Microfiltration of Lager Beer

The Grolsche Bierbrouwerij Nederland and Aquamar-ijn Micro Filtration started a collaboration in 1995 with the purpose of investigating the possibilities of replacing diatomaceous earth (kieselguhr) filtration of lager beer with microsieve filtration [115].

Nowadays tubular ceramic or polymeric membrane systems are being developed as an alternative for classical packed bed diatomaceous earth (see Fig. 1) dead-end filtration. Complications in using these type of filtration membranes are yeast cell clogging and protein adsorption leading to a fast flux decline and subsequent elaborate inline cleaning procedures. A pilot plant was built in which the performance of the microsieves was tested. Clarification of lager beer is an important operation during the brewing process. Rough beer is filtered in order to eliminate yeast cells and colloidal particles responsible for haze. Common beer-filtration systems are based on kieselguhr. However, the exploitation costs of these systems are rather high. See Figure 53.

Cross-flow microfiltration with polymeric or ceramic membranes may be an alternative. Several studies have been carried out, but often problems like poor permeate quality (i.e., high turbidities or protein and aroma retention) or insufficient fluxes are encountered [116, 117]. Experiments at the Grolsch breweries in which a microsieve was used for the filtration of beer [118] showed a permeate flux of 4 x 103 l/m2/hr during a period of at least 5 hours without any increase in transmembrane pressure (see Fig. 54).

This flux is one to two orders of magnitude higher than typical fluxes obtained with diatomaceous earth or other membranes. In the experiments the formation of a cake layer was diminished by using a cross-flow configuration in combination with backpulse techniques and a transmembrane pressure of only 20 cm H2O (0.02 bar). As the microsieve is made of an inert material it may be cleaned with aggressive chemicals or by steam sterilization. After cleaning the new permeate still has the tendency to foam. Many other membranes may give problems on this point, as it is difficult to remove cleaning agents from the large inner surface and from deadend pores. An additional advantage is the absoluteness of the filtration: the uniform pores do not permit a single yeast cell to pass through the sieve. See Figure 55.

Moreover, extensive cleaning procedures are required, as beer turns out to cause severe fouling [117, 119, 120]. Ceramic membranes have an advantage over polymeric membranes regarding fouling, as they can withstand harsh cleaning methods. However, the obtained fluxes are usually significantly lower. Ceramic membranes with a small flow resistance would therefore be highly desirable for beer filtration. Microsieves made with silicon micromachining consist of a thin microperforated silicon nitride membrane attached to a macroperforated silicon support. The membrane thickness is of the order of the pore size, thus allowing high fluxes and relatively simple cleaning procedures. Moreover, the membrane is optically flat and smooth (surface roughness typically below 10 nm), which hampers adsorption of foulants. Furthermore, the pores are uniform in size and distribution, which may be important for quality control.

Table 5. Interaction of albumin with membranes is also measured by looking at the difference between the initial albumin concentration and the final concentration; the differences can be attributed to both protein diffusion and adsorption [113].

Whatman Millipore Micromachined

(albumin concentration) (albumin concentration) (albumin concentration)

Time (min) Absolute (g/dl) Absolute (g/dl) Absolute (g/dl)

0 0.381 ± 3.88 ± 0.02 0.423 ± 0.003 4.31 ± 0.03 0.395 3.980 ± 0.003

420 0.002 3.58 ± 0.02 0.398 ± 0.003 4.05 ± 0.03 0.394 3.970 ± 0.002

concentration 0.004

Table 6. Membrane parameters and calculated diffusivities [113].





1 (asymmetric



thickness (^m)


Porosity (%)








area (mm2)






Absolute diffusivity





Membrane fouling during filtration of lager beer with microsieves can well be studied through in-line microscopic observations. Ph.D. student S. Kuiper [121] was willingly asked in 2000 by Aquamarijn to perform a number of experiments with lager beer from The Grolsche Bierbrouwerij at Enschede. All microsieves with a pore size less than 1.2 /m for this study were manufactured by Aquamarijn (with thanks to Dimes at Delft for the photolithography).

It was observed that the main fouling was caused by micrometer-sized particles, presumably aggregated proteins and/or polysaccharide/cellulose residues. These particles formed flocks covering parts of the membrane surface. Most of the flocks could be removed by a strong temporary increase in cross-flow. Underneath the flocks a permanent fouling layer was formed inside the pores. This made frequent removal of the flocks crucial in delaying the process of permanent in-pore fouling.

Besides the fouling process the influence of pore size on permeate flux and turbidity was investigated. Centrifuged beer appeared to give a significantly clearer permeate than rough beer. For centrifuged beer and a microsieve with a pore diameter of 0.55 / m a haze of 0.23 European brewery convention (EBC) was obtained during 10.5 hours of filtration at an average flux of 2.21 x103 l/m2 hr. For a sieve with slit-shaped perforations of 0.70 x 3.0 /m2 a haze of 0.46 EBC was obtained during 9 hours of filtration at an average flux of 1.43 x 104 l/m2hr. This flux is about two orders of magnitude higher than is commonly obtained with membrane filtration of lager beer. Concentration of the beer by a transmembrane pressure (bar) flux (l/m2*hour)

transmembrane pressure (bar) flux (l/m2*hour)

Figure 54. Behavior of flux and pressure in yeast-cell filtration of lager beer with a microsieve. Reprinted with permission from [115], C. J. M. van Rijn et al., Proc. EBC Congress 501 (1997). © 1997,

Figure 54. Behavior of flux and pressure in yeast-cell filtration of lager beer with a microsieve. Reprinted with permission from [115], C. J. M. van Rijn et al., Proc. EBC Congress 501 (1997). © 1997, factor of 12 in a 3 hour run hardly influenced the magnitude of the flux. See Table 7.

Was this article helpful?

0 0

Post a comment