Ind. Eng. Chem. Res. 2005, 44, 7337-7344
7337
MATERIALS AND INTERFACES Micronization of Phenanthrene Using the Gas Antisolvent Process. 1. Experimental Study and Use of FTIR Yousef Bakhbakhi, Sohrab Rohani,* and Paul A. Charpentier* Department of Chemical and Biochemical Engineering, University of Western Ontario, London, Ontario, Canada N6A 5B9
In this study the micronization of phenanthrene from toluene was studied using the gas antisolvent (GAS) recrystallization process. A systematic investigation of the influence of the key GAS process parameters, antisolvent addition rate (1, 20, 50, and 100 mL/min), temperature (25, 45, 55, and 65 °C), solute concentration (25%, 50%, 75%, and 100%), and agitation rate (500, 1000, 2000, and 3500 rpm), was investigated on the particle morphology, size, and size distribution. It was found using laser diffraction that increasing the antisolvent addition rate and the agitation rate, while decreasing the temperature and solute concentration, led to a decrease in the mean particle diameter. Furthermore, a unimodal particle size distribution was obtained at the higher agitation and antisolvent addition rates, but a particle size distribution of a bimodal nature was obtained at the higher temperatures and the lower agitation and antisolvent addition rates. The process parameters could be reproducibly tuned to give a mean particle diameter between 21 and 210 µm. The applicability of on-line attenuated total reflection (ATR) FTIR measurements for an improved understanding of the dynamics of the GAS process was investigated through peak analysis of the in situ ATR-FTIR spectra of phenanthrene. This work also demonstrated that ATR-FTIR on-line monitoring of the solute is a valuable technique for analyzing the GAS crystallization process. 1. Introduction Investigating particle formation techniques with the potential of producing small particles with controlled particle size distributions, ranging from nanometers to hundreds of micrometers, has attracted significant interest in the scientific and industrial communities with applications in the pharmaceutical, food, nutraceutical, chemical, paint/coating, and polymer industries.1-5 The important properties of these products are a narrow particle size distribution, a uniform morphology, and enantiomeric purity.6,7 The employment of supercritical fluid techniques has attracted considerable interest as an emerging “green” technology for the formation of particles in these size ranges.8,9 Supercritical fluids (SCFs) have several advantages over conventional liquid solvents/antisolvents as their physical properties such as density and solubility can be “tuned” within a wide range of processing conditions by varying both temperature and pressure.10 Low viscosity and high diffusivity in SCFs are considered highly effective for producing superior products of fine and uniform particles.11,12 Moreover, supercritical fluids can be easily separated from both organic cosolvents and solid products, providing a potentially clean, recyclable, and environmentally friendly technology.12,13 Carbon dioxide is by far the most widely used supercritical medium and has a number of distinct advantages. In * To whom correspondence should be addressed. Tel.: (519) 661-3466 (P.A.C.); (519) 661-4116 (S.R.). Fax: (519) 661-3498 (P.A.C.; S.R.). E-mail:
[email protected] (P.A.C.);
[email protected] (S.R.).
addition to the low critical point (Pc ) 74 bar, Tc ) 31.1 °C), CO2 is nonflammable, nontoxic, and readily available in high purity. Particle formation using SCFs can be performed using several different techniques, each of which has advantages and disadvantages. Drugs and oganic materials that are soluble in supercritical fluids are often considered to be best processed by the rapid expansion of supercritical solutions (RESS) process.14 As most pharmaceuticals/organic compounds have poor solubility in SCFs, antisolvent techniques are more attractive. These include the gas antisolvent (GAS) process, precipitation with compressed antisolvent (PCA) process (also known as the supercritical antisolvent (SAS) process and aerosol spray extraction system (ASES) process), and solution-enhanced dispersion by supercritical fluids (SEDS).4,5 Antisolvent techniques exploit the low solubility of most pharmaceutical compounds in the antisolvent, in particular CO2, which has to be miscible with the organic solvent.15,16 In the GAS process, highpressure CO2 is injected into the liquid-phase solution, which causes a sharp reduction of the solute solubility in the expanded liquid phase. As a result, precipitation of the dissolved compound occurs. The potential advantage of the GAS recrystallization process lies in the possibility of obtaining solvent-free, micrometer and submicrometer particles with a narrow size distribution. By varying the process parameters, the particle size, size distribution, and morphology can be “tuned” to produce a product with desirable qualities.17 Unfortunately, experimental studies on how to tailor the GAS micronization technique according to the characteristics
10.1021/ie050206e CCC: $30.25 © 2005 American Chemical Society Published on Web 08/20/2005
7338
Ind. Eng. Chem. Res., Vol. 44, No. 19, 2005
Figure 1. Schematic diagram of the GAS apparatus: (A) CO2 tank; (H1, H2, and H3) cooling/heating units; (PS) high-pressure syringe pump; (HPV) high-pressure vessel (crystallizer); (E) expansion vessel; (CHV) check valve; (V) on/off valve; (CV) control valve; (PI) pressure indicator; (TC) temperature controller.
of the final precipitate are rather limited. The micronization of organic compounds with well-defined physical properties and performance characteristics, suitable for specific applications, continues to be a significant challenge for the specialty chemical industry. Recent literature has shown that attenuated total reflection (ATR) FTIR spectroscopy has the technical feasibility for in situ measurement of concentration, solubility, and supersaturation in cooling crystallization.18-21 In this technique, the radiation beam is directed onto an angled crystal and reflected within the crystal until it emerges from the other “end”, where it is collected. The depth of penetration, as the beam of radiation penetrates a fraction of a wavelength beyond the reflecting surface, distinctively produces (a) little to no disturbing interference from the solid particles present in the solution and (b) a less intense solvent contribution to the overall infrared spectrum so the solvent spectra can be easily subtracted from the sample spectrum. For ATR-FTIR to be used for quantitative analysis of solute concentration, a calibration model must be employed. A few calibration approaches are reported in the literature including those by Dunuwila et al.,22 Togkalidou et al.,23 and Lewiner et al.20 This is a two-part study, with the first part investigating the applicability of the GAS recrystallization technique for the micronization of phenanthrene from toluene and the technical potential of ATR-FTIR spectroscopy for the in situ monitoring of the solute concentration profile during the GAS crystallization process. The second part of the study will present a theoretical model and simulation results for the GAS process. 2. Experimental Section 2.1. Materials. The solute-solvent model system investigated in this study was phenanthrene in toluene. The organic solvent (analytical grade) and solid solute (98% purity) were both purchased from the SigmaAldrich Co. GAS experimental work was carried out using instrument-grade CO2 (99.99% purity) obtained from BOC Canada and further purified by passage through columns containing molecular sieves (Aldrich) and copper oxide (Aldrich) to remove excess water and oxygen, respectively. 2.2. Apparatus. A schematic diagram of the constructed experimental setup to perform the GAS crystallization experiments on a laboratory scale is shown in Figure 1. Liquid CO2 was drawn from a dip tube cylinder (A) which was subcooled and compressed by
means of a syringe pump (Isco, 260D). To prevent backflow, a check valve (HIP) was connected to the carbon dioxide feed line. The syringe pump (PS) was outfitted with a pump-head cooling jacket connected through a coiled piece of tubing to a temperature-controlled water bath (VWR, 1180A). Before entering the crystallizer, the liquefied carbon dioxide was preheated in the water bath to the chosen process temperature. The highpressure vessel (HPV-Autoclave Engineers) is a stainless steel vessel of 100 mL volume, consisting of a sixbladed air-driven magnetic stirrer and a rupture disk. The vessel temperature was maintained within very close tolerance of (0.5 °C using a PID temperature controller (Fuji Electronic PXW-4). Pressure inside the vessel was monitored with an electronic pressure transducer (Omega Engineering Inc., PX302-10KGV). After precipitation, the fluid mixture was released through a sintered metal frit mounted in the outlet tube to avoid entrainment of particles. A water bath (VWR, 1180A) was used to heat the fluid mixture leaving the crystallization unit to avoid blockage of the pressure release control valve (CV-Badger Meter Inc.) due to JouleThompson effects during expansion. An expansion vessel (E) was used to collect the organic solvent from the vented CO2. In situ monitoring of the solution concentration in the stirred autoclave was performed using a high-pressure diamond immersion probe (Sentinel-ASI Applied Systems). The probe is attached to an ATR-FTIR spectrophotometer (ASI Applied Systems ReactIR 4000), connected to a microcomputer, supported by ReactIR software (ASI). 2.3. Characterization. Quantitative analysis of the precipitated particles was carried out using laser diffraction (Malvern Mastersizer 2000, Malvern Instruments Ltd.) which provided the particle size and particle size distribution (PSD) measurements for all samples. 2.4. Procedure. GAS crystallization of phenanthrene using compressed CO2 was performed by preparing a predetermined volume of phenanthrene solution (10 mL) at a saturated concentration for the given operating temperature and loading it into the 100 mL crystallization vessel. The agitator was turned on and set to the desired rpm. When the system had equilibrated thermally, the pressurization by injection of CO2 was initiated. A controlled CO2 flow rate was maintained until the full liquid volumetric expansion of 900% was achieved. Consequently, the CO2 supply feed was stopped, while mixing was continued for 1 h. Then, a rinsing step was performed by flushing the expanded liquid phase with CO2 at a constant flow rate, for a minimum period of 5 h. Finally, the crystallization vessel was depressurized by venting the entire fluid mixture of the vessel, and the dry solid powder was collected for off-line analysis. For ATR-FTIR experiments, to determine the effect of inert gas pressure on solute absorbencies, a weighed amount of phenanthrene and toluene was added to the crystallizer and then heated to the desired temperature. Subsequently, helium gas was introduced into the system until the desired pressure was obtained. Spectra were obtained at regular time intervals at the various studied pressures. To investigate the sorption and miscibility of the toluene with supercritical carbon dioxide, the crystallizer was first heated to the desired temperature. Then, CO2 was injected into the system until the desired pressure was obtained. Next, predetermined amounts of the liquid-solvent toluene were injected into the system successively. The spectra were
Ind. Eng. Chem. Res., Vol. 44, No. 19, 2005 7339 Table 1. Experimental Conditions Used for the GAS Experiments run
CO2 addition rate (mL/min)
concn ratio (%)
T (°C)
agitation rate (rpm)
A1 A2 A3 A4 C1 C2 C3 T1 T2 T3 G1 G2 G3
1 20 50 100 50 50 50 50 50 50 50 50 50
100 100 100 100 25 50 75 100 100 100 100 100 100
25 25 25 25 25 25 25 45 55 65 25 25 25
1000 1000 1000 1000 1000 1000 1000 1000 1000 1000 500 2000 3500
Table 2. Experimental Results of Phenanthrene Produced Using the GAS Process run
mean particle size (µm)
std dev (µm)
coeff of variation
A1 A2 A3 A4 C1 C2 C3 T1 T2 T3 G1 G2 G3
177.5 85.8 46.1 25.1 27.1 32.9 39.8 118.4 166.2 209.9 82.6 30.6 21.2
122.3 42.5 14.8 4.2 6.8 9.4 10.8 63.1 119.5 165.2 29.4 6.4 4.2
0.7 0.5 0.3 0.2 0.3 0.3 0.3 0.5 0.7 0.8 0.4 0.2 0.2
scanned at regular time intervals upon addition of the liquid solvent. 3. Results and Discussion Prior to embarking on the GAS micronization experiments, a comprehensive study of the solubility of phenanthrene in toluene using ATR-FTIR was investigated over a wide range of temperatures.24 The solubility of organic solute in toluene was found to be relatively high; consequently, toluene was chosen as the principle organic solvent in this study. The volumetric expansion, the pressure-temperature-volume behavior, of the investigated organic solvent, toluene, with the antisolvent, carbon dioxide, has already been studied.25,26 The literature indicates that large liquid volumetric expansions have been achieved at both subcritical and supercritical conditions. In this work, the effect of the process parameters such as the antisolvent addition rate, temperature, solute concentration, and agitation rate on the particle size, size distribution morphology, and crystallinity were investigated. The summary of the working conditions of the phenanthrene GAS processing experimental runs is reported in Table 1. Each run is represented by a label which indicates the type of the experiment according to the investigated parameter. The letters A, C, T, and G refer to the antisolvent addition rate, concentration, process temperature, and agitation rate, respectively. Table 2 provides the results of particle size measurements, mean particle size, standard deviation, and coefficient of variation of the particle size distribution. Particle size measurements were obtained from the quantitative results of the laser diffraction produced volume percent distributions using the Malvern Mastersizer 2000. For plots provided of these results below,
Figure 2. Normalized volume density distribution of phenanthrene particles determined by laser diffraction for runs A1, A2, A3, and A4. The experimental conditions are provided in Table 1.
the mean of three separate experiments is given, and the error bars represent 1 standard deviation. 3.1. Effect of the Antisolvent Addition Rate and FTIR Results. In this set of experiments, the effect of the antisolvent addition rate was investigated at four levels of the carbon dioxide addition rate, namely, 1, 20, 50, and 100 mL/min (see Table 1 for full experimental conditions). Particle size imaging by SEM was found problematic for phenanthrene due to its tendency to sublime. Hence, particle sizing was carried out exclusively by laser diffraction. Figure 2 shows the produced volume percent distributions for the experimental runs A1, A2, A3, and A4. For the lowest addition rate, i.e., 1 mL/min, the particle size distribution was bimodal due to a large degree of agglomeration, with a mean particle diameter of 177.5 µm. When the antisolvent addition rate was increased to 20 mL/min, a bimodal particle size distribution persisted, but with a smaller mean particle diameter of 85.7 µm. Unimodal particle populations with narrower size distributions were obtained by further raising the level of the antisolvent addition rate to 50 and 100 mL/min in runs A3 and A4. At the highest level of antisolvent addition rate (100 mL/min), the average particle size became far smaller, 25 µm, compared to the mean size produced at the low-level antisolvent addition rates. The final results are illustrated in graphical form in Figure 3, where the mean particle size is reported as a function of the antisolvent addition rate. It is evident that increasing the antisolvent addition rate directly lowers the mean particle size. It is evident that the nucleation and growth dynamics in the GAS crystallization process are strongly influenced by the rate of the volumetric expansion of the liquid phase. After the primary nucleation, as in the case of the low addition rate, the slow volumetric expansion rates tend to move the system toward lower levels of supersaturation, where growth is the prevailing mechanism, while the faster volumetric expansion rates tend to move the system toward higher levels of supersaturation, where large nucleation rates would be generated. This is evident from the fast expansion rate (100 mL/min), as the resulting mean particle size is almost 7 times smaller than that resulting from the slow expansion rate (1 mL/min). The extent of the bimodal nature of the particle size distribution observed in this investigation can be analyzed in terms of both (a) agglomeration and (b) the degree of competition between
7340
Ind. Eng. Chem. Res., Vol. 44, No. 19, 2005
Figure 3. Mean particle size of phenanthrene produced by the GAS process as a function of the antisolvent addition rate. The points are experimental data; the line is the best fit of the data points.
the supersaturation buildup force due to the continuous volumetric expansion and the supersaturation depletion force due to particle growth. Part 2 of this study will investigate the theoretical behavior in more detail. Study of Expansivitity Using In Situ ATR-FTIR. To investigate this phenomenon further experimentally, in situ ATR-FTIR spectroscopy was utilized for investigation of the expansivity phenomenon in the GAS crystallization process. Muller et al.17 proposed a theoretical framework to describe the dynamics of GAS crystallization in regard to the volumetric expansion rate. On the basis of this framework, we monitored the concentration trajectory of the solute during the GAS crystallization process. The phenanthrene peak at 812 cm-1 was chosen for analysis, which was found to be unobscured by other peaks.24 During a typical GAS experiment, the pressure increases to a set value as CO2 is added to the crystallizer. Hence, we first studied the effect of the inert helium head pressure (1000, 2000, and 3000 psig) at 35 °C on the absorption spectra. Helium was chosen as the inert gas to study as its solubility in organic solvents is very low and it does not have the ability to expand the liquid phase. We found that the ATR-FTIR absorption readings were not affected by the pressure, indicating that the ATR-FTIR probe’s calibration is not sensitive to the system pressure, only to concentration. We also found that the absorbance of the toluene peak increased linearly with concentration as toluene was added to supercritical CO2, indicating complete miscibility in the studied range. To demonstrate the effect of the antisolvent addition rate on the volumetric expansion of the liquid phase, and consequently supersaturation, different volumetric expansion rates (antisolvent addition rates) were investigated. The studied volumetric expansion rates are identified as fast expansion (10 mL/min), intermediate expansion (5 mL/min), and slow expansion (1 mL/min). Figure 4 shows the effect of the volumetric expansion rates on the concentration profiles for phenanthrene during GAS crystallization. In the case of the fast volumetric expansion rate, the concentration of solute remained high as the antisolvent was initially added
Figure 4. Effect of the volumetric expansion rate on phenanthrene ATR-FTIR concentration profiles. The points are experimental data, and the lines are smoothed to fit through the data. The experimental conditions are for CO2 addition rates of 1, 5, and 10 mL/min at T ) 35 °C.
to the crystallizer. After the solution of phenanthrene in toluene was expanded by 90% upon addition of CO2, a sharp decline was noted which started to level off around 500% expansion. It is likely that crystallization begins with a nucleation burst, almost catastrophic, at approximately 90% ∆V/V0. This is followed by a steep decline in supersaturation. After nucleation, the final crystal growth takes place at a low level of supersaturation. The concentration trajectories for the intermediate and slow volumetric expansion rates are significantly different compared to that for the fast expansion rate. Their initial concentration profiles decrease significantly with initial expansion. Hence, the initial supersaturation sharply decreases in both cases. The phenanthrene PSDs resulting from the different antisolvent addition rate experiments were measured and are provided in Figure 5. Table 3 gives the measured mean particle sizes and standard deviations. The results indicate that faster antisolvent addition rates led to a smaller mean particle size, with a narrower particle size distribution. The distributions are primarily unimodal, although some bimodality is evident. The results indicate that the magnitude of the supersaturation is a strong function of the applied volumetric expansion rate. A faster rate of antisolvent addition will generate higher levels of supersaturation, thus higher rates of nucleation, and consequently smaller particle sizes and narrower PSDs. 3.2. Effect of Temperature. A set of experiments was performed in a subcritical-supercritical range of temperatures and at a fixed antisolvent addition rate of 50 mL/min, while all other process parameters, namely, the antisolvent addition rate, solute concentration, and agitation rate, were fixed at constant values. As illustrated in Figure 6 and Table 2, an increase in the process temperature (25, 45, 55, and 65 °C, respectively, in runs A2, T1, T2, and T3) for the recrystallization process results in the enlargement of the phenan-
Ind. Eng. Chem. Res., Vol. 44, No. 19, 2005 7341
Figure 5. Normalized volume density distribution of phenanthrene particles determined by laser diffraction as a function of the antisolvent addition rate during ATR-FTIR experiments.
Figure 6. Mean particle size of phenanthrene produced by the GAS process as a function of the crystallization temperature. The points are experimental data; the line is the best fit of the data points. Table 3. Particle Size Measurements for Phenanthrene Samples Prepared from the GAS Process at Varying Volumetric Expansion Rates (for ATR-FTIR Experiments) CO2 flow rate (mL/min)
mean diam (µM)
std dev (µM)
1 5 10
35.0 77.3 101.5
9.5 8.7 14.0
threne mean particle size and a broadening of the size distribution. Figure 7 shows the produced volume percent distributions of the investigated runs. It is evident that the mean particle size increases and the distribution gets highly bimodal as the temperature is increased. Microparticles with mean particle sizes of 46.1 and 118.4 µm were generated at 25 and 45 °C, respectively (runs A3 and T1). The particles produced in run T2 had a mean particle diameter of 166.2 µm with a large standard deviation and coefficient of variation of 119.5 µm and 0.7, respectively. When the temperature was raised to 65 °C in run T3, a highly bimodel particle size distribution was obtained, with a far larger mean particle size of about 209.9 µm and a large standard deviation of
Figure 7. Normalized volume density distribution of phenanthrene particles determined by laser diffraction for runs A3, T1, T2, and T3. The experimental conditions are provided in Table 1. The mean particle size is indicated by an arrow at each temperature.
165.2 µm. The level of biomodality of the particle size distributions at 65 °C in run T3 is larger than that of the PSDs performed at 55 and 45 °C. The effect of temperature on the particle size can be explained in terms of the nucleation-growth dynamics during the crystallization process. Increasing the temperature will increase the solubility of the material in the organic solvent, hence moving the position of the saturation and critical supersaturation lines upward17 in addition to changing their shape. Hence, increasing the temperature lowers the magnitude of the generated supersaturation during the GAS process (analogous to lowering the volumetric expansion rate) as the profile moves closer to the saturation line. This is followed by a gradual decline-depletion in the supersaturation as the nuclei grow; i.e., a high growth rate follows. This may lead to multiple crossing of the critical metastable line, between the nucleation and metastable zones, resulting in increased bimodal behavior. Thus, a larger mean particle size with broad particle size distribution is expected. 3.3. Effect of the Solute Concentration. In this section, the influence of the initial solute concentration (varied between 25% and 100%) on the mean particle size and particle size distribution was investigated. All other operating conditions were fixed at the constant values indicated in Table 1. The concentration ratio was defined as the ratio between the actual concentration of the liquid solution and the saturation concentration. As indicated in Table 2 and illustrated in Figure 8, the higher the solute concentration, the higher the mean particle size. Figure 9 shows the laser diffraction produced volume percent distributions for the experimental runs A3, C1, C2, and C3. It is apparent that at the lowest solute concentration, i.e., 25% concentration ratio (run C1), the particle size distribution has the smallest mean particle diameter of 27.1 µm. When the concentration ratio was increased to 50%, 75%, and 100% in the experimental runs C2, C3, and A3, the mean particle size increased to 32.9, 39.8, and 46.1 µm. However, in all the investigated levels of the concentration ratio, the produced particle size distributions are rather unimodal. It is also obvious from the particle size distribution displayed in
7342
Ind. Eng. Chem. Res., Vol. 44, No. 19, 2005
Figure 8. Mean particle size of phenanthrene produced by the GAS process as a function of the solute concentration. The points are experimental data; the line is the best fit of the data points.
Figure 10. Normalized volume density distribution of phenanthrene particles determined by laser diffraction for runs G3, G2, A3, and G1. The experimental conditions are provided in Table 1. The mean particle size is indicated by an arrow at each rpm value.
Figure 9. Normalized volume density distribution of phenanthrene particles determined by laser diffraction for runs C1, C2, C3, and A3. The experimental conditions are provided in Table 1. The mean particle size is indicated by an arrow at each concentration ratio.
Figure 11. Mean particle size of phenanthrene produced by the GAS process as a function of the agitation rate. The points are experimental data; the line is the best fit of the data points.
Figure 9 that the higher the initial solute concentration, the broader the particle size distribution. The increase of particle size and the broadening of the particle size distribution with increasing solute concentration can be analyzed in terms of nucleation and growth processes. At higher solute concentrations, precipitation of the solute occurs earlier during the expansion process, resulting in increased time for crystal growth. At lower solute concentrations, precipitation of the solute is reached later during the expansion process; hence, nucleation is the prevailing mechanism, giving smaller particles. 3.4. Effect of the Agitation Rate. Experiments to study the effect of the agitation rate (500, 1000, 2000, and 3500 rpm) on the phenanthrene particle size and particle size distribution were investigated, while all other process parameters were kept at the constant values reported in Table 1. Figure 10 shows the laser diffraction produced volume percent distributions for the experimental runs A3, G1, G2, and G3. The particle size distribution generated
during with a low agitation rate, i.e., 500 rpm (run G1), is highly bimodal. The particles produced have an estimated mean particle diameter of 82.6 µm, with a large standard deviation and coefficient of variation of 29.4 µm and 0.4, respectively. When the agitation rate was increased to 1000 rpm in run A3, a unimodal particle size distribution with a mean particle size of about 46.1 µm was obtained. Unimodal particle populations with narrower size distributions were obtained by further raising the agitation rate up to 2000 and 3500 rpm in runs G2 and G3 (the mean particle sizes are 30.6 and 21.2 µm), respectively. The effect of the agitation rate on the mean particle diameter is quantitatively illustrated in Figure 11, where the mean particle size is reported as a function of the agitation rate. Hence, it is evident that the agitation rate has a significant effect on the particle size and particle size distribution, with the particle size distribution going from bimodal to a relatively narrow distribution as the agitation rate was increased. It is also clear that increasing the agitation rate directly lowers the mean particle size.
Ind. Eng. Chem. Res., Vol. 44, No. 19, 2005 7343
This study investigated the extent of the effect of the agitation rate on the volumetric expansion profile and, consequently, the characteristics of the final product. The agitation rate, in combination with the antisolvent addition rate, clearly has a strong influence on the volumetric expansion process, thus supersaturation, in terms of the required time to expand the liquid phase and achieve supersaturation. Therefore, the time needed to achieve a homogeneous volumetric expansion will be sharply reduced in the presence of high-quality mixing, i.e., high mass transfer efficiency generated by the high agitation rate. In contrast, a low agitation rate implicates a longer time for reaching supersaturation. Moreover, increasing the agitation rate can also increase the degree of crystal breakage through particle-impeller and particle-particle collisions, although this effect is believed secondary. 4. Conclusions This study showed that the micronization of phenanthrene can be obtained by means of the GAS crystallization process. Furthermore, this work demonstrated that the size and size distribution of the precipitated particles was strongly influenced by the GAS process through the manipulation of the process parameters, antisolvent addition rate, concentration, temperature, and agitation rate. Increasing the antisolvent addition rate gave smaller particles with a narrower size distribution. The rate of antisolvent addition was found to significantly affect the measured phenanthrene concentration profiles, as determined by the FTIR peak intensity. The solute concentration exhibited an opposite effect; as the initial solute concentration was reduced, an observed decrease in the particle size and particle size distribution was obtained. Higher temperatures led to an increase in the particle size and the level of bimodality. Higher agitation rates produced a smaller particle size and narrower size distribution. The achieved experimental results necessitate further investigation for a better understanding of the theoretical underpinnings of the dynamics of the GAS process and how the process parameters influence the volumetric expansion profile of the liquid phase, thus the magnitude of supersaturation, and consequently the characteristics of the final product. For the GAS process to be developed and utilized with high efficiency by the specialty chemical industry, quantitative knowledge will be the key factor in accomplishing this important futuristic task. Acknowledgment We acknowledge Dr. Todd Simpson at the University of Western Ontario (UWO) Photonics and Nanotechnology Laboratory for aid in the SEM characterization and Mr. Robert Harbottle for aid in the particle size characterization. We acknowledge the financial support of the Natural Sciences and Engineering Research Council (NSERC) of Canada, the Canadian Foundation for Innovation (CFI), and the UWO Academic Development Fund (UWO-ADF). Literature Cited (1) Cansell, F.; Chevalier, B.; Demourgues, A.; Etourneau, J.; Even, C.; Garrabos, Y.; Pessey, V.; Petit, S.; Tressaud, A.; Weill, F. Supercritical Fluid Processing: a new route for materials synthesis. J. Mater. Chem. 1999, 9, 67-75.
(2) Cooper, A. I. Recent Developments in Materials Synthesis and Processing Using Supercritical CO2. Adv. Mater. 2001, 13 (14), 1111-1114. (3) Subramaniam, B.; Rajewski, R. A.; Snavely, W. K. Pharmaceutical Processing with Supercritical Carbon Dioxide. J. Pharm. Sci. 1997, 86 (8), 885-890. (4) Tan, H. S.; Borsadia, S. Particle Formation Using Supercritical Fluids: Pharmaceutical Applicatons. Expert Opin. Ther. Pat. 2001, 11 (5), 861-872. (5) Ye, X.; Wai, C. M. Making Nanomaterials in Supercritical Fluids: A Review. J. Chem. Educ. 2003, 80 (2), 198-203. (6) Yu, B.; Shekunov, P.; York, P. Crystallization processes in pharmaceutical technology and drug delivery design. J. Cryst. Growth 2000, 211, 122-136. (7) York, P. Strategies for particle design using supercritical fluid technologies. PSIT 1999, 2 (11), 430-440. (8) Jung, J.; Perrut, M. Particle Design Using Supercritical Fluids: Literature and Patent survey. J. Supercrit. Fluids 2001, 20, 179-219. (9) Perrut, M. Supercritical Fluid Applications: Industrial Developments and Economic Issues. Ind. Eng. Chem. Res. 2000, 39, 4532-4535. (10) McHugh, M. A.; Krukonis, V. J. Supercritical Fluid Extraction: Principles and Practice, 2nd ed.; Butterworth-Heinemann: Boston, 1994. (11) Debenedetti, P. G. Homogeneous Nucleation in Supercritical Fluids. AIChE J. 1990, 36 (9), 1289-1298. (12) Tom, J. W.; Lim, G. B.; Debenedetti, P. G.; Prud’homme, R. K. Applications of SCF in the Controlled Release of Drugs. In Supercritical Fluid Engineering Science; Kiran, E., Brennecke, J. F., Eds.; ACS Symposium Series 514; Kiran, E., Brennecke, J. F., Eds.; American Chemical Society: Washington, DC, 1993; pp 238256. (13) Tom, J. W.; Debenedetti, P. G. Particle Formation With Supercritical FluidssA Review. J. Aerosol Sci. 1991, 22 (5), 555584. (14) Krukonis, V. Supercritical Fluid Nucleation of Difficult to Comminute Solids, AIChE Annual Meeting, San Francisco, November 1984; American Institute of Chemical Engineers: New York, 1984; p 140 f. (15) Gallagher, P. M. C., M. P.; Krukonis, V. J.; Klasutis, N. Gas antisolvent recrystallization: new process to recrystallize compounds insoluble in supercritical fluids. Supercritical Fluid Science Technology; ACS Symposium Series 406; American Chemical Society: Washingtion, DC, 1989; pp 334-354. (16) Gallagher, P. M.; Krukonis, V. J.; Botsaris, G. D. Gas antisolvent recrystallization: Application to particle design. In Particle Design via Crystallization; Ramanarayanan, R., Kern, W., Larson, M., Sidkar, S., Eds.; AIChE Sympsium Series 284; American Institute of Chemical Engineers: New York, 1991; pp 96-103. (17) Muller, M.; Meier, U.; Kessler, A.; Mazzotti, M. Experimental study of the effect of process parameters in the recrystallization of an organic compound using compressed carbon dioxide as anti-solvent. Ind. Eng. Chem. Res. 2000, 39, 2260-2268. (18) Dunuwila, D.; Berglund, K. A. ATR FTIR spectroscopy for in situ measurement of supersaturation. J. Cryst. Growth 1997, 179, 185-193. (19) Groen, H.; Roberts, K. J. Nucleation, growth and pseudopolymorphic behavior of citric acid as monitored in situ by attenuated total reflection Fourier transform infrared spectroscopy. J. Phys. Chem. B 2001, 105 (43), 10723-10730. (20) Lewiner, F.; F’evotte, G.; Klein, J. P.; Puel, F. Improving Batch Cooling Seeded Crystallization of an Organic Weed-Killer Using On-Line ATR FTIR Measurement of Supersaturation. J. Cryst. Growth 2001, 226, 348-362. (21) Feng, L.; Berglund, K. A. ATR FTIR for determination optimal cooling curves for batch crystallization of succinic acid. Cryst. Growth Des. 2002, 2 (5), 449-452. (22) Dunuwila, D. D.; Caroll, L. B.; Berglund, K. A. An investigation of the applicability of attenuated total reflection infrared spectroscopy for measurement of solubility and supersaturation of aqueous citric acid solutions. J. Cryst. Growth 1994, 137, 561-568.
7344
Ind. Eng. Chem. Res., Vol. 44, No. 19, 2005
(23) Togkalidou, T.; Fujiwara, M.; Patel, S.; Braatz, R. D. Solute concentration prediction using chemometrics and ATR-FTIR spectroscopy. J. Cryst. Growth 2001, 231, 534-543. (24) Bakhbakhi, Y.; Charpentier, P. A.; Rohani, P. The Solubility of Phenanthrene in Toluene: In-situ ATR-FTIR, Experimental Measurement, and Thermodynamic Modeling. Can. J. Chem. Eng. 2005, 83, 267-273. (25) Dixon, D. J.; Johnston, K. P. Molecular thermodynamics of solubilities in gas antisolvent crystallization. AIChE J. 1991, 37, 1441-1449.
(26) Berends, E. M.; Bruinsms, O. S. L.; Graauw, J. D.; Rosmalen, G. M. V. Crystallization of Phenanthrene from Toluene with Carbon Dioxide by the GAS Process. AIChE J. 1996, 42, 431435.
Received for review February 20, 2005 Revised manuscript received June 20, 2005 Accepted July 12, 2005 IE050206E
