CHAPTER 4 TRANSPORT COMPONENT DESIGN EFFECTS ON
4.6 Tendency Comparisons with Previous Works from Literature
The transport phenomena of various species in PEMFCs and the competition effect between electrons and reactant gas are elucidated in this chapter. Contrary to the earlier investigation, current modeling results on the transport component design effects show that the peak point of local current density is not located at the middle of channel but moves around shoulder and channel regions of the CLs. This phenomenon is explained by the combined effort of reactant concentration and activation overpotential that appeared in the Bulter-Volmer equation. Similar results can be found in the works of Sivertsen (2005) and Meng (2004). As regards the comparison with experimental results, literature survey shows that there is few experimental works concerning the polarization curves for the effects of AR and GDL thickness.
The work of Lee et. al. (1999a) on the effects of compression and gas diffusion layers may provide a clue to examine the effect of GDL thickness. Fig. 4.15 shows their experimental work of cell performance subjected to three different screw compression forces by using the ELAT GDL which is a flexible material. From the principle of material mechanics, material with lower compression force corresponds to a larger
dimension and vice versa. Therefore, this figure may be employed to the examination of GDL thickness effects. The data show that when the bolt torque is 100in-lb, which corresponds to the thickest GDL, the cell current density is the smallest at each operating voltage. This is consistent with current modeling result in Fig. 4.2(b) that the 352 μ GDL design has the worst performance for every cell voltage. m Furthermore, experimental results of 125 and 150 in-lb bore torques reveal that the better performance case is dictated by the operating condition. At cell voltage of 0.7V, the cell with 150 in-lb bore torque exhibits greater current density and that with 125 in-lb bore torque behaves better when the reaction rate increases. Apparently, this tendency is coincide with the data in Fig. 4.2(b), offering a strong support of the reliability of current work.
4.7 Summary
A multi-dimensional, multi-component, computational fluid dynamic model has been employed to elucidate transport phenomena and electrochemical reaction in PEMFCs. Two important transport component design parameters - channel aspect ratio and GDL thickness, are investigated in detail. The simulation and discussion support the following conclusions:
1. The designs of channel aspect ratio and GDL thickness affect the physical property distributions. At a high reaction rate, these two parameters have a strong influence on cell performance.
2. Three mechanisms of cell irreversibility are resolved locally from the model and are considered to determine variation in the macroscopic cell current density and performance.
3. Transverse current density distribution is governed by both electron conduction and activation overpotential in channel region or by reactant concentration in shoulder region of catalyst layer. The relative strengths of these two factors depend on the operating voltage and the transport geometry.
4. At a moderate reaction rate, the transverse direction current density in most regions is a convex function of position and is influenced by solid phase potential and activation overpotential, making the peak point of the concave pattern current density either close to the shoulder region or non-existent. Thus, a geometry design that facilitates electron transport such as a large channel aspect ratio or a thin GDL thickness causes a relatively larger current in the catalyst layer.
5. At a low cell voltage, the largest reaction rate location is close to the channel center, therefore, the electrochemical reaction in the majority regions is
dominated by the reactant transport and a smaller channel aspect ratio design is preferred.
6. Large GDL thickness positively increases oxygen transport under the shoulder region; therefore output current density elevates according to cell voltage decrease. However, at the lowest considered cell voltage of 0.14V, oxygen deficiency caused by long traveling length and liquid water clogging effect reverses this relationship.
7. Plots of axial velocity and oxygen concentration in cathode vertical direction indicate that the main mass transport mechanism in channel is the convection and that in electrode is the diffusion. Therefore, vertical distance change arising from the variation of channel aspect ration and GDL thickness cause a marked difference of oxygen concentration level at cathode CL.
8. The operating temperature has a positive effect on cell performance for various cases of transport component designs considered in this study at medium cell voltages due to its improvement of ohmic overpotential. However, this phenomenon is not sensible at high reaction rates for the design with a higher AR or GDL thickness as the concentration overpotential smears the effect of the higher operating temperature.
Table 4.1 Channel and shoulder geometries for various ARs (unit: mm)
AR Channel Height
(Hc)
Channel Width (Wc)
Half Shoulder Width (Wsh)
0.5 0.57 1.14 0.23
0.75 0.69 0.92 0.34
1 .080 0.80 0.40
1.25 .089 0.72 0.45
1.5 0.98 0.66 0.47
Wc Hc
AR = /
Fig. 4.1 Definition of channel aspect ratio AR
0.5 1 1.5 Channel Aspect Ratio
0.6 0.8 1 1.2 1.4
Current Density (A/cm2)
0.14V 0.22 V
0.32 V
0.42 V
0.52 V
0.62 V
Fig. 4.2(a) Current density distribution at various cell voltages as function of AR
100 200 300 400 GDL thickness (micro m)
0.6 0.8 1 1.2 1.4
Current Density (A/cm2)
0.14V 0.22 V
0.32 V
0.42 V
0.52 V
0.62 V
Fig. 4.2(b) Current density distribution at various cell voltages as function of GDL thickness
0 0.0004 0.0008 0.0012 0.0016 Y (m)
-0.16 -0.164 -0.168 -0.172 -0.176 -0.18
Membrane Phase Potential (V)
12.8 12.9 13 13.1 13.2
Membrane Conductivity (S/m)
Membrane Phase Potential Membrane Conductivity
AR 0.5 AR 1 AR 1.5
Fig. 4.3(a) Distributions of potential and conductivity of membrane in the transverse direction of the middle X-Y plane for three values of AR at cell voltages of 0.62V
0 0.0004 0.0008 0.0012 0.0016 Y (m)
-0.15 -0.2 -0.25 -0.3 -0.35 -0.4 -0.45
Membrane Phase Potential (V)
12 12.4 12.8 13.2 13.6 14
Membrane Conductivity (S/m)
Membrane Phase Potential Membrane Conductivity
AR 1 AR 1.5 AR 0.5
Fig. 4.3(b) Distributions of potential and conductivity of membrane in the transverse direction of the middle X-Y plane for three values of AR at cell voltages of 0.14V
0 0.0004 0.0008 0.0012 0.0016 Y (m)
-0.16 -0.164 -0.168 -0.172 -0.176 -0.18
Membrane Phase Potential (V)
12.8 12.9 13 13.1 13.2
Membrane Conductivity(S/m)
Membrane Phase Potential Membrane Conductivity
152 micro m 203 micro m 356 micro m
Fig. 4.4(a). Distributions of potential and conductivity of membrane in the transverse direction of the middle X-Y plane for three values of GDL thickness at cell voltage of 0.62V
0 0.0004 0.0008 0.0012 0.0016 Y (m)
-0.15 -0.2 -0.25 -0.3 -0.35 -0.4 -0.45
Membrane Phase Potential (V)
12.4 12.8 13.2 13.6 14
Membrane Conductivity(S/m)
Membrane Phase Potential Membrane Conductivity 152 micro m
203 micro m 356 micro m
Fig. 4.4(b). Distributions of potential and conductivity of membrane in the transverse direction of the middle X-Y plane for three values of GDL thickness at cell voltage of
0.14V
0 0.0004 0.0008 0.0012 0.0016 Y (m)
-0.4 -0.42 -0.44 -0.46 -0.48
Solid Phase Potential (V)
-0.11 -0.115 -0.12 -0.125 -0.13
Activation Overpotential (V)
Solid Phase Potential Activation Overpotential AR 0.5
AR 1 AR 1.5
Fig. 4.5 (a). Transverse distributions of solid phase potential and activation overpotential at the interface between the cathode catalyst layer and the GDL for three values of AR at cell voltage of 0.62V
0 0.0004 0.0008 0.0012 0.0016 Y (m)
-0.6 -0.7 -0.8 -0.9
Solid Phase Potential (V)
0 -0.2 -0.4
Activation Overpotential (V)
Solid Phase Potential Activation Overpotential AR 0.5
AR 1 AR 1.5
Fig. 4.5 (b). Transverse distributions of solid phase potential and activation overpotential at the interface between the cathode catalyst layer and the GDL for three values of AR at cell voltage of 0.14V
0 0.0004 0.0008 0.0012 0.0016 Y (m)
-0.4 -0.42 -0.44 -0.46 -0.48
Solid Phase Potential (V)
-0.11 -0.115 -0.12 -0.125 -0.13
Activation Overpotential (V)
Solid Phase Potential Activation Overpotential 152 micro m
203 micro m 356 micro m
Fig. 4.6 (a). Transverse distributions of solid phase potential and activation overpotential at the interface between the cathode catalyst layer and the GDL for three GDL thicknesses at cell voltage of 0.62V
0 0.0004 0.0008 0.0012 0.0016 Y (m)
-0.6 -0.7 -0.8 -0.9
Solid Phase Potential (V)
0 -0.2 -0.4
Activation Overpotential (V)
Solid Phase Potential Activation Overpotential 152 micro m
203 micro m 356 micro m
Fig. 4.6(b). Transverse distributions of solid phase potential and activation overpotential at the interface between the cathode catalyst layer and the GDL for three GDL thicknesses at cell voltage of 0.14V
0 0.0004 0.0008 0.0012 0.0016 Y (m)
0 0.05 0.1 0.15 0.2 0.25
Saturation Level
X 0.005m X 0.025m X 0.045m
AR 0.5 AR 1 AR 1.5
Fig. 4.7(a). Effect of channel aspect ratio on transverse water saturation of the cathode GDL along channel direction at cell voltage of 0.62V
0 0.0004 0.0008 0.0012 0.0016 Y (m)
0 0.1 0.2 0.3 0.4
Saturation Level
X 0.005m X 0.025m X 0.045m AR 0.5 AR 1 AR 1.5
Fig. 4.7(b). Effect of channel aspect ratio on transverse water saturation of the cathode GDL along channel direction at cell voltage of 0.14V
0 0.0004 0.0008 0.0012 0.0016 Y (m)
0 0.05 0.1 0.15 0.2 0.25
Saturation Level
x 0.005m x 0.025m x o.o45m 152 micro m 203 micro m 356 micro m
Fig. 4.8(a). Effect of GDL thickness on transverse water saturation of the cathode GDL along channel direction at cell voltage of 0.62V
0 0.0004 0.0008 0.0012 0.0016 Y (m)
0 0.1 0.2 0.3 0.4
Saturation Level x 0.005m
x 0.025m x o.o45m 152 micro m 203 micro m 356 micro m
Fig. 4.8(b). Effect of GDL thickness on transverse water saturation of the cathode GDL along channel direction at cell voltage of 0.14V
0 0.0004 0.0008 0.0012 0.0016 Y (m)
2000 4000 6000 8000
Current Density (A/m2)
0 0.04 0.08 0.12 0.16
O2 Mass Fraction
o2 Mass Fraction Current Density AR 0.5
AR 1 AR 1.5
Fig. 4.9(a). Transverse distributions of oxygen mass fraction and local current density at the interface between the cathode catalyst layer and the GDL for three values of AR at cell voltage of 0.62V
0 0.0004 0.0008 0.0012 0.0016 Y (m)
0 4000 8000 12000 16000 20000
Current Density (A/m2)
0 0.02 0.04 0.06 0.08 0.1
O2 Mass Fraction AR 0.5
AR 1 AR 1.5
O2 Mass Fraction Current Density
Fig. 4.9(b). Transverse distributions of oxygen mass fraction and local current density at the interface between the cathode catalyst layer and the GDL for three values of AR at cell voltage of 0.14V
0 0.0004 0.0008 0.0012 0.0016 Y(m)
2000 4000 6000 8000
Current Density (A/cm2)
0 0.04 0.08 0.12 0.16
O2 Mass Fraction O2 Mass Fraction
Current Density 152 micro m 203 micro m 356 micro m
Fig. 4.10(a). Transverse distributions of oxygen mass fraction and current density at the interface between the cathode catalyst layer and the GDL for three GDL thicknesses at cell voltage of 0.62V
0 0.0004 0.0008 0.0012 0.0016 Y(m)
0 4000 8000 12000 16000 20000
Current Density (A/cm2)
0 0.02 0.04 0.06 0.08 0.1
O2 Mass Fraction 152 micro m
203 micro m 356 micro m O2 Mass Fraction Current Density
Fig. 4.10(b). Transverse distributions of oxygen mass fraction and current density at the interface between the cathode catalyst layer and the GDL for three GDL thicknesses at cell voltage of 0.14V
0.04 0.08 0.12 0.16 0.2 Oxygen Mass Fraction
Z (m)
0 0.4 0.8 1.2 1.6 2
Axial Velocity (m/sec)
0.0024 0.0026 0.0028 0.003
Axial Velocity O2 Mass Fraction 0.62V
0.14V
Cathode Channel
Cathode GDL
Fig. 4.11(a). Vertical distributions of fluid velocity and oxygen mass fraction in cathode channel and GDL for AR value of 0.5
0.04 0.08 0.12 0.16 0.2 Oxygen Mass Fraction
Z (m)
0 0.4 0.8 1.2 1.6
Axial Velocity (m/sec)
2
0.0024 0.0028 0.0032
Axial Velocity O2 Mass Fraction 0.62V
0.14V
Cathode Channel
Cathode GDL
Fig. 4.11(b). Vertical distributions of fluid velocity and oxygen mass fraction in cathode channel and GDL for AR value of 1.5
0.04 0.08 0.12 0.16 0.2 Oxygen Mass Fraction
Z (m)
0 0.4 0.8 1.2 1.6 2
Axial Velocity (m/sec)
0.0022 0.0024 0.0026 0.0028 0.003
Axial Velocity O2 Mass Fraction 0.62V
0.14V Cathode Channel
Cathode GDL
Fig. 4.12(a). Vertical distributions of fluid velocity and oxygen mass fraction in cathode channel and GDL for GDL thickness of 152 micro m
0.04 0.08 0.12 0.16 0.2 Oxygen Mass Fraction
Z (m)
0 0.4 0.8 1.2 1.6 2
Axial Velocity (m/sec)
0.0024 0.0028 0.0032
Axial Velocity O2 Mass Fraction 0.62V
0.14V
Cathode Channel
Cathode GDL
Fig. 4.12(b). Vertical distributions of fluid velocity and oxygen mass fraction in cathode channel and GDL for GDL thickness of 356 micro m
0.5 1 1.5
Channel Aspect Ratio
0.4 0.8 1.2
C u rr ent D e ns it y ( A /c m 2 )
353 K 343 K 333 K 0.14 V 0.62 V
Fig. 4.13. Effects of operating temperature on cell performance at various values of AR
100 200 300 400 GDL thickness (micro m)
0.4 0.8 1.2
Current Density (A/cm2)
353 K 343 K 333 K 0.14 V 0.62 V
Fig. 4.14 Effects of operating temperature on cell performance at various values of GDL thickness
0.2 0.3 0.4 0.5 0.6 0.7 0.8
Cell Potential (V)
0.4 0.8 1.2 1.6
Current Density (A/cm2) Torque (in-lbf)
100 125 150
Fig. 4.15 Experimental results of Lee et. al. (1999a) on the effect of compression force on cell performance
CHAPTER 5
ELECTROCHEMICAL REACTION AND PERFORMANCE OF PROTON EXCHANGE MEMBRANE FUEL CELLS WITH NOVEL CATHODE FLOW
CHANNEL SHAPE
5.1 Introduction
In previous chapter the roles of the transport component designs on the performance of PEMFCs are investigated in detail. Two factors, channel aspect ratio and GDL thickness are applied in the parametric modeling research to find the distributions of various transport variables such as local reactant concentration, water saturation and potential fields. Also the trends of cell performance variation are demonstrated and elucidated according to different operating voltages. The investigation indicates that the cell performance is dictated by the competition effect between electric current and reactant transport. According to the cell voltage, the electrochemical reaction is dominated by either electric conduction or reactant transport. That is, when the cell operates at medium reaction rate such as 0.62V, the role of reactant transport is not so important than electric conduction. Therefore, the transport component design which is beneficent to the electron transport delivers more current in such scenario. Consequently, a cell transport geometry with large channel
aspect ratio or a thin GDL thickness generate a relatively greater current in the reaction cites. However, with the increase of reaction rate, this trend is substituted by the reactant transport process as the requirement of high concentration reactant becomes more important, making the designs with small channel aspect ratio or large GDL thickness perform better at low cell voltage.
In this chapter, the investigation of the role of dominant mechanism in PEMFCs is extended to the quest of a novel channel geometry with variable shoulder/channel (S/C) ratio along the cathode channel. This channel configuration differs greatly from traditional channel design with a fixed S/C ratio throughout the cell domain. This is motivated from the result of previous investigation that the roles of electron and reactant transports dominate cell reaction at different situations.
However, a straight channel design has no the flexibility of manipulating the local distributions of these two factors.
Figure 5.1 (a) illustrates the computational domain of current study. It consists of nine essential components of a single cell. A PEM is sandwiched by catalyst layers (CLs), gas diffusion layers (GDLs), flow channels and bipolar plates (BPs) on anode and cathode sides. As shown in Fig. 5.1 (b), one feature of this paper is that through the assignment of two parameters Ws and Wc at cathode outlet port, the widths of channel and shoulder are varied continuously along the main stream direction.
Employed data for various simulation cases are shown in Tab. 5.2. According to the S/C ratio parameter, the channel configuration can be cataloged into divergent (case A, B), straight (case C) and convergent (case E, F) channels.