HYDROGEN BONDING SUCCINIMIDE-DIMETHYL SULFOXIDE SYSTEM
area (“large” particles) area (“small” particles)
where the subscripts 1 and 2 refer to the “large” particles and the “small” particles, respectively. It should be noted that the area rat,io is calculated on the basis of spheres for ease of computation. A gaussian distribution of monoclinic, cubic, or odd-shaped crystals would not give a rat8iotoo different from 0.08. This
3779
can be verified by performing the integration numerically from the Micromerograph data. If we compare the ratio of the specific rate constants as observed in (3) with the calculated ratios of the corresponding areas in (7), we find that the ratio of the specific rate constants is only four times that of the calculated areas. Considering the many approximations made in calculating the ratio of the areas for these two samples, we believe that the results support our contention that certain specific sites, probably nonbasal dislocations, on the KClO3 grains become activated by the ultrasonic field thus leading to reaction with neighboring aldehyde molecules.
A Nuclear Magnetic Resonance Study of Hydrogen Bonding in the Succinimide-Dimethyl Sulfoxide System
by David M. Porter and Wallace S. Brey, Jr. Department of Chemistry, University of Florida, Gainesville, Florida S.8601 (Received March 17, 1967)
The hydrogen bonding between succinimide and dimethyl sulfoxide has been studied by nuclear magnetic resonance. The concentration and temperature dependence of the chemical shift of the NH resonance of succinimide shows that the succinimide-dimethyl sulfoxide interaction is probably of the “n-donor” type. The experimental chemical shift data were interpreted in terms of an equilibrium between a succinimide-dimethyl sulfoxide complex and a succinimide dimer plus solvent molecules. The equilibrium constant for this equilibrium, written for the dissociation of the succinimide-dimethyl sulfoxide complex, was 0.096 at 33” and determinations of the equilibrium constant at four temperatures gave a AH of 5.0 f 1.0 kcal/mole and a AS”, of 12 f 4 cal/mole deg. The chemical shifts for the hydrogen-bonded protons in the succinimide-dimethyl sulfoxide complex showed a regular trend to higher field with increasing temperature.
Introduction Proton magnetic resonance measurements have shorn that a significant chemical shift change accompanies the formation of a hydrogen The formation of hydrogen bonds displaces the magnetic resonance of the protons toward lower magnetic except in certain cases involving aromatic mO~eCuleS.3 If the
primary function of the electron donor atom is to produce a strong electric field in the vicinity of the hydrogen-bonded Proton, then a shift toward lower field is (1) U. Liddell and N. F. Ramsey, J . Chern. Phys., 19, 1608 (1951);
J. T.Arnold and M. E. Packard, ibid., 19, 1608 (1951).
(2) H. S. Gotowsky and A, Saika, ibid,, 21, 1688 (1953); C, M. Huggina, G.C.Pimentel, and J. N. Shoolery, ibid., 23, 1244 (1955).
Volume 71,Number 12 November 1967
DAVID31. PORTER AND WALLACE S. BREY,JR.
3780
quite rea~onable.~The electric field deforms the electron distribution about the proton in the hydrogen bond, decreasing the electron density in its vicinity and increasing its asymmetry; both of these effects decrease the proton hi el ding.^ I n the case of an equilibrium between hydrogenbonded and nonhydrogen-bonded species, if the lifetimes in the associated and unassociated states are short, the diamagnetic shielding is an average over the two states and the proton resonance will be observed at a frequency corresponding to the average shielding for the two states. Since both temperature and concentration affect the extent of association, the chemical shift of the proton varies with these conditions. The purpose of the present investigation is to study the nature of the hydrogen bonding in the succinimidedimethyl sulfoxide system. Of primary interest are the various chemical equilibria which may be present in solutions of succinimide in dimethyl sulfoxide and the calculation of the equilibrium constants and the changes in enthalpy and entropy for these equilibria from the nuclear magnetic resonance chemical shift data. Also of interest are the chemical shifts of the hydrogenbonded protons present in the species involved in these equilibria.
Experimental Section The spectra of the solutions of succinimide in dimethyl sulfoxide were obtained using a DP-60 Varian Associates nuclear magnetic resonance spectrometer operating a t 60.0 Mc/sec. The spectrum of the saturated solution of succinimide in deuterated chloroform was obtained using a IIodel A-60A Varian Associates nuclear magnetic resonance spectrometer operating a t 60.0 Mc/sec, with the aid of a Model C-1024 Varian Associates time-averaging computer. ' The dimethyl sulfoxide used was obtained from the J. T. Baker Chemical Co. and was dried over Type 4 6 Fisher Molecular Sieve. The succinimide, from Eastman Organic Chemicals, was dissolved in the dimethyl sulfoxide immediately after opening without additional purification. The solutions of succinimide in dimethyl sulfoxide were made up by weight. The chemical shift of the NH was determined by the usual side-band method with the audio frequency continuously monitored by a Model 523B HewlettPackard counter. The concentration dependence of the dimethyl sulfoxide peak was checked with respect to internal tetramethylsilane reference and found to be negligible. Small quantities Of Water added to Various samples confirmed that the presence of moisture has no measurable effect on the NH chemical shift. The chemical shift, A, is defined as the number of Th,e Journal of Physical Chemistry
cycles per second downfield from the dimethyl sulfoxide peak. All chemical shifts are the average of five individual determinations and are considered to be accurate to within 1 cps or less. This apparently large maximum error is a consequence of the uncertainty in estimating the position of the center of the broad NH peak, especially for the very dilute solutions. The temperature of the sample was regulated by the flow rate of dry nitrogen through a Varian V-4340 variable temperature probe assembly. A copper-constantan thermocouple positioned within the dewar insert was used to determine the temperature of the sample. The thermocouple was calibrated against a second thermocouple which was positioned in a sample tube in the spectrometer. All temperature measurements are considered to be accurate to within 1 deg. The experimental results are given in Table I.
Table I : Observed Chemical Shifts" of the NH of Succinimide in Dimethyl Sulfoxide Mole
fraction of succinimide
0.096 0.137 0.193 0.271 0.348 0.412
7 -
Temperature, oC--------
33
43
57
72
507.1 505.6 503,4 500.4 494.8 489.3
505.6 504.9 500.7 498.6 493.0 487,s
499.4 497.0 495.2 492.3 487.2 483.1
494.8 492.8 489.7 485.7 479.8 475.6
a Expressed in cycles per second downfield from dimethyl sulfoxide resonance.
Treatment of Data A saturated solution of succinimide in deuterated chloroform, in which succinimide is very sparingly soluble, was investigated using the A-60A and the timeaveraging computer in an attempt to estimate t,he chemical shift of the NH in the unassociated or selfassociated succinimide, whichever the case may be. The chemical shift obtained provides a downfield limit for the chemical shift of the NH in the succinimide monomer and an upfield limit for the chemical shift of the NH in the succinimide dimer. A study of the self-association of succinimide in an inert solvent would have aided this work considerably, but succinimide is not sufficiently soluble in inert sol(3) J. A. pople, w. G. Schneider, and H. J. Bernstein, "HighResolution Nuclear Magnetic Resonance," McGraw-Hill Book CO., Inc., New York, N. Y.,1959. (4) H. S. Gutowsky, Ann. N. Y . Acad. Sci., 7 0 , 786 (1958); P. J. Frank and H. s. Gutowsky, Arch. sci., 1 1 , 216 (1958).
HYDROGEN BOXDING SUCCINIMIDE-DIMETHYL SULFOXIDE SYSTEM
vents such as carbon tetrachloride and cyclohexane for nuclear magnetic resonance investigation. I n fact, dimethyl sulfoxide is the only solvent tried in which the NH peak of the succinimide was readily observable over a significantly large concentration range. Dimethyl sulfoxide is also a suitable choice in view of the work of Allerhand and S ~ h l e y e rwho , ~ concluded that in order to extend the spectral shift criterion to very weak proton donors, very strong proton acceptors must be used. They found that hydrogen bonding equilibria involving dimethyl sulfoxide are favorable and explained this phenomenon in terms of the availability of the two equivalent p-orbital bonding sites on the oxygen in dimethyl sulfoxide. They also reported that the infrared spectrum of dimethyl sulfoxide in carbon tetrachloride solution, up to 100yo dimethyl sulfoxide, was not concentration dependent. This is evidence against self-association of the solvent. Each solution of succinimide in dimethyl sulfoxide displayed only one proton resonance which could be attributed to the S H proton of succinimide. The -CH2CHz- resonance was concealed by the solvent peak. The chemical shift of the N H resonance was strongly dependent upon concentration and temperature, moving to lower fields with increased concentration of dimethyl sulfoxide and to higher fields with increased temperature. This direction of change of the chemical shift is consistent with the results of similar studies of other hydrogen-bonding systems‘ in which ‘h-donor” association occurs. Dimethyl sulfoxide hydrogen bonds to proton donors by use of the lone pairs of electrons on the oxygen, as pointed out by Allerhand and Schleyer.6 An attempt was made to fit the experimental results with either of two equilibria
3781
/CH3 2
0-s
(2)
\CH3
The equilibrium constant in mole fraction units for reaction 1 is given by
where N = moles of total succinimide, S = moles of total dimethyl sulfoxide, C = moles of succinimidedimethyl sulfoxide complex, and M = N - C = moles of succinimide monomer. The observed chemical shift, A, for the N H in the system is expressed as a weighted average of the contributions of the two YH-containing species present and is given, for this case, by
where A, = chemical shift of the NH in the succinimide monomer and A, = chemical shift of the N H in the succinimide-dimethyl sulfoxide complex. Solving eq 4 for C
Solving eq 3 for C, the physically reasonable root obtained is
C =
N
+ S - d ( N + ~ 5 -) ~4NS/(1 + K1) 2
(6)
The equilibrium constant in mole fraction units for the other possible equilibrium, as shown in eq 2, is given by
Kz
=
1/2(N - C ) ( S - C)Z - D ( S - cy C2P/2(N - C) XI C 2 ( D S )
+
+
(7)
where D = l/z(N - C) = moles of succinimide dimer. The observed chemical shift of the S H is given for this case by
( 5 ) A. Allerhand and P. von R. Schleyer,
J. Am. Chern. Soe., 8 5 , 1715
(1963).
Volume 71, Number 12 -Vozember 2367
DAVID14. PORTER AND WALLACE S. BREY,JR
3782
510
where A d = chemical shift of the NH in the succinimide dimer. Solving eq 8 for C (9)
I
I
I
1
0.1
0.2
0.3
0.4
505
Solving eq 7 for C gives the third degree equation
+ 2S)Cz + (S2+ 2NS)C - N S 2 = 0
(1 - KZ)C3- (1 - K 2 ) ( N
5 00
(10)
Method of Calculation An initial estimate of A, is obtained by extrapolating a plot of A vs. mole fraction of succinimide to infinite dilution. A value is assumed for the other limiting chemical shift (A, for reaction 1, A d for reaction 2 ) and C is then calculated for reaction 1 from eq 5 or for reaction 2 from eq 9. The equilibrium constant is then calculated (K1 from eq 3 or K z from eq 7) using the appropriate value of C. The limiting chemical shifts are varied over the chemically reasonable ranges for these shifts. The values of the equilibrium constant for the various solutions are averaged and from this average equilibrium constant there is then calculated a value of C from eq 6 for reaction 1 or from eq 10 for reaction 2. The new approximation to the value of C is then used to calculate a predicted chemical shift from eq 4 for reaction 1 or from eq 8 for reaction 2. The values of A, and Am or Ad are varied over their chemically reasonable ranges until the best least-squares fit between the observed A and the predicted A is obtained. This best leastsquares fit yields the values of A,, A,, and K1 when considering reaction 1 and A,, A d , and K z when considering reaction 2, for one temperature. This process is then repeated for each individual temperature. The logarithm of the equilibrium constant is plotted against the reciprocal temperature and AH is estimated from this plot. A S is then calculated from AH and the equilibrium constant. FORTRAN IV programs were written for these calculations and all calculations were carried out at the University of Florida Computing Center on an IBN 709 computer.
Results and Discussion Reaction 1 did not fit the experimental results for reasonable values of the limiting chemical shifts A, and Am. The results of the calculations based on reaction 2 are listed in Table I1 and are shown in Figures 1 and 2. Table I1 lists the limiting chemical shifts and the equilibrium constant as functions of temperature based on reaction 2. Figure 1 shows the concentration dependence of the observed chemical shift of the NH in The Journal of PhysicaE Chemistry
49 5