Tu Lee, PhD

The solid state of an active pharmaceutical ingredient (API), like that of all materials, possesses functional and processing properties. Although much attention is paid to APIs' functional properties such as aqueous solubility, pharmacokinetics, and bioavailability, many processing properties of the solid state of an API (e.g., solubility in organic solvents, shape factor, and particle size) are also important to the success of pharmaceutical manufacturing operations such as mixing, wet granulation, and tableting (see Table I) (1–4). Therefore, in addition to the two essential roles of crystallization (i.e., a purification-and-separation step for the upstream organic synthesis and an adapting-linkage step for various outsourcing and globalization schemes), one other indispensable role of crystallization is a bottom-up material-fabrication step, from polymorphism in a subnanometer scale to crystal habit at a micrometer level, which determines processing and functional properties and the fate of all of the following downstream steps from filtration to dissolution. 

Table I: Polymorphism and crystal-habit-related processing properties of downstream unit operations in pharmaceutical production.			

Because solvent-mediated chemical reactions, nucleation, and phase transformation are common bottom-up approaches to traditional batch crystallization, the solvent's effects on API solubility, polymorphism, and crystal habits have become crucial (5). Most of the small-scale and solvent-evaporation-based high-throughput screening methods can provide primary information on solubility and polymorphism. However, data on crystal habits of an API are considered to be secondary (6). To change this perception, the authors developed a scalable initial solvent-screening method for laboratory-scale research using 23 to 25 common solvents to systematically determine solubility, polymorphism, and crystal habits of acetaminophen, (R,S)-(±)-ibuprofen,7 (R,S)-(±)-sodium ibuprofen dihydrate, tris(8-hydroxyquinoline)aluminum(III) (Alq3), and sulfathiazole through cooling crystallization (7–10). 

After the success of the three binary and 10 tertiary solvent mixtures in enhancing the solubility and inducing polymorphs of sulfathiazole, the authors decided to continue exploring all 100 cosolvent combinations encompassed by the symmetrical form space of acetaminophen (paracetamol) to see how the solubility, polymorphism, and crystal habits of acetaminophen are affected by the cosolvent systems (7, 10). The cosolvent combinations are shown as navy blue boxes within a red triangle that represents the symmetrical form space of acetaminophen in Figure 1. The authors chose acetaminophen as a model API because of its worldwide commercial value in analgesic and antipyretic therapy, previous experiences with it, and the recent discovery of the thermodynamically metastable Forms II and III crystals (7, 11–15). 

Figure 1: Solvent miscibility, cosolvent, and antisolvent systems of acetaminophen Form I crystals. (ALL FIGURES ARE COURTESY OF THE AUTHORS)			

 All materials and methods were the same as in the authors' previous experiments (7). Acetone [CH

, high-performance liquid chromatography (HPLC)–spectrometry grade, 99.9%, Lot: 601109], 

, HPLC–spectrometry grade, 99.4%, Lot: 111110], isopropyl alcohol (IPA) [(CH

CHOH, HPLC–spectrometry grade, 99.8%, Lot: 808907], methanol (CH

OH, HPLC–spectrometry grade, 99.9%, Lot: 703416), methyl tert-butyl ether (MTBE) [(CH

, certified grade, 99.9%, Lot: 712032], and methyl ethyl ketone (MEK) (C

, ACS grade, 99.6%, Lot: 201021) were purchased from Tedia (Fairfield, OH). Chloroform (CHCl

, ACS grade, 99.8%, Lot: E705279], tetrahydrofuran (THF) (C

O, HPLC–spectrometry grade, 99%, Lot: E704198), and ethanol (CH3CH2OH, HPLC–spectrometry grade, 99.5%) were obtained from Echo (Miaoli, Taiwan). Acetonitrile (CH

CN, analytical grade, 99.96%, Lot: 0043 X29B30), toluene (C

, ACS grade, 100%, Lot: B46755), and xylenes [C

, ACS grade, 98.8%, Lot: E03B34) were all received from Mallinckrodt Baker (Phillipsburg, NJ). Benzene (C

, ACS grade, 99%, Lot: 310008) and benzyl alcohol (C

OH, gas chromatography grade, 99.9%, Lot: 70950) were purchased from Sigma-Aldrich Laborchemikalien (Seelze, Germany). Dimethyl sulfoxide (DMSO) [(CH

SO, HPLC grade, 99.8%, Lot: SU0155] was obtained from Scharlau Chemie (Barcelona). 

, ACS grade, 99%, Lot: A0213203001], nitrobenzene (C

, ACS grade, 99%, Lot: A0195683001), and 

 Approximately 50 mg of acetaminophen Form I crystals were weighed in a 20-mL scintillation vial. From 0.2 to 0.5 mL of cosolvent was titrated carefully by micropipette into the vial with intermittent shaking for 2 min until all API solids were just dissolved. The solubility of the API at a given temperature was calculated as the weight of the API in a vial divided by the total volume of the cosolvent added to a vial (i.e., the gravimetric method) (7, 9). Solubility of the API in the same cosolvent at temperatures of 15, 25, 40, and 60 °C was determined. All temperatures were maintained and controlled by a water bath. Although the gravimetric method appeared to be imprecise, its advantages were its robustness, simplicity, lack of the need for calibration, and lack of solvate formation. Polymorphism, crystallinity, and crystal habits of acetaminophen solids recrystallized by cooling from 60 to 25 °C were fully characterized by differential scanning calorimetry (DSC) and optical microscopy (OM). 

Out of the 100 cosolvent systems, the authors constructed only 92 solubility curves represented as van't Hoff plots because eight cosolvent systems (i.e., THF and 

-butyl alcohol, THF and DMF, methanol and DMSO, methanol and 1,4-dioxane, methanol and ethanol, ethanol and 1,4-dioxane, DMSO and benzyl alcohol, and DMF and benzyl alcohol) had exhibited light pink–orange color, thus indicating the possible formation of two side-products of acetaminophen oxidation (i.e., 

Table II: Theoretical yield, apparent heat of solution, Gibb's free energy of dissolution, aspect ratio, enthalpy of melting, and crystallinity of acetaminophen Form I crystals grown from or dissolved in cosolvent systems.			

In general, a cosolvent system was a better solubilizer than a single solvent system (7). The theoretical yield (i.e., the amount of crystalline solids that would result if the saturated solution at 60 °C were cooled to 25 °C) was approximated from each solubility curve using the following equation: 

 is the solubility (g/mL) at 60 °C, and 

 is the solubility (g/mL) at 25 °C. The cosolvent system of benzyl alcohol and acetonitrile and the cosolvent system of ethanol and acetone gave a maximum theoretical yield of 71.73% and a minimum of 17.94%, respectively (see Table II). These results implied that the interactions between protic benzyl alcohol and polar aprotic acetonitrile were relatively weak and more sensitive to temperature change than the relatively strong interactions between hydrogen-bond donating ethanol and hydrogen-bond accepting acetone (7). 

The apparent differential heat of solution, Δ

, of acetaminophen in the 92 cosolvent systems was calculated from the 92 solubility curves based on the following van't Hoff equation if acetaminophen and the cosolvent system were assumed not to behave ideally (17):  

 is the mole fraction of acetaminophen solute in the cosolvent solution, 

 is the gas-enthalpy constant (8.314 J/mol K), Δ

 is the apparent differential heat of solution, which was obtained from the slope of the plot of 

 is a constant. Using the solubility data, the standard Gibbs free energy of the dissolution process, Δ

, at 25 °C was calculated using the following equation (18): 

Because the cosolvent systems were all real systems, the calculated Δ

 are listed in Table II. The positive values of Δ

 indicated that dissolution was an endothermic process. The energy of attraction of acetaminophen-solute molecules with each other and the energy of attraction of cosolvent molecules with each other were lower than the energy attraction of the acetaminophen solute and the cosolvent molecules in the solution. Therefore, the solubility of acetaminophen in different cosolvent systems increased with temperature. The negative values of Δ

 at 25 °C also revealed that dissolution was a spontaneous process at 25 °C. Only three systems seriously deviated from the linearity of Equation 2. They were ethyl acetate and acetone, THF and acetonitrile, and MEK and 1,4-dioxane. These solvent pairs seemed to have a large difference in their Hansen polar parameter values only (8). 

Optical micrographs of acetaminophen crystals grown from 88 cosolvent systems were taken because four other systems (i.e., THF and 1,4-dioxane, MEK and acetone, IPA and DMF, and IPA and 1,4-dioxane) had failed to produce acetaminophen solids. Perhaps the high affinity of acetaminophen solutes for those four cosolvent systems through hydrogen bonding more than compensated for their crystal interactions. As expected, increases in temperature generally increased the solubility of acetaminophen solutes in those four cosolvent systems to a lesser extent than in the other 88 cosolvent systems, probably because it reduced the cosolvent–cosolvent interactions in those four cosolvent systems less than it reduced the other cosolvent–cosolvent interactions in the other 88 cosolvent systems (19). 

Crystal habits of acetaminophen were classified into four typical shapes of prismatic, rhombohedral, hexagonal, and needle with sizes ranging from 50 μm in the cosolvent system of THF and acetone (see Figure 2a) to 4.5 mm in the cosolvent system of MEK and 1,4-dioxane (see Figure 2b) and aspect ratios, defined as length divided by breadth, ranging from 1.07 in the cosolvent system of nitrobenzene and N-butyl alcohol (see Figure 2c) to 7.33 in the cosolvent system of THF and acetonitrile (see Figure 2d) (20). The composition of solvents is known to influence nucleation and growth rates (21). Because THF, acetone, nitrobenzene, 1,4-dioxane, and MEK are hydrogen-bond accepting solvents, they might have preferentially adsorbed at specific faces, thus inhibiting the growth of acetaminophen crystals, or have similar structures to acetaminophen, thus hindering the regular deposition of oncoming molecular layers (22). 

Figure 2: Optical micrographs of crystal habit of acetaminophen Form I crystals grown by cooling in (a) tetrahydrofuran and acetone, (b) methyl ethyl ketone and 1,4-dioxane, (c) nitrobenzene + N-butyl alcohol, and (d) tetrahydrofuran and acetonitrile (scale bar = 500 Î¼m).			

All acetaminophen crystals harvested from the 88 cosolvent systems exhibited the same typical DSC melting endotherm of about 171 °C for Form I acetaminophen as shown in Figure 3. No other polymorph was observed, even though a change in the morphology to a needle shape could be a strong indication of the presence of acetaminophen Form II crystals (14). This important case serves as a reminder that a definitive identification of polymorphism must be based on spectroscopy, X-ray diffraction, or thermal analysis, and not on microscopy alone (7). 

Figure 3: A typical differential scanning calorimetry scan of acetaminophen Form I crystals grown in nitrobenzene and N,N-dimethylformamide cosolvent systems.			

The percent of crystallinity of acetaminophen crystals was quantified by dividing the area of a sample melting endotherm by the largest area of melting endotherm of all samples, which turned out to be the acetaminophen crystals grown in the cosolvent system of nitrobenzene and DMF with the enthalpy of melting, Δ

, of 195.24 J/g (see Table II). The authors suggest two possible reasons that this particular sample had the highest crystallinity. First, the slope of the solubility curve was not steep, so that cooling recrystallization took place gradually. Second, the molecular structures of nitrobenzene and DMF were similar to that of acetaminophen, which might have slowed down the nucleation and growth rate of acetaminophen crystals to yield high-quality crystals (22). 

Cosolvent screening is a useful and beneficial method for finding ideal cosolvent systems to create the desirable processing properties of acetaminophen crystals because cosolvent systems offered a wide range of choices in solubility, theoretical yield, crystal habit, aspect ratio, and crystallinity. Although small crystal sizes are advantageous for decreasing the force exerted on the roll-bearing blocks for roll compaction and for enhancing the dissolution rate because they lower ε

 respectively (see Table I), small crystal sizes can lengthen the drying rate, the filtration rate, the mixing rate, and the growth rate of granules in the nuclear growth region because of the decreases in 

 (which resulted from strong capillary force), 

, respectively (see Table I). Crystals with an aspect ratio higher than 3 have needle shapes that are easily broken into pieces of smaller crystal and fines. Therefore, systems produced relatively large crystal sizes, aspect ratios of ~1, and high theoretical yield such as that of the ethyl acetate and acetonitrile system. The authors will carry out similar investigations in the near future for the antisolvent systems represented by the green boxes in Table II and the scale-up effect on the crystal-size distribution.  

This work was supported by a grant from the national Science Council of Taiwan, ROC (NSC 97-2113-M-008-006). Suggestions about DSC from Jui-Mei Huang, at Precision Instrument Center in National Central University are gratefully acknowledged. 

 is a graduate student at the Department of Chemical and Materials Engineering, National Central University, 300 Jhong-Da Rd., Jhong-Li City 320, Taiwan, ROC, tel. +886 3 422 7151 ext. 34204, fax +886 3 425 2296, 

*To whom all correspondence should be addressed. 

Submitted: Aug. 13, 2009. Accepted: Nov. 20, 2009. 

1.  A.H. Goldberg, M. Gibaldi, and J.L. Kanig, 

2.  M.D. Rawlins, D.B. Henderson, and A.R. Hijab, 

3.  J.B. Sotiropoulus, T. Deutsch, and F.M.  Plakogiannis, 

13. N. Al-Zoubi, K. Kachrimanis, and S.         Malamataris, 

16. M. L. Ramos, J.F. Tyson, and D.J. Curran, 

Articles

The authors used common solvents to develop an initial solvent-screening method for laboratory-scale research to determine the solubility, polymorphism, and crystal properties of various active ingredients.

Engineering Processing Properties of Acetaminophen by Cosolvent Screening

				Part 1 of this article, which appeared in the March 2010 issue of 

, described a robust miniaturized screening strategy to accelerate the preparation and characterization of spherical agglomerates produced by spherical crystallization (1). This approach can be readily adopted or automated using only a small amount of active pharmaceutical ingredient (API), so spherical crystallization can be performed before the submission of a new drug application. The authors described their approach for constructuring form spaces for the nonionizable APIs, carbamazepine, cimetidine, and phenylbutazone by initial solvent screening of 19 common organic solvents to evaluate the feasibility of spherical crystallization. The materials and methods for this study were outlined in Part I of this article (1). Part II of this article discusses the results.    

 The differential-scanning-calorimetry (DSC) identification-test scan of the purchased carbamazepine exhibited the first endothermic peak at 170–174 °C, which corresponded to the melting of Form III (2). An exothermic peak at 175 °C followed, which corresponded to the crystallization of Form I (2). A second endothermic peak at 193 °C corresponded to the melting of Form I (2). The DSC identification-test thermogram of the purchased phenylbutazone displayed a single endothermic peak around 105 °C, which corresponded to the melting of Form A (3). The polymorphism of the purchased carbamazepine and phenylbutazone, therefore, were Form III and Form A, respectively, at room temperature. The identification test for cimetidine was performed using Fourier transform infrared (FTIR) spectroscopy instead of DSC because the melting points of the different forms of cimetidine were very close to each other (4). 

The FTIR spectrum showed four significant infrared peaks for Form A cimetidine at 1204, 1156, 1077, and 954 cm

 and two well-resolved characteristic peaks at 1243 and 1228 cm

, which indicated the polymorphism of the purchased cimetidine at room temperature was Form A (5–7). Two cimetidine molecules formed a dimer unit in Form A with an eight-member ring through the hydrogen bonds between N(10)H of one molecule and the N(12) of the other molecule  (1). The cimetidine molecule could absorb water at the N(1)H group (8). The imidazolium peaks were at 1204 and 954 cm

  was attributed to a CH deformation and a ring deformation (8, 10). The peak at 1156 cm

 was either a coupled ring and a CH-bending mode or the deprotonation of the N(1)H group on the complexation with a water molecule (8, 10). 

 To construct the form space for an API, which contains the possible solvent combinations for discovering a new polymorph, the 19 common solvents for scale-up listed in the materials and methods section of Part I of this article were arranged from top to bottom vertically and from left to right horizontally in ascending order of the solvent's Hildebrand value (1). This arrangement resulted in a 19 × 19 matrix of 361 boxes (see Tables 1(a), 1(b), and 1(c)).  Each box represented a specific binary solvent combination except for the 19 boxes located diagonally in the matrix stepping down from the upper left-hand corner to the lower right-hand corner. The 19 diagonal boxes represented the 19 pure-solvent systems. The yellow boxes were the 

 (i.e, defined as a solvent that gave solubility of ≥ 5 mg/mL at 25 °C) and the red boxes the

 (i.e, defined as a solvent that gave solubility of ≤ 5 mg/mL at 25 °C) with respect to a particular API. The yellow and red boxes were the materials properties of the API.  The gray boxes also were determined experimentally. They represented the immiscible solvent pairs that were independent of the API. Once the locations of the red, yellow,  and gray boxes were fixed, the location of the green, blue, and white boxes representing the antisolvent system (i.e., a good solvent and a bad solvent), the good cosolvent system (i.e,  a good solvent and a good solvent) and the bad cosolvent system (i.e,. a bad solvent and a bad solvent), respectively, were deduced graphically from the diagonal boxes. Since the 19 × 19 matrix was symmetrical along the diagonal boxes, the actual possible solvent combinations were reduced from 361 boxes to a total of 190 boxes or less (i.e., derived from 19 + (361 – 19) ÷ 2 = 190) because it is impractical to perform experiments in the white-boxed areas.    

The actual form space of carbamazepine had 13 yellow, 6 red, 69 green, 78 blue, 14 gray, and 10 white boxes (see Table  I (a)). The actual form space of cimetidine contained 5 yellow, 14 red, 66 green, 10 blue, 14 gray and 81 white boxes (see Table I (b)). The actual form space  of phenylbutazone was composed of 17 yellow, 2 red, 22 green, 135 blue, 14 gray and 0 white boxes (see Table I (c)). The actual form spaces would have been expanded significantly if various solvent compositions of binary mixtures, temperatures, and ternary solvent systems were considered (10–12).

The solubility values of carbamazepine, cimetidine, and phenylbutazone in their corresponding solvents at 25 °C were listed in Table II. Most of the good solvents of Form III carbamazepine belonged to Class 1 (i.e., a protic hydrogen-bond-donating solvents [Lewis acids]), Class 2 hydrogen-bond-acceptor solvents [Lewis bases]), and Class 3 polar aprotic solvents, which were capable of disrupting the N–H•••O hydrogen bonds and the C–H•••O weak interactions of the carboxamide dimer units existing in the solid state of all forms of carbamazepine (2, 13). The good solvents for Form A cimetidine were mainly Class 1 solvents, which had the capability of interacting readily with the weakly basic imidazole ring  and disrupting the N–H•••N hydrogen bonds of the dimer units existing in the solid state of Form A cimetidine (8, 13, 14). The good solvents for Form A phenylbutazone were Classes 1, 2, and 3 solvents, Class 4 chlorocarbon solvents, and Class 5 hydrocarbon solvents. This result implied the intermolecular interactions in the crystal lattice of phenylbutazone were dominated by van der Waals forces (i.e., dipole-dipole interactions and London dispersion force) (15). 

 The spherical agglomeration (SA) method was used in spherical crystallization. Because all miscible pairs of a good solvent and a bad solvent (i.e., an antisolvent) were represented by the green boxes and all immiscible pairs of a bad solvent and a bridging liquid were designated by the gray boxes in the actual form space of each API, the total number of combinations of a good solvent  and a bad solvent and a bridging liquid was equal to (the sum of the number of green boxes in each column with a red box × the number of gray boxes in the same column) and (the number of green boxes in the bottom row with a red box × the number of gray boxes in the same bottom row), if water were also a bad solvent symbolized by a red box situated at the bottom right-hand corner of the actual form space. 

The number of combinations for the SA method based on Table 1(a) for carbamazepine is 183 (i.e., derivied from 10 × 4 + 12 × 2 + 13 × 1 + 13 × 1 + 13 × 1 + 8 × 10 = 183). The number of combinations for the SA method based on Table 1(b) for cimetidine is 100 (i.e., derived from 3 × 4 + 4 × 2 + 5 × 1 + 5 × 1 + 5 × 1 + 5 × 1 + 5 × 1 + 5 × 1 + 5 × 1 + 5 × 1 + 4 × 10 = 100). The number of combinations for the SA method based on Table 1(c) for phenylbutazone was 136 (i.e., derived from 14 × 4 + 8 × 10 = 136).    

Using the SA method, five types of behavior were observed:  					

Type 1: No solid production upon the addition of 10 mL of an antisolvent to 2 mL of an API solution at 25 °C 

Type 2: Dissolution of the solids upon the addition of the bridging liquid at 25 °C 

Type 3: No agglomeration upon the addition of the bridging liquid at 25 °C 

Type 4: Nonspherical agglomeration upon the addition of the bridging liquid at 25 °C 

Type 5: Spherical agglomeration upon the addition of the bridging liquid at 25 °C.    

Type-3 behavior was not observed with phenylbutazone. 

In Type-1 and Type-2 behaviors, the solubility and the width of a metastable zone of an API solid were profoundly related to the molecular interactions among API solutes and solvents in a real solution system, which were dependent on the molecular structure of an API, the physical nature of a solvent, temperature, and compositional ratios of the solvent combinations (10, 12, 16, 17). In Type-3, Type-4, and Type-5 behaviors, the agglomeration process was intricately related to the balance between the drag force and the liquid-bridge strength (18, 19). The drag force was dependent on the speed of agitation, the viscous force, and the growing diameter of an agglomerate (18, 19). The liquid-bridge strength relied on the low pressure in the bridging-liquid phase led by the curvature of the bridging-liquid interface. The curvature of the bridging liquid was a function of:  the interfacial tension of the bridging liquid and the mother liquor; the surface tension between the API crystals and the bridging liquid; the solubility of the API at the interface between the bridging liquid and the mother liquor; and the geometry of the API crystals (18–21).   

Only 10 solvent combinations from a total of 183 combinations produced decent spherical agglomerates of carbamazepine. Only 4 from a total of 100 solvent combinations produced decent spherical agglomerates of cimetidine and 4 of 136 solvent combinations produced decent spherical agglomerates of phenylbutazone. Water was present in 18 workable combinations as an antisolvent (i.e., derived from 10 + 4 + 4 = 18). This observation agreed with the many reported experimental conditions that water could be used as an antisolvent for nonionizable organic compounds (22–43). If only the solvent combinations with water as an antisolvent were considered,  the total number of solvent combinations would have been reduced from 183 to 80 for carbamazepine, from 100 to 40 for cimetidine and from 136 to 80 for phenylbutazone. The criterion of using water as an antisolvent, therefore, was essential to accelerate the screening activities. The aqueous-solution environment induced the formation of crystal surfaces with hydrophilic functionalities, which helped to sustain the curvature of the hydrophobic solvents (44). This result may explain why most of the bridging liquids in the 18 workable solvent combinations were Class-4 chlorocarbon or Class-5 hydrocarbon solvents.  

Figure 1 (ALL FIGURES ARE COURTESY OF THE AUTHORS)

 The 10 solvent combinations of good solvent, antisolvent, and bridging liquid for carbamazepine were: 1,4-dioxane, water, and chloroform; acetonitrile, water, and paraxylene; acetonitrile, water, and chloroform; acetonitrile, water, and 

-dimethylaniline; ethanol, water, and toluene; ethanol, water, and chloroform; ethanol, water, and 

-dimethylaniline; methanol, water, and xylene; methanol, water and toluene; and methanol, water, and chloroform. All carbamazepine spherical agglomerates (see Figure 1) were freshly made as the dihydrate, which was the thermodynamically stable form in the aqueous system (45, 46). They were first oven-dried at 100 °C for 4 h to become Form I before being subjected to characterization by a scanning electron microscope (SEM), whose high-vacuum system could be contaminated by the dehydration of the dihydrate form. Although the solubility of Form I carbamazepine was less than the solubility of Form III, it was advantageous to make Form I carbamazepine spherical agglomerates for several reasons (47). Form I transformed to dihydrate at a slower rate than Form III did in an aqueous environment, and the solubility of the dihydrate was known to be about 1.3 times less than the solubility of any of the anhydrous form (48). Also, the difference in the solubility values between Forms I and III was relatively small, and the area under the curve in bioavailability tests of the forms were in the order of Form I > Form III < dihydrate. Lastly, Form I was stable below 71 °C (47).

The four solvent combinations of good solvent, antisolvent, and bridging liquid for cimetidine were: methanol, water, and 

-heptane; methanol, water, and xylene; methanol, water, and paraxylene; and methanol, water, and toluene. All cimetidine spherical agglomerates (see Figure 2) were made of Form C crystals as verified by the FTIR spectrum with Form C characteristic bands at 1066, 1182, 1229, and 1347 cm

  (5). Even though Form A is thermodynamically less stable than Form C under normal pressure, Form A was preferable for medicine because of Form C's extraordinarily poor bulk flowability for tableting (4, 6). The feasibility of producing Form C cimetidine spherical agglomerates, however, might have made the industrial tablet production of Form C cimetidine possible. 

The four different solvent combinations of good solvent, antisolvent, and bridging liquid for phenylbutazone were: tetrahydrofuran, water, 				

-heptane. Spherical agglomerates (see Figure 2) of the first two cases were made of Form E crystals (see Figure 3), whereas the last two cases were made of Form A crystals as indicated by DSC. Form A showed only a single melting endotherm at 105 °C for Form A, but the DSC thermogram (see Figure 4) exhibited the first endothermic peak at 92.5 °C , which corresponded to the melting of Form E (3). This peak was followed by an exothermic peak at 95 °C, which  corresponded to the crystallization of Form E (3). A second endothermic peak at 104 °C corresponded to the melting of Form A (3). Because the equilibrium solubility and the dissolution rate of the phenylbutazone crystal forms were in the order of Form C > Form E> Form D >  Form B > Form A and Form C > Form E > Form B > Form D > Form A, respectively, spherical agglomeration might offer a way of tuning the dissolution profile of phenylbutazone by the change of polymorphism in addition to providing  good bulk flowability. 

 The polymorphism of the API solids remained unchanged before and after the addition of a bridging liquid in the previoulsy described 18 solvent systems. Other powder characteristics of the spherical agglomerates from the 18 solvent systems such as the percent yield, the length-mean diameter, the apparent density, the population density, the sphericity, and the friability index were fully analyzed based on the sample weight and digital-camera photographs and were calculated according to Equations 1–6 as follows (22, 26, 40): 

 is length-mean agglomerate diameter (cm); 

 is length-mean agglomerate diameter (cm) after 30 min, and (

 is length-mean agglomerate diameter (cm) at the beginning of the friability test. The volume of a bridging liquid and the angle of repose were entirely determined by experiments. The sample sizes varied greatly from numerous powder-like agglomerates to one large aggregate as revealed clearly in Figures 1, 2, and 3. No replicate was made for each test at this point due to the limitation of time and materials. All experimental and calculated results are summarized in Table III.  

Spherical agglomerates of carbamazepine dihydrate from an acetonitrile–water–paraxylene system, Form C cimetidine from a methanol–water–xylene system, Form E phenylbutazone from a tetrahydrofuran–water–

-heptane system, and Form A phenylbutazone from an acetonitrile–water–

-heptane system gave relatively high yields, high friability indices (i.e., greater strength) and low angles of repose (i.e., good bulk flowability) among the other spherical agglomerates produced by other systems. Therefore, they were first chosen as candidates for further scale-up or preformulation studies. Rotation in a cylindrical vessel alone did not encourage the formation of spherical agglomeration because a control experiment confirmed that spherical-shaped agglomerates could also be produced in a 500-mL sized crystallizer equipped with baffles and an overhead stirrer. The yield of each system was dependent on the compositional ratios of the solvent combinations. The yield can be easily raised during future process development by increasing the volume ratio of antisolvent to good solvent to values higher than 5:1. 

 SEM images of the cross-section of all spherical agglomerates of carbamazepine dihydrate, Form C cimetidine, Form E phenylbutazone, and Form A phenylbutazone were categorized into four different typical modes. All primary crystals were needles and were: tangled with fibrils (see Figure 5(a)); randomly stacked together and linked by solid nodes (see Figure 5(b)); bundled together in parallel (see Figure 5(c)) and radiated from a center core; and jointed by solid bridges to form many 

-shaped ladder structures (see Figure 5(d)) (49). 

Because the volume ratio of water (i.e., the antisolvent) to good solvent was fixed at 5:1, the presence of a small amount of bridging liquid immiscible with water also would be immiscible with the saturated aqueous API solution and the good solvent as confirmed experimentally. Under a high speed of stirring, the bridging liquid was dispersed in the saturated aqueous API solution and the good solvent as tiny suspended droplets. These tiny droplets were randomly drawn to near the points of contact between the preformed needle-like primary crystals by interfacial tension. The fibrils, nodes and joints (see Figures 5 (a)–5 (d)) suggested that the dissolution of API would take place initially at the crystal–bridging liquid interface because of the relatively higher solubility power of the bridging liquid than the surrounding saturated high water content of the mother liquor. As the saturated high water content of the mother liquor diffused into the bridging-liquid region, recrystallization of the API occurred inside the liquid bridges. 

The form space constructed from the common organic solvents by initial solvent screening provided a systematic way of finding ideal combinations of a good solvent, an antisolvent, and a bridging liquid for spherical agglomeration on  a miniaturized scale. The same working logic could be extended to other solvent techniques such as the spreading of cube-shaped particles, monodispersed double emulsions,  microemulsions,  and organic nanocrystal fabrication (50–53). Powder characteristics of spherical agglomerates such as percent yield, length-mean diameter, apparent density, population density, sphericity, friability index, and angle of repose served as a selection guide for narrowing down the final feasible solvent combinations for future scale-up. The initial solvent screening method made it possible to identify the right solvent combinations for spherical crystallization early in drug development. This strategy, therefore, increases the  possibility of direct tableting without performing granulation. 

This work was supported by a grant from the National Science Council of Taiwan, R. O. C. (NSC 98-2113-M-008-008-006). Suggestions from Jui-Mei Huang in DSC and Ching-Tien Lin in SEM, both at the National Central University's  Precision Instrument Center and High Valued Center are gratefully acknowledged. 

 were graduate students in the department of Chemical and Materials Engineering, and 

 was a graduate student at the Institute of Materials Science and Engineering, all at National Central University, 300 Jhong-Da Rd., Jhong-Li City 320, Taiwan, ROC, tel. +886 3 422 7151 ext. 34204, fax +886 3 425 2296, 

To whom all correspondence should be addressed.  

Submitted: Feb. 2, 2009; Accepted May 20, 2009.  

 3. M.D. Tuladhar, J.E. Carless, and M.P. Summers, 

In Part I of this article, which appeard in the March 2010 issue, the authors describe their approach for constructing form spaces for carbamazepine, cimetidine, and phenylbutazone by initial solvent screening to evaluate the feasibility of spherical crystallization. Part II of this article discusses their findings.

Spherical Crystallization for Lean Solid-Dosage Manufacturing (Part II)

Pharmaceutical manufacturing typically lags behind other industries in process efficiencies because production processes for new drug products are quickly developed in a laboratory for producing clinical-trial materials to enable the rapid submission of a new drug application (NDA) (1). The NDA locks in the documented manufacturing steps with a regulatory commitment and leaves no chance for the full optimization of processing steps such as crystallization, filtration, drying, dry blending, granulation, drying, and tableting (2). Granulation is the slowest of these processing steps, but by applying cross-performance relationships and mixing rules, the efficiency of granulation can be enhanced significantly (2–4). To achieve lean solid-dose manufacturing, the granulation step can be replaced with spherical crystallization in a common stirred tank at the crystallization step (5–26). This change may reduce the time-consuming and labor-intensive steps in solid-dose manufacturing for a high-dose formulation to only a few unit operations of spherical crystallization, filtration, drying, dry blending, and tableting.  

To the best of our knowledge, no literature has laid out a systematic way to identify the best solvent combinations for spherical crystallization. This article presents a robust miniaturized screening strategy to accelerate the preparation and characterization of spherical agglomerates produced by spherical crystallization. This strategy can be readily adopted or automated using only a small amount of active pharmaceutical ingredient (API), thus enabling spherical crystallization to be performed before the NDA submission. 

Spherical crystallization is a well-studied, solvent-based method that directly produces spherically agglomerated crystals to afford free flowability, uniform packability, and good compressibility for mixing, filling, direct tableting, and even burning characteristics (17–19). Spherical crystallization is categorized into two methods: emulsion solvent diffusion (ESD) and spherical agglomeration (SA) (13, 22, 24, 26). Both methods use a 

 (i.e., a solvent that gave solubility of ≥ 5 mg/mL at 25 °C ) to dissolve the API and a 

 (i.e., a solvent that gave solubility of < 5 mg/mL at 25 °C) miscible with the good solvent (i.e., the antisolvent) to generate the required supersaturation. For the ESD method, the solution containing the dissolved API is dispersed into the bad solvent to produce emulsion droplets. Crystals are generated inside the suspended droplets by the counterdiffusion of the good and bad solvents at the droplet interface (26). However, for the SA method, the formation of agglomerates is promoted by the addition of a small amount of the third solvent, namely the bridging liquid, which is immiscible with the bad solvent and drawn preferentially in the interstitial space between the newly formed API crystals (26). 

Spherical crystallization is a very versatile method. The operating conditions involve either: the recrystallization of a near-saturated solution by cooling (10, 12, 13, 15, 16, 19, 21, 22, 25, 26), by seeding (14), or by an antisolvent (7, 8, 10–13, 15, 16, 19–26) with (7, 8, 10–13, 15, 20–26) or without (19, 22) the presence of a bridging liquid or with the addition of an emulsifying agent (13, 15, 18) or a cross-linking agent (16, 23); or secondly, the dispersion of solids with the addition of a bridging solution (5, 6, 11, 17). The addition order of the antisolvent and the bridging liquid to the near-saturated solution can vary (11, 26). 

Spherical crystallization relies on the form space (or the solvent-miscibility plot) derived from an initial solvent-screening method performed in 20-mL scintillation vials (27–31). Because the composition ratio of an antisolvent to a good solvent at the time of nucleation determines the polymorphism of the API crystals, the solvents were added by a specific SA pathway (11, 12, 30). The antisolvent was first added to the near-saturated API solution to ensure the occurrence of recrystallization in a binary solvent system. The bridging liquid was introduced at the end point of recrystallization to induce the agglomeration of the existing primary crystals. The polymorphism of the primary crystals before and after spherical agglomeration was monitored by either differential scanning calorimetry (DSC) or transmission Fourier transform infrared (FTIR) spectroscopy. The volume of a bridging liquid used, the percent yield, the length-mean diameter, the apparent density, the population density, the sphericity, the angle of repose, and the friability index of spherical agglomerates were carefully measured and calculated (5, 9, 13, 23).  The internal microstructures of spherical agglomerates were examined by scanning electron microscopy (SEM). 

The feasibility of spherical crystallization by initial solvent screening was tested with three model APIs (32–42) (see Figure 1). These APIs were: carbamazepine (5H-dibenze (b, f) azepine-5-carboxamide), an anticonvulsant used to treat epilepsy and trigeminal neuralgia; cimetidine (N"-cyano-

'-[2-[[5-methyl-1H-imidazol-4-yl]methyl]thio]ethyl)guanidine), a specific competitive histamine H2-receptor antagonist used to  treat human peptic ulcers; and phenylbutazone (1,2-diphenyl-4-

-butyl-3,5-pyrazolidinedione), a nonsteroidal antiinflammatory drug with antipyretic and analgesic activity. These APIs were chosen because of  several key properties: their nonionizable nature between pH of 6 and 7; their worldwide commercial value; the abundance of characterization information in the literature; the lack of extensive solvent studies on their solubility, polymorphism, crystallinity, and crystal habits; and the variety of their naturally occurring polymorphs, hydrates,  and/or solvates. 

Figure 1: Molecular structures of (a) carbamazepine,(b) cimetidine, and (c) phenylbutazone.   (FIGURE I IS COURTESY OF THE AUTHORS.)

O), white, crystalline, and prismatic powders, molecular weight (mw) of 236.67 g/mol, melting point (mp) of 191–192 °C, reagent grade, Lot 036k1219, and cimetidine (C

S), white crystalline rod-like powders, mw of 252.34 g/mol, mp of 141–143 °C, reagent grade, Lot 088H1317 were purchased from Sigma-Aldrich (St. Louis, MO) and used as received. Phenylbutazone (C

), white, crystalline, and needle-shaped powders, mw of 308.38 g/mol, mp of 106–108 °C, purity of 99+ %, Lot A0230775 was purchased from Acros Organics (Morris Plains, NJ) and used as received. Identification tests for carbamazepine and phenylbutazone were carried out by DSC. The use test for cimetidine was performed by FTIR instead of DSC because DSC was unable to distinguish the polymorphs of cimetidine clearly (39). 

The following sovlents were obtained from Tedia (Fairfield, OH): acetone (CH

), high-performance liquid chromatographical/spectrometrical (HPLC/spectro) grade with purity of 99.5%, boiling point (bp) of 56 °C, mw of 58.08 g/mole, Lot 411050; 

OH),  American Chemical Society (ACS) grade with purity of 99.4%, bp of 117.7 °C, mw of 74.12 g/mole, Lot 205027; 

), ACS grade with purity of 99.8%, bp of 153 °C, mw of 73.10 g/mole, Lot 020505; 

), HPLC/spectro grade with purity of 99.4%, bp of 98 °C, mw of 100.21 g/mole, Lot HS-1712 ; isopropyl alcohol ((CH

CHOH), HPLC/spectro grade with purity of 99.8%, bp of 82.4 °C, mw of 60.1 g/mole, Lot 503027; methanol (CH

OH), HPLC/spectro grade with purity of 99.9%, bp of 64.7 °C, mw of 32.04 g/mole, Lot 411070; methyl ethyl ketone (C

), ACS grade with purity of 99.6%, bp of 81.6 °C, mw of 72.11 g/mole, Lot 201021; methyl 

), certified grade with purity of 99.9%, bp of 55.2 °C, mw of 88.15 g/mole, Lot 712032; tetrahydrofuran (C

O), HPLC/spectro grade with purity of 99%, bp of 65–67 °C, mw of 72.11 g/mole, Lot 411013; and xylene (C

), ACS grade with purity of 98.5%, bp of 137–144 °C, mw of 106.17 g/mole, Lot 305065. 

The following solvents were obtained from Acros Organics: 

), ACS grade with purity of  99%, bp of 193 °C, mw of 121.18 g/mole, Lot A0213203001 and 

), ACS grade with purity of 99%, bp of 138 °C , mw of 106.17 g/mole, Lot 48754/2. 1,4-Dioxane (C

), ACS grade with purity of 98%, bp of 100–102 °C, mw of 88.11 g/mole, Lot sp-3432R was obtained from Showa Chemical (Tokyo). Chloroform (CHCl

), HPLC/spectro grade with purity of 99.9%, bp of 60.5–61.5 °C, mw of 119.38 g/mole, Lot E554180 and ethanol (CH

OH), HPLC/spectro grade with purity of 99.5%, bp of 78 °C, mw of 46.7 g/mole were obtained from Echo Chemical (Taipei, Taiwan). Ethyl acetate (CH

), ACS grade with purity of 99.5%, bp of 76.5–77.5 °C, mw of 88.11 g/mole, Lot G43342 was obtained from Grand Chemical (Daejeon, South Korea). Acetonitrile (CH

CN), ACS grade with purity of 99.96%, bp of 81.6 °C, mw of 41.05 g/mole, Lot 0043X29B30, and toluene (C

), HPLC/spectro grade with purity of 99.8%, bp of 110.6 °C, mw of 92.14 g/mole, Lot B46755 was purchased from  Mallinckrodt Baker (Phillipsburg, NJ). Reversible-osmosis (RO) water was clarified by a water purification system (Milli-RO Plus, Millipore, Billerica, MA).  

 Under the initial solvent screening, the 19 solvents specified in the solvents section were chosen. The working procedures followed closely the details described in References 27–30. As previously defined, a 

was a solvent that gave solubility of ≥ 5 mg/mL at 25 °C. A 

 was a solvent that gave solubility of < 5 mg/mL at 25 °C). The form space was constructed in  a framework of 19 X 19 solvent matrix (27–30).  

 The working logics followed the procedures laid out in References 27–30.  

 Good solvents, antisolvents, and bridging liquids were chosen according to the form space of each API. For carbamazepine and phenylbutazone, approximately 30 mg of API were dissolved in 2 mL of a good solvent in a 20-mL scintillation vial. For cimetidine, 2 mL of saturated solution were prepared using a good solvent. Then, 10 mL of an antisolvent were added to all API solutions under a standardized agitation of 200 rpm by an orbital shaker (Model OS-701, TDK , Taipei). The 5:1 volume ratio of the antisolvent to the  good solvent was based on a specified cystallization method (11). Crystallization only took place in some of the solution systems. 

Two hours after the first appearance of solids, the solids were filtered, oven-dried under a mild condition at 40 °C overnight, and characterized for their purity, polymorphism, solvates, and crystallinity by DSC, thermal gravimetric analysis (TGA), and FTIR. 

Next, 0.02-mL portions of a bridging liquid were added in the remaining slurry every once in a while by a micropipette until all suspended crystals were transformed into agglomerates by sight. The amount of the bridging liquid used should be kept at minimum without causing any dissolution of the primary crystals or exceeding 80% volume (i.e., 16 mL in total solution volume) of the scintillation vial. All trials were performed at room temperature, and all agglomerates were filtered, oven-dried under mild conditions of 40 °C overnight, and characterized. For carbamazepine spherical agglomerates, oven-drying at 100 °C for 4 h was performed before any further characterization. For the solution systems that gave rise to spherical agglomerates, experiments were repeated again without the addition of any bridging liquid to produce primary crystals that were filtered, oven-dried at 40 °C overnight, and characterized for their polymorphism by either DSC or FTIR. 

 Thermal analytical data of 3 to 5 mg of solids placed in perforated-aluminum-sealed 60-µL pans were collected on a calorimeter (Perkin Elmer DSC-7, Perkin Elmer, Waltham, MA), with the following temperature scanning rates: 10 °C/min from 50 to 200 °C for carbamazepine; 5 °C /min from 50 to 170 °C for cimetidine (mainly for the crystallinity measurements); and 8 °C/min from 50 to 130 °C for phenylbutazone using nitrogen with purity of 99.990 % as a blanket gas (32, 33, 35, 39, 41). The temperature axis was calibrated with indium with purity of 99.999% (Perkin Elmer) with a melting onset at 156.6 °C. 

 Sample-weight loss as a function of temperature was monitored by a thermogravimetric analyzer (TGA 7, Perkin Elmer). The heating rate was 10 °C/min from 50 to 200 °C, 5 °C/min from 50 to 170 °C and 8 °C/min from 50 to 130 °C for carbamazepine, cimetidine, and phenylbutazone, respectively (32, 33, 39, 41). Weight loss was usually associated with either solvent evaporation close to the boiling point of a solvent, as in the case of solvates, or sample decomposition. The open platinum pan and stirrup were rinsed by ethanol and burned by a spirit lamp to remove all impurities. All samples were heated under nitrogen atmosphere to avoid oxidization. About 3 mg of sample were placed on the open platinum pan suspending in a heating furnace. 

 Transmission FTIR spectra were recorded on a spectrometer (Perkin Elmer Spectrum One, Perkin Elmer)  Approximately 1 mg of sample was ground gently with 99 mg of 50 °C oven-dried potassium bromide (KBr) in an agate mortar and pestle to avoid polymorphic transition possibly induced by extended grinding. The round KBr sample disk was prepared by a uniaxial press with a pressure of 7 tons. The disk was scanned with a scan number of 8 from 450 to 4000 cm

 A digital still camera (DSC-T7 Cyber-shot, Sony, Tokyo) was used to take photographs of millimeter-sized spherical agglomerates before and after the friability test. The number of spherical agglomerates and their individual diameters were measured based on the image analysis. 

 Because of the small amount of spherical agglomerates produced, the angle of repose was measured with the oven-dried spherical agglomerates inside the 20-mL scintillation vial. The vial was mounted on a lever that had one end pivoted to the benchtop and was originally laid flat on the benchtop surface.  The vial was tilted by raising the other end of the lever gradually. The angle made between the lever and the benchtop surface at which the spherical agglomerates inside the vial began to flow was the angle of repose. Each angle measurement was repeated at least 10 times with a protractor. 

 Spherical agglomerates harvested from each trial were placed in a 20-mL scintillation vial that was subjected to a rotating speed of 50 rpm on a ball-milling machine (MUBM-236-RTD, Shin Kwan Machinery, Taipei) for 30 min. Photographs of spherical agglomerates before and after this friability test were taken.  Image analysis was performed to calculate the friability index. 

 An SEM (Hitachi S-3500N, Hitachi Ltd., Tokyo) was used to observe the internal morphology of spherical agglomerates that were cross-sectioned by a razor blade. Because the cutting was done by hand, it might have crushed the agglomerates. To avoid imaging the artifacts, the cross-sectioned area was blown with compressed air to remove surface debris. Both secondary electron imaging and backscattered electron imaging were used for the SEM detector, and the magnification was 15- to 300,000-fold.  The operating pressure was 10

  Pa vacuum, and the voltage was 15.0 kV. All samples were mounted on a carbon conductive tape (Prod. No. 16073, TED Pella, Redding, CA) and sputter-coated with gold (Hitachi E-1010 Ion Spotter) with a thickness of about 6 nm. The discharge current used was about 0 to 30 mA, and the vacuum was around 10 Pa. 

 were graduate students in the Department of Chemical and Materials Engineering, and 

Submitted: Feb. 2, 2009; Accepted May 20, 2009. 

 7. Y. Kawashima, M. Okumura, and H. Takenaka, 

 9. Y. Kawashima, M. Okumura, and H. Takenaka, 

 18. U. Teipel, T. Heintz, and H.H. Krause, 

In Part I of this article, the authors describe the materials and methods used in developing a screening strategy to accelerate the preparation and characterization of spherical agglomerates by spherical crystallization.

Spherical Crystallization for Lean Solid-Dose Manufacturing (Part 1)

In Part 1 of this article, which appeared in the May 2009 issue of 

 the authors described characterization methods for the functional properties and the process-dependent polymorphic flow charts of the three soft pharmaceutical materials: carbamazepine, cimetidine, and phenylbutazone. This part presents several indispensable processing properties (i.e., solubility, polymorphism, crystallinity, and crystal habit) of those three active pharmaceutical ingredients. The underlying principles can be easily extended to other nonconventional sytems such as metal–organic framework materials, infinite coordination polymer particles, and cocrystals. 

 The differential scanning calorimetry (DSC) use-test scan of carbamazepine in Figure 1 exhibited the first endothermic peak at 170–174 °C, which corresponded to the melting of Form III, followed immediately by an exothermic peak at 175 °C, which corresponded to the crystallization of Form I, and a second endothermic peak at 193 °C, which corresponded to the melting of Form I (1). The DSC use-test thermogram of phenylbutazone in Figure 2 displayed a single endothermic peak around 105 °C, which corresponded to the melting of Form A (2). Thus, the polymorphisms of carbamazepine and phenylbutazone were Form III and Form A, respectively, at room temperature. The use test for cimetidine was performed by Fourier transform infrared spectroscopy (FTIR) because the melting points of the different forms of cimetidine are close to each other (3). 

Figure 1: Differential scanning calorimetry scan of the carbamazepine. (ALL FIGURES ARE COURTESY OF THE AUTHORS) 

The FTIR spectrum in Figure 3 showed four significant infrared peaks for Form A cimetidine at 1204, 1156, 1077, and 954 cm

  and two well-resolved characteristic peaks at 1243 and 1228 cm

  (4–5). Therefore, the polymorphism of cimetidine at room temperature was Form A. The cimetidine molecule has three hydrogen-donor centers (i.e., N1H, N10H, and N13H) and three hydrogen acceptor centers (i.e., N3, N12, and N17) (see Figure 4) (6). Two cimetidine molecules formed a dimer unit in Form A with an eight-member ring through the hydrogen bonds between N10H of one molecule and N12 of the other. The cimetidine molecule absorbed water at the N1H group (6). Cimetidine was weakly basic, and its reported pKa values were 6.80 and 6.93. The imidazolium peaks were 1204 and 954 cm

  was attributed to a CH deformation and a ring deformation (6, 7). The peak at 1156 cm

  could either be a coupled ring and a CH bending mode or the deprotonation of the N1H group on the complexation with a water molecule (6, 7). Two imidazole tautomers coexisted in the cimetidine structure: one with a hydrogen atom bound to the N1 atom and the other with the hydrogen atom bound to the N3 atom (see Figure 4). However, only the N1-H form was present in the crystal states of cimetidine and its monohydrates. In the solution, cimetidine also assumed two staggered (

) configurations of the cyanoguanidine group (8). 

Figure 2: Differential scanning calorimetry scan of the phenylbutazone. 

 To construct the form space for an active pharmaceutical ingredient (API), which contains the possible solvent combinations for discovering a new polymorph, 19 common solvents for scale-up were first arranged vertically and horizontally in ascending order of the solvent's Hildebrand value. This arrangement resulted in a 19 x 19 matrix of 361 boxes (see Tables I–III). Each box represented a specific binary-solvent combination, except for the 19 boxes descending diagonally from the upper left corner to the lower right corner. These 19 boxes represented the 19 pure solvent systems. The yellow boxes were the good solvents, and the red boxes the bad solvents, with respect to a particular API. The yellow and red boxes were the material properties of an API. Besides the red and yellow boxes, the gray boxes were also determined experimentally. They represented the immiscible solvent pairs that were independent from the nature of an API. Once the locations of the red, yellow, and gray boxes were fixed, the location of the green, blue, and white boxes representing the antisolvent system (i.e., a good solvent and a bad solvent), the good cosolvent system (i.e., a good solvent and a good solvent), and the bad cosolvent system (i.e., a bad solvent and a bad solvent), respectively, could be deduced graphically. Since the 19 x 19 matrix was symmetrical along the diagonal boxes, the actual possible solvent combinations were reduced from 361 boxes to a total of 19 + (361 – 19) ÷ = 190 boxes or fewer because it is normally impractical to perform experiments in the white-boxed areas. 

Figure 3: Fourier transform infrared spectrum of the cimetidine. 

The form space of carbamazepine had 13 yellow, 6 red, 69 green, 78 blue, 14 gray, and 10 white boxes (see Table I); the form space of cimetidine had 5 yellow, 14 red, 66 green, 10 blue, 14 gray, and 81 white boxes (see Table II); and the form space of phenylbutazone had 17 yellow, 2 red, 22 green, 135 blue, 14 gray, and 0 white boxes (see Table III). The actual form spaces would have been significantly bigger if various solvent compositions of binary mixtures,temperatures, and ternary solvent systems had been considered (9–11). 

 When the locations of the 19 solvents were assigned within the 3-D Hansen plot of a dispersion-force component (δ

) and all the good solvents and bad solvents were represented by yellow and red, respectively, based on the colors of the diagonal boxes in the form space of a given API, a cluster of yellow domains was formed. The contour of these yellow domains outlined a 3-D volume of solubility in space that excluded all the bad solvents by a solubility sphere with center coordinates (δ

) (12). For the solubility sphere of Form III carbamazepine at 25 °C, δ

 . For the solubility sphere of Form A cimetidine at 25°C, δ

 . Finally, for the solubility sphere of Form A phenylbutazone at 25°C, δ

) were close to the reported values of 17.5, 12.5, and 10.7 MPa

 , respectively, measured by dissolution calorimetry (13). 

Table I: Solvent systems of carbamazepine. 

With the solubility sphere of an API at hand, the solubility power of any new solvent other than the 19 solvents used with known values of δ

 can be predicted instantly. If the coordinates of a new solvent (δ

) are located inside the solubility sphere of an API, it is a good solvent with respect to that API. Otherwise, it is deemed a bad solvent. To verify the validity of the solubility sphere, each sphere of an API was tested with 

-propanol to be located inside the Form III carbamazepine's solubility sphere (because 

  for Form III cambamazepine), inside the Form A cimetidine's solubility sphere (because 

  for Form A cimetidine), and inside the Form A phenylbutazone's solubility sphere (because 

  for Form A phenylbutazone). These calculations agreed well with the experimental observations that 

The authors describe the importance of a rapid and an abbreviated screening strategy in initial solvent screening. This article contains bonus online-exclusive material.