CRYSTAL MORPHOLOGY AND GROWTH RATE OF
NAPHTHALENE IN VARIOUS PROCESSES INVOLVING
SUPERCRITICAL CARBON DIOXIDE
C. Y. TAI and C.-S. CHENG
Department of Chemical Engineering, National Taiwan University, Taipei, Taiwan
A
n experimental apparatus was designed to investigate the crystal morphology and growth rate of naphthalene crystals grown in various processes using supercritical carbon dioxide, including SCG (single crystal growth), SESS (slow expansion of supercritical solution), and RESS (rapid expansion of supercritical solution). Photo-graphs were taken to identify crystal habits. The growth rates of naphthalene in the SCG and SESS processes were measured and compared with those in a RESS process. Keywords: crystal morphology; growth rate; supercritical CO2; RESS process; SESSprocess; single crystal growth
INTRODUCTION
Recently, the RESS process was applied to produce small, monodisperse particles, which utilized the dis-tinctive features of a supercritical solution for its fast propagation of pressure perturbation and high saturation ratio. The former leads to uniform super-saturation and hence to narrow particle size distribution; the latter, to small particles. The crystals nucleate instantly with limited growth during the expansion step1,2. Thus, the phenomena of nucleation from a supercritical ¯ uid has been intensively investigated3±5, leaving the study of crystal growth neglected. However, Mohamed et al.3 produced fairly large naphthalene
crystals by expanding a supercritical carbon dioxide solution through a nozzle. The size of naphthalene crystals so obtained was over 200l m in some cases. Apparently, the naphthalene crystals were subject to substantial growth during the expansion process.
Meanwhile, Chimowitz and his co-workers6,7 devel-oped a separation process exploiting the cross-over region, which is a distinguishing feature of supercritical ¯ uids and concerns the solute solubility. In the process, the solutes deposited under a much weaker super-saturation as compared with the RESS process. There-fore, the particles produced have a chance to grow bigger. Later, using a batch crystallizer to study the crystallization mechanism in a supercritical carbon dioxide solution, Tavana and Randolph8con® rmed the validity of the conventional concept that the crystals are formed by nucleation and subsequent growth. The crystal growth rate of benzoic acid in supercritical carbon dioxide solution was estimated to be of the same order of magnitude as crystal growth in aqueous media.
To study the crystal growth from supercritical solution more systematically, Tai and Cheng9set up an appara-tus, which can be used to observe the growth phenom-enon and to measure the growth rate of naphthalene
crystal growing in a supercritical carbon dioxide solu-tion. The pressure, temperature, supersaturation, and relative velocity between solution and crystal was varied to investigate their e ects on the crystal morphology and growth rate. Their experimental work was later extended to SESS and RESS processes with a minor modi® cation of the apparatus. The supersaturation is generated by adjusting temperature in the SCG process and by releasing pressure in the SESS and RESS processes. The supersaturation level is high in the RESS process, and moderate to low in the SESS and SCG processes. Thus, a di erent crystal morphology and growth rate would be expected for the crystals generated from the various modes of operation. The aim of this report is to compare the results obtained in the series of growth experiments.
EXPERIMENTAL
The design principle of the apparatus used in this experiment is similar to that of growing crystals from liquid solution10. The set-up was able to measure the solubilities of solid materials, to observe growth phe-nomena, and to measure growth rate in supercritical carbon dioxide. The major part of the apparatus, shown schematically in Figure 1, is a closed loop consisting of a growth cell for growing crystals, an extractor for supplying solute, and a piston pump for circulating solution. In the extractor, a crystal bed of ground naphthalene (Wako, reagent grade) with a dimension 0.018 m
´
0.25 m was packed. The growth cell is a jacketed liquid-level gauge (Jerguson, Model 18-T-30) with viewing windows. The circulation pump is a variable-stroke dual-piston type, a Milton Roy Mini-pump and a Clark-Cooper Mini-pump for low and high circulation rates, respectively. The pipeline of the closedloop was built of 1/49 9 O.D. stainless steel tubes. In the SCG and SESS experiments, the valve V3was closed and
the total volume of the closed loop is about 3.5
´
10-4m3. On the other hand, the valves V4and V5
were closed for the RESS experiment, and the total volume of the closed loop excluding the growth cell is around 1.5
´
10-4m3.The seed crystals of naphthalene for the SCG experiment were prepared from ethanol solution; several small crystals obtained by evaporating alcohol were introduced into a slightly supersaturated ethanol solu-tion at room temperature to induce nucleasolu-tion, and the induced nuclei were allowed to grow without stirring. Well-grown crystals, as shown in Figure 2, so prepared were bounded by {001}, {20Å1} and {110} faces, the same
morphology as that reported by Wells11. A seed crystal of 1.5
´
10-3m was chosen and glued on the tip of aneedle, which was then screwed into the growth cell. The system was ready for the growth experiment. The
detailed procedure of measuring the solubility and the growth rate of a single crystal is described elsewhere9. It
should be noted that the supersaturation for single crystal growth was generated by heating, i.e., the temperature in the growth cell was higher than that in the extractor, because the system was operated in the retrograde region where higher temperature causes lower solubility.
In the SESS experiment, carbon dioxide was ® rst introduced to the closed loop and circulated to saturate the supercritical carbon dioxide solution with naphtha-lene at desired pressure and temperature. Then the valve V2 was opened to release the supercritical ¯ uid
through the expansion valve (Autoclave, Model 30 VRMM) at a rate of 2±5lit/min measured at 1 atm. During the expansion period, the pressure and tem-perature versus time were recorded and the solid formation in the growth cell, either on the tip of the needle or on the glass wall, was detected by visual examination. Then photographs of the crystal were
Figure 1. Apparatus for the experiments of crystallization from
supercritical carbon dioxide solution. (1) Carbon dioxide cylinder; (2) Silica-gelbed;(3) Filter; (4) Cooler;(5) Feeding pump;(6) Extractor;(7) Crystal growth cell; (8) Needle to ® x seed crystal; (9) Microscope; (10) Circulation pump; (11) Expansion valve; (12) Crystal collector; (13) Flowmeter; (14) Wet test meter; PI, Pressure indicator; TI, Tempera-ture indicator; TC, TemperaTempera-ture controller; V1, V2, V3, V4, V5, Ball
valves.
Figure 2. Seed crystal of naphthalene.
Figure 3. Development of individuals and sprouting of plates for
naphthalene crystals from supercritical CO2solution.
Figure 4. E ect of pressure and temperature on metastable
taken during a certain growing time using a stereo microscope (Nikon, Model SMZ-10 ). Thus, the crystal growth rates were estimated. In this experiment, various pre-expansion pressures, temperatures, and solution concentrations were tested.
In the RESS experiment with the valve V4 and V5
closed, carbon dioxide becomes saturated in the closed loop containing only the extractor under desired the temperature and pressure. Then the valve V2was opened
to expand the supercritical solution into the crystal collector. Meanwhile, the valve V1 was opened to
maintain the pressure and temperature steady in the extractor by supplying carbon dioxide through valve V1.
Photographs of the collected crystals were taken.
RESULTS AND DISCUSSION Crystal Morphology and Growth Phenomena Single crystal growth
There are two distinct features for the crystal growth of a naphthalene seed in a supercritical carbon dioxide solution, as seen in Figure 3. First, many small individual
crystals with light-re¯ ecting faces developed on the seed crystal, which had been partially dissolved to become a hemi-ellipsoid before the solution became saturated. Second, plates sprouted from some of the individual crystals, which located away from the pole of the hemi-ellipsoid.
The crystal faces of the individuals might change from facet growth to hopper growth when they grew bigger. Critical sizes for the hopper growth are in the range of 3
´
10-4to 7´
10-4m, depending on the saturation andsolution velocity. Higher supersaturation and lower solution velocity rendered the critical size smaller. One interesting growth phenomenon concerning the sprouted plates is that the lateral faces of the plates might grow stably to remain smooth, or unstably to become zigzag or dendritic.
The mechanism and kinetics of the two types of growth, individuals and sprouted plates, have been
Figure 5. E ect of depressurization rate on the metastable
super-saturation of naphthalene in near-critical and supercritical CO2
solution.
Figure 6. Circular macrostepson the nucleatednaphthaleneplatesfrom
supercriticalCO2solution in SESS experiment.
Figure 7. Nucleated naphthalenegranules(a), platesnd needles(b), and
discussed by Tai and Cheng9. They conclude that the growth of naphthalene crystals from supercritical solu-tion shows characteristics similar to liquid-solusolu-tion growth rather than vapour growth.
Slow expansion of supercritical solution
For the slow expansion experiment the metastable supersaturation, i.e., the lowest supersaturation for the formation of nuclei detected by visual examination, was plotted against temperature at which nuclei were detected, as shown in Figure 4. It is clear that the metastable supersaturation depends on the temperature and pressure. However, the metastable supersaturation is independent of depressurization rate between 0.5 and 2 bar/min as illustrated in Figure 5.
The morphology of naphthalene crystals, as shown in Figure 6, is of hexagonal plates with circular, or sometimes irregular, macrosteps to form a hopper structure, which is similar to that of an individual crystal developed from a seed crystal. The size of the hexagonal plate is between 0.25 to 1.0
´
10-3m. Apparently, thegenerated nuclei were subject to growth for a long period of time.
Rapid expansion of supercritical solution
The morphology of naphthalene precipitated in the crystal collector after rapid expansion is shown in Figure 7 and summarized in Table 1 for two di erent pre-expansion pressures. When the pre-expansion
pressure was set at 76.9 bar, crystal habits of granule, plate, needle, and dendrite form were observed independent of the pre-expansion temperature. How-ever, when the pre-expansion pressure was set at 83.8 bar, needle and plate crystals were missing for the pre-expansion temperature of 39.0°C and 44.9°C respec-tively. Thus, the crystal morphology was in¯ uenced by the pre-expansion pressure and temperature. The di erent crystal habits exhibited by RESS prepared particles implys varying growth rates on individual crystal faces, even during the extremely short time for growth. The crystal size of precipitated naphthalene is listed in Table 1, with the largest size up to 400l m. It is an unusual size of crystal obtained from a RESS process. For example, the ceramic powders produced from the expansion of supercritical water are sub-microns in size12; even organic materials precipitated from the expansion of supercritical carbon dioxide are usually small2. However, Mohamed et al.3 obtained fairly large naphthalene crystals, which is similar to the present result, up to 225l m. The crystal size and estimated growth rate of naphthalene of some organics precipitated from the expansion of supercritical solu-tion are tabulated in Table 2. It is noted that the crystal size and growth rate of naphthalene crystal are much larger than other crystals. One possible reason is that a liquid phase was formed during expansion to give larger particle size; the other may be that the solute clusters experienced an unusually high rate of agglom-eration in the expansion region due to high super-saturation.
Table 1. Crystal habits and granule size of nucleated naphthalenecrystals from supercritical CO2in the RESS experiments.
Crystal habita
P1 T1 C1 T2 Qave Granule size
(bar) (°C) (g/100g) (°C) (l/min) g p n d (l m) 76.9 35.5 0.265 28.8 2.8 20±200 39.9 0.175 30.0 3.5 50±400 45.0 0.169 30.5 2.0 100±150 83.8 39.0 0.605 27.0 2.2 ± 44.9 0.282 27.5 2.1
-ag=granule, p=plate, n=needle, d=dendrite P1pre-expansion pressure
T1pre -expansion temperature T2post-expansion temperature Qavedischarge rate
Table 2. Estiamted growth rate of various crystals in theRESS proces.
Nozzle diameter length Ya 1 Lm Rb Crystal (l m) (mm) (´10-3) (l m) (m s-1) Reference Lovastatin 25 0.25 0.055 0.26 0.026 [13] Benzoic acid 50 0.20 1.7 8.1 0.81 [4] Phenanthrene 25 0.25 2.1 15 1.5 [14] Naphthalene 25 0.25 26 225 22.5 [3]
metering valve ± 400 40 this work
amole fraction of solute before expansion bthe maximum particle sizes (L
m) divided by the residence time of
¯ uid in the expansion to obtain the growth rates; the residencetime was taken15as 10-5s.
Table 3. Crystal growth rate obtained by various growth techniques
from supercriticalCO2.
Growth rate
Crystal Technique (m/s) Reference Naphthalene SCG 3´10-9~9´10-8 this work
Napthalene, SESS 8´10-9~4´10-7 this work
Benzoic acid [8]
Napthalene, RESS 0.02~23 seeTable 2 Benzoic acid,
Phenanthrene, Lovastatin
Comparison of the Growth Processes Width of metastable region
There are two methods to generatethe supersaturation of a supercritical solution, i.e., by adjusting the system pressure (depressurization) or temperature (heating or cooling, depending on the nature of solubility). The former was employed in the SESS process and the later in the SCG process. However, the width of metastable region is quite di erent for the processes. At 45°C and 84 bar, the supersaturationr (relative supersaturation) was up to 0.1 for the depressurization, but 0.4 for the heating. This means that a pressure disturbance induces primary nucleation more easily than a temperature disturbance.
Crystal growth rate
Although the technique for estimating the crystal growth rate is di erent for the three processes, the crystal growth rates are all expressed as the change of a speci® ed linear dimension: the advancement of a crystal face in the SCG process; the change of a longest linear dimension in the SESS process; the maximum size divided by the growing time in the RESS process. A comparison of crystal growth rates obtained by various techniques from supercritical carbon dioxide solution is presented in Table 3. The growth rate of a naphthalene single crystal is approximately between 10-9and 10-7m s-1, which is
close to the rate of crystal growth from aqueous solution. The growth rates of naphthalene crystal obtained by the SESS process is of the same order of magnitude as that obtained by single crystal technique. Joined this group in the order of magnitude is the growth rate of benzoic acid crystals estimated from a batch experiment conducted by Tavana and Randolph8. Therefore, their batch experi-ment is probably similar to the slow expansion process. The growth rates for several organics precipitated in the RESS process are much higher, by 5 order of magnitude at least. Apparently, the growth mechanism of the RESS process is di erent from that of other processes.
CONCLUSION
An apparatus was designed to study the crystal growth of naphthalene from supercritical solution for di erent processes, including SCG, SESS, and RESS. The
metastable region for crystal growth was wider by adjusting the temperature than the pressure; the former was used to obtain the supersaturation in the SCG process and the later in the SESS process. Di erent crystal morphologies were obtained from various pro-cesses and under di erent operating conditions in the same process. The particle size of naphthalene crystal varied considerably at di erent operating conditions in the RESS process.
REFERENCES
1. Krukonis, V. J., 1984, `Supercritical¯ uid nucleation of di cult-to-comminutesolids, AIChE meeting, San Francisco, November, 1984, paper 140f.
2. Tom, J. W. and Debenedetti, P. G., 1991, J Aerosol Sci 22 (5): 555. 3. Mohamed, R. S., Debenedetti, P. G., and Prud’homme, R. K.,
1989, AIChEJ, 35 (2): 325.
4. Berends, E. M., Bruinsma, O. S. L., and van Rosmalen, G. M., 1993, J Crystal Growth, 128: 50.
5. Liang, M.-T., The Application of nucleationtheoriesto the particle formation in expansion of supercriticalsolutions, In third Int Symp
on SupercriticalFluids, Strasbourg, France October, 1994. Tome 3,
247±252,.
6. Chimowitz, E. H. and Pennisi, K. J., 1986, AIChEJ, 32 (10): 1665. 7. Chimowitz, E. H. and Kelley, F. D., 1989, AIChE J, 35 (6): 981. 8. Tavana, A and Randolph, A. D., 1989 AIChE J, 35 (10): 1625. 9. Tai, C. Y. and Cheng, C.-S., 1995, AIChE J, 41 (10): 2227. 10. Tai, C. Y., Cheng C.-S. , and Huang, Y.-C. , 1992, J Crystal
Growth, 123: 236.
11. Wells, A. F., 1946, Phil Mag, 37: 184.
12. Matson, D. W., Petersen, R. C., and Smith, R. D., 1987, Advances
in Ceramic, 21: 109..
13. Mohamed, R. S., Halverson D. S., Debeenedetti, P. G., and Prud’homme, R. K., 1989, ACS Symp Ser No 406, 355.
14. Shaub, G. R., Brennecke, J. F., and Mccready, M. J., 1991, In Proc
2nd Int Symp on Supercritical Fluids, 338.
15. Peterson, R. C., Matson, D. W., and Smith, R. D., 1986, J Am
Chem Soc, 108: 2100.
ACKNOWLEDGEMENT
The author gratefully acknowledge the ® nancial support of National Science Council of the Republic of China.
ADDRESS
Correspondence concerning this paper should be addressed to Dr C. Y. Tai, Department of Chemical Engineering, National Taiwan University, Tapei, Taiwan.
