Dewar avtd Lucken Chemical Appiications of 77. Chemical Applications of Nuclear Quadrupole Resona nce 8pectro-scopy. Part II? Chloro-derivatives of Maleic Anhydrides Thiophen and Anilinium Salts. By M. J. S. DEWARand E. A. C. LUCKEN. Measurements of chlorine nuclear quadrupole resonance frequencies are reported for p-chloroanilinium chloride and bromide for mono-and di- chloromaleic anhydrides and for 2-chloro- 2 :5-dichloro and 2 :3 4 5-tetrachlorothiophen. The results are discussed in terms of hydrogen bonding in the P-chloroanilinium salts carbon-chlorine n-bonding in the maleic anhydrides and inductive effects in thiophen due to an increased electro- negativity of sulphur brought about by promotion of a 3p electron to a 3d orbital in forming the appropriate valence state.INPart I we described a study of inductive and conjugative effects in chloro-derivatives of nitrogen heterocycles by nuclear quadrupole resonance spectroscopy; here we describe some further studies of compounds containing chlorine attached to unsaturated carbon atoms. EXPERIMENTAL The chlorine resonances were observed as described in Part 1.l Chlorothiophens were prepared by chlorination of thiophen and separated by fractional distillation. p-Chloro-anilinium chloride and bromide were prepared from p-chloroaniline and the corresponding acid and recrystallised from dilute hydrochloric or hydrobromic acid. Monochloromaleic anhydride (a commercial specimen) and dichloromaleic anhydride (supplied by Imperial Chemical Indus- tries Limited) were used without further purification.The observed frequencies are shown in the Table. Observed nuclear quadrupole resonance frequencies. Frequency (Mc./sec.) Compound 86" K 195"K Room temp. p-Chloroanilinium chloride ............... 35.448 35-201 34-939 (297") p-Chloroanilinium bromide ............... 34.752 34.636 34.440 (297") Chloromaleic anhydride .................. 37.159 36.750 36.283 (297") Dichloromaleic anhydride ............... 37.945 38-01 3 37.565 37.593 37.049 37.065 (297") 2-Chlorothiophen ........................... A seriesof medium intensity lines in the region 34-39 Mc/sec. 2 :5-Dichlorothiophen ..................... 36-669 36.875 36.696 37-046 36.840 37.078 2 3 4 5-Tetrachlorothiophen ...... 37.517 38-500 36.906 36.931 37.740 (294") Part I Dewar and Lucken J.1958 2653. * Conrad Hartough and Johnson. J. Amer. Chem. Soc. 1948 78 2564. [19591 Nuclear Quadrupole Resonance Spectrosco$y. Part 11. 427 DISCUSSION It is necessary to decide to what extent it is legitimate to ascribe to purely electronic effects the differences in nuclear quadrupole resonance frequencies between different molecules when it is known that electrostatic crystalline fields can cause an unpredictable lowering of frequency of up to several megacycleslsec. from that of the molecule in the gas phase.3 The number of molecules whose nuclear quadrupole coupling frequencies have been measured by microwave spectroscopy as well as by direct nuclear quadrupole resonance is unfortunately so small that no general conclusions can be drawn.It is therefore profitable to consider those compounds in which multiple resonances occur and in which no extra- ordinary factor such as hydrogen-bonding can account for this multiplicity. Results for compounds as varied as the Group IV tetrachlorides and the polychlorobenzenes show that splittings of up to 0.6 Mc./sec. are possible although in most cases they are not greater than 0.3 Mc./sec. A further indication of magnitude of solid-state effects can be obtained from the correlation between the Hammett function and the nuclear quadrupole resonance frequencies of substituted chlorobenzenes where the standard deviation between observed resonance frequencies and those predicted by the linear relationship derived by Bray and Barnes is 0.36 Mc./sec.These facts in conjunction with the success of pure quadrupole resonance measurements in demonstrating inductive effects in the alkyl halide^,^ indicate that a difference of more than 0.5 Mc./sec. between the chlorine nuclear quadrupole frequencies of two molecules is very probably due to electronic effects. This " limit of confidence " is considerably reduced when comparisons are being made between two series of different molecules. The chlorine nuclear quadrupole resonance frequencies for the two salts of p-chloro-aniline are much higher than that for the free base (34.146 Mc./sec. at 77"K). Protonation of the amino-group can have several consequencies (a) protonation will destroy the conjugative (-E) effect of the nitrogen which normally leads to an increased x-electron density at the para-carbon atom; (b)protoilation will increase the inductive (+I) effect of nitrogen and the resulting polarisation of the x-electrons will decrease the charge density at the para-carbon atom ; (c) the increased electronegativity of nitrogen on protonation will by a relayed (inductoelectromeric 8) effect increase the degree of x-bonding between chlorine and the ring.Now effects (a) and (b) will lead to a decrease in negative charge at the para-carbon atom and so to a decrease in a-bond polarity of the C-C1 bond and a consequent increase in nuclear quadrupole resonance frequency; effect (c) will lower 1 this frequency. The observed frequencies imply that the former effects predominate.It is interesting that the frequencies for the two salts differ markedly that for the hydrochloride being much the greater. This difference is most probably due to differences in hydrogen-bonding between the ammonium groups and halide anions in the salts. Such hydrogen-bonding will reduce the effective positive charge at the nitrogen atom and so lower its electronegativity; the argument of the previous section implies that there should be a corresponding increase in polarity of the C-Cl a-bond and a corresponding decrease in x-bonding between chlorine and the ring. Since the former effect will raise and the latter lower the quadrupole resonance frequency one cannot be sure what the net effect will be. One cannot tell from the observed frequencies whether hydrogen-bonding is greater in 9-chloroanilinium chloride or in the bromide-only that there must be a difference.We are investigating this problem further for such measurements offer a most interesting method for studying hydrogen bonds in such salts. Allen J. Phys. Chem. 1953 57 502. Livingstone J. Phys. Chem. 1963 67 496. Bray Barnes and Bersohn J. Chem. Phys. 1956 25 813. Bray and Barnes .I. Chem. Phys. 1957 27 551. ' Heal J. Amer. Chem. SOC.,1952 74 6121. 8 Brown and Dewar J. 1953 2406. 428 Nuclear Quadrufiole Reso fiance Sfiectroscopy. Part II. The frequencies of the mono- and di-chloromaleic anhydrides can be predicted approxim- ately as follows. The frequencies at 20"~ for vinyl chlorideg (33.6 Mc./sec.) and for cis-dichloroethylene * (35.0Mc./sec.) are known; the frequencies at 87"K will be about 0-2Mc./sec.lower. The effect of the a-carbonyl group in the anhydride can be estimated approximately from the difference (2.3Mc./sec.) between the nuclear quadrupole resonance frequencies for chloroacetic acid and methyl chloride; * and that of the p-carbonyl group from the corresponding difference (1.8 Mc./sec.) between o-chlorobenzoic acids and chloro- benzene.6 Together these comparisons suggest that the nuclear quadruple resonance frequency for a chloromaleic anhydride should be 2.3 + 1.8 =4.1 Mc./sec. greater than that for the corresponding chloroethylene. The predicted frequencies for mono- and di-chloromaleic anhydride at 86"K are therefore 37.6 Mc./sec. and 38.9 Mc./sec.respectively values which are greater than those observed (37.2 and 38.0 Mc./sec. respectively). These differences suggest that C-CI x-bonding is greater in the anhydrides for such x-bonding lowers the frequency; this is what one might expect in view of our work on chloro-derivatives of heterocycles for the more electrophilic character of the C% bond in maleic anhydride should lead to an increase in x-bonding to chloiine. The frequencies of the thiophen compounds are interesting in that they seem to be much higher than those for the corresponding benzene derivatives (chlorobenzene 34.622 Mc./sec. ; p-dichlorobenzene 34.779 Mc./sec. ; 1 :2 3 :4-tetrachlorobenzene 37.557 37.455 37.013 Mc./sec.).6 This is surprising for sulphur and carbon are thought to have closely similar electronegativities (2-5 on the Pauling scale).The difference cannot reasonably be ascribed to decreased x-bonding in the chlorothiophens for since vinyl chloride has a lower frequency (33.4Mc./sec.) than chlorobenzene any decrease in aromatic character in thiophen as compared with benzene should if anything lower the nuclear quadruple resonance frequency. The electronic structure of thiophen has been discussed by Longuet-Higgins,lO whose conclusions may be briefly summarised The x-molecular orbitals of benzene can be written as combinations of six ethylenic x-molecular orbitals three bonding and three antibonding whose symmetries are indicated in (I) and (11); the resulting molecular orbitals are identical with those given by the usual LCAO approach.Likewise the x-molecular orbitals of thiophen can be constructed from a set of six orbitals two pairs of ethylenic x-molecular orbitals and a pair of sulphur orbitals 3p (111)and 3d (IV). It is obvious from this diagram that the sulphur atomic orbitals resemble the ethylenic x-molecular orbitals in symmetry; they also have similar dimensions and binding energies. This representation therefore explains very clearly the close resemblance between benzene th ophen. As a result the x-electrons are uniformly shared between the six orbitals SO that the mean x-electron charge density is unity at each carbon atom and two at sulphur. This argument suggests that any differences between benzene and thiophen in x-electron distribution at the carbon atoms are small; the differences in nuclear quadrupole resonance frequencies for the chloro-derivatives cannot therefore be explained in terms of x-electron distributions.However this analysis requires the sulphur atom in thiophen to be in an excited valence state with one 3p electron promoted into a 3d orbital. Since the screening effect of 3d electrons is small the effective electronegativity of sulphur as regards the 3s and 3p electrons a Livingstone J. Chein. Phys. 1951 19 1613. lo Longuet-Higgins Trans. Faraday SOC.,1949 45 173. [1959] Peroxides of Elements other than Carbon. Part III. will be increased; indeed Slater’s rules imply an increase in effective nuclear charge of 0.35 unit for the 3s and 3p electrons. This will lead to a strong inductive polarisation (C +S) of the C-S 0-bond formed by a 3s-3p orbital and this by a relayed inductive effect will lead a decreased 0-bond polarity (and increased nuclear quadruple resonance frequency) for the C-Cl bonds in the chlorothiophens.A similar change in electronegativity must occur in other compounds where sulphur atoms use d orbitals to form bonds; the very strong inductive effect of the sulphonyl group may be due largely to this. Since the publication of Part I our attention has been drawn to the paper by Kojima et aZ.ll in which piezoelectric resonances in a variety of inorganic compounds were reported. The radiofrequency spectra here observed consisted of a series of sharp resonances occurring over a considerable frequency range-some tens of megacycles-and in which one or more series obeying an interval-rule could be detected.It is probable that the multiple resonances of 2 :4 :6-trichloropyrimidine and of 2-chlorothiophen should be ascribed to this cause rather than to any randomness in the orientation of molecules in the crystal. We thank the Royal Society and the University of London for grants for the purchase of apparatus and Imperial Chemical Industries Limited for a loan of dichloromaleic anhydride. One of us (E. A. C. L.) thanks the Department of Scientific and Industrial Research for a maintenance grant. QUEEN MARY COLLEGE UNIVERSITY OF LONDON, LONDON, E.1. [Received May 21st 1968.1 l1 Kojima Tsukada Ogawa and Shimauchi J. Phys. Soc. Japan 1955 10,265.