Organic Process Research & Development
Technical Note
Notes
contribution of a strategically placed π-bond can be just as
important as solid-phase hydrogen-bonding interactions in the
crystal lattice in the rational design of an appropriate cocrystal
partner in a supermolecular assembly from first principles.
⊥DRL-IPD Communication number: IPDO-IPM-00351
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
CONCLUSION
■
We thank the management of Dr. Reddy’s Laboratories Ltd. for
supporting this work. Co-operation from the colleagues of the
Analytical Research and Development Department and the
Process Research and Development Department of Active
Pharmaceutical Ingredients, IPDO, Dr. Reddy’s Laboratories
Ltd., is highly appreciated.
We have shown the results of our in-depth studies of the nature
of the host−guest interaction in finasteride 1 and dihydrofi-
nasteride 2 that led to a simple and unique purification of
4
structurally related undesired congeners. Knowledge gained
from these studies will be helpful in designing an appropriate
partner suitable for a molecular complex formation with the
host molecule involving noncovalent interactions. A further
systematic study to select appropriate cocrystallizable partners
to improve the bioavailability of APIs in drug development
employing these principles is in progress.
ADDITIONAL NOTES
■
a
The possibility of mere surface interaction is not obvious, as
the existence of solid solution is confirmed by several
techniques (see Supporting Information). However, solid
solution formation phenomenon is a kind of surface interaction
at molecular level.
EXPERIMENTAL SECTION
■
Polymorphic form III of finasteride 1 was used in all the
experiments. Solvents and regents were used for all the
reactions as received. Solid-state 13C NMR was recorded at 300
MHz. Infrared (IR) spectra were recorded as thin films on a
Mattson Galaxy series FTIR 3000 spectrometer referenced to
polystyrene standard. X-ray powder diffraction was collected on
the Rigaku D/Max-2200 model diffractometer equipped with a
horizontal goniometer in θ/2θ geometry. Cu Kα (λ = 1.5418
Å) radiation was used, and the samples were scanned between 3
and 45° 2θ. Differential scaning calorimetric (DSC) analyses
were carried out on Shimadzu DSC50. The ThermoGravi-
metric Analysis (TGA) was performed on Q500 of TA
Instruments. The thermogram was recorded from 25 to 250
°C under the nitrogen gas purge at a flow of 40 mL/min for
balance and 60 mL/min for a sample at a heating rate of 10 °C/
min.
b
Solid solution has been identified and observed in different
solvents such as EtOAc and EtOH. In order to have larger
crystals (for recording X-ray) we performed crystallization in
EtOH by adopting slow evaporation techniques.
c
Single-crystal X-ray diffraction: X-ray data have been collected
on the Rigaku AFC-7S diffractometer equipped with a Mercury
CCD detector using graphite monochromated Mo Kα radiation
(λ = 0.7107 Å). The crystal structures were solved by direct
methods (SIR 2004) and refined by least-squares procedures
(CRYSTALS) using the CrystalStructure (version 3.8) from
Rigaku/Molecular Structure Corporation, 9009 New Trials Dr.
The Woodlands, TX 77381-5209, USA. Crystal data for 2:
Formula C23H38N2O2, M = 374.57, monoclinic, a = 10.098(4)
Å, b = 7.705(3) Å, c = 28.585(11) Å, β = 94.240(4)°, V =
2217.9(15) Å3, T = 298 K, space group C2, Z = 4, ρcalc = 1.122
g cm−3, μ(Mo Kα) = 0.071 mm−1, 12464 reflections measured,
4154 unique reflections, 2976 observed reflections [I >
2.0σ(I)], Rint = 0.041, R1_obs = 0.049, wR2_all = 0.076.
Crystal data for solid solution of 1 and 2: Formula
C23H37.33N2O2, M = 370.53, monoclinic, a = 10.129(6) Å, b
= 7.686(4) Å, c = 28.564(17) Å, β = 94.255(7)°, V = 2217(2)
Å3, T = 298 K, space group C2, Z = 4, ρcalc = 1.11 g cm−3, μ(Mo
Kα) = 0.070 mm−1, 11992 reflections measured, 2455 unique
reflections, 1468 observed reflections [I > 2.0σ(I)], Rint = 0.045,
R1_obs = 0.056, wR2_all = 0.083.
DETAILS OF RECOVERY OF THE PURIFIED
■
MATERIAL THAT WAS USED TO PACK THE
COLUMN AND PURIFICATION OF IMPURE
FINASTERIDE
Ten grams of finasteride containing 3% DHF impurity was
dissolved in 800 mL of ethyl acetate to prepare a saturated
solution. This saturated solution obtained above was passed at a
rate of 0.4 mL/min through the 15 g of pure finasteride powder
(solid) packed in 1.1 cm × 10 cm glass column with 1−2 cm
glass beads of 90−150 μ packed from both the ends of column.
After completion of elution, either in continuous or batch
mode, the impurity level of finasteride 1 dropped down to
0.20−0.25% from the initial 3%. After evaporating the solvent,
dry powder was obtained in 97% yield (∼9.7 g), and the yield
of the solid powder recovered from the column (continuous
mode) or flask (batch mode) was about 102% (∼15.28 g).
4
It is understood that the pure finasteride column may
conceptually be used until it gets enriched with 10−15%
(This depends on concentration, length, width and flow rate; as
the preliminary data indicate.). Since the yield and purity data
were collected out of impure finasteride [contaminated with 3%
dihydrofinasteride (input)] and pure finasteride [ICH grade
(output)], we thus did not generate/include the meaningful
data from repeated use of the pure finasteride column (until it
becomes enriched with 10−15% of dihydrofinasteride).
However, there is much direct evidence (see SI) that the
solid solution can occur even with the 1:1 or 1:2 ratio of
finasteride and dihydrofinasteride; thus, there may be a
possibility of retaining even more than 15% of dihydrofinas-
teride on the pure coloumn of finasteride. Since the retention
percentage may vary as it depends not only on finasteride and
dihydrofinasteride equilibrium but also on several other factors
such as concentration, length, width, and flow rate, it seems a
ternary phase diagram may not help. Thus we do not include
the related data.
ASSOCIATED CONTENT
■
S
* Supporting Information
This material is available free of charge via the Internet at
AUTHOR INFORMATION
■
Corresponding Author
*Telephone: +91 8458 279 485. Fax: +91 8458 279 619. E-
601
dx.doi.org/10.1021/op300142a | Org. Process Res. Dev. 2013, 17, 599−602