Applicability and Medium-Resistance Performance of Hydrogenated Nitrile Butadiene Rubber with Varying Nitrile Levels Composites with Carbon Black in Supercritical CO₂

Abstract

The structural deterioration and performance instability of hydrogenated nitrile butadiene rubber (HNBR) under supercritical CO₂ (scCO₂) conditions remain major challenges for carbon capture, utilization, and storage (CCUS) applications, despite its widespread use as a sealing material in CO₂ injection wells. This study methodically examined the effects of hydrogenation degree and acrylonitrile (ACN) content on the structural development and service performance of commercial HNBR. The results demonstrate that the combined effects of ACN content and hydrogenation degree dictate the scCO₂ resistance of HNBR. By increasing residual unsaturation, lower hydrogenation degrees enhance the efficacy of peroxide cross-linking, improving structural stability and compression set resistance. Conversely, higher ACN concentrations increase polymer polarity and stiffness while worsening chain scission and mechanical degradation induced by scCO₂. By strengthening the cross-linked network and limiting CO₂ transport, the addition of N550 carbon black greatly improves thermal stability and inhibits scCO₂-induced deterioration. With outstanding mechanical property retention and dimensional stability, the 2020/N550 composite exhibits the greatest balanced performance with a moderate ACN content (∼36%) and a low hydrogenation degree (∼91%). These results provide practical recommendations and mechanistic insights for the logical design of HNBR sealing materials in scCO₂-rich downhole environments.

1. Introduction

Carbon capture, utilization, and storage (CCUS) has emerged as a crucial strategy for lowering greenhouse gas emissions amid international efforts to reach carbon peaking and carbon neutrality, particularly in process-intensive industries like petrochemicals where the marginal abatement potential is intrinsically limited. , Injecting CO₂ into oil and gas reservoirs for enhanced oil recovery (EOR) is one of the most widely used CCUS techniques since it not only promotes long-term geological sequestration but also boosts the productivity of low-permeability reservoirs. , Nevertheless, CO₂ injection wells function in challenging downhole conditions: CO₂ enters a supercritical state (scCO₂) when the temperature rises above 31.1 °C and the pressure rises above 7.38 MPa. − CO₂ can easily penetrate elastomeric sealing materials in packers and cause swelling, plasticization, and rapid gas decompression (RGD) damage in this state because to its high diffusivity and solvent-like solubility. , The safe and effective operation of seals in scCO₂ environments can be seriously hampered by these degradation mechanisms, which can eventually result in premature seal failure. , Because of its better mechanical durability, chemical resistance, and thermal stability than unmodified nitrile butadiene rubber (NBR), hydrogenated nitrile butadiene rubber (HNBR), which is created by selectively hydrogenating butadiene double bonds in NBR, has emerged as a preferred sealing material in downhole tools. , Acrylonitrile (ACN) concentration and hydrogenation level are two important structural characteristics that largely control its overall performance. Increasing the ACN content improves chemical stability, oil resistance, and polarity, but low-temperature flexibility is sacrificed in the process. ,

Similarly, larger levels of hydrogenation improve thermal-oxidative resistance but may reduce processability. , Recent studies have further demonstrated that the mechanical behavior of HNBR depends strongly on its network structure, including cross-link density, molecular chain orientation, and strain-induced crystallization. These effects are also influenced by the ACN content and deformation conditions. Therefore, changes in the mechanical properties of HNBR should not be attributed to chain rearrangement or network modification without direct characterization of its cross-link density and molecular orientation. − As a result, when exposed to scCO₂, HNBR formulations with varying ACN contents and hydrogenation levels may react quite differently, especially in terms of swelling, plasticization, and other degrading mechanisms. The behavior of HNBR in CO₂-rich settings has been extensively studied, offering a crucial basis for its implementation in CCUS systems. The mechanical property evolution of HNBR following scCO₂ exposure was the primary focus of early research. For example, Yang et al. found that while the tensile strength dropped by around 50%, immersion in scCO₂ (35 MPa, 90 °C) for 7 days produced low dissolution and minimal hardness modification. The gas phase showed more noticeable bubbling than the liquid phase, suggesting that bulk mechanical strength is more seriously compromised by scCO₂ than surface integrity. Environmental elements such hydrocarbon aging, pressure, and applied stress were included in later studies. According to Alcock et al., improvements in cross-link density brought on by age improved hardness and tensile stiffness while also decreasing CO₂ diffusivity and swelling. Zeng et al. further simulated wellbore conditions (140 °C, 35 MPa, 0.5 MPa CO₂ partial pressure) and revealed that compressive stress accelerates CO₂-induced corrosion, modifying penetration pathways and restricting chain mobility.

Comparative research has repeatedly shown that HNBR performs better than traditional elastomers. According to Zhu et al., HNBR shows much slower rates of tensile strength and elongation at break degradation than NBR when exposed to scCO₂. They also demonstrated that HNBR degradation is accelerated by compressive stress, with liquid-phase CO₂ causing more severe chemical reactions than gaseous CO₂. Investigations into thermomechanical and dynamic behavior have highlighted the crucial importance of filler–matrix interactions beyond chemical degradation. According to Lainé et al., adding nanofillers makes HNBR more vulnerable to rapid gas decompression (RGD) degradation, but CO₂ sorption decreases its compressive stiffness. These results are consistent with the findings of Ahmed et al., who observed blistering and cracking in HNBR seals after scCO₂ exposure, indicating potential structural degradation under such conditions. The combination of modeling and simulation has improved mechanistic understanding in addition to experimental observations. While Lainé et al. investigated the CO₂ swelling coefficients of fluoroelastomer (FKM) and HNBR, confirming the latter’s better dimensional stability, Jin et al. created a time-dependent Yeoh hyperelastic model to predict scCO₂-induced degradation in HNBR. More recently, multifactor studies have expanded this framework to extreme service environments, such as offshore injection wells, cryo-compressed CO₂ transport, and H₂S–CO₂ coupled conditions, offering a deeper understanding of HNBR behavior under practical operating conditions.

There are still a number of significant gaps in spite of these developments. First, a systematic study of the coupled effects of these two parameters (ACN: 18–45%; hydrogenation: 90–99%) on mechanical evolution, swelling behavior, and failure mechanisms under supercritical CO₂ (scCO₂) conditions is still lacking. Instead, the majority of current research focuses on HNBR with a single acrylonitrile (ACN) content or hydrogenation level. Second, the industrial relevance and applicability of current research findings are limited since commercially available HNBR grades, including Zeon Corporation’s Zetpol series, have well-defined ACN concentrations and hydrogenation levels but have not been thoroughly investigated. Third, a thorough characterization of the time-dependent evolution of critical mechanical propertiessuch as tensile strength, elongation at break, and compression setduring extended scCO₂ exposure is still lacking. This is crucial for precise service-life prediction. The current study systematically investigates the mechanical property evolution of commercial HNBR grades from the Zetpol series with different ACN contents and hydrogenation levels during scCO₂ exposure in order to solve these constraints. HNBR samples with hydrogenation levels of 91%, 95–96%, and 99% and ACN concentrations of 20%, 36%, and 49% were treated with scCO₂ for 72 h at 10 MPa and 100 °C. Following exposure, changes in tensile strength, elongation at break, shore hardness, and compression set were quantitatively assessed. Scanning electron microscopy (SEM), Fourier-transform infrared spectroscopy (FTIR), and tribological measurements were additionally employed to characterize the morphological, chemical, and surface-performance changes after scCO₂ exposure. These analyses provide complementary evidence for interpreting the observed property changes, rather than direct confirmation of molecular-scale degradation mechanisms. The findings should promote the safe and effective operation of CO₂ injection wells and CCUS systems by offering a solid theoretical and practical foundation for the selection and optimization of HNBR sealing materials in scCO₂-rich downhole environments.

2. Materials and Methods

2.1. Raw Materials

Zeon Corporation offered HNBR grades with varying ACN concentrations, including Zetpol4310, Zetpol3310, Zetpol2020, Zetpol2010, Zetpol2000, and Zetpol0020. Dicumyl peroxide (DCP) was purchased from Beijing Honghu United Chemical Products Co., Ltd., while Cabot Corporation supplied carbon black (N550) with a purity of >95%. Table summarizes the detailed compositions of HNBR composites. The HNBR composites design in this study was based on a systematic screening strategy rather than a Design of Experiments (DoE) optimization approach. The primary purpose was to investigate the influence of HNBR molecular structure, including acrylonitrile content and hydrogenation degree, on cross-link-network characteristics and scCO₂ resistance. Therefore, different HNBR grades were selected while keeping the curing system, reinforcing filler, and processing conditions constant to minimize the coupling effects of formulation variables.

1. Detail Composition of HNBR Composites.

samples	zetpol 0020	zetpol 2000	zetpol 2010	zetpol 2020	zetpol 3310	zetpol 4310	N550	DCP
4310						100		2
4310/N550						100	50	2
2020				100				2
2020/N550				100			50	2
2010			100					2
2010/N550			100				50	2
2000		100						2
2000/N550		100					50	2
0020	100							2
0020/N550	100						50	2

2.2. Fabrication and Characterization of HNBR Composite

The rubber was mixed with cross-linker and carbon black, rheometrically evaluated, vulcanized at 150 °C according to Tc 90, and then processed into sheets and standard dumbbell specimens for testing. The compounds were compression-molded at 150 °C for their respective tc90 values. No subsequent postcuring treatment was applied. Before CO₂ was introduced, the high-pressure vessel was briefly purged with nitrogen to remove air. The specimens were then exposed to scCO₂ at 100 °C and 10 MPa for 72 h, followed by depressurization at 0.1 MPa·min^–1. A rotorless rheometer was used to analyze the vulcanization behavior in order to determine the curing conditions. Changes in mass, volume, and compression set were computed after age. Mechanical properties were tested using tensile and hardness tests, thermal stability with a thermal gravimetric analyzer (TGA), microstructural changes with SEM, chemical structure with FTIR, and viscoelastic qualities with dynamic mechanical analysis (DMA) from −75 to 75 °C. Details can be found in Supporting Information, and raw rubber grade information is shown in Table S1.

2.3. Cross-Link Density Measurement

The cross-link density of the HNBR specimens before and after scCO₂ exposure was characterized using a low-field nuclear magnetic resonance analyzer (VTMR20-010 V–I, Niumag, China) with a resonance frequency of 21 MHz. The magnet temperature was maintained at 35.00 ± 0.01 °C. The specimens were cut into strips and placed within the homogeneous region of a 10 mm probe. Cross-link density was evaluated from proton relaxation behavior associated with residual dipolar interactions in the constrained rubber network. Increasing network constraint strengthens the residual dipolar interaction and shortens the proton relaxation time. The relaxation data were analyzed using the cross-link-density analysis software supplied with the instrument, and the results are expressed in 10^–4 mol/cm³. The reported values represent the mean ± standard deviation of repeated measurements.

3. Results and Discussion

3.1. Curing Results of HNBR Rubbers

Table summarizes the curing properties of HNBR compounds at a curing temperature of 150 °C. The minimum and maximum torque values (M _L and M _H) in the unfilled systems did not show a monotonic dependence on the amount of ACN. Higher ACN content improves intermolecular contacts and chain polarity, although its impact on torque generation is closely correlated with molecular weight and hydrogenation degree. ACN level alone does not control cure torque behavior, as evidenced by the fact that 0020 (high ACN) maintained a reasonably high M_H (12.5 dN·m), while 4310 (low ACN content) displayed the highest M_L (6.9 dN·m) and a comparably high MH (13.3 dN·m). The degree of hydrogenation was more important in determining the cure kinetics and torque level. In comparison to 2010 (96%) and 2000 (>99.5%), 2020 (91% hydrogenation) showed higher M_L and M_H and the shortest t_c90 (23.6 min) across grades with comparable ACN levels (∼36.2%). Less hydrogenated HNBR’s increased residual unsaturation offers more reactive sites for peroxide cross-linking, which speeds up curing and encourages the creation of a denser cross-linked network. On the other hand, despite having a similar Mooney viscosity, highly hydrogenated HNBR (2000) displayed lower torque values because of decreased unsaturation and restricted cross-linking efficiency.

2. Curing Results of HNBR Rubbers.

samples	M L (dN·m)	M H (dN·m)	t c 10 (min)	t c 90 (min)
4310	6.9	13.3	2.9	26.6
4310/N550	3.3	8.6	3.4	28.3
2020	6.0	12.6	2.5	23.6
2020/N550	4.8	12.1	2.8	25.0
2010	6.3	14.0	2.9	25.8
2010/N550	3.8	10.2	3.3	27.6
2000	3.5	8.9	2.4	25.0
2000/N550	3.8	9.9	4.0	28.1
0020	4.7	12.5	2.3	25.1
0020/N550	2.4	8.6	3.4	28.9

Mooney viscosity did not exclusively determine cross-link density, but it did have an impact on the baseline torque level. 2010 showed the highest M _H (14.0 dN·m) and raised M _L with a relatively high Mooney viscosity, indicating that effective network formation is favored by a moderate degree of residual unsaturation and enough molecular weight. The incorporation of N550 altered the curing behavior of the HNBR compounds, but the changes in rheometer torque were grade-dependent. The M _L value decreased after N550 addition for all investigated systems, whereas M _H decreased for 4310, 2020, 2010, and 0020 but increased slightly for 2000. The torque difference, ΔM = M _H – M _L, also did not show a uniform trend. In particular, ΔM increased from 6.6 to 7.3 dN·m for 2020 and from 5.4 to 6.1 dN·m for 2000 after N550 incorporation. These results indicate that the effect of N550 on rheometer torque cannot be interpreted simply as either a reduction in peroxide-curing efficiency or a suppression of its conventional reinforcing effect. Rheometer torque reflects the combined response of compound viscosity, filler dispersion, filler-induced network constraint, polymer–filler interactions, and chemical cross-linking. The relative contributions of these factors vary among the HNBR grades because of differences in molecular structure and residual unsaturation. Therefore, the changes in M _L, M _H, and ΔM are interpreted as changes in curing and network-development behavior rather than as direct evidence of weakened reinforcement. The mechanical-property results discussed later confirm that N550 still provides a significant reinforcing effect in the HNBR composites. In contrast, tc10 and tc90 were typically longer, indicating better scorch safety at the expense of curing rate. Nevertheless, the influence of hydrogenation degree remained evident. Among grades with comparable ACN content, 2020/N550 showed the smallest fall in MH and the shortest tc90, indicating that increased residual unsaturation reduces the decreased curing efficiency caused by carbon black. Overall, processing safety was improved by carbon black, but curing efficiency was somewhat compromised.

3.2. Thermal Performance of HNBR Rubbers

HNBR rubber is a critical sealing material for packer applications and is inevitably exposed to elevated temperatures during service, thereby requiring excellent thermal stability. TGA was used on both unfilled HNBR and N550-filled HNBR composites to assess their thermal behavior and the findings are displayed in Figure . In the temperature range of roughly 430–470 °C, the unfilled HNBR shows a typical single-step thermal degrading behavior that is characterized by a rapid mass loss of roughly 65–70% of the total mass. The HNBR polymer backbone’s thermal scission, which includes chain rupture and the volatilization of low-molecular-weight breakdown products, is responsible for this significant degradation stage. The intrinsic thermal resistance of the HNBR matrix is reflected in the small amount of residual char that remains after the mass loss rate significantly drops at higher temperatures. After incorporation of N550 carbon black, the thermal stability of HNBR is significantly enhanced. The onset of thermal degradation shifts to higher temperatures, and the mass-loss rate in the main decomposition region is reduced, indicating delayed thermal decomposition. Moreover, the residual mass at elevated temperatures increases substantially in the N550-filled composites, demonstrating the pronounced char-promoting effect of carbon black. This improvement results from N550 acting as a physical barrier that promotes the creation of a more stable carbonaceous residue during pyrolysis while also inhibiting heat transfer and the diffusion of volatile breakdown products. Furthermore, in both unfilled and filled systems, HNBR grades with a greater bound ACN concentration show better thermal stability. Higher residual mass at higher temperatures is the result of the presence of polar cyano groups, which enhance intermolecular contacts and promote the creation of thermally stable char formations during decomposition. This impact intensifies when N550 is present, indicating a synergistic interaction between carbon black and the polar HNBR matrix. Overall, the TGA results show that N550 carbon black significantly improves the thermal stability of HNBR by increasing char yield and postponing thermal deterioration, and that increased ACN content further boosts thermal resistance. These characteristics are particularly advantageous for HNBR-based sealing materials operating under high-temperature downhole conditions.

1.

Thermal performance of HNBR rubbers: (a) HNBR and (b) HNBR/N550 rubbers.

Quantitative TGA parameters are summarized in Table S2. The incorporation of N550 increased T_5% and T_10% by approximately 7–12 °C, indicating a modest delay in the initial mass-loss stage. However, the changes in Tmax were small for most grades, suggesting that N550 did not substantially alter the principal degradation temperature of the HNBR matrix. An exception was 0020/N550, which exhibited a lower Tmax than unfilled 0020, demonstrating that the effect of N550 was grade-dependent. The higher residual mass of the filled compounds primarily originates from the nonvolatile carbon-black fraction and therefore cannot be regarded alone as evidence of improved thermal stability. Overall, N550 provides a limited improvement in resistance to the onset of thermal mass loss, rather than a pronounced enhancement in the intrinsic thermal stability of HNBR.

3.3. Viscoelastic Performance of HNBR Rubbers

Figure shows the DMA curves for glass transition (T _g), storage modulus, and loss modulus for several HNBR grades (4310, 2020, 2010, 2000, and 0020) without N550 (Figure a–c). For 4310, 2020, 2010, 2000, and 0020, the corresponding T _g values are −24.32, −12.17, −11.40, −11.54, and 2.94 °C. Similar patterns can be seen in the storage modulus (E′) and loss modulus (E″). The highest E′ is 0020, which is followed by 2020, 2010, and 2000. The lowest E′ is 4310. The similar pattern is followed by the E″. These patterns are strongly correlated with elements including the material’s unsaturation, ACN concentration, and degree of hydrogenation. With the highest ACN concentration (49.2%), 0020 shows more stiffness and stronger intermolecular interactions, which raises the E′. Conversely, 2020, 2010, and 2000 exhibit lower storage moduli than 0020 but still greater than 4310 due to their comparatively lower ACN concentration and hydrogenation degrees. 4310, which has the lowest T _g and molecular rigidity, has the lowest loss modulus. The higher T _g of 0020 shows stronger molecular rigidity, which enables it to absorb and dissipate energy more effectively, resulting in a higher E′. In conclusion, T _g values, molecular stiffness, ACN concentration, and hydrogenation degree are all strongly correlated with changes in E′ and E″ of HNBR rubbers.

2.

DMA results of HNBR and HNBR/N550 rubbers: (a) Tan δ, (b) E′ and (c) E″ of HNBR; (d) Tan δ, (e) E′ and (f) E″ of HNBR/N550 rubbers.

The T _g values for the HNBR composite with N550 (Figure d–f) are −22.35 °C, −10.70 °C, −10.67 °C, −10.17 °C, and 3.69 °C for 4310/N550, 2020/N550, 2010/N550, 2000/N550, and 0020/N550, respectively. The highest E′ is ranked as follows: 0020/N550 > 4310/N550 > 2000/N550 > 2020/N550 > 2010/N550. Similarly, the highest E″ is ranked as follows: 0020/N550 > 4310/N550 > 2020/N550 > 2010/N550 > 2000/N550. Compared to the other composites, the T _g of 0020/N550 (3.69 °C) is noticeably higher, suggesting that its polymer chain is more stiff. Its increased ACN content and the additional reinforcing from the addition of N550 are responsible for this increased stiffness. 0020/N550 shows the highest value in terms of E′, indicating its higher capacity to store elastic energy. However, 2010/N550 has the lowest E′, which could be explained by its higher Mooney viscosity and lower degree of hydrogenation. Because of these characteristics, the structure becomes softer and less effective in storing energy. The E″ trend is similar to the E′, with 0020/N550 showing the highest E″ and 2010/N550 the lowest. This implies that 2010/N550 dissipates less energy under dynamic loading because to its poor filler dispersion and lower cross-linking degree, whereas 0020/N550 dissipates more energy. Because of its greater T _g, storage modulus, and loss modulus, 0020/N550 exhibits a relatively high dynamic mechanical response and thermal performance. Additionally, 4310/N550 and 2020/N550 exhibit outstanding qualities, especially in terms of wear resistance and heat resistance. On the other hand, 2010/N550 is ranked lower in all performance parameters, perhaps because of its weaker filler–polymer interactions and lower degree of hydrogenation. These variations demonstrate the substantial impact of filler dispersion, hydrogenation level, and molecular structure on the material characteristics of different HNBR grades. Overall, the DMA results indicate that ACN content and hydrogenation degree influence the intrinsic viscoelastic behavior of the HNBR matrix, whereas the response of the N550-filled systems is governed by a more complex combination of matrix viscoelasticity, effective network constraint, filler dispersion, filler–rubber interfacial interactions, and network heterogeneity. Therefore, the storage and loss moduli do not necessarily vary monotonically with ACN content or hydrogenation degree. In particular, the relatively high storage modulus of 4310/N550 cannot be conclusively attributed to the formation of a stable filler network or bound-rubber layer because these features were not directly characterized. Similarly, the nonmonotonic ranking of the loss modulus suggests the coexistence of several energy-dissipation mechanisms, including segmental motion, interfacial friction, and rearrangement of filler aggregates. Accordingly, the present DMA results should be interpreted as the combined response of multiple structural factors rather than as direct evidence of a specific filler-network mechanism.

3.4. Compression Set Performance of HNBR Rubbers

The performance of the compression set was evaluated by measuring the recovery height of standard specimens following 72 h of compression at 100 °C and these findings are shown in Figure . 2020, 2010, and 2000 show significantly varied compression set behaviors with the same ACN content (36.2%). These changes are mostly caused by variations in cross-linking network architectures brought on by variations in hydrogenation degree. 2020 offers more reactive sites for peroxide curing due to its comparatively low hydrogenation degree and thus higher residual unsaturation. Consequently, a denser three-dimensional network and a greater effective cross-link density are created, which more successfully prevent irreversible rearrangement and molecular chain slippage under prolonged compression at 100 °C, resulting in better elastic recovery and a lower compression set. Higher hydrogenation levels in 2010 and 2000, on the other hand, significantly lower the quantity of remaining double bonds, restrict the effectiveness of peroxide cross-linking, and produce relatively looser cross-linked networks. Such networks are more prone to stress relaxation and viscoelastic flow during prolonged high-temperature compression, which increases the buildup of permanent deformation and, consequently, the compression set values. Even though the three materials exhibit comparable Mooney viscosities and molecular weight effects are negligible, their combination with cross-linking structure may increase the buildup of deformation in highly hydrogenated systems at high temperatures. Overall, variations in the compression set of HNBR are mostly determined by the hydrogenation degree (residual unsaturation) when the ACN concentration is kept constant.

3.

Compression set performance of HNBR rubbers: (a) HNBR and (b) HNBR/N550 rubbers.

Following the addition of N550 carbon black, the composites showed significant variations in compression set, which were as follows: 4310/N550 (18.8%) < 2020/N550 (20.1%) < 0020/N550 (28.3%) < 2000/N550 (35.1%) ≈ 2010/N550 (36.7%). This trend indicates that compression set does not vary monotonically with ACN content or filler addition alone, but is determined by the synergistic effects of effective cross-link density and filler network/interfacial constraints. The low hydrogenation degree (∼91%) of 2020/N550 allows for enough residual unsaturation to maintain effective DCP cross-linking even after carbon black is added. This keeps the effective cross-link density high and prevents stress relaxation during compression at 100 °C for 72 h. On the other hand, 2010/N550 and 2000/N550, which have higher hydrogenation degrees (∼96% and >99.5%), show much less residual unsaturation and limited cross-linking efficiency. Their looser cross-linked networks, which are consistent with the longer tc₁₀ and tc₉₀ seen in vulcanization data, result in increased chain slippage and viscoelastic flow under high-temperature compression and, as a result, greater permanent deformation. Notably, 4310/N550 exhibits the lowest compression set despite having the lowest ACN content, indicating that ACN content is not the sole factor governing compression-set behavior. Compression set is a macroscopic response jointly influenced by matrix viscoelasticity, segmental relaxation, effective network constraint, and thermal stability during prolonged compression at elevated temperature. Therefore, the relatively low compression set of 4310/N550 indicates better recovery from compressive deformation under the present test conditions, although the individual contributions of these factors cannot be determined from the available results. Because of its greater ACN content (∼49.2%), which lowers compression set, 0020/N550 benefits from stronger filler–rubber interfacial contacts. However, its relatively high hydrogenation degree (∼96%) limits cross-link density, resulting in an intermediate compression set level. Overall, the experimental findings show that residual unsaturation controls cross-linking ability in filled systems, while matrix rheology and ACN concentration work together to control filler–rubber interfacial interactions, which in turn regulate the final compression set.

3.5. Corrosion Resistance Performance in scCO₂ Environment

A corrosion resistance test was conducted to assess HNBR rubber’s corrosion resistance performance in a supercritical CO₂ environment. The HNBR samples were put in a high-temperature, high-pressure reaction vessel with a CO₂ gas pressure of 10 MPa and a temperature of 100 °C. Before being gradually cooled to room temperature, the vessel was kept under these conditions for 72 h. To safely discharge the pressure, the gas was gradually released once the temperature had stabilized. Following this procedure, each HNBR rubber’s mass change rate and volume change rate were measured, as seen in Figure . The mass change rates of 4310, 2020, 2010, 2000, and 0020 for the unfilled HNBR rubbers were 0.874%, 0.158%, −0.167%, −0.247%, and −0.983%, respectively; the corresponding volume change rates were 0.463%, −0.076%, −0.397%, −0.394%, and −1.305%. 4310 showed concurrent positive changes in both mass and volume, suggesting that CO₂ uptake and swelling dominated its response during exposure. The simultaneous negative mass and volume changes of 0020 may be associated with the combined effects of mobile-component loss and postdepressurization structural contraction. However, because the potentially extracted species were not chemically analyzed, their nature and contribution cannot be conclusively determined from the present results. 2020 had the least volume change and one of the least mass changes among the grades with the same ACN content (36.2%), indicating the best balanced reaction to CO₂ exposure. This behavior can be explained by the combined effects of relatively low hydrogenation degree and moderate polarity, which both inhibit rapid CO₂ diffusion associated with more flexible, low-ACN systems and ameliorate excessive CO₂ affinity and swelling. Additionally, the increased residual unsaturation linked to reduced hydrogenation may moderate CO₂ dissolution and transport within the HNBR matrix by increasing chain stiffness and decreasing free-volume-assisted diffusion.

4.

Corrosion resistance performance of HNBR rubber in a supercritical CO₂ environment: (a) mass change rate of HNBR, (b) mass change rate of HNBR/N550 rubbers, (c) volume change rate of HNBR and (d) volume change rate of HNBR/N550 rubbers.

The mass change rates for the N550-filled HNBR composites were 0.595%, 0.960%, −0.065%, 0.667%, and 0.067% for 4310/N550, 2020/N550, 2010/N550, 2000/N550, and 0020/N550, while the corresponding volume change rates were 0.456%, 0.728%, −0.294%, 0.373%, and −1.020%. By inhibiting CO₂-induced extraction and encouraging CO₂ retention within the composite, the addition of N550 significantly changed the mass and volume evolution behavior when compared to the unfilled systems. Due to a favorable balance between residual unsaturation and filler–rubber interfacial interactions, 2020/N550 showed the biggest positive changes in both mass and volume, suggesting that CO₂ uptake and swelling became dominant in this grade. On the other hand, 2010/N550 exhibited a little mass loss along with volumetric shrinkage, indicating that even with carbon black present, inadequate cross-linking effectiveness at high hydrogenation levels allowed CO₂-induced chain rearrangement. While both 4310/N550 and 2000/N550 showed positive mass and volume changes, their responses were not as strong as those of 2020/N550, indicating variations in matrix rheology and ACN concentration that affect CO₂ transport and filler network formation. The nearly unchanged mass but pronounced volume shrinkage of 0020/N550 may reflect a balance between residual CO₂ uptake and mobile-component loss, together with postdepressurization contraction and structural relaxation under filler-related constraints. Overall, these findings show that the synergistic impacts of hydrogenation degree, ACN concentration, and filler–rubber interfacial structure control the mass and volume responses of HNBR to CO₂ exposure in filled systems. Overall, the 2020-based composite shows the best reaction to CO₂ exposure of all the N550-filled HNBR systems under investigation. The HNBR matrix and N550 carbon black exhibit a steady and well-balanced interaction, as evidenced by the biggest positive mass and volume changes, which indicate dominant CO₂ uptake without considerable extraction or structural collapse. Its moderate ACN content and comparatively low hydrogenation degree, which together guarantee adequate chain rigidity, efficient cross-linking capabilities, and strong filler–rubber interfacial contacts, give rise to this ideal behavior. As a result, the 2020/N550 system is the most appropriate formulation for applications under CO₂-rich service circumstances since it strikes the optimal balance between resistance to CO₂-induced extraction and regulated swelling. It should be noted that mass and volume changes alone cannot be directly used to evaluate scCO₂ resistance. A desirable sealing material should exhibit limited dimensional variation together with stable mechanical properties after CO₂ exposure.

3.6. Mechanical Performance Change in scCO₂ Environment

The changes in hardness, tensile strength, and elongation at break were examined for 4310, 2020, 2010, 2000, and 0020 under the circumstances of 100 °C, 10 MPa, and 72 h in order to examine the mechanisms of supercritical CO₂ on HNBR, as shown in Figure . The findings showed that the mechanical property evolution trends of the various HNBR grades varied significantly. The tensile strength of 4310 stayed constant at 3.0 MPa, however the hardness marginally dropped from 45A to 43A. This implies that while surface relaxation has place, there were no notable alterations to the overall cross-linking structure. The stability of 4310 during CO₂ exposure is attributed to its low ACN concentration and comparatively high hydrogenation degree. Further evidence that CO₂ had little effect on the chain segments, with just a little loss in material flexibility, comes from a slight fall in elongation at break from 382% to 370%.

5.

Mechanical performance change in supercritical CO₂ environment: (a) hardness of HNBR, (b) tensile strength of HNBR, (c)­elongation at break of HNBR, (d) hardness of HNBR/N550, (e) tensile strength of HNBR/N550 and (f) elongation at break of HNBR/N550.

The tensile strength increased from 8.1 to 11.1 MPa in 2020 despite a small decrease in hardness from 55A to 53A, while elongation at break increased from 447% to 477%. The simultaneous increase in tensile strength and elongation at break suggests that the cross-linked network remained well preserved during scCO₂ exposure. Such behavior may be associated with the combined effects of slight postcuring during prolonged thermal exposure, residual stress relaxation, and physical aging during depressurization. In 2010, the tensile strength dropped from 11.4 to 8.2 MPa, and the hardness dropped from 53A to 52A. The elongation at break, however, rose little from 594% to 599%. This implies that the material’s flexibility and ductility were not significantly impacted, even though CO₂ caused some relaxation of the cross-linked network and chain scission, enhancing chain mobility. In comparison to 2020, the material in 2010 showed considerable strength loss and a reduced degree of hydrogenation, although it still had good elongation capabilities. Furthermore, 2000 showed a significant increase in elongation at break (from 601% to 661%) and a slight decrease in tensile strength (from 10.4 to 9.9 MPa), but minimal change in hardness (from 52A to 51A). This implies that, despite a modest decrease in strength, chain mobility and flexibility were greatly increased, most likely as a result of CO2-induced physical rearrangement of the cross-linked network, which increased its elongation. The tensile strength of 0020 dropped dramatically from 12.3 to 5.6 MPa, while elongation at break increased slightly from 454% to 465%. However, the hardness of the material did not change at 56A. This behavior can be attributed to the high ACN content and lower hydrogenation degree, making it more susceptible to chain scission under CO₂ exposure. While the hardness remained unchanged, the significant reduction in tensile strength indicates substantial damage to the cross-linking network. The slight increase in elongation suggests localized network relaxation, which contributed to increased flexibility, but the significant loss of strength indicates substantial structural damage.

For the N550-filled HNBR composites, the changes in hardness, tensile strength, and elongation at break were tested for 4310/N550, 2020/N550, 2010/N550, 2000/N550, and 0020/N550 under the same conditions. For 4310/N550, the hardness decreased from 65A to 61A, while the tensile strength increased from 20.7 to 26.2 MPa, and elongation at break improved from 307% to 366%. This indicates that N550 reinforcement helped to mitigate CO₂-induced swelling and enhanced its strength and flexibility by promoting stress redistribution and preserving network integrity. The slight decrease in hardness suggests that CO₂ exposure led to some relaxation of the cross-linked network, but the presence of N550 effectively restricted this effect.

For 2020/N550, the hardness decreased from 74A to 71A, but the tensile strength significantly increased from 28.4 to 33.5 MPa, and the elongation at break increased from 312% to 333%. This result shows that 2020/N550 achieved the most favorable balance between strength and flexibility, with both tensile strength and elongation improving significantly. The increase in tensile strength can be attributed to the molecular chain reorientation and localized densification promoted by CO2, while the increase in elongation at break indicates enhanced physical cross-linking, forming a stronger, more flexible network structure. This combination of enhanced strength and flexibility makes 2020/N550 the optimal formulation among the tested composites. For 2010/N550, the hardness decreased from 72A to 69A, while the tensile strength increased from 24.2 to 28.3 MPa, and the elongation at break significantly increased from 422% to 513%. These improvements indicate that N550 effectively stabilized the cross-linked network, enhancing both the strength and flexibility of the material. The significant increase in elongation at break suggests that the addition of N550 facilitated stress relaxation while maintaining the cross-linked network, contributing to the improved ductility of the composite under CO₂ exposure. For 2000/N550, the hardness decreased from 72A to 68A, while the tensile strength increased from 20.0 to 22.7 MPa, and the elongation at break increased from 499% to 517%. These results highlight that while the strength of the material slightly weakened, its elongation at break was significantly enhanced, demonstrating that N550 improved chain mobility and flexibility under CO₂ exposure, likely due to CO₂-induced chain scission and reorientation.

For 0020/N550, the hardness decreased from 76A to 74A, while the tensile strength increased from 27.0 to 30.1 MPa, and the elongation at break increased from 331% to 360%. The increase in tensile strength and elongation at break suggests that N550 effectively mitigated the degradation effects of CO₂ exposure, but its performance was slightly less impressive compared to 2020/N550, likely due to the higher ACN content and lower hydrogenation degree of 0020, which made it more prone to chain scission and network breakdown. Overall, 2020/N550 demonstrated a favorable balance of mechanical properties after CO₂ exposure, with a simultaneous increase in tensile strength and elongation at break. The combination of moderate ACN content, relatively low hydrogenation degree, and N550 reinforcement provided the most effective stabilization mechanism, ensuring enhanced strength, flexibility, and resistance to CO₂-induced degradation.

3.7. Microscopic Morphologies of HNBR Rubbers in scCO₂ Environment

The HNBR materials showed varying degrees of surface roughening and increased striping upon exposure to the scCO2 environment (Figure ). Among them, 2020 showed good cross-sectional continuity even before corrosion, despite the presence of crack bifurcations and stepped lamellar features. After corrosion, no obvious through-thickness delamination cracks or large peeling bands were observed in the low-magnification view of 2020. At high magnification, the lamellae displayed mainly uniform undulations, the interlayer boundaries appeared relatively blunted, and no clear hollowing collapse or stacked blocks of thick sheet-like tearing were detected. These features suggest that, during CO₂ permeation and plasticization, defect evolution in 2020 is more likely to be dispersed into multiple microscale slip paths or fine wrinkles rather than concentrated into a few dominant fracture surfaces. As a result, 2020 exhibits stronger resistance to corrosion-induced morphological degradation. In contrast, 4310 developed more distinct directional stripes at low magnification and sharper lamellar ripples at high magnification, implying more localized interfacial slip. For 2010, the postcorrosion roughness increased at low magnification and the local texture scale became larger, indicating a trend toward more concentrated crack propagation. After corrosion, 2000 showed denser stripes/grooves and stronger overall undulations, reflecting greater structural rearrangement under CO₂ exposure. Meanwhile, 0020 exhibited pronounced banding after corrosion, suggesting that stress is more readily released through directionally oriented delamination. Overall, compared with 2020, these samples tended to display clearer, more penetrating, and/or more severe striping, layering, and roughening features after scCO₂ exposure. Therefore, 2020 maintained the best cross-sectional integrity and the mildest morphological evolution, with the lowest degree of defect amplification, demonstrating superior stability against scCO₂ action. This behavior further indicates that, in the carbon-black-reinforced system, 2020 achieves more effective matrix–filler synergistic constraint, which helps disperse CO₂-induced plasticization and stress relaxation more uniformly and thereby suppresses significant interlayer instability.

6.

Cross-sectional microscopic morphologies of HNBR rubbers in scCO₂ environment: (a) 4310, unexposed; (b) 4310, after corrosion; (c) 2020, unexposed; (d) 2020, after corrosion; (e) 2010, unexposed; (f) 2010, after corrosion; (g) 2000, unexposed; (h) 2000, after corrosion; (i) 0020, unexposed; and (j) 0020, after corrosion.

After exposure to the scCO₂ environment, the cross-sectional morphologies of the five HNBR/N550 composites exhibited different degrees of surface roughening and enhanced layered textures (Figure ). Among them, 2020/N550 retained the relatively good cross-sectional integrity after scCO₂ exposure: at low magnification, no obvious bubbling bulges or large collapsed undulations were observed, while at high magnification the microstructure was dominated by relatively uniform lamellae. In addition, defect boundaries appeared more blunted and crack paths were more dispersed, which prevented the formation of a continuous, through-thickness delaminated surface. In contrast, 4310/N550 developed pronounced lamellar wrinkles and interfacial slip marks after corrosion. Although 2010/N550 remained largely intact, distinct directionally oriented delamination zones were still evident, suggesting that locally weak interfaces can be amplified by CO₂ exposure. The corroded morphology of 2000/N550 was characterized by bubbly or blocky undulations, indicating a higher tendency for micropore nucleation and coalescence driven by CO₂-induced volume changes. Meanwhile, 0020/N550 showed the strongest banding and layering features, consistent with a greater propensity for rigid/brittle fracture associated with its high ACN content and a higher likelihood of interfacial separation after corrosion. Overall, considering the apparent porosity/foaming, layer connectivity, and cross-sectional continuity, 2020/N550 exhibited the most stable corroded cross-section. Therefore, within this HNBR group, 2020/N550 experienced the smallest morphological changes, maintained structural continuity most effectively, and showed the lowest degree of defect amplification after scCO₂ exposure, demonstrating superior resistance to scCO₂ corrosion. This performance is mainly attributed to the swelling–rigidity balance provided by a moderate ACN level, together with a more uniform cross-linked network and improved resistance to defect propagation resulting from faster and more complete vulcanization.

7.

Cross-sectional microscopic morphologies of HNBR/N550 rubbers in scCO₂ environment: (a) 4310/N550, unexposed; (b) 4310/N550, after corrosion; (c) 2020/N550, unexposed; (d) 2020/N550, after corrosion; (e) 2010/N550, unexposed; (f) 2010/N550, after corrosion; (g) 2000/N550, unexposed; (h) 2000/N550, after corrosion; (i) 0020/N550, unexposed; and (j) 0020/N550, after corrosion.

3.8. Chemical Structure of HNBR Rubbers in scCO₂ Environment

The FTIR spectra of HNBR and HNBR/N550 composites in supercritical CO₂ are shown in Figure . The unexposed pure HNBR samples (Figure a) exhibit typical HNBR characteristic bands. The absorptions at ∼2920 and ∼2850 cm^–1 correspond to the stretching vibrations of aliphatic –CH₂/–CH₃ groups, while the band at 2235–2240 cm^–1 is assigned to the stretching vibration of the nitrile group (CN). In addition, the –CH₂ scissoring bending and –CH₃ deformation vibrations are observed at ∼1450–1460 cm^–1 and ∼1370–1380 cm^–1, respectively. The peak positions are clear and stable, indicating that the polymer backbone and nitrile functionalities remain intact in the initial state. The spectral differences among grades are mainly reflected in the intensity of the –CN band and minor variations in band shape.

8.

FTIR of HNBR and HNBR/N550 composites in scCO₂ environment: HNBR (a) unexposed and (b) after corrosion; HNBR/N550 (c) unexposed and (d) after corrosion.

After exposure to scCO₂ (Figure b), the main characteristic peaks of pure HNBR remain at essentially the same wavenumbers. No obvious new absorption bands or pronounced peak shifts are observed for the –CH₂/–CH₃ (2920/2850 cm^–1) and –CN (2235–2240 cm^–1) bands after scCO₂ exposure. Within the sensitivity and spectral resolution of the present FTIR measurements, these results indicate that no large-scale chemical transformation or substantial functional-group change occurred under the investigated conditions. However, minor oxidation, localized chain scission, cross-link rearrangement, or other low-concentration chemical changes cannot be completely excluded because such changes may remain below the detection threshold of conventional FTIR. Notably, some grades show changes in band intensity, band shape, and baseline flatness after exposure, but no consistent emergence or strengthening of carbonyl (CO) absorption in the 1700–1750 cm^–1 region is detected. This pattern indicates that the spectral variations are more likely governed by physical effectssuch as CO₂ permeation–induced plasticization, swelling/shrinkage, and associated chain orientation or microdensity changesrather than chemical structural changes dominated by oxidative degradation.

For the carbon-black-reinforced HNBR/N550 system (Figure c), the HNBR bands at 2920/2850 cm^–1 (–CH₂/–CH₃) and 2235–2240 cm^–1 (–CN) are retained. Compared with pure HNBR, the overall baseline is more susceptible to elevation and tilting, which is consistent with the broad infrared absorption and scattering effects introduced by carbon black. After scCO₂ exposure (Figure d), the characteristic peak positions remain largely stable, indicating that the composite also does not undergo pronounced formation or loss of major functional groups. Therefore, the primary role of scCO₂ in the HNBR/N550 system is inferred to be physical rather than chemical, and these physical processes likely contribute to the observed macroscopic differences in banding, layering, and surface roughness. Overall, FTIR analysis shows that the principal characteristic bands of HNBR remain largely unchanged after scCO₂ exposure, and no major chemical transformation is evident within the sensitivity of the measurement. Combined with other results, the observed performance evolution is more consistent with physical processes such as CO₂ diffusion, plasticization, swelling, extraction, and cross-link-network adjustment. Nevertheless, FTIR alone cannot exclude minor oxidation, localized chain scission, or cross-link rearrangement occurring below its detection threshold. Therefore, the present spectroscopic results should be interpreted as evidence that physical effects are likely dominant, rather than as proof of the complete absence of chemical changes.

3.9. Changes in Cross-Link Density after scCO₂ Exposure

Figure presents the low-field NMR-derived cross-link density of the HNBR and HNBR/N550 specimens before and after scCO₂ exposure. Before exposure, the cross-link density of the unfilled HNBR grades increased from 8.19 × 10^–4 to 15.94 × 10^–4 mol/cm³ with increasing ACN content and residual structural differences among the commercial grades. A similar trend was observed for the N550-filled compounds, whose initial values ranged from 7.30 × 10^–4 to 15.37 × 10^–4 mol/cm³. After scCO₂ exposure, the apparent cross-link density increased for all formulations. The increases were approximately 4.0–11.2% for the unfilled HNBR specimens and 5.0–13.8% for the HNBR/N550 compounds. The largest relative increase was observed for 2010/N550, whereas 4310 exhibited the smallest increase among the unfilled samples.

9.

Low-field NMR-derived apparent cross-link density of (a) HNBR and (b) HNBR/N550 before and after scCO₂ exposure at 100 °C and 10 MPa for 72 h.

These results provide direct evidence that the average constraint imposed on mobile polymer segments changed after scCO₂ exposure. Limited postcuring during exposure at 100 °C may have contributed to this increase, particularly because the compounds were cured using a peroxide system. Physical aging, redistribution or removal of mobile low-molecular-weight species, and network relaxation during depressurization may also affect the low-field NMR relaxation response. Nevertheless, the NMR-derived cross-link density should be interpreted as an apparent or effective network parameter. It reflects changes in segmental mobility and network constraint but does not independently distinguish chemical cross-links from physical entanglements, filler–rubber interactions, or other mobility restrictions.

4. Conclusions

This study fully investigated the structure–property relationships of commercial HNBR with varying ACN concentrations and hydrogenation degrees under scCO₂ circumstances. The results demonstrate that rather than a single structural parameter, the combined effects of ACN concentration and hydrogenation degree govern the service behavior of HNBR in scCO₂. The overall performance of HNBR under scCO₂ conditions is greatly improved by the addition of N550 carbon black. By strengthening the rubber matrix, maintaining the cross-linked network, and reducing CO₂ transit under filler–rubber interfacial constraints, carbon black enhances thermal stability in filled systems while also reducing scCO₂-induced degradation. Consequently, the dimensional stability and mechanical property retention of HNBR are greatly enhanced, and the degrading behavior is controlled by the combined impacts of filler network structure, matrix rheology, and cross-link density. The 2020/N550 composite, with moderate ACN content (∼36%) and low hydrogenation degree (∼91%), exhibits a favorable balance of mechanical properties, dimensional stability, and scCO₂ adaptability, making it a promising candidate for further sealing-material development. These findings give mechanistic insight into the scCO₂-induced degradation of HNBR, as well as practical guidance for rationally designing and selecting sealing elastomers for CO₂-rich downhole and CCUS applications.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c06441.

• Experimental details for the fabrication and characterization of HNBR composites; properties of raw HNBR grades (Table S1); quantitative thermogravimetric parameters of HNBR and N550/HNBR compounds (Table S2) (PDF)

The authors declare no competing financial interest.