EELS & EDX spectrum imaging : pushing the limits

Transcription

1 EELS & EDX spectrum imaging : pushing the limits David W McComb Department of Materials Science and Engineering The Ohio State University Columbus, Ohio USA Note : Seven slides showing unpublished data removed

2 Acknowledgements Center for Electron Microscopy & Analysis at The Ohio State University Frank Scheltens, Robert Williams, Dan Huber, Hamish Fraser, Srini Rajagopalan FEI Company Anna Carlsson, Dmitri Klenov Imperial College London John Kilner, Stuart Cook, Monica Burriel CSIC-ICN, Barcelona A.Cavallaro, J.Roqueta, A. Apostolidis, A.Bernardi, A.Tarancón, J.Santiso Universidad Complutense, Madrid Jacobo Santamaria, Carlos Leon, Alberto Rivera 2

3 Motivation YSZ STO 10 nm STO

4 Surprising...but enormous potential if a real effect! Temperature ( C) Bi 2 O multilayer SDC/CeO 2 1nm YSZ log σ (Scm -1 ) YSZ/Y 2 O 3, Korte et al PCCP (2008) 10, GDZ/GDC on Al 2 O 3, Wang et al SSI (2006) 177, 1299 SDC/CeO 2 on MgO, Kosacki unpublished YSZ/STO, Garcia-Barriocanal et al Science (2008) 321, 676 thin film YSZ -12 YSZ thin films on MgO/Al 2 O 3, Karthikeyan et al APL (2006) 89, 183 YSZ thin films on MgO, Kosacki et al SSI (2005) 176, /T (K -1 ) Slide courtesy of John Kilner

5 PLD grown : SrTiO3 YSZ SrTiO3 multilayer Extensive mixing initially (consistent with in-situ RHEED). Change to layer-by-layer after 3 repeats (RHEED) A. Cavallaro et al. Solid State Ionics 181 (2010)

6 PLD : SrTiO3 YSZ SrTiO3 multilayer Coherent or semi-coherent chains Incoherent islands A. Cavallaro et al. Solid State Ionics 181 (2010)

7 PLD : EELS analysis the limits A. Cavallaro et al. Solid State Ionics 181 (2010)

8 EDX hyperspectral mapping HAADF O Sr Ti [110] SrTiO3

9 Care : There are non-local effects in spectroscopic imaging Split signal into elastically scattered electrons and thermally diffused electrons EDX map for oxygen K-edge in SrTiO 3 [001]: Absorptive Model O Ti/O Sr Ti/O Fractional intensity Fractional intensity Probe position (Å) Probe position (Å) Probe position (Å) Probe position (Å) Elastic signal Thermal signal Born-Oppenheimer Contribution of thermally scattered electrons to atomic resolution elemental maps. Forbes, McComb, Allen et al, Phys Rev B 86 (2012)

10 Oxygen and Titanium Signals Collected from Ti/O O Ti/O Columns 300 kv α = 24 mrad β = 39 mrad t/λ =.19λ

11 Quantification : entire multilayer C A C B = k AB I A I B Fit background remove Bremsstrahlung Fit peaks overlap issues Integrate to obtain net counts Quantify Cliff Lorimer k-factor

12 Why calculated k-factors are dangerous! A : atomic weight Q : ionization cross-section ρ : density µ : linear absorption coeff. a : relative transition probability t : thickness ω : fluorescence yield ε : detector effciency Calculated values much have higher errors than the experimental data.

13 Determine Experimental k-factors Bulk SrTiO 3 wafer/substrate Bulk Y 2 O 3 -ZrO 2 crystal : composition established by SEM- EDX with elemental standards and Laser Ablation Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS) C A C B = k AB I A I B

14 Coherent layers in RF sputtered STO-YSZ-STO Samples grown & provided by J Santamaria et al

15 RF sputtered STO-YSZ-STO HAADF Sr Zr Ti Four 1nm layers of YSZ EDX shows that substrate STO buffer interface is Sr-rich YSZ layers appear to be continuous Samples grown & provided by J Santamaria et al Y

16 EDX mapping of RF sputtered STO-YSZ-STO PLD Sr Ti Zr Y

17 Quant 1 : RF sputtered STO-YSZ-STO Entire Layer : 13.6 mol% Y 2 O 3

18 Quantification 2 : RF sputtered STO-YSZ-STO Divide sample into 10 sub-regions Sub-regions are (256 x 4) pixels i.e. each box is 0.4nm in height Quantify each box for Sr, Ti, Y & Zr

19 Part I : Conclusions, comments & caveats Samples grown by BOTH PLD and RF sputtering show Y/Zr rich layers that are coherent with STO Sr & Ti present throughout the layer Ti concentration is more reduced than Sr in the Zrlayer Averaged composition of layers is higher than nominal target values Lower STO-YSZ interface is smoother/flatter than upper YSZ-STO interface

20 Computational study of structures of Y 2 O 3 - stabilised ZrO 2 /SrTiO 3 multilayers Wei Li Cheah, Mike W. Finnis, David W. McComb Department of Materials, Imperial College London, UK

21 Chua et al, A genetic algorithm for predicting the structures of interfaces in multicomponent systems., Nat Mater, 9 (2010) 418 Genetic algorithm Crossover Initial configuration New generation The genetic algorithm Mutation Tournament selection Exchange

22 YSZ/STO supercell Initial configuration for the genetic algorithm STO z ZrO 2 with 2 Zr 4+ replaced with Y 3+ and one O 2- removed Supercell with SrO-terminated STO interfaces STO O Zr Ti Sr

23 Result from the genetic algorithm and CASTEP O Zr Ti Sr Y Fluorite YSZ structure unstable Spontaneously transforms into a different structure Low energy structure from genetic algorithm confirmed with energy calculations within CASTEP This structure only with excess SrO? What about stoichiometric STO/excess TiO 2?

24 Result from the genetic algorithm and CASTEP O Zr Ti Sr Y Intermixing of YSZ and STO: Y 3+ in Sr 2+ lattice site Sr 2+ in Zr 4+ lattice site Zr 4+ in Ti 4+ lattice site Ti 4+ in Zr 4+ lattice site Cheah WL, Finnis MW, Structure of multilayer ZrO2/SrTiO3, J. Mat Sci, 47 (2012) 1631

25 Conclusion O Zr Ti Sr Y The lowest energy YSZ structure in a YSZ/STO multilayer, determined using a combination of classical and first principles-based methods, is different from the expected fluorite lattice and appears to favour interdiffusion. Cheah WL, Finnis MW, Structure of multilayer ZrO2/SrTiO3, J. Mat Sci, 47 (2012) 1631

26 Conclusions & Comments EDX spectrum imaging is easier to quantify that EELS mapping. Quantification on the sub-nanometre scale achievable provided due care is given to determination of experimental k-factors (or z-factors) and channelling effects. Atomic scale chemical maps provide qualitative insights but. Quantification on the atomic scale requires simulation of the images/maps due to non-local contributions in both EELS & EDX How do we relate the role of compositional fluctuations on ionic conductivity properties.

27 Thanks to Ohio Research Scholar Award (ODOD Third Frontier program) A*STAR programme (Wei Li Cheah) UK Engineering and Physical Sciences Research Council (EPSRC)