Filtration Challenges in Non-Alcoholic and Low-Alcohol Beer Production with a Focus on Different Yeast Strains

Abstract

This study examines the impact of yeast selection on fermentation performance, filtration efficiency, and the stability of low-alcohol and non-alcoholic beer (NAB). Three yeast strains (LA-01, LoNa, and W-34/70) were evaluated for their effects on key NAB parameters like ABV, bitterness, color, haze readings, and filtration-linked performance. Filtration trials revealed that highly flocculating W-34/70 resulted in lower post-filtration turbidity, whereas LA-01 and LoNa (which flocculated less) required more effective filtration techniques to achieve clarity. Among the tested filter sheets, AF 31, AF 71, and AF 101 successfully reduced turbidity below 2 EBC. However, all NABs exhibited haze formation during storage, indicating the necessity of stabilization treatments. These findings highlight the importance of yeast strain selection and filtration strategies in optimizing NAB production. Additional pretreatment steps, such as centrifugation, may improve filterability for non-flocculating yeasts, while stabilization measures are essential for maintaining long-term clarity. This study provides valuable insights for improving industrial NAB processing and quality.

1. Introduction

The growing socio-cultural emphasis on health-conscious living has driven an increased demand for beverages that adhere to wellness principles [1]. This trend is particularly evident in the rising consumer demand for low-alcohol and non-alcoholic beer (NAB) offerings [2,3,4]. The expansion in NAB consumption is attributable to several key factors, including documented health benefits, such as isotonic properties advantageous for post-exercise recovery, evolving societal attitudes toward alcohol consumption, stricter regulations on drink-driving, and the changing palates of consumers [1,4,5].

The production of NABs leverages advanced techniques to preserve flavor while minimizing alcohol content, requiring precise adjustments to enhance mouthfeel and flavor complexity. Despite their reduced alcohol levels, NABs aim to closely replicate the sensory characteristics—taste, aroma, and mouthfeel—of conventional beer styles [6]. There are primarily two approaches for producing such beer: (1) microbiological approaches that suppress ethanol production during fermentation [4,7,8,9] and (2) physical approaches which remove ethanol from fully fermented beer [8,10]. Additionally, hybrid techniques that combine these different methods can be used to closely mimic the flavor profile of alcoholic counterparts in non-alcoholic beers [5,10]. The brewing of NABs provides extensive engineering possibilities for process optimization. This versatility highlights microbiological methods as a valuable approach for studying various yeast strains [4,7] and their applications in downstream processes, including filtration.

Filtration is a critical step in the production of beverages including NAB as it ensures product clarity, microbiological stability, and sensory quality [11,12]. However, the process presents unique challenges due to the interaction between yeast, particle size, and the physicochemical properties of the NAB matrix [11,12,13]. Yeast plays a pivotal role in both the fermentation and filtration stages, and its characteristics, including size and surface properties, significantly influence the efficiency of filtration processes [12].

Yeast cells in beer production in general are larger than many other (non)suspended particles in the beer matrix, with diameters ranging from 3 to 10 µm, depending on the species and strain [11,12]. The size distribution of these cells is an important factor affecting filtration [12]. Larger yeast cells can form a more permeable filter cake, enabling faster flow rates and extending the filtration process without cleaning (increased throughput) but potentially reducing the retention of smaller particles. Conversely, smaller yeast cells and fragments, such as those resulting from autolysis, may clog filter pores and decrease filtration efficiency and throughput. The challenge lies in achieving a balance between adequate yeast removal and the maintenance of desired beer characteristics, such as haze stability and flavor. The physicochemical properties of yeast, such as cell wall structure and surface hydrophobicity, further complicate filtration dynamics. On the one hand, yeast with a more robust cell wall or a higher degree of hydrophobicity (e.g., flocculent yeasts) tends to adhere to filtration surfaces, which can lead to fouling and reduced filter life, especially in crossflow processes [14,15]. On the other hand, yeast strains with weaker cell walls may release intracellular components, increasing the overall particle load and negatively impacting filtration and/or shelf life. Furthermore, dissolved compounds from raw materials (e.g., glucans, colloids, and/or gel forming substances like glucan gels) or the release of intercellular materials can increase the viscosity of beer and in turn decrease fluid flow and filtration capability [11,12,13,16,17].

In turn, the particle size distribution of the beer matrix is another critical parameter. Besides yeast, beer (and also NAB) contains a range of particles, including protein–polyphenol complexes, starch residues, and other colloidal substances [11,12,16,18]. These particles can interact with yeast cells during filtration, influencing the formation and compaction of the filter cake [11,12]. A high concentration of fine particles can result in a densely packed filter bed, leading to increased pressure drops and reduced throughput. Conversely, an optimal distribution of particle sizes can facilitate the formation of a more open filter structure, improving filtration performance.

The selection of an appropriate filtration method is crucial in beer processing. Depth filtration, utilizing media like diatomaceous earth (also called kieselguhr) or cellulose-based filters, is often favored for NABs produced via microbiological methods due to its capacity to manage higher particle loads and variations in yeast size [12]. Membrane filtration, although offering superior microbial retention, can encounter challenges such as clogging when handling heterogeneous particle size distributions or high particle loads, resulting in increased fouling [12,13,15]. The selection of an appropriate filtration system significantly impacts beer quality. Most breweries producing NABs using microbiological methods are limited in their ability to alternate between filtration technologies (e.g., dead-end or crossflow filtration). Consequently, beer clarification predominantly relies on cake filtration methods, such as kieselguhr filtration [11]. The kieselguhr plate-and-frame filter is composed of several filter plates alternately arranged on a frame. Each plate is coated with a filter layer, typically made of cellulose and kieselguhr, which acts as the primary filtration medium. The filter plates, layers, and frames are tightly compressed to ensure optimal performance. In modern applications, particularly on a smaller scale, ready-to-use filter sheets or modules are commercially available, designed for various beer styles and capable of handling different particle loads or turbidity levels [19].

Advances in filtration materials, such as customized pore sizes and surface modifications, have significantly expanded the potential for optimizing filtration processes in brewing. However, most of the existing literature predominantly focuses on yeast applications and filtration performance in traditional alcoholic beer production [11,12,13]. In contrast, research specifically addressing filtration efficiency in the production of microbiologically produced NABs is limited, particularly when using alternative yeast strains such as maltose-negative or non-Saccharomyces yeasts [4,7].

While there is a growing body of research on the sensory and chemical quality of NABs [9,11,12,20,21,22], these studies mainly concentrate on the impact of yeast strains on flavor development, leaving broader processing challenges insufficiently explored. Brewing is a complex, multi-stage process that extends far beyond fermentation, encompassing critical downstream steps such as filtration, stabilization, and packaging [11,12,23]. The successful implementation of novel yeast strains [20,21] in commercial NAB production thus depends not only on their fermentative performance and flavor contributions but also on their compatibility with these subsequent processing stages.

Furthermore, the industrial processing, especially filtration, of NABs presents specific challenges related to sustainability, cost, and food safety that warrant more critical discussion [4,10]. The selection and management of filtration media must consider environmental impacts and operational expenses including filter replacement and cleaning, as well as the maintenance of microbiological stability to ensure product safety [24,25]. Additionally, the sensory importance of filtration should be emphasized, as this process affects not only turbidity but also aroma, flavor, and mouthfeel, all of which are essential for consumer acceptance [11,23,26].

This study addresses a significant knowledge gap by systematically evaluating the fermentation characteristics of two maltose-negative Saccharomyces yeast strains (SafBrew™ LA-01 (Fermentis, Lille, France) and Lallemand LoNa™ (Lallemand Inc., Montreal, QC, Canada) [27,28]) alongside a commonly used bottom-fermenting yeast (SafLager™ W-34-70 (Fermentis) [29]). It further investigates the downstream processing behavior of various kieselguhr-based ready-to-use filter sheets [19] in a frame filter system within NAB production. Focusing on strains that differ markedly in morphology and physiology from classical brewing yeasts, this work examines how these differences influence filtration efficiency, yeast flocculation, and product stability.

Through this comprehensive approach, this study provides novel insights into the practical challenges and opportunities associated with using alternative yeast strains in commercial NAB brewing. It offers valuable guidance for brewers aiming to optimize both product quality and process efficiency. Finally, the practical relevance of this research is highlighted by examples and recommendations for matching different yeast–filter combinations to specific NAB market segments, such as clear lagers versus craft-style beers.

2. Materials and Methods

2.1. Wort and Beer Production

The NABs in this study were brewed, fermented, filtered, and analyzed at the University of Arkansas Department of Food Science Center for Beverage Innovation (Fayetteville, AR, USA). NAB production was conducted by brewing a single batch of wort, which was subsequently divided among three fermenters, each fermented with a different yeast strain (LA-01, LoNa, W-34/70 [27,28,29]).

2.1.1. Wort Production

Wort production utilized a 4-vessel brewing system (2 heating vessels, 2 mash vessels) from Ss Brewtech (Wildomar, CA, USA), producing a total of 120 L of wort. Barley malt (American Pale Two-Row; Rahr Malting Co., Shakopee, MN, USA) was milled using an Ss Brewtech Grain Mill with roller setting 0 and mashed into water at a 1:4 grist-to-water ratio at 74 °C. A final mash rest temperature of 72 °C was maintained for 60 min, and saccharification was confirmed using the iodine test (0.02N iodine; Merck KGaA, Darmstadt, Germany).

Lautering and sparging were performed until a kettle full concentration of 6.0 °P was achieved, monitored using an Anton Paar EasyDens with Brew Meister Software (iOS version 4.5.0; Anton Paar GmbH, Graz, Austria) [30]. During the 60 min boil, CTZ hops (14.2% α-acid; BSG Hops, Wapato, WA, USA) were added at the beginning to achieve a final bitterness of 11 BU. The boiling process concluded with adjustments to gravity (6.5 °P ± 0.1) and pH (4.8 ± 0.05, using 88% concentrated lactic acid (BSG, Shakopee, MN, USA)). Following 5 min of circulation and a 15 min whirlpool rest, the wort was cooled via a single-stage heat exchanger. Finally, 35 L of wort was transferred into each Ss Brewtech Unitank fermenter (53 L capacity).

2.1.2. Fermentation

The yeasts used in these trials were commercially available dry yeasts, stored at 4 °C and nearing the end of their labeled best-by date. Yeast pitching was performed with LA-01 and LoNa strains at a rate of 65 g/hL wort, while W-34/70 was pitched at 100 g/hL, following the medium-rate recommendations provided by the supplier [27,28,29]. Fermentation for all beers was conducted at 20 °C and continued until stable ethanol concentrations were achieved, verified over two consecutive days. After fermentation, the beers were cold-crashed at 2.5 °C for two days before further processing.

2.1.3. Filtration, Kegging, and Pasteurization

Two hours prior to filtration, the cones of the fermentation tanks were emptied to remove settled yeast, preventing resuspension during beer filtration. This step was repeated immediately before filtration began. The beer/yeast slurry was drained until visually clear beer was observed.

For filtration, FIBRAFIX^® AF depth filter media (Filtrox AG, St. Gallen, Switzerland) were employed, covering a range from coarse (AF 11), through fine (AF 31 and AF 71), to germ-reducing filtration (AF 101) [19]. The filtration setup utilized a Hobracol 200 Mikro filter (Hobra—Školník s.r.o., Broumov, Czech Republic) with a single inlet, two collecting frames (one outlet each), and two filter sheets with a total filter area of 0.064 m². Preparation involved assembling the filter frame with the sheets and tightening all screws. The filter was rinsed with 90 °C water to remove loose particles, facilitate filter sheet swelling, and sterilize the system. The hot water was held in the filter for 10 min for proper swelling, followed by flushing with sterilized, degassed water at 4 °C to cool the system.

The fermenter was then connected to the filter, and filtration commenced under 1 bar of pressure (food-grade CO₂) to transfer beer through the filter into KEGs. Before filling KEGs (19.5 L stainless steel), an initial 2 L of beer was discarded to avoid dilution with residual water. After collecting 5 L of filtered beer (filtered volume 78 L/m²), the process was stopped, and the filter was cleaned with hot water. The preparation and filtration steps were repeated with new filter sheets, resulting in 4 distinct KEGs of filtered beer for each yeast strain, yielding 12 KEG samples in total.

The kegged beer was pasteurized in boiling water to achieve 110 PU and then cooled, carbonated, and stored at 4 °C for stability testing over two weeks.

2.2. Analysis

2.2.1. Physical–Chemical Analysis

Beer analyses were conducted using standardized methods outlined by the European Brewery Convention (EBC)—density, original gravity, and apparent and real extract (EBC 9.4 [31]; EBC 9.43.2 [32]) as well as alcohol content (EBC 9.2.6) [33]—and ASBC standard methods Beer Method 9. pH (Hydrogen Ion Concentration) [34], Beer Method 10. Color [35], and Beer Method 23. Bitterness were applied [36].

The turbidity was measured using a VWR^® Turbidity Meter (VWR International, LLC., Radnor, PA, USA; measurement range 0–1000 NTU; NTU conversion to EBC turbidity units by dividing the results by a factor of 4) following the supplier’s instructions on handling, calibration, and measurement.

2.2.2. Volatile Analysis

Volatile compound analysis was performed using a Shimadzu GC-MS system (GCMS-TQ8050 NX, Shimadzu Corporation, Kyoto, Japan), equipped with a Shimadzu AOC-6000 Plus Autosampler. A 50/30 µm DVB/CAR/PDMS fiber (Supelco, Bellefonte, PA, USA) was employed for SPME. Samples were incubated at 65 °C for 10 min, extracted under the same conditions for an additional 10 min, and thermally desorbed for 3 min at 240 °C in splitless mode. The inlet temperature was maintained at 240 °C, with helium used as the carrier gas.

The GC oven was programmed to start at 35 °C with a 5 min hold and increase to 100 °C at 5 °C/min, then to 150 °C at 3 °C/min, then to 160 °C at 8 °C/min, and finally to 250 °C at 25 °C/min with a final hold of 5 min. The total runtime was 39.52 min. Separation was achieved using an HP-5MS UI column (30 m × 0.25 mm × 0.25 µm; Agilent J&W GC Columns, Santa Clara, CA, USA). The MS source temperature was set to 200 °C, and the interface transfer line was held at 280 °C.

Prior to quantitative analysis, standards were subjected to a full scan (m/z 40–400) followed by a product ion scan to optimize collision energy for achieving optimal transitions. The MRM method was established using the full scan as a template. For precise measurements, analytes were grouped into 11 major events, each consisting of multiple transitions specific to individual compounds, with a loop time of 0.0550 s per event. The detector was operated at an absolute voltage of 1.3 kV, with the ion source and interface temperatures maintained at 200 °C and 280 °C, respectively, while all other parameters, including oven programming and injection, mirrored those of the full scan method.

For quantification, an internal standard-based MRM calibration curve comprising 10 calibration points (10–10,000 µg/L) was prepared using a standard mixture of 41 compounds (purity > 95%, Sigma-Aldrich, St. Louis, MO, USA) spanning various chemical classes (esters, aldehydes, terpenes, terpene alcohols). Calibration solutions included 500 µg/L of five internal standards (hexanal-d₁₂, ethyl hexanoate-d₁₁, phenylacetaldehyde-d₅, linalool-d₅, and beta-myrcene-d₆; purity > 94%, AromaLAB GmbH, Planegg, Germany), each representing a distinct compound class. All analytes demonstrated linear calibration curves with R² > 0.999. Following calibration, unknown compounds in the samples were quantified relative to the internal standards assigned to their respective classes.

2.2.3. Microscopic Analysis

Morphological and viability (percentage of live, metabolically active cells in a yeast population) yeast cell evaluation was conducted using standardized methods. Yeast cell staining (Microbiology Yeast 3 yeast stains) and Methylene blue staining and counting (Microbiology Yeast 4 microscopic yeast cell counting) were performed as outlined by the ASBC [37,38].

Scanning Electron Microscopy (SEM) was performed using the University of Arkansas Institute for Nanoscience and Engineering FEI Nova Nanolab 200 SEM. All samples had been gold-plated to increase SEM imaging and resolution. Detailed information on the SEM system used can be found on the Arkansas Nano & Bio Materials Characterization Facility page (https://microscopy.uark.edu/nova-nanolab/, accessed on 13 December 2024).

3. Results and Discussion

3.1. Yeast Analysis

The selection of yeast strains (LA-01, LoNa, and W-34/70) was based on both their growing relevance to NAB production and their distinct fermentation characteristics [20,21]. LA-01 and LoNa are maltose-negative Saccharomyces cerevisiae strains that are commercially available and have been developed and investigated specifically for the production of NABs through restricted fermentation [27,28]. Their limited ability to metabolize maltose and maltotriose minimizes ethanol formation during fermentation but also results in the development of fermentation-derived aroma compounds. These characteristics render them appropriate model organisms for investigating flavor development- and process-related parameters of NABs [20,21]. W-34/70, in contrast, is a bottom-fermenting lager strain (Saccharomyces pastorianus) that has been widely used in the traditional brewing of full-strength beers [29]. However, when applied under constrained fermentation conditions, such as reduced fermentation temperatures, this strain can also be used to produce NABs. Its inclusion enables a comparison between specialized NAB strains and a conventional brewing yeast with modified fermentation parameters, thus providing a broader perspective on strain-dependent outcomes in NAB production.

Dry yeasts offer numerous advantages, including ease of storage, precise dosing, and extended shelf life. Nevertheless, limitations remain despite recent advancements in freeze-drying methodologies. One common issue is cell wall disruption, which limits viable cell counts upon pitching, particularly when the yeast has been stored for extended periods or is close to its best-by date [39]. SEM images of the three yeasts in their freeze-dried state before pitching show a significant degree of cell wall disruption (Figure 1), particularly for LoNa. In contrast, W-34/70 exhibits some irregularly shaped cells but does not display a high proportion of damaged yeast cells. This could be attributed to the yeast drying process itself or may also be affected by the yeast nearing the end of its best-before date, as stated on the packaging [39]. Although W-34/70 is not typically used in microbiological processes to produce NAB, it remains one of the most commonly employed lager yeast strains and demonstrates considerable potential for low-alcohol beer production. Fermentation with this yeast is generally recommended within a temperature range of 12 to 18 °C [29]. However, because fermentation activity and the generation of flavor compounds such as higher alcohols and esters are often limited by wort gravity during NAB production, a slight increase in the fermentation temperature to 20 °C was applied in this case.

Yeast cell counts at pitching were not determined in this study. Instead, pitching rates were applied according to the respective manufacturers’ recommendations [27,28,29], reflecting standard industry practice. In commercial brewing operations, craft brewers commonly follow these guidelines and often pitch dry yeast directly, without propagation or precise cell enumeration. As craft brewers typically do not quantify pitching cell counts but may implement beer filtration, the present study was intentionally designed to reflect practical brewing conditions. While the initial pitching rate does influence the yeast concentration at the end of fermentation, it does not necessarily correlate with the cell count at the filter inlet [11,40,41]. Several factors, including original gravity, nutrient availability, yeast growth kinetics, flocculation, sedimentation during fermentation, yeast cropping for reuse, and pre-filtration treatments such as centrifugation, can substantially affect the yeast load entering the filtration step [11,23,41,42,43]. To highlight the importance of monitoring yeast cells under these conditions, this analysis focused on determining the cell count at the filter inlet. This metric was chosen to enable an evaluation of filtration behavior within a realistic brewing framework.

Yeast viability (percentage of live, metabolically active cells in a yeast population) was assessed after 60 h of fermentation at 20 °C, followed by a 48 h cold crash at 2.5 °C and an additional 24 h at 4 °C in bottles before sample processing. Sampling for yeast cell count and viability assessment was performed immediately prior to the onset of filtration, ensuring that the measured values accurately reflected the yeast load at the filter inlet. To eliminate sampling-related variability, all samples were homogenized before microscopic staining. Consequently, the 24 h storage in bottles prior to staining analysis did not influence the total cell count. The measured concentrations were 4.7 × 10⁶ cells/mL for LA-01, 5.8 × 10⁶ cells/mL for LoNa, and 6.2 × 10⁶ cells/mL for W-34/70. According to standard brewing research [11,23], beers typically undergo cold maturation at 0 to −2 °C for 1 to 2 weeks to improve clarification, enhance filterability, and increase colloidal stability. During this phase, yeast concentrations are expected to decrease below 2 × 10⁶ cells/mL, with 5 × 10⁶ cells/mL considered the upper acceptable limit prior to filtration [23]. In the present study, the beers were subjected to a short cold-storage period of only 48 h at 2.5 °C. As a result, yeast concentrations remained slightly above the commonly recommended thresholds. Nevertheless, the observed values were considered acceptable for initiating the filtration trials, particularly given this study’s focus on realistically simulating practical brewing conditions.

The viability of the non-maltose-metabolizing yeasts was 55.6% for LA-01 and 53.4% for LoNa, compared to 88.4% for the bottom-fermenting yeast W-34/70. This reduced and varying viability observed in the specialized non-maltose-metabolizing yeasts may present challenges in brewing non-alcoholic beers. The differences in viability can be explained by initial damage and irregular yeast cell morphology (Figure 1) caused by the freeze-drying process. If yeast cells are not in a healthy condition at the outset, their long-term viability may be negatively affected, potentially impacting fermentation performance and product quality.

Furthermore, limited sugar metabolism may lead to earlier yeast starvation and decreased viability due to the restricted availability of fermentable sugars. The literature shows that non-alcoholic mashing procedures typically yield higher maltose and maltotriose levels while limiting glucose and fructose concentrations to limit the ethanol content in the final non-alcoholic products [9]. The initial glucose concentration in wort significantly impacts sugar consumption, as glucose repression prevents yeast from metabolizing other sugars until glucose is depleted. High glucose levels, whether from crystalline or liquid sugar extract, can allow the yeast’s enzymatic system to adapt to prioritize glucose. This adaptation reduces or halts maltose uptake, often resulting in incomplete fermentation [11].

Brewing processes after fermentation must be meticulously planned if yeast reuse or filtration is required. Yeast starvation can lead to autolysis, potentially releasing substances that negatively affect filtration [12,13]. In contrast, the bottom-fermenting yeast (W-34/70), capable of metabolizing all low-molecular-weight sugars, exhibited higher viability (88.4%). This can be attributed to increased sugar availability in wort, supporting more robust metabolism during fermentation.

3.2. Basic Beer Analysis

No statistical differences were seen in original extract (OG) and bitterness, indicating that no dilution occurred due to wort splitting into the three different fermenters (

Table 1. Basic analytical results for NABs after fermentation (constant alcohol readings over 2 days); superscript letters indicate LS mean groupings (Fisher LSD, p < 0.05).

	Original Extract [% w/w]	Bitterness [BU]	Alcohol [% v/v]	Apparent Degree of Fermentation [%]	pH [-]
LA-01	6.56 a ± 0.01	11.1 a ± 0.2	0.53 c ± 0.00	15.51 c ± 0.01	4.40 a ± 0.00
LoNa	6.59 a ± 0.00	11.0 a ± 0.0	0.56 b ± 0.00	16.26 b ± 0.00	4.36 a ± 0.01
W-34/70	6.56 a ± 0.00	11.2 a ± 0.1	2.33 a ± 0.01	68.91 a ± 0.00	3.99 b ± 0.00

). After fermentation, as expected, the maltose-negative yeast strains resulted in a lower alcohol by volume (ABV) compared to the W-34/70 strain. The LA-01 (ABV 0.53% vol.) and LoNa (ABV 0.56% vol.) strains produced ABV values generally aligned with the threshold for non-alcoholic products (ABV < 0.5% vol. [10,11,44]). In contrast, W-34/70 yielded a significantly higher ABV of 2.33% vol. [27,28,29]. The apparent degree of fermentation (ADF) further underlines these results, with LA-01 at 15.51% and LoNa at 16.26%, compared to 68.91% for W-34/70. This increased fermentation activity also affected the pH of the beers. During the main fermentation of alcoholic beers, pH can drop by up to one unit due to the formation of volatile (acetic, formic) and non-volatile organic acids (pyruvic, malic, citric, lactic) through amino acid deamination [11]. In the maltose-negative yeasts tested, limited carbohydrate metabolism resulted in final pH values of 4.40 (LA-01) and 4.36 (LoNa), comparable to typical beer values of 4.3–4.6 [11]. Due to the initial acidification of the wort with lactic acid (pH 4.8 ± 0.05), in contrast, the W-34/70 NAB exhibited a further pH drop due to its higher degree of fermentation and subsequent organic acid production. Overall, these results provided an optimal starting condition for the subsequent filtration tests, as the NABs demonstrated the desired characteristics for further processing.

3.3. Filtration Impact on NABs Brewed and Analyzed

3.3.1. NAB Filtration Performance and Impact on Fresh Products

The filtration parameters assessed in this study were limited in scope and do not encompass key technical variables typically used to evaluate filtration performance, such as flow rate, filtration time, pressure drop, or filter loading capacity. These parameters are ultimately necessary for assessing the efficiency and economic feasibility of filtration processes in an industrial brewing context. Moreover, filtration setups in breweries vary considerably, ranging from sheet and candle filters to crossflow membrane systems, and each of these techniques operates under distinct process conditions [12]. As such, the limited scale and simplified filtration setup used in this study (i.e., a two-sheet lab-scale filter with a fixed filtration volume of 5 L and a constant pressure of 1 bar) limit the comparability of the data with industrial filtration practices, and the results should be compared with those in the appropriate context.

To highlight the practical implications of employing various filter media and yeast strains in the production of low-alcohol and non-alcoholic beers, this study did not include a detailed analysis of haze composition. Instead, the focus was placed on monitoring haze reduction and observing potential changes in unstabilized beer stored at 4 °C over a two-week stability testing period. Turbidity levels in beers fermented with LA-01 and LoNa decreased over the course of the filtration trials, whereas W-34/70 exhibited a significant increase in turbidity, exceeding 250 EBC (Figure 2). This variation can be attributed to differences in yeast flocculation characteristics and the fermentation tanks used. LA-01 and LoNa are classified as non- to medium-flocculent yeasts [20,21], which facilitated effective sedimentation and subsequent removal during filtration. Since all beers were filtered from a single fermentation tank, the observed turbidity reduction can be explained by yeast sedimentation and retention within the filter matrix (AF 11 → AF 101). Conversely, W-34/70 is a highly flocculent yeast strain [22], and it accumulated on the inner cooling coils of the fermenter. Upon emptying the fermenter, aggregated yeast clumps were released into the beer stream, entering the filter system and contributing to the observed increase in turbidity.

Among the filter sheets tested, AF 11 exhibited the lowest filtration efficiency, as indicated by the highest turbidity levels across all beer samples (Figure 2). Despite the high yeast load in the inlet for LA-01 and LoNa, the final turbidity values remained elevated at 39.3 EBC and 71.5 EBC, respectively, demonstrating visibly turbid filter outlet samples. In contrast, beer fermented with the highly flocculent yeast W-34/70 showed the lowest turbidity (1.28 EBC) after filtration with AF 11, falling below the human eye’s turbidity perception threshold of >2 EBC [18,45], and thus would have been classified as non-hazy.

When considering only the turbidity results, the filtration efficiency of AF 31, AF 71, and AF 101 was notably superior, yielding outlet turbidity values well below 2 EBC, categorizing these beers as clear or nearly brilliant [18,45]. This finding is particularly relevant, as it indicates that all filter sheets effectively clarified beers fermented with medium- or non-flocculent yeasts (LA-01 and LoNa), while the presence of highly flocculent and potentially clumped yeasts (W-34/70) did not significantly impair filtration efficiency.

Although filtration time and flow rate were not quantitatively recorded, qualitative assessment suggested that AF 11 exhibited the highest flow rate, followed by AF 101, while AF 31 performed more slowly and AF 71 had the lowest filtration efficiency in terms of flow rate. To optimize filtration performance, particularly for beers fermented with non-flocculent, maltose-negative yeasts such as LA-01 or LoNa, additional pretreatment steps (e.g., centrifugation) may be necessary and should be explored in future trials.

In addition to turbidity-based filtration performance, the effects on color and bitterness were also evaluated (

Table 2. Color and bitterness results for NABs produced with the three different yeast strains and filtered using various filter sheet setups; superscript letters indicate LS mean groupings (Fisher LSD, p < 0.05).

	Filter Inlet	Filter Outlet AF 11	Filter Outlet AF 31	Filter Outlet AF 71	Filter Outlet AF 101
Color [EBC]
LA-01	5.5 b ± 0.0	Haze b,*	4.7 a ± 0.0	4.7 a ± 0.1	4.6 a ± 0.0
LoNa	5.4 c ± 0.0	Haze b,*	4.7 a ± 0.0	4.4 b ± 0.0	4.7 a ± 0.0
W-34/70	7.1 a ± 0.0	4.4 a ± 0.0	4.2 b ± 0.0	4.1 c ± 0.0	4.2 b ± 0.0
Bitterness [BU]
LA-01	11.1 a ± 0.2	11.3 a ± 0.0	11.3 a ± 0.0	11.4 a ± 0.1	11.4 a ± 0.0
LoNa	11.0 a ± 0.0	11.4 a ± 0.1	11.0 ab ± 0.0	11.1 b ± 0.1	11.0 b ± 0.0
W-34/70	11.2 a ± 0.1	9.6 b ± 0.1	10.3 b ± 0.3	9.5 c ± 0.1	9.7 c ± 0.0

* Haze—due to the elevated turbidity observed after filtration with AF 11 for LA-01 and LoNa, color measurements could not be conducted.

). For most of the NAB filtration trials, a statistically significant reduction in color was observed in the filter outlet samples. However, the color reduction was independent of the filter sheet used, indicating that all filtration setups performed similarly in this regard. This outcome aligns with expectations, although the existing literature does not provide a definitive conclusion regarding color changes due to filtration [46,47]. Notably, the most pronounced color reduction was detected in the low-alcohol beer fermented with W-34/70. Given that the wort composition was identical across all fermentation trials, this suggests a potential reduction in color during fermentation. This hypothesis is further supported by the observation that the filtered low-alcohol beer exhibited a lower color intensity compared to the NABs.

The reduction in bitterness observed after the filtration of the W-34/70 NAB (Table 2) is in line with the previous literature [48,49] and may be attributed to strain-specific interactions between the yeast and hop-derived bitter compounds, particularly iso-α-acids. W-34/70, a strongly flocculating Saccharomyces pastorianus strain [29], exhibited the highest yeast concentration (6.2 × 10⁶ cells/mL) at the filter inlet. At the same time, its pronounced tendency to aggregate during cold conditioning and clarification likely promoted the removal of suspended solids, including yeast-bound hop constituents. One plausible explanation is the adsorption of iso-α-acids onto the yeast cell wall [50]. Strains with elevated flocculation capacity and more hydrophobic cell surfaces, such as W-34/70, may display a greater ability to bind and sequester these bittering compounds, thus reducing their concentration in the final beer. Moreover, the elevated cell count (6.2 × 10⁶ cells/mL) at the point of filtration could further intensify the adsorption and physical retention of iso-α-acids within the filter cake. Collectively, these effects, strain-specific adsorption, flocculation-driven sedimentation, and filtration-related retention, likely contributed to the observed decline in bitterness in the W-34/70 sample.

Overall, the different filter sheets (except for AF 11) resulted in clear beers. Highly flocculent yeast strains, such as W-34/70, might lead to a minor reduction in color and bitterness post filtration. Conversely, NAB production with non- or medium-flocculent yeasts, such as LA-01 and LoNa, may require pretreatment (e.g., centrifugation) to improve filtration efficiency. The AF 11 filter sheet demonstrated inadequate performance, failing to produce acceptable clarity in the LA-01 and LoNa trials. While color decreased across all NAB filtration trials, the bitterness of NABs remained largely unaffected, in contrast to the low-alcohol beer made with W-34/70.

3.3.2. Filtration and Its Impact on Selected Volatile Compounds in Fresh NABs

The current volatile compound analysis is limited to a narrow range of substances, primarily focusing on aldehydes and ethyl esters. While these compounds are important markers of flavor [51], this restricted scope overlooks other critical flavor-active compounds that significantly influence beer quality. For instance, diacetyl—a basic flavor compound known for its buttery aroma [11,51]—plays a crucial role in sensory perception and will be included in further investigations to provide a more comprehensive flavor profile.

Volatile compounds in full-strength beer encompass various groups of flavor-related chemicals [5,26,52,53]. Modifications in production processes, such as dry hopping, metabolic regulation, or thermal treatments for non-alcoholic beer production, can further alter these compounds in fresh beer prior to filtration [4,5,10]. Since the aroma and flavor of non-alcoholic beers influence consumer perception and repurchase decisions, chemical volatile analysis serves as a valuable tool for product comparison. Selected aldehydes, hop aroma compounds, and esters were monitored before and after filtration (Figure 3, Figure 4 and Figure 5).

Aldehydes contribute to off-flavors that typically develop during aging, with specific compounds imparting distinct undesirable notes [5,22,52]. While aldehydes mainly originate from malt [54], modifications in production processes can alter their profile, potentially enhancing the fresh aroma impression while also improving the long-term stability of non-alcoholic beers [9,52]. A statistically significant reduction in 2-methylpropanal (2MP) and hexanal (C6 aldehyde) was observed (Figure 3). 2MP was included in the analysis due to its known contribution to grainy, varnish-like, and fruity notes, with a reported flavor threshold of 86 µg/L [55], whereas hexanal imparts bitter, vinous, and aldehydic sensory impressions, with a higher threshold of 88–350 µg/L [51,55]. Although the initial concentrations and the observed reductions in the NABs produced remained below these thresholds, making a noticeable impact on fresh flavor unlikely, previous studies have demonstrated that 2MP and hexanal levels can increase during storage, contributing to oxidative off-flavors in lager-style beers over time [56]. The reduction in these aldehyde observations is likely linked to yeast reduction, as it is anticipated that yeast can bind aldehydes [57]. For LA-01 and LoNa, the aldehyde concentrations decreased after filtration with AF 11 compared to the filter inlet yet remained higher than those observed for AF 31, AF 71, and AF 101. This finding is noteworthy, as it highlights a beneficial effect in reducing aldehydes, which can contribute to off-flavors in specific beer styles over time [22,55,56]. In contrast, the filtration of W-34/70 led to the highest turbidity reduction across all filter sheets (Figure 2) but no statistically significant differences in 2MP and hexanal levels among filtered samples (Figure 3).

Since no sensory evaluation was conducted in this study, only chemical correlations between certain compounds and their concentration-dependent off-flavors can be inferred. Notably, nonanal (C9 aldehyde) remained mainly unaffected by filtration (Figure 3), suggesting that its increased chain length prevents binding to yeast in the same manner as hexanal (C6 aldehyde). Consequently, reducing yeast content in NABs does not lead to the efficient extraction of nonanal. A similar pattern was observed for furfural (Figure 3), an aldehyde associated with off-flavors (paper, caramel, bready, sensory impressions [51,55]) in beer and known to increase with temperature during storage [52,55]. Due to furfural’s aldehyde group attachment at the second position of the furan ring, yeast binding is unlikely, explaining the statistically insignificant changes observed across all yeast strains and filter sheets tested.

In contrast to aldehydes, which are predominantly associated with off-flavors [55], both brewers and consumers seek specific volatile compounds that contribute desirable aromas to beer, including non-alcoholic beers [52]. These compounds primarily include hop-derived volatiles (Figure 4) [58] and fermentation-related esters (Figure 5) [52], both of which play a crucial role in defining the sensory characteristics of the final product [26].

Filtration had little to no statistical impact on hop aroma compounds in the NABs tested (Figure 4). The minor but statistically significant changes observed can likely be attributed to factors such as sampling variability, sample preparation inconsistencies, and the inherent measurement uncertainty of the analytical method. These factors were not accounted for in the ANOVA performed in this study, suggesting that the observed variations may not necessarily indicate a true filtration effect.

Given the importance of hop-derived volatiles in beer aroma [53,58,59], the minimal impact of filtration on these compounds is a valuable finding. It suggests that filtration does not significantly alter the hop profile of lager-style NABs. The hop-derived compounds analyzed in this study represent some of the most common and sensorially relevant volatiles found in beer [58]. Myrcene, for example, is a major contributor to fresh hop aroma, typically present in concentrations ranging from 30 to 100 µg/L, and is described as having resinous, pine-like, and herbal characteristics [58]. Although the concentrations detected in this study were below the typical flavor threshold, they remained unaffected by filtration.

Linalool, a key compound of hop oil [60], also contributes to hoppy beer aroma, with a reported flavor threshold below 10 µg/L (R-linalool, beer-specific) [58]. In this study, filtration had no statistically significant effect on linalool concentrations (Figure 4), suggesting that the impact of filtration on linalool-driven sensory impressions is negligible. Given the importance of linalool in defining hop character, this finding underscores the ability to preserve hop aroma during beer clarification processes.

β-Citronellol, a compound associated with yeast metabolism [58,61], showed elevated concentrations in the W-34/70 sample, likely due to the strain’s higher fermentation activity. While β-citronellol levels tended to decrease slightly upon filtration, they remained above its reported flavor threshold of <10 µg/L [58,62]. From a purely chemical perspective, this suggests that the compound’s lemon- or lime-like aroma [62] would still contribute to the sensory profile of the final product, even after filtration.

Similar conclusions apply to geraniol. Although the observed variation in geraniol concentrations was slightly greater, no filtration-induced reductions were statistically observed. This is particularly relevant, as the flavor threshold for geraniol is also <10 µg/L [58], and this value was not undercut in any of the filtered samples. Therefore, from a chemical standpoint, filtration does not appear to compromise the geraniol-derived sensory attributes of the NABs.

α-Terpineol (associated with lilac-like aroma [62]; threshold 2000 µg/L [51]), α-humulene (spicy, woody character; threshold 120–450 µg/L [63]), and β-caryophyllene (also described as spicy and woody; threshold 230 µg/L [63]) were all detected at concentrations well below their respective flavor thresholds. Moreover, their levels remained unaffected by filtration. From a chemical standpoint, this indicates that their already minimal contribution to the overall sensory profile of the NABs is not altered by the filtration process.

As previously mentioned, esters represent another crucial group of aroma compounds in beer [52]. Their significance lies in their synergistic interactions, where a combination of esters and other volatile compounds can activate flavors below their individual perception thresholds. Given their inherently low flavor thresholds [51], even slight changes in ester concentrations can alter the overall aroma and flavor profile of beer [53,59]. To assess potential filtration effects, ethyl hexanoate (apple, fruity, sweetish flavor impressions; threshold 230 µg/L [51]) and ethyl octanoate (apple, fruity, sweetish flavor impressions; threshold 900 µg/L [51]) were selected as representative fermentation-derived esters (Figure 5).

The results indicate that the filtration of LA-01 and LoNa with AF 11 led to statistically significant differences compared to treatments using AF 31, AF 71, and AF 101. While ester concentrations decreased when LA-01 and LoNa were filtered with AF 11, a much more pronounced reduction was observed in the other filtration treatments. In contrast, W-34/70 exhibited relatively consistent ester concentrations across all filtration treatments, mirroring the comparable turbidity readings obtained in these trials (Figure 2). When considering the sensory thresholds of the esters analyzed [51], the reductions observed due to filtration must be interpreted with caution. Although the absolute concentrations of esters such as ethyl hexanoate and ethyl octanoate remained below or near their individual flavor thresholds, their decline may still represent a critical sensory change—from a chemical standpoint. This is because esters are known to act synergistically with other volatile compounds, meaning that their combined presence can enhance fruity and floral flavor impressions even when individual concentrations are below perceptual limits [52]. Therefore, any reduction in ester levels may lead to a perceptible decrease in the overall aroma complexity and fruity character of the beer. In this context, ester retention becomes even more critical for achieving a well-rounded and appealing sensory profile [5,26]. From a chemical perspective, the ester losses observed in the filtration treatments, especially with finer filter grades (AF 31, AF 71, AF 101), suggest that there may be an impact on the sensory quality of the final product, particularly for beer styles where ester-driven fruitiness is desired, and future studies should investigate the sensory impact of these differences [5,26,52].

3.4. Storage-Related Haze Changes

The NABs produced and filtered in this study were not subjected to stabilization treatments. Haze formation in beer is predominantly influenced by two major components—(I) proteins and (II) tannins (polyphenols) [11,64]—both of which in combination tend to increase over time during storage. To mitigate this effect, adsorption techniques utilizing colloidal silica or bentonite can be employed to remove proteins, while polyvinylpyrrolidone (PVPP) is effective in stabilizing beer by reducing polyphenol content [11,64]. In this study, NABs were stored at a controlled temperature of 4 °C for a duration of two weeks, with periodic monitoring of turbidity changes (start/1 week/2 weeks).

According to the established literature on haze development in full-strength beers, an increase in turbidity over time is generally expected [11,64]. For all NABs produced and filtered, the changes in turbidity were tracked at both one and two weeks of storage (

Table 3. Turbidity changes in NABs filtered with various filter sheet setups and measured after different storage times; superscript letters indicate LS mean groupings (Fisher LSD, p < 0.05).

Filter Sheet	Storage Time @4 °C	LA-01	LoNa	W-34/70
AF 11	filter outlet	39.3 e ± 1.5	71.5 a ± 0.5	1.28 g ± 0.0
	1 week	37.0 e ± 0.3	62.0 c ± 1.0	3.62 fg ± 0.0
	2 weeks	43.8 d ± 1.3	68.4 b ± 0.6	4.02 f ± 0.0
AF 31	filter outlet	0.34 f ± 0.0	0.20 g ± 0.0	0.90 a ± 0.0
	1 week	0.87 a ± 0.0	0.44 e ± 0.0	0.71 cd ± 0.0
	2 weeks	0.80 b ± 0.0	0.68 d ± 0.0	0.75 c ± 0.0
AF 71	filter outlet	0.11 i ± 0.0	0.20 h ± 0.0	0.83 a ± 0.0
	1 week	0.60 d ± 0.0	0.47 g ± 0.0	0.70 c ± 0.0
	2 weeks	0.56 e ± 0.0	0.49 f ± 0.0	0.76 b ± 0.0
AF 101	filter outlet	0.19 f ± 0.0	0.21 f ± 0.0	0.79 ab ± 0.0
	1 week	0.76 bc ± 0.0	0.48 e ± 0.0	0.73 c ± 0.0
	2 weeks	0.82 a ± 0.0	0.60 d ± 0.0	0.83 a ± 0.0

). A comparison of different yeast strains and filtration approaches revealed distinct trends in haze formation. Specifically, NABs fermented with LA-01 and LoNa, which exhibited the highest initial turbidity (39.3 EBC and 71.5 EBC, respectively) when filtered with AF 11, displayed different behaviors over the two-week period. The non-flocculent yeast strain LA-01 [27] showed a continuous increase in turbidity, aligning with the expected trend based on previous research [11,64]. Conversely, LoNa, which possesses medium flocculation and sedimentation properties [28], initially experienced a decline in haze levels after one week, likely due to sedimentation effects. However, turbidity began to rise again after an additional week of storage. This fluctuation indicates a dynamic interplay between sedimentation and subsequent resuspension of haze-forming compounds, suggesting that medium-flocculent yeasts can contribute to variable turbidity patterns over time.

Haze development is influenced by both pH and ethanol content [65], which helps explain the more intense haziness observed in the LA-01 and LoNa beers. These beers had ethanol levels of 0.53% ABV (LA-01) and 0.56% ABV (LoNa) with corresponding pH values of 4.40 and 4.36, respectively. In contrast, the W-34/70 beer exhibited a higher ethanol content of 2.33% ABV and a lower pH of 3.99. The reduced haze formation observed with the AF 31–AF 101 filter sheets in the W-34/70 samples can thus be attributed to these differences in ethanol concentration and pH.

These findings warrant cautious interpretation, particularly in the context of beers displaying elevated initial turbidity levels. In practical brewing scenarios, such beers would typically undergo additional processing steps to enhance clarity and stability. Meanwhile, the NAB produced with W-34/70 and filtered with AF 11 exhibited a markedly different trend. Initially, this beer had a very low turbidity reading of 1.28 EBC, appearing visually clear [18]. However, over time, an increase in haze formation was observed, with turbidity rising to 4.02 EBC after two weeks. This finding is consistent with previous studies, as W-34/70 was not stabilized against protein- or polyphenol-induced haze formation [11,64]. The increase in turbidity suggests that, even in cases where initial haze levels are low, long-term stability can still be a concern if stabilization measures are not applied.

Further analysis of the NABs filtered with AF 31, AF 71, and AF 101 showed that all remained below the critical threshold of 2 EBC, beyond which beers typically exhibit visible haziness [18,45]. Interestingly, despite starting with lower turbidity values in these filtration trials, LA-01 and LoNa exhibited a greater increase in haze over time compared to W-34/70 (Table 3). The latter, which initially recorded turbidity readings between 0.8 and 0.9 EBC, ultimately reached similar final turbidity values after two weeks of storage at 4 °C. This suggests that yeast strain selection and filtration type play crucial roles in determining haze stability.

From a broader brewing perspective, the findings underscore the importance of filtration in maintaining haze levels below 2 EBC during short-term storage at 4 °C. However, stabilization methods are necessary to ensure prolonged haze stability and extended shelf life. Beer is rarely stored continuously at 4 °C in commercial distribution and retail settings, making it susceptible to varying temperature conditions that can accelerate haze formation. To better simulate real-world storage conditions, forced aging methods are commonly employed in brewing science to assess beer stability under stress conditions to simulate flavor variations [53,66] or turbidity-related changes [67]. Such methods for assessing turbidity-related changes involve exposing the beer to elevated temperatures for a defined period in combination with lowering the temperature [67] to force protein–polyphenol-related turbidity, accelerating the aging process and providing insights into long-term haze development.

3.5. Applied Strategies for Industrial Implementation

The findings of this study do provide relevant implications for industrial-scale NAB production, particularly in relation to yeast strain selection, filtration strategy, and product positioning. Depending on the targeted market segment and desired product style, brewers must adapt their process design to ensure both efficiency and quality. For clear, lager-style NABs, highly flocculent yeast strains such as W-34/70 [29] are particularly suitable. When used in combination with fine or germ-reducing filter media (e.g., AF 31 or AF 101), these strains reliably produce NABs with turbidity values below 2 EBC, aligning with consumer expectations for visual brilliance [18]. Conversely, for craft-style NABs, which often tolerate or even promote a hazier appearance and more pronounced fermentation character, the use of non- or medium-flocculent yeast strains such as LA-01 or LoNa may be advantageous. These strains showed good applications for NAB production in microbiological processes but will require additional processing considerations due to their limited sedimentation and lower cell viability. In such cases, pre-filtration steps like cold sedimentation or mechanical clarification (e.g., centrifugation) are recommended to reduce particle load before filtration. Additionally, selecting a filter medium with an adequate retention grade that avoids rapid clogging—such as a two-stage filtration train with a coarse followed by a fine filter sheet—can improve operational reliability.

Beyond filtration performance, the increase in turbidity observed during storage for all samples underlines the need for effective stabilization strategies to ensure long-term colloidal stability. Proteins and polyphenols, the main contributors to chill haze formation [18,23], can be selectively removed using silica gel or PVPP, respectively [11]. These treatments should be considered essential process steps, particularly for NABs with extended chemical shelf life or ambient storage requirements. Moreover, the combination of yeast strain and filtration conditions can serve as a means of product differentiation. While W-34/70 tends to produce cleaner beers with lower bitterness and ester retention post-filtration, ideal for neutral, light-flavored NABs, LoNa and LA-01 allow for more expressive aromatic profiles. This makes them suitable for NABs targeting craft-oriented consumers who seek different flavor profiles or haze-tolerant styles such as dry-hopped or wheat-based NABs. From an operational standpoint, the choice of yeast directly affects processing efficiency, material use, and equipment capacity [12]. Highly flocculent strains may reduce filter sheet consumption and filtration time, thereby improving throughput in larger breweries [12]. In contrast, maltose-negative, low-flocculent strains may require longer processing times and more intensive clarification but offer greater control over ethanol content and flavor modulation. Thus, brewers should weigh processing complexity and style flexibility against production efficiency and sensory outcomes.

Altogether, the results support a modular and strain-specific approach to NAB production. Yeast–filter combinations should be selected not only for technical compatibility but also with a clear alignment to target product attributes and consumer preferences. This process-integrated thinking enables brewers to increase consistency, reduce production variability, and position their NAB offerings more strategically in a competitive and diversifying beverage market.

4. Conclusions

The findings of this study demonstrate that yeast selection has a significant impact on fermentation performance, filtration efficiency, and the stability of non-alcoholic beer (NAB). The yeast strains investigated (LA-01, LoNa, and W-34/70) exhibited notable differences in fermentation activity, cell viability, and flocculation, which directly influenced key beer parameters. Post-fermentation cell viability varied considerably among strains, with maltose-negative yeasts LA-01 and LoNa showing significantly lower viability (55.6% and 53.4%, respectively) compared to the bottom-fermenting strain W-34/70 (88.4%). This difference can be attributed to the limited sugar uptake of maltose-negative yeasts, leading to early nutrient depletion and increased cell mortality. Consequently, LA-01 and LoNa produced substantially lower alcohol concentrations (0.53% and 0.56% vol., respectively) than W-34/70 (2.33% vol.).

Filtration trials revealed distinct differences in efficiency across filter sheet types (FIBRAFIX^® AF 11, AF 31, AF 71, and AF 101 filter sheets). Highly flocculating yeast, such as W-34/70 [29], resulted in lower turbidity in the filtrate, whereas the non- to moderately flocculating yeasts (LA-01 and LoNa [27,28]) exhibited significantly higher turbidity levels (Figure 2). The AF 11 filter proved particularly ineffective, yielding insufficient clarification. In contrast, filter pads AF 31, AF 71, and AF 101 successfully reduced turbidity below 2 EBC, ensuring visually clear beers [18,45].

Despite effective filtration, substantial changes in turbidity were observed during storage. All NABs exhibited an increase in haze over two weeks, particularly in beers fermented with LA-01 and LoNa, where an initial turbidity decrease due to potential sedimentation was followed by a subsequent rise. Even W-34/70, which exhibited low turbidity post-filtration, showed an increase in haze over time. These results highlight the necessity of stabilization measures (e.g., PVPP or silica gel) to maintain optical clarity in NABs [11,18,64].

Nevertheless, certain limitations must be acknowledged. The storage stability trials in this study were limited to a two-week period under refrigerated conditions, which does not fully reflect real-world distribution or retail scenarios. The absence of sensory analysis restricts conclusions regarding flavor retention or loss, particularly with regard to esters and hop volatiles. Furthermore, while turbidity measurements were used to evaluate filtration performance and storage stability, the composition of haze-forming particles (e.g., proteins, polyphenols, β-glucans) was not quantified, limiting mechanistic interpretation.

Future research should address these limitations by expanding storage durations, applying forced aging protocols, and integrating sensory and compositional analyses. Additionally, investigations into enzymatic treatments, alternative filter media (e.g., membranes, crossflow), and bio-based stabilizers may provide further avenues for optimizing NAB clarity, flavor, and shelf stability. Developing such integrated approaches will help brewers meet the increasing demand for high-quality NABs while ensuring consistent process performance and product appeal.