Results and Discussion

3.1. Vapor-liquid equilibrium phase diagram of anti-solvent + solvent system

the extremely minute solubility of pigment green 36 in supercritical carbon dioxide, the phase behavior of the mixtures in the precipitator should be very similar to that of the mixtures without containing the pigment, i.e., the binary mixtures of carbon dioxide + quinoline [6]. The vapor-liquid phase boundaries (including dew points and bubble points) of the binary system of carbon dioxide + quinoline were thus measured with the volume-variable phase equilibrium analyzer over a temperature range of 308.2 K to 328.2 K and up to critical pressures. Fig. 2 illustrates the experimental results in which xCO2 stands for the mole fraction of carbon dioxide. This phase diagram is divided into four phase regions, including vapor-liquid coexistence (V-L), compressed liquid (L), superheated vapor (V), and supercritical (SC) regions. The critical pressure and composition at each temperature were estimated from the maximum point of the smoothed isothermal phase boundary near the critical region. The estimated critical points are located at 308.2 K/23.9 MPa/x_CO2 = 0.843, 318.2 K/25.3 MPa/x_CO2= 0.842, and 328.2 K/26.9 MPa/x_CO2= 0.842, respectively.

3.2. Influence of phase behavior on morphology of resultant particles

The precipitation experiments were conducted at different volumetric flow rates of pigment solution (F, from 0.0100 cm³/s to 0.0667 cm³/s), precipitation temperatures (T, from 308.2 K to 328.2 K), and precipitation pressures (P, from 20 MPa to 31 MPa), but the concentration of the injected pigment solution was fixed at 0.08 kg/m³ through all the experimental runs due to the low solubility of pigment green 36 in quinoline.

The experimental conditions and the results are reported in Table 1, where the densities of carbon dioxide at precipitation temperature and pressure were taken from the NIST Chemistry WebBook [11].

The composition variations of fluid mixtures in the precipitation chamber during the SAS process were calculated, in this study, from unsteady-state material balance equations around the precipitator. The estimated results can be mapped onto the phase diagram of carbon dioxide + quinoline to identify the phase regions being passed during the SAS procedure. Similar to the work of Schmitt et al. [12], the precipitator was assumed as a continuously stirred tank reactor (CSTR) with two inlet and one outlet streams. With the initial condition of pure carbon dioxide (zCO2= 1.0, where zCO2

[12].

where M_CO2 (g/s) is the mass flow rate of the inlet carbon dioxide, x_S is the mass fraction of solvent (quinoline) in the injected pigment solution (x_S≒1.0), is the density of pigment solution (about 1.090 g/cm³), F (cm³/s) is the volumetric flow rate of the injected pigment solution, and M^sys_f (g) is the total mass of fluid phase in the precipitator which was assumed to be remained a constant with time and was approximated by the mass of pure carbon dioxide in the whole precipitator at the beginning of injection of pigment solution.

Upon ending injection (at t = t^o), a purge (or drying) operation was followed by continuously charging the pure carbon dioxide into the precipitator. The variation of the mass fraction of carbon dioxide (z’CO2) during this purge stage was also derived by Schmitt et al. [12] as follows.

  ^{ }  

where z^o_CO2 is the mass fraction of carbon dioxide at the end of injection (t = t^o). The detailed derivation for eqs (1) and (2) has been given by Schmitt et al. [12]. The mass fraction zCO2can be readily converted into mole fraction xCO2by



 



Fig. 3 is an illustrative example of the calculated mole fraction of carbon dioxide xCO2 in the precipitator varying with time, including injection and purge stages. The horizontal lines on the graph represent the vapor-liquid phase boundaries, i.e., the composition of the dew point at the corresponding precipitation conditions. The mole fraction of carbon dioxide decreases from unity to a minimum value (x^oCO2) at the end of pigment solution injection. After that, x’_CO2 increases with an increase of time and approaches to unity when the residual solvent in the precipitator was totally removed.

horizontal dashed line twice. It means that the fluid mixtures in the precipitator passed by the vapor-liquid coexistence region during the SAS process. While the mixing in precipitator was non-ideal, the actual minimum concentration of carbon dioxide x^oCO2

should be lower than the value calculated from eq (1).

The estimated minimum concentrations of carbon dioxide x^oCO2, at the end of pigment solution injection, were marked on Figs. 4, 5, and 6 for the experiments conducted at 308.2 K, 318.2 K and 328.2 K, respectively. The horizontal dashed line stands for the critical pressure at the precipitation temperature and the solid curve for the vapor-liquid phase boundary. The particles formation occurred in the supercritical region, if the points are above the horizontal dashed line, such as runs #1, #3 to #6, #10 to #13 (in Fig. 4), run #15 (in Fig. 5), and run #16 (in Fig. 6). The precipitation and purge stages were operated in the superheated vapor regions, if the points are located under the horizontal dashed line but well behind the right hand side of the vapor-liquid phase boundaries, such as run #2 (in Fig. 4), runs #8 and #14 (in Fig. 5), and run #9 (in Fig. 6). In the case of the points located on the left hand side or near the phase boundaries, such as run #17 (in Fig. 4), run #18 (in Fig. 5), and runs #19 and #20 (in Fig. 6), the SAS procedure was likely to crossing the vapor-liquid coexistence region.

Although the estimated minimum compositions of carbon dioxide of runs #17 and #20 are located in the superheated vapor region, actually these runs should pass by the two-phase coexistence region at the end of pigment solution injection because the recovered quinoline from the separator was found to turn into greenish. As noted earlier, this discrepancy may be attributed to the non-ideal mixing in the precipitator.

Fig. 7 shows the FESEM images of the samples prepared in the vapor-liquid coexistence region (run #18, Fig. 7(a)), in the supercritical region (run #10, Fig. 7(b)), and in superheated vapor region (run #14, Fig. 7(c)), respectively. Fig. 8 is the corresponding enlarged FESEM images. As seen from the images, the shape of the resultant primary particles of pigment green 36 is near sphere. The images also illustrate that micron-metric aggregated ball-like particles were formed when the SAS process passed by the vapor-liquid coexistence region. Nano-metric and less aggregated pigment particles were produced when the SAS process was operated in

cases the particles formed in “dry”environments.

3.3. Influences of process parameters on particle size

Figs. 9(a) to 9(c) present the mean particle size varying with flow rate of pigment solution (F), precipitation pressure (P), and precipitation temperature (T), respectively.

These graphs show that the differences of mean particle sizes among all the resultant pigment products are almost within the experimental uncertainty (3 nm) over the entire range of the individual process parameters, and the mean sizes of the majority cases are in the range of 40 nm to 50 nm. Additionally, the dependences of resultant particle sizes on both temperature and pressure can also be represented as a function of density. In the precipitation experiments, carbon dioxide is a predominant component in the fluid mixtures, and thus the mixture density should be very close to the density of pure carbon dioxide. Fig. 10 depicts the mean particle size varying with the density of carbon dioxide for those experiments operated in either supercritical or superheated region (runs #1 to #16). In these two homogeneous phase regions, a qualitative trend was found that higher densities of the fluid mixtures in the precipitator are favorable to reduce the mean size of resultant pigment particles. Nevertheless, nano-metric pigment particles of green 36 can be produced over a wide range of operating conditions, only if the precipitation is conducted in a homogeneous phase region, either at supercritical or at superheated vapor states. When the precipitation pressures are lower than the corresponding critical value, the relative flow rates between of carbon dioxide and the pigment solution should be sufficiently large to ensure that the precipitation was made in the superheated vapor region.

4. Conclusions

Nano-particles of pigment green 36 have been successfully prepared with a continuous supercritical anti-solvent (SAS) apparatus by using quinoline as a solvent and supercritical carbon dioxide as an anti-solvent. As evidenced from the experimental results, the morphology of the resultant products is closely dependent on the phase regions where the precipitations were implemented. Micro-meteric aggregated ball-like particles were produced as the SAS process passed by the

concentrations of fluid mixtures in the precipitator during the SAS process can be estimated by a simple CSTR model. The vapor-liquid equilibrium phase diagram together with this CSTR model provides valuable information for manipulating the particulate products. The experimental results also showed that higher densities of fluid mixtures in the precipitator may be favorable to reduce the particle size of the resultant products.

Acknowledgements

The authors gratefully acknowledged the financial support of the National Science Council, Taiwan, through grant No. NSC-94-2214-E011-009. The authors also thank Mr. H.Y. Chiu for the measurements of phase boundaries and Dr. J.T. Chen, Department of Chemical Engineering, Ming Hsin University of Science and Technology, for valuable discussion.

References

[1] Wu HT, Lee MJ, Lin HM. Nano-particles formation for pigment red 177 via a continuous supercritical anti-solvent process. J. Supercrit. Fluids 2005; 33: 173.

[2] Reverchon E. Supercritical antisolvent precipitation of micro- and nano-particles. J.

Supercrit. Fluids 1999; 15: 1.

[3] Jung J, Perrut M. Particle design using supercritical fluids: Literature and patent survey. J. Supercrit. Fluids 2001; 20: 179.

[4] Gao Y, Mulenda TK, Shi YF, Yuan WK. Fine particles preparation of Red Lake C Pigment by supercritical fluid. J. Supercrit. Fluids 1998; 13: 369.

[5] Hong L, Guo JZ, Gao Y, Yuan WK. Precipitation of microparticulate organic powders by a supercritical antisolvent process. Ind. Eng. Chem. Res. 2000; 39:

4482.

[6] Wu HT, Lee MJ, Lin HM. Precipitation kinetics of pigment blue 15:6 sub-micro particles with a supercritical anti-solvent process. J. Supercrit. Fluids 2006; 37:

220.

[7] Reverchon E, Adami R, De Macro I, Laudani CG, Spada A. Pigment red 60

micronisation using supercritical fluids based techniques. J. Supercrit. Fluids 2005;

[9] Wubbolts FE, Bruinsma OSL, van Rosmalen GM. Dry-spraying of ascorbic acid or acetaminophen solutions with supercritical carbon dioxide. J. Crystal Growth 1999;

198/199: 767.

[10] Reverchon E, Caputo G, De Marco I. Role of phase behavior and atomization in the supercritical antisolvent precipitation. Ind. Eng. Chem. Res. 2003; 42: 6406.

[11] “Isothermal properties for carbon dioxide,”NIST Chemistry WebBook, NIST Standard Reference Database No. 69 - March, 2003, Release, National Institute of Standard and Technology, USA (http://webbook.nist.gov/chemistry/).

[12] Schmitt WJ, Salada MC, Shook GG, Speaker III SM. Finely-divided powders by carrier solution injection into a near or supercritical fluid. AIChE Journal 1995; 41:

2476.

Table 1 Experimental conditions and results

C P T F CO2a Mean

size

^b

Run kg/m³ MPa K cm³/s g/cm³ nm nm

1 0.08 25 308.2 0.0167 0.902 44.5 6.5

2^c 0.08 23 308.2 0.0167 0.889 48.6 9.4

3 0.08 27 308.2 0.0167 0.913 45.2 6.2

4 0.08 28 308.2 0.0167 0.919 45.5 6.3

5 0.08 29 308.2 0.0167 0.924 40.0 5.4

6 0.08 31 308.2 0.0167 0.934 45.1 7.0

7 0.08 25 313.2 0.0167 0.879 49.0 6.9

8^c 0.08 25 318.2 0.0167 0.857 47.6 7.1

9^c 0.08 25 328.2 0.0167 0.811 50.8 6.6

10 0.08 25 308.2 0.0100 0.902 45.9 7.3

11 0.08 25 308.2 0.0250 0.902 42.7 6.5

12 0.08 25 308.2 0.0500 0.902 44.5 7.8

13 0.08 25 308.2 0.0667 0.902 45.7 7.5

14^c 0.08 23 318.2 0.0167 0.841 49.5 8.9

15 0.08 28 318.2 0.0167 0.878 49.7 7.0

16 0.08 28 328.2 0.0333 0.836 51.9 7.0

17^d 0.08 20 308.2 0.0500 0.866 -

-18^d 0.08 20 318.2 0.0500 0.813 -

-19^d 0.08 20 328.2 0.0500 0.755 -

-20^d 0.08 23 328.2 0.0167 0.791 -

-aDensity of carbon dioxide at precipitation temperature and pressure. The values were taken from the NIST Chemical WebBook [11].

bStandard deviation.

cOperated in superheated vapor region. Precipitations were conducted in supercritical region if not indicated.

dPassed by vapor-liquid coexistence region.

Fig. 1 Chemical structure of pigment green 36

Fig. 2 VLE phase diagram of carbon dioxide (1) + quinoline (2) system.

Fig. 3 Mole fractions of carbon dioxide in precipitator varying with time during the SAS process

Fig. 4 VLE phase boundary and the estimated compositions of carbon dioxide in precipitator at the end of pigment solution injection for the experimental runs at 308.2 K.

Fig. 5 VLE phase boundary and the estimated compositions of carbon dioxide in precipitator at the end of pigment solution injection for the experimental runs at 318.2 K.

Fig. 6 VLE phase boundary and the estimated compositions of carbon dioxide in precipitator at the end of pigment solution injection for the experimental runs at 328.2 K.

Fig. 7 FESEM images of the samples taken from the experiment of (a) run #18, at T

= 318.2 K, P = 20 MPa, F = 0.05 cm³/s, crossing the vapor-liquid coexistence region; (b) run #10, at T = 308.2 K, P = 25 MPa, F = 0.01 cm³/s, in

supercritical region; (c) run #14, at T = 318.2 K, P = 23 MPa, F = 0.0167 cm³/s, in superheated vapor region.

Fig. 8 The corresponding enlarged FESEM images of Figs. 7(a), 7(b), and 7(c).

Fig. 9 Mean particle size varying with the process parameter of: (a) flow rate of pigment solution, at P = 25 MPa and T = 308.2 K; (b) pressure, at T = 308.2 K and F = 0.0167 cm³/s; (c) temperature, at P = 25 MPa and F = 0.0167 cm³/s.

Fig. 10 Mean particle size varying with density of carbon dioxide at the precipitation conditions.

N

0 1000 2000 3000 4000 5000

0.8 0.84 0.88 0.92 0.96 1

(b)

(c)

(a)

(b)

0 0.02 0.04 0.06 0.08

F (cm³/s)

30 40 50

MeanSize(nm)

(a)

22 24 26 28 30 32

P (MPa)

30 40 50 60

MeanSize(nm)

(b)

305 310 315 320 325 330

T (K)

30 40 50 60

MeanSize(nm)

(c)

0.8 0.84 0.88 0.92 0.96

(g/cm

30 40 50 60

MeanSize(nm)

Fig. 10

第三章