최소 단어 이상 선택하여야 합니다.
최대 10 단어까지만 선택 가능합니다.
다음과 같은 기능을 한번의 로그인으로 사용 할 수 있습니다.
NTIS 바로가기다음과 같은 기능을 한번의 로그인으로 사용 할 수 있습니다.
DataON 바로가기다음과 같은 기능을 한번의 로그인으로 사용 할 수 있습니다.
Edison 바로가기다음과 같은 기능을 한번의 로그인으로 사용 할 수 있습니다.
Kafe 바로가기주관연구기관 | 전남대학교 Chonnam National University |
---|---|
연구책임자 | 이종숙 |
참여연구자 | J.G. Fisher , 신의철 , H.D. Nguyen Tran , Nguyen Dang Thanh , 김영헌 , 김지훈 , 문수현 , 조동춘 , 김계록 , 조승영 , 김영철 , 양진훈 , 김지수 , 안자민 , 전정욱 |
보고서유형 | 2단계보고서 |
발행국가 | 대한민국 |
언어 | 한국어 |
발행년월 | 2015-05 |
과제시작연도 | 2014 |
주관부처 | 미래창조과학부 Ministry of Science, ICT and Future Planning |
등록번호 | TRKO201700011931 |
과제고유번호 | 1711014627 |
사업명 | 첨단융합기술개발 |
DB 구축일자 | 2017-11-13 |
키워드 | 프로톤전도성세라믹.다종수송.교류모델 분석법.입계 및 전극반응.프로톤세라믹연료전지.계산과학.protonic ceramic conductor.multi-species transport.ac model analysis.grain boundary and electrode reaction.protonic ceramic fuel cells.computational science. |
○ 벌크에서의 다종수송현상을 실험적으로 모니터할 수 있는 고온 직류 van der Pauw 법과 교류모니터링법을 개발하고 온도의존성으로부터 물 첨가반응의 비단조전도도 이완을 moving boundary 확산 메커니즘으로 밝힘.
○ 벌크 및 프로톤 연료전지 박막 전해질의 여러 주파수 분산 효과를 AC 전도도와 Bode plot를 통해 검토하고 Havriliak-Negami 유전함수와 혼합전도를 나타내는 트랜스미션라인모델로 기술하는 새로운 해석모델을 개발하고 물리적인 메커니즘이 반영된 패러미터로 온도와 주파수에 양상을 전면적으로
○ 벌크에서의 다종수송현상을 실험적으로 모니터할 수 있는 고온 직류 van der Pauw 법과 교류모니터링법을 개발하고 온도의존성으로부터 물 첨가반응의 비단조전도도 이완을 moving boundary 확산 메커니즘으로 밝힘.
○ 벌크 및 프로톤 연료전지 박막 전해질의 여러 주파수 분산 효과를 AC 전도도와 Bode plot를 통해 검토하고 Havriliak-Negami 유전함수와 혼합전도를 나타내는 트랜스미션라인모델로 기술하는 새로운 해석모델을 개발하고 물리적인 메커니즘이 반영된 패러미터로 온도와 주파수에 양상을 전면적으로 기술할 수 있게 함. 병행된 전산모사는 프로톤전도체 벌크 및 입계에서의 프로톤의 전도도는 물론 의견이 분분했던 정공 전도 메커니즘 규명하였음.
○ 모델 연료전지를 구성하고 산소극, 수소극 수분에 따른 OCV 와 임피던스 양상을 dynamic 하게 모니터할 수 있는 측정시스템을 개발함. 연료전지 구동과 연관하여 프로톤의 표면 및 전극표면에서의 전도 및 반응의 원자적 메커니즘을 계산함.
(출처:요약서 3p)
Ⅳ. Results and Discussion
1. Nonmonotonic dc conductivity relaxation
Bar-geometry samples have been generally used for dc 4-probe conductivity measurements at high temperature. High temperature van der Pauw method with disk samples was developed and applied to BZY20(Ni) which simplifies th
Ⅳ. Results and Discussion
1. Nonmonotonic dc conductivity relaxation
Bar-geometry samples have been generally used for dc 4-probe conductivity measurements at high temperature. High temperature van der Pauw method with disk samples was developed and applied to BZY20(Ni) which simplifies the diffusion in one-dimension and minimizes catalytic effects by the electrodes. The high and low apparent chemical diffusivity for hydrogen and oxygen incorporation for reduction and oxidation conductivity effects turned out to show the similar temperature dependence. This can be explained by the moving boundary diffusion mechanism heuristically suggested, where the diffusion in the hydration front layer determines the hydrogenation process. The mechanism was shown to be valid for all the reported work on the non-monotonic relaxation behavior and also supported by a recent numerical calculation by MPI, Stuttgart. While humidity and oxygen activity dependence of the chemical diffusivities in the literature can be explained in view of the defect chemical consideration, temperature dependence observed cannot be explained by the ideal behavior of the defects but by hole trapping effects which is higher in zirconates than in cerates. The hole conduction behavior in proton ceramic conductors is least known. The contradictory theoretical results are closely related with the well-known poorly estimated valence band levels in DFT calculations. Dc conductivity is shown to include the grain boundary effects in the impedance spectra supporting the mixed conduction origin of GB impedance.
2. Non-monotonic relaxation by AC method
Since LCR meters cover the sample response at high temperature, relaxation was monitored by AC method. The diffusion direction and electrical field direction is same in contrast to DC methods. It has been shown recently, for the small deviation of the concentration, the integral conductivity becomes similar. The measurement for the spectra of 61 pts from 10Hz to 10MHz takes 40 seconds only, similarly to DC methods using different current values or electrode combinations, but yields much more information. Application to BZY15 showed that AC conductivities at different frequencies exhibit similar non-monotonic relaxations but the capacitance monotonic increase upon hydration. Two different kinetic regimes are distinguished in any case, which depend on the frequencies. The capacitance behavior is not consistent with themixed conduction model: The chemical capacitance should decrease with the decreased hole concentration upon hydration. Hydration capacitance appears to dominate. Mixed conduction model successfully deconvoluted the monotonic relaxation of the proton conductivity at 550℃. Warburg-like electrode response suggests the mixed conduction effects associated with the chemical capacitance of electron holes. SCYb exhibited weak non-monotonic effects consistent with dc measurements. Poorly distinguished GB resistance varies together with bulk resistance and the capacitance effects showed the dependence on the electron hole concentration. The electrode resistance measured at low frequencies however clearly decrease with humidity.
3. New impedance spectroscopy
Also supported by the high temperature AC behavior in comparison with dc conductivity studies, the sample resistance including GB effects should be attributed to the electron hole resistance of BZY15. The impedance behavior may appear similar in conventional brick-layer model with high GB potential and the transmission line model for the mixed conductors with proton blocking GB and the hole conduction more activated than the proton conduction. In brick-layer model, at high enough temperature, bulk conduction determines the total sample conduction. In the transmission line model, the total conduction is determined by the hole conduction at all temperature. The present interpretation suggests the cube-root-law dependence of the hole conduction activation energy on yttrium concentration as known as bandgap narrowing in heavily doped p-type semiconductors. Not only the conductivity traces but the frequency dispersions are characteristic of the respective models. The brick-layer-like characteristics observed in BZY15 can be explained by the capacitance due to the blocked proton conductors. The element shows well-defined capacitance effects in the series of capacitance Bode plots, not CPE-like, and yet the strong frequency dispersion. The breakthrough was the employment of the complex dielectric functions such as Cole-Cole, Cole-Davidson, or most generally Havriliak-Negami response. For BZY15 data two Cole-Davidson functions in parallel worked well where the exponents and capacitance effects are temperature-independent and the relaxation time constants indicate the bulk conduction mechanism. The transmission model for the mixed conduction was also modified where the chemical capacitance was represented by the parallel circuit of hydration capacitance (ideal) and CPE. Non-trivial CPE-like GB response in LSN can be similarly described by two dielectric functions. On the other hand, the GB response of NASICON, a sodium conductor, can be described with a Havriliak-Negami response with the opposite skewedness to that of Cole-Davidson response. Huge GB impedance in LSN in almost ideal semicircular shape suggests the effects of a Schottky barrier, which appears to be activated below the phase transition temperature. Not only these complications associated with the grain boundary response, strong dispersion is generally associated with bulk response. It has been already well established, though not widely recognized, the mobile charge carriers in the solid electrolytes lead to apparent dielectric dispersion. The response can be exactly formulated in the time domain as the stretched exponential function with shape factor 1/3, named as CK1 model by Macdonald. The model response was evidenced in beta alumina, AgI, NASICON, SCYb, and many oxide ion conductors. However, in BZY (and also) in BCY different dispersion behaviour was indicated. Most clearly observed in LSN and LCN, the admittance curve exhibits the frequency dispersion close to 1/3, instead of ~0.6 for CK1 behavior. The behavior was also indicated in LLZ, lithium ion conductor. Therefore another type of ‘universal’ dispersion response appears to exist. As recently proposed for CK1 model, Cole-Davidson dielectric function can represent the bulk frequency dispersion in combination with other circuit elements. The parameters of the dielectric functions indicate well-defined physical mechanism. The present new approach can simulate the AC response over the wide temperature and frequency range with a few parameters either temperature-independent or well-defined. The parameters may be still correlated and there can be also equivalent circuit models but the approach fundamentally and properly describes the electrical behavior and should be considered an ultimate solution to the problems plaguing the practical impedance analysis for long time.
4. AC behavior of various ceramic proton conductors
The AC behavior is poorly understood even for SCYb introduced long time ago. The small GB effects are ascribed to the current-constriction effects and the bulk dispersion is CK1 type. While the p-type characteristics are well-studied in oxygen and humidity, clear n-type conduction with low activation energy is indicated in reducing atmosphere at low temperature. BCY15(KIST) exhibits a similar mixed conduction GB origin but the hole conduction and proton conduction are similarly activated. The electron conduction is indicated in reducing atmosphere but less dominant than in SCYb. BaCe1-xYxO₃ showed the change in activation energy of total sample conductivity similarly as shown for BZY system. The detailed impedance characteristics appear to be complicated by the microstructural effects. The dominance of the proton conduction at high temperature was shown by Ac relaxation behavior. BZY15 sample examined in different atmosphere confirmed the mixed conduction origin. Interestingly, BZY15 exhibited much higher n-type conduction than BCY15. BZY20(Ni), used for high T van der Pauw method, showed smaller GB effects, which may be due to the higher Y concentration, larger grain size, and probably also Ni doping effects. To understand more the non-trivial electrode and GB response, Ni/Au contact was used for low temperature measurements and an ideal semicircular electrode was distinguished with the background of the dispersive responses. Poorly sintered BZY10 exhibited huge GB resistance. The mixed conduction origin was indicated by the humidity effects. CuO sintering aids increased the density and decreased the GB effects substantially. The sample showed very pronounced humidity effects on the proton conduction, while the overall resistance is hardly affected. Bi₂O₃ addition improved the density and correspondingly GB effects, without a significant change in the AC behavior. 3 wt% MnCO₃ addition worked well as sintering aids but completely destroyed the proton conduction. The materials behaves an insulator with characteristic bulk dispersion with exponent 1/2. Pristine BZO shows the activation energy of 1.6 eV at high temperature, which is used for the cube-root law behavior. The AC Arrhenius plots distinguished the presence of the high frequency response of a protonic character. The method can be applied to the post-mortem analysis of BCY(Cu) and BZCY thin film electrolytes of KIST PCFCs. Characteristic frequency dependences of different electrolytes were distinguished. An in-depth impedance modeling of LSN was described in the previous section. LSN was also prepared by a modified solid solution method using precursors of C₄H₄NNbO9.xH₂O 와 La₂(CO₃)₃.xH₂O. The AC behavior of the samples sintered at 1500℃ were more or less similar. The modified method increases the density above 97% already at 1350℃ in contrast to 70% for the conventional method, which is of a great significance for the fuel cell applications. LCN, with Ca, showed higher conductivity thant LSN. CuO additive was previously reported to increase the sinterability. Grain growth occurred extensively with large intragranular porosity. The conductivity is however greatly decreased. At low temperature characteristic 1/2 dependence is distinguished similarly as in Mn-BZY.
5. PCFC monitoring
A parametric analysis of impedance behavior of the state-of-the-art Ni-YSZ/YSZ/LSM cells has been recently performed as a function of bias (FC/EC) and temperature. It is no wonder PCFC impedance is far from being understood in view of the complications presented above. Model PCFC system was examined using Pt electrodes on BCY05 and BCY15 electrolyte disks. AC dynamic monitoring was performed with LCR meter by applying the dc bias values corresponding to the monitored OCV values. For PCFC/PCEC configuration, no appreciable relaxation, monotonic or non-monotonic, was observed. Polarization decreases with EMF, i.e. from fuel cell to electrolyzer mode. Bias and humidity independent peak frequencies were found in the imaginary admittance Bode plots, which is suggested to represent the RC series component in the transmission line model both for the electrolyte and electrode polarization.
Cermet-supported Ni-BCZY/BCZY/BSCF cells provided by KIST exhibited similar polarization characteristics in fuel cell/electrolyser operation. Gerischer-like single polarization component was distinguished which can be interpreted as the surface reaction rate increases with increasing hydrogen activity in the hydrogen electrode by electrolysis. The behavior could not be explained by the gas concentration impedance in the cermet layer the polarization of which should increase with decreasing humidity. Similarly as observed for the model PCFC with bulk electrolytes carbonates were formed on the proton conducting electrolytes, which is the origin of the degradation.
6. Computer simulations
The behavior of defects in BZO(BZY) system was investigated via simulation. Proton and oxygen vacancy concentrations were calculated using hydration enthalpy and entropy. Proton migration barrier was calculated for different configurations and the value of 0.4 eV, similar to the experimental observation was obtained. The conductivity obtained from the concentration and mobility was consistent with the experimental observation. The hole concentration is determined by the oxidation enthalpy. The oxidation enthalpy was estimated by 0.875 eV. Hole migration barrier was estimated to 0.02 eV, negligibly small. The hole conductivity experimentally observed is thus consistent with that of the hole concentration. The grain boundary core Σ3 (111)/[110] was found to consist of BaO₃. The segregation enthalpy for proton and oxygen vacancy was calculated and the proton and oxygen vacancy concentrations were calculated using the space charge model. Protons segregate easily on the grain boundary core. The proton migration barrier was estimated. The grain boundary conductivity thus obtained was consistent with the experimental results. BaO surface configuration of BaZrO₃ (001) plane was found most stable. The proton segregation enthalpy was estimated as –0.97 and –0.43, respectively. The incorporation of surface protons was found very difficult with the migration barrier of 1.10 eV. The surface conduction was estimated lower by 106 than in the bulk. Nickel electrode clusters were found stable on BaZrO₃ (001) plane with ZrO₂ configuration due to the larger oxygen concentration than BaO layer. Therefore, the incorporation of the proton is supposed more difficult than the protonation of hydrogen at the electrodes.
(출처:SUMMARY 15~22p)
과제명(ProjectTitle) : | - |
---|---|
연구책임자(Manager) : | - |
과제기간(DetailSeriesProject) : | - |
총연구비 (DetailSeriesProject) : | - |
키워드(keyword) : | - |
과제수행기간(LeadAgency) : | - |
연구목표(Goal) : | - |
연구내용(Abstract) : | - |
기대효과(Effect) : | - |
Copyright KISTI. All Rights Reserved.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.