Polycondensation and disproportionation of an oligosiloxanol in the presence of a superbase

Article abstract of DOI:10.1016/S0022-328X(02)01650-9

Kinetics of reactions of model oligosiloxanols, 1,1,3,3,3-pentamethyldisiloxane-1-ol (MDH) and 1,1,3,3,5,5,5-heptamethyltrisiloxane-1-ol (MD₂H), which occur in the presence of phosphazenium superbase, hexapyrrolidine-diphosphazenium hydroxide, in an acid-base inert solvent, toluene, was studied using sampling and gas chromatographic analysis method. In addition, kinetics of reactions of MDH and MD₂H with trimethylsilanol (MH) was studied. In the MDH and MD₂H systems the oligosiloxanol condensation competes with the oligosiloxanol disproportionation, the latter being the dominating process. The disproportionation products, i.e. MD_n+1H and MD_n-1H, n=1, 2, ? undergo analogous consecutive disproportionation and condensation reactions. The kinetic law was derived and rate parameters determined from initial rates and by computer simulation to the best agreement with experimental data. Both competing reactions, the disproportionation and the condensation, conform to the same general kinetic law being first internal order in substrate and first order in catalyst. Activation parameters of these reactions were determined. The results were interpreted in terms of a bimolecular mechanism in which nucleophilic attack of the silanolate anion directed to silicon of the silanol group causes the cleavage of one of its geminal bonds to oxygen, either the one to hydroxyl leading to condensation or the one to siloxane which leads to disproportionation. The latter is faster as the silanolate is a better leaving group compared with OH^-. Moreover, in the pentacoordinate silicon transition state (or intermediate) the siloxane substituent preferentially enters the apical position, thus driving the OH substituent into the unreactive equatorial position.

Full text of DOI:10.1016/S0022-328X(02)01650-9

Journal of Organometallic Chemistry 660 (2002) 14ꢀ26
/
www.elsevier.com/locate/jorganchem
Polycondensation and disproportionation of an oligosiloxanol in the
presence of a superbase
a
Agnieszka Grzelka , Julian Chojnowski *, Marek Cypryk , Witold Fortuniak , Peter
a,
a
a
b
C. Hupfield , Richard G. Taylor
b
a
Centre of Molecular and Macromolecular Studies, Sienkiewicza 112, 90-363 L o´ dz, Poland
b
Dow Corning Ltd., Cardiff Road, South Glamorgan, Wales CF63 2YL, UK
Received 9 April 2002; received in revised form 24 June 2002; accepted 24 June 2002
Abstract
Kinetics of reactions of model oligosiloxanols, 1,1,3,3,3-pentamethyldisiloxane-1-ol (MDH) and 1,1,3,3,5,5,5-heptamethyltrisi-
loxane-1-ol (MD H), which occur in the presence of phosphazenium superbase, hexapyrrolidine-diphosphazenium hydroxide, in an
acidꢀbase inert solvent, toluene, was studied using sampling and gas chromatographic analysis method. In addition, kinetics of
reactions of MDH and MD H with trimethylsilanol (MH) was studied. In the MDH and MD H systems the oligosiloxanol
2
/
2
2
condensation competes with the oligosiloxanol disproportionation, the latter being the dominating process. The disproportionation
products, i.e. MDnꢁ1H and MDnꢂ1H, nꢃ1, 2, . . . undergo analogous consecutive disproportionation and condensation reactions.
/
The kinetic law was derived and rate parameters determined from initial rates and by computer simulation to the best agreement
with experimental data. Both competing reactions, the disproportionation and the condensation, conform to the same general
kinetic law being first internal order in substrate and first order in catalyst. Activation parameters of these reactions were
determined. The results were interpreted in terms of a bimolecular mechanism in which nucleophilic attack of the silanolate anion
directed to silicon of the silanol group causes the cleavage of one of its geminal bonds to oxygen, either the one to hydroxyl leading
to condensation or the one to siloxane which leads to disproportionation. The latter is faster as the silanolate is a better leaving
ꢂ
group compared with OH . Moreover, in the pentacoordinate silicon transition state (or intermediate) the siloxane substituent
preferentially enters the apical position, thus driving the OH substituent into the unreactive equatorial position. # 2002 Elsevier
Science B.V. All rights reserved.
Keywords: Silanols; Polysiloxanes; Silanol polycondensation; Siloxanol disproportionation
1
. Introduction
conversion or monomer conversion was followed. Thus,
reactions of the siloxane bond cleavage could have
escaped the notice of researchers. However, studies
performed in the 80 s, in which some oligomer substrates
were explored as models, revealed that the polyconden-
sation carried out in the presence of strong bases is
greatly dominated by dismutation (disproportionation)
There has been considerable interest in the study of
processes involving silanol condensation as they are
commonly used in the synthesis of siloxane polymers,
for review see [1ꢀ4]. Strong bases such as alkali
hydroxides or silanolates and ammonium hydroxide or
silanolates, are known to be very efficient catalysts for
/
processes [9ꢀ12]. These processes involve siloxane
/
these processes [5ꢀ8]. For a long time it had been
/
cleavage and lead to randomisation of oligomers and
polymers by exchange of terminal siloxane units. The
dismutation of oligosiloxanol was so fast compared with
the condensation that kinetic studies of this dismutation
were performed neglecting the silanol condensation
accepted that the homofunctional condensation of
silanol end groups was the sole reaction in the poly-
condensation system. Monomeric silanol substrates
were exploited in kinetic studies and functional group
[
10,11].
Finding a model system to study the kinetics of silanolꢀ
silanol condensation and siloxane bond cleavage pro-
/
* Corresponding author
E-mail address: jchojnowski@bilbo.cbmm.lodz.pl (J. Chojnowski).
0022-328X/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 0 2 2 - 3 2 8 X ( 0 2 ) 0 1 6 5 0 - 9

A. Grzelka et al. / Journal of Organometallic Chemistry 660 (2002) 14ꢀ
/
26
15
cesses using basic catalysts is difficult. It has been
demonstrated that counter-cations play a considerable
role in the disproportionation, which proceeds particu-
larly fast in the presence of cations, which strongly
interact with the silanolate anion [10], such as Li . Thus,
a system where the interaction was weak seemed to be
2. Results and discussions
2.1. The description of model systems
ꢁ
The model substrates used for the kinetic studies
were 1,1,3,3,3-pentamethyl-1,3-disiloxane-1-ol, (MDH),
more plausible for the study of the silanolꢀ
condensation reaction. Most studies have been per-
formed using alkali metal bases in acidꢀbase reactive
solvents, which dissolve the base well, such as dioxaneꢀ
water [11,12], methanol [13,15], deuterated water [14] and
isopropanol [9]. These solvents are also known to be
/
silanol
and
(MD
1,1,3,3,5,5,5-heptamethyl-1,3,5-trisiloxane-1-ol,
H), while 1,1,1,3,3,3-hexapyrrolidinyl-1,3-dipho-
2
/
sphazenium hydroxide, Scheme 1, was used as the
catalyst. Bulky pyrrolidinyl substituents at phosphorus
should reduce the ability of the phosphazenium cation
to associate with nucleophilic species. The reaction was
performed in a toluene solution and was followed by
sampling and gas chromatographic analysis. The reac-
tion was quenched by addition of ethyldimethylchlor-
osilane in a mixture with triethylamine in THF solution.
Thus, the expected products of the primary reactions
analysed by gas chromatography are formed according
/
active in the Siꢀ/O bond cleavage leading to a very
complex reaction system. This is also the reason for which
they are not used in industrial synthesis of linear
polysiloxanes. Thus, a non-polar acidꢀbase inert solvent
/
would be preferable in this study. However, the solubility
of strong bases in these solvents is limited. Moreover,
intensive ion association, in particular, the formation of
strong aggregates of silanolate intermediates [15,16], and
the association of silanolate with silanol and water [17]
occurs. These association phenomena additionally com-
plicate the kinetics of the polycondensation leading to
fractional apparent order in catalyst [18].
Some years ago very strong bases, peraminooligopho-
sphazenium hydroxides, were synthesised [19,20]. Some
of them are soluble in aromatic hydrocarbon, such as
toluene, and proved to be very efficient catalysts of
siloxane polymerisation and silanol polycondensation
to Scheme 2, nꢃ
MD H, respectively. Commonly used quenching agent
Me SiClꢁEt N would convert MD H to MD M, thus
/
1 and 2 for substrates MDH and
2
/
3
3
n
n
would not permit the differentiation between condensa-
tion and disproportionation products.
Me
silanol condensation reaction in the absence of any
dismutation processes. The mixed systems MHꢁMDH
and MHꢁMD H were also explored.
The condensation reactions lead to generation of
water, which may form a separate phase. However, the
phase separation only occurs in the polysiloxanol
systems at a late stage of the reaction studied when
most of the reactant was consumed. Thus, it was
assumed that the whole reaction proceeds in a one-
phase system. Since the amount of water produced in
SiOH was used as the substrate to study the pure
3
/
/
2
[
21]. This makes it possible to investigate the polycon-
densation in a neutral solvent system close to that used
in industrial manufacturing of polydimethylsiloxanes.
These bases have bulky phosphazenium cations in which
the charge is strongly delocalised. The cations should
interact only very weakly with negatively charged base,
such as the silanolate ion, and the association phenom-
ena with other nucleophilic species in the system should
be of minor importance. Studies of the kinetics of silanol
condensation and siloxane disproportionation in the
presence of these catalysts are of particular interest.
Using a suitable model of oligosiloxanol substrates the
rates of the silanol condensation could be directly
compared with processes involving siloxane cleavage,
the reactions of MDH and MD H is low, it is also
2
assumed that the rates of these reactions are not
significantly affected by water. However, the effect of
water is considerable in the case of the trimethylsilanol
(MH) solution system, where condensation is the
exclusive process. Moreover, MH is a considerably
weaker acid than MD H [22], thus more comparable
n
to water. The reaction slows down as it proceeds and is
strongly suppressed when the solution becomes hazy
indicating phase separation (Fig. 1). Evidently most of
the catalyst passes to the water phase.
such as the disproportionation, in the same baseꢀ
/
siloxanol system. For this purpose it was necessary to
use analytical methods permitting the measurement of
various oligosiloxane and oligosiloxanol concentrations
formed in the reaction system. Gas chromatography is
particularly useful in this respect.
Siloxane hydrolysis, which is reverse to the MD H
n
condensation, is not considered as the condensation
eventually leading to almost full conversion of MD H
species. Primary products of disproportionation and
condensation are almost the only products in the initial
stages of the reaction. Only one secondary product,
MDM, appears early in the MDH reaction system
n
reflecting fast condensation of MDHꢁMH. The bal-
/
ance equations for D units in the substrate and in
primary products, Eq. 1 for the reaction of MDH and
Scheme 1.
Eq. 2 for the reaction of MD H, was kept fairly well up
2

16
A. Grzelka et al. / Journal of Organometallic Chemistry 660 (2002) 14ꢀ
/
26
Scheme 2.
Fig. 1. Substrate conversion-time dependence for the condensation of
trimethylsilanol catalysed by 1,1,1,3,3,3-hexapyrrolidinediphosphaze-
Fig. 2. Substrate conversion-time and product formation-time depen-
dencies for the disproportionation and condensation reactions of
1,1,3,3,3,-pentamethyldisiloxane-1-ol (MDH) catalysed by 1,1,1,3,3,3-
hexapyrrolidinediphosphazenium hydroxide in toluene at 25 8C,
ꢂ3
nium hydroxide in toluene at 25 8C, (^) [MH]
0
ꢃ
/
1.80 mol dm
ꢂ3
,
mol dm , x-marks
dmꢂ3, [cat]
ꢂ3
(
the moment when the water phase appears.
k) [MH]
0
ꢃ
/
1.41 mol×
/
0
ꢃ
/
3.60 10
ꢂ3
ꢂ3
ꢂ3
[
MDH]
0
ꢃ/1.56 mol×/dm
,
[cat]
0
ꢃ/
9.60×
/
10
mol dm
.
(k)
M].
MDH, (I) MD
2
H, (%) MD
2
M, (m) [MDH]ꢁ
/
2[MD H]ꢁ2[MD
2
/
2
to about 30% of the substrate conversion (Figs. 2 and 3).
Therefore, in this range secondary processes played a
minor role. Thus, the comparison of the initial rate of
formation of MD M and MD H, the latter is quantita-
densation of silanols formed according to Scheme 3.
Therefore, Eqs. 1 and 2 are approximate. The general
balance equation is as follows:
2
2
Et
tively transformed to MD M upon quenching, per-
2
X
X
i[MD H]ꢁ
i[MD M]ꢃn[MD H]
i n 0
mitted comparison of the rates of condensation and
disproportionation of MDH (Fig. 2). Similarly, from the
initial rates of formation of MD M and MD H
i
iꢃ1
iꢃ1
(3)
nꢃ1 or 2
4
3
Et
transformed to MD M , relative rates for the conden-
3
It should be pointed out that, although condensation
is slower than disproportionation, both reactions can be
followed in the MDH system, which is in contrast to
much higher disproportionation to condensation rate
ratios observed in the presence of some alkali metal
silanolates [10].
Since the processes in model systems are complex, two
approaches for the analysis of their kinetics were used.
First was a simplified analysis based on initial rates
neglecting any consecutive reactions. This analysis
sation and disproportionation of MD H could be
2
deduced (Fig. 3).
[
MDH]ꢁ2[MD M]ꢁ2[MD H]:[MDH]
0
(1)
(2)
2
2
2
[MD H]ꢁ[MDH]ꢁ3[MD H]ꢁ4[MD M]
2
3
4
:
2[MD H]0
2
The over-all process is more complex because of
consecutive reactions of disproportionation and con-

A. Grzelka et al. / Journal of Organometallic Chemistry 660 (2002) 14ꢀ
/
26
17
Fig. 4. The dependencies of the initial rates of substrate conversion
and product formations on the catalyst concentration for the
Fig. 3. Substrate conversion-time and product formation-time depen-
dencies for the disproportionation and condensation reactions of
disproportionation and condensation of MDH catalysed by
ꢂ
]ꢁ OH in toluene at 25 8C, [MDH]
[
(C
4
H
8
)
3
PꢀNꢀ
/
/
P(C
.
4
H
8
)
3
0
ꢃ/
1
1
2
,1,3,3,3,5,5,5-heptamethyltrisiloxane-1-ol (MD
2
H), catalysed by
,1,1,3,3,3-hexapyrrolidinediphosphazenium hydroxide in toluene at
ꢂ3
1
.52 mol dm
ꢂ3
ꢂ3
ꢂ3
5
8C, [MD
H, (\) MD
H]ꢁ1/2[MDH]ꢁ
2
H]
0
ꢃ
/
1.16 mol×
H, (k) MDH, (j) MD
2[MD M].
/
dm , [cat]
0
ꢃ/3.54×
/
10
mol dm
.
(
I) MD
[MD
2
3
4
M, (m) [MD
2
H]ꢁ3/
/
2
3
/
/
4
permitted determination of the kinetic and activation
parameters for primary reactions. The other approach
was a computer simulation of the kinetic curves to the
best agreement with experimental points. This analysis
permitted the interpretation of kinetic dependencies at
higher substrate conversions and the verification of the
assumptions used in the analysis of the initial rate
kinetics.
2
.2. Initial rate kinetics in MDH and MD H systems
2
Fig. 5. The dependence of the initial rates of substrate conversion and
products formation on initial substrate concentration for the dispro-
portionation and condensation of MDH in toluene at 25 8C catalysed
Results shown in Figs. 2 and 3 indicate that dis-
ꢁ
ꢂ
ꢂ3
ꢂ3
by [(C
4
H
8
)
3
Pꢀ
/
Nꢀ
/
P(C
4
H
8
)
3
]
OH , [cat]ꢃ
/
3.5×
/
10
mol dm . (k)
proportionation and condensation compete with each
other and both can be followed separately as they lead
to different products. Since the relative rates of the
formation of the primary condensation and dispropor-
total, (\) disproportionation, (I) condensation.
tionation products are not changed with the variation of
initial substrate and catalyst concentrations, Figs. 4 and
Scheme 3.

18
A. Grzelka et al. / Journal of Organometallic Chemistry 660 (2002) 14ꢀ26
/
5, it may be assumed that both these processes conform
to the same kinetic law.
any water. Second, the condensation processes occur in
a limited range. Moreover, the ability of water to form
complexes with silanolate may be comparable to that of
The initial rate kinetics were analysed assuming that
the component reactions in the system are first order in
1
the substrate . This assumption was based on previous
siloxanol. Thus, the replacement of silanol with 1/2H O,
2
which occurs as the condensation proceeds, may not
change the effect of the reaction medium. On the other
hand, the condensation of MH is slowed down by water
formed in the reaction, Fig. 1. Thus, MH is less effective
than water in decreasing the thermodynamic activity of
silanolate.
Both component reactions, disproportionation and
condensation, conform to the Arrhenius dependence
(Fig. 6). Activation parameters are shown in Table 1.
Values of the parameters for both reactions are similar.
It seems that the energy barrier for the condensation
process is similar to that for the disproportionation.
Entropies of activation are similar as well.
results for the kinetics of disproportionation of oligosi-
loxanol in the presence of strong bases [10]. The validity
of this assumption was confirmed in the computer
simulation approach.
The dependence of initial specific rates on the catalyst
concentration for substrate conversion and for the
disproportionation and condensation for MDH is pre-
sented in Fig. 4. The apparent order in the total
substrate conversion is first order to a good approxima-
tion and the orders of both component reactions are
fairly close to one. The kinetic law for the initial period
of the reaction is approached by general Eq. 4. This
kinetic law indicates that there is virtually no self-
association of the catalyst and no self-association of
ionic intermediate, which would lead to a fractional
order as the associates usually are unreactive forms of
the catalyst [16,18]. The observed order in catalyst does
not exclude the ion pair formation, which could be the
reactive form of the catalyst.
2
.3. Initial kinetics in mixed systems MHꢁ
/
MDH,
MHꢁMD H and MDHꢁD4
/
/
2
Kinetic runs were followed using a mixture of MH
and MDH which permits the comparison of hetero-
condensation, MHꢁMDH, according to Eq. 6, with
/
ꢂ1=2d[MD H]=dtꢃ(kdisp ꢁkcond)[cat][MD H]
homocondensation and disproportionation of MDH
according to Eq. 7.
n
n
(4)
nꢃ1; 2
cond:
Me SiOHꢁMe SiOSiMe OH 0
The dependence of the initial reaction rate on the
initial substrate concentration is very complex (Fig. 5)
and does not fit to the internal kinetics. The specific rate
increases with the decrease of the initial concentration of
SiOH. The reason for this strange dependence is the
medium effect on the catalytic activity of the base. The
nucleophilic reactivity of the naked silanolate ion in a
weakly solvating solvent, such as toluene, must be very
high. However, the interaction of the silanolate anion
with proton donating species, such as silanol and water
formed in the condensation, strongly decreases its
thermodynamic activity. Interaction of silanol with
silanolate, structure 5, reduces the density of negative
charge on silanolate, and effectively weakens its nucleo-
philicity.
3
3
2
(MH)
(MDH)
Me SiOSiMe OSiMe ꢁH O
(6)
3
2
3
2
(
MDM)
ð5Þ
This phenomenon is not expected to affect the kinetic
curves for the reactions in MDH and MD H systems
2
(
Figs. 2 and 3). First, the disproportionation processes
do not change the medium effect, as they do not change
the total concentration of silanol groups nor it produces
Fig. 6. Arrhenius plot for condensation and disproportionation of
ꢁ
MDH and MD
2
H in toluene catalysed by [(C
4
H
8
)
3
Pꢀ
mol dmꢂ3, [MDH]
dm . (k) substrate conversion, (\) disproportionation of MDH,
/
Nꢀ
0
/
P(C
4 8 3
H ) ]
ꢂ
ꢂ3
OH . For MDH: [cat]ꢃ
/
3.50×/10
ꢃ
/
1.52 mol
ꢂ3
1
ꢂ3
ꢂ3
In some kinetic experiments, first order was observed up to a
relatively high conversion, however, it was somehow accidental.
(I) condensation of MDH. For MD
2
H: [cat]ꢃ
/
5.10×
/
10 mol dm
,
ꢂ3
[MD
2
H]
0
ꢃ/
1.00 mol dm , (m) disproportionation of MD
2
H.

A. Grzelka et al. / Journal of Organometallic Chemistry 660 (2002) 14ꢀ
/
26
19
Table 1
Activation parameters for disproportionation and condensation reaction of MD
n
H nꢃ1, 2 in toluene, catalysed by [(C
4
H
8
)
3
4
PNP(C H
8
)
3
]ꢁOHꢂ
act (kcal molꢂ1
)
DH (Kcal molꢂ1)
%
DS (kcal mol
%
ꢂ1
degꢂ1
)
Substrate
Reaction
E
MDHꢁMDH
Substrate conversion (apparent values)
Condensation
Disproportionation
15.5
14.2
15.3
16.7
14.9
13.6
14.7
16.1
ꢂ17.5
ꢂ26.3
ꢂ18.4
ꢂ10.8
MD
2
HꢁMD
2
H
Disproportionation
Since initial rates of the formation of MDH and
MD H are almost the same (Fig. 8), both oligosilox-
3
ð7Þ
anols are the products of disproportionation according
to Eq. 9. Thus, the dismutation according to Eq. 8 is
negligible; consequently, the condensation 8 and dis-
proportionation 9 occur more or less at the same specific
initial rates (Fig. 8).
Unexpectedly, the heterocondensation MHꢁ
/
MDH
proceeds faster than the disproportionation of MDH
and much faster than the homocondensation of MDH.
See Fig. 7 and Table 2.
A kinetic experiment was also performed to compare
under the same condition the cleavage of the siloxane
bond adjacent to the hydroxyl group in MDH with that
of an unstrained ring oligomer octamethylcyclotetrasi-
loxane, D . The experiment revealed that D reacts
Similarly, kinetic runs in the system MHꢁMD H
/
2
gave us an idea about the relative rates of condensation
and disproportionation, according to Eqs. 8 and Eq. 9.
4
4
much more slowly (Fig. 9). Considering the statistical
factors and initial concentrations the disproportionation
is faster than the opening of the D ring by three orders
4
of magnitude, which points to a preferential cleavage of
siloxane bonds adjacent to hydroxyl groups by a strong
base.
ð8Þ
disp:
2.4. The simulation of kinetic dependencies
2
Me SiO(SiMe O) H 0
(MD H)
2
2
2
2
The gas chromatographic method used for studies of
Me SiOSiMe OHꢁMe SiO(SiMe O H)
(9)
3
2
3
2
3
(MDH)
(MD H)
siloxanol condensationꢀdisproportionation allows the
/
3
simultaneous analysis of the concentration of various
oligohomologues of the MD H and the MD M series.
n
n
This experimental data was used for simulations of
concentrationꢀtime dependencies for various species to
/
the best agreement with experimental data. This ap-
proach provides information for a better understanding
of this complex process. In particular, it leads to
interpretation of the kinetics of the reaction system at
much broader substrate concentration ranges as it
accounts for the formation of the second generation of
products, Scheme 3.
In the simulations the following assumptions were
used. A series of consecutive condensation and dispro-
portionation reactions, Eq. 10, Eq. 11ꢀ12, lead to
/
the formation of various MD H and MD M homo-
logues.
n
n
Condensations :
c
Fig. 7. Simulation of kinetic dependencies (curves) to the best
agreement with the data (points) for the condensation and dispropor-
tionation reactions in the mixture of MDH and MH in toluene at
k
m;n
MD HꢁMD H 0 MD
MꢁH O
(10)
n
m
nꢁm
2
m; nꢃ0; 1; 2; 3 . . .
ꢁ
ꢂ
2
1
5
0
8C, catalysed by [(C
mol dm , (k) MDH, (I) MD
4
H
8
)
3
Pꢀ
/
Nꢀ
/
P(C
4
H
8
)
3
]
OH ,[cat]
0
ꢃ
/
3.60×
/
ꢂ3
ꢂ3
2
H, (%) MD
2
M, (') MDM.

Table 2
Catalytic constant values, k mol
1
ꢂ1
sꢂ1 dmꢂ3, for disproportionation and condensation of permethyloligosiloxanols in toluene, 25 8C, [SiOH]
ꢂ3
0
ꢃ1.0 mol dm
Reaction
From simulation
From initial rates
Disproportionation
Condensation
System
relative rates kcond.
/
Disproportionation
Condensation
Substrate con-
Relative rates kcond./
k
disp.
k
cond.
k
cond. (MDH)
k
disp.
k
cond.
version
k
cond. (MDH)
MHꢁMH
0.41
0.40)
0.028
MHꢁMDH
230
0.46
307
(
(MHꢁMD
MHꢁMDH
MHꢁMD
2
H)
MHꢁMDH
16
7.8
1
0.021
0.015
0.0015
0.022
0.028
0.013
14
10
1
MHꢁMD
MDHꢁMDH 0.013
2
H
0.014
0.0018
2
H
B 0.001
0.012
MDHꢁMDH
(MHꢁMDH)
(0.012)
0.045 (0.032)
(0.0015)
0.0033 (0.0031)
MD
2
HꢁMD
2
H
MD
(
2
HꢁMD
2
H
1.8
0.042
0.003
0.043
2
MHꢁMD
2
H)

A. Grzelka et al. / Journal of Organometallic Chemistry 660 (2002) 14ꢀ
/
26
21
between the neutral, MD H, and the negatively charged
n
ꢂ
m
species MD . Thus, the following reactions between
ꢂ
MDm and MD H were taken into consideration
n
(
Scheme 4):
Fig. 8. Simulation of kinetic dependencies (curves) to the best
agreement with the data (points) for the condensation and dispropor-
2
tionation reactions in the mixture of MD H and MH in toluene at
ꢁ
ꢂ
2
5
0
8C catalysed by [(C
4
H
8
)
3
Pꢀ
/
Nꢀ
/
P(C
4
H
8
)
3
]
OH , [cat]
0
ꢃ/
5.10×
/
ꢂ3
ꢂ3
1
mol dm , (I) MD
M, (') MDM.
2
H, (j) MD
4
M, (\) MD H, (k) MDH,
3
(
%) MD
2
Scheme 4.
Disproportionation reactions for nꢃ
/
mꢁ
/
1
are
omitted, since they do not change the concentrations
of reactants.
If initial concentration of the base was [B], then the
following equations were assumed to hold:
ꢂ
ꢂ
2
ꢂ
3
[
MD ]ꢁ[MD ]ꢁ[MD ]ꢁ. . .ꢃ[B]
(13)
Fig. 9. Competition of condensation and disproportionation of MDH
and ring opening of octamethylcyclotetrasiloxane (D ) catalysed by
OH , in toluene, 25 8C, [D 1.70 mol
4
ꢁ
ꢂ
[
(C
4
ꢂ3
H
8
)
3
Pꢀ
dm , [MDH]
") D , (k) MDH, (%) MD
/
Nꢀ
/
P(C
4
H
8
)
3
]
4
ꢂ3
]
0
ꢃ
mol dm
/
ꢂ3
ꢂ3
0
ꢃ
/
0.71 mol dm , [cat]
0
ꢃ/
5.90ꢄ/10
.
(
4
2
M, (') MDM.
(
11)
12)
(
Fig. 10. Simulation of kinetic dependencies (curves) to the best
agreement with the experimental data (points) for the disproportiona-
tion and condensation of MDH in toluene at 25 8C catalysed by
ꢁ
ꢂ
ꢂ3
[
(C
6.73×
MDM.
4
H
8
)
3
ꢂ3
Pꢀ
/
Nꢀ
/
P(C
4
H
8
ꢂ3
)
3
]
OH , [MDH]
0
ꢃ/
1.52 mol dm , [cat]
0
ꢃ/
All reactions are bimolecular and, according to the
mechanism discussed in the next section, they occur
/10
mol dm , (k) MDH, (I) MD H, (%) MD M, (')
2
2

22
A. Grzelka et al. / Journal of Organometallic Chemistry 660 (2002) 14ꢀ
/
26
[
MD H]
i
ꢂ
[
MD ]ꢃ[B] X
(14)
i
[
MD H]
n
nꢃ1
Taking into account that the acidity of MH is lower
ꢂ
than that of MD H, n E
/
1, [M ] [22] was omitted in Eq.
3 in the simulation of the kinetics in the systems MDH
n
1
and MD H.
2
Eq. 13 corresponds to the assumption that the acidity
of the siloxanols is large enough to transform all the
catalyst introduced to the reaction system into silox-
anolates. The assumption is justified by the much higher
acidity of oligosiloxanols compared with that of water
and alcohols [10]. If this holds, first order with respect to
the total SiOH should be observed. The Eq. 14
corresponds to the assumption that the acidity of all
2
MD H species are the same, independent of n.
n
Other assumptions were those concerned with the
relative reactivities. Thus, it was assumed that rate
c
constants k
n,m
Fig. 11. Simulation of kinetic dependencies (curves) to the best
agreement with the experimental data (points) for the disproportiona-
for all reactions of condensation n ꢀ2,
/
m ꢀ2 are the same. The same assumption was used for
disproportionation process. Thus, it was generally
assumed that the reactivity is independent of chain
/
2
tion and condensation of MD H in toluene at 25 8C catalysed by
ꢁ
OHꢂ, [MD
ꢂ3
[(C
4
H
8
ꢂ3
)
3
Pꢀ
/
Nꢀ
mol dm . (I) MD
H, (j) MD M.
M, (%) MD
/
P(C
4
ꢂ3
H
8
)
3
]
2
H]
0
ꢃ/1.35 mol dm , [cat]
0
ꢃ/
6
MD
.0×
/
10
2 3
H, (k) MDH, (\) MD H, (2)
4
4
2
length except for the cases when n 0
/
2, m 02. Examples
/
of the simulation of both substrates MDH and MD H
2
2.5. General mechanism of the reaction
are shown in Figs. 10 and 11. In Table 2 are given the
rate constants for the simulations. Simulated dependen-
cies fit well to experimental data confirming the validity
of the assumptions in the kinetic analysis of the system.
The rate constants obtained from the simulation
correlate well with the catalytic rate constants obtained
from the analysis of initial rates confirming the validity
of the assumptions of the simulation. This clearly shows
that the complex character of the kinetics of the reaction
The first order of the reaction in substrate together
with the first order in catalyst, the same kinetic law for
the condensation and disproportionation, as well as
relative rates and activation parameters are rationalised
on the basis of a two step mechanism presented in
Scheme 5.
Since toluene is a solvent of low dielectric constant,
the silanolate must appear as an ion-pair. The charge on
the diphosphazenium counter-ion is effectively deloca-
lised and stabilised by six bulky pyrrolidinyl groups at
the phosphorus atoms, which should prevent any
specific interaction of the phosphazenium ion with the
silanolate. Thus, the ion pair is likely to be separated by
silanol molecules interacting with the silanolate anionic
centre (structure 15). The number of silanol molecules
solvating the anionic centre depends on the silanol
concentration.
of MDH and MD H is not only the result of consecutive
2
reactions but is also due to the variation of the substrate
conjugate base concentration with substrate conversion.
The substrate competes for the base with products of
disproportionation. The first order kinetics assumption
is valid only for the initial period of reaction, when the
substrate-conjugate base concentration is equal to the
total catalyst concentration in the system.
ð15Þ
2
This assumption is reasonable for the purpose of the simulation.
Taking into account that the acidity of MH is considerably lower than
H series the reaction of Mꢂ is
The disproportionation is significantly faster than
condensation in the MDH and MD H series; however,
that of other siloxanols of MD
disregarded in the analysis of the MDH and MD
n
H reaction systems
0). However, M is taken into consideration in
2
2
ꢂ
ꢂ
(
the stationary [M ]ꢃ
/
it is not the dominating process. Both competing
reactions exhibit very similar activation parameters
and the same kinetic law. There is evidence that
oligosiloxanol disproportionations and oligosiloxanol
condensations occur via a bimolecular transition state.
the analysis of the system with trimethylsilanol. The lower stationary
ꢂ
concentration of M resulting from the lower acidity of MH is partly
compensated by its higher nucleophilicity thus the net effect on the
ꢂ
contribution of M reactions in MD
significant.
n
H system should not be

A. Grzelka et al. / Journal of Organometallic Chemistry 660 (2002) 14ꢀ
/
26
23
Scheme 5.
Nucleophilic attack of the anion requires its desolvation
by removal of at least part of the silanol molecules
stabilising the anion. The nucleophilic attack is directed
to the terminal silicon bearing the hydroxyl group. It is
surprising that during this reaction the siloxane bond is
more easily cleaved, leading to disproportionation
[24].These effects are the reason for a great tendency to
cleavage of the terminal siloxane bond in polisiloxanols
resulting in dismutation.
The initial rate of the substrate conversion changes as
follows in the series MHꢁ
MD HꢀMDHꢁMDHBMD
This trend is strange and shows that the reactivity is
affected by several factors. First, the reaction MDHꢁ
/
MHꢀ
/
MHꢁ
/
MDHꢀ
H, Table 2.
2
/
MHꢁ
/
/
/
/
Hꢁ
/
MD
2
2
(Scheme 6), than cleavage of the bond to hydroxyl
occurs, which would lead to condensation.
/
It has been shown [10] that the siloxanol dispropor-
tionation is favoured when a counter-ion is capable of
MDH and MD Hꢁ
/
MD H may be considered as
2
2
models of a polymerꢀ
/
polymer reaction which, because
strongly interacting with silanolate, as the rate of the
ꢁ
of steric reasons, usually occur more slowly than the
corresponding polymerꢀmonomer process, the models
of which are MHꢁMDH and MHꢁMD H. The
reaction between two monomer molecules, model
MHꢁMH, is expected to still be faster, which is the
case. On the other hand, susceptibility to nucleophilic
attack increases in the series MHBMDHBMD H, but
this factor is partly compensated by the reverse order of
disproportionation decreases in series of cations Li ꢀ
/
/
ꢁ
ꢁ
ꢁ
Na ꢀ
/
K ꢀ
/
Cs . The phosphazenium counter-ion
/
/
2
used in the present study only loosely interacts with
silanolate, but still the domination of the disproportio-
nation over the silanol condensation is observed.
On the other hand, siloxane acquires the negative
charge more easily than hydroxyl and hence makes it a
better leaving group. It should also be remembered that
the siloxane bond is activated towards cleavage by the
adjacent hydroxyl [23]. As shown here, the siloxane
/
/
/
2
ꢂ
ꢂ
ꢂ
the nucleophilicity of the anion M ꢀ
/
MD ꢀMD .
/
2
Thus, the net electronic effect may be less important
than the steric effect, although it is manifested in a
bond in octamethylcyclotetrasiloxane (D ) is opened by
4
higher rate of MD H reaction as compared with MDH.
2
the same base much more reluctantly than the terminal
bond in the hydroxyl-ended polysiloxane chain. Recent
calculations showed that the silanol group, even par-
Another interesting feature of the kinetics is the fast
increase of the contribution of the disproportionation
reactions in series MHꢁ
/
MDHB
/
MHꢁ
/
MD HB
/
2
tially ionised, decreases the energy of the adjacent Siꢀ
bond by both negative hyperconjugation p(O)0s*(Siꢀ
O) effect and the p(O)0d(Si) orbital interaction effect
/
O
MDHꢁMDHBMD HꢁMD H. This observation
may be easily understood on the ground of the
bimolecular mechanism if structures of pentacoordinate
/
/
/
2
2
/
/
/
Scheme 6.

24
A. Grzelka et al. / Journal of Organometallic Chemistry 660 (2002) 14ꢀ26
/
silicon transition state, Eqs. 16 and 17, or a pentacoor-
dinate silicon intermediate are considered.
Quenching agent ethyldimethylchlorosilane was
synthesised by Grignlard reaction. Ethylmagnesium
chloride prepared from 10 g (0.41 mol) of magnesium
in 100 ml of dried Et O was slowly added to 52.9 g (0.41
2
mol) of dimethyldichlorosilane in 200 ml of dried Et O.
2
Ethyldimethylchlorosilane was isolated and purified by
distillation on column (15 th pl.). Only 15 g fraction
(30% yield) boiling at 88 8C was cut rejecting a large
ð16Þ
primary fraction. The purity of the product checked by
gas chromatography was 97%, but the only contamina-
tion was Et Me Si (about 3%) which did not interfere in
The nucleophilic attack by silanolate anion places the
siloxane substituent into the apical position and OH
group in the equatorial position. This is not only
because silanolate is a better leaving group, more
effectively stabilising the negative charge as compared
2
2
analyses in kinetic studies.
3.2. Preparation of catalyst
ꢂ
with OH , but also because the two bulky siloxane
Hexachlorodiphosphazenium hexachlorophosphate
1), was synthesised by reaction of phosphorus pen-
substituents at pentacoordinate silicon will tend to
occupy positions which are as far from each other as
possible. Whereas, the nucleophilic attack on MH
always directs the OH group to the apical position,
Eq. 17.
(
tachloride with ammonium chloride in POCl according
3
to the procedure similar to that described in Ref. [28]. It
was purified by washing with CH Cl and CCl and
2
2
4
ꢂ3
drying under vacuum 10
r.t.) (yield over 90%).
Compound 1 (6.94 g, 13 mmol) in 50 ml of anhydrous
torr at room temperature
(
toluene were placed under nitrogen in a flask equipped
with a condenser, dropping funnel, stirrer and neutral
gas inlet. The mixture was cooled down to ꢂ50 8C and
/
ð17Þ
13.9 g (195 mmol) of pyrrolidine with 19.7 g (195 mmol)
of triethylamine was added dropwise, during 1 h,
stirring the reaction mixture at ꢂ
stream of nitrogen through the reactor. The system was
allowed to warm to r.t. Then it was heated at ꢁ40 8C
with stirring for 26 h under nitrogen. The precipitate of
Et N×HCl was separated. The main low dark yellow
viscousꢀliquid phase was isolated and after dilution
with CH Cl was repeatedly shaken with a small amount
/
50 8C pushing a slight
/
3
. Experimental
/
3
3
.1. Chemicals
/
2
2
Solvents toluene and acetonitrile were purified ac-
cording to the standard methods described in Ref. [25].
,1,3,3,3-pentamethyldisiloxane-1-ol (MDH) was
of water. Solvents and volatile contaminations were
ꢂ3
removed by heating at 50 8C under vacuum 10 torr
for 5 h. 1,1,1,3,3,3-hexapyrrolidinyl-1,3-diphosphaze-
1
synthesised by hydrolysis of 1,1,1,3,3-pentamethyldisi-
loxane, which had been purchased as reagent grade
product of ABCR. The reaction was catalysed by 5%
Pd/C (Aldrich) purified by distillation at 52 8C/50 mm
Hg. The purity was checked by gas chromatography.
nium chloride (2) (6.2 g, 11.7 mmol) 90% yield, was
3
1
obtained. P-NMR spectrum showed only one signal at
about 2 ppm.
Compound 2 (1.3 g) dissolved in 2 ml of 1:1 V/V
methanolꢀ
1000 ml, dꢃ
Dowex 550A OH, using H Oꢀ
/
water was filtrated through a column (lꢃ
18 mm) packed with strongly anionic resin
MeOH 1/1 as an eluent.
/
1
,1,3,3,5,5,5-heptamethyltrisiloxane-1-ol (MD H) was
2
/
synthesised in an analogous way from 1,1,1,3,3,5,5-
heptamethyltrisiloxane. MH was synthesised by hydro-
lysis of hexamethyldisilazane [26] and purified by
distillation at 25 8C/50 mm Hg. It contained several
percent of hexamethyldisiloxane (MM) which was
determined by gas chromatography and regarded in
the preparation of solutions for kinetic experiments.
Standards for analysis: linear permethyloligosiloxanes
/
2
All operations were performed under nitrogen. Care-
fully decarbonized methanol and water were used.
Methanol and water were removed by heating the
ꢂ3
catalyst at 100 8C under vacuum 10
torr for 5 h.
ꢂ
Since traces of Cl anion were detected in the catalyst, it
was subjected to the repeated exchange procedure using
freshly regenerated Dowex 550 A OH. Then, 1 ml of
3
1
-
ane (MD M), dodecamethylhexasiloxane (MD M) were
octamethyltrisiloxane (MDM), decamethyltetrasilox-
pure dioxane was introduced. P-NMR disclosed the
presence of (Pyr) PNP(O)(Pyr) dꢃ3.66 (d), 9.09 ppm
/
2
3
3
2
prepared and purified using well known procedures [27].
Their purity was checked by gas chromatography.
(d). In order to remove this impurity about half of the
ꢂ3
added dioxane was distilled off under vacuum 10 torr

A. Grzelka et al. / Journal of Organometallic Chemistry 660 (2002) 14ꢀ
/
26
25
for 1 h and 1 ml of toluene was added to the
phosphazenium base. The upper liquid phase containing
toluene and impurities was removed. New portion of
toluene was repeatedly added and removed to wash out
The detector response factor for MD M oligohomo-
n
logue series was determined analysing a solution con-
taining a known amounts of compounds of this series
nꢃ
/
1ꢀ4 and the standard which was undecane. The
/
3
Et
the impurity. Then solvents were removed by heating at
ꢂ3
detector response factor for homologues of MD M
series was determined by the analysis of solutions in
Et
which MD M were quantitatively synthesised in situ
n
5
was dissolved in CH CN to obtain the catalyst stock
0 8C at 10
torr for 2 h. The phosphazenium base
3
n
3
1
solution. The P-NMR spectrum showed only one
signal at dꢃ2.46 ppm. The concentration of the catalyst
was determined by H-NMR analysis comparing the
by silylation of known amounts of MD H with an
n
/
excess of EtMe SiClꢁEt N mixture.
/
2
3
1
1
31
The H-NMR analyses and P-NMR analyses were
performed with Bruker 200 and Bruker 500 spectro-
meters.
ꢁ
ꢃ
2
intensity of the signal of (
/
CH ) CH NCH// with the
2 2 2
signal of CH NO added as internal standard. The other
3
2
method was potentiometric titration with HCl standard
water solution.
The potentiometric titration analyses were performed
using a 721 NET Titrino Methron analyser working
with electrode 6.0233.100.
3
.3. Kinetic studies
3.5. Computer simulation
The reaction was carried out in a glass 10 ml reactor
fitted with a three ways stopcock connected to a
reservoir with pure nitrogen being at a positive pressure.
Kinetics of copolymerisation were simulated by
numerical solving of the approximate differential kinetic
equations (Scheme 3) using the RungeꢀKuttaꢀFehlberg
/
/
The reactor was thermostated (90.1 8C) and fitted with
/
procedure [29]. Apparent rate constants were found with
the non-gradient Rosenbrock optimisation method, in
which the least-squares sum of deviations of the
calculated concentrations of both monomers from those
found experimentally was minimized [30].
a magnetic stirrer. After purging the reactor with
nitrogen the prepared substrate solution in toluene was
placed through the stopcock by means of the precision
Hamilton syringe. The solution was thermostated. The
zero time sample was withdrawn and the solution of
catalyst was introduced by means of a precision
Hamilton syringe. The time of initiator addition was
considered as the zero time of the reaction. Samples
were withdrawn at a suitable time interval by means of
the Hamilton syringe. Streaming of the nitrogen through
the stopcock prevented the reactor content from any
contact with atmosphere during the opening of the
stopcock. Samples were immediately quenched by
introducing them to the Ependorfer vessels containing
Acknowledgements
This research was supported by Dow Corning Ltd.
References
[
1] J. Chojnowski, M. Cypryk, in: R.G. Jones, W. Ando, J.
Chojnowski (Eds.), Silicon-Containing Polymers, Kluwer Aca-
demic Publishers, Dordrecht, 2000, p. 3.
the solution of an excess of 1:1 molꢀmol mixture of
/
ethyldimethylchlorosilane and triethylamine. The mix-
ture quantitatively silylated the silanol groups. The
solution was subjected to the gas chromatographic
analysis.
[
2] R. Drake, I. MacKinnon, R. Taylor, in: Z. Rappoport, Y.
Apeloig (Eds.), The Chemistry of Organic Silicon Compounds
(Chapter 38), Wiley, 1998, p. 2217.
[
[
3] J. Chojnowski, in: S.J. Clarson, J.A. Semlyen (Eds.), Siloxane
Polymers, Ellis Horwood-PTR Prentice Hall, 1993, p. 1.
4] J.J. Le Brun, H. Porte, in: G. Allen, J.C. Bevington, G.C.
Eastmond, A. Ledwith, S. Russo, P. Sigwalt (Eds.), Comprehen-
sive Polymer Science, vol. 5, Pergamon Press, Oxford, 1989, p.
3
.4. Analysis
Gas chromatographic analysis was performed with a
593.
5] T.C. Kendrick, B. Parbhoo, J.W. White, in: S. Patai, Z.
Rappoport (Eds.), The Chemistry of Organic Silicon Compounds
HewlettꢀPackard HP 6890 using TDC detector and
/
[
HP.1 capillary column HP 190592-023, length 30 m,
ꢂ1
diameter 0.53 mm, carrier gas He 5 ml min , detector
temperature 250 8C, injector temperature 250 8C, tem-
(Part 2), Wiley, Chichester, 1989, p. 1289.
[6] W.T. Grubb, J. Am. Chem. Soc. 76 (1954) 3408.
perature program: 60 8C isoth. 3 min, 60ꢀ
/
240 8C, 10 8
MDH and MHꢁMD H
systems the conditions of analysis were: temperature
[7] Z. Lasocki, S. Chrzczonowicz, J. Polym. Sci. 59 (1962) 259.
[8] K.A. Andrianow, P.L. Prikhodko, V.M. Kopylov, A.M. Gasa-
nov, L.M. Khananashvili, J. Gen. Chem. USSR 109 (1979) 49.
ꢂ1
min . In studies of MHꢁ
/
/
2
´
.Scibiorek,
[9] J. Chojnowski, S. Rubinsztajn, W. Sta n´ czyk, M.
ꢂ1
program: 40 8C isoth., 6 min, 40ꢀ
/
240 8C, 10 8 min
.
Makromol. Chem. Rapid Commun. 4 (1983) 703.
10] J. Chojnowski, K. Ka z´ mierski, S. Rubinsztajn, W. Sta n´ czyk,
[
Makromol. Chem. 187 (1986) 2039.
[11] W. Rutz, D. Lange, H. Kelling, Z. Anorg. Allg. Chem. 528 (1985)
98.
3
The method cannot be used if the content of the oxidation
product is too high.

26
A. Grzelka et al. / Journal of Organometallic Chemistry 660 (2002) 14ꢀ26
/
[
12] S. Bilda, G. Rohr, D. Lange, E. Popowski, H. Kelling, Z. Anorg.
Allg. Chem. 564 (1988) 155.
13] T. Suzuki, Polymer 30 (1989) 333.
[21] P.C. Hupfield, R.G. Taylor, J. Inorg. Organomet. Polym. 9 (1999)
17.
[
[
[22] P.D. Lickiss, Adv. Inorg. Chem. 42 (1995) 147.
[23] J. Chojnowski, in: R.G. Jones (Ed.), Silicon-Containing Polymers,
The Royal Society of Chemistry, Cambridge, UK, 1995, p. 59.
[24] M. Cypryk, J. Chojnowski, J. Organomet. Chem. 642 (2002) 163.
[25] D.D. Perrin, W.L.F. Armarego, D.R. Perrin, Purification of
Laboratory Chemicals, Pergamon Press, New York, 1980.
[26] R.O. Sauer, J. Am. Chem. Soc. 66 (1944) 1707.
14] E.R. Pohl, F.D. Osterholtz, in: D.E. Legden (Ed.), Silane,
Surfaces and Interfaces, Gordon and Breach, New York, 1986,
p. 481.
[
15] K. Weiss, Hoffmann, H.F. Grutzmacher, Chem. Ber. 103 (1970)
1190.
[
[
16] J. Chojnowski, M. Mazurek, Makromol. Chem. 176 (1975) 2999.
17] W.M. Kopylov, P.L. Prikhodko, V.A. Kovyazin, I.L. Dubchak,
Vysokomol. Soedin. A26 (1984) 1882.
´
[27] J. Chojnowski, M. .Scibiorek, Makromol. Chem. 177 (1976) 1413.
[28] D. Bougeard, C. Bremard, R. De Jueger, Y. Lemmouchi,
Phosphorus Sulfur Silicon 79 (1993) 147.
[
[
[
18] N.I. Martyakova, S.B. Dolgoplosk, E.G. Kagen, V.P. Mile-
shkevich, Vysokomol. Soedin. 1313 (1971) 579.
19] R. Schwesinger, M. Missfeldt, Angew. Chem. Int. Ed. 26 (1987)
[
29] G.E. Forsythe, M.A. Malcolm, C.B. Moler, Computer Methods
for Mathematical Computations, Prentice-Hall, New York, 1977.
30] H.H. Rosenbrock, Computer J. 3 (1960) 175.
1167.
[
20] R. Schwesinger, et al., Liebigs Ann. (1996) 1055.