A multinuclear NMR study of organosilances


Material Information

A multinuclear NMR study of organosilances
Physical Description:
Kanyha, Paul John, 1956-
Publication Date:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 20303550
oclc - 12203446
System ID:

Table of Contents
    Title Page
        Page i
        Page ii
    Table of Contents
        Page iii
        Page iv
        Page v
    Chapter 1. Introduction
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
    Chapter 2. Experimental
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
    Chapter 3. Novel NMR methods
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
    Chapter 4. Results and discussion
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
    Chapter 5. Conclusions
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
    Biographical sketch
        Page 144
        Page 145
        Page 146
Full Text








The author wishes to thank the member of his supervisory

committee and to express his sincere appreciation to Dr. W.S.

Brey, Jr., chairman of the author's supervisory committee,

for his constant guidance and help during the course of this

work. The author gratefully thanks Dr. R.W. King for his

assistance in mass spectral interpretation, Mr. D. Plant

for his assistance in the synthesis of the chloroaminosilanes,

glassblowers Mr. R. Strohschein and Mr. R. Moshier, and typist

Ms. P. Victor.

The author wishes to acknowledge PCR, Inc.,of Gaines-

ville, Florida, who kindly donated some of the organosilanes

to the Department of Chemistry and the Graduate School of

the University of Florida for a graduate assistantship and

to the Instrument Program, Chemistry Division, National

Science Foundation, for aid in the purchase of the Nicolet

NT-300 NMR spectrometer.



ACKNOWLEDGMENTS ...................................... ii

ABSTRACT .............................................. iv


I INTRODUCTION ................................... 1

Evidence of (d-p)w Bonding .................. 1
Summary of the Literature Findings .......... 40
Aim of the Investigation .................... 41

II EXPERIMENTAL ................................. 43

Materials .................................... 43
Spectrometer ................. .......... .... 47
Computer Calculations ....................... 49

III NOVEL NMR METHODS ............................. 50

Modified Selective Population Transfer
Experiments ............ ..................... 50
13C Spin-Echo J-Modulation Experiments ...... 64

IV RESULTS AND DISCUSSION ...................... 75

Trimethylsilyl Compounds .................... 75
Methylsilanes ................................ 100
Silylamines ................................. 110

V CONCLUSIONS .................................. 130

REFERENCES ......................................... 131

BIOGRAPHICAL SKETCH ............................... 144


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Paul John Kanyha

April 1984

Chairman: Dr. W.S. Brey, Jr.
Major Department: Chemistry

1 13 15 29
The H, 13C, 15N and 29Si chemical shifts and coupling
1 29 .13 1 29 15
constants including J( 29Si 13C) and J( 29Si 15N) are reported

for 26 trimethylsilanes, 48 methylsilanes, and 33 silylamines.

The chemical shift data are best interpreted in terms of cora-

peting (d-p) T and inductive effects. The J( 29Si 3C) and
1 29 15
J( 29Si 5N) are strongly dominated by the Fermi contact

interaction. For 29Si nuclei in silylamines, T2 scalar

coupling relaxation is dominant. From the chemical shift

and coupling constant data, it is concluded that there is

significant (d-p)ff bonding for Si-N and Si-O bonds, moderate

to negligible (d-p)rr bonding for Si-C (unsaturated carbon)

and negligible (d-p)f bonding for Si-S and Si-Cl bonds.

In addition, novel NMR pulse sequences are described.

One of these is a method to decouple selectively a group of

magnetically equivalent protons which are also used for

polarization transfer. The others are techniques to sup-

press strong unwanted signals and to differentiate between
13C multiplicities.


The question of (d-p)n bonding in organosilanes has long

been an area of controversy, partly because (d-p)r bonding has

often been invoked to explain the "unexplainable" (1-4). Most

cases where the participation of silicon d orbitals in bonding

has been cited have been concerned with physical and spectro-

scopic properties rather than chemical properties. Although

the (d-p)h bonding model has been employed most often in the

interpretation of Si-N and Si-O bonds, there is also some

evidence of (d-p)7r bonding in Si-F bonds and to a lesser

extent in Si-Cl, Si-S, and Si-C (unsaturated carbon) bonds.

Therefore, it is of considerable interest to determine the

extent of double-bond formation in these bonds.

Evidence of (d-p)T Bonding

Because of the limited commercial availability, high

volatility, and arduous synthesis of the fluorosilanes, only

compounds containing Si-C, Si-N, Si-O, Si-S, and Si-Cl bonds

were included in this investigation. In the next few pages,

the literature will be briefly reviewed for evidence of

(d-p)7 bonding in these bonds.

Bond Lengths and Angles

Participation of (d-p)n bonds is usually claimed when bond

lengths are shorter than predicted by the Schomaker-Stevenson

equation (5) in which the bond length is taken to be the sum

of atomic radii, less a correction for electronegativity.

Judging by bond lengths, an increase in bond orders due to

(d-p)ir bonding occurs for Si-N and Si-O bonds. A quantitative

estimation of the extent of 'T bonding from the bond shortening,

however, is difficult for several reasons. Firstly the single

bond reference length is only an estimate, and secondly it is

well known that, as in carbon compounds, inductive effects (-I)

and increased s character both result in bond shortening. There-

fore, the shortening of the Si-O bond length with increasing

electronegativity of X in the series X3SiOSiX3 can be attrib-

uted to increased bond-order due either to better (d-p)7 over-

lap or to inductive effects. Table 1-1 presents typical bond

lengths for simple compounds.

With the exception of the Si-N bond length for

Me 3SiN = NSiMe3, the bond lengths of Si-N and Si-O are 0.05

to 0.18 A shorter than that predicted for a "single" bond by

the Schomaker-Stevenson equation. The Si-N bond length of

1.81 A for Me3SiN = NSiMe3 (6) seems to indicate an absence

of (d-p)7 bonding. The lengths of Si-C, Si-S, and Si-Cl bonds

are within 0.04 A of the Schomaker-Stevenson bond length, indi-

cating that (d-p)ir bonding is weak or absent. For example, from

an electron-diffraction study (7) of (Me2N) 3SiCl a Si-N bond

length of 1.715 A was observed which is 0.08 A shorter than

Table 1-1. Typical Si-X Bond Length (A). a

Si-C 1.88b

Me 3SiCECH 1.825 SiC=C (8)c
1.865 SiMe

H 3SiCH=CH2 1.853 (10)

(Ph3Si)20 1.864 (12)

(Me3Si)2NH 1.876 (14)

(Me2Si)4 1.889 (16)

Me2(Cl)SiCH=CH2 1.838 SiC=C (9)
1.876 SiMe

Ph 4Si 1.872 (11)

(Ph3Si)2NH 1.878 (13)

(Me2SiNH)3 1.871 (15)

Cl3SiCH2SiCl3 1.866 (17)

Si-N 1.80b

H3 SiNMe2 1.715

C3 SiNMe2 1.657

F3 SiNMe2 1.654

(H3Si)2NH 1.724

(Me3Si) 2NH 1.735

(Ph3Si) 2NH 1.718

(H3Si)3N 1.734

(H3Si) 2N-N(SiH3)2 1.713

(H3Si) 2NCN 1.696










Si-O 1.76b

H SiOSiH3 1.631 (29,30)

Me 3SiOSiMe3 1.626 (29)

Ph 3SiOSiPh3 1.616 (12)

Cl 3SiOSiCl3 1.592 (32)




(H3Si) 2NMe

(Me2SiNH) 3

(Me2SiNPh) 2

(Me2N) 3SiCl

Me 3SiN=NSiMe

H 3SiN3



(Bu2SiO) 3

F 3SiOSiF3

1.703 (19)

1.646 (21)

1.714 (22)

1.726 (24)

1.728 (15)

1.744 (25)

1.715 (7)

3 1.81 (6)

1.719 (28)

1.640 (30,31)

1.580 (32)

1.654 (33)

1.580 (34)

Si-S 2.16 b

H3 SiSSiH3

2.136 (35)

Table 1-1 continued.

Si-Cl 2.05b

Cl 3SiCH 2SiCI3



2.027 (17)

2.014 (21)

2.082 (7)

Cl 3SiOSiCI3 2.011 (32)

Cl3 SiNMe2 2.023 (20)

Cl(Me2)SiCH=CH2 2.078 (9)

a Me = CH3, Bu = (CH3)3C, and Ph = C6H5.

bSchomaker-Stevenson (5) "single" bond length.

CNumbers in parentheses are references.

the "single" bond length predicted by the Schomaker-Stevenson

equation. The observed Si-Cl bond length (7) is considerably

longer than that in Cl3SiCH 2SiCl3 (17) and would seem to indi-

cate a lack of (d-p)r bonding to chlorine.

Structural data indicate that unusual bond angles occur

only for Si-N and Si-O compounds as compared to the bond

angles for the carbon analogs. In contrast, the bond angles

for Si-C and Si-S compounds are within 2 of the bond angles

for the corresponding alkyl derivatives; therefore, these

bond angles seem to indicate an absence of (d-p)r bonding.

Typical bond angles for simple compounds and the analogous

carbon derivatives are shown in Table 1-2.

Nitrogen is invariably planar when linked to two or three

Si atoms, as in (H3Si)3N (26), (H3Si)2NH (23), (H3Si)2NMe (24),

(H3Si)2NN(SiH3)2 (7), and (Me3Si)2NH (14). If, however, a

single silyl group is bonded to nitrogen, nitrogen will not

always be planar and can in fact be an approximately trigonal

pyramid, similar to that of the alkylamines. This is

demonstrated by H3SiNMe2 in which the Si-N bond is tilted out

of the CNC plane by 27.8 (18). In contrast, planar nitrogen

has been observed for F 3SiNMe2 (20), Cl 3SiNMe2 (20), and

(Me2N)3SiCl (7). An electron diffraction study (28) showed

H 3SiN3 to have a nonlinear SiN3 skeleton with a Si-N-N bond

angle of 123.8 while the molecule (H3Si)2NCN has a planar

Si 2NCN skeleton. This, therefore, suggests (d-p)r bonding

in the cyanamide and an absence of (d-p)Tr bonding in the

azide. As another example, the X-ray structure of

Table 1-2. Si-X-Si and Si-X-C Bond Angles: Comparison with
Carbon Compounds.a


H 3Si-CH 2-SiH3


H 3Si-O-CH3

Ph 3Si-O-SiPh3


(H3Si) 3N





114.4 (36)b

142.2 (29,30)

120.6 (30,31)

180.0 (12)

127.7 (23)

119.7 (24)

180 (19)

180 (22)

97.4 (35)


H 3C-CH2-CH3

H 3C-O-CH3



(H3C) 3N



H 3C-S-CH3


112.4 (36)

111.6 (30)

127.9 (37)

112.2 (38)

110.6 (36)

140.0 (39)

147.5 (39)

98.9 (40)

aPh = C6 H.
bNumbers in parenthesis are references.
bNubers in parenthesis are references.

Table 1-3.

Geometry of Some X 3SiOSiX3 Molecules.

Si-O Length (A)







Si-O-Si Angle (0)







aMe = CH3 and Ph = C6 H.








bis(trimethylsilyl) diimide has shown the Si-N=N bond angle to

be 120 with a Si-N bond length of 1.81 A, suggesting an ab-

sence of (d-p)7 bonding (6).

The Si-O-Si bond angle in disiloxanes is always larger

than the angle predicted by the valence shell electron pair

repulsion theory for single bonds. This increase in bond

angle at oxygen must result from (d-p)r bonding, since as

the a bond assumes increasing sp hybrid character, there will

be better pir (or pn') overlap with the 3d 2_y2 (or dz2)

orbital. Therefore, in the series X3SiOSiX3 (X=H, Cl, F),

the Si-O bond length decreases and the Si-O-Si bond angle

increases with increasing electronegativity of X, as shown

in Table 1-3. X-ray crystallographic studies have shown the

Si-O-Si bond angle for Ph3SiOSiPh3 (12) and(PhCH2)3SiOSi(CH2Ph)3

(41) to be 180.0. These large bond angles can be interpreted

in terms of strong (d-p)r bonding, since the carbon analog,

Ph 3COCPh3 has a C-O-C bond angle of 127.9 (37). Table 1-3

presents Si-O-Si bond angles and Si-O bond lengths for some

X3SiOSiX3 molecules.

Vibrational Spectroscopy

In theory, force constants calculated from infrared,

Raman, or microwave spectra through normal coordinate anal-

ysis should be excellent indicators of the extent of (d-p)n

bonding since force constants can be used to calculate bond

orders by means of theSiebert formula (42). The Siebert

formula is a linear correlation between the total bond order

and the ratio of the actual force constant to that for an
ideal single bond AB, this being defined as 7.2x(ZAZB/nAnB)

where Z is the atomic number and n the principal quantum

number. In practice, however, the calculation of force con-

stants is difficult especially for complex molecules and

often is uncertain because of the numerous assumptions made.

Therefore, even when force constants are available, the wide

variations observed in published values, e.g., 5.56 and 5.05

mdyne/A for the Si-Obond in H3SiOSiH3 (43,44) render any con-

clusions deduced from them exceedingly questionable.

Vibrational studies (45) have shown X4-nSi(NMe2)n (where

X = H, Cl, Me and n = 0 to 4) compounds to have Si-N and Si-Cl

bond orders greater than one. The calculated Si-Cl Siebert

bond order of 1.3 from the reported force constant (45) of

(Me2N)3SiCl seems to indicate moderate (d-p)ir bonding result-

ing in a corresponding shortening of the Si-Cl bond, when in

fact, the observed bond length (7) is 0.03 A longer than the

Schomaker-Stevenson "single" bond length. Therefore, this

further demonstrates the unreliability of force constant

data in determining bond orders.

H6fler calculated force constants for the Si-C bond for

Ph nSiXn (where X = H, Cl, F, OH for n = 0 to 4) and found
4-n n
the force constants ranged from 3.06 to 3.44 mdyne/A (46);

these force constants correspond to Siebert bond orders of

1.1 to 1.2. Although the reported force constant of 3.44

mdyne/A for the Si-C bond in PhSiF3 (46) seems to indicate

a moderate amount of (d-p)7 bonding in the related compound

MeSiF3 where no (d-p)n bonding is possible, the calculated

force constant for the Si-C bond is 3.47 mdyne/A (47). Con-

sequently, the increased force constant must result from

inductive effects (-I) rather than (d-p)7 bonding. Comparison

of the force constants for the Si-C bond for Ph 3SiH (46), 3.30

mdyne/A with Me3SiH (48), 2.97 mdyne/A indicates increased

bond order due to (d-p)7 bonding. Therefore, the large force

constants H6ffler (46) observed can be interpreted as arising

from both inductive effects and (d-p)h bonding.

A recent study of the molecular geometry of the series

Me Si(NCX) 4n (where n = 0 to 3, X = S, 0) by both infrared
n i(NX)
spectra and dipole moments indicates a linear structure for

Si-N-C-X chains (49). A C3v symmetry for monofunctional sili-

con compounds is unambiguously revealed by comparison of the

IR spectra with that of the carbon analogs (39), which have

structures that are bent at the N. The structural differences

between the silicon and carbon compounds can be attributed to

(d-p)7 bonding in the former.

The extent of delocalization of nonbonding electron

pairs can be determined by infrared measurements of acidity

and basicity. In the absence of (d-p)rr bonding, since sili-

con is more electropositive than carbon, silicon will attract

bonding electrons less than carbon; consequently SiNH, SiOH,

and SiSH compounds should be weaker acids than the correspond-

ing carbon derivatives, while N, 0, and S when linked to

silicon should exhibit stronger Lewis base properties.

However, if the nonbonding pairs on N, 0, or S are delocal-

ized by (d-p)7 bonding, the opposite behavior would result.

Ulbricht and coworkers (50) measured the oxygen basicity

of the series Me nM(OEt) (where M = C, Si) by observing

changes in the 0-Hstretching frequency of phenol. The re-

sults were that the oxygen basicity of Me3SiOEt is lower and

that of MeSiOEt3 and SiOEt4 is higher than the oxygen basi-

city of the corresponding carbon derivatives. The oxygen

basicity in both species declines with the increasing number

of ethoxy groups which is consistent with the inductive

effect (-I) of the ethoxy group. The smaller decrease in

the ethoxysilane series was ascribed to the lower transmis-

sion of the ethoxy electronic effect through the silicon and

was argued (50) to be a result of both inductive (+1) and

(d-p) r effects.

The basicity of silyamines and the corresponding carbon

analogs was measured from the observed shifts in the C-D

stretching frequencies of base and CDC13 (51). The results

showed that the silylamines have a lower basicity and in

fact (Me3Si)3N had no effect on the C-D stretching frequen-

cies at all; these results suggest that the nonbonding elec-

tron pair on nitrogen is engaged in (d-p)w bonding with

silicon. Another study found the basicities of the com-

pounds R 3Si(CH2) NH3 to be anomalously low when n is zero

(52). An explanation for this observation is delocalization

of the nitrogen lone pair by (d-p)7 bonding. Lastly, West

and coworkers (53) determined the thermodynamic constants

for hydrogen bonding of phenol to several siloxanes and

alkoxysilanes, as well as to alkythiosilanes, by variable

temperature infrared spectroscopy. The results indicated

that the siloxanes are weaker bases than the corresponding

ethers and the alkylthiosilanes are also significantly less

basic than the corresponding dialkyl sulfides. These re-

sults are consistent with (d-p)7 bonding between the Si-N

and Si-S bonds.

Quantum Chemical Calculations

Quantum chemical calculations can possibly be useful in

the determination of relative amounts of (d-p)7 bonding.

Parkanyi and coworkers (25) employing the Pariser-Parr-Pople

(PPP) SCF method calculated 7-bond orders of 0.325 and 0.346

for the C-N and Si-N bond, respectively, for tetramethyl

N,N'-diphenylcyclodisilazane. Using the PPP SCF method,

Perkins (54) calculated a (d-p)T bond energy of 67 kJ per

Si-N bond in trisilylamine.

PPP and Del Re calculations were carried out on

PhNMe2, (Me3Si)2NPh, and (Me2HSi)2NPh (55) and the relia-

bility of the results was evaluated by comparing calculated

ultraviolet spectra and dipole moments with experimental

values. The results of these calculations (55) indicated

that (d-p)h bonding between silicon and nitrogen reduces

the Ph-N7r-bond order from 0.346 in PhNMe2 to 0.275 in

(Me3Si)2NPh and (Me2HSi)2NPh.

Quantum calculations were reported on the effects of

silyl substituents on ethylene and benzene. Horn and

Murrell (56) employing ab initio SCFM calculations on ethylene,

vinylsilane, and allylsilane found that inclusion of d orbitals

in the basis set gave a better interpretation of the photoelec-

tron spectra than their exclusion. Similar results were found

by McLean (57) for the CNDO/2 calculation of the photoelectron

spectra of phenylsilane and trimethylphenylsilane. Zeeck (58)

showed that the inclusion of silicon d functions in ab initio

SCFM calculations gives a quantitatively more accurate descrip-

tion of the bathochromic shift in the U.V. spectra produced by

silicon substitution for vinylsilane and propene than does the

exclusion of d orbitals. All three groups agree that (d-p)7

interactions are significant.

A CNDO/2 calculation of the photoelectron spectrum of

Me2Si(SMe)2 showed that (d-p)w bonding is significant (59).

This result is consistent with a recent ultraviolet spectro-

scopic study (60) of ring and open chain compounds containing

S-Si and S-Si-S bonds in which the n-a* transition showed a

blue shift as compared with that in the analogous carbon

compounds. The blue shift was ascribed to a (d-p)7 interac-

tion between the lone electron pair of sulfur and the

d orbitals of silicon (60).

NMR Spectroscopy

It has been shown both experimentally and theoretically

that chemical shifts of 13C (61-64), 14N or 15N (65-68),
170 (69-72), and 29Si (73-75) can be strongly influenced both

by inductive effects and by the n-electron density at the

observed nucleus. Although the theory of nuclear shielding

is well established, it is not always possible to quantita-

tively relate observed chemical shifts, especially those of
29Si, to theoretical concepts because of the complexity of

the factors involved. The shielding constant of a nucleus

can be presented as a sum of local and long range effects (76):

a = a + Cy + [1]
=loc loc [

where the local contribution is further divided into a dia-
magnetic term, dloc and a paramagnetic term, a The long

range effects (a') are contributions from other atoms in the

molecule and include such factors as electric field, magnetic

anisotropy and dispersion effects. The local diamagnetic

term depends on the electron density at the nucleus and is

given by Lamb's formula (77):
d 02/1m)-l>
aoc = (1e /212m7) EP..i [2]
i i

where i0 is the permeability constant, e and m are the mag-

nitudes of the electronic charge and the electronic mass,

respectively, and P.. is the charge density in the atomic

orbital i which is at an average distance of r. from the
nucleus. Therefore, aocd is influenced by substituent

electronegativity; substitution by an electronegative atom

or group causes a reduction in the electron density at the

nucleus in question, resulting in a downfield shift of the

resonance. The local paramagnetic term may be written (78):

cp = (-1e2f2/67m2AE)[ P + D [3]
loc 0 p u dr-3du

where AE is an average excitation energy (AEE), and

d are the mean inverse cubes of the distance of the val-

ence shell p and d electrons, respectively from the nucleus,

and P and D represent the unbalance of the p and d electron

populations, respectively. The last terms, Pu and D include

elements of the charge density and bond order matrix.

Problems in theoretical calculation of chemical shifts

arise because of the need to employ a number of simplifying

assumptions. Firstly, it is usually assumed that local ef-

fects dominate in nuclear shielding. This assumption is not

always valid; for example, in nitrogen shielding, although
d is negligible, up can be appreciable, parti-
non-loc non-loc
cularly in cases of multiple bonding (79,80). Secondly,

although it is generally assumed that changes in op are
dominant for heavier nuclei as compared to loc' this assump-

tion has been strongly challenged (81-83). For nitrogen

shielding, ad is approximately constant for nitrogen
nuclei in a variety of molecular environments (84,85).

Thirdly, although most calculations of nuclear shielding

have employed the AEE approximation, this method is only

successful in accounting for gross chemical shift trends in

series of closely related molecules. Furthermore, a CNDO/2
calculation of 29Si shielding in Me 4-n SiX compounds has

shown the assumption of a constant AE in ap to be in-

valid (86). Fourthly, for 29Si chemical shifts, there is

the additional problem as to whether the d-orbital term can

be neglected since the d-orbital term becomes important if

there is (d-p)r bonding. Consequently, it is apparent from

the preceding discussion that a quantitative theoretical

treatment of nuclear shielding is often impossible because

of oversimplification of theory. Therefore, in most cases

chemical shifts can only provide a qualitative interpreta-

tion of bonding.

Only two theoretical interpretations of 29Si shielding

have appeared in the literature. Both approaches neglected

all the other terms in Eq. [1] except ao which was used
as given in Eq. [3] with the AEE approximation; neither con-

sidered the d orbitals of silicon.

The theory developed by Engelhardt and coworkers (87)
was able to qualitatively predict 29Si chemical shifts in
compounds of the type Me 4-n SiX by calculating relative

paramagnetic screening constants, a*. In this model, only

the P term in a with AE held constant is considered.
p u oc
In addition, the a bonding structure was described by four

localized orbitals with tetrahedral bond angles on silicon.

In order to be able to predict experimental 29Si chemical

shifts, an empirical correction factor, f, must be included

in the determination of a*. A mean deviation of 22.7 ppm
was obtained between calculated and observed values of 29Si

chemical shift. Although Engelhardt and coworkers concluded

that (d-p)r interactions are negligible, it is interesting

to note that the X substituents most likely to iT bond show

the largest deviation between calculated and experimental
results. In addition, a comparison of the 29Si shielding
results. In addition, a comparison of the Si shielding

data of trimethylsilyl derivatives with the 13C shielding data

of the analogous t-butyl derivatives shows that the trend with

electronegativity predicted by Engelhardt's theory is followed
13 29
in the 13C series but not in the 29Si series. An angular cor-

rection factor was later introduced to account for rehybridi-

zation of the silicon atom for silanes with substituents of

electronegatives significantly different from that of carbon

(88). Wolff and Radeglia (89) qualitatively confirmed this

model by CNDO/2 calculations for which the best results were

found by neglecting d orbitals.

Ernst and coworkers (90) adopted the approached used by

Letcher and Van Wazer (91,92) for 31P chemical shifts to

explain the different signs of the slopes observed in plots
of 29Si vs. Hammett a constants of X for a series of aryl-

silanes XC6H4SiY3; the signs of the slope depended upon the

nature of Y. This theory considers only changes in oc'

assumes the ratio /AE is constant, neglects (d-p)7

interactions, and derives chemical shift contributions from

p electrons as a function of substituent electronegativities.

Ernst successfully predicted the observed sign of the slope

of these plots but could not predict the electronegativity of

Y at which the sign changes; therefore, these results must be

considered qualitative. In addition, Ernst was able to use a

polynomial electronegativity summation over all the substitu-

ents in R R IIR IIIR IVSi compounds to generate the observed

U-shaped relationship between silicon-29 shielding and the

sum of substituent electronegativity. The calculations

generally reproduced the observed 29Si shielding to within

10 ppm. Unfortunately, Ernst and coworkers did not show how

this theory can account for differences between the trends in
29 ad13
29Si and 13C shieldings of the analogous carbon derivatives.
One of the earliest qualitative models of Si shielding

was proposed by Lauterbur (93) to account for the observed 13C

and 29Si shieldings in the series Me nM(OEt) (where M = C, Si).

Lauterbur observed that as n increased the 13C resonance was

shifted downfield progressively while the 29Si resonance exhib-

ited a "sagging" pattern with the opposite behavior (Fig. 1-1).
Lauterbur's model views 29Si shielding as a result of inductive

effects which should be similar for both nuclei and (d-p)7 in-

teractions in the silanes. Therefore, the significant differ-

ences in substituent effects between the two nuclei were

attributed to (d-p)7 bonding which increases the shielding

of silicon by increasing the electron density on the silicon


Hunter and Reeves (94) further elaborated on Lauterbur's

model by demonstrating graphically a linear dependence of the
29Si chemical shifts of Me3SiX compounds upon the electronega-

tivity of the X substituent, XX. The results showed that the

most shielded silicon nuclei appear in compounds with the least

electronegative substituents and that an additional shielding

effect is operative for X = N, 0, and F. Since these substitu-

ents have the highest capabilities to form 7 bonds, this addi-

tional shielding effect was ascribed to (d-p)7 bonding. Harris

and coworkers (95,96) in a more extensive study of trimethylsilyl

,, --40 8Si
8\_ / -30

60- / --20
I --
80 0
90 I
100- l20
LJ -I --I --I --- 1 20
0 I 2 3 4
i .

Figure 1-1. 13C and 29Si chemical shifts in the series
Me4_n M (OEt)n(where M = 13C, 29Si) (93). Shifts are in
ppm relative to Me4Si.

compounds found that 29Si chemical shifts correlate poorly with
substituent electronegativity (Fig. 1-2). Since 6 29Si does not

vary smoothly with X as Engelhardt's theory (87) predicts,

Harris and coworkers (95,96) suggested that his theory is

oversimplified. In addition, Harris found no correlation be-
29 sidigad13
tween 29Si shielding and 13C shielding of the trimethylsilyl

group. Rakita and Worsham (97) reported 13C chemical shift

data of Me nSiX compounds and found that a replacement of

a methyl by a first-row donor atom (C, N, 0) leads to an up-

field shift of the remaining methyls, regardless of the pres-

ence or absence of unshared pairs or 7-type orbitals on the
atom in question. These results suggest 29Si chemical shifts

in trimethylsilyl compounds depend on other factors beside

simple inductive effects including (d-p)r bonding.

Data showing the effects of multiple a substitution on
29Si shieldings for Me 4-n SiX compounds are presented in Fig.

1-3 and Fig. 1-4. Data are from ref. 74 and references

therein; data for X = Cl, CH2CH = CH2, CH = CH2, and H are

from refs. 98, 99, 100, and 98, respectively. Figure 1-3

shows the "sagging" pattern observed for electronegative

substituents. This "sagging" pattern has been attributed

to opposing (d-p)7 bonding and inductive effects for these

derivatives (93,94,101-104). The deshielding of the 29Si

nuclei with increasing n in the series Me nSi(SMe) is
4-n n
ascribed to dispersion forces and diamagnetic anisotropy of

the C-S bond (102). In contrast to Fig. 1-3, no "sagging"

pattern is observed for the substituents shown in Fig. 1-4.

o F

0 Cl


o SMe


o Me

o OEt

SOSi (OSiMe3)3

N(SiMe3) 2

0 H
0 SiMe3


Figure 1-2. Plot of 29Si
electronegativity, X, for

2.8 3 T\

T .0

chemical shifts vs. substituent
trimethylsilyl compounds, Me3SiX


Me nSiXn






Figure 1-3. 2"Si chemical shifts (in ppm relative to Me4Si)
for methylsilanes Me4-nSiXn.






Me nSiX
4-n n

S--------- CH CH=CH
0 2 2



f -M


f SiMe3
1 T- 1 3


Figure 1-4. 29S chemical shifts (in ppm relative to Me4Si)
for methylsilanes Me_~in

It has been suggested that this is caused by an absence of

opposing (d-p)7 bonding and inductive effects for these

derivatives (74).

Several groups have reported NMR studies of alkenylsilanes.
13 29
Schraml and coworkers (105) measured both 13C and 29Si chemical

shifts of a series of Me- X Si(CH 2) CH=CH2 (where m = 0, 1).
3-n n 2 m 2.
Their findings showed the 29Si nuclei in vinylic compounds are

shielded by several ppm more than in the corresponding ethyl

or methyl silane. In the allyl compounds, silicon is less

shielded and the olefinic carbons are more shielded than in

the vinyl silanes. These results were attributed to delocali-

zation of the T electrons of the vinyl group toward the silicon

atom; for allyl compounds, this delocalization is decreased

when compared to the vinylic compounds. Schraml also noted

that this delocalization of 7 electron density could be

assigned equally to (d-p)r bonding or to hyperconjugation

(a-r interaction). Pitt (106) from the results of CNDO/2

calculations of compounds containing Si, Ge, and Sn claimed

hyperconjugation is dominant over (d-p)7 bonding for delocali-

zation of i electron density in bonds containing Si, Ge, and

Sn. Since unfortunately, the NMR experiment cannot distinguish

between the two effects and the prevailing theories support the

(d-p)7 bonding model over the hyperconjugation model, all de-

localization phenomena involving silicon will be ascribed to

(d-p)ir bonding in this paper. In the series Me4-nSi(CH=CH2)n

(where n = 0 to 4), the 29Si nuclei are shielded about 6.5 ppm

per vinyl group, the protons of the vinyl group in the

organosilane are deshielded with respect to the vinylic pro-

tons in the carbon analog, and the a-carbon is shielded and

the B-carbon is deshielded as compared to the corresponding

carbons in the carbon analog (100). These results were attrib-

uted both to a mesomeric (-M) effect from the vinyl group to

the Si atom and to the electron-supplying inductive effect (+1)

of the Me Si group. A similar interpretation of the observed
13C shieldings of the vinylic carbon was offered in the series

CH2 = CHMX3 (where M = C, Si) (107); the results showed that

the a-carbon is shielded and the s-carbon is deshielded with
respect to the hydrocarbon analog. Lastly, a 29Si NMR study

(108) indicated that the phenyl derivatives are shielded in

comparison with the corresponding benzyl derivatives and that

both phenyl and benzyl silanes are shielded with respect to

the methyl derivatives. These results were discussed in terms

of (d-p)ir bonding with a minor contribution to shielding from

ring current effects.

Scholl and coworkers (109) were one of the first groups
to note the increased sensitivity of 29Si chemical shifts to

substituent effects in X3SiOR as compared to X3SiCH2R

compounds. In a later study of silicon shielding in com-

pounds containing a Me3SiOC fragment, Schraml and coworkers
(110) found the large sensitivity of 29Si shielding to sub-

stitution is comparable to substituent effects of multiply

bonded carbon. Schraml and coworkers (111) in a recent study

found the sensitivity of silicon shielding to substituent

effects (change in R) in Me4_n-mCl Si(OR) decreases with

increasing numbers of chlorines or oxygens bonded to silicon.

This result was interpreted in terms of competing (d-p)h back-

bonding to silicon from both chlorine and oxygen.
Harris and Kimber (103) reported the 170 chemical shifts

of methylethyoxysilanes which are shown in Table 1-4. The 170

shielding in the series Me Si(OEt)4n shows a U-shaped depen-

dence of 170 shielding upon n with a maximum for n = 2. These

results were interpreted in terms of the scheme proposed by

Jaffe (112) in which the effectiveness of 7r-bonding in reduc-

ing electron density at oxygen decreases as the number of
alkoxy groups increases. However, comparison of the 0

shielding of methylethoxysilanes with the corresponding carbon

analogs (113) (see Table 1-4) shows the silicon derivatives to

be more shielded. This implies the 170 shielding of siloxanes

is dominated by a strong inductive (+1) effect of silicon

with only a small contribution from (d-p)7 effects. A recent

determination of the 170 shielding of hexamethyldisiloxane

(43 ppm) (114) and di-t-butyl ether (90 ppm) (114) also

further supports this interpretation.

Table 1-4. Comparison of 17 Chemical Shifts of Siloxanes
and the Corresponding Ether.'

Siloxane 6 170/ppm Ref. Ether 6 017/ppm Ref.

Me SiOEt 10 (103) Me 3COEt -- --

Me2S (OEt) 2 25 (103) Me2C(OEt) 2 52.0 (113)

MeSi(OEt) 3 21 (103) MeC(OEt) 3 58.5 (113)

Si(OEt)4 9 (103) C(OEt)4 56.5 (113)

Me 3SiOSiMe3 43 (114) Me3COCMe3 90 (114)

aAll shift references to external H 170.

bMe = CH3 and Et = CH2CH .*

Several groups have reported NMR studies of silylamines.

Jancke and coworkers (115) measured the 29Si shielding of 17

silylamines; their results showed that increasing the number

of nitrogens bonded to silicon results in increased shielding.

Although no explanation was offered for this observation, it

is apparent that (d-p)n bonding is responsible for the higher

shielding. In a recent 15N and 29Si NMR study of the Si-N

bond in silylamines, Filleux-Blanchard and Nguyen Dinh An

(116) observed that increasing the number of chlorines bonded

to silicon in N,N-dimethylaminosilanes, N,N-diethylaminosilanes,

and N,N-diisopropylaminosilanes results in higher shielding of

the 29Si nuclei while the 15N nuclei become deshielded. These

results were attributed to increased nitrogen pair delocaliza-

tion which is enhanced through inductive effects (-I) of the

chlorine bonded to silicon. The first nitrogen NMR

investigation of the Si-N bond determined 14N chemical shifts of

six silylamines (117). The results, which were interpreted in

terms of the (d-p)7 bonding model, showed that an increase in
the number of alkoxy groups attached to silicon increases 14N

shielding; hence, the delocalization of the lone pair electrons

of nitrogen is decreased. N6th and coworkers (118) in a latter
14 11
study by 14N and lB NMR on 39 silylamines and 19 silylamino-
14 ad11
boranes interpreted both the 14N and B shieldings in terms
of the Si-N (d-p)ir bonding model. Comparison of 14N shieldings

of (Me 3Si) nNR3-n with HnNR3-n for n = 1, 2, 3 shows a deshield-

ing of the 14N resonance upon introduction of a trimethylsilyl

group. This observation was accounted for by contributions from

the well-known @ effect and (d-p)7 bonding. Increasing the num-
ber of chlorines bonded to silicon deshields the 14N nuclei.

Although N6th and coworkers (118) attributed this result to the

-I effect of the chlorine and to better (d-p)r overlap, it is

apparent from a theoretical viewpoint that the inductive effect

of chlorine will increase shielding of the nitrogen nucleus, and

hence (d-p)r bonding alone is responsible for this deshielding.

The high field shifts of the methoxysilylamines were interpreted

by N6th and coworkers (118) in terms of a weakening of the Si-N

bond which raises the electron density of the nitrogen through

the resonance of the methoxy groups. In addition, the

Me-6-Si = N Me-6 = Si-N
deshielding of the 14N nucleus in the series Me3SiNMe2 (-374

ppm), (Me3Si)2NMe (-370 ppm), and (Me3Si) 3N

(-344 ppm) was ascribed to increased delocalization of the

nitrogen lone pair of electrons. Lastly, the results of a

more recent study of 14N shieldings of trimethylsilylamines
(119) showed the 14N shieldings are influenced by the strong

+1 effect of the trimethylsilyl groups (shielding) with a

small contribution from (d-p)7T effects (deshielding).

Nuclear spin-spin coupling like nuclear shielding can

provide information on bonding since the indirect coupling is

transmitted via the electronic system of the molecule. The

calculation and interpretation are usually based on Ramsey's

formulation (120). Within this framework the spin-spin

coupling constant, J(A-B), between nuclei A and B is

expressed as a summation of contributions arising from the

J(A-B) = J(A-B)C + J(A-B)0 + J(A-B)D [4]

contact, orbital, and dipolar interactions, respectively.

The contact interaction arises from electrons which are in

orbitals having a finite density at the nucleus. The orbital

term accounts for the interaction between the nuclear magnetic

moments and the field produced by the orbital motions of the

electrons. The dipole interaction arises from a dipole-

dipole interaction between nuclear and electron spins.

Self-consistent perturbation theory (SCPT) calculations

employing the INDO-MO method have been used most frequently

and successfully in the calculations of indirect spin-spin

coupling. The expressions of various contributions to J(A-B)

using this method were developed by Blizzard and Santry (121).

Excluding d orbitals, the orbital term is given by

2 < -3>pp []
J(A-B) 0 3 -2 0BYAYBrA P[PzBBByBzB xBzB

where pB equals the Bohr magneton, A and yB represent the

magnetogyric ratios of nuclei A and B respectively, and the

P's refer to the first order elements of the charge-density,

o and iT bond-order matrix. The dipolar term excluding d

orbitals appears as
3 2
J(A-B)D -0 20BYAB
D l10Tr

[2P -p P +3P cc P [6
A p B p zBzB xBxB- yByB -xBzB-3QyBzB 61

where the superscripts a and $ refer to the spin orbitals with

a and a spins respectively and Qa is the imaginary part of

the first order matrix. The contact term is given by

2 2 2 a[7]
J(AB~c= -- O0 BYAYBSA(0)SB(0)?SS
J(A-B)B A B 55B [7]

2 2
where S 2(0) and S (0) are the electron density in the s-valence
orbitals at the nucleus of atoms A and B respectively and
P is a diagonal element of the first-order a bond order
2 2
matrix of orbitals S and S. The integrals SA (0)S (0) and
A B* B
are usually treated as empirical parameters which
A p B p
are evaluated from the best least-squares agreement between

the calculated and experimental coupling constants.

From Eqs. [5] and [6], it is apparent that the orbital

and dipolar terms are only important for atoms having valence

p electrons. Consequently, these contributions vanish if the

spin-spin coupling involves a proton. The contact contribu-

tion is also usually the dominant term for other nuclei which

are singly bonded. Since both the orbital and dipolar terms

depend upon the expectation value of the inverse cube of the

radius of the orbitals on atoms A and B and elements of the

it-bond order matrix, in cases of multiple bonding, these terms

increase in magnitude and can dominate the spin-spin interaction.

For example, INDO calculations have shown that J( 15N=3 C) and
J( 15NE 13C) are dominated by non-contact interactions (122-124)

whereas J( 15N- 13C) is dominated by the Fermi contact interac-

tion (122-127). A similar situation holds for 1J(13C-13C)

which is dominated by the contact contribution, where the bond

is single, as compared to multiply bonded carbon in which there

is a significant contribution from the orbital and dipolar

terms (128,129). Thus, significant contributions from the non-

contact terms to J ( 29Si13C) and J ( 29Si 5N) may be taken as

evidence of the (d-p)fr bonding in those compounds.

Equation [7] indicates that when the contact term is domi-

nant, the spin-spin coupling will depend upon the amount of s

character in the single bond joining the nuclei. Thus, based

on an INDO-MO calculation of hybridization parameters, the

following relationship was derived for 1J 3CIH) (130):

J( 13C H) = 5.7 (% s) 18.4 [Hz] [8]

An empirical expression was found for IJ3CX 13CY) (131)

J(13 C xCY) 7.3 (% sx) (% Sy) 17 [Hz] [9]


while an expression for J( 15N 3C) based upon the finite per-

turbation technique employing the INDO-MO method was derived


J( 15N- 13C) = -94 (% sN) (% sC) [10]

However, some deviations from the linearity implied by Eq.

[10] are observed for singly bound J( 15N- 13C) values even

when the Fermi contact term dominates the coupling. These

deviations are attributed to the effects of lone-pair electrons

in orbitals with s character on the coupled nuclei (124).

The presence of lone pairs in orbitals with p character does

not interfere withn the linear relationships given by Eq. [10].

Similar deviations have been revealed through finite perturba-

tion (FP) (132) and sum-over-states perturbation (SOS) (133)

calculations employing the INDO-MO method of some 15N-15 N

couplings. Since nitrogen is invariably planar (lone pair

electrons are contained in a p orbital) when bonded to two or
1 29 .15
three Si atoms, these types of deviations between the J( 29Si N)

and s character of the Si and N atoms are not expected. However,

these deviations may be observed for a single silyl group bonded

to nitrogen, since available structural data indicates the nitro-

gen is not always planar in these types of molecules.

If the Fermi contact term is dominant, removal of electron

density brought about by substitution by an electronegative

substituent will increase the effective nuclear charge of the

nucleus in question. This augments the probability of the val-

ence s electrons for nuclear contact and hence, an increase of

J(A-B) should result. Consequently, correlations of J(A-B)

with substituent inductive parameters, i.e., electronegativity,

are often taken to indicate the dominance of the contact

contribution. For example, a linear correlation was found
1 29 13
between the J( 29Si3 C) of Me3SiX compounds and substituent

electronegativity, XX, (95,96) which is shown in Fig. 1-5.

The relationship is

1 29 13i
I J( 29Si 3C) = 8.00 XX + 31.6 [11]

In addition, Harris and coworkers (95,96) showed that there is

a linear relationship between J( 29Si3 C) in trimethylsilyl

compounds and J( 13C 13C) in the corresponding t-butyl deriva-

tives (134) (see Fig. 1-6):

I J(29 Si13C) I = 1.65 (j( 13C 13C)) 6.46 [12]

From these results, Harris concluded that the Fermi contact

term is the dominant contribution to spin-spin coupling and

that the silicon d orbitals do not participate to any great

extent in the Si-C bondings, or, if they do, their participa-

tion varies smoothly with the same substituent parameters that
1 29 13
determine the coupling. The variation in J( 29Si3 C) for the

series of methylethoxysilanes Me, Si(OEt) was investigated

by the same authors(103). It was concluded that since these

changes were too large to be attributed solely to s-character

variations, another mechanism, possibly (d-p)7 bonding, con-

tributes to the coupling.

Other groups have investigated the dependence of J( 29Si3 C)

upon various theoretical and experimental parameters. Levy and

coworkers (135) were the first group to note a rough



SMe o

0 Me

0 SiMe3

1 I
1.6 2.2

I. P-. 8 73

Figure 1-5. Plot of 1J(29Si13C)j vs. substituent
negativity in Me3SiX compounds (95,96).



OSi (OSiMe3)2

0 Cl



36 38

I 0

Figure 1-6. Plot of Ij(29Si1'3C) I in Me3SiX vs. 'J('C3C)
in MesCX compounds (95,96). The value of 1J(13C13C) for NH2
has been used with 1J(29Si"C) for N(SiMe3)2 and NHSiMe3 and
the value of 1J(13C13C) for OMe has been used with 1J(29SiI3C)
for OEt, OSiMe3, and OSi(OSiMe3)3.


CD -



0 C6H5
0 Me

proportionality between the magnitude of J( 29Si 3C) and the

s character of the carbon nucleus in several organosilanes

containing only Si, C, and H. Kovacervi-6 and Maksib' (136),

through calculations based on the maximum overlap approxima-

tion method, reproduced this trend and derived a relationship
1 29 .13
between J( 29Si3 C) of these compounds and the s character of

the bonded carbon and silicon atoms:

J( 29Si 13C) = 5.554 (% sC) (% Ssi ) + 18.2 [Hz] [13]

These calculations neglected silicon d orbitals and assumed

the Fermi contact term is dominant. The results showed that

the hybridization of the Si atom remains approximately sp.

Dreeskamp and Hildenbrand (137) investigated couplings to

silicon in tetravinylsilane as well as in some vinyl- and

chloro-silanes. They found J( 29Si3 C) to be proportional to

the s character of the orbitals forming this bond whereas
2 29
J( 29SiCH) depends on the hybridization of both the silicon

atom and of the intervening carbon atom.

Summerhays and Deprez (138), employing the FP-INDO method,
1 29 13
calculated J( 29Si3 C) of trimethylsilyl compounds and compared

their results with the experimental data of Harris and Kimber

(95). These calculations assumed the contact term is dominant

and the silicon d orbitals were neglected. Although the cal-

culated results generally reproduce the experimental data,

the largest deviations (ca. 10% lower than experimental values)

are for substituents (N, 0) most likely to (d-p)fT bond. Conse-

quently, it was concluded that the contact term alone is ade-

quate to reproduce the trends of J( 29Si 13C) and that inclusion of

silicon d orbitals in the basis set will improve the overall

agreement between calculated and experimental constants. Beer

and Grinter (139) using an sp basis set in FP-INDO calculations
1 29 13
calculated 20 J( 29Si3 C) values. Their results indicated that

the orbital and dipolar terms are of opposite signs and that

the combined non-contact contributions are less than 1% of the

observed coupling, and that inclusion of silicon d orbitals is

not required. However, since their data set excluded siloxanes

and silylamines which are compounds most likely to have (d-p)7

bonds, the latter conclusion may not be necessarily valid.

Lastly, in a recent calculation of 30 J( 29Si 3C) by SCPT-INDO
1 29 131
calculations, Duangthai and Webb (140) showed that J( 29Si3 C)

is dominated by the contact interaction. The orbital terms

are larger and the dipolar terms are smaller than the previ-

ously reported values (139); the combined non-contact contri-

butions are less than 4% of the observed coupling. The authors
1 29 13
concluded that the J( 29Si C) data are satisfactorily repro-

duced by calculations employing a sp basis set. Since, however,

only three siloxanes and no silylamines were included in their

data set, this conclusion is questionable.

Randall and Zuckerman (141), from observations of moderate

variations of J(15N H) in Me3M-NH-Ar (with M = C, Si, Ge),

concluded that there is no stereochemical evidence for (d-p)7

bonding in these compounds. Their conclusion was based upon

the Randall equation (142):

% s N = 0.34 1 J(15NH) [Hz] [14]

which was derived independently of the Binsch equation (143):

% sN = 0.43 J( 15N H) 6 [Hz] [15]

Equations [14] and [15] were established on the basis of tetra-

hedral and trigonal-planar nitrogen geometries. In a later

study of J( 15N H) in N-P, N-As, N-S, and N-Si compounds,

Cowley and Schweiger (144) measured J( 15N H) in (Me3Si)2NH and

(H3Si) 2NH. On the basis of Eqs. [14] and [15], these compounds

are predicted to possess sp hybridized nitrogen bonding

orbitals. However, electron diffraction studies (14,23) of

these compounds indicate the nitrogens are planar (sp ).

Therefore, it was concluded that the Randall and Binsch equa-

tions are not applicable when the atoms bonded to nitrogen have

very different electronegativities from that of carbon.

No investigations of the dependence of J( 29Si 15N) upon

theoretical or experimental parameters have been reported. To

this author's knowledge, there are only six published values
1 29 .15
of J( 29Si 5N); these are listed in Table 1-5. Nitrogen-15

enriched compounds were used for the determination of spin-

spin couplings in these compounds.

Table 1-5. Published Values of J( 29Si15N).

Compound 1 ( 29Si 15N)/Hz Ref.
ClCH2Si(OCH2CH2) 3N 1.5 (145)

(H3Si) 3N +6 (146)

trans-[PtI(PEt3)2{H2SiN(SiH3)2}] +12 (147)

H3SiNCSe +15.1 (148)

H3SiNCS +16.8 (148)

H3SiNCO +18.6 (148)

Dynamic (variable-temperature) NMR spectroscopy can be

useful in the determination of inversion and rotational barriers.

Overlap of d-p orbitals may be involved in the lowering of the

C-N rotational barrier between cisoid and transoid formation

from 9.49 kJ/mol in HC(O)NHCMe3 to 83.2 kJ/mol in

HC(O)NHSiMe3 (149). In this case the electron-donating

inductive (+1) effect of silicon should favor the resonance

shown below, thus increasing the rotational barrier.

0 SiMe3 0 +SiMe3
X/ 3 +/ 3
\ N C = N
/ \ H \H

A H NMR study (150) of eight silylamines including bis-

(Dimethylamino)methylphenylsilane and trimethylsilylaziridine

showed equivalent N-methyl and methylene groups at temperatures

as low as 145K, corresponding to a free energy barrier of less

than 33.6 kJ/mole. Many phosphinoamines, on the other hand,

show much larger barriers ranging from 61.3 kJ/mole for

PhPClN(i-Bu)2 to less than 33.6 kJ/mole for F2PNMe2 (151).

Even higher free energy barriers are found for sulfenamines

ranging as high as 74.7 kJ/mole for ArSN(Ph)CH2Ph (152).

These results suggest lone pair repulsion as the dominant

factor affecting rotational barriers. This explains the low

rotational barrier in silylamines, where there is only one

lone pair, and the increase in the free energy barrier in

going from phosphinaomines to sulfenamines, increasing the

number of interacting lone pairs from two to three.

In another H NMR study (153) of barriers to torsion

around Si-N bonds in aminosilanes, the free energy barrier to

rotation for N,N-diisopropylaminophenyldichlorosilane was

shown to be 37 kJ/mole at the coalescence temperature of 176K.

Spectra of other N,N-diisopropylaminosilanes showed the iso-

propyl doublets broadened and apparently coalescing in the

region of 173K 5K but upon further cooling, the spectra

remained broad and did not split out into discrete frequencies.

Spectra of N,N-diethyl- and N,N-dimethyl-silylamines showed no

broadening at temperatures above 170K. In a recent 13C NMR

study (116) of the barrier to rotation in N,N-dimethyl-,

N,N-diethyl-, and N,N-diisopropyl-silylamines, no splitting

of signals could be observed down to temperatures as low as

145K for the N,N-dimethyl- and N,N-diethyl-silylamines. A

free energy barrier to rotation of 32 kJ/mole and 37 kJ/mole

was measured for N,N-diisopropylaminodiphenylchlorosilane,

and N,N-diisopropylaminophenyldichlorosilane, respectively.

These results suggest the barrier to rotation in silylamines

is largely steric in nature with only a minimal contribution

from (d-p)fr bonding.

Summary of the Literature Findings

Although there is a considerable amount of evidence for

(d-p)ff bonding in Si-N and Si-O bonds and to a lesser extent

in Si-C (unsaturated carbon), Si-S, and Si-Cl bonds, the

(d-p)w bonding model is controversial. From available struc-

tural data, the bond lengths and angles of Si-N and Si-O bonds

seem to indicate (d-p)7r bonding; for Si-C, Si-S and Si-Cl bonds,

there appears to be little or no (d-p)ir bonding. For silyla-

mines, nitrogen is invariably planar when linked to two or more

silicons but for a single silicon nitrogen may or may not be

trigonal planar. Vibrational studies have revealed evidence of

(d-p)T bonding in Si-C, Si-N, and Si-O bonds. Measurements of

acidity and basicity by infrared spectroscopy showed that com-

pounds containing Si-N, Si-O, and Si-S bonds are weaker bases

than the corresponding carbon analog; this is attributed to

(d-p)r bonding. Quantum calculations provide evidence of (d-p)w

bonding in Si-C (unsaturated carbon), Si-N, and Si-S bonds

through calculations of (d-p)n bond orders and calculated
29 17 ad14
photoelectron spectra. Shieldings of 29Si, 170, and 14N or
15N of organosilanes have been interpreted qualitatively in

terms of inductive and (d-p)7 bond effects. The one-bond
29 13
29Si- 13C coupling constant has been shown to be dominated by

the Fermi contact interaction and the results have been inter-

preted in terms of an absence of (d-p)ir bonding. There are
1 29 15
too few data for J( 29Si 5N) to permit conclusions to be

drawn. Dynamic NMR studies have attributed the barriers to

rotation in silylamines to steric effects with a negligible

contribution from (d-p)w bonding.

Aim of the Investigation

The present NMR investigation was undertaken to obtain

more information about the nature of Si-C, Si-N, Si-O, and

Si-S bonds in organosilanes. Two types of organosilanes,

namely methylsilanes and silylamines, were studied to see

if nuclear shielding and spin-spin coupling could provide some

insight into the extent of (d-p)ff bonding in these compounds.

Since numerous trimethylsilyl compounds are readily available

commercially and these are the easiest compounds for which to

interpret substituent effects, most of the methylsilanes that

were studied were trimethylsilyl compounds. In the NMR study
1 29 13
of methylsilanes particular emphasis was placed on J( 29Si3 C)

in providing information of (d-p)T bonding. In addition,

since few values of spin-spin coupling to silicon-29 have

been reported, it would be interesting to relate these coup-

lings to substituent effects. Finally, through a H, 13C,
15 29
15N, and 29Si NMR study of silylamines, it was hoped that

information on the nature of the Si-N bond could be obtained.
1 29 15
Of particular interest was relating the J( 29Si 5N) to theo-

retical concepts since there is a paucity of data for this




The compounds (Et2N) SiCl4n (where n = 1 to 3),

(Et2N) SiMeCl3_n (where n = 1, 2), and (Me2N)nSiCl4n

(where n = 1, 2) were synthesized. The other organosilanes

were kindly donated by PCR Research Chemicals, Inc., Gaines-

ville, Florida, or were purchased from Petrarch Systems, Inc.,

Bristol, Pennsylvania. These materials were used without any

further purification. The solvents, acetone-d6, benzene-d6,

chloroform-d,, and dimethyl-d6 sulfoxide, were obtained from

Merck Sharp & Dohme Isotopes, St. Louis, Missouri.


Since chloraminosilanes are extremely sensitive towards

atmospheric moisture (154), all synthesis experiments involv-

ing these compounds were carried out under a blanket of dry

nitrogen in dried glassware. Benzene was dried byazeotropic

distillation and anhydrous ether was used without further

purification. These solvents were obtained from Fisher

Scientific Company, Fair Lawn, New Jersey. Methyltrichloro-

silane and tetrachlorosilane,which were generously donated by

PCR Research Chemicals, Inc., Gainesville, Florida, were dis-

tilled before use. Anhydrous dimethylamine and diethylamines

obtained from Eastman Kodak Company, Rochester, New York,


were used as obtained. Proton NMR spectra were used for pur-

ity determination. Mass spectral data obtained on the AEI

MS30 interfaced with a KRATOS DS-55 data system operating at

an ionising potential of 70 eV were used for structural


N,N-Dimethylaminotrichlorosilane. The procedure for

synthesis of this material illustrates the general method

for preparation of chloraminosilanes. In a three-necked,

1-liter flask equipped with condenser, pressure-equalizing

addition funnel, and magnetic stirrer were placed 94.4 g

(0.55 mole) of tetrachlorosilane and 300 ml of anhydrous

ether. The solution was cooled to -50C in a dry-ice/

methyl alcohol bath and treated with a solution of 50.0 g

(1.10 mole) of anhydrous dimethylamine in 50 ml of anhydrous

ether over a one hour period, causing formation of a volumi-

nous precipitate of dimethylamine hydrochloride. Filtration

through a sintered glass filter gave a white cloudy filtrate

which was concentrated to ca. 150 ml by distillation at atmos-

pheric pressure and fractionated through a 20 cm column filled

with glass helices. The product N,N-dimethylaminotrichloro-

silane is a white cloudy oil-like liquid, b.p. 45-46C (19

mm) which fumes violently on contact with air. The mass

spectral data for this compound is listed in Table 2-1.

Compounds prepared in an analogous fashion are listed

in Table 2-2. Two equivalents of Me2NH or Et2NH were used

for each Si-Cl bond to be replaced.

Table 2-1. Mass Spectral Data of N,N-Dimethylaminotrichloro-
















Relative Abundance















Assigned Formula

C2H5NSiCl35 37C2
C 2H 6NSi 35Cl 237Cl

C2H5NSi 35Cl 237Cl

C2H6 NSi35Cl3


C2H 6NSi35Cl37Cl

Si35C137Cl 2

SiC35Ci 2Cl37C l


Si35C l
C2 H6 3


Table 2-2. Chloraminosilanes Prepared.






(Et2N) 2SiCl2

MeSiCl 3

MeSiCl 3


Me 2NSiCl3

(Me2N) 2SiCl2


(Et2N) 2SiCl2

(Et2N) 3SiCl

Et 2NSiMeCl2

(Et2N) 2SiMeCl

B.p. [C(mm)]

45-46 (19)

73 (10)

55-60 (17)

50-60 (2-4)

115 (5)

60-70 (20-25)

95-105 (22-25)




Et 20





(gas phase)

(gas phase)






Mass spectra data which were obtained on each of the

synthesized chioraminosilanes confirmed the structure of the

isolated product. Facile interpretation of the mass spectral

data was achieved through the recognition of the relative

35 37
isotopic abundance of 35Cl and 37Cl in the ions. The promi-

nent fragmentation pathways for N,N-dimethylaminochlorosilanes

are loss of H, Cl, and NMe2; for N,N-diethylaminochlorosilanes,

they are loss of Me, Cl, Et2N, and loss of both Et and Me to

form a rearranged ion.

Sample Preparation

Since aminosilanes (154-157) and thiosilanes (158) are

extremely sensitive to atmospheric moisture, solutions of

these samples were prepared in a homebuilt Plexiglas dry-box

under a dry N2 (g) environment. Only solvents thoroughly

dried over molecular sieve were used. The 80% v/v solutions

were degassed by the freeze-pump-thaw technique and sealed

under vacuum into 12 mm o.d. NMR sample tubes. These sealed

NMR sample tubes were stored on dry ice. For the other organo-

silanes, the solutions were prepared outside the dry-box and

contained in unsealed NMR sample tubes.

All experiments for aminosilanes and thiosilanes were

performed on a Nicolet NT-300 spectrometer equipped with a

293-C pulse programmer and a Nicolet 1280 computer. Spectra

of H, 13C, 15N, and 29Si NMR were obtained at 300.07, 75.45,

30.41, and 59.61 MHz, respectively. These spectra were

obtained at ambient temperature, 22C. For 15N and 29Si NMR

spectra, sensitivity enhancement was obtained through the

INEPT (159), INEPTD (160), and SPT (161-164) pulsed experiments.

In addition to these techniques, a method was developed to de-

couple selectively a group of magnetically equivalent protons

which are also used for polarization transfer; this method is

discussed in Chapter III. These techniques transfer spin polar-

ization from protons to nuclei such as 15N via the scalar coup-

lings between the nuclear spins involved, e.g., J( 15N H).

Theoretically, one can expect an enhancement factor of

y( H)/y( 15N) = 9.9 for 15N spectra and an enhancement factor
of 5.0 for 29Si spectra. Further gains in sensitivity can be

expected within these methods, since faster pulse repetition

rates are possible due to the shorter relaxation times of

protons. The amplitude, yH2/27, of the H decoupler output

was calibrated by the standard off-resonance decoupling method

(165) using a sample of chloroform for the 46 to 122 MHz broad-

band probe and formamide for the 21 to 46 MHz broadband probe.

All samples were contained in 12-mm-o.d. sample tubes.

For the other organosilanes, the 13C NMR spectra were

obtained at 25.16 MHz on a Varian XL-100-15 spectrometer

equipped with a NIC-80 32K computer system, a Varian Gyrocode

decoupler, MONA multinuclear accessory, and 18-mm probe, at

25.00 MHz on a Jeol FX-100 spectrometer equipped with a switch-

able dual band H/ 13C 5-mm probe, or at 75.45 MHz on the Nicolet

NT-300 spectrometer described above. The spectra were obtained

at ambient temperature, 22C for the FX-100 and NT-300 and 26C

for the XL-100-15. The 29Si NMR spectra were obtained at

59.61 MHz on the NT-300 and 19.88 MHz on the XL-100-15

spectrometer. Sensitivity enhancement techniques were
29 1
employed for the 29Si NMR spectra. The H NMR spectra were

obtained on the NT-300 or the FX-100 spectrometers at 300.07

and 99.55 MHz, respectively.

Computer Calculations

Analysis of the non-first order spectra was performed

using the ITRCAL program provided by Nicolet for the Nicolet

1280 computer system. This program is adapted from the LAOCN3

program written by Bothner-By and Castellano (166). Least

squares analysis of experimental data was accomplished employ-

ing a program written in Basic for the Nicolet 1280 computer

system. The levels of significance reported in these calcula-

tions are those obtained from a two-sided t test and give the

probability that a correlation exists between the two



During the multinuclear NMR investigation of organo-

silanes, it was desirable to develop a method to decouple

selectively a group of magnetically equivalent protons which

are also used for polarization transfer. In addition, tech-

niques to suppress strong unwanted signals and to differenti-

ate between 13C multiplicities were developed. The discussion

of these modified selective population transfer experiments

and spin-echo J-modulation experiments are presented in this

chapter. Portions of this material have already been

published (167,168).

Modified Selective Population
Transfer Experiments

Selective population transfer (SPT), induced by the

application of one or more (161) selective iT pulses to H

satellite spectra followed by a nonselective "read" pulse,

has been widely utilized as a method of enhancing the sensi-

tivity of spin-coupled NMR signals for less-receptive nuclei.

Examples include magnetization transfer to 13C (162), 15N
(163), and 29Si (164) NMR spin multiplets. The assignment

and relative signs of coupling constants may be obtained

simultaneously from such experiments (169). The ordinary

SPT method does not, however, lead to a net magnetization

transfer and the multiple lines will appear with an antiphase

relationship. Thus, the total integral of the multiple is

zero (provided the native magnetization is suppressed) and

a spin decoupled spectrum can be recorded only following

suitably delayed decoupling and acquisition (170). An alter-

native procedure for obtaining noise decoupled SPT spectra

has been suggested by Harris and coworkers (171). In general,

the drawback of the SPT method is the high selectivity required

for irradiation in the hydrogen spectrum. However, this has

been overcome with the advent of the INEPT (159), DEPT (172),

and JCP (173) pulse sequences which apply hard, nonselective

pulses but require a spectrometer with 90 proton phase shift-

ing hardware. More recently methods have been devisedwhich

avoid this hardware requirement and thus allow magnetization

transfer experiment using nonselective pulses to be performed

on older spectrometers (174). Enhanced proton coupled as well

as fully decoupled spectra with multiplicity spectra may be

obtained using these methods (159,170,172-175) or variants of

these experiments (176).

In many cases selective decoupling is desirable to sim-

plify the frequently rather complex proton coupled spectra

resulting from an SPT (or INEPT) experiment. Especially

this is valid when the J coupling utilized for SPT is of

similar magnitude to other coupling constants of the observed

nucleus, thus resulting in overlap between positively and

negatively enhanced lines in the coupled spectrum. In

other experiments it is of interest to observe directly

weak resonances hidden by other intense NMR signals. For

example, in the observation of satellite spectra of lines

obscured by a solvent resonance, suppression of the main

signal is often required. SPT combined with difference

spectroscopy has earlier proven useful in such experiments

(177) although at the expense of a loss in signal-to-noise

ratio. In the next few pages, some new and improved tech-

niques which extend the SPT method to cope with such experi-

mental situations are described.

Selective Population Transfer with
Selective Proton Decoupling

Two different situations must be considered when per-

forming selective decoupling experiments in a spin system

for which one of the spins is observed using the SPT method

or other polarization transfer techniques. In the first

case, the selective decoupling field is applied for the

elimination of coupling to a proton (or group of protons)

which is not involved in the polarization transfer process.

Experiments associated with this case are readily performed

using a second proton frequency source with selective irra-

diation applied continuously during the time of the experiment.

Alternatively, the proton irradiation source used for

polarization transfer may also be applied for the selective

decoupling (on during acquisition only) provided computer

control of power level/frequency changes is available for

this channel. This has been applied in an INEPT experiment

using a state-of-the-art spectrometer (178). Since these

experiments are in general straightforward, they will not be

further discussed. In the second situation, the protons

used for polarization transfer are also selectively decoupled;

this type of experiment is more difficult than the first situ-

ation, as is evident from the following discussion.

The observation of NMR spectra of low-y nuclei via polar-

ization transfer is especially advantageous when a group of

magnetically equivalent protons can be utilized for the

transfer. The reason is that large improvements in S/N are

obtained both for the multiplets in a coupled spectrum (162)

and for a proton decoupled spectrum (160). Examples of com-

monly encountered spin systems with groups of magnetically

equivalent protons and low-y X nuclei include molecular

fragments such as -29Si(CH3) 15N-CH, and -13C(CH).

However, in many such cases the J coupling between the

observed X nucleus and the group of protons used for polar-

ization transfer is similar to the magnitudes of the coup-

ling constants between the X nucleus and other protons in

the spin system. This gives rise to severe overlap between

positive and negative enhanced lines in the coupled spectrum.

Thus spectral analysis is rendered almost impossible not only

because of the complexity of the spectrum, but also because

the antiphase nature of the lines makes it difficult to

obtain exact line positions. Obviously, selective decoup-

ling of the magnetically equivalent protons, following their

use for polarization transfer, would be very useful for such

systems. However, application of the usual refocusing

technique (with two 180Q pulses) (175) to the coupled anti-

phase spectrum followed by selective decoupling does not work

because J modulation from other protons in the spin system

leads to extensive phase errors. Although simplification

of such polarization transfer enhanced spectra by selective

decoupling has not previously been demonstrated experimentally,

it has been suggested (179) that a selectively decoupled spec-

trum may be obtained by a mathematical processing of the

coupled spectrum.

The pulse sequence shown in Fig. 3-1 was applied for

selective decoupling of the group of protons used for polar-

ization transfer in an SPT experiment (i.e., case (b)).

The delay T1 following the polarization transfer is used

for refocusing of the antiphased multiple before the

selective decoupling field is turned on. This delay is

optimized according to the number of equivalent protons

(N) used for the polarization transfer, i.e., to give a

maximum for the term sin (rJT1){cos rJT1)1} (160,175,176)

and is shown in Table 3-1. J-coupling evolution arising

from scalar coupling between the X nucleus and other spins

(e.g., protons) within the system also occurs during the

period TI. The resulting defocused multiple lines, how-

ever, are focused after a delay T2 by means of the 180 X

pulse and by allowing for an addition T 2 delay before the

acquisition. Obviously, the condition T2 a T1 must be


t | Setective Decouping

90g 180,
--2 -- Acq

Figure 3-1. Pulse sequence for selective decoupling of a
group of magnetically equivalent protons used for polari-
zation transfer in an SPT experiment.


0 I H


Hi.,/ 'Si(CH,)3

Figure 3-2. 29Si NMR spectra (four scans each) of vinyltri-
methylsilane (75% v/v in acetone-d6). (a) Normal proton
coupled spectrum without NOE. (b) SPT-enhanced coupled
spectrum with polarization transfer from the trimethylsilyl
protons using yH2/2ir = 2.5 Hz, T = 0.2 sec (yH2T = r), and
applying the SPT r pulse to the high-frequency 29Si satel-
lite in the 1H spectrum (2j(29Si-CHa) = 6.63 Hz).
(c) SPT-enhanced coupled spectrum with selective decoupling
of the trimethylsilyl protons obtained using the sequence
shown in Fig. 3-1 (TI = 16.3 msec, T2 = 25 msec, and a
selective decoupling amplitude yH2/27 = 62 Hz). The SPT
conditions are similar to those in (b).

Table 3-1. Values of TI to Give the Maximum Decoupled
Enhancement as a Function of the Number of Scalar
Coupled Protons.

N 1 2 3 4 6 8 9
T 0.50 0.25 0.196 0.167 0.134 0.115 0.108

aIn units of J.

A spectrum resulting from the sequence in Fig. 3-1 is

enhanced with all the advantages of proton polarization

transfer as well as being simplified to permit more facile

interpretation. Furthermore, the selectively decoupled

spectrum is easily phased to give a spectrum with all mul-

tiplet lines in positive absorption. This is easily demon-
strated for the SPT-enhanced 29Si NMR spectra of vinyltri-

methylsilane shown in Fig. 3-2. The spectra were obtained

without (Fig. 3-2b) and with (Fig. 3-2c) selective decoupling

of the trimethylsilyl protons which were also used for SPT.

Clearly, it would be very difficult, if not impossible, to
29 1
extract values for the 29Si- H couplings involving the vinyl

protons from the polarization transfer enhanced spectrum re-

corded without selective decoupling (Fig. 3-2b). Of course

a precise determination of coupling constants from spectra

such as that recorded in Fig. 3-2c requires a complete analy-

sis based on the double resonance Hamiltonian. Thus the

strength of the selective decoupling field (yH2) and its fre-

quency offset from other resonances of other protons within
the spin system must be known. The 29Si- 1H coupling constants
the spin system must be known. The Si- H coupling constants

determined from analysis of the spectrum in Fig. 3-2c are
2 29 3 29 3{ 29
J( 29Si-Hl) = 6.42 Hz, 3J( 29Si-H2) = 15.26 Hz, and 3J( 29Si-H3)

= 8.59 Hz. The defocusing effect arising from the 29Si- H

(vinyl) J-coupling evolution is illustrated in the spectra

of Fig. 3-3 where the 180 refocusing pulse of Fig. 3-1 has

been omitted. From these experiments the length of the 2T2

delay has been varied; this results in a series of spectra

with different phase properties which can be predicted from

the J(29 Si-H) couplings.

Selective Population Transfer with
Suppression of Unwanted Signals

A number of experimental methods have been introduced

for the direct observation of weak NMR resonances hidden

below other strong signals, e.g., solvent peaks or "parent"

signals in satellite spectra. These techniques include dif-

ference spectroscopy combined with various double resonance

methods (177,180) (e.g., SPT difference spectroscopy (177,

181,182)), pulse sequences utilizing the creation of double

quantum coherence (183-186), and other pulse sequences for

the observation of X-nuclei satellite spectra in 13C NMR and
H NMR such as double SPT and reverse INEPT (161,187,188).

However, the methods based on "ordinary" difference spectros-

copy suffer from the disadvantage that subtraction of an

"unperturbed" spectrum in alternate scans will degrade the

S/N ratio by a factor of 2 compared to the methods which

utilize a "perturbed" spectrum in each scan. Furthermore,

these subtraction techniques involve acquisition of the

30 Hz

i i i' I
33 ms 9ms 6 ms 132 ms 198ms

Figure 3-3. SPT-enhanced proton-coupled 29Si spectra
(four scans each) of vinyltrimethylsilane with selective
decoupling recorded similar to that in Fig. 3-2c but
without the 180 refocusing pulse of Fig. 3-1. The
total 2T2 delay has been varied for each experiment and
is given below the individual spectra. These have been
phase corrected so as to display the low-frequency peak
in positive absorption.

strong unwanted signals leading to incomplete cancellation

of these unwanted signals and possible dynamic range


As an improvement of the SPT difference spectroscopy

technique a method was devised which proved useful for sup-

pression of strong signals with the simultaneous observation

of weak resonances hidden by the strong peaks. The basic

pulse sequence, in which SPT f pulses are applied alter-

nately to the low- and high-frequency satellites (e.g.,
13C satellites) in the H spectrum with a concurrent 180
phase shift of the 90 13C "read" pulse, may be written

SPT, 180 (I1H,v)-90 ( 13C)-Acq.-PD-SPT,

180 (H, +J)-90 (13C)-Acq.-PD- [E]

where PD is a pulse delay. Thus, the SPT-enhanced signals

add for each data acquisition. Although the strong signals

are acquired in each FID using sequence [1], all resonances

arising from native 13C equilibrium magnetization including

the unwanted strong signals will be canceled or greatly

attenuated in the final spectrum. Figure 3-4 demonstrates

the frequency-alternating SPT T pulses of sequence [1] to

a mixture of CH3CN (10% v/v) and CD3CN (90% v/v; 99% deu-

terium enriched). For this sample the proton decoupled
13C spectrum for the methyl region without NOE (Fig. 3-4a)

and the proton coupled SPT enhanced -1:-1:1:1 quartet

spectrum of CH3CN obtained using sequence [l](Fig. 3-4b)

illustrate the degree of suppression achieved for the



a b c

Figure 3-4. 13C NMR spectra for the methyl region of a
mixture of CH3CN (10% v/v) and CD3CN (90% v/v); 99%
deuterium enriched). (a) Normal proton-decoupled spectra
(32 scans); the resonance for the protio compound is
marked with an asterisk. (b) SPT-enhanced proton-
coupled spectrum (32 scans) with suppression of the
CD3CN solvent signal obtained using the pulse sequence [i].
(c) Proton-decoupled version (four scans) of the spectrum
in (b) obtained by introduction of refocusing delays before
the data acquisitions in sequence [1].

250 Hz

Figure 3-5. SPT-enhanced proton-coupled 13C NMR spectrum
(32 scans) for the mixture of CH3CN (10% v/v) and CDaCN
(90% v/v; 99% deuterium enriched) obtained using the
pulse sequence [2]. The SPT frequency has been shifted
slightly to lower frequency compared to the experiment in
Fig. 3-4b in order to optimize the enhancement of the six
observed resonance peaks (marked with asterisks) for the
CHD2CN isotopomer present in the CD3CN solvent (see text
for further discussion).

CD3CN solvent resonances by this simple method. The proton

decoupled version of the spectrum in Fig. 3-4b is shown in

Fig. 3-4c. This spectrum was obtained by extending the

sequence [1] with refocusing delays following the 90'
13C "read" pulses and applying broadband proton decoupling

during data acquisition.

A significant improvement in the suppression of the

CD3CN solvent peaks, as compared, for example, to the

residual signals observed in Fig. 3-4b, may be achieved

by preceding the SPT T pulses in sequence [1] by 90 13C

pulses for which the phase is alternated between positive

and negative x axes. Thus, this improved sequence, which

may be written

90x(13 C)-SPT, 180 H,)-90+x(13C)-AT-PD-900x ( 13C)-SPT,

1800(IH, V+J)-90 ( C)-AT-PD- [2]

avoids the acquisition of strong unwanted signals. Further-

more, the x axis phase cycling of the 90( 13C) pulses pre-

ceding the SPT T pulses ensures almost complete elimination

of any residual unsuppressed resonances arising from non-

ideal 90 (13C) pulses. The improvements of sequence [2]

over [1] are illustrated in Fig. 3-5 employing the same

sample as used for the spectra in Fig. 3-4, i.e., a mixture

of 10% v/v CH3CN and 90% v/v CD3CN. In addition to the

observation of the proton coupled -1:-1:1:1 quartet 13C

spectrum for the methyl carbon in CH3CN and almost perfect

suppression of the CD 3CN solvent peaks, the spectrum in Fig.

3-5 also demonstrates that the proton coupled 13C spectrum

for the CHD2CN isotopomer (six peaks marked with asterisks),

originating from the -99%-deuterium-enriched CD3CN solvent,

may also be uncovered. The different intensities (pairwise

negative/positive) for the six peaks for this isotopomer are

related to the relative sign for the J( 13C-D) and 2J( IH-D)

coupling constants and to the frequency of the SPT t pulse

(169). In the actual experiment of Fig. 3-5 the SPT fre-

quency was positioned at the second highest frequency line

in the ten line 1H spectrum for the 13CHD2CN isotopomer

(a doublet of pentets; J( 13C-H) = 136.2 Hz and 2J(1 H-D)

S2.6 Hz) as is shown in Fig. 3-6. Thus the spectrum (Fig.

3-5) shows that J( 13C-D) x 2J( H-D) < 0 in accordance with

the well-known absolute signs for these types of coupling

constants. It should be mentioned that the deuterium iso-

tope effect for the two deuterium atoms on the H chemical

shift for CH3CN (ca. 6.8 Hz to lower frequency at 300 MHz)

must be taken into account when optimizing SPT frequencies

for the CHD2CN isotopomer.


Materials. Vinyltrimethylsilane was generously donated

by PCR Research Chemicals, Inc., Gainesville, Florida. Ace-

tonitrile was obtained from Mallinckrodt, Inc., Paris, Ken-

tucky, and acetonitrile-d3, 99%-deuterium-enriched ("gold

label"), was obtained from Aldrich Chemical Co., Milwaukee,

13 CH3CN


di ilL

Figure 3-6. H NMR spectrum showing the irradiation
positions of the SPT 7 pulses employed for the gener-
ation of spectra in Fig. 3-4b(A) and Fig. 3-5(B).

Wisconsin. Acetone-d6 was obtained from Merck Sharp &

Dohme Isotopes, St. Louis, Missouri. These materials were

used without any further purification.

Spectrometer. All experiments were performed on a

Nicolet NT-300 spectrometer equipped with a 293-C pulse
1 13 29
programmer and a Nicolet 1280 computer. H, 13C, and 29Si

NMR spectra were obtained at 300.07, 75.45, and 59.61 MHz,

respectively. These spectra were obtained at ambient temp-

erature, 22C. The amplitude, yH2/2n, of the H decoupler

was calibrated using the standard off-resonance decoupling

method (165) or by optimization of a 13C-{ H} SPT experiment

using a sample of CHC13 in both cases. The switching of the

decoupler frequency and amplitude required for the pulse

sequences described in this work was performed by means of

the 293-C pulse programmer. All samples were contained in

12-mm-o.d. sample tubes.

13C Spin-Echo J-Modulation Experiments

Spin-echo J modulations, a well-known and established
technique in H homonuclear pulse NMR (189), has been shown

to be an elegant method for the determination of CH3, CH2,

CH, and quaternary C multiplicities in proton-decoupled
13C NMR (190-199). Thus, this technique represents a wel-

come replacement for the off-resonance decoupling experiment

(200). Burum and Ernst (190) have demonstrated the applica-

bility of the INEPT (159) pulse sequence with refocusing

pulses. However, the INEPT experiment cannot readily be

performed on many older and routine FT NMR spectrometers.

This has led to the design of alternative methods (192-199)

for the retrieval of this important multiplicity information

from decoupled 13C NMR spectra. These techniques employ

heteronuclear spin-echo J modulation, decoupling only dur-

ing the acquisition time (197,198), or various other gated

broadband proton decoupling patterns (192-196,199). Further-

more, several variations using subtraction or addition of

spectra resulting from such experiments also have been

reported (194-196,198).

With these simple spin-echo J-modulation experiments

(i.e., without the use of the INEPT sequence) the signals

for CH and CH3 groups, on the one hand, are easily distin-

guished from those of CH2 and quaternary carbons, on the

other. The resonances of CH2 and quaternary carbons may

also be readily differentiated from one another by perform-

ing a second experiment in which use is made of the much

slower time evolution of the quaternary carbon signals

( J( 13C H) > nJ( 13C 1H)). However, the CH and CH3 carbon

signals are not as easily differentiated (difference spec-

troscopy has been used (195,196)) as in the INEPT sequence,

a fact that may lead to ambiguities.

Two of the pulse sequences which may be used for the

observation of proton-decoupled spin-echo J-modulated 13C

spectra are shown in Fig. 3-7. Figure 3-7 represents the

heteronuclear J-modulation experiment with broadband proton

decoupling during the acquisition period. Figure 3-7b

13c 90 180. Ac
oL 180o.

o"C 18 Acq Acq

b oVvv
I//u// . i ,

Figure 3-7. Pulse sequences for the observation of proton
broadband-decoupled spin echo J-modulated 13C NMR spectra.
(a) Heteronuclear J-modulation experiment with proton
decoupling during acquisition. (b) Delayed decoupling
sequence with 1800 s1C spin-echo pulse.

illustrates the simple delayed decoupling sequence with a

180 13C spin-echo pulse to prevent phase errors. For both

sequences the time evolution of the J modulation occurs dur-

ing a total period of 2T sec, i.e., during the time in which

the broadband decoupler is turned off.

Analytical expressions for the intensity modulation of

the proton-decoupled singlets from CH, CH2, and CH3 carbons

observed using either of the sequences in Fig. 3-7 may be

calculated from simple vector pictures (Fig. 3-8). Neglect-

ing transverse relaxation and using normalized intensities

(i.e., taking the intensities of the normal fully relaxed

decoupled 13C spectrum as unity), the intensity modulation

for the CH, CH2, and CH3 carbon singlets is described by

I(CH) = -cos (27JT), [1]

I(CH) = -{1 + 2cos (4TrJT)}, [2]
2 2 2
1 3
I(CH3) = -{ cos (67TJT) + 2-cos (27TJT)}, [3]

where T refers to the evolution periods indicated in Fig.

3-7 and the minus signs result from the 180 13C refocusing

pulse. Figure 3-9 shows how the amplitudes of the detected

singlets depend on 2T (the minus signs of Eqs. [1] to [3]

have been omitted). It is clear from the figure that the

CH and CH3 lines are easily separated from the CH2 and qua-

ternary carbon signals for 2T = 1/J. Further, it is also

evident that all types of protonated carbon signals with

equal values for the one-bond 13C- H coupling constants can

Figure 3-8. Spin vector representation of the 13C spins for a 13C
coupled to a 'H using the pulse sequence in Fig. 3-7a. (a) At equi-
librium, the spin vectors a and 8 are aligned along the z-axis.
(b) Applying a 90x(13C) pulse causes the spin vectors to become
aligned along the Y axis. (c) After a T period, the spin vectors
begin to fan out in the X-Y plane. Here the spin vector a represents
the higher frequency transition of the doublet. (d) The 180x (13C)
pulse reverses the sign of the Y-component of the vectors. The 180+x
(1H) pulse reverses the labels of the spins since each carbon spin is
associated with a particular proton spin. (e) After another T period
the spin vectors are symmetrically aligned about the y-axis with an
angle of 4TTJT between them. At this point, the proton decoupler is
turned on causing the spin vectors to collapse into the Y-axis and
the receiver is turned on. Since the intensity of the observed sig-
nal is proportional to the magnitude of the y-component of the vectors,
for a CH, the detected intensity is proportional to cos(2fIjT).





48 J



90(13 C)

180 H)
+ H)



X /-~~~



0.5- ^,*" '

_____ ---\ *_-------------. /...< ----- -2TCJT

-0.-\ -----CH
^..^.......... CH,
-).0- ~ -" ---- CH,

Figure 3-9. Dependence of the detected intensities on the
evolution period 2T for CH, CH2, and CH3 spin systems in
proton-decoupled spin-echo J-modulated 13C spectra; i.e.,
Eqs. [1] to t3], where the minus signs have been omitted.
The "magic angle" of 54.74' is indicated by the vertical

be suppressed for 2T = 1/2J thereby allowing quaternary

carbons (or protonated carbons of appreciably different

carbon constants) to be selectively observed. In this way

CH2 and quaternary carbon signals may be distinguished from

one another. Thus, the remaining assignment problem is dif-

ferentiating the CH and CH3 carbon lines.

The most straightforward solution to the differentia-

tion between CH and CH3 singlets appears to be a determina-

tion of the angle 27rJT for which the maximum difference

between the intensities of these resonances is observed

(197). The difference between the normalized intensity

expressions [1] and [3] is given by
AI = I(CH) I(CH3) = -{cos (6irJT) cos (2TJT)}


AI = cos3 (27JT) cos (27JT), [4]

from which

d(AI) = sin (2irJT){l 3 cos2 (2TJT-r)} [5]

is obtained. The extrema for AI, determined from

d(AI)/d(27TJT) = 0, show that for 2IJT < r/2 the maximum

difference between the intensities for CH and CH3 of equal

coupling constants occurs at an angle of 54.74 (in solid-

state NMR known as the "magic angle" for the magic-angle

spinning experiment). This angle corresponds to an evolu-

tion time 2T = 0.304/J and to normalized absolute intensity

values for I(CH) = 0.58 and I(CH3) = 0.19. Thus, neglecting

relaxation effects, the CH and CH3 singlet intensities at the

"magic angle" are reduced by 42 and 81%, respectively, as com-

pared to those of the "normal" decoupled spectrum (or of the

spectrum observed for 2T = l/J). The magnitude of this in-

tensity difference is large enough to allow CH and CH3 sing-

lets to be distinguished from one another and thereby make

a complete multiplicity assignment of the decoupled 13C


This spin-echo J-modulation multiplicity assignment

method is illustrated for the proton-decoupled 13C NMR

spectra of a-pinene at 75.46 MHz (Fig. 3-10). In Fig.

3-10c, the CH and CH3 lines are separated from the CH2 and

quaternary carbon signals using the simple pulse sequence

in Fig. 3-7b with 2Tr = 1/J. Examination of the signal at

41.1 ppm in the spectrum presented in Fig. 3-lOc shows the

signal is composed of a positive shoulder (carbon 6) on a

large negative peak (carbon 5). Quaternary and CH2 carbon

signals are distinguished from one another in the spectrum

of Fig. 3-10d in which the quaternary carbon resonances are

selectively observed employing 2T = 1/2J. The spectrum

shown in Fig. 3-10b which was observed at the "magic angle"

of 54.74 (i.e., using 2T = 0.304/J) clearly shows the

expected pronounced reduction of the CH line intensities

relative to the CH singlets by comparing it with the in-

tensities of Fig. 3-10a or Fig. 3-lOc.

b i


........ i

55 50 .5 4,0 35 30 25 20 ppm

Figure 3-10. Proton-decoupled 75.46-MHz 13C NMR spectra
of S-pinene in (CD3)2C0. (a) "Normal" spectrum shown with
the assignment given by Stothers (201). (b)-(d) Spin-echo
J-modulated 13C spectra obtained using (b) 2T = 0.304/J =
2.4 msec (J = 125 Hz) (i.e., at the "magic angle" of
54.74), (c) 2T = 1/J = 8.0 msec (J = 125 Hz), and
(d) 2T = 1/2J = 4.0 msec (J = 125 Hz). Phase corrections
have been made so quaternary and CH2 carbon resonances
appear as positive signals in all spectra.

In conclusion, the method proposed in this communication

for the differentiation between CH and CH3 multiplicities

shows that assignment of 13C resonances to CH3, CH2, and

CH, and quaternary carbon sites in large molecules may be

performed routinely without the need of subtraction or addi-

tion of spectra. Finally, the spin-echo J-modulation pulse

sequences in Fig. 3-7 not only preserve the NOE intensity

gain which in 13C NMR is virtually as great as the INEPT

polarization transfer enhancement but can also be performed

on older and FT NMR spectrometers without phase-shifting



Materials. H-pinene was obtained from a shelf bottle

and the solvent acetone-d6 was obtained from Merck Sharp

& Dohme Isotopes, St. Louis, Missouri. These materials

were used without any further purification. The sample

was contained in a-12 mm-o.d. tube.

Spectrometer. The 13C NMR spectra were obtained at

75.46 MHz on a Nicolet NT-300 spectrometer equipped with

a 293-C pulse programmer. Spin-echo J-modulation pulse

sequences were generated through microprogramming of the

293-C pulse programmer for the NT-300 spectrometer. A

spectral width of 14085 Hz and 64K data points at 75.46

MHz was used.


For the purposes of discussion, the carbons and hydrogens
of vinyl, phenyl, and ethyl groups will be denoted as follows:

1H 2 x
/a \/A

x 3H 3
a e

The chemical shifts are reproducible to 0.05 ppm and
the coupling constants are reproducible to 0.05 Hz. The
1 13 29
H, 13C, and 29Si chemical shifts are all reported with ref-

erence to internal TMS; a positive value indicates deshielding.
The 15N shifts are reported with reference to external 80%

v/v nitromethane in the same solvent as the sample.

Trimethylsilyl Compounds
Chemical Shifts
The H, 13C, and 29Si compound shifts of the trimethyl-
silyl group for compounds of the type Me 3SiX are presented in

Table 4-1. The H and 13C chemical shift data of the X
group are listed for completeness in Table 4-2.

Table 4-1. Chemical Shiftsa of the Trimethylsilyl Group for
Compounds of the Type Me3Six.

X H/ppm 13C/ppm 29Si/ppm

Me 0.00 0.00 0.00
CH2 CH=CH2 0.20 -1.75 0.30
CH2C6H5 -0.05 -1.95 1.16
CH=CH2 0.05 -1.59 -6.96
C6H5 0.21 -0.63 -4.08
CH Br 0.13 -2.53 3.14
CH2C1 0.11 -3.12 3.08
CHC1CH3 0.09 -3.97 5.41
CECSiMe3 0.13 0.05 19.42
SSiMe 0.06 4.16 14.01
SEt 0.27 0.97 14.74
SC6H5 0.21 0.80 16.98
N(SiMe3)2 0.20 5.49 2.46
NHSiMe3 0.04 2.57 2.14
NMeSiMe3 0.07 1.28 6.05
NHC6 H5 0.18 0.08 1.71
NMe2 0.08 1.30 6.48
NEt2 0.03 0.16 3.67
N=C=NSiMe3 0.14 1.09 -1.03
N=C=S 0.32 0.00 5.84
N=C=O 0.23 0.65 4.16
*N=N=N 0.24 -1.05 15.81
OSiMe2H 0.07 1.64 8.99
OSiMe3 0.07 1.96 6.86
OMe 0.06 -0.97 17.35
OEt 0.06 -0.41 15.14

chemical shifts are relative to TMS.

Table 4-2.

IH and 13C Chemical Shiftsa of the X Substituent for Compounds of the
Type Me SiX.



1.69 (CH2)
5.93 (HI)



4.98 (H2)
5.00 (H3)

2.01 (CH2)
6.90-7.00 (3H)


6.13 (H1)
5.87 (H2)




3.34 (CHC1)


5.64 (H3)


25.21 (CH2)

26.98 (CH2)
140.24 (C1)
128.11 (C2)

135.30 (a)
113.19 (8)

128.25 (C3)
124.11 (C4)

140.22 (a)
130.99 (0)

140.56 (C1)
133.86 (C2)

128.40 (C3)
129.40 (C4)



1.47 (CH3)

45.33 (CHCl)

C=CSiMe3 113.43

2.49 (a)

1.24 ()

7.04-7.20 (3H)
7.32-7.42 (2H)

20.50 (a)

131.65 (C1)
135.13 (C2)

20.01 (CH3)

18.75 (8)

128.72 (C3)
126.80 (C4)

-_c (NH)




CH2 Cl



SC6 H5


Table 4-2 continued.


NMeSiMe 3


NMe 2


N=C=NSiMe 3



OSiMe2 H



1n b


4.10 (NH)
6.56-6.64 (3H)
6.96-7.05 (2H)


2.79 (a)

0.13 (SiMe2)


3.62 (a)

0.96 (0)

4.73 (SiH)

1.12 (f)

13 C/ppmb


147.88 (C1)
116.30 (C2)
129.06 (C3)

40.57 (a)

117.12 (C4)


16.27 (3)






58.15 (a)

18.83 (8)

aChemical shifts are relative to TMS.

For the a, 0, HI, H2, H3, C1, C2, C3, C4 notation see the text.

The NH resonance could not be detected, presumably because of excessive line broadening
caused by nitrogen quadrupolar effects.

An extremely crude correlation (80% level of signifi-
chance) is obtained between 29Si chemical shifts and Swain-

Lupton field effect parameters, F (202), which are measures

for the inductive effect exerted by the substituent, X

(see Fig. 4-1). The relationship is

629Si = 27F + 2.6 [ppm] [i]

(r = 0.29, a = 6.1, N = 26)

This correlation implies 29Si shieldings in trimethylsilyl

compounds are influenced by inductive effects. Although no
correlation can be found between Si chemical shifts for

compounds containing nitrogen and oxygen substituents (sub-

stituents most likely to (d-p)r bond) and the F values, a

correlation is obtained at the 90% level of significance

for the carbon and sulfur substituents (substituents least

likely to (d-p)h bond). The relationship is

6 29Si = 56F + 0.3 [ppm] [2]

(r = 0.56, a = 6.0, N = 12)

The higher shielding of vinyl (-6.96 ppm) and phenyl (-4.08

ppm) in comparison to methyl (0.00 ppm) substituted trimethyl-

silane seems to indicate (d-p)7 bonding. The unusually large

deshielding of silicon for CECSiMe3, SC6H5, SEt, and SSiMe3

substituents may be explained by bond polarization which



-n 0

CO o
C E _


5fl 0

W\J 0 0
W o o

C o

-.2 .0 .2 .6

Figure 4-1. Plot of 629Si vs. Swain-Lupton field effect
parameters, F, for Me3SiX compounds. Open circles are
for molecules with N and 0 substituents; closed circles
are for those with C and S substituents.

increases the apparent electron-withdrawing power of these
substituents. The Si chemical shifts of CH2Cl (3.08 ppm),

CH2Br (3.14 ppm), and CH3 (0.00 ppm) substituted trimethyl-

silane are not in the predicted order for inductive effects

alone. Therefore, these results suggest 29Si chemical shifts

in trimethylsilyl compounds depend on other factors beside

simple inductive effects, including (d-p)7 bonding.

No correlation can be found between 6 13C of the tri-

methylsilyl group and the substituent F value. The largest

deviations (deshielding) from this relationship are for sub-

stituents having a number of 6 methyls. Examples and the

observed shifts include N(SiMe3)2 (5.49 ppm), NHSiMe3 (2.57

ppm), NMeSiMe3 (1.28 ppm), and OSiMe3 (1.96 ppm). Comparison

of 13C shifts of the trimethylsilyl group in Me3SiX compounds

for X = CH3, CH2Br, and CH2Cl shows the 13C shifts are in

the opposite direction from that predicted from inductive

effects. These results suggest that other factors including

steric effects are operating in 13C shielding.

Although no overall correlation can be found between
S29S 13
6 29Si and 13C of the trimethylsilyl group for Me3SiX com-

pounds, the data in Fig. 4-2 can be divided into three groups.

The first contains the four data points corresponding to the

three sulfur substituents and C:CSiMe3; it appears there is

an additional deshielding mechanism operational in these com-

pounds in comparison to that for 29Si shieldings of the re-

maining carbon substituted compounds. The second group

contains all the nitrogen and oxygen substituents; there





U 0


c I


0 0

-6 -3 3

Figure 4-2. Plot of 629Si vs. 6 3C of the trimethylsilyl
group for Me3SiX compounds. Open circles are for molecules
with N and 0 substituents; closed circles are for those
with C and S substituents.

29 ad 13
is no significant correlation between 6 29Si and 6 13C for this

group. The third group contains the remaining carbon

substituents. If the two substituents most likely to (d-p)T

bond (CH=CH2 and C6H5) are eliminated from this group, then a

correlation at the 95% level of significance is obtained with

the following relationship:

629 Si = -1.4 6S13C 0.9 [ppm] [3]

(r = 0.82, o = 1.0, N = 6)

The negative sign of the slope is correctly predicted from

inductive effects. In addition, the two other groups of data
29 13
also have negative slopes. Thus, although 29Si and 13C shifts

are significantly influenced by inductive effects, other con-

tributions to shielding such as (d-p)r bonding are important.

An extremely crude correlation (80% level of significance)

is obtained between 6 H of the trimethylsilyl group and the

corresponding F value (see Fig. 4-3). The relationship is:

IH = 0.36F + 0.06 [ppm] [4]

(r = 0.30, a = 0.08, N = 26)

The positive sign of the slope is correctly predicted by assum-

ing inductive effects dominate H shielding. The deviations

of H chemical shifts from this relationship may be due to

anisotropic effects since the largest deviations are for groups

with the largest anisotropy (e.g., CH2CH=CH2, CH2C6H5, C6H5,

and NH(C6H5)). In addition, there may be differences in the

transmission of inductive effects through the silicon atom.



-.2 1.0 .2 1 -

Figure 4-3. Plot of
for Me3SiX compounds


of the trimethylsilyl group
substituent F values.





13 29
Data comparing 13C and 29Si chemical shifts in analogous

trimethylsilyl and tertiary butyl compounds are presented in

Table 4-3. For the t-butyl derivatives the 13C chemical

shifts increase with increasing electron-withdrawing power of

the substituents as would be expected on the basis of induc-

tive effects alone. On the other hand 629Si values vary

irregularly with increasing inductive withdrawing power of

the substituents. Comparison of the 13C shifts referenced to
13 29
Me4 13C with 29Si shifts clearly shows increased shielding for

C6H5, nitrogen, and oxygen substituents. Since the inductive
13 29
effects of the substituents are the same in both 13C and 29Si

compounds, the increased shielding of these substituents indi-

cates contributions from (d-p)r bonding.

In Table 4-4 are presented data comparing 13C shifts of

the methyl group of trimethylsilyl and tertiary butyl

compounds. These data show the 13C shifts of the trimethyl-

silyl compounds are dominated by the strong inductive effect

(+1) of the silicon since the silicon derivatives are

shielded by about 30 ppm in comparison to the carbon


Coupling Constants

The values of coupling constants involving nuclei of

the trimethylsilyl group for Me 3SiX compounds are given in
29 1 29 13
Table 4-5. 29Si- H and 29Si- 13C coupling constants of the

X substituent for Me3SiX compounds are listed in Table 4-6.
1 29 .13 29
The J( 29Si3 C) were measured from the 29Si satellites

in the 13C- H decoupled spectra with NOE. Figure 4-4 shows

13 29
Table 4-3. Comparison of 13C and 29Si Chemical Shifts for M
in Me 3MX Compounds, where M = 13C and 29Si.a



CH2 2

C 6H C5








26.7 (0.0)

30.74 (4.0)

29.5 (2.8)

33.0 (6.3)

44.9 (18.2)

53.15 (26.5)

54.3e (27.6)

58.4 (31.7)

72.1 (45.4)

72.2 (45.5)























chemical shifts are referenced to Me 4Si.

bMe = CH3 and Et = CH 2CH3.

CValues in parentheses are referenced to Me4 13C.
dData from this study.

eFor comparison purposes, the value of 613C for N=C=NCHMe2
has been used with 629Si for N=C=NSiMe3.

Table 4-4. Comparison of 13C chemical Shifts in ( CH3)3MX
(where M = C, Si).

X M=C Ref. M=Sia

CH2CH=CH2 29.38 204 -1.75

CH2 C6 H5 28.9 205 -1.95

SC6 H5 30.8 206 0.80

NMe2 25.62 207 1.30

N=C=NSiMe3 31.1b 208 1.09

N=C=S 30.9 209 0.00

OMe 26.7 203 -0.97

OEt 27.7 203 -0.41

aData from this study.

bFor comparison purposes, the value of 613C for
Me CN=C=NCHMe has been used with 613C for
3 .

Table 4-5. Coupling Constants Involving Nuclei of the Trimethylsilyl Group for Compounds
of the Type Me3SiX.

X 1J13CH/Hz 13jI3CSiCHI/Hz I1J29Si3C I/Hz 12J29SiCHI/Hz

Me 118.14 2.11 50.96 6.62
CH2 C=CH2 118.66 1.91 51.49 6.77
CH2 C6 H5 118.81 1.88 51.27 6.79
CH=CH2 118.84 2.03 52.37 6.73
C6115 119.02 2.10 52.11 6.62
CH Br 119.50 1.79 52.88 6.69
CH2Cl 119.47 1.73 53.06 6.68
CHCICH3 119.44 1.96 53.10 6.66
C=CSiMe 119.86 2.00 56.15 7.06
SSiMe3 120.06 1.89 54.10 6.86
SEt 120.02 1.87 53.72 6.84
SC6 H5 120.56 1.81 54.29 6.81
N(SiMe3)2 118.15 1.93 56.43 6.48
NHSiMe3 117.69 1.66 56.12 6.61
NMeSiMe3 117.87 1.85 56.57 6.67
NHC6H5 118.60 1.72 57.19 6.68
6 5
NMe2 117.80 1.76 56.73 6.52
NEt2 117.70 1.67 56.55 6.52
N=C=NSiMe3 118.96 1.66 59.35 6.92

Table 4-5 continued.





IJj29Si3c I/Hz




OSiMe2 H

Table 4-6.


CH 2CH=CH 2c
CH2 2


CH 2Br

CH2 Cl







OSiMe 2Hh



29Si- 1H and 29Si-13C Coupling Constantsa of the
X Substituent for Compounds of the Type Me3SiX.

i1j29Si3C I/Hz




















15.26 (H2)e
8.59 (H3)










acoupling Constants are reproducible to 0.05 Hz.

bMe=CH3 and Et=CH2CH .

cI4J29SiCCCHI = 2.15 Hz.

d 4j2 9SiCCCHI = 15J29SiCCCCHI = 16J29SiCCCCCHI = 0.63 Hz.
eFor H2, H3 notation see the text.

fI4J29SiCCCHi = 15J29SiCCCCHI = 0.80 Hz.

g2j29SiCECc = 12.74 Hz.
hIlJ29SiHI = 203.55 Hz.


I I' I I ' I ' I ' I '
115.5 115.0 11'.5 114.O 113.5 PPM

Figure 4-4. 'H-decoupled 13C spectrum with NOE of the acetylene carbon of bis-
(trimethylsily) acetylene.

satellites for both the J( 29Si3 C) and the 2J( 29Si 13C) to the

acetylene carbon of bis-(trimethylsilyl) acetylene.
1 29 .1
The J( 29Si3 C) is strongly correlated (99.9% level of

significance) with the Huggins'electronegativity (210) XX, of

the X-group atom directly bonded to silicon (see Fig. 4-5).

The relationship is:

J( 29Si 13C) I = 7.9 XX + 32.8 [Hz] [5]

(r = 0.73, o = 1.6, N = 26)

Even for a data set containing a large number of substituents

capable of (d-p)T bonding, this correlation further supports

Harris' conclusion (95,96) (see Fig. 1-5) that the Fermi con-
1 29 13
tact term dominates J( 29Si3 C) in trimethylsilyl compounds.

As a better measure of the inductive effect of a substituent,

Swain-Lupton field effect parameters, F, were used. Figure

4-6 shows a plot of I J( 29Si 3C) I vs. F parameters. The

interesting feature of this plot is that, with the exception

of the substituent C=CSiMe3, all the points fall on one of

two lines. The upper line contains data for nitrogen and

oxygen substituents which are the substituents most likely

to (d-p)7r bond. The relationship is:

iJ(29 Sil 3C) = 8.3F + 56.8 [Hz] [6]

(r = 0.71, a = 0.9, N = 14)

The lower line contains carbon and sulfur substituents which

are the least likely to (d-p) bond. The relationship is:

-4_ - -_________

2.5 2.8

1 1
3,1 3, I


Figure 4-5. Plot of | J(29Si13C) in trimethylsilyl
compounds vs. substituent electronegativity.



CoI. _M


UIr I I^

.2 0 2 .6

are for molecules with N and 0 substituents; closed
circles are for those with C and S substituents.

I J( 29Si C) I = 11F + 51.9 [Hz] [7]

(r = 0.73, o = 0.8, N = 12)

Both of these correlations are at the 99% level of significance.

These results suggest that there is an additional contribution

to the spin-spin coupling from the non-contact terms for sub-

stituents most likely to (d-p)h bond. Thus, this appears to

be evidence of (d-p)rr bonding in Si-N and Si-O bonds. The

unusual position of CECSiMe3 which is intermediate between the

two lines may indicate moderate (d-p)7 bonding; however, since

this F value was approximated from the value of C=CH, this

conclusion cannot be drawn with any degree of certainty.

Two separate correlations are also obtained in a plot of
2 29 1 29 13?
J( 29SiCH) vs. J( 29Si 13C) (see Fig. 4-7). For carbon and

sulfur substituents, the relationship is:

2 29Q 1 29Q 13"
2J( SiCH) I = 0.064 J( 29Si 13C) I + 3.4 [Hz] [8]

(r = 0.57, o = 0.09, N = 12)

The relationship for nitrogen and oxygen substituents is:

12j(2 9SiCH)I = 0.11 1 j(29Si3 C)I + 0.2 [Hz] [9]

(r = 0.56, a = 0.16, N = 14)

Both of these correlations are at the 95% level of significance.
2 29 1 29Q 13"
Since both J( 29SiCH) and J( 29Si3 C) are expected to be depen-

dent upon the hybridization of the silicon and carbon atoms

(137) (Fermi contact term is dominant), it appears the silicon

hydribidization differs significantly for the two types of

substituents. Therefore, these results suggest that there