Computational discovery of promising new n-type dopable ABX Zintl thermoelectric materials

Gorai, Prashun; Ganose, Alex M.; Faghaninia, Alireza; Jain, Anubhav; Stevanović, Vladan

doi:10.1039/d0mh00197j

Cited by 40 publications

(38 citation statements)

References 61 publications

(44 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One plausible route for finding new n-type dopable Zintl phases with high TE performance is by chemical replacements in structure prototype (CRISP). 19 Computational efforts in the past have utilized chemical substitution in known compounds to search for undiscovered phases. 20,21 In the last decade, various routes have been explored to search for as-yet-unknown materials for functional applications.…”

Section: Introductionmentioning

confidence: 99%

Doping by design: finding new n-type dopable ABX₄ Zintl phases for thermoelectrics

Stevanović

Ertekin

et al. 2020

J. Mater. Chem. A

Self Cite

View full text Add to dashboard Cite

show abstract

Section: Introductionmentioning

confidence: 99%

Doping by design: finding new n-type dopable ABX₄ Zintl phases for thermoelectrics

Stevanović

Ertekin

et al. 2020

J. Mater. Chem. A

Self Cite

View full text Add to dashboard Cite

show abstract

“…In this work, we build a hybrid CGCNN model to accurately predict the total energy of GS and HE structures by training the model on DFT total energy of ⇠16,500 ICSD structures from the NREL Materials Database 27 and ⇠11,000 hypothetical structures generated by the ionic substitution method. 32,33 The overall prediction accuracy of our hybrid model is at par with other CGCNN models (MAE = 0.04 eV/atom), with similar accuracy in predicting the total energy of GS and HE hypothetical structures. We demonstrate the model's capability to satisfactorily distinguish low-and higher-energy structures for a given composition.…”

Section: Introductionmentioning

confidence: 59%

“…We also leverage a dataset of ⇠11,000 hypothetical structures that were created by ionic substitutions in known prototype structures from the ICSD. 32,33 Upon ionic substitution, the decorated structures are relaxed and their total energy are calculated with DFT. The relaxed structures (as VASP POSCAR files) and the total energy are available through the GitHub repository.…”

Section: Data and Preparationmentioning

confidence: 99%

A Graph Neural Network for Predicting Energy and Stability of Known and Hypothetical Crystal Structures

Pandey¹,

Qu²,

Stevanović³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

The discovery of new inorganic materials in unexplored chemical spaces necessitates calculating total energy quickly and with sufficient accuracy. Machine learning models that provide such a capability for both ground-state (GS) and higher-energy structures would be instrumental in accelerating the screening for new materials over vast chemical spaces. Here, we develop a unique graph neural network model to accurately predict the total energy of both GS and higher-energy hypothetical structures. We use ~16,500 density functional theory calculated total energy from the NREL Materials Database and ~11,000 in-house generated hypothetical structures to train our model, thus making sure that the model is not biased towards either GS or higher-energy structures. We also demonstrate that our model satisfactorily ranks the structures in the correct order of their energies for a given composition. Furthermore, we present a thorough error analysis to explain several failure modes of the model, which highlights both prediction outliers and occasional inconsistencies in the training data. By peeling back layers of the neural network model, we are able to derive chemical trends by analyzing how the model represents learned structures and properties.

show abstract

“…This vast and relatively uncharted phase-space allows for not only opportunities to discover new Zintl phases, but also new TE materials with potentially high zT. To this end, Zintl phases with a layered structure have garnered attention from the thermoelectric community: the "1-2-2" family (e.g., EuZn 2 Sb 2 10 ) and "1-1-1" family (e.g., KSnSb 11 ) to name just a few. The "1-2-2" family [12][13][14] of layered Zintl 15 another new material with a layered structure, have shown moderate zT due to improved Seebeck coefficient in combination with the features inherited from the "parent" Yb 2 CdSb 2 structure.…”

Section: Introductionmentioning

confidence: 99%

New n-Type Zintl Phases for Thermoelectrics: Discovery, Structural Characterization, and Band Engineering of the Compounds A₂CdP₂ (A = Sr, Ba, Eu)

Balvanz

Baranets

et al. 2020

Chem. Mater.

Self Cite

View full text Add to dashboard Cite

Zintl phases, owing to their complex crystal structures and intricate chemical bonding, have recently been recognized as promising candidates for thermoelectric (TE) applications. Band engineering, including band convergence, has been shown to be an effective way to enhance the thermoelectric performance of such materials. In this work, a series of emerging TE materials, the isostructural Zintl phases with the general formula A2CdP2 (A = Sr, Ba, Eu), are presented for the first time. Their structures, established from single-crystal X-ray diffraction methods, show them to crystallize with the orthorhombic Yb2CdSb2 structure type, with first-principles calculations on phase stability confirming that Ba2CdP2 and Sr2CdP2 are thermodynamically stable. Computationally, it was found that both Ba2CdP2 and Sr2CdP2 have the potential to exhibit high n-type TE performance (0.6 and 0.7 relative to the n-type PbTe, a reference TE material). To optimize the TE performance, band engineering strategies, including isovalent substitution and cation mutations, were investigated. From the band engineering of Ba2CdP2 via isovalent substitution of Sr on a single Ba site, leading to the quaternary composition SrBaCdP2, it can be suggested that increasing the conduction band valley degeneracy is an effective way to improve the n-type TE performance by 3-fold. Moreover, first-principles defect calculations reveal that both Ba2CdP2 and SrBaCdP2 are n-type dopable, adding these compounds to a small list of rare n-type dopable Zintl phases. The band engineering strategies used in this work are equally applicable to other TE materials, either for optimization of existing TE materials or designing new materials with desired properties.

show abstract

Computational discovery of promising new n-type dopable ABX Zintl thermoelectric materials

Abstract:
We develop a chemical replacements in structure prototype (CRISP) approach for novel materials discovery with functional applications.

Cited by 40 publications

References 61 publications

Doping by design: finding new n-type dopable ABX₄ Zintl phases for thermoelectrics

Doping by design: finding new n-type dopable ABX₄ Zintl phases for thermoelectrics

A Graph Neural Network for Predicting Energy and Stability of Known and Hypothetical Crystal Structures

New n-Type Zintl Phases for Thermoelectrics: Discovery, Structural Characterization, and Band Engineering of the Compounds A₂CdP₂ (A = Sr, Ba, Eu)

Contact Info

Product

Resources

About

Computational discovery of promising new n-type dopable ABX Zintl thermoelectric materials

Abstract: We develop a chemical replacements in structure prototype (CRISP) approach for novel materials discovery with functional applications.

Cited by 40 publications

References 61 publications

Doping by design: finding new n-type dopable ABX4 Zintl phases for thermoelectrics

Doping by design: finding new n-type dopable ABX4 Zintl phases for thermoelectrics

A Graph Neural Network for Predicting Energy and Stability of Known and Hypothetical Crystal Structures

New n-Type Zintl Phases for Thermoelectrics: Discovery, Structural Characterization, and Band Engineering of the Compounds A2CdP2 (A = Sr, Ba, Eu)

Contact Info

Product

Resources

About

Abstract:
We develop a chemical replacements in structure prototype (CRISP) approach for novel materials discovery with functional applications.

Doping by design: finding new n-type dopable ABX₄ Zintl phases for thermoelectrics

Doping by design: finding new n-type dopable ABX₄ Zintl phases for thermoelectrics

New n-Type Zintl Phases for Thermoelectrics: Discovery, Structural Characterization, and Band Engineering of the Compounds A₂CdP₂ (A = Sr, Ba, Eu)