Effect of flow and packing properties on pneumatic conveying of powdered mixtures for direct compression

Transcription

1 Effect of flow and packing properties on pneumatic conveying of powdered mixtures for direct compression C. D'Andria a, G. Massimo a, P. L. Catellani a, P. Santi a, R. Bettini a, L. Bianco b, P. Colombo a * a Department of Pharmacy, University of Parma, Parco Area delle Scienze 27/a, Parma, Italy b CFM SpA, Via Porrettana, 1991, Zocca (MO), Italy Submitted to Pharmaceutical Technology Europe, July 1999 * Corresponding author. Tel , Fax address: farmac2@unipr.it

2 ABSTRACT The purpose of this work was to evaluate the effectiveness of the air conveyor 3VT 22XC (CFM SpA, Zocca, MO, Italy) when used for transferring powder mixtures for direct compression, which are sensible to segregation. True density and flow properties of powdered material were the variables evaluated. Their optimisation for pneumatic transport allows the homogeneity of the product to be maintained.

3 INTRODUCTION Pneumatic conveying is a method for transferring powdered or granular materials through a pipeline, using the motive force of air under a positive or negative pressure. It has advantages linked to flexibility, efficiency, safety and respect for the environment. However, the technique can experience material handling problems, such as adhesion, blocking and breakage of product 1. The air conveyor 3VT 22XC (Figure 1), by CFM SpA (Zocca, MO, Italy), was constructed in compliance with the peculiar requirements of the pharmaceutical, chemical and food sectors. In particular, the conveyor is ideally used on capsule or tablet machines in order to transfer powder mixtures. The machine is constituted by a vacuum pump that creates the void in a hopper by means of a connecting pipe. When the pick-up probe tube is immersed into the product, this is sucked up and deposited in the hopper. A discharge valve on the hopper allows the product to drop out by gravity. An optional vibration system facilitates the descent of less flowable products. A backwash of blown air cleans the primary filter after each discharge. The loading, discharge, and filter cleaning times can be adjusted, depending on the nature of the product and its flow ability. In pharmacy, product segregation during conveying determines a loss of product quality and process efficiency. For example, the segregation of a tablet mixture, such as in the case of a free flowing filler mixed with drug, can determine an incorrect dosage of drug in tablets. A properly engineered system and an optimisation of transported material contribute to avoid these problems 2-3. A basic concept in pneumatic transport is that the fluidised powder bed should quickly recover "close" packing conditions, in order to avoid segregation phenomena caused by size and density differences between the mixture components. The aim of this work was to evaluate the effectiveness of the air conveyor 3VT 22XC, when used with powders having borderline flow and packing properties. Product and machine variables were investigated. The product variables were the derived properties of powder, i.e. flow and packing. The machine was used in standard conditions, by comparing the presence or absence of discharging vibration system. Mixtures for direct compression were designed in order to have a progressive variation of size distribution of particles, flow properties and packing conditions. The evaluation procedure was performed on four powder mixtures containing 0.1 % w/w of a coloured marker. The content uniformity of the marker and the particle size distribution of mixtures as a result of transport were evaluated. Additionally, the work wished to establish a validation guideline procedure for the user.

4

5 MATERIALS AND METHODS Two excipients for direct compression, namely Di-Cafos (calcium phosphate dibasic dihydrate, Giusto Faravelli SpA, Milan, Italy) and Avicel PH 102 (microcrystalline cellulose, Prodotti Gianni SpA, Milan, Italy), and Amaranth (Red 2 E123, Variati & Co SpA, Milan, Italy) as marker material, were employed. The excipients were mixed with 0.1 % of the marker, previously sieved, in a V-shaped mixer (100 liters capacity) for 30 min at a rotation speed of 25 rpm. Magnesium stearate 1 % was added to the mixture and the mixing was restarted for 15 min. The composition of the four mixtures is illustrated in Table I. The particle size of powders was determined by sieve analysis. Geometric mean diameter, dg, and geometric standard deviation, σg, were calculated from the Log-probability plot of the cumulative percent undersize weight distribution. Packing properties were determined by measuring, in a graduated cylinder, the volume occupied by 100 g of powder. Bulk (V3) and tapped (V500) volumes were measured after 3 and 500 taps respectively. From these values the apparent densities (bulk and tapped) and Carr Index 4 were calculated. The flowability of the materials was determined by measuring the Flow Index by mean of Flodex Tester (Copley Instruments, UK). Flow Index (mm) corresponds to the minimum opening size of the device's orifice through which the powder freely flows and is inversely proportional to flowability of the powder. The pneumatic conveyor 3VT 22XC was used, manually commanding the discharge valve. The mixtures, stored in PET containers, were sucked up through a tube of 3 mt in the hopper placed 2 mt from the ground, with a loading time of 20 seconds. Then, the material was discharged, collecting four successive samples (coded A, B, C e D), corresponding to four different levels of the powder in the hopper. Samples A and D corresponded to the material at the top and the bottom of the hopper, respectively. Samples B and C corresponded to the intermediate material between the positions A and D. After the first discharge, a second loading cycle was carried out and four additional samples (coded A', B', C' and D') were collected. Furthermore, two samples were withdrawn from the beginning and the end of the pick-up tube (where "beginning" means the part of the tube dipping into the product and "end" the part close to the hopper). In order to verify the possibility of an influence of vibration system on the homogeneity of the mixtures, sampling and analysis of the material discharged with vibration system were performed as well. For such reason, after two successive loading cycles, four samples (coded Av, Bv, Cv and Dv) were collected during the second discharge with vibration. All the samples

6 collected were analysed spectrophotometrically at 522 nm in order to verify the colouring agent content. Finally, packing conditions of mixtures during transport were evaluated by measuring the weight and volume occupied by the amount of material in both the tube and hopper. From these volumes the apparent density of the mixture during transport was calculated.

7 RESULTS AND DISCUSSION The prepared mixtures were designed in order to obtain a continuous variation of packing and flow properties of powder bed. In particular, packing conditions were considered to be a critical factor for maintaining the physical stability of mixture. In fact, the stability of a random mixture, constituted by coarse (excipient) and fine (drug) components, is favoured by close packing of powder bed. In this situation, there are fewer empty spaces, through which the percolation of fine material can occur 5. Table II shows the physical and technological properties of the mixture components. Compared with Di-Cafos, Avicel PH 102 presented lower particle size, larger size distribution, higher values of bulk and tapped volumes and less favourable flow properties. Di-Cafos presented a higher true density. Table III shows how technological properties were modulated in the four mixtures prepared. Mix #1 and #4 showed characteristics very similar to the individual excipients, revealing a beneficial influence of the two fine powders (marker and magnesium stearate) on flow and packing of powder bed. Mix #2 and #3 showed values of size distribution, apparent density and flow properties intermediate between those of the two main components. Before the transportation experiments, the mixtures were tested for segregation possibility by gently vibrating a sample on a sieve with an aperture of 38 µm. After 5 minutes of vibration the powder remaining on the sieve was three times less concentrated in marker. This propensity to segregation indicated that such mixtures are at risk pharmaceutical preparations which demand a concerted effort to maintain uniformity. In order to evaluate the stability of the mixtures, we used as reference the limits established by US Pharmacopeia for the dosage form uniformity, which is met if the amount of marker lies within the range of 85 percent to 115 percent. Then, the four different mixtures were transferred by pneumatic conveyor 3VT 22XC. Figure 2 shows the content uniformity of colouring marker among samples collected after transport. Mix #1, constituted only by Di-Cafos, showed that the samples corresponding to the starting of pneumatic conveying, had a significantly low colouring agent content. However, such content improved with the transport. The samples of Mix #2 were not significantly different from the value measured before transport. Mix #3 appeared more variable than Mix #2, but the differences were not significant. Finally, Mix #4, constituted only by Avicel, presented a certain reduction of colour content, especially in correspondence with the first loading cycle. It was in general noted that the content of colouring agent was lower than expected. This indicated that not a segregation process but rather a loss of marker occurred. We verified that part of

8 the marker was captured by the pick-up tube (clearly coloured in red) and vacuum filters (intensively coloured). Samples of powder recovered from filters had a colour content 7-8 times higher. This reduction in colour content was very clear with the borderline Mix #1 and #4, that must be considered quite unusual in pharmaceutical practice. In addition, Mix #1 (Di-Cafos 98%), which had a true density of 2.35 g/cc, was the heaviest for the same volume transported. Mix #4 (Avicel 98%), which was the lightest (true density of 1.55 g/cc), presented the least favourable flow properties. Therefore, true density and flowability appeared crucial variables for transportation. In fact, the mixtures of Di-Cafos with Avicel (Mix #2 and #3) showed improved behaviour compared to Mix #1 and #4. The result of combining the two direct compression excipients was the reduction of the true density while maintaining acceptable flow properties. The influence of the density of the material to be transported was also indicated by the improvement of marker content during the progression of transport in the two loading cycles. In fact, the denser Mix #1, showed a dependence of marker content on the moment of transport, since it increased proceeding from samples A to D and from sample A' to D'. Therefore, this mixture underwent certain loss of marker during the starting phase of pneumatic transport, less evident during the stationary transport. This phenomenon was not shown by the Mix #2 and #3 (with intermediate density), in which the homogeneity was significantly maintained during the transport. With regard to Mix #4, with its lower density but less favourable flowability, a certain loss of marker content was measured in particular in the first loading cycle. The transport conditions respected the composition and physical integrity of powder particles, which though granular, did not undergo abrasion because of friction generated during pneumatic transport. No significant difference was observed in terms of particle size distribution of transported product. The geometric mean diameter and geometric standard deviation of mixtures, in particular the fine particle fractions, were not significantly changed after transport. Figure 3 shows the homogeneity of the mixtures that remained in the pick-up tube after the second loading cycle. The difference of homogeneity was very evident in the case of Mix #4, the least flowable. In particular, a sensible loss of marker material was observed at the end of the pick-up tube. It must be underlined that the powder in correspondence with this position was very intensively exposed to up and down movements, due to material fluidisation during transport and packing back into the tube when aspiration stopped. The vibration system facilitating discharge did not change the homogeneity of transported mixtures. For example, the samples of Mix #4 (Figure 4) presented a lower variability compared with

9 preceding transport tests. Mix #2 and #3 showed, even in these conditions, favourable behaviour. Results concerning the heaviest Mix #1, showed even in this transport test a dependence of homogeneity on the transport sequence. Finally, the vibration system did not cause any significantly change to particle size distributions of the transported mixtures (data not shown). The measurement of apparent density of mixtures in the hopper and pick-up tube (Table IV) allowed evaluation of the packing conditions of the powder bed during pneumatic transport. In particular, packing into the tube was referred to the powder bed during the aspiration phase, whereas, after powder loading into the hopper, packing was in an "at rest" situation. In fact, the apparent density (and porosity) values calculated for the portion of mixture contained in the pick-up tube, showed that during the transport the mixtures were in "fluidised" conditions. The porosity of powder bed during transport was higher than the bulk value (V3), indicating an incipient fluidisation of the bed, without reaching the bubbling or slugging fluidisation 6. On the contrary, the values of the mixture deposited in the hopper were close to tapped density values (V500). These results showed that the mixtures, conveyed fluidised into the pick-up tube, quickly recovered a close packing condition in the hopper. This was attributed to the insertion of the pick-up tube in a central position in the hopper allowing the powder to deposit directly without cycloning in the hopper 7. The permanence of the material in fluidised conditions was shorter than 2-3 seconds and the close packing recovery was instantaneous. Such behaviour helps in preventing segregation phenomena, always possible with random mixtures which have significant size and density differences between components.

10 CONCLUSIONS In conclusion, the air conveyor 3VT 22XC, manufactured by CFM SpA (Zocca, MO), showed considerable capabilities for conveying powdered mixtures for direct compression, which are usually critical in terms of homogeneity. In fact, the conveyor succeeds in transferring mixtures at borderline of the normally employed pharmaceutical preparations. For example, the mixture containing only Di-Cafos, which is substantially "heavy" in relation to the mixtures normally used for capsule or tablet manufacturing, despite a higher variability of the tracer content compared to the other "lighter" mixtures, was transferred with satisfactory results. The mixture containing only Avicel PH102 was also easily transferred, despite a certain loss of marker. This mixture contained an excessive amount of microcrystalline cellulose as compared to the usual composition of preparations containing this excipient. In this case, the major problem was the less favourable flow properties. As such properties were optimised, including in the composition a portion of calcium phosphate granulation, the result was improved flow and transport with refound homogeneity. True density and flow properties are, therefore, the most critical characteristics to be evaluated for the pneumatic transport of powdered material, when the homogeneity conditions of the product must be maintained, in particular in the case of mixtures with a propensity to segregation. The air conveyor fluidised the powder bed during transport, but the powder quickly recovered favourable close packing conditions in the hopper. In conclusion an acceptable level of homogeneity of the conveyed material was obtained, despite the variations of critical properties of the powders. However, partial adhesion of product to the equipment and an aspiration of more aerodynamic fractions in the machine filters, was observed and should be taken into consideration when using the conveyor. Finally, the conveyor did not cause wear on the particles of conveyed material, even when these particles were granular, as in the case of calcium phosphate dihydrate.

11 REFERENCES 1. R. Farnish, "Efficient Pneumatic Conveying: A key to Commercial Advantage", Pharmaceutical Technology Europe, 10 (12), (December 1998). 2. Y. Morikawa, "Transportation", in Powder Technology Handbook, Marcel Dekker, New York, NY, (1997). 3. T. Yokohama, "Fluidity", in Powder Technology Handbook, Marcel Dekker, New York, NY, ). 4. R.L. Carr, "Evaluating flow properties of solids", Chem. Eng. 72, (1965). 5. J. N. Staniforth, "Advances in Powder Mixing and Segregation in Relation to Pharmaceutical Processing". Int. J. Pharm. Tech. And Prod. Mfr. 3 (suppl), 1-12 (1982). 6. D. M. Jones, "Air suspension coating", in J. Swarbzick, J. C. Boylan, Eds., Encyclopædia of Pharmaceutical Technology, (Marcel Dekker, Inc, NY) Vol. 1, (1988). 7. C. Gualandi, Personal Communication, CFM SpA, Zocca (MO), July 1999.

12

13 Figure 1: Schematic representation of 3VT 22XC pneumatic conveyor. 1. pick-up probe tube; 2. suction tube; 3. connecting pipe; 4. safety filter; 5. vacuum pump; 6. hopper. Figure 2: Marker content in the samples after conveying. The point Mix represents the marker content and its variability before conveying. Points A, B, C e D refer to the samples of first discharge, whereas points A', B', C' and D' to the second. Figure 3: Marker content in the samples of mixtures collected at beginning ( ) and end ( ) of the pick-up tube. Figure 4: Marker content in the samples of mixtures discharged with vibration system.

14

15 Table I: Composition of mixtures (%). Table II: Physical and technological properties of excipients. Table III: Technological properties of mixtures. Table IV: Apparent densities (g/cc) of the mixtures before and after transportation.

16 Mix #1 Mix #2 Mix #3 Mix #4 Di-Cafos Avicel PH Mg Stearate Amaranth

17 dg (µm) σg True density (g/cc) V3 (ml) V500 (ml) CI Flow Index (mm) Di-Cafos (65)* (59)* Avicel (78) 225 (71) * Into brackets values of porosity per cent. Apparent volumes have been determined on powder weight of 100g.

18 dg (µm) σg V3 (ml) V500 (ml) CI Flow Index (mm) Mix # (60)* 97.5 (51)* Mix # (67) 130 (62) Mix # (70) (65) Mix # (73) (67) * Into brackets values of porosity per cent.

19 Mix ρ (V3) ρ (V500) ρ (hopper) ρ (tube) # (60)* 1.03 (51)* 0.99 (57)* 0.62 (73)* # (67) 0.77 (62) 0.75 (63) 0.51 (75) # (70) 0.63 (65) 0.61 (66) 0.41 (78) # (73) 0.51 (67) 0.50 (67) 0.32 (79) * Into brackets values of porosity per cent

20

21