1 Introduction
Proton exchange membrane fuel cells (PEMFCs) are efficient and clean electrochemical energy conversion devices that utilize renewable hydrogen energy [
1]. In the hydrogen economy, fuel cell vehicles (FCVs) are regarded as a key solution for low-carbon transportation. When hydrogen is produced from renewable energy sources, greenhouse gas emissions are expected to be reduced to near zero [
2,
3]. However, the performance, cost, and durability of PEMFC stacks greatly affect the large-scale commercialization of FCVs. The U.S. Department of Energy (DOE) has set ambitious cost and durability targets for FCV systems. For light-duty vehicles (LDVs), the system must achieve over 8000 h of stable operation, while for heavy-duty vehicles (HDVs), the target exceeds 25000 h by 2030 [
4]. Additionally, cost targets of 30 $/kW for LDVs and 60 $/kW for HDVs must be met at an annual production scale of 100000 systems. Within the stack, platinum (Pt) catalysts alone account for approximately 40% of the total cost at a production volume of 500000 per year [
5]. Therefore, meeting performance and durability requirements while decreasing platinum group metal (PGM) loading remains a fundamental challenge for the widespread commercial viability of fuel cells [
6].
Traditionally, a catalytic layer (CL) is composed of carbon supports, PGM catalysts, and ionomers [
7]. At high power densities, a low Pt loading results in a reduced reaction surface area and high reaction rates per unit area. This condition demands optimized mass transfer pathways to achieve efficient gas-liquid transport and minimize mass transfer losses, especially in the cathode of PEMFCs, where the oxygen reduction reaction (ORR) exhibits sluggish kinetics [
4]. Furthermore, high surface current densities accelerate Pt corrosion and migration, necessitating strong bonding between the carbon support and the PGM catalyst [
8]. Extensive research has focused on refining ink formulations to optimize mass transfer in the CL, including adjustments to the solution type [
9], ionomer concentration [
10], dispersion methods [
11], and drying temperatures [
12]. Additionally, advanced techniques allow for precise control of catalyst and ionomer distribution to form patterned CLs and membranes, thereby improving catalyst utilization efficiency [
13,
14]. Despite these advances, oxygen transport often remains hindered due to the random and tortuous paths caused by carbon support stacking. Moreover, durability improvements are still limited, largely due to the insufficient corrosion resistance of traditional carbon supports.
An ideal carbon support should exhibit a high graphitization for corrosion resistance, a high surface area to ensure uniform PGM nanoparticle dispersion, and a regular morphology to establish an ordered mass transport network in low-Pt loaded CLs [
4,
15]. These characteristics contribute to maximizing the utilization of highly active catalysts. Beyond surface modification of conventional carbons [
16,
17], a variety of specially shaped carbon supports, including mesoporous carbon [
18], graphene [
19], carbon nanotubes (CNTs) [
20], and carbon nanofibers (CNFs) [
21] have been developed to tailor CL microstructures. Furthermore, as a subclass of metal-organic frameworks (MOFs), 2-methylimidazole zinc salt (ZIF-8) can be used to generate transition metal and nitrogen co-doped carbons (M-N-C) through calcination [
22]. ZIF-8-derived carbons offer high surface area, tunable porosity, and excellent electrical conductivity. When mass production processes incorporate solvent recycling, low-cost solvents, and concentrated formulations, ZIF-8-derived carbons can achieve competitive pricing compared to other candidates. Therefore, they are considered promising carbon supports for low-Pt loaded CLs following controlled structural growth [
23].
In this paper, a tree-like nitrogen-doped carbon support (T-NC) is reported, developed using low-cost raw materials, operating under mild conditions with low energy consumption, and achieving the desired structure and functionality through a simple multi-step process. When applied at the cathode with a Pt loading of 0.1 mgPt/cm2, this design significantly reduces mass transfer losses and increases the durability of PEMFCs.
As illustrated in Fig. 1, acid washing increases the concentration of oxygen-containing functional groups and carbon defects on multi-walled carbon nanotubes (MWCNTs). Introducing a small amount of pretreated MWCNTs into the precursor solution facilitates the nucleation and growth of ZIF-8 on their surfaces. During high-temperature calcination under inert gas flow, zinc (Zn) is evaporated, introducing porosity, and resulting in the formation of a unique tree-like carbon morphology. In this structure, MWCNT serves as the conductive backbone, while ZIF-8-derived carbon acts as the branches to offer nitrogen (N) coordination and additional porosity.
The T-NC architecture is universally applicable for supporting various advanced PGM catalysts. CLs containing T-NC have abundant pore space, facilitating improved reactant supply and water removal. Additionally, the structure exhibits enhanced corrosion resistance and benefits from an ordered conductive network formed by highly graphitized MWCNTs. This strategy presents a promising solution to overcome the intrinsic limitations of conventional carbon supports, leading to significant improvements in both the performance and durability of low-Pt loaded fuel cells.
2 Experimental materials and methods
2.1 Chemical materials
Polyvinyl pyrrolidone (PVP, MW = 58000), isopropanol (IPA), and 2-methyl imidazole (2-MIM) were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. Analytical-grade zinc nitrate hexahydrate (Zn(NO3)2.6H2O) was obtained from Sigma-Aldrich (Shanghai) Trading Co., Ltd. MWCNTs were sourced from Chengdu Organic Chemicals Co., Ltd. Ethylene glycol (EG) and chloroplatinic acid hexahydrate (H2PtCl6·6H2O) were obtained from Shanghai Macklin Biochemical Co., Ltd. The commercial Pt/C catalyst (TEC10E50E, 50 wt.%) was supplied by TANAKA. The 5 wt.% Nafion solution (D520) was purchased from Dupont and the membrane (M775.15, 15 μm thickness) was provided by Gore.
2.2 Preparation of ZIF-8, ZIF-8@MWCNT, and T-NC precursor
The highest product formation rate was observed when the molar ratio of Zn(NO3)2·6H2O to 2-MIM was 1:4 (Fig. S1 in Electronic Supplementary Material). After ultrasound treatment for 20 min at room temperature, Zn(NO3)2·6H2O (2.975 g) and PVP (1.16 g) were dissolved in 100 mL of methanol. Separately, 2-MIM (3.284 g) was dissolved in another 100 mL of methanol. The Zn2+ solution was subsequently injected into the 2-MIM solution. After stirring for 30 min, the resulting mixture was transferred to a water bath at 30 °C and remained for 48 h. The high temperature resulted in particle cluster growth and agglomeration, as shown in Fig. S2. The supernatant was discarded, and the remaining solid was collected by centrifugation, washed with methanol several times, and dried under vacuum at 60 °C overnight to obtain ZIF-8.
For the preparation of ZIF-8@MWCNT, a similar procedure was followed. First, Zn(NO3)2·6H2O (2.975 g) and PVP (1.16 g) were completely dissolved in 100 mL of methanol. MWCNTs (125 mg) were added to 100 mL of methanol and dispersed for 10 min in an ice bath using an ultrasonic disruptor. Then, 2-MIM (3.284 g) was dissolved in the MWCNT-containing methanol after 20 min of ultrasound treatment at room temperature. The Zn2+-containing solution was then injected into the MWCNT/2-MIM solution. After stirring for 30 min, the mixture was incubated in a 30 °C water bath for 48 h. The final product was collected by centrifugation, washed with methanol several times, and dried under vacuum at 60 °C overnight to obtain ZIF-8@MWCNT.
The T-NC precursor was synthesized using the same procedure as ZIF-8@MWCNT, except that the MWCNTs were pre-oxidized using potassium permanganate (KMnO4) in sulfuric acid (H2SO4) solution prior to use.
2.3 Preparation of NC, NC@MWCNT, and T-NC and Pt reduction
The powder of ZIF-8, ZIF-8@MWCNT, and the T-NC precursor were placed in a tube furnace and heated to 1150 °C for 2 h at a heating rate of 5 °C/min under flowing argon gas, followed by natural cooling to room temperature to obtain the representative samples. The addition of PVP effectively prevented structural collapse of ZIF-8-derived carbon during calcination, as shown in Fig. S3. The Zn element was completely evaporated, as confirmed by Figs. S4 and S5. Larger ZIF-8 particle sizes resulted in more pronounced volume shrinkage after calcination. When the molar ratio of Zn(NO3)2·6H2O to 2-MIM was 1:4, the size of a single ZIF-8-derived carbon particle was approximately 240 nm, as shown in Figs. S6 and S7. The agglomerated carbon obtained after calcination was subjected to wet milling, vacuum filtration, and drying. Then, 100 mg of the sample was dispersed in 10 mL of IPA, ultrasonically treated for 1 h and analyzed for particle size distribution using a laser particle size analyzer (Bettersize2600).
For Pt loading, 160 mg of NC, MWCNT, NC@MWCNT, and T-NC powders were each dispersed in a mixed solution of EG (60 mL) and ultrapure water (20 mL) via ultrasonic treatment for 30 min, respectively. A 10 g/L H2PtCl6·6H2O solution (8.4 mL) was then added to each dispersion. The mixtures were stirred at 140 °C for 2 h. The resulting products were centrifuged, washed with ultrapure water several times, and dried under vacuum at 60 °C overnight.
For high Pt-loading catalysts used in the PEMFCs, 100 mg of T-NC powder was dispersed in a mixture of EG (200 mL) and ultrapure water (67 mL) via ultrasonic treatment for 30 min. A 10 g/L H2PtCl6·6H2O solution (40 mL) was then added. The mixture was stirred at 140 °C for 3 h. The products were centrifuged, washed with ultrapure water several times, and dried under vacuum at 60 °C overnight.
2.4 Material characterizations
The powdered samples were used for the structure characterization. Scanning electron microscopy (SEM) images were acquired using a HITACHI Regulus 8100 microscope. Transmission electron microscopy (TEM) images, high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images, and elemental mapping were performed on a JEOL JEM-F200 microscope. Nitrogen adsorption‒desorption measurements were performed with a Micromeritics ASAP 2460, and surface areas were calculated using the BET model. Power X-ray diffraction (XRD) patterns were collected using a Rigaku SmartLab SE equipped with Cu Ko radiation, using a scanning rate of 5 °/min. X-ray photoelectron spectroscopy (XPS) measurements were performed using a Thermo Scientific K-Alpha surface analysis system with a monochromatic Al Ka source. Raman spectroscopy was performed on a Horiba LabRAM HR Evolution spectrometer with a 633 nm laser. The ID/IG ratio of the prepared electrocatalysts was calculated based on the intensities of the G and D peaks.
2.5 Material electrochemical test
For the ink preparation for the rotating disk electrode (RDE) test, 4 mg of catalyst powder was mixed with 100 μL of water and 900 μL of IPA under ultrasonication for approximately 10 min. Then, 1 mL of Nafion ionomer solution diluted in IPA was added to the ink. After an additional 30 min of ultrasonication, the final ink was ready for use. A 20 μL aliquot of the prepared ink was dropped onto a 5 mm diameter glassy carbon electrode and dried in air. Then, the electrode was ready for electrochemical testing. Electrochemical measurements were performed in a three-electrode system. The working electrode was the catalyst-loaded glassy carbon electrode described above. A saturated Ag/AgCl electrode and a graphite rod were used as the reference and counter electrodes, respectively. ORR tests were conducted in an O2-saturated 0.1 mol/L HClO4 solution, with a linear voltage scan at a sweep rate of 10 mV/s.
Membrane electrode assembly (MEA) preparation and fuel cell testing were conducted as follows. Catalyst ink was prepared by ultrasonically dispersing the catalyst powder with ionomer solution diluted in an IPA/water solvent mixture for 2 h. The solid content in the ink was 1 wt.%. The resulting ink was then applied to a membrane to form the CLs via ultrasonic spraying (Sonotek ExactaCoat). Pt loading was detected by an X-ray fluorescence (Bruker S1TITAN). The anode CL, Pt loadings of 0.05 or 0.1 mgPt/cm2 was prepared using commercial Pt/C (TEC10E50E). The cathode CL, with Pt loadings of 0.1 or 0.2 mgPt/cm2, was prepared using the experimental catalysts. The prepared cathode and anode were assembled with gas diffusion layers (GDLs) of 168 μm thickness to make the final MEA. The active area was 4 cm2. Fuel cell testing was conducted on a fuel cell test station (NBT, PEM-400W) and an electrochemical workstation (Zahner, Zennium pro, and PP241). The cell temperature was maintained at 70 °C. Fully humidified hydrogen and air at 70 °C were supplied to the anode and cathode, respectively. Both inlet pressures were set to 150 kPa. The hydrogen flow rate was maintained at 400 sccm, while air was supplied at 800 sccm.
Accelerated durability test (ADT) was performed for 5000 triangular wave cycles, each cycling from 1.0 to 1.5 V and back to 1.0 V in two seconds. For fuel cell tests using H2/O2, the hydrogen flow rate was maintained at 300 sccm, while the oxygen flow was supplied at 600 sccm.
3 Results and discussion
3.1 Material characteristics comparison
The T-NC integrates the high graphitization of MWCNTs with the high specific surface area and N enrichment derived from ZIF-8-derived carbon. Bare MWCNT exhibits very few mesopores on the surface, with a specific surface area of only 88 m2/g (Figs. 2(a), S8–S11). During the synthesis of NC@MWCNT, excessive addition of MWCNT reduces the specific surface area of the product due to particle aggregation. However, after ZIF-8 grew and carbonized on the MWCNT with an appropriate amount, the specific surface area of NC@MWCNT significantly increases to 625 m2/g. Acid washing further introduces oxygen-containing functional groups and carbon defects on the MWCNT, increasing the ID/IG ratio from 0.66 to 0.77 (Fig. S12, Table S1). This treatment provides sufficient nucleation sites for ZIF-8 growth.
After calcination, T-NC retains the dodecahedron structure of ZIF-8 with a specific surface area of 699 m2/g. It features a hierarchical architecture comprising interconnected micropores, mesopores, and macropores. The markedly higher proportion of mesopores and macropores facilitates more efficient mass transfer (Fig. S13), and the morphology of T-NC shows a high degree of consistency (Fig. S14). According to Raman spectra, the ID/IG ratio of T-NC is reduced to 0.9, indicating higher graphitization than NC produced solely from ZIF-8 (Fig. 2(b)).
Following the same Pt reduction process, RDE test results (Fig. 2(c)) show that Pt/T-NC exhibits a higher limiting diffusion current and a half-wave potential (E1/2) of 0.88 V versus RHE, outperforming Pt/MWCNT, Pt/NC, and Pt/NC@MWCNT (0.834, 0.85, and 0.865 V, respectively). It demonstrates that the self-supporting structure of T-NC alleviates poor mass transfer caused by the stacking of carbon support particles, thereby enhancing ORR efficiency.
Correspondingly, XPS peak analysis of Pt 4f (Fig. 2(d)) reveals that Pt 4f7/2 in Pt/MWCNT has a relatively higher binding energy of 71.78 eV, indicating a more positive Ptδ+ valence state. In contrast, the Pt 4f7/2 peak of Pt/T-NC shifts to a lower binding energy of 71.3 eV, corresponding to improved catalytic activity for ORR. The binding energy differences relate to Pt particle size variations on different carbon supports. XRD patterns show negligible changes in Pt crystalline structure across supports due to the same reduction process (Fig. 2(f)).
TEM images (Fig. 2(e)) reveal that Pt particles tend to be larger and unevenly distributed on MWCNT surfaces. In contrast, on T-NC, Pt tends to attach to the NC surface rather than the MWCNT (Fig. S15). The evaporation of Zn during preparation results in T-NC, NC, and NC@MWCNT possessing more pores and N content, providing more Pt attachment sites and further limiting particle growth. Pt particles on T-NC are approximately 2 nm. TEM and HADDF-STEM images confirm uniform distribution of N and Pt (Fig. 2(g)). This is also corroborated by XRD analysis, where the smaller full width at half maximum (FWHM) of the Pt diffraction peaks in Pt/MWCNT suggests a larger particle size. Additionally, the diffraction peak at 26.3° in Pt/T-NC corresponds to the (002) plane of graphitic carbon, confirming the preservation of the highly graphitized MWCNT structure, which is expected to enhance long-term operational stability.
3.2 PGMs reduction and high loading on T-NC
In ORR catalyst research, although some single-atom catalysts demonstrate excellent mass activity, their PGM loading is usually low. When these catalysts are prepared as fuel cell membrane electrodes, the increased thickness of the catalytic layer often leads to inefficient charge transfer and mass transfer. To minimize internal cell resistance and shorten mass transfer paths, the CL in the MEA needs to be very thin. This, in turn, requires high PGM loadings (≥ 40 wt.%) in novel catalyst development to ensure practical application in devices [
24].
In the PGM precursor reduction and high-loading process, the particle size of carbon support is a critical factor governing the performance of the product. For T-NC, different grinding methods were used before polyol reduction, controlling the medium diameter (D50) at 8.33, 6.72, and 2.66 μm as measured by a laser particle analyzer (Fig. S16). MEAs with the same Pt loading were prepared and tested, with a cathode Pt loading of 0.2 mgPt/cm2 set for clear comparison. The peak power densities were 0.048, 0.531, and 1.146 W/cm2, corresponding to 8.33, 6.72, and 2.66 μm, respectively (Fig. 3(a)). The cell output performance increased by a remarkable 23.8 times with well-controlled T-NC size, attributed to the synergistic effects of reduced CL thickness and enhanced catalytic activity. The carbon support particle size directly influences the CL thickness and thus the mass transfer path. Corresponding cathode CLs thicknesses were 14, 7, 5 μm for the respective particle sizes (Figs. S17, S18). The thinnest CL was achieved with T-NC at 2.66 μm D50 for the same Pt loading.
Additionally, the carbon support particle size affects the properties of the PGMs. XRD patterns show negligible changes in the crystalline index of Pt on T-NC with different sizes (Fig. 3(b)). However, smaller particle size exposes more pore structure and elemental N on the T-NC surface, favoring control of Pt particle size. Pt particles supported on T-NC with a D50 of 2.66 μm exhibited the smallest average particle size of approximately 3.73 nm (Fig. 3(c)), demonstrating that even at high Pt loading (> 50 wt.%), T-NC still enables excellent dispersion and effective size control of Pt nanoparticles (Fig. S19).
Correspondingly, XPS Pt 4f peak splitting data reveal that Pt 4f peaks of Pt/T-NC shift to lower binding energy with decreasing T-NC particle size (Fig. 3(d)). In the N 1s spectrum of Pt/T-NC (Fig. 3(e)), four peaks at 398.1, 398.7, 399.8, and 401.9 eV correspond to pyridinic N, Pt-N, pyrrolic N, and graphitic N, respectively [
25,
26]. Only the sample with T-NC at 2.66 μm shows Pt-N coordination bonds, indicating interaction between N atoms and Pt
δ +, which has been reported to strengthen the ORR activity of Pt and improve PEMFC performance [
23]. RDE tests (Fig. S20) show
E1/2 of 0.886, 0.905, and 0.915 V for Pt/T-NC with D50 values of 8.33, 6.72, and 2.66 μm, respectively. The highest
E1/2 observed for Pt supported on T-NC with 2.66 μm D50 confirms that the ORR enhancement results directly from reduced Pt particle size and Pt-N coordination.
3.3 Electrochemical performance
For low-Pt loaded fuel cells, Pt/T-NC electrodes demonstrate significantly enhanced electrochemical performance compared to conventional Pt/C electrodes in PEMFC testing under H
2/Air conditions, especially at high current densities. When the Pt loading is 0.05 mg
Pt/cm
2 at the anode and 0.1 mg
Pt/cm
2 at the cathode, the peak power density of the PEMFC with Pt/T-NC reaches 0.93 W/cm
2, representing a 12.7% increase over that of the commercial Pt/C catalyst (Fig. 4(a)). The performance improvement becomes even more pronounced with increased Pt/T-NC loading on the cathode (Figs. S21, S22). Moreover, the internal resistance (iR)-corrected Tafel plot demonstrates that the fuel cell mass activity (MA) of Pt/T-NC at 0.9 V
iR-free is 0.217 A/mg
Pt (Figs. S23, S24). Compared to state-of-the-art Pt alloy catalysts tested under similar operating conditions (Table S2), the PEMFC with Pt/T-NC exhibits competitive power performance despite the intrinsic limitations of pure Pt nanoparticles in ORR activity (Fig. 4(b)). This highlights the significant role of T-NC in modifying the microstructure of low-Pt loaded CLs. Notably, T-NC is a versatile support material that can replace conventional carbons such as Vulcan XC-72R [
27], Cabot BP2000 [
28], and Ketjenblack EC300J [
31] in a wide range of synthesis routes, enabling the translation of highly active, mass-produced catalysts into CLs with superior performance.
Multiple electrochemical impedance spectroscopy (EIS) tests were performed to decouple activation, ohmic, and concentration losses, validating the microstructural improvements of the cathode CL under practical operating conditions. The activation loss and MA of the Pt/T-NC electrode closely match those of the Pt/C electrode (Figs. 4(c), S25), confirming the feasibility of the Pt reduction method employed. The ohmic loss for Pt/T-NC is lower than that of Pt/C, especially at low current densities, likely due to improved membrane hydration from better water management. At higher current densities, as water removal becomes limited, the ohmic losses of Pt/C approach those of Pt/T-NC (Fig. 4(d)). The concentration loss trends further support these observations (Fig. 4(e)). Despite increased water production at high currents, the concentration loss of the Pt/T-NC electrode decreases relative to Pt/C, showing a 30% reduction, compared to the Pt/C electrode at 2.0 A/cm
2. This demonstrates that T-NC facilitates the formation of an effective mass transport network in low-Pt loaded fuel cells, enhancing water management capabilities. Limiting current density tests at different inlet pressures were performed to assess oxygen transport resistance [
37]. The pressure-independent resistance term (
RNP) of the PEMFC with Pt/T-NC is significantly lower than that with commercial Pt/C, indicating reductions in Knudsen diffusion resistance and permeation resistance through ionomer/water film, consistent with uniform ionomer coverage of the electrode.
3.4 Electrode durability
An accelerated durability test (ADT) for carbon corrosion was conducted to evaluate the durability of the T-NC support in PEMFCs. The test followed the DOE-recommended protocol [
38], involving triangular wave potential cycling between 1.0 and 1.5 V with a period of 2 s. After 5000 cycles, T-NC exhibited superior corrosion resistance, attributed to its high degree of graphitization (Fig. 5(a)). The peak power density of the PEMFC with Pt/T-NC decreased by 49.2% to 0.472 W/cm
2, whereas the PEMFC with commercial Pt/C suffered a larger reduction of 62%.
Further comparison of beginning-of-test (BOT) and end-of-test (EOT) resistances at 0.8 A/cm2 (Figs. 5(b) and 5(c)) showed that the Pt/T-NC electrode exhibited increases of 28.8%, 66.5%, and 40.0% in ohmic, activation, and concentration losses, respectively. In contrast, the Pt/C electrode showed substantially higher increases of 49.8%, 82.2%, and 146%. The electrochemically active surface area (ECSA) of Pt/T-NC electrode retained 51.5% of its initial value at EOT, more than 2.2 times that of Pt/C electrode (Fig. 5(d)). TEM analysis of Pt particle sizes before and after (Figs. S26 and S27) revealed only a 6.7% increase on T-NC, compared to 14.1% growth on commercial carbon support (Figs. 5(e) and 5(f)).
Performance degradation in Pt/C-based PEMFCs primarily stems from CL collapse due to carbon corrosion, leading to reduced conductivity, Pt particle agglomeration, and severe ionomer redistribution, which exacerbate ohmic, activation, and concentration losses [
39,
40]. In contrast, electrodes containing T-NC largely maintain their original morphology and exhibit strong PGM anchoring, significantly enhancing PEMFC durability.
4 Conclusions
In summary, a facile synthesis method for T-NC is proposed, where MWCNTs serve as a conductive skeleton and ZIF-8-derived carbon provides surface porosity and N doping to anchor catalysts. The T-NC electrode creates a highly graphitized conductive network and efficient mass transfer pathways, fulfilling the durability and performance requirements of low-Pt loaded PEMFCs. Compared to conventional Pt/C electrodes, the T-NC electrode achieves a 12.7% increase in peak power density, a 21.6% reduction in pressure-independent oxygen transport resistance, and a 30% reduction in concentration loss at 2.0 A/cm2.
During carbon degradation tests, the T-NC electrode demonstrates significantly enhanced corrosion resistance. The Pt/T-NC electrode exhibits significantly lower increases in ohmic, activation, and concentration losses relative to Pt/C. At the end of the test, the ECSA retention is 2.2 times higher, indicating superior structural stability and electrochemical durability.
This paper offers a promising technological route to bridge the gap between theoretical catalytic activity and practical low-Pt loaded fuel cell performance through enhanced mass transfer. It contributes to the translation of highly active, mass-produced catalysts into CLs with superior performance, leading to substantial improvements in both fuel cell performance and durability.
PDF
(4227KB)
