Article pubs.acs.org/JPCB
Molecular and Vibrational Dynamics in the Cholesterol-Lowering Agent Lovastatin: Solid-State NMR, Inelastic Neutron Scattering, and Periodic DFT Study Paweł Bilski,*,†,‡ Kacper Druzḃ icki,†,‡ Jacek Jenczyk,§ Jadwiga Mielcarek,∥ and Jan Wąsicki†,§ †
Faculty of Physics, Adam Mickiewicz University, Umultowska 85, 61-614, Poznan, Poland Frank Laboratory of Neutron Physics, Joint Institute for Nuclear Research, 141980, Dubna, Russia § NanoBioMedical Centre, Adam Mickiewicz University, Umultowska 85, 61-614, Poznan, Poland ∥ Department of Inorganic and Analytical Chemistry, Poznan University of Medical Sciences, Grunwaldzka 6, 60-780, Poznan, Poland ‡
ABSTRACT: Molecular and vibrational dynamics of a widely used cholesterol-lowering agent, lovastatin, have been studied by combining nuclear magnetic resonance relaxation experiments (1H NMR) with inelastic neutron scattering (INS) and periodic density functional theory modeling (plane-wave DFT). According to a complementary experimental study, lovastatin shows no phase transitions down to cryogenic conditions, while a progressive, stepwise activation of several molecular motions is observed below room temperature. The molecular packing and intermolecular forces were analyzed theoretically, supported by a 13C NMR study and further correlated with observed molecular dynamics. The NMR relaxation experiments combined with theoretical calculations disclose that molecular dynamics in solid lovastatin is related to methyl group motions and conformational disorder in the methylbutanoate fragment. This is precisely assigned and analyzed quantitatively from both experimental and theoretical perspectives. The neutron vibrational spectroscopy further corroborates that the methyl rotors have a classical nature. In addition to the intramolecular reorientations, the vibrational dynamics was analyzed with an emphasis on the low-wavenumber range. For the first time, the terahertz response of lovastatin was studied by confronting neutron and optical techniques and clearly illustrating their complementarity. The consistent picture of the molecular dynamics is provided, which may support further considerations on alternative drug formulations and the amorphization tendency in this important lipidlowering drug.
■
INTRODUCTION Inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A reductase, HMG-CoA, also commonly known as statins, are the most effective and the most widely used cholesterol-lowering drugs. Statin monotherapy is generally well tolerated, with a rather low frequency of adverse effects. The most important adverse effects can be associated with myopathy and an asymptomatic increase in hepatic transaminases, both of which, however, occur infrequently.1 The first FDA-approved HMG-CoA inhibitor is lovastatin (hereafter LOV), sold as Mevacor, Altocor, and Altoprev, and primarily introduced to control hypercholesterolemia. It is interesting to note that LOV is a natural product (referred to as Monacolin K), being a secondary metabolite of several fungal species, including Aspergillus terreus, Penicillium citrinum, Pleurotus spp., Monascus ruber, or the soil fungus Aspergillus sclerotiorum.2 LOV occurs in two molecular forms, namely, as an unstable hydroxy acid and as a lactone (see Figure 1). The latter form is converted in vivo into the corresponding β-hydroxy acid, which is the biologically active form.3 It inhibits HMG-CoA reductase, © XXXX American Chemical Society
which is the rate-limiting enzyme involved in cholesterol biosynthesis. As other statins, LOV reduces the levels of “bad” cholesterol (low-density lipoprotein, or LDL) while increasing levels of “good” cholesterol (high-density lipoprotein, or HDL). Serum cholesterol level is strongly associated with coronary heart disease. For that reason, LOV is given to lower the risk of stroke, heart attack, and other heart complications in people with diabetes or other risk conditions. LOV is also the most effective lipid-lowering substance for transplant patients, and it is generally well tolerated by most of them. While traditionally used for lowering the serum cholesterol levels, there is also some evidence indicating its potential chemotherapeutic potential in cancer therapy by inducing tumor cell apoptosis and inhibiting invasiveness of tumor cells. These properties are collectively referred to as nonlipid or pleiotropic effects. Pleiotropism of LOV is the protective activity of endothelium, Received: February 6, 2017 Revised: March 10, 2017 Published: March 10, 2017 A
DOI: 10.1021/acs.jpcb.7b01090 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
Such a conclusion motivates us to examine the performance of modern solid-state calculations in physical-chemical analysis of statin drugs. This paper fills the gap and opens the discussion on molecular dynamics and low-wavenumber phonon excitations, starting with the case of the first FDA-approved agent from this family, LOV. To this end, the low-wavenumber vibrational response is explored with neutron vibrational spectroscopy and confronted with the earlier reported THzTDS data.6 For the first time, the low-wavenumber phonon spectrum of the model statin is explored with the first-principles calculations, proving high reliability and accuracy of high-end solid-state computations and illustrating the complementarity of neutron and optical terahertz spectroscopy techniques. So far, crystalline LOV has not been studied so extensively as SIM, leaving its structural considerations open. By focusing on the LOV in bulk form, we thoroughly examine an influence of the crystal packing on its molecular dynamics by employing NMR relaxometry combined with first-principles computations. This work is oriented toward drawing an accurate model of the local relaxation in the reference bulk form of LOV, and searching for possible peculiarities in the low-wavenumber vibrational response. Such results will facilitate future exploration of LOV in the amorphous state and other alternative formulations. Despite the enormous chemical size of the considered system, a remarkable synergy between the theory and experiment was found.
Figure 1. Molecular formula of lovastatin (LOV; (1S,3R,7S,8S,8aR)-8{2-[(2R,4R)-4-hydroxy-6-oxooxan-2-yl]ethyl}-3,7-dimethyl1,2,3,7,8,8a-hexahydronaphthalen-1-yl (2S)-2-methylbutanoate) along with the labeling adopted for methyl groups (I−IV) and carbon atoms.
plaque stabilization, anti-inflammatory effect, and the impact on coagulation and fibrinolysis. Because of these properties, LOV may find use for treatment of autoimmune diseases and neurodegenerative diseases or in transplantation. LOV is an antihyperlipidemic and lipophilic statin, which penetrates the central nervous system (CNS). It has extensive first pass metabolism and is biotransformed in the liver primarily by cytochrome P450. As a consequence of the extensive hepatic metabolism of LOV, its bioavailability is low and variable due to general circulation. Although statin drugs have been widely known on the pharmaceutical market, relatively little attention has been given so far to understand their structural properties and molecular dynamics in their solid formulations. What is more, the low aqueous solubility of statins, and thus their reduced bioavailability, calls for the search of alternative formulations with the amorphous state in focus. Therefore, a deep understanding of the bulk properties of statins becomes an essential query to address. However, the structure of LOV in bulk form is not a trivial aspect, as it is complicated by a subtle balance of intermolecular forces, conformational flexibility, and local disorder. An understanding of these contributing factors can be partially achieved by employing novel experimental tools supported by theoretical solid-state techniques, for instance, nuclear magnetic resonance (NMR) or terahertz (THz) spectroscopy. The proper understanding of these experiments becomes, however, a great challenge, mainly because of the enormous size of the unit cells and molecular ensembles considered. All in all, definitely much more attention already has been paid to the analogous HMG-CoA inhibitor, simvastatin (SIM). Recently, Nunes et al. have presented an extensive study of SIM in the supercooled and glassy states,4 illustrating the power of dielectric spectroscopy and nuclear magnetic resonance techniques and pointing out the role of local relaxation events in amorphization of statins. The power of the time-domain terahertz spectroscopy (THz-TDS) in exploration of solid-state transformations of SIM has also been recently illustrated by Tan and Zeitler.5 It was, however, concluded that the state-of the-art solid-state calculations are not practicable at present resources available, remaining an open challenge. This is because of the large size of the unit cell in SIM and the fact that computational time scales roughly with the cube of the unit cell volume.
I. EXPERIMENTAL AND COMPUTATIONAL DETAILS Sample. The high-purity LOV sample was kindly provided by BIOFARM pharmaceutical company, Poznan, Poland. Prior to extensive experimental research, the integrity, crystallinity, and thermal stability of the sample were verified in a broad range of temperatures. Such an analysis did not reveal any trails of phase instability, which stays in line with the literature reports (see, e.g., refs 7 and 8). Solid-State Spectroscopy. Inelastic neutron scattering (INS) measurements were performed with an invertedgeometry spectrometer NERA, set at the high flux pulsed nuclear reactor IBR-2 at JINR Dubna, Russia.9 The incident neutron energies were determined by measuring the neutron time-of-flight (TOF) across the ∼110 m distance from the water moderator to the sample. The INS spectra were recorded at the final energy of the scattered neutrons of Ef = 4.65 meV, fixed by the crystal analyzers and the beryllium filters. The INS spectra were collected at both ambient and highpressure conditions. In the former case, the ∼5 g sample was placed in a flat aluminum container and measured approximately for 16 h at 50 K. The sample cooling was achieved via the use of a closed-cycle helium refrigerator. The high-pressure INS measurements were performed with a dedicated chamber, working in the pressure regime of 1−400 MPa. The lowtemperature down to 90 K was accessed with the liquidnitrogen cooling system. The ∼7 g sample was placed in an aluminum pressure tube. In each case, the final time-of-flight spectra were transformed into the form of generalized density of states (GDOS).10 The 13C CP/MAS NMR spectrum of LOV was acquired using the Agilent spectrometer, operating at a Larmor frequency of 400 MHz for protons. The sample was placed within a 4 mm diameter zirconia rotor and spun at 5 kHz, with a cross-polarization contact time of 1300 μs. Two-pulse phasemodulated decoupling was utilized during the acquisition period. 1024 transients were accumulated so as to achieve a B
DOI: 10.1021/acs.jpcb.7b01090 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
II. RESULTS Crystal Packing. To date, only a few reports on the structural properties of LOV exist. The crystal structure was solved in the 1980s by Alberts et al.23 and then by Sato et al.11 LOV was found to crystallize in the orthorhombic P212121 space group, with Z = 4.11 The existence of polymorphism in LOV was excluded by Yoshida et al.,8 who have proven that the P212121 space group is highly stable. While LOV did not show any classical structural polymorphism, it presents a distinctive morphology depending on the type of solvent used. The roomtemperature single-crystal and powder X-ray data provided the following cell constants: a = 22.15 Å, b = 17.32 Å, c = 5.97 Å (XRD) and a = 22.59 Å, b = 17.65 Å, c = 6.08 Å (PXRD), according to Sato et al.11 and Yoshida et al.,8 respectively. For comparison, the equilibrium PBE-TS structure at 0 K (see Figure 2a) is defined by the following cell parameters: a = 21.73
satisfactory signal-to-noise (S/N) ratio, with a total suppression of spinning sidebands (TOSS) used to detect solely isotropic signals. The 13C chemical shifts were referenced to the 38.3 ppm signal of adamantane. The measurement of the second moment of the proton NMR line was carried out as a function of temperature by a continuous wave spectrometer, operating on proton nuclei at a frequency of 28 MHz. 1H NMR measurements of the spin− lattice relaxation times T1 were performed using laboratory made pulsed NMR spectrometers, working at 25 and 58.9 MHz. T1 times were determined using the saturation−recovery sequence. The temperature of the sample was controlled by means of a gas-flow cryostat, being monitored with a Cernox resistor to a maximum uncertainty of 0.1 K. Computational Methodology. All of the theoretical calculations were performed in periodic boundary conditions (PBC), starting from the P212121 crystal structure, reported by Sato et al.11 (CSD ref code: CEKBEZ). Plane-wave/pseudopotential formulation of density functional theory (PW-DFT) was employed as implemented in CASTEP.12,13 Exchange and correlation were defined with the generalized gradient approximation (GGA) exchange-correlation functional from Perdew−Burke−Ernzerhoff (PBE) and augmented with pairwise dispersion-corrections from Tkatchenko and Scheffler.14,15 Core electrons were represented by norm-conserving pseudopotentials, while the electronic wave functions were expanded using a PW basis set, with a kinetic energy cutoff of 1050 eV. The Monkhorst−Pack grid was kept to maintain the constant k-spacing of 0.07 Å−1. The convergence criteria in variation of the total energy, maximum force, external stress, maximum displacement, and SCF iterations were defined as 1 × 10−10 eV/atom, 1 × 10−5 eV/ Å, 0.0001 GPa, 1 × 10−6 Å, and 1 × 10−12 eV/atom, respectively. Prior to vibrational analysis, the crystal structure was fully relaxed at the atmospheric pressure. The phonon frequencies were obtained by diagonalization of the dynamical matrices computed using density functional perturbation theory (DFPT).16,17 DFPT also provided dielectric permittivity and Born charge tensors, directly used to compute the terahertz activities. The INS spectra were modeled with the help of the aClimax program, based on the DFPT eigenfrequencies and eigenvectors.18 The 13C NMR calculations were performed with the gaugeincluding projector-augmented wave (GIPAW) method.19 Allelectron information was reconstructed using scalar relativistic ultrasoft pseudopotentials generated on-the-f ly (USPP).20 Adiabatic activation barriers for the internal reorientations were calculated at 0 K with PBE-TS. The same, rather restricted, numerical conditions were used for constrained geometry optimization, where both the equilibrium volume and selected dihedral angles were fixed during relaxation. The potential energy surface (PES) was also scanned with PW-DFT in a rigid manner (without subsequent structure relaxation) and further compared to the results of calculations performed with use of atom−atom potentials.21 The analysis of the second moment of the 1H NMR line was based on the Monte Carlo iterative scheme, according to the procedure earlier described in ref 22. The calculations were performed for the {3 × 3 × 3} supercell built from the equilibrium PBE-TS geometry.
Figure 2. Molecular packing of in P212121 phase of lovastatin (a) projected toward the c-axis along with close contacts and normalized Hirshfeld surface reflecting noncovalent forces (b).
Å, b = 17.47 Å, c = 5.77 Å. Although omitting the finitetemperature effects (i.e., cell expansion), the PBE-TS equilibrium cell constants agree very well with the singlecrystal data. One can consider three main parts of the LOV molecule, namely, the lactone ring (lct), the naphthalene moiety (nph), and the methylbutanoate chain (mbt). According to Figure 2a, the molecules are clearly bound via the O···HO hydrogen bonds propagated to infinity toward the crystallographic axis b. A hydrogen bond links the lct rings (hydroxyl group) with the mbt moieties (carbonyl group). The second carbonyl group present in the each lct ring remains unbound in the crystal state. C
DOI: 10.1021/acs.jpcb.7b01090 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
These results are fully interpreted with the help of theoretical C NMR computations within the GIPAW methodology. The reference isotropic values were obtained with the help of the iso iso expression δiso calc = δref − aGIPAW, where both the reference, δref, and scaling factor, a, were taken from ref 24. The full collection of the experimental and theoretical resonances is given in Table 1. Additionally, the influence of the crystal environment is
The O···HO is the only obvious noncovalent force present here. To further elucidate the noncovalent forces, one can refer to Figure 2b, which presents the close contacts along with the Hirshfeld surface. This enables identification of the regions of particular importance to intermolecular interactions present. The Hirshfeld surface is visualized in the normalized (−1.00 to 1.00), dnorm, projection, measuring whether the close contact is greater (blue), the same (white), or shorter (red) than the sum of the van der Waals radii for the atoms considered. Through more detailed inspection of the regions of particular importance, one can note that the crystal structure is mainly stabilized by the T-shaped stacking interactions between the naphthalene rings, further supported by the dihydrogen interactions of attracted nph rings via the CH···HC contacts. The dihydrogen bonds also attract the lct rings through the CH2 contacts. Red dots indicate the presence of weak CH···O forces, formed by free carbonyl groups in the lct rings and both CH and CH2 groups in the neighboring lct and nph parts, respectively. These attracting forces, along with large steric barriers arising from the highly anisotropic shape of the LOV molecules, preserve their reorientations in the bulk. Nevertheless, some molecular fragments may still be in a considerable thermal motion, which has been thoroughly analyzed further down below. The lct rings and ester groups (in mbt) are evidently immobilized by the O···HO hydrogen bonds. Nevertheless, there is still a large probability for the methyl group motion and dynamics of the sec-butyl part. In the latter case, both the internal conformational change as well as rotation of the whole sec-butyl fragment are equally possible. According to the projection of the Hirshfeld surface, there is a considerable crowding in the crystal, resulting in the close contacts between methyl group nos. II and III, while methyl group nos. I and IV remain relatively free. The methyl group no. II is further affected by CH2 from the neighboring lct ring. It suggests that close contacts may further affect the dynamics of the methyl rotors, which is discussed further down below. 13 C CP/MAS NMR. The assumed model of the crystal structure with only one possible conformation has been further verified with 13C CP/MAS NMR measurements (see Figure 3).
13
Table 1. Collection of Experimental (13C CP/MAS NMR at 298 K) and Theoretical (GIPAW PBE-TS at 0 K) Chemical Shifts (δ in ppm) for Lovastatina 13
δ
exp
179 171 136 134 128 76 69 62 41 39 37 37 34 33 30 28 25 24 14 12
C chemical shift δ (ppm)
Δδ
δGIPAW
atom
groupb
2.1 0.3 2.9 2.4 −1.6 −0.4 0.3 1.0 −0.5 −0.6 0.3 0.3 −0.4 0.8 1.0 0.2 −0.7 0.5 1.2 0.7 1.1 −2.3 0.1 0.3
179.5 169.0 137.6 135.5 130.5 128.6 78.5 70.9 63.8 41.1 37.1 36.8 36.7 34.9 32.0 31.8 30.8 30.5 28.9 22.1 21.4 11.9 11.5 9.9
C(1″) C(6′) C(6) C(4a) C(4) C(5) C(2′) C(1) C(4′) C(2″) C(8) C(5′) C(8a) C(3′) C(2) C(10) C(7) C(3″) C(3) C(9) C(II) C(III) C(I) C(IV)
CO···H (mbt) CO (lct) CH (nph) CH (nph) CH (nph) CH (nph) CH (lct) CH-mbt (nph) CHCH2 (nph) CHCH3 (mbt) C-mbt (nph) CH2 (lct) CH (nph) CH2 (lct) CH2 (nph) CH2CH2 CCH3 (nph) CCH3 (nph) CCH3 (mbt) CH2CH2 CH3 (nph) CH3 (mbt) CH3 (nph) CH3 (mbt)
Atom numbering according to Figure 1. Δδ refers to the difference w.r.t. liquid NMR data reported by Moore et al.25 blct - actone; nph naphtalene; mbt - methylbutanoate. a
traced with use of the Δδ parameter, which is simply defined as the difference in the chemical shift between the CP/MAS and the liquid 13C NMR data reported elsewhere.25 A nearly perfect match has been achieved in the upper spectral range, excluding a negligible (∼1 ppm) splitting of the signal at 128 ppm (carbon nos. 4 and 5 in the nph moiety). Such a discrepancy, although negligible, can be attributed to some inaccuracy in the description of the T-shaped stacking forces between the nph rings. The presence of the singlets at 171 and 179 ppm confirms an existence of the well-defined Hbond architecture of only one type. More noticeable but still negligible deviations are found in the range of ∼30−40 ppm. However, we believe that such a discrepancy is the result of the sum of errors due to a large number of resonances present, rather without any serious physical meaning. Finally, the highest NMR frequency is exhibited by the CH3 resonances, contributing as a triple signal at 24, 14, and 12 ppm, respectively. Analysis of the Δδ confirms that methyl group nos. I and IV are expected to be relatively unaffected by the crystal environment. In contrast, methyl group no. III seems to be highly deshielded. An opposite trend was found for group II. It
Figure 3. Room-temperature experimental 13C CP/MAS NMR spectrum of lovastatin along with the theoretical one, calculated with GIPAW (PBE-TS) for the P212121 crystallographic structure. D
DOI: 10.1021/acs.jpcb.7b01090 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B hence suggests that these two moieties are coupled to each other. Vibrational Dynamics. While the theoretical calculations confirm a high appropriateness of the adopted model, further conclusions can be derived from the computationally supported low-wavenumber vibrational spectroscopy. To this end, INS spectroscopy has been employed. An advantage of the INS spectroscopy over photon-based techniques comes from the lack of selection rules. Unlike optical vibrational spectroscopy, INS does not follow the symmetry of a given vibration, as all modes are in principle allowed for observation. What INS measures is the dynamical structure factor S(Q, ω), generally defined as Sinc(Q , nωi) ∝
(QUi)2n exp( − (Qu Total)2 )σinc n!
(1)
where Ui stands for the mass-weighted amplitude of the vibration in the ith mode, ωi, and n denotes the transition order, i.e., n = 1 for a fundamental transition and n > 1 for an overtone. The exponential term is known as the Debye−Waller factor, which can be reduced by cooling down the sample.26 Formally, the INS spectra can be presented in the form of the so-called generalized density of states (GDOS), which can be expressed as
GDOS =
S(Q , ω)ω Q 2[nB + 1]
(2) 1
where nB stands for the Bose factor, nB = βℏω , and the e −1 scattered intensity is summed at fixed energy transfer over the total range of angles defined by the instrument geometry.10 Hence, the spectral intensity generally depends directly on the phonon displacement amplitude and the incoherent neutron scattering cross section, σinc, of a given atom, which is an element specific property. The latter is particularly large for protium (σiH = 80.26 barns), which far exceeds that for any other species. Therefore, INS is superbly sensitive to hydrogen dynamics. The INS spectrum of LOV, recorded at 50 K, is presented in Figure 4 versus the theoretical one calculated in the harmonic approach. Despite the enormous size of the system, the predicted envelope matches very well to the experimental spectrum. Here, the INS spectrum can be defined as the amplitude weighted density of hydrogenous vibrational states. The theoretical spectrum extracted for CH3 and CH2 (mbt) groups is presented versus the remaining contributions, coming from both lct and nph fragments. It becomes clear that the INS spectrum is covered by different CH modes from the molecular framework. These modes completely cover the range above ∼325 cm−1 as being assigned to multiple different CH/CH2 deformations. Other modes, like out-of-plane H-bond deformations, γ(OH), are buried and could not be clearly defined experimentally. Nevertheless, direct access to the vibrational dynamics of hydrogen allows one to probe vibrations not accessible to optical spectroscopy, i.e., methyl torsions. These vibrations are of very high amplitude, and therefore dominate intensity in the lower-wavenumber range. These can be clearly identified thanks to the theoretical calculations (see the blue features in Figure 4) and have been further decomposed in the upper panel of Figure 4 (200−350 cm−1).
Figure 4. Experimental (NERA at 50 K) and theoretical (harmonic, DFPT/PBE-TS) INS spectra of lovastatin along with a general band assignment.
The complete assignment of the INS spectrum in the lower wavenumber range has been further provided in Table 2. From inspection of both Figure 4 and Table 2, one can clearly see that the INS confirms our previous findings from structural and NMR study, that methyl group nos. I and IV are similar and their torsional modes contribute to the most intense band at 225 cm−1. Since INS intensity is proportional to the displacement amplitude, one can assume that these groups are relatively free in motion. The second pair of methyl groups contributes at higher wavenumbers, namely, at 246 (no. II) and 264 cm−1 (no. III), which suggests that their rotational barriers are higher. It is interesting to note that the lowest intensity for the τ(CH3) mode is observed for methyl no. III, while NMR shows a prominent density outflow, resulting in a considerable deshielding. Moreover, one can note that the predicted frequencies of the τ(CH3) modes agree very well with the experimental ones, and also that their wavenumbers are relatively high and not coupled to the low-wavenumber modes. It suggests their classical, nonquantum behavior. According to Figure 4, CH3 modes also contribute considerably below 200 cm−1. Nevertheless, these are contributions from the in-plane, δ, and wagging motions, ω. The temperature and pressure dependence of the modes under interest was examined with further experiments, as shown in Figure 5. The difference in the intensity distribution for the higher pressure spectra partially comes from different technical conditions (cryostat, pressure chamber, different E
DOI: 10.1021/acs.jpcb.7b01090 J. Phys. Chem. B XXXX, XXX, XXX−XXX
Article
The Journal of Physical Chemistry B
Table 2. Collection of Phonon Wavenumbers for Lovastatin (in cm−1) according to INS at 50 K (This Work) and THz-TDS at 298 K (from Nowak et al.6)a ν (cm−1) THz-TDS THz-TDS6 (298 K)
INS DFPT (0 K)
84
83
74
77
66
65
44
NERA (50 K)
DFPT (0 K)
tentative band assignmentb
329 308 291 283 264 246 225 220 188 172 148 116 101
323 298 283 277 258 243 212 210 183 168 150 128 102
80
82
68
67
56 50
59 55
43 38
48 37
27
28
20
22
τ(CH3)[III] + τ(CH2)[III] τ(CH3)[I] + τ(CH3)[IV] τ(CH3)[I] τ(CH3)[III] + δ(C−C−CH3)[IV] τ(CH3)[II] + τ(CH3)[III] τ(CH3)[II] τ(CH3)[I] + τ(CH3)[IV] τ(CH3)[I] + τ(CH3)[IV] δ(C−C−CH3)[IV] τ(C−CH2−CH2−C) [lct] ω(C−CH3)[I] + ω(C−CH3)[II] + ω(nph) τ(lct) ω(C−CH3)[I] + ω(C−CH3)[II] τ(lct) + ω(C−O) + ν(O···O) τ(nph) + τ(mbt) ω(lct) ω(nph) + ω(mbt) + ω(lct) ω(C−O) + τ(lct)(nph) + ω(mbt) τ(mbt) τ(mbt) δ(lct)(nph) + δ(COO) ω(lct)(nph) + τ(O)-(mbt) τ(nph) τ(lct) ω(mbt) τ(lct) τ(lct)
48
31
36
24
22
a
The experimental data are presented against the computational results obtained with harmonic plane-wave DFT calculations (PBE-TS). Tentative band assignment is given for each normal mode. blct - lactone; nph - naphtalene; mbt - methylbutanoate.
and that the band positions do not shift within the pressure range commonly accessible to the drug formulation processing. The INS spectra become typically blurred with temperature, which is due to the Debye−Waller factor contributing to the thermal background. Nevertheless, by comparing the set of data between 90, 200, and 280 K, some influence of molecular dynamics seems to be noticeable. This is particularly visible above 200 cm−1, where the bands due to τ(CH3) modes become less intense and partially split. The spectral range around ∼300 cm−1 is more interesting, as it clearly becomes considerably broadened. This can be attributed mainly to the methylbutanoate fragment and dynamics of the sec-butyl part within this moiety. Such a broadening may suggest that there is a noticeable increase of conformational disorder within the methylbutanoate fragment, which generally starts between 90 and 200 K. The δ(C−C−CH3) and ω(C−CH3) bands over ∼100−200 cm−1 remain generally unaffected by the thermodynamic conditions, as they are sparely interesting, while the terahertz spectrum (
