4152
A. LOEWENSTEIN AND Y. MARGALIT
Nuclear Magnetic Resonance Studies of Nitriles and Isocyanides: Acetonitrile and Methyl Isocyanide
by A. Loewenstein and Y. Margalit Department of Chemistry, Technion, Israel Institute of Technology, Haifa, Israel
(Received June 10, 1066)
~
The positions and line widths of HI, CI3, and NI4 resonances in CHaNC and CHaCN are reported and discussed. The positions of the H' resonances of the OH group in methanol in mixtures with CH3CN, CHaNC, and CC1, were measured. From these measurements the enthalpies of hydrogen-bonded complex formation with CHsCN and CH3NC are estimated. The values obtained are AH = 0.9 and 2.0 kcal./mole for CH3CN and CH3NC, respectively.
Introduction Acetonitrile and methyl isocyanide are isoelectronic symmetric top molecules which show close similarities in their molecular parameters such as the bond l e n g t h ~ l - and ~ electric dipole moment^.^ The significant difference between CH3CN and CH3NC is that whereas the former can be presented in terms of classical bond structures, the latter structures such as CH3N+=C- or CH3--R=C: must be p ~ s t u l a t e d . ~This report presents a study of CH3CN and CHaNC using the HI, W4,and C13nuclear magnetic resonance spectra. The major emphasis is on the study of the hydrogen bond formation of both molecules to CHBOH. Additional information on the electric field gradient at the N14 nucleus is obtained through the study of the NI4 1i.m.r. line widths. Infrared investigations0+' have indicated that in acetonitrile, hydrogen bonds are formed to the nitrogen atom whereas in methylisonitrile, the carbon acts as a proton acceptor. The change in the OH stretching frequency due to H bond formation was found to be smaller in CH3CN as compared to CH3NC,6indicating a stronger H-bond formation in the latter. Mitra' estimates a value of AH = 2.25 kcal./mole (at 30-50") for the complexation constant between CH3CN and CH30H; no data for the corresponding CHsNC system are available. Nuclear magnetic resonance has been applied extensively to the study of hydrogen bonding and complex formation'" in solutions. Similar procedures are used in the present work to evaluate the complexaThe Journal of Physical Chemistry
tion constants and enthalpies between CHsCN or CH3NC and methanol. The N14 n.m.r. line width is proportional to the electric field gradient a t the N14 nucleus and to a correlation time. The latter quantity is approximately equal in liquids of similar viscosity and molecular diameter. A variety of nitrogen-containing compounds has recently been studied in this manner." I n a previous study of the NI4n.m.r. of methyl isocyanide,I2 under low resolution conditions, the resonance was found to be relatively sharp. This indicates a very (1) M. Kessler, H. Ring, R. Trambarulo, and W. Gordy, Phys. Rw., 79, 54 (1950). (2) J. B. Moffat, Can. J. Chem., 42, 1323 (1964). (3) C. C. Costain, J. Chem. Phys., 29, 864 (1958). (4) 9. N. Ghosh, R. Trambarulo, and W. Gordy, ibid., 21, 308 (1953). (5) L. Pauling, "The Nature of the Chemical Bond," 3rd Ed., Cornell University Press, Ithaca, N. Y., 1960,p. 270. (6) A. Allerhand and P. von R. Schleyer, J. Am. Chem. soc., 85, 866 (1963). (7) S. S. Mitra, J. Chem. Phys., 36, 3288 (1962). (8) W.H.Fletcher, C. S. Shoup, Jr.. and W. T. Thompson, Spectrochim. Acta, 20, 1065 (1964). (9) L.L. Ferstrandig, J. Am. Chem. soc., 84, 1323, 3553 (1962). (10) Cf. (a) C. M. Huggins, G. C. Pimentel, and J. N. Shoolery, J. Chem. Phy$., 23, 1244 (1965); (b) H. A. Christ and P. Diehl, Helu. Phys. Acta, 36, 170 (1963); (c) E. Gore and 9. 9. Danyluk, J. Phys. Chem., 69, 89 (1965). (11) (a) W. B. Monitz and H. S. Gutowsky, J. Chem. Phys., 38, 1155 (1963); (b) D. Herbison-Evans and R. E. Richards, Mol. Phys., 7, 515 (1964). (12) J. D.Ray, L. H. Piette, and D. P. Hollis, J. Chem. Phys., 29, 1022 (1968).
4153
N.M.R.STUDIESOF ACETONITRILE AND METHYLISOCYANIDE
Table I: N.m.r. Parameters of CHaNC and CHsCN' Solvent
7
HI (ref. TMS)Position
-
Widthb
N " (ref. HNOi 70%)-Position Width
Neat CHCla (1:1 V O ~ . ) CClr ( 1 :1 vol.) CH30HC
-3.17 -3.12 -3.17 -3.17 to -3.01
CHaNC 0.5 196 0.5 ... 0.5 196 0.5 197
Neat CCl, (1 : 1 vol.) HzO (1:1 V O ~ . ) CHaOH (1 :1 V O ~ . )
-2.03 -2.03 -2.05 -2.03 to -1.98
CHaCN 0.5 0.5 0.5 0.5
'
104 104 to 101" 104 to 114" 104 to 107.5"
Experimental Section Materials. Commercially available C.P. acetonitrile and carbon tetrachloride were used without further purification. Methanol was distilled over CaHz to remove traces of water. Methyl isocyanide was prepared by a procedure similar to that given by Lifshitz, Carroll, and Bauer.14 The product contained about 4% acetonitrile which was not removed. Other solvents were all C.P. products. Spectrometer. Hydrogen-1 spectra were taken with a Varian A-60 spectrometer equipped with a variabletemperature accessory. Tetramethylsilane was used as an internal reference. Carbon-13 spectra, a t natural abundance, were taken with a Varian DP-60 spectrometer equipped with a V4210A variable-frequency unit operating a t 15 Mc. Nitrogen-14 spectra were taken with a Varian DP-60s pectrometer equipped with a V4311 fixed-frequency unit operating a t 4.33 Mc.
(ref. CHs in CHaCN)-Position
CHI
0.5'
...
0.5b 0.5' 81 100 113 81
Resolution determined by field inhomogeneity. P o ~ i t i o mare given in p.p.m. and width in C.P.S. Chemical shift between Cla resonances of CH3- and -NC groups. was studied; see Figures 1 and 2.
low electric field gradient a t the W4nucleus in CH3NC which has been further proven by the ob~ervation'~ of a sharp H1 n.m.r. triplet in CH3NC with JNH = 2.7 C.P.S. These measurements show that with respect to the electric field gradient a t the IV4 nucleus, CH3NC displays a similar behavior to symnietrical species such as the ammonium or tetramethylammonium ions. We have extended these observations by high-resolution N14 n.m.r. measurements and attempted to measure the N14 resonance in solid CH3NC. We are a t present unable to offer any quantitative explanation of this rather surprising phenomenon. Proton n.m.r. work on other is on it rile^'^ seems to indicate that a low electric field gradient is not a unique characteristic of CH3NC but may apply to other isocyanides as well. Further investigation of this problem is planned.
-CIS
NC -157.5 G C H ~ N C ~ = 133.5 G C H , N C ~ = 131.5 Sca3~cd = 133.5 to 108.5"
-24
CH3 0
- 123
CN
0 0 0
- 123 - 123
- 121
Concentration dependence
The samples for both N14and C13spectra were contained in 15-mm. 0.d. Pyrex tubes. The C13resonances were measured from an internal reference (the CH3 resonance of CH3NC or CH3CN), while the N14 resonance were measured relative to an external reference (70% HN03, or saturated "$1 solution) contained in an inner 8-mm. 0.d. tube. No susceptibility corrections were applied. The attempts to detect the N14 resonance in solid CH&C were performed with the V4210A variable-frequency unit operating at 4.3 Mc. and a specially constructed dewar insert.
Results and Discussion A . Measurements on CH3CN and CH3NC. The positions and widths of the H', C13, and N14 resonance in CH3NCand CH3CN are given in Table I. 1. H 1Resonances. The methyl resonance in CH3NC is shifted about 1 p.p.m. lower than the corresponding resonance in CH3CN. This shift is probably due to different bond anisotropy and to the partial positive charge on the nitrogen atom in CH3NC. The methyl resonance in CH8NC displays a triplet of equal intensi= 2.35 C.P.S. This value is slightly ties with JNH lower than the value of 2.7 C.P.S. reported previously.18 The C13-H spin-spin interaction in the methyl group of CH3CN has been measured previously16 and found to be 136 C.P.S. We have measured the same quantity for CH3NC and found it to be equal in magnitude t o that in CHSCN. (13)I. D.Hunte, P. von R. Schleyer, and A. Allorhand, J . Chem. Phys., 35,1533 (1961). (14) A. Lifshitz, H.F. Carroll, nnd 5. H. Bauer, J. Am. Chem. SOC., 86, 1488 (1964). (15) N. Muller and D. E. Pritchard, J . Chem. Phys., 31, 1471 (1959); H. Dreeskamp, E. Sackmann, and G. Stegmeier, Ber. Bunsages. Physik. Chem., 67, 860 (1963).
Volume 69, Xumber 18 December 1966
A. LOEWENSTEIN AND Y. MARGALIT
4154
114
A
IIO
fi
Y
$106 Lu
102
0.2
0.4
0.6
Mole fraction of
0.8 CH,CN
1.0
Figure 1. Chemical shift of the NI4 resonance in CHsCN as a function of its mole fraction in various solvents. Shifts are measured relative to the N14 resonance in 70% aqueous HNOa a t 30 f 2'.
2. N14 Resonance. The N14 resonance in CHsNC is shifted to higher field than in CH3CN by about 90 p.p.m. It displays a 1:3 :3 :1 spin-spin quadruplet with J N H = 2.3 C.P.S. which agrees well with the value obtained from the H1spectra. The position of N14 resonance in CH3NC does not change on dilution with polar or nonpolar solvents. I n CH3CN however, such a dependence is observed and is particularly large in aqueous solutions. The results are presented in Figure 1. The shift in hydrogen bond forming solvents is to higher fields, opposite to HI shifts which are to lower fields. This suggests that the hydrogen bonds are formed to the nitrogen atom in CH3CN. A larger shift, to higher field, of the "4 resonance in pyridine-methanol solutions was observedle and attributed to hydrogen-bond formation. The shifts to lower fields in ccI4 or CeH6 are solvent effects not related to intermolecular hydrogen bonding. The line width in CH3NC is only about I/IOO to l/m as large as it is in CH3CN. We may use the relationship"
where TI is the longitudinal relaxation time, e2Qq/h the quadrupole coupling constant (in Mc.), and rq is the correlation time (assuming the asymmetry factor 17 = 0), to estimate the upper limit of e2Qq/hin CHaNC. Taking Tz = TI, where TZis the transverse relaxation time, and assuming equal rq values for both CH3NC and CH3CN and e2Qg/A = 4.35 Mc. for CH3CN,11 we obtain e2Qq/A < 0.3 Mc. for CH3NC. From microwave spectra,1 an upper limit of 0.5 Mc. was estimated for this quantity. If the solid, state line width is comThe Journal of Physical Chemistry
pletely determined by the quadrupolar interaction, we should be able to observe this resonance, provided Mc. the interaction is smaller than about 6 X (corresponding to about 20-gauss line width). Attempts to observe the N14 resonance in solid CH3NC (at -70") failed. I n solid NH4CI, where the electric field gradient is zero, N14 resonances were observed, which suggests that dipolar broadening is not the cause for not observing the resonance in solid CHsNC. The N14 line width in CH3CN is slightly solvent-dependent and probably related to small changes in Tq. Similar observations have been noticed before. lib 3. C13 Spectra. Line widths could not be measured with our experimental setup. The chemical shift between the CH3- and the -CN or -NC is larger in methylisocyanide than in acetonitrile by about 10 p.p.m. Also, all resonances are shifted to higher fields in the latter. The most significant, result however, is that in a hydrogen bond forming solvent (methanol) the C13 resonance of the -NC group is concentrationdependent (shifts to higher fields) whereas the resonance position of the -CN group in CH3CN remains unchanged. This strongly indicates (cf. N14 shifts) H-bond formation to the C atom in CH3NC. The detailed shifts and the calculated curve (which shall be discussed later) are given in Figure 2. B. Measurements on CH30H. The equilibrium constants of hydrogen-bonded complexes of CH&N and CH3NC with CH30H were estimated from the H' resonance shifts of the OH group in the methanol. The results are shown in Figures 3 and 4. In order to take into account the shifts due to changes in the self-association of the methanol we have measured the OH shift of methanol in cc14 (Figure 5 ) . The curves given in Figures 3-5 were analyzed by a procedure analogous to that given by Huggins, Pimentel, and Shoolery.loa The observed chemical shift, 6obad, is the sum of two terms
where x is the concentration of the complex, c is the concentration of the methanol, and 6, and 6, are the shifts of pure complex and pure methanol, respectively. Assuming that by subtracting from 6obsd the shift of CH30H in CC1, a correction for the self-association effects is applied, we obtain n.
(16) H. S a i 6 and K. Nukada, Tetrahedron Letters, 2, 111 (1965).
N.M.R.STUDIES OF ACETONITRILE AND METHYL ISOCYANIDE
4155
I
I
0.2
0.4
I 0.6
1
OB
ID
MOLE FRACTDN OF CH,W
I
1
02
I
I
I
04 06 Q8 Mole fraction of CH,NC
Figure 4. The HI chemical shift of the OH group in methanol in mixtures with CH3NC at various temperatures. Reference is the same as in Figure 3.
I
LO
I I
Figure 2. The C'3 chemical shift of the NC group in CHaNC as a function of its mole fraction in methanol. Shifts are measured relative to the Cl8 resonance of the CH3 group in CHaCN. The line is the calculated curve (see text); temperature, 30 rrt 2'.
CCI,-
I
I
CH,OH
b
5.0-
-s m
I
I
I
I
-z
b
4.0-
CH,CN-CH,OH
_i Y 1
I
>
-, '/
3.0
,/
I 30"
/
r.
I
-
A -18.50 0 0 0
/
1
1
/
/
I
1
1
I
where 610bed is the difference between the OH shift in CH8CN (or C H 8 C ) and CCL at the same concentra tion of methanol. Assuming further that as c --t 0, also z --t 0, one 6, =
(6obsd)c-0
f
1 K[(fiobsd)c+O
-
sc]
(4)
Volume 69,Number 18 December 1966
4156
A; LOEWENSTEIN AND Y. hlARGALIT
where (6obsd)G-0 is the extrapolated value of the OH shift in infinite dilutions in CHaCN (or CH3NC) and is taken from Figure 3 or 4. The equilibrium constant K is defined as
K =
x(b (c
+ c - x)
- x ) ( b - z)
where b is the concentration of the CH&N (or CHaNC). Combining eq. 3, 4, and 5, one may write 6'obsd in terms of K and the concentrations of methanol and CH3CN (or CHQNC) in the solution. The values of K were then computed by a trial and error procedure. The best values of K thus obtained are given in Table 11. The temperature dependence of K can be represented by AH = 0.9 and 2.0 kcal./mole for CH3CN and CH3NC,respectively. Table I1 : Equilibrium Constants for Hydrogen-Bonded CompIexes in CHsCN and CHaNC
temperature. Therefore, no estimate for the limits of errors was calculated. The orders of magnitude of the data are, however, very plausible and roughly agree with infrared The stronger H bonds to CHBNCas compared to CH&N are evident from the larger value of AH for the former. Also, this result is consistent with the data obtained from infrared spectra. A correction for self-association was applied in a study of hydrogen bonding of alcohols in various solvents.18 This method, which involves computer curve fitting, should, in principle, give more reliable results. However, the low precision of our measurements and the other approximations still uncorrected for put in doubt the justification for the use of this method in our case. An independent check on the consistency of the results was performed in the following manner: The CY3 shifts of the NC group in CH3NC solutions in methanol (Figure 2) can be written as aCls(obsd)
Temp.,
-K----.-----,
7
OC.
CHsCN
CHsNC
30 8
2.6 2.4 2.0
2.1 1.6
-18.5
3.0
Obviously, the model on which this calculation is based involves some rather over-simplifying approximations. The worst assumptions probably are the correction applied for self-association and the neglect of specific solvent effects. Another uncertainty lies shown in in the extrapolation to obtain (&,bsd)c--+O Figure 3 or 4. At relatively high methanol concentrations its spectrum already turns into an ABatype which makes the determination of the OH position rather difficult. Furthermore, 6, might be temperaturedependent17and this may account, in part, for the variation of the calculated equilibrium constant with the
The JOUTWEof Physical Chemistry
=
z(oomplex)6CIa(oomplex)
+
z(CHsNC)80qCHaNC)
(6)
where the z's denote mole fractions. The x values at 30" were calculated from the equilibrium constant (Table 11) and the value of 6cxa(complex) was chosen so as to give best fit with the measured points. We thus obtained 6Cis(oomplex) = -124 p.p.m. and calculated the curve which is shown in Figure 2. Measurement of the Cla resonance in the range of XCH~NC< 0.4 would serve as a critical test to this calculation. Unfortunately, however, C18 resonances at these low concentrations could not be measured with sufficient accuracy due to experimental difficulties. Similar calculations on the N f 4 resonance in the CHZCNCH30H system were not performed due to the small over-all shift of the NI4. (17) Cj. N. Muller and R. C. Reiter, J. Chem. Phys., 42, 3265 (1965); we are indebted t o a referee for pointing out this reference. (18) C. Lussan, J . chim. phys., 60, 1100 (1963); we are indebted t o a referee for pointing out this reference.