Resolving the Ti-V Phase Diagram Discrepancy with First-Principles Calculations and Bayesian Learning

2025-06-21

AI Summary

Long-standing experimental controversy surrounds the existence of a miscibility gap in the body-centered cubic (BCC) phase of Ti–V binary alloys. To resolve this, we propose an *ab initio*–machine learning hybrid framework that requires no experimental prior: it integrates first-principles calculations, active-learning-driven moment tensor potential construction, and Bayesian thermodynamic inference to construct the free energy surface with quantified uncertainty across the full composition range. Our results conclusively demonstrate a genuine BCC miscibility gap, with a critical point at *T* = 980 K and *c*_V = 0.67. Crucially, we identify oxygen impurities as the primary cause of spurious “pseudo-gaps” observed in prior experiments—thereby correcting the erroneous CALPHAD assessment of complete mutual solubility. This work not only resolves a decades-old controversy but also establishes a generalizable, impurity-aware paradigm for predicting thermodynamically sensitive phase transformations.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Constraint Satisfaction and Optimization: Mixed Discrete/Continuous Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Web Mining and Content Analysis: Web data provenance, reliability, and authenticity
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Conflicting experiments disagree on whether the titanium-vanadium (Ti-V) binary alloy exhibits a body-centred cubic (BCC) miscibility gap or remains completely soluble. A leading hypothesis attributes the miscibility gap to oxygen contamination during alloy preparation. To resolve this controversy, we use an ab initio + machine-learning workflow that couples an actively-trained Moment Tensor Potential to Bayesian thermodynamic inference. Using this workflow, we obtain Ti-V binary system across the entire composition range, together with confidence intervals in the thermodynamic limit. The resulting diagram reproduces all experimental features, demonstrating the robustness of our approach, and clearly favors the variant with a BCC miscibility gap terminating at T = 980 K and c = 0.67. Because oxygen was excluded from simulations, the gap cannot be attributed to impurity effects, contradicting recent CALPHAD reassessments.

Problem

• Resolve Ti-V phase diagram discrepancy using first-principles calculations
• Determine if Ti-V alloy has BCC miscibility gap or complete solubility
• Assess impact of oxygen contamination on Ti-V alloy phase behavior

Innovation

• First-principles calculations resolve Ti-V phase discrepancy
• Bayesian learning enhances thermodynamic inference accuracy
• Machine-learning potential excludes oxygen impurity effects