J. CHEM. SOC. DALTON TRANS. 1987 1593 Nitrogen-I 5 and Oxygen-I7 Nuclear Magnetic Resonance Chemical Shifts and 15N-29Si, 13C-29Si, and I5N-l H Coupling Constants in Cyclosilazoxanes Eriks KupCe, Edvards Liepigs, and Edmunds Lukevics Institute of Organic Synthesis, Latvian S.S. R. Academy of Sciences, Riga, U.S.S.R. Boris Astapov Institute of Organo- Element Compounds, U.S.S.R. Academy of Sciences, Moscow, U.S.S.R. The effects of ring size on 15N, 170, and 29Si n.m.r. chemical shifts and 15N-29Si, 13C-2gSi, l5N-'H, and 29Si-29Si coupling constants in cyclosilazoxanes have been studied. The opposite shielding sequence in 8(15N) and 6(29Si) and the decrease in 15N-29Si coupling with diminishing ring size are consistent with higher Si-N bond ionicity (reduced d,-p, interaction) in smaller rings.The 170 chemical shift is practically independent of ring size in cyclosiloxanes. It has been proposed that the factors governing the 170 chemical shift act in opposite directions and are balanced. A linear relationship between N-H bond stretching frequency and 15N n.m.r. parameters (15N-lH coupling and 15N chemical shifts) has been established. Although a large amount of data is available on 'H, 13C, and 29Si n.m.r. chemical shifts ( 6 ) of cyclosiloxanes, cyclosilazanes, and cycl~silazoxanes,~-~ data have not been reported for the "N and "0 chemical shifts or the one-bond coupling constants ('4 1sN-29Si, 13C-29Si, and 15N-'H for these types of compounds. I t has been that 6(29Si) reflects variations in d,-p, interaction along Si-X (X = 0 or N). In this respect measurement of 6( 5N), 6(' 70), and especially ' J ( 5N-29Si) can give a more definite result.In this paper we report the 15N and 1 7 0 chemical shifts and the 5N-29Si, 3C-29Si, and "N-IH coupling constants, together with some two-bond 29Si-0-29Si couplings, in six- to twelve-membered cyclosilazoxanes and in some acyclic derivatives, selected partly on the basis of their availability. It has been reported that a linear relationship exists between the NH stretching frequency (vNH) and the "N n.m.r. parameters [6( 'N) and ' J ( "N-IH)] in saturated primary amines and anilines. In order to examine the existence of the analogous relationships in cyclosilazoxanes and to compare the factors influencing i.r. and n.m.r. parameters in organic amines with those in cyclosilazoxanes, we have also measured vNH.Results and Discussion 5N und 7O Chemical Shifis.--It appears from Table 1 that 6(I5N) and 6("O) are less sensitive to the nature of the substituents at the silicon atom than 6(29Si). Nevertheless, the influence over the two bonds (p effect) is important. Data obtained for the acyclic compounds (1)-(5) indicate that 15N and ' 7O shielding decreases with increase in electronegativity of the substituents at adjacent silicon atoms. For compound (5) this allows us to observe the 170 signal of the OSiMe, end- group separately. In cyclosilazoxane rings of equal size, substitution of one nitrogen atom by oxygen decreases the shielding of the adjacent I5N by an average value of 2 p.p.m. A similar tendency is characteristic of 7O n.m.r.spectra. The influence of ring size on 6(15N) is opposite to that of 6(29Si). In the cyclosilazoxanes studied 6(29Si) moves to higher field with increasing ring size, which is in compliance with previous investigations.'-3 It has been suggested that this is the result of an increase in the d,-p, interaction along the Si-X (X = 0 or NH) bond. The tendency of 6("N) to move downfield with increase in ring size may be understood in (CIMezSi ),X [Me,Si(OSiMe,), 1,NH ( Me3Si ),X ( 1 ) X=NH ( 3 ) X = NH ( 5 ) ( 2 ) x = o ( 4 ) x = o Me2 X I I Me2 X /si\x I I Ph2 o/si\o I I ( 6 ) X = NH (8) X=NH,Y=O ( 7 ) x = 0 ( 9 ) X = 0, Y =NH Me2Si - X - SiMe, I I i I X X H (10) H MqSi- N -SiMe2 I I 0 0 I I MezSi- X -SiMez MezSi- X -Siblet (11) X = NH (12) x = 0 (13) X = NH (14) x = 0 Me2 Me, I I I I I I I I 0- Si - Y - SiMe, MezSi -0-Si-NH SiMe, 0 HN Me, Si X I I MezSi 0- Si - N- SiMe, Me, H 0- Si - N - SiMe, Me, (15) X = NH, Y = O (16) X= 0, Y = NH (17) similar terms.Thus, the opposite shielding sequence in the nitrogen and silicon resonances reflects an increase in Si-N bond ionicity (decrease in d,-p, interaction) in smaller rings. Unexpectedly, 6( I 70) is practically independent of ring size in 'pure' cyclosiloxanes. This indicates that the factors governing 6( 70) act in opposite directions and are balanced.1594 J. CHEM. SOC. DALTON TRANS. 1987 Table 1. N.m.r. parameters and NH stretching frequencies of cyclosilazoxanes and acyclic analogues " Compound 6( "N) (1) - 354.2 (2) -334.3 (3) (4) (5) -335.7 (6) - 347.3 (7) (8) - 345.4 - - - (9) - 342.9 (15) - 336.7 (17) - 334.8 ~ ( 1 7 0 ) - 39.0 70.0 71.4' 72.0 60.1 63.0 56.0 71.4 68.1 70.9 72.3 70.5 - - - 69.6 67.3 6(29Si) 2.I 6.5 13.3 6.9 - 11.9 ( x ) ~ . ~ - 4.6 - 8.7 -4.0 (NN) -4.3 (ON) -3.4 (ON) -9.0 (00)' - 1.2 (ON) - 8.2 - 19.4 - 11.3 - 10.9 (ON) - 19.8 (00) -8.3 (NN) - 12.6 (ON) - 22.6 (00) - 12.6 (NN) - 13.2 (ON) -23.6 (00)' - 12.3 'J( 'sN-29Si) lJ('3C-29Si) 13.4 56.3 59.5 14.4 68.1 72.6 17.4 69.4 15.4 63.0 74.2 15.8 62.3 15.8 68.5 15.5 68.4 74.0 15.7 68.2 16.9 64.0 75.4 15.9 69.8 15.9 69.7 75.4 17.7 64.5 17.7 69.4 74.9 17.3 69.3 17.1 69.6 74.7 17.2 69.2 - - - - ' J ( "N-' H) 69.6 68.1 - - - 69.4 69.5 69.5 69.6 67.0 67.1 67.3 66.3 - - 66.7 65.9 VNn 3 402 3 364 3 376 3 403 3 402 3 400 3 396 3 385 3 383 3 384 3 377 - - - - 3 380 3 373 " ' J in Hz; 6 in p.p.m.relative to CH,N02 ("N), H 2 0 ("O), or SiMe, (29Si); vNH in cm-'. * Ref. 9. (2-p), 3.1 (p-y), 2.2 (y-6), 2.5 ( 6 - ~ ) , 1.3 (E-<) Hz. 2J(29Si-29Si) = 1.02 Hz. ' zJ(29Si-29Si) = 1.50 Hz. 56.1 p.p.m. (OSiMe,). 2J(29Si-29Si) = 2.1 2J(29Si-29Si) = 1.54 Hz. HN-SiaMe,OSieMe2OSiYMe2OSi6Mez0Si"Me2OSi~Me3; -22.3 (p), -22.4 (y), - 22.5 (6), -21.6 (E), + 7.1 p.p.m. (6). 74.4 (p), 74.4 (y), 74.3 (S), 74.3 ( E ) Hz. Table 2. Si-X-Si (X = 0 or N) valence angles (") in somc cyclosiloxanes, cyclosilazanes, cyclosilazoxanes, and acyclic analogues Compound (Me,SiO), (Me,SiO), (Me,Si),O (Me,SiNH), (Me,SiNH), (Me,Si),NH (Ph,SiO),( Ph,SiNEt) (PhMeSiO),(NH),- (PhMeSiO), Si-X-Si 131.6 144.8 148.8 126.8 132/ 131.3 126.3 130.0 Method" Ref.ED h ED h X,ED c * , d ED 6' X R ED h X i X i " ED = electron diffraction, X = X-ray analysis. H. Oberhammer, W. Zeil, and G. Fogaresi, J. Moi. Sfrucr., 1973, 18, 309. ' M. J. Barrow, E. A. V. Ebsworth, and M. M. Harding, Actu Crj*stullogr., Sect. B, 1979, 35, 2093. B. Csakovari, Z. Wanger, P. Gomory, F. C. MijlholT, B. Roszandai, and I. Hargittai, J. Orgunomet. Cliem., 1976, 107, 287. B. Roszondai, I. Hargittai, A. Golubinski, L. V. Vilkov, and V. S. Mastryukov, J. Mol. Struct., 1975,28,339. Mean value. G. Smith and L. E. Alexander, Actu Crj~.srullogr., 1963, 16, 1015. T. Fjeldberg, J . Mol. Struct., 1984, 112, 159. W. Fink and P. J. Wheatley, J. Cheni. Soc. A, 1967, 1517. V. E. Shklover, N. G. Bokij, Yu. T. Struchkov, K. A. Andrianov, A. B. Zachernyuk, and E.A. Zhdanova, Zk. Strukt. Khim., 1974, 15. 864. The main contributor to the chemical shift is the variation in the paramagnetic term (oPA) of the shielding of nucleus For a second-row element this is approximated by equation ( l ) , where po is the permeability of free space, pe is the Bohr magneton, and ( Y - ~ ) ~ , , is the radial-expansion term for the 2p electrons; the Q terms express the imbalance of charge in the -350 t 3: -340 P z Y ro VV" 3370 3380 3390 3400 v,,/cm-' Figure 1. Plots of corrected ' 'N chemical shifts [6( "N*)] against NH stretching frequencies. Regression analysis gave the equation S("N*) = 884.5 - 0.361vN,, (number of points, n = 8, correlation co- efficient, r = 0.996) valence shell of A and AE is the mean energy of electronic excitation.As has been demonstrated by CND0/2 calculations,' 6( 'N) in primary aliphatic amines is governed by the AE term, but in aniline derivatives by the ZQ and ( Y - ~ ) ~ , terms. Furthermore, the linear relationships of 6( "N) UPYSUS vNH were observed with a negative slope ( - 1.7601 ) in the former case and a positive one (0.3855) in the latter. Generally no correlation has been found between 6( ' 'N) and vNH in the cyclosilazoxanes studied. This indicates the existence of the factors which influence 6( 5N) but not vNH. Indeed, if the p effect is excluded by introducing the correction of - 2 p.p.m. forJ. CHEM. SOC. DALTON TRANS. 1987 1595 70 69 N I I 68 I 5 -' 67 66 3370 3380 3390 3 4 0 0 vNH /cm" - Figure 2. Plots of 15N-'H coupling constants against NH stretching frequencies.Regression analysis gave the equation 'J(I5N-'H) = 0.1278 vNH - 365.5 ( n = 10, r = 0.994) the substitution of one nitrogen atom by oxygen in the p position, the resultant 'corrected' I5N chemical shifts [6(' 5N*)] show a good correlation with vNH (Figure 1). The slope of the line obtained (-0.3610) is ca. five times less steep than that found in aliphatic amines ( - 1.7601) and is close to that in aniline derivatives (0.3855), but with opposite sign. It seems, therefore, that 6( "N*) in cyclosilazoxanes are also governed by the A€ term, but the contribution of the CQ and ( r - 3 ) 2 p terms is significantly increased, possibly, due to the presence of dn-p, interactions along the Si-N bonds. The lack of correlation between 'non-corrected' 6( "N) and vNH may be explained by significant influence of fJ substitution on the CQ term.It may also be suggested that independence of 6("O) of ring size in cyclosiloxanes (and cyclosilazoxanes) is a consequence of equal contributions of AE on the one hand, and CQ and terms on the other. Coupling Constants.-Coupling constants to 29Si are very sensitive to the electronegativity of the substituents at the silicon atom and usually increase with increasing electronegati~ity.~.'~ Similar dependences exist for ' J ( ' sN-29Si) and 'J(' 3C-29Si) in the acyclic compounds (1)-(5) (Table 1). In cyclosilazoxanes the 1J(29Si-'3C) values for the fragments N-Si-N (62.3-64.5), N-Si-0 (68.2-69.8), and 0-Si-0 (74XL75.4 Hz) lie in a very narrow, non-overlapping and characteristic range, which allows us to use these couplings for analytical purposes.In contrast to ' J ( 13C-29Si), the 'J("N- 24 Si) values are practically independent of the substituents (N or 0) at the silicon atom. This may be understood in terms of Bent's rule,' ' which suggests that major changes in s-orbital overlap proceed in bonds with more electropositive substituents (Si-C). The decrease of ' J ( ' 5N-29Si) with diminishing ring size gives further evidence of higher Si-N bond ionicity in smaller cyclosilazoxane rings. The effect of p substitution on ' J ( I5N-'H) in the compounds studied seems to be negligible. The dependence of the e.vocyclic coupling 'J("N-'H) on ring size is opposite to that of 1J(29Si-'3C), for example 'J(15N-'H) decreases on going to larger cyclosilazoxane rings, whilst 1J(29Si-' %) slightly in- creases.This fact is consistent with the suggested increase in ionicity of Si-X (X = 0 or N) bonds in smaller rings. Alter- 10 Hz H I * Figure 3. An example of observation of 15N-'H coupling in a 'H spectrum via multiple quantum n.m.r. Compound (lo), ca. 20% solu- tion in CDCI,, 5-mm sample tube, sweep width 600 Hz, 200 scans. A signal marked * is due to an incompletely suppressed signal from protons bonded to 14N natively, the observed changes in exocyclic couplings may be explained by changes in appropriate valence angles. A greater sensitivity of 'J("N-'H) to changes in ring size as compared with ' J ( I 3C-29Si) is consistent with the observation" that the valence angles Si-X-Si depend on ring size to a greater extent than X-Si-X.However, it appears from the data present in Tables 1 and 2 that 'J("N-'H) decreases with an expansion of the Si-N-Si valence angle. This contradicts the situation normally observed for organic amines.' 3p1s We have found a linear relationship between ' J ( I5N-'H) and vNH (Figure 2). As has been demonstrated by CND0/2 calculations,6 such a relationship exists because both 'J( "N-' H) and vNH depend primarily on the s character of the N-H bond. It should be noted that simultaneous decrease of 'J("N-'H) and vNH may be explained by an increase in the N-H bond length. We have also measured some two-bond 29Si-29Si couplings (Table 1). It appears that 2J(29Si-29Si) decreases with decreasing ring size, but the information is as yet insufficient.Unfortunately, the measurement of 2J(29Si-29Si) in cyclo- silazoxanes is difficult to perform because the values of these couplings are small ( 1 .&3.1 Hz). Further complications arise from broadening of the 29Si signals of the silicon atoms bonded to nitrogen due to the fast quadrupolar relaxation of I4N nuclei. Nevertheless, the 2J(29Si-29Si) values measured for compound ( 5 ) allowed us to make a complete assignment of the signals in the 29Si n.m.r. spectrum (Table 1). Experimental Allcompounds used were synthesized as described previ~usly.~,' ' The "N, "0, and 29Si n.m.r. spectra were recorded on a Bruker WM-360 spectrometer at 36.5, 48.8, and 71.5 MHz, respectively. CDCI, was dried over 4A molecular sieves and used as the solvent and internal deuterium lock material.All spectra were measured for ca. 20--50% solutions at 303 K in 10- mm sample tubes. The "N and 29Si chemical shifts were measured using the INEPT sequence. '' The 5N-'H couplings were measured from the 360-MHz 'H spectra viu multiple quantum n.m.r. in 5-mm sample tubes (Figure 3). The pulse sequence used, 90',( ' H)- 1 /( 2JNH)- 9 0 3 ' 5N)-t1 -90",( 'N)-acquire( H), has been previously reported." ' 3C-2'Si couplings were measured from the ' 3C n.m.r. spectra. 5N-29Si couplings were measured at natural abundance of isotopes in the 29Si n.m.r. spectra for 0-Si-N1596 J. CHEM. SOC. DALTON TRANS. 1987 fragments and in the ‘’N n.m.r. spectra for N-Si-N fragments as described elsewhere.10*’9 1.r. spectra were recorded on a Perkin-Elmer 580 B spectrometer using the same samples as for n.m.r.spectra. Acknowledgements The authors are grateful to Miss Rita Upmacis of Nottingham University for assistance in the preparation of the English version of the manuscript. References I G. Engelhardt, H. Jancke, M. Magi, T. Pehk, and E. Lippmaa, J. 2 H. Jancke,G. Engelhardt, M. Magi,and E. Lippmaa,Z. Chem., 1973, 3 B. D. Lavrukhin, B. A. Astapov, A. V. Kisin, and A. A. Zhdanov, Izv. 4 B. D. Lavrukhin, K. A. Andrianov, and E. I. Fedin, Org. Magn. 5 B. D. Lavrukhin, B. A. Astapov, A. A. Zhdanov, G. Engelhardt, and 6 A. Takasuka and Y. Torui, J. Chem. SOC., Perkin Trans. 2,1984,1545. Orgunomet. Chem., 1971, 28, 293. 11, 435. Akad, Nuuk SSSR, Ser. Khim., 1983, 1059. Reson., 1975, 7, 298. H. Jancke, Org. M a p .Reson., 1982, 18, 71. 7 J. Mason, Chem. Rer., 1981, 81, 205. 8 J. P. Kintzinger, ‘Oxygen NMR. Characteristic Parameters and Applications,’ Springer-Verlag, Berlin, 198 1, p. 14. 9 H. Marsmann, ‘29Si NMR Spectroscopic Results,’ in ‘NMR Basic Principles and Progress,’ Springer-Verlag, Berlin, 1981, vol. 7, p. 90. 10 l?. Kupee, E. LiepipS, 0. Pudova, and E. Lukevics, J. Chem. SOC.. Chem. Commun., 1984, 58 1. I 1 H. A. Bent, Chem. Rev., 1961,61, 275. 12 W. Fink, Angew. Chem., Int. Ed. Engl., 1966, 5, 760. 13 R. E. Wasylishen and T. Schaefer, Can. J. Chem., 1973, 51, 3087; 14 J. Kowalewsky, Prog. Nucl. Magn. Reson. Spectrosc., 1977, 11, 1. M. D. Beer and R. Grinter, J. Magn. Reson., 1977, 26, 421. 15 16 17 18 19 W. Freyer, Z. Chem., 1981, 21, 47; G. J. Martin, M. L. Martin, and J.P. Gouesnard, ” 5N NMR Spectroscopy,’ Springer-Verlag, Berlin, 1981, p. 195; G. C. Levy and R. L. Lichter, ‘Nitrogen-15 NMR Spectroscopy,’ Wiley, New York, 1979, ch. 4. V. E. Shklover, Yu. T. Struchkov, G. V. Solomatin, A. V. Zachernyuk, and K. A. Andrianov, Zh. Strukt. Khim., 1979,20, 309. G. A. Morris and R. Freeman, J. Am. Chem. SOC., 1979, 101, 160. A. Bax, R. H. Griffey, and B. L. Hawkins, J. Magn. Reson., 1983,55, 301. E. Kuge, E. LiepipS, and E. Lukevics, Angew. Chem., Inf. Ed. Engl., 1985, 24, 568. Received 10th February 1986; Paper 61286 J. CHEM. SOC. DALTON TRANS. 1987 1593 Nitrogen-I 5 and Oxygen-I7 Nuclear Magnetic Resonance Chemical Shifts and 15N-29Si, 13C-29Si, and I5N-l H Coupling Constants in Cyclosilazoxanes Eriks KupCe, Edvards Liepigs, and Edmunds Lukevics Institute of Organic Synthesis, Latvian S.S.R. Academy of Sciences, Riga, U.S.S.R. Boris Astapov Institute of Organo- Element Compounds, U.S.S.R. Academy of Sciences, Moscow, U.S.S.R. The effects of ring size on 15N, 170, and 29Si n.m.r. chemical shifts and 15N-29Si, 13C-2gSi, l5N-'H, and 29Si-29Si coupling constants in cyclosilazoxanes have been studied. The opposite shielding sequence in 8(15N) and 6(29Si) and the decrease in 15N-29Si coupling with diminishing ring size are consistent with higher Si-N bond ionicity (reduced d,-p, interaction) in smaller rings. The 170 chemical shift is practically independent of ring size in cyclosiloxanes. It has been proposed that the factors governing the 170 chemical shift act in opposite directions and are balanced.A linear relationship between N-H bond stretching frequency and 15N n.m.r. parameters (15N-lH coupling and 15N chemical shifts) has been established. Although a large amount of data is available on 'H, 13C, and 29Si n.m.r. chemical shifts ( 6 ) of cyclosiloxanes, cyclosilazanes, and cycl~silazoxanes,~-~ data have not been reported for the "N and "0 chemical shifts or the one-bond coupling constants ('4 1sN-29Si, 13C-29Si, and 15N-'H for these types of compounds. I t has been that 6(29Si) reflects variations in d,-p, interaction along Si-X (X = 0 or N). In this respect measurement of 6( 5N), 6(' 70), and especially ' J ( 5N-29Si) can give a more definite result. In this paper we report the 15N and 1 7 0 chemical shifts and the 5N-29Si, 3C-29Si, and "N-IH coupling constants, together with some two-bond 29Si-0-29Si couplings, in six- to twelve-membered cyclosilazoxanes and in some acyclic derivatives, selected partly on the basis of their availability.It has been reported that a linear relationship exists between the NH stretching frequency (vNH) and the "N n.m.r. parameters [6( 'N) and ' J ( "N-IH)] in saturated primary amines and anilines. In order to examine the existence of the analogous relationships in cyclosilazoxanes and to compare the factors influencing i.r. and n.m.r. parameters in organic amines with those in cyclosilazoxanes, we have also measured vNH. Results and Discussion 5N und 7O Chemical Shifis.--It appears from Table 1 that 6(I5N) and 6("O) are less sensitive to the nature of the substituents at the silicon atom than 6(29Si).Nevertheless, the influence over the two bonds (p effect) is important. Data obtained for the acyclic compounds (1)-(5) indicate that 15N and ' 7O shielding decreases with increase in electronegativity of the substituents at adjacent silicon atoms. For compound (5) this allows us to observe the 170 signal of the OSiMe, end- group separately. In cyclosilazoxane rings of equal size, substitution of one nitrogen atom by oxygen decreases the shielding of the adjacent I5N by an average value of 2 p.p.m. A similar tendency is characteristic of 7O n.m.r. spectra. The influence of ring size on 6(15N) is opposite to that of 6(29Si). In the cyclosilazoxanes studied 6(29Si) moves to higher field with increasing ring size, which is in compliance with previous investigations.'-3 It has been suggested that this is the result of an increase in the d,-p, interaction along the Si-X (X = 0 or NH) bond.The tendency of 6("N) to move downfield with increase in ring size may be understood in (CIMezSi ),X [Me,Si(OSiMe,), 1,NH ( Me3Si ),X ( 1 ) X=NH ( 3 ) X = NH ( 5 ) ( 2 ) x = o ( 4 ) x = o Me2 X I I Me2 X /si\x I I Ph2 o/si\o I I ( 6 ) X = NH (8) X=NH,Y=O ( 7 ) x = 0 ( 9 ) X = 0, Y =NH Me2Si - X - SiMe, I I i I X X H (10) H MqSi- N -SiMe2 I I 0 0 I I MezSi- X -SiMez MezSi- X -Siblet (11) X = NH (12) x = 0 (13) X = NH (14) x = 0 Me2 Me, I I I I I I I I 0- Si - Y - SiMe, MezSi -0-Si-NH SiMe, 0 HN Me, Si X I I MezSi 0- Si - N- SiMe, Me, H 0- Si - N - SiMe, Me, (15) X = NH, Y = O (16) X= 0, Y = NH (17) similar terms.Thus, the opposite shielding sequence in the nitrogen and silicon resonances reflects an increase in Si-N bond ionicity (decrease in d,-p, interaction) in smaller rings. Unexpectedly, 6( I 70) is practically independent of ring size in 'pure' cyclosiloxanes. This indicates that the factors governing 6( 70) act in opposite directions and are balanced.1594 J. CHEM. SOC. DALTON TRANS. 1987 Table 1. N.m.r. parameters and NH stretching frequencies of cyclosilazoxanes and acyclic analogues " Compound 6( "N) (1) - 354.2 (2) -334.3 (3) (4) (5) -335.7 (6) - 347.3 (7) (8) - 345.4 - - - (9) - 342.9 (15) - 336.7 (17) - 334.8 ~ ( 1 7 0 ) - 39.0 70.0 71.4' 72.0 60.1 63.0 56.0 71.4 68.1 70.9 72.3 70.5 - - - 69.6 67.3 6(29Si) 2. I 6.5 13.3 6.9 - 11.9 ( x ) ~ .~ - 4.6 - 8.7 -4.0 (NN) -4.3 (ON) -3.4 (ON) -9.0 (00)' - 1.2 (ON) - 8.2 - 19.4 - 11.3 - 10.9 (ON) - 19.8 (00) -8.3 (NN) - 12.6 (ON) - 22.6 (00) - 12.6 (NN) - 13.2 (ON) -23.6 (00)' - 12.3 'J( 'sN-29Si) lJ('3C-29Si) 13.4 56.3 59.5 14.4 68.1 72.6 17.4 69.4 15.4 63.0 74.2 15.8 62.3 15.8 68.5 15.5 68.4 74.0 15.7 68.2 16.9 64.0 75.4 15.9 69.8 15.9 69.7 75.4 17.7 64.5 17.7 69.4 74.9 17.3 69.3 17.1 69.6 74.7 17.2 69.2 - - - - ' J ( "N-' H) 69.6 68.1 - - - 69.4 69.5 69.5 69.6 67.0 67.1 67.3 66.3 - - 66.7 65.9 VNn 3 402 3 364 3 376 3 403 3 402 3 400 3 396 3 385 3 383 3 384 3 377 - - - - 3 380 3 373 " ' J in Hz; 6 in p.p.m. relative to CH,N02 ("N), H 2 0 ("O), or SiMe, (29Si); vNH in cm-'. * Ref. 9. (2-p), 3.1 (p-y), 2.2 (y-6), 2.5 ( 6 - ~ ) , 1.3 (E-<) Hz.2J(29Si-29Si) = 1.02 Hz. ' zJ(29Si-29Si) = 1.50 Hz. 56.1 p.p.m. (OSiMe,). 2J(29Si-29Si) = 2.1 2J(29Si-29Si) = 1.54 Hz. HN-SiaMe,OSieMe2OSiYMe2OSi6Mez0Si"Me2OSi~Me3; -22.3 (p), -22.4 (y), - 22.5 (6), -21.6 (E), + 7.1 p.p.m. (6). 74.4 (p), 74.4 (y), 74.3 (S), 74.3 ( E ) Hz. Table 2. Si-X-Si (X = 0 or N) valence angles (") in somc cyclosiloxanes, cyclosilazanes, cyclosilazoxanes, and acyclic analogues Compound (Me,SiO), (Me,SiO), (Me,Si),O (Me,SiNH), (Me,SiNH), (Me,Si),NH (Ph,SiO),( Ph,SiNEt) (PhMeSiO),(NH),- (PhMeSiO), Si-X-Si 131.6 144.8 148.8 126.8 132/ 131.3 126.3 130.0 Method" Ref. ED h ED h X,ED c * , d ED 6' X R ED h X i X i " ED = electron diffraction, X = X-ray analysis. H. Oberhammer, W. Zeil, and G. Fogaresi, J. Moi. Sfrucr., 1973, 18, 309. ' M. J. Barrow, E.A. V. Ebsworth, and M. M. Harding, Actu Crj*stullogr., Sect. B, 1979, 35, 2093. B. Csakovari, Z. Wanger, P. Gomory, F. C. MijlholT, B. Roszandai, and I. Hargittai, J. Orgunomet. Cliem., 1976, 107, 287. B. Roszondai, I. Hargittai, A. Golubinski, L. V. Vilkov, and V. S. Mastryukov, J. Mol. Struct., 1975,28,339. Mean value. G. Smith and L. E. Alexander, Actu Crj~.srullogr., 1963, 16, 1015. T. Fjeldberg, J . Mol. Struct., 1984, 112, 159. W. Fink and P. J. Wheatley, J. Cheni. Soc. A, 1967, 1517. V. E. Shklover, N. G. Bokij, Yu. T. Struchkov, K. A. Andrianov, A. B. Zachernyuk, and E. A. Zhdanova, Zk. Strukt. Khim., 1974, 15. 864. The main contributor to the chemical shift is the variation in the paramagnetic term (oPA) of the shielding of nucleus For a second-row element this is approximated by equation ( l ) , where po is the permeability of free space, pe is the Bohr magneton, and ( Y - ~ ) ~ , , is the radial-expansion term for the 2p electrons; the Q terms express the imbalance of charge in the -350 t 3: -340 P z Y ro VV" 3370 3380 3390 3400 v,,/cm-' Figure 1.Plots of corrected ' 'N chemical shifts [6( "N*)] against NH stretching frequencies. Regression analysis gave the equation S("N*) = 884.5 - 0.361vN,, (number of points, n = 8, correlation co- efficient, r = 0.996) valence shell of A and AE is the mean energy of electronic excitation. As has been demonstrated by CND0/2 calculations,' 6( 'N) in primary aliphatic amines is governed by the AE term, but in aniline derivatives by the ZQ and ( Y - ~ ) ~ , terms. Furthermore, the linear relationships of 6( "N) UPYSUS vNH were observed with a negative slope ( - 1.7601 ) in the former case and a positive one (0.3855) in the latter.Generally no correlation has been found between 6( ' 'N) and vNH in the cyclosilazoxanes studied. This indicates the existence of the factors which influence 6( 5N) but not vNH. Indeed, if the p effect is excluded by introducing the correction of - 2 p.p.m. forJ. CHEM. SOC. DALTON TRANS. 1987 1595 70 69 N I I 68 I 5 -' 67 66 3370 3380 3390 3 4 0 0 vNH /cm" - Figure 2. Plots of 15N-'H coupling constants against NH stretching frequencies. Regression analysis gave the equation 'J(I5N-'H) = 0.1278 vNH - 365.5 ( n = 10, r = 0.994) the substitution of one nitrogen atom by oxygen in the p position, the resultant 'corrected' I5N chemical shifts [6(' 5N*)] show a good correlation with vNH (Figure 1).The slope of the line obtained (-0.3610) is ca. five times less steep than that found in aliphatic amines ( - 1.7601) and is close to that in aniline derivatives (0.3855), but with opposite sign. It seems, therefore, that 6( "N*) in cyclosilazoxanes are also governed by the A€ term, but the contribution of the CQ and ( r - 3 ) 2 p terms is significantly increased, possibly, due to the presence of dn-p, interactions along the Si-N bonds. The lack of correlation between 'non-corrected' 6( "N) and vNH may be explained by significant influence of fJ substitution on the CQ term. It may also be suggested that independence of 6("O) of ring size in cyclosiloxanes (and cyclosilazoxanes) is a consequence of equal contributions of AE on the one hand, and CQ and terms on the other.Coupling Constants.-Coupling constants to 29Si are very sensitive to the electronegativity of the substituents at the silicon atom and usually increase with increasing electronegati~ity.~.'~ Similar dependences exist for ' J ( ' sN-29Si) and 'J(' 3C-29Si) in the acyclic compounds (1)-(5) (Table 1). In cyclosilazoxanes the 1J(29Si-'3C) values for the fragments N-Si-N (62.3-64.5), N-Si-0 (68.2-69.8), and 0-Si-0 (74XL75.4 Hz) lie in a very narrow, non-overlapping and characteristic range, which allows us to use these couplings for analytical purposes. In contrast to ' J ( 13C-29Si), the 'J("N- 24 Si) values are practically independent of the substituents (N or 0) at the silicon atom.This may be understood in terms of Bent's rule,' ' which suggests that major changes in s-orbital overlap proceed in bonds with more electropositive substituents (Si-C). The decrease of ' J ( ' 5N-29Si) with diminishing ring size gives further evidence of higher Si-N bond ionicity in smaller cyclosilazoxane rings. The effect of p substitution on ' J ( I5N-'H) in the compounds studied seems to be negligible. The dependence of the e.vocyclic coupling 'J("N-'H) on ring size is opposite to that of 1J(29Si-'3C), for example 'J(15N-'H) decreases on going to larger cyclosilazoxane rings, whilst 1J(29Si-' %) slightly in- creases. This fact is consistent with the suggested increase in ionicity of Si-X (X = 0 or N) bonds in smaller rings. Alter- 10 Hz H I * Figure 3.An example of observation of 15N-'H coupling in a 'H spectrum via multiple quantum n.m.r. Compound (lo), ca. 20% solu- tion in CDCI,, 5-mm sample tube, sweep width 600 Hz, 200 scans. A signal marked * is due to an incompletely suppressed signal from protons bonded to 14N natively, the observed changes in exocyclic couplings may be explained by changes in appropriate valence angles. A greater sensitivity of 'J("N-'H) to changes in ring size as compared with ' J ( I 3C-29Si) is consistent with the observation" that the valence angles Si-X-Si depend on ring size to a greater extent than X-Si-X. However, it appears from the data present in Tables 1 and 2 that 'J("N-'H) decreases with an expansion of the Si-N-Si valence angle. This contradicts the situation normally observed for organic amines.' 3p1s We have found a linear relationship between ' J ( I5N-'H) and vNH (Figure 2).As has been demonstrated by CND0/2 calculations,6 such a relationship exists because both 'J( "N-' H) and vNH depend primarily on the s character of the N-H bond. It should be noted that simultaneous decrease of 'J("N-'H) and vNH may be explained by an increase in the N-H bond length. We have also measured some two-bond 29Si-29Si couplings (Table 1). It appears that 2J(29Si-29Si) decreases with decreasing ring size, but the information is as yet insufficient. Unfortunately, the measurement of 2J(29Si-29Si) in cyclo- silazoxanes is difficult to perform because the values of these couplings are small ( 1 .&3.1 Hz). Further complications arise from broadening of the 29Si signals of the silicon atoms bonded to nitrogen due to the fast quadrupolar relaxation of I4N nuclei.Nevertheless, the 2J(29Si-29Si) values measured for compound ( 5 ) allowed us to make a complete assignment of the signals in the 29Si n.m.r. spectrum (Table 1). Experimental Allcompounds used were synthesized as described previ~usly.~,' ' The "N, "0, and 29Si n.m.r. spectra were recorded on a Bruker WM-360 spectrometer at 36.5, 48.8, and 71.5 MHz, respectively. CDCI, was dried over 4A molecular sieves and used as the solvent and internal deuterium lock material. All spectra were measured for ca. 20--50% solutions at 303 K in 10- mm sample tubes. The "N and 29Si chemical shifts were measured using the INEPT sequence. '' The 5N-'H couplings were measured from the 360-MHz 'H spectra viu multiple quantum n.m.r.in 5-mm sample tubes (Figure 3). The pulse sequence used, 90',( ' H)- 1 /( 2JNH)- 9 0 3 ' 5N)-t1 -90",( 'N)-acquire( H), has been previously reported." ' 3C-2'Si couplings were measured from the ' 3C n.m.r. spectra. 5N-29Si couplings were measured at natural abundance of isotopes in the 29Si n.m.r. spectra for 0-Si-N1596 J. CHEM. SOC. DALTON TRANS. 1987 fragments and in the ‘’N n.m.r. spectra for N-Si-N fragments as described elsewhere.10*’9 1.r. spectra were recorded on a Perkin-Elmer 580 B spectrometer using the same samples as for n.m.r. spectra. Acknowledgements The authors are grateful to Miss Rita Upmacis of Nottingham University for assistance in the preparation of the English version of the manuscript. References I G. Engelhardt, H. Jancke, M. Magi, T. Pehk, and E. Lippmaa, J. 2 H. Jancke,G. Engelhardt, M. Magi,and E. Lippmaa,Z. Chem., 1973, 3 B. D. Lavrukhin, B. A. Astapov, A. V. Kisin, and A. A. Zhdanov, Izv. 4 B. D. Lavrukhin, K. A. Andrianov, and E. I. Fedin, Org. Magn. 5 B. D. Lavrukhin, B. A. Astapov, A. A. Zhdanov, G. Engelhardt, and 6 A. Takasuka and Y. Torui, J. Chem. SOC., Perkin Trans. 2,1984,1545. Orgunomet. Chem., 1971, 28, 293. 11, 435. Akad, Nuuk SSSR, Ser. Khim., 1983, 1059. Reson., 1975, 7, 298. H. Jancke, Org. M a p . Reson., 1982, 18, 71. 7 J. Mason, Chem. Rer., 1981, 81, 205. 8 J. P. Kintzinger, ‘Oxygen NMR. Characteristic Parameters and Applications,’ Springer-Verlag, Berlin, 198 1, p. 14. 9 H. Marsmann, ‘29Si NMR Spectroscopic Results,’ in ‘NMR Basic Principles and Progress,’ Springer-Verlag, Berlin, 1981, vol. 7, p. 90. 10 l?. Kupee, E. LiepipS, 0. Pudova, and E. Lukevics, J. Chem. SOC.. Chem. Commun., 1984, 58 1. I 1 H. A. Bent, Chem. Rev., 1961,61, 275. 12 W. Fink, Angew. Chem., Int. Ed. Engl., 1966, 5, 760. 13 R. E. Wasylishen and T. Schaefer, Can. J. Chem., 1973, 51, 3087; 14 J. Kowalewsky, Prog. Nucl. Magn. Reson. Spectrosc., 1977, 11, 1. M. D. Beer and R. Grinter, J. Magn. Reson., 1977, 26, 421. 15 16 17 18 19 W. Freyer, Z. Chem., 1981, 21, 47; G. J. Martin, M. L. Martin, and J. P. Gouesnard, ” 5N NMR Spectroscopy,’ Springer-Verlag, Berlin, 1981, p. 195; G. C. Levy and R. L. Lichter, ‘Nitrogen-15 NMR Spectroscopy,’ Wiley, New York, 1979, ch. 4. V. E. Shklover, Yu. T. Struchkov, G. V. Solomatin, A. V. Zachernyuk, and K. A. Andrianov, Zh. Strukt. Khim., 1979,20, 309. G. A. Morris and R. Freeman, J. Am. Chem. SOC., 1979, 101, 160. A. Bax, R. H. Griffey, and B. L. Hawkins, J. Magn. Reson., 1983,55, 301. E. Kuge, E. LiepipS, and E. Lukevics, Angew. Chem., Inf. Ed. Engl., 1985, 24, 568. Received 10th February 1986; Paper 61286