Beer Filtration Foam Retention Limitations Following Propylene Glycol Alginate Dosing

In bright beer operations where propylene glycol alginate (PGA, E405) is used as a foam-positive stabilizer, the sequence of filtration unit operations after additive injection determines whether the expected head retention improvement is actually observed at packaging. PGA is typically metered as a dilute aqueous solution at 20–60 mg/L final beer concentration under cold conditions of 0–2 °C, at a product pH of 4.0–4.4. Under these conditions the propylene glycol ester side chains increase hydrophobic association with foam-positive protein fractions, while residual carboxylate sites remain pH-responsive and capable of interacting with dissolved calcium. If beer is then passed through a precoated filter, a turbidity-clear filtrate measured by ISO 7027:2016 may retain only a fraction of the high-molecular-weight foam-active complexes because depth filter media discriminate by size exclusion, adsorption, and cake charge. The operational consequence is that bulk viscosity and foam stability measured by EBC 9.42 or ASBC Beer-22 can diverge: a beer may be analytically within turbidity specification while its NIBEM time after filtration is 15–40 s below the value obtained immediately after dosing. Published data for exact polymer recovery across production-scale beer filters is limited, but plant investigations consistently identify pre-filtration dosing as the main variable that disrupts the otherwise predictable dose-response curve of PGA. The interaction is compounded when inline dosing is completed less than 60 s before the filter inlet, because incomplete homogenization produces concentration streaks that challenge both filter permeability and foam-active polymer availability. Regulatory identity for PGA is established under FDA 21 CFR 172.858 in the United States and under Regulation (EC) No 1333/2008 Annex II as E405 in the European Union; neither designation exempts the substance from the physical removal mechanisms that occur in downstream filtration. Consequently, the practical limitation is not the additive itself but the placement of filter operations relative to the dosing point and the adsorptive character of the media selected.

Why Does Upstream Dosing Before Diatomaceous Earth Create a Foam Retention Liability?

The pressure leaf and candle diatomaceous earth filter is designed for removal of suspended yeast and colloidal haze, not for selective passage of anionic polysaccharides. A typical precoat load of 700–1000 g/m² and body feed of 0.8–1.5 g/L produce a permeable cake whose depth channels can blind rapidly when hydrated PGA films deposit at the cake surface. Inline dosing at 30–50 mg/L less than 60 s before the filter inlet produces a bimodal distribution of polymer concentration; this causes localized regions of high PGA that increase filter cake compressibility and raise differential pressure from a clean-cake baseline of 0.6–0.8 bar toward 2.0–2.2 bar over a 4–6 h run. The pressure rise is frequently interpreted as a yeast load problem, though the actual mechanism is cake blinding by hydrated PGA films. Simultaneously, the filter aid surface and cellulose fractions can adsorb the foam-positive PGA-protein complex through hydrogen bonding and hydrophobic interaction, lowering the concentration of foam-stabilizing material reaching the bright beer tank. Downstream NIBEM measurement may show that foam retention is only 70–80 % of the pre-filtration value even when no turbidity breakthrough is detectable. Because esterified alginate is anionic, it can bind to exposed calcium salts in the cake or to positively charged filter aid impurities, further reducing the amount of foam-active polymer that survives filtration. Published data for this exact configuration is limited, and most manufacturers of PGA and filter aids recommend post-filtration addition to avoid this interaction. Where pre-filtration dosing is unavoidable due to tank logistics, the body feed should be reduced to ≤1.0 g/L, differential pressure should be terminated at ≤2.0 bar, and the filtration run should be shortened so that the cake does not accumulate successive PGA layers. Even with these adjustments, the foam retention deficit must be accounted for by dose correction, because the filter medium is not inert to charged polysaccharides.

Across ceramic and polymeric crossflow membrane skids configured with 0.45 µm or 0.65 µm nominal pore ratings, the addition of PGA before the membrane introduces a colloidal fouling mechanism distinct from ordinary particle removal. In this arrangement, the esterified alginate and its associated beer proteins form a compressible concentration-polarization layer at the membrane surface under transmembrane pressures of 0.8–1.5 bar, and the observed flux may decline from 120–180 L/m²/h to 30–60 L/m²/h within 60–90 min depending on crossflow velocity and temperature. The shear rate in the retentate channel can reduce the size of PGA-protein aggregates, but high shear also disrupts the hydrophobic interfacial association responsible for foam stabilisation; the result is a filtered beer with low turbidity and limited foam retention. Ceramic membranes with titanium dioxide or zirconia surfaces show less hydrophobic adsorption than hydrophobic polyethersulfone membranes, but they still retain large polymeric aggregates by size exclusion at the pore entrance. Even if the membrane is backflushed and chemically cleaned with alkaline detergent at 60–80 °C, the loss of foam-active material cannot be reversed because the protein-polymer complex is discarded with the retentate. The operational boundary is therefore strict: post-filtration dosing is preferred, or if sterile filtration is required, the membrane must be positioned before PGA addition and the downstream package must be maintained under aseptic conditions without further filtration. Process monitoring by turbidity alone is insufficient because the shift in molecular weight distribution occurs while filtrate clarity remains acceptable. For membrane filtration following PGA dosing, tangential-flow operation with crossflow velocity ≥4 m/s and transmembrane pressure ≤1.2 bar is the minimum control envelope to delay gel-layer collapse, but foam retention may still fall below specification because the shear history itself modifies the interfacial functionality of the polymer.

Sheet Filtration with PVPP and Selective Adsorption of Esterified Alginate

Sheet filters containing cellulose or regenerated cellulose depth pads with embedded polyvinylpolypyrrolidone are used in brewpubs and smaller production lines for combined cold stabilization and polyphenol adjustment. When PGA is dosed before such pads, the high internal porosity of the PVPP fraction is not selective for polyphenols alone; it also adsorbs hydrophobic protein-polymer complexes and removes the very fraction that contributes to head retention. A typical single-pass PVPP contact dose of 15–30 g/hL in a pad configuration can reduce post-sheet NIBEM values by 10–25 s relative to the same beer dosed after the sheet filter, even when the pads are prewashed with deaerated water and the differential pressure remains below 1.0 bar. The limitation is often invisible in standard bright tank turbidity because the sheet filter removes both haze and the foam-stabilizing PGA complex simultaneously; the filtered beer appears bright but lacks lacing. Pad manufacturers generally recommend that foam stabilizers be added after all polishing and stabilization media because depth pads are designed to remove high-molecular-weight species and are not neutral to charged polysaccharides. If inline dosing must be conducted upstream of a sheet filter due to tank logistics, the dose must be increased by an empirical retention correction factor, and the filter run length must be shortened to prevent breakthrough of unbound PGA, which can create post-filtration haze. The operating boundary for sheet filtration after PGA addition is a filter load ≤2.5 hL/m²/h, with differential pressure termination at ≤1.0 bar and sequential checks of foam retention by EBC 9.42 or ASBC Beer-22 at the initial, middle, and terminal stages of the pad run. These checks are necessary because pad uptake is not linear; once adsorption sites become saturated, PGA begins to pass through the pad in increasing concentration, and the relationship between filtrate volume and foam-active polymer delivery changes. Published data on the exact adsorption isotherm for PGA on PVPP-containing beer pads is limited, but the directional loss of foam stability under this sequence is consistently reported in brewing process audits.

When a sterilising membrane is placed downstream of the PGA injection point in a cold-sterile or flash-pasteurised beer line, the critical limitation is no longer the removal of yeast or haze but the integrity of the filter under a load of high-molecular-weight foam stabiliser. A 0.45 µm polyethersulfone membrane cartridge validated for bacterial retention per ASTM F838-20 may exhibit a bubble point shift from a clean-water baseline of 1.8–2.0 bar to 2.2–2.5 bar after only 2–3 h of beer service when PGA has been added upstream at 25 mg/L; the shift indicates pore wetting or surface coating by the anionic polysaccharide, and the resulting diffusive flow may still pass an integrity test while the actual beer flux is far below the manufacturer’s rated capacity. The foam retention limitation arises because the membrane surface and depth structure selectively retain the largest foam-positive complexes while allowing lower-molecular-weight polymer and simple beer components to pass; the sterile filtrate therefore has a shifted molecular weight distribution and lower surface viscosity at the bubble interface. If the dosing point is moved to the sterile filtrate side, the line after the membrane must remain aseptic and the downstream filler must provide final microbial control. Published data for this specific configuration is limited, but membrane fouling studies and brewery commissioning records indicate that polymer-assisted sterile filtration is feasible only when the membrane pore size is at least twice the median aggregate diameter and when crossflow or tangential flow is used instead of dead-end filtration. Dead-end cartridge arrangements are especially vulnerable because the entire polymer load must pass through a static membrane surface, increasing the rate of irreversible surface coating. Where sterile filtration after PGA cannot be avoided, the dosing rate should be reduced to the lowest level capable of meeting foam specification, and pre-filtration particle size distribution should be monitored by laser diffraction per ISO 13320:2020 to ensure that the median aggregate diameter does not approach the membrane pore rating.

When Recarbonation Dilution Follows PGA Injection

In high-gravity beer production where concentrate at 14–16 °P is diluted with deaerated water to sales gravity, PGA is sometimes added to the concentrate before dilution because this simplifies dosing and mixing. The subsequent dilution step can reduce the effective concentration of foam-active polymer below the minimum required for bubble film reinforcement, particularly when the concentrate has been filtered after dosing and the recovered stabiliser is already depleted. If PGA is metered at 40 mg/L on concentrate volume and the beer is then diluted 1:1, the final concentration falls to 20 mg/L, which may be below the observed foam retention threshold for that particular beer matrix. In addition, the addition of cold deaerated water at 2–4 °C reduces bulk viscosity and disrupts the weakly formed interfacial equilibrium between PGA, hydrophobic proteins, and iso-alpha acids; the foam half-life may then decline even though the analytical concentration of polymer in the final beer appears satisfactory. Carbonation adjustments after dilution introduce further shear at the carbonation stone or venturi, and the resulting CO₂ bubble nucleation can strip surface-active compounds from the bulk liquid. The practical boundary is that PGA should be dosed on final package volume after dilution, recarbonation, and final filtration. If this is not possible, the dose must be calculated on final volume rather than concentrate volume, and the recarbonation step should be completed before inline dosing so that the stabiliser is not exposed to high CO₂ bubble shear. This is especially important in lines where dissolved carbon dioxide is adjusted to 4.8–5.2 g/L and the product is then subjected to a final guard filter; the guard filter may remove PGA aggregates that formed during recarbonation, producing variability between packaged batches even when tank readings remain within specification.

Calcium concentrations above 80 mg/L in brewing water or in beers treated with calcium sulfate in the mash and kettle subject residual unesterified carboxylate groups on propylene glycol alginate to divalent bridging, even though propylene glycol esterification reduces calcium sensitivity compared with unesterified sodium alginate. The degree of esterification of commercial PGA is often in the range of 60–85 %, leaving 15–40 % of the carboxyl sites free to interact with calcium at beer pH 4.0–4.4. This interaction can produce hydrated gel domains that appear as post-filtration haze and that deposit on final sterile membranes or packaging bowl filler surfaces. The fouling is not removed by ordinary cold-water rinse and may require hot alkaline detergent at 60–70 °C with chelating agents. When calcium-hard water is used for final dilution after PGA dosing, the local pH can be slightly higher than the product pH, further increasing the ionised carboxyl fraction and promoting precipitation. Foam retention in such systems is limited because the polymer is sequestered into insoluble aggregates rather than being available at the gas-liquid interface; the NIBEM time can be 10–30 s lower than in the same beer with calcium below 50 mg/L. The control point is therefore the calcium-to-PGA ratio and the degree of esterification, not simply the total PGA dose. Where calcium concentrations exceed 80 mg/L, a PGA product with a degree of esterification above 80 % and a final addition point after all filtration is required to avoid simultaneous haze and foam defects.