The thraustochytrid isolate used in this study is Aurantiochytrium sp. SJB1AcBl1 is obtained from a mangrove swamp in Lian, Batangas, Philippines. The isolates are provided by the research group of the Department of Science and Technology - Philippine Council for Agriculture, Aquatic, and Natural Resources Research and Development (DOST-PCAARRD) funded project entitled “Thraustochytrid Cultivation in Wastewater for Polyunsaturated Fatty Acid Production as Alternative Fish Feed/Ingredient”. The isolates are cultured on GYPSA medium [10g/L glucose (CHEM-SUPPLY), 1.2g/L yeast extract (HIMEDIA), 1.2g /L peptone (HIMEDIA), 1,000 mL filtered 100% natural seawater or NSW, 10g/L agar] with antibiotics (100 ppm streptomycin and 100 ppm ampicillin) and antifungals (nystatin) to prevent contamination. The cultures are incubated at 25-30 °C for four days. The pure isolates are kept on GYPSA slants and subcultured every month to ensure growth and survival.

The seed culture for inoculation is prepared in sterile GYPS (30 g/L glucose, 10 g/L yeast extract, 1.25 g/L peptone, 1,000 mL NSW). The broth is inoculated with 1% v/v thraustochytrid from old stock cultures and cultivated at 150 rpm for 2 days.

The general procedure of the bench scale experiment is based on the methods of dos Reis et al. (2024) with a few modifications^[17]. The inoculum is created in a 500 mL flask with 200 mL of culture broth, prepared and cultivated according to the inoculum preparation section. To this, 2 L of adjusted DWW are added in the 5 L bioreactor vessel to operate in a batch system. The DWW characteristics are presented in Table 1. The process is non-axenic; thus, no sterilization is done beforehand to reduce costs in the process.

The inoculum (10% v/v) is added using flame ring inoculation. To mitigate foam formation, 0.5 mL of Antifoam B is added before cultivation, and also an automatic feeding of Antifoam B is also done to ensure foam generation is mitigated.

After 96 h of cultivation, the fermentation broth is harvested from the sampling port of the bioreactor, which is immediately processed through the Benchmark Scientific, C3100-E LC-8 3500 Centrifuge, China at a speed of 3,500 rpm for 8 min. The recovered biomass is washed twice using sterile distilled water at the same centrifuge conditions to remove residual medium components. The biomass is lyophilized using the table-top Martin Christ Gefriertrocknungsanlagen GmbH, Christ Beta 2-8 LSC Freeze-Dryer, Germany for 36 h. The procedure of lyophilization includes freezing for 1 h then drying for 35 h or until fully free from moisture. The total biomass yield is achieved by measuring the dry cell weight per volume of fermentation medium. Additionally, the wastewater is analyzed by its Chemical Oxygen Demand before and after thraustochytrid cultivation.

While the simulated cultivation process is modeled as non-axenic to eliminate sterilization costs, the simulation does not account for contamination risks or biological competition. In a physical bioreactor, thraustochytrids like Aurantiochytrium sp. SJB1AcBl1 can face competition from fast-growing bacterial or fungal strains in non-sterile wastewater. However, this risk is mitigated in the process design by two ways: (1) the rapid inoculation strategy (10% v/v), which gives the target strain a significant initial biomass advantage, and (2) the high salinity of the baseline medium (utilizing natural seawater), against standard freshwater wastewater bacteria.

In evaluating the environmental and economic impacts of utilizing DWW, a gate-to-gate system boundary approach is implemented. Under this approach, DWW is classified as a production waste stream and all environmental burdens and upstream processing costs from the distillery operations are cut off at the point of generation. This boundary choice is justified because DWW is an unavoidable secondary effluent that does not displace any primary market product. However, the system boundary strictly accounts for all post-generation burdens required to make the feedstock viable within the plant gate. While raw DWW enters at zero cost and footprint, the material consumption and environmental impacts associated with their subsequent in-plant pre-treatment, specifically the intensive pH adjustment using NaOH are fully quantified and integrated into the total process footprint.

The total resource consumption is quantified in terms of material and energy components. Material inputs include DWW, pre-treatment reagents, and water, while energy consumption includes steam, cooling water, electricity, and diesel. The consumption rate for each component is extracted directly from the converged DWSIM simulation.

Moreover, the unit costs for each material and energy component are obtained from literature and industry data, relevant to the Philippine setting. The total daily cost for each component obtained is aggregated to determine the total material, energy, and overall daily production cost.

For each material and energy component, there is an appropriate emission factor, in terms of kg CO₂/unit, that is sourced from the Greenhouse Gas Protocol Emission Factor Guide^[23] and other relevant literature. The footprint calculation for each component is obtained by multiplying its consumption rate by the emission factor. The individual footprints are summed up to establish the total carbon footprint for the entire process, expressed in kg CO₂/day.

The carbon footprint intensity (CI) method is employed to identify the most impactful area for emission reduction by relating the carbon footprint with each component cost. Each category (material and energy) is calculated for CI, and whichever component has a higher ratio, it is the primary target for carbon footprint reduction efforts, as it represents the largest emission source per peso spent. A goal of 7.5% reduction from the total carbon footprint is set which is based on the Philippines’ commitment to reduce 75% of total greenhouse gas emission from 2020 to 2030^[24].

To further pinpoint carbon footprint hotspots, the target component is disintegrated into internal (generated within the plant, e.g., water) and external (generated outside the plant, e.g., electricity) sources.

The COD values of DWW before and after TB cultivation are presented in Supplementary Tables 1 and 2, while the summary of the yield results is presented in Supplementary Table 3. For the simulation, the resulting DWSIM flowsheet is found in Supplementary Figure 1. The biomass yield of the DWSIM-based simulated process is compared with the experimental results done in a pilot-scale 5 L bioreactor, is shown in Table 3.

The comparison gave an error of 12.03%, which may be attributed to the difference in centrifuge split fractions due to the type of centrifuge used. The model may still be used since it is comparable with bench-scale processes and gives a percent difference of less than 20%.

The simulated model in DWSIM is solved for the process’s material and energy balances. The unit cost of each material is obtained from market surveys and literature reviews. The computation for material and energy costing is shown in Tables 4-6.

As shown in Table 4, while the utilization of DWW eliminates raw water and baseline medium costs, the cost is still heavily sensitive to the supplements: glucose, yeast extract, and peptone. This current study reflects a conservative benchmark scenario based on unoptimized laboratory-scale parameters designed to ensure maximum cellular growth. However, in an optimized industrial bioreactor cluster, these commercial supplements may be substituted for more cost-effective raw materials^[25]. Future process configurations may implement other strategies by blending DWW with other high-carbon raw materials, such as sugarcane molasses or coconut water, and utilizing localized nitrogen sources like spent brewer’s yeast to minimize dependency on the current analytical-grade components.

Material cost analysis showed that having a low-cost substrate for cultivating thraustochytrids can significantly reduce the cost of TB production. Anticipated expenses for consuming distillery wastewater are only on transportation and storage, however, in this preliminary analysis, the wastewater provides no cost. On the other hand, electricity garners the bulk of the energy cost. This is due to the use of high-cost equipment such as a centrifuge and a freeze-dryer, which consume 245.73 kWh and 480 kWh, respectively.

Table 7 shows the indicative price of selected feed ingredients in the Philippines with comparison to the TB product. The total TB product price is estimated to be Php 39.58 per gram, which is higher compared to the current prices of the bulk feed ingredients in the Philippines. This cost analysis is preliminary and includes only laboratory grade material and energy costs. Capital expenditure, labor, maintenance, depreciation, quality control, packaging, storage, transportation, and other overhead costs are not included. Currently, the price is not yet competitive. However, these figures must be contextualized by the product’s ultimate application: TB is not intended to replace bulk feed, but rather as a premium, high-value nutrient additive. Because it serves as a concentrated source of essential omega-3 fatty acids and high-quality protein, TB only needs to be blended into ordinary feeds in small quantities. The current study presents the potential for optimizing the price of TB as an omega-3 fatty acid as a protein source that can be added to ordinary feeds as nutrient additive.

The economic analysis reveals a major discrepancy between the free feedstock and the expensive supplements required by the process. While using DWW successfully eliminates baseline substrate costs, adding commercial glucose, yeast extract, and peptone to the medium accounts for 99.07% of the total daily material expenses (6,354,000 Php/day out of 6,414,000 Php/day). This shows that the economic success of this waste-to-value biorefinery is limited not by the wastewater itself, but by the expensive nutrients needed to optimize cell growth^[27]. With regards to energy consumption, electricity dominates the utility expenses, making up 98.93% of total energy costs. This concentration of expense creates a major operational bottleneck caused by high-energy equipment.

The available carbon footprint (CF) for each component is calculated from literature values of the emission factor. The computation of the carbon footprint analysis for the material and energy components is shown in Tables 8-10.

The solids contributing to the carbon footprint that are used for pretreatment and inoculum preparation is based mostly on its production process and transportation emissions. In correlation to the cost, components with high cost emit larger carbon footprints. The largest contributor to the process’s CF is electricity consumption. The Philippines is still heavily reliant on coal for its electricity production, comprising of 52% from the Philippine energy mix in 2018, totaling of 691 g CO₂ e/kWh^[31]. Efforts to switch to renewable energy sources and promoting low-carbon activities can ensure corporate sustainability to lessen emissions from product generation.

The carbon footprint composite curves are plotted from the material- and energy-based footprints against the corresponding costs of the overall process. The demand curve is based on the goal of 7.5% reduction from the total carbon footprint, corresponding to 0.45 t CO₂/day.

As shown in Figure 3A, the material-based CF (3,877.80 kg CO₂/day) is higher compared to the energy-based CF (2,097.16 kg CO₂/day), the carbon intensity of the energy segment is much higher than the material segment, as represented by their slopes. To formulate an emission mitigation plan, the study must focus on the energy segment. Figure 3B compares the composite curves based on the internal (water and steam) and external (electricity) sources of energy.

The resulting plot demonstrates that the entire source composite curve is positioned above the demand composite curve. However, adjusting the internal segment alone cannot achieve the carbon intensity target since the external footprint already exceeds this threshold. Instead, reducing the external segment’s gradient through low-carbon electricity sources presents a more viable approach to meet the reduction of electricity supply to 1.94 t CO₂/day.

The proposed carbon footprint reduction strategy to lower electricity consumption is to replace the freeze-dryer with a high-speed industrial homogenizer for cell disruption in preparation for oil extraction. This approach produces a less dried product; however, this may lead to a much moist product. Aurantiochytrium sp. wet biomass is processed using a high-speed homogenizer to disrupt cells and extract lipids simultaneously, where 80% of the lipids are removed in 10 min at a speed of 15,000 rpm^[32]. Since the TB product aims to be a feed additive, the moisture of the product will not matter as it is subjected to further processing. The homogenizer is expected to require 120 kWh of power, a significant drop from the power requirement of the freeze-dryer at 480 kWh. This substitution had a total of 4.16% of the total carbon footprint reduced, achieving the desired 7.5% reduction as shown in Figure 3C and Table 11.

The process generates a total carbon footprint of 5,974.96 kg CO₂/day, from supply chain emissions and direct electricity use. The material segment accounts for 3,877.80 kg CO₂/day, which represents the embodied carbon footprint from manufacturing glucose and peptone. In contrast, the energy segment accounts for 2,097.16 kg CO₂/day and represents the active operational carbon footprint. This operational impact is heavily multiplied by the local energy mix. Because the Philippine national grid relies on coal for over 52% of its power generation^[31], high-energy equipment directly causes intense carbon emissions. Using the CI method, the steeper slope of the energy segment proves that emissions reduction efforts must focus on cutting electricity use rather than changing materials. This target justifies replacing the energy-intensive freeze-dryer with a mechanical homogenizer. This strategy improves the overall carbon footprint of the process.

However, replacing the freeze-dryer with a mechanical homogenizer changes the final product to a wet thraustochytrid biomass paste rather than a stable, shelf-ready dry feed pellet. Mechanical homogenization via the industrial high-speed homogenizer is intended as a low-energy upstream cell disruption mechanism to maximize the polyunsaturated fatty acid (PUFA) for aquaculture species^[33]. While this unit operation does not reduce moisture content for extended shelf life, the resulting wet paste is targeted directly for immediate, localized blending within an integrated or nearby aquafeed manufacturing facility.

This study assumes an idealized scale-up from the 5 L bench-scale bioreactor to the 1,000 L commercial reactor and is modeled using a simplified, deterministic volumetric scaling approach. The model does not dynamically account for real-world bioreactor constraints such as oxygen transfer limitations, gas-liquid mass transfer gradients, mixing time delays, and impeller shear stress. The model also assumes stable culture growth and uses a fixed stoichiometric growth equation based on optimized laboratory parameters to calculate conversion metrics. It does not consider variations, time-dependent microbial population kinetics, and contamination. The system boundary assumed in the study for the mass and energy balances is gate-to-gate, the DWW enters at zero cost and footprint.

For the process change recommendation, switching from a freeze-dryer to a homogenizer results in 4.16% operational energy reduction within the plant gate, there is a trade-off downstream. The resulting product is a wet biomass paste rather than a dry powder. The shelf-life stabilization and thermal dewatering burdens required for long-term storage following the recommendation are excluded from the scope of this study. The economic costing performed in this study serves as a preliminary evaluation focused exclusively on operational energy and material consumption. It excludes capital expenditures (CAPEX), labor, maintenance, equipment depreciation, etc. In addition, the baseline economic and environmental viability assumes an integrated biorefinery or co-location of the plant to justify a zero-cost and zero-footprint transport for the raw distillery wastewater effluent. The costs used in this study is based on the actual costs from the laboratory scale. It is expected that the costs will decrease when the economies of scale is applied.