Liquid-Liquid membrane contactors for sustainable ammonia recovery and valorization: experimental insights, novel approaches and applications

PhD program in Chemical Process Engineering Liquid-Liquid Membrane Contactors for Sustainable Ammonia recovery and Valorization: Experimental Insights, novel approaches and Applications. Doctoral thesis by: Miguel Aguilar i Moreno Thesis Advisors: Jose Luís Cortina Pallás César Alberto Valderrama 2 KEYWORDS Ammonia recovery; Wastewater treatment; Membrane contactors; Coagulationflocculation; Aeration; Aluminum sulphate; Liquid-liquid membrane contactors (LLMC); Sustainable technology; Hollow fiber membrane contactors; Circular economy AKNOWLEDGEMENTS This study was conducted at Universitat Politècnica de Catalunya (UPC), Barcelona. The research received support from the R2MIT project (CTM2017-85346-R) funded by the Spanish Ministry of Economy and Competitiveness (MINECO) and the Govern de Catalunya (ref. 2017-SGR-312). I would like to express my sincere gratitude to all the institutions involved for providing me with the opportunity to carry out this doctoral thesis. Their support has been essential in the development of this research work. I would like to express my sincere gratitude for the continuous support and dedication I have received over these four years from Elena Guillén Burrieza, Mònica Reig, and my thesis advisors, José Luis Cortina and Cesar Valderrama. Their commitment, guidance, and expertise have been fundamental pillars in my academic journey. I am deeply thankful for the opportunity to benefit from their mentorship, which has been invaluable to the development and completion of this research project. Furthermore, I want to extend my gratitude to my fellow doctoral collegues who shared this academic journey with me. Their ideas, discussions, and camaraderie have enriched my experience and created an environment conducive to mutual learning. En últim lloc, voldria agrair a la meva família tot l'esforç i dedicació que m'han lliurat durant aquests anys, als meus amics de tota la vida que m'han fet costat i acompanyat quan ho necessitava, i sobretot agraïr i possar de manifest la paciencia i ajuda de la Gisela que m'ha donat tot el suport tant a les dures com a les madures, t’estimo, a tots vosaltres moltes gràcies. 3 ABSTRACT This comprehensive research wants to represent a significant stride in the exploration of innovative strategies aimed to enhance ammonia recovery within diverse wastewater streams. The study is structured into distinct phases, each addressing crucial aspects of the ammonia recovery process. In the initial phase, the research focuses on augmenting membrane contactor performance, employing coagulation-flocculation (C/F) and aeration as preliminary treatments. The outcomes of this phase demonstrate substantial increases in both the mass transfer coefficient and overall efficiency of ammonia recovery, increasing of 7.80x10-7 to 1.04 x 10-5 m·s-1 and from 8 to 67%, respectively, particularly notable when treating the sidestream centrate. A pivotal finding underscores the efficacy of dosing aluminum sulfate (Al2(SO4)3) at 30 mg Al+·L-1 in the C/F process, yielding remarkable efficiencies (>50%) in the removal of chemical oxygen demand (COD), turbidity, and total suspended solids (TSS). Into the second phase, the study delves into the application of liquid-liquid membrane contactors (LLMC) for ammonia recovery. An array of experimental conditions is meticulously explored, with the results illuminating the considerable impact of replacing the acid washing liquid between steps on the overall performance of the LLMC. Additionally, the study highlights the nuanced relationship between the initial ammonia concentration and the subsequent recovery, providing valuable insights. This phase effectively showcases the potential versatility and efficiency of LLMCs in the valorization of ammonia within wastewater streams. The third and final phase introduces a novel asymmetric hollow fiber liquid-liquid membrane contactor (HF-LLMC) with distinctive selectivity for ammonia over water 𝑃𝑁𝐻3 = 87 - 180 L m-2 h-1 bar-1 and 𝑃𝑤 = 1.2 -1.4·10-3 L m-2 h-1 bar-1 at NTP conditions. The investigation entails a comprehensive examination of various operational parameters, including feed and acid flow rates, mass transfer coefficients, and acid consumption. Notably, the results affirm the high selectivity of the HF-LLMC for ammonia, coupled with minimal water transfer. This establishes the HF-LLMC as a promising technology for the recovery and concentration of ammonium in diluted urban and industrial streams. The combination of these findings, considered from a global perspective, significantly contributes not only to the advancement of sustainable nutrient recovery technologies but 4 also underscores their pragmatic feasibility for implementation within the context of the circular economy and efficient resource management. 5 LIST OF CONTENTS AKNOWLEDGEMENTS .............................................................................................. 2 ABSTRACT ..................................................................................................................... 3 GLOSSARY ................................................................................................................... 14 CHAPTER 1 .................................................................................................................. 16 1.1. The Global Impact of Nitrogen Fertilizers: Haber-Bosch process ......... 16 1.2. From nitrogen removal to nitrogen recovery in wastewater effluents . 20 1.2.1. Conventional Nitrogen removal in WWTP’s .................................... 20 1.2.2. Nutrients, Energy and Resources: From WWTPs to W&RRFs ............... 23 1.2.3. Alternative ammonia recovery techniques from wastewater ......... 25 1.2.4. A chemical process for nitrogen recovery: recovery ammonium as struvite …………………………………………………………………………...26 1.2.5. Physico-chemical processes for nitrogen recovery ........................... 27 1.2.6. Biological process for nitrogen recovery ............................................ 35 1.3. Research challenges and objectives ............................................................ 35 1.4. References ...................................................................................................... 37 CHAPTER 2 .................................................................................................................. 52 2.1. Objectives ........................................................................................................... 52 2.2. Thesis overview ................................................................................................. 53 CHAPTER 3 .................................................................................................................. 55 3.1. Abstract .............................................................................................................. 55 3.2. Introduction ....................................................................................................... 55 3.3. Materials and Methods .................................................................................... 58 3.3.1. Chemical reagent and wastewater source ............................................. 58 3.3.2. Experimental design .................................................................................. 59 3.3.3. Experimental set-up .................................................................................. 59 3.3.4. Analytical methods.................................................................................... 65 3.3.5. Economic analysis...................................................................................... 65 6 3.4. Results and discussion ..................................................................................... 70 3.4.1. Coagulant and dosage selection for the C/F process ............................ 70 3.4.2. Optimisation of the operating conditions for the C/F process ............ 72 3.4.3. Flocculation stage ...................................................................................... 74 3.4.4. Aeration Stage ............................................................................................ 75 3.4.5. Flat-sheet membrane contactor stage ..................................................... 76 3.4.6. Economic Analysis .................................................................................... 79 3.5. Conclusions ........................................................................................................ 83 3.6. References .......................................................................................................... 85 CHAPTER 4 .................................................................................................................. 93 4.1. Abstract .............................................................................................................. 93 4.2. Introduction ....................................................................................................... 93 4.3. Materials and Methods .................................................................................... 96 4.3.1. Reagents ...................................................................................................... 96 4.3.2. Wastewater Solution ................................................................................. 96 4.3.3. Experimental Set-Up ................................................................................. 96 4.3.4. Experimental Design ................................................................................. 98 4.3.5. Data Analysis ............................................................................................. 99 4.4. Results and Discussion ................................................................................... 100 4.4.1. Effect of Changing Feed or Acid Stripping Solution between LLMC Process Steps to Increase Ammonia Recovery .............................................. 101 4.4.2. Initial Ammonia Concentration Effect on the Overall LLMC Performance ........................................................................................................ 105 4.4.3. Study of the Effect of Initial Feed Volume on the LLMC Trials ....... 108 4.4.4. Wastewater Temperature Effect on the LLMC Process ..................... 109 4.5. Conclusions ...................................................................................................... 111 4.6. References ........................................................................................................ 113 CHAPTER 5 ................................................................................................................ 116 5.1. Abstract ............................................................................................................ 116 7 5.2. Introduction ..................................................................................................... 117 5.3. Materials and methods ................................................................................... 121 5.3.1. Chemicals and analytical techniques .................................................... 121 5.3.2. Membrane module .................................................................................. 122 5.3.3. Experimental set-up ................................................................................ 123 5.3.4. Experimental design ................................................................................ 124 5.3.5. Mass transfer modelling in LLMC operations .................................... 125 5.3.5. Ammonia transport ................................................................................. 130 5.3.6. Experimental determination of the overall ammonia mass transfer coefficient ............................................................................................................ 131 5.3.7. Water transport ........................................................................................ 131 5.3.8. Permeate side chemical equilibrium ..................................................... 132 5.4. Results and discussion ................................................................................... 132 5.4.1. Ammonia recovery and CF .................................................................... 133 5.4.2. Acid consumption in the permeate side............................................... 136 5.4.3. Ammonia and water transport .............................................................. 138 5.4.4. Asymmetrical membrane structure related effects ............................. 142 5.4.5. Mass transfer coefficient and resistance regime .................................. 144 5.5. Conclusions ...................................................................................................... 148 5.6. References ........................................................................................................ 150 CHAPTER 6 ................................................................................................................ 158 8 List of Figures CHAPTER 1 Figure 1. The relationship between ammonia and agriculture, energy and environmental emissions. Solid lines show existing relationships. Dashed lines show the possible future relationships. Only environmentally significant emissions are shown (Razon, 2018). Figure 2. Schematic of typical wastewater treatment plant configuration with anaerobic digestion (AD) stage. Figure 3. Figure 3. (a) Publications of “ammonia recovery technologies” [Scopus 2020] (b) Percentage of disciplines in total publications [Scopus 2020]. Figure 4. Publications of “Membrane contactors” and “ammonia”. [Scopus 2020] Figure 5. (a) Schematic representation of ammonia recovery by means gas permeable membrane (W. Lee et al., 2021). (b) Schematic description of ammonia transport through the hydrophobic membrane (E. E. Licon Bernal et al., 2016). Figure 6. Schematic reproduction of Hollow Fiber Membrane Contactor (3M Liqui-cell) Figure 7. HFMC “open loop” set up (a) and HFMC “closed loop” set up (b) (Darestani et al., 2017; Vecino et al., 2019) CHAPTER 2 Figure 1. Graphical abstract of the project and how the chapters are interconnected in reference to the objectives. CHAPTER 3 Figure 1. General scheme of the different anaerobic side-stream treatment stages used in the present study. Figure 2. Schematic representation of the nitrogen recovery scheme. Figure 3. Theoretical TSS, turbidity and COD removal values for different mixing times and mixing speeds, at a fixed settling time of 30 min (graphics obtained from the Design Expert 11 software). 9 Figure 4. Removal of COD (%) and turbidity (%) from anaerobic centrate after Fe3O4(s)/SiO2(s) addition. Figure 5. Variation of pH and the efficiency of HCO3- removal with time in the aeration stage. Figure 6. Membrane contactor results during operation: (A) TAN concentration evolution in the feed tank for pre-treated and untreated centrate, (B) TAN recovery and pH variation and (C) TAN concentration evolution and concentration factor in the acid tank. Figure 7. Gross cost, revenues and net present value for the nitrogen recovery scenario under study. Figure 8. Gross cost contribution of the nitrogen recovery scenario under study for: (A) the different processes and (B) for the different capital and operating costs. Figure 9. Sensitivity analysis for a ±30% variation of the main economic parameters. Figure 10. Sensitivity analysis for the NH4NO3 prices and mass transfer coefficient (Km). CHAPTER 4 Figure 1. Ammonia concentration evolution over time in the feed tank when changing (a) the acid and (b) feed between steps (up). Nitrogen concentration achieved in the liquid fertilizer by changing the (c) acid or (d) feed solution between steps (down). Orange color implies one stage of LLMC (triangle for the feed side and circle referring to the fertilizer solution), while yellow color refers to experiments with two LLMC stages (triangle for the feed side and circle referring to the fertilizer solution). Figure 2. Ammonia recovery after each step, and the global results changing the acid or the feed solution between steps. Figure 3. Comparison between working with the sidestream (high NH3 concentration) and mainstream wastewater (low NH3 concentration): (a) ammonia concentration evolution in the feed tank, (b) ammonia recovery, (c) concentration factor and (d) %N- NH4 concentration in the liquid fertilizer. High ammonia concentration is indicated by the color orange, while the color yellow implies working at low ammonia concentrations. 16 CHAPTER 1 Introduction 1.1. The Global Impact of Nitrogen Fertilizers: Haber-Bosch process One of the most important challenges that developed societies face nowadays is to transform the economy to a efficient economic model with more restricted and selective access to the necessary natural resources aiming to avoid resource shortage and other related issues, and at the same time increasing competitiveness without generating externalities in the supply chain. In this context moving from linear thinking (cradle to grave) to a circular model by closing the material loops is a primary objective (Suzanne et al., 2020). Throughout urban and industrial cycles, waste process is considered potentially as secondary resources, especially when they contain elements included in the list of critical raw materials (European Commission, 2023). In the urban water cycle, wastewater has been identified as having great potential through different components: i) carbon present in organic matter in wastewater streams to transform wastewater treatment plants (WWTP) in energy self-sufficient (energy positive WWTP’s), the carbon present in these wastewater streams serves as a food source for the microorganisms found in wastewater treatment plants. Through biological processes, these microorganisms break down the carbon-rich organic matter, producing biogas as a byproduct, ii) nitrogen and phosphorus for fertilizers production and/or direct soil applications, and iii) water to be recover using reuse schemes for industrial, agricultural, and environmental applications. The role of nutrients, such as P or N, significantly influences the intensive farming industry, where fertilizers are extensively used, leading to a substantial demand for this type of fertilizer (Robles et al., 2020). Fertilizers are categorized into two groups: singlenutrient fertilizers and multi-nutrient fertilizers. The former consists of one essential nutrient, while the latter combines two or more essential nutrients (K, P, and N) (Vecino et al., 2019). Specifically, focusing on Nitrogen (N), the global reactive nitrogen cycle has doubled over the last century due largely to population growth. Ammonia is the second most produced chemical in the world (Beckinghausen et al., 2020; Lee et al., 2021; Razon, 2018). 17 Ammonia production plants, typically utilizing the Haber-Bosch process (Eq. 1) for the synthesis of industrial ammonia, play a pivotal role in fertilizer generation (Kirova- Yordanova, 2004). This method incurs an energy consumption of approximately 35-50 MJ/kg N. Producing one ton of N-fertilizer requires almost 1000 m3 (NTP) of natural gas. Due to the high resource and energy costs, there is a growing exploration of alternative sources for ammonia (Beckinghausen et al., 2020). Globally, approximately 85% of ammonia production is allocated to fertilizers, with the remaining 15% used in various industrial applications, such as plastics and fibers, underscoring agriculture's predominant role in ammonia usage (Vecino et al., 2019). 𝑁2(𝑔)+3𝐻2(𝑔) →2𝑁𝐻3(𝑔) (1) In general terms, more than 100 million tons of fertilizer are generated throughout the globe from the Haber-Bosch process, out of a total of approximately 160 million tons of ammonia. Asia is the region where the largest ammount of ammonia is produced by the Haber-Bosch method, accounting for about 50% (80 Mt) of total production (González Montiel, 2008), which entails the consumption of 1-2% of all the worldwide energy annually. It is estimated that 50% of the natural gas used in the industry corresponds to the production of fertilizer with the Haber-Bosch method (Smith et al., 2020). In addition, the amount of CO2(g) emitted when when natural gas is used as the primary source for NH3 generation, is estimated to be 1.6 t of CO2 /t NH3, reaching 3.2t CO2(g)/t NH3 if the raw material is coal (Osorio-Tejada et al., 2022; Yüzbasıoglu et al., 2021). The direct emissions from ammonia production currently amount to 450 million tons of CO₂ (IEA 2021) about 1.8% of global carbon dioxide emissions (Bird et al., 2020). Additionally, the market predictions for fertilizers indicate a significant growth, with a Compound Annual Growth Rate (CAGR) of approximately 9.1% from 2023 to 2030 (Company et al., 2023). During this period, it is expected that NPK fertilizers will lead the market, occupying the largest share. A breakdown of the nitrogen fertilizer supply, demand and balance worldwide is listed in Table 1. (FAO, 2019). In addition, in Table 2 is collected the nitrogen consumption by region (Yara International, 2022). 18 Table 1. World nitrogen supply (in milions of tonnes), demand and balance 2016-2022 (FAO, 2019). 2016 2017 2018 2019 2020 2021 2022 World Ammonia-capacity 180 184 187 189 187 189 190 Ammonia-supply capability 153 155 157 161 160 161 163 Nitrogen-other uses 36 37 38 39 39 40 40 Nitrogen-avalible for fertilizers 116 117 119 120 119 121 122 Nitrogen-fertilizer demand 105 105 180 107 108 110 111 Nitrogen-potential balance 11 12 13 15 12 11 11 Table 2. Nitrogen use by region (% en each region) (Yara International, 2022). Fertilizer % of nitrogen-based fertilizers used USA Brazil West/Central Europe India China Urea 24 57 20 79 34 Ammonia 27 - - - - Nitrates 2 13 41 - - UAN 26 - 12 - - NPK 6 1 12 3 54 DAP/MAP 6 13 4 11 7 Other 12 4 11 1 6 As - 12 - - - ABC - - - - - Total year 2020 (milion tonnes) 12.1 5.3 11.1 23.7 20.4 UAN: Urea Ammonium Nitrate, NPK: Nitrogen, Phosphorus, Potassium fertilizer, DAP/MAP: Di-Ammonium Phosphate/MonoAmmonium Phosphate, AS: Ammonium Sulfate, ABC: Ammonium BiCarbonate. The most widespread fertilizer used is urea, but the other varieties of fertilizer vary according to the requirements of each region as it is clearly shown in Table 2. This emphasizes the importance of considering the target market of the product when defining recovery strategies. According to Beckinghausen et al. 2020, a potential solution would be the most appropriate as it would encourage a relationship between farmers and the WWTPs if N and P recovery options need to be implemented (Beckinghausen et al., 2020). 19 A critical review by Razon (Razon, 2018) delves into the reactive nitrogen cycle, encompassing its generation, emission into the environment, and its integral connection to food security, environmental degradation, climate change, and alternative energy, as illustrated in Figure 1. Figure 1. The relationship of ammonia to agriculture, energy and environmental emissions. Solid lines show existing relationships. Dashed lines show the possible future relationships. Only environmentally significant emissions are shown (Razon, 2018). Figure 1 shows that the global food supply depends on anthropogenic nitrogen fertilizer and that its production and overuse has contributed to the deterioration of the environment through the different options identified in the N-cycle. It is being promoted from different regulatory bodies that biofuel production and carbon capture schemes may also increase demand for reactive nitrogen. In addition, ammonia is being proposed as an alternative fuel and new technologies to replace the Haber–Bosch process must be developed. (Kehrein et al., 2020). Although the exact magnitude of future requirements is uncertain, a large demand for reactive nitrogen may be inevitable and recovery from waste must be pursued. This 20 recovery of reactive nitrogen from waste streams is imperfect because natural processes tend to return reactive nitrogen to the more stable state N2(g) and then in practice is lost. In this perspective, a change of paradigm on the waste water treatment cycles should be implemented: from N removal as N2(g) to N recovery. (Galloway et al., 2008; Geissdoerfer et al., 2017). 1.2. From nitrogen removal to nitrogen recovery in wastewater effluents 1.2.1. Conventional Nitrogen removal in WWTP’s Wastewater streams serve as significant sources of N, with raw urban wastewater containing nearly 60 to 80% of total nitrogen in the form of ammoniacal nitrogen. In industrial wastewater, the NH3 content varies depending on the industry and specific processes, ranging approximately from 0.005 to 5 g·L-1. Urban wastewater streams, on the other hand, typically exhibits NH3 concentrations around 0.01 to 1.20 g·L-1 depending on the WWTP configuration (Sheikh et al., 2023). It is noteworthy that WWTPs are designed to mitigate NH3 levels in waterways, primarily due to stringent total ammonia nitrogen content (TAN) discharge limits imposed to the discharge to environmental compartments. The main-stream and side-stream components have been identified as the major sources contributing to elevated NH3 concentrations in WWTPs. (Azreen et al., 2017; Fowler et al., 2013; Salomon et al., 2016). A prevalent technic employed for the mitigation of nitrogen in urban WWTP, where it predominantly exists in the form of ammoniacal nitrogen, in both inorganic and organic forms, is the Biological Nitrogen Removal (BNR) process (Ren et al., 2020). This involves the application of nitrification/denitrification techniques, including autotrophic oxygen-limited nitrification-denitrification, aerobic nitrification-denitrification using Bacillus, and decontamination techniques such as Anaerobic Ammonium Oxidation (anammox), Completely Autotrophic Nitrogen Removal Over Nitrite (CANON), as well as combinations like anammox and denitrification or bacterial consortiums involving algae and nitrification. Nitrification involves the conversion of ammonium to nitrite and then to nitrate, while denitrification is the reduction of nitrate to gaseous nitrogen. The autotrophic process 21 refers to the ability of certain microorganisms to obtain energy from inorganic sources, such as ammonium or nitrite. Anaerobic Ammonium Oxidation (anammox) is a process where specific bacteria oxidize ammonium directly with nitrite, without the need for oxygen. CANON involves the autotrophic conversion of ammonium to nitrate over nitrite. (Mukarunyana et al., 2018; Thakur & Medhi, 2019; Yue et al., 2023). A simplified urban WWTP is depicted in Figure 2. In the primary settling stage, which is part of the primary treatment in a WWTP, sedimentable solids from the mainstream are reduced. Subsequently, the mainstream moves to the biological treatment, where aerobic nitrogen removal takes place, following the stages previously described depending on the plant configuration. The most conventional process of carbon removal using conventional areared sludge (CAS) involves constant aeration to supply the necessary oxygen. During the biological stage, activated sludge (microbial biomass) is generated and then separated in the secondary settling. Advanced WWTPs where more restringing discharges of N are requested, biological removal processes need to be included. In this phase, both the sludge from the primary settling and the secondary settling are treated in the absence of oxygen to produce biogas. The resulting liquid from this process, known as digestate, and the reject water from digestate is the remaining liquid phase after dehydrating anaerobically digested sludge, either through centrifugation or dewatering, are key aspects in the management of wastewater treatment. (Faragò et al., 2021; Fernández, 2008; Ma et al., 2020). If the WWTP does not have an AD stage, the centrate is the water from the solid-liquid separation. Thirty percent of the nitrogen load in the treatments can be provided by centrate water. (Beckinghausen et al., 2020; C. H. Guo et al., 2010). The streams resulting from these stages are known as side-streams to differentiate them from the mainstream. The properties of these types of streams can vary dramatically depending on the type of sludge handled and the methods employed for pretreatment, digestion, and dewatering. Due to this complexity, categorizing the side-stream proves challenging (Aguilar-Moreno et al., 2022). 22 Figure 2. Schematic of typical wastewater treatment plant configuration with anaerobic digestion (AD) stage. One of the greatest controversy and challenge of BNR processes is the high energy and resource demand to transform, for example, ammonia nitrogen into a low-valuable product such as N2(g) (Ren et al., 2020). The aeration stage represents 50% of the usual energy consumption in WWTP wastewater treatment (Nowak, 2003). An extra 4% of electricity is used in the WWTP for the removal of nutrients in wastewater, i.e. 45 MJ/kg N, in addition, this leads to an emission of 0.9 kg CO2/m3 to the atmosphere, (Xie et al., 2016). Nowak 2003, quantified that the specific energy demand for aeration to remove nitrogen is 0.5 kw·h·kg-1 O2 consumed, which is equivalent to 40% of the O2(g) consumption load for carbon disposal. (Nowak, 2003). If the endeavours to recover NH3 in WWTPs were strategically oriented towards a circular ecnomy, it could potentially yield significant energetic and economic benefits. For instance, if we take as case study the WWTP located at El Prat de Llobregat (Barcelona, Spain) one of the most modern and largest wastewater treatment plants in Europe (all the data in Table 3.), which treats 420.000 m3/d, equivalent to 36% of all wastewater generated in the Barcelona metropolitan area, is a localized source of nitrogen. 23 Table 3. Metrics for Ammonia Recovery Potential at WWTP El Prat de Llobregat. WWTP El Prat de Llobregat Value Volume treated by the WWTP 420,000 m³·d-1 Equivalent population served by the WWTP 1.26 million inhabitants Nitrogen emission from the WWTP 2,500 ton N·year-1 Nitrogen production per inhabitant 5.5 g N·day-1 Price of one ton of ammonia €180·ton-1 Total value of ammonia produced €450,000·year-1 The WWTP of El Prat de Llobregat has an average input of 1.26 million equivalent inhabitants, with an emission of nitrogen into the environment of 2,500 t N/y (Àrea Metropolitana de Barcelona, 2002; Rufí-Salís et al., 2020). This implies that each inhabitant produces 5.5 g/N/d. According to data from the EC ((European Commission, 2019), the price of a ton of ammonia fluctuates around 180 €, which implies a value of 450,000 €/y considering only the WWTP of El Prat de Llobregat (Table 3). 1.2.2. Nutrients, Energy and Resources: From WWTPs to W&RRFs The recovery of nutrients not only prevents their discharge into the environment but also facilitates their safe reintroduction into environmental compartments. This represents a significant step towards a circular economy (Robles et al., 2020). As a result, ammonium and phosphate are no longer viewed solely as pollutants to be eliminated but also as sustainable resources to be recovered (Darestani et al., 2017). Water sources rich in ammonium can be used as a food source for both animals and humans, it can also be used for algae growth and/or bacteria that are used for the generation of biogas or biofuels or directly as fertilizers (Matassa et al., 2015; Thorin et al., 2018; WIDANARNI et al., 2012; Wuang et al., 2016). In the context of fertilizers, micro-organisms growing in wastewater cannot be directly applied due to the high concentration of metals and other pollutants that can appear in the biomass along with nitrogen. It is necessary to refine the biomass to guarantee a safe product. As an 24 alternative to this, direct application of nitrogen-rich recovery solutions to agricultural fields could be considered (Walsh et al., 2012). It is still necessary to demonstrate the economic viability of this recovery, i.e. whether the value of the products recovered outweighs the economic effort involved (Perera et al., 2019). WWTPs can be transformed into Water and Resource Recovery Facilities (W&RRFs) by recovering resources such as water, energy, biosolids, and nutrients, simultaneously reducing operating costs (Geissdoerfer et al., 2017; Lin et al., 2015). The interconnection between water and energy has become an integral component of modern economies. In developed countries, WWTPs contribute to almost 3% of the total electrical energy load of a country (Capodaglio & Olsson, 2020). Municipal wastewater contains approximately 5 to 10 times more energy (chemical and thermical) than is needed for wastewater treatment processes (Bauer, 2014; International Energy Agency, 2022; Tarallo, 2015). Wastewater is a highly concentrated source of organic matter (OM) and can be considered a carrier of chemical energy (Nasr Esfahani et al., 2022). The total energy potential in wastewater is estimated at 898·1015 J annually worldwide, with 80% being thermal energy and the remaining 20% chemical energy (Barnard & Stensel, 2014). According to data from the Water Environment Research Foundation (WERF) (Tarallo, 2015), it is possible to produce approximately 3.2 ·109 m3 of biogas globally, equivalent to 72 ·1015 J annually (Holmgren et al., 2016; Verstraete et al., 2009). However, biogas alone would only account for 8% of all the energy available in wastewater, highlighting the need to enhance current technologies and develop new and more efficient technologies that allow the recovery and use of the energy contained in wastewater. Resource recovery poses a crucial challenge all the countries, particularly those lacking modern facilities. Taking the Macro Metropolis of Sao Paulo, Brazil, as an example, only 26% of wastewater treatment plants (WWTPs) implement resource recovery techniques (Chrispim et al., 2020). Another example is Italy, where a detailed analysis of 600 plants showed that over 60% exhibit no significant signs of resource recovery (Papa et al., 2017). 25 These results emphasize the need to transition towards more sustainable WWTPs globally. There is a substantial gap in the implementation of resource recovery practices worldwide, underscoring the urgency to adopt more sustainable practices such as Water and Resource Recovery Facilities (W&RRFs) (Coats & Wilson, 2017). These facilities not only align with circular economy objectives but also address fertilizer costs, generate energy, and produce purified water. The necessity for improvements in tertiary treatment and the advancement towards W&RRFs may lead to these facilities being considered innovative biofactories (self-sufficient or even energy-producing water treatment plants) producing materials like biomass, biofuels, biofertilizers, and bioplastics (Sheikh et al., 2023). 1.2.3. Alternative ammonia recovery techniques from wastewater As previously mentioned, N (or the energy derivate) recovery is usually applied as a tertiary treatment divided into three categories: biological (BES, microalgae, duckweed and macrophyte westlands) (Zubair et al., 2020), physical-chemical and hybrid, a treatment train where two or more of these technologies would be combined (Cherif et al., 2023; Reig et al., 2022). Commonly, within physical-chemical, air stripping and chemical precipitation both could be regarded as the tertiary treatment technology with the greatest presence in WWTPs. New Alternative nitrogen recovery techniques are implemented with the aim of avoiding the use of air stripping stage given that this is an energy intensive technology and thus making the system very economically feasible as well as guaranteeing the generation of a valuable product from nitrogen (Yan et al., 2018). Nutrient recovery is a promising strategy to reduce natural resource exploitation (Robles et al., 2020) in front of Biological Nitrogen Removal (BNR) (Mao et al., 2020) which focuses on the removal. During the past three decades, there has been a consistent rise in literature focusing on ammonia recovery, as depicted in Figure 3a and 3b. This surge in publications is driven by the necessity to explore alternative sources of ammonia. Urban wastewater streams, for instance, often exhibit low concentrations of nutrients but are present in substantial volumes, posing challenges for recovery. Consequently, there's a critical need to develop technologies that selectively target these nutrients or are capable of concentrating such streams (X. Guo et al., 2023; Tao et al., 2019). 32 This implies that the feed flows in one direction, while the stripping solution flows in the opposite directionThe movement of ammonia toward the acid compartment is driven by the partial vapor pressure difference controlled by the pH (Gabelman & Hwang, 1999; Hasanoĝlu et al., 2010). It is crucial to maintain a pH in the feed effluent above the pKa of ammonia (pKa = 9.25) to enable the flow of NH3 in the gas phase through the pores of the membrane (Licon et al., 2015). A critical aspect in membrane contactors is achieving affordable mass transfer and NH3 flow through the membrane (Reig et al., 2021). For effective membrane performance, a deep understanding of membrane transport mechanism as illustrated in Figure 5b, becomes essential. Membranes used are crafted from hydrophobic polymers, including materials such as polypropylene (PP), polyethylene (PE), polytetrafluoroethylene (PTFE), poly(tetrafluoroethylene-co-perfluorovinyl ether) (PFA), or polyvinylidene fluoride (PVDF). Notably, the innovative membranes from SEPAREL® are constructed from Polytetramethylpentene (PMP). In the presence of an polar liquid on one side of the hydrophobic microporous membrane, the membrane remains unwetted due to surface tension, and the liquid is unable to permeate the pores, a phenomenon dictated by the contact angle with the surface (Pabby & Sastre, 2013). In aqueous mediums, the likelihood of pore wetting increases this occurs because polar liquids like water has the natural tendency to fill and occupy empty spaces on a surface, including the pores of a membrane. It is crucial to prevent membrane wetting, a challenge addressed through the control of surface tension facilitated by the chemical composition of the membrane surface and the pore geometry (Bavarella et al., 2022; McLeod et al., 2015). Hydrophobic membranes are the basis of membrane contactors, with various configurations such as flat sheet, spiral-wound, rotary annular, and hollow fiber, among which the hollow fiber configuration is most commonly used (Bazhenov et al., 2018) (Figure 6). Typically arranged in a shell and lumen configuration, this design offers a robust structure, minimizing size and protecting the membranes. However, tubular configurations have been reported particularly useful in processes prone to fouling and requiring regular cleaning (Majd & Mukhtar, 2013). 33 Figure 6. Schematic reproduction of Hollow Fiber Membrane Contactor (3M Liqui-cell) Both flat sheet and hollow fiber configurations facilitate fluid exchange (gas/gas, gas/liquid, or liquid/liquid). Comparatively, the hollow fiber configuration boasts several advantages, including greater membrane area per unit volume, enhanced mechanical strength due to a more rugged and compact structure, and increased convenience in module manufacturing, repair, and membrane operation (Bazhenov et al., 2018). In a patented method (United States Patent, Patent No.: US 9,708,200 B2) by Szogi et al. 2017, passive ammonia capture from liquid effluent is achieved by passing ammonia gas molecules through a gas-permeable membrane immersed in the effluent (Szogi et al., 2017). In some studies, Vanotti et al. 2017, configured a setup for ammonia recovery using gas-permeable membranes immersed in an effluent containing 2.3g NH4+·L-1, transferring the recovered ammonia to a stripping stream to obtain an ammonium salt (Vanotti et al., 2017). In terms of operation the membrane contactor configuration used should be considered, "open loop" configuration (Figure.7 (a)) where the feed passes through the contactor and is collected in a treated water tank, or if the feed solution is recirculated and recycled to the feed tank after passing through the membrane is named "closed loop" configuration (figure 7. (b)). It is possible to assemble a set up with one or more membrane contactors, 34 (Vecino et al., 2019) using Hollow fiber liquid-liquid membrane contactor in a two-step closed loop configuration with satisfactory results. The advantages of ammonia recovery by means of membrane contactors are high selectivity, i.e., the pores of the membrane, being full of air, provide an optimum barrier for any other constituents contained in the feed water. This must always depend on the control of fouling and wetting.(Tan et al., 2006; Wang et al., 2023; Zarebska et al., 2014). For this reason, this leads to a series of advantages when it concerns the valorisation of Figure 7. HFMC “open loop” set up (a) and HFMC “closed loop” set up (b) (Darestani et al., 2017; Vecino et al., 2019) a) b) 35 the collected product since it allows it to reach a very high level of purity. (Al-Juboori et al., 2022). In the case of membrane contactors, the control of the partial pressure for mass transfer is carried out by means of scrupulous control of the pH (Yu et al., 2021). Recent developments in membrane technology have led to result a high density of membrane packingfor example 30 cm2/cm3 of density packing (Yang et al., 2013) thus optimising the space and a reduced portion of the membrane is a large effective area which results in a higher recovery. This allows Membrane Contactor Technology to be a suitable option at different levels, on a domestic as well as a metropolitan scale (Lee et al., 2021). 1.2.6. Biological process for nitrogen recovery In recent years bioelectrochemical (BES) systems has been postulated as a potential reliable alternative for nutrient recovery in wastewater. In these systems the nitrogen is concentrated under the influence of an electrical current in the cathode area for further recovery (Kuntke et al., 2018), these systems are based on the use of microorganisms that function as catalysts for the reactions that take place in either the anode or the cathode by transforming chemical energy into electrical energy (or vice versa), just as a battery does (Rodríguez Arredondo et al., 2015). Within the BES there are mainly Microbial Electrolysis Cells (MEC) and Microbial Fuel Cells (MFC) (Jatoi et al., 2022; Kelly & He, 2014). The main difference between the both is the direction of the thermodynamic reaction, in the case of the MFC, the reaction is favoured in the anode producing energy when the oxygen is reduced, on the other hand, the MEC needs an extra energetic application due to the fact that the reaction is thermodynamically unfavourable in the cathode (Yan et al., 2018). Certainly, the most outstanding feature of this technology is the ability to generate energy (by producing electricity or hydrogen) while facilitating nutrient recovery. One of the main disadvantages of this technology is the fouling of the membrane, which generates a huge limitation (Mondor et al., 2009; Rahmani et al., 2020). 1.3. Research challenges and objectives However, despite the relatively considerable amount of studies, and the number increasing each year, there is a certain gap in the knowledge of membrane contactors. This gap is not in the understanding of the operating mechanism but rather in the 36 application capabilities, such as the optimal type of influent for this technology, the operating conditions that contribute to higher performance, and the possibilities for scaling up. While the literature boasts an extensive repertoire of articles showcasing promising results with membrane contactors, these outcomes often pertain to very specific conditions that are not easily transferable to other scenarios, especially when dealing with real wastewater rather than synthetic water. The challenge lies in determining where to direct the recovery efforts whether to focus on the mainstream (0.1 g·L-1), the sides-tream or centrate from anaerobic digestion (1-3 g·L-1), agri-food wastewater effluents (>5 g·L-1), industrial streams (>5 g·L-1), or other waste streams sources of ammonia (Sheikh et al., 2023). The successful application and integration of membrane contactors in these diverse streams require careful consideration of the water quality and pre-treatment requirements. This involves employing techniques like coagulation/flocculation, as well as implementing filtration stages such as Microfiltration (MF) and Ultrafiltration (UF) to reduce particulate material, suspended solids, and the dissolved and colloidal chemical demand of total oxygen (COD), thereby avoiding membrane fouling scenarios. Additionally, the selection of membranes involves evaluating critical properties such as hydrophobicity, contact angle, and configuration. The question arises whether the commonly used symmetrical PP fibers are the most ideal solution, or if other polymers such as PMP would achieve similar performance. Moreover, exploring the potential benefits of an asymmetrical membrane structure adds another layer to this consideration. Compounding these challenges is the absence of standardization in result representation within the literature. This lack of consistency in reporting methods poses difficulties in comparing findings across different articles, hindering the extrapolation of conclusions and the identification of best practices. Consequently, the establishment of standards in result presentation becomes crucial for enhancing the overall understanding and evaluation of the effectiveness of membrane contactors in ammonia recovery applications. 37 1.4. References Agrahari, G. K., Shukla, S. K., Verma, N., & Bhattacharya, P. K. (2012). Model prediction and experimental studies on the removal of dissolved NH 3 from water applying hollow fiber membrane contactor. Journal of Membrane Science, 390–391, 164–174. https://doi.org/10.1016/j.memsci.2011.11.033 Aguilar-Moreno, M., Vinardell, S., Reig, M., Vecino, X., Valderrama, C., & Cortina, J. L. (2022). Impact of Sidestream Pre-Treatment on Ammonia Recovery by Membrane Contactors: Experimental and Economic Evaluation. Membranes, 12(12). https://doi.org/10.3390/membranes12121251 Ahmadijokani, F., Molavi, H., Peyghambari, A., Shojaei, A., Rezakazemi, M., Aminabhavi, T. M., & Arjmand, M. (2022). Efficient removal of heavy metal ions from aqueous media by unmodified and modified nanodiamonds. Journal of Environmental Management, 316(April), 115214. https://doi.org/10.1016/j.jenvman.2022.115214 Al-Juboori, R. A., Uzkurt Kaljunen, J., Righetto, I., & Mikola, A. (2022). Membrane contactor onsite piloting for nutrient recovery from mesophilic digester reject water: The effect of process conditions and pre-treatment options. Separation and Purification Technology, 303(September), 122250. https://doi.org/10.1016/j.seppur.2022.122250 Ammonia Technology Roadmap. (2021). Ammonia Technology Roadmap. https://doi.org/10.1787/f6daa4a0-en Àrea Metropolitana de Barcelona. (2002). Depuradora i estació de regeneració d ’ aigua del Prat de Llobregat. Azreen, I., Lija, Y., & Zahrim, A. Y. (2017). Ammonia nitrogen removal from aqueous solution by local agricultural wastes. IOP Conference Series: Materials Science and Engineering, 206(1). https://doi.org/10.1088/1757-899X/206/1/012077 Barnard, J. L., & Stensel, H. D. (2014). The activated sludge process in service of humanity. DSD International Conference, Figure 1. http://www.archive.org/details/cu31924005016419 38 Barros, A., Vecino, X., Reig, M., & Cortina, J. L. (2022). Coagulation and Flocculation Optimization Process Applied to the Sidestream of an Urban Wastewater Treatment Plant. Water (Switzerland), 14(24), 1–26. https://doi.org/10.3390/w14244024 Bauer, D. (2014). Water-Energy Nexus : Challenges and Opportunities. Bavarella, S., Luqmani, B., Thomas, N., Brookes, A., Moore, A., Vale, P., Pidou, M., & McAdam, E. J. (2022). CO2 absorption into aqueous ammonia using membrane contactors: Role of solvent chemistry and pore size on solids formation for low energy solvent regeneration. Separation and Purification Technology, 290(March), 120786. https://doi.org/10.1016/j.seppur.2022.120786 Bazhenov, S. D., Bildyukevich, A. V., & Volkov, A. V. (2018). Gas-liquid hollow fiber membrane contactors for different applications. Fibers, 6(4). https://doi.org/10.3390/fib6040076 Beckinghausen, A., Odlare, M., Thorin, E., & Schwede, S. (2020). From removal to recovery: An evaluation of nitrogen recovery techniques from wastewater. Applied Energy, 263(February), 114616. https://doi.org/10.1016/j.apenergy.2020.114616 Bird, F., Clarke, A., Davies, P., & Surkovic, E. (2020). Ammonia: zero-carbon fertiliser, fuel and energy store. In The Royal Society. https://royalsociety.org/- /media/policy/projects/green-ammonia/green-ammonia-policy-briefing.pdf Bonmatí, A., & Flotats, X. (2003). Air stripping of ammonia from pig slurry: Characterisation and feasibility as a pre- or post-treatment to mesophilic anaerobic digestion. Waste Management, 23(3), 261–272. https://doi.org/10.1016/S0956- 053X(02)00144-7 Bouropoulos, N. C., & Koutsoukos, P. G. (2000). Spontaneous precipitation of struvite from aqueous solutions. Journal of Crystal Growth, 213(3–4), 381–388. https://doi.org/10.1016/S0022-0248(00)00351-1 Burks, T. F., Shearer, S. a, Fulton, J. P., & Sobolik, C. J. (2004). L Iterature R Eview. 20(3), 269–276. https://doi.org/10.1016/j.profnurs.2010.10.006 Capodaglio, A. G., & Olsson, G. (2020). Energy issues in sustainable urban wastewater management: Use, demand reduction and recovery in the urban water cycle. Sustainability (Switzerland), 12(1). https://doi.org/10.3390/su12010266 39 Cherif, H., Risse, H., Abda, M., Benmansour, I., Roth, J., & Elfil, H. (2023). Nanofiltration as an efficient tertiary wastewater treatment for reuse in the aquaponic system in Tunisia. Journal of Water Process Engineering, 52(February), 103530. https://doi.org/10.1016/j.jwpe.2023.103530 Chrispim, M. C., Scholz, M., & Nolasco, M. A. (2020). A framework for resource recovery from wastewater treatment plants in megacities of developing countries. Environmental Research, 188(June), 109745. https://doi.org/10.1016/j.envres.2020.109745 Coats, E. R., & Wilson, P. I. (2017). Toward Nucleating the Concept of the Water Resource Recovery Facility (WRRF): Perspective from the Principal Actors. Environmental Science and Technology, 51(8), 4158–4164. https://doi.org/10.1021/acs.est.7b00363 Company, T. M., Group, H., Group, E., States, U., Fertilizers, P., Treatment, S., Overview, R. I., Intelligence, M., Analysis, C., & Data, H. (2023). With 9 . 1 % CAGR , Chemical Fertilizers Market Size to Reach USD 23 . 1 Billion by 2030 | Global Chemical Fertilizers Industry Trends , Share , Value , Analysis & Forecast Report by Zion Market. Darestani, M., Haigh, V., Couperthwaite, S. J., Millar, G. J., & Nghiem, L. D. (2017). Hollow fibre membrane contactors for ammonia recovery: Current status and future developments. Journal of Environmental Chemical Engineering, 5(2), 1349–1359. https://doi.org/10.1016/j.jece.2017.02.016 El Diwani, G., El Rafie, S., El Ibiari, N. N., & El-Aila, H. I. (2007). Recovery of ammonia nitrogen from industrial wastewater treatment as struvite slow releasing fertilizer. Desalination, 214(1–3), 200–214. https://doi.org/10.1016/j.desal.2006.08.019 Eskicioglu, C., Galvagno, G., & Cimon, C. (2018a). Approaches and processes for ammonia removal from side-streams of municipal effluent treatment plants. Bioresource Technology, 268(May), 797–810. https://doi.org/10.1016/j.biortech.2018.07.020 Eskicioglu, C., Galvagno, G., & Cimon, C. (2018b). Approaches and processes for ammonia removal from side-streams of municipal effluent treatment plants. Bioresource Technology, 268(May), 797–810. 40 https://doi.org/10.1016/j.biortech.2018.07.020 Etter, B., Tilley, E., Khadka, R., & Udert, K. M. (2011). Low-cost struvite production using source-separated urine in Nepal. Water Research, 45(2), 852–862. https://doi.org/10.1016/j.watres.2010.10.007 European Commission. (2019). Fertilisers in the EU - Prices, trade and use. EU Agricultural Markets Briefs, 15(15), 1–8. European Commission. (2023). Organic farming in the EU - A decade of organic growth. Organic Farming in the EU - A Decade of Organic Growth, 20, 32. https://ec.europa.eu/info/sites/info/files/food-farming- fisheries/farming/documents/market-brief-organic-farming-in-the- eu_mar2019_en.pdf FAO. (2019). World fertilizer outlook and trends to 2022. Faragò, M., Damgaard, A., Madsen, J. A., Andersen, J. K., Thornberg, D., Andersen, M. H., & Rygaard, M. (2021). From wastewater treatment to water resource recovery: Environmental and economic impacts of full-scale implementation. Water Research, 204(July). https://doi.org/10.1016/j.watres.2021.117554 Fernández, B. (2008). Tratamiento anaerobio de residuos de la industria alimentaria. II Congreso Iberoamericano Sobre Seguridad Alimentaria. V Congreso Español de Ingeniería de Alimentos, 1–8. https://upcommons.upc.edu/bitstream/handle/2117/97917/Flotats y Fernandez_CESIA2008.pdf Fowler, D., Coyle, M., Skiba, U., Sutton, M. A., Cape, J. N., Reis, S., Sheppard, L. J., Jenkins, A., Grizzetti, B., Galloway, J. N., Vitousek, P., Leach, A., Bouwman, A. F., Butterbach-Bahl, K., Dentener, F., Stevenson, D., Amann, M., & Voss, M. (2013). The global nitrogen cycle in the Twentyfirst century. Philosophical Transactions of the Royal Society B: Biological Sciences, 368(1621). https://doi.org/10.1098/rstb.2013.0164 Gabelman, A., & Hwang, S. T. (1999). Hollow fiber membrane contactors. Journal of Membrane Science, 159(1–2), 61–106. https://doi.org/10.1016/S0376- 7388(99)00040-X 41 Galloway, J. N., Townsend, A. R., Erisman, J. W., Bekunda, M., Cai, Z., Freney, J. R., Martinelli, L. A., Seitzinger, S. P., & Sutton, M. A. (2008). Transformation of the Nitrogen Cycle : Science, 320(May), 889–892. Gao, L., Li, J. De, Yang, G., Zhang, J., & Xie, Z. (2020). De-ammonification using direct contact membrane distillation – An experimental and simulation study. Separation and Purification Technology, 250(February), 117158. https://doi.org/10.1016/j.seppur.2020.117158 Geissdoerfer, M., Savaget, P., Bocken, N. M. P., & Hultink, E. J. (2017). The Circular Economy – A new sustainability paradigm? Journal of Cleaner Production, 143, 757–768. https://doi.org/10.1016/j.jclepro.2016.12.048 González Montiel, J. M. (2008). Acetato de etilo en la Industria. Journal of Chemical Information and Modeling, 53(9), 287. Guaya, D., Mendoza, A., Valderrama, C., Farran, A., Sauras-Yera, T., & Cortina, J. L. (2020). Use of nutrient-enriched zeolite (NEZ) from urban wastewaters in amended soils: Evaluation of plant availability of mineral elements. Science of the Total Environment, 727, 138646. https://doi.org/10.1016/j.scitotenv.2020.138646 Guaya, D., Valderrama, C., Farran, A., Sauras, T., & Cortina, J. L. (2018). Valorisation of N and P from waste water by using natural reactive hybrid sorbents: Nutrients (N,P,K) release evaluation in amended soils by dynamic experiments. Science of the Total Environment, 612, 728–738. https://doi.org/10.1016/j.scitotenv.2017.08.248 Guida, S., Van Peteghem, L., Luqmani, B., Sakarika, M., McLeod, A., McAdam, E. J., Jefferson, B., Rabaey, K., & Soares, A. (2021). Ammonia recovery from brines originating from a municipal wastewater ion exchange process and valorization of recovered nitrogen into microbial protein. Chemical Engineering Journal, 427(April 2021), 130896. https://doi.org/10.1016/j.cej.2021.130896 Guo, C. H., Stabnikov, V., & Ivanov, V. (2010). The removal of nitrogen and phosphorus from reject water of municipal wastewater treatment plant using ferric and nitrate bioreductions. Bioresource Technology, 101(11), 3992–3999. https://doi.org/10.1016/j.biortech.2010.01.039 Guo, X., Chen, J., Wang, X., Li, Y., Liu, Y., & Jiang, B. (2023). Sustainable ammonia 48 Smith, C., Hill, A. K., & Torrente-Murciano, L. (2020). Current and future role of Haber- Bosch ammonia in a carbon-free energy landscape. Energy and Environmental Science, 13(2), 331–344. https://doi.org/10.1039/c9ee02873k Suzanne, E., Absi, N., & Borodin, V. (2020). Towards circular economy in production planning: Challenges and opportunities. European Journal of Operational Research, 287(March), 168–190. https://doi.org/10.1016/j.ejor.2020.04.043 Szogi, A. A., Data, R. U. S. A., & Ariel, L. (2017). ( 12 ) United States Patent. 2(12). Tan, X., Tan, S. P., Teo, W. K., & Li, K. (2006). Polyvinylidene fluoride (PVDF) hollow fibre membranes for ammonia removal from water. Journal of Membrane Science, 271(1–2), 59–68. https://doi.org/10.1016/j.memsci.2005.06.057 Tao, W., Bayrakdar, A., Wang, Y., & Agyeman, F. (2019). Three-stage treatment for nitrogen and phosphorus recovery from human urine: Hydrolysis, precipitation and vacuum stripping. Journal of Environmental Management, 249(July), 109435. https://doi.org/10.1016/j.jenvman.2019.109435 Tarallo, S. (2015). A Guide to Net-Zero Energy Solutions for Water Resource Recovery Facilities. In Water Intelligence Online (Vol. 14). https://doi.org/10.2166/9781780407685 Tarpeh, W. A., Udert, K. M., & Nelson, K. L. (2017). Comparing ion exchange adsorbents for nitrogen recovery from source-separated urine. Environmental Science and Technology, 51(4), 2373–2381. https://doi.org/10.1021/acs.est.6b05816 Thakur, I. S., & Medhi, K. (2019). Nitrification and denitrification processes for mitigation of nitrous oxide from waste water treatment plants for biovalorization: Challenges and opportunities. Bioresource Technology, 282(January), 502–513. https://doi.org/10.1016/j.biortech.2019.03.069 Thorin, E., Olsson, J., Schwede, S., & Nehrenheim, E. (2018). Co-digestion of sewage sludge and microalgae – Biogas production investigations. Applied Energy, 227(January 2017), 64–72. https://doi.org/10.1016/j.apenergy.2017.08.085 Ukwuani, A. T., & Tao, W. (2016). Developing a vacuum thermal stripping – acid absorption process for ammonia recovery from anaerobic digester effluent. Water Research, 106, 108–115. https://doi.org/10.1016/j.watres.2016.09.054 49 Ulbricht, M., Schneider, J., Stasiak, M., & Sengupta, A. (2013). Ammonia recovery from industrial wastewater by transMembranechemiSorption. Chemie-Ingenieur- Technik, 85(8), 1259–1262. https://doi.org/10.1002/cite.201200237 Uludag-Demirer, S., Demirer, G. N., & Chen, S. (2005). Ammonia removal from anaerobically digested dairy manure by struvite precipitation. Process Biochemistry, 40(12), 3667–3674. https://doi.org/10.1016/j.procbio.2005.02.028 Vanotti, M. B., Dube, P. J., Szogi, A. A., & García-González, M. C. (2017). Recovery of ammonia and phosphate minerals from swine wastewater using gas-permeable membranes. Water Research, 112, 137–146. https://doi.org/10.1016/j.watres.2017.01.045 Vecino, X., Reig, M., Bhushan, B., Gibert, O., Valderrama, C., & Cortina, J. L. (2019). Liquid fertilizer production by ammonia recovery from treated ammonia-rich regenerated streams using liquid-liquid membrane contactors. Chemical Engineering Journal, 360(December 2018), 890–899. https://doi.org/10.1016/j.cej.2018.12.004 Vecino, X., Reig, M., Gibert, O., Valderrama, C., & Cortina, J. L. (2020). Integration of liquid-liquid membrane contactors and electrodialysis for ammonium recovery and concentration as a liquid fertilizer. Chemosphere, 245, 125606. https://doi.org/10.1016/j.chemosphere.2019.125606 Verstraete, W., Van de Caveye, P., & Diamantis, V. (2009). Maximum use of resources present in domestic “used water.” Bioresource Technology, 100(23), 5537–5545. https://doi.org/10.1016/j.biortech.2009.05.047 Walsh, J. J., Jones, D. L., Edwards-Jones, G., & Williams, A. P. (2012). Replacing inorganic fertilizer with anaerobic digestate may maintain agricultural productivity at less environmental cost. Journal of Plant Nutrition and Soil Science, 175(6), 840– 845. https://doi.org/10.1002/jpln.201200214 Wang, P., Cheng, W., Zhang, X., Li, J., Ma, J., & Zhang, T. (2023). Engineering a protective surface layer to resist membrane scaling and scale-induced wetting in membrane distillation for the treatment of hypersaline wastewater. Chemical Engineering Journal, 452(July 2022), 1–10. https://doi.org/10.1016/j.cej.2022.139167 50 WIDANARNI, EKASARI, J., & MARYAM, S. I. T. I. (2012). Evaluation of Biofloc Technology Application on Water Quality and Production Performance of Red Tilapia Oreochromis sp. Cultured at Different Stocking Densities. HAYATI Journal of Biosciences, 19(2), 73–80. https://doi.org/10.4308/hjb.19.2.73 Wuang, S. C., Khin, M. C., Chua, P. Q. D., & Luo, Y. D. (2016). Use of Spirulina biomass produced from treatment of aquaculture wastewater as agricultural fertilizers. Algal Research, 15, 59–64. https://doi.org/10.1016/j.algal.2016.02.009 Xie, M., Shon, H. K., Gray, S. R., & Elimelech, M. (2016). Membrane-based processes for wastewater nutrient recovery: Technology, challenges, and future direction. Water Research, 89, 210–221. https://doi.org/10.1016/j.watres.2015.11.045 Yan, T., Ye, Y., Ma, H., Zhang, Y., Guo, W., Du, B., Wei, Q., Wei, D., & Ngo, H. H. (2018). A critical review on membrane hybrid system for nutrient recovery from wastewater. Chemical Engineering Journal, 348(December 2017), 143–156. https://doi.org/10.1016/j.cej.2018.04.166 Yang, D., Le, L., Martinez, R., & Morrison, M. (2013). Hollow fibers structured packings in olefin/paraffin distillation: Apparatus scale-up and long-term stability. Industrial and Engineering Chemistry Research, 52(26), 9165–9179. https://doi.org/10.1021/ie400126y Yara International. (2022). Fertilizer Industry Handbook 2022. December, 32. https://www.yara.com/siteassets/investors/057-reports-and- presentations/other/2022/fertilizer-industry-handbook-2022.pdf/ Ye, Y., Ngo, H. H., Guo, W., Liu, Y., Chang, S. W., Nguyen, D. D., Liang, H., & Wang, J. (2018). A critical review on ammonium recovery from wastewater for sustainable wastewater management. Bioresource Technology, 268(June), 749–758. https://doi.org/10.1016/j.biortech.2018.07.111 Yu, S., Qin, Y., Zhao, Q., Li, M., Yu, H., Kang, G., & Cao, Y. (2021). Nafion-PTFE hollow fiber composite membranes for ammonia removal and recovery using an aqueous-organic membrane contactor. Separation and Purification Technology, 271(April), 118856. https://doi.org/10.1016/j.seppur.2021.118856 Yue, X., You, A., Liu, Y., Lai, M., & Zhang, K. (2023). Low-concentration methanol 51 effect on the microorganisms, nitrogen removal, and recovery of the completely autotrophic nitrogen removal over nitrite. Water Science and Technology, 87(1), 130–143. https://doi.org/10.2166/wst.2022.417 Yüzbasıoglu, A. E., Tatarhan, A. H., & Gezerman, A. O. (2021). Decarbonization in ammonia production, new technological methods in industrial scale ammonia production and critical evaluations. Heliyon, 7(10). https://doi.org/10.1016/j.heliyon.2021.e08257 Zarebska, A., Nieto, D. R., Christensen, K. V., & Norddahl, B. (2014). Ammonia recovery from agricultural wastes by membrane distillation: Fouling characterization and mechanism. Water Research, 56, 1–10. https://doi.org/10.1016/j.watres.2014.02.037 Zhu, Y., Chang, H., Yan, Z., Liu, C., Liang, Y., Qu, F., Liang, H., & Vidic, R. D. (2024). Review of ammonia recovery and removal from wastewater using hydrophobic membrane distillation and membrane contactor. Separation and Purification Technology, 328(September 2023), 125094. https://doi.org/10.1016/j.seppur.2023.125094 Zubair, M., Wang, S., Zhang, P., Ye, J., Liang, J., Nabi, M., Zhou, Z., Tao, X., Chen, N., Sun, K., Xiao, J., & Cai, Y. (2020). Biological nutrient removal and recovery from solid and liquid livestock manure: Recent advance and perspective. Bioresource Technology, 301(October 2019), 122823. https://doi.org/10.1016/j.biortech.2020.122823 52 CHAPTER 2 Objectives and Thesis overview 2.1. Objectives The primary objective of this PhD thesis is to advance the development of a N recovery process aimed at producing a liquid fertilizer as a high valuable product, by leveraging the high selectivity of hydrophobic membrane contactors. These contactors facilitate the transport of solutes only in vapor or gas phases through the membrane. The study will focus on different types of streams: urban wastewater generated in sewage sludge anaerobic digesters (side-streams) with concentrations of up to 1.5 g·L-1, wastewater from the chemical industry, ranging from diluted streams (0.1 g·L-1) to concentrated ones (>10 g·L-1) containing ammonium/ammonia and digested effluents from the food industry with a NH3 concentration between 3.5 and 10 g·L-1 and a very high organic load. To achieve this goal, laboratory-scale tests for ammonia recovery will be conducted using membrane contactors (hollow fiber and flat sheet) and to a lesser extent membrane distillation. The experimental approach will be tailored to generate results that align with the initially defined objectives. Specifically, the research will address the following key aspects: - Identify sample pre-treatment needs to reduce suspended solids (SS), chemical oxygen demand (COD) and alkalinity to ensure safe operation, determination and optimization of pre-treatment trains: comparison of coagulation/sedimentation stages with sand filtration or UF/NF membrane technologies. - To analyse the performance of novel hollow fiber membrane polymer; polypropylene and polymethylpentene (SEPAREL®), to quantify and compare in terms of nitrogen recovery and the transport of water, using synthetic and real solution to identify physical and chemical fouling and scaling events; to determine mass transfer parameters influencing membrane performance and compare with existing literature. 53 - To obtain and evaluate the purity, composition and concentration of liquid fertilizers as a pure product in the form of fertilizer that helps to mitigate the overwhelming demand for fertilizers due to the exploitation of extensive agriculture with commercial solutions and to quantify and economically evaluate the cost of the by-product. - To estimate the economic feasibility to determine the potential role of proven technology as an effective alternative to the current fertilizer industry, reducing the depletion of raw materials, cost and impact on the environment, all within the framework of a circular economy. 2.2. Thesis overview This section aims to give a more concrete view of the chapters of this PhD thesis in line with the objectives assigned to this work to provide a better understanding, monitoring, and ease of reading the manuscript. • Chapter 3. Impact of side-stream pre-treatment on ammonia recovery by membrane contactors: Experimental and economic evaluation • Chapter 4. Ammonia Valorization by Liquid–Liquid Membrane Contactors for Liquid Fertilizers Production: Experimental Conditions Evaluation • Chapter 5. Ammonium recovery and concentration from synthetic wastewater using a poly(4-methyl-1-pentene) (PMP) liquid-liquid membrane contactor: flux performance and mass transport characterization 54 Figure 1. Graphical abstract of the project and how the chapters are interconnected in reference to the objectives. he six topics marked in this paper, along with their corresponding subtopics, appear in different chapters, outlining their relationships with each other. 55 CHAPTER 3 Impact of sidestream pre-treatment on ammonia recovery by membrane contactors: Experimental and economic evaluation 3.1. Abstract Membrane contactor is a promising technology for ammonia recovery from anaerobic digestion centrate. However, high suspended solids and dissolved organic matter concentrations can reduce the effectiveness of the technology. In this study, coagulation-flocculation (C/F) and aeration pretreatments were evaluated to reduce chemical oxygen demand (COD), turbidity, suspended solids and alkalinity before the ammonia recovery stage using a membrane contactor. The mass transfer coefficient (Km) and total ammonia (TAN) recovery efficiency of the membrane contactor increased from 7.80x10-7 to 1.04 x 10-5 m·s-1 and from 8 to 67%, respectively, after pre-treating the sidestream centrate. The pre-treatment results showed that dosing aluminium sulphate (Al2(SO4)3) at 30 mg Al+·L-1 was the best strategy for the C/F process, providing COD, turbidity and TSS removal efficiencies of 50±5, 95±3 and 90±4%, respectively. The aeration step reduced HCO3- content by 51±6% and allowed reducing alkaline consumption by increasing the pH before the membrane contactor. The techno-economic evaluation showed that the combination of C/F, aeration and membrane contactor can be economically feasible for ammonia recovery. Overall, the results of this study demonstrate that C/F and aeration are simple and effective techniques to improve membrane contactor performance for nitrogen recovery from the anaerobic digestion centrate. Keywords: Gas permeable membrane; coagulation-flocculation; resource recovery; circular economy; techno-economic evaluation; 3.2. Introduction Nutrient pollution is one of the major environmental problems due to excessive discharge of nitrogen and phosphorus into the environment. Anthropogenic activities and population growth have increased the amount of nitrogen contained in wastewater. The recovery of nitrogen is particularly important considering that ammonia is the second most produced chemical in the world (Beckinghausen et al., 2020; Lee et al., 2021; Razon, 2018). Ammoniacal-Nitrogen recovery has the potential (i) to reduce the dependency of Haber-Bosch process to obtain nitrogen-based fertilizers, (ii) to produce a fertilizer (e.g. NH4NO3, (NH4)2HPO4, (NH4)2SO4) suitable for commercialisation and (iii) to reintroduce 56 nitrogen into its cycle contributing to circular economy (Darestani et al., 2017; González Montiel, 2008). For this reason, it is important to develop efficient technologies for nitrogen recovery to support the transition of wastewater treatment plants (WWTPs) towards water resource recovery facilities (WRRF) (Puyol et al., 2017). Several technologies have been proposed to recover nitrogen from wastewater treatment plants (WWTPs), such as ion exchange (IX) technologies (Kurniawan et al., 2006), membrane contactors (MC) (Hasanoĝlu et al., 2010; Licon Bernal et al., 2016; Reig et al., 2021), or ultrafiltration (UF) (Hermassi et al., 2017). For instance, Wan et al. 2017, effectively recovered nutrients from the sludge fermentation liquor in a WWTP (N-NH4+ and P-PO43-) using natural zeolites and proposed a model to predict that a maximum recovery of 94% ammonium and 98% phosphate could be achieved (Wan et al., 2017). Among them, ammoniacal nitrogen recovery through membrane contactors has been reported as a suitable technology to achieve high nitrogen recovery efficiencies with relatively low energy inputs (Darestani et al., 2017). By this technology, ammonia in gas form diffuses through a porous hydrophobic membrane from the feed solution to the acidic stripping solution. Subsequently, it can be recovered in ammonium form as a nitrogen-rich fertilizer. (Serra-Toro et al., 2022). Vecino et al. 2019, used a membrane contactor for ammonium recovery as a nutrient-based fertilizer product and achieved a maximum ammonium recovery of 94% using a regenerated stream with ion exchange from an initial sidestream wastewater (Vecino et al., 2019). Sheikh et al. 2022, also achieved similar values (>95%) of recovery using synthetic water and liquid-liquid hollow fibre MC (LL-HFMC) (Sheikh et al., 2022). Additionally, both membrane contactors and ion exchange technologies can be combined as proposed by Sancho et al. 2017. In that study, a concentrated ammonium stream was generated by means of liquidliquid membrane contactors, by previously passing it through zeolites, achieving a recovery of 95% (Sancho et al., 2017). Thus, these publications highlight that membrane contactors have potential to achieve high recovery efficiencies and to obtain ammoniumfree streams. However, membrane contactors still need to overcome some challenges when using streams with high concentration of organic matter. Membrane fouling, caused by organic matter and/or suspended solids, can lead to the deposition of solids as a thin cake layer and increase pore clogging (Leiknes, 2009). This phenomenon generates a reduction of the flux during long-term operation. Thus, to maintain adequate flux levels, it is necessary 57 to increase energy and chemical consumption with a direct impact on the membrane lifetime and economic feasibility (Zarebska et al., 2014). In this regard, some pretreatment strategies have been proposed to reduce fouling of membrane contactors, such as UF (Jiang, 2015), coagulation-flocculation (C/F) processes (Leiknes, 2009) or ion exchange (Rivadeneyra et al., 2007). For example, Rivadeneyra et al. used ion exchange technology and observed a maximum chemical oxygen demand (COD) removal efficiency of 70% with an initial COD load of 4500 mg O2·L-1. Raghu et al. 2007, combined ion exchange with coagulation-flocculation and achieved a COD removal of 80% from an industrial wastewater effluent (Raghu & Ahmed Basha, 2007). C/F consists of destabilization of colloids by surface modification. This reduces the electrostatic repulsive forces between the particles and leads to the formation of larger flocs with improved settling properties (Dosta et al., 2008). The most common coagulants and flocculants used are iron and aluminium salts because these chemicals have demonstrated their effectiveness to reduce the chemical oxygen demand (COD) of liquid streams (Krupińska, 2020; Postolachi et al., 2016). C/F has been widely applied in wastewater treatment applications as it allows removing organic and inorganic matter with relatively low costs (Jiang, 2015; Taboada-Santos et al., 2020). For instance, Al- Juboori et al. 2022 (Al-Juboori et al., 2022) evaluated the use of PAX/polymer or starch as a coagulant to pre-treat the centrate before a membrane contactor. Besides C/F, aeration could also be a useful pre-treatment to reduce the amount of chemicals needed to increase the pH before the membrane contactor stage. Garcia- Gonzalez et al. 2015 applied low flow-rate aeration and increased the pH above 8.5 before the membrane contactor, which allowed reducing the operating costs of ammonia recovery by 57% (García-González et al., 2015). However, to the best of the authors’ knowledge, the combination of C/F technology with aeration has not yet been used to pretreat anaerobic digester centrate prior to a membrane contactor. Therefore, an experimental and economic study is needed to understand how C/F pre-treatment impacts the technical and economic competitiveness of implementing a membrane contactor system for nitrogen recovery. The aim of this work is to evaluate the combination of C/F, aeration and membrane contactor processes to recover ammoniacal nitrogen from the effluent of an anaerobic digester (centrate). To this end, different operating conditions and chemical reagents were 64 Thus, these experiments allowed bicarbonate conversion to CO2(g) (aq) (Eq 4) due to the aeration process promoting the removal of dissolved CO2(g) (aq) as CO2(g) (Eq. 5) and consequently increasing the pH. Subsequently, the pH increased at levels close to pKa. allowed the conversion of NH4+ into ammonia. (Eq. 6). The aeration experiments were performed in duplicate. 3.3.3.6. Flat-sheet membrane contactor The different pre-treatment processes aimed at conditioning the centrate to reduce fouling and clogging in the membrane contactor. A flat-sheet membrane contactor similar to the one used by Hasanoĝlu et al. 2010, was used in this study (Hasanoĝlu et al., 2010). The polytetrafluoroethylene (PTFE) membrane had a surface area of 90 cm2 and a pore size of 0.2 µm. The pH of the feed solution was increased up to 10.2 with NaOH 1 M, to shift the equilibrium towards NH3. The feed solution was stored in a 5 L tank, whereas the acid stripping solution (Concentration of 0.4 M of nitric acid) was stored in a 1.5 L tank. Both tanks were continuously agitated, while nitric acid was continuously added to maintain the pH of the stripping solution in the acidic regime (pH < 2). The feed and stripping solutions were circulated at 450 ml/min in counter current mode towards both sides of the membrane. Further details of the membrane contactor set-up can be found elsewhere (Reig et al., 2021). The ammonia flux through the membrane is driven by the difference between the partial pressure on both sides of the membrane, (pNH3,f-pNH3,s) and the mass transfer coefficient (Km(NH3)) (Eq. (7)). Where pNH3,s is the partial pressure of ammonia in the shell side (atm), pNH3,f is the partial pressure in the feed side (atm), Km(NH3) is the ammonia mass transfer coefficient (m·s-1), R is the universal gas constant coefficient (0.082 atm·m3·K-1 mol·K) and T is the temperature of the system (K). Subsequently, Eq. (7) can be expressed as Eq. (8) considering that: (i) the partial pressure of ammonia on both sides of the membrane can be assumed as the concentration of ammonia on either side, (ii) the pH does not vary during the experimental procedure, meaning that the concentration of ammonia is proportional to the TAN concentration in JNH3=Km(NH3)(pNH3,f-pNH3,s) RT (7) 65 the feed solution and (iii) the ammonia partial pressure in the stripping side is negligible (Licon Bernal et al., 2016; Vecino et al., 2019). Where Am is the membrane area (m2), C0(NH3)f and Ct(NH3)f are the feed ammonia concentration (mg·L-1) at the initial time and at the experimental time, respectively, and 𝑉𝑓 is the feed volume (m3). The tests were conducted for both untreated and pre-treated centrate to evaluate and compare the membrane contactor performance before and after pre-treatment implementation. 3.3.4. Analytical methods The anions and cations were analysed by an ion chromatography system (Dionex ICS- 1000 and ICS-1100 Thermo-Fisher Scientific, USA) equipped with a cationic detector (ICS-1000) and an anionic detector (ICS-1100) and controlled by Chromeleon® chromatographic software. A CS16 column (4 x 250 mm) and an AS23 column (4 x 250 mm) (Phenomenex, Barcelona, Spain) were used for cation and anion determination and quantification, respectively. The mobile phase was a 0.03 mol·L-1 CH3SO3H solution for the cation system, and a mixture of 0.8 mmol·L-1 NaHCO3 and 4.5 mmol·L-1 Na2CO3 for the anion system. The COD was analysed through the Standard Method 5220C (Matthews, 2014) using a multiparametric photometer HI83224 (Hanna Instruments, Italy), whereas TSS were analysed through the Standard Method 2540D (APHA, 2012). A turbidimeter HI 93703 (Hanna instruments, Italy) was used to measure the turbidity. Total alkalinity was measured by titration following the Standard Method 2320B (APHA Method 2320, 1992)and using a T70 titrator (Mettler Toledo, United States). 3.3.5. Economic analysis An economic analysis was conducted to evaluate the techno-economic implications of implementing a membrane contactor system for ammonia recovery from the anaerobic digester centrate. Figure 2 shows the configuration evaluated in the economic analysis, ln C0(NH3)f Ct(NH3)f =Km(NH3)Am Vf t (8) 66 which included four different stages: (i) C/F with Al2(SO4)3 to enhance solids sedimentation, (ii) precipitation for suspended solids removal, (iii) aeration to desorb part of the solubilised CO2 and reduce the alkalinity, and (iv) membrane contactor system for nitrogen recovery. Figure 3. Schematic representation of the nitrogen recovery scheme. The membrane contactor system was operated using a HNO3 trapping solution and considering a relation between the feed and trapping solution flow rate of 1:1. The pH of the feed solution was adjusted to 10.2 with NaOH to shift the NH4+/NH3 equilibrium towards NH3. The trapping solution was continuously recirculated from the acid tank to the membrane contactor and replaced when the pH increased up to 6 (Richter et al., 2020). The mass balance was obtained considering that the WWTP generated 150 m3· day-1 of centrate, containing TAN and TSS concentrations of 0.71 g N·L-1 and 0.24 g TSS·L-1, respectively. Detailed information on the mass balance can be found in Table 5. Table 5. Main flow data for the nitrogen recovery scenario under study. (A) Anaerobic digester centrate Unit Value Flow rate m3·day-1 150 TSS g·L-1 0.24 TAN g N·L-1 0.71 pH - 8.1 (B) Al2(SO4)3 solution Flow rate m3·day-1 9 Al2(SO4)3 mg·L-1 500 (C) Coagulation-flocculation effluent Flow rate m3·day-1 159 67 TSS g·L-1 0.22 TAN g N·L-1 0.67 pH - 8.1 (D) Solid fraction precipitator Flow rate m3·day-1 8 TSS g·L-1 3.98 (E) Liquid fraction precipitator Flow rate m3·day-1 151 TAN g N·L-1 0.67 pH - 7.9 (F) Air Flow rate m3·day-1 2288 (G) Effluent aeration tank Flow rate m3·day-1 151 TAN g N·L-1 0.67 pH - 8.9 (I) NaOH solution Flow rate m3·day-1 0.3 NaOH mol·L-1 1 (H) Feeding solution MCt=0 Flow rate m3·day-1 151.3 TAN g N·L-1 0.67 pH - 10.3 (K) Feeding solution MCt=F Flow rate m3·day-1 151.3 TAN g N·L-1 0.24 pH - 8.42 (M) HNO3 solution Volume/cycle m3/cycle 0.52 Number of cycles cycles/day 22 HNO3 mol·L-1 0.4 pH - 0.4 (N) NH4NO3 solution Volume/cycle m3/cycle 0.52 Number of cycles cycles/day 22 NH4NO3 mol·L-1 0.4 pH - 6 The capital costs, operating costs and revenues were calculated using both lab-scale data and literature average values. The capital costs accounted for membrane contactor, tanks, stirrers, blowers and pumps, whereas the operating costs accounted for energy consumption, sludge disposal, equipment replacement and chemicals’ purchase (i.e., 68 Al2(SO4)3, NaOH and HNO3). Finally, the revenues were obtained considering (i) the commercialisation of the produced NH4NO3 and (ii) the lower nitrogen load to be treated in the mainstream of the WWTP. Table 6 and 7 of the summarise the main design and cost parameters used for the economic analysis. The present value (PV) of the gross cost and revenues were calculated for the nitrogen recovery configuration by using Eq. (9) and Eq. (10), respectively. Subsequently, Eq. (11) was used to calculate the net present value (NPV): PVGC=CAPEX+∑OPEXt (1+i)t T t=1 (9) PVR=∑Rt (1+i)t T t=1 (10) NPV=∑Rt-OPEXt (1+i)t-CAPEX T t=1 (11) Where CAPEX is the capital expenditure (€), OPEXt is the operating expenditure at year t (€), Rt is the revenue at year t (€), PVGC is the PV of the gross cost (€), PVR is the PV of the revenues (€), NPV is the net present value (€), i is the discount rate (5%) and T is the plant lifetime (20 years). Table 6. Main design parameters used for the economic evaluation. Parameter Value Source Coagulationflocculation Specific Al2(SO4)3 consumption (g Al2(SO4)3 ·L-1centrate) 0.19 Lab-scale data Retention time (h) 0.1 Lab-scale data Mixer revolutions (rpm) 100 Lab-scale data Precipitation Retention time (h) 0.5 Lab-scale data Q0/QE (%) 95 Lab-scale data Aeration Retention time (h) 1 Lab-scale data Specific air consumption (NL·h-1·L-1tank) 364 Lab-scale data Membrane contactor Km (m·s-1) 1.04×10-5 Lab-scale data TAN recovery (%) 64 Lab-scale data 69 Flow rate trapping solution:Flow rate feed solution 1:1 Lab-scale data Specific NaOH consumption (mol NaOH·mol-1 TANrecovered) 0.067 Lab-scale data pHt=0 HNO3 trapping solution (-) 0.4 Lab-scale data pHt=F HNO3 trapping solution (-) 6 (Richter et al., 2020) Table 7. Main economic parameters used for the economic evaluation. Parameter Value Source Tank cost (€/m3) 220 (Verrecht et al., 2010) Settler cost (€/m3) 100 (Noriega-Hevia et al., 2021) Pump cost (€/m3/h) 12.1 (Verrecht et al., 2010) Stirrer cost (€/m3Tank) 27.8 (Vinardell et al., 2020) Blower cost (€/Nm3/h) 4.15 (Verrecht et al., 2010) Membrane cost (€/m2) 49 (Noriega-Hevia et al., 2021) Al2(SO4)3 cost (€/kg) 0.16 (Vu et al., 2020) NaOH cost (€/kg) 0.62 (Bouzas et al., 2019) HNO3 cost (€/kg) 0.38 (Das et al., 2018) Electricity cost (€/kWh) 0.1445 (Eurostat, 2021) Sludge disposal cost (€/t TS) 373 (Foladori et al., 2015) Lifetime membrane (years) 10 (Noriega-Hevia et al., 2021) Lifetime auxiliary equipment (years) 10 (Noriega-Hevia et al., 2021) NH4NO3 price (€/kg) 0.43 (Ministerio de Agricultura Pesca y Alimentación, 2022) Energy consumption mainstream N removal (kWh/kg N) 2.38 (Horstmeyer et al., 2018) 70 3.4. Results and discussion The following sections discuss the results concerning the application of C/F and aeration pre-treatments before a membrane contactor. Table 8 shows the COD, TSS, turbidity and ions concentrations of the centrate wastewater used for these tests. Table 8. Initial centrate characterization. The errors represent standard deviation (n=3). Parameter value Units Calcium 90.5±26.8 mg·L-1 Carbonates 3366.7±792.5 mg·L-1 Chlorine 348.0±15.4 mg·L-1 COD 786.0±126.7 mg O2·L-1 Magnesium 33.6±13.4 mg·L-1 Nitrate 30.7±8.8 mg·L-1 Phosphate 138.1±30.2 mg·L-1 Potassium 146.6±7.6 mg·L-1 pH 8.2 ± 0.1 -- Sodium 474.4±18.4 mg·L-1 Sulphate 37.5±10.8 mg·L-1 TAN 650 ±64.5 mg·L-1 TSS 235.0±104.7 mg·L-1 Turbidity 275.1±106.2 NTU 3.4.1. Coagulant and dosage selection for the C/F process Table 9. Presents the COD and turbidity removal efficiencies for the three coagulants (FeCl3, Al2(SO4)3 and Derypol® HT20) analysed in this study. The Al2(SO4)3 coagulant achieved the best COD removal efficiencies (50.2 ± 1.1%), followed by FeCl3 (38.9 ± 0.3%) and Derypol HT20 (36.0 ± 0.3%). Thus, Al2(SO4)3 and FeCl3 were selected for the next set of experiments. The turbidity removal efficiencies ranged from 74.2 to 84.7 %. The lowest turbidity values were obtained by using FeCl3 (74.2 mg·L-1) and they were similar to those achieved by Abdessemed et al. 2000, which achieved turbidity removal values of 66.1% using FeCl3 (Abdessemed et al., 2000). 71 Table 9. Results obtained on COD removal (%) and Turbidity reduction for the coagulation assays coagulant test. The errors represent standard deviation (n=3). Coagulant COD removal (%) Turbidity reduction (%) Al2(SO4)3 50.2±1.1 82.3±1.1 Derypol HT20 36.0±0.3 84.7±0.4 FeCl3 38.9±0.3 74.2±1.7 Table 10. Results of water quality improvement for the coagulation experiments (COD removal (%), Turbidity reduction (%)) as a function of coagulant type and coagulant dose. The errors represent standard deviation (n=3). Dosage Al2(SO4)3 FeCl3 COD removal Turbidity reduction pH COD Turbidity reduction pH (mg·L- 1) 10 51.5 ± 1.2 80.4 ± 2.8 8.0 42.5 ± 0.7 60.3 ± 1.2 8.0 30 56.2 ± 1.0 85.5 ± 4.4 7.7 48.0 ± 0.9 71.2 ± 1.2 7.9 50 50.1 ± 1.7 82.3 ± 3.5 7.4 38.9 ± 1.6 74.2 ± 2.4 7.7 100 41.1 ± 0.9 76.7 ± 1.2 7.1 41.5 ± 1.9 80.6 ± 3.4 7.4 200 62.1 ± 1.2 86.6 ± 4.0 6.9 45.1 ± 2.1 87.9 ± 3.3 7.1 400 66.7 ± 2.5 82.2 ± 1.7 6.1 50.0 ± 1.9 90.4 ± 4.1 6.7 600 64.7 ± 2.1 55.5 ± 2.4 4.3 52.5 ± 1.8 95.7 ± 3.4 6.4 800 66.9 ± 1.0 27.3 ± 3.3 4.1 51.8 ± 1.7 97.0 ±3.0 5.8 72 Table 10 lists the COD and turbidity removal efficiencies of Al2(SO4)3 and FeCl3 for concentrations ranging from 10-800 mg·L-1. The results showed that Al2(SO4)3 provided better COD removal performance in comparison to FeCl3, which reinforces the idea that Al2(SO4)3 is the most favourable coagulant-flocculant to be used as a membrane contactor pre-treatment. On the one hand, the COD removal efficiency increased from 42.5 ± 0.7 to 51.8 ± 1.7% as the FeCl3 concentration increased from 10 to 800 mg·L-1, respectively. On the other hand, the COD removal efficiency increased from 51.5 ± 1.2 to 62.1 ± 1.2% as the Al2(SO4)3 concentration increased from 10 to 200 mg·L-1, respectively. However, in the case of Al2(SO4)3, dosages above 200 mg·L-1 only led to minimal improvements in the COD removal efficiency. This behaviour is due to the fact that applying coagulant dosages above the optimal does not lead to considerable improvements (Duan & Gregory, 2003). The results also showed that the pH progressively decreased as the coagulant dosage increased. In the case ofAl2(SO4)3, when the metal ion (Al+3) hydrolyses in water, it reacts to form complexes (Al(OH)n +(n-3)) compounds. This leads to the formation of CO2(g), which increases the acidity of the solution (Krupińska, 2020). From the results of Table 6, it can be concluded that dosing 30 mg·L-1 of Al (Al2(SO4)3) can be considered as the optimum strategy because this dosage achieved similar COD removal efficiencies than those achieved above 200 mg·L-1, while reducing the coagulant dosage more than seven times. 3.4.2. Optimisation of the operating conditions for the C/F process After selecting the optimum coagulant and dosage (Al2(SO4)3, 30 mg Al+·L-1), the impact of the operational conditions (i.e. mixing time, mixing speed and settling time) on the C/F efficiency was evaluated. Seventeen experiments were tested based on the outputs provided by the Design Expert 11 software (Table 11). Table 11. Experiments set of Design Expert 11 software. Run M. Time (min) M. Speed (rpm) S. Time (min) Run M. Time (min) M. Speed (rpm) S. time (min) 1 25 100 45 10 25 250 15 73 2 15 175 30 11 5 100 45 3 5 250 15 12 5 250 45 4 15 175 30 13 15 175 37.5 5 25 250 45 14 10 175 30 6 25 100 15 15 5 100 15 7 15 137.5 30 16 15 175 30 8 15 175 22.5 17 15 212.5 30 9 20 175 30 These experiments were conducted changing the mixing time, the mixing speed and the settling time. Figure 3 shows the theoretical TSS, turbidity and COD removal values obtained from the Design Expert 11 software for the different mixing time and mixing speed conditions at a fixed settling time of 30 min. It is worth mentioning that only the results of 30 min settling time are illustrated because this condition provided the best results when compared with the other settling times. The results highlighted that reducing the mixing time to 5 minutes and the mixing speed to 100 rpm, would theoretically increase removal values up to 100% in turbidity and suspended solids and up to 70% in COD. Accordingly, the software revealed that there was better removal when mixing time and speed were reduced to the minimum tested values. This behaviour was in agreement with Kan et al. 2002, who reported that higher mixing speed did not give a better coagulation performance (Kan et al., 2002). Figure 3. Theoretical TSS, turbidity and COD removal values for different mixing times and mixing speeds, at a fixed settling time of 30 min (graphics obtained from the Design Expert 11 software). 80 Figure 7. Gross cost, revenues, and net present value for the nitrogen recovery scenario under study. The membrane contactor system represented the most expensive process (55%), followed by aeration (36%) and coagulation-flocculation (9%) (see Figure 8A). The high cost of the membrane contactor system is mainly associated with the intensive consumption of HNO3 and, to a lesser extent, NaOH. In this regard, chemicals’ consumption features the highest cost contribution, representing 57% of the gross cost (Figure 8B). Energy consumption also represents an important fraction of the gross cost (34.1%), which can be attributed to the high energy requirements of the air blower system. These results highlight that chemical consumption and aeration requirements are two important operational factors influencing the economic competitiveness of the system. 81 Figure 8. Gross cost contribution of the nitrogen recovery scenario under study for: (A) the different processes and (B) for the different capital and operating costs of all treatment train (C/F, aeration and Membrane contactor) 3.4.6.2. Sensitivity analysis Figure 9 shows the sensitivity analysis for a ± 30% variation, this variation implies the effects of increasing or decreasing these economic parameters by 30% are being assessed to comprehend how this would impact the financial outcome of the system. The results illustrate that NH4NO3 price featured the highest impact on the NPV. This is particularly 82 important considering that the cost of fertilizers is expected to increase in the future due to the progressive increase in fuel and electricity costs according to Panos & Desing, 2019 (Panos & Densing, 2019). To better understand how NH4NO3 price impacts the economic balance of the system, a sensitivity analysis was conducted for NH4NO3 prices between 0.30 and 0.70 €/kg of dry weight (Figure 9). The results show that the NPV of ammoniacal nitrogen recovery increased from -350,000 to 300,000 € as the NH4NO3 price increased from 0.30 to 0.70 €/kg, respectively. This implies that a positive NPV was achieved at NH4NO3 prices above 0.52 €/kg. Overall, these results highlight that the commercialisation of the produced NH4NO3 fertilizer has the potential to make membrane contactor configuration economically feasible. Figure 9. Sensitivity analysis for a ± 30% variation involves evaluating the consequences of both increasing and decreasing these economic parameters by 30%, aiming to understand how this would affect the financial outcome of the project. Nitric acid and electricity costs also feature a noticeable impact on the NPV of the system (Figure 9). This reinforces the idea that chemicals consumption and aeration requirements are two important aspects influencing the economics of this configuration. Conversely, membrane purchase cost variation showed the least influence in the NPV compared with the others factors studied. The low impact of the membrane purchase cost on NPV can be attributed to the high Km coefficient (1.04×10-5 m·s-1) achieved in this study, which is substantially higher than in other studies (Darestani et al., 2017; Noriega-Hevia et al., 2020) . However, it is worth mentioning that the Km could be substantially lower during long-term membrane contactor operation due to organic and inorganic membrane fouling 83 development on the membrane surface. For this reason, a sensitivity analysis was conducted to evaluate the impact of Km on the economic balance of the nitrogen recovery scheme under study (Figure 10). Figure 10. Sensitivity analysis illustrating the impact of mass transfer coefficient (Km) variations and ammonium nitrate (NH4NO3) price on the net present value (NPV) of the nitrogen recovery scheme. The results show that the NPV slightly decreased from -140,000 to -260,000 € as the Km decreased from 1×10-5 to 1×10-6 m·s-1, respectively (Figure 10). However, a sharp decrease of the NPV was observed at Km values below 1×10-6 m·s-1. These results highlight that Km could have a large influence on the economic balance due to its impact on the membrane requirements of the system. For this reason, it is important to look for suitable physical and chemical cleaning strategies able to achieve effective control of long-term membrane fouling without excessive consumption of chemicals and energy. 3.5. Conclusions This study evaluated the implementation of C/F and aeration pre-treatments prior to a membrane contactor stage to recover nitrogen from the anaerobic digester centrate. The results revealed that dosing Al2(SO4)3 at 30 mg Al+·L-1 was the best strategy for the coagulation process. The maximum COD, turbidity and TSS removals (58 and 95 and 90%, respectively) were achieved with a mixing speed of 100 rpm, a mixing time of 5 84 minutes and a settling time of 30 min. The flocculation stage using Fe3O4(s)/SiO2(s) (30- 70% (w/w)) according to our assessment, did not appear to introduce discernible improvements in the removal efficiencies. The aeration stage reduced HCO3- content up to 51% and increased the pH up to 9, without the addition of external chemicals. Subsequently, the effluent from the C/F and aeration stages was fed to the membrane contactor for nitrogen recovery. The membrane contactor recovered 67% of TAN and achieved a concentration factor of 3.8. in the acid solution Finally, although is highly depending on the market price the techno-economic evaluation showed that the combination of C/F, aeration and membrane contactor has potential to be an economically competitive alternative for nitrogen recovery. 85 3.6. References Abdessemed, D., Nezzal, G., & Ben Aim, R. (2000). Coagulation-adsorption- ultrafiltration for wastewater treatment and reuse. Desalination, 131(1–3), 307–314. https://doi.org/10.1016/S0011-9164(00)90029-8 Al-Juboori, R. A., Uzkurt Kaljunen, J., Righetto, I., & Mikola, A. (2022). Membrane contactor onsite piloting for nutrient recovery from mesophilic digester reject water: The effect of process conditions and pre-treatment options. Separation and Purification Technology, 303(September), 122250. https://doi.org/10.1016/j.seppur.2022.122250 APHA. (2012). Standard Methods for the Examination of Water and Wastewater. 1496. APHA Method 2320. (1992). Zigurat Roche.Pdf. American Public Health Association, 552, 2–25 to 2–28. https://law.resource.org/pub/us/cfr/ibr/002/apha.method.2320.1992.pdf Apostol, G., Kouachi, R., & Constantinescu, I. (2011). Optimization of coagulationflocculation process with aluminum sulfate based on response surface methodology. UPB Scientific Bulletin, Series B: Chemistry and Materials Science, 73(2), 77–84. Aygun, a, & Yilmaz, T. (2010). Improvement of Coagulation-Flocculation Process for Treatment of Detergent Wastewaters Using Coagulant Aids. International Journal of Chemical and Environmental Engineering, 1(2), 97–101. Beckinghausen, A., Odlare, M., Thorin, E., & Schwede, S. (2020). From removal to recovery: An evaluation of nitrogen recovery techniques from wastewater. Applied Energy, 263(February), 114616. https://doi.org/10.1016/j.apenergy.2020.114616 Bouzas, A., Martí, N., Grau, S., Barat, R., Mangin, D., & Pastor, L. (2019). Implementation of a global P-recovery system in urban wastewater treatment plants. Journal of Cleaner Production, 227, 130–140. https://doi.org/10.1016/j.jclepro.2019.04.126 Darestani, M., Haigh, V., Couperthwaite, S. J., Millar, G. J., & Nghiem, L. D. (2017). Hollow fibre membrane contactors for ammonia recovery: Current status and future developments. Journal of Environmental Chemical Engineering, 5(2), 1349–1359. https://doi.org/10.1016/j.jece.2017.02.016 Das, S., Gaustad, G., Sekar, A., & Williams, E. (2018). Techno-economic analysis of supercritical extraction of rare earth elements from coal ash. Journal of Cleaner Production, 189, 539–551. https://doi.org/10.1016/j.jclepro.2018.03.252 86 Devesa-Rey, R., Bustos, G., Cruz, J. M., & Moldes, A. B. (2012). Evaluation of nonconventional coagulants to remove turbidity from water. Water, Air, and Soil Pollution, 223(2), 591–598. https://doi.org/10.1007/s11270-011-0884-8 Dosta, J., Rovira, J., Galí, A., Macé, S., & Mata-Álvarez, J. (2008). Integration of a Coagulation/Flocculation step in a biological sequencing batch reactor for COD and nitrogen removal of supernatant of anaerobically digested piggery wastewater. Bioresource Technology, 99(13), 5722–5730. https://doi.org/10.1016/j.biortech.2007.10.021 Duan, J., & Gregory, J. (2003). Coagulation by hydrolysing metal salts. Advances in Colloid and Interface Science, 100–102(SUPPL.), 475–502. https://doi.org/10.1016/S0001-8686(02)00067-2 Dube, P. J., Vanotti, M. B., Szogi, A. A., & García-González, M. C. (2016). Enhancing recovery of ammonia from swine manure anaerobic digester effluent using gaspermeable membrane technology. Waste Management, 49, 372–377. https://doi.org/10.1016/j.wasman.2015.12.011 Eskicioglu, C., Galvagno, G., & Cimon, C. (2018). Approaches and processes for ammonia removal from side-streams of municipal effluent treatment plants. Bioresource Technology, 268(May), 797–810. https://doi.org/10.1016/j.biortech.2018.07.020 Eurostat. (2021). Electricity price statistics. Ferreira, S. L. C., Bruns, R. E., Ferreira, H. S., Matos, G. D., David, J. M., Brandão, G. C., da Silva, E. G. P., Portugal, L. A., dos Reis, P. S., Souza, A. S., & dos Santos, W. N. L. (2007). Box-Behnken design: An alternative for the optimization of analytical methods. Analytica Chimica Acta, 597(2), 179–186. https://doi.org/10.1016/j.aca.2007.07.011 Foladori, P., Andreottola, G., & Ziglio, G. (2015). Sludge Reduction Technologies in Wastewater Treatment Plants. In Sludge Reduction Technologies in Wastewater Treatment Plants. IWA Publishing. https://doi.org/10.2166/9781780401706 Fragoso, R. A., Duarte, E. A., & Paiva, J. (2015). Contribution of Coagulation- Flocculation Process for a More Sustainable Pig Slurry Management. Water, Air, and Soil Pollution, 226(5), 4–9. https://doi.org/10.1007/s11270-015-2388-4 García-González, M. C., Vanotti, M. B., & Szogi, A. A. (2015). Recovery of ammonia from swine manure using gas-permeable membranes: Effect of aeration. Journal of 87 Environmental Management, 152, 19–26. https://doi.org/10.1016/j.jenvman.2015.01.013 González Montiel, J. M. (2008). Acetato de etilo en la Industria. Journal of Chemical Information and Modeling, 53(9), 287. Guimarães, N. R., Dörr, F., Marques, R. de O., Pinto, E., & Ferreira Filho, S. S. (2020). Removal efficiency of dissolved organic matter from secondary effluent by coagulation-flocculation processes. Journal of Environmental Science and Health - Part A Toxic/Hazardous Substances and Environmental Engineering, 56(2), 161– 170. https://doi.org/10.1080/10934529.2020.1856580 Hasanoĝlu, A., Romero, J., Pérez, B., & Plaza, A. (2010). Ammonia removal from wastewater streams through membrane contactors: Experimental and theoretical analysis of operation parameters and configuration. Chemical Engineering Journal, 160(2), 530–537. https://doi.org/10.1016/j.cej.2010.03.064 Hermassi, M., Valderrama, C., Gibert, O., Moreno, N., Querol, X., Batis, N. H., & Cortina, J. L. (2017). Recovery of nutrients (N-P-K) from potassium-rich sludge anaerobic digestion side-streams by integration of a hybrid sorption-membrane ultrafiltration process: Use of powder reactive sorbents as nutrient carriers. Science of the Total Environment, 599–600, 422–430. https://doi.org/10.1016/j.scitotenv.2017.04.140 Horstmeyer, N., Weißbach, M., Koch, K., & Drewes, J. E. (2018). A novel concept to integrate energy recovery into potable water reuse treatment schemes. Journal of Water Reuse and Desalination, 8(4), 455–467. https://doi.org/10.2166/wrd.2017.051 Hu, H., Ding, L., Geng, J., Huang, H., Xu, K., & Ren, H. (2016). Effect of coagulation on dissolved organic nitrogen (DON) bioavailability in municipal wastewater effluents. Journal of Environmental Chemical Engineering, 4(2), 2536–2544. https://doi.org/10.1016/j.jece.2016.04.036 Huang, A. K., Veit, M. T., Juchen, P. T., Gonçalves, G. D. C., Palácio, S. M., & Cardoso, C. D. O. (2019). Sequential process of coagulation/flocculation/sedimentationadsorption - Microfiltration for laundry effluent treatment. Journal of Environmental Chemical Engineering, 7(4), 103226. https://doi.org/10.1016/j.jece.2019.103226 Jiang, J. Q. (2015a). The role of coagulation in water treatment. Current Opinion in Chemical Engineering, 8, 36–44. https://doi.org/10.1016/j.coche.2015.01.008 88 Jiang, J. Q. (2015b). The role of coagulation in water treatment. Current Opinion in Chemical Engineering, 8, 36–44. https://doi.org/10.1016/j.coche.2015.01.008 Kan, C., Huang, C., & Pan, J. R. (2002). Time requirement for rapid-mixing in coagulation. Colloids and Surfaces A: Physicochemical and Engineering Aspects, 203(1–3), 1–9. https://doi.org/10.1016/S0927-7757(01)01095-0 Krupińska, I. (2020). The effect of the type of hydrolysis of aluminum coagulants on the effectiveness of organic substances removal from water. Desalination and Water Treatment, 186(September 2019), 171–180. https://doi.org/10.5004/dwt.2020.25248 Kurniawan, T. A., Lo, W. H., & Chan, G. Y. S. (2006). Physico-chemical treatments for removal of recalcitrant contaminants from landfill leachate. Journal of Hazardous Materials, 129(1–3), 80–100. https://doi.org/10.1016/j.jhazmat.2005.08.010 Lee, W., An, S., & Choi, Y. (2021). Ammonia harvesting via membrane gas extraction at moderately alkaline pH: A step toward net-profitable nitrogen recovery from domestic wastewater. Chemical Engineering Journal, 405(May 2020), 126662. https://doi.org/10.1016/j.cej.2020.126662 Leiknes, T. O. (2009). The effect of coupling coagulation and flocculation with membrane filtration in water treatment: A review. Journal of Environmental Sciences, 21(1), 8–12. https://doi.org/10.1016/S1001-0742(09)60003-6 Licon Bernal, E. E., Maya, C., Valderrama, C., & Cortina, J. L. (2016). Valorization of ammonia concentrates from treated urban wastewater using liquid-liquid membrane contactors. Chemical Engineering Journal, 302, 641–649. https://doi.org/10.1016/j.cej.2016.05.094 Liu, H., & Wang, J. (2016). Separation of ammonia from radioactive wastewater by hydrophobic membrane contactor. Progress in Nuclear Energy, 86, 97–102. https://doi.org/10.1016/j.pnucene.2015.10.011 Matthews, J. A. (2014). Chemical Oxygen Demand (Cod). Encyclopedia of Environmental Change, 5000, 14–19. https://doi.org/10.4135/9781446247501.n623 Ministerio de Agricultura Pesca y Alimentación. (2022). Índices y Precios Pagados Agrarios. Noriega-Hevia, G., Serralta, J., Borrás, L., Seco, A., & Ferrer, J. (2020). Nitrogen recovery using a membrane contactor: Modelling nitrogen and pH evolution. Journal of Environmental Chemical Engineering, 8(4), 103880. 89 https://doi.org/10.1016/j.jece.2020.103880 Noriega-Hevia, G., Serralta, J., Seco, A., & Ferrer, J. (2021). Economic analysis of the scale-up and implantation of a hollow fibre membrane contactor plant for nitrogen recovery in a full-scale wastewater treatment plant. Separation and Purification Technology, 275(June), 119128. https://doi.org/10.1016/j.seppur.2021.119128 Panos, E., & Densing, M. (2019). The future developments of the electricity prices in view of the implementation of the Paris Agreements: Will the current trends prevail, or a reversal is ahead? Energy Economics, 84. https://doi.org/10.1016/j.eneco.2019.104476 Postolachi, L., Rusu, V., & Lupascu, T. (2016). Effect of aluminium sulphate aging on coagulation process for the Prut river water treatment. Chemistry Journal of Moldova, 11(1), 27–32. https://doi.org/10.19261/cjm.2016.11(1).03 Preston, K., Lantagne, D., Kotlarz, N., & Jellison, K. (2010). Turbidity and chlorine demand reduction using alum and moringa flocculation before household chlorination in developing countries. Journal of Water and Health, 8(1), 60–70. https://doi.org/10.2166/wh.2009.210 Puyol, D., Batstone, D. J., Hülsen, T., Astals, S., Peces, M., & Krömer, J. O. (2017). Resource recovery from wastewater by biological technologies: Opportunities, challenges, and prospects. Frontiers in Microbiology, 7(JAN), 1–23. https://doi.org/10.3389/fmicb.2016.02106 Raghu, S., & Ahmed Basha, C. (2007). Chemical or electrochemical techniques, followed by ion exchange, for recycle of textile dye wastewater. Journal of Hazardous Materials, 149(2), 324–330. https://doi.org/10.1016/j.jhazmat.2007.03.087 Razon, L. F. (2018). Reactive nitrogen: A perspective on its global impact and prospects for its sustainable production. Sustainable Production and Consumption, 15, 35–48. https://doi.org/10.1016/j.spc.2018.04.003 Reig, M., Vecino, X., Gibert, O., Valderrama, C., & Cortina, J. L. (2021). Study of the operational parameters in the hollow fibre liquid-liquid membrane contactors process for ammonia valorisation as liquid fertiliser. Separation and Purification Technology, 255(July 2020), 117768. https://doi.org/10.1016/j.seppur.2020.117768 Richter, L., Wichern, M., Grömping, M., Robecke, U., & Haberkamp, J. (2020). Ammonium recovery from process water of digested sludge dewatering by membrane contactors. Water Practice and Technology, 15(1), 84–91. 96 g·L-1), (iii) feed volume (60 and 5 L) and (iv) feed temperature (25 and 35 °C) to evaluate the effects of them on the PP HF-LLMC efficiency of ammonium recovery from wastewater streams. In fact, to the best of our knowledge, some of these parameters have not been thoroughly studied previously, such as the feed volume effect on the PP HFLLMC performance. 4.3. Materials and Methods 4.3.1. Reagents Nitric acid (65%, HNO3) was used as the acid stripping solution for ammonium salts production. Additionally, methanesulfonic acid (CH3SO3H, 99%), sodium hydrogen carbonate (NaHCO3, 99%) and anhydrous sodium carbonate (Na2CO3, 99%) were used for the ionic chromatography analysis. All chemicals used in this work were analytical grade reagents and were supplied by Sigma-Aldrich®, Spain. 4.3.2. Wastewater Solution Pretreated urban wastewater from the Vilanova i la Geltrú wastewater treatment plant (WWTP) in Barcelona, Spain, was utilized in this study. The wastewater subsequent treatment in a pilot plant situated within the same facility to decrease ammonium levels using zeolites (Reig et al., 2021; Sancho et al., 2017; Vecino et al., 2019). However, a notable drawback of this final treatment was the generation of a more concentrated stream with high ammonia content, resulting from the elevated pH (≈12). Consequently, this treated stream was employed as the feed solution for the present study. 4.3.3. Experimental Set-Up Although the experimental setup has been described in other studies (Reig et al., 2021; Vecino et al., 2019, 2020), the most relevant details about the lab-scale LLMC experimental set-up are described in this section. A pair of Liqui-Cel® 2.5 × 8 Series Membrane Contactor X-50 polypropylene provided by 3MTM, USA LLMC modules were used located in a series and the vertical position. The above mentioned modules and their characteristics (e.g., membrane configuration, active area, hydrophobicity, pore diameter or the number of fibers) have also been described previously (Licon Bernal et al., 2016), the configuration was hollow fibers with a membrane area of 1.4 m2. Two tanks were used to introduce the feed solution (60 L) and the acid stripping solution (0.5 L) of 0.4 M of HNO3, connected to the LLMC modules through PVC flexible tubes. All tests were 97 conducted under contra-current mode in a closed-loop (i.e., recirculating both streams and introducing the feed solution through the shell side and the acid stripping one through the lumen (inside the fibers). Furthermore, it should be noted that both flow rates were kept constant following the optimal results previously determined: 450 mL/min for the feed stream and 700 mL/min for the acid stripping side. In summary, the hydrophobic LLMCs used in this work only allowed the passage of ammonia gas (pH > pKa(NH4+/NH3) = 9.25) from the feed solution through the hollow fibers. Then, the NH3 reacts with the acid stripping solution, producing ammonium salts, which can be used in agriculture as liquid fertilizers. In this case, HNO3 was used as an acid stripping solution, keeping the stripping solution between pH 2 and 3 (Licon Bernal et al., 2016) by a concentrated HNO3 (14M) addition. For that, 0.5 L of 0.4 M HNO3 was prepared as the initial acid stripping solution, which, after the trials, was converted into ammonium nitrate salts, as follows in Equation 1. NH3(g)+ HNO3↔ NH4NO3 (Eq.1) Moreover, to enhance ammonia recovery and fertilizer concentrations, the valorization process was conducted in two steps. Once the feed concentration reached a plateau, indicating no further ammonia transport to the stripping side, the experiment was halted, and either the acid or the feed solution was replaced with a fresh one. Previously, as per Reig et al. 2021, only the acid stripping solution was changed between steps to further decrease the ammonia concentration of the feed solution, albeit without increasing the ammonium salt concentration. However, in this study, changing the feed solution was also considered to increase fertilizer concentrations and determine which option would optimize overall performance. Furthermore, not only the replacing of the acid stripping or the feed solution between both stages was studied, but also other variables, such as the initial ammonia concentration in the feed solution (mainstream or side, around 1 or 4.5 g NH3·L-1, respectively), feed volume to be treated (60 or 5 L) and feed temperature (around 25 °C (room) or 35 °C (maximum allowed acoording to manufacturer (3MTM LLMC)) (Reig et al., 2021). During the experiments, 112 samples were collected from both tanks over time. Therefore, these samples were analyzed to determine their compositions (mainly the concentration of ammonia in the feed solution and the concentration of ammonium and 98 nitrate in the acid stripping solution). All experiments were replicated twice, and duplicate samples were taken at each experimental point, in order to report data with higher precision and confidence. Thus, data were reported as the mean ± standard deviation of replicate determinations. 4.3.4. Experimental Design Five experiments were designed to study the effects of the abovementioned parameters. Table 1 summarizes the experimental design, where one parameter was varied trial after trial. Table 1. Experimental Design of the Five Conducted Experiments, detailing which tank is changed if acid or feed tank before the plateau. Initial theoric Concentrations, Feed Volume, and Feed Temperature. Exp. 1 Exp. 2 Exp. 3 Exp. 4 Exp. 5 Tank change after plateau acid feed acid acid acid Feed [NH3] (g·L-1) 4.5 4.5 1 1 1 Volume (L) 60 60 60 5 5 Temperature (°C) 25 25 25 25 35 As can be seen in Table 1, the initial experiments (1 and 2) were designed to study the effect of changing the acid or feed solution after the plateau stage. In this case, 60 L of sidestream wastewater at room temperature were used (4.5-5 g NH3·L-1) following the already published conditions (Reig et al., 2021; Vecino et al., 2019). Once the best option was found, the next experiment (Exp. 3) was conducted to study the effect of the initial ammonia concentration by using mainstream wastewater (≈ 1 g NH3·L-1), keeping the other parameters as in the first experiments. Next, the feed volume was varied from 60 to 5 L (Exp. 4) to determine its effect on the overall performance. Finally, the feed temperature was studied by increasing the feed solution temperature up to 35 °C (Exp. 5). 99 4.3.5. Data Analysis Several main (measured and calculated) parameters were determined to analyze the LLMC efficiency, depending on the analyzed parameters, to valorize ammonia from wastewater and recover it as liquid fertilizers: two for the feed side and two more for the acid stripping side. Thus, the final ammonia concentration and the ammonia recovery were the analyzed parameters for the feed side. The former was directly analyzed by analytical methodologies, and the latter was calculated by Equation 2.: where Cfeed,0 and Cfeed,final are the initial and final ammonia concentrations (mg·L-1), respectively, in the feed tank (Vecino et al., 2019). On the other hand, the ammonia concentration factor (CF) and the final nitrogen concentration in the stripping side were also determined. Thus, the CF was calculated following Equation 3.: where C0(NH3,feed tank) and Cf(NH3,acid tank) are the initial and final NH3 concentrations (mg·L- 1) in the feed and acid stripping tanks, respectively (Vecino et al., 2019). The ammonium salt composition was expressed by the percentage of N-NH4 present in the liquid fertilizer solution, as described by Equation 4.: where Cacid stripping,final is the final ammonium concentration in the acid stripping tank (g N-NH4/g solution (%, w/w)) (Vecino et al., 2019). The %N in the liquid fertilizer is a common parameter that fertilizer companies consider when describing the composition of their liquid fertilizers instead of ammonium salt amount or concentration. Thus, this parameter was used to determine the composition of the obtained ammonium salts (fertilizer). Finally, analysis of variance (ANOVA) is a statistical test that evaluates the hypothesis that the means of two or more populations are equal. Therefore, the null hypothesis Ammonia recovery (%)=Cfeed,0−Cfeed,final Cfeed,0 ·100 (Eq.2) CF (−)= Cf(NH3,acid tank) C0(NH3,feed tank) (Eq.3) N-NH4 concentration (%, w/w) = Cacid stripping,final (Eq.4) 100 establishes that all the analyzed measures are identical, while the alternative hypothesis confirms that at least one value is different from the rest. The result of this test provides the statistical significance value, p; if this value is lower than the established significance level of 0.05 (<95%), it is concluded that at least one mean of the analyzed values is different from the rest of the values. (Barros et al., 2022). 4.3.5.1. Analytical Methodology During the experiments, the pH was monitored and measured online by a GLP 22 pH meter (Crison®, Spain) which has a pH measuring range of 0-14, and the conductivity was measured by an EC-Metro GLP 31 (Crison®, Spain) (Vecino et al., 2019). The total carbon (TC) was determined by a TOC-VCPH meter (Shimadzu®, Japan). Moreover, the sample compositions were determined by ionic chromatography (IC). In this case, two apparatuses from Thermo-Fisher Scientific, USA were used for cation and anion quantifications: (i) Dionex ICS-1000 equipped with a CS16 column (5 × 250 mm), a pre-column CG16 (5 × 50 mm) and cationic detector ICS-1000 and (ii) Dionex ICS- 1100 equipped with a AS23 column (4 × 250 mm), pre-column AG23 (4 × 50 mm) and an anionic detector ICS-1100. Thus, 0.03 mol·L-1 of the CH3SO3H solution was used as the mobile phase for the cations equipment and a mixture of 0.8 mmol·L-1 of NaHCO3 and 4.5 mmol·L-1 of Na2CO3 for the anions system. Both devices were controlled by Chromeleon® chromatographic software. 4.4. Results and Discussion First of all, pretreated wastewater from the WWTP was analyzed by ionic chromatography to determine the ions concentration and other parameters, such as pH or conductivity (Table 2). Parameter Value Units Ammonium (NH4+) 4.60 ± 0.14 mg·L-1 Calcium (Ca2+) 0.04 ± 0.02 Chloride (Cl−) 0.35 ± 0.14 Magnesium (Mg2+) 0.03 ± 0.01 101 Table 2. Initial sidestream wastewater composition As can be seen in Table 2, the sidestream wastewater used as the feed solution in this work was mainly composed of sodium and ammonium ions mixed with dissolved organic matter at a high pH, in this scenario, N is predominantly present in the form of NH3(>pKa = 9.25). Then, ammonium was present as ammonia in gas form. Moreover, other elements were found, such as potassium, chloride, nitrate or sulphate, but at trace levels. On the other hand, apart from the sidestream wastewater, mainstream wastewater was also used in this work. In this case, the major difference between wastewaters was the ammonium concentration, being approximately 1 g·L-1 for the mainstream. After each experiment, the remaining feed solution and, also, the ammonium salt produced were both analyzed to corroborate that only ammonium passed through the LLMC but not other elements (data not shown). 4.4.1. Effect of Changing Feed or Acid Stripping Solution between LLMC Process Steps to Increase Ammonia Recovery As previously mentioned, two scenarios were studied: (i) changing the acid solution between steps and (ii) replacing the feed solution with a new one between steps. In the first case, the idea was to further decrease the final ammonia concentration of the feed solution while obtaining two liquid fertilizer solutions of a similar concentration. In the second scenario, the aim was to achieve a more concentrated liquid fertilized in the acid stripping side, although not able to decrease the ammonia concentration of the feed solution so much. Figure 1 shows the ammonia evolution in the feed solution and the Nitrate (NO3−) 0.33 ± 0.11 Phosphate (PO43−) 0.05 ± 0.02 Potassium (K+) 0.46 ± 0.05 Sodium (Na+) 12.70 ± 0.01 Sulphate (SO42−) 0.38 ± 0.11 Conductivity (25ºC) 66.30 ± 0.99 mS/cm pH (25ºC) 13.13±0.24 - Total carbon (C) 57.93 ± 0.87 mg·L-1 102 nitrogen concentration evolution in the ammonium salts solution over time for both scenarios. Figure 1. Ammonia concentration evolution over time in the feed tank when changing (a) the acid and (b) feed between steps (up). Nitrogen concentration achieved in the liquid fertilizer by changing the (c) acid or (d) feed solution between steps (down). Orange color implies one stage of LLMC (triangle for the feed side and circle referring to the fertilizer solution), while green color refers to experiments with two LLMC stages (triangle for the feed side and circle referring to the fertilizer solution). The errors bars represent standard deviation (n=4). As can be seen in Figure 1, the ammonia concentration in the feed solution decreased over time from around 4.5 to 1.6 g NH3·L-1 during the first step (Figure 1 a,b). Afterwards, replacing the fertilizer for a new acid solution (0.4 M HNO3) in the stripping side, it was (A) (B) (C) (D) 0 1 2 3 4 5 010 20 30 NH3concentration (g·L-1) Time (h) 0 1 2 3 4 5 010 20 30 NH3concentration (g·L-1) Time (h) 0 1 2 3 4 5 6 7 010 20 30 N-NH4concentration (%) Time (h) 0 1 2 3 4 5 6 7 010 20 30 N-NH4concentration (%) Time (h) 103 possible to further decrease the ammonia concentration down to 0.7 g·L-1 (Figure 1a), whereas a similar behaviour to that observed in step 1 was achieved when changing the feed solution for a new one with approximately 4.5 g NH3·L-1, reducing its concentration to 1.6 g NH3·L-1 (Figure 1b). On the other hand, comparing the nitrogen concentration evolution in the stripping side (Figure 1c,d), it can be seen that two ammonium salt solutions (around 5.4% N-NH4 and 3.9% N-NH4) were produced when changing the acid between steps (Figure 1c), whereas just one liquid fertilizer was produced when changing the feed solution, although its concentration was almost not even increased (from around 5.5 to 5.6% N-NH4) (Figure 1d). Additionally, the ammonia recovery was calculated after each step and was also determined for the overall process, taking into account both steps (Figure 2). Figure 2. Ammonia recovery after each step, and the global results changing the acid or the feed solution between steps. The errors bars represent standard deviation (n=4). Figure 2 illustrates, the first step for both experiments had the same performance, achieving ammonia recovery values of around 64.4%. The results for the second step demonstrated that the NH3 recovery obtained was better when changing the feed solution after reaching a plateau (62.2±1.3% vs. 54.7±3.2% N-NH4) since the initial ammonia concentration was again the same as in the beginning, so more ammonia ions could react with the remaining nitric acid of the stripping side. Nevertheless, the overall and 0 10 20 30 40 50 60 70 80 90 100 Step 1 Step 2 Global NH3recovery (%) ACID FEED 104 maximum ammonia recovery achieved was higher (≈83.9±0.7%) when changing the acid stripping solution between steps. The difference between both cases can be attributed to the fact that during operation, the saline concentration on the stripping side is practically zero or very low, which eliminates the effect of osmotic distillation. However, as the saline concentration on the stripping side increases, this phenomenon begins to occur, leading to a greater water passage and reducing the concentration capacity on the acid side. By changing the stripping solution, the saline concentration essentially returns to nearly zero, allowing for greater selectivity towards ammonia and therefore higher concentration capacity. On the other hand, when replacing the feed tank with a new one, a gradient in saline concentration is also generated, but it is lesser. This is why the ammonia concentration in the feed decreases, but it does not increase on the acid side. This is because the acidic solution already contains a high presence of salts from the previous feed tank that were transported, resulting in higher water transport due to osmotic distillation, preventing higher concentration on that side. Few papers can be found in the literature studying the effects of working with different steps of LLMCs. Indeed, preliminary experiments by two-stage LLMC changing the acid stream were previously done by our research group (Reig et al., 2021). The results demonstrated that similar results were obtained by one or two steps. For this reason, in the present work, the feed solution was changed between steps to try to enhance the overall performance. However, changing the acid solution between steps was selected as optimal regarding the results. In fact, it allowed to obtain a feed solution with less ammonia concentration and two liquid fertilizer solutions with a similar nitrogen concentration, and also, a higher ammonia recovery (around 25% more) could be reached. Furthermore, Zhang et al. 2021, proposed a three-stage LLMC process, changing the acid stripping solution by passing the feed stream through the three LLMC in the series. The main objective was to recover ammonia from human urine as ammonium nitrogen (J. Zhang et al., 2021). The results are in agreement with the one found in this article Zhang et al. 2021, since the average ammonia removal percentage was much higher as a global value (over 99%), than taking into account LLMC by LLMC (between 80 and 83%). Additionally, Yan et al. 2018, studied a four-stage LLMC system, recirculating both streams, feed and acid between steps. Again, the results demonstrated that the ammonia recovery could be increased by including more LLMC stages, being able to enhance the recovery value from 65 up to >98% (H. Yan et al., 2018). 105 4.4.2. Initial Ammonia Concentration Effect on the Overall LLMC Performance Sidestream wastewater (4.5-5 g NH3·L-1) and mainstream wastewater (≈1 g NH3·L-1) were used as the feed solution in the LLMC. The results indicated that both wastewaters could be treated by LLMC, although several parameters were determined to establish the optimal performance. Figure 3 shows the ammonia concentration evolution over time and its recovery on the feed tank (up) and, also, the concentration factor and the nitrogen concentration achieved in the ammonium salt solution (down). Figure 3. Comparison between working with the sidestream (high NH3 concentration) and mainstream wastewater (low NH3 concentration): (a) ammonia concentration evolution in the feed tank, (b) ammonia recovery, (c) concentration factor and (d) %N-NH4 concentration in the liquid fertilizer. High ammonia concentration is indicated by the color orange, while the color yellow implies working at low ammonia concentrations. The errors bars represent standard deviation (n=4). (A) (B) (C) (D) 0 1 2 3 4 5 010 20 30 NH3concentration (g·L-1) Time (h) 0 20 40 60 80 100 010 20 30 NH3recovery (%) Time (h) 0 10 20 30 40 50 Step 1 Step 2 Concentration factor (−) 0 1 2 3 4 5 6 Step 1 Step 2 N-NH4concentration (%) 112 nuances of the observed results. Finally, the differences in temperature tested were not enough to have significant improvements in the LLMC performance. All in all, LLMC proved to be a versatile technique to treat wastewater with low (1 g·L-1) and high (4.5 g·L-1) initial ammonia concentrations, where the most influential parameter was the change of the stripping solution in the formulation of ammonium salts. Therefore, LLMC could be an easy technique to be implemented in biofactories for the recovery of nutrients from main or side wastewater streams. 113 4.6. References Ahn, Y. T., Hwang, Y. H., & Shin, H. S. (2011). Application of PTFE membrane for ammonia removal in a membrane contactor. Water Science and Technology, 63(12), 2944–2948. https://doi.org/10.2166/wst.2011.141 Ashrafizadeh, S. N., & Khorasani, Z. (2010). Ammonia removal from aqueous solutions using hollow-fiber membrane contactors. Chemical Engineering Journal, 162(1), 242–249. https://doi.org/10.1016/j.cej.2010.05.036 Barros, A., Vecino, X., Reig, M., & Cortina, J. L. (2022). Coagulation and Flocculation Optimization Process Applied to the Sidestream of an Urban Wastewater Treatment Plant. Water (Switzerland), 14(24), 1–26. https://doi.org/10.3390/w14244024 Darestani, M., Haigh, V., Couperthwaite, S. J., Millar, G. J., & Nghiem, L. D. (2017). Hollow fibre membrane contactors for ammonia recovery: Current status and future developments. Journal of Environmental Chemical Engineering, 5(2), 1349–1359. https://doi.org/10.1016/j.jece.2017.02.016 Guidelines, S. (n.d.). 3M TM Liqui-Cel TM EXF Series Membrane Contactors. Kartohardjono, S., Damaiati, G. M., & Rama, C. T. (2015). Effects of absorbents on ammonia removal from wastewater through hollow fiber membrane contactor. Journal of Environmental Science and Technology, 8(5), 225–231. https://doi.org/10.3923/jest.2015.225.231 Kartohardjono, S., Fermi, M. I., Yuliusman, Elkardiana, K., Sangaji, A. P., & Ramadhan, A. M. (2015). The removal of dissolved ammonia from wastewater through a polypropylene hollow fiber membrane contactor. International Journal of Technology, 6(7), 1146–1152. https://doi.org/10.14716/ijtech.v6i7.1845 Li, K., Kong, J., & Tan, X. (2000). Design of hollow fibre membrane modules for soluble gas removal. Chemical Engineering Science, 55(23), 5579–5588. https://doi.org/10.1016/S0009-2509(00)00193-7 Licon Bernal, E. E., Maya, C., Valderrama, C., & Cortina, J. L. (2016). Valorization of ammonia concentrates from treated urban wastewater using liquid-liquid membrane contactors. Chemical Engineering Journal, 302, 641–649. https://doi.org/10.1016/j.cej.2016.05.094 114 Majd, A. M. S., & Mukhtar, S. (2013). AND FIELD-SCALE STUDIES. 56(Lm), 1951– 1958. Moradihamedani, P. (2021). Recent developments in membrane technology for the elimination of ammonia from wastewater: A review. Polymer Bulletin, 78(9), 5399– 5425. https://doi.org/10.1007/s00289-020-03386-y Naim, R., Ismail, A. F., & Mansourizadeh, A. (2012). Effect of non-solvent additives on the structure and performance of PVDF hollow fiber membrane contactor for CO 2 stripping. Journal of Membrane Science, 423–424, 503–513. https://doi.org/10.1016/j.memsci.2012.08.052 Reig, M., Vecino, X., Gibert, O., Valderrama, C., & Cortina, J. L. (2021). Study of the operational parameters in the hollow fibre liquid-liquid membrane contactors process for ammonia valorisation as liquid fertiliser. Separation and Purification Technology, 255(July 2020), 117768. https://doi.org/10.1016/j.seppur.2020.117768 Samani Majd, A. M., & Mukhtar, S. (2013). Ammonia diffusion and capture into a tubular gas-permeable membrane using diluted acids. Transactions of the ASABE, 56(5), 1943–1950. https://doi.org/10.13031/trans.56.10218 Sheikh, M., Reig, M., Vecino, X., Lopez, J., Rezakazemi, M., Valderrama, C. A., & Cortina, J. L. (2022). Liquid–Liquid membrane contactors incorporating surface skin asymmetric hollow fibres of poly(4-methyl-1-pentene) for ammonium recovery as liquid fertilisers. Separation and Purification Technology, 283, 120212. https://doi.org/10.1016/j.seppur.2021.120212 Uzkurt Kaljunen, J., Al-Juboori, R. A., Mikola, A., Righetto, I., & Konola, I. (2021). Newly developed membrane contactor-based N and P recovery process: Pilot-scale field experiments and cost analysis. Journal of Cleaner Production, 281, 125288. https://doi.org/10.1016/j.jclepro.2020.125288 Vecino, X., Reig, M., Bhushan, B., Gibert, O., Valderrama, C., & Cortina, J. L. (2019). Liquid fertilizer production by ammonia recovery from treated ammonia-rich regenerated streams using liquid-liquid membrane contactors. Chemical Engineering Journal, 360(December 2018), 890–899. https://doi.org/10.1016/j.cej.2018.12.004 115 Vecino, X., Reig, M., Gibert, O., Valderrama, C., & Cortina, J. L. (2020). Integration of liquid-liquid membrane contactors and electrodialysis for ammonium recovery and concentration as a liquid fertilizer. Chemosphere, 245, 125606. https://doi.org/10.1016/j.chemosphere.2019.125606 Yan, H., Wu, L., Wang, Y., Shehzad, M. A., & Xu, T. (2018). Ammonia capture by water splitting and hollow fiber extraction. Chemical Engineering Science, 192, 211–217. https://doi.org/10.1016/j.ces.2018.07.040 Yu, S., Qin, Y., Zhao, Q., Li, M., Yu, H., Kang, G., & Cao, Y. (2021). Nafion-PTFE hollow fiber composite membranes for ammonia removal and recovery using an aqueous-organic membrane contactor. Separation and Purification Technology, 271(April), 118856. https://doi.org/10.1016/j.seppur.2021.118856 Zhang, J., Xie, M., Yang, D., Tong, X., Qu, D., Feng, L., & Zhang, L. (2021). The design of multi-stage open-loop hollow fiber membrane contactor and its application in ammonia capture from hydrolyzed human urine. Water Research, 207(October), 117811. https://doi.org/10.1016/j.watres.2021.117811 Zhu, Z., Hao, Z., Shen, Z., & Chen, J. (2005). Modified modeling of the effect of pH and viscosity on the mass transfer in hydrophobic hollow fiber membrane contactors. Journal of Membrane Science, 250(1–2), 269–276. https://doi.org/10.1016/j.memsci.2004.10.031 116 CHAPTER 5 Ammonium recovery and concentration from synthetic wastewater using a poly(4-methyl-1-pentene) (PMP) liquid-liquid membrane contactor: flux performance and mass transport characterization 5.1. Abstract Hollow fiber membrane contactors are a promising technology for the removal and recovery of ammonia from liquid effluents. However, a better understanding of the process engineering (e.g. mass transport of ammonia over water) and performance optimization is required. In this study, the performance of a hollow fibre liquid-liquid membrane contactor (HF-LLMC), incorporating a new polymer chemistry (i.e., poly (4- methyl-1-pentene (PMP) and an asymmetric fibre structure, for the recovery and concentration of ammonium from synthetic aqueous solutions, was investigated. The influence of the feed and acid flow rates was evaluated experimentally by determining the overall mass transfer coefficient (𝑘𝑜𝑣), the ammonium recovery as a function of time and the acid consumption. In addition, the experimental results were fitted to a mathematical model to determine the membrane permeabilities to ammonia (𝑃𝑁𝐻3) and water (𝑃𝑤) and identify the mass transfer resistance regime. The highest 𝑘𝑜𝑣 values experimentally obtained were in the range of 3·10-3 to 3.51·10-3 m h-1 with corresponding ammonia recovery rates of 94 to 96.2% after 10h, operating at a feed and acid flow rates of 180 L h-1 and 500 L h-1, respectively, which are in the upper range of the HF-LLMC literature. The overall results of this study a much-lower water transport was confirmed indirectly by the concentration factor (CF) values obtained experimentally. The remarkable selectivity of the membrane towards ammonia over water (i.e., 𝑃𝑁𝐻3 = 87 - 180 L m-2 h-1 bar-1 and 𝑃𝑤 = 1.2 -1.4·10-3 L m-2 h-1 bar-1 at NTP conditions) could be attributed to the asymmetrical membrane structure and the polymer chemistry (i.e., PMP). The proven high ammonia selectivity of the HF-LLMC renders it a promising technology for the recovery and concentration of ammonium from urban (e.g. wastewaters) and industrial (e.g. soda ash and fertilizers production) diluted streams. 117 Keywords: Asymmetric membrane; Ammonia recovery and concentration; Ammonia membrane permeability. 5.2. Introduction The growing world population results in increase demand for food, leading to an increased need for fertilizer production (Chagas et al., 2023; W. Lee et al., 2021; Mayor et al., 2023), and consequently an increase in energy consumption due to an almost exclusive dependence on the Haber-Bosch process (HB) for the synthesis of nitrogenous fertilizers (Silva & Quesada, 2010). The HB process consumes approximately 35- 50 MJ kg-1 N (i.e., 1000 m3 of natural gas per ton of N-based fertilizers) which represents approximately 1-2% of the global energy consumption and 50% of the total natural gas used in the chemical industry (Beckinghausen et al., 2020). Approximately, 85% of the worldwide synthesized ammonia is employed in fertilizer production (PotashCorp, 2014) and out of the 160 Mt of fertilisers used annually worldwide, 100 Mt come from the HB process (González Montiel, 2008). It is therefore necessary to identify alternative sources of ammonia in order to reduce the energy impact of fertilizer production (Razon, 2018). On the other hand, nitrogenous fertilizers (N-fertilizers) have very low application efficiencies and only 16% of the applied quantity is converted into proteins usefull por animals/humans while the rest is lost into water bodies and the atmosphere aggravating phenomena such as eutrophication (Deng et al., 2021). About 10 - 40% of the lost N- fertilizers ends up denitrified (i.e., released as N2) by the vegetation or in wastewater treatment plants (WWTPs) (Matassa et al., 2015; Mayor et al., 2023). Nitrogen is traditionally removed in WWTPs during the secondary treatment using biological nutrient removal methods, and although the annamox technology has gained much interest recently (Beckinghausen et al., 2020; Vineyard et al., 2020), the nitrificationdenitrification (N/DN) process remains the most widely. The N/DN is an energy intensive process (i.e., 45 MJ per kg-1 N eliminated) (Ledezma et al., 2015) whose aeration requirements represent 50% of the total energy consumption of a WWTP (Nowak, 2003). For instance, WWTPs in the United States account for 4% of the nation's yearly electricity consumption and generate a carbon footprint of 0.9 kg CO2 per m3 of treated water (Xie et al., 2016). In the new circular economy paradigm, there is a growing interest in transforming WWTPs into "bio-factories" and recovering valuable resources such as nutrients and 118 energy. Under this paradigm, anaerobic digestion (AD) and the recovery of ammonium and phosphate from digestates and side-streams will be pivotal in the valorisation of organic matter as biogas and fertilizers (Darestani et al., 2017; Rongwong & Goh, 2020). As a result, new technologies and processes with high selectivity such as ion exchange (IX) (Kurniawan et al., 2006; Wan et al., 2017), its combination with membrane technologies, e.g., ultrafiltration (Hermassi et al., 2017) and membrane contactors (MC) (Hasanoĝlu et al., 2010; Licon Bernal et al., 2016; Reig et al., 2021) among others, are being studied for ammonium recovery from several liquid waste streams (e.g., digestates, landfill leachate, industrial waste streams, etc.). However, IX technology has some limitations including the requirement for a chemicalintensive and time-consuming regeneration process, lower ammonia recovery rates and competence of unwanted ions (Darestani et al., 2017). For example, (Wan et al., 2017) recovered nutrients from liquid supernatant from municipal WWTP excess sludge using natural zeolites (IX) reaching ammonium and phosphate recoveries of 70.5% and 84.3%, respectively after five hours of operation plus another five hours (100% extra time) for the regeneration step. Hermassi et al. 2017, used mixtures of calcium and sodium zeolites for the simultaneous recovery of N-P-K from anaerobic digestion side-streams using a hybrid sorption/filtration system and found an important decrease in the NH4+ ion exchange capacity due to the presence of competing ions (any divalent cations) in real waters (Hermassi et al., 2017). The necessity for significant quantities of zeolite material can also be a challenge in IX. For instance, according to a study by (Eskicioglu et al., 2018a) for a medium-sized WWTP (20,000 m3 d-1) over 8,000 kg of zeolite would be required to treat a side-stream (representing only a 0.5% of the total water input) with an average ammonium concentration of 1.2 g NH4+ L-1. As stated by the authors, the expenses related to transporting, maintaining and storing such a substantial amount of zeolite could be prohibitive unless a nearby source is readily accessible. Among membrane technologies for ammonium recovery, MC stands out for its selectivity. The basis of MC technology is the utilization of a microporous hydrophobic membrane that separate two fluid phases (gas/liquid or liquid/liquid) to promote mass transfer of certain components from one phase to the other while avoiding dispersion (Gabelman & Hwang, 1999; Pabby & Sastre, 2013). MC has been already studied to recover ammonia from wastewater streams in the laboratory (Amaral et al., 2016; García- 119 González et al., 2015; Pauzan et al., 2021; Reig et al., 2021) and even some medium-scale implementation examples have been reported. For example, Richter et al. 2020 studied the feasibility of implementing a Polypropylene (PP) hollow fiber liquid-liquid membrane contactor (HF-LLMC) stage to treat a centrate water stream of 342 m3d-1 and 0.9 g NH4+ L-1 in an urban WWTP plant in Münster (Germany) to produce an ammonium sulphate fertilizer solution. The authors reported a recovery efficiency >95% and a concentration factor (CF) calculated as the final NH4+ permeate concentration divided by the initial NH4+ feed concentration, of 4.1 (i.e., ammonium sulphate solution of 3.69 g NH4+ L-1) (Richter et al., 2020). There are several polymeric MC module configurations, yet the most used ones are hollow fibre modules because of their high packing density and total surface area (Bazhenov et al., 2018). However, these modules are more sensitive to membrane fouling (Darestani et al., 2017). In fact, fouling is reportedly the most challenging drawback of membrane technologies, including MC, and depending on the feed water, pre-treatment might be mandatory. For example, in a previous work (Aguilar-Moreno et al., 2022) we successfully optimized a coagulation/flocculation pre-treatment prior a LLMC module used for the recovery of ammonium from a WWTP centrate stream with moderate fouling potential (i.e., 0.8 g COD L-1). By dosing 30 mg of aluminium sulphate (Al2(SO4)3) per litre of centrate, the COD was reduced by 60% while maintaining a relatively high pH (i.e., pH>10) and increasing the ammonium recovery from 8% to 66%. The combination of the two technologies (IX and MC) can help in this direction since the IX stage not only pre-concentrates the ammonium but also acts as a pre-treatment as ammonium is selectively removed. For example, Sancho et al. 2017 used a LLMC to treat the ammonium-rich regenerated stream from the elution of ammonium-selective zeolites with a 2 g L-1 NaOH solution and produced liquid nitrogenous fertilizers. The authors reported removal rates of 95% and final ammonium nitrate and di-ammonium phosphate fertilizer solutions of concentrations ranging 2–5% w N (Sancho et al., 2017). Likewise, Vecino et al. 2019, used MC to obtain a fertilizer product from the regeneration stream of ammonium-selective zeolites. They reported a maximum ammonia recovery of 94% under the following conditions: initial ammonium concentration of 4 g L-1, pHfeed of 13.5, feed and acid flow rates of 27 L h-1 and two modules in series of 2.8 m2 each (Vecino et al., 2019). 120 HF-LLMC are usually fabricated using polymers of the polyolefin family as they are hydrophobic, cheaper to produce, ecologically more sustainable since they can be recycled and have a relatively low environmental footprint, compared to those containing fluor, mechanically and chemically stable (Moradihamedani, 2021; Twarowska-Schmidt & Wlochowicz, 1997; Zhang et al., 2022). They have been mostly produced from polypropylene (PP), polyvinylidene difluoride (PVDF) and polytetrafluoroethylene (PTFE) (Chabanon et al., 2013; Tan et al., 2006; Yu et al., 2021) and with symmetrical membrane structures. A promising (i.e., high hydrophobic nature and good chemical resistance) new polymer chemistry: poly(4-methyl-1-pentene) (PMP) which is being commercialised for degasification of industrial production processes (i.e., O2(g) in ink production, CO2(g) in water treatment schemes of the microelectronic and hydrogen production) in gas-liquid (GL) applications, has shown lower water transport and therefore higher selectivity than conventional PP modules (Ignatenko et al., 2020; Z. Yang et al., 2010 (Sheikh et al., 2022)). The PMP polymer has methyl side chains on the surface (Markova et al., 2020) that prevent the formation of hydrogen bridges between the water and the membrane surface generating a very hydrophobic surface. The downside of PMP HF-LLMCs is their low fluxes, consequence of their asymmetric membrane structure containing a thick selective layer (Markova et al., 2020). However, several companies such as Mitsui Chemical Inc. (Japan), Dainippon Ink and Chemicals Co. (Japan) and DuPont Water Solutions (USA) are currently commercialising them under the trademarks TPX, Separel® and Ligasep™, respectively. While HF-LLMCs have been proposed and tested as an innovative solution for ammonium removal/recovery from liquid streams, there is a need for optimizing their operation in order to maximize the ammonia flux and recovery and characterizing their technical performance to facilitate their industrial implementation (Noriega-Hevia et al., 2021). Furthermore, symmetrical PP hollow fibres are the most used membranes in MC while there are very few references of the use of asymmetrical PMP hollow fibres. Moreover, the previous studies of this new polymer chemistry incorporated into an asymmetric HF membrane, were devoted to studying the performance of MC originally designed for GL applications in liquid-liquid (LL) applications to recover ammonia as liquid fertilizers. The objective of the present work is to investigate the influence of the feed and acid flow rates combination and the circulation configuration (i.e., counter-current/co-current) on 121 the performance of a commercial PMP HF-LLMC module (Separel®) for recovering ammonia from simulated industrial wastewaters containing low-levels of dissolved organic matter. Moreover, the experimental results were fitted to a mathematical model to study the transport of ammonia and water through the hydrophobic membrane by calculating the corresponding permeabilities and mass transfer resistances. Thus, the optimization of the MC operation towards the maximum ammonia mass transfer coefficient and the identification of the mass transfer bottlenecks (whether in the feed, membrane or permeate sides) with a mathematical tool were the objectives of the present study. 5.3. Materials and methods 5.3.1. Chemicals and analytical techniques Artificial feed solutions were prepared with a 30% NH3(l) solution provided by Sigma- Aldrich®,Spain and deionised water in concentrations ranging 6 to 10 g NH3 L-1 simulating those of regeneration zeolites streams (Sancho et al., 2017; Vecino et al., 2019) and certain industrial side-streams like, for example, those of the Solvay process (Trypuć & Białowicz, 2011). The acid stripping solution was prepared by diluting analytical grade sulphuric acid (H2SO4, 98%) from Sigma-Aldrich®, Spain in deionised water to a concentration of 0.01M and during the experiment, automatically dosed in concentrated form (H2SO4, 98%) to maintain a constant pH. Sulphuric acid was selected based on budget considerations and not on the interest or added value of the final product. Feed and acid samples were analysed by ion chromatography (IC) (Dionex ICS-1000 and ICS-1100 Thermo-Fisher Scientific, USA) equipped with CS16 and AS23 columns for measuring cations and anions, respectively. The mobile phase was 30 mmol L-1 methanesulphonic acid (CH3SO3H) for cations and a mixture of 0.8 mmol L-1 NaHCO3 and 4.5 mmol L-1 Na2CO3 for anions. Analytical grade methanesulfonic acid (CH3SO3H, 99%), sodium hydrogen carbonate (NaHCO3, 99%) and anhydrous sodium carbonate (Na2CO3, 99%) supplied by Sigma-Aldrich (Spain) were used in the analyses. Samples were properly diluted and filtered (0.22 µm) prior to their IC analysis. The pH of the samples was determined using a pH-meter (GLP22 Crison, Spain). 128 129 Figure 3. NH3 concentration (mol L-1) in the feed tank (shell) and percentage of ammonia recovery in the feed tank; (A) Experiment 30-250; (B) Experiment 100-250; (C) Experiment 500- 180; (D) Experiment 100-50 (E) Experiment 180-500-A; (F) Experiment 180-500-B. Points: experimental data; Lines: model simulation. The bars indicate the absolute errors of the IC measurements. It is observed that the model is in good agreement with the experimental results. The errors bars represent standard deviation (n=3). Figure 4. NH4+ and SO42- concentration (mol L-1) in the acid stripping tank (lumen; (A) Experiment 30-250; (B) Experiment 100-250; (C) Experiment 500-180; (D) Experiment 100-50 (E) Experiment 180-500-A; (F) Experiment 180-500-B. Points: experimental data; Lines: model simulation. The bars indicate the absolute errors associated with the NH4+ and SO42- IC measurements. It is observed that the model is in good agreement with the experimental results in all the experiments with the exception of the SO42- of experiment 100-250 (4.B). The errors bars represent standard deviation (n=3). 130 5.3.5. Ammonia transport The gaseous transport of ammonia across the membrane can be described as follows (Equation 1) (Licon Bernal et al., 2016): 𝐽𝑁𝐻3,𝑚 =𝑘𝑁𝐻3,𝑚 𝑅·𝑇 ·(𝑝𝑁𝐻3,𝑓 𝑚−𝑝𝑁𝐻3,𝑎 𝑚) (Eq. 1) where 𝐽𝑁𝐻3,𝑚 is the ammonia flux (mol m-2 s-1), 𝑘𝑁𝐻3,𝑚 is the membrane ammonia mass transfer coefficient (L m-2 s-1), 𝑝𝑁𝐻3,𝑓 𝑚 and 𝑝𝑁𝐻3,𝑎 𝑚 are the ammonia partial pressures (bar) at the membrane surface of the feed and acid sides, respectively, 𝑅 is the ideal gas constant (0.083 bar L mol-1 K-1) and 𝑇 is the temperature (K). The ammonia partial pressures (𝑝𝑁𝐻3,𝑓 𝑚 and 𝑝𝑁𝐻3,𝑎 𝑚) are calculated based on the ammonia concentrations at the membrane surface (i.e., 𝑐𝑁𝐻3,𝑓 𝑚 and 𝑐𝑁𝐻3,𝑎 𝑚) using Henry´s law. The ammonia transport from the feed bulk to the membrane surface and from the membrane surface to the acid bulk can be described using Eq. 2.a and Eq. 2.b, respectively: 𝐽𝑁𝐻3,𝑓 =𝑘𝑁𝐻3,𝑓 ·(𝑐𝑁𝐻3,𝑓 −𝑐𝑁𝐻3,𝑓 𝑚) (Eq. 2.a) 𝐽𝑁𝐻3,𝑎 =𝑘𝑁𝐻3,𝑎 ·(𝑐𝑁𝐻3,𝑎 𝑚−𝑐𝑁𝐻3,𝑎) (Eq. 2.b) where 𝑐𝑁𝐻3 and 𝑐𝑁𝐻3 𝑚 are the ammonia concentrations (mol L-1) in the bulk and at membrane surface and 𝑓 and 𝑎 subscripts refer to the feed and acid sides, respectively. 𝑘𝑁𝐻3,𝑓 and 𝑘𝑁𝐻3,𝑎 (L m-2 s-1) are the ammonia mass transfer coefficients from the bulk to the membrane surface of the feed and acid sides, respectively. Considering that the ammonia transport is conservative (i.e., 𝐽𝑁𝐻3,𝑓 =𝐽𝑁𝐻3,𝑚 =𝐽𝑁𝐻3,𝑎), a general expression for the 𝐽𝑁𝐻3 can be derived (Equation 3): 𝐽𝑁𝐻3=𝑈𝑁𝐻3·(𝑐𝑁𝐻3,𝑓 −𝑐𝑁𝐻3,𝑎) (Eq. 3) where 𝑈𝑁𝐻3 is an overall ammonia mass transfer coefficient (m s-1), which considers the transport resistances of the feed, membrane and acid sides in series and has been defined according to Equation 4: 𝑈𝑁𝐻3=1 1 𝑘𝑁𝐻3,𝑓 +1 𝑘𝑁𝐻3,𝑎 +𝑅·𝑇 𝑘𝑁𝐻3,𝑚 (Eq. 4) 131 Different correlations can be used to calculate the mass transfer coefficients (i.e., 𝑘𝑁𝐻3,𝑓 and 𝑘𝑁𝐻3,𝑎) as a function of the module geometry in the lumen (i.e., acid) and in the shell (i.e., feed). The ones used in this study can be found in (X. Yang et al., 2013) and (Sheikh et al., 2022) and, in general, they are proportional to the velocity in the different channels (shell and lumen). 5.3.6. Experimental determination of the overall ammonia mass transfer coefficient In scientific literature, the overall mass transfer coefficient (𝑘𝑜𝑣 or 𝑘𝑚) is used as a parameter to compare the performance of different membranes and modules (Pauzan et al., 2021; Yu et al., 2021; Zhang et al., 2022). The higher the 𝑘𝑜𝑣, the greater the capacity of the membrane to transfer a species through the membrane and therefore the better the performance. The 𝑘𝑜𝑣 (m3 m-2 s-1) can be calculated experimentally from the total ammonia mass balance in the feed side, as follows (Licon Bernal et al., 2016): ln(𝑐𝑁𝐻3,0 𝑐𝑁𝐻3,𝑡)=𝑘𝑜𝑣 ·𝐴𝑚 𝑉·𝑡 (Eq. 5) where 𝑐𝑁𝐻3,0 and 𝑐𝑁𝐻3,𝑡 are the NH3 concentrations in the feed tank at the beginning of the experiment and at time 𝑡, respectively, 𝐴𝑚 is the membrane area (m2), 𝑉 is the feed solution volume (m3) and 𝑡 is the time of the experiment (h). By representing the experimentally measured values of ln(𝑐𝑁𝐻3,0 𝑐𝑁𝐻3,𝑡 ⁄) versus the time (𝑡) a linear relationship is found which slope is 𝑘𝑜𝑣·𝐴𝑚 𝑉 and from which 𝑘𝑜𝑣 can be calculated. This 𝑘𝑜𝑣, as an overall mass transfer coefficient, should be in principle equivalent to 𝑈𝑁𝐻3 (Equation 3). 5.3.7. Water transport Water vapour can be transported across the membrane, provided that a transmembrane water partial pressure gradient exists according to the following equation: 𝐽𝑊=𝑃𝑊 (𝑝𝑊,𝑓 𝑜·𝑎𝑤,𝑓 − 𝑝𝑊,𝑎 𝑜·𝑎𝑤,𝑎) (Eq. 6) Where 𝐽𝑊 is the water flux (mol m-2 s-1), 𝑃𝑊 is the water permeability (mol m-2 s-1 bar-1), 𝑝𝑊 𝑜 is the pure water vapour pressure (bar) at a given temperature and calculated via Antoine’s equation (Technologist et al., 1968) and 𝑎𝑊 is the dimensionless water activity, 132 calculated using the Norrish equation for NaCl solutions (Barbosa-Cánivas et al., 2007) for the respective sides. Due to the lack of correlations for the NaOH-NH3 and NH4+- H2SO4 systems, it was decided to use Norrish equation. The implemented mathematical model aimed to fit the experimental data by adjusting the water and ammonia permeability values (𝑃𝑊 and 𝑃𝑁𝐻3, respectively) in order to minimize the RMSE between the model values and the experimental data, namely the measured ammonia concentrations. 5.3.8. Permeate side chemical equilibrium In order to maintain and enhance the ammonia transport, the NH3 partial pressure in the stripping side (𝑝𝑁𝐻3,𝑎 𝑚 in Eq. 1) must be kept as low as possible. When the ammonia gas (NH3) dissolves in water, it behaves like a weak base (reaction 1 in Table 4) consuming the protons from the acid side and increasing the pH. For that reason, the pH in the acid side was maintained below 2.5 by dosing concentrated H2SO4 automatically throughout the experiment. Bearing in mind that the second dissociation of the H2SO4 is weak, chemical equilibrium reactions must be considered to model the species formed in the acid side. According to this equilibrium, different species, namely: ammonium bisulphate (NH4HSO4), ammonium sulphate ((NH4)2SO4) or a mixture of both can be found depending on the pH ranges of the stripping solution (Uzkurt Kaljunen et al., 2021). Table 4 collects the equilibrium constants (pKa) values of the chemical reactions involved in the acid stripping side. Considering that the three species in solution are pH dependant, they must be accounted for in order to model properly the chemical equilibrium of the permeate side, and therefore the acid consumption. Table 4. Chemical equilibrium reactions that take place in the acidic stripping solution. Number Reaction pKa (25ºC) (Puigdomenech, 2001) 1 𝑁𝐻3+𝐻+⇄𝑁𝐻4 + 9.24 2 𝐻++𝑆𝑂4 2− ⇄𝐻𝑆𝑂4 − 1.98 3 𝑁𝐻3+𝐻++𝑆𝑂4 2− ⇄𝑁𝐻4𝑆𝑂4 − 10.36 5.4. Results and discussion The results of the experiments, named according to the feed and acid flow rate in L h-1 as: 30-250 ,100-250, 500-180, 100-50, 180-500-A and 180-500-B, were analysed in terms 133 of (i) the ammonia recovery and the CF, (ii) the acid consumption, (iii) the ammonia and water transport, (iv) the possible polarization effects induced by the asymmetrical membrane structure and (v) the experimentally calculated and modelled mass transfer coefficients. Letters A and B in the experiments 180-500-A and 180-500-B refer to the acid flow direction (A: co-current and B: counter-current). 5.4.1. Ammonia recovery and CF The ammonia recovery percentages (measured and simulated) as a function of time for the first four experiments 30-250, 100-250, 500-180 and 100-50 are shown in Figure 5 where dots represent the experimental data and lines the model simulation. Figure 5. Ammonia recovery (%) as a function of time for experiments 30-250, 100-250, 500-180 and 100-50. Points: experimental data; Lines: model simulation. The experimental recovery values obtained for all the experiments reached values of 90% after ~15 h (Figure 3). However, two experiments (500-180 and 100-250) were faster (i.e., ~10-11h versus 15h) in reaching 90% recovery than the other two (30-250 and 100- 50). In experiment 500-180 the feed concentration decreased from 6.26 g NH3 L-1 to 0.24 g NH3 L-1 after 13h, achieving a recovery of 96% and a CF of 2.1. In experiment 100- 250, despite the higher initial concentration (i.e., 10.01 g NH3 L-1) a recovery and a CF of 96% and 2.3, respectively, were achieved after 15h. On the other hand, during 0 20 40 60 80 100 0 5 10 15 20 Ammonia Recovery [%] Time [h] Exp 30-250 Exp 100-250 Exp 500-180 Exp 100-50 Exp 30-250 M Exp 100-250 M Exp 500-180 M Exp 100-50 M 90% recovery 134 experiment 30-250, the initial concentration decreased from 6.90 g NH3 L-1 to 0.46 g NH3 L-1 in 17h, which resulted in a recovery of 93% and a CF of 2.1 while in experiment 100- 50 the concentration decreased from 6.56 g NH3 L-1 to 0.47 g NH3 L-1 in 16h reaching only a 92.5% recovery and a final CF of 2.2. Thus, results showed that the relatively higher feed flow rates of experiments 500-180 and 100-250 led to faster recovery rates, while a higher initial concentration had a negligible influence. Regarding the CF, the values for all the experiments are very similar (2.1-2.2) suggesting that the flow rates have a minimal influence and only the feed/stripping volume ratio (i.e., 2.14) is determinant. The higher initial ammonia concentration of experiment 100-250 did lead to a slightly higher CF (i.e., 2.3). However, these minor differences could be attributed to the lack of water flux measurements during the experiments which contribute to the uncertainties in the overall mass balance. Additionally, the higher dilution factors employed for the samples of experiment 100-250 could have introduced additional uncertainties in the extrapolation of the NH4+ quadratic calibration curve of the IC measurements. In fact, the calculated absolute errors of the IC measurements associated with the NH4+ calibration curves (being the relative mean error in the range of 5.4 - 6.8%) are greater at higher concentrations (see Figures S2-S3 in the appendix). The experimental results were compared to the literature. For instance, Vecino et al. 2019 using a PP HF-LLMC module (1.44 m2) and a feed stream of 4 g NH3 L-1 reached a CF of 26 ± 3.3 after 15h of operation. Although the CF was much higher than the results of the present study, the authors reported an ammonium recovery of only a 76%. A 2-stage scheme was proposed by the authors and experiments were extended beyond the 30h, as a result, the recovery was increased to 94%. The higher CF obtained in comparison to our results can be attributed to the higher feed/stripping volume ratio (i.e., 60 versus 2.14). However, assuming no water transport and a 94% recovery, Vecino et al. 2019 should have been able to achieve a CF closer to 60. The CF value obtained (26±3.3) far from 60, meant that water transport occurred and diluted the final product by a factor of 2.3 (Vecino et al., 2019), unlike the current study where the achieved CF was very close or even higher than the used feed/stripping volume ratio (i.e., 2.14) and therefore an almost 100% ammonia selectivity (i.e., negligible water passage) scenario can be assumed. 135 Shi et al. 2022, evaluated the ammonia recovery from an anaerobic digestate using a 0.165 m2 PP HFMC module, where a temperature of 60°C and an aeration flow to maintain a high pH (pH > 8.1) on the feed side were applied (Shi et al., 2022). Authors reported that 82% of the ammonia was recovered from an initial concentration of 3.8 g NH4+ L-1 after only 6h of operation. Temperature accelerated the ammonia transfer but also enhanced the water transport which was in the range of 0.09 g h-1 m-2 and limited the CF to a value of around 8.9. In the 500-180 experiment, similar recovery values were achieved after 6h, but with the significant difference that the 500-180 experiment was operated at room temperature and a negligible water transport can be assumed. Reig et al. 2021 also studied the influence of flow rate (16.0 - 46.2 L h-1) on ammonium recovery using a PP HFMC (1.44 m2). Authors found that lower flow rates entailed lower recoveries and observed that the highest recovery values were obtained when at least one of the two flowrates (e.g., stripping acid or feed) was the highest value tested (Reig et al., 2021), like in the present study. However, they also reported CF values of 20-29 working at a feed/stripping volume ratio of 120 which pointed towards an important water passage. Likewise, Hasanoĝlu et al. 2010, concluded that increasing the flow rate of both streams (47 - 120 L h-1) reduced the mass transfer resistance due to the boundary layer effect and increased the recovery values (Hasanoĝlu et al., 2010). In all the experiments, a feed pH decrease with time was measured (Figure 6). Because of the transport of NH3 from the feed to the stripping side, the equilibrium NH3/NH4+ was shifted to the deprotonation of NH4+ (reaction 1 in Table 3) which decreased the feed pH from 11.8 (max value) to 9.9 (minimum value in Experiment 100-50), but in all cases it remained higher than the ammonia pKa (9.33 at 25°C) making NH3(g) the dominant species in the feed. Very similar results were obtained and reported previously (Aguilar- Moreno et al., 2022). 136 Figure 6. Experimental data of the evolution of pH in the feed tank; (A) Experiment 30-250; (B) Experiment 100-250; (C) Experiment 500-180; (D) Experiment 100-50. Similarly, Jang et al. 2022 experimented with ammonia recovery at different concentrations and pH values from 9-12 using a PP-HFMC module and concluded that increasing the pH resulted in higher removal percentages. They reported values of 20% removal at a pH of 9, which later increased to 50% at a pH of 12, concluding that increasing the initial pH above 10 resulted in a slight improvement of the removal (Jang et al., 2022). 5.4.2. Acid consumption in the permeate side As explained before, because of the transfer of NH3 to the acid side and the subsequent conversion of NH3 into NH4+, H+ was consumed and so 98% concentrated H2SO4 was added to maintain the pH < 2.5 and ensure the lowest ammonia vapour pressure on the stripping side. According to this, each experiment needed different acid dosing rates, but the total amount of acid was very similar and proportional to the total recovered ammonia. For instance, during experiments 30-250, 500-180 and 100-50 a total of 921.8, 997.8 and 909.8 g H2SO4 were correspondingly added while for experiment 100-250 (36% higher NH4+ feed concentration) approximately a 27% more acid (i.e., 1,287.5 g) was used. 137 The average flow rate of the acid dosing pump was in the range of 30-47 ml h-1 and the average total added acid volume per experiment was around 540 ml. Considering that the permeate tank volume was 28 l the total acid addition is less than a 2% volume increase, therefore barely affecting the feed/stripping volume ratio. Evolution of the sulphate (SO42-) and ammonium (NH4+) molar concentrations in the acid side were also measured and simulated according to the chemical equilibrium described by the reactions in Table 3. It was found that for experiments 30-250, 500-180 and 100- 50 the average value of the NH4+/SO42- molar concentration ratio was around 1.7:1, which is between the ammonium sulphate ((NH4)2SO4) and ammonium bisulphate (NH4HSO4) ratios, 2:1 and 1:1, respectively. Further analysis with the speciation software Hydra/Medusa developed by Puigdomench 2001, (Figure 7) shows that at the operating pH of 2.5, around a 74% of the sulphuric acid is dissociated in the form of SO42- while the remaining 26% stays as HSO4- which matches well with the NH4+/SO42- molar concentration ratio measured (Uzkurt Kaljunen et al., 2021). The only exception is experiment 100-250, where a lower ratio was measured (i.e., 1.4) although the acid pH (i.e., 2.5) was the same. However, it is possible that the high acid concentrations reached in this particular experiment and the corresponding higher dilution factor needed for the IC measurements introduced more uncertainties in the measurements. The calculated mean relative error for the SO42- IC measurements was in the range of 1.8-2.8% (Figure S3). Figure 7. Dissociation of the sulfuric acid and corresponding sulphate (SO42-) bisulphate (HSO4- ) molar fractions as a function of pH simulated with the Hydra/Medusa speciation software (Puigdomenech, 2001). 0,76 0,24 0,0 0,2 0,4 0,6 0,8 1,0 012345678910 Molar Fraction [-] pH 144 to experiment 500-180, ammonia recovery and therefore ammonia transfer were faster (e.g., in experiment 500-180 a recovery of 96% was registered after 13h) but not significantly. The results of experiment 180-500-B are also shown in Figure 10a. In this case, the feed decreased from 6.10 g NH3 L-1 to 0.22 g NH3 L-1 (i.e., 96% recovery) after 10h, the feed pH went down from 11.7 to 11.5 and a CF of 2.2 was reached. A total of 952 g of H2SO4 were added and the final NH4+/SO42- molar concentration ratio was also around 1.6. In comparison to experiment 180-500-A, the ammonia recovery rate was again, very similar. In terms of ammonia flux (Figure 10b), the highest maximum flux was measured in experiment 180-500-B (i.e., 19.6 g NH3 h-1 m-2) which was very similar to that of experiment 180-500-A (i.e., 18.6 g NH3 h-1 m-2) and both were comparatively higher than that of experiment 500-180 (i.e., 15.4 g NH3 h-1 m-2) but only in the beginning of the experiment, after 2-3 h the ammonia fluxes were practically the same for the three experiments according to the model. Table 5 shows the simulated 𝑃𝑤 and 𝑃𝑁𝐻3 for the three experiments. Again, the 𝑃𝑁𝐻3 values were very similar for the three experiments (i.e., average value of 175.15 ± 6.8 L h-1 m-2 Pa-1) supporting the idea that a polarization effect due to the asymmetry of the polymeric membrane was not necessarily happening under these working conditions. Regarding co-current versus counter current operation, in light of these results, a similar conclusion can be reached since no significant differences were found for experiments 180-500-A and 180-500-B. Table 5. Membrane permeabilities to water (Pw) and ammonia (PNH3) calculated via the mathematical model for experiments 500-180, 180-500-A and 180-500-B. Experiment 𝑷𝒘 [L h-1 m-2 Pa-1] 𝑷𝑵𝑯𝟑 [L h-1 m-2 Pa-1] 500-180 1.2·10-3 180.2 180-500-A 1.4·10-3 165.5 180-500-B 1.4·10-3 179.8 5.4.5. Mass transfer coefficient and resistance regime In the MC literature, the 𝑘𝑜𝑣 is generally accepted as an overall mass transfer coefficient that collectively accounts for the mass transfer phenomena in the feed, membrane and 145 acid sides and therefore, it would be expected to be equivalent to 𝑈𝑁𝐻3. It is also the most used parameter to compare the performance of different MC membranes, modules and processes (Lauterböck et al., 2012; Moradihamedani, 2021; Tan et al., 2006; Uzkurt Kaljunen et al., 2021; Vecino et al., 2019; Yu et al., 2021). Table 6 summarizes the 𝑘𝑜𝑣 values for all the performed experiments calculated as described in section 3.1. (Figure 11) and the modelled 𝑈𝑁𝐻3, 𝑘𝑁𝐻3,𝑓, 𝑘𝑁𝐻3,𝑚 and 𝑘𝑁𝐻3,𝑎 values, according to the mathematical model and expressed in m h-1. Table 6. Experimentally calculated 𝑘𝑜𝑣 values referred to the internal membrane area (Table 1) and modelled 𝑈𝑁𝐻3, 𝑘𝑁𝐻3,𝑓, 𝑘𝑁𝐻3,𝑚 and 𝑘𝑁𝐻3,𝑎 values for each experiment. Experiment 𝒌𝒐𝒗 (m s-1) 𝑼𝑵𝑯𝟑 (m s-1) 𝒌𝑵𝑯𝟑,𝒇 (m s-1) 𝒌𝑵𝑯𝟑,𝒎 (m s-1) 𝒌𝑵𝑯𝟑,𝒂 (m s-1) 30-250 (4.56±0.06) ·10-7 1.99·10-8 7.5·10-5 1.64·10-10 1.76 ·10-6 100-250 (6.28±0.08) ·10-7 7.39·10-8 1.25·10-4 2.21·10-10 1.76·10-6 500-180 (7.39±0.11) ·10-7 1.10·10-7 2.47·10-4 3.37·10-10 1.58·10-6 100-50 (4.61±0.19) ·10-7 5.61·10-8 9.44·10-5 1.68·10-10 1.30·10-6 180-500-A (8.33±0.22) ·10-7 1.03·10-7 1.61·10-4 3.12·10-10 2.22·10-7 180-500-B (9.75±0.26) ·10-7 1.11·10-7 1.62·10-4 3.37·10-10 2.22·10-7 Figure 11. Representation of the natural logarithm ln(cNH3,0 cNH3,t ⁄) versus the time (t) a linear relationship is found which slope is kov·Am V and from which kov is calculated. Points: experimental data; lines: linear fit to eq. 1. 146 As expected, the highest 𝑘𝑜𝑣 values corresponded to the experiments performed at higher flow rates and amongst them, experiment 180-500-B (i.e., co-current configuration configuration) with a 𝑘𝑜𝑣 of 3.51±0.10 ·10-3 m h-1 was the highest. As the flow rates increased, the thickness of the respective boundary layers decreased and their mass transfer resistances became lower, resulting in an overall improvement of the 𝑘𝑜𝑣 (Moradihamedani, 2021). On the other hand, experiment 100-250, whose initial ammonia feed concentration was higher (i.e., 10 g l-1) and having showed higher ammonia fluxes than the rest of the experiments, did not exhibited a comparatively higher 𝑘𝑜𝑣 value, in fact it was among the lowest (i.e., 2.26 ± 0.03·10-3 m h-1). This is in principle, consistent with other authors results (Lauterböck et al., 2012; Uzkurt Kaljunen et al., 2021), who reported decreasing 𝑘𝑜𝑣 values with higher initial ammonia feed concentrations. Conversely, other studies (Tan et al., 2006) support the opposite hypothesis and conclude that the 𝑘𝑜𝑣 is in fact independent of the initial concentration. In any case, it must be pointed out that the feed and acid flow rates combination in experiment 100-250 was in the lower range. When compared to other values reported in literature, these 𝑘𝑜𝑣 values are higher to those reported by Uzkurt Kaljunen et al. 2021 who used a PTFE module of 10 mm diameter membrane fibres (i.e., 0.612·10-3 and 0.36·10-3 m h-1 working with 0.75 and 3 g L-1 NH3) (Uzkurt Kaljunen et al., 2021) but very similar to those reported by (Reig et al., 2021) (i.e., 2.05 - 3.96 ·10-3 m h-1) and (Vecino et al., 2019) (i.e., 2.3 - 3.1 ·10-3 m h-1) working with a 1.4 m2 PP Liqui-Cel® Membrane Contactor X-50 and ammonia feed streams in the range of 4-4.5 g NH3 L-1. Sheikh et al. 2022, using the same membrane module as in the present study and 5 g NH3 L-1 in the feed, reported 𝑘𝑜𝑣 values between 1.8·10-4 and 1.04·10-3 m h-1 (i.e., below the lowest 𝑘𝑜𝑣 obtained here). Nevertheless, they worked at much lower flow rates (i.e., 27 L h-1) and noted a significant reduction in the 𝑘𝑜𝑣 values when the feed/acid volume ratio was reduced from 60 to 10 (Sheikh et al., 2022). Compared to other novel polymer chemistries, such as Nafion, which is claimed to be an ammonia selective membrane with high flux (Tricoli & Cussler, 1995), Yu et al. 2021, have recently evaluated the effect of different operational parameters in the 𝑘𝑜𝑣 of a new Nafion-PFTE HF membrane. By maintaining a constant flow velocity of 0.037 m s-1 (the maximum flow velocity in the present study was 1.21 m s-1) and an initial concentration 147 of 0.1 g NH3 L-1 at a pH 12, the authors reported 𝑘𝑜𝑣 values of 1.2·10-2 m h-1, one order of magnitude higher that the ones obtained in this study (Yu et al., 2021). However, Yu et al. operated the HF system quite differently since the feed was circulated through the lumen side and the fibre bundle was simply immersed in a stirred beaker with the receiving solution. As discussed earlier in the introduction section, HFMC have been already tested at pilot scale. At pilot or even industrial scale the ideal operation is not batch but continuous flow or open-loop configuration. In this direction, Licon et al. 2015, evaluated the removal of ammonia traces from water used for hydrogen production by electrolysis, operating at a feed flow rate between 10 and 82 L h-1 in open loop configuration. The authors obtained an ammonia removal rate of 78% for a single step process and from an initial concentration of 15 mg NH3 L-1 but reported a 𝑘𝑜𝑣 value of 1.47 m h-1. In the end, they calculated that a total of three modules in series were needed to reach a 95% removal (Licon et al., 2015). Regarding the 𝑈𝑁𝐻3values, they followed the same trend as the 𝑘𝑜𝑣 but they differ by an order of magnitude. As mentioned before, it is generally accepted that the 𝑘𝑜𝑣 represents the overall mass transfer coefficient, however, as pointed out by Wang L. K. et al. 1993, the 𝑘𝑜𝑣 can be used as the overall mass transfer coefficient only if no mass transfer resistance across the membrane and the stripping side exists (Wang & Cussler, 1993), which is not the case in LL applications. In fact, when comparing the different transport resistances calculated via the mathematical model (i.e., 𝑘𝑁𝐻3,𝑓, 𝑘𝑁𝐻3,𝑚 and 𝑘𝑁𝐻3,𝑎 in Table 5) the system seems to be dominated primarily by the 𝑘𝑁𝐻3,𝑚 (i.e., membrane resistance regime) which is in agreement with the semi volatile nature of the ammonia (i.e., Henry constant between 10-4 and 3·103 Pa L mol-1) (S. Lee & Straub, 2022) and secondly by the 𝑘𝑁𝐻3,𝑎 which is often considered as a minor limiting factor in MC literature. The fact that this particular MC has been design for gas-liquid (GL) applications where the gas flows in the lumen side (i.e., acid side in our application) and where negligible resistances on this side and across the membrane are often found (Wang & Cussler, 1993), points out the need for further lumen side flow conditions optimization (i.e., flowrates, fiber/module geometry) if used in LL configuration. Regarding the different 𝑘𝑁𝐻3,𝑚 values obtained via modelling (Table 5), just like the 𝑃𝑁𝐻3 values, these should be 148 independent of the flow conditions (i.e., constant for all the experiments) and yet the mathematical model calculated different values. A possible explanation for these results has been already provided in section 4.3. Finally, the maximum 𝐽𝑁𝐻3 (at time = 0h in graphs 4a and 6b) versus the calculated 𝑘𝑜𝑣 and 𝑈𝑁𝐻3values (Table 5) has been plotted in Figure 7. As previously mentioned, the maximum ammonia flux does depend on the initial concentration, for this reason, the maximum 𝐽𝑁𝐻3 has been normalized by the initial ammonia concentration. Figure 12. Maximum JNH3 (at time = 0h in graphs 12A and 12B) and normalized by the initial ammonia feed concentration in g L-1 versus kov and UNH3 calculated for all the experiments. Figure 12 shows a steady increase of the 𝐽𝑁𝐻3 with increasing 𝑘𝑜𝑣 or 𝑈𝑁𝐻3. Thus, it suggests that even further improvements can be made to enhance the module's performance in terms of ammonia flux. In view of the mass transfers coefficients calculated in Table 5 further operational improvements should be focussed on the acid/stripping side. 5.5. Conclusions In this study PMP LL-HF asymmetric MC membrane has been tested and proved as a promising technology for the selective recovery and concentration of ammonia from rather diluted waste streams. 0 1 2 3 4 5 0,0001 0,001 0,01 J NH3_nor [g h-1 m-2 g/l -1] Kov / UNH3 [m s-1] U NH3 Kov 149 All experiments performed in this study reached recovery values above 90% after 15h. The results showed that increasing the flowrate (experiments 500-180; 180-500-A and 180-500-B) decreased the recovery time, while the initial concentration had a minimal influence (experiment 100-250). Likewise, the CF values obtained (i.e., 2.1-2.2) were also independent from the flow rate combinations and rather determined by the feed/stripping volume ratio (i.e. 2.14). Regarding which species were formed on the acid side, the mean value of the measured NH4+/ SO4-2 molar ratio was 1.7:1 in most experiments, agreeing with the acid dissociating analysis of the Hydra/Medusa speciation software and confirming that the operating acid pH determined the ammonium salts fractions. It was observed that for this module, low flowrate values (i.e., <100 L h-1) on either side reduced greatly the ammonia mass transfer rate due to a combination of a greater boundary layer effect and a poorer flow distribution. Highest ammonia flux was that of experiment 100-250 due to the higher initial ammonia concentration. But also, and according to the model, the higher ammonia initial concentration and the consequent final dissolved species in the permeate side led to the highest water flux (i.e., 18.6 g H2O h-1 m-2) attributed to a larger OD effect. The mathematical model was used for determining and comparing the NH3 and H2O membrane permeabilities of the module. It was confirmed that the membrane used in this study was highly selective towards ammonia with a 𝑃𝑁𝐻3 between 86.9 and 180.2 L h-1 m-2 bar-1 and an average 𝑃𝑤 of 1.29 ± 0.105 ·10-3 L h-1 m-2 bar-1. Additionally, the similar ammonia flux and 𝑃𝑁𝐻3 values of experiments 500-180, 180-500-A and 180-500B, supported the idea that a polarization effect due to the asymmetry of the polymeric membrane is not necessarily happening under these working conditions or is negligibly small. The highest 𝑘𝑜𝑣 value experimentally obtained was 3.51±0.10·10-3 m h-1 (experiment 180-500-B) which was in the upper range of the HF-LLMC literature. However, it was proven that the 𝑘𝑜𝑣 might not be the most appropriate parameter to compare different MC modules operating in LL mode and particularly inadequate in capturing the influence of the acid side parameters. Finally, a deeper analysis of the mass transfer resistances on each of the channels showed further potential improvement on the acid side hydrodynamics. 150 5.6. References Aguilar-Moreno, M., Vinardell, S., Reig, M., Vecino, X., Valderrama, C., & Cortina, J. L. (2022). Impact of Sidestream Pre-Treatment on Ammonia Recovery by Membrane Contactors: Experimental and Economic Evaluation. Membranes, 12(12). https://doi.org/10.3390/membranes12121251 Ahn, Y. T., Hwang, Y. H., & Shin, H. S. (2011). Application of PTFE membrane for ammonia removal in a membrane contactor. Water Science and Technology, 63(12), 2944–2948. https://doi.org/10.2166/wst.2011.141 Alkhudhiri, A., Darwish, N., & Hilal, N. (2012). Membrane distillation: A comprehensive review. Desalination, 287, 2–18. https://doi.org/10.1016/j.desal.2011.08.027 Amaral, M. C., Magalhães, N. C., Moravia, W. G., & Ferreira, C. D. (2016). Ammonia recovery from landfill leachate using hydrophobic membrane contactors. Water Science and Technology, 74(9), 2177–2184. https://doi.org/10.2166/wst.2016.375 Bazhenov, S. D., Bildyukevich, A. V., & Volkov, A. V. (2018). Gas-liquid hollow fiber membrane contactors for different applications. Fibers, 6(4). https://doi.org/10.3390/fib6040076 Beckinghausen, A., Odlare, M., Thorin, E., & Schwede, S. (2020). From removal to recovery: An evaluation of nitrogen recovery techniques from wastewater. Applied Energy, 263(February), 114616. https://doi.org/10.1016/j.apenergy.2020.114616 Chabanon, E., Roizard, D., & Favre, E. (2013). Modeling strategies of membrane contactors for post-combustion carbon capture: A critical comparative study. Chemical Engineering Science, 87, 393–407. https://doi.org/10.1016/j.ces.2012.09.011 Chagas, G. O., Coelho, L. C., Darvish, M., & Renaud, J. (2023). Modeling and solving the waste valorization production and distribution scheduling problem. European Journal of Operational Research, 306(1), 400–417. https://doi.org/10.1016/j.ejor.2022.06.036 Darestani, M., Haigh, V., Couperthwaite, S. J., Millar, G. J., & Nghiem, L. D. (2017). Hollow fibre membrane contactors for ammonia recovery: Current status and future developments. Journal of Environmental Chemical Engineering, 5(2), 1349–1359. 151 https://doi.org/10.1016/j.jece.2017.02.016 Deng, Z., van Linden, N., Guillen, E., Spanjers, H., & van Lier, J. B. (2021). Recovery and applications of ammoniacal nitrogen from nitrogen-loaded residual streams: A review. Journal of Environmental Management, 295, 113096. https://doi.org/10.1016/j.jenvman.2021.113096 DIC Corporation. (2016). Instruction Manual for SEPAREL PF Modules. 1–22. DiC, S. (n.d.). Dic_Separel.Pdf. Eskicioglu, C., Galvagno, G., & Cimon, C. (2018a). Approaches and processes for ammonia removal from side-streams of municipal effluent treatment plants. Bioresource Technology, 268(May), 797–810. https://doi.org/10.1016/j.biortech.2018.07.020 Eskicioglu, C., Galvagno, G., & Cimon, C. (2018b). Approaches and processes for ammonia removal from side-streams of municipal effluent treatment plants. Bioresource Technology, 268(May), 797–810. https://doi.org/10.1016/j.biortech.2018.07.020 Gabelman, A., & Hwang, S. T. (1999). Hollow fiber membrane contactors. Journal of Membrane Science, 159(1–2), 61–106. https://doi.org/10.1016/S0376- 7388(99)00040-X García-González, M. C., Vanotti, M. B., & Szogi, A. A. (2015). Recovery of ammonia from swine manure using gas-permeable membranes: Effect of aeration. Journal of Environmental Management, 152, 19–26. https://doi.org/10.1016/j.jenvman.2015.01.013 González Montiel, J. M. (2008). Acetato de etilo en la Industria. Journal of Chemical Information and Modeling, 53(9), 287. Hasanoĝlu, A., Romero, J., Pérez, B., & Plaza, A. (2010). Ammonia removal from wastewater streams through membrane contactors: Experimental and theoretical analysis of operation parameters and configuration. Chemical Engineering Journal, 160(2), 530–537. https://doi.org/10.1016/j.cej.2010.03.064 Hermassi, M., Valderrama, C., Gibert, O., Moreno, N., Querol, X., Batis, N. H., & Cortina, J. L. (2017). Recovery of nutrients (N-P-K) from potassium-rich sludge 152 anaerobic digestion side-streams by integration of a hybrid sorption-membrane ultrafiltration process: Use of powder reactive sorbents as nutrient carriers. Science of the Total Environment, 599–600, 422–430. https://doi.org/10.1016/j.scitotenv.2017.04.140 Ignatenko, V. Y., Anokhina, T. S., Ilyin, S. O., Kostyuk, A. V., Bakhtin, D. S., Antonov, S. V., & Volkov, A. V. (2020). Fabrication of microfiltration membranes from polyisobutylene/polymethylpentene blends. Polymer International, 69(2), 165–172. https://doi.org/10.1002/pi.5932 Jang, D., Tran, T. N., Ko, K., Park, D., Park, S., & Kang, S. (2022). Parametric studies during the removal of ammonia by membrane contactor with various stripping solutions. Chemosphere, 309(P1), 136648. https://doi.org/10.1016/j.chemosphere.2022.136648 Kurniawan, T. A., Lo, W. H., & Chan, G. Y. S. (2006). Physico-chemical treatments for removal of recalcitrant contaminants from landfill leachate. Journal of Hazardous Materials, 129(1–3), 80–100. https://doi.org/10.1016/j.jhazmat.2005.08.010 Laguntsov, N. I., & Karaseva, M. D. (2019). Pressure effect at permeability and deformation of the hollow fiber membrane gas separation module. Journal of Physics: Conference Series, 1189(1). https://doi.org/10.1088/1742- 6596/1189/1/012028 Lauterböck, B., Ortner, M., Haider, R., & Fuchs, W. (2012). Counteracting ammonia inhibition in anaerobic digestion by removal with a hollow fiber membrane contactor. Water Research, 46(15), 4861–4869. https://doi.org/10.1016/j.watres.2012.05.022 Lawson, K. W., & Lloyd, D. R. (1997). Membrane distillation. Journal of Membrane Science, 124(1), 1–25. https://doi.org/10.1016/S0376-7388(96)00236-0 Ledezma, P., Kuntke, P., Buisman, C. J. N., Keller, J., & Freguia, S. (2015). Sourceseparated urine opens golden opportunities for microbial electrochemical technologies. Trends in Biotechnology, 33(4), 214–220. https://doi.org/10.1016/j.tibtech.2015.01.007 Lee, S., & Straub, A. P. (2022). Analysis of Volatile and Semivolatile Organic Compound 153 Transport in Membrane Distillation Modules. ACS ES and T Engineering, 2(7), 1188–1199. https://doi.org/10.1021/acsestengg.1c00432 Lee, W., An, S., & Choi, Y. (2021). Ammonia harvesting via membrane gas extraction at moderately alkaline pH: A step toward net-profitable nitrogen recovery from domestic wastewater. Chemical Engineering Journal, 405(May 2020), 126662. https://doi.org/10.1016/j.cej.2020.126662 Licon Bernal, E. E., Maya, C., Valderrama, C., & Cortina, J. L. (2016). Valorization of ammonia concentrates from treated urban wastewater using liquid-liquid membrane contactors. Chemical Engineering Journal, 302, 641–649. https://doi.org/10.1016/j.cej.2016.05.094 Licon, E., Reig, M., Villanova, P., Valderrama, C., Gibert, O., & Cortina, J. L. (2015). Ammonium removal by liquid–liquid membrane contactors in water purification process for hydrogen production. Desalination and Water Treatment, 56(13), 3607– 3616. https://doi.org/10.1080/19443994.2014.974216 Markova, S. Y., Gries, T., & Teplyakov, V. V. (2020). Poly(4-methyl-1-pentene) as a semicrystalline polymeric matrix for gas separating membranes. Journal of Membrane Science, 598(December 2019). https://doi.org/10.1016/j.memsci.2019.117754 Matassa, S., Batstone, D. J., Hülsen, T., Schnoor, J., & Verstraete, W. (2015). Can direct conversion of used nitrogen to new feed and protein help feed the world? Environmental Science and Technology, 49(9), 5247–5254. https://doi.org/10.1021/es505432w Mayor, Á., Beltran, E., Cortina, J. L., & Valderrama, C. (2023). Nitrogen flow analysis in Spain: Perspectives to increase sustainability. Science of the Total Environment, 858(March 2022), 10–14. https://doi.org/10.1016/j.scitotenv.2022.160117 Moradihamedani, P. (2021). Recent developments in membrane technology for the elimination of ammonia from wastewater: A review. Polymer Bulletin, 78(9), 5399– 5425. https://doi.org/10.1007/s00289-020-03386-y Noda, I., Brown-West, D. G., & Gryte, C. C. (1979). Effect of flow maldistribution on hollow fiber dialysis - experimental studies. Journal of Membrane Science, 5(C),