ADR-12886; No of Pages 14

Advanced Drug Delivery Reviews xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Advanced Drug Delivery Reviews

journal homepage: www.elsevier.com/locate/addr

Recent developments in the Raman and infrared investigations of

amorphous pharmaceuticals and protein formulations: A review☆
Alain Hédoux
Université Lille Nord de France, F-59000 Lille, France
USTL UMET UMR, 8207F-59655 Villeneuve d'Ascq, France

a r t i c l e i n f o a b s t r a c t

Article history: The success rate for drug discovery and the development of innovative therapeutic strategies are intimately re-
Received 15 October 2015 lated to the physical properties of the solid-state condensed matter, which have direct influence on the bioavail-
Received in revised form 26 November 2015 ability of Active Pharmaceutical Ingredients. In order to transform a new molecule in efficient drug, the material is
Accepted 30 November 2015
brought into an amorphous state using various manufacturing processes including freeze drying, spray drying,
Available online xxxx
hot melt extrusion and loading in different delivery devices. The infrared and Raman spectroscopic analyses
Keywords:
used for exploring disordered and amorphous states, for the monitoring of the drug physical stability in drug de-
Low-frequency Raman spectroscopy livery systems are described in this review.
Chemometrics © 2015 Elsevier B.V. All rights reserved.
Physical stability
Drug delivery systems

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
2. Description of the different kinds of motions in molecular systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
3. Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
3.1. Raman spectrometers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
3.2. Infrared spectrometers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
4. Spectrum analysis of molecular compounds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
4.1. Analysis of the low-frequency spectrum (LFRS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
4.2. Analysis of the high frequency IR or Raman spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
5. Spectroscopic signatures of amorphous states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
5.1. Amorphous solid state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
5.2. Amorphous liquid state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
6. Using LFRS to analyze disordered and amorphous solid states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
6.1. Structural description of disordered states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
6.2. Evidence of disorder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
7. Using IR and Raman spectroscopy for the in-process monitoring and for analyzing solid dispersions . . . . . . . . . . . . . . . . . . . . . . . 0
8. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0

1. Introduction
It is well recognized that the physical state of an active pharmaceu-
tical ingredient (API) is closely related to their bioavailability, and
☆ This review is part of the Advanced Drug Delivery Reviews theme issue on “SI:
Amorphous pharmaceutical solids”.
knowledge about its stability conditions are crucial to respect the drug
dosage. Numerous new drug candidates are characterized by a poor sol-
ubility in crystalline states and require amorphization in special dosage
forms for enhanced bioavailability [1]. However, the amorphous state is
metastable with respect to the crystalline state, and may spontaneously
recrystallizes under stress induced by manufacturing processes, or by
storage conditions [2–4] (humidity, temperature). The determination

http://dx.doi.org/10.1016/j.addr.2015.11.021
0169-409X/© 2015 Elsevier B.V. All rights reserved.

Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
2 A. Hédoux / Advanced Drug Delivery Reviews xxx (2015) xxx–xxx
of the stability conditions of amorphous formulations, requires to un-
derstand excipient — API interactions, and also the monitoring through
the several stages of processing as solid-dosage forms.
Raman and near infrared (NIR) spectroscopies are recognized as
non-destructive tools for both process analytical technology (PAT) [5,
6] applications and in quality control laboratories for stability analysis.
Both spectroscopic techniques enable the enhancement of process
knowledge for solid, liquid and biopharmaceutical forms [7, 8].
Combined with chemometric tools, they are very sensitive and com-
plementary indirect structural probes, which provide information on
molecular conformation and molecular packing in disordered and com-
plex systems.
Interaction between light and matter can result in absorption, scat-
tering or no interaction of the incident photon with the material
which may pass through it without any change. Absorption and scatter-
ing processes are presented in Fig. 1. Absorption requires that the ener-
gy of the incident photon corresponds to the energy gap between the
ground state and a higher energy level of the molecule. The process of
absorption spectroscopy is used in various techniques including infra-
red spectroscopy. In this technique, infrared energy covering a range
of frequencies is used as radiation. By contrast, the scattering phenom-
enon does not require that the photon energy matches the difference
between two energy levels of the molecule. Two types (elastic and in-
elastic) of scattering can occur, in which the energy photon is trans-
ferred to the molecule, promoted in a short-lived (called virtual)
vibrational energy level. In the case of elastic scattering, the photon en-
ergy is spontaneously re-emitted without energy change; this phenom-
enon is called Rayleigh scattering. In the inelastic scattering process the
energy photon is re-emitted with an energy lower or a higher than the
incident radiation energy, corresponding to the anti-stokes or stokes
Raman scattering. Raman scattering requires a monochromatic radia-
tion from a laser source. Infrared absorption and Raman scattering are
two vibrational spectroscopies providing similar information on molec-
ular vibrations. However, the use of two different types of radiation has
important consequence on the selection rules of the vibrational modes.
Infrared absorption requires that a vibrational mode of the molecule in-
duces a change in the dipole moment, i.e. in the charge distribution
within the molecule, to be IR active. In contrast the electromagnetic ra-
diation associated with the laser source momentary distorts the elec-
tronic cloud distributed around a bond within a molecule. As a result,
the vibrational mode of the bond will be Raman active if the vibration
induces a distortion of the electronic distribution around the bond,
and causes large polarization change, i.e. a polarizability change. The
Raman activity of vibrational modes is dependent on parameters
influencing polarizability, and marking pronounced differences with
IR spectroscopy. Symmetric vibrations induce the largest polarizability
change and give the greatest Raman scattering, while asymmetric vibra-
tions cause more significant change in the charge distribution and then
IR absorption activity. This difference is responsible for advantages or
Fig. 1. Description of Raman effect, fluorescence, resonant Raman effect, IR and NIR
absorption.
Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
disadvantages of one of two techniques, with interesting use in the do-
main of pharmaceutical applications. Water vibrations give an intense
contribution to the IR spectra, which forces to eliminate the presence
of moisture even in low proportion. Additionally, the analysis of aque-
ous solution by IR spectroscopy is highly perturbed by the stretching
and bending bands of water. On the other hand, Raman scattering can
generate specific excitations when the virtual states are overlapping
with excited states (Fig. 1). This can induce interesting enhanced
Raman intensity of specific vibrational bands (resonant Raman scatter-
ing) or intense interference signal (fluorescence) depending on the
timescale of the virtual state (~10−14 s for scattering and ~ 10−8 s for
fluorescence). Polarizability is proportional to the number of electrons,
to the molecular size, and inversely proportional to the bond strength.
Generally, Raman scattering intensity is much larger for π systems
than σ bonded structures, making Raman spectroscopy very sensitive
for analyzing APIs compared with excipients.
For many years the IR spectroscopy has been widely used with re-
spect to the Raman spectroscopy, largely due to the cost of the lasers.
Nowadays, the advances in notch-filter and diode laser technologies
have resulted in new Raman spectrometers accessible to non-expert
users, and this technique becomes industry standard with advantages
with respect to IR spectroscopy. The most significant advantage of
Raman spectroscopy is that no specific sample preparation is required
without care of moisture, making possible the analysis of tablets or
powders as obtained from manufacturing process, or almost in-situ
from in line manufacturing process. Raman spectroscopy gives the op-
portunity to investigate low-frequency domains down to ~ 2 cm−1,
while far-IR spectroscopy is limited to the 100–400 cm−1 range. It is
an undeniable advantage of Raman spectroscopy which makes it possi-
ble the analysis of low-frequency excitations distinctive of the amor-
phous state. Selecting the use of one of the two spectroscopic
techniques was discussed by De Beer et al. [5].
In this review, the capabilities of Raman and NIR spectroscopy to
carefully describe the physical state of different kinds of disordered sys-
tems, especially amorphous states and mixed amorphous — (macro,
micro, nano) crystalline states, will be shown. The contribution of IR
and Raman spectroscopy to the analysis of protein formulations in solu-
tions and in the solid state will be presented. The description of the THz
spectroscopy methods was deliberately omitted since it's the subject of
another review.
2. Description of the different kinds of motions in molecular systems
Considering a molecule as a rigid body, three different kinds of mo-
tions can exist in disordered organic solids, and only the access of the
low-frequency region (below 100 cm−1) gives the opportunity to ana-
lyze these different kinds of motions.
The internal motions correspond to vibrations of interatomic bonds
within the molecule. They are mainly distinctive of the molecular con-
formation in the fingerprint region, usually lying between 500 and
1800 cm−1, and corresponding to the region where the vibrations of
the molecular skeleton are active. Generally, a molecule composed of
N atoms has 3 N-6 vibrations, 3 N-5 for linear molecules. If the molecule
is placed in crystalline lattice, the 3 N-6 vibrations generates a number
of Raman and IR vibrational bands higher than 3 N-6, depending both
on the crystalline and the molecular symmetries, i.e. the so-called site
symmetry. The increase of the degree of disorder in a crystal induces
higher lattice symmetry, and thus less and broader vibrational bands.
Consequently, the spectrum of internal vibrations can be used as a sen-
sitive probe of the local molecular environment, i.e. to the degree of dis-
order, by counting the vibrational modes, or/and by analyzing the band-
shape of Raman or IR active modes. The spectrum of an amorphous state
corresponds to the spectrum of the isolated molecule, and then is
usually composed of 3 N-6 bands. Internal Raman or IR spectrum
makes it possible the evidence of molecular associations via hydrogen
bonds (H-bonds). Polymorphism in molecular compounds including
 A. Hédoux / Advanced Drug Delivery Reviews xxx (2015) xxx–xxx
pharmaceuticals is characterized by different types of molecular associ-
ations via H-bonding. Raman or IR spectroscopy can be used to deter-
mine molecular associations existing in the amorphous state in order
to predict and to analyze the crystallization mechanism. Additionally,
spectroscopy is a powerful tool to reveal the existence of H-bonding,
taking into account the limit of X-ray diffraction to accurately determine
H-atom positions.
At lower frequencies, two types of bands can be observed which are
mainly analyzed by Raman spectroscopy, since far-infrared spectrosco-
py is commonly limited down to frequencies around 100 cm−1.
The external motions between a molecule and its nearest molecular
neighboring. In ordered crystalline phases, external motions give rise
to the phonon spectrum corresponding to the lattice vibrations.
Disordering induces an inhomogeneous broadening of the phonon
spectrum, and the low-frequency spectrum of the amorphous state is
observed to be a very close representation of the vibrational density of
states, usually determined by inelastic neutron scattering. Molecular
compounds are characterized by the pronounced contrast between
the strong covalent intramolecular interactions and soft van der Waals
intermolecular interactions or hydrogen bond associations. This con-
trast is responsible for the frequency gap between internal and external
vibrations. Intermolecular interactions are responsible for specific phys-
ical properties of molecular compounds (polymorphism, low melting
temperature,…) while covalent bonds maintain the cohesion within
the molecule. Molecular flexibility favors disorder and reduces the gap
between external and internal vibration modes.
The semi-internal (or semi-external) motions mainly exist in disor-
dered states. These motions usually correspond to large amplitude rota-
tions of the molecule or a group of atoms within the molecule, giving a
Raman signature in the very low-frequency range (2–~ 70 cm− 1).
This low-frequency component corresponds to fast local motions, called
β-fast relaxational motions, typically on the picosecond timescale.
These motions are thermally activated and their temperature depen-
dence is closely related with the mechanism of phase transitions.
The simultaneous investigation of these three types of motions pro-
vides structural description of disordered states (crystalline, micro-
nano crystalline, and amorphous states), and the understanding of the
mechanism of phase transformation in relation with the molecular con-
formation. This type of analysis requires investigations down to the very
low frequencies (~ 5–2 cm−1) and is only possible by Raman scattering
and Terahertz pulsed spectroscopy well detailed in a previous review
[9]. The development of a new generation of filters makes it possible
the access to the low-frequency range with routine dispersive Raman
spectrometers. This feature opens important perspectives to Raman
spectroscopy for analyzing the physical state in molecular systems. A
significant part of this review will focus on the experimental con-
figuration and on the method of data analysis required to perform
low-frequency investigations in disordered states. Additionally, new
technical advances in Raman spectroscopy will be presented with
their expected contribution to the analysis of the physical state of phar-
maceutical materials.
3. Equipment
3.1. Raman spectrometers
Raman spectrometers are usually composed of four main compo-
nents: (1) the laser source of monochromatic radiation, (2) a sample il-
lumination system and light connection optics, (3) a device for the
analysis of the scattered light (spectrometer), (4) and a sensitive detec-
tor such as a charge-coupled device (CCD). Two kinds of spectrometers
can be used for collecting Raman spectra, the Fourier transform (FT-
Raman) or dispersive Raman spectrometers. The difference between
both types of spectrometer bases essentially on the device (3): a grating
dispersing system for dispersive Raman spectrometers, or an interfer-
ometer for FT-Raman spectrometers. Most of FT-Raman spectrometers
Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
3
use a Nd:YAG laser (1.064 μm) with a near-infrared interferometer
coupled to either a liquid nitrogen cooled germanium or indium gallium
arsenide detector. The main advantages are (i) the reduction in the
number of samples that exhibit laser-induced fluorescence, (ii) the
ease of operation, (iii) and the high spectral resolution with good wave-
length accuracy. However, the use of FT-Raman spectrometers presents
the disadvantage of being relatively unsuitable for sample environment
(temperature, pressure, humidity,…). Dispersive Raman spectrometers
use multi-channel (CCD) detectors, with extremely low intrinsic noise
and high quantum efficiency. The benefits for dispersive Raman are far
higher sensitivities and far lower detection limits than FT-Raman. The
shorter wavelengths of lasers associated with dispersive Raman in-
crease the sensitivity of the spectrometer, because of the ν4 scattering
efficiency dependency. As a consequence, the data acquisition times
are much shorter than for FT-Raman methods, that is important for
the study of fast transformation kinetics. The performances of dispersive
Raman spectrometers result from the adequacy between a good resolu-
tion and rejection of the laser line, and a sufficient scattered intensity,
imposed by the dispersive system. Modern instrumentation almost uni-
versally employs notch or edge filters for laser rejection, and spectro-
graph (monochromator or grating). However, in this case the analysis
of the low-frequency range (b150 cm−1) is generally prohibited. The
use of multiple dispersive stages (double and triple monochromators)
mounted with a long focal length allow the detection of Raman-active
modes with frequencies as low as 2–5 cm−1. New advances in ultra-
narrow band notch filter technology [10] with holographic gratings en-
able measurements of low-frequency spectra from a relative compact,
easy to use and cost effective system [11]. The development of this
kind of dispersive system composed of a single stage spectrograph
opens new perspectives for the qualitative and quantitative analysis of
the physical state of various pharmaceutical forms by Raman spectros-
copy with respect to IR spectroscopy.
Depending on the nature of the analysis (large volumes or micro-
scopic samples, or heterogeneities in multi-component systems) two
types of installation can be distinguished; the conventional installation
or the macroanalysis, and the installation micro Raman suitable for mi-
croanalyses. Detailed information on spatial resolution accessible by
micro-Raman investigations can be found in a recent review of Smith
et al. [12]. Raman investigations under pressure are typically performed
on microscopic samples located in diamond anvil cells [13, 14] (DACs),
and then micro-Raman installation is systematically used for pressure
analyses. For micro Raman, the sample holder is a XYZ table equipped
with a microscope composed of high numerical aperture objectives. Mi-
croscopic investigations are also very suited to analyze different kinds of
heterogeneities in samples with pharmaceutical applications: tablets
[15], extrudates [16], freeze-dried [17–19] formulations. Online analysis
can be also performed, by using a fiber optic probe, during the
manufacturing process such as holt-melt extrusion [20, 21] or freeze-
drying [22]. Micro Raman analyses can also be performed using a fiber
optic Raman probe to investigate the solid state transformation of hy-
drate forming drugs, in-situ during dissolution tests [23].
3.2. Infrared spectrometers
Infrared spectrometers can be schematically described by the basic
components including a light source, a monochromator, a sample and
a detector. Tungsten lamp, or more generally an inert solid that are heat-
ed to generate thermal emission of radiation in the infrared region of
the electromagnetic spectrum, are commonly used. The monochroma-
tor is used to separate the polychromatic broad IR spectrum into narrow
monochromatic lines. Different optical configurations exist, using grat-
ing or interferometer corresponding to dispersive or FT spectrometers
respectively. Detectors convert the analog spectral output into an elec-
trical signal. Different types of detectors are used characterized by dif-
ferent spectral domains of sensitivity and acquisition rates. The more
commonly used are silicon, and InGaAS, PbSe, InSb which are more
4 A. Hédoux / Advanced Drug Delivery Reviews xxx (2015) xxx–xxx
sensitive in the NIR domain. Detailed descriptions of the different types
of IR spectrometers can be found in specific literature dedicated to infra-
red spectroscopy [5, 24, 25].
4. Spectrum analysis of molecular compounds
4.1. Analysis of the low-frequency spectrum (LFRS)
In a crystalline state characterized by a unit cell composed of Z mol-
ecules, 6Z external motions are expected including 3 acoustic vibrations
corresponding to collective translation motions of molecules, not de-
tectable by Raman spectroscopy. The number of Raman active external
motions is usually lower than 6Z-3 and is depending on the crystalline
symmetry. LFRS is an indirect structural probe of the long-range order
in crystalline states. Disorder in the molecular environment of a mole-
cule induces an inhomogeneous broadening of phonon peaks, and the
LFRS of amorphous states results in the envelope of phonon peaks of
the underlying crystalline state. This broadening is related to the exis-
tence of different nearest molecular neighborings, i.e. different molecu-
lar cages.
The analysis of the low-frequency spectrum (LFRS) of disordered
molecular materials is very difficult, because of the overlapping contri-
butions of the quasi-elastic intensity (IQES) and the vibrational intensity
(IVIB), as shown in Fig. 2. However, it can be considered as containing
rich information (i) about the molecular organization in disordered
Fig. 2. Representations of the low-frequency spectrum of triphenyl phosphite: a) reduced
intensity and description of the fitting procedure of experimental spectrum. b) Raman sus-
ceptibility compared with the VDOS (G(ν)) obtained by neutron scattering.
Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
systems and (ii) on the mechanism of (order–disorder) phase transfor-
mation, confined in a narrow frequency range (3–200 cm−1). The
Raman intensity (IRaman(ν, T)) in the LFRS is related to the vibrational
density of states (VDOS, G(ν)) according to [26, 27]
½nðν; TÞ þ 1
IRaman ðν; TÞ ¼ CðνÞGðνÞ ð1Þ
ν
and
CðνÞ
χ}ðνÞ ¼ GðνÞ ð2Þ
ν
where C(ν) is the light-vibration coupling coefficient, and χ″(ν) is the
Raman susceptibility. To obtain the χ″(ν) spectrum (plotted in
Fig. 2b), which contains the structural information, the Raman intensity
is firstly transformed into reduced intensity Ir(ν) plotted in Fig. 2a using
IRaman ðν; TÞ
Ir ðνÞ ¼ : ð3Þ
½nðν; TÞ þ 1ν
In this representation the QES intensity is the dominant contribution
of the LFRS. This component is usually determined from a fitting proce-
dure using a Lorentzian shape centered at ν = 0. The vibrational compo-
nent is usually well described with a lognormal function. After
subtracting the QES component, Ir(ν) can be converted into Raman sus-
ceptibility by:
χ}ðνÞ ¼ ν  Ir ðνÞ: ð4Þ
The frequency dependence of C(ν) has been hotly debated [27–31].
Several studies have converged into a linear dependence of C(ν) sug-
gesting that χ " (ν) ∝ G(ν) [32–34]. This relation was corroborated by
the similarity between the χ″(ν) and G(ν) spectra plotted in Fig. 2b
the molecular glass-forming liquid TriPhenyl Phosphite (TPP) [33].
This figure confirm that χ″(ν) spectra can be considered as a close rep-
resentation of the VDOS, and demonstrates that the spectral resolution
of χ″(ν) spectra is significantly higher than that of G(ν) spectra. Fig. 3a
shows the temperature dependence of χ″(ν) spectra in TPP, plotted in
the liquid, undercooled liquid and glassy states. The temperature de-
pendence over about 100 K is relatively weak, reflecting the quasi-
harmonic behavior of collective motions.
The temperature dependence of the QES component is analyzed
using the fitting procedure described in Fig. 2. This contribution to the
LFRS corresponds to fast anharmonic motions, the so-called β-fast
relaxational motions. The Ir(ν) spectra are normalized by the integrated
intensity of the vibrational component, which has a weak temperature
dependence, and plotted at various temperatures (in the liquid,
undercooled liquid and glassy states) in Fig. 3b. This figure indicates a
strong temperature dependence distinctive of anharmonic motions,
contrasting with Fig. 3a. IQES(T) is similar to the temperature depen-
dence of the mean-square displacement (MSD) [35] bu2N (T), usually
determined by neutron scattering experiments. The temperature de-
pendence of these kinds of anharmonic motions is commonly deter-
mined as the driving force in phase transitions. Fragile molecular
liquids can be distinguished from strong liquids by comparing the tem-
perature dependence of MSD [35], and assuming the empirical relation
1= u2 ðTÞ∝ logηðTÞ ð5Þ
experimentally determined in different glass-forming systems [36, 37].
The temperature dependence of logη is plotted in the inset of Fig. 3b
with the fitting curve corresponding to the Vogel–Fulcher–Tammann
(VFT) function given below:
expðA  T0 =ðT−T0 ÞÞ ð6Þ
 A. Hédoux / Advanced Drug Delivery Reviews xxx (2015) xxx–xxx
Fig. 3. Temperature dependence of the LFRS of TPP. a) In the representation of reduced in-
tensity; the inset shows the temperature behavior of bu2N versus Tg/T. b) In the represen-
tation of Raman susceptibility.
where A reflects the curvature of the plot distinctive of the fragility of
the system.
4.2. Analysis of the high frequency IR or Raman spectrum
Raman investigations are commonly performed at frequencies
higher than 200 cm−1, i.e. in the domain which does not require any
specific high-dispersive equipment (triple monochromator) used for
the high rejection of the elastically scattered light. This region can be
also easily analyzed using routine IR spectrometers. For rigid molecules,
internal modes are detected above 150–200 cm−1. Two regions are usu-
ally distinguished in the spectrum of internal vibrations. (i) the finger-
print region covering the 500–1800 cm−1 range, distinctive of the
molecular conformation and (ii) the intramolecular C–H, N–H, O–H
stretching region, lying from 2200 up to 3800 cm−1, recognized to be
very sensitive to molecular associations via H-bonding [38], OH
and NH groups being the most commonly involved in intermolecular
H-bonding. In these two regions, no intensity correction is required,
since the Bose–Einstein factor becomes close to 1 above 100 cm− 1.
However, to obtain information on the formation of H-bonds, related
to the Raman intensity of X–H stretching bands [38] (with XO, N, C),
the spectrum must be normalized with respect to Raman bands which
are temperature independent.
The fingerprint region is widely used to characterize molecular con-
formations both in small API molecules and biomolecules, to monitor
Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
5
the unfolding process in protein formulations, to identify polymorphism
and to distinguish amorphous from crystalline states. Univariate spec-
tral analyses are based on peak height at a specified frequency of the un-
treated Raman data. The selected spectral response must be not biased
by contributions from other sources (presence of water, buffer compo-
nents…). Multivariate data analysis (MVA) takes into account the
whole spectra of information, and is widely used to reveal very weak
changes in the molecular conformation or in the structure of the physi-
cal state.
Data pretreatments are the first stage in MVA. This preprocessing
stage consists to eliminate or to minimize the effect of various uncon-
trollable parameters on the intensity. The baseline correction is com-
monly performed, since its contribution may be different from one
spectrum to another. Many methods exist for baseline correction [39],
second-order derivatives and polynomial fitting, being the most com-
monly used. In a second step a normalization procedure is performed
in order to correct the overall intensity variations which could be in-
duced by weak differences in focusing distances because of the rough-
ness of the sample surface. This correction aims to remove intensity
differences unrelated to the sample composition. The most widely
used normalization procedure is the standard normal variate (SNV)
where a normalized spectrum is obtained by subtracting the mean of
its spectrum and divided by the standard deviation.
MVA consists to used newly defined variables that are linear combi-
nations of the original ones, and are mutually independent. The aim is to
reduce the number of correlated variables existing in IR or Raman spec-
tra in a few uncorrelated variables containing the relevant information
for a qualitative or/and quantitative analysis of inhomogeneous sam-
ples… The best known and probably most widely used method to deter-
mine the number of linear independent components in a given system
is the principal component analysis (PCA).
The basic idea of PCA is to represent the original data matrix D by a
product of two smaller matrices T and P, respectively the scores matrix
and the loadings matrix, according to:
D ¼ T  PT ð7aÞ
schematically presented in Fig. 4a. PCA builds a new coordinate system,
called principal components (PC), as a linear combination of the old co-
ordinates (Fig. 5). D matrix has the dimensions of number of spectra
(n) by intensity at particular frequencies (m). The score matrix has
the dimensions of the number of spectra (n) by principal components
(p). The loading matrix has the dimensions of intensity (m) by principal
components (p). The most important information is contained in the
first two or three PCs. The rest of PCs mainly corresponds to noise, and
Fig. 4b describes the decomposition of D matrix depending on the num-
ber of significant PCs, according to:
D ¼ T  PT þ E: ð7bÞ
Fig. 4b shows that the loading matrix can be decomposed into load-
ing spectra which look like spectra composed of negative and/or posi-
tive bands representing spectral information negatively or positively
correlated through particular PC. The presence of negative bands
makes it difficult the interpretation of loading spectra in terms of chem-
ical information. In order to achieve loadings more representative of
chemical information, different MVA methods have been developed.
The widely used MCR technique is a two-step multivariate regression
method, in which PCA of the data is firstly carried out, in order to deter-
mine the number of components and their contribution in the sample,
used to decompose the D matrix. In a second step various regression
methods can be used, including alternating least squares (ALS) method
[40] which is probably the most widely used. Using the ALS algorithm, D
matrix is decomposed in such a way that a component corresponds
chemically to relevant distinctive feature of the sample. Several con-
straints are applied, especially the non-negative constraint [40, 41]. No
6 A. Hédoux / Advanced Drug Delivery Reviews xxx (2015) xxx–xxx
Fig. 4. Schematic diagrams for describing the decomposition of data matrix: a) as a function of the score data matrix. b) as function of the significant PCs and noise.
prior information is required in this method. The T matrix represents
the distribution of the components contained in the sample, whereas
the PT matrix corresponds to the spectra of components.
When the pure components are known and their spectra can be col-
lected, direct classical least square (DCLS) method [41] can be applied.
In DCLS method the PT matrix (spectra of pure components) is available.
DCLS method finds the linear combination of spectra from the pure
components that most closely matches the Raman spectra collected
during the mapping (D matrix). Quantitative information can be obtain-
ed by calculating the concentration matrix (T matrix) according to:
 −1
T ¼ DP PT P : ð7cÞ
DCLS is undoubtedly the most unambiguous method to display the
distribution of components from a Raman mapping data collection.
Using Raman micro-spectroscopy for quantitative analyses in a
multi-components system may have some limitation because of sub-
sampling [42]. The backscattered Raman signal is mainly dependent
on the size of a laser spot (Ø ~ 100 μm). Consequently, Raman spectra
collected in this spot size are representative of only a small area of the
sample. Different methods have been developed to avoid subsampling,
described in two reviews [6, 43]. It was shown [44] that subsampling
can be reduced by moving the sample over a larger area during the
data collection. This method was tested on the quantitative analysis of
tablets with satisfactory results [44]. A second method consists to in-
crease the spot size. The so-called wide area illumination (WAI) method
was used for a quantitative determination of a drug in liquid formula-
tions by direct measurement through a plastic bottle [45], in solid for-
mulations in an intact capsule [46] and in tablets [47]. In these
methods only the surface of samples are mainly analyzed. Transmission
Raman spectroscopy becomes widely used for investigation of drugs
more deeply inside the different kinds of samples [48, 49] including cap-
sules [49]. Spatially offset Raman scattering is also a technique leading
to more representative sampling volumes with pharmaceutical applica-
tions [50]. In this method the imaging optics collecting the scattered
light are moved away from the illuminated area [6, 42, 43, 51]. However,
in these different configurations, the microscopic information about
heterogeneities in the sample are lost.
Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
5. Spectroscopic signatures of amorphous states
The universal signature of a solid amorphous state is an excess of
VDOS in the low-frequency region, mainly detected by inelastic neutron
scattering and LFRS. This enhancement of the low-frequency intensity
of the VDOS in the amorphous states reflects the deviation from the ν2
behavior below 50 cm− 1, i.e. the deviation of the Debye behavior of
VDOS commonly observed in crystalline states. The excess of VDOS
gives rise to a low-frequency peak, the so-called Boson peak (BP),
which is considered as the universal signature of a solid amorphous
state [30, 52]. Among the many controversial interpretations of the ori-
gin of the BP, an emerging hypothesis is related to the confinement of
molecular vibrations within domains formed by different nearest mo-
lecular neighborings [52–55]. The designation of BP is extended to the
low-frequency band detected in liquids, corresponding to the “caging
effect”, i.e. the frequency distribution of vibrations that molecules expe-
rience within the cage formed by their neighbors. The frequency of BP is
inversely proportional to the size of these domains [52] both in glasses
and liquids. The origin of BP is only related to the disorder existing in
the amorphous state, and should have no correspondence in the low-
frequency spectrum of the underlying crystalline state.
5.1. Amorphous solid state
The detection of BP in the solid state is highly depending on the
physical properties of the liquid. It was shown that the LFRS can be
used to identify the nature of the liquid state (fragile or strong in the
Angel classification [56]), and estimate the degree of fragility of a liquid
[35]. The ratio between the maximum and minimum reduced intensity
([Ir]min/[Ir]max) reflects the relative contributions of relaxation and vi-
bration to the low-frequency spectrum. Fig. 6, plotted from data pub-
lished by Sokolov et al. [35], shows that this ratio is significantly
higher for a fragile liquid (see LFRS of m-tricresyl phosphate — m-TCP,
Fig. 6b) than for stronger liquid (see LFRS of glycerol in Fig. 6a). This fea-
ture is usually considered as an empirical criterion to characterize the
nature of the liquid. Comparing Fig. 6a and b clearly reveals that BP is
easily detected for strong liquid whereas it is rarely observed for most
of fragile liquids, i.e. in most of organic molecular glasses. An alternative
to detect BP is to plot χ " (ν)/ν2 to point out the deviation of χ " (ν) from
 A. Hédoux / Advanced Drug Delivery Reviews xxx (2015) xxx–xxx
Fig. 5. Relation between spectroscopic data and spectrum projection in the PCs coordinates.
the ν2-behavior. The LFRS has much unexplored potential for sensitive
detection of low crystal content in solid dosage forms. However, there
are only a few studies performed in the domain of pharmaceutical sci-
ences. X-ray powder diffraction is often used, since amorphous material
is characterized by a halo background and crystal by Bragg peaks, pro-
viding clear and direct information on the nature of the physical state.
Low-frequency Raman spectroscopy also offers the same type of con-
trast between the spectrum of phonon peaks in the crystalline state
and the very broad hump of the Raman susceptibility, as shown in
Fig. 7 for indomethacin (IMC). At higher frequencies, Raman active
bands in amorphous materials exhibit broadening and frequency shift
of Raman bands active in crystals, and there is no signature distinctive
of an amorphous state, making it distinguishable from a very disordered
crystalline state. Additionally, it was shown that the LFRS was much
more sensitive to detect and to identify the very early stages of crystal-
lization from a ground amorphous powder of IMC than the high-
frequency range [57]. A quantitative analysis has led to determining a
very accurate kinetics curve, making possible the interpretation of the
crystallization mechanism as corresponding to a two-dimensional (sur-
face) crystallization from pre-existing nuclei of γ-IMC. This result sug-
gests that grinding the γ-phase reduces the size of crystallites into an
amorphous powder, preserving nuclei of γ-phase. The LFRS is also
very suited to the detection of a very small amount of amorphous via
the detection of BP. Fig. 7 shows the different representation of the
LFRS of the different physical states in IMC. It is observed that BP is de-
tected in the very low-frequency region, where no phonon peak of the
crystalline states is Raman active. The analysis of LFRS at different stages
of grinding γ-IMC [58] has revealed very small amount of amorphous
material after only two minutes of grinding, via the detection of a
clear enhancement of χ″(ν) below 20 cm−1 localized by an arrow in
Fig. 7e. Applying an analytical method, which can be only used from
Fig. 6. Description of the criterion used to characterize the degree of fragility of liquids in
two cases from data taken in Ref. [34]: a) for a relatively strong molecular liquid: glycerol.
b) for a fragile molecular liquid: m-TCP.
Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
7
the knowledge of the whole low-frequency spectra of crystalline and
amorphous states, about 1% of amorphous content can be detected in
a crystalline matrix, idem for the degree of crystallinity detected by
analyzing a ground powder of IMC. A chemometric analysis [59] from
FT-Raman investigation in the CO stretching region, has led to the deter-
mination of 2% of amorphous or crystalline IMC. Such a small amount of
amorphous material can be determined only because of the high sensi-
tivity of 1697 cm−1 band to the lost of long-range order, especially in
the γ-form in which the band is very intense and sharp. Several other
groups have used MVA analysis in combination with high-frequency
Raman spectroscopy to quantify crystallinity in amorphous systems
[59–61], but a recent paper [62] reports the first study of a low-
frequency investigation combined with a MVA compared with classical
high-frequency methods to quantify crystallinity in amorphous
grisefulvin tablets. This paper clearly demonstrates that LFRS is very
useful in the quantification of crystallinity, regarding higher frequency
regions analyzed by Raman or IR spectroscopy.
The LFRS was recently used [63], combined with broadband
terahertz time domain spectroscopy to determine the type H-bonding
in the glassy state of IMC from analyzing intermolecular vibrations.
This study has revealed the existence of H-bonded cyclic dimers in the
glassy state of IMC and the intermediate correlation length of the glassy
structure was estimated to about 2.5 nm from analyzing BP in glassy
IMC.
Different preparations (quenching the liquid below the temperature
of glass transition (Tg), milling crystals below Tg, pressurizing, freeze-
drying) can produce amorphous forms with different energies resulting
in variable properties, including bioavailability and the stability condi-
tions. The classical experimental methods are relatively unsuitable for
the direct characterization of different amorphous forms. PCA method
was used by Karmwar et al. [64] to analyze spectral variation between
the different amorphous states of indomethacin prepared via different
transformation methods: transformation via melt quenching, spray dry-
ing and by ball- and cryo-milling the α and γ crystalline forms. Spectra
of each sample have been recorded in the 1000–1720 cm−1 and 2800–
3100 cm−1 regions including the CO stretching region containing a sig-
nature distinctive of each state of IMC. Karmwar et al. [64] have shown
the existence of three clusters in the scores plot resulting from PCA anal-
ysis, that mirror the differences observed in the diffractograms of the
differently prepared samples. Despite a strong similarity between spec-
tra collected during a Raman mapping procedure, this study demon-
strates the capability of PCA to point out subtle structural differences
in amorphous IMC samples prepared by different transformation
methods. Sample differences have been interpreted using the spectral
loading plots and the assignment of Raman bands [65] in the spectral
regions where spectral differences are observed, suggesting that differ-
ent amorphous states could be related to different intermolecular
associations.
Similar analysis using PCA method was applied to IMC glassy states
prepared by cooling the liquid with different cooling rates [66]. This
8 A. Hédoux / Advanced Drug Delivery Reviews xxx (2015) xxx–xxx

Fig. 7. Representations of the low-frequency Raman spectrum of the different states of indomethacin: a) Reduced spectrum (Ir) of the liquid state at 443 K, the glassy state at room tem-
perature obtained by grinding (ground amorphous state, GAS) and from quenching the glassy state (thermal glass, TG). b) Raman susceptibility (χ″) of the liquid state, TG and GAS sam-
ples. c) Raman susceptibility divided by ν2 to point out the deviation from the ν2-behavior. d) Raman susceptibility of the solid states: glassy state, and α-form, δ-form, γ-form. e) Raman
spectra collected at various times of grinding from γ-phase; the insert displays the very low-frequency range of spectra collected at time t = 0, 2 min, 4 min, and 6 min to show the en-
hancement of the Raman susceptibility in the very low-frequency range; the arrows localize this low-frequency signature of the amorphous state.

analysis has revealed significant structural changes in the different
amorphous samples, in the fingerprint region and the high frequency
spectrum, suggesting different local organizations depending on the
cooling rate used to undercool the liquid until the glass.
Pressure-induced amorphization, using MDAC can be easily moni-
tored using Raman micro-spectroscopy. It was shown from a low- and
high-frequency investigation that two different amorphous forms of
IMC can be produced by a thermal route (quenching the liquid below
Tg) and by pressurization [14]. This feature combined with the evidence
of two different mechanisms of amorphization, revealed by Raman in-
vestigations performed during pressurization and at different stages of
cryogrinding, suggests the existence of a polyamorphism [67] in IMC,
i.e. the existence of two different amorphous forms linked via a first-
order transition, as observed in water. Application of high pressure
could place molecular compounds in an amorphous state characterized
by higher stability than an amorphous state obtained by simple quench

Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
 A. Hédoux / Advanced Drug Delivery Reviews xxx (2015) xxx–xxx
of the liquid as observed in ibuprofen [68]. Raman analyses under high
pressure can provide interesting information about mechanisms re-
sponsible for possible enhanced stability of amorphous states.
5.2. Amorphous liquid state
The Analysis the liquid state has important applications in the bio-
pharmaceutical area, since a considerable amount of vaccines including
influenza vaccines are formulated as liquids. In aqueous solution they
are subjected to physical and chemical degradation, and then were
stored and distributed under refrigeration. One strategy is to bring liq-
uid formulations into the dry state using suitable bioprotectants and
drying biotechnologies [69, 70]. An alternative to the use of these very
expensive drying processes is to stabilize the aqueous formulations at
room temperature and against temperature variations during storage
and shipping. In this context deciphering the thermostabilization mech-
anisms of protein in aqueous solutions is probably a necessary step to
determine innovative methods for stabilizing vaccines in liquid solu-
tions. Raman and IR spectroscopy are widely recognized as suitable
techniques for monitoring protein unfolding and were used to analyze
the mechanism of protein denaturation/stabilization.
There is no difference between the spectra of internal modes in the
liquid and solid amorphous states. In the low-frequency region, Fig. 7
clearly shows significant changes between the Ir spectra of liquid and
solid amorphous IMC, mainly resulting from the strong contribution of
relaxational motions in the liquid state. χ″(ν) spectra of liquid and
solid amorphous states are very similar and only distinguishable be-
cause of the slight temperature dependence of the VDOS. The spectra
are dominated by a broad band reflecting the molecular interactions
of a molecule with its disordered nearest neighboring. This indicates
that the short-range order is similar in the liquid and glassy states. In
this context, it can be considered that the caging effect in the liquid
state and BP in the solid amorphous state have probably the same origin.
Low-frequency Raman spectroscopy has been widely used to analyze
solution of proteins [8]. These solutions are mainly composed of
water, and the analysis of protein solutions is based on the description
of the spectra of water and dry proteins. The analysis of the low-
frequency spectrum of a protein formulation is described in Fig. 8. Con-
trasting with the spectrum of IMC, χ″(ν) spectrum of water is composed
of two broad bands [71]. The low-frequency band (I) was assigned to
the caging effect [71–74] and the high frequency band (II) is associated
with the intermolecular O–H stretching vibrations in the H-bonded
Fig. 8. Raman susceptibility of aqueous solution of lysozyme (LW) compare to that of
water (W) and dry lysozyme (dry-L).
Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
9
network [71–74]. Addition of a protein, e.g. lysozyme, has two conse-
quences on the χ″(ν) spectrum of water [75–78]. (i) A shift of band
(I) reflecting a change in the close neighboring of a water molecule,
and (ii) an intensity decrease of band (II) induced by the disruption of
the H-bonding network of water molecules. Comparison of χ″(ν) spec-
tra of water, dry protein and protein solution [76] in Fig. 8 shows that
band (I) mainly reflects the dynamics of the protein with a weaker con-
tribution of the coupling between dynamics of protein and solvent. On
the other hand, band (II) gives the opportunity to probe the organiza-
tion of the H-bonding network in the solvent. Moreover, it was shown
that the temperature dependence of low-frequency region of the Ir(ν)
spectrum was similar to that of the mean square displacements (MSD,
bu2N). The combination of the analysis of the low-frequency region
and that of the amide I band, distinctive of the secondary structure [79],
provide a detailed description of the thermal denaturation process [80].
An important result is the discrimination between the transformation of
the native tertiary structure into a molten globule state, detected from a
break in bu2N (T) [35, 37, 38], i.e. in the temperature dependence of
Ir(ν), and the unfolding process of the secondary structure detected by
a frequency upshift of the amide I band. The combination of low and
high-frequency Raman investigations is the first experimental method
providing a relation between the dynamic transition at high tempera-
ture and the thermal denaturation process of proteins in solution. The
analysis of the temperature dependence of χ″(ν) spectrum has shown
the influence of the H-bond network of water on the denaturation
mechanism. Same types of analysis in presence of denaturants [81]
(urea, hydrochloride guanidine) and sugars [77, 78, 80, 82] (disaccha-
rides) have led to the understanding of protein stabilization mecha-
nisms in aqueous solution, against various kinds of stress [13, 80, 83]
(low and high temperature, high-pressure). The influence of the differ-
ent types of solutes (urea, guanidine hydrochloride, disaccharides) on
the H-bond network the solvent was clearly shown by Raman spectros-
copy. The preferential hydration [84, 85] of proteins in presence of di-
saccharides [86] or other sugars [87] was clearly evidenced at room
temperature [80]. The use of D2O instead of H2O and deuterated treha-
lose, has led to a detailed description of sugar – protein – water interac-
tions, suggesting that the preferential hydration hypothesis is not valid
to explain biopreservation mechanisms of disaccharides against high
temperature [80]. Combining Raman spectroscopy with molecular dy-
namics simulations [71, 75, 88] has bring out crucial information at
the molecular level, about the influence of disaccharides on the
hydrogen-bond network of water, explaining the exceptional capabili-
ties of trehalose to stabilize proteins.
Investigations for characterization of therapeutic proteins at high
concentration were recently performed using a Raman spectrometer
and dynamic light scattering system combined in single platform [89].
This kind of study provides information on the protein secondary and
tertiary structures and hydrodynamic size in real time during heating,
making possible the rigorous comparison between structural changes
and protein aggregation.
6. Using LFRS to analyze disordered and amorphous solid states
6.1. Structural description of disordered states
Forms I and II of anhydrous caffeine (C8H10N4O2) constitute an
enantiotropic system [90]. The commercial form II is thermodynamical-
ly stable at room temperature and transforms upon heating in form I at
153 °C [91]. Form I is an orientationally [92, 93] and dynamically [93]
disordered phase, called rotator phase, where molecules are slowly ro-
tating around their molecular C6 axis. Such a rotator phase is considered
as dynamical analogous of under-cooled liquid state [94]. Refinements
of X-ray powder data collected in form I [92] have conducted to a struc-
tural description of the molecular packing of caffeine molecules, very
similar to that determined in theophylline [12]. The main differences
between both compounds belonging to the family of methylxanthines,
10 A. Hédoux / Advanced Drug Delivery Reviews xxx (2015) xxx–xxx
is the absence of disorder and a tilt of molecules out of the molecular
plane in theophylline. Form II is quite intriguing, since despite more
than 50 years of investigations [95] its structural description remains
unclear. Indeed, the last refinements from X-ray powder diffraction
data [96, 97] performed without any consideration of disorder,
have led to the determination of an unusually very large unit cell
(~ 4277 Å3) composed of 20 molecules, suggesting that the existence
of disorder similar to that existing in phase I [98] should be considered.
After purification by cold sublimation, caffeine is in the metastable form
I at room temperature. The I → II transformation is very hindered
around room temperature and faster transformations are observed
around 90 °C [99, 100]. Consequently, Raman spectra of both caffeine
forms are directly plotted in Fig. 9 at 90 °C, after isothermal transforma-
tion of form I into form II. Fig. 9 provides original and important infor-
mation on structural description of disordered phases. First, the LFRS
of form I is composed of a single and broad band and characterized by
the absence of phonon peaks, contrasting with the observation of
sharp Bragg peaks in the X-ray powder diffraction diagram of form I
[92, 97]. Bragg peaks only result from the periodic organization of the
molecular mass centers, while lattice modes are related to intermolecu-
lar atom-atom potentials. Consequently, the orientational molecular
disorder induces an inhomogeneous broadening of phonon peaks corre-
sponding to a librational density of states [100]. χ″(ν)-spectrum of form
II is characterized by a splitting of the broad low-frequency band
into two components [100] which are too broad to be considered as
Fig. 9. Raman spectra of crystalline phases in caffeine in different representations.
a) Raman susceptibility; the inset shows the close relationship between Rama spectra of
phase II and theophylline. b) Raman intensity of crystalline and liquid phases of caffeine.
Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
phonon peaks, suggesting the existence of a substantial disorder. Sec-
ond, both polymorphic forms of caffeine can be only distinguished
below 100 cm− 1; the fingerprint regions of both forms, plotted in
Fig. 9b, are very similar and does not display highlights with respect to
that of the liquid. These considerations indicate similar local order in
both forms, and then similar description of the disorder in forms I and
II. The close relationship between the fingerprint regions of polymor-
phic phases and the liquid state confirm the high degree of orientational
disorder of caffeine molecules. χ″(ν) spectra of theophylline and form II
of caffeine are plotted in the inset of Fig. 9a at room temperature. This
figure reveals that the Raman susceptibility of form II of caffeine corre-
sponds to the rigorous envelope of the low-frequency phonon peaks in
theophylline. The close relationship between the LFRS of caffeine and
the low-frequency region of the phonon spectrum of theophylline can
be related to the predominant contribution of librational motions in ro-
tator phases. The relationship between both spectra indicates that in
form II, caffeine molecules are located in positions similar to those
occupied by theophylline molecules, and rotate around their molecular
axis, probably more slowly than in form I, in agreement with previous
studies [98]. The I → II transformation can be interpreted as the tilt of
molecules out of the hexagonal plane in the lattice of form I, giving mo-
lecular positions similar to those determined by last X-ray investiga-
tions [96, 97].
6.2. Evidence of disorder
It was shown that racemic ibuprofen (IBP) transforms into a meta-
stable crystalline phase (phase II) upon heating from a deeply quenched
state [101]. Only LFRS has revealed a substantial disorder in phase II
[102], not directly detected by X-ray diffraction [103]. The χ″(ν) spec-
trum of phase II was observed typical of an amorphous state. However,
Raman features distinctive of the transformation of undercooled IMC
into phase II correspond to the signatures of the stable phase I. The spec-
trum of phase II was interpreted as reflecting the VDOS of the amor-
phous state where the first trace of the long-range order of phase I are
detected. Phase II was then interpreted as an early stage in crystalliza-
tion process in agreement with the “Oswald rule of stages” [104].
7. Using IR and Raman spectroscopy for the in-process monitoring
and for analyzing solid dispersions
It is well known that pharmaceutical manufacturing may induce dis-
order and (expected or not) amorphization. Grinding, holt-melt extru-
sion, spray-drying, freeze-drying are often used in order to improve
bioavailability of poorly water soluble drugs. Freeze-drying is a standard
but costly procedure, and then mainly used for preserving biological
systems in the solid state, including various types of vaccines which
are marginally stable in solutions.
During a freeze drying cycle proteins are exposed to different ex-
treme conditions (low temperature, pH, dehydration), and protective
additives whose effects are still not completely understood are used to
alleviate these conditions. Several hypotheses have been proposed
to explain the preservation of protein functions in the dry state
[105–112]. Raman and IR spectroscopy can be used to get a better un-
derstanding on biopreservation mechanisms in the dry state. Dong
et al. [113] have analyzed by Raman microscopy frozen protein formu-
lations prepared using various conditions of supercooling the solvent,
to determine the distribution of the excipient in the frozen medium in
relation with secondary structural changes and aggregation. Both IR
and Raman spectroscopy are used as process analytical (PAT) tools [5,
114–116], especially to monitor the primary drying by determining
the endpoint from the analysis the sharp and intense OH stretching
band in ice. De Beer et al. [22] have clearly demonstrated the comple-
mentarity of NIR and Raman spectroscopy as PAT tools during a freeze
drying cycle. Pressure-temperature devices can be used to mimic a
freeze drying cycle on microscopic volume, analyzed by Raman
 A. Hédoux / Advanced Drug Delivery Reviews xxx (2015) xxx–xxx
microspectroscopy to monitor simultaneously the secondary structure
of the protein via the analysis of the amide I band and the freeze-
drying process [18, 19, 117]. This analysis has shown that the water de-
sorption during the primary stage of freeze drying was the stress re-
sponsible for a distortion of the protein structure, in absence of
stabilizing agents. In presence of sugar (trehalose), Raman mapping
performed at the end of the first (congelation) and second (primary
drying) stages of freeze drying, have shown that the rigidity of the
amorphous water-trehalose matrix was the main stabilizing effect of
trehalose to preserve the secondary structure of the protein during the
freeze-drying process, in agreement with the vitrification hypothesis
proposed by Green et al. [108]. The results of these investigations are
in agreement with IR experiments [118], suggesting that embedding
the protein in a rigid amorphous water-disaccharide matrix mainly ex-
plain the stabilization of the protein, making it less flexible. Raman spec-
troscopy also gives the opportunity to monitor the physical state of
excipients during a freeze-drying cycle which can be crucial for the
properties of the freeze-dried product. Cao et al. [17] have analyzed
the influence of the operating parameters on the transformations of
mannitol. The presence of remaining water at the end of the freeze-
drying process is crucial because of the plasticizing effect of water. IR
spectroscopy is very sensitive to a small amount of moisture, and is
used to monitor the presence of water in the amorphous freeze-dried
product. IR spectroscopy has been widely used to analyze freed-dried
products since Raman spectroscopy requested high laser power, re-
sponsible for protein degradation. Nowadays, the development of de-
tector technology makes possible analyses freeze-dried products with
low laser power and short time of laser exposure. Consequently, IR
and Raman spectroscopy combined with MVA are widely used to deter-
mine slightly changes in the secondary structure of freeze-dried pro-
teins [119]. A quantitative estimate of the protein denaturation can be
obtained using both techniques, by calculating a degree of overlap
area of the amide I band collected in the denatured and native states
[19, 120]. Raman spectroscopy was also used for in-situ secondary
structure determination of proteins in ice for frozen storage [121].
The demand for drug delivery systems is gradually growing, and the
use of IR and micro Raman spectroscopy is crucial to understand the ki-
netics of drug release. Understanding the mechanisms of drug dissolu-
tion and drug release requires the knowledge of the component
distribution within delivery device. Raman mapping is widely used for
providing the characterization of distribution and the solid state of the
drug [122–125].
Hot melt extrusion (HME) is now widely used for the rapid produc-
tion of solid dispersions. Wilson et al. [126] have used Rama spectrosco-
py as PAT tool for monitoring of the concentration of drugs in the
extrudate, of the physical state of extrudates, and for analyzing drug —
excipient interactions. Saerens et al. [127, 128] also used Raman spec-
troscopy for in-line monitoring of the physical state of extrudates, for
various formulations, and various operating parameters. Raman spec-
troscopy gives direct and rapid information to choose operating condi-
tions of HME and formulation to obtain amorphous extrudates.
Haaser et al. [129] have used Raman microscopy to investigate dif-
ferent extrudate formulations in terms of component distribution and
structural changes during dissolution testing. Raman maps of cross-
section of the extrudate corresponding to a binary system, based on a
tripalmitin matrix containing equal amount of the model drug theoph-
ylline anhydrate, were performed after a dissolution period of 120 min.
It was shown that the distribution of the drug particles in some places
completely surrounded by lipid, made dissolution impossible.
Grinding is basically used to enhance bioavailability of poorly water
soluble drugs by reducing the particle size of powder. The resulting
ground powder can be amorphous or not, mainly depending on the
temperature of grinding (below or above Tg) and other operating pa-
rameters (intensity, time) [130]. Raman spectroscopy is a fast acquisi-
tion technique and does not require any specific sample preparation,
giving the opportunity to analyze almost in-line the solid-state
Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
11
amorphization mechanism by collecting Raman spectra at different
grinding times. Using this method, it was shown that the amorphization
of IMC by grinding directly results from the “easy” breaking of the
H-bond network of the γ-form (after short time of grinding), while
it was observed that this H-bond network remains stable at very high
hydrostatic pressures and collapses in a high density state prior
amorphization [14].
It is recognized that amorphous drugs are significantly more soluble
than their crystalline counterparts [131]. However, a major issue
associated with amorphous drugs is their inherent low stability, leading
to unexpected crystallization. It was found that co-grinding a drug
with a polymer enhances the physical stability of the drug (e.g.
indomethacin-PVP [65, 132–134]), without detecting intermolecular
interactions between drug and polymer [57, 135]. Polymers enhance
the physical stability during the storage and also during dissolution
when the drug is introduced to aqueous media. Raman spectroscopy
was used to detect phase transformation of amorphous pharmaceuticals
exposed to an aqueous environment, to understand the behavior of
amorphous pharmaceuticals during dissolution, in presence or not of
polymers [136]. More recently co-amorphous drugs were found to be
promising for both enhancing solubility and physical stability of the
amorphous solid state [137]. FTIR was used to determine the origin of
the stability enhancement [138, 139]. The use of density functional the-
ory calculations to analyze the IR spectra of naproxen-indomethacin
glassy systems has revealed the formation of heterodimers (naproxen-
indomethacin) in a single phase co-amorphous mixture [140].
It is recognized that grinding usually induces disorder [130]. Intrigu-
ingly, it was shown from LFRS investigations, that under grinding at
room temperature, metastable form I transforms into form II, and
form II transforms into form I [141]. After a long enough grinding time
of form I or form II, a metastable driven state is reached, only stabilized
under grinding. This metastable state was described by LFRS as corre-
sponding to a partially transformed form I into form II or inversely as
a partially transformed form II toward form I.
8. Conclusion
It was shown that IR and Raman spectroscopy are very suited for
characterizing the physical state, the distribution of pharmaceuticals
and biopharmaceuticals in drug delivery systems, and then is a crucial
aid for predicting the kinetics of drug release and drug dissolution. Now-
adays, Raman spectroscopy also focuses on interactions of cells with
drug molecules and carrier systems [142]. Interactions of drugs with
cells in suspension in an aqueous media can be analyzed by using optical
tweezers for cell trapping [143–145].
If the most significant signature of an amorphous state is localize in
the very low-frequency region, the internal vibration region is widely
used to characterized the physical state of pharmaceuticals with rou-
tine IR and Raman spectrometers, including to distinguish the amor-
phous state from crystalline phases. However, to obtain structural
information on very disordered states or to distinguish the amorphous
state from disordered crystalline states or micro/nano crystallized
states, low-frequency data are requested requiring the use of complex
multi-grating equipment. It is often requested to investigate the very
low-frequency region, to perform quantitative analyses, or separate
quasielastic scattering from collective vibrations. It can be expected
that this type of investigation becomes possible using routine spec-
trometers with the development of a new generation of holographic
filters [11]. In this context, very precise structural information on
amorphous states will become easily accessible with routine Raman
spectrometers. Combining low-frequency investigations with chemo-
metric analyses could provide very detailed structural information on
amorphous, especially making possible to distinguish amorphous
states which are characterized by different calorimetric signatures of
the glass transition.
12 A. Hédoux / Advanced Drug Delivery Reviews xxx (2015) xxx–xxx
Acknowledgments
The author is indebted to MMT group in Unité Matériaux Et Trans-
formations (UMR CNRS 8207), especially to Professor Y. Guinet and
Dr L. Paccou.
References
[1] K. Kawakami, Current status of amorphous formulation and other special dosage
forms as formulation for early clinical phases, J. Pharm. Sci. 98 (2009) 2875–2885.
[2] D.Q.M. Craig, P.G. Royall, V.L. Kett, M.L. Hopton, The relevance of the amorphous
state to pharmaceutical dosage forms, Int. J. Pharm. 179 (1999) 179–207.
[3] C. Ahlneck, G. Zografi, The molecular basis of moisture effects on the physical and
chemical stability of drugs in the solid state, Int. J. Pharm. 62 (1990) 87–95.
[4] M. Yoshioka, B.C. Hancock, G. Zografi, Crystallization of indomethacin from the
amorphous state below and above its glass transition temperature, J. Pharm. Sci.
83 (1994) 1700–1705.
[5] T. De Beer, A. Burggraeve, M. Fonteyne, L. Saerens, J. Remon, C. Vervaet, Near infra-
red and Raman spectroscopy for the in-process monitoring of pharmaceutical pro-
duction processes, Int. J. Pharm. 417 (2011) 32–47.
[6] A. Paudel, D. Raijada, J. Rantanen, Raman spectroscopy in pharmaceutical product
design, Adv. Drug Deliv. Rev. 89 (2015) 3–20.
[7] A. Bunaciu, H. Aboul-Enein, V. Dang Hoang, Raman spectroscopy for protein anal-
ysis, Appl. Spectrosc. Rev. 50 (2015) 377–386.
[8] Z.-Q. Wen, Raman spectroscopy of protein pharmaceuticals, J. Pharm. Sci. 96
(2007) 2861–2878.
[9] J.A. Zeitler, P.F. Taday, D.A. Newnham, M. Pepper, K.C. Gordon, T. Rades, Terahertz
pulsed and spectroscopy imaging in the pharmaceutical setting — a review, J.
Pharm. Pharmacol. 59 (2007) 209–223.
[10] C. Moser, F. Havermeyer, Ultra-narrow band tunable laserline notch filter, Appl.
Phys. B Lasers Opt. 95 (2009) 597–601.
[11] R. Heyler, A. Randy, J. Carriere, F. Havermeyer, THz-Raman — accessing molecular
structure with Raman spectroscopy for enhanced chemical identification, analysis
and monitoring, in: M. Druy, R. Crocombe (Eds.),Conference on Next-generation
Spectroscopic Technologies VI, Book Series: Proceedings of SPIE, Baltimore, 2013
(UNSP 87260 J).
[12] E.D.L. Smith, R.B. Hammond, M.J. Jones, K.J. Roberts, J.B.O. Mitchell, S.L. Price, R.K.
Harris, D.C. Apperley, J.C. Cherryman, R. Docherty, The determination of the crystal
structure of anhydrous theophylline by X-ray powder diffraction with a systematic
search algorithm, lattice energy calculations, and 13C and 15N solid-state NMR: a
question of polymorphism in a given unit cell, J. Phys. Chem. B 105 (2001)
5818–5826.
[13] A. Hédoux, Y. Guinet, L. Paccou, Analysis of the mechanism of lysozyme pressure
denaturation from Raman spectroscopy investigations, and comparison with ther-
mal denaturation, J. Phys. Chem. B 115 (2011) 6740–6748.
[14] A. Hedoux, Y. Guinet, F. Capet, L. Paccou, M. Descamps, Evidence for a high-density
amorphous form in indomethacin from Raman scattering investigations, Phys. Rev.
B 77 (2008) 094205-094201-094210.
[15] S. Hubert, S. Briancon, A. Hédoux, Y. Guinet, L. Paccou, H. Fessi, F. Puel, Process in-
duced transformations during tablet manufacturing: phase transition analysis of
caffeine using DSC and low frequency micro-Raman spectroscopy, Int. J. Pharm.
420 (2011) 76–83.
[16] R. Fule, P. Amin, Hot-melt extruded amorphous solids dispersion of posaconazole
with improved bioavailability: investigating drug-polymer miscibility with ad-
vanced characterization, Biomed. Res. 2014 (2014) (146781-146716).
[17] W. Cao, Y. Xie, S. Krishnan, H. Lin, M. Ricci, Influence of process conditions on the
crystallization and transition of metastable mannitol forms in protein formulations
during lyophilization, Pharm. Res. 30 (2013) 131–139.
[18] A. Hédoux, L. Paccou, S. Achir, Y. Guinet, In situ monitoring of proteins during ly-
ophilization using micro-Raman spectroscopy: a description of structural changes
induced by dehydration, J. Pharm. Sci. 101 (2012) 2316–2326.
[19] A. Hédoux, L. Paccou, S. Achir, Y. Guinet, Mechanism of protein stabilization by tre-
halose during freeze-drying analyzed by in-situ micro-Raman spectroscopy, J.
Pharm. Sci. 102 (2013) 2484–2494.
[20] V.S. Tumuluri, M.S. Kemper, I.R. Lewis, S. Prodduturi, S. Majumdar, B.A. Avery, M.A.
Repka, Off-line and on-line measurements of drug-loaded hot-melt extruded films
using Raman spectroscopy, Int. J. Pharm. 357 (2008) 77–84.
[21] L. Dierickx, L. Saerens, A. Almeida, T. De Beer, J. Remon, C. Vervaet, Co-extrusion as
manufacturing technique for fixed-dose combination mini-matrices, Eur. J. Pharm.
Biopharm. 81 (2012) 683–689.
[22] T. De Beer, P. Vercruysse, A. Burggraeve, T. Quinten, J. Ouyang, X. Zhang, C. Vervaet,
J. Remon, W.R.G. Baeyens, In-line and real-time process monitoring of a freeze dry-
ing process using Raman and NIR spectroscopy as complementary process analyt-
ical technology (PAT) tools, J. Pharm. Sci. 98 (2009) 3430–3446.
[23] J. Aaltonen, P. Heinanen, L. Peltonen, H. Korteiarvi, V.P. Tanninen, L. Christiansen, J.
Hirvonen, J. Yliruusi, J. Rantanen, In-situ measurement of solvent-mediated phase
transition during dissolution testing, J. Pharm. Sci. 95 (2006) 2730–2737.
[24] G. Reich, Near-infrared spectroscopy and imaging: basic principles and pharma-
ceutical applications, Adv. Drug Deliv. Rev. 57 (2005) 1109–1143.
[25] J.M. Chalmers, P.R. Griffiths, Handbook of Vibrational Spectroscopy, John Wiley &
Sons, New York, 2002.
[26] F.L. Galeener, P.N. Sen, Theory for the first-order vibrational spectra of disordered
solids, Phys. Rev. B 17 (1978) 1928–1933.
Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
[27] R. Shuker, R. Gammon, Raman-scattering selection-rule breaking and the density
of states in amorphous materials, Phys. Rev. Lett. 25 (1970) 222–225.
[28] A.J. Martin, W. Brenig, Phys. Status Solidi B 64 (1974) 163.
[29] A.P. Sokolov, A. Kisliuk, D. Quitmann, E. Duval, Evaluation of density of vibrational
states of glasses from low-frequency Raman spectra, Phys. Rev. B 48 (1993)
7692–7695.
[30] A. Fontana, F. Rocca, M.P. Fontana, B. Rossi, A.J. Dianoux, Low-frequency dynamics
in superionic borate glasses by coupled Raman and inelastic neutron scattering,
Phys. Rev. B 41 (1990) 3778–3785.
[31] N.V. Surovtsev, S.V. Adichtchev, E. Rossler, M.A. Ramos, Density of vibrational states
and light-scattering coupling coefficient in the structural glass and glassy crystal of
ethanol, J. Phys. Condens. Matter 16 (2004) 223–230.
[32] T. Achibat, A. Boukenter, E. Duval, Correlation effects on Raman scattering from
low-energy vibrational modes in glasses. II. Experimental results, J. Chem. Phys.
99 (1993) 2046–2051.
[33] A. Hedoux, P. Derollez, Y. Guinet, A.J. Dianoux, M. Descamps, Low-frequency vibra-
tional excitations in the amorphous and crystalline states of triphenyl phosphite: a
neutron and Raman scattering investigation — art. no. 144202, Phys. Rev. B 6314
(2001) 4202 (NIL_4169-NIL_4175).
[34] V.N. Novikov, E. Duval, A model of low-frequency Raman scattering in glasses:
comparison of Brillouin and Raman data, J. Chem. Phys. 102 (1994) 4691–4698.
[35] A.P. Sokolov, E. Rossler, A. Kisliuk, D. Quitmann, Dynamics of strong and fragile
glass formers: differences and correlation with low-temperature properties,
Phys. Rev. Lett. 71 (1993) 2062–2065.
[36] U. Buchenau, R. Zorn, A relation between fast and slow motions in glassy and liquid
selenium, Europhys. Lett. 18 (1992) 523–528.
[37] M.C.C. Ribeiro, Low-frequency Raman spectra and fragility of imidazolium ionic
liquids, J. Chem. Phys. 133 (2010) 24503–24506.
[38] A. Hédoux, Y. Guinet, L. Paccou, P. Derollez, F. Danede, Vibrational and structural
properties of amorphous n-butanol: a complementary Raman spectroscopy and
X-ray diffraction study, J. Chem. Phys. 138 (2013) (214506-214501 – 2145068).
[39] G. Qun, L. Yan, L.H. Chen, C. Yifeng, L. Feng, Comparison of several chemometric
methods of libraries and classifiers for the analysis of expired drugs based on
Raman spectra, J. Pharm. Biomed. Anal. 94 (2014) 58–64.
[40] J.H. Wang, P.K. Hopke, T.M. Hancewicz, S.L. Zhang, Application of modified alter-
nating least squares regression to spectroscopic image analysis, Anal. Chim. Acta
476 (2003) 93–109.
[41] N. Scoutaris, V. Kapilkumar, I. Slipper, B. Chowdhry, SEM/EDX and confocal Raman
microscopy as complementary tools for the characterization of pharmaceutical
tablets, Int. J. Pharm. 470 (2014) 88–98.
[42] S.J. Fraser, K.C. Gordon, Raman spectroscopy in the study of pharmaceuticals; the
problem and solutions of subsampling and data analysis, Eur. Pharm. Rev. 19
(2014) 3–8.
[43] C.M. McGoverin, T. Rades, K.C. Gordon, Recent pharmaceutical applications of
Raman and Terahertz spectroscopies, J. Pharm. Sci. 97 (2008) 4598–4621.
[44] J. Johansson, S. Pettersson, S. Folestad, Characterization of different laser irradiation
methods for quantitative Raman tablet assessment, J. Pharm. Biomed. Anal. 39
(2005) 510–516.
[45] M. Kim, H. Chung, Y. Woo, M.S. Kemper, A new non-invasive, quantitative Raman
technique for the determination of an active ingredient in pharmaceutical liquids
by direct measurement through a plastic bottle, Anal. Chim. Acta 587 (2007)
200–207.
[46] M. Kim, J. Noh, H. Chung, Y. Woo, M.S. Kemper, Y. Lee, Direct, non-destructive
quantitative measurement of an active pharmaceutical ingredient in an intact
capsule formulation using Raman spectroscopy, Anal. Chim. Acta 598 (2007)
280–287.
[47] M. Kim, H. Chung, Y. Woo, M.S. Kemper, New reliable Raman collection system
using the wide area illumination (WAI) scheme combined with the synchroneous
intensity correction standard for the analysis of pharmaceutical tablets, Anal. Chim.
Acta 579 (2006) 209–216.
[48] J. Villaumie, H. Jeffrey, Revolutionising Raman with the transmission technique,
Eur. Pharm. Rev. 20 (2015) 41–45.
[49] P. Matousek, A.W. Parker, Non-invasive probing of pharmaceutica capsules using
transmission Raman spectroscopy, J. Raman Spectrosc. 38 (2007) 563–567.
[50] C. Eliasson, P. Matousek, Passive signal enhancement in spatially offset Raman
spectroscopy, J. Raman Spectrosc. 39 (2008) 633–637.
[51] P. Matousek, I.P. Clark, E.R. Draper, M.D. Morris, A.P. Goodship, N. Everall, M.
Towrie, W.F. Finney, A.W. Parker, Subsurface probing in diffuselely Scattering
media using spatially offset Raman Spectroscopy, Appl. Spectrosc. 59 (2005)
393–400.
[52] E. Duval, A. Boukenter, T. Achibat, Vibrational dynamics and the structure of
glasses, J. Phys. Condens. Matter 2 (1990) 10227–10234.
[53] V.K. Malinovski, V.N. Novikov, A.P. Sokolov, Log-normal spectrum of low-energy
vibrational excitations in glasses, Phys. Lett. A 153 (1991) 63–66.
[54] V.K. Malinovski, A.P. Sokolov, The nature of boson peak in Raman scattering in
glasses, Solid State Commun. 57 (1986) 757–761.
[55] I. Pocsik, M. Koos, Cluster size determination in amorphous structures using the
Boson peak, Solid State Commun. 57 (1990) 1253–1256.
[56] C.A. Angell, Relaxation in liquids, polymers and plastic crystals — strong/fragile pat-
terns and problems, J. Non-Cryst. Solids 131–133 (1991) 13–31.
[57] A. Hédoux, L. Paccou, Y. Guinet, J.-F. Willart, M. Descamps, Using the low-frequency
Raman spectroscopy to analyze the crystallization of amorphous indomethacin,
Eur. J. Pharm. Sci. 38 (2009) 156–164.
[58] A. Hédoux, Solid state transformations of APIs during manufacturing by Raman
analysis of pharmaceutical molecules and dosage forms, Eur. Pharm. Rev. 16
(2011) 31–38.
 A. Hédoux / Advanced Drug Delivery Reviews xxx (2015) xxx–xxx
[59] T. Okumura, M. Otsuka, Evaluation of the microcrystallinity of a drug substance, in-
domethacin, in a pharmaceutical model tablet by chemometric FT-Raman spec-
troscopy, Pharm. Res. 22 (2005) 1350–1357.
[60] W. Sinclair, M. Leane, G. Clarke, A. Dennis, M. Tobyn, P. Timmins, Physical stability
and recrystallization kinetics of amorphous ibinabant drug product by Fourier
Transform Raman spectroscopy, J. Pharm. Sci. 100 (2011) 4687–4699.
[61] A. Heinz, M. Savolainen, T. Rades, C.J. Strachan, Quantifying ternary mixtures of dif-
ferent solid-state forms of indomethacin by Raman and near-infrared spectrosco-
py, Eur. J. Pharm. Sci. 32 (2007) 182–192.
[62] P.T. Mah, S.J. Fraser, M.E. Reish, T. Rades, K.C. Gordon, C.J. Strachan, Use of low-
frequency Raman spectroscopy and chemometrics for the quantification of crystal-
linity in amorphous giseofulvin tablets, Vib. Spectrosc. 77 (2015) 10–16.
[63] T. Shibata, T. Mori, S. Kojima, Low-frequency vibrational properties of crystalline
and glassy indomethacin probed by terahertz time-domain spectroscopy and
low-frequency Raman scattering, Spectrochim. Acta A 150 (2015) 207–211.
[64] P. Karmwar, K.A. Graeser, K.C. Gordon, C.J. Strachan, T. Rades, Investigations of
properties and recrystallisation behaviour of amorphous indomethacin samples
prepared by different methods, Int. J. Pharm. 417 (2011) 94–100.
[65] L.S. Taylor, G. Zografi, Spectroscopic characterization of interactions between PVP
and indomethacin in amorphous molecular dispersions, Pharm. Res. 14 (1997)
1691–1698.
[66] P. Karmwar, J.P. Boetker, K.A. Graeser, C.J. Strachan, J. Rantanen, T. Rades, Investiga-
tions on the effect of different cooling rates on the stability of amorphous indo-
methacin, Eur. J. Pharm. Sci. 44 (2011) 341–350.
[67] M.C. Wilding, M. Wilson, P.F. McMillan, Structural studies and polymorphism in
amorphous solids and liquids at high pressure, Chem. Soc. Rev. 35 (2006)
964–986.
[68] K. Adrjanowicz, A. Grzybowski, K. Kaminski, M. Paluch, Temperature and volume
effect on the molecular dynamics of superccooled Ibuprofen at ambient and elevat-
ed pressure, Mol. Pharm. 8 (2011) 1975–1979.
[69] J.-P. Amorij, A. Huckriede, J. Wilschut, H.W. Frijlink, W.L.J. Hinrichs, Development of
stable influenza vaccine powder formulations: challenges and possibilities, Pharm.
Res. 25 (2008) 1256–1273.
[70] Y.-F. Maa, M. Ameri, C. Shu, L.G. Payne, D. Chen, Influenza vaccine powder formu-
lation development: spray-drying and stability evaluation, J. Pharm. Sci. 93 (2004)
1912–1923.
[71] A. Lerbret, F. Affouard, P. Bordat, A. Hédoux, Y. Guinet, M. Descamps, Slowing down
of water dynamics in disaccharide aqueous solutions, J. Non-Cryst. Solids 357
(2011) 695–699.
[72] G.E. Walrafen, M.R. Fisher, M.S. Hokmabadi, W.H. Yang, Temperature dependence
of the low- and high-frequency Raman scattering from liquid water, J. Chem.
Phys. 85 (1986) 6970–6982.
[73] G.E. Walrafen, Y.C. Chu, G.J. Piermarini, Low-frequency Raman scattering from
water at high pressures and high temperatures, J. Phys. Chem. 100 (1996)
10363–10372.
[74] J.A. Padro, J. Marti, An interpretation of the low-frequency spectrum of liquid
water, J. Chem. Phys. 118 (2003) 452–453.
[75] A. Lerbret, F. Affouard, P. Bordat, A. Hédoux, Y. Guinet, M. Descamps, Low-
frequency vibrational properties of lysozyme in sugar aqueous solutions: a
Raman scattering and molecular dynamics simulation study, J. Chem. Phys. 131
(2009) 245103–245110.
[76] A. Hédoux, R. Ionov, J.F. Willart, A. Lerbret, F. Affouard, Y. Guinet, M. Descamps, D.
Prevost, L. Paccou, F. Danède, Evidence of a two-stage thermal denaturation pro-
cess in lysozyme: a Raman scattering and differential scanning calorimetry inves-
tigation, J. Chem. Phys. 124 (2006) 14703–14709.
[77] J.-A. Seo, A. Hedoux, Y. Guinet, L. Paccou, F. Affouard, A. Lerbret, M. Descamps, Ther-
mal denaturation of beta-lactoglobulin and stabilization mechanism by trehalose
analyzed from paman spectroscopy investigations, J. Phys. Chem. B114 (2010)
6675–6684.
[78] A. Hedoux, J.F. Willart, L. Paccou, Y. Guinet, F. Affouard, A. Lerbret, M. Descamps,
Thermostabilization mechanism of bovine serum albumin by trehalose, J. Phys.
Chem. B 113 (2009) 6119–6126.
[79] R. Williams, K. Dunjer, Determination of the secondary structure of proteins from
the amide I band of the laser spectrum, J. Mol. Biol. 152 (1981) 783–813.
[80] A. Hédoux, L. Paccou, Y. Guinet, Relationship between β-relaxation and structural
stability of lysozyme: microscopic insight on thermostabilization mechanism by
trehalose from Raman spectroscopy experiments, J. Chem. Phys. 140 (2014)
225102–225107.
[81] A. Hédoux, S. Krenzlin, L. Paccou, Y. Guinet, M.-P. Flament, J. Siepmann, Influence of
urea and guanidine hydrochloride on lysozyme stability and thermal denaturation;
a correlation between activity, protein dynamics and conformational changes,
Phys. Chem. Chem. Phys. 12 (2010) 13189–13196.
[82] A. Hédoux, J.F. Willart, R. Ionov, F. Affouard, Y. Guinet, L. Paccou, A. Lerbret, M.
Descamps, Analysis of sugar bioprotective mechanisms on the thermal denatur-
ation of lysozyme from Raman scattering and differential scanning calorimetry in-
vestigations, J. Phys. Chem. B 110 (2006) 22886–22893.
[83] A. Hédoux, J.-A. Seo, Y. Guinet, L. Paccou, Analysis of cold denaturation mechanism
of β-lactoglobulin and comparison with thermal denaturation from Raman spec-
troscopy, J. Raman Spectrosc. 43 (2011) 16–23.
[84] S.N. Timasheff, Protein-solvent preferential interactions, protein hydration, and the
modulation of biochemical reactions by solvent components, Proc. Natl. Acad. Sci.
U. S. A. 99 (2002) 9721–9726.
[85] S.N. Timasheff, Protein hydration, thermodynamic binding, and preferential hydra-
tion, Biochemistry 41 (2002) 13473–13482.
[86] G. Xie, S.N. Timasheff, The thermodynamic mechanism of protein stabilization by
trehalose, Biophys. Chem. 64 (1997) 25–43.
Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
13
[87] G. Xie, S.N. Timasheff, Mechanism of the stabilization of ribonuclease A by sorbitol:
preferential hydration is greater for the denatured than for the native protein, Pro-
tein Sci. 6 (1997) 211–221.
[88] R. Ionov, A. Hédoux, Y. Guinet, P. Bordat, A. Lerbret, F. Affouard, D. Prevost, M.
Descamps, Sugar bioprotective effects on thermal denaturation of lysozyme: in-
sights from Raman scattering experiments and molecular dynamics simulation,
J. Non-Cryst. Solids 352 (2006) 4430–4436.
[89] C. Zhou, W. Qi, E.N. Lewis, J.F. Carpenter, Concomitant Raman spectroscopy and dy-
namic light scattering for characterization of therapeutic proteins at high concen-
trations, Anal. Biochem. 472 (2015) 7–20.
[90] H. Bothe, H.K. Cammenga, Phase transitions and thermodynamic properties of an-
hydrous caffeine, J. Therm. Anal. 16 (1979) 267–275.
[91] A. Cesaro, G. Starec, Thermodynamic properties of caffeine crystal forms, J. Phys.
Chem. 84 (1980) 1345–1346.
[92] P. Derollez, N. Correia, F. Danede, F. Affouard, J. Lefebvre, M. Descamps, Ab initio
structure determination of the high-temperature phase of anhydrous caffeine by
X-ray powder diffraction, Acta Crystallogr. B61 (2005) 329–334.
[93] M. Descamps, N. Correia, P. Derollez, F. Danede, F. Capet, Plastic and glassy crystal
states of caffeine, J. Phys. Chem. B 109 (2005) 16092–16098.
[94] P. Lunkenheimer, U. Schneider, R. Brand, A. Loidl, Glassy dynamics, Contemp. Phys.
41 (2000) 15–36.
[95] D.J. Sutor, The structures of pyrimidines and purines. VII. The crystal structure of
Caffeine, Acta Crystallogr. 11 (1958) 453–458.
[96] C. Lehmann, F. Stowasser, The crystal structure of anhydrous beta-caffeine as de-
termined from x-ray powder-diffraction data, Chem. Eur. J. 13 (2007) 2908–2911.
[97] G. Enright, V. Terskikh, D. Brouwer, J. Rpmeester, The structure of two anhydrous
polymorphs of caffeine from single-crystal diffraction and ultrahigh-field solid-
state C NMR spectroscopy, Cryst. Growth Des. 7 (2007) 1406–1410.
[98] J. Moura Ramos, N. Correia, H. Diogo, M. Descamps, Dielectric study of the slow
motional processes in the polymorphic states of anhydrous caffeine, J. Phys.
Chem. B110 (2006) 8268–8273.
[99] V.P. Lehto, E. Laine, A kinetic study of polymorphic transition of anhydrous caffeine
with micocalorimeter, Thermochim. Acta 317 (1998) 47–58.
[100] A. Hédoux, A.-A. Decroix, Y. Guinet, L. Paccou, P. Derollez, M. Descamps, Low- and
high-frequency investigations on caffeine: polymorphism, disorder and phase
transformation, J. Phys. Chem. B 115 (2011) 5746–5753.
[101] E. Dudognon, F. Danede, M. Descamps, N. Correia, Evidence for a new crystalline
phase of racemic ibuprofen, Pharm. Res. 25 (2008) 2853–2858.
[102] A. Hédoux, Y. Guinet, P. Derollez, E. Dudognon, N. Correia, Raman spectroscopy of
racemic ibuprofen: evidence of molecular disorder in phase II, Int. J. Pharm. 421
(2011) 45–52.
[103] P. Derollez, E. Dudognon, F. Affouard, F. Danede, N.T. Correia, M. Descamps, Ab
initio structure determination of phase II of racemic ibuprofen by X-ray powder
diffraction, Acta Crystallogr. B 66 (2010) 76–80.
[104] W. Oswald, Studien uber die Bildung und Umwandlung fester Korper, Z. Phys.
Chem. 22 (1897) 289–330.
[105] J.H. Crowe, S.B. Leslie, L.M. Crowe, Is vitrification sufficient to preserve liposomes
during freeze-drying? Cryobiology 31 (1994) 355–366.
[106] J.H. Crowe, L.M. Crowe, J.F. Carpenter, C. Aurell Wistrom, Stabilization of dry phos-
pholipid bilayers and proteins by sugars, Biochem J. 242 (1987) 1–10.
[107] J. Crowe, J.F. Carpenter, L. Crowe, The role of vitrification in anhydrobiosis, Annu.
Rev. Physiol. 60 (1998) 73–103.
[108] J.L. Green, J. Fan, C.A. Angell, the protein-glass analogy: some insights from
homopeptide comparisons, J. Phys. Chem. 98 (1994) 13780–13790.
[109] T. Arakawa, S.N. Timasheff, Stabilization of protein structure by sugars, Biochemis-
try 21 (1982) 6536–6544.
[110] T. Arakawa, S.N. Timasheff, Preferential interactions of proteins with salts in con-
centrated solutions, Biochemistry 21 (1982) 6545–6552.
[111] T. Arakawa, S.J. Prestrelski, W.C. Kenney, J.F. Carpenter, Factors affecting short-term
and long-term stabilities of proteins, Adv. Drug Deliv. Rev. 46 (2001) 307–326.
[112] L. Chang, M. Pikal, Mechanisms of protein stabilization in the solid state, J. Pharm.
Sci. 98 (2009) 2886–2908.
[113] J. Dong, A. Hubel, J. Bischof, A. Aksan, Freezing-induced phase separation and spa-
tial microheterogeneity in protein solutions, J. Phys. Chem. B 113 (2009)
10081–10087.
[114] R.L. Remmele Jr., C. Stushnoff, J.F. Carpenter, Real-time in situ monitoring of lyso-
zyme during lyophilization using infrared spectroscopy: dehydration stress in
the presence of sucrose, Pharm. Res. 14 (1997) 1548–1555.
[115] A. Kauppinen, M. Toiviainen, J. Aaltonen, O. Korhonen, K. Jarvinen, M. Juuti, R.
Pellinen, J. Ketolainen, Microscale freeze-drying with Raman spectroscopy as a
tool for process development, Anal. Chem. 85 (2013) 2109–2116.
[116] T. De Beer, M. Alleso, F. Goethals, A. Coppens, Y. Van der Heyden, H. Lopez De Diego, J.
Rantanen, F. Verpoort, C. Vervaet, J. Remon, W. Baeyens, Implementation of a process
analytical technology system in a freeze-drying process using Raman spectroscopy for
in-line process monitoring, Anal. Chem. 79 (2007) 7992–8003.
[117] T. De Beer, M. Alleso, F. Goethals, A. Coppens, Y. Van der Heyden, H. Lopez De
Diego, J. Rantanen, F. Verpoort, C. Vervaet, J. Remon, W. Baeyens, Implementation
of a process analytical technology system in a freeze-drying process using Raman
spectroscopy for in-line process monitoring, Anal. Chem. 79 (2007) 7992–8003.
[118] L. Cordone, G. Cottone, A. Cupane, S. Giuffrida, M. Levantino, Proteins in saccharides
matrices and trehalose peculiarity: biochemical and biophysical properties, Curr.
Org. Chem. 19 (2015) 1684–1706.
[119] S. Pieters, Y. Van der Heyden, J.-M. Roger, C. Vervaet, T. De Beer, Raman spectros-
copy and multivariate analysis for the rapid discrimination between native-like
and non-native states in freeze-dried protein formulations, Eur. J. Pharm.
Biopharm. 85 (2013) 263–271.
14 A. Hédoux / Advanced Drug Delivery Reviews xxx (2015) xxx–xxx
[120] B.S. Kendrick, A. Dong, S.D. Allison, M.C. Manning, J.F. Carpenter, Quantitation of
the area overlap between second-derivative amide I infrared spectra to determine
the structural similarity of a protein in different states, J. Pharm. Sci. 85 (1996)
155–158.
[121] U. Roessl, S. Leitgeb, S. Pieters, T. De Beer, B. Nidetzky, Spectroscopy-based process
analytical tool for frozen storage of biopharmaceuticals, J. Pharm. Sci. 103 (2014)
2287–2295.
[122] A. Balogh, G. Dravavoelgyi, K. Farago, A. Farkas, T. Vigh, P.L. Soti, I. Wagner, J.
Madarasz, H. Pataki, G. Marosi, Z.K. Nagy, Plasticized drug-loaded melt electrospun
polymer mats: characterization, thermal degradation, and release kinetics,
J. Pharm. Sci. 103 (2014) 1278–1287.
[123] E. Karavas, M. Georgarakis, A. Doscoslis, D. Bikiaris, Combining SEM, TEM, and
micro-Raman techniques to differentite between the amorphous molecular level
dispersions and nanodispersions of a poorly water-soluble drug within a polymer
matrix, Int. J. Pharm. 340 (2007) 76–83.
[124] B. Vajna, H. Pataki, Z.K. Nagy, I. Farkas, G. Marosi, Characterization of melt
extrudedand conventional isoptin formulations using Raman chemical imaging
and chemometrics, Int. J. Pharm. 419 (2011) 107–113.
[125] I. Sievens-Fegora, A. Bhakay, J.I. Jerez-Rozo, N. Pandya, R.J. Romanach, B. Michniak-
Kohn, Z. Iqbal, E. Bilgili, R.N. Dave, Preparation and characterization of hydroxypro-
pyl methyl cellulose films containing stable BCS class II drug nanoparticles for
pharmaceutical applications, Int. J. Pharm. 423 (2012) 496–508.
[126] M. Wilson, M.A. Williams, D.S. Jones, G.P. Andrew, Holt-melt extrusion technology
and pharmaceutical application, Ther. Deliv. 3 (2012) 787–797.
[127] L. Saerens, D. Ghanam, C. Raemdonck, K. Francois, J. Manz, R. Kruger, S. Kruger, C.
Vervaet, J. Remon, T. De Beer, In-line solid state prediction during pharmaceutical
holt-melt extrusion in a 12 mm twin screw extruder using Raman spectroscopy,
Eur. J. Pharm. Biopharm. 87 (2014) 606–615.
[128] L. Saerens, L. Dierickx, B. Lenain, C. Vervaet, J. Remon, T. De Beer, Raman spectros-
copy for the in-line polymer-drug quantification and solid state characterization
during a pharmaceutical holt-melt extrusion process, Eur. J. Pharm. Biopharm. 77
(2011) 158–163.
[129] M. Haaser, M. Windbergs, C.M. McGoverin, P. Kleinebudde, T. Rades, K.C. Gordon,
C.J. Strachan, Analysis of Matrix dosage forms during dissolution testing using
Raman Microscopy, J. Pharm. Sci. 100 (2011) 44524459.
[130] J.-F. Willart, M. Descamps, Solid state amorphization of pharmaceuticals, Mol.
Pharm. 5 (2008) 905–920.
[131] B. Hancock, M. Parks, What is the true solubility advantage for amorphous pharma-
ceuticals, Pharm. Res. 17 (2000) 397–404.
[132] L.S. Taylor, G. Zografi, The quantitative analysis of crystallinity using FT-Raman
spectroscopy, Pharm. Res. 15 (1998) 755–761.
Please cite this article as: A. Hédoux, Recent developments in the Raman and infrared investigations of amorphous pharmaceuticals and protein
formulations: A review, Adv. Drug Deliv. Rev. (2015), http://dx.doi.org/10.1016/j.addr.2015.11.021
[133] A.G. Schmidt, S. Wartewig, K.M. Picker, Polyethylene oxides: protection potential
against polymorphic transitions of drugs? J. Raman Spectrosc. 35 (2004) 360–367.
[134] T. Matsumoto, G. Zografi, Physical properties of solid molecular dispersions of
indomethacin with poly(vinylpyrrolidone) and poly(vinylpyrrolidone-co-vinyl-
acetate) in relation to indomethacin crystallization, Pharm. Res. 16 (1999)
1722–1728.
[135] A. Hédoux, Y. Guinet, M. Descamps, The contribution of Raman spectroscopy to the
analysis of phase transformations in pharmaceutical compounds, Int. J. Pharm. 417
(2011) 17–31.
[136] D.E. Alonzo, G.G. Zhang, D. Zhou, Y. Gao, L.S. Taylor, Understanding the behavior of
amorphous pharmaceutical systems during dissolution, Pharm. Res. 27 (2010)
608–618.
[137] M. Alleso, N. Chieng, S. Rehder, J. Rantanen, T. Rades, J. Aaltonen, Enhanced disso-
lution rate and synchronized release of drugs in binary systems through formula-
tion: amorphous naproxen-cimetidine mixtures prepared by mechanical
activation, J. Control. Release 136 (2009) 45–53.
[138] K. Lobmann, R. Laitinen, H. Grohganz, K.C. Gordon, C.J. Strachan, T. Rades,
Coamorphous drug systems: enhanced physical stability and dissolution rate of in-
domethacin and naproxen, Mol. Pharm. 8 (2011) 1919–1928.
[139] K. Lobmann, C.J. Strachan, H. Grohganz, T. Rades, O. Korhonen, R. Laitinen, Co-
amorphous simvastatin and glipizide combinations show improved physical sta-
bility without evidence of intermolecular interactions, Eur. J. Pharm. Biopharm.
81 (2012) 159–169.
[140] K. Lobmann, R. Laitinen, H. Grohganz, C.J. Strachan, T. Rades, K.C. Gordon, A theo-
retical and spectroscopic study of co-amorphous naproxen and indomethacin,
Int. J. Pharm. 453 (2013) 80–87.
[141] A. Hédoux, Y. Guinet, L. Paccou, F. Danede, P. Derollez, Polymorphic transformation
of anhydrous caffeine upon grinding and hydrostatic pressurizing analyzed by
low-frequency Raman spectroscopy, J. Pharm. Sci. 102 (2013) 162–170.
[142] B. Kann, H.L. Offerhaus, M. Windbergs, C. Otto, Raman microscopy for cellular in-
vestigations — from single cell imaging to drug carrier uptake visualization, Adv.
Drug Deliv. Rev. 89 (2015) 71–90.
[143] D.G. Grier, A revolution in optical manipulation, Nature 424 (2003) 810–816.
[144] A. Ashkin, J.M. Dziedzic, T. Yamane, Optical trapping and manipulation of single
cells using infrared laser beams, Nature 30 (1987) 769–771.
[145] S. Rao, S. Raj, S. Balint, C. Bardina Fons, S. Campoy, Single DNA molecule detection
in an optical trap using surface-enhanced Raman scattering, Appl. Phys. Lett. 96
(2010) 213701–213703.