Study on failure mechanism of rubber sealing rings for high pressure equipment in ultra high H₂S gas fields

Abstract

In this study, the failure mechanisms of fluoroether rubber (FM) and hydrogenated nitrile butadiene rubber (HNBR) seal rings employed in two feedstock gas coalescers operating within an ultra-high H₂S gas field were investigated. Both seal rings experienced breakage and failure after 30 days of operation. To analyze the failure mechanisms under ultra-high H₂S conditions, a high-temperature and high-pressure (HTHP) dynamic autoclave, along with an O-ring sealing fixture, was employed to replicate on-site conditions. Corrosion experiments were conducted on FM and HNBR under simulated conditions for 7 days. The progression of performance and failure behavior was assessed through comparative analyses of mechanical performance, fracture morphology, Fourier transform infrared spectroscopy (FTIR), and thermogravimetric (TG) results. These analyses were carried out on samples in their initial state, after 7 days under simulated conditions, and following 30 days of field operation. Initial evaluations revealed ductile fracture characteristics for both FM and HNBR. After 7 days of simulated testing, no breakage was observed, and the thermal stability and mechanical properties of both materials remained largely unchanged. The tensile fractures maintained their ductile nature, although resistance to compression set declined, and porous defects were detected. Additionally, HNBR exhibited a slight increase in tensile strength. After 30 days of service, both FM and HNBR demonstrated brittle fracture failure accompanied by significant reductions in mechanical properties and thermal stability. The tensile strength decreased to 1.96 MPa for FM and 4.09 MPa for HNBR, while elongation at break declined to 46.2% and 59.9%, respectively. FTIR analysis indicated that the molecular chains of FM experienced breakdown and degradation, while those of HNBR underwent additional crosslinking. The deterioration in mechanical properties is primarily attributed to molecular chain degradation, crosslinking, and breakage, highlighting the impact of ultra-high H₂S exposure on these materials.

Introduction

The recent development of H₂S-rich gas fields in South China, with H₂S concentrations reaching 16%, meets the criteria for ultra-high H₂S gas fields as defined by the GBT/26,979–2011 standard (H₂S content between 10 and 50%)¹. Compared to high H₂S gas fields, which exhibit H₂S concentrations between 2 and 10%, the elevated H₂S content in these ultra-high fields poses substantial challenges to the corrosion resistance of pipeline and equipment materials^2–5. To maintain safe and continuous production in such extreme environments, ultra-high H₂S gas fields increasingly rely on advanced materials with intrinsic corrosion-resistant properties. For wellbore applications, nickel-based alloy tubing is primarily utilized, whereas nickel-based alloy bimetal composite tubing is employed in gathering and transportation pipelines to mitigate corrosion risks. However, a standardized approach to selecting effective sealing materials and methods for high-pressure equipment, such as feedstock gas coalescers and separators, has yet to be established^6–8.

Rubber seals are commonly used in oil and gas fields due to their cost-effectiveness, ease of processing, and simple disassembly and replacement processes^9–11. Fluoroelastomer and hydrogenated nitrile butadiene rubber (HNBR) are preferred materials for sealing in high-pressure equipment within ultra-high H₂S fields due to their favorable mechanical properties and resistance to aging, pressure, and corrosion^12–14. Despite these advantages, practical applications have revealed failures in fluoroether rubber (FM) and HNBR sealing rings. Specifically, two feedstock gas coalescers operating at 10 MPa and 16% H₂S concentration experienced total seal failure within 30 days, with the sealing rings showing extensive surface cracks and bubbles. These surface disruptions compromised both continuous production stability and sealing performance, increasing the risk of ultra-high H₂S gas leakage and subsequent safety hazards^15,16.

Research on rubber sealing reliability and degradation mechanisms in harsh environments has expanded significantly. For instance, Shi et al. studied the thermal oxidative aging behavior of FM 600S under different aging conditions, demonstrating that with prolonged aging time, 600S exhibits decreased heat resistance, damaged TAIC crosslink points¹⁷. Similarly, Quirinus de Ruijter et al. evaluated HNBR aging at elevated temperatures, observing that hydrogenation affects C = C double bonds and that degradation of HNBR becomes heterogeneous at temperatures of 140 °C and 170 °C¹⁸. Anton et al. examined HNBR degradation in simulated oil and gas environments, finding that chain rupture predominantly occurs in the later stages of hydrocarbon aging, with significant changes in the chemical structure, including the transformation of the C≡N group into an amide group¹⁹. However, the failure mechanisms of rubber materials in H₂S-rich environments, particularly in ultra-high H₂S conditions, are still not comprehensively understood.

A field investigation of rubber seal application in an ultra-high H₂S gas field in southwest China revealed a maximum operational lifespan of less than three months, as summarized in Table 1. To address this, in the present study, an analysis was conducted on the performance changes and failure mechanisms of FM and HNBR seals under ultra-high H₂S exposure, drawing insights from field failure incidents involving these materials. Corrosion experiments were conducted on the rubber rings under simulated conditions over a 7 day period. A comparative analysis of the mechanical properties, fracture morphology, and changes in rubber chain structure of FM and HNBR in their initial state, post-experimentation, and following 30 days of field service was performed. The failure mechanisms of FM and HNBR in ultra-high H₂S environments are examined in this study, providing technical insights for the selection of sealing materials and the development of performance optimization strategies in ultra-high sulfur H₂S fields.

Table 1.

Failure incidence statistics.

Equipment	Pressure and Temperature	Gas	Materials	Failure time	Failure behavior
Feedstock gas filter (Wet Gas)	13 MPa 37 ℃	H2S 16%vt CO2 8%vt	FKM	18	Breakage
			FM	45	Bubbling and cracking
			HNBR	82	Bubbling and cracking
Dry gas coalescer (Dry Gas)	13 MPa 37 ℃		HNBR	86	Breakage
			NBR	62	Breakage
Feedstock gas coalescers (Wet Gas)	10 MPa 32 ℃		FM	30	Breakage
			HNBR	30	Breakage

Results

Mechanical properties

Each property test was conducted with five parallel samples, and the arithmetic mean of the measured values for each property was used as the final test result. Mechanical property results are presented in Fig. 1.

Fig. 1.

Variation in (a) Hardness, (b) Density, (c) Tensile strength, (d) Elongation at break, and (e) Compression set of rubber at different service stages.

For FM, after 7 days of exposure, density remained largely stable, while hardness, tensile strength, and elongation at break decreased, and compression set increased. Following 30 days of service, both density and hardness exhibited slight declines. Specifically, compared to the initial state, tensile strength decreased from 16.1 MPa to 1.96 MPa, elongation at break reduced from 180% to 46.2%, and compression set increased from 15 to 45%.

For HNBR, after 7 days of exposure, density, hardness, and elongation at break exhibited varying degrees of reduction, tensile strength showed a slight increase, and compression set increased markedly. After 30 days of service, density further declined. Relative to the initial state, hardness increased to 88.1 HA, elongation at break decreased to 59.9%, tensile strength fell to 4.09 MPa, and compression set rose to 61%.

The data showed no significant outliers, and each group displayed an approximately normal distribution with homogeneous variances. All P-values were below 0.05, indicating a statistically significant difference in the performance of the two rubbers across states.

The test results indicate that the mechanical properties of FM deteriorate progressively over extended service time, while HNBR exhibits minimal degradation in the initial stages, followed by significant declines. After 30 days of service, both rubber types display substantial reductions in mechanical performance relative to their initial states.

According to the GB/T 3452.5–2022 standard²⁰, FM must meet an international rubber hardness (IRHD) level of 80, with a tensile strength exceeding 10 MPa, elongation at break of at least 125%, and a compression set not exceeding 25%. For HNBR, the IRHD level is 85, with a minimum tensile strength of 16 MPa, elongation at break of at least 125%, and a maximum compression set of 45%. Comparing these standards with the test results reveals that, after 7 days of exposure, both rubber types meet the required standards without compromise in integrity. However, after 30 days of service, both materials fall significantly short of these requirements, suggesting that the integrity of the rubber seals had likely been compromised prior to removal.

Morphology

Figure 2 illustrates the morphological changes in the two types of rubber after different service durations. From a macroscopic perspective, both rubber rings maintained their standard circular shape after 7 days of exposure, displaying only minor deformation. Surface characteristics remained smooth and retained a degree of luster, without signs of swelling, stickiness, or other degradation phenomena.

Fig. 2.

Morphological analysis of rubber at different service stages (5000 ×).

Micromorphological analysis, conducted using scanning electron microscopy (SEM), revealed that the surfaces of both samples were generally flat, though small micropores with an approximate diameter of 0.5 μm were present. FM samples exhibited a higher density of these micropores, approximately 2700 per square millimeter, while that of HNBR is about 1300 per square millimeter. Additionally, both samples showed long strip-shaped extrusion marks on their surfaces, likely resulting from the inherent stiffness of these polar rubbers. This stiffness impedes their full recovery to an undeformed state after compression, indicating the presence of a compression set. Notably, a lower compression set is associated with enhanced material elasticity.

After 30 days of service, substantial morphological changes were observed in both rubber types. FM exhibited extensive surface bubbling, while HNBR displayed numerous large cracks along its edges. SEM analysis further revealed that the FM surface had developed a rough and porous texture, with the density of micropores amounts to 8000 per square millimeter. Some pores ranging from 3 to 5 μm in diameter. In contrast, the HNBR surface had fewer porous defects, with the density about 4000 per square millimeter. And the pores measuring approximately 1 μm in diameter. Additionally, HNBR showed prominent, regular strip-shaped extrusion marks, indicating a further reduction in its recovery ability after 30 days of service²¹.

Fracture morphology

The O-ring exhibited characteristics of tensile fracture, while the on-site sealing ring’s fracture was identified as failure-induced.

Figure 3a presents the tensile fracture morphology of FM in its initial state. The fracture originated at the edge, with the propagation area displaying a rough texture that gradually transitioned to a smoother surface, characteristic of a typical ductile fracture pattern²².

Fig. 3.

Fracture morphology of FM at different service stages.

Figure 3b shows the fracture morphology of FM after 7 days of exposure. The fracture section is distinguishable into three regions: the fracture source, the fracture growth zone, and the fast growth zone. The fracture source was located at the sample’s edge, and the fracture growth zone had a rough texture, while the fast growth zone appeared smooth. Fluvial streaks were evident, with radiating linear streaks extending from the fracture source in the direction of fracture propagation, suggesting a reduction in FM toughness following 7 days of simulated exposure.

Figure 3c illustrates the fracture morphology of FM after 30 days of service. Multiple fracture sources were observed, and representative sections were selected to allow a comprehensive examination of the fracture characteristics. Brittle fracture features, including mirror regions, fluvial streaks, and arc rib zones, were identified in various areas of the fracture section. The final fracture zone showed a transition from smooth to rough surfaces, indicative of brittle fracture characteristics²³.

The comparative analysis of FM fracture morphology across different exposure durations suggests a progressive reduction in toughness with extended service time under ultra-high H₂S conditions. The fracture morphology observed after 30 days indicates a complete transition in FM behavior from ductile to brittle, signifying significant material embrittlement.

Figure 4 presents the fracture morphology of HNBR at different service stages. In its initial state, the fracture source for HNBR was located at the edge, with the fracture section divided into a rough slow-growth zone and a smooth fast-growth zone, characteristics typical of ductile fractures²⁴.

Fig. 4.

Fracture morphology of HNBR at different service stages.

Figure 4b illustrates the fracture morphology of HNBR after 7 days of exposure. The overall fracture topography closely resembles that shown in Fig. 3a, also exhibiting features consistent with ductile fracture.

Figure 4c shows the fracture morphology of HNBR after 30 days of service. At this stage, the entire fracture surface appeared smooth, with no secondary cracks, and displayed a distinct step-tear layer, indicative of brittle fracture behavior²⁵.

Comparative analysis of HNBR fracture morphology across these stages indicates that, under ultra-high H₂S conditions, HNBR retains good toughness during the initial service period (up to at least 7 days). However, a significant decline in toughness occurs with extended service time, completing the "ductile-to-brittle transition" between 7 and 30 days of exposure.

FTIR analysis

The FTIR spectra of FM samples at different stages are shown in Fig. 5a. Key absorption peaks were identified as follows: the peaks at 2916 cm⁻¹ and 2850 cm⁻¹ correspond to the antisymmetric and symmetric stretching vibrations of methylene (–CH₂–) groups²⁶. The peak at 1630 cm⁻¹ is attributed to the stretching vibration of C = C bonds, while a distinct peak at approximately 1387 cm⁻¹ reflects the bending vibration of CH₂ groups²⁷. An absorption peak at 1125 cm⁻¹ is associated with the characteristic C–F bond, and the peak at 870 cm⁻¹ represents the characteristic absorption of CH₂ = CF₂ bonds²⁸.The prominent peak near 1726 cm⁻¹ indicates the presence of a carboxyl group (C = O), which is observed only after 30 days of service, with a area of 6.83A·cm⁻¹.The peak at 1130 cm⁻¹ is attributed to ether linkages (C-O). The observed increase in peaks corresponding to carbonyl groups (C = O) and ether linkages (C-O) suggests oxidative processes during service, leading to the formation of carbonyl or ester groups within the FM molecular chains²⁹. Additionally, a decrease in CH₂ = CF₂ absorption intensity after service indicates possible breakage and recombination of CH₂ = CF(CF₃) bonds, suggesting an increase in branching³⁰. This behavior is likely a result of degradation processes during use, where unsaturated bond positions undergo oxidation^31–33.

Fig. 5.

FTIR of HNBR and FM at different service stages.

The FTIR spectra of HNBR samples at different stages are illustrated in Fig. 5b. The absorption peak at 2910 cm⁻¹ corresponds to the characteristic C-H absorption, with a slight increase of 21.6% in area after 7 days, indicating further cross-linking of HNBR during the initial service stage in the H₂S environment. The peak at 2120 cm⁻¹ corresponds to the nitrile (C≡N) group in HNBR, and its area gradually decreases over time. After 7 days of the experiment, it decreases by 1.6%, and after 30 days of service, it decreases by 17.2%. This suggests that the unsaturated C≡N bonds in the HNBR chains react with free radicals generated by H₂S in a humid environment, resulting in a reduction of nitrile group concentration³⁴. Peaks at 1631 cm⁻¹ and 1538 cm⁻¹ represent stretching vibrations of C = C bonds, indicating some level of unsaturation within the HNBR molecular chain both before and after service. However, after 30 days, the absorption peak area of C = C decreased by 27.3% compared to the initial state, which indicates that addition reactions occurred at unsaturated double bonds in the later stages of service.

TG analysis

The TG analysis results for FM at different service stages are presented in Fig. 6a. In both the initial state and after 7 days of exposure, FM exhibited a stable, low rate of thermal decomposition below 400 °C, with a maximum decomposition temperature of approximately 490 °C. Residual mass remained largely unchanged, at 29.3% initially and 29.2% after 7 days. However, after 30 days of service, FM showed a significantly increased thermal decomposition rate below 400 °C, resulting in a mass loss of about 11% in this temperature range. The maximum decomposition temperature decreased to around 477 °C, and residual mass decreased to 20.6%.

Fig. 6.

TG analysis results of HNBR and FM at different service stages.

The TG analysis for HNBR at different stages, shown in Fig. 6b, revealed two distinct mass-loss stages for both the initial state and after 7 days of exposure. During the first stage, between 231 °C and 416 °C, initial mass loss was approximately 10.9%, decreasing slightly to 9.5% after 7 days. In the second mass-loss stage, between 420 °C and 496 °C, the maximum decomposition rate occurred around 475 °C, with residual mass at 58.9% in the initial state and 58.4% after 7 days, indicating minimal molecular chain degradation in this period. After 30 days of service, the first mass-loss stage occurred at a lower range (220 °C to 343 °C), with an increased mass loss of 14.1%. The second mass-loss stage, occurring from 343 °C to 487 °C, showed a reduced maximum decomposition temperature of about 448 °C, with residual mass further reduced to 41.1%.

A comparison of TG results for FM and HNBR highlights several key differences. The C-F bond in FM provides superior protection to the main chain due to the high electronegativity and atomic radius of the fluorine (F) atom, contributing to the material’s high chemical inertness and stability. Consequently, FM displayed a higher decomposition temperature than HNBR.

Discussion

The aging process of rubber involves chemical reactions, such as cross-linking and molecular chain degradation, as well as physical changes, including the loss of reinforcing agents and small molecular substances. Under ultra-high H₂S conditions, the primary factor contributing to the decline in mechanical properties and eventual failure of FM and HNBR over prolonged service periods is their interaction with H₂S. Other factors, such as ambient pressure, the presence of CO₂, and the concentration gradient across the sealing rings, further accelerate the degradation process^35,36.

Microscopic and fracture morphology analysis revealed that both types of rubber develop internal porous defects and a gradual loss of toughness during service, leading to brittle fracture after 30 days. This behavior is associated with the polar groups (–C–F and –C≡N) in FM and HNBR, which alter internal energy distribution and affect the solubility of non-polar molecular chains. These changes result in physical and chemical modifications, causing additives to detach from the rubber matrix, form pores, and create additional stress-sensitive points, thereby reducing tensile properties and toughness.

TG results show that FM’s thermal decomposition rate significantly increases below 400 °C after 30 days of service, suggesting the formation of more terminal groups in the branch chain of the molecular structure, which decompose first and cause mass loss at lower temperatures³⁷. Both the maximum decomposition temperature and residual mass decrease, likely due to the further oxidation and decomposition of FM degradation products, which disrupt the material’s cross-linked network and reduce its thermal stability³⁸.

Combining FTIR results with the corrosion environment, it can be inferred that the following reactions predominantly occur during FM service in ultra-high H₂S conditions, leading to performance degradation and eventual failure:

1. FM exhibits some unsaturation in its initial state, with C = C double bonds present. Following service, an increase in C = C double bonds suggests that FM undergoes further defluorination and dehydrogenation reactions.

2. In the liquid medium, H₂S dissociates into H⁺ and HS⁻ ions, with HS⁻ further dissociating into H⁺ and S²⁻ ions³⁹. Likewise, CO₂ hydrolyzes to form H⁺ and HCO₃⁻ ions. Both H⁺ and HS⁻ ions can interact with the carbon groups at FM cross-linking points. FTIR analysis shows a reduction in the absorption intensity of CH₂ = CF₂ after service, indicating that molecular chain fragmentation, reconfiguration, and oxidation of unsaturated bonds occur due to interactions with H⁺ and HS⁻ ions.

3. Post-service, an increase in the C = O and C-O peaks in FM suggests oxidation, branch chain breakage, and the formation of carbonyl or ester groups.

After 7 days of exposure, the initial weight-loss temperature of HNBR decreases, indicating further cross-linking of HNBR in the initial stages of H₂S exposure⁴⁰. Comparison of test results across the three stages reveals that while HNBR retains some C = C bonds in the cross-linked network, the corrosive environment impacts its structural integrity. The original cross-linked network is disrupted, accelerating the reaction rate within the rubber and reducing its thermal stability. Consequently, HNBR experiences a lower initial decomposition temperature and a decrease in the maximum decomposition rate temperature⁴¹.

The mechanical property test results indicate that the tensile strength of HNBR slightly increases during the initial stages of service, followed by significant degradation. Analysis combining TG and FTIR results suggests that the cross-linking density of HNBR increases in the ultra-high H₂S environment during these initial stages. A moderate increase in cross-linking density can enhance tensile strength; however, this also shortens the chain segments between cross-linking points, limiting their mobility and leading to a decrease in elongation at break.

With prolonged service, the interaction of H₂S with nitrile groups and C = C bonds induces additional cross-linking within HNBR. H₂S reacts with unsaturated bonds in the rubber chains to generate mercaptan substituents, which subsequently interact with nitrile groups to form S–C = N structures (Fig. 7)^42,43. The formation of this new three-dimensional network structure intensifies intermolecular forces, resulting in a denser rubber chain configuration and increased hardness. However, this dense cross-linking creates stress concentration points under external forces, contributing to a decline in tensile strength and elongation at break⁴⁴.

Fig. 7.

Reaction mechanism of C = C bonds under H₂S conditions.

Furthermore, as illustrated in Fig. 8, free H⁺ ions react with unpaired electrons on the surface of CB particles, disrupting free radical reactions. This interaction weakens the chemisorption of CB, reducing its binding affinity with the rubber matrix and, consequently, impairing the mechanical properties of the rubber^45,46.

Fig. 8.

Reaction mechanism between H⁺ and unpaired electrons.

H⁺ ions, upon combining with electrons from the CB surface or metal, are reduced to hydrogen atoms. Under typical conditions, these hydrogen atoms have a natural tendency to combine and form H₂ gas, which subsequently escapes from the solution. However, HS⁻ and S²⁻ ions act as inhibitors, preventing the hydrogen atoms from readily combining into H₂^47–50. The rubber chains remain in a state of micro-Brownian motion, generating and eliminating transient pores between molecules. This continuous movement allows hydrogen atoms to permeate through the rubber material^51,52.

The FTIR results confirm that both FM and HNBR retain C = C double bonds before and after service, indicating a degree of unsaturation in their molecular structures. The presence of these double bonds facilitates the penetration of hydrogen into the molecular chain. Hydrogen atoms are more likely to interact with weaker bonds in the carbon chain, such as –C = C– and –C = O, leading to a reduction in bond energy and increasing the susceptibility of the molecular chain to breakage. On the other hand, when sufficient hydrogen atoms accumulate in localized regions within the rubber, they eventually combine to form H₂ gas in areas where HS⁻ and S²⁻ ions are absent, allowing the atoms to combine uninhibited. This localized accumulation of H₂ gas generates internal hydrogen pressure, which compromises the structural integrity of the rubber by causing chain breakage and material degradation (Fig. 9).

Fig. 9.

Mechanism of hydrogen atom permeation in rubber molecular chain.

The FM and HNBR examined in this study do not represent the highest-performing rubber materials. Additional research is warranted to investigate rubber materials with higher fluorine and acrylonitrile content, which exhibit superior corrosion resistance. Such research would help ascertain whether these materials are suitable for ultra-high H₂S environments and whether hard sealing methods are necessary.

Methods

Simulation experiments

Material and conditions

The FM and HNBR materials used in both field and laboratory experiments were provided by Shenyangxin Petrochemical Equipment Co., Ltd. Elemental composition analysis performed using energy dispersive spectroscopy (EDS) indicated that FM primarily contains carbon (C), oxygen (O), and fluorine (F), while HNBR is composed of carbon (C), oxygen (O), and nitrogen (N) (refer to Fig. 10). The primary fillers in these materials include carbon black (CB), white CB, and calcium carbonate (CaCO₃). Samples were fabricated into O-rings with dimensions of φ47.2 × 3.6 mm, adhering to ISO 3601.1–2012 standards⁵³.

Fig. 10.

EDS spectra of samples in the initial state (10,000 ×).

The experimental conditions, summarized in Table 2, replicate the actual operational environment. In the gas phase, H₂S concentration was maintained at 16%, CO₂ at 8%, with the balance comprising nitrogen (N₂), under a total pressure of 10 MPa. The liquid phase medium was formulated based on on-site water analysis, containing a chloride ion (Cl⁻) concentration of 1893.3 mg/L. To ensure complete deoxygenation, nitrogen (N₂) was continuously injected for 24 h prior to testing. During the experiment, the rubber rings were fully immersed in the liquid phase medium.

Table 2.

Experiment conditions and parameters.

Materials	Gas	Solution	Period	Pressure	Temperature
FM	H2S 16%vt CO2 8%vt N2 76%vt	Cl− 1893.3mg/L	7 days	10 MPa	32 ℃
HNBR

Experimental apparatus

The experimental setup employed a high-temperature high-pressure (HTHP) autoclave (refer to Fig. 11) to simulate the ultra-high H₂S and high-pressure conditions found in field operations.

Fig. 11.

Schematic diagram of HTHP dynamic autoclave.

A sealing fixture (refer to Fig. 12) was used to simulate the actual sealing state of the rubber ring, with the assembly process illustrated in Fig. 13. Each component of the apparatus was manufactured with precision, adhering to tolerances not exceeding 0.8 μm, to ensure structural integrity.

Fig. 12.

Schematic diagram of the O-ring sealing fixture.

Fig. 13.

Assembly of O-ring in compressed state.

During assembly, the rubber O-ring was positioned within the groove of the sealing cover (Fig. 13b). After coupling the sealing cover with the sealing body, the compression rate was set to 10%, corresponding to a compressive interference of approximately 1.5 to 2 mm. This configuration established a fully sealed state using the rubber O-ring. The assembled sealing fixture was then placed in the designated position within the HTHP autoclave.

Experimental process

Initially, FM and HNBR samples were divided into two groups, each consisting of five samples. One group was tested to establish baseline properties of the materials in their initial state, while the other group was designated for corrosion experimentation and subsequent performance evaluation.

Next, the assembled sealing fixture and the prepared simulated solution were placed into the HTHP autoclave. Following autoclave sealing, was continuously injected for 2 h to achieve thorough deoxygenation. After this, the autoclave temperature was gradually increased to 32 °C. Pressure levels were then adjusted to 1.6 MPa for H₂S, 0.8 MPa for CO₂, and supplemented with N₂ to reach a total pressure of 10 MPa, initiating the experimental process.

The experiment was conducted over a duration of 168 h. Upon completion, samples were retrieved for detailed analysis, including fracture morphology observation, mechanical property testing, examination of rubber chain structure changes via Fourier-transform infrared spectroscopy (FTIR), and thermal stability assessment through thermogravimetric analysis (TG). Results were then compared with baseline measurements and data from samples subjected to 30 days of service.

To prevent “rapid gas decompression failure,” a controlled pressure relief rate of approximately 0.3 MPa/min was maintained when depressurizing the simulated experimental autoclave and the on-site feedstock gas coalescers during rubber seal removal.

Testing and characterization of mechanical properties

Mechanical properties testing

Following the NORSOK-M710-2014 standard⁵⁴, key indicators for evaluating the performance of rubber seals included density, hardness, and tensile properties. Tensile properties, comprising tensile strength and elongation at break, provide essential insights into the sealing performance of rubber seals after extended use.

The mechanical properties testing adhered to the specifications outlined in GB/T 5720–2008⁵⁵ standard. For tensile properties, a universal electronic tensile test machine (model Y-191) equipped with a standard tensile fixture was utilized, per GB/T 5720–2008 guidelines. During testing, O-rings were converted to linear samples, while on-site sealing rings were shaped into dumbbell samples.

Density measurements were conducted using the drainage method, in compliance with the GB/T 533–2018 standard⁵⁶, with the testing temperature controlled at the standard laboratory temperature of 23 °C ± 2 °C. Hardness was measured according to the ASTM D1414-2022 standard⁵⁷ using an LX-A-type durometer.

The compression set was measured using a dedicated compression set apparatus comprising a compression plate, spacer, and fastening screws (Fig. 14). As specified in the GB/T 5720–2008 standard, the spacer height was set to 85% of the sample’s cross-sectional diameter, with testing conducted at 27 °C ± 2 °C. The compression set was calculated using the following formula:

1

where C₁ is the compression set percentage (%); d₁ is the initial cross-sectional diameter (mm); d₂ is the cross-sectional diameter after recovery (mm); and h is the height of the spacer (mm).

Fig. 14.

Schematic diagram of the compression set device.

After test, the independent sample t-tests were performed to compare the performance of FM and HNBR across different service states.

Fracture morphology and composition analysis

The fracture morphology of rubber materials provides critical information about the fracture process, enabling the identification of fracture characteristics and underlying failure mechanisms. Observing and analyzing fracture morphology is essential for the development and application of durable materials^58,59. In this study, the fracture morphology of rubber samples was examined using a scanning electron microscope (ZEISS EVO MA15, Japan) to elucidate the fracture process. Prior to observation, a small segment of the fracture sample was extracted to preserve the fracture structure. Surface impurities and contaminants were removed via ultrasonic cleaning, and the sample was then coated with gold for scanning electron microscopy, performed at an acceleration voltage of 5 kV.

The surface chemical composition of the rubber samples was analyzed using infrared (IR) spectroscopy. Measurements were obtained with a Nicolet Nexus 670 FTIR-ATR spectrometer, capturing spectra over 64 scans at a resolution of 2 cm⁻¹, within a spectral range of 400 cm⁻¹ to 4000 cm⁻¹.

TG analysis was performed to assess the impact of service duration on the thermal stability of the materials. Samples, with an initial mass of 5 to 10 mg, were heated in a nitrogen atmosphere up to 600 °C at a heating rate of 10 °C/min, using a thermogravimetric analyzer (TA 0550–0381).

Conclusion

The investigation led to the following conclusions:

The mechanical properties of FM progressively declined over prolonged service time, whereas those of HNBR remained relatively stable during the initial service stages before experiencing a marked decrease. After 30 days of service, both FM and HNBR showed significant mechanical deterioration and ultimately failed by brittle fracture.

In ultra-high H₂S environments, FM underwent dehydrogenation and defluorination reactions. Additionally, H⁺ and HS⁻ ions interacted with carbon groups at cross-linking points, resulting in molecular chain breakage and reorganization, along with oxidation and rupture of branch chains.

In the initial stages of service, cross-linking occurred among HNBR molecular chains. With extended service duration, H₂S reacted with unsaturated bonds to form thiol substituents and further reacted with cyano groups, generating S–C = N cross-linking structures. This resulted in a more compact molecular chain configuration, which increased the hardness of HNBR while reducing its mechanical properties.

Throughout the service process, polar groups on the molecular chains of both rubbers caused additives to detach from the matrix, while free H⁺ ions weakened the chemical adsorption of CB. The migration, aggregation, and combination of hydrogen atoms within the rubber chains contributed to molecular chain breakage and compromised structural integrity.

The FM and HNBR studied in this manuscript are not suitable for long-term service in ultra high H₂S environment. Further research is required to define rubber materials with enhanced corrosion resistance, characterized by higher fluoride and propylene cyanide content, or alternatively, hard sealing methods should be considered.

Author contributions

Y Z conceived the experiment(s), S Z and T M conducted the experiment(s), Y Z, Y F, X Y and J Y analysed the results. All authors reviewed the manuscript.

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.