SubscribeOf the many methods available for particle size measurement, laser diffraction is perhaps the most commonly used.
Suitable for the characterization of sprays, emulsions, suspensions and dry powders, it is a robust, non-destructive technique ideal for both the laboratory and process environments. Laser diffraction offers many advantages compared with other particle sizing technologies, such as speed of measurement, dynamic range and ability to measure porous and non-spherical particles. Consequently it has become the method of choice for a broad range of industries.
With the development of modern automated instruments, laser diffraction analysis has become routine. However, the need for appropriate sample preparation remains poorly understood. In this article, we explore requirements for good powder sampling and dispersion, for laboratory-based laser diffraction measurement and on-line applications in the process environment.
We describe methods available for representative sample extraction from within a process stream, and highlight techniques that ensure powder dispersion, but which do not affect the measured primary particle size.
Representative Sampling
One of the most important aspects of dry powder analysis is the need for representative sampling. Typically only a small fraction of the powder being produced will be analyzed and therefore the results obtained must be representative of the bulk material. Various different methods exist for extracting part of the process stream for on-line real-time analysis, or for withdrawing a grab sample for laboratory or at-line measurement.
When measuring a free-flowing material it is best to sample the powder whilst it is in motion, rather than from a storage vessel where segregation is more likely. It is also preferable to sample in vertical (gravity-fed flow) rather than horizontal or angled pipes - and after a randomizing screen if possible.
If the sample must be extracted from a pneumatic pipe it is essential to break up any roping inside the line. Intrusive flow conditioners solve this problem but more recently-developed non-intrusive methods are preferable, particularly in certain manufacturing environments.
One approach that eliminates any requirement for sampling is direct, in-line measurement. This method, which is attractive in terms of simplicity of set-up, is an option with laser diffraction systems for lean phase applications (< 20 Kg/h), where the mean size is 20 microns or greater.
It allows the whole batch to be measured but can only indicate the size of particles as they exist in the process stream, because no sample preparation is involved. For some applications this may be ideal, but for others where primary, as opposed to agglomerated, particle size is of interest, or where the degree of particle dispersion is variable, in-line analysis may not be appropriate. For alternative continuous measurement options an effective process interface is key.
Sampling for on-line measurement
One approach for on-line measurement is to extract sample across the entire cross-section of the pipe. However, complex sampling devices based on this principle are intrusive, and can cause downstream problems as they interrupt flow during collection. Furthermore, collection time may be significant relative to the response time of the process and therefore the requirement for real-time monitoring is not met.
The use of sampling probes strategically placed into the process stream simpler and more effective. These allow samples to be extracted in real-time, providing process engineers with the data required to directly control the process. An understanding of the nature of the stream before the sampling point helps in selecting an appropriate device.
When sampling from a homogeneous stream, it is acceptable to use a J-tube probe with a single sample hole (see figure 1), typically 5 to 12 mm in diameter. In general, J-tube probes are installed in vertical pipes with the hole positioned to capture falling material. Homogeneity of the sample flow can be verified during installation by measuring particle size as a function of J-tube position across the pipe diameter. An inconsistent particle size measurement indicates the need for an alternative sampling strategy (see figure 2).
Figure 2: Size parameters as a function of probe position within the process pipe. Large changes in the Dv50 and Dv90 are observed, showing that the powder flow is not homogeneous.
If the sample stream is not homogeneous, a flute (a tube with sampling holes drilled at selected points along its length) may be an appropriate choice (figure 1).
Figure 1: J-Tube and Flute sample probes used with the Malvern Insitec on-line laser diffraction system.
The total area of holes required can be calculated from the amount of sample required for optimum transmission. For the final design, a balance is struck between the number and size of the holes, and the need for the extracted sample to be representative of the bulk powder.
An alternative option is static flow conditioning, which involves the installation of a spool piece designed to deliver homogeneous flow when upstream powder distribution is uneven. A suitable flow conditioner, most easily selected with the assistance of CFD modelling, allows extraction with a downstream J-tube probe. With any on-line sampling solution, especially where non-homogeneity can occur, it is always recommended that a representative sample is collected and measured off-line in order to verify the suitability of the selected design.
Sampling for laboratory measurements
When extracting material from a process line for analysis in the laboratory, the same requirements for good sampling apply. An additional concern is that powders delivered to a distant laboratory will segregate during transportation. If segregation is not reversed, biased results will be obtained, so care must be taken to ensure that there is complete re-mixing of any fine and coarse particles before analysis.
This is particularly important for samples containing large particles (>70 microns in size), or having broad and/or bimodal size distributions. Laser diffraction is a volume-based measurement technique and therefore sensitive to small changes in the number of large particles in the sample. One 100 micron particle has the same volume as one million 1 micron particles and will therefore give the same scattering response.
A variety of sampling techniques is available for off-line analysis (table 1).
| Method | Estimated | Effciency (%) |
| Cone & Quartering | 22.7 | 0.013 |
| Scoop Sampling | 17.1 | 0.022 |
| Table Sampling | 7.00 | 0.130 |
| Chute Riffling | 3.40 | 0.560 |
| Spinning Riffling | 0.42 | 36.30 |
| Random Variation | 0.25 |
Simple scoop sampling (where the powder is directly sampled from the top of a powder container) can results in large errors, as segregation frequently occurs during vessel filling or storage. Use of a spinning riffler, where samples of the correct mass are obtained by continuously sampling a moving particle stream, is the most efficient method of minimizing sample bias.
Powder dispersion
Control of powder dispersion is the other aspect of sample preparation that has a major impact on the results achieved with laser diffraction. Malvern Instruments experience gained over 23 years in the design and optimal use of dispersion devices is important in providing deliverable solutions even for the most difficult of applications.
Dispersion of agglomerates occurs when the adhesion forces between particles in a sample are overcome. At large particle sizes (>10 microns) these forces are quite small (figure 3) and dispersion can be achieved by simple tumbling of the sample. However, below 10 microns the forces of attraction between particles can be very great and the energy required therefore increases dramatically.
Figure 3: Force of adhesion as a function of particle size.
Energy for dispersion is frequently supplied in the form of compressed air, used to drive a venturi in which agglomerate break-up is caused by high shear, particle-particle and particle-wall interactions. The key to a good dry measurement is to apply enough energy to cause particle dispersion, while avoiding particle fracture or milling.
Performing a "pressure or flow rate titration" is the only way to assess whether dispersion or milling is occurring [REF - ISO13320-1]. This simply involves measuring particle size as a function of the compressed air pressure or flow rate used to drive the venturi, in order to determine which setting is most suitable.
The correct pressure is selected by comparing the dry results with those obtained from a well-dispersed wet sample. A good dry measurement result can never be finer than one obtained using the correct wet dispersion method. This is because the energy and time available for dispersion in a slurry are much greater than can be achieved when particles are placed under shear in an air stream.
Laboratory dispersion case study
A pressure titration for a typical crystalline powder is shown in figure 4. As dispersion pressure is increased from 0 to 4 bar, a large decrease in particle size is observed. This is most evident for the reported Dv90 value, where a shift from almost 50 microns to below 20 microns is seen. What is unknown is whether this is the result of agglomerate dispersion or the break up of primary particles.
Figure 4: Pressure titration for a pharmaceutical powder.
Comparison with a well-dispersed wet result allows the dry dispersion process to be understood. In this case, a good wet-dry correlation is observed at 0.2 bar (figure 5), suggesting that at higher pressures particle milling occurs, distorting the measurement.
Figure 5: Comparison of the particle size distribution reported at 0.2bar with that obtained using wet dispersion.
On-line powder dispersion
The techniques employed in on-line dispersion units are similar to those in the laboratory. A gas-driven eductor draws sample from the process stream into the measurement zone. Eductor operation is based on the venturi principle, with gas flow rate used to control the rate of sample extraction and dispersion. Eductor geometry and gas choice are tailored for each application, to meet shear rate and safety requirements respectively.
As with laboratory measurements, care must be taken to avoid the break-up of primary particles. Generally, suitable operating conditions are identified by studying particle size as a function of gas flow rate, and then comparing the results with verified laboratory data.
The standard eductor (figure 6) used with the Malvern Insitec system generates relatively high shear during sampling because the gas flow is introduced at right angles to the particle stream. It is ideal for dispersing robust agglomerates and can be used with fine or coarse powders.
Figure 6: Standard eductor - air/nitrogen is introduced at 90 degrees to the particle stream through pneumatic fitting in top right hand side of photo.
For fragile particles, an alternative eductor geometry is offered which promotes gentle dispersion. With this coaxial design, gas flow is introduced and aligns in the same direction as powder flow, producing efficient powder transportation at much lower shear rates. This option is ideal for larger fragile particles, or for well-dispersed samples containing cohesive particles prone to sticking either to each other or to pipe walls.
With either, eductor flow rate titrations will confirm the suitability of the selected design and operating conditions. Data relating to the use of a coaxial eductor for fragile particles, in this case coarse lactose, are shown in figure 7. A Mastersizer 2000 (using a very low air pressure of 0.5 Bar) was used to confirm the measured size (< 2% difference in the Dv(50)).
Figure 7: Coaxial eductor flow rate titration.
Data obtained using the two eductor types to measure the same sample, a pharmaceutical excipient, are shown in figure 8. With the standard eductor, measured particle size decreases with increasing flow rate, but with the coaxial design it remains constant.
Figure 8: Change in size with gas flow rate for each eductor type.
A comparison with laboratory data indicates that, in this case, the standard eductor system, operated at a flow rate of 8.4 m3/hr, provides optimal dispersion without particle break-up. Data for several grades of the excipient (see figure 9) show that here the coaxial eductor is failing to disperse the samples, with the exception of Grade 1 material.
Figure 9: Measurements made on different material grades using the the Insitec with two different on-line eductors (8.4 m3/hr) compared to the Mastersizer 2000 (1.5bar dispersion pressure). Complete dispersion is seen using the standard eductor for each grade.
Conclusions
The importance of representative sampling and correct dispersion must be appreciated fully in order to develop effective particle size measurement systems for the laboratory or process environments.
On-line systems deliver real-time data, which are invaluable for process control. In addition they offer safety benefits, in the form of sample containment and minimal exposure of either personnel or the environment. Laboratory systems, on the other hand, are powerful research tools and essential for day-to-day QC applications.
Comparisons of laboratory and on-line data are useful for validating the design of an on-line system and understanding the analytical process in its entirety. Consistency between well-established QC measurements and a new on-line system significantly enhances confidence in any associated process control strategy.
It is most important to note that no one single solution provides the answer for all processes. Care and attention must be paid to a number of factors to generate an optimal system for any given application.
Malvern Instruments has the depth of knowledge, range of hardware and proprietary software to provide solutions for both on-line and laboratory particle sizing applications for coarse/fine, cohesive or free flowing materials.