Knowledge integration to promote waste-to-PHA biorefineries

INTERNATIONAL DOCTORAL SCHOOL OF THE USC Mateo Saavedra del Oso PhD Thesis Knowledge integration to promote waste-to-PHA biorefineries Santiago de Compostela, 2023 Doctoral Programme in Chemical and Environmental Engineering TESIS DE DOCTORADO KNOWLEDGE INTEGRATION TO PROMOTE WASTE-TO-PHA BIOREFINERIES Mateo Saavedra del Oso ESCUELA DE DOCTORADO INTERNACIONAL DE LA UNIVERSIDAD DE SANTIAGO DE COMPOSTELA PROGRAMA DE DOCTORADO EN INGENIERÍA QUÍMICA Y AMBIENTAL SANTIAGO DE COMPOSTELA 2023 DECLARACIÓN DEL AUTOR/A DE LA TESIS D. Mateo Saavedra del Oso Título de la tesis: Knowledge integration to promote waste-to-PHA biorefineries Presento mi tesis, siguiendo el procedimiento adecuado al Reglamento y declaro que: 1) La tesis abarca los resultados de la elaboración de mi trabajo. 2) De ser el caso, en la tesis se hace referencia a las colaboraciones que tuvo este trabajo. 3) Confirmo que la tesis no incurre en ningún tipo de plagio de otros autores ni de trabajos presentados por mí para la obtención de otros títulos. 4) La tesis es la versión definitiva presentada para su defensa y coincide la versión impresa con la presentada en formato electrónico. Y me comprometo a presentar el Compromiso Documental de Supervisión en el caso que el original no esté depositado en la Escuela. En Helsinki, 20 de abril de 2023. Firma electrónica AUTORIZACIÓN DEL DIRECTOR / TUTOR DE LA TESIS Knowledge integration to promote waste-to-PHA biorefineries Dª. Almudena Hospido Quintana D. Miguel Mauricio Iglesias INFORMA/N: Que la presente tesis, se corresponde con el trabajo realizado por D. Mateo Saavedra del Oso, bajo mi dirección/tutorización, y a utorizo su presentación , considerando que reúne l os r equisitos exigidos en el R eglamento de Estudios de Doctorado de la USC, y que como director de esta no incurre en las causas de abstención establecidas en la Ley 40/2015. De acuerdo con lo indicado en el Reglamento de Estudios de Doctorado, declara también que la presente tesis doctoral es idónea para ser defendida en base a la modalidad de COMPENDIO DE PUBLICACIONES, en los que la participación del doctorando/a fue decisiva para su elaboración y las publicaciones se ajustan al Plan de Investigación. En Santiago de Compostela, 21 de abril de 2023 INDEX Acknowledgements .............................................................................................................. i List of publications ............................................................................................................. iii Abstract.............................................................................................................................. vii Resumo ............................................................................................................................. xiii 1 Introduction ...................................................................................................................... 1 1.1. A new horizon for polyhydroxyalkanoates ........................................................... 3 1.2. PHA production based on mixed microbial culture systems ................................ 4 1.3. PHA downstream processing ................................................................................ 5 1.4. Challenges for PHA production based in MMC systems ..................................... 6 1.5 Thesis structure and objectives .............................................................................. 9 2 Materials & Methods ..................................................................................................... 13 2.1. Mathematical modelling ..................................................................................... 15 2.2. Upscaling Framework ........................................................................................ 16 2.3. Life Cycle based methods .................................................................................. 17 3 Scientific Publications derived from this thesis ............................................................ 21 Paper I ......................................................................................................................... 23 Paper II ....................................................................................................................... 35 Paper III ...................................................................................................................... 51 Paper IV ...................................................................................................................... 75 Paper V ....................................................................................................................... 91 4 Discussion .................................................................................................................... 115 4.1. Consistency of the thesis .................................................................................. 117 4.2. Critical analysis of the methods used in the knowledge integration approach . 120 4.3. A framework to engage with different interests and available knowledge ...... 122 4.4. Prospects for the development of waste-to-PHA biorefineries ........................ 125 5 Conclusions .................................................................................................................. 129 References ....................................................................................................................... 133 vii ABSTRACT Plastics are ubiquitous and essential components of modern society. Its production is linked to two of today’s society major concerns: climate change and marine ecosystems spoilage. Although recycling rates have increased 117% compared to 2006, the plastics industry is still based on non-renewable resources and a linear economy approach, as only 8.5% of plastics packaging total production were post-consumer recycled. In the recent years, the European Commission has tried to tackle plastics related issues by legislating against single-use plastics and creating a new legal framework within the European Green Deal and the new Circular Economy Action Plan for a cleaner and more competitive Europe. This new legal framework includes actions to increase circularity of plastics and to foster the production, labelling, and use of biobased, compostable and biodegradable plastics. Among bioplastics, polyhydroxyalkanoates (PHA) present an outstanding opportunity for substituting oil-based plastics. PHA are biobased and biodegradable polymers, even in marine environments. PHA are produced through microbial fermentation from a diverse range of substrates, including organic wastes and byproducts. Although PHA present advantages over oil-based plastic, their market implementation is still hindered by their high production cost and uncertain environmental performance. In fact, high energy requirement during feedstock cultivation, sterilization in pure culture fermentation and PHA downstream processing were reported as the main hotspots from both environmental and economic perspectives. Using organic wastes as feedstock and non-sterile conditions would significantly reduce the cost of PHA production. The major driver for PHA development at industrial scale would be demonstrating a better environmental performance than conventional plastics. H2020 project USABLE Packaging proposes cost-effective and sustainable PHA production routes based on mixed microbial culture (MMC) systems. MMC systems enable operating under non-sterile conditions and using organic wastes as feedstocks. MMC systems comprise a 3-step system: (i) anaerobic fermentation, where complex organic feedstocks are hydrolyzed and fermented into volatile fatty acids (VFA); (ii) an enrichment mixed culture, where PHA-storing bacteria are selected by imposing a feast/famine regime under nutrient excess for ensuring growth and; (iii) an accumulation step, where enriched biomass is fed VFA under nutrient limitation to facilitate PHA accumulation. After intracellular accumulation, the PHA needs to be recovered from the PHA enriched biomass. PHA downstream processing consists of several steps: firstly, PHA-enriched biomass may be concentrated by physical separation (filtration, centrifugation, sedimentation) and pretreated (heating, grinding, freezing) to enhance the PHA recovery; then, PHA can be recovered by two different methods: (i) solubilizing the non-cellular PHA mass (NCPM) through chemical digestion, enzymatic digestion, mechanical disruption or a combination of them or (ii) solubilizing the PHA through solvent extraction. Finally, and depending on the quality requirements, PHA is purified by redissolution with water or ethanol, or by a bleaching treatment employing sodium hypochlorite or hydrogen peroxide. PHA production based in MMC systems tackles the main drawbacks of PHA production based on pure cultures and integrates PHA production into the circular economy and resource recovery paradigm. However, its development at large scale still faces multiple challenges, from the variability of anaerobic fermentation products to guaranteeing its environmental performance. MATEO SAAVEDRA DEL OSO viii Anaerobic fermentation is one of the key steps in the PHA production, as the copolymer blend composition and thus, its properties and potential applications are determined by the fermented stream employed, i.e. the VFA composition. The anaerobic fermentation has a very variable stoichiometry which depends, among others, on (i) pH, (ii) substrate composition and (iii) hydraulic retention time. Having a better understanding of anaerobic fermentation stoichiometry variability would enable driving VFA production towards desired compositions, and mathematical models can be of great help describing anaerobic fermentation of organic wastes. Unfortunately, current models cannot predict the selectivity and productivity of VFA from a given substrate due to the variability of acidogenic stoichiometry which depends, additionally, on previous disintegration and hydrolysis steps. PHA downstream processing is one of the main bottlenecks in the development of a costefficient and sustainable PHA production. Concretely, it can account for up to 50% of the production costs and environmental impacts, due to significant chemicals and energy consumption. The development and scaling up of PHA downstream processing pose a challenge due to uncertainties regarding the polymer type, culture type and required properties. Life cycle assessment (LCA) and life cycle costing (LCC) can be used to assist the decision making during the process development and select a cost-efficient and sustainable method. However, there is a lack of environmental and economic assessments of PHA downstream processing. PHA production based on MMC systems has already achieved a technology readiness level (TRL) of 5-6, i.e. pilot scale. However, further steps for pushing these emerging processes require guidance to ensure its sustainable performance at industrial scale. Prospective LCA (pLCA) enables the upscaling of emerging technologies using scenarios of future performance at industrial scale, and the comparison of those future processes with the industrial ones. Deriving scenarios on how these technologies could develop in the future would facilitate their environmental optimization. The interest for PHA is likely to grow as high-performance biodegradable plastics with a potentially reduced environmental footprint. However, it is required to engage and connect all stakeholders among the value chain. Waste producers may not have the resources or the knowledge to screen the potential of the organic wastes to be used as feedstock. How can PHA producers explore the operational conditions and substrates without lab and/or pilot research? How can waste managers be connected to PHA producers with specific polymer composition requirements? Likewise, there might be other value chains that could employ these wastes as feedstock constraining their availability. Thus, PHA production needs to prove their environmental and economic feasibility against other valorization technologies. Proving the environmental and economic feasibility of PHA value chain might not be enough to promote its implementation. PHA value chains need to demonstrate their environmental performance at product level also against their commercial counterparts. Most LCA studies on PHA production by MMC focused on analyzing the environmental performance of particular stages of the life cycle (e.g., PHA downstream processing) or comparing MMC systems with other valorization technologies. Therefore, they excluded the final stages of the plastic life cycle: namely, the shaping and compounding, the use, and the end-of-life (EoL). Excluding the EoL disregards the benefits derived from PHA biodegradability and ignores the long-term environmental impacts caused in marine ecosystems by plastic pollution. The aim of this thesis is to assist the decision-making process in the development of new PHA value chains based on MMC systems. By integrating the knowledge available and ABSTRACT ix different methodologies and tools, such as pLCA or mathematical modelling, the challenges of PHA production can be overcome. The identified gaps regarding the production and selectivity of VFA by anaerobic fermentation of organic wastes are covered in Paper I. Additionally, this work proposes a framework to connect multiple levels of actors in the value chain with different interests, expertise levels, available data, and knowledge. This framework can be the key to foster circular economy and bridge the development risks, allowing an early-stage evaluation of process performance. It is implemented as a computer-aided design tool, which comprises: (i) a library of substrates including their characterization and appropriate kinetic parameter selection, (ii) an integral metabolic model which solves both identified gaps regarding the disintegration mechanisms and the acidogenic stoichiometry variability in the anaerobic mono and cofermentation of complex organic wastes, and (iii) a set of indicators to interpret simulation results and assist decision making. The tool applications on connecting multiple stakeholders in the value chain with different interests and expertise are proven. These stakeholders may include from those involved in the optimization and design of acidification processes to those studying the feasibility of valorizing wastes from an economic an environmental perspective, are proven. Thus, this framework enables the identification of technical bottlenecks, and proposes innovative solutions prior to expensive and time-consuming lab and pilot research. Paper II aims to find the economic and environmental hotspots in the PHA downstream processing and to provide insights for its optimization. Eight PHA downstream alternative processes were selected after a systematic review of available PHA downstream methods and related LCA. These processes are evaluated from a techno-economic and an environmental perspective, assessing scale-up possibilities and challenges. Extraction processes are hindered by high energy consumption and the use of large amounts of solvents. However, the integration of PHA recovery within biorefineries where surplus of utilities is produced, and the use of so-called green solvents may decrease the environmental impacts up to a 50%. Mechanical disruption was pointed out as the most promising PHA recovery method from both economic and environmental perspective. Thus, Paper II contributes to increase the available knowledge regarding the PHA downstream processing and its modular character enables its further integration of LCA flowchart and inventories into the value chain. Once the local bottlenecks for the development of waste-to-PHA biorefineries were addressed, Paper III depicts how these biorefineries could develop in the future, ensuring that environmental guidance is included while there are still opportunities for major alterations. The inherent uncertainties of the prospective LCA were covered by deriving scenarios following methods on systematic scenario development. Influencing parameters were firstly identified based on literature review as well as both individual meetings and a workshop with stakeholders. Based on the information collected, four future scenarios were derived (and validated by stakeholders) considering both surrounding (e.g., scale, environmental or bioeconomy policies) and technological parameters (e.g., acidification yield, PHA content in biomass or recovery yield). The scenarios derived under ambitious environmental and bioeconomy policies showed up to 50% lower environmental impacts than those under business-as-usual policies, mainly due to the different environmental burdens of background processes (e.g., electricity mix with low renewable energies share) and the higher consumption of chemicals and utilities. Sensitivity analysis results pointed out at the MATEO SAAVEDRA DEL OSO x recovery yield and PHA content as the parameters that influence most the environmental performance, being responsible for up to 60% of variance in environmental performance. These parameters determine the chemicals and utilities consumption in PHA downstream processing, which is confirmed as the main environmental hotspot. By quantifying the relationship between technical parameters and the process environmental performance with a prospective perspective, Paper III goes beyond previous LCA studies on PHA production and serves effectively as development guidance of waste-to-PHA biorefineries. Paper IV integrates the insights given on the prospective LCA of waste-to-PHA biorefineries and expands the framework developed in Paper II to build a comprehensive framework (CarboxyLCA) for a preliminary holistic assessment of organic waste valorization within the carboxylate platform. CarboxyLCA enables a fast evaluation of a wide range of substrate and conditions, with a minimum requirement of experimental data. Initially, it covers different valorization pathways (anaerobic digestion, VFA production and PHA production) and products (electricity, district heat, VFA and PHA), but is ready to be enlarged in the future if other pathways or products gain interest. This framework is implemented as a computer-aided tool and consists of the following components: (1) a substrate library, (2) a mechanistic kinetic and stoichiometric model, (3) an upscaling and flowchart design module, (4) a pLCA module, (5) a LCC module, and (6) a set of comprehensive indicators for interpreting simulation results and facilitating decision making. CarboxyLCA functionalities are proven through case studies that showcase its utility on screening optimal substrates, technologies, and products within the carboxylate platform. The findings from this study highlight the significance of the framework in advancing the principles of circular economy and fostering collaboration among various stakeholders in the value chain. With a wide range of expertise and interests, this tool serves as a bridge that connects these stakeholders, enabling the promotion of sustainable practices in the valorization of organic waste. Finally, Paper V compares the environmental performance of PHA-based items intended for food contact (reusable plates, frozen bags and biscuit bags), developed within USABLE Packaging, to their commercial counterparts. To the best of our knowledge, this is the first time that a cradle-to-grave LCA, including the long-term environmental impacts caused by microplastics, is performed. The system is modeled integrating all available knowledge, from pilot-scale data to process simulators, and following upscaling frameworks. PHA-based items outperform their commercial counterparts from an environmental perspective. Indeed, they show environmental benefits due to the avoided electricity obtained in the cogeneration heat and power unit within the PHA production. However, these environmental benefits are very sensitive to the substituted electricity environmental burdens. PHA downstream processing remains as the main environmental hotspot due to solvents and utilities usage. EoL has a negligible contribution to overall impact categories. Recommendations are provided: using the biogas produced within PHA production to produce utilities and substituting dimethyl carbonate-ethanol solvents by acetone-water. The different technologies’ maturity levels and market implementation are seen as critical and thus, it is important to focus on how upscaling affects the results and interpretation of life cycle assessment findings. This serves as a valuable reference point for the development of processes and products. The PHA value chain based on MMC systems could be depicted as a sort of puzzle comprised by different pieces. Within this thesis the different pieces are assembled by ABSTRACT xi integrating the available knowledge, mathematical models, tools, and methodologies. The consistency of this thesis on bridging the bottlenecks that hinder the development of waste-toPHA biorefineries is proven. This thesis assists the decision-making during the development of this innovative value chain at multiple levels: from solving the local bottlenecks (zoom in) within the anaerobic fermentation (Paper I) and the PHA downstream processing (Paper II) to forecast the development of waste-to-PHA biorefineries (Paper III) and developing a framework (zoom out) that connects stakeholders across the value chain (Paper II and IV). These papers provide insights that aid on assembling the different pieces that are necessary to validate the sustainable performance of waste-to-PHA biorefineries. Hence, Paper V acts as the keystone that integrates all the knowledge generated in former papers, the literature and the pilot-scale data provided by the USABLE Packaging partners. By applying knowledge integration, stakeholders across the value chain can strengthen relationships between them, gain valuable insights and take informed decisions during the early-stage development of waste-to-PHA biorefineries. Hence, it facilitates the implementation of circular-based solutions and mitigating both environmental and economic risks. It is discussed how the different stakeholders (food industry and waste management, PHA manufacturing, packaging users, designers and converters, policymaking, and academia and R&D) can benefit from the knowledge integration approach. As conclusion, the knowledge integration approach adopted in the former document contributes to the development of a sustainable PHA value chain within a circular economy approach. xiii RESUMO Os plásticos son parte esencial e omnipresente nas sociedades modernas. A súa produción está vencellada a dúas das maiores preocupacións da sociedade actual: o cambio climático e a deterioración dos ecosistemas mariños. Aínda que as taxas de reciclaxe aumentaron un 117% con respecto a 2006, a industria do plástico segue baseándose en recursos non renovables seguindo un enfoque de economía lineal, xa que só o 8,5% da produción total de envases de plástico procedía de reciclaxe posterior ao consumidor. Nos últimos anos, a Comisión Europea tentou abordar os problemas relacionados cos plásticos lexislando de forma restritiva os plásticos dun só uso e creando un novo marco xurídico dentro do Pacto Verde Europeo e o novo Plan de Acción de Economía Circular para unha Europa máis limpa e competitiva. Este novo marco xurídico inclúe accións para aumentar a circularidade dos plásticos e fomentar a produción, a etiquetaxe e o uso de plásticos baseados en recursos biolóxicos, compostables e biodegradables. Entre os bioplásticos, os polihidroxialcanoatos (PHA) representan unha oportunidade excepcional para substituír aos plásticos derivados do petróleo. Os PHA son polímeros de base biolóxica e biodegradables, mesmo nos ambientes mariños. Os PHA prodúcense mediante fermentación microbiana a partir dunha ampla gama de substratos, como residuos orgánicos e subprodutos. Se ben os PHA presentan vantaxes sobre os plásticos derivados do petróleo, a súa implantación no mercado segue a verse obstaculizada polo seu elevado custo de produción e un desempeño ambiental aínda non establecido con certeza. De feito, as altas necesidades enerxéticas durante o cultivo da materia prima, a esterilización na fermentación de cultivos puros e o procesamento posterior do PHA considéranse os principais puntos conflitivos desde o punto de vista ambiental e económico. Empregar residuos orgánicos como materia prima e realizar a fermentación en condicións non estériles reduciría significativamente o custo da produción de PHA. A principal forza que pode impulsar o desenvolvemento de PHA a escala industrial será demostrar un mellor desempeño ambiental que os plásticos convencionais. O proxecto H2020 USABLE Packaging propón cadeas de produción de PHA rendibles e sostibles baseadas en sistemas de cultivos microbianos mixtos (MMC polo seu acrónimo en inglés). Os sistemas MMC permiten operar en condicións non estériles e utilizar residuos orgánicos como materia prima, constando, para a produción de PHA, de tres pasos: (i) fermentación anaerobia, na que as materias primas orgánicas complexas son hidrolizadas e fermentan dando luar a ácidos graxos volátiles (AGV); (ii) un cultivo mixto de enriquecemento, no que as bacterias que almacenan PHA selecciónanse impoñendo un réxime de fartura/fame con exceso de nutrientes para garantir o crecemento; e (iii) un paso de acumulación, no que a biomasa enriquecida aliméntase con AGV con limitación de nutrientes para facilitar a acumulación de PHA. Tras a acumulación intracelular, cómpre recuperar o PHA da biomasa enriquecida. O procesamento posterior do PHA consta de varios pasos: en primeiro lugar, a biomasa enriquecida en PHA é concentrada mediante separación física (filtración, centrifugación, sedimentación) e pretratada (quecemento, trituración, conxelación) para mellorar a recuperación do PHA; a continuación, o PHA pode recuperarse mediante dous métodos diferentes: (i) solubilizando a masa non celular de PHA (NCPM) mediante dixestión química, dixestión enzimática, ruptura mecánica ou unha combinación delas ou (ii) solubilizando o PHA mediante extracción con disolventes. Finalmente, e dependendo dos requisitos de calidade, o MATEO SAAVEDRA DEL OSO xiv PHA é purificado mediante redisolución con auga ou etanol, ou mediante un tratamento de branqueo empregando hipoclorito sódico ou peróxido de hidróxeno. A produción de PHA baseada en sistemas MMC aborda os principais inconvenientes dos cultivos puros e permite integrar a produción de PHA no paradigma da economía circular e a recuperación de recursos. Con todo, o seu desenvolvemento a gran escala aínda se enfronta a múltiples retos, desde a variabilidade dos produtos da fermentación anaerobia ata garantir o seu rendemento ambiental. A fermentación anaerobia, na que substratos complexos son convertidos en AGV, é un dos pasos chave na produción de PHA, xa que a composición dos polímeros ou copolímeros e, por tanto, as súas propiedades e aplicacións potenciais veñen determinadas pola corrente fermentada empregada, é dicir, a composición de AGV. A fermentación anaerobia ten unha estequiometría moi variable que depende, entre outras cousas, de (i) o pH, (ii) a composición do substrato e (iii) o tempo de retención hidráulica. Coñecer mellor a variabilidade da estequiometría da fermentación anaerobia permitiría conducir a produción de AGV cara ás composicións desexadas, e os modelos matemáticos poden ser de gran axuda para describir a fermentación anaerobia de residuos orgánicos. Desafortunadamente, os modelos actuais non poden predicir a selectividade e produtividade dos AGV a partir dun substrato dado debido á variabilidade da estequiometría acidogénica que depende, ademais, dos pasos previos de desintegración e hidrólise. As etapas de separación e purificación do PHA, o chamado downstream processing, é un dos principais colos de botella no desenvolvemento dunha produción de PHA rendible e sostible. Concretamente, pode supoñer ata o 50% dos custos de produción e do impacto ambiental, debido ao importante consumo de produtos químicos e enerxía. O desenvolvemento tecnolóxico do downstream processing e o seu escalado supoñen un reto xa que depende fortemente do tipo de polímero, o tipo de cultivo e as propiedades requiridas. A análise de ciclo de vida (LCA) e o cálculo do custo do ciclo de vida (LCC) poden utilizarse para facilitar a toma de decisións durante o desenvolvemento do proceso e seleccionar un método rendible e sostible. Malia a súa importancia na rendibilidade económica e ambiental do proceso en conxunto faltan avaliacións de ciclo de vida do downstream processing da produción de PHA. A produción de PHA baseada en sistemas MMC xa alcanzou un nivel de madurez tecnolóxica (TRL) de 5-6, é dicir, a escala piloto. Impulsar estes procesos emerxentes máis aló require un acompañamento baseado en ferramentas prospectivas para garantir o seu rendemento sostible cando se chegue escala industrial. Á análise de ciclo de vida prospectiva (pLCA) permite ampliar a escala das tecnoloxías emerxentes mediante escenarios de rendemento futuro a escala industrial. Concibir escenarios de desenvolvemento das tecnoloxías emerxentes no futuro facilita a súa optimización ambiental dende o presente. É probable que aumente o interese polos PHA se continúa o seu camiño ata seren plásticos biodegradables de alto rendemento cunha pegada ambiental potencialmente reducida. Para este fin, é necesario implicar e conectar a todas as partes interesadas da cadea de valor. É posible que os produtores de residuos non dispoñan dos recursos ou os coñecementos necesarios para examinar o potencial dos residuos orgánicos como materia prima. Como poden os produtores de PHA explorar as condicións operativas e os substratos sen investigación de laboratorio ou piloto? Como se pode poñer en contacto aos xestores de residuos cos produtores de PHA que teñen necesidade de acadar propiedades específicas? Igualmente, podería haber outras cadeas de valor que poderían empregar estes residuos como materia prima, o que limitaría a súa RESUMO xv dispoñibilidade. En conclusión, a produción de PHA necesita demostrar a súa viabilidade ambiental e económica fronte a outras tecnoloxías de valorización. Demostrar a viabilidade ambiental e económica da cadea de valor de PHA poder non abondar para promover a súa implantación. As cadeas de valor de PHA necesitan demostrar o seu rendemento ambiental a nivel de produto fronte aos seus homólogos comerciais. A maioría dos estudos de LCA sobre a produción de PHA mediante MMC centráronse en analizar o comportamento ambiental de etapas concretas do ciclo de vida (por exemplo, o downstream processing do PHA) ou en comparar os sistemas de MMC con outras tecnoloxías de valorización. Por tanto, excluíron as etapas finais do ciclo de vida dos plásticos: a conformación e a composición, o uso e o final da vida útil. Excluír o final da vida útil non ten en conta os beneficios derivados da biodegradabilidade dos PHA e ignora os impactos ambientais a longo prazo causados nos ecosistemas mariños pola contaminación plástica. O obxectivo desta tese é axudar á toma de decisións no desenvolvemento de novas cadeas de valor de PHA baseadas en sistemas MMC. Integrando o coñecemento dispoñible e diferentes metodoloxías e ferramentas, como a LCA prospectiva ou os modelos matemáticos, pódense superar os retos da produción de PHA. As preguntas de investigación identificadas en relación coa produción e selectividade de AGV por fermentación anaerobia de residuos orgánicos cóbrense no Artigo I. Ademais, este traballo propón un marco de traballo para conectar múltiples niveis de actores na cadea de valor con diferentes intereses, niveis de experiencia, datos dispoñibles e coñecementos. Este marco pode ser a chave para fomentar a economía circular e reducir as dificultades no desenvolvemento tecnolóxico, permitindo unha avaliación temperá do rendemento do proceso cando se selecciona entre un abano de substratos posibles. Este paper proporciona tamén unha ferramenta de deseño asistido por computador, que comprende: (i) unha biblioteca de substratos, incluíndo a súa caracterización e a selección de parámetros cinéticos adecuados, (ii) un modelo metabólico integral que describe matematicamente os mecanismos de desintegración e a variabilidade da estequiometría acidoxénica na mono e cofermentación anaerobia de residuos orgánicos complexos, e (iii) un conxunto de indicadores para interpretar os resultados da simulación e axudar á toma de decisións. Demostrouse a utilidade da ferramenta para conectar a múltiples actores da cadea de valor con diferentes intereses e coñecementos, desde os implicados no deseño dos procesos de acidificación ata os que estudan a viabilidade de valorizar os residuos desde unha perspectiva económica e ambiental. Así pois, este marco permite identificar os colos de botella técnicos e propoñer solucións innovadoras antes de levar a cabo investigacións onerosas en escala de laboratorio ou mesmo piloto e con grandes necesidades de tempo. O obxectivo do Artigo II é atopar os puntos críticos desde o punto de vista económico e ambiental no downstream processing de PHA e proporcionar ideas para a súa optimización. Seleccionáronse oito procesos alternativos de concentración e purificación de PHA tras unha revisión sistemática dos métodos dispoñibles e a LCA correspondente. Estes procesos avaliáronse desde unha perspectiva tecnoeconómica e ambiental, valorando as posibilidades de escalado. Os procesos baseados en disolventes necesitan grandes cantidades dos mesmos ademais dun elevado consumo de enerxía. Neste contexto, a integración da recuperación de PHA dentro de biorrefinarías nas que se producen excedentes de vapor ou calor e o uso dos chamados disolventes verdes poden diminuír o impacto ambiental ata un 50%. A ruptura mecánica sinalouse como o método de recuperación de PHA máis prometedor tanto desde o punto de MATEO SAAVEDRA DEL OSO 4 Although PHA present advantages over oil-based plastic, their market expansion is still hindered by their high production cost and uncertain environmental performance (Tan et al., 2021). Demonstrating a better environmental performance than conventional plastics would foster PHA implementation (Yadav et al., 2020). High energy requirement during feedstock cultivation, sterilization in pure culture fermentation and PHA downstream processing were reported as the main hotspots from both environmental and economic perspectives (Yadav et al., 2020). Using organic wastes as feedstock and non-sterile conditions would significantly reduce the cost of PHA production (Atasoy et al., 2018). The H2020 project USABLE Packaging (USABLE Packaging, 2019), launched in June 2019 and finished in November 2022, aimed to overcome the PHA production drawbacks, building an entirely new value chain from food industry byproducts and wastes, and using innovative PHA production routes. 1.2. PHA PRODUCTION BASED ON MIXED MICROBIAL CULTURE SYSTEMS USABLE Packaging proposes cost-effective and sustainable production routes, such as the one based on mixed microbial culture (MMC) systems. MMC systems enable not only operating under non-sterile conditions and using organic wastes as feedstocks (Mannina et al., 2020). MMC systems comprise a 3-step process (Nguyenhuynh et al., 2021) (Figure 1.1). i. anaerobic fermentation, where complex organic feedstocks are hydrolyzed and fermented into volatile fatty acids (VFA). ii. MMC enrichment, where PHA-storing bacteria are selected by imposing a feast/famine regime under nutrient (nitrogen, phosphorus, and micronutrients) excess for ensuring growth. iii. PHA accumulation, where enriched biomass is fed VFA under nutrient limitation until sufficient PHA has been accumulated. Figure 1.1 Cradle-to-gate flowchart of PHA production based on MMC systems. PHA production based on MMC system has been tested at lab-scale in a wide range of substrates, from wastewater and organic fraction of the municipal solid waste (OFMSW) to food industry byproducts and side streams such as cheese whey or mussel canning wastewater (Sabapathy et al., 2020). The development of MMC-PHA production has gone a step further during the last eight years by proving its feasibility at pilot scale (Table 1.1). These pilot-scale studies employed mostly industrial wastewater (Bengtsson et al., 2017; Morgan-Sagastume et al., 2015; Werker et al., 2018), OFMSW and sewage sludge (Conca et al., 2020; Lorini et al., 2022; Moretto et al., 2020; Valentino et al., 2019, 2018) and fruit waste as feedstocks (Matos et al., 2021a, 2021b; Silva et al., 2022). Anaerobic fermentation MMC enrichment PHA accumulation PHA downstream processing Organic waste MMC-PHA production VFA-rich PHA-enriched biomass Enriched biomass VFA-rich PHA powder CHAPTER 1: INTRODUCTION 5 Table 1.1. Summary of peer-review studies on PHA production by MMC at pilot scale, presented in chronological order. Title Reference Integrated production of polyhydroxyalkanoates (PHAs) with municipal wastewater and sludge treatment at pilot scale (MorganSagastume et al., 2015) A process for polyhydroxyalkanoate (PHA) production from municipal wastewater treatment with biological carbon and nitrogen removal demonstrated at pilot-scale (Bengtsson et al., 2017) Organic fraction of municipal solid waste recovery by conversion into added-value polyhydroxyalkanoates and biogas (Valentino et al., 2018) Consistent production of high quality PHA using activated sludge harvested from full scale municipal wastewater treatment – PHARIO (Werker et al., 2018) Pilot-scale polyhydroxyalkanoate production from combined treatment of organic fraction of municipal solid waste and sewage sludge (Valentino et al., 2019) Pilot-scale production of poly-3-hydroxybutyrate-co-3-hydroxyvalerate from fermented dairy manure: Process performance, polymer characterization, and scale-up implications (Guho et al., 2020) Biopolymers from urban organic waste: Influence of the solid retention time to cycle length ratio in the enrichment of a mixed microbial culture (MMC) (Moretto et al., 2020) Mixed-culture polyhydroxyalkanoate (PHA) production integrated into a foodindustry effluent biological treatment: a pilot-scale evaluation (MorganSagastume et al., 2020) Long-term validation of polyhydroxyalkanoates production potential from the sidestream of municipal wastewater treatment plant at pilot scale (Conca et al., 2020) Combined strategies to boost polyhydroxyalkanoate production from fruit waste in a three-stage pilot plant (Matos et al., 2021a) Sludge retention time impacts on polyhydroxyalkanoate productivity in uncoupled storage/growth processes (Matos et al., 2021b) An integrated process for mixed culture production of 3-hydroxyhexanoate-rich polyhydroxyalkanoates from fruit waste (Silva et al., 2022) Sewage sludge as carbon source for polyhydroxyalkanoates: a holistic approach at pilot scale level (Lorini et al., 2022) The high organic content (175 g COD·L-1) points out fruit waste as a promising substrate for MMC systems, while the wide availability of both OFMSW and sewage sludge may foster its deployment at industrial scale. Even though MMC systems cannot reach the same cells density or PHA content in biomass as pure cultures (Mannina et al., 2020), their overall yield of PHA on substrate, which varies from low to high values (0.17-0.55 g COD-PHA per g COD-VFA-1) seems promising. 1.3. PHA DOWNSTREAM PROCESSING PHA downstream processing is one of the main bottlenecks in the development of a costefficient and sustainable PHA production (Dietrich et al., 2017). Concretely, it can account for up to 50% of the production costs and environmental impacts (Pérez-Rivero et al., 2019), largely due to chemicals and high energy consumption (Heimersson et al., 2014; Narodoslawsky et al., 2015). PHA downstream processing consists of several steps (Figure 1.2). PHA-enriched biomass may be concentrated by physical separation (filtration, centrifugation, sedimentation) and pretreated (heating, grinding, freezing) to enhance the PHA recovery (Mannina et al., 2020; Pérez-Rivero et al., 2019). PHA can be recovered by two different methods: MATEO SAAVEDRA DEL OSO 6 i. solubilizing the non-cellular PHA mass (NCPM) through chemical digestion, enzymatic digestion, mechanical disruption or a combination of them. PHA is recovered by physical separation, i.e. filtration and centrifugation. ii. solubilizing the PHA through solvent extraction. NCPM is separated from PHA-rich solvent using either filtration or centrifugation. PHA is precipitated afterwards by employing an antisolvent and recovered through filtration/centrifugation. Figure 1.2 Common steps within PHA downstream processing. Depending on the quality requirements, PHA is purified by redissolution with water or ethanol, or by a bleaching treatment employing sodium hypochlorite or hydrogen peroxide (Koller et al., 2013; Mannina et al., 2020; Pérez-Rivero et al., 2019). Unlike the PHA production, the PHA downstream processing from MMC biomass has not been widely proven at pilot-scale and thus, hinders PHA development (Werker et al., 2020). 1.4. CHALLENGES FOR PHA PRODUCTION BASED IN MMC SYSTEMS MMC-PHA production systems tackle the main drawbacks of PHA production based on pure cultures and integrates PHA production into circular economy and resource recovery frameworks. However, its implementation at large scale still faces multiple challenges, from the acidogenic fermentation complexity to the validation of its environmental performance, as detailed below. 1.4.1. Anaerobic fermentation complexity Anaerobic fermentation is one of the key steps in the PHA production, as the copolymer blend composition and resultant properties and potential applications are determined by the fermented stream employed, i.e. the VFA composition. The acidogenic fermentation has a very variable stoichiometry (Domingos et al., 2017; Zhou et al., 2018) which depends on: i. the pH, which can alter radically the obtained VFA composition (Bevilacqua et al., 2021; Jin et al., 2019; Lu et al., 2020) ii. the substrate composition (carbohydrates, proteins, lipids, inert) can modify metabolic pathways and determines the VFA composition (Alibardi and Cossu, 2016; Bevilacqua et al., 2021) iii. the hydraulic retention time (HRT), which affects the hydrolysis degree of compounds with different hydrolysis rates and thus, the composition of the acidified substrate (Atasoy et al., 2018; Jankowska et al., 2018; Strazzera et al., 2018). Also, high retention times are associated with chain elongation reactions that modify the VFA spectrum obtained. Biomass separation PHA-enriched biomass Pretreatment PHA recovery PHA separation PHA powder Purification CHAPTER 1: INTRODUCTION 7 Having a better understanding of anaerobic fermentation stoichiometry variability would enable driving VFA production towards desired compositions (Regueira et al., 2021). However, laband pilot research is expensive and time consuming (Varghese et al., 2022), so the use of mathematical models can be the key to assist the decision making during the development of MMC systems. Mathematical models can be a valuable tool to screen a large set of organic waste as substrates for VFA production, identify the optimal conditions and propose solutions (Regueira et al., 2020a).Unfortunately, current models cannot forecast the selectivity and productivity of VFA from a given substrate due to the variability of acidogenic stoichiometry which depends, itself, on previous disintegration and hydrolysis steps (Regueira, 2020). Most published mathematical models for MMC consider a fixed stoichiometry (Alexandropoulou et al., 2018; Bai et al., 2017) and focus on methane production. Only recently models at metabolic level predict the VFA composition depending on substrate and operational conditions, hence predicting the effect of pH on the VFA selectivity in glucose mixed culture fermentation (MCF) (González-Cabaleiro et al., 2015), protein MCF (Regueira et al., 2020b) and cofermentation with sugars (Regueira et al., 2020a). These models focused on the MCF stoichiometry for model substrates and can envisage the actual VFA production in monoor cofermentation of substrates ready to be acidified, e.g., glucose or casein. Nevertheless, organic wastes are composed of different fractions that undergo, at different rates, several transformation steps before being available for acidification. Other difficulties in modelling the MCF real substrates are (i) the characterization of organic substrates and their representation as model variables; and (ii) different (and unknown) disintegration and hydrolysis rates that change the substrate composition capable of being acidified. Therefore, new models that address all these challenges are required to enable the valorization of organic wastes into VFA. 1.4.2. Cost-efficient and sustainable PHA downstream processing The development and scaling up of PHA downstream processing pose a challenge due to uncertainties regarding the feedstock, polymer type, culture type and required properties. Life cycle assessment (LCA) and life cycle costing (LCC) may be used to assist the decision making during the process development and select a cost-efficient and sustainable method. However, there is a lack of environmental and economic assessments of PHA downstream processing. Most available LCA and LCC studies focus on assessing the feasibility of PHA production in comparison to oil-based plastics (Akiyama et al., 2003; Yu and Chen, 2008), whereas recent studies have concentrated on comparing PHA production from different feedstocks (Kachrimanidou et al., 2021; Kookos et al., 2019; Morgan-Sagastume et al., 2016). To the best of my knowledge, only three LCA studies have focused on the PHA downstream (Fernández-Dacosta et al., 2015; López-Abelairas et al., 2015; Righi et al., 2017), although their scope is concentrated in specific cases, e.g., PHA downstream processing from MMC using wastewater as feedstock. Therefore, there is a gap regarding the upscaling and environmental and economic assessment of PHA recovery processes. 1.4.3. Sustainable development of waste-to-PHA biorefineries As previously stated in section 1.2, PHA production based on MMC systems has already achieved a technology readiness level (TRL) of 5-6, i.e. pilot-scale (Moretto et al., 2020; MATEO SAAVEDRA DEL OSO 8 Morgan-Sagastume et al., 2020; Silva et al., 2022). However, further steps for deploying these emerging processes are challenging. The use of life cycle-based tools can ensure that environmental guidance is included while major alterations are still possible (Arvidsson et al., 2018). Existing LCA studies on the production of PHA by MMC evaluated the environmental benefits of integrating the PHA production within urban (Morgan-Sagastume et al., 2016) and industrial wastewater treatment plants (Fernández-Dacosta et al., 2015; Roibás-Rozas et al., 2020). Morgan-Sagastume et al. (2016b) compared four alternative configurations that integrate PHA production with the benchmark wastewater treatment configuration. Likewise, RoibásRozas et al. (2020) compared the benchmark scenario for saline wastewater treatment and a circular approach that integrates PHA production and anaerobic digestion. Fernández-Dacosta et al. (2015) assessed different PHA recovery technologies (chemical digestion using alkalisurfactant and surfactant-hypochlorite, and solvent extraction using dimethyl carbonate) from wastewater. These studies were based on lab-scale data and low TRL technologies which leads to very uncertain projections on the environmental performance of future PHA production. Prospective LCA (pLCA) enables the upscaling of emerging technologies using scenarios of future performance at industrial scale, and the comparison of the future process with the industrial processes (Cucurachi et al., 2018). However, its prospective character implies, to some degree, lack of data and considerable uncertainty (Igos et al., 2019). Thus, deriving scenarios on how these technologies will develop in the future is a feasible approach (Arvidsson et al., 2018; Bergerson et al., 2020). Langkau et al. (Unpublished work) proposed a scenario methodology framework on biobased products, which enables the systematic and documented development of scenarios. Developing scenarios projecting how waste-to-PHA biorefineries could develop in the future considering both foreground (extraction yield, VFA production yield, etc.) and background (environmental policies, renewable energies share, etc.) parameters would facilitate their environmental optimization. 1.4.4. Connecting and engaging stakeholders Even if technical viability has been demonstrated for various organic wastes at pilot scale, a better understanding of the value chain and its different stakeholders is required. The availability of certain organic wastes as feedstock might constrain the supply chain. For instance, if the OFSMW collected separately yearly in the EU (about 38.8 Mt in 2018) was employed to produce PHA, about 394 kt PHA could be obtained per year (Estévez-Alonso et al., 2021). Only 2% of the current production scale and demand of plastics in packaging applications in the EU could be covered from this hypothetical supply chain. However, OFMSW is just a fraction of the organic wastes produced in the EU. Hence, it is important to develop tools for mapping both feedstock availability and competitive uses. The interest for PHA is likely to grow as high-performance biodegradable plastics with a potentially reduced environmental footprint. (Estévez-Alonso et al., 2021). However, to implement an innovative value chain it is required to engage and connect all stakeholders among the value chain. Waste producers may not have the resources or the knowledge to screen the potential of the organic wastes to be used as feedstock and, therefore disregard interest productive routes for their streams. PHA manufacturers may not have the time or the ability to screen a large set of potential feedstocks and the optimal operational conditions and, thus ignore the chance to find cheaper and more sustainable substrates. These constrains and related CHAPTER 1: INTRODUCTION 9 consequences could be bridge if the stakeholders were connected across the value chain. Thus, the only way to evaluate the value chain is to involve and integrate every stakeholder. 1.4.5. Validating the environmental performance of whole PHA value chain Demonstrating a better environmental performance when comparing products with than their commercial counterparts would be the largest driver for the deployment of PHA production based on MMC systems. Most LCA studies on PHA production by MMC focused on analyzing the environmental performance of particular stages of the life cycle (e.g., PHA downstream processing) or comparing MMC systems with other valorization technologies. Therefore, they excluded the final stages of the bioplastic life cycle: namely, the shaping and compounding, the use, and the end-of-life (EoL). Even if the packaging use normally adds negligible burdens to the life cycle impacts, as it does not involve significant energy consumption nor emissions to the environment (Nessi et al., 2021) excluding the EoL disregards the benefits derived from PHA biodegradability and ignores the long-term environmental impacts caused in marine ecosystems by plastic pollution (Roibás-Rozas et al., 2022). Recent developments on plastic leakage and microplastics modelling (Quantis, 2020), and impact methods (Corella-Puertas et al., 2022) and characterization factors (Maga et al., 2022) make possible the inclusion of the EoL phase in the LCA of plastics and bioplastics (even though these methodologies are at early development level). 1.5. THESIS STRUCTURE AND OBJECTIVES This thesis addresses the principal challenges for the deployment of new PHA value chains based on MMC systems by integrating knowledge from different methodologies and tools: mathematical modelling, (prospective) life cycle assessment and life cycle costing (Figure 1.3). It was conducted in the timespan of three years with the financial support of the USABLE Packaging project, which aimed to unlock the potential of new packaging value chains from organic wastes and industrial side streams. The hypothesis of this thesis is that the main bottlenecks and constraints for deploying new PHA value chains based on MMC system can be overcome through the knowledge integration of different tools. If confirmed, stakeholders can apply these tools in the decision-making during the development phases when there are still chances for major alterations. Thus, the main objectives of this research are: i. Covering the identified gaps regarding the production and selectivity of VFA by anaerobic fermentation of organic wastes and providing a framework for screening organic wastes potential to be converted into VFA (Paper I). ii. Finding the economic and environmental hotspots in the PHA downstream processing and providing insights for its optimization (Paper II). iii. Providing insights on how waste-to-PHA biorefineries could develop in the future and identifying which are they key parameters for their environmental performance (Paper III). iv. Developing the framework (Paper II) for the holistic assessment (technical, environmental, and economic) of organic wastes valorization within carboxylate platform that connects stakeholders and assists the decision-making (Paper IV). MATEO SAAVEDRA DEL OSO 10 Figure 1.3 Thesis structure and objectives. Feedstock production PHA production PHA downstream processing Shaping and compounding Use EoL Zoom in Zoom out Cost-efficient and sustainable PHA downstream processing Paper I Anaerobic fermentation complexity Covering anaerobic fermentation identified gaps and providing a framework for screening wastes Finding economic and environmental hotspots and providing insights for optimization Paper II Paper V Sustainable development of waste-to-PHA biorefineries Connecting and engaging stakeholders Demonstrating a better environmental performance at product level and integrating plastics long-term environmental impacts Expanding the framework for the holistic assessment of organic wastes valorization Paper IV Projecting how waste-to-PHA biorefineries could develop in the future and identifying the key parameters Paper III Proving the sustainability of PHA production based on MMC at product level CHAPTER 1: INTRODUCTION 11 v. Proving the PHA production based in MMC systems sustainability at product level by comparing it with its oil-based commercial counterparts and integrating long-term environmental impacts caused by microplastics (Paper V). This thesis is structured as a compendium of scientific papers that aim to assist the decision level at different granular levels within the PHA value chain (Figure 1.3). The first part of the thesis, comprised by chapters 1 and 2, provides an introduction and a methodological overview of this research: i. Chapter 1 provides an overview of the PHA production by MMC systems and the main challenges for its deployment at industrial scale. ii. Chapter 2 describes the tools and methodologies employed in the knowledge integration: mathematical modelling and life cycle-based tools (LCA, pLCA and LCC). Then, Chapter 3 includes the Scientific Publications (Paper I-V) derived from this thesis: i. Paper I proposes a framework implemented as a computer-aided design tool for screening organic wastes potential to be valorized as VFA. Based on mathematical modelling of MMC fermentation at metabolic level, it t covers the gap regarding anaerobic fermentation stoichiometry variability and connects stakeholders within the carboxylate platform. ii. Paper II examines the sustainability of the PHA downstream processing. It reviews the existing processes for PHA recovery, upscales and evaluates their environmental and economic performance. It provides insights for its optimization by identifying the environmental and economic hotspots within PHA downstream processing. iii. Paper III investigates how waste-to-PHA biorefineries could develop in the future. By deriving future scenarios considering both background and foreground parameters, it covers all possible scenarios for this process deployment at industrial scale. Thus, it enables the identification of the most sensitive parameters for environmental performance. iv. Paper IV expands the framework developed in Paper II and enables the holistic assessment of different valorization pathways for organic wastes within the carboxylate platform. The implemented tool follows a modular approach that enables its usage even for non-expert practitioners. vi. Paper V proves and evaluates the environmental performance of the PHA production based on MMC at product level. It upscales the processes carried out at pilot scale within the USABLE Packaging project and compares the environmental performance of the PHA-based prototypes with their commercial counterparts. Besides, it integrates the long-term environmental impacts caused by microplastics generated during the manufacturing and mismanaged EoL. Finally, the main outcomes and impact of the research are discussed (Chapter 4) and the main conclusions are summarized (Chapter 5). 2 MATERIALS & METHODS SUMMARY In this chapter, an overview of the materials & methods used in this research is presented. It this divided into three sections, covering the mathematical modelling of anaerobic fermentation, the upscaling frameworks followed to project industrial scale scenarios and the life cycle-based methods. MATEO SAAVEDRA DEL OSO 20 estimated considering a Lang Factor of 5.03 (characteristic from a solids-fluids processing plant) and the Chemical Engineering Plant Cost Index (CEPCI) for year 2021. The annual depreciation (AD) was calculated according to Eq. 2.6., assuming an interest rate (i) of 5% and a payback time (PB) of 20 years. The total annual costs (TAC) were calculated according to Eq. 2.7, where utilities (U) and materials (m) costs were estimated based on the LCI and costs data (Seider et al. 2016) while maintenance costs (M) and labor costs (L) were assumed to account for the 3% and 10% of the total capital investment respectively. 𝐴𝐷 = 𝐶𝑇𝐶𝐼 ·[𝑖(1 + 𝑖)𝑃𝐵] [(1 + 𝑖)𝑃𝐵 − 1]⁄ Eq. 2.6 𝑇𝐴𝐶 = 𝐴𝐷 + 𝑈 + 𝑚 + 𝐿 + 𝑀 Eq. 2.7 3 SCIENTIFIC PUBLICATIONS DERIVED FROM THIS THESIS SUMMARY In this chapter, the five scientific papers derived from the thesis are included. Three of them are already published and the remaining two are under revision at the time of writing the present manuscript. 23 PAPER I FOSTERING THE VALORIZATION OF ORGANIC WASTES INTO CARBOXYLATES BY A COMPUTER-AIDED DESIGN TOOL Waste Management, 2022, 142, 101-110 Mateo Saavedra del Oso, Alberte Regueira, Almudena Hospido, Miguel Mauricio-Iglesias CRediT author statement Mateo Saavedra del Oso: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Software; Supervision; Visualization; Roles/Writing - original draft; Writing - review & editing. This is an open access article distributed under the terms of the Creative Commons CC-BY license, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. https://www.elsevier.com/about/policies/copyright/permissions Waste Management 142 (2022) 101–110 Available online 17 February 2022 0956-053X/© 2022 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Country report Fostering the valorization of organic wastes into carboxylates by a computer-aided design tool Mateo Saavedra del Oso a , * , Alberte Regueira a , b , c , Almudena Hospido a , Miguel Mauricio-Iglesias a a CRETUS, Department of Chemical Engineering, Universidade de Santiago de Compostela, 15782 Santiago de Compostela, Spain b Center for Microbiology Ecology and Technology (CMET), Ghent University, Coupure Links 653, Ghent 9000, Belgium c CAPTURE (www.capture-resources.be), Coupure Links 653, Ghent 9000, Belgium ARTICLE INFO Keywords: Early-stage design Aided decision making Resource recovery VFA production Mathematical modelling ABSTRACT The carboxylate platform has the potential to constitute an outstanding opportunity for converting organic wastes into chemicals and other value-added products within a circular economy framework. However, its development is still hampered by technological and financial constraints due to difficulties at forecasting the carboxylates yields by different wastes. This work provides a framework that can be the key to foster circular economy and bridge the development risks, allowing early-stage evaluation of process performance. This framework, which is implemented as a computer-aided design tool, is comprised by: (i) a library of substrates including their characterization and appropriate kinetic parameter selection, (ii) an integral kinetic and stoichiometric model which solves both identified gaps regarding the disintegration mechanisms and the acidogenic stoichiometry variability in the anaerobic mono and cofermentation of complex organic wastes, and (iii) a set of indicators to interpret simulation results and assist the decision making; and presents a showcase of applications supported by two case studies. These case studies show that the optimal conditions to steer VFA spectrum towards odd-chain VFA in MCF of regrind pasta are neutral pH (6.5–7) and a relatively low HRT (3–4 days), while cofermentation of tuna canning wastewater and regrind pasta follows interactive mechanisms that cannot be captured by a “naïve approach”, i.e. by adding up the individual contributions. Finally, it is discussed how value chain actors with different interests can benefit from the proposed tool: identifying technical, economic, and environmental bottlenecks, and proposing innovative solutions prior to costly lab research and piloting. 1. Introduction Growing attention to resource recovery is reflected in the new circular economy action plan, adopted by the European Commission in March 2020 (European Commission, 2020). A biorefinery approach would allow the transition towards circular upcycled value chains, thereby contributing to the sustainability growth goals (Teigiserova et al., 2019). Given its growing market demand and wide range of possible substrates, as well as a cost-effective and environmentally friendly approach, the carboxylate platform is gaining attention in the resource recovery framework (Ramos-Suarez et al., 2021). In this biorefinery approach, waste, side streams and wastewaters are initially converted into volatile fatty acids (VFA), organic solvents (e.g., ethanol) and hydrogen through mixed culture fermentation (MCF) (Naresh Kumar et al., 2022). These intermediates can be further purified or employed as a platform for the obtention of a wide range of high-value products with multiple applications, such as biofuels, chemicals, pharmaceuticals or bioplastics, in particular polyhydroxyalkanoates (Atasoy et al., 2018). H2020 project USABLE Packaging (H2020 USABLE Packaging, 2019) aims to overcome the polyhydroxyalkanoates (PHA) production drawbacks, building an entirely new value chain from food industry byproducts and wastes, and using innovative PHA production routes. The first step in the conversion of organic wastes to PHA is an acidogenic fermentation yielding a mixture of VFA. The performance of the PHA depends highly on the obtained VFA composition. The advantage of this Abbreviations: CH, Rapidly biodegradable carbohydrates; COD, Chemical oxygen demand; CSTR, Continuous stirred tank reactor; I, Inert; LCFA, Long chain fatty acids; LI, Lipids; MCF, Mixed culture fermentation; PHA, Polyhydroxyalkanoates; PR, Proteins; sCH, Slowly biodegradable carbohydrates; VFA, Volatile fatty acids. * Corresponding author. E-mail address: [email protected] (M. Saavedra del Oso). Contents lists available at ScienceDirect Waste Management journal homepage: www.elsevier.com/locate/wasman https://doi.org/10.1016/j.wasman.2022.02.008 Received 22 November 2021; Received in revised form 25 January 2022; Accepted 3 February 2022 Waste Management 142 (2022) 101–110 102 first acidification step is that MCF can handle the complexity and variability of complex organic substrates such as food wastes and has both economic and environmental advantages (Naresh Kumar et al., 2022), e. g., MCF allows continuous operation at non-sterile conditions and thus, reducing substantially the process operating costs. In contrast, a major technical challenge of the carboxylate platform is that acidogenic fermentation has a very variable stoichiometry (Bhatia and Yang, 2017; Domingos et al., 2017; Zhou et al., 2018), depending on its design parameters, i.e., (i) the pH, which can change radically the obtained VFA spectrum (Bevilacqua et al., 2021a; Jin et al., 2019; Lu et al., 2020); (ii) the substrate composition (carbohydrates, proteins, lipids, inert) can modify metabolic pathways and determines the VFA composition (Alibardi and Cossu, 2016; Bevilacqua et al., 2020; Yin et al., 2016) and, (iii) the hydraulic retention time (HRT), which affects the hydrolysis degree of compounds with different hydrolysis rates and thus, the composition of the acidified substrate (Atasoy et al., 2018; Jankowska et al., 2018; Strazzera et al., 2018). The use of computer-aided design tools, such as mathematical models for predictive, environmental, and techno-economic assessment, can be the key to facilitate process development decisions. Concretely, it is possible to assess a large set of alternatives, identify technical bottlenecks and propose innovative solutions prior to expensive lab research and piloting (Varghese et al., 2022). Mathematical models can be a valuable tool to screen complex organic wastes as potential VFA producers. However, current models cannot predict the selectivity and productivity of VFA from a given substrate due to the variability of acidogenic stoichiometry which depends, itself, on previous disintegration and hydrolysis steps. Most published mathematical models for open mixed cultures in anaerobic conditions are characterized by fixed stoichiometry (Alexandropoulou et al., 2018; Bai et al., 2017) and in many cases focus on methane production. Only recently, models at metabolic level have provided predictive capabilities on the VFA stoichiometry in function of substrate and operational conditions, hence predicting the effect of pH on the VFA selectivity in glucose MCF (Gonz´ alez-Cabaleiro et al., 2015), protein MCF (Regueira et al., 2020b) and cofermentation with sugars (Regueira et al., 2020a). These models focus on the MCF stoichiometry for model substrates and can predict the actual VFA production in monoor cofermentation of substrates ready to be acidified, e.g., glucose or casein. However, real substrates are composed of different fractions that undergo, at different rates, several transformation steps before being available for acidification. Other challenges in modelling the MCF real substrates are (i) the characterization of organic substrates and their representation as model variables; and (ii) different (and unknown) disintegration and hydrolysis rates that change the substrate composition capable of being acidified. Therefore, new models that address all these challenges are required to enable the valorization of organic wastes into VFA. The current work covers the identified gaps regarding the production and selectivity of VFA by acidogenic fermentation of organic wastes. Additionally, our contribution goes beyond and proposes a framework to connect multiple levels of actors in the value chain with different interests, expertise levels, available data and knowledge. The development of the framework is described in Section 2: (i) substrate library, (ii) integral kinetic and stoichiometric model, and (iii) set of indicators of MCF performance. Section 3 presents its implementation as a computeraided design tool as well as a showcase of its functionality. Then, the limitations and applicability of the framework within the carboxylate value chain are discussed in Section 4, and final remarks and conclusions are provided in Section 5. 2. Framework development The framework developed in this section aims to cover a gap in the early-stage development of the carboxylate platform and includes: (i) a catalogue of complex organic wastes that includes their characterization and representation as model variables; (ii) a mathematical model which allow the MCF of different complex organic substrates and captures the variability on acidification stoichiometry; and (iii) a set of indicators that can be employed in assisting decision making in the early-stage design of MCF processes. 2.1. Substrate library The substrate library characterizes several real substrates as variables that are represented in a mathematical model, i.e., (i) the substrate composition, i.e., primary data referred to moisture, proteins (and aminoacids), carbohydrates, lipids and ash content; (ii) the substrate chemical oxygen demand (COD) fractionation, i.e., the characterization of substrate chemically degradable fractions in terms of COD, and (iii) the description of the disintegration kinetic behavior. Basic composition data, i.e., moisture, carbohydrates, proteins, lipids, and ash content, provided by USABLE Packaging partners or extracted from food databases is employed as input to obtain the substrate fractionation factors on COD (f i [=] g i-COD⋅ g total-COD -1 ) for particulate (f Xc ) and soluble compounds (f S ), carbohydrates (f CH and f sCH ), proteins (f PR ), lipids (f LI ) and inert (f I ). Given their different hydrolysis rates into sugars (García-Gen et al., 2015), carbohydrates are decoupled into rapidly biodegradable (CH, e.g., starch) and slowly biodegradable (sCH, e.g., cellulose). Their fractions on COD basis (f CH and f sCH ) are calculated assuming a carbohydrates COD factor of 1.18 g COD/g, while protein fraction (f PR ) is calculated according to their AA composition. As some substrates could already contain most of its organic matter partially or totally solubilized, COD is also fractionated into particulate (f Xc ) and soluble (f S ) components. Both fractions are assumed to have the same carbohydrates, proteins, and lipids composition, i.e., a particulate substrate which is partially solubilized would have the same proportion of carbohydrates/sugars or proteins/aminoacids in both particulate and soluble fraction. However, when complex organic substrates such as household food waste are considered, particulate and soluble fractions can have different proportions of carbohydrates/sugars or proteins/aminoacids and thus, a characterization of these fractions may be necessary. Substrate specific kinetic parameters (e.g., kinetic parameter for the disintegration of particulate matter into macromolecule compounds) were gathered through a literature review of anaerobic digestion and MCF studies. For that purpose, the following query string was introduced in Scopus web search engine: ‘TITLE-ABS-KEY (((anaerobic AND digestion) OR (anaerobic AND fermentation) OR (VFA AND production)) AND ((food AND waste) OR (solid AND waste) OR (complex AND organic AND substrates) OR (fruits AND vegetables) OR (wastewater)) AND kinetic)’. Among the document results sort by relevance, those containing information regarding the disintegration of particulate substrates and kinetic parameters were selected. The procedure to choose disintegration kinetic parameters follows a hierarchical order: (1) parameters that are specific to a particular substrate in our library, (2) parameters of substrates with similar composition and physicochemical characteristics, (3) general parameters for complex organic wastes. Besides, substrate library can be easily expanded to any complex organic substrate (if there is no data available) following the method proposed by (Fisgativa et al., 2020) for the characterization of complex organic substrates and their representation as model variables for the anaerobic digestion model N◦1. 2.2. Mathematical model The mathematical model follows an integral approach, i.e., the kinetic model which describes the production of the different VFA (acetate, propionate, butyrate and valerate), hydrogen, methane and other subproducts (e.g., ammonia) during the MCF of complex organic substrates is coupled to a bioenergetic model which predicts the acidogenic stoichiometry in function of substrate composition and operational conditions (Regueira et al., 2021). In subsequent subsections the kinetic M. Saavedra del Oso et al. Waste Management 142 (2022) 101–110 103 model structure and biochemical reaction involved are presented, while the insights on how the bioenergetic model is integrated into the kinetic model are provided later in Section 3.2. 2.2.1. Kinetic model structure and mass balances The kinetic model is built on the mass balances in a continuous stirred tank reactor (CSTR) of the thirty different compounds which make up the states of the model. Two different compartments are considered: reactor bulk and gas phase. The 30 states represent the mass hold-up of different soluble compounds (13), macromolecule compounds (5), particulate compounds (3), gaseous compounds (3) and biomass (6). The mass balances of the system are defined by the following equations (Eqs. (1)–(3)). Soluble, macromolecule and particulate compounds dSi dt =Dliq⋅(Si,inlet −Si)+Ri+Ri,T(1) where S i is the concentration (g COD⋅L -1 ), D liq is the dilution rate of the liquid fraction (d -1 ), R i and R i,T are, respectively, the reaction rate and liquid–gas transport rate of volatile compounds (g COD⋅L -1 ⋅d -1 ). Biomass dXi dt = − Dliq⋅Xi+Ri,ana −Ri,decay (2) where X i is the biomass concentration (g COD⋅L -1 ), R i,ana is the anabolism rate and R i,decay the decay rate. Gas compounds dGi dt = − Dgas⋅Gi+Ri,T(3) where G i is the concentration (g COD⋅L -1 ) and D gas is the dilution rate of the gas fraction (d -1 ). The related concentrations and processes are referred with respect to the headspace volume of the reactor (V gas ). The kinetic model of the reactor is solved as a system of 30 non-linear algebraic equations by integration with MATLAB (R2021a) command ode15s to steady state (pseudo time stepping). Steady state was assumed when all the state absolute derivatives values were under 1⋅10 −12 g COD⋅L −1 ⋅d −1 . 2.2.2. Biochemical processes This kinetic model covers all biochemical reactions from the disintegration of particulate components to the acidogenesis, including also the acetogenesis and methanogenesis to evaluate the risk of part of the COD being converted to methane, possibly decreasing the VFA yield (Fig. 1). Biomass decay, degradation of composites into macromolecules, and their hydrolysis into monomers, which are extracellular reactions, follow a first-order kinetics (Eq. (4)). The considered intracellular biochemical reactions are divided in several sets: (i) acidogenesis from sugars, (ii) acidogenesis from aminoacids, (iii) acetogenesis from long chain fatty acids (LCFA), (iv) acetogenesis from propionate, (v) acetogenesis from butyrate and valerate, (vi) acetoclastic methanogenesis, and (vii) hydrogenotrophic methanogenesis. These reactions are considered to follow Monod kinetics (Eq. (5)), with the inclusion of an inhibition term. Several mechanisms of inhibition and kinetic control were considered: (i) pH inhibition, which affects all intracellular processes (with different severity for acidogens and acetogens, hydrogenotrophic methanogens, and acetoclastic methanogens); (ii) free ammonia and hydrogen inhibition, which affect acetoclastic methanogens and acetogens respectively; (iii) substrate (butyrate and valerate) uptake competition (for C4 biomass); (iv) inorganic nitrogen limitation, which is modelled as a secondary substrate inhibition and affects all uptake reactions. A detailed and comprehensive description of stoichiometric matrix, kinetic rate equations, parameters and inhibition forms is available in Supplementary materials. ρ i=ki⋅Xi(4) ρ i=kmax⋅Si KS+Si ⋅Xi⋅I(5) where ρ i is the specific consumption rate (g COD⋅g COD biomass -1 ⋅d -1 ), ρ max is the maximum specific consumption rate (g COD⋅g COD biomass - 1 ⋅d -1 ), S i is the soluble component concentration (g COD⋅L -1 ), K S is the half-saturation constant (g COD⋅L -1 ), and X i is the biomass concentration (g COD biomass⋅L -1 ), and I is the inhibition term. 2.3. Indicators to support decision-making Mathematical model outputs must be transformed into easy interpretable indicators that provide valuable information to both expert and non-expert user, since these indicators determine the technical, environmental, and economic performance of MCF processes. The following indicators were considered relevant here: Fig. 1. Biochemical reactions of the MCF of complex organic substrates. M. Saavedra del Oso et al. Waste Management 142 (2022) 101–110 104 i) Substrate conversion (%), which measures what percentage of inlet COD is converted into biomass and products and indicates substrate biodegradability. ii) VFA yield (g VFA-COD⋅g feed-COD -1 ) and productivity (g VFACOD⋅L -1 ⋅d -1 ) indicate the ratio of substrate converted to VFA and the amount of VFA produced per volume and time, respectively, and are relevant indicators to reduce capital costs. iii) Selectivity measures the fraction of each VFA within the spectrum (e.g., g COD-acetate⋅g COD-VFA -1 ) for targeted production of a given VFA. iv) As VFA can be employed as precursor for PHA production an additional equivalent hydroxyvalerate (HV) ratio indicator is proposed which measures the molar ratio of precursors of HV, i.e., acetate plus propionate and valerate, within the total precursors of both HV and hydroxybutyrate (HB), i.e., acetate plus acetate and butyrate. 3. Computer-aided design tool utility assisting the process earlydesign Following the proposed framework, a modular tool for computeraided design was implemented. The purpose of this section is to provide the results of the built substrate library (Section 3.1), insights on how these modules are interconnected (Section 3.2), and present a showcase of applications of this computer-aided design tool (Sections 3.3 and 3.4), e.g., screening substrates as potential precursors of VFA that are employed to produce PHA. 3.1. Substrate library The substrate library (Table 1) covers a wide range of substrates with diverse compositions and characteristics. Additionally, it can be easily expanded to any kind of complex organic substrates following the method described in Section 2.1. The substrate library majorly contains fruit and vegetable wastes (spinach, cucumber, tomato, and lettuce), food industry byproducts (regrind pasta, wheat bran and bread crust) and side streams (vinasses and tuna canning wastewater). Kinetic parameters for disintegration were chosen according to criteria establish in Section 2.1: a value of 1.7 d -1 for fruit and vegetable wastes (García-Gen et al., 2015), 0.41 d -1 for wheat byproducts and vinasses (Vavilin et al., 1998), and 1.4 d -1 for fishery wastes (García-Gen et al., 2015). 3.2. Solution strategy: Bioenergetic model integration into the kinetic model To account for the well-known variability of the acidification stoichiometry with respect to substrate composition and operating conditions the yields of acidification products are estimated by a bioenergetic model which determines the most likely metabolic pathways in MCF and has been already validated for the cofermentation of sugars and aminoacids (Regueira et al., 2021). The bioenergetic model is integrated with the kinetic model by providing the stoichiometry of the acidification products given the acidification substrates, i.e., the fraction of the substrate that can be directly acidified and the products of hydrolysis (Fig. 2). For a given substrate (or substrate mixture) and operational conditions (i.e., pH and HRT), in first place, the information related to its fractioning in terms of components (i.e., soluble fraction, rapidly and slowly biodegradable carbohydrates, proteins, lipids, inert and aminoacids molar composition) and its disintegration kinetics is taken from the substrate library. After that, acidogenic stoichiometric coefficients, which do not depend on disintegration and hydrolysis, are determined considering only the pH, substrate(s) aminoacids composition and the actual consumed fraction of sugars and aminoacids (HRT is not considered since its influence on acidogenic stoichiometry is negligible). However, as the different organic fractions comprising the wastes can have different fractioning between soluble and particulate matter, with Table 1 Substrate library comprised by basic characterization of substrates (TS, VS, TCOD) and inputs to the kinetic model (COD fractionation, AA molar fraction composition and disintegration kinetic parameter). Acronyms: S: spinach, C: cucumber, T: tomato, L: lettuce, V: vinasses, RP: regrind pasta, WB: wheat bran, BC: bread crust, TWW: tuna canning wastewater. Substrate S C T L V RP WB BC TWW Gross substrate characterization TS (% ww) 28.0 3.0 4.2 4.0 49.0 94.0 91.0 93.0 1.0 VS (% ww) 28.0 2.7 3.4 3.5 40.0 74.0 74.0 75.0 1.0 tCOD (g COD⋅L -1 ) 35.8 33.3 48.5 43.0 485.1 896.0 941.9 968.2 15.2 Disintegration kinetic parameters k dis (d -1 ) 1.7 1.7 1.7 1.7 0.42 0.42 0.42 0.42 1.4 Substrate fractionation f Xc 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 0.26 f S 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.74 f CH 0.40 0.49 0.43 0.56 0.10 0.78 0.27 0.70 0.00 f PR 0.29 0.26 0.19 0.27 0.09 0.15 0.18 0.10 0.92 f sCH 0.22 0.22 0.11 0.14 0.75 0.02 0.41 0.05 0.00 f LI 0.10 0.04 0.28 0.04 0.07 0.05 0.14 0.15 0.08 Aminoacid composition Arginine (% molar) 5.1 6.5 2.0 5.0 7.1 2.1 2.1 2.1 4.8 Alanine (% molar) 8.8 6.9 4.9 7.5 7.9 4.0 4.0 4.0 9.5 Aspartate (% molar) 1.0 3.1 7.9 5.0 0.9 0.6 0.6 0.6 9.7 Lysine (% molar) 6.6 5.1 3.0 6.9 2.6 0.9 0.9 0.9 8.8 Glutamate (% molar) 3.8 9.8 16.6 4.6 8.4 2.1 2.1 2.1 12.8 Serine (% molar) 5.5 4.9 4.0 4.5 7.8 6.5 6.5 6.5 5.5 Threonine (% molar) 5.7 4.1 3.7 5.9 3.9 3.8 3.8 3.8 5.2 Cysteine (% molar) 2.2 1.2 1.6 2.3 3.0 1.5 1.5 1.5 1.7 Glycine (% molar) 9.9 8.2 4.1 9.4 6.6 12.2 12.2 12.2 9.0 Proline (% molar) 5.4 3.3 2.1 5.0 4.8 15.2 15.2 15.2 4.3 Valine (% molar) 7.6 4.8 2.5 7.2 4.0 3.8 3.8 3.8 6.2 Isoleucine (% molar) 6.2 4.1 2.2 7.7 1.1 3.1 3.1 3.1 4.9 Leucine (% molar) 9.4 5.7 3.1 7.2 2.7 6.2 6.2 6.2 8.7 Methionine (% molar) 2.5 1.3 0.8 1.7 4.7 0.8 0.8 0.8 3.6 Glutamine (% molar) 9.2 24.6 31.3 10.6 15.7 34.9 34.9 34.9 1.4 Asparagine (% molar) 9.0 4.9 8.6 7.9 14.8 1.1 1.1 1.1 1.1 Histidine (% molar) 2.3 1.7 1.5 1.7 4.0 1.4 1.4 1.4 2.7 M. Saavedra del Oso et al. Waste Management 142 (2022) 101–110 105 intrinsic distinct disintegration kinetics, the resulting macromolecules are also hydrolyzed at different rates. All of this makes determining a priori the consumed sugars and AA in the acidification step a nonobvious task. To solve this challenge, a preliminary run of the kinetic model coupled with a mass balance function which determines the net consumption of monomers (glucose and aminoacids) to the bioenergetic model is carried out. When only one substrate is fermented, this function calculates the net consumption of monomers in the MCF. If two substrates are fermented, then this function estimates the corresponding net fractions of consumed AA, and after that, determines the AA mixture that is acidified in MCF (an example is included in Fig. 3). Once the function calculates the corresponding net fractions of consumed monomers, the yields of acidification products are estimated by the bioenergetic model. To determine the product spectrum, the bioenergetic model main hypothesis states that the most efficient microorganisms harvesting energy from the substrate are dominant in an anaerobic mixed microbial culture in a CSTR. Namely, microorganisms employ the metabolic branches that yield the maximum net ATP, i.e., the sum of the ATP produced in catabolism and the ATP spent in the active transport of compounds, ensuring a neutral net electron balance (Table 2), since anaerobic process lack external electron acceptors that could act as electron sinks (Regueira et al., 2020b). Once the process stoichiometry is determined, one function transforms the bioenergetic model outcome into the stoichiometric coefficients needed by the kinetic model and loads them into the stoichiometric matrix. Finally, the kinetic model is solved again considering the determined stoichiometric coefficients and the concentrations at steady state are transformed into easy interpretable indicator to assist decision-making: substrate conversion, VFA yield, productivity, and selectivity. 3.3. Steering the VFA spectrum by modifying operational conditions: A case study on MCF of regrind pasta Regrind pasta is a byproduct of pasta production, which is currently employed as animal feed. As carbohydrate-rich substrate, it could be a suitable substrate to produce VFA and therefore the application of the developed tool can support the selection of the appropriated values for the operational conditions. A fast evaluation of the range of MCF design parameters, i.e., pH and HRT, was carried out. HRT and pH ranges were from 2 to 6 days and, from 5 to 8 respectively according to MCF literature (Bevilacqua et al., 2020; Domingos et al., 2017; Strazzera et al., 2018). Then, the program was run for these design parameters ranges and model outputs were transformed into a map for each indicator defined in Section 2.3 (Fig. 4). Employing an Intel(R) Core (TM) i58250U (@3.4 GHz) the average time employing per simulated pH and HRT was 400 s. Regrind pasta is mainly composed by rapidly biodegradable carbohydrates (78% of COD) such as starch, which are hydrolyzed at higher rates than other macromolecules. As both disintegration and hydrolysis are the MCF limiting steps and independent to pH, substrate conversion map (Fig. 4a) only shows dependency on the HRT. Substrate conversion map also confirms the good biodegradability degree of regrind pasta. Regarding the VFA yield (Fig. 4b) and productivity (Fig. 4c), both maps show a dependency on the HRT and pH, especially at neutral to basic pH. Operating at a high HRT (>5 days) and a neutral to basic pH (from 6.5 to 8) leads to methanization (see methanization graph enclosed in Supplementary materials), while acidic pH values inhibit the methanogenesis and thus, enhance the VFA yield. Therefore, it is preferable to operate at low HRT, which also implies a higher productivity due to an increased feeding flow rate. With regards to the equivalent HV precursors ratio (Fig. 4d), high values of HV are obtained at neutral pH values. In this region, the VFA spectrum is dominated by both acetate and propionate, which are precursors of HV, and whose production is favored by this pH range. However, highest values, which are obtained at high HRT (>5 days), must not be considered since VFA concentration is rather low due to methanization (see total VFA concentration graph enclosed in Supplementary materials). HV ratio values at acid pH are low due to these conditions favor the butyrate production, which is a precursor of HB. If Fig. 2. Structure of computer-aided design tool developed in this research. M. Saavedra del Oso et al. Chemical Engineering Journal 412 (2021) 127687 Available online 12 November 2020 1385-8947/© 2020 Elsevier B.V. All rights reserved. Evaluation and optimization of the environmental performance of PHA downstream processing M. Saavedra del Oso, M. Mauricio-Iglesias * , A. Hospido CRETUS Institute, Department of Chemical Engineering, Universidade de Santiago de Compostela, 15782 Santiago de Compostela, Spain ARTICLE INFO Keywords: Polyhydroxyalkanoate extraction Biorefinery Biobased materials Process optimisation Life cycle assessment ABSTRACT Biobased and biodegradable materials such as polyhydroxyalkanoates (PHA) have great potential as an alternative for conventional oil-based plastics in consumer goods and medical applications, but their total market share is still marginal due to their high production costs. Downstream processing, with high energy demand and significant requirements in oil-derived solvents and chemicals, has been identified as one bottleneck in the PHA value chain. Hence, a thorough study of the environmental performance of PHA recovery processes is essential to promote their applicability. This work provides valuable insights on PHA downstream processing environmental hotspots and how to optimize them accordingly. Eight PHA downstream alternative processes for both highgrade and low-grade purification are evaluated from a techno-economic and an environmental perspective, assessing scale-up possibilities and challenges. To reach this goal, both scenario definition and process design were supported by a systematic review of available PHA downstream methods and related life cycle assessments. Methods relying on solvent extraction require large amounts of energy for solvent recovery, and thus, their higher performance in impurity removal also entails larger costs and impacts in all categories, when compared to mechanical disruption or chemical digestion. Therefore, solvent extraction is only recommended for those cases where a higher quality is required, or solvents can be reasonably obtained from an integrated biorefinery. Chemical digestion can be optimized by adding a chemicals recovery unit, while mechanical disruption appears to be the most promising technology in terms of environmental performance. Through this technoeconomic and environmental assessment, it is proved that PHAs can be attractive materials for a sustainable bioeconomy if the process and product design incorporate life cycle assessment such as the developed in this work. 1. Introduction The environmental impact derived from the ubiquitous use of plastics has been linked to two of society’s present main environmental concerns: climate change and the spoilage of marine environments caused by improperly disposed plastics. In this situation replacing oilbased non-biodegradable plastics with alternative bio-based plastics that have similar properties could have notable benefits. Among bioplastics, polyhydroxyalkanoates (PHAs) are biobased, biodegradable –even in marine environments by ASTM standards– and biocompatible polymers with multiple applications [1,2]. They can be produced, through bacterial fermentation, from a feedstock other than starch/ glucose such as a diverse range of complex organic substrates including by-products from agriculture and food industry [3–5]. PHAs display thermoplastic properties, which depend on the choice of substrate, bacteria and fermentation conditions. Thus, PHAs are ideal substitutes for conventional oil-based plastics such as polyethylene (PE), polyethylene terephthalate (PET) or polypropylene (PP) [6]. Abbreviations: AD, Annual depreciation; AP, Acidification potential; ATP, Aquatic toxicity potential; CH, Switzerland; COD, Chemical Oxygen Demand; C TCI , Total capital investment; DMC, Dimethyl carbonate; EP, Eutrophication potential; FAETP, Ecotoxicity for aquatic fresh water; FD, Fossil depletion; FE, Freshwater eutrophication; FEX, Freshwater ecotoxicity; FOFP, Photo-oxidant formation potential; FU, Functional unit; GWP, Global warming potential; HT, Human toxicity; HTPE, Human toxicity potential by either inhalation or dermal exposure; HTPI, Human toxicity potential by ingestion; L, Labor costs; LCA, Life Cycle Assessment; LCC, Life Cycle Cost; LCI, Life Cycle Inventory; M, Maintenance costs; m, Materials costs; NCPM, Non-cellular PHA mass; NREU, Non-renewable energy use; ODP, Ozone depletion potential; OFP, Photochemical ozone formation; PE, Polyethylene; PET, Polyethylene terephthalate; PHA, Polyhydroxyalkanoates; PHB, Polyhydroxybutyrate; PP, Polypropylene; RER, Europe; SDS, Sodium dodecyl sulfate; TA, Terrestrial acidification; TAC, Total annual costs; TTP, Terrestrial toxicity potential; U, Utilities costs. * Corresponding author. E-mail address: [email protected] (M. Mauricio-Iglesias). Contents lists available at ScienceDirect Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej https://doi.org/10.1016/j.cej.2020.127687 Received 24 July 2020; Received in revised form 4 November 2020; Accepted 6 November 2020 Chemical Engineering Journal 412 (2021) 127687 2 The largest driver for replacing conventional plastics by PHA would be a better environmental performance [7]. Studies based on life cycle assessment (LCA), a technique that identifies and quantifies the potential environmental impacts associated with a product, process or service [8], have reported that conventional plastics may have lower carbon footprint [9,10] than PHA albeit biodegradable and based on renewable resources. High energy requirement was reported in overall PHA production life cycle, especially during feedstock cultivation, sterilization equipment within PHA fermentation and the PHA downstream processing [2,11]. Furthermore, PHAs production at large scale is still limited due to its high production cost –2.2 to 5 € /kg–compared with conventional oil-based plastic –less than 1 € /kg– [12]. Within PHA production, the downstream processing of PHAs has been reported as a cost and environmental hotspot for PHAs production. Specifically, it can account for up to 50% of the production costs [13], largely due to the use of great amounts of solvents and high energy requirements [11,14]. PHAs downstream processing is commonly comprised by several steps (Fig. 1), starting with the physical separation of the biomass, and probably followed by a pretreatment step to enhance the yield and purity in the extraction step. In the extraction step, PHA can be recovered by two different ways: (1) by solubilizing the non-cellular PHA mass (NCPM) through chemical digestion or mechanical disruption or (2) by solubilizing the PHA through solvent extraction. Then, PHA is separated from the NCPM by basic operations such as precipitation, filtration or sedimentation, or more alternatives for higher-value products such as liquid–liquid extraction or air classification. Finally, in most cases and regardless of the final product requirements, PHA should be purified: redissolution with water or ethanol can be enough, although sometimes a bleaching step with hydrogen peroxide or sodium hypochlorite takes place; additionally, ozone treatment or activated charcoal seems to be effective too in order to obtain a purified PHA [13,15,16]. The objective of this study is to provide valuable insights on PHA downstream processing optimization by finding economic and environmental hotspots for improve its performance. For this endeavor, a literature review of LCA on PHA downstream processing was carried out providing the basis for a techno-economic and environmental assessment of eight potential downstream approaches –both low and high grade PHA purification–. Firstly, the most promising methods for PHA downstream processing were identified and, through a design and scaleup supported by articles and patents, downstream processes were classified as corresponding to low or high grade PHA purification. Then, an in-depth revision of selected LCA of PHA production and downstream processes was carried out to evaluate how the key LCA elements –such as objective, functional unit, system boundaries, impact categories …– were handled. Finally, the economic and environmental performance of the eight selected downstream processes was performed by using life cycle cost (LCC) and LCA, respectively, and best practices guidelines for PHA downstream processes design were proposed. 2. Methods 2.1. Literature review In order to identify the most promising methods for PHA downstream processing, a systematic review of the literature available using Scopus web search engine for articles and Google Patents for patents was performed. To do so, the following keywords were used: ‘biopolymers’, ‘PHA’, ‘polyhydroxyalkanoates’, ‘PHB’, ‘polyhydroxybutyrate’, ‘PHBV’, ‘poly(3–hydroxybutyrate–co–3–hydroxyvalerate’, ‘downstream processing’, ‘extraction’, ‘green solvents’, ‘recovery’, ‘chemical digestion’, ‘isolation’, ‘industrial production’, ‘sustainable’. Among the 200 resulting elements, twelve peer-reviewed articles and patents which collect the methods for PHA recovery and purification as well as technical specifications were screened [2,13,15,17–25], being the primary source in relation to the methods used for scenarios definition and process design, described later in Section 3. For the second literature review focused on LCAs and LCC on PHA production and downstream processes, the following keywords were added: ‘technical assessment’, ‘techno-economic assessment’, ‘life cycle costing’, ‘LCC’, ‘life cycle assessment’, and ‘LCA’ in Scopus web search engine. Thirteen LCA peer-review articles on PHA production and downstream processing [9,10,19,26–35] which include specific data about PHA recovery and purification were screened. 2.2. Process simulation The process simulator Aspen HYSYS –provided by Aspen Technology, Inc., US– was used to estimate the process requirements in terms of fresh solvent and energy duties in heat exchangers and solvent recovery units –mostly evaporators and distillation columns–. Fluid packages were chosen according to Carlson [36] and Elliot and Lira [37] recommendations, selecting the NTRL model for acetone–water, acetoneethanol, ethyl acetate-heptane and isoamyl alcohol-water mixtures. 2.3. Life cycle assessment and life cycle costing LCA is a systematic and standardized methodology which determines the process or product’s environmental impacts in the design and operational phase, and it is comprised by four steps: i) the goal and scope states the intended application, the system, its function and the related functional unit, the system boundaries, the impact categories selected and the impact assessment methodology; ii) the inventory analysis involves data collection and calculation procedures to quantify relevant inputs and outputs of a product system; iii) the impact assessment transforms the inventory results into potential environmental impacts; and iv) the interpretation phase, which involves critical review, determination of data sensitivity, and result presentation [8]. Life cycle costing (LCC), which is one pillar of the full life cycle sustainability assessment, is an economic assessment aligned with LCA in terms of system boundaries, functional unit, and methodological steps, is performed [38]. In general, costs can be divided into capital and operational costs. Regarding the former, delivered equipment costs were Fig. 1. Conventional steps in the PHA downstream processing and most common methods for each step. M. Saavedra del Oso et al. Chemical Engineering Journal 412 (2021) 127687 3 estimated using correlations based on individual equipment’s characteristics size [39]: total capital investment was estimated applying a Lang Factor of 5.03, characteristic from a solids-fluids processing plant, and a 2019 CEPCI index of 607.5; annual depreciation (AD) was calculated –see Eq. (1)– based on the total capital investment (C TCI ) with an interest rate (i) of 5% and a payback time (PB) of 20 years. Concerning the latter, total annual costs (TAC) were estimated using Eq. (2): utilities (U) and materials (m) costs were based on the inventories built and literature information [39], while maintenance costs (M) were assumed to account for the 3% of the total capital investment, and the labor costs (L) were assumed to be 10% of the total annual costs. AD =CTCI ⋅[i(1+i)PB ]/[(1+i)PB −1](1) TAC =AD +U+m+L+M(2) 3. Scenarios definition and process design To define the scenarios studied in this paper, the most promising methods of PHA downstream processing were identified. To do so, twelve peer-review articles and patents were screened by the following criteria: those which include a revision of the state-of-the-art [2,13,15,17,18], and those which include methods with technical specifications applied at lab or pilot scale [19–25]. The processes were designed according to articles and patents guidelines. Other relevant information includes the process background (substrate or type of culture employed for PHA production), the technology readiness level (TRL) and the quality of the PHA obtained (Table 1). Also, solvents selection guides and technologies maturity level were considered when designing solvent-based processes [40,41]. The requisite quality of a recovered PHA is coupled to the intended use of the bioplastic. Most of studies employ only recovered polymer purity and molecular weight as the indicator of the polymer quality [19,20,42]. However, there are other aspects than should be considered such as the consistency of the material quality, and rather than the purity, the chemical impurities that are present –e.g. the immunostimulatory lipopolysaccharide present in most Gram-negative bacteria, which is considered a endotoxin– [43]. Indeed, loss of molecular weight during melt processing is predominantly caused by trace impurities, and thus, a higher molecular weight may not mean a higher quality. In this sense, we define here high-grade PHA as the one used for pharmaceutical and food grade applications. It has to comply with the following requirements: high purity, high molecular weight and chemical compatibility according to the European Commission regulation No 10/2011 related to plastic materials and articles intended to come into contact with food [44]. On the other hand, PHA obtained through processes which do not comply these requirements, especially those related to impurities –e.g. salts and lipopolysaccharide– and substrate employed –e.g. methane and wastewater– were classified as low-grade PHA notwithstanding stringent mechanical and structural property requirements. A brief description of the selected processes for each type, highor low-grade, is presented below plus a summary of key model assumptions (Table 2 and Table 3), while mass balances and detailed schemes along with process specifications and design outcomes are collected in Supplementary information (sections S1.1, S2.1, S3.1 and so on; note that processes H1 and L2 flow diagrams are also included there for completeness). 3.1. Downstream processes for high-grade PHA In process H1 (Fig. 2), which is based on a patent applied by Tepha Inc. US [24], P(4HB) enriched biomass –obtained through fermentation of glucose by Escherichia coli [45]– is extracted by acetone. Before being submitted into extraction, biomass is dried and washed with ethanol to remove its toxins content. After extraction, NCPM is separated by ultrafiltration, and dissolved PHA is recovered by adding an antisolvent –ethanol in this case– until the PHA precipitates. Additionally, P(4HB) is purified through ethanol washing prior to spray drying. The pretreatment and purification step with ethanol along the extraction with acetone, which is a Class 3 solvent –i.e. those solvents with low human toxic potential according to the European Medicines Agency [54]– allows that the obtained P(4HB) powder can be employed in medical and food grade applications. Process H2 is based on NCPM mechanical disruption and chemical digestion. Developed by Bio-On (Italy) as a patent [25], it is already being applied at full scale by them [46]. P(3HB) enriched biomass –produced through microbial fermentation of food industry byproducts– is sent to a high-pressure homogenization where a surfactant is added to cause simultaneously the NCPM disruption and digestion. Then P(3HB) is easily separated from dissolved biomass by a liquid–solid separation. Additionally, P(3HB) is purified by bleaching prior to spray drying. Similarly, process H3 is based on NCPM chemical digestion to extract the copolymer of P(3HB-co-4HB). Based on a study performed at pilot plant scale [47], PHA enriched biomass previously dewatered is sent to a NaOH and Lysol treatment, to digest NCPM. After that, a posttreatment by water washing and a spray drying is carried out. Process H4 is based on a large-scale study [21], where the P(3HB-co3HHx) enriched biomass is obtained by a pure culture of Aeromonas hydrophile that uses only glucose as a carbon source. The downstream process comprises dry biomass solvent extraction with ethyl acetate, heptane precipitation and post-treatment consisting of ethanol washing followed by a spray drying step to obtain the final PHB copolymer. 3.2. Downstream processes for low-grade PHA Process L1 is based on a research scale and a technoeconomic study [5,48], which employs solvent extraction to isolate the copolymer P (3HB-co-4HB). PHA enriched biomass, obtained through microbial fermentation of methane by Methylocystis hirsute, is previously dried prior to hot acetone extraction. After NCPM separation, the stream is refrigerated and water is added to precipitate the PHB. Finally, PHB is Table 1 Processes definition of PHA downstream processes, accounting their quality grade, substrate, type of culture, recovery method, technology readiness level and references in which are based. Grade PHA Feedstock Culture Microorganism Recovery method TRL** Reference High (H1) P(4HB) Glucose Pure Escherichia coli Acetone extraction 9 [22,24,45] High (H2) P(3HB) Food industry byproducts Pure Ralstonia eutropha* HPH +SDS digestion 9 [25,46] High (H3) P(3HB-co-4HB) Oleic acid, ɣ-butyrolactone Pure Cupriavidus sp.* NaOH +Lysol digestion 4 [45,47] High (H4) P(3HB-co-3HHx) Glucose, soybean oil Pure Aeromonas, Wauteria. Ethyl acetate extraction 6 [21] Low (L1) P(3HB-co-4HB) Methane Pure Methylocystis hirsuta Acetone extraction 4 [5,48] Low (L2) P(3HB) Canning industry waste Halophilic bacteria Halomonas boliviensis Osmotic shock +SDS digestion 4 [49-51] Low (L3) P(3HB) Wastewater Mixed culture Not applicable. NaClO +SDS digestion 4 [32,33] Low (L4) P(3HB) Sugar molasses byproducts Pure Alcaligenes eutrophus* Fusel alcohols extraction 8 [23,52] * Ralstonia eutropha, Cupriavidus sp. and Alcaligenes eutrophus are the same microorganism. ** TRL was assigned according to Humbird [53]. M. Saavedra del Oso et al. Chemical Engineering Journal 412 (2021) 127687 4 recovered and spray dried. Process L2 (Fig. 3), based on a patent developed by Repsol S.A. (Spain) [51], takes advantage of halophilic culture characteristics and uses an osmotic shock to disrupt NCPM [49,50]. PHA enriched biomass is partially dewatered and heated prior a recovery step where an osmotic shock combined with SDS digestion is carried out to digest NCPM. Then, PHA is separated from dissolved NCPM and spray dried after a washing step with water. Chemical digestion of the NCPM is the extraction method employed in process L3 to recover P(3HB). Based on a paper of PHA isolation from a mixed microbial culture [32,33], this process uses a chemical digestion which combines the high oxidizing potential of NaClO with a surfactant such as SDS. Then a washing step with fresh water is carried out before spray drying. Process L4 has the peculiarity of employing a byproduct of ethanol Table 2 Summary of key model assumptions for high-grade PHA downstream processes. Process Step Assumptions H1 Biomass treatment 10,000 t of P(4HB) in enriched biomass are recovered for 330 days/year 11.5 kg⋅m −3 biomass concentration with 68% P (4HB) weight content A drying step –less than 5 wt% water content – prior to 95 wt% ethanol washing in a 4:1 ratio ethanol-biomass in R-101 b An additional drying step to reach up to 99 wt% solids content b Extraction A 94% P(4HB) recovery yield c Acetone extraction in a 20:1 acetone-biomass ratio in R-201 b Solvent recovery in a distillation column Recovery and purification 99.7% P(4HB) purity b P(4HB) precipitation with ethanol 95 wt% in a 1:1 ethanol-acetone ratio in R-301 a H2 Biomass treatment 10,000 t of P(3HB) in enriched biomass are recovered for 330 days/year 80 kg⋅m −3 biomass concentration with 70% P (4HB) weight content d A 10 wt% H 2 SO 4 addition obtain a pH 4.5 d Extraction A 83% P(3HB) recovery yield d A high pressure homogenization step (1000 bar) at room temperature d A 50 wt% NaOH addition to obtain a pH 10.0 and a 4 g⋅L -1 SDS aqueous solution addition in R-201 d A water dilution (50% of FT-201 liquid stream is recycled) in a 1:4 stream-water ratio in R-202 d Recovery and purification A 94% P(3HB) recovery yield and 99% purity d 30 wt% H 2 0 2 bleaching treatment in a 1:8 streamH 2 0 2 ratio in R-301 d Water dilution (85% of FT-301 liquid stream is recycled) in a 1:3 stream-water ratio in R-302 d H3 Biomass treatment 10,000 t of P(3HB-co-4HB) in enriched biomass are recovered for 330 days/year 11.5 kg⋅m −3 biomass concentration with 70% P (4HB) weight content ae A dewatering step to reach up to 80 wt% solids content Extraction A 90% P(3HB-co-4HB) recovery yield e A 0.15 M NaOH and 2 vv% Lysol addition in R-201 e Recovery and purification 93% P(3HB-co-4HB) purity e A water dilution in a 1:3 stream-water ratio in R301 H4 Biomass treatment 10,000 t of P(3HB-co-3Hxx) in enriched biomass are recovered for 330 days/year 50 kg⋅m −3 biomass concentration with 50% P (3HB-co-3Hxx) weight content f A drying step –up to 99 wt% solids content – prior to grinding f Extraction A 94% P(3HB-co-3Hxx) recovery yield f Ethyl acetate extraction in a 20:1 ethyl acetatebiomass ratio in R-201 f Solvent recovery in a distillation column Recovery and purification 95% P(3HB-co-3Hxx) purity P(3HB-co-3Hxx) precipitation with heptane in 3:1 heptane-ethyl acetate ratio in R-301 f A P(3HB-co-3Hxx) purification step with ethanol in a 4:1 ethanol-biomass ratio in R–302 f References for assumptions: a-[45], b-[24], c-[22], d-[25], e-[47], f-[21]. Table 3 Summary of key model assumptions for low-grade PHA downstream processes. Process Step Assumptions L1 Biomass treatment 100,000 t of P(3HB-co-4HB) in enriched biomass are recovered for 330 days/year 60.7 kg⋅m −3 biomass concentration with 50% P (3HB-co-4HB) weight content A drying step –less than 5 wt% water content– Extraction A 92% P(3HB-co-4HB) recovery yield b Hot acetone extraction –3 bar and 90 ◦C– in a 9:1 acetone-biomass ratio in R-201 a Solvent recovery in a distillation column Recovery and purification 98% P(3HB-co-4HB) purity a A P(3HB-co-4HB) precipitation step comprised by cooling (E-301) and water addition (R-301) a A separation by microfiltration and spray drying to obtain P(3HB-co-4HB) powder L2 Biomass treatment 1,500 t of P(3HB) in enriched biomass are recovered for 330 days/year 8 kg⋅m −3 biomass concentration with 50% P(3HBco-4HB) weight content c A dewatering step –up to 60 wt% solids content – followed by a heating step (60 ◦C) de Extraction A 98% P(3HB) recovery yield d 0.1 wt% SDS chemical digestion at 60 ◦C –less than 5 wt% solids content– de Recovery and purification 94% P(3HB) purity d Water washing in two step maintaining up to 5 wt % solids content –water from second washing step is recycled into first washing step– A separation by centrifugation and spray drying to obtain P(3HB) powder L3 Biomass treatment 1,500 t of P(3HB) in enriched biomass are recovered for 330 days/year 60.8 kg⋅m −3 biomass concentration with 70% P (3HB) weight content f A heating step (55 ◦C) f Extraction A 94% P(3HB) recovery yield SDS chemical digestion in a 3:1 SDS-NCPM ratio in first recovery step f 15 wt% NaClO solution chemical digestion in a 8:1 NaClO-NCPM ratio in second recovery step f SDS recovery step comprised by: refrigeration (9 ◦C) of the liquid stream resulting of first recovery step, precipitation and microfiltration –80% yield– f Recovery and purification 94% P(3HB) purity Water washing in two step maintaining up to 5 wt % solids content –water from second washing step is recycled into first washing step– Separation by centrifugation and spray drying to obtain P(3HB) powder f L4 Biomass treatment 100,000 t of P(3HB) in enriched biomass are recovered for 330 days/year 150 kg⋅m −3 biomass concentration with 70% P (3HB) weight content g A heating step –up to 95 ◦C– h Extraction A 95% P(3HB) recovery yield h Fusel alcohols extraction at 115 ◦C in a 75:1 solvent-biomass ratio in R-201 –80% of the solvent stream is introduced as a liquid stream at 105 ◦C while the rest of the stream is added as vapor stream at 135 ◦C– h Solvent recovery in a distillation column h Recovery and purification 98% P(3HB-co-4HB) gh P(3HB-co-4HB) precipitation step by cooling (E301) h Separation by microfiltration and spray drying to obtain P(3HB) powder h References for assumptions: a-[48], b-[22], c-[50], d-[49], e-[51], f-[32,33], g-[52], h- [23] M. Saavedra del Oso et al. Chemical Engineering Journal 412 (2021) 127687 5 biorefinery, such as fusel alcohols, as solvent. This process, which is based on a PHB Industrial S.A. patent [23], is comprised by a extraction step where the P(3HB) is extracted by fusel alcohols. Then, after some solid–liquid separations to separate the NCPM and concentrate the polymer, the stream is cooled to precipitate and recover the polymer prior to spray drying. 4. Environmental and economic evaluation of the selected processes 4.1. State-of-the-art of LCA and LCC studies on PHA downstream processing Even when a lack of LCAs which assess only the PHA downstream processing has been detected, useful information can be collected from the selected LCAs studies on PHA production (Table 4). In terms of motivation studies, most research in the period 2000–2010 focused on evaluating the feasibility of PHA production comparing it to oil-based plastics, while in the present decade, studies were more focused on comparing PHA production from different feedstocks, evaluating different processes configuration such as PHA production using mixed cultures, assessing more impact categories or evaluating some specific phases of bioplastic value chain such as downstream processing [19,32,34]. Concerning to the recovery method, chemical digestion is the most evaluated option –9 out of the 13 papers–. Indeed, chemical digestion methods have an advantage over solvent extraction, especially when the PHA content in dry cell basis is high (greater than80%) as it happens in these studies [28,30,31]. Mechanical disruption, which has been evaluated by Gerngross [9], can be considered as competitive as chemical digestion. Solvent extraction, which was evaluated by few authors –6 out of the 13 papers–, is a highly energy intensive process due to solvents recovery step. Thus, a heat integration within a biorefinery may solve this dependency. Regarding the selection of the functional unit (FU), a mass-based approach is the most common option, being 1 kg or 1 ton of PHA –or specifically PHB– the preferred option –i.e. 10 of the 13 papers–. When looking at the system boundaries, 7 studies are classified as cradle-togate –i.e. from raw material extraction to factory gate– and the remaining 6 as gate-to-gate –i.e. from one defined point along the life cycle to a second defined point further along the life cycle–. Looking at the inventory phase it can be concluded that nondisclosure agreements prevent the publication of primary data by industry actors [55]. Therefore, most of primary data comes from own estimations, computer simulation and previous literature. Indeed, a dense correlation among the 13 papers has been observed and an important part of primary in LCI share the same origin –two earlier studies [9,28] are used as reference for the data collection of other six references [10,26,29–31,34]–, so there is a lack of primary sources and overdependency in LCI construction. Regarding secondary data, both Ecoinvent and GaBi databases are the most employed for data collection. Concerning the impact assessment stage, most studies followed a midpoint approach. Global warming potential (GWP) is so far the most evaluated category –11 of 13, followed by non-renewable energy use (NREU) –6 of 13–. In fact, they were the main focus in early studies (1999–2008): the goal of these studies was to compare the biopolymers carbon footprint with their respective oil-based. These days, as some solvents are considered hazardous for the ecosystem and human health, more impact categories –such as acidification, eutrophication and human toxicity potential– are being included in the assessment [10,26,34,35]. Half of the selected LCA studies include an LCC assessment. Most of the studies which assess the overall PHA production process pointed out fermentation as responsible of major operating costs –up to 75% of total–, which range from 1.18 to 6.12 € ⋅kg −1 [28,29,33,35]. Substrates represented up to 50% of fermentation costs in pure culture fermentation [28,35], while energy was the major cost in mixed culture fermentation processes [29,33]. Concerning to LCCs which asses PHA downstream processing, methods based on chemical digestion, especially those which employ NaOH and surfactants, have lower costs –which range from 1.02 to 5.23 vs 1.95 to 6.61 € ⋅kg −1 – than those based on solvent extraction[18,31]. Within PHA downstream processing by Fig. 2. H1 process flow diagram. Fig. 3. L2 process flow diagram. M. Saavedra del Oso et al. Chemical Engineering Journal 412 (2021) 127687 6 Table 4 Chronological summary of characteristics of reviewed LCA studies. 1 NREU: non-renewable energy use. GWP: global warming potential, OFP: photochemical ozone formation, FAETP: ecotoxicity for aquatic fresh water, AP: acidification potential, EP: eutrophication potential, FOFP: photo-oxidant formation potential, ODP: ozone depletion potential, TTP: terrestrial toxicity potential, ATP: aquatic toxicity potential, HTPI: human toxicity potential by ingestion, HTPE: human toxicity potential by either inhalation or dermal exposure. 2 PHA production costs including raw materials, fermentation and downstream processing. 3 PHA downstream processing costs. Source Recovery method Solvent Functional Unit System boundaries Origin of primary data Origin of secondary data LCIA method and impact categories 1 LCC [9] Mechanical disruption – 1 kg PHA Cradle-to-gate Calculations and estimates based on US Department of Energy (DOE) and United States Department of Agriculture (USDA); literature Literature Midpoint level NREU – [27] Solvent extraction C4-C11 alcohol 1 kg PHA Cradle-to-gate Monsanto’s data and assumptions; literature; Air Chief, EPA 1997; Ontario Ministry of Agriculture, Food and Rural Affairs, OMAFRA Economic Research Service (ERS) of the United States Department of Agriculture-USDA; Ecobalance’s DEAM TM database Midpoint level GWP – [28] Chemical digestion SDS + NaClO 5000 tons PHA Cradle-to-gate USDA and DOE; literature; own calculations; computer simulation using SuperPro Designer v4.5 Own calculations; literature; computer simulation using SuperPro Designer v4.5 Midpoint level GWP, NREU 3.53–4.77 € ⋅kg −1 PHA 2 [29] Chemical digestion NaClO 1 kg COD in the feed Gate-to-gate All the environmental figures are taken from Australian based source (local utility companies, the Australian Greenhouse Office); literature Literature Midpoint level GWP 4.33 € ⋅kg −1 PHA 2 [30] Chemical digestion NaOH + NaClO 1 kg PHA Gate-to-gate Literature; computer simulation (data of a simulated ethanol plant that are based on laboratory and pilot-plant results) Data from Agricultural Resource Management Survey (ARMS); Literature; average performance of chemical industry in the U.S. Midpoint level GWP, NREU – [31] Mechanical disruption + Chemical digestion / Solvent extraction SDS + NaClO / C4-C11 alcohol 1 kg PHA Cradle-to-gate Literature GaBi Professional database End-point level Ecosystem quality, human health, supply of resources – [33] Chemical digestion SDS + NaClO 1 kg PHB Gate-to-gate Laboratory and pilot scale data and literature data integrated with process modelling in ASPEN Plus software EcoInvent v2.2 Midpoint level GWP, NREU 1.18 € ⋅kg −1 PHA 2 [32] Chemical digestion (2) / Solvent extraction SDS + NaOH / SDS + NaClO / DCM 1 kg PHB Cradle-to-gate Laboratory and pilot scale data and literature data integrated with process modelling in ASPEN Plus software EcoInvent v2.2 Midpoint level GWP, NREU 1.40–1.95 € ⋅kg −1 PHA 3 [19] Chemical digestion (3) / Solvent extraction NaOH / NaClO / DCM / H 2 SO 4 1 kg PHB Gate-to-gate Laboratory and pilot scale data and literature data integrated with process modelling in ASPEN Plus software Not defined Midpoint level GWP 1.02–6.61 € ⋅kg −1 PHA 3 [34] Solvent extraction DMC 1 kg PHB Gate-to-gate (only downstream processing) Scale-up of laboratory data; estimates; literature Gabi Professional Database; Ecoinvent Database Midpoint level GWP, OFP, FAETP – [26] Solvent extraction Acetone Not stated Cradle-to-gate (polymer rich biomass) Literature data GaBi 6 Professional and Ecoinvent 3.1 databases Midpoint level GWP, AP, EP freshwater, EP marine, EP terrestrial, POFP – [35] Chemical digestion SDS + NaOCl 1 kg PHA Gate-to-gate Laboratory data; literature WAR database Midpoint level GWP, AP, ODP, TTP, ATP, HTPI, HTPE 5.77–6.12 € ⋅kg −1 PHA 2 [10] Chemical digestion NaOCl 1 kg PHB Cradle-to-gate Literature Literature; Gabi Professional database; Ecoinvent Midpoint level GWP, NREU, EP, AP – M. Saavedra del Oso et al. Chemical Engineering Journal 412 (2021) 127687 7 NCPM digestion and PHA extraction, chemicals and heat were the major contributors to total operating costs respectively –up to 80 and 70%– [19,32]. 4.2. LCA and LCC of selected processes: Goal and scope The goal of the current LCA is the identification of the main contributors, i.e. hotspots, both from an environmental and economic perspective within the four PHA recovery processes for both groups –highand low-grade– as well as the cross comparison within each group to identify the best alternatives. As the function to be covered by the final product is not equivalent for all the cases, two functional units were defined: 1 kg of high-grade PHA powder and 1 kg of low-grade PHA powder. The system boundaries are common, with a gate-to-gate approach covering from the PHA enriched biomass to the purified PHA. Being a comparative assessment, both the upstream substrate and PHA production phase and afterwards bioplastic compounding and shaping, use and end-of-life are assumed to be equivalent for all considered downstream processes [56]. Indeed, not all the downstream processes are interchangeable within each group –high and low-grade PHA–, i.e. some processes are based on an halophilic culture which can be submitted to osmotic shock or a solvent which is produced in the facility. Nevertheless, identified best practices can guide the proposal and adoption of improvement actions in different configurations, as it is commonly done in process. As almost all alternatives are still at lab or pilot scale, some well-founded assumptions were taken for extrapolation to full scale production –see Supplementary information–. Data for the processes of the background system come from the Ecoinvent v3.3 database [57]. Regarding the geographical boundaries and the time horizon, all processes are placed in the EU, including background processes such as chemicals and energy production –whenever possible–, and long-term emissions are excluded. Capital goods were excluded for the environmental assessment, and therefore only operational inputs and outputs have been collected. The analysis of the environmental impacts followed a midpoint approach, being global warming (GWP), terrestrial acidification, freshwater eutrophication, human toxicity, freshwater ecotoxicity and fossil depletion potential the selected impact categories, according to the latest reviewed LCAs on PHA downstream processing [10,26,34,35]. Excluding GWP, where the last version of the IPCC method –for a 100year time horizon– was used, other impacts categories were assessed using the Hierarchist ReCiPe(H) v1.13. To do so, version 8.3 of SimaPro software –PR´ e Sustainability, NL– was selected. Finally, to overcome the lack of specific background data, the following assumptions were formulated: (1) Biomass waste in processes H1, H4, L1 and L4 was assumed as composting biowaste. (2) Wastewater stream was assumed as urban wastewater. (3) The production process for SDS, used in processes H2, L2 and L3, was assimilated to another chemical with similar function: alkylbenzene sulfonate. (4) The production process for Lysol, used in process H2, was approximated to its main component: o-cresol. (5) The production of isoamyl alcohol from ethanol biorefinery, as is the case in process L4, was assimilated as the standard isoamyl alcohol chemical production process, which a priori is considered the worst case scenario. (6) Solvent losses in spray drying are burnt and complete combustion is assumed so fossil CO 2 is produced. 4.3. LCA and LCC of defined processes: Results for High-grade PHA Inputs of raw material and energy as well as emissions to air, water and solid waste or wastewater per kg of obtained PHA for the four highgrade PHA processes are summarized in Table 5. The selected background processes from Ecoinvent v3.3 are also reported for transparency. esults from the characterization stage are displayed on Table 6 and Fig. 4, and further individual information is reported on the Supplementary Information (sections 1.2.1, 2.2.1, 3.2.1 and so on). Results from the life cycle costing are displayed on Fig. 5. Capital investment and operating costs calculation is collected in Supplementary material. Processes based on NCPM disruption show a better environmental and economic performance than those based on solvent extraction, Table 5 LCI per kg of PHA for high grade processes: H1, H2, H3 and H4. Item H1 H2 H3 H4 Ecoinvent v3.3 Resources Water (m 3 ) – 0.028 0.005 – Water, cooling, unspecified natural origin, Europe without Switzerland Acetone (kg) 0.321 – – – Acetone, liquid {RER 1 }| oxidation of butane Materials/fuels Ethanol (kg) 0.678 – – 0.223 Ethanol, without water, in 99.7% solution state, from ethylene {RER}| ethylene hydration SDS (kg) – 0.049 – – Alkylbenzene sulfonate, linear, petrochemical {RER}| production Lysol (kg) – – 0.429 – o-cresol {RER}| Production Sodium hydroxide (kg) – 0.013 0.172 – Sodium hydroxide, without water, in 50% solution state {RER}| chlor-alkali electrolysis, membrane cell Hydrogen peroxide (kg) – 0.280 – – Hydrogen peroxide, without water, in 50% solution state {RER}| hydrogen peroxide production, product in 50% solution state Sulfuric acid (kg) – 0.015 – – Sulfuric acid {RER}| production Ethyl acetate (kg) – – 0.803 Ethyl acetate {RER}| production Heptane (kg) – – 0.185 Heptane {RER}| molecular sieve separation of naphtha Electricity/heat Electricity (kJ) 309 1,581 128 382 Electricity, medium voltage {Europe without Switzerland} market group for Heat duty (kJ) 62,972 3,171 3,166 110,082 Heat, in chemical industry {RER} market for Cooling duty (kJ) 48,217 – – 93,989 Steam, in chemical industry {RER}| steam production in chemical industry Emissions to air Carbon dioxide (kg) 2.484 – – 0.595 Carbon dioxide, fossil Waste and emissions to treatment Solid waste (kg) 0.577 – – 1.113 Biowaste {CH 2 } treatment of, composting Wastewater (m 3 ) 0.136 0.032 0.171 0.041 Wastewater, average {Europe without Switzerland}| treatment of wastewater, average 1 RER =Europe. 2 CH =Switzerland M. Saavedra del Oso et al. Chemical Engineering Journal 412 (2021) 127687 8 while chemicals and heat production are the main contributors to environmental impacts and total operating costs in high grade processes. Going further down and looking at elements contribution (Fig. 4), most environmental impacts of both H1 and H4 processes, which are based on solvent extraction, are caused mainly by heat production, and to a lesser extent by solvent production. However, freshwater eutrophication is predominately caused by ethanol production. From an economic perspective, heat represents a 20% and a 30% of total operating costs in H1 and H4 respectively, while chemicals represent more than a 50% in both cases. Respecting H2 process, based on mechanical disruption, heat, electricity and peroxide hydrogen production have a similar weight in most impact categories. Economically, peroxide hydrogen represents almost the 50% of total operating costs. Ortocresol production is behind the major share of environmental impacts in most categories in process H3 except freshwater eutrophication, which is caused by waste stream treatment, similarly to process H2. Likewise, surfactant is responsible for 80% of the total operating costs. 4.4. LCA and LCC of defined processes: Results for low-grade PHA Inputs of raw material and energy as well as emissions to air, water and solid waste or wastewater per kg of obtained high-grade PHA for the four processes are summarized in Table 7. The selected background processes from Ecoinvent v3.3 are also reported for completeness. In current subsection, results of characterization of several alternative processes of PHA downstream processing for high value Table 6 Comparative of characterization results of high-grade PHA downstream processes per kg of PHA. Shading colors from red to white indicate a better environmental performance of each process compared to the others in each impact category. Impact category Unit L1 L2 L3 L4 IPCC GWP 100a kg CO 2 eq 9.259 0.806 2.394 12.956 Terrestrial acidification kg SO 2 eq 0.023 0.003 0.009 0.044 Freshwater eutrophication kg P eq 4.950⋅10 - 4 6.770⋅10 - 5 2.370⋅10 - 4 5.600⋅10 - 4 Human toxicity kg 1,4DB eq 0.172 0.036 0.068 0.331 Freshwater ecotoxicity kg 1,4DB eq 2.470⋅10 - 3 4.400⋅10 - 4 2.920⋅10 - 3 5.890⋅10 - 3 Fossil depletion kg oil eq 2.675 0.290 1.223 4.503 Fig. 4. Characterization of high-grade processes H1, H2, H3 and H4 respectively, and contributions of each LCI component. Fig. 5. Economic evaluation of high grade processes H1, H2, H3 and H4 respectively, per kg of PHA. AD, U, m, L and M are referred to annual depreciation, utilities, materials, labor and maintenance respectively. Table 7 LCI per kg of PHA for processes L1, L2, L3 and L4. Items L1 L2 L3 L4 Ecoinvent v3.3 Resources Water (m 3 ) 0.0002 0.314 0.039 – Water, cooling, unspecified natural origin, Europe without Switzerland Materials/ fuels Acetone (kg) 0.487 – – – Acetone, liquid {RER}| oxidation of butane Isoamyl alcohol (kg) – – – 0.962 3-methyl-1-butanol {RER}| hydroformylation of butane SDS (kg) – 0.252 0.303 – Alkylbenzene sulfonate, linear, petrochemical {RER}| production Sodium hypochlorite (kg) – – 0.414 – Sodium hypochlorite, without water, in 15% solution state {RER}| sodium hypochlorite production, product in 15% solution state Electricity/ heat Electricity (kJ) 206 1,372 237 353 Electricity, medium voltage {Europe without Switzerland} market group for Heat (kJ) 24,684 43,847 4,380 54,570 Heat, in chemical industry {RER} market for Cooling duty (kJ) 17,470 – 1,659 48,189 Cooling energy {RER}| from natural gas, at cogeneration unit with absorption chiller 100 kW Emissions to air Carbon dioxide (kg) 1.109 – – 1.578 Carbon dioxide, fossil Waste and emissions to treatment Solid waste (kg) 1.172 – – 0.710 Biowaste {CH} treatment of, composting Wastewater (m 3 ) 0.006 0.316 0.065 0.009 Wastewater, average {Europe without Switzerland}| treatment of wastewater, average, capacity 1E9l/year 1 RER = Europe. 2 CH = Switzerland M. Saavedra del Oso et al. Chemical Engineering Journal 412 (2021) 127687 9 applications are displayed on Table 8 and Fig. 6. Results from the life cycle costing are displayed on Fig. 7: Process L3 and L4, based on NCPM digestion and PHA solvent extraction, show the best environmental and economic performance respectively within low-grade processes, while chemicals and heat production are the main contributors to environmental impacts and total operating costs. Regarding processes contribution to low-grade PHA processes (Fig. 6), most environmental impacts of L1, which is based on solvent extraction, are caused mainly by heat production, and to a lesser extent by acetone production. Similarly, total operating costs are caused by heat (49%) and acetone (44%). Respecting L2 process, based on osmotic shock and chemical digestion, heat production dominates almost all impact categories, except freshwater eutrophication, which is mainly caused by wastewater treatment. With regards to total operating costs, heat (36%) and annual depreciation (25%) –the plant scale determines higher AD costs– have a relevant weight. Sodium hypochlorite and SDS production are responsible by major of environmental impacts in most categories in process L3 except freshwater eutrophication, which is caused by waste stream treatment, similarly to process L2. Economically, sodium hypochlorite (36%) and annual depreciation (27%) –note that, similarly to L2, the size scale influences AD costs– represent the major costs. As regards process L4, isoamyl alcohol and heat production are the major responsible of all environmental impacts in any category. In economic assessment, and due to fusel alcohols are a byproduct of ethanol biorefinery, heat cost represents 80% of total operating costs. Similarly, Lysol and heat production are the principal contributors to H3 environmental impacts. 4.5. LCA and LCC of defined processes: Interpretation of results The fourth stage of the life cycle based assessments aims at deriving conclusions and assessing their robustness of conclusions. To do so, firstly a comparison among the processes assessed in this work and those from previous studies, secondly a sensitivity analysis to assess the robustness of the LCA results is carried out. 4.5.1. Comparison of results According to results showed in Table 6 and Fig. 4, methods relying Table 8 Comparative of characterization results of low-grade PHA downstream processes per kg of PHA. Shading colors from red to white indicate a better environmental performance of each process compared to the others in each impact category. Impact category Unit L1 L2 L3 L4 IPCC GWP 100a kg CO 2 eq 3.931 4.159 1.177 10.249 Terrestrial acidification kg SO 2 eq 0.011 0.015 0.005 0.032 Freshwater eutrophication kg P eq 6.170⋅10 - 5 3.710⋅10 - 4 1.020⋅10 - 4 2.270⋅10 - 4 Human toxicity kg 1,4DB eq 0.075 0.176 0.115 0.234 Freshwater ecotoxicity kg 1,4DB eq 1.300⋅10 - 3 1.880⋅10 - 3 6.350⋅10 - 4 2.427⋅10 - 3 Fossil depletion kg oil eq 1.132 1.467 0.577 3.413 Fig. 6. Characterization of low-grade PHA processes L1, L2, L3 and L4 respectively, and contributions of each LCI component. Fig. 7. Economic evaluation of low-grade processes L1, L2, L3 and L4 respectively, per kg of PHA. AD, U, m, L and M are referred to annual depreciation, utilities, materials, labor and maintenance respectively. Fig. 8. Comparative of characterization of high-grade PHA processes. Fig. 9. Comparative of characterization of low-grade PHA processes. M. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 Available online 29 November 2022 0959-6526/© 2022 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Prospective LCA to provide environmental guidance for developing waste-to-PHA biorefineries Mateo Saavedra del Oso a , b , c , * , Miguel Mauricio-Iglesias a , Almudena Hospido a , Bernhard Steubing b a Cross-disciplinary Research Center in Environmental Technologies (CRETUS), Department of Chemical Engineering, Universidade de Santiago de Compostela, 15782, Santiago de Compostela, Spain b Institute of Environmental Sciences (CML), Leiden University, 2300, Leiden, the Netherlands c Centre of Excellence for Packaging Sustainability (CEPS), Division of Packaging Materials, Stora Enso, 00180, Helsinki, Finland ARTICLE INFO Handling Editor: Kathleen Aviso Keywords: Polyhydroxyalkanoates Mixed culture fermentation Life cycle assessment Environmental assessment Bioprocess scale-up ABSTRACT Polyhydroxyalkanoates (PHA) production from waste streams using mixed microbial cultures (MMC) can unlock the potential of PHA to substitute oil-based plastics. However, these processes are still at low technology readiness level (4–6). Demonstrating a better environmental performance would boost their deployment at industrial scale. Hence, including environmental guidance during their development, when there are still opportunities for major alterations, is essential. To the best of our knowledge, this work elucidates for the first time how waste-to-PHA biorefineries could develop in the future by combining prospective LCA with scenario methodology and where the attention of stakeholders should be focused. Four future scenarios were derived considering both surrounding (e.g., scale, environmental or bioeconomy policies) and technological parameters (e.g., acidification yield, PHA content in biomass or recovery yield). Those scenarios derived under ambitious environmental and bioeconomy policies shop up to 50% lower environmental impacts than those under businessas-usual policies. These differences are caused by the different background processes’ environmental burdens (e. g., electricity mix with low renewable energies share) and the higher consumption of chemicals and utilities. However, the environmental impacts caused by lower yields can be partially mitigated by valorizing the intermediate waste streams into biogas. Sensitivity analysis results pointed out recovery yield and PHA content as the parameters that influence most the environmental performance, being responsible for up to 60% of variance in environmental performance. These parameters determine the chemicals and utilities consumption in PHA downstream processing, which is confirmed as the main environmental hotspot. This work goes beyond previous LCA studies on PHA production and quantifies the influence of different parameters on the environmental performance. 1. Introduction Bioeconomy is expected to play a significant role in the mitigation and adaptation to climate change across the European Union, targeting at reducing the pressure on biological resources as well as reducing CO 2 emissions and fossil fuel use in the chemical sectors according to its action plan launched by the European Commission in 2018 (Bell et al., 2018). In this endeavor, using biomass and especially organic wastes as feedstock to produce these chemicals and materials is a priority. For instance, polyhydroxyalkanoates (PHA), which are biodegradable polymers produced through microbial fermentation from diverse feedstocks (Khatami et al., 2021), could substitute petrochemical plastics in multiple applications (Dietrich et al., 2017). However, their high production cost and uncertain environmental performance hamper their industrialization (Tan et al., 2021). Demonstrating a better environmental performance than conventional plastics would foster PHA deployment (Yadav et al., 2020). High energy requirement during the feedstock cultivation, sterilization in pure culture fermentation and PHA downstream processing were reported as the main hotspots from both environmental and economic perspectives (Saavedra del Oso et al., 2021). Coupling PHA production with the carboxylate platform, where * Corresponding author. Cross-disciplinary Research Center in Environmental Technologies (CRETUS), Department of Chemical Engineering, Universidade de Santiago de Compostela, 15782, Santiago de Compostela, Spain. E-mail address: [email protected] (M. Saavedra del Oso). Contents lists available at ScienceDirect Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro https://doi.org/10.1016/j.jclepro.2022.135331 Received 5 August 2022; Received in revised form 8 November 2022; Accepted 24 November 2022 Journal of Cleaner Production 383 (2023) 135331 2 organic waste streams are converted into volatile fatty acids (VFA) as PHA precursors, would significantly reduce the cost and could improve the environmental performance of PHA production (Atasoy et al., 2018). H2020 project USABLE Packaging (H2020 USABLE Packaging, 2019) seeks the development of new value chains from food industry wastes by solving the bottlenecks in the PHA production. This project proposes cost-effective and sustainable production routes, such as the one based on the mixed microbial cultures (MMC) systems which comprises a 3-step system: (1) an anaerobic fermentation, where feedstock organic carbon is converted into VFA, (2) an enrichment of mixed culture, where PHA-storing bacteria are selected by imposing a feast/famine regime, and (3) an accumulation step, where enriched biomass is fed VFA until sufficient PHA has been accumulated (Nguyenhuynh et al., 2021). The performance of these systems at pilot scale has been evaluated for a diverse range of feedstocks, such as wastewater (Morgan-Sagastume et al., 2020), household waste (Moretto et al., 2020) or food industry side-streams (Silva et al., 2022). However, scaling up these processes to a commercial full-scale PHA production still faces multiple challenges, from low substrate conversion and poor mechanical properties to deploying a cost-effective and sustainable extraction strategy at pilot scale (Est´ evez-Alonso et al., 2021). Demonstrating a better environmental performance than petrochemical plastics would promote PHA development and market expansion. Previous LCA studies on the production of PHA by MMC assessed the environmental benefits of integrating the PHA production within urban (Morgan-Sagastume et al., 2016) and industrial wastewater treatment plants (Fern´ andez-Dacosta et al., 2015; Roib´ as-Rozas et al., 2020). Fern´ andez-Dacosta et al. (2015) evaluated the environmental performance of different PHA recovery technologies from wastewater. However, none of these articles addressed the inherent uncertainties of upscaling processes at laband pilot-scale or quantified the influence of process parameters on the environmental performance (Igos et al., 2019). Prospective life cycle assessment (LCA) can play a key role in the development and optimization of MMC systems by providing environmental guidance (Arvidsson et al., 2018). Prospective LCA facilitates the upscaling of emerging processes employing scenarios of future performance at industrial scale, and the comparison of the future process with the incumbent processes (Cucurachi et al., 2018). However, its prospective character entails, to some degree, lack of data and considerable uncertainty (Igos et al., 2019). For instance, the market share of these biopolymers and the feedstock availability will determine the production scale, or the environmental policies will influence how electricity is produced (Saavedra del Oso et al., 2021). Yet, both future market share and environmental policies entail a high uncertainty. A feasible approach to evaluate the environmental performance in a future framework is to propose scenarios projecting how emerging processes will develop in the future (Arvidsson et al., 2018; Bergerson et al., 2020). Contributions on how to combine scenario methodology and prospective LCA have been published (Delpierre et al., 2021; Thomassen et al., 2019) but did not cover biobased emerging technologies, which face different challenges regarding the choice of functional unit, allocation approaches, or process upscaling. Only recently, Langkau et al. (Unpublished work) have proposed a scenario methodology framework on biobased products, which enables the systematic and documented development of scenarios, rather than the implicit approach that is often behind prospective LCA studies. Analyzing the LCA results of the proposed scenarios allows inferring the relationship between the characteristics of the scenarios (e.g., type of substrate, yield or renewable energies share) and the environmental performance. Thus, it is possible to identify the most relevant parameters to optimize during the technology development, enabling effective environmental guidance. The objective of this work is to elucidate how waste-to-PHA biorefineries could develop in the future and ensure that environmental guidance is included in their development. To do so, influencing parameters were firstly identified based on literature review as well as both individual meetings and a workshop with stakeholders. Based on the information collected, scenarios were then developed and validated with stakeholders. Finally, their environmental impacts were quantified and analyzed, allowing the identification and quantification of the parameters influence on the environmental performance. 2. Methodology In this section, prospective LCA (section 2.1) and scenario methodology (section 2.2) are described. 2.1. Prospective LCA Prospective LCA is a systematic methodology which determines the environmental impacts of an emerging/incumbent product/process in a future framework where the production system is modelled (Cucurachi et al., 2018). Like conventional LCA, prospective LCA is comprised by four steps (see Fig. A.1): (i) the goal and scope state of the system function, functional unit, system boundaries, technology readiness level (TRL), temporal boundaries, foreground and background data source, identification of alternative systems, impact assessment method and impact categories; (ii) the inventory analysis involves the data collection and the development of scenarios and its implementation as inputs and outputs of a product system; (iii) the impact assessment transforms the inventory results into potential environmental impacts; and (iv) the interpretation phase, which involves a critical review, results presentation and determination of data uncertainty and sensitivity (Cucurachi et al., 2022). Abbreviations and nomenclature BAU business-as-usual CHP cogeneration unit plant COD chemical oxygen demand FDP fossil depletion potential FEP freshwater eutrophication FETPinf freshwater ecotoxicity FU functional unit GDP gross domestic product GMA general morphological analysis GWP100 climate change HTPinf human toxicity LCA life cycle assessment LCI life cycle inventory MMC mixed microbial culture OFMSW organic fraction of the municipal solid waste PHA polyhydroxyalkanoates R&D research and development RES Renewable energies share SSP shared socio-economic pathway TAP terrestrial acidification TRL: technology readiness level VFA volatile fatty acids Y PHA/VFA accumulation yield Y VFA/S acidification yield Y X/VFA selection yield M. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 3 2.2. Scenario methodology Scenario development is a systematic and documented methodology (Thomassen et al., 2019) that involves both the goal & scope definition and inventory analysis phases. The key choices for scenario development are made during the goal & definition, i.e., the time horizon, the current and expected TRL and the scenario approach (i.e. whether a predictive, explorative, or normative approach is followed; it aims to answer “how will/could/should the future develop” respectively (Langkau et al., Unpublished work)). During the inventory analysis, a 4-step procedure is followed to develop the scenarios: (i) identification of influencing parameters, (ii) construction of sub-scenarios for each parameter, (iii) creation of scenarios from sub-scenarios and (iv) implementation of scenarios in the life cycle inventory (LCI) model. 2.2.1. Identification of influencing parameters Scopus web search engine was employed to screen scientific papers on waste-to-PHA biorefineries. To do so, the following query string was used: TITLE-ABS-KEY ((pha OR polyhydroxyalkanoates OR polyhydroxybutyrate OR phb OR phbv OR p3hb OR poly3-hydroxybutyrate) AND ((mixed AND microbial AND cultures) OR (mixed AND culture) OR (open AND mixed AND culture) OR (mmc)) AND (pilot-scale OR scaleup OR (large AND scale))). Among the 77 resulting elements, only studies on PHA production by MMC at pilot scale were selected (i.e., the 11 peer-review studies listed in appendix A). The insights and data gathered in this review were used to develop the scenarios, both identifying the influencing parameters and upscaling the processes. 2.2.2. Individual meetings with stakeholders Individual meetings with stakeholders (feedstock providers, technology developers, packaging producers and end users) were carried out after the literature review to support the scenario development. The procedure was the following: (i) presentation of the PHA production flowchart, (ii) identification and validation of the process influencing parameters, (iii) identification and validation of the parameters influencing the surroundings, (iv) creation and validation of sub-scenarios for each parameter. The stakeholders feedback enabled the influencing parameters and sub-scenarios definition. Fig. 1. Waste-to-PHA biorefinery system boundaries. M. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 4 2.2.3. Workshop with stakeholders An on-line workshop with the previously identified stakeholders was organized to gather their individual feedback and to jointly create, develop, and validate the scenarios derived. The workshop was structured in the following steps: (i) presentation of the prospective LCA and scenario methodology, (ii) overview of the waste-to-PHA biorefinery system, (iii) validation of the influencing parameters, (iv) validation of parameters’ sub-scenarios, (iv) creation and validation of scenarios from sub-scenarios, (v) discussion and final remarks. 3. Scenario development This section depicts the goal and scope definition (section 3.1) and the inventory analysis (section 3.2). 3.1. Goal and scope definition The main goal of this prospective LCA is to quantify the environmental impacts of the waste-to-PHA biorefinery in a future context when its large-scale production is already implemented. As the system function is to produce PHA from complex organic wastes, the functional unit (FU) was defined as 1 kg of PHA powder (a common approach in previous LCA of PHA (Roib´ as-Rozas et al., 2022)). A cradle-to-gate approach covering all the unit processes within the waste-to-PHA biorefinery is defined in Fig. 1; including within the system boundaries the PHA production (i.e. anaerobic fermentation, the VFA separation, the PHA enrichment and accumulation, as well as the downstream processing) and the further energy valorization of the residual intermediate streams by anaerobic digestion (AD) and cogeneration unit plant (CHP); and excluding then the gate-to-grave phases (i.e., compounding and shaping, use and end-of-life). The multifunctionality of the system was addressed by allocating the environmental burdens from VFA separation and from CHP to PHA and electricity production, respectively. This choice was made following the guidelines on the LCA of alternative feedstock for plastics production (Nessi et al., 2021). Given the actual TRL, i.e., from 5 to 6, the temporal boundaries were established for 2030, when market level maturity is expected to be reached (Lorini et al., 2022). An explorative scenario approach (B¨ orjeson et al., 2006) was followed with the aim of envisioning how the PHA production by MMC could develop. The foreground system was modelled based on the data collected from the literature review and using an upscaling framework for emerging technologies (Tsoy et al., 2020). For the background system, a futurized version of the ecoinvent 3.7.1 database was used that includes scenario data derived from the IMAGE integrated assessment model (Stehfest et al., 2014). The latter model’s global future scenarios are based on the Shared Socio-Economic Pathway (SSP) scenarios (O’Neill et al., 2014) and representative concentration pathways (van Vuuren et al., 2011). Two scenarios, the SSP2-base and the SSP2-RCP2.6 were derived here via the PREMISE framework (Sacchi et al., 2022). Both scenarios represent the SSP “middle of the road”, although they differ substantially in terms of climate change mitigation. In the SSP-base scenario a warming of 3.5 ◦C is modelled by 2100, while in the SSP2-RCP2.6 the temperature increase is limited to just below 2 ◦C (Sacchi et al., 2022). The foreground scenarios were modelled together with the background scenarios using the superstructure approach (Steubing and de Koning, 2021), as implemented in the open source LCA software Activity Browser (Steubing et al., 2020). The analysis of the environmental impacts followed a midpoint approach, being climate change (GWP100), terrestrial acidification (TAP), freshwater eutrophication (FEP), human toxicity (HTPinf), freshwater ecotoxicity (FETPinf) and fossil depletion (FDP), for the selected impact categories, according to previous LCA on PHA production (Roib´ as-Rozas et al., 2022). All impact categories were assessed using the Hierarchist ReCiPe (H) v1.13. 3.2. Inventory analysis The core part of the applied methodology concerns the scenario construction during the inventory analysis phase (Langkau et al., Unpublished work). Scenarios represent both descriptions of possible future states and descriptions of developments (B¨ orjeson et al., 2006); affecting both foreground and background data of the LCI. The main results of the 4-step procedure are presented in the next subsections (3.2.1, 3.2.2, 3.2.3 and 3.2.4). 3.2.1. Identification of influencing parameters The identified influencing parameters are listed in Table B.1. A 60% of them are surrounding parameters, which are related to political actions, societal concerns, technological and environmental aspects, but are not intrinsically present in the LCI model. The rest of identified parameters are technological parameters (foreground and background), i.e., those that affect directly the LCI model flows. Foreground parameters affect the mass and energy balances, while background parameters modify the background processes. The hybrid causal loop flowchart diagram shown in Fig. 2 depicts the influences and correlations between the identified influencing parameters and the LCI model. This diagram helps the analyst to consider surrounding parameters (left side), but also to understand how the technological parameters can be modelled within the LCI model (right side), and contains both quantitative/qualitative parameters, unit processes and intermediate/elementary flows for the LCI model. The process performance parameters (i.e., the acidification yield, productivity, PHA-storing biomass selection and PHA accumulation yield, PHA content in enriched biomass and extraction yield) depend on the R&D on these processes, but also on the feedstock employed (Saavedra del Oso et al., 2022). The choice of wastes as feedstock, e.g., organic fraction of the municipal solid waste (OFMSW) or sewage sludge, may hinder the applicability of the polymer for high purity applications such as food contact materials. The implementation of ambitious bioeconomy policies (Fritsche et al., 2020), fostered by higher environmental awareness and the impact of climate change, may increase the demand of PHA as well as the use of waste or industrial side-streams as feedstocks. Ambitious bioeconomy policies would increase not only the production scale but the R&D funding, whose outcomes can contribute to the reduction of the production costs. A positive feedback loop between the production scale and costs may be possible, i.e., by decreasing the production costs, the production scale increases, which in turn decreases them. The production of PHA by MMC from waste streams can benefit from the adoption of new environmental policies. The implementation of these legislations may increase the RES and introduce new emissions regulations, modifying then the background processes such as electricity production or chemicals production. This causality diagram allows not only to develop the model by identifying correlations between parameters and deal with their complexity, but also to communicate to stakeholders the influence of parameters. Thus, this tool adds transparency to the study. As the number of identified influencing parameters is high and the effects of some surrounding parameters are either intrinsically present in others or cannot be easily implemented in the LCI model, only environmental policies, bioeconomy policies and feedstocks were chosen, among those for the construction of sub-scenarios. 3.2.2. Construction of sub-scenarios for each parameter For all the selected surrounding parameters (bioeconomy policies, environmental policies, feedstock, production scale) a set of three levels or sub-scenarios are defined, except for the feedstock type where only two options were fixed. Business-as-usual (BAU), moderately ambitious and ambitious were defined as the three sub-scenarios for both environmental and bioeconomy policies. Related the former, since environmental policies determine both background parameters, i.e., RES and M. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 5 emissions regulations, these three sub-scenarios were implemented in the LCI model according to the RES and the SSP (O’Neill et al., 2014): BAU (current RES and SSP2-base), moderately ambitious (50% RES and SSP2-RCP2.6) and ambitious (60% RES and SSP2-RCP2.6). Concerning the later, BAU at bioeconomy policies means focus mainly on waste treatment and resource recovery, but the market share of these biobased products is still small. Moderately ambitious and ambitious bioeconomy policies consider waste as resources that can be transformed widely into added-value products with a big market share. These sub-scenarios are qualitative and are not directly implemented in the LCI model. Regarding the production scale, the sub-scenarios were proposed according to waste availability and technological developers’ knowledge: 500, 2000 and 7500 t PHA/y. The sub-scenarios for the feedstock and process performance parameters were chosen according to the literature review on PHA production by MMC at pilot scale. Only two sub-scenarios were created for the feedstock, i.e., fruit waste and the mixture of OFMSW and sewage sludge, since they are widely available (Yadav et al., 2020) and were employed in most of the waste-to-PHA pilot scale studies (Moretto et al., 2020; Silva et al., 2022). The sub-scenarios for the selected technological parameters are listed in Table 1. The sub-scenarios for the acidification yield, selection yield, accumulation yield, productivity and PHA content in biomass were created based on the literature review and feedback provided by the stakeholders in the individual meetings and the workshop. The recovery yield values were selected according to previous authors’ work on the optimization of the environmental performance of the PHA downstream processing (Saavedra del Oso et al., 2021). 3.2.3. Creation of scenarios from sub-scenarios Since the total number of combinations of sub-scenarios is very high, creating scenarios from those sub-scenarios is not a straightforward task. A cross-consistency assessment was performed to discard combinations of sub-scenarios that were not consistent (Appendix B.2). Regarding the environmental policies, it is inconsistent to have an ambitious bioeconomy policy or medium and large-scale production under the SSP2base sub-scenario, as oil-based chemicals and materials would still have a higher market share compared to biobased alternatives. Likewise, the PHA production from OFMSW & sewage sludge seems unfeasible at industrial scale under BAU bioeconomy policies, as conventional waste treatments such as energy valorization would be preferred. However, the use of concentrated, carbon-rich, easily fermentable, and localized feedstocks such as side-streams of food processing facilities could be feasible, e.g., fruit waste from juice beverage industries. As showed in Fig. 2, the implementation of ambitious bioeconomy policies would lead to a higher investment in the R&D of these technologies and thus, to higher values of process performance parameters. Nevertheless, these parameters are influenced by the type of feedstock too. For instance, a value of 0.75 g COD-VFA/g COD-F for the anaerobic fermentation yield of OFMSW & sewage sludge seems inconsistent, as these substrates are not easily hydrolyzed and fermented into VFA. The Fig. 2. Hybrid causal loop flowchart diagram depicting the interrelations between surrounding and technological parameters and the LCI model. Table 1 Sub-scenarios created for selected technological parameters. Parameter Subscenario 1 Subscenario 2 Subscenario 3 Units Y VFA/F 0.35 0.55 0.75 g COD/g COD Y X/VFA 0.40 0.45 0.55 g COD/g COD Y PHA/VFA 0.50 0.65 0.80 g COD/g COD Productivity 3.00 6.00 8.00 g PHA/ (L⋅d) PHA content in biomass 0.46 0.59 0.80 g/g Recovery yield 0.65 0.75 0.85 g/g M. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 6 cross-consistency check also allows disregarding combinations that would lead to technological inconsistencies, such as low PHA-storing biomass selection yield together with high PHA accumulation yield. As the remaining number of consistent scenarios is still high, general morphological analysis (GMA) was applied (see Table B.2). GMA is a method for structuring and investigating the total set of relationships contained in multi-dimensional, nonquantifiable problem complexes [51]. Different narratives for feasible scenarios were derived from the GMA. The eight narratives derived from the GMA, which are summarized in Appendix B.2, were presented, and discussed at the stakeholders’ workshop. As a result of the discussion, four of them (i.e., scenarios A, D, E and H) were chosen to be implemented in the LCI, as they can be considered as the cornerstone scenarios and thus, it may allow assessing scenarios at the edges of the solution space. From now, scenarios A, D, E and H will be referred as scenarios 1, 2, 3 and 4; and Table 2 lists their specific parameters sub-scenarios. 3.2.4. Implementation of scenarios in the LCI model The foreground system, i.e. the waste-to-PHA biorefinery, was upscaled using detailed process calculations (Tsoy et al., 2020). The process calculations, which follow a scale-up framework for chemical processes in LCA studies (Piccinno et al., 2016), are detailed in the Appendix C. Background process “market for electricity, high voltage (RER)” was chosen for scenario 1 from the ecoinvent 3.7.1 IMAGE database, which is the database employed for the rest of background processes. Electricity production was modelled for scenarios 2, 3 and 4 according to the project report Roadmap 2050 (included in Appendix D). Regarding the lack of specific background data, the following considerations were made: (i) the production process for sodium dodecyl sulphate (SDS), used in downstream processing, was assimilated to another chemical with similar function: non-ionic surfactant; (ii) the wastewater stream produced in the downstream processing was assumed as urban wastewater, (iii) potassium phosphate was assumed as sodium phosphate and (iii) the digestate treatment was assumed as treatment of raw sewage sludge by municipal incineration with fly ash extraction. Finally, the following assumptions related to process system (Fig. 1) were formulated: •The electricity produced by the AD is not locally consumed in the process and, instead, it is exported to the net. Thus, all the electricity consumed in the system comes from the European Union electricity mix. •The heat produced by the AD is internally consumed for heating the digestor, and should heat surplus be produced, it is assumed to be used by the PHA production section. All the environmental burdens of AD have been allocated to electricity production, which is the function of the waste valorization section. •The residual gas stream produced in the anaerobic fermentation, with negligible flowrate and composed by (biogenic) carbon dioxide with a small fraction of hydrogen, is assumed to be released to the atmosphere. Inputs of raw material and energy as well as emissions to air, water, and wastewater per kg of obtained PHA powder (i.e., defined functional unit) are summarized in Table 3. 4. Results This section presents the results of scenarios life cycle impact assessment and analyzes the parameters influence on the environmental performance (sections 4.1 and 4.2). Furthermore, the allocation methods are compared in Appendix E.2. 4.1. Environmental evaluation Results from the characterization stage for PHA production, where are displayed on Fig. 3. Heat production is the main contributor to GWP100 (52 and 51%) and FDP (62 and 58%) in scenarios 1 and 2. However, its contribution is negligible to all impacts categories in scenarios 3 and 4. This discrepancy is caused by the heat employed in these scenarios, which comes from the CHP (scenario 3, 100%; scenario 4, 90%), and has no environmental burdens as explained earlier. Thus, the heat source determines the related environmental impact, as shown in Fig. 3. Electricity contributes significantly to GWP100, HTPinf and FDP categories, especially in scenario 3. Regarding the electricity consumption, the downstream processing, the PHA accumulation and the biomass selection are pointed out as the unit processes that consume the most electricity (see Fig. E.1 and E.2). PHA downstream processing is caused by the high-pressure homogenization employed for disrupting the cells and the aeration systems used in both biomass selection and PHA accumulation. Significant differences were found between scenarios, as the values for extraction, selection and accumulation yield were different. For instance, extreme electricity consumption values in PHA downstream processing are 0.620 and 1.608 kWh/FU for scenarios 2 and 3 respectively (see Table 3), being the correspondent recovery yield 0.85 and 0.65 g/g respectively, and the biomass PHA content 0.80 and 0.46 g/g. Electricity consumption values in biomass selection and PHA accumulation for scenarios 2 and 3 are 0.140 and 1.209 and 0.414 and 1.260 kWh/FU respectively (see Table 3). These differences are explained by the differences in biomass selection (0.55 vs 0.4 g COD/g COD) and PHA accumulation yield (0.8 vs 0.5 g COD/g COD) as well as productivity (8 vs 3 g PHA/(L⋅d)), which affect the reactors dimensions and the required aeration. However, background parameters such as the RES must be considered, as they have a significant influence on the environmental Table 2 Description of selected parameter sub-scenarios for the created scenarios. Parameter Scenario 1 Scenario 2 Scenario 3 Scenario 4 Units Environmental policy SSP2-Base SSP2-RCP2.6 SSP2-RCP2.6 SSP2-RCP2.6 Current RES 60% RES 50% RES 60% RES Bioeconomy policy BAU Ambitious Ambitious Ambitious Production scale 500 7500 7500 7500 t PHA/y Feedstock Fruit waste Fruit waste OFMSW & SS OFMSW & SS Y VFA/F 0.55 0.75 0.35 0.55 g COD/g COD Y X/VFA 0.45 0.55 0.40 0.45 g COD/g COD Y PHA/VFA 0.65 0.80 0.50 0.65 g COD/g COD Productivity 6.00 8.00 3.00 6.00 g PHA/(L⋅d) PHA content in biomass 0.59 0.80 0.46 0.59 g/g Recovery yield 0.75 0.85 0.65 0.75 g/g M. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 7 impacts too. The GWP100 environmental burdens for the current, 50% and 60% RES are 0.268, 0.154 and 0.126 kg CO 2 -eq/kWh. The steam production is responsible for up to 25% environmental impact in GWP100, FEP and FDP. The steam is consumed in the last step of the PHA downstream processing, where the PHA cake is spray dried into powder. Regarding the consumption, scenario 1 employs a 44% more steam than scenario 2, while scenario 3 doubles scenario 1 and is 70% higher than scenario 4. These notable differences can be explained by the different PHA content and the recovery yield (see Tables 2 and 3). Chemicals contribute considerably to TAP, FEP, HTPinf and FETPinf. SDS, which is employed to disrupt the biomass in the downstream processing, has the highest contribution within chemicals (up to a 60% chemicals’ environmental burdens). Its consumption, which is related with the PHA content in the biomass, is substantially higher in scenario 1 than scenario 2, although scenario 3 has the highest. The other chemicals are used as additives in the biomass selection and their rates per FU depend essentially on the selection yield, being scenario 3 the one with the highest consumption (see Tables 2 and 3). Table E.1 summarizes the technological parameters influence on the background processes’ environmental burdens. With regards to the background parameters, the RES is determinant for the electricity production environmental burdens as previously stated. However, the environmental burdens reduction caused by using the SSP2-RCP2.6 background scenario instead SSP2-Base for 2030 is almost negligible (less than a 5% in all impact categories), as these scenarios consider a slow progress in achieving the sustainable development goals until 2050. Regarding the foreground parameters, recovery yield and PHA content in biomass are the parameters that influence most the environmental burdens of electricity, steam, and chemicals. 4.2. Uncertainty and sensitivity analysis Applying the scenario methodology allows the characterization of model structure and context uncertainty (Igos et al., 2019), i.e., accounting the uncertainty in the representativeness of the reality of the LCA model and the normative choices of LCA analysts respectively. Uncertainty analysis, shown in. Fig. 4, was performed to evaluate both epistemic (i.e. lack of knowledge) and ontic (i.e. deterministic and randomness) uncertainty related to the process performance parameters, considering a 10% variation for the lower and upper limits. The pressure drop in the highpressure homogenizer was included in the analysis, as there is a significant uncertainty related to this parameter (variance interval is 500–1500 kPa) and this step is one of the hotspots regarding the electricity. Uncertainty analysis was carried out through a Monte Carlo analysis (1000 iterations). Then, a global sensitivity analysis (GSA) (Cucurachi et al., 2022) was performed to evaluate the influence of each parameter on the results uncertainty. The GSA, shown in Fig. 5, was carried out following the standardized regression coefficients method. Significant differences were found among scenarios and categories when analyzing the uncertainty and GSA results together (Figs. 4 and 5). However, recovery yield is the parameter responsible for at least 50% Fig. 3. Characterization of PHA production within a waste-to-PHA biorefinery and contributions of background processes (FU: 1 kg of PHA). M. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 8 Table 3 LCI per kg of PHA powder for scenarios 1, 2, 3 and 4. Items Scenario 1 Scenario 2 Scenario 3 Scenario 4 Ecoinvent process Anaerobic fermentation Intermediate products VFA þnon fermented (kg COD) 14.05 5.81 39.95 12.71 Technosphere inputs Feedstock (kg COD) 14.41 5.97 39.97 12.72 Water (m 3 ) 0.07 0.01 0.02 0.01 market for tap water (Europe without Switzerland) Electricity (kWh) 0.11 0.04 0.40 0.13 market for electricity, high voltage a (RER b ) Heat (MJ) 30.35 12.70 0.00 1.27 heat production, natural gas, at industrial furnace >100 kW (Europe without Switzerland) Heat (MJ) 0.45 0.00 42.79 12.42 Sodium hydroxide (kg) 0.10 0.06 0.17 0.09 chlor-alkali electrolysis, membrane cell (RER) VFA separation Intermediate products VFA rich stream (kg COD) 6.13 3.46 10.8 5.41 Solid-rich wastewater stream (kg COD) 7.92 2.35 29.13 7.30 Technosphere inputs VFA þnon fermented (kg COD) 14.05 5.81 39.95 12.71 Electricity (kWh) 0.17 0.07 0.72 0.23 Biomass selection Intermediate products Biomass (kg) c 1.13 0.57 2.04 0.99 Technosphere inputs VFA rich stream (kg COD) 2.90 1.00 6.37 2.56 Electricity (kWh) 0.36 0.14 1.21 0.39 market for electricity, high voltage (RER) Heat (MJ) 1.74 3.81 0.00 4.00 heat production, natural gas, at industrial furnace >100 kW (Europe without Switzerland) Heat (MJ) 0.03 0.00 7.81 0.41 Ammonium chloride (kg) 0.04 0.02 0.08 0.03 market for ammonium chloride (GLO d ) Potassium phosphate (kg) 0.03 0.02 0.07 0.03 market for sodium phosphate (RER) Calcium chloride (kg) 0.02 0.01 0.04 0.01 market for calcium chloride (RER) Emissions to air Biogenic CO 2 (kg) 1.07 0.37 2.35 0.95 PHA accumulation Products PHA enriched biomass (kg) 2.26 1.57 3.34 1.99 Technosphere inputs VFA rich stream (kg COD) 3.23 2.46 4.46 2.85 Electricity (kWh) 0.49 0.41 1.26 0.61 market for electricity, high voltage (RER) Heat (MJ) 0.06 0.00 0.00 0.00 heat production, natural gas, at industrial furnace >100 kW (Europe without Switzerland) Heat (MJ) 0.00 0.00 0.04 0.02 Emissions to air Biogenic CO 2 (kg) 1.19 0.91 2.55 1.05 PHA downstream processing Products PHA (kg) 1.00 1.00 1.00 1.00 Technosphere inputs PHA enriched biomass (kg) 2.26 1.57 3.34 1.99 Electricity (kWh) 0.93 0.62 1.61 0.83 market for electricity, high voltage (RER) Steam (MJ) 4.77 3.31 7.06 4.21 steam production, as energy carrier, in chemical industry (RER) Sodium hydroxide (kg) 1.20⋅10 −4 8.37⋅10 −5 1.78⋅10 −4 1.06⋅10 −4 chlor-alkali electrolysis, membrane cell (RER) Sulfuric acid (kg) 7.77⋅10 −5 5.39⋅10 −5 1.15⋅10 −4 6.86⋅10 −5 market for sulfuric acid (RER) SDS (kg) 0.06 0.04 0.09 0.05 market for non-ionic surfactant (GLO) Waste to treatment Wastewater (m 3 ) 0.04 0.52 0.98 0.47 market for wastewater, average (Europe without Switzerland) Anaerobic digestion Intermediate products Biogas (m 3 ) 2.12 0.40 13.98 2.76 Technosphere inputs Solid-rich wastewater stream (kg COD) 7.92 2.35 29.13 7.30 Wastewater (kg COD) 1.89 0.86 3.49 1.46 Electricity (kWh) 0.43 0.18 1.81 0.58 Heat (MJ) 0.00 9.10 0.00 0.00 heat production, natural gas, at industrial furnace >100 kW (Europe without Switzerland) Heat (MJ) 23.49 4.57 40.04 14.83 Waste to treatment Digestate (kg) 35.35 15.16 32.38 20.91 Emissions to air CH 4 (g) 0.87 0.16 5.72 1.13 H 2 S (g) 1.41⋅10 −2 2.7⋅10 −3 9.35⋅10 −2 1.85⋅10 −2 NH 3 (g) 1.41⋅10 −3 2.7⋅10 −4 9.35⋅10 −3 1.85⋅10 −3 CHP e (continued on next page) M. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 9 and up to a 70% of results variance. As previously stated in the section 4.1, this parameter is strongly related to the steam, electricity, and chemicals consumption and thus, it influences considerably the environmental performance of the process. Similarly, the PHA content is responsible for an average 25% of results variance, as it is related to both chemicals and electricity consumption. The pressure drop in the highpressure homogenizer has a negligible effect on the variance of the environmental performance results, except for scenario 1 in freshwater eutrophication and human toxicity. Other parameters such as acidification yield are responsible for up to 30% variance in GWP100 and FDP, as this parameter determines the amount of intermediate waste that is sent to anaerobic digestion and further transformed into heat (free of environmental burdens). Therefore, it affects the environmental burdens of the employed in the process. 5. Discussion In this section, the current LCA results are compared with the available literature on LCA of PHA production from organic wastes. Those LCA studies on PHA production from organic wastes published recently (2014 onwards) were selected (Table F.1). Five out of eight were based on MMC, while the rest employed pure culture. Half of Table 3 (continued) Items Scenario 1 Scenario 2 Scenario 3 Scenario 4 Ecoinvent process Products Electricity (kWh) 5.33 1.02 35.20 4.66 Heat (MJ) 23.97 4.57 158.38 31.27 Technosphere inputs Biogas (m 3 ) 2.12 0.40 13.98 2.76 Emissions to air CH 4 (g) 1.12 0.21 3.91 0.98 CO 2 (kg) 4.04 0.77 26.69 5.27 CO (g) 2.37 0.45 15.65 3.09 N 2 O (g) 0.12 0.02 0.80 0.16 NOx (g) 0.73 0.14 4.81 0.95 NMVOC (g) 0.10 0.02 0.65 0.13 SO 2 (g) 1.24 0.24 8.22 1.62 a Refers to current, 50% and 60% RES electricity mix production processes. b RER =Europe. c Biomass contains a 15% weight of PHA. d GLO =Global. e Note that these values are reported per FU, 1 kg of PHA powder. Fig. 4. Uncertainty results for PHA production in the four scenarios defined and across the evaluated impact categories. M. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 16 C.3 Electricity consumption Upscaling methods used in ex-ante LCA of emerging technologies were followed to estimate utilities consumption (Piccinno et al., 2016; Tsoy et al., 2020). C.3.1 Stirring energy Stirring electric energy is required by the reactors (Eq. ((8)), where N p is a dimensionless number specific to a certain type of impeller (axial or radial) and constant at turbulent flow. N p is 0.79 in the case of axial flow (our case) and the density of the reaction mixture ( ρ mix ) is assumed equal to 1000 kg/m 3 . Estir[J] = Np⋅ ρ mix⋅N3⋅d5⋅t η stir (8) N is the rotational velocity of stirring, calculated as in Eq. (9), where v t is the tip speed assumed equal to 1.66 m/s. N[s−1]=vt π ⋅d(9) The diameter of the impeller (d) is calculated based on the assumption that it measures one third of the reactor diameter (D). Since some energy loss occurs, for example through friction, from the electricity input to the actual stirring, an efficiency value ( η stir ) is included and assumed equal to 0.9. C.3.2 Aeration energy Aeration is required in both enrichment and accumulation reactors. Energy required by the aeration is calculated in Eq. (10), where W ΔS is the minimum isentropic gas work and n air is the molar flow and the compressor efficiency ( η aer ), assumed equal to 0.75. The required air molar flow was estimated assuming an extra 25% of the required air for the VFA consumed in the enrichment and accumulation reactors. The minimum isentropic gas work is calculated in Eq. C.11, where γ is the adiabatic compression factor (dimensionless, 1.4 for ideal gases), Z the compressibility factor (1 for ideal gas), T 1 the temperature before compression (K) and, P 1 and P 2 the in an out pressure of the gas respectively (Pa) Eaer[J] = WΔS⋅˙nair⋅t η aer (Eq. C.10) WΔS[J⋅mol−1]=γ⋅Z⋅R⋅T1 (γ−1)⋅[(P2 P1)(γ−1) /γ−1](Eq. C.11) C.3.3 Filtration An electricity consumption of 10 kWh per ton of dry material was considered in all filtration units (Piccinno et al., 2016). C.3.4 Pumping An electricity consumption of 50 kWh per ton of dry material was considered in PHA production section (Vea et al., 2021), while a 0.155 kWh per cubic meter of stream was considered in the anaerobic digestion (Rodriguez-Verde et al., 2014). C.4 Heat consumption The heating energy required in the reactors is the energy necessary to raise the reaction mixture to a certain temperature and to keep it for the duration of the reaction (Q react ). This is calculated as the sum of the energy for raising the temperature (Q heat ) and the heat loss on the reactor surface (Q loss ) divided by the efficiency of the heating device ( η heat ) (Eq. C.12). To calculate the Q heat (Eq. C.13), the specific heat capacity (C p ), the mass of the reaction mixture (m mix ) and the reactor (T r ) and inlet (T 0 ) temperature of the stream are required. To calculate the heat loss (Eq. C.14), he surface area of the reactor (A), the thermal conductivity of the insulation material (k a ), the thickness of the insulation (s), the temperature difference between the inside and outside of the reactor (ΔT =T r −T out ) and the time of the reaction (t) are needed. Qreact =Qheat +Qloss η heat (Eq. C.12) Qheat =CP⋅mmix⋅(Tr−T0)(Eq. C.13) Qloss =A⋅ka s⋅(Tr−Tout)⋅t(Eq. C.14) C.5 Steam consumption The required steam (Q dry ) in the spray dryer is calculated based on Eq. C.15. For that, the drying efficiency ( η dry ) the liquid’s specific heat capacity (C p,liq ), the mass of the liquid (m liq ), the temperature difference between the boiling and the initial temperature (ΔT =T boil −T out ), the enthalpy of vaporization (ΔH vap ) and the mass of vaporised liquid are required (m vap ). Qdry =CP,liq⋅mliq⋅(Tboil −T0) + ΔHvap⋅mvap η dry (Eq. C.15) M. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 17 C.6 Chemical consumption Chemical consumption is summarized in Table C.2. Sodium hydroxide is consumed in acidogenic fermentation to maintain the pH and it is estimated according to G´ alvez-Martos et al. (2021). Chemical consumption in enrichment reactor is calculated according to Serafim et al. (2004). Chemical consumption in PHA downstream processing (DSP) is estimated according to Saavedra del Oso et al. (2021). Table C.2 Summary of chemical consumption. Operational unit Chemical Conditions/Amount Units Reference Acidification NaOH 0.880 mol NaOH per mol C consumed G´ alvez-Martos et al. (2021) Enrichment NH4Cl 0.160 g/L Serafim et al. (2004) K2HPO4 0.137 g/L Serafim et al. (2004) CaCl2 0.070 g/L Serafim et al. (2004) DSP H2SO4 pH 4 Saavedra del Oso et al. (2021) NaOH pH 10 Saavedra del Oso et al. (2021) SDS 4.00 g/L Saavedra del Oso et al. (2021) C.7 Emissions Emissions are listed in Table C.3. Biogenic carbon dioxide emissions in PHA production were estimated based on COD consumed, while emissions due to leakage in the anaerobic digestion were calculated based on the volume of biogas produced. Emissions in the cogeneration unit. Table C.3 Summary of emissions. Operational unit Flow Conditions/Amount Units Reference Enrichment & Accumulation CO 2 0.37 g/g COD Vea et al. (2021) Anaerobic digestion CH 4 6.50⋅10 −2 %vv biogas Rodriguez-Verde et al. (2014) H 2 S 5.00⋅10 −4 %vv biogas Rodriguez-Verde et al. (2014) NH 3 1.00⋅10 −4 %vv biogas Rodriguez-Verde et al. (2014) CHP CO 2 1.91 kg/m 3 biogas Rodriguez-Verde et al. (2014) CO 1.10⋅10 −3 kg/m 3 biogas Rodriguez-Verde et al. (2014) N 2 O 5.74⋅10 −5 kg/m 3 biogas Rodriguez-Verde et al. (2014) CH 4 5.29⋅10 −4 kg/m 3 biogas Rodriguez-Verde et al. (2014) NO x 3.44⋅10 −4 kg/m 3 biogas Rodriguez-Verde et al. (2014) NMVOC 4.62⋅10 −5 kg/m 3 biogas Rodriguez-Verde et al. (2014) SO 2 5.88⋅10 −4 kg/m 3 biogas Rodriguez-Verde et al. (2014) Appendix D. Electricity mix life cycle inventories For the electricity mix LCI we refer to the provided excel file (APPENDIX D). Appendix E. Environmental evaluation E.1 Characterization Fig. E.1. Electricity consumption within the PHA production. M. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 18 Fig. E.2. Characterization of PHA production within a waste-to-PHA biorefinery and contributions of each process unit (FU: 1 kg of PHA). Table E.1 Qualitative assessment of the technological parameters influences on the background processes’ environmental burdens. Empty cells mean negligible influence. Parameter Electricity Heat 1 Steam Chemicals RES ++ Emissions regulations Y VFA/F + + Y X/VFA + + ++ Y PHA/VFA + Productivity + PHA content in biomass ++ ++ ++ Recovery yield +++ +++ +++ 1 Heat environmental burdens depend mainly on the heat source. E.2 Comparison of allocation methods Allocation method selection can also affect the results. To evaluate the uncertainty related to this selection, the baseline scenarios (100% allocation to PHA and electricity) were compared with substitution and other allocation approaches: COD and primary energy savings (PES) (Frangopoulos, 2012) allocation in VFA separation and CHP respectively, and only minor differences were found among allocation approaches (see Fig. E.3). As biogas production and combustion CO 2 emissions are biogenic, the related environmental impacts are caused mainly by the digestate incineration and the N 2 O, NOx and NMVOC emissions. The related environmental impacts of heat production in the CHP are seven times lower than heat district production from natural gas. Thus, PES allocation results present a significant variance compared to the baseline. Similarly, as VFA separation related environmental impacts are low compared to other units, differences with the baseline allocation method are lower than a 10%. However, substitution method results differ significantly compared to allocation methods. The high amount of avoided electricity produced in scenario 3 favors its environmental performance in both GWP100 and FDP, although the emissions in the AD and CHP hamper its environmental performance in the rest of impact categories. Indeed, the environmental impacts are strongly related to these emissions. Substitution may lead to misunderstandings in decision-making. For instance, if policy makers prioritize the GWP100, scenario 3 could be chosen over other scenarios even though the PHA production environmental performance is worse. Thus, this approach should be avoided. M. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 19 Fig. E.3. Comparison of different approaches to handle multifunctionality: baseline (solid waste stream and heat have no environmental burdens), allocation COD (on VFA separation), allocation PES (on CHP) and substitution (avoided electricity production). Appendix F. Discussion Table F.1 Summary of LCA studies on PHA production from organic wastes published 2014 onwards. Number Year Title Reference 1 2014 Methodological issues in life cycle assessment of mixed-culture polyhydroxyalkanoate production utilising waste as feedstock Heimersson et al. (2014) 2 2015 Microbial community-based polyhydroxyalkanoates (PHAs) production from wastewater: Techno-economic analysis and ex-ante environmental assessment Fern´ andez-Dacosta et al. (2015) 3 2016 Techno-environmental assessment of integrating polyhydroxyalkanoate (PHA) production with services of municipal wastewater treatment Morgan-Sagastume et al. (2016) 4 2020 How sustainable are biopolymers? Findings from a life cycle assessment of polyhydroxyalkanoate production from rapeseed-oil derivatives Nitkiewicz et al. (2020) 5 2020 Environmental assessment of complex wastewater valorization by polyhydroxyalkanoates production Roib´ as-Rozas et al. (2020) 6 2021 Environmental life cycle assessment of polyhydroxyalkanoates production from cheese whey Asunis et al. (2021) 7 2021 Techno-economic evaluation and life-cycle assessment of poly(3-hydroxybutyrate) production within a biorefinery concept using sunflower-based biodiesel industry by-products Kachrimanidou et al. (2021) 8 2021 Inclusion of multiple climate tipping as a new impact category in life cycle assessment of polyhydroxyalkanoate (PHA)-based plastics Vea et al. (2021) References Arvidsson, R., Tillman, A.M., Sand´ en, B.A., Janssen, M., Nordel¨ of, A., Kushnir, D., Molander, S., 2018. Environmental assessment of emerging technologies: recommendations for prospective LCA. J. Ind. Ecol. 22, 1286–1294. https://doi.org/ 10.1111/JIEC.12690. Asunis, F., de Gioannis, G., Francini, G., Lombardi, L., Muntoni, A., Polettini, A., Pomi, R., Rossi, A., Spiga, D., 2021. Environmental life cycle assessment of polyhydroxyalkanoates production from cheese whey. Waste Manag. 132, 31–43. https://doi.org/10.1016/J.WASMAN.2021.07.010. Atasoy, M., Owusu-Agyeman, I., Plaza, E., Cetecioglu, Z., 2018. Bio-based volatile fatty acid production and recovery from waste streams: current status and future challenges. Bioresour. Technol. 268, 773–786. https://doi.org/10.1016/J. BIORTECH.2018.07.042. Bell, J., Paula, L., Dodd, T., N´ emeth, S., Nanou, C., Mega, V., Campos, P., 2018. EU ambition to build the world’s leading bioeconomy—uncertain times demand innovative and sustainable solutions. Nat. Biotechnol. 40, 25–30. https://doi.org/ 10.1016/J.NBT.2017.06.010. Bengtsson, S., Karlsson, A., Alexandersson, T., Quadri, L., Hjort, M., Johansson, P., Morgan-Sagastume, F., Anterrieu, S., Arcos-Hernandez, M., Karabegovic, L., Magnusson, P., Werker, A., 2017. A process for polyhydroxyalkanoate (PHA) production from municipal wastewater treatment with biological carbon and nitrogen removal demonstrated at pilot-scale. N Biotechnol 35, 42–53. https://doi. org/10.1016/J.NBT.2016.11.005. Bergerson, J.A., Brandt, A., Cresko, J., Carbajales-Dale, M., MacLean, H.L., Matthews, H. S., McCoy, S., McManus, M., Miller, S.A., Morrow, W.R., Posen, I.D., Seager, T., Skone, T., Sleep, S., 2020. Life cycle assessment of emerging technologies: evaluation techniques at different stages of market and technical maturity. J. Ind. Ecol. 24, 11–25. https://doi.org/10.1111/JIEC.12954. B¨ orjeson, L., H¨ ojer, M., Dreborg, K.H., Ekvall, T., Finnveden, G., 2006. Scenario types and techniques: towards a user’s guide. Futures 38, 723–739. https://doi.org/ 10.1016/J.FUTURES.2005.12.002. M. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 20 Corella-Puertas, E., Guieu, P., Aufoujal, A., Bulle, C., Boulay, A.M., 2022. Development of simplified characterization factors for the assessment of expanded polystyrene and tire wear microplastic emissions applied in a food container life cycle assessment. J. Ind. Ecol. https://doi.org/10.1111/JIEC.13269. Cucurachi, S., Blanco, C.F., Steubing, B., Heijungs, R., 2022. Implementation of uncertainty analysis and moment-independent global sensitivity analysis for fullscale life cycle assessment models. J. Ind. Ecol. 26, 374–391. https://doi.org/ 10.1111/jiec.13194. Cucurachi, S., van der Giesen, C., Guin´ ee, J., 2018. Ex-ante LCA of emerging technologies. Procedia CIRP 69, 463–468. https://doi.org/10.1016/j. procir.2017.11.005. Delpierre, M., Quist, J., Mertens, J., Prieur-Vernat, A., Cucurachi, S., 2021. Assessing the environmental impacts of wind-based hydrogen production in The Netherlands using ex-ante LCA and scenarios analysis. J. Clean. Prod. 299, 126866 https://doi.org/ 10.1016/J.JCLEPRO.2021.126866. Dietrich, K., Dumont, M.J., del Rio, L.F., Orsat, V., 2017. Producing PHAs in the bioeconomy — towards a sustainable bioplastic. Sustain. Prod. Consum. 9, 58–70. https://doi.org/10.1016/j.spc.2016.09.001. Elginoz, N., Owusu-Agyeman, I., Finnveden, G., Hischier, R., Rydberg, T., Cetecioglu, Z., 2022. Application and adaptation of a scale-up framework for life cycle assessment to resource recovery from waste systems. J. Clean. Prod. 355, 131720 https://doi. org/10.1016/J.JCLEPRO.2022.131720. Est´ evez-Alonso, ´ A., Pei, R., van Loosdrecht, M.C.M., Kleerebezem, R., Werker, A., 2021. Scaling-up microbial community-based polyhydroxyalkanoate production: status and challenges. Bioresour. Technol. 327, 124790 https://doi.org/10.1016/J. BIORTECH.2021.124790. Fern´ andez-Dacosta, C., Posada, J.A., Kleerebezem, R., Cuellar, M.C., Ramirez, A., 2015. Microbial community-based polyhydroxyalkanoates (PHAs) production from wastewater: techno-economic analysis and ex-ante environmental assessment. Bioresour. Technol. 185, 368–377. https://doi.org/10.1016/j.biortech.2015.03.025. Frangopoulos, C.A., 2012. A method to determine the power to heat ratio, the cogenerated electricity and the primary energy savings of cogeneration systems after the European Directive. Energy 45, 52–61. https://doi.org/10.1016/J. ENERGY.2011.12.044. Fritsche, Uwe, Brunori, Gianluca, Chiaramonti, David, Galanakis, C.M., Hellweg, S., Matthews, Robert, Panoutsou, Calliope, 2020. Future Transitions for the Bioeconomy towards Sustainable Development and a Climate-Neutral Economy - Knowledge Synthesis Final Report. Publications Office of the European Union, Luxembourg. G´ alvez-Martos, J.L., Greses, S., Magdalena, J.A., Iribarren, D., Tom´ as-Pej´ o, E., Gonz´ alezFern´ andez, C., 2021. Life cycle assessment of volatile fatty acids production from proteinand carbohydrate-rich organic wastes. Bioresour. Technol. 321, 124528. https://doi.org/10.1016/J.BIORTECH.2020.124528. Guho, N.M., Pokhrel, D., Abbasi, M., McDonald, A.G., Alfaro, M., Brinkman, C.K., Coats, E.R., 2020. Pilot-scale production of poly-3-hydroxybutyrate-co-3hydroxyvalerate from fermented dairy manure: Process performance, polymer characterization, and scale-up implications. Bioresour. Technol. Rep. 12, 100588. https://doi.org/10.1016/j.biteb.2020.100588. H2020 USABLE Packaging, 2019. USEABLE packaging – bio-based raw materials [WWW Document]. USABLE Packaging – Bio-Based Raw Materials. URL. https://www. usable-packaging.eu/. accessed 10.1.21. Heimersson, S., Morgan-Sagastume, F., Peters, G.M., Werker, A., Svanstr¨ om, M., 2014. Methodological issues in life cycle assessment of mixed-culture polyhydroxyalkanoate production utilising waste as feedstock. Nat. Biotechnol. 31, 383–393. https://doi.org/10.1016/j.nbt.2013.09.003. Igos, E., Benetto, E., Meyer, R., Baustert, P., Othoniel, B., 2019. How to treat uncertainties in life cycle assessment studies? Int. J. Life Cycle Assess. 24, 794–807. https://doi.org/10.1007/S11367-018-1477-1/TABLES/3. Kachrimanidou, V., Ioannidou, S.M., Ladakis, D., Papapostolou, H., Kopsahelis, N., Koutinas, A.A., Kookos, I.K., 2021. Techno-economic evaluation and life-cycle assessment of poly(3-hydroxybutyrate) production within a biorefinery concept using sunflower-based biodiesel industry by-products. Bioresour. Technol. 326, 124711 https://doi.org/10.1016/J.BIORTECH.2021.124711. Khatami, K., Perez-Zabaleta, M., Owusu-Agyeman, I., Cetecioglu, Z., 2021. Waste to bioplastics: how close are we to sustainable polyhydroxyalkanoates production? Waste Manag. 119, 374–388. https://doi.org/10.1016/J.WASMAN.2020.10.008. Langkau, S., Steubing, B., Mutel, C., Permana Aije, M., Erdmann, L., Janssen, M., n.d. Guidelines for the Construction of Scenarios for Prospective LCA. Lorini, L., Munarin, G., Salvatori, G., Alfano, S., Pavan, P., Majone, M., Valentino, F., 2022. Sewage sludge as carbon source for polyhydroxyalkanoates: a holistic approach at pilot scale level. J. Clean. Prod. 354, 131728 https://doi.org/10.1016/J. JCLEPRO.2022.131728. Matos, M., Cruz, R.A.P., Cardoso, P., Silva, F., Freitas, E.B., Carvalho, G., Reis, M.A.M., 2021a. Combined strategies to boost polyhydroxyalkanoate production from fruit waste in a three-stage pilot plant. ACS Sustain Chem Eng 9, 8270–8279. https://doi. org/10.1021/acssuschemeng.1c02432. Matos, M., Cruz, R.A.P., Cardoso, P., Silva, F., Freitas, E.B., Carvalho, G., Reis, M.A.M., 2021b. Sludge retention time impacts on polyhydroxyalkanoate productivity in uncoupled storage/growth processes. Sci. Total Environ. 799, 149363. https://doi. org/10.1016/j.scitotenv.2021.149363. Moretto, G., Lorini, L., Pavan, P., Crognale, S., Tonanzi, B., Rossetti, S., Majone, M., Valentino, F., 2020. Biopolymers from urban organic waste: influence of the solid retention time to cycle length ratio in the enrichment of a Mixed Microbial Culture (MMC). ACS Sustain. Chem. Eng. 8 https://doi.org/10.1021/ acssuschemeng.0c04980. Morgan-Sagastume, F., Bengtsson, S., de Grazia, G., Alexandersson, T., Quadri, L., Johansson, P., Magnusson, P., Werker, A., 2020. Mixed-culture polyhydroxyalkanoate (PHA) production integrated into a food-industry effluent biological treatment: a pilot-scale evaluation. J. Environ. Chem. Eng. 8, 104469 https://doi.org/10.1016/j.jece.2020.104469. Morgan-Sagastume, F., Heimersson, S., Laera, G., Werker, A., Svanstr¨ om, M., 2016. Techno-environmental assessment of integrating polyhydroxyalkanoate (PHA) production with services of municipal wastewater treatment. J. Clean. Prod. 137, 1368–1381. https://doi.org/10.1016/j.jclepro.2016.08.008. Morgan-Sagastume, F., Hjort, M., Cirne, D., G´ erardin, F., Lacroix, S., Gaval, G., Karabegovic, L., Alexandersson, T., Johansson, P., Karlsson, A., Bengtsson, S., ArcosHern´ andez, M. v., Magnusson, P., Werker, A., 2015. Integrated production of polyhydroxyalkanoates (PHAs) with municipal wastewater and sludge treatment at pilot scale. Bioresour. Technol. 181, 78–89. https://doi.org/10.1016/J. BIORTECH.2015.01.046. Nessi, S., Sinkko, T., Bulgheroni, C., Garcia-Gutierrez, P., Giuntoli, J., Konti, A., SanyeMengual, E., Tonini, D., Pant, R., Marelli, L., Ardente, F., 2021. Life Cycle Assessment (LCA) of Alternative Feedstocks for Plastics Production Part 1: the Plastics LCA Method. Publications Office of the European Union. https://doi.org/ 10.2760/693062. Nguyenhuynh, T., Yoon, L.W., Chow, Y.H., Chua, A.S.M., 2021. An insight into enrichment strategies for mixed culture in polyhydroxyalkanoate production: feedstocks, operating conditions and inherent challenges. Chem. Eng. J. 420, 130488 https://doi.org/10.1016/J.CEJ.2021.130488. Nitkiewicz, T., Wojnarowska, M., Sołtysik, M., Kaczmarski, A., Witko, T., Ingrao, C., Guzik, M., 2020. How sustainable are biopolymers? Findings from a life cycle assessment of polyhydroxyalkanoate production from rapeseed-oil derivatives. Sci. Total Environ. 749 https://doi.org/10.1016/j.scitotenv.2020.141279. O’Neill, B.C., Kriegler, E., Riahi, K., Ebi, K.L., Hallegatte, S., Carter, T.R., Mathur, R., van Vuuren, D.P., 2014. A new scenario framework for climate change research: the concept of shared socioeconomic pathways. Clim. Change 122, 387–400. https:// doi.org/10.1007/S10584-013-0905-2/TABLES/2. Piccinno, F., Hischier, R., Seeger, S., Som, C., 2016. From laboratory to industrial scale: a scale-up framework for chemical processes in life cycle assessment studies. J. Clean. Prod. 135, 1085–1097. https://doi.org/10.1016/J.JCLEPRO.2016.06.164. Quantis, 2020. The plastic Leak project guidelines [WWW Document]. URL. https://qua ntis.com/report/the-plastic-leak-project-guidelines/. accessed 10.13.22. Rodriguez-Verde, I., Regueiro, L., Carballa, M., Hospido, A., Lema, J.M., 2014. Assessing anaerobic co-digestion of pig manure with agroindustrial wastes: The link between environmental impacts and operational parameters. Sci. Total Environ. 497–498, 475–483. https://doi.org/10.1016/J.SCITOTENV.2014.07.127. Roib´ as-Rozas, A., Mosquera-Corral, A., Hospido, A., 2020. Environmental assessment of complex wastewater valorisation by polyhydroxyalkanoates production. Sci. Total Environ., 140893 https://doi.org/10.1016/j.scitotenv.2020.140893. Roib´ as-Rozas, A., Saavedra del Oso, M., Zarroli, G., Mauricio-Iglesias, M., MosqueraCorral, A., Fiore, S., Hospido, A., 2022. How can we validate the environmental profile of bioplastics? Towards the introduction of polyhydroxyalkanoates (PHA) in the value chains. Assessing Progress Towards Sustainability 405–429. https://doi. org/10.1016/B978-0-323-85851-9.00010-9. Saavedra del Oso, M., Mauricio-Iglesias, M., Hospido, A., 2021. Evaluation and optimization of the environmental performance of PHA downstream processing. Chem. Eng. J. 412, 127687 https://doi.org/10.1016/J.CEJ.2020.127687. Saavedra del Oso, M., Regueira, A., Hospido, A., Mauricio-Iglesias, M., 2022. Fostering the valorization of organic wastes into carboxylates by a computer-aided design tool. Waste Manag. 142, 101–110. https://doi.org/10.1016/J.WASMAN.2022.02.008. Sacchi, R., Terlouw, T., Siala, K., Dirnaichner, A., Bauer, C., Cox, B., Mutel, C., Daioglou, V., Luderer, G., 2022. PRospective EnvironMental Impact asSEment (premise): a streamlined approach to producing databases for prospective life cycle assessment using integrated assessment models. Renew. Sustain. Energy Rev. 160, 112311 https://doi.org/10.1016/J.RSER.2022.112311. Serafim, L.S., Lemos, P.C., Oliveira, R., Reis, M.A.M., 2004. Optimization of polyhydroxybutyrate production by mixed cultures submitted to aerobic dynamic feeding conditions. Biotechnol. Bioeng. 87, 145–160. https://doi.org/10.1002/ BIT.20085. Silva, F., Matos, M., Pereira, B., Ralo, C., Pequito, D., Marques, N., Carvalho, G., Reis, M. A.M., 2022. An integrated process for mixed culture production of 3-hydroxyhexanoate-rich polyhydroxyalkanoates from fruit waste. Chem. Eng. J. 427 https://doi. org/10.1016/j.cej.2021.131908. Stehfest, E., van Vuuren, D., Kram, T., Bouwman, L., Alkemade, R., Bakkenes, M., Biemans, H., Bouwman, A., den Elzen, M., Janse, J., Alkemade, R., Bakkenes, M., Biemans, H., Bouwman, A., den Elzen, M., Janse, J., Lucas, P., van Minnen, J., Muller, C., Prins, A.G., 2014. Integrated Assessment of Global Environmental Change with IMAGE 3.0, Model Description and Policy Applications. PBL Netherlands Environmental Assessment Agency, The Hague (The Hague). Steubing, B., de Koning, D., 2021. Making the use of scenarios in LCA easier: the superstructure approach. Int. J. Life Cycle Assess. 26, 2248–2262. https://doi.org/ 10.1007/S11367-021-01974-2/FIGURES/7. Steubing, B., de Koning, D., Haas, A., Mutel, C.L., 2020. The Activity Browser — an open source LCA software building on top of the brightway framework. Software Impacts 3, 100012. https://doi.org/10.1016/J.SIMPA.2019.100012. Tan, D., Wang, Y., Tong, Y., Chen, G.Q., 2021. Grand challenges for industrializing polyhydroxyalkanoates (PHAs). Trends Biotechnol. 39, 953–963. https://doi.org/ 10.1016/J.TIBTECH.2020.11.010. Thomassen, G., van Dael, M., van Passel, S., You, F., 2019. How to assess the potential of emerging green technologies? Towards a prospective environmental and technoM. Saavedra del Oso et al. Journal of Cleaner Production 383 (2023) 135331 21 economic assessment framework. Green Chem. 21, 4868–4886. https://doi.org/ 10.1039/C9GC02223F. Tsoy, N., Steubing, B., van der Giesen, C., Guin´ ee, J., 2020. Upscaling methods used in ex ante life cycle assessment of emerging technologies: a review. Int. J. Life Cycle Assess. 25, 1680–1692. https://doi.org/10.1007/S11367-020-01796-8/FIGURES/3. Valentino, F., Gottardo, M., Micolucci, F., Pavan, P., Bolzonella, D., Rossetti, S., Majone, M., 2018. Organic fraction of municipal solid waste recovery by conversion into added-value polyhydroxyalkanoates and biogas. ACS Sustain. Chem. Eng. 6, 16375–16385. https://doi.org/10.1021/acssuschemeng.8b03454. Valentino, F., Moretto, G., Lorini, L., Bolzonella, D., Pavan, P., Majone, M., 2019. PilotScale polyhydroxyalkanoate production from combined treatment of organic fraction of municipal solid waste and sewage sludge. Ind. Eng. Chem. Res. 58, 12149–12158. https://doi.org/10.1021/acs.iecr.9b01831. van Vuuren, D.P., Edmonds, J., Kainuma, M., Riahi, K., Thomson, A., Hibbard, K., Hurtt, G.C., Kram, T., Krey, V., Lamarque, J.F., Masui, T., Meinshausen, M., Nakicenovic, N., Smith, S.J., Rose, S.K., 2011. The representative concentration pathways: an overview. Clim. Change 109, 5–31. https://doi.org/10.1007/S10584011-0148-Z/TABLES/4. Vea, E.B., Fabbri, S., Spierling, S., Owsianiak, M., 2021. Inclusion of multiple climate tipping as a new impact category in life cycle assessment of polyhydroxyalkanoate (PHA)-based plastics. Sci. Total Environ. 788, 147544 https://doi.org/10.1016/J. SCITOTENV.2021.147544. Werker, A., Bengtsson, S., Korving, L., Hjort, M., Anterrieu, S., Alexandersson, T., Johansson, P., Karlsson, A., Karabegovic, L., Magnusson, P., Morgan-Sagastume, F., Sijstermans, L., Tietema, M., Visser, C., Wypkema, E., van der Kooij, Y., Deeke, A., Uijterlinde, C., 2018. Consistent production of high quality PHA using activated sludge harvested from full scale municipal wastewater treatment -. PHARIO. Water Sci. Technol. 78, 2256–2269. https://doi.org/10.2166/wst.2018.502. Woods, J.S., Verones, F., Jolliet, O., V´ azquez-Rowe, I., Boulay, A.M., 2021. A framework for the assessment of marine litter impacts in life cycle impact assessment. Ecol. Indicat. 129, 107918 https://doi.org/10.1016/J.ECOLIND.2021.107918. Yadav, B., Pandey, A., Kumar, L.R., Tyagi, R.D., 2020. Bioconversion of waste (water)/ residues to bioplasticsA circular bioeconomy approach. Bioresour. Technol. 298, 122584 https://doi.org/10.1016/J.BIORTECH.2019.122584. M. Saavedra del Oso et al. 75 PAPER IV CARBOXYLCA: ASSISTING DECISION-MAKING IN WASTE VALORIZATION BY COMPUTER-AIDED DESIGN TOOL Manuscript to be submitted to Journal of Industrial Ecology Mateo Saavedra del Oso, Almudena Hospido, Miguel Mauricio-Iglesias CRediT authorship contribution statement Mateo Saavedra del Oso: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Supervision, Visualization, Writing – original draft, Writing – review & editing. PAPER IV 77 CarboxyLCA: Assisting decision-making in waste valorization by computer-aided design tool Mateo Saavedra del Oso*, Miguel Mauricio-Iglesias and Almudena Hospido Cross-disciplinary Research Center in Environmental Technologies (CRETUS), Universidade de Santiago de Compostela, Santiago de Compostela, Spain. * Corresponding Author ([email protected]) Abstract The carboxylate platform presents an excellent opportunity to turn organic waste into valuable chemicals and other products as part of a circular economy. Despite its tremendous potential, the growth of the platform is hindered by technological and financial challenges in linking the key players involved. To tackle these obstacles, this study extends the framework previously established by the authors, broadening its scope to allow for a comprehensive techno-economic and environmental evaluation of potential substrates, technologies, and products. This framework implemented as computer-aided tool consists of the following components: (1) a substrate library, (2) a thorough kinetic and stoichiometric model, (3) an upscaling module, (4) a prospective life cycle assessment module, (5) a life cycle cost module, and (6) a set of comprehensive indicators for interpreting simulation results and facilitating decision making. The effectiveness of the tool is demonstrated through three case studies, showcasing its use in selecting the optimal technology for valorizing a specific organic waste, optimizing anaerobic fermentation for polyhydroxyalkanoates (PHA) production, and choosing the best substrate for PHA production. Finally, the study highlights the tool's utility in supporting decision-making, connecting the key players and its versatility for customization. Keywords Carboxylate platform; volatile fatty acids; resource recovery; prospective life cycle assessment; life cycle cost; industrial ecology. 1. INTRODUCTION Organic waste is a significant challenge in the European Union (EU), where over 118 million tons are generated annually (European Commission, 2019c). The European Circular Economy Action Plan (European Commission, 2020) aims to increase the circularity of waste flows by utilizing residues as feedstock, and by altering consumption patterns and the product policy framework. Also, this plan seeks to attain climate neutrality by promoting greener alternatives that are economically competitive, and the implementation of upcycled value chains, such as the carboxylate platform, can contribute to these goals (Naresh Kumar et al., 2022). Within this platform, organic wastes are valorized into intermediate products, i.e., volatile fatty acids (VFA), alcohols, methane, and hydrogen, that are further converted into a wide range of final products, such as biofuels, biochemicals, or bioplastics (Varghese et al., 2022). Among the carboxylate platform, polyhydroxyalkanoates (PHA) production is gaining attention. PHA production from organic wastes is based on mixed microbial culture (MMC) systems. Concretely, it is a 3-step process where (1) organic wastes are anaerobically fermented into VFA, (2) MMC is enriched in PHA-storing bacteria by imposing a feast/famine regime under nutrient excess, and (3) PHA accumulation where the enriched biomass is fed VFA (Nguyenhuynh et al., 2021). The effectiveness of the MMC system has been tested on various feedstocks such as wastewater (Morgan-Sagastume et al., 2020), household waste (Moretto et MATEO SAAVEDRA DEL OSO 84 The production of PHBV at pH 6.5 has both greater environmental benefits (-67 vs -6 kg CO2-eq) and lower costs (212 vs 247€) per FU (1000 kg COD) compared to pH 5.5. The higher concentration of hydroxyvalerate precursors at pH 6.5, i.e., propionic and valeric acid, results in a reduction of added propionic acid (37.5 vs 0 kg), leading to a decrease in carbon footprint and production costs. Environmental benefits are caused by the avoided district heat (-64 kg CO2-eq), electricity (-45 kg CO2-eq) and PET (-164 kg CO2-eq) production. The carbon footprint is caused by district heat (38 and 49% of GWP100a) and electricity (11 and 15% of GWP100a) consumption. Surfactant employed in the PHA downstream processing accounts for 6% and 7% of GWP100a for both scenarios. Sodium hydroxide, which is used in anaerobic fermentation, is responsible for 5% of GWP100a for pH 6.5. Regarding the LCC contribution, utilities account for 36 and 42% of costs per FU for pH 5.5 and pH 6.5 respectively. Annual depreciation accounts for 32 and 38%, while chemicals contribute a 19 and 7% respectively. Using electricity and district heat produced own site would decrease both environmental impacts and costs. 3.4. Case study 2: screening the environmental performance of valorization routes for tuna canning wastewater Tuna canning wastewater is a side stream of fishery industry. It has a low organic matter content (15.2 g COD/L), although a high content in soluble protein. The environmental performance of three alternative valorization routes was screened. The operational conditions, an HRT of 2 days, an organic loading rate of 7.5 g COD/(L·d) and a pH of 5.5, were chosen according to Saavedra del Oso et al. (2022). Figure 4 depicts the GWP100a characterization results per FU (1000 kg COD). Figure 4. Environmental indicators (GWP100a) for tuna canning wastewater valorization. 29 191 1,037 -800 -600 -400 -200 0 200 400 600 800 1000 1200 1400 (A) Anaerobic digestion (B) PHA production (C) VFA production GWP100a (kg CO2-eq) Avoided acetic acid Avoided electricity Avoided heat district Emissions Digestate treatment Avoided PET Ethyl acetate Surfactant Steam Wastewater treatment Sodium phosphate Ammonium chloride Sodium hydroxide Electricity Heat district Total PAPER IV 85 The low organic matter content (15.2 g COD/L) hampers the environmental performance of the three valorization routes considered. The low organic loading rate in both anaerobic digestion3 and anaerobic fermentation leads to a low biogas production and VFA concentration respectively. Thus, the environmental credits received by avoided production of electricity, district heat, PET and acetic acid are low compared to the environmental impacts. Anaerobic digestion shows the lowest GWP100a (28 kg CO2-eq), as utilities and chemicals consumption are lower than the other valorization routes. The low VFA concentration obtained (8 g COD/L) hampers the valorization into PHA (route B) or pure VFA (route C). Then, the environmental burdens related to the district heat required in the anaerobic fermentation and PHA production are high (338 and 168 kg CO2-eq). Similarly, steam consumption in the solvent recovery within VFA extraction is high (723 kg CO2-eq) as well as the solvent employed (224 kg CO2-eq). Therefore, a co-substrate rich in COD would be necessary for considering other valorization routes such a PHA production. 3.5. Case study 3: screening substrates for PHA production In this third case study, different substrates (organic fraction of the municipal solid waste (OFMSW), tuna canning wastewater and regrind) are screened as potential substrates for PHA production (Figure 5). The evaluation of the environmental and economic performance of these substrates was carried out considering the following operational conditions (used at laband pilot scale) in anaerobic fermentation: pH 5.5, HRT of 6, 2 and 4 days for OFMSW, tuna canning wastewater and regrind pasta respectively, and a scale of 10000 t COD per year. Conditions were chosen according to pilot-scale studies and prior work (Moretto et al., 2020; Saavedra del Oso et al., 2022; Silva et al., 2022). The other parameters are included in Supplementary Information S1. Both OFMSW and regrind pasta show environmental benefits (-65 and -73 kg CO2-eq respectively) when added as co-substrates. Tuna canning wastewater environmental performance in GWP100a (191 CO2-eq) is hampered by its high heat consumption per FU. This high consumption is caused by tuna canning wastewater’s low COD content (15 g COD·L-1), as higher volume is used for FU (1000 kg COD). However, tuna canning wastewater has higher PHA yield than OFMSW and regrind pasta (18 and 6% respectively). District heat is the main contributor to GWP100a, being responsible for 52%, 76% and 51% of environmental impacts in OFMSW, tuna canning wastewater and regrind pasta respectively. Finding an alternative source for district heat could decrease overall GWP100a substantially. With regards to the LCC, OFMSW and regrind pasta costs are almost a 40% lower than tuna canning wastewater. Utilities consumption (district heat) represent 66% of total costs for tuna canning wastewater. 3 An HRT of 12 days for anaerobic digestion is fixed. MATEO SAAVEDRA DEL OSO 86 Figure 5. Environmental indicators (GWP100a) and economic indicators (cost per FU) for substrates screening as feedstock for PHA production. 4. DISCUSSION CarboxyLCA enables the holistic assessment of substrates potential to be valorized within the carboxylate platform. As demonstrated in section 3.3., process developers can screen which are the optimal operational conditions for a substrate and technology from a technical, environmental, and economic perspective. Similarly, it is possible to explore the different valorization pathways for a certain substrate considering technical, environmental, and economic indicators (section 3.5.). Namely, the developed framework enables the decision support, identification of hotspots and design improvement. Besides, its modular structure -89 191 -74 -400 -300 -200 -100 0 100 200 300 400 500 OFMSW Tuna caning wastewater Regrind pasta GWP100a (kg CO2-eq) Avoided electricity Avoided PET Avoided heat district Emissions Digestate treatment Surfactant Steam Wastewater treatment Sodium phosphate Ammonium chloride Sodium hydroxide Electricity Heat district Total 0 € 100 € 200 € 300 € 400 € OFMSW Tuna caning wastewater Regrind pasta AD Chemicals Utilities Maintenance Labor PAPER IV 87 enables the individual use of the modules. For instance, modules 2, 3 and 4 could be deactivated for screening the technical feasibility of organic waste to be valorized into VFA. Similarly, module 2 could be deactivated when enough information on the anaerobic fermentation is available. Although this tool does not cover all valorization pathways within carboxylate platform (Atasoy et al., 2018), its modular approach allows for flexible customization. Thus, it is possible to add new downstream processes for the transformation of VFA into other added-value compounds. It is also possible to adjust or deactivate modules based on the desired outcome. For instance, the upscaling of certain valorization routes can be deactivated when the aim is to assess potential substrates or optimize the process. This computer aided design tool also enables practitioners to integrate their own data. For instance, PHA process developers can include parameters related to the PHA production or downstream processing or study the effects of these ones in the process environmental or economic performance. 5. CONCLUSIONS In conclusion, the present study presents a comprehensive framework for a preliminary holistic assessment of organic waste streams valorization within carboxylate platform. This framework provides valuable insights into the carboxylate platform, addressing the gaps by incorporating the environmental and economic perspective in the decision-making process. The tool's utility is demonstrated through case studies that showcase its ability to screen optimal substrates, technologies, and products within the carboxylate platform. The findings from this study highlight the significance of the developed tool in advancing the principles of circular economy and fostering collaboration among various stakeholders in the value chain. With a wide range of expertise and interests, this tool serves as a bridge that connects these stakeholders, enabling the promotion of sustainable practices in the valorization of organic waste. ACKNOWLEDGMENTS This work has been financially supported by the USABLE Packaging project (Call: H2020BBI-JTI-2018, EU ID: 836884). Mateo Saavedra del Oso, Miguel Mauricio-Iglesias and Almudena Hospido belong to a Galician Competitive Research Group (GRC ED431C 2021/37). The programme is co-funded by FEDER (UE). REFERENCES Atasoy, M., Owusu-Agyeman, I., Plaza, E., Cetecioglu, Z., 2018. Bio-based volatile fatty acid production and recovery from waste streams: Current status and future challenges. Bioresour Technol 268, 773–786. https://doi.org/10.1016/J.BIORTECH.2018.07.042 Cucurachi, S., Van Der Giesen, C., Guinée, J., 2018. Ex-ante LCA of emerging technologies. Procedia CIRP 69, 463–468. https://doi.org/10.1016/j.procir.2017.11.005 European Commission, 2020. A new Circular Economy Action Plan for a cleaner and more competitive Europe, COM(2020) 98 final. European Commission, 2019. Guidance for separate collection of municipal waste. Publications Office. https://doi.org/10.2779/691513 MATEO SAAVEDRA DEL OSO 88 Hauschild, M.Z., Rosenbaum, R.K., Olsen, S.I., 2017. Life Cycle Assessment: Theory and Practice. Life Cycle Assessment: Theory and Practice 1–1216. https://doi.org/10.1007/978-3-319-56475-3/COVER Igos, E., Benetto, E., Meyer, R., Baustert, P., Othoniel, B., 2019. How to treat uncertainties in life cycle assessment studies? International Journal of Life Cycle Assessment 24, 794– 807. https://doi.org/10.1007/S11367-018-1477-1/TABLES/3 Matos, M., Cruz, R.A.P., Cardoso, P., Silva, F., Freitas, E.B., Carvalho, G., Reis, M.A.M., 2021. Combined Strategies to Boost Polyhydroxyalkanoate Production from Fruit Waste in a Three-Stage Pilot Plant. ACS Sustain Chem Eng 9, 8270–8279. https://doi.org/10.1021/acssuschemeng.1c02432 Moretto, G., Lorini, L., Pavan, P., Crognale, S., Tonanzi, B., Rossetti, S., Majone, M., Valentino, F., 2020. Biopolymers from urban organic waste: Influence of the solid retention time to cycle length ratio in the enrichment of a Mixed Microbial Culture (MMC). ACS Sustain Chem Eng 8. https://doi.org/10.1021/acssuschemeng.0c04980 Morgan-Sagastume, F., Bengtsson, S., De Grazia, G., Alexandersson, T., Quadri, L., Johansson, P., Magnusson, P., Werker, A., 2020. Mixed-culture polyhydroxyalkanoate (PHA) production integrated into a food-industry effluent biological treatment: A pilot-scale evaluation. J Environ Chem Eng 8, 104469. https://doi.org/10.1016/j.jece.2020.104469 Naresh Kumar, A., Sarkar, O., Chandrasekhar, K., Raj, T., Narisetty, V., Mohan, S.V., Pandey, A., Varjani, S., Kumar, S., Sharma, P., Jeon, B.H., Jang, M., Kim, S.H., 2022. Upgrading the value of anaerobic fermentation via renewable chemicals production: A sustainable integration for circular bioeconomy. Science of The Total Environment 806, 150312. https://doi.org/10.1016/J.SCITOTENV.2021.150312 Nguyenhuynh, T., Yoon, L.W., Chow, Y.H., Chua, A.S.M., 2021. An insight into enrichment strategies for mixed culture in polyhydroxyalkanoate production: feedstocks, operating conditions and inherent challenges. Chemical Engineering Journal 420, 130488. https://doi.org/10.1016/J.CEJ.2021.130488 Nouaille, R., Pessiot, J., 2017. Process for extracting carboxylic acids produced by anaerobic fermentation from fermentable biomass. https://patents.google.com/patent/US20170022446A1/en Piccinno, F., Hischier, R., Seeger, S., Som, C., 2016. From laboratory to industrial scale: a scale-up framework for chemical processes in life cycle assessment studies. J Clean Prod 135, 1085–1097. https://doi.org/10.1016/J.JCLEPRO.2016.06.164 Rodriguez-Verde, I., Regueiro, L., Carballa, M., Hospido, A., Lema, J.M., 2014. Assessing anaerobic co-digestion of pig manure with agroindustrial wastes: The link between environmental impacts and operational parameters. Science of The Total Environment 497–498, 475–483. https://doi.org/10.1016/J.SCITOTENV.2014.07.127 Saavedra del Oso, M., Mauricio-Iglesias, M., Hospido, A., Steubing, B., 2023. Prospective LCA to provide environmental guidance for developing waste-to-PHA biorefineries. J Clean Prod 383, 135331. https://doi.org/10.1016/J.JCLEPRO.2022.135331 Saavedra del Oso, M., Regueira, A., Hospido, A., Mauricio-Iglesias, M., 2022. Fostering the valorization of organic wastes into carboxylates by a computer-aided design tool. Waste Management 142, 101–110. https://doi.org/10.1016/J.WASMAN.2022.02.008 PAPER IV 89 Seider, W.D., Lewin, D.R., Seader, J.D., Widagdo, S., Gani, R., Ng, K.M., 2016. Product and process design principles : synthesis, analysis, and evaluation. Fourth ed. Wiley Siew Ng, K., Farooq, D., Yang, A., 2021. Global biorenewable development strategies for sustainable aviation fuel production. https://doi.org/10.1016/j.rser.2021.111502 Silva, F., Matos, M., Pereira, B., Ralo, C., Pequito, D., Marques, N., Carvalho, G., Reis, M.A.M., 2022. An integrated process for mixed culture production of 3hydroxyhexanoate-rich polyhydroxyalkanoates from fruit waste. Chemical Engineering Journal 427. https://doi.org/10.1016/j.cej.2021.131908 Stehfest, E., van Vuuren, D., Kram, T., Bouwman, L., Alkemade, R., Bakkenes, M., Biemans, H., Bouwman, A., den Elzen, M., Janse, J., Alkemade, R., Bakkenes, M., Biemans, H., Bouwman, A., den Elzen, M., Janse, J., Lucas, P., van Minnen, J., Muller, C., Prins, A.G., 2014. Integrated assessment of global environmental change with IMAGE 3.0, Model description and policy applications, The Hague: PBL Netherlands Environmental Assessment Agency. The Hague. Steubing, B., de Koning, D., Haas, A., Mutel, C.L., 2020. The Activity Browser — An open source LCA software building on top of the brightway framework. Software Impacts 3, 100012. https://doi.org/10.1016/J.SIMPA.2019.100012 Valentino, F., Moretto, G., Lorini, L., Bolzonella, D., Pavan, P., Majone, M., 2019. Pilot-Scale Polyhydroxyalkanoate Production from Combined Treatment of Organic Fraction of Municipal Solid Waste and Sewage Sludge. Ind Eng Chem Res 58, 12149–12158. https://doi.org/10.1021/acs.iecr.9b01831 Varghese, V.K., Poddar, B.J., Shah, M.P., Purohit, H.J., Khardenavis, A.A., 2022. A comprehensive review on current status and future perspectives of microbial volatile fatty acids production as platform chemicals. Science of The Total Environment 815, 152500. https://doi.org/10.1016/J.SCITOTENV.2021.152500 SUPPORTING INFORMATION Supporting information S1 Supporting information S2 91 PAPER V COMPARATIVE LIFE CYCLE ASSESSMENT OF PHA-BASED CONSUMER ITEMS FOR DAILY USE Manuscript to be submitted to Resources, Conservation and Recycling Mateo Saavedra del Oso, Rakesh Nair, Miguel Mauricio-Iglesias, Almudena Hospido CRediT authorship contribution statement Mateo Saavedra del Oso: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Software; Supervision; Visualization; Roles/Writing - original draft; Writing - review & editing. PAPER V 93 Comparative life cycle assessment of PHA-based consumer items for daily use Mateo Saavedra del Oso1,*, Rakesh Nair2, Miguel Mauricio-Iglesias1 and Almudena Hospido1 1 CRETUS, Department of Chemical Engineering, Universidade de Santiago de Compostela, 15782, Santiago de Compostela, Spain 2 Bio Base Europe Pilot Plant, Rodenhuizekaai 1, 9042 Ghent, Belgium * Corresponding Author ([email protected]) Abstract Polyhydroxyalkanoates (PHA) face an outstanding opportunity to substitute oilbased polymers in food packaging applications. H2020 project USABLE Packaging has developed an innovative value chain based on mixed microbial cultures (MMC) using food industry wastes and byproducts as feedstock to produce prototypes of PHA-based items intended for food contact (reusable plates, frozen bags and biscuit bags) that were compared to their commercial counterparts, so the expected better environmental performance can be checked. To do so, a cradle-to-grave life cycle assessment (LCA), including the long-term environmental impacts caused by microplastics, was carried out. The system was modeled integrating all available knowledge, from pilot-scale data to process simulators, and following upscaling frameworks. PHA-based items outperform their commercial counterparts since, they show environmental benefits thanks to the avoided electricity obtained in the cogeneration heat and power (CHP) unit within the PHA production; however, these environmental benefits are very sensitive to the substituted electricity environmental burdens. PHA downstream processing remains as the main environmental hotspot due to solvents and utilities usage. Recommendations are provided: using the biogas produced within PHA production to produce utilities and substituting DMC-ethanol solvents by acetone-water. However, the different levels of development and market implementation are seen as critical and therefore attention should be paid on how upscaling affects the results and interpretation of the LCA while serving as a guide for process and product development. Keywords Bioplastics; cradle-to-grave; end-of-life; life cycle assessment; upscaling Abbreviations CHP: combined heat and power EoL: end-of-life FU: functional unit LCA: life cycle assessment LDPE: low-density polyethylene LR: loss rate MMC: mixed microbial culture OTR: oxygen transfer rate PBS: polybutylene succinate PDF: potentially disappeared fraction of species PEF: product environmental footprint PHA: polyhydroxyalkanoates PP: polypropylene MATEO SAAVEDRA DEL OSO 100 Stage Unit process Product Amount Unit Durum wheat pasta Intermediate products Regrind pasta 79.1 t Technosphere inputs Semolina production 79.8 t Tap water 24.5 m3 Natural gas 2000.2 m3 Electricity 12649.2 kWh Heavy fuel oil 1.3 t Emissions to air Particulate Matter, < 2.5 µm 1.19·10-3 kg Sulfur dioxide 3.64·10-2 kg Nitrogen oxides 1.58·10-2 kg Dinitrogen monoxide 6.17·10-3 kg Carbon monoxide, fossil 1.09·10-3 kg Carbon dioxide, fossil 8.4 t Hydrocarbons 1.26·10-3 kg PHA production Anaerobic fermentation Technosphere inputs Regrind pasta 79.1 t Tap water 268.8 m3 Heat 379.8 MJ Electricity 53.5 kWh Sodium bicarbonate 2489.8 kg VFA separation Technosphere inputs Electricity 1212.0 kWh Biomass selection Technosphere inputs Tap water 3334.7 m3 Heat 831.7 MJ Electricity 2049.0 kWh Sodium hydroxide 2742.8 kg PHA accumulation Intermediate products PHA-enriched biomass 2000.0 kg Technosphere inputs Heat 84.7 MJ Electricity 1380.2 kWh Propionic acid 841.9 kg Anaerobic digestion Technosphere inputs Heat 1022.8 MJ Electricity 72.0 kWh Emissions to air Methane, non-fossil 25.0 kg Hydrogen sulfide 0.4 kg Ammonia 4.08·10-2 kg Waste to treatment Digestate 40.7 t CHP Byproducts and coproducts Heat (byproduct) 678962.7 MJ PAPER V 101 Stage Unit process Product Amount Unit Electricity (coproduct) 150880.6 kWh Emissions to air Methane, non-fossil 33.2 kg Carbon monoxide, non-fossil 70.4 kg Carbon dioxide, non-fossil 120.0 t Nitrogen oxides 3.6 kg NMVOC 2.9 kg Nitrogen oxides 21.6 kg Sulfur oxides 36.9 kg PHA downstream processing Pretreatment Technosphere inputs Heat 1310.0 MJ Electricity 50.3 kWh Solvent extraction Technosphere inputs Heat 15.8 MJ Electricity 10.3 kWh Dimethyl carbonate 903.1 kg PHA recovery Products PHA 1000.0 kg Technosphere inputs Steam 770.2 MJ Electricity 11.6 kWh Cooling energy 13890.4 MJ Ethanol 2709.2 kg Solvent recovery Technosphere inputs Steam 131039.8 MJ Cooling energy 130387.5 MJ Emissions to air Dimethyl carbonate 903.1 kg Ethanol 2709.2 kg Table 2. Life cycle inventory of both PHA-based (P) and commercial (C) items (gate-to-grave) for food contact: reusable plates, frozen bags and biscuits bag. Sta ge Product Reusable plates (P) Reusable plates (C) Frozen bags (P) Frozen bags (C) Biscuits bag (P) Biscuits bag (C) Un it Compounding and shaping Products Item 1 1 1 1 1 1 uni t Technosphere inputs PHA 1.26·10-2 1.25·10-2 2.21·10-2 kg PS 1.26·10-2 kg PBS 5.34·10-3 5.15·10-2 kg LDPE 1.78·10-2 kg PP 5.88·10-2 kg Paper 1.48·10-2 kg MATEO SAAVEDRA DEL OSO 102 Sta ge Product Reusable plates (P) Reusable plates (C) Frozen bags (P) Frozen bags (C) Biscuits bag (P) Biscuits bag (C) Un it Electricity 0.53 0.53 0.53 0.53 0.53 0.53 kW h Emissions to water Microplastics 1.52·10-7 1.52·10-7 2.13·10-7 2.13·10-7 8.83·10-7 7.06·10-7 kg EoL Waste to treatment Biowaste 1.26·10-2 7.36·10-2 kg PS waste 1.26·10-2 kg PP waste 5.88·10-2 kg LDPE waste 1.78·10-2 kg Paper waste 1.48·10-2 kg Emissions to water Microplastics 9.68·10-7 9.68·10-7 3.41·10-5 3.41·10-5 2.32·10-4 1.86·10-4 kg 4. RESULTS In this section, results of the environmental assessment are presented: characterization, normalization, and weighting of environmental impacts (section 4.1) and contribution analysis (section 4.2). 4.1. Characterization, normalization, and weighting Table 3 summarizes: (i) the characterization results for EF impact categories and physical effects on biota, (ii) the relevance of the impact category on the normalized and weighted results and (iii) the EF single score for each item. Normalized and weighted results per impact category are included in Appendix C. PHA-based items outperform their commercial peers from an environmental perspective, even showing environmental benefits (Table 4). Ionizing radiation, freshwater eutrophication, water use, and use of mineral and metal resources are the most relevant categories for PHAbased items based on the normalized and weighted results (see Appendix C). Climate change and particulate matter only show relevance for PHA-based biscuit bags. Nevertheless, climate change and use of fossil resources are the most relevant categories for their commercial counterparts. Besides, PHA-based items show substantially lower physical effects on biota caused by microplastics than their commercial counterparts. Table 3. Characterization, impact category relevance after normalization and weighting, and overall single scores of PHA-based and commercial items for food contact (bold values indicates the best performance between pairs). Midpoint indicator Reusable plates (P) Reusable plates (C) Frozen food packaging (P) Frozen food packaging (C) Biscuits bag (P) Biscuits bag (C) Unit EF Characterization Climate change -2.86·10-4 6.83·10-2 5.85·10-2 8.30·10-2 0.57 9.25·10-2 kg CO2eq PAPER V 103 Midpoint indicator Reusable plates (P) Reusable plates (C) Frozen food packaging (P) Frozen food packaging (C) Biscuits bag (P) Biscuits bag (C) Unit Ozone depletion 3.24·10-8 3.03·10-11 3.32·10-8 3.59·10-10 6.88·10-8 1.50·10-9 kg CFC11-eq Human toxicity, cancer effects 1.54·10-10 1.08·10-11 2.00·10-10 2.43·10-11 7.27·10-10 6.67·10-11 CTUh Human toxicity, non-cancer effects 3.51·10-9 1.11·10-10 3.77·10-9 4.19·10-10 9.11·10-9 9.16·10-10 CTUh Particulate matter 1.85·10-8 1.88·10-9 1.98·10-8 2.21·10-9 4.74·10-8 4.38·10-9 disease incidenc e Ionising radiation, human health -0.32 1.20·10-5 -0.32 4.87·10-3 -0.52 1.06·10-2 kBq U235-eq Photochemical ozone formation – human health 6.93·10-4 1.28·10-4 8.12·10-4 2.29·10-4 2.45·10-3 4.20·10-4 kg NMVOCeq Acidification 4.64·10-4 1.75·10-4 6.39·10-4 2.28·10-4 2.56·10-3 4.09·10-4 mol H+- eq Eutrophication, terrestrial 2.30·10-3 3.07·10-4 2.62·10-3 4.86·10-4 7.45·10-3 1.18·10-3 mol N-eq Eutrophication, freshwater -3.74·10-4 5.56·10-7 -3.56·10-4 1.40·10-5 -5.38·10-4 1.09·10-4 kg N-eq Eutrophication marine 3.95·10-4 3.22·10-5 4.18·10-4 5.15·10-5 9.63·10-4 1.99·10-4 kg N-eq Land use 3.77 0.03 3.82 0.15 7.58 0.48 CTUe Ecotoxicity freshwater 4.78 5.05·10-3 4.82 0.19 9.41 11.11 pt Water use 0.79 2.92·10-2 0.82 3.79·10-2 1.78 4.84·10-2 m3 Resource use, fossils -3.71 1.05 -2.90 1.53 0.89 1.74 MJ Resource use, minerals and metals 4.55·10-6 6.88·10-9 4.65·10-6 1.89·10-7 9.60·10-6 2.97·10-7 kg Sb-eq physical effects on biota* 2.03·10-5 8.14·10-3 0.04 0.25 0.50 1.36 CTUe EF Normalization and Weighting EF single score -7.85·10-15 5.94·10-16 -7.21·10-15 1.05·10-15 -6.61·10-15 2.06·10-15 * Physical effects on biota are not included in the normalized and weighted EF single score results 4.2. Contribution analysis Contribution analysis of processes for PHA-based items is showed in Figure 2. PHA-based items show similar contribution analysis results as they have equivalent EoL and are composed of PHA. Differences are caused by the different composition of PHA and PBS. The (avoided) electricity produced at the CHP provides environmental benefits in all impact categories, but specially in ionizing climate change, human toxicity cancer effects, ionizing radiation, acidification, freshwater eutrophication, freshwater ecotoxicity, water use and use of fossil resources. Utilities (district heat, steam and cooling energy) production contribute substantially to climate change, ozone depletion, human toxicity cancer effects, particulate matter, acidification, freshwater ecotoxicity, water use, use of fossil resources, and use of mineral and metals resources. Solvents (DMC and ethanol) production cause a relevant environmental burden in climate change, human toxicity non-cancer effects, particulate matter, acidification, MATEO SAAVEDRA DEL OSO 104 terrestrial eutrophication, marine eutrophication, land use, freshwater ecotoxicity and use of mineral and metal resources. Chemicals (propionic acid, sodium bicarbonate, sodium hydroxide) production shows a relevant environmental burden in ozone depletion, particulate matter, acidification, terrestrial eutrophication and use of mineral and metal resources. Own emissions are the main contributor to photochemical ozone formation. Copolymer (PBS) production has a significant environmental burden in climate change, particulate matter and use of fossil resources for biscuit bags. Other processes such as electricity production, feedstock provision, waste treatment or end-of-life treatment have negligible contribution to overall impact categories. Regarding the commercial items, polystyrene production is the main contributor within reusable plates production to climate change (67%), carcinogenic human toxicity (37%), particulate matter formation (96%), photochemical oxidant formation (96%), acidification (97%), terrestrial eutrophication (93%) and marine eutrophication (81%). Waste polystyrene treatment contributes to climate change (32%), carcinogenic human toxicity (36%), noncarcinogenic human toxicity (53%) and freshwater ecotoxicity (63%). Similarly, polyethylene production is the main contributor within frozen bags manufacturing to most impact categories, being waste management especially relevant to climate change (37%). Environmental burdens within biscuit bags production are shared by paper production, propylene production and propylene waste management. Paper and its waste management dominate some categories such as terrestrial eutrophication (41%), land use (41%) or freshwater ecotoxicity (36%). An endpoint approach (ecosystem quality) was chosen to measure the contribution of physical effects on biota. As shown in Figure 3, microplastic impacts in marine ecosystem might contribute to up to 81% of overall environmental impacts in ecosystem quality for commercial frozen bags and biscuit bags. Microplastic impacts in marine ecosystems contribution to ecosystem quality for prototypes are almost negligible (reusable plates and frozen bags) and low (biscuits bags) due to lower CF (2.49·10-4 - 2.93·10-2 vs 4.12·10-2-9.98·102), as PHA-based items are composed by biodegradable polymers. However, these results are very uncertain as these methodologies are still in its early development. For further details see Appendix A. PAPER V 105 Figure 2. Contribution analysis of PHA-based items for food contact: reusable plates (a), frozen bags (b) and biscuit bags (c). -100% -50% 0% 50% 100% Climate change Ozone depletion Human toxicity, cancer… Human toxicity, non-… Particulate matter Ionising radiation,… Photochemical ozone… Acidification Eutrophication, terrestrial Eutrophication, freshwater Eutrophication marine Land use Ecotoxicity freshwater Water use Resource use, fossils Resource use, minerals… own emissions electricity utilities chemicals solvents feedstock copolymer waste treatment end-of-life treatment avoided electricity -100% -50% 0% 50% 100% Climate change Ozone depletion Human toxicity, cancer… Human toxicity, non-… Particulate matter Ionising radiation,… Photochemical ozone… Acidification Eutrophication, terrestrial Eutrophication, freshwater Eutrophication marine Land use Ecotoxicity freshwater Water use Resource use, fossils Resource use, minerals… own emissions electricity utilities chemicals solvents feedstock copolymer waste treatment end-of-life treatment avoided electricity -100% -50% 0% 50% 100% Climate change Ozone depletion Human toxicity, cancer… Human toxicity, non-… Particulate matter Ionising radiation,… Photochemical ozone… Acidification Eutrophication, terrestrial Eutrophication, freshwater Eutrophication marine Land use Ecotoxicity freshwater Water use Resource use, fossils Resource use, minerals… own emissions electricity utilities chemicals solvents feedstock copolymer waste treatment end-of-life treatment avoided electricity a) b) c) MATEO SAAVEDRA DEL OSO 106 Figure 3. Contribution of physical effects on biota to ecosystem quality endpoint impact category. 5. DISCUSSION In this section, the environmental hotspots of PHA-based items’ value chain (section 5.1) and the limitations of the environmental assessment (section 5.2) are discussed. 5.1. Hotspot analysis Ionizing radiation, freshwater eutrophication, water use, and use of mineral and metal resources were pointed out as the most relevant impact categories in section 4.1. Figure 4 depicts the hotspot analysis of the life cycle stages of PHA-based items for these impact categories. Climate change was included to enable the comparability with other research studies. CHP is presented separately from PHA production to enable the visibility of PHA production environmental burdens. Independent of the item assessed, PHA downstream processing is the main environmental hotspot in climate change, water use and use of minerals and metal resources. These environmental burdens are due to solvent production and solvent recovery, as high amounts of steam (131 MJ per kg PHA) and cooling energy are required there (Table 1). These results are aligned with previous LCA on PHA downstream processing (Saavedra del Oso et al., 2023, 2021). When disregarding avoided electricity credits, PHA production contributes to climate change, ionizing radiation, freshwater eutrophication and use of mineral and metal resources. Propionic acid, sodium hydroxide and sodium bicarbonate, which are consumed in biomass selection and anaerobic fermentation respectively, are the main contributors within PHA production. The shaping and compounding stage contribution to these impact categories depends on the item and its copolymer content. Namely, frozen bags and (especially) biscuit bags’ shaping, and compounding contribution is relevant to climate change, freshwater eutrophication and use of mineral and metal resources due to PBS production. 0% 20% 40% 60% 80% 100% Reusable plates (C) Reusable plates (U) Frozen food packaging (C) Frozen food packaging (U) Biscuits bag (C) Biscuits bag (U) ecosystem quality microplastics PAPER V 107 Feedstock provision and EoL contribution to these impact categories is negligible. Physical effects on biota are caused by microplastics generated from the plastic leakage in the EoL (see Appendix A). However, it is important to quantify their environmental burdens of these two life cycle stages. Another consideration is that the environmental impacts caused by chemicals and additives during the mismanaged EoL of these items is not addressed. Figure 4. Hotspot analysis of PHA-based items for food contact: reusable plates (a), frozen bags (b) and biscuit bags (c). a) b) c) -100% -80% -60% -40% -20% 0% 20% 40% 60% 80% Climate change Ionising radiation, human health Eutrophication, freshwater Water use Resource use, minerals and metals Feedstock production PHA production PHA downstream processing Shaping and compounding EoL CHP -100% -80% -60% -40% -20% 0% 20% 40% 60% 80% Climate change Ionising radiation, human health Eutrophication, freshwater Water use Resource use, minerals and metals Feedstock production PHA production PHA downstream processing Shaping and compounding EoL CHP -100% -80% -60% -40% -20% 0% 20% 40% 60% 80% Climate change Ionising radiation, human health Eutrophication, freshwater Water use Resource use, minerals and metals Feedstock production PHA production PHA downstream processing Shaping and compounding EoL CHP MATEO SAAVEDRA DEL OSO 108 5.2. Limitations and recommendations In this work, the environmental performance of PHA-based items for food contact is analyzed and compared to their commercial peers. However, limitations regarding the process upscaling and the system modelling should be considered. PHA-based items outperform their commercial peers in some impact categories such as climate change, where the lower environmental impacts (or even environmental benefits) are due to the avoided electricity produced at the CHP and, therefore, are very sensitive to the substituted background activity. Here the European average electricity (“electricity production (high voltage) | RER”) was selected and other electricity mixes with a higher share on renewable energies (e.g., Sweden) may affect the overall result. Although PHA downstream processing was carried out at pilot scale, the solvent-biomass ratio was not optimized (ratio of 40:1), and solvent recovery was not tested. Previous findings show that the environmental performance of downstream processes is very sensitive to the extraction yield (Saavedra del Oso et al., 2023). Hence, the obtained results are uncertain as the assumed solvent-biomass ratio is relatively low (9:1) and the extraction yield (95%) has not yet been proven at pilot scale. However, measures can be taken to minimize the environmental impacts. The biogas obtained in the valorization of intermediate wastes could be used in a boiler to produce utilities that are used in the solvent recovery. Thus, environmental impacts could be reduced up to a 40%. Another reported solution would be using acetone and water as solvent and antisolvent (Vermeer et al., 2022), as solvent recovery would entail lower utilities consumption and it would avoid using ethanol, whose production has high environmental burdens (Saavedra del Oso et al., 2021). Finally, shaping and compounding were modelled assuming that both PHA-based and commercial items weight the same. However, the weight of PHA-based prototypes depends on the items’ required thickness and the grammage of the compounds. Thus, there is certain uncertainty regarding this assumption that unfortunately could not be solved here due to confidentiality issues. 6. CONCLUSIONS The environmental performance of PHA-based items for food contact has been evaluated and compared to their conventional pairs, and recommendations to improve it are provided. To the best of our knowledge, this study evaluates for first time the environmental performance of PHA-based items produced at pilot-scale using mixed microbial cultures. Besides, physical effects on biota caused by microplastics generation from plastic leakage into marine environments are included. As conclusions, valuable insights are extracted from this work: • PHA-based items outperform their commercial peers from an environmental perspective. Indeed, they show environmental benefits. • Those environmental benefits are caused by avoided electricity produced at CHP. Thus, these environmental benefits are very sensitive to the substituted electricity’s environmental burdens. • PHA downstream processing is the main environmental hotspot due to solvents and utilities usage. Results are also very uncertain due to assumptions and sensitiveness to solvent biomass ratio and extraction yield. PAPER V 109 • Recommendations are provided: using the obtained biomass to produce utilities consumed in the solvent recovery and changing DMC-ethanol by acetone-water as solvent and antisolvent. • Due to their biodegradable character, PHA-based items show lower physical effects on biota caused by microplastics generated in the manufacturing and EoL. By evaluating a selection of PHA-based food packaging items and integrating the information provided by key stakeholders, this work contributes to the development of a sustainable PHA value chain within a circular economy approach. CREDIT AUTHOR STATEMENT Mateo Saavedra del Oso: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Software; Supervision; Visualization; Roles/Writing - original draft; Writing - review & editing. Rakesh Nair: Data curation, Writing - review & editing. Miguel MauricioIglesias: Conceptualization; Funding acquisition; Project administration; Supervision; Validation; Visualization; Roles/Writing - original draft; Writing - review & editing. Almudena Hospido: Conceptualization; Funding acquisition; Methodology; Project administration; Supervision; Validation; Visualization; Roles/Writing - original draft; Writing - review & editing. DECLARATION OF COMPETING INTEREST The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ACKNOWLEDGMENTS This work has been financially supported by the USABLE Packaging project (Call: H2020BBI-JTI-2018, EU ID: 836884). Mateo Saavedra del Oso, Miguel Mauricio-Iglesias and Almudena Hospido belong to a Galician Competitive Research Group (GRC ED431C 2021/37), programme co-funded by FEDER (UE). Finally, USABLE Packaging project partners are deeply recognized for providing data. REFERENCES Asunis, F., de Gioannis, G., Francini, G., Lombardi, L., Muntoni, A., Polettini, A., Pomi, R., Rossi, A., Spiga, D., 2021. Environmental life cycle assessment of polyhydroxyalkanoates production from cheese whey. Waste Management 132, 31–43. https://doi.org/10.1016/J.WASMAN.2021.07.010 Bevilacqua, M., Braglia, M., Carmignani, G., Zammori, F.A., 2007. Life cycle assessment of pasta production in Italy. J Food Qual 30, 932–952. https://doi.org/10.1111/j.17454557.2007.00170.x Broeren, M.L.M., Kuling, L., Worrell, E., Shen, L., 2017. Environmental impact assessment of six starch plastics focusing on wastewater-derived starch and additives. Resour Conserv Recycl 127, 246–255. https://doi.org/10.1016/J.RESCONREC.2017.09.001 CHAPTER 4: DISCUSSION 117 4.1. CONSISTENCY OF THE THESIS The development of biobased and circular value chains entails multiple opportunities and challenges (Santagata et al., 2021) and involves multiple actors across the value chain with different interests and available knowledge. The PHA value chain based on MMC systems could be depicted as a sort of puzzle, whose different pieces are assembled in this thesis by integrating the available knowledge, tools, and methodologies (Figure 4.1). Paper I covered the first two stages of the PHA value chain (i.e. feedstock production and PHA production) and focused on the anaerobic fermentation of organic wastes into VFA. From the identification of gaps regarding the stoichiometry variability and the disintegration and hydrolysis steps, it integrated the existing metabolic and kinetic models as well as the available information regarding substrates composition into a framework implemented as a computeraided design tool. Such tool is constituted by (i) a library of substrates, (ii) an integral kinetic and stoichiometric model, and (iii) a set of indicators to interpret simulation results and assist the decision making. The library of substrates and the mathematical model parameters was stored in a spreadsheet, while the mathematical model was implemented in MATLAB (R2021a). The framework can connect stakeholders with different interests and expertise; so, for instance, food industry actors can screen their wastes and side stream as potential substrates for VFA production; or PHA manufacturers can screen the substrates that are suitable for obtaining a specific PHA. The framework assembled the pieces needed to build the upstream of the PHA value chain focusing on the technical performance of the production of VFA in terms of product selectivity, productivity and titer. Moving along the PHA value chain, Paper II focused on the evaluation and optimization of the environmental performance of PHA downstream processing. Information gathered thorough literature review of available PHA recovery processes at industrial and lab-scale was integrated with process simulation and detailed process calculations for scenarios definition and process upscaling. The upscaled processes were assessed using LCA and LCC. Mechanical disruption was pointed out as the most promising technology for recovering PHA from both environmental and economic perspective. Solvent-based PHA downstream processes are highly energy intensive due to solvent recovery and their environmental burden could only be decreased when using biofuels or integrating the utilities production within a biorefinery. The LCA and LCC results can assist PHA manufacturers to choose the appropriate recovery technology. Paper III integrated prior Paper I and Paper II scopes to a cradle-to-gate approach, covering the feedstock provision, the PHA production and PHA downstream processing, and expanded the temporal boundaries. It depicted a future framework where waste-to-PHA biorefineries are implemented at industrial scale and aimed to identify the key parameters for their sustainable development. To do so, pLCA was combined with scenario methodology to derive four scenarios. The scenario methodology and choice of technologies were supported by stakeholders’ inputs and learnings obtained in Paper I and II. Concretely, mechanical disruption was chosen for the PHA downstream processing and the parameters for PHA recovery were set according to the outcomes of the literature review. Similarly, the substrates and the scenario parameters for anaerobic fermentation were chosen based on the pilot-scale data available and the learnings from Paper I. MATEO SAAVEDRA DEL OSO 118 Figure 4.1 Knowledge integration as a puzzle that bridges the bottlenecks that hinder the development of waste-to-PHA biorefineries. Feedstock production PHA production PHA downstream processing Shaping and compounding Use EoL Paper IV Connecting and engaging stakeholders Covering gaps and developing a framework for fostering carboxylate platform Paper III How waste-to-PHA biorefineries could develop in the future? Identifying the key parameters for the sustainable development of waste-to-PHA biorefineries Paper V Addressing long-term environmental impacts due to plastic mismanagement and demonstrating a better environmental performance that commercial counterparts Validating the environmental performance of waste-to-PHA biorefineries from cradle-to-grave Paper I Anaerobic fermentation complexity Covering gaps and developing a framework for fostering carboxylate platform Paper II PHA downstream processing as an environmental and economic hotspot Analyzing PHA downstream processing alternatives and proposing improvement actions CHAPTER 4: DISCUSSION 119 The findings of Paper III are aligned with the outcomes of Paper I and II. The choice of substrate has a strong influence on the process environmental performance, as it determines the yield in acidification, biomass selection and PHA accumulation steps. Likewise, the recovery yield and PHA content were found as the key parameters for the environmental performance, accounting for up to 50% variance. As highlighted in Paper II, using the biogas, which is produced in the valorization of intermediate waste streams, to produce utilities that are used within the biorefinery could improve overall system’s environmental performance. This work also provided additional insights on how the different background scenarios could influence on the development of waste-to-PHA biorefineries. Technical performance of anaerobic fermentation is not the only criterion to select a waste stream to produce PHA or VFA. Paper IV, which was built on the basis of Paper I, provided a whole framework to screen both organic wastes and valorization technologies from a holistic perspective. This framework was implemented as a computer-aided design tool by including 3 additional modules, and environmental and economic indicators. The tool is constituted by (1) a substrate library, (2) a kinetic and stoichiometric model, (3) an upscaling and flowchart design module, (4) a prospective life cycle assessment module, (5) a life cycle cost module, and (6) a set of comprehensive indicators for interpreting simulation results and facilitating decision making. Paper IV integrated the mathematical model (Paper I) into the upscaling and LCA flowchart module, which includes three alternative value chains for valorizing organic wastes. The projected scenarios and LCA flowcharts were defined based on the literature and the learning from Papers II and III. This work enabled the holistic screening of organic wastes as substrate for different valorization pathways. The findings from Paper IV highlighted the significance of the developed tool in advancing the principles of circular economy and fostering collaboration among various stakeholders in the value chain. With a wide range of expertise and interests, this tool serves as a bridge that connects these stakeholders, enabling the promotion of sustainable practices in the valorization of organic waste. Finally, and covering the complete waste-to-PHA value chain, Paper V validated the environmental performance of the Usable Packaging value chain. This work expanded its scope integrating both compounding and shaping and EoL, considering the whole value chain from cradle-to-grave. Paper V also covered the identified gap regarding the long-term environmental impacts caused by microplastics generated during manufacturing. It integrated the latest methodological developments. Paper V integrated the learnings from former papers in the upscaling of pilot-scale data from USABLE Packaging partners using advanced process calculations and process simulation. Paper V demonstrated the environmental sustainability of this innovative value chain, unlocking its potential to be deployed at industrial scale. Besides, it proposed improvements actions for the sustainability of waste-to-PHA biorefineries. The consistency of this thesis on bridging the bottlenecks that hinder the development of waste-to-PHA biorefineries is certainly proven. This thesis assisted the decision-making during the development of this innovative value chain at multiple levels: from solving the local bottlenecks (zoom in) within the anaerobic fermentation (Paper I) and the PHA downstream processing (Paper II) to forecast the development of waste-to-PHA biorefineries and developing a framework (zoom out) that connects stakeholders across the value chain (Paper II and IV). These papers provided insights that aid on assembling the different pieces that are necessary to validate the sustainable performance of waste-to-PHA biorefineries. Hence, Paper IV and Paper V acted as the keystone that integrated all the knowledge generated in former papers, the literature and the pilot-scale data provided by the USABLE partners. MATEO SAAVEDRA DEL OSO 120 4.2. CRITICAL ANALYSIS OF THE METHODS USED IN THE KNOWLEDGE INTEGRATION APPROACH The knowledge integration approach applied in this thesis can be a useful approach to assist the decision-making process during the early-stage development. However, its outcomes should be critically reviewed to avoid biased conclusions, and this is why it is discussed here the strengths and limitations of the methodologies and tools developed and used. 4.2.1. Mathematical Modelling The mathematical model developed in this thesis (Paper I) integrated an existing bioenergetic model and a kinetic model that describes the production of VFA and other products in the mixed culture fermentation of organic substrates. Concretely, it predicts the acidogenic stoichiometry regarding the pH, HRT and substrate composition and covers the disintegration and hydrolysis steps to macromolecules (carbohydrates, proteins, and lipids) and their monomers (glucose, amino acids, and long chain fatty acids). This mathematical model was implemented as a module of the computer-aided design tool in MATLAB (2021a). The substrate library and the kinetic model parameters are stored in a spreadsheet. Likewise, the bioenergetic model parameters are stored in a different spreadsheet. This enables practitioners with higher expertise or available knowledge to tune these parameters. Its modular implementation enables different functionalities: (i) screening a substrate or mixture of substrates for given operational conditions (pH and HRT), (ii) mapping operational conditions (pH and HRT) for a given substrate or substrates mixture, (iii) mapping the pH and substrates mixture for a given HRT. When the stoichiometric coefficients for the acidogenic fermentation are known, it is possible to switch off the bioenergetic model. The results are stored as .mat files and transformed into easy interpretable indicators and graphs that are saved as spreadsheet and images respectively. The mathematical model also presented certain limitations with regards to the reactor configuration, stoichiometry, and description of reactions. It was implemented for a CSTR configuration and thus, its accuracy predicting the acidogenic stoichiometry in other configurations such sequencing batch reactors (SBR) might be lower. The bioenergetic model disregarded the lipid cofermentation. Hence, its accuracy predicting the acidogenic stoichiometry of fat-rich substrates is expected to be low. It was unable to depict the chain elongation reactions, as the drivers are not sufficiently known. Besides, the fermentation yield sensitivity to kinetic parameters (specially disintegration and hydrolysis) might constrain the results accuracy. Therefore, it requires some expertise on anaerobic fermentation and modelling when higher accuracy is needed. Nevertheless, the aim of the mathematical model is to provide data and indicators for assisting the decision-making. Its modular character allows modifying and expanding the modules without affecting the overall performance. Therefore, it is possible to implement other reactor configurations, add the description of other reactions and so on. Paper IV integrated the mathematical model with the upscaling framework, pLCA and LCC. Within the expanded computer-aided design tool, the outcomes of the mathematical model are transformed into easy interpretable technical indicators, but also used in the upscaling. Concretely, the concentrations and steady state, operational conditions and chemical consumption are employed for the projected data estimation. CHAPTER 4: DISCUSSION 121 4.2.2. Upscaling Framework This thesis followed the existing framework on upscaling (Tsoy et al., 2020), which is three-step method: (i) projected scenario definition, (ii) preparation of a projected LCA flowchart and (iii) projected data estimation (Tsoy et al., 2020). Paper II carried out a literature review for all the three stages, complementing the projected data estimation with process simulation and reference books on chemical engineering process design (Seider et al., 2016; Ulrich et al., 2003). In Paper III, the scenario development methodology for pLCA (Langkau et al., Unpublished work) was used for stage (i). This four-step procedure (identification of influencing parameters, construction of sub-scenarios for each parameter, creation of scenarios from sub-scenarios and implementation of scenarios in the LCA model) was supported by the literature review of pilot-scale PHA production and validated by USABLE Packaging partners. The learnings obtained in Paper II were employed in the preparation of the projected LCA flowchart (ii) and the projected data estimation (iii). Concretely, mechanical disruption was chosen as the PHA downstream technology and the LCA flowchart for this step was designed as in Paper II. Electricity and utilities consumption were estimated according to Piccinno et al. (2016) based on the TRL, technology and available data. Hence, Paper III combined data at laband pilot scale and advance process calculations. Paper IV upscaling was implemented in MATLAB (2021a) as a module of the computeraided design tool. The projected scenario definition and projected LCA flowchart integrated the learnings from former Papers II and III. With regards to the projected data estimation, it combined laband pilot-scale data, mathematical modelling, process simulation and detailed process calculations. The substrate-related parameters (composition, disintegration kinetic parameters, scale and operational conditions) are stored in a .xml file. Similarly, all parameters related to the anaerobic fermentation downstream processes, i.e., VFA separation, anaerobic digestion, PHA production, PHA downstream processing and VFA downstream processing, are stored in a separated .xml file. Thus, its modular approach enables changing these parameters without affecting the overall tool function. The upscaling module combines the inputs from the mathematical model and this .xml file and solves the mass and energy balances. Based on the balances, units are dimensioned, and utilities are estimated according to Piccinno et al. (2016). Paper V upscaling integrates the pilot-scale data from USABLE Packaging partners and the knowledge from former papers. The projected scenario definition and projected LCA flowchart is carried out based on the pilot-scale configuration and the learnings from Paper II and III. Concretely, solvent recovery was modelled in Aspen Plus based on assumption from Paper II. Likewise, the valorization of intermediate wastes into biogas and its transformation into electricity and district heat was upscaled according to Paper III. Thus, the upscaling throughout this thesis developed by integrating data and learnings from papers. 4.2.3. Lifecycle based tools The purpose of any LCA is to identify and quantify the environmental impacts of a product/process/service over its entire life cycle and assist decision-makers on developing sustainable alternatives. It involves a comprehensive analysis of the inputs and outputs of the system defined and considers various environmental impact categories. Besides, LCC is a life cycle-based method used in the economic assessment of a product/process/service to account for the monetary costs associated, including capital costs, operation, maintenance, and disposal costs. MATEO SAAVEDRA DEL OSO 122 In Paper II, LCA and LCC enabled comparing different technologies and finding the environmental and economic hotspots of PHA downstream. Regarding the former, results robustness was evaluated through a local sensitivity analysis and improvement actions were proposed. Concerning the latter, capital and operational costs are estimated based on mass and energy balances and using reference books on chemical engineering process design (Seider et al., 2016; Ulrich et al., 2003). As the system boundaries only cover the PHA downstream processing, revenue was not considered. The learnings from the LCC are applied to Paper IV, where the LCC was implemented in MATLAB 2021a as module of the computer-aided design tool. Materials and utilities cost are stored in a spreadsheet, while the capital costs estimation is based on the upscaling module outputs and the equations from literature (Seider et al., 2016; Ulrich et al., 2003). The main shortcoming of LCC is the data uncertainty and the limited availability of cost databases. However, the modular configuration enables practitioners to add their own cost data or include an uncertainty analysis. Going one step further, pLCA determines the environmental impacts of an innovative/current product or process in a future framework. It can then be a useful tool for identifying potential environmental impacts early in the development process, but it should be used in conjunction with other methods and with a clear understanding of its limitations. Paper III combined the scenario methodology with pLCA to derive systematically scenarios on how waste-to-PHA biorefineries could develop in the future. As waste-to-PHA biorefineries represent a multifunctional system, Paper III investigated the effects of different methods to handle multifunctionality, i.e. substitution and allocation based on COD and primary energy savings. For the selected impact categories, negligible differences were found between the different allocation methods, while substitution approach decreases the environmental impacts in climate change and fossil resources depletion. The goal of the computer-aided design tool developed in Paper IV is to compare different valorization technologies within the carboxylate value chain and thus, substitution approach is employed to handle the multifunctionality system. The chosen substituted activities entail different degrees of uncertainty. For instance, electricity is modeled as avoided electricity using the Europe without Switzerland mix. Results might differ depending on the electricity profile. Likewise, PHA and VFA are modeled respectively as avoided PET and acetic acid production in Europe. These assumptions are based on the available activities in the ecoinvent database. However, these products are not equivalent. Similarly, Paper V environmental results depend on the avoided electricity produced within the PHA biorefinery. Although this assumption may lead to biased conclusions, the influence of this decision has been discussed. It is expected than in the future waste-to-PHA biorefinery the biogas produced in anaerobic digestion would be used in the cogeneration heat and power (CHP) unit to cover the utilities requirements and produce electricity. Thus, the related environmental impacts due to utilities production would be substantially lower than employing natural gas. 4.3. A FRAMEWORK TO ENGAGE WITH DIFFERENT INTERESTS AND AVAILABLE KNOWLEDGE By applying knowledge integration, stakeholders across the value chain can establish relationships with one another, gain valuable insights and take informed decisions during the early-stage development of waste-to-PHA biorefineries. It facilitates then the implementation of circular-based solutions and mitigating both environmental and economic risks. The current section discusses how the different stakeholders (food industry and waste management, PHA CHAPTER 4: DISCUSSION 123 manufacturing, packaging users, designers and converters, policymaking, and academia and R&D) can benefit from the knowledge integration approach. 4.3.1. Food industry and waste management stakeholders Food industry represents a diverse and heterogeneous group of subsectors: dairy industry, agriculture, fish caning industry, food beverage industry, etc. Thus, a wide range of waste and side streams are produced, e.g. rich-protein waste streams such as tuna canning wastewater, sugar-rich waste stream such as fruit waste, etc. These waste and side streams, which are currently treated or valorized into energy or animal feed, could be valorized into higher addedvalue products, such as VFA or PHA (Yadav et al., 2020). Likewise, wastewater treatment plants and specially, urban waste facilities produce two wastes with high organic content: primary and secondary sludge, and the organic fraction of the municipal solid waste. These wastes, which are currently valorized into energy, composted, or landfilled, could be used as a substrate for the carboxylate platform (Yadav et al., 2020). The framework developed in this thesis (Figure 4.1) could be used to evaluate the feasibility of integrating these organic waste streams and side streams into the carboxylate platform. Practitioners can screen the technical potential of their organic wastes to be converted into VFA. Based on the technical indicators provided by the computer-aided design tool, i.e. substrate conversion, VFA yield, selectivity and equivalent hydroxyvalerate ratio, they can take an informed decision on whether to proceed with labor pilot-scale experiments or establish synergies with PHA manufacturers or academia and R&D stakeholders for exploring new substrates. Besides, the tool implemented in Paper IV enables them to estimate the potential environmental impacts and costs. Demonstrating a better environmental and economic performance would boost the development of these valorization technologies. 4.3.2. PHA manufacturers Currently PHA is produced at an industrial scale using only pure cultures. However, there are some projects that have explored the production of PHA based on MMC systems. Indeed, there are two demonstration plant biorefineries under construction in Spain and Italy (Circular Biocarbon, 2023). This thesis validates the environmental performance of waste-to-PHA biorefineries. PHA manufacturers might be tempted to explore these production systems, as they offer cheaper and lower environmental burdens compared to pure cultures and food competing feedstocks. PHA manufacturers could use the tools developed within this thesis to screen which organic wastes or mixture of them satisfy their requirements to obtain a PHA with specific properties. Concretely, they could perform a fast holistic assessment to a wide catalogue of organic waste streams prior to any labor pilot-scale experiment, obtaining not only technical indicators but also environmental and economic. The knowledge generated in both Paper II and Paper III helps identifying which PHA downstream technologies might be adequate or which key parameters they should pay attention to ensure a sustainable performance. Practitioners could use these learnings in the upscaling and development of their own processes. Thus, optimizing the technologies when there are still opportunities for major alterations. MATEO SAAVEDRA DEL OSO 124 4.3.3. Packaging users, designers, and converters Packaging users, designers and converters play a key role in the deployment and market expansion of materials such as PHA. One of the main bottlenecks for PHA deployment at industrial scale is finding niche applications where the volumes produced can fulfill the demand (Estévez-Alonso et al., 2021). Usually packaging users, especially in the food sector, are also producers of substantially high amounts of organic wastes. For instance, pasta industry actors produce side streams, such as regrind pasta, and control the upstream of the raw materials used in the manufacturing process. It may be possible for them to become their own suppliers of raw materials for packaging, increasing the circularity within their value chain. Attracted by its environmental performance, packaging designers could consider the inclusion of PHA as candidate to substitute conventional plastics as mono material for packaging or coating barrier for paperboard solutions. The findings of Paper V regarding the EoL of PHA-based items versus commercial counterparts highlight the environmental benefits of PHA. Thus, packaging designers and converters could develop packaging solutions that are fully renewable, compostable, and biodegradable. 4.3.4. Policymakers Over the past years, the European Commission has taken steps to establish a new legal framework within the European Green Deal (European Commission, 2019b) and the Circular Economy Plan (European Commission, 2020) for a cleaner and more comprehensive Europe. These initiatives include as some key points sustainable design and production, improving waste management, encouraging the production of biobased products and the circular use of bioresources. The knowledge integration approach applied within this thesis pursues both the sustainable development of a biobased value chain and an efficient circular use of bioresources. Policymakers may benefit from the framework proposed in this thesis. The computer-aided design tool developed in Paper I and IV enables a fast screening of the potential valorization routes for organic wastes and side streams within the carboxylate platform. Likewise, the insights provided in Paper III on how the policies implemented may affect the development of waste-to-PHA biorefineries can be useful in decision-making. The findings of Paper V highlighted the relevance of including the EoL and the need for initiatives such as PLP or MariLCA. 4.3.5. Academia and R&D The knowledge integration approach to assist the decision-making in the development of waste-to-PHA biorefineries can be particularly helpful for academia and R&D actors. Likewise, academia and R&D actors can be useful providing experimental data. Practitioners can employ the computer-aided design tool implemented in Paper I and Paper IV to screen the potential of a substrate to be converted into VFA and other products. They can also test the optimal operational conditions prior to any experimental work, hence as a tool for experimental design. It is possible to study which substrates may be complementary in cofermentation to obtain a targeted product composition. For instance, tuna canning wastewater, which is a protein-rich substrate but has a low COD content, might require a co-substrate to reach a certain VFA composition and concentration. Practitioners can study the feasibility of a broad range of substrates and mixing ratios. Besides, they can obtain economic and environmental indicators of all cofermentation scenarios. CHAPTER 4: DISCUSSION 125 The outcomes of Paper II and III can serve to select, design and upscale new extraction methods. Practitioners can use the available information to upscale and evaluate the environmental performance of their processes. For instance, they can focus on the optimization of the key parameters for the sustainable performance of PHA downstream processing, e.g., the biomass-solvent ratio content. Practitioners can screen which solvents enable a lower ratio and easier recovery while fulfilling the purity and properties requirements. 4.4. PROSPECTS FOR THE DEVELOPMENT OF WASTE-TO-PHA BIOREFINERIES This thesis provides insights on how to promote waste-to-PHA biorefineries. Concretely, it assists the decision-making process during the development of an innovative PHA value chain and addresses the main challenges. The purpose of this section is to discuss the prospects for the development of PHA production based on MMC systems. 4.4.1. Building a supply chain upstream: risks and opportunities Organic waste is available in very large quantities in the EU. The environmental and economic potential of PHA, which could benefit from biomass as a cost-effective source of raw materials, is subjected to various regional and seasonal factors that impact its feasibility, such as availability (and its current valuation), geographical location, transportation distance, and processing technology. Geographic proximity plays a significant role in determining the technical and economic feasibility of valorization efforts, and legal and traceability issues may also impact the selection of the most suitable organic wastes. To mitigate concerns about seasonality and storage capacity, using multiple sources of raw materials may be a viable option if it does not result in additional logistics costs. Additionally, focusing on food processing residues in areas with a strong agro-industrial presence can help to reduce collection and pre-treatment difficulties. It is essential to note that increasing demand for a particular organic waste could transform it into a valuable commodity, leading to increased costs. As more projects and studies focus on valorizing biomass and co-products for a wide range of applications, particularly energy production, it is crucial to monitor emerging valorization technologies when considering a specific organic waste. This thesis depicts how waste-to-PHA biorefineries could develop in the future and provides a framework for evaluating their feasibility from a holistic perspective. It also enables comparing this waste-to-PHA biorefinery approach with other valorization technologies. Further research should incorporate regional LCA. Further research should include methodologies such as Ïnput-Output LCA or Material Flow Analysis to quantify the supply chain constrains due availability, geographical location, and transportation distance. 4.4.2. Future waste-to-PHA biorefineries The insights provided within this thesis aim to lessen the bottlenecks that hinder the implementation of waste-to-PHA biorefineries. Figure 4.2 depicts the flowchart of how the waste-to-PHA biorefinery approach would look like. The biorefinery should be able to process different organic wastes. This flexibility would have several benefits, from the possibility of cofermenting substrates to obtain a specific VFA composition or obtaining different VFA streams from individual fermentation, to handle substrate availability and seasonality.