*3.2. Characterization of the Drug–CD Binary Systems*

DSC curves of pure HCT and CDs and of their binary systems obtained with the different techniques were recorded in order to investigate the effect of the CD type and preparation technique towards drug amorphization and complexation (Figure 2A). Thermograms of pure drug submitted to the different treatments used for the preparation of the binary systems were also recorded, in order to evidence any effect of the preparation technique on drug solid-state properties.

**Figure 2.** DSC curves of pure HCT (**A**) untreated or submitted to kneading (KN), grinding (GR) or dissolution-solvent evaporation (EV) and its binary systems with βCD (**B**), HPβCD (**C**), HEβCD (**D**), SBEβCD (**E**), and RAMEB (**F**), obtained by physical mixing (i.e., PM), co-grinding (GR), kneading (KN), and co-evaporation (COE).

The drug alone thermogram showed a sharp melting peak (Tpeak <sup>=</sup> 274.4 ◦C and <sup>Δ</sup><sup>H</sup> <sup>=</sup> 152.8 J·g<sup>−</sup>1), as expected for an anhydrous crystalline compound. The kneading does not seem to affect the thermal behavior, causing just a slight decrease in Tpeak (272.2 ◦C) and RDC (98.96%).

The thermal curve of the drug submitted to the grinding process showed an exothermic phenomenon at 134.5 ◦C which is attributed to the recrystallization of the drug fraction amorphized during the grinding procedure; it was followed by the melting peak at 271.2 ◦C, whose reduced intensity (ΔHfus128.1 Jg−1) indicated some loss of crystallinity (RDC 83.23%). In the evaporated product, the drug melting peak at 274.1 ◦C was characterized by a sensible reduction of fusion enthalpy (72.1 Jg<sup>−</sup>1). The calculated RDC of 46% suggested an appreciable drug amorphization, as a consequence of its dissolution in the water-ethanol mixture and the following solvent evaporation.

In Figure 2B, the thermal profiles of HCT binary systems with βCD are reported. The DSC curve of βCD in the examined temperature range was characterized by an intense and broad endothermic effect that ranged between 50 to 130 ◦C, due to its dehydration. The PM showed the CD dehydration band, followed at higher temperatures by an endothermic event characterized by the presence of two peaks at 265 ◦C and 269 ◦C. These can be attributed to the partial superimposition of the drug melting and the CD thermal decomposition, both shifted at lower temperature than the corresponding pure components, due to their co-presence, as observed also by other authors [27,28]. In fact, CD alone started to decompose usually over 300 ◦C [29]. A thermal behavior substantially similar to that of PM was found for GR, KN, or COE products, suggesting poor host–guest interactions. A rather similar behavior (Figure 2C) was observed for the series of products with SBEβCD. As previously observed by Cirri et al., 2017, the thermal profile of this CD showed a broad endothermic effect between 60 to 110 ◦C due to its dehydration, and another endothermic band between 250 and 280 ◦C, due to degradation phenomena [30]. All binary systems prepared with the different techniques showed the superimposition of HCT melting and CD decomposition phenomena, similar to that observed in the simple PM. As for the thermal profiles of binary systems with HPβCD, HEβCD and RAMEB (Figure 2D–F, respectively) after the initial broad endothermic effect, due to the amorphous CD dehydration, the drug and CD endothermic decomposition phenomena in the range 250–260 ◦C was still barely detectable only for the PM with HPβCD, while the drug melting peak completely disappeared in all other systems, where only the CD decomposition band was observed.

In order to better elucidate the DSC findings and, in particular, to evidence any possible artifact of the technique, due to a heating-induced interaction between the components as a consequence of the thermal energy supplied to the sample during the DSC scans [31], XRPD analysis was conducted. As shown in Figure 3A, the patterns of pure drug and βCD presented several sharp peaks, characteristic of crystalline substances, whereas all βCD derivatives, as shown, as example, for HPβCD, presented an almost flat pattern typical of amorphous substances. Representative drug peaks were clearly detectable in the patterns of all PMs (Figure 3B), even if reduced in intensity, particularly, in combinations with the amorphous partners, indicating that any solid-state interaction occurred during the simple mixing of drug and CDs. However analogous results to those of PMs were also obtained for all binary systems with βCD and different CD derivatives, with the only exception for COE products obtained with RAMEB and SBEβCD (Figure S1, Supplementary Material). Thus, the disappearance of the drug melting peak, observed in the other cases can be attributed to heating-induced interactions due to the thermal energy supplied to the sample during the scan. However, the completely amorphous pattern exhibited by the COE products with RAMEB and SBEβCD proved that the actual drug amorphization and complexation was achieved only by this preparation technique with these CDs (Figure 3C).

**Figure 3.** XRPD patterns of pure HCT and CDs (**A**) and of equimolar physical mixtures (PM) (**B**) with all CDs and the co-ground (GR), kneaded (KN), and co-evaporated (COE) products with RAMEB (**C**).

The results of dissolution rate studies performed on pure drug, both untreated and submitted to the different techniques and on the different binary systems with all the CDs, are presented in Figure 4.

**Figure 4.** Dissolution profiles of pure HCT (**A**) untreated ( red line) or submitted to kneading ( blue line), grinding ( green line), or dissolution-solvent evaporation (x violet line) and from its binary systems with βCD (**B**), SBEβCD(**C**), and RAMEB (**D**), obtained by physical mixing (PM) ( red line), kneading (KN) ( blue line), co-grinding (GR) ( green line), and co-evaporation (COE) (x violet line). Each value represents the mean of 3 experiments.

As shown, HCT alone (Figure 4A) always reached a solubility of 1 mg/mL, even if the treatments, especially the grinding one, initially allowed a more rapid release, probably due to a particle size reduction of the powder. In binary systems with βCD (Figure 4B) a better profile was also observed for the simple PM that reached a plateau level of 1.45 ± 0.5 mg/mL, due to the improved drug wettability and possible in situ complexation phenomena. A little more favorable effect on the drug dissolution rate, particularly in the first minutes, was observed for all treated products, in virtue of the more intimate contact between the components brought about by the sample preparation method. Similar results were obtained for drug systems with HEβCD and HPβCD (data not shown), regardless of the sample preparation technique. Instead, significant differences were observed for the series of products with SBEβCD and RAMEB (Figure 4C,D) especially for COE systems, which achieved plateau levels around 5.5 ± 0.3 and 6.8 ± 0.2 mg/mL, respectively. It should be pointed out that the COE systems with SBEβCD and RAMEB were the only ones actually containing the drug in an amorphous/complexed status, as revealed by XRPD analysis. Moreover, the results were in accordance with phase solubility studies, where SBEβCD and RAMEB showed the higher complexing and solubilizing efficiency towards HCT.

Taking into account these results, co-evaporation was selected as the most effective preparation technique and RAMEB as CD that lead to the best drug dissolution profile.
