Managing Sustainable Nitrogen Removal from Wastewater Carbon Sources for Denitrification Master’s thesis in Infrastructure and Environmental Engineering & Industrial Ecology LINNEA KJELLÉN & SOFIA SJÖSTEDT DEPARTMENT OF ARCHITECTURE AND CIVIL ENGINEERING CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2025 www.chalmers.se MASTER’S THESIS ACEX30 Managing Sustainable Nitrogen Removal from Wastewater - Carbon Sources for Denitrification Master’s Thesis in Infrastructure and Environmental Engineering and Industrial Ecology LINNEA KJELLÉN SOFIA SJÖSTEDT Department of Architecture and Civil Engineering Division of Water and Environment Technology CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2025 Managing Sustainable Nitrogen Removal from Wastewater Carbon Sources for Denitrification Master’s Thesis in Infrastructure and Environmental Engineering & Industrial Ecology LINNEA KJELLÉN SOFIA SJÖSTEDT © LINNEA KJELLÉN & SOFIA SJÖSTEDT, 2025 Examensarbete ACEX30 Institutionen för arkitektur och samhällsbyggnadsteknik Chalmers tekniska högskola, 2025 Department of Architecture and Civil Engineering Division of Water and Environment Technology Chalmers University of Technology SE-412 96 Göteborg Sweden Telephone: + 46 (0)31-772 1000 Cover: Cover picture taken by photographer Emelie Asplund, used with consent. Department of Architecture and Civil Engineering Göteborg, Sweden, 2025 I Managing Sustainble Nitrogen Removal from Wastewater Carbon Sources for Denitrification Master’s thesis in Infrastructure and Environmental Engineering & Industrial Ecology LINNEA KJELLÉN SOFIA SJÖSTEDT Department of Architecture and Civil Engineering Division of Water and Environment Technology Chalmers University of Technology ABSTRACT External carbon sources are essential for effective nitrogen removal in biological post- denitrification at wastewater treatment plants (WWTPs). However, the commonly used fossil-based methanol contributes significantly to the climate impact of WWTPs. This study examined both technical and sustainability aspects of transitioning to more sustainable carbon sources, focusing on methanol and ethanol. Denitrification performance during carbon source transitions was evaluated in laboratory-scale moving bed biofilm reactors (MBBRs), designed to replicate full-scale conditions. Methanol, ethanol, and a methanol/ethanol mixture were tested in a system acclimated to methanol. Results showed comparable performance between the mixture and methanol alone. No major differences were observed when switching from methanol to ethanol. Switching back to methanol resulted in a temporary decrease in performance, suggesting a short acclimation period may be needed. In parallel, a sustainability assessment using a multi-criteria analysis framework was conducted to develop a basis for future evaluation of fossil-free methanol and ethanol alternatives. Relevant options and criteria across ecological, social, and economic dimensions were defined, and both qualitative and quantitative data were collected. Based on the findings, three of the 11 criteria are considered ready for direct application in future assessments. One criterion was excluded due to a confirmed lack of impact, while the remaining seven require additional data or further investigation. All alternatives are recommended for continued evaluation, except recycled ethanol, which requires further testing due to uncertainties related to impurities. Key words: wastewater treatment, nitrogen removal, denitrification, moving bed biofilm reactor, external carbon source, multi-criteria analysis II Hantering av hållbar kväverening från avloppsvatten Kolkällor för denitrifikation Examensarbete inom masterprogrammet Infrastruktur och miljöteknik samt Industriell ekologi LINNEA KJELLÉN SOFIA SJÖSTEDT Institutionen för arkitektur och samhällsbyggnadsteknik Avdelningen för Vattenmiljöteknik Chalmers tekniska högskola SAMMANFATTNING Externa kolkällor används för effektiv kväverening vid biologisk efterdenitrifikation vid avloppsreningsverk. Fossilbaserad metanol som ofta används bidrar dock avsevärt till verkens klimatpåverkan. Denna studie undersökte både tekniska aspekter och hållbarhetsaspekter av övergången till mer hållbara kolkällor, med fokus på metanol och etanol. Denitrifikationsförmågan mellan övergångar av kolkällor utvärderades i laboratorieexperiment som använde sig av Moving bed biofilm reactors (MBBR), utformade för att replikera fullskaliga förhållanden. Metanol, etanol och en blandning av metanol och etanol testades i system som acklimatiserats till metanol. Resultaten visade jämförbar förmåga mellan blandningen och enbart metanol. Inga större skillnader observerades vid övergång från metanol till etanol. Övergången tillbaka till metanol resulterade tillfälligt i lägre denitrifikationsförmåga, vilket tyder på att en kort acklimatiseringsperiod kan behövas. Parallellt genomfördes en hållbarhetsanalys, som delvis följer ramverk för multikriterieanalys, för att utveckla en grund för framtida utvärdering av fossilfria metanol- och etanolalternativ. Relevanta alternativ och kriterier över ekologiska, sociala och ekonomiska dimensioner definierades, och både kvalitativ och kvantitativ data samlades in. Baserat på resultaten anses endast tre av de elva kriterierna vara redo för direkt tillämpning i framtida bedömningar. Ett kriterium bekräftades inte innebära någon påverkan och kan exkluderas framöver, medan de återstående sju kriterierna kräver ytterligare data eller vidare utredning. Alla alternativ rekommenderas för fortsatt utvärdering, förutom återvunnen etanol, som kräver ytterligare testning på grund av osäkerheter relaterade till föroreningar. Nyckelord: avloppsrening, kväveborttagning, denitrifikation, moving bed biofilm reactor, extern kolkälla, multikriterieanalys III Acknowledgements We would like to express our sincere gratitude to our supervisor at Chalmers and our mentors at Gryaab for their enthusiasm, valuable insights, and continuous support throughout the project. Linnea Kjellén & Sofia Sjöstedt, Gothenburg, 2025 IV Contents 1 INTRODUCTION 1 1.1 Background 1 1.2 Aim and objectives 2 1.3 Limitations 3 2 THEORY 4 2.1 Wastewater treatment 4 2.1.1 Biological nitrogen removal 5 2.1.2 Moving bed biofilm reactor systems 8 2.1.3 Rya wastewater treatment plant 8 2.2 Fossil free versions of methanol and ethanol 10 2.3 Multicriteria analysis 11 3 METHOD 12 3.1 Laboratory experiments 12 3.1.1 Experimental setup 12 3.1.2 Sampling, laboratory analysis and data evaluation 16 3.2 Sustainability analysis 18 3.2.1 Aim formulation 18 3.2.2 Specification of alternatives and prerequisites 19 3.2.3 Selecting and defining criteria 19 3.2.4 Data acquisition 20 3.2.5 Results and reflections of the evaluation basis 21 4 RESULTS AND DISCUSSION 22 4.1 Laboratory experiments 22 4.1.1 Influent and carbon dosage flows 22 4.1.2 Organic carbon 23 4.1.3 Ammonium 27 4.1.4 Nitrite 27 4.1.5 Nitrate 28 4.1.6 Phosphorus 32 4.1.7 Suspended solids 33 4.1.8 Uncertainties 33 4.1.9 Relevance to full-scale systems and previous research 35 4.1.10 Contributions to Research and Future Directions 36 4.2 Multicriteria analysis 36 4.2.1 Available data from suppliers 37 4.2.2 Alternatives and prerequisites 37 4.2.3 Chosen criteria 40 4.2.4 Evaluation basis 42 4.2.5 Uncertainties and limitations 55 V 4.2.6 Further directions relating to the evaluation basis 57 5 CONCLUSION 59 6 REFERENCES 61 7 APPENDICES 67 Appendix A 67 Appendix B 68 Appendix C 71 Appendix D 72 Appendix E 75 Appendix F 76 1 1 Introduction Today, wastewater treatment plants (WWTPs) face the dual challenge of adapting to more stringent effluent requirements while lowering their carbon footprint. A significant part of this challenge involves rethinking the use of fossil-based chemicals in treatment processes. This thesis explores how such use can be improved through both practical experiments and a broad sustainability assessment with environmental, social and economic perspectives. The work revolves around a specific WWTP but still provides insights for other treatment plants. 1.1 Background Nitrogen removal is a critical component of modern wastewater treatment, mainly for preventing eutrophication of oceans and to return nutrients to agricultural processes (Swedish Environmental Protection Agency, 2022). The transformation of nitrogen in wastewater into nitrogen gas can be achieved through nitrification and denitrification processes, such as those implemented in moving bed biofilm reactors (MBBRs) at Rya WWTP (Gryaab, n.d.-a). It is the treatment plant in Gothenburg, operated by Gryaab, that treats wastewater in the larger Gothenburg region. WWTPs must manage increased loads driven by climate change and population growth while adapting to stricter discharge regulations and striving for greater sustainability (Neth et al., 2022; Sapmaz et al., 2022). Decreasing the use of fossil-based carbon sources is crucial for decreasing the carbon footprint of a WWTP (Gustavsson & Tumlin, 2013). Fossil-based methanol is commonly used as carbon source, for instance in MBBR systems, due to its low cost and high efficiency but poses environmental concerns and potential toxicity risks at high concentrations (Sapmaz et al., 2023; Wang et al., 2021). According to Gryaab (2025), the fossil-based methanol used as external carbon source is the third largest contributor to the climate impact of Rya WWTP, corresponding to 6 000 tons of CO₂-e per year. Reducing the amount of fossil-based methanol required can be achieved by switching to a non-fossil-based carbon source entirely, or by introducing it in a mix with the methanol. There are many different types of carbon sources you can use, but easily biodegradable chemicals, low-molecular weight are preferable (Fu et al., 2022). In line with this and what is believed to work best at Gryaab, only methanol and ethanol are relevant to this context. In particular, bio-based ethanol emerged as the best potential fossil free carbon source from previous projects including market analysis and simpler sustainability comparisons. Since ethanol is a well-established carbon source (Fu et al., 2022), there is no point in direct comparisons of methanol and ethanol. Further, the technical function with bio-based or fossil based should be the same, hence not requiring testing how bio-based versions compare to fossil versions. How the presented solution with switching to ethanol or to a mix of ethanol and methanol works in practice remains to be determined. To the authors’ knowledge this has not been examined in MBBR systems before. Examining the transition from ethanol back to methanol, is also relevant to know more about the implications of switching overall and can potentially help WWTPs make more informed choices of carbon source in case of market availability shifts. 2 There are many different product alternatives of fossil-free methanol and ethanol available on the market. While the previous comparisons at Gryaab brought forth potential candidates, and to some extent included comparisons between aspects such as cost, technical performance and safety, a more in-depth analysis of the performance and environmental, social and economic sustainability is needed. Ultimately, the challenge is to determine how switching between methanol and ethanol works in a biological system acclimated to methanol as well as determining which products and sustainability aspects should be considered. This is what this thesis will cover and therefore aid in the research of more sustainable wastewater treatment. 1.2 Aim and objectives The aim of this thesis is to aid WWTPs in transitioning to sustainable external carbon source usage in post-denitrification processes, by examining process performance when switching carbon sources and assessing relevant carbon source alternatives through sustainability perspectives. There are two objectives. The first objective is to evaluate, through lab-scale experiments, whether ethanol can be successfully introduced into a post-denitrification process originally designed for methanol, followed by a return to methanol. The comparison includes methanol acting as a reference, alternating between methanol and ethanol, and alternating between a mix of methanol and ethanol. The sub-objectives related to this part are: • Evaluate key performance indicators for the carbon sources alternatives such as denitrification efficiency, denitrification rate and carbon utilization. • Assess whether the carbon source transitions have considerable impact on the treatment process. • Compare and relate performance indicators to existing literature and full-scale WWTP data. • Provide recommendations for further research on carbon source substitution in denitrification processes. The second objective is to evaluate and compare several carbon source alternatives through a sustainability analysis, covering environmental, social, and economic dimensions. This is performed using a multicriteria analysis (MCA) approach, with the aim of developing an evaluation basis for future selection. The related sub-objectives are: • Identify and define relevant external carbon source alternatives. • Establish a comprehensive list of sustainability criteria for evaluation. • Gather qualitative and quantitative data to create the evaluation basis. • Identify key areas for further research and provide guidance on which aspects WWTPs should prioritize for future investigation. 3 1.3 Limitations The thesis focuses on carbon source alternatives best suited for Rya WWTP. The process and results are therefore influenced by local considerations relevant for Gryaab. The sustainability analysis includes carbon source alternatives selected from a prior market analysis at Gryaab and does not cover all available products on the market. While the multicriteria analysis (MCA) structure was followed, actual weighting, scoring, and aggregation of results were not performed within the scope of this study. The evaluation basis developed with the MCA methodology in this thesis aim to support future evaluation and decision-making processes. The laboratory scale experiments focused on the nitrogen removal efficiency. Internal reactor conditions (e.g., pH, temperature variations) and external conditions (e.g., inflow fluctuations) were not fully monitored. Sampling was limited to grab samples. 4 2 Theory This section presents the theoretical background relevant for the thesis. It gives an overview of wastewater treatment in general, and with a focus on biological nitrogen removal. It also describes the treatment process design of Rya WWTP including the treatment system moving bed biofilm reactor (MBBR). The post-denitrification MBBR system of Rya WWTP is presented in more detail. The subject regarding different types of fossil-free versions of the carbon sources methanol and ethanol is briefly introduced. Finally, the general methodology of multicriteria analysis is outlined to give context to the sustainability analysis in this thesis. 2.1 Wastewater treatment Water is an important resource in human societies, and it is thus important to maintain its quality. The water taken from ground- or surface waters will ultimately be returned to the environment after human use (Metcalf & Eddy, 2004; Sonune & Ghate, 2004). This used water is called wastewater. To avoid environmental pollution, wastewater must be treated before being released back into water bodies. Wastewater can originate from domestic facilities such as households, offices, and schools, but also from industries (Sonune & Ghate, 2004). Other types are stormwater, urban runoff caused by precipitation, and water infiltrating sewage pipes through cracks or other defects. These are part of the municipal sewer system. Wastewater generated from industrial processes may need pre-treatment before reaching the municipal sewer system or require separate handling. Generally, the main constituents of wastewater are different types of organic material, nutrients, pathogens and metals (la Cour Jansen et al., 2021). To remove these, a combination of physical, chemical and biological processes is used (Sonune & Ghate, 2004). To represent different levels of treatment at WWTPs, the terms preliminary, primary, secondary, and tertiary treatment are commonly used (Meng et al., 2024; Sonune & Ghate, 2004). Which constituents that are removed at which treatment level, and examples of treatment methods for different constituents are summarised in Table 2.1. Table 2.1. Summary of the conventionally used treatment levels, which wastewater constituents that is removed and which main processes that are commonly used. Compiled with information from Metcalf & Eddy (2004), Meng et al. (2024), Jansen et al. (2019), & Sonune & Ghate (2004). Level of treatment Constituent/pollutant removed Main treatment process Preliminary Larger, solid constituents such as grit and sticks, as well as grease Physical/mechanical Primary Suspended solids Physical Secondary Colloidal and dissolved organic material and nutrients Biological and chemical Tertiary Remaining small particles, pathogens Varies. Combinations of the above or other methods. 5 Biological wastewater treatment is based on the principle of microorganisms, mostly bacteria, transforming or removing compounds such as organic matter and nutrients present in the wastewater (Metcalf & Eddy, 2004). According to la Cour Jansen et al. (2019), the two main biological treatment methods are activated sludge and biofilters. In activated sludge, reactors contain a sludge consisting of living biomass (bacteria and other microorganisms) as well as inorganic and organic particles. The sludge is kept suspended by stirring or aeration. In biofilter methods, the bacteria are instead densely fixated onto a surface in the form of a so-called biofilm. Sufficient contact between the biofilm and the wastewater is important for pollutant removal and there are different techniques as to how this can be ensured. One is where the medium onto which the biofilm grows is fixed and the water is distributed onto the medium, as in the technology trickling filters; another is where the medium is moving through water, as in the technology moving bed biofilm reactor (MBBR). These biological methods can be used for removal of organic matter and the nutrient nitrogen. Municipal wastewater is along with agriculture one of the main sources of excessive environmental concentrations of nitrogen (Rahimi et al., 2020). High concentrations of nitrogen being released into freshwater environments can lead to eutrophication. This can in turn lead to algal blooms and oxygen depletion, which ultimately negatively affect aquatic life. Hence, removal of the nitrogen from wastewater is of importance. 2.1.1 Biological nitrogen removal Nitrogen naturally occurs in several forms and oxidation states, and is a crucial element for microorganisms, plants and animals (Metcalf & Eddy, 2004). Its most commonly occurring forms are ammonia (NH3), ammonium (NH4 +), nitrogen gas (N2) and nitrate (NO3 -). There are multiple routes in which the nitrogen species are converted from one form to another. The different forms of nitrogen and the pathways between them are typically described with the nitrogen cycle, a simplified version of this is depicted in Figure 2.1. Utilisation of the two processes nitrification and denitrification is commonly used in biological nitrogen removal methods. 6 Figure 2.1. Simplified version of the nitrogen cycle, inspired by la Cour Jansen et al. (2021). Nitrification is a biological process consisting of two steps, where the first step involves ammonium, NH4 +, being oxidised to nitrite, NO2 -, by ammonia-oxidising bacteria (AOB) (Rahimi et al., 2020). Secondly, nitrite-oxidising bacteria (NOB) oxidise nitrite to nitrate, NO3 -, in the presence of oxygen. The second step is faster than the first, making the first reaction the limiting step. The bacteria performing the nitrification steps are groups of autotrophic AOB and NOB, using oxygen as the final electron acceptor (Rahimi et al., 2020). They gain their energy from the ammonia oxidation and acquire carbon from CO2. In the denitrification process, bacteria convert nitrate into nitrogen gas, N2, through the creation of several intermediate products (Meng et al., 2024). The reaction chain can be summarised as follows: 𝑁𝑂3 − → 𝑁𝑂2 − → 𝑁𝑂 → 𝑁2𝑂 → 𝑁2 The denitrification process is anaerobic, since nitrate is used as electron acceptor (Jansen et al., 2019). Many of the denitrifying bacteria can use oxygen as terminal electron acceptor as well, and therefor anoxic conditions are required for the denitrification process to take place (Rahimi et al., 2020; Jansen et al., 2019). Denitrification can be performed by both autotrophic and heterotrophic bacteria, however, heterotrophic denitrifies are more commonly occurring (Rahimi et al., 2020). Heterotrophic bacteria, in contrast to autotrophs, use organic carbon for growth of bacterial cell material (Metcalf & Eddy, 2004). This organic carbon is necessary for denitrification by heterotrophs and cell growth of the bacteria (Rahimi et al., 2020). The organic carbon used in the denitrification process can either be already present in the incoming wastewater, generated within the wastewater treatment system (at the plant) due to bacterial cell decay, or externally added to the system (Swinarski et al., 2009). External source of organic carbon usage is usually needed when the organic matter content in the influent water is too low, or when the denitrification treatment process is separated from and placed after an aerated treatment process, since the organic material gets consumed in these processes (Fu et al. 2022; la Cour Jansen et al., 7 2021). Post-denitrification is a used term for when the denitrification treatment process is placed after the aerated nitrification process. That carbon source substrates have low molecular weight is an important property, since it generally correlates with higher biodegradability by bacteria (Fu et al., 2022). Commonly used carbon sources are for example methanol, ethanol, acetate and glucose (Rahimi et al., 2020). The populations of denitrifying bacteria vary depending on what carbon source is used (Meng et al., 2024). Some populations are more abundant using ethanol, methanol or acetate while others are only more abundant when glycerol is used. The ratio between organic material and nitrogen (C/N ratio) is one of the important parameters affecting the denitrification process (la Cour Jansen et al., 2021). It directly affects the denitrification rate, since it represents the ratio between electron donor and electron acceptor (Güven, 2009). Generally, a C/N-ratio of about 3.5–4.5 is considered sufficient for complete denitrification. Three commonly used measures of the organic content in water are chemical oxygen demand (COD), biological oxygen demand (BOD) and total organic carbon (TOC) (Metcalf & Eddy, 2004): • COD is defined as the oxygen equivalent to the organic material available for chemical oxidation in a water sample (Metcalf & Eddy, 2004). There is no clear definition of soluble and particulate fractions, however, a general partitioning can be made through filtering of the sample. The COD-concentration can potentially include oxidation of inorganic substances as well, due to the measuring method. • BOD is defined as the amount of dissolved oxygen consumed by microorganisms for biochemical oxidation of organic material (Metcalf & Eddy, 2004). The measurement is dependent on time, with a measurement taken over 5 days being labelled as BOD5 (Jouanneau et al., 2014). In Sweden, BOD7 is the standard, whereas in most other parts of Europe, BOD5 is more common (Swedish Agency for Marine and Water Management, 2016). • TOC is defined as the total organic carbon present in a water sample (Metcalf & Eddy, 2004). Dissolved organic carbon (DOC) is the measured organic content in which the sample have been filtered through a 0.45 µm pore size filter. Too low C/N-ratio in the denitrification process medium can lead to reduced denitrification rates or only a partial reaction (la Cour Jansen et al., 2021). This can also result in an increase in intermediate products such as dinitrogen oxide and a decrease in treatment of nitrate. Increased emissions of dinitrogen oxide are something to be avoided, since it is a potent greenhouse gas significantly stronger than carbon dioxide (Zhou et al., 2025). Each intermediate in the denitrification reaction chain can act as electron acceptor, creating competitive conditions for available electron donors (Güven, 2009). The reduction rates of the different reaction steps, such as the rate of nitrate being transformed into nitrite, are affected by the bacterial culture and the wastewater composition. Generally, however, the nitrite reduction rate (NO2 - to NO) is slower than the nitrate reduction rate (NO3 - to NO2 -). This inconsistency can cause accumulation of nitrite in the denitrification medium. Microbial species composition affects nitrite accumulation in the denitrification process (Du et al., 2016; Meng et al., 2024; Rahimi et al., 2020). Some denitrifying bacteria only inhibit the enzymes for 8 reduction of nitrate to nitrite, and not further steps of the reaction chain. If these are abundant, nitrite accumulation is more likely. Other factors affecting nitrite accumulation are limited carbon, pH, nitrate concentration, oxygen concentration and toxic compounds (Rahimi et al., 2020). 2.1.2 Moving bed biofilm reactor systems Moving bed biofilm reactor (MBBR) is one type of biofilm system that can be used for biological wastewater treatment. The main incentive for its development was the need of a compact technology with limited mechanical issues, which earlier biofilm technologies were prone to having (Dezotti et al., 2017; di Biase et al., 2019; Madan et al., 2022). In biofilm, microorganisms such as bacteria grow on surfaces in a cluster of bacterial cells and extracellular polymeric substances (Dezotti et al., 2017). This creates a protective environment for the microbial community of the biofilm. Additionally, it contains a high variety of microbial functional groups which generally can imply a greater ability for wastewater pollutant removal. The MBBR technology consists of biofilm growing on small carriers, usually plastic in a hollow cylinder form, which are kept suspended in the water of the reactor (Madan et al., 2022). Reactors with anoxic conditions need mechanical stirring, while in aerobic reactors biocarrier movement can often be caused by the aeration process. In contrast to treatment processes where the biomass is suspended directly in the water (activated sludge), there is no need for recirculation of biomass due to the biofilm being fixed onto the carrier surface and hence remains in the system (Dezotti et al., 2017). This allows the microbiome to become specialised at the system’s intended purpose, such as nitrogen removal. The MBBR technology have other advantages against methods where the biomass moves freely in the reactor water. A high biomass concentration leads to a higher treatment capacity per volume, resulting in a compact process where smaller reactor volumes can be used (Madan et al., 2020). Additionally, biofilm systems are more resistant to varying conditions such as fluctuations in temperature, pH, and influent characteristics. However, both operation and investment costs of the MBBR system can be high. Maintenance can be expensive if aeration is used for both providing of oxygen and stirring of biocarriers, as well as the initial costs of biocarriers and reactor construction. Stagnant zones can appear if the hydrodynamics of the reactor is not given enough thought, but to which extent these zones affect the treatment capacity is uncertain (di Biase, 2019). 2.1.3 Rya wastewater treatment plant Rya wastewater treatment plant (WWTP) receives wastewater from 7 municipalities in the Västra Götaland-province, approximately 800 000 people (Gryaab, n.d.-a). The facility removes BOD, phosphorus and nitrogen from the wastewater. Figure 2.2 show the processes and how they are connected. 9 The incoming wastewater passes through several treatment processes before being released into the recipient Göta Älv river. Firstly, bar screens remove larger parts such as paper and plastics. The water then goes through primary settling before the biological treatment of BOD and nitrogen and simultaneous precipitation of phosphorus. The biological treatment consists of activated sludge, nitrifying trickling filters, post- nitrifying moving bed biofilm reactor (MBBR) and post-denitrifying MBBR. The final treatment step is disc filters. An overview of the treatment processes is seen in Figure 2.2. Sludge recirculation occurs within the activated sludge process. Some sludge is processed again through the treatment plant. The produced sludge that leaves the process steps is treated and can be used for agricultural purposes. Figure 2.2. Process diagram of the water treatment processes at Rya WWTP. The post-denitrification is a MBBR system which consists of six basins, so-called lines, each divided into three zones, shown in Figure 2.3. The zones are divided by strainers which allows flow of wastewater but hinders biocarriers. Nitrified water from the trickling filters and nitrifying MBBR is mixed and then distributed into the six lines of the post-denitrification. 10 Figure 2.3. Illustration of the lines of the denitrification MBBR process at Rya WWTP, including dosage pumps (PU). The zones have a filling ratio of 55–58 % carriers, corresponding to about 350–370 m3. The active area of the biocarriers, that is, the area of the biocarriers onto which the biofilm can grow, is 500 m2/m3. Phosphoric acid and external carbon source, currently methanol, is added to the system. These are crucial for the denitrifying bacteria and hence the function of the denitrification process. The external carbon source flow is regulated by the incoming wastewater flow, the incoming nitrate concentration and a set effluent nitrate concentration, as well as the C/N-ratio. Both the carbon source and the phosphoric acid is added into Zone 1. The largest removal of nitrate happens in Zone 1. 2.2 Fossil free versions of methanol and ethanol Bio-based methanol and ethanol can be produced from biomass that is converted to biofuels through various process. These are categorized as first-generation biofuels which are produced from edible biomass and second-generation biofuels which are produced from non-edible biomass (Bhaskar & Pandey, 2015; Darda et al., 2019; El- Araby, 2024). The first-generation biofuels can be based on for example corn, wheat or sugarcane as feedstock. The second-generation biofuels are made of lignocellulosic biomass from agricultural and forest residues and crops grown for biofuel purposes (Bhaskar & Pandey, 2015; Darda et al., 2019). The forest residues may stem from logging operations or industrial processing of forest biomass (Bhaskar & Pandey, 2015). Methanol and ethanol can also originate from recycled chemicals, by recovering and processing solvents used in other industries, such as the pharmaceutical sector (Chea et al., 2019). Furthermore, an increasingly common procedure used in the chemical industry is mass-balancing (ISCC, n.d.). It allows the bio-based or recycled feedstocks 1 2 3 4 5 6 Zone 1 Zone 2 Zone 3 PU PU PU PU PU PU 11 to be tracked through the whole value chain despite it being mixed with other types of feedstocks. 2.3 Multicriteria analysis A multicriteria analysis (MCA), or multicriteria decision analysis (MCDA), is a structured decision-making tool used to compare different alternatives based on multiple criteria (Belton & Stewart, 2002; Neth et al., 2023; UK Government, 2009). It is commonly used in sustainability assessments and decision-making processes where options must be evaluated from several perspectives, such as environmental, economic, and social perspectives. The general MCA process involves defining the problem and objectives, followed by identifying a set of alternatives to achieve those objectives (Belton & Stewart, 2002; Neth et al., 2023; UK Government, 2009). A set of criteria is selected to assess the performance of each alternative. Information of the alternatives is acquired so that each alternative can be scored according to its performance in each criterion, typically using a consistent numerical scale. Weighting is then applied through choosing numerical coefficients to reflect the relative importance of the criteria or differences in performance. An overall score for each alternative is achieved by multiplying its score in each criterion by the corresponding weight, then summing across all criteria. This provides a basis for ranking or selecting alternatives. Finally, the results should be analysed. It is recommended to conduct a sensitivity analysis to tests how changes in scores, weights, or assumptions affect the results which in turn can help assess the robustness of the conclusions. How an MCA process is carried out or by whom can vary. Typically, the process is supported by facilitated workshops involving key stakeholders, especially during scoring and weighting stages. 12 3 Method This section presents the methodology used to compare different external carbon sources through laboratory experiments and a sustainability analysis. The experiments are primarily technical while the sustainability analysis covers several perspectives. The alternatives examined in each part differ, but the approaches together offer a comprehensive basis for studying possibilities of switching to lower climate impact carbon source. 3.1 Laboratory experiments The laboratory experiments investigated the effects of methanol and ethanol as carbon sources in the post-denitrification process. The setup used aimed to reflect the first zone of the post-denitrification process at Rya WWTP as most of the denitrification occurs in zone 1. The methodology for the setup, sampling and analysis and evaluation of the carbon source performance are described below. 3.1.1 Experimental setup The experiments were conducted using three laboratory-scale continuous stirred tank reactors (CSTR), specifically Dolly Biogas Reactors from Belach Bioteknik AB. Each reactor had a total volume of 10 litres. Biocarriers (AnoxKaldnes K1) acclimated to methanol from the first zone of the full-scale MBBR were added. A filling fraction of approximately 50% was used, consistent with full-scale conditions. The reactors were continuously supplied with influent used in the post-denitrification process. The water passed through an influent buffer tank before being distributed to the reactors via three peristaltic pumps. Each reactor was equipped with an upper and lower stirring blade which was set at constant mixing. All effluent directed to the plant’s wastewater treatment system. See Figure 3.1 for the actual setup and Figure 3.2 for a schematic overview. 13 Figure 3.1. The experimental setup with reactors and influent arrangement. Figure 3.2. Schematic illustration of experimental setup showing what was done in lab-scale and how it relates to the full-scale wastewater treatment plant (WWTP). Including reactor labels (GA1, GE1 and GE2) and flow arrows for water (blue) and chemicals (black). The experiments were conducted over a period of 82 days between February and May 2025. The average water temperature during the experiment was 16–17°C. The carbon dosing timeline is summarized in Table 3.1. Initially all reactors were first dosed with methanol to ensure comparability and stable reactor conditions. After this, the first reactor alternated between methanol and ethanol (GA1), the second reactor continued receiving methanol (GE1) and the third reactor alternated between methanol 14 and a mixture of methanol and ethanol (GE2). Transitions from methanol to alternative carbon sources and back were performed once stable reactor performance had been confirmed through analytical results. Table 3.1. Timeline of experiment showing the duration each reactor was dosed with each carbon source. Period Duration GA1 GE1 GE2 Period 1 Day 1 to 22 Methanol Methanol Methanol Period 2 Day 22 to 50 Ethanol Methanol Methanol/ethanol mix Period 3 Day 50 to 73 Methanol Methanol Methanol Period 4 Day 73 to 82 Increased methanol dosage The reactor GE1 which only received methanol, served as the reference in the experiment. This was done to approximate the standard operation at the full-scale treatment plant, where methanol is the carbon source. Since the experimental setup did not include online measurements nor chemical dosing based on incoming nitrate, the reference reactor was critical for interpreting the results. It provided a baseline against which the other configurations used in GA1 and GE2 could be compared to. The reactor GA1 that switched entirely to ethanol is intended to reflect a direct implementation of ethanol in the post-denitrification process. For the mixture in reactor GE2, a methanol/ethanol ratio of 60/40 was chosen to reflect an incremental implementation instead where methanol is still the main carbon source, but the amount is decreased. To understand the impacts of these transitions more and investigate the possibility to go back to methanol, it is switched to methanol once again. Methanol and ethanol with concentrations >99% were used, sourced from both fossil- based and bio-based origins depending on availability. Due to their high purity, no functional differences in performance were expected between the fossil-derived and bio-based variants of either methanol or ethanol. Additionally, phosphate (85% ortho- phosphoric acid) was added throughout the experiment to ensure biofilm growth. All chemicals were diluted with deionized water and stored in containers beneath the reactors and prepared throughout the experiment. The carbon sources were diluted to 25% as a standard but were increased to 50% in the last period to investigate the effects of increased carbon dosage. The phosphoric acid was diluted to 5%. The influent and chemical flow rates were determined based on full-scale design criteria. This meant aiming for a hydraulic retention time around 30 minutes and a C/N- ratio of 4.7. Based on this, the influent flow was set to 250 ml/min. The chemical dosing was controlled via computer-regulated peristaltic pumps. They operated at a constant speed of 27 rpm and activated intermittently based on pre-set operator parameters. Preliminary test runs using water and various operational settings were conducted to calibrate the dosage pumps. To match the full scale, the diluted phosphoric acid required a flow of 0.001 ml/min, but the minimum flow the pumps provided were 0.006 ml/min. This flow was used, resulting in excess phosphorous in the reactors which was assumed not to affect the process. For carbon dosage, a flow of 15 0.06 ml/min calculated for methanol diluted to 25% was used for all carbon source configurations. An explanation for this is given at the end of this sub-section. The corresponding added COD concentrations and C/N ratios that are achieved with the setup are calculated with Equation 3.1 and Equation 3.2 respectively. 𝐶𝑂𝐷 𝑎𝑑𝑑𝑒𝑑 [ 𝑚𝑔 𝑙 ] = 𝑄𝑐 ∙ 𝜌𝐶 ∙ 𝑑𝑖𝑙𝑢𝑡𝑖𝑜𝑛 ∙ 𝐶𝑂𝐷 𝑐𝑜𝑛𝑡𝑒𝑛𝑡 𝑄 3.1 𝐶 𝑁 = 𝑄𝐶 ∙ 𝜌𝐶 ∙ 𝑑𝑖𝑙𝑢𝑡𝑖𝑜𝑛 ∙ 𝐶𝑂𝐷 𝑐𝑜𝑛𝑡𝑒𝑛𝑡 (𝑐𝑁𝑂3 −−𝑁,𝑖𝑛 + 𝑐𝑁𝑂2 −−𝑁,𝑖𝑛) ∙ 𝑄 3.2 𝑄 is the target flow to each reactor [l/min], 𝑄𝐶 is the target flow of carbon source to each reactor [l/min], 𝜌𝐶 is the density of carbon source [mg/l], 𝑐 is the incoming concentration of NO3 —N and NO3 --N [mg/l]. The dilution is 0.25 (25%) for Period 1-3 and 0.5 (50%) for Period 4. The density and COD-content of ethanol and methanol is seen in Appendix A. The density and COD content of the mix are approximated with the methanol/ethanol proportions (60/40). Assuming the correct influent and dosage flows are achieved, and that the incoming nitrogen is constant at 13 mg/l, based on WWTP data for the same time period as the experiments, the theoretical COD and C/N values can be seen in Table 3.2. Each parameter is the same for all reactors in Period 1 and 3 when they received methanol (with 25% dilution). The C/N ratio of 4.9, is slightly above the target ratio to have enough carbon. It is the same for GE1 in Period 2 since it continued with methanol. The COD and C/N ratio are doubled when the methanol dilution instead is 50% in Period 4. Because the flows were not altered for the reactors receiving ethanol in Period 2, the added COD and C/N ratio are higher than in the methanol reactor. In this period the C/N ratio theoretically reaches 7.3 for GA1 which received ethanol and 5.8. for GE2 that received the mixture. Table 3.2. Theoretical COD added mg/l and theoretical C/N-ratio corresponding to the target carbon dosage flow for each reactor in each period. Period 1 = all reactors on methanol, Period 2 = GA1 – ethanol, GE1 – methanol, GE2 – mixture of methanol and ethanol. Period 3 = all reactors on methanol. Period 4, all reactors on methanol with increased concentrations. GA1 GE1 GE2 Period COD C/N COD C/N COD C/N Period 1 63.2 4.9 63.2 4.9 63.2 4.9 Period 2 94.4 7.3 63.2 4.9 75.7 5.8 Period 3 63.2 4.9 63.2 4.9 63.2 4.9 Period 4 126.4 9.7 126.4 9.7 126.4 9.7 The dosing with ethanol does in fact require lower flows due to ethanol having higher COD content than methanol. Specifically, the required flow rate for ethanol and the mixture corresponding to the methanol flow rate was 0.041 ml/min and 0.52 ml/min, respectively. Although they differ, the flow rates for each carbon source were not adjusted. The dosing pumps lacked the precision needed to accurately deliver such 16 small flow differences because there were inconsistencies with the provided flows. The intention was to keep the flows constant throughout the experiment, including during the increased dosage phase, where only the dilution of the carbon source was altered. Efforts were made to operate as close to the intended flows as possible. This included tubing maintenance and measuring the carbon flow rates at several points of the experiment. Additionally, to achieve target flows with the determined dilutions, parameters had to be set so that the chemicals were added every 16 minutes, and therefore not continuous dosing. 3.1.2 Sampling, laboratory analysis and data evaluation Grab samples were collected each morning, five times per week. Samples were collected directly from the reactor outlets and the main influent stream and subsequently filtered through a 0.45 µm filter. The inflow to each reactor was measured during sampling. The analytical parameters and corresponding methods are summarized in Table 3.3. Table 3.3. Analytical parameters and the analytical methods used in the experiment. Parameter Analytical method Nitrate (NO₃-N) Using Hach Lange LCK 339/340 Nitrite (NO₂-N) Using Hach Lange LCK 341 Ammonium (NH₄-N) Using Hach Lange LCK 304 Phosphate (PO₄-P) According to SS-EN ISO 6878:2005 Chemical Oxygen Demand (COD) Using Hach Lange APC114/814. Dissolved Organic Carbon (DOC) Using Shimadzu TOC-V analysis. Total suspended solids (TSS) According to SS-EN 872:2005 with modification (microwave drying) All methods required filtration, including COD. Nitrogen compounds were analysed 3– 5 times per week. During transitions between carbon sources, additional samples for nitrogen analysis were taken in the afternoon over several days to monitor stabilization. DOC was analysed 2–3 times per week. COD, PO₄-P and suspended solids were analysed once per week. After sampling, COD- and PO₄-P-samples were sent to the Rya WWTP laboratory for analysis. Samples for nitrogen species and DOC were stored in a freezer at -18°C and were slowly thawed and shaken prior to analysis. The analysis of total suspended solids (TSS) includes using Equation 3.3. TSS were measured as an indicator of biomass production, to see if there was a difference between methanol and ethanol. 𝑇𝑆𝑆 [ 𝑚𝑔 𝑙 ] = 𝑑𝑟𝑦 𝑤𝑒𝑖𝑔ℎ𝑡 𝑜𝑓 𝑟𝑒𝑠𝑖𝑑𝑢𝑒 𝑎𝑛𝑑 𝑓𝑖𝑙𝑡𝑒𝑟 − 𝑑𝑟𝑦 𝑤𝑒𝑖𝑔ℎ𝑡 𝑜𝑓 𝑓𝑖𝑙𝑡𝑒𝑟 𝑠𝑎𝑚𝑝𝑙𝑒 𝑣𝑜𝑙𝑢𝑚𝑒 3.3 Phosphorous were only analysed to control if the reactors received enough nutrient. Concentrations measured from the other analyses were used in calculations for further 17 interpretation of the results: denitrification rate, nitrate reduction and parameters related to carbon content. Denitrification rates of the biofilm in the reactors were calculated according to Equation 3.4 and Equation 3.5. 𝑟𝑁𝑂3 − = (𝑐𝑁𝑂3 −−𝑁,𝑖𝑛 − 𝑐𝑁𝑂3 −−𝑁,𝑜𝑢𝑡) ∙ 𝑄 𝑉 ∙ 𝑎𝑠 3.4 𝑉 = 𝜑 ∙ 𝑉𝑟 3.5 Where rNO3 - is the denitrification rate [g/m2·d], 𝑄 is the flow rate [l/d], 𝑐 is the in-and outflow NO3 --N concentrations [g/l], 𝑉 the wet volume of the bioreactor [m3], 𝑎𝑠 the specific surface area of the biocarriers [m2/m3], 𝜑 the filling degree of the biocarriers and 𝑉𝑟 the total volume of the reactors [m3]. Nitrate reduction was calculated in percentage, as demonstrated in Equation 3.6. 𝑁𝑂3,𝑅𝑒𝑑𝑢𝑐𝑡𝑖𝑜𝑛 − = (𝑐𝑁𝑂3 −−𝑁,𝑖𝑛 − 𝑐𝑁𝑂3 −−𝑁,𝑜𝑢𝑡) 𝑐𝑁𝑂3 −−𝑁,𝑖𝑛 ∗ 100 3.6 Where 𝑐 is the in-and outflow NO3 --N concentrations [g/l]. The COD input concentrations and C/N-ratio that are theoretically achieved in the setup (presented above in Table 3.2), was calculated with Equation 3.1 and 3.2. To better analyse the carbon content achieved in the experiment, the same equations are used to determine the added COD and C/N ratio depending on measured values. This means the same equations are used but with a change of input parameters where 𝑄 instead is the measured flow to each reactor [l/min], 𝑄𝐶 is the measured flow of carbon source to each reactor [l/min] and 𝑐 is the measured incoming concentration of NO3 —N and NO3 - -N [mg/l]. To get the corresponding DOC input concentrations a COD/DOC ratio calculated with the measured concentrations of COD and DOC. It is only the average ratios in each reactor within Period 2 that are used to get the COD/DOC ratios for methanol, ethanol and the mixture. Thereafter the estimated DOC is calculated with Equation 3.7 𝐷𝑂𝐶 𝑎𝑑𝑑𝑒𝑑 [ 𝑚𝑔 𝑙 ] = 𝐶𝑂𝐷 𝑎𝑑𝑑𝑒𝑑 𝐶𝑂𝐷 𝐷𝑂𝐶 𝑟𝑎𝑡𝑖𝑜 3.7 Concerning COD or DOC input, the added concentration in each reactor is added to the respective incoming concentration to achieve a total carbon input of either COD or DOC in each reactor. This is compared to the effluent COD or DOC concentrations to study how much carbon is used related the total carbon content. 18 The results of the experiment were supplemented with current and historical data from the full-scale WWTP, obtained through internal databases and targeted measurement campaigns. These data were used to complement the experimental results regarding influent conditions. Access to the WWTP’s routine influent monitoring allowed for a greater number of data points than those captured through the experiment’s grab samples alone, as it included measurements from days without experimental sampling. Additionally, it was used for evaluating the findings by assessing expected variability in incoming wastewater, as well as what can be considered typical effluent values, particularly in the first post-denitrification zone this setup aimed to replicate. In addition, a literature review was conducted to further interpret and contextualize the experimental results by comparing them to other studies involving methanol and ethanol in MBBR systems or other biological nitrogen removal processes. 3.2 Sustainability analysis The methodology to assess carbon source alternatives through sustainability perspectives partially follows a multicriteria analysis (MCA) approach, and it will be referred to as such throughout the thesis. The key steps include formulation of the aim, specification of alternatives, definition of the criteria set, and data acquisition. Finally, the findings are summarized and reflected upon. The outcome is intended to serve as a structured evaluation basis for comparing the proposed alternatives. While a complete MCA typically includes scoring the alternatives, and weighting criteria to produce a composite sustainability score, this thesis focuses on establishing the foundation for such an evaluation rather than performing the full MCA. The alternatives, prerequisites, and criteria were gradually developed in parallel with data collection. This iterative structure meant that the different components of the assessment evolved together. Data gathering informed the selection of criteria, while the selected criteria and alternatives also guided which types of information were sought. Therefore, more details on methodology are presented in the Results and discussion section, but the main methodology and overall approach are outlined here. Note that the alternatives in the MCA are separate from the configurations tested in the laboratory experiment. The experimental results were not part of the main basis for the MCA, although they provided some points of discussion for certain alternatives. 3.2.1 Aim formulation The aim formulated specifically for the sustainability analysis was to develop an evaluation basis for comparing bio-based or recycled alternatives to the fossil-based methanol currently used in post-denitrification at the Rya WWTP, considering ecological, social, and economic aspects. The analysis concerns consumption over a one-year period, with 2024 used as the reference year. 19 3.2.2 Specification of alternatives and prerequisites Initial concepts for potential alternatives (various forms of ethanol and methanol) were identified based on previous studies and a market analysis performed at Gryaab 2024, and consultations with Gryaab personnel. This analysis mainly covers specific product alternatives but alternatives reflecting design options were considered as well, where representative products were chosen. The alternatives were determined and further specified during the project through continued consultations and reviewing previous internal reports regarding the carbon source alternatives. This included determining the specific products to be included in each alternative and how they are meant to be implemented at the plant and other prerequisites. Similar products for a certain supplier or from different suppliers were combined into one alternative and representative information were chosen for each alternative with more than one product. The alternatives are described in the thesis, but the suppliers are not mentioned here for confidentiality reasons. 3.2.3 Selecting and defining criteria The development of the criteria list involved reviewing Gryaab’s prior studies and market analysis, as well as academic literature regarding MCA applications in water treatment contexts. This literature review served to understand commonly used criteria in comparable assessments and provided a conceptual basis for structuring and justifying the selection of criteria in this study. Initially, a broader list of criteria was compiled (Appendix B), capturing general sustainability aspects that could be of interest when choosing carbon source at a WWTP. This list draws significant inspiration from Backeström and Ceder (2022) who established a comprehensive set of criteria for sustainability assessments in the water and environmental sectors. Their list together with criteria used previously at Rya WWTP (when comparing technologies for pharmaceutical removal and future effluent compliance) served as key references for identifying relevant criteria in this study. The criteria proposed by Backeström and Ceder (2022) were intended for water-related MCAs, and some were applied specifically in their evaluation of infiltration and water reuse. Besides expanding that list with criteria and applying own interpretations, the context of the included studies is also emphasised here. Authors such as Malmquist (2006), Marques et al. (2015), and Zheng et al. (2016) provided comprehensive criteria lists suggested for water services and sustainable urban water management. Criteria were also identified in more specific MCA contexts, including soil remediation alternatives (Rosén et al., 2015), retrofitting of WWTPs (Machado et al., 2020) and resource recovery from water bodies (Johannesdottir et al., 2021). Additional contributions in the current work were found related to sustainability assessment of sludge management (Sabet et al., 2025), MCAs comparing WWTP technologies (Castillo et al., 2016; Omran et al., 2021), and an MCA on industrial water reuse (Isaac et al., 2022). Although some studies focus on contexts other than WWTPs the criteria are considered relevant to the topic and may offer valuable insights for sustainability assessments when evaluating carbon sources. 20 For selecting the criteria to be included in this thesis, the relevance of each criterion in the general criteria list was evaluated based on comparability and significance, meaning that alternatives could be meaningfully assessed using the same criteria and focusing on aspects expected to vary across the alternatives. The selection was also based on available data for the products provided by the suppliers (see Appendix C). The criteria selection process is detailed further in the Results and discussion, where the included and excluded criteria are discussed under each sustainability dimension. Each selected criterion was then clearly defined and categorized under one of the three sustainability dimensions environmental, social, or economic. 3.2.4 Data acquisition Data to support the developed evaluation basis were collected through a combination of methods. Consultations were held with Gryaab personnel to gain operational insights into how carbon sources are handled at the facility, including aspects of implementation, daily operations, and supply logistics. Discussions also addressed how new products and sustainability efforts are communicated internally and externally, particularly in relation to media, public interest, and stakeholder engagement. Relevant internal studies conducted by Gryaab were reviewed, including a market analysis. These documents provided foundational information for evaluating existing options and supplier characteristics. To complement existing knowledge and gather updated or missing information, targeted questionnaires were distributed during this thesis work. These were adapted to gather both new information and follow-up data where prior responses were incomplete. Appendix C summarizes the content of the market analysis and the questionnaires. A targeted literature review was conducted to supplement practical insights and provide a broader context. This review covered topics including: • Denitrification processes and external carbon source application (particularly within MBBR systems), • Environmental and safety impacts of various carbon source types • Usage trends and reported preferences from other WWTPs, • Environmental implications such as recipient water impacts, sludge characteristics, and resource use • Health and safety considerations relevant to WWTP operations. The first topic was reviewed in parallel with the literature work supporting the experimental component of the thesis. While some sources contributed to both aspects, they are applied differently depending on context. Calculations were carried out based on supplier data, literature, and other collected inputs. These calculations were used to support selected criteria with assumptions stated where necessary. Due to the evolving nature of the analysis, calculations cannot be 21 described upfront, as they are closely tied to the iterative definition of criteria. They are brought up in the Results and discussion and in presented in detail in Appendix D. 3.2.5 Results and reflections of the evaluation basis The data were gathered using structured templates to ensure systematic compilation. However, some qualitative and discussion-based data could not be easily quantified. In such cases, these findings were included to reflect uncertainties, highlight aspects that are difficult to assess at present, or identify areas where further clarification is expected in future evaluations. Although formal MCA scoring and weighting were not performed, practical considerations for applying the evaluation basis in Gryaab’s continued work or other WWTPs considering alternative external carbon sources are discussed. 22 4 Results and discussion This chapter presents the results and discussions for the laboratory experiments and the steps of the MCA. Discussions are integrated throughout the results where relevant but are further developed in dedicated sections at the end of each part. In the MCA section it is highlighted how the findings related to the laboratory experiments informed the sustainability assessment. 4.1 Laboratory experiments The results from the laboratory experiments are presented and discussed. It includes measured parameters and calculated performance indicators. Data points where the inflow occasionally was completely stopped were removed. The results are compared to findings in the literature and to data from the full-scale WWTP. The influence of the experimental setup on the results is discussed, along with sources of uncertainty. Based on the findings, recommendations are provided for future research on the transition of carbon source in denitrification processes. 4.1.1 Influent and carbon dosage flows Influent flows to the reactors are shown in Figure 4.1. The target flow was 0.25 l/min, and most inflows were within a range of 0.24–0.27 l/min, apart from some outliers. Reactor GE1 had the most operational problems with regards to inflow, as can be seen in the diagram below. Figure 4.1. Influent flow to the reactors. Vertical lines indicate changed carbon dosage (Switch to ethanol means for GA1 - methanol to ethanol, GE1 - Methanol 0.05 0.10 0.15 0.20 0.25 0.30 0 7 14 21 28 35 42 49 56 63 70 77 84 [l /m in ] Days Influent flow GA1 GE1 GE2 Switch to ethanol Switch to methanol Increased C 23 reference, GE2 - Methanol to mix of ethanol and methanol. All reactors are dosed with methanol after the second line). Table 4.1 show that the measured carbon dosage flows. The measurements are quite uncertain due to difficulties in accurately setting and measuring the small flows, so these should be seen as estimated averages. However, it can be stated that the flows generally are slightly lower than the target flow of 0.06 ml/min for all reactors. It was seemingly always lower in reactor GA1. Only measurements after the prior testing were included in these calculations. The average flows in Period 1 might be a bit higher than the averages presented here, depending on how long the initial pump calibration stayed consistent. For Period 4, adjustments of the tubing finally gave flows closer to the target flow. Table 4.1. Average carbon dosage flows for each reactor and period in ml/min. GA1 GE1 GE2 Period 1 0.038 0.048 0.050 Period 2 0.039 0.045 0.047 Period 3 0.047 0.055 0.051 Period 4 0.054 0.065 0.055 4.1.2 Organic carbon Figure 4.2 and Figure 4.3 show the measured COD and DOC concentrations in the influent, the effluent of the reactors and the calculated total carbon input. The average COD/DOC ratios calculated for Period 2 that was used for calculating the added DOC were 2.75 mg/l for methanol, 2.56 mg/l for ethanol and 2.71 for the mix. No substantial changes in effluent COD or DOC concentrations were observed after switching carbon sources. If carbon was being used inefficiently after a switch, one would expect clear patterns of increased organic content in the effluent compared to before the switch, but this was not seen in the transition points. 24 Figure 4.2. Concentrations of chemical oxygen demand in influent and reactor effluent. Including total input of carbon (incoming concentration + added carbon). Vertical lines indicate changed carbon dosage (Switch to ethanol means for GA1 - methanol to ethanol, GE1 - Methanol reference, GE2 - Methanol to mix of ethanol and methanol. All reactors are dosed with methanol after the second line). Figure 4.3. Concentrations of dissolved organic carbon in influent and reactor effluent. Including total input of carbon (incoming concentration + added carbon). Vertical lines indicate changed carbon dosage (Switch to ethanol means for GA1 - methanol to ethanol, GE1 - Methanol reference, GE2 - Methanol to mix of ethanol and methanol. All reactors are dosed with methanol after the second line). 0 20 40 60 80 100 120 140 160 180 0 7 14 21 28 35 42 49 56 63 70 77 84 [m g /l ] Days Chemical oxygen demand (COD) concentration In GA1 GE1 GE2 Switch to ethanol Switch to methanol Increased C In+added, GA1 In+added, GE1 In+added, GE2 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 0 7 14 21 28 35 42 49 56 63 70 77 84 [m g /l ] Days Dissolved organic carbon (DOC) In GA1 GE1 GE2 Switch to ethanol Switch to methanol Increased C In+added, GA1 In+added, GE1 In+added, GE2 25 Occasional peaks in reactor effluent concentrations of COD and DOC are observed at specific times. These correspond to periods of reduced flow due to operational issues with the influent pumps. While these outliers exist, the general trends are more relevant. DOC concentrations are typically between 10–20 mg/L and COD concentrations are generally around 37 mg/L. For each reactor and period, DOC and COD show parallel fluctuations patterns in the influent and effluent on the dates where both are sampled, see Figure 4.4. Figure 4.4. Concentrations of dissolved organic carbon (DOC) and chemical oxygen demand (COD) in the influent and each reactor. The y-axis is the concentrations in mg/l and x-axis the days. Vertical lines indicate changed carbon dosage (Switch to ethanol means for GA1 - methanol to ethanol, GE1 - Methanol reference, GE2 - Methanol to mix of ethanol and methanol. All reactors are dosed with methanol after the second line). The COD measurements are consistently higher than DOC because COD captures all oxidizable substances, including inorganic and non-dissolved organics, whereas DOC only includes dissolved organic carbon. Note that both DOC and COD analyses were performed on filtered samples. As a result, COD does not provide information about particulate matter. Effluent COD concentrations are typically just above the background level, indicating that most of the added carbon was used effectively and that there is enough carbon. Effluent DOC, on the other hand, is sometimes lower than in the influent. A lower effluent DOC than influent suggest excessive consumption of dissolved available carbon, potentially indicating carbon limitation. There are also instances where effluent DOC exceeds the estimated total influent DOC. There are not measurements for both DOC and COD on these days to compare them. They are likely related to lower flows or possibly measurement variability. 26 To see if the carbon concentrations in the reactors varied due to the intermittent dosing regimen, additional sampling of DOC at several times during the 16-minute dosing cycle was performed. It was observed that the added carbon reaches a peak in the first few minutes and is then consumed during the cycle, ending up just above or just below the incoming concentration (see Appendix E). Unlike the reactor-samples, the influent was just analysed once for this date. However, the incoming concentration is not expected to vary in such a short time frame. Following this test, samples were consistently taken just before new carbon dosing, to capture conditions near the end of the cycle. This change in sampling strategy was intended to reduce variability and possibly clarify both carbon and nitrogen trends, though this was not clearly reflected in the data. While short-term variability is visible, these differences likely do not impact the overall conclusions. The calculated C/N-ratios are seen in Figure 4.5. Considering the WWTP guideline of 4.7, values well below this likely indicate underdosing. In Period 4, the ratio increased significantly when switching to higher carbon dosing. The decreasing trend that follows is due to slightly higher incoming flow and nitrogen. Figure 4.5. Estimated added carbon to incoming nitrogen ratio. Vertical lines indicate changed carbon dosage (Switch to ethanol means for GA1 - methanol to ethanol, GE1 - Methanol reference, GE2 - Methanol to mix of ethanol and methanol. All reactors are dosed with methanol after the second line). It was established in Section 4.1.1, that each period is affected by variations in influent and carbon dosage flows and that they differ from the target flows. This causes some uncertainties with all calculations dependent on these flows (added COD, added DOC and C/N ratio). 0 1 2 3 4 5 6 7 8 9 10 11 0 7 14 21 28 35 42 49 56 63 70 77 84 C /N Days C/N ratio GA1 GE1 GE2 Switch to ethanol Switch to methanol Increased C 27 4.1.3 Ammonium Influent and effluent concentration of NH4 +-N for the reactors can be seen in Figure 4.6. These concentrations stayed relatively low for large parts of the experiment. Higher effluent concentrations seem to be connected to higher influent concentrations. Figure 4.6. Influent and effluent NH4+-N-concentrations for the reactors. Vertical lines indicate changed carbon dosage (Switch to ethanol means for GA1 - methanol to ethanol, GE1 - Methanol reference, GE2 - Methanol to mix of ethanol and methanol. All reactors are dosed with methanol after the second line). 4.1.4 Nitrite Influent and effluent concentration of NO2 --N for the reactors can be seen in Figure 4.7. Most of the measured concentrations are between 0.035 and 1.7 mg/l. This variation is similar to an earlier measurement campaign in the full scale MBBR denitrification from 2018–2019, where nitrite concentrations in Zone 1 varied between 0.023 and 1.8 mg/l. The range is therefore considered reasonable, but the results also show that effluent nitrite concentrations increased throughout the test period for all three reactors. After increasing the external carbon source dosage, the concentrations started declining. This indicates that the nitrite accumulation was caused by a too low carbon source dosage. 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 0 7 14 21 28 35 42 49 56 63 70 77 84 [m g /l ] Days NH4-N Concentration In GA1 GE1 GE2 Switch to ethanol Switch to methanol Increased C 28 Figure 4.7. Influent and effluent NO2--N-concentrations for the reactors. Vertical lines indicate changed carbon dosage (Switch to ethanol means for GA1 - methanol to ethanol, GE1 - Methanol reference, GE2 - Methanol to mix of ethanol and methanol. All reactors are dosed with methanol after the second line). Rocher et al. (2015) explains differences in nitrite accumulation at different C/N ratios. While nitrite accumulation occurs whenever there is a lack of carbon, it may be high residual NO3 - and little NO2 - in cases with strong carbon deficiencies. This is because both the denitrification rate and the production of intermediate species is slow. With moderate underdosing, the concentration of NO2 - is higher. This ratio is supposedly sufficient for removing nitrate but not for eliminating intermediate NO2 - in the process. However, Du et al. (2016) found that the C/N-ratio did not influence the nitrite accumulation. The authors note that this differs from earlier experiments that have shown increased nitrite accumulation for high and low C/N-ratios. Since this study switched between methanol and ethanol, it is difficult to know if the increasing nitrite concentrations were only caused by a low C/N-ratio. Nitrite accumulation could be caused by a combination of factors. For instance, it can involve aspects such as pH, nitrate concentration and presence of oxygen or toxic compounds (Rahimi et al., 2020). However, the results strongly indicate that the increase in nitrite concentration over time mainly was caused by a low dosage of carbon source. 4.1.5 Nitrate Figure 4.8 show the influent and effluent concentrations of NO3 --N. The nitrate concentration is consistently reduced from the influent to effluent. Although the reduction varies, this indicates the biofilm have denitrification capacity throughout the entire experiment. 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 0 7 14 21 28 35 42 49 56 63 70 77 84 [m g /l ] Days NO2-N Concentration In GA1 GE1 GE2 Switch to ethanol Switch to methanol Increased C 29 Figure 4.8. Influent and effluent NO3--concentrations for the reactors. Vertical lines indicate changed carbon dosage (Switch to ethanol means for GA1 - methanol to ethanol, GE1 - Methanol reference, GE2 - Methanol to mix of ethanol and methanol. All reactors are dosed with methanol after the second line). Looking at the reduction of NO3 --N-concentration shown in Figure 4.9, a decline in reduction can be observed in the reactor operating on only ethanol, GA1, after the switch back to methanol. There is also a reduction after the switch to ethanol, however, there is large variation in values throughout that period. For the reactor receiving a mix of ethanol and methanol, the nitrate reduction seems to be unaffected by the switches. Another observation is dips in reduction occurring prior to any changes in dosing, which are believed to be caused by flow, stirring or dosing inconsistencies. One aspect to consider is that all switches occurred on Mondays, where these issues were usually worse since there were no maintenance of the reactors during the weekends. This potentially makes these concentrations less representative of the impacts of switching carbon source. 0 2 4 6 8 10 12 14 16 18 0 7 14 21 28 35 42 49 56 63 70 77 84 [m g /l ] Days NO3-N Concentration In GA1 GE1 GE2 Switch to ethanol Switch to methanol Increased C 30 Figure 4.9. Removal efficiency (%) of NO3--N-concentration for the reactors. Vertical lines indicate changed carbon dosage (Switch to ethanol means for GA1 - methanol to ethanol, GE1 - Methanol reference, GE2 - Methanol to mix of ethanol and methanol. All reactors are dosed with methanol after the second line). The lower reduction in GA1, the reactor that switched between methanol and ethanol dosing, seems to persist for two to three days until it increases again, as seen in Figure 4.10. This finding is put in relation to other studies, described further down in Section 4.1.9. Figure 4.10. Removal efficiency (%) of NO3-N concentration for the reactors. Focus on the days around the switch back to methanol. GA1 - Ethanol, GE1 - Methanol reference, GE2 - Mix of ethanol and methanol. 0 10 20 30 40 50 60 70 80 90 100 0 7 14 21 28 35 42 49 56 63 70 77 84 [% ] Days NO3-N Reduction GA1 GE1 GE2 Switch to ethanol Switch to methanol Increased C 0 20 40 60 80 100 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 [% ] Days NO3-N Reduction, Switch back to methanol GA1 GE1 GE2 Switch to ethanol Switch to methanol Increased C 31 Figure 4.11 show how the denitrification rates changed over time for the different reactors. A decrease in the denitrification rate was observed in GA1 when switching from ethanol back to methanol, following a pattern comparable to the one seen in nitrate reduction. Figure 4.11. Denitrification rates for the reactors. Vertical lines indicate changed carbon dosage (Switch to ethanol means for GA1 - methanol to ethanol, GE1 - Methanol reference, GE2 - Methanol to mix of ethanol and methanol. All reactors are dosed with methanol after the second line). The mean values of the denitrification rate along with standard deviation are summarized in Table 4.2. Table 4.2. Mean denitrification rate in NO3-N/m2d and standard deviation (std). Period 1 - All reactors on methanol, Period 2 - GA1 on ethanol, GE1 on methanol, GE2 on ethanol- and methanol mix. Period 3 - All reactors on methanol. Period 4 - All reactors on methanol, increased dosage. GA1 GE1 GE2 Period Mean Std Mean Std Mean Std Period 1 1.3 0.2 1.3 0.5 1.1 0.5 Period 2 1.6 0.4 1.6 0.4 1.6 0.4 Period 3 1.4 0.4 1.6 0.3 1.4 0.3 Period 4 1.8 0.3 1.5 0.6 1.5 0.7 The denitrification rates seem to increase from Period 1 to Period 2, however, they increase for all reactors and not only the ones receiving ethanol. Thus, it is unlikely that the change in carbon source is the main cause for the increase in denitrification rates. For GA1 and GE2, the reactors receiving ethanol and mix of ethanol and methanol, the denitrification rates decrease when switching back to methanol. In an earlier pilot MBBR study, denitrification rates were reported to be higher for ethanol than methanol 0.00 0.50 1.00 1.50 2.00 2.50 3.00 7 14 21 28 35 42 49 56 63 70 77 84 [g N O 3 -N m ⁻² d ⁻¹ ] Days Denitrification rates GA1 GE1 GE2 Switch to ethanol Switch to methanol Increased C 32 (Bill et al., 2009). However, as seen in Figure 4.11 and the standard deviation values (Table 4.2), there is quite a large variation in values during the experiment. The mean values of denitrification rate for the whole experiment were 1.55 g NO3- N/m2d for GA1, 1.57 for GE1 and 1.50 for GE2. The mean value of denitrification rate in Zone 1 of the full-scale denitrification MBBR for 2024 was 1.9 g NO3-N/m2d. Based on this, it seems like the denitrification rate of neither of the reactors are reaching the levels of the full-scale denitrification process. However, given the small scale of the laboratory reactors and other differences between the lab- and full-scale, the results indicate a relatively high level of treatment performance. 4.1.6 Phosphorus Phosphorus, PO4-P, was measured once a week to make sure there was enough for bacterial growth. The results can be seen in Figure 4.12. Figure 4.12. Measured PO4-P for the three reactors. As mentioned in the methodology, phosphoric acid was overdosed in the reactors. This was assumed to not impact the results. For the first week, the difference between the influent and effluent concentrations are quite large. After the second week and the rest of the experiment, the difference is smaller. This could indicate that the biomass adapted to the higher dosage of phosphorus and started using higher amounts. However, the lack of consistent samples makes the indications uncertain. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 0 7 14 21 28 35 42 49 56 63 70 77 84 m g /l Days PO4-N GA1 GE1 GE2 In 33 4.1.7 Suspended solids Suspended solids were measured during the experiment and some variations were observed, see Figure 4.13. However, due to the limited number of data points and the lack of clear trends or correlations, no meaningful conclusions could be drawn. Other studies using lab-scale MBBRs found that more biomass was produced when using ethanol as carbon source (Cheng et al., 2023; Torresi et al., 2017). This could not be seen from the TSS parameter in this study. It was visually observed though, and through the increase in reactor maintenance required, that biomass increased over time in all reactors. Figure 4.13. Concentration of suspended solids in the effluent. Vertical lines indicate changed carbon dosage (Switch to ethanol means for GA1 - methanol to ethanol, GE1 - Methanol reference, GE2 - Methanol to mix of ethanol and methanol. All reactors are dosed with methanol after the second line). 4.1.8 Uncertainties Uncertainties in this study arise from the absence of statistical analysis and certain limitations in the experimental setup. Statistical analysis was not performed due to a limited number or samples for certain periods and analytical parameters, the overall exploratory nature of the study, and time constraints. This limits the certainty of the results, particularly with respect to variations within and between reactors and during transition periods. The experimental design aimed to reflect realistic operational conditions at a real WWTP and to test whether added carbon could be fully utilized, particularly during carbon source transitions. However, in practice, it became evident that a more robust and precise setup would have been beneficial. A main challenge was to achieve accurate and continuous carbon dosage due to technical limitations of the equipment. The flow suitable for methanol was therefore used for all carbon sources. However, ensuring the exact same COD content between carbon sources were not considered a priority, given that the focus was on studying the transitions of carbon sources. Dosing flows were measured during the experiment at different occasions and showed varying flows and were generally lower than intended. Fine-tuning was difficult due to both equipment limitations and the inherently small 0 10 20 30 40 50 60 70 80 0 7 14 21 28 35 42 49 56 63 70 77 84 [m g /l ] Days Suspended soilds GA1 GE1 GE2 Switch to ethanol Switch to methanol Increased C 34 volumes involved. Calculations involving carbon input (e.g., C/N ratios, COD added) which are dependent on the carbon dosage flows were uncertain as a result. To avoid altering too many parameters mid-experiment, increasing the methanol concentration was chosen as a practical approach to investigate its effect, particularly on nitrite accumulation, which appeared to decrease with higher carbon availability. While further dilution of methanol to allow higher flows might have been ideal, to also have more continuous dosing, limitations in the pump and tubing setup made this infeasible. Higher flows caused displacement of the dosing tubes during testing prior to the experiments. There were issues with the influent as well, as the pumps stopped providing flow, particularly the one connected to reactor (GE1). Flow inconsistencies were particularly evident in the morning. Several attempts were made to fix the pump or tubing, but issues remained throughout the experiment. Other operational issues include stirring inefficiencies and technical constraints of the reactors. The reactors are quite small, and not originally designed to house biocarriers. The internal pipes and such hindered free movement of the carriers. Stuck biocarriers caused stops in stirring, but there were also some technical issues with the stirrers. It possibly contributed to stagnant zones and low flows. As mentioned in the theoretical background, stagnant zones can appear in MBBR systems but the knowledge on whether this affect the overall nitrogen removal performance is uncertain (di Biase, 2019). However, it is possible that stagnant zones in this study could have affected the process due to the small size of the reactors. Additionally, biomass seemed to build up over time in all the reactors, which eventually caused issues with blockages in the outlets, requiring increased maintenance. The results showed variability during and between days across the reactors. It may be partly due to natural variations but is also likely tied to operational fluctuations and measurement uncertainties. Sudden low or high concentrations of nitrogen species right before or after a switch are considered less representative of the impacts of carbon transitions. This is because the changes were made after the weekends (without reactor maintenance). Concerning measurement frequencies, nitrogen was the most frequently analysed parameter and was the only one measured twice daily during transition phases. In contrast, DOC was still only measured once per day during these periods, and overall, it was analysed less frequently than nitrogen. This limited the ability to capture fluctuations in DOC concentrations. This along with the limited number of samples for COD, only once weekly, may not provide sufficient data to accurately assess carbon dynamics. The resolution for suspended solids was also low, with measurements conducted only on a weekly basis, making it difficult to draw reliable conclusions from the data. Furthermore, the study relied solely on grab sampling, which even for parameters measured twice daily may not fully reflect the system's variability. The frequency of the analyses was constrained by the scope of this thesis. Certain NO2 - measurement values were outside of the measuring range of the measuring kit used for the spectrophotometry instrument. For some samples, the 35 spectrophotometer showed an approximate concentration outside of the measuring range, while for other samples the absorbance was too high for it to provide any value at all. The samples with too high absorbance were diluted and run in the instrument again. The samples outside of the measuring range, but still given an approximate value of the concentration, were too many to be able to dilute and measure again within the time scope of the study. Therefore, a set of representative samples were diluted and measured again and compared with the approximative values. In this aspect, the approximate values were deemed to be sufficiently accurate and were consequently used as data points in the laboratory results. Nonetheless, these nitrate concentration values have only been used for general discussions on concentration fluctuations over time. For more detailed analysis and use of specific values, all samples could be diluted and re-measured. 4.1.9 Relevance to full-scale systems and previous research The results indicated an effect when switching carbon source from ethanol to methanol when going back to methanol. Several studies have observed that methanol generally requires a longer acclimatization period compared to ethanol (Bill et al., 2009; Christensson et al., 1994; Gavazza Dos Santos et al., 2004; Zhang et al., 2024). This means how long it takes for the bacteria to get used to the new substance. The laboratory experiments showed a noticeable drop in denitrification capacity during the first few days of transitioning from ethanol back to methanol. Only two studies were found that explored the ability of switching between methanol and ethanol. Cherchi et al. (2009) tested various acclimated and non-acclimated biomass with several carbon sources in laboratory scale Sequencing Batch Reactors (SBR). Among them are methanol, ethanol, and a commercial alcohol mix that contains methanol. It was suggested that both methanol and ethanol could be used right away in the biomass acclimated to the alcohol mixture. Ethanol could be used directly in the methanol acclimated sludge. Methanol was however not tested on ethanol acclimated sludge. Mokhayeri et al (2008) found that ethanol could be utilized on methanol- acclimated biomass with a similar denitrification rate to that of methanol on methanol- acclimated biomass. Using methanol on ethanol-acclimated biomass resulted in lower denitrification rates but still worked and showed quite fast acclimatisation. They suggest that the two substrates could be interchanged to grow respective populations with a short acclimatisation period. While these experiments used another technique, their results also point to possibilities with carbon source interchangeability, at least in the short-term response. This, together with the results in the current thesis, may mean that the acclimatisation period to methanol is not as long as expected. Assuming that the acclimatization time is a few days, implementing a switch of carbon source in the large-scale process might be deemed acceptable to do once. If the plan instead is to be able to switch back and forth between different carbon sources, it could potentially affect the yearly mean value too much. The results are considered highly relevant to the full scale. Although denitrification performance is somewhat lower in the lab, the results come quite close when considering differences in size and setup between the lab-scale and full-scale systems. 36 4.1.10 Contributions to Research and Future Directions Previous experiments at Gryaab or the reviewed literature have primarily employed batch testing using denitrification capacity tests, with few efforts directed at continuous-flow MBBR systems operating with real wastewater. While both methanol and ethanol are well-established carbon sources for denitrification, comparative studies rarely focus solely on transitions between the two. Instead, the research tends to evaluate alternative or bio-based carbon sources or combining conventional sources with emerging ones. The dynamics of switching between methanol and ethanol, especially in continuous systems, remain underexplored. This study therefore contributes to the growing body of knowledge on post-denitrification in continuous biofilm systems by addressing an area that has seen limited experimental focus. Despite the practical limitations of the current experimental setup, this study provides valuable insights. Several key areas are proposed for further study. Future experiments should prioritize securing a stable and representative setup, with continuous and accurately controlled inflow and carbon dosing. This includes using appropriate tubing and pumps capable of handling the correct flow rates, and ensuring consistent carbon availability (i.e., excess COD) to evaluate performance under ideal conditions. More frequent or even continuous (online) monitoring is recommended to better capture short-term variations, transition dynamics, and potential lag effects. Grab sampling in its current form may miss critical fluctuations, particularly around carbon source switches. Additionally, incorporating gas monitoring (e.g., N₂, N₂O) would allow for a more detailed understanding of denitrification performance and pathways, particularly with regard to environmental impacts. Future studies should incorporate statistical analysis to increase confidence in the results. Analysing the microbial communities could provide deeper insights into their response to changes in carbon source. This could help determine whether bacteria adapt over time, how communities shift during repeated or short-interval transitions, and how such shifts correlate with performance. Lastly, future work should use different methanol-to-ethanol ratios in the mixture and observe whether performance stabilizes after repeated switches with all options. 4.2 Multicriteria analysis The following sections present the results and discussion regarding the MCA. First an overview of what information was available regarding product alternatives is presented. Afterwards the sub-chapters cover the main steps of selecting and defining alternatives and criteria and developing the evaluation basis. Given the iterative nature of the MCA, these sections contain methodology details as well. The final section summarizes key findings of the evaluation basis and discusses the relevance of each criterion and alternative. It includes guidance for WWTPs on interpreting the results and determining the next steps for further evaluation. 37 4.2.1 Available data from suppliers Appendix C summarizes what information has been requested and provided in the previous market analysis or in this work. Overall, in the previous assessment, sufficient product data were available to categorize the carbon sources into different alternatives and to support calculations and evaluations in several criteria. For the current assessment, updated or additional information was requested from suppliers. This included details on transport (location and means), types and quantities of raw materials, associated production processes (e.g., feedstocks, methods, and energy consumption), full product specifications, safety data sheets, pricing, and deliverable volumes. Comments are also included in the appendix, regarding the data availability and usage of the information in the MCA. Further discussion about data availability is held in Section 4.2.5. 4.2.2 Alternatives and prerequisites The alternatives that have been chosen for evaluation in this thesis are seen in Table 4.3. Table 4.3. The chosen alternatives, including description of type of product and how many products and suppliers are involved. Alternative Description 1. Fossil-based methanol Fossil methanol, supplied by the current provider and represents continued operation as at the present Rya WWTP. 2. First-generation bio-ethanol Bio-ethanol derived from cereal-based feedstocks, primarily wheat. It includes two products from two different suppliers. 3. Second-generation bio-ethanol Bio-ethanol derived from forestry industry residue. Includes two products from two different suppliers. 4. Recycled ethanol Ethanol used in other industries, such as pharmaceutical industry, prior to recycling. It includes two products from the same supplier. 5. Mass-balanced bio-methanol Includes two mass-balance certified products from two different suppliers. 6. Second-generation bio-methanol Bio-methanol derived from forestry industry residue. It concerns only one product from a single supplier. 7. Methanol and ethanol in separate lines It involves using three dosing lines with first-generation bio-ethanol (as in Alternative 2) and three lines with fossil methanol (as in Alternative 1). Some product characteristics are presented in Appendix F. All alternatives, except the recycled ethanol has purities above 99%. Besides varying concentration, this alternative contains impurities. One of the included products contains 1–30% isopropanol (propan- 2-ol), depending on batch, while the other contains 3.1% acetone. One of the included products in Alternative 3 potentially contains a small amount of metal concentrations. The products originate from seven different companies, including both Swedish and international actors. These represent a range of sectors such as chemical industries, 38 agricultural and forestry cooperatives, biorefineries, and companies specializing in solvent recycling and processing. All alternatives are initially regarded as technically and economically feasible for application at Rya WWTP because the assessment focuses solely on variants of methanol and ethanol, which in general are well-established carbon sources. Alternative 7 does not include any additional products to those evaluated in the other alternatives. While various product combinations could be theoretically assessed in this alternative, only fossil-based methanol and biobased ethanol was chosen. This option is included rather to represent design options for using ethanol and methanol simultaneously. This alternative could also be applied for representing a transition period from fossil-based methanol to biobased ethanol. The COD demand is assumed to be split equally between the two carbon sources, unlike the other alternatives in which the sole carbon source should make up the whole COD demand. The assumptions underlying the evaluation are the following: • All alternatives are assumed capable of achieving the same annual average treatment efficiency after a sufficient acclimatization period, although the required quantity of carbon source may vary. • Fossil-free methanol and fossil-free ethanol with high purity (>99%) are assumed functionally equivalent to their fossil-based counterparts. • The alternatives are evaluated as if they were already implemented at the treatment plant. The first assumption is necessary to enable a fair, structured comparison between the options within an MCA framework. This approach is consistent with similar MCA studies in the wastewater sector, where certain performance variables are held constant to avoid introducing case-specific complexity or undermining comparability. In practice there are expected differences in denitrification effectiveness among the alternatives, but here the focus is instead of the amounts required to achieve the same level of nitrogen removal. With the second assumption, the recycled ethanol (Alternative 4) with lower purity, requires assumptions about the technical function. The COD-content is comparable to methanol (see Appendix F). Assuming the COD is easily available, this provides some indication of functionality of this alternative. This underlines the choice of including it in the analysis. However, it is further assessed in the evaluation basis, including other relevant assumptions. The third assumption is needed to make simplifications. This thesis does not address detailed infrastructure-specific adjustments which are dependent on further investigation and expert review and thus fall outside the current scope. However, the implementation for each alternative were conceptualized, see Figure 4.14. Their application is assumed to take place in the existing post-denitrification facility using current carriers. In Alternative 1–6 a single carbon source is added to all treatment lines using existing dosing pumps. Both storage tanks are used for the same carbon source. 39 Alternative 7 likely requires facility adaptation to handle both carbon sources independently, such as: • Separate delivery logistics and transport handling. • Separate storage tanks for each carbon source. • Potentially additional pumps and structural support for unloading. • New or modified physical infrastructure (valves, pipes etc.) or operational schemes to manage separate dosing system from the tanks to the dosing pumps. Furthermore, if two tanks are no longer jointly used (as in dual carbon source scenarios), the existing booster pump may no longer be necessary. Unused pipes or other equipment under such configurations may also require decommissioning. While the implementation is not part of the analysis, it needs to be mentioned that for any alternative, the facility must be compatible with the selected carbon source. As Rya WWTP is already configured for methanol, modifications (e.g., physical adjustments, inspections and permits) would be necessary to accommodate ethanol. During consultations on how to apply the alternatives it was stressed that construction works require shutting down the facility. The more complicated reconstruction, the longer the shutdown, which could significantly affect treatment results. Figure 4.14. Conceptual overview of how the alternatives can be implemented at the WWTP, containing the six post-denitrification lines, along with pumps (P) for dosing, pumping from the truck and a booster pump in the middle. The pipe systems led between the lines and the two storage tanks are shown. On the left, each carbon source is used in the entire system. On the right, methanol and ethanol are used separately. Green indicates new changes and dashed lines the structures that potentially becomes unused in this alternative. Other alternatives that represent the design or transition aspects, like Alternative 7, were initially considered. Further discussions on this along with other choices for alternatives is held in Section 4.2.5. 40 4.2.3 Chosen criteria The general criteria that can be relevant when studying carbon sources was studied and summarized, see Appendix B. From this, the chosen criteria and the definitions included in this assessment are presented in Table 4.4. The explanations of included or excluded criteria follows bellow the table. Table 4.4. Chosen criteria for the multi-criteria analysis. Criteria Definition Ecological dimension Impact on recipient Impact on recipient through changes of BOD- and copper concentrations in effluent. Climate impact Total g