Talanta 71 (2007) 1382–1386

Analysis and stability of polymorphs in tablets: The case of Risperidone

I. Karabas a,b , M.G. Orkoula a,b , C.G. Kontoyannis a,b,∗
a Department of Pharmacy, University of Patras, Patras, Greece
b Institute of Chemical Engineering and High Temperature Chemical Processes,
FORTH, P.O. Box 1414, Patras GR-26500, Greece
Received 23 March 2006; received in revised form 13 June 2006; accepted 6 July 2006
Available online 5 September 2006

Abstract
Identification of the crystal phase of an active pharmaceutical ingredient (API) in a pharmaceutical tablet is of outmost importance since
different polymorphs exhibit different physicochemical properties. Furthermore, some of the crystal phases are protected by patents. Identification
of Risperidone polymorph A in film coated commercial tablets was attempted using IR spectroscopy, Raman spectroscopy and X-ray powder
diffraction (XRPD). The stability of this polymorph through time and during the manufacturing process was also examined. The inability of IR and
Raman techniques to identify the presence of polymorph A in the tablets, despite their lower detection limits for Risperidone, left the XRPD as the
only technique that could be used for identifying the presence of Risperidone A against the other crystal phases in the presence of the excipients.
Polymorph A was proved to be stable during the manufacturing process and after a storage period of 2 years.
© 2006 Elsevier B.V. All rights reserved.

Keywords: X-ray powder diffraction; FT-IR; Raman; Risperidone; Polymorphs; Tablets

1. Introduction Since an amorphous form is thermodynamically less stable than

Many pharmaceutical solids exhibit polymorphism, which
is frequently defined as the ability of a substance to exist as
two or more crystalline phases that have different arrangements
and/or conformations of the molecules in the crystal lattice [1].
Polymorphism may also include solvation or hydration products
(also known as pseudopolymorphs) and amorphous forms [2].
Polymorphs and/or solvates of an active pharmaceutical ingre-
dient (API) can have different chemical and physical properties
such as melting point, chemical reactivity, apparent solubility,
dissolution rate, optical and electrical properties, vapor pressure
and density. These properties can have a direct impact on the
process-ability of drug substances and the quality/performance
of drug products, such as stability, dissolution and bioavailabil-
ity.
One polymorph may convert to another during manufactur-
ing and storage, particularly when a metastable form is used.
∗ Corresponding author at: Institute of Chemical Engineering and High
Temperature Chemical Processes, FORTH, P.O. Box 1414, Patras GR-26500,
Greece. Tel.: +30 2610 997727/969328; fax: +30 2610 997658.
E-mail address: cgk@iceht.forth.gr (C.G. Kontoyannis).
any crystalline form, inadvertent crystallization from an amor-
phous drug substance may occur. As a consequence of the higher
mobility and ability to interact with moisture, amorphous drug
substances are also more likely to undergo solid-state reactions.
In addition, phase conversions of some drug substances are
possible when exposed to a range of manufacturing processes
[3]. Milling/micronization operations may result in polymorphic
form conversion of a drug substance. In the case of wet gran-
ulation processes, where the usual solvents are aqueous, one
may encounter a variety of interconversions between anhydrates
and hydrates, or between different hydrates. Spray-drying pro-
cesses have been shown to produce amorphous drug substances.
The most stable polymorphic form of a drug substance is often
used because it has the lowest potential for conversion from one
polymorphic form to another while the metastable form may be
used to enhance the bioavailability. Therefore, it is essential to
ensure that polymorphic differences are addressed via design
and control of formulation and process conditions to physical
and chemical stability of the product over the intended shelf-
life.
Based on these, some regulatory agents may refuse to approve
an application for a new dug referencing a listed drug if the

0039-9140/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.talanta.2006.07.009
 I. Karabas et al. / Talanta 71 (2007) 1382–1386
application contains insufficient information to show that the
drug substance is the “same” and remains the “same” as that of
the reference listed drug during the expected shelf-life. Even in
the cases where the regulatory agent accepts the crystal phases as
equivalent, an additional complication stems from the patented
protection of some metastable polymorphs.
Physicochemical measurements and techniques that are com-
monly used to determine the crystal phases are [4]: melting point
(including hot-stage microscopy), solid-state IR, X-ray powder
diffraction, thermal analysis procedures (DSC, TGA and DTA),
Raman spectroscopy, optical microscopy and solid-state NMR.
Microscopy, thermal analysis methodology and solid-state NMR
are generally considered as sources of supporting information.
The definite criterion for the existence of polymorphism is via
demonstration of a non-equivalent crystal structure, usually by
comparison of the X-ray diffraction patterns. Although most of
these techniques can be easily applied to pure crystal phases
it is mostly Raman spectroscopy and X-ray powder diffraction
(XRPD) that have been used for rapid screening and identifi-
cation of new crystal phases in high-throughput crystallization
[5]. Despite the large array of available techniques the identifi-
cation of an active ingredient in a formulation in the presence of
numerous excipients is neither easy nor straightforward.
In this work, an attempt was made to identify the presence of
the crystal phases of the active ingredient Risperidone, a ben-
zisoxazole derivative that is used as an antipsychotic drug, in
commercially available film coated tablets, using Raman and IR
spectroscopy as well as XRPD. Furthermore, the possibility of
the transformation of the API from one crystal phase to another
with time or due to the manufacturing process was examined,
through these techniques. There are three known polymorphs of
Risperidone (A, B and E) for which only the XRPD spectra are
available [6].
2. Experimental
2.1. Materials used
Risperidone is commercially available in film-coated tablets
containing 0.5, 1, 2, 4, 6 and 8 mg of the active ingredient.
Film coated Risperidone A tablets with 8 mg strength as well
as placebo tablets and pure Risperidone form A were kindly
provided by Specifar Pharmaceuticals SA, Athens, Greece.
The tablets contained the following excipients: lactose, maize
starch, microcrystalline cellulose, Hypromellose 2910, magne-
sium stearate, colloidal anhydrous silica, sodium lauryl sulphate
and propylene glycol. Each tablet weighted 185 mg. The film
coating was carefully removed from the tablets with a scalpel.
Attempts were made to prepare Risperidone B and E fol-
lowing the recipes given in the relevant patent [4]. One recipe
yielded the form B while several others yielded mixtures of A
and B. Polymorph E could not be obtained.
2.2. Characterization of the samples
The X-ray powder diffraction analysis was performed
(Philips 1830/40) on finely powdered samples using the Cu K␣
1383
Fig. 1. XRPD spectra of the three known Risperidone polymorphs (A, B and
E).
(40 kV and 30 mA) radiation and Ni filter with a scanning speed
of 0.005◦ 2θ s−1 . Time constant was set at 2 s. XRPD spectra
of the pure A and B Risperidone polymorphs are shown in
Fig. 1.
The Raman spectra were recorded using a FRA-106/S
FT-Raman (Bruker) with the following characteristics: the laser
excitation line used, was the 1064 nm of a Nd:YAG laser. A
secondary filter was used to remove the Rayleigh line. The
scattered light was collected at an angle of 180◦ . The system
was equipped with a liquid N2 cooled Ge detector (D 418).
The power of the incident laser beam was about 370 mW on
the sample’s surface. Typical spectral line width was 0.5 cm−1
while the recorded spectra were the average of 300 scans.
Raman spectra from the samples that were in the form of
powders, e.g. raw Risperidone, were recorded by placing in
aluminum cylindrical cups with 10 mm diameter and 4 mm
height having a cavity in the center of 2 mm diameter and 1 mm
depth. Homogeneity of the powders was verified by obtaining
five Raman spectra for each mixture, focusing the laser beam at
randomly selected spots of the surface. Spectra of uncoated and
the as-received intact film coated tablets were recorded by plac-
ing directly into laser’s beam path. Raman spectra of the pure
A and B Risperidone polymorphs are shown in Fig. 3A and B,
respectively.
All infrared spectra were recorded from 4000–400 cm−1 ,
using the EQUINOX 55 (Bruker) FT-IR spectrometer. Each
spectrum was the average of 200 scans using a spectral reso-
lution of 4 cm−1 . The IR spectra were recorded and stored using
a spectroscopic software (OPUS, Version 2.0). FT-IR spectra
of the pure substances and tablets were obtained by grounding
and mixing 2 mg in an agate mortar with 200 mg of desiccated
IR grade potassium bromide (Merck) in a dry box. Pellets were
obtained by compression of the powdered mixtures in a stainless-
steel die at 8 MPa for 5 min. IR spectra of the pure A and B
Risperidone polymorphs are shown in Fig. 5A and B, respec-
tively.
1384 I. Karabas et al. / Talanta 71 (2007) 1382–1386

3. Results and discussion

3.1. Comparison of spectra of Risperidone polymorphs

3.1.1. X-ray powder diffraction spectra

The XRPD spectra of the synthetically prepared A and B
crystal phases of Risperidone can be seen in Fig. 1 and are in
accordance with the spectra quoted in the relevant patent [4]. In
Fig. 1, the XRPD spectrum of polymorph E, from ref. [4], can
also be seen. It is apparent that there are significant differences
in the spectra which can be used to identify their appearance
in the sample. For example, Risperidone A exhibits a major
reflection at 21.05◦ 2θ while Risperidone B and E at 21.6◦ 2θ.
Furthermore, the peak at 18.83◦ and the double peak at 23.03◦
and 23.43◦ 2θ can differentiate A polymorph against B and E.
The presence of the other two crystal phases (B and E) can
be verified by existence of peaks 17.4◦ and the double peak at
15.5◦ and 16.4◦ , respectively. Purity of A crystal form can be
ascertained by absence of any peak at the area of 17.5◦ and 27◦
2θ.
Though differentiation between the two polymorphs in a mix-
ture of two appears to be feasible, the discrimination should also
be checked in the case of a formulation where a number of other
substances are present. An XRPD spectrum of the placebo of the
commercial tablet is quoted in Fig. 2 as well as the spectrum of a
powdered tablet containing 8 mg Risperidone in a total amount
of 185 mg of the tablet. It is noted that the API cannot be detected
in the spectrum of the tablet which is overwhelmed by the pres-
ence of placebo, indicating that XRPD detection limit in the case
of Risperidone is higher than 8 mg/tablet.
3.1.2. Raman spectra
The Raman spectra of the synthetically prepared A and B
crystal phases of Risperidone can be seen in Fig. 3. There are
numerous differences between Risperidone A and B such as the
Fig. 3. FT-Raman spectra of: (A) Risperidone A; (B) Risperidone B; (C) tablet
with 25% (w/w) Risperidone; (P) Risperidone placebo; (T) commercial tablet
with 8 mg Risperidone (uncoated).
double peak at 3051 and 3069 cm−1 for Risperidone A, having
the 3069 cm−1 vibration stronger than the former, while Risperi-
done B exhibits a single peak at 3055 cm−1 with a shoulder
at 3069 cm−1 . An additional strong peak for polymorph B can
be seen at 2963 cm−1 . Details of some of the additional dif-
ferences between Risperidone A and B can be seen in Fig. 4.
Polymorph A exhibits a double peak at 432 and 440 cm−1 ,
a double peak at 1260 and 1271 cm−1 and a single peak at
1397 cm−1 as opposed to polymorph B whereas single peaks
appear at 434 cm−1 , 1266 cm−1 and a double peak at 1390 and
1402 cm−1 . The Raman spectra that were excited from 8 mg
intact commercial tablets, after the removal of the coating, as
well as from tablets of placebo, can also be seen in Fig. 3. It
can be seen that though the concentration of the API is only
4.3% (w/w) in the commercial tablet, its presence can be certi-
fied easily as the spectra (Fig. 3T and P) differ in three regions,
marked with arrows. These differences are probably sufficient
to identify the presence of API in the tablet though, as shown
earlier, it is not possible to distinguish between A and B crystal
forms. At this point it is worth noting that although the detection
limit of Raman spectroscopy for Risperidone is lower than the

Fig. 2. XRPD spectra of: (A) Risperidone A; (B) Risperidone A recorded from
the same powder 2 years after obtaining spectrum A; (P) placebo of Risperidone;
(C) powdered tablet of commercial 8 mg Risperidone; (D) powdered tablet with
25% (w/w) Risperidone A; (E) powdered tablet with 25% (w/w) Risperidone A Fig. 4. FT-Raman spectra of: (A) Risperidone A; (B) Risperidone B; (C) tablet
recorded from the same powdered tablet 2 years after obtaining spectrum D. with 25% (w/w) Risperidone; (P) Risperidone placebo.
 I. Karabas et al. / Talanta 71 (2007) 1382–1386
Fig. 5. FT-IR spectra of: (A) Risperidone A; (B) Risperidone B; (C) powdered
tablet with 25% (w/w) Risperidone; (T) powdered commercial tablet with 8 mg
Risperidone (uncoated); (P) Risperidone placebo.
respective for XRPD this advantage is not useful for the task at
hand.
3.1.3. IR spectra
The IR spectra of the synthetically prepared A and B crystal
phases of Risperidone, the placebo and the uncoated powdered
commercial tablet with 8 mg API can be seen in Fig. 5. The
major bands for both polymorphs appear at 1535 and 1650 cm−1 ,
but the major difference between them is that the vibration at
1650 cm−1 appears as single peak in polymorph A whereas in
polymorph B multiple peaks can be observed. In the spectrum
obtained from the powdered tablet (Fig. 5T) the presence of the
API could not be detected with the exception of two minor noise-
like wide bands, marked with arrows, which appear at around
1535 and 1650 cm−1 . Unfortunately, these bands cannot be used
for identification of the API.
3.2. Polymorph A stability tests
3.2.1. Twenty-five percent Risperidone tablets
Among the targets of this work was to verify if Risperidone
A in tablets transforms with time or through the manufacturing
process to a different crystal phase. Since neither the FT-IR nor
the Raman was capable of differentiating between the crystal
phases, despite their better detection limits than the XRPD, a
different approach was followed. A labscale batch of Risperi-
done tablets with 25% (w/w) Risperidone was manufactured by
the pharmaceutical company following the same manufactur-
ing process as with the commercially available concentrations.
Working with a 25% (w/w) of Risperidone was chosen after test-
ing mixtures of 10 and 15% (w/w). In the XRPD spectra obtained
from the 10 and 15% (w/w) mixtures (not shown) only the main
reflection of Risperidone A could be barely observed at 21.0◦
2θ on the top of a placebo reflection indicating that the XRPD
detection limit for Risperidone A was approximately 10% (w/w).
The Raman and IR spectra recorded from the same mixtures (not
1385
shown) were not more helpful than those recorded from the tablet
with 8 mg API. Although a 10% (w/w) mixture could have been
used for testing the stability of A polymorph based on XRPD
data, a 25% (w/w) was thought to be more helpful for the analy-
sis due to the appearance of additional Risperidone reflections,
as well as to have Raman and IR spectra that might permit the
identification of Risperidone A.
The Raman spectrum obtained from the tablet with 25%
(w/w) Risperidone A can be seen in Fig. 3C. The increased
quantity of the active ingredient allows the positive identification
of Risperidone A in these tablets. The wide band at 436 cm−1
(Fig. 4C) stems from the addition of polymorph A double peak
at 432 and 440 cm−1 and the placebo vibration at 436 cm−1
(Fig. 4P). The polymorph A vibrations at 1271 and 1397 cm−1
(Fig. 4C) as well as at 3051 and 3069 cm−1 (Fig. 3C) are unal-
tered due to the absence of the respective placebo vibrations.
It should be noted though that the presence of these vibrations
does not rule out the possibility of the simultaneous presence of
small quantity of B crystal phase.
The FT-IR spectrum obtained from the tablet with 25% (w/w)
Risperidone A can be seen in Fig. 5C. The presence of a single
peak at 1650 cm−1 indicates the presence of Risperidone A but
this does not also exclude the possibility of the simultaneous
presence of small amounts of B crystal phase since the multiple
peaks of B at 1650 cm−1 can be hidden under the apparently
single peak at 1650 cm−1 .
3.2.2. Manufacturing process stability test
In the sample with 25% (w/w) Risperidone (Fig. 2D) the char-
acteristic reflections of Risperidone polymorph A (Fig. 2A), as
well as the diffraction peaks of the compounds consisting the
placebo (Fig. 2P), were detected. Furthermore, no additional
reflections were detected including the most prominent diffrac-
tion peaks at 21.6◦ 2θ of polymorphs B and E. This indicates that
during the manufacturing process no transformation of Risperi-
done A took place. Therefore, in the case of Risperidone, XRPD
proved to be the only reliable technique for identifying the pres-
ence of A polymorph and simultaneously exclude the possibility
of the presence of the other two polymorphs.
3.2.3. Stability tests with time
In order to test the stability of Risperidone A with time,
the XRPD spectra of raw Risperidone A in powder form were
recorded with a time interval of 2 years (Fig. 2A and B). The
XRPD spectra of tablets with 25% (w/w) Risperidone A were
also recorded shortly after their manufacture (Fig. 2D) and after
2 years (Fig. 2E). In both cases, and in order to simulate real-
life scenario, in the interim period the powders were placed in a
laboratory cabinet for protection against the light but neither the
temperature nor the humidity was controlled. No difference was
observed indicating there is no transformation or degradation of
the active ingredient after a storage period of 2 years.
4. Conclusions
From the application of X-ray powder diffraction, FT-IR and
FT-Raman spectroscopy on the above mentioned specimens and
1386 I. Karabas et al. / Talanta 71 (2007) 1382–1386
raw materials as well as on tablets, it is apparent that there is
no straightforward process that can be applied in order: (a) to
identify the presence of different crystal phases of the active
ingredient and/or (b) to test their stability. Product oriented
methodologies should be devised using all available techniques
depending on a number of factors including the percentage of
the active ingredient in the formulation, the problems arising
from the presence of the excipients, the detection limits of the
different techniques for the active substance in question and the
availability of all crystal forms, including the patented.
For Risperidone A, the case study of this work, it can be
safely concluded that no transformation takes places during the
manufacturing process or during a storage period of 2 years.
None of the techniques used could positively identify the pres-
ence of polymorph A in the commercial tablets with 8 mg API.
In tablets with increased API strength IR and Raman proved to
be incapable of excluding the simultaneous presence of other
crystal phases although positive identification of polymorph A
was accomplished. Only the information obtained from XRPD
could be used for the identification of form A and the exclusion
of forms B and E.
Acknowledgments
Authors would like to thank Specifar Pharmaceuticals SA
for its help and in particular Ms. Eirini Vasilopoulos from the
Department of Product Development.
References
[1] D.J. Grant, in: H.G. Brittain (Ed.), Polymorphism in Pharmaceutical Solids,
Marcel Dekker, Inc., New York, 1999, pp. 1–31 (Chapter 1).
[2] International Conference on Harmonization Q6A Guideline: Specifications
for New Drug Substances and Products: Chemical Substances, October
1999.
[3] H.G. Brittain, E.F. Fiese, in: H.G. Brittain (Ed.), Polymorphism in Pharma-
ceutical Solids, Marcel Dekker, Inc., New York, 1999, pp. 279–330 (Chapter
8).
[4] H.G. Brittain, in: H.G. Brittain (Ed.), Polymorphism in Pharmaceutical
Solids, Marcel Dekker, Inc., New York, 1999, pp. 227–271 (Chapter 6).
[5] S. Morissette, O. Almarsson, M. Peterson, J. Remenar, M. Read, A.
Lemmo, S. Ellis, M. Cima, C. Gardner, Adv. Drug Deliv. Rev. 56 (2004)
275.
[6] B. Krochmal, D. Diller, B.-Z. Dolitzky, J. Aronhime, Preparation of Risperi-
done, U.S. Patent No. 6,750,341 and PCT/WO 02/12200 A1, Feb 14,
2002.