Power Generation Characteristics of Polymer Electrolyte Fuel Cells Using Carbon Nanowalls as Catalyst Support Material

: We evaluated the power generation characteristics of a polymer electrolyte fuel cell (PEFC) composed of Pt-supported carbon nanowalls (CNWs) and a microporous layer (MPL) of carbon black on carbon paper (CP) as catalyst support materials. CNWs, standing vertically on highly crystallizing graphene sheets, were synthesized on an MPL/CP by plasma-enhanced chemical vapor deposition (PECVD) using inductively coupled plasma (ICP). Pt nanoparticles were supported on the CNW surface using the liquid-phase reduction method. The three types of voltage loss, namely those due to activated polarization, resistance polarization, and diffusion polarization, are discussed for the power generation characteristics of the PEFC using the Pt/CNWs/MPL/CP. The relationship between the height or gap area of the CNWs and the voltage loss of the PEFC is demonstrated, whereby the CNW height increased with the extension of growth time. The three-phase interface area increased with the increase in the CNW height, resulting in mitigation of the loss due to activated polarization. The gap area of the CNWs varied when changing the CH 4 /H 2 gas ratio. The loss due to diffusion polarization was reduced by enlarging the gap area, due to the increased diffusion of fuel gas and discharge of water. The secondary growth of the CNWs caused the three-phase interface area to decrease as a result of platinum aggregation, impedance of the supply of ionomer dispersion solution to the bottom of the CNWs, and inhibition of fuel gas and water diffusion, which led to the loss of activated and diffuse polarizations. The voltage losses can be mitigated by increasing the height of CNWs while avoiding secondary growth.

In recent years, carbon nanowalls (CNWs) were discovered, and fundamental research on them has been progressing [68][69][70][71].CNWs have been grown using various chemical vapor deposition (CVD) methods, and several applications using CNWs have been reported, such as fuel cells [72][73][74][75][76][77], biofuel cells [78], strain sensors [79], electrochemical sensors [80][81][82], scaffolds [83], and substrates for surface-assisted laser desorption/ionization mass spectrometry [84,85].CNWs are composed of stacked graphene sheets and are stranded perpendicular to the substrate.CNWs are expected to be alternative catalyst support materials due to their high electric conductivity, very large specific surface area, and physically robust shape due to having a high aspect ratio.The electrochemical characterization or high durability of CNWs for a fuel cell has been reported [77][78][79][80][81].A test cell unit using a Pt-supported CNW/carbon fiber paper was fabricated and the V-I curve characterized for assessment of proton exchange membrane (PEM) fuel cell application.However, the PEM fuel cell exhibits a voltage drop to 0.2 V due to activated polarization, because no ionomer binder is incorporated in the catalyst layers.The three-phase boundary region near the surface of the catalyst layers and the proton conduction between the catalyst layer and membrane are both essential for improving the power generation efficiency.Since CNWs have a unique structure, the gap area and height of CNWs can be controlled by changing the deposition conditions.
In this study, CNWs were synthesized on a microporous layer (MPL) of carbon black on carbon paper (CP) by plasma-enhanced chemical vapor deposition (PECVD) using inductively coupled plasma (ICP) to improve the power generation of the PEFC.The relationship between the CNW height or gap area and the PEFC power generation characteristics was investigated.

Plasma-Enhanced Chemical Vapor Deposition for CNW Growth
Figure 1 shows a schematic diagram of the ICP reactor used for CNW growth.RF (13.56 MHz) power was applied to the coil antenna.Two types of CNWs with different gap spaces between walls were synthesized on an MPL of carbon black on the CP (MPL/CP) (SGL carbon GmbH; GDL 29BC, thickness: 235 µm, void fraction: 40-41%).Condition (1) for the small gap was 550 W RF power, 20/44 sccm Ar/CH 4 gas flow rates, 22 mTorr total pressure, and 720 • C growth temperature.Condition (2) for the large gap was 600 W RF power, 20/9/8 sccm Ar/CH 4 /H 2 gas flow rates, 16 mTorr total pressure, and 720 • C growth temperature.

Structure of a Single PEFC Using CNWs
Figure 2 shows a schematic diagram of the single PEFC, which was assemble a membrane electrode assembly (MEA) and separators.The MEA consisted Pt-supported CNWs on MPL/CP, ionomer dispersion solution, and non-treated io membrane (Nafion 117) for proton exchange.The surface of the CNWs on the M was treated with atmospheric pressure plasma to endow hydrophilicity.The Ar ga rate was 2.0 sccm, the applied voltage to the electrodes was 6 kV, and the treatme tance was 5 mm.Pt nanoparticles were supported on the surface of the CNWs/M by the liquid-phase reduction method using chloroplatinic acid solution (H2PtCl6, 8 in H2O) and NaBH4 as the reducing agents.A total volume of 0.1 mL ionomer disp solution (Nafion DE202) diluted with 2-propanol was applied dropwise o Pt/CNWs/MPL/CP.The MEA was fabricated by hot-pressing at 135 °C with a pr pressure of 5.0 MPa and pressing time of 90 s.The MEA with an electrode area of was assembled to the standard single PEFC (Electro Chem, Inc., EFC-05-02, W MA, USA).The power generation characteristics were investigated using a fuel ce system (Scribner Associates Inc., AutoPEM, Southern Pines, NC, USA) at an ano gas flow rate of 0.

Structure of a Single PEFC Using CNWs
Figure 2 shows a schematic diagram of the single PEFC, which was assembled with a membrane electrode assembly (MEA) and separators.The MEA consisted of the Ptsupported CNWs on MPL/CP, ionomer dispersion solution, and non-treated ionomer membrane (Nafion 117) for proton exchange.The surface of the CNWs on the MPL/CP was treated with atmospheric pressure plasma to endow hydrophilicity.The Ar gas flow rate was 2.0 sccm, the applied voltage to the electrodes was 6 kV, and the treatment distance was 5 mm.Pt nanoparticles were supported on the surface of the CNWs/MPL/CP by the liquid-phase reduction method using chloroplatinic acid solution (H 2 PtCl 6 , 8 wt.% in H 2 O) and NaBH 4 as the reducing agents.A total volume of 0.1 mL ionomer dispersion solution (Nafion DE202) diluted with 2-propanol was applied dropwise on the Pt/CNWs/MPL/CP.The MEA was fabricated by hot-pressing at 135 • C with a pressing pressure of 5.0 MPa and pressing time of 90 s.The MEA with an electrode area of 5 cm 2 was assembled to the standard single PEFC (Electro Chem, Inc., EFC-05-02, Woburn, MA, USA).The power generation characteristics were investigated using a fuel cell test system (Scribner Associates Inc., AutoPEM, Southern Pines, NC, USA) at an anode H 2 gas flow rate of 0.1 L/m, cathode air flow rate of 0.2 L/m, and a temperature of 80 • C.

CNWs on MPL/CP
Figure 3 shows the height of the CNWs as a function of the growth time for condition (1).The height of the CNWs increased from 1.7 µm to 7.2 µ m as the growth time increased.The growth rate was estimated as 28 nm/min.Figure 4 shows the SEM images of the CNWs for various growth times.The CNWs were directly formed on the MPL of carbon black as shown in Figure 4a.The growth mechanism of CNWs on the substrates was detailed in a previous report [80].The growth was as follows: (1) A very thin amorphous carbon layer formed on the substrate.(2) A nucleation site was formed there by ion irradiation.(3) Nanoislands were formed by aggregating carbon radicals at the nucleation site.(4) The nanoislands were irradiated with ions, and the absorption of carbon radicals was enhanced.( 5) Nanographene grew preferentially in the height direction to form carbon nanowalls.In this study, the flake-like CNWs grew on the MPL, and some branches grew from the flake-like CNWs after 3 h.

CNWs on MPL/CP
Figure 3 shows the height of the CNWs as a function of the growth time for condition (1).The height of the CNWs increased from 1.7 µm to 7.2 µm as the growth time increased.The growth rate was estimated as 28 nm/min.Figure 4 shows the SEM images of the CNWs for various growth times.The CNWs were directly formed on the MPL of carbon black as shown in Figure 4a.The growth mechanism of CNWs on the substrates was detailed in a previous report [80].The growth was as follows: (1) A very thin amorphous carbon layer formed on the substrate.(2) A nucleation site was formed there by ion irradiation.
(3) Nanoislands were formed by aggregating carbon radicals at the nucleation site.(4) The nanoislands were irradiated with ions, and the absorption of carbon radicals was enhanced.
(5) Nanographene grew preferentially in the height direction to form carbon nanowalls.In this study, the flake-like CNWs grew on the MPL, and some branches grew from the flake-like CNWs after 3 h.

CNWs on MPL/CP
Figure 3 shows the height of the CNWs as a function of the growth time for condition (1).The height of the CNWs increased from 1.7 µm to 7.2 µ m as the growth time increased.The growth rate was estimated as 28 nm/min.Figure 4 shows the SEM images of the CNWs for various growth times.The CNWs were directly formed on the MPL of carbon black as shown in Figure 4a.The growth mechanism of CNWs on the substrates was detailed in a previous report [80].The growth was as follows: (1) A very thin amorphous carbon layer formed on the substrate.(2) A nucleation site was formed there by ion irradiation.(3) Nanoislands were formed by aggregating carbon radicals at the nucleation site.(4) The nanoislands were irradiated with ions, and the absorption of carbon radicals was enhanced.(5) Nanographene grew preferentially in the height direction to form carbon nanowalls.In this study, the flake-like CNWs grew on the MPL, and some branches grew from the flake-like CNWs after 3 h.

Relationship between CNW Height and Power Generation Characteristics
Figure 5 shows the voltage-current density (V-J) curve of the single PEFC for the various growth times for deposition under condition (1).Three gradients, due to activated polarization, resistance polarization, and diffusion polarization, were observed on the PEFC using Pt/CNW/MPL/CP.The maximum current density increased with the increase in the growth time up to 3.5 h, since the reaction area at the three-phase boundary region among Pt on CNW, ionomer dispersion solution, and supplied fuel gas increased with the increase in the CNW height.The maximum current density, however, decreased after more than 3.5 h, because the dispersion of the Pt and ionomer dispersion solution to the bottom of CNWs was suppressed due to the secondary growth (branches) from the sidewalls of the flake-like CNWs.

Relationship between CNW Height and Power Generation Characteristics
Figure 5 shows the voltage-current density (V-J) curve of the single PEFC for the various growth times for deposition under condition (1).Three gradients, due to activated polarization, resistance polarization, and diffusion polarization, were observed on the PEFC using Pt/CNW/MPL/CP.The maximum current density increased with the increase in the growth time up to 3.5 h, since the reaction area at the three-phase boundary region among Pt on CNW, ionomer dispersion solution, and supplied fuel gas increased with the increase in the CNW height.The maximum current density, however, decreased after more than 3.5 h, because the dispersion of the Pt and ionomer dispersion solution to the bottom of CNWs was suppressed due to the secondary growth (branches) from the sidewalls of the flake-like CNWs.
The magnitude of the three types of voltage loss, activated polarization, resistance polarization, and diffusion polarization, was evaluated with consideration to the power generation characteristics and how improvements could be achieved by reducing this loss [86].Figure 6 shows the voltage loss due to activated polarization for the various growth times.Activated polarization is the potential difference beyond the value of equilibrium needed to generate currents and is dependent on the energy activation of a redox reaction and its reaction area, and it can be calculated using the following equations: where i is the total current (A), a is the coefficient of migration, R is the gas constant (8.314J/K mol), T is the temperature (K), i 0 is the exchange current (A), n is the number of electrons involved in the reaction (mol), F is the Faraday constant (96,485 C/mol), E a is the activation energy (J/mol), k 0 is the reaction rate constant (s −1 ), and A is the reaction area (m 2 ).
C 2022, 8, x FOR PEER REVIEW 6 of 15 The magnitude of the three types of voltage loss, activated polarization, resistance polarization, and diffusion polarization, was evaluated with consideration to the power generation characteristics and how improvements could be achieved by reducing this loss [86].Figure 6 shows the voltage loss due to activated polarization for the various growth times.Activated polarization is the potential difference beyond the value of equilibrium needed to generate currents and is dependent on the energy activation of a redox reaction and its reaction area, and it can be calculated using the following equations: where i is the total current (A), a is the coefficient of migration, R is the gas constant (8.314J/K mol), T is the temperature (K), i0 is the exchange current (A), n is the number  The loss due to activated polarization decreased as the growth time was increased from 2 to 3.5 h, followed by an increase from 3.5 to 4 h.The activation energy remained constant in this study because the catalytic activity did not change for any of the experimental conditions.As shown in Figure 3, the three-phase interface region increased with the increase in CNW height observed from 2 to 3.5 h, resulting in an increase in reaction area .However, the Pt particles aggregated due to CNW secondary growth from 3.5 to 4 h.Accordingly, the decrease in the three-phase interface led to the loss of activated polarization.
Figure 7 shows the voltage loss due to resistance polarization for the various growth times.Resistance polarization is a part of electrode polarization arising from an electric current through an ohmic resistance within the electrode or the electrolyte.The  The loss due to activated polarization decreased as the growth time was increased from 2 to 3.5 h, followed by an increase from 3.5 to 4 h.The activation energy remained constant in this study because the catalytic activity did not change for any of the experimental conditions.As shown in Figure 3, the three-phase interface region increased with the increase in CNW height observed from 2 to 3.5 h, resulting in an increase in reaction area A. However, the Pt particles aggregated due to CNW secondary growth from 3.5 to 4 h.Accordingly, the decrease in the three-phase interface led to the loss of activated polarization.
C 2022, 8, 44 7 of 14 Figure 7 shows the voltage loss due to resistance polarization for the various growth times.Resistance polarization is a part of electrode polarization arising from an electric current through an ohmic resistance within the electrode or the electrolyte.The resistance polarization can be calculated using the following equation: where j is the current density (A/cm 2 ) and R is the internal resistance of the fuel cell (Ωcm 2 ).The ionic conductivity of the electrolyte, the electrical conductivity of the CNWs, and the contact resistance did not change regardless of the CNW growth time.
imental conditions.As shown in Figure 3, the three-phase interface region increased with the increase in CNW height observed from 2 to 3.5 h, resulting in an increase in reaction area .However, the Pt particles aggregated due to CNW secondary growth from 3.5 to 4 h.Accordingly, the decrease in the three-phase interface led to the loss of activated polarization.
Figure 7 shows the voltage loss due to resistance polarization for the various growth times.Resistance polarization is a part of electrode polarization arising from an electric current through an ohmic resistance within the electrode or the electrolyte.The resistance polarization can be calculated using the following equation: where j is the current density (A/cm 2 ) and R is the internal resistance of the fuel cell (Ωcm 2 ).The ionic conductivity of the electrolyte, the electrical conductivity of the CNWs, and the contact resistance did not change regardless of the CNW growth time.Figure 8 shows the voltage loss due to diffusion polarization for the various growth times.The diffusion polarization of an electrode is a result of the formation of a diffusion layer for fuel gas at the three-phase boundary or the water as a byproduct of an oxygen reduction reaction.Diffusion polarization can be calculated using the following equation: where C OX is the concentration of oxygen at the three-phase boundary and C • OX is the concentration of bulk oxygen.The loss due to diffusion polarization decreased with the increase in growth time from 2 to 3.5 h, since the oxygen concentration (C OX ) at the threephase boundary increased along with the CNW height.However, from 3.5 to 4 h, the loss due to diffusion polarization decreased.The supply of fuel gas to the three-phase interface or the discharge of water molecules as the byproduct of the oxygen reduction reaction would be suppressed by CNW secondary growth.It was found that the best power generation characteristics were obtained at the growth time of 3.5 h, as shown in Figure 5, due to an increase in the three-phase interface area and the improvements such as the increased diffusion of fuel gas and water molecules.
C 2022, 8, 44 8 of 14 the loss due to diffusion polarization decreased.The supply of fuel gas to the three-phase interface or the discharge of water molecules as the byproduct of the oxygen reduction reaction would be suppressed by CNW secondary growth.It was found that the best power generation characteristics were obtained at the growth time of 3.5 h, as shown in Figure 5, due to an increase in the three-phase interface area and the improvements such as the increased diffusion of fuel gas and water molecules.

Relationship between CNW Gap Area and Power Generation Characteristics
Figure 9 shows plane or cross-sectional SEM images of the CNWs for different gap areas.The CNWs with a small gap area (0.37 µ m 2 ) were deposited under condition (1), and those with a large gap area (0.79 µ m 2 ) were obtained under condition (2).The growth time to achieve a height of 1.6 µ m was 2 h for condition (1) and 70 min for condition (2) since the influence of the secondary growth can be suppressed under relatively low height.Here, the gap area of CNWs was evaluated from the area enclosed by the walls as their reciprocal [87].

Relationship between CNW Gap Area and Power Generation Characteristics
Figure 9 shows plane or cross-sectional SEM images of the CNWs for different gap areas.The CNWs with a small gap area (0.37 µm 2 ) were deposited under condition (1), and those with a large gap area (0.79 µm 2 ) were obtained under condition (2).The growth time to achieve a height of 1.6 µm was 2 h for condition (1) and 70 min for condition (2) since the influence of the secondary growth can be suppressed under relatively low height.Here, the gap area of CNWs was evaluated from the area enclosed by the walls as their reciprocal [87].Figure 10 shows the V-J curve of a single PEFC for the different gap areas.Three gradients, due to activated polarization, resistance polarization, and diffusion polarization, were observed on the large gap area.The voltage of approximately 0.91 V at a current density of 0 mA/cm 2 was obtained for a large gap area under relatively low height.Moreover, the maximum current density for a large gap area was larger than that with a C 2022, 8, 44 9 of 14 Figure 10 shows the V-J curve of a single PEFC for the different gap areas.Three gradients, due to activated polarization, resistance polarization, and diffusion polarization, were observed on the large gap area.The voltage of approximately 0.91 V at a current density of 0 mA/cm 2 was obtained for a large gap area under relatively low height.Moreover, the maximum current density for a large gap area was larger than that with a small gap area and was increased from approximately 0.15 to 0.3 A/cm 2 at the same height.Figure 11 shows the voltage loss due to activated polarization for the different gap areas.The voltage loss due to the activated polarization for the large gap was smaller than that for the small gap.The difference was approximately 0.2 V regardless of the current density.These results indicate that the three-phase interface region increased with the increase in gap area since the supply of platinum or ionomer dispersion solution to the bottom of the CNWs may have been enhanced.Figure 12 shows the voltage loss due to diffusion polarization for the different gap areas.The voltage loss due to diffusion polarization for the large gap was improved to be approximately 4 times smaller than that for the small gap.This result indicates that the supply of fuel gas to the three-phase interface and the discharge of water molecules were drastically improved by enlarging the gap of CNWs.Gap area: Small Gap area Large    The voltage loss due to activated polarization and diffusion polarization was improved by an increase in gap area, although the surface area of the CNWs was decreased.The improvement in the diffusion polarization indicates that the supply of fuel gas to the three-phase interface and the discharge of water molecules were enhanced.Accordingly, the loss due to the activated polarization was improved, since the effective area of the three-phase interface was increased by the supply of fuel gas.Moreover, the Pt particle and the ionomer dispersion solution would be supplied to the bottom of CNWs in the case of the large gap.It was found that the power generation characteristics were improved by enlarging the gap of CNWs, since the loss due to activated polarization and diffusion polarization was mitigated.

Conclusions
CNWs as catalyst support materials were synthesized on an MPL/CP by ICP-PECVD.The power generation of the single PEFC using Pt/CNWs/MPL/CP as an MEA was demonstrated.The power generation characteristics of the PEFC were evaluated based on the three types of voltage loss due to activated polarization, resistance polarization, and diffusion polarization.

•
An increase in the height of the CNWs increased the three-phase interface area with a reduction in the loss due to activated polarization.

•
An increase in the gap area of the CNWs resulted in improvements due to increases in fuel gas diffusion and water discharge, and the loss due to diffusion polarization was reduced.

•
The secondary growth of the CNWs caused a reduction in the three-phase interface area due to platinum aggregation, impedance of the supply of ionomer dispersion solution to the bottom of CNWs, and inhibited fuel gas and water diffusion, which led to the loss of activated and diffuse polarizations.
The voltage losses can be mitigated by increasing the height of the CNWs and expanding the gap area.The data obtained in this study are useful for the fabrication of PEFCs using CMWs.

Figure 1 .
Figure2shows a schematic diagram of the single PEFC, which was assemble a membrane electrode assembly (MEA) and separators.The MEA consisted Pt-supported CNWs on MPL/CP, ionomer dispersion solution, and non-treated io membrane (Nafion 117) for proton exchange.The surface of the CNWs on the M was treated with atmospheric pressure plasma to endow hydrophilicity.The Ar ga rate was 2.0 sccm, the applied voltage to the electrodes was 6 kV, and the treatme tance was 5 mm.Pt nanoparticles were supported on the surface of the CNWs/M by the liquid-phase reduction method using chloroplatinic acid solution (H2PtCl6, 8 in H2O) and NaBH4 as the reducing agents.A total volume of 0.1 mL ionomer disp solution (Nafion DE202) diluted with 2-propanol was applied dropwise o Pt/CNWs/MPL/CP.The MEA was fabricated by hot-pressing at 135 °C with a pr pressure of 5.0 MPa and pressing time of 90 s.The MEA with an electrode area of was assembled to the standard single PEFC (Electro Chem, Inc., EFC-05-02, W MA, USA).The power generation characteristics were investigated using a fuel ce system (Scribner Associates Inc., AutoPEM, Southern Pines, NC, USA) at an ano gas flow rate of 0.1 L/m, cathode air flow rate of 0.2 L/m, and a temperature of 80 °

Figure 2 .
Figure 2. Schematic diagram of the single PEFC using CNWs.

Figure 3 .Figure 2 .
Figure 3. Height of CNW as a function of growth time.

C 15 Figure 2 .
Figure 2. Schematic diagram of the single PEFC using CNWs.

Figure 3 .Figure 3 .
Figure 3. Height of CNW as a function of growth time.

Figure 4 .
Figure 4. Cross-sectional view of the SEM images of the CNWs at a growth time of (a) 2 h, (b) 3 h, (c) 3.5 h, and (d) 4 h.

Figure 4 .
Figure 4. Cross-sectional view of the SEM images of the CNWs at a growth time of (a) 2 h, (b) 3 h, (c) 3.5 h, and (d) 4 h.

Figure 5 .
Figure 5. Voltage-current density (V-J) curve of single PEFC cells of Pt/CNW/MPL/CP for various growth times of the CNWs.

Figure 5 .
Figure 5. Voltage-current density (V-J) curve of single PEFC cells of Pt/CNW/MPL/CP for various growth times of the CNWs.C 2022, 8, x FOR PEER REVIEW 7 of 15

Figure 6 .
Figure 6.Activated polarization for various growth times.

Figure 6 .
Figure 6.Activated polarization for various growth times.

Figure 8 .
Figure 8. Diffusion polarization for various growth times.

Figure 8 .
Figure 8. Diffusion polarization for various growth times.

Figure 9 .
Figure 9. (a,c) Plane view and (b,d) cross-sectional view of the SEM images of CNWs grown on MPL with (a,b) a small gap area (0.37 µm 2 ) and (c,d) a large gap area (0.79 µm 2 ).

Figure 9 .
Figure 9. (a,c) Plane view and (b,d) cross-sectional view of the SEM images of CNWs grown on MPL with (a,b) a small gap area (0.37 µm 2 ) and (c,d) a large gap area (0.79 µm 2 ).

C 15 Figure 10 .
Figure 10.Voltage-current density (V-J) curve of single PEFC cells of Pt/CNWs/MPL/CP for different gap areas.

Figure 11 .
Figure 11.Activated polarization for different gap areas.

Figure 11 .
Figure 11.Activated polarization for different gap areas.

Figure 11 .
Figure 11.Activated polarization for different gap areas.

Figure 11 .
Figure 11.Activated polarization for different gap areas.

Figure 12 .
Figure 12.Diffusion polarization for different gap areas.

Figure 12 .
Figure 12.Diffusion polarization for different gap areas.