- Email: [email protected]
Preparation and Characterization of Polymorphs for an LTD4 Antagonist, RG 12525
Preparation and Characterization of Polymorphs for an LTD4 Antagonist, RG 12525 ROBERT A. CARLTON, THOMAS J. DIFEO‡, TORY H. POWNER, IVAN SANTOSX,
AND
MICHAEL D. THOMPSONX
Received June 26, 1995, from the Rhoˆne-Poulenc Rorer, 500 Arcola Road, Collegeville, PA 19426. Final revised manuscript ‡ Present address: Warner-Lambert, received January 23, 1996. Accepted for publication January 23, 1996X. Morris Plains, NJ 07950. Abstract 0 This report describes the preparation and characterization of two polymorphic forms of RG 12525, a leukotriene D4 (LTD4) antagonist. Polymorph I is prepared by recrystallization from methanol or titration of the sodium salt of RG 12525 with citric acid. Polymorph II is prepared by recrystallization from methanol or titration of the ammonium salt of RG 12525 with citric acid. The polymorphic system is enantiotropic, with pure form I melting at 154 °C, 3 deg less than the melting temperature of form II. Form I is thermodynamically more stable than form II at room temperature. These polymorphic forms are differentiated using microscopy, differential scanning calorimetry (DSC), infrared spectroscopy (IR), and powder X-ray diffraction (XRD) analysis. Solubility properties from 31 to 72 °C were determined to be similar for both forms. The calculated solubilities at 25 °C are 7.6 and 9.8 µM for forms I and II, respectively. The free energy change from form II to form I at 25 °C is −0.15 kcal/ mol. Thermodynamic properties of the system are summarized using a schematic free energy diagram.
Introduction RG 12525, an LTD4 antagonist, has been shown to have activity against leukotriene-induced bronchoconstriction. Further study of the physical properties of this compound revealed that RG 12525 has two stable polymorphic forms. This report describes the preparation and characterization of each polymorphic form and presents the results of studies undertaken to establish the relationship between the two forms.
Both forms have been prepared by crystallization from methanol. In addition, a titration method for the crystallization of each form is reported. The preparative conditions and experiments preceding their discovery are described. Solid state characteristics of the polymorphic forms are also included. Each form exhibits unique properties when examined through optical microscopy, differential scanning calorimetry (DSC), infrared spectroscopy (IR), solid state studies, and powder X-ray diffraction (XRD). The solubility properties of the two polymorphs are quite similar. Information gathered from solubility and DSC studies indicates that polymorph I is the more thermodynamically stable form at room temperature. These studies include a determination of the thermodynamically stable temperature ranges, identification of the high- and low-temperature forms, enantiotropism vs monoX
Abstract published in Advance ACS Abstracts, March 1, 1996.
© 1996, American Chemical Society and American Pharmaceutical Association
tropism, and a schematic free energy diagram. A comparison of optical and thermal properties is also given.
Experimental Section MaterialssRG 12525, 2-[[4-[[2-(1H-tetrazol-5-ylmethyl)phenyl]methoxy]phenoxy]methyl]quinoline, was prepared1 and subsequently used to investigate polymorph preparations. All other chemicals used were reagent grade. General Recrystallization MethodsRG 12525 was dissolved in refluxing solvent. The solution was seeded (if required) and then cooled to 0 °C. The mixture was filtered and dried, giving product. General Titration MethodsA solution of acid in water and cosolvent was added slowly to a solution of RG 12525 and aqueous base solution with a cosolvent at the given temperature. After crystallization, the mixture was cooled to room temperature, filtered, and dried, giving product. Preparation of RG 12525 Form I by Methanol RecrystallizationsRG 12525 (16.0 g) was dissolved in refluxing methanol (400 g). The solution was cooled to 50 °C and immediately seeded with form I. After 1 h, the mixture was cooled to 0 °C and filtered. Drying gave white needle-like fibers (15.0 g, 94% yield). Preparation of RG 12525 Form I by TitrationsTo a 7% solution of aqueous citric acid (12 g) in 2-propanol at 55 °C were added seed crystals of form I. To this mixture was slowly added a solution of water (13.0 mL), 2-propanol (11.0 mL), 50% aqueous sodium hydroxide (5.8 g), and RG 12525 (3.0 g). After addition, the mixture was cooled to room temperature, filtered, and dried, giving white needle-like fibers (2.7 g, 90% yield). Preparation of RG 12525 Form II by Methanol RecrystallizationsRG 12525 (3.0 g) was dissolved in refluxing methanol (75 g). The solution was cooled to 50 °C and seeded with form II. After cooling to 0 °C, the mixture was filtered and dried, giving white blocklike crystals (2.2 g, 75% yield). Preparation of RG 12525 Form II by TitrationsTo a solution of water (25.0 mL), PEG-200 (5.6 g), concentrated ammonium hydroxide (1.0 mL), and RG 12525 (3.0 g) at 80 °C was slowly added a 20% solution of aqueous citric acid until crystallization was complete at pH 5.3-5.4. After cooling to room temperature, the mixture was filtered and dried, giving white block-like crystals (2.9 g, 98% yield). MicroscopysSamples were viewed at magnifications ranging from 100× to 400× under cross-polarization using a Zeiss Axiophot polarizing light microscope. The refractive indices were determined by the immersion method using Cargille refractive index liquids. No attempt was made to align the indices with the interference figure. Rather the two indices represent the highest and lowest indices determined from crushed fragments. A Linkam hot stage (Model THM 600) was used with the optical microscope to observe melting behavior. The heating rate through the melting temperature was 4 °C/min for all the samples. The scanning electron microscope photographs were taken using a JEOL 6400F field emission SEM. The samples were sputter-coated with gold/palladium to improve conductivity. Differential Scanning CalorimetrysThe DSC thermograms were measured with a Perkin-Elmer DSC-7 thermoanalyzer at a heating rate of 5 °C/min under nitrogen. IR SpectroscopysIR spectra of samples from 4000 to 400 cm-1 were generated at 4 cm-1 resolution on a Nicolet Model 740 IR spectrophotometer as dispersions in a KBr pellet which were pressed at approximately 20 000 psi on a Carver press. Powder X-ray DiffractometrysA Siemens D5000 diffractometer with a Cu radiation source (1.8 kW, 45 kV, and 40 mA) was used to
0022-3549/96/3185-0461$12.00/0
Journal of Pharmaceutical Sciences / 461 Vol. 85, No. 5, May 1996
Table 1sSolvents Used to Recrystallize RG 12525 Solvent
Cooling Rate
Seed
Samplea
Methanol Methanol Isobutanol Isobutanol Acetonitrile Acetonitrile Dimethylformamide/water Dimethoxyethane Dimethoxyethane/water Dimethoxyethane/heptane
Slow Slow Fast Slow Fast Slow Fast Fast Fast Fast
Yes Yes No Yes No Yes No Yes Yes Yes
I (98%) II (98%) I (90%) II (80%) I (98%) I (60%) I (90%) Pure I Pure I Pure I
a I: form I. II: form II. Percentages are estimated from a combination of DSC and microscopy.
Table 2sTitration Procedures Used To Crystallize RG 12525 Cosolventa
Acid
Base
Temp (°C)
Sampleb
2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol PEG 200 Tween 80
HCl H2SO4 H2SO4 H2SO4 HOAc HOAc H2SO4 H2SO4 HOAc Citric Acid Citric Acid Citric Acid Citric Acid
NaOH NaOH K2CO3 Na2CO3 K3PO4 Na3PO4 K3PO4 LiOH NaOH NaOH NH4OH NH4OH NH4OH
50 35 30 30 34 45 35 25 50 55 80 80 80
Pure I I (98%) II (99%) II (80%) I (80%) 50% I/50% II I (80%) II (99%) Pure I Pure I Pure II Pure II II (98%)
a Other solvent is water. b I: form I. II: form II. Percentages are estimated from a combination of DSC and microscopy.
scan powder samples from 3 to 40° (2θ). Samples were gently crushed prior to the measurement to reduce particle size effects on the peak intensities. Approximately 60 mg of sample was loaded onto a 1.5 × 1.0 cm sample holder and scanned in the 3-40° 2θ range with a step size of 0.04° and a total exposure of 1 s per step size. Equilibrium SolubilitysThe equilibrium solubility determinations of the two polymorphs of RG 12525 were carried out in 50 mM sodium phosphate, pH ) 7.0. [Sodium phosphate was chosen as the buffer since temperature dependence of the activity (and hence pH) is minimal for sodium phosphate. Constant pH conditions are important to the experiments as the solubility of RG 12525 is highly dependent upon pH.] The solubility experiments were performed in duplicate at the following temperatures: 31, 38, 44, 50, 62, and 72 °C. Sample solutions were analyzed by HPLC at a time point sufficiently long to ensure that equilibrium had been attained. Examination of the solid samples at the end of the experiment indicated that no interconversion had taken place.
Figure 1sIR examples of each polymorphic form. Table 3sX-Ray Powder Diffraction Data for RG 12525 Polymorph I
Polymorph II
2θ (deg)
d Spacing (Å)
Rel Intensity (I/I0 × 100)
2θ (deg)
d Spacing (Å)
Rel Intensity (I/I0 × 100)
17.4 24.4 25.5 22.2 12.6 21.4 32.2 26.0 29.1 28.3
5.10 3.64 3.50 4.00 7.03 4.14 2.78 3.42 3.06 3.15
100.0 79.7 77.1 66.0 40.8 40.5 39.9 37.5 37.2 37.1
26.4 25.4 25.8 17.9 16.9 15.0 24.0 24.5 29.3 22.3
3.38 3.51 3.45 4.96 5.23 5.88 3.70 3.63 3.05 3.99
100.0 72.2 54.9 44.8 40.6 29.8 27.5 25.0 24.8 20.0
Results and Discussion Preparation of the RG 12525 PolymorphssTwo polymorphs of RG 12525 were first noted while recrystallization studies were being performed.2 Interestingly, two similar methanol recrystallization procedures were performed, but the crystals obtained from each procedure had very different filtration characteristics. Further investigation revealed that each product was a unique RG 12525 polymorphic form. A thorough discussion of physicochemical properties is included in the next section. Laboratory studies were performed to determine the key parameters governing crystallization of the isolated forms. Specific conditions were discovered for the reproducible preparation of each form. One of these key parameters was the presence of seed crystals at the onset of crystallization, and the other was a specific cooling rate. A crystallization procedure was developed for the preparation of form I, but usually a trace (up to 5%) of form II was present 462 / Journal of Pharmaceutical Sciences Vol. 85, No. 5, May 1996
as a contaminant. The key parameter for the preparation of form I was the use of seed crystals. Slow cooling was necessary. Under slightly modified conditions the preparation of form II also required seeding. In this procedure, slow cooling was necessary. Form II prepared from methanol crystallization was usually contaminated with form I (from 2 to 10%). Since pure form I and pure form II were not available using methanol recrystallization, a search for alternate procedures to prepare each form as a pure polymorph was initiated. A variety of crystallization solvents were tried, and the results are summarized in Table 1. Analogous to the methanol system, polymorphic mixtures arose from isobutanol. The primary form generated depends on seeding and the cooling rate. Other solvent systems which were used for recrystallization tended to favor the formation of form I. For instance,
Figure 2sPowder XRD examples of each polymorphic form.
when dimethoxyethane was used as a neat solvent, or in combination with another solvent, pure form I resulted on recrystallization. Titration techniques were also found to be successful in preparing each polymorphic form. However, the preparation of pure form II and pure form I was highly dependent on the chosen acid, base, cosolvent, and crystallization temperature. A summary of results is listed in Table 2. Different acids used for titration included hydrochloric, sulfuric, acetic, and citric acid. Several bases were also used including sodium and lithium hydroxide, sodium carbonate, sodium and potassium phosphate, and ammonium hydroxide. While sodium hydroxide seemed to cause preferred formation of form I, no
conclusive pattern emerged for the preferential formation of one polymorphic form using other acids and bases. Formation of pure form I was achieved by titrating a solution of the sodium salt of RG 12525 (generated with sodium hydroxide) with either hydrochloric acid, acetic acid, or citric acid. The use of 2-propanol as a cosolvent at a crystallization temperature of approximately 50-55 °C was found to be important for the formation of pure form I. These conditions do not require seeding, and spontaneous crystallization is preferred. The formation of pure form II was best achieved by titrating a solution of the ammonium salt of RG 12525 (generated with ammonium hydroxide) with citric acid. The use of seed crystals was important to prevent the
Journal of Pharmaceutical Sciences / 463 Vol. 85, No. 5, May 1996
Figure 3sDSC of each polymorphic form.
compound from oiling out of the mixture. Either PEG or 2-propanol could be used as a cosolvent to form pure form II. A crystallization temperature of 80 °C was necessary for exclusive formation of form II. Physicochemical Properties of the RG 12525 PolymorphssInfraredsEach polymorphic form exhibits unique spectral characteristics. The IR spectra (KBr) of the two forms are shown in Figure 1. No interconversion was observed due to the pressure needed to form a KBr pellet. This was determined from IR-microscopy (Spectra Tech IR PLAN) spectra obtained on crystals that were compressed and uncompressed in a Carver press. Distinguishing IR absorption peaks for form I are at 3145, 1070, 1017, and 821 cm-1. Distinguishing peaks for form II are a 1239/1225 doublet, an 825/814 doublet, and two broad bands at 2545 and 1960 cm-1. Each spectrum is different from the other and can be used for quick polymorphic identification. However, the differences 464 / Journal of Pharmaceutical Sciences Vol. 85, No. 5, May 1996
are not distinctive enough to allow for a sensitive and precise quantitative method. Powder XRDsThe powder XRD for each form is shown in Figure 2, and Table 3 lists information about some relevant peaks (2θ angles, d spacings, and relative intensities) for both forms in order of decreasing relative intensity. Significant differences are seen. Form I has peaks at 2θ angles of 8.7° and 32.2° which are not present for form II. The pattern for form II shows unique peaks at 2θ angles of 10.5°, 12.2°, 16.9°, 17.8°, and 26.4°. As can be seen from Figure 2 and Table 3, several other differences are also observed, thus making the diffraction patterns useful for form identification. Polymorphic Composition of MixturessA variety of techniques used to quantify the polymorphic composition were evaluated (such as microscopy, IR, Raman spectroscopy, powder XRD, and DSC). DSC became the method of choice as it could detect as little as 1% of form II in form I. The
Figure 4sDSC quantitation plot.
DSC thermogram of each polymorph is shown in Figure 3. The average melting points as recorded by DSC are 153.8 ( 0.8 °C and 157.0 ( 0.5 °C for forms I and II, respectively. When each form is pure, only one endotherm is noted when scanned from 150 to 180 °C. However, when a small amount of the other form is present, two distinct endotherms are noted. A method for quantifying form II in the presence of form I was developed and validated. A small amount of form II was placed into a tared DSC pan and then weighed; then form I was placed into the pan and the contents were gently mixed; and then the pan was again weighed. The entire pan was then carefully crimped, and placed into the DSC, and run, thus eliminating any inhomogeneity due to subsampling. The samples of forms I and II were sieved in an attempt to control the particle size of the fractions used for testing. The instrument was calibrated with respect to temperature and heat of fusion using indium and zinc. All samples were analyzed immediately after calibration. The thermograms were then analyzed for onset temperature and peak areas. From this was generated a calibration curve over the area of interest (see Figure 4) that allows form II to be quantified to a significantly low level. Prior to the analysis of the mixtures, both pure forms were analyzed separately. Plots of amount of pure form versus endotherm area were found to behave linearly, and no changes in the form were observed upon heating. Linear regression of the data in Figure 4 yielded an equation with an r2 value of 0.99, an intercept value of 0.0023, and a slope of 0.0144. MicroscopysPolymorph I is composed of long needles (see Figure 5) with length to width (aspect) ratios greater than 10:1 as estimated by microscopy. The needles appear to be stronger in length than breadth since they break lengthwise with the application of mechanical force. In most reductiontype operations, such as milling and micronizing, the fibrous character is maintained. Polymorph II is composed of rectangular blocks (see Figure 5). In reduction-type operations the block-like character is lost and the fragments are conchoidal. Some of these fragments are elongated with aspect ratios on the order of 4:1. Polymorphs I and II can often be distinguished on the basis of shape alone. Polymorphs I and II have different values of refractive index and birefringence (see Table 4). Polymorph I has low, firstorder white birefringence colors whereas form II has high second-order and higher order colors. This is evident from the numerical birefringence values. Polymorph I has a birefringence of 0.060 while II has a birefringence of 0.210. The individual indices of each polymorph are also quite different. The γ ′ values for forms I and II are 1.720 and 1.750, respectively. The R′ indices are even more distinctive, with
Figure 5sPhotomicrograph of each polymorphic form. Table 4sOptical Properties of RG 12525 Forms I and II Property
Form I
Form II
Refractive index, R′ Refractive index, γ′ Birefringence Extinction angle (deg) Sign of elongation Morphology
1.660 1.720 0.060 30−45 negative thin needles
1.540 1.750 0.210 45 negative blocks
form I having a value of 1.660 and form II having a value of 1.540. Birefringence and refractive index can be used to identify the polymorphs in a mixture. The hot-stage melting point of the two polymorphs differs by about 3 °C. Polymorph I has an average corrected melting point of 154.6 °C, whereas the average corrected melting point for polymorph II is 157.8 °C. This compares well with the average DSC melting points. Thermodynamic Properties of the RG 12525 PolymorphssThe heat of fusion rule can be used to determine whether a system is enantiotropic or monotropic. The rule states, “If the higher melting form has the lower heat of fusion, the two forms are usually enantiotropic, otherwise they are monotropic.” 3 From DSC data were determined the heats of fusion for each RG 12525 form. The heats of fusion, replicates of at least three runs for each pure form, were found to be 11.2 ( 0.1 kcal/mol and 10.3 ( 0.2 kcal/mol for forms I and II, respectively. In this system, the higher melting form II has the lowest heat of fusion. Therefore, according to the heat of fusion rule, the RG 12525 polymorphic system is expected to be enantiotropic.
Journal of Pharmaceutical Sciences / 465 Vol. 85, No. 5, May 1996
Figure 6sSolubility plots for polymorphs I and II in 50 mM sodium phosphate buffer (pH ) 7.0). Table 5sEquilibrium Solubility Data (n ) 2) for Polymorphs I and II of RG 12525 in 50 mM Sodium Phosphate Buffer (pH ) 7.0) Form I
Form II
Temp (°C)
Concn (µM)
SD
Concn (µM)
SD
31 38 44 50 62 72
11.84 23.67 41.58 78.21 275.9 578.8
0.49 0.03 1.13 0.61 4.50 17.30
14.88 26.52 48.54 90.62 292.0 591.2
0.65 0.09 0.02 1.89 0.78 6.00
One of the prime objectives of the comparison between the two polymorphs was to determine the more thermodynamically stable form at room temperature. The temperature/ dissolution relationship4 of the two polymorphs and the solution phase transformation test5 were used to determine that polymorph I is the more thermodynamically stable polymorph at room temperature. Solubility Determinations: Results and CalculationssThe temperature range of thermodynamic stability and various thermodynamic parameters can be determined from measurements of the individual polymorph equilibrium solubilities.6,7 Figure 6 shows a plot of the solubility of each polymorph as a function of temperature. The standard deviation for each set of measurements is generally too small to be shown graphically. The plot displays similar equilibrium solubilities for each polymorph throughout the temperature range studied, with polymorph I being slightly less soluble than polymorph II. Table 5 shows the solubility data for each polymorph at various temperatures. From the solubility data collected at various temperatures, classical van’t Hoff type plots6 were employed to determine thermodynamic values for the dissolution of each polymorph. For example, the solution enthalpies can be calculated from measurements of the equilibrium polymorph solubilities and the following:
ln Ci ) -∆Hisln/RT + constant
(1)
where C is the equilibrium solubility, i is polymorph I or II, and the constant is the y-intercept, which is related to the entropy of solution for each polymorph. The heat of solution, ∆Hisln, can be calculated from a plot of ln Ci vs 1/T, the van’t Hoff plot (see Figure 7). Table 6 shows the thermodynamic 466 / Journal of Pharmaceutical Sciences Vol. 85, No. 5, May 1996
Figure 7sLog of solubility of polymorphs in 50 mM sodium phosphate buffer (pH ) 7.0) vs reciprocal temperature. Table 6sSolution Thermodynamic Values Calculated for Polymorphs I and II of RG 12525 Derived from Solubility Measurements in 50 mM Sodium Phosphate Buffer (pH ) 7.0) and Calculated Values for the Transformation of Polymorph II to Polymorph I at 25 °C Polymorph
Enthalpy (kcal/mol)
I II II f I
19.99 (SE ) 0.3) 19.09 (SE ) 0.3) −0.91 (SE ) 0.4)
Free Energy at 25 °C (kcal/mol) 3.04 (SE ) 0.6) 2.90 (SE ) 0.6) −0.14 (SE ) 0.8)
Entropy (eu) 56.90 (SE ) 1.0) 54.33 (SE ) 1.0) −2.6 (SE ) 1.4)
Table 7sThermodynamic Values Calculated, from Solubility Measurements, for the Transformation of Polymorph II to Polymorph I of RG 12525 Temp (°C)
Free Energy (kcal/mol)
Temp (°C)
Free Energy (kcal/mol)
31 38 44
−0.14 −0.07 −0.10
50 62 72
−0.09 −0.04 −0.01
values derived from the equilibrium solubility measurements for each polymorph and their standard errors as determined from the slopes of the van’t Hoff plots. The values of the free energy of solution for each polymorph are given for room temperature (25 °C). Once the thermodynamic values from the solubility studies were determined, an estimate of the thermodynamic parameters for the transformation of polymorph I to polymorph II was made. Since the two polymorphs are indistinguishable in solution, the following equation applies:
∆HIfII ) ∆HIsln - ∆HIIsln
(2)
where ∆HIsln and ∆HIIsln are the heats of solution for polymorphs I and II, respectively, from Table 6. According to theory, the transformation from the low-temperature form to the high-temperature form is an endothermic reaction.8 Application of the above mentioned theory for this system, ∆HIfII ) +0.9 kcal/mol, indicates that I is the low-temperature form and II is the high-temperature form. From the solubility measurements and eq 3 one can calculate the free energy change at the measured temperatures (see Table 7). The free
∆GIIfI ) RT ln(CI/CII)
(3)
energy values indicate the order of thermodynamic stability of the two forms over the temperatures studied and indicate that the spontaneous direction is II f I over this temperature range. The free energy at 25 °C can be calculated by applying a temperature correction to the solubility measured at 31 °C. This can be done using a van’t Hoff type correction by applying eq 4 at the two temperatures:
ln χ ) -(∆Hf/R)((Tm - T)/TTm)
(4)
where χ is the mole fraction, T is the measured temperature, ∆Hf is the heat of fusion, and Tm is the melting point for a given polymorph. The calculated solubilities for forms I and II at 25 °C are 7.6 and 9.8 µM, respectively, and are in good agreement with the measured values of 5.9 and 7.1 µM, respectively, for forms I and II at 20 °C, thus placing the calculated values at 25°C, as expected, in between the measured values at 20 and 31 °C. From eq 3 one can then obtain:
∆GIIfI ) -0.15 (kcal/mol) at 25 °C
SD ) 0.03
which is in good agreement with the value of -0.14 kcal/mol calculated at 25 °C from the van’t Hoff plots (see Table 6). This confirms that at room temperature the most thermodynamically stable form is polymorph I and that the spontaneous reaction is from form II to form I. The entropy change for the polymorphic transformation (∆SIIfI) can be calculated from the entropy change of dissolution for each polymorph obtained from the y-intercepts of the van’t Hoff plots. Solution Phase TransformationsA method described by Haleblian and McCrone was used to determine the most stable polymorph at room temperature and to estimate the transition temperature.5 This method utilizes the fact that the more stable polymorph will also be the less soluble at a given temperature and pressure. If crystals of both polymorphs are present in a saturated solution of RG 12525, the more stable form will grow at the expense of the less stable one. The small difference in solubility between the two forms of RG 12525 presented a major experimental difficulty. The long time periods required for the transformations forced the development of sealed-cell methods for conducting these experiments. Two experiments for solution phase transformation were conducted in benzyl alcohol, one at room temperature and the other at 30 °C. In each experiment, polymorph I crystals appeared to be growing while polymorph II crystals were decreasing in size. After 1.5 h at room temperature, there was obvious growth of form I at the expense of form II. After being in a sealed cell for 1.5 weeks, all of the form II particles were surrounded and enclosed by form I needles, yet some form II did remain. At 30 °C, the transformation proceeded more rapidly with almost complete dissolution of form II. The kinetics for the conversion were not determined. Schematic Energy DiagramsFigure 8 presents a semiempirical graph of the energy vs temperature relationship of the two polymorphs. A free energy vs temperature graph is one of the best ways of showing the relationship among the various phases and their thermodynamic properties.3 The phase with the lowest free energy at any particular temperature will be the more thermodynamically stable form. The rationale for the lines in Figure 8 is as follows. From the DSC data the relative placement of the enthalpy lines for forms I and II relative to the enthalpy line for the liquid can be determined. Since the HI line is below the HII line, and since a good assumption3,8 is that these lines will have the same order from 0 K to the melting point, then their relative order at 0 K can
Figure 8sSchematic energy diagram for RG 12525 polymorphs I and II.
be determined. Again from the melting data it is known that polymorph II melts at a higher temperature than polymorph I and hence the GI line will cross the GL line at a lower temperature than the GII line. The solubility experiments allowed the free energy between the two forms at various temperatures to be determined and gave the relative order of those curves from 31 to 70 °C.
Conclusions Two polymorphs of RG 12525 have been identified and studied. Two preparative methods are described for each polymorphic form. Methods to differentiate the two forms are described, with DSC providing a way to quantitate polymorphic purity. It has been determined that form I is the more thermodynamically stable form at room temperature and is the low-temperature form. Polymorph II is the high-temperature form and exists in an enantiotropic system with form I. The solubilities of these two forms are very close to each other. The free energy difference between the two forms at room temperature is 0.15 kcal/mol.
References and Notes 1. Huang, F.-C.; Galemmo, R. A., Jr.; Johnson, W. H., Jr.; Poli, G. B.; Morrissette, M. M.; Mencel, J. J.; Warus, J. D.; Campbell, H. F.; Nuss, G. W.; Carnathan, G. W.; Van Inwegen, R. G. J. Med. Chem. 1990, 33, 1194-1200. 2. Goetzen, T. Unpublished results. 3. Burger, A.; Ramberger, R. Mikrochim. Acta 1979, II, 259-271. 4. Wadke, D. A.; Reier, G. E. J. Pharm. Sci. 1972, 61, 868-871. 5. Haleblian, F.; McCrone, W. J. Pharm. Sci. 1969, 58, 911-929. 6. Aguiar, A. J.; Zelmer, J. E. J. Pharm. Sci. 1969, 58, 983-987. 7. Bettinetti, G. P. Farmaco 1988, 43, 71-99. 8. Buerger, M. J. Crystallographic Aspects of Phase Transformations. In Phase Transformation in Solids; Smoluchowski, R., Mayer, J. E., Weyl, W. A., Eds.; Proceedings of Symposium at Cornell on Aug 23-26, 1948; Chapter 6, pp 183-211.
Acknowledgments The authors would like to thank Patricia Logue (powder XRD), Walter Hause, Jr. (DSC quantification), and Ed Orton (IR spectroscopy), for their contributions to this report and David J. W. Grant for his helpful comments during the preparation of this paper. The authors would also like to express appreciation to Tom Goetzen and Jean-Paul Richard for their contributions in the early stages of this investigation. The authors would also like to thank Jim Kerner of Lehigh University for his help with the scanning electron microscopy.
JS950265P
Journal of Pharmaceutical Sciences / 467 Vol. 85, No. 5, May 1996
AND
MICHAEL D. THOMPSONX
Received June 26, 1995, from the Rhoˆne-Poulenc Rorer, 500 Arcola Road, Collegeville, PA 19426. Final revised manuscript ‡ Present address: Warner-Lambert, received January 23, 1996. Accepted for publication January 23, 1996X. Morris Plains, NJ 07950. Abstract 0 This report describes the preparation and characterization of two polymorphic forms of RG 12525, a leukotriene D4 (LTD4) antagonist. Polymorph I is prepared by recrystallization from methanol or titration of the sodium salt of RG 12525 with citric acid. Polymorph II is prepared by recrystallization from methanol or titration of the ammonium salt of RG 12525 with citric acid. The polymorphic system is enantiotropic, with pure form I melting at 154 °C, 3 deg less than the melting temperature of form II. Form I is thermodynamically more stable than form II at room temperature. These polymorphic forms are differentiated using microscopy, differential scanning calorimetry (DSC), infrared spectroscopy (IR), and powder X-ray diffraction (XRD) analysis. Solubility properties from 31 to 72 °C were determined to be similar for both forms. The calculated solubilities at 25 °C are 7.6 and 9.8 µM for forms I and II, respectively. The free energy change from form II to form I at 25 °C is −0.15 kcal/ mol. Thermodynamic properties of the system are summarized using a schematic free energy diagram.
Introduction RG 12525, an LTD4 antagonist, has been shown to have activity against leukotriene-induced bronchoconstriction. Further study of the physical properties of this compound revealed that RG 12525 has two stable polymorphic forms. This report describes the preparation and characterization of each polymorphic form and presents the results of studies undertaken to establish the relationship between the two forms.
Both forms have been prepared by crystallization from methanol. In addition, a titration method for the crystallization of each form is reported. The preparative conditions and experiments preceding their discovery are described. Solid state characteristics of the polymorphic forms are also included. Each form exhibits unique properties when examined through optical microscopy, differential scanning calorimetry (DSC), infrared spectroscopy (IR), solid state studies, and powder X-ray diffraction (XRD). The solubility properties of the two polymorphs are quite similar. Information gathered from solubility and DSC studies indicates that polymorph I is the more thermodynamically stable form at room temperature. These studies include a determination of the thermodynamically stable temperature ranges, identification of the high- and low-temperature forms, enantiotropism vs monoX
Abstract published in Advance ACS Abstracts, March 1, 1996.
© 1996, American Chemical Society and American Pharmaceutical Association
tropism, and a schematic free energy diagram. A comparison of optical and thermal properties is also given.
Experimental Section MaterialssRG 12525, 2-[[4-[[2-(1H-tetrazol-5-ylmethyl)phenyl]methoxy]phenoxy]methyl]quinoline, was prepared1 and subsequently used to investigate polymorph preparations. All other chemicals used were reagent grade. General Recrystallization MethodsRG 12525 was dissolved in refluxing solvent. The solution was seeded (if required) and then cooled to 0 °C. The mixture was filtered and dried, giving product. General Titration MethodsA solution of acid in water and cosolvent was added slowly to a solution of RG 12525 and aqueous base solution with a cosolvent at the given temperature. After crystallization, the mixture was cooled to room temperature, filtered, and dried, giving product. Preparation of RG 12525 Form I by Methanol RecrystallizationsRG 12525 (16.0 g) was dissolved in refluxing methanol (400 g). The solution was cooled to 50 °C and immediately seeded with form I. After 1 h, the mixture was cooled to 0 °C and filtered. Drying gave white needle-like fibers (15.0 g, 94% yield). Preparation of RG 12525 Form I by TitrationsTo a 7% solution of aqueous citric acid (12 g) in 2-propanol at 55 °C were added seed crystals of form I. To this mixture was slowly added a solution of water (13.0 mL), 2-propanol (11.0 mL), 50% aqueous sodium hydroxide (5.8 g), and RG 12525 (3.0 g). After addition, the mixture was cooled to room temperature, filtered, and dried, giving white needle-like fibers (2.7 g, 90% yield). Preparation of RG 12525 Form II by Methanol RecrystallizationsRG 12525 (3.0 g) was dissolved in refluxing methanol (75 g). The solution was cooled to 50 °C and seeded with form II. After cooling to 0 °C, the mixture was filtered and dried, giving white blocklike crystals (2.2 g, 75% yield). Preparation of RG 12525 Form II by TitrationsTo a solution of water (25.0 mL), PEG-200 (5.6 g), concentrated ammonium hydroxide (1.0 mL), and RG 12525 (3.0 g) at 80 °C was slowly added a 20% solution of aqueous citric acid until crystallization was complete at pH 5.3-5.4. After cooling to room temperature, the mixture was filtered and dried, giving white block-like crystals (2.9 g, 98% yield). MicroscopysSamples were viewed at magnifications ranging from 100× to 400× under cross-polarization using a Zeiss Axiophot polarizing light microscope. The refractive indices were determined by the immersion method using Cargille refractive index liquids. No attempt was made to align the indices with the interference figure. Rather the two indices represent the highest and lowest indices determined from crushed fragments. A Linkam hot stage (Model THM 600) was used with the optical microscope to observe melting behavior. The heating rate through the melting temperature was 4 °C/min for all the samples. The scanning electron microscope photographs were taken using a JEOL 6400F field emission SEM. The samples were sputter-coated with gold/palladium to improve conductivity. Differential Scanning CalorimetrysThe DSC thermograms were measured with a Perkin-Elmer DSC-7 thermoanalyzer at a heating rate of 5 °C/min under nitrogen. IR SpectroscopysIR spectra of samples from 4000 to 400 cm-1 were generated at 4 cm-1 resolution on a Nicolet Model 740 IR spectrophotometer as dispersions in a KBr pellet which were pressed at approximately 20 000 psi on a Carver press. Powder X-ray DiffractometrysA Siemens D5000 diffractometer with a Cu radiation source (1.8 kW, 45 kV, and 40 mA) was used to
0022-3549/96/3185-0461$12.00/0
Journal of Pharmaceutical Sciences / 461 Vol. 85, No. 5, May 1996
Table 1sSolvents Used to Recrystallize RG 12525 Solvent
Cooling Rate
Seed
Samplea
Methanol Methanol Isobutanol Isobutanol Acetonitrile Acetonitrile Dimethylformamide/water Dimethoxyethane Dimethoxyethane/water Dimethoxyethane/heptane
Slow Slow Fast Slow Fast Slow Fast Fast Fast Fast
Yes Yes No Yes No Yes No Yes Yes Yes
I (98%) II (98%) I (90%) II (80%) I (98%) I (60%) I (90%) Pure I Pure I Pure I
a I: form I. II: form II. Percentages are estimated from a combination of DSC and microscopy.
Table 2sTitration Procedures Used To Crystallize RG 12525 Cosolventa
Acid
Base
Temp (°C)
Sampleb
2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol 2-Propanol PEG 200 Tween 80
HCl H2SO4 H2SO4 H2SO4 HOAc HOAc H2SO4 H2SO4 HOAc Citric Acid Citric Acid Citric Acid Citric Acid
NaOH NaOH K2CO3 Na2CO3 K3PO4 Na3PO4 K3PO4 LiOH NaOH NaOH NH4OH NH4OH NH4OH
50 35 30 30 34 45 35 25 50 55 80 80 80
Pure I I (98%) II (99%) II (80%) I (80%) 50% I/50% II I (80%) II (99%) Pure I Pure I Pure II Pure II II (98%)
a Other solvent is water. b I: form I. II: form II. Percentages are estimated from a combination of DSC and microscopy.
scan powder samples from 3 to 40° (2θ). Samples were gently crushed prior to the measurement to reduce particle size effects on the peak intensities. Approximately 60 mg of sample was loaded onto a 1.5 × 1.0 cm sample holder and scanned in the 3-40° 2θ range with a step size of 0.04° and a total exposure of 1 s per step size. Equilibrium SolubilitysThe equilibrium solubility determinations of the two polymorphs of RG 12525 were carried out in 50 mM sodium phosphate, pH ) 7.0. [Sodium phosphate was chosen as the buffer since temperature dependence of the activity (and hence pH) is minimal for sodium phosphate. Constant pH conditions are important to the experiments as the solubility of RG 12525 is highly dependent upon pH.] The solubility experiments were performed in duplicate at the following temperatures: 31, 38, 44, 50, 62, and 72 °C. Sample solutions were analyzed by HPLC at a time point sufficiently long to ensure that equilibrium had been attained. Examination of the solid samples at the end of the experiment indicated that no interconversion had taken place.
Figure 1sIR examples of each polymorphic form. Table 3sX-Ray Powder Diffraction Data for RG 12525 Polymorph I
Polymorph II
2θ (deg)
d Spacing (Å)
Rel Intensity (I/I0 × 100)
2θ (deg)
d Spacing (Å)
Rel Intensity (I/I0 × 100)
17.4 24.4 25.5 22.2 12.6 21.4 32.2 26.0 29.1 28.3
5.10 3.64 3.50 4.00 7.03 4.14 2.78 3.42 3.06 3.15
100.0 79.7 77.1 66.0 40.8 40.5 39.9 37.5 37.2 37.1
26.4 25.4 25.8 17.9 16.9 15.0 24.0 24.5 29.3 22.3
3.38 3.51 3.45 4.96 5.23 5.88 3.70 3.63 3.05 3.99
100.0 72.2 54.9 44.8 40.6 29.8 27.5 25.0 24.8 20.0
Results and Discussion Preparation of the RG 12525 PolymorphssTwo polymorphs of RG 12525 were first noted while recrystallization studies were being performed.2 Interestingly, two similar methanol recrystallization procedures were performed, but the crystals obtained from each procedure had very different filtration characteristics. Further investigation revealed that each product was a unique RG 12525 polymorphic form. A thorough discussion of physicochemical properties is included in the next section. Laboratory studies were performed to determine the key parameters governing crystallization of the isolated forms. Specific conditions were discovered for the reproducible preparation of each form. One of these key parameters was the presence of seed crystals at the onset of crystallization, and the other was a specific cooling rate. A crystallization procedure was developed for the preparation of form I, but usually a trace (up to 5%) of form II was present 462 / Journal of Pharmaceutical Sciences Vol. 85, No. 5, May 1996
as a contaminant. The key parameter for the preparation of form I was the use of seed crystals. Slow cooling was necessary. Under slightly modified conditions the preparation of form II also required seeding. In this procedure, slow cooling was necessary. Form II prepared from methanol crystallization was usually contaminated with form I (from 2 to 10%). Since pure form I and pure form II were not available using methanol recrystallization, a search for alternate procedures to prepare each form as a pure polymorph was initiated. A variety of crystallization solvents were tried, and the results are summarized in Table 1. Analogous to the methanol system, polymorphic mixtures arose from isobutanol. The primary form generated depends on seeding and the cooling rate. Other solvent systems which were used for recrystallization tended to favor the formation of form I. For instance,
Figure 2sPowder XRD examples of each polymorphic form.
when dimethoxyethane was used as a neat solvent, or in combination with another solvent, pure form I resulted on recrystallization. Titration techniques were also found to be successful in preparing each polymorphic form. However, the preparation of pure form II and pure form I was highly dependent on the chosen acid, base, cosolvent, and crystallization temperature. A summary of results is listed in Table 2. Different acids used for titration included hydrochloric, sulfuric, acetic, and citric acid. Several bases were also used including sodium and lithium hydroxide, sodium carbonate, sodium and potassium phosphate, and ammonium hydroxide. While sodium hydroxide seemed to cause preferred formation of form I, no
conclusive pattern emerged for the preferential formation of one polymorphic form using other acids and bases. Formation of pure form I was achieved by titrating a solution of the sodium salt of RG 12525 (generated with sodium hydroxide) with either hydrochloric acid, acetic acid, or citric acid. The use of 2-propanol as a cosolvent at a crystallization temperature of approximately 50-55 °C was found to be important for the formation of pure form I. These conditions do not require seeding, and spontaneous crystallization is preferred. The formation of pure form II was best achieved by titrating a solution of the ammonium salt of RG 12525 (generated with ammonium hydroxide) with citric acid. The use of seed crystals was important to prevent the
Journal of Pharmaceutical Sciences / 463 Vol. 85, No. 5, May 1996
Figure 3sDSC of each polymorphic form.
compound from oiling out of the mixture. Either PEG or 2-propanol could be used as a cosolvent to form pure form II. A crystallization temperature of 80 °C was necessary for exclusive formation of form II. Physicochemical Properties of the RG 12525 PolymorphssInfraredsEach polymorphic form exhibits unique spectral characteristics. The IR spectra (KBr) of the two forms are shown in Figure 1. No interconversion was observed due to the pressure needed to form a KBr pellet. This was determined from IR-microscopy (Spectra Tech IR PLAN) spectra obtained on crystals that were compressed and uncompressed in a Carver press. Distinguishing IR absorption peaks for form I are at 3145, 1070, 1017, and 821 cm-1. Distinguishing peaks for form II are a 1239/1225 doublet, an 825/814 doublet, and two broad bands at 2545 and 1960 cm-1. Each spectrum is different from the other and can be used for quick polymorphic identification. However, the differences 464 / Journal of Pharmaceutical Sciences Vol. 85, No. 5, May 1996
are not distinctive enough to allow for a sensitive and precise quantitative method. Powder XRDsThe powder XRD for each form is shown in Figure 2, and Table 3 lists information about some relevant peaks (2θ angles, d spacings, and relative intensities) for both forms in order of decreasing relative intensity. Significant differences are seen. Form I has peaks at 2θ angles of 8.7° and 32.2° which are not present for form II. The pattern for form II shows unique peaks at 2θ angles of 10.5°, 12.2°, 16.9°, 17.8°, and 26.4°. As can be seen from Figure 2 and Table 3, several other differences are also observed, thus making the diffraction patterns useful for form identification. Polymorphic Composition of MixturessA variety of techniques used to quantify the polymorphic composition were evaluated (such as microscopy, IR, Raman spectroscopy, powder XRD, and DSC). DSC became the method of choice as it could detect as little as 1% of form II in form I. The
Figure 4sDSC quantitation plot.
DSC thermogram of each polymorph is shown in Figure 3. The average melting points as recorded by DSC are 153.8 ( 0.8 °C and 157.0 ( 0.5 °C for forms I and II, respectively. When each form is pure, only one endotherm is noted when scanned from 150 to 180 °C. However, when a small amount of the other form is present, two distinct endotherms are noted. A method for quantifying form II in the presence of form I was developed and validated. A small amount of form II was placed into a tared DSC pan and then weighed; then form I was placed into the pan and the contents were gently mixed; and then the pan was again weighed. The entire pan was then carefully crimped, and placed into the DSC, and run, thus eliminating any inhomogeneity due to subsampling. The samples of forms I and II were sieved in an attempt to control the particle size of the fractions used for testing. The instrument was calibrated with respect to temperature and heat of fusion using indium and zinc. All samples were analyzed immediately after calibration. The thermograms were then analyzed for onset temperature and peak areas. From this was generated a calibration curve over the area of interest (see Figure 4) that allows form II to be quantified to a significantly low level. Prior to the analysis of the mixtures, both pure forms were analyzed separately. Plots of amount of pure form versus endotherm area were found to behave linearly, and no changes in the form were observed upon heating. Linear regression of the data in Figure 4 yielded an equation with an r2 value of 0.99, an intercept value of 0.0023, and a slope of 0.0144. MicroscopysPolymorph I is composed of long needles (see Figure 5) with length to width (aspect) ratios greater than 10:1 as estimated by microscopy. The needles appear to be stronger in length than breadth since they break lengthwise with the application of mechanical force. In most reductiontype operations, such as milling and micronizing, the fibrous character is maintained. Polymorph II is composed of rectangular blocks (see Figure 5). In reduction-type operations the block-like character is lost and the fragments are conchoidal. Some of these fragments are elongated with aspect ratios on the order of 4:1. Polymorphs I and II can often be distinguished on the basis of shape alone. Polymorphs I and II have different values of refractive index and birefringence (see Table 4). Polymorph I has low, firstorder white birefringence colors whereas form II has high second-order and higher order colors. This is evident from the numerical birefringence values. Polymorph I has a birefringence of 0.060 while II has a birefringence of 0.210. The individual indices of each polymorph are also quite different. The γ ′ values for forms I and II are 1.720 and 1.750, respectively. The R′ indices are even more distinctive, with
Figure 5sPhotomicrograph of each polymorphic form. Table 4sOptical Properties of RG 12525 Forms I and II Property
Form I
Form II
Refractive index, R′ Refractive index, γ′ Birefringence Extinction angle (deg) Sign of elongation Morphology
1.660 1.720 0.060 30−45 negative thin needles
1.540 1.750 0.210 45 negative blocks
form I having a value of 1.660 and form II having a value of 1.540. Birefringence and refractive index can be used to identify the polymorphs in a mixture. The hot-stage melting point of the two polymorphs differs by about 3 °C. Polymorph I has an average corrected melting point of 154.6 °C, whereas the average corrected melting point for polymorph II is 157.8 °C. This compares well with the average DSC melting points. Thermodynamic Properties of the RG 12525 PolymorphssThe heat of fusion rule can be used to determine whether a system is enantiotropic or monotropic. The rule states, “If the higher melting form has the lower heat of fusion, the two forms are usually enantiotropic, otherwise they are monotropic.” 3 From DSC data were determined the heats of fusion for each RG 12525 form. The heats of fusion, replicates of at least three runs for each pure form, were found to be 11.2 ( 0.1 kcal/mol and 10.3 ( 0.2 kcal/mol for forms I and II, respectively. In this system, the higher melting form II has the lowest heat of fusion. Therefore, according to the heat of fusion rule, the RG 12525 polymorphic system is expected to be enantiotropic.
Journal of Pharmaceutical Sciences / 465 Vol. 85, No. 5, May 1996
Figure 6sSolubility plots for polymorphs I and II in 50 mM sodium phosphate buffer (pH ) 7.0). Table 5sEquilibrium Solubility Data (n ) 2) for Polymorphs I and II of RG 12525 in 50 mM Sodium Phosphate Buffer (pH ) 7.0) Form I
Form II
Temp (°C)
Concn (µM)
SD
Concn (µM)
SD
31 38 44 50 62 72
11.84 23.67 41.58 78.21 275.9 578.8
0.49 0.03 1.13 0.61 4.50 17.30
14.88 26.52 48.54 90.62 292.0 591.2
0.65 0.09 0.02 1.89 0.78 6.00
One of the prime objectives of the comparison between the two polymorphs was to determine the more thermodynamically stable form at room temperature. The temperature/ dissolution relationship4 of the two polymorphs and the solution phase transformation test5 were used to determine that polymorph I is the more thermodynamically stable polymorph at room temperature. Solubility Determinations: Results and CalculationssThe temperature range of thermodynamic stability and various thermodynamic parameters can be determined from measurements of the individual polymorph equilibrium solubilities.6,7 Figure 6 shows a plot of the solubility of each polymorph as a function of temperature. The standard deviation for each set of measurements is generally too small to be shown graphically. The plot displays similar equilibrium solubilities for each polymorph throughout the temperature range studied, with polymorph I being slightly less soluble than polymorph II. Table 5 shows the solubility data for each polymorph at various temperatures. From the solubility data collected at various temperatures, classical van’t Hoff type plots6 were employed to determine thermodynamic values for the dissolution of each polymorph. For example, the solution enthalpies can be calculated from measurements of the equilibrium polymorph solubilities and the following:
ln Ci ) -∆Hisln/RT + constant
(1)
where C is the equilibrium solubility, i is polymorph I or II, and the constant is the y-intercept, which is related to the entropy of solution for each polymorph. The heat of solution, ∆Hisln, can be calculated from a plot of ln Ci vs 1/T, the van’t Hoff plot (see Figure 7). Table 6 shows the thermodynamic 466 / Journal of Pharmaceutical Sciences Vol. 85, No. 5, May 1996
Figure 7sLog of solubility of polymorphs in 50 mM sodium phosphate buffer (pH ) 7.0) vs reciprocal temperature. Table 6sSolution Thermodynamic Values Calculated for Polymorphs I and II of RG 12525 Derived from Solubility Measurements in 50 mM Sodium Phosphate Buffer (pH ) 7.0) and Calculated Values for the Transformation of Polymorph II to Polymorph I at 25 °C Polymorph
Enthalpy (kcal/mol)
I II II f I
19.99 (SE ) 0.3) 19.09 (SE ) 0.3) −0.91 (SE ) 0.4)
Free Energy at 25 °C (kcal/mol) 3.04 (SE ) 0.6) 2.90 (SE ) 0.6) −0.14 (SE ) 0.8)
Entropy (eu) 56.90 (SE ) 1.0) 54.33 (SE ) 1.0) −2.6 (SE ) 1.4)
Table 7sThermodynamic Values Calculated, from Solubility Measurements, for the Transformation of Polymorph II to Polymorph I of RG 12525 Temp (°C)
Free Energy (kcal/mol)
Temp (°C)
Free Energy (kcal/mol)
31 38 44
−0.14 −0.07 −0.10
50 62 72
−0.09 −0.04 −0.01
values derived from the equilibrium solubility measurements for each polymorph and their standard errors as determined from the slopes of the van’t Hoff plots. The values of the free energy of solution for each polymorph are given for room temperature (25 °C). Once the thermodynamic values from the solubility studies were determined, an estimate of the thermodynamic parameters for the transformation of polymorph I to polymorph II was made. Since the two polymorphs are indistinguishable in solution, the following equation applies:
∆HIfII ) ∆HIsln - ∆HIIsln
(2)
where ∆HIsln and ∆HIIsln are the heats of solution for polymorphs I and II, respectively, from Table 6. According to theory, the transformation from the low-temperature form to the high-temperature form is an endothermic reaction.8 Application of the above mentioned theory for this system, ∆HIfII ) +0.9 kcal/mol, indicates that I is the low-temperature form and II is the high-temperature form. From the solubility measurements and eq 3 one can calculate the free energy change at the measured temperatures (see Table 7). The free
∆GIIfI ) RT ln(CI/CII)
(3)
energy values indicate the order of thermodynamic stability of the two forms over the temperatures studied and indicate that the spontaneous direction is II f I over this temperature range. The free energy at 25 °C can be calculated by applying a temperature correction to the solubility measured at 31 °C. This can be done using a van’t Hoff type correction by applying eq 4 at the two temperatures:
ln χ ) -(∆Hf/R)((Tm - T)/TTm)
(4)
where χ is the mole fraction, T is the measured temperature, ∆Hf is the heat of fusion, and Tm is the melting point for a given polymorph. The calculated solubilities for forms I and II at 25 °C are 7.6 and 9.8 µM, respectively, and are in good agreement with the measured values of 5.9 and 7.1 µM, respectively, for forms I and II at 20 °C, thus placing the calculated values at 25°C, as expected, in between the measured values at 20 and 31 °C. From eq 3 one can then obtain:
∆GIIfI ) -0.15 (kcal/mol) at 25 °C
SD ) 0.03
which is in good agreement with the value of -0.14 kcal/mol calculated at 25 °C from the van’t Hoff plots (see Table 6). This confirms that at room temperature the most thermodynamically stable form is polymorph I and that the spontaneous reaction is from form II to form I. The entropy change for the polymorphic transformation (∆SIIfI) can be calculated from the entropy change of dissolution for each polymorph obtained from the y-intercepts of the van’t Hoff plots. Solution Phase TransformationsA method described by Haleblian and McCrone was used to determine the most stable polymorph at room temperature and to estimate the transition temperature.5 This method utilizes the fact that the more stable polymorph will also be the less soluble at a given temperature and pressure. If crystals of both polymorphs are present in a saturated solution of RG 12525, the more stable form will grow at the expense of the less stable one. The small difference in solubility between the two forms of RG 12525 presented a major experimental difficulty. The long time periods required for the transformations forced the development of sealed-cell methods for conducting these experiments. Two experiments for solution phase transformation were conducted in benzyl alcohol, one at room temperature and the other at 30 °C. In each experiment, polymorph I crystals appeared to be growing while polymorph II crystals were decreasing in size. After 1.5 h at room temperature, there was obvious growth of form I at the expense of form II. After being in a sealed cell for 1.5 weeks, all of the form II particles were surrounded and enclosed by form I needles, yet some form II did remain. At 30 °C, the transformation proceeded more rapidly with almost complete dissolution of form II. The kinetics for the conversion were not determined. Schematic Energy DiagramsFigure 8 presents a semiempirical graph of the energy vs temperature relationship of the two polymorphs. A free energy vs temperature graph is one of the best ways of showing the relationship among the various phases and their thermodynamic properties.3 The phase with the lowest free energy at any particular temperature will be the more thermodynamically stable form. The rationale for the lines in Figure 8 is as follows. From the DSC data the relative placement of the enthalpy lines for forms I and II relative to the enthalpy line for the liquid can be determined. Since the HI line is below the HII line, and since a good assumption3,8 is that these lines will have the same order from 0 K to the melting point, then their relative order at 0 K can
Figure 8sSchematic energy diagram for RG 12525 polymorphs I and II.
be determined. Again from the melting data it is known that polymorph II melts at a higher temperature than polymorph I and hence the GI line will cross the GL line at a lower temperature than the GII line. The solubility experiments allowed the free energy between the two forms at various temperatures to be determined and gave the relative order of those curves from 31 to 70 °C.
Conclusions Two polymorphs of RG 12525 have been identified and studied. Two preparative methods are described for each polymorphic form. Methods to differentiate the two forms are described, with DSC providing a way to quantitate polymorphic purity. It has been determined that form I is the more thermodynamically stable form at room temperature and is the low-temperature form. Polymorph II is the high-temperature form and exists in an enantiotropic system with form I. The solubilities of these two forms are very close to each other. The free energy difference between the two forms at room temperature is 0.15 kcal/mol.
References and Notes 1. Huang, F.-C.; Galemmo, R. A., Jr.; Johnson, W. H., Jr.; Poli, G. B.; Morrissette, M. M.; Mencel, J. J.; Warus, J. D.; Campbell, H. F.; Nuss, G. W.; Carnathan, G. W.; Van Inwegen, R. G. J. Med. Chem. 1990, 33, 1194-1200. 2. Goetzen, T. Unpublished results. 3. Burger, A.; Ramberger, R. Mikrochim. Acta 1979, II, 259-271. 4. Wadke, D. A.; Reier, G. E. J. Pharm. Sci. 1972, 61, 868-871. 5. Haleblian, F.; McCrone, W. J. Pharm. Sci. 1969, 58, 911-929. 6. Aguiar, A. J.; Zelmer, J. E. J. Pharm. Sci. 1969, 58, 983-987. 7. Bettinetti, G. P. Farmaco 1988, 43, 71-99. 8. Buerger, M. J. Crystallographic Aspects of Phase Transformations. In Phase Transformation in Solids; Smoluchowski, R., Mayer, J. E., Weyl, W. A., Eds.; Proceedings of Symposium at Cornell on Aug 23-26, 1948; Chapter 6, pp 183-211.
Acknowledgments The authors would like to thank Patricia Logue (powder XRD), Walter Hause, Jr. (DSC quantification), and Ed Orton (IR spectroscopy), for their contributions to this report and David J. W. Grant for his helpful comments during the preparation of this paper. The authors would also like to express appreciation to Tom Goetzen and Jean-Paul Richard for their contributions in the early stages of this investigation. The authors would also like to thank Jim Kerner of Lehigh University for his help with the scanning electron microscopy.
JS950265P
Journal of Pharmaceutical Sciences / 467 Vol. 85, No. 5, May 1996