Electrical–Thermal Aging Performance of PAH-Modified Interfacial Coating Agent for HVDC Cable Accessory

Abstract

A novel interfacial coating agent was developed by modifying silicone oil with polycyclic aromatic hydrocarbons (PAHs) to enhance the insulation performance of HVDC cable accessories. This study investigates the effects of corona and hot–cold cycle aging on the DC breakdown characteristics of the Cross-Linked Poly Ethylene and Ethylene Propylene Diene Monomer (XLPE/EPDM) interface. Interfacial breakdown tests, infrared spectroscopy, and a microstructural analysis were employed to investigate aging mechanisms. The results show that PAH-modified silicone oil significantly increases the breakdown voltage, with 2,4-dihydroxybenzophenone (C₁₃H₁₀O₃) identified as the optimal additive via quantum chemical calculations (QCCs). Even after aging, the modified interface maintains its superior performance, confirming the long-term reliability of the coating.

1. Introduction

China has abundant offshore wind energy, which is a key pillar for achieving the country’s energy transition and advancing the “dual carbon” strategy. The optimal method for transmitting large-scale, reliable power from renewable sources is flexible high-voltage direct current (HVDC) transmission. Among these, the subsea cable terminal, which connects subsea cables to onshore equipment, is a critical component. Due to operational conditions and differences in insulation material performance, the terminal interface is prone to breakdowns, limiting large-scale transmission development. Improving the insulation of these HVDC cables directly increases the feasibility of new large-scale projects, as well as reducing the maintenance costs of existing projects. The main challenges faced by the existing technology can be summed down to the insulation layer of the cable terminal ends degrading with time. Factors such as extreme temperatures (both high and low) and extremely high pressures, which the cables are subjected underground, cause accelerated mechanical wear. Chemical interactions between the multiple layers of materials used to construct these cable terminals interact with each other, causing each layer to provide suboptimal insulation [1,2]. Although these effects are important, the most important factor is the constant high voltage (100 kV~800 kV) that these cables are handling. This causes a multitude of problems, such as space charge accumulation, which in turn causes partial discharges and electrical treeing. It is important to study the effect of high electric fields on these insulation layers in order to improve them [2,3,4].

In modern HVDC cable accessory insulation, a go-to solution is using Cross-Linked Poly Ethylene (XLPE) and Ethylene Propylene Diene Monomer (EPDM). These materials are favored for their excellent electrical and thermal resilience, although they do have their limits. Silicone oil has shown promising results when used as an insulation layer between an XLPE and EPDM interface [5,6]. Despite this, the performance of these systems eventually degrades with time, mainly due to long-term thermal and electrical stress [4,7,8,9,10].

The use of PAH small molecules as an additive to silicone oil to improve its electrothermal properties has also proven itself as an effective method. Studies have shown that doping with aromatic hydrocarbons at different mass ratios can increase the electrical breakdown properties of silicone oil at the oil-coated interfaces. Among these, 2,4-dihydroxybenzophenone has shown potential due to its molecular and thermal stability. However, the long-term behavior of such modified silicone oils under corona and hot–cold cycle aging has not been thoroughly examined. This is a considerable achievement when considering the scope of the industry and all possible applications. Since the existing research in this area is severely lacking, further testing on the limits of 2,4-dihydroxybenzopheone-doped silicone oil interfaces is necessary [11,12].

This study aims to evaluate the impact of corona discharge and hot and cold cycle aging on the microstructure, electrical breakdown characteristics, and molecular integrity of silicone oil-coated XLPE/EPDM interfaces. Both unmodified and 3% 2,4-dihydroxybenzophenone-doped silicone oils are investigated through DC breakdown testing, infrared spectroscopy, quantum chemical calculations, and microstructural changes at the oil-coated interface after both corona aging and hot–cold cyclic aging. This research gives insight into the effects of aging on the silicone-coated interface, the mechanisms involved, and the performance difference between standard silicone oil and silicone oil doped with 2,4-dihydroxybenzophenone.

1.1. Sample Preparation

1.1.1. XLPE Sample Preparation

The custom mold was cleaned with anhydrous ethanol and placed on PET film. The mold was covered with another layer of PET film. The plate vulcanizing machine was preheated to 110 °C for 5 min. The mold was slowly transferred into the machine, the upper plate was adjusted to lightly contact the mold (0 MPa pressure), and preheating was conducted for 2 min. After the particles were melted, the temperature was increased to 180 °C, and a pressure of 20 MPa was applied for 15 min. The mold was allowed to cool naturally under pressure to room temperature. The specimen was then removed and placed in a vacuum oven at 80 °C for 8 h to eliminate cross-linking byproducts. The final specimens were prepared as circular discs (300 μm thick, 100 mm diameter) for space charge measurements and square sheets (40 mm × 40 mm × 0.3 mm) for interface breakdown testing.

1.1.2. EPDM Sample Preparation

The roller distance was set to 1 mm and the temperature was set to 90 °C. EPDM particles were preheated in the roller mixer for 5 min until fully melted. DCP was gradually added to the mixer and blended for 5 min as a vulcanizing agent. The plate vulcanizer was preheated to 180 °C. The cleaned mold was placed between two PET films, the compound was added, and it was transferred to the vulcanizer. A pressure of 10 MPa was applied at 180 °C for 5 min. The mold was cooled under pressure to room temperature before demolding. The specimens were prepared as circular discs (300 μm thick, 100 mm diameter) and square sheets (40 mm × 40 mm × 0.3 mm). Figure 1 shows the appearance of the EPDM and XLPE samples.

1.1.3. Preparation of Coating Modifying Agent

The coating agent (silicone oil) was modified via physical blending with a polycyclic aromatic hydrocarbon (PAH) compound. To ensure uniform dispersion and prevent agglomeration, 100 g of silicone oil was weighed in a beaker. The required amount of PAH, based on the desired mass percentage of the compound, was added, and the mixture was stirred magnetically for 40 min. The beaker was sealed with plastic wrap and sonicated for 10 min using an ultrasonic cleaner. The mixture was degassed in a vacuum chamber (10⁻⁴ Pa, 15 min) to remove air bubbles. The beaker was then sealed to prevent contamination. Figure 2 presents a simple flowchart of the coating agent modification process.

The commercial XLPE particles used in this study were produced by Borealis (LS4201EHV), including a wire and cable extrusion-grade material with a density of 1.08 g/cm³, a melt flow index of 20–30 g/min, and a molecular weight of 28.0532 g/mol. EPDM particles were sourced from Mitsui Chemical Co., Ltd. in Tokyo, Japan. (4045M), with a density of 1.85 g/cm³ and a molecular weight of 190.32448 g/mol. The vulcanizing agent was dicumyl peroxide (DCP, purity 98%) produced by Nanjing Xinlian Chemical Co., Ltd. in Nanjing, China. The interface coating agent was dimethyl silicone oil ([-Si(CH₃)₂O-]_n, density: 0.963 g/mL) purchased from Shanghai Aladdin Reagent Co., Ltd. in Shanghai, China.

1.2. Experiment Procedures

1.2.1. Corona Aging Treatment

A needle and plate electrode pair was attached to the XLPE sample, and silicone oil was carefully applied over the XLPE sample carefully. The sample was placed on a horizontal surface and left undisturbed to settle evenly. To maintain consistent oil thickness, the initial mass of the XLPE sample (M₀) was measured using an electronic balance. After dripping the silicone oil, the mass (M₁) was recorded again. For a 0.3 mm thick sample, the mass difference (M₁ − M₀) was controlled to approximately 0.2 g. The mass difference (M₁ − M₀ ≈ 0.2 g) was selected based on pre-calibration trials to achieve a consistent silicone oil layer thickness of approximately 0.3 mm across the 40 mm × 40 mm interface area. This mass-to-thickness mapping was determined assuming uniform density and surface distribution. Thickness uniformity was ensured by using a leveled deposition surface and visually inspecting for pooling or gaps. Although precise thickness measurements were not taken for each sample, samples with noticeable inconsistencies in surface coverage were discarded.

The needle electrode was used for corona discharge. The experimental apparatus can be seen in Figure 3. A maximum potential difference of 20 kV was applied to the two electrodes with a 5 mm gap in between the samples. The sample aging duration was 96 h. Ambient temperature during the aging process was 25 ± 2 °C, with relative humidity maintained at ~60%.

1.2.2. Hot and Cold Cycle Aging Treatment

The samples were placed in a custom-made high–low temperature test chamber shown in Figure 4. This is essential in order to maintain pressure on the interface while changing the temperature in a way that does not squeeze out the silicone oil in the interface. The temperature was alternated between a maximum of 90 °C and a minimum temperature of 0 °C, with each temperature interval maintained for 30 min. A total of 30 thermal cycles were performed. The goal of applying pressure during thermal cycling was to ensure consistent contact at the XLPE/EPDM interface while preventing excessive pressure that could displace the silicone oil layer. A constant pressure of approximately 0.05 MPa was maintained across all samples using the threaded screw mechanism shown in the apparatus, with real-time monitoring from a pressure sensor. Although the number of thermal cycles (30) was selected due to experimental time constraints, it is sufficient to represent a short-term accelerated aging condition and allows for meaningful comparative analysis between unmodified and modified silicone oils.

1.2.3. FT-IR Analysis

Fourier Transform Infrared Spectroscopy (FTIR) is a technique that is used to analyze the chemical composition of materials. The method employs a Michelson interferometer to split light from a source, create an optical path difference, and generate interference patterns. The resulting data, encompassing all frequencies and intensities of the light source, undergoes a Fourier transformation via computer processing. This produces a frequency-dependent intensity spectrum. This enables us to identify the chemical composition by comparing the vibration frequency ranges by comparing it with known data [13,14].

To conduct this test, a Thermo Fisher Nicolet Nexus 470 FTIR spectrometer was used in attenuated total reflectance (ATR) mode. Measurement wavenumber range was set to 400–4000 cm⁻¹, with a resolution of 4 cm⁻¹. Since we were conducting the testing on liquid samples, a dry environment was maintained, a droplet of the sample was placed on a crystal surface, and the infrared spectra were obtained. The data was processed using the Origin 2019b software. To confirm the molecular structure of the selected interfacial coating agent, its IR spectrum is shown in Figure 5.

Dimethicones, also known as polydimethyloxanes, contain multiple silicone-oxygen (Si-O) units in their molecules as well as methyl (CH₃) clusters on the terminal and side groups. The infrared spectrogram of dimethicone oil has several significant vibrational absorption peaks, as shown in

Table 1. Dimethyl silicone oil infrared absorption peaks and corresponding functional groups.

Assignment	Wave Number (cm−1)
Si(CH3)3	700
Si(CH3)2	780–840
C-H in Si(CH3)3	2960
Si-CH3	1240–1280
Si-O-Si	1000–1130

[15]. By comparing the features of the IR spectra of the coating agent and dimethicone shown in Figure 5, it can be seen that the major vibrational absorption peaks of the selected coating agent coincide with those of dimethicone. This result verifies that the coating agent used is a high-purity dimethylsiloxane oil with no impurity components, further confirming the accuracy of its chemical structure and the singularity of its composition.

1.2.4. DC Breakdown Testing

To closely simulate real-world HVDC submarine cable termination conditions (temperature: 10–90 °C, pressure: 0.05–0.3 MPa), two test setups were designed.

DC Breakdown Voltage Measurement: A needle–plate electrode configuration was used. Copper foil (50 μm thick) was cut and attached to the XLPE surface. The needle electrode (30° angle at the needle tip) and plate electrode were spaced 1 mm apart. Pressure and temperature were controlled via sensors and a thermal regulator. Voltage was applied incrementally at a ramp rate of 0.5 k V/s until breakdown occurred.

Pre-stressed DC Breakdown: A vertical DC pre-stress (15 kV/mm, 20 min) was applied to simulate charge accumulation. Post–pre-stress breakdown testing followed the same protocol. Figure 6 shows the experimental setup used for the DC breakdown voltage testing; on the top, we can also see a representation of the cross-sectional view of the EPDM/XLPE silicone oil-coated sandwich. A mechanical pressure press was used in conjunction with a pressure sensor to vary the pressure. A flat heating pad, as shown in the cross-sectional representation, was used to vary the temperature. Tests were repeated 10 times for each condition combination, and an average result was obtained. In all DC breakdown voltage measurements, the needle electrode was connected to the negative terminal of the DC power supply, and the plate electrode was grounded. Therefore, breakdown voltages are reported as negative values (e.g., −20.28 kV) to reflect the polarity of the applied electric field with respect to the ground electrode.

1.2.5. Microstructure Analysis

A polarizing microscope was used to observe the microstructural changes in the samples before and after aging. The experimental setup is shown in Figure 7. The samples were placed on a transparent acrylic plate, illuminated with light, and observed through a computer-integrated S-EYE video system, which enabled us to record the findings.

1.2.6. Quantum Chemical Computing

Quantum chemical calculations (QCCs) offer a powerful theoretical framework for analyzing the electronic structure and reactivity of molecules. In the context of dielectric materials, understanding how small molecular additives (such as PAHs) interact with high-energy charge carriers is essential for predicting their insulation performance under high-voltage stress. The ability of a molecule to trap high-energy electrons, as well as its ionization potential and electron affinity, are key indicators of its effectiveness in suppressing breakdown phenomena. Therefore, QCC was selected in this study to model and compare the frontier molecular orbitals (HOMO/LUMO levels) of candidate PAH modifiers and to assess the impact of aging-induced structural changes on their charge trapping capabilities.

With the significant increase in computational power, QCC has risen in popularity as a solution to the electronic structures of molecules, energy levels, and simulating properties of atoms and molecules. Phenomena such as polymer aging, breakdown, and electrical treeing are closely related to the movement and transfer of high-energy charge carriers. Quantum chemical calculations not only provide deep insights into the behavior of these carriers but also reveal their mechanistic roles in material performance evolution. Software such as VASP, Gaussian, and Q-Chem has proven itself as reliable software for QCC. In this experiment, we have used the Density Function Theory (DFT) proposed by Kohn and Hohenberg as the computational basis method for our platform [16,17].

In this simulation, we used Gaussian 09W as the computational platform and B3LYP functional with the 6-31G(d) basis set, in order to simulate the HOMO-LUMO energy levels of both aged and non-aged silicone oil. Analyzing the energy levels would give us a clear view of the underlying phenomenon.

2. Results

2.1. Effect of Corona Aging on Insulation Performance

Microstructure Observation

The sample was subjected to corona aging by exposing it to an electric field that was strong enough to ionize the surrounding air. A microscope was used to observe the surface of the specimens before and after corona aging, as shown in Figure 8. The results indicate that after being coated with silicone oil, the surface of the XLPE sample was smooth with no visible mechanical scratches or impurities. However, after 96 h of corona aging, the silicone oil layer dried out, and a few fine cracks appeared on the surface. Additionally, several circular pits were observed in the needle electrode region. This can be attributed to the high temperature and ozone generated during the corona discharge process.

Before corona aging, the surface of the modified silicone oil-coated specimens exhibited uniformly distributed 2,4-dihydroxybenzophenone crystals, which were embedded in the silicone oil matrix, forming a relatively stable structure. However, after corona aging, circular pits also appeared on the specimen’s surface, and some doped crystals precipitated. This can be observed in Figure 9.

2.2. DC Breakdown Characteristics

The DC breakdown voltage after being subjected to corona aging was tested. This showed that the trends with varying temperature and pressure are the same as before being subjected to corona aging. The breakdown voltage decreases with increasing temperatures and increases with increasing pressures. The DC breakdown voltages are given in Figure 10a,b.

It was observed that after subjecting the samples to corona aging, the breakdown voltage generally decreased. With the unmodified silicone oil under the conditions of 90 °C and 0.05 MPa, the breakdown voltage dropped from −20.28 kV to −9.87 kV, representing a 51.55% reduction. The modified silicone oil also displayed similar behaviors, where its breakdown voltage decreased from −29.73 kV to −12.35 kV, representing a 58.46% reduction.

2.3. Aging Mechanism Analysis

Following this, an infrared spectroscopy analysis was conducted with the corona-aged samples. Figure 11 shows the infrared spectra of unmodified silicone oil before and after corona aging.

After comparing the values in Figure 11 with Table 1, along with the dimethyl silicone oil infrared absorption peaks and corresponding functional groups presented in Section 1.2.3 on FT-IR analysis, we can identify the functional groups present in the unmodified silicone oil. By comparing the two spectra “before aging” and “after aging” in Figure 11, we can observe whether any functional groups have been oxidized into other functional groups. The peaks at 2962.29 cm⁻¹ and 2962.61 cm⁻¹ correspond to the absorption vibration peaks of the saturated C-H bonds in -CH₃, while 1257.40 cm¹ and 1257.61 cm⁻¹ correspond to the vibration absorption peaks of the Si-CH₃ bond. The peaks at 1007.74 cm⁻¹ and 1007.48 cm⁻¹ correspond to the vibration absorption peaks of the Si-O bond in Si-O-Si. The peaks at 784.97 cm⁻¹ and 784.79 cm⁻¹ correspond to the vibration absorption peaks of the Si(CH₃)₂ group. The peaks at 700.45 cm⁻¹ and 696.07 cm⁻¹ correspond to the vibration absorption peaks of the Si(CH₃)₃ group. After the process of corona aging, although no new peaks appeared in the spectra, the intensity levels of the major absorption peaks were significantly decreased.

The vibration absorption peak of C-H bonds decreased from 0.155 to 0.095. The vibration absorption peak of Si-CH₃ groups decreased from 0.552 to 0.451. The vibration absorption peak of Si-O bonds decreased from 1.182 to 1.106. The vibration absorption peak of Si(CH₃)₂ groups decreased from 1.625 to 1.514. The vibration absorption peak of Si(CH₃)₃ groups decreased from 0.223 to 0.169.

These results indicate that during the corona aging process, the aforementioned molecular chains and their functional groups and bond strengths have undergone a significant amount of degradation. However, no additional vibration absorption peaks were observed after the corona aging process, which also shows that the process has not formed or introduced new functional groups. Such degradation will have a notable impact on the insulation performance and cause a decrease in the breakdown voltage.

Figure 12 shows the infrared spectra of silicone oil doped with 3% 2,4-dihydroxybenzophenone by weight. A similar comparison can be done here with Table 1. The aforementioned molecular chains and their functional groups, and their bond strengths, have undergone a significant amount of degradation. From the figure, no new vibrational absorption peaks were observed after the corona aging process, which also indicates that no new functional groups were formed or introduced during the process. However, a degradation reaction was produced, resulting in some degree of structural changes in both the silicone oil molecule and the 2,4-dihydroxybenzophenone small molecule, which may lead to a decrease in the breakdown voltage. However, it can be observed that the IR spectra showed new vibrational absorption peaks in the range of 3100–3350 cm⁻¹, indicating the introduction of hydroxyl (-OH) groups. The peak of this group showed a different trend from the other peaks after corona aging. This is due to the aging of the -H functional groups on the benzene ring, which have been oxidized to hydroxyl (-OH) groups, causing an incremental change in absorbance from 0.320 to 0.711. No new vibrational groups were detected here, which confirms that no new functional groups have formed during aging. Therefore, corona aging leads to a decrease in the interfacial breakdown voltage of coated modified silicone oils in two ways:

• The molecular chains of modified silicone oil undergo decomposition due to the impact of electrons from the needle electrode during corona aging.
• New hydroxyl groups that form on the benzene rings are altering the chemical properties of the composition.

Figure 12. Infrared spectra of 3% modified silicone oil before and after corona aging.

Figure 12. Infrared spectra of 3% modified silicone oil before and after corona aging.

In order to further investigate whether the newly formed hydroxyl group (-OH) would have an effect on the properties of modified silicone oils, we analyzed 2,4-dihydroxybenzophenone before and after corona aging using quantum chemical calculations. The calculation results are shown in Figure 13.

According to Figure 13, which shows the electron energy level distribution of the C₁₃H₁₀O₃ molecules before and after aging, the HOMO and LUMO levels of the aged C₁₃H₁₀O₃ molecule are, respectively, −6.04 eV and −1.05 eV. The aged C₁₃H₁₀O₃ small molecule has an electron affinity energy of 1.05 eV and an ionization energy of 6.04 eV, which gives it a lower electron affinity energy and a wider energy gap compared to the original molecule. These results suggest a weakened ability of the aged molecule to trap high-energy electrons; therefore, when electrons enter the interface, the result is an increased probability of collisions between the high-energy electrons and the silicone oil molecule. At the same time, the increase in the energy gap means that the C₁₃H₁₀O₃ molecules in the excited state are difficult to excite, which means that the energy absorbed by the high-energy electrons cannot be safely released, and the captured high-energy electrons cannot be recycled again. The available C₁₃H₁₀O₃ molecules are reduced and their efficiency is lowered, and the high-energy electrons remaining increase, which makes it easier for them to collide with the silicone oil molecules. This aggravates the damage to the chain of the silicone oil molecules, and ultimately leads to the occurrence of breakdowns. It is worth noting that, although the LUMO energy level of the C₁₃H₁₀O₃ molecules increases and the HOMO energy level decreases after corona aging, the interfacial breakdown voltage is still higher than that of the unmodified silicone oil-coated interface, and there is still the ability to improve the interfacial breakdown voltage.

2.4. Effect of Thermal Cycling Aging on Insulation Performance

Microstructure Observation

Figure 14, presented below, shows the microscopic image of the interface of the specimens coated with unmodified silicone oil, both before and after hot–cold cycles. When the samples were subjected to hot–cold cycle aging, in the vicinity of the needle electrode, it was evident that while the EPDM/XLPE interface coated with undoped silicone oil pre-hot–cold cycle aging contained none to very few air bubbles, impurities could be seen. When it came to the aged samples, there was a considerable amount of bubble formation between the needle and the plate electrode.

Figure 15, presented below, shows the microscopic image of the interface of the specimens coated with modified silicone oil (3% W/W 2,4-dihydroxybenzophenone), both before and after the hot–cold aging cycles. We could observe uniformly distributed 2,4-dihydroxybenzophenone crystals and minimal to very little impurities and air bubbles. After being subjected to hot–cold cycle aging, we observed a considerable amount of air bubble formation, where we observed bubbles of varying sizes, while the number of embedded 2,4-dihydroxybenzophenone crystals significantly decreased.

2.5. DC Breakdown Characteristics

The results of DC breakdown testing of cold and hot cycle aged samples coated with modified and unmodified silicone oil interfaces are shown in Figure 16a and Figure 16b, respectively. Despite undergoing hot and cold cyclic aging, the trends of the interfacial breakdown voltages of unmodified and modified silicone oil coatings under different temperature and pressure conditions are consistent, with both being inversely proportional to the temperature and positively proportional to the pressure. However, after hot and cold cyclic aging, the breakdown voltages of both the modified and unmodified silicone oil-coated samples decreased. It was also observed that under all the conditions, the silicone oil modified with 3% 2,4-dihydroxybenzophenone displayed the highest breakdown voltage readings.

These experimental results indicate that hot and cold cyclic aging significantly hinders the breakdown characteristics of silicone oil-coated interfaces. However, modifying the interface with 3% 2,4-dihydroxybenzophenone showed that it can mitigate this effect. It is noteworthy to mention that the breakdown voltage changes are more significant at lower interfacial pressures, while being less significant at higher interfacial pressures. This phenomenon is closely related to the more complete gas expulsion from the interface at higher pressures.

2.6. Aging Mechanism Analysis

To further understand the effects of hot and cold cyclic aging on the silicone oil-coated XLPE/EPDM interface, infrared spectroscopy was used to conduct a deep dive into the structural changes that contributed to the changes in the breakdown voltage. Infrared spectroscopy was performed on both unmodified and modified 3% 2,4-dihydroxybenzophenone, each for 30 cycles of heat and cold. The results can be seen in Figure 17 and Figure 18.

Observing Figure 17, which is presented above, showing the infrared spectra of the unmodified silicone oil before and after being subjected to hot–cold cycle aging, it was observed that the main vibrational absorption peaks exhibit slight changes after the process. The Si-O-Si vibrational absorption peak (1000–1100 cm⁻¹) decreased slightly from 1.181 to 1.165. The Si(CH₃)₂ vibrational absorption peak (780–840 cm⁻¹) decreased from 1.624 to 1.6. The vibrational peaks of C-H (around 2960 cm⁻¹), Si-CH₃ (1240–1280 cm⁻¹), and Si(CH₃)₃ (around 700 cm⁻¹) bonds remained unchanged. No new vibrational absorption peaks were observed. This indicates that the hot and cold cyclic aging had minimal effects on the degradation of the silicone-oil molecular chains.

Figure 18 contains the infrared spectra of the silicone oil modified with 3% 2,4-dihydroxybenzophenone. It can be observed that, unlike in the previous spectra in Figure 18, we can see a considerable change after the hot and cold cyclic aging process. The intensity of the main vibrational peak has decreased, which indicates molecular degradation of both the silicone oil and 2,4-dihydroxybenzophenone particles. Since no new absorption vibrational peaks were found, no new molecular structures were generated. The hydroxyl (-OH) vibrational absorption peak (3000–3300 cm⁻¹) increased from 0.291 to 0.322, indicating a slight increase in the hydroxyl group content. Although it is known from the above analysis of the corona aging mechanism that the increase in hydroxyl groups on the benzene ring may affect its quantum chemical properties and thus lead to changes in the properties of the modified silicone oils, we found that the slight increase in hydroxyl groups does not have a significant impact on the properties under hot and cold cyclic aging conditions. Despite a slight degradation, we conclude that this has a negligible influence on the DC breakdown at the interface. Thus, this is not the primary cause of the decrease in the breakdown voltage.

3. Discussion

Mechanisms of Bubble Formation and Interface Breakdown Development

From the microscopic morphology analysis of the XLPE/EPDM composite silicone-coated interface, a significant amount of bubble formation was observed after the hot and cold cyclic aging process. This phenomenon is attributed to the thermal expansion and contraction of the XLPE and EPDM. As the temperature changes, the contact between the two materials weakens, creating voids in between the two materials at the interface, and the process of silicone oil filling the interface is not perfectly efficient. The infrared spectra analysis showed that in both the modified and unmodified silicone oils, no new substances were formed, and only minor degradation was displayed. Therefore, it is inferred that the main reason for the decrease in the interfacial breakdown voltage is the formation of voids rather than chemical degradation.

The localized cavity discharge damage process and degradation mechanism analysis at the XLPE/EPDM interface are shown in Figure 19. As the breakdown field strength of air is much smaller than that of liquid and solid dielectrics, when bubbles appear in the XLPE/EPDM composite interface, the bubbles change the electric field distribution at the interface, making the electric field strength around the bubbles higher; at this time, the bubbles around the tip electrode begin to deform and move, as shown in the first stage of Figure 19. With the increase in voltage, the gas inside the bubble begins to ionize, and the partial discharge phenomenon initially occurs. Due to the self-recovering nature of air, the whole discharge briefly subsides after the partial discharge begins to occur at the bubble, as shown in the second stage of Figure 19. With the development of the discharge phenomenon and the continuous accumulation of electrons and ions within the bubble, the electric field within the bubble gradually loses stability until the bubble is filled with a large number of charged particles, and it loses of its insulating properties; at the same time, because of the concentration of the electric field near the bubble, the development of the discharge is intense, and the local temperature continues to rise, resulting in an accelerated aging process for the silicone oil and the two materials on both sides. The material’s dielectric properties are gradually reduced, and the high temperatures further promote increases and expansion of the interface cavities, allowing for electronic kinetic energy to accumulate more easily, and finally, the interface breakdown phenomenon occurs as shown in Figure 19c.

While this study investigated corona aging and hot–cold cycling independently to isolate their respective mechanisms, it is important to acknowledge that in real-world HVDC cable applications, these stress factors often occur concurrently. In such environments, the presence of thermal cycling-induced voids may exacerbate corona-induced degradation by providing localized weak points that concentrate electric fields and vice versa. In this study, the experiments were performed and the mechanisms of breakdown were studied in isolation for simplification purposes. A synergistic effect between these conditions is very likely, where these effects accelerate each other’s impact on insulation breakdown. A great direction for future studies would be to investigate these synergistic effects.

4. Conclusions

In this study, we systematically investigated the mechanism of XLPE/EPDM interface property changes due to corona aging and thermal cycling aging. Emphasis was placed on confirming the excellent insulating properties of 2,4-dihydroxybenzophenone-doped modified silicone oil under aging conditions. Quantum chemical calculations revealed that the mechanism involves 2,4-dihydroxybenzophenone molecule trapping high-energy electrons and thus inhibiting the interfacial breakdown.

Corona Aging: Both unmodified and modified silicone oils exhibited a reduced breakdown voltage (51.55% and 58.46%, respectively) due to molecular chain degradation and the oxidation of benzene rings. FTIR revealed decreased bond strengths in the Si–O, Si–CH₃, and C–H groups, while quantum chemical calculations showed a decrease in the electron affinity energy and an increase in the energy gap in the aged C₁₃H₁₀O₃, resulting in a decrease in the ability of the molecule to trap high-energy electrons.

Hot and Cold Cyclic Aging: Breakdown voltage degradation was primarily attributed to mechanical void formation at the interface rather than chemical degradation. Hot–cold thermal expansion–contraction cycles formed large air bubbles, leading to partial discharges and localized electric field distortion. Hot and cold cyclic aging emphasized the critical role of physical defects (e.g., voids) over chemical changes, highlighting the need for improved interface quality and new bubble suppression strategies. Despite aging-induced performance declines, the modified silicone oil maintained superior breakdown characteristics compared to the unmodified oil, validating its potential for enhancing HVDC cable terminal reliability under the existing harsh operational conditions. These findings confirm that 2,4-dihydroxybenzophenone-modified silicone oil not only enhances the initial interface insulation performance but also significantly mitigates aging-related degradation, making it a promising candidate for long-term reliability in HVDC cable accessories.