Silicon-29 NMR spectra of tert-butyldimethylsilyl and trimethylsilyl derivatives of some non-rigid diols

Article abstract of DOI:10.1135/cccc19970761

²⁹Si NMR spectra of trimethylsilyl (TMS) and tert-butyldimethylsilyl (TBDMS) derivatives of selected diols were measured under standardized conditions (i.e., in diluted CDCl₃ solutions). Application of the recently reported correlation between the chemical shifts in TMS and TBDMS derivatives revealed considerable and systematic deviations which exceeded experimental errors and error estimates from the correlation. Two possible explanations of the deviations are considered: interaction between the two bulky substituent groups and invalidity of the reported correlation for simple hydroxy derivatives. An independent study of analogous derivatives of monohydroxy compounds has shown that the linear correlation holds but the slope and intercept are significantly different from those reported previously on the basis of a study of amino acid derivatives. The data obtained for the diol derivatives fit the new correlation very well and no indication of an interaction between the bulky TBDMS groups was noticed. However, deviations do occur in branched diol derivatives in which branching reduces accessibility of the oxygen atoms surface to associate with proton donors. The largest deviation was found when intramolecular hydrogen bond was formed.

Full text of DOI:10.1135/cccc19970761

Silicon-29 NMR Spectra
761
SILICON-29 NMR SPECTRA OF tert-BUTYLDIMETHYLSILYL AND
TRIMETHYLSILYL DERIVATIVES OF SOME NON-RIGID DIOLS
a1
a2
b
Magdalena KVICALOVA , Vratislav BLECHTA , Krzysztof KOBYLCZYK ,
b
a3
Ryszard PIEKOS and Jan SCHRAML
a
Institute of Chemical Process Fundamentals, Academy of Sciences of the Czech Republic,
65 02 Prague 6-Suchdol, Czech Republic; e-mail: kvicalova@icpf.cas.cz,
blechta@icpf.cas.cz, schraml@icpf.cas.cz
Department of Inorganic Chemistry, Medical University of Gdansk, 80 416 Gdansk, Poland
1
1
2
3
b
Received December 9, 1996
Accepted February 7, 1997
2
9
Si NMR spectra of trimethylsilyl (TMS) and tert-butyldimethylsilyl (TBDMS) derivatives of se-
lected diols were measured under standardized conditions (i.e., in diluted CDCl solutions). Applica-
3
tion of the recently reported correlation between the chemical shifts in TMS and TBDMS derivatives
revealed considerable and systematic deviations which exceeded experimental errors and error esti-
mates from the correlation. Two possible explanations of the deviations are considered: interaction
between the two bulky substituent groups and invalidity of the reported correlation for simple hy-
droxy derivatives. An independent study of analogous derivatives of monohydroxy compounds has
shown that the linear correlation holds but the slope and intercept are significantly different from
those reported previously on the basis of a study of amino acid derivatives. The data obtained for the
diol derivatives fit the new correlation very well and no indication of an interaction between the
bulky TBDMS groups was noticed. However, deviations do occur in branched diol derivatives in
which branching reduces accessibility of the oxygen atoms surface to associate with proton donors.
The largest deviation was found when intramolecular hydrogen bond was formed.
Key words: NMR; Silicon-29 NMR; Trimethylsilyl derivatives; tert-Butyldimethylsilyl derivatives;
Diols.
While trimethylsilylation (TMS) has been extensively used for a number of purposes
since 1950’s, tert-butyldimethylsilyl (TBDMS) group has been introduced by Corey
1
and Venkateswarlu only in 1972. Since then the bulkier and hydrolytically more stable
tert-butyldimethylsilyl group has become the most popular silicon protecting group in
2
,3
synthetic chemistry . Despite that little has been reported on NMR properties of the
compounds containing TBDMS group, the data being usually hidden in experimental
4
29
parts of synthetic publications. Only Ralph has initiated a limited Si NMR study in
an attempt to analyze polyhydroxy compounds contained in lignin. The lack of suitable
2
9
data for comparison has become acute in our study of Si NMR spectra of TBDMS
derivatives of amino acids , which was supposed to be a useful alternative of Si tagg-
ing based on TMS derivatives of amino acids, which proved to be rather unstable .
5
29
6
7
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

7
62
Kvicalova, Blechta, Kobylczyk, Piekos, Schraml:
The studies of 29Si NMR spectra of TMS (ref. ) and TBDMS (ref. ) derivatives of
7
5
amino acids produced a surprisingly good correlation (r = 0.999 for 20 data points)
2
9
between the Si chemical shifts in the two classes of compounds measured under
6
identical standard conditions . The correlation was surprisingly good in the sense that it
involved both O- and N-bonded silicon atoms and covered a large chemical shift range
from 8.1 to 25.8 ppm (for TBDMS groups). The correlation could be used to predict
chemical shifts of e.g. TBDMS derivatives from the known shifts in TMS derivatives
with an estimated error of 0.11 ppm. Such accuracy suggested that a study of 2 Si NMR
spectra of disilylated diols could reveal interactions between the bulky groups as devi-
ations from the correlation. It is the goal of the present work to demonstrate that the
correlation could be applied to diol derivatives with no steric interaction between the
bulky groups.
9
EXPERIMENTAL
1
TBDMS derivatives of alcohols and diols were prepared by Corey’s procedure . A flask was charged
successively with dry alcohol (0.4–1.0 g), tert-butyldimethylchlorosilane (TBDMSCl, 1.2 mol per 1 mol
of OH), imidazole (2.5 mol per 1 mol OH), and dry dimethylformamide (DMF, 1–2 ml). The reac-
tion mixture was stirred at 50–70 °C under dry inert atmosphere for ca 1.5–2 h, cooled to room
temperature and the product was extracted into dry ether. Pure TBDMS derivative was obtained by
fractional distillation.
Two different procedures were used for preparation of TMS derivatives of alcohols.
A) Larger scale preparations were performed by refluxing 10–20 g of alcohol with HMDS (hexa-
methyldisilazane, 1.25 mol per 1 mol OH) until the evolution of ammonia ceased (about 2 days)
followed by fractional distillation.
B) Small scale derivatizations were carried out by stirring ca 0.5 g of alcohol with a silylating
reagent (N,O-bis(trimethylsilyl)acetamide, BSA, trimethylsilyldiethylamine, TMSDMA, trimethyl-
silyldimethylamine, TMSDEA, chosen to maximize boiling point difference, 1.3 mol per 1 mol OH)
under dry inert gas at 60–80 °C for 2 h. Pure TMS derivatives were isolated by fractional distillation.
NMR spectra were measured in dry CDCl solutions containing 1% (v/v) of hexamethyldisilane
3
(
HMDSS) as a secondary reference. The concentration of the sample in the measured solution was
1
3
reduced until the C chemical shift of HMDSS was δ –2.48 ± 0.02, relative to the central line of the
solvent at 76.99 ppm (see ref. for the details of this standard procedure).
6
All the NMR spectral measurements were performed on a Varian UNITY-200 spectrometer (oper-
1
3
29
ating at 50.3 MHz for C and at 39.7 MHz for Si NMR measurements), using standard software
APT and INEPT pulse sequences). The spectra were recorded in the temperature range 22–24 °C.
(
2
9
6
The Si NMR spectra were measured by the INEPT with the pulse sequence optimized for TMS
derivatives, i.e., for coupling to 9 protons and coupling constant of 6.5 Hz. The signal loss in the
8
case of TBDMS derivatives was negligible . Acquisition (1.0 s) was followed by a relaxation delay
of 5 s. During the acquisition period WALTZ decoupling was used and FID data (8 K) were sampled
for the spectral width of 4 000 Hz. Zero filling to 32 K and a mild exponential broadening were used
2
9
1
in the data processing. The Si π/2 pulses were at the maximum 17 µs long whereas H π/2 were 10 µs
2
9
in a 5 mm switchable probe. The Si spectra were referenced to the line of HMDSS at δ –19.79. The
C NMR spectra were measured using a spectral width of 16 000 Hz. WALTZ decoupling was ap-
plied both during acquisition (1 s) and relaxation delay (2–5 s). Zero filling to 64 K and 1–3 Hz line
1
3
2
9
broadening were used in data processing. The Si lines of diol derivatives were assigned by de-
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

Silicon-29 NMR Spectra
763
8
9
10
scribed variants of SPINEPTR (ref. ), selective INEPT and selective decoupling . These methods
2
9
1
were also used in the determination of the silylation site in a monosilylated diol. Long-range Si- H
1
0
coupling constants were determined either from 1D spectra or selective 2D J-resolved spectra .
The diastereoisomers of butane-2,3-diol and its derivatives were identified by a comparison of the
spectra of the mixtures with the spectra of an authentic R,R diastereoisomer (Aldrich). All com-
pounds were identified by their H and C NMR spectra; the latter being reported here for all com-
pounds studied. For the identification of coupling constants in compounds with two diastereotopic
1
13
CH protons the protons are labelled a and b. The proton a resonates at a lower magnetic field. The
2
two protons differ substantially in their coupling constants.
Solvent accessible surface (A) was calculated for oxygen atoms exactly as described pre-
viously1
2,13
.
The isolated compounds and their 13C NMR chemical shifts are as follows:
1
,10-Bis(trimethylsilyloxy)decane: 62.73 (CH O); 32.73, 29.56, 29.42, 25.82 (CH ); –0.46
2 2
(
CH Si).
3
1
,10-Bis(tert-butyldimethylsilyloxy)decane: 63.33 (CH O); 32.89, 29.59, 29.43, 25.80 (CH );
2 2
2
5.99 (CH C); 18.39 (C); –5.25 (CH Si).
3
3
1
1
,5-Bis(trimethylsilyloxy)pentane: 62.58 (CH O); 32.50, 22.10 (CH ); –0.48 (CH Si).
2 2 3
,5-Bis(tert-butyldimethylsilyloxy)pentane: 63.21 (CH O); 32.67, 22.14 (CH ); 25.98 (CH ); 18.37 (C);
2
2
3
–
5.28 (CH Si).
3
1
1
,4-Bis(trimethylsilyloxy)butane: 62.44 (CH O); 29.10 (CH ); –0.48 (CH Si).
2 2 3
,4-Bis(tert-butyldimethylsilyloxy)butane: 63.11 (CH O); 29.33 (CH ); 25.97 (CH ); 18.35 (C); –5.27
2
2
3
(
CH Si).
3
1
1
,3-Bis(trimethylsilyloxy)propane: 59.17 (CH O); 35.53 (CH ); –0.51 (CH Si).
2 2 3
,3-Bis(tert-butyldimethylsilyloxy)propane: 59.68 (CH O); 35.92 (CH ); 25.94 (CH ); 18.32 (C);
2
2
3
–
5.35 (CH Si).
3
1
1
1
,2-Bis(trimethylsilyloxy)ethane: 63.88 (CH O); –0.43 (CH Si).
2 3
,2-Bis(tert-butyldimethylsilyloxy)ethane: 64.66 (CH O); 25.96 (CH ); 18.42 (C); –5.26 (CH Si).
2
3
3
-Butoxy-2-trimethylsilyloxyethane: 72.04, 71.17, 62.05 (CH O); 31.77, 19.29 (CH ); 13.91
2
2
(
CH ); –0.41 (CH Si).
3 3
1
-Butoxy-2-(tert-butyldimethylsilyloxy)ethane: 72.15, 71.19, 62.75 (CH O); 31.83, 19.27 (CH );
2 2
2
5.92, 13.91 (CH ); 18.37 (C); –5.26 (CH Si).
3
3
1
,2-Bis(trimethylsiloxy)propane: 69.13 (CHO); 68.28 (CH ); 20.41 (CH ); 0.20 (CH Si); –0.47
2 3 3
(
CH Si).
3
1
,2-Bis(tert-butyldimethylsilyloxy)propane: 69.44 (CHO); 69.00 (CH O); 25.98, 25.89 (CH );
2 3
1
8.39, 18.20 (C); –4.57, –4.73, –5.27, –5.36 (CH Si).
3
1
,3-Bis(trimethylsilyloxy)butane: 65.29 (CH); 59.35 (CH ); 42.34 (CH ); 24.03 (CH ); 0.22, –0.50
2 2 3
(
(
CH Si).
3
1
,3-Bis(tert-butyldimethylsilyloxy)butane: 65.50 (CHO); 60.18 (CH O); 42.82 (CH ); 25.98, 25.93
2 2
CH C); 24.05 (CH ); 18.29, 18.12 (C); –4.37, –4.80, –5.30 (CH Si).
3
3
3
(
2R,3R)-2,3-Bis(trimethylsilyloxy)butane: 72.09 (CH); 18.18 (CH ); 0.25 (CH Si).
3 3
(
2R,3R)-2,3-Bis(tert-butyldimethylsilyloxy)butane: 71.36 (CH); 25.87 (CH C); 18.10 (CH ); 18.06 (C);
3
3
–
4.34 (CH Si).
3
(
2R,3S)-2,3-Bis(trimethylsilyloxy)butane: 73.04 (CH); 20.05 (CH ); 0.23 (CH Si).
3 3
(
2R,3S)-2,3-Bis(tert-butyldimethylsilyloxy)butane: 73.32 (CH); 25.93 (CH C); 20.17 (CH ); 16.32 (C);
3
3
–
2
4.62 (CH Si).
3
2
,4-Bis(trimethylsilyloxy)-2-methylpentane: 73.35 (CHO); 65.90 (CO); 54.09 (CH ); 31.74, 29.63,
2
5.57 (CH ); 2.64, 0.50 (CH Si).
3
3
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

7
64
Kvicalova, Blechta, Kobylczyk, Piekos, Schraml:
2
,4-Bis(tert-butyldimethylsilyloxy)-2-methylpentane: 73.05 (CO); 66.12 (CH); 54.76 (CH ); 31.76,
2
2
2
2
9.57 (CH ); 25.97, 25.88 (CH C); 18.05 (C); –3.82, –4.52 (CH Si).
3 3 3
4
-Trimethylsilyloxy-2-methylpentan-2-ol: 70.25 (CO); 67.77 (CHO); 49.95 (CH ); 31.20, 27.86,
2
4.75 (CH ); 0.58 (CH Si).
3
3
4
-tert-Butyldimethylsilyloxy-2-methylpentan-2-ol: 70.30 (CHO); 68.10 (CO); 50.03 (CH ); 31.20,
2
7.73, 24.96 (CH ); 25.80, 25.56 (CH C); 17.80 (C); –2.49 (CH Si).
3
3
3
RESULTS AND DISCUSSION
The 29Si NMR experimental data are summarized in Table I, 13C NMR chemical shifts
are given in Experimental for each of the compounds studied.
As expected, the 29Si chemical shifts of both TMS and TBDMS groups in the deriva-
tives of linear diols reach the values found for the derivatives of simple alcohols as the
distance between the two oxygen atoms sufficiently increases. For example, the chemi-
cal shifts in TMS and TBDMS derivatives of decane-1,10-diol differ from those in
1
3
butanol derivatives by +0.03 and +0.08 ppm, respectively. Shortening of the alkyl
1
3
chain increases these differences (e.g., the maximum shift differences are found be-
tween ethanediol and ethanol derivatives, +0.29 and +0.22 ppm, respectively). This
trend and the sign of the differences are in agreement with the reported dependence of
1
3,14
these shifts upon polar effects
(electronegative substituents R on the oxygen atom
′
cause deshielding of the silicon atom in the R–O–SiR , moiety if the substituents on the
3
silicon atom R′ are alkyl groups) and with the different polar effects of (CH ) SiO
3
3
groups and hydrogen atoms.
Branching has similar effects on 29Si chemical shifts in TMS and TBDMS deriva-
1
3
tives, its effects are in agreement with the trends observed for other nuclei, e.g., for
C
chemical shifts. Thus, branching at carbon α to the oxygen atom has a larger effect
(
about –2 ppm) than branching at β position (about –0.5 ppm).
Even for the shortest –(CH ) – chain, substitution on one of the oxygen atoms has
2
n
2
9
little effect on the Si chemical shift of the silicon atom on the other oxygen atom.
This observation follows from comparison of chemical shifts in 1,2-bis(trimethylsilyl-
oxy)ethane (18.52), 1-butoxy-2-trimethylsilyloxyethane (18.56) and 2-trimethylsilyl-
1
3
oxy-1-methoxyethane (18.79) (ref. ). Violation of this rule will be discussed later in
connection with intramolecular hydrogen bond in a monosilylated diol.
6
The chemical shifts reported here exhibit large deviations from the described corre-
lation between the shifts in the two classes (TMS and TBDMS) of derivatives. These
deviations are expressed in Table I as differences ∆ between the observed experimental
chemical shift and the shifts calculated for the TBDMS derivatives according to the
6
correlation . In most cases (including derivatives of both linear and branched diols), the
deviations are about 10 times larger than the error estimate due to the correlation (±0.11 ppm)
6
(
ref. ) and since they are all negative they suggest some systematic error. In view of the
above discussion of the shifts in linear diols fitting the general trends, it seems most
probable that the large deviations are due to the correlation employed. This observation
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

Silicon-29 NMR Spectra
765
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

7
66
Kvicalova, Blechta, Kobylczyk, Piekos, Schraml:
has triggered an independent study of TMS and TBDMS derivatives of simple alcohols
1
3
which is reported elswhere . The study of derivatives of 24 alcohols yielded correla-
tion
δ(TBDMS) = 0.9899 δ(TMS) + 1.422
(
with correlation coefficient 0.998 and error estimate ±0.19 ppm) which is significantly
different (on 90% significance level) from the correlation reported earlier for amino
6
acid derivatives . The chemical shifts predicted according to this correlation and their
differences from the experimental values are given in the last two columns of Table I.
Obviously, the differences are within the expected error of estimate for all the linear
and some of the branched diol derivatives. Larger differences are found only in the last
5
rows of Table I which comprise the most branched compounds studied here. Similar
1
3
negative differences are, however, noticed in the source correlation for branched com-
pounds like tert-butyl alcohol derivatives. Negative differences (i.e., the calculated
shifts being larger than the experimental ones) are probably due to steric hindrance to
1
1,12
association with the proton of the solvent. As described previously
such an associ-
ation leads to higher chemical shift values. The different association capabilities of
TMS and TBDMS derivatives are well ilustrated in Fig. 1 for association of 2-methyl-
–3
pentane-2,4-diol derivatives with phenol. The derived association constants (K in mol dm )
4
2
roughly follow the order in solvent accessible surface (A in 10 pm ) of the correspond-
ing oxygen atoms. In TMS and TBDMS derivatives the K(CHOSi) values are 2.5 and
0
.5, respectively, while the respective A(CHOSi) values are 0.6 and 0.1 and K(COSi)
ones are 2.0 and 0.4, while the surfaces A(COSi) values are 0.1, and 0.0.
The large positive deviation in the monosilyl derivatives (the last line in Table I) is caused
by intramolecular hydrogen bonds in these compounds as illustrated by (i) limited asso-
2
2
1
δ, ppm
2
4
3
1
1
1
8
4
0
FIG. 1
2
9
Dependence of Si chemical shift δ
ppm) on the concentration c (mol l ) of
–
1
(
phenol added to carbon tetrachloride sol-
utions of 4-tert-butyldimethylsiloxy-2-
methylpentan-2-ol (1), 4-trimethylsilyl-
oxy-2-methylpentan-2-ol (2), 2,4-bis(tri-
methylsilyloxy)-2-methylpentane (CHOSi
5
6
3
, COSi 6), 2,4-bis-(tert-butyldimethyl-
silyloxy)-2-methylpentane (CHOSi 4,
COSi 5)
6
0
1
2
3
4
–1
c, mol l
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

Silicon-29 NMR Spectra
767
ciation of both mono-TMS and mono-TBDMS derivatives with phenol (Fig. 1) and (ii)
large downfield shifts of both TMS and TBDMS monosilyl derivatives relative to their
disilyl counterparts (despite the insensitivity of the shifts to substitution on the other
oxygen atom as discussed above). The calculated solvent accessible surface is 0.6 and
0
.1 for the oxygen atom in the CH–O–Si moieties in these two compounds. The com-
pounds can form a six-membered ring through the hydrogen bond formation. Since the
hydroxy proton is much more acidic than the proton of chloroform, this hydrogen bond
2
9
is stronger and shifts the Si resonance downfield more than the association with chlo-
roform in the model compounds from which the correlation was derived. Hence, we
observe the large positive differences between the experimental and calculated shifts.
(
CH )
CH
H
CH
3
3
2
2
C
O
CH
O
Si
The large shift due to intramolecular hydrogen bond observed in this case brings an
important warning. In analyses of complicated mixtures of polyfunctional compounds
2
9
utilizing Si NMR tagging technique in which the spectral lines can be assigned only
by comparison with the tabulated chemical shifts the assignment can be in error if
similar association (not necessarily intramolecular) occur due to e.g. incomplete silyla-
tion.
Summarizing – the correlation between (TMS) and (TBDMS) gives good prediction
for the 29Si chemical shifts in the latter derivatives. The deviation exceeding the error
of the estimate may indicate specific interactions.
The authors are grateful to the Grant Agency of the Czech Republic for a grant (No. 203/94/1026)
which made this work possible. We would also like to thank several colleagues, Dr R. Rericha and Dr J.
Cermak (from this Institute), and Dr M. Tkadlecova (Prague Institute of Chemical Technology) for their
extensive help.
REFERENCES
1
2
. Corey E. J., Venkateswarlu A.: J. Am. Chem. Soc. 94, 6190 (1972).
. Greene T. W., Wuts P. G. M.: Protective Groups in Organic Synthesis (2nd ed.), p. 77. Wiley,
New York 1991.
3
4
5
6
7
8
. Kocienski P.: Protecting Groups, p. 33. Thieme, Stuttgart 1994.
. Ralph J.: NMR and Wood Science Workshop. University of British Columbia, Vancouver 1985.
. Kubec R., Velisek J., Kvicalova M., Cermak J., Schraml J.: Magn. Reson. Chem. 33, 458 (1995).
. Schraml J.: Prog. Nucl. Magn. Reson. Spectrosc. 22, 289 (1990).
. Schraml J., Kvicalova M., Schwarzova I., Velisek J.: Magn. Reson. Chem. 32, 591 (1994).
. Blechta V., Cermak J., Kvicalova M., Schraml J.: XVIth Int. Conf. Organometal. Chem., p. 189.
The Royal Society of Chemistry, Brighton 1994.
9. Schraml J.: J. Magn. Reson. 59, 515 (1984).
Collect. Czech. Chem. Commun. (Vol. 62) (1997)

7
68
Kvicalova, Blechta, Kobylczyk, Piekos, Schraml:
1
1
1
1
1
0. Blechta V., Schraml J.: Unpublished results.
1. Schraml J., Jakoubkova M., Kvicalova M., Kasal A.: J. Chem. Soc., Perkin Trans. 2 1994, 1.
2. Kasal A., Schraml J., Cermak J., Magn. Reson. Chem. 32, 394 (1994).
3. Schraml J., Kvicalova M., Blechta V., Cermak J.: J. Magn. Reson., submitted.
4. Schraml J., Chvalovsky V., Magi M., Lippmaa E.: Collect. Czech. Chem. Commun. 46, 377
(
1981).
Collect. Czech. Chem. Commun. (Vol. 62) (1997)