COMMUNICA'IIONS TO THE EDITOR 709
are indicative of the vinyl protons of a 4-vinylpyridinium end group. The integration of these absorptions indicates a =of 15-20.
-
\
0
-fCH2CHz-Q?& IVa, X = CF3C0,b, X = HSO4C, X = p.CH3C6H4S03d, X = CI-
To substantiate the structure of IVa we have independently synthesized IVd ( [ q ]0.13 in 0.01 M KBr) by the method of Meisenheimer, et a / . , 8from the intermolecular condensation of 4-(B-~hloroethyl)pyridine. The nnir spectrum of 1Vd was identical with Figure l a except for the diflerent end groups of the two polymers. All the products obtained from the reactions given in Table I were shown by nmr t o contain varying degrees of monomeric 4-vinylpyridinium salts (Ha-c) and ionene polymers (IVa-c). There were n o indications of poly(4-vinylpyridinium) salts under the conditions employed. The structure of the polymers resulting from reaction of 4-vinylpyridine with strong acids supports the proposed initiation by reaction 2, since the propagation step must be similar to reaction 2 in order to obtain pyridinium units in the backbone. Propagation could occur by the addition of a neutral pyridine t o the double bond of a 4-vinylpyridinium ion to give a zwitterion, followed by proton transfer to regenerate a neutral pyridine. Since the growing polymer probably has a vinylpyridinium group and a neutral pyridine end group it might be expected to have some of the kinetic features of a condensation polymerization. As further evidence for the proposed initiation and hydrogen transfer, the dimeric model compound V was prepared by the reaction of 4-vinylpyridine with pyridine hydrochloride in pyridine. The method employed was similar t o that of Cislak and Sutherlandg for the related compound from 2-vinylpyridine. The structure of V was
-c1-
Figure 1, Nmr spectra of (a) poly( 1,4-pyridiniumdiylethylene trifluoroacetate) (IVa) in D?O; (b) poly(4-vinylpyridine) in DC1-DIO. The sharp singlets cu. 60 and 120 cps from the H20peaks are spinning side bands. The absorptions of the vinyl protons of this salt in DrO showed the expected ABX pattern at 6 7.02, 6.42, and 5.98. Preliminary results indicate that IIa is also stable in nitromethane solution for several days. However, when pyridine or 4-vinylpyridine is added to the solution, ionene formation occurs. We hope to report in the near future further studies on the polymerization of 4-vinylpyridinium salts.
Acknowledgment. The author are grateful t o Professor C. G. Overberger for his support and encouragement of this work.
* To whom inquiries niay be addressed at the Department of Chemistry, Lowell Technological Institute, Lowell, Mass. 01854. J. C. Salamone,* B. Snider, W. L. Fitch Mucromoleculur Research Ceriter Imtitute of'Science and Technology The Utiicersity of' Michigarz Ann Arbor, Micliigatz 48105 Receiced July 7, 1970
V
confirmed by the elemental analysis and the nmr spectrum of its hydrochloride salt. The nmr absorptions for the methylene protons of the hydrochloride of V and of the ionenes (IV) were identical. We have previously suggested that stable 4-vinylpyridinium salts could be prepared if quaternization were sufficiently rapid, if the resulting counterion were a weak nucleophile and if the salt were not exposed to excess 4-vinyIp~ridine.~The strong acids employed in this investigation were expected t o fulfill these requirements. Indeed, by the addition of 4-vinylpyridine t o excess trifluoroacetic acid in dry ether, crystalline 4vinylpyridinium trifluoroacetate (IIa), m p 178-1 79', could be obtained in nearly quantitative yield. The nmr spectrum of IIa in D u oor trifluoroacetic acid showed n o evidence of polymerization, even after several days. (8) J. Mciscnheimer, J. Neresheimer, 0. Finn, and W. Schneider, JustusLiebigs Anri. Chem., 420, 190 (1920). (9) E. Cisldk iind L . Suthcrland, U S .Patent 2,512,789 (1950).
Intermolecular Interactions in Crystalline Poly(viny1 chloride) Intermolecular forces between polymer chains can be studied by their effects on the vibrational spectrum of the crystalline polymer.' These effects are generally of two kinds: a splitting of bands associated with internal modes of vibration, and the appearance of lowfrequency absorption bands which result directly from intermolecular interactions. Both of these effects are observed, for example, in the spectrum of polyethylene, where it is well established that the intermolecular interaction arises primarily from dispersion forces between the chains.' Such intermolecular interactions have not as yet been demonstrated spectroscopically in crystalline syndiotactic poly(viny1 chloride). A search for splitting of ( I ) S . Krimm, Fortschr. Hochpolj,m.-Forsch.,2 , 5 1 (1960). ( 2 ) M. Tasumi and S. Krimm, J . Chem. Phys., 46, 755 (1967).
710 COMMUNICATIONS TO THE EDITOR
Mucrornalecules
TABLE I
LOW-FK~QUENCY ABSORPTION BAVD OF CRYSTALLINE POLY(VISYL CHLORIDE) -Frequency, Polymer
-.
(CH2CHCl)n (CD2CHCIh (CH2CDCl)n (CD2CDCl)n
Room temp 64.1 64.0
cm-'-Liquid N? temp
63.1
67.1 67.1 65.8
61.5
64.3
modes is not favorable, since both components of such doublets are not active in a given spectrum (infrared or Raman).' Evidence which we have obtained from far-infrared spectroscopy, however, does indicate the presence of such a n interaction. We wish to show furthermore that it is in part different from the dispersion type of interaction characteristic of polyethylene. It has been shown3 that a band occurs near 67 cm-I in the infrared spectra of liquid secondary chlorides which can be attributed t o a localized C-H . . .CI intermolecular vibration. A study was made of the absorption of poly(viny1 chloride) in this region. For a commercial sample, no absorption band was found. For a sample polymerized at low temperature or for one polymerized in urea complex an absorption band was observed, of relatively greater intensity in the latter case than in the former. The frequencies of this band for such urea complex polymers and their deuterated derivatives are given in Table I. The samples were in the form of disks (pressed from the powdered polymer at about 5000 Ibs), and spectra were obtained at room and at liquid nitrogen temperatures on a Perkin-Elmer 301 spectrometer. The absence of a n absorption band in the spectrum of a commercial sample of poly(viny1 chloride), which is known to be poorly crystalline in comparison with a urea complex polymer,4 indicates that the -67-cm-' band is to be associated with a vibrational mode in the crystalline phase This is not likely to be a pure translational lattice mode, for several reasons First, such modes are forbidden in the infrared spectrum by symmetry.' Second, the frequency shift on deuteration is inconsistent with this type of motion For the singly, doubly, and triply deuterated molecules in Table I the observed ratios of hydrogen to deuterium frequencies (in the liquid Ne temperature spectra) are 1.020, 1.000, and 1.044, respectively, whereas the predicted ratios for a translational mode would be 1.008, 1.016, and 1.024. Third, a band is observed in a number of (3) A. V. R. Warrier and S. Krimm, J . Chem. Phjs., 52, 4316 ( 1970).
(4) S. Krimm, A . R. Berens, V. L. Folt, and J. J. Shipman, Chem. Ind. (London), 1512 (1958); 433 (1959).
liquid secondary chlorides at this poisition.a Since in the latter case the mode cannot be strictly a n intramolecular vibration,3 we suppose that in crystalline syndiotactic poly(viny1 chloride) this band arises from a n intermolecular interaction The suggestion3 that the -67-cm-1 band arises from a localized vibrational mode involving a C-H. . .C1 interaction analogous to a hydrogen bond is supported by the present results Such an interaction is feasible in the structure of crystalline syndiotactic poly(viny1 chloride), which places the (C1)C-H bond of one molecule essentially colinear with a C1-C bond of the second molecule in the unit cell.; There is, furthermore, spectroscopic evidence6 for the possibility of forming such weak hydrogen bonds. That the vibrational mode is localized is indicated by the deuteration results. Replacement of the two hydrogens on the CH2 group has no effect on the frequency, whereas replacement of the CHCl hydrogen drops the frequency by a factor of 1.020 (compared to a predicted value of 1.027 if the vibration were restricted to the C--H. . .C1 triatomic unit). The further drop in frequency for the fully deuterated molecule is probably a result of a change in the detailed form of the mode, the magnitude of the ratio (riz. 1.044) being a further indication that a translational lattice mode is not involved. (An assignment t o the allowed' rotatory lattice mode is not indicated, despite the ratio of 1.042 for the moments of inertia of (CDeCDCI), and (CH2CHCl),: such an assignment would predict a large frequency change for (CDyCHCI), and a small change for (CH2CDC1),, whereas in fact the opposite is true.) These and previous studies3 strongly suggest the presence of an intermolecular C-H . .CI interaction analogous to a hydrogen bond in crystalline syndiotactic poly(viny1 chloride). This lends support to the suggestion' that such a bond may be responsible for similar intramolecular interactions in 2,4-dichloropentane. +
Acknowledgment. This research was supported by a grant from the National Science Foundation. We are indebted to V. L. Folt of the B. F. Goodrich Co. for providing us with the polymer samples. ( 5 ) G. Natta and P. Corradini, J . Polym. Sci., 20,25 1 (1956). (6) A . Allerhand and P. von R. Schleyer, J . Amer. Chem. Sac., 85, 1715 (1963). (7) S. Krimm, Pure A p p l . Chem., 16,369 (1968). * To whom correspondence should be addressed.
A. V. R. Warrier, S. Krimm*
Harrisori M . Ratidall Laboratory of Physics Upiicersity of Michigan, Ant? Arbor, Michigarr Receiced April 2, 1970
