Nucleophilic attack of R-lithium at tetrahedral silicon in alkoxysilanes. An alternate mechanism

Article abstract of DOI:10.1246/bcsj.20160039

The currently accepted mechanism for nucleophilic attack at silicon in tetraalkoxysilanes, e.g. Si(OEt)₄ is suggested to involve formation of penta- and then hexacoordinated intermediates as supported by the apparent exclusive formation of R₃SiOR′ and R₄Si from nucleophilic attack by RLi and RMgX. Our recent discovery of a direct route from biogenic silica to tetraalkoxyspirosiloxanes prompted us to revisit this reaction as a potential route to diverse silicon-containing species with single SiC bonds as early studies demonstrate that spirosiloxanes form quite stable pentacoordinated alkoxysilane compounds. As anticipated, Si(2-methyl-2,4-pentanediolato)₂ (SP) reacts with RLi (R = Ph, anthracene, phenylacetylene, etc.) at -78 °C to form pentacoordinated Si, e.g. LiPhSP equilibrates with the starting reagents even at 3:1 ratios of PhLi:SP with no evidence for formation of hexacoordinated species by mass spectral, NMR and quenching studies. Thus, quenching with MeI or Me₃SiCl allows isolation of monosubstituted products from RLi:SP; RSi(OR′)₃ including some ring-opened oligomers. Comparative studies of reactions of PhLi with Si(OEt)₄ allows isolation of mono- and disubstituted products again even at 1:1 ratios of PhLi:Si(OEt)₄. However, on standing at -78 °C for long periods of time or on warming to 0 °C, the primary product for both reactions is Ph₄Si even with 0.5 equivalents of PhLi. At reaction temperatures ≥0 °C the primary product is again Ph₄Si. These results suggest that hexacoordinated intermediates are not part of the substitution mechanism and may suggest that the higher-substituted compounds arise from disproportionation processes. We also briefly describe the conversion of anthracenylSP and 9,9-dimethylfluoreneSP to silsesquioxanes.

Full text of DOI:10.1246/bcsj.20160039

Nucleophilic Attack of R-lithium at Tetrahedral Silicon
in Alkoxysilanes. An Alternate Mechanism
Joseph C. Furgal1 and Richard M. Laine*2,3
1Department of Chemistry, University of Michigan, Ann Arbor, Michigan 48109-2136, USA
2Department of Materials Science and Engineering, University of Michigan,
Ann Arbor, Michigan 48109-2136, USA
3Department of Macromolecular Science and Engineering, University of Michigan,
Ann Arbor, Michigan 48109-2136, USA
E-mail: talsdad@umich.edu
Received: January 31, 2016; Accepted: February 22, 2016; Web Released: February 29, 2016
Richard M. Laine
Richard M. Laine received his Ph.D. in Chemistry (1973) from the University of Southern California
supervised by Professor Robert Bau. He did postdoctoral work before moving to Stanford Research Institute
(1976–87). He moved to the Dept of Materials Science and Engineering, University of Michigan in 1990 and
became a full professor in 1995. Most recently he served as director of Macromolecular Science and
Engineering Center (2006–2015). He is an ACS Fellow (2015) and an International Fellow of the Society for
Polymer Science of Japan (2013). His research interests include the synthesis and processing of organic/
inorganic hybrid materials.
Abstract
pounds arise from disproportionation processes. We also briefly
describe the conversion of anthracenylSP and 9,9-dimethyl-
fluoreneSP to silsesquioxanes.
The currently accepted mechanism for nucleophilic attack
at silicon in tetraalkoxysilanes, e.g. Si(OEt)4 is suggested to
involve formation of penta- and then hexacoordinated inter-
mediates as supported by the apparent exclusive formation of
R3SiOR¤ and R4Si from nucleophilic attack by RLi and RMgX.
Our recent discovery of a direct route from biogenic silica to
tetraalkoxyspirosiloxanes prompted us to revisit this reaction
as a potential route to diverse silicon-containing species with
single Si-C bonds as early studies demonstrate that spirosilox-
anes form quite stable pentacoordinated alkoxysilane com-
pounds. As anticipated, Si(2-methyl-2,4-pentanediolato)2 (SP)
reacts with RLi (R = Ph, anthracene, phenylacetylene, etc.) at
¹78 °C to form pentacoordinated Si, e.g. LiPhSP equilibrates
with the starting reagents even at 3:1 ratios of PhLi:SP with
no evidence for formation of hexacoordinated species by mass
spectral, NMR and quenching studies. Thus, quenching with
MeI or Me3SiCl allows isolation of monosubstituted products
from RLi:SP; RSi(OR¤)3 including some ring-opened oligo-
mers. Comparative studies of reactions of PhLi with Si(OEt)4
allows isolation of mono- and disubstituted products again
even at 1:1 ratios of PhLi:Si(OEt)4. However, on standing at
¹78 °C for long periods of time or on warming to 0 °C, the
primary product for both reactions is Ph4Si even with
0.5 equivalents of PhLi. At reaction temperatures ²0 °C the
primary product is again Ph4Si. These results suggest that
hexacoordinated intermediates are not part of the substitution
mechanism and may suggest that the higher-substituted com-
Introduction
Silicon chemistry is dominated by compounds made start-
ing from the products of the direct process, reaction (1).1
Other products can be made from PhSiCl3 available via
reaction (2).2 Finally, hydrosilylation as exemplified by reac-
tion (3), provides access to other types of monofunctional
silanes.3,4
Cu=Sn cat=300ꢁC
Simet þ MeCl ꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀ! MeSiCl3 þ Me2SiCl2
þ MeHSiCl2 ðpolysiloxane precursorsÞ
ð1Þ
ð2Þ
ð3Þ
BCl cat=350ꢁC
3
Ph-H þ SiCl4 ꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀ! PhSiCl3
cat
HSiCl3 þ CH2=CH-R ꢀꢀ! R-CH2CH2SiCl3
However, if one wants to make organosilicon compounds
starting from organic compounds that do not contain an acces-
sible double bond or compounds without phenyl or methyl
groups, very few synthetic avenues remain. These typically
involve reactions of nucleophiles with chloro- or alkoxysilanes
as illustrated in reactions (4) and (5).
solvent
SiCl4 þ RLi ðRMgBrÞ ꢀꢀꢀꢀ! R4ꢀxSiClx
þ ðx ¼ 0{3Þ þ LiCl ðMgBrClÞ
ð4Þ
Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan | 705

R2R¤SiOEt.7 They could never obtain RR¤Si(OR)2, again
supporting the likelihood that pentacoordinate species react
faster than tetracoordinate ones. Although their reactions were
initiated at ¹78 °C, they were allowed to warm to room tem-
perature prior to quenching with water.
1. 4 RMgX
2. LiAlH4
O
O
O
O
Na2
SiR3H
50% Et, 60% Bu
Si
Et
O Reflux
2
In relatively recent work, Manoso et al.10 and later Jung
et al.11 found that ¹78 °C aryl lithium and ¹30 °C Grignard
reagents react with TEOS or Si(OMe)4 (TMOS) to produce
monofunctional alkoxysilanes, RSi(OR)3. In these studies it was
found that low-temperature reaction conditions, excess TEOS
(>2 equiv) and low-temperature quenching are key aspects in
obtaining monofunctional products. Manoso et al. also found
that nucleophilic substitution reactions worked best with elec-
tron poor as opposed to electron-rich aryl reagents, which are
more reactive and tended to form R2Si, R3Si, and R4Si as
found by Corriu. They also reported that reaction time has little
influence on reaction products and yields, regardless of temper-
ature and RLi(MgBr):TEOS stoichiometries (e.g. even at 1:3
ratios) always providing some quantities of di- and trialkoxy-
silanes. In contrast to the work of Tour, these researchers
quenched with water at ¹78 °C prior to allowing the reactions
to warm to room temperature. It appears that warming before
quenching makes quite a difference in the spectrum of products.
Our interest in this area is driven by the need to develop
new ways to make arylsilsesquioxanes given our extensive
work on their properties.17-22 We have previously reported the
facile synthesis of pentacoordinated spirosilicates from any
silica source in ethylene glycol.23 The glycolato silicate of
reaction (6) can be isolated in quantitative yield and appears to
be quite stable suggesting that spirosiloxanes might be better
candidates for stabilizing the monofunctionalized species as
suggested by reaction (7).
RMgX or RLi
SiR3OR
50% Et, 60% Bu
Si(OR)4
Scheme 1. Typical reactions of triscatecholatosilicate or
tetralkoxysilanes with Grignard and lithium reagents.5-16
:Nu
+ X-
Nu
R3Si
R3SiX
Cat
:Nu/-Cat
X
R
R
Si
R
Cat
Cat = F-, RCO2 , HMPA, etc
-
Scheme 2. Proposed mechanism for nucleophilic substitu-
tion at Si via pentacoordinate intermediates.15,16
solvent
SiðORÞ4 þ R0Li ðRMgBrÞ ꢀꢀꢀꢀ! R04ꢀxSiðORÞx ðx ¼ 0{3Þ
þ LiOR ðMgBrORÞ
ð5Þ
Reaction (4) proceeds reasonably well but one must contend
with chlorosilanes before and after reaction and their hydrolysis
products, which require extra care and often afford unwanted
complications including the formation of intractable polymers.
Consequently, multiple groups have explored nucleophilic
attack at Si(OEt)4 (TEOS) in particular using strong nucleo-
philes, however no others have targeted the synthesis of
silsesquioxane precursors [RSi(OR¤)3] and more importantly
difunctional siloxanes [-R2SiO- or R1R2SiO-] where R is not
methyl.
-3H2O/200°C
SiO2 + MOH + excess
HO
OH
M = Li, Na, K, etc
Reaction (5) also suffers from an important complication
because nucleophilic substitution often does not stop at the
monofunctional silane but rather commonly continues to the
tri- and tetrafunctional products even when the alkoxysilane is
used in excess, as suggested in Scheme 1.5-11 Thus, nucleo-
philic attack at tetrahedral alkoxysilanes differs considerably
from that of carbon-containing compounds.
O
O
M
Si
OH
O
2
20
ð6Þ
This surprising result has prompted multiple studies on
nucleophilic attack at tetrahedral silicon. Extensive studies by
the Corriu group provide considerable evidence that nucleo-
philic attack occurs much faster at penta- and hexacoordinated
rather than tetrahedral silanes.12-16 They have proposed a
O
O
Si
O
O
¹
general mechanism as illustrated in Scheme 2.15,16 Here F
¹
or RCO2 would be a catalyst for reactions not involving
PhLi
O
Grignard or Li reagents.
O
Si
ð7Þ
-78 °C/Et2O
This mechanism seems to provide an explanation for many
of the results reported to date.5-11 Corriu et al. argue that in the
absence of catalyst, the nucleophile takes the position “Cat” in
Scheme 2; thus, double and triple functionalization can be
expected. Many of the Corriu et al. studies were run in ether
at or near reflux temperatures (μ30 °C).
Tour et al. demonstrated that it was possible to get two of
the same R groups to add to Si using alkyl-/aryllithium rea-
gents and thereafter add a different R¤Li group to produce
O
O
Li
Thus, our recent discovery of the direct depolymerization of
biogenic silica sources such as rice hull ash (RHA) to produce
spirosiloxanes per reactions (8) and (9) offered an exceptional
opportunity to develop greener routes to special mono- and
difunctional silicon-containing compounds.24 We report here
our first successful efforts to develop such routes.
706 | Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan

¹78 °C (dry ice/acetone bath) for 10 min under argon (room
temperature for phenylacetylide). Then 10 mL of dry argon
purged diethyl ether was added to the reaction and allowed to
cool for another 10 min. This was followed by 2.2 mL (5.5
mmol) of n-butyllithium solution in diethyl ether. The reaction
was allowed to stir for 30 min to complete the halide-exchange
reaction, at which time the reactions turned a yellow color and
salt precipitates formed. Then 1 g (3.8 mmol) of spirosiloxane
I dissolved in 5 mL of cooled dry argon purged diethyl ether
was added by syringe to the reaction mixture and allowed to
stir for an additional 2 h. The reaction was run till the yellow
color subsided and was then quenched with excess Me3SiCl
and allowed to stir cold for >30 min before warming up. The
reaction was then worked up by a quick water wash (2 times) to
remove salts and then dried over MgSO4. The reaction was then
filtered through Celite to remove salts and solvent removed
in vacuo. Yellow-orange viscous oils were obtained. Character-
ization is given in tabular form in the text.
OH
OH
-H2O
RHA + 10 mol % NaOH +
2-Methyl-2,4 pentanediol
O
O
ð8Þ
Si
O
O
I
H
O
200 °C/-H2O
RHA + 10 mol % NaOH +
O
H
Tetraethoxysilane: PhLi Reaction. To a flame dried, 3
times evacuated and argon purged 25 mL Schlenk flask was
added 1.07 mL (4.8 mmol) of TEOS and a magnetic stir bar.
The flask was then cooled to either ¹40 [dry ice/ethylene
glycol (0.6):ethanol (0.4)] or ¹78 °C (dry ice/acetone) bath for
10 min under argon. Then 15 mL of dry argon purged diethyl
ether was added to the reaction and allowed to cool for another
10 min. Then 4.8 mmol of a PhLi solution was added dropwise
by syringe to the reaction mixture, which became yellow in
color. The reaction was run till the yellow color subsided and
was then quenched with excess Me3SiCl or MeI and allowed to
stir cold for >30 min before warming up. The reaction was then
worked up by a quick water wash (2 times) to remove salts
and then dried over MgSO4. The mixture was then filtered and
the solvent removed in vacuo. A clear semi-viscous oil was
obtained. Characterization is given in tabular form in the text.
Silsesquioxane Synthesis (Ph-Example).21,22 Briefly, 150
mg of oligomeric monoPh-I were added to a 50 mL round
bottom flask with a magnetic stirrer. Then 30 mL of CH2Cl2
was added to the reaction, followed by 20 ¯L of water and
25 ¯L of 1 M TBAF solution in THF. The reaction was left to
stir for 16 h, and was then quenched with 50 mg of CaCl2 to
remove fluoride. The reaction was then filtered through Celite,
and solvent removed in vacuo. The product was then dissolved
in a small amount of CH2Cl2 and precipitated into methanol
to remove partial cage by-products. The precipitate was then
filtered, giving 62 mg of isolated PhSQ products in a mixture of
cage sizes (T8, T10, and T12). Other cages were synthesized in a
similar manner. Characterization data is given in the text.
Analyical Methods. NMR Analyses: 1H and 29Si NMR
were measured in diethyl ether or chloroform-d with TMS (0.00
ppm) as the internal reference on a Varian VNMRS 500 spec-
O
O
O
O
ð9Þ
Si
Experimental
Materials. Chemicals and solvents were obtained from
commercial suppliers and used without further purification,
unless otherwise indicated. Spirosiloxane I was synthesized as
described elsewhere.24 All other chemicals were purchased from
Fisher Scientific or Aldrich Chemical and used as received.
Spirosiloxane: R-lithium Reaction. To a flame dried and
argon purged 100 mL Schlenk flask was added 2.5 g (10 mmol)
of spirosiloxane I and a magnetic stir bar. The flask was then
evacuated under vacuum 3 times and purged with argon. The
flask was then cooled to either ¹40 [dry ice/ethylene glycol
(0.6):ethanol (0.4)] or ¹78 °C (dry ice/acetone bath for 10 min
under argon. Then 75 mL of dry argon purged diethyl ether was
added to the reaction and allowed to cool for another 10 min.
Then 10 mmol of R-lithium solution in diethylether was added
dropwise by syringe to the reaction mixture, at which time the
reaction became a yellow color. The reaction was run till the
yellow color subsided and was then quenched with excess
Me3SiCl or MeI and allowed to stir cold for >30 min before
warming up. The reaction was then worked up by a quick water
wash (2 times) to remove salts and then dried over MgSO4. The
reaction was then filtered through Celite to remove salts and sol-
vent removed in vacuo. A clear semi-viscous oil was obtained.
Further purification on select samples was achieved by bulb-to-
bulb distillation under reduced pressure. Characterization is
given in tabular form in the text. Reaction conversions were
calculated by integrating the NMR peaks for each species and
determining the ratio of the desired product versus the total area.
1
trometer. H NMR spectra were collected at 500 MHz using a
7998.4 Hz spectral width, a pulse width of 45°, relaxation delay
of 0.5 s, 65K data points. 13C NMR spectra were collected at
100 MHz using a 25000 Hz spectral width, a pulse width of 40°,
relaxation delay of 1.5 s, and 75K data points. 29Si NMR spectra
were collected at 99.35 MHz using a 4960 Hz spectral width, a
pulse width of 7°, a relaxation delay of 15 s, and 4K data points.
Mass Spectroscopy (MS): Electron impact (EI) analyses
were conducted using a VG 70-250-S magnetic sector instru-
Spirosiloxane: R-arylbromide Reaction.
To a flame
dried, 3 times evacuated and argon purged 25 mL Schlenk flask
was added 4.6 mmol of aryl bromide (2-bromo-9,9-dimethyl-
fluorene, 9-bromoanthracene, and 1-bromo-4-methylnaphtha-
lene) and a magnetic stir bar. The flask was then cooled to
Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan | 707

ment (Waters) by electron impact ionization (EI). The instru-
ment was calibrated with perfluorokerosene-H. The samples
were run in EI mode at 70 eV electron energy with an ion
source temperature of 240 °C. The mass range was scanned
from m/z 1000 to 35. High-resolution mass spectrometry
(HRMS) analysis was conducted on the same instrument.
Electrospray ionization (ESI) was performed on an Agilent Q-
TOF system with a dual ESI ion source. The mobile phase con-
sisted of a 9:1 mixture of acetonitrile:water with 0.1% formic
acid. Lockmass correction was used to obtain mass accuracy.
Matrix-assisted laser desorption ionization time-of-flight
(MALDI-TOF) was performed on a Micromass TofSpec-2E
equipped with a 337 nm nitrogen laser in positive-ion reflectron
mode using poly(ethylene glycol) as calibration standard,
dithranol as matrix, and AgNO3 as ion source. Sample was
reagents (i.e. phenylacetylenyl, anthracenyl, etc.) with I and
characterization studies. The reactions of PhLi with tetra-
ethoxysilane (TEOS) are also discussed. Lastly, several exam-
ples are given using R-I to synthesize novel silsesquioxanes by
fluoride catalyzed cage formation.
Reactions with Spirosiloxane I. Our first steps in these
efforts were to extend the work of Manoso et al. and Jung et al.
to our system to compare the reactions of TEOS with those of
I produced per reaction (8).10,11 As such, studies were initiated
using stoichiometric PhLi/Et2O/¹78 °C to arylate I resulting
in facile formation of monosubstituted PhILi (Reaction 10).
Figure 1 presents the MALDI-TOF spectrum taken by spotting
the ¹78 °C unquenched reaction mixture directly on a sample
plate just prior to analysis.
O
¹1
O
prepared by mixing solution of 5 parts matrix (10 mg mL in
0.5 or 1 equiv PhLi
Si
¹1
THF), 5 parts sample (1 mg mL in THF), and 1 part of
-78 oC, Et2O
O
¹1
O
AgNO3 (2.5 mg mL in water) and then spotting the mixture
I
on a stainless steel target plate.
Gel Permeation Chromatography (GPC): GPC analy-
ses were run on a Waters 440 system equipped with Waters
Styragel columns (7.8 © 300, HT 0.5, 2, 3, 4) with a Waters
2410 refractometer for detection, with THF as the solvent. The
system was calibrated using polystyrene standards and toluene
as reference. Analyses were performed using PL Caliber 7.04
software (Polymer Labs, Shropshire UK).
O
O
O
ð10Þ
Si
O
Li
In contrast, if the above reaction of I with PhLi (1:1 ratio) is
allowed to warm to ambient without quenching then a quite
different product appears, which precipitates out of diethyl
ether allowing easy separation from remaining I. The 29Si NMR
(Figure 2) of this product gives a single peak at ¤ = ¹14.4
ppm, the literature value,25 typical for tetraarylsilanes. The
EI-MS (Figure S1) provides further evidence showing the tetra-
phenyl product peak at 336.0 m/z. If the same reaction is run
with a ratio of 0.5:1 similar peak distributions are observed in
the EI-MS mass spec (Figure S2). This product distribution is
truly unique, since it tells us something about the reaction itself.
Since only monophenyl products are observed at low temper-
ature (Figure 1), it suggests that PhLi reacts with I to form a
pentacoordinated species, which then transfers the phenyls to
other similar species multiple times until Ph4Si forms as the
final product. Corriu et al. report that reacting hexacoordinated
catecholato spirosiloxanes with PhMgBr at 35 °C (Scheme I)
Fourier Transform Infrared Spectroscopy (FTIR):
Diffuse reflectance Fourier transform (DRIFT) spectra were
recorded on a Nicolet 6700 Series FTIR spectrometer (Thermo
Fisher Scientific, Inc., Madison, WI). Optical grade, random
cuttings of KBr (International Crystal Laboratories, Garfield,
NJ) were ground (400 mg) with 5 mg of sample to be analyzed.
For DRIFT analyses, samples were packed and smoothed off to
ensure little scattering. The FTIR sample chamber was flushed
continuously with N2 prior to data acquisition in the range
¹1
¹1
4000-400 cm with a precision of «4 cm
.
Results and Discussion
In the following sections we first discuss the substitution
reactions of PhLi with I, and analyze the effects of temperature
and quenching on the final structures. This is followed by
studies on the stoichiometric reactions of various other RLi
Figure 1. MALDI-TOF of 1:1 spirosiloxane (I):PhLi (¹78 °C), m/z μ 344.4 is the parent ion for LiPhSi(2-methyl-2,4-diolato)2,
which loses PhLi to give I at m/z = 259.
708 | Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan

Figure 2. 29Si NMR of synthesized tetraphenylsilane, ¹14.4 ppm, isolated from a 1:1 PhLi:I reaction run at ¹78 °C then warmed to
room temperature without quenching, and then rinsed with diethyl ether to remove excess I.
O
O
Si
O
O
OH
O
O
Si
O
O
O
O
OH
O
Si
S i
OH
O
O
O
O
S i
O
O
OH
Figure 3. Negative ion ESI mass spectrum of 1:1 PhLi:I after quenching with MeI.
gives only Ph4Si.11,12,21,26,27 It seems plausible to suggest that
based on their structural similarity to T1 species observed in
the synthesis of silsesquioxanes.28,29 The peaks upfield corre-
spond to Si(OR)4 species, including residual I. Table 1 shows
literature values for 29Si NMR of various alkoxysilanes.
these reactions occur via at least a heptacoordinated transition
state rather than a penta- or hexacoordinated one. Running reac-
tions at low temperatures and quenching with an electrophile
before warming, circumvents this reaction pathway, see below.
Quenching the 1:1 PhLi:I reaction with MeI at ¹78 °C results
in products that include ring-opened oligomers per the Figure 3
negative ion ESI-MS. The 352.9 m/z peak is the targeted
quenched product II (Reaction 11), while the other peaks can
be ascribed to ring-opened products with structures suggested in
Figure 3. By-products of the substitution reaction such as dimer
III are also observed in EI-MS at 457.9 m/z (Figure S3).
The Figure 4 29Si NMR shows product peaks at ¹57 ppm
likely corresponding to PhSi(OR)3 [PhSi(OEt)3, Table 1] and at
¹63 to ¹67 ppm likely peaks from dimeric/oligomeric species
containing Si-O-Si linkages (i.e., [Ph(RO)2Si-O-Si(OR)2Ph])
O
O
PhLi
-78 oC, Et2
Quench: MeI
Si
O
O
O
I
O
O
O
O
O
O
O
Si
Si
Si
And/Or
+ Higher Oligomers
O
ð11Þ
O
Me
II
III
A small peak appears at ¹30 ppm that likely corresponds
to Ph2Si(OR)2 [Ph2Si(OEt)2 in Table 1]. Note that even at a 1:1
Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan | 709

Figure 4. 29Si NMR spectrum of the MeI quenched ¹78 °C 1:1 PhLi:I of Figure 2.
Table 1. Literature 29Si NMR chemical shifts of selected
To examine the influence of reaction times; we ran 1:1
30
alkoxysilanes in CDCl3
PhLi:I reactions for 3, 4.5, and 6 h. Figure S4 shows the GPC
of this set of reactions with similar peak retention times and
shapes for the first two samples, while the 6 h reaction shows a
broader peak, perhaps indicative of more extensive ring-
opening polymerization.
Formula
Ph4Si27
Me3SiOEt
29Si NMR/ppm
¹14.0
14.5
¹4.3
Me2Si(OEt)2
Ph2Si(OEt)2
MeSi(OEt)3
MeSi(OPr)3
EtSi(OEt)3
CH2=CHSi(OEt)3
CH¸CSi(OEt)3
PhSi(OEt)3
O
O
¹32.4
¹43.3
¹43.3
¹44.6
¹58.7
¹74.7
¹57.9
¹57.2
¹64.9
PhLi
-78 oC, Et2O
Si
O
O
Quench: Me3SiCl
I
O
O
O
O
O
O
4-MePhSi(OEt)3
+ Higher Oligomers
Si
Si
Si
Si
And/Or
Ph(OEt)2SiOSi(EtO)2Ph28,29
O
V
IV
ð12Þ
reaction ratio, small quantities of unreacted I remain. This is
more prevalent in the TMSCl quenched samples discussed
below. Though Manoso et al. and Jung et al. both found good
yields (>50%) by quenching the TEOS systems with water;
we find that H2O quenching of nucleophilic reactions of I give
many ring-opened by-products, likely due to the formation of
hydroxide ions. To overcome these drawbacks, our quenching
studies used TMSCl and MeI, see reaction (12) and Scheme 3.
Table 2 summarizes the PhLi:I reaction conditions, % conver-
sions, 29Si NMR, and MS results.
Mass spectral analysis reveals two new peaks at 310 m/z
(V, -Me peak at 295 m/z) and 357 m/z (high-resolution mass
spectral analysis indicates a formula of C15H25Si2O6, however
we are unable to assign a structure for this formula). The 357
m/z peak is also more intense after 6 h vs. the 4.5 h reaction.
Figure 7 shows an exemplary EI-MS of the 4.5 h sample. The
Figure 8 29Si NMR for the 4.5 h reaction presents strong peaks
in the ca. ¹65 ppm region, suggesting oligomers similar to
those found in the MeI quenching studies.
TMSCl quenching is complete in half the time needed for
MeI (<30 min) vs. MeI (>1 h). Figure 5 provides an EI-MS
for a 3 h reaction quenched with TMSCl. The desired quenched
product appears at 410 m/z (IV) with the representative
29Si NMR seen in Figure 6. The peak at ¹58 ppm corresponds
to a PhSi(OR)3 product (IV) per Table 1, the peak at ¹82 ppm
is remaining I and the peak at ¹4.8 ppm is TMS-benzene, from
quenching PhLi. ATMS-O- signal would appear at ca. 18 ppm
but is not seen.
One possible explanation for the formation of V and the Si-
O-Si linkages in oligomeric derivatives is through cleavage of
C-O bonds on the diol subunits during quenching. This would
¹
produce Si-O or Si-OH, which thereafter reacts with an
electrophile (TMS), and a chloro-substituted alkyl (from diol)
is made as a by-product. Such reactions are observed for
Michaelis-Arbuzov rearrangement reactions with phospho-
rus,31 and as suggested in the synthesis of phenylsilsesquioxane
from phenyltrichlorosilane.32
710 | Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan

O
O
O
Si
O
O
O
Si
O
O
Si
+
PhLi
PhLi
<-40 oC, Ether, 1h
Quench: SiMe3
<-40 oC, Ether, 1h
Quench: MeI
O
Si
Si
O
O
O
O
Si
O
O
PhLi
PhLi
Any oC, Ether, 1h
Unquenched
>-40 oC, Ether, 1h
Quenched
Si
Si
Scheme 3. Example reaction conditions for the substitution of I with PhLi and their dominant substituted products from Table 2.
Table 2. Reaction conditions and observed characterization data for PhenylLi reactions with spirosiloxane, all reactions were conducted
at ¹78 °C for 2 h (I: spirosiloxane)
Observed MS
29Si NMR
/ppm
Conversiona)/%
(29Si NMR)
Rxn ratio
Structures
Quenched
No
(m/z)
1:1 PhLi:I
(¹78 °C)
336.0
¹14.4
100
Si
1:1 PhLi:I
(¹78 °C)
352.9
¹57.0
MeI
100
Ph1I + PhSi(O)x
O
O
Si
O
O
457.9
678.9
¹65
O
O
O
O
Si
O
ca. (¹65)
O
OH
O
Si
O
O
O
OH
O
N/A
¹30.6
N/A
O
O
Si
344.4
O
O
Si
O
O
Continued on next page:
Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan | 711

Continued:
Rxn ratio
Observed MS
29Si NMR
/ppm
Conversiona)/%
(29Si NMR)
Structures
Quenched
MeI
(m/z)
0.5:1 PhLi:I
(¹78 °C)
336.1
N/A
N/A
Si
260.1, 243.2
O
O
Si
(M+ ¹ Me)
O
O
337.1
O
O
Si
O
O
0.5:1 PhLi:I
(¹78 °C)
336.1
N/A
No
N/A
Si
260.1,
O
O
Si
243.2 (M+ ¹ Me)
O
O
0.6:1 PhLi:I
(¹40 °C)
410.1
18.1, ¹58.9
Me3SiCl
36
Ph1I + PhSi(O)x
O
O
Si
O
O
Si
260.1
¹82.1
O
O
Si
O
O
N/A
¹64 to ¹68
O
OH
O
Si
O
O
O
OH
O
1:1 PhLi:I
410.1
¹58.9
Me3SiCl
46
(¹40 °C, 30 min)
Ph1I + PhSi(O)x
O
O
O
Si
O
Si
260.1
¹82.1
O
O
O
Si
O
N/A
ca. ¹64
O
O
O
O
Si
O
Oligomers
N/A
¹4.8
Si
Continued on next page:
712 | Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan

Continued:
Rxn ratio
Observed MS
29Si NMR
/ppm
Conversiona)/%
(29Si NMR)
Structures
Quenched
(m/z)
1:1 PhLi:I
410.1
18, ¹57
Me3SiCl
47
(¹78 °C, 3 h)
Ph1I + PhSi(O)x
O
O
Si
O
O
Si
259.1
¹82
O
O
Si
O
O
N/A
ca. ¹65
O
O
O
O
Si
O
Oligomers
N/A
¹4.8
Si
1:1 PhLi:I
310.1
9.2, ¹65
Me3SiCl
88
(¹78 °C, 4.5 h)
PhSi(O)x
O
O
Si
Si
O
245.1 (M+ ¹ Me)
N/A
¹82
O
O
O
Si
O
¹66, ¹69
O
O
O
O
Si
O
Oligomers
N/A
¹4.8
Si
Continued on next page:
A general observation that can be made from all of our
studies (see SI for more details) is that there is little change
in the product distribution and starting materials no matter
what reaction conditions are chosen (see below). These results
suggest equilibration between starting materials and products,
is the first step in nucleophilic attack and a significant barrier
exists to a second such attack, contrasting greatly with the work
and mechanisms of reaction suggested by Tour et al., Corriu
et al., Manoso et al., and Jung et al.7,10-12
In all reaction studies done at ¹40 °C or below, 1:1 PhLi:I
reactions almost always contain I in a nearly 1:1 PhI:I ratio
by 29Si NMR. This is especially evident for TMSCl-quenched
samples. To better understand this relationship; non-stoichio-
metric PhLi:I, reactions were run at 0.5 and 3.0 equivalents of
PhLi. At ratios of 0.5:1 and 1:1, nearly identical product distri-
butions are found for TMSCl-quenched samples. At 3 PhLi
equivalents (¹78 °C), the main product while still only mono-
functionalized is now V, where one diol subunit is displaced
by an -OSiMe3 unit as detailed above. Though the nature of
the monosubstituted products differ, the ratio of products to
starting materials changes only slightly and significant amounts
of I and TMS-benzene are found in the 29Si NMR (Figure S5).
This further suggests that a very large excess of PhLi is needed
to push the equilibrium toward monosubstituted products at
¹78 °C, whereas at room temperature this excess would result
exclusively in the tetraphenylsilane derivative. The EI-MS is
similar to that for 1:1 samples, see SI.
Given that Corriu et al. suggest that these types of reac-
tions must take place through hexacoordinated intermediates,
which may still be true at ambient; our ability to generate only
monophenyl-substituted products strongly suggests that the
primary reaction intermediates generated at low temperatures
Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan | 713

Continued:
Rxn ratio
Observed MS
29Si NMR
/ppm
Conversiona)/%
(29Si NMR)
Structures
Quenched
(m/z)
1:1 PhLi:I
(¹78 °C, 6 h)
310.1
¹65
Me3SiCl
73
PhSi(O)x
13
O
O
Si
Si
Ph2I
O
245.1 (M+ ¹ Me)
N/A
¹82
O
O
O
Si
O
¹66, ¹69
O
O
O
O
Si
O
Oligomers
N/A
¹4.8
Si
C15H25O6Si2
357.1
310.1
ca. ¹66
¹65
3:1 PhLi:I
(¹78 °C, 3 h)
Me3SiCl
68
PhSi(O)x
8
O
O
Si
Si
Ph2I
O
245.1 (M+ ¹ Me)
N/A
¹82
O
O
O
Si
O
¹65, ¹69
O
O
O
Si
O
O
Oligomers
N/A
¹4.8
Si
C15H25O6Si2
357.1
ca. ¹66
ca. ¹32
N/A
O
Si
O
a) Percent conversion calculated by 29Si NMR integration ratio of product over total alkoxysilane present.
are pentacoordinated. Reactions run at ¹60, ¹40, 0, and 20 °C.
Those run at ¯¹40 °C gave monophenyl products at different
reaction times, whereas warmer temperatures gave, tetraphenyl-
or triphenylalkoxysilanes.
proposition is significantly different from what Corriu et al.
have argued in the past. It also greatly affects the product
distribution achievable in simple nucleophilic reactions.
Other Nucleophiles.
Based on our desire to generate
Hence it is clear that two mechanisms are operating. Given
that it appears difficult to form large amounts of even the mono-
substituted species, without using a large excess of the nucleo-
phile, it is likely that polysubstituted species arise from dispro-
portation rather than sequential addition. The latter mechanism
should happen easily in the presence of excess nucleophile. This
novel silsesquioxanes, we sought to extend these mechanistic
studies to other nucleophiles, including: MeLi, n-butylLi, 3-
pyridinylLi, Li-TMS-acetylide, Li-acetylide (ethylenediamine
complex), thianaphthenylLi, 2-thienylLi, Li-phenylacetylide,
9-anthracenylLi, 9,9-dimethylfluorenylLi, and Me-naphthal-
enylLi. Trial reactions were run with 1.2:1 RLi:I stoichiometry,
714 | Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan

I
IV
Figure 5. EI-MS of 1:1 PhLi:I after quenching with TMSCl; peaks: 410.3 m/z is IV, 259.1 is I.
Figure 6. 29Si NMR spectrum of the TMSCl quenched ¹78 °C 1:1 PhLi:I. Peaks: ¹4.8 ppm, TMS-benzene; ¹59 ppm PhSi(RO)3;
¹65 ppm Ph-Si-OSiMe3(OR)2.
are un-optimized and are in support of the utility of these reac-
tions. The first five [MeLi, n-butylLi, 3-pyridinylLi, Li-TMS-
acetylide, Li-acetylide (ethylenediamine complex)], do not
provide observable/isolable monomeric products. Only starting
material I, ROP by-products and TMS-substituted starting
materials are observed by mass spectral analysis and/or
29Si NMR. Grignard reagents were also explored; however
either no reaction or unidentifiable products were observed
likely resulting from formation of stable pentacoordinated
complexes with I.33,34
In contrast, 2-thienylLi, Li-phenylacetylide, anthracenylLi,
9,9-dimethylfluorenylLi, and Me-naphthalenylLi all gave prod-
ucts that could be characterized by both 29Si NMR and EI-MS
(Table 3). Most reactions were run using methods akin to those
of Manoso et al.,10 Jung et al.,11 and those for PhLi above, in
which an arylbromide is first converted to a lithium reagent via
exchange with n-butylLi, followed by addition of I (structures
VIII-X). The 29Si NMR and EI-MS for each of these com-
pounds are presented in SI, Figures S6-S12.
The nucleophilic reactions with the best conversions from I
were (ca. 100%) for the anthracenylLi:I (Reaction 13) and Li-
phenylacetylide:I reactions; no I remained in either case by
29Si NMR. The 2-thienylLi:I (VI) shows ca. 18% conversion
of I by 29Si NMR and two types of products as discussed
above for PhI reactions (13.2, 6.2, ¹62 ppm), in which the
TMS quenching group is either attached to the diol, or directly
attached to the central Si atom (Si-O-TMS) (Table 3). The Li-
phenylacetylide:I (VII) reaction with I at room temperature
gives ca. 100% conversion by 29Si NMR (¹78 ppm), with no
remaining spirosiloxane observed, and with product peaks in
Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan | 715

Figure 7. EI-MS of 1:1 PhLi:I (4.5 h, ¹78 °C), quenched with TMSCl; I ¹ Me at 245.1 m/z, V at 310.1 m/z, and V ¹ Me at
295.1 m/z.
Figure 8. 29Si NMR spectrum of 1:1 PhLi:I (4.5 h, ¹78 °C) quenched with TMSCl, Peaks μ55-65 ppm are PhSi(O)x units, peak at
9.17 ppm is Si-O-SiMe3, peak at ¹82 ppm is I.
EI-MS at 334.0 and 434.1 m/z for the two monomeric struc-
ture types respectively (Table 3). The 9,9-dimethylfluorenyl-
Li:I (IX) reaction shows ca. 15% conversion of I by 29Si NMR
(¹52.6 and ¹65.0 ppm) and a clear peak in EI-MS at 426.1 m/z
for an Si-O-TMS-structured monomeric product (Table 3).
Lastly, Me-naphthalenylLi:I (X) shows ca. 29% conversion
of I to Me-naphthalenyl products by 29Si NMR (¹46.1, ¹57.5
ppm). The peak at ¹57.5 ppm corresponds to the expected
structure formed by lithium-bromide exchange on 1-bromo-4-
methylnaphthalene starting material, followed by direct attack
of I (Table 1 for examples). The peak at ¹46.1 ppm, is attrib-
uted to the product from deprotonation of the methyl group by
either n-butylLi or 4-methylnaphthalenylLi (due to its lower
pKa), and attack by the methylenylLi group on I. Alkyltrieth-
oxysilanes typically give 29Si NMR peaks in the ¹45 ppm
region (Table 1). By EI-MS only one product mass is observed
at 402.2 m/z. This peak could correspond to both products
observed by 29Si NMR since they would have the same molec-
ular formula. Note also that the observed mass is for a proton-
quenched derivative, instead of being quenched with TMS
(Table 3). This transfer may occur during the aqueous workup
used to remove salts or is a result from incomplete quenching.
As an example, Figure 9 shows the EI-MS for anthracenylI
(VIII) quenched with Me3SiCl, with the product peak [M] at
410.2 m/z and [M ¹ TMS] at 338.1 m/z. Figure 10 shows the
29Si NMR of VIII, with the product peaks at 10.0 and ¹57.4
ppm, and TMS-anthracene at ¹2.87 ppm. Silsesquioxanes of
VIII will be discussed in that section.
716 | Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan

Table 3. Reaction conditions and observed characterization data for RLi reactions with I, all reactions were conducted at ¹78 °C for 3 h
and quenched with Me3SiCl unless otherwise noted
Observed MS
29Si NMR
/ppm
Conversion/%
(29Si NMR)
Rxn ratio/Conditions
Structure
(m/z)
1.2:1 2-thienylLi:I
316.2
6.66, ¹62.58
18
S
(VI)
0 °C
O
Si
Si
O
O
416.8
13.24, ¹62.06
S
O
O
O
Si
O
Si
S
156.3
¹7.2
¹82
SiMe3
245.1 (M+ ¹ Me)
O
O
O
O
Si
1.2:1 Li-
phenylacetylide:I
(VII)
334.0
434.1
¹78.8
100
O
O
Si
Si
O
25 °C
¹78.3
O
O
O
Si
Si
O
245.1 (M+ ¹ Me)
¹82
O
O
O
Si
O
1.2:1 anthracenylLi:I
(VIII)
272.3
¹53.4, 10.0
100
O
Si
Si
O
O
250.1
438.2
¹2.9
SiMe3
¹53.9
O
O
O
Si
HO
Continued on next page:
As with the PhI derivatives, no simple method was found to
isolate the monomeric products from I. For example bulb-to-
bulb distillation attempts led to distillation of I and function-
alized products simultaneously. Note this is also similarly
observed in the distillation of I from diol.24 Reactions that give
low conversion also tend to show TMS-R by-products from
quenching, likely due to equilibration between I and RLi as
found with PhLi. Future work will look at developing effective
separation techniques and/or optimizing reaction conditions to
make separation unnecessary.
n-BuLi
-78 oC, Et2O
30 min
Br
Li
O
O
O
Si
O
ð13Þ
+
O
-78 oC, Et2O
Quench: Me3SiCl
Si
O
Si
I
TEOS Systems. For comparative purposes, we also ran
the same studies with Si(OEt)4. These reactions gave slightly
O
Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan | 717

Continued:
Rxn ratio/Conditions
Observed MS
29Si NMR
/ppm
Conversion/%
(29Si NMR)
Structure
(m/z)
1.2:1 9,9-
dimethylfluorenylLi:I
(IX)
334.0
N/A
16
O
Si
Si
O
O
434.1
¹52.6
O
O
O
Si
O
Si
Dimer/Oligomers
N/A
251.0 (M+ ¹ Me)
¹65.8
¹4.8
SiMe3
245.1 (M+ ¹ Me)
¹82.4
O
O
Si
O
O
1.2:1 Me-
naphthalenylLi:I
(X)
402.2
14.7, ¹57.5
23
O
O
Si
HO
O
402.2
¹46.1
O
O
Si
HO
O
N/A
N/A
¹5.8
¹82
SiMe3
O
O
Si
O
O
different product distributions also highly temperature depend-
ent. For example, Ph:I reactions favor mono- or tetrafunction-
alization with little di- or trifunctionalization regardless of
starting material ratio at temperatures below ¹40 °C. Ph:TEOS
reactions result in a mixture of mono- and difunctional
products, with unreacted TEOS also remaining in 1:1 PhLi:
TEOS systems at ¹78 °C. In general, reactions run at or below
¹40 °C favor mono- and difunctionality, reactions at 0 °C
favor di- and trifunctionality, and temperatures above 0 °C
favor tetraphenylsilane. As with Ph:I, Ph:TEOS reactions were
quenched with MeI or TMSCl before warming per condi-
tions listed in Table 4. Thus, a ¹78 °C 1:1 PhLi:Si(OEt)4
reaction quenched with MeI after 4 h gives the EI-MS shown
in Figure 11.
As seen in Figure 11, the PhSi(OEt)3 parent ion appears at
m/z = 240, Ph2Si(OEt)2 appears at m/z = 272.3 and PhSi-
(OEt)2 at m/z = 194 after losing ethoxy and the peak at m/z =
227 corresponds to the diphenyl species missing an EtO group.
Unfortunately, peak intensities in mass spectra do not permit
quantification of the individual species. However, the 29Si NMR
provides a somewhat more accurate estimation of the species
present. Thus, Figure 12 indicates that at ¹78 °C, at a 1:1 ratio
of Si(OEt)4:PhLi, the products are PhSi(OEt)3 (¹58 ppm),
Ph2Si(OEt)2 (¹32 ppm), and Si(OEt)4 (¹82 ppm). The diphenyl
718 | Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan

Figure 9. EI-MS of anthracenylLi:I (3 h, ¹78 °C), quenched with TMSCl; VIII at 410.2 m/z, VIII ¹ Me at 395.2 m/z,
VIII ¹ TMS at 338.1 m/z, 438.2 is VIII + diol without TMS group.
Figure 10. 29Si NMR of 1:1 anthracenylLi:I (3 h, ¹78 °C, VIII) quenched with TMSCl, peak at 53.97 ppm is R-Si(OR)3, peak at
53.40 is R-Si(OR)2(OSiMe3), peak at 10.06 ppm is Si-O-SiMe3, peak at ¹2.87 is TMS-anthracene, I is unobserved at ¹82 ppm.
product seems to form in smaller amounts at this temperature
compared to ¹40 °C, reinforcing the disproportionation mecha-
nism. The ratio of species can be controlled to some degree.
Thus, 0 °C favors formation of Ph3SiOEt. At ¹78 °C/4 h
quenching with MeI (30 min) or TMSCl gives similar results
(Figures S13 and S14).
One might argue based on Corriu’s work that we should not
see any mono- or diphenyl products, yet that is all that is seen.
However, Manoso et al.’s work suggests we might expect to
see the monophenyl products. Note that Manoso et al. used
excess Si(OEt)4 at ¹78 °C as such, we also explored the use of
excess alkoxysilanes discussed further below.
¹78 °C, with the peak at 240.1 m/z corresponding to PhSi-
(OEt)3 and the peak at 208.1 m/z corresponding to unreacted
TEOS. The peak heights suggest that the ratios are μ1:1, which
is expected for a 0.5:1 reaction of PhLi:TEOS. If this reac-
tion is not quenched with MeI or TMSCl before warming up
(Figure S17), the mass spectrum shows Ph2Si(OEt)2 at 272.3
m/z as the main phenyl product, as well as unreacted TEOS at
208 m/z. This shows the importance of quenching the reaction
at low temperature before workup.
We also explored 0.6:1 PhLi:TEOS systems at ¹40 °C, the
results of which are similar to those at ¹78 °C, but at a frac-
tion of the reaction time (30 min vs. 4 h). We see Ph2Si(OEt)2:
PhSi(OEt)3:TEOS distributions similar to those for the 1:1
Figure S16 shows the EI spectrum of 0.5:1 PhLi:TEOS at
Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan | 719

Table 4. Reaction conditions and observed characterization data for PhenylLi reactions with TEOS, all reactions were conducted at ¹78
or ¹40 °C for 30 min-2 h (TE: TEOS)
Observed MS
29Si NMR
/ppm
Conversion/%
(29Si NMR)
Rxn ratio/Conditions
Structures
Quenched
MeI
(m/z)
1:1 Ph:TE
239.9
¹57.9
45
(¹78 °C)
PhSi(OEt)3,
18
O
O
O
Si
Ph2Si(OEt)2
271.9
¹32.2
O
O
Si
N/A
¹81.6
¹57.9
O
Si
O
O
O
0.5:1 Ph:TE
240.1
MeI
34
(¹78 °C)
PhSi(OEt)3,
12
O
O
O
Si
Ph2Si(OEt)2
208.1
272.3
¹81.6
N/A
O
Si
O
O
O
0.5:1 Ph:TE
No
N/A
(¹78 °C)
O
O
Si
208.1
O
Si
O
O
O
Continued on next page:
reaction at ¹78 °C (Figure S18) in the 29Si NMR spectra.
Furthermore, at 0.5:1 PhLi:TEOS at 0 °C we observe Ph3Si-
(OEt) as the dominant species by EI-MS (Figure S19). This is
expected, as higher temperatures favor its formation, and since
for the TEOS reactions once one phenyl reacts, the overall
reactivity increases.
The most important observation is that even at low temper-
atures and sub-stoichiometric PhLi:TEOS ratios, the reaction
products still contain Ph2Si(OEt)2 in considerable quantities, as
found by Manoso et al.10 These results seem to at least partially
support a hexacoordinated silicon mechanism of nucleophilic
attack. However, PhSi(OEt)3 is still present so there likely
exists an equilibrium between the three species as discussed for
the PhI samples (Ph2:Ph:TEOS).
Summary of Findings. In these studies on nucleophilic
substitution of I and TEOS, we find that reaction tempera-
ture, and low-temperature quenching are extremely important
to achieve monosubstitution. However, contrary to previous
studies, we find that I reacting with excess (3 equiv) RLi still
generates monofunctional products. We also find that at
low temperatures, monofunctional-alkoxysilanes form exclu-
sively, and then on warming PhLi must transfer to the other
monofunctional-alkoxysilanes and then to difunctional-
alkoxysilanes, etc. until only Ph4Si and I remain. This sug-
720 | Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan

Continued:
Rxn ratio/Conditions
Observed MS
29Si NMR
/ppm
Conversion/%
(29Si NMR)
Structures
Quenched
(m/z)
0.6:1 Ph:TE
240.1
¹58.7
Me3SiCl
26
(¹40 °C)
PhSi(OEt)3,
18
O
O
O
Si
Ph2Si(OEt)2
N/A
¹32.8
O
O
Si
208.1
240.1
¹82.4
¹58.7
O
Si
O
O
O
1:1 Ph:TE
Me3SiCl
49
(¹78 °C)
PhSi(OEt)3,
15
O
O
O
Si
Ph2Si(OEt)2
N/A
¹32.8
¹82.4
O
O
Si
208.1
O
Si
O
O
O
gests that temperature influences the nature of attack on I; at
<¹40 °C unsubstituted I is most reactive and forms a stable
pentacoordinated intermediate, and at >¹40 °C R-I is most
reactive toward nucleophilic attack suggesting that dispropor-
tionation occurs much more readily.
1 mol% TBAF
CH2Cl2/H2O
O
O
O
Si
O
Si
Silsesquioxane (SQ) Synthesis. To verify the utility of the
materials made by nucleophilic substitution of I, we explored
the synthesis of phenylSQs, anthracenylSQs, and 9,9-dimeth-
ylfluoreneSQs (Reaction 14). The first step was to learn to
purify the products. I and IV can be recovered in a mixture
after vacuum distillation, but are very difficult to separate
further. However, if the 1:1 PhLi:I reaction is run longer
than 3 h/¹78 °C then oligomeric materials result that do not
distill. Treating these oligomeric materials with catalytic TBAF
(see Experimental) provides cage compounds.21,22 Thus GPC
(Figure 13) and MALDI-TOF (Figure 14) analyses verify the
formation of a mixture of PhT8,10,12 and some partial cages,
which could be separated by selective solubility.
T8
+
+
+ Partials
T10
T12
22
ð14Þ
AnthracenylSQs and 9,9-dimethylfluoreneSQs were made
by the same method described for phenylSQs. These are
Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan | 721

-OEt, Ph1
-OEt, Ph2
PhSi(OEt)
3 Ph2Si(OEt)2
Figure 11. EI-MS of 1:1 PhLi:TEOS after quenching with MeI at ¹78 °C.
Figure 12. 29Si NMR of 1:1 Ph:TEOS, ¹32.2 (18% Ph2TE), ¹57.9 (Ph1TE), ¹81.6 (TEOS) after quenching with MeI.
exciting systems since no known anthracenylSQs or 9,9-
dimethylfluoreneSQs with direct attachment have been report-
ed. The isolated cage yields were ca. 58 and 82% respec-
tively. Figures 15 and 16 show the MALDI-TOF spectra of
the anthraceneSQ and 9,9-dimethylfluoreneSQ cage mix-
tures respectively. Due to the large size of the anthracene
and 9,9-dimethylfluorene substituents, the cage formation
process results in many partial cages, with T7 monohydroxide
as the predominant structures in both cases. Note the formation
of the TMS derivative as a by-product from the quenching
process in the anthracene derivative (1525.6 m/z); alternative-
quenching methods such as MeI would alleviate these by-
products.
The synthesis of phenylacetyleneSQs is also explored since
the monofunctional phenylacetyleneI is easily made at RT.
However, it proved difficult to make cages by F -catalyzed
rearrangement due to the sensitivity of the Si-C bond and its
Figure 13. GPC trace comparison of oligomeric materials
from 1:1 PhLi:I (6 h) compared with an overnight TBAF
catalyzed PhSQ synthesis.
¹
722 | Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan

Figure 14. MALDI-TOF spectrum of PhSQs made by TBAF-catalyzed reaction of PhI oligomers; all peaks are Ag+ ions.
Figure 15. MALDI-TOF spectrum of anthraceneSQs made by TBAF-catalyzed reaction of anthracenylI (VIII).
Figure 16. MALDI-TOF spectrum of 9,9-dimethylfluoreneSQs made by TBAF-catalyzed reaction of 9,9-dimethylfluoreneI (IX), all
peaks are Ag+ ions.
¹
propensity to undergo F -catalyzed cleavage (deprotection)
destroying the starting material and resulting in precipitated
silica as a by-product. At 0 °C it was possible to obtain a very
low yield <ca. 2% of the T9-(OH)3 derivative. Other methods
of cage formation including acid catalysis were unsuccessful
in generating the desired SQ products.
AnthracenylSQs and 9,9-dimethylfluoreneSQs are new
compounds with potentially interesting photophysical proper-
Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan | 723

ties even though the cages are incompletely condensed.35 We
explored the absorption and emission behavior (Figures S24
and S25). For anthracenylSQ the absorption and emission show
maxima at 371 and 422 nm, and are both red-shifted from the
parent anthracene by 10 and 25 nm respectively.36 For 9,9-
dimethylfluoreneSQs the absorption and emission maxima are
at 268 and 322 nm, each shifted by ca. 7 nm.37,38 The parent
phenylacetylene shows absorption and emission peaks at 235
and 248 nm. To the contrary, phenylacetyleneSQ (Figure S26)
absorption and emission peaks are present at 284/366 and 427
nm respectively, suggesting further conjugation and/or excimer
formation once attached to the cage. These and other exciting
new cage systems derived from these studies will be explored
in the future.
References
1
2
E. G. Rochow, J. Am. Chem. Soc. 1945, 67, 963.
W. H. Daudt, Production of Phenyltrichlorosilane, U.S.
Patent 2,576,448, Nov. 27, 1951.
T. J. Barton, P. Boudjouk, Organosilicon Chemistry: A
3
Brief Overview, in Silicon-Based Polymer Science, in Advances
in Chemistry, 1989, Vol. 224, Chap. 1, pp. 3-46. doi:10.1021/ba-
1990-0224.ch001.
4
The Chemistry of Organic Silicon Compounds, ed. by Z.
Rappoport Y. Apeloig, John Wiley & Sons Ltd., London, 1998,
Parts 1-3.
5
E. T. McBee, C. W. Roberts, G. F. Judd, T. S. Chao, J. Am.
Chem. Soc. 1955, 77, 1292.
6
P. D. George, L. H. Sommer, F. C. Whitmore, J. Am. Chem.
Soc. 1955, 77, 6647.
7
J. M. Tour, J. A. John, E. B. Stephens, J. Organomet. Chem.
Conclusion
1992, 429, 301.
We have illustrated an important development in siloxane
chemistry, in which we can go from a biosourced silica (RHA)
to functional alkoxysilanes. Our method uses RLi nucleophiles
at low temperatures to functionalize spirosiloxanes made from
RHA, which when quenched with an electrophile such as
MeI or Me3SiCl give mono-R-alkoxysilanes; and unquenched
give tetra-R-silanes with reasonable conversions. We propose
that the mechanism of nucleophilic substitution takes place
through a pentacoordinate intermediate below ¹40 °C, which
if unquenched and warmed to room temperature is capable
of transferring PhLi. We find that aryl halides can be easily
converted to Li reagents, which can then undergo nucleophilic
substitution on spirosiloxanes, while alkylLi reagents and
Grignard reagents do not give clear substitution. Alkynyl deriv-
atives such as phenylacetylene also show clean substitution
with ca. 100% conversion of the spirosiloxane to a monofunc-
tional derivative. We have further demonstrated that silses-
quioxanes can be made from monofunctional spirosiloxanes,
with a never before observed anthraceneSQ.
8
G. Silverman, P. Rakita, Handbook of Grignard Reagents,
Marcel Dekker, 1996, pp. 667-675. doi:10.1201/b16932-33.
9
B. A. Klokov, Org. Proc. Res. Dev. 2001, 5, 234.
10 A. S. Manoso, C. Ahn, A. Soheili, C. J. Handy, R. Correia,
W. M. Seganish, P. DeShong, J. Org. Chem. 2004, 69, 8305.
11 K.-H. Jung, S.-Y. Kim, W. Tan, D.-S. Shin, C. Ahn, Bull.
Inst. Basic Sci. 2005, 17, 79.
12 A. Boudin, G. Cerveau, C. Chuit, R. J. P. Corriu, C. Reye,
Angew. Chem., Int. Ed. Engl. 1986, 25, 474.
13 A. Boudin, G. Cerveau, C. Chuit, R. J. P. Corriu, C. Reye,
Angew. Chem., Int. Ed. Engl. 1986, 25, 473.
14 A. Boudin, G. Cerveau, C. Chuit, R. J. P. Corriu, C. Reye,
Organometallics 1988, 7, 1165.
15 R. J. P. Corriu, J. C. Young, in Chemistry of Organic
Silicon Compounds, ed. by S. Patai, Z. Rappaport, Wiley,
Chichester, 1989, Chap. 20. doi:10.1002/0470025107.ch20.
16 R. Corriu, Pure Appl. Chem. 1988, 60, 99.
17 R. M. Laine, M. F. Roll, Macromolecules 2011, 44,
1073.
18 J. C. Furgal, J. H. Jung, T. Goodson, III, R. M. Laine,
J. Am. Chem. Soc. 2013, 135, 12259.
19 J. H. Jung, R. M. Laine, Macromolecules 2011, 44,
7263.
With this class of materials, we can envision making new
functionalized silsesquioxanes directly from most any aryl-
halide, as well as the possibility for other hybrid polymeric
materials. Control to achieve disubstituted (mono- and difunc-
tional) materials selectively can also be imagined for making
new classes of siloxane polymers, all this while avoiding the
use of chlorosilanes.
20 J. H. Jung, J. C. Furgal, T. Goodson, III, T. Mizumo, M.
Schwartz, K. Chou, J.-F. Vonet, R. M. Laine, Chem. Mater. 2012,
24, 1883.
21 J. C. Furgal, T. Goodson, III, R. M. Laine, Dalton Trans.
2016, 45, 1025.
The authors would like to thank Intel Corporation through
MSR-Task 2170.001 for partial support of this work. We would
also like to thank the University of Michigan Rackham
Graduate School for a Predoctoral Fellowship. We would also
like to thank Mayaterials Inc. for starting materials for I
synthesis. We would like to thank Clemens Krempner for a
fruitful discussion about the failure of Grignard reactions due
to complexation in these studies. Lastly, we would like to thank
James Windak and Paul Lennon in the UM Mass Spec lab for
all their help.
22 J. C. Furgal, J. H. Jung, S. Clark, T. Goodson, III, R. M.
Laine, Macromolecules 2013, 46, 7591.
23 R. M. Laine, K. Y. Blohowiak, T. R. Robinson, M. L.
Hoppe, P. Nardi, J. Kampf, J. Uhm, Nature 1991, 353, 642.
24 R. M. Laine, J. C. Furgal, P. Doan, D. Pan, V. Popova, X.
Zhang, Angew. Chem., Int. Ed. 2016, 55, 1065.
25 N.-D. Chuy, V. Chvalovský, J. Schraml, M. Mägi, E.
Lippmaa, Collect. Czech. Chem. Commun. 1975, 40, 875.
26 F. Carre, C. Chuit, R. J. P. Corriu, A. Fanta, A. Mehdi, C.
Reye, Organometallics 1995, 14, 194.
27 N. Kano, N. Nakagawa, T. Kawashima, Angew. Chem., Int.
Ed. 2001, 40, 3450.
Supporting Information
28 Y. Sugahara, T. Inoue, K. Kuroda, J. Mater. Chem. 1997, 7,
53.
29 T. Jermouni, M. Smaihi, N. Hovnanian, J. Mater. Chem.
1995, 5, 1203.
Additional characterization including, Mass Spectrometry,
NMR, and FTIR for synthesized compounds are included in the
supporting information. This material is available on http://
dx.doi.org/10.1246/bcsj.20160039.
724 | Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan

30 E. Liepiņš, I. Zicmane, E. Lukevics, J. Organomet. Chem.
1986, 306, 167.
31 A. K. Bhattacharya, G. Thyagarajan, Chem. Rev. 1981, 81,
415.
32 a) S.-G. Kim, S. Sulaiman, D. Fargier, R. M. Laine, Sim-
ple Syntheses of Octaphenyloctasilsesquioxane and Polyphenyl-
silsesquioxane. Starting Point for Aromatic Nanocomposites with
Complete Control of Physical and Chemical Properties at Nano-
meter Length Scales, in Materials Syntheses: A Practical Guide,
ed. by U. Schubert, N. Hüsing, R. Laine, Springer-Verlag,
Weinheim, 2008, pp. 179-182. doi:10.1007/978-3-211-75125-
1_24. b) R. M. Laine, unpublished work.
33 A. L. Spek, P. Voorbergen, G. Schat, C. Blomberg, F.
Bickelhaupt, J. Organomet. Chem. 1974, 77, 147.
34 E. C. Ashby, M. B. Smith, J. Am. Chem. Soc. 1964, 86,
4363.
35 Y. Chujo, K. Tanaka, Bull. Chem. Soc. Jpn. 2015, 88, 633.
36 C. A. Parker, W. T. Rees, Analyst 1960, 85, 587.
37 H. Qi, J. Chang, S. H. Abdelwahed, K. Thakur, R. Rathore,
A. J. Bard, J. Am. Chem. Soc. 2012, 134, 16265.
38 R. G. Harvey, P. P. Fu, P. W. Rabideau, J. Org. Chem. 1976,
41, 2706.
Bull. Chem. Soc. Jpn. 2016, 89, 705–725 | doi:10.1246/bcsj.20160039
© 2016 The Chemical Society of Japan | 725