ASM Online Member Community

Expand all | Collapse all

Density of austenite vs. nickel fraction

  • 1.  Density of austenite vs. nickel fraction

    Posted 05-05-2021 15:17

    Does anyone know of a paper that demonstrates a plot or mathematical model of the density of austenitic steel as a function of nickel content at STP? 

    I am working on a paper where I need to know the density of austenitic steel at standard temperature and pressure (STP) as a function of nickel content

    The question seems at a glance like it should be trivial, but it is not that simple.  Nickel atoms are slightly heavier and larger than iron.  As nickel replaces iron in austenite, the FCC unit cell gets heavier but also gets more spacious.  The increasing mass is probably linear with nickel content, while the increasing volume is probably not linear w/[Ni].  Further complicating the question is that austenite with very low [Ni] is unstable below the transition temperature curve, while the data I need must be at atmospheric temperature and pressure.  So simple, yet so complex! 

    I made a plot (below) of data taken from x-ray diffraction PDF cards from the 04-XXX-XXXX subfile, about 28 cards total of Fe-Ni alloys of point group 225 at STP, and the scatter is huge regardless of a linear or nonlinear best-fit curve.  I suspect some of the cards were not measured at STP, confused mass% with mole%, or contained other unspecified elements in more than trace amounts.  I am hoping to find something a little more definitive than this, but I have had no luck finding any such paper.

    Richard Chinn
    National Energy Technology Laboratory

    Richard Chinn
    Research Materials Engineer
    National Energy Technology Lab
    Albany OR
    (541) 967-5986

  • 2.  RE: Density of austenite vs. nickel fraction

    Posted 05-06-2021 03:22
    Do you care what is stabilizing the Austenite? In a simple Fe-Ni system you would not have Austenite at low Ni concentrations. Some old work I did at the Bureau of Mines showed a somewhat linear relationship for the simple pure Fe-Ni system. The density was actually fairly constant from 0 to ~ 20 wt pt Ni. Presumably this was all bcc, above this the structure was more likely fcc. I didn't have X-ray diffraction done to see what phases were present. This was for some old school wear testing so the densities would have been calculated from wt./volume of the sample. The densities were not reported in, but this is from "Influence of Composition on Abrasive Wear for Several Binary Systems", and one of system studied was the Fe-Ni system. 

    Joseph Tylczak
    Albany OR
    (541) 928-2193

  • 3.  RE: Density of austenite vs. nickel fraction

    Posted 05-07-2021 19:52
    That is interesting data; I didn't even think that Richard would get an answer.  He obviously came to the right place!  Some comments:
    You point out some aspects of the original question, related to Austenitic steel.  I found an old reference, footnoted, showing the Fe-Ni binary system.  The way I interpret it (not a guarantee of accuracy!) is that the field is ferritic-austenitic-mixed and not going to be fully austentic at equilibrium at room temperature, probably, though the curves start being dashed below 500 degrees C and about 8% Ni, the limit shown for the fully alpha field.  At equilibrium below 50 w/o Ni there should be some alpha present if I use the lever rule.

    There is a lot of research on Fe-Ni due to its being present in meteoric rocks and I recall statements being made about how the Fe-Ni system is particularly sluggish in diffusion, such that true equilibrium in some parts is unlikely to be reached within the current age of the universe....  (!)

    Also, Richard's statement of "austenitic steels" naturally limits the data to less than 50% Ni, beyond which value the material is not technically a steel any more.  Perhaps using 'Fe-Ni binary' would be better.

    That phase diagram also suggests that there is a gamma prime at the nickel rich side, looks like maybe a peritectic (forgive inaccuracies in terminology, it's been a LONG time) at 500 C and 77 w/o Ni.  The lack of any data below 500 C suggests that answers below this temperature can't be found within the period of a budgetary calendar year or residence time of a grad student thesis, whichever is greater.  Then, to the sluggishness of formation of a true equilibrium, add the ordering that takes place because of the Curie temperature.  Though I do not know if this causes a change in the unit cell/equilibrium atomic spacing, or just an energy absorption/release from the ordering.  Is there also a martensitic / BCT chance of formation?  If samples were quenched from the fully gamma field could there be effects?

    So overall, using a 100% FCC cell as the assumption might be a poor one and this might be the source of some of the non-linearity of Richard's source data, as well as the other possible causes he cites.  I am not experienced with XRD methods to know how the samples would have been prepared or what the detection limits for alternate crystals would be.

    (reference used was " Determination of the iron rich portion of the iron - nickel - carbon phase diagram", auth. Romig, Alton D., 1977.  Found in the Lehigh Preserve repository.  This paper included a referenced source for the Fe-Ni diagram, "J. I. Goldstein and R. E.- Ogilvie: A Re-Evaluation of the Iron-Rich Portion of the Fe-Ni System, Trans. TMS-AIME, 233, (1965), pp. 2083-2887. ")

    Paul Tibbals, P.E.

  • 4.  RE: Density of austenite vs. nickel fraction

    Posted 05-06-2021 10:36
    I suspect this might be more nearly linear with atomic % (atom fraction).
    Using X-ray lattice parameter measurements for the pure elements (yes, you need one for gamma iron, but it has been published) you can write a program (or use a spreadsheet) to get your answer. 
    That will ignore effects including:
    • volume changes associated with the "Invar effect"
    • interstitial elements
    • changes in vacancy (and self interstitial) contents - these are very small
    • possible ordering effects at some compositions
    Experimentally measured densities (from weight in air and weight in water) are not especially accurate. A very small air bubble can reduce the calculated density by more than 0.01 g.cm3. Because the errors are not going to be normally distributed, simple averaging may not give you the best answer.
    X-ray lattice parameter measurements are far more accurate. 
    If you've done the X-ray density calculation carefully, those numbers may be more accurate than experimental densities.

    John Grubb
    New Kensington PA
    (724) 448-5272

  • 5.  RE: Density of austenite vs. nickel fraction

    Posted 05-06-2021 16:39
    Richard is this plot for a single composition of steel, or a collection of values for different steels of different chemistries?   if you collect data for several different austenitic steels of different Ni content, I don't think you would expect to get a smooth a curve because the density also depends on other elements not being considered on the x-axis. This is something that may be possible to calculate using a CALPHAD database for steels which could be used to calculate the density variation as a function of chemistry taking into consideration variation in the other alloying elements. 

    Paul Mason
    Thermo Calc Software Incorporated
    McMurray PA