We have reached the maximum registration capacity of 500 seats for this webinar series. Due to popular demand and interest, folks can still register by sending an email to sivakumar.sekharan@xtalpi.com expressing interest and provide their full name, organization, country, job title and email address to be added to the waiting list and subject to drop outs and cancellations, their names will be added to the registration portal.
For busy professionals working in pharmaceutical solid form, we have created a 7-week webinar series to distill state-of-the-art solutions to key challenges. Each webinar will feature a leading expert who understands your daily workflow and will share his/her real-world experience in successfully supporting preclinical and clinical projects. The series will catalyze discussions aimed at ensuring a stable form is selected with optimal bioavailability and manufacturing properties.
We have reached the maximum registration capacity of 500 seats for this webinar series. Due to popular demand and interest, folks can still register by sending an email to sivakumar.sekharan@xtalpi.com expressing interest and provide their full name, organization, country, job title and email address to be added to the waiting list and subject to drop outs and cancellations, their names will be added to the registration portal.
We have reached the maximum registration capacity of 500 seats for this webinar series. Due to popular demand and interest, folks can still register by sending an email to sivakumar.sekharan@xtalpi.com expressing interest and provide their full name, organization, country, job title and email address to be added to the waiting list and subject to drop outs and cancellations, their names will be added to the registration portal.
The Role of Mechanical Properties of Drugs on Tableting Performance – Mechanism and Engineering
The Role of Mechanical Properties of Drugs on Tableting Performance – Mechanism and Engineering
Q1. Typically, the tableting issue comes at a later stage of drug development where the solid form of API may have been fixed, when do you see the crystal engineering can be applied for the typically pharmaceutical compound? Should we elevate the tableting earlier in the dev?
Yes, we should evaluate the tableting performance of drugs as early as possible and use that information to guide the solid form screening and API engineering. Useful tools include nanoindentation (only single crystals are required to quantify mechanical properties) and compaction simulator.
Q2. Do you know some solutions or experimental work on AI based formulators for tableting APIs and Excipients? If yes, could you please indicate some players?
A few major pharmaceutical companies are working on AI based formulation development approaches. A key challenge with this approach is to have a very large set of high quality data (training set). It will take major efforts to have such data available. Some of the work we propose to do through the CIMSEPP platform will also contribute to this.
Q3. Co-crystal formation is a great approach to improve mechanical properties. Is it possible that making a co-crystal may change solubility so the dissolution profile-kinetics? What is your opinion on making a co-crystal changing the crystal structure (e.g. avoid exposing polar group on the outer face) and limit the sticking?
This is something we are actively working on. Cocrystals can be used to improve drug dissolution. This has been well documented in the literature. We recently showed that cocrystal can also be used to alleviate punch sticking (https://doi.org/10.1021/acs.cgd.0c00492).
Q4. In early drug development, enough amount of material is sometimes unavailable. What investigation or analysis do you recommend to mitigate mechanical properties as early as possible?
We fully understand the challenge of material availability in an early stage of development. Our approaches to address this challenge include 1) predicting mechanical properties from crystals structure (e.g., plasticity); 2) nanoidentation (only single crystals are required); 3) compaction simulation using an instrumented simulator.
Q5. You mentioned cohesive energy of crystal can affect the bonding strength. Could you talk about if surface bonding effects are more important than the bulk cohesive energy of crystal.
Surface energy is more directly relevant to tablet tensile strength than bulk cohesive energy, but they are also intrinsically related. We recently observed that dispersive surface energy of crystals played a more important role in tablet tensile strength than polar surface energy (DOI : 10.1007/s11095-020-02856-2). This is a topic that deserves further research.
Small Molecule Structure Determination by MicroED
Small Molecule Structure Determination by MicroED
Q1. what would be the 'minimum' configuration of instrument for microED and its estimated associated price?
The minimum configuration needed to solve easy targets like paracetamol and biotin, vs the minimum configuration needed for a more difficult target like teniposide can be a big increase in price. I would say that the most important things are (1) a good camera, (2) a very accurate goniometer and (3) a coherent beam, preferably with a FEG or an XFEG, through structures have been published using a LaB6 source. The camera should have rolling shutter mode with a fast readout so that you don’t miss reflections. A thick scintillator will ensure you can accurately measure low intensity reflections. A detector with high gain is also preferable. We saw huge improvements in data quality when upgrading from the CETA to the CETA-D, and we suspect this upgrade was the make-or-break element for many of the structures we have been able to solve. You should contact microscope and camera providers for up to date cost estimates or consider outsourcing to a CRO.
Q2. How long does a typical data collection take?
The actual data collection for a single dataset is very fast, typically 30-60s for small molecule samples rotated the full 120°. The most time consuming part of the process is the time associated with crystal identification/targeting, moving to the crystal and centering in X,Y and Z. We typically spend ~2 hours per sample to ensure we get many, high-quality datasets that can be combined to get a final dataset with high completeness and multiplicity.
Q3. Serial Electron Diffraction combined with 3DED/MicroED can be used to do phase identification quite well. It is possible to crystallize on the grid?
(1) The easiest way to phase microED data is by ab initio methods (SHELXT and SHELXD) or molecular replacement for more difficult cases (PHASER). Our in-house team has never failed to phase a small molecule dataset by these methods as long as the sample has had diffracting crystals. (2) It is possible to crystallize samples directly on the grid, but we have found that this often leads to crystals that are too large for data collection. We prefer to crystallize elsewhere and mechanically grind these crystals to produce the nanocrystals required for optimal data collection.
Q4. What is the resolution of MicroED compared to X-ray crystallography?
Resolution is generally limited by the quality of the crystal, not the quality of our microscope and/or camera. The highest resolution microED structure solved to date is of a peptide solved to 0.6 Å (6KJ3). Our highest resolution structure is of biotin at 0.85 Å, but generally client samples achieve ~1-1.2 Å, with some as good as 0.9 Å and some as low as 1.7 Å. We suspect that crystals that are unable to grow large enough to perform single crystal X-ray diffraction experiments may be more likely to contain lattice defects and disordered regions that limit the resolution of these crystals.
Q5. Is the data monitored continuously or intermittently in MicroED software?
Data is uploaded to our online viewer continuously and real-time raw data transfer options are available to our clients. For clients purchasing data collection only packages for multiple samples in one day blocks, we do not currently process the data while it is being collected. We do monitor data quality and can adjust collection parameters if we notice a problem such as radiation sensitivity or ice contamination. Clients may also reach out to our microscopists during collection and request data collection adjustments based on a list of approved settings. For clients purchasing small molecule microED research studies in which NIS will be processing the data and providing an initial phased structure, we do process the data “on-the-fly” to ensure that the highest quality data is being collected and that we efficiently use our microscope time.
Q6. Do you think microED will ever replace single crystal crystal XRD in routine structure solution?
No. Single crystal XRD is a robust technique that is significantly easier to implement in terms of data collection and data processing, and it is significantly less expensive. If you have crystals large enough for single crystal XRD, we recommend that you use that technique. MicroED should generally be reserved for difficult targets for which large crystals are not available.
Q7. Is it possible to determine mesocrystalline structure by SDED?
Ideal samples for microED are composed of single nanocrystals. If a mesocrystalline array can be broken up into single crystals or if data can be collected from regions of uniform crystallinity, then this should be possible. Our beam diameter is very small, ~600nm, and may be able to find such regions.
Q8. How about the completeness for the full dataset collection?
We are limited by our goniometer such that we can only collect ~120° of data from a single crystal (assuming minimal radiation damage). For high symmetry crystal systems this may be enough to reach 100% completeness. Only 90° of data is required for full completeness for orthorhombic systems. For lower symmetry crystal systems, such as triclinic and sometimes monoclinic depending on how the crystal is orient relative to the rotation axis, 180° of data is needed to reach full completeness. This is why we generally collect data from many crystals and combine data from the best ones. Random orientation of the crystals on the grid can allow us to sample unique regions of reciprocal space and boost completeness. We have seen cases though where preferred orientation of the crystals on the grid limits our ability to generate high completeness. This generally does not hinder structure solution but it can lead to streaky (anisotropic) maps and less reliable quality metrics. We have also found that increasing multiplicity through combing data from multiple crystals is very beneficial for increasing the quality of the data and can sometimes be essential for phasing by ab initio methods.
Q9. Can you show/comment about the validation of MicroED measurements?
We meticulously calibrated our camera length(s) and elliptical distortion with powder diffraction from a metal alloy. We also monitor this over time to ensure that these calibrated values are still accurate. We also determined the structures of three compounds that had already been solved by X-ray diffraction (paracetamol, progesterone and biotin) to ensure that our structures align with previous X-ray studies. We can also anecdotally share that one of our clients shared with us that he initial thought he had not opened his microED structure correctly because it lined up so perfectly with a structure he had previously solved by X-ray diffraction. The proof is in the high-quality maps/models; don’t let the high Rs discourage you too much.
Q10. Do you have in-house expertise in indexing powder X-ray diffraction data?
At the moment we do not, but we are currently working to build this expertise to better ensure successful outcomes for our clients.
Q11. Remark from two participants: Might be worth passing on to Jessica that I believe absolute configuration by 3D-ED is possible, as demonstrated by Lukas Palatinus in his Science paper from last year. Platinus et al. Science 2019, Xu et al 2019.
Yes, the 2019 Science paper from Brázda, Palatinus and Babor caused quite a lot of excitement for the possibility of accessing the absolute configuration with an electron diffraction method. I would like to point out that the technique described in that paper is not strictly speaking “microED.” MicroED and the method described in Guene et. al.’s 2018 Angewandte Chemie paper are based on the rotation method of data collection where the beam is constant, but the crystal rotates relative to the beam in order to sample different parts of reciprocal space. The 2019 Science paper describes data collection using precession electron diffraction tomography (PEDT), which involves tilting the beam during collection. These data cannot be processed in standard single crystal X-ray diffraction programs such as DIALS and XDS because those programs were built for data collected using the rotation method. We do not offer precession data collection at NIS and therefore cannot offer absolute structure determination based strictly on the diffraction data.
Dome vs. Basket Paste Extrusion Techniques for Solid Formulations
Dome vs. Basket Paste Extrusion Techniques for Solid Formulations
Abstract
A water dispersible granule (WG) formulation is a dry, solid formulation containing active ingredients intended to be diluted/dispersed in water prior to application. If the co-formulants are properly selected, WGs typically have desirable physical and chemical stability. In addition, storage and transport of dry formulations are inherently easier than liquid formulation types. WG formulations of agricultural products offer the added advantage of low dust and safer handling compared to wettable powders (WP). WG formulations can be prepared via several types of granulation processes, with paste extrusion being the most common in crop protection products. Granules created by paste extrusion generally offer improved resistance to attrition, and higher and more consistent bulk density compared to WP formulations. Rapid and complete dispersion in a variety of water conditions are key performance parameters used to evaluate WG formulations. Additionally, quality criteria for WG products include dispersion, bulk density, suspensibility, wet sieve, and moisture sorption. *Adapted from Abstract of: Zukowski, S. R., et al. (2020). "Performance comparison of dome and basket extrusion granulation." Chemical Engineering Research and Design.
Biography
Sam has worked in Formulation Development and Formulation Process Development for 3.5+ years at Corteva Agriscience (formerly Dow AgroSciences) in Indianapolis, Indiana (USA). He received his PhD in Physical Chemistry at Purdue University in 2016 (Thesis title: “Characterization of Water Mediated Hydrophobic and Ionic Interactions Using Raman Spectroscopy”). While at Corteva, Sam has primarily worked on solid formulations, such as coated granules (GR) and water dispersible granules (WG), leading both formulation and formulation process development projects. Currently, he leads several liquid formulation developments, including emulsifiable concentrates (EC) and suspension concentrates (SC), but continues to be heavily involved in solid formulation development throughout Corteva.
Next Generation Models for Crystal Morphology Prediction
Next Generation Models for Crystal Morphology Prediction
Q1.How does the choice of intermolecular energy model affect morphology predictions? Do systems containing intermolecular hydrogen bonds, or strongly polarized systems represent difficulties in terms of correctly modelling intermolecular energy contributions?
The models depend on the bond energies at kink sites (these are the most favorable docking sites on the crystal surface for solute molecules). The bond energies are calculated from the atom-atom force field – so it is very important to have a good force field in order to obtain good predictions. One way that we test the “goodness” of a force field is to compute the lattice energy (or sublimation enthalpy) for the crystal and compare against experimentally measured values.
Q2. How expensive is an ADDICT prediction for a given crystal structure?
Once the input information is loaded (cif file and mol2 file) the program takes about 5 minutes to make its predictions.
Q3. Is the steady state crystal morphology invariant to temperature?
No.
Q4. Is the growth rate for NaCl an average of all facets?
All facets are identical for NaCl therefore, we calculated the growth rate for one of them which gave us the growth rate for all of them.
Q5. Any experience or attempt to predict crystal growth under external field (electric, magnetic, etc.)?
No.
Q6. Can you predict the solvent or temperature impact for crystal morphology?
Yes, it is a standard part of the output.
Opportunities and Challenges for Solubility Improvement Approaches of Solid Forms in Formulations
Opportunities and Challenges for Solubility Improvement Approaches of Solid Forms in Formulations
Q1. Are you concerned with stable form switch due to nanocrystallization?
Nanocrystal technique usually requires to use the most stable form in the formulation environment (the most stable polymorph or hydrate) so that we don’t need to concern with the polymorph stability. The nanocrystal stability usually refers to change of the particle sizes, we need to select the appropriate excipient(s) in order to avoid the aggregation of nanoparticles and monitor the aggregations in manufacture process and final drug product.
Q2. If there is no solid form found through polymorph, salt and co-crystal screening, how to isolate the API at large scale?
Usually with extensive study (and knowledge/experience of experts), the API could crystallize in a way by forming crystalline salt or cocrystal. For a compound exists as a liquid form, it will be very challenge in development, e.g. the quality control and chemical stability issues. There are only a few examples that we isolated a drug candidate as a liquid form or a gel-like solid in the early stage, but eventually we found a way to get a solid form by forming a salt, cocrystal, or cyclodextrin inclusion compound. I would recommend to think about the similar strategy for developing amorphous form, it would be help for developing a liquid molecule. Sometimes, we use also a “prodrug” to achieve the crystallization of a drug substance, and it turns to the active molecule once it is in aqueous environment (e.g. WO 2020/038983).
Q3. In relation to the co-amorphous formulation using a small molecule as opposed to a polymer, what steps are necessary to be taken to ensure that the screening is thorough and a more traditional co-crystal will not pop up at a later stage?
First we use software package (Virtual cocrystal screening) developed by CosmoTherm to get suggested top 10-20 co-formers from 300-800 small molecules to ensure that we have the best interacts of the API and the small molecules. After an experimental screening of these top 10-20 conformers like cocrystals, we focus on the amorphous material that not crystallized as API (If it is cocrystal, we evaluate cocrystal as an option). We usually conduct the thermal study first to evaluate the physical stability of co-amorphous form, and then assess the solubility/dissolution assessments. If there is no sign of crystallization of the crystalline API (or cocrystal) during the assessments, we conduct more “crystallization” studies to evaluate the physical stability. Long term stability of co-amorphous form (especially in a high RH environment) is necessary, however, if the co-amorphous form survived in solubility/dissolution assessments and “crystallization” studies in solvent (include water) mediated environments, it usually suggests a low risk of physical stability for storage as a solid form. In case a co-crystal was identified (other than crystalline API), we should assess the solubility improvement of the co-crystal and consider it as an option. Developing a cocrystal is usually preferred compared to developing a co-amorphous form if the solubility improvement can be achieved.
Q4. In the Tricor slide where micronized DS was discussed, can you comment a bit more on the coated-API? How was this achieved?
The formulation of the coated-micronized DS is a commercial product of fenofibrate, and the technology was developed by the company that specialized in this area.
Q5. What would be you recommendation in selecting crystalline micronized API vs amorphous API for fomulation development of amorphous solid dispersion product? and Why?
Crystalline micronized API is preferred since developing a crystalline material is much easier and the particle size reduction (micronization) is a routine process for conventional formulation. In fact, the micronization of API is the first option that we usually consider if the crystalline free form can not achieve the desired exposure.
Q6. What are the techniques to characterize a mestable API polymorph vs stable polymorph? How do you determine which polymorph is stable?
Powder X-ray diffraction is the gold tool to identify various polymorphs. Form the stability, there are three studies that we used to compare the physical stability between polymorphs, 1. Thermal Study, assessment with the Berg’s rule. 2. Slurry study in various solvents (competitive slurry study) to check the form conversion. 3. Solubility measurement. It should be noted that we can only say that we have the most stable form among the identified forms from experimental studies. There is always possibility that we will identify a more stable form with more extensive polymorph/crystal studies. We need to use in silico Crystal Structure Prediction (CSP) to outline the polymorph energy landscape to ensure the most stable is identified. With the predicted crystal structures, we can also understand why a polymorph is more stable than the other at the molecular level to rationalize the polymorph form selection.
Q7. How to decide the particle size of API for solid dosage form development?
If the average of particle sizes of API from crystallization is about 20 micron or above, it will impact the dissolution rate so that impact the performance. If the low drug exposure is caused by the dissolution rate, we usually consider to micronize the API to <10 micron.
Q8. What is the biggest challenge of developing a cocrystal and how to overcome it?
I think the biggest challenge of cocrystal is the identification of a cocrystal with suitable performance. We know co-crystal screening (weak molecular interaction and too many potential co-formers) and manufacture (with narrow crystallization condition to form a cocrystal) is not straightforward. On the other hand, sometimes it is easy to get a co-crystal through the conventional crystallization process, usually in this case the solubility improvement is usually very limited (sometimes decrease the solubility). We often use cocrystal approach to optimize other pharmaceutical properties, e.g. converting to liquid/semi-solid form of a neutral molecule to a solid form, improving solid state property, robust manufacture of the drug substance, improving chemical stability by form a crystalline material from the amorphous form, et.al.
Q9. How to make the decision of developing a metastable form or an enabling formulation of a stable form?
Sometimes the metastable polymorph is stable and won’t easily convert to the more stable during manufacture and storage, we can develop the meta-stable form since these may metastable forms crystallize first and retain the same form in various conditions. There are some examples to a metastable form as the drug substance form, e.g. Avibactam, Ticagrelor. Usually it is not an issue if the conversion of the meta-stable form to the more stable form will not impact the drug exposure (not a poor soluble compound). On the other hand, when we try to use the meta-stable polymorph to achieve the solubility advantage for a poor soluble compound, it is very challenge since the meta-stable form would eventually convert to the more stable and impact the performance. So far it is still very difficult to predict the conversion of the meta-stable form to the more stable form. There are various factors impact the form conversion, and the very small amount (even not detected by conventional methods) of stable polymorph (as seeds) may cause significant issue to physical stability of the drug product. So that the regulatory agency requires extra experiments to prove that the meta-stable drug substance form is physically stable and no performance impact in the final drug product during the whole shelf life period.
Material Properties to Support Modern-Day Tablet Manufacturing
Material Properties to Support Modern-Day Tablet Manufacturing
Q1. What is the drug load limit for Parteck technology?
Assuming this question is addressing the maximum load of our mesoporous silica (Parteck® SLC), in general, the load factor depends on the individual small molecule. Based on our experience of loading more than 30 different APIs onto Parteck® SLC, a drug load of 30% (w/w) is a reasonable target.
Q2. Does your representation minimize the loss of API in the formulation?
Assuming this question addresses the topic of particle optimized mannitol (e.g. Parteck® M 200), in low-dose application case studies – like the 0.4% API formulation case study example shared in the presentation – we have excellent recovery rates in the final tablet formulation and did not experience significant API loss during manufacturing trials.
Q3. What is, according to your experience, the higher drug loading achievable with Parteck SLC?
We have seen examples of up to 50% (w/w); however, on average, a 30% (w/w) drug load is realistic.
Q4. For the amorphous silica carrier, what is the max percentage api loading that can be achieved as a percent of the total tablet weight?
Depending on the desired characteristics of the tablet, we have formulated up to 40% (w/w) of loaded Parteck® SLC into the formulation mixture. In total this amounts to 10-15% (w/w) of API in the final tablet (which is ideal for a 50 – 100 mg API single dose application).
Q5. For the amorphous silica, what does the product stability look like - would expect it to be worse due to higher surface area and lack of crystallinity?
API stability in formulations of mesoporous silica can even improve compared to other technologies. Please be referred to a recent head-to-head comparison on “Opportunities for successful stabilization of peer glass-forming drugs: A stability-based comparison of mesoporous silica versus hot melt extrusion technologies” (https://doi.org/10.3390/pharmaceutics11110577).
Q6. Besides Silica, any other inorganic excipient can be applied? Like TiO2, ZnO, CaCO3?
Due to its multi-compendial specification and existing safety data (e.g. GRAS status) silica is an ideal material for mesoporous carriers. In addition, we have also shown that other inorganic carrier materials can feature similar effects (e.g. magnesium carbonate, Parteck® Mg DC), however, silica shows the best performance.
Q7. My question is for chronic diseases, there is no way we can use solid orals, what are the technologies/challenges to develop solid orals that are implemented against parenteral for such scenario?
There is even precedence that chronic diseases can also be treated by administration of oral solids (e.g. thyroid hormone replacement). In treatment regimens of chronic diseases (but also in treatment of curable diseases), patient adherence and compliance is paramount. We support formulators to develop patient-convenient formulations by providing tailored excipients for modified release kinetics (e.g. ranging from Parteck® ODT for fast orally disintegrating tablets to Parteck® SRP 80 for sustained release tablets).
Q8. Parteck M200 is a SD or granulated or which process is used for its manufacture?
Parteck® M excipients are manufactured to feature an open and filamentary particle structure with high surface areas to best support content uniformity, compressibility and flowability.
Q9. How is SLC compared to other mesoporous silica, e.g., Aeroperl. What are the main differences among the different grades of silica?
In the development of our product various different powder properties (including pore size) were assessed. For silica, there is a trade-off between loading properties and tableting/powder properties. We believe Parteck® SLC represents an optimal balance between pore size and surface area; and particle size. This allows effective loading and excellent tableting properties.
Post Rationalisation of Disorder, Metastable Polymorphs and Extreme Crystallisation Conditions
Post Rationalisation of Disorder, Metastable Polymorphs and Extreme Crystallisation Conditions
Q1. If you calculated free energies for the Form I and Form II ranked CSP structures and if it explains any role of entropy? Also, did you try higher level theory - since DFT-D may not correctly rank order the polymorphs?
The configurational free energy of the thermodynamically stabilising Form I disordered structure is lower than the energy of the two corresponding ordered crystal structures (rank 11 & 45), the energy decrease is dominated by the lower lattice energy of the alternating arrangement, with a small entropic contribution. Whereas, the configurational free energy of the frozen-in Form II disordered structure falls in between the lattice energies of the two corresponding ordered structures (rank 1 & 58, separated by 1.2 kcal/mol). Only DFT-D was used to calculate energies; no other level of theory was applied. Applying the experimental occupancy of 0.47 for Form II conformer 2 (rank 58) and the energy difference between the rank 1 and 58 using the isolated site model, an internal energy of 0.94 kcal/mol and a free energy of 0.53 kcal/mol for the frozen-in disordered Form II structure results. The internal energy difference obtained for Forms II and I is 0.54 kcal/mol, matching well the experimentally measured crystallisation exotherm (by DSC), of 0.33 kcal/mol, allowing for the fact that we completely neglected the contribution of phonon dispersion.
Q2. In your opinion, are high Z' structures (in general) higher in energy compared to lower Z' polymorphs?
The short answer is no. We have seen Z’ = 2 structures through to Z’ = 5 structures having lower lattice energies compared to higher symmetry forms, however, it is system dependent and difficult to generalise.
Q3. How is COSMO-QR calculation performed?
The quantum chemistry package TURBOMOLE was used in the ab initio electronic structure calculations for the conformational analysis of loratadine. The final COSMO level was BP/TZVPD/COSMO-FINE level and the FINE parameterisation in COSMO-RS. COSMOlogic’s COSMOtherm X was used to calculate the temperature dependent Boltzmann distributions and relative energies of the defined conformations to highlight the difference that solvents and solvent mixtures can have on the conformer population.
Q4. How do you predict crystal morphology for a disordered crystal structure?
Exactly the same way for any ordered structure, be it BFDH or an attachment energy model. The point is to have the correct structure represented and not 50% of the structure, as the configuration of the second conformer may influence the surface chemistry... incidentally affecting overall hydrophobicity and mechanical properties.
Q5. Among all disordered structures, how much have lower energy than the ordered structure?
Only one. The thermodynamically disordered Form I structure, which is almost degenerate in terms of energy to the predicted (ordered) rank 1 structure (Form II, conformer 1). However, rank 1 does not crystallise as an ordered structured, instead the frozen-in disordered Form II structure does, represented by an energy in between rank 1 and 58, approximately 0.5 kcal/mol above the thermodynamically disordered Form I structure.