I am new to free lime determination, and have been doing a bit of background reading. I have a question for you. I was looking at a brochure from Thermo for equipment to carry out determination of free lime (CaO) in clinkers by combined XRD and XRF. I include the brochure in the link. In their figure 1 it shows the peak for a series of samples with varying CaO. There are problems with this diagram (scale, spelling) which I shall ignore for the moment, but when I looked at my own cement data I noticed that the there is potential problem with the 100% 200 peak from CaO (2.405A) being overlapped by the 18% 221 (2.403A) and 13% 112 (2.410A) lines of C2S. Now the intensities are much lower in C2S, but the proportion of that phase in the mix is vastly greater than CaO. I would have thought that this would create a problem with this type of calibration. They show the peak resolved from C3S - I presume this is the triclinic polymorph - but ignore any mention of C2S. Any comments on this? Does anyone have any XRD data of clinker with a known amount of free lime that I could measure?
https://www.thermoscientific.com/content/dam/tfs/ATG/CAD/CAD%20Documents/Application%20%26%20Technical%20Notes/Cement,%20Coal,%20Minerals%20Sampling%20and%20Online%20Analysis/Cement%20Online%20Elemental%20Analyzers/Free-Lime-Determination-Clinker.pdf
Hello!
I am new to free lime determination as well, but have been helping Proto develop an instrument with software which is able to automatically conduct the analysis you are asking about. I can provide you with some sample data.
A few comments that I have:
- Even if peaks belonging to multiple phases overlap, a program such as Jade can allow you to fit the peaks for all phases and then extract peak information, assuming you know which phases are contributing. By fitting the peaks of all phases present, you can extract the peak information for free lime only.
- If you choose to attempt Rietveld (which I believe could help you determine which phases are present and in approximately which quantities), bear in mind that C2S itself can have multiple polymorphs, and that will need to be considered. see more here: https://books.google.com/books?id=1BOETtwi7mMC&pg=PA16&lpg=PA16&dq=c2s+xrd&source=bl&ots=6Yw8MC2Uwz&sig=X3QSa0JrdaP8B0tbEBGyIULnwrQ&hl=en&sa=X&ved=0ahUKEwictqjlyJXKAhWD2SYKHTVLA5gQ6AEINDAD#v=onepage&q=c2s%20xrd&f=false
- I was able to find a paper which seems directly related to your question. Perhaps this can help you: http://link.springer.com/article/10.1631%2Fjzus.A1300215
If you are interested in data, please contact me and I will find what I have that may help you. Good luck!
Hello Lindsay,
Thank you very much for the links. Yes I am using quantitative Rietveld method, but with TOPAS rather than Jade. I am very familiar with this approach in the field of natural geological materials. I have also looked at a lot of concrete so I am familiar with the polymorphs, but this is for clinkers which is new. We only have access to lab X-rays rather than synchrotron, this is for commercial purposes so to an extent it needs to be quick and dirty.
If you have some data with free lime then yes I would be very interested, thanks! I will message you with my email address.
Thank you Fabrizio but my samples are bulk powders of clinkers and cements, so SAED ad TEM are not appropriate for this problem. I have no access to neutron or synchrotron diffraction - this is for routine commercial work, rather than a research project, so costs are at a minimum. Yes the two patterns are different, but in this region there is a peak overlap which raised my original query (whether 1% CaO would be 'seen' over 15% C2S at 2.40A).
I have had a helpful response from Thermo to my original question - here is what they said.
"Thanks for the comments on the figure itself; I guess some editing errors were introduced when the figure was processed for publishing. The scale is meant to display 2.48 instead of 2.58.
You are also correct that C2S has an overlap on free lime, but I would compute the actual effect of the overlap in a slightly different way.
In your original example, if C2S is constant at 15%, then its contribution to Free lime overlap is constant. Under that constant assumption, it is mostly irrelevant if C2S has a peak overlap contribution equivalent 1% 2% or 5% Lime. As C2S does not change, it gets eliminated in the “offset at origin” in the calibration curve for Free lime.
What is more interesting is the effect of a change of C2S concentration, say from 15% to 10%. I ran some quick simulations below. Intensity unit axis is arbitrary.
1) sample with 1% lime + 15% C2S has a peak of 8500
2) sample with 0% lime + 15% C2S has a peak of 3000
3) sample with 1% lime + 10% C2S has a peak of 7800
Roughly, sensitivity of 1% lime is (8500-3000)/1 = 5500
Overlap contribution of 1% C2S is (8500-7800)/5 = 140
So, with a change of 1% in C2S, ignoring the C2S overlap contribution will cause an error on Free lime reading of around (140/5500) = 0.025 %
A change of 5% in C2S -> 0.12% Lime error.
See some graphs below."
-------------
Additionally I had a very helpful response from Lindsay Young who has supplied me with some test data that I can use with this technique. It seems that lab XRD and careful Rietveld analysis is sufficient to solve the problem.
-------------
Hi Ian,
We have the same ARL 9900 in my previous laboratory (College of Materials Science and Engineering, Nanjing Tech University), and I am one of the users of this machine. Some of my publications are based on ARL 9900 (X. Li, et al. Industrial & Engineering Chemistry Research, 2014 (53): 1954-1964 ).
There is some experience from me on this machine might helpful to you.
-This machine is OK for the determination of the f-CaO content in clinker. See Figure 1 in my paper [X. Li, et al. Industrial & Engineering Chemistry Research, 2014 (53): 1954-1964 ]. For commercial purpose, I think the precision is good enough.
-The overlap of f-CaO and beta-C2S is not the problem if we apply the Rietveld analysis. The techniques for the ARL 9900 is similar to Rietveld but different.
- Advantages: automatic sample loading with 12 sample stages; XRD and XRF double out put.
-Problems for the ARL 9900 (mainly caused by the sampling with pressure):
(a). For normal OPC clinker, there is no problems. If the SO3 content is very high, i.e. 2.0% in the clinker, then there will be very strong preferred orientation for the C3S as the pressure applied to the sample. Luckily, such high SO3 content rarely occurred in the production line. In my case, I want to research the SO3 impact on the clinker phases, then failed.
(b). For cement, there will problem from gypsum which is very easy to be preferred orientated during sampling.
(c). This machine is not good for hydrates. With the pressed sample, the quantification of portlandite, Aft and AFm phase are problematic.
Good luck!
Xuerun
Article Stability of Tricalcium Silicate and Other Primary Phases in...