Fundamental Electrochemical Materials Aspects of Solid State Fuel Cells

Fundamental Electrochemical Materials Aspects of Solid State Fuel Cells

Info iconThis preview shows pages 1–13. Sign up to view the full content.

View Full Document Right Arrow Icon
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 2
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 4
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 6
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 8
Background image of page 9

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 10
Background image of page 11

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 12
Background image of page 13
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: .43 ‘ r $2.2. 422 Ionics 4 (1998) Fundamental Electrochemical Materials Aspects of Solid State Fuel Cells W. Weppner Chair for Sensors and Solid. State Ionics, Faculty of Engineering Christian—Albrechts University, KaiserstraBe 2, D—24143 Kiel, Germany Abstract. The development of fuel cells and water electrolysis cells requires careful materials engineering in order to overcome the present limitations of degradation and high cost. For this reason, the fundamental parameters of operation of the galvanic cells are discussed. It has to be taken into account that the voltage drops over a very narrow region in the nanometer range which requires a high degree of materials stability. Theelectn'cal current depends en the redox processes at the interfaces and the electronic properties of the employed materials. The possibilities to control these parameters by dopants are discussed. Some general rules are concluded. Furthermore, long- term effects of cell polarization on the charge transfer resistance are reported. The described aspects of fuel cells may also provide a general guidance for the selection of appropriate materials for other solid state ionic devices. 1 . Introduction The principles of fuel cells and water electrolysis cells appear to be quite simple considering Nernst's equation relating the cell voltage E to the activities a of the mobile electroactive component M at both sides ‘ and " of the electrolyte: ln—.—.—- (1) where k, T, 21'“ and q are Boltzmann’s constant, the absolute temperature, charge number of the mobile com- ponent and elementary charge, respectively. When a current is passed, ohmic and various other polarization effects related to diffusion, adsorption, dissociation and re- dox processes will occur, which reduce the open circuit cell voltage E in the case of fuel cells or require higher voltages in the case of electrolysis cells U=E-Ep01=E-ip£-Ep01 (2) where i, p and E are the current density, specific resistance and length of the sample, respectively. Although it is generally known that fuel cells provide much higher energy conversion efficiencies compared to any other technology and provide the best environmental performance, nevertheless fuel cells did not yet reach a major breakthrough in replacing conventional techniques. There have been so far only some niches of applications, e.g. in submarines or in space crafts when costs do not play an important role. In view of the present difficulties in the development of solid state fuel cells and electrolysis cells, it appears essential to consider the underlying principles of ope- ration of such a device and the materials parameters which control the performance. Depending on the functional requirements of the materials being employed for the gal— vanic cell, appropriate and systematic materials develop- ment is necessary. Furthermore, we need to consider ma- terials combinations and interfacial processes for the appropriate operation of the galvanic cells. Long term chemical stabilities of the materials and contacts are re- quired together with low polarisation losses. Some of the most important fundamental aspects of fuel cell and electrolysis cell'operation will be considered. The generation of cell voltages and currents is different Ionics 4 (1998) from the case of conventional liquid electrolyte devices and the formation of internal electrical fields in semiconductor junctions when solid electrolytes and elec- trodes are being used. Another difference to conventional liquid electrolytes is the predominance of one type of ions in solid electrolytes with electrons as minority charge carriers which play a fundamental role for the redox pro; cesses. At the operating temperature of solid oxide fuel cells, also electrodes with extraordinarily enhanced chemi- cal diffusion rates are being considered. In view of developing fuel and electrolysis cells in which the electrolyteand electrodes are combined into a single material (“monolithic fuel cell“), the electronic properties have to be controlled and established appro— priately under the reducing and oxidizing conditions to be n—and p—conducting, respectively. The role of dopants with regard to the electronic minority charge carriers needs to‘ be better understood in view of major differences between ionic compounds and common semiCOnductors. In addi- tion to dopants, also electrical currents being passed across the galvanic cell may result in chemical changes of the surface composition of the electrolyte which has an effect on the interfacial polarisations. Long-term effects have been observed which show up only after several weeks and may have an important influence on the per— formance. 2. Generation of Cell Voltages and Electrical Currents The e1ectrodes and the electrolyte of a galvanic cell are good electronic and ionic conductors, respectively. The concentrations of the mobile charge carriers, i.e. electrons and ionic defects, are generally quite high. Accordingly, there may be no major gradient in the chemical potential of the mobile species according to: Lu = it? + leI‘l(’YiCi) (3) where m, ci and it? are the activity coefficient, con- centration, and chemical potential in the standard state (i.e. yici = l) of the mobile species i = e-, h+, M“, respectively. As a result, there will be no (or only a negligible) electrical field within the electrodes and elec- trolyte. Any potential electrical field will be extinguished by a small motion of the large number of mobile charge carriers. Under open circuit conditions, the electrostatic potential gradient compensates precisely the (negligibly small) concentration gradient as driving force for the 423 (b) —-— (111111 +- Fig. 1; Local variation of the chemical potentials p of neutral oxygen 01, oxygen ions 0“, electrons e" and holes ' h+ and the electrostatic potential q); 20 across the entire gal- vanic cell and b) across an electrodelelectrolyte interface. charge carriers. In the case of an electrical current in the electrical circuit, a small electrostatic potential drop re— mains which is responsible for the continuous motion of the ions in the electrolyte and electrons or holes in the electrodes. The chemical (pi) and electrical (ziqrp) energy are commonly linearly superimposed and com— bined to the electrochemical potential: Tii = i'l'i + Ziq LP (4) In equilibrium, any gradient of this electrochemical po- tential will disappear. 424 The'important electrostatic potential drops occur prac- tically exclusively at the interfaces between the electrodes and the electrolyte. Both electrons (or holes) and ions are being exchanged there. The electrical field is formed to compensate the driving force by the concentration gra— dients of all mobile species across the interface. In view of large differences in the concentrations of the mobile charge carriers in the two phases, large potential drops occur within the space charge region. The local variation of the electrostatic and chemical potentials of the various species across the galvanic cell is presented in. Fig. 1a. ‘ A magnification of the potential drops at one of the interfaces is shown in Fig. 1b. In view of thermodynamic equilibrium, the chemical potential of neutral oxygen (or hydrogen) is the same at both sides of the interface. Also, the electrochemical potentials of electronic and ionic charge carriers go through. Within the bulks of the elec- trolyte and electrodes, the chemical and electrochemical potentials of the various species are related to each other by ionization equilibria: ' 1 '2‘ H02 = llO" + 2Tlh+ = 710" ' 2T1e- = llo-- + 2Hh+ = NO" ‘ lee- 1 (5) ‘5 MHz = TlH+ ' Tlh+ = TlH+ + Tle- = HH+ ‘ llh+ = lLH+ + lie- The eleCtrostatic and chemical potential profiles in the intermediate region are controlled by the mutual compen- sation of the diffusional and electrical fluxes under open circuit conditions; 0 = JD¥JE = D§_E_d_cp (6) 3x zq dx where D, o and (p are the diffusion coefficient, electIiCaI conductivity and electrostatic potential, respectively. The electrical field at the interfaces has to “compensate simultaneously the concentration differences of all mobile species, i.e. ions and electrons. The concentrations of the ionic and electronic charge carriers will change accordingly by an exchange between both materials as much as necessary to fulfil this requirement. There may be major numbers of charge carriers necessary in order to establish the equilibrium. The second fastest species are responsible Ionics 4' (1998) for this process [1]. These are often the electrons or holes in solid electrolytes which may show extremely low diffusivities [2]. The transport of oxygen ions or protons is rate determining for the electrodes. Diffusion through pores of an electrode may provide pathes to short—circuit this process. The region over which the electrostatic potential drop occurs at the interface between a solid electrolyte and elec- trode is rather small. If one assumes the validity of De— bye's formula aao kT qszz+ LI): (.7) (e and so are the dielectric constant of the material and the permittivity in vacuum, respectively) as a qualitative first approach, the thickness LD of the space charge layer is of the order of 0.1 nm for common solid electrolyte and electrode materials. This very small value of only a few atomic‘layers is caused by the large number of charge carriers in a fast conducting solid electrolyte and metallic conducting electrode. Any electrical field will be readily shielded by the large concentration of existing charge carriers. This situation is quite different from the case of common liquid electrolytes where the conducting ions are dissolved in a dielectric material, e.g. water or or— ganic solvent. The concentration of mobile ions is commonly very small which allows to build up a large Space charge region. I _ The narrow region of electrostatic potential drop at the interface is also unique in comparison with semiconduc— ting materials. In spite of the same order of conductivity as in solid electrolytes [3], semiconductors are characte- rized by small numbers of (highly mobile) electrons or holes. The thickness of the space charge layer is accor- dingly in this case several orders of magnitude larger than in the case of solid electrolytes and is commonly in the um range. Otherwise, the situation of forming electro- static potential drops at semiconductor interface junctions is similar to electrode/electrolyte junctions. However, electronic junctions do not allow to build up voltages which are measurable at the outside and do not produce a current in an outer electrical circuit. Electronic junctions are “short—circuiting themselves“, while solid electrolytes are blocking the electronic charge carriers which therefore have to move through the outer electrical circuit in order to compensate the local charge displacement of the ionic charge carriers in the electrolyte. Ionics 4 (1998 Electrolyte — — — — — - Electrode Fig. 2. Cumulative Brouwer diagram, i.e. relationships between the chemical potentials (log concentrations) of oxygen ions, electrons, holes-and neutral oxygen, for the electrodes (broken lines) and the electrolyte (solid lines). ‘ Depending on the type of disorder, different concentration changes with the oxygen partial pressure are observed. Relationships for proton conductors are analogous. The electrostatic potential drop at the interface depends not only on the type of materials combination but also on the chemical potential of oxygen or hydrogen which results in a variation of the electrostatic potential drops at the two electrolytefelectrode interfaces and change in Nernst's voltage (eq. 1). Different local distributions of the chemical potentials are commonly observed for the electrolyte and the elec- trodes. While the chemical potential of oxygen ions or protons is approximately constant in the electrolyte, these quantities are changing with the oxygen and hydrogen ac— tivity in less disordered electrodes. As a consequence, the chemical potentials of electrons and holes change diffe- rently with the oxygen activity, typically with a slope of 1/4 and 1/6 in electrolytes and electrodes, respectively [4]. These relationships are shown schematically in Pig. 2 in a cumulative Brouwer diagram. This diagram illustrates also that the same material may have quite different majority and minority charge carrier properties depending on the oxygen activity of the operation coaditions. The local distribution of the electrostatic and chemical potentials changes when a current is passed across the gal- vanic cell (Fig. 3). The chemical potentials of the majo— rity charge carriers remain nearly unchanged in all phases while those of the minority charge carriers, i.e. electrons in the solid electrolyte and oxygen ions in the electrodes, are changed. As a consequence, a small electrostatic. po- tential drop occurs in the electrolyte to drive the ions. The 425 with current _ .. .. — — — without current Fig. 3. Variation of the chemical potentials of oxygen ions and electrons and the electrostatic potential across the gal- vanic cell when a current is passed as compared to open cir— cuit conditions. electrostatic potential drop at the cathode side is decreasod while that at the anode side is increased. The amount depends in addition on the polarisation processes 'at the interfaces. For improving the performance of the inter- faces, both ions and electrons have to be taken into consideration. 3. Electronic Minority Charge Carriers The interfaces are sinks and sources of electronic and ionic charge carriers in the case of an electrical current by closing the electrical circuit. This is tied to the consump- tion and generation of neutral oxygen or hydrogen de- pending on the type of ionic conductor. Both electrons'and ions have to be transferred at the interfaces. They originate from neutral gaseous species and recombine to those again at a lower chemical energy. Any transport in gradients of the chemical composi— tion_occurs by a chemical diffusion process. This includes both the contribution by concentration gradients and the effect of internal electrical fields generated by the coupling of the motion of the two fastest species [5]. The faster. diffusing species accelerate the slower ones in order to maintain local electrical neutrality. The effect of the internal electrical field is described by the Wagner factor W '[1] which is a generalization of 'the thermodynamic factor known from metallic systems. By making use of 426 Ionics 4' (1998) this factor, the flux density of ions i, i.e. oxygen ions or protons, is described by i ' dlna. dc W=t 3 'i =—WD1 —- J a 8qux (8) where Di is the diffusivity in the absence of a compo- sitional gradient which is related to the general mobility hi and electrical mobility u; by Nernst—Einstein's relation— Ship kT Izi Iq Di = bikT = u; (9) t6 is the transference number of the electrons. The upper index x in eq. (8) indicates neutral components. While the thermodynamic factor d ln aix/d 1n cix is commonly a slightly larger number than 1 in metallic systems, the Wagner factor may be much larger or smaller than 1 in the case of semiconducting electrodes and solid electrolytes, respectively. In fact, fuel cells and other ionic devices are based on the availibility of materials with appropriate Wagner factors. Under the simplifying assumption of concentration in- dependent acitivity coefficients, the Wagner factors were calculated from eq. (8) as a function of the logarithms of a: A 5 ""h- I] :1 U! 62 Fig. 4. Wagner factor (indicated as parameter for each curve) as a function of the logarithms of the ratios of the mobilities u and concentrations 0 of the electrons e and ions i. The broken line separates electrodes with predo- minant electronic conduction (upward to the right) from electrolytes with predominant ionic conduction (downward to the left). the ratios of mobilities ue/ui and concentrations ce/ci of the electrons e and ions i. The result is plotted in Fig. 4 with the Wagner factor indicated as parameter of each curve. For fast chemical diffusion it is required that the electrons have a much higher mobility than the ions but are present in much smaller numbers. In some cases, Wagner factors as large as about 105 were observed experi- mentally [6]. In such a case, the transport originates from a concentration gradient, but by far the major driving force for the ions is the local electrical field which results from fast electronic charge carriers pulling the ions behind. These considerations are also applicable for two-phase electrodes in which the electrons move in a metallic conducting phase and the ions along the surface of the second phase. Such a system is similar to a metallic con- ducting material with a thermodynamic factor close to 1. However, the diffusion path is restricted to the surface- near region. Solid electrolytes require smaller electronic than ionic conductivity, i.e. larger mobility-concentration products for the ions than for the electronic charge carriers. In the case of high electronic mobility, the concentrations of electrons and holes have to be sufficiently small for any activity of the electroactive component. The limit that separates electrodes from electrolytes is shown in Fig. 4 by the broken line. In order to fulfil the requirement, solid electrolytes should have commonly rather low electronic mobilities in view of the presence of intrinsic or extrinsic electrons or holes. Large Wagner factors may be present if the electronic concentration is sufficiently small (see Fig. 4). The chemical diffusion is much enhanced, but the transport rate for the ions is nevertheless small because of the rather small concentration gradient which corresponds to that of the electronic charge carriers for the reason of charge neutrality. Since the concentrations of electrons or holes are small, these species do not allow to build up a large concentration gradient. The enhanced diffusion of the ions may be very advantageous on the other hand for the charge u‘ansfer electrode processes where electrons are available from the electronic leads. So far fast ion conductors have been examined for their electronic properties, low mobilities of the electronic charge carriers were observed. Most of this work has been performed on yttria—doped zirconia. The results of diffusion coefficient measurements of electrons and holes are plotted in Fig. 5. In addition, the diffusivity of oxygen vacancies is plotted as calculated from the ionic con- ductivity and the number of oxygen vacancies. ‘il Ionics 4 (1998 T ['C] -—-—-—- 900 850 800 750 700 electrons ——-— to‘ 1/7 [rt-1] Fig. 5. Diffusion coefficients of electrons and holes of cubic ZrO2 (+10 m/o Y203) in the temperature range from 700 to 900 “C. The activation energies are = 0.55 and.= 1.4 eV for electrons and holes, respectively. In comparison, the diffusion coefficient of the oxygen vacancies is 5-10J cmzls at 700 “C with an activation energy of = 1 eV. Figure 6 is a plot of the oxygen ion, electron and hole concentrations of zirconia as a function of the oxygen partial pressure. While the concentration of oxygen va- cancies remains constant, the concentrations of holes and electrons change according to a 1/4 power law. The pre- sence of electronic charge carriers 'at the electrolyte/elec- trode junction is therefore changing largely with the oxy— gen or hydrogen activity. Redox processes occur at the interfaces between the electrodes and the electrolyte in which both electronic and ionic charge carriers are involved when a current is passed through the external circuit. These are the result of fluxes of electrons and ions across the interfaces which require gradients in their electrochemical potentials. The decrease in the electrical field will provide diffusional fluxes of the ions and electrons. This will have an influence on the equilibria and concentrations of charge carriers. Both ions 427 I I . 20 ‘5 -IO -l§ 120 -25 '30 '35 ‘—- loq p02 [aim] Fig. 6. Variation of the concentrations of electrons and holes of cubic ZrO2 (+10 rn/o Y203) as a function of the oxy— gen partial pressure in the temperature range from 700 to 900 °C. The relationship follows a 1/4 power law. For com- parison, the concentration of oxygen vacancies is 2.7-1021 cm'3 independent of the oxygen partial pressure and tempe- rature. and electrons will control the performance of the inter— facial processes. Accordingly, this is an important issue which resulted in the attempt to control the concentration of electronic species in the electrolyte in view of mini- mizing the degree of polarisation of the fuel or electro— lysis cell. Electronic concentrations are more readily changed than mobilities by solid solution of dopants. 4. The Electronic Effect of Dopants It is known from semiconductors that the electronic con— centration may be readily changed by aliovalent dopants in approximately the same amount as the dopant concentra- tion. In the case of compound semiconductors, such as solid electrolytes, the situation is different. Dopants are not single elements but compounds. Only the difference in cations and anions, i.e. the nonstoichiometry, counts for the change in n- and p-conduction. This difference is usually very small and the amount of a dopant which may be dissolved in a solid electrolyte is usually small. It is therefore understandable that we have to deal with a minor effect. Nevertheless, such doping should allow to show the general influence of the electronic minority charge carriers on the electrode kinetics. 428 Gas ' Air (H2, CH4, I ' (02) (9 1m Fig. 7. Principle of a monolithic fuel and electrolysis cell. The separator material is predominantly an excess electron conductor at the fuel/hydrogen side and predominantly a hole conductor at the air/water side. Inbetween, the material is predominantly an ionic conductor which forces the electrons to be passed through the outside electrical circuit and perform the electrical work. Compared to a pure ionic O N conductor the voltage is (slightly) reduced because of the 7 mixed conducting regimes. Another important issue of the electronic conductivity of solid electrolytes is the potential development of monolithic fuel and electrolysis cells. This would drasti— cally reduce the cost and will improve the stability since the material will be in thermodynamic equilibrium with the oxygen or hydrogen at the activities of the gas en- vironments. Such a system is shown in principle in Fig. 7. The cell voltage will be somewhat reduced by the mixed conduction at both electrode sides. However, this energy loss may be recovered as heat, tag. for combined power and heat generators. Also, this influence may be I /I '12P(ss) + ZrTiO4 i i I Prepared by coprecipitation method . Prepared by Pechini method 0 s to is 20 as I mol% TiOz Fig. 8. Variation of. the lattice parameters a and c of tetra- gonal zirconia with 3 m/o Y203 as a function of the amount of TiO2 doparit concentration. ‘- 4 bulk conductivity [S/cm] Ionics 4 (1998) small compared to other losses in three-materials cells. Since the chemical potential of oxygen or hydrogen drops continuously across the single-material cell, there will be always a region of pure ionic conduction which allows blocking of the electronic current and building up the cell voltage according to the generalized Nernst‘s equation [7]: (10) where " and ‘ stand for the right and left hand side elec- trode, respectively. Most experimental work was per- formed on tetragonal zirconia with the composition Zr02 (+2—3 m/o Y203) as host material. This material was doped with a large number of n— and p-type conducting binary oxides, such as TiOz, MnOx, Tb407 and Fe203. The solubility ranges may be observed in many cases by the change in the lattice parameters as a function of the amount of depant. Figure 8 shows, e.g., the variation of the a and c lattice parameter as a function of the TiO2 dopant concentration. The solubility of 20 11110 is one of the largest observed for tetragonal zirconia. In contrast, the solubility of Fe203 is rather small. The addition of Fe203 does not change the lattice para- meters significantly. Therefore, XRD is not suitable to determine the solubility limit of FeQO3 in tetragonal zirconia polycrystals. However, conductivity measure- ments allowed to estimate the solubility limit to be 0.8 Info F8203. Figure 9 shows the bulk conductivity as a function of the Fe203 content at 650 "C. There is a rather steep decrease up to about 0.3 m/o, followed by a small decrease at higher FezO3 contents [8]. 10:10" awn“ solubility limit of F6203 in TZP (SS) 2.0m 0‘ mol% F6203 Fig. 9. Bulk (intragranular) conductivity of FelO3 doped tetragonal zirconia polycrystals as a function of the ‘r'ieZO3 concentration. Ionics 4 1998) 4.5 undoped TZP 650°C 3 5 ZZY1T655°C ‘ Z2Y5T550"C 22Y1GT570°C H 2.5 E :r E 1.5 D) 2. 0.5 ~05 -15 tIL_r_r_I|i|:|L_L_i_._IIL_l_|_LLI ' 0 0.5 i 1.5 2 . 2.5 E [V] Fig. 10. Hebb-Wagner polarization curves of tetragonal ZrO2 (—r- 3 mlo Y203) doped with O, l. 5 and 10 Info TiO2 at about 670 “C. The reference electrode is air. The diffusional I currents of holes and electrons are shown as a function of the oxygen partial pressure at the inert electrode side which is imposed by the applied voltage E according to Nernst‘s law. The measurement of the conductivity of electrons and holes requires careful experimentation because the conduc- tivity of these minority charge carriers is commonly several orders of magnitude lower than the predominant oxygen ion conductivity. The ionic current has to be carefully blocked by an inert electronically conducting electrode in order to pass only the electrons and holes (Hebb-Wagner technique [9, 10]). A reversible reference electrode has to be used on the opposite side.‘The current is determined by Fick's law rather than Ohm's law. Con- centration gradients instead of electrical fields are the driving force for the electronic species since the ionic ma- jority charge carriers shield any potential electrical field. Depending on the experimental condition whether the con— centrations of electrons and holes are decreased or increaSed by the applied voltage at the blocking inert electrode, a current plateau or exponential increase of the current is observed, respectively. Results are shown in Fig. 10 for several amounts of TiOz dopant in Zr02 (+2 info Y203) at a temperature of z 670 °C. The hole and elec- tron conductivity at the reversible reference electrode side are determined by comparison with the Heb'trWagner equation ‘ kT E ‘E baggage ui)+of1(1—e “)1 - (11) 429 T : 650°C TZP pure log (c[S/cm1) -5. D 5 10 15 20 25 30 35 -log poalPa] Fig. 11. Variations of the partial electron, hole and oxygen ion conductivities of tetragonal ZrO2 with 3 mlo Y203 by the amount of Tit)2 dopant at 650 “C. Zirconia becomes predominantly electronically conducting by the addition of 5 mlo Tio2 at the fuel side. ' G erh, i and L’ are the electron, hole conductivity at the reversible side of the sample, the current density and thickness of the sample, respectively. More precisely, a modification of this relationship has to be employed which takes valence changes of the dopants into account as derived in [8, 11]. For the oxygen partial pressure dependence of the con- 114 02 [4] under the assumption of a large constant concentration ductivity it is reasonable to assume a p relationship of oxygen vacancies. Figure 11 shows this dependence for various amounts of TiOz dopant at 650 °C. While the electronic concentration varies drastically with the amount of Ti02, the hole conductivity remains practically un- changed. The electronic concentration may be accordingly easily modified by the amount of Ti02 at the anode side of a fuel cell. The material may be even made predomi— nantly electronically conducting at this side. The amount of dopant does neither change the activation energies of electrons and holes (Fig. 12) nor that of the ions (Fig. 13).-It is therefore concluded that the dopant has an effect on the entropy term of the motion, i.e. on the attempt frequency and not on the barrier height of each jump. The studies on the doping of zirconia solid electrolytes have resulted in several rules with regard to the effect of the dopant on the transport and phase stability properties of the electrolyteTS]: log (a [Slcm]) _ 20 . 0.95 1 1.05 1.1 1.15 1.2 1.25 1.3 iooorr [K4] Fig. 12. Temperature dependence of the electron and hole conductivity of tetragonal ZrO2 (+ 3 mfo Y203) for different amounts of TiOz dopant. The activation energy remains nearly unchanged. (i) p— or n—type conducting oxides as dopants of zirconia increase the partial hole and electron conductivity, res- pectively. This is independent of the different crystal Structure and chemical environment. The binary oxides T102 and Fe203 are n-type conductors and increase the partial excess electron conductivity. (ii) The ionic conductivity decreases by the addition of further dopants because of local lattice distortions. The magnitude of solid solubility is not directly related to the lattice distortion, however. Total conductivity [S/crn] s r E 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.0 1.7 1.0 x lOOO/TUIK] Fig. 13. Temperature dependence of the total (ionic) con- ductivity of tetragonal ZIO2 (+ 3 Info YZOB) for different amounts of TiO2 dOpant. . Ionics 4' (1998) u 23Y16T 1000 o Z3Y8T I ZBY 1° Charge-transfer resistance [9 cm2] m 0.10 0.05 1.00 1.05 1.10 1.15'120 us IUOOITDIK] Fig. 14. Charge transfer resistance of tetragonal ZrO2 (+ 3 info Y203) as a function of temperature for different amounts of TiO2 dopant for the air electrode, as determined from impedance analysis. (iii) The diffusion coefficients of the electronic species and mobile ions are correlated. Both decrease with in— creasing amounts of dopant. (iv) The electrochemical redox potentials are not ob- viously changed by the dopant as compared to the undoped electrolyte. (v) The dopant controls the phase formation of zirconia (similar to Y203 or CaO) depending on the valence of the dopant. Valencies lower than four increase the oxygen va— cancy concentration and may convert the monoclinic phase into the tetragonal or even cubic phase, while va- lencies higher than four reduce the oxygen vacancy con; centration which favors the formation of the tetragonal or even monoclinic phase. Less expensive dopants may be used for stabilizing the better ionically conducting phases. These rules may be seen as a guideline for the selec- tion of the types and concentrations of dopants in the case of zirconia, but may not be necessarily valid in general for other compounds. The charge transfer resistance of TiO2 doped tetragonal zirconia was determined by impedance analysis from the low frequency are for various doped samples in air. The results are plotted in Fig. 14 as a function of the inverse temperature and show that doping by TiO2 in fact in- creases the charge transfer resistance which may be related to the lower ionic conductivity of the doped samples. However, the activation emergy is slightly reduced by the dopant. Further studies will be necessary to get a better understanding of the rate determining processes. Ionics 4 1998 431 I ZBY.5F.2T l [Man] 0 23Y-.5F2T ......... 231m: 7 ---- --23Y theorellcal UM Fig. 15. Hebb-Wagner polarization curves of pure, iron oxide and iron oxide/titania co-doped tetragonal ZrO2 (+3 m/o YZOH). Valence changes of the dopant cations may Show a si- milar influence as a variation of the hole concentration at the reversible electrode for various co—doped and “pure“ tetragonal zirconia samples. The Hebb~Wagner polariza- tion curves are shown in Fig. 15. Distinct plateau regions are no longer observed. Furthermore, the current shows a characteristic exponential increase at low voltages below about 0.5 V [8, ll]. 5. Electrode'Process and Long-term Interfacial Electrolyte Polarisation Effects Passing a current across a solid electrolyte does not only result in instantaneous polarisation losses as commonly known from physical chemistry. There are also effects Igtpm). pm [Pa] Fig. 16. Electrode conductance per unit length of the 3- phase electrode 02, Pt I cubic ZrO2 as a function of the oxygen partial pressure. which are only observed after extended periods of time of several weeks of operation at elevated tempera— tures [12-14]. These effects were investigated by using a line electrode as working electrode and a large-area counter electrode together with small reference electrodes. The line electrode was a platinum wire spring loaded against the electrolyte which allows a considerably larger exchange area as compared to point electrodes. The electrode conductance per unit wire length is plot— ted for the temperature range from 773 to 1173 K in Fig. 16 as a function of the oxygen partial pressure in the range from 10‘1 to 106 Pa. The Slope is +0.55 and —0.45 for low and high oxygen partial pressures, respectively. These values indicate a process limited by the surface coverage of oxygen. Assuming a Langmuir isotherme, the coverage of the platinum surface with atomic oxygen is described by Kllz 1/2 P 6: ad 02 (12) 112 112 1+K ad p02 and the diffusional flux of the oxygen along the surface is given by the product of the occupied and empty sides: joc a (1—9) (13) Substitution of eq. (12) into eq. (13) provides the follow— ing relationship of a diffusion controlled flux of oxygen along the platinum surface toward the triple phase boun— dary zirconia/Pt/gas 112 112 K ad p02 [2 112 ad 1’02)2 joc (14) (la-{(1 A maximum is Observed at a coverage of 6 = 1/2. The limiting cases are logj cc + § log p02 (15) for high oxygen partial pressures and 10ng — ; log p02 (16) for low oxygen partial\pressures. These relationships are observed experimentally with reasonable good agreement. 432 .L. o l .— M 1l u .— uh - - -ID|325PI “ a loos“). on [Sem' 3 -l8 u 101331’: - 10:3 P. 101?: ‘20 mm 1?: -22 0.8 0.9 1.0 1.1 1.2 lOOO/I‘ [11K] Fig. 17. Arrhenius plot of the maximum electrode conduc- tance per unit length of the 3-phase electrode 02, Pt I cubic ZrO2 at the corresponding oxygen partial pressure. This indicates clearly that the surface diffusion of disso- ciatively adsorbed oxygen along platinum toward the three-phase boundary is rate determining. The maximum is shifted for the employed platinum electrode to lower oxygen partial pressures with decreasing temperature [12]. Information about the thermodynamics of the dissociative adsorption of oxygen on platinum in the temperature range from 773 to 1173 K may be derived from these measurements. Figure 17 shows an Arrhenius plot of the electrode conductivity of an air/platinum electrode in combination with yttria Stabilized cubic zirconia at the conductivity maximum. The result provides the activation energy of the surface diffusion and the diffusion length. Similar measurements were also performed with gold electrodes. The electrode resistance per unit length is pre- sented in Fig. 18 for the temperature range from 873 to 1173 K. No maximum is observed for the partial pressure range from-104’ to 106 Pa. The slope increases with in— creasing po2 from +0.25 to +0.50. Single crystals and polycrystalline material showed the same results [12}. The considerably higher polarisation resistances in the case of gold electrodes indicate the lower catalytic activity compared to Ft. At high oxygen partial pressures; the slope +l/2 indicates again that dissociatively adsorbed oxygen is the rate limiting step of the electrode reaction. Diffusion of oxygen along the zirconia surface is less likely since the degree of coverage of oxygen was found to be independent of the oxygen partial pressure at least in the case of undoped cubic zirconia. The slope +1/4 at low Ionics 4 (1998) lstpm). Pm [Pa] Fig. 18. Electrode conductance per unit length of the 3- phase electrode 02, An I cubic ZrO2 as a function of the oxygen partial pressure. oxygen partial pressures indicates that the charge transfer of the dissociativer adsorbed oxygen is the rate deter- mining step. Considering the exchange current densities and assuming equal cathodic and anodic transfer coeffi— cients, the following relationship is obtained (see [14]): log i0 0: + i log p02 _(l7) This is in agreement with the experimental findings. The application of high current densities to Pt elec— trodes of solid electrolytes at the operating temperature causes a significant decrease in the polarization resistance [15] while pronounced long-term annealing leads to an increase in the electrode polarization resistance [16]. These results were attributed to changes in the electrode mor- phology. Recent studies of combined polarization and heat treatments conclude the formation of additional metastable reaction sites at the zircoma surface with a leng lifetime at intermediate temperatures [14]. These data show the importance of the catalytic activity of the zirconia surface. The samples were equilibrated in air at 600 “C and then polarized both cathodically and anodically by using Pt wire electrodes and applying voltages of —2.0 and +0.5 V against air, respectively, for 1 h. After this treatment, the voltage dropped back to zero. The impedance was subsequently measured for 300 h. The polarization re- sistances of these samples remained nearly unchanged over the entire period of time. Figure 19 shows the impedance plots of cathodic'ally and anodically pretreated samples .at 300 h after the polarization. The spectra are similar to Ionics 4 (1998) ' 433 15 -Z" [kg] 0.5 0 l 2 3 4 2’ [k9] Fig. 19. Impedance spectra at 600 “C of cathodically (2 V, top) and anodically (+0.5 V, bottom) pretreated Pt elec- trodes applied to cubic zirconia electrolytes at 300 h. after the polarization (for 1 h at 600 “(3) at p02 = 1013 Pa and 101 Pa, respectively. those of samples without pretreatment at low and high oxygen partial pressures. A more distinct pattern is observed at intermediate oxygen partial pressures. I The specific electrode conductivity is plotted in Fig. 20 as a function of the oxygen partial pressure for different temperatures of the polarized sample. Both in the case of cathodic and anodic petreatment, an increase of the electrode conductance per unit length by nearly 3 orders of magnitude is observed compared to the electrode without pretreatment. When the applied voltage is relaxed, even 2000 hours after the polarization, the resistance has increased again by only about 20 %. At 1173 K, however, the conductance is recovered within only about 200 hours. The interpretation of these results is based on the assumption of local va— riations in the composition of the electrolyte within the interfacial region. 6. Conclusions Materials understanding will be the key for developing appropriate fuel cell and electrolysis cell materials which 'e U a be" a: .3” -I 0 1 2 3 4 5 6 13(1),”). no2 [Pa] -3 I—u—I—r—r—r I'l—l—l—X l—l—l—l—l_1 I r—l—!_l'—| 7'" -4 , . E I u a E -5 o . ' o b: o I . F: ‘6 I I o o 3‘ - 873K - .39 -7 ' 323K I II I 773 K _8 I .I—l—I_‘ I IJ_I_:._I I LI_I_I._.I | L_|_I_-l |_I_I__I |__|_l_(_1_.—K -1 0 l 2 3 4 5 6 Is(p0,), 90, [Pa] Fig. 20. Electrode conductance per unit length of the 3- phase electrode 02, Pt I cubic ZrO2 after cathodic pretreat- ment (top) and anodic pretreatment (bottom) at 873 K for 1 . houL may overcome present limitations ‘of insufficient long- term stability and unacceptable high‘cost. Both electrons and ions play important roles in the performance of the materials and junctions. The solution of the present problems is complicated by the very narrow range of space charge at the interfaces and chemical reac— tions within this regime. Long-term application of cell voltages results in modi— fications of the chemical surface composition which in- fluences the exchange Current density and electrostatic po— tential drop. The presented considerations may also provide a guidance for the selection of appropriate materials for many other applications in the field of solid state ionics. 7. References [l] W. Weppner, R.A. Huggins, J. Electrochem. Soc. 124, 1569 (1977). [2] W. Weppner, Z. Naturforsch. 313, 1336 (1976). [3] W. Weppner, Ionics 1, 1 (1995). 434 [4] [5] [6] [7] [8] [9] [10] H. Rickert, Electrochemisz of Solids, Springer— Verlag, Berlin, Heidelberg, New York, 1982. W. Weppner, in: Transport-structure Relations in Fast Ion and Mixed Conductors (EW. Poulsen, N. Hessel Andersen, K. Clausen, S. Skaarup, O. Toft S¢rensen, Eds), Rise Natl. Lab., Roskilde, DK, 1985, p. 139. W. Weppner, R.A. Huggins, J. Solid State Chem. 22, 297 (1977). ' H. Schmalzried, Z. Phys. Chem. NF 38, 87 (1963). U. Abend, X.J. Huang, W. Weppner, Ionics 3, 427 (1997). M. Hebb, J. Chem. Phys. 20, 185 (1952). C. Wagner, Proc. 7th Meeting, Int]. Committee on Electrochemical Thermodynamics and Kinetics, Lin- dau 1955, Butterworths,'London 1956, p. 361. Ionics 4 1998) [11] X.J. Huang, W. Weppner, Ionics 1, 220 (1995). [12] C. Schwandt, W. Weppner, Ionics 2, 113 (1996). [13] C. Schwandt, W. Weppner, J. Electrochem. Soc. 144, 3728 (1997). [14] C. Schwandt, W. Weppner, Solid State Ionics, in print. [15] IE. Bauerle, J. Phys. Chem. Solids 30, 2657 ' (1969. [16] S.P.S. Badwal, F.T. Ciacchi, Solid State Ionics 188519, 1054 (1986). Paper presented at the 97th Xianshan Science Conference on New Solid State Fuel Cells, Xianshan, Beijing, China, June 14-17, 1998. Manuscript rec. Aug. 15, 1998; acc. Oct. 10, 1998. A12... ...
View Full Document

This note was uploaded on 04/24/2008 for the course EMA 6446 taught by Professor Wachsman during the Fall '08 term at University of Florida.

Page1 / 13

Fundamental Electrochemical Materials Aspects of Solid State Fuel Cells

This preview shows document pages 1 - 13. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online