Process-based VOC emission characteristics, environmental impact and health risk assessment in typical breweries in Beijing

Abstract

Brewing is an understudied but influential source of VOC emissions in the food manufacturing industry. In this study, we conducted a first comprehensive analysis of process-based VOC concentration characteristics, ozone formation potential (OFP), secondary organic aerosol formation potential (SOAFP) and health risks in two typical breweries in Beijing that use malted barley, hops, water, and yeast. In Brewery A, 35 to 53 distinct VOC species were detected, with total mass concentrations ranging from 148.17 ± 18.64 µg m⁻³ to 15 225.91 ± 1912.51 µg m⁻³. Brewery B demonstrated comparable patterns, with 28 to 49 species detected at concentrations between 104.49 ± 8.48 µg m⁻³ and 10 368.87 ± 879.47 µg m⁻³. Process-stage analysis identified boiling and fermentation stages as the key stages with the highest VOC concentrations, dominated by oxygenated VOCs (OVOCs) such as acetaldehyde, acetone, ethyl acetate, and 2-butanone, as well as the alkane isobutane. Atmospheric dispersion modeling (AERMOD) indicated negligible public health risks from organized stack emissions. In contrast, occupational health assessment revealed significant risks for workers from fugitive emissions, with the cumulative hazard index (HI) far exceeding the threshold. The OFP and SOAFP results, representing the secondary pollutant formation potential of the source mixtures, highlighted OVOCs and aromatics as priority control species for mitigating the secondary pollution potential. The findings demonstrate that VOC control strategies must be differentiated, with large-scale breweries prioritizing organized emissions, while small breweries urgently need to control fugitive emissions. This study aims to promote the implementation of VOC regulations and occupational health protection strategies within the brewing industry.

---

Environmental significance

Volatile organic compounds (VOCs) from food and beverage manufacturing represent significant yet inadequately characterized sources of atmospheric oxidants and particulate matter. This investigation provides the first process-based analysis of brewery emissions, with boiling and fermentation identified as critical sources characterized by OVOCs. The study reveals substantially different risk profiles between emission types, with organized releases showing minimal community impact compared to the substantial occupational hazards created by fugitive emissions. Additionally, aromatic hydrocarbons display disproportionate contributions to secondary organic aerosol formation relative to their minor mass fractions. These findings establish a scientific foundation for developing differentiated control strategies that concurrently address secondary pollution formation and workplace health risks, thereby facilitating precision environmental management in food and beverage industry operations.

1. Introduction

Volatile Organic Compounds (VOCs) are crucial precursors of PM_2.5 and O₃ that participate in photochemical reactions that lead to photochemical smog and secondary organic aerosol (SOA) pollution, which are the main contributors to haze pollution.^1,2 Moreover, exposure to VOCs can induce cancer and other diseases, with long-term exposure to high concentrations having the most significant effects on the respiratory tract and lungs,³ thereby posing a major threat to human health. Numerous previous studies have shown that industrial sources are the dominant anthropogenic contributors to VOC emissions in China.^4,5 Consequently, understanding the sources and emission characteristics of VOCs is critically important for the implementation of effective control strategies.

Research findings consistently identify industrial sources as the primary contributors to VOC emissions. Industrial emissions stand out due to their multifaceted origins, high intensity, complex composition, and elevated concentrations.⁶ In the past few years, Chinese scholars have paid much attention to quantifying and characterizing VOC emissions from specific industries, and previous studies have focused on large scale and heavily polluting industries such as the coking and coating industries. However, only limited studies have been conducted on the emission characterization of VOCs from other light industrial sources, such as the food manufacturing industry and the chemical materials industry.⁷ In the 21st century, China's food industry has experienced rapid and steady growth,^8,9 leading to increased VOC emissions from enterprises. These emissions predominantly consist of malodorous releases and organic by-products of fermentation.¹⁰ The food industry occupies a prominent position among the sources of VOC emissions, after manufacturing, pharmaceuticals, and transportation.^11–13 While current research has primarily focused on odorous emissions and organic by-products in the food industry, there is a noticeable dearth of studies examining VOCs at different process stages.¹⁴ Therefore, addressing VOC emissions from the food industry necessitates a more comprehensive understanding of the specific stages involved in food processing.

Currently, there is almost no research concerning the detection and analysis of breweries. Based on previous research, odor gases are released during the production of beer,^15,16 which involves processes such as malting, saccharification, fermentation, boiling, filtration, filling.¹⁷ A general idea of the type of pollutant can be obtained, but it is not sufficiently clear and the range of concentrations is not clear. Previous studies related to the brewing industry have focused on the characterization of odorants of beer itself and the pollution characteristics of beer wastewater and have paid more attention to individual facilities or generic industry average emission values, without explicitly examining the impacts of different scales and technologies on VOC emissions. There is a lack of research on VOC emission characteristics and health risks at different process stages, so studying the emission characteristics of VOCs at different scales is essential for developing more targeted control strategies.

In this study, VOC emissions from organized and fugitive sources of the brewing process were analyzed by comparing two breweries with the same core technologies but different scales. The study aimed to characterize VOC emission characteristics, quantify their atmospheric impacts through OFP and SOA generation assessments, model dispersion patterns using AERMOD simulations, and evaluate health risks following the U.S. Environmental Protection Agency (U.S. EPA) risk assessment method. The key objectives included: (1) comparative analysis of organized versus fugitive emission characteristics in brewing processes; (2) mechanistic evaluation of VOC-driven photochemical ozone and SOA formation; (3) quantifying comprehensive health risks for both non-carcinogenic and carcinogenic risks and analyzing the dispersion of organized VOCs using the AERMOD model; (4) comparing breweries of different scales and focusing on their own differences to determine the factors affecting the emission characteristics of VOCs. The results of the study advance the scientific understanding of pollution characteristics and health risks in the brewing industry, and offer differentiated emission reduction pathways for breweries of varying scales, effectively circumventing the limitations of "one-size-fits-all" environmental regulatory approaches.

2. Methodology

2.1. Sampling sites and VOC sampling

In this study, two breweries with different scales of production and technological levels were selected in Beijing, China, designated as Brewery A and Brewery B, respectively. Both breweries primarily produce lager beer, using malted barley, hops, water, and yeast as the main raw materials. The location is shown in Fig. S1. The differences between Brewery A and Brewery B in terms of production scale, process details, and equipment types are summarized in Table S1. Brewery A employs fully automated production lines, enclosed fermentation systems, and regenerative thermal oxidation (RTO) for exhaust gas treatment. Brewery B retains semi-automated equipment, open fermentation processes, and activated carbon adsorption for waste gas treatment. Involving different company scales, the rated capacities of Brewery A and Brewery B are 60 000 and 10 000 tons per year, respectively. During the sampling period, both breweries were operating at or near their full rated capacity. By comparing the two breweries, it is possible to compare the impact of the production mode and treatment technology of the enterprises on VOC emissions.

The sampling period was June–July 2022, and the sampling cycle included two complete production batches, mainly focusing on the peak season. A total of 44 samples were collected during each sampling period. The typical beer production process and VOC emission points are shown in Fig. 1. The process and pollution emission links were carefully investigated to select appropriate sampling sites for VOCs. In addition, background monitoring points were set up 500 meters upwind and downwind of the breweries. The organized emission sampling points included the stages of gelatinization, saccharification, filtration, boiling and sedimentation. The fugitive sampling points include five stages of fermentation, filling, labeling and coding, adhesives and emission of VOCs from the biobox at the sewage station. The selected sample sites covered almost all target factories' mission units. During the sampling period, the production status and processes of the breweries operated normally, so the sampling was well representative.

Fig. 1 Brewing process flowchart and VOC emission points.

Air samples were taken using 3.2 L fused silica stainless steel canisters (Entech Instruments, Inc., Simi Valley, CA, USA). The containers should be pre-cleaned three times with high purity nitrogen and evacuated to <10 Pa.^18,19 Samples were collected using polytetrafluoroethylene (Teflon) tubing connected from the particulate filter to the SUMMA canister.²⁰ A flow restrictor valve was connected to the inlet of the canister to control the flow rate to 20 mL min⁻¹, and the sampling time was 120 min. Before sampling, a portable photoionization detector (PID, ppbRAE 3000, RAE Systems, San Jose, CA, USA) was used to make initial measurements of VOC concentrations at each sampling site. Fugitive emission samples were collected in the workshop at a height of 1.5 m above the ground, positioned within 1–2 m of the primary emission point for each process stage to capture representative process-specific concentrations. Workshop air samples were collected at a height of 1.5 meters above the ground near primary production units, The canisters were immediately sent to the laboratory and analyzed within 15 days.²¹

2.2. Sample analysis and quality assurance

Samples need to be diluted with high purity nitrogen prior to detection, and the dilution ratio was determined by the PID value of each sample. The analysis was performed using a gas chromatograph equipped with a mass selective detector and a flame ionization detector (7890B-5977B GC-MSD; Agilent Technologies Inc., USA) and coupled with a pre-concentrator system (model 7200, Entech Instruments Inc., USA).²² The system incorporated cryogenic preconcentration to eliminate water and carbon dioxide interference. Given the analytical challenge of separating C2–C12 compounds, we implemented a dual-column strategy. For light hydrocarbon separation, including ethane, ethene, acetylene, propane, and propene, we utilized a Plot Q column with FID detection following established techniques.²³ The remaining 112 compounds were separated using a DB-1 column with MS detection. The initial temperatures were held at 5 °C for 6 min, increased at a rate of 5 °C min⁻¹ to 200 °C, and kept for 8 minutes until all target compounds came out finally.²⁴

The standard gas used for measurement included the Photochemical Assessment Monitoring Stations (PAMs) standard mixture (56 NMHCs, Linde Spectra Gases, Inc., USA) and TO-15 standard mixture (65 compounds, Linde Spectra Gases, Inc., USA).^25,26 Based on these standard gases, calibration curves were calculated for each target VOC species using a serial concentration gradient (7 levels: 0.5, 1.0, 2.0, 5.0, 10.0, 15, and 20 ppb). The correlation coefficients (R²) of calibration curves were more than 0.99. Bromochloromethane, 1,2-difluorobenzene, chlorobenzene-d5, and 1,4-difluorobenzene of known concentrations were used as internal standards for the calibration of the system. The measurement accuracy of each species was within 5%. In addition, the daily calibration results of the measuring system were consistent, indicating that the system was stable.

For the quality assurance (QA) and quality control (QC), no detectable VOCs were present when ten laboratory blank samples were processed using the same method as the real samples. No discernible contamination was found in the field blanks. The total recovery efficiencies of VOCs ranged from 76% to 128%, and the variance coefficient was <20%. The detection limits were 0.17–2.57 µg m⁻³. The relative correlation for the standard curve was in the range of 0.992–0.999.

2.3. Estimation of the secondary pollution formation potential

In this study, both the OFP and SOAFP were calculated based on VOC mass concentrations. This approach aims to evaluate the intrinsic chemical reactivity and aerosol formation potential of VOCs at each process stage. The types of measured VOCs, chemical categories, MIR values, and SOAP values are shown in Table S2. It is crucial to note that these values cannot be directly equated to the absolute mass of ozone or SOA ultimately formed in the atmosphere. The latter depends on total emission rates, subsequent atmospheric dispersion and chemistry, which are affected by factors beyond the scope of this study, including emission transport efficiency, concentrations of nitrogen oxides, and aerosol concentration levels.

2.3.1. Estimation of the ozone formation potential. The concentration and reactivity of VOCs are two key factors influencing ozone formation.^27,28 In 1994, Carter first proposed the Maximum Incremental Reactivity (MIR) of the OFP,²⁹ taking into account both the kinetic and mechanistic activities of VOCs. In this study, the OFP was calculated using MIR values based on the VOCs produced at various stages of brewery's processes, which can be calculated using formula (1):where OFP_j represents the OFP of the jth process, gO₃ (g VOCs)⁻¹; MIR_i denotes the MIR value of substance i, gO₃ (g VOCs)⁻¹; x_ij signifies the weighted percentage of the ith VOC in the jth process.³⁰

2.3.2. Estimation of SOA formation. SOA constitutes a critical atmospheric pollutant contributing significantly to haze formation. As key precursors in SOA generation, VOCs participate in aerosol formation through atmospheric oxidative transformation and gas–particle partitioning mechanisms. The secondary organic aerosol formation potential (SOAFP) metric quantifies the contribution of VOCs to this process. Current SOAFP estimation methodologies encompass multiple approaches, including the fractional aerosol coefficient (FAC) method, EC tracer technique, OC/EC ratio analysis, CMB-MM receptor modeling, and empirical formula calculations.

In this study, the contributions of VOC species to SOA were assessed based on Derwent's SOAP coefficient method.³¹ The SOAP index quantifies the relative SOA formation capacity of organic compounds normalized to toluene, where toluene's SOAP value is standardized at 100 under equivalent mass emission conditions. Toluene was selected as the reference compound due to its emission characteristics and its widespread recognition as a predominant anthropogenic SOA precursor.^32,33 The methodology for calculating SOAP is shown in eqn (2).

In this study, the SOAP values for each VOC species were obtained from Derwent. For all VOC species, SOAFP was estimated by multiplying the SOAP values with the emissions of individual VOC species. The total SOAP was calculated as the sum of the SOAP values for all VOC species, as shown in eqn (3).

SOAFP_i = VOC_i × SOAP_i × FAC_toluene (3)

where VOC_i represents the mass concentration of i (µg·m⁻³)⁻¹ and FAC_toluene is the aerosol conversion factor of toluene, which is 5.4%.³⁴

2.4. Atmospheric diffusion simulation and health risk assessment

In this study, the assessment of health risks adopts a combined approach of the U.S. EPA Inhalation Risk Assessment Model and the AERMOD atmospheric dispersion model.

2.4.1. AERMOD simulation setup and input parameters. To qualitatively assess the dispersion patterns of VOCs from organized stacks and identify potential zones of influence around the breweries, the U.S. EPA's AERMOD model was employed. AERMOD is a steady-state Gaussian plume model recommended for regulatory assessments of pollutant dispersion from point sources. For organized emission sources, the AERMOD model was run using the estimated emission rates derived from the measured stack concentrations and the designed volumetric flow rate of the exhaust gas. The emission rate (Q, g s⁻¹) for a stack was calculated using the following equation:

Q = C × F (4)

where C is the measured VOC mass concentration (µg m⁻³) at the stack sampling point and F is the design volumetric flow rate (m³ s⁻¹) of the exhaust gas, which was obtained from the breweries' engineering design specifications.

The simulation utilized hourly surface meteorological data (wind speed, wind direction, temperature, cloud cover, etc.) from a nearby national weather station throughout 2022. Covering a 500 m × 500 m area centered on the brewery with a receptor spacing of 10 m, the analysis examined vertical dispersion processes from 0 to 80 m. AERMOD input parameters are shown in Table S3.

2.4.2. Health risk assessment method. The health risks to workers from fugitive emissions were assessed for occupational exposure within the plant premises using the directly measured indoor concentrations. The inhalation risk assessment model quantitatively describes the relationship between human exposure doses and adverse reactions,³⁵ utilizing the mass concentration of pollutants in ambient air as exposure assessment parameters. The exposure concentration calculation formula is shown in formula (4):In the equation, CA represents the concentration of VOCs in air, µg m³; ET denotes the exposure time, with the worker's daily working hours considered for exposure assessment, calculated at 8 hours per day; EF stands for the exposure frequency, based on the worker's annual working time over 270 days; ED signifies the exposure duration, set at 15 years; AT represents the average exposure time, equivalent to the average lifespan of 74.8 × 365 × 24 hours.^36,37

Inhalation risk assessment involves both non-carcinogenic risk and carcinogenic risk,^36,37 expressed, respectively, as the Hazard Quotient (HQ) and the Lifetime Cancer Risk (LCR). The Hazard Index (HI) is the sum of individual HQs for various pollutants. HQ is classified into five risk levels, the HQ risk level is divided into five levels, with HQ <0.1 indicating negligible risk, 0.1 < HQ < 0.5 indicating low risk, 0.5 < HQ < 1 indicating moderate risk, 1 < HQ < 2 indicating high risk, and HQ ≥2.0 indicating extremely high risk. Carcinogenic risk is divided into four levels, with compounds greater than 10⁻⁴, between 10⁻⁵ and 10⁻⁴, between 10⁻⁶ and 10⁻⁵, and below 10⁻⁶ considered as "definite," "probable," "possible," and "negligible risk," respectively. The calculation formulas are (5–7):

HI = ∑HQ_i (7)

LCR_i = EC × IUR_i (8)

where Rfc_i represents the reference mass concentration of pollutant i, mg m⁻³, and IUR_i refers to the inhalation unit risk, (µg·m⁻³)⁻¹.³⁰

3. Results and analysis

3.1. Overview of VOC concentrations

Quantifying VOC concentrations and emission contributions across production stages enables the systematic identification of dominant species and distribution characteristics. The 116 analyzed VOCs were classified into six primary groups based on their functional groups: alkanes, halogenated hydrocarbons, alkenes, oxygenated hydrocarbons, aromatic hydrocarbons, and other species including CS₂ and acetylene.³⁸ The concentrations of VOCs at different stages of the breweries and the contribution rates of various components are illustrated in Fig. 2.

Fig. 2 The concentrations of VOCs and the percentage composition of each component by mass at different stages in Brewery A (a) and Brewery B (b).

The organized emission stages include five processes of gelatinization, saccharification, filtration, boiling, and sedimentation, with significant variations observed across stages. Brewery A detected 35, 38, 47, 53, and 37 VOC species across these stages, while Brewery B demonstrated 28, 42, 28, 48, and 38 types, respectively. From Fig. 2, it can be observed that both breweries displayed peak VOC concentrations during the boiling stage, with Brewery A registering 15 225.91 ± 1912.51 µg m⁻³ and Brewery B 4752.01 ± 432.41 µg m⁻³. The high temperature during the boiling stage accelerates the escape of VOCs, and the added hops contain a large quantity of volatile terpenes, which are highly volatile compounds at high temperatures. OVOCs dominated emissions across all organized production stages, accounting for 56.6–83.1% of total VOCs in Brewery A and 62.3–83.6% in Brewery B.

The proportional contributions of the top ten VOC species in both organized and fugitive emission phases for each brewery are illustrated in Fig. S2–S5. In Brewery A, the priority pollutants include acetaldehyde, acetone, isobutane, ethyl acetate, and 2,3-dimethylbutane, with concentrations of 9913.76 ± 257.21 µg m⁻³, 2662.36 ± 98.11 µg m⁻³, 1922.89 ± 66.64 µg m⁻³, 639.84 ± 19.62 µg m⁻³ and 581.90 ± 14.31 µg m⁻³, respectively. In contrast, the characteristic species in Brewery B are acetaldehyde, ethyl acetate, 2-butanone, isobutane, and 1,1-dichloroethene, with concentrations of 2729.79 ± 80.27 µg m⁻³, 1134.16 ± 24.54 µg m⁻³, 513.70 ± 18.04 µg m⁻³, 253.74 ± 12.21 µg m⁻³, and 189.19 ± 6.39 µg m⁻³, respectively.

Fugitive emissions predominantly occur during non-enclosed operational phases, exhibiting irregular release patterns and intermittent temporal characteristics. The stages of fugitive emissions include five sources identified as fermentation units, filling stations, inkjet labeling operations, adhesive volatilization, and biological treatment tanks in wastewater facilities. Brewery A demonstrated VOC species counts of 44, 43, 40, 40, and 38 across these stages, while Brewery B exhibited 49, 39, 40, 35, and 36. The fermentation stage emerged as the dominant emission source, displaying both the highest VOC diversity and concentration levels due to elevated airflow rates, complex substrate compositions, and abundant byproduct generation. OVOCs and halogenated hydrocarbons constituted the primary components of fugitive emissions. In Brewery A, OVOCs accounted for 55.7–97.1% of total emissions, with ethyl acetate reaching 7837.77 ± 153.22 µg m⁻³ during fermentation. Acetaldehyde and 2-butanone dominated inkjet labeling emissions at 705.39 ± 21.14 µg m⁻³ and 2259.19 ± 44.63 µg m⁻³, respectively. Filling stages primarily released ethyl acetate, acetaldehyde, and 2-butanone. Brewery B showed OVOC concentrations ranging from 54.5% to 96.4%, accompanied by halogenated hydrocarbons from 1.6% to 28.1%. Fermentation emissions were dominated by acetaldehyde and ethyl acetate, collectively representing 97.0% of total OVOCs. The filling stage primarily released acetaldehyde, 1,2-dichloropropane, and ethyl acetate, while 2-butanone exhibited peak concentrations during inkjet labeling operations. The specific characteristic pollutants for fugitive emissions can be observed in Fig. S3 and S5.

The analysis results show that the overall emission characteristics of VOCs from the two breweries are similar, with OVOCs dominating and the key emission stages being boiling and fermentation. At the concentration level, Brewery A emissions are dominated by organized emissions, while brewery B is dominated by fugitive emissions. The reason for this is that Brewery A has a high capacity and a high degree of process confinement, so the organized emissions are concentrated, whereas Brewery B has a high degree of equipment openness, so there is a high percentage of fugitive emissions. Essentially, it is the difference in the degree of industrialization and traditional techniques. It is confirmed that the difference in VOC emissions from breweries are determined by a combination of production scale and process technology. The consistency in the dominant VOC species across both breweries, despite their scale and technological differences, points to their role as characteristic pollutants for the brewing industry. The prevalence of acetaldehyde and ethyl acetate is a direct signature of yeast metabolism during fermentation, while isobutane is likely a characteristic volatile released from hops during the high-temperature boiling process. Therefore, these compounds serve as key chemical indicators for brewery emissions.

3.2. Secondary pollution formation potential results

3.2.1. Ozone formation contribution. As is well known, VOCs are ozone precursors, and their contribution to chemical reactivity can be assessed through the OFP.^39,40 The OFP provides practical guidance for the prioritized control of VOCs.⁴¹ When the components and concentrations of VOCs are known, the OFP can be calculated using a formula.^42,43 It is worth noting that due to the lack of MIR values for some substances, the OFP values for certain VOCs cannot be calculated. However, these substances have concentrations small enough to be negligible and contribute minimally to the OFP. Therefore, their impact on ozone formation can be disregarded. Fig. 3 demonstrates that boiling and fermentation stages exhibit the highest OFP, driven predominantly by elevated OVOC emissions. Analysis of VOC class contributions to OFP reveals the following hierarchy: OVOCs > alkanes > halogenated hydrocarbons > alkenes > aromatic hydrocarbons > others. These findings highlight OVOC emission reduction as a critical priority for OFP mitigation strategies.

Fig. 3 OFP contribution of VOCs from the whole processes in Brewery A (a) and Brewery B (b).

OFP analysis identified distinct key reactive species contributions across emission types. During the organized emission stages in Brewery A, six compounds, including isobutane, ethyl acetate, tetrahydrofuran, acetone, 1,1-dichloroethylene, and 2,3-dimethylbutane, collectively accounted for 89.0% of the total OFP. Fugitive emissions demonstrated higher OFP dominance from ethyl acetate, 2-butanone, tetrahydrofuran, isobutane, methyl methacrylate, and 2-hexanone, contributing 93.6% of the total potential. Brewery B demonstrated different patterns with five major compounds in organized emissions, such as isobutane, ethyl acetate, tetrahydrofuran, acetone, and 1,1-dichloroethylene, driving 68.75% of the OFP. Fugitive phase emissions involved six compounds comprising 2-butanone, ethyl acetate, isobutane, tetrahydrofuran, 1,1-dichloroethylene, and cis-2-pentene, representing 63.7% of the total potential. From Fig. 3, it is evident that fugitive emissions exhibit significantly higher OFP than organized emission sources in Brewery B, which is consistent with the concentration characteristics.

This finding emphasizes the urgency of managing fugitive emission stages of VOCs in traditional semi-automated breweries. Analytical results reveal a notable discrepancy between the ten species exhibiting peak OFP values and those with maximum VOC concentrations identified in this investigation. These results indicate that VOC-induced ozone generation depends on both ambient concentration levels and compound-specific MIR coefficients. These findings highlight that reducing OVOCs from these key processes should be a priority, as they constitute the chemical mixtures with the highest intrinsic OFP per unit volume.

3.2.2. Analysis of the SOAFP results. The SOAFP of VOCs in different processes was calculated based on the concentration of each component of VOCs and SOAP, and the results are shown in Fig. 4. Comparative analysis between two breweries (Fig. S6) shows consistent patterns. Both facilities demonstrate peak SOAFP values during boiling and fermentation operations. Compositional analysis identifies OVOCs and aromatic compounds as primary contributors, with aromatics accounting for 44.89 to 75.31% of the total SOAFP from the measured VOCs despite constituting merely 0.19 to 0.70% of TVOCs (Total VOCs). Therefore, reducing the emissions of OVOCs and aromatics is more likely to reduce the generation of SOA. Fig. 5 shows the top ten species contributing to SOAFP, and it can be seen that acetaldehyde, styrene and benzene are the main contributors. Notably, high-concentration compounds like acetone and ethyl acetate exhibit limited SOA generation due to their low yield coefficients. Comparison of OFP and SOAFP rankings shows four species, including acetaldehyde, acetone, 2-butanone, and ethyl acetate, all appearing in both top 10 lists. These compounds simultaneously drive ozone and particulate matter formation, suggesting that coordinated control strategies could achieve dual air quality improvements.

Fig. 4 Comparison of the SOAFP of VOCs from the whole processes in Brewery A (a) and Brewery B (b). The colored bars (left y-axis) show the percentage contribution of each VOC category to the total SOAFP. The stars (right y-axis) indicate the absolute SOAFP value for each process.

Fig. 5 The top 10 VOC species contributing to the SOAFP in Brewery A (a and b) and Brewery B (c and d). Panels (a) and (c) show contributions during organized emission stages, while panels (b) and (d) show contributions during fugitive emission stages.

3.3. Exposure, dispersion simulation, and health risk assessment

3.3.1. AERMOD simulation results and dispersion analysis. The AERMOD simulation results based on the estimated emission inputs, provide a visualization of the dispersion pattern and relative impact of VOCs from the organized stacks. The model predicted that VOC concentrations increased from the ground level (0 m) up to a height of approximately 30 meters, reaching a maximum at this elevation. This phenomenon is primarily attributed to the stack height and the positive buoyancy and momentum of the hot exhaust plume. The plume initially rises after exiting the stack before beginning to disperse and descend due to atmospheric turbulence and wind shear. The area around 30 m height represents the zone where the plume is most likely to interact with the ground for the first time. Beyond 30 m, concentrations gradually decreased due to further dilution. Fig. 6 illustrates the simulated ground-level VOC concentration contours within a 500 m × 500 m area downwind of the breweries. The results show an elliptical pollution zone extending downwind from the source. The highest ground-level concentrations occurred a few hundred meters downwind of the stacks, rather than immediately adjacent to the breweries. Compared to Brewery B, Brewery A exhibits a more pronounced spatial extent and concentration gradients.

Fig. 6 AERMOD simulation results of VOC concentrations (µg m⁻³) within the 0–80 m height range over a 500 m × 500 m area of Brewery A (a) and Brewery B (b).

3.3.2. Health risk assessment. Using the AERMOD-predicted concentrations at the 30 m height receptor where the maximum impact was predicted, an exploratory health risk assessment for the surrounding population was conducted. The calculated HI for both breweries was below 1, and the LCR for all individual species was below 10⁻⁶. These results suggest that under the modeling assumptions and estimated emission rates, the non-carcinogenic and carcinogenic risks to the public from organized stack emissions are likely negligible. However, it is crucially important to emphasize that this is an exploratory assessment, and the absolute values of the predicted risks are highly dependent on the accuracy of the estimated emission rates.

Carcinogenic risk: the Lifetime Cancer Risk (LCR) for all individual species was below 10⁻⁶. The HQ and LCR values were used to assess the health risk of fugitive emissions to workers. For individual species, at Brewery A, acrolein and acetaldehyde had HQ values of 33.67 and 3.88, respectively, while the rest of the substances had negligible HQ values of less than 0.1 and an HI of 37.55. At Brewery B, the substances posing a non-carcinogenic risk were 1,2-dichloropropane, acrolein, and acetaldehyde, with HQs of 3.42, 25.54, and 29.06, respectively, and an HI of 58.02. The HI was 58.02. Overall, inhalation of VOCs by workers during working hours posed a non-cancer risk. For the fugitive phases, an inhalation risk assessment model was employed to calculate the risk of VOC exposure during working hours for workers. Data for non-carcinogenic risk-related substances are presented in Table 1. The HI values for both breweries are greater than 1, indicating non-carcinogenic risks. As for cancer risk, carcinogenic risk data for relevant substances are shown in Table S2. From Fig. 7, it can be observed that, except for dichloromethane, carbon tetrachloride, and ethylbenzene, the LCR for other substances is greater than 10⁻⁶. In other words, exposure to VOCs in the workplace will pose a clear risk of chronic cancer for workers. Brewery B has a greater cancer risk than Brewery A, and the health risk of the open traditional technique is seriously underestimated.

Table 1 Non-carcinogenic risk assessment results of the VOCs

Species	RFC(mg m^−3)	HQ
		Brewery A	Brewery B
Chloromethane	9.00 × 10^−2	1.45 × 10^−3	1.54 × 10^−3
Dichloromethane	6.00 × 10^−1	1.75 × 10^−4	—
1,1-Dichloroethene	2.00 × 10^−1	9.52 × 10^−3	3.78 × 10^−2
Carbon tetrachloride	1.00 × 10^−1	7.21 × 10^−5	2.02 × 10^−4
1,2-Dichloropropane	4.00 × 10^−3	3.09 × 10^−2	3.42 × 10
Acrolein	2.00 × 10^−5	3.37 × 10^1	2.55 × 10^1
Acetaldehyde	9.00 × 10^−3	3.88 × 10	2.91 × 10^1
4-Methyl-2-pentanone	3.00	1.73 × 10^−3	6.83 × 10^−4
2-Hexanone	3.00 × 10^−2	6.46 × 10^−2	6.48 × 10^−2
Propionaldehyde	8.00 × 10^−3	4.64 × 10^−2	1.45 × 10^−2
Ethyl acetate	2.00 × 10^−1	3.99 × 10^−3	3.08 × 10^−3
Methyl tert-butyl ether	3.00 × 10	2.53 × 10^−5	—
2-Butanone	5.00 × 10	2.23 × 10^−2	3.34 × 10^−2
Tetrahydrofuran	2.00 × 10	3.96 × 10^−3	3.63 × 10^−3
1,4-Dioxane	3.00 × 10^−2	1.39 × 10^−2	1.26 × 10^−2
Methyl methacrylate	7.00 × 10^−1	6.52 × 10^−4	6.43 × 10^−4
m-Xylene	1.00 × 10^−1	4.73 × 10^−4	8.51 × 10^−4
o-Xylene	1.00 × 10^−1	—	5.11 × 10^−4
Benzene	3.00 × 10^−2	5.99 × 10^−4	7.45 × 10^−3
Toluene	5.00 × 10	5.51 × 10^−5	4.08 × 10^−6
Ethylbenzene	1.00 × 10	1.57 × 10^−4	1.56 × 10^−4
Styrene	1.00 × 10	8.67 × 10^−4	8.26 × 10^−4
Carbon disulfide	7.00 × 10^−1	3.60 × 10^−3	2.61 × 10^−3
HI	—	37.75478	58.20087

---

Fig. 7 The carcinogenic risk assessment result diagram of fugitive emissions.

3.4. Recommendations and strategies for VOC control

Summarizing the analysis above, this study proposes recommendations and strategies for VOC control in two different models of breweries. For Brewery A, the following integrated strategies are proposed to enhance VOC control. (1) It is recommended to implement process optimization measures, such as introducing low-temperature yeast strains to suppress acetaldehyde formation at the source, as acetaldehyde is the primary contributor to the OFP. This strategy targets a key reactive VOC and is expected to reduce the OFP of the fermentation emissions. This preventive approach aligns with the priority of cleaner production and achieves synergistic carbon reduction by lowering the energy demand for subsequent air pollution control devices. (2) Upgraded abatement technology by adding a palladium-based catalyst bed (Pd/Al₂O₃) in the rear section of the RTO to enhance the oxidation efficiency of acetaldehyde. (3) Deploy an online monitoring system for VOCs to dynamically track key species and auto-adjust RTO operational parameters. For fugitive emissions, adopt a closed fermentation tank process to reduce the escape of VOCs, which is a foundational measure for advancing green and low-carbon development in industry. The comprehensive implementation of these strategies not only enables breweries to comply with VOC emission regulations but also advances corporate sustainability and social responsibility initiatives, contributing to the national strategic goals of Beautiful China and Carbon Peaking and Carbon Neutrality.

The comparative analysis of the two breweries demonstrates that Brewery A requires a primary focus on capacity-linked gross emission control, whereas Brewery B necessitates targeted management of fugitive process sources emitting highly reactive VOCs. This empirical evidence provides a scientific foundation for implementing differentiated regulatory frameworks within China's environmental policy, challenging the conventional one-size-fits-all approach and advocating for technology-informed, precision governance in the food industry.

3.5. Limitations

This study has several limitations that should be considered when interpreting the results. The most significant limitation is the inability to convert the measured VOC concentrations into emission rates due to a lack of air exchange rate or exhaust flow rate data. Therefore, the environmental impact potentials are presented as intrinsic properties of the source gas mixture and not as absolute contributions to the regional atmosphere. Future studies should incorporate direct emission rate measurements to accurately quantify the brewery's contribution to ambient air quality.

4. Conclusion

This study presents the first process-based analysis of VOC concentration profiles, environmental impacts and health risks in breweries of contrasting scales and technologies. The two breweries exhibited highly similar VOC species and key emission processes, with boiling and fermentation phases being identified as the dominant sources of high concentration and chemically reactive VOC mixtures. OVOCs such as acetaldehyde, ethyl acetate and ketones were the predominant groups across all stages. The assessment of chemical reactivity revealed that OVOCs possess the highest OFP per unit volume. Although aromatic hydrocarbons account for 0.19–0.70% of the mass concentration of TVOCs, they contribute as much as 44.89–75.31% to the SOAFP. This indicates that reducing OVOCs is the most effective approach to lowering the OFP, while targeting aromatic compounds is key to reducing the SOAFP. The health risk assessment yielded sharply different conclusions for workers versus the general public. AERMOD dispersion simulation results indicated negligible impacts from organized stack emissions on surrounding communities, with both carcinogenic and non-carcinogenic risks falling below thresholds of concern. However, the assessment based on direct in-plant measurements revealed significant health risks. Workers are exposed to fugitive emissions that far exceed acceptable levels for both non-carcinogenic and carcinogenic effects, with acrolein, acetaldehyde, and 1,2-dichloropropane identified as the primary drivers of risk. Consequently, VOC control strategies must be differentiated. For large modern facilities, the primary task is managing the total mass and reactivity control of concentrated organized emissions. For smaller traditional facilities, the immediate priority is controlling fugitive emissions to protect worker health. The study provides a scientific basis for developing VOC-targeted risk management strategies that prioritize the most significant and actionable exposures.

Author contributions

Huan Zhao: conceptualization, investigation, writing – original draft. Hongna Ren: investigation, visualization. Hailin Wang: formal analysis, software. Yiwen Wang: investigation. Yiming Lu: data curation. Jie Cheng: methodology, writing – review & editing, acquisition. Guoxia Jiang: writing –review & editing, investigation. Zhengping Hao: supervision.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The datasets generated during this study, including raw GC-MS chromatograms, AERMOD input parameters, and health risk calculations, are available from the corresponding authors upon reasonable request.

Supplementary information (SI): geolocation data of breweries due to confidentiality agreements with participating enterprises, process-specific emission factors, SOAP/OFP coefficients, and supplementary tables (S1–S6). See DOI: https://doi.org/10.1039/d5em00567a.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (52370123) and the Fundamental Research Funds for the Central Universities.

References

1. W. Wu, B. Zhao, S. Wang and J. Hao, Ozone and secondary organic aerosol formation potential from anthropogenic volatile organic compounds emissions in China, J. Environ. Sci., 2017, 53, 224–237.
2. E. Nie, G. Zheng, Z. Shao, J. Yang and T. Chen, Emission characteristics and health risk assessment of volatile organic compounds produced during municipal solid waste composting, Waste Manag., 2018, 79, 188–195.
3. Y. S. Huang and C. C. Hsieh, Ambient volatile organic compound presence in the highly urbanized city: source apportionment and emission position, Atmos. Environ., 2019, 206, 45–59.
4. X. Liang, X. Chen, J. Zhang, T. Shi, X. Sun and L. Fan, et al., Reactivity-based industrial volatile organic compounds emission inventory and its implications for ozone control strategies in China, Atmos. Environ., 2017, 162, 115–126.
5. C. Hsu, H. Chiang, R. Shie, C. Ku, T. Lin and M. Chen, et al., Ambient VOCs in residential areas near a large-scale petrochemical complex: Spatiotemporal variation, source apportionment and health risk, Environ. Pollut., 2018, 240, 95–104.
6. H. Xie, W. Gao, W. Zhao, Y. Han, Y. Gao and B. Liu, et al., Source profile study of VOCs unorganized emissions from typical aromatic devices in petrochemical industry, Sci. Total Environ., 2023, 889.
7. Z. Liang, Y. Yu, B. Sun, Q. Yao, X. Lin and Y. Wang, et al., The underappreciated role of fugitive VOCs in ozone formation and health risk assessment emitted from seven typical industries in China, J. Environ. Sci., 2024, 136, 647–657.
8. W. Xu, Y. Kuang, Y. Bian, L. Liu, F. Li and Y. Wang, et al., Current Challenges in Visibility Improvement in Southern China, Environ. Sci. Technol. Lett., 2020, 7(6), 395–401.
9. M. Wen, W. Deng, J. Huang, S. Zhang, Q. Lin and C. Wang, et al., Atmospheric VOCs in an industrial coking facility and the surrounding area: Characteristics, spatial distribution and source apportionment, J. Environ. Sci., 2024, 138, 660–670.
10. Z. Q. Gao, G. J. Hu, H. Wang and B. Q. Zhu, Characterization and assessment of volatile organic compounds (VOCs) emissions from the typical food manufactures in Jiangsu province, China, Atmos. Pollut. Res., 2019, 10(2), 571–579.
11. B. Lin and X. Lei, Carbon emissions reduction in China's food industry, Energy Policy, 2015, 86, 483–492.
12. L. Wang, Z. Xiang, S. Stevanovic, Z. Ristovski, F. Salimi and J. Gao, et al., Role of Chinese cooking emissions on ambient air quality and human health, Sci. Total Environ., 2017, 589, 173–181.
13. G. Zhang, S. Li, G. Pan, L. Zhu and Y. Sun, Determination and formation mechanism of precedence-controlled VOCs pollutants in chemical plant, China Environ. Sci., 2019, 39(4), 1380–1389.
14. N. R. Passant, S. J. Richardson, R. P. J. Swannell, N. Gibson, M. J. Woodfield and J. P. van der Lugt, et al., Emissions of volatile organic compounds (VOCs) from the food and drink industries of the European community, Atmos. Environ. Part A Gen. Top., 1993, 27(16), 2555–2566.
15. A. Mellouki, T. J. Wallington and J. Chen, Atmospheric Chemistry of Oxygenated Volatile Organic Compounds: Impacts on Air Quality and Climate, Chem. Rev., 2015, 115(10), 3984–4014.
16. Z. Ni, J. Liu, M. Song, X. Wang, L. Ren and X. Kong, Characterization of odorous charge and photochemical reactivity of VOC emissions from a full-scale food waste treatment plant in China, J. Environ. Sci., 2015, 29, 34–44.
17. N. B. Gibson, G. T. Costigan, R. P. J. Swannell and M. J. Woodfield, Volatile organic compound (VOC) emissions during malting and beer manufacture, Atmos. Environ., 1995, 29(19), 2661–2672.
18. H. Wang, R. Hao, L. Fang, L. Nie, Z. Zhang and Z. Hao, Study on emissions of volatile organic compounds from a typical coking chemical plant in China, Sci. Total Environ., 2021, 752.
19. R. Wang, X. Wang, S. Cheng, K. Wang, L. Cheng and J. Zhu, et al., Emission characteristics and reactivity of volatile organic compounds from typical high-energy-consuming industries in North China, Sci. Total Environ., 2022, 809, 151134.
20. G. You, H. Liu, R. Sun, Z. Jin, J. Li and X. Zhou, et al., Characterizing VOCs emissions of five packaging and printing enterprises in China and the emission reduction potential of this industry, J. Clean. Prod., 2023, 420, 138445.
21. C. Wang, W. Wang, W. Deng, S. Zhang, S. Shao and M. Wen, et al., Distribution characteristics, air-water exchange, ozone formation potential and health risk assessments of VOCs emitted from typical coking wastewater treatment process, Sci. Total Environ., 2023, 862, 160845.
22. G. Li, W. Wei, X. Shao, L. Nie, H. Wang and X. Yan, et al., A comprehensive classification method for VOC emission sources to tackle air pollution based on VOC species reactivity and emission amounts, J. Environ. Sci., 2018, 67, 78–88.
23. Z. Gao, X. Zhang, W. Zhao, X. Zhang, Z. Xue and J. Du, Characteristic Analysis of VOCs Emitted from a Typical Coking Plant, Res. Environ. Sci., 2019, 32(9), 1540–1545.
24. J. Shi, H. Deng, Z. Bai, S. Kong, X. Wang and J. Hao, et al., Emission and profile characteristic of volatile organic compounds emitted from coke production, iron smelt, heating station and power plant in Liaoning Province, China, Sci. Total Environ., 2015, 515, 101–108.
25. G. Wang, S. Cheng, W. Wei, Y. Zhou, S. Yao and H. Zhang, Characteristics and source apportionment of VOCs in the suburban area of Beijing, China, Atmos. Pollut. Res., 2016, 7(4), 711–724.
26. C. M. Da Silva, B. Siciliano, G. Dantas and G. Arbilla, An improvement of method TO-15A, aided by heart-cutting multidimensional gas chromatography, for the analysis of C2-C12 hydrocarbons in atmospheric samples, Microchem. J., 2022, 183, 108008.
27. J. Zheng, M. Shao, W. Che, L. Zhang, L. Zhong and Y. Zhang, et al., Speciated VOC Emission Inventory and Spatial Patterns of Ozone Formation Potential in the Pearl River Delta, China, Environ. Sci. Technol., 2009, 43(22), 8580–8586.
28. A. Kumar, D. Singh, K. Kumar, B. B. Singh and V. K. Jain, Distribution of VOCs in urban and rural atmospheres of subtropical India: Temporal variation, source attribution, ratios, OFP and risk assessment, Sci. Total Environ., 2018, 613–614, 492–501.
29. W. P. L. Carter, Development of Ozone Reactivity Scales for Volatile Organic Compounds, Air Waste, 1994, 44(7), 881–899.
30. W. P. L. Carter, Development of the SAPRC-07 chemical mechanism, Atmos. Environ., 2010, 44(40), 5324–5335.
31. R. G. Derwent, M. E. Jenkin, S. R. Utembe, D. E. Shallcross, T. P. Murrells and N. R. Passant, Secondary organic aerosol formation from a large number of reactive man-made organic compounds, Sci. Total Environ., 2010, 408(16), 3374–3381.
32. M. Kanakidou, J. H. Seinfeld, S. N. Pandis, I. Barnes, F. J. Dentener and M. C. Facchini, et al., Organic aerosol and global climate modelling: a review, Atmos. Chem. Phys., 2005, 5(4), 1053–1123.
33. D. Johnson, S. R. Utembe and M. E. Jenkin, Simulating the detailed chemical composition of secondary organic aerosol formed on a regional scale during the TORCH 2003 campaign in the southern UK, Atmos. Chem. Phys., 2006, 6(2), 419–431.
34. D. GROSJEAN, INSITU ORGANIC AEROSOL FORMATION DURING A SMOG EPISODE - ESTIMATED PRODUCTION AND CHEMICAL FUNCTIONALITY, Atmos. Environ. Part A Gen. Top., 1992, 26(6), 953–963.
35. L. Hui, X. Liu, Q. Tan, M. Feng, J. An and Y. Qu, et al., Characteristics, source apportionment and contribution of VOCs to ozone formation in Wuhan, Central China, Atmos. Environ., 2018, 192, 55–71.
36. H. Zhang, H. Li, Q. Zhang, Y. Zhang, W. Zhang and X. Wang, et al., Atmospheric Volatile Organic Compounds in a Typical Urban Area of Beijing: Pollution Characterization, Health Risk Assessment and Source Apportionment, Atmosphere, 2017, 8(3), 61.
37. H. Zhang, Y. Ji, Z. Wu, L. Peng, J. Bao and Z. Peng, et al., Atmospheric volatile halogenated hydrocarbons in air pollution episodes in an urban area of Beijing: Characterization, health risk assessment and sources apportionment, Sci. Total Environ., 2022, 806, 150283.
38. X. Huang, D. Han, J. Cheng, X. Chen, Y. Zhou and H. Liao, et al., Characteristics and health risk assessment of volatile organic compounds (VOCs) in restaurants in Shanghai, Environ. Sci. Pollut. Res., 2020, 27(1), 490–499.
39. L. Li, J. Y. An, Y. Y. Shi, M. Zhou, R. S. Yan and C. Huang, et al., Source apportionment of surface ozone in the Yangtze River Delta, China in the summer of 2013, Atmos. Environ., 2016, 144, 194–207.
40. L. Hui, X. Liu, Q. Tan, M. Feng, J. An and Y. Qu, et al., VOC characteristics, chemical reactivity and sources in urban Wuhan, central China, Atmos. Environ., 2020, 224, 117340.
41. C. Li, Y. Liu, B. Cheng, Y. Zhang, X. Liu and Y. Qu, et al., A comprehensive investigation on volatile organic compounds (VOCs) in 2018 in Beijing, China: Characteristics, sources and behaviours in response to O3 formation, Sci. Total Environ., 2022, 806, 150247.
42. Q. Wang, S. Li, M. Dong, W. Li, X. Gao and R. Ye, et al., VOCs emission characteristics and priority control analysis based on VOCs emission inventories and ozone formation potentials in Zhoushan, Atmos. Environ., 2018, 182, 234–241.
43. Q. Li, G. Su, C. Li, M. Wang, L. Tan and L. Gao, et al., Emission profiles, ozone formation potential and health-risk assessment of volatile organic compounds in rubber footwear industries in China, J. Hazard Mater., 2019, 375, 52–60.

---