6138
NORMAN J. ROSEAND RUSSELLS. DRACO
vicinity of the ring. This model has the merit that increasing electronegativity or vacant d orbitals leads to increased transfer of n-cloud in the direction of the metal ion. This postulated model can account for the stabilities found. Cowdrey and DaviesZ3 have stated that the diazotized anisidines are more stable than benzenediazonium salts which are more stable than diazotized o-toluidine. Nesmeyanov’s3 data indicate that those compounds having a group on the ring which allows delocalization of an electron pair are generally the most thermally stable. A metaldiazonium ion interaction as postulated above would assist in the flow of electrons from the amino nitrogen and increase stability. However, as such an interaction increases in strength (increasing electronegativity of the metal atom) flow of electrons from the vicinity of the C-Na+ bond should also increase with the positive charge on the diazonium group resisting this. The result is decreased electron density in the C-N2+ bond and its becoming weakened. The delocalization of the amino group electron pair and the decrease of electron density in the above bond have opposite effects. n’hen the pull of electrons away from the C-N2 + bond becomes large the net effect is destabilization. Considering the effect of water, only the iron compound is less stable than the p-dimethylaminobenzenediazonium chloride. In this study only the iron compound has incomplete outer orbitals (3d). These are well known for their coordination of electrons. (23)
nT.A.
Cowdrey and D. S. Davics, Qiiarl. Rea., 6 , 358 (1952).
[CONTRIBUTION FROM
THE
VOl.
s1
The variation of the frequency for the out-ofplane ring hydrogen wag suggests an interaction as proposed. An explanation of Table 111 is offered by the assumption that the metal ions position themselves over the center of the ring and in the case of the two most stable compounds the metal ions are closer to one side of the ring than the other. TABLE
111
. 4 B S O R P l I O N I k E Q U E N C I E S FOR R I X G
Metal of t h e compd.
Kone Fe Bi Zn(2H20) Zn(H20) Sn 1% g Zn
cCl
Cm.
HYDR0t;ES \VA(;GIN(: Trans.
-1
818 817 820 823
11 G 10.0 25.0 26.0 18 8 37.0 11.5 38.5 45.0 42.0 57.5
824
826
817 82 1 833 818 829
In spite of the failure to “stabilize” with some of the metal chlorides used, it seems that the present data should be extended to include more metal ions. I t is also suggested that greater insight into the interaction of metal ion with diazonium ion may be available through nuclear magnetic rcsonance spectroscopy. ~ - E WORLEASS,LA.
b’.A. NOYESLABORATORY OF CHEMISTRY, U N I V E R S I ~OFY ILLITOIS]
Molecular Addition Compounds of Iodine. I. An Absolute Method for the Spectroscopic Determination of Equilibrium Constants BY NORMAN J. ROSEAND RUSSELL S. DRAGO RECEIVED APRIL 1, 1959
A general equation is derived for evaluating acid-base equilibria from spectroscopic data. The equation is a general one and can be applied t o any acid-base system t h a t absorbs in some region of the spectra. Application t o base-iodine systems is described. The treatment requires fewer assumptions than are involved in the Benesi-Hildebrand treatment or several reported modifications of it. In this article the equation is derived, its application illustratcd and the interpretation of reported data evaluated. T h e advantages of employing our treatment are outlined.
Introduction The interaction between Lewis bases and iodine provides an ideal system for investigating the basicity of various electron pair donors. The addition compounds formed with iodine are soluble in nonpolar solvents so their heats of formation can be determined in the absence of crystal lattice effects or extensive solvation by studying the equilibrium constant of the reaction [B Iz S BI2] as a function of temperature. These data usually are obtained from the ultraviolet spectra of the solutions. The equilibrium constant often is calculated by the method proposed by Benesi and Hildebrandl or some tritsdificatiori of it. 2 - 4
+
(1) H. A. nenesi and J. €1. IIildellrarid, ‘I’IILS J o r r a ~ a i 71, , 2703 (1949). (2) J. A. A . K e t e l a a r , et ai., Rec. trazr. cizinz., 71, 1104 (1952). (:I) 5. X a g a k u r a , T H I S J O t J R N A l > , 7 6 , 3070 (1954); 8 0 , 520 (19Z8). (4) It, .\I. Keefer a n d 1.. J. Andrews, i b i d , 7 4 , 18‘31 ( l $ l 5 ? ) .
In many of these studies, conclusions and explanations for observed trends are based on small differences in the heats of formation. The slight curvature obtained in the plot of In I< vs. 1/T was used5 as evidence to support existence of several geometrical isomers. Scott6 has been concerned with the effect of solvent on these equilibria arid has obtained slightly dEerent values for the equiTz Ft CsH0.12 librium constant for the reaction C& in carbon tetrachloride and in pure benzene as the solvents. DeiMaine,’ on the other hand, reports n o solvation effects for the equilibrium betwTeen ethanol and iodine. I n view of the importance of conclusions 1)ased o t i sinall diirereiices in the values for the c o ~ ~ s t u i t
+
(5) L.E. Orgel and R. S . Li~illiken,ibid.,79. 4839 (19.j7) (6) T. A I . Cromwell a n d R . I.. S c o t t , i b i d . . 72, 3825 (19.jU); Scott, X P C l y n x ( I ~ I J ’ 75, Z . , 787 (l95fi). ( 7 ) 1’. A. A I k R I a i n e , J , ( h c m . P k y s . , 2 6 , 1 l ! l l (.I!Ii7)
I! I! I 1t j ,O(I( t
Preliminary calculations indicated that A ,'tc expressed in mole fraction units could be neqlected from the general equation 12. Thus CI, C, and tc can be based on moles/l. A curve il€1
Vol.
s1
lustrating the dependence of K on ec under the above set of experimental conditions thus is obtained. Similar calculations for other sets of experimental data on this system yield a series of curves the coordinates of whose intersections represent solutions for K and ec (see Fig. 1). The majority of the intersections occur in a small area indicating good experimental precision. The formation of a 1:1 complex is given strong support. Equilibrium constants and the molar absorptivity for the complex were determined for all intersections, and the average deviation was calculated. A11 values that lay outside the average deviation by a factor of 2.5 were eliminated and a new average determined. The significant data are reported in Table I1 and values for the equilibriumg constant ( K , = 1.78 + 0.05) and absorptivity of the coinplex ( t c = 13,400 f 6001. mole-' cin-') were obtained. These results are in good agreement with the values reported by Benesi and Hildebrand K, = 1.73 and cC = 15,4001. mole-1 cm.-'. TABLE I1 Sets of CD 0.0433-0.213 .08G2- , 2 1 3 ,0433- ,619 ,0862- ,019 , 2 1 3 - ,610 .043:3- , 8 1 2 .08B2- , 8 1 2 , 0 2 4 - .G19
K , -1 0.520 ,572 ,571 ,GO2
,m ,590 n20 .>!I I
e.2
14,000 15,000 15,275 15,676 15,890 1;7,7x 1A,O,5O
13..jOO
Sets of C D 0.924-0.0433 ,924- .08Ii2 ,924- ,213 1 . 0 0 - ,924 1 . 0 0 - ,0433 1 . 0 0 - ,019 1 . 0 0 - .08ii% 1 00 - . ? I 3
K , -1 0.574 .600 .(;lo ,500 ,872 ,572 595 .iiO5
6,
15,350 15.800 15,700 14,575 15,275 16,275 15,500 15,580
Scott6 has recalculated the above datal using a modified extrapolation procedure. The values K , = 1.9 and ec = 14,000 were obtained. Using a value of E , = 14,000,and Benesi and Hildebrand's data in pure benzene, he recalculated K , obtaining a value of 2 3. The difference between the value of 2.3 and 1.9 was attributed to solvent effects. Within experimental error examination of the data in Table I1 fails to show any concentration dependence to K,. When our value of eo = 15,400 is used to solve the same data in pure benzene the values 1.74, 171 and 1.71 are obtained for K , in good agreement with the values in carbon tetrachloride. The value for K , in pure benzene is much more dependent on the accuracy of eC than those in dilute solution and accounts for the differences obtained by Scott. This indicates that any solvent effect by benzene in the concentration ranges of mole fraction 0 02 to 1 lies inside the experimental error and cannot be detected. There is a very definite possibility that a solvent effect may be observed when results obtained when the benzene is not in large excess of the iodine are compdred with those obtained at large benzeneiodine ratios. This treatment indicates that for dilute iodine solution, 0.02 i4.f solutions of benzene are nearly as effective as pure benzene in causing solvation The above system is an exatnple of conditions under which the general equation reduce to the B-H equation. The results obtained are in good agreement when the B-H curve is constructed using a least squares treatment. In the second (q) J
