190
`
`COMMUNICATIONS
`
`J. Pharm. Pharmacol. 1984,36: 190-191
`Communicated August 8, 1983
`Polymorphic behaviour of chloroquine diphosphate
`
`0 1984 J. Pharm. Pharmticol.
`
`PH. VAN AERDE*, J. P. REMON, D. DE RUDDER, R. VAN SEVEREN, P. BRAECKMAN,
`Laboratory of Pharmacognosy and
`Galenical Pharmacy, State University of Ghent, Harelbekesrraat 72, 8-9000, Ghent, Belgium
`
`two different batches of
`The crystalline structure of
`chloroquine diphosphate was studied. Differential thermal
`analysis showed the existence of two polymorphic modifica-
`tions in one of the batches, whilst spectral analysis did not
`reveal any differences. This discrepancy is attributed to the
`fact that the second modification is formed during the
`transition phase when analysed by differential thermal
`analysis. This recrystallization process is initiated by the
`presence of seed material. The ratio of transition heat of
`both modifications is dependent on particle size and heating
`rate during analysis.
`During preformulation studies of chloroquine diphosp-
`hate, a crystallographic characterization of two batches
`of the drug, purchased from two different suppliers, was
`carried out. Chloroquine diphosphate is reported to
`exist in two different polymorphic forms (Merck Index
`1976; Paulini 1953). To identify the polymorphic type of
`the batches under investigation, we used differential
`thermal analysis (DTA), X-ray diffraction, Raman
`spectroscopy and ir spectroscopy.
`Because polymorphic behaviour was only supported
`by the results of DTA, further investigation was made to
`evaluate these findings.
`
`Materials and methods
`Chloroquine diphosphate was purchased from RhBne-
`Poulenc, France (CDP1) and Sigma, U.S.A. (CDP2).
`X-ray diffraction patterns were obtained with a
`Philips PA 25, equipped with a copper anticathode
`(25 mA, 40 KV).
`Raman spectroscopy was performed using a SPEX
`1401 spectrometer, equipped with an Ar+ laser (v =
`19 430 cm-1). Ir spectra were recorded on a Beckman
`IR spectrometer type 7, using KBr discs. Fluorescence
`analysis was with a Philips fluorescence spectrometer
`equipped with an Au anticathode (40mA, 60KV.).
`Spectra of light elements were recorded using a penta-
`erythritol crystal (002) with gasflow detection. A LiF
`crystal (200) with a scintillation counter was used for the
`determination of heavy elements.
`DTA was carried out with a Mettler TA 2000
`analyser. Samples of 7 mg were heated in open alu-
`minium crucibles. The sensitivity range was 200 yV.
`The powders were sized on test sieves (Haver and
`Boecker) with an air-jet sifter (Alpine).
`
`Results and discussion
`DTA of CDPl shows two endothermic peaks respec-
`tively at 196°C and 216°C while the DTA trace of
`* Correspondence.
`
`CDP2 reveals only one endotherm at 196 "C (Fig. 1).
`These results indicate the existence of a mixture of two
`polymorphic modifications in CDP1.
`To confirm the difference in crystalline structure
`between both products, X-ray powder diffraction was
`carried out. The resulting values of d, 28 and 1/Imax are
`reported in Table 1. Neither pattern could be differen-
`tiated from the other. An attempt was made to gain
`additional information by vibrational spectroscopy.
`Analysis of the ir absorption spectra did not reveal any
`complementary data from overlapping of the broad
`phosphate bands. The Raman spectra specifications are
`shown in Table 2. Again no difference between CDPl
`and CDP2 could be detected.
`To elucidate the observed discrepancy between the
`thermal and spectral analyses, further DTA-data were
`obtained. Fluorescence analysis was performed on both
`products to exclude erroneous conclusions due to
`sample impurity. No other elements besides phosphorus
`and chlorine could be detected.
`Because electron scanning microscopy revealed a
`difference in crystal habit only between the bigger
`
`196 216
`Temp ("C 1
`FIG. 1. Differential thermograms of CDPl and CDP2.
`Heating rate: 5 "C min-I.
`
`Argentum EX1038
`
`Page 1
`
`

`

`191
`COMMUNICATIONS
`Table 1. X-Ray diffraction pattern of chloroquine diphos-
`Table 2. Raman spectroscopy of chloroquine diphosphate.
`phate.
`
`20
`7.8
`9.4
`11.2
`11.8
`13.4
`14.0
`16.0
`16.8
`18.6
`19.4
`21.1
`23.0
`23.5
`24.5
`25.2
`26
`27
`
`d
`11.10
`9.40
`8.10
`7.48
`6.62
`6.30
`5.50
`5.25
`4.78
`4.55
`4.20
`3.87
`3.75
`3.59
`3.51
`3.41
`3.28
`
`I/Imax
`21
`24
`18
`30
`8
`12
`8
`39
`26
`96
`100
`33
`12
`68
`14
`22
`96
`
`28
`d
`28.2
`3.14
`-. . - _ _
`29.5
`3.00
`31.8
`2.80
`32.6
`2.73
`33.8
`2.64
`34.6
`2.58
`33.6
`2.51
`36
`2.48
`37.2
`2.41
`38.4
`2.33
`39.8
`2.25
`40.8
`2.21
`41.8
`2.15
`42.8
`2.10
`44.8
`2.01
`47.2
`1.92
`
`W n a x
`11
`1 7
`_ .
`13
`12
`7
`7
`8
`9
`8
`9
`7
`5
`9
`8
`7
`9
`
`particles of both powders, we presumed CDPl to be a
`mixture of crystals originating from two different
`crystallization processes. Therefore we wondered if
`there was any connection between particle size and
`polymorphic behaviour so the following sieve fractions
`of CDPl were prepared: 32-63, 63-100, 100-125,
`125-180 and 180-250 wm. The DTA of these fractions
`showed a clear decrease of the transition heat quantity
`(AQ) of the second modification with increasing par-
`ticle size, while no change was noticed in the AQ of the
`first modification. When the 150-250 pm fraction was
`crumbled with pestle and mortar and sieved again in
`fractions as described above, an identical relation was
`obtained. These results indicate that the influence of
`particle size on the AQ of both endotherms cannot be
`caused by a difference in crystalline structure between
`the fractions.
`Again X-ray patterns were identical in all cases.
`With a heating microscope, it was possible to study
`visually the melting behaviour of both powders. For
`
`1.2-
`
`-
`0.8
`
`
`
`-
`
`.-
`0
`c e
`0.4- :
`
`\
`
`v (cm-1)
`600 w
`770 mw
`900 w
`1070 w
`1100ms
`i i 7 0 ~
`1200 w
`1250 w
`::y:]dblt
`
`v (cm-1)
`1460 w
`1550 s
`1600 w
`1650 w
`2875 w ~. .
`2925 w
`2975 w
`3050 w
`
`w =weak
`mw = medium weak
`ms = medium strong
`
`s = strong
`dblt = doublet
`
`CDP2 this process was completed at 200°C, above
`which no further changes were observed. But for CDPl
`between 200-210 "C the formation of small regular
`crystals occurred and these melted at 216"C, which
`corresponds with the second peak in its thermogram.
`It is postulated that, during the first transition phase
`of CDP1, the second modification is formed by partial
`recrystallization, characterized by the exothermic peak
`occurring between both endotherms and initiated by the
`presence of seed material. Indeed, when equal amounts
`of CDPl and CDP2 were mixed and analysed by DTA,
`no decrease of the AQ of the second modification was
`noticed. Moreover, removal of potential seed material
`by recrystallization of CDPl from water-methanol, led
`to the disappearance of the second endothermic peak.
`If it is assumed that the second modification results
`from a conversion of the first modification during the
`melting process, the transition kinetics should be
`dependent not only on the particle size as demon-
`strated, but also on the available transition time
`between both endotherms. In fact, increasing the
`heating rate had no effect on the AQ of the first
`modification whilst a distinct decrease of AQ of the
`second modification was observed.
`The combined effect of both parameters is visualized
`in Fig. 2. As expected, the influence on the heating rate
`increases with smaller particles, whilst the effect of
`particle size is the most pronounced at the lowest
`heating rate. The enthalpy change of the first modifica-
`tion is 14.9calg-l
`(63Jg-l). Because the actual
`amount of the second form depends upon how much has
`crystallized from the first melting process, the enthalpy
`change could not be calculated.
`As the observed polymorphic behaviour is only
`established during the melting process, it is obvious that
`this cannot be detected by spectral analysis at ambient
`temperature.
`
`REFERENCES
`The Merck Index (1976) in: Windholz M. (ed.) Merck &
`Co. Inc. Rahway, N.J., U.S.A., 9th edition, p. 276
`Paulini, E. (1953) Rev. Brasil, Malariol. D. Trop. 5, 1:
`49-54
`
`O
`
`L
`32
`
`, -
`I
`
`I
`63
`180
`'."J 125
`Particle size (pm)
`FIG. 2. The effect of particle size and heating rate on the
`ratio of transition heat of the second modification towards
`the first during DTA of CDP1. Heating rate "C min-1: A =
`2.5; 0 = 5; = 10.
`
`I
`
`Page 2
`
`