5218
"N Nuclear Magnetic Resonance Spectroscopy. XIII. Pyridine-"" Robert L. Lichter and John D. Roberts* Contribution No. 4196 ,from the Gates and Crellin Laboratories of Chemistry, California Insfitute qf Technologj, Pasadena, California 91 109. Received January 18, 1971 Abstract: From high-resolution proton, 13C,and JjNnuclear magnetic resonance spectra of ljN-enriched pyridine and its hydrochloride, all of the chemical shifts, IjN-H, and l5N-I3C coupling constants have been obtained. Attention is centered on the coupling constants, which are compared with other values and discussed in light of the known inapplicability to pyridine of the average energy approximation. The coupling 2Jskr exhibits the normal dependence on geometry, and the algebraic increase observed on protonation parallels the behavior exhibited by other acyclic compounds with unsaturated nitrogen. The coupling lJscin pyridine increases 30-fold on protonation, which accords with qualitative predictions. In contrast to the absence of C-N coupling over more than one bond in saturated systems (glycine, alanine, 1,l-dimethylhydrazine),pyridine and pyridinium ion display substantial C-N couplings over two and three bonds. Because large two-bond C-N couplings are also observed in dimethylnitrosamine as well as in some amides, the presence of a 7~ system is felt to be important for the longer range couplings. Substantial solvent effectsare observed on these couplings which may be attributed to increasing hydrogen bonding at nitrogen. On the basis of changes in liN chemical shifts and IjN-H coupling constants, lJSc.is shown to very likely undergo a change in sign when the nitrogen is protonated.
As
the availability of IjN-labeled compounds has grown, measurement of spin-spin couplings between nitrogen and other magnetic nuclei has evoked increasing interest, not only for theoretical reasons, but also because of the obvious utility in structure elucidation. The available data permit some generalizations as to the relationships between structure and nitrogenproton coupling constants. Typically, two-bond couplings in saturated systems are of the order 0--1 H z , ' , ~ while unsaturation at either carbon or nitrogen leads to values with enhanced magnitude. 2 , 4 Three-bond couplings in saturated systems are generally larger than two-bond couplings, 2 . 3 and the enhancement on unsatMore signifiuration is not as great.s,3b'~b.c,e,f,~c,fi cantly, in compounds with fixed geometry, both twobond and three-bond couplings are larger in magnitude than that proton which lies closer i n space to the lonepair electrons on nitrogen ;oc,d this observation has been exploited extensively in assigning preferred conforniations of nitrosamines and related compounds.' In par( I ) (a) Part XII:
R. L. Lichter and J. D. Roberts, J . Amer. Chem.
Soc., 93, 3200 (1971); (b) supported by the Public Health Service,
Grant No. GM-11072-07, from the Division of General Medical Sciences. (2) G. Binsch, J. B. Lambert, B. W. Roberts, and J. D. Roberts, J . Amer. Chem. Soc., 86, 5564 (1964). (3) (a) E. W. Randall and D. Shaw, Spectrochim. Acta, Purt A , 23, 1235 (1967); (b) W. McFarlane, J . Cheni. SOC.A , 1660(1967); ( c ) F. G . Riddell and J. M . Lehn, ibid., B , 1224 (1968); (d) R . L. Lichter and J. D. Roberts, Spectrochim. Acta, Part A , 26, 1813 (1970). (4) (a) A. J. R. Bourn and E. W. Randall, J . Mol. Spectrosc., 13, 29 (1964); (b) A. I acliieoiis \\orI,-tip and distillation in a nitrogen atmosphere, airorded I .I-dimeth! Iii~~lraii1ie-2-~"h', Proton spectra \4ere tahen on a Varian A 56/60 spectrometer, and linc po\ition\. rncastired either w i t h respect t o internal cqclopentane (neat p)ridine) or internal methanol (pyridine and pyridinium ion?l j by tlie audio-sideband calibration technique, are the results of at least three determinations alternating in sweep direction at sweep rates 01' 0.05 or 0.1 Hz,sec. The amino acids were run as acltteoiis soltilions with water as the internal reference, while Suc., 90, 3622 (1968); (b) T. Tokuhiro miti G . Fracnkcl, J . Atuer. Chem. Soc.., 91, 5005 (1969). (19) W. Adam, A. Grimsoii, and G . Rodriguez, J . Cheni. Ph>.s., 50, fi45 .~( l9fi91. (20) For a general discussion, see M. Barfield and D. M . Grant, ~
Adc.a/i. !Magti. Resotmire, I ,
149 (1965).
(21) (21) A. J. R. Bourn anti E. W. Randall, Mol. Ph.is., 8, 567 (1964); (I?) E . W . Randall a n d D. Ci. Gillics, i n ref 12, Vol. 6, i n pi'sss. ( 2 2 ) H . H . H a l t , "Organic Synthescs," Collect. Vol. 11, Wiley, New YorL, h. Y., 1943, p 211. (23) D. b l , Lenial, F. Mengcr, and E. Coats, J . A m e r . Chem. Soc., 86, 2395 (1964).
(24) Pcrdi.uteriomc.th~iiiol (containing some rcsidual protiated niatcrial) \\:IS uscd instcad of ordinar) iiicthanol bccause thc resonance of the cxcha~igcahlcprotons i n thc latter solwilt obscured the lower field rcgioii of interest.
Jourml o j IIIC Aturricun Cliet?iical Societj. 1 93:20
Figure 2. Pmr spectrum of pyridine-I5N, H,. As in Figure 1. the upper and lower traces are the experimental and calculated spectra, respectively.
N-iiitrosodimethylainitie and 1,l-dimethylhydrazine were neat liquids containing a trace of TMS. Carbon-13 and nitrogen-15 spectra were taken o n a Varian DFS-60 spectrometer operating at 15.08 (l:IC) and 6.07 (15N) M H z under conditions of complete proton decottpling.25~zGAt sweep rates of 2-4 Hzjsec, carbon line positions were measured with respect t o the carbon resonances
of cyclopentaneZi and methanol, as above, while nitrogen shifts were measured with respect t o external 15N-enri~hednitric acid. Because of long-term field instabilities, it was not possible t o obtain high-resolution spectra without proton decoupling.
Results Proton Spectra. The spectra of neat pyridine and pyridinium ion are shown in Figures 1-4, together with calculated spectra resulting from iterative fitting of the experimental line positions, using the program LAOCN3."' The best values of the coupling constants obtained for these systems, a s well as those for the 30% (v/v) niethanol solution, are given i n Table I t . At worst, the calculated probable errors were not greater than 0.05 H z , and most were less than 0.02 H z . Visual coniparisons of calculated and experimental spectra are satisfactory except for the high-field region of the pyridinium ion spectrum (Figure 4). Although calculated line positions agree well with experimental values, line intensities are not always reproduced. The appearance of the calculated spectrum was markedly insensitive to detailed assignments of experimental t o calculated frequencies, and to the signs of the N-H coupling constant^.?^ The reliability of values obtained from this tightly coupled part of the spectrum is thus questionable;'hb nevertheless, the protowproton coupling constants agree well with published values,"' (25) (a) F. J. Weigert and .I.D. Roberts, J . Amer. Chem. Soc., 89, 2967 (1967); (b) F. J. Weigert and J. D. Roberts, ibid., 90, 3543 (1968). (26) F. J. Wcigert, M . Jautelat, and J . D. Roberts, Proc. Nor. Acad. Sci, U . S.,60, 1152 (1968). (27) External cyclohexane was used as standard for the amino acids. (28) (a) A. A. Bothner-By and S. M. Castellano i n "Computer Programs for Chemistry," Vol. 1, D. F. DeTar, E$., w. A. Benjamin, New York, N. Y . , 1968, p I O ; (b) see R. J. Abraham and S . M. Castellano ( J . Chem. SOC.B, 49 (1970)) for a discussion of uncertainties in spectral analysis. (29) The signs of the proton-proton couplings were gencrally kept in accord with those iii thc Iiterature.l0" Rcversing the sign of Jia was found to have no efect on the appearance of the spectruiii.
October 6 , 1971
5221
5 Hz
Figure 4. Pmr spectrum of pyridinium-15N hydrochloride in methanol, Hn. The labels have the same significance as in Figure 3.
Figure 3. Pmr spectrum of pyridinium-16N hydrochloride in methanol, H I a n d Ha:(a) experimental spectrum; (b) calculated spectrum based on parameters in Table 11; (c) calculated spectrum as in b but with positive.
and except for JI4, the nitrogen-proton couplings are in accord with those observed with trifluoroacetic acid as solvent.gc In none of the cases is the iterative analysis capable of unambiguous determination of the signs of the 15N-H coupling constants. These have been assigned in analogy to results of spin-tickling experiments with quinoline- 15N.8,9c
Figure 5. Natural-abundance 13C spectrum of pyridinium-16N hydrochloride in methanol. The spectrum was obtained with complete proton decoupling after 23 scans at a sweep rate of 5 Hzisec.
,
1Hz
,
Table 11. Proton-Proton a n d Proton-Nitrogen Coupling Constantsa in Pyridine-16N H, H8
H6&HJ H, Hz
M
N,
Figure 6. Natural-abundance 13C resonance of C-2 in pyridine-IbN. The signal was obtained with complete proton decoupling after a single scan a t a sweep rate of 0.1 Hz/sec.
H 6 h H J HJ
Hb
H,
N,
I
computer program sNARE~O revealed a doublet separated by ~ J N C= 0.45 f 0.1 Hz. This value was confirmed when the sweep rate was reduced to 0.1 Hz/sec (Figure 6 ) . Table 111 summarizes the coupling
H J2 3
JZ
4
Jz
5
J2 6 J3 4 J3 j
J12 J13 J14
4.88 1.83 0.97 -0.12 7.62 1.34 -10.76 -1.53 10.21
4.97 1.81
5.94
0.90
0.67
-0.16 7.83 1.38 -10.06 -1.56 *O. 18
1.51 -0.55 7.93 1.64 -3.01 (-2.2)d -3.98 ( - 4 . 3 ) d f 0 . 6 9 (0.3)d
I n hertz. Neat. 30% (v/v) solutjon in methanol. ence 9c, trifluoroacetic acid solvent.
Table 111. Carbon-Nitrogen Coupling Constants in Pyridine-l6Na
Refer-
I 'JuI Carbon Spectra. With the initial exception of that derived from C - 2 in neat pyridine, all 13C-15N coupling constants were directly measurable from the protondecoupled 13C spectra (Figure 5 ) . In the case of C - 2 , a sweep rate of 1 Hz/sec allowed observation of only a singlet, but application of the sensitivity-enhancement
0.45
2.4 3.6
12J~aI
I
'JibI
In hertz, k 0 . 1
Hz.
* Neat.
0.7 2.6 3.8
12.0 2.1 5.3
3 0 z (v/v) in methanol.
(30) (a) G. A. Peterson, Ph.D. Dissertation, California Institute of Technology, 1970; (b) G. A. Peterson, R. L. Lichter, and J. D. Roberts, manuscript in preparation.
Lichter, Roberts
Nuclear Magnetic Resonance Spectroscopy, Pyridine-15N
5222
constants for all pyridine systems; the chemical shifts agreed with those previously reported, with minor differences attributable to differences in measurement conditions. 15N Spectra. The nitrogen chemical shifts obtained directly from the proton-decoupled 15N spectra (Table 1V) agree with published valuesUg The proton-deTable IV. Nitrogen Chemical Shifts of Pyridine-16N 8 ~ ppma , Pyridine, neat Pyridine-methanol Pyridinium hydrochloride
56.8 74.4 170.1
isotropy of the N-nitroso group,31 but is more likely to arise from electric-field effects, as have been suggested to explain the 5-ppm difference in the methyl carbon chemical shifts of dimethylformamide. 3 2
Discussion Comparison of the proton-nitrogen coupling constants of pyridine and quinoline with those of the oximes and hydrazones (Table I) shows the same kind of correlation between lone-pair orientation and coupling exists for these types of compounds. The relationships are emphasized by considering structures 3-8. Thus, the coupling pathways in pyridine and quinoline (3)
Upfield with respect t o external H 1 5 N 0 3 ; estimated error, 0.2 PPm.
coupled signals derived from both neat and methanolic pyridine were not inverted. Because the nuclear Overhauser effect on 16N resonances on proton irradiation normally produces inverted 15N signals,la the absence of inversion implies a negligible nuclear Overhauser effect in these systems. Apparently, the exchange lifetime of the hydrogen-bonded proton in the pyridinemethanol complex is too short to provide an effective relaxation pathway for the 15Nnucleus.la Other 15N-13C Couplings. Table V shows the measured coupling constants for a few nonaromatic 15N-containing compounds. The 15N splittings in Table V. Carbon-Nitrogen Coupling Constants in Some Nonaromatic Compounds 'JNc, H z H3N+CHzCOOpH 6.33 p H 13.68 H3N+C(CHa)HCOOp H 6.37 p H 13.64 N-Nitrosodimethylamine (2)*
6.2 4.8 5.6 4.4
l,l-Dimethylhydra~ine-2-~~N
'JNc, Hz
0 0 0 00 7 . 5 (anti) 1 . 4 (syn)
