Analysis of the aging mechanism and life evaluation of elastomers in simulated proton exchange membrane fuel cell environments

2.1 Samples and experimental environments

In this study, silicone rubber, which is a commercial elastomeric material (Dongjue, Inc., China), was chosen for the tests. The material is crosslinked by polymethylsiloxane with vinyl and polymethylsiloxane with a hydrosilylated functional group. It is composed of siloxane by hydrosilylation, and the fillers are mostly composed of silica, quartz, and calcium carbonate. The silicone rubbers are shaped into squares (10 mm by 10 mm; 6 mm thick) and cylinders (12.5 mm in diameter; 6 mm thick). Two test solutions with different acid concentrations are prepared. One test solution is called the regular solution (RS). It approximates the actual PEMFC operating conditions (6,7,19,21). The solution includes 12.5 ppm sulfuric acid (H₂SO₄) and 1.8 ppm hydrofluoric acid (HF). The other solution, with 1 mol·L⁻¹ H₂SO₄, 30 ppm HF, is marked as ADT for the accelerated durability test in a short time. The aging temperature is controlled at 70°C.

2.2 Characterization methods

The samples were individually exposed to two different experimental conditions for a period of time. Then, the aged samples were taken out and cleaned before the tests. The modification of the surface morphology of the samples was analyzed using optical microscopy. The ATR-FTIR and XPS techniques can help qualitatively and quantitatively determine changes in the chemical structure. The ATF-FTIR was conducted to obtain the absorbance of the chemical group on the surface of specimens with the help of a Nexus Model 670 device, with a resolution of 0.1 cm⁻¹. XPS using a Thermo Fisher Scientific K-Alpha spectrometer with a monochromatic Al Kα X-ray source was employed to analyze the chemical changes on the surface of specimens. The survey spectra ranged from 0 to 1,350 eV and were maintained at 1 eV steps for each specimen. The stress relaxation test was performed, and the compression Young’s modulus was chosen to characterize the changes in the mechanical properties.

3.1 Morphology analysis

The micro characteristic changes on the surface of the silicones are displayed in Figure 1. The magnification is 200×. Figure 1a and c show the micro topography of the unexposed samples. Figure 1b and d show the micro topography of the samples immersed into RS and ADT test solutions for 33 days at 70°C individually. It can be summarized from Figure 1 that the surface morphologies of the samples taken out of the RS test environment changed, but no cracks appeared (Figure 1c). The ADT test environment accelerates the aging of the samples from outside to inside. The surface becomes rough, and many cracks appear. The samples degrade severely in the ADT test solution. This aging process exhibits both acid-dependent and time-dependent processes. This conclusion is in agreement with the results found in previous research (7).

Figure 1

Micromorphology of samples exposed to test solutions: (a) unexposed, (b) for 33 days-RS, (c) unexposed, and (d) for 33 days – ADT.

3.2 ATR-FTIR and XPS results

ATR-FTIR and XPS, advanced material analysis techniques, were chosen to investigate the chemical composition changes on the surface of the silicone rubber specimens before and after aging under the experimental conditions. Figures 2 and 3 show the spectra of specimens measured using the ATR-FTIR method. The spectra ranging from 650 to 1,500 cm⁻¹ are displayed in Figures 2a and 3a, and spectra from 2,500 to 3,500 cm⁻¹ are displayed in Figures 2b and 3b. The spectra A in Figures 2 and 3 belong to the specimens before exposure. The spectra B in Figures 2 and 3 belong to the specimens after being immersed into two different solutions, respectively.

Figure 2

IR spectra for (A) unexposed specimens and (B) aged specimens in RS for 33 days at 70°C. (a) Wavenumbers from 650 to 1,500 cm⁻¹, and (b) wavenumbers from 2,500 to 3,500 cm⁻¹.

Figure 3

IR spectra for (A) unexposed specimens and (B) aged specimens in ADT solution for 33 days at 70°C. (a) Wavenumbers from 650 to 1,500 cm⁻¹ and (b) wavenumbers from 2,500 to 3,500 cm⁻¹.

For unaged specimens (Figures 2a and 3a), the peak that appears at 1,412 cm⁻¹ is ascribed to the rocking vibration of the CH₂ groups in the crosslinking area. The peak that appears at 1,257 cm⁻¹ is caused by the Si–CH₃ bending vibration. A broadened peak from 1,057 to 1,007 cm⁻¹ is strong and represents the characteristic of stretching vibrations of the backbone Si–O–Si. The peak that appears at 864 cm⁻¹ is caused by the rocking vibration of the Si–CH₃ groups. A prominent peak near 785 cm⁻¹ is ascribed to the coupling of the CH₃ chemical bonds rocking vibration and Si–C stretching vibration. Two peaks that appear at 692 and 2,962 cm⁻¹ are due to CH₃ chemical bond stretching vibrations. All these vibration modes at different spectra are analyzed based on prior references (7,23).

For specimens immersed in the RS test environment for 33 days (shown in Figure 2b), the intensity of all absorption peaks decreases compared with that of the unexposed specimens (Figure 2a). This phenomenon demonstrates that the chemical composition on the surface of the specimens changes after they are aged in the RS solution.

For specimens exposed to the ADT test solution for 33 days (Figure 3a), the intensity of the peaks weakens sharply with exposure time compared with that of the unexposed specimens (Figure 3b). The peaks at 2,962 cm⁻¹ (CH₃) and 864 cm⁻¹ (Si–CH₃) disappear for the specimen after 33 days of exposure. The intensities of the peaks at 1,257 cm⁻¹ (Si–CH₃), 785 cm⁻¹ (CH_3, Si–C), and 692 cm⁻¹ (CH₃) decrease dramatically. This may be due to the breakage of the side chain Si–CH₃. The peak at 1,412 cm⁻¹ (CH₂) disappears after 33 days of exposure. This may be due to the hydrolysis of the crosslink point –CH₂–CH₂– in the crosslink area. The intensity of the peak from 1,057 to 1,007 cm⁻¹ decreases sharply, which suggests decomposition of the silicone rubber backbone Si–O–Si. Based on the comparison of the changes of peaks for specimens immersed in two different test solutions, the conclusion is that the degradation process of specimens is highly dependent on the concentration of the acid in the exposure environments. The solution with a higher acid concentration could make the Si–O–Si, –CH₂–CH₂–, and CH₃ structures decompose more seriously and accelerate the aging of silicone rubber materials.

For further analysis, the surface elemental compositions of specimens were determined using XPS. As shown in Figure 4a, survey XPS shows the presence of elements of silicon (Si), oxygen (O), and carbon (C) in the unexposed specimens. In addition, there is little fluorine (F). The high-resolution spectra for Si2p, O1s, and C1s (24,25,26), as shown in Figure 4b–d, respectively, in specimens before and after immersion into two different test solutions for 83 days are chosen to further study the chemical compositions. The high-resolution spectra for elements Si, O, and C are saved at each 0.1 eV step by deconvolution using curve-fitting. From Figure 4b–d, it can be seen that the binding energies of Si2p, O1s, and C1s for a specimen immersed in the RS test solution do not obviously change. For the specimens immersed in the ADT test solution, the binding energy (approximately 101.3 eV) of Si2p shifts 2 eV (Figure 4b) toward a higher binding energy level (approximately 103.2 eV). The binding energy of O1s (approximately 531.5 eV) shifts 0.9 eV (Figure 4c) toward a higher binding energy level (approximately 532.4 eV). The binding energy of C1s (approximately 283.75 eV) has no obvious shift.

Figure 4

XPS spectrum and high-resolution spectra analysis for the specimens: (a) XPS spectrum for unexposed specimens, (b) high-resolution spectra-Si2p, (c) high-resolution spectra-O1s, and (d) high-resolution spectra-C1s.

Atomic concentrations of four types of elements are calculated by integrating the relevant peak intensities and are listed in Table 1. The proportion of C and Si atomic concentration (C/Si) and the proportion of O and Si atomic concentration (O/Si) for specimens over time are calculated and listed in Table 1. The atomic concentration ratio of C/Si decreases dramatically (1.663 → 1.556 → 0.657), while the atomic concentration ratio of O/Si increases (0.919 → 0.958 → 1.416) as the acid content increases in the solutions over time. The decrease in the C/Si atomic concentration ratio may be due to the hydrolysis of the crosslink bond –CH₂–CH₂–, which transforms it into Si–O under the action of the acid and temperature, and then it forms a different crystal structure SiO₂ from the original –Si–O–Si– backbone structure. An increase in the O/Si atomic concentration ratio may occur because the side chain Si–CH₃ is transformed to Si–O under the action of oxidation and H⁺. First, the backbone Si–O–Si is transformed to an unstable Si–OH bond under the action of the acid, and then the Si–OH is transformed to a stable Si–O bond under the action of H⁺. In summary, these results may be due to the destruction of the –CH₂–CH₂– and CH₃ chemical bonds. They may be oxidized and then transformed into Si–O bonds. Additionally, hydrolytic breakage of the backbone (–Si–O–Si–) may also lead to an increase in the O/Si ratio. The ATR-FTIR results are consistent with the XPS results.

Combining the ATR-FTIR and XPS analyses, it is determined that the changes in the chemical composition on the surface of the silicone rubber materials are caused by the decomposition of the main chain Si–O–Si in the backbone to form Si–O, the hydrolysis of crosslinks –CH₂–CH₂–, and the fracture of the side chain Si–CH₃ because of the oxidation and H⁺ attack in the acidic experimental environments. A solution with a higher acid concentration causes more serious decomposition of the chemical structures and accelerates the aging of silicone rubbers.

3.3 Compression mechanical properties

Compression stress relaxation tests were performed on the silicone rubbers at 25% strain levels. The variations of the time-dependent Young’s modulus are shown in Figure 5. The Young’s modulus E(t) increases more for the samples aged in the ADT test solution than the samples in the RS test solution and the unexposed samples, which represents a decrease in the elasticity. In general, the stress relaxation curves show a nearly exponential decay of stress with time. A higher acid concentration contributes to a greater increase in modulus and a faster decay of stress in the first, which indicates a poorer sealing property.

Figure 5

Stress relaxation modulus E versus time curves.

To better understand the viscoelastic behavior of the silicone rubbers, a generalized Maxwell model shown in the Eq. 1 is adopted to describe the stress relaxation behavior of elastomeric materials. This model has multiple springs and dashpots. Because the components of the generalized Maxwell model are connected in parallel, all branches have the same strain at all times, and the overall stress of the system is the sum of the stresses in each branch (27,28):

where σ ∞  – stress after a long time, Pa; σ i  – stress, Pa; t – time, s; τ i  – a material constant, s.

The equation can also be rewritten as follows (Eq. 2):

where ε 0  – the strain.

When the Young’s modulus is normalized, it becomes Eq. 3, which is called the Prony series:

where g ( t ) = E ( t ) E 0 and g i = E i E 0 .

The normalized stress relaxation modulus, called g(t), versus time is shown in Figure 6. The curves displayed in Figure 6 show a similar trend. At the beginning, the modulus decreases sharply, indicating a quick relaxation of the stress. Afterwards, the modulus declines gradually over time. This test was stopped after 6 h due to the low relaxation rate. By comparing the two curves in Figure 6, it can be seen that the stress relaxes faster for samples from the ADT test solution than the samples from the RS test solution. Figure 6 shows the exponential decay nature of the stress with time similar to that described by the Maxwell model. It also shows that a higher temperature contributes to a faster stress relaxation. Here, a method similar to the TTS principle, which represents the equivalence between time and temperature, is applied to study the viscoelasticity of the materials and predict the lifetime. The equivalence of time and acidic concentration are discussed and this method is named Time-Acid Superposition (TAS) principle. The mathematical expression is shown in Eq. 1. A master curve is constructed by shifting the curves a certain amount, α T , horizontally to a particular reference test solution with acid concentration A ₀. Using the RS as the reference test solution, the curves of the samples in ADT solution with the acid concentration A are shifted to the reference test solution of RS in the logarithmic scale, as shown in Figure 7. The two curves overlap well. It can be seen that the time is longer than the actual experimental time. The stress relaxation master curve can be constructed using this principle of shifting to any reference test condition. The stress relaxation behavior for a long time period can be predicted in a short test time under a severe test condition by this method. The coefficients of the three-term Prony series have been fitted and listed in Table 2. In this experiment, the curve of samples exposed to the ADT test solution has been shifted by 0.898, fitted by the least square method, horizontally to the right on a logarithmic scale. The lifetime of the specimens is estimated to be about 27 h, while the test time under the ADT environment is only 6 h. The result points out that via the time-concentration superposition theory, the life extrapolating time is about 4 times of the experimental time at a harsh test environment. Cui et al. (26) studied the stress relaxation behavior of silicone rubbers aged in simulated PEM fuel cell environments and predicted a service life of approximately 6,000 h using the TTS principle. However, they also only studied the equivalence between time t and temperature T, without referring to the acid concentration. This work studied the equivalence between time and the acid concentration. The two curves overlap well via the time-concentration superposition principle. If the lifetime of seals compressed in a PEM fuel cell must be predicted, one can use the test data of the seals exposed to a severer test environment than its operating condition in a laboratory for a short time to construct the master curve. This is the advantage of equivalence between time and acid concentration. This work will provide a foundation for future studies, and this method will be investigated further.

where E(A, t) – stress relaxation modulus at time t and temperature T; A ₀, A – two different acid concentrations; t  – time; α T  – shift factor.

Figure 6

Normalized stress relaxation modulus for 2-week exposure samples at 70°C.

Figure 7

Master curve of stress relaxation for 2-week exposure samples at a reference solution of RS at 70°C.