LA-UR-15-23317 (Accepted Manuscript)

Aging mechanisms in RTV polysiloxane foams

Labouriau, Andrea
Robison, Tom
Meincke, Linda Jeanne
Wrobleski, Debra
Taylor, Raymond Dean
Gill, John T.

Provided by the author(s) and the Los Alamos National Laboratory (2018-09-20).

To be published in: Polymer Degradation and Stability

DOI to publisher's version: 10.1016/j.polymdegradstab.2015.08.013

Permalink to record: http://permalink.lanl.gov/object/view?what=info:lanl-repo/lareport/LA-UR-15-23317

Disclaimer:
Approved for public release. Los Alamos National Laboratory, an affirmative action/equal opportunity employer, is operated by the Los Alamos
National Security, LLC for the National Nuclear Security Administration of the U.S. Department of Energy under contract DE-AC52-06NA25396.
Los Alamos National Laboratory strongly supports academic freedom and a researcher's right to publish; as an institution, however, the
Laboratory does not endorse the viewpoint of a publication or guarantee its technical correctness.
 Aging Mechanisms in RTV Polysiloxane Foams

Andrea Labouriau1*, Tom Robison2, Linda Meincke1, Debra Wrobleski1, Dean Taylor1 and
John Gill1
1
Los Alamos National Laboratory, Los Alamos, New Mexico, 87545
2
Honeywell FM&T, Kansas City, Missouri, 64147

Abstract.
The long-term use of polysiloxane foams requires a good understanding of the diverse factors
that may influence their performance. The aim of the present work was to understand the
interplay between in-service environments and reactive species that reside in the foams. Two
foams (SX358 and S5370) were thermally aged under compressive strain to gauge the onset of
compression set, an indicator of degradation. Experiments were performed by either aging the
foams in open air containers or in high humidity environments. Our results indicated that foams
aged in open air environments exhibited less compression set than when aged in presence of
moisture. These results were analyzed in terms of post-curing reactions, hydrolysis and
rearrangement of the network, which are promoted by active residues that are left in the foam
from the curing process. In particular, the residues from tin octoate, the catalyst used in the
cure of the foams, were identified and their role in promoting aging was determined. Reactions
were either catalyzed by tin(II) species or by a combination of trace amounts of water and
octanoic acid, a product of the hydrolysis of tin octoate. More precisely, this work
demonstrated that reactive residues promote chemical changes in the polymer, which resulted
in compression set. On the other hand, the oxidized residues of the tin octoate show no
catalytic activity towards promoting condensation reactions between silanol groups. Our work
established that, depending on service conditions, certain residues are more harmful to the
foam’s lifetime than others.

Keywords: Polysiloxanes, hydrolysis, stress relaxation, RTV foams, tin octoate, octanoic acid
Corresponding author: A. Labouriau (andrea@lanl.gov), phone: 505-667-7216.

1
© 2015. This manuscript version is made available under the Elsevier user license
http://www.elsevier.com/open-access/userlicense/1.0/
Introduction.

Polysiloxane elastomers are extensively used in diverse applications due to their superior
properties like good thermal and chemical stability.1 Nevertheless, exposure to harsh
environmental conditions may adversely impact their long-term performance. For example, it is
well known that the stability of polysiloxanes is unfavorably affected by oxygen, water, fillers,
traces of acidic or basic impurities, and residual catalyst.2-7 Case in point, decomposition
products of catalysts are known to affect the stability and mechanical properties of the silicone
rubber.8 On the other hand, the addition of stabilizers, such as zirconium and cerium 2-ethyl-
hexanoates, has been shown to inhibit thermal degradation.9 Therefore, detailed knowledge of
aging mechanisms is important not only to predict the lifetime of the material, but also to
provide guidance on service conditions to be avoided, and lead to improvements that should be
considered when developing replacement materials.

Identifying aging mechanisms that result in changes of the mechanical properties of room
temperature vulcanized (RTV) polysiloxane foams was the primary concern. Silica-filled foamed
polysiloxanes are engineered as stress-absorbing materials to be used in applications that call
for good elastic recovery properties during their lifetime. A set of published studies have shown
that these materials undergo compression set when accelerated aged under compressive
load.10,11 In particular, moisture present in the service environment is likely to react with the
polymer producing lower molecular weight siloxanes.12 By the same token, residual catalyst is
expected to promote post-curing reactions that may increase cross-link density. Generally,
changes in the polymer network linkages while the polymer is under strain are associated with
permanent set or loss of elastic recovery as has been indicated by several experimental and
13,14
molecular dynamics studies. Data reported in this work provides valuable insights on the
presence of reactive sites and the influence of residual catalyst and its by-products in
promoting post-curing reactions, chain scission and bond-exchange reactions. All of these
changes are likely to degrade the RTV foam over its lifetime.

2
 2. Experimental Section.

2.1. Materials.

Two chemically similar polysiloxane foams, S5370 and its replacement foam known as
SX358 were artificially aged under compressive strain.

SX358 foams were newly synthesized from a polysiloxane elastomeric gumstock that
contained 75wt% dihydroxyl-terminated polydimethylsiloxane (HO-PDMS-OH), 3wt%
polymethylhydrosiloxane (PMHS), 2wt% tetrapropyl orthosilicate (TPS), and 5wt%
diphenylmethylsilanol (DPMS); all of which were purchased from NuSil Technologies. The
dihydroxyl-terminated polydimethylsiloxane consisted of a mixture of about 17wt% low
molecular weight (Mw=2,350), 14.5 wt% medium molecular weight (Mw=29,400), and 43.5 wt%
high molecular weight (Mw=54,300). The uncured resin also contained 15 wt% of diatomaceous
earth filler to provide reinforcement to the cured polymer. SX358 foams were cured by
condensation reactions between silanols on dihydroxyl-terminated PDMS, and DPMS and
silanes on PMHS, producing hydrogen gas. Condensation between silanols and TPS also
contributed to cross-linking, with evolution of n-propyl alcohol. Both condensation reactions
were catalyzed by tin octoate (also known as tin 2-ethylhexanoate). The procedure consisted of
mixing 5 wt% tin octoate into the polysiloxane gumstock, and allowing it to cure at room
temperature into a mold to obtain a chosen thickness (the average thickness of pristine
samples was 0.102 cm) and density (about 0.4 g/cm3). This amount of tin catalyst is necessary in
order to produce foams of high-quality. The foams were post-cured by heating at 115°C in air
for 3.5 hours. SX358 samples were artificially aged in air at temperatures of 30, 50, 70, 90 and
110°C in convection ovens. There were three specimens per aging condition that were
periodically withdrawn to measure changes in thickness. Then, the samples were placed back
into the compression fixtures and ovens for continued aging. Stackable compression fixtures
were made from aluminum sheet stock with sample wells machined for the appropriate depth
to accommodate 5.0 cm diameter sample discs at 25% compressive strain. The fixtures used in
this work also had two small indents that allowed for air to diffuse through the samples.

3
 S5370 foams were originally prepared from a commercial kit (Dow Corning Silastic5370)
that is no longer available.10 These materials were synthesized in 1985 and stored under low
humidity conditions at 20 to 22°C. Thus, S5370 foams were already naturally aged for 28 years
before being artificially aged and tested in this work. The reasoning to use naturally aged S5370
foams was to investigate the role of long-lived residues in the long-term aging of RTV foams.
S5370 samples had an average thickness of 0.108 cm and density of 0.4g/cm 3. Samples were
cut into approximately 5 cm diameter discs and thermally aged at 70°C, under fixed
compressive loading (about 30%). Specimens were aged for up to 2 years at controlled
humidity levels of 10 and 25% using commercial humidity chambers (ESPEC North America Inc.,
Michigan, USA). A set of two samples per aging condition were analyzed at 3-months intervals
(newly aged samples were never returned to the humidity chambers).

Bond exchange reactions, or rearrangement of the network, were evaluated by probing

variations in the molecular weight distribution of tri-methyl-terminated PDMS (Mn=10,000,
American Polymer Laboratories) when mixed with 5 wt% tin octoate (Aldrich-Sigma). The tin
119
catalyst sample was initially analyzed by Sn Mössbauer spectroscopy, confirming that it
contained mostly tin(II) species (85% tin(II) and 15% tin (IV)). For a period of up to 14 weeks,
aliquots were periodically taken and investigated by gel permeation chromatography (GPC).
Since no significant changes in the PDMS molecular weight or in its polydispersity index
(Mw/Mn) were observed for the first 8 weeks, the sample was heated to 70°C for an additional 6
weeks; aliquots were analyzed by GPC at 3 and 6 weeks.

Acid-catalyzed hydrolysis reactions were investigated by monitoring changes in the

molecular weight of tri-methyl-terminated PDMS (Mw=97,200, Mn=51,900, Sigma-Aldrich, 200
mg of polymer for each aging condition) when mixed with 5 wt% of octanoic acid (also known
as 2-ethyl-hexanoic acid, Sigma-Aldrich) and 10 μl of water. These samples were aged at 40, 70,
90, and 110°C for up to 103 days; aliquots were periodically investigated by GPC.

The ability of the tin catalyst for promoting condensation reactions or chain extension
reactions was examined by mixing low molecular weight dihydroxyl-terminated PDMS (NuSil
Technologies) with 5 wt% of tin octoate (these were new and degraded batches of the tin

4
 29
catalyst). Samples were dissolved in chloroform and readily analyzed by liquid-state Si NMR
17
spectroscopy (or within one hour after sample preparation). O insertion in the polymer
backbone was studied by artificially aging post-cured and non-post-cured SX358 foams in
17
excess of isotopically enriched O water (17O, 20% Cambridge Isotope Laboratories, Inc.) in
closed containers at 60°C for 23 weeks. These samples were analyzed by solid-state 17O NMR
spectroscopy.

Residues from the tin catalyst present in RTV foams were evaluated by extracting the
samples (380 mg) in deuterated chloroform (CDCl3, 99.6% Acros Organics) for 4 hours at room
temperature. The soluble fraction was about 3 to 5% of the initial mass of the sample. The
extracted material (0.75ml) was analyzed by liquid-state 1H NMR spectroscopy.

2.2 Techniques.

Liquid-state 1H and 29Si NMR experiments were carried out using a Bruker Avance NMR
spectrometer operating at 300.13 MHz for 1H and at 59.62 MHz for 29
Si, proton and silicon
17
signals were referenced to internal tetramethylsilane (TMS). O NMR spectroscopy was
17
performed on RTV foams that were artificially aged with excess of O isotopically enriched
17
water. O NMR experiments were performed using a Bruker solid-state NMR spectrometer
operating at 54.24MHz. Samples were spun to 5kHz using a 4 mm Magic Angle Spinning (MAS)
probe. 1H MAS NMR experiments were performed on pristine SX358 foams using a Bruker
Avance NMR spectrometer operating at 400.13 MHz equipped with a Bruker (4 mm) Magic
Angle Spinning (MAS) probe. Samples were spun at about 9 KHz. 1H MAS NMR was used to
characterize the foam’s network, especially to detect unreacted chemical groups such as
silanols and silanes.

119
Sn Mössbauer spectroscopy was performed at low temperature (75K) using an
integral cryostat and spectrometer. A commercial Ca119mSnO3 source provided the 23.87 keV
gamma rays for the Mössbauer spectroscopy. Spectra were fitted using the appropriate

5
Hamiltonian and least square-fitting program. The statistical errors associated with the fitting
procedure are of the order of ±0.02 mm/s in velocity or 0.5% in transmission. The inherent
recoil-free fraction for Sn(IV) is slightly greater than that for Sn(II), and a correction has been
applied to reflect this difference as it applies to determining the Sn(IV)/Sn(II) relative
abundances.15
Mass spectroscopy was performed to detect volatile species and permanent gases in the
headspace of thermally aged SX358 foam. The foam was aged at 70°C under N2 for 1 year in a
sealed aluminum canister. A sample of headspace gas was collected into a small gas bottle
using a vacuum manifold. An aliquot of a sample gas was admitted to the inlet chamber of a
Finnigan 271 mass spectrometer. From the inlet chamber, the gas was allowed to leak slowly
into the mass spectrometer’s electron-beam ionizer region during an analysis. Ions were
electrostatically accelerated and mass filtered by an electromagnetic sector to a resolution of
m/Δm ~ 1400. The partial pressures of inlet gases that produced the observed ion data were
deduced from detector sensitivities. The sum of the molecular gas partial pressures was
ratioed to the total measured inlet pressure, and the results were within a few percent of
100%. The NIST mass spectral database was consulted for assignment of hydrocarbons in the
headspace analysis.
GPC provided molecular weights and molecular weight distributions of PDMS standards
(Aldrich-Sigma) when mixed either with tin octoate or octanoic acid. The GPC instrument
consisted of an Alliance 2690 pump equipped with a Wyatt Rex Differential Refractive Index
Detector and utilized three Polymer Labs PL Mixed B GPC columns. The solvent used was
toluene at 70°C and at a flow rate of 1.0 ml/min. The molecular weights were determined using
narrow molecular weight PDMS calibration standards.
The foams investigated in this work were aged under compressive load and compression
set was used to evaluate aging. Compression set represents the reduction in thickness after a
material is aged in compression, and in this work is a fraction of the applied compression. In the
experiments described above, S5370 foam samples of initial thickness Hi were aged under
compressive strain (30%) for long periods of time. Every 3 months, two specimens per humidity
level were removed from the compression fixtures for measurement of their thickness and

6
other characterizations. The percent compression set at age t is given by S=100*(Hi-Ht)/(Hi-hind),
where S is the compression set, Hi is the initial thickness, Ht is the recovered material thickness
at age t, hind is the height of the indent of the compression fixture. The depth on the
compression fixture determines the strain that is applied to the RTV test sample. The
thicknesses of the samples were measured 2 hours after compression fixtures were removed
from the humidity chambers and the samples were removed from the compression fixtures. In
these experiments, a height gauge (Browne and Sharpe Micro-Hite 350) was used to accurately
measure the sample thickness.

3. Results and Discussion.

3.1. Artificially Aged SX358 foams.

SX358 foams were artificially aged under 25% compressive strain at temperatures
varying from 30 to 110°C in an open air environment for up to six years. Figure 1 shows that
compression set values increased with temperature and aging time. It is important to point out
that this behavior is comparable to other data previously reported in the literature for similar
foams. For example, Coons and co-workers analyzed the mechanical response of S5370 foams
that were aged for 9 years in air with 50% compressive strain at temperatures varying from 23
to 70°C.10 They assumed that compression set followed a first order kinetic model with
Arrhenius temperature dependence. The activation energy was estimated to be approximately
29 kJ/mol. Patel and co-workers investigated compression set in comparable polysiloxane
foams that were aged in dry nitrogen atmospheres at temperatures ranging from 25 to 70°C
with 25% compressive strain.11 The authors suggested that compression set was not necessarily
associated with significant changes in cross-link density, but more likely due to a reorganization
of the network. In this work, the role of residues from catalytic active species was investigated
in order to explain the compression set behavior observed by our team and others. Specifically,
tin(II) residues, which are a product of the foam curing, are of particular interest and are
discussed in detail in the next section.

7
 35 14 30C
30 30C
12 50C
% Compression Set

50C

% Compression Set
25 10 70C
70C
20 8 90C
90C
15 6 110C
110C
10 4
5 2
0 0
0 1000 2000 3000 0 50 100 150 200 250 300
days of aging days of aging

Figure 1. Compression set of artificially aged SX358 foam as a function of aging time: left plot
shows data points for up to 6 years, whereas the plot on the right shows data points up to 7
months (the linear fit is just to guide the eye).

3.2. Oxidation of Tin(II) Octoate Residues.

RTV polymers are always manufactured by combining a hydroxyl-terminated silicone

resin, a cross-linker and a catalyst, with organotin carboxylates as the catalysts of choice since
these are less likely to promote side-reactions.16 Case in point, the network of SX358 and
S5370 foams are made by reactions between dihydroxyl-terminated PDMS chains and cross-
linkers (TPS and PMHS). The foams are cured using tin octoate, an effective silanol
condensation catalyst routinely used in the silicon industry for curing organosilicon polymers.
The curing of SX358 and S5370 foams requires the use of relatively large amounts of tin octoate
(5 wt%), leaving significant amounts of residues in the foam. Interestingly, even though
organotin carboxylate catalysts have been used for decades and in a number of applications,
the polymerization mechanism is still a subject of consideration.17
The work by Severnyi and collaborators provided the first interesting insight on the nature of
the active tin center by showing that the silicone curing does not proceed under anhydrous
conditions.16 This observation led to the work by van der Weij, which suggested that the
hydrolysis products are the actual catalysts.18 We and others have demonstrated that the tin

8
catalyst is prone to oxidation and/or hydrolysis, and these chemical reactions have a direct
impact on the catalytic activity and on the final properties of the rubber.15,19 In its pristine form,
the neat catalyst consists primarily of tin octoate with small amounts of oxidized tin species and
octanoic acid.15 Distillation of tin octoate does not remove all acid impurities, and since trace
amounts of water are always present, these species are readily formed by hydrolysis of the
catalyst.20 Interestingly, the chemical makeup of tin residues in pristine RTV foams is very
different from the composition of the neat tin octoate. Tin residues found in pristine RTV foams
119
are tin(II) and tin(IV) species, often in approximately equal concentrations as shown by Sn
Mössbauer spectroscopy (see Figure 2). Consequently, tin(II) residues remain in the pristine
foams, and are likely to promote post-curing reactions amongst residual active groups, such as
silanols and silanes. These reactive species are definitely observed in the 1H MAS NMR
spectrum of pristine SX358 foams as shown in Figure 3, even though solid-state NMR is not a
particularly sensitive technique. The NMR spectrum was truncated to better display the small
peaks associated with unreacted SiH and/or SiOH (4.7 ppm) and partially unreacted TPS (-OCH2
groups at 3.6 ppm). The larger peaks in the spectrum are due to methyl protons from PDMS
(0.1 ppm), phenyl groups from DPMS (from 7 to 8 ppm) and residues from the tin catalyst
(small peaks from 0.8 to 1.6 ppm). Post-curing reactions in the presence of these reactive
species will likely modify the foam’s original network structure by creating new siloxane
linkages. The net result will be an increase in the cross-link density. Thus, post-curing reactions
are expected to induce changes in the mechanical response of the artificially aged foam.

9
Figure 2. 119Sn Mössbauer spectrum obtained for pristine SX358 foam showing the presence of
tin (IV) species (an unresolved doublet centered at 0 mm/s), and tin (II) species (a doublet
centered at a Doppler velocity of 3 mm/s). Curve fittings for tin (II) and tin(IV) sites are also
shown by dotted lines. Tin relative concentrations are 45% tin(II) and 55% tin(IV).

Figure 3. 1H MAS NMR spectrum obtained for pristine SX358 foam showing small peaks in the
baseline, which are associated with reactive species such as silanols and silane groups.

10
 Compelling evidence showing that post-curing reactions actually take place during aging
of the SX358 foam is provided by analysis of the evolved gases. A SX358 sample that was aged
in a sealed container under N2 afforded the means to evaluate permanent gases and volatiles
generated due to thermal aging. The SX358 sample was artificially aged at 70°C for a period of 1
year. As listed in Table 1, characterization of the headspace by mass spectroscopy clearly shows
H2 evolution. This gas is a by-product of condensation reactions between silanol and silane
groups. It is worth pointing out that such chemical reactions are promoted by tin octoate, and
will lead to new Si-O-Si linkages. Clearly, this result indicates that tin(II) residues are
catalytically active and able to promote post-curing reactions of reactive species, which in turn
may lead to compression set. Cyclic siloxanes are also detected in the headspace, albeit in very
small concentrations. These species could be formed through back-biting reactions promoted
by residues of the tin catalyst.

Table 1. Major constituents detected in the headspace of a canister containing 12 g of SX358

foam aged under inert conditions at 70°C for 1 year. Concentrations are in ppm. D3 and D4
refer to hexamethylcyclotrisiloxane and octamethylcyclotetrasiloxane, respectively. A control
sample shows only 22 ppm H2.

H2 CH4 Ar CO2 C3H8O C6H6 D3 D4

3,000 10.8 182 144 1.4 8.6 26.7 0.7

Since tin octoate is prone to oxidation reactions, the question arises as to when tin(II)
residues become completely oxidized. Changes in the oxidation state of tin(II) residues in
artificially aged SX358 foams were investigated by 119Sn Mössbauer spectroscopy. It was found
that higher aging temperatures and longer aging times resulted in an increase of the relative
concentration of tin(IV) species, and likewise a decrease in the concentration of tin (II) species.
Figure 4 illustrates the change in the tin(IV) relative concentration as a function of time. This

11
master curve was obtained by using the time-temperature superposition (TTS) approach.21 The
activation energy for the tin(II) oxidation process was found to be approximately 50 kJ/mol.

1
4
0.9

0.8 3
30C
Sn(IV)

log aT
50C
0.7 2
70C
%

0.6 90C
1
110C
0.5
0 5000 10000
0
days 3 3.2 3.4 3.6 3.8 4
104/kT

Figure 4. Increase in the relative concentration of tin(IV) species due to oxidation reactions as a
function of time and temperature (left) and TTS shift factors (right).

Based on these results, it was of interest to investigate procedures that could rapidly
oxidize tin(II) residues. One of the most effective methods found by our team consisted of
treating the pristine foam with 10% hydrogen peroxide solution for up to 30 minutes, followed
by washing in water and drying in air. The final treatment consisted of post-curing the
119
“peroxide washed” foam at 115°C in air for about 3 hours. Sn Mössbauer data obtained for
samples treated by this method show no tin(II) species after the post-curing treatment (see
29
Figure 5). Furthermore, Si MAS NMR spectrum of the treated foam showed no detectable
damage to the polymer. Another method for oxidizing tin(II) residues consisted of exposing the
foam to gamma irradiation in air. However, this approach has some limitations since the RTV
foam is susceptible to radiolysis. 22

12
 1.000 1.00 1.000
0.995
Transmission

Transmission
0.995 0.98

Transmission
0.990
0.990 0.96
0.985
0.985 0.94
(a) (b) 0.980 (c)
0.980 0.92 0.975
-8 -6 -4 -2 0 2 4 6 8 -8 -6 -4 -2 0 2 4 6 8 -8 -6 -4 -2 0 2 4 6 8
velocity (mm/s) velocity (mm/s) velocity (mm/s)

Figure 5. Evolution of the oxidation state of tin residues as the RTV foam receives different
treatments: (a) pristine foam (non-post-cured) has 60% tin(II) species, (b) pristine foam treated
with hydrogen peroxide has 12% tin(II) species, and (c) foam treated with hydrogen peroxide
and then post-cured has no tin(II) species, only tin(IV) species. Curve fittings of tin(IV) species
(green trace), and tin(II) species (red trace) are represented as well.

3.3. Artificially aged S5370 foams.

This work demonstrated that specific residues associated with the tin catalyst promote
post-curing reactions that, in the long term, could be detrimental to the life of the foam and
lead to mechanical failure. Once tin(II) residues are oxidized, other aging mechanisms may still
take place and influence the mechanical performance of the RTV foam. Thus, S5370 foams that
were already naturally aged for 28 years under standard shelf storage and never used in service
119
were chosen for the second part of this aging study. Sn Mössbauer experiments confirmed
that S5370 foams had very small amounts of tin(II) residues, namely about 12% on average. As
shown in Figure 6, by the end of the aging study the foams had only 4 to 6% tin(II) species (the
remaining being tin(IV) species).

13
Figure 6. 119Sn Mössbauer spectrum obtained for 28 year-old S5370 foam that was aged for 21
months at high humidity showing that most tin(II) residues are oxidized. There are only 6%
Sn(II) species. Curve fittings of tin(IV) species (green trace), and tin(II) species (red trace) species
are represented as well. No significant differences in the tin (II) oxidation rate were observed
with humidity levels.

In addition to tin(II) residues, it was of interest to investigate other non-network species

that reside in the 28 year-old foam. The soluble fraction not only has a plasticizer effect, but it
may also contribute to the aging of the foam. Analysis of the extractable material obtained
from S5370 foams shows the presence of octanoic acid as detected by liquid-state 1H NMR. This
acid is produced when tin octoate is exposed to trace amounts of moisture. Figure 7 shows that
extracts consist mostly of a large peak corresponding to methyl groups from PDMS chains, and
smaller peaks that are associated with organic residues from the tin catalyst and octanoic acid
(1.085 to 0.75 ppm are methyl groups, 1.2 to 1.6 ppm are due to methylene protons, and peaks
from 2.2 to 2.3 ppm are methyne protons). Residues from unreacted DPMS and possibly by-

14
products like dimers are also present in the extracts; aromatic protons show up as a set of
peaks at lower field (7.2 to 7.6 ppm), whereas the methyl protons appear at 0.5 to 0.7 ppm.

Figure 7. 1H NMR spectrum of extractable material obtained from S5370 foams (bottom). The
sharp peak at 7.26 ppm is due to residual protons from CDCl3, the solvent used to extract the
foam, and the peak at 1.56 ppm is due to residual water. The top 1H NMR spectrum is of neat
octanoic acid dissolved in CDCl3.

S5370 foams were artificially aged under compression at 70°C for up to two years. These
specimens were aged in high humidity conditions in order to probe the effects of environments
other than air. The reasoning for choosing this set of aging conditions was to probe the
combined effect of octanoic acid residues and moisture on the degradation of the foam. Thus,
S5370 foams afforded the opportunity to gauge the detrimental effects of other reactive
species other than tin(II) residues, since these were mostly oxidized in the 28 year-old foam. It
has been previously suggested that residual octanoic acid contributes to the long term
compression set observed in S5370 foams, and that improved post-curing procedures (longer

15
post-curing times under light vacuum) may reduce their concentration.23 Figure 8 shows the
averaged compression set observed for S5370 foams that were thermally aged under
compressive load (30%) in presence of moisture. It is clear that compression set increases
linearly with aging, and it reaches very high values for the foams aged for the longest amount of
time. Compression set is also more pronounced for the samples aged under higher humidity
levels than for those aged at lower humidity levels. These results indicate that, even in the
absence of tin(II) residues, S5370 foams continue to age when exposed to high humidity
environments. In order to elucidate such compression set behavior, the catalytic activity of
species originated from the degradation of tin octoate residues was evaluated.

Figure 8. Compression set of S5370 samples aged at 70°C at 10 (blue trace) and 25% RH (red
trace) for up to 2 years.

3.4. Catalytic Activity of Tin Residues.

As previously mentioned, tin octoate slowly oxidizes producing tin(IV) species and
undergoes hydrolysis producing octanoic acid. 15,24,25 The presence of the acid in the RTV foam
may lead to the polymer degradation by protonation of the siloxane chain, which eases the

16
barrier for hydrolysis.26 In order to demonstrate this effect, trimethyl-terminated
polydimethylsiloxane samples were thermally aged in presence of octanoic acid and trace
amounts of water. These samples were mixed with 5 wt% octanoic acid and aged at 40, 70, 90
and 110°C for various periods of time (up to 103 days for the 40°C sample, less for the other
samples). Changes in the molecular weight of the artificially aged PDMS samples were
periodically assessed by GPC measurements. A significant drop in the molecular weight was
observed as shown in Figure 9. The activation energy for the hydrolysis reaction, promoted by
trace amounts of moisture and carboxylic acid, was estimated by the time-temperature
superposition approach (TTS). It was found to be approximately 30 kJ/mol, which is consistent
with results obtained by other authors for polydimethylsiloxane networks.7,8,27,28

140 2.5
110C
140
90C 2
120
120 70C 1.5

log aT
103 * Mw

40C
100 100 1
w
10 X M

pristine
0.5
80
3

80 0
3 3.2 3.4 3.6 3.8 4
60 4
10 /kT

40 60
0 20 40 60 80 100
days aging 40
0 5 10 15 20
log (days)

Figure 9. Decrease in the average molecular weight of linear PDMS when thermally aged with
octanoic acid (left), and TTS master curve (right) with shift factors (insert).

Residues from the tin catalyst are also thought to promote bond-exchange reactions,
which are rearrangements of the network (basically an active unit attaches to an existing unit
by replacing a pre-existing bond with a new one). More specifically, trace amounts of water and

17
acidic species (such as octanoic acid) can attack the polar O-Si-O backbone and induce chain
scission with formation of silanols. The tin catalyst then promotes the condensation of the
newly formed silanol groups, resulting in a network with a broader molecular weight
distribution than the original one. This bond-exchange effect was investigated by monitoring
the molecular weight distribution of tri-methyl-terminated PDMS samples that were mixed with
the tin catalyst. Significant changes in the molecular weight distribution were observed for the
aged samples when compared with the control sample, as seen in Figure 10. The polydispersity
of the 10,000 Dalton sample changed from 1.2 to 1.6 and 1.9 after being thermally aged for 3
and 6 weeks at 70°C; respectively.

300

400
DRI (toluene)

500

600

700

800

900
3 3.5 4 4.5 5
Time (min)

Figure 10. GPC chromatograms obtained for the 10,000 Dalton tri-methyl terminated PDMS
samples after being aged with 5 wt% tin octoate for 3 weeks at 70°C (blue trace), and the same
sample after aged for 6 weeks at 70°C (red trace). The green trace is the control sample. (DRI
stands for the differential refractometer index).

18
 Determining if bond-exchange reactions could take place in RTV foams was of interest
as well. Thus, the experiment performed here consisted of artificially aging SX358 foams at 60°C
for up to 23 weeks with excess of isotopically enriched H217O. The 17O NMR signal was used as a
17
signature to probe the insertion of O in the polymer chain. Two sets of foams were aged by
this procedure: foams that had been post-cured and those non-post-cured. NMR results
obtained for these artificially aged samples are shown in Figure 11. The large peak at 0.0 ppm is
associated with 17O signal from free water, whereas the peak at 71.2 ppm is due to –Si-17O-Si-
groups, confirming the proposed chemical reaction described in Figure 11. 29,30 It is interesting
to note that the non-post-cured sample also shows a peak at 43.9 ppm, which we assume to be
associated with –Si-17OH groups. The presence of silanol groups in the non-post-cured foam is
not unexpected, since this foam is likely to contain relatively large amounts of unreacted
silanols from the dihydroxyl-terminated PDMS chains. These reactive groups may have
exchanged with H217O, producing the small peak observed in the NMR spectrum.

H217O

-Si-17O-Si-

19
 H3C H3C CH3 CH3 CH3 CH3
CH3 H

Si O Si O Si + O Si O Si OH + Si O Si OH

H3C H3C CH3 H CH3 CH3 CH3 CH3

Sn(II)
H3C H3C CH3

Si O Si O Si
+ H2O
H3C H3C CH3

Figure 11. 17O MAS NMR spectra for post-cured (bottom) and non-post-cured (top) SX358
foams that were thermally aged with 17O isotopically enriched water. The scheme for the 17O
insertion in the polymer backbone is shown as well (O represents 17O). The condensation of
silanol groups is promoted by the tin catalyst.

It was previously established that even in the absence of a cross-linker, tin octoate is
able to catalyze chain extension reactions of dihydroxyl-terminated PDMS.31-33 The catalytic
activity of oxidized tin residues was then assessed by combining low molecular weight
dihydroxyl-terminated PDMS with two different batches of tin octoate catalyst: one obtained
from an unused bottle and a second one from an opened bottle that had degraded naturally.
119
Both samples were initially characterized by Sn Mössbauer spectroscopy to determine the
relative concentration of tin(II) and tin(IV) species. The pristine sample contained about 16%
tin(IV) and 84% tin(II), whereas the degraded material contained 71% tin(IV) and 29% tin(II).
Once the dihydroxyl-terminated PDMS samples were mixed with one of these two catalysts, the
29
resulting silanol condensation reactions were followed by liquid-state Si NMR spectroscopy.
29
Si NMR is well suited for probing this type of reaction, since the peak associated with end
groups, i.e, HO-Si groups (at -11.1 ppm), is clearly distinguishable from peaks associated with
silicons in the polymer backbone, -O-Si-O- (from -21 to -22.5 ppm). Since the dihydroxyl-
terminated PDMS used in this experiment has a low averaged molecular weight, silicons have
29
distinct chemical shifts depending on their proximity to the end groups. Figure 12 shows Si
NMR spectra obtained for the PDMS samples that contained the pristine and naturally aged tin

20
octoate. This spectrum is exactly like the one for the neat polymer. As can be seen, the tin
catalyst with a high tin(IV) content does not promote the condensation chemistry, whereas the
sample with a very low tin(IV) content promptly catalyzes the reaction. This is verified by the
disappearance of the Si-OH groups for the PDMS sample that had the catalyst with low tin(IV)
content.

Figure 12. 29Si NMR spectra obtained for dihydroxyl-terminated PDMS mixed with tin octoate
containing either high or low tin(IV) concentrations. (a) PDMS sample mixed with tin octoate
containing 18% tin(IV) species, and (b) PDMS sample mixed with tin octoate containing 71% of
tin (IV) species. This spectrum is exactly as the one of the neat polymer. NMR experiments were
performed at room temperature.

4. Conclusions.

The goal of the present work was to identify aging mechanisms that degrade RTV
polysiloxane foams when subjected to aging under compressive strain. Two sets of foams were
artificially aged either in high humidity environments or ambient conditions. It was found that

21
compression set was more pronounced in foams aged in presence of moisture than in foams
aged in an open air environment. Since the RTV foams are synthesized using relatively large
amounts of a tin(II) catalyst, it was important to identify chemical reactions that could be
catalyzed by its residues. The degradation of the tin catalyst, which occurs through oxidation
and hydrolysis processes, produces tin(IV) species and hexanoic acid. Our results showed that a
degraded catalyst sample that contained a large ratio of tin(IV) to tin(II) species no longer
readily catalyzed silanol condensation reactions in dihydroxyl-terminated PDMS, indicating loss
of catalytic activity with oxidation. The carboxylic acid, once combined with trace amounts of
water, was found to promote chain scission reactions in trimethyl-terminated PDMS as
demonstrated by GPC experiments. Furthermore, tin residues combined with trace amounts of
water were able to catalyze bond-exchange reactions in trimethyl-terminated PDMS. The same
17 17
reaction was also observed by O NMR in RTV foams thermally aged with excess of O
isotopically enriched water. Thus, chain scission and bond-exchange reactions were identified
as the main reactions promoted by trace amounts of water when combined with tin residues
and octanoic acid. Based on these observations, S5370 foams containing mainly oxidized tin
catalyst were artificially aged at 70°C under compressive load, and in high humidity
environments. These foams were 28 years-old, and were chosen because of their low tin(II)
residue concentration. It was determined that accelerated aged S5370 foams exhibited large
compression set, which was attributed to chemical relaxation of the polymer chain largely
induced by chain scission reactions. On the other hand, SX358 foams aged under lab conditions
showed compression set that was mostly attributed to post-curing reactions. These were
catalyzed by tin(II) residues and accompanied by H2 evolution. In conclusion, the picture that
emerged is a detrimental one regarding the long-term use of RTV foams under compressive
strain and in high humidity environments.

Acknowledgments: This work was funded by the Enhanced Surveillance Campaign, and
the US Department of Energy’s National Nuclear Security Administration under contract DE-
AC52-06NA25396.

22
 References.

(1) Mark, J. E. Accounts Chem Res 2004, 37, 946.

(2) Grassie, N.; Macfarlane, I. G. Eur Polym J 1978, 14, 875.
(3) Grassie, N.; Francey, K. F.; Macfarlane, I. G. Polym Degrad Stab 1980, 2, 67.
(4) Grassie, N.; Francey, K. F. Polym Degrad Stab 1980, 2, 53.
(5) Visser, S. A.; Hewitt, C. E.; Binga, T. D. J Polym Sci Pol Phys 1996, 34, 1679.
(6) Thomas, T. H.; Kendrick, T. C. J Polym Sci A2 1969, 7, 537.
(7) Stein, J.; Prutzman, L. C. J Appl Polym Sci 1988, 36, 511.
(8) Osthoff, R. C.; Bueche, A. M.; Grubb, W. T. J Am Chem Soc 1954, 76, 4659.
(9) Lin, S. B. Acs Sym Ser 1995, 603, 37.
(10) Coons, J. E.; McKay, M. D.; Hamada, M. S. Polym Degrad Stab 2006, 91, 1824.
(11) Patel, M.; Chinn, S.; Maxwell, R. S.; Wilson, T. S.; Birdsell, S. A. Polym Degrad Stab 2010,
95, 2499.
(12) Gilbert, A. R.; Kantor, S. W. J Polym Sci 1959, 40, 35.
(13) Andrews, R. D.; Tobolsky, A. V.; Hanson, E. E. J Appl Phys 1946, 17, 352.
(14) Rottach, D. R.; Curro, J. G.; Grest, G. S.; Thompson, A. P. Macromolecules 2004, 37, 5468.
(15) Labouriau, A.; Taylor, D.; Stephens, T. S.; Pasternak, M. Polym Degrad Stab 2006, 91,
1896.
(16) Severnyi, V. V.; Minasyan, R. M.; Makarenko, I. A.; Bizyukova, N. M. Vysokomol Soedin
a+ 1976, 18, 1276.
(17) Cervantes, J.; Zarraga, R.; Salazar-Hernandez, C. Appl Organomet Chem 2012, 26, 157.
(18) Vanderweij, F. W. Makromol Chem 1980, 181, 2541.
(19) Cypryk, M.; Apeloig, Y. Organometallics 2002, 21, 2165.
(20) Leenslag, J. W.; Pennings, A. J. Makromol Chem 1987, 188, 1809.
(21) Ferry, J. D. Viscoelastic properties of polymers; Wiley: New York, 1980.
(22) Labouriau, A.; Cady, C.; Gill, J.; Taylor, D.; Zocco, A.; Stull, J.; Henderson, K.; Wrobleski, D.
Polym Degrad Stab 2015, 117, 75.
(23) Robinson, M. W. C.; Swain, A. C.; Khan, N. A. Polym Degrad Stab 2015, 116, 88.
(24) Blair, M. W.; Muenchausen, R. E.; Taylor, R. D.; Labouriau, A.; Cooke, D. W.; Stephens, T.
S. Polym Degrad Stab 2008, 93, 1585.
(25) Hubert, C.; Ziemons, E.; Rozet, E.; Breuer, A.; Lambert, A.; Jasselette, C.; De Bleye, C.;
Lejeune, R.; Hubert, P. Talanta 2010, 80, 1413.
(26) Lai, S. K.; Batra, A.; Cohen, C. Polymer 2005, 46, 4204.
(27) Kantor, S. W.; Grubb, W. T.; Osthoff, R. C. J Am Chem Soc 1954, 76, 5190.
(28) Johnson, D. H.; Mcloughlin, J. R.; Tobolsky, A. V. J Phys Chem-Us 1954, 58, 1073.
(29) Alam, T. M. Rad Phys Chem 2001, 62, 145.
(30) Patel, M.; Morrell, P.; Cunningham, J.; Khan, N.; Maxwell, R. S.; Chinn, S. C. Polym
Degrad Stab 2008, 93, 513.
(31) He, X. W.; Widmaier, J. M.; Herz, J. E.; Meyer, G. C. Eur Polym J 1988, 24, 1145.
(32) Normand, F.; He, X. W.; Widmaier, J. M.; Meyer, G. C.; Herz, J. E. Eur Polym J 1989, 25,
371.
(33) Clarson, S. J.; Wang, Z. M.; Mark, J. E. Eur Polym J 1990, 26, 621.

23