)JJIGXWSJβ+PYGER7LIEVMRKERH)RZMVSRQIRXEP *EGXSVWSRXLI8YVFMHMX]SJ;SVXERH&IIV R. Alex Speers 1,2, Yu-Lai Jin 1, Allan T. Paulson 1 and Robert J. Stewart 3 %&786%'8
J. Inst. Brew. 109(3), 236–244, 2003 The presence of added ␤-glucan in wort caused increased turbidity levels, which increased at higher molecular weights and concentrations of the polymer. Levels of pH, maltose and ethanol, and shear experienced in a brewery also influenced the turbidity of wort and beer. Haze levels of beer after 0.45 µm membrane filtration were found to decrease due to the removal of non-␤glucan particles. Cold storage at 4°C for two weeks was found not to lower the turbidity caused by high concentrations of high molecular weight ␤-glucan polymers. Key words: Beer, beta-glucan, haze, turbidity, wort.
-2863(9'8-32 The first two assessments of beer quality are the quality of its head and its brilliance or clarity. Clarity is usually measured by light scattering of beer at a certain wavelength. The majority of the instruments for haze determination are designed to measure turbidity of light scattered at 90° to the incident light at wavelengths of 450–860 nm 17. Although several polymers are available to calibrate wort and beer turbidity, formazin is used in ASBC and EBC standard methods 1,6. However, these brewing organizations have defined formazin turbidity with different scales. One EBC Formazin Haze Unit (FHU) is equivalent to 69 ASBC Formazin Turbidity Units (FTU). The ASBC turbidity scale was used in this study. The clarity of beer can be categorized into five levels: brilliant (276 FTU) 6,9. A visual haze threshold has recently been reported to be 0.38–0.82 nephelos turbidity units (equivalent to 6.6–14.1 FTU) determined by 90° scattering of white light 21. Beer turbidity can be caused by microbial cells or colloidal particles of biological and mineral origin. The microbiological stability of beer is easily achieved by either pasteurization or by sterile filtration. In order to achieve col-
of Food Science and Technology, Dalhousie University, 1360 Barrington Street, D401, Halifax, NS B3J 2X4 Canada. Phone: +1(902) 494-3146. Fax: +1(902) 420-0219. 2 Corresponding author. E-mail: [email protected]
3 Labatt-ITW Technology Department Americas, 197 Richmond Street, London, ON Canada N6A 4M3. Present address: Ocean Nutrition Canada Ltd., 1721 Lower Water Street, Halifax, NS B3J 1S5, Canada. 4YFPMGEXMSRRS+ 8LI-RWXMXYXI +YMPHSJ&VI[MRK
loidal stability, non-microbial hazes (and their precursors) must also be removed to prevent haze formation after beer packaging. It should also be noted that beer hazes present at both warm and cold temperatures are referred to as ‘permanent hazes’ whereas those that appear only at refrigerated temperatures are termed ‘chill hazes’. Some might argue that the prevention or removal of non-microbiological hazes poses the largest challenge to the industry. The removal of haze precursors (e.g., haze-active proteins and haze-active polyphenols) is one of the techniques used to prevent the development of protein-polyphenol hazes in packaged beer. Colloidal haze particles originating from proteins and polyphenols have been recently reviewed by Bamforth 3 and Siebert 20. Malt-derived hazes usually contain high levels of ␤glucans in addition to proteins, polyphenols and pentosans 4,7,10,11,14,16,23,25,27. Gelatinous particles of barley ␤-glucans dispersed in beer can also instantly clog beer filters 13. Under certain circumstances such as high pressure or elevated filtration temperatures, these polymers pass through the filters and aggregate later in the packaged beer leading to formation of problem hazes 7,27. Since the 1960s, it has been recognized that low ␤-glucan levels are necessary to avoid the risk of precipitates in packaged beer 5. The obvious solution to ␤-glucan hazes, the use of ␤-glucanases, is restricted in many markets due to potential consumer concerns. Even though ␤-glucans may not always form gelatinous precipitates in beer, they can still cause haze problems. Particles below 0.1 µm scatter substantial light 90° to the incident beam and result in so-called “pseudo-” or “invisible” hazes 2,3,11,15,26. When added to beer, ␤-glucans having molecular weights (MWs) of 31–443 kDa are primarily 0.01–0.1 µm in apparent diameter 12. The purpose of this study was to investigate how ␤-glucan MWs and concentrations affect the turbidity of wort and beer. The effects of shearing, shearing temperature, pH, maltose content of the wort and ethanol content of beer on turbidity were also examined.
1%8)6-%07%2(1)8,3(7 ;SVXWXYHMIW A “low ␤-glucan” Harrington pale malt (i.e., producing a Congress wort of 100 mg/L ␤-glucan) was used to prepare wort using the EBC mashing temperatures 1. The mash was filtered and sparged with hot water followed by boiling for 1 hour to allow 25% (v/v) evaporation. The hot wort was filtered through Whatman No. 1 filter paper to remove hot trub. Nine litres of 16.9°P wort was collected.
This wort was further treated with a commercial ␤-glucanase (Filtrase BTM, containing at least 180 U/g with maltodextrin as carrier, DSM Food Specialties, Seclin, FRA) to reduce its ␤-glucan to an undetectable level by Congo red assay 12. This ␤-glucan-free wort was dispersed into 1 L Erlenmeyer flasks and covered with 4 layers of aluminum foil. The wort was autoclaved at 121°C for 15 min in order to inactivate the remaining ␤-glucanase. The wort was then filtered through a Kimax Büchner “M” funnel (Fisher Scientific Co. Ltd., Nepean, ON) with 1.0 g/100 mL of diatomaceous earth (non-washed, product No. D-5509, Sigma-Aldrich Canada Ltd., Oakville, ON) as a filter aid to remove coagulated and suspended particles. After such a filtration, the wort sample was brilliant and thus the interference on ␤-glucan turbidity due to other large haze particles was minimized. The clarified wort (SG = 1.0695 at 20°C) was preserved with 100 mg/L of sodium azide (NaN3 ) and stored at room temperature until used (up to 16 weeks) without any microbial growth or haze development by microscopic and visual observations. Barley ␤-glucans (purity ≥ 97% on dry basis) purchased from Megazyme International Ireland Ltd. (Bray, IRL) with reported molecular weights of 31, 137, 250, 327 and 443 kDa (Lot numbers 41101, 90401, 60501, 40301 and 90501, respectively) were dissolved in double distilled and deionized water (DDW) containing 100 mg/L of NaN3 to make 0.500 g/100 mL stock solutions. The pH value of the high gravity wort (16.9°P) was 5.28 and was adjusted to pH 5.40 by adding 0.5 mL of 1 N NaOH per liter of wort. An adequate amount of ␤-glucan stock solution was mixed with the pH 5.4, ␤-glucan-free wort (16.9°P), and DDW (whose pH was adjusted to 5.4) to prepare ␤-glucan solutions of 0, 50, 100, 200, 400, 600, 800 and 1000 mg/L in 12.0°P wort at pH 5.4. Duplicate wort samples (at five MWs and seven concentrations) were sheared. A Lourdes blender (Model MM-1B, Lourdes Instrument Corp., Brooklyn, NY) was employed to shear wort and beer samples (10 mL) for 35 s at a speed setting of “60”, approximately equivalent to an average shear rate of 1.3 ± 0.2 × 10 4 s–1 which was calculated from the power dissipation of the liquid 18. To minimize the occurrence of wort and beer oxidation, the headspace of the blender cup was flushed with N2 for 10 s prior to shearing. After shearing, samples were transferred into 50 mL plastic centrifuge tubes and flushed with N2 for 10 s followed by sealing with caps. Sheared and unsheared wort samples were then examined for their turbidity. Differing shear temperatures were achieved by equilibrating wort samples at the required temperature for 15 min and maintaining the jacketed shear cell at the same temperature with a circulator. To investigate the effect of ␤-glucans at various pH values and maltose levels, the 16.9°P, ␤-glucan-free wort and a 443 kDa (the highest MW available) ␤-glucan solution (0.500 g/100 mL) were used to prepare 8°P wort containing 600 mg/L of ␤-glucan. Maltose (Sigma Chemical Co., St. Louis, MO) was incorporated into the 8°P wort at 4.0% and 10.0% (w/w) of dry matter to prepare 12°P and 18°P worts, differing only in maltose content from the 8°P wort. The maltose content of the 8°P wort was 6.1% (w/w) as determined by Fehling’s method 1. Therefore, the 8°P, 12°P and 18°P worts contained 6.1%, 10.1% and 16.1% (w/w) of maltose, respectively. Each wort sample was divided into three
equal portions and the pH value was adjusted from 4.95 to 4.0 (with 5.40 mL of 1.0 N HCl per liter of wort), 5.4 (with 0.16 mL of 1 N NaOH per liter of wort) and 6.8 (with 9.60 mL of 1 N NaOH per liter of wort), respectively. The ionic strengths of worts differed between 4.2– 9.4 mM, due to pH adjustment. This minor difference in ionic strength was assumed not to affect experimental results. Adequate amounts of 1 M sodium azide solution were incorporated to adjust the final concentration of NaN3 to 100 mg/L. Measurements for each sample were completed within 5 days after sample preparation although no microbial growth was observed by microscopic examination for at least 6 months.
&IIVWXYHMIW A commercial lager beer of ethanol content 5% (v/v), obtained from Oland Brewery Limited (Halifax, NS) was treated with a ␤-glucanase and used as the “beer base”. The real extract was determined to be 3.3% (w/w) 1. The beer was first degassed by filtering through Whatman No. 1 filter paper under a vacuum of 20 mm Hg. The degassed beer (25 L) was then boiled for 2 hours to remove ethanol and other volatile components (which resulted in 55 g of wet precipitate). The concentrated beer was cooled and kept at 50°C for 30 min after the addition of lichenase at 0.1 U/mL (Megazyme, Bray, IRL) to hydrolyze ␤-glucan. The residual ␤-glucans were undetectable by Congo red dye. The enzyme was then inactivated by autoclaving of concentrated beer in Erlenmeyer flasks covered with 4 layers of aluminum foil at 121°C for 15 min. The concentrate, having an extract content of 7.14% (w/w), was kept at room temperature until used (up to 2 months). There were no stability problems observed during storage. Obviously, the boiling process followed by sterilization caused denaturation and precipitation of some beer components such as haze precursors. This heat treatment also provided a lower level of turbidity in the control beer sample. Beta-glucans (of 31, 137, 250, 327 and 443 kDa) were dissolved in DDW at a concentration of 0.300 g/100 mL. Degassed “beer” samples were prepared by mixing adequate amounts of the ␤-glucan stock, beer concentrate, anhydrous ethanol and DDW. To investigate the effect of MW and concentration of ␤-glucans on beer turbidity, ␤-glucans (MWs of 31, 137, 250, 327 and 443 kDa) were included at concentrations of 0, 50, 100, 200, 400, 600, 800 and 1000 mg/L in beer, which contained 5.0% (v/v) ethanol and 3.3% (w/w) real extract at pH 4.2. Beer samples were also subjected to shearing and their turbidity examined. Beer samples containing 600 mg/L of 443 kDa ␤-glucan were prepared to investigate the effect of pH (3.8, 4.2 and 5.4) and ethanol (0, 5 and 10% (v/v)) on beer turbidity. These samples were also subjected to shearing at different temperatures (0, 5 and 10°C). Pre-equilibrated beer samples at desired temperatures for 15 min were sheared in a jacketed shear cell, which was equilibrated at the same temperature with a circulator. Freshly prepared samples were stored at 5°C and analyzed within 3 days. All samples were stable until used. The turbidity of wort and beer was determined by 90° light scattering at 580 nm. This technique was based on and modified from the official ASBC method 1. Three milliliters of sample were placed in a 10 × 10 mm four-sided :3023
clear styrene cuvette (Fisher Scientific Co. Ltd., Nepean, ON). The intensity of the 90° scattered light was measured with an LS 50 spectrophotofluorimeter (Perkin-Elmer Ltd., Beaconsfield, Buckinghamshire, GBR) with both the excitation and emission wavelengths set at 580 nm. Readings of scattered light at a 90° angle (I580 ) were then calibrated with formazin turbidity standards and expressed as formazin turbidity units (FTU) 1. Double distilled and deTable I. Experimental design summary. Sample
Shearing MW of ␤-glucan Concentration of ␤-glucan ␤-Glucan (443 kDa) Shearing temperature pH Maltose
Sheared vs. unsheared 31–443 kDa 50–1000 mg/L 0, 600 mg/L 20, 48, 76°C 4.0, 5.4, 6.8 6.1, 10.1, 16.1% (w/w)
Shearing MW of ␤-glucan Concentration of ␤-glucan ␤-Glucan (443 kDa) Shearing temperature pH Ethanol Cold storage
Sheared vs. unsheared 31–443 kDa 50–1000 mg/L 0, 600 mg/L 0, 5, 10°C 3.8, 4.2, 4.6 0, 5, 10% (v/v) 4°C for 2 weeks
(12°P) at pH 5.4. at pH 4.2 containing 3.3% (w/w) of real extract and 5.0% (v/v) of ethanol.
ionized water (DDW) was filtered through a 0.45 µm “Acetateplus” plain membrane (Cat. No. A04sp02500, Batch No. 088439; Osmonics Inc., Minneapolis, MN) before being used to prepare the formazin haze standards. A mixture of 10.0 mL of 1% hydrazine sulfate (Sigma Chemical Co., St. Louis, MO) and 10.0 mL of 10% hexamethylenetetramine (Sigma Chemical Co.) was stirred for 24 h at room temperature to form stable haze particles. This colloidal dispersion had a haze level of 69,000 FTU and was diluted to 6,900 FTU. One mL of the 6,900 FTU suspension was then diluted to 50.0 mL with DDW to make a 138 FTU standard (equivalent to 2 EBC haze units). Standards of 0, 13.8, 27.6, 41.4, 55.2, 69.0, 82.8, 96.6, 110.4, 124.2 and 138.0 FTU were further prepared to obtain a calibration curve of the intensity of scattered light at 580 nm (I580 ) versus FTU (r 2 ≥ 0.99). Duplicate experiments were carried out and results are given as mean values ± one standard deviation. Statistical analyses of the data were undertaken with SYSTAT version 5.05 (SPSS Inc., Chicago, IL). The General Liner Model was used to perform a forward stepwise regression with ␣-to-enter = 0.15 and ␣-to-remove = 0.15. Each of the terms in a model had p ≥ 0.05. The overall p-values of the models are reported. Models with and without a constant were compared and the better fit with a higher determination coefficient (i.e., R2 – adjusted) was chosen to be reported. Experimental conditions are listed in Table I.
Fig. 1. Effect of MW and concentration of ␤-glucans on the turbidity of (A) unsheared and (B) sheared wort at 20°C.
6)79087%2((-7'977-32 )JJIGXSJWLIEVMRK1;ERHGSRGIRXVEXMSR SJbKPYGERWSR[SVXXYVFMHMX] The ␤-glucan-free wort (12°P, pH 5.4) initially had a very low turbidity level of 22.2 FTU (Fig. 1a). The addition of ␤-glucans increased wort turbidity at higher MWs and concentrations (p < 0.001). The increase in turbidity was proportional to the increase in MW and concentration 2 (R – adj = 0.851; n = 80; p < 0.001). Beta-glucan complexes of higher apparent MW and concentration have 8,12 larger apparent sizes , which scatter more light at 90°. However, wort samples were all brilliant since the turbidity values were lower than 35 FTU (Fig. 1a). After being sheared at 20°C, wort samples exhibited higher turbidity values (p < 0.001; Fig. 1b). The ␤-glucan-free wort had a turbidity of 33.8 FTU after shearing in contrast to 22.2 FTU with the unsheared sample. The turbidity of wort was found to increase with MW and concentration of ␤-glucans (p < 0.001). Wort turbidity level can be described by the following model (R2 – adj = 0.954; n = 160; p < 0.001): FTU = 23.46 + 3.02 × 10–3 MW + 4.79 × 10–3 C + 7.251 S + 0.0318 S × C + 6.236 × 10–6 MW × C
the majority of them have pI values around 5 according to Sørensen and Ottesen 24. Most of the proteins in wort samples would have more net positive charges at pH 4.0 than at pH 5.4 (net negative charges), and more negative charges at pH 6.8 than at pH 5.4. Shearing of wort may have possibly enhanced the unfolding of protein molecules and the exposure of hydrophobic or charged groups. At pH 4 (lower than pI), the net positive charge of proteins in-
where FTU is the haze in formazin turbidity units; MW is the molecular weight (kDa) of ␤-glucans; C is the concentration of ␤-glucans (mg/L); and S is shearing (S = 0 for unsheared wort and S = 1 for sheared wort); S × C and MW × C are the interactions of concentration with shearing and MW, respectively. The increased turbidity by shearing was previously explained by an increase in ␤glucan particle size after shearing 12. This result confirmed previous reports from this laboratory that noted ␤-glucan turbidity increases after shearing 12,18,19.
)JJIGXSJbKPYGERO(EEXQK0 QEPXSWIPIZIPT,ERHWLIEVMRKXIQTIVEXYVI SR[SVXXYVFMHMX] Control (unsheared) wort samples containing 0 and 600 mg/L of a 443 kDa ␤-glucan at pH 4.0–6.8 and 6.1–16.1% (w/w) of maltose as well as sheared samples at 20, 48 and 76°C were examined for their turbidity at 20°C. The presence of 600 mg/L of 443 kDa ␤-glucan increased wort turbidity (p < 0.001; Fig. 2). This supports the results reported in Fig. 1. Turbidity values decreased at higher maltose levels and pH values (p < 0.001). Higher maltose concentrations may have inhibited the interchain association of ␤-glucan and protein-polyphenol molecules. Previously, sucrose at 200 mg/L has been shown to cause a decrease in turbidity formed by protein and polyphenol (at pH 4.2) 22. It has also been found that maltose at high concentrations lowered the fraction of ␤-glucan particles >0.01 µm formed after shearing at 48°C and 76°C 12. The observed higher turbidity at lower pH (p < 0.001) was hypothesized to be caused by proteins. The highest turbidity level produced by protein and polyphenol interactions has been achieved at pH 4.1–4.2 in a model system 22. One could argue that the effect of pH on wort turbidity was related to the electrostatic properties of proteins. Although the isoelectric point (pI) of beer proteins vary from 4–10,
Fig. 2. Effect of ␤-glucan (443 kDa at 600 mg/L), pH and shearing temperature on the turbidity of wort (20°C) at maltose levels of (A) 6.1%; (B) 10.1% and (C) 16.1% (w/w). N20, N48 and N76 represent wort samples containing no ␤-glucan sheared at 20, 48 and 76°C, respectively; B20, B48 and B76 represent wort samples containing 600 mg/L of 443 kDa ␤-glucan sheared at 20, 48 and 76°C, respectively.
creases the intermolecular electrostatic repulsion. At pH 5.4 (close to the pI of beer proteins), protein molecules possess less net polarity charges and thus have less repulsive force. At wort pHs above the pI (i.e., 6.8), proteins are more negatively charged. As the ␤-glucan-free wort had an increased turbidity after shearing, it is possible that protein contributed to the observed haze. Shearing temperature also affected the turbidity of wort measured at 20°C (p < 0.001). Shearing wort at higher temperatures decreased the turbidity of worts at all pHs and maltose concentrations (Fig. 2). This shearing at higher temperatures resulted in a smaller apparent particle size of the 443 kDa ␤-glucan 12. A model was found to describe the changes in turbidity of sheared worts (R2 – adj = 0.772; n = 108; p < 0.001): FTU = 419.5 + 0.292 C – 13.291 Mal + 0.0453 Ts × Mal + 0.156 Ts × pH + 1.781 Mal × pH –1.764 S × Ts – 56.830 pH – 3.460 × 10–3 C × Mal
where Mal is maltose level % (w/w) and Ts is the shearing temperature (°C). Wort turbidity was lower after shearing at higher temperatures (p < 0.001). When sheared at a low temperature (20°C), the interchain interactions may have promoted linkages leading to higher turbidity values. This was evidenced by the lower reactivity of ␤-glucan polymers with Congo red dye when the samples were sheared at 20°C (data not shown). When sheared at high tem-
peratures such as 76°C (which presumably exposed more Congo red reactive groups), the ␤-glucan molecules may be dispersed and forced into an extended conformation because of the disruption of hydrogen bonds, resulting in lower turbidity values. It is worth noting that wort samples at pH 6.8 were all brilliant in clarity despite the effects of the physical conditions discussed above (Fig. 2). Worts at pH 5.4 (a “normal” value for a production wort), were brilliant whereas wort containing 600 mg/L of 443 kDa ␤-glucan became slightly hazy after being sheared at 20°C. Wort samples at pH 4.0 had much higher turbidity than at pH 5.4 and 6.8. At pH 4.0, the ␤-glucan-free worts were slightly hazy while worts containing 600 mg/L of 443 kDa ␤-glucan became hazy with turbidity values as high as 350 FTU (Fig. 2). Hazes formed in the ␤-glucan-free worts probably consisted of proteins whereas the hazes in worts containing 443 kDa ␤-glucan were derived from both proteins and ␤-glucans.
)JJIGXSJWLIEVMRK1;ERHGSRGIRXVEXMSR SJbKPYGERWSRFIIVXYVFMHMX] Similar to the turbidity of wort, the ␤-glucan-free beer was brilliant (20 FTU). The addition of 31-443 kDa ␤glucans increased beer turbidity (p < 0.001; Fig. 3a). Turbidity significantly increased with both MW and concentration of ␤-glucans (p < 0.001). Although ␤-glucans in beer resulted in higher turbidity, the samples were still in a
Fig. 3. Effect of MW and concentration of ␤-glucans on the turbidity of beer (A) unsheared and (B) sheared at 5°C. .3962%03*8,)-278-898)3*&6);-2+
satisfactory range of clarity (brilliant and almost brilliant as practiced by the brewing industry). It is noteworthy that the 250 kDa ␤-glucan shows unexpected high levels of turbidity in both wort and beer (Figs. 1a and 3a). This behavior could be a result of the variation in impurity compositions and/or the difference among ␤-glucan samples in their polydispersity although these species had highest purities commercially available. After being sheared at 5°C, turbidity increased significantly (p < 0.001). Turbidity of the ␤-glucan-free beer was increased from 20 to 59 FTU after shearing. Beer samples containing ␤-glucans became slightly hazy after shearing. Macromolecules such as proteins and ␤-glucans in beer were presumably able to “unfold” during shearing thus allowing aggregation and the formation of larger particles, which scatter more light. Shearing was found to increase the apparent particle size of the 443 kDa ␤-glucan molecules in beer 12. Turbidity values of the sheared beer samples increased with higher MWs and concentrations of ␤-glucans (p < 0.001). The increase in beer turbidity by molecular weight, concentration and shearing can be described by the following model (R2 – adj = 0.987; n = 160; p < 0.001): FTU = 0.142 MW + 0.0632 C + 55.58 S + 0.0254 S × C – 2.5 × 10–4 MW 2 – 4.0 × 10–5 C 2
temperature did not affect beer turbidity (p > 0.05) over the narrow temperature range studied (0–10°C).
'PEVMJ]MRKFIIV[MXL£QQIQFVERIJMPXVEXMSR Sheared beer samples (10 mL) were filtered through 0.45 µm membranes in order to examine the decrease in turbidity after membrane filtration (Fig. 5). The 0.45 µm membrane filtrates had lower turbidity (p < 0.001) com-
)JJIGXSJbKPYGERO(EEXQK0 T, IXLERSPGSRXIRXERHWLIEVMRKXIQTIVEXYVISR FIIVXYVFMHMX] The presence of 600 mg/L of 443 kDa ␤-glucan increased beer turbidity because this colloid had an apparent particle size of 0.01–0.1 µm that is able to effectively scatter light 12. This result supports the findings presented in Fig. 3. At pH 4.2–4.6, beer samples were more turbid (p < 0.001; Fig. 4) than at pH 3.8. In theory, pure ␤-glucan polymers are neutral in charge and their intermolecular interactions are not affected by pH. However, the commercial ␤-glucans contained a small amount of protein (0.72–9.35%)12 although it is unclear how ␤-glucans associate with these proteins. As the majority of beer proteins have pI values around pH 5, pH 4.2–4.6 may have led to fewer net charges on the proteins than pH 3.8, resulting in less repulsion among proteins. The addition of ethanol at 5% (v/v) decreased the turbidity (p < 0.001; Figs. 4a and 4b). However, the addition of 10% (v/v) of ethanol led to higher turbidity levels than 5% (v/v) ethanol (p < 0.001; Fig. 4c). This finding is similar to the effect of ethanol on protein-polyphenol haze formation in model systems studied by Siebert et al.22. The decrease in turbidity caused by 5% (v/v) of ethanol was hypothesized to be due to weaker hydrophobic interactions between protein molecules caused by ethanol 22. The increased turbidity at high ethanol levels, however, was presumed to be due to ethanol lowering the dielectric constant of the medium, leading to lower protein solubility. Among the sheared treatments, turbidity can be described by the pH, ethanol and β-glucan concentration (R2 – adj = 0.965; n = 108; p < 0.001): FTU = 0.0253 C + 15.088 pH – 0.868 E
where E is the ethanol content of beer (% (v/v)). Shearing
Fig. 4. Effect of ␤-glucan (443 kDa at 600 mg/L), pH and shearing temperature on turbidity of beer containing ethanol at (A) 0%; (B) 5.0% and (C) 10.0% (v/v). N0, N5 and N10 represent beer samples containing no ␤-glucan sheared at 0, 5 and 10°C, respectively; B0, B5 and B10 indicate beer samples containing 600 mg/L of 443 kDa ␤-glucan sheared at 0, 5 and 10°C, respectively.
pared to that of sheared beer as shown in Fig. 3b. The reduction in turbidity was mainly due to the removal of hazes of non-␤-glucan origin (Figs. 3b and 5). The turbidity of the ␤-glucan-free beer dropped from 59 FTU to 13 FTU after filtration whereas the effects of varying molecular weights and concentrations of ␤-glucan on turbidity were similar to that of unfiltered beer (Fig. 3b; Eq. 4). The turbidity of sheared beer before and after membrane filtration can be described by the following relationship (R2 – adj = 0.921; n = 480; p < 0.001): FTU = 72.280 + 0.0219 MW + 0.0325 C – 72.42 MF
where MF is a 0.45 µm membrane filtration “dummy variable” (MF = 0 for un-filtered beers and MF = 1 for filtered samples). It is assumed that the 0.45 µm membrane filtration did not remove the ␤-glucan turbidity completely although the hazes of non-␤-glucan origin (presumably proteins) were reduced remarkably after filtration. It is notable that filtrate of beers containing the 443 kDa ␤glucan had lower turbidity than 250 kDa and 327 kDa ␤glucan (Fig. 5) because the 443 kDa ␤-glucan had larger particle sizes 12. The larger ␤-glucan particles were easily retained by the 0.45 µm membranes.
'LERKIWSJFIIVGPEVMX]HYVMRKGSPHWXSVEKI The sheared beer samples were stored at 4°C for two weeks to simulate the clarification of beer during lagering, which is usually achieved at 0 to –1°C in the breweries. The turbidity of the ␤-glucan-free beer decreased from 59 FTU to 24 FTU during storage (Fig. 6). Turbidity decreased at low ␤-glucan MWs of 31 and 137 kDa and low concentrations of 50–200 mg/L. However, turbidity increased with beers containing 250–443 kDa ␤-glucans at 400–1000 mg/L after cold storage, leading to higher haze levels (Fig. 6). A multiple linear regression analysis indicated that the increase in beer turbidity after cold storage responded in a quadratic fashion to increasing ␤-glucan MWs and concentrations (Eq. 6). The chemical interactions between lagering, ␤-glucan MW and concentration may in part explain the observed changes in beer turbidity after the cold storage (R2 – adj = 0.960; n = 160; p < 0.001): FTU = 0.434 MW + 0.218 C + 4.1 × 10–4 Lg × MW × C – 6.8 × 10–4 MW 2 – 1.3 × 10–4 C 2 – 0.0611 Lg × MW – 0.0955 Lg × C (6) – 1.3 × 10–4 MW × C where Lg represents the treatment of simulated lagering (i.e., cold storage). The above finding was unexpected
Fig. 5. Turbidity of sheared beer samples after 0.45 µm membrane filtration.
Fig. 6. Turbidity of sheared beer samples after storage at 4°C for 2 weeks.
because ␤-glucans after shearing had increased particle size at higher MWs and concentrations 12. According to Stokes’ law, larger particles sediment faster and more readily than smaller ones, resulting in lower turbidity. However, this rapid sedimentation may not have occurred to the ␤-glucan particles that were 0.01 to 0.1 µm 12, a colloidal size range where Stokes’ law does not apply. It has been observed that, beers containing no ␤-glucan, low MW ␤-glucans or low concentrations of ␤-glucan exhibit reduced turbidity after storage (Fig. 6). Presumably, precipitation of proteins and/or other polymers has caused the partial removal of ␤-glucans. It is hypothesized that some of the ␤-glucan particles are adsorbed or bound to the settling proteins. This binding/adsorption may have removed some of the ␤-glucans from beer (i.e., by settling) (Fig. 7). Since only a given amount of proteins settled, the removed ␤-glucan was limited to a certain level. The remaining ␤glucans are hypothesized to suspend well in beer because of their small size (0.01–0.1 µm). The increased turbidity of beer at higher molecular weights and concentrations may be caused by aggregation of the particles.
'32'097-327 The experiments were carried out with clear wort and beer, which had very low initial turbidity levels (20–22 FTU). The addition of ␤-glucans increased turbidity of both wort and beer at higher MWs and concentrations (p < 0.001) although samples still remained brilliant and clear. Shearing increased turbidity of both wort and beer, especially those containing high MW ␤-glucans at higher concentrations (p < 0.001). The increased turbidity is hypothesized to be due to polymer aggregation and larger particle sizes. Changes in the number of particles, however, need to be further investigated. The turbidity of wort decreased at both higher maltose levels and higher pH values (p < 0.001). Shearing wort at higher temperature decreased the turbidity of samples (p < 0.001) compared to shearing at lower temperatures. The hazes formed in the ␤-glucan-free worts were believed to be caused by proteins, whereas the increased hazes in worts containing the 443 kDa ␤-glucan were believed to be derived from both proteins and ␤-glucans.
Beer turbidity was increased with ␤-glucans, shearing, and higher pH values (p < 0.001), but was not affected by shearing temperature (p > 0.05). The addition of ethanol decreased beer turbidity (p < 0.001). After filtration through 0.45 µm membranes, beer samples were clarified due to removal of the non-␤-glucan haze particles. However, filtration with 0.45 µm membranes removed only very small amounts of the hazes caused by ␤-glucans. After lagering, turbidity of the sheared beers containing low MW ␤-glucans (31 and 137 kDa) or low concentrations (50–200 mg/L) of ␤-glucans decreased (p < 0.001). However, turbidity of beer containing high MW (250, 327 and 443 kDa) ␤-glucans at concentrations higher than 400 mg/L increased (p < 0.001). In conclusion, hazes caused by high MW ␤-glucans at high concentrations cannot be thoroughly removed by 0.45 µm membrane filtration or lagering at 4°C for 2 weeks by gravity sedimentation. %'/23 ;0)(+)1)28 7
A strategic grant awarded by the Natural Sciences and Engineering Research Council of Canada to RAS and ATP is acknowledged. Graduate Fellowships awarded by the American Association of Cereal Chemists (1999–2001) to YLJ are also appreciated. The financial support of Canada Malting Company Ltd. and hosting of YLJ in their laboratories in Calgary are gratefully acknowledged. The donations of commercial malt by Canada Malting Company Ltd. and ␤-glucanase by Gist-Brocades Frances S.A. are also acknowledged. 6)*)6)2')7
1. American Society of Brewing Chemists. Methods of Analysis, 8th ed. Malt 4 Extract; Beer 5 Real Extract (B. Gravimetrically); Beer 12 Reducing Sugars (Copper Reducing Substances); Beer 26 Formazin Turbidity Standards. The ASBC: St. Paul, MN, 1992. 2. Bamforth, C.W., Biochemical approaches to beer quality. J. Inst. Brew., 1985, 91(5), 154-160. 3. Bamforth, C.W., Beer haze. J. Am. Soc. Brew. Chem., 1999, 57(3), 81-90. 4. Coote, N. and Kirsop, B.H., A haze consisting largely of pentosan. J. Inst. Brew., 1976, 82(1), 34. 5. Erdal, K. and Gjertsen, P., ␤-Glucans in malting and brewing II. The fate of ␤-glucans during mashing. Proceedings of the European Brewing Convention Congress, Madrid, IRL Press: Oxford, 1967, pp. 302–319.
Fig. 7. The percentage of ␤-glucans remaining in suspension in the sheared beer after storage at 4°C for 2 weeks. :3023
6. European Brewery Convention. Analytica-EBC, 4th ed. 9.16 Haze. The EBC: Brauerei- und Getränke-Rundschau : Zurich, CHE, 1987. 7. Gjertsen, P., Beta-glucan in malting and brewing. I. Influence of beta-glucan on the filtration of strong beer. Proc. Am. Soc. Brew. Chem., 1966, pp. 113–119. 8. Grimm, A. and Krüger, E., Determination of high molecular beta-glucan. 23rd EBC Monograph: Symposium Malting Technology, Andernach, 1994, pp. 94–109. 9. Hough, J.S., Briggs, D.E., Stevens, R. and Young, T.W., Chemical and physical properties of beer. In: Malting and Brewing Science. Vol. II. Hopped Wort and Beer. Chapman and Hall Ltd.: New York, 1982, pp. 776–838. 10. Igarashi, H. and Amaha, M., Frozen beer precipitates. II. Chemical structure of the ␤-glucan isolated from the precipitates. J. Inst. Brew., 1969, 75(3), 292–299. 11. Jackson, G. and Bamforth, C.W., Anomalous haze readings due to ␤-glucans. J. Inst. Brew., 1983, 89(3), 155–156. 12. Jin, Y.-L. Effect of β-glucan and environmental factors on the physical and chemical properties of wort and beer. Ph.D. Thesis. Dalhousie University: Halifax, NS, 2002. 13. Leedham, P.A., Savage, D.J., Crabb, D. and Morgan, G.T., Materials and methods of wort production that influence beer filtration. Proceedings of the European Brewing Convention Congress, Nice, IRL Press: Oxford, 1975, pp. 201–215. 14. Letters, R., Beta-glucans in brewing. Proceedings of the European Brewing Convention Congress, Amsterdam, DSW: Dordrecht, NDL, 1977, pp. 211–224. 15. Letters, R., Carbohydrate gels: A problem with newer brewing technologies. 5th Proc. Scientific and Technical Conv. Inst. Brew. (Central and S. African Sect.), 1995, pp. 115–121. 16. Moll, M., Colloidal Stability of Beer. In: Brewing Science. Vol. 3. J.R.A. Pollock, Ed., Academic Press, Inc.: New York, 1987, pp. 2–293. 17. Mundy, A.P. and Boley, N., A survey of instrumentation used for the determination of haze in beer. J. Inst. Brew., 1999,
105(2), 75–78. 18. Patelakis, S., The effect of ␤-glucan polymers on the rheological and filtration properties of beer. M.Sc. Thesis. Dalhousie University: Halifax, NS, 1999. 19. Patelakis, S., Speers, R.A., Paulson A.T. and Stewart, R., Effect of ␤-glucan polymers on the rheological and filtration properties of beer. Paper presented at the 65th ASBC Annual Meeting, June 19–23, 1999, Phoenix, AZ. Abstract O-9. ASBC Newsletter 59(2), 16. 20. Siebert, K.J., Effects of protein-polyphenol interactions on beverage haze, stabilization, and analysis. J. Agric. Food Chem., 1999, 47(2), 353–362. 21. Siebert, K.J., Relationship of particle size to light scattering. J. Am. Soc. Brew. Chem., 2000, 58(3), 97–100. 22. Siebert, K.J., Carrasco, A. and Lynn, P.Y., Formation of proteinpolyphenol haze in beverages. J. Agric. Food Chem., 1996, 44(8), 1997–2005. 23. Skinner, K.E., Hardwick, B.C. and Saha, R.B., Characterization of frozen beer precipitates from single packages. J. Am. Soc. Brew. Chem., 1993, 51(2), 58–63. 24. Sørensen, S.B. and Ottesen, M., Fractionation and characterization of beer proteins. Carlsberg Res. Commun., 1978, 43(3), 133–144. 25. Takayanagi, S., Amaha, M., Satake, K., Kuroiwa, Y., Igarashi, H. and Murata, A., Frozen beer precipitates. I. Formation and general characters. J. Inst. Brew., 1969, 75(3), 284–292. 26. Vårum, K.M. and Smidsrød, O., Partial chemical and physical characterisation of (1-3), (1-4)-␤-D-glucan from oat (Avena sativa L.) aleurone. Carbohydr. Polym., 1988, 9(2), 103–117. 27. Whitear, A., Maule, D.R. and Sharpe, F.R., Methods of mash separation and their influence on wort composition and beer quality. Proceedings of the European Brewing Convention Congress, London, IRL Press, Oxford, 1983, pp. 81–88.
(Manuscript accepted for publication September 2003)