Solid-State NMR and Dielectric Studies on Dynamic Heterogeneity of

Article pubs.acs.org/molecularpharmaceutics

A Stable Amorphous Statin: Solid-State NMR and Dielectric Studies on Dynamic Heterogeneity of Simvastatin Teresa G. Nunes,*,† M. Teresa Viciosa,‡ Natália T. Correia,§,∥ F. Danède,§ Rita G. Nunes,⊥ and Hermínio P. Diogo† †

Centro de Química Estrutural and ‡Centro de Química-Física Molecular and INInstitute of Nanoscience and Nanotechnology, Instituto Superior Técnico/Universidade de Lisboa, Avenida Rovisco Pais, 1049-001 Lisboa, Portugal § Unité Matériaux et Transformation (UMET), UMR CNRS 8207, UFR de Physique, BAT P5, Université Lille 1, 59655 Villeneuve d’Ascq, France ∥ REQUIMTE/CQFB, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal ⊥ Instituto de Biofísica e Engenharia Biomédica, Faculdade de Ciências, Universidade de Lisboa, Campo Grande, 1749-016 Lisboa, Portugal ABSTRACT: Statins have been widely used as cholesterollowering agents. However, low aqueous solubility of crystalline statins and, consequently, reduced biovailability require seeking for alternative forms and formulations to ensure an accurate therapeutic window. The objective of the present study was to evaluate the stability of amorphous simvastatin by probing molecular dynamics using two nondestructive techniques: solid-state NMR and dielectric relaxation spectroscopy. Glassy simvastatin was obtained by the melt quench technique. 13C cross-polarization/magic-angle-spinning (CP/ MAS) NMR spectra and 1H MAS NMR spectra were obtained from 293 K up to 333 K (Tg ≈ 302 K). The 13C spin−lattice relaxation times in the rotating frame, T1ρ, were measured as a function of temperature, and the correlation time and activation energy data obtained for local motions in different frequency scales revealed strong dynamic heterogeneity, which appears to be essential for the stability of the amorphous form of simvastatin. In addition, the 1H MAS measurements presented evidence for mobility of the hydrogen atoms in hydroxyl groups which was assigned to noncooperative secondary relaxations. The complex dielectric permittivity of simvastatin was monitored in isochronal mode at five frequencies (from 0.1 to 1000 kHz), by carrying out a heating/cooling cycle allowing to obtain simvastatin in the supercooled and glassy states. The results showed that no dipolar moment was lost due to immobilization, thus confirming that no crystallization had taken place. Complementarily, the present study focused on the thermal stability of simvastatin using thermogravimetric analysis while the thermal events were followed up by differential scanning calorimetry and dielectric relaxation spectroscopy. Overall, the results confirm that the simvastatin in the glass form reveals a potential use in the solid phase formulation on the pharmaceutical industry. KEYWORDS: amorphous simvastatin, solid-state NMR, DSC, TGA, DRS changes that may influence the drug properties,3 for example, the increase of their shelf-life time. To characterize pharmaceutical solids, we may use various techniques. Spectroscopic techniques such as Fourier transform Raman spectroscopy (FT-RS), Fourier transform infrared spectroscopy (FT-IR), and solid-state nuclear magnetic resonance spectroscopy (ss-NMR) are primarily intramolecular methods, probing the sample at the molecular level. Intermolecular information is gained by directly employing

1. INTRODUCTION The growing number of active pharmaceutical ingredients (APIs) with poor aqueous solubility has led to development of various strategies to improve dissolution rates. The transformation from their crystalline state to a more soluble amorphous, nanocrystalline solid dispersion and/or solid solution represents the most promising ways as the bioavailability of APIs depends on solubility in human fluids and permeability in the gastrointestinal tract.1 As the amorphous form of drugs is thermodynamically less stable than their crystalline counterparts, different methods for stabilization of these forms are emerging to profit from their solubility and dissolution rate advantages.2 Hence, a detailed solid-state study can be very useful to predict any structural © 2014 American Chemical Society

Received: Revised: Accepted: Published: 727

August 1, 2013 January 31, 2014 February 5, 2014 February 5, 2014 dx.doi.org/10.1021/mp400455r | Mol. Pharmaceutics 2014, 11, 727−737

Molecular Pharmaceutics

Article

separately and provides information about the global αrelaxation as well as the local relaxations. Hence, studying mobility at different frequencies by ss-NMR allows not only obtaining information on time scales of both the structural αrelaxation and the secondary relaxations but also identifying the chemical groups involved in localized dynamic processes. In this context, the present study reports an evaluation on the mobility of amorphous simvastatin in different frequency scales using conventional 1H MAS and 13C CP/MAS NMR techniques over temperatures ranging from 293 to 333 K (below and above the glass transition temperature region), and line shape simulations as already described elsewhere.9 Moreover, it is well-known that NMR relies on local environment rather than long-range order. In fact, broad lines are expected from an amorphous compound thus reflecting the presence of a distribution of local electronic shielding; however, under similar experimental conditions, a line narrowing indicates that crystallization has occurred, in agreement with evidence for short-range order. Thus, NMR was also used to probe the stability of the amorphous simvastatin, by periodically evaluating any NMR spectral change. In fact, the amorphous state is a high-energy state resulting in enhanced dissolution rate, but from a thermodynamic approach it is a metastable state and crystallization is an event with a high probability to occur. Literature data show examples where the kinetics of the crystallization processes is dependent on the route to obtain the amorphous state.10 Possible differences in physicochemical properties and stability between amorphous forms of simvastatin prepared by (a) cryomilling and (b) melting and quench-cooling were investigated using several techniques; in particular, it was found that cryomilled samples are less disordered, have a lower stability, and have decreased recrystallization enthalpy.10 Therefore, the object of the present study was amorphous simvastatin prepared by the melting and quench-cooling technique. The present study also allowed information to be obtained on the thermal stability of simvastatin using TGA while the thermal transitions were followed up by DSC and DRS which in addition enabled us to detect if crystallization had occurred.

techniques such as differential scanning calorimetry (DSC), thermogravimetric analysis (TGA), and X-ray powder diffraction (XRPD), which analyze the sample on a global level. Among all these methods NMR is considered to play an important role in the study of polymorphic and solid amorphous forms.4 An example of a poor aqueous soluble drug is simvastatin, a cholesterol-lowering agent commonly used to treat hypercholesterolemia (see Scheme 1), which acts as a prodrug. Scheme 1. Structural Formula of Simvastatin

Simvastatin presents three crystalline polymorphs that have been already identified with well-defined temperature regions (enantiotropic system);5 the main difference between the three solid phases was attributed to a conformationally disordered part of the side chain ethyl group (see Scheme 1). Form I of simvastatin (the stable form at room temperature) is used in pharmacology, and it is characterized by very low solubility in water (0.03 mg·mL−1) and therefore low oral bioavailability (5%).6 Thus, it is important to focus on increasing the absorbency of this substance.1 For example, high improvement in solubility and in vitro drug release was observed in solid dispersion prepared from simvastatin and PVP by a solvent evaporation method.7 It is therefore of interest to evaluate the molecular dynamics of amorphous simvastatin since it is believed that mobility is one of the main factors that govern the physical stability.8 Amorphous pharmaceutical drugs generally exhibit higher molecular mobility than the corresponding crystalline forms, which is normally assumed to be an appropriate indicator of lower stability. Simvastatin, in the amorphous state, is an excellent example where high mobility does not imply low stability, and this issue is a highlight pulled out from the studies performed in the present investigation. Therefore, the main goal of this study is to obtain an elucidation for the high stability of the amorphous simvastatin, using nondestructive techniques such as dielectric relaxation spectroscopy (DRS) and solid-state NMR (ss-NMR). DRS probes the molecular dynamics through the motions of dipoles which fluctuate under the influence of an external oscillating electric field, allowing inferences to be made about localized (βrelaxation, γ-relaxation, etc.) and cooperative or global (αrelaxation) molecular motions. Nuclei with spin number different from zero can be observed by NMR spectroscopy, producing a free induction decay signal when submitted both to a high static magnetic field (B0) and to a radiofrequency field; different relaxation mechanisms determine the return of spin states’ populations to thermodynamic equilibrium under the presence of B0. The 13C ss-NMR relaxation technique, for example, allows each carbon of a molecule to be observed

2. EXPERIMENTAL SECTION 2.1. Material and Sample Preparation. Simvastatin was a kind gift from Mepha Lda. “Investigaçaõ , Desenvolvimento e Fabricaçaõ Farmacêutica” (Porto Salvo, Portugal). The purity was higher than 99% in accordance with the supplier. It is a white, crystalline powder, and it was used without further purification. The amorphous sample for ss-NMR studies was obtained by the melting and quench-cooling method: simvastatin as received was heated at ca. 10 K·min−1 up to 423 K (∼10 degrees above the melting temperature) for approximately 3 min, under inert atmosphere, and the melt was cooled down to room temperature.11 The glassy sample was gently powdered in a mortar and placed in the solid-state NMR sample holder. Both samples, either crystalline or amorphous, were kept inside a desiccator over P2O5 several days before NMR measurements. 2.2. Methods. 2.2.1. Solid-State NMR. NMR spectra were recorded on a Tecmag Redstone/Bruker 300 WB spectrometer operating at 75.47 and 300.13 MHz for 13C and 1H observations, respectively. Powdered samples were packed into standard 4 and 7 mm outer diameter ZrO2 rotors, and double-resonance probes were used. 13C cross-polarization/ 728

dx.doi.org/10.1021/mp400455r | Mol. Pharmaceutics 2014, 11, 727−737


Article

analyzer from Novocontrol Technologies GmbH. More details about the used equipment and methodology to analyze dielectric data can be found elsewhere.14 It should be noted that during all dielectric measurements samples are kept under a constant nitrogen gas flux allowing a dried atmosphere to be maintained. A crystalline sample was placed in parallel plate geometry between two gold-coated electrodes with a diameter of 10 mm. Two silica spacers of 50 μm thickness were used in order to avoid contact between the electrodes when the crystalline simvastatin melts and to ensure a constant thickness. In order to study the thermal transitions of simvastatin, the complex dielectric permittivity was measured in isochronal mode, ε*(T), at five frequencies sufficiently high to allow a reliable measurement (0.1, 1, 10, 100, 1000 kHz) during heating/cooling cycles performed at a rate of around 10 K· min−1, in the temperature range between 153 and 433 K.

magic-angle-spinning (CP/MAS) spectra were obtained with 3.4 kHz rate, 1 or 3 ms contact time (for the observation of crystalline or amorphous samples, respectively), 4 μs RF-pulse duration (90° magnetization tip angle), 1H decoupling RF-field of 50 kHz using continuous wave irradiation at the nominal frequency of protons, 3 s recycling delay, and 400 scans. 1H spectra were obtained with a MAS rate of 6.3 kHz. Glycine (δ13CO = 176.03 ppm) and ethanol (δCH3 = 1.23 ppm) were used as external references for 13C and 1H chemical shifts, respectively. The significant contributions of the 13C spinning side bands, particularly in the aromatic and carbonyl regions, were evaluated by comparing spectra run with and without the SELTICS sequence (Sideband Elimination by Temporary Interruption of Chemical Shift).12 Temperature was controlled with a Bruker variable-temperature unit B-VT 1000E. In general, ten kelvin increments were selected, and experiments started from the lowest temperature. Thirty minutes waiting time allowed each temperature to stabilize; the probe was always tuned before starting any experiment. The spin−lattice in the rotating frame of carbon nuclei (CT1ρ) was measured at 293, 298, 303, 311, 318, 325, and 333 K, under a MAS rate of 3005 ± 4 Hz using a standard CP pulse sequence, with 1 ms contact time, 5 μs for the 90° pulse duration, with proton spin-locking interruption from 2 to 16 ms over the cross-polarization period, and 5 s recycling delay. The probe ring down was 8 μs. The CT1ρ values were determined by the least-squares fit using the experimental data and the exponential function that characterize the rate of decay of the spin-locked carbon magnetization: M(t ) = M 0 exp( −t / CT 1ρ)

3. RESULTS AND DISCUSSION 3.1. Thermal Transitions Studied by DSC and Isochronal DRS Measurements. The thermal behavior of simvastatin was first studied by thermogravimetric analysis (TGA) under nitrogen atmosphere. The corresponding TGA and derivative curves are shown in Figure 1 that reveals that

(1)

where M(t) is the magnetization at time t and the initial magnetization at t = 0, M0, was obtained without proton spinlocking interruption. The rotational correlation time τC of the motion responsible for each CT1ρ (see section 3.3.2 and Table 2) was obtained by finding the roots of eq 3 using the “fzero” algorithm of the MATLAB computer program (MathWorks, USA). 2.2.2. Thermogravimetric Analysis. A sample of 2.784 mg was placed in an open platinum sample pan and the TGA measurements were carried out with a TGA 7 apparatus from Perkin-Elmer, at a heating rate of 10 K·min−1 under highly pure nitrogen atmosphere with a flow rate of 20 mL·min−1. The temperature reading was calibrated using the Curie points of alumel and nickel standards, while the mass reading was calibrated using balance tare weights provided by Perkin-Elmer. 2.2.3. Differential Scanning Calorimetry. Differential scanning calorimetry (DSC) measurements were performed with a 2920 MDSC system from TA Instruments Inc., used in the conventional mode. The measuring cell was continuously purged with high purity helium gas at 30 mL·min−1. An empty aluminum pan, identical to that used for the sample, was used as reference. Details of the calibration procedures are given elsewhere.13 Samples of around 5−7 mg of crystalline simvastatin were placed on the base of aluminum pans, heated up to 373 K inside a vacuum oven during 2 h, cooled down to around 303 K, and finally, hermetically encapsulated using a lid. The mass of the samples was weighed with a precision of ±0.1 μg in a Mettler UMT2 ultramicro balance. 2.2.4. Dielectric Relaxation Spectroscopy. Dielectric measurements were carried out using an Alpha-N impedance

Figure 1. Thermogravimetric curve (left-hand vertical axis; black line) and derivative curve (right-hand vertical axis; gray line) obtained on heating at 10 K·min−1 the as received simvastatin. The onset of thermal decomposition occurs around 483 K. In the inset a close-up of the temperature region below 383 K is shown: a small weight loss is observed due to water evaporation (∼0.3 wt %).

decomposition of simvastatin occurs in two weight loss steps, in agreement with results reported in the literature for TGA studies of simvastatin also carried out in nitrogen atmosphere.15 Crystalline simvastatin is thermally stable until around 483 K; below this temperature, only a close inspection of the thermogravimetric data for temperatures below 383 K allows identifying a small weight loss (∼0.3 wt %) due to water loss (inset in Figure 1). This quantity of water, in spite of being small, prevented a reliable analysis of the melting process, at least when carried out in hermetically sealed pans. Consequently, before all studies (particularly by NMR), the samples were dried and kept inside a desiccator over P2O5 (see Material and Sample Preparation). Thermal transitions of simvastatin were investigated by DSC and DRS measurements (in the isochronal mode). DSC thermograms obtained upon two heating scans carried out at 10 K·min−1 are shown in Figure 2. In the first heating scan (run 1) the as received simvastatin (after being dried) was heated up to 729



Article

ε″, of the dielectric permittivity are plotted in panels a and b, respectively.

Figure 2. DSC thermograms of simvastatin obtained upon heating at 10 K·min−1. The first heating scan of the as received crystalline form (I) is shown in the main graph (run 1), where melting is observed at Tm. The inset enlarges the glass transition region, at Tg, obtained on cooling (run 2) from the melt down to 213.15 K, and on the subsequent heating (run 3).

433 K. A sharp endothermic peak due to melting of simvastatin (polymorphic form I) is observed, with an onset at 412.9 K, peak temperature at Tm = 414.7 K, and an estimated enthalpy of melting, ΔHm = 74.5 J·g−1, very close to values reported in the literature (Tm,onset = 412.32 K, ΔHm = 68.33 J·g−1;10 Tm = 412.65 K, ΔHm = 77.39 J·g−1;16 and Tm,onset = 412.2 ± 0.2 K, Tm,max = 414.1 ± 0.2 K, ΔHm = 72.6 ± 0.4 J·g−1 17). In the subsequent cooling of the melt (run 2), crystallization of simvastatin is avoided and a clear signal of the glass transition is observed in the thermogram. The second heating scan (run 3) is illustrated in the inset of Figure 2: the glass transition, with the characteristic enthalpy overshoot, can be identified with an onset temperature Tg,onset = 303.2 K and heat capacity jump ΔCp = 0.386 J·(gK)−1, in good agreement with results reported by Graeser et al. for amorphous simvastatin also obtained by the melt and quenching-cooling method (Tg,midpoint = 302.05 K, ΔCp = 0.334 J·(gK)−1, obtained from modulated temperature DSC measurements carried out at a heating rate of 3 K· min−1).10 No cold crystallization is observed on further heating up to 433 K (data not shown), which is also evidence of the high resistance of the amorphous form to crystallization. DRS probes the molecular dynamics through the reorientational motions of dipoles, which fluctuate under the influence of an outer oscillating electric field. At a given temperature, the measured dielectric permittivity, ε*( f) = ε′( f) − i ε″(f) ( f: the frequency of the electric field), is a complex quantity whose real part (ε′) decreases as a sigmoid curve and whose imaginary part (ε″) passes through a maximum at f max, when the dipoles are no longer able to follow the electric field oscillations. DRS has been proven to be a suitable tool to study thermal transitions in low molecular weight compounds,14,18 including pharmaceutical drugs.19 The complex dielectric permittivity of simvastatin was monitored in isochronal mode, ε*(T), at five frequencies (0.1, 1, 10, 100, 1000 kHz), by carrying out a heating (run 1)/ cooling (run 2) cycle allowing simvastatin to be obtained in the supercooled and glassy states (similar thermal protocol used in the DSC experiments). Results at some representative frequencies are shown in Figure 3, where the temperature dependence of the real part, ε′, and that of the imaginary part,

Figure 3. Isochronal measurements at representative frequencies (see legend): (A) real part of dielectric permittivity and (B) imaginary part as a function of temperature. Run 1 (symbols): ε′(T) obtained upon heating the as received crystalline simvastatin. Run 2: ε′(T) (lines in A) and ε″(T) (symbols in B) measured during cooling of the melt from 433 K down to 153 K. Melting temperature, Tm, and the glass transition temperature, Tg, are indicated. ss-NMR studies presented in this work have been performed in the glass transformation region identified in the figure by the gray area (293 K < T < 333 K).

In order to monitor melting, a slightly compressed powder of crystalline simvastatin was heated from room temperature up to 433 K (run 1, open symbols in Figure 3A). When melting occurs, ε′(T) shows a steep increase whose position is independent of the measuring frequency (typical behavior of a first order transition); the melting temperature taken at midpoint is 412 K, in good agreement with calorimetric measurements (Tm,onset = 412.87 K). Absolute values for ε′(T) measured during run 1 are omitted due to the fact that both thickness and volume of the sample change during melting. After being kept at 433 K for ∼10 min in order to ensure complete melting, the sample was cooled down to 153 K at ∼8 K·min−1 (run 2). The ε′(T) trace (lines in Figure 3A), after a linear increase with the temperature decrease, shows a marked fall just before 363 K (at 100 kHz) indicating the transformation, at the measuring frequency, from the supercooled liquid to the glass. This step in the ε′(T) trace is frequency dependent, and in the ε″(T) trace (Figure 3B), it corresponds to an intense peak, i.e., the main relaxation process associated with the dynamical glass transition. At temperatures below Tg, 730



Article

Figure 4. 13C CP/MAS NMR spectra obtained at 293 K from different simvastatin samples using the SELTICS sequence:12 crystalline (bottom) and amorphous (top). The inset shows an expanded resonance frequency region of the carbonyl groups in the amorphous simvastatin.

moment is lost due to immobilization that would occur if crystallization had taken place. 3.2. 13C CP/MAS NMR at 293 K from Crystalline and Amorphous Simvastatin. In general, the isotropic NMR signals of a solid cannot be identified based only on solution NMR data comparisons. 13C CP/MAS NMR spectra depend on crystal rearrangements, the presence of interactions of either inter- or intramolecular origin, and molecular conformation. Full unambiguous assignment of 13C resonances was achieved from correlation spectroscopic techniques24 used for the study of simvastatin, both in solution and as a crystalline powder, which is followed in the present study. Moreover, the complete assignment of 13C signals obtained from three simvastatin polymorphs (the stable polymorph at room temperature (I) and two lower temperature polymorphs, II and III) was recently reported.5 Figure 4 shows typical 13C CP/MAS spectra which were obtained from different simvastatin samples at 293 K: crystalline (polymorph I) and glass. Table 1 displays the 13C chemical shifts of carbon nuclei obtained for both crystalline and amorphous simvastatin. Five frequency regions are identified in the spectra obtained from crystalline samples, which correspond to the carbon numbered according to Scheme 1 (within parentheses): 185−165 ppm (18, 1), 140−120 ppm (10, 12, 11, 17), 80−60 ppm (5, 14, 3), 45−20 ppm (19, 2, 4, 8, 13, 6, 15, 20, 9, 16, 22, 7, 23, 24) and 20−10 ppm (25, 21). The full spectral assignment is also shown in Figure 4. Due to the presence of superimposed signals, the chemical shifts of the following nuclei remain ambiguous: (a) 2, 4, 8, 13 and (b) 7,

broad and low intense peaks are observed associated with secondary relaxations. The influence of preparation methods on the stability against crystallization has been already studied.20 The absence of any discontinuity in the ε′(T) trace (and also in ε″(T)) upon cooling until the region where the main relaxation evolves led us to conclude that crystallization of simvastatin was circumvented. The increase observed in ε′ with the temperature decrease is caused by the expected increase of the dielectric strength and can be quantified according to the Fröhlich−Kirkwood equation:21,22 N

εs − ε∞ =

μ0 2 g V εs(ε∞ + 2)2 9ε0kBT (2εs + ε∞)

(2)

where μ0 is the dipolar moment of the isolated dipole; g (the well-known Kirkwood factor) takes into account the dipole− dipole correlation (for parallel or antiparallel correlations between neighboring dipoles, g > 1 or 0 < g < 1, respectively, while for a random orientation distribution of dipoles, g = 1); εs and ε∞ are the limits of the real part of the dielectric permittivity at low and high frequencies respectively, the latter being approximately the permittivity of the glass;23ε0 is the vacuum permittivity; N/V is the number of dipoles per unit of volume; and kB is the Boltzmann constant. From eq 2, the dielectric strength (Δε = εs − ε∞) is proportional to μ02g/ 9ε0kBT. If it is assumed that the Kirkwood factor g is constant, an increase in Δε from 413 to 358 K of 13.4% is predicted. The observed increase is 15.4% (ε′(T = 413.8 K) = 6.79 and ε′(T = 358.3 K) = 8.03 for f = 10 kHz), confirming that no dipolar 731



Article

23. The resonances from 8 and 13 nuclei were also not resolved by Brus et al.24 in spite of using a higher magnetic field than this study. The good spectral resolution of the crystalline sample is a consequence of short distance order and rigidity of the simvastatin molecules. The X-ray structure of simvastatin (room temperature polymorph) suggests that molecular packing is based on only one hydrogen bond, C3O−H··· OC18, which facilitates the formation of infinite chains of molecules along the b axis.25 Accordingly, the corresponding 13 C CP/MAS spectrum (Figure 4, bottom) presents a single narrow resonance for C18 at 180.23 ppm. On the other hand, the spectrum of the amorphous sample (the inset in Figure 4) shows significant asymmetrical line broadening of C18 and C1 resonances: for C1, ranging from about 168 to 175 ppm, with a maximum intensity at 172 ppm, and for C18, ranging from about 175 to 182 ppm, with a maximum intensity at 177 ppm. This observation is consistent with the presence of a distribution of conformations. For example, an asymmetric shape of line was already reported on C1,1′ of glassy α−α′ tetrahalose, and the main source of structural disorder was assigned to the distribution of glycosidic torsion angles.26 A similar observation was mentioned for C1 in amorphous gentiobiose.9 Although the glassy state is structurally rigid, it retains part of the structural distribution present in the liquid state, allowing accessing the conformational degrees of freedom reachable to the simvastatin molecule. Consequently, the spectrum obtained from amorphous simvastatin shows broad lines, in agreement with lack of order at short distance. The comparison of amorphous and crystalline simvastatin NMR spectra (Figure 4) enables observing the following: (a) different relative intensity of some signals, in particular from CH groups (60−80 ppm),

Table 1. Chemical Shifts (ppm) of the Carbon Nuclei, Numbered As Shown in Scheme 1, Obtained for Crystalline and Amorphous Simvastatin simvastatina

a

carbon atom

crystalline (from lit.)24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

170.53 (170.66) 37.3** (38.31) 62.18 (62.35) 37.3** (37.17) 76.73 (76.83) 35.00 (35.16) 24.0* (24.78) 37.3** (37.34) 29.51 (29.65) 135.92 (135.91) 128.16 (128.26) 133.98 (133.98) 37.3** (37.34) 70.91 (70.91) 32.10 (32.80) 27.57 (27.76) 127.84 (128.05) 180.23 (180.28) 42.77 (43.02) 32.10 (32.16) 9.78 (9.95) 25.63 (25.52) 24.0* (24.25) 22.71 (22.95) 14.30 (14.30)

amorphous 172 (asym) 37.1** 62.1 37.1** 76.9 24.2 37.1** 31.0 (shoulder) 135 (very broad) 128.9 (broad) 132.4 (broad) 37.1** 68.8 33.4*** 27.3 (shoulder) 177 (asym) 42.8 33.4*** 9.4 24.2* 24.2* or 19.4 13.8 (broad)

*, **, *** superimposed signals.

Figure 5. 13C CP/MAS NMR spectra obtained from amorphous simvastatin at the indicated temperatures: (A) magnetization without interruption of proton spin-locking during the CP period (with SELTICS);12 (B) with interruption of proton spin-locking over 16 ms during CP. 732



Article

groups), C10, C11, C12, and C17 signals also broadened with temperature (the frequency range 120−185 ppm is not shown here). 3.3.2. Carbon Spin−Lattice Relaxation in the Rotating Frame. Carbon spin−lattice relaxation time in the rotating frame (CT1ρ) enables probing the random orientational motion of dipolar coupled CH nuclei that occurs in the rotating frame Larmor frequency ω1, that is, in the kHz frequency range. In general, CT1ρ for amorphous solids is primarily determined by molecular motion; strong static dipolar H−H effects (spin− spin diffusion) are absent because of molecular motion or large interproton distances. Figure 5B presents 13 C CP/MAS NMR subspectra (frequency regions of the whole spectra) of amorphous simvastatin obtained at 293 K, 303 K, 325 K, and 333 K, respectively, which were acquired with proton spin-locking interruption over 16 ms during the cross-polarization period. It is clearly observed that carbons 19, 21, 22, 23, and 7 must have longer CT1ρ than other nuclei because of the lower decrease of the corresponding signal intensities with temperature. As a consequence, for example, C19 resonance which was shown to be superimposed with other signal at 293 K becomes distinctly resolved at 333 K. Overall, the influence of temperature was not particularly noticed for the quaternary carbon C19 and the methyl groups in the ester tail (C21, C22, and C23). Such observation could be an indication that no motion takes place at the kHz frequency in the temperature range studied or, considering the chemical structure of simvastatin and reasonably assuming that the rotation of these methyl groups is favored, most probably these nuclei are involved in a mobility regime with rates well above 50 kHz (the 1H spin-locking frequency) and related to a fast secondary relaxation process. Therefore, the influence of temperature on mobility in the probed frequency range was particularly noticeable for other simvastatin carbon nuclei (for example 2, 4, 8, 13 and 6, 15, 20) due to the strong line shape variation, which points to an intermediate regime. The signals of aromatic (C10,11,12,17) and carbonyl (C1,18) nuclei (not shown here) were strongly broadened by increasing the temperature. Figure 6 shows CT1ρ measured from 293 till 333 K for several carbon nuclei with the following chemical shifts (carbon numbers within parentheses): 77 (5), 69 (14), 62 (3), 33 (6, 15, 20), and 24 ppm (7, 22, 23). Single-exponential decays were used to fit the experimental data recorded at each temperature. At first observation, it is noticed that the variation of CT1ρ with the inverse of temperature presents two different trends: (a) no significant variation for carbons 7, 22, 23 and (b) a decrease up to 333 K for carbons 3, 5, 6, 14, 15, and 20, which show similar C T1ρ above Tg (0.00331 K−1). Hence, this fact may indicate that heating the amorphous sample up to a temperature higher than Tg liberated a molecular mobility and possibly a subsequent cooperative motion. For example, cooperative torsional oscillations of the rings may explain the observed CT1ρ decrease at T > Tg: (CT1ρ)−1 = K2 sin2 θ J(ω), where θ is the dipolar fluctuation amplitude due to motion orthogonal to the RF field, K is a constant that includes powder averaging, and J(ω) is the spectral density function (the Fourier transform of the correlation function of the carbon−proton bond, that is, the probability function of finding motions at the frequency ω).32 The long CT1ρ of C19 (from about 50 ± 6 ms at 293 K to 33 ± 2 at 333 K) is consistent with the fact that the dipolar relaxation is more efficient when the carbon has attached protons like methyne and methylene carbon nuclei. Thus, this

and (b) large broadening of C24 and C25 (methyl groups) signals. The observed line broadening of C24 and C25 (methyl groups) resonances is explained as follows. In general, motions with rates between 10 kHz and 1 MHz (representing the intermediate regime, at which the inverse of the time scale of the motion is comparable with the magnitude of the interaction under observation) produce a broadening in the CP/MAS spectra due to the motion interference with the dipolar decoupling.27,28 Therefore, line broadening in 1H decoupled 13 C CP/MAS spectra is due to the presence of molecular motions that occur on a time scale comparable to the inverse of the 1H decoupling frequency (50 × 2π × 103 rad·s−1), then of the order of about 10−6 s. Under this condition, the molecular motion interferes with the decoupling frequency, resulting in the observed line broadening.27,28 Furthermore, molecular motions occurring in such intermediate regime affect the crosspolarization efficiency,29 producing a decrease in the overall signal intensity. This result shows that C24 and C25 are the less stereochemically hindered methyl groups in simvastatin. In contrast, geometrically restricted motions taking place with rates well above the 1H decoupling frequency produce partial averaging of the effective dipolar coupling without decoupling interference, producing a spectral renarrowing; this is the case of the C21 signal. In the next section it will be shown that other nuclei in the ester group undergo motions with rates in a similar frequency scale. Thus, these results show evidence for local mobility in different frequency regimes which could be a factor for increasing the configurational entropy. This fact may prevent recrystallization explaining the high stability of glassy simvastatin. In this context, it is noteworthy that a reasonable correlation between the thermodynamic parameters and the stability above Tg was found by investigating thermodynamic and kinetic parameters as potential predictors of physical stability of amorphous drugs, including simvastatin, with the configurational entropy exhibiting the strongest correlation.30 Moreover, it was also hypothesized that the stability and very low enthalpy recovery of glassy simvastatin could be attributed to strong intermolecular hydrogen bonding.31 3.3. 13C CP/MAS and 1H MAS NMR from Amorphous Simvastatin: Influence of the Temperature. 3.3.1. 13C CP/ MAS at 293 - 333 K. 13C CP/MAS signals arise mainly from molecules that are rigid or execute very restricted dynamics. Therefore, the CP/MAS sensitivity to molecular dynamics can be explored to obtain information about moving molecular groups or segments. Figure 5A shows 13C CP/MAS subspectra of amorphous simvastatin obtained at 293 K, 303 K, 325 K,and 333 K, respectively. It is clearly observed that the intensity of most of the signals decreases by increasing the temperature while line broadening increases. In particular, it may be noticed that the intensities of the resonances in the region 30−40 ppm, which were assigned to CH2 (carbons 2, 4, 6, 15, 20) and CH (carbons 8, 9, 13, 16) groups, have decreased with temperature, when compared with the two signals at lower frequency. Line narrowing of signals recorded from C19, C21, C22, C23 (in the ester tail) and from C7 is observed when the sample is heated up, showing that the mobility is fast enough to produce the averaging of the CH dipolar couplings. However, since there is still an efficient cross-polarization, the motional amplitudes should be restricted.29 Conversely, large broadening is noticed for signals of C3, C5 and C14. Otherwise, C1, C18 (CO 733



Article

J(ω1) = τC/(1 + ω12τC 2) J(ωC) = τC/(1 + ωC 2τC 2) J(ωH) = τC/(1 + ωH 2τC 2) J(ωH − ωC) = τC/[1 + (ωH − ωC)2 τC 2] J(ωH + ωC) = τC/[1 + (ωH + ωC)2 τC 2]

Here γC and γH are the gyromagnetic ratios for the 13C and 1H nuclei, respectively, N is the number of directly bound atoms, and r is the carbon−proton internuclear distance (the typical value for the CH bond length is 0.109 nm), where h is the Planck’s constant, μ0 is the magnetic permeability of free space, ωH and ωC are the Larmor frequencies of 13C and 1H nuclei, respectively, and ω1 (50 × 2π × 103 rad/s) is the spin-lock field. Therefore, for each temperature and the corresponding measured CT1ρ, we have obtained τC using eq 3. Table 2 shows Figure 6. Spin−lattice relaxation time in the rotating frame of the indicated carbon nuclei measured from 293 till 333 K. The signals were observed at the following chemical shifts (carbon number within parentheses): 77 (5), 69 (14), 62 (3), 33 (6, 15, 20), and 24 (22, 7, 23) ppm. The dashed line indicates the glass transition temperature.

Table 2. Rotational Correlation Time (τC) at the Mentioned Temperature and Activation Energy (Ea) Obtained for the Motions of the Indicated Carbon Atoms carbon number

result confirms that, in simvastatin, dipole−dipole interaction is the driven mechanism for spin−lattice relaxation in the rotating frame; quaternary carbons present higher CT1ρ. The frequencies of motions are expected to increase with T, and a minimum relaxation time should be observed when they became equal to the characteristic frequency of CT1ρ.33 The variation of CT1ρ with temperature (Figure 6) shows that C3, 5, 14, 20 undergo motions on the slow-frequency side of the CT1ρ minimum, under slow motion conditions ϕ12τC2 ≫ 1, where τC is the rotational correlation time. The parameter τC is the time that a chemical species takes to rotate one radian and, therefore, is a direct measure of the rate of motion. Hence, the decrease in CT1ρ with T represents an increase in mobility at higher temperatures for these carbons. Thus, in the probed temperature range (which was not extended due to technical constraints) amorphous simvastatin is rigid as far as CT1ρ is concerned. It was reasonable to hypothesize that below the glass transition temperature (Tg ≃ 302 K), besides methyl groups 24 and 25, was the ester tail that presented the highest mobility. In fact, it is the disorder of this group that mainly explains the existence of three crystalline polymorphs.5 However, it is shown here that the mobility of the rings must also be taken into account. C T1ρ can be expressed in terms of an isotropic τC as an approximate model for the molecular motions by the wellknown BPP (Bloembergen−Purcell−Pound) function,34,35 which is only valid in the so-called “weak collision” case where τC is less than the spin−spin−lattice relaxation time T2:36 C

T1ρ

−1

τC/ 10−4 s

Ea/kJ· mol−1

3 (CH)

5 (CH)

6, 15, 20 (CH2)

14 (CH)

293 298 303 311 318 325 333

4.7 ± 0.6 6.2 ± 0.7 5.8 ± 0.5 6.9 ± 0.5 3.4 ± 0.5 2.8 ± 0.5 1.2 ± 0.3 27.3 ± 9.3

6.2 ± 0.8 8.5 ± 0.8 8.0 ± 0.3 6.3 ± 0.3 4.9 ± 0.6 3.9 ± 0.6 3.1 ± 0.4 18.1 ± 4.3

3.3 ± 0.4 3.6 ± 0.3 3.7 ± 0.2 3.2 ± 0.2 2.9 ± 0.1 2.2 ± 0.2 1.4 ± 0.1 16.1 ± 4.2

9.9 ± 2.3 11.9 ± 0.9 10.1 ± 0.6 8.1 ± 0.3 7.0 ± 0.6 4.9 ± 0.8 2.4 ± 0.7 27.6 ± 5.7

τC values from 293 to 333 K obtained for some carbon atoms. Although we have used all the spectral density functions in eq 3 to obtain the data presented here, it must be pointed out that J(ω1) is the function that mostly determines CT1ρ. It is usually assumed that the theoretical temperature dependence of the rate of rotational motions is a simple Arrhenius expression, τC = τ0 exp(Ea/RT), where Ea is activation energy for the molecular motions and R is the molar gas constant (8.314 J·K−1·mol−1). Accordingly, a plot of the natural logarithm of the correlation time as a function of the inverse temperature is linear with a slope that is proportional to Ea for motion. Therefore, further insight into simvastatin dynamics can be gained by examining the activation energies associated with the various carbons, which were obtained using Table 2 data. Table 2 shows Ea values for the motions of carbons in methyne (3, 5, 14) and methylene (6, 15, 20) groups. Clearly, there are two different data sets: about 28 kJ·mol−1 and 17 kJ·mol−1, respectively for C3/C14 and C5/C6, 15, 20. That is, Ea for C3/C14 motions is about twice as high as the values obtained for C5/C6, 15, 20 and may be assigned to slow and fast βsecondary processes. A methodology similar to the one described here was followed to study, for example, the dynamics of poly(methyl acrylate-co-sodium methacrylate) ionomer.37 3.4. 1H MAS NMR Spectra of Amorphous Simvastatin: Influence of the Temperature. Molecular dynamics of

= (N /20)(γCγHh/2πr 3)2 (μ0 /4π )2 [4J(ω1) + J(ωH − ωC) + 3J(ωC) + 6J(ωH + ωC)+6J(ωH)]

T/K

(3)

where, if a single exponential correlation function with correlation time τC is implicit, the spectral density of the motion correlation functions J(ω) are 734



Article

Figure 7. (a) 1H MAS NMR spectra obtained at a rate of 6.3 kHz from amorphous simvastatin at T (K): 293, 308, 316, 320, 326, and 333. The insets show the spectra obtained at 293 K, 316 and 333 K, and the Lorentzian curves used to deconvolute the central lines, respectively (dashed lines represent the line fitting residues). * denotes spinning side bands. (b) Top: simulation of the variation of fwhmMAS with τC according to eq 5 for ωr equal to 39584 rad·s−1 and M2, from bottom to top, equal to 1, 2, 3, 4, and 5 × 108 s−2 (solid line). Bottom: 1H MAS magnetization decay simulations using eq 4 and different τC values (1, 3, 5, and 7 μs, from top to bottom) and subsequent Fourier transform (τC of 1, 3, 5, and 7 μs, from bottom to top), for ωr and M2 equal to 39584 rad·s−1 and 5 × 108 s−2, respectively (see text for details).

chains of molecules along the b axis, as already mentioned; the general conformational features are closely related to those of other reported crystal structures of statins.5 The strong narrowing with temperature of the resonance at higher frequency under a MAS rate of only 6.3 kHz (Figure 7a) is considered to be due to a dominant “inhomogeneous” interaction being canceled. This observation agrees well with the proposed assignment considering that the spectral line shape from hydrogen bonded OH groups depends mainly on an inhomogeneous magnetic dipolar interaction. Moreover, it will be demonstrated below that the line width reduction is in agreement with the shortening of the correlation time τC with temperature. Estimating the activation energy Ea for proton mobility implies obtaining the variation of the correlation time τC with temperature. For this purpose, we have followed a methodology similar to that previously used to study the dynamics of amorphous gentiobiose hydroxyl groups.9 Theory about NMR line width variation with temperature under major interactions was reported on39 (a) inhomogeneous magnetic dipolar interaction and (b) chemical shift anisotropy. The following equation expresses the magnetization decay when inhomogeneous magnetic dipolar interaction is the main contribution and a unique τC describes mobility:39

amorphous simvastatin as a function of temperature was also probed using 1H MAS solid-state NMR. Figure 7a presents representative 1 H MAS spectra run from amorphous simvastatin under a spinning rate of 6.3 kHz in the temperature range 293 K to 333 K. The spectra show at least two components as well as spinning side bands. These facts demonstrate that there are several inequivalent proton sites. The most intense line is assigned to protons involved in anisotropic interactions such as dipolar couplings and chemical shift anisotropy not completely suppressed by MAS. Subsequent to baseline correction, the spectra obtained at 293, 316, and 333 K were deconvoluted using Lorentzian functions, and the following chemical shifts and full width at half-maximum (fwhm, within parentheses) were obtained: (a) 0.950 ppm (1230 Hz) and 3.802 ppm (1200 Hz), at 293 K, (b) 0.850 ppm (1515 Hz) and 3.840 ppm (690 Hz), at 316 K and (c) 1.100 ppm (1710 Hz) and 3.830 ppm (450 Hz), at 333 K. A tentative assignment of the resonances implies considering the eventual existence of water. Signals at about 5 ppm are generally assigned to free water molecules, but hydrogen bond breaking that occurs when the sample is heated up induces an upfield shift of the resonance, which it was not the present case. Other resonances, like methylene signals, could not be resolved under the selected experimental conditions (6.3 kHz MAS rate) because the dipolar coupling splitting (about 50 kHz) is very high due to the short interproton distance (0.17 nm). Therefore, the signal at higher frequency of resonance is assigned to the hydroxyl group. Eckert et al. reported on the relationship between the 1H chemical shift and the hydrogen bond strength:38

⎧⎛ ⎛ 2M ⎞⎢ F(2ωr , t ) ⎥⎞⎫ φMAS(t ) = exp⎨⎜ −⎜ 2 ⎟⎢F(ωr , t ) + ⎥⎟ ⎬ ⎦⎠ ⎭ 2 ⎩⎝ ⎝ 3 ⎠⎣ ⎪

⎪

⎪

⎪

(4) −2

where M2 (s ) is the second moment for the isotropic peak, ωr (rad·s−1) is the spinning rate of the sample, and

δiso(ppm) = 79.05 − 255d(O − H···O) (nm)

F(ωr , t ) = tτC[1 + (ωrτC)2 ]−2 − [1 − (ωrτC)2 ]τC 2

where δiso is an isotropic chemical shift and d(O−H···O) is the O−H···O distance. We have obtained 0.295 nm for the high frequency signal recorded at 293 K, in good agreement with intermolecular hydrogen bonding X-ray diffraction data involving the hydroxyl group and C(18)O.25 In fact, in simvastatin, which is almost isostructural with lovastatin, the only hydrogen bond present facilitates the formation of infinite

+ [1 + (ωrτC)2 ]−2 [1 − e−t / τC cos ωrt ] − 2ωrτC3[1 + (ωrτC)2 ]−2 e−t / τC sin ωrt

Equation 4 gives an exponential decay when the acquisition time is much longer than τC, the NMR line being then 735



Article

carbons in the ester group undergo a fast process, the mobility of methyl groups in the ring occurs at about 50 kHz at an intermediate rate. Moreover, at least other four local motional processes could be distinguished, which involve carbon atoms, and the corresponding activation energy was determined (27.3 ± 9.2, 18.1 ± 4.3, 16.1 ± 4.2, and 27.6 ± 5.7 kJ·mol−1). Furthermore, evidence was obtained for another process involving intermolecular hydrogen bonding (24 kJ·mol−1). Overall, the dynamic heterogeneity of amorphous simvastatin is considered here to contribute for the configurational entropy increase therefore being determinant in order to avoid recrystallization from the amorphous state. The present study demonstrated that crystalline simvastatin could be physically transformed to its amorphous form without chemical degradation by quench-cooling under inert atmosphere. This observation jointly with the lack of tendency to recrystallize exhibited by the amorphous form (information confirmed from the different techniques here employed) led to consider the glassy form a suitable candidate for potential application on the solid state formulations in the pharmaceutical area.

represented by a Lorentzian function. Under such conditions, eq 5 gives the line width at half-maximum (fwhm) obtained with MAS and motion (fwhmMAS):39 FWHMMAS =

⎤ ⎡ ⎤ ⎛ M 2 ⎞⎡ τC 2τC ⎜ ⎟⎢ ⎢ ⎥ ⎥ + ⎝ 3π ⎠⎣ 1 + (ωrτC)2 ⎦ ⎣ 1 + (ωrτC)2 ⎦ (5)

Equation 5 enables distinguishing the effect on fwhmMAS induced only by molecular mobility. With ωr 39584 rad·s−1 and M2 in the range 108 to 5 × 108 s−2 were obtained the corresponding fwhmMAS data using eq 5 (Figure 7b). The curves present a maximum as a function of τC and a decrease both at smaller (line narrowing caused by thermal motion) and at higher values of τC (influence of MAS) and also an increase with M2. Figure 7b shows 1H MAS spectra simulated using eq 5 with correlation times in the range 1 to 7 μs, that is, under line narrowing only controlled by molecular motion (see Figure 7b), the same value of ωr and M2 equal to 5 × 108 s−2, which, according to eq 4 and Figure 7b, gives fwhmMAS similar to the experimental data obtained at higher temperature (Figure 7a). The simulated magnetization decays for τC equal to 1, 3, 5, or 7 μs were obtained using eq 4, and subsequently to Fourier transforms, the fwhmMAS corresponding to each τC value were found: 288, 795, 1212, and 1518 Hz, respectively (Figure 7b). Considering again the curves that represent the variation of fwhmMAS with τC (Figure 7b), it may be concluded that the following equation applies for short τC values: FWHMMAS = constant·τC

■

AUTHOR INFORMATION

Corresponding Author

́ *Centro de Quimica Estrutural, Instituto Superior Técnico/ Univ. Lisboa, Av. Rovisco Pais, 1049-001 Lisboa, Portugal. Tel:. +351-21-8419043. Fax: +351-21-8464455. E-mail: teresa. [email protected].

(6)

Notes

The authors declare no competing financial interest.

■

In the present study the constant reflects mainly the resonance frequency difference between the C3(O−H··· OC18) and C18(O−H···OC3) sites. Using eq 6 and the fwhmMAS data of the simulated curves corresponding to different τC values it was possible to obtain the τC values related to the fwhmMAS of the resonances at about 3.8 ppm, which were found by deconvoluting the experimental central lines (Figure 7a): 1200 Hz at 293 K, 690 Hz at 316 K, and 450 Hz at 333 K are in agreement with τC changing from 5.2 μs to about 2.7 μs and then to 1.6 μs, respectively. These data enabled obtaining an approximate activation energy Ea for proton mobility using the Arrhenius relation, τC = τ0 exp(Ea/ RT), where R is the gas constant: Ea = 24 kJ·mol−1. In this context the mobility of hydroxyl hydrogen atoms undergoing secondary relaxations may explain the results presented here.

ACKNOWLEDGMENTS This work has been carried out with financial aid from “Fundaçaõ para Ciência e Tecnologia” (FCT, Portugal) through the projects RECI/QEQ-QIN/0189/2012, PEst-OE/ QUI/UI0100/2013, PEst-OE/CTM/LA0024/2013, PEst-C/ EQB/LA0006/2011 and PTDC/CTM/098979/2008. M.T.V. thanks FCT for the postdoc grant SFRH/BPD/39691/2007. Thanks are also due to FCT (“Infra-estruturas de C&T”) and IST for recently funding the upgrade of the solid-state NMR spectrometer.

■

REFERENCES

(1) Murtaza, G. Solubility enhancement of simvastatin: a review. Acta Pol. Pharm. 2012, 69, 581−90. (2) Laitinen, R.; Löbmann, K.; Strachan, C. J.; Grohganz, H.; Rades, T. Emerging trends in the stabilization of amorphous drugs. Int. J. Pharm. 2013, 453 (1), 65−79. (3) For example: Brittain, H. G.; Grant, D. J. W. Polymorphism in Pharmaceutical Solids; Marcel Dekker, Inc.: New York, 1999; pp 279− 330. (4) Harris, R. K. NMR studies of organic polymorphs & solvates. Analyst 2006, 131, 351−373. (5) Hušaḱ , M.; Kratochvíl, B.; Jegorov, A.; Brus, J.; Maixner, J.; Rohlíček, J. Simvastatin: structure solution of two new lowtemperature phases from synchrotron powder diffraction and ssNMR. Struct. Chem. 2010, 21, 511−518. (6) Ellison, D. K.; Moore, W. D.; Petts, C. R. Simvastatin. In Analytical Profiles of Drug Substances and Excipients; Brittain, H. G., Ed.; Academic Press: San Diego, 1993; Vol. 22, pp 359−388. (7) Patel, R.; Patel, M. Preparation, Characterization, and Dissolution Behavior of a Solid Dispersion of Simvastatin with Polyethylene Glycol 4000 and Polyvinylpyrrolidone K30. J. Dispersion Sci. Technol. 2008, 29, 193−204.

4. CONCLUSIONS It is generally accepted that the high mobility observed in the amorphous state is one of the main factors favoring crystallization. The DRS technique, which is based on the measurement of complex dielectric permittivity, allowed inferring about molecular mobility persisting in the glass form of simvastatin through the detection of secondary relaxations usually associated with localized motions (besides the main relaxation associated to the glass transition dynamics). However, experimental evidence for the high stability of the amorphous simvastatin, when obtained by the quench-cool method, demands another explanation to be found. For this purpose, we have used ss-NMR to perform a detailed local mobility study in the glass transformation region. This study shows that simvastatin undertakes local motions in different processes, in the kHz frequency scale. While methyl 736



Article

(8) For example: Aso, Y.; Yoshioka, S.; Kojima, S. Molecular Mobility-Based Estimation of the Crystallization Rates of Amorphous Nifedipine and Phenobarbital in Poly(vinylpyrrolidone) Solid Dispersions. J. Pharm. Sci. 2004, 93, 384−391. (9) Nunes, T. G.; Diogo, H. P.; Pinto, S.; Moura Ramos, J. J. Molecular dynamics of amorphous gentiobiose studied by solid-state NMR. J. Phys. Chem. B 2010, 114, 15976−84. (10) Graeser, K. A.; Strachan, C. J.; Patterson, J. E.; Gordon, K. C.; Rades, T. Physicochemical Properties and Stability of Two Differently Prepared Amorphous Forms of Simvastatin. Cryst. Growth Des. 2008, 8, 128−135. (11) Simões, R. G.; Diogo, H. P.; Dias, A.; Oliveira, M. C.; Cordeiro, C.; Bernardes, C. E. S.; Minas da Piedade, M. E. Thermal Stability of Simvastatin under Different Atmospheres. J. Pharm. Sci. 2014, 103, 241−248. (12) Hong, J.; Harbison, G. Sideband Elimination by Temporary Interruption of Chemical Shift. J. Magn. Reson., Ser. A 1993, 105, 128− 136. (13) Moura Ramos, J. J.; Taveira-Marques, R.; Diogo, H. P. Estimation of the fragility index of indomethacin by DSC using the heating and cooling rate dependency of the glass transition. J. Pharm. Sci. 2004, 93, 1503−1507. (14) Merino, E. G.; Rodrigues, C.; Teresa Viciosa, M.; Melo, C.; Sotomayor, J.; Dionísio, M.; Correia, N. T. Phase Transformations Undergone by Triton X-100 Probed by Differential Scanning Calorimetry and Dielectric Relaxation Spectroscopy. J. Phys. Chem. B 2011, 115, 12336−12347. (15) Souza, M. A. F.; Conceiçaõ , M. M.; Silva, M. C. D.; Soledade, L. E. B.; Souza, A. G. Thermal and kinetic study of statins. Simvastatin and lovastatin. J. Therm. Anal. Calorim. 2007, 87, 859−863. (16) Ambike, A. A.; Mahadik, K. R.; Paradkar, A. Spray-Dried Amorphous Solid Dispersions of Simvastatin, a Low Tg Drug: In Vitro and in Vivo Evaluations. Pharm. Res. 2005, 22, 990−998. (17) Simões, R. G.; Bernardes, C. E. S.; Diogo, H: P.; Agapito, F.; Minas da Piedade, M. E. Energetics and Structure of Simvastatin. Mol. Pharmaceutics 2013, 10, 2713−2722. (18) Viciosa, M. T.; Correia, N. T.; Salmerón Sanchez, M.; Gómez Ribelles, J. L.; Dionísio, M. Molecular Dynamics of Ethylene Glycol Dimethacrylate Glass Former: Influence of Different Crystallization Pathways. J. Phys. Chem. B 2009, 113, 14196−14208. (19) Brás, A. R.; Noronha, J. P.; Antunes, A. M. M.; Cardoso, M. M.; Schönhals, A.; Affouard, F.; Dionisio, M.; Correia, N. T. Molecular motions in amorphous ibuprofen as studied by broadband dielectric relaxation spectroscopy. J. Phys. Chem. B 2008, 112, 11087−11099. (20) Zhang, F.; Aaltonen, J.; Tian, F.; Saville, D. J.; Rades, T. Influence of particle size and preparation methods on the physical and chemical stability of amorphous simvastatin. Eur. J. Pharm. Biopharm. 2009, 71, 64−70. (21) Schönhals, A.; Kremer, F. Analysis of Dielectric Spectra. In Broadband Dielectric Spectroscopy; Kremer, F., Schönhals, A., Eds.; Springer Verlag: Berlin, 2003; Chapter 3. (22) Fröhlich, H. Theory of Dielectrics; Oxford Univ. Press: 1958. (23) Massalska-Arodz, M.; Williams, G.; Smith, I. K.; Conolly, C.; Aldridge, G. A.; Dabrowski, R. Molecular dynamics and crystallization behaviour of isopentyl cyanobiphenyl as studied by dielectric relaxation spectroscopy. J. Chem. Soc., Faraday Trans. 1998, 94, 387−394. (24) Brus, J.; Jegorov, A. Through-Bonds and Through-Space SolidState NMR Correlations at Natural Isotopic Abundance: Signal Assignment and Structural Study of Simvastatin. J. Phys. Chem. A 2004, 108, 3955−3964. (25) Xejka, J.; Kratochvíl, B.; Císaρová, I.; Jegorov, A. Simvastatin. Acta Crystallogr. 2003, C59, 428−430. (26) Zhang, P.; Klymachyov, A. N.; Brown, S.; Ellington, J. G.; Grandinetti, P. J. Solid-state NMR investigations of the glycosidic linkage in [alpha]-[alpha]’ trehalose. Solid State NMR 1998, 12, 221− 225.

(27) Rothwell, W.; Waugh, J. Transverse relaxation of dipolar coupled spin systems under rf irradiation: Detecting motions in solids. J. Chem. Phys. 1981, 74, 2721−2733. (28) Takegoshi, K.; K. Hikichi, K. Effects of blending on local chain dynamics and glass transition: Polystyrene/poly(vinyl methyl ether) blends as studied by high-resolution solid-state 13C NMR spectroscopy. J. Chem. Phys. 1991, 94, 3200−3207. (29) Cobo, M. F.; Malináková, K.; Reichert, D.; Saalwächter, K.; deAzevedo, E. R. Intermediate motions and dipolar couplings as studied by Lee−Goldburg cross-polarization NMR: Hartmann−Hahn matching profiles. Phys. Chem. Chem. Phys. 2009, 11, 7036−7047. (30) Graeser, K. A.; Patterson, J. E.; Zeitler, J. A.; Gordon, K. C.; Rades, T. Correlating thermodynamic and kinetic parameters with amorphous stability. Eur. J. Pharm. Sci. 2009, 37, 492−498. (31) Ambike, A. A.; Mahadik, K. R.; Paradkar, A. Physico-Chemical Characterization and Stability Study of Glassy Simvastatin. Drug Dev. Ind. Pharm. 2005, 31, 895−899. (32) Abragam, A. In The Principles of Magnetic Resonance; Oxford: New York, 1961; p 571. (33) Farrar, T. C.; Becker, E. D. Pulsed and Fourier Transfom NMR; Academic Press: New York, 1971. (34) Bloembergen, N.; Purcell, E. M.; Pound, R. V. Relaxation Effects in Nuclear Magnetic Resonance Absorption. Phys. Rev. 1948, 73, 679− 712. (35) Jones, G. P. Spin-Lattice Relaxation in the Rotating Frame: Weak-Collision Case. Phys. Rev. 1966, 148, 332−335. (36) Harris, R. K. Nuclear Magnetic Resonance Spectroscopy; Longman Scientific and Technical: New York, 1986. (37) Lim, A. R.; Kim, J.-S. Poly(methyl acrylate-co-sodium methacrylate) ionomer studied by solid state 13C T1ρ NMR. Solid State Commun. 2000, 115, 179−183. (38) Eckert, H.; Yesinowski, J. P.; Silver, L. A.; Stolper, E. M. Water in Silicate Glasses: Quantitation and Structural Studies by ’H Solid Echo and MAS-NMR Methods. J. Phys. Chem. 1988, 92, 2055−2064. (39) Fenzke, D.; Gerstein, B. C.; Pfeifer, H. Influence of thermal motion upon the line shape in magic-angle-spinning experiments. J. Magn. Reson. 1992, 98, 469−474.

737


Solid-State NMR and Dielectric Studies on Dynamic Heterogeneity of

Recommend Documents