PEMFCs CO Tolerance and Direct Methanol Operation

Transcription

1 PEMFCs CO Tolerance and Direct Methanol Operation J. M. Fenton 11/20/02 1

2 Cross Section of Proton Exchange Membrane Fuel Cell 2

3 Performance Characteristics of a Fuel Cell 3 Cell Voltage (V) Polarization (V) Polarization (V) Cathode Loss Cathode Loss Membrane Internal Resistance Loss Membrane Internal Resistance Loss Anode Loss Current Density (ma/cm 2 )

4 Key Overpotential Sources Electrode losses - Kinetic losses - Transport losses - Resistive losses: ionic resistance Membrane internal resistance (IR) losses - Includes contact resistances Transport losses within the diffusion layers Objective: Minimize Losses; Ensure Longevity 4

5 Drawbacks of Direct Hydrogen Operation Hydrogen infrastructure not yet available Systems issues (automobile): - fuel storage - weight and volume for reasonable range - ease of refueling? - safety? Direct hydrogen (pure hydrogen) does not seem feasible (and / or economically viable) in the near future 5

6 Other Options?? Indirect Hydrogen Reform readily available liquid fuel (or natural gas for stationary applications) Use H 2 thus synthesized in the fuel cell In-situ, continuous generation Mature technology (reforming) Direct Methanol Feed methanol (and water) directly into fuel cell anode Oxidize methanol (instead of hydrogen) at the anode Methanol may be fed as a liquid or as a vapour Greatly simplifies system and infrastructure issues 6

7 Gottesfeld 7

8 Indirect Hydrogen Effect of CO Reformate stream contains CO and CO 2 CO can be minimized by shift conversion (0.4 2 % CO), and further reduced by partial oxidation (< 100 ppm CO) However, even 10 ppm of CO detrimental to electrocatalyst adsorbs on active sites - increases anode overpotential Clearly, need better fuel processing (and) CO tolerant electrocatalysts (and) a technique to minimize CO adsorption 8

9 Effect of CO in Reformate on Performance Pressurized Operation?? Gottesfeld 9

10 Effect of CO in Reformate on Performance Perf.-H 2 Cell Voltage(V) Resistance-H 2 Perf.-H ppmCO Resistance-H ppmCO Perf.-H ppmCO Resistance-H ppmCO Perf.-H ppmCO Resistance-H ppmCO Resistance(Ohm-cm 2 ) Current density(ma/cm 2 ) Cell Performance with Various CO Concentrations at 80 o C and 1atm Anode:H 2 +CO at 90 o C, 40%Pt-Ru/C; Cathode: O Si 10

11 Why the Performance Loss? E cell = E cathode E anode Thermodynamically: E cathode = 1.23 V (O 2 +4 H + + 4e - = 2H 2 O) E anode = 0 V (H 2 = 2H + +2e - ) Therefore E cell (max) = 1.23 V In the presence of CO, two electrochemical reactions occur: Pt + CO CO/Pt H Pt 2 H/Pt (rate-limiting) H /Pt Pt + H + +e - CO/Pt + OH ads Pt + CO 2 + H + + e - Gottesfeld 11

12 Anode overpotential determined by the relative contribution of CO oxidation Low currents hydrogen requirements met by adsorbed H 2 High currents (or large CO concentrations) adsorbed H 2 insufficient for faradaic requirements Therefore, CO removal (via electrooxidation) has to occur higher anode overpotential, higher E anode lower E cell Key Strategy: minimize CO adsorption on catalyst 12

13 Anode Overpotentials of 40% wt Pt-Ru/C (Pt/Ru = 1/1) with H containing 2 CO (10.4 to 485 ppm) at 80 o C (100% R.H.). Over-potential (mv) H 2 H ppm CO H ppm CO H ppm CO H ppm CO H ppm CO High-potential 200 High-slope 100 Recall: Performance characteristics chart contribution of anode overpotential (slide 3) Current Density(mA/cm 2 ) Low-potential Note: increasing overpotential with: increasing CO concentration and increasing currents Si, Jiang 13

14 How to Deal with CO? Air Bleed (or oxygen bleed) Better (more CO tolerant) anode electrocatalysts Elevated temperature operation Each of above briefly discussed in forthcoming slides Note improvements in reforming, shift conversion and partial oxidation steps are also of great use however, these approaches are not discussed in this lecture 14

15 Air / Oxygen Bleed CO surface coverage biggest problem Bleeding oxygen (as oxygen or as air) into the fuel stream helps: - CO poisoning Pt sites are oxidized in the presence of free platinum to give CO 2 CO/Pt +O 2 + Pt O/Pt + CO/Pt ½ O 2 + H 2 O/Pt + CO/Pt CO 2 +2Pt H 2 O - The strong preferential adsorption of CO on Pt actually helps this mechanism! Gottesfeld 15

16 Demonstration of Air Bleed Efficacy Gottesfeld 16

17 Demerits of Air Bleed Technique Explosive limit of O 2 in H 2 is 5%. This limits amount of CO tolerated to ~100ppm Some loss in fuel efficiency due to chemical oxidation of hydrogen This loss increases as amount of oxygen introduced increases (2 fold increase) thus the larger the amount of CO in stream, the larger the loss of fuel efficiency 17

18 CO Tolerant Electrocatalysts Recall: mechanism of CO and H 2 oxidation in a mixed stream: Pt + CO CO/Pt H Pt 2 H/Pt (rate-limiting) H /Pt Pt + H + +e - CO/Pt + OH ads Pt + CO 2 + H + + e - Now, for CO electrooxidation, the catalyst site must be hydrated (have an attached hydroxyl group on its surface) Ru + H 2 O Ru-OH ads + H + + e- Pt + H 2 O Pt-OH ads + H + + e- The potential at which this group is generated varies from catalyst to catalyst: V for Pt - ~0.2 V for Ru! 18

19 Thus, catalysts with Ru will have a lower anodic overpotential at high currents (or high CO concentrations) than catalysts containing Pt Ru not a very good catalyst for H 2 oxidation Pt-Ru alloys have been successfully used However limitations such as increasing anode overpotential with increasing CO concentration and increasing currents are not eliminated by this approach 19

20 The extent of CO tolerance depends greatly on the catalyst structure and formulation Even for the best catalysts, the improved CO tolerance all but vanishes for high currents However, a combination of precisely formulated catalyst (typically 1:1:: Pt:Ru) and air / oxygen bleed has been found to be effective at high currents as well 20

21 Elevated Temperature Operation CO adsorption on Pt is an exothermic process By the Le-Chatlier Braun principle, increasing the system temperature favours the endothermic CO desorption reaction Effect of increasing system temperature is to lower the fraction of catalyst covered with CO, thereby lowering anode overpotential The effect has been clearly demonstrated 21

22 CO+Pt = CO-Pt (associative adsorption) H 2 +2 Pt = 2 H-Pt (dissociative adsorption) Fractional coverage (f) of CO and H given by: f CO = K CO P CO /[1+ K CO P CO +K 0.5 H P 0.5 H ] f H = K 0.5 H P 0.5 H /[1+ K CO P CO +K 0.5 H P 0.5 H ] K = equilibrium constants, P = partial pressures As T increases, f H increases as H adsorption is less exothermic than CO adsorption, and because H adsorption requires 2 sites as opposed to one for CO adsorption Yang et. al 22

23 Chemisorption at T/chemsorption at 35 o C(%) JM 40%wt Pt/C E-Tek 40%wt Pt-Ru/C Temperature ( o C) Effect of temperature on CO chemisorption on potential Anode catalysts note also the exceptional CO tolerance of Pt-Ru alloys when compared to Pt Si 23

24 CO Coverage on the Pt-Ru Catalyst Surface at Various Temperatures o C 105 o C 120 o C CO Coverage (%) CO concentration in H 2 (ppm) Si 24

25 Over-potential (mv) H 2 H ppm CO H ppm CO H ppm CO H ppm CO H ppm CO High-potential 200 High-slope 100 Over-potential (mv) H 2 H ppm CO H ppm CO H ppm CO H ppm CO H ppm CO High-Potential 200 High- Slope 100 Over-potential (mv) H 2 H ppm CO H ppm CO H ppm CO H ppm CO H ppm CO High-potential 200 High-slope 100 Low-potential Current Density(mA/cm 2 ) Low-Potential Current density (ma/cm 2 ) Low-potential Current Density (ma/cm 2 ) Anode overpotentials vs. current density and CO concentration on Pt-Ru catalyst Similar scales, increasing temperature from left to right (80C, 105C and 120C) Performance curves follow Si 25

26 IR-free Voltage (V) pure H 2 H ppm CO H ppm CO H ppm CO H ppm CO H ppm CO Current Density (ma/cm 2 ) IR-Free H 2 /O 2 Cell Performance with Different CO Concentrations at 120 o C on Pt-Ru Si 26

27 Pure H2 H2+10.4ppm CO Cell Voltage (V) Perf.-80 o C Resistance-80 o C Perf.-105 o C Resistance-105 o C Perf.-120 o C Resistance-120 o C Resistance (Ohm-cm 2 ) Cell Voltage (V) Perf.-80 o C Resistance-80 o C Perf.-105 o C Resistance-105 o C Perf.-120 o C Resistance-120 o C Resistance (Ohm-cm 2 ) Current Density (ma/cm 2 ) 1.0 H2+104ppm CO Current Density (ma/cm 2 ) H2+485ppm CO Cell Voltage (V) Perf.-80 o C Resistance-80 o C Perf.-105 o C Resistance-105 o C Perf.-120 o C Resistance-120 o C Resistance (Ohm-cm 2 ) Cell Voltage (V) Perf.-80 o C Resistance-80 o C Perf.-105 o C Resistance-105 o C Perf.-120 o C Resistance-120 o C Resistance (Ohm-cm 2 ) Current Density (ma/cm 2 ) 1 atm, H 2 /O 2, NTZHP membrane Si Current Density (ma/cm 2 ) 27

28 Can we extend this infinitely? NO!!! Materials issues rise to the fore especially the ionomeric membrane in a PEMFC Note: previous figures indicated 105 C to be a better CO tolerant operating temperature than 120 C contrary to expectation based on Le-Chatlier- Braun principle This apparent contradiction effect of membrane resistance, cathode overpotential and system water content. These issues will be discussed in the following slides Note membrane conductivity (p) determines its resistance at any given condition for a given thickness (t) and active area (A)(R = p t/a) 28

29 Temperature and Relative Humidity Linked to one another Maintaing a saturated environment above 100 C requires system pressurization Leads to parasitic power losses and complex systems Need exists to develop membranes for high temperature / low relative humidity operations Is proton conductivity influenced by temperature and water content?? 29

30 Limitations of Current PEM Technology Conductivity strong function of water content Drops in under saturated environments Increased membrane and electrode resistance at High T / Low RH Conductivity vs. T and RH - Nafion 112 Fractional RH Conductivity (S/cm) Temperature (C) Zawodzinski/Gottesfeld, Ramani 30

31 Conductivity Mechanisms Vehicular mechanism Proton attached to solvent ( vehicle ) molecule e.g. H 3 O + Moves at rate of vehicular diffusion Vehicle counter diffusion Net proton transport governed by vehicle diffusion rates Grotthuss mechanism Also called hopping mechanism Stationary vehicles (only local motion) Proton hops from vehicle to vehicle Always within H bond environment Solvent reorientation provides H + pathway Continuous motion Columban, Kreuer 31

32 Gierke Cluster Network Model for Nafion Columban 32

33 Conductivity in Nafion LT / 100% RH High water uptake Combined vehicular / Grotthuss mechanisms Large water content symmetric environment Easy, quick reorientation Large cluster diameters (4 nm); large interconnecting pores (~ 1nm) Good diffusional transport Fast hopping High conductivity! HT / LRH Low water uptake Cluster shrinks (~ 2.4 nm) Hopping difficult Proton transport vehicular mechanism Pore narrowing Poor diffusional transport Low conductivity! Giereke/Hsu, Kreuer, Columban 33

34 Alternate Strategy Nafion composite membranes Incorporation of inorganic additives to Nafion matrix Additives used Heteropolyacids (HPAs), layered phosphates, metal oxides, etc. Conductivity (S/cm) Nafion atm. pressure Nafion PTA STA SMA Nafion / PTA 80 C 100 C 120 C Membrane Nafion vs. composite membranes Ramani 34

35 Cell Voltage (V) Perf.-Zr(HPO 4 ) 2-80 o C Resistance-Zr(HPO 4 ) 2-80 o C Perf.-Nafion o C Resistance-Nafion o C Perf.-Zr(HPO 4 ) o C Resistance-Zr(HPO 4 ) o C Perf.-Nafion o C Resistance-Nafion o C Resistance (Ohm-cm 2 ) Current Density (ma/cm 2 ) H 2 /O 2, 1 atm Effectiveness of composite membranes at high temperatures and low relative humidities Si 35

36 The development of such composite membranes permits operation at higher temperatures though resistive losses are still greater than at 80 C The temperatures currently attainable at ambient pressure (130 C) allow operation (in conjunction with CO tolerant catalysts) with up to 100ppm CO with minimal losses (when compared to operation with pure H 2 at 130 C) This approach can be combined with techniques such as air-bleed for greater efficacy Further improvement hinges on improved membranes and electrocatalysts 36

37 Effect of Water Content on CO Tolerance Recall: CO oxidation requires the generation of hydroxyl (OH - ) groups on the catalyst surface Such groups are generated by the oxidation of water Thus, better CO tolerance can be achieved under well hydrated conditions Trade off exists between Temperature (and lower surface coverage) and humidity (and more hydroxyl groups generated on catalyst)!! 37

38 Effect of CO on Membrane Resistance Resistance: Membrane resistance(ohm-cm 2 ) o C-100%R.H 105 o C-50%R.H. 120 o C-30%R.H CO concentration in H 2 (ppm) - Constant at 100% RH, -Increases slightly with CO concentration at 50% RH - Increases perceptibly at 30% RH Si 38

39 Why Does Resistance Increase? The oxidation of CO to CO 2 will occur at a rate determined by the current output of the cell Thus, all available water is used up (to generate hydroxyl groups) at a particular CO concentration Any increase in CO concentration will result in water being sucked out from the membrane to support CO oxidation thereby increasing membrane resistance The CO concentration at which this starts to occur is lower at lower relative humidities 39

40 Effect of CO 2 on PEMFC Performance CO 2 neither chemically nor electrochemically inert! Can be chemically reduced to give CO (reverse water gas shift: CO 2 + H 2 = CO+ H 2 O) Can be electrochemically reduced to give CO CO 2 + 2H e - = CO+ H 2 O Approaches similar to those adopted for CO tolerance have been shown to improve CO 2 tolerance as well 40

41 Effect of CO 2 Treatment Using Air-Bleed Gottesfeld 41

42 Direct Methanol Fuel Cells 42

43 Rationale for Direct Methanol Operation Greatly simplified system design Readily available fuel infrastructure High fuel energy density lowered system weight and volume Ideal for mobile applications such as laptops, cellular phones Also of great interest to the military to power individual soldiers electronics 43

44 Gottesfeld 44

45 Gottesfeld 45

46 Comparison of Performance between H 2 PEMFC and DMFC with Nafion 117 at 60 o C and 1atm 1.0 Cell Voltage(V) DMFC-4mg/cm 2 catalyst 1M MeOH/Air H 2 PEMFC-0.4mg/cm 2 catalyst H 2 /Air Current density(ma/cm 2 ) Si 46

47 Principle Challenges in DMFC Operation Sluggish methanol oxidation (anode) kinetics: - 6 electron transfer as opposed to 2 electron transfer for H 2 oxidation - formation of CO as an intermediate in the multi-step methanol electrooxidation mechanism poisoning of catalyst Large methanol crossover through the membrane: - linked to the electro osmotic drag - has detrimental effect on fuel efficiency - may poison the cathode - creates mass transport problems at cathode layer by wetting hydrophobic gas channels, leading to increased flooding. CO 2 removal at anode 47

48 Breakup of DMFC Losses E cell = E cathode - E anode = = 1.2 V (Thermodynamic) Hoogers 48

49 Anode Kinetics Thermodynamically, methanol oxidation and hydrogen oxidation occur at nearly the same potential (0.046 V and 0 V respectively) However, hydrogen oxidation is a 2 electron process, while methanol oxidation is a 6 electron process. It is very unlikely that all 6 electrons are transferred at the same time Therefore, transfer occurs step by step, leading to the formation of intermediates 49

50 Proposed Methanol Oxidation Mechanisms Pt + MeOH = Pt-MeOH = Pt-CO ads (methanol adsorption through a series of steps, see figure) Pt + H 2 O = Pt-OH ads + H + + e - (generation of hydroxyl groups on catalyst) Pt-CO ads + Pt-OH ads = 2Pt +CO 2 + H + + e - (Oxidation of CO to CO 2 - similar to CO oxidation in direct hydrogen systems) Hoogers 50

51 Postulated Methanol Adsorption Mechanism Note: Final stage is CO adsorbed on Pt sites Hoogers 51

52 Below 450 mv, Pt surface entirely poisoned by CO No further methanol adsorption Further adsorption requires CO electrooxidation induces overpotential of 450 mv or greater for Pt catalysts Thus, E cell = ~0.45 = ~ 0.8 V max. even at low currents! Effect of anode overpotential contributes to poor methanol performance seen in performance curve 52

53 Breakup of DMFC Losses E cell = E cathode -E anode = = 0.8 V (small currents) Hoogers 53

54 Direct Methanol vs. Direct Hydrogen Direct Methanol Pt-CO formation due to adsorption of methanol and subsequent intermediate formation Pt-CO inhibits further methanol adsorption Large currents requires CO electrooxidation to free catalyst sites for further methanol adsorption- thereby inducing large overpotentials Direct Hydrogen Pt-CO formation is due to the adsorption of CO from the feed stream Pt-CO inhibits further hydrogen adsorption Large currents requires CO electrooxidation to free catalyst sites for further hydrogen adsorption thereby inducing large overpotentials End Result identical POOR ANODE KINETICS 54

55 Improving Methanol Oxidation Kinetics Similar approaches to those taken for H 2 / CO operation: - Better electrocatalysts for enhanced efficacy of CO electrooxidation surface hydroxyls generated at lower potential (see figure) - High temperature operation (> 100 o C) for improved anode kinetics Note, this approach also pays significant dividends by reducing methanol crossover (cathode kinetics???) 55

56 Enhanced Activity of Alloy Catalysts Clearly, for a given Specific Activity (current density at a high voltage / unit active catalyst area),, Alloy catalysts have lower overpotentials for methanol oxidation Hoogers 56

57 Anode kinetics alone does not account for the significant performance loss seen in DMFC systems The effect of methanol crossover is equally The effect of is equally important and affects parameters such as fuel efficiency, cathode kinetics, and mass transport in the cathode layer. Methanol crossover, and techniques to limit it are discussed in the following slides Another important aspect of DMFCs (not discussed in this lecture) is the efficiency of CO 2 removal at the anode. 57

58 Methanol Crossover Ionomeric membranes have a complex water distribution during operation: Anode Cathode Electro osmotic Drag H 2 (or) Methanol H + (H 2 O) n Water Back Diffusion O 2 /Air H 2 2 H e - (or) CH 3 OH + H 2 O CO 2 +6H + +6e - PEM O H + +4 e - 2H 2 O (water generation) Si 58

59 Why (& How) does Methanol Crossover? As water moves from anode to cathode (osmotic drag), methanol migrates along with the water On the cathode side, it is adsorbed onto the cathode electrocatalyst, thereby reducing efficiency The flux of methanol across the membrane depends upon: - methanol concentration in fuel stream - current density - membrane selectivity (ratio of protonic conductivity to methanol permeability; property of membrane) - membrane thickness 59

60 Detrimental Effects of Methanol Crossover Reduced fuel efficiency Cathode mixed potential (due to competing methanol oxidation and oxygen reduction) lowering of open circuit voltage Cathode poisoning CO adsorption on cathode catalyst, lowered cathode activity (see figure) Mass-Transport limitations at cathode (especially for air based applications) methanol wets the hydrophobic gas channels, and permits flooding thereby increasing diffusional resistance 60

61 Effect of Methanol on Oxygen Reduction (Cathode) Kinetics Hoogers 61

62 Effect of Methanol Concentration and Membrane Thickness on Crossover Methanol crossover(ma/cm 2 ) mil 1.7mil 0.5M 1M Evidently, crossover increases as methanol concentration increases and as membrane thickness decreases 2M Si 62

63 Why the Concentration and Thickness Effect? Methanol flux ~ concentration gradient High concentration, low thickness maximum concentration gradient (dc( dc/dx) C left C right x 63

64 Note Cannot use very low concentrations of methanol this is because the energy density of the fuel goes down with dilution, and very low methanol concentrations will increase system weight and volume Can however use neat methanol as fuel, and dilute using water tapped from the cell (cathode) prior to injection into anode approach adopted by LANL 64

65 Effect of Current Density and Concentration on Methanol Crossover Gottesfeld 65

66 Current plays two competing roles: - increased faradaic electrooxidation of methanol at anode (lowers concentration at anode, reduces flux) positive effect - increased protonic current increased water transport to cathode by electro osmotic drag larger methanol crossover negative effect Clearly, the former effect dominates at low methanol concentrations, and the latter at high methanol concentrations 66

67 Effect of Temperature on Methanol Crossover Methanol Crossover (A/cm 2 ) Methanol crossover Resistance Liquid-fed MeOH Temperature ( o C) Vapor-fed MeOH Resistance (Ohm-cm 2 ) Si 67

68 Methanol crossover increases with temperature up to a temperature of ~ 100 C This is because the diffusion coefficient increases with temperature Above ~ 100 C, there is a precipitous drop in crossover This drop occurs due to reduced water transport through the membrane as liquid water does not exist at ambient pressure recall water uptake chart Note membrane resistance also increases with increasing temperature (decreasing relative humidity) therefore, a tradeoff exists! 68

69 High Temperature Membranes Recall discussion on elevated temperature membranes These membranes can be used for high temperature DMFCs The liquid water (and hence methanol) transport rates through the membrane remain minimal However, the resistance at elevated temperatures is greatly reduced Comparative performance data reveals that this technique permits superior performance. Reasons include: - lower crossover (above discussion) - lower resistance (above discussion) - improved anode kinetics (CO desorption favoured) 69

70 Low Temperature DMFC Approaches Certain applications (cellular phones) require operation at room temperature (or low temperatures) The high temperature approach is clearly invalid under these conditions Lowering methanol concentration is achieved by tailoring the membrane microstructure Proven technique using a polymer that does not permit methanol transport (e.g. PVDF) 70

71 Unfortunately Such polymers do not conduct protons well either! Therefore, it is important to look at the membrane selectivity A tradeoff clearly exists between eliminating crossover and retaining protonic conductivity Illustrated in the following diagrams 71

72 Methanol permeability decreases with PVDF content Methanol Permeability (cm 2 /s) 1e-5 1e-6 1e-7 1e-8 1e-9 Nafion Membrane (50 micron) Nafion-PVDF (10 micron) 1e PVDF Content in Nafion-PVDF (% wt) Proton Conductivity (S/cm) o C PVDF Content in Nafion-PVDF (% wt) So does protonic conductivity Optimal PVDF Content will determined experimentally, depends upon thickness Si 72

73 Alternate Design Use a thin layer of methanol barriers such as Nafion PVDF blends sandwiched between proton conducting Nafion layers Recall resistance ~ thickness/conductivity Therefore, very small thickness of barrier acceptable increase in resistance The thin layer is reasonably effective in keeping methanol out Thickness of barrier layer determined experimentally 73

74 Methanol Concentration Profile Through Layered membrane C left Nafion layer C right (<< C left ) Barrier Layer Nafion / PVDF 74

75 Proton Conductivity (S/cm) Nafion Tri-layer Membrane (50 micon) Nafion-PVDF Membrane (10 micron, measured) Nafion-PVDF Membrane (10 micron, calculated ) Methanol Permeability (cm 2 /s) 1e-5 1e-6 1e-7 1e-8 1e-9 Nafion-PVDF (10 micron) Tri-layer Membrane (50 micron) Nafion Membrane (50 micron) PVDF Content (% wt) Proton Conductivity of 50 µm Tri-layer Membrane Containing a 10 µm Nafion - PVDF Barrier Layer as a Function of PVDF in the Barrier Layer. 1e PVDF Content in Nafion-PVDF (% wt) Methanol Permeability Methanol Permeability of 50 µm Tri-layer Membrane Containing 10 µm Nafion - PVDF Barrier Layer as a Function of PVDF Content in Nafion -PVDF. Si 75

76 Effect of Barrier Layer Thickness on Methanol Crossover 2 Methanol Crossover (ma/cm ) 2 ) Methanol Crossover (ma/cm micron Nafion-PVDF Layer PVDF Content (% wt) PVDF Content (% wt) 0.5 M Methanol 1 M Methanol 2 M Methanol Nafion117-1M MeOH 6 micron Nafion-PVDF Layer Nafion117-1M MeOH Note 1. The comparison is made against a Nafion membrane 175 µm thick, and for a 1M methanol concentration Note 2. Note 2. for a thinner barrier layer, a greater fraction of methanol barrier component (namely PVDF) is required to attain a comparable crossover Si 76

77 Effect of Barrier Layer Thickness on Proton Conductivity Proton Conductivity (S/cm) Tri-layer with 6 micron Nafion-PVDF layer Tri-layer with 10 micron Nafion-PVDF Layer Nafion-PVDF Note thinner barrier layer larger conductivity for a given PVDF content PVDF Content (% wt) Si 77

78 Performance with Layered Membranes Cell Voltage(V) 0.8 Nafion112 (50micron) Tri-layer (50 micron) Nafion117 (175 micron) Current density(ma/cm 2 ) Performance and Power Density with 50 µm Nafion 112 and 50 µm Tri-Layer Membrane and 175 µm Nafion 117 at 60 o C with 1M Methanol/Air under 1atm Pressure. Power density (mw/cm 2 ) Current density (ma/cm 2 ) The efficacy of the layered membrane approach is evident Si 78

79 Other Techniques to Reduce Crossover Air/oxygen bleed to oxidize the CO formed at the anode High oxidant flow rates: - enhances oxidant crossover oxidizes CO - may reduce methanol crossover by inducing a suitable pressure profile Pressurized cathodes pressure based disincentive for methanol crossover 79

80 DMFC Summary (Anode) Hoogers 80

81 DMFC Summary (Cathode) Hoogers 81

82 Recap of Challenges - Electrodes Operation of a fuel cell with Reformate (w / CO and CO 2 ) and on direct methanol have similar challenges on the electrode front: - better catalysts for efficient CO electro oxidation at low noble metal loadings - in-situ techniques for enhanced CO electrooxidation and reduced CO adsorption - better materials to facilitate high temperature operation (to permit lower CO adsorption) 82

83 Recap of Challenges - Membranes From a water transport point of view, direct reformate fuel cell (DRFC) and DMFC operation require diametrically opposite material properties: - DRFCs require membranes with enhanced water transport (at higher temperatures) - DMFCs require membranes with next to zero water transport (to reduce methanol crossover) However, from a conductivity point of view, the requirements are the same high protonic conductivity 83

84 Suggested Approaches to Membrane Design Recall different conductivity mechanisms (slide 31) From the requirements point of view (previous 2 slides), the best possible approach would be to develop membranes (electrolytes) that function exclusively via the Grotthuss Mechanism!! Such membranes would have high conductivities and Good Luck!!!!! low methanol permeabilities exceptional selectivities! 84

85 References Ph.D. Thesis, Yongchao Si, University of Connecticut (2002) Unpublished Results Vijay Ramani and Ruichun Jiang, University of Connecticut Polymer Electrolyte Fuel Cells by S. Gottesfeld and T. A. Zawodzinski in Advances in Electrochemical Science and Engineering (Vol. 5), R. Alkire et. al. ed. Fuel Cell Technology Handbook, G. Hoogers ed. (2003) K-D. Kreuer, Chem. Mater., (1996) Proton Conductors, Columban (ed) (1992) Hsu and Gierke, Macromolecules (1982) Malhotra et. al, J. Electrochem. Soc. (1997) Tazi et. al., Electrochim. Acta (2000) Yang et. al, J. Power Sources (2001) Fenton Group: Work in progress 85