Phase diagram plays a central role in materials discovery and development. Conventional way of phase diagram construction, featured by synthesizing and characterizing one sample at a time, is costly, time-consuming, unsystematic, and, given these disadvantages, insufficient for the current era of accelerating technology. In contrast, combinatorial materials chip technology, including high-throughput synthesis and characterization of materials libraries composed of 102–104 samples on one small substrate, has demonstrated great potential to accelerate the process significantly. In this study, we attempt to establish isothermal section of a ternary phase diagram by combinatorial materials chip preparation technique, combined with rapid, accurate and systematic compositional and structural characterization by synchrotron radiation and hierarchical clustering analysis. The multi-layered thin film precursors covering the entire composition range of Fe-Co-Ni ternary system was prepared using a high throughput combinatorial ion beam deposition system (HTC-IBD) with a continuously moving shutter. The composition of each sample (pixel) is governed by the relative thickness of the three constituent layer on the pixel. The overall thickness of the film stack was 100 nm. The as-deposited chip was sealed into an evacuated quartz tube and crystallized isothermally in a muffle furnace at the designated temperature for 2 hours. Micro-beam X-ray diffraction was performed pixel-by-pixel on beamline 11-1D-D of APS/ANL at Argonne National Laboratory. The diffracted signal was collected using a Pilatus 2M area detector by Dectris Inc at a rate of 1 s/pattern. Composition spread of the chip was characterized by micro-beam X-ray fluorescence on beamline BL 7-2 at SSRL, calibrated by WDS analysis. A total number of over 400 effective X-ray diffraction patterns were automatically processed, phase-identified and categorized by machine learning process. Peak positions instead of the full patterns were used as the input for hierarchical clustering analysis by CombiView. The isothermal section of the Fe-Co-Ni ternary phase diagram constructed by the combinatorial chip approach in this study is consistence with the phase diagram in the ASM Alloy Phase Diagram DatabaseTM. Details of the constructed phase diagram indicates that classification by machine learning exceeds those by human classification.
Rapid Construction of Fe-Co-Ni Phase Diagram by Combinatorial Materials Chip with a Continuous Composition Spread
Type Of Event:
Lanting Zhang, Materials Genome Initiative Center, Shanghai Jiao Tong University
Wednesday, April 26, 2017