Workshop: Advanced sorption methods for particle characterization in pharma research

Yes. So good morning and good afternoon, depending on where you are. I think we will start now with that webinar. With that workshop. It will take the next two hours. You are welcome. We are very, very happy to get you in that workshop. It will be divided how the program said in three talks. We will ask you or I will ask you to submit the queries and your question in the question tab you have in front of the screen or at the end of each talk to raise your hand and to ask your question to the presenter. Let me introduce myself also. I'm Thomas Schmidt from Sales, responsible for Germany, Switzerland and Austria from Surface Measurement Systems. And we are starting directly, I would say, with the first talk. And I ask welcome Christian Lubert. Dr. Christian Lubert from the University, the technical University here in Germany. And may I give a small introduction of him? So, Dr. Christian Lubet studied chemical engineering at TU Dortmund University and finished his PhD in 2018. During his PhD he successfully predicted the long term stability of hundreds of pharmaceutical formulations via thermodynamic modeling. He developed methods to measure and to predict unwanted changes in amorphous solid dispersions during long term storage, amorphous phase separation and crystallization. His work was awarded as a best German PhD thesis in thermodynamics in 2018. Moreover, he received the PhD award. Today he's managing director of the company Amorphor, which offers formulation development for active ingredients in pharmaceutical, aquashemical and life size application. And I welcome Dr. Christian Lubet. His talk will be about next level amorphous solid dispersion development by combination of DVS measurements and thermodynamic modeling. We know each other for a few years. He has in his lab different DVS and DVS resolution. And it's every time a very great pleasure to work with him and to work with his team. And the stage is yours, please. Christian, take over.

Thank you for your kind introduction, Thomas. I hope that you see my screen now. Is that right? I think so. Right. So I just go ahead. Yeah. Thanks for this kind introduction. Thanks for giving me the opportunity here to talk with you today about the here shown topic and introduce you to next level ASD development by combination of DVS measurements together with thermodynamic modeling. What can be achieved if we combine these two applications in the development of pharmaceutical formulations. And it's a great pleasure that I have this opportunity now to share this with you. So good afternoon, good morning to all of you.

Perfect, we can see your screen.

Okay. And you see also the next slide now Right, Yes. Yeah. Excellent. Very good. So that works perfect. Yeah. I'd like to start with you and have a look with you at something you are certainly aware of if you are working in the field of DVS measurements. If you are aware of DVS measurements, you certainly know these kind of diagrams. In the development of a new formulation of a new drug, we often characterize how much water is absorbed at a certain relative humidity after a certain time. And a typical characterization method in the DVS is, for example, the one which is shown here. The result is taken from one of the application nodes of surface measurement systems. And you see that the sample, which might be a crystalline material or polymer or active ingredient, amorphous or crystalline forming hydrates, no hydrates that this is exposed in different steps to different humidity. Starting for example, at a very low humidity and then going stepwise up. And then the water uptake is characterized as function of time. And usually there are certain cycles applied, meaning that you are first of all going up with the relative humidity, going down with the relative humidity and checking how much water is absorbed, what is maybe the hysteresis. So the difference between the absorption and desorption step and will there be a hydrate formed, is there difference between different polymorphs, for example? All these kind of characterizations are usually performed in a typical DVS measurement in a typical cycle DVS measurement at early stage of drug development. And I'd like to have a look at what else is in this kind of data hidden. What can be derived from these data? Once we have this information, we are able to access much, much more than just the equilibrium points as such or the velocity as such. We can learn many, many more things just from these measurements. If we combine that with a strong thermodynamic modeling tool, for example. Let us first have a look at the question at the development of pharmaceutical formulation. What is actually an amorphous solid dispersion? Most of the formulators that are working in the pharmaceutical industry are facing the problem and the challenge that increasing number of drug candidates shows a poor bioavailability because the aqueous solubility is very low. Thus, we have here a permanent challenge for formulators to find a suitable formulation technique to enhance and improve the bioavailability of these poorly soluble drugs in that field. We can say that today ASDs are established enabling technique. This is a standard tool in many pharmaceutical companies to enhance the bioavailability and to increase the bioavailability of these poorly soluble drug candidates, which is shown here on the right. Side, schematically we have the API molecular dispersed dissolved in a polymer matrix. In order to being really successful in the development of these formulations, we need to understand what is the best polymer for a given API. Means we need to understand, for example, what are the API polymer interactions in a specific API polymer combination and finally, what is the shelf life and the physical stability of such a formulation. Because we want to maintain of course, the amorphous state for the entire shelf life and want to make sure that the amorphous state is maintained during the entire storage stability. Unfortunately, we have to say that there is no standard polymer that can be applied for such a formulation. There's not even a polymer as such the best solution. And we first of all have to know and identify for a tailored new API, the best polymer for and this is usually identified in development by trial and error. This means we have a huge amount of different screenings, for example, where we check what is the best and best performing polymer and just make trial and error and check based on that, what is giving us a good dissolution profile, what is giving us a good stability. This is always only working for one polymer for one API and we have to restart that again and again for each new coming API. And it's of course very helpful if we have a in silicon method that allows us to narrow and focus really on the most promising formulations just from the beginning on, so that we only make the few experiments and few formulations that are really the most promising candidates. And therefore we have the following approach. We start first of all with a thermodynamic modeling, have a look at the intermolecular interactions in silico and check how strong is a certain interaction to stabilize the API. Then we make a few key experiment validation measurements that are not necessarily the ones that I showed initially on the last slide. And thus we obtain hopefully the final asd. The entire development process covers many, many different analytical techniques and I'd like to have a look at one specific aspect, namely the prediction of the shelf life of an asd. Now in this webinar, because it requires so many different aspects, so many different analytical techniques, if we try to predict the shape life of an asd, this means if we try to predict how long remains the amorphous state stable and and what is the time until we see first crystals occurring in the asd, which is usually considered as the end of shelf life of an asd, we need to know many, many different physical properties of the formulation. To assess that. You see the main driving factors summarized here on this slide, this is for example, the viscosity of the polymeric matrix. The polymeric matrix as such is a viscoelastic fluid which has a solid appearance. So it's a powder in the end from the haptic appearance, but it is a glassy formulation. And therefore we need to know the viscosity of this polymer matrix, ideally even under consideration of the thermal history. This means how did we manufacture the formulation and how did we process that. Then we need to have thermodynamic information that means how much water is being absorbed at a certain relative humidity, what is the driving force to crystallization from a thermodynamic point of view, so that we also understand how strong is the driving force and how high is the risk to crystallization at a certain drug load. Next thing we need to know is the glass transition as such, because this is the main influence on the viscosity. And depending on the glass transition we have to say that yeah, if we are above or below the glass transition, there is a completely different viscoelastic behavior. And then we need to know and understand the crystallization regime of the API as such. There are some APIs that crystallize quickly, there are some APIs that are poor glass, very strong glass forms that crystallize only very slowly. And we need to understand the nucleation in crystal growth regime in order to understand also how fast crystallizes the API when it is dissolved and dispersed in a polymeric matrix. And once we know all these different contributions, we can estimate the shelf life of the ASDs. And for all these different aspects, among the different, among the different models and equations that are behind. We also need a high experimental effort. We need, for example, for the viscosity dynamic mechanic analysis need to understand the viscoelastic rheological behavior of the formulations. Then we need DSC measurements if we want to assess, for example, the glass transition. If we want to understand where is the glass transition of the formulation, we need spectroscopic or X ray GUM characterizations in order to understand the crystallinity, the crystal onset, crystal growth kinetics. And so we need X ray or spectroscopy for that. And if we want to know how much water is in the formulation, we need of course a DVS or similar method to understand the amount of absorbed water. So you see, there is a huge bunch of different analytical techniques required. And I'd like to discuss with you today now how we can replace some of these measurements with by just simple DVS measurements before we even start with any ASD development or before we even do Any experiment, we first of all ask ourselves, if we obtain a new API coming from drug development, what is actually the best enabling formulation for a poorly soluble drug? So what can we do in principle to find out what is potentially best performing excipient class? And it's not always necessarily the polymer. It's not, yeah, as I said, there is no standard method for that. So there's maybe even a different class as the best polymer that we need to consider. So first of all we have a look before we start with the ASD development, what are the intermolecular interactions with the different excipient classes? In order to assess potentially feasible excipients and therefore we first of all make a screening for different classes like amino acids, which are then giving us the hint to follow. So called co amorphous formulations in literature, are they called co amorphous formulations? We have the option to make lipid based formulations and dissolve the API or stabilize the API in the lipid. There is the potential to look for sugars that stabilize my API. Or we can take consider polymers which are then leading us to the ASDs. And we have to check what is is the ASD the best formulation type and then first of all check what is maybe the best for my given API. Therefore we make an excipient screening, predict the thermodynamic API activity and all these different excipients and first of all only check what is a potential excipient for my given API. We have here now considered phenofibrate and Piroxicam made an activity prediction in those formulations and see for example here that Piroxicam has a very low activity in amino acids. This means Piroxicam would be a very good candidate for making co amorphous formulations or for making ASDs as well. Whereas lipids and sugars are not that good. In contrast to that, if we have a look at phenofibrate for example, the entire scheme just turns upside down and we see now that the amino acids and polymers are the worst candidates. So the activities are the highest and therefore there are very strong repulsive interactions. And we should avoid making these formulations. We should avoid making an ASD with phenofibrate. Phenofibrate is an API where we go best for lipids. Therefore we first of all screen just the excipient class and check what is potentially the best excipient. Once we know that, we can then go into the details and have a look at the detailed interactions and calculate and compute what is the best polymer for. For example, if we go into the ASD development track, therefore we can make the same ranking again and just have a look at different polymers. These are standard polymers that are considered in drug development and we simply predict what is the thermodynamic activity in the different polymers that are available. Therefore, we first of all have a look at the piroxic arm. Again we see that there are the best beneficial interactions with PvP and PvPva 64, whereas EuroDragID is not that good compared to those. And if we have a look at for example Eudragit, there are very weak interactions, as I said, there is a completely different shape. If we have a look at RetroNERVR, this is very well stabilized and even stronger stabilized by PvP, PvPva 64 and Peg. And if we go for Itraconazole, we see again a completely different behavior. And we see now that the celluloses are the best stabilizing polymers for the ultraconazole. So you see the ranking of the different polymers differs from API to API and we have always a different order and always obtain a different best performing API for a given formulation. Very interesting to see if you compare that to the marketed formulations. So if you have a look at what is on the market already, Itraconazole and Ritonavia are typical formulations in the ASD field. We can have a look at what is there inside. In case of Itraconazole we have Sporanox as marketed product by Janssen and this is indeed formulated with hpmc. So one of the excipients that indeed shows one of the lowest activities compared to for example or PvPVA 6 form. And as a contra example on that we have Ritonavir. This is the formulation Calytra by epv. And we see here that this is formulated in the market product with PvPva 64 where we indeed see the strongest interactions also with the insulin model. So we see that there is a certain correlation between the marketed formulations and the strength of the intermolecular interactions. So it's always good to select the formulations with the strongest interactions. But I have to say that also the dissolution behavior, for example, plays a critical role, which is not scope of this investigation. This is here mainly what we are discussing here. Physical stability against crystallization. If we want to now understand is what we are predicting and doing actually right, we need to have a look at ASD phase diagrams. The people here in the webinar who know US our publications on that. They are certainly aware of the fact that we very often have a look at phase diagrams and very often consider phase diagrams as they give us a hint on the stability as such. And this is a way to validate these phase diagrams. We can directly say whether a thermodynamically stable asd, whether it will never crystallize, whether it will never change, whatever we have as a mobility as molecular mobility. For example, if we have a phase look at the phase diagram and consider the drug load and see that we are undersaturated and below the equilibrium solubility of the API in the polymer. This is shown us here, phase diagram of a typical API, PvP, ASD. And we see here that the drug load of for example 20% is stable. On the other hand, if we have a drug load that is above the critical equilibrium solubility, we will definitely see crystallization after infinite storage time. It's only a matter of time. This means we have to ask ourselves how far are we below the glass transition and how far are we away from the equilibrium solubility? And then we know whether there will be crystallization in a short period of time or in a long period of time. And usually in order to validate this kind of diagrams, what usually is done is we have a look at the C data. This means we make a formulation in a highly supersaturated state and determine then in a slowly heating environment where is the melting offset of this formulation, where is the last API molecule dissolved. And this so called melting point depression method is the, I would say, gold standard in the characterization of these phase diagrams. Although many people who are working on that are definitely aware of the fact that there are certain drawbacks that are facing us as for example thermal degradation. The problem is furthermore that we are measuring phenomena at very high temperatures which are not relevant, and we have a very high viscosity, which leads us to the problem of low diffusivity and a kinetic retardation of this phenomenon. So the melting point depression is certainly in DSC measurements, not always perfectly covered. As an alternative, I'd like to have a look at what can be done with DVS measurements in that field. We want to characterize in the end the API polymer interactions and want to learn what is the phase diagram. If we can calculate the phase diagram, for example, Naproxene and PVPVA 64, this is a nice example because we have also much data on that. The phase diagram might maybe look like that. So we don't know that right now. Maybe we have a phase diagram looking like that at room temperature, which is the dashed line. Here we have for example 20% solubility equilibrium, maximum drug load and the corresponding absorption diagram. The amount of water being absorbed at this condition, 25 degrees, 60% relative humidity, depends on the drug load and decreases with increasing mass fraction of the asd. So this is directly correlating and this is something we directly obtain from the in silico model. But it might actually also be the case like that. So maybe the phase diagram is completely different and the solubility is much much lower. So I just modified the interactions between the naproxene and PVPVO in the model and based on that I obtain a very low solubility at room temperature and I also obtain a very much stronger watersorption behavior. So the water absorption is much higher at a certain drug load. And if the solubility is very high, we have maybe very strong interactions in the water free asd. In that case we have a very low water sorption behavior and the absorption profile would theoretically look like that. Well now if we have a look at the reality at the real data on the DSC data, this is available in literature and we see that the melting point offsets, they are looking like that. So apparently the orange line is the correct one. And if we are not able to perform these measurements in the DSC due to the aforementioned reasons, we might think of alternatively measuring just the watersorption of certain formulations that are here in the, in this region where we have the highest difference and the highest sensitivity to water sorption. That means if we just make a DVS measurement of for example a 40% drug load formulation, we will directly learn how strong are the interactions at this temperature. So this is really a very high advantage compared to the DSC measurements We are measuring at the temperates temperatures that we are interested in. And now we see, if we compare that to the measurements, the data that is available, we see that the sorption diagrams indeed fit perfectly to the prediction which is coming from the interaction profile that predicted for the orange line here. So we can validate and then go back by measuring water sorption back to the solubility. So we can indirectly determine crystalline solubilities of APIs and polymers only based on DSC, only based on DVS data. And this is something very, very helpful for understanding API polymer interactions better. Next thing we have to understand is the ASD viscosity. ASD viscosity is giving us indirectly also coming ASD viscosity Is coming indirectly also from the sorption profiles. We see here. Sorption profile for toluene polystyrene system. Where we have here on the Y axis the relative solvent partial pressure. This is. Yeah. If we are talking about water. The relative humidity giving us the uptake of solvent as function of relative humidity. If we have a look at the sorption data at the sorption curves of this polymer. We see that the sorption velocity depends strongly on the amount of absorbed solvent. And the first steps that are all below the glass transition are very slow. And the sorption velocity increases with increasing molecular mobility as well. As soon as we exceed the glass transition. As soon as we have a more viscous flow in the formulation and in the polymer. And we have a lower molecular, higher molecular mobility. Interesting to see is also the axis, the x axis. The scale is here the time to the root of time. This means the first equilibrium was reached after 35 days and the last equilibrium was reached within two hours. So there is a dramatic difference in the sorption velocity. And this is something that gives us an information on the diffusivity in the formulation. If we now make a model on that and model the sorption profiles and fit diffusion coefficients to the given geometry. We are able to determine the viscosity only based on the measured water solption. So we measure how much water is absorbed at a certain extent at a certain partial pressure. How fast is the water absorbed. And then we can convert that to diffusion coefficients and to the viscosity here as shown on the right hand side. So we learn directly what is the viscosity of the asd. If we consider the mass uptake, the DVS behavior, soption behavior at a certain partial pressure of a polymer. This is an information that is hardly to obtain from other methods. Of course, what we are working together with is the company Melt prep. Here we need viscosity measurements for biosoption experiments. And we need a disk clearly defined surface with a clearly known surface area in order to assess the diffusion coefficient properly. And we work with this kind of samples here and make the samples on that way. The last thing we need to know and understand is the crystallization kinetics of my material. If we again have a look at the sorption application nodes from surface measurement systems. This is also published in a publication in International Journal of Pharmaceutics. As shown here we see that the crystallization is determinable by the amount of absorbed water. Here is an example shown for amorphous lactose that Converts from the amorphous to the crystalline state at different humidities. And we have a very handsome method here to determine that in situ and online just by DVS measurements. Because the crystalline lattice incorporates almost no or hardly no water, Whereas the amorphous material can absorb very high amount of water. In case of a pure material, we would always expect that the water absorption goes down during crystallization and that the water absorption falls down. The problem with the amorphous solid dispersions is that we have sometimes polymers that are very hydrophilic. And we have sometimes not the typical fall down of water sorption due to crystallization. But sometimes we see a behavior as shown here. This is giving us the water absorption as function of drag load. And in that case we would expect that the formulation, as long as it is amorphous, absorbs water according to the blue line. And once we have crystallization in the system, it will go up up instead of normally going down. We observed that also in the experiments. After five days, we saw that the experimental values were closely lying to the blue line, closely lying to the absorption prediction of the pure fully amorphous state. And once we waited for a very long time, 550 days we see, or 500 days we see now that the water content increased. So there is not any more the clear stating possible crystallinity increases, watersorption goes down. This is only true in some cases, as for example here in the Philippine pvac. But if we understand the thermodynamic phase behavior behind and know what do we expect Mass increase, mass decrease. If we understand that, we can also work with the mass increase and quantify the crystallinity with the water soption measurement. Based on the predictions and the DVS measurement.

Again.

Yeah, the crystallinity was also confirmed. We know this is definitely crystalline in X ray also seen. And we might say that the water content depends on the physical state of the ASD and that it might change during time. And if we now convert that to an application and want to learn, as I said initially, want to learn about the crystallization behavior nucleation crystal growth in an asd. We need to understand the sorption profile of the fully amorphous and crystalline state. For example, 70% drug load, we will first of all have a mass increase to the blue line and then a decrease to the orange line. And this means then directly if we have a blue line measured, we have fully amorphous. If we are measuring something in the vicinity of the orange line, we are fully crystalline. And now we can convert that so first of all we know that the water content goes up and then with increasing crystallization, water content is here in the shown case going down again because we have also very hydrophobic polymer that is investigated here. Then we convert that to the crystallinity and measure just with the sorption velocity, the sorption behavior as function of time, the crystallization kinetics. And this is something we then determine as here shown for the neferipine formulations. We determine the crystallization as function of time and assess how fast crystallizes by formulation. So we see watersorption measured as well, goes up and then down, convert that, see it fits and then we convert it to the crystallinity and we see the crystallinity quantitatively in C2 during time. And then we can play a bit around and check for example what is the influence of drug load. Investigate different drug loads from 90 to 70, see the decrease in crystallization velocity with decreasing drug load. We can have a look at, for example, different temperatures going from 40 to 30 degrees centigrade. We see also a lake in the crystallization velocity. And we can also have a look at the influence of humidity and see what happens at high and low humidities and take that as well into account. Thus we have here an approach that allows us for a calibration free and easy method to determine in situ the API specific crystallization mechanism. As shown here. We learn from a DVS measurement about the nucleation of the crystals, about the crystal growth and for example the glass forming ability is a typically low or fast or slow glass former will it crystallize instantaneously? And if we have that all and combine that all together with the model and bring all these DVS contributions together with a model, we are able to have a look at the shelf life of the formulations. I have an example for Phylodpine and PvPv A shown here. And I'm going to plot always the predicted versus the true measurements and and the shelf life as we observed in different storage conditions 40, 75, 25, 60, 25, 0. The ones at the highest and harshest stability study@4075 open dish conditions crystallized quite quickly. And if we then decrease the harshness of the systems we will see the following. These are all different drug loads, all ASDs with different drug load from 30 to 90%. And if we go to 20, 60, 25, 60, we see that the crystallization velocity indeed goes down and that this is all covered across the entire temperature and humidity range and also dark load range covered adequately. By the model. So we are able to set up a model that allows us estimating the shelf life of a metastable ASD just based on this theoretical model and a dvs. And then we can take this for example, go to more practical applications, practical stability conditions. For example, a formulation with 25% dried load and investigate and predict what is going to happen at 40 degrees centigrade, 40 degrees, 60% relative humidity. We expect a shape life of 8 months. At 25 and 16 this increases already. So just going temperature down increases to almost three years. And if we have a look at the zero relative humidity, no humidity in the system, yeah, we have a huge increase in shelf life going up to 15 years in the water Free State at 40 degrees and at 25 degrees Centigrade, which is the room temperature around us, we have a shelf life of 87 years. So really a huge shelf life. And you see this is all the same drug load, only different, very narrow changes in relative humidity and temperature. And we see this broad and extreme scattering of the predicted shelf life. So this is what I hope to show you. Now that is more possible to obtain from a normal DVS measurement that there is more inside typical sorption experiment than just the normal equilibrium value. As we are used to that, we learn much, much more from that. If we combine that with a strong computational approach and learn from the basic interactions, the basic intermolecular interactions in the formulation, put that into in silicon model validated with DVS and with other analytical techniques. Therefore, by that confirm the predictions which are made on the intermolecular interactions, we can assess the excipient compatibility, the shelf life, miscibility, solubility, stuff like that. So completely different things, all just coming from the key validation measurements that are performed with the dvs. And thus we are able to assess and predict stability and the shelf life of the desired formulation, which is of course the key goal for the final marketed product. This is what I understand with next level formulation development that we are doing really targeted experiments with the dvs. We combine that with the calculations with meaningful experiments and by that get a really, really good understanding of the ASD and what will happen so that we can in the end sleep well and do not have to be afraid of crystallization anymore in the asds. This was it. I thank you all for your attention and for listening. Happy to take your questions and yeah, thanks for listening.

Thank you Christian for this very interesting talk. You have several questions. I will read you the question. The first question is from Joseph. Will you cover Phase separation too?

Yeah, definitely. I did not have the time and yeah, I didn't have the time in this talk now to discuss that. You see that or you saw that in the sorption profile? Just roughly, I made the sorption prediction for the fully amorphous state, for the phase separated state and for the crystallized state. It is covered in the model and it is considered. So it's phase separation might accelerate crystallization. And yeah, due to that you might have a different shelf life than just from the normal crystallization mechanism, but it's considered. Okay, thank you.

Second question. How is thermodynamic activity being measured or calculated?

The thermodynamic activity is very hard. It depends on which thermodynamic dynamic activity we are talking about. If we want to measure the water activity in a formulation, we can directly take the dvs to measure that. And if we have for example not the water activity, I mean IGC and DVS is a typical, is a typical water activity measurement. IGC for example, at infinite dilution, activities at infinite dilution, DVS at a certain composition. If we want to know for example the activity of the API in the polymer, we need a help quantity which is the crystalline solubility, for example and the crystalline solubility, this can be measured also in organic solvents and need to combine that with a thermodynamic model. And then we have the thermodynamic activity, but it is very often only a theoretical quantity. Okay, thank you.

Next question. How did you determine the API polymer intermolecular interaction?

Yeah, this was determined by asking the in silico model and making the prediction for this specific mixture. For example, for a given API polymer combination. This is based on the in silico model of the API, for example, is based on real solubility data that is taken to fit and train the model. And then we consider this interaction and calculate for example the activity and the interactions in a different environment, like in a polymeric matrix. And by that we obtain these interactions. They are then first of all fully predicted and just a blind guess to make the first ranking. And if we want to make the validation for example of a lead formulation, we go for DSC or DBS measurements and confirm that with these selected measurements and yeah, also make sure that the computations are right. And I showed also in this phase diagram composition that there is certainly a certain scattering. So depending on the model might be lying between in a certain range this predicted solubility. And in order to really make sure what is the correct One. Yeah. This is then a question for the DVS measurement. Okay.

Yeah. So there's a question about the recording of that workshop. Yes, the workshop is recorded and we get the right to send you the recording or the slides if you would like. So we, we have another question. How do you account for the polymorphic mixture crystallizing out of ISD for fraction crystallized using dvs?

Well, this is a very specific question for the polymorphic transitions. In the end, we have to understand then we have different solubilities. We have the solubility of the one polymorph and the other polymorphism and they differ slightly depending on the different crystal properties because the crystals, the polymorphs, have different, slightly different melting enthalpies. And then we need to understand both. But yeah, in polymorphic transitions, I must admit it is very hard to really clearly and precisely account for that. In the end, the solubility of the different polymorphs in the polymeric matrix does not change that strongly, usually because the melting points are scattering something between 20 degrees centigrade, I would say, and the melting entropies are in the range of also plus minus 10 to 20%. And therefore there is not a huge influence coming from the different polymorphic transitions. But if that is occurring, and if that happens, I would definitely add a spectroscopic measurement as well and utilize for example, DVS plus Raman to observe the polymorphs as well and see how fast crystallizes the one and the other polymorph. In the end, in thermodynamic equilibrium, there is only one polymorph remaining. So there should be only one thermodynamically stable polymorph at a given storage temperature and condition.

Okay, Ira has also a question. Knowing the overall interaction landscape is important, but shouldn't knowing specific interaction that are playing a role to determine the landscape? Maybe there are coexistence or competitive interactions in isd.

What do you mean specifically by that? This is right? Do you mean for example, that a very strongly interacting polymer is beneficial for making a stable asd, but not beneficial for having a good dissolution profile? Or how do you.

Okay, I will go to the next question and perhaps we have the possibility to unmute. Ira, to specify a question, how do you account for the randomness of nucleation process in prediction of shelf life?

Now I'm unmute. Yeah. Thanks. This is a random nucleation process. We are always working with an induction or considering the crystallization in harsh conditions where we have almost no induction time and measure the crystallinity for example in pure API or in a highly drug loaded formulation where the nucleation occurs spontaneously and quickly and then calculate to other conditions and we always consider the same nucleation time which is then zero and assume that this occurs almost instantaneously. This is then so to say the the worst case scenario that can be imagined and if you are lucky and you can avoid this nucleation by chance then you then you have a higher stability, a certain higher shelf life. But this is uncontrollable, this is not predictable and therefore we are always working with the worst case scenario. No induction time.

Answer to your question after the workshop. So okay, we'll do like that. Please re ask or specify your question after the workshop. Will you forward your question to Christian so you can talk together so you can answer it. I would like to thank Christian for a very interesting talk. Thank you very much and we will pass to our second talk. It is given by Maishan so thank you. Christian Maishan is an application engineer in surface measurement systems. She joined SMS in 2018. She has been working on surface characterization of solid materials including fiber, powder and film samples using dynamic vapor absorptions or DVS and IGC inverse gas chromatography, other techniques. Meishan obtained her master degree in Chemical Engineering at Imperial College London in December 2017. Her research project was focused on investigating the properties of fragrance molecules with solid substrates by diffusion adsorption. She also received a bachelor degree in Chemical Technology at the Hong Kong Polytechnic University and in Chemistry at Sun Yat sen University in June 2016 in China. I presume my share the stage is yours and the talk is about dynamic vapor absorption technique for particle engineering.

Thank you very much Thomas for the kind introduction. I'm just going to share my screen so I hope everybody can see my screen now. Is that all right, Thomas?

Yes, we can see your screen for everyone. Question Chat is still open for the second talk on the right side on your desktop. If you have any question, please ask.

It in the question tab.

Thank you.

Okay, so thank you very much for Dr. Christine's presentation about the DBS application. It's really informative and nice and I do have another general DBS presentation. It's more related to the basic technique and more general applications in the pharmaceutical industry. So we will focus on the basic technique and instrument in my presentation and here is the outline. We will first briefly introduce the DVS technique itself and then we will focus on the two most typical methods we normally use with the DVS technique and then some other General applications for the pharmaceutical industry. So we will talk about how to measure the bets of this area, the formation of hydrate or solvate, as well as some other applications like the Amalfi quantum determination and some variation determination. So the dynamic vapor assumption basically is a gravimetric method based on a really accurate microbalance that will put the sample of interest into our balance. And then we will change the concentration of the vapor o site surrounding the sample itself. So this should be a water vapor or the organic vapors. And then we weigh the sample inside the chamber to reach an equilibrium to. And then we record the mass change of the sample itself. So basically we are introducing different levels of relative humidity or partial pressure of organic vapors. We will wait the sample to reach a fuller equilibrium until we get a full picture about the sorption and the desorption. So from this kinetic information from the math plot, we can then get equilibrium mass change at certain level of relative humidity and and plotting these sorption and desorption isotherms. So we can then predict loss of performance or the physical chemical properties of the solid material from this isoscerm. So for example, we can calculate the BET surface area, the pore size distribution, the diffusion coefficient or the permeability of the thin film sample. And we can also measure the water capacity stability or chemical stability, physical stability of the shampoo at different conditions as well as. So just give you a brief idea about how the DBS instrument is working. So basically we have a micro balance with the resolution of 0.1 microgram. And we will put sample in the latent side chamber for the sample site. And we keep the resonant side as empty for the reference. And as you can see we have two bottles at the back. So we have one for water vapor and the other side for the organic vapor. So basically the mass controller inside the system will control the flow of decayer gas which could be dry nitrogen or dry air with the saturated vapor in a ratio to reach a certain level of relative humidity. And we also have our sensor in the front here to determine whether this mixture is reaching the target value we required. And then this mixture will go to both the sample chamber and the reference chamber at constant rate. And then we can determine what's the mass change at a certain level of relativity. So we have the whole setup like this one sitting in an equivalent which has the temperature range from 5 to 85 degree. And with the DDS we can also associate with some other accessories like preheater. So this will allow you to do the in situ preconditioning up to 200 degree. Or we can also use some other video or Raman spectra, Nir spectra as additional accessory pointed to the sample which you can get some video image for the Rama spectra along with the DBS data. So for example here we have a set of data with the freeze dry sample at different humidity. You can see clearly changes of the nature going to the liquid site. And we also have some cellulose sample along with the humidity changing. You can see a swelling and then shrinking behavior of this cellulose material. So with the video function it's really useful to study any kind of changing in the nature of the material or changing in the color of the material along with the DVS data. And another option as Dr. Christine also mentioned before, that DBS can also be integrated with Raman. So we can put a Raman gun mounted to the sample of interest and then we can collect Raman spectra at the end of each humidity step. So here we have the amorphous electrode sample from the left hand side we have the DVS data which we changing the humidity from 0 to 90% RH with different steps. And we noticed a significant drop in the mass for this lactose sample which correlates to the amorphous to the crystalline phase change. And we also collected the Raman spectra at the end of each step. So at 0 and 40% RH where brought pig from the Raman trans showing the atmorphous phase above 60% RH with new pics showing up here just confirming with the crystalline state of the material. So the combination of the Raman, the video camera as well as the DBS data itself will really give you a full idea about how the sample will change with the temperature as well as with the humidity. And then I'm going through two typical methods we will normally use with the DVS technique. So it's exactly the same method Dr. Christine showed before. So I'm going to just go through very quickly and then we can go through some case studies. So the method basically you will wait the sample to reach equilibrium at each stage and then it will move to the next stage automatically. So we have the equilibrium mass change at each route humidity and then you can perform multiple cycles of both sorption and desorption on the same sample to see whether it is a reversible or irreversible process. So the case study I have here with the same step method with watersorption is on a potassium citrate sample. So normally if the sample has very good Uptake of water. We don't need to use a lot of sample because the balance is really sensitive. Here we only use about 8 milligram of the sample. Still we can get some good data of the watersorption desorption. So we set up the method from 0 to 95% RH with 10% intervals. And we get this kinetics data from the math plot. So as you can notice, the sample didn't pick up any of the moisture at all before 70% RH. So basically the sample is really stable up to 60% water humidity and then gradually pick up loss of water with the humidity going up in the high humidity range. And the mass change could be even 200% mass of change. And the same here. We can get the equilibrium mass change from the isotherm report and then we plotting the isotherm from each of the steps. So we have the change of mass as a function of relative humidity. So it's probably more clear from this isotherm the sample is really stable up to 60% rhythm and then pick up lots of water at high humidity level. And the blue line for the desorption. As you can see, at 30% RH there is a sudden stabilize of the shampoo and at the end of the desorption still some of the water molecule trapped inside the material in the structure that didn't go back to zero for an irreversible watersorption behavior. So we do take some of the images during the running of the experiment. So we have some particulate sample at the beginning for the potassium citrate. This sample keeps the signature up to 60% RH as shown in the DVS data. No water pickup at all. And then gradually dissolve with the water increasing and fully dissolve at higher humidity range. And this liquid paddle keeps the same up to 40% RH during the dissoption. And there is a sudden crystallization of the material becoming a crystalline nature at the end of the 30% RH. And this crystalline material keep the same nature during the desorption back to 0% RH. So the original sample is quite different compared to the last sample after disruption behavior. So basically here you can study how the morphology changes along with the humidity from the combination of the DVS technique as well as the video cameras. The next study I have here is a little bit more complicated. So we perform the same watersoption experiment at different temperature. And we want to study the temperature stability of the material for the macro crystalline cellulose. And because this cellulose sample have good uptake as well. So we use about 10 to 20 milligram. And we perform the same experiment with water from 5 degree up to 60 degree. And there is a really nice overlapping of all the data at different temperature. So for the red line we have here is for the microcrystalline cellulose at 5° and up to 25°, which is the pink line here, the uptake of the water is hardly changing almost the same level, about 14% change in the mass at higher humidity. So basically the uptake of the moisture cellulose is kind of constant in this temperature range. But the kinetics of the equilibrium state is much faster at a higher temperature. And when we keep increasing the temperature from 40 up to 60 degree here, at elevated temperature, the water suction capacity of this microcrystallite cellulose drops a lot, which could be due to a structural rearrangement inside the material of this crystalline cellulose. So basically the stability of these cellulose is not ideal at higher temperature. And here we can see the effect of the temperature on the hysteresis loop as well. So basically at higher temperature, which we have the green line here for 60 degree and we have the red line here for 25 degree, the high temperature sample has a reduced hysteresis as the water molecule can move more freely and the diffusion barrier for the desorption is less. At higher temperature and from the water uptake at different stages. We can also apply the theory to calculate the diffusion coefficient of the water into the structure of the material by using the radius of the particle size of this microcrystalline cellulose sample. So here we summarize the temperature effect on the diffusion coefficient of this crystalline cellulose that at higher temperature the diffusion coefficient improves a lot. So basically water will have sorry temperature will help the water molecule with the better diffusion rate into the structure of the material, which give a higher diffusion coefficient as well as a faster kinetics for the moisture absorption. And then we have another typical method with the DVS technique. So here we didn't change the shampoo to reach equilibrium, but this time we ramping the humidity from 0 to 90% RH with a constant ramping rate. So this study we perform on the lactose sample to see any kind of critical changes inside the structure of this lactose material. So at drying stage here we had the right line losing the mass, which is just the flow of the carrier gas, to remove any kind of free water on the structure. And then when we begin introducing the humidity to the Sample the sample pick up the humidity slowly because of because of the water is only absorbed on the surface of the material. And as you can see, after a certain point of humidity, the water active will become dramatically because the water molecule could act as the plasticizer that will induce some vapor induced phase change happening inside the material. So basically, the transformation from the surface option to the box option is because of the glass transition of the material. And from the slope before and after this change, we can easily determine the glass transition humidity for this lacto sample at 25 degrees here. And when we keep increasing the humidity, the sample can also go through the recrystallization step. Basically, the sample goes through the amorphous phase to the crystalline phase with less soft capacity and less surface area. So basically from this ramping method we can get two critical conditions for the material. One is for the glass transition point of the material and the other is for the crystallization point of the material. And if we perform the same experiment at different temperature here, we can get a nice correlation between the temperature and relative humidity for the prediction of the stability of this material. So at each temperature we can get two critical points. And with a wide range of different temperature, we can get a blue line here for the glass transition and the red line here for the crystallization. So for example, if we have a sample at a certain condition for the storage or for the application, we can then easily predict this sample is kind of amorphous phase, a crystalline phase, or just in the middle for the structure to change anytime during the experiment. So this method at different temperature is really useful to predict authority, the stability of the shampoo at different conditions with the temperature and relative humidity. And then I will go through some other applications with DBS technique as well. So basically with DVS you can changing the PRO molecule from water to the organic vapors. And then you can apply different methods or similar method as I introduced previously to to measure the BET surface area of the material. So basically, the method here for the theory here we calculate the BET surface area is exactly the same as the volumetric technique. We measure the amount of the solvent or the molecule absorbed by the surface. And then we measure the pressure pressure above the surface. Then we can calculate the monolayer volume of the sample of the PRO molecule we have here. And with the cross section area of the PRO molecule we are using here, we can then calculate the BET surface area of the sample of interest. And normally we will use some organic vapors to form the monolayer with the sample, which gave you a type 2 or type 4 isotherm to calculate the surface area of the material. We don't really recommend it to use water isotherm to do the calculation because most of the time the water will make some cluster on the surface. It's not dispersion very well on the surface of the material that's giving a type 3 or type 5 isotherm. So for example, here we have a sample as alpha lactose monohydrate. And because the surface of this lactose sample is very small, we need to use a large amount of sample to get a good signal to noise ratio. So we perform the Octiischamp at 25° with a full range of sorption. So from 0 to 90% or 95% RH or partial pressure with smaller intervals in the BET range. So with the apply of the equation we introduced in the previous slide, we can get a good correlation in the BET plot. And from the slope and intercept of this plot as well as the bet, the cross section area of the PRO molecule we are using, we can then get the monolayer volume absorption constant as well as the BET surface area of this lacto sample with a very good linear regression. So the BET surface area of lactose is about 0.26 square meter per gram, which is really highly constant. With the literature reported value for the lactose material and the major difference between the gravimetric method and the volumetric method, it's more like the condition of the experiment we are applying. So basically we didn't expose the sample to some extreme condition like liquid nitrogen temperature or the vacuum condition that will sometimes will destroy the structure of the sample of interest. So we perform the experiments more like in the real world applications, real world conditions to give a very practical value for the precising of the storage. The next case study is about the formation of hydrate. So we have the water structure of different kind of hydrate and the hydrate state of the material may influence several physical chemical properties, including the physical and chemical stability of the material. So by voltage option of different material we can study this kind of hydroformation is a reversible or irreversible pore size as simply doing two different cycles on the same material. And for the structure information, if you have the molecular weight of the sample of interest, we can also calculate how many moles of molecule is hydrated in the structure. So here we have the hydrate of this sample that we perform a watersorption from 0 to 90% RH with two different cycles and there is a significant increase of the water of the mass change at certain level of humidity, which is 40% RH here, which indicates a transformation step with the formation of hydrate. And if we plot the hydrogen together, we can see a very nicely overlapping between the first cycle and the second cycle, which gives the information about the two multiple isotherm are identical to each other, indicating that it's a reversible hydrated formation here. And if we have the molecular weight of the sample, we are doing the test for the experiment. We can also calculate the STO metric of the structure of the hydrate formation and the case of the irreversible hydro formation here also for the two different cycles. And now you can see the blue and the red line for the first cycle and the pink and the green line for the second cycle is not overlapping. It's two different cycles for the water sorption behavior for this material. So basically mean during the desorption of the first cycle, the sample changed the structure chemically by giving a different absorption behavior on the second cycle. So this means it's an irreversible hydrogen formation for this kind of material. And another case here we have is the nadochromyosodium trihydrate we are doing so we use about 5 milligram of the shampoo and we use the water moisture soften behavior from 0 to 95% RH. And during the drying step, which just the flowing of the carrier gas about 8% of water mass change is losing which indicating of two moles of molecule hydrated. So basically it's from a trihydriform into a monohydriform during the drying step. And when we keep introducing the moisture up to 20% RH, this trihydriform becomes a gain during the structure and is kind of stable up to 90 or 95% RH. So maybe it's more clear in this isotherm. So during the drying stage, the sample forms a monohydrate and at 20% RH is trihydride formation here. And at 90 or 95% RH there is a peptohymihydrate in the structure of the material. And this structure is relative quite stable during desorption step that go back to the monohydrate at the end of the desorption without reformation of the trihydrate during the desorption stages. So with the DBS technique it's really simple and it's very simple to study the formation of hybrids, the structure information as well as the kinetics information. The next story I'd like to introduce here is how to determine the amorphous content with the DBS technique. We do have many different methods to narrow the amorphous content. And this is just a method based on the fact that the sample will squeeze out any of the absorbed molecules after the crystallization. So we have a kind of solvent that will induce the crystallization inside the material. And we need a kind of reference standard with known amorphous content for the reference value. So here we know this acton at 60% partial pressure will induce the crystallization inside the reference material. And then we design the method with three different stages. So first we expose the sample of interest into 30% partial pressure of acetone for both the amorphous and the crystalline state. And then we expose the sample up to 85% partial pressure of the acetone to crystallize all the amorphous phase inside the structure. And at the end we expose the sample again to 30% partial pressure of acetone. Here is for the fully crystalline state of the material. So we will notice there is a difference between the first and the last stage of 30% partial pressure of acetone. And this difference is related to the amorphous content of the material. So we have a reference material with no amorphous content, the percentage to that reference material. We can then calculate how much of the amorphous content inside the sample of interest. So I have a case study here with an API sample. We use about 50-100mg of the sample for the analysis. And here we use the ethanol as a PRO molecule will induce the crystallization inside the material. So the same stages we introduced previously 20% ethanol and then 90% ethanol to induce the crystallization. And we compare the difference uptake between the first 20% stage and the last 20% stage for the calculation of the amorphous content. So we have the crystalline material in the pink color here. And we have the 2% amorphous standard as a reference material. And we calculate for the red line for the sample of entries, which is about 6.3% amorphous content inside this material of interest. And if you have different kind of material or different kind of solvent that induce the crystallization inside the material, we do have some other method to suit your sample and your pro molecule to calculate the amorphous content as well. The advantage of this method with the induce of the crystallization before and after for the mass uptake is more relevant or more accurate for the low concentration of the low amorphous content inside the material. So we get a really good reproducibility and sensitivity for those low amorphous content. And we can determine as low as 0.2% amorphous content inside the material. But the requirement for this method is we need standards with more amorphous content and also we need a vapor that will induce the amorphous to the crystallization. And some other slides we have here is for the variation of different batches that CBS can have a very good sensitivity to determine the hydrophobic drugs with very low uptake of water only about 0.07 and 0.05%. This variation in different batches could be related to the particle size of different batches that give you a different surface area. And this low concentration of water uptake is related to the surface of this water molecule for different size in different batches. And the last slide here is we can also determine the variation in the polymorph so maybe the same sample it gave different watersorption behavior for different polymorphs. The first polymorph A is a kind of high uptake about 4.5% change in the mass for the 90% water stage and is a kind of reversible hydrate formation at 70 and 80% RH and the polymorph B has lower uptake less than 0.2% only the surface option of water is more indicating some hydrophobic behavior of this polymorphism polymorph material. So with the DBS technique, just by simply marrying the water isotherm of different polymorph material we can determine or identify which kind of polymorph is sitting for this material of interest. And that goes through my last slide about the summary of EBS technique. It's a very versatile and sensitive technique. Determine the water sorption or the organic vapor sorption with different applications like bet, surface area, the formation of hydrate, the batch batch variation as well. So if you have thank you very much for your attention and I hope DBS could be useful for your future research. If you have any questions, please feel free to let me know. Thank you.

Thank you. Thank you Maishan for this typical example and this good overview. We have questions in the question tab and interesting question. First question how does the surface area from DVS compare with values from N2 nitrogen adsorption? Does the value change with solvent chosen octane or other alkanes?

So basically from our experience most of the inorganic material we do have very similar values between the surface the volumetric technique like nitrous option or the DBS technique like silica or some other kind of porous material, like some organic material. If there is a difference between the DVS technique and nitrogen sorption could be related to the structure changes at dish harsh conditions for the volumetric technique. So during the low temperature of the degassing in the vacuum condition, some of the sensitive structure of the material could be collapsed or losing the hydration water, which is most of the cases for the pharmaceutical powders. That's why you might notice a difference between the volumetric and the gravimetric method. But most of the cases if the sample didn't change the structure in different conditions, the value should be comparable.

Okay, there's another question going in the same direction. Can you explain more on how to measure BET using the dvs?

It's a very simple method that you only changing the probe of water to the organic vapor. So most of the cases we want to use octane or ethanol based on the type of the isotherm. So you want to get a type 2 or type 4 isotherm. And we need the points in the monolayer range to do the calculation. So. So you probably need to use a small intervals between 0.05% 5% to 35%. So just the step method with octane from 0 to maybe 50% with smaller intervals. Then you can do the calculation for the binky survey.

Okay, thank you. The other question on amorphous content. What is the sensitivity of this DVS to detect unstable amorphous. Amorphous backslash. Amorphous. Amorphous phase separation.

We do have different methods. We have application, no summary, different method we can use to determine the amorphous content. And the method I introducing today is a method based on the crystallization of the material is sensitive for the low amorphous quantum material. We have some papers or some application notes related to that. If you're interested, I can share you after the meeting with some more documents.

Thank you. Maisha. We can say, I would say generally that the sensitivity is 0.5% or below 0.5% depending on the method, depending on the vapor used. Other question on amorphous phases, does size of the sample particle. So does particle size, I think affect the detection of amorphous phase separation.

For the particle size of the material for the amorphous content, do you mean the correlation between them? I think the uptake will be different, but because the amorphous phase is more related to the box structure of the material. So basically, if you determine. If you determine different uptake from the different particle size of the amorphous content. I don't think you will have quite big differences which is related to the box structure. That's a good question.

Okay, thank you very much, Maishan. Thank you to the audience for these interesting questions. I think we will pass to the next talk in view of time. So the next talk will be Conor Hewson from Surface Measurement System. So Connor is also a member of the application science team at SMS. He joined SMS in 2022 and has been using inverse gas chromatography and dynamic vapor sorption to characterize the surface property and sorption kinetics of powders, fibers and films. Connor obtained his master degree also in chemistry from Nottingham University in 2019. His research project was investigating the kinetic and mechanic features of the heat operation of nitrile using novel Mg based catalysts in graduation. He has worked as an R and D chemist of a company specializing in suspension polymerization of meta crelates and as a quality control chemist for Europe's largest air liquids producers. And we will hear the talk about surface energy for particle engineering in pharmaceutical solids processing. The stage is yours, Conan.

Thank you very much for that kind introduction Thomas. And yeah, I'm going to be talking about a very different technique to our other two users from today and we're going to be looking at a technique called inverse gas chromatography. So just before I go, can you guys see my screen?

Thomas yes, we can see it.

Okay, that's it.

Great.

So yeah, like dbs, inverse gas chromatography is a technique that's going to be using probe molecules in order to analyze a material as a whole and a surface as a whole. But now we're not as observing the uptake of those in the sense that it changes its mass. We're looking how to adhere to the surface and we and what that does on an energetic basis. So I'm just going to be talking about an introduction to inverse gas chromatography as a method and how we can use molecules as probes. Talk in detail about how we can use this to determine the BET surface area of our samples. Once we have the BET surface area, we can go on to find out its surface energy, map its surface energy and then I'm going to talk about some case studies of how we can relate some of these surface properties to physiochemical properties. So inverse gas chromatography was first developed in the 1960s as a technique and in the past IGC experiments were performed on your classical analytical gas chromatography pieces of equipment. Now these pieces of Equipment were not optimized to perform inverse gas chromatography experiments. They would often experience problems with uniform column heating and and had problems with injecting gas of known volumes and of small volumes. So this technique came first commercially available in 1999, and the IGCSEA, which stands for Surface Energy Analyzer, was released in 2009. And it was the first specially designed piece of equipment to use this kind of technique in order to evaluate the energy and the surface heterogeneity of a sample. So on this slide here, you can really see the difference between your classic analytical gas chromatography and inverse gas chromatography. In the case of our analytical gas chromatography, we are passing over a mixture of gases which is our unknown, and this is acting as our mobile phase. And we're passing it over a column packed with a solid phase, our stationary phase of which is known. What this is hoping to do is to separate our mixture of gases into fractions. From the location of the peaks in this type of chromatography, we identify what the gas is, and from the area underneath these curves, we find out how much of that gas there is. In inverse gas chromatography, what we're looking to do is swap the role our stationary and our mobile phase. So now our stationary phase is what we are packing our column with. It is our sample of interest. And now what we are injecting is unknown entity. It is going to be a gas injected in a single injection of a known volume. So what we will do is inject this in a range of different concentrations and in a range of different dilutions. And that will change the characteristic of these retention peaks. And as these retention peaks change, we can plot our isotherms for every probe that we've injected, and we can relate these changes to differences in their physiochemical properties. What's also important to do is to find out the dead volume of our columns that we are packing off sample with. So it's important to do an injection of an inert gas, which is typically something like methane. And what this is doing is calculating the dead volume. And the dead volume is just the amount of area within our column that can be attributed to the volume of the column itself, which is empty, and the volume of the sample that is occupying the column. So we can take away this dead volume of from our retention time volumes that we get from our other probes in order to get our net retention time. And we can relate this net retention time directly to the strength of the Interaction between our probe molecule we've injected and the solids that we are analyzing. So this is what a typical IGCSE column looks like. It's about 30 centimeters in length, has an outer diameter of 6 millimeters, and has a range of different inner diameters which have different suitability for different side samples. Now the amount that we're packing into each of these columns highly depends on the surface area of our sample and the affinity of this sample in order to interact with our probe molecules. So if we have a very low surface area material, we will need to pack to the higher end of what we can fit into our column. Whereas very high surface area materials will interact heavily with our probe molecules, leading to very large elution times. Then we will pack a lower quantity into these columns. And here you can see the range of different columns that are needed. So something like a very fine powder where we're putting in a very small amount of samples. A smaller column would be preferred. And then you have larger ones all the way to a 10 millimeter column, which will be better for using pills as a whole or tablets as a whole. An important thing to note about this type of analysis is that there's no special preparation required to your sample before we begin. And it's non destructive analysis. So your sample is completely recoverable at the end of the experiment. So our typical mythology when we're going through an experiment like this is to pack our column. Then we go through a range of selecting our probes. We will split them into two different camps. We have our non polar solvents, so this will be something like a straight line alkane, hexane, hexane, octane, all the way up to straight line alkanes of higher order. These will be injected in order to work out the surface interaction with respect to dispersive forces. So its likelihoods to react with our probes due to van der Waals interactions. We will then inject a range of different polar solvents. And what this will do is it will find out how polar our surface is. So its affinity towards reacting to charged bodies and to dipoles. We will inject all of these probes in a range of different dilutions. And from this we can assess the high energy sites and the low energy sites of our surface. From these different concentrations, we will get different retention times and we'll be able to plot a load of different isotherms for every single vapor that we're injecting. And from there we can work out some physiochemical parameters of our surface. So this Is what a IGC surface Energy Analyzer looks looks like. You'll notice we have two vapor reservoir drawers either side of an oven in the middle. In each vapor reservoir drawer we can hold six vapor reservoirs of all different organic solvents. So we can have 12 in a system at any one time. What this will do is it will inject vapor from the headspace of these solvent reservoirs. It will introduce it through a concept column with our carrier gas. Here you can see how the column sits within the oven. It will inject our carrier gas and our probe molecules through the top of this column over our sample, come out the ends to the detector. This oven will uniformly heat our column all the way through the top to the bottom with a strict temperature control. So we're having true isotherm analysis of our sample. During the experiment. You'll see that we can run temperatures anywhere from 20 degrees C all the way to 150 degrees C. And you'll notice that we can do our experiments as a function of relative humidity. So we can find out the BET and the surface energy whilst also exposing it to a range of different humidities to see how this might affect it. Our flame ionization detector is what we're using to detect our organic gases coming through our injection system. And this is directly will make an electrical current which is directly proportional to the concentration of the gas that we're injecting, but completely inert to our humidity. We're introducing and inert to our carrier gases. And here you can see in this picture here the range of different samples that we can apply to IGCSEA that can be analyzed from BET to surface energy. And here is just a range of different thermodynamic properties that we can observe using this technique. So we can measure its BET surface area from that go on to find out its energy, Split its energy into dispersive and polar contributions. We can map its high energy sites and its low energy sites and then go on to find out some other physiochemical properties from a few more injections into the system. So we can find out its acid and base properties on the surface. So it's likelihood to be an electron donoring surface or electron accepting surface. We can find out the heat of entropy and heat of enthalpy of every probe we inject. And then go on to do some other stuff like phase transitions by varying the temperature. And find out solubility and diffusion coefficients also. So first I'd like to start with measuring the surface area, which is vital to do before we can go on and determine surface energies. So on this slide you can really see where DVs as Meshan presented previously and IGC stands in their performance and their ranges for determining bet of various other materials. And you'll also see that we put the nitrogen absorption technique on here also. So here you can see at the low range of BET surface areas, you can see that our two techniques of IGTSEA and DBS really begin to excel for the low surface area materials. You'll see that our nitrogen absorption technique has a very large relative standard device deviation for some of these lower order materials, just because of the conditions of the technique and the very small size of the probe that it's using. So for the nitrogen absorption technique, which is a volumetric technique, these experiments are typically done at something like 77° Kelvin. These low temperatures and the vacuum conditions that are often applied for to the samples are going to be likely to dry out hydrates. If it's a pharmaceutical sample, it can easily collapse fragile crystal structures. And all of this is going to change the active bet that we're looking to observe. To mitigate some of this large relative standard deviation that we get for the nitrogen sorption technique, it is advisable to use an alternative gas, something like krypton or argon, some of the more expensive gases for these very low surface area materials. So some of the advantages of using something like DBS or using something like IGC is that you're determining the bet at ambient temperature and at ambient pressures. So you're analyzing your sample under the conditions at which they're going to be used in real life conditions. So we can analyze this with a range of different probes. We can find out its bet surface area as a result of hexane octane or something a little bit different like cyclohexane or ethanol. If some of these other, if some of these other probes were causing phase transitions or dissolving your sample. We can find out bet as a result of different relative humidities that we introduced. And this method has a very low relative standard deviation of about 1% with surface areas as low as 0.1 square meter per gram. So here we can see on this slide how this method performs at measuring the bet of a sample with a known BET sample. So this is a bet standard CRM170. This has a reported value of 1.05 square meter per gram. And what we did was we injected, we packed a column with 500mg of this sample. We then preconditioned our sample at 100 degrees C for 6 hours prior to running our experiment. Our experiment, our experimental conditions were then ran at 30 degrees C. And we did a range of different octane injections all the way from high surface coverages, from about 4, 40% surface coverage, all the way down to as low to zero as possible. So as Michan said before, it's important to find an area on this isotherm where we exhibit very linear behavior. And we selected that between 12% and between 36% of surface coverage. We did this because in this zone we can guarantee that we're going to get a mono layer forming of our octane profile molecule because we know the cross sectional area of our probe molecule and because we know the monolayer capacitor, because we know the uptake of this probe molecule, we can then find out its monolayer capacity times that with Avogadro's number and we get our bet surface area. And you can see that this was in very good agreement with its reported value. You'll see it's only.05.0.01 compared to 0.05. So we did the same kind of experiment on a different CRM sample. It's got a slightly higher value of 2.95 square meters per gram. And what we wanted to do was see how this would vary as a function of relative humidity. So we ran the exact same experiment at 0% RH and increased it to 50% relative humidity and 19% relative humidity. And here you can see we have a steady decline in its surface area. And we can attribute that to the water competing for some of these surface sites on the surface of the Illumina sample. So this is going to be blocking some of our octane uptake and it's going to make our apparent bet surface area decrease. We performed the same kind of experiment on a very different surface sample. So here we're looking at a sample of magnesium stearate and you can see the same kind of a trend is observed with relative humidity increasing. However, what this slide is showing is the clear effect that a pre treatment condition to a more extreme extent can have on the bet observed. You'll see that the bet for our last column here is now 5.95 square meter per gram. And this is because of the increased temperature that we did our pre conditioning step at. This is going to likely have increased. Going to remove some tightly bound water perhaps would have activated some high energy sites on this surface and also could have significantly changed the sample surface area by inducing a phase change. So this is just Showing how IGCSEA can be used to analyze surface area over a range of different ways, from pre treatment condition to dry conditions to wet conditions. When we apply this kind of methodology to a slightly different type of sample, something like microcrystallized cellulose, again we do a plot from high concentration to low concentration. Look for an area where we have have linear behavior. And what we do here is we plot its bet surface area as a function of relative humidity. Again, you'll notice that it remains fairly, fairly stable all the way up until about 60% relative humidity, at which point we begin to see its bet surface area decrease as a function of relative humidity. Now this could be due to effects of infiltration of the water into the structure, causing it to swell, or it could be due to us dissolving some of our sample, causing it to crystallize in a different way, causing a slight phase change, and that causing its surface area to change also. So you can see how this type of methodology can be applied to a range of different materials in order to analyze how they change in a range of different ways. So now that we have the surface area of a sample, we can analyze its surface energy. And this is the most common property measured using inverse gaps chromatography. The surface energy of a sample is quite analogous to the surface tension of a liquid. And it's just our sample's difference between its surface being able to react to a sample or its bulk. And from its surface energy, we can find out a load of different thermodynamic properties. So a regular topography on a surface is likely to have higher surface energy and higher activity. And we can relate this higher activity to other properties. So we would expect a higher surface energy material to have better wet ability or a better ability to react with polar molecules from its total surface energy, we can directly relate that to the work of cohesion. So this is useful for predicting the likelihood of a sample to aggregate or clump together. And then we can go a step further to see how a sample's likelihood to adhere to samples of different compositions. So it's likelihoods to lump or aggregate, but now in a mixture.

So.

Sorry, all solid surfaces are going to be innately heterogeneous. And this just means that there's going to be a difference between our highest energy sites and our lowest energy sites. And we expect these differences from our solids having a variety of different topographies, different functionalities on the solid surface, impurities in the formulation and adsorbates onto their surfaces. And with igcsea we can map the surface energy profile of a sample and we can translate this into a surface energy distribution. So our homogeneous surface are ones that typically have a flat energy profile and they have a narrow energy distribution. So that will be our surface in pink here. A heterogeneous surface is likely to be rough. It's going to have a curved energy profile we see in blue here and it's going to have a wide energy distribution. So we can also split our surface energy into its different constituents. We can split it into its dispersive component. And we do this from a range of different straight line out injections. And this is as a result of those van der Waals interactions. We can then find out its Lewis acid base component, so its likelihood to react to dipoles charged particles. And we do this by injecting two different monofunctional polar solvents, so something like dichloromethane or ethyl acetate and a range of different concentrations. And this will work out specific component. And we can add these two components together to get our total surface energy. So next we're going to be talking about how we can change a surface properties. We can change its surface energy as a total and we can change its distribution of the surface energies. So how do we modify a surface in order to influence its surface energy properties? We can lower a surface energy by physically covering it via waxing or coating, or we can introduce uniform functionalization by selenization or hydrogenation. We can go the other way and increase the surface energy by introducing high energy functional groups. We can oxidise the surface, or we can change the surface's topography or roughness by chemical etching or sand. So what we've done in this case study is we've taken a sample of mannitol, which you can see at the bottom right of the screen, and you can see that for one of our samples, we have uniformly cylinized the entire surface of the sample. So this is represented in our plot in red here. And you can see that its dispersive surface energy is noticeably lower than an untreated demonitol sample. So not only has this lowered the dispersive surface energy of this sample, but it's also made it much more homogeneous. So now its surface energy no longer changes as a function of the surface coverage of the probes that we expose to this surface. On the other hand, our untreated sample is comparatively heterogeneous. Here we have a large difference between our highest energy sites and our lowest energy sites. And this can perhaps be a bit more clearly observed with our surface energy distribution. Here we have a very wide distribution in the case of untreated and a very narrow distribution in the case of our selenised he Manitol. So with this differences in their.

Their.

Distribution of surface energies, we can relate this to their flowability. So the light how easily our powder will flow over each other when in a mixture. So here we can see now our selenide sample is our blue plot here. And you can see that it is a much lower energy for total flow required at these low air velocities. When we compare this to our untreated sample, you can see that it requires much higher total flow flow energy at very low air velocities in order to encourage it to flow. So because we have those high energy sites within this sample, it's going to have a much higher activity and a much higher cohesive strength and stickiness, which is meaning that it has a higher energy of flow and poorer flowing characteristics, which we have been able to predict just by mapping in surface energy and its heterogeneity via igcsea. So for our next case study, we're going to be looking at how we can measure the surface energy of a sample and predict that to something like dissolution rates. So in this case study, we had a sample of powdered aspirin and we ball milled this sample for 16 hours and then separated them into two groups. We have our as received group, which is just our unmilled group, and our milled group. So each of these groups, we separated them further into two different particle size fractions. So our smaller particle size fraction and our larger particle size fraction. All four of these particle size fractions and milled variations were pressed into 50 milligram pellets. Then their dissolution rates were observed by placing them in an acidic solution and by measuring their UV absorbance of that solution. And here you can see the same trend being observed in both particle size fractions. The milled aspirin samples were showing much faster dissolution rates when compared to their un milled samples. The same trends can be observed over both particle size fractions, which help support the conclusion that the difference in these dissolution rates was not just due to volumetric effects, but was due to surface behaviors. It's also important to note that this milling process did not have an amorphous phase change, which also could have contributed to the difference in the dissolution rates. So then to each of the samples we measured their distressive set energy. And you can see a clear difference for both of our milled Samples we can see we clearly have high energy sites as a result of this milling effect. This milling effect was likely to have exposed some high energy facets and activating some high energy sites on the surface. We can also see that the surface above coverages of 25% start to when as we're approaching 25%, sorry, start to get to the same value. So we've only affected about one quarter of this surface by activating those high energy sites. And that is what is encouraging it to get to some of those faster dissolution rates. So this study is really showing how we can use something like surface energy in order to predict useful physiochemical properties of pharma samples. So we can use this in order to find out how different suppliers offering different active ingredients or different excipients may have similar ways of producing these parts of the farmer material, but can have very large impacts on their dissolution rate just from the length at which they were milled. Next we're going to be looking how we can relate the total surface energy of our sample directly to the work of cohesion. And this can be very useful property to know about a sample, particularly when trying to predict the interaction between multi multiple bodies in a multi component system. So it's common for pharmaceutical formulations to have components with mainly APIs mixed with excipients. And the performance of an API of course isn't just linked to its dosage, but how heavily linked to its interaction with other ingredients and blends. So it's really useful to measure the work of adhesion between these active ingredients and between their excipients. So in a surface energy experiment we measure a sample in ISO, we want to measure each sample in isolation. So therefore its total surface energy can translate into the work of cohesion, which is how well a particle will will stick and interact with particles of the same composition. So there we can predict its likelihood of aggregation. We can then split the work of cohesion into its dispersive and its specific contributions which you can see here. Our dispersive is shown here with the D above and we have our positive contribution and negative contribution for our specific contribution. So we can split this for a excipient and for an active ingredient and then go on to predict our work of adhesion. And that's what we did in this study. We're looking at two different APIs. We have butersonides and we have salbutamol sulfate. And using an IGCSEA experiment we want to predict their compatibility with an excipient of our choice. Which in this case is alpha lactose monohydrate. We do this by packing a column of each sample in isolation. We perform an experiment calculating their surface energy of each separate sample. We calculate their work cohesion and then we go on to predict the work of adhesion between the API and with the excipient. So here we can see how the work of cohesion of each sample in isolation compares with the work of adhesion between the API and the excipient. In the case of busodenized, you can see that the most favorable interaction in the system is going to be between the API and the API rather than between the API and the excipient. Meaning that we would expect there to be aggregation within this mixture due to the energetic favorability of API API interactions. Therefore, we would expect this to have a poor blending performance. In the case of our other API salbutamol sulfate, we can see that the API excipient interaction is now more favorable than our API API interaction. And here we would expect it to be a much better approach performance with regards to blending. And we wanted to back up this prediction by measuring another aspect of this formulation. And we did this by measuring its blending performance by investigating its uniform content. Its content uniformity for both mixture and you can see we have a lower content uniformity. Relative standard deviation was observed for our salbutamol lactose mixtures, meaning that there was much less aggregation and much less lumping in this blend, which is in agreement with our adhesion prediction. And if anyone would like any more information, particularly on these adhesion studies at the end of the of this presentation, then I will happily send them over at the end. Now, I do actually have one more study, but I'm just going to go through it very quickly just as we get our slowly running out of time. So what we can do in our last study is that we can analyze a stability map between our different components that we can measure in a system. So we can map how likely a sample would be to react to temperature and react to humidity to see whether or not a sample would be stable in certain conditions. So the way that we did this in this study we took a sample of alpha D maltose monohydrate and we analyzed it at 0% relative humidity. We used Dycane as our probe molecule. And what we did is we did a series of dykane injections at the same volume, but we did it at a variety of different temperatures. The retention behavior was plotted against the inverse of the temperature. And typically we Would see plots like this would produce a straight line and the gradient of this line all yields the heat absorption of. For our probe molecule, which in this case is dikane. When we have phase transitions occurring, we can identify these phase points by the maxima or the minima of the inflection points that we observe. What we can then do is perform the same experiment at a range of different relative humidities. And from that we can plot a stability map. And from the regions of this plot, you can see areas in which you would expect our sample to see high stability as a respect of temperature and of relative humidity. And you'll see that an increase of only 15% relative humidity can have a really large impact on the glass transition temperature, the glass transition point of our sample and can lower it by something by nearly 30 degrees. And with that I'd like to summarize. So our surface energy analyzer using inverse cache chromatography is a very versatile technique. It can measure a wide variety of different samples, including farmer material. We can observe the BET surface area, look at some of the surface energetics, work out its uniformity of those energetics, and go on to find out some other characteristics of the surface. So I'd like to thank all of you for your time and hand back to Thomas for any questions.

Yes, thank you. Thank you, Connor. So I'm unmuting everyone. So you are able to ask questions or you should be able hopefully. So you, you need to unmute your, your microphone to ask questions. Of course you, you will get the presentation also from. By, by email. So the, the last presentation.

Yeah, I'm looking at.

The question.

Yeah.

Here, get a question. So can you detect multiple TGS glass transition in the single sample?

Yes.

You would be able to predict seeing multiple glass transitions or phase transitions within the same sample just by doing a similar methodology to something like this. If you were to notice multiple inflection points while doing this type of experiment, you could find different glass transition temperatures with different glass transition humidities. As long as these inflection points were and glass transition temperatures were quite far away from each other and weren't going to be interacting with each other or interfering with the gradients that we could observe either point of this, either side of this maxima and this minima. So as long as you have a low scanning temperature lot and you do a lot of different temperature steps for the same probe molecule in the same volume, then yes, you would be able to find multiple different blast transition temperatures. Thank you.

Other question from audience Surface energy. Would you recommend using Giga, Volpe or VOC scale.

Can you repeat that please, Thomas?

Yeah, the question is, for surface energy, would you recommend using De la Volpe or VOCG scale?

Oh, so yeah, the scales at which you want to be using for measuring your surface energy really just depends kind of on the temperature that you're doing your experiments at. So with our software that we use in order to analyze our surface energies, we have a range of different methods that we can apply. And the two most common ones that we actually use is going to be our Schultz method, Doris Gray method, and a hybrid between the two. Because our polarization methods and the way that we choose our selection process is more to do with the trends we receive in our data. And when we account for the temperature at which we ran our temperature at. So something like the Schulz method, we see that this was done using a range of different quoted values in the literature that were performed at 25 degrees C. So this was the surface tensions and the surface energy of a bunch of liquids under ambient conditions. Whereas something like the Doris Gray method, this has a temperature term accounted for it within. So this is more appropriate to use at something like elevated temperatures or if you're using some slightly harsh conditions. And the same kind of modeling applies when using the two methods that you mentioned earlier. It does really depend on the temperature that you're applying and the conditions at which you're performing the experiment in as to which one is the most most appropriate.

Okay, here I have a more specific question. I'm not sure that you. But I ask when I'm dealing with particles in the size range below 20 micro or smaller, the relationship between surface energy and dissolution didn't really follow the relation just shared. Would there be another factor that I can look into to predict dissolution rate powders from IGC measurements?

So porous materials, particularly ones with extremely high surface areas, are something that is quite challenging in the sense of IGCSEA problems. When we come across really high surface area materials is the fact that they take really long times to elute our probe molecules. So it's something to do with perhaps elevating the temperature or decreasing the amounts of the sample that we're packing into our columns in order to get really complete elution data is something that will not only improve the data, but improve the trends that you're seeing amongst them. But when it comes to really, really fine materials and really small materials.

You.

Would expect still for the higher surface energy. So stuff with really, really high exposed facets to have those same kind of dissolution rates, but I'm not sure if that would be expected from a sample that perhaps isn't as soluble in a certain solution. So different types of experiments in a range of different solvents may yield the kind of relationship that you're looking for.

Yeah. Yeah. Okay, Sangha, thank you very much. I think we can discuss your specific sample and application after the meeting. I've seen that we are five minutes behind, so I would like to thank you everyone for your time and also to announce already the next workshop. So it is already next week. Advanced Method for Determining Physical Chemical Properties of solid state material tales. That is next week and we have an IGCSEA week in the week of 23 till 27 of May 22nd. And if you would like to inform you regularly, please look also on our homepage you will find different webinars and workshop and different meetings we we have. So I would like to thank all the speakers. If you have questions, don't hesitate to send us an email for example to inforsurfacemeasurementsystems.com we will forward it also to Christian Lubert from Amerfor if it's a question for him. And of course we will pass all the information you would like to pass either to to Christian or one to our application scientist. Thank you for your time. We wish you a great day, a good afternoon and a good evening. See you next time. Goodbye.

Thank you everyone. Thank you Thomas thank you Christian.