A Straightforward Gradient-Based Approach for High-Tc Superconductor Design: Leveraging Domain Knowledge via Adaptive Constraints

Akihiro Fujii; Anh Khoa Augustin Lu; Koji Shimizu; Satoshi Watanabe

← Back to search

A Straightforward Gradient-Based Approach for High-Tc Superconductor Design: Leveraging Domain Knowledge via Adaptive Constraints

Akihiro Fujii, Anh Khoa Augustin Lu, Koji Shimizu, Satoshi Watanabe

arXiv:2403.13627 · 2024-03-20

T31 citations

Active bibliographic source — not scientific approval

Bibliographic access preserves source history; it does not approve extracted materials or validate reported claims. Review warnings on each occurrence separately.

Abstract

Materials design aims to discover novel compounds with desired properties. However, prevailing strategies face critical trade-offs. Conventional element-substitution approaches readily and adaptively incorporate various domain knowledge but remain confined to a narrow search space. In contrast, deep generative models efficiently explore vast compositional landscapes, yet they struggle to flexibly integrate domain knowledge. To address these trade-offs, we propose a gradient-based material design framework that combines these strengths, offering both efficiency and adaptability. In our method, chemical compositions are optimised to achieve target properties by using property prediction models and their gradients. In order to seamlessly enforce diverse constraints, including those reflecting domain insights such as oxidation states, discretised compositional ratios, types of elements, and their abundance, we apply masks and employ a special loss function, namely the integer loss. Furthermore, we initialise the optimisation using promising candidates from existing dataset, effectively guiding the search away from unfavourable regions and thus helping to avoid poor solutions. Our approach demonstrates a more efficient exploration of superconductor candidates, uncovering candidate materials with higher critical temperature than conventional element-substitution and generative models. Importantly, it could propose new compositions beyond those found in existing databases, including new hydride superconductors absent from the training dataset but which share compositional similarities with materials found in literature. This synergy of domain knowledge and machine-learning-based scalability provides a robust foundation for rapid, adaptive, and comprehensive materials design for superconductors and beyond.

Source-reported materials — not catalogue approval

Formula	Reported Tc (K)	Pressure (GPa)	Type
CeFeAsF0.2O0.8 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	39.8	Pressure not reported	unknown
Ti0.1807Sm0.8193FeAsF0.2O0.8 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	57.08	Pressure not reported	unknown
GdFeAsF0.2O0.8 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	48.3	Pressure not reported	unknown
LaFeAsO1.0 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	13.4	Pressure not reported	unknown
Nd0.6077Gd0.3239Tb0.0684FeAsO Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	53.64	Pressure not reported	unknown
Sm0.65Th0.35FeAsO Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	52.9	Pressure not reported	unknown
SrFe2As2 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	17.3	Pressure not reported	unknown
Mg0.2089V0.0183La0.243Nd0.251Eu0.2014Dy0.0774Fe2As2 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	47.59	Pressure not reported	unknown
Ba0.66K0.34Fe2As2 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	38.6	Pressure not reported	unknown
Bi2CaSr2Cu2O8 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	81	Pressure not reported	unknown
Hf0.2381Pt0.5835Bi1.1784CaSr2Cu2O8 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	95.88	Pressure not reported	unknown
Bi1.2Pb0.8CaSr2Cu2O8 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	92.9	Pressure not reported	unknown
CeNiC2 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	3	Pressure not reported	unknown
N0.1514S0.1395V0.086Zr0.2171Rh0.3626Pr0.0434NiC2 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	20.91	Pressure not reported	unknown
Th0.1Y0.9NiC2 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	20.91	Pressure not reported	unknown
La0.9999Ce0.0001NiC2 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	25.24	Pressure not reported	unknown
La0.9999Ce0.0001NiC2 Archive — visibility unverified Source-occurrence policy only; no material identity or catalogue acceptance is inferred from the formula.	25.24	Pressure not reported	unknown