Norfloxacin–NSAID Pharmaceutical Salts: Structure–Property Relationships for Enhanced Solubility and Stability Authors: Jeannette Carolina Belmont-Sánchez, Carolina Alarcón-Payer, Antonio Frontera, Francisco Javier Acebedo-Martínez, Cristóbal Verdugo-Escamilla, Alicia Domínguez-Martín, Duane Choquesillo-Lazarte Published: 2026 ABSTRACT The use of antibiotics is usually associated with mucosal damage and antimicrobial resistance due to inefficient dosing. This issue is particularly relevant for norfloxacin (NOR), whose amphoteric nature leads to strongly pH-dependent solubility and low oral bioavailability. To overcome these limitations, this work presents three novel drug–drug pharmaceutical salts of NOR with the anti-inflammatory drugs ketoprofen (KET), dexketoprofen (DKT), and niflumic acid (NIF). Salts were synthesized by mechanochemical and slurry strategies and fully characterized using thermal, spectroscopic, and X-ray diffraction techniques. The crystal structure, along with DFT and QTAIM computational analyses, showed that salts are governed by tetrameric assembly formation. However, while NOR–KET and NOR–DKT are stabilized by robust charge-assisted hydrogen bonds, NOR–NIF packing arrangement is dominated by weaker hydrogen bonds together with π–π and CF3···π interactions. These supramolecular differences were directly reflected in the pharmaceutical performance. The salts suppressed the rapid hydration of NOR under aqueous environments and showed improved stability at physiological pH. Moreover, NOR–KET and NOR–DKT significantly enhanced the solubility of both NOR and the corresponding NSAIDs in PBS at pH 6.8, while all salts provided a more controlled NOR dissolution profile under acidic conditions. These findings demonstrate that salt formation is an effective strategy to overcome the solubility and stability limitations in NOR and NSAIDs, offering a framework for improving the performance not only of NOR but also of other APIs facing similar challenges. Introduction Over the last few decades the World Health Organization has warned about the continuous rise in the incidence of urinary and gastrointestinal infections, − mainly driven by the indiscriminate use of antibiotics and the emergence of antimicrobial-resistant bacteria. − This global health challenge highlights the urgent need for strict control and rational use of antibiotics. Besides antibiotic regulation, a major issue associated with many active pharmaceutical ingredients (APIs) is their poor oral bioavailability, which often necessitates the administration of high doses to achieve therapeutic plasma concentrations. , Antibiotics such as norfloxacin (NOR, Figure ) exemplify this problem. NOR is a fluoroquinolone antibiotic with potent activity against Gram-negative bacteria, widely prescribed for urinary, renal, and gastrointestinal infections. , Despite its efficacy, NOR belongs to Class II in the Biopharmaceutics Classification System, characterized by low solubility but high permeability. This results in a low oral bioavailability (30–40%), primarily due to its amphoteric nature and limited solubility under physiological pH conditions. Although NOR is highly soluble in acidic media and readily dissolves in the stomach, its absorption predominantly occurs in the intestine (duodenum and jejunum), where the near-neutral pH reduces its solubility, limiting intestinal absorption and requiring higher doses to achieve therapeutic levels, which can be problematic in acute infections that require a rapid therapeutic response. Furthermore, precipitation of NOR in the intestinal lumen increases its local concentration, potentially irritating the intestinal mucosa, disrupting the microbiota, and contributing to the development of bacterial resistance. , Molecular diagrams of norfloxacin (NOR), ketoprofen (KET), dexketoprofen (DKT) and niflumic acid (NIF).. Improving NOR solubility at physiological pH therefore represents a promising strategy to enhance its oral bioavailability and mitigate pharmaceutical drawbacks. In this context, crystal engineering, particularly the development of pharmaceutical multicomponent materials (PMMs), has emerged as an effective and versatile approach. , These systems are crystalline solids composed of an API and one or more coformer molecules, that are stabilized by noncovalent interactions, including hydrogen bonding, van der Waals forces, and π–π stacking interactions. , The resulting crystal structure allows the modulation of key physicochemical properties, including solubility, thermal stability, and mechanical behavior without altering the molecular structure of the individual components, thereby preserving their pharmacological activity. − One step forward in this strategy involves the use of a second API as coformer, resulting in dual-drug multicomponent systems with potential therapeutic synergy. − In the case of antibiotics, nonsteroidal anti-inflammatory drugs (NSAIDs), such as ketoprofen (KET), dexketoprofen (DKT), or niflumic acid (NIF) (Figure ) represent promising coformers, given their frequent coadministration in clinical settings where infectious diseases are accompanied by inflammation. − In fact, NSAIDs are among the most widely consumed drugs for the management of acute and chronic pain, the attenuation of inflammatory responses, and the reduction of fever. , Despite their extensive prescription, NSAIDs are also associated with a wide range of dose-dependent side effects, many of which are related to their limited aqueous solubility. Consequently, PMMs offer a rational strategy to simultaneously modulate the physicochemical properties of both NOR and NSAIDs while combining antibacterial and anti-inflammatory properties within a single crystalline phase. In this work, we present the synthesis and comprehensive characterization of three novel NOR–NSAID salts. A detailed study of the crystal structures and the noncovalent interactions within the crystal lattice is conducted to understand the physicochemical properties exhibited by the developed materials. The structure–property relationships established in this study provide a solid basis for the rational design of PMMs with improved solubility under physiological pH conditions, thereby promoting intestinal absorption and oral bioavailability, as well as potential synergistic therapeutic effects arising from the coadministration of both drugs within a single solid-state material. Experimental Section NOR, DKT, KET, NIF, and all solvents were purchased from Sigma-Aldrich (purity >98%, Sigma-Aldrich, St. Louis, MO, USA) and were used directly as received. NOR–NSAIDs Salt Synthesis The mechanochemical synthesis of NOR–KET and NOR–DKT salts was carried out via liquid-assisted grinding (LAG) in a RETSCH MM2000 ball mill, operating at 25 Hz frequency for 30 min. For this purpose, equimolar mixtures of NOR (0.5 mmol, 159.66 mg) and the corresponding NSAID (0.5 mmol, 127.14 mg) were weighed and placed into stainless steel jars of 20 mL of volume, along with two stainless steel balls of 7 mm diameter and 100 μL of methanol (MET). NOR–NIF was obtained via slurry operations, in which equimolar mixtures of NOR (0.5 mmol, 159.66 mg) and NIF (0.5 mmol, 141.11 mg) were weighed and placed in 5 mL vials along with 1 mL of MET and one magnetic stirrer. After 72 h of agitation, the suspension was centrifuged and the pellet air-dried. All solid products obtained were analyzed by PXRD to determine the formation of the novel materials. Single crystals of the phases were obtained by slow solvent evaporation at room temperature, using saturated solutions of the material obtained from the synthesis in ethyl acetate (ETA). Suitable crystals for SCXRD appeared after 2 days of slow solvent evaporation at room temperature. Characterization Techniques X-ray Diffraction Analysis Powder X-ray diffraction (PXRD) analysis was carried out at room temperature on a Bruker D8 Advance Vαrio diffractometer (Bruker-AXS, Karlsruhe, Germany) equipped with a LYNXEYE detector and CuKα1 radiation (1.5406 Å). Generator operating conditions were 40 kV and 40 mA and measurements were performed in transmission geometry between Mylar foils with the beam focused in the detector. The diffractograms were obtained over an angular range of 5–40° (2θ) with a step size of 0.02° (2θ) and a constant counting time of 5 s per step. Suitable crystals or Single-Crystal X-ray diffraction (SCXRD) were prepared under inert conditions using perfluoropolyether as protective oil for manipulation. Crystals were placed on MiTeGen Micromounts and data were collected with a Bruker D8 Venture diffractometer (Bruker-AXS, Karlsruhe, Germany) with graphite monochromated CuKα radiation (λ = 1.54178 Å). Integration and scaling of the intensity data were accomplished using the APEX4 suite. The structures were solved by intrinsic phasing using the ShelXT program, which revealed the position of all non-hydrogen atoms. These atoms were refined on F2 by a full-matrix least-squares procedure using the anisotropic displacement parameter. All hydrogen atoms were located in difference Fourier maps and included as fixed contributions riding on attached atoms with isotropic thermal displacement parameters 1.2 or 1.5 times those of the respective atom. The Olex2 software was used as a graphical interface. Intermolecular interactions were calculated using PLATON. Molecular graphics were generated using Olex2 and Mercury. The crystallographic data for the reported structures were deposited with the Cambridge Crystallographic Data Center as supplementary publication no. CCDC 2550127, 2550128 and 2550129. Computational Methods All quantum chemical calculations were computed either Gaussian-16. The initial geometries for the molecular assemblies (dimers, trimers, and tetramers) were taken directly from their experimental SCXRD structures. This approach was chosen to analyze supramolecular interactions as they exist in the solid state, thereby preserving the subtle balance of forces and molecular conformations present in the crystalline environment, rather than using optimized gas-phase geometries. Interaction energies were computed using Density Functional Theory (DFT) with the PBE0-D3/def2-TZVP model chemistry. The PBE0 hybrid functional provides a balanced description of both covalent and noncovalent interactions. Grimme’s D3 empirical dispersion correction was included to accurately describe the critical long-range dispersion forces, and the def2-TZVP basis set was used to ensure high accuracy. All interaction energies were corrected for Basis Set Superposition Error (BSSE) using the standard counterpoise method. Topological analysis of the electron density was conducted using the Quantum Theory of Atoms in Molecules (QTAIM) and the Non-Covalent Interaction (NCI) index, , as implemented in the AIMAll program package. The QTAIM analysis identified bond critical points (BCPs) and bond paths for intermolecular contacts, with hydrogen bond strengths estimated using the local energy density predictor, (EH-bond ≈ 0.5 × V). The NCI analysis provided visual confirmation of these interactions through reduced density gradient (RDG) isosurfaces. Finally, Molecular Electrostatic Potential (MEP) surfaces were generated using GaussView to assess intrinsic electron donor–acceptor propensities and identify favorable sites for hydrogen bonding. Fourier Transform Infrared Spectroscopy (FT-IR) Fourier transform infrared (FTIR) spectroscopic measurements were performed on a Bruker Tensor 27 FT-IR instrument (Bruker Corporation, Billerica, MA, USA) equipped with a single-reflection diamond crystal platinum ATR unit and an OPUS data collection program. The scanning range was 4000–400 cm–1, with a resolution of 4 cm–1. Thermal Analysis Simultaneous differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA) were performed in a NETZSCH STA 449 F5 calorimeter (NETZSCH Group, Germany). Experimental conditions: alumina (Al2O3) crucibles of 85 μL volume, atmosphere of dry nitrogen with 250 mL/min flow rate, and heating rates of 5 °C/min ranging from 25 to 300 °C. The calorimeter was calibrated with indium of 99.99% purity (m.p.: 156.4 °C; DH: 28.14 J/g). Stability Studies Stability in aqueous solution of the APIs and their derived PMMs was evaluated through slurry experiments, in which, an excess of the solid was added to 10 mL of buffer PBS pH 6.8 or of KCl pH 1.2. After 24 h of stirring at 25 °C in sealed vials, the solids were collected, filtered, dried at 35 °C and further analyzed by PXRD. The stability of the samples was also assessed under accelerated aging conditions (40 °C and 75% RH), placing powder samples in a Memmert HPP110 climate chamber (Memmert, Schwabach, Germany) up to 3 months. The samples were analyzed by PXRD periodically to evaluate possible phase transformation during the time of the experiment. Powder Dissolution Profile Powder dissolution profile studies were undertaken following the shake-flask method in KCl pH 1.2 and PBS pH 6.8. Saturated solutions of NOR, NSAIDs and NOR–NSAIDs salts, were prepared by adding an excess of solid to 10 mL of each buffer. During 24 h of stirring at 25 °C, aliquots of the solutions were filtered through 0.22 μm syringe filters, diluted to achieve a measurable concentration and analyzed via high-performance liquid chromatography (HPLC). HPLC studies were performed using a HPLC Agilent 1200 series and a C18 column (4.6 mm × 150 mm, 4 μm particle size, Phenomenex) set at 40 °C. The mobile phase consisted of (A) PBS, 0.04 M, pH 2.5 and (B) Acetonitrile (purity >99%, Sigma-Aldrich, St Louis MO, USA). The initial mobile-phase composition was 85% A and 15% B, which was held isocratically for 4.4 min. A linear gradient was then applied, increasing solvent B from 15% to approximately 70% between 4.4 and 5.5 min. This composition was maintained isocratically until the end of the run (11 min). Flow rate was adjusted to 1 mL/min and 10 μL of sample was injected. The solubility (expressed in mg/mL) was evaluated at 254, 278, and 289 nm maximums of absorbance of KET/DKT, NOR and NIF. Results and Discussion NOR–NSAID Salts Synthesis Initial attempts to obtain the NOR–NSAID phases were carried out using neat grinding to prioritize organic solvent-free synthesis routes. However, as shown in Figure S1, the PXRD patterns revealed only physical mixtures of NOR and the respective NSAID. To induce the formation of novel solid phases, LAG was employed. Mechanochemical approaches, and particularly LAG, have become widely used methodologies for the rapid, efficient, and sustainable preparation of pharmaceutical salts and cocrystals. Compared with conventional solution-based strategies, LAG drastically reduces solvent consumption, shortens reaction times, and often improves reaction selectivity and reproducibility. In this sense, the use of LAG is closely aligned with the principles of green chemistry, minimizing waste generation and e environmental impact. Moreover, mechanochemical strategies have been recognized as sustainable chemical processes aligned with the principles of green chemistry and the objectives of the United Nations Sustainable Development Goals. LAG reactions were first performed using water as liquid additive to maintain green processing conditions. Nevertheless, as shown in Figure S1 these reactions predominantly yielded a mixture of unreacted NSAID and an unknown phase, most likely a hydrated form of NOR, due to the already reported disposition of NOR to form hydrates. However, none of the NOR hydrates reported in the CCDC matches the phase observed in the aqueous LAG. , At this point, different organic solvents were screened. For NOR–KET and NOR–DKT, positive results were achieved using MET. Figure shows the PXRD patterns of the reaction products demonstrating the formation of new phases, with no signals of the parent components. PXRD patterns of the products obtained from LAG (NOR–KET and NOR–DKT) and slurry (NOR–NIF) reactions compared with the parent APIs and the calculated (Calc) PXRD pattern of the phases derived from their corresponding crystal structures. In contrast, LAG of NOR with NIF using MET and a variety of organic solvents produced physical mixtures of the components or poorly crystalline solids that could not be isolated (see Figure S2a). Poor crystalline products indicated amorphization, attributed to an excess of energy during the LAG reaction. For this reason, a less energetic and more thermodynamic approach was adopted: slurry crystallization. In this method, the components were suspended in a small volume of MET (to minimize dissolution of the APIs) and allowed to equilibrate under continuous stirring. Experiments were performed at 24, 48, 72, and 96 h (Figure S2b), but only after 72 h, a novel crystalline phase was obtained, without any signals of the parent APIs (Figure . This indicates that NOR–NIF system does not benefit from more energetic techniques such as LAG. Furthermore, the absence of product formation at 24 and 48 h suggests that the phase emerges only after a prolonged equilibration period, indicating that the higher volume of solvent used in slurry (compared to LAG) is not the main reason for the formation, but rather the time of equilibrium and reaction. At this stage, PXRD alone was insufficient to distinguish whether the materials corresponded to a new salt, cocrystal, or a derived phase from an API (such as solvates or polymorphs). For this reason, FT-IR spectroscopy was used as complementary technique. Shifts in characteristic bands of functional groups that are involved in supramolecular interactions supported the formation of PMMs and provide information about the nature of the materials. Figure S3 shows the FT-IR spectra of the NOR–NSAID products, while the most relevant FT-IR bands are summarized in Table for clarity. 1Main FT-IR Bands (cm–1) Ascribed to the Principal Functional Groups of NOR and NSAIDs, along with the Shift of the Bands in the Respective PMMscompoundυ(N–H)υ(O–H)acid υ(C=O)acid υ(C–F) NOR 3600–32003090–242217271102 KET 3216–23281696 DKT 3408–25601730 NIF 3391–32413136–238716661109 NOR–KET 3060235317091139 NOR–DKT 3055235017141134 NOR–NIF 3055235617071105 Main FT-IR Bands (cm–1) Ascribed to the Principal Functional Groups of NOR and NSAIDs, along with the Shift of the Bands in the Respective PMMs For KET, the υ­(O–H) stretching is observed between 3216–2328 cm–1, while the υ­(C = O) band of the carboxylic acid appears at 1696–1655 cm–1. In DKT, the υ­(O–H) band is detected at 3408–2560 cm–1 and the υ­(C = O) stretching associated with the acid group at 1730–1650 cm–1. Similarly, in NIF, the υ­(O–H) stretching lies within 3136–2387 cm–1, and the carboxylic υ­(C = O) band appears at 1666 cm–1. In NOR–KET, NOR–DKT, and NOR–NIF, the carboxylic υ­(C = O) bands of the NSAIDs appear at lower wavenumbers compared to the pure APIs, being observed at 1709 cm–1, 1714 cm–1, and 1707 cm–1, respectively. These shifts are consistent with the deprotonation of the carboxylic acid and the formation of a carboxylate stabilized through interaction with NOR. The disappearance or marked alteration of the υ­(O–H) stretching bands in the resulting solids aligns with the loss of the acidic proton. Meanwhile, NOR exhibits the characteristic signals of amine and amide functionalities, with the υ­(N–H) stretching at 3600–3200 cm–1, the amide υ­(C = O) at 1620 cm–1, and the υ­(C–N) stretching around 1250 cm–1. The amine and amide groups of NOR also exhibit changes after interaction with the NSAIDs. The υ­(N–H) stretching band, originally located at 3600–3200 cm–1, appears at 3060 cm–1 in NOR–KET and at 3055 cm–1 in both NOR–DKT and NOR–NIF. Likewise, the υ­(C–N) stretching band of NOR, initially at 1250 cm–1, shifts to 1273 cm–1 in NOR–KET, 1271 cm–1 in NOR–DKT, and 1273 cm–1 in NOR–NIF. These results point to the formation of pharmaceutical salts, stabilized by NH2 +···OOC–hydrogen bonds. Crystal Structure Analysis To corroborate PXRD and FT-IR results, single crystals were grown and subjected to SCXRD analysis. The resulting crystal structures confirmed that all NOR–NSAID are true pharmaceutical salts formed between protonated NOR (NOR +) and the deprotonated NSAID (KET –, DKT – or NIF –). In addition, the crystal structures allowed the calculation of PXRD patterns, which, upon comparison with the experimental ones (Figure ), verified the purity and reproducibility of the synthesis reactions. NOR–KET and NOR–DKT display closely related crystal structures due to their chemical relationship. KET is a racemic mixture containing both R- and S-enantiomers, whereas DKT is the enantiopure S-form. This difference is relevant for understanding the crystallographic description of both salts, particularly regarding the content of the asymmetric unit and the role of symmetry operations within the crystal structure. NOR–KET crystallizes in the triclinic system, P1̅ space group, with a 1:1 stoichiometric ratio (the asymmetric unit is composed of one NOR + and one KET –). The ionic pair is stabilized by a strong charge-assisted hydrogen bond (CAHB) between the deprotonated carboxylate group of KET – and the piperazine ammonium group of NOR +. This main NH2 +···OOC– interaction is described by an D 1 1(2) graph set (Figure a). In the asymmetric unit, an additional intramolecular hydrogen bond within NOR + allows to improve the cohesion of the ionic pair. Because NOR–KET crystallizes in the centrosymmetric P1̅ space group, the inversion center generates a second symmetric ionic pair containing the opposite KET enantiomer. These two ionic pairs assemble through equivalent N–H···OOC interactions to generate a tetrameric unit, resulting in a ring-type hydrogen-bond framework described by an R 4 2(8) graph set (Figure b). Within this synthon, each NOR + contributes the two hydrogen atoms of its protonated piperazine, and each KET – uses one oxygen atom of its carboxylate group to form two N–H···OOC hydrogen bonds. To generate the three-dimensional structure, tetramers are further connected through reciprocal C–H···O interactions between adjacent NOR + molecules, generating zigzag chains of NOR + cations that extend along the ac axis, while KET – anions remain laterally disposed (Figure b). Structural features of NOR–KET. a) Major D 1 1(2) graph set established between NOR + and KET – in the asymmetric unit. b) Supramolecular organization of NOR + (orange) and KET – (blue). Green boxes highlight the connection of tetrameric units by R 4 2(8) graph sets. Atoms involved in the interactions are presented as balls, while the rest are presented as sticks. NOR–DKT also crystallizes in the triclinic system but in the noncentrosymmetric P1 space group. Consequently, the asymmetric unit contains two independent NOR + cations and two DKT – anions, all involving the S-enantiomer (DKT). These two ionic pairs are directly present in the asymmetric unit and assemble into a tetrameric architecture (R 4 2(8)) similar to that observed in NOR–KET. Therefore, although the Z values presented in Table differ between the two salts, both unit cells contain two NOR–NSAID ionic pairs. The three-dimensional organization in NOR–DKT follows the same structural pattern described for NOR–KET. The close structural similarity between both salts is supported by molecular overlay analysis (Figure ), which gives root-mean-square deviation (RMSD) values of 0.027 Å for NOR + and 0.1098 Å for the NSAIDs –, indicating only minor deviations in molecular conformation, arising from the use of enantiopure DKT. This high degree of similarity is also reflected in the comparable unit-cell parameters summarized in Table . Overlap between two independent molecules of NOR + and NSAIDs – , in the crystal structures of NOR–KET (blue) and NOR–DKT (red). Hydrogen atoms are omitted for clarity. 2Crystallographic Data of NOR–KET, NOR–DKT and NOR–NIF compound nameNOR–KETNOR–DKTNOR–NIFCCDC number 2550127 2550128 2550129 crystal systemtriclinictriclinictriclinicspace group P1̅(2) P1(1) P1̅(2) a [Å]8.3289(2)8.3654(13)8.056(4) b [Å]13.1499(2)13.068(2)11.090(6) c [Å]13.7364(2)13.747(3)15.503(7)α [°]103.4850(10)103.009(9)101.59(2)β [°]106.0490(10)106.138(7)92.70(2)γ [°]91.3220(10)91.476(6)93.43(2)volume [Å3]1399.79(5)1400.3(4)1351.9(11) Z 212ρcalc[g cm–3]1.3611.3601.478 F(000)604604624reflections collected242333738420807independent reflections491292384727 R int = 0.0393 R int = 0.0552 R int = 0.0565data/restraints/parameters4912/0/4029238/3/7744727/78/419goodness-of-fit on F 2 1.0491.0681.052final R indexes R 1 = 0.0556 R 1 = 0.0565 R 1 = 0.0483[I ≥ 2σ(I)]wR 2 = 0.1415wR 2 = 0.1435wR 2 = 0.1247 Crystallographic Data of NOR–KET, NOR–DKT and NOR–NIF Similarly, NOR–NIF crystallizes in the triclinic system, P1̅ space group with an asymmetric unit containing one NOR + and one NIF –. The ions are connected through the same CAHB, NH2 +···OOC– (D 1 1(2) graph set) observed in the previous salts. However, in this case the ionic pair is oriented, creating a right angle of almost 90 ° (Figure a). Within the asymmetric unit, two additional intramolecular interactions are observed. In NOR + , a strong O–H···O hydrogen bond is established between the carboxylic acid hydroxyl group and the adjacent quinolone oxygen atom, allowing the carboxylate group’s orientation to be locked. In NIF –, the interaction is an N–H···O H hydrogen bond between the carboxylic acid hydroxyl group and the adjacent nitrogen atom, which is also key to maintaining the planar conformation of the molecule’s core. As a consequence of the centrosymmetric the centrosymmetric P1̅ space group, the ionic pair is related to a second NOR + /NIF – pair by inversion symmetry, which leads to the formation of a tetrameric assembly inside the unit cell (Figure b). However, in this particular case, the two ionic pairs are connected via π···π interactions between the aromatic rings of NIF + . As shown in Figure c, these assemblies propagate along the b axis through N–H···O interaction between adjacent NOR +, generating one-dimensional chains. In this arrangement, NOR + molecules remain approximately parallel to the chain direction, while NIF – anions are disposed nearly perpendicular. Finally, the one-dimensional chains are connected through additional π–π interactions involving NOR+ molecules, leading to the final three-dimensional supramolecular structure. Structural features of NOR–NIF. a) Asymmetric unit and the major D 1 1(2) graph set established between NOR + and NIF –. b) Tetrameric formation inside the unit cell. c) Supramolecular organization of NOR + (orange) and NIF – (green). Atoms involved in the interactions are presented as balls, while the rest are presented as sticks. Gray box represents π···π stacking interaction areas. Computational Studies Following the successful synthesis and SCXRD characterization of the three pharmaceutical salts, we performed a detailed theoretical analysis of the observed solid-state supramolecular architecture. To understand the intrinsic electron donor–acceptor ability of the coformers and rationalize the observed CAHB preferences, we computed the MEP surfaces for the isolated ionic species at the same level of theory used for the full interaction energy calculations, specifically PBE0-D3/def2-TZVP. The computed MEP surfaces for the DKT – (serving also as model for the KET –), NIF –, and NOR + are shown in Figure . The MEP surfaces for DKT – and NIF – , are negative across the entire surface due to the overall negative charge, confirming their nature as electron-rich hydrogen bond acceptors. The most electron-rich regions (MEP minima) are quantified as follows: for DKT – , the absolute MEP minimum is located at one of the carboxylate oxygen atoms with a value of – 148.0 kcal/mol, with a second significant negative region found at the nearby carbonyl oxygen atom, with a MEP value of – 83.5 kcal/mol. The MEP minimum for NIF – is also located at the carboxylate oxygen atom, measuring – 137.8 kcal/mol. Another important electron-rich region is observed at the pyridine nitrogen atom, with a value of – 77.8 kcal/mol. The less negative MEP minimum observed for NIF – compared to DKT – (−137.8 vs – 148.0 kcal/mol) suggests that DKT – /KET – anion is intrinsically a stronger hydrogen bond acceptor. This decrease in the MEP value at the NIF – carboxylate group is most likely attributable to the already described intramolecular N–H···O hydrogen bond, which partially shields or disperses the negative charge, thus rationalizing why DKT – /KET – are better CAHB acceptors than NIF – . The positive MEP values for NOR + confirms its strong electron-poor hydrogen bond donor ability, with the MEP maxima located at the hydrogen atoms of the protonated piperazine ring’s ammonium group: the axial N–H bond exhibits a slightly larger MEP maximum of 161.3 kcal/mol, and the equatorial N–H bond is slightly less positive, with a value of 160.6 kcal/mol. These data confirm that the ammonium group is the primary site for the formation of N–H···OOC CAHBs, supporting the SCXRD observations. Importantly, the MEP values at the oxygen atoms of the NOR + carboxylic acid group are slightly negative, specifically – 15.7 kcal/mol for the carbonyl oxygen and – 6.7 kcal/mol for the hydroxyl oxygen. This residual negative potential provides a rational explanation to the NOR + ⋯NOR + chain formation via N–H···O interactions observed in the crystal structure of NOR–NIF. MEP surfaces for the isolated ionic species: a) DKT – , b) NIF – , and c) NOR + . The surfaces map the potential onto the 0.001 au isodensity surface, showing electron-rich regions (red/negative) and electron-poor regions (blue/positive). Key MEP minimum (negative) and maximum (positive) values are labeled, indicating the primary hydrogen bond acceptor sites (carboxylate oxygens in a, b) and donor sites (ammonium N–H groups in c). The influence of the intramolecular interactions in NOR + and NIF – leads us to perform a topological analysis of the electronic density, based on the QTAIM. The analysis of the isolated NOR + and NIF – molecules is visualized in Figure . The strength of these interactions, estimated using an energy predictor based on the potential energy density at the bond critical point (E ≈ 0.5 × V, see theoretical methods). In NOR + the O–H···O presented an estimated energy of – 12.5 kcal/mol, which is significantly stronger than the C–H···F interaction (−3.9 kcal/mol). Similarly, in NIF – the N–H···O contact presented an estimated energy of – 7.9 kcal/mol, while the second C–H···N interaction is – 3.2 kcal/mol. In contrast to NOR + and NIF – , DKT – (and thus KET –) shows no evidence of intramolecular hydrogen bonds, meaning its conformational flexibility in the gas phase is solely governed by rotational barriers around single bonds. QTAIM analysis for the optimized gas-phase structures of a) NOR + and b) NIF – . Intramolecular interactions are visualized by dashed brown lines (bond paths) and characterized by a small red sphere (BCP). The corresponding stabilization energies for the interactions, are labeled in kcal/mol. The primary supramolecular synthons observed in the crystal structures were analyzed using a combination of QTAIM and NCIplot methods to quantify the contributions of the individual interactions. Figure illustrates the results for the dominant assemblies in the three salts. SXCRD demonstrated that NOR–KET and NOR–DKT are governed by the formation of the R 4 2(8) graph set, which involves two NOR + cations and two KET – /DKT – anions, resulting in an eight-membered supramolecular ring. Each N–H···O hydrogen bond is characterized by a BCP, a corresponding bond path, and a localized blue RDG isosurface, consistent with strong, CAHB. Furthermore, the QTAIM analysis identified two ancillary C–H···O hydrogen bonds, involving the carboxylate oxygen atoms that do not participate in the CAHBs. Due to the inherent ion-pair nature of the assembly, the calculated total formation energies for the tetramers are substantial: – 252.4 kcal/mol for NOR–KET and – 251.3 kcal/mol for NOR–DKT. To isolate the intrinsic strength of the interactions from these large, nondirectional Coulombic effects, the QTAIM energy predictor (E ≈ 0.5 × V) was applied. This analysis revealed very large stabilization energies for the individual N–H···O contacts forming the R 4 2(8) synthon: – 34.6 kcal/mol in NOR–KET and – 35.5 kcal/mol in NOR–DKT. Such high values confirm the crucial role and dominance of these CAHBs in dictating the solid-state structure of both salts. The secondary C–H···O contacts provide only minor stabilization, with energies of – 2.9 kcal/mol (NOR–KET) and – 2.8 kcal/mol (NOR–DKT). NOR–NIF exhibits a distinct trimeric assembly with a calculated total formation energy of – 125.7 kcal/mol. These trimeric units are the main source of union between the tetramers described in the crystal structure section. Here, the QTAIM analysis shows a bifurcated N–H···O,O hydrogen bond between the two cationic units, characterized by two BCPs, bond paths, and localized green RDG isosurfaces and established between the equatorial N–H bond and both O-atoms of the carboxylic group. The connectivity between the cation and anion is achieved through three BCPs and bond paths: one primary N–H···O CAHB, and two corresponding to bifurcated C–H,H···O contacts. The N–H···O interaction in NOR–NIF is very strong, evidenced by the intense blue color of its RDG isosurface. The estimated strength of all three NH···O contacts is – 25.2 kcal/mol, which is noticeably smaller than those found in NOR–KET and NOR–DKT tetramers. In contrast, the energetic contribution of the secondary C–H···O contacts in NOR–NIF trimer is very similar to the other salts, at – 2.6 kcal/mol. QTAIM and NCIplot analysis of the dominant assemblies in the pharmaceutical salts. a,b) R 4 2(8) synthon formed by the tetrameric unit observed in NOR–KET and NOR–DKT. c) Trimeric assembly in NOR–NIF. The interactions are characterized by BCPs (small red spheres) and Bond Paths (brown dashed lines). The RDG isosurfaces are mapped onto the sign of the second density Hessian eigenvalue (sign­(λ2)­ρ), where intense blue represents strong CAHBs and green represents weaker van der Waals interactions. The hydrogen-bond networks observed in NOR–KET and NOR–DKT constitute the main source of stabilization. In contrast, although the tetrameric assemblies in NOR–NIF are propagated into chains through the N–H···O hydrogen-bonding interactions described above (trimeric assemblies), the association between adjacent chains is mainly governed by nonclassical noncovalent interactions (Figure ). While only π···π interactions were described in the crystal-structure section, the computational analysis reveals that this stabilization also involves LP···π interactions, leading to a large formation energy of – 116.5 kcal/mol. This value is comparable to that calculated for the hydrogen-bonded trimeric assembly (Figure c), confirming the relevance of these contacts in controlling the final solid-state packing of NOR–NIF. DFT calculations also revealed CF3···π interactions that are topologically characterized by three BCPs and corresponding bond paths, interconnecting the three fluorine atoms of the CF3 group to the π–system of the quinoline ring (Figure ). The NCIplot method further characterizes this interaction through localized green RDG isosurfaces, consistent with weak, dispersion-driven contacts. The central motif involves a strong π–stacking interaction between the two adjacent NOR + molecules. In this region, the green RDG isosurfaces extend over the entire space between the molecules, indicating a strong, stabilizing complementarity between the aromatic rings, supporting the crystal structure analysis. It is noteworthy that the stacking is not limited to the aromatic cores; the aliphatic piperazine ring of the NOR + unit is also stacked over the carboxylic group of the quinoline ring of the adjacent NOR + unit. This additional C–H···O interaction is topologically confirmed by the presence of three BCPs and bond paths connecting C–H groups of the piperazine ring to the oxygen atoms of both the carboxylic and quinone functionalities of the neighboring NOR + cation, thereby maximizing the total cohesive energy of the assembly. Topological analysis, using QTAIM and NCIplot method, of the noncovalent interactions present in the supramolecular assembly of NOR–NIF tetrameric units. BCPs are represented as small red spheres and Bond Paths as brown dashed lines. The RDG isosurfaces are mapped onto the sign of the second density Hessian eigenvalue (sign­(λ2)­ρ). Stability Studies Having established the supramolecular organization and quantified the interaction energetics, the next step is to evaluate how these structural features impact the physicochemical properties of the novel salts. The thermal behavior was evaluated by simultaneous DSC–TGA analysis. As shown in Figure , the DSC thermograms display a single, well-defined endothermic event for each compound, corresponding to their melting point. The absence of additional thermal events in the DSC, along with the lack of mass loss before melting in the TGA traces (Figure S4), discards polymorphic transitions, or dehydration events before melting, while confirming phase purity. DSC traces of the NOR–KET, NOR–DKT and NOR–NIF along with the melting point of the APIs (dotted lines). The melting point of the salts fall in between the melting point of pure NOR (221 °C) and the respective NSAIDs, which is a common behavior of PMMs. , NOR–KET and NOR–DKT displayed significantly increased melting points (192.7 and 195.9 °C, respectively) relative to KET (75 °C) and DKT (96 °C). This enhancement correlates directly with the high stability of the R 4 2(8) tetrameric assembly identified in the salts, thus requiring higher thermal input to disrupt when compared with the dimeric association present in the native NSAIDs. Interestingly, NOR–NIFexhibited the lowest melting point (153.6 °C), despite NIF presenting the higher melting point (203 °C) among the NSAIDs. Unlike the tetrameric organization in NOR–KET and NOR–DKT, NOR–NIF adopts a trimeric assembly with weaker CAHBs due to the presence of an intramolecular hydrogen bond in NIF –, which delocalizes acceptor electron density and diminishes CAHB efficiency. The stability of the salts under accelerated aging conditions was also evaluated by storing the samples at 40 °C and 75% RH for three months (Figure ). Under these conditions, pure NOR transformed into its known sesquihydrate form (NOR·1.5H2O, CCDC 720746) in less than 24 h. In contrast, all three NOR–NSAID salts remained stable, retaining their original PXRD patterns and crystallinity. The strong hydrogen bonding between the NSAIDs and the protonated nitrogen atom in NOR’s piperazine ring blocked the hydration pathway, as the same nitrogen is involved in the coordination of water in NOR sesquihydrate. PXRD of the salts and the parent APIs under accelerated aging conditions. NOR·1.5H2O, CCDC number 720746, is also represented for comparison. The thermodynamic stability in aqueous solution was assessed under two physiologically relevant conditions, KCl at pH 1.2 and PBS at pH 6.8, simulating the stomach and intestinal environments. As shown in Figure , at pH 1.2, NSAIDs remained stable under both pH conditions, whereas NOR transformed into the same unidentified phase observed during water-LAG, suggesting the formation of a NOR hydrate. In parallel, all salts underwent dissociation, with PXRD patterns revealing precipitated NSAIDs. This behavior is consistent with the amphoteric nature of NOR, which becomes protonated in acidic media, disrupting the salt and increasing NOR solubility while leaving the NSAID undissolved. PXRD patterns of the parent APIs and the novel NOR–NSAID salts after 24 h in aqueous suspension in KCl at pH 1.2 and PBS at pH 6.8. Despite the instability of the salts observed at acidic pH, this does not represent a pharmacological limitation, as both NOR and NSAIDs are primarily absorbed in the intestinal tract and gastroprotective formulations could be employed. At pH 6.8, NOR–NSAID salts remained chemically stable, with only NOR–NIF showing partial loss of crystallinity, which is a consequence of the trimeric assemblies that led to a less cohesive hydrogen bonding network, as demonstrated by the DFT calculations. These results demonstrated that salt formation with NSAIDs effectively protected NOR against hydration, not only under environmental humidity but also in aqueous solution. Powder Dissolution Profile The solubility behavior was evaluated under physiological conditions (PBS pH 6.8) and acidic medium (KCl, pH 1.2) (Figure ). The good agreement between the reported solubility values of the APIs and those obtained in this work confirmed the reliability of the experimental data. − In PBS, NOR–KET allowed a marked enhancement in the solubility of both components, reaching concentrations of 0.96 mg/mL for NOR and 0.89 mg/mL for KET within the first 50 min, followed by a sustained increase over 24 h, reaching equilibrium concentrations of 1.42 mg/mL (NOR) and 1.20 mg/mL (KET). In the same conditions, native NOR showed a 5-fold lower equilibrium solubility (0.294 mg/mL) along with a less stable and more erratic supersaturation profile in the first minutes. Pure KET also displayed very limited solubility (∼0.2 mg/mL), reaching equilibrium within the first minutes of the experiment. A similar trend was observed for NOR–DKT, consistent with the structural similarity between both salts. The system reached equilibrium solubilities of 1.52 mg/mL for NOR (x5 increase) and 1.12 mg/mL for DKT (x4 increase). Solubility profiles of the APIs and the salts at KCl pH 1.2 and PBS pH 6.8. In contrast, NOR–NIF achieved 0.068 mg/mL for NOR and 0.089 mg/mL for NIF. This downregulation (x4) is consistent with the intrinsically poor solubility of NIF (∼0.02 mg/mL), which limits the water uptake in the crystal lattice and the dissolution of the salt. No pronounced spring–parachute phenomenon was observed for NOR–NIF, despite NOR and NIF presenting supersaturation phenomena in the early stages. In KCl pH 1.2, as demonstrated in the stability section, the salts underwent dissociation, thus the solubility of the NSAIDs remained comparable to that of the native drugs (∼0.09–0.10 mg/mL). Native NOR displayed very high solubility, reaching an initial supersaturation peak of 8.58 mg/mL and stabilizing at approximately 9.40 mg/mL. Interestingly, all salts exhibited a more controlled dissolution profile for NOR, with no supersaturating effects, lower initial concentrations and a gradual increase over time. For NOR–KET, concentrations increased from ∼ 2 mg/mL to ∼ 5 mg/mL over 48 h, while NOR–DKT showed a faster initial release (∼4 mg/mL) followed by a slower increase toward similar equilibrium values. Salt formation and further dissociation in acid media allow the salts to act as reservoirs that modulate NOR release and the final equilibrium concentrations of NOR. Notably, NOR–NIF led to the lowest solubility of NOR (∼2 mg/mL). In this case, the structure is dominated by hydrophobic π···π stacking and CF3···π interactions with a reduced contribution from CAHBs, which decrease the affinity of the affinity of the system for aqueous molecules, leading to reduced dissolution. These results are particularly relevant from a pharmacological perspective, as controlling the high solubility and supersaturation of NOR at acidic pH, where absorption is minimal, may reduce local accumulation and associated side effects. Conclusions This work demonstrates that drug–drug salt formation between NOR and NSAIDs is a powerful strategy to simultaneously tune stability, solubility, and release behavior. Structural and computational analyses revealed that robust CAHB-stabilized tetrameric architectures in NOR–KET and NOR–DKT lead to highly stable crystal packings, whereas π-stacking and CF3···π interactions dominate the structure of NOR–NIF, resulting in weaker cohesion. The structural differences directly translate into pharmaceutical performance. The cohesive tetrameric structures of NOR–KET and NOR–DKT significantly improved the thermal stability of the NSAIDs, approximately doubling their melting points, whereas the less cohesive NOR–NIF structure led to less thermal stability. Additionally, salt formation suppressed NOR hydration by blocking its primary water coordination site, enhancing the stability under both accelerated aging and aqueous conditions. At the same time, the salts modulated the dissolution behavior, enhancing NOR solubility under physiological conditions while reducing excessive supersaturation in acidic media through a reservoir-like release mechanism. Interestingly, the lower number of CAHBs combined with the greater contribution of hydrophobic π–π interactions in NOR–NIF resulted in a less water-accessible structure and reduced NOR solubility. Overall, this study highlights how rational crystal engineering enables direct control of macroscopic drug performance. Beyond combining therapeutic functionalities, this approach provides a robust framework for designing solid forms with optimized stability, bioavailability, and controlled release, establishing a solid foundation for future pharmaceutical development. Supplementary Material