Abstract:authorVarious combinations of characteristic temperatures, such as the glass transition temperature, liquidus temperature, and crystallization temperature, have been proposed as predictions of the glass forming ability of metal alloys. We have used statistical approaches from machine learning to systematically explore a wide range of possible characteristic temperature functions for predicting glass forming ability in the form of critical casting diameter, Dmax. Both linear and non-linear models were used to learn on the largest database of Dmax values to date consisting of 747 compositions. We find that no combination of temperatures for features offers a better prediction of Dmax in a machine learning model than the temperatures themselves, and that regression models suffer from poor performance on standard machine learning metrics like root mean square error (minimum value of 3.3±0.1 mm for data with a standard deviation of 4.8 mm). Examination of the errors vs. database size suggest that a larger database may improve results, although a database significantly larger than that used here would likely be required. Shifting a focus from regression to categorization models learning from characteristic temperatures can be used to weakly distinguish glasses likely to be above vs. below our database's median Dmax value of 4.0 mm, with a mean F1 score of 0.77±0.02 for this categorization. The overall weak results on predicting Dmax suggests that critical cooling rate might be a better target for machine learning model prediction.

A Relationship Between Glass-Forming Ability and Reduced Glass Transition Temperature Near Eutectic Composition

Structural transformation and rheological processes in bulk metallic glass Zr41Ti14Cu12.5Ni10Be22.5

Glass transition temperature of bulk metallic glasses: A linear connection with the mixing enthalpy

Correlation between glass transition temperature and melting temperature in metallic glasses

Simulation Of Reduced Glass Transition Temperature Of Cu-Zr Alloys By Molecular Dynamics

Effects of configurational changes on electrical resistivity during glass-liquid transition of two bulk metal-alloy glasses

Glass-forming ability correlated with the liquid-liquid transition in Pd42.5Ni42.5P15 alloy

Eutectic growth kinetics as an indicator for glass-forming ability

Connecting glass-forming ability of binary mixtures of soft particles to equilibrium melting temperatures

On the formation of metallic glass

The effect of minor alloying on the glass forming ability and crystallization reaction of Ti32.8Zr30.2Cu9M5.3Be22.7 (M=Fe, Co, and Ni) bulk metallic glass

Non-isothermal crystallization kinetics and glass-forming ability of Ti41Zr25Be28Fe6 bulk metallic glass investigated by differential scanning calorimetry

Thermodynamics of Equilibrium and Partitionless Solidifications in Glass Forming Binary Eutectic Alloys

Unusual relation between glass-forming ability and thermal stability of high-entropy bulk metallic glasses

Exploration of characteristic temperature contributions to metallic glass forming ability

Quantifying the origin of metallic glass formation

Ultrastrong behaviors of a non-equiatomic Ti-Zr-Ni-Cu-Be high entropy metallic glass with ultrahigh glass forming ability

Effects of Alloying Elements on Crystallization Kinetics of Ti–Zr–Be Bulk Metallic Glass

Proposed Thermodynamic Method to Predict the Glass Formation of the Ternary Transition Metal Systems

Mg and Al mixed effects on thermal properties in aluminosilicate glasses

On low-temperature strength and tensile ductility of bulk metallic glass composites containing stable or shape memory β-Ti crystals