Donald E. Leyden and Ronald E. Channell
1562
primary isotope effect, giving rise to a maximum in the curved plotted in Figure 1, pKa(HzO) us. pKa(DzO) pKa(Hz0). Change in medium (acid molarity) affect the solvent isotope effect. As acidity is increased, the solvent isotope
effect should decrease due to solute-solvent solvation changes but increase due to a net loosening of H2O (or D2O) bonds. The only factor which may be sorted out is that nitro groups are strongly and specifically solvated in strongly acidic medium.
Protolysis and Nitrogen inversion of Anilines in Sulfuric Acid' Donald E. Leyden" and Ronald E. Channell Department of Chemistry, University of Georgia, Athens, Georgia 30601 (Received August 17, 1972)
The protolysis and nitrogen inversion kinetics of several substituted anilines have been studied using water-sulfuric acid as the solvent. Line shape analysis of the high-resolution nmr spectra of the amines was used to determine the rate of reaction. The studies were performed in the 50-90" range to reduce line broadening because of the high solution viscosity. An acidity function was used as a measure of acidity. Results indicate the principal mechanism of proton exchange is similar to that of alkyl amines, but with larger rate constants. The rate constant for breaking of an amine-water hydrogen bond decreases with increasing size of the N-alkyl substituents. The rate of the nitrogen inversion process is closely related to the protolysis rate.
Introduction Recently the kinetic analysis of proton exchange between ammonium ions and aqueous acid has been extensively studied.2-8 Both high-resolution2J,5>7,8 and pulsed4,6,7 nuclear magnetic resonance techniques have been employed, and data are available over a wide range of acid and ammonium ion concentrations. A rather extensive study has also been conducted in the area of nitrogen inversion in aliphatic amines.gJO However, there has been very little study of the proton exchange of aromatic amines such as anilines,ll and essentially no work has been done in the area of nitrogen inversion of aniline compounds.12 In previous studies, a similar kinetic scheme appears to prevail over a wide range of acid and/or amine concentrations6 as well as amine structures. Apparently, the variations in the rate constants of the exchange and inversion reactions are not simply related to the basicity of the amines or steric factors.6 This paper is a report of the results of an investigation of the proton exchange kinetics of N-methylaniline (I), N,N-dimethylaniline (11), N-methyl-N-ethylaniline (III), N-benzylaniline ( W ) , N-methyl-N-benzylaniline (V), N-ethyl-N-benzylaniline (VI), and N-methyl-N-benzyl-p-anisidine (VII).The results of an investigation of nitrogen inversion of V, VI, and VI1 in sulfuric acid and deuteriosulfuric acid are also discussed. Although extensive use of high-resolution nmr for studies of the type reported here have been described repeatedly, a brief discussion of the basis of the method is given for clarity. In the case of proton exchange, it is most convenient to use the broadening and eventual coalescence of a multiplet resulting from spin-spin coupling between the labile N-H proton and protons of an N-alkyl group such as a methyl group.2 In the case of tertiary methyl amines, 9 1 3
The Journalof Physical Chemistry, Vol. 77, No. 72, 1973
a simple doublet is observed for the methyl protons. To detect inversion of the nitrogen atom, some type of diastereotropic probe is required. Most convenient is the benzyl group because the methylene protons give rise to a clean AB pattern in the limit of slow interchange between the enantiomers. Because the magnetic environments of the two methylene protons interchange upon inversion of the amine nitrogen atom, the AB pattern will broaden and eventually coalesce in the limit of fast inversion. Because nonlabile protons are observed, and coupling to the exchanging proton is not desired, these studies may be performed in deuterium solvents.
Experimental Section All anilines with the exception of VI1 were obtained from Aldrich Chemical Co. or Eastman Organic Chemical Co. and were purified by vacuum distillation. VI1 was prepared from p-anisidine which was obtained from Eastman Organic Chemical Co. N-Benzyl-p-anisidine was prepared This investigation was supported in part by Public Health Service Grant No. GM-13935 from the National institutes of Health. E. Grunwald, A. Loewenstein, and S. Meiboom, J. Chem. Phys., 27, 630 (1957). A. Loewenstein and S. Meiboom, J. Chem. Phys., 27, 1067 (1957). E. K. Ralph, I l l , and E. Grunwald, J. Amer. Chem. Soc., 89, 2963 (1967). R. J. Day and C. N. Reilley, J. Phys. Chem., 71, 1588 (1967). E. Grunwald and E. K. Ralph, I l l . J. Amer. Chem. SOC., 89, 4405 (1967). E. Grunwald and A. Y . Ku, J. Amer. Chem. Soc., 90, 29 (1968). D.E. Leyden and W. R. Morgan, J. Phys. Chem., 73,2924 (1969). W. R. Morgan and D. E. Leyden, J. Amer. Chem. Soc., 92, 4527 (1970). D. E. Leyden and W. R. Morgan, J. Phys. Chem., 75,3190 (1971). E. Grunwaid, R. L. Lipnick, and E. K. Ralph, J. Amer. Chem. SOC., 91,4333 (1969). H. Kessler and D.Leibfritz, Tetrahedron, 25, 5127 (1969). M. Emerson, E. Grunwald, M. Kaplan, and R. Krornhout, J. Amer. Chem. SOC., 82, 6307 (1960).
Protolysis and Nitrogen Inversion of Anilines by treating p-anisidine with benzaldehyde to prepare the corresponding Schiff base, which subsequently was reduced with sodium b0r0hydride.l~VI1 was then prepared by treating N-benzyl-p-anisidine with formic acid,l5 followed by the reduction of the resulting amide with lithium aluminum hydride. The product was confirmed by nmr. Stock solutions were prepared from the anilines and sulfuric acid and were diluted to the desired aniline and/ or acid concentrations. Each solution was then analyzed by potentiometric titration. Aniline concentrations ranged from 1.0 to 0.25 M with excess sulfuric acid present from 60 to 40%. Deuteriosulfuric acid was prepared by reaction of phosphorus pentoxide with fuming sulfuric acid under a nitrogen atmosphere. The resulting sulfur trioxide was bubbled through deuterium oxide. The per cent H in the solution was determined by standard addition of H2O to the solution followed by integration of the HDO nuclear magnetic resonance line. Back extrapolation of the plot of area us. added H2O yielded the per cent H present in the D2SO4 which was found to be less than 0.3%. Dissociation constants were determined at 50" by differential potentiometric titration.16 Nmr measurements were made at 50" for I, 11, and 111, a t 70" for V and VII, and a t 90" for IV and VI, These temperatures were selected in order to avoid viscosity broadening of the nmr lines. Nuclear magnetic resonance data were obtained on a Hitachi Perkin-Elmer R-20 nuclear magnetic resonance spectrometer operated under slow passage conditions. Routine checks were made to ensure against saturation. The temperature was regulated to within f l " using a standard R-20 variable temperature probe. Exchange rate parameters were obtained by comparison of computer-simulated line shapes with experimental ones. The simulated mean residence time, r , of a proton on a given aniline molecule was adjusted so that a minimum standard deviation between computed and experimental points on the spectrum was obtained. The program used was prepared using equations similar to those given by Arnold1? for spin-coupled systems in which the coupling is small compared with the chemical shifts of the coupled nuclei. The simulated residence time of a proton before A-B interchange was used to obtain the inversion rate parameters and was also adjusted so that a minimum standard deviation between computed and experimental points in the spectra was obtained. The program was prepared utilizing the equations of Alexander.ls In both cases the natural line width was taken from the line width at half height of the group under study in solutions of low acidity in which the exchange or inversion is very rapid and was found to approximately 0.5 Hz in most cases. This measurement was made for each experimental point and the value was assumed to be controlled by inhomogeneties in the magnetic field. Although this line width is not a measurement of T2 as desired, there were reasons for this choice as an approximation of Tz. First, in sufficiently acidic solutions to achieve "nonexchanging" conditions for more accurate T2 measurements, viscosity caused serious line broadening. It was apparent that attempts to correct this phenomenon would introduce considerable errors. Experimental points were not taken in this region of acidity. Second, a comparison of the line widths of aliphatic amines in the slow exchange limit in aqueous acid gave values within 0.1 Hz. This observation implies that factors other than a pure T2 relaxation are controlling the line width. Finally, considering that data
1563
TABLE I: Nmr Spectral Parameters for Compounds Studied Amine
JAXa
I II Ill IV V VI VI1 a
5.67 5.16 5.15 4.96 5.16 3.89,(7.44c)
5.00
All in Hz. All in Hz at 60 MHz. J,x
JAB0
ALABb
12.70 12.84 12.75
11.48 15.41 13.47
in Hz.
were taken only from the middle of the nmr kinetic window ( i e . , extensive line broadening), errors caused by T2 estimation should be minimal. Thus, the choice was made to use the line width at the fast exchange limit as the best available source of a natural line width. The spin-coupling constants J between the N-H proton and the group under study in the various compounds are given in Table I. The spectral parameters for the A-B pattern are also given in Table I. The sulfuric acid concentration required for these studies was in the range of 40-60%. Rate data were plotted us. the Hammett acidity for tertiary amines given by Arnett and Mach.19 All data were treated with standard leastsquares techniques. The precision of the data for both exchange and inversion processes varied between 5 and 10% relative standard deviation. When data repeated on different runs are included, an estimate of 15% relative standard deviation is obtained. This is reasonably typical of nmr kinetic data.
Results There are several possible mechanisms of proton exchange in the H2S04-H20 solvent system that have evolved in previous investigations.6,7J3,20However, eq 1 and 2 represent those which repeatedly have been shown to be the predominant mechanisms in acidic solutions.6~8J3 R3NH+ R,NH+
+
+
H20
+
9-H H
NR3
-
R,N
+
H30f
(1)
k2
R3N
+
H-0
I
+
lHNR3 (2)
H Reaction 1 has been shown to occur in two steps.6J3 These are given as R,NH+---H20
+
H20
R3N---H20
+
H30+
k-a
(34
R3N---H20 R3N + H 2 0 (3b) By applying a steady-state approximation to the concen(14) S.Yamadaand S.Ikegami, Chern. Pharrn. Bull., 14,1382 (1966). (15) "Organic Synthesis," Collect. Vol. I l l , Wiley, New York, N. Y., 1964,p 590. (16) A. L. Bacarella, E. Grunwald, H. P. Marshall, and E. L. Purlee, J. Org. Chern., 20,747 (1955). (17) J. T. Arnold, Phys. Rev., 102,136 (1 956). (18) S.Alexander, J. Chern. Phys., 37, 967 (1962). (19) E. M. Arnett and G. W. Mach, J. Arner. Chern. SOC.,86, 2671 (1964). (20) D. E. Leyden and J. F. Whldby, J. Phys. Chern., 73, 3076 (1969) The Journal of Physical Chemistry, Vol. 77,No. 12, 7973
Donald E. Leyden and Ronald E. Channel1
1564
tration of R3N- - -Ha0 in reaction 3 and combining the result with reaction 2 a total rate equation may be obtained.6>8,9J3.20This is given by
(4) This equation has been widely and successfully used to explain the protolysis kinetics of a variety of amines. Although many combinations of values of the parameters of eq 4 are possible, the most general case predicts first an increase of the proton exchange rate with a decrease in acidity, followed by an acid-independent region where k H >> k-a[H+] and the second-order term is negligible, and finally an increase in rate with a decrease in acidity as the second term in the equation becomes large compared with ka. However, for each aniline compound studied, a plot of 1 / US. ~ l/hO"' was linear over the entire acid range employed and none of the anilines exhibited second-order dependence in this acid range. The exchange phenomenon may be represented by eq 5 where ka/k-, is assumed to equal K,, the dissociation constant for the anilinium ion. The value of [H+] must be represented in terms of an acidity function.
A scheme for the nitrogen inversion of amines in acid solution has been proposed as9 ~ 0 3 2 1 R3"+---H20
+ H,O
ka
k-a
R3N---H20 3- H,O+
pQ+
R,N
k-H
(6)
+ HzO R&
where R3*N represents an inverted species of R3N. The rate of inversion is described in eq 7 in which l/rl represents the observed rate of A-B interchange, k, is the inversion rate constant and k-H is the pseudo-first-order rate constant of rehydration (and/or) reprotonation of the amine. In eq 7, f = k,/(ki k-H) or the fraction of dehy-
+
drated amine molecules which undergo nitrogen inversion before reprotonation or rehydration. In this scheme it is assumed that neither protonated or hydrated (RaN. HzO) amine molecules can invert at a rate comparable to the free amine. In concentrated sulfuric acid K H