Version of Record: https://www.sciencedirect.

com/science/article/pii/S0141391017302963
Manuscript_a8f711aec47770e756ea7a7ec9ba75e1

Silicone Rubber Aging in Electrolyzed Aqueous Salt Environments

Authors: M. Bleszynski and M. Kumosa*
University of Denver
NSF I/UCRC for Novel High Voltage/Temperature Materials and Structures
Denver, Colorado, USA
*Correspondence to: mkumosa@du.edu

Abstract

In this work we suggest an entirely new and highly transformative model of

silicone rubber aging caused by the presence of electrolyzed aqueous salt, leading to the

formation of hypochlorous acid in-service on energized High Voltage (HV) transmission

lines in coastal environments. The model was subsequently verified by aging one-

component Room Temperature Vulcanized (RTV-1) rubbers in commercially produced

hypochlorous acid and an electrolyzed aqueous salt environment at room temperature.

Molecular dynamics models of the rubbers with silica fillers were conducted to offer at

least partial explanations of the types and extents of aging observed in the RTV-1

samples tested in the aqueous salt solutions. We have shown that the above environments

are highly damaging to the silicone rubber polymer network and significantly more

destructive than non-electrolyzed standard aqueous salt solutions at higher temperatures.

The main observations of this research should greatly contribute to a better understanding

of complex degradation mechanisms of modern HV silicone rubber compounds, which

are utilized in some of the most critical lines and substations all over the world.

Keywords: silicone rubber aging, HV transmission, molecular dynamics, modeling

© 2017 published by Elsevier. This manuscript is made available under the Elsevier user license
https://www.elsevier.com/open-access/userlicense/1.0/
1. Introduction

1.1 Application of silicone rubbers on high voltage transmission lines

Polymers and their composites can be commonly found on modern high voltage

(HV) transmission lines (1-30). For example, room temperature vulcanized (RTV)

silicone rubbers (SIR) (1-14) are used in Non-Ceramic Insulators (NCIs) (15-24)

supporting some of the most critical lines across the world. Mechanical, electrical and

chemical catastrophic failures of the polymer based HV transmission components have

been occurring sporadically in different parts of the world (15, 16, 22), including the US,

bringing down energized transmission lines and creating blackouts, ground fires, and

many other disasters. Due to the exceptionally aggressive nature of the HV transmission

line environment, it has been extremely difficult to predict and prevent these failures

when designing state of the art modern polymer based components for HV applications.

The primary objectives of this work are to properly explain SIR aging in aqueous salt

environments and associated potential in-service failures of modern transmission systems

using highly interdisciplinary and transformative approaches.

1.2 Silicone rubber aging in high voltage environments

In service aging of SIR sheds and sealants used in HV applications is well known

(1, 2, 3-14). Damage can result from various environmental and electrical factors such as

salt, acid, UV, corona discharges, leakage currents, etc., which routinely cause damage to

HV NCIs. RTV-1 is a one-component based caulk and is similar to RTV-2 two-

component SIR. RTV-1 and RTV-2 are both found on NCIs and denote two chemical

2
methods for the production of cross-linked SIR materials. Both RTV materials are

comprised of large amounts of polydimethylsiloxane (PDMS): a silicon based polymer

with hydrophobic methyl groups and commercial versions include large amounts of silica

fillers as binding agents. RTV-1 is commonly used to seal the end fittings of high voltage

NCIs, while HV sheds and housings are comprised mainly of RTV-2. RTV-1 makes up

the primary portion of the triple seal system (figure 1a), which prevents water ingress into

the fiberglass core through the fitting (1).

Figure 1 (a-c). (a) Failure of RTV-1 sealant in a HV NCI causing catastrophic insulator
and line failure by brittle fracture (b) and electrical failures (c) of a 500kV transmission
line in the Monterey Bay area in California, USA (3, 11, 22).

3
 When an RTV-1 sealant is compromised by aging (figure 1a), improper

manufacturing, vandalism, etc., water will penetrate into the fittings causing either

insulator failures by brittle fracture (figure 1b) and line drops of energized lines, or

electrical failures (figure 1 c,) (3, 11, 22) also leading to major catastrophic failures of the

lines. The processes illustrated in figure 1 a-c are very complex with continuously

changing mechanical, electrical and environmental stresses.

1.3 Silicone rubber aging in the presence of aqueous salt

Salt contamination in coastal environments is an identified cause of damage to

RTV SIR materials in NCIs, and salt exposure can decrease the hydrophobicity of RTVs

over time (3). Salt contamination can also cause increased corona discharge, corona

arcing, and a loss of hydrophobicity, resulting in permanent material damage to silicone

rubber RTV (i.e., chain scission) (1, 2). Historically NCI failures in coastal environments

have generally been assumed to be solely caused by chlorine damaging the PDMS

network (1 – 4).

Our previous aging experiments have indicated that permanent and significant

material damage to RTV-1 can occur after ten weeks in 3% NaCl saltwater at 80°C. This

condition caused surface degradation, which occurred in the form of voids beginning at

two weeks aging time (1, 2). This reduced the hardness of the material by 50% after ten

weeks. Mass losses and significant declines in surface hydrophobicity were also

recorded. RTV-1 material damage was more extensive in salt water than under individual

and combined UV, moisture, and nitric acid exposures (1, 2). Chain scission of the Si-O-

Si backbone was determined to be the primary mechanism responsible for the RTV’s
material loss in aqueous NaCl, which resulted in loss of the hydrophobic methyl groups

(1). Current literature indicates chain scission in PDMS produces cyclic and linear low

weight chains, resulting in low molecular weight (LMW) silicone oil (2). This oil can

account for some of the healing properties seen in SIRs, and is the primary explanation

for the recovery of hydrophobicity seen in experimental results (1, 2).

2. Proposed New Model of Aging of Silicone Rubbers in Electrolyzed Aqueous Salt

Environments

Under the right conditions, a fully drenched insulator may create an electrolyzed

saltwater environment in the presence of ocean salt, moisture, and leakage currents.

Because RTVs are commonly used on energized HV insulators where corona arcing and

leakage currents are prominent, additional aging due to secondary reactions could also

occur. Leakage currents (20, 24) on contaminated transmission lines could easily exceed

~ +1.63 V on or near the energized ends, the voltage necessary for reactions described

below. This unique energized environment could lead to a secondary reaction in standing

NaCl salt water, resulting in the formation of highly oxidizing hypochlorous acid (HOCl)

(1).

The electrolysis of salt water creates products through the overall reaction: 2NaCl

+ 2H2O → Cl2 + 2Na+ + H2 +2OH-. This results in the potential formation of highly

oxidative hypochlorous acid as chlorine hydrolyses in water through the following

equation: Cl2 + H2O → HClO + HCl (31, 32). At present, however, there are no known

technologies that could be employed to detect such acids in-service on energized HV

5
transmission lines. Therefore, the model suggested in this work is the only available

indication that HOCl and its byproducts could be a major cause of in-service aging of HV

NCIs in the presence of aqueous NaCl, and, perhaps other salts (e.g., MgCl, commonly

used for deicing purposes). Depolymerization as a result of hydrolysis in NaCl aqueous

environments is responsible for the degradation of RTV type polymers and monomer

chains, resulting in the decrease of their average molecular weight and structural integrity

(1, 2). Available literature suggests that HOCl can cause damage to polypropylene pipes

in the presence of heated chlorinated water as a result of the polymer backbone breaking

at the tertiary carbon sites (33). In our work, we compare for the first time the response to

aging of an RTV-1 compound under HOCl and aqueous NaCl salt exposures.

The acid ionization constant (Ka), determines the equilibrium constant of a weak

acid in solution, and can be used to predict the disassociation of an acid into its

constituents. The Ka of HOCl is 3.5 × 10-8, and its power of oxidation is 1.1 compared to

that of chlorine of 1.0 (32-34). HOCl therefore rapidly dissociates into damaging

hydrogen and hypochlorite ions OCl- and H+ in solution, which can re-associate back into

the HOCl original form under ambient conditions (32, 33).

To verify the new model of SRI aging, a commercially available RTV-1

compound was subjected to commercially available hypochlorous acid and electrolyzed

3% aqueous salt solutions to mimic a typical coastal environment. The presence of

electrolyzed salt solutions on HV NCIs cannot be questioned. What can still be

questioned, however is the presence of the hypochlorous acid on the lines. Therefore, it is

important to compare individually the effects of hypochlorous acid and energized salt

6
solutions on silicone rubbers of NCIs and determine how they degrade the rubbers in

comparison with the traditional aqueous salt solution.

If the effects of the acid and energized salt on silicone rubbers are very similar

and much stronger than the effect of non-energized aqueous salt, this will show for the

first time that the electric field (voltage) is critical in the degradation processes of HV

SIRs. This will also provide a significant amount of credibility to the main claims of this

work that either hypochlorous acid or electrolyzed salt solutions are responsible for the

failure of RTVs in coastal environments and not just chlorine which has been the only

explanation available so far (1-4).

3. Molecular Dynamics Simulations of PDMS

3.1 Polymeric adsorption, absorption, and diffusion

Initial sorption of a solvent or permeating compound through polymeric RTV

materials occurs by the stages of absorption, adsorption, and pooling of the penetrant

molecule relative to the bulk material (35, 36). Absorption in RTV occurs in the bulk

material where voids within the polymer network are occupied by permeating

compounds, and is modeled using Henry’s law of absorption (35). Adsorption occurs via

the molecular interactions of the permeating compound on the substrate surface, resulting

in adhesion or immobilization of the permeating compounds. For silicone rubber

materials such as RTV-1, this can result in more Langmuir adsorption due to the highly

hydrophilic silica fillers, and may cause permeating compounds to pool (35). Diffusion

theory within polymeric materials is more complex, and the diffusion rate of permeating

7
compounds can be determined by various properties such as the concentration of the

penetrating compounds relative to the polymer, void concentration, and molecular

interactions (Van der Waals) (35, 36).

Because the diffusion pathways for penetrating compounds in filled RTV-1

materials are highly complex, a simplistic computational model was developed to assess

the diffusion coefficients of various compounds in unfilled PDMS using Mean Square

Displacement (MSD). MSD, based upon Brownian motion, is calculated according to the

equation shown below, where Nα is the number of diffusive atoms in the model, ri(t) is

the position of molecule i at time t and D is the diffusion coefficient of a molecule (i) (1).

3.2 Molecular Dynamics models

Two molecular dynamics (MD) diffusion models were developed using Materials

Studio to assess diffusion rates of HOCl, as well as other products (water and Cl, Na

ions) through unfilled (figure 2) and silica filled (figure 3) RTV-1 polymer networks.

Similar to our previous model, molecular dynamics simulations were performed to better

assess the interaction of various possible compounds diffusing within the PDMS chain

network and help predict possible future aging effects in the presence of the silica fillers.

Although these models are simplified representations of a PDMS network and do not

precisely represent a real in-service material, they intended to identify the molecular

interactions of various compounds which may preferentially diffuse and thus potentially

damage the silicone rubber material. Identification of a highly diffusive compound on a

8
molecular level using molecular dynamics therefore may permit more precise evaluation

of molecular interactions and diffusion rates.

3.3 Diffusivities of aqueous salt byproducts in unfilled PDMS

Two separate iterations of the unfilled PDMS diffusion model were performed for

each permeating compound and each diffusion molecule or ion was modeled over a

period of 10 picoseconds. Molecular dynamics calculations were done at NVE for a

constant number of molecules (N), constant volume (V), and constant energy (E). The

initial density of the amorphous cell structure was set at 0.9 g/cm3, and a mass ratio of

99% Si-O-Si backbone to 1% diffusion molecule was set. Annealing was done at NVT

for a constant number of molecules (N), a constant volume (V), with temperature (T) set

for consistency with the performed experimental aging.

The first run was modeled at 80°C and the second at 25°C with an amorphous cell

density of ~0.95 g/cm3 after optimization, mimicking the density of physical pure PDMS

(36). Forcite Mechanics tools were set to group based calculations for both electrostatic

and van der Waals interactions. Additionally, each PDMS and diffusion molecule or ion

variation underwent geometry optimization and annealing at constant temperature and

energy for 10,000 cycles.

Forcite calculations were done at each interval for all previously mentioned

molecules and ions to generate the Mean Squared Displacement (MSD) data for a period

of 10 picoseconds. Forcite analysis was then conducted to generate the Mean Square

Displacement (MSD) data. The slopes of the curves in figure 2, which represent the MSD

9
rates of the modeled ions and molecules through the PDMS network, can be correlated to

molecular diffusion rates. The curves show a similarity in the diffusion rates for both

25°C and 80°C temperatures, although the diffusion at 25°C was a lower overall rate.

Even at the lower temperature, the model indicates that HOCl will likely diffuse through

the polymer network at a much faster rate than water, or Na and Cl ions.

Figure 2 (a, b) Left (a) Simulated molecular dynamics MSD diffusion rates for
permeating compounds: HOCl, H2O, Na and Cl ions at 25°C and 80°C, Right (b)
Materials Studio molecular dynamics model of PDMS in Amorphous Cell with HOCl
molecules.

3.4 Diffusivities of aqueous salt byproducts in silica filled PDMS

A second molecular dynamics simulation was conducted to assess the diffusion

rates of the previously mentioned permeating compounds (HOCl, H2O, Na and Cl ions)

through a PDMS polymer network filled with silica nanoparticles in order to more

realistically resemble molecular diffusion through filled RTV-1 silicone rubber. Two

SiO2 nanoparticles approximately 10Å in size were constructed independently and

optimized for geometry before being combined with previously constructed PDMS

chains using amorphous cell.

10
 Similar to the previous unfilled PDMS model, the molecular dynamics of the

silica filled PDMS model was modeled over a period of 10 picoseconds. All MD

calculations were done at NVE, annealing at NVT and the initial density of the

amorphous cell structure was set at 1.19 g/cm3. A mass ratio of 79% Si-O-Si backbone to

20% silica nanoparticles was assumed and the migrating species were assumed to be 1%

by volume. These parameters were set for consistency with the previous model.

The silica filled PDMS MD model was conducted at 25°C with a final cell density

of ~1.24 g/cm3 after optimization to more closely mimic the physical density of silica

filled PDMS (36). The final silica filled PDMS structure underwent a fine geometry

optimization at constant temperature and energy for 10,000 cycles. The final Forcite

calculations and analysis were conducted to generate MSD data for a period of 10

picoseconds, similar to the previous model. It was observed that molecules exhibited

faster diffusion rates when in closer proximity to the silica particles; therefore the MSD

data was averaged for all permeating molecules in the system to create the final MSD

diffusion curves.

Figure 3 (a, b) Left (a) Simulated molecular dynamics MSD diffusion rates for
permeating compounds: HOCl, H2O, Na and Cl ions at 25°C through filled PDMS
network Right (b) Materials Studio molecular dynamics model of PDMS in Amorphous

11
 The MSD slopes of the curves in figure 3a, show similar diffusion rates to the

unfilled modeled diffusion rates for both 25°C and 80°C temperatures, with some notable

differences. The overall diffusion curves are similar to the unfilled PDMS model;

however most of them exhibit higher final MSD diffusion rates, which may be due to the

presence of the hydrophilic silica particles in the model. A significant difference is the

diffusion curve for water, which closely follows that of the Cl ion, showing overall faster

diffusion in the silica filled PDMS compared to the unfilled PDMS. HOCl was observed

to rapidly diffuse near silica particles. Overall, HOCl still exhibits significantly faster

diffusion through the filled polymer network compared to water, Na and Cl ions.

4. Experimental

4.1 Materials tested

To verify our model experimentally, a previously investigated RTV-1 SIR

compound was chosen (1, 2). This material is a proprietary RTV-1 silicone rubber used in

the high voltage industry, and is dark grey in color. The material undergoes an amine

cure process, which was supported by FTIR analysis showing a lack of a carboxylic

(C=O) bond peak expected for an acid cure, and peaks for amine cure were confirmed

(1). The material was selected as it had above average recovery rates of hydrophobicity

after ocean concentration salt-water aging (1). For this study, 3mm thick samples were

created, and cured in a mold for 72 hours before aging. The measured density of the

cured samples was ~1.31 g/cm3.

12
5. Silicone Rubber Aging in Commercially Produced Hypochlorous Acid

5.1 Methods

For this study, the RTV-1 silicone rubber material was subjected to HOCl at RT

and compared to the previously aged samples in aqueous 3% NaCl at 80°C from our

previous study (1). The acid was commercially produced and contained electrolyzed

water with 500ppm free chlorine for a total concentration of 0.046% HOCl. The 3% NaCl

salt solution would have 30 times more chlorine than the HOCl solution. The

experimental setup is identical to the one presented in (1). After aging, the samples were

assessed for changes in hardness, mass, and hydrophobicity. Hardness readings were

recorded using a standard Shore A durometer (1). Hydrophobicity was evaluated the

contact angle of a single water droplet on dried aged sample surfaces. The tests were

performed on dry aged samples in order to discount temporary absorption or swelling of

the sample due to the immersion solution. SEM analysis was also conducted on the virgin

and aged samples (1, 2, 37).

5.2 Results

Cross-section SEM images of five week aged samples in HOCl at RT show

extensive material damage in the form of voids and cavities (figure 4b) compared with

the virgin material (figure 4a). Additionally, a five week aged sample in 3% NaCl

aqueous solution at 80°C (figure 4c) is provided for comparison, showing significant

crack propagation (semi- straight faint lines in figure 4c), but substantially fewer voids

and cavities compared to the five week HOCl aged sample. Damage in the HOCl aged

samples appeared after a week of aging as a band of voids close to the sample surface,

13
which was not observed in the salt aged samples even after five weeks of exposure at

80°C. Initial void penetration was approximately 90 µm from the sample surface after

one week. The voids progressed through the bulk of the 3mm thick sample material, with

the damage encompassing the entire interior after five weeks, as seen in figure 4b.

Figure 4 (a-c). SEM cross sections of a RTV-1 (a) virgin, (b) subjected to 0.046% HOCl
for five weeks at room temperature, (c) aged 5 weeks in 3% NaCl aqueous solution at
80°C.

FTIR results already reported in (1) confirmed the presence of Si-O-Si and CH3 in

small flakes of the RTV-1 material produced during testing in the 3% 80°C salt

immersion solution. Material shedding in the form of flakes was also determined to be

the primary cause responsible for the observed mass loss seen in figure 5b (1). FTIR

14
analysis of virgin and 0.046% RT HOCl RTV-1 aged samples also demonstrated

significant chemical changes to the material caused by the acid (37).

Hardness readings were approximately 40HA for all virgin RTV-1 samples

(figure 5a). The subsequent hardness tests on the aged samples showed greater decreases

in hardness over five weeks time for the acid aged samples at RT (22HA) compared to

the salt solution aged samples (29HA) at 80°C. After five weeks the samples lost about

50% of their hardness in the acid and about 25% in the salt solution. Five hardness

measurements were conducted per condition and the scatter in the hardness data was very

low irrespective of the testing conditions.

The mass recordings for the RTV-1 sample aged in the acid indicated changes in

percent mass compared to RTV-1 aged in the salt solution, as seen in figure 5a. After five

weeks of exposure, the mass of the fully dried RTV-1 sampled aged in the acid decreased

0.80% more compared to the 3% NaCl aged sample. The mass readings for the salt

solution gradually decreased in a stepwise fashion, however the sample aged in the acid

declined in an almost linear manner. The mass loss of the RTV in the salt solution

occurred predominantly by surface shedding (1) as opposed to rapid void nucleation and

progression in the acid aged samples, which may be due to non-uniform surface

shedding.

15
Figure 5 (a-b). (A) Comparison of durometer hardness readings of aged RTV-1 samples
over five weeks aging time in 0.046% HOCl at 80°C and 3% NaCl at 25°C, (B) and
comparison of mass changes of RTV-1 samples aged in 0.046% HOCl at 80°C and 3%
NaCl at 25°C over five weeks aging time.

Hydrophobicity measurements performed on the virgin RTV-1 samples yielded a

contact angle of 85.6°, and lower contact angles for the aged samples. The angle for the

sample aged in the acid at RT for five weeks was 43.1°, while a ten week aged sample in

the salt solution at 80°C exhibited a larger contact angle of 52.2˚ (figure 6). The surface

contact angle of the five week aged samples in the acid therefore indicates a significantly

greater loss in hydrophobicity: 42.5° loss in contact angle in the acid vs. 33.4° loss in the

salt solution. Standard deviation for the contact angles of the virgin sample was minimal,

at 2%. For HOCl and NaCl aged samples the standard deviations were higher at 8% and

6.5% respectively, which was likely caused by surface irregularities in the aged samples.

Figure 6 (a-c). Contact angles of for (a) virgin RTV, (b) aged for five weeks in HOCl at
RT, (c) and after ten weeks in 3% 80ºC NaCl.

16
6. Silicone Rubber Aging in an Electrolyzed Salt Environment

To further support our claim that either hypochlorous acid or aqueous

electrolyzed salt HV conditions may form in-service, the same type of RTV-1 was

subjected to aging in an electrolyzed ocean concentration 3% NaCl solution. The RTV-1

sample was placed in a dish containing 300ml of 3% NaCl solution. Subsequently, the

sample was flanked at either end by wires connected to a 9V battery to produce ample

current for electrolysis to occur. The electrodes were placed without a physical contact

with the sample. Wire placement relative to the sample is shown in figure 7. The NaCl

salt water across the polymer surface completed the circuit, resulting in the electrolysis of

the NaCl solution. Hydrogen and chlorine gas bubbles emanating from the electrodes at

the cathode and anode ends, respectively, confirmed active electrolysis. The sample was

left submerged for a total of 8 hours.

After removal from the solution, approximately 50% of the sample was visibly

damaged with an ‘orange peel’ surface, as seen in figure 8a. The damaged portion of the

sample extended approximately 1mm into the bulk material, and was soft to the touch.

The rest of the sample appeared uniform and undamaged. Hydrophobicity and hardness

measurements were attempted on the damaged portion of the sample surface, but were

deemed too inconsistent to be accurately determined due to the irregularity of the surface.

17
 +

Figure 7. Virgin RTV-1 sample before aging, with markings showing position of
the electrodes.

Figure 8 (a-c). (a) RTV-1 sample after 8 hours of aging in electrolyzed 3% NaCl solution
(b) optical microscope image showing cross section of damaged portion of the RTV-1
sample (c) SEM cross-section of the damaged portion of the RTV-1 sample.

After drying the sample under a heat lamp for 48 hours, a cross section of the

sample was observed under an optical microscope (figure 8b). SEM images of the cross

sections spanning the damaged and undamaged portions of the sample were also taken

showing extensive material damage in the form of voids and cavities (figure 8c). The

voids and cavities seen in the electrolyzed salt sample are comparable in size and quantity

as those found in the sample aged for five weeks in commercially produced HOCl (figure

4b). Although large concentrations of voids and cavities were found in the damaged

section of the sample, SEM analysis of multiple cross sections showed voids and cavities

throughout the sample. This suggests that although upon initial inspection the entire

18
sample did not seem affected by the electrolyzed salt solution, the entire polymer was

damaged at the molecular level.

7. Discussion

The MSD diffusion calculations performed in this study offer partial explanations

of the types and extents of aging observed in the RTV-1 samples tested in the

hypochlorous acid and salt solutions. HOCl, water, sodium and chlorine ions were

modeled for comparison. All models at 25°C and 80°C indicate that the MSD diffusion

curves are higher for HOCl, compared to chlorine ions, water and sodium ions. Although

the MSD diffusion is lower at 25°C than at 80°C for the unfilled case, the diffusion

curves for each permeating compound retain the same shape. The filled model also

showed the fastest diffusion for HOCl. The high MSD diffusion rate of HOCl, coupled

with a higher power of oxidation, therefore makes HOCl a potentially more aggressive

aging compound than chlorine.

In-service, damage to the PDMS silicone rubber may be multifold. Hydrolysis

and oxidation damage due to HOCl may account for some of the voids and cavities

observed in figures 4b and 8c. While chain scission in the polymer is evident due to the

loss of material, the specific cause is still unknown. Since HOCl molecules were

observed to readily diffuse near silica particles in the MD models, it is possible that the

silica particles present in the RTV-1 allow for rapid diffusion and concentration of HOCl,

which may result in areas of pooling sites (36). Because HOCl is a known oxidizer, the

19
pooling sites could be the sites of rapid polymer degradation resulting in the formation of

large concentration of micro-voids (35, 36).

Sample damage appears to be more extensive in the electrolyzed 3% NaCl salt

sample, which may be due to the higher concentration of chlorine, and thus hypochlorous

acid, compared to the commercially produced HOCl concentration of 0.046%.

Additionally, although the majority of the damage was found at the cathode end in the

electrolyzed salt experiment, voids and cavities were found throughout the sample,

indicating the damage is not localized and thus not dependent upon the proximity of the

energized wires.

While it is clear that an electrolyzed salt environment is more damaging than

aqueous NaCl, it is not clear if the damage is also caused by the by-products of

electrolysis, or from the by-products of hypochlorous acid itself, such as H+ or OCl-. By

itself, hypochlorite (OCl-) is the strongest oxidizing agent of the chlorine constituents,

often used as a bleach solution, and could be partially responsible for the observed

polymer damage. Additional electrolysis products such as chlorine gas, hydrogen gas,

and hydroxide ions may also be partially responsible for the chain scission damage seen

in the cross sections of both RTV-1 experiments, including the formation of free radicals.

PDMS, in the presence of water and dry band arcing (4) has been shown to cause free

radicals within the polymer chain, and hydrogen transfer from the CH3 methyl groups

may occur just in the presence of dry band arcing alone, causing chain scission (12, 13).

20
 Aging by aqueous NaCl salt, as shown in figure 4c, indicates a different aging

mechanism due to the significant formation of cracks along with fewer voids and cavities.

Crack propagation in salt solution could increase the possibility of macro scale pieces

(flakes) shedding from the silicone rubber, while low molecular weight leaching could be

the main failure process in HOCl. On transmission lines, this could result in the leaching

of siloxane products, should a NCI be continuously exposed.

8. Conclusions

It has been shown in this work that very complex aging mechanisms can exist in

modern transmission systems, which can be difficult to determine. We suggest a new

method of aging of PDMS RTV-1 in the presence of aqueous NaCl and voltage, which

could result in the formation of hypochlorous acid. We question the previous models

based solely upon chlorine, which may not account for all damage present in silicone

rubbers used in coastal NCIs. In this new model, the effect of hypochlorous acid and

electrolyzed aqueous salt on RTV-1 at room temperature was shown to be much more

damaging compared to standard ocean concentration aqueous NaCl salt at high

temperature. Commercially produced HOCl and electrolyzed aqueous NaCl salt produced

very similar polymer damage in the form of voids and cavities within the polymer

network. This was demonstrated despite the fact that the HOCl acid tests were performed

at room temperature, the electrolyzed salt experiments were conducted only for 8 hours,

and the commercially produced HOCl had a chlorine concentration approximately 30

times lower in comparison with standard ocean salt concentrations.

21
Acknowledgments:

This work was funded by the National Science Foundation I/UCRC Center for Novel
High Voltage Materials and Structures under #IIP 1362135 and by the NSF Grant
Opportunities for Academic Liaison with Industry program under #CMMI-123252. The
authors are grateful to Drs. Paul Predecki and Joseph Hoffman from the HVT Center for
their very useful comments on this paper.

References:
1. Bleszynski M, Kumosa M. (2016) Silicone Rubber RTV-1 Aging in the Presence of Aqueous Salt.
IEEE Transactions on Dielectrics and Electrical Insulation
2. Allen B, Bleszynski M., Kumosa, M., & Willis, E. (2015) Investigation into the effects of
environmental stresses on RTV-1 silicone-based caulk materials. IEEE Transactions on
Dielectrics and Electrical Insulation 22:2978-2986.
3. R. M. Fernando M, Gubanski S (2010) Ageing of silicone rubber insulators in coastal and inland
tropical environment. IEEE Transactions on Dielectrics and Electrical Insulation 17:326-333
4. Kim S (1992) “Electrical performance and surface analysis of RTV silicone rubber coatings for
HV outdoor insulators.".
5. Teverovsky A (1999) Chlorine contamination diffusion in silicones (EEE LINKS, NASA)
6. Amin M, Salman M (2006) Aging of polymeric insulators (an overview). Reviews on Advanced
Material Science 2:93-116.
7. Homma H, Mirley C, Ronzello J, Boggs S (2000) Field and laboratory aging of RTV silicone
insulator coatings. IEEE Transactions on Power Delivery 15:1298-1303.
8. R. A. Bernstorf, et al. (2000) Polymer compounds used in high voltage insulators. Hubbell power
Systems.
9. R. A. Bernstorf, Ryan, D (2007). Silicone Compounds for High-Voltage Insulators: Compounding
Silicone Rubber. Hubbel Power Syst. Inc, 1-8.
10. Kim S, Cherney E, Hackam R (1992) Hydrophobic behavior of insulators coated with RTV
silicone rubber. IEEE Transactions on Electrical Insulation 27:610-622
11. Tsarabaris P, Karagiannopoulos C, Theodorou N (2005) A model for high voltage polluted
insulators suffering arcs and partial discharges. Simulation Modelling Practice and Theory
13:157-167
12. Seog-Hyeon Kim, Cherney E, Hackam R, Rutherford K (1994) Chemical changes at the surface of
RTV silicone rubber coatings on insulators during dry-band arcing. IEEE Transactions on
Dielectrics and Electrical Insulation 1:106-123
13. Wright D (2001) Failure of polymer products due to thermo-oxidation (Rapra Technology
Limited)
14. Jahromi N et al. (2005) Effect of acid immersion on RTV silicone rubber coatings in inclined
plane tests in Annual Report Conference on Electrical Insulation and Dielectric Phenomena
15. Kumosa M, Qiu, Q. (1996) Failure Analysis of Composite Insulators (Failure Investigation of 500
kV Non-ceramic Insulators for Pacific Gas & Electric Company). Final Report to the Pacific Gas
and Electric Company
16. Kumosa M, Narayan H, Qiu Q, Bansal A (1997) Brittle fracture of non-ceramic suspension
insulators with epoxy cone end-fittings. Composites Science and Technology 57:739-751.
17. Kumosa M, Han Y, Kumosa L (2002) Analyses of composite insulators with crimped end-fittings:
Part I—non linear finite element computations. Composites Science and Technology 62:1191-
1207

22
18. Kumosa L, Benedikt B, Armentrout D, Kumosa M (2004) Moisture absorption properties of
unidirectional glass/polymer composites used in composite (non-ceramic) insulators. Composites
Part A: Applied Science and Manufacturing 35:1049-1063.
19. Kumosa M, Kumosa L, Armentrout D (2004) Can water cause brittle fracture failures of
composite non-ceramic insulators in the absence of electric fields?. IEEE Transactions on
Dielectrics and Electrical Insulation 11:323-333
20. Armentrout D, Kumosa M, Kumosa L (2004) Water diffusion into and electrical testing of
composite insulator GRP rods. IEEE Transactions on Dielectrics and Electrical Insulation
11:306-322
21. Kumosa L, Kumosa M, Armentrout D (2005) Resistance to Brittle Fracture of Glass Reinforced
Polymer Composites Used in Composite (Nonceramic) Insulators. IEEE Transactions on Power
Delivery 20:2657-2666.
22. Kumosa M, Kumosa L, Armentrout D (2005) Failure analyses of nonceramic insulators. Part I:
Brittle fracture characteristics. IEEE Electrical Insulation Magazine 21:14-2
23. Kumosa M, Kumosa L, Armentrout D (2004) Causes and potential remedies of brittle fracture
failure of composite (nonceramic) insulators. IEEE Transactions on Dielectrics and Electrical
Insulation 11:1037-1048
24. Kumosa L, Armentrout D, Benedikt B, Kumosa M (2005) An investigation of moisture and
leakage currents in GRP composite hollow cylinders. IEEE Transactions on Dielectrics and
Electrical Insulation 12:1043-1059
25. Burks B, Armentrout D, Kumosa M (2010) Failure prediction analysis of an ACCC conductor
subjected to thermal and mechanical stresses. IEEE Transactions on Dielectrics and Electrical
Insulation 17:588-596.
26. Kumosa M, Burks B (2008) DU Report on ACCC In-Service Failures in Poland in 2008
(Department of Energy, USA under contract WAPA/HDR 36532A-272803).
27. Middleton J et al. (2013) The effect of ozone and high temperature on polymer degradation
in polymer core composite conductors. Polymer Degradation and Stability 98:2282-2290.
28. Hoffman J, Middleton J, Kumosa M (2015) Effect of a surface coating on flexural performance of
thermally aged hybrid glass/carbon epoxy composite rods. Composites Science and Technology
106:141-148.
29. Hakansson E, Predecki P, Kumosa M (2015) Galvanic Corrosion of High Temperature Low Sag
Aluminum Conductor Composite Core and Conventional Aluminum Conductor Steel Reinforced
Overhead High Voltage Conductors. IEEE Transactions on Reliability 64:928-934.
30. Hakansson E, Ricker R, Predecki P, Kumosa M (2016) Electrochemical In-situ Assessment
Method for Galvanic Corrosion in Bare Overhead Transmission Line Conductors. CIGRE Science
and Engineering 1:38-49
31. Fair G (1949) The behavior of chlorine as a water disinfectant (Harvard Univ., [Cambridge,
Mass.])
32. Eryilmaz MPalabiyik I (2013) Hypochlorous Acid - Analytical Methods and Antimicrobial
Activity. Tropical Journal of Pharmaceutical Research 12
33. Backman, A. Effects of Chlorinated Water on Polymeric Water Distribution Systems in Plumbing
Contractor News.
34. Swain Crist D (1972) Mechanisms of chlorination by hypochlorous acid. The last of chlorinium
ion, Cl+. Journal of the American Chemical Society 94:3195-3200
35. Harley, Stephen J., et al. "Advances in modeling sorption and diffusion of moisture in porous
reactive materials." ChemPhysChem 15.9 (2014): 1809-1820.
36. Sun, Yunwei, Stephen J. Harley, and Elizabeth A. Glascoe. "Modeling and uncertainty
quantification of vapor sorption and diffusion in heterogeneous polymers." ChemPhysChem 16.14
(2015): 3072-3083.
37. Pang S, (2016) FTIR Analysis of RTV-1 PDMS in Internal report by the University of Illinois at
Urbana Champaign for the NSF HVT Center.
38. Nakade M, Kameyama K, Ogawa M (2004) Synthesis and properties of
dioxide/polydimethylsiloxane hybrid particles. Journal of Materials Science 39:4131-4137.
39. Harris, Daniel C. Quantitative chemical analysis. Macmillan, 2010
40. Ohno K (2008) Nano- and micromaterials (Springer, Berlin)
41. Karimi, M (2011) Diffusion in polymer solids and solutions (INTECH Open Access Publisher)

23
42. Benjamin M (2014) Water Chemistry (Waveland Press, Long Grove, IL)
43. Southall N, Dill K, Haymet A (2002) A View of the Hydrophobic Effect. The Journal of Physical
Chemistry B 106:521-533

24
• A new transformative model of RTV aging in aqueous salt environments is proposed
• Molecular dynamics model of RTVs are constructed to predict diffusivities of salt based
compounds
• Electrolyzed aqueous salt solution appears to be most damaging to RTVs in costal environments