Site hosted by Oak Ridge National Laboratory / Disclaimers / Contact the webmaster  (Bruce Warmack)

Isotopic Diffusion Databases for Magnesium Integrated Computational Materials Engineering (Mg-ICME)

DiffusionInterdiffusionGB DiffusionModelingTheoryCommunicationsLiterature


The objective of this project is to create an isotopic (tracer) diffusion database in the Mg-rich phases of the Mg-Al-Zn-Mn system. This database and a thermodynamic database that is being continuously updated will be provided to participants involved in various tasks in the Mg-ICME program. This website represents efforts in progress toward gaining critical diffusion data.

Principal Investigator: Nagraj Kulkarni, Oak Ridge National Laboratory
    (865) 576-0592; e-mail:

Industrial Partners: U.S. Automotive Materials Partnership Integrated Computational Materials Engineering (ICME) Team, Magnesium Elektron North America



Oak Ridge National Laboratory
   Bruce Warmack, (865) 574-6202; e-mail:
     Balasubramaniam Radhakrishnan, (865) 241-3861;
     Ethan Ambroziak [gb diffusion in thin films, summer 2012]
     Peter Todd [retired]
Argonne National Laboratory
     John Mundy [retired]
Virginia Tech
     Jerry Hunter, e-mail:
       Jay Tuggle
University of Central Florida
     Yongho Sohn,
       Catherine Kammerer
     Kevin Coffey,
     Edward Dein
University of Newcastle, Australia
    Graeme Murch, e-mail:
     Irina Belova, e-mail:

Industrial Partner:

Magnesium Elektron North America
    Bruce Davis


The approach for measuring the tracer diffusion coefficient is based on the thin-film approach. The procedure first requires the preparation of homogeneous single phase alloy samples in the desired Mg-Al-Zn-Mn system. This is followed by deposition of stable isotopes or radioisotopes of these elements in the form of thin films on the sample surfaces. After diffusion annealing at various temperatures below the melting temperatures of these alloy samples, the isotopic diffusion depth profiles in these samples are measured using SIMS. Analysis of the diffusion depth profile data using the thin-film solution provides the tracer diffusivity for the selected sample composition at the annealing temperature. At higher temperatures the tracer diffusion in polycrystalline alloy samples is likely to be dominated by volume diffusion, while at lower temperatures there will likely be an additional contribution from grain boundary diffusion. The SIMS diffusion data will be analyzed to extract both types of diffusion contributions, though in this FY we have focused on volume diffusion measurements in large-grained samples.  By repeating such measurements for different compositions and temperatures, a significant amount of tracer diffusion data for Mg and Zn in the single phase Mg-Al-Mn-Zn system is obtained. Since Al and Mn are monoisotopic elements, their tracer diffusivities will be computed indirectly using diffusion theory (Darken-Manning relations) that connects interdiffusion coefficients (obtained from diffusion couples) with tracer diffusion coefficients and thermodynamics. The collection of tracer diffusion data for all the components in the Mg-Al-Zn-Mn system will be then fitted using suitable functions to generate the tracer diffusion database.


  • Obtained Mg self-diffusivities in pure polycrystalline Mg samples using our SIMS-based thin-film stable-isotope technique, validating and extending historic radiotracer measurements to lower temperatures.
  • Synthesized a matrix of Mg-alloy compositions in the Mg-Al-Zn system for tracer diffusion and interdiffusion studies, and confirmed their composition uniformity across representative cross-sections.  Mg and Zn tracer diffusion measurements for select compositions within this matrix are in progress.
  • Developed a superior annealing system and procedure based on the Shewmon-Rhines approach that (a) minimizes Mg and Zn vapor phase loss during diffusion annealing of Mg-alloy samples, and (b) allows measurement of precise sample temperatures to enable full numerical corrections associated with heat-up and cool-down times during diffusion annealing.
  • Optimizing SIMS parameters (Cameca ims7f-Geo system) for accurate isotopic diffusion depth profiles in Mg alloys.
  • Applied Molecular Dynamics (MD) simulations for determining grain boundary diffusivities for select bicrystal grain boundaries in Mg.
  • Conducted interdiffusion studies in the Mg-Al and Mg-Zn systems using solid-to-solid diffusion couples that were annealed at various temperatures and times.
  • Designed and constructed an Ultra-High Vacuum (UHV) sputter-deposition system for high-purity (very low oxygen) deposition of stable isotopes.

Future Direction

  • Measure Mg and Zn tracer diffusivities in the Mg-Al-Mn and Mg-Al-Zn-Mn systems, similar to measurements carried out in Mg-Al-Zn. Using diffusion couples, determine interdiffusion coefficients in these systems, and use diffusion theory to extract the tracer diffusivities of Al and Mn in these alloys.
  • Obtain tracer diffusivities of Mg and X (X = rare earths: Nd, Ce) as a function of composition and temperature in Mg-Al-X rare earth systems at various temperatures (150°C-400°C).
  • Grain-boundary diffusion experiments in magnesium alloys.


U.S. Department of Energy Assistant Secretary for Energy Efficiency and Renewable Energy Office of Vehicle Technologies as part of the Automotive Lightweight Materials Program under contract DE-AC05-00OR22725 with UT-Battelle, LLC.

Special thanks

John Allison and Bob McCune: Mg-ICME Program
Carol Schutte (Team Lead) and William Joost (Materials Engineer): Vehicle Technologies Program, DOE
Joe Carpenter: Former program manager, Automotive Lightweight Materials Program, DOE
Phil Sklad, Dave Warren: Automotive Lightweight Materials Program, ORNL

Example of diffusion at 350°C for one hour of 25Mg tracer isotope into a single Mg grain. Nonlinear fitting of the data is used to measure the diffusion coefficient.

Arrhenius plot of non-linear fit data shown at left, including Shewmon's 1954 radio-
nuclide results on polycrystalline samples. The straight line is a fit to the entire data set.
For details, see Mg Self-Diffusion Summary.

·        last updated: 01/30/2013