We use microscopy technique and by Image analysis software we can find the phase fraction. 

In my case, the product is temper martensite, i.e. combination of ferrite and fine dispersed cementite. It is really difficult to identify the volume fraction of cementite from SEM micrographs. 

BTW, which image analysis software do you use? Is it possible to estimate the volume fraction of bright particles in the attached image?

Thank you

 If you don't have Image analysis software as far I know, you can use point count method to determine the volume fraction. But still I'm also not getting proper idea about it. If you get clear idea about it, please share it. 

Dear Ajeet Babu Kumar Parasumanna Sir, What is the Image analysis software to calculate volume fraction. Can I get it? Please give me your contact.

We use Olympus microscope and the software is their product. We will try to calculate the % and let you know .

Regards

Ajeet

Please, I'm waiting for reply..

Dear Dulal Chandra Saha 

I do not have a direct answer to your original question. Nevertheless I dare to make some remarks.

1. Dilatometry seems to be non-equivalent to quantitative microscopy.

2. Supposing that the Fe-C alloy under study can be approximated by a binary Fe-C alloy,the equilibrium fractions of the ferrite and cementite phases follows from the application of the balance rule for the Fe-C phase diagram.

3. The Bain transformation offers the possibility for the transformation from fcc --> bcc.

4. Measuring the martensitic tetragonal lattice parameters before tempering and comparing to the final ferrite lattice parameter allows to estimate the volume effect of the transformation from martensite to ferrite.

5. It seems reasonable to predict that the effect in one direction as measured during dilatometry will be of the order of one third of this total volume effect. 

6. Supposing that there is no residual austenite, the progress of the transformation from martensite to ferrite can be estimated by the dilatometry effect at a time t divided by the total effect as denoted under point 5.

7. Taking this fraction as the volume fraction of the ferrite formed, it seems reasonable in view of the assumption under point 5 that the balance is the volume fraction of the cementite formed.

8. The ulitmate check of this approach lies in the correspondence with the data obtained from the Fe-C phase diagram after completed transformation.

Dear Dr.Van Mourik, 

I am not able to understand you remark( The Bold one) "It seems reasonable to predict that the effect in one direction as measured during dilatometry will be of the order of one third of this total volume effect.". Request you to elaborate further for better understanding. Thank you.

Regards

Ajeet

Dear Ajeet Babu Kumar Parasumanna 

Assuming random orientation of the grains in the dilatometry specimen and assuming that the dilatometry measurement is only in one direction (the test specimen is assumed being a  bar), then  it is reasonable to predict that the relative length effect as measured during dilatometry is about one third of the relative volume effect.

Re Image analysis

Online you will find the ImageJ software (freeware), this is the descendant of NIH-Image developed in the US, by a team at their health institute (originally it was for Macs not PCs). After a learning curve you will be able to do a lot of what the commercial packages can do. However, the commerical packages have considered a lot of the 'thorny' issues involved in such work and made their software user friendly. You will just have to work a bit harder. The key tasks are thresholding and segmenting the image to reveal the metrics you desire. Often a combinatorial approach is required (e.g. you may use a 'roundness filter' to select/de-select certain particles in the image). John C Russ's book 'The Image Processing Handbook' is very good primer , well it is much more than a primer, more like the bible on this topic. Also John used to sell  some software (via 3rd party vendors e.g. www.reindeergraphics.com , which was very good value, not bank busting, i.e. the Image Processing Tool Kit, IPTK. I think his son ran the QA support forum, very helpful guy. I've used these packages on my desktop PC, but we have both the Zeiss AxioVision, and the Kontron KS400 Image analysis systems for this kind of work when linked to dedicated optical microscopes.

A bit of a tongue in cheek answer, but if the structure is fully tempered, put the steel's composition in ThermoCalc, MTDATA or FactSage and use the equlibrium volume fraction predictions for these phases.

As to what you might observe on the dilatometer during the tempering reaction, take a look at this article and its references.

Metallurgical and Materials Transactions A

March 2009, Volume 40, Issue 3, pp 551-559

Date: 28 Jan 2009

Microstructural and Dilatational Changes during Tempering and Tempering Kinetics in Martensitic Medium-Carbon Steel

Minsu Jung, 

Seok-Jae Lee, 

Young-Kook Lee

http://rd.springer.com/article/10.1007/s11661-008-9756-2

For one view on how to do the dilatation calculations/interpretation look here:-

http://phase.yonsei.ac.kr/accom_data/Tempering%20Kinetics%20of%20S45C%20Martensitic%20Steel.pdf

Also here

http://www.journalamme.org/papers_vol46_1/4611.pdf

My usual advice:- read a few articles first, and then make your own judgement on what the authors are saying and which method you should adopt.

Dilatometric effects during tempering of martensite (when considered separately from independent information) in no way allow you to estimate the cementite fraction. Indeed: 1) These effects are visible as carbon atoms leave the martensite lattice by precipitating in some carbides. However, depending on tempering T and duration, there may be various types- so called transient carbides (like epsilon), the cementite of interest, carbides of alloy elements. Moreover, they may transform each to other, when competeing for carbon. 2) Transformation of retained austenite also takes place giving opposite (volume increse)  dilatometric effect. 3) Some of carbides are practically invisible (too fine for optic microscopy and only randomly found by very local TEM technique). What to do ? Employ a set of methods and preexisting knowledge on certain steel and processing schemes.  BTW, an above advice to consider the equilibrium Fe-C diagram is certainly naïve, since the usual tempering still does not remove unequilibrium state of martensite. Otherwise the latter would get ferrite with a low dislocation density, that is not the case.

Hi Dulal,

I suggest you see this paper: http://www.sciencedirect.com/science/article/pii/S0040603113005972