Effects of Polymeric Additives on the Crystallization and Release Behavior of Amorphous Ibuprofen

Some polymeric additives were studied to understand their effects on the amorphous phase of ibuprofen (IBU), used as a poorly water soluble pharmaceutical model compound.The amorphous IBU in bulk, as well as in nanopores (diameter ∼24 nm) of anodic aluminum oxide, was examined with the addition of poly(acrylic acid), poly(N-vinyl pyrrolidone), or poly(4-vinylphenol). Results of bulk crystallization showed that they were effective in limiting the crystal growth, while the nucleation of the crystalline phase in contact with water was nearly instantaneous in all cases. Poly(N-vinyl pyrrolidone), the most effective additive, was in specific interaction with IBU, as revealed by IR spectroscopy.The addition of the polymers was combined with the nanoscopic confinement to further stabilize the amorphous phase. Still, the IBU with addition of polymeric additives showed sustained release behavior. The current study suggested that the inhibition of the crystal nucleation was probably the most important factor to stabilize the amorphous phase and fully harness its high solubility.


Introduction
Enhancing the bioavailability of active pharmaceutical ingredients (APIs) is becoming increasingly important because of the mounting problem of their sparing solubility.The solubility problem has escalated in the recently developed APIs, as it occurs in about 60% of them [1].To overcome this problem, the enhancement of bioavailability is realized by the increase of the dissolution rate.In the dissolution rate, the importance of the surface area () and solubility (  ) is well recognized, for example, from the Noyes-Whitney equation: (dissolution rate) =   (  −  ), where   is the rate constant and   is the bulk concentration [2][3][4].
The increase of the surface area has been investigated through nanoformulation.Nanocrystals of the poorly soluble APIs as well as polymeric nanoparticles containing the APIs have been extensively studied in the recent years [5,6].Also, the increase of the solubility has been explored by changing the intermolecular interactions of the APIs.Amorphous phases, cocrystals, and polymorphs are some of the representative examples [7][8][9].
We have previously reported the formation of an amorphous API through the confinement within nanoporous anodic aluminum oxide (AAO) using ibuprofen (IBU) as a model API [10].In the case of IBU, the kinetically stabilized amorphous phase rapidly changed into a crystalline form during the release experiment, as was similarly observed in the amorphous IBU stabilized within polymer matrices (IBU/polymer = ca.20 : 80) [11,12].These results suggested that the stability of the amorphous phase in contact with water would be the key component to successfully utilize the high solubility of the amorphous phase [13].In addition, the crystallization of IBU is implicated in the possibly damaging effects on the stomach membrane [12].In the present study, some polymeric additives (IBU/polymer = 90 : 10) were tested in their ability to stabilize the amorphous phase in bulk, as well as in nanopores (diameter ∼24 nm) of anodic aluminum oxide.

Fabrication of AAO with Nanopores.
AAOs with nanopores were prepared from the Al foil (99.999%,Goodfellow; thickness 0.5 mm) through the two-step anodization technique [10,14,15].Before anodization, Al was electropolished First anodization was performed in 0.3 M sulfuric acid (aq) for 12 h (4 ∘ C, 25 V).Then, the anodized alumina layers on both sides of Al were removed in chromic acid (aq) (18 g CrO 3 and 40.4 mL phosphoric acid in 1 L water) for 3 h (65 ∘ C), after which well-ordered dimples were made on Al surfaces.Second anodization on the dimpled Al was conducted in 0.3 M sulfuric acid (aq) for 24 h (4 ∘ C, 25 V) to obtain AAO with nanopores.The one side of the AAO substrate was masked with a teflon tape, and the alumina on the other side was removed in 5 wt% NaOH (aq).This was followed by the removal of the remaining Al in cupric acid (aq) (17 g CuCl 2 in 500 mL HCl (aq) and 500 mL water) to finally obtain AAO substrates with nanopores.The average diameter of the nanopores was calculated by measuring about 240 pores from SEM micrographs (JSM-6701F, JEOL).Sulfuric acid (95-98%) and phosphoric acid (>85%) were purchased from Sigma-Aldrich (ACS grade).Sodium hydroxide, hydrochloric acid (35-37%), and copper(II) chloride were obtained from Daejung Chemicals and Metals (extra pure grade; Siheung, Korea).Chromium trioxide (99.5%) was from Acros.
IBU and each polymer (9 : 1 by weight) were well mixed for 30 min using a ball mill (MM200, Retsch; frequency 5 Hz, two stainless steel balls of diameter 9 mm).The mixed powder (ca. 1 mg) was then placed on the AAO substrate, and it was melted at 130-165 ∘ C using a temperature controlled hot plate (HS180, Misung Scientific Co., Seoul, Korea).The melt filled the nanopores by the capillary force assisted by a slight mechanical force (pushed through a polyimide (DuPont) film).The external surfaces of AAO substrates were carefully cleaned using cotton swabs for any residual material; the absence of which was microscopically confirmed afterwards.
The amount of material filled in the AAO was measured by weight (CP225D, Sartorius).Then, the infused AAO substrates were placed in the cells of differential scanning calorimeter (DSC: 821e, Mettler-Toledo) and precalibrated using indium for enthalpy and temperature.Thermal treatment in DSC was performed by cyclic heating and cooling (2 times) from 25 ∘ C to 130-150 ∘ C (heating and cooling rates at 10 and 20 ∘ C/min, resp.).Also, the same experiment was performed on neat IBU.
After the thermal treatment in DSC, the AAO samples were immediately used for the release test at room temperature (20-24 ∘ C).Each sample was immersed in deionized water (2 mL in a quartz UV cell, pH∼6), and the IBU release was monitored by measuring the UV absorbance at 214 nm using a UV/Vis spectrophotometer (Agilent 8453, Agilent).The wavelength was selected for the maximum UV absorbance for IBU, and the IBU concentration was calculated from the absorbance using a preconstructed calibration curve.The UV cell was covered with a paraffin film to prevent any water evaporation during the measurements.Deionized water (resistivity > 18.2 MΩ-cm) was obtained from a Direct-Q (Millipore).The release experiments were independently repeated three times.

Optical Microscopy and Fourier Transform
Infrared Spectroscopy.The effect of polymers on the crystallization of IBU was studied using optical microscopy (OM: BX51, Olympus).Ball-milled samples of IBU and IBU/polymers (3-5 mg) were placed on glass slides and shielded with cover glass.They were melted at 140-165 ∘ C for 10 min using a hot stage (FP90, Mettler-Toledo).The melt was withdrawn from the hot stage, and its glass cover was removed.It was naturally cooled to form amorphous solid during this process.Then, a drop (30 L) of deionized water was placed at the center, and the cover glass was restored.OM images were taken after 1, 2, and 3 min of water contact in the transmission mode under crosspolarization and with a 530-nm retardation plate (U-TP530, Olympus).
Fourier transform infrared spectroscopy (FTIR: Cary 660, Agilent Technologies) was also used in the transmission mode (400-4000 cm −1 ) to analyze any structural change with addition of polymers.Ball-milled and melt/cooled samples of IBU and IBU/polymers were ground with KBr (>99%, Sigma-Aldrich) to prepare the IR specimen.

Effects of Polymeric Additives on IBU Crystallization.
When IBU or IBU/polymer (9 : 1) mixtures were melted and solidified on glass slides, clear solid films were obtained.These did not show birefringence under cross-polarization, which indicated the formation of amorphous phases.When the amorphous phases were exposed to water, regions of birefringence showed up nearly instantaneously.The nucleation of the crystalline phase seemed not very much suppressed by the addition of polymers in this bulk experiments.In all cases, the crystalline phase could be found within 1 min of water contact (Figure 2).By contrast, the growth of the crystals was significantly affected by the presence of polymers.The growth rate was in the following order: IBU (73.9 ± 4.2 m/min) > IBU/PVPh (59.4 ± 5.4 m/min) ≈ IBU/PAA (57.2 ± 7.5 m/min) > IBU/PVP (41.7 ± 7.6 m/min).The crystal growth was measured from the increased distance of the crystal edge from the center of the disks or assemblies, and the directions were chosen to avoid possible external hindrance such as void and adjacent crystalline phases.PVP was most effective in reducing the growth speed of the crystalline phase.Also, its effect was unique in that the crystalline formation of the spherulitic shape was least complete.
IR spectra were examined to probe possibly strong intermolecular interactions between IBU and polymers.Spectra of IBU with and without polymeric additives were shown in Figure 3(a), and those of neat polymers were displayed in Figure 3(b).Overall, there are only subtle changes from which the intermolecular interactions could be inferred.The peaks related to IBU, the main component (90 wt%), remained the same.The interpretation on the intermolecular interaction was from the peaks related to polymers, the minor component (10 wt%).First, the C=O stretching peak of PAA (1710 cm −1 ) overlapped too much with that of IBU (1721 cm −1 ).However, the lack of significant shoulder formation, which could occur from the downshift of the PAA peak [16], suggested insignificant intermolecular interactions between IBU and PAA.Second, the C=O stretching peak (marked with asterisks) of PVP (1664 cm −1 ) showed significant red shift (1640 cm −1 ), which suggested strong interaction, probably hydrogen bonding with the OH group of IBU.Similar shifts were previously observed for PVP, when intimate interactions were established [17][18][19].Finally, the broad OH stretching peaks of PVPh were at about 3380 (peak) and 3522 (shoulder) cm −1 : the former for the hydrogen-bonded groups and the latter for the free groups [20,21].The overlap with peaks of IBU made it difficult to accurately assess the possible changes of the PVPh peaks, but the general shapes and positions of the PVPh peaks appeared not significantly altered.Overall, the IR experiments showed that PVP interacted more strongly with IBU than PAA and PVPh did at this level of concentration.(Further increase of the polymer concentration was not pursued because of the practicalities of drug formulation.)Note that the strong interaction of PVP is in good agreement with the OM observation of bulk crystallization, where the crystalline phase with PVP seemed less mature than with other polymers, and its growth rate was most significantly reduced (Figures 2(g)-2(i)).

IBU in the Nanoporous AAO Substrates.
The pore diameter of the AAO substrates utilized in the present study was 24 ± 2 nm, of which representative image was shown in Figure 4(a).The two-step anodization technique, as described in the experimental section, successfully formed the nanopores.The conditions to generate this particular pore diameter were chosen based on the results of the previous study, where pore diameter less than or equal to 55 nm was necessary to form and stabilize amorphous IBU during solidification [10].
After IBU/polymer infusion in AAO, the thermal behavior of the infused material was observed during resolidification using DSC.No melting was observed, which indicated successful formation of amorphous phase (Figure 5).This result was as expected because our previous study showed the kinetic stabilization of the amorphous phase when the spatial confinement was less than or equal to 55 nm [10].Note that the IBU not in subject of such confinement showed melting behavior at ca. 74-76 ∘ C [10].Representative image of AAO filled with IBU, after the thermal treatment, was shown in Figure 4(c).Release behavior of IBU from AAO was summarized in Figure 6.In all cases, the cumulative release (Figure 6(a)) increased slowly with time, probably due to the low solubility of IBU (e.g., 1.54 × 10 −3 mol/L at 25 ∘ C and pH 6.1 for crystalline IBU) [10,27].The IBU concentration after 180 min was less than 10 −4 mol/L in all cases; it was much smaller than the equilibrium concentration indicating that the current results were from the early stage of the IBU dissolution.The addition of polymeric additives decreased the cumulative release.Also, it generally increased the variation among the repeated release experiments, suggesting that the variation increase was probably due to the inhomogeneous distribution of polymer molecules.In spite of the variation, all repeats showed the same release trends.The IBU flux (Figure 6(b)) also showed the similar trends.The flux decreased dramatically as a function of time in all cases, and it was diminished with addition of polymers.Note that the UV absorption of the polymers at the wavelength used to monitor IBU release (214 nm) was determined negligible as follows.First, the aqueous solutions of PVP, PAA, and PVPh had maximum absorbance at 195, 195, and 190 nm, respectively.Second, assuming that the release of IBU and polymer was proportional to the composition, the corresponding polymer solutions were prepared, and their UV absorbance was measured at 214 nm.The UV absorption of the polymers was at the most two hundredth of the total absorbance observed for the IBU release.
The reason of the decreased flux with time was already reported in our previous study [10].The initially high flux (0 to 5 min) was because of the large solubility of amorphous IBU, and the decrease with time was due to the crystallization of IBU in contact with water [10].The water contact was known to start the crystallization of the amorphous IBU almost instantaneously [11,28], and this was also confirmed in our bulk crystallization (Figure 2), where the initiation of the crystallization was found in less than 1 min.
The basis of the decreased flux with polymer addition could be found from two different sources.First, the polymer molecules associated with IBU appeared to act as diffusion barriers, as previously observed with other APIs [29].In particular, PVPh with the most hydrophobic characteristics showed the most pronouncedly diminished release behavior.Second, the addition of polymer did not stabilize the amorphous IBU enough to increase its release rate.In the bulk crystallization study (Figure 2), the polymers decreased the growth rate of the crystals.However, they did not significantly hinder the nucleation of the crystalline phase.Since IBU was confined in the pores of 24 nm diameter, nucleation followed by growth in nanoscale might be enough to form a crystalline "cap" that essentially stopped the unobstructed release from any remaining amorphous phase [11].

Conclusions
In summary, the AAO substrates were employed to generate amorphous IBU with polymeric additives by confining them within the nanopores (diameter, 24 nm).The amorphous IBU had tendency to rapidly crystallize upon water contact, which was the origin of the temporal decrease of the IBU release flux [10,11,28].The polymeric additives, which retarded the crystal growth, were less effective in inhibiting the nucleation of the crystalline phase.In addition, the IBU with polymeric additives showed the sustained release behavior, probably because of the polymer molecules acting as the diffusion barriers.Since the temporal decrease of the flux was insignificantly affected by the polymers, the present study suggests that the inhibition of the crystal nucleation is the key aspect of stabilizing the amorphous IBU in the nanoscopic confinement, which is essential to make full use of its high solubility (ca.6 times higher than that of the bulk crystalline phase) [10].
Figure 2 shows the temporal evolution of the amorphous-to-crystalline transformation for IBU ((a)-(c)), IBU/PAA ((d)-(f)), IBU/PVP ((g)-(i)), and IBU/PVPh ((j)-(l)).The initial form of the crystalline phase was oval shaped, as indicated by white arrowheads at the various places of Figure2.Then, the crystalline phase became disk shaped probably through the assembled formation of the ovals sharing the center of the disk as their common corner.Also, the peripheries of the disk-shaped crystalline phase grew outward with time.

Figure 3 :Figure 4 :
Figure 3: IR spectra of (a) IBU with and without polymeric additives and (b) neat polymers.