The Pharma Lab Show: Triaging Powder Samples with the Progeny Handheld Raman 1064 nm and Other Tools

     

    Failure of GMP XRD method can be very costly. Can triaging of incoming material help avoid this?

    In this episode, Simon discussed:

    • Triaging incoming material as a sample suitability test
    • The Progeny 1064 nm handheld Raman analyzer and polarized light microscopy as triaging tools
    • General observations of sample properties for 'categorical' method data

    Welcome to the Pharma Lab Show live broadcast. 

     It's been some time since we were last together. We took a break for the holiday period and for the last COVID wave. But now we're back, and we're beginning our journey again through elements of method development. And today we're going to be talking about something that's really important; and that's the failure of methods, so OOS. And we're going to look into in particular how triaging and the assignment of categorical variables to your samples can help you avoid OOS situations of method failures. 

     Now, I mean people typically don't like to talk about method failures, but the reality is that all methods fail. But it really depends on how they fail. Whether this can be addressed through triaging and categorical variables or not. And I view the failures really in terms of two types. There's what I call in-method failures. And this is just the normal statistics in the method itself. So the system is fine, the sample is fine but just due to the accumulation of random statistics during the analysis, the method fails. So I call these in-method failures. And if your method is truly diagnostic, it will actually, it will fail due to in-method failures. And those are very predictable. Once you do the method development and the method design, you should be able to predict the frequency of these in-method failures. Just you do statistics; whether it's one in one tenth, one in a hundred, one in a thousand, however the method is actually set up. But then there are out-of-method failures. These are completely random and unpredictable failures. And these come from events outside, so potentially, something wrong with the instrument, or something is wrong with the sample. The sample is not representative and this should not be something that's released. And it's addressing those types of situations or out-of-method failures, that triaging and the assigning of categorical variables can really help you a lot. 

     So what is triaging and what are categorical variables? Well, triaging is essentially a pre-study that you do on your incoming materials using orthogonal techniques. So if you're developing a method that is X-ray diffraction method, let's say, then you might use Raman or polarized light microscopy or some other orthogonal technique to assign categorical variables. And a simple example of a categorical variable which we'll go through in a few minutes, is one of color or visual appearance of a sample. So as you assign these categorical variables through your triaging techniques, whether it's microscopy, Raman, whether it's assigning a color, whether it's looking at flow or particle size, these allow you to assign categorical variables. And these are variables that don't have necessarily a value. They could be red, blue, green, or flows easily, clumpy, sticky, things like these. So general observations you make of your sample while you're doing your normal sample preparation as you're building up the method itself. Then depending on the sophistication of the method, you can use these categorical variables in different ways. For simple methods, you can say that: if the sample is blue then we're not going to run this method at all; the sample has to flow easily during the sample preparation in order for it to be acceptable for the method. So, in some ways, this is like a sample suitability test. Like a system suitability test is being put in place to make sure the instrument is performing in the right way. You can have a sample suitability test, such as the sample must be blue, it has to flow easily in order for it to be acceptable for this particular method, it has to have a particular Raman signature, the PLM shows an average particle size or habit for the material and these can be input quality gates or sample suitability tests. Now for more sophisticated methods, you can actually build categorical variables inside the method itself. So as the method starts up, you will load in, this sample was blue, this sample was green, this was red, and the method will actually embark along different analytical pathways depending on the categorical variables that you've given it. 

     So with that introduction, I'd like to go through some demonstrations of that. And we'll start off looking at color and flow and we'll then move on to looking at some Raman methods and finally into some polarized light microscopy. So the first thing I'd like to look at 

    is this issue of color. Now, color is quite obvious to us when we have something to compare it to. So I'd just like to pull out a few samples, in which I'll look at the color and how they flow. And we can note these down as categorical variables. So this is the first sample that I'm going to make, the sample. I'm going to take it out of the holder. We'll put it onto the slide and we'll try and make some observations about how it flows, the appearance of it. So straight away, I can see it's sort of clumpy. It's a bit sticky too as we put it on the slide as I try and spread it out in a suitable way for the microscope. That basically is not happening. So this is an extremely clumpy sticky material, not showing good flow properties. And clearly the color is sort of a bright orange color. So let me put that one out of the way, clean the spatula. This one is very difficult to clean off, which tells you that the particle size is probably very small. Now we're going to our next sample, which we'll put down. And after cleaning the spatula to make sure that no cross contamination is possible or using different spatulas maybe. This material is somewhat different. As you see, it separates easily on the slide. No problems with flow, a little bit sticky on the sample holder, on the spatula. Nevertheless, it spreads out very easy. When you put it down, it disperses easily, spreads out, nice flow properties, a little sticky, it has very fine particles in appearance and it is white. Let me clean the spatula again. Now we come to the final sample. Find a microscope slide, and for this sample we use something slightly different. And then we put it on the microscope slide. This again flows easy, there's no clumping, it comes off, and as I move it around, as one of my colleagues calls these things, they're very crackly. So this has definitely got large grains in it. As I move it around, you can see some extremely large grains inside there. It's very crackly and the material almost appears like translucent as opposed to white. Now, with samples like these, it's relatively easy to tell the color difference, particularly with the orange sample on the top. But now with a lot of a APIs, they're all going to be sort of white color. And so then, you know, there's off-white and then there's yellow-white and gray-white. And so that can be a little confusing to assign all of those different categories based on color. So then something that's really advantageous at that point is being able to take a picture of the sample itself.  

     We'll have a look at these samples later on the polarized light microscope, but what I want to introduce right now is the Raman unit. And the beauty of this handheld Raman is it has a built in camera as well. So as we're going through the samples, we can take pictures, we can store those pictures online in the limb system and then go back and analyze them for color. If we believe that color is truly a diagnostic categorical variable. Where the Raman really comes in useful, is if you have, you know, as you're making multiple samples, you want to make sure that you're making the same type of APIs, is it a hydrate, anhydrous, a solvate or such like or in this case here, I've got two vials of material that I'd extracted earlier. And now, of course, I can't remember. You know, at my advanced age, my memory is terrible. I can't remember which one is which. And the handheld Raman I really think is one of those triage tools that pretty much should be in every lab. Again, whether you're doing solid form screening, we have a lot of materials incoming outgoing; it's a very simple, quick, easy test, very powerful for checking similarity between materials whether what you're making today is the same or different from material from before and the same type of polymorph. So I just loaded in that vial straight into the holder. I can set up a scan. I'm going to do analysis this time because I've really forgotten completely what it is. If I'm trying to look for the same this type of material that I can do an ID test instead. And we just arm the laser, and set the scan running. So while it's doing this ID test, I'll just move that across to the side, out of the way. And what I want to do very quickly is we'll make a transmission sample, we'll go back over that again. We had a number of requests to show how to make these samples in a little more detail. And for the Raman unit we have a holder, especially for transmission X-ray samples. So the trick that we found for making these samples very robust is to use some of the parylene. So again let me just bring it over the camera. So I just stretched the parylene around the edge of the holder. It's actually a little trickier with gloves than it is with your with your bare hands. And in this case parylene has just snapped, but that's no problem. We can just move the parylene around the outside, and this allows you to get a much tighter fit when you make the other samples themselves. So let's put some parylene on the back and we stretch it around the holder. So this is just to give us a tighter fit then we'll put the films over the top. And again these are specially-cut films for the size of our sample holders. So they're very easy to make, very little clean up at the end. Just position one film on the top. We'll put some of the sample in the middle. So we'll just sprinkle some sample in the middle here. Again, you don't really need to spread it out too much. Making the sample itself will do that for you. And then we put the top film on, and then we put the plug through the center with the parylene, and that'll keep it very tight. Push it through. And then, as we hold it down, the rings will break off, and gives us very little clean-up and a very clean sample itself. So here is an example again of a transmission X-ray holder. And we're going to run that through the Raman system. Let's move that sideways. So now we've got a match. This is trehalose dihydrate. Material clearly identified by looking at the correlation coefficient. And we just turn that sideways. So you can read the screen there where the correlation coefficient is displayed and how well this sample matches, test diagnostic sample of trehalose dihydrate. And it's very easy to add to the databases on these machines. So if you're working with new materials and you're doing polymorph screening or solid-form screening, when you make something new, you can just add them to the database very easily.  

     Now after we've measured that material, we can say, well I need a picture. Let me load a picture along with that. Move that out of the way. Background is a little cluttered. So then I can bring that in to auto focus, and once they get the positioning right, I can then just take a picture of that and it'll add that and attach it to the record. This has a full audit trail on there. So when you upload it wirelessly to the limb system, it will have all the date, time, whoever did the measurements and then add the pictures into the Raman spectra. And you can also export the Raman spectra for further analysis if you want to do that. Now what I'll do is I'll take off the bottle measuring attachment. I'm going to put on another attachment here. And this is the one that's designed specially to do the X-ray transmission sample holders. It's very easy. I just clip on the nozzle on the end. and I bring over the sample I've just made, and we just click it into place. It's as easy as that. And we can do another measurement. Scan, we'll do analysis again, arm laser, scan. So I'm going to move that out of the way while it's doing its analysis. And now we'll take these samples across, the ones that we put on the slide here. Let's have a look at these on the PLM, and see what we can learn a little bit more about these materials. So here's the first sample I'm putting on. And, so you can see from  this one. It's pretty much as our visual interpretation gave us. It's very clumpy with extraordinarily fine particulates inside there. You can see the measured diagnostic on the bottom right hand side, that's set to 50 microns. Usually for X-ray diffraction, we want to keep it around about 40 microns as like a mean particle size, maybe between 10 and 40 is probably ideal. These particles are teeny weeny, I mean they are really tiny. And I doubt we're going to get any response to the polarizer at all, but it's very small. So that's told us a lot about this material. It's  sort of clumpy. Extraordinarily fine particle size, slightly orange in color. And again, if this is the material you're using consistently to develop your method; If you suddenly notice you're getting big crystals coming in, then clearly that's you know, a violation of that particular category. So let's have a look at the second sample. This one flowed very well. The sample particles dispersed easily on the holder. Let's have a look at those. There's very little that we can see on that one, and we turn down there brightness a little bit, there we go. So that is really interesting. So these are almost spherical particles with a little blip on, they're all almost identical in appearance and shape. They're all, I would estimate about 60, 70 microns in size, and I think from just their appearance, you can pretty much guess how these were made. But now that would be a really good categorical variable. The size is very uniform, appearance is very uniform. So clearly when you're doing your method development, if you start to see changes to that or you're running the method in a production environment and the incoming material appears to be very different, then that can be a warning to you that this material should not pass our incoming sample suitability test. Now let's have a look at the last sample. This one again was very granular and crunchy. And you can see straight away massive crystals in there; very crystalline. Each particle is essentially an individual crystal. We can swing the polarizer around and you can clearly see some birefringence setting in on these materials, quite pretty. So I would describe these as sort of chunky or they're almost like three dimensional cubic, I mean column-type of chunky columns type of thing or thick blades. So we can use these, we can take pictures of these, we can add that to the sample description and use this as another categorical variable. But essentially what we're trying to do with categorical variables is take note of things that you may observe during your routine operation and setting up samples and preparing the method. Just make note of these because they can be very important down the line.  

     If you do have an OOS, you already have this information gathered, so you can go back to that and verifying or was there any changes? Any differences that might have led to the OOS? As I said, if you built it into a method beforehand, then you can avoid the OOS altogether. You could say this material is not suitable for the method and then go back to where the source of the material came from and investigate that without actually triggering a full O0S investigation. So let's go back to the to the Raman again. Let's have a look at what we got out. So this one came in as actually quite a poor match on this case. Let me just check the sample position. Then we can run something again. Let's just check where it's sitting on the Raman unit and we can run something again on that. So what I look for is the correlation coefficient, and I'm looking for numbers above like 90%. In that case it came in a little low. So I found that it didn't really identify the material itself. And it was probably how I loaded up the sample at that point. Okay, so I think we've reached the end of what I wanted to talk about today, a very quick session. 

     Let's go to the Q and A. And if there's any questions from the audience, please feel free to post them over. Now you can find links to the systems that we've talked about here for the handheld Raman unit and for the microscope, they will be in in the notes. And so you can find all the information about the materials there. You can reach out to me on LinkedIn if you wish. Or at my Rigaku email, simon.bates@rigaku.com or on LinkedIn it's Simon_Bates. I look forward to meeting you all in the next Pharma Lab Show live. Bye for now. 


    Progeny 1064nm handheld Raman analyzer featured in this episode

    800x610 Progeny with bottle_picture_2018.12.01_0 Purchase the Progeny 1064 nm handheld Raman analyzer or learn more about it.

    Catch up on episodes of The Pharma Lab Show on our Learning Center

    Watch The Pharma Lab Show on select Fridays
    For more insights into the pharmaceutical industry, subscribe to the podcast on Apple Podcasts, Spotify, or wherever podcasts are found.
    Check out the Rigaku Pharmaceutical Technologies Showcase for more information on Rigaku's commitment to the pharmaceutical industry

    Simon Bates, Ph. D.
    Simon Bates serves customers as the VP of Science and Technology with Rigaku Americas. Simon Bates received his PhD in Applied Physics from the University of Hull, utilizing Neutron diffraction to study the magnetic properties of rare earth materials. The neutron diffraction work was performed at the Institute Laue Langevin in Grenoble. For his postdoctoral work in the Dept. of Physics at the University of Edinburgh, Simon helped design and build high-resolution triple axis X-ray diffraction systems for the study of solid-state phase transformations. Simon continued his work on high resolution X-ray diffraction systems at both Philips NV and Bede Scientific where he was focused on the development of X-ray diffraction and X-ray reflectivity methods for the measurement and modeling of advanced materials. Before moving to Rigaku, Simon spent the last 15 years working in contract research organizations (SSCI and Triclinic Labs) studying solid state pharmaceutical materials. In particular, he was directly involved in the development of advanced characterization methods for formulated pharmaceutical products based on the analysis of structure (crystalline, non-crystalline, meso-phase, polymorph, salt, co-crystal..), microstructure (texture, strain, crystal size, habit..) and their functional relationships in the solid state. Simon also holds an appointment as an Adjunct Professor at LIU in the Division of Pharmaceutical Sciences where he helps teach a graduate course on solid state materials analysis.