Dynamic avalanching accurately assesses flowability and quality

Cefaclor is a β-lactam cephalosporin antibiotic that has a wide particle size distribution. Because of the nonporous nature of the material, the specific surface area value accounts for a significant amount of fine particles possibly present in the samples under analysis. In light of this, dynamic avalanching measurements were performed on Cefaclor samples characterized by comparable particle size distribution curves, but different specific surface area values.

What is dynamic avalanching?

Dynamic avalanching is a technique for investigating powder flow properties under dynamic conditions and has, in the last decade, received growing attention from the scientific community.^1,2 Powder avalanching in a rotating drum enables the evaluation of dynamic powder flow characteristics on the basis of deterministic chaos theory.³

Table 1: Tested powder samples.

In the last few years, the commercial availability of automated powder flowability analysers has contributed to the widening acceptance of this analytical technique in the pharmaceutical industry.^4,5 The first instrument launched on the market, AeroFlow (TSI Inc. Particle Instruments/Amherst, MA, USA), was based on a rotating plexiglas drum that contained the powder to be tested. The powder movement was monitored by a light and photocell arrangement; the photocell, located behind the drum, collected the light passing through the drum and recorded light variations resulting from cyclic formation of avalanches. Kaye has reported the theoretical basis for this commercial device.^3,6,7

Table 2: Fine particle content of the analysed samples.

Because of the limits intrinsic to its own arrangement, AeroFlow has been replaced by the more versatile Revolution powder analyser (Mercury Scientific Inc., SC, USA), which features a more powerful detection system. In this instrument, a video camera, with the assistance of a cold cathode back-light illumination, captures images of the whole powder contained in the drum during the rotation process. With AeroFlow, however, information regarding the powder sample was limited to just a small portion of it.

Table 3: Specific surface area of the analysed samples.

Cefaclor suitability for formulation purposes is assessed on the basis of both particle size distribution data and specific surface area (SSA) values obtained from air permeability measurements. These measurements are essential to discriminate between 'suitable' and 'unsuitable' samples because, as a nonporous material, Cefaclor's SSA values are related to the content of particles with an average diameter in the 0.2–50.0 µm range.⁸ If these fine particles are present in a significant quantity, then they can negatively affect successive formulation operations.

Table 4: Most conventional flowability descriptors measured after fluidization.

In our study, Revolution was used to test the flowability of Cefaclor; the analyses were performed on the samples with similar particle size distribution, but with slightly different fine (≤ 10 µm) particle contents. The aim of the dynamic avalanching measurements was to see if differences among the analysed samples influenced their flowability characteristics and, if so, which among the typical descriptors that can be derived from these measurements (e.g., avalanche time, avalanche power, surface fractal, etc.) was more informative to reveal these differences.

Figure 1: Overlay of power attractions for sample 1 and sample 4.

Materials and methods Materials

Four different samples of Cefaclor (ACS Dobfar SpA, Vimercate site, Italy) were selected as test materials on the basis of their different fine particles contents.

Particle size distribution

All determinations were performed on a Coulter Model LS 200 analyser (Coulter Electronics, FL, USA), and each sample was dispersed in ethyl acetate until the obscuration was between 8 and 12% before the reading was performed. All the tested powder samples showed similar rightshifted distribution curves, mainly differing from each other by fine particle contents (Table 1).

Table 5: Surface fractal data measured after fluidization.

SSA measurements

SSA measurements were conducted using the Fisher Sub-Sieve Sizer (FSSS) Model 95 (Instrument Division of Fisher Scientific Company, PA, USA) powder analyser, which operates on the air permeability principle for measuring the average particle size of powders. The system is based on the early theoretical work of Carman,⁹ which was later implemented with the standardization according to Gooden and Smith,¹⁰ and operates in accordance with ASTM Standards, Method B-330.⁸

Avalanche testing

The tests were conducted using the Revolution system. Because of the large amount of material available for each sample, the standard (100 mm diameter, 35 mm wide) borosilicate glass drum was used for all measurements. The rotating drum was filled with 100 mL of powder, which was carefully measured using a Class A glass cylinder, and the flowability tests were then conducted at a rotation rate of 0.6 rpm. Instead of selecting a time duration for the experiment, a preestablished number of 200 avalanches was set, assuring a large dataset for meaningful statistical analysis.⁵ Prior to initiating flowability analysis, the samples were fluidized according to the following scheme:

• preparation: 100 rpm for 1 min
• pause: 1 min
• fluidization: from 20 to 100 rpm with steps of 5 rpm each and a final equilibrium
• time of 5 s.

The fluidization step was included to 'normalize' the samples before the flowability test (200 avalanches at 0.6 rpm), which was then performed with a preparation time of 90 s at 0.6 rpm ahead of it. As the sample drum consisted of an aluminum body and two borosilicate glass sides, no special treatments with antistatic agents were required to prevent powder from adhering to the windows.

Results and discussion

As anticipated in the experimental section, for the purpose of this investigation, flowability was determined for four samples characterized by progressively increasing fine (≤ 10 µm) particle contents (Tables 1 and 2) and SSA values (Table 3). The most conventional flowability sample data, obtained after fluidization, are presented in Table 4.

In the first two columns of Table 4, avalanche time data (the sum of the time of all avalanches, divided by the total number of avalanches in the test) and their corresponding standard deviations are reported. This parameter was considered first because it relates directly to the ability of the powder to flow under the test conditions; the shorter the mean avalanche time, the more flowable the powder is under evaluation. When comparing the avalanche time values, however, no remarkable differences or specific trends were observed within the data set.

The avalanche median is defined as the avalanche power multiplied by the avalanche time for each avalanche. The avalanche power, measured in cubic centimeters time height (cch), is representative of the powder's potential energy as it rises in the sample drum. The avalanche median is the sum of the avalanche power values multiplied by the avalanche time for each avalanche divided by the sum of avalanche power for all avalanches measured during the test. The avalanche median is measured in seconds.

The avalanche power and avalanche median values are similar for the first three tested samples, but lower for Sample 4. The higher avalanche power and median values for the first three samples indicate they are building to a larger avalanche before avalanching compared with Sample 4. These data do not reveal any specific trend clearly differentiating the analysed samples; however, a trend can be observed for the power standard deviation values, which progressively increase from Sample 1 to Sample 4. This indicates a progressive variation in the size of each avalanche size from Sample 1 to Sample 4, which reflects the increasing potential difficulty of controlling powder flowability in a processing situation.

The power slope values, which are the slopes of the least squares regression curves obtained from the corresponding avalanche power values, did not reveal any specific trend. This descriptor, which indicates a change in the avalanche size as the flowability test progresses, exhibited negative values for all tested samples, suggesting a decrease in the sizes of the avalanches as the test progresses.

Because of the strong similarity among the tested samples, the use of attractors,³ which has proved to be extremely helpful in many circumstances, ^2,11,12 was of no use in this case. In Figure 1, the power attractors for Sample 1 and Sample 4 (the two samples with the greatest difference in avalanche power data) are overlaid. The two clouds almost overlap completely with the centroids and the moments of the two point sets that are representative of the data scatter being very close to each other. The power attractor is a phase space diagram produced by the Revolution plotting the 'avalanche power_n' versus the 'avalanche power_n+1' for a series of successive avalanches.

If power attractors are poorly informative, surface fractal values, on the other hand, exhibit a trend similar to that of fine particle contents and SSA values by increasing from Sample 1 to Sample 4. Table 5 summarizes values for this powder descriptor measured on the tested samples together with the corresponding relative standard deviations.

The surface fractal measured using the Revolution corresponds to what mathematicians define as "fractal dimension"; an explanation of the mathematical and physical significances of fractal dimension can be obtained in references 6 and 7, but, for the purposes of this simple investigation, it has just been considered a parameter that provides a quantitative measure of the degree of heterogeneity of a powder surface. Powders have 'fractal properties' (a rough or fragmented shape that can be subdivided in parts; each of which is approximately a reduced/size copy of the whole),¹³ which means fractal geometry is more suitable for describing them than classical Euclidean geometry. According to the manufacturer of the Revolution, surface fractal values can range from 1 to 11, with a value of 2 corresponding to a powder with a smooth, even surface and the value progressively increasing as the powder surface becomes more rough and fragmented. As shown in Table 5, surface fractal values progressively increase from the Cefaclor sample with the lowest fine material content (Sample 1) to that with the highest (Sample 4) — similar to the corresponding SSA values. This trend of surface fractal values could be related to a progressive increase in powder surface roughness because of the presence of fine particles, suggesting that surface fractal measurements could represent an alternative to FSSS determinations to discriminate samples with similar particle size distributions, but different fine particle contents.

What can we learn from this?

Quick and simple flowability measurements conducted using the Revolution powder analyser on Cefaclor samples with similar particle distribution, but progressively increasing fine particle contents and SSA values, have demonstrated surface fractal values that show a similar trend: values increase as the SSA value in the sample increases. This suggests a relationship between the surface fractal value and the amount of fine particles present in a sample that is worthy of further investigations. The findings also suggest the possibility of using fractal dimension values as opposed to SSA measurements to assess the suitability of Cefaclor samples for successive formulation operations.

Acknowledgements

The Authors would like to thank Mr Marco Falciani, President of ACS Dobfar SpA for his support and continuous encouragement in investigating antibiotics solid state properties.

Riccardo Bonfichi is Corporate Quality Control Director and Qualified Person at ACS Dobfar SpA (Italy). Tel. +39 33 4656 1731 Fax +39 02 9069 3226 riccardo.bonfichi@acdobfar.net

Giulia Cloralio is Head Quality Control ACSD4 at ACS Dobfar SpA.

Andrea Rainoldi is Analyst Quality Control ACSD4 at ACS Dobfar SpA.

References

1. S.A. Howard, "Solids flow properties," in Encyclopedia of Pharmaceutical Technology, 3rd Edition (Informa Health Care, New York, New York, USA, 2007) pp 3275–3296.

2. F. Lavoie, L. Cartilier and R. Thibert, Pharm. Res., 19(6), 887–893 (2002).

3. B.H. Kaye, Part. Part. Syst. Charact., 14, 53–66 (1997).

4. Y.S.L. Lee et al., AAPS PharmSci. Tech., 1(3) article 21 (2000).

5. B.C. Hancock et al., J. Pharm. Biom. Anal., 35(5), 979–990 (2004).

6. B.H. Kaye, Part. Part. Syst. Charact.,10(3), 99–110 (1993).

7. B.H. Kaye, Part. Part. Syst. Charact.,10(4), 91–200 (1993).

8. ASTM Standards — ASTM B330. www.astm.org

9. P.C. Carman, J. Soc. Chem. Ind.,57, 225–234 (1938).

10. E.L. Gooden and C.L. Smith, Ind. Eng. Chem., Anal. Ed.,12, 479–482 (1940).

11. J.W. Hobbs and D. Rhodes, The use dynamic avalanching and fractal analysis to characterize uranium oxide powders, International conference "Scientific research on the back-end of the fuel cycle for the 21st century", Avignon (France), Poster 2-22, 1-4 (2000). www-atalante2004.cea.fr/home/liblocal/docs/atalante2000

12. T.M. Crowder et al., Part. Part. Syst. Charact., 16(4), 191–196 (1999).

13. B. Mandelbrot, Les object fractals, forme, hasard et dimension (Flammarion, Paris, France, 1975).