Dear Janki Patel,

This data should pose no problem for PhaseID. For further analysis, I would discard the angular range below 7 degrees 2Theta. There is no information in there for QPA and it will just make it harder to determine a good background line.

You can zoom in on the "good/essential" range and then use the function "Treatment => Clip Range" in HighScore to get rid of the artifact.

Best Regards,

Thomas

Hi Thomas Degen Thank you for explaining how to clip the data, I was wondering if that is possible, I have one follow up question, but is it normal to get multiple minerals which could fit the data? Everyone has told me I should do XRD to accurately determine the minerals within the sample, but I still end up with multiple minerals that could fit the data after executing a search match

Hi Janki Patel, XRPD is no silver bullet to solve all mineral-related PhaseID problems. But it comes pretty close often. Phases with totally different chemistry can produce the same pattern or the other way, Carbon can occur as Graphite or Diamond, 2 distinct patterns. Also, minor phases can be hard to spot sometimes. Then indeed some mineralogy knowledge comes in handy to decide what would be the solution that makes the most sense (is most probable). Also having the information about elemental composition helps to make the right choice between multiple well-fitting patterns. Can you give an example of which pattern you have to choose? Usually, you chose the least amount of pattern to fit most of your sample Peaks.

HI Thomas Degen thank you for explaining that, it seems there really are no silver bullets to solve all my problems, but alas. I have used the images and EDS spectra gathered from SEM to help me narrow down potential minerals, but in regards to some samples, even from SEM-EDS, I cannot quite figure out which would be the best minerals that suit the phases as the search-match outputs a large range of potential matches, and many of them could theoretically fit.

Hi Janki Patel, f you can share the XRD data (and Chemistry, if known) I can try to make an educated guess

Hey Thomas Degen Thank you so much, I would really appreciate it. I've uploaded the Xrdml file without any corrections I have made and the HPF file where I attempted to treat the background and determine peaks and a profile fitting. As far as chemistry goes the sample contains mainly Si, O, Al, and then slightly lower amounts of K, Ca, Na, Fe, Mg, Cl in varying quantities depending on the mineral. My issue is from the images I took on SEM I was unable to determine any of the minerals or even narrow them down to possibilities.

Please see the attachment, these 3 phases seem to explain the majority of peaks well and are valid looking at your given chemistry. I will have another look tomorrow and then send you the HPF document

Thomas Degen Wow, thank you so much. I appreciate the help a lot.

Hi Janki Patel, see attached my final guess. You should try to make a better measurement with better signal to noise. Either measure longer or add up some scans. In the attached HPF I used the COD as a Database and I exported/downgraded the HPF to V3.x. Let me know if you can open this document in your HighScore(Plus) version.

Thank you so much Thomas Degen for all your help! and I was able to open the file up, thank you for that. I really appreciate it.

Hi Janki,

Search/Match has never worked for me properly, especially when you have different kinds of clays with little peaks in your XRD pattern. Your XRD measurement seems fine to me. I see some clays (around 10 degrees), albite (around 28 degrees) and quartz (around 26.5 degrees), etc. in your sample XRD pattern. Then, you need to know where your endmembers peaks appear in the pattern.

Also, the procedure I usually follow before search and match is as follows (it is based on Bruker manual):

1. Background noise caused by various factors, e.g., inelastic scattering, electron noise, etc [Suevalov and Kaplan, 2005] were removed using the "background subtraction" function from EVA software.

2. Kα2 were removed using the "strip kalpha2" function.

3. Displace all the peaks using the "displacement" function based on the quartz peaks in the sample. This step is required because the software needs to be able to distinguish patterns and match the sample peaks with reference data/library.

4. Using “match/search” function I compared the dust sample data to a database of minerals to identify peaks in the sample.

The following Steps 5 and 6 are when you want to derive Semi-quantitative abundances using DIFFRAC.EVA: (they should be done after the 4 first steps)

5. The height of the largest peak of each mineral in the sample is matched with its corresponding standard height from the library.

6. Using the “S-Q” function, mineral percentages (abundances) are determined for the sample.

Hope this helps,

Hi Mohammad Reza Sadrian thank you for your advice, and I am so glad that I am not the only one who search match didn't work for properly. But those were some super helpful tips, I will try them out. Cheers.

Hi Janki,

I am glad to hear it. And yes, it is hard to trust search/match when you have too many little peaks (especially related to the clay minerals).

Good luck,

Hello all, for such a complex material, with so many peaks, it is absolutely crucial to achieve a decent signal to noise ratio.

Especially for the, mostly un-overlapped, low angle peaks. These peaks are most important for a PhaseID, there should be no doubt if a feature is a peak or just noise. The given scan was for example a 30-minute scan (reading from the timestamp), the highest low angle peak at 2Theta 9.8 having a net peak height of only 838 counts. To get the noise level down and positively identify all phases I would measure 5 to 10 times longer or measure 5 to 10 similar scans and then add them up.