Use of RVF technology to achieve rough beer clarification and cold-sterilisation of beer.

HAL Id: hal-02332550

https://hal.archives-ouvertes.fr/hal-02332550

Submitted on 4 Jun 2020

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Use of RVF technology to achieve rough beer clarification and cold-sterilisation of beer.

Luc Fillaudeau, Stanislas Ermolaev, Nicolas Jitariouk, Antoine Gourdon

To cite this version:

Luc Fillaudeau, Stanislas Ermolaev, Nicolas Jitariouk, Antoine Gourdon. Use of RVF technology to achieve rough beer clarification and cold-sterilisation of beer.. 28. Congress of the European Brewery Convention, May 2001, Budapest, Hungary. �hal-02332550�

27

USE OF RVF TECHNOLOGY TO ACHIEVE ROUGH BEER CLARIFICATION AND

COLD-STERILISATION OF BEER

FILLAUDEAU L.

, ERMOLAEV S.

, JITARIOUK N.

and GOURDON A.

Institut National de la Recherche Agronomique^(a) (INRA)

Laboratoire de Génie des Procédés et Technologie Alimentaires (LGPTA)

369, rue Jules Guesde – BP39 - F-59651 VILLENEUVE D’ASCQ cedex – FRANCE Tél : +33.(0).3.20.43.54.36 – Fax : +33.(0).3.20.43.54.26

Email : [email protected]

ProFiltra SARL ^(b)

78, rue d’Aguesseau - F-92100 BOULOGNE-BILLANCOURT – FRANCE Tél : +33.(0).1.46.05.10.72 – Fax : +33.(0).1.46.05.67.60

Descriptors:

Filtration, RVF technology, beer, clarification, cold-sterilisation SUMMARY

Over the last two years, RVF technology (Rotating and Vibrating Filtration) has been studied for rough beer clarification and cold-sterilisation of beer. RVF laboratory module is divided into parallel cells including flat disc membrane fixed onto porous substrate which drain the permeate, and impeller shaped rotating bodies attached to a central shaft. The membranes used were metalo-ceramic flat discs with mean pore diameters in microfiltration. The results show the quantitative (flux, hydraulic resistance, retention) and qualitative (pH, haze, colour, bitterness, polyphenol, protein, dry matter and yeast concentration) performances. The conclusions demonstrate an interesting potential for rough beer clarification and cold-sterilisation of beer.

RVT-Technologie zur Bierfiltration und Kaltsterilisation

Deskriptoren : Ausbeute, Bierfiltration, Bierqualität, Sterielfiltration

ZUSAMMENFASSUNG

In den letzten beiden Jahren wurde die RVT-Technologie (Rotating and Vibrating Filtration) zur Bierfiltration und Kaltsterilisation von Bier untersucht. Das RVT-Labormodul ist in parallele Zellen unterteilt. Die flachen Scheibenmembranen, die an durchlässiges Material

1 Corresponding author

fixiert sind, sorgen für den Ablauf des Permeats. Der wie ein Laufrad geformte rotierende Körper ist an einer Welle befestigt. Die verwendeten Membranen bestehen aus flachen, metallkeramischen Scheiben; der durchschnittliche Porendurchmesser ermöglicht eine Mikrofiltration. Die Ergebnisse zeigen die quantitative (Durchfluss, hydraulischer Widerstand, Retention) und qualitative (pH, Trübung, Farbe, Bitterkeit, Polyphenole, Proteine, Trockensubstanz und Hefekonzentration) Leistung. Die Ergebnisse zeigen das Potenzial der RVT-Technologie für die Bierfiltration und die Kaltsterilisation von Bier.

Utilisation de la technologie RVF pour obtenir la clarification de la bière brute et la stérilisation à froid de la bière

Mot-clés :

Filtration, RVF technologie, bière, clarification, sterilisation à froid RESUME

Au cours des deux dernières années, la technologie RVF (Rotating and Vibrating Filtration) a été étudiée pour clarifier la bière de garde et stériliser à froid la bière. Le module RVF se compose de deux cellules parallèles comprenant des membranes planes fixées sur un support poreux permettant de collecter le perméat, et d'un système d'agitation fixé sur un axe central. Les membranes utilisées étaient des disques metalo-céramique avec un diamètre moyen de pore en microfiltration. Les résultats présentent les performances quantitative (flux, rétention) et qualitative (pH, trouble, couleur, amertume, polyphenol, protein, matière sèche et concentration en levure). Les conclusions démontrent la potentialité de cette technologie pour clarifier et stériliser la bière.

INTRODUCTION

Membrane separation technology has become widely used in the food processing industry [5]. Important developments have taken place in the dairy industry, in drinkable water production as well as in wastewater treatment. In the cross-flow microfiltration (CFMF) of fermented food products (beer, wine), the fouling mechanisms and local phenomenology associated with fouling are largely unknown and still unidentified. In consequence, industrial applications of CFMF encounter two main problems: (i) the control of fouling mechanisms and (ii) the enhancement of permeate quality.

In the brewing industry, clarification and pasteurisation are among the most important operations. At the final step of beer processing, beer clarification (i.e. elimination of yeast cells, permanent and chill hazes after beer maturation) is currently obtained by conventional dead-end filtration with filter-press using filter-aids (Kieselguhr).

Pasteurisation is then necessary to ensure microbiological stability of the final product. Brewers are very concerned that the finishing techniques they use are the best, in terms of product quality and cost effectiveness. Considerations of the brewing process indicate two areas where MF might play useful roles [10]: (i) loss reduction in the brewing process, and (ii) as a technological alternative to the conventional solid-liquid separations. The aim of this work is closely related to the clarification of rough beer and the pasteurisation of clarified beer, hence a rapid description of these operations and their objectives.

Clarification of rough beer

Beer clarification is probably one of the most important operations: rough beer is filtered in order to eliminate yeast and colloidal particles responsible for haze (chill and permanent hazes); in addition, this operation should also ensure the biological stability of the beer. This operation should comply with the haze specification of a lager beer in order to produce a clear bright beer (European Brewery Convention norm). Standard filtration consists in the retention of solid particles (yeast cells, macro-colloids, suspended matter) and solutes responsible for haze. The variety of compounds (chemical diversity, large size range) to be retained makes this operation one of the most difficult to control. The use of MF is to provide an alternative to conventional dead-end filtration with filter-aids such as diatomaceous earth. However this operation must satisfy the same economic and qualitative criteria [16, 19]. MF should be able: (i) to produce a clear and bright beer with similar quality to a Kieselguhr filtered beer, (ii) to perform separation in a single-step without additives, (iii) to operate at low temperature (0°C) and (iv) to achieve economic flux.

Cold sterilisation of clarified beer

Clarification is usually followed by pasteurisation (with a plate heat exchanger).

Pasteurisation is necessary to ensure the microbiological stability of the final product.

A stability of 3 to 6 months can be ensured when the retention of beer spoilage organisms (bacteria, yeast) is obtained. The level of one yeast cell per millilitre represents an admissible standard with common filtration techniques [1]. Currently, heat treatment is mainly performed by flash pasteurisation before conditioning or tunnel-pasteurisation of bottles. Sterile filtration (by CFMF) appears interesting to replace pasteurisation by heat treatment and allows the elimination of the organoleptic problems induced by thermal processing [13, 14]. CFMF is examined in order to (i) produce a microbial free beer without deterioration in beer quality by operating at low

temperature (close to 0°C), (ii) ensure beer stability (biological, colloidal, colour, aroma and flavour, foam stability), (iii) achieve economical flux and (iv) indicate the viability of MF as a commercial alternative to pasteurisation and dead-end filtration with cartridges. Cold-sterile filtered beer (draught beer or bottled beer) corresponds to a strong demand from consumers for quality and natural products. The objective of eliminating thermal treatment of the finished product is achieved with membrane cartridge systems (dead-end filtration) installed directly upstream of the filling system [8, 9]. However, cold-sterilisation by cross-flow membrane is under study and appears feasible [18].

For both these operations, the choice of filtration process (dead-end, cross-flow or dynamic filtration) and membranes (chemical nature, mean pore diameter) are essential. Over the last two years, a new filtration process call RVF technology (Rotating and Vibrating Filtration – patent n°FR-97-14825) has been studied with microfiltration flat disc membranes. Clarification of rough beer and cold-sterilisation of clarified beer were performed at pilot scale under severe and well-controled operating conditions (temperature close to 0°C, product under CO₂ pressure) in order to satisfy product quality and industrial process requirements. The results show the quantitative (flux, hydraulic resistance, retention) and qualitative (pH, haze, colour, bitterness, polyphenol, protein, dry matter and yeast concentration) performances of the process. Most of these analyses correspond to EBC methods which are widely used in the brewing industry. Analytical profiles of conventional dead-end filtration (brewery reference) and permeate (RVF technology) were compared. The experimental results provided invaluable indications about permeate quality, membrane selection and the most effective operating conditions through a hydrodynamic study of the technology. The main conclusions highlight that RVF technology may be a satisfactory method of rough beer clarification and cold- sterilisation of beer for accurate “membrane – operating condition” choices.

EXPERIMENTS Pilot plant

Experiments were carried out using a pilot plant composed of two major parts: the feed (retentate) and the permeate loops. The retentate loop and feed tank had respectively a tubular heat exchanger and a double-jacketed vessel including temperature regulation (1°C±1°C). The installation was kept under CO₂ pressure (under 150kPa) in order to maintain identical conditions to industrial process. The pressure could reach 300 kPa inside the apparatus and the retentate flow rate 3.0m³.h^-1 in maximum values. Measurements of data were made with the following sensors:

flow-meters, temperature gauges, relative and differential pressure sensors as shown in Figure 1. After calibration, the precision of measurements was better than 1%.

Filtration module and membranes

RVF laboratory module (filtration area 0.048m²) is divided into parallel cells including flat disc membrane fixed onto porous substrate which drain the permeate, and impeller shaped rotating bodies attached to a central shaft. This simple mechanical device runs continuously and maintains a high shear rate as well as a hydrodynamic perturbation at the membrane surface. The impeller is made of three blades (∅ext=142mm, thickness = 8mm) included in a horizontal plane between two membranes. The gap between the blade and the membrane is small and equal to 3mm. Transmembrane

pressure (up to 300kPa) and rotation frequency (up to 50Hz=50rd/sec) can be adjusted to reach the optimal conditions for cross-flow filtration. The two separated cells makes it possible to work with two different membranes in similar conditions.

TM2

Motor

FQP1 FQP2

TERT TPT

PP

PRE

F

T TSR TM1

IMPELLER

MEMBRANE

RETENTATE PERMEATE

QBR, QP1 and QP2 : flow-meter TER, TSR, TP : Temperature gauge PRE and PP : Relative pressure sensor DP, TM1 and TM2 : Differential pressure sensor F : Tachometer

F QPR

DP

Figure 1 : Schematic flow diagram of the RVF technology and instrumentation.

We worked with stainless membranes (coarse membranes) with a ceramic selective layer (TiO2, ZrO2 or SiO2). The membranes used were flat discs with an external diameter of 142mm and a thickness of 0.25mm. Filtration area per membrane is equal to 0.012m² and forms a crown shape with an internal diameter of 62.25mm and an external diameter of 135.5mm (including corrections due to gasket). In this work, a large range of pore diameters (0.15µm, 0,20µm, 0.40µm, 0.60µm, 1.10µm and 1.50µm) used in microfiltration were investigated.

Experimental protocol and measurements

All filtrations were carried out at between 0°C and 2°C to ensure that the beer was representative of that being transferred from cold conditioning to the filter line. A transmembrane pressure of between 20 and 200 kPa and an impeller rotational speed between 10 and 50Hz were investigated during different trials. During the experiment, the permeate was recycled through the feed tank to maintain constant inlet concentration. In this work, three trials were dedicated to the cold-sterilization of clarified beer (MFT n°1, 2 and 3) and four trials to the clarification of rough beer (MFT n°4, 5, 6 and 7). The experiments were made up of two parts. The first part was classic cross-flow microfiltration under constant operating conditions, lasting 3 to 4 hours in order to reach the quasi steady state flux. In the second part, we investigated other conditions (pressure, velocity). In each case we waited for the quasi steady-state flux whatever the time required (5 to 40 min).

Fluids and analyses Experimental fluids

A local brewery (Terken GBM, Roubaix, France) supplied concentrated rough and clarified beers (Lager beer type, equivalent to 15°Plato) and a new beer was used for

each experiment. Rough beer (RB) was taken from the maturation tank and clarified beer (CB) came from the downstream side of a Kieselguhr sheet filter. Specific attention should be paid to experimental fluids. Rough and clarified beers are real products; consequently a large and inevitable variability of product quality is noticeable (from brand to brand, from batch to batch and even within a tank due to stratification) [2, 11, 12, 17]. Table 1 shows the average analysis of the different beers used for the microfiltration runs. The beer composition variability is evident but variations in dry matter, protein and polyphenol concentrations between rough beers and clarified beers are no greater than the differences observed between the rough beers. Thus, some trends can be drawn from the comparison of all the filtration run results, even if the effect of the composition of beers [4] used for different runs can not be eliminated.

Rough beer Clarified beer

Mean σσσσ Mean σσσσ

Protein (Bradford) [mg/l] 270 20 215 30

Protein (Lowry) [g/l] 6.5 0.4 6.0 0.4

Dry matter [%] 5.0 0.2 4.9 0.1

Colour [EBC] / / 10.5 0.2

Carbohydrate [g/l] 40.1 3.5 40.0 4.4

Polyphenol [mg/l] 219 21 202 17

Bitterness [EBC] 24.3 3.1 22.2 1.3

Haze [EBC] 17.8 3.7 0.44 0.04

pH [/] 4.4 0.3 4.2 0.1

Yeast cells [cell/ml] 1.0E⁺⁶ 5.1E⁺⁵ / /

Table 1 : Mean composition of rough and clarified beers used for filtration runs: mean values and standard deviation (σσσσ).

Physical and chemical analyses

The physical properties (density and viscosity) versus temperature have been characterised. Several analyses were carried out to characterise rough and clarified beers before each filtration according to the European Brewery Convention methods.

Analyses were achieved by collecting permeate and retentate samples during experiments. Most of these analyses corresponded to EBC methods which are widely used in the brewing industry (pH, dry matter, haze, polyphenols). Proteins and their derivatives affect much of the brewing process and the quality of beer. As a consequence samples were analysed using two colorimetric methods: the Bradford and Lowry methods. Bradford’s assay is based upon the binding of Coomassie Brillant Blue and has been reported to be optimal for beer protein analysis. This method enables the variation in concentration of high molecular weight nitrogen compounds (MW>5000) to be followed and displays minimal interference. The Lowry method is based on the Biuret reaction and this method may be considered as being complementary to the Bradford one, in that it is sensitive to nitrogen-containing compounds of lower molecular weight.

RESULTS & DISCUSSIONS

Hydrodynamic study of the RVF module

Dynamic or high-shear cross-flow filtration consists in creating a relative motion between the membrane and the impeller in order to produce high shear rates at membrane surface. The main advantage of this system is that it allows operation at a

low transmembrane pressure and high shear rates, a combination that limits the growth and the compression of a cake on the membrane.

One of the widely used theory for modelling flux in pressure independent, mass- transfer-controlled systems is the film theory. Dimensional analysis for mass transfer (analogy to heat transfer) lead to establish a general correlation between the Sherwood (Sh), Schmidt (Sc) and Reynolds (Re) numbers. It makes possible an evaluation of the mass-transfer coefficient, K and provides insight into how membrane geometry and fluid flow conditions can be specified to optimise flux. In laminar flow condition, Lévêque's solution may be used to evaluate the mass-transfer coefficient where the laminar-parabolic velocity profile is assumed. It shows that flux (mass transfer coefficient) may be increase by increasing the velocity or by decreasing the cross- section. In more general terms, any fluid management technique which increases shear rate at the membrane surface will increase flux.

However, RVF technology exhibits a complex hydrodynamic system and the modelling of shear-rate appears difficult. The lack of an accurate model to describe this module drives us to make the following comparison : RVF technology may be defined as (i) the flow of a Newtonian liquid in a duct with a complex shape [7] and as (ii) a stirred vessel with a continuous feed and extract [6]. In the first assumption, we do not have the criteria necessary to choose the characteristic geometrical dimension (d_h, e, w, L) and to define the impact of rotational velocity of the impeller.

In the second assumption, we should take into account the retentate flow-rate and to integrate the shear-rate all along the blades and versus time. Both these analogies are defined by dimensionless numbers relating to fluid flow and mixing (theoretical system) in laminar flow. In each case, experimental measurements (pressure drop, power consumption) enable a dimensionless geometrical parameter (ξ, Kp) to be determine, which will allow the mean shear rate to be quantified.

Firstly, we assume a complex shape duct (Table 2) in which we assume a laminar regime. Figure 2 shows the evolution of pressure drop divided by viscosity versus retentate flow-rate for different rotational velocities. The linear evolution characterises a laminar regime. It enables us to evaluate a slope equivalent to a general geometrical parameter but it does not permit the values of d_h, or L to be identified in order to calculate a Reynolds number.

Dimensionless relation:

ξ

=

⋅Re f

2

Dimensional relation:

S Q d

L µ

P

h

⋅









⋅

⋅ ⋅

∆ =

2

ξ 4

γ µ τ µ

ρ ρ τ

τ

&

h p

h p

p

d Re U

L d P U

f

⋅ =

= ⋅

⋅

⋅

= ∆

= ⋅

, , 4

2 ² In laminar flow

S Q d_h ⋅









⋅ ⋅

=ξ 1

γ&

Table 2 : Relationship between dimensionless numbers and experimental parameters in a complex shape duct.

Secondly, we assume a linear model describing the geometrical parameter versus rotational speed (F) as shown in Figure 3. As a consequence, we propose a linear relationship between the shear rate versus operating condition (retentate flow-rate, rotational velocity). This value may be considered as an estimation of the axial shear- rate but radial shear-rate need to be identified.

0.0E+00 2.0E+06 4.0E+06 6.0E+06 8.0E+06 1.0E+07 1.2E+07

0.0E+00 2.0E-05 4.0E-05 6.0E-05 8.0E-05 1.0E-04 1.2E-04 1.4E-04 Flow-rate [m³/s]

Pressure Drop / Viscosity - DP/µ [1/s]

F=50Hz F=40Hz F=30Hz F=25Hz F=20Hz F=10Hz

Linear increase DP vs Flow rate => Laminar regim

Slope for F=10Hz P [1/m3]

Figure 2 : Evolution of Retentate pressure drop divided by viscosity versus retentate flow-rate for different rotational speeds.

Slope vs F [Hz]

y = 1.56E+09x + 2.01E+10 R2 = 9.86E-01 0.0E+00

5.0E+10 1.0E+11 1.5E+11 2.0E+11 2.5E+11

0 10 20 30 40 50 60

Frequency F [Hz]

Slope [1/m3]

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

(DP/DP0)=(γγγγ/γγγγ0) Slope : P vs F

Reduced shear-rate [/]

Modelling of reduced shear-rate versus frenquency (F) for a given flow-

rate (Q)

S Q d

L µ

P

h

⋅









⋅

⋅ ⋅

∆ =

2

ξ 4

with A B F

S d

L

h

⋅ +

=









⋅

⋅ 4₂⋅

ξ

then

( )



 





⋅ ⋅

⋅

⋅ +

= L

Q d F B

A ^h

tot 4

γ&

Figure 3 : Evolution of the geometrical parameter versus rotational speeds and modelling of reduced shear-rate.

Clarification of rough beer

The quantitative performances of CFMF of rough beer correspond to the experimental steady-state flux versus operating conditions and mean pore diameter. The flux values confirm traditional levels of performances according to the operational variables. We can summarise them as follows: (i) flux is closely correlated to mean pore size, (ii) flux increases in a non-linear fashion with transmembrane pressure (Figure 5) and with shear-rate (Figure 4). We can clearly distinguish the membrane with a high mean pore diameter (superior to 1µm) which exhibits a flux value of above 100 l.h^-

1.m^-2 whereas small mean pore diameters remain inferior to 80 l.h^-1.m^-2. Flux values appear to be close to those mentioned in existing literature [3, 12, 15].

The chemical analyses and retention rate were determined for the following elements:

(i) beer haze, (ii) colour, (iii) bitterness, (iv) polyphenols, (v) carbohydrates dry matter, (vi) pH and (vii) proteins. Table 3 and 4 sum up the mean composition of permeate and retention rate versus mean pore diameter. We note that the permeate quality and retention rate evolve with the nominal pore size. Permeate quality becomes acceptable with 1.10µm and 1.50µm membranes even for haze. In this case the global retention rate is inferior to 10% which means that there is not an excessive loss of essential compounds. Yeast cell retention is close to 100% and residual concentration is inferior to 10cell/ml.

0 50 100 150 200 250 300

0.0E+00 2.0E+06 4.0E+06 6.0E+06 8.0E+06 1.0E+07 1.2E+07 (A+B.N).Q [1/s]

Flux [l.h-1.m2]

MFT n°5: 1.10µm, TMP=1400mBar MFT n°5: 1.50µm, TMP=1400mBar MFT n°4 : 0.60µm, TMP=2750mBar MFT n°4 : 0.80µm, TMP=2900mBar

0 50 100 150 200 250 300

0 500 1000 1500 2000 2500 3000 3500

Transmembrane pressure [mbar]

Steady-state flux [l.h-1.m-2]

MFT n°4 : 0.60µm, F=50Hz MFT n°4 : 0.80µm, F=50Hz MFT n°6 : 1.10µm, F=50Hz MFT n°7 : 1.10µm, F=50Hz MFT n°6 : 1.50µm, F=50Hz MFT n°7 : 1.50µm, F=50Hz

Figure 4 : Steady-state flux versus







 ⋅ ⋅

dh

L

γ& 4 and mean pore diameter.

Figure 5 : Steady-state flux versus transmembrane pressure and mean pore

diameter with F=50Hz.

Cold-sterilisation of beer

Qualitative performances during cold-sterilisation of clarified beer indicate the same level of performances than for rough beer. With 0.60µm membrane, flux is slightly superior to rough beer (common value inferior to 100l.h^-1.m^-2). We observed a regular increase of flow rate with an increase in mean pore diameter for similar operating conditions. This remark is also valid for qualitative and retention rates as shown in Table 3 and 4. A critical pore size close to 0.50µm appears. Below 0.50µm, the retention rates are too great which will induce a downgraded beer. In addition we check whether high molecular weight nitrogen compounds (MW>5000) suffer from a high retention rate with 0.15µm and 0.20µm membranes.

Cold-sterilisation of CB Clarification of RB Mean pore diameter [µm] 0.15 0.20 0.40 0.60 0.60 0.80 1.10 1.50 Protein (Brad.) [mg/l] 99 89 163 160 192 212 251 252 Protein (Lowry) [g/l] 3.69 4.05 4.73 5.02 5.78 5.87 5.92 5.96 Dry matter [%] 3.07 3.19 4.24 4.28 4.10 4.29 4.61 4.59 Colour [EBC] 6.7 6.8 8.7 9.1 8.4 8.8 9.8 9.7

Carbohydrate [g/l] 24 26 32 34 35 35 39 38

Polyphenol [mg/l] 140 146 181 180 170 185 184 190 Bitterness [EBC] 17.5 18 20 20.5 20 20.5 21 20.6 Haze [EBC] 0.26 0.23 0.33 0.30 0.35 0.38 0.59 0.55 pH [/] 4.24 4.29 4.29 4.25 4.38 4.37 4.36 4.37

Yeast cells [cell/ml] / / / / 0 0 1-10 1-10

Table 3 : Mean composition of permeate during (i) the cold-sterilisation of clarified beer and (ii) the clarification of rough beer versus mean pore diameter.

Cold-sterilisation of CB Clarification of RB Mean pore diameter [µm] 0.15 0.20 0.40 0.60 0.60 0.80 1.10 1.50 Protein (Brad.) [mg/l] 56.2 60.6 15.5 14.4 24.7 16.9 5.8 5.4 Protein (Lowry) [g/l] 39.2 33.3 14.5 10.9 15.7 14.4 6.5 5.9 Dry matter [%] 37.0 34.6 13.4 12.3 19.4 15.7 4.2 4.7

Colour [EBC] 35.4 34 15.5 12.9 / / / /

Carbohydrate [g/l] 34.8 27.9 11.9 10.3 10.3 10.3 7.9 8.1 Polyphenol [mg/l] 33.5 30.6 11.3 10.7 30.9 24.8 16.4 13.5

Bitterness [EBC] 19.5 17.2 11.1 9.9 20.0 18.0 5.7 7.3 Table 4 : Mean retention rate composition of permeate during (i) the cold-sterilisation of clarified beer and (ii) the clarification of rough beer versus mean pore diameter.

CONCLUSIONS

Potential applications of microfiltration in the beer industry are clarification (elimination of yeast cells and suspended matter) and cold-sterilisation. If the objective of the filtration is clarification, large pore membranes (superior to 1µm) should be used because of the higher permeate flow rates and the low retention of essential beer compounds. We observed that 1.10µm and 1.50µm may achieve these criteria in presence of a high shear rate due to the rotational velocity of impeller. If the objective of the filtration is pasteurisation, this study shows that no membrane can satisfy flux requirement and beer quality criteria at the same time. Moreover, the permeate flux obtained with these membranes is still too low to make this technique economically viable. The cake layer erosion due to hydrodynamic perturbation seems to improve membrane selectivity in regard with retention rate versus mean pore diameter. Finally, we should understand and model the hydrodynamics of this system in relation to permeate flux and quality. The knowledge of the average shear-rate will permit to appreciate its impact on filtration performances.

NOMENCLATURE :

A geometrical constant, [m^-3] TMP Transmembrane pressure, [Pa] or [mbar]

B constante, [s.m^-3] U mean velocity, [m/s]

dh hydraulic diameter, [m] S cross section, [m²] f Fanning friction coefficient, [/] Sh Sherwood number, [/]

F Rotation frequency, [Hz] or [rd/s] Greek letters

L length, [m] ξ dimensionless geometrical parameter, [/]

P power consumption, [W] µ viscosity, [Pa.s]

Q flow rate, [m³.s^-1] ∆^P pressure drop, [Pa]

Re Reynolds number, [/] γ& shear rate, [s^-1] τp wall shear stress, [Pa]

REFERENCES :

1. Blanpain P., Fillaudeau L., Lalande M., Trans IchemE, 77, Part C, (1999), 75-89.

2. Burrell K., Gill C., McKetchnie M., Murray J., MBAA Technical Quarterly, 31, (1994), 42-50.

3. Burrell K.J., Reed R.J.R., Filtration & Separation, 31 (4), (1994), 399-405 4. Czekaj P., Lòpez F., Güell C., J. Membrane Sci., 166 (2), (2000), 199-212.

5. Daufin G., René F, Aimar P., Ed. Lavoisier Tech & Doc, Paris (France), ISBN 2-7430-0228-X, (1998), 592p.

6. Delaplace G., PhD thesis, Université de Nancy I (France), (1998), 148p.

7. Delplace F., Delaplace G., Lefebvre S. and Leuliet J.C., 7^th International Congress on Engineering and Food, Brighton (UK), (1997), 36-39.

8. Dickmann H., Neradt F., Brewer's Guardian, June 1995, (1995), 27-30.

9. Dunn A.E., Leeder G.I., Molloy F., Wall R., Ferment, 9 (3), (1996), 155-161.

10. Fillaudeau L., Brauwelt International, 17 (5), (1999), 414-421.

11. Fillaudeau L., PhD thesis, Université de Compiègne (France), (1998), 150p.

12. Gan Q., Field R.W., Bird M.R., England R., Howell J.A., McKetchnie M.T., O'Shaughnessy C.L., TransIChemE, 75, Part A, (1997), 3-8.

13. Gaub R., Brauwelt International, (5), (1993), 448-457.

14. Leeder G., Brauwelt International, (4), (1993), 372-373.

15. McKetchnie M.T., Burrell K.J., Gill C., Kotzian R., O'Sullivan P., Food Process Engineering, University of Bath (UK), (1994), 47-53.

16. O’Reilly S.M.G., Lummis D.J., Scott J., Molzahn S.W., Proceeding of the 21^st EBC Congress, Madrid (Spain), (1987), 639-647.

17. O'Shaughnessy C.L., Durosinmi-Etti O., Proceeding of the 26^th European Brewery Convention, Masstricht (The Netherlands), (1997), 681-690.

18. Stewart D.C., Hawthorne D., Evans D.E., J. Inst.Brewing, 104 (6), (1998), 321-326.

19. Wackerbauer K., Evers H., Brauwelt International, (2), (1993), 128-133.