As duplex stainless steel consists of austenite and ferrite, when it is cold deformed austenite transforms into martensite.
Martensite formed from austenite has BCT crystal structure while ferrite has BCC crystal structure.
Both this BCT and BCC give same reflections in xrd as BCT structure is very similar to BCC due to low carbon in alloy.
There might be several option of duplex alloy in the described case.
(1) The steel consists of martensite and austenite. One has bcc lattice, another one has fcc lattice. Peaks of each are easily distinguished.
(2) The steel consists of tempered martensite and of quenched martensite. The first has stable cubic lattice, the second has unstable tetragonal lattice. They also may be identified using one of the known XRD data bases.
(3) The steel consists of two cubic phases, one stable, and one strain induced. Then, if software for analyzing XRD patterns using Fourier filtering is applied, peaks of two phases can be split. Provided that stresses are elastic, shift of peaks of the strained phase from Bragg positions is calculated by the formula sigma = (E/nu)ctgTheta/Delta d(Theta), where E is Young modulus, nu is Poisson coefficient, Delta d(Theta) is shift of an XRD peak under elastic stress. Fraction of strain-induced martensitic phase can be estimated approximately, for example, by PowderCell software by the ratios of relative intensities of respective lines of each phase, provided that the analyzed material has no crystallographic texture.
By using voight profile, which is powerful technique to calculate crystallite size and lattice strain for single peak analysis
(i) First of all the Martensite that forms during deformation has BCC structure not BCT to my knowledge. Please correct me if i am wrong.
(ii) To determine martensite volume fraction, Compare the peaks before and after cold deformation. (Calculation can be done using peak intensity)
(iii) An easier tool for determining strain induced martensite is Ferritoscope a portable instrument.
There might be several option of duplex alloy in the described case.
(1) The steel consists of martensite and austenite. One has bcc lattice, another one has fcc lattice. Peaks of each are easily distinguished.
(2) The steel consists of tempered martensite and of quenched martensite. The first has stable cubic lattice, the second has unstable tetragonal lattice. They also may be identified using one of the known XRD data bases.
(3) The steel consists of two cubic phases, one stable, and one strain induced. Then, if software for analyzing XRD patterns using Fourier filtering is applied, peaks of two phases can be split. Provided that stresses are elastic, shift of peaks of the strained phase from Bragg positions is calculated by the formula sigma = (E/nu)ctgTheta/Delta d(Theta), where E is Young modulus, nu is Poisson coefficient, Delta d(Theta) is shift of an XRD peak under elastic stress. Fraction of strain-induced martensitic phase can be estimated approximately, for example, by PowderCell software by the ratios of relative intensities of respective lines of each phase, provided that the analyzed material has no crystallographic texture.
Alignment: In paragraph (3) two martensitic phases (bcc) with different unit cell parameters are meant
XRD peaks of fcc-Fe, and bcc/bct martensite (depending on carbon content) can be differentiated. To calculate the relative volume fraction, the ratio of 100% intense peak of fcc and bcc phases can be calculated. You can see my paper in the attachment where I have adopted such a procedure.
Ferritoscope may not reveal bcc/bct Fe if its volume fraction is low.
"By using voight profile, which is powerful technique to calculate crystallite size and lattice strain for single peak analysis" is wrong for several reasons:
a) it should be spelled Voigt
b) peaks of real materials are NOT Voigtians
c) you need several peaks to be able to correctly estimate the strain contribution from the trend with the diffraction order
Regarding the original question, the answer is not very easy. Unfortunately the difference between bcc and bct are really slight. Plus in some cases the specimen is not ideal (e.g. you see preferred orientation or you see the effects of heavy faulting). If the pattern is well behaving (you must be sure and try to make even a very rough pole scan to exclude the preferred orientation), then you can use the Rietveld method and cross your fingers. You better use a highly penetrating radiation (e.g. Mo) to be sure you sample a sufficiently large volume and you have a clean signal (if you use Cu radiation you can have problems of fluorescence)
E Borisenko i believe gave the most complimentary answer. At the best of my knowledge the new generation XR diffrectometers can conduct such kind of measurements and pick splits quite easily.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Material Characterization