Preparation of Porous cellulose-based ODFs for drug delivery application by thermally induced porogen decomposition in solvent-cast films

Preparation of porous cellulose-based ODFs for drug delivery application by thermally induced porogen decomposition in solvent-cast films Dominik ˇ Sv´ ara a , Petr Jelínek a , Anna Kluk b , Miroslav ˇ So´ oˇ s a,* a Department of Chemical Engineering, University of Chemistry and Technology, Technick´ a 3Prague 6Dejvice, 166 28, Czech Republic b Zentivak.s., U Kabelovny 130Prague 10, 102 00, Czech Republic ABSTRACT Orally dispersible films (ODFs) represent a promising pharmaceutical formulation aimed at enhancing ease of use and, consequently, improving patient compliance with therapy. In this study, we developed a novel method for preparing porous cellulose-based films by modifying the solvent casting technique. Unlike the original method, which is limited to water-insoluble polymers, our approach successfully integrated a porogen insoluble in an organic solvent. Subsequently, this porogen was removed through film heating and thermal decomposition. Several parametric studies were conducted to investigate the impact of various components of the formulation on film properties, including porogen, drug, and disintegrant content, and the mass ratio of the film former to thickener. A notable finding was significant reduction of disintegration time by an increase of the porogen content in the ODF, where the disintegration time decreased from 53.9 ±2.4 s for samples with a porogen mass fraction of 0.27–4.3 ±0.1 s for samples with a porogen mass fraction of 0.72. Furthermore, our results confirm that the presence of the porogen and its subsequent removal did not interfere with active pharmaceutical ingredient (API), which was Tadalafil, nor produce any toxic byproducts, underscoring the viability of prepared material as an alternative to current pharmaceutical forms. 1. Introduction Orally dispersible films (ODFs) are a novel type of pharmaceutical formulation that recently gained significant research interest. This is largely due to their inherent advantages for patients who experience swallowing difficulties, such as small children and elderly individuals, or those with limited access to drinking water, thus significantly improving patient compliance ([1], Bayor et al., 2019). However, despite these benefits, developing ODFs with optimal drug characteristics often presents significant formulation challenges. These challenges include sufficient drug loading while maintaining film integrity, ensuring physical stability (e.g., against moisture or brittleness) during storage and handling, and precisely controlling the dissolution rate of poorly soluble drugs to achieve the desired bioavailability ([2], Kim et al., 2017). There are several approaches to improve the solubility and consequently the bioavailability of prepared materials. This includes the introduction of dissolution enhancers in classical formulations, or improvement of material structure, e.g., by introducing porosity. Although various methods exist for creating porous structures to enhance drug properties, each presents its own set of limitations, particularly when considering pharmaceutical applications and the stringent requirements for patient safety and product quality. One such method is high internal phase emulsion (PolyHIPE) used for the preparation of porous emulsion-templated polymers. This method involves using high internal phase emulsions as templates to prepare porous materials ([3], Krajnc et al., 2011). The main limitation that keeps this method to be useable for pharmaceutical application is related to potential toxic residues that may result from the polymerization process. Another method is foam templating, which works by creating foam from the starting solution. The foam is then solidified into a final porous structure. Solidification can be done by polymerization of continuous phase or removal of solvent and solidification of film former ([4], Andersson et al., 2013, [5], Xu et al., 2016, [6], Lee et al., 2020). The next viable method is electrospinning ([7], Kopˇ rivov´ a et al. 2022, [8], Filipov´ a et al. 2023). It involves the application of a strong electric field between two electrodes. One electrode acts as a collector and collects produced fibers. The other acts as a source of fiber. The mechanism could be described in a simple way by an electric force drawing the starting material and creating nanofibers as the solvent evaporates ([9], Wu et al., 2019). However, there are some limitations to this method. One is a limitation of the starting material, that needs to be overcome by, for example, solvent combination ([10], Nangrejo et al., 2010). It needs to have the right thickness and electrical conductivity to form fibers successfully. Additionally, it is important to note that the electrospinning process is sensitive to the surrounding environment, * Corresponding author. E-mail address: [email protected] (M. ˇ So´ oˇ s). Contents lists available at ScienceDirect Journal of Drug Delivery Science and Technology journal homepage: www.elsevier.com/locate/jddst https://doi.org/10.1016/j.jddst.2025.107668 Received 3 April 2025; Received in revised form 3 October 2025; Accepted 17 October 2025 Journal of Drug Delivery Science and Technology 115 (2026) 107668 Available online 19 October 2025 1773-2247/© 2025 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). where even small changes in humidity can greatly affect productivity and the properties of the resulting material ([11], Van Camp et al., 2009). This is the main disadvantage compared to other methods such as solvent casting. The solvent casting technique is commonly used in the tissue engineering field for the generation of porous materials that then serve as a scaffold ([12] and Mikos 2007). An extension of this method towards formation of porous film is combination of solvent casting technique with particulate leaching. The simplest porogenic agent used in the literature is sodium chloride. Due to its cost and easy control over crystal size that can be used to modulate porosity of final material ([13], Pavasant et al., 2014, [14], Bertacchini et al., 2019). In this case, porogen and film forming material is mixed together. The polymer is dissolved in an organic solvent, while the salt remains undissolved, thus creating a suspension of sodium chloride particles. After homogenization, the mixture is poured into the mold, and the organic solvent is let to evaporate. After that, the material is immersed in a solvent, which dissolves the porogen and leaves the material, forming a porous structure [15]. Modification of this process can also use two polymers with different solubility based on the pH of the aqueous solution. In this case, the one dissolves faster. The remaining solvent should be removed from the material either by heating the material or by freezing. As can be seen, there are two main limitations of this process. The first is the need to remove the solvent after the washing process. This can be unsuitable for certain applications, e.g., drug delivery, where water with an adjusted pH can leave residues of chemicals used to adjust the acidity of the solvent for selective porogen removal. The second limitation is the loss of API from the structure. This happens by dissolution of API that is in contact with washing solution. This limitation is the most important one. Taking into account these challenges and the specific requirements for pharmaceutical formulations, such as avoiding the use of excessive surfactants, minimizing potential toxic residues from polymerization, and eliminating the need for solvent-based porogen removal, a more suitable approach is presented in this work. This study focuses on a novel method where a solid porogen is removed by simply using heat, specifically thermally induced porogen decomposition. This technique offers significant advantages over traditional solvent casting and other methods by completely omitting the need for a liquid leaching step. This eliminates concerns about residual water or pH-adjusting chemicals that could affect the film’s stability, or lead to unwanted impurities in the final product. Furthermore, unlike methods that rely on complex polymerization or electrospinning, this approach simplifies the manufacturing process and reduces the potential for environmental sensitivity during production. The porogen could be chosen from organic or inorganic chemicals. However, the choice must satisfy two parameters. One is that the porogen or its degradation products cannot be toxic. The second is the low temperature of decomposition, which needs to be lower than the decomposition temperature of other film components. When these two requirements are taken into account, most of the organic components are discarded from selection because of the chance of formation of toxic byproducts and higher decomposition temperatures. On the contrary, inorganic compounds such as sodium hydrogen carbonate and ammonium carbonate are used in food preparation as a leavening agent. They are also used in the pharmaceutical industry. Therefore, their possible residues in products are safe to ingest and have no negative longand short-term effects on the human body. Because they are simple in structure, their decomposition products are only gases that would not stay in a prepared material. When comparing the two mentioned chemicals, ammonium carbonate is the preferred choice, since it has lower decomposition temperature (~36 ◦C) than sodium hydrogen carbonate (~80 ◦C). Therefore, this means that lower temperatures and shorter times are needed and there is less potential to create toxic by-products from other components of the formulation, mainly from incorporated API. 2. Materials and methods 2.1. Chemicals For our experiments, we used Hypromellose 2910 (also referred to as HPMC 2910 or by the commercial name Tylopur 645, kindly provided by SE Tylose (Germany)) as a film former because of its biocompatibility, chemical stability, and ability to form a clear and strong film. It was chosen over lower grades to ensure uniformity and processability, which could not be met by lower grades. Ammonium carbonate as a porogen and ethanol as a solvent were purchased from Lach-ner (Czech Republic). Before use, porogen particle size was reduced with the Retsch MM 400 vibration mill (Germany) using 20 stainless steel balls with 3 mm diameter per milling vessel. Grinding vessel had a nominal volume of 50 ml with material volume to be 20 ml. The reduced particles were then separated according to their size to be below 71 μ m. This specific size was chosen as the smallest practically achievable fraction, as attempts to obtain smaller particles with sufficient yields were not successful with the available equipment and separation methods. For separation AS 200 by Retsch (Germany) was used. Tadalafil (model API) was kindly provided by Zentiva k.s. Sodium dodecyl sulfate (SDS) as a surfactant was purchased from Sigma-Aldrich. PEG 600 was chosen as the commonly used plasticizer to modulate the mechanical properties of the films after casting, crospovidone (Kollidon CL) as the disintegrant to modulate the disintegration time, and PVP k90 as the thickening agent to prevent sedimentation of porogen particles by slightly increasing the viscosity of the solution were kindly provided by BASF (Germany). This grade of PVP was selected because of its high molecular weight, enabling significant enhancement of viscosity at low concentrations, thereby minimizing the required amount in the formulation. The buffer solution used for the experiments with a set pH of 6.8 was prepared with KH 2 PO 4 and NaOH. Both components were dissolved in 1000 ml of deionized water. The weight of components was 0.9 g of NaOH and 6.81 g of KH 2 PO 4 . All components were thoroughly mixed and dissolved in water. The final pH value was validated by the SevenCompact Duo pH meter with the InLab Expert Pro-ISM, both by Mettler-Toledo (Switzerland). 2.2. Porous material preparation All samples were prepared as follows. First, API, film former, plasticizer, disintegrant, surfactant, and thickening agent were weighed into the mixing vessel. Then ethanol was added, and the resulting suspension was stirred with a magnetic stirrer at a rate of 300 RPM for 15 h. The porogen was then added and the suspension was allowed to homogenize for another 2 h at a rate of 300 RPM. Consequently, the suspension was cast into prepared molds. The volume of suspension was 3.3 ml per 1 prepared strip. The cast suspension was then dried for 12 h to remove the organic solvent. The prepared samples were then placed in an oven at 60 ◦C for 24 h to remove the remaining organic solvent and porogen from the prepared samples. The molds were made of polytetrafluoroethylene (PTFE). This material was chosen for its chemical resistance and ease of removal of samples from molds after drying. The inner dimension of the molds was 25x32 ×5 mm. They were manufactured with a CNC machine 4MILL300 by P2J Technology (Czech Republic). The film preparation steps are shown schematically in Fig. 1. 2.3. Composition studies We prepared several sample groups to examine the effect of each component. The first study was to find a suitable amount of porogen that should be used to prepare the porous structure of solvent-casted films. The total amount of porogen was gradually increased while maintaining the amount of other components. This means that after the porogen is completely removed, the composition is the same for all samples. This was done to find an optimal amount of porogen and explore the effects of D. ˇ Sv´ ara et al. Journal of Drug Delivery Science and Technology 115 (2026) 107668 2 porogen on film properties. For exact compositions, refer to Table 1 (for composition before and after porogen removal). The role of porogen in this study was to create a porous structure and to find an optimal amount of porogen to use in preparation of the study material. Table 1 also contains mass corresponding to one strip. We prepared 20 strips from each sample from this study. The volume of pure ethanol used to prepare one strip was equal to 3.3 ml. In the next study, we investigated the impact of the API. There, we gradually increased the amount of API while maintaining initial amounts of other components. This was done to isolate the effect of API and investigate how starting composition will affect the properties of the prepared films. Please see Table 2 (for composition before and after porogen removal) for the exact composition of the samples. The mass fraction of porogen in these samples was fixed and corresponds 0.72. The role of the porogen in this study was to generate a porous structure similar to that found to be optimal in the porogen study. Please note that other components mass fraction is changing in the sample due to fixing the initial amount of other components excluding API, which was increased. Table 2 also contains mass corresponding to one strip. We prepared 20 strips from each sample from this study. The volume of pure ethanol used to prepare one strip was equal to 3.3 ml. Another study examined the effect of the disintegrant. Similar to the API study we started with a composition that corresponds to the optimal sample in the porogen study, and we gradually increased the amount of disintegrant, while the amount of other components was fixed to isolate effect of disintegrant. The amount of porogen in these samples was fixed and corresponds to the last sample in the porogen study (that is, sample 6 with a porogen mass fraction of 0.72). Please see Table 3 (before and after porogen removal) for exact compositions. This study was carried out to assess properties in the changes of the samples and to examine if the disintegration of the samples can be further improved by other components, in addition to the porogen. The role of the porogen in this study was to generate a porous structure similar to that found to be optimal in the porogen study. Table 3 also contains mass corresponding to one strip. We prepared 20 strips from each sample from this study. The volume of pure ethanol used to prepare one strip was equal to 3.3 ml. The last study was dedicated to investigating the impact of the thickener/film former ratio while the amount of all other components remained the same. For exact composition, see Table 4 (before and after porogen removal). This was done in order to examine the effect of plasticizer on mechanical stability and potential interchangeability of these two components in the prepared films. The porogen content was the same in all samples, equal to the amount in the last sample of porogen study (that is, sample 6 with a porogen mass fraction of 0.72). The role of the porogen in this study was to generate a porous structure similar to that was found to be optimal in the porogen study. Table 4 also contains mass corresponding to one strip. We prepared 20 strips from each sample from this study. The volume used to prepare one strip was equal to 3.3 ml of pure ethanol. 3. Analytic techniques 3.1. Scanning electron microscopy The prepared samples were analyzed by scanning electron microscope (SEM). Two small pieces were cut from each strip to investigate the structure of both parts of each sample. Consequently, all samples were coated with Emitech K550X sputter coater (Quorum Technologies, UK). The coating current was 2 mA, and the coating time was 3 min. SEM images were obtained by Vega3 from Tescan (Czech Republic). Images were processed prior to analysis. The image processing was performed using ImageJ as follows. First, the images were transformed into an 8-bit grayscale image. The images were divided into segments with uniform illumination levels. Even lighting is essential for accurate segmentation of pores and porosity quantification. This is because unFig. 1. Illustration of porous films preparation process. Table 1 Composition of samples in porogen study before and after porogen removal. Porogen Tadalafil SDS PEG 600 Kollidon CL PVP k90 Tylopur 645 (NH 4 ) 2 CO 3 Mass corresponding to one strip [mg] b (API) [−] a (Surfactant) [−] a (Plasticizer) [−] a (Disintegrant) [−] a (Thickener) [−] a (Film Former) [−] a (Porogen) [−] a 1 0.036/ 0.050 0.036/0.050 0.036/0.050 0.036/0.050 0.145/0.200 0.434/0.600 0.276/- 159/116 2 0.028/ 0.050 0.028/0.050 0.030/0.050 0.028/0.050 0.115/0.200 0.342/0.600 0.429/- 203/116 3 0.024/ 0.050 0.024/0.050 0.024/0.050 0.023/0.050 0.094/0.200 0.282/0.600 0.529/- 246/116 4 0.020/ 0.050 0.020/0.050 0.020/0.050 0.020/0.050 0.081/0.200 0.241/0.600 0.598/- 289/116 5 0.018/ 0.050 0.017/0.050 0.018/0.050 0.017/0.050 0.070/0.200 0.209/0.600 0.651/- 333/116 6 0.014/ 0.050 0.014/0.050 0.014/0.050 0.014/0.050 0.055/0.200 0.165/0.600 0.724/- 419/116 a Before porogen removal/after porogen removal. b Corresponds to an average mass of one strip calculate as a mean of 20 strips (including porogen/excluding porogen). D. ˇ Sv´ ara et al. Journal of Drug Delivery Science and Technology 115 (2026) 107668 3 even lighting can mask the contrast between pores and solid material, making it difficult to accurately segment pores. The auto-testing algorithm was applied to distinguish between pores and solid regions. The chosen algorithm was Otsu because it provided the best results of all standard tested algorithms. The regions with black color were interpreted as void regions, and regions with white color were interpreted as bulk material. The surface void ratio (VR) was then calculated as percentage from all pixels in an image according to the following formula. VR =NB NB+NW *100 (1) Where VR is surface void ratio [%], N B are counts of black pixels and N w are counts of white pixels. 3.2. FTIR spectroscopy FTIR (Fourier transform infrared) spectroscopy was used to measure Table 2 Composition of samples in API study before and after porogen removal. API Tadalafil SDS PEG 600 Kollidon CL PVP k90 Tylopur 645 (NH 4 ) 2 CO 3 Mass corresponding to one strip [mg] b (API) [−] a (Surfactant) [−] a (Plasticizer) [−] a (Disintegrant) [−] a (Thickener) [−] a (Film Former) [−] a (Porogen) [−] a 1 0.014/ 0.050 0.014/0.050 0.014/0.050 0.014/0.051 0.055/0.202 0.164/0.597 0.725/- 419/116 2 0.017/ 0.062 0.014/0.050 0.014/0.049 0.014/0.050 0.055/0.197 0.164/0.592 0.722/- 421/117 3 0.021/ 0.073 0.014/0.049 0.014/0.049 0.014/0.049 0.055/0.195 0.164/0.585 0.718/- 422/119 4 0.027/ 0.096 0.014/0.047 0.014/0.048 0.014/0.047 0.054/0.191 0.163/0.571 0.714/- 425/122 5 0.040/ 0.136 0.013/0.046 0.013/0.045 0.013/0.046 0.054/0.182 0.161/0.545 0.706/- 432/127 6 0.053/ 0.174 0.013/0.044 0.014/0.044 0.013/0.043 0.053/0.174 0.159/0.521 0.695/- 437/133 a Before porogen removal/after porogen removal. b Corresponds to an average mass of one strip calculate as a mean of 20 strips (including porogen/excluding porogen). Table 3 Composition of samples in disintegrant study before and after porogen removal. Disintegrant Tadalafil SDS PEG 600 Kollidon CL PVP k90 Tylopur 645 (NH 4 ) 2 CO 3 Mass corresponding to one strip [mg] b (API) [−] a (Surfactant) [−] a (Plasticizer) [−] a (Disintegrant) [−] a (Thickener) [−] a (Film Former) [−] a (Porogen) [−] a 1 0.014/ 0.051 0.014/0.050 0.014/0.052 0.014/0.051 0.055/0.200 0.164/0.596 0.725/- 419/116 2 0.014/ 0.050 0.014/0.049 0.014/0.050 0.017/0.062 0.055/0.198 0.164/0.591 0.722/- 421/117 3 0.014/ 0.049 0.014/0.049 0.014/0.049 0.021/0.073 0.055/0.195 0.164/0.585 0.718/- 422/119 4 0.014/ 0.047 0.014/0.048 0.014/0.048 0.027/0.095 0.054/0.190 0.163/0.572 0.714/- 424/120 5 0.013/ 0.045 0.013/0.046 0.014/0.046 0.040/0.136 0.054/0.182 0.161/0.545 0.705/- 425/122 6 0.013/ 0.043 0.013/0.044 0.013/0.043 0.053/0.173 0.053/0.174 0.159/0.523 0.696/- 437/133 a Before porogen removal/after porogen removal. b Corresponds to an average mass of one strip calculate as a mean of 20 strips (including porogen/excluding porogen). Table 4 Composition of samples used in film former thickener study before and after porogen removal. Sample Tadalafil SDS PEG 600 Kollidon CL PVP k90 Tylopur 645 (NH 4 ) 2 CO 3 Mass corresponding to one strip [mg] b (API) [−] a (Surfactant) [−] a (Plasticizer) [−] a (Disintegrant) [−] a (Thickener) [−] a (Film Former) [−] a (Porogen) [−] a 1 0.014/ 0.051 0.014/0.049 0.015/0.053 0.014/0.052 0.022/0.079 0.198/0.716 0.723/- 419/116 2 0.014/ 0.050 0.014/0.050 0.014/0.049 0.014/0.050 0.055/0.200 0.165/0.601 0.724/- 419/116 3 0.014/ 0.050 0.014/0.051 0.014/0.050 0.014/0.051 0.110/0.400 0.110/0.399 0.724/- 419/116 4 0.014/ 0.050 0.014/0.050 0.014/0.051 0.014/0.050 0.166/0.601 0.055/0.198 0.723/- 419/116 5 0.014/ 0.050 0.014/0.050 0.014/0.050 0.014/0.050 0.198/0.719 0.022/0.081 0.724/- 419/116 a Before porogen removal/after porogen removal. b Corresponds to average mass of one strip calculate as a mean of 20 strips (including porogen/excluding porogen). D. ˇ Sv´ ara et al. Journal of Drug Delivery Science and Technology 115 (2026) 107668 4 the final composition of the prepared samples. The Nicolet iS10 FTIR instrument from Thermo Fisher (USA) was used for measurements. The instrument utilized mid-infrared Ever-Glo ceramic light source and a HeNe laser as internal reference for interferometer setting. The resolution was set to 0.4747 cm −1 , and the scanning range was set from 400 to 4000 cm −1 . The sampling interval was set to 0.482 cm −1 . A total of 32 scans were acquired per spectrum. To ensure homogeneous distribution of material, the samples were ground before analysis. Before measurement, all samples were stored under laboratory conditions (relative humidity 50 % at 25 ◦C). 3.3. X-ray diffraction (XRD) XRD measurements were performed in order to confirm the removal of porogen from samples, since the porogen is an inorganic crystal salt. All measurements were made with a PANalytical X’Pert3 Powder diffractometer, Cu K α radiation, λ ≈1.5406 Å from Malvern (UK). The measured results from the prepared samples were compared with those of the pure components. 3.4. Mechanical properties All mechanical properties were performed using a CT3 texture analyzer CT3 from Brookfield Amatek (USA). This instrument is designed to measure force, and its maximum force capacity is 10 N. To accommodate this limitation, the samples were cut into pieces approximately 4 ×30 mm in size. The dimensions of each sample were measured and recorded for later normalization of the data. The texture analyzer was set to a test speed of 1 mm/s and a target distance of 10 mm. The measured force values were recalculated to obtain stress values in the sample. These stress values were then plotted against the corresponding strain values, which were calculated based on the set deformation of the sample and its original length. The slope of the linear portion of the stress-strain curve was determined, and this value corresponds to the Young’s modulus of the prepared material. 3.5. Disintegration The novel nature of the prepared material required a different approach to obtain disintegration time. Since the repeatability of disintegration time measurement was required, the holder for samples was constructed. The holder comprises three parts. Bottom part, where samples are placed, spacer which serves to hold the sample in defined position and top cover (see Fig. SI 1 in the supplementary material for a detailed view of the assembled holder). All parts were printed with Prusa i3 3D printer (Prusa Research, Czech Republic). The material used for printing was polylactic acid (PLA) ±added color. All models are included as “.stl” files in supplementary material. The device used to record disintegration was a Samsung Galaxy Note 10 by Samsung (South Korea). The phone was set to capture disintegration with its camera set to high-speed frame rate. It was set to 240 FPS. The video was then processed by the MATLAB script and separated into individual images. The images were then mass-processed by cropping to contain only areas unobscured by parts of holder. It also helped decrease the resolution of images and slightly reduce the computational requirements. Finally, all images were analyzed by Python script. This script obtained grayscale histograms for all images. The average value of each histogram was then calculated. The change in the average grayscale value was then used to determine the disintegration time. A similar analysis method was used in our previous work (ˇ Sv´ ara, Kopˇ rivov´ a et al., 2022, [8], Filipov´ a et al., 2023). All scripts are included as additional supplementary material. 3.6. Dissolution kinetics Before experiments, the calibration curve was obtained. This was done to obtain the dissolved concentration of tadalafil during dissolution. It was obtained as follows. Tadalafil was first dissolved in pure ethanol to obtain a set concentration of 15 mg/ml. This solution was then added to a 500 ml aqueous buffer solution (pH =6.8) with a step of 0.1 ml, and the absorbance was measured from 500 ml to 501 ml. Absorbance was measured for each point. The calibration curve was used in the following measurements to obtain the dissolved ratio of declared dose. All experiments were carried out as follows. The samples were placed and enclosed in a small basket to prevent them from floating on the surface of the liquid. After the start of the experiment, approximately 3 ml of the solution was withdrawn from the dissolution vessel with syringe. The syringe content was than filtered through PTFE filter with a porosity of 0.45 μ m. This was done to remove solid particles of undissolved material because they would interfere with the measurement. Change in volume was accounted in final calculations of concentration. The sampling was done at time intervals of 1 min, 2 min, 5 min, 10 min, 15 min, 20 min, 30 min, 40 min, 50 min, and 60 min from the start of the experiment. All measurements were done with a UV/Vis spectrophotometer Cary 60 by Agilent (USA). The scanning speed was set to 600 nm/min. The measured interval ranged from 500 to 200 nm. The peak used to determine concentration of dissolved tadalafil was at the position of 291 nm. This peak was chosen because of the highest absorbance of tadalafil at this position. All experiments were carried out in buffer solution prepared with pH of 6.8. The volume of solvent was the same in all experiments, and it was 800 ml. The stirring was performed with a magnetic stirrer with stirring speed of 150 RPM. The mass of samples was established to contain approximately 2.97 mg of tadalafil. The final mass of the sample was calculated on their prepared composition to contain a set amount of tadalafil. The value was determined to correspond to 20 % of the solubility of tadalafil in the volume used. 4. Results and discussions 4.1. Scanning electron microscopy Images from scanning electron microscopy for all investigated effects are presented in Fig. 2. As shown in Fig. 2-A, in the case of an increase in the porogen content in the starting mixture, we observed gradual changes in the porous structure. This is attributed to the overall larger content of porogen in the starting mixture, which leads to greater porosity after removal. The specific porogen particle size below 71 μ m was selected as the smallest practically achievable fraction. Although finer particles were initially desired to potentially enhance pore characteristics, obtaining significant yield of porogen particles smaller than 71 μ m was with our ball vibration mill and subsequent separation methods prohibited. Furthermore, the thermal sensitivity (decomposition at ~36 ◦C) also played a role in limiting the intensity and duration, preventing the generation of much finer particles without degradation of the material. Therefore, the fraction of porogen particles smaller than 71 μ m represented the most practical and stable compromise for consistent film production. As shown in Fig. 2-B, an increase in the amount of API in the mixture did not significantly alter the structure of the prepared samples. It also supports the assertion that the API is incorporated into the material structure rather than being present in particle form. For a detailed image, see Fig. SI 2 in the supplementary material where the detailed structure of the film structures from the API study is presented. Similarly, in the case of the disintegrant (see Fig. 2-C), we did not observe any significant changes in the structure of the prepared materials. This was expected because the amount of the disintegrant was significantly lower than that of the other components and, therefore, it would have D. ˇ Sv´ ara et al. Journal of Drug Delivery Science and Technology 115 (2026) 107668 5 minimal effect on the porous structure. This outcome is interpreted positively as it suggests the potential for application in larger amounts without affecting the prepared structure. A similar lack of change was observed in the study of the film former/thickener ratio (see Fig. 2-D). The porosity of prepared samples was further characterized by examining the surface void ratio, and the results are depicted in Fig. 3. The porogen study (see Fig. 3-A) demonstrates a significant increase in the surface void ratio with the increased amount of porogen. Upon the assumption that there is a positive correlation between the surface void ratio and film porosity, the observed trend confirms an increase in film porosity. To be able to compare the impact of other components, the amount of porogen in all other samples was kept the same and equal to 0.72 mass fraction in the starting material, corresponding to the highest achieved value in this study. As can be seen in Fig. 3-B-D within the confidence interval the surface void ratio of all other tests was similar, supporting the dominant role of the porogen on the porosity of the prepared samples. The small differences are due to imprecision of the measurement method. On the remark of precise porosity, we tried to obtain direct measurement of porosity by mercury porosimeter, which was the suitable method for our sample. However, the sample was too brittle to withstand the pressure that is applied to force mercury into pores of the material. Therefore, we decided to use surface void ratio obtained from 2D surface images as approximation of porosity of prepared materials. It must be noted that this parameter is limited since it does not provide direct information about bulk porosity. It only provides information about surface porosity. However, we were not able to obtain bulk porosity so we used this one to compare prepared samples. Based on these findings, we can conclude that the critical parameter for the film porosity is the amount of porogen while other parameters have negligible effect and thus could be used to fine-tune the other film properties without altering the porous structure. Fig. 2. SEM image of all prepared samples. Examples of structures observed in studies of API effect, porogen effect, disintegrant effect and film former/thickener effect. D. ˇ Sv´ ara et al. Journal of Drug Delivery Science and Technology 115 (2026) 107668 6 4.2. FTIR spectroscopy FTIR spectroscopic measurements yielded several results. One of the positive outcomes was observed in the study of the porogen mass fraction. The measured data are presented in Fig. 4. The spectrum did not indicate any changes in the location of the most intensive peaks of the porogen. The strongest peaks at 702 and 835 cm −1 are characteristic for ammonium subunits wagging and rocking vibration are not present in any of our prepared samples. This strongly indicates the absence of ammonium carbonate in the prepared material, since this group is not present in any other component of prepared material. Furthermore, absence of characteristic peaks at 1588 cm −1 (C=O asymmetric stretch) also support this claim. This is a positive finding, suggesting that all porogen was removed from the prepared samples, minimizing the potential for unpleasant taste. It also shows no formation of toxic residues in the prepared samples as a result of the interaction with API. Other compositions studies also revealed no significant changes in the API, confirming the absence of toxic residues from API decomposition. We can see that by absence of changes at 1340 cm −1 (O-H group bending) also at peaks at 1590 cm −1 and 1560 cm −1 (aromatic C=C stretches) which are the only ones that do not overlap with other components. All remaining data are presented in the supplementary material; please refer to Fig. SI 4 and Fig. SI 5 for exact FTIR spectra. Data from the film former/thickener study shows gradual changes in two characteristic peaks for the components studied. The peaks of interest are at 1657 cm −1 (PVP 90F) and 1066 cm −1 (Tylopur 645). The peak at 1657 cm −1 shows a gradual increase in the prepared samples, corresponding to the increase in PVP 90F in the samples, while the peak at 1066 cm −1 shows a gradual decrease, corresponding to the decrease of Tylopur 645 in the sample. This indicates changes in the ratio between these two samples. However, no other changes were observed in this study (see Fig. SI 6 in the supplementary material). For comparison of all components in all samples, the FTIR spectra of pure compounds are also shown in Fig. SI 3 in the supplementary material. 4.3. X-ray diffraction XRD measurements were performed to confirm the complete removal of the porogen through heat treatment. Since the porogen ((NH 4 ) 2 CO 3 ) possessed a crystalline structure and its peaks are prominently visible on the obtained diffractogram (see Fig. 5A), their detection would indicate presence of porogen in the ODF samples prepared in this work. However, as can be seen in (see Fig. 5B–D), independent to the amount of used porogen there is absence of major diffraction peaks at 24◦and 30◦together with some other minor peaks, supporting the absence of porogen residues in the samples. This confirm that the selected heating duration effectively eliminates the porogen from the samples. While the porogen itself is not toxic, its complete removal also eliminates the potential for unpleasant taste. 4.4. Mechanical properties Measurement of Young’s modulus reveals several factors regarding the mechanical properties of the samples. In the case of the porogen study (see Fig. 6-A), the maximum value of Young’s modulus was observed in the sample with a mass fraction of 0.43 compared to the other components, followed by a steep decrease. In this case, the changes in Young’s modulus correspond to changes in the inner structure of the material, where more porous materials exhibit lower values of Young’s modulus. Thus, the results also confirm the increase in Fig. 3. Surface void ratio of prepared samples. Part A contains data for porogen study. Part B contains data for API study. Part C contains data for disintegrant study. Part D contains data for Thickener/Film former study. D. ˇ Sv´ ara et al. Journal of Drug Delivery Science and Technology 115 (2026) 107668 7 porosity observed in the analysis of SEM images (see Fig. 3-A in Section 3.1. Scanning electron microscopy). However, the sample with a porogen mass fraction of 0.28 was identified to have the lowest Young’s modulus value within the studied range. This unexpected behavior, particularly considering its relatively low porosity compared to other samples, deviates from the general trend where lower porosity typically correlates with higher stiffness. This anomaly could be attributed to several microstructural factors beyond simple volumetric porosity. For instance, in very dense or minimally porous structures, specific processing conditions are known to sometimes lead to localized stresses or an unfavorable packing arrangement of the polymer chains [16]. Additionally, it is plausible that this specific porogen concentration resulted in non-uniform pore distribution with localized weaknesses, or perhaps even induced subtle microcracking or interfered with the optimal polymer chain organization within the matrix. These phenomena, which may not present visible macroscopic defects in standard imaging, can collectively contribute to a lower-than-expected Young’s modulus. While a direct comparison to a film prepared without porogen was not available for this study (due to long disintegration time – see text below), these findings highlight the complex interplay of processing, microstructure, and mechanical properties. Overall, the rest of the results show a decrease in Young’s modulus that is important for better handling of the prepared material, which is crucial for the possible large-scale preparation of examined films. The API study showed a very weak maximum in mechanical properties for API mass fraction of about 0.12 (see Fig. 6-B). Tadalafil (M w ~389.4 g/mol) is a relatively small organic molecule with various functional groups (e.g., nitrogen atoms, oxygen atoms within cyclic structures) that could potentially engage in intermolecular interactions with the polymer matrix (e.g., such as hydrogen bonding or dipoledipole interactions). Its molecular size and structure are within the range observed for many effective pharmaceutical plasticizers, which are typically low molecular weight compounds, often with ester or ether linkages. Given these characteristics, it is plausible that tadalafil molecules could intercalate within the polymer chains, disrupt existing polymer-polymer interactions, and thereby increase chain mobility, leading to the observed peak in Young’s modulus (indicating initial toughening or optimized flexibility), followed by a decrease at higher concentrations due to excessive plasticization or phase separation. Although direct evidence of glass transition temperature (T g ) reduction or specific molecular interactions is beyond the scope of current data, this observed trend in mechanical properties is consistent with a plasticizing role of API present in the literature. For example ([17], Le Brun et al., 2006), reported Metoprolol tartrate, chlorpheniramine maleate and ibuprofen as an effective plasticizers for Eudragit RS. Another Fig. 4. FTIR spectra of samples after porogen removal. Presented data shows spectra of samples corresponding to various starting porogen mass fraction, from 0.28 to 0.78. In addition, it also shows spectra of pure porogen (ammonium carbonate). Fig. 5. X-ray diffractograms of porogen samples. Diffractograms show selection of samples from porogen study. Part A shows diffractogram of pure porogen (ammonium carbonate). Part B shows sample with mass fraction of 0.28 of porogen at the beginning. Part C shows sample with mass fraction of 0.62 of porogen at the beginning. Part D shows sample with mass fraction of 0.72 of porogen at beginning. D. ˇ Sv´ ara et al. Journal of Drug Delivery Science and Technology 115 (2026) 107668 8 example is ([18], Viau et al., 2020) who reported ionic liquid of lidocainum ibuprofenate to act as plasticizer for zein based material. This finding is particularly interesting as the impact of API amount variation did not show a significant effect on the surface void ratio or on the SEM analysis, suggesting that mechanical changes are driven by molecular-level interactions rather than macroscopic structural changes. A similar situation was also observed for the disintegrant study (see Fig. 6-C), where the increase of Young’s modulus was observed up to a sample with a mass fraction of 0.14, again followed by a decrease. Similarly, as before, it points to an increase in the brittleness of samples up to a certain point, followed by the sample becoming more plastic. This could be explained by the insolubility of crospovidone in an organic solvent used in the preparation of films. The thickener/film former ratio study was excluded due to samples becoming too brittle after a certain point. 4.5. Disintegration All the obtained values of disintegration time are presented in Fig. 7. The porogen study shows a decrease in disintegration time with an increase of the porogen mass fraction (as depicted in Fig. 7-A). More precisely, two samples with porogen mass fraction of 0.65 (17.59 ± 0.62 s) and 0.72 (4.29 ±0.03 s) met the required criteria of 30 s specified by the pharmacopoeia for orally disintegrating formulations. All other samples disintegrate after a minute, which is larger than the pharmacopoeia limit. Please note, that preliminary results for films without porogen indicate even longer disintegration time than a minute (data not shown). This finding correlates with the porosity of the material obtained from the image analysis of the SEM pictures (see Fig. 2Fig. 6. Young’s modulus of prepared samples. Part A shows results for samples from the porogen study. Part B shows results for samples from API study. Part C shows results for samples from the disintegrant study. Fig. 7. Measured disintegration time of the samples. Part A shows results for the porogen study. Part B shows results for API study. Part C shows the results for the disintegrant study. Part D shows results for the film former and thickener study. D. ˇ Sv´ ara et al. Journal of Drug Delivery Science and Technology 115 (2026) 107668 9