MASS TRANSPORT ENHANCEMENT IN PROTON EXCHANGE MEMBRANE FUEL CELL POH HEE JOO A thesis submitted for the degree of Doctor of Philosophy Department of Mechanical Engineering National University of Singapore 2009 Acknowledgement To Professor Arun S. Mujumdar, a great teacher, mentor and supervisor. I am very grateful to Professor Mujumdar for his constant encouragement, advice on the research direction and commitment to critically reviewing my thesis drafts. His in-depth knowledge to industrial and academic research, particularly in the field of Drying Technology and Heat & Mass Transfer has motivated me to pursue this study. Regular sessions with Prof Mujumdar’s Transport Process Research (TPR) Group kept the work progressing well and were very useful in stimulating new and innovative research ideas. To Dr Erik Birgersson, Asst. Professor in NUS with whom I discussed many ideas and learnt the CFCD and its model validation, especially during the most critical thesis writing period. To Rina Lum, Xing XiuQing, Wu Yanling and Narissara Bussayajarn from SERCPEMFC group, Agus Sasmito from NUS, Singapore, and Shaoping Li from ANSYS. Inc. I am grateful for the many discussions and email correspondences on the fuel cell issues. To my colleagues at IHPC (Dr Alex Lee, Dr LouJing, Mr. George Xu, Dr. Cary Turangan, Dr Chew Choon Seng and etc). I am thankful to all of you for helping me in one way or another in the work relating to computational modeling. To Dr Kurichi Kumar, as he has motivated me greatly during the first year of this study. His leaving IHPC was a great loss to me professionally. To my church cell group members, as they constantly offered prayer and encouragement to me. Last but not least, I would like to express my love and appreciation to my wife, Cherrie, for her endless support throughout this study and our whole life. i Contents Page Acknowledgement i Contents ii Summary viii List of Figures x List of Tables xviii Nomenclature xx Chapter One: Introduction 1.1 Fuel Cell Overview 1.2 Fuel Cell Thermodynamics and Electrochemistry 3 1.2.2 Fuel Cell Performance 1.3 1.2.1 Theoretical Limit Proton Exchange Membrane (PEM) Fuel Cell 1.3.1 Components 1.3.2 Operating Principles 1.3.3 Water Transport 11 1.3.4 Mass Transport Limitations 12 1.4 Research Objectives and Methodology 13 1.5 Thesis Outline 15 Chapter Two: Literature Review, Objectives and Methodology 16 2.1 Review of Prior Publications on PEMFC Models 16 2.2 Computational Fuel Cell Dynamics (CFCD) 18 ii 2.3 Multiphase Model in PEMFC 23 2.4 Prior Work on CFD Model Validation 25 2.5 Mass Transport Enhancement Techniques 30 2.6 Air Breathing PEMFC 33 33 2.6.2 Computational Modeling Studies 34 2.6.3 Experimental Studies 36 2.7 2.6.1 Analytical Studies Closing Remarks 37 Chapter Three: PEMFC Modeling 39 3.1 Model Assumptions and Simplifications 39 3.2 Governing Conservation and Constitutive Equations 42 3.2.1 Conservation of Mass 43 3.2.2 Conservation of Momentum 43 3.2.3 Conservation of Energy 44 3.2.4 Conservation of Non-Charged Species 46 3.2.5 Conservation of Charged Species 48 3.2.6 Conservation of Liquid Water Saturation 49 3.2.7 Phenomenological Membrane Water Transport Equations 50 3.3 Coupling of the Transport Equations 53 3.4 Boundary Conditions 58 3.5 Closing Remarks 59 Chapter Four: PEMFC Model Validation 4.1 60 Introduction 60 iii 4.2 Results of Parametric Study 60 4.1.1 Geometric Model 60 4.1.2 Effect of Electrochemical Parameters 61 4.1.3 Effect of Operating Pressure 64 4.1.4 Multiphase Model Results For Liquid Water Saturation 64 4.1.5 Effect of cathode bipolar plate electrical conductivity 66 4.1.6 Effect of membrane thickness 68 4.1.7 Constant stochiometric ratio 69 4.1.8 Constant Relative Humidity 71 4.3 Experimental Uncertainty and Reproducibility 73 4.3.1 Comparison with Temasek Poly data 73 4.3.2 Comparison with experimental data of Wang et al (2003) 76 4.3.3 Comparison with experimental data of Ticianelli et al (1988) 77 4.3.4 Comparison with experimental data of Noponen et al (2004) 78 4.3.1.1 3D Simulation model 78 4.3.1.2 Simulation model for low thermal conductivity of porous net flow distributor 4.4 80 Model Validation with Experimental Data from Noponen et al (2004) 84 4.4.1 Geometrical and Computational Model 84 4.4.2 Model Parameters 86 4.4.3 Boundary Conditions 89 4.4.4 Mesh Independence Study 90 4.4.5 Validation Results 93 iv 4.4.5.1 Global Polarization Curves 93 4.4.5.2 Local Current Density Variations and Drop at Entrance Region 96 4.4.5.3 Local Temperature Distributions and Liquid Saturation Factor 103 4.5 4.4.5.4 Multiphase Model Results 105 4.4.5.5 Effect of Cathode Relative Humidity 106 Closing Remarks 107 Chapter Five: Enhanced Performance Using Impinging Jet Concept in PEMFC 109 5.1 Introduction 109 5.2 2D Cathode Side Single Impinging Jet Design 112 5.2.1 Effect of GDL and net flow distributor permeability 113 5.2.1.1 Results for Net Flow Distributor permeability 1e-10m2 (Case and 2) 122 5.2.1.2 Results for Net Flow Distributor permeability 1e-08m2 (Case and 4) 126 5.2.2 Effect of Cathode Stoichiometric Ratio, c 131 5.2.3 Effect of Relative Humidity 142 5.2.4 Effect of Porous Net Flow Distributor Thickness and Impinging Jet Width 144 5.2.5 Effect of Anode/Cathode Impinging Jet 5.3 2D Multiple Impinging Jet Configuration 147 149 5.3.1 Effect of Jet Multiplicity 149 5.3.2 Effect of Stoichiometric Ratio in Multiple Impinging jet 152 5.3.3 Effect of Alternating Jet Impingement Inlet and Suction Outlet 154 v 5.4 2D Cross Flow Jet 155 5.5 Closing Remarks 157 Chapter Six: Enhanced Performance for Self Air Breathing PEM Fuel Cells 159 6.1 Overview 159 6.2 Physical and Mathematical Aspects in ABFC 160 6.2.1 Working Principle 160 6.2.2 Operational Problems 161 6.2.3 Transport Phenomena 162 6.2.4 Modeling Challenges 164 6.3 Motivation and Objective 166 6.4 Model Validation 168 6.5 Simulation Methodology 174 6.6 2D CFCD Simulation 175 175 6.6.2 Results and Discussion 176 6.7 6.6.1 Geometry and Computational Model 6.6.2.1 Global and Local Results 176 6.6.2.2 Effect of Channel Height and Length 186 6.6.2.3 Effect of Device Orientations 190 6.6.2.4 Performance Durability 194 3D CFCD Simulation 196 6.7.1 Geometry and Computational Model 196 6.7.2 Comparison between 2D and 3D 198 6.7.3 Effect of ambient temperature 202 vi 6.7.4 Comparison between channel and planar ABFC with perforations 203 6.7.5 Effect of orientation for planar ABFC with perforations 211 6.7.6 Effect of bipolar plate thickness in planar ABFC with perforations 215 6.7.7 Comparison between planar ABFC full and segmented perforations 219 6.8 Closing Remarks 225 Chapter Seven: Conclusions and Suggestions for Further R&D 227 References 231 Relevant Publications 243 Appendix 1: Summary of PEMFC publications with commercial CFCD software 246 Appendix 2: FLUENT User Defined Function for Modified Heat Source Term 248 Appendix 3: FLUENT User Defined Function for Membrane Properties Adaptation 250 Appendix 4: FLUENT User Defined Function for Constant RH Boundary Condition 252 Appendix 5: Geometrical Description for Experimental Data from TP, Singapore (2008) 254 Appendix 6: Experimental Work associated with undergraduate students vii 256 Summary This thesis presents Computational Fuel Cell Dynamics (CFCD) approaches to analyze the enhanced performance of typical forced convection and self air-breathing PEM fuel cells (ABFC). The mathematical framework used in the simulation is a comprehensive two/three dimensional, multi-component, multiphase, non-isothermal, time-dependent transport computation model, performed using the commercial CFD software (FLUENT 6.3.16) with a PEMFC add-on module and self-developed user subroutines. User Defined Functions are developed for the simulation code for constant relative humidity, stoichiometric ratio and entropy irreversibility heat source generation. This model is validated on the basis of close agreement with relevant published experimental data for both forced and free convection PEMFC. For forced convection fuel cells, a flow structure which delivers the reactant transversely to the membrane electrode assembly (MEA) using an impinging jet configuration on the cathode side is proposed. The flow structure is modeled to examine its effectiveness to enhanced fuel cell performance, especially at high current densities. Larger flow rate is found to deteriorate PEMFC performance due to membrane dehumidification. A single impinging jet outperforms the conventional channel flow configuration by 80% at high current densities. A multiple impinging jet design is further suggested as an effective way to achieve flow and species uniformity; this results in a more uniform and higher catalyst utilization. It can also lower the fuel cell temperature and alleviate flooding as the fresh reactant from each jet can remove excess water vapor. viii Compared to a single impinging jet, a multiple jet gives up to 14% predicted enhancement at a high current density of about A/cm2. For the self air-breathing PEMFC (ABFC), the effect of geometric factors (e.g. channel length and height), device orientation (horizontal, vertical or an inclined angle), and O2 transfer configuration (channel vs. planar) have been investigated using the validated model. When anode inlet is fully humidified, electro-osmotic drag (EOD) outweighs back-diffusion for water transport across the membrane. The planar airbreathing fuel cell can outperform the channel design by about 5%. The channel airbreathing fuel cell prefers larger openings whereas the planar prefers the opposite. This new finding establishes the relationship between dominant mass transport modes with the length scale of fuel cells. Based on the simulation results, an optimum design for the airbreathing fuel cell is proposed. Finally, this thesis seeks to give a better understanding of design for the enhanced performance of PEMFC (both forced convection and air-breathing fuel cells). This requires the optimal combination of improved reactant mass transport for the electrochemical reaction and keeps the right membrane water content for ionic transfer without causing flooding of the gas diffusion layer. ix Publication CFD Packages Atul Kumar and R. G. Reddy “Modeling of polymer electrolyte FLUENT 5.5 membrane fuel cell with metal foam in the flow-field of the bipolar/end plates” Journal of Power Sources, Volume 114, Issue 1, 2003, pp 54-62 Hua Meng and Chao-Yang Wang, “Large-scale simulation of polymer STAR-CD 2001 electrolyte fuel cells by parallel computing” Chemical Engineering Science, Volume 59, Issue 16, August 2004, Pages 3331-3343 10 Phong Thanh Nguyen , Torsten Berning and Ned Djilali “Computational CFX 4.3 model of a PEM fuel cell with serpentine gas flow channels” Journal of Power Sources, Volume 130, Issues 1-2, May 2004, Pages 149-157 11 S. M. Senn and D. Poulikakos, “Polymer Electrolyte Fuel Cells With CFD-ACE Porous Materials as Fluid Distributors and Comparisons With Traditional Channeled Systems” J. Heat Transfer 126, 410 (2004) 12 S. Shimpalee, S. Greenway, D. Spuckler and J. W. Van Zee, “Predicting STAR-CD water and current distributions in a commercial-size PEMFC”, Journal of Power Sources, Volume 135, Issues 1-2, September 2004, Pages 7987 13 N. P. Siegel, M. W. Ellis, D. J. Nelson and M. R. von Spakovsky CF-Design “Single domain PEMFC model based on agglomerate catalyst geometry” Journal of Power Sources, Volume 115, Issue 1, 27 March 2003, Pages 81-89 14 B.R. Sivertsen and N. Djilali, CFD-based modelling of proton exchange FLUENT membrane fuel cells Journal of Power Sources, Volume 141, Issue 1, 2005, Pages 65-78 247 University/Institutes Department of Metallurgical and Materials Eng, The University of Alabama, Tuscaloosa, USA Department of Mechanical and Nuclear Engineering, Electrochemical Engine Center (ECEC), The Pennsylvania State University, University Park, PA 16802, USA Institute for Integrated Energy Systems, University of Victoria, Victoria, Canada BC V8W 3P6 Institute of Energy Technology, Swiss Federal Institute of Technology Zurich, ETH Zentrum, CH-8092 Zurich, Switzerland Department of Chemical Engineering, University of South Carolina, Columbia, SC 29208, USA Department of Mechanical Engineering, Virginia Polytechnic and State University, Blacksburg, VA, USA Institute of Integrated Energy Systems, University of Victoria Appendix 2: FLUENT User Defined Function for Modified Heat Source Term RT dE rev The following heat source T ; E rev  1.229  0.83 x10 3 T  298  ln xO2 has dT 4F been used to replace the heat formation (reaction) of water. The procedure consists of the following steps: 1. Change the fraction of the energy released in the chemical reaction for the formation of water as heat energy from 0.2 to 0.0. Original setting real heat_apportionment_factor(cell_t c, Thread *t) { return 0.2; Modified Setting real heat_apportionment_factor(cell_t c, Thread *t) { return 0.0; 2. Compile the pem_user.c under fuelcellcells2.2 directory using this command  nmake /f makefile_master-client.nt (in Windows OS)  make –f Makefile-client FLUENT_ARCH=lnamd64 (in LINUX with AMD 64 bit processor) 3. Hook up the relevant UDF file using DEFINE-User Defined-Function-Manage 4. Write the following UDF for the heat source term relevant to the reversible cell potential: /* UDF for specifying heat generation in the active layers related to the reversible cell potential*/ DEFINE_SOURCE(energy_rev,c,t,dS,eqn) { real C1 = 8.3e-4; /*constant is E_rev*/ real C2 = 2.1542e-5; /*constant in E_rev, R/4F*/ 248 real MWO2 = 0.032; /*molecular weight of oxygen*/ real MWH2O = 0.018; /*molecular weight of water vapor*/ real MWN2 = 0.028; /*molecular wieght of nitrogen*/ real ave_MW; real mol_frac_O2; real source; ave_MW = 1/((C_YI(c,t,O2)/MWO2) + (C_YI(c,t,H2O)/MWH2O) (C_YI(c,t,N2)/MWN2)); mol_frac_O2 = ave_MW*C_YI(c,t,O2)/MWO2; source = C_T(c,t)*(C1 + C2*log(mol_frac_O2))*C_UDMI(c,t,13); dS[eqn] = (C1 + C2*log(mol_frac_O2))*C_UDMI(c,t,13); return source; } + 5. Hook up the relevant source term file using DEFINE-User Defined-FunctionInterpreted (make sure the c file is under the same directory as your case file) 6. Run the program 249 Appendix 3: FLUENT User Defined Function for Membrane Properties Adaptation In this simulation, m is 0.625, and UDF has been implemented for this adaptation. The original value of membrane proton conductivity and water content diffusivity in pem_user.c is multiplied by 0.625 for necessary adaptation from default NAFION® to GORE-SELECT membrane. real Membrane_Conductivity(real lam, cell_t c, Thread *t) { real T = C_T(c,t); real eee = 1268.0*(T-303.0)/(T*303.0); lam = MAX(1.0,lam); /* Springer et al 1991: The following * correlation works only for lam>=1 * below 1.0, use constant value. */ return 0.625*alpha_m*exp(eee)*pow((0.514*lam-0.326),beta_m); } real Water_Content_Diffusivity(real lam, real T, real mem_mol_density, cell_t c, Thread *t) { real diff_w; if (1) /* reference --cited by CY Wang */ { if (lam[...]... catalyst layer is a thin agglomerate-type structure where electrochemical reactions occur The catalyst in PEMFC is usually made of platinum and its alloys Fine particles of catalyst are dispersed on a high-surface area carbon in the active layer of the electrode in order to minimize platinum loading Its performance is characterized by surface area of platinum by mass of carbon support and the typical... porosity Flow channels (parallel, serpentine, inter-digitated and etc) are machined in graphite plates to feed the reactant gases to the GDL Optimum flow channel area can be determined, as in some cases a larger channel area is required for minimal gas transport pressure loss However, a larger land area contact between bipolar plate and GDL is necessary for minimum electrical contact resistance and ohmic... certain disadvantage of fuel cells lies in making it commercially available for consumer usage These include high cost of fuel cell, low volumetric power density compared to I.C engines and batteries, and the various issues regarding safety, availability, storage and distribution of pure hydrogen fuel 2 1.2 Fuel Cell Thermodynamics and Electrochemistry A study of thermodynamics uncovers the “ideal case”... Comparison of local current density, Iave = 1 A/ cm2, along anode bipolar plate for single impinging jet with different c Comparison of local O2 concentration, Iave = 1 A/ cm2, along cathode catalyst/GDL for single impinging jet with different c Comparison of membrane water content, Iave = 1 A/ cm2, along cathode catalyst /membrane for single impinging jet with different 127 c Comparison of impinging... 40-55% 45-60% Generally, fuel cell applications have the unique advantage of being quiet (no moving parts) and clean (reduced air pollution and green house emissions such as NOx and SOx) They also enable improved efficiency for transportation, allow independent scaling between power (determined by fuel cell size) and capacity (determined by fuel storage size), and have a low temperature start-up (e.g 600C... (Larminie and Dirks, 2003) Practically, in a portable PEMFC, the thickness of bipolar plate is about 3mm, while the footprints that sandwiching MEA is approximately 5cm x 5cm 1.3.2 Operating Principles From Figure 1.2, the basic operating principle involves four physical transport phenomena described as follow Many illustrations can be found in Larminie and Dicks (2003), Barbir (2005), O'Hayre et al... Comparison of O2 concentration between Cases 8 and 10 (a) O2 contour (b) O2 profile along cathode GDL/catalyst Comparison of water content at cathode catalyst /membrane interface between Cases 8 and 10 Comparison of current density along anode bipolar plate between Cases 8 and 10 Geometry and boundary conditions description for full and segmented planar perforated ABFC (a) planar ABFC with full perforation... case” for fuel cell performance while an analysis of electrochemistry reveals its the kinetic limitations and defines the “practical case” for fuel cell performance This section summarizes the important equations used in fuel cell thermodynamics and electrochemistry and gives a basic understanding of fuel cell performance and their characteristics 1.2.1 Theoretical Limit Fuel cell thermodynamics provides... of water by evaporation from the membrane Membrane thickness is also important as a thinner membrane minimizes ohmic resistance losses but risks hydrogen cross-over to cathode, producing parasitic currents Typical membrane thickness is in the range of 5 – 200m (Kolde et al, 1995) The role for membrane in PEMFC is to provide ionic conduction, reactant separation and water transport 2 Catalyst Layer... fuel cell deterioration 6 1.3 Polymer Electrolyte Membrane (PEM) Fuel Cells 1.3.1 Components Figure 1.2 shows a 2D schematic diagram to illustrate various components and operating principles of a single PEMFC Load e- C Anode Catalyst Anode Bipolar Plate Anode Electrode (GDL) Cathode Catalyst B H2 channel A Cathode Cathode Electrode Bipolar Plate (GDL) Membrane C B H+ H2 D Air A channel H2O H2O O2 D . validation, especially during the most critical thesis writing period. To Rina Lum, Xing XiuQing, Wu Yanling and Narissara Bussayajarn from SERC- PEMFC group, Agus Sasmito from NUS, Singapore, and. results in a more uniform and higher catalyst utilization. It can also lower the fuel cell temperature and alleviate flooding as the fresh reactant from each jet can remove excess water vapor (CFCD) approaches to analyze the enhanced performance of typical forced convection and self air-breathing PEM fuel cells (ABFC). The mathematical framework used in the simulation is a comprehensive