Low-hysteresis shape-memory ceramics designed by multimode modelling.

Zirconia ceramics exhibit a martensitic phase transformation that enables large strains of order 10%, making them prospects for shape-memory and superelastic applications at high temperature1-5. Similarly to other martensitic materials, this transformation strain can be engineered by carefully alloying to produce a more commensurate transformation with reduced hysteresis (difference in transformation temperature on heating and cooling)6-11. However, such 'lattice engineering' in zirconia is complicated by additional physical constraints: there is a secondary need to manage a large transformation volume change12, and to achieve transformation temperatures high enough to avoid kinetic barriers6. Here we present a method of augmenting the lattice engineering approach to martensite design to address these additional constraints, incorporating modern computational thermodynamics and data science tools to span complex multicomponent spaces for which no data yet exist. The result is a new zirconia composition with record low hysteresis of 15 K, which is about ten times less transformation hysteresis compared to typical values (and approximately five times less than the best values reported so far). This finding demonstrates that zirconia ceramics can exhibit hysteresis values of the order of those of widely deployed shape-memory alloys, paving the way for their use as viable high-temperature shape-memory materials.

Structural stability of Co-V intermetallic phases and thermodynamic description of the Co-V system.

The Co-V system has been reviewed. Density functional theory (DFT) calculations using the generalized gradient approximation (GGA) were used to obtain the energies for the end-members for all three intermediate phases, Co3V, σ and CoV3. Results from DFT calculations considering spin polarization were used to evaluate the CALPHAD (Calculation of phase diagrams) model parameters. The method to evaluate the contribution of the magnetism to the energies of Co-rich compounds that was introduced in our previous work is presented in more detail in the present work. For the description of the σ phase, the magnetic part of the total energy is included in the description of the pure Co end-member compound resulting in a non-linear description of the magnetic contribution over composition. The calculated phase diagram obtained from the present CALPHAD description is in good agreement with the experimental data. The metastable FCC-L12 phase diagram was calculated and compared with experimental data.

Thermodynamic analysis of the topologically close packed σ phase in the Co-Cr system.

Density functional theory (DFT) calculations show that it is essential to consider the magnetic contribution to the total energy for the end-members of the σ phase. A more straightforward method to use the DFT results in a CALPHAD (Calculation of phase diagrams) description has been applied in the present work. It was found that only the results from DFT calculations considering spin-polarization are necessary to obtain a reliable description of the σ phase. The benefits of this method are: the DFT calculation work can be reduced and the CALPHAD description of the magnetic contribution is more reliable. A revised thermodynamic description of the Co-Cr system is presented which gives improved agreement with experimental phase boundary data for the σ phase.

Thermodynamic assessment of the Co-Ta system.

The Co-Ta system has been reviewed and the thermodynamic description was re-assessed in the present work. DFT (density functional theory) calculations considering spin polarization were performed to obtain the energies for all end-member configurations of the C14, C15, C36 and µ phases for the evaluation of the Gibbs energies of these phases. The phase diagram calculated with the present description agrees well with the experimental and theoretical data. Considering the DFT results was essential for giving a better description of the µ phase at lower temperatures.

Prediction of seebeck coefficient for compounds without restriction to fixed stoichiometry: A machine learning approach.

The regression model-based tool is developed for predicting the Seebeck coefficient of crystalline materials in the temperature range from 300 K to 1000 K. The tool accounts for the single crystal versus polycrystalline nature of the compound, the production method, and properties of the constituent elements in the chemical formula. We introduce new descriptive features of crystalline materials relevant for the prediction the Seebeck coefficient. To address off-stoichiometry in materials, the predictive tool is trained on a mix of stoichiometric and nonstoichiometric materials. The tool is implemented into a web application (http://info.eecs.northwestern.edu/SeebeckCoefficientPredictor) to assist field scientists in the discovery of novel thermoelectric materials. © 2017 Wiley Periodicals, Inc.