(172c) Process Control and Characterization of Porous Structure of Pharmaceutical Microparticles Prepared by the W/O/W Double Emulsion Technique

Proposal to present 5^th World Congress on Particle Technology

Introduction Injectables are a widely used pharmaceutical delivery form, especially for proteins, peptides or other water soluble drugs. These therapeutics have a relative rapid clearance from the bloodstream. If a treatment over a longer time period is necessary, the patients require multiple injections in order to ensure a certain drug level. In such cases a long acting, controlled release of a drug substance is highly desirable. This can be achieved by embedding the drug substance into a polymer matrix made out of biodegradable polymers, e.g. Poly(lactide-co-glycolide). To encapsulate these water-soluble drug substances the water-in-oil-in-water (W/O/W) double emulsion microparticle preparation process has been established during the past decade [1].

Due to this two step emulsification process the final microparticles are highly porous and the release of drug substance is strongly dependent on this porous structure. The critical pore forming process step is the preparation of the primary W/O emulsion. Thus, in order to control drug product quality, drug release profile and batch to batch reproducibility, there is a need for on-line monitoring of the W/O emulsification process. This holds especially during development, were scale up experiments from small lab scale to production scale are performed. For this purpose Ultrasound Extinction spectroscopy (USE) was evaluated [2]. The method was used to monitor scale up experiments from 0.5 l lab scale to 5 l pilot scale. Beside evaluation of a suitable droplet size on-line analysis method, the experiments aimed to compare the emulsification properties of different dispersion devices. Beside process control there is also a need for suitable analytical methods to characterize the porosity of the final microparticles. So far scanning electron microscopy (SEM) on fractured microparticles was used to get a qualitative impression about morphology and size of the pores. However this method is very time consuming and only few particles can be analyzed, thus representative sampling may cause problems. Mercury porosimetry is a well established analytical method to measure powder porosity. However in the case of the microparticles the main assumption i.e. cylindrical pores, is not fulfilled. This makes the applicability of this method questionable. As alternative method nitrogen gas sorption may be applied. But the typical pore size is in the range up to tens of micrometer. Therefore this method is also not applicable for a quantitative pore size analysis. [3]

Nevertheless microparticles with distinct porosities were manufactured and analyzed with these two classical methods. It turned out that the total pore volume can be measured by mercury intrusion. In addition nitrogen gas adsorption according to BET is a sensitive method for the analysis of the specific surface area of the microparticles. Furthermore a general combination of these data, together with particle size distribution and true density values, is suggested which allows a quantitative evaluation of a mean pore diameter for spherical particles with spherical pores. The pore size data obtained in this manner for different microparticles are compared with SEM data and critical points of this evaluation are discussed.  

Results and Discussion Part I: On-line monitoring of the W/O emulsion process The figures 1A and 1B show USE on-line monitoring data of a typical lab-scale W/O emulsion process over a time period of two hours. In this experiment a gear pump was used as dispersing device. As indicated by the plots the gear pump dispersed the water phase down to an endpoint within the first 20 min. In the following the droplet size remained almost constant, whereas the attenuation spectra increased continuously. Similar data were also obtained using ultra turrax and ultrasound processor as alternative dispersing devices. In addition the application of the method on scale-up experiments from 500ml to 5l will be presented. Figure 1A. On-line monitoring of W/O emulsion process using an USE apparatus, Sympatec OPUS. The emulsion was prepared similarly as described in [1]. The attenuation spectra were recorded in the range of 0.2 MHz ? 40 MHz. 5 characteristic frequencies are plotted in this graph.  

Figure 1B. Same process data as shown in figure 1A. Here characteristic particle size values (x16, x50, x84) are shown. The droplet size distributions were obtained by transformation of the USE attenuation spectra according to an emulsion model.  

It could be shown that the emulsification intensity can be varied in a wide range using different dispersing devices, whereby each in turn is adjustable within a certain range. The data indicate that emulsification yields stable minimum size of droplets. However the attenuation spectra revealed that a final equilibrium state could not be achieved. Thus on-line monitoring of the process is necessary, not only for scale-up purposes but also for process steering, in order to control batch to batch reproducibility and quality.  

Part II: Characterization of the porous structure of W/O/W microparticles The final microparticles are of spherical shape with an outer diameter of about 100 µm. The microparticles are highly porous. The diameter of the pores is in the range of 1 µm up to 20 µm or even more, depending on process conditions. Figure 2 shows SEM pictures of a typical sample.  

Figure 2. SEM pictures of spherical microparticles (top) and of a fracture (buttom) showing the inner porous structure of several individual microparticles.  

Microparticles of four different porosity classes were prepared and analyzed: (1) microparticles mainly small pores, (2) with small as well as large pores, thus exhibiting a broad pore size distribution, (3) with mainly large pores and (4) with a dense structure.

Mercury intrusion data of these samples are shown in figure 3. The low pressure range can be interpreted as inter-particle volume. The high pressure range above 1000 kPa for sample (1), (2) and (4) is characteristic for the intra-particle volume. As will be discussed in the paper, this clear separation of the two regions enables to calculate the total inner pore volume of the microparticles.

The samples were also investigated by nitrogen gas sorption. Ad- and desorption isotherms were recorded over the full p/p0 range, but could not be evaluated in terms of porosity, as already expected from the size of the pores. However the specific surface area according to BET could be obtained.

Figure 3. Mercury intrusion data of the different microparticle samples, (1) (2) (3) (4) see text. Due to the horizontal course between low and high pressure region, the inter- and intra-particle volume can be well separated.  

The values of the four samples were in the range of 0.1 m²/g to 2.3 m²/g and the different particle qualities could be well distinguished. Thus specific surface area measurements provide a second sensitive method to characterize these particles. In addition particle size distributions of the samples were obtained by laser light diffraction (LLD) measurements. The particle size distributions were similar for all 4 samples, with x50 values in the range of 80 µm ? 120 µm. Furthermore the solid state density, or true density of these samples was measured by helium pycnometry, yielding 1.4 g/cm³ for all four samples.

All these methods together provide a sound basis for a quantitative characterization of the size and structure of microparticles. However there is no single method which can provide a measure for the pivotal characteristic of these particles, the pore size. This can be achieved by combining the specific surface area data obtained by nitrogen adsorption and particle size analysis, together with the specific intra-particle volume and the solid state density. Assuming spherical particles with spherical pores the pore size can be evaluated according to the following equation:

with

Applying this equation to the data set obtained yielded the following pore sizes of the microparticles: 11 µm for (1), 12 µm for (2) and 50 µm for (3). For sample (4) no meaningful number could be given due to the very low BET specific surface area of this sample. These values could be confirmed by SEM pictures of particle fractures. In the paper the equation itself and some critical aspects of this kind of evaluation will be discussed. Despite the initial doubts about the applicability of mercury porosimetry and BET gas adsorption, both methods turned out to be highly sensitive to the porous structure of microparticles. The pore size values obtained by the proposed combination of the techniques may not only be used for quality control purposes. It may also help to get a better understanding of the relation between pore size and drug release profile.  

References 1.        Lambert, O., Nagele, O., Loux, V., Bonny, J.-D., Marchal-Heussler, L., (2000). Poly(ethylene carbonate) microspheres: Manufacturing process and internal structure characterization. Journal of Controlled Release, 67, 89-99.

2.        Riebel U., F. Löffler, F., (1989). The fundamentals of particle size analysis by means of ultrasonic spectrometry. Part. Part. Syst. Charact., 6, 135-143.

3.        Lowell, S., Shields, J.E., (1991). Powder surface area and porosity. 3^rd edition, Chapman and Hall, 1991.