The advent of rapid and reliable measurement technologies, together with the FDA’s PAT (Process Analytical Technologies) initiative, has increased the use of particle shape analysis within the pharmaceutical industry. Particle shape, like particle size which is routinely measured and controlled, can directly influence product performance and its measurement can lead to improved process and product understanding. Here we consider the importance of particle shape measurement for the pharmaceutical industry, with reference to the aims of the PAT initiative, and highlight the modern image analysis techniques available for sensitive size and shape characterization.
Why measure particle shape?
Often, manufacturers producing a particulate product need to identify and understand the differences between batches, either for product development reasons or for quality control purposes. For some applications particle size analysis generates enough data for sample differences to be fully rationalized, but for applications where samples are very close in size, measurement of subtle variations in shape may be necessary.
Figure 1 shows two different samples. The particle size distributions for each material could be the same, but they are clearly not identical. It is likely that these two materials would behave differently during processing, or in their final product form. For example, their flow and abrasion characteristics would be dramatically different. Particle size data alone would not allow differentiation between them.
Fig. 1: Two different samples could be reported as identical using a size-only distribution.
The FDA’s PAT initiative, an effort to improve cGMP by providing a regulatory framework for the introduction of new manufacturing technologies for the pharmaceutical industry, is ultimately designed to improve process control in the sector. Improved process control delivers greater efficiency, less waste and lower production costs. It will therefore allow the industry to respond more effectively to environmental and economic challenges.
Currently, many manufacturing operations are based on time-defined endpoints; for example ‘blend for 10 minutes’ or ‘mill for 1 hour’. The spirit of the PAT initiative is to move away from this approach, to one where endpoint is defined in relation to a property that is closely linked to product quality - granule size, morphic form or blend uniformity for example. Material with the desired properties is then produced more consistently and waste is minimized. This approach requires identification of an appropriate variable, with effective monitoring and control of the selected parameter.
Particle characterization using image analysis
Particle shape and size data can be generated using automated image analysis techniques, complementing both microscopy and laser diffraction for particle characterization. In contrast to manual microscopy, image analysis generates statistically relevant data with no subjective bias, allowing shape, and its effects, to be studied systematically. Image analysis generates number-based distributions and is therefore extremely sensitive to the presence of fines or small numbers of foreign particles. In addition, individual particle images are recorded, allowing visual detection and verification of agglomerates or contaminants.
Image analysis procedures involve the capture of images using transmitted or reflected light, a lens system and a CCD. Movement between the sample and the magnification lens allows scanning of a large number of particles for the production of statistically relevant data; typically several thousand particles are measured per minute. Multiple shape parameters are calculated for each individual particle and collated into distributions with all the associated distribution parameters.
Particle orientation is critically important for effective characterization of particle shape by image analysis. Figure 2, which shows an analysis of a sample of monodisperse needle-shaped particles, clearly illustrates the problem associated with random orientation. The shape and particle size data produced shows a polydisperse sample. The bank of images illustrates why. The camera and software are seeing a selection of different 2D views of similar particles – the random orientation is hiding the genuine primary morphology of the sample.
Fig. 2: Shape analysis of monodisperse needle shaped particles.
Consistent orientation is critical for the identification of real morphological differences. Particles may be presented showing their largest surface area, their smallest surface area or something in between. Which area is analyzed is less important than the consistency of presentation. However, as the largest area orientation is more closely correlated with surface area and volume-based data - and easier to achieve - this approach tends to be adopted.
Defining particle shape
Various different aspects of particle shape are of interest and a range of descriptors has been devised to allow particle shape to be quantifiably described. No single shape descriptor is suitable for all applications. The following three parameters, which are all normalised (defined to have values lying in the range 0 – 1) are frequently used to quantify different aspects of particle shape.
Elongation provides an indication of the length/width ratio of the particle and is defined as (1-[width/length]). Shapes symmetrical in all axes, such as circles or squares, will tend to have an elongation close to 0 whereas needle-shaped particles will have values closer to 1. Elongation is more an indication of overall form than surface roughness (see figure 3) - a smooth ellipse has a similar elongation to a ‘spiky’ ellipse of similar aspect ratio.
Fig. 3: Elongation.
Convexity is a measurement of the surface roughness of a particle and is calculated by dividing the particle area by a ‘total area’, best visualized as the area enclosed by an imaginary elastic band placed around the particle. A smooth shape, regardless of form, has a convexity of 1 while a very ‘spiky’ or irregular object has a convexity closer to 0 (see figure 4).
Fig. 4 : Convexity.
Circularity is a measurement of the ratio of the actual perimeter of a particle to the perimeter of a circle of the same area. A perfect circle has a circularity of 1 while a very ‘spiky’ or irregular object has a circularity closer to 0. Intuitively, circularity is a measure of irregularity or the difference from a perfect circle. Figure 5 shows how circularity is sensitive to both overall form (like elongation) and surface roughness (like convexity). This shape factor is particularly useful for applications where perfectly spherical particles are the desired end product.
Fig. 5 : Circularity.
A further parameter frequently used in particle characterization is:
Circle equivalent diameter
Circle equivalent diameter is calculated by measuring the area of a 2D image of a particle and back-calculating the diameter of a circle with the same area. It is one of many equivalent values used to define particle size and is calculated easily from image analysis data. Circle equivalent diameter calculation depends upon which 2D view is captured and hence may not be directly comparable with alternative particle size measuring techniques, particularly if the particles are not spherical.
Practical example of sensitivity to shape
The following exemplifies the sensitivity of one pharmaceutical process to particle shape. One of four batches of a pharmaceutical excipient was continuously failing at the tabletting stage of a manufacturing process. This was proving to be highly expensive since the tabletting process was at the very end of the manufacturing process where all the value has been locked into the product.
The tablet producer wanted some way of identifying the failed batch much earlier – ideally as a raw material. Traditional microscopy or ensemble sizing methods could not distinguish between the four batches being used.
Automated image analysis was used to evaluate the average convexity of each of the four batches. Convexity is a measure of surface roughness or ‘spikeyness’ of the particle surface and the failed batch was found to consistently exhibit a lower average convexity than the other three good batches (Figure 6).
The need for higher quality, higher sensitivity analytical techniques to increase process understanding within the pharmaceutical industry has been highlighted through the PAT initiative. Image analysis, an increasingly accessible option thanks to advances in PC processing power and digital camera technology, is particularly suited to analyzing size and shape and is a valuable tool for the sector. With particle shape and size data readily available it becomes possible to more effectively define process end-point and rationalize differences in the behaviour of different batches. In this way image analysis technology is delivering significant improvements in both process efficiency and product quality