Aug. 5 , 1956
3625
1,3,5-TRINITROBENZENEAND PYRIDINE-N-OXIDES I N CHLOROFORM [CONTRIBUTION FROM
THE
RESEARCH LABORATORIES O F THE SPRAGUE ELECTRIC COS]
Molecular Compounds. VII. Interactions between 1 ,j,S-Trinitrobenzene and Pyridine-N-Oxides in Chloroform Solution BY SIDNEY D. Ross, DONALD J. KELLEYAND MORTIMERM. LABES RECEIVED JANUARY 12, 1956 The color formation observed in chloroform solutions of 1,3,5-trinitrobenzene and pyridine-N-oxide or methyl substituted pyridine-N-oxides is discussed in terms of complex formation and physical perturbation of the trinitrobenzene spectrum.
Pyridine-N-oxide is a strong electron donor, and 1,3,5-trinitrobenzene is an electron acceptor. It might, therefore, be expected that these two molecules would form a charge-transfer complex in solution, and, in fact, coloration is observed in chloroform solutions of l93,5-trinitrobenzene with pyridine-N-oxide, 2-picoline-N-oxide, 4-picoline-N-o~ide and 2,6-lutidine-N-oxide. However, when the spectroscopic data were treated in the usual mannerl to determine equilibrium constants for complex formation, good linear plots were obtained, but these intersected the Y-axis slightly below the origin. The plots shown in Fig. 1 for solutions of pyridine-N-oxide and 1,3,5-trinitrobenzene in chloroform a t 24.8 f 0.1' are typical of those obtained with all of the amine oxides. Moreover, attempts to prepare a solid complex from these reactants were unsuccessful. This suggests the possibility that something other than 1:l complex formation may be responsible for the color formation observed in these solutions. This possibility is strengthened by the fact that chloroform solutions of pyridine-N-oxide and 1,3,5-trinitrobenzene do not show a new absorption maximum in the spectral region from 250450 mp, although from 320-480 mp, the measured total optical density is greater than the sum of the optical densities due to the two individual components. This absence of a characteristic chargetransfer maximum raises the question of whether a molecular complex is present in these solutions. It becomes pertinent, therefore, in this instance, to consider whether or not the observed intensification of the absorption is due to a purely physical perturbation. This question of differentiating between weak complex formation and physical perturbation arises in other cases where the molecular complex is not isolable as a definite chemical en tit^.^-^ It is our present purpose t o consider two types of physical perturbation and to consider their applicability to the present data.
Experimental Solvents and Reagents.-Baker and Adamson reagent grade chloroform containing 0.75% ethanol as stabilizer was used without further purification. All of the chloroform was of the same lot number. 1,3,5-Trinitrobenzene, Eastman Kodak White Label, was crystallized from chloroform as almost colorless crystals, m.p. 120.8-122°. The amine oxides were all obtained from the Reilly Tar (1) R. M. Keefer and L. J. Andrews, THISJOURNAL,74, 1891 (1952): S. D. Ross, M . Bassin, M . Finkelstein and A. Leach, ibid., 76, 69 (1954).
W.
(2) W. Haller, G . Jura and G. C. Pimentel, J. Chcm. Phys.. 22, 720 (1954). (3) D. E. Schiller and R . H. Schuler, THISJOURNAL,76, 3092 (1964).
(4) N. S. Bayliss and C. J. Brarkenridge, ibid., 77, 3950 (1955).
0
0.5
1 .o 1/(Ao
Fig. 1.-Plot
1.5
2.0
2.5
+ To). 1 upper line, 421 A Q+ To' lower line, 416 mp.
AoTo of A . + T O D O us.
-*
mp; middle line, 420 mp;
and Chemical Corp. and purified as follows. PyridineN-oxide was either sublimed a t 60' and 0.1 mm., and dried over sulfuric acid in an evacuated desiccator, or distilled in vacuo, not allowing the temperature to rise above 130°6; b.p. 82" a t 0.1 mm. The distillate solidified and was crystallized from a benzene-hexane mixture in a drybox and dried in an evacuated desiccator; m.p. 65-66' in a sealed capillary. 4-Picoline-N-oxide was crystallized from benzene-chloroform and vacuum dried; m.p. 181182". 2-Picoline-N-oxide was distilled; b.p. 89-90' a t 0.8-0.9 mm. The distillate crystallized on seeding; the solid was dissolved in boiling benzene; addition of hexane t o the solution in a dry-box gave small waxy crystals which were filtered and vacuum dried.6 2,6-Lutidine-N-oxide was distilled; b.p. 83-84' a t 1-1.5 mm. It was not obtained as a crystalline solid. The Absorption Spectra Measurements.-A Beckman model DU spectrophotometer was used throughout. Stoppered Corex absorption cells were used, and the cell housing was maintained a t constant temperature by means of two Beckman thermospacers, through which water from a constant temperature bath was circulated. Measurements from 410-480 mp were made a t a t least seven different sets of concentrations of the two molecular species. For the pyridine-N-oxide-l,3,5-trinitrobenzenemeasurements, the amine oxide concentrations were varied from 1.4-0.15 M , and the trinitrobenzene concentrations were varied from 0.010.20 M . The concentrations of 2-picoline-N-oxide were varied from 1.5-0.5 M a s the trinitrobenzene concentrations varied from 0.20-0.02 M . The 4-picoline-N-oxide concentrations varied from 0.8-0.2 M as the trinitrobenzene concentrations were varied from 0.02-0.10 M . I n the experiments with 2,6-lutidine-N-oxide, the amine oxide concentrations were varied from 1.3-0.6 M and the trinitrobenzene concentrations from 0.02-0.10 M .
Results and Discussion Chloroform solutions of 193,5-trinitrobenzene and pyridine-N-oxide or methyl substituted pyridine-N-oxides show absorption, in the region from (5) Org. Syntheses, 53, 79 (1953). ( 6 ) E. Matsumura, J. Chcm. SOC.J a p a n , P a r e C k e m . Sccf., 74, 3G3 (1953); C. A , , 48, 644B (1954).
3626
SIDNEY D. Ross, DONALD J. KELLEYAND MORTIMER hi. LABES
Vol. 7 8
TABLE I PERTURBATIOS O F THE S P E C T R V M O F 1 , 3 , 5 - T R I N I T R O B E S z E N E B Y P Y R I D I N E - 3 - O X I D E AND
Amine A7-oxide
OC
Ad, A [Amine oxide]o[Trinitrobenzeneja
.
0 24.8 30.4 40.3
Pyridine-
METHYLSUBSTITUTED PYRIDISE-N-OXIDES
Temp., 416 mp
420 mfi
7.95 i 0.30 9 , 1 6 + .54
5.57&0.19 6 . 5 3 i .35 (?.75=k ,23 7.30 i .18
11.35 =t . 3 % 9.86 i .27
424 m p
3 . 6 8 i 0.09 4.70i .28 4.82i .%4 5 . 1 2 i .14
430 nip
434 mp
.440 111p
444 mp
?-Picoline-
24.8
10.3 i 0 . 3
7.79=t0.17
4.91i 0 . 1 4
3.63 i0.08
4-Picoline-
24.8
35.7 i 0.7
2,B-Lutidine-
24.8
10.7 i 0 . 1
420 mp
424 mp
430 xnp
434 m p
2 7 . 6 =k 0 . 7
18.5 i 0 . 4
14.1i0 . 4
444 mfi
440 miL
8.33
* 0.11
4.10
I,,&,
2 32 i 0 OG
4 j O mp
454 mfi
4G0 m p
5 . 5 5 i0.03
4 . 2 1 i 0 05
2 . 7 3 i0 . 0 5
320-480 mp, in addition to that ascribable to the two solutes individually; Le., where dT is the total
nitrobenzene in chloroform a t 24.8 i 0.1" are shown in Fig. 2 . The slopes of these lines, Adc/ AAoTo,are pertinent to the subsequent discussion, d T > e.A'4o ETTo and in Table I we have presented some repremeasured optical density, A. is the initial amine data, To indicate the accuracT, of the oxide concentration and To is the initial 1,3,5- sentative results, the values are given as the slopes trinitrobenzene concentration. This additional ab- + the average deviations. In all measuresorption, J,, may be defined as ments were made from 410-480 mG, and in the case d , = t1.y - t L . 1 0 - til'" uf pyridine-N-oxide a t 24.8", the wave length range -1s can be Seen froin Fig. 1, when the spectra- covered was 3?0-380 xnp. The selected values in scopic data are plotted in the usual manner for 1 :I Table I are those where the differences between d, complex formation, the resultant straight lines and d~ were of conwiient magnitudes. cannot be used t o evaluate the equilibrium conThe plots shown in Fig. 2 are in accord with two stant, since they intersect the Y-axis below the possible modes of Perturbation. The first type origin. It is, of course, possible that the intersee- has been discussed in detail by Bayliss and Brackention points on the \'-axis for these plots are, Tvithin r.idge,'J who treat the perturbation as a solvent experimental error, indistinguishable fronl zero and effeCt resulting fro111 active solvent molecules in that this corresponds to a case of weak complexing the cage surrounding the perturbed solute. The with the extinction coefficient for the complex very exact nature of this perturbation is not specified, large. Nevertheless, in this particular case, since although it is suggested that it may depend "011 it was not possible t o isolate or prepare the corn- favorable mutual orientation between the solute plex, since a characteristic charge-transfer maxi- and active solvent 1lloledes." nlum could not be identified in the spectruln froi11 The second type of perturbation is related to that 250-450 1np and since, as will be showI1, the discussed by Bayliss and Brackenridge but is more ditional absorption, dc, increases mith increasing specific in that the effect is attributed entirely to temperature, it Seems appropriate to consider collisions between the solute and the perturbing whether some explanation other than 1 : l complex molecules. For fither type a Plot of d e 8s. AoTo formation may not account more satisfactorily for would be linear and would pass through the origin as shown in Fig. 2. However, if the temperature the observed de's. \\Then d, is plotted against &l'ua straight line is varied, i t becomes possible to distinguish bepassing through the origin is obtained, Some tween an effect which depends on mutual orientatypical data for pyridine-N-oxide and 1,3,3.tri. tions and an effect which depends on collisions. If we focus our attention on the slopes, Ad,/AAoTo, we would expect a decrease in the slopes as the 0.8 temperature is increased, if the effect is an orientational one, but an increase in the slopes if the effect is collisional. Since the slopes increase with in0.6 creasing temperature (Table I), a purely orienta tional perturbation of the type described by Bay%liss and Brackenridge is eliminated.8 0.4 -111 of the observations reported herein are fully in accord with the collisional perturbation hypothesis. In particular, plots of d, us. AaTo are lin0.2
+
average
0
0.02
u 04
(1
06
0 08
0.10
0.12
:I ,,I " 2 . -.Plot kif dc v.7. A L y ' i : 1 1 ~ ~ lilies ~ 1 -41ti I W ; iiiiddlr ]in?, 420 n i ~ ;t ~ i \ v r rline, 424 inp.
(7) See also N. S Bayliss and E. G. McRne, J . P h y s . Chew!., 68, 1002, 1006 (1954). (8) Collision frequencies are proportional t o t h e square root of the absolute temperature. I t might, therefore, be expected t h a t the slopes would be proportional t o t h e square root of t h e absolute temperature. if t h e obser\ ed color formation were d u e to collisional perturbation In actual fact, plots of Ad,/AAoTo us. d? are linear H o w ~ v e rt,h c ~ ~ , , , ~ (,I . r stl,r variations temperat,lre are so small th;,, temperature rive equally good linear plots.
3627
KINETICSOF ANTIGEN-ANTIBODY REACTIONS
Aug. 5 , 1956
ear and pass through the origin, and the slopes increase with increasing temperature. Nevertheless, the possibility that a charge-transfer complex is present in these solutions is not fully eliminated. Even the increase in absorption with increasing temperature is possible depending upon the extent and energies of solvation of the donor, acceptor and c~mplex.~
FROM THE
+ Acceptor
Donor.S.x
+ Accept0r.S.y
=
Complex
Comp1ex.S.z
+ ( x + y - z)S
where S i s the solvent and x , 5' and a are the numbers of solvent molecules solvating each species, respectively.
KORIr< ADAMS,MASS.
(9) T h e reaction in solution is not simply
[COMMUNICATION No. 38
Donor but
DEPARTMENT OF BIOPHYSICS, FLORENCE R. SABINLABORATORIES, UNIVERSITY OF COLORADO MEDICAL CENTER]
Kinetics of the Antigen-Antibody Reaction. Effect of Salt Concentration and pH on the Rate of Neutralization of Bacteriophage by Purified Fractions of Specific Antiserumlavb BY JOHN R. CANNAND EUGENE W. CLARK RECEIVED JANUARY 16, 1956 The kinetics of neutralization of bacteriophage T2r by purified fractions of specific antiserum indicate an electrostatically controlled diffusion process. The effects of salt concentration and PH on the rate of neutralization are interpreted in terms of changes in the electrostatic interactions between oppositely charged antigen and antibody combining sites, modified by changes in collision freauencv between virus uarticles and antibody molecules. It appears that one or more carboxylate groups are involved in
