Effects of electrical-hydrothermal aging degradation on dielectric and trap properties of high temperature vulcanized silicone rubber materials

In this paper, a new aging platform combined the high voltage electric field and the hydrothermal environment was built. To investigate the aging mechanism, physicochemical, dielectric and trap properties of HTV SR before and after electrical-hydrothermal aging for 24 days were discussed. The results indicated that, compared with hydrothermal aging, more cracks and holes appeared on the surface of HTV SR after electrical-hydrothermal aging, and the content of flame retardant decreased significantly. Due to the main chain and side chain scission of PDMS, lots of low weight molecular (LWM) substances and free radicals were produced. And the tensile strength and elongation at break significantly decreased. Various physical and chemical defects appeared in the HTV SR specimen in the process of electrical-hydrothermal degradation, as a result of which, the dielectric constant significantly increased and the peak trap density increased by about 2.5 times compared to the virgin sample. The increase in trap density in turn accelerated the charge accumulation and enhanced the breakdown probability, resulting in the electric field strength decrease from 21.8 kV mm−1 to 16.1 kV mm−1 and severe degradation of HTV SR.


Introduction
Composite insulators are widely used in power systems because of their light weight and strong resistance to pollution ashover. [1][2][3][4][5][6][7][8] As an important material for the sheaths and sheds of composite insulators, high temperature vulcanized silicone rubber (HTV SR) plays an important role. However, as the operating life increases, the aging problem becomes more and more serious. Due to long-term exposure to the harsh environment, external insulation of composite insulators can suffer from multiple stresses such as electrical discharge, temperature and mechanical stress et al. [9][10][11][12][13] In hot and humid areas all over the world, they are also accompanied by the problem of hydrothermal aging. Therefore, it is of great signicance to carry out multiple stress aging research on external insulation materials of composite insulators for mastering the aging rule and evaluating the insulation state.
Degradation studies of HTV SR have been a topic for research from the last few decades. Z. D. Jia et al. 14 studied the thermooxidative aging process of HTV SR and calculated the activation energy of the aged specimens. Z. Wang et al. 15 investigated the absorption characteristics of HTV SR in deionized water, NaCl solution and HNO 3 solution, analyzing the effects of the three liquids on the electrical properties. Mansab Ali and Reuben Hackam 16 studied the loss of hydrophobicity of HTV SR aer immersion in saline solutions at different temperatures, revealing that the contact angle will decrease sharply during immersion. Yong Zhu et al. 17 described the degradation of HTV SR resulting from corona discharges under atmospheric pressure, showing that the corona discharges can cause physical and chemical damages to the surface of HTV SR. Recent studies focus on corona aging, thermal aging, electrical branch aging and other single physical eld aging mechanism. Comparative analysis of hydrothermal and electrical-hydrothermal aging has not been reported. However, as sheds of composite insulators, HTV SR has been exposed to multiple stresses, and the aging rate is much higher than the single eld aging. Therefore, studying the electrical-hydrothermal aging properties to analyze the aging mechanism is imperative.
In the present work, the aging mechanism of HTV SR exposed to multiple stresses has not been clearly claimed. In this paper, AC breakdown eld strength, dielectric spectroscopy, thermally stimulated current (TSC), Fourier transform infrared (FTIR) spectroscopy, scanning electron microscopy (SEM) and energy dispersive X-ray (EDAX) analysis were conducted to analyze changes of electrical characteristics and physical properties aer hydrothermal aging and electrical-2 Experimental details

Sample preparation
The HTV SR specimen was provided by Dongguan Gaoneng Company in China. It should be emphasized that the composite insulator manufactured by the Dongguan Gaoneng Company has been running successfully in hot and high humidity areas in southern China for nearly 20 years. The formulation of the HTV SR specimen was 100 parts per hundred (phr) methyl vinyl silicone rubber, 40 phr meteorological silica, 125 phr aluminum powder, 6 phr silicone oil, 2 phr vulcanizing agent and some iron oxide colorant. The methyl vinyl silicone rubber, meteorological silica, aluminum powder, silicone oil and colorant are added to an open mill in proportion and mixed uniformly, then heated in an oven at 150 C for 2 hours and added vulcanizing agent at room temperature. Aer the vulcanizing agent is mixed in the open mill, the compound is vulcanized and formed on a at vulcanized machine. Finally, the HTV SR with a diameter of 100 mm and a thickness of 2 mm is prepared. The detailed product information (code, supplier) of each component for the formulation of the HTV SR is listed in Table 1. Outside of China, the similar HTV SR specimen can be easily procured in factories that produce composite insulators. For instance, Mar-Bal company in America, Maschinenfabrik Reinhausen company in German, etc.

Test apparatus and test method
The experimental set up for investigating the electricalhydrothermal aging effect is shown in Fig. 1. It is composed of AC power box, rod-plate electrode, quartz tube and heating plate. Saline solutions were prepared as the conductive solution and the salt contaminant. The high voltage electrode, when immersed in the saline solutions, will provide a uniform and stable electric eld on the HTV SR. To verify the electric eld uniformity, the simulation model was established using the COMSOL and the results were shown in Fig. 2. As shown in Fig. 2, the HTV SR specimens were in a uniform electric eld. In the electrical-hydrothermal aging test, the temperature of the heating plate was controlled at 30 C, 60 C, 80 C and the electric eld intensity was 0 kV cm À1 (equivalent to hydrothermal aging), 15 kV cm À1 and 25 kV cm À1 . For comparison, the temperature of the saline solutions in the hydrothermal aging test was also controlled at 30 C, 60 C and 80 C.

Characterization
2.3.1 Rate of weight change. The initial weight of the HTV SR was measured using an electronic precision balance with a measurement accuracy of 0.1 mg. Aer aging for several days, the HTV SR with a diameter of 100 mm and a thickness of 2 mm was taken out and cleaned with anhydrous ethanol. Then, the weights of the HTV SR were measured at room temperature and humidity. When the indication of the electronic precision balance is stable, the weights of each HTV SR specimen were recorded. The rate of weight change can be calculated as the following equation: where m 1 and m 0 are the weight at aging time t and the initial weight.

Measurements of electrical characteristics.
To analyze the electrical characteristics under different aging behavior, the AC breakdown voltage, the complex dielectric constant and the thermally stimulated current were measured before aging and at different states of the aging process.
The AC breakdown voltage measurements were performed on the hydrothermal and electrical-hydrothermal aging  specimens. In the measurement of AC breakdown voltage, a set of 5 samples with a diameter of 100 mm and a thickness of 2 mm were tested to calculate the breakdown eld strength. The experimental setup for the AC breakdown voltage was shown in Fig. 3. The aging specimens are placed in the plate to plate arrangement, and placed into a vessel containing lling with insulation oil to avoid surface discharges. During the test, the voltage is increasing step by step until the breakdown occurs. The voltage was rising at the rate of 500 V s À1 . When the breakdown occurs, the AC breakdown stress can be calculated according to the equation E ¼ V/d, where V is the breakdown voltage and d is the thickness of the aging specimen in the breakdown point.
For the investigation of the frequency domain dielectric spectroscopy, the dielectric constant and loss of dielectric materials at different frequencies were analyzed. In this paper, the dielectric spectroscopy of different aging specimens with a diameter of 40 mm and a thickness of 2 mm was measured by ALPHA-ANB type broadband dielectric spectrometer produced by Guangzhou Huayu Electronic Service Co., Ltd. Each HTV SR specimen in the study, the dielectric response in the range of frequency at 10 0 Hz to 10 5 Hz was measured at room temperature.
For the measurement of TSC, the HTV SR specimen with a thickness of 2 mm and a diameter of 40 mm were vacuum vaporized with an Au electrode on both sides. Firstly, the HTV SR specimen was loaded between the electrodes and heated to 50 C. Secondly, the DC voltage of 3 kV was applied to the HTV SR and kept warm for 20 min. Then the HTV to À50 C quickly with liquid nitrogen. Aer keeping the temperature at À50 C and the voltage at 3 kV for 2 min, the polarization voltage can be removed. When the current decays to zero, the specimen can be heated at a rate of 5 C min À1 . In the process, the depolarization current was measured as a function of temperature.
2.3.3 Measurements of physicochemical characteristics and mechanical characteristics. In the measurements of physicochemical properties, FTIR spectroscopy (VERTEX 70) was applied to study the surface chemical groups of initial samples, hydrothermal aging and electrical-hydrothermal aging samples. To study the surface aging degree, SEM (LEO 1530VP), outtted with the EDAX detector to recognize the elements, was employed to compare the effects in different aging mode. The above testing samples are all 5 mm in diameter and 2 mm in thickness. Tensile properties were measured using the Electro-Mechanical Universal Testing Machines. The stretching rate was about 50 mm min À1 . Each HTV SR sample is 20 mm in length, 10 mm width and 2 mm in thickness.  Fig. 4, in the absence of an electric eld, the weights of HTV SR specimens immersed in saline solutions at 30 C increased rapidly at rst and then slowly increased within 24 days. However, when immersed in saline solutions at 60 C and 80 C, the weight of HTV SR increased rapidly and then decreased. This phenomenon might result from the increase of temperature and the invasion of water accelerating the hydrothermal aging effect of the HTV SR specimen. The relative weight change of HTV SR specimen depended on the competition between the invasion of water and the extraction of the low-molecular-weight (LMW) substances and the ller particles. Both process was inuenced by water temperature. During the immersion for 24 days at 30 C, the weight increased because of the invasion of water over counteracted the weight decrease caused by the extraction of the LMW substances and the ller particles. However, at 60 C and 80 C, the weight of HTV SR decreased rapidly because the oxidation decomposition of HTV SR specimen and the extraction of the LMW substances and the ller particles over counteracted the weight increase due to the diffusion of water into HTV SR.

Relative weight change
When considering the electric eld, the weight of the HTV SR samples was obviously decreased. With the increase of electric eld intensity, the nal mass loss rate was signicantly higher  This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 3805-3816 | 3807 than that of the HTV SR samples without an electric eld. The nal relative weight change of the hydrothermal aging samples at 80 C was 0.27%, while the relative weight change rate reached À1.71% under the 80 C and 25 kV cm À1 electricalhydrothermal aging condition. Therefore, there was a synergistic effect of mutual enhancement of electrical-hydrothermal aging process, which signicantly accelerated the weight reduction of HTV SR specimens. The presence of an electric eld accelerated the hydrothermal aging process, which in turn accelerated the deterioration of the electric eld to HTV SR samples.

AC breakdown studies
The Weibull distribution, proposed by the Swedish physicist Weibull in 1939, has been widely used to represent breakdown voltages. [18][19][20] It reected the failure probability of a dielectric being broken down at a certain electric eld strength E. In this test, the two-parameter Weibull distribution was applied to evaluate the statistical law of the breakdown behavior of HTV SR under alternating electric elds. The function can be expressed as: 15 where P(E) was the cumulative failure probability of breakdown; E was the breakdown eld strength of HTV SR; b was the shape parameter related to the shape or the width of the distribution; E 0 represented the breakdown eld strength at P(E) ¼ 63.2%. To get E 0 , the eqn (2) can be written as follows: To nd the value of P(E), the Bernard estimator was given by ref. 21, where i was the number of measurements in the ascending order of E; n was the total number of breakdown eld strength tests per sample, and each HTV SR specimen was tested 5 times in this experiment. The AC breakdown characteristic of all samples under Weibull distribution is presented in Fig. 5. According to eqn (3), the value of the shape parameter b was calculated and listed in Table 2. Coppard 22 et al. believed that the shape parameter b represented the distribution of defects and lling particles. However, as seen in Table 2, the value of b did not show obvious  regularity but presented a great uctuation. It indicated that there were complex reactions in HTV SR aer hydrothermal and electrical-hydrothermal aging processes. Due to high temperature and high electric eld, severe degradation occurred in HTV SR and various physical and chemical defects were produced. Therefore, the change in the shape parameter b showed large variation in dispersity of electrical weak points in the aging samples. Then the value of breakdown eld strength under different aging conditions is shown in Fig. 6. As shown in Fig. 6, compared to the virgin sample and hydrothermal aging sample, the breakdown eld strength of the electrical-hydrothermal (80 C -25 kV cm À1 ) sample was signicantly decreased from 21.8 kV mm À1 to 16.1 kV mm À1 . It indicated that there was obvious synergistic aging effect of electric, water and temperature. The change of breakdown eld strength caused by electrical-hydrothermal aging was much greater than that caused by hydrothermal aging. The combined effect signicantly accelerated the decrease of breakdown eld strength. To further explore the combined aging effect, the microscopic analysis was studied below.

Physicochemical properties and mechanical properties analysis
3.3.1 SEM and EDAX analysis. Fig. 7 and 8 present the SEM photos of HTV SR specimens before and aer electricalhydrothermal aging at varying electrical eld intensity and varying temperatures for 24 days. It showed that the surface of the virgin HTV SR was smooth, on which there were no holes, cracks and obvious particles. Under the hydrothermal aging condition of 80 C, the surface of HTV SR samples became rough and appeared a few holes. However, cracks appeared on HTV SR with the addition of electric eld. When the electric eld strength reached 25 kV cm À1 , derived cracks developed on the surface of HTV SR in different orientation. It reected that water tended to diffuse into the HTV SR at 80 C, making it easier for electrons to inject into the material and accelerating the chain scission and the degradation process. Due to the irregularity of the self-diffusion of water, cracks appeared and    To examine the cross-sections of the HTV SR, SEM photos of specimens aging in different conditions to verify the depth of the damage was presented in Fig. 9. It indicated that the crosssections of the virgin HTV SR was relatively smooth and presented little cracks, holes and particles compared to the aging samples. Then the cross-section of the HTV SR specimen became rough aer hydrothermal aging. With the addition of electric eld, a large number of microscopic cracks and holes appeared in the HTV SR specimen in Fig. 9(c). It can be concluded that the electrical-hydrothermal aging degradation occurred not only on the surface but also in the interior. However, the degradation on the surface showed much more serious than the interior. It indicated that the initial deterioration process began on the surface of the samples.
At the same time, EDAX was conducted to recognize the elements present over the HTV SR surface. Fig. 10 presents the EDAX of the virgin HTV SR (spectrum 1) and the electricalhydrothermal aged (80 C -25 kV cm À1 ) HTV SR (spectrum 2). Because the test mainly studied the variation rules of surface elements on HTV SR aer electrical-hydrothermal aging, various trace elements in the ller were neglected. As shown in Fig. 10, silicon (Si), carbon (C), oxygen (O) and aluminum (Al) were the main elements on the surface of the HTV SR specimen. Aer electrical-hydrothermal aging (80 C -25 kV cm À1 ) for 24 days, the content of elements C and Al decreased signicantly. It represented the side chain scission of PDMS molecules, and the LMW substances would be extracted, resulting in a decline of C content. Moreover, Al content was also signicantly reduced due to acidic substances generated by discharge. The generated acidic substances would react with the ame retardant aluminum hydroxide. The analysis could be further proved in the FTIR below.
3.3.2 FTIR analysis. Aer hydrothermal aging and electrical-hydrothermal aging for 24 days, the IR absorbance spectra for the fresh and aer aged specimens were obtained ( Fig. 11 and 12). Table 3 lists the characteristic IR absorption bands in HTV. Fig. 11 shows the FTIR of HTV SR before and aer electrical-hydrothermal aging at 80 C for 24 days. As shown in Fig. 11, with the increase of electric eld intensity, the   peak of the main chain Si-O-Si, side chain Si-CH 3 and Si-(CH 3 ) 2 of PDMS in HTV SR specimen showed a downward trend. When the electric eld intensity reached 25 kV cm À1 , the absorption peak of the main chain and each side chain showed a remarkable reduction. For the hydroxyl absorption peak in the wavenumber range from 3200 cm À1 to 3700 cm À1 , it uctuated with the increase of eld intensity. It was caused by the complex reaction processes such as oxidation, 14 cracking, 23 crosslinking 24,25 and condensation 26 that occurred in the electricalhydrothermal aging process of HTV SR samples, resulting in the uctuating tendency of hydroxyl peak. Fig. 12 shows the FTIR of HTV SR before and aer electrical-hydrothermal aging at 25 kV cm À1 for 24 days. Similar to the downward trend in Fig. 11, when the electric eld intensity was 25 kV cm À1 , the peak of each organic group decreased signicantly with the increase of the temperature in electrical-hydrothermal aging processes. Therefore, it indicated that the PDMS on the HTV SR surface severely degraded when the electric eld intensity or the temperature increased.
3.3.3 Tensile properties analysis. The mechanical property is an important index to evaluate the aging characteristics of insulating materials. Fig. 13 and 14 present the change of the tensile strength, the elongation at break and the Young's modulus in different aging conditions aer hydrothermal and electrical-hydrothermal aging. As shown in Fig. 13 and 14, the tensile strength and the elongation at break decreased and the Young's modulus increased with the increase of the electric eld strength. In the presence of an electric eld, the change of the tensile strength, the elongation at break and the Young's modulus of the aging HTV SR specimens were much greater than that of the hydrothermal aging samples. The deterioration in tensile properties aer aging is normally related to linkages scission and main chain scission. According to FTIR results, the peak of the main chain Si-O-Si, side chain Si-CH 3 and Si-(CH 3 ) 2 of PDMS in HTV SR specimen showed a downward trend. The crosslinking structure was destroyed aer electrical-hydrothermal aging. However, the change of tensile strength and Young's modulus at varying temperature presented different trends in electrical-hydrothermal aged specimens. At 30 C -15 kV cm À1 and 60 C -15 kV cm À1 , the tensile strength and the Young's modulus did not show obvious change. When temperature reached 80 C, the tensile strength signicantly decreased and the Young's modulus signicantly increased. It revealed that when the temperature did not reach the threshold, the temperature would not promote the electrical aging process.

Frequency response analysis
In this section, the data of 80 C and 15 kV cm À1 aer electricalhydrothermal aging for 24 days were selected for analysis. Fig. 15(a) and (b) shows the change of dielectric constant and the dielectric loss of HTV SR at varying temperature aer electrical-hydrothermal aging at 15 kV cm À1 for 24 days.
As shown in Fig. 15(a), it indicated that the dielectric constant of the aging samples in the whole frequency range  signicantly increased with the rise of temperature, especially at low frequency. However, in the high frequency band (10 4 Hz to 10 5 Hz), the curve tended to a constant, which might be caused by electron polarization or molecular chain vibration of HTV SR media under the inuence of the electric eld. Fig. 15(b) shows that the dielectric loss of HTV SR specimen increased with the rise of temperature in the middle and high frequency band, while in the low frequency band, the dielectric loss of the electrical-hydrothermal aging samples was lower than that of the initial samples.
This phenomenon can be attributed to the following aspects. First of all, according to Debye relaxation theory, the dielectric loss of HTV SR samples at low frequency band is mainly caused by the DC conductance loss. With the increase of frequency, the polarization loss of different relaxations tends to play an important role and gradually becomes dominant. Secondly, in literature, 27 the dielectric properties of silicone rubber sleeve were measured by the non-contact dielectric spectrum, and the various relaxation processes of silicone rubber were analyzed in detail by using the modied Cole-Cole model. The polarization processes were divided into macroscopic interface polarization, microscopic interface polarization and dipole polarization. Considering that the contact dielectric spectrum was selected in this test, the macroscopic interface polarization process can be neglected. Therefore, the dielectric loss of silicone rubber samples was mainly caused by the microscopic interface polarization and dipole polarization. According to the SEM analysis, the polarization loss at the microscopic interface enlarged with the increase of physical and chemical defects such as holes and cracks on the surface of HTV SR and the loss of packing. Combined with the infrared spectrum characteristics, it can be seen that the rise of temperature aggravated the damage degree of the molecular main chain and decomposed to produce more LMW substances, leading to the increase of dipoles and dipole polarization loss. Fig. 16(a) and (b) shows the change of dielectric constant and the dielectric loss of HTV SR under different electric eld intensity aer electrical-hydrothermal aging at 80 C for 24 days. As seen in Fig. 16(a), the dielectric constant behave similarly for the HTV SR of the aged and the virgin sample, and increased with decreasing frequency. Especially in the low frequency band, the increase became faster because of the space charge polarization or the effect of interfacial polarization. Due to the appearance of various physical and chemical defects on the surface and inside of HTV SR, as well as the increase of particles shown by SEM results, the interface polarization was signicantly enhanced. The highest observed values (dielectric constant) were for the samples aer electricalhydrothermal aging at 80 C -25 kV cm À1 . As compared to the  virgin sample, there was an obvious increase in dielectric constant almost along the entire frequency range for hydrothermal and electrical-hydrothermal aged HTV SR specimen. This might be due to the increase in mobility of the polymer chains by the breakage of bonds caused by applied electric and hydrothermal stress.
As regards the dielectric loss angle tangent, in Fig. 16(b), it presented different and interesting phenomenon. In the middle frequency region (10 Hz to 10 3 Hz), there was an obvious dielectric loss peak in the HTV SR specimen aer electricalhydrothermal aging at 80 C -25 kV cm À1 for 24 days. It might be caused by the orientation of greater dipolar groups. A large amount of broken PDMS molecular chains led to the increase of free radicals in the material and the enhancement of dipole polarization, presenting the apparent loss peak in the dielectric frequency response. However, in the low frequency region, the dielectric loss angle tangent of the electricalhydrothermal aging samples showed different change rules. Because there was another dielectric relaxation process in the dielectric loss curve at low frequency band in addition to the DC conductivity, 27 and they worked together in the low frequency region of the dielectric loss curve.

Trap properties analysis
It is commonly acknowledged that the aging performance of the polymer material is related to the accumulation of space charge or the trap characteristics. Fig. 17 shows the TSC curve of unaged, the hydrothermal aged and electrical-hydrothermal aged samples. In order to analyze trap properties, the amount of trap charge can be calculated by eqn (5). Then the amount of trap charge, the peak current and the peak current temperature can be listed in Table 3.
As shown in Fig. 17, it indicated that the TSC curve of all HTV SR samples presented a main current peak. With the increase of the temperature or the electric eld strength of electricalhydrothermal aging, the peak current increased continuously and tended to dri towards the high temperature region. As listed in Table 4, the trap charge of the electrical-hydrothermal aging samples was signicantly more than that of the unaged and aer hydrothermal aged samples. Similarly, the trap charge also increased with the rising temperature and electric eld Fig. 16 The dielectric constant (a) and the dielectric loss angle tangent (b) of HTV SR before and after electrical-hydrothermal aging at 80 C for 24 days (varying electric field intensity). strength. Because there were more physical and chemical defects generated in the HTV SR specimen, and the deterioration degree was deepened, which led to the increase of trap and trap charge.
However, it is almost difficult to obtain the trap density and trap energy level distribution of the HTV SR specimen directly before and aer aging from Fig. 17. According to the TSC theory, the TSC caused by the de-trapping process of the captured charges can be expressed as the following function: 28 where N t (E) is the distribution function, which is spatially uniform deep into the sample with a distance l. e is electronic charge quantity, f 0 is the initial occupancy of a trap level and is a constant, E is the trap energy (trap depth). e n (E,T) ¼ v exp(ÀE t / kT) is the emission rate of electrons at trap level E and temperature T. k is the Boltzmann constant, d is the sample thickness and b is the heating rate. v is commonly called the frequency factor or attempt-to-escape frequency, typically 10 12 s À1 as suggested by Mott. 29 E v and E c are the valence band top energy and conduction band bottom energy, respectively. Then, the trap level distribution from TSC measurement can be calculated by eqn (6) based on MATLAB numerical calculation. Fig. 18 shows the trap level distributions of the virgin and aging samples. It shows that the trap density of the virgin was relatively low. With the rise of temperature and electric eld intensity, the trap density increased and the trap energy level corresponding to the peak of trap density became deeper. The peak density of traps could reach about 2.5 times larger than that of the virgin sample under the condition of 80 C -25 kV cm À1 electrical-hydrothermal aging. The phenomenon can be explained by the following aspects. Firstly, with the combination of electrical-hydrothermal aging, more defects were produced inside the HTV SR, which enhanced the electron trapping ability and increased the trap density. Secondly, it is well known that traps can be divided into deep traps and shallow traps according to the trap level. The current peak at low temperature is mainly contributed by the charge decapitation in shallow traps, while the current peak at high temperature corresponds to the charge decapitation in deep traps. As shown in Fig. 18, compared to the virgin sample, more deep traps were generated in the process of electrical-hydrothermal aging. Deep traps bound carrier migration, leading to local charge accumulation and rising local electric eld strength. When the threshold was exceeded, the polymer breakdown might occur nally.

Discussion
According to the trap theory of solid dielectric, the energy band of insulators can be divided into the conduction band, forbidden band and valence band. The charges injected by the electrode can pass through the potential barrier into the conduction band through the Schottky effect and be captured by the trap in the forbidden band. 30 The relaxation time of captured charges in traps is mainly determined by trap energy level and trap energy density. The larger trap energy level and density, the longer the relaxation time would be. Electrons captured by traps must overcome the larger trap barrier to be released. Due to the increase of trap energy level and trap density aer electrical-hydrothermal aging, it is easier to form charge accumulation. When charges accumulate to a certain critical value, the breakdown happens.  Combined with the test results above, the mechanism of electrical-hydrothermal aging can be summarized in Fig. 19. First of all, when the HTV SR specimen is subjected to hydrothermal aging effects, the main and side chains scission on PDMS molecular chain occurred, producing free radical and LMW substances. With the aging process going on, more physical and chemical defects are formed in the material. The oxidation decomposition reaction in the aging process promotes the formation of more free molecules and traps, leading to the increase of dielectric constant, loss tangent tan d, trap energy level and trap density.
Then, when the electric eld is added, electrons would be injected into the HTV SR from the electrode, and some electrons are trapped by the polymer. The trapped electron becomes the recombination center of the hole, releasing large amounts of energy and transferring it to other electrons to produce thermions. The hydrothermal aging condition provides convenience for electron injection. More thermions will bombard the molecular chain, creating more traps. On the other hand, the dipole polarization loss generates thermal power, which aggravates the deterioration of insulating materials. Therefore, according to the trap properties during the aging process, producing nanocomposites can be an available way. The addition of nanoparticles can reduce the number of traps in the material and enhance breakdown strength. By adjusting the charge transport characteristics, the breakdown probability can be lowered down.

Conclusions
This study focuses on the electrical-hydrothermal aging experiment of HTV SR based on the high temperature and high humidity environment all over the world. Some important conclusions can be summarized as follows: (1) Compared with hydrothermal aging samples, the relative weight, tensile strength, elongation at break and breakdown eld strength signicantly decreased aer electricalhydrothermal aging. The surface and interior of HTV SR were severely degraded, which was reected in the scission of a large number of PDMS molecular chains, and the loss of ame retardant Al(OH) 3 .
(2) The dielectric constant and dielectric loss of HTV SR samples signicantly increased with the rise of the aging temperature and electric eld strength. When aged at 80 C and 25 kV cm À1 , there was an obvious dipole polarization loss peak that occurred due to the breaking of molecular chains and the increase of dipoles.
(3) The electrical-hydrothermal aging has a synergistic effect. The presence of temperature and water facilitates the electron injection and deepens the trap energy level and trap density of HTV SR, which in turn increases the probability of electron entrapment, accelerating the deterioration process of HTV SR.

Conflicts of interest
There are no conicts to declare.