,3,5-trinitrobenzene with hexamethylbenzene, pentamethylbenzene, durene, mesitylene, or benzene dissolved in CCh, n-hexane, n-heptane, cyclohexane, or CHC13 were calculated from spectroscopic data collected at 20 and 45” for a dozen wave lengths between 2800 and 4200 A. K values for all complexes vary with solvent in the order: cyclohexane > n-heptane N n-hexane > CC14 > CHC13, with a 10 to 20-fold variation in K in passing from cyclohexane to CHCls at 20”. K is independent of wave length near the band maximum, but in some systems K increases at longer wave lengths, probably because of the simultaneous formation of 1:1 and higher order complexes.
Introduction Although the effects of the solvent on formation constant, absorptivities, and heat of formation of charge-transfer complexes2have been noted for several systems, few quantitative studies of these effects have reference has been made been made. to recent literature pertaining to charge-transfer, contact charge-transfer, simultaneous complex formation, and solvent-shift theories. Literature concerned with complexes between aromatic nitro compounds and various donors has also been n0ted.3”)~ As a natural consequence of the valence bond approach, Mulliken2 has emphasized the importance of ionic contributions in charge-transfer complexes. Thus for a given complex the formation constant ( K ) , absorptivities (ac), and heat of formation (AH) should all increase if the electrical properties (dipole moment, dielectric constant, etc.) of the solvent are increased. In the Dewar-Lepleys molecular orbital theory of charge-transfer, back coordination can nearly equal the charge transfer, and thus K , ac, and AH can be virtually unaltered by a change in solvent. The Journal of Phgsical Chemistry
In this work, complexes of 1,3,5-trinitrobenzene with each of benzene, mesitylene, durene, pentamethylbenzene, and hexamethylbenzene are studied with carbon tetrachloride, chloroform, n-hexane, n-heptane, and cyclohexane as solvents. Precise analytical treatment of the data was achieved by use of the self-judgment method of curve fitting. ~
~~~
(1) Taken from the Ph.D. Thesis of C. C. Thompson, Jr., University of Mieaissippi, 1964. (2) R. S. Mulliken, J. Am. Chem. Soe., 74, 8 1 1 (1952); J . Phys. Chem., 56, 801 (1952). (3) (a) C. C. Thompson, Jr., Ph.D. Thesis, University of Mississippi, 1964; (b) C. C. Thompson, Jr., and P. A. D. de Maine, J . Am. Chem. SOC.,85, 3096 (1963). (4) N. B. Jurinski and P. A. D. de Maine, 8id., 86, 3217 (1964). (5) N. B. Jurinski, Ph.D. Thesis, University of Mississippi, 1963. (6) V. Ramakrishnan and P. A. D. de Maine, J . Miss.Acad. Sci., 10, 82 (1964). (7) N. B. Jurinski, C. C. Thompson, Jr., and P. A. D. de Maine, J. Inorg. Nucl. Chem., in press. (8) M. J. S. Dewar and A. R. Lepley, J . Am. Chem. SOC.,83, 4560 (1961). (9) P. A. D. de Maine and R. D. Seawright, “Digital Computer Programs for Physical Chemistry,” The Macmillan Co., New York, N. Y.: (a) Vol. I, 1963; (b) Vol. 11, 1965.
COMPLEXES O F
2767
1,3,5-TRINlTROBENZE,NE
Table I : Details of the Systems Studied a t 20 and 45"
Systema
Wave lengthb range, A.
TNB-CClr TNB-n-C?Hia TNB-n-CeHir TNB-c-CeHie TNB-CHCla HMB-CClr HMB-n-C?Hia HMB-n-CsH14 HMB-c-CsH11 HMB-CHCls PMB-CClr PMB-n-C?His PMB-n-CaHlr PMB-c-CsHiz PMB-CHCls DurCCL Dur-n-C?H~s Dur-n-CeHl' Dur-c-CsH1~ Dur-CHCls Mea-CClr Mes-n-C~Hls Mes-n-CaHir Mes-c-CaHlz Mes-CHC18 Bz-CClr Bz-c-CsHiz HMB-TNB-CClr
2700-4300 2650-4300 2650-4300 2650-4300 2750-4200 3600-4200 3600-4200 3600-4200 3600-4200 3600-4200 3300-4000 3300-4000 3300-4000 3300-4000 3400-4000 3200-3800 3200-3800 3200-3800 3200-3800 3250-3800 3100-3700 3100-3700 3100-3700 3 100-3700 3100-3700 2750-3300 2750-3300 3650-4200
HMB-TNB-dhHls
3650-4200
HMB-TNB-n-CeHlr
3650-4200
HMB-TNB-c-CsHn
3650-4200
HMB-TNB-CHCla
3650-4200
No. of samples
C ~ n c n range, .~ M 3.80 X 10-L1.90 X 1.63 X 10-L8.17 X 1.71 X 10-C8.55 X 1.75 X 1 0 - L 8 . 7 5 X 4.51 X 10-L2.26 x 0.138-0.276 0.171-0.342 0.185-0.371 0.152-0.305 0.131-0.21Q 3.96 x 10-Lo.198 4.22 X 10-Lo.211 4.09 X 10-LO.204 3.95 X 10-2-0.188 4.53 x 10-e0.227 5.56 X 10-LO.278 5.45 X 10-eO.273 5.43 X 10-e0.271 5 . 5 1 X 10-Lo.275 6.30 X 10--1-0.315 0.144-0.719 0,144-0,719 0.144-0.71Q 0.144-0.719 0.143-0.717 0.281-2.81 0.281-2.81 5.75 x 10-4 1.62 x 10-1-0.121 5.58 x 10-4 1.53 X 10-Lo.115 5.69 x 10-4 1.53 X 10-e0.115 5.72 x 10-4 7.42 X 1 0 - e O . l l l 1.61 X 10-1 9.70 x 10-Lo.146
10-1 10-4 10-4 10-4 10-1
System"
Wave lengthb range, A.
PMB-TNB-CCL
3400-4000
PMB-TNB-wC~HM
3400-4000
PMB-TNB-n-CsHlr
3400-4000
PMB-TNB-c-CsHia
3400-4000
PMB-TNB-CHClr
3400-4000
Q Q
DurTNB-CCL
3250-3800
9
D U P T N B - ~ I - C ~ H I ~ 3250-3800
9
Q 9
Q 9 3d 3d 3d 34 3d
Q DurTNBlt-CaHic
3250-3800
Dur-TNB-c-CaHlz
3250-3800
Dux-TNB-CHCls
3250-3800
9 9
Mes-TNB-CC14
3100-3700
Q
Mas-TNB-n-C7Hls
3100-3700
Mes-TNB-n-CsHic
3100-3700
Mes-TNB-c-CsHu
3100-3700
Mes-TNB-CHCls
3100-3700
Bz-TNB-CC4
2750-3300
Bz-TNB-c-CsH1z
2750-3300
9
Q 9
Q 9
Q
9 9 10 10 18 18 18 18
C ~ n o n range, .~ M 5.55 x 10-4 1.53 X 10-Lo.230 5.68 x 10-4 1.52 X 1 0 - W . 2 2 8 5.67 x 10-4 1.44 X 10-Lo.217 5.81 X 10-4 1.47 X 10-Lo.220 1.08 X 10-8 2.17 X 10-Lo.325 5.64 X 10-4 1.61 X 10-W.242 5.81 x 10-4 1.68 X 10-Lo.251 5.82 X 10-4 1.75 X 10-Lo.262 5.67 x 10-4 1.71 X 1 0 - W . 2 5 6 1.17 X 10-1 2.13 X 10-eO.320 5.66 x 10-4 3.59 X 10-"0.539 5.94 x 10-4 3.60 X 10-e0.540 5.61 x 10-4 3.60 X 10-e0.539 5.65 x 10-4 3.60 X 10-e0.539 1.02 x 10-1 7.19 X 10-L1.08 8.35 x 10-4 0.113-1.13 5.71 x 10-4 0.113-1.13
No. of samples 18 18 18 18 18 18 18 18 18 18 18 18
18 18 18 15 15
18
a Symbols used are: TNB, 1,3,5-trinitrobenzene; HMB, hexamethylbenzene; PMB, pentamethylbenzene; Dur, durene; For ternary Mes, mesitylene; Bz, benzene. Measurements were made at 50-A.intervals over the indicated wave length range. Measured absorbance below 0.005 absorbance unit. systems the first concentration refers to TNB and the second to the donor.
'
Experimental All solid materials were Eastman Kodak reagent grade chemicals. 1,3,5-Trinitrobenzene (m.p. 121122') was recrystallized from Fisher Spectrograde methanol. Durene (m.p. 79-80') and pentamethylbenzene (52-53.5') were recrystallized from absolute ethanol. Hexamethylbenzene (m.p. 166-167') was recrystallized from CC1,. Fisher Spectrograde benzene (b.p. 79.5-80.0') and Eastman Kodak reagent mesitylene (b.p. 163.0163.5') were distilled through a Vigreux column in a system protected from atmospheric moisture. The distilled liquids were then purged with oxygen-free dry nitrogen (dew point less than -40'). Fisher Spect'rograde CCl,, n-hexane, n-heptane, and cyclohexane were purged with oxygen-free dry nitrogen immediately before use. Fisher Spectrograde CHCla was purified as described p r e v i o ~ s l y ~ and ~ ~then ' ~ was
purged with nitrogen immediately before use. Solvent purity was checked with an Aerograph Hy-Fi Model 600 gas chromatograph. The techniques of solution preparation and of measurement of the absorption spectra at 20 and 45' for selected wave lengths between 2650 and 4200 A. have been described elsewhere.ab The calibrated Beckman DU spectrophotometer^^^ were equipped with photomultiplier tubes, temperature control accessories and four matched 1-cm. stoppered quartz cells. Pure solvent served as the reference. To within the limits of reproducibility, all systems except pentamethylbenzene-trinitrobenzene-chloroform were stable for at least 24 hr. Experimental details for the systems studied are given in Table I. For each two-solute system, duplicates were pre(10) P. A. D. de Maine, J. Chem. Phys., 26, 1036 (1957).
Volume 69,Number 8 August 1966
C. C. THOMPSON, JR.,AND P. A. D.
2768
pared from separate portions of the components for six of the eighteen solutions. On the basis of reproducibility of data and calibration results for the two spectrophotometers, an experimental error of less than 1%is claimed for the data here reported.
Data Processing Method Data for the trinitrobenzene (TNB)-inert solvent and TNB-donor-inert solvent systems were processed by the self-judgment m e t h ~ as d ~already ~ ~ ~ described.3b In the spectral region considered, donor-solvent solutions show low absorptivities, and values a t each temperature and wave length were calculated without use of the self-judgment principle. The data for the TNB-solvent systems were compatible with the equation
+ ai [AI
(1)
UA = O ~ A
where U A is the absorptivity of the solute, A, whose molar concentration is denoted by [A] and where O ~ and A al are constants for each temperature and wave length. K
For the reversible reaction, A equations can be derived CC =
[KCA
+ B e C, the following
-t KCB + 1 -
~ ( K ' ( C A-- CB)' Ac = @Cc = A0
CACB/AC =
+ 2KC.4 + ~ K C Bf 1)]/2K
-
- Cc) - UB(CB- Cc) 1/K@ + (CA + CB - Cc)/a~
(2)
(3)
(4)
In these equations, CA and CB are initial concentrations of A and B, respectively; CC is the equilibrium concentration of C; K is the formation constant; uA, uB, and ac are absorptivities of A, B, and C, respectively; and Ac and A. are the absorbance due to the complex C and the measured absorbance, respectively. In the derivation of eq. 2, 3, and 4, it is assumed only that the law of mass action is obeyed. The assumptions common to earlier methods (CA >>> CB; A and B obey Beer's law) are not made. In our calculations we have assumed that the complex, C, obeys Beer's law. Equations 2, 3, and 4 were solved for K and ac by an iterative method3b with data a t each wave length and temperature. limits of experimental error ( i e . , DEVFl = DEVF2 = 0.01) were used to compute an error zone for eq. 4 and data points outside this zone were discarded. From the error zone and the points within it was calculated the maximum permitt,ed error (MPE) for the computed values of K and ac. The Journal of Ph,ysicaE chemistry
DE
MAWE
In this paper the original definition of MPE3bs49a has been retained. The more recent definitiongb makes the MPE values appreciably larger.
Results The absorbance data for TNB-solvent systems are compatible with eq. 1 to within il%. Details of the computer analysis for the TNB-solvent and donorsolvent systems have been given elsewhere. Attempts to study benzene-carbon tetrachloride mixtures were not entirely successful. For mixtures containing more than 50% by volume of benzene, oxygen absorbed during even careful handling produced erratic results.13 With less than 20% benzene band enhancement due to absorbed oxygen was not observed, even in samples prepared 24 hr. earlier. Thus only data for solutions with less than 20% benzene were used in our calculations. The absorption for pentamethylbenzene-TNBCHC13 solutions gradually increases with time. This phenomenonois especially pronounced a t wave lengths below 3750 A. but it is not found for the single solute systems. Thus only data from the 3750 to 4000-A. region are used in calculating values reported in this paper. New absorption bands with a single maximum in the 2800 to 3950-A. region were observed in all TNBdonor-solvent systems. At each temperature and wave length the new data are compatible to within 1%with the 1:1reaction 3m112
TNB
+ donor
K
complex
Mean formation constants ( K ) are given in Table 11. Near the band maxima the values for K are independent of wave length for all donor-acceptor pairs studied. For most systems, K increases slightly a t longer wave lengths, especially in paraffin solvents. Detailed numerical results for the ternary systems studied are given elsewhere. l4 Mean formation constants for the wave length invariant regions are given in Table 111. I n Table IV are given values for the mean heats of formation computed with the van't Hoff equation from values in Table 111. Plots of ac, absorptivity of the complex, vs. wave length for each temperature and each system yielded a 3&1
(11) P.A. D.de Maine and R. D. Seawright, Ind. Eng. Chem., 5 5 , 29 (1963). (12) C. C. Thompson, Jr., and P. A. D. de Maine, J . Miss. Acad. Sci., 10, 123 (1964). (13) D. F. Evans, J. Chem. Soc., 345 (1953); 1351, 3885 (1957); 2753 (1959). (14)C. C. Thompson, Jr., and P. A. D. de Maine, J . Miss. Acad. Sci., IO, 137 (1964).
2769
COMPLEXES OF 1,3,!%TRINITROBENZENE
Table 11: Average Formation Constanta (in 1. mole-') for All Wave Lengths Studied Donor"
Temp., OC.
CClr
n-CvH~a
n-CsHu
c-CaHn
CHCls
HMB
20 45
4.86 f 0.Mb 2.76 f 0.10
14.69 f 0.20 8.70 f 0.12
15.41 f 0.39 9.41 f 0.10
17.50 f 0.28 9.78 f 0.16
0.90 f 0.15 1.01 f 0.14
PMB
20 45
3.09 f 0.07 2.28 f 0.12
8.69 f 0.12 5.91 f 0.08
8.91 f 0.16 5.98 f 0.14
10.45 f 0.10 6.48 f 0.10
1.02 f 0.04 0.85 f 0.03
Dur
20 45
2.29 f 0.07 1.57 f 0.13
5.77 f 0.19 3.91 f 0.09
5.30 f 0.10 4.10 f 0.11
6.02 f 0.14 4.18 f 0.15
0.97 f 0.26 0.70 f 0.25
MW
20 45
1.36 =t0.07 1.00 f 0.05
3.02 f 0.12 2.18 f 0.05
2.76 f 0.19 2.26 f 0.06
3.51 f 0.06 2.60 f 0.06
0.17 f 0.04 0.12 f 0.03
Bz
20 45
0.56 f 0.06 0.45 f 0.09
" Symbols for the donors are given in Table I.
0.88 f 0.03 0.77 f 0.07
' Square-root-mean-square deviations.
Table 111: Mean Formation Constanta (in 1. mole-l) for Wave Length Invariant Regions Donor"
Temp., OC.
CClr
n-C~Hls
n-CsHir
HMB
20
4.87 f 0.08b (9/W 2.77 f 0.08 (8/12)
14.82 f 0.17 (7/W 8.92 f 0.12 (6/12)
15.34 i 0.18 (6/12) 9.19 f 0.13 (8/W
17.50 f 0.20 (8/W 9.70 f 0.12 (6/12)
0.92 f 0.10 (5/12) 0.87 f 0.09 (6/12)
3.08 f 0.05 (7/13) 2.22 f 0.14
8.82 f 0.11 (5/13) 5.84 f 0.09 (7/13)
8.82 f 0.12 (6/13) 5.90 f 0.09 (7/13)
10.39 f 0.13 (6/13) 6.43 f 0.09 (8/13)
0.98 f 0.07 (5/7) 0.85 f 0.08 (7/7)
2.24 f 0.09 (8/W 1.66 f 0.11 (6/12)
5.76 f 0.10 (6/12) 3.86 f 0.07 (7/12)
5.33 f 0.09 (6/12) 3.96 f 0.07 (6/12)
5.97 f 0.09 (7/12) 4.13 f 0.08 (5/12)
0.84 f 0.11 (5/11) 0.54 f 0.11 (5/11)
1.34 f 0.06 (8/13) 0.92 f 0.04 (5/13)
3.06 f 0.05 (4/13) 2.07 f 0.04
2.81 f 0.05 (5/13) 2.21 f 0.04 (6/13)
3.32 f 0.06 (5/13) 2.44 f 0.05 (7/13)
0.13 f 0.03 (611' 3) 0.08 f 0.03 (5/13)
45
PMB
20 45
Dur
20 45
Mea
20 45
Bz
20 45
0.47 & 0.02 (5/12) 0.36 f 0.03 (5/12)
" Symbols for donors given in Table I. total number of wave lengths studied.
Discussion For each ternary system, the data near the chargetransfer band maximum are consistent with the existence of a single 1 :1 complex. I n part Iabit was shown that the variation in the value of the formation con-
CHCls
0.81 f 0.02 (5/12) 0.66 f 0.05 (4/12)
'Mean maximum permitted errors.
single broad band with a poorly defined maximum. In Table V are given the maximum value of a~ and the corresponding wave length for each system.
C-CSHIZ
Number of wave lengths used in calculating K and
stant, K , with wave length for a fixed temperature cannot be attributed to neglect of unsuspected 1:l interactions between donor and acceptor molecules themselves or with the solvent. Jurinski16 and Hayman16 have shown mathematically that the simultaneous formation of higher order complexes and (15) N. B.Jurinski, J. Miss. Aead. Sci., 10,74 (1964). (16) H.J. G.Hayman, J. C h .Phys., 37, 2290 (1962).
vohnt?69,Number 8 August 1966
C. C. THOMPSON, JR.,AND P. A. D.
2770
DE
MAINE
Table IV : Heats of Formation“ (kcal. mole) CClr
Dono@
HMB PMB Dur Mes Bz
-4.18 -2.44 -2.24 -2.81 -2.03
n-C~Hls
f 0.18‘ f 0.61 f 0.80 f 0.69 f 0.92
-3.77 -3.05 -2.96 -2.90
n-CsHic
f 0.18 f 0.20 f 0.27 f 0.25
-3.80 -2.97 -2.20 -1.78
CHCh
C-CEHIZ
f 0.19 f 0.21 f 0.26 f 0.27
a Calculated from mean formation constants a t 20 and 45” listed in Table 111. permitted errors.
-4.37 -3.55 -2.72 -2.30 -1.47
f 0.18 f 0.20 f 0.25 f 0.27 f 0.70
-0.44 -1.06 -3.26 -3.14
f 1.63 f 1.20 f 1.95 f 4.65
Symbols for donors given in Table I.
Maximum
Table V : Maximum Molar Absorptivities and Wave Lengths of Band Maxima
-Temp., 7 C C l r -
A.
-n-C7Hl-
A.
-n-CsHuac
x
~ A. ~
-c-CeHla--
, ac
hn.*,
A.
CHCla-
7
Xmm aC
A.
2374 f 206 2089 f 180
3890 3880
DonorG
OC.
aC
HMB
20 45
2565 f 21’ 2565 f 57
3920 3920
2270 f 1 2175 i 8
3860 3860
2229 f 8 2190 f 8
3880 3860
2391 f 38 2197 f 24
3870 3870
PMB
20 45
2394 f 16 2260 f 92
3700 3690
2140 f 1 2041 f 5
3620 3610
2169 f 1 2079 f 6
3630 3620
2180 f 1 2061 f 4
3650 3640
Dur
20 45
1966 f 39 1886 f 67
3430 3430
1883 f 3 1860 f 6
3350 3360
1940 f 3 1831 f 4
3350 3350
1977 f 1 1836 f 5
3380 3380
1537 f 110 1899 f 219
3390 3390
Mes
20 45
2280 f 49 2398 f 60
3350 3350
2220 f 3 2229 f 7
3300 3300
2296 f 4 2188 f 8
3300 3300
2205 f 2 2082 f 7
3300 3300
6036 f 955 7978 f 1795
3350 3350
Bz
20 45
3893 f 80 4133 i 152
2830 2800
3867 f 6 3775 f 94
2800 2770
A-?
UC
,,X,
. . .d
. . .d
...
...
’
a Symbols ofor donors are given in Table I. Due to the broad abaorption banda, the values given here are considered accurate t o within f 2 0 A. ‘ Maximum permitted errors. Time effect at short wave lengths.
1 : l species can account for the variation in value of K with wave length. Although exact computer methodsebhave been devised for resolution of data for a system with both 1:l and 2 : l complexes, limited computer storage capacity prevented the treatment of the present data in this way. For each donor-acceptor-solvent system, a change of the “inert” solvent effects drastic changes in the computed K values (Tables I1 and 111) which, except for CHCI3, are not reflected in the values for the heat of formation (Table IV). The absorptivity of the assumed 1: 1 complex near the absorption maxima are approximately constant (Table V) . Competitive donor-solvent complexes in CHCls, and possibly CCL, can block either the LN or T N donor orbital and thus reduce the apparent formation constant (KQ) in these solvents.3b Certainly, the variation in KQ with solvent cannot be attributed solely to the assumption of ideality, to the neglect of higher complexes, or to changes in the average properties of the solvent environment. 3b The Journal of Physical Chemistry
Variations in KQ for hydrocarbon solvents can be explained qualitatively if it is supposed that the “short-range lattice structures” of the solvent, the molecular dimensions and the orbital symmetry of the interacting donor-acceptor pair, and the magnitude of the forces between the solute molecules themselves or with the solvent each play an important role in complex formation.’’ Constants for the naphthalene-trinitrobenzene complex a t 20’ were foundab to be 5.16 f 0.07, 9.15 f 0.17, and 1.82 f 0.08 in cc14, cyclohexane, and CHC13, respectively. The KQ for the hexamethylbenzene complex (Table 111) is considerably greater than that for the naphthalene complex in cyclohexane (and other hydrocarbons), while for ccI4and CHCls the order is reversed. These results indicate that CHC13 and CCL are more effective in blocking complex sites of hexamethylbenzene. The importance of considering (17) P.A. D.de Maine, J . Chem. Phys., 26, 1042, 1192 (1957); J. MG8. ACd. SCi., 9, 160 (1963).
COMPLEXES O F 1,3,6-TRJNITROBENZEJYE
solvent effects in comparisons of relative donor strengths is thus demonstrated. For each solvent, except CHCl,, and a fked temperature the absorptivities of the complex (ac) (Table V) vary with donor thus: benzene > hexamethylbenzene > pentamethylbenzene > mesitylene > durene. The corresponding order for the formation constants (Table 111) and (except for cc14 systems) for AH, heat of formation, (Table IV) is hexamethylbenzene > pentamethylbenzene > durene > mesitylene > benzene. If the comparatively large uncertainties in AH are considered it is possible that this order also applies for C C 4 systems. Thus except for benzene, and either mesitylene or durene, the relative absorptivities are in the predicted order.2 Higher values for ac with mesitylene and benzene have been cited as evidence of contact charge transfer.18-20 Contact charge-transfer theories predict that ac should increase with temperature. I n only four of the twentytwo systems studied did % increase with temperature (Table V) and in the remainder ac was unchanged or decreased slightly. The presence of isomeric 1:l complexes can explain the temperature variations of ac if it is supposed that back coordination charge-transfer bands8 are formed. Arguments to support this suggestion have already been presented. 3b Plots of % vs. wave length (not shown) indicate that the durene complex has the broadest chargetransfer band with the lowest maximum absorptivity. Of the donors used in this study durene has the largest difference in energy of the el, orbitals.21 Thus, transitions involving these orbitals should occur at considerably different wave lengths to form a broad absorption band composed of the separate transitions. Several workers17~z~~22~2a have pointed out that integrated band intensities should be used in comparing the relative absorption for a series of complexes. For the systems reported here, the wave length of maximum absorption for a given complex varies with solvent thus: heptane S hexane < cyclohexane < CHCI, < cc14. Fosterz4and B h a t t a c h a r ~ ahave ~ ~ re-
2771
ported similar results for various complexes in CC14, CHCla, and cyclohexane or heptane. The results cannot be explained in terms of average solvent properties alone, but they can be interpreted by the supposition that the formation of one isomeric complex is favored by the “structure of the solvent.” Time effects in the pentamethylbenzene-trinitrobenzenechloroform system are similar to those reported previously for naphthalenetrinitrobenzeneab and iodinenitromethane2scomplexes. For the sep& rate components dissolved in CHCI, no changes in the measured absorbance with time are observed. This behavior can be attributed to reactions of the pentamethylbenzenechloroform complex to yield products which subsequently interact with the trinitrobenzene molecules. It seems that theories which predict more extensive complexing in polar solvents and which emphasize the importance of nonspecific solvent-solute interactions are incorrect in principle. In conclusion, it can be said that charge-transfer theories which do not consider the role of the solvent are inadequate.
Acknowledgments. Financial support for this work was provided by the United States Air Force under Grant No. AF-AFOSR 62-19, monitored by the Directorate of Chemical Sciences, Air Force Office of Scientific Research. C. C. T. wishes to thank the National Science Foundation for the award of Cooperative Graduate Fellowships. (18) L. E. Orgel and R. 8. Mulliken, J. Am. Chem. SOC.,79, 4839 (1957). (19) 8. P.McGlynn, Chem. Rev., 58, 1113 (1958). (20) J. N. Murrell, Quart. Rev. (London), 15, 191 (1961). (21) L. E.Orgel, J . C h a . Phys., 23, 1352 (1955). (22) 8.F. Mason, Quart. Rev. (London), 15, 287 (1961). (23) R. 8. Mulliken and W. B. Person, Ann. R H . Phys. Chem., 13, 107 (1962). (24)R. Foster, J. Chem. SOC.,1075 (1960). (25) R.Bhattacharya, J. C h a . Phys., 30, 1367 (1959). (26) P.A. D.de Maine and W. C. Ahlers, J. Chem. SOC.,211 (1960).
Volume 69,Number 8 August 1966
