HighScore Plus Tutorial - Phase Identification X-ray Diffraction - Long Version - JIAM Diffraction

Summary

In this instructional video, Michael Koehler, lab manager at the JIAM Diffraction Facility, demonstrates how to utilize HighScore Plus software for phase identification in a sample. He guides viewers through loading data, background determination, peak marking, and refining the search for matching diffraction patterns. The tutorial emphasizes optimizing parameters for background adjustment, manually correcting peak identification errors, and narrowing down potential matches using chemical composition filters. The goal is to accurately match sample peaks with known diffraction patterns, concluding with phase identification.

Takeaways

• 📚 Michael Koehler introduces himself as the lab manager for the JIAM Diffraction Facility at the University of Tennessee, Knoxville.
• 📹 The video is a tutorial on using HighScore Plus software for phase identification in samples.
• 🔍 HighScore Plus supports various file types for data analysis, including XRDML, which was used in the demonstration.
• 🛠 The first step in phase identification is to load a data file and determine the background of the data, which can be done automatically or manually.
• 📉 The background determination involves adjusting parameters like 'bending factor' and 'granularity' to accurately represent the data's baseline.
• 🔍 'Search Peaks' is a feature that helps identify and differentiate peaks from background noise, refining the data for better analysis.
• 👀 It's important to review and correct the peak list generated by the software to ensure accuracy in phase identification.
• 📊 The 'Default Profile Fit' function helps improve the fit between the calculated pattern and the actual data, enhancing peak accuracy.
• 🔎 'Search Match' is used to find possible matches for the identified peaks against a database of patterns, with the results sorted by a scoring system.
• ⚙️ Refining search parameters, such as subfiles, chemistry, and crystallography, can significantly reduce the number of patterns to search through, speeding up the process.
• 🗂 The final step is to match the peaks to the correct phases by evaluating the candidate list and ensuring that all peaks have corresponding matches in the selected patterns.

Q & A

Who is the speaker in the video and what is his role?

The speaker is Michael Koehler, and he is the lab manager for the JIAM Diffraction Facility at the University of Tennessee, Knoxville.

What is the purpose of the video?

The video's purpose is to demonstrate how to use HighScore Plus software for phase identification in a sample.

What is the first step in phase identification using HighScore Plus according to the video?

The first step in phase identification is to load a data file into HighScore Plus.

What file types are compatible with HighScore Plus?

HighScore Plus can accommodate different types of files, including XRDML files, which are used when data is collected using a Panalytical instrument.

How does one determine the background in HighScore Plus?

The background can be determined in HighScore Plus by going to 'Treatment' and selecting 'Determine Background.' Options include 'Automatic,' 'Manual,' and 'By Search Peaks.'

What is the significance of the 'Bending Factor' and 'Granularity' in determining the background?

The 'Bending Factor' determines how much the background curve bends up into the peak, while 'Granularity' changes the distance between points of inflection in the background curve, affecting its ability to bend.

Why might one choose not to subtract the background in HighScore Plus?

Subtracting the background can lead to negative intensity values, which Rietveld refinement does not handle well, and it may also obscure certain data points that are important for analysis.

What is the purpose of the 'Search Peaks' function in HighScore Plus?

The 'Search Peaks' function helps identify and distinguish peaks from noise in the data, allowing the software to better understand and analyze the sample's diffraction pattern.

How can the 'Search Match' feature in HighScore Plus assist in phase identification?

The 'Search Match' feature compares the identified peaks with a database of patterns to find potential matches, which can then be evaluated to determine the phases present in the sample.

What is the importance of refining search parameters in the 'Search Match' process?

Refining search parameters helps to narrow down the number of patterns that need to be searched through, making the matching process more efficient and targeted to the specific characteristics of the sample.

How does the 'Chemistry' tab in HighScore Plus help in narrowing down potential matches?

The 'Chemistry' tab allows the user to specify the elements believed to be present in the material and to ignore patterns that do not contain these elements, significantly reducing the number of potential matches.

What does the 'Score' column in the 'Search Match' results represent?

The 'Score' column represents the probability of a match between the pattern and the sample's peaks, with higher scores indicating a better match, although it does not guarantee the presence of the phase in the sample.

How can one verify if a potential match from the 'Search Match' is accurate?

Verification is done by clicking on the potential match and observing if the peaks in the diffraction pattern align with the peaks in the sample's data curve.

Outlines

00:00

🧪 Introduction to HighScore Plus for Phase Identification

Michael Koehler, the lab manager at the JIAM Diffraction Facility, introduces a tutorial on using HighScore Plus software for phase identification in samples. He mentions other videos for different purposes like phase quantification and crystallite size determination. The process begins with loading an XRDML data file from a Panalytical instrument, highlighting the software's compatibility with various file types. The focus then shifts to determining the background of the data, with options for automatic, manual, and search peak methods, and adjusting parameters like bending factor and granularity to refine the background curve.

05:01

🔍 Background and Peak Identification Techniques

The tutorial continues with techniques for refining the background curve and identifying peaks in the data. The presenter explains the implications of adjusting the bending factor and granularity, and demonstrates how to accept or modify the automatically determined background. He also discusses the option to subtract the background, but advises against it due to issues with negative intensity values and Rietveld refinement. The 'Search Peaks' feature is introduced to help identify true peaks versus noise, with a walkthrough of correcting any mistakes in peak identification and using the zoom functionality to scrutinize peak markers.

10:03

📊 Refining Peak Markers and Automatic Profile Fit

This section delves into refining peak markers and using the automatic profile fit to improve the match between the calculated pattern and the actual data. The presenter shows how to insert missing peaks and remove incorrect ones, ensuring that all actual peaks are marked. He also explains how to use the difference plot to assess the fit between the red data curve and the blue calculated pattern. After demonstrating the 'Default Profile Fit' to reduce differences and improve peak accuracy, the tutorial moves on to phase identification, including using the 'Search Match' feature and editing restriction sets to narrow down potential matches.

15:07

🔬 Phase Identification and Search Parameter Refinement

The final part of the tutorial focuses on phase identification using the 'Search Match' feature. The presenter discusses the importance of refining search parameters to reduce the number of patterns to search through, using criteria such as material type, quality, and crystallography. He also explains how to use the 'Chemistry' tab to specify elements present in the sample to further narrow down the search. After performing the search, the presenter reviews the candidate list, emphasizing the importance of the score column and how to verify matches by comparing the diffraction pattern peaks with the sample data. The tutorial concludes with a demonstration of how to confirm phase matches and complete phase identification.

📝 Conclusion and Invitation for Further Inquiry

Michael Koehler concludes the video with a summary of the phase identification process using HighScore Plus and invites viewers to explore additional tutorials linked in the video description. He encourages viewers to ask questions in the comment section if they have any, ensuring a supportive learning environment. The presenter expresses his hope for a good day for the viewers, wrapping up the tutorial on a positive and engaging note.

Keywords

💡HighScore Plus

HighScore Plus is a software application used for the analysis of X-ray diffraction (XRD) data, which is essential for identifying and characterizing the phases in a sample. In the video, it is the primary tool demonstrated for phase identification, with various features such as background determination and peak search explained through its use.

💡Phase Identification

Phase identification is the process of determining the different phases present in a sample based on their X-ray diffraction patterns. It is the central theme of the video, where the script describes the steps to identify these phases using HighScore Plus, including loading data files and analyzing peak patterns.

💡XRDML File

An XRDML file is a type of data file format used in X-ray diffraction studies. In the script, it is mentioned as the file type used by the presenter, collected from a Panalytical instrument, and opened in HighScore Plus to begin the phase identification process.

💡Background Determination

Background determination is a crucial step in XRD data analysis where the software identifies and subtracts the background signal from the raw data to better visualize and analyze the peaks. The video script explains different methods of background determination within HighScore Plus, such as 'Automatic', 'Manual', and 'By Search Peaks'.

💡Bending Factor

The bending factor is a parameter in HighScore Plus that influences the curvature of the background line in the XRD data. It is used to refine the background determination process, as described in the script, to ensure that the background line accurately represents the baseline without interfering with the peaks.

💡Granularity

Granularity, in the context of the script, refers to the resolution setting for the background determination in HighScore Plus. It affects the spacing between points of inflection in the background curve, allowing for a more precise fit to the data, as explained by the presenter.

💡Peak Search

Peak search is a feature in HighScore Plus that helps identify and mark the peaks in the XRD data. The script describes how to use this feature to distinguish peaks from noise and to correct any misidentifications, which is vital for accurate phase identification.

💡Rietveld Refinement

Rietveld refinement is a method used in crystallography to refine the profile of XRD data to improve the fit between the observed and calculated patterns. Although not the main focus of the video, the script mentions that subtracting the background can lead to negative intensities, which Rietveld refinement does not accommodate.

💡Search Match

Search match is a function in HighScore Plus that compares the peaks in the sample's XRD data with a database of known diffraction patterns to identify the phases present. The script explains how to use search match effectively by refining search parameters and interpreting the results.

💡PDF-4+ Database

The PDF-4+ database is a comprehensive collection of diffraction patterns used for phase identification in XRD analysis. The script mentions this database as the source for the search match function in HighScore Plus, highlighting its importance for identifying phases based on their diffraction patterns.

💡Chemistry Tab

The Chemistry tab in HighScore Plus is a feature that allows users to specify the elements they expect to find in their sample, which helps narrow down the search for matching diffraction patterns. The script describes how to use the Chemistry tab to refine the search match process by focusing on patterns containing specific elements.

Highlights

Introduction to HighScore Plus software for phase identification in materials science.

Explanation of the JIAM Diffraction Facility and its role in the University of Tennessee, Knoxville.

Overview of different HighScore Plus functionalities for phase quantification and crystallite size determination.

Step-by-step guide on loading a data file for phase identification.

Demonstration of the background determination process in HighScore Plus.

Description of 'Automatic', 'Manual', and 'By Search Peaks' background determination methods.

Adjustment of background parameters like bending factor and granularity for data refinement.

Importance of not subtracting the background to avoid negative intensity issues in Rietveld refinement.

Utilization of 'Search Peaks' feature to identify and correct peak markers in the data.

Techniques for refining the search for peaks and correcting any misidentified data points.

Use of 'Default Profile Fit' to improve the fit between calculated and actual data patterns.

Process of refining search parameters to narrow down potential matches in the PDF-4+ database.

Explanation of how to use the 'Chemistry' tab to filter patterns based on elemental composition.

Method of selecting 'at least one of' or 'none of' elements to further refine the search for phase matches.

Discussion on the significance of the score column in the search match results.

Visualization of how to match diffraction pattern peaks with sample peaks for phase identification.

Finalization of phase identification by confirming all peaks have been matched to their respective phases.

Conclusion of the tutorial with a reminder of available resources and an invitation for questions.

Transcripts

Hello, my name is Michael Koehler, and I am the  lab manager for the JIAM Diffraction Facility,  

located at the University of Tennessee, Knoxville.  This video will cover how to use HighScore Plus to  

identify the phases in the sample. I have a number  of other videos that show and explain how to use  

HighScore Plus for different purposes, such as  phase quantification and the determination of  

crystallite size and micro strain, and the links  of these videos can be found in the description  

below. The first step in phase identification is  to load a data file. To do this, we need to go to  

"File," "Open," here we see an XRDML file, that's  because I used a Panalytical instrument in order  

to collect this data. If you have a different  type of file, don't worry, HighScore Plus can  

accommodate different types of files. Let's open  it, and we see our data in red. Now, the first  

thing we want to do is determine the background,  and we can do that in a couple of different ways.  

We can go to "Treatment," and under treatment, we  see a list of different ways that we can treat the  

data, one of which is determine background. In  order to save clicks, I just like to right click  

determine background. Here we see three options,  "Automatic," "Manual," and "By Search Peaks." I  

don't really use "By Search Peaks." "Manual" I  use rarely, that's if the background is rather  

complicated. But most of the time, "Automatic"  is just fine. We see that we have two parameters  

here that we can change to try to improve our  background, and here we see the background is  

represented as the green line. Let me turn this  down to start. Now, bending factor is exactly  

what it sounds like, it determines how bendy your  background is. So as we increase, we see that the  

green curve begins bending up into the peak  more and more. That is not something we want.

Granularity changes the distance between  points of inflection in the background curve,  

so a smaller granularity means smaller  distance between inflection points. So if  

we were to increase, it's no longer able to bend  up into the peak because the distance from one  

point of inflection to the next is too great. So  typically, I like to go around 24 granularity and  

then play with bending factor to see what looks  best. Let me get this out of the way too. Now,  

in this example, the bending factors all  look pretty similar, so I'll just leave it  

somewhere around seven. You'll notice that you  have the option to subtract your background,  

which will then make the data curve more flat  and closer to the x-axis. I don't like to do  

this for a couple of reasons, one of which  is because when you subtract the background,  

any data points beneath the green line turn into  a negative intensity and Rietveld refinement  

does not like that, so that's the main reason.  Another reason I will touch on here shortly,  

but for now I will just accept the background as  we have it, and we see that the background has  

changed to a dark green line. The next step is  not really required but it is nice when you're  

trying to do phase identification. If we right  click and go to "search peaks," we can start to  

try to tell the software what is a peak and what  isn't a peak, or we'll let the software determine  

that. We have some parameters here that we  could change in order to help the software  

figure out what is a peak and what's just noise.  I don't usually change these, I'll just "Search  

Peaks." We see a bunch of data markers along  the top, so a lot of peaks have obviously been  

found. So for now, I will just accept. We  see the peak list now here under peak list.

And we will want to go through and make sure that  the software didn't make any mistakes, it didn't  

mark something as a peak that's not or maybe it  just missed a peak. Right off the bat, I see that  

there is an error here at low angles. It's marking  this as a peak, but that's really just background  

noise. If we want to get rid of that, we put our  cursor over the data marker and hit delete on the  

keyboard, and now it's gone. We'll right-click,  zoom back out, and now I will go through and just  

make sure that no other mistakes have been made.  Now, a lot of these peaks are rather small, so we  

want to zoom in a little bit, but you'll notice  that when we try to zoom, it doesn't really allow  

us to zoom according to that zoom box that we're  drawing. It's making sure that these peaks remain  

fully in the window. If you want to avoid that,  you can just click this button, zoom intensity, it  

is likely somewhere else on your screen than here;  this is a highly customizable program, so just  

look for that button. But now if we zoom, it will  zoom in according to the box we draw. So let's  

start to scroll to the right, look at how well  the peaks are marked. A couple of items of note,  

this blue curve is the calculated pattern, so  that's where the computer thinks there are peaks.  

Red again is your data. You see that we have  one peak marker here, but it looks like two.  

If you have these sharp peaks that are about  half the intensity of the main peak, then that  

is likely caused by the two different wavelengths  of radiation. This is a K alpha 1 peak, this is K  

alpha 2. All you really need to have marked is  one peak for alpha 1, so it's okay that there's  

not a second data marker for this one. Obviously,  the calculated and the data patterns don't match  

up all that well. That's okay for now, we will  fix that shortly. We just want to make sure that  

all of the peaks are actually marked. Here we  come into one problem. So this peak is marked,  

this one is not, and this is very obviously a  peak. If we right-click, we can "Insert Peak" and  

then tell the computer that there is a peak right  there. And now that blue line even matches up  

better with that red line. If you want to get rid  of that option of inserting a peak, all you do is  

right click, "Insert Peak" again, and it's gone.  So let's continue on, make sure that no other  

mistakes are found. Here, this might be a peak but  it might not, it's kind of noisy so it's hard to  

tell. This most likely is a peak, so we will just  insert one there. Right-click, "Insert" again.

This might be a peak, hard to tell, so  I'll just leave it off for now. Now,  

if we zoom back out, we see here at the bottom  a difference plot. If you don't see it there,  

you can always just right-click, show graphics,  and then choose difference plot. But we see that a  

lot of the peaks don't match up, as we saw when we  zoomed in. What we get down here is the difference  

between the red curve and the blue curve. If we  want to improve this fit, all we do is go up here,  

make sure we're in automatic mode, and then  choose "Default Profile Fit," and you'll see  

those differences really start to go away.  These peaks are much better fit now, peak  

locations are more accurate. We still have some  spikes of difference here around the tall peaks,  

but that's perfectly normal. They are still very  nicely fit. So now that we have all of our peaks  

identified as peaks, we can go in and try to  determine what phases those peaks belong to. So  

if we right click, we can go to "Search Match,"  and then we want to edit a restriction set.

The first thing I'm going to point out  is that we have resulting hits of four  

hundred and twelve thousand patterns. So that  means if we closed and searched right away,  

we would have to search through four  hundred and twelve thousand patterns to  

find which ones match best with our peaks. Now  we use the PDF-4+ database, the 2019 version,  

so it's fully up-to-date; that's why we have so  many patterns. Now, that takes a little while,  

that's a lot of patterns to search through, so  we want to begin refining our search parameters  

so that we can search through fewer patterns.  One way we can do this is the Subfiles tab. If  

we know that we have a ceramic material and  all of our phases are ceramic, we can choose  

that. That cuts down the possible patterns to  about 17,000. If we know we had a carbohydrate,  

that decreases the number of searchable patterns  to 657, so already we're getting rid of a lot  

of possibilities and that will help us in  our search. You can refine searches based  

on quality. Star quality is your highest, so if  you only want those patterns, you can click that.  

We see that it greatly decreased our possible  patterns. I don't usually do that because you  

might miss something important. Same thing for  "skip marked as deleted by ICDD," that might be  

good a vast majority of the time, but I just like  to be able to see all possible patterns. If you  

know that your sample was...or if your data were  collected at ambient pressure and temperature, you  

can click these to avoid any patterns that were  collected at non-ambient pressure and temperature.

Crystallography...you can put in  crystallography search parameters.  

Strings... you can search based off of  compound name or mineral name, formula,  

mineral, or zeolite class. A lot of these  are useful, but I don't really use them.

Subfiles I use occasionally if I'm dealing  with something more difficult like a cement  

or a mineral. Most of the time, I just use  this Chemistry tab, and here we can tell it  

what elements we believe we have in our material  and to ignore all other elements. For instance,  

when everything is gray like this, it means  that everything is possible, so we are searching  

through every possible pattern. If we start  clicking some, one click brings us to this bluish  

color, so that means only patterns that have at  least carbon and/or oxygen have to be searched,  

so that reduces it by quite a bit. If we make them  green, then we will only search through patterns  

that have both carbon AND oxygen. If we made them  red, then we would search through all patterns  

except for those containing the carbon and oxygen.  So in this case, I know that I made a sample that  

had manganese oxide in it and silicon oxide in it,  so I will just choose these three as possibilities  

or "at least one of," and I will add the rest to  "none of." So now we will ignore anything in red,  

any pattern that has an element in red, and only  search through patterns that have manganese,  

silicon, or oxygen in them. And we see that we  have 1,428 patterns, so I will close and search.  

Sometimes this is really quick, sometimes it's a  little slow. We do see here "elements from XRF."  

Panalytical makes XRF instruments, so you can  test it with that, import the results, and then  

use that to help you in your search match, but we  see that the search match is done. So we need to  

make sure that we say "OK," and here we have a  list of possible matches. This is our candidate  

list. First thing to point out is that we have  this score column, and it automatically sorts it  

based off of this column. A high score means that  it's a high probability that this pattern matches  

what's in your sample, but just because it's a  high pattern does not mean that it's guaranteed to  

be in your sample; you need to check and make sure  that it really is. We do that by clicking on the  

pattern that we're interested in, and we go over  here and we see these little lines appear beneath  

our markers or beneath our read data curve, and  that represents the peaks that are contained in  

this diffraction pattern. We see that, sure  enough, each peak that is in that diffraction  

pattern matches up with a peak in our sample. This  one looks like it's not, like we're missing it,  

but we see that the intensity is so low, we  wouldn't really expect to see it in our pattern.  

Here, this line extends all the way up and matches  about what's in our pattern. So this one it's not  

a big deal if we don't see it; if we didn't see  this peak, that would be a problem because it's  

one of the main peaks of that diffraction pattern.  Now, if we click it, left click, and drag up, we  

notice that the markers have now turned dark blue,  but we also noticed that these little V markers  

on top of the data markers or the peak markers  have disappeared, but only those corresponding  

to the peaks of silicon dioxide. So one reason  that search peaks is such a nice feature is that  

it helps us more quickly determine what peaks we  still have to find a match for. We see that after  

we accepted this as a candidate, it reorganized  this list, and now manganese oxide best matches  

the peaks that remain. If we left-click, hold that  down, and then drag it up, we now see that those  

peaks correspond to green, which we see here,  and now all of our data markers...the V markers  

on top have disappeared. So now we know that all  of our peaks have been matched, these are the two  

phases that are in our material, and we have  completed phase identification. Now that does  

conclude this video. As a reminder, if you would  like to learn more about using HighScore Plus,  

links to my other tutorials can be found in the  description below. If you have any questions,  

please feel free to ask in the comment section.  Thank you, and I hope you have a good day!