Amorphous Ibuprofen Confined in Nanostructured Silica Materials: A

ARTICLE pubs.acs.org/JPCC

Amorphous Ibuprofen Confined in Nanostructured Silica Materials: A Dynamical Approach Ana R. Bras,† Esther G. Merino,† Paulo D. Neves,† Isabel M. Fonseca,† Madalena Dionísio,† Andreas Sch€onhals,‡ and Natalia T. Correia*,† †

REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal ‡ BAM Federal Institute for Materials Research and Testing, D-12205 Berlin, Germany ABSTRACT: The molecular mobility of condensed matter confined to nanometer dimensions can be dramatically changed from those of the bulk state in such a way that, when the guest is a drug, it can be advantageously used in pharmaceutical applications. We show by dielectric relaxation spectroscopy that the molecular mobility of the important ibuprofen drug embedded in nanoporous SBA-15 is significantly influenced by the confinement. An evidence of the existence of two families of molecules with different molecular mobilities is provided and investigated in their temperature dependence. One family is due to molecules close to the pores' center with a higher mobility compared with the bulk at low temperatures, and another family with slower dynamics originated from molecules interacting with the pore walls. The work reports the simultaneous manifestation of true confinement and surface effects in this nanostructured silica host for a drug. For future applications in drug delivery systems, the dynamics determined by the guest-host interplay and the one of the bulklike molecules can be tuned to achieve a desired release profile.

’ INTRODUCTION Pharmaceutical compounds in their amorphous form can play a crucial role concerning the therapeutic activity.1-4 Recently, confinement of drugs in nanoporous host systems emerged as a strategy to stabilize the otherwise unstable glassy/supercooled state and to manipulate the crystalline state of pharmaceuticals.5,6 Moreover, it is well known that the glass transition and the glassy dynamics (R-relaxation, dynamic glass transition) of guest molecules is affected by confining it in nanoscaled geometries, which tend to lower the glass transition temperature Tg.7-10 This is called the confinement effect.8 Even the molecular motions responsible for the R-relaxation can change dramatically for pore sizes lower than a critical value, which allows estimating a minimal length scale for cooperativity.8 In addition to this scenario, molecular dynamics in a confining space is also determined by surface effects resulting from interactions of the guest molecules with the walls of the porous host. These interactions take place at the interface of both and slow down the molecular dynamics.8,11,12 Surface effects become important for small pores and strong surface interaction as has been shown for systems forming hydrogen bonds between the guest molecules and the inner pore wall.13-16 As a result, the molecular dynamics of molecules confined to nanoporous hosts is controlled by a counterbalance between confinement and surface effects. Among the different nanoporous matrices, mesoporous silicabased materials attract a great interest due to their particular physical properties, such as their biocompatibility, ordered pore network, high internal surface area, silanol-containing surface, and chemical and mechanical stability.17,18 r 2011 American Chemical Society

Small organic molecules have been confined to such matrices19-23 and pharmaceutics as well24-29 from which particular relevance was given to the analgesic, antipyretic, and anti-inflammatory poorly water-soluble model drug, ibuprofen, as a potential drug delivery system.30-33 The crystallization of bulk ibuprofen (racemic form) can be easily circumvented on cooling from the melt,34-36 which allowed some of us to investigate the molecular mobility in the supercooled and the glassy state35 using dielectric relaxation spectroscopy (DRS),37 over a wide frequency and temperature range. Multiple relaxation processes were identified: the main Rrelaxation associated with the dynamic glass transition, a βJG Johari-Goldstein process taken as the precursor of the Rrelaxation, and a more localized secondary γ-relaxation. Moreover, in the supercooled liquid state (T > Tg), an additional relaxation mode with a Debye-like shape (D-process) was found assigned to the dynamics of hydrogen-bonded aggregates. Regarding the molecular mobility in confinement, NMR showed that ibuprofen, when confined to MCM-4138-41 and SBA-15,27 is extremely mobile and it is not in a crystalline or glassy state. Nevertheless, most of these studies were carried out at room temperature. Therefore, it is necessary to investigate the molecular mobility in a much wider temperature and frequency range. DRS is a powerful tool to get a detailed knowledge regarding the molecular mobility of guests, among which only few studies refer to organic low-molecular-weight materials confined Received: August 12, 2010 Revised: December 27, 2010 Published: February 25, 2011 4616

dx.doi.org/10.1021/jp107631m | J. Phys. Chem. C 2011, 115, 4616–4623

The Journal of Physical Chemistry C

ARTICLE

to silica-based mesoporous hosts (MCM42 and SBA42-44); some studies refer also to water45,46 confined to the same kind of mesopores. To get a deeper insight in the behavior of a pharmaceutical model system, ibuprofen (Ibu) is confined to the pores of SBA15 with a pore diameter of 8.6 nm (100% silica). DRS is applied (from 10-1 to 106 Hz) over a wide temperature range (more than 150 K) to characterize in detail its molecular mobility in the glassy and supercooled liquid states compared to the corresponding states of the bulk.

’ EXPERIMENTAL DETAILS Materials. Ibuprofen. Ibuprofen ((2RS)-2[4-(2-methylpropyl)phenyl]propanoic acid, C13H18O2) was purchased from Sigma (catalogue no. I4883 (CAS 15687-27-1), lot number 026H1368, 99.8% GC assay) with a molar mass of 206.28 g 3 mol-1. It is a racemic mixture of S(þ)-ibuprofen and R(-)ibuprofen, and it was used without further purification. The crystallographic structure of racemic ibuprofen has been solved from X-ray diffraction by McConnell47 (ref code in Cambridge Structural Database, CSD, CSD-IBPRAC) and from single-crystal pulsed-neutron diffraction by Shankland et al.48 (ref code CSDIBPRAC02): it was established that crystalline ibuprofen exists only as cyclic dimers formed by one molecule in the R configuration and other in the S configuration through a strong double hydrogen bond involving the COOH acid groups, organized in the monoclinic P21/c space group. An X-ray powder diffraction characterization of the crystalline sample studied in the present work was recently performed by Derrolez et al.,49 confirming the crystallographic structure reported by those authors47,48 (in this reference, the crystalline form analyzed in this work is called “phase I”). For the sake of simplicity, the studied (RS)-ibuprofen is referred to in this work as ibuprofen. At room temperature, ibuprofen is a white crystalline powder (phase I) that is characterized by an onset melting temperature of 347 K and melting enthalpy of 25.5 kJ 3 mol-1, obtained by DSC at a heating rate of 5 K min-1.35 The onset of the glass transition occurs at Tg = 228 K, with a heat capacity step of ΔCp = 76 J 3 K-1 3 mol-1. SBA-15. The mesoporous SBA-15 (100% Si) matrix was hydrothermally synthesized and characterized according to the procedures described in the literature.50,51 Briefly, the block copolymer poly(ethylene oxide)-poly(propylene oxide)-poly(ethylene oxide) was used as a template in acidic conditions, and tetraethyl orthosilicate was used as a silica source in that case. For calcination, the templates were heated with a rate of 1 K/min to 773 K, kept there under dry nitrogen for 1 h, and after that, kept under dry air for 6 h. Nitrogen absorption analysis was used to obtain the textural features. The specific surface area was determined from the linear portion of the BET plots, giving a value of 612.8 m2/g; the DFT total pore volume is 1.047 cm3/g. The pore size distribution determined by the DFT method has an average diameter of 8.6 nm (see Figure 1); by the BJH (desorption) method, the average pore width is 7.65 nm. SBA-15 materials were routinely characterized by XRD and electron microscopy (TEM), which confirmed the uniform structure of two-dimensional hexagonally ordered cylindrical pores. Ibuprofen Loading. Ibuprofen was confined according to a procedure adapted from refs 30 and 52. After a cleaning step to remove water and impurities (7 h at 10-4 mbar and 573 K),42 self-supported pellets of finely ground SBA-15 were immersed

Figure 1. Pore size distribution of SBA-15 as determined by the DFT method; by the BJH (desorption) method, the average pore width is 7.65 nm.

Figure 2. Thermogravimetric curves obtained on heating at 10 K 3 min-1 for the host mesopores SBA-15 (SBA) and confined ibuprofen (Ibu/ SBA). The behavior of bulk ibuprofen (Ibu) is given for comparison.

under vacuum with a solution of ibuprofen in ethanol (∼0.05 g 3 cm-3). The pores were filled by capillary wetting, and the solvent was removed by heating the sample to 333 K at normal pressure for ca. 12 h. Five successive impregnation steps were carried out. The samples were carefully washed with a small amount of ethanol to remove the excess ibuprofen that crystallized at the outer surface of the pellet. The dielectric measurements were performed immediately after preparation, and samples were handled in a desiccator. Experimental Techniques. Thermogravimetric Analysis. Thermogravimetric analysis (TGA) was used to estimate the content of ibuprofen inside the mesoporous silica, as described previously for other guests confined in the same kind of hosts.42 TGA measurements were carried out by a Seiko TG/DTA 220 apparatus, at a heating rate of 10 K 3 min-1 under a dry synthetic air atmosphere using a part of the sample prepared for the dielectric measurements. Figure 2 shows the thermogravimetric curves obtained for the empty host SBA-15 matrices (SBA), ibuprofen/ SBA (Ibu/SBA), and for bulk ibuprofen (Ibu). For SBA, a first small step is observed until 423 K, which is related to water mass loss (ca. 1 wt %); no additional weight loss occurs, being thermally stable up to 1073 K. The decomposition 4617

dx.doi.org/10.1021/jp107631m |J. Phys. Chem. C 2011, 115, 4616–4623


ARTICLE

of bulk ibuprofen starts at 473 K, as revealed by an abrupt mass loss in the thermogram. In Ibu/SBA, the first small step until 423 K gives the amount of absorbed water (3%) and it was not considered for the determination of loaded ibuprofen within the host matrix. The content of ibuprofen loaded within the system, mIbu, was determined from these measurements, and the filling degree, Θ, was defined as the ratio42 Θ¼ ¼

mass measured maximal mass for complete pore filling mIbu mSBA VP FCon Ibu

ð1Þ

where FCon Ibu is the density of Ibu in confinement and mSBA the mass of the empty molecular SBA guest. Assuming that FCon Ibu ≈ FBulk Ibu , the filling degree Θ can be calculated. For this calculation, the density of liquid ibuprofen at 298 K (1.00 g 3 cm-3) was obtained from NPT Molecular Dynamic Simulations53,54 because no experimental value was found. A value of 27 wt % of confined ibuprofen was obtained. ATR-FTIR Analysis. The attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectra of the samples were recorded by a Thermo-Nicolet (Nexus 670) FTIR spectrometer at room temperature. The measurements were carried out in the wavenumber range from 400 to 4000 cm-1 accumulating 64 scans having a resolution of 4 cm-1. Dielectric Relaxation Spectroscopy. The equipment used to measure the complex dielectric function ε*( f ) = ε0 ( f) - iε00 ( f ) ( f, frequency; ε0 , real part; ε00 , loss part) from 10-1 to 106 Hz, and the methodology to analyze dielectric data is described in detail elsewhere.35,42 Samples were prepared in parallel plate geometry between two gold-plated electrodes with a diameter of 10 mm. First, the samples were cooled to 153 K with a rate of 10 K min-1 and dielectric spectra were collected isothermally from 153 to 373 K in increasing steps of 2 K (called “run 1” in the next section). The temperature was then decreased in steps of 2 K in the range of 373-159 K (“run 2”). All measurements were carried out isothermally with a temperature stability of 0.1 K.

’ RESULTS AND DISCUSSION ATR-FTIR spectroscopy was used to get further evidence of ibuprofen loading in the host mesopores. Figure 3a presents the spectra of loaded SBA (open circles), crystalline ibuprofen (solid line), and empty SBA (dashed line). The detection of a band centered around 1710 cm-1 in the SBA/Ibu spectrum, as observed in bulk Ibu due to the CdO stretching (H-bonded) vibration,31 as well as the C-C (aromatic) stretching55 (∼1512 cm-1) and C-H3 antysimmetric deformation55 (∼1465 cm-1), confirms the successful guest loading in the silica matrix; the latter does not absorb in this frequency region. Moreover, the CdO band emerges broadened compared with the one detected for crystalline Ibu. A more detailed analysis on its frequency location is out of the scope of this paper; however, it is related to the distribution population of ibuprofen n-mers (dimers, trimers, etc.), as discussed in refs 35 and 56. The widening of the CdO band together with a shift of the C-C and C-H3 bands upward was also previously found for bulk amorphous versus crystalline Ibu by FTIR analysis,57 replotted in Figure 3b. Therefore, this can be taken as an indication that ibuprofen is in the amorphous state inside pores.

Figure 3. (a) ATR-FTIR spectra of Ibu/SBA (open circles) compared with crystalline ibuprofen, Ibucrys (solid line), and empty SBA. The spectrum of Ibu/SBA after being submitted to the dielectric measurements is also shown (gray circles). (b) FTIR spectra of crystalline (Ibucrys, solid line) and amorphous (Ibuam, dashed line) bulk Ibuprofen.57

To evaluate the physical state of confined ibuprofen, the temperature dependence of the real part of the complex dielectric permittivity, ε0 , at a fixed frequency is considered; in fact, ε0 is a valuable property to probe phase transitions, as recently shown for bulk ibuprofen35 and for other pharmaceuticals.34,58,59 Figure 4 presents ε0 (T) for ibuprofen/SBA (Ibu/SBA) taken at 1 kHz from the isothermal measurements carried on heating to 373 K (run 1, gray circles); the measurements were preceded by a cooling ramp from room temperature to 159 K (data not shown). ε0 1 kHz(T) collected isothermally in a second cooling run is also displayed. Contrary to bulk crystalline ibuprofen (full circles), no steplike frequency independent increase of ε0 35 occurs for confined ibuprofen, in the whole studied temperature range. This confirms, first, that the residual crystallized ibuprofen outside the pores was efficiently removed during the washing step. Otherwise, a bulklike melting transition would be observed, as found by DSC.32,60 Second, if confined ibuprofen would undergo crystallization, a melting transition should be observed at a lower temperature than the bulk one, Tm, as found for ibuprofen inside mesoporous silicon with larger pores61 and also for several other materials.5,20,62 These results confirm that all loaded ibuprofen is confined inside the pores of SBA-15 and, more importantly, exists in its amorphous state, confirming the ATR-FTIR analysis. The appearance of a broad peak in ε0 for temperatures higher than 250 K in the first heating run is due to water reorientation 4618



Figure 4. Temperature dependence of the real part of the complex dielectric function, ε0 (isochronal plot at 1000 Hz, in arbitrary units, a.u.), for confined ibuprofen, Ibu/SBA: data obtained on heating, run 1 (gray circles), and on the subsequent cooling, run 2 (open circles). Data were taken from isothermal measurements carried out between 153 and 353 K. For comparison, the melting of the as-received bulk crystalline ibuprofen (Ibu) at Tm (black circles)35 is shown. The lines are guides for the eyes.

and evaporation, which is absent in the subsequent second run. Similar dielectric responses were observed for water adsorbed in porous glasses63 and MCM-41.64 This frequency-independent behavior is caused by the evaporation of water molecules (not bulk water) adsorbed at the outer surface of the silica particles or at the extra pores when the sample is taken out from the vacuum chamber, after the first impregnation step. Only a small amount of water (ca. 3 wt %) was calculated from the TGA measurements; nevertheless, it highly influences the spectra due to its high dipole moment. Because of the effect of thermal treatment (heating up to 373 K) and the influence of the dry nitrogen flux used for the temperature control, the water molecules easily desorbed. As a result, in the subsequent cooling (run 2), the ε0 (T) trace presents the normal steplike behavior typical for a relaxation process. Therefore, data obtained from run 2 for dried Ibu/SBA will be taken for further discussion and compared with bulk amorphous ibuprofen measured during cooling.35 To evaluate the stabilization of the confined guest upon thermal treatment, the sample submitted to the dielectric spectroscopy measurements was further analyzed by ATR-FTIR. Figure 3a shows the similarity between the spectra collected before and after (gray symbols) both temperature and frequency scans, evidencing that confined ibuprofen did not undergo crystallization as observed for the amorphous bulk material under the same thermal treatment.34,65 To investigate in detail the different relaxation processes, the imaginary part of the permittivity (dielectric loss, ε00 ) is analyzed. Figure 5 gives the dielectric loss at 1 Hz for Ibu/SBA in comparison with bulk ibuprofen. It must be noted that the dielectric loss of the dry SBA matrix is negligibly small compared to that observed for the (liquid) filled sample (squares in Figure 5), as also found for related hosts;66 therefore, the detected dielectric response is due to the confined molecules. The dielectric loss shows several dielectric relaxation processes, as already discussed for bulk ibuprofen, but one realizes that the mobility of ibuprofen embedded in the pores of the SBA mesoporous matrix is significantly influenced by the confinement. At the lowest temperatures, the βJG Johari-Goldstein

ARTICLE

Figure 5. Dielectric loss of ibuprofen versus temperature at a frequency of 1 Hz confined to the pores of SBA-15 (black circles), Ibu/SBA, during cooling in comparison to the bulk (Ibu, open circles, right side) and to the dry empty SBA (squares). The heavy solid line is the overall fitting function to data by a sum of three Gaussians (dashed lines). The thin solid lines are guides for the eyes.

process (βIbu/SBA) is detected, which is discussed as the precursor of the dynamic glass transition that is observed in its hightemperature side (RIbu/SBA). At this low frequency (1 Hz), both processes are significantly shifted to lower temperatures compared with bulk ibuprofen (ca. 20 K for the RIbu/SBA-relaxation). In addition to these bulklike relaxations, two further processes, SIbu/SBA and DS-Ibu/SBA, are observed. The multimodal nature of the spectra measured for confined ibuprofen with the processes mentioned above is detected also in the dielectric loss isotherms (frequency domain, Figure 6a). The model function of HavriliakNegami (HN-function) is used to analyze the data and to separate the different relaxation processes (examples of the fitting procedure are given in Figure 6b,c, together with the individual HN-functions used to separate the spectra). The data analysis was extended to the isochronal representation of the dielectric loss, that is, ε00 versus temperature at constant frequencies (ε00 (T; f = const)). A superposition of k Gaussians35,67 (dashed lines in Figure 5) was fitted to ε00 (T) to obtain the maximum temperature of peaks, Tmax,k, for each measured frequency. Figure 7a gives the temperature dependence of the relaxation time extracted from the HN-fitting, together with the isochronal analysis for each relaxation process. It is worth noting that, first, the two sets of relaxation time values obtained by both kinds of analysis coincide. This coincidence validates the use of isochronal data in cases when the loss peaks maxima are ill defined in the frequency domain, out of the accessible frequency range, or to distinguish multiple processes.67,68 Second, the assignment of the relaxation processes given above is supported. The process associated with the glassy dynamics of confined ibuprofen (RIbu/SBA) superimposes with that of the bulk (RIbu) at the highest temperatures. For bulk ibuprofen, the data follow a VFTH law, as typical for the cooperative glassy dynamics (Figure 7a). Moreover, the respective dielectric strength decreases with increasing temperature, also in agreement with an interpretation as a R-relaxation (Figure 7b). With decreasing temperature, the relaxation times for confined ibuprofen deviate more and more from that of the bulk in such a way that they are shorter at the same temperature. Moreover, the temperature dependence seems to change from a VFTH to an 4619



ARTICLE

Figure 7. (a) Relaxation time, τmax, versus 1/T for the processes detected above Tg: open symbols, bulk isothermal loss data collected during cooling; filled symbols, obtained from isothermal loss data collected during cooling for the confined sample; gray symbols, τ obtained from the isochronal plots for all studied frequencies. Lines are fits of the Arrhenius and VFTH equation to the corresponding data. Vertical dashed lines indicate the dielectric glass transition temperature Tdiel g (τ = 100 s). (b) Dielectric strength, 4ε, versus 1/T for all processes shown in (a). Lines are guides for the eyes.

Figure 6. (a) Dielectric loss versus frequency at T = 219, 235, 253, and 287 K for ibuprofen confined to SBA-15. Panels (b) and (c) are two examples of the fitting procedure used to analyze the spectra at T = 235 K and T = 287 K, respectively, where an additive superposition of three HN-functions is assumed (dashed lines). In all graphs, the solid lines are the overall HN fitting curves to data.

Arrhenius law with the following parameters: pre-exponential factor, τ¥ = 4.5 10-30 s, and activation energy, Ea = 122 kJ 3 mol-1. The change from VFTH toward Arrhenius dependence was found for other glass formers under confinement8,69 and is conventionally rationalized as an effect of the interference of the imposed geometry, that is, the pore dimension, with the characteristic length for the dynamic glass transition;70 the latter is defined as the dimension of the cooperative rearranging regions in light of the Adam-Gibbs model.71 Therefore, the transition to an Arrhenius behavior, and ultimately the suppression of the glass transition itself, allows estimating a minimal length scale for cooperativity.

In this context, the behavior here reported for ibuprofen inside nanopores reflects a strong confinement effect that could point to an inherent length scale relevant for the glassy dynamics. Taking the temperature where the relaxation time is τ = 100 s as Tg and assuming a linear temperature dependence, a value of 204 K is estimated. Compared with bulk ibuprofen (Tg = 226 K),35 a Tg depression of more than 20 K is observed for confined ibuprofen (see also Figure 5). Moreover, this depression of the glass transition temperature in confinement is confirmed by a similar shift of the glass transition temperature deduced from the temperature dependence of ε0 measured at the lowest frequency (0.1 Hz) in Figure 8. Whereas the onset of the step in the ε0 (T) trace taken as a signature of the dynamic glass transition occurs around 225 K for bulk ibuprofen, for the confined material, it is observed slightly above 200 K. By this way, the location of the glass transition is estimated using the same onset criteria as in DSC. This result is deduced independently of any data treatment. The sharpness of the transition region in bulk ibuprofen is due to the fact that mainly the R-process is probed by ε0 (T), whereas for the confined system, both the RIbu/SBA-process and the S-process contribute; the latter is more intense. Both contributions are indicated in Figure 8. This turned out to be a way to obtain the 4620



Figure 8. Isochronal plots of (a) ε0 evidencing the shift of the glass transition to lower temperatures of Ibu/SBA-15 relative to bulk; the “glass transition” of the ibuprofen molecules adsorbed in the pore walls, SIbu/SBA-process, exhibits a significant shift to higher temperatures relative to the glass transition of molecules confined in the middle of the pore. Panel (b) presents the correspondent ε00 representation that helped in the assignment of the onset of the SIbu/SBA-process and that also makes clear the difference between both R-bulklike processes; the DS-Ibu/SBA-process is observed at the high-temperature flank of the latter. The heavy solid line is the overall fit by a sum of three Gaussians to the data (red lines). The thin solid lines are guides for the eyes.

glass transition of the Ibu confined to SBA. Calorimetric studies have been reported by Mellaerts et al.32 for ibuprofen confined to the same type of SBA-15 nanopores with a diameter of 8.4 nm and with an equivalent loading degree (20-30 wt %). The authors pointed out that it is impossible to characterize the glass transition for these guest-hosts systems unambiguously by DSC due to the influence of phase transitions of water adsorbed inside the pores. It is interesting to note that the molecular mobility of the nematic liquid crystalline mixture E7 with an average length of the molecule of ca. 1.74 nm,72 which is longer than that of ibuprofen (1.13 nm72), is less influenced by confinement in the same SBA-15 host.42 The enhanced confinement effect on the glassy dynamics of ibuprofen probably originates from its strong tendency to form hydrogen-bonded aggregates35 involving a larger length scale of the cooperative dynamics. In fact, no evidence of free CdO vibration was found in the ATR-FTIR spectrum of Ibu/SBA, which can be taken as an indication that confined ibuprofen, like the amorphous bulk one, has a strong tendency for aggregation as dimers and trimers. The same argument might be true for salol.8 At lower relaxation times (higher temperatures) than the RIbu/SBA-process, the dominant S-process is observed for the confined system. A similar process was observed for confined guests

ARTICLE

interacting with the pore surface of the host13,42,73 having a slower molecular dynamics. For ibuprofen confined to the same mesoporous matrix, recent multinuclear solid-state NMR studies27 suggested the existence of two families of molecules, with different mobilities: one that does not interact with the surface and another one interacting with the Si-OH present at the pore wall, although only through weak hydrogen bonds. Moreover, the dielectric strength, Δε, of the S-process increases in general with increasing temperature (see Figure 7b), which can be due to a decrease of the interaction strength of ibuprofen molecules with the pore wall, leading to the reorientation of greater parts of the molecular dipole vector (increased fluctuation angle) or an increased number of fluctuating dipoles. A similar behavior was observed for the surface process detected in E7 confined in the same host.42 At the highest temperatures, Δε nearly reaches a plateau, as observed for the liquid crystal 8CB confined to an MCM material (Al-Si) due to a counterbalance between the described effect and the predicted decrease with the temperature increase due to thermal energy (Fr€ohlich-Kirkwood equation).37 Therefore, the S-process is assigned to the dynamic glass transition of the ibuprofen molecules (e.g., monomers, linear dimers) likely localized close to the pore walls and interacting with the surface silanol groups via weak hydrogen bonds. The temperature dependence of relaxation times of the S-process follows a VFTH law,74-76 τ(T) = τ¥ exp[(B)/(T T0)] with the following parameters: τ¥ = 10-14 s, B = 2530 K and T0 = 163 K. In Figure 7a, the solid line is the respective VFTH fit from which parameters a Tg value (τ = 100 s) of 231 K was estimated. The existence of two families of ibuprofen molecules with different mobilities, suggested previously by NMR at room temperature, is thus, first, confirmed by DRS and, second, characterized over a wide frequency and temperature range, strengthening the potential of DRS to evaluate mobility even in complex systems as confinement in nanostructured channels. At the low-frequency side of the S-process (Figure 6c) in the isothermal loss curves and at the high-temperature flank in the isochronal representation (Figures 5 and 8b), a further relaxation process with a weak intensity is observed, here designated as the DS-Ibu/SBA-process. As already discussed for bulk ibuprofen,35 a weak Debye-shaped relaxation appears in the low-frequency site of the R-relaxation, having some properties similar to the Debye process found in a large class of associating liquids, such as monohydroxy alcohols.77 The molecular assignment of the Debye-type relaxation and how it contributes to the structural relaxation or its viscosity continue to be a matter of debate.78,79 Recently using molecular dynamics simulations, Affouard and Correia56 demonstrated from calculations of the single molecule and total dipole autocorrelation functions that, in ibuprofen, the single molecule and total relaxation times are comparable. This result shows that the behavior of the long time total dipole correlation should be dominated by the self-autocorrelation function, being dominated by the internal cis-trans conversion of the OdC-O-H group coupled to the change of the intermolecular linear/cyclic hydrogen-bonded structures. It was demonstrated that the effective rotational potential energy barriers of the OdC-O-H groups due to the surroundings are averaged and the dipolar relaxation follows a simple exponential decay responsible by the Debye-like nature of the observed relaxation peak in DRS. For the bulk system, the temperature dependence of relaxation times of the Debye process follows nearly that of the 4621


The Journal of Physical Chemistry C R-relaxation; a correlation between the dynamics of both processes was proposed.35 The D-process associated with the bulklike RIbu/SBA-relaxation in the confined system was not detected due to the overlapping with the strong S-process. Nevertheless, a Debye-shaped relaxation is observed with a temperature dependence of the relaxation times nearly parallel to that of the S-process (see Figure 7a). Moreover, the almost temperatureindependent behavior of the dielectric strength, 4ε, for DS,Ibu/SBA is similar to that observed for the bulk D-process (see Figure 7b). The magnitude of the dielectric strength was found to be determined by the population of ibuprofen molecules in the trans conformation, having a higher dipole moment, and to the variation of the Kirkwood correlation factor.56 Both effects decrease with decreasing temperature and counterbalance the increase predicted by the Fr€ohlich-Kirkwood equation.37 Therefore, the temperature dependence of both relaxation times and dielectric strength are consistent with a D-process that, in this particular case, is associated with the surface relaxation, that is, a DIbu/SBA-process correlated to the cis-trans conversion of the OdC-O-H group of the slower family of ibuprofen molecules adsorbed at the pore wall.

’ CONCLUSIONS In summary, dielectric relaxation spectroscopy is applied to the study of the molecular mobility of an important pharmaceutical drug confined in a nanostructured silica material for the first time. The reported results, by directly probing the mobility of the drug inside the porous host, are of crucial importance for future applications aiming to develop drug delivery systems. The existence of two families of molecules with different molecular mobilities was shown unambiguously and characterized over a large range of frequencies and temperatures, for which the respective glass transition temperatures are provided. The mobility of the molecules located more in the center of the pores, on the one hand, and the slower dynamics of the molecules adsorbed at the pore walls that unravels the guest-host interplay, on the other hand, can be advantageously used to tune the release of the drug to achieve a desired kinetic profile. Besides its practical interest, this work provides an example of a low molecular guest hydrogen-bonded liquid undergoing deviations from its bulk dynamical behavior due to finite size effects when confined to a nanostructured silica matrix. ’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT N.T.C. and A.R.B. thank the BAM Federal Institute for Materials Research and Testing for the use of its research facilities. Financial support from the Fundac-~ao para a Ciência e Tecnologia (FCT, Portugal) through the projects PTDC/ CTM/64288/2006 and PTDC/CTM/098979/2008 is acknowledged. A.R.B. acknowledges the FCT for a Ph.D. grant SFRH/BD/23829/2005. ’ REFERENCES (1) Hancock, B. C.; Zografi, G. J. Pharm. Sci. 1997, 86, 1–12. (2) Hancock, B. C.; Parks, M. Pharm. Res. 2000, 17, 397–404.

ARTICLE

(3) Roberts, C. J.; DeBenedetti, P. G. AIChE J. 2002, 48, 1140– 1144. (4) Zhou, D.; Zhang, G. G. Z.; Law, D.; Grant, D. J. W.; Schmitt, E. A. J. Pharm. Sci. 2002, 91, 1863–1872. (5) Beiner, M.; Rengarajan, G. T.; Pankaj, S.; Enke, D.; Steinhart, M. Nano Lett. 2007, 7, 1381–1385. (6) Rengarajan, G. T.; Enke, D.; Steinhart, M.; Beiner, M. J. Mater. Chem. 2008, 18, 2537–2539. (7) Kranbuehl, D.; Knowles, R.; Hossain, A.; Hurt, M. J. Phys.: Condens. Matter 2003, 15, S1019–S1029. (8) Kremer, F.; Huwe, A.; Sch€onhals, A.; Roz_ a nski, S. A. Molecular Dynamics in Confining Space. In Broadband Dielectric Spectroscopy; Sch€onhals, A., Kremer, F., Eds.; Springer-Verlag: Berlin, 2003. (9) Leys, J.; Sinha, G.; Glorieux, C.; Thoen, J. Phys. Rev. E 2005, 71, 051709. (10) Koppensteiner, J.; Schranz, W.; Carpenter, M. A. Phys. Rev. B 2010, 81, 024202. (11) Sch€uller, J.; Richert, R.; Fischer, E. W. Phys. Rev. B 1995, 52, 15232–15238. (12) Sch€onhals, A.; Goering, H.; Schick, Ch. J. Non-Cryst. Solids 2002, 305, 140–149. (13) Arndt, M.; Stannarius, R.; Groothues, H.; Hempel, E.; Kremer, F. Phys. Rev. Lett. 1997, 79, 2077–2080. (14) Dadmun, M. D.; Muthukumar, M. J. Chem. Phys. 1998, 98, 4850–4852. (15) Arndt, M.; Stannarius, R.; Gorbatschow, W.; Kremer, F. Phys. Rev. E 1996, 54, 5377–5390. (16) See the Supporting Information of ref 6. (17) Ruiz-Hitzky, E.; Ariga, K.; Lvov, Y. M., Eds. Bio-inorganic Hybrid Nanomaterials: Strategies, Syntheses, Characterization and Applications; Wiley-VCH: Weinheim, Germany, 2008. (18) Salonen, J.; Kaukonen, A. M.; Hirvonen, J.; Lehto, V. P. J. Pharm. Sci. 2008, 97, 632–653. (19) Morineau, D.; Alba-Simionesco, C. J. Phys. Chem. Lett. 2010, 1, 1155–1159. (20) Xia, Y.; Dosseh, G.; Morineau, D.; Alba-Simionesco, C. J. Phys. Chem. B 2006, 110, 19735–19744. (21) Morineau, D.; Xia, Y.; Alba-Simionesco, C. J. Chem. Phys. 2002, 117, 8966–8972. (22) Morishige, K.; Kawano, K. J. Chem. Phys. 2000, 112, 11023– 11029. (23) Azais, T.; Hartmeyer, G.; Quignard, S.; Laurent, G.; Babonneau, F. J. Phys. Chem. C 2010, 114, 8884–8891. (24) Qu, F. Y.; Zhu, G. S.; Huang, S. Y.; Li, S. G.; Sun, J. Y.; Zhang, D. L.; Qiu, S. L. Microporous Mesoporous Mater. 2006, 92, 1–9. (25) Nunes, C. D.; Vaz, P. D.; Fernandes, A. C.; Ferreira, P.; Rom~ao, C. C.; Calhorda, M. J. Eur. J. Pharm. Biopharm. 2007, 66, 357–365. (26) Vallet-Regí, M.; Balas, F.; Arcos, D. Angew. Chem., Int. Ed. 2007, 46, 7548–7558. (27) Izquierdo-Barba, I.; Sousa, E.; Doadrio, J. C.; Doadrio, A. L.; Pariente, J. P.; Martínez, A.; Babonneau, F.; Vallet-Regí, M. J. Sol-Gel Sci. Technol. 2009, 50, 421–429. (28) Van Speybroeck, M.; Barillaro, V.; Do Thi, T.; Mellaerts, R.; Martens, J.; Van Humbeeck, J.; Vermant, J.; Annaert, P.; Van Den Mooter, G.; Augustijns, P. J. Pharm. Sci. 2009, 98, 2648–2658. (29) Lebold, T.; Jung, C.; Michaelis, J.; Br€auchle, C. Nano Lett. 2009, 9, 2877–2883. (30) Vallet-Regí, M.; Ramila, A.; del Real, R. P.; Perez-Pariente, J. Chem. Mater. 2001, 13, 308–311. (31) Mu~ noz, B.; Ramila, A.; Perez-Pariente, J.; Díaz, I.; Vallet-Regí, M. Chem. Mater. 2003, 15, 500–503. (32) Mellaerts, R.; Jammaer, J. A. G.; Van Speybroeck, M.; Chen, H.; Van Humbeeck, J.; Augustijns, P.; Van den Mooter, G.; Martens, J. A. Langmuir 2008, 24, 8651–8659. (33) Zheng, J. P.; Luan, L.; Wang, H. Y.; Xi, L. F.; Yao, K. D. Appl. Clay Sci. 2007, 36, 297–301. (34) Johari, G. P.; Kim, S.; Shanker, R. M. J. Pharm. Sci. 2007, 96, 1159–1175. 4622


The Journal of Physical Chemistry C (35) Bras, A. R.; Noronha, J. P.; Antunes, A. M. M.; Cardoso, M. M.; Sch€onhals, A.; Affouard, F.; Dionisio, M.; Correia, N. T. J. Phys. Chem. B 2008, 112, 11087–11099. (36) Dudognon, E.; Danede, F.; Descamps, M.; Correia, N. T. Pharm. Res. 2008, 25, 2853–2858. (37) Kremer, F.; Sch€onhals, A., Eds. Broadband Dielectric Spectroscopy; Springer Verlag: Berlin, 2002. (38) Babonneau, F.; Yeung, L.; Steunou, N.; Gervais, C.; Ramila, A.; Vallet-Regí, M. J. Sol-Gel Sci. Technol. 2004, 31, 219–223. (39) Babonneau, F.; Camus, L.; Steunou, N.; Ramila, A.; Vallet-Regí, M. Mater. Res. Soc. Symp. Proc. 2003, 775, 77–82. (40) Azaïs, T.; Hartmeyer, G.; Quignard, S.; Laurent, G.; TournePeteilh, C.; Devoisselle, J.-M.; Babonneau, F. Pure Appl. Chem. 2009, 81, 1345–1355. (41) Azaís, T.; Tourne-Peteilh, C.; Aussenac, F.; Baccile, N.; Coelho, C.; Devoisselle, J.-M.; Babonneau, F. Chem. Mater. 2006, 18, 6382–6390. (42) Bras, A. R.; Frunza, S.; Guerreiro, L.; Fonseca, I. M.; Corma, A.; Frunza, L.; Dionísio, M.; Sch€onhals, A. J. Chem. Phys. 2010, 132, 224508. (43) Frunza, L.; Kosslick, H.; Sch€onhals, A.; Pitsch, I.; Frunza, S. Dynamic behavior of hybrid organic-inorganic composites: Liquid crystals confined in SBA-15 mesoporous materials. In Liquid Crystals VI; Khoo, I. C., Ed.; SPIE: Bellinham, 2002; Vol. 4799, pp 230-238. (44) Frunza, L.; Kosslick, H.; Frunza, S.; Sch€onhals, A. Microporous Mesoporous Mater. 2006, 90, 259–270. (45) Frunza, L.; Kosslick, H.; Pitsch, I.; Frunza, S.; Sch€onhals, A. J. Phys. Chem. B 2005, 109, 9154–9159. (46) Sliwinska-Bartkowiak, M.; Jazdzewska, M.; Huang, L. L.; Gubbins, K. E. Phys. Chem. Chem. Phys. 2008, 10, 4909–4919. (47) McConnell, J. F. Cryst. Struct. Commun. 1974, 3, 73–75. (48) Shankland, N.; Wilson, C. C.; Florence, A. J.; Cox, P. C. Acta Crystallogr. 1997, C53, 951–954. (49) Derollez, P.; Dudognon, E.; Affouard, F.; Danede, F.; Correia, N. T.; Descamps, M. Acta Crystallogr., Sect. B 2010, 66, 76–80. (50) Oliveira, P.; Machado, A.; Ramos, A. M.; Fonseca, I. M.; Braz Fernandes, F. M.; Botelho do Rego, A. M.; Vital, J. Catal. Commun. 2007, 8, 1366–1372. (51) (a) Corma, A. Chem. Rev. 1997, 97, 2373–2419. (b) Corma, A.; Grande, M. S.; Gonzalez-Alfaro, V.; Orchilles, A. V. J. Catal. 1996, 159, 375–382. (52) Charnay, C.; Begu, S.; Tourne-Peteilh, C.; Nicole, L.; Lerner, D. A.; Devoisselle, J. M. Eur. J. Pharm. Biopharm. 2004, 57, 533–540. (53) Affouard, F. DSMM, Universite Lille, France. Private communication. (54) Tanis, I.; Karatasos, K. J. Phys. Chem. B 2009, 113, 10984– 10993. (55) Vueba, M. L.; Pina, M. E.; Batista de Carvalho, L. A. E. J. Pharm. Sci. 2008, 97, 845–859. (56) Affouard, F.; Correia, N. T. J. Phys. Chem. B 2010, 114, 11397– 11402. (57) Only the broadening of the CdO band for amorphous Ibu is reported in ref 35; however, the shift of the bands centered in crystalline Ibu at 1508 and 1462 cm-1 was also observed. (58) Gopalakrishnan, T. R.; Beiner, M. Lett. Drug Des. Discovery 2006, 3, 723–730. (59) Johari, G. P.; Kim, S.; Shanker, R. M. J. Pharm. Sci. 2005, 94, 2207–2223. (60) Shen, S. C.; NG, W. K.; Chia, L.; Dong, Y. C.; Tan, R. B. H. J. Pharm. Sci. 2010, 99, 1997–2007. (61) Riikonen, J.; M€akil€a, E.; Salonen, J.; Lehto, V. P. Langmuir 2009, 25, 6137–6142. (62) Jackson, C. L.; McKenna, G. B. Chem. Mater. 1996, 8, 2128–2137. (63) Kinka, M.; Banys, J.; Naberezhnov, A. Ferroelectrics 2007, 348, 67–74. (64) Banys, J.; Kinka, M.; Macutkevic, J.; Volkel, G.; Bohlmann, W.; Umamaheswari, V.; Hartmann, M.; Poppl, A. J. Phys.: Condens. Matter 2005, 17, 2843–2857.

ARTICLE

(65) Bras, A. R.; Dionísio, M.; Correia, N. T. Complex Systems; Complex Systems Book Series: AIP Conference Proceedings; Tokuyama, M., Oppenheim, I., Nishiyama, H., Eds.; 2008, 982, 91-96. (66) Frunza, S.; Sch€onhals, A; Frunza, L.; Zubowa, H.-L.; Kosslick, H.; Fricke, R.; Carius, H. Chem. Phys. Lett. 1999, 307, 167–176. (67) Hartmann, L.; Kremer, F.; Pouret, P.; Leger, L. J. Chem. Phys. 2003, 118, 6052–6058. (68) Viciosa, M. T.; Rodrigues, C.; Fernandez, S.; Matos, I.; Marques, M. M.; Duarte, M. T.; Mano, J. F.; Dionísio, M. J. Polym. Sci., Part B 2007, 45, 2802–2812. (69) Sch€onhals, A.; Goering, H.; Schick, Ch.; Frick, B.; Zorn, R. Eur. Phys. J. E 2003, 12, 173–178. (70) W€ubbenhorst, M.; Lupascu, V. Proceedings of the ISE-12; Brazil, 2005. (71) Adam, G.; Gibbs, J. H. J. Chem. Phys. 1965, 43, 139. (72) Estimated for a conformer with lowest energy using MOPAC. http://openmopac.net/. (73) Frunza, L.; Frunza, S.; Kosslick, H.; Sch€onhals, A. Phys. Rev. E 2008, 78, 051701. (74) Vogel, H. Phys. Z. 1921, 22, 645–646. (75) Fulcher, G. S. J. Am. Ceram. Soc. 1925, 8, 339–355. (76) Tammann, G.; Hesse, W. Z. Anorg. Allg. Chem. 1926, 156, 245–257. (77) Huth, H.; Wang, L.-M.; Schick, C.; Richert, R. J. Chem. Phys. 2007, 126, 104503. (78) Johari, G. P.; Dannhauser, W. J. Chem. Phys. 1969, 50, 1862–1876. (79) Wang, L.-M.; Shahriari, S.; Richert, R. J. Phys. Chem. B 2005, 109, 23255–23262.

4623


Amorphous Ibuprofen Confined in Nanostructured Silica Materials: A

Recommend Documents