On the interactions of alkyl 2-hydroxycarboxylic acids with alkoxysilanes: Selective esterification of simple 2-hydroxycarboxylic acids

Article abstract of DOI:10.1002/chem.200601859

The interactions of a range of monocarboxylic acids with tetramethoxysilane Si(OMe)₄ (TMOS), in methanol (MeOH), have been investigated by using ¹H, ¹³C and ²⁹Si solution-phase NMR spectroscopy and electrospray mass spectrometry (ESMS). Si(OMe)₄ acts as a catalyst/reagent in the selective methylation of 2-hydroxycarboxylic acids (2HOAs) in MeOH at room temperature: glycolic acid, lactic acid and 2-hydroxybutyric acid are esterified more than a hundred times faster in MeOH and Si(OMe)₄ than in MeOH alone. No acceleration of methylation is observed for carboxylic acids lacking the 2-hydroxy group. Methylation of the 2HOAs is associated with the condensation of individual siloxane units to form oligomers. A mechanism is proposed in which 2HOAs attach to silicon via the alkoxy group, then subsequently via the carboxyl group in an intramolecular rearrangement to form an unstable and reactive cyclic intermediate. This intermediate may lead to accelerated methylation of the carboxylic acid via nucleophilic attack of MeOH at the carbonyl group, while a separate reaction pathway leads to condensation of silanols and/or alkoxysilanes leading to oligosiloxanes. The mechanism has implications for the use of 2HOAs as templates in sol-gel silica preparation.

Full text of DOI:10.1002/chem.200601859

DOI: 10.1002/chem.200601859
On the Interactions of Alkyl 2-Hydroxycarboxylic Acids with Alkoxysilanes:
Selective Esterification of Simple 2-Hydroxycarboxylic Acids
RichardJ. Ansell,* [a] Simon A. Barrett,[a] Jonathan E. Meegan,[a, b] and
Stuart L. Warriner[a]
Abstract: The interactions of a range
of monocarboxylic acids with tetrame-
than in MeOH alone. No acceleration
of methylation is observed for carbox-
ylic acids lacking the 2-hydroxy group.
Methylation of the 2HOAs is associat-
ed with the condensation of individual
siloxane units to form oligomers. A
mechanism is proposed in which
2HOAs attach to silicon via the alkoxy
group, then subsequently via the car-
boxyl group in an intramolecular rear-
rangement to form an unstable and re-
active cyclic intermediate. This inter-
mediate may lead to accelerated meth-
ylation of the carboxylic acid via nucle-
ophilic attack of MeOH at the
carbonyl group, while a separate reac-
tion pathway leads to condensation of
silanols and/or alkoxysilanes leading to
oligosiloxanes. The mechanism has im-
plications for the use of 2HOAs as
templates in sol–gel silica preparation.
thoxysilane Si(OMe)4 (TMOS), in
G
methanol (MeOH), have been investi-
gated by using 1H, 13C and 29Si solu-
tion-phase NMR spectroscopy and
electrospray
mass
spectrometry
(ESMS). Si(OMe)4 acts as a catalyst/re-
ACHTREUNG
agent in the selective methylation of 2-
hydroxycarboxylic acids (2HOAs) in
MeOH at room temperature: glycolic
acid, lactic acid and 2-hydroxybutyric
acid are esterified more than a hundred
Keywords: homogeneous catalysis ·
mass
spectrometry
·
NMR
spectroscopy · silicates · siloxanes
times faster in MeOH and Si(OMe)4
A
Introduction
and 98% yields after 18 h at room temperature, whilst suc-
cinic acid was methylated <5%. The authors proposed a
mechanism involving a five-membered intermediate with
both the carboxy and alkoxy oxygen atoms bonded to
boron. The following year, Yamamoto et al.[3] reported im-
proved yields in the same reaction by replacing boric acid
with N-methyl-4-boropyridinium iodide.
Esterification of carboxylic acids is one of the most widely
used and best understood reactions in synthetic chemistry.[1]
Nonetheless, new catalysts or reagents capable of regio-,
stereo- or chemoselectively esterifying acids are of continu-
ing interest. For example, in 2004 Houston et al.[2] reported
that boric acid (B(OH)3, 10–20 mol%) was effective as a
catalyst for the chemoselective esterification of 2-hydroxy-
carboxylic acids (2HOAs) with excess alcohol as solvent.
Glycolic, lactic and tartaric acids were methylated in 80, 65
It is shown here that tetramethoxysilane, TMOS or Si-
A
similar selectivity and efficiency to boric acid (Table 1). This
work focuses particularly on the mechanism of this reaction,
which is associated with the condensation of monomeric si-
loxane units to form oligosiloxanes. Alkoxysilanes Si(OR)4
are commonly used as precursors in the sol–gel process and
their complicated chemistry has been the subject of exten-
sive investigations.[4] The generation of silica from mixtures
of alkoxysilane, alcohol, acid and water has been the subject
of numerous studies.[5] 2HOAs have been widely used in
such mixtures as templates or structure-directing agents,[6]
with some interaction with the polymerising siloxane centres
assumed but not well understood. Understanding the
chemistry of siloxanes, silanols or silicates with small organic
ligands is also important in bioinorganic chemistry[7] and
theories of the chemical origins of life.[8]
[a] Dr. R. J. Ansell, S. A. Barrett, Dr. J. E. Meegan, Dr. S. L. Warriner
School of Chemistry
University of Leeds
Woodhouse Lane, Leeds LS29JT (UK)
Fax : (+44)113-343-6465
E-mail: chmrja@leeds.ac.uk
[b] Dr. J. E. Meegan
Atomic Weapons Establishment
Aldermaston, Reading, RG7 4PR (UK)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
4654
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 4654 – 4664

FULL PAPER
Table 1. Results for mixtures of acids (0.45m) with TMOS (0.53m) left in
CD3OD at room temperature for 24 h. Yields calculated from 1H NMR
spectra as described in the Results section.
whilst for AA and 3HBA (carboxylic acids without a 2-hy-
droxy group) the acceleration due to Si(OMe)4 is minimal
A
or zero. This selectivity suggests direct involvement of the
hydroxy group, probably as a ligand to the silicon, which for
the 2HOAs but not for 3HBA serves to accelerate steps 1a,
1b or 1c.
2HOAs are known to form stable complexes in aqueous
solution with a range of metals,[17] but their complexes with
silicon are apparently less stable. Mehrotra et al. in the
1960ꢁs reacted salicyclic, mandelic and lactic acids with Si-
Acid
R=
Yield [%]
1
acetic (AA)
(R)-(ꢀ)-b-hydroxybutyric (3HBA)
glycolic (GA)
0
71
84
81
(OEt)4 in benzene with distillation of the EtOH/benzene
isotrope and isolated products that contained residual
alkoxy ligands together with salicylate, mandelate or lactate
ligands, either divalent (via both carboxy and hydroxy) or
monovalent (via the hydroxy), depending on the exact con-
ditions.[16] However, structures were assigned purely on the
basis of the measured quantity of alcohol eliminated and el-
emental analysis of the residue. Tacke et al. isolated a series
of penta- and hexavalent silicon-HOA complexes with SiO5
(S)-(+)-lactic (LA)
(S)-(+)-a-hydroxybutyric (2HBA)
Although the reaction of carboxylic acids with alkoxysi-
lanes has been known to produce carboxylate esters, as well
as causing siloxane oligomerisation, dependent on the condi-
tions, for many years,[9,10] there has been no report of the se-
lectivity of this reaction for 2HOAs. In fact the production
of carboxylate esters has been observed not only in this re-
action, but also in those of carboxylic anhydrides with alkox-
ysilanes,[11] alcohols with acyloxysilanes[12,13] and ternary re-
actions of carboxylic acids with chlorosilanes and alco-
hols;[14] however, in each case, the literature is divided over
the actual mechanisms. The chemistry of alkoxy and acylox-
ysilanes can be broadly understood in terms of a reactivity
or SiO6 cores by treating Si(OMe)4 with BA, GA, citric acid
N
or malic acid in MeCN or THF and recrystallising the pre-
cipitated product.[18] For most of these complexes, crystal
structures have been obtained which show that the 2HOA
ligands chelate the silicon via both the carboxyl and hydrox-
yl groups. By using alkylalkoxysilanes instead of Si(OMe)4,
G
ꢀ
similar complexes were isolated incorporating one Si C
bond, which in some cases were reported to be water
stable.[19]
When 2HOAs have been used as templates or structure-
directing agents in the sol–gel synthesis of silica from alkox-
ysilanes it has usually been in alcohol/water cosolvent sys-
tems.[6] Hence, the question arises whether under these con-
ditions the 2HOAs and Si centre only interact via “outer
sphere” interactions, such as hydrogen bonding to the
alkoxy ligands, or whether “inner sphere” interactions occur
via covalent complexation. Moreover, if such complexes can
form in protic solvents, do they include coordinating interac-
tions via both the hydroxyl and carboxyl groups, or single
point interactions? In an attempt to shed light on these
issues and to investigate further the mechanism of the car-
boxylic acid–alkoxysilane reaction we applied a range of
modern analytical methods to the reactions of different car-
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ ꢀ
series
Si Cl>Si OOCR>Si OR>Si OH>Si O Si.
Thus, acyloxysilanes are susceptible to alcoholysis and hy-
drolysis, whilst the substitution of an alkoxy ligand by an
acyloxy one is difficult. The reaction of carboxylic acids with
alkoxysilanes is generally believed to include (at least) two
steps, carboxylation (step 1a), followed by reaction of the
acyloxysilane group with alcohol (step 1b) and/or an alkoxy-
silane group (step 1c) to give the ester.[5,13]
L3SiOR1 þ R2COOH ! L3SiOOCR2 þ R1OH
L3SiOOCR2 þ R1OH ! L3SiOH þ R2COOR1
L3SiOOCR2 þ L3SiOR1 ! L3SiOSiL3 þ R2COOR1
ð1aÞ
ð1bÞ
ð1cÞ
boxylic acids with Si(OMe)4 in MeOH.
U
Results
These reactions generally proceed in rigorously dry condi-
tions and at a high temperature.[15,16] Driving off the alcohol
created in step 1a promotes the formation of acyloxysilanes,
rather than esters and oligosiloxanes/siloxane gels.
13C and 1H NMR spectroscopy of acid/TMOS mixtures in
[D4]methanol—the acidandester signals : When carboxylic
acids are added to Si(OMe)4 in CD3OD, ligand exchange
G
In the course of studying the action of 2HOAs in the sol–
gel process, we found that these acids readily undergo
(L3SiOCD3 for L3SiOCH3) is accelerated, as well as the hy-
drolysis and condensation reactions of L3SiOMe (the initial
steps of the sol–gel process), leading to changes in the silox-
esterification in dilute solutions of Si(OMe)4 in MeOH, at
A
1
room temperature and without rigorous exclusion of water.
The acids compared in this work are shown in Table 1. The
rate of esterification for GA, LA and 2HBA is accelerated
to 100–10000 times the background rate in MeOH alone,
ane signals of the H and 13C spectra and the growth of sig-
nals due to CH3OD. For certain carboxylic acids investigat-
ed, the rapid conversion of the acid into the methyl ester
was observed (step 2a).
Chem. Eur. J. 2007, 13, 4654 – 4664
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4655

R. J. Ansell et al.
R2COOD þ R3OCðD3=H3Þ $ R2COOCðD3=H3Þ þ R3OD
ð2aÞ
Assuming complete exchange of the acidic protons for the
more abundant deuteriums of the solvent. The methoxy
group transferred to the acid may come from free solvent
(R3 =D), or from the methoxy side chains of the siloxane
species (R3 =L3Si)). Figure 1 shows selected regions of the
1H and 13C spectra after mixing TMOS+GA for 7 h,
Figure 2. Concentrations of product esters versus time for the mixtures of
acids (0.45m) with TMOS (0.53m, filled symbols and bold lines) or
CH3OH (2.1m, open symbols and dotted lines) in CD3OD, determined
1
~
~
&
from H NMR spectra as outlined in text. a) AA (
,
), 3HBA ( , &).
~
,
~
&
^
,
^
), LA (
b) GA (
), 2HBA ( , &). Note different scales are used
in a) and b).
by comparing the integrals of the CHO
free acid and ester. This calculation was easiest for GA and
LA, in which the CHO(H/D) is a singlet (Figure 1a), but
A
AHCTREUNG
more complicated for 2HBA, for which it required the mea-
surement of overlapping multiplets. For 3HBA, the signals
do not change as there is no ester produced. For AA, the in-
ꢀ
tegrals of the CH3 C signal for the free acid and ester were
Figure 1. Selected regions of the 1H and 13C spectra after mixing
measured, and, being singlets, the calculation was again
simple. Signals attributed to the esters were confirmed by
spiking the reactions after one week with authentic samples
of the esters which were commercially available. Assign-
ments of all the observed signals are given in Table 2.
TMOS+GA for 7 h. a) CHO(D/H) region, b) OCH3 region, c) COOMe
G
region, d) CHOH region and e) OCH3 region.
whence approximately 50% of the GA has been esterified.
The CHO
A
In the absence of TMOS, the rate of esterification increas-
es in the order 3HBA<AA<2HBA<LA<GA, suggesting
this process is governed by a combination of acid pKa and
steric accessibility. When the acids are mixed with TMOS in
CD3OD, only minor changes to the background esterifica-
tion rate are observed for the non-2HOAs. The rate remains
insignificant for 3HBA+TMOS and is accelerated <10 fold
for AA+TMOS. For all of the 2HOAs, when added to
TMOS in CD3OD, dramatic acceleration of esterification is
observed. The rate enhancements are ꢁ100-fold over the
background rate for GA, ꢁ1000-fold for LA and >10000-
ester are clearly seen (the COOCH3 and COOCH3 peaks,
however, are very small because most of the ester created is
in the form ꢀCOOCD3). Signals attributed to methyl GA
were confirmed by spiking the mixture with pure ester.
Esterification occurs when carboxylic acids are incubated
in CD3OD even in the absence of TMOS, but usually at an
extremely slow rate. The rate of esterification was measured
for a series of acids in CD3OD in the presence or absence of
TMOS (either TMOS or CH3OH added, Figure 2). For GA,
LA and 2HBA, the extent of esterification was quantified
4656
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 4654 – 4664

2-Hydroxycarboxylic Acids
FULL PAPER
Table 2. Assignment of signals in 1H and 13C spectra. Chemical shifts (d) are measured in ppm and coupling constants (J, shown in parentheses) in Hz.
Acid
Form a H
b H
g H
Ester CH3 a C
b C
Carboxy C g C
Ester CH3
AA
3HBA free
free 1.99, s
–
–
–
–
20.76
–
175.25
–
–
–
2.44, dd (15.24, 7.08) 4.15, dqd, (ca. 6)
2.38, dd (15.36, 5.92) merged
1.21, t (6.50)
44.56 65.54 175.37
23.25
GA
LA
free
ester
free
4.08, s
4.11, s
4.22, q (6.95)
4.25, q (7.00)
–
–
–
–
–
–
–
60.71
60.92–
67.60 20.71 178.34
67.83 20.56 176.83
72.60 28.47 177.74
–
176.13
174.79
–
–
3.73, s
–
3.72, s
–
–
51.59
1.377, d (7.00)
1.36 d, (6.90)
–
–
9.69
–
ester
51.60
–
2HBA free
ester
4.060, dd (7.07, 4.72) 1.674, dqd (14.21, 7.10, 7.08) 0.98, t (7.35)
1.806, dqd (13.99, 7.55, 4.63)
4.09, dd (7.25, 4.75)
1.66, dqd (14.31, 7.19, 7.16)
1.78, dqd (14.44, 7.20, 4.76)
0.95, t (7.50) 3.72, s
72.89 28.47 176.35
9.69 51.60
fold for 2HBA. Thus, TMOS appears to act as either reagent
or catalyst in the methylation of 2HOAs, whilst being less
effective, or ineffective, in the methylation of carboxylic
acids lacking 2-hydroxy groups.
Each of these processes leads to further signals due to
L3SiOCH3 or L3SiOCD3 within hydrolysed or oligomeric
species which can also be seen in Figure 1b and e.
Figure 3 shows how the concentrations of L3SiOCH3,
L3SiOCD3 and CH3OD change after adding TMOS to LA
or H2O in CD3OD. Changes occur ꢁ1000 fold faster with
LA than with H2O. TMOS+AA and TMOS+3HBA be-
haved essentially identically to TMOS+H2O, whilst mix-
tures of TMOS with GA or 2HBA behaved essentially iden-
tically to TMOS+LA. Initial growth of the CH3OD and
The observed selectivity suggests that a covalently bound
intermediate, either via the 2HOA hydroxyl ligand, carboxyl
group or both, may be involved in the mechanism, although
1
no signals were ever identified in the H and 13C NMR spec-
tra that could be assigned to such intermediates. This does
not preclude their existence, as they may be short-lived on
the NMR timescale and their nuclei may possess chemical
shifts only slightly different from the free acid.
13C and 1H NMR spectroscopy of acid/TMOS mixtures in
[D4]methanol—the siloxane signals: When Si(OMe)4 is incu-
A
bated alone in CD3OD, slow processes of ligand exchange
(step 3a) and hydrolysis (step 3b) are observed.
L3SiOCH3 þ CD3OD ! L3SiOCD3 þ CH3OD
ð3aÞ
ð3bÞ
ðL3SiOCH3 or L3SiOCD3Þ þ D2O
! L3SiOD þ ðCH3OD or CD3ODÞ
D2O is derived from H2O, which is initially present or which
ingresses into the tube and exchanges its protons with the
more abundant CD3OD. Both processes lead to the loss of
the L3SiOCH3 (d=3.56 ppm, singlet; Figure 1b) and
L3SiOCH3 (d=51.6 ppm, singlet; Figure 1e) signals, but the
growth of new CH3OD (d=3.35 ppm, singlet) and CH3OD
(d=49.9 ppm, singlet) signals, almost overlaying the
CHD2OD and CD3OD signals. The ligand exchange process
(step 3a) also leads to a L3SiOCD3 signal (d=50.8 ppm,
septet). In the presence of added H2O, these processes are
accelerated and, because L3SiOD is formed in larger quanti-
ties, condensation reactions to form L3SiOSiL3 also become
significant (steps 3c and 3d).
ðL3SiOCH3 or L3SiOCD3Þ þ L3SiOD
ð3cÞ
! L3SiOSiL3 þ ðCH3OD or CD3ODÞ
^
),
Figure 3. Changes in the concentration of TMOS L3SiOCH3
(
~
&
L3SiOCD3 ( ) and CH3O
A
L3SiOD þ L3SiOD ! L3SiOSiL3 þ D2O
ð3dÞ
CD3OD containing a) H2O (0.90m) and b) LA (0.45m).
Chem. Eur. J. 2007, 13, 4654 – 4664
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4657

R. J. Ansell et al.
L3SiOCD3 signals due to ligand exchange is followed by a
slow decrease of the L3SiOCD3 signal due to processes 3b
and 3c. The time at which the L3SiOCD3 concentration
reaches a maximum can be taken as a measure of the rates
of processes 3a to 3d combined: this time is approximately
one week for TMOS with H2O, AA or 3HBA, and ꢁ10 min
for TMOS with GA, LA or 2HBA.
FTIR analysis of 2HOA+TMOS mixtures in [D4]methanol:
Short-lived species may be resolved more easily by FTIR
than by NMR spectroscopy. Thus, FTIR was used to study
the reactions of the acids with TMOS in CD3OD. However,
the results were complicated by the similar frequencies of
absorbances due to groups in CD3OD and L3SiOCD3, and
1
in CH3OD and L3SiOCH3. The data confirmed the H and
13C NMR findings that SiOCD3 groups were more abundant
after one week in the mixtures of TMOS with AA and
3HBA than in mixtures of TMOS with the 2HOAs. No evi-
dence was found for transient siloxane–acid complexes (see
the Supporting Information).
29Si NMR of 2HOA+TMOS mixtures in [D4]methanol:
29Si NMR was also used to probe the species present after
mixing the acids with TMOS in CD3OD (Figure 4). The as-
signments shown are based on reported studies of the hy-
drolysis of alkoxysilanes by using 29Si NMR spectrosco-
py.[20,21] Q0 corresponds to silicon with four OC or OD li-
gands, Q1 to silicon with one of these ligands replaced by
OSi and Q2 to silicon with two ligands replaced with OSi.
The major Q0 peak is in each case a singlet at d=
ꢀ78.2ppm, which by comparison with the spectrum of pure
TMOS and with the literature can be assigned to Si(OCH3
or OCD3)4. For the mixture of TMOS+AA and TMOS+
3HBA, an additional Q0 peak appears at d=ꢀ76.0 ppm
which can be assigned to (CH3O or CD3O)3SiOD.[21] This
peak also appears after 24 h, for TMOS+H2O. However,
the spectra observed for TMOS with the 2HOAs are quite
different: the only Q0 peak observed is that due to Si(OCH3
or OCD3)4, and this is observed to diminish over time while
Q1 and Q2 signals grow. The Q1 species present are chain-
end L3SiOSi
(OR3), and the Q2 species are mid-chain
A
Figure 4. 29Si NMR spectra of TMOS (0.53m) in CD3OD, mixed with
a) H2O (0.90m), b) AA (0.44m) and c) and d) GA (0.45m). a) recorded
7 h after mixing, b) and c) 1 h after mixing and d) 16 h after mixing.
Number of scans for d) was reduced to 128.
L3SiOSi(OR)2OSiL3. Different length chains and cycles, as
well as different combinations of OH and (OCH3 or OCD3)
ligands split the Q1 and Q2 signals. In the spectrum of GA+
TMOS one hour after mixing, there are just two Q1 signals
at d=ꢀ85.6 and ꢀ85.8 ppm, but after 16 h these are joined
by two Q2 signals at d=ꢀ93.5 and ꢀ93.8 ppm, indicating the
evolution of a more complex array of oligosiloxane prod-
ucts. Similar spectra are observed with TMOS+LA and
TMOS+2HBA.
substituted with 2HOAs. Substitution of an alcohol with an
acyl group can lead to large changes in the 29Si chemical
shift, but in the opposite direction to that observed here.[22]
Substitution via the 2HOA hydroxyl group would lead to
much smaller changes[23] and so might account for the ob-
served splitting of the Q1 and Q2 signals.
29Si NMR spectra thus show a clear difference between
the initial stages of the AA+TMOS and 3HBA+TMOS re-
actions (which lead to slow production of L3SiOD) and the
2HOA+TMOS reactions (which lead to no measurable
L3SiOD, but relatively fast production of L3SiOSiL3). It
could be conjectured that the signals assigned as Q1 and Q2
species actually correspond to complexes, that is, siloxanes
MS analysis of 2HOA+TMOS mixtures in methanol:
GCMS has been used previously to identify intermediates in
the sol–gel process of TEOS.[24] It is by no means obvious,
however, that ions identified under these conditions truly re-
4658
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 4654 – 4664

2-Hydroxycarboxylic Acids
FULL PAPER
flect the species present in solution at a given stage. A more
accurate picture of the inherent ions might be expected by
using electrospray ionisation. Thus, Cooney et al.[25] used
electrospray with a quadrupole mass analyser to obtain a
spectrum for TEOS in EtOH in positive-ion mode (domi-
nated by Na(Si
negative-ion mode by using an inherently ionisable siloxane,
3-sulfanylpropyl-triethoxysilane (HS(CH2)3Si(OEt)3). Schüth
(OEt)4)2+), but could only obtain spectra in
A
ACHTREUNG
et al.[26] used a similar setup to study solutions of silica dis-
solved in tetraethylammonium hydroxide in MeOH/H2O.
Under these alkaline conditions, they identified a range of
oligomeric species with various levels of OH/OCH3 ex-
change. Marshall et al.[27] used electrospray with a Fourier
transform ion-cyclotron resonance mass analyser to charac-
terise alkaline TEOS/EtOH/H2O mixtures, and used the ac-
curate mass resolution of their instrument to assign the oli-
gosiloxane structures present with confidence. The limita-
tions of their approach, however, are that the solution was
diluted in acetonitrile prior to electrospraying, which may
have altered the equilibria between different species pres-
ent, and the instrument was operated in positive-ion mode,
with species being identified as cationic metal–siloxane com-
plexes. Subsequently Woenckhaus et al.[28] employed electro-
spray with a quadrupole analyser to characterise acidic
TMOS/water and TEOS/water mixtures. They directly in-
jected the unadulterated reaction mixtures and compared
positive and negative-mode ionisation; only the latter gave
useful spectra, enabling them to assign peaks to anionic sili-
cate chains (5 min after mixing) cylcles and polyhedra (after
longer intervals). As the quadrupole instrument used had
limited mass resolution, they identified the various ions on
the basis of integral m/z ratios and with the use of H/D ex-
change.
In the current work, unadulterated samples were directly
injected into a TOF instrument. They were ionised by nega-
tive electrospray and the structures of the resulting ions
were assigned on the basis of accurate masses. Spectra are
shown in Figure 5 and assignments of selected peaks in
Table 3 (for more complete assignments see the Supporting
Information). The first spectrum shown is for TMOS+H2O
in CH3OH and was recorded 26 hours after mixing. The
same mixture analysed immediately after mixing gave only
a very weak spectrum, confirming the observation via
29Si NMR spectroscopy that only TMOS itself is present in
the early stages. The sample after 26 hours contains oligo-
meric ions of one to seven silicon atoms, which can be as-
signed as shown to either chains in which some of the silicon
atoms are 3-coordinate with one L2Si=O group, or to cycles
and polyhedra. The ions are similar to those observed by
Schuth et al.[26] and by Woenckhaus et al.,[28] except that the
ions in our spectrum, recorded with a 56:1 (v/v) ratio of
MeOH/H2O, are more extensively methylated. 3-Coordinate
silicon species containing L2Si=O groups are certainly not
expected to be present in the actual solution. Woenckhaus
et al. proposed that these arose due to dehydration of
Figure 5. ES-MS spectra obtained in negative-ion mode with direct infu-
sion of samples. TMOS (0.53m) in CH3OH mixed with a) H2O (0.90m),
b) 3HBA (0.45m), c) GA (0.45m), d) LA (0.45m), e) 2HBA (0.45m).
ꢀ
ꢀ
methoxylation of oligomers containing Si(OH)(OMe) .
A
Thus included in Table 3 are the “parent ion” linear, saturat-
ed oligomers which could yield the observed ions by loss of
one or more CH3OH.
The remaining spectra were recorded approximately two
hours after mixing. The spectrum for TMOS+AA in
MeOH was extremely weak with no signals above 104 (not
1
shown). H, 13C and 29Si NMR spectroscopy has shown that
L3SiOH is formed in the TMOS+AA reaction, as in
TMOS+H2O; however, the acid presumably suppresses the
ionisation of the silanols under electrospray conditions.
All of the hydroxy acids studied when mixed with TMOS
in MeOH generated spectra containing peaks that could be
assigned to complexes between monomeric or oligomerised
siloxanes and the acids. However, the spectra for TMOS
with the 2HOAs, GA, LA and 2HBA, were very similar
(Figure 5c–e), and quite different to that for TMOS+3HBA
(Figure 5b). The peaks in the spectrum for TMOS+3HBA
are of comparable intensity to those in the spectrum for
ꢀ
ꢀ
oligomers containing Si(OH)2 during the electrospray
process, in the current work, they might also arise from de-
Chem. Eur. J. 2007, 13, 4654 – 4664
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4659

R. J. Ansell et al.
Table 3. Selected ions observed in ESI-TOF MS experiments (for more complete assignments see the Supporting Information).
Sample
Observed
ions
Assigned
Calcd m/z
(difference
[ppmꢀ1])
Possible structure
Suggested parent ion
H2O+TMOS,
1568 min
ꢀ
ꢀ
229.0202
303.0027
Si2C4H13O7
229.0205 (+1.3)
303.0029 (+0.8)
A, n=2 , 4”R=CH3, 1”R=H
Si3C5H15O9
A, n=3, 6”R=CH3, 1”R=H
B, n=1, 5”R=CH3
B, n=2 , 6”R=CH3, 1”R=H
ꢀ
ꢀ
394.9956
468.9788
Si4C6H19O12
394.9959 (+0.8)
468.9783 (ꢀ1.1)
A, n=4, 7”R=CH3, 2 ”R=H
A, n=5, 9”R=CH3, 2 ”R=H
Si5C7H21O14
C, n=1, 7”R=CH3
E, n=1, m=1, R=H
3HBA+TMOS,
122 min
ꢀ
177.0228
SiC5H9O5
177.0225 (ꢀ1.6)
F, n=1, m=1, 1”R=CH3,
2”R=H
ꢀ
189.0768
249.0441
C8H13O5
189.0768 (+0.1)
249.0436 (ꢀ2.2)
G, m=2
ꢀ
SiC8H13O7
E, n=1, m=2 , R=H
F, n=1, m=2 , 1”R=CH3,
2”R=H
ꢀ
275.1146
335.0813
C12H19O7
275.1136 (ꢀ3.6)
335.0804 (ꢀ2.6)
G, m=3
E, n=1, m=3, R=H
ꢀ
SiC12H19O9
F, n=1, m=3, 1”R=CH3,
2”R=H
F, n=1, m=4, 1”R=CH3,
2”R=H
ꢀ
421.1184
SiC16H25O11
421.1172 (ꢀ3.0)
E, n=1, m=4, R=H
GA+TMOS,146 min
ꢀ
133.0142C
4H5O5
133.0142( +0.1)
H, all X=H
ꢀ
192.9807
SiC4H5O7
192.9810 (+1.6)
I, n=0, all X=H, 1”R=H
I, n=1, all X=H, 3”R=CH3
I, n=2, all X=H, 5”R=CH3
ꢀ
ꢀ
313.0051
419.0121
Si2C7H13O10
Si3C9H19O13
313.0053 (+0.5)
419.0139 (+4.3)
ꢀ
LA+TMOS, 129 min 221.0123
SiC6H9O7
221.0123 (ꢀ0.2)
235.0280 (ꢀ5.9)
327.0209 (+0.3)
I, n=0, all X=CH3, 1”R=H
I, n=0, all X=CH3, 1”R=CH3
I, n=1, all X=CH3, 2 ”R=CH3,
1”R=H
I, n=1, all X=CH3, 3”R=CH3
I, n=2, all X=CH3, 4”R=CH3,
1”R=H
ꢀ
235.0293
327.0208
SiC7H11O7
ꢀ
ꢀ
Si2C8H15O10
341.0377
433.0299
Si2C9H17O10
Si3C10H21O13 433.0295 (ꢀ0.8)
341.0366 (ꢀ3.4)
ꢀ
ꢀ
447.0447
Si3C11H23O13 447.0452( +1.2)
I, n=2, all X=CH3, 5”R=CH3
4660
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 4654 – 4664

2-Hydroxycarboxylic Acids
Table 3. (Continued)
FULL PAPER
Sample
Observed
ions
Assigned
Calcd m/z
(difference
[ppmꢀ1])
Possible structure
Suggested parent ion
ꢀ
ꢀ
2HBA+TMOS,
136 min
249.0429
SiC8H13O7
SiC9H15O7
249.0436 (+2.8)
263.0593 (ꢀ7.7)
I, n=0, all X=CH2CH3, 1”R=H
263.0613
355.0512Si
I, n=0, all X=CH2CH3, 1”R=CH3
I, n=1, all X=CH2CH3, 2 ”R=CH3,
1”R=H
I, n=1, all X=CH2CH3, 3”R=CH3
I, n=2, all X=CH2CH3, 4”R=CH3,
1”R=H
ꢀ
2C10H19O10 355.0522 (+2.8)
ꢀ
369.0699
461.0600
Si2C11H21O10 369.0679 (ꢀ5.4)
ꢀ
Si3C12H25O13 461.0608 (+1.9)
ꢀ
475.0747
Si3C13H27O13 475.0765 (+3.8)
I, n=2, all X=CH2CH3, 5”R=CH3
TMOS+H2O after 26 hours, and can be assigned to dimers
and trimers of 3HBA, and to mono- and disiloxanes with
varying levels of OH/OCH3 exchange and between one and
four 3HBAs attached. The peak at m/z: 189.0768 could be
the 3HBA anhydride with a deprotonated hydroxyl, but
given the existence of the peak at m/z: 275.1146, the homo-
ester structures shown seem more probable. These ions also
appear in the spectrum of 3HBA+CH3OH in the absence
of TMOS, so must form spontaneously in solution. Various
structures might account for the acid–siloxane complexes. It
has been assumed that as the ionisation of SiOH is sup-
pressed for the AA+TMOS mixture, the same is true here
and ions must be due to ionised carboxylate groups. Com-
plexes in which silicon is multiply substituted with single
3HBA ligands might appear more probable than those
singly substituted with oligomeric 3HBA as illustrated by
structures E or F. However, the ion with m/z: 421.1172
shows that some of the ligands must indeed be oligomeric.
This ion cannot result from a complex with only monomeric
ligands.
The spectra for each of the 2HOAs with TMOS are differ-
ent from that for 3HBA+TMOS in several ways: the peaks
are ꢁ20 times more intense; a dimer is observed for GA
but only weakly for LA, and not at all for 2HBA, there are
no trimers for any of the 2HOAs and the silicon complexes
present can all be assigned with monomeric ligands only;
the oligosiloxanes are larger with up to four silicon atoms;
and however many silicon atoms are present, there are
always exactly two 2HOA ligands. The greater extent of oli-
gomerisation is in agreement with the 29Si spectra. The fact
that all of the silicon complexes in these three spectra can
be assigned with a general structure I, with exactly two
2HOA ligands, has implications for the mechanisms of com-
plex formation and esterification as discussed below.
acid–siloxane complexes as intermediates, such as that out-
lined in the introduction (steps 1a–1c), originate with An-
drianov et al. who, in the 1950ꢁs, proposed that during the
reaction of Et2Si
ene,[29] the acid displaced an alkoxy ligand giving Et2Si-
(OEt)(OAc). A second cycle gave Et2Si(OAc)2, which was
A
A
N
ACHTREUNG
the final product if EtOH was removed by distillation, but
otherwise might react with EtOH to give EtOAc, and, by
condensation, polydiethylsiloxanes. Andrianovꢁs reactions
took place under water-free conditions at an elevated tem-
perature, and the only alcohol present was that produced in
step 1a. Also, the alkylsiloxane groups should stabilise the
resulting complex with the less electron-donating carboxyl
ligand. Sanchez and Livage et al. showed later by using
NMR spectroscopy that acetic acid will displace ethoxy
groups from Si(OEt)4, although they only observed this to
U
occur for the 2-component mixture under reflux.[30] Thus, it
is unclear whether the mechanism in steps 1a–1c (outlined
more fully by Sharp[13]) could actually apply for acids react-
ing with tetraalkoxysilanes, such as TEOS, at room tempera-
ture.
In our studies, we observed esterification of AA (very
slow) in CD3OD in the absence of TMOS. For AA+TMOS
in CD3OD, only very slight acceleration of esterification was
observed, perhaps because in the presence of excess alcohol
and at temperatures less than reflux, the reaction in step 1a
is unfavourable. NMR spectroscopy and MS both failed to
identify any complexes of silicon with AA.
For 3HBA no esterification was observed at all, which
may be attributed to steric hindrance (the pKa for 3HBA is
almost identical to that for AA), even in the presence of
TMOS. Consideration of the CH3OD, SiOCD3 and Si signals
suggests that the processes in steps 3a–3d occur at a similar-
ly slow rate for 3HBA+TMOS as for AA+TMOS. Howev-
er, MS shows a clear difference between AA+TMOS and
3HBA+TMOS: in the latter case, a series of complexes are
observed involving one or two siloxane units with some or
all of the methoxy ligands hydrolysed or replaced with
3HBA (in its mono-, di- or trimeric form). It is proposed
that these complexes involve 3HBA complexing the silicon
centre via the alkoxy group as in structure F. Such com-
plexes can arise by simple alkoxy exchange as in step 4a.
Discussion
The reaction of carboxylic acids with alkoxysilanes has been
studied since at least 1928[9] and been suggested as a method
for the preparation of esters since 1956,[10] but the mecha-
nism was for a long time unclear. Mechanisms involving
Chem. Eur. J. 2007, 13, 4654 – 4664
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4661

R. J. Ansell et al.
SiðOCH3Þ4 þ HOꢀYꢀCOOH $
ðCH3OÞ3SiOꢀYꢀCOOH þ CH3OH
lived, so not detected by NMR spectroscopy. If 5 forms,
however, it could react with nucleophiles, either at the sili-
con centre (which is more electropositive than in TMOS) or
at the carbonyl carbon atom. Attack by MeOH at the sili-
con, would lead, by steps 4c, 4b and 4a in reverse, to the
rapid exchange of OCD3 and OCH3 groups observed in
CD3OD. Attack by H2O (or D2O) at the silicon would cause
hydrolysis, by steps 4d and 4e, as observed by the rapid loss
of SiOCD3 in the 13C NMR spectra. As 7 does not appear in
the 29Si NMR spectra, it is proposed that it reacts rapidly,
probably with 5 as shown in steps 4k and 4l, leading to the
acceleration of condensation, and Q1 species, such as 14, ob-
served soon after mixing in the 29Si spectra. There is also the
possibility of 5 reacting with another molecule of 2HOA as
in 4 f, by means of acid-catalysed alkoxy exchange, to form
the 2:1 complex 8. Compound 8 could undergo elimination
to form 9, which may happen particularly under ES condi-
tions such that 9 is detected in the spectra of TMOS with all
of the 2HOAs. Compound 9 is reminiscent of the stable
structures isolated from solution by Tacke et al.[18,19] and is
presumably stabilised by the exactly sufficient amount of
electron donation from the five ligands to the silicon.
ð4aÞ
In the case of 3HBA (Y=(CH2)3)), the ligand is proposed
to have no effect on the reactions of the other three groups,
which can subsequently undergo the reactions in steps 3a–
3d (or in step 4a again). The 3HBA-siloxane complexes are
present in very small amounts, such that they are not detect-
ed in the NMR spectra, and the MS signals are very weak.
Even smaller amounts of complexes with two siloxane units
are formed (the two peaks in the MS spectrum are yet
smaller, and no Q1 signal is detected in the 29Si NMR spec-
trum). With the 2HOAs, GA, LA and 2HBA, however, the
situation is quite different. 1H and 13C NMR spectroscopy
confirm esterification is hugely accelerated, as are processes
in steps 3a–3d. MS indicates that complexes with silicon
form and they are present at higher concentrations and/or
more readily ionised than those with 3HBA. To explain
these observations, it is proposed that the complexes with
2HOAs can cyclise and eliminate MeOH as in steps 4b
and 4c in Scheme 1 below, whereas complexes with 3HBA
do not. The initial complex 3 would again be in equilibrium
with the starting materials and present in very small
amounts. The complexes 4 and 5 may be unstable and short-
It is proposed that for the 2HOAs mixed with TMOS,
esterification occurs via the attack of MeOH at the carbonyl
carbon of 5, forming a tetrahedral intermediate 10 as shown
Scheme 1. Proposed mechanism for reaction of 2HOAs+TMOS in MeOH.
4662
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 4654 – 4664

2-Hydroxycarboxylic Acids
FULL PAPER
in step 4h. This may be accelerated relative to the equiva-
lent attack on free 2 for two reasons: firstly, the inductive
effect of the silicon and secondly, L3SiOꢀ is a much better
leaving group than OHꢀ.
tached to the same silicon atom, as in 17, will the ES condi-
tions result in an ionic pentavalent species, which is stabi-
lised in the same way as 9 as discussed above.
As species, such as 15, are postulated to form, it is neces-
sary to explain why these do not ionise, even though the
equivalent species do give peaks in the MS for the mixture
of TMOS+H2O in CH3OH (Figure 5, Table 3). Firstly, these
compounds may ionise more easily in the mixture including
H2O than in the less-polar CH3OH-based solution. Secondly,
the peaks in the spectrum for TMOS+H2O are in any case
very much weaker than the major peaks observed for the
mixtures of TMOS+2HOAs.
Alternative mechanisms might equally account for our ob-
servations. Thus, esterification and SiOSi bond formation
could be concerted as in Sharpꢁs mechanism for non-2-hy-
droxy-carboxylic acids.[13] This is particularly appealing
since, for GA for example, the concentration of SiOSi link-
ages (estimated from the integrals of Q1 and Q2 species in
Figure 4c and d) appears to increase in line with the concen-
tration of ester (in Figure 2b). However, such a mechanism
would be necessarily more complex. Scheme 1 is sufficient
to explain the simultaneous production of SiOSi and ester, if
monomeric 7 (a byproduct of esterification) reacts faster
with 5 (or similar oligomers) due to the increased electro-
philic nature of the silicon, than it would with TMOS or a
silicon with four other alkoxy or hydroxy ligands. The
scheme does not yet, however, explain the details of the MS
spectra observed. Why, unlike the complexes with 3HBA,
do all the siloxane complexes with 2HOAs in the MS in-
clude exactly two 2HOA moieties and between one and
four siloxane units?
Repeated cycles of the hydrolysis, esterification and con-
densation steps can rapidly lead to a diverse array of oligosi-
loxane species—some examples are indicated in Scheme 2.
It can be explained why no 1:1 (siloxane:2HOA) complexes
are observed. Under the vacuum conditions of the electro-
spray source, complexes 3 and 4 undergo elimination of
MeOH forming 5, which is uncharged. This route is less fa-
vourable for TMOS+3HBA, as it would require formation
of a six-membered cycle, hence 1:1 complexes are observed
with 3HBA. Further, although oligomeric species, such as
16, are postulated to form under ES conditions they also
eliminate MeOH forming cyclic species as shown, which
also do not ionise. Only when two 2HOA moieties have at-
Conclusion
It has been shown that 2HOAs are methylated selectively
by TMOS in MeOH at room temperature, and a mechanism
consistent with data obtained by means of 1H, 13C and
29Si NMR spectroscopy as well as FTIR spectroscopy and
ESMS has been proposed. The reaction may form a useful
route for the preparation of 2-methoxy carboxylic acids, as
it proceeds at room temperature, without rigorous exclusion
of water, and the reagent TMOS is very inexpensive.
In the proposed mechanism, 2HOAs attach to silicon cen-
tres via the alkoxy group, forming L3Si(OCHXCOOH).
U
Such species are not detected in NMR spectra because they
are present at low concentrations in equilibrium with L3Si-
(OCH3) and the chemical shifts for the silicon, a-carbon and
a-hydrogen may scarcely change from those in the starting
materials. Subsequently the bound 2HOA also interacts with
the silicon via the carboxyl group in an intramolecular rear-
rangement to form an unstable and reactive cyclic inter-
mediate. This intermediate may lead to accelerated methyla-
tion of the carboxylic acid via the nucleophilic attack of
methanol at the carbonyl group, while a separate reaction
pathway leads to condensation of silicon centres leading to
oligosiloxanes. The cyclic intermediate is considered to be
short-lived, such that it cannot be detected spectroscopically,
but its existence is supported by the presence of 1:1 siloxa-
ne:acid complexes in the MS of TMOS+3HBA, but not the
MS of TMOS+2HOAs. The existence of ions in the MS
spectra for TMOS+2HOAs with always two 2HOA moiet-
ies is considered to be due to the stabilisation of pentavalent
silicon with two chelating 2HOAs. These are probably pres-
ent only at low concentration in solution, but other com-
plexes do not ionise.
The existence of the five-membered cyclic intermediates,
even transiently, may have implications for the use of
2HOAs as templates in the sol–gel synthesis of structured
silica. Further work is being conducted on the interactions
of TMOS with di- and tricarboxylic acids containing the
2HOA group.
Scheme 2. Examples of some of the oligomeric species that may be
formed in the reactions of 2HOAs with TMOS by repeated hydrolysis,
esterification and condensation steps, and the “daughter species” that
may arise under ES conditions.
Chem. Eur. J. 2007, 13, 4654 – 4664
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4663

R. J. Ansell et al.
[9] A. W. Dearing, E. E. Reid, J. Am. Chem. Soc. 1928, 50, 3058–3062.
[10] G. Sumrell, G. E. Ham, J. Am. Chem. Soc. 1956, 78, 5573–5575.
[11] a) C. Friedel, J. M. Crafts, Ann. Chim. Phys. 1866, 4, 5–51; b) H. W.
Post, C. H. J. Hofrichter, J. Org. Chem. 1940, 5, 443–448; c) R. P.
Nahrain, R. C. Mehrotra, J. Indian Chem. Soc. 1962, 39, 855–859.
[12] K. C. Chen, T. Tsuchiya, J. D. Mackenzie, J. Non-Cryst. Solids 1986,
81, 227–237.
[13] K. G. Sharp, J. Sol–Gel Sci. Technol. 1994, 2, 35–41.
[14] a) R. Nakao, K. Oka, T. Fukumoto, Bull. Chem. Soc. Jpn. 1981, 54,
1267–1268; b) M. A. Brook, T. H. Chan, Synthesis 1983, 201–203.
[15] a) C. M. Langkammerer (E. I. du Pont de Nemours, US), US
2490691, 1949; b) B. E. Ostberg (Polaroid Corp, US), US 2397287,
1946; c) H. G. Emblem, K. Hargreaves, C. E. Oxley, J. Appl. Chem.
1968, 18, 97–99.
[16] a) R. C. Mehrotra, B. C. Pant, J. Indian Chem. Soc. 1963, 40, 62 3–
628; b) R. C. Mehrotra, B. C. Pant, Indian J. Chem. 1963, 1, 380–
381; c) R. C. Mehrotra, R. P. Narain, Indian J. Chem. 1968, 6, 110–
111.
[17] R. Portanova, L. H. J. Lajunen, M. Tolazzi, J. Piispanen, Pure Appl.
Chem. 2003, 75, 495–540.
[18] a) R. Tacke, C. Burschka, I. Richter, B. Wagner, R. Willecke, J. Am.
Chem. Soc. 2000, 122, 8480–8485; b) I. Richter, M. Penka, R. Tacke,
Inorg. Chem. 2002, 41, 3950–3955; c) R. Tacke, R. Bertermann, C.
Burschka, S. Dragota, Z. Anorg. Allg. Chem. 2004, 630, 2006–2012;
d) R. Tacke, M. Penka, F. Popp, I. Richter, Eur. J. Inorg. Chem.
2002, 1025–1028.
[19] a) M. Muehleisen, R. Tacke, Organometallics 1994, 13, 3740–3742;
b) M. Muehleisen, R. Tacke, Chem. Ber. 1994, 127, 1615–1617; c) R.
Tacke, A. Lopez-Mras, P. G. Jones, Organometallics 1994, 13, 1617–
1623; d) R. Tacke, R. Bertermann, C. Burschka, S. Dragota, M.
Penka, I. Richter, J. Am. Chem. Soc. 2004, 126, 14493–14505; e) R.
Tacke, M. Muehleisen, P. G. Jones, Angew. Chem. 1994, 106, 1250–
1251; Angew. Chem. Int. Ed. Engl. 1994, 33, 1186–1188.
[20] a) F. Brunet, B. Cabane, M. Dubois, B. Perly, J. Phys. Chem. 1991,
95, 945–951; b) J. Sanchez, A. McCormick, J. Phys. Chem. 1992, 96,
8973–8979.
Experimental Section
1H NMR spectra were recorded on a Bruker Avance 500 spectrometer
operating at 500 MHz. Proton-decoupled 13C NMR spectra were record-
ed on a Bruker DPX 300 spectrometer operating at 75.5 MHz. Proton-
decoupled 29Si NMR spectra were recorded on a Bruker DRX 500 spec-
trometer operating at 99.4 MHz, an aquisition time of 0.43 s with a delay
of 5 s was used and the number of scans was 256, unless otherwise stated.
Acid (0.375 mmol) was dissolved completely in [D4]methanol (0.75 mL),
and 1H and 13C spectra were recorded. Subsequently Si
(0.442mmol) was added immediately before placing the sample into the
(OMe)4
magnet and further spectra of the acid+Si(OMe)4 mixtures recorded at
T
increasing intervals. 1H and 13C spectra were recorded in 0.5”18 cm boro-
silicate glass tubes. Amounts for 29Si NMR spectra are four times the
quantities above and spectra were recorded in 1.0”21 cm teflon tubes.
29Si NMR spectra were recorded only after mixing acid+Si
ACHTREUNG
were compared to a Si
ACHTREUNG
tive to SiMe4 as internal standard. NMR spectra were recorded in a labo-
ratory maintained at 188C. Between measurements, samples were left at
RT in a laboratory without temperature control (15–258C).
Mass spectra were recorded on a Bruker micrOTOF instrument by using
direct injection electrospray introduction in negative-ion mode with a
total capillary voltage of 4500 V and a capillary exit voltage of ꢀ100 V.
Prior to injecting the sample, LiOOCH (10 mm) in MeOH/iPrOH (9:1 v/
v) was injected at 9 mLminꢀ1 to provide a calibration signal. The capillary
was then flushed with H2O/MeCN (1:1 v/v) at 0.3 mLminꢀ1 before the
sample was introduced at 4 mLminꢀ1. Data were collected for at least
2min or until a constant signal was obtained. Mass spectra were recorded
in a laboratory maintained at 188C. Prior to measurements, samples were
left at RT in a laboratory without temperature control (15–258C).
Acknowledgement
[21] M. Mazur, V. Mlynarik, M. Valko, P. Pelikan, Appl. Magn. Reson.
1999, 16, 547–557.
[22] J. Schraml, J. Pola, V. Chvalovsky, M. Magi, E. Lippma, J. Organo-
met. Chem. 1973, 49, C19–C21.
The authors are grateful to Tanya Marinko-Covell for technical assistance
with the MS, and acknowledge the support of the EPSRC via grant
number EP/C010973/1.
[23] T. N. M. Bernards, M. J. van Bommel, A. H. Boonstra, J. Non-Cryst.
Solids 1991, 134, 1–13.
[1] E. Haslam, Tetrahedron 1980, 36, 2409–2433.
[2] T. A. Houston, B. L. Wilkinson, J. T. Blanchfield, Org. Lett. 2004, 6,
679–681.
[3] T. Maki, K. Ishihara, H. Yamamoto, Org. Lett. 2005, 7, 5047–5050.
[4] C. J. Brinker, G. W. Scherer, Sol–gel science: The physics and
chemistry of sol–gel processing, Academic Press, London, 1990.
[5] E. J. A. Pope, J. D. Mackenzie, J. Non-Cryst. Solids 1986, 87, 185–
198.
[24] a) L. M. Ng, Macromolecules 1995, 28, 6471–6476; b) G. Zehl, S.
Bischoff, F. Mizukami, H. Isutzu, M. Bartoszek, H. Jancke, B.
Lucke, K. Maeda, J. Mater. Chem. 1995, 5, 1893–1897.
[25] T. Løver, W. Henderson, G. A. Bowmaker, J. M. Seakins, R. P.
Cooney, J. Mater. Chem. 1997, 7, 1553–1558.
[26] P. Bussian, F. Sobott, B. Brutschy, W. Schrader, F. Schuth, Angew.
Chem. 2000, 112, 4065–4069; Angew. Chem. Int. Ed. 2000, 39, 3901–
3905.
[6] a) C. B. Hurd, J. Phys. Chem. 1936, 40, 21–26; b) E. I. Klabunovskii,
A. E. Agronomov, L. M. Volkova, A. A. Balandin, Izv. Akad. Nauk
SSSR Ser. Khim. 1963, 228–234; c) H. Izutsu, F. Mizukami, Y. Kiyo-
zumi, K. Maeda, J. Am. Ceram. Soc. 1997, 80, 2581–2589; d) H.
Izutsu, F. Mizukami, T. Sashida, K. Maeda, Y. Kiyozumi, Y. Akiya-
ma, J. Non-Cryst. Solids 1997, 212, 40–48; e) J. B. Pang, K. Y. Qiu,
Y. Wei, J. Non-Cryst. Solids 2001, 283, 101–108; f) J.-Y. Zheng, J. B.
Pang, K. Y. Qiu, Y. Wei, J. Inorg. Organomet. Polym. 2001, 11, 103–
113; g) H. Nakamura, Y. Matsui, J. Am. Chem. Soc. 1995, 117,
2651–2652; h) F. Miyaji, S. A. Davis, J. P. H. Charmant, S. Mann,
Chem. Mater. 1999, 11, 3021–3024.
[27] R. E. Bossio, S. D. Callahan, A. E. Stiegman, A. G. Marshall, Chem.
Mater. 2001, 13, 2097–2102.
[28] K. Eggers, T. Eichner, J. Woenckhaus, Int. J. Mass Spectrom. 2005,
244, 72–75.
[29] N. S. Leznov, L. A. Sabun, K. A. Andrianov, Zh. Obshch. Khim.
1959, 29, 1518–1522; K. A. Andrianov, in Metalorganic Polymers,
Vol. 8, Wiley, New York, 1965, pp. 168–177.
[30] a) A. Campero, R. Arroyo, C. Sanchez, J. Livage in Ultrastructure
processing of advanced ceramics (Eds.: J. D. Mackenzie, D. R.
Ulrich), Wiley, New York, 1988; b) C. Sanchez, J. Livage, M. Henry,
F. Babonneau, J. Non-Cryst. Solids 1988, 100, 65–76.
[7] C. M. Zaremba, G. D. Stucky, Curr. Opin. Solid State Mater. Sci.
1996, 1, 425–429.
[8] a) H. G. M. Hill, J. A. Nuth, Astrobiology 2003, 3, 291–304; b) I.
Parsons, M. R. Lee, J. V. Smith, Proc. Natl. Acad. Sci. USA 1998, 95,
15173–15176.
Received: December 21, 2006
Published online: April 19, 2007
4664
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 4654 – 4664