ferrite, cementite, austenite and martensite are enough to describe phase constituent of a vast number of steels. But has there been any extensive research on partition coefficient of impurities and alloying elements in these phases, by a two-by- two basis ?(e.g. ferrite and cementite in plain carbon steel, ferrite and martensite in martensitic steel, ferrite and austenite in duplex stainless steel) What would be the possible crystallographic sites of substitution (and possibly interstitial solid solution) can be inferred for the metal and ceramic crystal structure upon knowing this issue.
There are a few things to distinguish here, i.e.
1) full equilibrium partitioning among all phases (and making sure that all phases present are really known and not overlooked, for instance because they are very small), considering full thermodynamic equilibration for all of the elements (viz the fast ones such as C, N, H etc and the slow ones such as Mn and Si etc) - this does not happen very often. Most low temperature equilibrium partitioning coefficients are extremely difficult to measure.
2) equilibrium partitioning among phases only for fast diffusers; rest same as above
3) partial equilibrium partitioning only at the interface, but decaying concentrations far away from the interface, due to kinetics.
4) equilibrium partitioning NOT only among the two (or more) phases, but also among the interface itself and the adjacent phases (the interface is a thermodynamic trapping site of its own, due to the adsorption isotherm, and may elements in steel also partition to them (and not only to the phases)).
All of the resulting partitioning coefficients depend substantially on kinetics and composition of course.
Some of them can be obtained directly from multi-component phase diagrams (full equilibrium required, is practically never the case in steels) and some not (kinetics, then u have to use dictra of phase field etc).
In all of these cases exact values for multi-component partitioning coefficients can experimentally usually only be obtained by using atom probe tomography.
Taking partitioning coefficients from the concentration profiles very close across the hetero-interfaces is particularly helpful as the atoms can maintain local equilibrium (equal chem. potential across the interface) with only a few atomic diffus. jumps - when probing far away from the interface partitioning coefficients are usually wrong, due to kinetics.
Good luck.
Here are some references, but this is a wide and really complex field…
http://www.dierk-raabe.com/atom-probe-tomography-on-steels/
http://www.dierk-raabe.com/qp-quench-partitioning-steels/
http://www.dierk-raabe.com/medium-mn-steels/
some example pictures are also attached, just to give a rough impression
etc
No. The question doesn't arise.For alloys chemical composition, phases and their stoichiometry matter.
There are a few things to distinguish here, i.e.
1) full equilibrium partitioning among all phases (and making sure that all phases present are really known and not overlooked, for instance because they are very small), considering full thermodynamic equilibration for all of the elements (viz the fast ones such as C, N, H etc and the slow ones such as Mn and Si etc) - this does not happen very often. Most low temperature equilibrium partitioning coefficients are extremely difficult to measure.
2) equilibrium partitioning among phases only for fast diffusers; rest same as above
3) partial equilibrium partitioning only at the interface, but decaying concentrations far away from the interface, due to kinetics.
4) equilibrium partitioning NOT only among the two (or more) phases, but also among the interface itself and the adjacent phases (the interface is a thermodynamic trapping site of its own, due to the adsorption isotherm, and may elements in steel also partition to them (and not only to the phases)).
All of the resulting partitioning coefficients depend substantially on kinetics and composition of course.
Some of them can be obtained directly from multi-component phase diagrams (full equilibrium required, is practically never the case in steels) and some not (kinetics, then u have to use dictra of phase field etc).
In all of these cases exact values for multi-component partitioning coefficients can experimentally usually only be obtained by using atom probe tomography.
Taking partitioning coefficients from the concentration profiles very close across the hetero-interfaces is particularly helpful as the atoms can maintain local equilibrium (equal chem. potential across the interface) with only a few atomic diffus. jumps - when probing far away from the interface partitioning coefficients are usually wrong, due to kinetics.
Good luck.
Here are some references, but this is a wide and really complex field…
http://www.dierk-raabe.com/atom-probe-tomography-on-steels/
http://www.dierk-raabe.com/qp-quench-partitioning-steels/
http://www.dierk-raabe.com/medium-mn-steels/
some example pictures are also attached, just to give a rough impression
etc
Adding the following to Dierk Raabe:
-consider searches like "https://scholar.google.de/scholar?hl=de&as_sdt=0%2C5&q=partitioning+cementite&btnG="
-Generally there are different types of works:
(i) experimental works by spatially resolved techniques by to some extent also X-ray diffraction. As indicated atom probe tomography is the method (like all methods with its specific problems) with highest spatial resolution which can simultanesously statistiscally relevant element compositions. Note that states of different degrees of chemical equilibrium can be attained
(ii) theory based works: using DFT methods you can construct "cemenite" in very funny compositions (Au3C, Fe2AuC...) and you can obtain structure and energy parameters. The main task is to draw appropriate conclusions (e.g. partition coefficients) for experimentally relevant situations from the structure and energy characteristics from DFT. This may be more tricky than doing the actual DFT calculations.
(iii) studies on the kinetics of carbide etc formation which may draw conclusions on partitioning from the kinetics of phase formation (which phase develops at which rate). This is of course indirect, but the conclusions, e.g. with respect to the effect of Si are very correct.
As it concerns specifically cementite: in 3d metals, if you go to the left (lower atomic number) you will stabilise cementite, e.g. the elements will partition (if they can kinetically, diffusion!) into cementite from alpha (or gamma). This is especially true for Mn and Cr, whereas form Cr on (V, Ti) there is the trend to form Sonderkarbide (carbides of other structures than cementite). The corresponding elements generally form themselves stable carbides. If you go to the right, you will gradually destabilise the cementite and the alloying elements will deplete in the cementite (if they can). Al and Si as main group elements do the same and actually more extremely (these elements are hardly detectable in cementite e.g. by APT). Especially Si has a stable carbide but its chemical bonding is incompatible with that of most carbides (including cementite), which can lead to peculiar carbide structures (if they are allowed to form, which does not happen in steel). This is an opportunity to advertise an own work (Article The iron silicocarbide in cast irons revisited
). Hence Si is an efficient tool to suppress carbide formation.As it concerns crystallographic sites for substitution: first try metal by metal and non-metal by non-metal. Cementite has two Fe sites, one "seeing" more C than the other. This is the site, which is according to DFT works (as much as I remember), the C-loving elements left from Fe prefer. Upon a quick look there is apparently no experimental evidence for metal ordering in ternary cementite ( e.g. none in the neutron work
Transactions of the Metallurgical Society of AIME (1966) 236, p. 1342-1346). In a Ni-Pd-P phase with cementite structure, order has, however, been encountered, which is however not easily interpreted in terms of group number (https://www.scientific.net/MSF.443-444.353.pdf).
A lot on Fe3C can also be found in https://www.tandfonline.com/doi/pdf/10.1080/09506608.2018.1560984?needAccess=true .
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials Chemistry