Distribution of Alumina in Aluminum

Significance Statement

Until in the recent past, the solubility of oxygen in aluminum has been considered to be less than 1 atom in 1035-1040 aluminum atoms. This is however equivalent to less than 1 atom of oxygen in the entire world supply of aluminum since the metal was extracted. On the contrary, it has been known that actual form of the oxygen in aluminum is entirely aluminum oxide. This paradox has then been taken to show that the oxide in the metal may not have occurred by chemical reaction between oxygen in solution and the metal, but could have been incorporated mechanically.

It is now known that the surface film of oxide on the liquid metal, arising from contact with air, is entrained into the bulk metal during casting. This entrainment process results in the form of double films that are very detrimental to the mechanical properties.  Similar populations of incorporated double oxide films, known as ‘bifilms’, can be observed in other consolidation processes including spray forming and powder metallurgy. Bifilms seem to be important: recent studies taking account of bifilms is beginning to be able, for the first time, to fully explain fracture in metals.

A number of experimental studies have been carried out in an attempt to establish the oxygen content in liquid aluminum saturated by alumina. Unfortunately, unreliable results have been reported so far owing to unclear experimental conditions. This could be attributed to the very low oxygen content in liquid aluminum that normally falls below analysis limits, or could be referenced to the fact that suspended alumina particles from other sources as described above lead to inaccurate analysis.

Youn-Bae Kang at the Pohang University of Science and Technology in South Korea in collaboration with John Campbell at University of Birmingham in the UK demonstrated that the solubility of oxygen in liquid aluminum (in terms of mole fraction)  is low, in the region of only 10-7 but is vastly greater than the 10-35 and 10-40 values as mentioned previously. Therefore, this non-trivial revision in oxygen solubility necessitates a re-examination of the current assumptions. It then becomes interesting to speculate on the form this newly discovered source of alumina will take and if it will have an impact on the existing alumina bifilm theory. Their study provides an examination of this speculation. Their research work is published in Metallurgical and Materials Transactions A.

Thermodynamic computation that will provide a precise estimate for the oxygen content is based on sophisticated solution thermodynamics and supporting experimental data. However, thermodynamic modeling of deoxidation equilibria is complex owing to the strong interaction between the deoxidizing agent and oxygen. The conventional method assumes random mixing between the elements implementing a large number of adjustable parameters, but normally fails to  explain the phenomena.

When liquid metal solidifies, the advancing solidification front  can take a cellular, planar, or dendritic morphology. In  these circumstances, with an almost pure aluminum with a small concentration of oxygen which is highly segregated, it is to be expected that the aluminum will freeze on an almost planar front. The oxygen will be rejected ahead of the front and will build up there forming a steady-state sideways displacement where its additional concentration will gradually lower the freezing point of the alloy.

The consequence would be the formation of a domed hexagonal array of cells, dimpled at the corners of the hexagons where the oxygen concentration increases to form alumina. The almost planar front, partially indented by the hexagonal cellular network of segregated regions would be the structure anticipated of a eutectic solidification where the amount of second phase is quite small.  The experimental confirmation of this structure, its fine alumina strings coalescing under the action of interfacial tension, mixed together with the normally present population of entrained alumina bifilms, is not expected to be easy.

Distribution of Alumina in Aluminum Prediction Based on Thermodynamic and Diffusion Analysis. Advances in Engineering

About the author

Dr. Youn-Bae Kang earned his Ph.D. in material science and engineering at Pohang University of Science and Technology (POSTECH), Republic of Korea in 2005. He then joined Centre for Research in Computational Thermochemistry (CRCT), Department of Chemical Engineering, Ecole Polytechnique de Montreal, Canada, as a post-doc, later a research associate.
He worked on solution thermodynamic modeling and thermodynamic database development for oxide, oxysulfide, and metallic phases, which have been integrated into a thermodynamic computing system FactSage.  In 2009, he joined Graduate Institute of Ferrous Technology (GIFT) at POSTECH as an assistant professor, and now he holds a position associate professor.  He had been a visiting researcher in Technical Research Laboratories, POSCO, Rep. of Korea in 2011.  His research interest is clean steel production, novel refining process, recycling of ferrous scrap, theoretical solution thermodynamics and modeling of various solutions, and application of computational thermodynamics and reaction kinetics to process control of metallurgy and alloy design.
He has authored 100 journal articles, 60 conference proceedings, and one book chapter.  He was honoured by a Young Scientist Award by the Korean Institute of Metals and Materials (KIMM), Marcus A. Grossman Young Author Award by ASM International, Leader of Future Technology Development by National Academy of Engineering of Korea, Best Poster Award by Calphad, for his contribution to the field of high temperature processing of steel/oxide systems with the aid of computational thermodynamics, kinetic analysis, and experimental techniques.

About the author

John Campbell

An Englishman of Scottish descent, educated at Cambridge, Sheffield and Birmingham as a physicist and metallurgical engineer, he has spent most of his life in foundries where he thinks he has made more castings than he has had hot dinners.

He developed the Cosworth Casting Process for producing Al alloy cylinder heads and blocks for Formula One racing engines, using counter-gravity filling of molds by electromagnetic pumps.  Later still, after 15 years as Professor of Casting Technology in the University of Birmingham, UK, he has worked for over ten years with his partner in the USA, John Grassi, to set up a new casting operation, Alotech Inc, helping to develop the new and exciting Ablation Casting Process. He gave the American Foundry Society annual Hoyte lecture in 2012 “Stop Pouring; Start Casting”.

His “COMPLETE CASTINGS HANDBOOK” published in 2011, revised 2015, is not light bed-time reading, but is a bargain for the determined and fearless reader with an open mind. The new casting concepts promise greatly improved quality together with reduced costs, targeting a revolution in the metallurgical and mechanical engineering worlds.


Youn-Bae Kang and John Campbell. Distribution of Alumina in Aluminum Prediction Based on Thermodynamic and Diffusion Analysis. Metallurgical and Materials Transactions A, 2017, Volume 48, Issue 6, pp 2697–2700


Go To Metallurgical and Materials Transactions A

Check Also

Directed Self-Assembly of Liquid-Crystalline Molecular Building Blocks for Sub-5 Nm Nanopatterning. Advances in Engineering

Directed Self-Assembly of Liquid-Crystalline Molecular Building Blocks for Sub-5 Nm Nanopatterning