Introduction

A polymer electrolyte membrane fuel cell (PEMFC) is environmentally friendly while maintaining high energy conversion efficiency at low temperatures [1,2,3]. Because it can produce a wide range of output, it has various applications; however, its market expansion is inhibited by its high price and short lifespan [4, 5]. PEMFCs require durability of 5000–40,000 h depending on the field of application; however, this goal has not been achieved owing to the deterioration of the membrane electrode assembly (MEA) triggered by long-time operations [6].

Degradation of polymer membranes is largely classified into electrochemical and mechanical degradation [7]. Electrochemical degradation occurs when radicals generated within the cell attack the polymer membrane, causing the membrane's thickness to thin and forming pinholes or cracks, which trigger the degradation of the polymer membrane. Mechanical degradation is caused by tearing of the polymer membrane under mechanical stress as the polymer membrane is repeatedly contracted and expanded by humidification (wet) and drying (dry) [8, 9]. To reduce the durability evaluation time of polymer membranes, the U.S. Department of Energy (DOE) and the Japan New Energy Industry Technology Development Organization (NEDO) adopt an accelerated stress test (AST) technique. Electrochemical AST, which utilizes the open circuit voltage (OCV) holding method under high-temperature, low-humidity conditions with a high radical generation rate [10], and mechanical AST, which employs the wet-dry cycle method, are employed to evaluate the durability of polymer membranes [11].

In general, electrochemical and mechanical ASTs were evaluated separately; however, during actual operation, the two degradations occurred simultaneously. Therefore, DOE proposed a novel evaluation protocol in 2016 to evaluate the durability of polymer membranes similar to actual vehicle driving conditions [12]. This novel AST is a method for repeating wet-dry in OCV state by supplying H2 to the anode and air to the cathode at high temperature. As an advantage, it can simultaneously evaluate the electrochemical and mechanical durability of the polymer membrane. However, during the AST process, as wet and dry states are repeated, changes in OCV voltage emerge. This is similar to the electrode catalyst degradation protocol and can lead to electrode catalyst degradation. In addition, electrode catalyst degradation is significantly influenced by the number of voltage change cycles. The DOE AST employs a short cycle time for wet (45 s)-dry (30 s) conditions, which can accelerate electrode catalyst degradation owing to a high number of voltage alterations [13, 14].

When evaluating Nafion XL (reinforced membrane) using the DOE AST protocol, the time spent was excessive, up to 408 h, and more electrode catalyst degradation occurred than polymer membrane degradation. Therefore, the durability evaluation of the MEA can be terminated due to electrode catalyst degradation [15]. Furthermore, because electrochemical and mechanical durability are evaluated simultaneously, it is challenging to determine whether the durability of the polymer membrane is electrochemically or mechanically weaker.

Here, we attempt to develop an improved AST protocol based on the polymer membrane DOE AST protocol, with the objective of reducing the durability evaluation time, mitigating the impact of electrode catalyst degradation, and determining the cause of degradation during electrochemical/mechanical durability evaluations.

Experimental

Cell Preparation

The polymer membrane employed in this experiment is a 27.5-μm-thick Nafion XL, including a reinforced membrane with an expanded-polytetrafluoroethylene (e-PTFE) support between the ionomer layers. An MEA with a 25-cm2 electrode area was manufactured by coating Pt/C electrode (BIMaterials Co., Korea) particles on both sides using the decal method. The Pt loading amount is 0.4 mg/cm2. The MEA and gas diffusion layer (GDL, SGL 39BB, SGL Carbon Co., Wiesbaden, Germany) were assembled in a unit cell with 80 torque. The ratio of the flow field area to the gas diffusion layer in the bipolar plates (separator) can affect the permeability of the gas through the polymer membrane. In this experiment, a separator with a flow field area ratio of 66.7% was employed. The cell temperature, flow rate, and relative humidity (RH) were controlled using a station (CNL energy Co., Korea).

AST Method

The 2016 DOE AST protocol, which combines the electrochemical/mechanical deterioration of polymer membranes, supplies anode H2 and cathode air at 1 L/min each at 90 °C and then proceeds with wet (100% RH, 45 s)-dry (0% RH, 30 s) conditions and 1.0 bar in the OCV state [16]. To improve the DOE AST protocol and reduce the evaluation time of the polymer membrane, oxygen was supplied to cathode. To mitigate the impact of electrode catalyst degradation, the evaluation was conducted by increasing the wet and dry time to 60 s and 300 s, respectively, thereby reducing the number of voltage change cycles. To distinguish between electrochemical and mechanical durability evaluations, the electrochemical degradation, OCV holding, was conducted for 72 h before proceeding with the electrochemical/mechanical degradation, OCV/Wet-Dry cycle. Each step of OCV/wet-dry cycling and OCV holding was repeated for 12 h until the durability evaluation was completed (Fig. 1). At this point, the durability evaluation was terminated when the hydrogen crossover current density (HCCD) increased by more than 15 mA/cm2. The method developed by improving DOE AST was named Sunchon National University (SCNU) AST.

Fig. 1
figure 1

AST protocol for experimental process a DOE AST, b SCNU AST

Full size image

Cell Characterization Analysis

Here, all performance evaluations were conducted at 70 °C and 100% RH. After activating the MEA, I-V curves were measured under conditions of anode H2 1.5 stoichiometry and cathode air 2.0 stoichiometry [17].

The HCCD was measured using a potentiostat (Solatron, SI 1287) and the DOE linear sweep voltammetry (LSV). DOE LSV measured current in the voltage range of 0.05–0.4 V at a scan rate of 1 mV/s while supplying anode H2 and cathode N2 at 40 mL/min and 200 mL/min, respectively. The HCCD chose a current density value of 0.3 V, which is not affected by the limit current density (0.30–0.35 V).

Short resistance (SR) was measured using a potentiostat via the NEDO LSV method. NEDO LSV measured current in the voltage range of 0.2–0.5 V at a scan rate of 0.5 mV/s while supplying anode H2 and cathode N2 at 200 mL/min and 200 mL/min, respectively [18]. SR was calculated as the reciprocal of the slope by extrapolating a straight line of 0.4–0.5 V.

The electrochemical surface area (ECSA) was measured using a potentiostat via cyclic voltammetry (CV). CV involved injecting gas in the same manner as DOE LSV, measuring current while varying the voltage at a scan rate of 30 mV/s, and selecting the value measured after 14 cycles.

High frequency resistance (HFR) and charge transfer resistance (CTR) were measured using an impedance analyzer (Solatron, SI 1260) via an impedance method. Impedance was measured by supplying anode H2 93 mL/min and cathode air 296 mL/min, and DC current 1 A and AC amplitude 100 mA in the frequency range of 100,000–0.1 Hz.

To verify the thickness change of the polymer membrane and electrode layer due to deterioration after AST, scanning electron microscope (SEM, JSE-7100F, Jeol, Japan) was utilized, and the beam potential was 5 kV.

Results and Discussion

OCV Change During AST

Figure 2 illustrates the OCV changes during the AST process. The voltage is increased in the middle of the OCV graph in the AST process because of the point at which activation was performed to evaluate the degree of MEA degradation, which is a result of the MEA performance being restored by the activation effect. In the OCV graph, the upper line (upper limit line, ULL) represents the OCV when dry and the lower line (lower limit line, ILL) is humidified. In the DOE AST, up to 216 h, the OCV gradually decreased while the variation ranges of the ULL and ILL was consistently maintained. Subsequently, the variation ranges gradually increased, and at 408 h (19,581 cycles), the ILL decreased to 0.70 V, thereby terminating the durability evaluation. SCNU AST exhibited an initial OCV of 1.0 V or more by supplying oxygen to the cathode to shorten the evaluation time, and the OCV decreased by approximately 0.15 V from the initial value at 72 h of OCV holding, which is an electrochemical durability evaluation method. Because OCV/wet-dry and OCV holding were repeated for 12 h each, the overall OCV decreased uniformly; however, the ILL decreased to 0.40 V at 127 h (310 cycles), and the durability evaluation was completed. In addition, it can be noted that significant degradation of the polymer membrane occurred as evidenced by the voltage of the ILL at the termination point of the durability evaluation.

Fig. 2
figure 2

Change in OCV during chemical/mechanical cycle at AST protocol conditions a DOE AST, b SCNU AST

Full size image

The durability evaluation ended at 408 h for the DOE AST and 127 h for the SCNU AST, reducing the evaluation time by approximately 1/3. Evidently, the decrease in OCV was more significant during mechanical degradation than electrochemical degradation, thus indicating that the influence of mechanical durability on the polymer membrane was more significant than its electrochemical durability.

Hydrogen Permeability and Short Resistance Change During AST

Fig. 3 illustrates the DOE LSV and HCCD changes during the AST process. In the DOE AST, HCCD decreased from an initial value of 1.084 mA/cm2 to 0.978 mA/cm2 at 360 h. However, it increased significantly to 39.68 mA/cm2 at 408 h, exceeding the durability evaluation termination criterion of 15 mA/cm2, thereby terminating the durability evaluation. In the SCNU AST, there were negligible changes up to 120 h; however, at 127 h, HCCD increased abruptly to 23.05 mA/cm2, thereby terminating the durability evaluation. Therefore, at the termination point of the durability evaluation, it was verified that the MEA degradation rate in SCNU AST was accelerated by more than three times compared to DOE AST, thus resulting in a shortened evaluation time. In addition, a correlation was observed between the rapid decrease in OCV and the significant increase in HCCD during the AST process.

Fig. 3
figure 3

Change in LSV by DOE method and HCCD during chemical/mechanical cycle at AST protocol conditions a DOE AST, b SCNU AST, c HCCD comparison

Full size image

Fig. 4 illustrates the NEDO LSV and SR changes during the AST process. Therefore, reaching below the DOE durability evaluation termination criterion of 1 kΩcm2, the evaluation was also concluded based on the SR criterion. As the polymer membrane undergoes degradation, resulting in a thinning of the membrane thickness and a reduction in the distance between the electrodes, it can be observed that the SR decreases, leading to shorting [19]. In addition, the change in SR exhibited the same trend as the change in HCCD, cross-confirming that the polymer membrane degradation rate of SCNU AST was accelerated compared to that of DOE AST.

Fig. 4
figure 4

Change in NEDO　LSV and short resistance during chemical/mechanical cycle at AST protocol conditions a DOE AST, b SCNU AST, c Comparison of short resistance

Full size image

ECSA and Resistance Change During AST

Figure 5 illustrates the CV and ECSA change during the AST process. At the termination point of the durability evaluation, the ECSA reduction rates for DOE AST and SCNU AST were 96% and 30%, respectively, illustrating a significantly larger difference of over three times. This is attributed to the degradation of the electrode catalyst due to the voltage changes during wet and dry conditions in the AST process [11]. DOE AST had 19 581 cycles at 408 h, while SCNU AST had 310 cycles at 127 h, thus indicating that SCNU AST, which distinguished the evaluation method, proceeded with significantly fewer voltage change cycles. Therefore, it was confirmed that the impact of electrode degradation was reduced in the SCNU AST supplemented with the number of voltage changes in which electrode degradation occurred.

Fig. 5
figure 5

Change in CV and ECSA during chemical/mechanical cycles at AST protocol conditions a DOE AST, b SCNU AST, c Comparison of ECSA

Full size image

Figure 6 presents the impedance, HFR, and CTR change during the AST process. The HFR increased by 20.5% in DOE AST and 54.6% in SCNU AST at the termination point of the durability evaluation owing to the progression of polymer membrane degradation. This result indicates that SCNU AST with increased drying time was more affected by polymer membrane degradation. CTR increased by 19.7% in DOE AST and 27.9% in SCNU AST, thus indicating simultaneous degradation in both the polymer membrane and electrode. In addition, DOE AST exhibited a similar increase rate of HFR and CTR, while SCNU AST exhibited an increase rate of HFR approximately twice that of CTR, confirming that polymer membrane degradation occurred more than electrode catalyst degradation.

Fig. 6
figure 6

Change in impedance, HFR, and CTR during chemical/mechanical cycle at AST protocol conditions a DOE AST, b SCNU AST, c Comparison of HFR and CTR

Full size image

Change in Polymer Membrane and Electrode Thickness After AST

Figure 7 compares the thickness alterations of the polymer membrane and the electrode layer via the SEM cross-sectional analysis after the AST process. MEA comprises an anode electrode in the upper layer, a polymer membrane in the middle layer, and a cathode electrode in the lower layer.

Fig. 7
figure 7

Comparison of SEM image of MEA after chemical/mechanical cycle at AST protocol conditions a DOE AST, b SCNU AST

Full size image

In the SEM cross-sectional image, the thickness of the cathode electrode layer decreased by 15.4% for DOE AST (6.6 μm) and 1.3% for SCNU AST (7.7 μm) from the initial 7.8 μm, thus confirming that the electrode deterioration in DOE AST was significant. As voltage variations are repeated, Pt dissolves and is discharged into the polymer membrane and cathode effluent, thereby reducing the electrode thickness [20]. Because the number of voltage changes under wet-dry conditions in the AST process is higher for DOE AST than for SCNU AST, it can be observed that significant electrode deterioration occurred.

The polymer membrane thickness decreased by 13.1% in DOE AST (23.9 μm) and by 18.2% in SCNU AST (22.5 μm) from the initial 27.5 μm. Hence, the polymer membrane degradation occurred more significantly in SCNU AST. Consequently, the drying time in SCNU AST increases, polymer membrane degradation occurs more in SCNU AST than in DOE AST due to the formation of low humidification conditions in which polymer membrane degradation is accelerated [6]. Therefore, SCNU AST confirmed that polymer membrane degradation occurred more than electrode catalyst degradation.

Conclusion

The results of the SCNU AST, which addressed the challenges of the polymer-membrane electrochemical/mechanical AST developed by DOE, are summarized as follows.

  1. (1)

    The evaluation of Nafion XL with the DOE AST required a long duration of 408 h, and electrode catalyst degradation occurred due to voltage variations from OCV/Wet-Dry cycles. In addition, it was difficult to determine whether the durability of the polymer membrane is electrochemically or mechanically weak.

  2. (2)

    The SCNU AST, which improved the DOE AST, supplied oxygen to the cathode instead of air and increased the drying time under low humidity conditions to accelerate the degradation rate of the polymer membrane. In addition, separate electrochemical and mechanical evaluation methods were developed.

  3. (3)

    In SCNU AST, the rate of radical generation increased as the high oxygen pressure and drying time increased, and the evaluation was completed within 127 h, thereby reducing the time required by more than 1/3.

  4. (4)

    By extending the time of one cycle to 60-s wet and 300-s dry conditions and distinguishing between electrochemical and mechanical evaluation methods, the number of voltage changes was reduced, thereby mitigating the impact of electrode catalyst degradation.

  5. (5)

    When the electrochemical/mechanical evaluation methods were distinguished in the polymer membrane, the OCV reduction was more significant during mechanical degradation than during electrochemical degradation. It was confirmed that the polymer membrane is more influenced by mechanical durability than electrochemical durability.