Print page

Email page

Scaling Up and Controlling Crystallization

By Michael Dunne, ITOMS Ltd

PharmaManufacturing.com

Keywords: crystallization, Focused Beam Reflectance Measurement, FBRM, process tomography, Michael Dunne, ITOMS Ltd, François Ricard, GlaxoSmithKline and GSK

Process Tomography can offer insights into the onset of crystallization, allowing for improved process control.

Crystallization is a key pharmaceutical manufacturing operation. While the target is always to grow the desired crystals within a narrow size distribution, the nature of the crystallization process can make it difficult to achieve this level of control. One particular challenge is scaling up production to achieve the ideal particle size (which influences the final drug’s performance in the patient) and the ideal crystalline form (which can drastically affect biological performance).

Figure 1. ERT electrodes around a 7-liter vessel in eight planes, 16 per plane

Achieving perfect crystals involves perfect mixing and the control of concentration and temperature. It isn’t easy. “Hydrodynamics are simply not scaleable. There’s a tendency to assume that things are perfectly mixed when they are not,” observes Steve Stanley, technology manager for Nexia Solutions which is shortly to form the core of the U.K. National Nuclear Laboratory.

A key to achieving ideal crystals is the ability to measure the formation of crystals at each stage of manufacture. One popular tool for keeping tabs on crystallisation is Focused Beam Reflectance Measurement (FBRM): in this process, laser light hits tiny crystals as they grow and is reflected and measured. One limitation of FBRM is that it only notices particles above a certain size and so does not detect the onset of crystallization.

An alternative technology – Electrical Resistance Tomography (ERT) – offers some advantages. Developed by Manchester, U.K.-based Industrial Tomography Systems, it detects changes in conductivity and ion concentrations, both of which fall during crystallization. Thus, ERT can detect the onset of crystallization.

Gathering Data From the Entire Vessel

The technology can provide data throughout the vessel rather than at a specific point. For example, when electrodes are arranged around the circumference of the vessel – in Figure 1 there are eight planes with 16 electrodes per plane – ERT can monitor conductivity at more than 200 points at 20 to 40 times a second. At this speed, it e ectively builds a clear real-time picture of mixing patterns and ion concentrations throughout the vessel.

Independent work by François Ricard and colleagues at GlaxoSmithKline has compared the performance of ERT and FBRM for paracetemol crystallization. Figure 2 shows their results: the upper ERT curve detected the onset of crystals a er about 700s, while the lower FBRM curve didn’t react until some 150s later, as particles began to reach 9μm. PAT APPLICATIONS If crystal formation can be detected early by ERT, it may also be possible to control the temperature from an earlier stage in the meta-stable zone of the crystal form required, providing a Process Analytical Technologies (PAT) solution to the pharmaceutical industry. Steve Stanley at the National Nuclear Laboratory has obtained similar results using ERT to investigate the reprocessing of spent nuclear fuel and the performance of filtration and ion exchange beds.

The features of ERT are particularly benecial in assessing quantitatively whether scale-up from the laboratory to production is satisfactory. Take a simple example: barium sulphate precipitating from barium chloride and sodium sulphate solutions. It
was proposed to scale-up production from 7 to 170 liters by replicating addition times and mixing power per unit volume. ERT can provide further analysis which confirms satisfactory scale-up. Figure 3 shows relative conductivity with time between positions at the center of the vessel and close to a baffle at the two scales during rapid mixing; Figure 4 has the same at slow mixing. The two graphs are similar. Finally, everyone would expect the variation in conductivity between vessel center and wall to get smaller as mixing speed increases. ERT shows this and, in Figure 5, quantifies it.