Generation of Silk Fibroin Nanoparticles via Solution-Enhanced

Feb 21, 2013 - ... or precipitation temperature can increase PS and PSD of SF nanoparticles. The nanoparticle formation mechanism was elucidated throu...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/IECR

Generation of Silk Fibroin Nanoparticles via Solution-Enhanced Dispersion by Supercritical CO2 Zheng Zhao,† Yi Li,*,† Ai-Zheng Chen,†,‡ Zi-Jian Zheng,† Jun-Yan Hu,† Jia-Shen Li,† and Gang Li† †

Institute of Textiles and Clothing, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong College of Chemical Engineering, Huaqiao University, Xiamen, Fujian 361021, China

‡

S Supporting Information *

ABSTRACT: A solution-enhanced dispersion by supercritical CO2 (SEDS) was employed to prepare silk fibroin (SF) nanoparticles. The results of 24 full factorial experiment indicated that SF nanoparticles with particle size (PS) from 52.5 to 102.3 nm and particle size distribution (PSD) from 0.32 to 0.66 can be fabricated successfully. Moreover, reducing precipitation pressure or increasing concentration of SF solution, flow rate of SF solution, or precipitation temperature can increase PS and PSD of SF nanoparticles. The nanoparticle formation mechanism was elucidated through the formation and growth of SF nuclei in the gaseous miscible phase evolved from initial droplets generated by liquid−liquid phase split. Mass transfer between supercritical CO2 and SF solution superimposed on supersaturation was the most important process parameter affecting nanoparticle formation. Furthermore, Fourier transform infrared spectroscopy and X-ray powder diffraction analysis revealed that SF nanoparticles exhibited predominant random coil and α-helix structure with minor β-sheet conformation. microspheres.14 Moreover, spray drying requires higher temperatures and may induce denaturation of active substances. Precipitation processes based on supercritical fluids (SCFs) are green technologies to produce nanosized materials mainly for pharmaceutical, polymer, and catalyst processing. Compared to conventional micronization techniques, such as milling, spray drying, and solvent evaporation, the SCF-based micronization processes present many advantages, including mild operating temperature, less or no solvent residue, efficient separation, and the fact that it is environmentally benign.22 Among all the possible SCFs, supercritical CO2 (scCO2) is the most widely used and has been shown to have great potential in particle engineering due to its favorable critical conditions (Tc = 31.1 °C, Pc = 7.38 MPa), nontoxicity, nonflammability, and lower price.23 So far, the most common precipitation processes using supercritical CO2 include the rapid expansion of supercritical solutions (RESS), particles from gas-saturated solutions or suspensions (PGSS), and gas or supercritical fluid antisolvents (GAS or SAS). In particular, the SAS process, due to the lower operational temperature, resultant particles thatare smaller in size as well as having more controlled morphology, and less or no solvent residue or impurities, has been widely used to prepare nanoparticles. In the SAS process, supercritical CO2 acts as an antisolvent and the particle size and particle size distribution of the resulting product can be controlled by adjusting the process parameters, including the concentration of the initial solution, temperature, pressure, and flow rates of a solution.24 In the present study, solution-enhanced dispersion by supercritical CO2 (SEDS) was used to prepare silk fibroin

1. INTRODUCTION Silk fibroin (SF) derived from Bombyx mori is a protein-based biomacromolecule composed of 5507 amino acids.1 This native biopolymer is made almost entirely of the amino acids glycine, alanine, and serine (-Gly-Ala-Gly-Ala-Gly-Ser-) leading to the formation of antiparallel β-pleated sheets in the fibers.2,3 Owing to its characteristic features such as biocompatibility and degradability, moderate moisture absorption and retention properties, plus adequate ultraviolet absorption properties, silk fibroin has been widely used in the food and cosmetics industry, as well as in biomedical fields in the form of films, three-dimensional scaffolds, hydrogels, electrospun fibers, and microspheres.4−8 Particularly, silk micro/nanoparticles have been demonstrated to have beneficial characteristics for drug delivery and controlled release because of their mechanical stability, mild processing conditions, high drug encapsulation efficiency, long-term sustained release, and slow degradability.9 Recently, study of the production of silk fibroin nanoparticles has become more and more important. Many investigations have emphasized the significance of size and revealed the advantages of nanoparticles over microspheres in controlled drug delivery systems.10 In general, nanoparticles possess better intracellular uptake ability and higher drug carrier capacity and feasibility through various routes of administration, including oral application and inhalation.11−14 The methods used to prepare silk fibroin micro/nanoparticles include mechanical milling,15 the water-in-oil emulsion solvent evaporation method,16 spray drying,17 self-assembly,18 laminar jet breakup,19 lipid templating20 and two-phase microfluidic flowfocusing devices.21 However, the major disadvantage of these methods is the high residual solvent content in the resultant protein powders. In addition, spray drying and other methods often produce particles that are not small enough for either pulmonary delivery or encapsulation into controlled-release © 2013 American Chemical Society

Received: Revised: Accepted: Published: 3752

July 18, 2012 February 18, 2013 February 21, 2013 February 21, 2013 dx.doi.org/10.1021/ie301907f | Ind. Eng. Chem. Res. 2013, 52, 3752−3761

Industrial & Engineering Chemistry Research

Article

from a CO2 cylinder was cooled down to around 0 °C by a cooler to ensure the liquefaction of the gas before it entered the pump, thus avoiding cavitations. Then a high-pressure meter pump was used to deliver liquefied CO2 to the high-pressure vessel. The precipitation temperature was controlled and maintained at a constant level by placing the high-pressure vessel in a gas bath during the experiment. A heat exchanger was used to preheat the liquefied CO2 to the desired operating temperature after it left the pump head. When the desired pressure and temperature that kept the CO2 in a supercritical state were reached, a steady flow of CO2 was maintained, and the system pressure was controlled by adjusting a downstream valve and monitored by a pressure gauge to keep the pressure constant. The silk fibroin solution dissolved in HFIP was then injected into the high-pressure vessel via a stainless steel coaxial nozzle simultaneously with the supercritical CO2 by use of an HPLC pump, and precipitation takes place. The silk fibroin solution was pumped through the inner part of the nozzle, while the CO2 went through the external route. Improved mixing between the two flows was obtained because of the geometry of the nozzle. When the spraying was finished, fresh CO2 was used to wash the products for about 30 min to remove the residual organic solvent from precipitated particles. During the process of washing, the system operating conditions were maintained as described before. After the washing process, the CO2 flow was stopped and the pressure of CO2 in the high-pressure vessel was slowly reduced to atmospheric pressure. The silk fibroin nanoparticle products were then collected on the filter and stored in a desiccator at room temperature in powder form for characterization. 2.2.3. Full Factorial Design. By analyzing the SEDS process, the four key variables were identified as follows: flow rate of silk fibroin solution, precipitation pressure, concentration of silk fibroin solution, and precipitation temperature. To investigate the influence and significance of the four variables in the SEDS process on morphology, particle size, and particle size distribution of silk fibroin nanoparticles, a 24 factorial experiment was designed and conducted, as shown in Table 1. According to the factorial design there were 16 experiments

nanoparticles. It is a modified SAS process, in which the polymer solution in organic solvent and supercritical fluid are sprayed together by use of a specially designed coaxial nozzle. A full factorial experiment was designed to perform a systemic investigation into the actual effects of process parameters, including flow rate of silk fibroin solution, precipitation pressure, concentration of silk fibroin solution, and precipitation temperature, on particle size (PS) and particle size distribution (PSD). Fourier transform infrared (FTIR) spectroscopy measurement and X-ray powder diffraction (XRPD) were used to study chemical and physical changes during the SEDS process. Furthermore, experimental and theoretical investigations of the nanoparticle formation mechanism are presented.

2. MATERIALS AND METHODS 2.1. Materials. Cocoons of Bombyx mori were purchased from Jiangsu Wujing China Eastern Silk Market Co. Ltd. (China). CO2 with a purity of 99.9% was supplied by Hong Kong Specialty Gases Co. Ltd. (Hong Kong). The solvent, 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP, 99.5%), was purchased from Dupont. All other compounds were of analytical purity. 2.2. Methods. 2.2.1. Preparation of Pure Silk Fibroin Raw Material. Cocoons of B. mori were degummed three times to remove sericin and other impurities enveloping silk fibroin (SF), in pressurized hot water at 120 °C for 60 min. Afterward, deionized water was used to wash the fibroin fibers and then they were dried at room temperature. Degummed SF fibers were dissolved in a solution of calcium chloride, water, and ethanol (CaCl2:water:ethanol = 1:8:2, molar ratio) for 6 h at 70 °C to obtain the SF solution. Then the solution was dialyzed in distilled water for 3 days to remove the neutral salts by use of semipermeable cellulose tubing (molecular weight cutoff 12 000−14 000) and obtain the pure silk fibroin solution. Subsequently, dry samples of pure silk fibroin were obtained by lyophilization with a freeze-dryer. 2.2.2. Preparation of Silk Fibroin Nanoparticles by the Solution-Enhanced Dispersion by Supercritical CO2 Process. The experimental apparatus of the SEDS process used for the formation of nanoparticles consists of three major sections: a CO2 supply system, an organic solution delivery system, and a high-pressure vessel.25 A schematic diagram of the SEDS process is shown in Figure 1. In the SEDS process, the CO2 fed

Table 1. Experimental Factors and Levels coded level symbols

factors

−1

+1

A B C D

flow rate of SF solution (mL·min−1) precipitation pressure (MPa) concn of SF solution (mg·mL−1) precipitation temp (°C)

0.5 10 0.5 35

1 20 1 45

(shown in Table 2). Analysis of variances was performed on the experimental data by use of MINITAB software version 15. 2.2.4. Surface Morphology, Particle Size, and Particle Size Distribution. The surface morphology of the samples was investigated by using a field emission scanning electron microscope (FE-SEM; JEOL JSM-6490). Before observation, the samples were directly adhered onto an aluminum stub with a thin self-adherent carbon film and then coated with a thin layer of gold. The particle size and particle size distribution of the samples dispersed in ethanol by sonication were analyzed by use of a laser diffraction particle size analyzer with a liquid module (LS 13320, Beckman Coulter).

Figure 1. Schematic diagram of apparatus for the SEDS process. 3753

dx.doi.org/10.1021/ie301907f | Ind. Eng. Chem. Res. 2013, 52, 3752−3761

Industrial & Engineering Chemistry Research

Article

Figure 2. FE-SEM photos of silk fibroin nanoparticles prepared by the SEDS process in different runs as shown in Table 2 (a, run 11; b, run 12; c, run 3; d, run 7; e, run 14; f, run 16; g, run 2; h, run 10).

Table 2. Experimental Design and Results of Full Factorial Design particle size and particle size distribution run

A

B

C

D

mean size (nm)

SD

D10 (nm)

D50 (nm)

D90 (nm)

span (D90− D10)/D50

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

−1 1 −1 1 −1 1 −1 1 −1 1 −1 1 −1 1 −1 1

−1 −1 1 1 −1 −1 1 1 −1 −1 1 1 −1 −1 1 1

−1 −1 −1 −1 1 1 1 1 −1 −1 −1 −1 1 1 1 1

−1 −1 −1 −1 −1 −1 −1 −1 1 1 1 1 1 1 1 1

64.9 76.0 52.5 70.2 78.5 90.4 76.1 81.6 66.5 80.5 54.8 72.6 81.4 102.3 77.8 87.5

0.5 1 0.9 0.8 2.5 1 0.9 0.6 0.5 0.5 0.4 1 1.5 2 0.8 1.4

55.9 57.5 52.7 64.4 56.9 75.9 64.6 64.8 53.8 68.9 52.1 68.7 61.2 71.1 46.5 63.3

62.4 74.8 50.6 68.5 78.2 88.9 74.2 81.2 65.8 80.5 54.3 72.5 80.8 100.5 76.9 86.7

79.7 92.7 68.9 91.8 98.4 132.8 102.5 108.6 79.5 109.2 71.6 98.4 109.7 137.4 94.2 114.5

0.38 0.47 0.32 0.40 0.53 0.64 0.51 0.54 0.39 0.50 0.36 0.41 0.60 0.66 0.62 0.59

2.2.5. Fourier Transform Infrared Spectroscopic Analysis. The samples were mixed with KBr and pressed into a thin tablet. FTIR spectra for the silk fibroin samples were obtained

on an FTIR Perkin-Elmer 1720 (Perkin-Elmer) in the transmission mode with the wavenumber ranging from 4000 to 400 cm−1. 3754

dx.doi.org/10.1021/ie301907f | Ind. Eng. Chem. Res. 2013, 52, 3752−3761

Industrial & Engineering Chemistry Research

Article

Figure 3. Normal plot of the standardized effect of the factors on particle size (a) and the main effects plot for particle size (b).

experimental design and results of 24 full factorial design experiments. Particle size and particle size distribution, defined as (D90 − D10)/D50, varied markedly under the conditions tested. From these figures and Table 2, it can be seen that the silk fibroin nanoparticles possess an irregular spherical morphology with a mean particle size from about 52.5 to 102.3 nm and particle size distribution (span) from 0.32 to 0.66. Various morphologies indicated that the process parameters in the SEDS process have a significant effect on the morphology, particle size, and size distribution. Furthermore, by adjusting the different process parameters, silk fibroin nanoparticles with a controlled particle size and size distribution can be obtained. 3.2. Influence of Process Parameters on Particle Size and Particle Size Distribution. The full factorial design can cover the main effects of the parameters including the flow rate of silk fibroin solution (A), precipitation pressure (B), concentration of silk fibroin solution (C), and precipitation

2.2.6. X-ray Powder Diffraction Analysis. XRPD was carried out on an X-ray diffractometer with Cu Kα (λ = 1.5405 Å) radiation (D8 Advance, Bruker AXS). The measurement was performed in a 2θ range of 5−45° with 0.02° step size and 10°·min−1 scan speed with a 2D detector at 40 kV and 40 mA.

3. RESULTS 3.1. Surface Morphology, Particle Size, and Particle Size Distribution of Silk Fibroin Nanoparticles. Figure 2 shows FE-SEM photos of silk fibroin nanoparticles prepared by the SEDS process in different runs as shown in Table 2. It can be seen that some of the silk fibroin nanoparticles had a tendency to aggregate together. This may be caused by the small particle size and surface activity of the high reunion.26 However, these aggregates can be separated when the powder is suspended in ethanol and dispersed by sonication (see Supporting Information, Figure S1). Table 2 presents the 3755

dx.doi.org/10.1021/ie301907f | Ind. Eng. Chem. Res. 2013, 52, 3752−3761

Industrial & Engineering Chemistry Research

Article

Figure 4. Normal plot of the standardized effect of the factors on particle size distribution (a) and the main effects plot for particle size distribution (b).

temperature (D) on particle size and particle size distribution of silk fibroin nanoparticles within the whole range of those selected. A smaller span defined as (D90 − D10)/D50 indicated a narrower particle size distribution.The quantitative data analysis for the results shown in Table 2 was carried out with Minitab software version 15 (estimated effects and coefficients for particle size and particle size distribution are shown in Tables S1 and S2 of Supporting Information). Figure 3 shows the normal plot of the standardized effect of the factors on particle size (a) and the main effects plot for particle size (b) . Figure 4 shows the normal plot of the standardized effect of the factors on particle size distribution (a) and the main effects plot for particle size distribution (b). As shown in Tables S1 and S2 (Supporting Information) and Figures 3a and 4a, A, B, C and D (P < 0.001) significantly affect the particle size (nanometers) and particle size distribution, and the order of importance of the factors for particle size can be

summarized as follows: C > A > B > D. This result was also demonstrated by the slopes of the lines shown in Figures 3b and 4b. Within the range of parameters studied, the particle size and particle size distribution increased with increasing concentration or flow rate of the solution and decreasing precipitation pressure. The influence of temperature alone on particle size did not appear obvious, and decreasing the temperature reduced the particle size of silk fibroin nanoparticles slightly. 3.3. Fourier Transform Infrared Spectroscopic Analysis. FTIR can reveal chemical changes by producing an infrared absorption spectrum. Protein materials show characteristic vibrational bands at 1630−1650 cm−1 for amide I, at 1540−1520 cm−1 for amide II, and at 1270−1230 cm−1 for amide III.27 Figure 5 shows the FTIR spectra of SF (before the SEDS process) and SF nanoparticle (after the SEDS process). It can be seen that these two kinds of samples possessed the 3756

dx.doi.org/10.1021/ie301907f | Ind. Eng. Chem. Res. 2013, 52, 3752−3761

Industrial & Engineering Chemistry Research

Article

of SF (before the SEDS process) and SF nanoparticles (after the SEDS process). On the XRPD curves, a broad peak at 2θ=20.4° and a sharp peak at 2θ = 12.8°, which are attributed to amorphous structure and silk I, respectively, indicate silk fibroin before the SEDS process was mainly in random coil and α-helix. After the SEDS process, a minor crystalline peak at 9.6° appears that denotes the silk II (β-sheet structure) of silk fibroin nanoparticles. It indicated a minor transformation from random coil and α-helix to β-sheet. However, the broad peak at 20.4° and a sharp peak at 2θ = 12.8° are still predominant. Therefore, silk fibroin nanparticle prepared by the SEDS process consisted of random coil, α-helix, and β-sheet forms with predominantly random coil and α-helix. In a word, there was no significant difference about the crystal phase of the samples before and after the SEDS process. These results also have confirmed those from FTIR analysis.

4. DISCUSSION The SAS process involves fluid dynamics, surface tension variations, mass transfer, vapor−liquid equilibria (VLE), and nucleation and growth mechanisms, which are complementary and, when taken together, can give a comprehensive description of the SAS process. However, the nanoparticle formation mechanism is still not readily understood. The 24 full factorial design experiments can investigate the effect of key process parameters systematically, including concentration, flow rate of solution, precipitation temperature, and pressure on particle size and particle size distribution. It will be very useful to study the nanoparticle formation mechanism. Figure 2 shows FE-SEM images of silk fibroin nanoparticles prepared by the SEDS process. As shown in these images, all silk fibroin nanoparticles possessed an irregular spherical shape. Therefore, it can be concluded that the resulting nanoparticles were generated from a gaseous miscible phase, not by droplets drying directly. Otherwise, the surface tension of the droplets will confer on them a perfectly spherical shape.31 Therefore, the nucleation and growth mechanism can be used to explain the generation of silk fibroin nanoparticles. However, at the moment that the solution was sprayed into the particle formation vessel, the solution and the scCO2 in the vessel were not in equilibrium. So within an extremely short initial period, the solution will exist in the form of primary droplets due to a liquid−liquid phase split. Therefore, the mutual mass transfer between scCO2 and the solution will take place until the concentration variance of the solution is completely dissipated.32 4.1. Effect of Flow Rate of Silk Fibroin Solution. In the SAS process, an increase in solution flow rate may cause competing effects.33 On one hand, the degree of mixing may be improved, resulting in a higher supersaturation, and thus smaller particles are expected. On the other hand, it also could reduce the CO2 mole fraction in the precipitator and decrease the mass-transfer rate of organic solvent out of the droplets, consequently inhibiting particle formation. In this study, an increase in the liquid flow rate of silk fibroin solution led to larger particle size and particle size distribution, supported by the morphology of silk fibroin nanoparticles shown in Figure 2a,b. The lower flow rate of the solution (0.5 mL/min for run 11 in Table 2) yielded smaller particles of 54.8 nm with a particle size distribution of 0.36, and the higher flow rate (1 mL/min for run 12 in Table 2) generated larger particles of 72.6 nm with a particle size distribution of 0.41. Therefore, in the experimental system, enhancement of mass

Figure 5. FTIR spectra of SF (before the SEDS process) and SF nanoparticles (after the SEDS process).

main characteristic peaks of amide I at 1654 cm−1, amide II at 1534 cm−1, and amide III at 1241 cm−1, assigned to α-helix or random coil conformation28 It indicated that SF (before the SEDS process) and SF nanoparticle (after the SEDS process) exhibited mainly α-helix or random coil conformation. On the curve of FTIR spectra of SF nanoparticles (after the SEDS process), two minor peaks have been found at 1621 cm−1 (amide I) and 1265 cm−1 (amide III), attributed to the β-sheet conformation of SF nanoparticles.28 Therefore, after the SEDS process, a minor transformation from α-helix or random coil conformation of SF to β-sheet conformation occurs. However, the main secondary structure of silk fibroin nanoparticles is still α-helix or random coil conformation. 3.4. X-ray Powder Diffraction Analysis. X-ray powder diffractometry (XRPD) is an important technique to determine the crystal phase of the samples. B. mori silk fibroin (SF) has crystalline and amorphous states in different conditions, and the crystalline state includes silk I (α-helix) and silk II (βsheet).29 On the XRPD curves, the main diffraction peaks of silk I were 12.2°, 19.7°, 24.7°, 28.2°, etc., and peaks of silk II were 9.1°, 18.9°, 20.7°, etc.30 Figure 6 shows the XRPD pattern

Figure 6. XRPD pattern of SF (before the SEDS process) and SF nanoparticles (after the SEDS process). 3757

dx.doi.org/10.1021/ie301907f | Ind. Eng. Chem. Res. 2013, 52, 3752−3761

Industrial & Engineering Chemistry Research

Article

distribution of silk fibroin nanoparticles from 102.3 nm and 0.66 to 87.5 nm and 0.59, respectively. This trend was also observed in surface morphology of silk fibroin nanoparticles as shown in Figure 2e,f. Jin et al.40 also reported similar results. An increase in pressure led to an increase in the CO2 density and mole fraction, thus enhancing the antisolvent effect of the CO2 and mass-transfer rate between scCO2 and the solution. Therefore, supersaturation within the liquid phase was reached more quickly and it prevented the crystals from growing and induced particle formation with smaller size and narrower size distribution.40 However, reducing precipitation pressure also could decrease solubility, which is inclined to generate higher maximum supersaturation, resulting in smaller particles with a narrow size distribution.41 Within the operating conditions used in the present study, the first effect proved to be predominant, owing to the postive effect of precipitation pressure on particle size and particle size distribution. Therefore, mass transfer between scCO2 and the solution exerted a more significant effect on the particle size and particle size distribution of silk fibroin nanoparticles in the SEDS process. 4.4. Effect of Precipitation Temperature. Precipitation temperature, like precipitation pressure, can affect the density of supercritical fluid and thus mass transfer between scCO2 and solution during precipitation. Especially near the critical point, a small change in temperature will cause considerable density changes.41 Figures 3 and 4 suggest that temperature exerts a comparatively negligible influence on particle size and particle size distribution, and decreasing the temperature can reduce the particle size and particle size distribution of silk fibroin nanoparticles. Runs 2 and 10 in Table 2 show that increasing the temperature from 35 to 45 °C can result in a slight increase in particle size and particle size distribution of silk fibroin nanoparticles from 76 nm and 0.47 to 80.5 nm and 0.50, respectively. The morphologies shown in Figure 2g,h demonstrate this trend. The density of supercritical fluid depended on the temperature and pressure parameters of the fluid; an increase in temperature caused a decrease in density of CO2, resulting in reduced mass-transfer rate between scCO2 and the solution, and thus slowing the achievement of supersaturation. Therefore, a larger particle with wider particle size distribution will form. Lee et al.42 also indicated a similar phenomenon. However, increasing the temperature could reduce the solubility and thus enhance the maximum supersaturation, so that smaller particles may be obtained. In the present study, the slight increase in particle size and particle size distribution with increasing temperature indicated the effect of mass transfer between CO2 and solvent superimposed on that of solubility, which also is driven by the former. 4.5. Silk Fibroin Nanoparticle Formation Mechanism. In summary, the nonperfect spherical morphology of silk fibroin nanoparticles prepared under the operating conditions of the SEDS process indicated that nanoparticles formed due to nucleation and growth in the gaseous miscible phase evolved from initial droplets. Analysis of the main effects of process parameters on particle size and particle size distribution suggested that mutual mass transfer between scCO2 and the solution was superimposed on the supersaturation and was the most important process parameter affecting nanoparticle formation. Lower concentration and flow rate of the solution, lower temperature, and higher pressure enhance the masstransfer rate between scCO2 and the solution and generate a

transfer between scCO2 and the solution with decreasing flow rate was observed as a dominant factor in relation to the flow rate effect. This result can be explained by the nucleation and growth mechanism induced by supersaturation, which is determined by the mass transfer between scCO2 and the solution. A higher flow rate of the solution could generate much solution in the precipitator within the same time, leading to an increase of the liquid film thickness and decreasing the antisolvent effect of CO2 and the mass-transfer rates of organic solvent out of the droplets, thus reducing the achievable supersaturation ratio.34 Lower supersaturation ratios engendered fewer nuclei, which in turn yielded larger particles. The micronization process will shift toward the growth process, and larger particles with a broad PSD are produced.34 Therefore, under the condition of higher flow rate of the solution, lower mass transfer between scCO2 and the solution induced abatement of supersaturation and caused the final particles to become larger and nonuniform. Similar experimental results have been observed by other authors.35,36 Obviously, this can demonstrate that the mass transfer between scCO2 and solution plays a crucial role in silk fibroin nanoparticle formation. 4.2. Effect of Concentration of Silk Fibroin Solution. The initial concentration of the solution could have two opposite effects on particle size and particle size distribution.33 According to the classical nucleation and growth theory, higher concentration can induce higher supersaturation, which is inclined to decrease the particle size and particle size distribution. However, a lower concentration with lower molecular number of silk fibroin would reduce the viscosity and surface tension of the liquid solution at the moment that the solution is sprayed into the particle formation vessel37 and therefore produce smaller initial droplets, which could enhance mass transfer between scCO2 and the solution. Thus the supersaturation would occur more rapidly. This obviously favors the formation of smaller homogeneous nuclei and size distribution. Therefore, nucleation is the prevailing mechanism, producing smaller particles. In the present study, Figures 3 and 4 show that the concentration of the solution has the greatest significant effect on particle size and particle size distribution. The higher the concentration of the solution, the larger the particles and wider the particle size distribution. When the concentration was raised from 0.5% to 1% as indicated in Figure 2c,d, the particle size and particle size distribution of silk fibroin nanoparticles increased respectively from 52.5 nm and 0.32 for run 3 to 76.1 nm and 0.51 for run 7 in Table 2. A similar positive effect of concentration on particle size and particle size distribution has been reported by many researchers.38,39 Therefore, this result indicated mass transfer between scCO2 and solution superimposed on the supersaturation and that it was also the single most important process parameter affecting nanoparticle formation. Otherwise, the higher supersaturation induced by higher concentration will result in smaller particle size and particle size distribution. The analysis thus demonstrated the effect of the flow rate of a solution. 4.3. Effect of Precipitation Pressure. Precipitation pressure also influences particle formation in the supercritical CO2 antisolvent technique. The results shown in Figures 3 and 4 indicate that particle size and particle size distribution were reduced by increasing the precipitation pressure. Runs 14 and 16 in Table 2 show that increasing the pressure from 10 to 20 MPa can lead to a decrease in particle size and particle size 3758

dx.doi.org/10.1021/ie301907f | Ind. Eng. Chem. Res. 2013, 52, 3752−3761

Industrial & Engineering Chemistry Research

Article

helix or random coil conformations had been transformed into water-insoluble β-sheet crystalline structure of silk fibroin nanoparticles. However, the main secondary structure of silk fibroin nanoparticles was still water-soluble α-helix and random coil conformation. Generation of silk fibroin particles have attracted much attention in recent years. The secondary structure of silk fibroin particles prepared by Srisuwan et al.16 via a water-in-oil emulsion solvent evaporation method is similar to that of silk fibroin nanoparticles fabricated by the SEDS process. But emusification could result in high residual solvent content in the resultant particles and cause toxicity problems. Laminar jet break-up,1 lipid templating,20 and a two-phase microfluidic flow-focusing device have also been developed to prepare water-soluble silk fibroin particles. Methanol or ethanol treatment or exposure to water vapor have been applied in these methods to induce the formation of a β-sheet structure and result in water insolubility. However, the resultant silk fibroin particles with large particle size also limit their application. Silk fibroin particles prepared by mechanical milling15 and self-assembly18 predominantly possess β-sheets. These methods avoid using organic solvent to induce water insolubility; however, the particle size is not small enough. Espeically for mechanical milling, no drug or active substance can be encapsulated in the preparation of silk fibroin particles. Besides, Yeo et al.17 prepared a silk fibroin microsphere 2 ± 10 μm in size using a spray dryer. The structure transition from the random coil or silk I to β-sheet structure occurs during spray drying. However, the high temperature in this process could result in denaturation of active substances. Therefore, compared to the methods above, the SEDS process displays many advantages in the generation of silk fibroin nanoparticles, such as mild experimental conditions and resultant particles with small particle size and particle size distribution. Furthermore, there is no significant physical and chemical change of silk fibroin before and after the SEDS process.

higher degree of supersaturation. Therefore, a large number of smaller nuclei form. Meanwhile, the rate of nucleation and growth process will become higher, which will result in a lower degree of growth process and thus a smaller particle size distribution. Based on the analysis above, a possible silk fibroin nanoparticle formation mechanism for the SEDS process can be proposed. Figure 7 shows a schematic diagram of this

Figure 7. Schematic diagram of silk fibroin nanoparticle formation mechanism by the SEDS process.

possible mechanism. In the process of SEDS, at the moment that the solution is sprayed into the particle formation vessel through a coaxial nozzle, the solution and the scCO2 in the vessel are not in equilibrium. So within an extremely short initial period, the solution will exist in the form of droplets that form as a result of the formation of a liquid−liquid phase split. Then the mutual mass transfer between the solution and scCO2 makes the droplets expand rapidly and the strength of the solvent reduce significantly. Thus, the solvent-rich phase and CO2-rich phase are formed. During the addition of polymer solution into scCO2, very small amounts of solute will be dissolved in scCO2 with the solvent acting as a cosolvent until saturation of the polymer in the mixture of scCO2 and solvent is reached. Continued feeding of the solution into scCO2 results in crossing over the equilibrium boundary and supersaturation of the polymer in the mixture of scCO2 and solvent. Subsequently, a phase transition takes place in the solvent-rich phase, leading to the formation of polymer nuclei and a polymer-rich phase. With the continued mass transfer between droplets and scCO2, the nuclei formed in the supersaturated polymer solution quickly grow into larger polymer nanoparticles. After precipitation, the solvent/CO2 phase can be flushed out to obtain pure polymer nanoparticles, and remaining traces of the solvent are finally removed from the precipitate by a continuous flow of supercritical carbon dioxide. 4.6. Physical and Chemical Properties of Silk Fibroin Nanoparticles. In our study, the silk fibroin nanoparticles prepared by the SEDS process possessed a spherical morphology with a small mean particle size from about 52.5 to 102.3 nm and a narrow particle size distribution (span) from 0.32 to 0.66. The results of FTIR and XRPD analysis indicated that silk fibroin before the SEDS process possessed random coil and α-helix structure mainly. After the SEDS process, a few α-

5. CONCLUSIONS Silk fibroin (SF) nanoparticles were successfully prepared via solution-enhanced dispersion by supercritical CO2 (SEDS). Investigation with a 24 full factorial design indicated that increasing the concentration and flow rate of the silk fibroin solution and precipitation temperature increased the particle size and particle size distribution of silk fibroin nanoparticles, while reducing the precipitation pressure decreased the particle size and particle size distribution. The nanoparticle formation mechanism can be explained by the formation and growth of silk fibroin nuclei in miscible mixtures of the solution and supercritical CO2 (scCO2) due to supersaturation, which was driven and determined by mass transfer between scCO2 and solution droplets generated by the liquid−liquid phase split. Furthermore, characterization with FTIR and XRPD indicated that the physical and chemical structure of silk fibroin underwent no significant change before and after the SEDS process. Thus the SEDS process could be a promising technique to prepare silk fibroin nanoparticles applied in textiles, cosmetics, and biomedical fields. 3759

dx.doi.org/10.1021/ie301907f | Ind. Eng. Chem. Res. 2013, 52, 3752−3761

Industrial & Engineering Chemistry Research

■

Article

(12) Schafer, V.; Von Briesen, H.; Andreesen, R.; Steffan, A. M.; Royer, C.; Troster, S.; Kreuter, J.; Rubsamen-Waigmann, H. Phagocytosis of nanoparticles by human immunodeficiency virus (HIV) infected macrophages: a possibility for antiviral drug targeting. Pharm. Res. 1992, 9, 541. (13) Sundar, S.; Kundu, J.; Kundu Subhas, C. Biopolymeric nanoparticles. Sci. Technol. Adv. Mater. 2010, 11, No. 014104. (14) Berthold, A.; Cremer, K.; Kreuter, J. Preparation and characterization of chitosan microspheres as drug carriers for prednisolone sodium phosphate as model for anti inflammatory drugs. J. Controlled Release 1996, 39, 17. (15) Rajkhowa, R.; Wang, L.; Kanwar, J.; Wang, X. Fabrication of ultrafine powder from Eri silk through attritor and jet milling. Powder Technol. 2009, 191, 155. (16) Srisuwan, Y.; Srihanam, P.; Baimark, Y. Preparation of silk fibroin microspheres and its application to protein adsorption. J. Macromol. Sci., Part A 2009, 46, 521. (17) Yeo, J. H.; Lee, K. G.; Lee, Y. W.; Kim, S. Y. Simple preparation and characteristics of silk fibroin microsphere. Eur. Polym. J. 2003, 39, 1195. (18) Cao, Z. B.; Chen, X.; Yao, J. R.; Huang, L.; Shao, Z. Z. The preparation of regenerated silk fibroin microspheres. Soft Matter 2007, 3, 910. (19) Wenk, E.; Wandrey, A. J.; Merkle, H. P.; Meinel, L. Silk fibroin spheres as a platform for controlled drug delivery. J. Controlled Release 2008, 132, 26. (20) Wang, X.; Wenk, E.; Matsumoto, A.; Meinel, L.; Li, C.; Kaplan, D. L. Silk microspheres for encapsulation and controlled release. J. Controlled Release 2007, 117, 360. (21) Breslauer, D. N.; Muller, S. J.; Lee, L. P. Generation of monodisperse silk microspheres prepared with microfluidics. Biomacromolecules 2010, 11, 643. (22) Wang, Q.; Guan, Y. X.; Yao, S. J.; Zhu, Z. Q. Controllable preparation and formation mechanism of BSA microparticles using supercritical assisted atomization with an enhanced mixer. J. Supercrit. Fluids 2011, 56, 97. (23) Chen, A. Z.; Zhao, Z.; Wang, S. B.; Li, Y.; Zhao, C.; Liu, Y. G. A continuous RESS process to prepare PLA-PEG-PLA microparticles. J. Supercrit. Fluids 2011, 59, 92. (24) Reverchon, E.; De Marco, I.; Adami, R.; Caputo, G. Expanded micro-particles by supercritical antisolvent precipitation: Interpretation of results. J. Supercrit. Fluids 2008, 44, 98. (25) Zhao, Z.; Chen, A. Z.; Li, Y.; Hu, J. Y.; Liu, X.; Li, J. S.; Zhang, Y.; Li, G.; Zheng, Z. J. Fabrication of silk fibroin nanoparticles for controlled drug delivery. J. Nanopart. Res. 2012, 14, 736. (26) Zhao, X. H.; Chen, X. Q.; Zu, Y. G.; Jiang, R.; Zhao, D. M. Recrystallization and micronization of taxol using the supercritical antisolvent (SAS) process. Ind. Eng. Chem. Res. 2012, 51, 9591. (27) Ha, S. W.; Tonelli, A. E.; Hudson, S. M. Structural studies of Bombyx mori silk fibroin during regeneration from solutions and wet fiber spinning. Biomacromolecules 2005, 6, 1722. (28) Zhang, H.; Li, L. L.; Dai, F. Y.; Zhang, H. H.; Ni, B.; Zhou, W.; Yang, X.; Wu, Y. Z. Preparation and characterization of silk fibroin as a biomaterial with potential for drug delivery. J. Transl. Med. 2012, 10, 117. (29) He, J.; Wang, Y.; Cui, S.; Gao, Y.; Wang, S. Structure and properties of silk fibroin/carboxymethyl chitosan blend films. Polym. Bull. 2010, 65, 395. (30) Magoshi, J.; Magoshi, Y. Physical properties and structure of silk. Physical properties and structure of silk. J. Polym. Sci.: Polym. Phys. 1977, 15, 167. (31) Reverchon, E.; De Marco, I. Mechanisms controlling supercritical antisolvent precipitates morphology. Chem. Eng. J. 2011, 169, 358. (32) He, W. Z.; Jiang, Z. H.; Suo, Q. L.; Li, G. M. Mechanism of dispersing an active component into a polymeric carrier by the SEDSPA process. J. Mater. Sci. 2010, 45, 467. (33) Tenorio, A.; Gordillo, M. D.; Pereyra, C. M.; Martínez de la Ossa, E. J. Screening design of experiment applied to supercritical

ASSOCIATED CONTENT

S Supporting Information *

One figure with FE-SEM photo of separated silk fibroin nanoparticles dispersed in ethanol by sonication, and two tables with estimated effects and coefficients for particle size and span. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

* Telephone: +852 2766 6479. Fax: +852 2773 1432. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Hong Kong Innovation and Technology Commission and Hong Kong Research Institute of Textile and Apparel for providing funding support to this research through Projects ITP/001/07TP and ITP/031/08TP and the Hong Kong Research Grant Council and the Hong Kong Polytechnic University through Projects PolyU5242/09E, G-YX1M, JBB6Q, and GU942. Also, we acknowledge with thanks the support of Guangdong Provincial Department of Science and Technology through the Guangdong−Hong Kong International Textile Bioengineering Joint Research Center with project code 2011B050300023, as well as the sponsorship from Hong Kong Jockey Club Sports Medicine and Health Science Center.

■

REFERENCES

(1) Zhou, C. Z.; Confalonieri, F.; Medina, N.; Zivanovic, Y.; Esnault, C.; Yang, T.; Jacquet, M.; Janin, J.; Duguet, M.; Perasso, R.; Li, Z. G. Fine organization of Bombyx mori fibroin heavy chain gene. Nucleic Acids Res. 2000, 28, 2413. (2) Kundu, J.; Chung, Y. I.; Kim, Y. H.; Tae, G.; Kundu, S. C. Silk fibroin nanoparticles for cellular uptake and control release. Int. J. Pharm. 2010, 388, 242. (3) Kim, U. J.; Park, J.; Kim, H. J.; Wada, M.; Kaplan, D. L. Three dimensional aqueous-derived biomaterial scaffolds from silk fibroin. Biomaterials 2005, 26, 2775. (4) Kundu, J.; Patra, C.; Kundu, S. C. Design, fabrication and characterization of silk fibroin-HPMC-PEG blended films as vehicle for transmucosal delivery. Mater. Sci. Eng. C 2008, 28, 1376. (5) Nazarov, R.; Jin, H. J.; Kaplan, D. L. Porous 3D scaffolds from regenerated silk fibroin. Biomacromolecules 2004, 5, 718. (6) Kim, U. J.; Park, J.; Li, C.; Jin, H. J.; Valluzzi, R.; Kaplan, D. L. Structure and properties of silk hydrogels. Biomacromolecules 2004, 5, 786. (7) Zhang, F.; Zuo, B.; Bai, L. Study on the structure of SF fiber mats electrospun with HFIP and FA and cells behavior. J. Mater. Sci. 2009, 44, 5682. (8) Lammel, A. S.; Hu, X.; Park, S. H.; Kaplan, D. L.; Scheibel, T. R. Controlling silk fibroin particle features for drug delivery. Biomaterials 2010, 31, 4583. (9) Wenk, E.; Wandrey, A. J.; Merkle, H. P.; Meinel, L. Silk fibroin spheres as a platform for controlled drug delivery. J. Controlled Release 2008, 132, 26. (10) Meclean, S.; Processer, E.; O’Malley, D.; Clark, N.; Ramtoola, Z.; Brayden, D. Binding and uptake of biodegradable poly-DL-lactide micro- and nanoparticles in intestinal epithelia. Eur. J. Pharm. Sci. 1998, 6, 153. (11) Panyam, J.; Labhasetwar, V. Biodegradable nanoparticles for drug and gene delivery to cells and tissue. Adv. Drug Delivery Rev. 2002, 55, 329. 3760

dx.doi.org/10.1021/ie301907f | Ind. Eng. Chem. Res. 2013, 52, 3752−3761

Industrial & Engineering Chemistry Research

Article

antisolvent precipitation of amoxicillin. J. Supercrit. Fluids 2008, 44, 230. (34) Wang, Q.; Guan, Y. X.; Yao, S. J.; Zhu, Z. Q. Microparticle formation of sodium cellulose sulfate using supercritical fluid assisted atomization introduced by hydrodynamic cavitation mixer. Chem. Eng. J. 2010, 159, 220. (35) Reverchon, E.; De Marco, I. Supercritical antisolvent precipitation of cephalosporins. Powder Technol. 2006, 164, 139. (36) Reverchon, E.; Della Porta, G.; De Marco, I. Role of phase behavior and atomization in the supercritical antisolvent precipitation. Ind. Eng. Chem. Res. 2003, 42, 6406. (37) Braeuer, A.; Dowy, S.; Torino, E.; Rossmann, M.; Luther, S. K.; Schluecker, E.; Leipertz, A.; Reverchon, E. Analysis of the supercritical antisolvent mechanisms governing particles precipitation and morphology by in situ laser scattering techniques. Chem. Eng. J. 2011, 173, 258. (38) Montes, A.; Tenorio, A.; Gordillo, M. D.; Pereyra, C. M.; de la Ossa, E. J. M. Supercritical antisolvent precipitation of ampicillin in complete miscibility conditions. Ind. Eng. Chem. Res. 2011, 50, 2343. (39) Wang, Q.; Guan, Y. X.; Yao, S. J.; Zhu, Z. Q. Controllable preparation and formation mechanism of BSA microparticles using supercritical assisted atomization with an enhanced mixer. J. Supercrit. Fluids 2011, 56, 97. (40) Jin, H. Y.; Hemingway, M.; Xia, F.; Li, N. S.; Zhao, Y. P. Production of β-carotene nanoparticles by the solution enhanced dispersion with enhanced mass transfer by ultrasound in supercritical CO2 (SEDS-EM). Ind. Eng. Chem. Res. 2011, 50, 13475. (41) Kalani, M.; Yunus, R. Application of supercritical antisolvent method in drug encapsulation: a review. Int. J. Nanomed. 2011, 6, 1429. (42) Lee, B. M.; Kim, S. J.; Lee, B. C.; Kim, H. S.; Kim, H.; Lee, Y. W. Preparation of micronized β-HMX using supercritical carbon dioxide as antisolvent. Ind. Eng. Chem. Res. 2011, 50, 9107.

3761

dx.doi.org/10.1021/ie301907f | Ind. Eng. Chem. Res. 2013, 52, 3752−3761