Oxygen Reduction Kinetics Enhancement on a Heterostructured Oxide Surface for Solid Oxide Fuel Cells Ethan J.Crumlin,†,þEva Mutoro,†,þSung-Jin Ahn,†Gerardo Jose la O',†Donovan N.Leonard,‡Albina Borisevich,‡Michael D.Biegalski,§Hans M.Christen,§and Yang Shao-Horn*,†
†Electrochemical Energy Laboratory,Massachusetts Institute of Technology,77Massachusetts Avenue,Cambridge, Massachusetts02139,United States,‡Materials Science and Technology Division,and§Center for Nanophase Materials Sciences,Oak Ridge National Laboratory,Oak Ridge,Tennessee37831,United States
ABSTRACT Heterostructured interfaces of oxides,which can exhibit transport
and reactivity characteristics remarkably different from those of bulk oxides,are
interesting systems to explore in search of highly active cathodes for the oxygen
reduction reaction(ORR).Here,we show that the ORR of∼85nm thick La0.8Sr0.2-
CoO3-δ(LSC113)films prepared by pulsed laser deposition on(001)-oriented yttria-
stabilized zirconia(YSZ)substrates is dramatically enhanced(∼3-4orders of magni-
tude above bulk LSC113)by surface decorations of(La0.5Sr0.5)2CoO4(δ(LSC214)with
coverage in the range from∼0.1to∼15nm.Their surface and atomic structures were
characterized by atomic force,scanning electron,and scanning transmission electron
microscopy,and the ORR kinetics were determined by electrochemical impedance
spectroscopy.Although the mechanism for ORR enhancement is not yet fully under-
stood,our results to date show that the observed ORR enhancement can be attributed to
highly active interfacial LSC113/LSC214regions,which were shown to be atomically
sharp.
SECTION Surfaces,Interfaces,Catalysis
T he efficiency of solid oxide fuel cells(SOFCs)is limited primarily by the oxygen reduction reaction(ORR)at
the cathode,and there is a need to search for electrode materials with enhanced ORR activity,particularly for SOFCs operated at intermediate temperatures.Mixed electronic and ionic conductors,such as ABO3perovskites1-3and A2BO4 Ruddlesden-Popper materials,4are promising cathode mate-rials due to their high oxygen ion diffusivity and surface exchange properties.Recently,heterostructured oxide inter-faces have shown surprisingly high transport or oxygen sur-face exchange properties.5-14In particular,Sase et al.13have reported enhanced ORR kinetics of∼3orders of magnitude confined to an interfacial region of∼20μm in width between La0.6Sr0.4CoO3-δand(La,Sr)2CoO4(δrelative to their bulk values.Subsequently,these authors demon-strated∼1order of magnitude enhancement in activity for the composite cathode screen-printed with these two oxide materials.14More recently,we have reported that epitaxial La0.8Sr0.2CoO3-δthin films grown on YSZ exhibit enhanced ORR kinetics up to∼2orders of magnitude relative to bulk.15
In this study,we examine whether(La0.5Sr0.5)2CoO4(δ(LSC214)surface decoration on epitaxial La0.8Sr0.2CoO3-δ(LSC113)thin films would lead to any further enhancement in ORR kinetics.We here report that LSC214surface decoration on the LSC113thin films provides at least one additional order of
magnitude of enhancement in ORR kinetics,achieving ∼3-4orders of magnitude enhancement over the entire electrode surface relative to bulk LSC113.16Using geometri-
cally well-defined thin film microelectrodes,particularly help-
ful to provide insight into ORR mechanisms,17-24we show that the active regions responsible for the ORR enhancement
are the interfacial regions between LSC113and LSC214,which
is in good agreement with enhanced ORR activity observed at
the interface of La0.6Sr0.4CoO3-δ/(La0.5Sr0.5)2CoO4(δ.13 Epitaxial thin films of LSC113of∼85nm thickness were
prepared by pulsed laser deposition(PLD)on yttria-stabilized
zirconia(YSZ(001))single crystals with gadolinium-doped ceria(GDC)as the buffer layer(Figure1a)using conditions similar to those reported previously.15A surface decoration with LSC214having different thicknesses,from partial to full coverage(∼0.1,∼0.8,∼5,∼15nm),was deposited subse-quently on top of
LSC113/GDC/YSZ(001),while a reference sample with surface decoration of∼0.1nm LSC113(LSC113 reference)was made for comparison.Scanning electron micro-scopy(SEM)and atomic force microscopy(AFM)images showed that the LSC113reference and LSC214surface decorated films were dense,having low surface roughness of∼1nm (Figures1b and S1,Supporting Information).Electrochemical impedancespectroscopy(EIS)measurementswereusedtoprobe ORR kinetics on geometrically well-defined LSC113and LSC113/ LSC214microelectrodes(Figure1b)created by photolithography
Received Date:August28,2010
Accepted Date:October1,
2010
Published on Web Date:October15,2010
and acid etching,where sintered porous Pt served as the counter electrode.EIS data were analyzed using an equivalent circuit depicted in Figure 1c,from which the ORR resistance (R ORR )and surface oxygen exchange rate were obtained.
Similar to our previous studies,15normal X-ray diffraction (XRD )data (Figure 2a )clearly show the presence of the (00l )pc (l is integer )or (00l )cubic (l is even )peaks of LSC 113,GDC,and YSZ,which indicates (001)pc LSC 113//(001)cubic GDC//(001)cubic YSZ.The subscript “pc ”denotes the pseudocubic notation.With increasing LSC 214coverage to ∼5and ∼15nm in thickness,the (00l )tetragonal peaks (l is even )of LSC 214become visible,corresponding to c tetragonal =12.5Å,comparable to
published
Figure 1.(a )Schematic of a LSC 214/LSC 113/GDC/YSZ (001)/porous Pt sample and electrochemical testing configuration (not drawn to scale;dimensions provided in the Experimental Methods ),(b )optical image of a micropatterned LSC 113reference sample and magnified surface topography (SEM/AFM )with a full height range of ∼6nm in the AFM image,(c )equivalent circuit (R 1=YSZ electrolyte resistanc
e,R 2=electrode/electrolyte interface resistance,15R ORR =ORR resistance,CPE =constant phase element )used to extract ORR kinetics,and (d )characteristic Nyquist plot schematic (color key:orange =YSZ/bulk transport,green =GDC/interface,blue =LSC/ORR )
.
Figure 2.X-ray diffraction analysis.(a )Normal XRD of the LSC 113reference and the LSC 214-covered
LSC 113films;LSC 113peaks are labeled in the pseudocubic (pc )notation,15GDC and YSZ peaks are marked by open circles (O )and open triangles (3),respectively.(b )Off-normal XRD of a similarly prepared sample with a thicker (∼45nm )LSC 214coverage,(c)schematic of the crystallographic rotational relationships among the LSC 214(001)tetragonal ,LSC 113(001)pc ,GDC (001)cubic ,and YSZ (001)cubic .
values.25,26Off-normal phi-scan analysis of a similarly prepared sample with a thicker (∼45nm )LSC 214coverage (Figure 2b )allowed us to identify the in-plane crystallographic relationships between GDC and YSZ (a cube-on-cube alignment ),LSC 113and GDC (a 45°rotation having [100]pc LSC 113//[110]cubic GDC//[110]cubic YSZ ),and LSC 113and LSC 214(no rotation having [100]pc LSC 113//[100]tetragonal LSC 214),shown in Figure 2c and Figure S2,Supporting Information.The relaxed lattice param-eters of the LSC 113films with and without LSC 214surface decoration in this study were found to be in the range of 3.84-3.85Å,having in-plane tensile strains and compressive strains in the direction normal to the film surface at room temperature (T able S1,Supporting Information ),which are in reasonably good agreement with those reported previously .15The origin of these strains is not unde
rstood fully but might be a consequence of different thermal expansion coefficients between YSZ (∼11Â10-6°C -127)and LSC films (∼17Â10-6°C -1for bulk 28).Experiments are ongoing to examine how these strains change upon heating to high temperatures.
EIS data of all samples used in this study were found to be very similar in shape,and typical features in the Nyquist plots are shown in the schematic in Figure 1d.The predominant semicircle (assigned to ORR impedance )was found to increase with decreasing p (O 2).Representative EIS data collected at 550°C in 1%p (O 2)are shown in Figure 3a,b.All films with LSC 214surface decoration were found to have markedly smaller real impedance relative to the LSC 113films reported previously 15and the LSC 113reference film of this study .The ORR area specific resistance (ASR ORR =R ORR 3Area electrode )was used to calculate the electrical surface exchange coefficient,k q29,30(Figure S3,Supporting Information ).As shown in Figure 3c,d,all of the LSC 214-decorated films showed enhanced k q values compared to the LSC 113reference,electrodes with low coverage demon-strating the highest surface exchange.It is interesting to note that the surfaces with low coverage exhibit up to an order of magnitude higher conductance at high p (O 2)than that observed for a screen-printed electrode composed of porous LSC 214on top of porous La 0.6Sr 0.4CoO 3-δreported previously 14(Figure
S4,
Figure 3.Electrochemical impedance spectroscopy (EIS )results of microelectrodes (∼200μm )for the LSC 113reference and the LSC 113films with ∼0.1,∼0.8,∼5,and ∼15nm LSC 214surface coverage at 550°C;(a )Nyquist plot at 1%p (O 2);the inset and (b )show a magnification;(c )oxygen partial pressure dependency of the surface exchange coefficients k q of LSC 113/LSC 214films and bulk LSC extrapolated from the data reported by De Souza et al.16(red ();and (d )average k q values versus thicknesses of LSC 214surface coverage.Error bars represent the maximum and minimum from three independently measured electrodes.
Supporting Information).Assuming that k q can be approxi-mated as k*,31it is noted that the highest k q values of(∼0.1nm) LSC214-decorated LSC113films obtained from this study are 3-4orders of magnitude higher than that of bulk LSC11316at 1atm and are comparable to those of the most active cathode materials such as thin film Ba0.5Sr0.5Co0.8Fe0.2Co3-δ[k*=∼1Â10-6cm s-1at500°C and p(O2)=0.5bar32]and bulk La2CoO4[k*=∼3Â10-6cm s-1at500°C and0.2bar33].
To evaluate what is responsible for the observed ORR enhancement,we here discuss factors that may influence the ORR kinetics.We note that there is no change in the oxygen vacancy concentration of the entire film with increasing LSC214 surface coverage.The oxygen nonstoichiometry(δ)was esti-mated from the volume-specific capacitance(VSC)of EIS data using a method published previously15,34(Figure S5a and b, Supporting Information),from which thermodynamic param-eters of the thin film samples were extracted from VSC dependence on p(O2)(T able S2,Supporting Information). Therefore,the enhanced ORR activity of LSC214-covered LSC113films cannot result from increased oxygen vacancy in the bulk of the films.Moreover,the origin of enhanced ORR activity of LSC214-covered LSC113films is not apparent as the LSC214reference film composed of LSC214(∼15nm)/GDC/ YSZ(001)/porous Pt has comparable activity to the LSC113 reference film(Figure S6,Supporting Information).These results suggest that the presence of LSC113and LSC214inter-faces on the surface of LSC214-covered LSC113films is critical to the observed ORR enhancement.
To obtain further insight into the ORR kinetics of LSC214-covered LSC113films,high-angle annular dark field scanning transmission electron microscopy(HAADF STEM)cross sections of three samples,LSC113reference and a partially(∼0.8nm) and a fully(∼15nm)covered LSC214film were compared,
from which active regions for surface oxygen exchanges are
proposed(Figure4).Cross-sectional STEM imaging(Figure4a) demonstrated that the LSC113reference film had no second-phase inclusions.By the first approximation,the ORR
kinetics on these films of∼85nm thickness are oxygen surface exchange limited,having the entire electrode sur-face ORR active(highlighted in red in Figure4a right panel) as the film thickness is much lower than the critical thickness
estimated from extrapolated bulk LSC113values(using t crit= D bulk/k bulk at550°C≈1μm).35This is further supported by the fact that the p(O2)dependency of k q(k qµp(O2)m,with m=∼0.93)is similar to that reported previously for bulk LSC,36particularly in the p(O2)range from10-4to10-2atm. Partial LSC214coverage on the LSC113film led to the forma-
tion of nanoislands distributed on top of the LSC113film, which is supported by the STEM cross-sectional imaging in Figure4b,showing uneven contrast at the film surface (compared with Figure4a).Nanoislands result in a large number of interfacial LSC113/LSC214boundaries on the film surface,which are correlated with the highest ORR activity values observed for partial coverage of LSC214on the LSC113 film.Increased k q values can decrease the critical thickness (t crit),assuming a
constant diffusion coefficient,D,which can lead ORR impedance to be influenced by bulk diffusion in addition to surface oxygen exchange.The much de-creased p(O2)dependency of the partially covered samples at high p(O2)(Figure3c)may suggest that the surface oxygen exchange kinetics of these thin films are so large that they are limited by bulk diffusion in the thin films,where bulk diffusion coefficients typically have a weak dependence on p(O2).
With increasing LSC214coverage to∼15nm,a fully dense surface layer of LSC214was developed,as shown in the cross-sectional STEM image in Figure4c,where the LSC214layer appears darker due to a lower physical density of the structure than LSC113.25,37This reduced the triple-phase boundary (LSC113/LSC214/gas phase),which is correlated with lower activity of fully covered LSC214films compared to the partially covered films.Even assuming that the edge of the LSC214thin layer had ORR activity comparable to that of the boundaries of LSC214nanoislands on the LSC113film,a large active area at the interface between LSC214and LSC113extending from the microelectrode edge toward the microelectrode center would be needed to explain the ORR enhancement of these films relative to that of the LSC113and LSC214reference samples (Figures4c,5,and S6,Supporting Information).Evidence to support this hypothesis is provided by evaluating how the ASR ORR changes with the microelectrode diameter,d.As shown in Figure5a,all partially covered LSC214electrodes show
no apparent ASR ORR dependency on the electrode diameter,which indicates that the enhanced ORR kinetics are comparable for all electrodes of different sizes.In
contrast, Figure4.Proposed ORR active regions(red)for the(a)LSC113 reference sample,(b)partially covered LSC214samples,and (c)fully covered LSC214sample(HAADF STEM reveals∼15nm LSC214coverage).
the ASR ORR of the fully covered LSC 214film decreases (i.e.,k q increases )with reducing electrode size,approaching a similar value to that for partially covered electrodes.This result suggests that an interfacial area extending from the micro-electrode edge to the electrode center with a finite radial distance,r ,on the order of ∼20μm (as depicted in Figure 5b )can contribute to enhance ORR kinetics.By decreasing the electrode diameter from ∼160to ∼35μm,the area of the highly active circular ring becomes greater than the remaining non-ORR-enhanced area near the electrode center (Figure 5b ),and finally ,the entire electrode is highly activated.This can explain why the ASR ORR of this sample approaches those of partially covered LSC 214samples.
Our results indicate that interfacial LSC 113/LSC 214regions or LSC 214thin layers of a few nanometers in thickness on LSC 113are responsible for the observed ORR enhancement.Atomic-resol
ution HAADF STEM imaging (Figure 6)revealed that the LSC 214layer of ∼15nm grew epitaxially with (001)-orientated LSC 113,which is in good agreement with XRD results in Figure 2.The interface is atomically sharp without
any amorphous or disordered interface regions,having atomic columns of the La/Sr layers of LSC 214and LSC 113clearly shown.It is not clearly understood why interfacial LSC 113/LSC 214boundaries or LSC 214thin layers of a few nanometers in thickness on LSC 113exhibit enhanced ORR activity.We hypothesize that the enhancement can be attributed to inter-facial properties such as strain,9,15space charge effects,8an increase in oxygen vacancy concentration,38or a change in electronic structure.39Since La 2CoO 4has a high k *33which is comparable to the k q found for the partially covered LSC 113/LSC 214samples,the enhanced ORR activity could also be explained by the absence of Sr at the interface.Further studies are needed to elucidate the mechanism of ORR kinetics enhancement of these heterostructured interfaces.
In summary,we show ORR activity enhancement up to ∼1-3orders of magnitude on the thin film LSC 113surfaces with LSC 214surface decoration and up to ∼3-4orders of magnitude with respect to bulk LSC 113.Our results indicate that interfacial LSC 113/LSC 214regions are responsible for the observed ORR enhancement.The enhancement in ORR kinetics cannot be attributed to changes in oxygen vaca
ncy concentration of the entire film nor a disordered interfacial LSC 113/LSC 214structure,and future studies are needed to elucidate the origin of enhanced ORR kinetics at interfacial LSC 113/LSC 214regions.This study illustrates the potential of utilizing heterostructured oxide surfaces/interfaces to develop highly active surface oxygen exchange materials for applica-tions in the field of solid-state electrochemistry such as solid-electrolyte-based sensors,oxygen-conducting membranes,and SOFC cathodes.
EXPERIMENTAL METHODS
Pt ink (#6082,BASF )counter electrodes were painted on one side of YSZ (001)single crystals (9.5mol %Y 2O 3,Princeton Scientific )with dimensions of 10Â10Â0.5mm and sintered for 1h at 800°C in air .PLD deposition of GDC (500laser pulses,∼5nm thickness ),LSC 113(15000pulses,(∼85nm;determined by AFM )and the surface coverages (either 25pulses of LSC 113(reference sample )or 25(∼0.1nm ),150(∼0.8nm ),900(∼5nm ),or 2700(∼15nm;determined by TEM cross section )pulses of LSC 214)was performed using self-synthesized stoichio-metric targets under the following PLD conditions:KrF excimer laser ,λ=248nm,10Hz pulse rate,∼50mJ pulse energy ,p (O 2)PLD and cooling =10mTorr ,growth temperatures,T GDC =450°C,T LSC 113,LSC 214=675°C,t cooling =1h.Thin film high-resolution XRD was performed in normal and off-normal configuration with a four-circle diffractometer (
Panalytical and Bruker D8).Surface morphology was examined by optical microscopy (Carl Zeiss ),AFM (Veeco ),and SEM (Carl Zeiss ).Microelectrodes (diameters ∼75,∼100,∼150,and ∼200μm,exact diameter determined by optical microscopy )were pre-pared using photolithography .EIS measurements (electrodes contacted by Pt-coated tungsten carbide probes )were carried
out using a microprobe station (Karl S €u
ss )connected to a frequency response analyzer (Solartron 1260)and dielectric interface (Solartron 1296).Data were collected between 1MHz and 1mHz using a voltage amplitude of 10mV under Ar/O 2mixtures in the p (O 2)range of 1-10-4atm.The
temperature
Figure 5.(a )ORR area-specific resistance (ASR ORR )versus elec-trode diameter,d ,at 550°C in 1%O 2.(b )Schematic of proposed enhanced active regions (red )for different electrode sizes and different amounts of LSC 214surface coverage;r =radial distance of active area.
reported
of 550°C was measured on the sample surface and controlled with a heating stage (Linkam TS1500).ZView software (Scribner Associates )was used to analyze the EIS data.STEM (VG501UX or Nion UltraSTEM100)was utilized to examine thin film cross sections and the atomic-resolution interface of LSC 113/LSC 214.Additional experimental details can be found in the Supporting Information.
SUPPORTING INFORMA TION A VAILABLE Details about
experimental methods,a table containing thin film measured in-plane/normal/calculated relaxed lattice parameters and in-plane/normal strain,SEM/AFM images of all samples,schematic relating the structur
e of LSC 113to LSC 214,ASR ORR versus p (O 2),information regarding the k q calculation,conductance and k q literature comparison,VSC versus p (O 2),δversus p (O 2),thermodynamic parameters,and Nyquist plot comparing thin film EIS of LSC 214reference (∼15nm )to LSC 113reference (∼85nm ).This material is available free of charge via the Internet at
AUTHOR INFORMATION Corresponding Author:
*To whom correspondence should be addressed.E-mail:shaohorn@mit.edu.
Author Contributions:
þ
These authors contributed equally to this work.
ACKNOWLEDGMENT This work was supported in part by the NSF
(CBET 08-44526),DOE (SISGR DE-SC0002633),and King Abdullah University of Science and Technology.E.M.is grateful for financial support from the German Research Foundation (research schol
arship ).The portion of research performed at the Center for Nanophase Mate-rials Sciences as well as FIB instrument access via ORNL's ShaRE user facility was sponsored by the Scientific User Facilities Division,Office of
Basic Energy Sciences,U.S.DOE.The STEM work was sponsored by the Materials Sciences and Engineering Division,Office of Basic Energy Sciences of the U.S.DOE.
REFERENCES
(1)Adler,S.B.Factors Governing Oxygen Reduction in Solid Oxide Fuel Cell Cathodes.Chem.Rev.2004,104,4791–4843.(2)Adler,S.B.;Chen,X.Y .;Wilson,J.R.Mechanisms and Rate Laws for Oxygen Exchange on Mixed-Conducting Oxide Surfaces.J.Catal.2007,245,91–109.
(3)
Lee,Y .L.;Kleis,J.;Rossmeisl,J.;Morgan, D.Ab Initio Energetics of LaBO 3(001)(B=Mn,Fe,Co,and Ni )for Solid Oxide Fuel Cell Cathodes.Phys.Rev.B 2009,80,224101/1–224101/20.
(4)
T arancon,A.;Burriel,M.;Santiso,J.;Skinner,S.J.;Kilner,J.A.Advances in Layered Oxide Cathodes for Intermediate Tem-perature Solid Oxide Fuel Cells.J.Mater.Chem.2010,20,3799–3813.
(5)
Garcia-Barriocanal,J.;Rivera-Calzada,A.;Varela,M.;Sefrioui,Z.;Iborra,E.;Leon,C.;Pennycook,S.J.;Santamaria,J.Colossal Ionic Conductivity at Interfaces of Epitaxial ZrO 2:Y 2O 3/SrTiO 3Heterostructures.Science 2008,321,676–680.(6)
Garcia-Barriocanal,J.;Rivera-Calzada,A.;Varela,M.;Sefrioui,Z.;Iborra,E.;Leon,C.;Pennycook,S.J.;Santamaria,J.Response to Comment on “Colossal Ionic Conductivity at Interfaces of Epitaxial ZrO 2:Y 2O 3/SrTiO 3Heterostructures.Science 2009,324,5926.
(7)
Guo,X.Comment on Colossal Ionic “Conductivity at Inter-faces of Epitaxial ZrO 2:Y 2O 3/SrTiO 3Heterostructures.Science 2009,324,5926.
(8)Guo,X.X.;Maier,J.Ionically Conducting Two-Dimensional Heterostructures.Adv.Mater.2009,21,2619–2631.
(9)
Korte,C.;Schichtel,N.;Hesse,D.;Janek,J.Influence of Interface Structure on Mass Transport in Phase Boundaries between Different Ionic Materials Experimental Studies
and
Figure 6.HAADF STEM micrograph of the LSC 113/LSC 214interface for the ∼15nm LSC 214-covered sample;the schematic shows the unit cell of both materials and the structure at the interface (blue atoms =La/Sr;purple atoms =Co ).
发表评论