 Hello everyone, thank you very much Vito for the kind introduction and also for the invitation to be here. It's a huge pleasure and this is a very good initiative of Lynx webinars. So and also, of course, thank you all the organizers for this event. So in today's presentation, I'm going to talk about crystalline and amorphous solid forms of pharmaceuticals. And in particular, I'm going to show you a couple of strategies that we may use in order to address the structural challenges. So starting a little bit with the introduction itself, so why we are so interested on drug solid forms, you know that to bring a new drug into the market is cost intensive, it take a couple of years in fact to have our new drug into the market. And then during this process, 40% of the drug candidates have poor biopharmaceutical properties, and up to 90% are too insoluble in water. And of course, these effects the availability of the drug. So in order to overcome this issue, we can take our drug molecule, in other words, our active pharmaceutical ingredient API, and then combine it with a second molecule, which should be also biological compatible. And we call it like a co former. And when we do this combination, we can prepare co crystals or salts. And here is just a representation of the difference between the two multi components solid forms. So in co crystals, you have your API interacting with your co former molecule via hydrogen bond interactions. And then on salts, what you do have is that one of the functional group of your API or co former could be acidic or basic. And then you have a kind of AC base reaction. And then you have your charged assisted hydrogen bond interactions. So now the point is that we wanted to locate the hydrogen atom position involved in these multi component solid forms, because this is important in case you wanted to patent a co crystal or a Paul or a salt, you needed to know and categorize these according to the specific multi component solid forms. Then of course, you can think about bringing our co crystals or salts into the amorphous state and we can prepare co amorphous. But also we have single component, single component solid forms. And these include polymorphs. We know that different polymorphs have different physical chemical properties. That means that we have the same drug molecule interacting in different ways or packing in different ways. Another category that we may introduce is poly amorphs more than one amorphous phase. And I hope I can convince you about these new world and the potentialities of actually including different poly amorphs in new drug formulations. I will show you a couple of case studies on this. So in order to actually categorize all these multi components solid forms according to the names I was telling co crystal salts and even the single components polymorphs, we needed to access the structure and also the physical chemical properties as well. But to access the structure, we can think about a couple of strategies that we may combine in order to get our structure of the material. So we can think in general as a chixel puzzle where we have a couple of pieces of the puzzle where we bring them together in order to get the full characterization. So we can start with the x-ray and with the x-ray we have information about atomic positions, the distances, the angles within the molecule. So in other words we solve the structure of our material. But then of course if we are thinking about the location of hydrogen atoms and considering that x-ray does not see hydrogen atoms so easily because there is only one electron involved, then we might consider using complementary methods and this could be using solid state nuclear magnetic resonance where you have information about the local structure and where you can also locate the position of hydrogen atoms and with these you can also study the interactions between the molecules. Of course, solid state NMR it does not provide you very sharp signals and you will see the reason why is that. Then we need to use computational models like for instance density functional theory calculations where we optimize our structure and then we calculate the NMR parameters that will be further compared with the experimental ones. So if you come up with these three pieces of the puzzle then you have your crystalline structure solved. But now the question is when we are working with amorphous materials what would be the strategy indeed to access to the local information or molecular level information of this material. So I will leave this question to the towards the end of the presentation. I will show you case study on this and now what I would like you is to show you a couple of case studies where the use of x-ray diffraction alone already is quite powerful and then I will move on for other couple of case studies where the combination of different techniques will be an advantage. So starting with the x-ray diffraction. So when we prepare a new drug or prepare a new molecule the first thing we wanted to know is the structure and it will be easier if we can grow single crystals of our material. So we have a powder in principle if you do organic synthesis and then we try to recrystallize our powder to get nice single crystals and this will be easier because then you can use single crystal x-ray diffraction analysis and then now we have softwares that already have routine they are very well routinely implemented where you can indeed solve the structure in an almost automatic way. So of course if you have some disordered in certain functional groups you may need the input of the user so you need to have knowledge on crystallography to actually solve properly the structure but we have quite good software that can do this for us. The problem is when we have materials that we try to recrystallize and grow nice single crystals and we are not capable of doing that. So then we are dealing with powders and we know these powders are crystalline so now the question is can we actually solve the structure of materials only having our powder sample? So in reality we can we can use powder x-ray diffraction analysis and we use the ebbing issue structure solution by powder x-ray diffraction. So there are a couple of steps that we need to consider in order to get access to the structure and these are just a summary of those. We need to index so in this case we wanted to know the cell parameters and the space group of your material then we do the poly refinement so we refine the cell parameters and the space group and we see if this refinement is correct and then after this we can move on to the structure solution and here we have a couple of algorithms we can use similar to the kneeling and also parallel tempering they are routinely implemented in softwares like dash and fox sorry and then you can use them for access to structure. So then you evaluate the results and now you are looking as well to parameters of the refinement so you will see the structures that would better match the experimental powder pattern so technically for each model there will be a theoretical powder pattern that we need to compare with and then it's not only based on these specific parameters but also because there is a figure of merit that is a value that we can look at it but also based on the structure itself so we need to open and have a look and see the structure so let's say you have two molecules that's supposed to be interacting via hydrogen bonding interactions and you have functional groups that are not pointing in a way where favors these interactions even though if the figure of merit is good then we may discard this specific model because we need something where the functional groups are pointing to each other in order to favor the hydrogen bonding interactions so based on this chemical knowledge as well then we can select the best structure then we do the ritual refinement which means that we refine all the structures the distance the positions and the angles and then at the end we have our final model structure so it's a couple of steps that we need to consider the other advantage of using powder diffraction is that we can usually when we prepare co crystals and salts in this case we use a technique called wall milling and this belongs to mechanochemistry it's a green unsustainable method and the advantage is that you can bring these jars as you see over here into a synchrotron and then you can follow in-situ experiments let's say you want to follow organic reaction even it doesn't only apply for co-crystallization let's say you want to follow that and you wanted to detect intermediates of reaction or you wanted to know if there is a phase transition between different crystallographic structures then you can access these very easily with the synchrotron so then that is a combined package of these two so now moving to case studies that I wanted to show you we started with barbiturate and melamine co crystals so this project is being developed in collaboration with the University of Cambridge with Anna she's in charge of this project so we have been preparing a couple of co crystals of barbiturate and melamine and we know from the literature since 1919 that there are the preparation of different polymorphs of these specific co crystals in different shapes like rosette and precludes or linear tape and it is in certain cases it's very easy to isolate the rosette form but the other ones are not so easy so then we decided as well to investigate how will be this polymorphic transitions using powder diffraction and especially because there are no structures on the linear tape for instance rosette we have one for a particular system which is the tazine with barbitol but we don't have for the other polymorphs and the same happens for any combination between barbiturate and melamine derivatives so what Anna did was to combine tazine with barbitol in a setonitrile and then she got one polymorph that has the rosette shape and then we investigated further what would happen if we use now instead of a setonitrile water so we got a different form from form A and what is interesting is that if you take a form B and add a setonitrile then you go back to form A and when you take form A and add water you go back to form B so that is a reversible thing so at this point we didn't know the structure of form B so we went further and solved it by powder x-ray diffraction and here is how it looks like a ritwell refinement plot where you have your experimental data on these blue circles you have your theoretical data in red and the difference curve is in gray as you can see it's quite flat which means there are no big variations between the calculated and the experimental data here is the space group as you can see and here you have the structure which actually is a linear tape so we could isolate this polymorph and even solve the structure so now moving to case study number two about Exito kinetic investigations of co-crystallizations so this was the work that I was doing while at BAM Federal Institute for Material Research and Testing we have been working with Carbamazpin and we try to produce different co-crystals using different conformers you see here there is the two four to five and two six the hydroxy benzoic acids the difference is that you have the hydroxy group located in different positions in the aromatic ring and we wanted to investigate if there was an influence of the type of the conformer in the co-crystallization rate and as you can see here we have the plot of all the kinetics of this co-crystallization reaction and as you can see for the two four and for the two six we have a quite fast co-crystallization but the four to two five is a little bit less faster than the other two and I just wanted to point it out because these are Exito studies each data point corresponds to an independent experiment means that we mill independently each of the data points and then of course we also measure independently via PXRD each of the materials and here is how it looks also the ritual refinement for one of the co-crystals so we solve all of them by powder diffraction because it was not easy to get the single crystals of these materials and here is how it looks like when we do the quantitative phase analysis so if we have so during each of the data points we basically have a powder pattern where we can refine the structure of our co-crystall that we saw previously and we also have access to the structure of the starting materials so if we find the three of them here you will see there are tip lines with different colors corresponding to the different materials and then you can quantify how much do you have in each of the data points uh respecting to the product or starting materials then we investigate a little bit further looking at the structure itself what could be the reason why for the two five we have a little bit less like the rate could be less than than the other two so as you can see here we came up with the rational so we have a look at the carbon spin itself we can see there are interactions between the amide groups as you see over here then for the co-formers the two five we also have this kind of motif over here and then we have an OHO interaction so when we get our co-crystal between these two components then we see that the motifs initially present on the starting materials are no longer on the product so we have different motifs so this could be one of the reason why it could be delayed a little bit that is a tendency that is a trend but we cannot make sure that this will be the reason why it will be less fast than the other ones but so we can see that is a trend so if you look to the two six now uh we don't keep the same motifs but instead we have a direct interaction between the amide and the carboxylic group over here but now when we move to the two four that we consider to be the fastest one then we see that the initial motifs on the starting materials are preserved on the product and we have actually an OHO interaction bringing together these two parts of the carbon must be in and hydroxybenzoic acid so now moving to another case study in-situ measurements so we saw polymorphism ex-situ now in-situ uh we have been doing this project in collaboration with the University of Cagliari and the University of Montpellier we were performing a chlorination reaction of the three ethyl 5-5-D methylydentroin so here is what we did uh we did we used carcel apocloride so we did the reaction in situ and then we have a probe for Raman and then we also uh collect x-rays so we did both at the same time there's another advantage that we can do so in this specific study we were interested on seeing the rate of this reaction when varying the size of the ball mill used so we go from two millimeters to eight millimeters here you can see uh some of the examples two of them uh when we use the two millimeter balls as you can see over here just to guide you do you have the time over here of milling and then you have the starting materials so as you can see there are reflections of the starting materials are all the way until completing the 60 minutes and this also happens after two hours because in this particular case the reaction was not complete and you can also see for instance the reflections of the products coming up to here so one reason why is that these balls are quite small the surface area is also smaller and the efficacy of mixing the components inside of the jar is not so good so therefore of course the reaction will be incomplete in these cases now if you move on to the four millimeter then now the the scenario changes a bit because now we managed to complete our reaction even after 30 minutes as you can see the reflections coming down like and disappear and then we have the reflections of the product over here so if you now plot all the kinetics of these reactions depending on the ball size used you can see here for the two millimeters we have an induction period where the reaction didn't occur for over six to seven minutes and then we start to see some product formation and then for the four it goes faster five six and eight and of course we have a speed increase with the increase of the ball size and of course the efficiency of mixing the components is also higher in those conditions so now moving to another case study where we were able to detect reaction intermediates this can be very useful as well so we were investigating a conagall-nagall conversation between the nitroganzaldehyde and melanin trial starting materials and we were performing the reaction in different conditions so we use nitriding which means absence of solvent and then we use more polar solvents like ethanol and DMF and then a more a polar solvent like octane what we were able to do is that in nitriding and octane we were able to favor the reaction intermediate which is this product over here as you can see that is an OH group attached here in this position and when we use more polar solvents then we favor the product which is basically elimination of water and then the formation of this double bond so it was interesting because we were able to kind of slow down a little bit the reaction we perform at four hertz but we also decrease it until we could actually isolate the intermediate and here is the powder pattern as you can see so the interesting part is that we were able to solve the structure of the intermediate even if there is a presence of contaminations of starting material as you can see these stars over here they correspond to the nitrobenzaldehyde so we could quantify as well how much we have it from this starting material so basically what we did we exclude these peaks index and solve the structure based on the peaks that we are sure that are from the product so now moving to our cheek cell puzzle we have been seeing a lot of case studies on using only x-ray now I'm moving for other case studies where the combination of solid state NMR and DFT can be an advantage so but before even going into the case studies I would like to briefly introduce you what is solid state NMR very briefly so there are a couple of interactions between the nuclei and an external magnetic field and there are two particular interactions dipole-dipole interactions means interactions between two nuclei and the external magnetic field and that is also the chemical shift which is the interaction between the nuclei and the external magnetic field through the electrons so these two interactions are involved in an effect called an esotropy and what is an esotropy so if you have your powder sample your crystalline powder sample you don't dissolve it so you place it inside of a sample holder and now you place this router that contains your sample inside of the NMR equipment and you measure an NMR and you will see that each molecules on these crystallites they are positioned in specific ways so what happened is that the interaction between each of the nucleus in the molecule will depend on the position of the molecule and because that happens you will see an envelope like this so you don't see defined signals you see like a convoluted signals so if you now look to the case of the liquids so in liquids we don't have static conditions like we see in solids so the molecules are moving quite fast and therefore the interactions with the external magnetic field are no longer depending on the orientation of the molecules and therefore the an esotropy effect does not exist and then you get these nice signals in the NMR spectrum so the idea now is that how can we bring something like this into something closer like we have on solution NMR so we can use a technique called magic angle spinning and these actually allow us to suppress this an esotropy effect that comes directly from these two components like these two interactions so this equation accounts with the dipole-dipole interaction chemical shift and esotropy and then this is the router where you have your sample so if you now you take your sample and place it inside of the NMR equipment and if this router is placed precisely at 54.7 degrees with respect to your external magnetic field then this term comes into zero and then you can suppress this effect but of course this is not so simple as it seems so you need to also account with other factors one of those being the frequency of spinning so the frequency of spinning must be larger than the magnitude of those interactions so in order to actually suppress as much as possible this effect so now just very pictorically if you have your sample and if you don't spin it there you see this envelope as I was explaining before now you start to spin your router slowly and you will see that you start to get some sharp lines so these lines are coming from the spinning side bends from the fact that we are rotating the the sample holder but then of course we must have our isotropic value so the value from our nuclei must be placed somewhere so in this way we cannot distinguish where it comes but then as soon as we started to spin it faster and faster this spinning side bends goes towards zero basically and then you can see the isotropic value that you need so now moving to the case studies this was a project that was developed during my PhD in University of Lisbon and in Aboedo so we have been working with asalic acid we combine with different conformance to produce salts and co crystals and this is only one little example where we have as like acid combined with morpholine we managed to produce the salt as you can see here from this crystal structure but then what was interesting is that one of the protons it was shifted towards very low value so low field means like 20.1 ppm or 19.3 this is very a typical for this type of materials then we decided to investigate what could be the reason why it was soldish shielded this proton so we went back to the DFT calculations we geometry optimized our structure then we calculate the NMR parameters and then we found out that the proton is indeed sitting in between the interaction basically so this is one reason why it's also soldish shielded so this is a way where you can track like unusual positions on hydrogen atoms so then about direct protonation in hydro atoms of course you you can also perform acid NMR in other nuclears you can do in carbon and nitrogen and here is an example we were working with adamantilamine once more we were preparing several salts with different amino acids and several other acids and as you can see here adamantilamine alone is not protonated in this case and as you can see has a chemical shift of 48.0 but now when it's protonated it shifted towards lower field so as you can see you have the signals places over here so this is an indication now that we have a protonated system because here the mine is protonated you can also do the same using directly a nitrogen and then you can detect the direct protonation you can see the same thing so when it's not protonated has these minus 317 and now when it's protonated it goes to minus 323 in average then of course you can also detect weak interactions and here is when we can use the 2D NMR so we have been looking at each of the different nuclears we can use on on organic systems but now we can also do see like a proton or carbon etrocorrelation and then with these we can even identify weak interactions this is just an example so this is DOHO interactions and we can spot it out even here on the structure then we have NH3O interactions as well and now the interesting part comes for this aliphatic proton that is actually engaged on nitrogen bond interaction according to our X-ray structure and what is interesting is that we can also see the chemical shift of these which is different from the one that is not involved in hydrogen bond interaction that is placed in the higher value higher field so now for the last case study of this series of crystallographic structures this is another study that was performed in collaboration with Umbolt University where we were able to collect solid state NMR data so we were working with salicylic acids and once more we solved the structure by powder diffraction so here we just omit the energy and atoms just for clarity but you have the space group you have the structure solved but now the most important question here is that when we are doing this we are not considering the fact that this proton may jump to the amine group and then we have a salt in fact if you take the structure and then you put the proton in the other position you would we will not see the difference by powder X-ray diffraction this is why we need to go further so then in order to understand if this is a co-crystalline salt we went for our periodic DFT calculations we did a geometry optimization first to see if the structure was correctly solved and here you can see the optimized in blue coming directly from DFT and non-optimized one coming from the ritwell refinement as I showed you before as you can see the positions are correct the angles and the distance are also correct in the structure and now if you expand a little bit further on the hydrogen atom in the hydrogen bond interactions you can see that in fact the proton that was initially located on the carboxylic group it migrates for the nitrogen of the midazolium so we have a salt according to our calculations then to validate these of course we use the solid state NMR and here is just a representation of the two so we have the proton where we can correlate isotropic chemical shieldings which are the values we obtained from the DFT and then when using this equation you can calculate the theoretical chemical shifts which are the ones we in fact compare with our experimental data so here is just the correlation between the chemical shielding and the chemical shift it is closer to one when when those values are like this means that we have actually a good correlation and here is just a representation of the proton NMR from this specific material as you can see the resolution was not so good in this case we're more spinning at 10 kilowatts if you spin it higher we will be having a better resolution as well on our data but the fact that we use the DFT calculation allow us to identify the signals and even spotting out the CH pi interaction which is over here so now we go back to our jigsaw puzzle we have been seeing a couple of case studies where we can solve the structure of crystalline materials how about amorphous so we can still access that information and we can still use x-ray but not in the normal way so in x-rays you are measuring blood depraction but we can also use what we call a pair distribution function analysis which is a probability of finding a pair of atoms at a given distance so this also gives you information about the local structure we are going to see in a second how this can work but then of course if you only use pdf you don't have a model of your structure so we need to come up with another method that can help on elucidating this which could be molecular dynamic simulations this is what we have been doing then we get our molecular models we can actually calculate the theoretical pdf and then we compare with experimental one so here i'm going to show a case study for that but before this just briefly about pair distribution function analysis so here is how it looks like the distribution when you get the molecule you have atoms at the specific distances so then you have the probability of having those atoms at those specific distances and then you have a plot like this and usually to collect these data we need to have high energy so we need to have extremely low or short wavelengths so we need to go to the synchrotron facilities and when we do that we collect what we call a total scattering so we have the BRAC contribution plus the diffuse contributions because BRAC has information about the average of the structure and the diffuse component or the pdf at the end will have the information about the local structure so we need to get a high resolution data as possible to indeed get a better pdf so you have here how it looks like the curve when we go to the synchrotron and then we get our s of q then we correct this data the background and then we sort of have a scattering factor so we have a kind of transformation of these s of q into a f of q through an equation and then from these data because we are in a reciprocal space we needed to bring it back to the real space we employ a Fourier transformation and then we get our pdf data so now moving to the case study and the last case study of this presentation it's about hydrochlorothiazide so this is a project I'm currently developing here at the University of Copenhagen so I've been working with this molecule and the idea is to produce different amorphous forms using different pathways so we consider the thermodynamic pathway where we can use spray drying and then we get one amorphous phase after spray drying you can use the quench cooling or you can use the kinetic pathway where you use ball milling so what is interesting in this study is that when you use the spray drying we obtain one amorphous phase called polyamorph one with a glass transition temperature of 88.7 so the glass transition temperature is the temperature at which a material goes from the glass to the rubbery state so if we have detected these on our differential scanning colorimetry that also means that we are in the presence of an amorphous form so we did this and we did these for the quench cooling material and as you can see there is a gap of 40 degrees which is quite big and then when you do the ball milling you don't have such a big gap but we have also a different glass transition temperature and here is how it looks our differential scanning colorimetry data as you can see different tg's but also different recrystallization temperatures now the question is are the crystallographic form after the recrystallization temperature the same or different so then the answer is it is in fact the same so you can see our hydrochlorothiazide crystalline here in green and then if you compare with the other three of them that were submitted at different temperature conditions after recrystallization point you can see that the diffraction peaks match very well our crystallographic hydrochlorothiazide so now what we did afterwards was to investigate how would be if we quench cool now our spray drying sample are we going to get the quench cool poly amorph then in fact we got it but now the things got interesting when we tried to spray drying our quench cool material which is called poly amorph too then at the end we got poly amorph too still so these puzzled us a little bit because now you are taking our amorphous material put it back in solution spray drying and then you expect that you will get back to poly amorph one but then we didn't as you can see here we have a glass transition of 119.1 and the recrystallization temperature is also the same for the quench cool material so that means that we didn't change this form then we also ball milled poly amorph two and we got poly amorph two still and quench cooling poly amorph three we still get poly amorph two so it seems that this phase seems to be the most stable amorphous form produced now the question is that do we have any degradation on our sample that may not be exactly this poly amorph so we did HPLC and then we confirmed that there is no degradation on these materials so what we did as well was to ball mill our spray drying sample and spray drying our ball mill sample and in both cases we got two amorphous phases and this is something quite interesting because usually you get like a phase separation when you have a polymer with a drug and then you see two TGs but not in a single component type of material so this raises a quite interesting point so now just very briefly to show you the real results from all the skin you see quench cooling spray drying we have our glass transition of 86 87 then it goes to 119 then when quench cooling the ball mill we also have a changing on the glass transition temperature towards the quench cool samples and then when we sprayed right the ball mill or ball mill the spray dry we have two glass transition temperatures we are now investigating these a bit better and these will come up with another publication that we are preparing because this is quite something surprising it's not supposed to happen or at least there is nothing reporting this type of behavior so now what we did as well after all these measurements we also evaluate the relaxation properties of the different amorphous forms so we measure for the spray drying ball mill sorry quench cool and ball mill and as you can see they have different relaxation times so we have 12 4 4 the spray drying which has higher mobility that it goes faster towards relax quite quickly then we have the quench cool one that it has a longest relaxation one and then we have the ball mill which is the intermediate one so that is a correlation between the glass transition temperature and the relaxation per meter low glass transition temperature means low relaxation per meter and high glass transition temperature means a higher one relaxation so now that we know that we have these different properties so different glass transition temperatures different relaxation properties now we wanted to evaluate what happened at the structural level so we went to the synchrotron to collect the pdf data of all our samples and surprisingly we get the same pdf for the three phases so here it's just an example of the three so we collected the data and as you can see we have resolution until eight angstroms roughly and the green one is just the modeling one i'm going to show you in a second how we did this but what is important to highlight the three are similar and there is no other signal that can be detected after eight angstroms so this also means that this is a completely fully amorphous completely deorganized there is no kind of seems to be no no kind of kind of clusters that could in fact repeat in a certain point or at least that we could detect it via this method and then what we thought now that we were a little bit shocked about the fact that you are similar we went back to the structure and then when you look at the structure we have two places where there be variations could be these diadral angle that can vary can rotate it and we also have this configuration here that can be up or below the plane so we took these different molecules and we simulate the pdf of all of them considering the rotation of this angle and as you can see they look pretty much similar there are only very little differences going on which means that if you sum up all of these you ended up getting a pdf that looks like this but the fact that we don't see the differences here that doesn't mean that we don't have different populations with different type of diadral angle distributions so then we decided to evaluate that we went a bit further so we did the simulations in this case this is the example of the spray drying and this is of the quench cool but for the quench cool we also consider a supercell that we actually create defects to get it into the amorphous states or emalted and that we cool it down but we also consider an approach where we place our 100 molecules into a box randomly we increase the temperature until the melting and then we decrease until 10 degrees to get our final material and then on the spray drying we have also a box of 100 molecules then we add a thousand ethanol molecules we spray dry with different velocities where we can remove five molecules or 10 molecules of ethanol at the time at every 10 or 100 picoseconds so we have a data set of four type of approaches and here is just the results of the simulated pdf coming up from these two molecular dynamics experiments and as you can see they are also very similar until eight angstroms so these us give us some kind of confidence that the models are approaching the reality because at this range we not only see the intermolecular components but you also see the intermolecular components that is a mixture but we cannot differentiate what are exactly the intermolecular components at this point so then we were moving and then we analyzed the diagonal angle distribution that was internal to the molecule so we know we have our hct over here we have this angle and this one on the sulfonamide so we have a 2d plot and as you can see for this particular angle when we spray drying then we favor the negative value and then for this one we have a 50 50 distribution of positive and negative angles because this is a group that is freely rotating so now if you do the quench cooling and if you plot the distribution now you have kind of a good surprise because now you don't have a favor of the negative now we have a 50 50 distribution so now we can see okay we have some kind of differences in these different amorphous materials now we go back to this scheme because i wanted to highlight this particular transformation you remember i said quench cooling spray drying sample will yield the quench cool and spray drying the quench cool we will keep the same quench cool sample so now we wanted to understand if the glass transition can be somehow correlated with this diagonal angle distribution and if that is true then in in fact if we quench cool computationally if you quench cool our spray drying material then we should have the distribution closer to the quench cool and when we sprayed right our quench cool material you should have the distribution closer to the quench cool and in fact the good surprise is that we have the distribution of the quench cool so if you take the quench cool you had the ethanol you sprayed right then you still have 50 50 and now if you take your spray dry sample and you quench cool it then you have 50 50 distribution so it seems that this diagonal distribution may play some kind of role in the understanding of these different properties of those type of amorphous materials so now going back to finalizing the presentation going back to our jigsaw puzzle i've been telling about these two specific pieces of the puzzle but of course we need to add much more pieces so now here i'm just adding these two they are not directly related with structural oxidation but they give important pieces of information to understand the molecular level of distribution of these materials but also we need to consider other techniques could be electron pdf or other methods that we may need to explore in order to get more details about these amorphous materials or the structure of these amorphous materials so now to finish the presentation i have a couple of people i would like to acknowledge of course tomas because he has been a lot of support during this journey on the polyamorphism project he has been a good a good journey and very good data that we are getting out of this project then yuka also for the scientific support all people involved in this project so we have anas matzen here from a department of pharmacy kirsten from the chemistry department she also works with pdf and we have also anas larson and olivia and i also wanted to acknowledge eric chaperson from desi so he has given us support to collect all these data nordic pop for funding our trips to the synchrotron and also for the scientific divulgation and also a dff of course for funding my postdoc position and i also wanted to acknowledge francisco from bomb in berlin for all the support he has given to me during my time there working with all these very nice projects all people involved on those projects here is just a nice picture of all of us before corona here so we have been in the christmas and we went for a nice dinner together i also wanted to thank my supervisor from lisbon my pg supervisor treza she also she has also given me a lot of support doing my phd then of course martin where i've been spending five months and learning how to solve structures by powder diffraction it was a really really nice time in frankfurt then of course my co-supervisor from avaro and all the team for supporting also these solicited nmr results that i've been showing the funding from portugal and also the collaborative people so people that were collaborating in all the works i've been showing good room from on both university anna julian pet from university of cambridge francisco delogu and maria cata from university of cagliari and evalina colatino from university of montpellier and then of course i want you to thank you for your kind attention as well