The Slow Relaxation Dynamics in the Amorphous Pharmaceutical Drugs Cimetidine, Nizatidine, and Famotidine

Journal of Pharmaceutical Sciences xxx (2016) 1-12

Contents lists available at ScienceDirect

Journal of Pharmaceutical Sciences journal homepage: www.jpharmsci.org

Pharmaceutics, Drug Delivery and Pharmaceutical Technology

The Slow Relaxation Dynamics in the Amorphous Pharmaceutical Drugs Cimetidine, Nizatidine, and Famotidine M. Teresa Viciosa 1, Joaquim J. Moura Ramos 1, Hermínio P. Diogo 2, * 1 2

Centro de Química-Física Molecular and Institute of Nanoscience and Nanotechnology, Instituto Superior T
ecnico, Universidade de Lisboa, Lisboa, Portugal Centro de Química Estrutural, Complexo I, Instituto Superior T
ecnico, Universidade de Lisboa, Lisboa, Portugal

a r t i c l e i n f o

a b s t r a c t

Article history: Received 4 July 2016 Revised 29 July 2016 Accepted 15 August 2016

The slow molecular mobility in the amorphous solid state of 3 active pharmaceutical drugs (cimetidine, nizatidine, and famotidine) has been studied using differential scanning calorimetry and the 2 dielectric-related techniques of dielectric relaxation spectroscopy and thermally stimulated depolarization currents. The glass-forming ability, the glass stability, and the tendency for crystallization from the equilibrium melt were investigated by differential scanning calorimetry, which also provided the characterization of the main relaxation of the 3 glass formers. The chemical instability of famotidine at the melting temperature and above it prevented the preparation of the amorphous for dielectric studies. In contrast, for cimetidine and nizatidine, the dielectric study yielded the main kinetic features of the a relaxation and of the secondary relaxations. According to the obtained results, nizatidine displays the higher fragility index of the 3 studied glass-forming drugs. The thermally stimulated depolarization current technique has proved useful to identify the JoharieGoldstein relaxation and to measure tbJG in the amorphous solid state, that is, in a frequency range which is not easily accessible by dielectric relaxation spectroscopy. © 2016 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.

Keywords: amorphous pharmaceuticals cimetidine nizatidine famotidine dynamic fragility DSC DRS TSDC

Introduction The vast majority of modern active pharmaceutical ingredients (APIs) are made up of complex molecules with very low water solubility in their most stable crystalline state. The fact that the amorphous solid drugs are in a higher energy state compared to their crystalline counterparts gives them a higher solubility and a faster dissolution rate, which leads to a higher bioavailability. It is thus in the best interest to take advantage of this behavior in the formulation of poorly water-soluble drugs.1-5 The use of APIs in the amorphous solid state is however limited by the glass instability (tendency to crystallization). The stability of the amorphous solid form of a drug greatly varies from substance to substance. Furthermore, the stability of an amorphous solid depends on the amorphization method (melt quenching, freeze drying, ball milling, cryomilling).6,7 A close relationship seems to exist between molecular mobility and chemical and physical stability of the amorphous.8,9 In fact, the cooperative mobility was found to be responsible for instability,9,10 and correlations were found * Correspondence to: Hermínio P. Diogo (Telephone: 351-218419232; Fax: 351218464457). E-mail address: [email protected] (H.P. Diogo).

between the relaxation time of the a relaxation and the crystallization onset time in amorphous pharmaceuticals.11 It was also shown that the JoharieGoldstein relaxation can facilitate the main mobility even if it is not directly the cause of the amorphous instability.8 The importance of the local mobility in determining the chemical stability of macromolecules was highlighted,12-15 and there is also slight evidence about the destabilizing role of the fast secondary motions.16 These fast secondary motions probably also contribute to instability in situations such as the crystallization of amorphous indomethacin and celecoxib, which may occur at several tens of degrees below the glass transition temperature, Tg, conditions where molecular diffusion is too restricted for nucleation and crystal growth to take place.17,18 Some recent works are valuable contributions to the understanding of the relationship between molecular mobility, glass-forming ability, and glass stability.19,20 When the amorphous drug is unstable, stabilization of the glassy form is important to take advantage of the solubility and bioavailability benefits of the amorphous state. Stabilizing an amorphous API is achieved by obtaining a glassy mixture of the API in a polymer matrix (the amorphous solid dispersion)21-24 or through the so-called coamorphous mixtures.25-27 In any event, understanding the molecular mobility in the pure API and in the

http://dx.doi.org/10.1016/j.xphs.2016.08.019 0022-3549/© 2016 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.

2

M.T. Viciosa et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-12

mixtures is important for finding the best stabilization method and for designing the most appropriate mixtures. This work is a study by differential scanning calorimetry (DSC), thermally stimulated depolarization currents (TSDC), and dielectric relaxation spectroscopy (DRS) of the slow molecular mobility in the amorphous solid state of 3 APIs: cimetidine,28 nizatidine,29 and famotidine.30 These are H2 antagonists or H2 blockers, a class of medications that block the action of histamine at the histamine H2 receptors of the parietal cells in the stomach, decreasing the production of stomach acid. They are also used to treat peptic ulcer disease, gastroesophageal reflux disease, and hypersecretory syndromes such as the ZollingereEllison syndrome. In addition, these drugs are currently being considered for alternative indications. We will use DSC first to characterize the thermal behavior: glass transition, melting, crystallization, glass-forming ability, glass stability, and tendency for crystallization on cooling from the equilibrium melt. Then, we will use DSC to characterize the structural relaxation of the 3 drugs. With TSDC and DRS, we will analyze the slow molecular mobility in the amorphous solid form of cimetidine and nizatidine; the decomposition of famotidine that occurs very close to the melting point makes difficult the production of the glass by quench cooling from the melt, which prevented the dielectric study of this substance. The 2 dielectric techniques provide quantitative and complementary information on the distribution of relaxation times of the different slow motional modes. In fact, although the low equivalent frequency of TSDC leads to an enhancement of the resolution power of the different relaxations, the very wide frequency range of DRS constitutes an enormous advantage of this experimental technique. DRS is unanimously considered as a powerful technique for the study of molecular mobility in glassforming substances. On the other hand, an important advantage of TSDC is the possibility, using the partial polarization (PP) procedure (see Experiments section), to experimentally resolve a broadly distributed relaxation process into its narrowly distributed components; this enables the calculation of the temperaturedependent relaxation time, t(T), associated to individual or very narrowly distributed motional modes. These are reasons, among others, that make TSDC and DRS very useful complementary tools for slow molecular dynamic studies. Experiments

Figure 1. The chemical structures of cimetidine, nizatidine, and famotidine.

Materials Cimetidine, empirical formula C10H16N6S, CAS number: 51481-61-9, molar weight Mw ¼ 252.34 g/mol, was purchased from TCI (purity > 99%); the melting temperature, taken as the endothermic peak’s maximum, was found to be Tfus ¼ 143.1 C, in agreement with published values in the literature,31-35 and the melting enthalpy was determined as DHfus ¼ 40.4 kJ/mol, compared to the published value of DHfus ¼ 35.2 kJ/mol.33 Nizatidine, empirical formula C12H21N5O2S2, CAS number: 76963-41-2, molar weight Mw ¼ 331.46 g/mol, was purchased from TCI (purity > 97%); the melting temperature was found to be Tfus ¼ 132.5 C, in agreement with published values in the literature,31,36,37 and the melting enthalpy was determined as DHfus ¼ 36.9 kJ/mol, compared to the published value of DHfus ¼ 45 kJ/mol.36 Famotidine, empirical formula C8H15N7O2S3, CAS number: 76824-35-6, molar weight Mw ¼ 337.45 g/mol, was purchased from TCI (purity > 98%); the melting temperature was found to be Tfus ¼ 165.5 C, in agreement with values published in the literature,31,37-39 and the melting enthalpy was determined as DHfus ¼ 47.4 kJ/mol, compared to the published values DHfus ¼ 43.5,40 49.9,39 50.6,41 51.3,42 and 55.4 kJ/mol.38

These substances were used without further purification. The chemical structures are shown in Figure 1. Techniques Differential Scanning Calorimetry The calorimetric measurements were performed with a 2920 MDSC system from TA Instruments Inc. The samples of ~5-10 mg were introduced in aluminum pans. The measuring cell was continuously purged with high purity helium gas at 30 mL/min. An empty aluminum pan, identical to that used for the sample, was used as the reference. Details of the calibration procedures are given elsewhere.43 Thermally Stimulated Depolarization Currents TSDC experiments were carried out with a TSC/RMA spectrometer (TherMold, Stamford, CT) covering the range from 170 C to 400 C. For TSDC measurements, the sample (thickness of ~0.5 mm) was placed between the disc-shaped electrodes (7 mm diameter) of a parallel plane capacitor and immersed in an atmosphere of high purity helium (1.1 bar).

M.T. Viciosa et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-12

The samples of cimetidine and nizatidine in the amorphous solid state were prepared using 2 different procedures. (1) A tablet of the crystalline as received powder was prepared in the form of a compressed disc of ~0.5 mm thickness, shaped under the pressure of ~735 MPa, and was dried in a vacuum oven for 2 h at a temperature of ~(Tfus 20) under reduced pressure. After cooling to room temperature, the tablet was placed in contact with the electrodes inside the machine chamber under helium atmosphere. The sample was then heated to 10 degrees above the melting temperature, left isothermally for 5 min, and the amorphous state was obtained by fast cooling (~20 C/min) down to temperatures well below Tg. (2) A given amount of the crystalline sample is placed on one of the electrodes, all placed in a vacuum oven, submitted to reduced pressure and heated to ~(Tfus þ10). The assembly electrode plus melted sample is then cooled under vacuum to room temperature, transferred to the TSDC sample chamber in contact with the second electrode, and placed under helium atmosphere. Note that, because the cooling in vacuum oven is necessarily slow, this procedure wherein the amorphous sample is prepared outside the TSDC chamber is only achievable if the test substance exhibits a high resistance to crystallization, which is the case for cimetidine and nizatidine. However, famotidine decomposes at temperatures very close to the melting temperature, making it difficult to obtain the amorphous solid state and thereby preventing the TSDC study of its mobility. TSDC has a low equivalent frequency (~2  103 Hz), so that, it has a high resolution power and is sensitive to slow molecular motions (1-3000 s). Furthermore, the PP experimental procedure (see later discussion) allows probing narrow regions of the TSDC spectrum, that is, narrowly distributed motional modes. The fact that the relaxation time of the motional processes is temperature dependent and becomes longer as temperature decreases allows to make it exceedingly long (freezing process) compared to the timescale of the experiment. As will be seen next, this is the very core of the TSDC technique, which relies on the possibility of producing stable electrets at low temperatures by cooling down to those temperatures in the presence of a polarizing electric field; a clear and concise explanation of the experimental procedures provided by the techniques of thermally stimulated currents is available44 and may be useful for the reader unfamiliar with this technique. Two important parameters in a TSDC experiment are the polarization temperature, TP, at which the polarizing electric field is turned on, and the temperature TP’ < TP at which the field is turned off (see Fig. 2).44 The difference TP  TP’ is the width of the polarization window of the experiment. If it is wide, the retained polarization (and of course the current peak, i.e., the result of a TSDC experiment) will correspond to a complex set of energy-distributed motional modes. In contrast, the PP experiment where the polarizing field is applied in a narrow temperature interval allows probing more narrowly distributed relaxation modes. In the conceptual limit of a very narrow polarization window, the experimental depolarization current peak is supposed to correspond to a single mode of relaxation.45 In the present work and in most of our previous studies, we used polarization windows 2 degrees wide (DT ¼ Tp  Tp’ ¼ 2 C), and we tacitly assume that this window isolates single modes of motion. This assumption is based on the observation that similar PP experiments with polarization windows of 0.5, 1, or 2 degrees lead essentially to the same results. The temperature-dependent relaxation time, t(T), associated to each PP peak is obtained by a standard treatment explained in Section 2 of the study by Diogo et al.44 To summarize very briefly the main aspects of this, we will say that the analysis of the PP peaks is based on the Debye relaxation concept where the assumption is that, at each temperature of the linear heating ramp,

3

Figure 2. Schematic diagram of the experimental procedure for a TSDC experiment. The width of the polarizing window is DT ¼ TP  T'P, and it is typically between 0 and 5 degrees in a narrow window PP experiment. The electric field is on in steps 1 and 2 (thicker lines), and the depolarization current is recorded during the constant rate heating process (step 6).

the decay of the polarization with time is a first-order rate process. For a single motional process, we can thus write

dPðTÞ PðTÞ ¼ dt tðTÞ

(1)

where P(T) ¼ P(t) is the remaining polarization at temperature T of the heating ramp (at time t, such that T ¼ T0 þ r  t, where T0 is the temperature at the beginning of the heating ramp [at t ¼ 0, Fig. 2], and r is the heating rate), and t(T) is a temperature-dependent relaxation time, characteristic of the mode of motion under consideration. Because the depolarization current density [current intensity per unit area, J(T)] is the rate of decreasing of the polarization, it comes out that

JðTÞ ¼

PðTÞ tðTÞ

(2)

As reported before, an important feature of the TSDC technique is that it allows the study of single modes of motion using the so-called thermal sampling (PP) procedure. The importance of Equation 2 is that it allows the calculation of the temperaturedependent relaxation time of a motional mode from the direct experimental result of the corresponding PP experiment.44 The physical foundations of the TSDC experimental technique are presented in classical publications,46-48 whereas more recent review articles indicate various applications.49-53

Dielectric Relaxation Spectroscopy The complex dielectric function ε*(f) ¼ ε0 (f)  iε00 (f) (f, frequency; ε0 , real part; ε00 , imaginary part) associated with reorientational motions of dipoles was measured by an Alpha-N impedance analyzer from Novocontrol Technologies GmbH, covering a frequency range from 101 Hz to 1 MHz. Approximately, 10 mg of the as-received nizatidine crystalline powder was slightly compressed between 2 gold-plated electrodes (upper electrode 10 mm diameter) of a parallel plate capacitor, with 2 silica spacers of 50 mm thickness. The silica spacers were used to avoid contact between the disposable electrodes when the crystalline sample melts and to ensure, on the following measurements, a constant distance between the 2 electrodes, that is,

4

M.T. Viciosa et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-12

constant sample geometry (capacitance of empty capacitor C0 ¼ 13.91 pF). The sample capacitor was inserted in BDS 1200 sample holder, mounted on a cryostat BDS 1100, and exposed to a heated nitrogen gas stream being evaporated from a liquid nitrogen Dewar. The temperature of the sample was controlled by a Quatro Cryosystem with temperature stability better than 0.5 C. Novocontrol GmbH supplied all these modules. After the first temperature scanning leading to melt, the contact between all elements was checked, and the top electrode of the sample holder was screwed again allowing fixing the sample thickness to 50 mm. The sample was heated at 140 C, a temperature slightly above Tfus of nizatidine and kept 5 min at this temperature to ensure complete melting. Then, it was cooled down to 120 C at around 7.6 C/min. The complex dielectric permittivity, ε*(T), was recorded upon cooling at 5 selected frequencies (102, 103, 104, 105, and 106 Hz). Next, isothermal dielectric spectra were collected between 120 C and 150 C at different increasing temperature steps: from 120 C to 60 C in steps of 2 C and in the remaining temperature range, every 5 C. Dielectric Data Analysis To analyze the isothermal dielectric data, the model function introduced by HavriliakeNegami54 was fitted to imaginary component of ε*(u). Because multiple peaks are observed in the available frequency window, a sum of HavriliakeNegami (HN) functions was used:

ε* ðuÞ ¼ ε∞ þ

X j





Dεj

1 þ iutHN j

aHN j bHN j

(3)

where j is the index over which the relaxation processes are summed, Dε ¼ εs  ε∞ is the dielectric strength, that is, the difference between the real permittivity values at, respectively, the low and high frequency limits, tHN is the characteristic HN relaxation time, and aHN and bHN are fractional parameters (0 < aHN < 1 and 0 < aHN bHN < 1) describing, respectively, the symmetric and asymmetric broadening of the complex dielectric function. From the estimated values of tHN, aHN, and bHN parameters, a model-independent relaxation time, t ¼ (2pfmax)1, was calculated according to55

  31=aHN HN bHN p sin a2þ2b HN 5 ¼ tHN 4  aHN p sin 2þ2b 2

tmax

(4)

HN

A term is/ucε0 was added to the previous equation to take into account the direct current (dc) conductivity contribution in the low frequency side; in this term, ε0 is the vacuum permittivity and s and c are fitting parameters (s is related to the dc conductivity of the sample and c describes the broadening of the relaxation time distribution for the dc conductivity). Because of the marked influence of conductivity even at relatively low temperatures, imaginary part of permittivity, ε00 (u), was determined from the frequency dependence of the derivative of the real part of the complex dielectric constant, ε' (u), according to the following56,57:

p vε0 ðuÞ ε deriv z  2 vlnu 00

versus temperature at constant frequencies, ε00 (T, f ¼ const). A Gaussian function was used to determine the temperature of the maximum for each frequency. Results and Discussion General Aspects of the Thermal Behavior The amorphous solid of cimetidine and nizatidine were easily obtained by cooling from the equilibrium melt; no tendency to crystallize on cooling the liquid was detected, so that these glass formers display a high glass-forming ability.58 On the other hand, the thermograms of successive cooling/heating cycles showed no signals of crystallization on heating above Tg (cold crystallization), and this was observed for cooling/heating rates between 2 C/min and 20 C/min. Moreover, the value of DCp at Tg at the end of each batch of successive cycles (which lasted for about 10 h) is equal to the value at 0 time, which indicates that the amount of amorphous remains constant. We can therefore conclude, in line with other authors,31,36,37 that, in addition to their high glass-forming ability, cimetidine and nizatidine have also high glass stability. The calorimetric glass transition temperature of cimetidine and nizatidine, here taken as the extrapolated onset temperature obtained on heating at 10 C/min, was determined respectively as Tgcal ¼ 35.9 C and 7.3 C, and the amplitude of the heat capacity step was found to be DCp ¼ 0.65 and 0.51 (J K-1 g-1), respectively (Table 1). The thermal behavior of famotidine is reported to be very different from that described previosuly.31 In fact, as noted previously, we found that famotidine decomposes at temperatures above but very close to the melting temperature, or even during melting, which probably explains the high dispersion of the melting enthalpy values published in the literature (Materials section). Moreover, this behavior makes it more difficult to obtain the amorphous solid state by fast cooling from the equilibrium melt. The strategy for obtaining the amorphous famotidine for DSC measurements consisted of heating the crystalline sample to temperatures T such that 163 C < T < 170 C, in the melting region (recall that Tfus ¼ 166 C), and then cool immediately and rapidly to below Tg. The objective was to find the temperature T that maximizes the amorphization process and minimizes decomposition. This strategy has been successful so that the calorimetric glass transition temperature of famotidine was determined as Tgcal ¼ 40.5 C (onset at 10 C/min), and the amplitude of the heat capacity step was DCp ¼ 0.52 (J K-1 g-1); it should be noted that the only Tg value published in the literature for famotidine (74.5 C37; Table 1) is very different from ours because the experimental protocol for Tg determination described in the study by Pajula et al.37 would certainly lead to sample decomposition. Our reported values for famotidine refer indeed to the fully amorphous sample because a final heating ramp up to above the melting temperature revealed no evidence of endothermic melting process. Furthermore, we noticed that once avoided decomposition, famotidine displays, as cimetidine and nizatidine, a high glass-forming ability and a high glass stability. Dynamic Characterization of the Glass Transition by DSC

(5)

This equation is an alternative to the numerical KramerseKronig relations to obtain ε00 (u) that allows the suppression of the conductivity contribution. The analysis of dielectric data was extended by the isochronal representation of the dielectric loss, that is, ε00 deriv

The activation energy of the structural relaxation, Ea(Tg), can be determined from DSC data using the relationship59

  Ea Tg dðln qÞ ¼ dð1=Tx Þ R

(6)

M.T. Viciosa et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-12

5

Table 1 Thermodynamic and Dynamic Properties of Cimetidine, Nizatidine, and Famotidine Determined in the Present Work, Compared to Values Reported in the Literature Parameter

Cimetidine

Nizatidine

Famotidine

Tgcal/ Cb TgTSDC ¼ TM/ C TgDRS/ C DCp/J K1 g1 mDSC mTSDC mDRS

35.9a 31a,c 4370 0.65 ± 0.02a,e; 0.6997 45a,h; 65a,i 54a,c 6970

7.3a; 1336; 10.369; 11.3037 3.6a,d 5.6a 0.51 ± 0.03a,f 62a,h; 56a,i 91a,d 84a

40.5a; 74.537 d d 0.52 ± 0.02a,g 43a,h; 68a,i d d

a b c d e f g h i

Determined in the present work. Extrapolated onset temperature at 10 C/min. Mean over 28 determinations. Mean over 18 determinations. Mean over 51 determinations. Mean over 71 determinations. Mean over 10 determinations. Estimated from the heating rate dependence of the overshoot peak temperature location, Tov. Estimated from the heating rate dependence of the extrapolated onset temperature, Ton.

where q is the heating or cooling rate of the DSC scan, Tx is a temperature defining the position of the glass transition signal, and R is the gas constant. The approach is thus based on the analysis of the effect of the heating or cooling rate on the temperature location of the DSC glass transition signal.60,61 However, different cooling/heating cycles corresponding to different temperatureetime histories are currently used, some of them inadequate to properly accede to the activation energy of the structural relaxation. The most frequently used methodologies have been recently reviewed,61-64 and their advantages and disadvantages have been discussed. Taking into account the conclusions of those analyses, we chose the following experimental procedure: (1) The temperature location of the glass transition step signal was determined on heating, given that the quality of the DSC signals is much better in this mode because of experimental difficulties of temperature calibration and control on cooling. (2) The protocol of cooling/heating cycles was such that each heating scan was immediately preceded by a cooling scan from the melt at the same rate; in such conditions, little relaxation is expected to occur on heating, and the TooleNarayanaswamyeMoynihan phenomenological model is obeyed.59,65,66 (3) The successive cooling/heating cycles were performed between 50 degrees below and 40 degrees above Tg to warrant a good definition of the baseline on both sides of the step signal. (4) The location of the glass transition step signal was defined in 2 different ways, one as the extrapolated onset temperature, Ton, and the other as the temperature of the endothermic overshoot peak, Tov. Either of these 2 temperatures can be determined with higher accuracy compared to the midpoint, the extrapolated end set, and the fictive temperatures. The obtained results are displayed on Figure 3 for cimetidine, nizatidine, and famotidine. The activation energies of the structural relaxation were determined from the slope of the regression lines using Equation 6. The dynamic fragilities, defined as67,68

m¼

  Ea Tg ðln 10Þ  RTg

(7)

were then calculated from those values. The obtained results are displayed on Table 1 and will be compared later with fragility values obtained by other techniques.

The Molecular Mobility Studied by TSDC and DRS The Main Relaxation and the Relaxation Map The tendency of famotidine for decomposition in the vicinity of Tfus prevented the study of the mobility of the drug in the amorphous solid state by the 2 dielectric techniques. The TSDC study of the molecular mobility in the amorphous cimetidine and nizatidine was performed using the usual procedure of partial polarization that allows the characterization of the narrowly distributed components of a broad and widely distributed relaxation. The PP experimental protocol was briefly accounted for in the Experiments section and explained in more detail in Section 1 of the study by Diogo et al.44 The experimental result of a PP experiment is a current peak, I(T), that corresponds, as was stated, to a narrow slice of the whole wide continuous distribution. Examples of such current peaks obtained on nizatidine are shown in Figures 4 and 8. (We do not present here similar results for cimetidine in order to not excessively lengthen this work.) Figure 4 shows a set of depolarization peaks that are the result of PP experiments performed in the temperature range of the glass

Figure 3. “Arrhenius plots” of the logarithm of the heating rate, qþ, as a function of 1000/ Tov (full symbols) and of 1000/Ton (empty symbols), where Tov is the temperature of the overshoot peak and Ton is the temperature of the extrapolated onset of the glass transition signature. The circles refer to cimetidine, the triangles to nizatidine, and the squares to famotidine (black). The experiments were designed in such a way that the ratio between the heating rate, qþ, and the previous cooling rate, q, was unity: q/qþ ¼ 1.

6

M.T. Viciosa et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-12

Figure 4. PP components of the glass transition relaxation of amorphous nizatidine. The polarization temperatures, Tp, varied between 4 and 8 C (dashed lines correspond to higher temperature components were the intensity decreases with increasing Tp). The insert shows the log10 t(T) versus 1000/T lines of the motional modes corresponding to the peaks displayed in the main figure. In all the experiments the width of the polarization window was DT ¼ 2  C, the polarizing electric field strength was E ¼ 300 V mm1, and the heating rate was qþ ¼ 4  C min1.

transformation of nizatidine. It can be seen that, among the PP peaks obtained in the glass transformation region, there is one that shows a higher intensity (red peak in the online edition, with maximum intensity IM ¼ 2.03  10-9 A at TM ¼ 3.6 C). The temperature of maximum intensity is in this case written as TM, to distinguish from the general designation of the temperature location, Tm, of the other unspecified PP peaks. This peak with higher current intensity, located at TM, shows the higher dielectric strength (bigger area) and contains the higher temperature (higher activation energy) modes of the glass transition relaxation. For these reasons, the temperature TM is regarded as the glass transition temperature provided by the TSDC technique.71,72 We found TgTSDC ¼ 31 C for cimetidine and TgTSDC ¼ 3.6 C for nizatidine (see Table 1). The insert of Figure 4 shows the log10 t(T) versus 1000/T lines of the motional modes corresponding to the peaks in the main, and it can be seen that these lines are nearly parallel, so that the activation energies of the corresponding motions, Ea(Tm), are not very different

Figure 5. TSDC relaxation map of cimetidine and nizatidine: activation energy, Ea(Tm), of the PP components of the mobility as a function of the temperature location, Tm, of the corresponding current peak. The circles correspond to cimetidine, and the triangles correspond to nizatidine. The black square is the point of coordinates (T, Ea) obtained by DRS at t ¼ 100 s for nizatidine. The dotted line is the zero entropy line.

from each other. Fitting the t(T) line associated to a given PP peak (i.e. corresponding to a narrowly distributed motional mode) to an appropriate equation (Arrhenius or Eyring for the secondary relaxations, Vogel or WLF for the highly cooperative motions) provides the kinetic parameters characterizing the mobility. Figure 5 shows the obtained kinetic information displayed in the form of a relaxation map designed as a representation of the activation energy, Ea(Tm), as a function of Tm, where Tm is the temperature of maximum intensity of a given PP peak. Let us recall that each point of coordinates [Tm, Ea(Tm)] on this relaxation map corresponds to a PP peak, that is, to a single or narrowly distributed mobility component. In Figure 5, circles refer to cimetidine and triangles to nizatidine. For cimetidine, the points (circles) in the temperature region from 25 C to 32 C correspond to motional modes of the a relaxation and display the common behavior of the glass transition of an amorphous solid as studied by TSDC, often referred as the compensation behavior: strong and progressive deviation of the points [Tm, Ea] of the relaxation map relative to the zero entropy line as the temperature increases in the glass transition temperature range. The term “compensation” reflects the fact that the strong and progressive increase in activation energy is concomitant with an equally strong and progressive increase of the activation entropy (increase of the deviation from the zero entropy line), such that the Gibbs energy of activation undergoes a slight increase as the temperature increases.73 It is often said in this sense that the glass transition of a fragile glass is characterized by a wide distribution of energies and entropies. Looking at another angle, we will say that the slope at Tm of the log t(T) versus 1000/T lines of the glass transition motional modes gradually increases as the temperature increases in the glass transition temperature range.74 Nizatidine has a slightly different behavior compared with that described previously for cimetidine. In fact, the TSDC results obtained in the temperature range of the a relaxation can be described as follows: (1) The log10 t(T) versus 1000/T lines in the insert of Figure 4 are roughly parallel and close to each other, such that no mobility modes are found with low activation energy, so that no points exist in Figure 5 lying between 400 kJ/mol and the zero entropy line. (2) The values of the activation energy calculated for the mobility modes of the glass transition are thus high (strong deviation from the zero entropy line), distributed between ~415 and ~515 kJ/mol as shown in Figure 5 (triangles in the temperature interval between 0 C and 10 C). (3) Because the amplitude of the deviation from the zero entropy line is a measure of the cooperativity of the molecular motions, we conclude that the mobility of the glass transition of nizatidine is markedly cooperative but displays a relatively narrow distribution of relaxation times. Nizatidine was also studied by DRS. As it was referred in the Experiments section, after cooling the melt down to 120 C, isothermal spectra were collected at increasing temperature steps. Figure 6a shows some representative spectra obtained between 10 C and 60 C, in the supercooled liquid range, where the conductivity contribution was suppressed by using the derivative analysis (Eq. 5). In this temperature region, the spectra are dominated by the a peak that corresponds to molecular motions associated to the dynamic glass transition. Furthermore, very close to this main peak, another one appears, partially submerged as an excess wing in its high frequency flank. At lowest temperatures, well below the calorimetric Tg, 2 secondary relaxations are also detected. The features of these secondary relaxations will be analyzed in the section The Secondary Relaxations. Figure 6b compares some ε00 (T) of the main relaxation in the isochronous representation, with a TSDC peak obtained with a large polarization window capable of activating the a relaxation in its entirety (in its entire distribution of relaxation times). The

M.T. Viciosa et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-12

7

Figure 6. (a) Imaginary part of dielectric permittivity of nizatidine versus frequency in the metastable liquid state from 10 C to 60 C, in steps of 4 degrees. (b) Left axis: isochronal 00 representation of permittivity (ε deriv ) at the frequencies of 10 ( ), 102 (D), 103 (,), and 104 (B) Hz. Right axis: TSDC thermogram obtained from a wide polarization window experiment with polarization temperature, Tp ¼ 5 C, and freezing temperature, T0 ¼ 125 C. The other relevant experimental parameters are strength of the polarizing electric field, Ep ¼ 350 V/mm; polarization time, tp ¼ 5 min; and heating rate, qþ ¼ 8 C/min.

maximum of the ε00 peaks shifts to lower temperatures with decreasing frequency, and the temperature position of the TSDC a peak is even lower as expected because the equivalent frequency for this technique is close to 103 Hz. The a peak is characterized by the HN shape parameters aHN increasing from 0.84 at 16 C to 1 at 50 C and bHN ¼ 1. The less intense peak at higher frequencies is well fitted with aHN ¼ 0.56 ± 0.01 and bHN ¼ 1. The dielectric strength (Dε) for the a relaxation slightly decreases from 56.0 to 54.4 in the analyzed temperature range, whereas for the secondary process, it increases from 0.38 to 1.28 from 2 C to 32 C. The most important parameter that is extracted from the analysis of isothermal dielectric spectra is the relaxation time. Figure 7 shows the temperature dependence of the relaxation time for all relaxation processes detected by DRS in nizatidine. For the a relaxation, the corresponding values of t(T) obtained from the isochronal plots (t ¼ 1/(2pf), 1/Tmax) are also displayed. The 2 data sets coincide well, and the inclusion of this second method of analysis allows enlarging the temperature range where the a process is characterized. In this figure, relaxation times obtained from

Figure 7. DRS relaxation map of nizatidine: decimal logarithm of the relaxation time, log10 (t/s) versus 1000/T for the a relaxation (,), the bJG ( ), the b (B), and g (D) secondary relaxations. The open and filled symbols correspond respectively to the isothermal and isochronal analysis. Lines are fits of the corresponding data to the Arrhenius or Vogel equations. Vertical arrow indicates the glass transition temperature. The star indicates the relaxation time of the bJG relaxation at Tg estimated according to the coupling model. The dashed rectangle demarcates the temperature region where ε00 derivative was analyzed.

TSDC measurements were also included for comparison (see the upper part of the figure). The temperature dependence of the relaxation time for the a relaxation, ta(T), was found to be well described close to Tg by the empirical Vogel equation75-77:

 ta ðTÞ ¼ t∞ exp

 B ; T  T0

(8)

where t∞, B, and T0 are empirical parameters: t∞ is the limit of the relaxation time at high temperature (corresponding to a vibrational frequency of ~1011-1014 Hz) and T0 is the so-called Vogel temperature whose value is found to be about 50-70 degrees below Tg. The fit of the experimental points with the Vogel equation lead to the following values of the parameters: t∞ ¼ 4.4  1012 s, B ¼ 1358 K, and T0 ¼ 235.6 K. Based on these parameters, we can predict that the temperature of the maximum of the ε00 (T) peak at 103 Hz would be 4.9 C, coincident with the temperature of maximum intensity of the TSDC a peak (Tm ¼ 5.0 C) shown in Figure 6b. On the other hand, in the context of DRS, the glass transition temperature is conventionally defined as the temperature at which ta(T) ¼ 100 s,

Figure 8. PP components of the secondary relaxation of amorphous nizatidine. In the main figure, the polarization temperatures, Tp, of the experiments varied between 20 C and 12 C (with intervals of 2 degrees) with a polarizing electric field strength of E ¼ 300 V/mm. In the inset, Tp varied from 34 C to 22 C (also with intervals of 2 degrees), and the polarizing electric field strength was E ¼ 400 V/mm. The heating rate was qþ ¼ 4 C/min for all the experiments.

8

M.T. Viciosa et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-12

and according to this criterion, the value TgDRS ¼ 5.6 C was obtained, in good agreement with the previously reported TgDSC and TgTSDC (Table 1). The Dynamic Fragility or Steepness Index The methodology proposed to determine the dynamic fragility from TSDC data refers to the PP peak in the glass transition region with maximum intensity at TM ¼ TgTSDC. From an extensive experimental work, we have obtained a high number of those peaks, whose analysis draws the following conclusions: (1) The log t(T) versus 1000/T line associated to each one of those peaks44 was determined and fitted to a Vogel type equation. The values obtained for the relaxation time at TM and for the activation energy at TM are t(TM) ¼ 39 s and Ea(TM) ¼ 316 kJ/mol for cimetidine and t(TM) ¼ 20 s and Ea(TM) ¼ 474 kJ/mol for nizatidine. (2) The dynamic fragility was calculated at TgTSDC ¼ TM using Equation 7, and it was found to be mTSDC ¼ 54 for cimetidine and 91 for nizatidine (Table 1). The activation energy of nizatidine estimated at Tg from DRS data, Ea(Tg), was calculated as

  RBTg2   vln tðTÞ Ea Tg ¼ R ¼ 2 v1=T Tg Tg  T0

(9)

using the parameters of the Vogel equation previously reported. The value Ea(TgDRS) ¼ 450 kJ/mol was obtained, and the dynamic fragility was then calculated using Equation 7, leading to mDRS ¼ 84; the point of coordinates TgDRS ¼ 5.6 C and Ea(TgDRS) ¼ 450 kJ/mol (corresponding to t ¼ 100 s) is displayed as a black square in Figure 5. For cimetidine, the value mDRS ¼ 69 is reported in a brief published study70 (Table 1), so that nizatidine clearly appears as a fragile liquid, whereas cimetidine displays a moderately fragile behavior. Let us note the good mutual agreement between the values of the glass transition temperature, TgTSDC and TgDRS, and of the dynamic fragility, mTSDC and mDRS, obtained by the 2 dielectric techniques. This reinforces the idea of a close complementarity between the 2 techniques in the study of molecular mobility associated with the glass transition. Let us remember that, for the study of the dynamics of the glass transition, the TSDC technique operates directly in the amorphous solid state, whereas DRS operates in the supercooled liquid and can only access the dynamics at Tg by extrapolation. Let us note also that the DSC procedure used to determine the fragility index does not distinguish these substances in view of their strong or fragile behavior (see Differential Scanning Calorimetry section and Table 1). The Secondary Relaxations Low-intensity secondary relaxations were observed by TSDC in the glassy state of cimetidine and nizatidine, which were activated by electric fields applied at temperatures above 15 C and 50 C, respectively, and merged with the glass transition relaxation at higher temperatures. The points in the relaxation map of Figure 5 that are close to the zero entropy line correspond to partial polarization peaks that are the manifestation of this mobility. In the case of nizatidine, the observed secondary mobility extends from 50 C up to 10 C with activation energies distributed between 60 and 70 kJ/mol. Figure 8 shows TSDC PP modes of the secondary relaxation of nizatidine. The coalescence with the main relaxation is evident. On the other hand, the dielectric loss spectra, ε00 (f), of nizatidine obtained in the glassy state show clearly 2 secondary relaxations as it can be seen in Figure 9 where some representative spectra are included. A sum of 2 HN functions was fitted to the isothermal spectra of ε00 (f) to characterize these processes which have been named g and

Figure 9. Imaginary part of the dielectric permittivity of nizatidine versus frequency in the glassy state from 100 C to 40 C, in steps of 10 degrees, where the g and b secondary processes are evidenced. The arrows indicate the evolution of the maxima of ε00 peaks of the b and g relaxations with increasing temperature. The inset shows the isothermal spectra at 16 C in the logelog scale. The solid line represents the overall fit obtained by the sum of 3 HN functions; the dashed lines correspond to the fitting of the individual a, bJG, and b processes. The full circles refer to the ε00 raw data, and open circles to the ε00 derivative (Eq. 5).

b in order of increasing temperature (decreasing frequency). Reliable fitting parameters are obtained in the temperature range from 106 C to 50 C for the g relaxation and from 70 C to 16 C for the b relaxation. The aHN and bHN shape parameters obtained are summarized in Table 2. The dielectric strength, Dε, slightly increases with temperature from 0.15 to 0.23 for the g relaxation, whereas it decreases from 0.52 to 0.2 for the b process. The relaxation times obtained from the fitting, tHN, have been converted to the model-independent relaxation times, tmax, which have been included in Figure 7 (triangles, g, and circles, b). The temperature-dependent relaxation times, t(T), of these processes, which are obtained in the glassy state, display Arrhenius temperature dependence:   Ea tðTÞ ¼ t0 exp RT

(10)

where t0 is the preexponential factor, Ea is the activation energy, and R is the ideal gas constant. The Arrhenius parameters obtained for the g relaxation were Eag ¼ 36 kJ/mol and t0g ¼ 7.2  1016 s and, for the b relaxation, Eab ¼ 53 kJ/mol and t0b ¼ 9.8  1016 s (Table 2). For the b process, the relaxation times have been also estimated from the isochronal analysis, and these results are included in Figure 7 as filled circles. The clear separation of both secondary processes relative to the a relaxation, visible in the permittivity spectra and in the relaxation map, suggests that the underlying mechanisms are independent of the cooperative motions associated to the dynamical glass transition. They probably correspond to very localized internal motions, so that the activation of the cooperative motions approaching the glass transition temperature region is not expected to lead to significant changes of the kinetics of these secondary relaxations. The previous results show that the isochronal analysis in the temperature range between 106 C and 60 C is suitable to characterize the different kinds of mobility present in amorphous nizatidine. A relaxation which appears partially merged in the high frequency flank of the a process is shown in the inset of Figure 9,

M.T. Viciosa et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-12

9

Table 2 HavriliakeNegami and Arrhenius Parameters for the Secondary Detected Process in Amorphous Nizatidine Parameter

Temperature Range ( C)

aHN

bHN

t0 (s)

Ea (kJ/mol)

g b bJG

106, 50 70, 16 2, 30

0.63 ± 0.04 0.24 ± 0.01 0.56 ± 0.01

0.7 / 1 0.73 ± 0.04 1

7.2  1016 9.8  1016 ~1045

36 53 230

00

where the permittivity spectrum (ε deriv ) taken at 16 C is represented. In this insert, the solid line corresponds to the global fitting curve resulting of the sum of 3 individual HN functions also included as dashed lines. The fitting parameters were obtained in the temperature range from 2 C to 30 C, and the HN and Arrhenius parameters are also displayed on Table 2. According to the coupling model (CM),78 the relaxation time of the JoharieGoldstein relaxation, tJG, is correlated with that of the a process, ta, by

tJG ðTÞzt0 ðTÞ ¼ tcn ½ta ðTÞ1n

(11)

where t0 is the relaxation time of the primitive process postulated in the CM, tc is a time characterizing the crossover from localized to cooperative fluctuations found to be close to 2  1012 s for molecular glass formers,79 and n is the coupling parameter, directly related to the nonexponentiality of the relaxation in the glass transition region. A common way of obtaining the coupling parameter is based on the analysis of the permittivity spectra obtained close to Tg not in the frequency domain (as done with the HN equation) but in the time domain. In this case, the well-known Kohlrausch-Williams-Watts (KWW) function80,81 is used:

" 4ðtÞ ¼ exp 



t tKWW

bKWW # (12)

where tKWW is a characteristic relaxation time and bKWW is the stretching parameter (0 < bKWW  1) related to the coupling parameter by n ¼ 1  bKWW. The isothermal spectrum of nizatidine collected at 16 C, where the a peak is visible, was fitted by the 1-side Fourier transforms of the KWW, giving the value of bKWW ¼ 0.83, and hence n ¼ 1  bKWW ¼ 0.17, which indicates a considerable degree of nonexponentiality. From Equation 11 written at TgDRS (i.e., for t ¼ 100 s), the value of tJG(TgDRS) ¼ 0.47 s which is represented in the relaxation map of Figure 7 by a star. This result is in good agreement with the experimental value of the relaxation time of the secondary process under analysis, which constitutes an argument to identify it as a JoharieGoldstein b process. The kinetic parameters estimated for the different processes allow comparing results from the DRS and TSDC techniques. Consider in Figure 7 the line for the bJG (asterisks) obtained by DRS and extend up this line to the higher frequencies of the TSDC technique (102-103 Hz). It is easy to see that this extension leads precisely to the temperature range in which the secondary relaxation is detected in TSDC (see upper part of Fig. 7), which justifies the identification of it with the JoharieGoldstein process. This may cause some strangeness in that the kinetic parameters of secondary relaxation detected by TSDC (Fig. 5) are substantially different from those of the bJG observed by DRS (Table 2). However, it should be noted that the bJG process is detected in DRS at temperatures above Tg (in the metastable liquid), whereas in TSDC, it is observed in the glassy state and that the energy of activation EabJG suffers a marked variation across Tg.82 In fact, in DRS, it often happens that the relaxation times of the a and bJG relaxations, ta and tJG, are very close to each other near and above Tg, making it impossible to determine tJG without using an approximate procedure. The

favorable situation to allow accurate determination of tJG(T) is when the JG relaxation signal is well resolved and separated from that of the a-relaxation. The analysis of different systems in such case (including low-molecular-weight glass formers83 and their mixtures,84-86 amorphous polymers and copolymers,87,88 ionic liquids,89 and also studies under high pressure85,86,89) provides wide and strong experimental evidence showing that bJG persists above Tg and that the temperature dependence of the relaxation time tJG above Tg is much stronger than this dependence below Tg.82 In conclusion from what has been argued, we can say that the secondary relaxation observed by TSDC in nizatidine, in the temperature range between 10 C and 50 C and obeying the zero entropy line, corresponds to the bJG process. The technique of TSDC thus appears as a very useful technique for the analysis of the temperature dependence of tJG in the glassy state. However, in the case of cimetidine and nizatidine, the intensity of the TSDC signals of the secondary relaxations is low, which does not allow the analysis of the temperature dependence of tbJG at low temperatures and the possible occurrence of a secondary glass transition at TJG y 0.65Tg whose existence has been recently strengthened by increasing experimental evidence.90-92 The fast secondary relaxations could not be detected by TSDC. Indeed, the extension of the DRS line of log10 tg versus 1000/T (open triangles in Fig. 7) to t values in the TSDC time window leads to temperatures much lower than the lower limit of the temperature range accessible to the common TSDC equipment, which uses liquid nitrogen as the cold source. On the other hand, the similar extension of the line log10 tb versus 1000/T (open circles in Fig. 7) leads to temperature values in the vicinity of ~ 120 C. This is a low temperature not far from the lower temperature limit of our equipment, and in this temperature region, we were not able to detect depolarization peaks with appreciable intensities. The Secondary Relaxations and Aging The nature of the secondary relaxations observed by TSDC in both cimetidine and nizatidine was analyzed by looking at the influence of physical aging on the corresponding mobility modes, and Figure 10 displays clear results on this subject. The experiments had wide polarization windows, allowing exciting both the secondary mobility and a significant number of the main relaxation cooperative modes. Figure 10a refers to cimetidine and shows results of the same experiment carried out on amorphous samples with different degrees of aging: as expected,93 the peak intensity decreases as the aging degree increases indicating that, as the glass equilibrates, the total polarization created by the same electric field, during the same polarization time, decreases. The inset of Figure 10a, on the other hand, shows curves 1(fresh, nonaged sample) and 4 (sample aged during one hour) of the main figure, together with that obtained by subtracting 4 from 1 (labeled 5). Curve 5 is therefore the expression of the mobility lost as a result of 1 h of physical aging. Careful observation of curve 5 indicates that this loss of mobility occurs down to temperatures as low as 0 C, showing that some components of the secondary mobility of cimetidine detected by TSDC appear as sensitive to aging. Figure 10b refers to nizatidine and leads to similar conclusions. The higher intensity peak in the main figure (curve 1) is the result

10

M.T. Viciosa et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-12

Figure 10. Effect of physical aging on the molecular mobility in amorphous cimetidine (a) and nizatidine (b). The probe experiments were wide polarization window. In panel (a), the polarizing field was applied between Tp ¼ 23 C and Tp’ ¼ 50 C. Curve 1 is the result of an experiment carried out on the fresh, nonaged sample, whereas curves 2, 3, and 4 were obtained performing the same experiment on a sample previously aged at Tag ¼ Tp ¼ 23 C for 10, 30, and 60 min, respectively. The inset in panel (a) shows curves 1 and 4 and that obtained by subtracting curve 4 from 1 (labeled 5). In panel (b), the polarizing field was applied between Tp ¼ 12 C and Tp0 ¼ 50 C. Curve 1 is the result of an experiment carried out on the fresh, nonaged sample, whereas curve 2 was obtained performing the same experiment on a sample previously aged for 150 min at Tag ¼ Tp ¼ 12 C. Curve 3 is obtained by subtracting curve 2 from curve 1. In the inset, curve 1 is the same as in main (b) and curve 4 was obtained from 3 by shifting horizontally and multiplying by a factor so as to match the peak maxima with that of 1.

of an experiment carried out on the fresh, nonaged sample, whereas curve 2 was obtained performing the same experiment on an aged sample. Comparing the 2 curves shows that both primary and secondary mobilities are affected by physical aging and that the secondary modes are affected down to temperatures in the vicinity of 35 C. Curve 3, on the other hand, was obtained by subtracting curve 2 from curve 1 and represents the mobility lost as a result of aging; it corroborates the severe loss of mobility, even in the temperature range from 35 C to 10 C, which is the region where most of the secondary motional modes appear in the TSDC thermogram (Figs. 5 and 8). Furthermore, the inset with curve 1 and modified curve 3 (normalized to the same intensity and temperature location and labeled 4) shows that the low temperature mobility (secondary) suffered a stronger relative loss caused by physical aging compared to the main relaxation. In earlier work using TSDC technique, it was shown that the kinetic parameters of the secondary relaxations obey the zero entropy line94-96 (or, what is the same, have Arrhenius prefactors close to t0 ¼ 1013 s44,97). It seems moreover be noted that this secondary mobility unfolds in 2 different types of mobility whose kinetic parameters both obey the zero entropy line: a secondary mobility that is affected by physical aging and another that is not. It was suggested in this context that the first type of mobility is the slow b or JoharieGoldstein (bJG) process, whereas the second one corresponds to the fast secondary relaxations (b, g …).74,98-101 These results on the influence of physical aging on the secondary relaxation of nizatidine seen by TSDC give more strength to the previous suggestion that this is a JoharieGoldstein mobility. Because the TSDC secondary relaxation of cimetidine is also affected by aging (Fig. 10a), it is reasonable to assume that it has the same nature as that of nizatidine. TSDC is therefore confirmed as a good technique for the kinetic characterization of the bJG process below Tg, and this is important for understanding the physical stability of the amorphous state given that it is suspected that this mobility can be a source of glass instability. Indeed, there is increasing evidence that the bJG process is at the origin of physical instability below Tg.8,18,19,102-104 From the reported TSDC results, we have EabJG ¼ 68 kJ/mol for nizatidine and EabJG ¼ 74 kJ/mol for cimetidine (Fig. 5), with a prefactor t0 ¼ 1013 s in both cases. Introducing these kinetic parameters of bJG of nizatidine in the Arrhenius equation, and using TgTSDC ¼ 3.6 C ¼ 276.7 K, we obtain tJG(Tg) ¼ 0.68 s, in

reasonable agreement with the experimental DRS value of the relaxation time at Tg found for bJG [tJG(Tg) y 0.01 s] and with the value tJG(Tg) ¼ 0.47 s, previously reported and obtained from the CM. These results show a good compatibility between the results obtained by the 2 dielectric techniques.

Conclusions In this work DSC, TSDC, and DRS were complementarily used to analyze the molecular dynamics in the amorphous state of cimetidine, nizatidine, and famotidine. The thermal behavior of the 3 pharmaceuticals was characterized by DSC, and it was found that cimetidine and nizatidine have high glass-forming ability and high glass stability. On the other hand, because famotidine decomposes at temperatures above but very close to the melting temperature, obtaining the substance in the pure amorphous state is not an easy task. However, we noticed that, once avoided decomposition, famotidine displays, as cimetidine and nizatidine, a high glassforming ability and a high glass stability. In the case of cimetidine, DSC and TSDC were used to study the main relaxation, and the results are in reasonable agreement with those previously reported and obtained by DRS. TSDC results also revealed a secondary relaxation that was attributed to a JoharieGoldstein process. The molecular dynamics in the supercooled and glassy state of nizatidine was thoroughly analyzed using DSC, DRS, and TSDC. Multiple relaxation processes have been observed by DRS: the a relaxation, the slow b or JoharieGoldstein relaxation, bJG, of intermolecular origin, that appears as an excess wing in the high frequency side of the main relaxation, and 2 fast secondary relaxations b and g. The TSDC technique showed the bJG relaxation reasonably separated from the a relaxation and proved to be a very useful technique to analyze the kinetics of the bJG process in the glassy state. However, it did not detect the b and g processes because they appear in the TSDC time window at temperatures below the available temperature range. The results of the 2 dielectric techniques (DRS and TSDC) relative to the main relaxation indicate that the nizatidine is a fragile liquid, whereas cimetidine has a moderately fragile behavior. The molecular mobility in amorphous famotidine could not be studied by these techniques because of the ease with which it decomposes

M.T. Viciosa et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-12

near the melting temperature; however, DSC allowed characterizing the structural relaxation. Acknowledgments ^ncia e This work was partially supported by Fundaç~ ao para a Cie a Tecnologia (FCT), Portugal (projects UID/NAN/50024/2013 and UID/QUI/00100/2013). M.T.V. acknowledges FCT for the scholarship SFRH/BPD/110151/2015. We also acknowledge the COST action CM1402. References 1. Hancock BC, Zografi G. Characteristics and significance of the amorphous state in pharmaceutical systems. J Pharm Sci. 1997;86:1-12. 2. Craig DQM, Royall PG, Kett VL, Hopton ML. The relevance of the amorphous state to pharmaceutical dosage forms: glassy drugs and freeze dried systems. Int J Pharm. 1999;179:179-207. 3. Yu L. Amorphous pharmaceutical solids: preparation, characterization and stabilization. Adv Drug Deliv Rev. 2001;48:27-42. 4. Hancock BC. Disordered drug delivery: destiny, dynamics and the Deborah number. J Pharm Pharmacol. 2002;54:737-746. 5. Kaushal AM, Gupta P, Bansal AK. Amorphous drug delivery systems: molecular aspects, design, and performance. Crit Rev Ther Drug Carrier Syst. 2004;21: 133-193. 6. Karmwar P, Graeser K, Gordon KC, Strachan CJ, Rades T. Investigation of properties and recrystallisation behaviour of amorphous indomethacin samples prepared by different methods. Int J Pharm. 2011;417:94-100. € bmann K, Grohganz H, Rades T, Chieng N. Investigation of physical 7. Lim AW, Lo properties and stability of indomethacin-cimetidine and naproxen-cimetidine co-amorphous systems prepared by quench cooling, coprecipitation and ball milling. J Pharm Pharmacol. 2016;68:36-45. 8. Bhattacharya S, Suryanarayanan R. Local mobility in amorphous pharmaceuticals-characterization and implications on stability. J Pharm Sci. 2009;98:2935-2953. 9. Kothari K, Ragoonanan V, Suryanarayanan R. Influence of molecular mobility on the physical stability of amorphous pharmaceuticals in the supercooled and glassy states. Mol Pharm. 2014;11:3048-3055. 10. Aso Y, Yoshioka S, Kojima S. Explanation of the crystallization rate of amorphous nifedipine and phenobarbital from their molecular mobility as measured by 13C nuclear magnetic resonance relaxation time and the relaxation time obtained from the heating rate dependence of the glass transition temperature. J Pharm Sci. 2001;90:798-806. 11. Bhardwaj SP, Arora KK, Kwong E, Templeton A, Clas SD, Suryanarayanan R. Correlation between molecular mobility and physical stability of amorphous itraconazole. Mol Pharm. 2013;10:694-700. 12. Yoshioka S, Miyazaki T, Aso Y, Kawanishi T. Significance of local mobility in aggregation of beta-galactosidase lyophilized with trehalose, sucrose or stachyose. Pharm Res. 2007;24:1660-1667. 13. Yoshioka S, Aso Y. Correlations between molecular mobility and chemical stability during storage of amorphous pharmaceuticals. J Pharm Sci. 2007;96: 960-981. 14. Chang L, Shepherd D, Sun J, et al. Mechanism of protein stabilization by sugars during freeze-drying and storage: native structure preservation, specific interaction, and/or immobilization in a glassy matrix? J Pharm Sci. 2005;94: 1427-1444. 15. Chang L, Shepherd D, Sun J, Tang X, Pikal MJ. Effect of sorbitol and residual moisture on the stability of lyophilized antibodies: implications for the mechanism of protein stabilization in the solid state. J Pharm Sci. 2005;94: 1445-1455. 16. Xu Y, Carpenter JF, Cicerone MT, Randolph TW. Contributions of local mobility and degree of retention of native secondary structure to the stability of recombinant human growth hormone (rhGH) in glassy lyophilized formulations. Soft Matter. 2013;9:7855-7865. 17. Yoshioka M, Hancock BC, Zografi G. Crystallization of indomethacin from the amorphous state below and above its glass-transition temperature. J Pharm Sci. 1994;83:1700-1705. 18. Grzybowska K, Paluch M, Grzybowski A, et al. Molecular dynamics and physical stability of amorphous anti-inflammatory drug: celecoxib. J Phys Chem B. 2010;114:12792-12801. 19. Mehta M, Ragoonanan V, McKenna GB, Suryanarayanan R. Correlation between molecular mobility and physical stability in pharmaceutical glasses. Mol Pharm. 2016;13:1267-1277. 20. Tu W, Li X, Chen Z, et al. Glass formability in medium-sized molecular systems/pharmaceuticals. IeThermodynamics vs. kinetics. J Chem Phys. 2016;144:174502-174512. 21. Vasconcelos T, Sarmento B, Costa P. Solid dispersions as strategy to improve oral bioavailability of poor water soluble drugs. Drug Discov Today. 2007;12: 1068-1075. 22. Janssens S, Van den Mooter G. Review: physical chemistry of solid dispersions. J Pharm Pharmacol. 2009;61:1571-1586.

11

23. Van den Mooter G. The use of amorphous solid dispersions: a formulation strategy to overcome poor solubility and dissolution rate. Drug Discov Today Tech. 2012;9:e79-e85. 24. Vasconcelos T, Marques S, das Neves J, Sarmento B. Amorphous solid dispersions: rational selection of a manufacturing process. Adv Drug Deliv Rev. 2016;100:85-101. €bmann K, Strachan CJ, Grohganz H, Rades T. Emerging trends in 25. Laitinen R, Lo the stabilization of amorphous drugs. Int J Pharm. 2013;453:65-79. 26. Riekes MK, Engelen A, Appeltans B, Rombaut P, Stulzer HK, Van den Mooter G. New perspectives for fixed dose combinations of poorly water-soluble compounds: a case study with ezetimibe and lovastatin. Pharm Res. 2016;33: 1259-1275. € bmann K. Recent advances in co27. Dengale SJ, Grohganz H, Rades T, Lo amorphous drug formulations. Adv Drug Deliv Rev. 2016;100:116-125. 28. Bavin PMG, Post A, Zarembo JE. Cimetidine. In: Florey K, ed. Analytical Profiles of Drug Substances. Vol. 13. San Diego: Academic Press; 1984:127-182. 29. Wozniak TJ. Nizatidine. In: Florey K, ed. Analytical Profiles of Drug Substances. Vol. 19. San Diego: Academic Press; 1990:397-427. 30. Al-Omar MA, Al-Mohizea AM. Famotidine. In: Brittain HG, ed. Profiles of Drug Substances, Excipients and Related Methodology. Vol. 34. San Diego: Elsevier; 2009:115-151.
tuo GL, Ga
lico DA, Fugita RA, et al. Thermal behavior of some antihis31. Perpe tamines. J Therm Anal Calorim. 2013;111:2019-2028.
giewicz M, Ciesielski A. Application of vibrational spectroscopy, thermal 32. Jamro analyses and X-ray diffraction in the rapid evaluation of the stability in solidstate of ranitidine, famotidine and cimetidine. J Pharm Biomed Anal. 2015;107: 236-243. 33. Bauer-Brandl A. Polymorphic transitions of cimetidine during manufacture of solid dosage forms. Int J Pharm. 1996;140:195-206. 34. Calvo NL, Maggio RM, Kaufman TS. A dynamic thermal ATR-FTIR/ chemometric approach to the analysis of polymorphic interconversions. Cimetidine as a model drug. J Pharm Biomed Anal. 2014;92:90-97. 35. Otsuka M, Kato F, Matsuda Y. Physicochemical stability of cimetidine amorphous forms estimated by isothermal microcalorimetry. AAPS PharmSciTech. 2002;3:32-44. €m CAS. Physical stability of drugs 36. Alhalaweh A, Alzghoul A, Mahlin D, Bergstro after storage above and below the glass transition temperature: relationship to glass-forming ability. Int J Pharm. 2015;495:312-317. 37. Pajula K, Taskinen M, Lehto VP, Ketolainen J, Korhonen O. Predicting the formation and stability of amorphous small molecule binary mixtures from computationally determined Flory-Huggins interaction parameter and phase diagram. Mol Pharm. 2010;7:795-804.
met Z, Hegedüs B, Sza
ntay Jr C, Sztatisz J, Pokol G. Pressurization 38. Ne effects on the polymorphic forms of famotidine. Thermochim Acta. 2005;430: 35-41. 39. Lin SY, Cheng WT, Wang SL. Thermodynamic and kinetic characterization of polymorphic transformation of famotidine during grinding. Int J Pharm. 2006;318:86-91. 40. Ali Ibrahim AS, Hassan MA, Abd El-Mohsen MG, El-Shanawani SM. Compatibility study of famotidine with some pharmaceutical excipients using different techniques. Bull Pharm Sci Assiut Univ. 2004;27(2):281-291.
valos JZ, Jime
nez P. Effect of pressure on the polymorphic forms 41. Roux MV, Da of famotidine. Thermochim Acta. 2002;394:19-24.
ter I, Ka
lma
n A, Pa
rka
nyi L. Comparison of 42. Hegedüs B, Bod P, Hars
anyi K, Pe the polymorphic modifications of famotidine. J Pharm Biomed Anal. 1989;7: 563-569. 43. Moura Ramos JJ, Taveira-Marques R, Diogo HP. 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. 44. Diogo HP, Viciosa MT, Moura Ramos JJ. Thermally stimulated currents: principles, methods for data processing and significance of the information provided. Supplementary Data Thermochim Acta. 2016;623:29-35. 45. Teyssedre G, Lacabanne C. Some considerations about the analysis of thermostimulated depolarization peaks. J Phys D Appl Phys. 1995;28:14781487. 46. van Turnhout J. Thermally Stimulated Discharge of Polymer Electrets. Amsterdam: Elsevier Sci. Pub. Co; 1975. 47. Chen R, Kirsh Y. Analysis of Thermally Stimulated Processes. Oxford: Pergamon Press; 1981. 48. van Turnhout J. Thermally stimulated discharge of electrets. In: Sessler GM, Ascheron C, Dresslhaus MS, eds. Electrets, Topics in Applied Physics. Vol. 33. Berlin/Heidelberg: Springer; 1987:81-215. 49. Gun'ko VM, Zarko VI, Goncharuk EV, et al. TSDC spectroscopy of relaxational and interfacial phenomena. Adv Colloid Interface Sci. 2007;131:1-89. 50. Sauer BB. Thermally stimulated currents: recent developments in characterisation and analysis of polymers. In: Cheng SZD, ed. Applications to Polymers and Plastics, 653e711, in Vol. 3 of Handbook of Thermal Analysis and Calorimetry by Gallagher PK, ed. Amsterdam: Elsevier; 2002. 51. Vassilikou-Dova A, Kalogeras IM. Dielectric analysis (DEA). In: Menczel JD, Prime RB, eds. Thermal Analysis of Polymers: Fundamentals and Applications. 1st ed. Hoboken, New Jersey: John Wiley; 2009:497-613. 52. Boutonnet-Fagegaltier N, Lamure A, Menegotto J, et al. The Use of Thermally Stimulated Current Spectroscopy in the Pharmaceutical Sciences. In: Craig DQM, Reading M, eds. Thermal Analysis of Pharmaceuticals. Boca Raton: CRC Press; 2007:359-382.

12

M.T. Viciosa et al. / Journal of Pharmaceutical Sciences xxx (2016) 1-12

53. Baker S, Antonijevic MD. Thermal analysisddielectric techniques. In:
n I, eds. Solid State Characterization of Pharmaceutics. ChiStorey RA, Yme chester, UK: Wiley; 2011:187-206. 54. Havriliak S, Negami S. A complex plane representation of dielectric and mechanical relaxation processes in some polymers. Polymer. 1967;8:161-210. €nhals A, Kremer F. Analysis of dielectric spectra. In: Kremer F, 55. Scho €nhals A, eds. Broadband Dielectric Spectroscopy. Berlin: Springer-Verlag; Scho 2003:59-98. €ttcher CJF, Bordewijk P. Theory of electric polarization. In: Dielectrics in time 56. Bo dependent fields. Vol. II. Amsterdam: Elsevier; 1978. 57. Wübbenhorst M, van Turnhout J. Analysis of complex dielectric spectra. I. One-dimensional derivative techniques and three-dimensional modelling. J Non Cryst Sol. 2002;305:40-49. 58. Baird JA, Van Eerdenbrugh B, Taylor LS. A classification system to assess the crystallization tendency of organic molecules from undercooled melts. J Pharm Sci. 2010;99:3787-3806. 59. Moynihan CT, Easteal AJ, Wilder J, Tucker J. Dependence of the glass transition temperature on heating and cooling rate. J Phys Chem. 1974;78:2673-2677. 60. Simatos D, Blond G, Roudaut G, Champion D, Perez J, Faivre AL. Influence of heating and cooling rates on the glass transition temperature and the fragility parameter of sorbitol and fructose as measured by DSC. J Therm Anal. 1996;47: 1419-1436. 61. Weuts I, Kempen D, Six K, et al. Evaluation of different calorimetric methods to determine the glass transition temperature and molecular mobility below Tg for amorphous drugs. Int J Pharm. 2003;259:17-25.
lek J. Description of enthalpy relaxation dynamics in terms of 62. Svoboda R, Ma TNM model. J Non Cryst Sol. 2013;378:186-195. 63. Svoboda R. How to determine activation energy of glass transition. J Therm Anal Calorim. 2014;118:1721-1732. 64. Svoboda R. Novel equation to determine activation energy of enthalpy relaxation. J Therm Anal Calorim. 2015;121:895-899. 65. Crichton SN, Moynihan CT. Dependence of the glass transition temperature on heating rate. J Non Cryst Sol. 1988;99:413-417. 66. Crowley KJ, Zografi G. The use of thermal methods for predicting glass-former fragility. Thermochim Acta. 2001;380:79-93. 67. Angell CA. Relaxation in liquids, polymers and plastic crystalsdstrong/fragile patterns and problems. J Non Cryst Sol. 1991;131-133:13-31. €hmer R, Ngai KL, Angell CA, Plazek DJ. Nonexponential relaxations in strong 68. Bo and fragile glass formers. J Chem Phys. 1993;99:4201-4209. 69. Urpayil S, Thayyil MS. Thermal and spectroscopic studies on nizatidine-PVP mixture. ISOR J Pharm. 2012;2(6):17-23. 70. Urpayil S, Thayyil MS. Fragility of cimetidine drug probed by broadband dielectric spectroscopy. Transl Med. 2014;4:138. 71. Correia NT, Alvarez C, Moura Ramos JJ. Fragility in side-chain liquid crystalline polymers: the TSDC contribution. Polymer. 2000;41:8625-8631. 72. Correia NT, Alvarez C, Moura Ramos JJ, Descamps M. Molecular motions in molecular glasses as studied by thermally stimulated depolarisation currents (TSDC). Chem Phys. 2000;252:151-163. 73. Moura Ramos JJ, Mano JF, Sauer BB. Some comments on the significance of the compensation effect observed in thermally stimulated current experiments. Polymer. 1997;38:1081-1089. 74. Mora E, Diogo HP, Moura Ramos JJ. The slow molecular dynamics in amorphous probucol. Thermochim Acta. 2014;595:83-88. €ngigketsgesetz der viskosita €t von flüssig75. Vogel H. Das temperature-abha keiten. (The temperature dependence law of the viscosity of fluids). Physik Z. 1921;22:645-646. 76. Fulcher GS. Analysis of recent measurements of the viscosity of glasses. J Am Ceram Soc. 1925;8:339-355. €ngigkeit der Viscosita €t von der Temperatur 77. Tammann G, Hesse W. Die Abha bie unterkühlten Flüssigkeiten. Z Anorg Allg Chem. 1926;156:245-257. 78. Ngai KL. An extended coupling model description of the evolution of dynamics with time in supercooled liquids and ionic conductors. J Phys Condens Matter. 2003;15:S1107-S1125. 79. Ngai KL, Paluch M. Classification of secondary relaxation in glass-formers based on dynamic properties. J Chem Phys. 2004;120:857-873. 80. Kohlrausch R. Theorie des elektrischen Rückstandes in der Leidener Flasche. Annalen der Physik Chem (Poggendorff). 1854;91:179-214.

81. Williams G, Watts DC. Non-symmetrical dielectric relaxation behaviour arising from a simple empirical decay function. Trans Faraday Soc. 1970;66: 80-85. 82. Ngai KL. Relaxation and Diffusion in Complex Systems. New York: Springer; 2011:364-380. 83. Paluch M, Roland CM, Pawlus S, Ziolo J, Ngai KL. Does the Arrhenius temperature dependence of the Johari-Goldstein relaxation persist above T(g)? Phys Rev Lett. 2003;91:115701. €ssler EA. Beta relaxation versus high frequency wing in the 84. Blochowicz T, Ro dielectric spectra of a binary molecular glass former. Phys Rev Lett. 2004;92(225701):4. € plinger J, Kasper G, Hunklinger S. Glass transition in chlor85. Ko obenzeneedecalin under pressure. J Chem Phys. 2000;113:4701-4706. 86. Capaccioli S, Kessairi K, Prevosto D, Lucchesi M, Rolla PA. Correlation of structural and Johari-Goldstein relaxations in systems vitrifying along isobaric and isothermal paths. J Phys Condens Matter. 2007;19:205133. 87. Nogales A, Sanz A, Ezquerra TA. On the role of the beta process as precursor of the alpha relaxation in aromatic polyesters. J Non Cryst Sol. 2006;352:46494655. 88. Casalini R, Snow AW, Roland CM. Temperature dependence of the JohariGoldstein relaxation in poly(methyl methacrylate) and poly(thiomethyl methacrylate). Macromolecules. 2013;46:330-334. 89. Jarosz G, Mierzwa M, Ziolo J, Paluch M, Shirota H, Ngai KL. Glass transition dynamics of room-temperature ionic liquid 1-methyl-3-trimethylsilylmethylimidazolium tetrafluoroborate. J Phys Chem B. 2011;115:12709-12716. 90. Capaccioli S, Ngai KL, Thayyil MS, Prevosto D. Coupling Caged Molecule Dyn JG B-Relaxation. J Phys Chem B. 2015;119:8800-8808. 91. Ngai KL, Capaccioli S, Prevosto D, Wang LM. Coupling of caged molecule dynamics to JG b-relaxation II: polymers. J Phys Chem B. 2015;119:12502-12518. 92. Ngai KL, Capaccioli S, Prevosto D, Wang LM. Coupling of caged molecule dynamics to JG b-relaxation III:van der Waals glasses. J Phys Chem B. 2015;119: 12519-12525. 93. Correia NT, Alvarez C, Moura Ramos JJ, Descamps M. Glass transition relaxation and fragility in the molecular glass forming m-toluidine: a study by thermally stimulated depolarization currents. J Chem Phys. 2000;113:32043211. 94. Starkweather HW. Simple and complex relaxations. Macromolecules. 1981;14: 1277-1281. 95. Starkweather HW. Noncooperative relaxations. Macromolecules. 1988;21: 1798-1802. 96. Starkweather HW. Aspects of simple, non-cooperative relaxations. Polymer. 1991;32:2443-2448. 97. Correia NT, Moura Ramos JJ. On the cooperativity of the b-relaxation: a discussion based on dielectric relaxation and thermally stimulated depolarisation currents data. Phys Chem Chem Phys. 2000;2:5712-5715. 98. Moura Ramos JJ, Diogo HP, Pinto SS. Effect of physical aging on the JohariGoldstein and alpha relaxations of D-sorbitol: a study by thermally stimulated depolarization currents. J Chem Phys. 2007;126:144506. 99. Moura Ramos JJ, Diogo HP, Pinto SS. Local motions in L-iditol glass: identifying different types of secondary relaxations. Thermochim Acta. 2008;467: 107-112. 100. Pinto SS, Moura Ramos JJ, Diogo HP. The slow molecular mobility in poly(vinyl acetate) revisited: new contributions from thermally stimulated currents. Eur Polym J. 2009;45:2644-2652. 101. Diogo HP, Moura Ramos JJ. Contribution of the technique of thermostimulated currents for the elucidation of the nature of the Johari-Goldstein and other secondary relaxations in the vitreous state. IEEE Trans Dielect Elect Insul. 2014;21:2301-2309. 102. Okamoto N, Oguni M. Discovery of crystal nucleation proceeding much below the glass transition temperature in a supercooled liquid. Solid State Commun. 1996;99:53-56. 103. Okamoto N, Oguni M, Sagawa Y. Generation and extinction of a crystal nucleus below the glass transition temperature. J Phys Condens Matter. 1997;9: 9187-9198. 104. Johari GP, Kim S, Shanker RM. Dielectric relaxation and crystallization of ultraviscous melt and glassy states of aspirin, ibuprofen, progesterone, and quinidine. J Pharm Sci. 2007;96:1159-1175.