Ultra clean machine
Large scale production of Penicillin V-acid by Swedish company DSM has been optimised by eliminating bulk carbon and replacing it with Cuno's ZetaCarbon filters. The implementation of an effective method to decolourise an antibiotic process stream is described by R Jansson1 and M Weaver2
Large scale production of Penicillin V-acid by Swedish company DSM has been optimised by eliminating bulk carbon and replacing it with Cuno's ZetaCarbon filters. The implementation of an effective method to decolourise an antibiotic process stream is described by R Jansson1 and M Weaver2
Antibiotics have been produced in large volumes for many decades, during which time production processes have been optimised to provide higher yields and to reduce production costs. Penicillin V-acid (phenoxymethylpenicilloic acid) is produced by fermentation and is subsequently precipitated using sulphuric acid. The precipitate is then dissolved in butyl acetate and ethanol. Sodium sulphate is added to remove the water, and filter aid is added to help clarify the solution, along with activated carbon to decolourise. The slurry is then filtered through a 150µm filter to remove all solid particles.
One problem that often occurs using bulk carbon in this process is the migration of activated carbon particles downstream. This results in carbon contamination in the solvent recovery plant that causes problems when the butyl acetate, ethanol and water are separated. The reworking of the material results in expensive inefficiency.
Swedish company DSM Anti-infectives has eliminated the problem in its large scale production of Penicillin V-acid by replacing bulk carbon with Cuno ZetaCarbon filters – lenticular cartridges containing carbon immobilised in the filtration medium.
Several bench scale tests were studied to assess the possible use of the ZetaCarbon media. The bench scale trials were intended to:
Cuno's engineers optimised the carbon by performing small-scale laboratory tests using 47mm discs of ZetaCarbon medium. The filters can incorporate any type of carbon to minimise product regulatory licence changes. The choice of carbon is dependent on:
Laboratory studies were performed to remove the yellow colour found in a Penicillin solution. Studies were performed on 50cm³ production samples and the results showed that with this Penicillin V application, an activated carbon made from peat, steam activated and acid-washed, gave the greatest reduction in colour. This carbon is used in a standard grade medium called R53S that has a filtration rating of ~0.8µm.
decolourisation efficiency
To determine the efficiency and the lifetime of the R53S medium, 47mm discs were used in laboratory studies. The filter discs were challenged with 3l of Penicillin V acid solution (95% Butyl acetate + 5% ethanol) at a flow rate of 10 cm³/min-1 at ambient temperature. The decolourisation efficiency was determined by ultra-violet spectrophotometry.
The ability of the medium to remove colour from the solution decreases with increasing volumes. This is expected because the colour is adsorbed onto the active sites within the porous structure; as the active sites become full, the ability of the carbon to remove colour at the constant flow rate is reduced.
The bulk carbon in the process provides 10-25% decolourisation (75-90% colour remaining). The filter study on 47mm discs showed that 3000cm³ could be processed (3000l m-² at a flux 10l min-1m-²) before exceeding the limit of the quality criteria. This result was favourable and was scaled up.
Scaling-up techniques apply very well to filtration systems, since filtration media are generally available in a wide variety of formats and surface areas. Typically in filtration, scaling up is achieved by keeping the following parameters constant:
Bench scale tests were usually carried out at the process temperature and with a representative sample of product, so the most relevant parameter to keep constant when scaling up is the flux across the medium.
Penicillin V acid is produced in 4,500l batches and the colour removal by the bulk carbon takes one hour. Taking into account the optimum conditions for the ZetaCarbon medium and the requirements of DSM's process, it was agreed that a 16ZPC4 system – 4 x 16 inch cartridges, 12.8m² of media – would be required. This provides a flow rate of 100l min-1, with a flux across the media of 7.8l min-1m-² (lower than bench scale flux rates). This processed the batch in 45 minutes, reducing the process time for prodcuing penicillin by 15 minutes.
To verify these parameters, a smaller pilot-scale test was designed based upon these conditions, while keeping the flux constant (see table 1).
Two tests were performed with the pilot scale system: one with a solution containing a high concentration of Penicillin V (90,000 units) and the other with a solution containing a low concentration (54,000 units). These two batches reflect product variation from one batch to another.
Decolourisation efficiency of the pilot scale system was followed during the filtration. The absorbance was measured at the start and every 45 minutes, which corresponds to a complete batch of filtered product at pilot scale.
significant improvement
The graph above shows the number of batches of penicillin that the filters can process, with the two curves representing the yellow colour left after filtering the penicillin. Both curves follow a logarithmic trend against the number of filtered batches.
The following observations can be made:
A small amount of Penicillin V was manufactured on a pilot scale using the R53S filter. This showed good stability and passed the required tests over a 3-month period. From a technical point of view, the use of ZetaCarbon medium in place of bulk activated carbon proved to be feasible and economically viable at the pilot scale. The filter system also generated good quality penicillin. Given these results, it was decided to validate and implement a production scale trial.
The system was validated on a production scale using six sets of 4 x 16 inch ZetaCarbon R53S cartridges in a 16ZPC4 filter housing. It appeared that the pilot scale results were scaled up linearly to production since similar filter capacity and decolourisation efficiency was found.
As observed at the pilot scale, the production test also showed that up to 40 batches can be filtered with decolourisation efficiency in the range of 15-25%. Those results are in conformance with DSM specifications. It was finally decided to implement the system into production.
As a result, DSM has seen significant improvement in its Penicillin V manufacturing process. The advantages of using these filters instead of bulk activated carbon include:
The CUNO ZetaCarbon filtration system has been shown to be an efficient and cost-effective alternative to the use of bulk carbon because of its reliability and reproducibility, which eliminate the need for a final product reworking and quality variations. Furthermore, the filtration system has shown itself to be easy to scale up, allowing detailed investigations and complete process optimisation to be undertaken before investing in full scale process changes.
Cartridges cut contamination
ZetaCarbon filters are lenticular style cartridges composed of cellulose fibres and activated carbon immobilised by a cationic resin. As the carbon is fixed the process fluid cannot be contaminated by loose carbon powder. The filter medium can be made from any type of activated carbon and, therefore, can be suitable in any decolourisation application where bulk activated carbon is commonly used. The lenticular cartridges are installed inside a housing, which makes them easy to use, easy to replace and much cleaner compared with activated carbon powder.