Electron-Rich 0 = PR3 compounds: Catalysts for alcohol silylation

Article abstract of DOI:10.1002/1098-1071(2001)12:1<21::aid-hc5>3.0.co;2-a

The catalytic effect of a group of R₃P = O compounds was studied in a mild procedure for the silylation of primary alcohols, secondary alcohols, hindered secondary alcohols, and of hindered phenols in the presence of t-butyldimethylsilyl chlori

Full text of DOI:10.1002/1098-1071(2001)12:1<21::aid-hc5>3.0.co;2-a

Heteroatom Chemistry
Volume 12, Number 1, 2001
Electron-Rich OסPR3 Compounds: Catalysts
for Alcohol Silylation
Xiaodong Liu and John G. Verkade
Department of Chemistry, Iowa State University, Ames, IA 50011
Received 18 July 2000; revised 29 August 2000
tistep syntheses [1]. Among the many trialkylsilyl
ABSTRACT: The catalytic effect of a group of R3PסO
compounds was studied in a mild procedure for the
silylation of primary alcohols, secondary alcohols,
hindered secondary alcohols, and of hindered phenols
in the presence of t-butyldimethylsilyl chloride
reagents used to protect this functionality, t-butyl-
dimethylsilyl chloride (TBDMSCl) and t-butyldi-
phenylsilyl chloride (TBDMSCl) are two of the most
popular [2]. A variety of methods has been reported
for the derivatization of alcohols with the TBDMS
and TBDPS moieties [3]. These reactions have been
most satisfactorily achieved by reacting the alcohol
with a molar excess of imidazole using dimethyl
formamide (DMF) as a solvent [4,5], or with cata-
lysts such as DMAP [2f], 1,1,3,3-tetramethylguani-
dine [2a], and ethyldiisopropylamine [2e]. More re-
cently, the silylation of primary and secondary
alcohols in 69–99% yields using TBDPCl in DMF
with catalysts, such as AgNO3, NH4NO3, or NH4ClO4
has been described [3d].
Proazaphosphatranes of type 1 [6] have been
shown to be very strong nonionic bases that function
as superior deprotonating agents [7], as superior cat-
alysts [8], and as efficient promoters [9] in a variety
of synthetically useful organic transformations. For
the very effective and mild silyl protection of a wide
variety of OH-containing organic substrates cata-
lyzed by 1 [8b], a mechanism involving intermedi-
ates 2a and 2b detected with 1 was postulated on the
basis of NMR evidence. In an earlier study of the
chemistry of 3, we discovered that the phosphoryl
group of this compound is capable of catalyzing the
conversion of isocyanates to isocyanurates [10] and
that it is also a good donor to Lewis acids including
silanes, forming cationic adducts such as 4a–f [11].
This prompted us to evaluate the catalytic activity of
3 in alcohol silylation reactions.
(TBDMSCl)
and
t-butyldiphenylsilyl
chloride
(TBDPSCl). It was found that R3PסO is an efficient
catalyst in such reactions when R is a good electron-
donating group, such as Me2N or n-Bu and as an
NMe(CH2) moiety in N(CH2CH2NMe)3PסO (3). How-
ever, R3PסO is a weak or ineffective catalyst when R
is a poor electron-donating group, such as Ph or O-n-
Bu or as a CH2N-o-CH2C5H4N moiety in N(CH2CH2N-
o-CH2C5H4N)3PסO. Compound 3, synthesized by ox-
idation of commercially available N(CH2CH2NMe)3P,
displayed the best catalytic properties for alcohol sily-
lation in terms of efficiency, stability, and safety.
᭧ 2001 John Wiley & Sons, Inc. Heteroatom Chem
12:21–26, 2001
INTRODUCTION
Protection of the organic hydroxyl group is neces-
sary for avoiding undesired reactions with oxidizing
agents and electrophiles during the course of mul-
Correspondence to: John Verkade
Contract Grant Sponsor: Petroleum Research Fund, administered
by the American Chemical Society.
Contract Grant Sponsor: ISU Center for Advanced Technology
Development.
᭧ 2001 John Wiley & Sons, Inc.
21

22 Liu and Verkade
quired for 99% conversion. In the nonpolar solvent
C6D6, 11 hours were required for 99% silylation of 6
using 0.5 equiv. of 3 as a catalyst, while no reaction
was observed during 12 hours in the absence of a
catalyst. When the coordinating solvent DMF was
used in the presence of 0.5 equiv. of 3, silylation of 6
was complete in 1.9 hours. Although DMF seemed
to be a somewhat better solvent than CH3CN in
terms of reaction rate, CH3CN was the solvent of
choice because of its lower boiling point and its abil-
ity to provide yields comparable with those obtained
in DMF. For a more complete comparison, reactions
were carried out on a preparative scale for three dif-
ferent hydroxyl compounds including primary al-
cohol 6, secondary alcohol 10, and hindered phenol
14 (Table 1) in the presence of one of the six phos-
phoryl compounds shown in this table. In addition,
DMAP, a commonly used silylation catalyst [2f], was
also compared.
Table 1 shows that OסPR3 wherein R is a good
electron-donating group, such as n-Bu, NMe2, or
NMe(CH2) (in 3), considerably accelerates silylation
for all three substrate alcohols, whereas the lack of
a good electron-donating group leads to a poor cat-
alyst (OסPPh3 and 5) or an ineffective one OסP(O-
n-Bu)3). In general, the catalytic efficiency for alco-
hol silylation of these phosphoryl compounds
follows the increasing electron donor ability of the
phosphoryl oxygen in the order OסP(O-n-Bu)3 Ͻ
OסPPh3 Ͻ 5 Ͻ OסP(n-Bu)3, OסP(NMe2)3, 3. Of all
the phosphoryl compounds tested, 3 seems most ef-
fective, although the advantage is admittedly some-
what marginal compared with OסP(n-Bu)3 or
OסP(NMe2)3. We believe that the slight superiority
of 3 in this respect may be associated with the
stronger donor character of the oxygen in this com-
pound than that in its acyclic analogue OסP(NMe2)3
owing to an Nax → P transannular interaction that
can occur in an intermediate or transition state [11].
However, the PסO group is more sterically hindered
in the rigid cage structure of 3 by the upwardly di-
rected Me groups, each of which resides on a planar
nitrogen. Such a bulk effect may compromise the
higher donor character of 3 to some extent. Thus 3
does not show a remarkable advantage over its acy-
clic analogues OסP(n-Bu)3 and OסP(NMe2)3. The
poorer performance of compound 5 is attributed to
the withdrawing nature of the pyridyl groups in the
CH2NCH2-o-C5H4N moieties and the large cone angle
swept out by the CH2-o-C5H4N segment of the CH2-
o-C5H4N groups.
Herein we report on the silyl protection of a wide
variety of alcohols, including primary alcohols, sec-
ondary alcohols, hindered secondary alcohols, and
of hindered phenols using 3, OסP(NMe2)3, and
OסP(n-Bu)3 as catalysts (equation 1) under mild
conditions. Among these catalysts, 3 displays the
best overall catalytic properties in terms of effi-
ciency, stability, and safety. A comparison of the ef-
ficiency of these catalysts with the commonly used
catalyst DMAP [2f] is also presented. The phosphine
oxides OסPPh3 and OסP(O-n-Bu) and 5 [12] were
found to be poor to nonfunctioning catalysts for al-
cohol silylation.
(1)
RESULTS AND DISCUSSION
In preliminary NMR monitoring reactions (see Ex-
perimental section), we found that TBDMS silylation
of benzyl alcohol (6) is accelerated in the presence
of 3. Thus in CD3CN, 2.1 hours were required to ef-
1
fect 99% silylation of 6 (according to H NMR inte-
gration) in the presence of 0.5 equiv. of 3, while 10.5
hours were required to obtain the same conversion
in the absence of a catalyst. When OסP(NMe2)3, an
acyclic analogue of 3, was used as the catalyst in
CD3CN in the same reaction, only 1.3 hours were re-
Compared with 3, the catalytic activity of DMAP
in the silylation of alcohols is about the same for the
primary alcohol 6 and the hindered phenol 14, but
is less efficient for the hindered secondary alcohol

Electron-Rich OסPR3 Compounds 23
TABLE 1 Comparison of Seven Catalysts for Alcohol Silylation
aSee Experimental section for conditions.
bBased on 1H integrations in which the error limit is about 1% absolute. Conversions are reproducible within this error limit for at least two
separate runs on each substrate.
10 (see later for additional discussion). At room tem-
perature, the silylation of 6, 10, and 14 using 2 equiv.
of imidazole was faster than the combination of 10
mol % 3 and 1.1 equiv. of Et3N (98% conversion vs.
90% for 6 in 0.4 hours, 97% vs. 88% for 10 in 4 hours,
and 67% vs. 50% for 14 in 8 hours, respectively).
However, when 10 mol % of imidazole and 1.1 equiv.
of Et3N was used, much slower conversions at room
temperature were observed than with 10 mol % of 3
and 1.1 equiv. of Et3N (67% vs. 90% for 6 in 0.4
hours, 53% vs. 88% for 10 in 4 hours, and 25% vs.
50% for 14 in 8 hours, respectively). Thus, on a mole-
for-mole basis, 3 is more efficient than imidazole.
That a base such as Et3N as well as a catalyst is nec-
essary for efficient silylations was shown by the lack
of detectable silylation of 6 by TBDMSCl in the pres-
ence of 10 mol % of 3 or OסP(NMe2)3 when Et3N
was absent.
In Table 2, it is shown that 3 catalyzes the
TBDMS silylation of primary alcohols 6–8 and phe-
nols 12 and 13 within 0.5 hours at room temperature
in excellent isolated yields (Ͼ91%), while the sec-
ondary alcohols 9–11 require a longer reaction time
(6 hours) to give excellent yields (Ͼ91%) of silylated
products. The hindered phenol 14 gave only a mod-
erate yield of silyl ether (55%) in 12 hours. The ter-
tiary alcohol 15 is resistant to silylation with
TBDMSCl, giving no detectable yield after 48 hours.
For each substrate, silylations catalyzed by DMAP
and in the absence of catalyst were also conducted
for comparison. It should be noted that for the pri-
mary alcohol 6, the less hindered secondary alcohol
11, and phenols (12–14), 3 shows about the same
efficiency as DMAP. However, for the primary alco-
hols 7 and 8, and the hindered secondary alcohols 9
and 10, 3 is somewhat marginally more efficient
than DMAP. Except for the TBDMS silylation of phe-
nols 12 and 13, silylations in the absence of catalysts
proceed in considerably lower conversions (20–
70%).
Table 3 shows that with 0.10 equiv. of 3 as a cat-
alyst at 35ЊC, the primary alcohols (6 and 7) and phe-
nol (12 and 13) are silylated with TBDPSCl within 6
hours and 4 hour, respectively, in high conversions
(95–99%), while the secondary alcohol 9 is more dif-
ficult to silylate, giving 74% conversion over 24
hours. It is noted that 11 and the acid-sensitive al-
cohol 8 require longer reaction times but give good
conversions (94% and 97%, respectively) to silylated
products. The hindered secondary alcohol 10 is re-
luctant to silylate with TBDPSCl, giving only a 30%
conversion of product in 24 hours.
Although several mechanistic pathways can be
considered that rationalize the ability of electron
rich OסPR3 compounds to catalyze hydroxyl silyla-
tion, we believe the one shown in Scheme 1 is the
most plausible on the basis of present evidence. This
pathway could be facilitated in the case of the some-
what superior catalyst 3, by transannulation of the
bridgehead nitrogen to the phosphorus to form in-
termediates B and/or C wherein phosphorus is five-
coordinate. An analogous pathway has been sug-
gested as a working hypothesis in the ring opening
of epoxides with SiCl4 promoted by OסP(NMe2)3
[13]. Further support for the pathway in Scheme 1
comes from our previously reported isolation and

24 Liu and Verkade
TABLE 2 Comparison of 3 and DMAP in Alcohol Silylation with TBDMSCla
aSee Experimental section for conditions.
bBased on 1H integrations of characteristic resonances. The conversions are reproducible for at least two separate runs on each substrate. The
error limit is about 1% absolute.
cAfter column chromatography, the purity was Ͼ95% by 1H NMR spectroscopy. The yields are the highest values observed in each case.
d10 mol% catalyst.
eNo detectable reaction.

Electron-Rich OסPR3 Compounds 25
known nasal carcinogen and should be avoided if a
substitute is available.
EXPERIMENTAL
CH3CN and CD3CN were distilled from CaH2, and
Et2O and benzene were dried with sodium. All sol-
vents were freshly distilled before use, and all reac-
tions were carried out under Ar. Catalysts 3 [6] and
5 [12] were prepared according to methods devel-
oped in our laboratories. Silica gel sheets were pur-
chased from J. T. Baker. All other chemicals were
purchased from Aldrich Chemical Co. and were used
as received.
TABLE 3 TBDPSCl Alcohol Silylations Catalyzed by 3a
SCHEME 1
1
characterization of 4e and 4f [11]. H and 31P NMR
spectroscopic data suggest the presence of transan-
nulation in these compounds as well as in 4c and 4d
[11]. Further supporting the cycle in Scheme 1 in
which ion formation is involved is the fact that the
catalyzed silylations occur in the polar solvent
CH3CN. By contrast, silylation is much slower in
benzene (see first paragraph of this section). There
is considerable evidence in the literature indicating
that pentacoordinate silicon compounds tend to be
more reactive to nucleophilic substitution than four-
coordinate silicon species [14]. This evidence also
supports the cycle in Scheme 1.
CONCLUSION
Compounds of the type OסPR3, in which R is a good
electron-donating group, are excellent catalysts for
alcohol and phenol protective silylations under mild
conditions using TBDMSCl and TBDPSCl. The ad-
vantages of these catalysts are (1) the yields or con-
versions of the silylated alcohols and phenols are
generally superior, (2) acetonitrile can be used in-
stead of the often employed but comparatively non-
volatile DMF, (3) the OסPR3 catalysts are highly sta-
ble under the reaction conditions employed, (4)
these catalysts are soluble in both polar and nonpo-
lar solvents, and (5) 3, though more expensive than
OסP(NMe2)3, can be recovered in good yield and is
less volatile than OסP(NMe2)3, which is a well-
aSee Experimental section for conditions.
bBased on 1H integrations in which the error limit is about 1% absolute.
Conversions are reproducible within this error limit for at least two
separate runs on each substrate. New compounds were character-
ized by 1H, 13C NMR, and HRMS(EI) spectroscopies.
c10 mol% catalyst.
dND, No detectable reaction.

26 Liu and Verkade
NMR Monitoring Experiments for the Catalytic
Silylation of 6
and evaporating the solvent under vacuum, 3 was
recovered as a white solid in 60–75% yields. The 31P,
1H, and 13C NMR spectra were identical to those of
an authentic sample of 3.
In a 5 mm NMR tube was dissolved 0.05 mmol of
catalyst (when catalysts were used) in 0.75 mL of
solvent (CD3CN,C6D6 or DMF). To this solution was
added 0.11 mmol of t-BuMe2SiCl followed by the ad-
dition of NEt3 (15 lL, 0.11 mmol). After shaking the
tube for 2 min, 6 (10 lL, 0.10 mmol) was added fol-
lowed by recording 1H NMR spectra at various time
intervals. The reaction temperature was 20ЊC. The
time interval between each spectrum was 1 min for
spectra 1–20, 10 min for spectra 21–30, 30 min for
spectra 31–40, 1 hour for spectra 41–45, and 4 hours
for each spectrum thereafter. One minute was re-
quired to complete each spectrum.
REFERENCES
[1] Greene, T. W.; Wuts, P. G. M. Protective Groups in
Organic Synthesis, 3rd ed.; Wiley & Sons: New York,
1999.
[2] (a) Kim, S.; Chang, H. Syn Commun 1984, 14 (10),
899; (b) Kim, S.; Chang, H. Bull Chem Soc Jpn 1985,
58, 3669; (c) Azipuraua, J. M.; Palono, C. Tetrahedron
Lett 1985, 26, 475; (d) Lissel, M.; Weiffer, J. Syn Com-
mun 1981, 11 (7), 545; (e) Lombardo, L. Tetrahedron
Lett 1984, 25, 227; (f) Chaudhary, S. K.; Hernandez,
O. Tetrahedron Lett 1979, 20, 99; (g) Olah, G. A.;
Gupta, B. G.; Narang, S. C.; Malhotra, R. J Org Chem
1979, 44, 4272; (h) Kita, Y.; Haruta, J.; Fuji, T.; Se-
gawa, J.; Tamura, Y. Synthesis 1981, 451.
[3] (a) Golinski, M.; Brock, C. P.; Watt, D. S. J Org Chem
1993, 58, 159; (b) Mazur, P.; Nakanishi, K. J Org
Chem 1992, 57, 1047; (c) Gooding, O. W.; Beard,
C. C.; Jackson, D. Y.; Wren, D. L.; Cooper, G. F. J Org
Chem 1991, 56, 1083; (d) Hardinger, S.; Wijaya, N.
Tetrahedron Lett 1993, 34, 3821; (e) Detty, M. R.; Sei-
dler, M. R. J Org Chem 1981, 46, 1283.
General Procedure for Silylations with t-
BuMe2SiCl (TBDMSCl) or t-BuPh2SiCl
(TBDPSCl)
In a 10 mL test tube capped with a rubber septum
was dissolved 0.1 equiv. of a catalyst in 2 mL of
CD3CN. To this was added 1.0 mmol of the alcohol
followed by the addition of NEt3 (0.15 mL, 1.1
mmol). After stirring the mixture for 5 min, 1.1
mmol of the silylating agent was added with contin-
uous stirring at room temperature (25ЊC) for t-
[4] Corey, E. J.; Venkateswarlu, A. J Am Chem Soc 1972,
94, 6190.
[5] Hanessian, S.; Lavallee, P. Can J Chem 1975, 53, 2975.
[6] Schmidt, H.; Lensink, C.; Xi, S. K.; Verkade, J. G. Z.
Anorg Allg Chem 1989, 578, 75.
1
BuMe2SiCl and at 35ЊC for t-BuPh2SiCl. H NMR
[7] (a) Tang, J.-S.; Verkade, J. G. J Org Chem 1994, 59,
7793; (b) Tang, J. S.; Verkade, J. G. J Org Chem 1996,
61, 8750; (c) Mohan, T.; Arumugam, S.; Wang, T.; Ja-
cobson, R. A.; Verkade, J. G. Heteroat Chem 1996, 7,
455; (d) Arumugam, S.; Verkade, J. G. J Org Chem
1997, 62, 4827; (e) Arumugam, S.; McLeod, D.; Ver-
kade, J. G. J Org Chem 1998, 63, 3677; (f) Wang, Z.;
Verkade, J. G. Heteroat Chem 1998, 9, 687.
[8] (a) Tang, J.-S.; Verkade, J. G. Angew Chem Int Ed
Engl 1993, 32, 896; (b) D’Sa, B.; Verkade, J. G. J Am
Chem Soc 1996, 118, 12,832; (c) D’Sa, B.; McLeod,
D.; Verkade, J. G. J Org Chem 1997, 62, 5057; (d) D’Sa,
B. A.; Kisanga, P.; Verkade, J. G. J Org Chem 1998,
63, 3961; (e) Kisanga, P.; D’Sa, B.; Verkade, J. G. J Org
Chem 1998, 63, 10057; (f) Kisanga, P. B.; McLeod, D.;
D’Sa, B. A.; Verkade, J. G. J Org Chem 1999, 64, 3091;
(g) Ilankumaran, P.; Verkade, J. G. J Org Chem 1999,
64, 3087; (h) D’Sa, B.; Kisanga, P.; Verkade, J. G. J
Org Chem 1998, 63, 3691; (i) Liu, X.; Verkade, J. G. J
Org Chem 1999, 64, 4840.
spectra were taken to obtain the conversion based
on 1H NMR integration of characteristic resonances,
and the product identity was confirmed by gas chro-
matography–mass spectroscopy. After the reaction
time stated in the tables, 0.02 mL of H2O was added
with stirring. The mixture was filtered, and the res-
idue was washed with Et2O (2 ן 5 mL) followed by
evaporating ca. 95% of the solvent under vacuum.
The resulting crude silyl ether was purified chro-
matographically on a silica gel column using a mix-
ture of 95% hexane and 5% ethyl acetate as the
eluent. The product was obtained upon drying over
anhydrous MgSO4 and evaporation of the eluent.
1
The identifying H and 13C NMR spectra compared
favorably with those in the literature (see supporting
information available from the author upon
request).
[9] (a) D’Sa, B. A.; Verkade, J. G. J Org Chem 1996, 61,
2963; (b) Kisanga, P. B.; Verkade, J. G. J Org Chem
1999, 64, 4298; (c) Wang, Z.; Verkade, J. G. J Org
Chem (in press).
General Procedure for the Recovery of 3
[10] Tang, J. S.; Mohan, T.; Verkade, J. G. J Org Chem
1994, 59, 4931.
[11] Liu, X.; Verkade, J. G. Inorg Chem 1998, 37, 5189.
[12] Liu, X.; Verkade, J. G. (in press).
[13] Denmark, S. E.; Barsanti, P. A.; Wong, K.; Stavenger,
R. A. J Org Chem 1998, 63, 2428.
After chromatographic separation of the silyl ether
product, the silica gel column was washed with an
additional 100 mL of a solution of 90% hexanes and
10% ethyl acetate followed by washing with 100 mL
of CH3OH. After collecting the pure CH3OH fraction