Smooth flow from the feed hopper is a prerequisite, so flow properties are important, but arguably the most important issue
here is the powder's response to a vacuum. Is the formulation drawn into the port effectively? Will it stay in place as the
drum rotates? If a relatively strong vacuum is needed to ensure success, then a correspondingly large positive pressure is
needed to eject the dose, increasing the risk of powder spillage.
Characterisation data for two forms of lactose allow comparative predictions of their inprocess behaviour for this application
(Figure 1 and Figure 2). These data allow comparative predictions of the in-process behaviour of the two forms of lactose, for this application.
The spray dried lactose is more heavily influenced by the application of a vacuum than the milled alternative, with flow energy
rising to a much higher level. The permeability data (Figure 3) suggest why. The spray dried lactose bed is relatively porous so an applied vacuum induces a uniform pressure drop across
the full height of the sample. With the less porous milled lactose, the pressure drop is concentrated near the vacuum's source.
In both, the pressure drop is similar, but has different effects on the powder bulk. The flow energy of the spray dried lactose
increases more markedly because the vacuum acts on the entire bed. With the milled lactose only a portion of the bed is affected,
making the rise in flow energy more modest.
The results suggest that the spray dried lactose has more appropriate properties in relation to response to vacuum for this
application. With this excipient the vacuum will work more efficiently, during port filling and retaining the powder plug
in place. Only a relatively low vacuum will be needed, with a concomitant reduction in positive pressure.
This approach suggests how formulators and process designers, together, can use powder characterisation data to engineer a
better process. While this can be done without referencing other information, existing manufacturing experience adds an extra
dimension. Those routinely operating a similar vial filling process may, for example, notice that processing formulation 1
is easy, while formulation 2 presents problems. By characterising these formulations, this experience can be converted into
more useful knowledge, such as "formulations with a permeability of x and a flow energy of y, under an applied vacuum of 5
kPa, process well; those with differing values of x and y do not". This sets more specific guidelines for formulators targeting
a similar process.
So far, the focus has been formulation, but what about the next steps: detailed process design and ongoing manufacture?
Once clinical development is complete, the designer is faced with the task of developing a robust manufacturing process, which
may be difficult because the formulation is now fixed and may not be ideal from a manufacturing stance. Here, the exact same
powder understanding permits intelligent process development. Returning to the vial filling process, if permeability and flow
energy are suboptimal then it is clear where problems will arise, and decisions can be made at this relatively early stage
to include additional operations, such as a granulation step. Adding shear data aids optimal design of the feed hopper and
selection of the best material of construction. All these steps help to reduce postinstallation modifications.
Lastly, manufacturing success relies on the skills brought to bear during daytoday operation. Batch processing is the norm
in the pharmaceutical industry and often relies on an individual's operating experience. Batchtobatch variability and compromised
yields are relatively common, but rationalising experience helps develop the understanding essential for a knowledge-led approach.
By understanding and quantifying why formulation 1 works well and formulation 2 is more difficult, more sensitive and relevant
feed specifications can be set, and a poorly performing batch can be detected before entering the plant. This helps to avoid
downtime resulting from, for example, a change in excipient supplier, or variability in an upstream unit.
In summary, powder characterisation data provide information to support every phase, from the knowledgeled product and process
development that underpins manufacturing success over the longterm, to the intelligent daily operation that ensures consistent
product quality with minimum waste.
1. T. Freeman and R. Price, Drug Delivery Technology (May 2009).
The author says...
- With the majority of APIs delivered as powders, the importance of effective powder characterisation has long been recognised
by the pharma industry.
- Many variables influence powder behaviour; however, it is not possible to predict the performance from all contributing factors.
- Dynamic powder characterisation is appealing when it comes to powder handling because it can characterise powders in consolidated,
conditioned, aerated and fluidised states.
- Used in combination, dynamic testing and traditional powder measurement techniques can generate a data set that helps identify
the variables that best describe powder behaviour in any situation.