Volume fraction of multiple phases

AI Thread Summary
The discussion centers on a final year project examining the effects of heat treatment on the IN718 alloy, specifically the challenges in identifying multiple phases using XRD scans, which only reveal matrix peaks. The user is exploring the use of SEM images and software like ImageJ to determine volume fractions of at least four phases, but faces limitations with programs that typically identify only two phases based on grayscale. There is a suggestion to produce cross sections to analyze potential compositional differences between the sample's surface and interior. The user anticipates an increase in delta and gamma prime phases while expecting a decrease in the laves phase over time. Overall, the conversation highlights the difficulties in phase identification and the need for suitable analytical tools.
jblakes
Messages
15
Reaction score
0
Afternoon all,

Hopefully somebody can help me, I'm doing my final year project and it's looking at the effect of heat treatment on in17, when I run an XRD scan I found that I all the phases sort of hid behind the matrix and so can't really make them out.

So I've been looking at using the SEM images I have and using imagen or similar to find out VF, problem is I have at least Four phases within the material. Does anybody know a program that you can work out the four phases? Most seem to only allow two as it just identifies white/grey not the different shapes of grey.

Kind regards
James
 
Engineering news on Phys.org
Are you able to produce a series of cross sections in the material? Does the surface of your sample have a different composition than the interior?

Might find something here.
 
Last edited:
Yeah I can do, I currently have a sample that's 20x40mm.

The treatment is for homonegisation of the material so should be the same, over time I am expecting to see an increase in delta phases and gamma prime and a decrease in laves phase.

I have tried XRD but for some reason only showed the matrix peaks even though there should be 15/20% of at least two other phase. Gonna try edsd and hope that picks it up.

Anybody used imageJ? Can you find more than two phases with the colour changes?

Regards James
 
Back
Top