Paul Kippax is product manager at Malvern Instruments.

Paul Kippax

Do I need particle size specifications for my pharmaceutical product?

Specifications define the critical to quality factors to which a pharmaceutical product must conform to be acceptable for its intended application. Quality by Design highlights the importance of identifying and controlling 

 the parameters that can influence final product performance: ICH Topic Q6A identifies particle size as a potentially important variable.

  The specific need for a particle size specification is determined by assessing if and how this parameter affects performance. 

For solid or suspension delivery systems, bioavailability is often directly related to particle size because it controls dissolution/solubility characteristics. Dissolution rate is directly proportional to particle surface area (Noyes-Whitney equation), so a finer particle size promotes faster drug dissolution. Particle size distribution is also relevant as a narrow distribution produces more uniform dissolution. Formulations with even a small number of relatively large particles may take some time to dissolve completely, but this may be the design intent. 

For suspensions, stability is an important issue because if the active ingredient settles there is a greater chance of non-uniform delivery. Stokes' law relates settling velocity to the physical characteristics of the fluid and the size of particles in the suspension. The relationship here is a strong one: velocity correlates with the square of particle diameter. For suspension stability, a very low settling velocity is preferable and is more easily achieved with finer particles. 

Perhaps less obviously, particle size may also affect a formulation's behaviour during processing and, ultimately, its content uniformity, which is critical. Take, for example, the widespread operation of direct compression tableting; particle size can influence segregation behaviour, the ease with which powder flows through the press and the compressibility of a formulation. In turn, these factors affect the consistency of tablet weight and composition, how the press operates and the mechanical properties of the finished product.  

Figure 1 shows a near-infrared chemical image of a recalled tablet. The active ingredientrich domains (the red areas) are much larger in the right side of the tablet than in the left, demonstrating that composition is not uniform. The product was recalled in this case as the larger particle size may have caused some unit doses to fail potency specifications. Poor control of particle size clearly resulted in a substandard product.  

As analytical techniques become more sophisticated, researchers are increasingly able to show that particle shape, as well as size, can be influential. The tableting example illustrates this point as segregation, flow properties and compressibility are a function of both size 

 shape, not size alone. It is not uncommon to find particle size specifications supplemented with complementary shape criteria. 

In summary, if any critical performance attributes depend on particle size or shape, a suitable specification is needed. If other variables can fully define performance, then none is necessary. In practice, particle size specifications are common within the industry for tablets, dry powder inhalers, suspensions and creams. 

What issues should be considered when selecting a particle sizing technique/instrument?

Once the need for a particle size specification is established, it must be defined in the most relevant way. Here, an understanding of different particle size measurement techniques (operating principles and the reported variable) is essential. There are many ways of defining particle size, but only perfect spheres, which are rare in the manufacturing environment, can be described fully using a single number (a diameter). For particles with irregular shapes, there are multiple, equally valid measures of size including:  			

equivalent volume diameter (diameter of a sphere with the same volume)

circle equivalent diameter (diameter of a circle with the same projected area)

longest dimension (arguably most relevant when particles are highly irregular — long and thin).

The ideal technique will report data that correlate closely with a given performance attribute. Measured parameter is one aspect of this, but the basis for the distribution is also important. With a number-based particle size distribution, the number of particles in each size range is reported; with a volume-based distribution, it is the volume, or for particles of constant density, the mass of material. In the case of pharmaceutical products, a mass or volumeweighted distribution is often the most relevant descriptor of the content of active ingredient as a function of particle size.

  Volumebased methods may be more sensitive to the presence of larger particles and aid the detection of the end point for unit processes, such as milling. However, number-based methods may improve detection of small amounts of fine particles, which may impact properties such as powder flowability and dissolution rate. Therefore, the choice of technique depends on the application.  

More pragmatic issues include the speed of measurement and how easy it is to transfer the technique from the laboratory to the manufacturing floor. During the early stages of a development, longer analytical times may be acceptable, particularly if the results are information rich. For particle characterization, image analysis can be an excellent choice because it provides statistically relevant size and shape data as well as visual images, which in combination give detailed insight into the product. Measurement times are relatively long (approximately 15–20 min), but still much shorter than using manual microscopy. 

In contrast, PAT must offer much more rapid analysis because the aim is measurement of a dynamic process. Where extremely fast analysis is required, other techniques become more valuable. Laser diffraction, for instance, provides a complete volumebased distribution in milliseconds. This makes it extremely productive for particle characterization in the laboratory and equally valuable for the realtime monitoring of processes, including spray events. 

Whatever the analytical method, many of the issues regarding instrument choice are shared — ease of use, flexibility and reproducibility are all key. For laser diffraction, decades of development have streamlined analytical methods; the best instruments now offer fully automated operation driven by standard operating procedures (SOPs). The repeatability of such systems is exceptionally high. Advances in the core technology have also delivered improvements in fundamental performance, broadening measurement ranges. A single instrument can be used for many different samples, building greater value into the investment. 

How do I develop and validate a methodology for particle size analysis?

If a particle size specification is required, then so too is an analytical method that will accurately reflect whether or not a material meets the specification. Measurement should be repeatable, reproducible and robust.  

If a method is highly repeatable it means that running the 

 instrument produces very similar results. Repeatability is determined by instrument performance and the state of dispersion of the sample. Typically, the aim is primary particle size measurement, in which case it is necessary to disperse any agglomerates/aggregates prior to analysis. Research has shown that inadequate dispersion is the greatest source of measurement error for particles with a diameter of 20 μm or less (Figure 2).

Laser diffraction can be used for either wet or dry analysis. Wet sample dispersion involves the use of a suitable dispersant to separate the particles, followed by stabilization with admixtures or surfactants to prevent recombination of the particles during measurement. Vigorous stirring or ultrasound treatment promotes initial dispersion. With a dry system, the air pressure used to draw the sample into the measurement zone can be manipulated to control the degree of dispersion. 

A wellengineered instrument will produce repeatable measurements provided dispersing conditions are appropriate and well defined, and the concentration of particles in the measurement zone is suitable. However, reproducible measurement is a more exacting requirement. A highly reproducible procedure gives accurate and precise results for different samples, so sampling methodology is a contributory factor. The US and European pharmacopoeias specify a target reproducibility of 10% for the volume median diameter (Dv(50)) for particles >10 μm in diameter and 20% for those that are smaller.

  This figure takes into account all the errors associated with the entire analytical process. 

Many analysts underestimate errors that arise from sampling. Laboratory analysis involves measuring a small amount of material to determine the properties of a much larger bulk. Ensuring the sample is truly representative is essential. It is estimated that for systems containing particles >75 μm in diameter, sampling is the greatest source of error.

  As with dispersion, it is sensible during method development to investigate the impact of sampling technique on size measurement and to closely define a procedure.  

Comparison of different measurement techniques can provide valuable insight for method development; for example, image analysis provides size and shape data that verify effective dispersion. With dispersion, the aim is to apply sufficient energy to disperse the particles without breaking them up. Image analysis is a relatively straightforward way of determining whether or not this aim has been achieved. Comparing image analysis and laser diffraction data is also a good way of confirming that the optical properties being used for diffraction analysis are correct. 

Validation is the final step in the development of a particle size specification and an associated measurement method. FDA guidance states that, for particle sizing, "methods validation usually involves an evaluation of intermediate precision and robustness".

  Intermediate precision is related to both repeatability and reproducibility, and is tested by measuring the same sample, on a different day, using a different instrument. Robustness is assessed by evaluating the impact of small changes to the methodology on the results. 

FDA guidance also states that "assurance should be provided that the data generated are reproducible and control the product's quality".

  Effective method development, particularly when implemented through SOPs, will give reproducible analysis. For the particle specification to control product quality, it essential to have effective definition that considers the variability associated with measurement. Understanding the particle size range that gives acceptable performance boundaries for a product enables the specification to be set. The specification will be narrower than the true window of acceptability by an amount that depends on the variability of the analytical method (Figure 3). If the product meets this specification then it will deliver the required performance.  

is Product Manager Diffraction Systems at Malvern Instruments Ltd (UK).

1. ICH Topic Q6A — Specifications: Test Procedures and Acceptance Criteria for New Drug Substances and New Drug Products: Chemical Substances, December 2000. 

3. H. Merkus "Quality Assurance in Particle Size Measurement" from Improving Standards in Particle Size Distribution Measurement, Engineering Research Centre for Particle Science and Technology (Gainsville, FL, USA, February 1997). 

4. General Chapter <429> (2008), "Light Diffraction Measurements of Particle Size," United States Pharmacopeia. 

5. European Pharmacopoeia, Chapter 2.9.31, Laser Diffraction Measurement of Particle Size, Supplement 5.6 (2006), p 4429. 

6. FDA Draft Guidance — Analytical Procedures and Methods Validation, August 2000. 

How to Establish Manufacturing Specifications

Articles

An expert on diffraction systems offers insight on particle size specifications and analysis.

Particle size analysis

Nasal drug delivery depends on many factors, including the conditions of use by the patient, the drug formulation, and the spray pump and aerosol characteristics. In recent years, the types of drug administered 

 the nasal route have expanded from locally acting drugs, such as those for allergic rhinitis, to delicate molecules for systemic activity, such as vaccines, proteins and peptides, which can be difficult to administer noninvasively. While the nasal cavity provides a delivery pathway for these large molecules, the rate of mucociliary clearance in the nasal cavity may hinder the extent of absorption. Therefore, formulators must develop mechanisms that improve absorption for high molecular weight compounds. 

 preparations, nasally administered drugs require a consistent and reproducible dose. Unlike a tablet, the dosing of a nasal spray must be achieved using a mechanical device. In addition, the device may deliver up to 200 doses compared with a single dose contained in a tablet. Therefore, enhancing the consistency of dose delivery to nasal sites of action requires the formulation characteristics and delivery device capabilities to be harmonized. Formulation researchers have recently explored new methods of delivering macromolecules nasally by using delivery systems such as liposomes, microspheres and gels, as well as applying bioadhesive additives such as chitosan and other ingredients to improve absorption 

 prolonged retention on the nasal mucosa.

 Other ingredients have also been used to increase the viscosity of formulations, with the goal of improving product stability during storage and increasing residence time of the formulation within the nasal passages following delivery.  

For the formulator, the broad range of options that may be included as ingredients in a nasal spray presents many challenges; for instance, the addition of excipients can aid drug retention or absorption following delivery, but may also have a significant effect on the properties of the spray plume, affecting the deposition site of the aerosol. Guo demonstrated using a nasal cast that nasal sprays with a low viscosity provided greater surface coverage than higher viscosity formulations, with this being related to the particle size and, possibly, the plume geometry.

Other studies have also confirmed the importance of the geometry of the nasal passages and the properties of the spray plume in influencing spray deposition.

 Guo concluded that "spray pattern, plume geometry and droplet size distribution all appeared to provide critical data for the assessment of nasal pump performance,"

 as it related to reproducible delivery from the device. Therefore, droplet size measurements, along with plume pattern and geometry assessments provide an

 test that can be used to understand the performance of the delivery device and how this is affected by changes to the formulation. These methods (with the exception of plume geometry) are also used as quality control tests.  

performance of different polymers, which may be used to improve stability or prolong retention, in model nasal spray formulations. Viscosity and droplet size were used to assess performance from a commercially available nasal spray device. 

This study concentrates on two areas of measurement that are key concerns when developing a nasal spray product: rheological measurement (including viscosity and yield stress) and droplet size distribution. Two different formulations were studied to examine how the rheological properties affect liquid atomization. One of these was a model formulation where the concentration of polyvinylpyrrolidone (PVP) in water was changed to yield different plume characteristics. PVP was used in this study as it increases the Newtonian viscosity with only a small change to the elasticity of the sample, allowing measurement of the effect of viscosity change on the droplet size. The second formulation related to an over-the-counter product where carboxymethylcellulose and microcrystalline cellulose (CMC/MCC) had been added to impart thixotropic behaviour to prevent the product from dripping back out of the nose after administration. Each formulation was delivered using a VP3 nasal spray pump actuator (Valois, France). The two formulations were characterized using a Spraytec laser diffraction particle sizing system and a Bohlin Gemini HR Nano Rheometer (both Malvern Instruments, UK).  

The FDA's draft guidance for industry, 'Bioavailability and Bioequivalence Studies for Nasal Aerosols and Nasal Sprays for Local Action', recommends how nasal sprays should be measured 

 This includes the use of laser diffraction particle sizing to assess how spray atomization is achieved. FDA also recommends that 

 measurements be conducted using an automatic means of actuating the nasal pump, as it eliminates the possibility of differences in results caused by human actuation. 

Very rapid data acquisition rates are possible using a laser diffraction system for nasal spray analysis. This allows changes in particle size during single device actuations to be measured. From this, the different 'phases' of atomization can be assessed.

 FDA recommends that the fully developed phase of atomization, where the concentration and particle size of the spray is relatively constant, be identified. This can then be compared for different formulations.

Although rheological assessment of nasal spray formulations is not referenced in the FDA guidance document, it is essential to the development of an efficient and effective product, and is included in FDA's Chemistry, Manufacturing and Controls

 For some formulations, especially those based on suspensions, it can be advantageous if the formulation displays high viscosity at low shear rates, as it can confer stability during storage and improve the spray content uniformity. Upon the application of high shear rates, such as can be experienced during shaking or atomization, the product should exhibit thixotropic properties to provide a low viscosity liquid that may be easily atomized.

On deposition in the nasal cavity, the formulation should quickly become viscous again to slow the rate of clearance from the nose. Using a rheometer, it is possible to mimic each of these situations by measuring the changes in viscosity that occur as the shear rate applied to a formulation is increased. This, in turn, can be used to understand how the addition of polymers to a product affects atomization characteristics.

To understand how changes in the Newtonian viscosity of a formulation affect nasal spray performance, the atomization performance of PVP solutions was monitored for actuation using the VP3 pump actuator. PVP concentrations of 0.25–1.5% were studied. Actuation of the device was controlled using a velocity-controlled actuator (SprayVIEW NSx; Proveris Scientific, MA, USA) using a velocity of 40–100 mm/s and an acceleration of 5000 mm/s

The rheology test results for the model formulations produced using PVP confirm that the viscosity of the solution increases with the concentration of PVP. No shear dependence of the viscosity is observed, which confirms that these solutions are Newtonian. 

The median (Dv50) droplet size produced during the stable phase of actuation of pumps containing each of the PVP solutions is shown in Figure 1 for different actuation velocities. For actuation velocities of ≥70 mm/s, a steady increase in the observed particle size is seen, mirroring the changes in viscosity. A similar trend is seen at 40 mm/s for the low viscosity solutions (≤0.5% PVP). However, a significant increase in particle size is observed at 40 mm/s when increasing the PVP concentration from 0.5% to 1.0%. This suggests that complete atomization of the liquid is not achieved at high PVP concentrations when using a low actuation velocity. 

The poor atomization performance observed at high polymer concentrations can be easily seen in the size history profiles obtained for actuation of the pump at 40 mm/s (Figure 2). At polymer concentrations of ≤0.5%, an obvious 'fully-developed' phase is observed, suggesting that liquid break-up is well controlled. However, at higher concentrations the flow velocity of the formulation through the pump nozzle is not high enough to overcome the liquid's viscosity. In the case of the atomization of the 1% PVP solution, the flow velocity is only sufficient to achieve complete break-up towards the end of the actuation process, whereas at 1.5% complete break-up is not observed at all. 

Adding CMC and MCC to a formulation increases the viscosity and leads to the formation of structure within the liquid. This causes the formulation to become thixotropic, with the viscosity at low shear rates being significantly greater than at high shear rates. This can have a significant effect on the atomization performance of a formulation. 

An example of the changes that occur upon addition of CMC/MCC can be seen in Figure 3. Here, the change in viscosity is shown for two nasal spray products — these are identical apart from the fact that one (Formulation 2) contains ~3% of a CMC/MCC blend. Addition of the polymer significantly modifies the formulation so that it becomes thixotropic instead of Newtonian. This suggests the formation of some form of structure within the liquid during storage. 

The nature of the structure present in Formulation 2 can be understood by observing how the viscosity changes when it is subjected to rapid changes in shear, as shown in Figure 4. For this experiment, the formulation was initially placed under shear at a rate of 1 s

. After 30 s, the shear rate was increased to 100 s

, causing the viscosity to decrease significantly to only 2% of its low shear value. This suggests that a rapid breakdown in the structure occurs when this product is placed under high shear. After a further 30 s, the shear rate was reduced to 1 s

 again, at which point the viscosity increased. The viscosity did not immediately return to the previously recorded low shear value and, instead, followed an initial rapid change. The viscosity slowly increased such that after 2 min it attained a level that was approximately two thirds of the initial value measured at a shear rate of 1 s

. This is indicative of the rebuilding of the polymer structure within the liquid following the application of high shear. 

The effect of the above on atomization performance using the VP3 pump actuator is significant. Formulation 1 shows good atomization performance with a range of actuation velocities, indicating that droplet break-up is easily achieved (Table 1). By contrast, Formulation 2 shows extremely poor atomization at low actuation velocities. However, once the velocity reaches a critical point, complete atomization is immediately observed (Figure 5). This correlates to the point where the shear experienced during atomization is able to break down the structure formed by the CMC/MCC polymers. The particle size remains larger than for Formulation 1 because of the higher viscosity of this formulation, even at high shear rates. 

These studies demonstrate a significant relationship between the rheological properties of a formulation and the way in which atomization occurs during actuation of a nasal pump spray. A combination of laser diffraction droplet sizing and rheological testing could aid researchers in understanding how the function of nasal spray products alters with changes in formulation. 

Table 1 Median particle size recorded for the stable phase of actuation for Formulations 1 and 2. Formulation 2 contains CMC. 

The results show that, for both Newtonian and thixotropic formulations, increasing the viscosity increases the droplet size. In addition, rheological changes have an impact on the duration of the stable phase, and affect the sensitivity of the measured stable phase particle size to changes in the actuation velocity. This, in turn, may influence the achieved site of deposition for different patient groups. Selecting the correct concentration of polymer for use as a viscosity modifier is critically important in delivering consistent droplet size, and to the production of a reliable and efficacious product.

is Product Manager, Diffraction Products at Malvern Instruments (UK). 

is President at Next Breath LLC (MD, USA). 

is Director R&D Laboratory Services at Valois SAS (France). 

7. J.D. Suman, B.L. Laube and R.N. Dalby, 

8. S.P. Newman, G.R. Pitcairn and R.N. Dalby, 

Nasal drug delivery depends on many factors, including the conditions of use by the patient, the drug formulation, and the spray pump and aerosol characteristics. In recent years, the types of drug administered via the nasal route have expanded from locally acting drugs, such as those for allergic rhinitis, to delicate molecules for systemic activity, such as vaccines, proteins and peptides, which can be difficult to administer noninvasively. While the nasal cavity provides a delivery pathway for these large molecules, the rate of mucociliary clearance in the nasal cavity may hinder the extent of absorption. Therefore, formulators must develop mechanisms that improve absorption for high molecular weight compounds.

Enhancing the in vitro assessment of nasal sprays

Laser diffraction is an important technique for particle size analysis. This article highlights the many advantages this tool has to offer.

Issues in the Appraisal of Laser Diffraction Particle Sizing Techniques

The size distribution of pharmaceutical aerosols is important in determining the deposition site of the inhaled drug particles in the respiratory tract. Cascade impaction is the industry's preferred method for particle size analysis, but measurements can be time-consuming and provide no information regarding the dynamics of the aerosol formation. Laser diffraction can accurately provide this information, allowing different drug formulations to be rapidly screened for delivery efficacy. Here, the two techniques have been combined to show that, in dry powder inhaler formulations, drug particles produced using supercritical-fluid crystallization disperse to a smaller particle size than micronized particles, providing more reproducible drug delivery.

Figure 1 Schematic of the supercritical-fluid process.			

Respirable drug delivery is becoming increasingly popular because it provides a non-invasive route with rapid drug uptake, not only for the treatment of respiratory complaints, but also for the systemic delivery of substances that cannot be delivered orally.1 Of the respirable delivery technologies available, dry powder inhalers (DPIs) are rapidly gaining acceptance because they overcome many of the disadvantages associated with devices such as metered dose inhalers (MDIs), particularly with regard to the Montr, Protocol covering propellant use.2	

However, the dispersion of drug particles to a respirable particle size during inhalation from a DPI is a concern because this must occur using the force associated with the patient's inhalation alone. DPI research has, therefore, concentrated on ensuring good particle dispersion, either through inhaler design or by engineering drug particles with specific properties. DPI research3 has shown that crystalline drug particles with smooth surfaces provide improved dispersion during inhalation compared with particles with rough surfaces. This is the conclusion of laser-diffraction-based size distribution studies performed on particles produced using a new pharmaceutical manufacturing process called supercritical-fluid (SCF) crystallization. SCF crystallization produces uniform crystalline particles with smooth surfaces, whereas the industry-standard practice of dry milling or micronizing provides less control of particle diameter and no control of surface morphology. Micronization also produces particles with rough surfaces.	

Drug particle dispersion and size distribution are key parameters in defining the efficiency of respiratory aerosols because they predict the deposition site for the drug within the respiratory tract.4 Laser diffraction was used to rapidly determine particle size distributions in DPI formulations containing micronized and SCF-crystallized particles. Because of their rough surfaces, the micro-nized drug particles will bind more strongly with excipients during storage, thus affecting the particle size of the powder plume emitted by the inhaler. Excipient–drug interactions are minimized in the case of the smooth-surfaced SCF-crystallized particles leading to increased dispersion.		

Figure 2 SEM of salmeterol xinafoate particles produced using the supercritical-fluid process; Figure 3 SEM of micronized salmeterol xinafoate particles.			

An SCF is any material maintained above a critical pressure and temperature that has the properties of both a liquid and a gas. SCF crystallization is a single-stage process that simplifies the particle manufacturing operation and can improve production efficiency. SCF crystallization produces micron-sized particles in a highly pure, crystalline, non-cohesive and solvent-free form. In addition, the physical and surface properties, as well as the crystal forms of the drug particles, can be easily regulated by varying the process parameters.				

SCF crystallization can achieve a yield of greater than 90% and is a self-contained environmentally friendly process that minimizes the risk of contamination and the use of organic solvents. Conventional multistage processes can create a multitude of solid-state manufacturing problems, including batch variation, crystal damage, processing inefficiency and out-of-specification product runs. SCF crystallization, however, gives fine, aerodynamic drug particles of optimal size (1–3 mm) and dispersibility by adjusting bulk powder properties, including geometric particle size and distribution, particle density, morphology, surface roughness and surface energy. Whereas the initial capital cost is high, because there are not many companies in the market that manufacture the specialized components, daily running costs are low because carbon dioxide (CO2) gas is inexpensive.	

One variation of SCF crystallization is solution-enhanced dispersion by SCFs.5 This uses an SCF, such as CO2, as a dispersing antisolvent and as a medium for extracting the solvent. In this process (Figure 1), the drug is dissolved in an organic solvent and the solution is co-introduced with the SCF through the annuli of a two-fluid coaxial nozzle. The two streams are thoroughly mixed in the nozzle to ensure efficient dispersion and mixing. To overcome problems with processing water-soluble compounds, a second solvent, such as ethanol, can be used. This is miscible with both water and supercritical CO2 and is introduced through the annuli of a three-fluid coaxial nozzle.		

Figure 4 Experimental set-up for particle size distribution measurements, showing Malvern's inhalation cell attached to an Andersen cascade impactor.			

This technique is distinguished by three features not available with other SCF techniques:	

Supercritical CO2 and the solvent solution are co-introduced into a jet in which the high velocity, turbulent CO2 flow accelerates both the mixing and particle formation processes. 

The composition of solvent, CO2 and precipitating material in the jet flow is fixed at the mixing point, providing the optimum conditions for uniform and continuous particle formation. 

The temperature and pressure in both the nozzle and vessel are kept constant throughout the process so that stable conditions for the formation of solid forms are established.

Dry particles, formed after solvent extraction by the SCF, are collected in a vessel maintained at a constant temperature and pressure. Unlike other methods using SCFs, this process results in particles with a size range of 0.5–30.0 mm that have a narrow size distribution, controlled solid-state properties and a low residual solvent content.	

Scanning electron micrograph (SEM) images of particles produced by the SCF process show that the particles have a plate-like morphology with defined and smooth edges (Figure 2). By contrast, SEM images of micronized particles of the same drug show irregular, agglomerated particles with less defined edges (Figure 3). Inverse gas chromatography results (Table I) also show that micronized particles have a higher surface energy than particles produced by SCF crystallization. This is possibly related to the surface roughness of the particles whereas the lower surface energy of SCF-produced powder may be because of the smoothness of the particle surfaces. 	

To analyse how this particle morphology affects the dispersion of a DPI formulation, a laser diffraction particle sizer (Spraytec; Malvern Instruments, Malvern, UK) and an Andersen-type cascade impactor (Copley Scientific, Nottingham, UK) were used (Figure 4). A cascade impactor discriminated particles based on their behaviour in airflow, reporting an aerodynamic particle size. The results are believed to correlate well with in vivo performance, hence the acceptance of impaction as the standard method for pharmaceutical aerosol analysis. However, measurements using impactors are time-consuming and provide no information regarding the dynamics of aerosol formation. Multiple actuations of the inhaler under test are also required to collect enough material for analysis.		

Table I Surface energy properties of salmeterol xinafoate particles produced by micronization and by the supercritical-fluid process. 						

Laser diffraction discriminates particles according to their volume. It utilizes the principle that particles of different sizes will scatter the light from a laser beam at an angle that is inversely proportional to their size. This non-invasive technique is rapidly gaining acceptance in the pharmaceutical industry. It allows the dynamics of each actuation of an inhaler to be followed, with much faster result reporting and manipulation compared with cascade impaction.	

In this experiment, the two techniques were combined. An inhalation cell (Malvern Instruments) was used to couple the laser diffraction instrument to the cascade impactor. This cell is specifically designed to allow laser diffraction measurements to be performed under controlled airflow conditions during testing of inhaler systems, enabling measurements in-line with a cascade impactor. This is achieved without causing a bias in the cascade impactor results.6	

Using this set-up, the standard impaction parameters such as the mass median aerodynamic diameter (MMAD), geometric standard deviation and fine particle fraction can be obtained for a series of actuations. The powder delivery dynamics and reproducibility can also be assessed for each actuation using laser diffraction. Data acquisition speeds of up to 2500 Hz (one measurement every 0.4 ms) are possible using laser diffraction, allowing the discrimination of the diffraction patterns produced by different parts of aerosol cloud such as primary drug particles and aggregates, as well as carrier particles and drug particles delayed by adhesion to the inhaler walls. These results can then be compared with the time-averaged aerodynamic size results obtained using the cascade impactor.		

Figure 5 Time-history showing the particle size measured during the actuation of the inhaler containing the micronized salmeterol xinafoate formulation.			

The DPI formulations tested in this study consisted of powders of salmeterol xinafoate produced by micronization and by the solution-enhanced dispersion SCF process. The formulations were prepared with DMV Pharmatose 325M 	

(DMV International, Veghel, The Netherlands) inhalation grade a-lactose monohydrate excipient and tested in a 13 mm3 Clickhaler inhalation device (Innovata Biomed, St Albans, UK). The laser diffraction measurements were done following the passage of the powder aerosol through a USP throat, immediately before the cascade impactor preseparator.	

Figure 5 shows the time-dependence of the particle size measured using laser diffraction for a single actuation of a DPI containing the micronized salmeterol xinafoate formulation. In this case, the powder was released from the device during a period of 400 ms. The results show that the delivered particle size remained relatively constant during the release of the powder, with a mixture of excipient and drug being observed at all times.				

Figure 6 shows the results obtained under the same measurement conditions for a DPI containing drug particles produced using the solution-enhanced dispersion technique. Separation of the fine drug particles and the lactose carrier was observed, suggesting increased deagglomeration of the fine drug particles. Dispersed, fine drug particles initially appeared in the measurement zone (up to 0.07 s). This was as expected because these particles are rapidly entrained in the airflow. During the mid-point of the spray plume (0.07–0.40 s), a mixture of lactose carrier and fine drug particles was observed. Finally, the tail of the spray plume (0.40–0.45 s) contained mainly fine drug particles. It is believed that these particles were delayed by adhesion to the inside of the inhaler.			

Figure 6 Time-history showing the particle size measured during the actuation of the inhaler containing the salmeterol xinafoate formulation produced by supercritical-fluid crystallization.			

The increased dispersion observed with the formulation containing drugs produced by SCF crystallization is believed to be related to the drug particle morphology and lower surface energy. The drug particles produced by SCF crystallization are individual crystals that are smooth-surfaced, thus aiding dispersion. By contrast, micronized drugs are generally rough-surfaced, which causes the lactose and drug particles to bind more strongly during storage, thus affecting the particle size of the powder plume emitted by the inhaler.	

Table I shows inverse gas chromatography data for salmeterol xinafoate samples produced by micronization and by the solution-enhanced dispersion SCF process. The results show that the dispersive component of surface energy (gsd) and the specific component of surface free energy (2Â³GASP) values of SCF-crystallized powder are lower than the micronized sample. Similarly, acid (KA) and base (KD) parameters of solution-enhanced dispersion powder are much lower than micronized material. Therefore, the results have demonstrated considerable changes in the surface properties of the powder prepared by the SCF crystallization process. These changes have possibly contributed towards better dispersion of powder and, as a result, a greater in vitro dose can be delivered to the lower part of lungs compared with micronized material.	

Table II shows in vitro cascade impactor data for salmeterol xinafoate samples produced by micronization and the solution-enhanced dispersion SCF process. These were analysed with a lactose blend (3.8% w/w). The cascade impactor measurements demonstrated that the SCF-processed salmeterol xinafoate has doubled the fine particle fraction compared with micronized salmeterol xinafoate. This correlates well with the increased dispersion observed for the SCF formulation during the laser diffraction measurements. 	

The laser diffraction and cascade impactor particle size measurements described in this article show that particles produced with the solution-enhanced dispersion SCF process will disperse better than micronized particles. This can lead to improved dose uniformity and more reproducible targeting of the correct deposition site for the drug within the respiratory tract.	

Table II Cascade impactor results showing the percentage emitted dose of salmeterol xinafoate particles prepared by micronization and the SCF process.						

Thus, SCF-produced particles are more likely to be deposited in the lower respiratory tract than micronized particles. The laser diffraction results show that the technique is a valuable tool in the rapid prototyping of new formulations in the pharmaceutical industry, with the results showing a good correlation with cascade impactor studies. 	

The cascade impactor provides a useful aerodynamic measure of particle size distribution, which can be utilized to compare different devices and formulations. The deposition pattern of the drug in the respiratory tract is determined by a complex interaction between the inhaler device, the formulation and the patient's inhalation technique.	

The improved performance of the SCF salmeterol xinafoate can be related to the lower surface energies of the particles and the low bulk density of SCF-processed material compared with the micronized particles. Therefore, the drug/carrier adhesion forces are of sufficient magnitude to enable the carrier particles to transfer the drug particles from the inhaler to the lower part of respiratory tract.	

This work demonstrates that products produced by the solution-enhanced dispersion SCF process improve aerosol performance compared with micronized material and give better dose uniformity. The drug particles produced by the SCF process are deposited where they are required in the lung and not in the throat; therefore, less drug is wasted. Considerable savings can be made because of a reduction in overages per device and increased patient compliance because of less frequent dosing.	

1.	J.S. Patten, J. Bukar and S. Nagarajan, "Inhaled Insulin," Adv. Drug Del. Rev. 35, 235–247 (1999).				

2.	United Nations Environment Program (UNEP) - Montr, Protocol on Substances that Deplete the Ozone Layer (1987).	

3.	P.G. Kippax, D.M.J. Higgs and M. Rehman, "Analysis of Dry Powder Aerosols Using Laser Diffraction," paper presented at Drug Delivery to the Lungs XIII  (December 2002, London, UK).	

4.	Task Group on Lung Dynamics, "Deposition and Retention Models for Internal Dosimetry of the Human Respiratory Tract," Health Physics 12, 173–208 (1966).	

5.	P. York and M. Hanna, "Particle Engineering by Supercritical Fluid Technologies for Powder Inhalation Drug Delivery," in proceedings of the Conference on Respiratory Drug Delivery V (Phoenix, Arizona, USA, 1996) pp 231–239.	

6.	C.E. Holmes et al., "Simultaneous Analysis of Respirable Aerosols via Laser Diffraction and Cascade Impaction," paper presented at Drug Delivery to the Lungs XII (December 2001, London, UK).	

Respirable drug delivery is becoming increasingly popular because it provides a non-invasive route with rapid drug uptake, not only for the treatment of respiratory complaints, but also for the systemic delivery of substances that cannot be delivered orally.