(686a) Molecular Mobility As a Tool for Understanding the Impact of Polyvinylpyrrolidone (polymer) and Tpgs (surfactant) in Crystallization Kinetics of Amorphous Celecoxib

Taylor, L., Purdue University

Formulating drugs (active pharmaceutical ingredients) as amorphous solid dispersions is an alternative for improving the solubility of poorly water soluble drugs. An amorphous solid dispersion ideally consists of the amorphous drug homogeneously distributed and thermodynamically stable within a polymer matrix. In addition to the polymer, the formulation may contain other materials such as surfactants to improve processability, dissolution and flow behavior, among others. Understanding the crystallization tendency of such compounds is crucial when assessing it’s feasibility for a solid dispersion formulation, as re-crystallization of the drug is an undesired effect and may lead to potential market withdrawal of a commercial drug product. Celecoxib is a poorly water soluble (BCS class II) non-steroidal anti-inflammatory drug. When compared to other BCS class II compounds, it has relatively low molecular mobility (high viscosity) and crystal growth rate upon cooling from the undercooled melt state [1,2]. The presence of surfactants or polymers has demonstrated to significantly increase or inhibit the radial crystal growth rate of Celecoxib, respectively. Interestingly but not necessarily advantageous, the combination of both, results in a complex behavior where the polymer is able to reduce the growth rate acceleration caused by the surfactant resulting in an increase or decrease of the overall crystal growth rate depending on the miscibility of Celecoxib and the surfactant [3]. Even though, there is a large number of surfactants and polymers available for amorphous solid dispersions, few studies elucidate the impact of all of these in the crystallization kinetics of drugs. While the impact of surfactants and polymers in the crystallization growth rates of pure amorphous Celecoxib has been studied and reported, this impact analyzed in terms of the molecular diffusivity is a knowledge still not available. Crystallization of a drug is a diffusive process where the nucleation rate is dictated by the fluidity of the liquid (viscosity). Fundamental knowledge may be obtained when the crystallization tendency (or lack thereof) is analyzed in terms of the “glass forming ability” of the liquid, i.e.kinetic fragility. Here, the liquid’s viscosity with temperature may follow Arrhenius or Vogel-Tamman-Fulcher (VTF) dependence, depending on the relaxation of the liquid molecules. It is commonly known that glassy liquids at high temperature follow Arrhenius dependence (linear), however upon lowering the temperature deviates from Arrhenius and follows VTF law (non-linear). Interestingly, some speculate that the VTF behavior transitions back to an Arrhenius dependence at low temperatures [4]. The viscosity is basically the resistance of a fluid to flow under an external applied stress, it is then related to the molecular relaxation and hence mobility of a material. We have characterized the rheological properties of Celecoxib with small amounts of PVP (polymer) and/or TPGS  (surfactant) with the purpose of understanding the impact on the molecular diffusivity of the Celecoxib during crystallization from the undercooled melt state. We expect these results to complement the knowledge available of molecular mobility of Celecoxib with regards to thermodynamic quantities for solubility prediction [5]. Additionally, in general for when assessing a drug as a candidate for a solid dispersion formulation during developmental stages of a drug product. Amorphous samples where prepared using a cryogenic mill. Binary and ternary mixtures consisted of 10 wt.% TPGS or 10 wt.% PVP dispersed in a 90wt.% Celecoxib, and ternary mixtures of 10wt.% PVP, 10 wt.% TPGS and 80 wt.% Celecoxib. Samples were characterized in an Anton Paar MCR 301 series rheometer, using 20mm diameter parallel Peltier Plates. Temperature dependent Dynamic Viscosity was measured while cooling the sample from melting point at a cooling rate of 10oC/min. We measured the dynamic rheological properties of the Celecoxib samples. The dynamic viscosity data was analyzed in terms of molecular mobility, Activation Energy (Ea), and fragility (Strength Parameter) (D) of the drug near the glass transition (Tg) and melt temperature (Tm), this was possible by analyzing the viscosity data with Arrhenius and Vogel-Tamman-Fulcher (VTF) equations. The temperature dependent dynamic viscosity curves resulted in overall viscosity increase upon cooling from the melting temperature, which was expected due to the nature of the measurement. The rate at which the viscosity increased with temperature varied with sample. Generally speaking, pure Celecoxib and Celecoxib/TPGS resulted in a small degree of undercooling compared to Celecoxib/PVP and Celecoxib/PVP/TPGS, this due to sample solidification during measurement. Pure Celecoxib resulted in an increase in viscosity of approximately one decade at roughly 10oC degrees of undercooling; at this point the viscosity suddenly increased dramatically almost three orders in magnitude. This sample could not be measured more than 30oC degrees of undercooling. Celecoxib/TPGS resulted in a similar behavior as pure Celecoxib, with a reduced degree of undercooling, as the sample showed a marked increase in viscosity near the melt temperature. Celecoxib/PVP resulted in a significant increase in the degree of undercooling (~95oC) as the dynamic viscosity could be measured up to the glass transition temperature of the sample with an increase of roughly one decade of viscosity every 10oC in temperature. Interestingly, almost 10oC before the glass transition, the viscosity increase rate with temperature decreased significantly. Near the glass transition temperature, this sample showed roughly a viscosity increment of seven orders of magnitude. Similar behavior was observed for Celecoxib/PVP/TPGS, having similar viscosity values as the Celecoxib/PVP sample near melt temperature, as the glass transition approached during cooling, the Celecoxib/PVP/TPGS sample showed a decrease of one decade in viscosity compared to Celecoxib/PVP. Celecoxib/PVP/TPGS viscosity increase rate also was reduced near the glass transition. Analyzing the viscosity data in terms of diffusivity indicates that the molecular mobility of the Celecoxib is impeded by the presence of the PVP, this occurs by the hindering of local arrangements of the Celecoxib molecules within a crystal lattice (crystallization inhibition). This was evidenced by the significant increase in viscosity and degree of undercooling compared to pure Celecoxib. This induced hindering of Celecoxib molecules allowed for viscosity measurements near the glass transition temperature. On the other hand, the TPGS acted as a plasticizer where miscibility between TPGS and Celecoxib aided in the molecular mobility of Celecoxib molecules assisting therefore the crystallization near the melting temperature. The combination of Celecoxib with PVP and TPGS resulted in interesting results, generally a decrease of molecular mobility similar to the effect of PVP alone, however the plasticizing effect of TPGS was able to increase the mobility, hence lower viscosity values near the glass transition compared to the effect of PVP. The viscosity curves showed mostly linear dependences with regard to temperature, however some deviation from linearity were observed most significantly for the samples containing PVP and PVP/TPGS. Fitting the viscosity curves with the Arrhenius and VTF equations evidenced these dependence. The hindering of local arrangements of the Celecoxib molecules (crystallization inhibition) by PVP was evidenced by the increase in activation energy of Celecoxib/PVP (115.96kJ/mol) compared to pure Celecoxib (92.28 kJ/mol) near the melt temperatures, whereas TPGS decreased the activation energy necessary for crystallization to occur (60.52kJ/mol), evidencing thus the plasticizing effect of TPGS. Interestingly, the combination of PVP and TPGS was not capable of reducing the activation energy (98.89kJ/mol), as significantly as PVP alone complementing the similar viscosity values near the melt temperature. Deviations from the Arrhenius equation were observed as the degree of undercooling progressed (10oC-80oC), evidenced by a decrease in the R2 value of up to 12% for the Celecoxib/PVP/TPGS sample, compared with R2 = 10% of Celecoxib/PVP sample. These deviations could be successfully corrected with the VTF model, increasing the R2 value to 99.9% for both samples within the same temperature range. A marked VTF-to-Arrhenius transition near the glass temperature was not observed for these two samples within the temperature range studied, however Arrhenius equation resulted in higher R2values compared to the VTF model (10% difference between both). VTF fitting demonstrated that all samples had a calculated strength parameter (D) of less than 10, confirming the kinetic fragility (crystallinity) of the liquids. However, no correlation could be observed between the impact in strength parameter, and the crystallization tendency and molecular mobility. The impact on the crystallization of Celecoxib by PVP (polymer) and TPGS (surfactant) was studied in terms of molecular mobility. Over all it was demonstrated that the polymer decreases the mobility of the drug, whereas the surfactant acts as a plasticizer. PVP hinders the mobility of Celecoxib molecules inhibiting to some degree the crystallization of the drug, evidenced by the significant increase in measured degree of undercooling, dynamic viscosity and calculated activation energy. The surfactant acted as a plasticizer aiding the mobility of the drug allowing therefore the crystallization of the drug to occur at earlier, and lowers the activation energy necessary for crystallization to occur. The dynamic viscosity of the Celecoxib with polymer and/or surfactant follows Arrhenius temperature dependence near the melting point (10oC of undercooling), however the VTF model results in higher agreement as the undercooling degree increases.