1. Introduction
PEMFCs (proton exchange membrane fuel cells) are gaining attention as an eco-friendly energy device. In PEMFC, a chemical reaction between hydrogen and oxygen takes place at the MEA (membrane electrode assembly), which is the interface between the electrolyte and the electrodes. This reaction converts chemical energy into electrical energy through an electrochemical process, producing electricity.
When hydrogen gas, the fuel, is supplied to the anode, it is split into hydrogen ions and electrons. The negatively charged electrons flow through an external circuit to generate electricity, while the positively charged hydrogen ions pass through the electrolyte membrane to the cathode, where they react with oxygen to produce water (H2O)[1,2].
Generally, the membrane electrode assembly (MEA) consists of a gas diffusion layer (GDL) and a catalyst layer (CL) on both the cathode and anode sides, with a proton exchange membrane in between. The CL, one of the components of the MEA, is a critical element that determines the performance of the proton exchange membrane fuel cell (PEMFC) as it acts as the electrode. It is composed of a carbon support, catalyst Pt/C (platinum on carbon), and ionomer. The CL is manufactured by mixing the Pt/C catalyst, ionomer, and a liquid solvent to create a catalyst ink, resulting in a porous structure comprising platinum nanoparticles supported on carbon and a thin ionomer film[3-5].
While PEMFCs (proton exchange membrane fuel cells) have many advantages, they also have the disadvantage of high costs, making it important to address this issue. Among the material costs of PEMFCs, the membrane electrode assembly (MEA) accounts for a significant portion. Specifically, Pt/C makes up 41% of the material costs of PEMFCs. Therefore, recycling Pt could lead to cost reductions in materials[ 6-9].
As of now, the methods researched for recovering Pt from waste membrane electrode assemblies (MEAs) in PEMFCs can be categorized into four types: 1) High-temperature combustion process; 2) acid dissolution process; 3) electrochemical process; and 4) alcohol treatment process.
Firstly, the high-temperature combustion process requires a lot of energy due to the high temperatures and releases toxic HF. Secondly, the acid dissolution process uses highly corrosive acids, necessitating special infrastructure, and emits hazardous substances such as HCl vapor and Cl2, as well as NOx. Thirdly, the electrochemical process uses corrosive or toxic electrolytes, resulting in the complete loss of the proton exchange membrane. Finally, the alcohol treatment process offers an environmentally friendly way to recover the membrane but has the disadvantage of recovering Pt in large particle sizes. Excluding the alcohol treatment process, all the other methods release harmful substances during the processing. Currently, since high-temperature combustion and acid dissolution processes are commonly used, research into environmentally friendly Pt recycling methods is essential[10,11].
For example, Sharma et al. aimed to recycle materials from used MEA using the alcohol treatment process. They separated the membrane and Pt from the waste MEA using ethanol and dissolved Pt with HCl, subsequently embedding Pt onto carbon to generate Pt/C, which was then recycled as a catalyst[12].
Xu et al. attempted to recycle materials from waste MEA using sulfuric acid. They submerged the catalyst coated membrane (CCM)- type MEA in sulfuric acid and separated Pt and ionomer solutions through centrifugation. The ionomer solution was then used to manufacture a membrane, and the amorphous carbon nanoparticles acting as the catalyst layer were oxidized to recover Pt[13].
This study differs from the aforementioned literature by recycling the catalyst in the form of Pt/C rather than Pt, resulting in a simple process without a synthesis step, which also saves costs associated with synthesis. In this research, our main goal was to recover Pt/C through an environmentally friendly method that does not emit harmful substances that would contaminate the environment during the recycling process of the recovered MEA from used PEMFCs. Furthermore, we intended to regenerate the Pt/C by removing the ionomer contained in it. We investigated the properties of the regenerated Pt/C through evaluations of electrode performance and analyses using TGA, SEM-EDS, TEM, FT-IR, and XRD.
2. Experimental
The proton exchange membrane used for MEA manufacturing was Nafion 211 (EW 1100, thickness 0.05 mm) from Du Pont, and the Gas Diffusion Layers (GDL) were Sigracet 22 BB (thickness 215 μm) from SGL Carbon. The electrolyte solution utilized was a 5 wt% Nafion solution from Chemours in a sprayed state. The catalyst Pt/C used for standard MEA manufacturing was 40% Pt on Vulcan XC-72 from Premetek. The fuel gases used were 99.9% H2 and air.
The reagents included ethanol (C2H5OH), 1-butanol (C4H10O), IPA (isopropyl alcohol, C3H8O), DMSO (dimethyl sulfoxide, C2H6OS), chloroform (CHCl3), hydrogen peroxide (H2O2) (99.5%, Samchun Pure Chemical Co., Ltd., Pyeongtaek, Gyeonggi-do, Republic of Korea), and sulphuric acid (H2SO4, Matsunoen) were used as received. Water consisting of triple-distilled water was supplied from our laboratory.
A thermogravimetric analyzer (TGA, Q50, TA Instrument, Newcastle, DE, USA) was used to analyze the presence of ionomers in Pt/C after the process, which was heated in an air atmosphere from room temperature to 800 °C at a rate of 10 °C/min. A transmission electron microscope (TEM, JEOL; JEM-ARM200F) was performed to analyze the distribution of Pt/C with a resolution of 0.08 nm and a maximum acceleration voltage of 200 kV. X-ray Diffraction (XRD, Rigaku; Smart Lab SE) was employed for the analysis of Pt particle size and crystallinity, using Cu Kα X-rays with a maximum output of 2.2 kW. Fourier transform infrared spectroscopy (FT-IR, Thermo Fisher Scientific; Nicolet iS50) was conducted to analyze the presence of polymer bonds, recorded over a wavenumber range of 399~3999 cm-1 with a resolution of 16 cm-1 and 64 scan repetitions.
After separating the GDL from the used MEA, it was cut into a size of (8 × 8) cm2 and placed in a water: IPA (5:5) solvent. It was then subjected to ultrasonic treatment for 10 minutes to separate the catalyst Pt/C from the proton exchange membrane. After that, the membrane was removed from the solution, and the solution was dried in an oven at 60 °C to retrieve the Pt/C. At this stage, the separation principle is similar to the polymer chain swelling process illustrated in Figure 1. When the MEA is placed in the solvent, solvent molecules penetrate the polymer chains of the polymer membrane, causing it to swell. While the polymer membrane is deformed by the swelling from the solvent, the catalyst layer remains unchanged, leading to the separation of the catalyst layer and the polymer membrane.
The recovered used Pt/C from the separation process had the ionomer that was removed using two methods. The first method involved the thermal treatment for ionomer removal, where the separated used Pt/C was heat-treated at 400, 500, and 600 °C for 1 hour in a nitrogen gas atmosphere to remove residual ionomer. The second method involved ionomer dissolution for removal. In this method, the separated used Pt/C and solvent were added to an autoclave, which was then placed in a box furnace and heated at the investigated temperature for 1 hour. Afterwards, the ionomer solution was separated from Pt/C through vacuum filtering to eliminate any remaining ionomer.
The MEA manufacturing and fuel cell performance testing followed the methods outlined in previous studies. First, the proton exchange membrane (Nafion 211) underwent pretreatment to remove impurities. Initially, it was treated for 1 hour in a 5 wt% H2O2 solution at 80 °C to eliminate organic impurities, followed by washing with distilled water at 80 °C for 1 hour. Then, to remove metal ions, it was treated for another hour in a 0.5 M H2SO4 solution at 80 °C and washed again with distilled water at the same temperature[14-16].
The process for manufacturing the MEA is as follows: Pt/C, isopropyl alcohol (hereafter referred to as IPA), and distilled water were placed in a vial and subjected to ultrasonic treatment for 10 minutes, maintaining a ratio of 6:4 for distilled water and IPA. Afterwards, a 5% Nafion solution was added to the vial, and it was stirred for 10 minutes. The ratio of Pt/C to ionomer was 8:2 at this stage. The resulting catalyst ink was sprayed to coat the catalyst layer on the membrane for the cathode and on the GDL for the anode. The electrode size was set at 25 cm2, and the Pt loading was 0.3 mg/cm2. After drying the catalyst coated GDL and the membrane, they were hot-pressed at 120 °C, 1 m, and 150 kgf/cm2 to form the MEA. The completed MEA was then installed into the cell, overlaying the GDL on the cathode.
For performance evaluation, hydrogen (H2) was used as the anode gas and air as the cathode gas. The humidification temperatures were set at 80 °C for the cell, 90 °C for the anode, and 70 °C for the cathode. The gas temperatures were set to be 2 °C higher than the humidification temperatures, at 92 °C and 72 °C, respectively. The flow rates of hydrogen and air were set to exceed the stoichiometric coefficients of the current by 1.5 and 3.0 times, respectively. Voltage values were measured against the current, and the test was stopped when the voltage reached 0.3 V.
3. Results and discussion
To select a solvent for the separation process, six different solvents were used. Figure 2 shows the changes in the used MEA after the separation process with each solvent. The solvents employed included ethanol, butanol, water, IPA, chloroform, and DMSO. As a result, water, chloroform, and DMSO did not facilitate the separation of Pt/C and the membrane, while butanol, IPA, and ethanol made the separation of Pt/C from the membrane easier. Swelling of polymer materials is the process of penetration of solvent molecules into a polymer matrix causing a change in the volume. Swelling can be seen as an equilibrium between the entropy of polymer chains and the enthalpy of mixing. The reason of different swelling ratios in different solvents are beyond of this work.
The membrane experiences swelling due to the interaction with the solvent and subsequently goes into contraction during the drying process. To confirm this, the dimensional reduction of membrane areas for each solvent are shown in Figure 3. When the solvent penetrates the polymer membrane, solvent molecules get interspersed among the polymer chains, causing swelling. As the solvent dries, the surface tension pulls the polymer chains, leading to a decrease in the membrane area. As illustrated in Figure 3, the dimensional reduction was in the order of ethanol (EtOH), butanol (BuOH), IPA, DMSO, water, and chloroform, showing the largest values first. Since DMSO, water, and chloroform did not allow for the separation of the catalyst Pt/C from the membrane, they were excluded. Later, to allow for the possibility of regenerating the membrane, IPA was selected as the solvent because it easily separated the catalyst Pt/C from the membrane while causing the least reduction in the membrane area.
Since the recovered Pt/C contains an ionomer, the ionomer was removed using both solvent treatment and heat treatment processes. First, the solvent treatment was carried out by increasing the temperature from 140 to 250 °C, and additionally, the process was continued at a lower temperature (100 °C) for an extended time. As a result, the ionomer removal rate was maintained at similar values around 160~170 °C, leading to the conclusion that 160 °C is the minimum temperature for ionomer solubilization. The ionomer removal rate showed values between 61~63% when the temperature was above 160 °C. Figure 4 reveals the average values of ionomer removal rates from four experiments.
The heat treatment process utilized a tube furnace in a nitrogen atmosphere, gradually increasing the temperature from 400 to 600 °C. The results showed similar ionomer removal rates of 65~66% in the temperature range of 500~600 °C. The ionomer removal rates shown in Figure 4 are the average values from two experimental data sets.
To analyze the presence or absence of the ionomer, TGA (thermogravimetric analysis) was conducted. In Figure 5, it can be observed that in the case of the solvent treatment process (Figure 5 (b)), there is a weight loss around 300~400 °C, which is different from the commercial Pt/C shown in Figure 5 (a). This indicates the presence of the ionomer and suggests that the ionomer is not completely removed by the solvent treatment[17]. In the case of the heat treatment process shown in Figure 5 (c), no weight loss is observed around 300~400 °C. Additionally, in Figure 5 (a), the TGA profile of the commercial 40% Pt/C + ionomer after heat treatment differs from that after the solvent treatment. This is likely due to the ionomer being of a different type.
FT-IR analysis was conducted to examine the presence of polymer bonding. Figure 6 (a) shows the FT-IR graphs of samples subjected to both heat treatment and solvent treatment in a simulation sample where the ionomer was applied to the commercial 40% Pt/C. For the simulation sample, it can be confirmed that the F-C-F bond has disappeared after both the heat treatment and the solvent treatment. In Figure 6 (b) and (c), the untreated spent catalyst shows a peak at 1261 cm⁻1 indicating an F-C-F bond[18,19]. In the case of heat treatment, this bond disappeared at all temperatures except for 400 °C; however, in the solvent treatment, the bond remained.
XRD analysis was conducted to analyze the particle size and crystallinity of the recovered Pt/C. Figure 7 presents the XRD analysis graphs of the standard Pt/C as well as before and after the pre-treatment. The results of the XRD analysis confirm that the Nafion peak was eliminated after the pre-treatment of the recovered catalyst. However, this does not mean that the ionomer has been completely removed; it is possible that the ionomer remains in a polymeric form rather than as Nafion[20].
Table. 1 expresses the results of the XRD analysis conducted to determine the size of Pt particles. It was confirmed that the particle size of Pt increased for both the solvent treatment and the thermal treatment. In the case of the solvent treatment, compared to the Pt/C before pre-treatment, the particle size increased by 20% at 100 °C (3 h), 21% at 160 °C, 16% at 200 °C, and 33% at 250 °C. For the thermal treatment, the particle size growth was 26% at 400 °C, 22% at 500 °C, and 49% at 600 °C.
For comparison, after applying 30% ionomer to the commercial Pt/C and then conducting a solvent treatment at 160 °C, the Pt particle size increased by 66% compared to the commercial Pt/C. The Pt/C that underwent thermal treatment at 600 °C exhibited a 267% increase. The size of Pt particles increased in both of the Pt/C regeneration methods; however, the rate of increase was higher in the thermal treatment process. Additionally, it was confirmed that the particle size of Pt in the spent Pt/C before pre-treatment had increased by 9% compared to that of commercial Pt/C. This may be due to degradation or the effects of the alcohol process.
TEM was conducted to analyze the distribution of Pt/C. Figure 8 shows the TEM images of Pt/C before and after the ionomer removal process. The black areas represent the Pt particles, and it was confirmed that the Pt was relatively evenly distributed on the carbon for all samples.
To analyze the surface and elemental composition of Pt/C, SEM-EDS was performed. Figure 9 presents the SEM images of Pt/C before and after the ionomer removal process, while Table 2 summarizes the EDS analysis results in tabular form. No significant structural differences were observed in the SEM images. The EDS analysis indicated that the fluorine weight percentage in the solvent-treated samples was 2-5%, while in the thermally treated samples, it was less than 1% except for the sample treated at 400 °C. This suggests that the ionomer remained in the sample treated with solvent and was not completely removed.
Figure 10 is a graph displaying the electrode performance of MEA manufactured with commercial 40% Pt/C compared to MEA made with solvent-treated and thermally treated Pt/C. The solvent-treated MEA reached a performance of 59.4% at 0.5 V and 60.7% at 0.3 V compared to the commercial catalyst. In contrast, the thermally treated MEA showed a performance of 14.1% at 0.5 V and 26.8% at 0.3 V relative to the commercial catalyst. When comparing the two values, the solution treatment performed 45.3% better at 0.5 V and 33.9% better at 0.3 V than the thermal treatment. It is believed that the thermal treatment catalyst exhibited lower performance due to an increase in the Pt particle size caused by the high temperature, which reduced the active surface area of Pt. Ionomer reside was not clearly eliminated at 400 °C by the results of EDS in Table 2. Additionally, the decrease in performance of the MEA at 250°C during solvent treatment is attributed to larger Pt particle sizes compared to those at 160 °C, 100 °C (3 h), and 200 °C.
Furthermore, MEAs were manufactured from a mimic sample, coated with ionomer on commercial Pt/C, and subjected to solvent treatment and thermal treatment before testing. The solvent-treated mimic sample MEA achieved a performance of 84.4% at 0.5 V and 82.1% at 0.3 V compared to the commercial catalyst, while the thermally treated MEA reached 54.5% at 0.5 V and 43.8% at 0.3 V relative to the commercial catalyst. The solvent-treated process of the mimic sample achieved a performance of 84.4% compared to the commercial catalyst, demonstrating the potential for environmentally friendly catalyst reuse.
5. Conclusions
Used MEA of PEMFC was environmentally treated to recover Pt/C. The recovered Pt/C was remanufactured into MEA, and its performance was tested and analyzed through various methods. First, we separated the membrane and Pt/C from the used MEA using isopropyl alcohol (IPA), and we conducted solvent (IPA) treatment and thermal treatment in N2 atmosphere to remove the ionomer from the separated Pt/C. As a result, the solvent treatment resulted in an ionomer removal rate of 61-63%, while the thermal treatment showed a removal rate of 65-66%. The solvent treatment indicated a weight loss at 300-400 °C in the TGA graph, unlike the commercial catalyst, and the presence of F-C-F bonds was confirmed in the FT-IR graph. Additionally, EDS analysis revealed that the weight percent of fluorine in the solvent treatment remained at 2~5%, suggesting that the ionomer was not completely removed. Using the recovered Pt/C, we remanufactured MEA and conducted performance evaluations. The solvent treatment achieved a performance of 60.7% compared to the commercial catalyst. The thermal treatment exhibited a performance that was 45.3% lower than the solvent treatment, which is believed to be due to an increase in Pt particle size at high temperatures that reduced the active surface area of Pt. For further comparison, we performed solvent treatment on a mimic sample (commercial 40% Pt/C + ionomer) and evaluated the performance of the MEA manufactured from it, achieving 84.4% relative to the commercial catalyst, thereby demonstrating the potential for environmentally friendly catalyst reuse. The recovery or recycling process of Pt/C from used PEMFC stacks in the industry is not environmentally friendly, as it uses toxic solvents and leads to the evolution of carbon dioxide. Therefore, this eco-friendly solvent treatment could offer environmental advantages in the recycling industry.
