Autotrophic nitrogen removal in granular sequencing batch reactors

UNIVERSIDADE DE SANTIAGO DE COMPOSTELA Departamento de Ingeniería Química Autotrophic nitrogen removal in granular sequencing batch reactors Memoria presentada por José Ramón Vázquez Padín Para optar al grado de Doctor por la Universidad de Santiago de Compostela Santiago de Compostela, Junio de 2009 UNIVERSIDADE DE SANTIAGO DE COMPOSTELA Departamento de Ingeniería Química Ramón Méndez Pampín, Catedrático de Ingeniería Química y Anuska Mosquera Corral, Profesora Contratada Doctora de Ingeniería Química de la Universidad de Santiago de Compostela, Informan: Que la memoria titulada “Autotrophic nitrogen removal in granular sequencing batch reactors”, que para optar al grado de Doctor de Ingeniería Química, Programa de Doctorado en Ingeniería Química y Ambiental, presenta Don José Ramón Vázquez Padín, ha sido realizada bajo nuestra inmediata dirección en el Departamento de Ingeniería Química de la Universidad de Santiago de Compostela. Y para que así conste, firman el presente informe en Santiago de Compostela, 24 de junio de 2009. Ramón Méndez Pampín Anuska Mosquera Corral Table of contents Objetivos y resumen 1 Obxectivos e resumo 7 Objectives and summary 11 Chapter 1. Introduction 15 1.1. Water scarcity and water pollution in the world 16 1.2. Legislation framework 16 1.3. Presence of nitrogen compounds in the environment 19 1.3.1. Toxicity of nitrogen compounds 21 1.3.2. Eutrophication 21 1.4. Nitrogen removal from wastewater 22 1.5. Nitrification-denitrification processes 23 1.5.1. Nitrification 23 1.5.2. Partial nitrification 26 1.5.3. Denitrification 28 1.5.4. Anaerobic ammonium oxidation (Anammox) 28 1.5.5. Combination of processes 30 1.6. Nitrogen removal treatment plants 31 1.6.1 Conventional WWTP configuration 31 1.6.2. Main limitations of WWTP: management of reject water and excess sludge 32 1.6.3. Bio-augmentation 33 1.6.4. Partial Nitrification/Denitrification 34 1.6.5. Autotrophic nitrogen removal: Partial nitrification/Anammox 36 1.6.6. Comparison of the available technologies to treat reject water 39 1.7. Systems based on biofilms/granules 40 1.7.1. Modelling of granular systems 41 1.8. References 43 Chapter 2. Materials and methods 49 2.1. Liquid phase 50 2.1.1. Total organic carbon 50 2.1.2. Ammonia nitrogen 50 2.1.3. Inorganic anions: NO2-, NO3-, Cl-; PO43and SO4251 2.1.4. Other control parameters 52 i Table of contents 2.2. Biomass characterisation 52 2.2.1. Sludge volume index 52 2.2.2. Granules density 53 2.2.3. Average diameter of the granules 53 2.2.4. Specific Anammox activity assays 54 2.2.5. Electron microscopy and micro-analysis 55 2.3. Microbiological determinations 56 2.3.1. Identification of bacteria populations by FISH 56 2.3.2. Cryosectioning and confocal microscopy 58 2.4. Microelectrodes 59 2.4.1. O2 microelectrode 60 2.4.2. NO2microelectrode 61 2.4.3. Experimental setup 62 2.5. References 63 Chapter 3. Nitrifying granular systems: a suitable technology to obtain stable partial nitrification at room temperature 65 3.1. Introduction 66 3.2. Objectives 67 3.3. Materials and methods 67 3.3.1. Reactor description 67 3.3.2. Operational conditions 68 3.3.3. Analytical methods 69 3.3.4. Biomass activity measurements 70 3.3.5. Calculations 70 3.4. Results and discussion 72 3.4.1. Reactor operation 72 3.4.2. Biomass physical properties 74 3.4.3. Identification of nitrifying populations 76 3.4.4. Physical stability of the granules 77 3.4.5. Partial nitrification 77 3.4.6. Effects of DO 78 3.4.7. Effects of free ammonia and free nitrous acid 80 3.4.8. Granular SBR as an alternative system for partial nitrification 81 3.5. Conclusions 83 3.6. References 83 ii Table of contents Chapter 4. Population dynamics of nitrite oxidizers in nitrifying granules 87 4.1. Introduction 88 4.2. Objectives 88 4.3. Materials and methods 88 4.3.1. Reactor description 88 4.3.2. Operational conditions 89 4.3.3. Analytical methods 89 4.3.4. Calculations 90 4.4. Results and discussion 91 4.4.1. Biomass characteristics 91 4.4.2. Reactor performance 92 4.4.3. Microbial characterization of the biomass 93 4.5. Conclusions 95 4.6. References 96 Chapter 5. Granular systems to improve Anammox biomass retention 99 5.1. Introduction 100 5.2. Objectives 100 5.3. Materials and methods 101 5.3.1. Reactor description 101 5.3.2. Composition of the feeding media 101 5.3.3. Inocula 102 5.3.4. Analytical methods 102 5.3.5. Specific Anammox activity tests and specific nitrogen loading rate 102 5.4. Results and discussion 103 5.4.1. Reactor operation 103 5.4.2. Biomass retention 104 5.4.3. Reasons for the improvement of biomass retention 105 5.4.4. Applications 106 5.5. Conclusions 107 5.6. References 108 Chapter 6. Development of the CANON process in a continuously aerated granular SBR and an air pulsing SBR to treat anaerobic digester supernatants at room temperature 109 6.1. Introduction 110 6.2. Objectives 111 6.3. Materials and methods 112 iii Table of contents 6.3.1. Reactors description 112 6.3.2. Operational conditions 112 6.3.3. Inocula 113 6.3.4. Operational strategy 113 6.3.5. Analytical methods 114 6.3.6. Calculations 115 6.4. Results and discussion 117 6.4.1. Nitrogen removal 117 6.4.2. Biomass physical properties 120 6.4.3. Identification of bacteria populations by FISH 123 6.4.4. Partial nitrification and spontaneous Anammox growth in the granular SBR 124 6.4.5. Comparison of the start up strategies of CANON reactors 126 6.4.6. Comparison of nitrogen removal performances in different reactor configurations 127 6.5. Conclusions 129 6.6. References 129 Chapter 7. Microbial community distribution and activity dynamics of granular biomass in a CANON reactor 133 7.1. Introduction 134 7.2. Objectives 135 7.3. Materials and methods 135 7.3.1. Reactor description 135 7.3.2. kLa measurement 136 7.3.3. Microscale experiments 136 7.3.4. NO2and O2 microsensors 137 7.3.5. Analytical methods 137 7.3.6. Calculations 138 7.4. Results and discussion 140 7.4.1. Operation of the CANON SBR 140 7.4.2. Identification of bacteria populations by FISH 142 7.4.3. Microprofiles measurements 143 7.4.4. From microscale results to granular CANON operation 149 7.5. Conclusions 151 7.6. References 151 iv Table of contents v Chapter 8. Modelling aerobic granular SBR at variable COD/N ratios including accurate description of total solids concentration 155 8.1. Introduction 156 8.2. Objectives 157 8.3. Materials and methods 157 8.3.1. Granular Sequencing Batch Reactor operation 157 8.3.2. Analytical methods 158 8.4. Model development 158 8.4.1. Definition of the SBR operation 159 8.4.2. Biological processes 159 8.4.3. Oxygen transfer 161 8.4.4. Granules description 161 8.4.5. Modelling the total solids concentration and density of the granules 162 8.4.6. Porosity of granules 162 8.5. Results and discussion 165 8.5.1. Model validation at different COD/N ratios 165 8.5.2. Distribution of the particulate compounds at different COD/N ratios 168 8.5.3. Predicted profiles of particulate compounds and DO concentrations inside the granules 169 8.6. Conclusions 171 8.7. References 172 8.8. Appendixes 175 Conclusiones 181 Conclusións 185 Conclusions 187 List of symbols 189 List of publications 193 Obxectivos e resumo Esta tese encádrase no marco da depuración de augas e máis concretamente na eliminación de nitróxeno das augas residuais. A eliminación de nutrientes (principalmente N e P) nas augas residuais é necesaria para evitar a proliferación de algas ou eutrofización das augas receptoras. Os procesos biolóxicos son os máis utilizados para levar a cabo a eliminación de nutrientes debido ao seu menor custo en comparación cos fisicoquímicos. O proceso típico de eliminación de nitróxeno consiste na oxidación de amonio a nitrato, en dous pasos: 1) oxidación de amonio a nitrito por medio de bacterias oxidantes de amonio (BOA) e 2) oxidación de nitrito a nitrato por medio de bacterias oxidantes de nitrito (BON). Unha vez o amonio oxidado ata nitrato, este é reducido a nitróxeno gas por bacterias heterótrofas en ausencia de osíxeno. Con todo, este proceso non é o máis axeitado para tratar augas con alta carga nitroxenada, é dicir augas con baixa relación DQO/N. Neste caso débese engadir unha fonte externa de carbono que encarece o proceso. Unha alternativa para eliminar nitróxeno destes efluentes, xurdida nos últimos anos, é a eliminación autótrofa de nitróxeno levada a cabo por bacterias Anammox que combinan amonio e nitrito para dar nitróxeno gas e unha pequena cantidade de nitrato en condicións anaerobias. Este proceso resultou interesante para tratar augas de saída de dixestores anaerobios xa que estas conteñen altas concentracións de nitróxeno e pouca materia orgánica biodegradable. Con todo, o lento crecemento das bacterias autótrofas, en comparación coas heterótrofas dificulta a posta en marcha e posterior desenvolvemento do proceso. Dúas configuracións alternativas son posibles para levar a cabo o proceso de eliminación autótrofa de nitróxeno: 1) proceso en dúas etapas: nitrificación parcial do 50% do amonio a nitrito no primeiro reactor que alimentaría un posterior reactor Anammox, 2) ambos procesos levados a cabo nun único reactor, proceso coñecido como CANON (“Complete Autotrophic Nitrogen Removal Over Nitrite”). A eliminación autótrofa de nitróxeno da corrente de saída dos dixestores anaerobios nas plantas de tratamentos de augas residuais presenta unha serie de vantaxes entre as que se atopan:  O aumento da capacidade de eliminación de nitróxeno da planta.  A redución da cantidade de lodos producidos xa que a produtividade da biomasa autótrofa (nitrificante e Anammox) é menor que a heterótrofa (desnitrificante).  A produción de inferiores cantidades de gases de efecto invernadoiro (N2O e NO).  Un aumento do aproveitamento enerxético no dixestor anaerobio xa que ao eliminar o nitróxeno da corrente de rexeitamento pódese aumentar o tempo de residencia hidráulico nos decantadores primarios das EDAR. Isto diminue a necesidade de aireación mantendo a eficacia de eliminación de nitróxeno da planta. Os procesos biolóxicos pódense levar a cabo en sistemas con biomasa en suspensión ou en forma de biopelículas (biomasa adherida ou granular). O uso de sistemas baseados en biopelículas permite o tratamento de maiores cargas en menores volumes de reactor e, xa que logo, está recomendado para o tratamento de efluentes con elevadas cargas. Ademais o feito de que a biomasa creza desta forma permite reter maiores cantidades da mesma, o cal é beneficioso no caso de utilizar microorganismos de lento crecemento como é o caso das bacterias nitrificantes e Anammox. Obxectivos e resumo Nos anos 90 desenvolveuse nos Países Baixos unha tecnoloxía baseada na formación de biomasa granular en condicións aerobias. Esta tecnoloxía fundaméntase no desenvolvemento de biopelículas sen necesidade de material de soporte. Os primeiros traballos desenvolvéronse estudando a autoagregación de bacterias heterótrofas alimentadas cunha fonte de carbono, formando agregados cuxa principal característica era a súa elevada velocidade de sedimentación. Para iso, utilízase un reactor operado de forma descontinua (SBR: Sequencing Batch Reactor) cunha elevada relación altura/diámetro e sométese a biomasa a períodos cíclicos de saciedade e fame. O emprego de sistemas granulares presenta as seguintes vantaxes:  Permiten alcanzar maiores concentracións de biomasa no reactor, diminuíndo a produtividade dos lodos e aumentando a capacidade do reactor.  Minimízase o lavado de biomasa, feito fundamental ao traballar con bacterias de lento crecemento.  A velocidade de decantación da biomasa é moito maior que a dos lodos activos, diminuíndo a superficie necesaria para o decantador (pódese chegar inclusive a non necesitar decantador). En base a todo o anteriormente citado, na presente tese estudáronse as condicións de operación axeitadas para a autoagregación de poboacións de lento crecemento como as nitrificantes (BOA e BON) e as Anammox, cada unha delas por separado (Capítulos 3, 4 e 5). Posteriormente, combináronse poboacións de BOA e Anammox nunha única unidade para levar a cabo o proceso CANON, este proceso foi desenvolvido a temperatura ambiente en dous tipos de reactores (Capítulo 6). A continuación, co fin de analizar a operación dos gránulos a nivel microscópico levouse a cabo un estudo por medio de microelectrodos para o seguimento de perfís de concentración no interior de devanditos gránulos (Capítulo 7). Cos datos experimentais obtidos nun traballo previo neste tipo de sistemas, desenvolveuse un modelo que permitiu simular o comportamento de gránulos aerobios nos que se levaba a cabo a eliminación de nitróxeno a través do proceso convencional de nitrificación-desnitrificación (Capítulo 8). A continuación, detallaranse os contidos principais de cada un dos capítulos da presente tese. No capítulo 1, preséntase unha revisión bibliográfica actualizada dos estudos realizados ata a data sobre a eliminación de nitróxeno nas augas residuais facendo especial fincapé na eliminación autótrofa de amonio (proceso Anammox). Preséntase tamén información relativa aos sistemas de granulación aerobia, a microbioloxía dos procesos implicados e o estado da arte do modelado destes sistemas. No capítulo 2, desenvólvense os materiais e métodos utilizados ao longo dos experimentos realizados nos capítulos posteriores. No capítulo 3 pártese de gránulos nitrificantes operando a temperatura ambiente (18―24 ºC) que foran previamente desenvolvidos con acetato como fonte de carbono e aos que se lles retirou progresivamente dita fonte de carbono da alimentación. A carga de amonio aplicada ao comezo da experimentación era de 0,4 g N L-1 d -1. Ao analizar a evolución das especies nitroxenadas durante o ciclo observáronse pequenas acumulacións de nitrito durante o mesmo que indicaron que as velocidades das dúas etapas da nitrificación (oxidación de amonio e oxidación de nitrito) eran diferentes ao comezo do ciclo, é dicir, cando a concentración de amonio era máxima. A carga máxima que se conseguiu nitrificar a nitrato foi de 0,4 g N L-1 d-1. Ao dobrar a carga nitroxenada aplicada, observouse unha acumulación permanente de nitrito deixando patente unha limitación na velocidade de crecemento das BON malia traballar cunha concentración de osíxeno disolto 8 Obxectivos e resumo próxima á de saturación. Co obxectivo de evitar altas concentracións de amonio que puidesen ser inhibitorias, pasouse a alimentar o reactor durante todo o período aireado do ciclo. Da estimación das velocidades das distintas etapas da reacción, determinouse que a transferencia de osíxeno desde a fase líquida ao interior do gránulo constituía a etapa limitante. Dado que as BOA teñen maior afinidade polo osíxeno que as BON, estas últimas víronse máis afectadas pola limitación de osíxeno, acumulándose nitrito no sistema. Unha vez descuberto o potencial dos gránulos nitrificantes para producir nitrito, fixouse como obxectivo a obtención da nitrificación parcial, o que implica a oxidación do 50% do amonio a nitrito evitando a oxidación a nitrato. A nitrificación parcial conseguiuse diminuíndo a concentración de osíxeno disolto no medio líquido. Nestas condicións, a máxima carga de nitróxeno parcialmente oxidada foi de 1 g N L-1 d-1 cunha concentración de osíxeno disolto no rango 2,0―3,5 mg O2 L-1. Os gránulos mantiveron o seu diámetro medio, a seu integridade e as súas boas características de sedimentación, cun baixo índice volumétrico e unha alta velocidade de sedimentación (100 m h-1) durante a operación do reactor. No capítulo 4 estudouse a granulación de BON co fin de corroborar que a acumulación de nitrito debíase á competición por osíxeno e non a que as bacterias oxidantes de nitrito presentan máis dificultades para crecer formando agregados que as oxidantes de amonio. O concepto granulación aerobia asóciase en xeral a bacterias heterótrofas, con todo, observouse no capítulo 4 do mesmo xeito que no 3 que as bacterias autótrofas tamén poden desenvolverse formando agregados ou gránulos sen necesidade de ser sometidas a períodos de saciedade e fame. Dado que as BON resultan indesexables nun reactor CANON, pois compiten coas bacterias Anammox polo nitrito, hase de poñer especial atención para evitar o seu desenvolvemento. O obxectivo deste capítulo foi o estudo da evolución da estrutura dos gránulos e das poboacións bacterianas ao cambiar a fonte de nitróxeno alimentada de amonio a nitrito. Os gránulos, cuxo tamaño era de 0,8 mm, estaban colonizados por Nitrosomonas (oxidantes de amonio) e Nitrospira (oxidantes de nitrito). As bacterias do xénero Nitrospira caracterízanse por ser estrategas da k, é dicir, bacterias que teñen maior afinidade polo substrato pero menor velocidade máxima de crecemento que as estrategas da r (Nitrobacter no caso das NOB). Unha vez cambiada a fonte de nitróxeno alimentada de amonio a nitrito e tras 250 días de operación non se observou un deterioro da biomasa granular e a poboación de BON maioritaria neste reactor resultou ser a Nitrobacter. No capítulo 5 estúdase o proceso de granulación de bacterias Anammox no tratamento dun efluente sintético salino simulando efluentes industriais, por exemplo as augas residuais xeradas na industria conserveira de peixe. A presenza de concentracións de sal (NaCl) de ata 10 g L-1 non só non producía efectos negativos significativos sobre a velocidade de reacción destes microorganismos senón que fomentaba as interaccións entre agregados favorecendo a granulación. O diámetro medio dos gránulos aumentou un 24% e a concentración de biomasa aumentou un 60% logo de engadir NaCl, nunha primeira etapa engadíronse 5 g NaCl L-1 na alimentación e na segunda etapa engadíronse 10 g NaCl L-1, proceso que tivo unha duración total de 54 días. As actividades máximas específicas rexistradas foron de 0,45 g N (g VSS)-1 d -1. No capítulo 6, estúdase o desenvolvemento do proceso CANON en dous sistemas diferentes: unha columna de borboteo con aireación continua e outra con aireación pulsada. Ambas unidades puxéronse en marcha como reactores nitrificantes nos que se estableceron as condicións de operación axeitadas para levar a cabo a oxidación parcial de amonio nun 50% a unha temperatura de ao redor de 20 ºC. No reactor con aireación continua a biomasa Anammox desenvolveuse no interior dos gránulos nitrificantes nos que se levaba a cabo o proceso de nitrificación parcial. Malia as condicións adversas presentes no medio líquido: alta concentración de osíxeno disolto e de nitrito, a elevada actividade oxidante de amonio nas capas externas dos gránulos elimina por completo o osíxeno no interior dos mesmos creando 9 Obxectivos e resumo 10 unha zona anóxica rica en amonio e nitrito, propicia para o desenvolvemento de biomasa Anammox (proceso CANON). Nesta unidade a máxima carga tratada foi de 1,1 g N L-1 d-1 cando o tamaño dos gránulos era de 3,2 mm. No reactor con aireación pulsada estudouse o arranque dun reactor CANON minimizando o consumo de aire para o aporte do osíxeno requirido. O osíxeno a este reactor aportouse mediante aire pulsado cunha frecuencia de 0.09 s-1. Unha vez desenvolvidos os gránulos nitrificantes cun diámetro de 1,6 mm que levaban a cabo a nitrificación parcial inoculouse o reactor con biomasa pouco enriquecida en bacterias Anammox. Nesta unidade a biomasa Anammox mantívose en forma de agregados mentres que a poboación de BOA permaneceu fundamentalmente en forma de flóculos aínda que unha pequena parte atopábase nas capas máis externas dos gránulos Anammox. O arranque do proceso CANON foi moi rápido e en 35 días a carga de nitróxeno eliminada foi de 0,25 g N L-1 d -1, sendo a máxima carga tratada de 0,45 g N L-1 d -1 cunha concentración de osíxeno de 0,3―0,5 mg O2 L-1. No capítulo 7, levouse a cabo un estudo microscópico dos gránulos do reactor CANON con aireación continua utilizando microelectrodos para medir a concentración de osíxeno disolto e de nitrito (os substratos limitantes no proceso CANON). Esta ferramenta combinouse coa técnica FISH de identificación de poboacións bacterianas. Os resultados indicaron unha clara estratificación de poboacións na profundidade dos gránulos con BOA concentradas nunha capa externa de 400 µm, mentres que as bacterias Anammox estaban situadas entre as 400 e as 1000 µm de profundidade. Os resultados da determinación dos perfís de concentración de osíxeno disolto nos gránulos, en experimentos levados a cabo a diferentes concentracións deste no medio líquido, indicaron que o fluxo de osíxeno cara ao interior dos gránulos aumentaba ao aumentar a concentración de osíxeno disolto no medio. Isto demostra que a actividade oxidante de amonio está limitada pola concentración de osíxeno disolto malia que o reactor operouse en condicións próximas ao 100% de saturación con aire. Do mesmo xeito, os perfís de concentración de nitrito realizados variando a concentración deste no medio líquido corroboraron que unha concentración próxima aos 9 mg N L-1 é necesaria para que a biomasa Anammox non estea limitada por nitrito. Coa información obtida neste traballo deduciuse que existen uns valores óptimos de concentración de amonio e concentración de nitrito no medio líquido que permiten optimizar a estratexia de control e a operación do proceso CANON. No capítulo 8, desenvolveuse un modelo capaz de simular para un reactor granular aerobio tanto as concentracións de nutrientes no medio líquido como as propiedades físicas dos gránulos formados, é dicir, os sólidos en suspensión volátiles e a densidade dos mesmos. Para conseguilo, utilizouse un modelo nunha dimensión empregando o software Aquasim e baseado no ASM3 modificado porque se describiu a nitrificación como un proceso en dúas etapas. Con respecto aos demais modelos xa publicados neste tema, incluíronse modificacións necesarias para simular adecuadamente as concentracións e densidade de biomasa que foron a introdución dun perfil de porosidade variable na profundidade do gránulo e a asignación de valores de densidade distintos para bacterias heterótrofas (350 g DQO (Lgránulo)-1) e autótrofas (150 g DQO (Lgránulo)-1). Definíronse no modelo cinco tipos de sustancias sólidas: bacterias heterótrofas, BOA e BON, polímeros de almacenamento e materia inerte e seis tipos de substratos disoltos no medio liquido: osíxeno, DQO, alcalinidade, amonio, nitrito e nitrato. Cos traballos realizados nesta tese conseguiuse información valiosa para a posible aplicación do proceso autótrofo CANON para eliminar o nitróxeno de efluentes a temperatura ambiente. Objectives and summary This thesis is framed in the field of wastewater treatment and more specifically in the field of nitrogen removal from wastewater. The nutrients removal (mainly N and P) from wastewater is necessary in order to avoid the overgrown of algae and the eutrophication of the receiver waters. Nutrients removal is commonly performed by means of biological processes due to the lower cost in comparison with the physicochemical ones. The typical process of nitrogen removal consists on the oxidation of ammonium to nitrate in two steps: 1) the oxidation of ammonium to nitrite by means of ammonia oxidizing bacteria (AOB) and 2) the oxidation of nitrite to nitrate by means of nitrite oxidizing bacteria (NOB). Once the ammonium has been oxidized to nitrite and nitrate, it is reduced to nitrogen gas by the heterotrophic bacteria in absence of oxygen. However, this process is not the most appropriated to treat high nitrogen loaded wastewaters, i.e., wastewaters characterized by low COD/N ratios. In this case, an external carbon source has to be added which increases the cost of the process. An alternative to remove nitrogen from these effluents, which appeared in the last years, is the autotrophic removal of nitrogen carried out by Anammox bacteria which combine ammonium and nitrite to produce nitrogen gas and a small amount of nitrate in anaerobic conditions. This process revealed to be interesting for the treatment of reject water from anaerobic digesters with high nitrogen concentrations and low biodegradable organic matter content. Nevertheless, the slow growth of the AOB in comparison with the heterotrophic ones makes the start-up and the later development of the process difficult. Two alternative configurations are possible to carry out the autotrophic nitrogen removal process: 1) process in two stages: partial nitrification of 50% ammonium to nitrite in the first reactor which fed the second Anammox reactor 2) both processes carried out in a single reactor, process known as CANON (“Complete Autotrophic Nitrogen Removal Over Nitrite”). The autotrophic nitrogen removal from the reject water of anaerobic digesters in wastewater treatment plants presents several advantages such as:  The increase of nitrogen removal capacity in the plant.  The reduction of the amount of sludge produced since the productivity of the autotrophic biomass (nitrifying and Anammox) is smaller than the heterotrophic one (denitrifying).  The production of smaller amounts of greenhouse gases (N2O and NO).  The increase of the energy production in the anaerobic digester since, once the nitrogen is removed from the reject water the hydraulic retention time in the primary settlers of the WWTP can be increased. This reduces the aeration energy at similar overall nitrogen removal. The biological processes can be carried out in systems with suspended biomass or in the form of biofilms (adhered to a carrier material or as granular biomass). The use of systems based on biofilms allows the treatment of larger loads in smaller reactor volumes and therefore, it is recommended for the treatment of effluents with high loads. Furthermore the fact that the sludge is grown as biofilms allows the retention of larger amounts of biomass which is beneficial in the case of slow growing microorganisms such as nitrifying and Anammox bacteria. During the 90s in The Netherlands, a technology based in aerobic granular biomass was developed. This technology is based on the development of biofilms without the need for support material. The first works Objectives and summary were performed to study the autoaggregation of heterotrophic bacteria fed with a carbon source and the formation of aggregates whose main characteristic is their large settling velocity. To achieve the biomass granulation, a discontinuously operated reactor with a high height/diameter ratio is used (SBR: Sequencing Batch Reactor) and the biomass is exposed to cyclical periods of feast and famine. The use of granular systems presents the following advantages:  It allows the obtaining of larger biomass concentrations diminishing the sludge yield and increasing the removal capacity of the reactor.  The biomass washout is minimized, this aspect is important when low growing bacteria are used.  The biomass settling velocity is much higher than that of activated sludge, with a consequent decrease of the surface needed for the settler (the settler can even be avoided). On the basis of all the aforementioned, in the present thesis the appropriated operational conditions to obtain the autoaggregation of slow growing organisms like nitrifiers (AOB and NOB) and Anammox in a separated way was studied (Chapters 3, 4 and 5). Later, the AOB and Anammox populations were combined in a single unit to carry out the CANON process in two different kinds of reactors (Chapter 6). Then with the aim of going deeper in the operation of those granules a microscopic study was performed by means of microelectrodes to follow the concentration profiles inside the aggregates (Chapter 7). With the experimental results obtained in a previous work, a mathematical model has been developed to simulate the dissolved and particulate compounds of aerobic granules where the nitrogen removal was carried out by nitrification-denitrification (Chapter 8). The main content of each chapter of the present thesis will be detailed in the following sections. In Chapter 1 an actualized literature review about the performed studies up to date in the field of nitrogen removal from wastewater is presented. Special attention is paid to autotrophic nitrogen removal (Anammox process). Information regarding aerobic granulation systems, the microbiology of the involved processes and the state of the art of the modelling of this kind of systems is presented. In Chapter 2, the material and methods used during the different experiments performed along the following chapters are described. In Chapter 3, nitrifying granules were operated at room temperature (18―24 ºC). These granules were developed in a previous work as heterotrophic granules using acetate as carbon source and then, the organic carbon content in the feeding media was progressively removed until the obtaining of nitrifying granules. With the increase of the nitrogen load applied to the reactor, nitrite accumulated during the operational cycle. This indicated that the rates corresponding to the two nitrification steps (ammonia and nitrite oxidation) were uncoupled at the beginning of the cycle, i.e., when the ammonia concentration had the maximum value. The maximal nitrogen load oxidized to nitrate was of 0.4 g N L-1 d-1. When the nitrogen load applied was doubled, a permanent accumulation of nitrite was observed indicating the limitation of the growth rate of NOB in spite of working at dissolved oxygen concentration close to air saturation. With the aim of avoiding high ammonium concentrations in the bulk liquid that could be inhibitory, the reactor was fed during the whole aerated period. Estimations of the rates of the different reaction stages revealed that the oxygen mass transfer from the bulk liquid to the granules was limiting. Since AOB have higher affinity for oxygen than NOB, the last ones were 12 Objectives and summary more affected by oxygen limitation which caused nitrite accumulation in the system. Once the potential of the nitrifying granules to produce nitrite was established, the obtaining of partial nitrification was the objective. Partial nitrification consists of the oxidation of 50% of the ammonium to nitrite avoiding its further oxidation to nitrate. This was achieved by means of decreasing the dissolved oxygen concentration in the bulk liquid. Under these conditions, the maximal nitrogen load partially oxidized was of 0.8 g N L-1 d-1 with a dissolved oxygen concentration ranging from 2.0 to 3.5 mg O2 L-1. As a result, the granules kept their mean diameter (around 3 mm), physical integrity and good settling properties, with a low sludge volumetric index and high settling velocity (100 m h-1). In Chapter 4, the granulation of NOB is studied to corroborate that the nitrite accumulation was caused by a competition for oxygen and that NOB do not present more difficulties to grow in the form of aggregates than AOB. The concept “aerobic granulation” is generally associated to heterotrophic bacteria; however in Chapter 4 as well as in Chapter 3 autotrophic bacteria grew forming aggregates or granules without being exposed to feast and famine periods. Taking into account that NOB are not desired in CANON reactors since they compete for nitrite with Anammox bacteria, special effort has to be made to avoid their growth. The aim of this chapter is the study of the evolution of the granule structure and the bacterial populations after the change of the nitrogen source fed from ammonium to nitrite. Granules with a mean diameter of 0.8 mm were initially colonized by Nitrosomonas (AOB) and Nitrospira (NOB). Bacteria from the genus Nitrospira are “k-strategists”, i.e., bacteria with higher substrate affinity but slower maximum growth rate than “r-strategists” bacteria (Nitrobacter in the case of NOB). Once the nitrogen source was changed from ammonium to nitrite and after 250 days of operation no damage of the granular biomass was observed and the main population of NOB resulted to be Nitrobacter. In Chapter 5, the granulation process of Anammox bacteria is studied for the treatment of a saline synthetic medium simulating industrial effluent such as wastewater produced in the fish canning industry. The presence of salt concentration (NaCl) up to 10 g L-1 did not produce a negative effect on the reaction rate of these microorganisms but it promoted the interactions between aggregates enhancing the granulation. The mean diameter of the granules and the biomass concentration in the reactor increased 24% and 60% respectively after the addition in the feeding of 5 g NaCl L-1 and 10 g NaCl L-1 in two steps along a period of 54 days. The maximal specific activities registered were of 0.45 g N (g VSS)-1 d-1. In Chapter 6 the development of the CANON process in two different bubble columns is studied: one with continuous aeration and another one with pulsing aeration. Both units were started up as nitrifying reactors where the suitable operational conditions to perform the partial oxidation of 50% of the ammonium to nitrite at around 20 ºC were established. In the reactor with continuous aeration, the Anammox biomass was developed inside the nitrifying granules where the partial nitrification was also taking place. Despite the adverse conditions present in the liquid media: high dissolved oxygen and nitrite concentrations; the high ammonium oxidation activity in the external layers of the granules fully removed the oxygen in their inner core creating an anoxic zone rich in ammonium and nitrite and suitable for the development of Anammox biomass (CANON process). In this unit the maximal nitrogen removal rate was of 1.1 g N L-1 d-1 with granules whose mean diameter was of 3.2 mm. The start-up of a CANON reactor was also studied in a bubble column with pulsing aeration in order to minimize the air consumption in relation to the required oxygen supply. The oxygen was supplied to this reactor by means of air pulsations with a frequency of 0.09 s-1. Once the nitrifying granules were developed with a mean diameter of 1.6 mm and the partial nitrification was established, the reactor was inoculated with biomass 13 Objectives and summary 14 poorly enriched in Anammox bacteria. In this unit, the Anammox biomass was kept in the form of aggregates, while the population of AOB was present mainly in the form of flocs and also in external layers of the Anammox granules. The start-up of the CANON process was fast and in 35 days the nitrogen removal rate was of 0.25 g N L-1 d-1. The maximum nitrogen removal rate was of 0.45 g N L-1 d-1 at an oxygen concentration of 0.3―0.5 mg O2 L-1. In Chapter 7, a microscopic study of the granules inside the CANON reactor with continuous aeration was performed using microelectrodes to measure dissolved oxygen and nitrite concentrations profiles (the two limiting substrates of the CANON process). This tool was combined with the FISH technique which was used to identify the bacterial populations. Obtained results indicated a clear stratification of bacterial populations inside the granules with the AOB concentrated in an external layer of 400 µm, while the Anammox bacteria were placed between 400 and 1000 µm of depth. Experiments performed at different dissolved oxygen concentrations in the bulk liquid revealed that the oxygen flux to the granules increased with the increase of dissolved oxygen concentration in the medium. This result demonstrates that the AOB activity is limited by the dissolved oxygen concentration despite the reactor was operated at conditions close to 100% air saturation. In the same way, the nitrite concentration profiles performed changing the concentration of nitrite in the bulk liquid corroborate that concentrations around 9 mg N L-1 are needed to avoid the limitation of Anammox bacteria by nitrite. With the information obtained from the present work it is inferred that there exist optimal values of ammonia and nitrite concentrations which allow the optimization of the control strategy and operation of the CANON process. In Chapter 8, a model able to simulate not only the nutrients concentrations in the bulk liquid but also the physical properties of the formed granules, in terms of volatile suspended solids and density, was developed. The software Aquasim was used to implement a one dimension model developed on the basis of the ASM3 model but modified to describe the nitrification as a two stages process. In comparison to other published models in this field, several modifications were implemented to accurately simulate the concentration and density of the biomass. These modifications consisted of the introduction of a porosity profile in depth inside the granule and the assignation of density values different for heterotrophic (350 g DQO (Lgranule)-1) and autotrophic bacteria (150 g DQO (Lgranule)-1). In the model five types of solid substances: heterotrophic bacteria, AOB and NOB, storage polymers and inert matter; and six types of substrates in the bulk media: oxygen, chemical oxygen demand, alkalinity, ammonium, nitrite and nitrate were defined With the work performed in this thesis, important information for the possible application of the autotrophic CANON process for nitrogen removal from wastewater at room temperature was obtained. Chapter 1 Introduction Summary In this chapter, the scope and the motivation of this thesis are detailed and the existent options applied to wastewater treatment are analyzed. This work has been developed in the field of water pollution, specially focused on nitrogen contamination and removal. It is known that the nitrogen cycle is highly affected by anthropogenic activities due to its massive fixation from the atmosphere to produce fertilizers. The release of ammonium in natural systems provokes eutrophication, i.e., proliferation of algae and oxygen depletion in natural waters; this is the reason why nitrogen has to be removed in wastewater treatment plants (WWTP). The traditional way to remove nitrogen from wastewaters is the conventional nitrification-denitrification process. More and more stringent effluent quality requirements and the increasing complexity of wastewaters seriously challenged the efficiency of the classical wastewater treatments. Recently new technologies have arisen to improve the nitrogen removal from wastewater: bioaugmentation, partial nitrification/denitrification and partial nitrification/Anammox. These processes can be applied to treat reject water (the effluent of the anaerobic digester of the WWTP). The nitrogen removal of the reject water is highly beneficial for the overall nitrogen removal efficiency of the WWTP instead of its direct recirculation to the head of the plant. These new processes will be suitable to implement in new WWTP but they can also be used to improve already built WWTP. The treatment of reject water by the autotrophic nitrogen removal (which combines partial nitrification and Anammox processes) is highly recommended due to the characteristics of these wastewaters: low concentration of biodegradable carbon source and high concentration of ammonium. The main drawback of the microorganisms involved in the process is their slow growth rate in comparison with heterotrophic bacteria. The application of the granular concept will be analyzed to achieve high nitrogen removal rates due to the high sludge retention times achievable in such systems. Chapter 1 1. Risks linked to the presence of organic matter: the treatment of raw water with high levels of organic matter is always technically difficult. It can lead to the production of carcinogenic by-products (trihalomethanes, other chlorinated components or ozonides) as a result of their reaction with disinfectants. 2. Risks linked to the presence of specific cyanobacteria in fresh waters: when eutrophication leads to the development of cyanobacteria that are potentially toxic, the elimination of these toxins is complex. 1.4. Nitrogen removal from wastewater The nitrogen compound mainly present in wastewaters is ammonium (NH4+) which can be removed by physicochemical or biological processes. The selection of the best alternative is generally based on costeffectiveness issues. However, in practice the selection of either a biological or a physiochemical method is determined by the nitrogen concentration of the wastewater. Three concentration ranges can be distinguished (Mulder, 2003): a) Diluted wastewater with ammonium concentration up to 100 mg N L-1 (e.g. domestic wastewater). In this range biological processes such as activated sludge are preferred processes based on cost-effectiveness issues. b) Concentrated wastewater with ammonium concentrations in the range of 100-5000 mg N L-1. A typical example is sludge reject water for which, after extensive investigations, biological treatment is to be preferred. The main reason to select a biological process is the lower price compared to the physicochemical methods. Van Dongen et al., (2001) estimated a cost of 4.5–11.3 euros per kg N removed with physicochemical techniques and 2.3–4.5 euros per kg N removed using conventional biological processes (nitrification-denitrification) for the case of The Netherlands. Recently new biological nitrogen removal processes for these concentrated streams have been under study as it is the case of the autotrophic nitrogen removal which can be carried out in two steps (e.g., the SHARON-Anammox process, van Dongen et al., 2001), or one single step (e.g., the CANON process, Third et al., 2001). c) Concentrated wastewater with ammonium concentrations higher than 5000 mg N L-1. In this range physicochemical methods are technically and economically feasible. The main physicochemical processes applied for ammonium removal are air stripping; breakpoint chlorination and selective ion exchange. The last two options have the inconvenient of involving technologies with high operational costs and complex control. Ammonia stripping is widely used because of its simple operation and high efficiency. When biological processes are considered, besides the ammonium concentration in wastewater, its COD/N ratio will determine the most suitable biological process to carry out nitrogen removal: 1) COD/N > 20: in this case, the assimilation of nitrogen by heterotrophic bacteria is sufficient to remove nitrogen. 2) 20 < COD/N < 5: removal of nitrogen by assimilation and nitrification-denitrification pathway. 3) COD/N < 5: in this case, the nitrification-denitrification process is not suitable since an additional carbon source is needed. Nitrogen removal by “nitrite-route” processes such as partial nitrificationdenitrification or partial nitrification-Anammox are being implemented to optimize WWTP operation. The removal of nitrogen from these “problematic” wastewaters will be the aim of the thesis. 22 Introduction 1.5. Nitrification-denitrification processes Biological processes in wastewater treatment are mainly carried out by bacteria. Conventionally, nitrogen removal from WWTP is carried out by the nitrification-denitrification process. It is based on a sequence of aerobic and anoxic conditions where the ammonium is firstly oxidized to nitrate in the presence of oxygen by nitrification and then reduced using an organic carbon source as electron donor during the denitrification process. 1.5.1. Nitrification Nitrification, which converts ammonia first to nitrite and then to nitrate, is the initial step of biological nitrogen removal processes carried out by two phylogenetically independent groups of autotrophic aerobic bacteria, namely, ammonium oxidizing bacteria (AOB) and nitrite oxidizing bacteria (NOB). The nitrification process is carried out in two sequential stages: the ammonium oxidation to nitrite and the subsequent oxidation of nitrite to nitrate. The oxidation of ammonium to nitrite by AOB produces hydroxylamine as intermediate. The enzymes involved are ammonia monooxygenase for ammonium oxidation (Eq. 1.2) and hydroxylamine oxidoreductase to produce nitrite (Eq. 1.3) with the complete process as described in Eq. 1.4. NH4+ + O2 + H+ + 2 e-  NH2OH + H2O (1.2) NH2OH + H2O  NO2- + 4 e- + 5 H+ (1.3) NH4+ + 1.5 O2  NO2- + H2O + 2 H+ (1.4) The carbonate system is usually the pH buffer available in the wastewater which neutralizes the production of protons through CO2 stripping (Eq. 1.5). NH4+ + 1.5 O2 + 2 HCO3-  NO2- + 3 H2O + 2 CO2 (1.5) The whole metabolism of the bacteria including their growth (anabolism and catabolism) is described by means of the following stoichiometric equation where the fixation of inorganic carbon and its equilibrium are represented in Eq. 1.6. NH4+ + 1.382 O2 + 1.982 HCO3- 0.018 C5H7O2N + 0.982 NO2- + 2.927 H2O + 1.891 CO2 (1.6) As represented in the Eq. 1.6, 2 moles of alkalinity are removed due to CO2 stripping per mole of ammonium oxidized due to the release of protons during the ammonium oxidation (Eq. 1.5). In the cases when this amount of buffer is not available in the water, the pH of the medium drops and the ammonium oxidation rate decreases sharply. In the second stage of the nitrification process, nitrite is oxidized to nitrate by NOB by means of the enzyme nitrate oxidoreductase (Eq. 1.7). NO2- + 0.5 O2  NO3 - (1.7) When including bacterial growth in the stoichiometry the Eq. 1.8 is obtained. NO2- + 0.003 NH4+ + 0.488 O2 + 0.010 H2CO3 + 0.003 HCO3- 0.003 C5H7O2N + NO3- + 0.008 H2O (1.8) Taking into account the two steps of the nitrification process, several aspects have to be pointed out (by comparison of Eq. 1.6 and 1.8): 1) ammonium oxidation to nitrite consumes more oxygen than nitrite oxidation; 23 Chapter 1 for the first step of ammonium oxidation 3.16 g O2 are required to oxidize 1 g NH4-N to nitrite while for the second step 1.11 g O2 are needed to oxidize 1 g NO2-N to nitrate. 2) nitrite oxidation produces more bacteria per mole of nitrogenous compound oxidized. 3) ammonium oxidation to nitrite produces protons, to oxidize 1 mole (14 mg N) ammonium to nitrate 2 mol of alkalinity are consumed. 1.5.1.1. Parameters affecting nitrification The main parameters affecting the nitrification process are: the concentrations of dissolved oxygen (DO) in the bulk liquid, the concentrations of substrates: ammonium and nitrite, the temperature and the pH. The growth rate (μ) of the microorganisms can be described by a Monod equation that indicates the dependency of the bacteria growing on a limiting substrate S, including a bacteria decay coefficient (Eq. 1.9). b CK C μμ SS S max    (1.9) where: μ = growth rate (d-1) μmax = maximum microorganisms growth rate (d-1) KS = half saturation constant (g L-1) CS = limiting substrate concentration (g L-1) b = decay constant (g g-1 d-1) The activity of the nitrifying bacteria begins at temperatures around 4 ºC, its maximum growth rate increases with the temperature up to the maximum value at temperatures around 35-40 ºC. At temperatures higher than 40 ºC the activity falls sharply. The concentrations of ammonium and nitrite can either limit or inhibit the nitrification process and for this reason they have to be controlled. The half saturation constants of ammonium and nitrite for AOB and NOB are 1 mg NH4+-N L-1 and 1 mg NO2--N L-1, respectively (Henze et al., 2000). Therefore ammonium and nitrite concentrations higher than 1 mg N L-1 would be necessary in order to avoid substrate limitation of the nitrification process. However, too high concentrations of ammonium and nitrite and specially their unionized forms: free ammonia and free nitrous acid inhibit the process and therefore they must be controlled, further information about AOB and NOB inhibitions will be given in the next section. The significance of pH decrease during the nitrification process relies on the fact that the reaction rates are rapidly depressed as the pH is reduced below 7.0 (Fig. 1.5). Therefore, in cases where the alkalinity of the wastewater will be depleted by the acid produced by nitrification, the proper amount of alkalinity must be supplemented by a chemical addition (EPA, 1993) to avoid a drop in the nitrifying activity. 0.0 0.2 0.4 0.6 0.8 1.0 567891 pH μ max (d -1 ) 0 Figure 1.5. a) Representation of b) Representation of the µmax of AOB and NOB vs pH at 20 ºC (obtained from data published Jubany et al., 2008). 24 Introduction The pH buffer normally available in wastewaters is the inorganic carbon. The equilibrium of the carbonate system in water plays a key role in both processes: nitrification and denitrification and furthermore it depends strongly on the pH of operation. Carbonic acid (H2CO3) is a diprotic acid which corresponds to dissolved carbon dioxide; this acid can yield more than one proton in different equilibriums (Eq. 1.10). CO2 (g) + H2O ⇌ H2CO3 (aq) ⇌ H+ + HCO3- (aq) ⇌ 2 H+ + CO32- (aq) (1.10) The ratio of the different acid/base species at different pH is represented (Fig. 1.6). Since bacterial activity occurs around neutral pH, the main carbonate compound in wastewater will be the bicarbonate ion. At pH 7, around 20% of the inorganic carbon in solution corresponds to H2CO3 which is in equilibrium with gaseous CO2. The hydration equilibrium constant of CO2 and the Henry constant both at 25°C are 1.70·10−3 and 29.76 atm l mol-1, respectively: hence, the majority of the carbon dioxide is not converted into carbonic acid and stays as CO2 molecules. 0.0 0.2 0.4 0.6 0.8 1.0 4681012 pH α i a) 0.0 0.2 0.4 0.6 0.8 1.0 0 10203040 Volume (mL) α i 3 5 7 9 11 13 pH pH 7 b) Figure 1.6. a) Molar fractions of H2CO3 (- - -), HCO3- (▬) and CO32- (▬) in solution at different pH values (T = 25 ºC). b) Distribution of acid/base species: H2CO3 (- - -), HCO3- (▬) and CO32- (▬) and pH (○) along a titration curve of 20 ml of CO320.1 M with HCl 0.1 M (performed with the computational freeware program CURTIPOT available in Internet). 1.5.1.2. Microbiology of AOB and NOB Although the basic metabolism is more or less uniform for all ammonia-oxidizing bacteria, different physiological requirements exist among the different strains (Wagner et al., 1995; Koops and PommereningRoser, 2001). For example, the substrate affinity, salt requirement and salt tolerance differ significantly among ammonia oxidizers. AOB are classified in two phylogenetic groups. One group is related to the γ subclass of the Proteobacteria. Its only genus, Nitrosococcus, is represented by two described marine species (Koops and Pommerening-Roser, 2001). The second group belongs to the β subclass of the Proteobacteria. Two clusters exist within this assemblage, the Nitrosospira cluster (with three genera) and the Nitrosomonas cluster. All members of the three genera of the Nitrosospira cluster are very closely related to each other, whereas the Nitrosomonas cluster reveals at least five distinct lineages of descent (Koops and Pommerening-Roser, 2001). Community studies on wastewater treatment systems have indicated that Nitrosomonas strains form one of the most important populations. Three lineages (Nitrosomonas europaea, Nitrosomonas oligotropha, and Nitrosomonas communis lineages) are routinely observed as dominant organisms in wastewater treatment plants (Gieseke et al., 2001). 25 Chapter 1 Nitrite oxidizing bacteria also differ in their physiological requirements and capabilities; they have been classified into four groups. The major group, which belongs to the α subclass of Proteobacteria, is represented by a single genus, Nitrobacter, with four described species (Sorokin et al., 1998). The two species of the genus Nitrospira, Nitrospira marina and Nitrospira moscoviensis, are members of a distinct phylum close to the δ subclass of Proteobacteria. In contrast to textbook knowledge, Nitrospira-like bacteria and not Nitrobacter spp., are the dominant nitrite oxidizers in most full-scale wastewater treatment plants, in laboratory scale reactors and also in different environmental samples (Daims et al., 2000). These findings seem to suggest that Nitrospira are widely distributed in nature and probably contribute significantly to global nitrite oxidation. 1.5.2. Partial nitrification In the last years, new strategies arose focused on the development of a shortcut in the nitrogen cycle named the “nitrite route”, which avoids the nitrite oxidation to nitrate. In this case, the further denitrification of the produced nitrite to nitrogen gas could be performed under autotrophic or heterotrophic conditions. To obtain partial nitrification, the ammonium has to be converted into nitrite by the AOB while the oxidation of nitrite to nitrate carried out by NOB has to be avoided. The AOB and NOB are two phylogenetically unrelated groups whose different growth rates and the way their growth rates are affected by parameters like temperature, pH, dissolved oxygen (DO), etc. can be used to outcompete NOB and to uncouple both reaction rates. The oxidation of high ammonium concentrations causes a significant pH-decrease, which limits further ammonium conversion as aforementioned. Sludge reject water typically contains equimolar amounts of bicarbonate and ammonium, so half of the produced protons are neutralized by CO2-stripping. As a result, for streams containing bicarbonate and ammonium in equimolar amounts and without additional pH control in the reactor, typically half of the ammonium is converted before a significant pH-decrease occurs, that prevents further ammonium oxidation. Since the different growth parameters for AOB and NOB are differently affected by the operational conditions, the conditions where AOB grow faster than NOB can be selected (μA>μN). According to this, several strategic parameters have been controlled in WWTP to reach partial nitrification: 1) Temperature and SRT: At the usual temperatures of operation of WWTP no nitrite build-up is registered. This is due to higher growth rate of NOB in the range of temperatures from 10 to 20 ºC. However the oxidation of ammonia has higher activation energy than the oxidation of nitrite. Therefore, the operation at temperatures above 25 ºC would allow the washout of NOB which grow slower than AOB (Hellinga et al., 1998) (Fig. 1.7). 2) Dissolved oxygen: AOB have lower oxygen affinity than NOB (Table 1.3). By controlling the DO concentration to low values, the oxidation of nitrite to nitrate can be controlled. Nitrite accumulation was, for instance, obtained manipulating the dissolved oxygen concentration in biofilm systems (Garrido et al., 1997; Bernet et al., 2005). Cecen and Gonenc, (1995) found that the bulk oxygen to bulk ammonia ratio is the most crucial parameter in the accumulation of nitrite. In nitrification, these researchers found a considerable degree of nitrite accumulation at bulk O2/bulk NH4+ ratios lower than 5 mg O2 (mg N)-1. Bernet et al., (2005) found that using O2/NH4+ ratio set points of 0.05 and 0.1 mg O2 (mg N)-1 it was possible to oxidize up to 80% of the inflow NH4+ into NO2-. 26 Introduction 0 1 2 3 4 10 15 20 25 30 35 Temperature (ºC) μ A,max , μ N,max (d -1 ) Figure 1.7. Comparison of the evolution of the µmax of AOB (- - -) and NOB (▬) with the temperature for a pH value of 7.5 (obtained from data published by Jubany et al., 2008). Table 1.3. Comparison of oxygen affinity constants of AOB (KO2A) and NOB (KO2N) at different temperatures. KOA (mg L-1) KON (mg L-1) T (ºC) Reference 0.60 1.30 20 Wiesmann, (1994) 0.30 1.10 25 Wiesmann, (1994) 0.72 1.75 25 Guisasola et al., (2005) 0.50 1.00 30 Pambrun et al., (2006) 0.24 1.50 30 Wyffels et al., (2004) 0.25 0.50 35 Dold et al., (2007) 3) Free ammonia (FA) and free nitrous acid (FNA) concentrations: Despite the mechanisms of inhibition of FA and FNA remain unclear (Vadivelu et al., 2007), Nitrobacter have been described to be more sensitive to FA and FNA inhibitions than Nitrosomonas. Anthonisen et al., (1976) proposed the expressions from Eq. 1.11 and 1.12 to calculate the concentrations of NH3 and HNO2 at the operational temperature (T) from the NH4+, NO2concentrations and the pH value in the bulk liquid.           1 10 e C C pH NH NH 273 6344 4 3T (1.11)   273 2300 2 2e10 C CpH NO HNO    T (1.12) Anthonisen et al., (1976) reported that FA initiated inhibition on Nitrobacter at about 0.1–1.0 mg NH3-N L-1, while the threshold value for Nitrosomonas was about 10–150 mg NH3-N L-1. They also reported that FNA at a concentration of as low as 0.22 mg HNO2-N L-1 inhibited the activities of Nitrobacter. More recently, Vadivelu et al., (2006a, 2006b, 2007) uncoupled the effect of both inhibitors in the anabolism and catabolism of Nitrosomonas and Nitrobacter performing tests in the presence or absence of inorganic carbon. Again, Nitrobacter growth (anabolism) was inhibited at much lower values of FA and FNA than Nitrosomonas. The biosynthesis was completely stopped at a FNA concentration of 0.40 mg HNO2-N L-1 27 Chapter 1 for Nitrosomonas and 0.023 mg HNO2-N L-1 for Nitrobacter. FA concentrations up to 16.0 mg NH3-N L-1 (the highest concentration tested) did not have any inhibitory effect on either the catabolic or anabolic processes of the Nitrosomonas culture while Nitrobacter likely ceased to grow at a FA level of above 6 mg NH3-N L-1. The adaptation of bacteria to inhibitions after long term operation is the main drawback of a strategy based exclusively on this factor (Turk and Mavinic, 1989). 1.5.3. Denitrification In the denitrification process, the nitrate and/or nitrite present in the wastewaters is reduced to molecular nitrogen in anoxic conditions by the action of heterotrophic bacteria. The process requires the presence of a source of organic carbon as electron donor, e.g. acetic acid or methanol, and nitrate acts as the last electron acceptor in the respiratory chain substituting the O2 molecule. The reduction is carried out by subsequent steps through different oxidation states of nitrogen (Eq 1.13 to 1.16) and the global stoichiometry with acetic acid as organic carbon source is represented in Eq. 1.17. 2 CH3COOH + 8 NO3-  4 CO2 + 8 NO2- + 4 H2O (1.13) CH3COOH + 8 NO2- + 2 H2O  2 CO2 + 8 NO + 8 OH- (1.14) CH3COOH + 8 NO  2 CO2 + 4 N2O + 2 H2O (1.15) CH3COOH + 4 N2O  2 CO2 + 4 N2 + 2 H2O (1.16) 5 CH3COOH + 8 NO3-  10 CO2 + 4 N2 + 6 H2O + 8 OH- (1.17) Rewriting equation 1.17 taking into account the equilibrium of CO2 gives Eq. 1.18. NO3- + 0.625 CH3COOH  HCO3- + 0.25 CO2 + 0.5 N2 + 0.75 H2O (1.18) From the stoichiometry it can be inferred that the denitrification process originates an increase in the medium alkalinity and that 40% of the organic matter needed is used to reduce nitrate to nitrite. 1.5.3.1. Microbiology of Denitrifiying bacteria This process makes use of N oxides (e.g. nitrate or nitrite) as terminal electron acceptors under anaerobic, microaerophilic, and occasionally aerobic conditions. Important advances in the biochemical characterization of denitrification and the underlying genetics have been achieved with various organisms (like species from the genera Pseudomonas, Paracoccus, Ralstonia, and Rhodobacter). Around 50 genes are required within a single bacterium for the synthesis of the biochemical denitrification apparatus (Zumft, 1997). 1.5.4. Anaerobic ammonium oxidation (Anammox) Another possibility to remove ammonium from wastewaters with low COD/N ratios consists of the autotrophic nitrogen removal combining AOB and Anammox bacteria. For a long time, it was thought that ammonium oxidation could only take place aerobically. Broda (1977) predicted, using thermodynamic calculations, the existence of chemolitoautotrophic bacteria capable to oxidize ammonium using nitrite as electron acceptor. That prediction would be experimentally confirmed two decades later by Mulder et al., (1995) in a denitrifying pilot plant, treating wastewaters from a yeast plant. Anammox bacteria convert ammonium together with nitrite (electron acceptor) directly to dinitrogen gas in the absence of any organic carbon source, following the reaction described in Eq. 1.19 (Strous et al., 1998). In this process a small amount of nitrate is also produced in the anabolism of Anammox bacteria. NH4+ + 1.32 NO2- + 0.066 HCO3-+ 0.13 H+→ 1.02 N2+ 0.26 NO3-+ 0.066 CH2O0.5N0.15 + 2 H2O (1.19) 28 Introduction Anammox bacteria have been detected in several wastewater treatment plants all around the world; they belong to the genus Plantomycetes and their optimal temperature and pH of operation are 35 ºC and 8, respectively. These bacteria are characterised by a low productivity: 0.038 g VSS (g N)-1 and a slow growth rate with large doubling times as long as 11 d (Strous et al., 2002). The advantage of this low biomass productivity is the reduction of operation costs related to sludge handling in the WWTP. On the other hand, the slow growth rate of Anammox bacteria makes the start up of the Anammox processes long and difficult. It is therefore mandatory to start-up Anammox processes in reactors with good biomass retention. Another important characteristic of Anammox bacteria is the fact that they are inhibited by both O2 and NO2-. Anammox bacteria are inhibited by an oxygen partial pressure of only 0.5% of air saturation, however, this inhibition is reversible once anaerobic conditions are re-established (Strous et al., 1997). Related to NO2inhibition, no uniformity is found in the literature about the threshold values. For instance, Strous et al., (1999a) reported a value of 100 mg N L-1 as completely inhibitory whereas Dapena-Mora et al., (2007) reported that concentrations of nitrite of 350 mg N L-1 corresponded only to 50% inhibition of Anammox bacteria. Further studies are needed in order to clarify this point. What is known is that, it is mandatory to maintain nitrite to low values in order to avoid the total breakdown of the process since once nitrite begins to accumulate there is high risk of failure in the reactor (Dapena-Mora et al., 2004a). 1.5.4.1. Microbiology of Anammox bacteria Since the discovery of Anammox process in Delft (The Netherlands), evidence for Anammox activity has been obtained, in a variety of laboratory and engineered systems (Schmid et al., 2005) besides the natural environments aforementioned. Until now, several Anammox organisms were identified: Candidatus “Brocadia Anammoxidans”, Candidatus “Kuenenia stuttgartiensis”, Candidatus “Scalindua sorokinii”, Candidatus “Scalindua brodae”, Candidatus “Scalindua wagneri”, Candidatus “Brocadia fulgida” and Candidatus “Anammoxoglobus propionicus” and their 16S rRNA sequences were determined (Schmid et al., 2000; Fujii et al., 2002; Schmid et al., 2003; Kartal et al., 2007; Kartal et al., 2008). The studies about enrichment or operation of the Anammox process at laboratory or pilot scale reactors up to date were performed mainly with bacteria of the genus C. “Brocadia” or C. “Kuenenia”, so the major part of the available information on Anammox is about these types of bacteria. The main difference between Candidatus “K. stuttgartiensis” and Candidatus “B. anammoxidans” is their Anammox activity: 26.5 nmol N2 (mg protein)-1 min-1 at pH 8 and 37 ºC for C. “Kuenenia” (Egli et al., 2001) and 55 nmol N2 (mg protein)-1 min-1 at pH 8 and 40 ºC for “Brocadia” . However, Candidatus “K. Stuttgartiensis” has higher tolerance to nitrite, it is more active in low cell density cultures and it is less inhibited by phosphate compared to Candidatus “B. anammoxidans” (Egli et al., 2001). The Anammox organisms resemble each other in the phylogenetic analyses of their 16S rRNA sequences, which show they form a monophyletic branch, which consists of five distinct genera with about 90% sequence similarity to each other, within the phylum Planctomycetes (Fig. 1.8). As the rest of the species within the order Planctomycetales, they lack of peptidoglycan, an almost universal polymer found within the domain Bacteria. Instead, protein is the major constituent of their cell walls. Among the domain Bacteria, this lack of peptidoglycan is a characteristic shared only with the Chlamydiae and the cell-wall-free Mycoplasms (Lindsay et al., 2001). As all known Planctomycetes, Anammox bacteria have an ultrastructure atypical for bacteria, with a compartmentalized cytoplasm (Fuerst, 2005). Electron microscopy analysis revealed that Anammox bacteria 29 Chapter 1 have a separate specific membrane bound compartment: the anammoxosome. Hydroxylamine oxidoreductase (HAO) enzymes are present exclusively inside this anammoxosome, an organelle-like body that made up more than 30% of the cell volume (van Niftrik et al., 2004). This dedicated intracytoplasmic compartment has been found to be surrounded by a membrane nearly exclusively composed of unconventional membrane lipids: ladderane lipids (Sinninghe Damste et al., 2002). The structure of ladderane lipids is unique in nature, they have been found so far only in Anammox bacteria. These lipids are composed of pentacycloanammoxic acids, which contain five linearly concatenated cyclobutane rings (Mascitti and Corey, 2004). Due to the very slow metabolism of Anammox bacteria, a very dense and impermeable membrane is required to maintain concentration gradients during the Anammox reaction. Such a membrane also protects the cell from the toxic intermediates. Figure 1.8. Anammox phylogenetic tree (Kuenen, 2008). The genome of Candidatus "K. stuttgartiensis" has been sequenced (Strous et al., 2006). The genome data illuminate the evolutionary history of the Planctomycetes and Anammox bacteria in particular found to be related to a clade of intracellular parasites known as the Chlamydiae. Candidate genes responsible for ladderane biosynthesis and biological hydrazine metabolism were identified. Besides, an unexpected metabolic versatility was discovered, this versatility could justify why although slow and specialized, Anammox bacteria are global players in the biological nitrogen cycle (Arrigo, 2005). 1.5.5. Combination of processes In order to perform the nitrogen removal from wastewater, the processes previously described can be combined according to two main different configurations based on nitrification-denitrification or partial nitrification-Anammox. 1.5.5.1. Nitrification-denitrification Due to the different operation conditions of these processes, two different tanks are needed: one stirred but not aerated where denitrification is carried out, and another aerated, where ammonium and organic matter are simultaneously oxidized. Combined nitrification-denitrification processes are effective in maintaining neutral pH level in the reactor, without the addition of external acid/base source. During nitrification alkalinity is consumed, but alkalinity is produced during denitrification. 30 Introduction Combining the equations obtained for the nitrification with those corresponding to the denitrification and taking into account the carbonate system equilibrium, the stoichiometry of the complete nitrogen removal by nitrification-denitrification with three different organic matter sources: methanol, acetate or acetic acid is as follows (Eq. 1.20 - 1.22). Nitrification-denitrification over NO3NH4+ + 0.83 CH3OH + 2 O2 + HCO3-  0.5 N2 + 4.17 H2O + 1.83 CO2 (1.20) NH4+ + 0.625 CH3COO- + 2 O2 + 0.375 HCO3-  0.5 N2 + 3.125 H2O + 1.625 CO2 (1.21) NH4+ + 0.625 CH3COOH + 2 O2 + HCO3-  0.5 N2 + 3.75 H2O + 2.25 CO2 (1.22) 1.5.5.2. Partial nitrification-Anammox When the partial nitrification and the Anammox processes are combined certain considerations have to be taken into account. From the stoichiometry (Eq. 1.19) it can be inferred that Anammox bacteria need ammonium and nitrite in a ratio around 1:1.3. To reach this objective, half of the ammonium fed has to be converted in nitrite by AOB and therefore, the oxidation of nitrite to nitrate carried out by NOB has to be avoided. Using this combined process of partial nitrification-Anammox, two alternatives are available to obtain autotrophic nitrogen removal. Using a two reactor configuration where half of the ammonium is oxidized to nitrite in the first reactor and a second Anammox reactor in anoxic conditions where ammonia and nitrite are removed producing a small amount of nitrate. Autotrophic nitrogen removal can also be achieved in a one reactor configuration where both populations: AOB and Anammox bacteria coexist under controlled aerobic conditions. The global stoichiometry of the total process combining partial nitrification and Anammox is represented in Eq. 1.23. NH4+ + 0.85 O2 + 1.11 HCO3- → 0.44 N2 + 0.11 NO3- + 2.56 H2O + 1.11 CO2 (1.23) In comparison with the conventional nitrification-denitrification process, in the case of the Anammox based processes less oxygen is required and no organic matter must be present or added as external carbon source what makes this process suitable to treat wastewaters with low COD/N ratio, e.g., reject water. These processes are under research to be applied in WWTP for the removal of nitrogen. 1.6. Nitrogen removal treatment plants 1.6.1 Conventional WWTP configuration Nowadays, activated sludge is the most used process for the biological treatment of domestic and industrial wastewaters. In most WWTP, wastewater is initially led to a stirred denitrifying basin with no air supply to avoid organic matter limitation in the denitrifying zone, and afterwards to a basin with aeration where nitrification occurs (Fig. 1.9). Nitrate containing wastewater from the aeration basin is recycled and mixed with the organic carbon rich wastewater entering the denitrifying zone. The nitrogen removal efficiency depends on the amount and presence of suitable organic substrates in the wastewater. This strategy has several disadvantages such as: a) the simultaneous growth of autotrophic nitrifiers together with heterotrophic biomass which leads to relatively low nitrification rates, since heterotrophs grow much faster and b) the need of recirculation between both reactors. To ensure a sufficient nitrification rate, the 31 Chapter 1 Full scale reactors with the Sharon-Anammox configuration have already been implemented in the Netherlands. The low growth of Anammox microorganisms delayed the start-up of the process but the design capacity of nitrogen removal (750 kg N d-1) was finally achieved after 3.5 years (Table 1.7). Dehidratation system Influent Effluent Sludge digester Sludge thickener Anoxic Aerobic Activated sludge reactor Dehidratated sludge Secondary settler Water line Sludge line Water line Sludge line Sludge treatment SHARON Reactor Anammox Reactor Figure 1.11. Sharon-Anammox process to treat reject water. 1.6.5.1. One reactor configuration The partial nitrification and the Anammox processes can also be carried out together in a single reactor. This combined process has been called different names, CANON: Completely Autotrophic Nitrogen removal Over Nitrite process (Third et al., 2001), OLAND: Oxygen-Limited Autotrophic Nitrification-Denitrification (Kuai and Verstraete, 1998), deammonification (Hippen et al., 1997; Helmer et al., 2001) and SNAP: Single-stage Nitrogen removal using Anammox and Partial nitritation (Furukawa et al., 2006). Once the single unit option is chosen the process occurs only under controlled aerated conditions. Under oxygen-limited conditions a coculture of aerobic and anaerobic ammonium-oxidizing bacteria can be established in a single unit. With the development of biofilms, aerobic and anoxic zones can be present within the biofilm due to the gradients of oxygen generated with oxygen consumption by aerobic microorganisms. This will allow the development of AOB in the aerobic layers and Anammox bacteria in the anoxic ones (Fig. 1.12). In those systems, NOB growth has to be avoided since the oxidation of nitrite to nitrate is not desired. NOB compete for oxygen with the aerobic AOB and for nitrite with Anammox bacteria, and thus its growth (and subsequent nitrate production) can be prevented, e.g., controlling the DO concentration in the bulk liquid (Rosenwinkel and Cornelius, 2005). 38 Introduction AEROBIC PARTIAL NITRIFICATION O 2 NH 4+ NO 2N 2 + some NO 3ANAMMOX ANOXIC Figure 1.12. CANON process in biomass aggregates. The one reactor configuration has been developed at full scale to treat reject water in different reactors and biofilm configurations but all of them applied to effluents with temperatures ranging between 30-40 ºC since this is the optimum range of temperatures for Anammox bacteria. Wett (2006, 2007) developed the autotrophic nitrogen removal process working with aggregates: the DEMON® system which used pH to control the two steps of the nitrogen removal. This process was first implemented in Austria and then in Switzerland. The availability of inoculum and the experience learned reduced the start-up time from 2.5 years to 2 months (Table 1.7). The autotrophic nitrogen removal was also developed at full scale in biofilms grown on Kaldness rings in MBBR (Moving Bed Biofilm Reactor) in Hattingen (Rosenwinkel and Cornelius, 2005) and Sweden (personal communication) (Table 1.7). Table 1.7. Anammox plants at industrial scale (adapted from Abma et al., 2007 and Wett, 2007). Project Application Volumea (m3) Capacity achieved (kg N d-1) Start-up time Waterboard Hollandse Delta, The Netherlands (2 units) Municipal (reject water) 72 750 3.5 year Strass, Austria (1 unit) Municipal (reject water) 500 350 2.5 year IWL, The Netherlands (2 units) Tannery 100 150b 1 year Waterstromen, The Netherlands (1 unit) Potato processing 600 700b 6 months Himmerfjärdsverket c, Sweden (1 unit) Municipal (reject water) 700 240 6 months Glanerland, Switzerland (1 unit) Municipal (reject water) 400 250 2 months Semiconductor Plant, Japan (2 units) Semiconductor 58 220 2 months a For two units systems the volume corresponds to the Anammox unit. b No more nitrogen available. c Personal communication. 1.6.6. Comparison of the available technologies to treat reject water The bio-augmentation, partial-denitrification and partial nitrification-Anammox technologies can be applied to optimize the nitrogen removal in WWTP. The selection of one of the strategies will depend on the specific limitations in the nitrogen removal of each plant according to the decision tree represented in Fig. 1.13. The first decision concerns whether the limiting step in the nitrogen removal in the main plant is the nitrification 39 Chapter 1 or the denitrification. In case the nitrification or denitrification are limited by the volume of the aerobic or the anoxic tank, the bio-augmentation is the recommended technology. However, if the aeration or the organic matter content is limiting the nitrogen removal, the “nitrite” route processes are recommended. The ammonium counter-ion plays an important role and will determine the most suitable technology to be applied. In most WWTP this counter ion will be the bicarbonate ion which works as pH buffer in the nitrification process and the recommended technology would be the partial nitrificationAnammox. Finally, if the sludge is dried the counter-ion will be the acetate ion; in this case, the partial nitrificationdenitrification will be the most suitable technology. Besides these aspects, the start-up time, the risk of failure, the flexibility of the process, etc. will also determine the decision of which technology to implement. Supernatant Limiting process Limiting factor Limiting factor Nitrification Denitrification Counter-ion NH 4+ Aeration Capacity Acetate HCO 3SRT Denitrification SRT Nitrification Bio-augmentation Partial Nit/Denit.Partial Nit/Denit. Partial Nit/ Anammox Partial Nit/ Anammox Organic Matter Figure 1.13. Diagram of selection of the best technology to treat reject water. 1.7. Systems based on aerobic granulation As aforementioned the main drawback of autotrophic microorganisms is their slow growth rate in comparison to heterotrophic bacteria. This makes the start-up of biological reactors based on autotrophic bacteria difficult. To overcome this drawback, the withdrawal of microorganisms has to be avoided and therefore, the SRT in the bioreactors has to be high enough to accumulate the slowest microorganisms, e.g., Anammox biomass. In the last years, aerobic granulation arose as a new technology to accumulate high biomass concentrations and consists of the development of aggregates that grow without the need of carrier material as a special kind of biofilm. Biofilms are ubiquitous life forms consisting of microorganisms embedded in a matrix of biological origin called extracellular polymeric substances. Aerobic granular sludge consists of compact and dense microbial aggregates with a spherical outer shape, which represent a particular or special case of biofilm development (Liu and Tay, 2004). The aerobic formation of such structures is linked with the stressing operating conditions imposed for its development, resulting in a self microbial adhesion process, which initiates the granule formation (Liu and Tay, 2004). The interest in developing aerobic granular sludge systems is mainly related to the compactness of their design in comparison to conventional activated sludge systems due to the higher biomass concentration achieved inside 40 Introduction the biological unit and to the lower settling size required. The application of this kind of systems can also contribute to the reduction of sludge production by three ways: a) Reduction of the amount of sludge generated: Tay et al., (2001) observed that the sludge production of granular systems was 30% lower than in activated sludge systems. A possible explanation is that microorganisms which grow forming granules have a higher percentage of exopolymers in their composition to maintain their structure. The production of these compounds implies a change of their metabolism and more energy is used in this process compared to flocculant microorganisms. b) Reduction of sludge volume: during the treatment of sludge generated most of the processes applied are focused on reducing its volume by decreasing its water content. Granules have denser and more compact structures than flocs which caused directly a lower sludge volume. On the other hand, during granulation cells increase their hydrofobicity from 50 to 80% which could favour dewatering processes (Tay et al., 2001). c) Improvement of anaerobic sludge digestion: the biodegradability of aerobic granules is also higher than flocs due to its high content of polyhydroxybutyrate (Fang et al., 2009). Recent research showed that it is possible to grow Granular sludge in Sequencing Batch Reactors (GSBR) at large dissolved oxygen concentrations using either synthetic wastewater (Beun et al., 1999; Mosquera-Corral et al., 2005b) or industrial effluents (Arrojo et al., 2004; Tsuneda et al., 2006). The operational sequence of GSBR is specially designed to enhance the formation of fast settling granules in which two different zones can be distinguish: (i) the outer aerobic shell, in which heterotrophic and autotrophic microorganisms will remove COD and oxidize ammonium into nitrate, respectively, and (ii) the inner anoxic core in which denitrification can take place. In addition, the phosphate removal is possible due to the use of anaerobic/aerobic phases in the sequential operation, which enable the development of specific biomass, the so called polyphosphate accumulating organisms (PAO). The aim of this thesis is linked to the application of the autotrophic nitrogen removal. The granulation concept is interesting since by having favourable conditions for granulation, stable Anammox population would develop in the inner core of aerobic granules. The consumption of DO in the external layers will allow the presence of internal anoxic layers where oxygen is not available. 1.7.1. Modelling of granular systems The complexity of modelling biofilm systems such as granular biomass thrives in the need of combining diffusion and reaction kinetics. A wide range of parameters influence the characteristics and performance of biofilms or granules. Therefore mathematical models represent an important and fundamental tool to know in detail the processes developed in the granules. Furthermore they can provide a solid foundation for design and operation without the time consumption and materials expense of the experimental approach. Anyway models, to be valuable, must be calibrated and validated with experimental data. For the conventional floc-based activated sludge systems, the Activated Sludge Model (ASM) established by the International Water Association (IWA), provides a consistent framework for the description of biological processes. The ASM can be used as a basis to develop the aerobic granule models but introducing several modifications. The IWA developed four Activated Sludge Models, the ASM1, ASM2, ASM2d and ASM3 (Henze et al., 2000). The ASM1 is the structure and platform for further development. The ASM1 allows the simulation of organic matter removal and biological nitrification and denitrification (N-removal) in activated sludge systems. The ASM2 is an extension of the ASM1 and includes biological phosphorous 41 Chapter 1 removal processes, and the ASM2d includes denitrifying PAOs. The ASM3 is a new modelling platform based on recent developed knowledge of the activated sludge processes and includes the possibility of following up the concentrations of internal storage compounds. The decay process from ASM1 is replaced by an endogenous respiration process in ASM3, a more realistic approach under a microbiological point of view (Van Loosdrecht and Jetten, 1998). The matrix structure and organization of ASM models makes an easy integration of these models into several simulation programs feasible: AQUASIM, WEST, Plan-It STOAT… and also the developement of the model using mathematical tools as MATLAB. In the ASM, biomass is assumed in suspension in the bulk liquid. However, this approach is not suitable when bacteria grow forming colonies that arise in biofilms, aggregates or aerobic granules. The aggregation of bacteria creates a separation between the bulk liquid and the biological media (Fig. 1.14). It can be assumed that the bulk liquid is well mixed and therefore homogeneous, however, in the bacterial phase biological processes and substrate diffusion velocities provoke concentration gradients of substrates: DO, COD, ammonium, nitrate, nitrite, phosphates… Those concentration gradients are influenced by many factors, e.g., diffusion coefficients, conversion rates, granule size, biomass spatial distribution, density, etc. There are several influences between every factor, thus the effect of separate factors cannot be studied experimentally, and the simulation models can give more insight at a microscopic scale by solving differential equations. Figure 1.14. Relevant processes considered in an aerobic granule. Lübken et al., (2005) researched if the model ASM3 could be used as a first simplification to simulate nutrient removal with aerobic granular sludge in a SBR. The model proved to be capable of describing the performance of a lab-scale SBR reactor concerning COD, ammonium and nitrate in the bulk liquid. However, for describing the processes inside of a granule, biofilm processes should be included in the model. The mathematical modelling of biofilms has gained increasing interest in the last decade and a wide range of biofilm models of different complexity applied to different types of benchmark problems has been developed. Aerobic granules can be viewed as a special type of biofilm but without carriers for the biofilm attachment. For this reason models of aerobic granular systems have been based on biofilm models. Modelling of biofilm or granular structures is possible using one dimensional models (1-D) or multidimensional models (2D and 3-D). One dimensional models are simpler and therefore require less computational effort. Nevertheless, in 1-D models several assumptions are made. It is assumed that the substrates gradients are some orders of 42 Introduction magnitude higher in the perpendicular direction to the attachment surface than in the parallel plane to it. Important characteristics derived from the dynamics of biofilm structure must be taken into account once assumed 1-D transport. Some examples of properties that must be explicitly defined are: external and internal mass transfer coefficients, changes in pore volume and motility of bacterial species inside the biofilm matrix (Xavier et al., 2005). Extensive research is being performed to model granular systems. Beun et al., (2001) described the COD and N removal in a granular sludge batch reactor using a 1-D model implemented in AQUASIM. These authors showed that nitrification, denitrification and COD removal can occur simultaneously in a granular sludge SBR. Su and Yu, (2006) established a generalized model for aerobic granular SBR taking into account the removal of nitrogen and organic compounds, reactor hydrodynamics, oxygen transfer and substrates diffusion. Recently Ni et al., (2008) used a modification of the ASM3 model to describe the simultaneous autotrophic and heterotrophic growth in aerobic granular SBR, implementing the model in AQUASIM. De Kreuk et al., (2007) introduced the removal of phosphate in addition to the removal of COD and nitrogen as a process in the model. They used a 1-D model to study the influence of different parameters (oxygen concentration, temperature, granule diameter, sludge loading rate and cycle configuration) on the granular SBR operation. The development of multidimensional (2-D or 3-D) models offers more potential to predict local compositions of particulate and dissolved variables. Several approaches to multidimensional modelling of biofilm can be found in the literature (Xavier et al., 2005). In general 2-D and 3-D biofilm models could be divided in two classes according to the way of describing the biomass: discrete individual particles or a continuum body. Discrete particle models are the most suitable for extrapolation to granular biomass. Individual-based models (Ibm) are, discrete particle models, widely applied to study effects of spatially multidimensional gradients in biofilms. Xavier et al., (2007) presented a multiscale model of a granular sludge batch reactor describing the complex dynamics of populations and nutrient removal. There is an extensive research to perform in the field of nitrogen removal from the point of view of the involved processes themselves if autotrophic processes are required and of the diffusion limitations in case of working with biofilms or granules. Information obtained from modelled results is of great interest in order to reduce the number of experiments to be performed. 1.8. References Abma, W.R., Schultz, C.E., Mulder, J.W., van der Star, W.R., Strous, M., Tokutomi, T., and van Loosdrecht, M.C. (2007) Full-scale granular sludge Anammox process. Water Science and Technology 55: 27-33. Aiyuk, S., Forrez, I., Lieven, D.K., van Haandel, A., and Verstraete, W. (2006) Anaerobic and complementary treatment of domestic sewage in regions with hot climates - A review. Bioresource Technology 97: 2225-2241. Anthonisen, A.C., Loehr, R.C., Prakasam, T.B.S., and Srinath, E.G. (1976) Inhibition of Nitrification by Ammonia and NitrousAcid. Journal Water Pollution Control Federation 48: 835-852. Arrigo, K.R. (2005) Marine microorganisms and global nutrient cycles. Nature 437: 349-355. Arrojo, B., Mosquera-Corral, A., Garrido, J.M., and Mendez, R. (2004) Aerobic granulation with industrial wastewater in sequencing batch reactors. Water Research 38: 3389-3399. Berends, D.H.J.G., Salem, S., van der Roest, H.F., and van Loosdrecht, M.C.M. (2005) Boosting nitrification with the BABE technology. Water Science and Technology 52: 63-70. Bernet, N., Sanchez, O., Cesbron, D., Steyer, J.P., and Delgenes, J.P. (2005) Modeling and control of nitrite accumulation in a nitrifying biofilm reactor. Biochemical Engineering Journal 24: 173-183. 43 Chapter 1 Beun, J.J., Hendriks, A., Van Loosdrecht, M.C.M., Morgenroth, E., Wilderer, P.A., and Heijnen, J.J. (1999) Aerobic granulation in a sequencing batch reactor. Water Research 33: 2283-2290. Beun, J.J., Heijnen, J.J., and van Loosdrecht, M.C.M. (2001) N-removal in a granular sludge sequencing batch airlift reactor. Biotechnology and Bioengineering 75: 82-92. Broda, E. (1977) Two kinds of lithotrophs missing in nature. Zeitschrift für Allgemeine Mikrobiologie 17: 491-493. Campos, J.L., Arrojo, B., Vazquez-Padin, J.R., Mosquera-Corral, A., and Mendez, R. (2009) N2O Production by Nitrifying Biomass Under Anoxic and Aerobic Conditions. Applied Biochemistry and Biotechnology 152: 189-198. Cecen, F., and Gonenc, I.E. (1995) Criteria for Nitrification and Denitrification of High-Strength Wastes in 2 Upflow Submerged Filters. Water Environment Research 67: 132-142. Daims, H., Nielsen, P.H., Nielsen, J.L., Juretschko, S., and Wagner, M. (2000) Novel Nitrospira-like bacteria as dominant nitrite-oxidizers in biofilms from wastewater treatment plants: diversity and in situ physiology. Water Science and Technology 41: 85-90. Dapena-Mora, A., Campos, J.L., Mosquera-Corral, A., Jetten, M.S.M., and Mendez, R. (2004a) Stability of the ANAMMOX process in a gas-lift reactor and a SBR. Journal of Biotechnology 110: 159-170. Dapena-Mora, A., Van Hulle, S.W.H., Campos, J.L., Mendez, R., Vanrolleghem, P.A., and Jetten, M. (2004b) Enrichment of Anammox biomass from municipal activated sludge: experimental and modelling results. Journal of Chemical Technology and Biotechnology 79: 1421-1428. Dapena-Mora, A., Fernandez, I., Campos, J.L., Mosquera-Corral, A., Mendez, R., and Jetten, M.S.M. (2007) Evaluation of activity and inhibition effects on Anammox process by batch tests based on the nitrogen gas production. Enzyme and Microbial Technology 40: 859-865. De Kreuk, M.K., Picioreanu, C., Hosseini, M., Xavier, J.B., and van Loosdrecht, M.C.M. (2007) Kinetic model of a granular sludge SBR: Influences on nutrient removal. Biotechnology and Bioengineering 97: 801-815. Dold, P.L., Fairlamb, P.M., Jones, R., Takacs, I., and Murthy, S. (2007) Sidestream Modelling incorporated into whole plant simulation. In CLONIC Workshop. Barcelone, pp. 114-122. Dosta, J., Gali, A., El-Hadj, T.B., Mace, S., and Mata-Alvarez, J. (2007) Operation and model description of a sequencing batch reactor treating reject water for biological nitrogen removal via nitrite. Bioresource Technology 98: 2065-2075. Duce, R.A., LaRoche, J., Altieri, K., Arrigo, K.R., Baker, A.R., Capone, D.G., Cornell, S., Dentener, F., Galloway, J., Ganeshram, R.S., Geider, R.J., Jickells, T., Kuypers, M.M., Langlois, R., Liss, P.S., Liu, S.M., Middelburg, J.J., Moore, C.M., Nickovic, S., Oschlies, A., Pedersen, T., Prospero, J., Schlitzer, R., Seitzinger, S., Sorensen, L.L., Uematsu, M., Ulloa, O., Voss, M., Ward, B., and Zamora, L. (2008) Impacts of atmospheric anthropogenic nitrogen on the open ocean. Science 320: 893-897. Egli, K., Fanger, U., Alvarez, P.J.J., Siegrist, H., van der Meer, J.R., and Zehnder, A.J.B. (2001) Enrichment and characterization of an Anammox bacterium from a rotating biological contactor treating ammonium-rich leachate. Archives of Microbiology 175: 198-207. EPA (1993) Manual of nitrogen control. Washington DC, USA: US environmental protection agency. Fang, F., Liu, X.W., Xu, J., Yu, H.Q., and Li, Y.M. (2009) Formation of aerobic granules and their PHB production at various substrate and ammonium concentrations. Bioresource Technology 100: 59-63. Fuerst, J.A. (2005) Intracellular compartmentation in planctomycetes. Annual Review of Microbiology 59: 299-328. Fujii, T., Sugino, H., Rouse, J.D., and Furukawa, K. (2002) Characterization of the microbial community in an anaerobic ammonium-oxidizing biofilm cultured on a nonwoven biomass carrier. Journal of Bioscience and Bioengineering 94: 412-418. Furukawa, K., Lieu, P.K., Tokitoh, H., and Fujii, T. (2006) Development of single-stage nitrogen removal using Anammox and partial nitritation (SNAP) and its treatment performances. Water Science and Technology 53: 83-90. Fux, C., Velten, S., Carozzi, V., Solley, D., and Keller, J. (2006) Efficient and stable nitritation and denitritation of ammoniumrich sludge dewatering liquor using an SBR with continuous loading. Water Research 40: 2765-2775. Gali, A., Dosta, J., van Loosdrecht, M.C.M., and Mata-Alvarez, J. (2007) Two ways to achieve an Anammox influent from real reject water treatment at lab-scale: Partial SBR nitrification and SHARON process. Process Biochemistry 42: 715720. 44 Introduction Galloway, J.N., Townsend, A.R., Erisman, J.W., Bekunda, M., Cai, Z.C., Freney, J.R., Martinelli, L.A., Seitzinger, S.P., and Sutton, M.A. (2008) Transformation of the nitrogen cycle: Recent trends, questions, and potential solutions. Science 320: 889-892. Garrido, J.M., vanBenthum, W.A.J., vanLoosdrecht, M.C.M., and Heijnen, J.J. (1997) Influence of dissolved oxygen concentration on nitrite accumulation in a biofilm airlift suspension reactor. Biotechnology and Bioengineering 53: 168178. Gieseke, A., Purkhold, U., Wagner, M., Amann, R., and Schramm, A. (2001) Community structure and activity dynamics of nitrifying bacteria in a phosphate-removing biofilm. Applied and Environmental Microbiology 67: 1351-1362. Gijzen, H.J. (2001) Anaerobes, aerobes and phototrophs - A winning team for wastewater management. Water Science and Technology 44: 123-132. Guisasola, A., Jubany, I., Baeza, J.A., Carrera, J., and Lafuente, J. (2005) Respirometric estimation of the oxygen affinity constants for biological ammonium and nitrite oxidation. Journal of Chemical Technology and Biotechnology 80: 388396. Hellinga, C., Schellen, A.A.J.C., Mulder, J.W., van Loosdrecht, M.C.M., and Heijnen, J.J. (1998) The SHARON process: An innovative method for nitrogen removal from ammonium-rich wastewater. Water Science and Technology 37: 135-142. Helmer, C., Tromm, C., Hippen, A., Rosenwinkel, K.H., Seyfried, C.F., and Kunst, S. (2001) Single stage biological nitrogen removal by nitritation and anaerobic ammonium oxidation in biofilm systems. Water Science and Technology 43: 311320. Henze, M., Gujer, W., Mino, T., and van Loosdrecht, M. (2000) Activated sludge models ASM1, ASM2, ASM2d and ASM3. London: IWA Publishing. Hippen, A., Rosenwinkel, K.H., Baumgarten, G., and Seyfried, C.F. (1997) Aerobic deammonification: A new experience in the treatment of wastewaters. Water Science and Technology 35: 111-120. Janus, H.M., and vanderRoest, H.F. (1997) Don't reject the idea of treating reject water. Water Science and Technology 35: 27-34. Jetten, M.S.M., Schmid, M., Schmidt, I., and Van Dogen, U. (2002) Improved nitrogen removal by application of new nitrogen-cycle bacteria. Reviews in Environmental Science and Biotechnology 1 51-63. Jubany, I., Carrera, J., Lafuente, J., and Baeza, J.A. (2008) Start-up of a nitrification system with automatic control to treat highly concentrated ammonium wastewater: Experimental results and modeling. Chemical Engineering Journal 144: 407-419. Kartal, B., Rattray, J., van Niftrik, L.A., van de Vossenberg, J., Schmid, M.C., Webb, R.I., Schouten, S., Fuerst, J.A., Damste, J.S., Jetten, M.S.M., and Strous, M. (2007) Candidatus "Anammoxoglobus propionicus" a new propionate oxidizing species of anaerobic ammonium oxidizing bacteria. Systematic and Applied Microbiology 30: 39-49. Kartal, B., van Niftrik, L., Rattray, J., van de Vossenberg, J.L.C.M., Schmid, M.C., Sinninghe Damste, J., Jetten, M.S.M., and Strous, M. (2008) Candidatus "Brocadia fulgida": an autofluorescent anaerobic ammonium oxidizing bacterium. FEMS Microbiology Ecology 63: 46-55. Konneke, M., Bernhard, A.E., de la Torre, J.R., Walker, C.B., Waterbury, J.B., and Stahl, D.A. (2005) Isolation of an autotrophic ammonia-oxidizing marine archaeon. Nature 437: 543-546. Koops, H.P., and Pommerening-Roser, A. (2001) Distribution and ecophysiology of the nitrifying bacteria emphasizing cultured species. FEMS Microbiology Ecology 37: 1-9. Kuai, L.P., and Verstraete, W. (1998) Ammonium removal by the oxygen-limited autotrophic nitrification-denitrification system. Applied and Environmental Microbiology 64: 4500-4506. Kuenen, J.G. (2008) Anammox bacteria: from discovery to application. Nature Reviews Microbiology 6: 320-326. Kuypers, M.M.M., Sliekers, A.O., Lavik, G., Schmid, M., Jorgensen, B.B., Kuenen, J.G., Damste, J.S.S., Strous, M., and Jetten, M.S.M. (2003) Anaerobic ammonium oxidation by Anammox bacteria in the Black Sea. Nature 422: 608-611. Kuypers, M.M.M., Lavik, G., Woebken, D., Schmid, M., Fuchs, B.M., Amann, R., Jorgensen, B.B., and Jetten, M.S.M. (2005) Massive nitrogen loss from the Benguela upwelling system through anaerobic ammonium oxidation. Proceedings of the National Academy of Sciences of the United States of America 102: 6478-6483. 45 Chapter 1 Lindsay, M.R., Webb, R.I., Strous, M., Jetten, M.S., Butler, M.K., Forde, R.J., and Fuerst, J.A. (2001) Cell compartmentalisation in planctomycetes: novel types of structural organisation for the bacterial cell. Archives of Microbiology 175: 413-429. Liu, Y., and Tay, J.H. (2004) State of the art of biogranulation technology for wastewater treatment. Biotechnology Advances 22: 533-563. Lübken, M., Schwarzenbeck, N., Wichern, M., and Wilderer, P.A. (2005) Modelling nutrient removal of an aerobic granular sludge lab-scale SBR using ASM3. In Aerobic Granular Sludge. Bathe, S., de Kreuk, M., McSwain, B., and Schwarzenbeck, N. (eds). London: IWA Publishing, pp. 63-70. Mascitti, V., and Corey, E.J. (2004) Total synthesis of (+/-)-pentacycloanammoxic acid. Journal of the American Chemical Society 126: 15664-15665. Metcalf L. and Eddy H. (1995). In spanish: Ingeniería de aguas residuales. Tratamiento, vertido y reutilización. McGrawHill/Interamericana de España S.A. (Third Edition). Mosquera-Corral, A., Gonzalez, F., Campos, J.L., and Mendez, R. (2005a) Partial nitrification in a SHARON reactor in the presence of salts and organic carbon compounds. Process Biochemistry 40: 3109-3118. Mosquera-Corral, A., Vázquez-Padín, J.R., Arrojo, B., Campos, J.L., and Méndez, R. (2005b) Nitrifying granular sludge in a sequencing batch reactor. In Aerobic Granular Sludge. Bathe, S., de Kreuk, M., McSwain, B., and Schwarzenbeck, N. (eds). London: IWA Publishing, pp. 63-70. Mulder, A., Vandegraaf, A.A., Robertson, L.A., and Kuenen, J.G. (1995) Anaerobic Ammonium Oxidation Discovered in a Denitrifying Fluidized-Bed Reactor. FEMS Microbiology Ecology 16: 177-183. Mulder, J.W., van Loosdrecht, M.C.M., Hellinga, C., and van Kempen, R. (2001) Full-scale application of the SHARON process for treatment of rejection water of digested sludge dewatering. Water Science and Technology 43: 127-134. Mulder, A. (2003) The quest for sustainable nitrogen removal technologies. Water Science and Technology 48: 67-75. Ni, B.J., Yu, H.Q., and Sun, Y.J. (2008) Modeling simultaneous autotrophic and heterotrophic growth in aerobic granules. Water Research 42: 1583-1594. Pambrun, V., Paul, E., and Sprandio, M. (2006) Modeling the partial nitrification in sequencing batch reactor for biomass adapted to high ammonia concentrations. Biotechnology and Bioengineering 95: 120-131. Revsbech, N.P., Risgaard-Petersen, N., Schramm, A., and Nielsen, L.P. (2006) Nitrogen transformations in stratified aquatic microbial ecosystems. Antonie Van Leeuwenhoek International Journal of General and Molecular Microbiology 90: 361-375. Rosenwinkel, K.H., and Cornelius, A. (2005) Deammonification in the moving-bed process for the treatment of wastewater with high ammonia content. Chemical Engineering & Technology 28: 49-+. Salem, S., Berends, D., Heijnen, J.J., and van Loosdrecht, M.C.M. (2002) Model-based evaluation of a new upgrading concept for N-rernoval. Water Science and Technology 45: 169-176. Schmid, M., Twachtmann, U., Klein, M., Strous, M., Juretschko, S., Jetten, M., Metzger, J.W., Schleifer, K.H., and Wagner, M. (2000) Molecular evidence for genus level diversity of bacteria capable of catalyzing anaerobic ammonium oxidation. Systematic and Applied Microbiology 23: 93-106. Schmid, M., Walsh, K., Webb, R., Rijpstra, W.I.C., Pas-Schoonen, K., Verbruggen, M.J., Hill, T., Moffett, B., Fuerst, J., Schouten, S., Damste, J.S.S., Harris, J., Shaw, P., Jetten, M., and Strous, M. (2003) Candidatus "Scalindua brodae", sp nov., Candidatus "Scalindua wagneri", sp nov., two new species of anaerobic ammonium oxidizing bacteria. Systematic and Applied Microbiology 26: 529-538. Schmid, M.C., Maas, B., Dapena, A., Pas-Schoonen, K.V., de Vossenberg, J.V., Kartal, B., van Niftrik, L., Schmidt, I., Cirpus, I., Kuenen, J.G., Wagner, M., Damste, J.S.S., Kuypers, M., Revsbech, N.P., Mendez, R., Jetten, M.S.M., and Strous, M. (2005) Biomarkers for in situ detection of anaerobic ammonium-oxidizing (Anammox) bacteria. Applied and Environmental Microbiology 71: 1677-1684. Siegrist, H. (1996) Nitrogen removal from digester supernatant - Comparison of chemical and biological methods. Water Science and Technology 34: 399-406. Siegrist, H., Salzgeber, D., Eugster, J., and Joss, A. (2008) Anammox brings WWTP closer to energy autarky due to increased biogas production and reduced aeration energy for N-removal. Water Science and Technology 57: 383-388. 46 Introduction Sinninghe Damste, J.S., Strous, M., Rijpstra, W.I., Hopmans, E.C., Geenevasen, J.A.J., van Duin, A.C.T., van Niftrik, L.A., and Jetten, M.S.M. (2002) Linearly concatenated cyclobutane lipids form a dense bacterial membrane. Nature 419: 708-712. Solbe, J.F.D.G., and Shurben, D.G. (1989) Toxicity of Ammonia to Early Life Stages of Rainbow-Trout (Salmo-Gairdneri). Water Research 23: 127-129. Sorokin, D.Y., Muyzer, G., Brinkhoff, T., Kuenen, J.G., and Jetten, M.S.M. (1998) Isolation and characterization of a novel facultatively alkaliphilic Nitrobacter species, N-alkalicus sp. nov. Archives of Microbiology 170: 345-352. STOWA (1996) One reactor system for ammonia removal via nitrite (In dutch). In Report no 96-01. STOWA, Utrecht, The Netherlands. Strous, M., vanGerven, E., Kuenen, J.G., and Jetten, M. (1997) Effects of aerobic and microaerobic conditions on anaerobic ammonium-oxidizing (Anammox) sludge. Applied and Environmental Microbiology 63: 2446-2448. Strous, M., Heijnen, J.J., Kuenen, J.G., and Jetten, M.S.M. (1998) The sequencing batch reactor as a powerful tool for the study of slowly growing anaerobic ammonium-oxidizing microorganisms. Applied Microbiology and Biotechnology 50: 589-596. Strous, M., Kuenen, J.G., and Jetten, M.S.M. (1999a) Key physiology of anaerobic ammonium oxidation. Applied and Environmental Microbiology 65: 3248-3250. Strous, M., Fuerst, J.A., Kramer, E.H.M., Logemann, S., Muyzer, G., Pas-Schoonen, K.T., Webb, R., Kuenen, J.G., and Jetten, M.S.M. (1999b) Missing lithotroph identified as new planctomycete. Nature 400: 446-449. Strous, M., Kuenen, J.G., Fuerst, J.A., Wagner, M., and Jetten, M.S.M. (2002) The Anammox case - A new experimental manifesto for microbiological eco-physiology. Antonie van Leeuwenhoek 81: 693-702. Strous, M., and Jetten, M.S.M. (2004) Anaerobic oxidation of methane and ammonium. Annual Review of Microbiology 58: 99-117. Strous, M., Pelletier, E., Mangenot, S., Rattei, T., Lehner, A., Taylor, M.W., Horn, M., Daims, H., Bartol-Mavel, D., Wincker, P., Barbe, V., Fonknechten, N., Vallenet, D., Segurens, B., Schenowitz-Truong, C., Medigue, C., Collingro, A., Snel, B., Dutilh, B.E., Op den Camp, H.J.M., van der Drift, C., Cirpus, I., van de Pas-Schoonen, K.T., Harhangi, H.R., van Niftrik, L., Schmid, M., Keltjens, J., van de Vossenberg, J., Kartal, B., Meier, H., Frishman, D., Huynen, M.A., Mewes, H.W., Weissenbach, J., Jetten, M.S.M., Wagner, M., and Le Paslier, D. (2006) Deciphering the evolution and metabolism of an Anammox bacterium from a community genome. Nature 440: 790-794. Su, K.Z., and Yu, H.Q. (2006) A generalized model for aerobic granule-based sequencing batch reactor. 1. Model development. Environmental Science & Technology 40: 4703-4708. Tay, J.H., Liu, Q.S., and Liu, Y. (2001) The effects of shear force on the formation, structure and metabolism of aerobic granules. Applied Microbiology and Biotechnology 57: 227-233. Third, K.A., Sliekers, A.O., Kuenen, J.G., and Jetten, M.S.M. (2001) The CANON system (completely autotrophic nitrogenremoval over nitrite) under ammonium limitation: Interaction and competition between three groups of bacteria. Systematic and Applied Microbiology 24: 588-596. Trimmer, M., Nicholls, J.C., and Deflandre, B. (2003) Anaerobic ammonium oxidation measured in sediments along the Thames estuary, United Kingdom. Applied and Environmental Microbiology 69: 6447-6454. Tsuneda, S., Ogiwara, M., Ejiri, Y., and Hirata, A. (2006) High-rate nitrification using aerobic granular sludge. Water Science and Technology 53: 147-154. Turk, O., and Mavinic, D.S. (1989) Maintaining Nitrite Buildup in a System Acclimated to Free Ammonia. Water Research 23: 1383-1388. Vadivelu, V.M., Keller, J., and Yuan, Z.G. (2006a) Effect of free ammonia and free nitrous acid concentration on the anabolic and catabolic processes of an enriched Nitrosomonas culture. Biotechnology and Bioengineering 95: 830-839. Vadivelu, V.M., Yuan, Z.G., Fux, C., and Keller, J. (2006b) The inhibitory effects of free nitrous acid on the energy generation and growth processes of an enriched Nitrobacter culture. Environmental Science & Technology 40: 4442-4448. Vadivelu, V.M., Keller, J., and Yuan, Z.G. (2007) Effect of free ammonia on the respiration and growth processes of an enriched Nitrobacter culture. Water Research 41: 826-834. 47 Chapter 2 c ) b ) a ) Figure 2.4. a) Original image of a sample of granules b) Image of the granules converted to black and white c) Area recognized by the software in red once the threshold levels are defined by the user. The average diameter obtained from the programme corresponded to the mean feret diameter of the granules. The feret diameter is calculated as an average value from the shortest and the longest measured segment in the granule (Fig. 2.5). a b Figure 2.5. Longest and shortest segments in a granule to estimate the feret diameter of a granule. 2.2.4. Specific Anammox activity assays The batch assays used to estimate the Anammox activity were performed according to the methodology described by Dapena-Mora et al. (2007). Completely closed vials with a total volume of 38 mL with 25 mL of liquid volume were used to perform the Anammox batch assays (Fig. 2.6). 1. Washing with phosphate buffer Ar 2. Flushing with Argon2. Flushing with Argon 3. Thermic stabilization 4. Addition of substrates 5. P (t) measurement Stirring 23.1 23.1 Figure 2.6. Experimental procedure to determine specific Anammox activity of the biomass. 54 Materials and methods Biomass concentration at the beginning of the experiment was fixed around 1.0 g VSS L-1. Before the beginning of the batch test the biomass was washed three times with phosphate buffer (0.143 g KH2PO4 L-1 and 0.747 g K2HPO4 L-1). The pH value was fixed at 7.8 and the temperature was fixed at a value T depending on the conditions to be analyzed. Gas and liquid phases were purged with argon gas to remove O2. The vials were placed in a thermostatic shaker, at 150 rpm and the temperature T until stable conditions were reached. Initial concentrations of substrates were 70 mg NH4+-N L-1 and 70 mg NO2--N L-1. The production of N2 was determined in the gas phase as the increment of pressure in the headspace of the vials, measured by means of a pressure transducer device. Maximum Specific Anammox Activity (SAA) was estimated from the maximum slope of the curve described by the cumulative N2 production along the time and related to the biomass concentration in the vials. The N2 gas production rate (moles N2 min-1) was calculated from the maximum slope of the curve describing the pressure increase in the vial along time () (atm min-1) (Eq. 2.5). T·R V α dt dN G2  (2.5) being VG the volume of the gaseous phase (L), R the ideal gas coefficient (atm L mol-1 K -1) and T the temperature (K). The SAA (g N2-N g VSS-1 d -1) is calculated from the N2 gas production rate and the biomass concentration in the vial (g VSS L-1): d min 1440 N mol N g 28 V·X dtdN SAA 2L 2  (2.6) being VL the volume of the liquid phase (L). Since the values of the affinity constant of the anammox bacteria for ammonium and nitrite are lower than 10 M and 5 M, respectively (Strous et al., 1999), it can be considered that the activity measured is the maximum activity for the range of nitrite and ammonium concentrations used. 2.2.5. Electron Microscopy and Micro-analysis Morphological studies of the biomass were performed with a scan electron microscope (Digital SEM Leica 440 at 20 kV) controlled with a computer system and with a magnification capacity ranging from 15 to 290000 folds. The sludge sample was washed three times for 10 minutes with phosphate buffer 0.05 N at pH 7.4 and subsequently fixed with a solution of glutaraldehyde 3% in phosphate buffer overnight. After fixation the sample was dehydrated using ethanol solutions with increasing ethanol concentrations (30, 50, 70 and 100%). Later the sample was shaded with gold and observed under the scan electron microscope. To investigate the elemental composition of the granules a micro-analysis was carried out. The instrument used was the SEM LEO-435VP with a system of micro-analysis (EDX) at voltages varying in the range of 5 kV, 20 kV and 30 kV. A typical count of the different atoms detected in a sample is represented in Fig. 2.7. 55 Chapter 2 Figure 2.7. Representation of the different atoms present in a sample of biomass detected in a micro-analysis. 2.3. Microbiological determinations 2.3.1. Identification of bacteria populations by FISH The abundance of the different populations of microorganisms present in the sludge samples of the reactors was researched by Fluorescent In Situ Hybridization (FISH). With this technique specific regions in 23S or 16S rRNA are detected with fluorescently labelled probes. If the corresponding domain, phylum, genus or species is present, the probe hybridizes to the targeted sequence and can later be detected microscopically. According to Amann et al. (1995) a typical FISH protocol includes four steps (Fig. 2.8): the fixation and permeabilization of the sample; hybridization of the targeted sequence to the probe; washing steps to remove unbound probe; and the detection of labelled cells by microscopy or flow cytometry. This protocol must be applied to disrupted biomass; therefore, the granules must be disintegrated before starting the procedure. To achieve the granular biomass breakage, biomass is sonicated for 1 min at 65% of amplitude using a probe sonicator (UP200s, Dr. Hielscher). Fixation of cells Immobilization of biomass Hybridization Total Cells (DAPI) Hybridized Cells (Cy3) Hybridized Cells (FLUOS) Microscopic analysis Bacteria Identification Figure 2.8. Different steps of the typical FISH protocol 56 Materials and methods During hybridization the cells are exposed to high temperatures, detergents and osmotic gradients. Thus fixation of the cells is essential in order to maintain the morphological integrity of the cells. Fixation of cells with glutaraldehide results in considerable autofluorescence of the specimen. Autofluorescence is minimized by fixation in freshly prepared (not older than 24 h) 4% paraformaldehyde solution in PBS. After fixation, the cells are immobilized on a microscopic slide and used for hybridization with 16S rDNA probes. In order to avoid non-specific binding of the rDNA probes, the hybridization is done at stringent conditions (46 ºC, 0-65% formamide) and specimens are washed with wash buffer (48 ºC). The targeted organisms can be detected by the characteristic fluorescence. The fluorochromes used to detect the hybridized rRNA were FLUOS (5(6)-carboxyfluorescein-Nhydroxysuccinimide ester) and Cy3 (indocarbocyanine). To visualize all cells in a sample the stain 4,6diamidino-2-phenylindole (DAPI) was used. Its application can provide insight into the existence of archaeobacteria and eukaryotes, like e.g. protozoa. For analysis of the slides an epifluorescence microscope (Axioskop 2 plus, Zeiss) in combination with a digital camera (Coolsnap, Roper Scientific Photometrics) was used. The phylogenetic tree reflecting the different probes applied in this study indicating the bacteria detected by each probe are shown in Fig. 2.9. The probes applied in this study are listed and detailed in Table 2.2. The three probes for the domain of eubacteria (EUB338, EUB338 II and EUB338 III) were applied together in all samples to get an impression of the relative abundance of the microorganisms detected by more specific probes. In comparison with DAPI they provided evidence of non-eubacteria present in the sample. For further discussion it has to be kept in mind that samples can never be 100% representative. Thus the fact that no bacteria of a certain kind were present in the sample can always be attributed to unrepresentative sampling as well. Still it was tried to keep this error small. Nitrosospira briensis (AOB) Nitrosovibrio tenuis (AOB) Nitrosolobus multiformis (AOB) Nitrosomonas europaea (AOB) Nitrosomonas eutropha (AOB) Nitrosococus mobilis (AOB) Nitrosomonas C56 (AOB) BACTERIA Proteobacteria Nitrospirae β-Proteobacteria Planctomyces α-Proteobacteria γ-Proteobacteria Nitrobacter ssp. (NOB) Nitrospira ssp. (NOB) Candidatus Anammoxoglobus propionicus (AB) Candidatus Jettenia asiatica (AB) Candidatus Scalindua sp. (AB) Candidatus Kuenenia stugartiensis (AB) Candidatus Brocadia anammoxidans (AB) Candidatus Brocadia fulgida (AB) EUB338 ALF1b BET42a GAM42a NIT3 NEU 653 NSM 156 NSV 443 NSO 190 NTSPA712 AMX 820 AMX 368 PLA46 Nitrococcus (NOB) Nitrosococcus halophilus (AOB) Nitrosococcus oceani (AOB) KST1275 BAN162 Figure 2.9. Different probes applied and the main bacteria detected by each probe (AOB: ammonium-oxidizing bacteria, NOB: nitrite-oxidizing bacteria and AB: Anammox bacteria). 57 Chapter 2 Table 2.2. Probes used for fluorescent in situ hybridisation and the formamide (FA) concentration used during hybridization. Probe Target site 16S Probe sequence (5’3’) % FA Target organisms Ref.a EUB 338 338-355 GCT GCC TCC CGT AGG AGT 0-50 Bacteria domain [1] EUB 338 II 338-355 GCA GCC ACC CGT AGG TGT 0-50 Planctomycetales [2] EUB 338 III 338-355 GCT GCC ACC CGT AGG TGT 0-50 Verrucomicrobiales [2] ALF1B 19-35 CGT TCG YTC TGA GCC AG 20 α-proteobacteria, some δ-proteobacteria, Spirochaetes [3] BET42a 1027-1043 GCC TTC CCA CTT CGT TT 35 β-proteobacteria [3] GAM42a 1027-1043 GCC TTC CCA CAT CGT TT 35 γ-proteobacteria [3] NEU653 653-670 CCC CTC TGC TGC ACT CTA 40 Most halophilic and halotolerant Nitrosomonas spp. [4] Competitor TTC CAT CCC CCT CTG CCG Nso190 189-207 CGA TCC CCT GCT TTT CTC C 55 Ammonio-oxidizing-- Proteobacteria [5] Nsm156 156-174 TAT TAG CAC ATC TTT CGA T 5 Nitrosomonas spp., Nitrosococcus mobilis [5] NIT3 1035-1052 CCT GTG CTC CAT GCT CCG 40 Nitrobacter spp. [6] Competitor CCT GTG CTC CAG GCT CCG Ntspa712 712-732 CGC CTT CGC CAC CGG CCT TCC 35 Most members of phylum Nitrospira [7] Competitor CGC CTT CGC CAC CGG GTT CC PLA46 46-63 GAC TTG CAT GCC TAA TCC 20 Planctomycetes [8] Amx820 820-841 AAA ACC CCT CTA CTT AGT GCC C 35 Candidatus “Brocadia anammoxidans” [9] Kst157 157-174 GTT CCG ATT GCT CGA AAC 25 Candidatus Kuenenia stuttgartiensis [9] Ban162 162-179 CGG TAG CCC CAA TTG CTT 40 Candidatus Brocadia anammoxidans [9] Pae997 997-1014 TCT GGA AAG TTC TCA GCA 0 Pseudomonas spp. [10] PAR1244 1244-1262 GGA TTA ACC CAC TGT CAC C 20 Paracoccus [11] a[1] Amann et al., (1990) [2] Daims et al., (1999) [3] Manz et al., (1992) [4] Wagner et al., (1995) [5] Mobarry et al., (1996) [6] Wagner et al., (1996) [7] Daims et al., (2001) [8] Neef et al., (1998) [9] Schmid et al., (2001). [10] Amann et al., (1996) [11] Neef et al., (1996). 2.3.2. Cryosectioning and confocal microscopy In order to determine the stratification of bacteria along the granule, some aggregates were frozen and cut in slices. Entire granules were embedded in OCT reagent (Tissue-Tek; Miles, Ind.) prior to their cryosectioning at -35 °C. Slides with a thickness ranging form 14 to 25 µm were cut at -16 °C, and these single sections were placed on the surface of poly-L-lysine coated microscopic slides. Hybridization was performed with the protocol aforementioned. The used probes for in situ hybridization were 5’ labelled with the fluorochromes FLUOS, Cy3 and Cy5 (Indodicarbocyanine). After in situ hybridization cells were stained with DAPI (0.5 µg mL-1) for 10 minutes. A TCS-SP2 confocal laser scanning microscope (Leica, Germany), equipped with HeNe laser for the detection of the fluorochromes Cy3 and Cy5 and Ar laser for the detection of the fluorochrome FLUOS, was used to analyze the sliced samples. 58 Materials and methods Quantification of the bacterial population was based on the procedure published by Crocetti et al., (2002). The quantification was performed by comparison of the positive area obtained with a probe with the area corresponding to the control: DAPI or EUBmix (a mixture of EUB338, EUB338 II and EUB338 III). The digital image analysis software Image ProPlus® was used to quantify the areas. When analyzing images with different probes, the software Leica Confocal Software Lite® was used to isolate each fluorochrome which corresponded to a specific probe. The area corresponding to the fluorescence of each FISH probe was obtained as the area of all pixels above a manually determined minimum pixel intensity to exclude empty spaces and background fluorescence (the sequence of analysis is represented in Fig. 2.10). a) f) e) c) b) d) g) Figure 2.10. a) Confocal image from a triple hybridization with the probes EUBmix (blue), NEU653 (green) and Amx820 (red). b) c) d) Images of each individual probe separated with the Leica Software. e) f) g) Images of the area detected for each probe by the Image Pro Plus software. 2.4. Microelectrodes By the introduction of few micrometer thick microsensors for analysis of a wide range of chemical species in microbial ecology, it became possible to study on a small scale the chemical gradients and 59 Chapter 2 metabolism in stratified microbial communities (e.g., Revsbech and Jorgensen, 1986). Microelectrodes are widely used to determine concentrations gradients of different compounds inside biofilms, core sediments, soils, etc. due to their small sizes. Due to the availability of very thin sensors, it is possible to perform experiments that cause minimal disturbance to the analyzed medium. Two different microelectrodes were used in this thesis: an electrochemical oxygen microsensor and a nitrite microbiosensor. 2.4.1. O2 microelectrode A Clark-type O2 microsensor (Fig. 2.11) equipped with a guard cathode was used for microscale analysis of DO; a detailed description of the functioning of this electrochemical microelectrode is given in Revsbech, (1989). The O2 microelectrode is an amperometric sensor; which detects the current caused by electrochemical reactions of the analyte at the sensor tip. A potential difference between the sensing electrode and the reference drives the reaction, and the current, measured by a sensitive picoammeter, is proportional to the analyte concentration (Schramm, 2003). These microsensors are characterized by absence of interferences, perfectly linear calibration curves, fast response times and low effects on the signal (<1–2%) from changes in stirring or diffusivity of the external medium. The sensors used in this study were constructed with tip diameters of 10 µm and they had 90% response time lower than 1 s. The very fast response time makes it possible to record detailed concentration profiles over distances of millimeters within a few minutes. The oxygen microsensor was calibrated at two points, the zero was established introducing the microelectrode in an anoxic alkaline ascorbic acid solution and the saturation was obtained both with an air saturated solution or with pure oxygen saturated solution depending on the DO concentration to be analyzed. The O2 microelectrodes used in this work were built by the technical assistant Preben Sorensen, a member of the microbiology department of Aarhus University (Denmark). a) b) Figure 2.11. a) Clark-type O2 microelectrode (adapted from Revsbech, 1989) b) Schematic of the O2 measuring principle in the microelectrode (from Kuhl and Revsbech, 2001). 60 Materials and methods 2.4.2. NO2microelectrode For some sensors a biological reaction precedes the electrochemical reaction, and these sensors are labeled as microbiosensors. The NO2microbiosensor was based on a combination of biological reactions catalyzed by bacteria and an electrochemical N2O microsensor (Fig. 2.12). The NO2microbiosensor (a detailed description is available in Nielsen et al., 2004) consisted of an immobilized pure culture of Stenotrophomonas nitritireducens, which specifically reduced NO2to N2O (they lack the enzyme N2O reductase), coupled to a Clark type N2O microsensor (Andersen et al., 2001). A medium reservoir supplied the culture with all essential growth constituents except the electron acceptor. NO2diffused freely through the ionpermeable membrane into the bacterial biomass of the reaction chamber, where it is denitrified to N2O and subsequently reduced to N2 at the cathode of the N2O sensor. By placing the bacteria in a capillary in front of an electrochemical N2O microsensor, the signal of the N2O electrode was proportional to the concentration of NO2-. The 90% response time was 30-50 s, slower than the O2 microsensor but also making possible to perform concentration profiles over distances of millimeters within a few hours. Figure 2.12. NO2microbiosensor (Kuhl and Revsbech, 2001). The microelectrodes were calibrated before every profile determination. The NO2microbiosensors were calibrated in a vial containing NO2--free water with approximately the same ionic strength and temperature as the environment where the measurements were to be performed. A calibration curve was obtained by plotting the current values from the N2O electrode against various NO2concentrations obtained by repeated addition of NO2from a concentrated stock solution (Fig. 2.13). Tungstate was added inside the biosensor in order to measure NO2concentrations higher than 7 mg N L-1. The tungstate increased the stability of the microsensor and slowed down the nitrite consumption rate of the bacteria (Nielsen et al., 2005). The lifetimes of microscale sensors containing immobilized bacteria are from days to weeks. The NO2microelectrodes used in this work were built by the Professor N.P. Revsbech, member of the microbiology department of Aarhus University (Denmark). 61 Chapter 2 y = 0.036x - 0.337 R 2 = 1.000 0 1 2 3 4 0 50 100 150 Signal (pA) C NO2 (mg N L -1 ) a) y = 0.472x - 1.919 R 2 = 0.999 0 10 20 30 40 050 Signal (pA) C NO2 (mg N L -1 ) b) 100 Figure 2.13. Calibration curves of two NO2microelectrode a) without addition of tungstate b) with addition of tungstate at different signal intensities. 2.4.3. Experimental setup A setup for measuring O2 and NO2in the granules is shown schematically in Fig. 2.14. The microelectrode is held by a micromanipulator, which is used to introduce the microelectrode tip into the granule. The current in the circuit is measured by a very sensitive ammeter with a range down to 1 pA (10-12 A). The amplified signal is then recorded on a strip-chart recorder. The granules were fixed on a mesh grid by a needle at the bottom of the aerated cell. Granules were kept for 1 h inside the cell as a pre-incubation period to create pseudo steady state conditions, the air was provided by an aeration pump. Concentration profiles were recorded by introducing the sensors into the granules at different depth positions using a manual micromanipulator. A dissection microscope was used to visually estimate the position of the granule/water interface by visual observation. For each granule and experimental condition tested several microprofiles were measured. Micromanipulator Aeration pump Pico ammeter + - 0.75 V 0 mg N L -1 1 2 3 4 Voltage source Microelectrode Reference electrode Recorder Cell with granule Figure 2.14. Circuit used for the obtaining of profile measurements with microelectrodes (Adapted from Revsbech and Jorgensen, 1986). 62 Materials and methods 2.5. References Amann, R. I., B. J. Binder, R. J. Olson, S. W. Chisholm, R. Devereux, and D. A. Stahl. (1990). Combination of 16S rRNAtargeted oligonucleotide probes with flowcytometry for analyzing mixed microbial populations. Applied and Environmental Microbiology 56: 1919–1925. Amann, R., Ludwig, W., and Schleifer, K.H. (1995) Phylogenetic identification and in-situ detection of individual microbialcells without cultivation. Microbiological Reviews 59: 143-169. Amann, R., Ludwig, W., Schulze, R., Spring, S., Moore, E., and Schleifer, K.H. (1996) rRNA-targeted oligonucleotide probes for the identification of genuine and former pseudomonads. Systematic and Applied Microbiology 19: 501-509. Andersen, K., Kjaer, T., and Revsbech, N.P. (2001) An oxygen insensitive microsensor for nitrous oxide. Sensors and Actuators B-Chemical 81: 42-48. APHA-AWWA-WPCF (1998) Standard methods for the examination of water and wastewater. Washington DC, USA: American Public Health Association/American Water Works Association/Water Environment Federation. Beun J.J., van Loosdrecht M.C.M. and Heijnen J.J. (2002) Aerobic granulation in a sequencing batch airlift reactor. Water Research 36: 702-712. Crocetti, G.R., Banfield, J.F., Keller, J., Bond, P.L., and Blackall, L.L. (2002) Glycogen-accumulating organisms in laboratoryscale and full-scale wastewater treatment processes. Microbiology 148: 3353-3364. Daims H., Brühl A., Amann R., Schleifer K.-H. and Wagner M. (1999) The domain-specific probe EUB338 is insufficient for the detection of all Bacteria: Development and evaluation of a more comprehensive probe set. Systematic and Applied Microbiology 22: 434-444. Daims H., Nielsen J. L., Nielsen P. H., Schleifer K. H. and Wagner M. (2001) In Situ characterization of Nitrospira-like nitriteoxidizing bacteria active in wastewater treatment plants. Applied and Environmental Microbiology 67: 5273-5284. Dapena-Mora, A., Fernandez, I., Campos, J.L., Mosquera-Corral, A., Mendez, R., and Jetten, M.S.M. (2007) Evaluation of activity and inhibition effects on Anammox process by batch tests based on the nitrogen gas production. Enzyme and Microbial Technology 40: 859-865. De Kreuk, M.K., McSwain, B.S., Bathe, S., Tay, S.T.L., Schwarzenbeck, N. and Wilderer, P.A. (2005) Discussion outcomes. In Aerobic granular sludge. Bathe, S., de Kreuk, M., Mc Swain, B.S., and Schwarzenbeck, N. (eds). London: IWA, pp. 165-169. Kuhl, M., and Revsbech, N.P. (2001) Biogeochemical microsensors for boundary layer studies. In The Benthic Bounday Layer. Boudreau, B.P., and Jorgensen, B.B. (eds). New York: Oxford University Press, pp. 180-210. Manz W., Amann R., Ludwig W., Wagner M. and Schleifer K.-H. (1992) Phylogenetic oligodeoxynucleotide probes for the major subclasses of Proteobacteria: problems and solutions. Systematic and Applied Microbiology 15: 593 - 600. Mobarry, B., M. Wagner, V. Urbain, B. Rittmann, and D. Stahl. (1996) Phylogenetic probes for analyzing abundance and spatial organization of nitrifying bacteria. Applied and Environmental Microbiology 62: 2156–2162. Neef, A., Zaglauer, A., Meier, H., Amann, R., Lemmer, H., and Schleifer, K.H. (1996) Population analysis in a denitrifying sand filter: conventional and in situ identification of Paracoccus spp. in methanol-fed biofilms. Applied and Environmental Microbiology 62: 4329-4339. Neef A., Amann R., Schlesner H. and Schleifer K.-H. (1998) Monitoring a widespread bacterial group: in situ detection of planctomycetes with 16S rRNA-targeted probes. Microbiology 144: 3257-3266. Nielsen, M., Larsen, L.H., Jetten, M.S.M., and Revsbech, N.P. (2004) Bacterium-based NO2biosensor for environmental applications. Applied and Environmental Microbiology 70: 6551-6558. Nielsen, M., Bollmann, A., Sliekers, O., Jetten, M., Schmid, M., Strous, M., Schmidt, I., Larsen, L.H., Nielsen, L.P., and Revsbech, N.P. (2005) Kinetics, diffusional limitation and microscale distribution of chemistry and organisms in a CANON reactor. FEMS Microbiology Ecology 51: 247-256. Revsbech, N.P. (1989) An Oxygen Microsensor with a Guard Cathode. Limnology and Oceanography 34: 474-478. Revsbech, N.P., and Jorgensen, B.B. (1986) Microelectrodes - Their Use in Microbial Ecology. Advances in Microbial Ecology 9: 293-352. 63 Chapter 3 Table 3.2. Targeted organisms and the corresponding formamide (FA) percentages for the used oligonucleotide probes. Probea Probe sequence (5’3’) FA (%) Targeted organisms EUB338 GCT GCC TCC CGT AGG AGT 0-50 Bacteria domain EUB338 II GCA GCC ACC CGT AGG TGT 0-50 Planctomycetales EUB338 III GCT GCC ACC CGT AGG TGT 0-50 Verrucomicrobiales ALF1b CGT TCG Y(C/T)TC TGA GCC AG 20 α-proteobacteria, some δ-proteobacteria, spirochaetes BET42ab GCC TTC CCA CTT CGT TT 35 - Proteobacteria GAM42a GCC TTC CCA CAT CGT TT 35 - Proteobacteria Nso190 CGA TCC CCT GCT TTT CTC C 55 Ammonium-oxidizing - Proteobacteria NEU653b CCC CTC TGC TGC ACT CTA 40 Most halophilic and halotolerant Nitrosomonas spp. Ntspa712 CGC CTT CGC CAC CGG CCT TCC 50 most members of the phylum Nitrospirae NIT3b CCT GTG CTC CAT GCT CCG 40 Nitrobacter spp. aDetails on oligonucleotide probes are available at probeBase (Loy et al., 2007). bUsed with an equimolar amount of corresponding unlabelled competitor oligonucleotide probe. 3.3.4. Biomass activity measurements Measurements of the nitrogen compounds in the liquid phase were performed during single operational cycles (Stage III, Cycle B) fixing the DO concentration at the selected values of: 2, 4, 16, 22 and 30 mg O2 L-1, respectively. These DO concentrations were reached for values above 8 mg O2 L-1 by flushing pure O2 or for values under the cited value by flushing N2 and proportionally reducing the air flow in order to maintain the total gas flow constant . The AOR was estimated calculating the ammonium consumption as the difference between the theoretical value of ammonium concentration if no biological activity would take place and the value measured experimentally (Mosquera-Corral et al., 2005c). The volume variation was taken into account since the feeding process took place during the whole aeration period. The nitrite oxidizing rate (NOR) was estimated using a similar procedure but referred to the nitrate formation. 3.3.5. Calculations 3.3.5.1. Free ammonia and free nitrous acid The concentrations of NH3 and HNO2 were calculated at the operational temperature from the NH4+, NO2concentrations and the pH in the bulk liquid using Eq. 3.1 and 3.2 according to the expressions proposed by Anthonisen et al., (1976):               1 10 e C C pH 273T 6344 NH NH 4 3 (3.1)             273T 2300 pH NO HNO e10 C C2 2 (3.2) 70 Nitrifying granular systems: a suitable technology to obtain stable partial nitrification at room temperature 3.3.5.2. Maximal oxygen and ammonium fluxes through the interface liquid-granule The oxygen was transferred from the air to the granules firstly through the gas-liquid interface and then through the liquid-granule interface. Since the dissolved oxygen concentration in the bulk liquid is determined experimentally the oxygen flux through the surface of the granules can be calculated according to Eq. 3.3.  sLcO CCAkJ 2   (3.3) being JO2 the oxygen flux (g O2 d-1); kc the mass transfer coefficient (m d-1), A the total surface of the granules (m2); CL the DO concentration in the bulk liquid (g O2 m-3) and CS the DO concentration at the surface of the granule. In order to estimate the value of kc, an empirical relation, the Sherwood number (Sh) (Eq. 3.4), obtained for spherical particles and valid for Reynolds number lower than 3·105, was used (Hooijmans et al., 1990; Garrido et al., 1997).       1.2 1.7 31 3 61 O mc Pe1 Pe Sc0.861 0.66 2 D D k Sh 2   (3.4) being Sc and Pe the dimensionless modules of Schmidt and Peclet calculated according to Eq. 3.5 and 3.6, respectively. 2 O D ν Sc  (3.5) 2 O mg D D v ScRePe  (3.6) being ν the kinematic viscosity: 10-6 m 2 s -1, DO2 the oxygen diffusion coefficient in water 1.7·10-9 m 2 s -1 (Picioreanu et al., 1997), vg was estimated as the settling velocity of the granules and Dm the weighted mean diameter of the granules obtained from a sample of n particles (n ranged between 300 and 400 granules). This mean diameter was obtained by means a weighted mean volume according to Eq. 3.7.   n 2D 2D n 1i 3 i 3 m    (3.7) being Di the measured diameter of each granule of the sample. The specific surface of the granules a (m2 m-3) was obtained as the total surface of a sample of “n” granules divided by their total volume (Eq. 3.8).       n 1i 3 i n 1i 2 i 2Dπ 3 4 2Dπ4 a (3.8) The total surface of the granules A (m2) was calculated by multiplying the total volume of the granules in the reactor by the specific surface (Eq. 3.9). RR granule VX Aa ρ  (3.9) being VR the reactor volume (m3), XR the granules concentrations (g VSS L-1) and ρgranule the density of the granules (g VSS (Lgranules)-1). The maximal flux of oxygen, JO2max (g O2 d -1), was obtained considering the dissolved oxygen concentration at the granule surface as zero according to Eq. 3.10. 71 Chapter 3 Lc max OCA kJ 2 (3.10) The maximal flux of ammonium towards the granules surface, JNmax (g N d-1), was estimated according to Eq. 3.11.  R NH4 max N VC JHRT i  (3.11) being (CNH4)i the ammonium concentration in the influent (g N L-1) and HRT the hydraulic retention time (d). 3.4. Results and discussion 3.4.1. Reactor operation Previous to this work the SBR was operated during two years in different operational conditions (data not shown, Mosquera-Corral et al., 2005a). Afterwards it was operated during 880 days in six different operational stages as indicated in Table 3.1. 0.0 0.1 0.2 0.3 0.4 0 100 200 300 400 500 600 700 800 900 Time (d) NH 4+ , NO 2- , NO 3- (g N L -1 ) a) III III IV VVI 0.0 0.4 0.8 1.2 1.6 0 100 200 300 400 500 600 700 800 900 Time (d) NLR, AOR, NOR (g N L -1 d -1 ) b) III III IV VVI Figure 3.3. a) Nitrifying granular SBR operation in terms of NH4+-N concentration in the feeding (─), and NH4+- N (○), NO2--N (●) and NO3--N (─) concentrations in the reactor. b) Nitrogen Loading Rate (NLR) (─) Ammonium Oxidizing Rate (AOR) (○) and Nitrite Oxidizing Rate (NOR) (─) along the operational period. 72 Nitrifying granular systems: a suitable technology to obtain stable partial nitrification at room temperature During Stage I (Fig. 3.3a) the inlet ammonium concentration was of 0.1 g NH4+-N L-1 and total nitrification to nitrate was reached representing a nitrate production rate of 0.4 g NO3--N L-1 d-1 (Fig. 3.3b). On day 62 the inlet ammonium concentration was doubled to 0.2 g NH4+-N L-1 (Stage II). From this date on, nitrite was always present in the effluent at concentrations up to 0.1 g NO2--N L-1 meaning that the NOR was kept around 0.4 g NO3--N L-1 d-1 while the ammonium was almost fully depleted. During these two stages, since the feeding phase took place during 3 minutes (Cycle A), the ammonium and subsequently the estimated free ammonia concentrations were maximal during the first minutes of the cycle and concentrations as high as 10 mg NH3-N L-1 were reached which have been reported as inhibitory for the nitrite oxidation process (Anthonisen et al., 1976). In order to determine the evolution of the nitrogen compounds several cycle measurements were performed corresponding to operational stages I and II (Fig. 3.4a and 3.4b). Results indicated that only when the ammonium concentration in the reactor felt to values close to zero and the free ammonia concentration reached a value close to 0.5 mg NH3-N L-1 the amount of nitrite accumulated experienced a decrease. Depending on the applied nitrogen load, the concentration of free ammonia and the amount of nitrite accumulated were different. 0.00 0.03 0.06 0.09 0.12 0.15 0.18 0 30 60 90 120 150 180 Time (min) NH 4 + , NO 2 - , NO 3 - (g N L -1 ) 0.000 0.002 0.003 0.005 0.006 0.008 0.009 NH 3 (g N L -1 ) a) 0.00 0.03 0.06 0.09 0.12 0.15 0.18 0 30 60 90 120 150 180 Time (min) NH 4 + , NO 2 - , NO 3 - (g N L -1 ) 0.000 0.002 0.003 0.005 0.006 0.008 0.009 NH 3 (g N L -1 ) b) Figure 3.4. Concentrations of NH4+ (○), NO2- (▲), NO3- () and NH3 () during two cycle measurements (g N L-1) a) on stage I, day 52 and b) on stage II, day 77. To diminish the free ammonia concentrations during the first minutes of the cycle, the feeding period was extended, from day 143 on (Stage III), to 171 minutes coinciding with the aeration phase (Cycle B). The concentrations of substrates in the bulk liquid remained practically constant during the entire cycle due to the 73 Chapter 3 continuous feeding pattern performed during the reaction phase. Unexpectedly, at this stage ammonium and nitrite were measured in the effluent at concentrations around 0.05 g N L-1 each, while nitrate concentrations remained around 0.095 g NO3--N L-1. After day 160 the system evolved in such a way that ammonium and nitrate concentrations decreased while nitrite concentrations increased (Fig 3.3a). Between days 250 and 390, a stable operational period was achieved with average values in the effluent of 0.02 g NH4+-N L-1, 0.13 g NO2-- N L-1 and 0.04 g NO3--N L-1. Stable nitrite accumulation was obtained although the bulk dissolved oxygen concentration was 8 mg O2 L-1. On day 400, granules coming from another granular SBR (where complete nitrification to nitrate occurred) were added to the SBR. The amount of inoculated granules represented the 25% in dry weight of the total biomass contained in the reactor before their addition. Nitrate concentration immediately increased and after 20 days the NOR was of 0.7 g NO3--N L-1 d-1. The composition of the effluent was 0.01 g NH4+-N L-1, 0.01 g NO2--N L-1 and 0.17 g NO3--N L-1. This total nitrification remained stable until day 460. From this day on, the nitrite concentration began to rise and 20 days later the system recovered the previous state with nitrite accumulations over 0.1 g NO2--N L-1. These results confirmed that, in this system, full oxidation of ammonium into nitrate could not be maintained at a NLR of 0.8 g NH4+-N L-1 d-1 under stable conditions for long periods of time. As it was observed that nitrite was easily accumulated in the system, even at high DO concentrations in the bulk liquid, the achievement of partial nitrification limiting the ammonium oxidation to 50% was the following objective. In order to obtain partial nitrification the DO was lowered to a medium value of 2.7 mg O2 L-1 during stages V and VI. The decrease in the DO concentration avoided the activity of the NOB and lowered the activity of the AOB. Partial nitrification was achieved and an effluent with a molar NO2/NH4 ratio of 1.0 ± 0.3 was obtained during stages V and VI when applied NLR ranged from 0.8 to 1.6 g N L-1 d-1. During stage VI the ammonium concentration in the influent was stepwise increased up to 0.4 g NH4+-N L-1. The mean temperature increased from 20 ºC to 23 ºC, coinciding with the summer station. The applied NLR was increased up to 1.6 g N L-1 d-1 and a molar NO2/NH4 ratio of 1.4 ± 0.3 was obtained. 3.4.2. Biomass physical properties The biomass concentration at the beginning of the work was of 1.4 g VSS L-1 and it steadily increased to values around 6 g VSS L-1 at the end of the experiment (Fig. 3.5a). An increase of 25% of the biomass concentration was registered on period IV due to the addition of biomass from another granular SBR (indicated with the arrow in Fig. 3.5a). The percentage of ISS in the suspended solids was of 10% during the whole operational time. The concentration of solids in the effluent ranged from values as low as 0.005 g VSS L-1 to values up to 0.05 g VSS L-1. During the discontinuous operational stages the estimated sludge retention time (SRT) value remained below 25 d (Stages I and II) and increased up to 190 d during Stage IV, remaining over 100 d during period V and decreasing again to values around 60 d on period VI. These large SRT values were possible due to the good settling properties of the granules which guarantee the appropriated solids retention conditions to keep nitrifying bacteria inside the reactor. 74 Nitrifying granular systems: a suitable technology to obtain stable partial nitrification at room temperature 0 2 4 6 8 10 12 0 100 200 300 400 500 600 700 800 900 Time (d) Reactor (g VSS L -1 ) 0.00 0.01 0.02 0.03 0.04 0.05 0.06 Effluent (g VSS L -1 ) a) III III I V VVI 0 30 60 90 120 150 0 100 200 300 400 500 600 700 800 900 Time (d) Density (g VSS (L granule ) -1 ) 0 20 40 60 80 100 SVI 5 (mL (g VSS) -1 ) b) III III IV VVI Figure 3.5. a) Concentrations of volatile suspended solids in the reactor (●) and effluent (○) (g VSS L-1) of the SBR, and b) density (●) (g VSS (Lgranules)-1) and sludge volumetric index (SVI5) (○) (mL (g VSS)-1) of the granules. Regarding the settling properties of the biomass the SVI5 and the density of the granules experienced opposite evolutions. The SVI5 values around 100 mL (g VSS)-1 and densities around 30 g VSS (Lgranule)-1 were measured during stage I. The trend until stage IV was a significant improvement in the settling properties since the values of density of the granules reached values higher than 80 g VSS (Lgranules)-1 whereas the SVI5 decreased to values under 50 mL (g VSS)-1 (Fig. 3.5b). These changes could be attributed to the increase in the concentration of AOB, which are known to form dense biofilms, after the two fold increase of the applied NLR. A slight decrease in the density together with a slight increase in the SVI5 were registered during stages V and VI probably caused by the decrease in the DO concentration which affected the physical properties of the granules. Regarding the formed granules no significant differences in their sizes distributions measured as percentages of total granules surface were observed during the 880 days of operation (Fig. 3.6). The main contributors to the total surface of the granules were those with diameters between 3 and 5 mm. The mean diameter of the granules ranged during the whole operational period being between 1.9 and 3.0 mm. The 75 Chapter 3 obtained granules presented large diameters which caused important limitations of oxygen due to the slight penetration of oxygen inside the granules. a) b) c) 0 5 10 15 20 25 0123456 Diameter (mm) % by volume Figure 3.6. Images of the granules (zoom: 6.5x) in a) Stage I and b) Stage VI; and c) corresponding distribution of sizes in Stages I (■) and VI (□). 3.4.3. Identification of nitrifying populations In order to gain insight in the distribution of nitrifying populations inside the granules, the FISH technique was used. Specific probes for nitrifying populations present in disaggregated and sliced granules were applied to biomass samples (Table 3.2). By application of a set of general (ALF1b, BET42a, GAM42a and EUBmix which is a mixture of EUB338, EUB338 II and EUB338 III) and specific probes (NSO190 and NEU653) to two samples of disaggregated granules collected in stages I and II, respectively, it was concluded that the genus Nitrosomonas was the dominant AOB present. The genera Nitrobacter (NIT3) and Nitrospira (Ntspa712) were the NOB detected, but in low concentrations in both cases. On day 768 (stage VI) a sample of granules was collected and sliced. The NEU653, NIT3 and Ntspa712 probes were applied to these sliced granules and the results indicated that Nitrosomonas were located in the outermost layers of the granule and these ones were the dominant specie. Nitrobacter and Nitrospira were present but in very low amounts widespread in the outer layer of the granule (Fig. 3.7). The width of the external zone where the AOB population was present corresponded approximately to 100 µm. 76 Nitrifying granular systems: a suitable technology to obtain stable partial nitrification at room temperature Figure 3.7. In situ hybridization of a cross-section of a granule, bright zones represent fluorescent signal of a combination of probes NEU653 and EUBmix. 3.4.4. Physical stability of the granules During the 880 days of operation the granules kept their physical integrity and no breakage events were observed in spite of the changes in the feeding strategy. The characteristics of the granules formed were comparable to those obtained with aerobic granules grown on carbon sources. Averaging 20 different reports on aerobic granules, the mean values of SVI, density, and diameter were: 64 ± 30 mL (g VSS)-1; 44 ± 20 g VSS (Lgranule)-1 and 1.7 ± 1.1 mm, respectively (de Kreuk et al., 2005). Therefore, the long term operation of granules treating an autotrophic medium and operating at DO concentrations in the range 2 – 4 mg O2 L-1 allowed maintaining the large diameter, high density and excellent settling properties of the granules. In the literature, nitrifying aggregates grown without carrier material have already been reported: Campos et al., (2000), accumulated 12 g VSS L-1 of nitrifying aggregates carrying out complete nitrification to nitrate with a SVI of 17 mL (g VSS)-1 and a density of 100 g VSS (Lbiomass)-1; Kim and Seo, (2006) accumulated 3.5 g VSS L-1 of AOB with a SVI of 40 mL (g VSS)-1 in a Sequencing Batch Airlift Reactor. However, in both cases, the mean diameter of the particles obtained was lower than 0.5 mm. By working with granules with high diameter the settling velocities of the granules is increased. Jin et al., (2008) obtained nitrifying granules with a diameter of 1.54 mm and settling velocities of 82 m h-1, values significantly higher than those of activated sludge flocs (lower than 9 m h-1, (Campos et al., 2000)) but lower than those of the granules from the present study which reached values up to 150 m h-1. 3.4.5. Partial nitrification The occurrence of nitrite accumulation was obtained in the granular SBR after doubling ammonium load and applying a NLR of 0.8 g N L-1 d-1. No evolution was observed in the reactor to cope with this load increase and to produce complete nitrification. As it was previously indicated, several factors are involved in the uncoupled performance of the two steps of the nitrification, being the main ones: temperature, DO concentration and inhibition by NH3 and HNO2. 77 Chapter 3 The nitrite accumulation registered in the system was obtained at temperatures around 20 ºC. The maximal growth rate of AOB and NOB is differently affected by the temperature due to the difference in the activation energy of both catabolic reactions. According to Wiesmann, (1994) the value of the µmax corresponding to the AOB is of 0.76 and 1.97 d-1 at 20 ºC and 30 ºC, respectively; whereas the µmax of the NOB is of 1.04 and 1.88 d-1 at 20 ºC and 30ºC, respectively. By working at temperatures higher than 30 ºC, the AOB have higher duplication velocities and therefore they outcompete the NOB by regulating the SRT and guarantee the occurrence of the partial nitrification. This is the operational strategy applied for example in the Sharon process (van Dongen et al., 2001). The use of SBR systems represents also an alternative to obtain partial nitrification at high temperature and it has been successfully applied, i.e., by Ganigue et al., (2007) at 36 ºC. However, partial nitrification is not only restricted to temperatures above 30 ºC and efforts have been made in order to establish partial nitrification at temperatures lower than 30 ºC. When operating at 20 ºC the AOB present maximal growth rates smaller than those of the NOB, meaning that the latter would be in advantage. In order to explain why AOB outcompete NOB in the granular SBR the possible effects of DO and the HNO2, NH3 inhibitions were analyzed. 3.4.6. Effects of DO In the granular SBR the effect oxygen limiting conditions on the growth of the NOB is expected to cause the observed stable nitrite accumulation. By working with biofilms, the importance of the biofilm surface to volume of bulk liquid ratio was highlighted as one of the critical factors in the oxygen transport from the gas phase to the granule surface. The oxygen penetration depth in biofilms varies typically in a range from 75 to 200 µm and therefore it is important to maximize the surface area of the biofilm or granule to maximize the reactor capacity (van Loosdrecht and Heijnen, 1993). During stage III, the DO concentration in the bulk liquid was varied and monitored during several cycles in order to determine the effect of the DO on the biomass activity. By working at DO concentrations from 4 to 30 mg O2 L-1 total oxidation of the ammonium fed was achieved; however, the AOB activity decreased 75% when working at DO concentration of 2 mg O2 L-1 (Fig. 3.8). The NOB activity increased considerably when increasing DO concentrations up to 8 mg O2 L-1. Maximal NOB activity was not reached until attaining DO concentrations of 22 mg O2 L-1. These batch experiments clearly demonstrated the difference in the oxygen affinity constant between AOB and NOB and pointed out the role of the mass transfer limitations involved when working with granules with a so large diameter. In the present work, the ratio between the total surface of the granules and the volume of the bulk liquid ranged between 58 and 216 m2 m-3 during the reactor operation. These values are much lower than those obtained in particle based biofilm reactors of around 3000 m2 m-3 (Garrido et al., 1997). It is therefore expected that the oxygen mass transport from the bulk liquid to the granule limited the oxidation rates. In order to evaluate the maximal flux of oxygen that could be transferred from the bulk liquid to the granule, JO2max was calculated as the maximal flux that could be transferred through the liquid-granule interface (i.e., the flux of oxygen when the DO concentration at the surface of the granule was zero) (Eq. 3.10). The values of the mass transfer coefficient (kc) for oxygen and the Sh module, obtained applying Eq. 3.4, 3.5 and 3.6, were 4·10-5 m s-1 and 60, respectively. The maximal fluxes of oxygen and nitrogen that can be transferred to the granule were calculated from Eq. 3.7, 3.8, 3.9, 3.10 and 3.11. 78 Nitrifying granular systems: a suitable technology to obtain stable partial nitrification at room temperature 0.0 0.2 0.4 0.6 0.8 1.0 248162230 DO (mg O 2 L -1 ) AOR, NOR (g N L -1 d -1 ) Figure 3.8. AOR (□), NOR (■) obtained at different DO in the bulk liquid of the granular SBR. The values of the JO2 (JN)-1 ratio are presented in Fig. 3.9. The value of JO2 (JN)-1 ratio of 4.5 g O2 (g N)-1 corresponded to the stoichiometric value required to obtain complete nitrification which was only reached in stage I. There was a correlation between the percentage of nitrite accumulated and the value of the JO2 (JN)-1 ratio meaning that partial nitrification occurred mainly due to oxygen mass transfer limitations. Only during stages I and IV this ratio was close to the stoichiometric one necessary to obtain complete nitrification as reflected in the low percentages of nitrite accumulated. The duplication of the NLR applied caused a decrease in the JO2 (JN)-1 ratio causing nitrite accumulation. With the introduction of granules from another SBR (Stage IV) the available surface augmented with the consequent increase of the JO2 to values close to the stoichiometric ratio allowing significant nitrate production while almost no nitrite was present. During stages V and VI, the low oxygen flow available due to the low DO concentration in the bulk liquid allowed reaching stable partial nitrification. 0.0 1.5 3.0 4.5 6.0 0 100 200 300 400 500 600 700 800 900 Time (d) J O 2 (J N) -1 0 25 50 75 100 NO 2- (%) I II III IV VVI Figure 3.9. Evolution of the JO2 (JN)-1 ratio (●) and nitrite percentage in the effluent (□) along the operational period. 79 Chapter 4 Population dynamics of nitrite oxidizers in nitrifying granules1 Summary The competition between Nitrospira and Nitrobacter species was analyzed in this work under conditions of excess of nitrite. A population of nitrite oxidizing bacteria (NOB) was developed from nitrifying biomass grown as granules with a mean diameter of 0.8 mm, by means of switching the composition of the feeding media from ammonium to nitrite. The initial population distribution of the granules was: 60% Nitrosomonas and 30% Nitrospira and it evolved to 45% Nitrobacter and 40% Nitrospira measured 177 days after the change in the feeding. The disappearance of Nitrosomonas allowed the development of an important population of Nitrobacter demonstrating that these microorganisms, characterized by being r-strategists NOB, are poor competitors when oxygen is the limiting substrate. Interestingly, the physical structure of the granules was not altered by the change of its microbial composition during the 220 days of operation. 1Vázquez-Padín J.R., Figueroa M., Campos J.L., Mosquera-Corral A. and Méndez R. (2009) Population dynamics of nitrite oxidizers in nitrifying granules. Water Science and Technology 60(10): 2529-2536. Chapter 4 4.1. Introduction Due to the slow growth rate of the nitrifying bacteria populations, nitrification is the controlling process during the biological nitrogen removal from wastewaters. Large efforts have been focused on the development of strategies to keep large nitrifying biomass concentrations inside the reaction systems by promoting the formation of biofilms or granules (Vázquez-Padín et al., 2009). In these systems both external and internal mass transfer phenomena are important and can determine the overall reaction rate (Wilen et al., 2004). Significant gradients of substrates around and inside the aggregates occur which may be responsible for the stratification of the spatial distribution and diversity of bacterial populations. In nitrifying microbial aggregates, ammonium-oxidizing bacteria (AOB) are generally found throughout the whole aggregate while nitrite-oxidizing bacteria (NOB) are placed in the inner zones of the biofilms (Okabe et al., 1999; Vázquez-Padín et al., 2009). In this kind of systems Nitrosomonas are generally the dominant AOB (Schramm et al., 2000; Han et al., 2003). However, both Nitrospira and Nitrobacter are simultaneously detected but no clear reason for that has been identified (Daims et al., 2001). Traditionally, Nitrobacter was considered to be the most important nitrite oxidizer in WWTP. The development of microbiological tools revealed that Nitrospira-like bacteria are widely distributed in natural and engineered ecosystems (Burrell et al., 1998; Hovanec et al., 1998; Okabe et al., 1999; Schramm et al., 1999; Daims et al., 2001; Regan et al., 2002; Daims et al., 2006). This predominance of Nitrospira over Nitrobacter in most WWTP could be a reflection of their different survival strategies. Schramm et al., (1999) postulated that Nitrospira-like bacteria are Kstrategists and can exploit low amounts of nitrite and oxygen much more efficiently than Nitrobacter. In contrast, Nitrobacter species are r-strategists that can grow faster than Nitrospira, but depend on significantly higher nitrite and oxygen concentrations. This K/r-hypothesis could explain the predominance of Nitrospira-like bacteria in activated sludge, where nitrite concentrations are usually low. Under these conditions the growth rate of Nitrospira-like bacteria is sufficiently high to maintain stable populations, while Nitrobacter cannot proliferate fast enough and are washed out of the continuously operated bioreactors. This fact has been already observed by Nogueira and Melo, (2006) in nitrite oxidizing chemostats and by Kim and Kim, (2006) in biofilm systems. Nevertheless other authors indicated that Nitrospira and not Nitrobacter can be the predominant populations of nitrite oxidizers in both biofilms (Okabe et al., 1999; Han et al., 2003) and granular nitrifying systems (Carvalho et al., 2006; Wang et al., 2007) even when high nitrite concentrations are present. This fact would indicate that the oxygen concentration rather than the nitrite concentration could be the key factor for the selection of nitrite oxidizing populations (Schramm et al., 2000). 4.2. Objectives ● The objective of this work was to determine the effect of oxygen availability on the distribution of nitrite oxidizing populations in a nitrifying reactor further operated as a nitrite oxidizing reactor. The performance of the system and the evolution of both physical characteristics and bacterial populations of nitrifying granules were studied. 4.3. Materials and methods 4.3.1. Reactor description A sequencing batch reactor (SBR) with a working volume of 2.5 L, an internal diameter of 0.1 m and a height of 0.27 m was used, the height to the diameter ratio was of 2.7. The exchange volume was fixed at 20%. The hydraulic retention time (HRT) was kept at 1.25 d. Air was supplied from the bottom of the reactor through 88 Population dynamics of nitrite oxidizers in nitrifying granules a sparger. The duration of the operational cycles was of 6 h distributed in: aeration and feeding (345 min); settling (10 min) and withdrawal (5 min). A programmable logic controller Siemens model S7-224CPU controlled the actuations of the pumps and valves, and regulated the different phases of the operational cycles. Temperature, pH and dissolved oxygen (DO) concentrations in the reactor were not controlled being their mean values: 21.8 ± 1.4 ºC, 8.0 ± 0.3 and 7.9 ± 0.4 mg O2/L, respectively. The composition of the feeding medium was, in g L-1: 0.084 NaHCO3, 0.092 K2HPO4, 0.036 KH2PO4, 0.049 MgSO4, 0.019 KCl, 0.5 mL L-1 of a trace solution and the nitrogen source: ClNH4 (1.49 g L-1) or NaNO2 (0.25 - 2.71 g L-1) (Table 4.1). The composition of the trace solution was in g L-1: 1.50 FeCl3·6 H2O, 0.15 H3BO3, 0.15 CoCl2·6 H2O, 0.12 MnCl2·4 H2O, 0.12 ZnSO4·7 H2O, 0.06 NaMoO4·2 H2O, 0.03 CuSO4·5 H2O and 0.03 KI. Table 4.1. Main operational conditions in the different stages of the SBR reactor. Stage Time (days) N-source Inlet concentration (mg N L-1) I 0-64 Ammonium 390 II 65-134 Nitrite 100 III 135-220 Nitrite 360 IV 221-285 Nitrite 500 4.3.2. Operational conditions The reactor was previously operated for a year in order to develop nitrifying granules from an activated sludge collected from a municipal wastewater treatment plant (data not shown). At the beginning of the experiment, the nitrifying granules had a mean feret diameter of 0.8 mm and the biomass concentration was 0.7 g VSS L-1. An ammonium loading rate of 0.3 g N L-1 d-1 was applied during 64 days. At day 65, the nitrogen source of the autotrophic medium was change by nitrite to suppress the activity of AOB. The nitrite loading rate applied to the systems was stepwise increased from 0.05 to 0.55 g N L-1 d-1. 4.3.3. Analytical methods The pH and the concentrations of DO, ammonium, volatile suspended solids (VSS) and sludge volumetric index (SVI5) were determined according to the Standard Methods (APHA-AWWA-WPCF, 1998). Nitrite and nitrate concentrations were determined by capillary electrophoresis (Vilas-Cruz et al., 1994). The morphology and size distribution, the aspect ratio and the roundness of the granules were measured regularly by using an image analysis procedure (Tijhuis et al., 1994) with a stereomicroscope (Stemi 2000-C, Zeiss) incorporating a digital camera (Coolsnap, Roper Sicientific Photometrics). For the digital image analysis the programme Image Pro Plus (Media Cybernetics) was used. Samples of suspended biomass were collected from the reactor on operating days 0, 77, 176 and 220. The FISH technique was performed according to the procedure described by Amann et al., (1995) with 4% paraformaldehyde solution. To achieve the granular biomass breakage, biomass was sonicated for 1 min at 65% of amplitude using a probe sonicator (UP200s, Dr. Hielscher). The specific oligonucleotide probes used are represented in Table 4.2. The oligonucleotide probes were labelled with the fluorochromes Cy3 and FLUOS obtained from ThermoHybaid (Ulm, Germany). After in situ hybridization cells were stained with DAPI (0.5 µg mL-1) for 10 min. Fluorescence signals were recorded with an acquisition system coupled to an Axioskop 2 epifluorescence microscope (Zeiss, Germany). Quantification of the bacterial population was based on the procedure published by (Crocetti et al., 2002). For each hybridization experiment at least 20 randomly 89 Chapter 4 chosen images were recorded, and the ratio of the area of those cells labelled by the specific probe to the area of all bacteria stained by DAPI was determined by digital image analysis using Image ProPlus. Further information about the analytical methods is provided in Chapter 2. Table 4.2. Targeted organisms and the corresponding formamide (FA) percentages for the used oligonucleotide probes. Probea Probe sequence (5’3’) FA (%) Targeted organisms EUB338 GCT GCC TCC CGT AGG AGT 0-50 Bacteria domain EUB338-II GCA GCC ACC CGT AGG TGT 0-50 Planctomycetales EUB338-III GCT GCC ACC CGT AGG TGT 0-50 Verrucomicrobiales ALF1b CGT TCG Y(C/T)TC TGA GCC AG 20 Most - Proteobacteria and other bacteria Bet42ab GCC TTC CCA CTT CGT TT 35 - Proteobacteria Pae997 TCT GGA AAG TTC TCA GCA 0 Pseudomonas spp. PAR1244 GGA TTA ACC CAC TGT CAC C 20 Paracoccus NEU653b CCC CTC TGC TGC ACT CTA 40 Most halophilic and halotolerant Nitrosomonas spp. Ntspa712 CGC CTT CGC CAC CGG CCT TCC 50 most members of the phylum Nitrospirae NIT3b CCT GTG CTC CAT GCT CCG 40 Nitrobacter spp. aDetails on oligonucleotide probes are available at probeBase (Loy et al., 2007). bUsed with an equimolar amount of corresponding unlabelled competitor oligonucleotide probe. 4.3.4. Calculations The maximal oxygen amount transferred from the bulk liquid to the granule surface was calculated according to Eq. 4.1.  SLcO CCAkJ 2 (4.1) being JO2 the oxygen flux (g O2 d-1); kc the mass transfer coefficient (m d-1), A the total surface of the granules (m2); CL the DO concentration in the bulk liquid (g O2 m-3) and CS the DO concentration at the surface of the granule. The total surface of the granules A (m2) was calculated by multiplying the total volume of the granules in the reactor by the specific surface (Eq. 4.2). mgranule RR D 6 ρ XV A (4.2) being VR the reactor volume (m3), XR the biomass concentration in the reactor (kg VSS m-3), granule the granules density (kg VSS mgranule-3) and Dm the mean diameter of the granule (m). kc was calculated following the procedure described in Chapter 3 (from Garrido et al., 1997). The oxygen required to oxidize the nitrogen loading rate (JO2required, g O2 d-1) applied was calculated according to the stoichiometry of the ammonium oxidation (NH4+ + 1.5 O2 → NO2- + 2 H+, i.e., 4.57 g O2 required per g of NH4+-N oxidized) and nitrite oxidation (NO2- + 0.5 O2 → NO3-, i.e., 1.14 g O2 required per g of NO2--N oxidized) processes (Eq. 4.3).  HRT VC4.57 JR inf NH I required O2 4    HRT VC1.14 JR inf NO IVIII,II, required O2 2   (4.3) being HRT the hydraulic retention time (d-1) and CNH4+inf and CNO2-inf the concentrations of ammonium and nitrite in influent (g N L-1). 90 Population dynamics of nitrite oxidizers in nitrifying granules In the same way, the oxygen consumed (JO2cons, g O2 d-1) during the different stages was calculated according to Eq. 4.4, being CNO3eff and CNO2eff the concentrations of nitrate and nitrite in the effluent (g N L-1).    HRT VC433C4.57 JR eff NO eff NO I removed O2 23   .  HRT VC1.14 JR eff NO IVIII,II, removed O2 3   (4.4) 4.4. Results and discussion 4.4.1. Biomass characteristics The biomass concentration in the reactor remained practically constant in the range 0.5-0.7 g VSS L-1 during the whole operational stage in spite of changes of feeding composition. During stage I, the VSS concentrations in the effluent were around 18 mg VSS L-1 but those values decreased to 6 mg VSS L-1 when the nitrogen source was changed in feeding media (stages II-IV). The low value of VSS in the effluent allowed working at sludge retention times around 100 d from day 40 on, indicating the good retention capacity of the SBR. The obtained biomass concentration was similar to the values obtained by Kim and Kim, (2006) in nitrate oxidizing biofilm reactors (0.39 – 0.54 g VSS L-1) and higher than the values obtained by Nogueira and Melo, (2006) in a chemostat (5.3 - 5.8 mg VSS L-1). The granules maintained their physical structure once nitrite was fed in the place of ammonium. No breakage was observed as demonstrated by the only slight variations registered in the granules physical properties (Fig. 1). The volume weighted mean diameter registered a slight decrease from 0.8 mm to 0.75 mm due to the disappearance of the biggest particles, larger than 2.8 mm, and the increase in importance of particles from 0.6 to 1.0 mm (Fig. 4.1). The aspect ratio and the roundness did not vary significantly remaining both values around 1.5. The SVI5 and the density of the granules were kept at around 20 mL (g VSS)-1 and 77 g VSS (Lgranule)-1, respectively during the whole experiment. These values are similar to those obtained by Campos et al., (2000) who measured a SVI of 17 mL (g VSS)-1 and a density of 100 g VSS (Lbiomass)-1 for nitrifying aggregates with 0.36 mm of diameter carrying out complete nitrification. Therefore, it can be concluded that the long term operation of nitrifying granules fed either with both ammonium or nitrite allow the maintaining of their integrity and their good settleability. 0 5 10 15 20 25 0 0.4 0.8 1.2 1.6 2 2.4 2.8 3.2 Diameter (mm) % by volume Figure 4.1. a) Image of the nitrite oxidizing granules on day 215 b) Size distribution of the granules on day 0 (■) and 270 (□). 91 Chapter 4 4.4.2. Reactor performance Ammonium was oxidized into nitrate with an efficiency of 83% during stage I. The nitrite concentration remained around 12 mg NO2--N L-1 in the effluent (Figure 4.2a). From day 65 on, a feeding media containing nitrite instead of ammonium was supplied to the system. As the NLR was increased the efficiency of nitrite oxidation decreased from 90 to 68% and, consequently, nitrite concentration in the effluent increased from 12 to 160 mg NO2--N L-1 (Stages II-IV). Since biomass concentration, physical characteristics of granules and DO concentration in the liquid bulk did not change significantly during whole operational stage, the maximal oxygen flux through the surface of the granules was expected to keep practically constant (Figure 4.2b). However, the amount of oxygen required to oxidize all substrate applied changed due to the modifications of feeding composition. During stage I the maximal oxygen flux through the surface of the granules calculated had a similar value to that of the oxygen consumption rate which would indicate that the capacity of the system was probably limited by external oxygen transfer rate. This fact would cause that oxygen concentration inside granules were lower than during stages II-IV when the oxygen consumption rate did not exceed external oxygen transfer rate. During these stages no important DO limitation was expected to occur. 0 150 300 450 600 I II III IV NH 4+ , NO 2- , NO 3- (mg N L -1 ) a) 0 1 2 3 4 I II III IV Period J O2 (g d -1 ) b) Figure 4.2. a) Concentrations of the nitrogen species in the influent (ammonium (Stage I) and nitrite (Stages II, III and IV)) ( ), and in the effluent: nitrate (□), nitrite (■) and ammonium (■). b) JO2max ( ), JO2required (□) and JO2removed (■). 92 Population dynamics of nitrite oxidizers in nitrifying granules 4.4.3. Microbial characterization of the biomass The evolution of the main microbial nitrifying communities in the granules was followed by FISH (Fig. 4.3). Figure 4.3. Bacterial populations on day 0: a) Nitrosomonas (NEU653, pink) and DAPI (blue) b) Nitrospira (Ntspa712, pink) and DAPI (blue). Bacterial population on day 77: c) Nitrosomonas (NEU653, pink) and DAPI (blue) d) Nitrospira (Ntspa712, green); Nitrobacter (NIT3, pink) and DAPI (blue). Bacterial populations e) on day 176 and f) day 220: Nitrospira (Ntspa712, green); Nitrobacter (NIT3, pink) and DAPI (blue). The bar represents 25 µm. 93 Chapter 4 Samples collected on day 60 (stage I) were disaggregated and analyzed by FISH to identify the AOB and NOB populations. Analysis with the probe NEU653 showed that the granules were mainly constituted by clusters of densely packed Nitrosomonas that represented 60% of the bacterial population (Fig. 4.3 and 4.4). The probes Nit3 and Ntspa712 were also applied and only bacteria belonging to the phylum Nitrospirae were detected. This NOB represented 30% of the whole bacterial populations. 0 20 40 60 80 I II III IV Stage % Population Figure 4.4. Evolution of the bacterial populations distribution of Nitrosomonas (NEU653, ■), Nitrospira (Ntspa712, □), Nitrobacter (NIT3, ) and others (the difference between the DAPI signal, which represents 100% of bacterial populations, and the sum of the positive signals detected with NEU653, Ntspa712 and NIT3, ■). During stage II, the amount of AOB in the biomass drastically decreased. The performed FISH analysis showed that less of the 10% of the microbial population hybridized with probe NEU653. Compared to the first sample taken on day 60, the majority of the Nitrosomonas were no longer present in the reactor or at least not active according to the FISH results. The active 10% fraction could have survived carrying out another processes, e.g., metabolizing organic compounds (Ahn, 2006). With regards to NOB population, an increase in the Nitropira percentage up to 40% was observed and Nitrobacter were detected but only as a 5% of the total. The remaining percentage (45%) could be composed by inactive or dead Nitrosomonas trapped in the granules or by heterotrophs which likely grew on cell soluble microbial products (humic and fulvic acids, polysaccharides, proteins, amino acids, etc.) released during Nitrosomonas endogenous respiration (Barker and Stuckey, 1999). Although probes Par1244 and Pae997 for Paracoccus and Pseudomonas were applied in order to identify possible heterotrophs, no positive results were observed. During stage III, the percentage of AOB was practically less of 1% while the Nitrobacter percentage in the total biomass increased to values around 45%, remaining Nitrospira at 40%. No significant changes of percentages of the different populations were observed on Stage IV. The nitrite and oxygen affinity constants for Nitrospira-like bacteria (0.14 mg NO2--N L-1 and 0.13 mg O2 L-1) are lower than those of Nitrobacter (7 mg NO2--N L-1 and 1.98 mg O2 L-1) (Schramm et al., 1999; Manser et al., 2005) which cause that the former have competitive advantage at low substrates concentrations. Nitrospira are generally the NOB predominant in continuous biofilm systems even when nitrite is present at high concentrations, i.e., even with stable nitrite accumulation during the reactor operation (Table 4.3). In these 94 Population dynamics of nitrite oxidizers in nitrifying granules systems oxygen concentrations rather than nitrite concentrations could be the selective parameter for NOB. During continuous operation the substrates consumption rates are almost constant meaning that their profiles inside the biofilm remain also constant. Oxygen is generally consumed by ammonium or heterotrophic biomass located in the outer layers being scarcely present in the deeper layers of biofilm. On the other hand in SBR systems, substrates concentrations change along the time according to the operational phases which provoke concentration profiles. This fact could provoke that, in some cases, higher oxygen concentrations could be available for bacteria located in deeper layers during some stages of time. These alternating conditions could explain that there is not a clear predominant NOB type in this kind of systems but coexistence between populations. Table 4.3. Nitrospira and Nitrobacter distribution in different systems. System Flow NOB detected Nitrite accumulation Reference Biofilm Continuous Nitrospira ------- Cortes-Lorenzo et al., 2006 Biofilm Continuous Nitrospira ------- Daims et al., 2001 Biofilm Continuous Nitrospira Stable (up to 600 mg N L-1) Han et al., 2003 Biofilm Continuous Nitrospira Stable (up to 45 mg N L-1) Okabe et al., 1999 Biofilm SBR Nitrospira + Nitrobacter ------- Daims et al., 2001 Granular SBR Nitrobacter -------- Wilen et al., 2004 Granular SBR Nitrospira Transitory (up to 24 mg N L-1) Carvalho et al., 2006 Granular SBR Nitrospira Stable (up to 15 mg N L-1) Wang et al., 2007 Granular SBR Nitrobacter Stable (up to 120 mg N L-1) Shi et al., 2009 In this work, a SBR system was operated with a long feeding/reaction stage which could be considered as a continuous system. This implies that substrates concentrations were constant along the whole operational cycle. The nitrite concentration was kept over 12 mg N L-1 to avoid nitrite to become the limiting substrate Nogueira and Melo, (2006). Oxygen was the limiting substrate during stage I when Nitrospira was the only NOB present. As soon as oxygen was not limited Nitrobacter appeared. These results agree with those of Schramm et al., (2000) who observed that Nitrobacter was the dominant NOB in the oxic part of a membranebound biofilm system operated under no limitation of nitrite whereas Nitrospira was dominant in the oxic-anoxic part. 4.5. Conclusions ● The nitrifying granules are able to maintain their structure fed with nitrite as energy source and only minor changes were registered in their physical properties. ● By switching the nitrogen source from ammonium to nitrite, a change in the biomass populations of the nitrifying granules was observed. At the beginning of the operation, Nitrosomonas was identified as the main population in a proportion of 60%, Nitrospira being the only NOB representing a 30%. After 220 days of operation with nitrite as nitrogen source Nitrospira and Nitrobacter coexisted in similar proportions around 4045%. 95 Chapter 5 Table 5.1. Characteristics of the feeding media used in the different operational stages. Stages Days NaCl (g L-1) NH4+-N (mg L-1) NO2--N (mg L-1) NLR (g N L-1 d-1) I 0-18 0 400 400 0.4 II 18-50 5 400 400 0.4 III 50-80 10 400 400 0.4 The composition of the synthetic autotrophic medium was as follows in g L-1: 1.25 KHCO3, 0.005 NaH2PO4, 0.30 CaCl2·2H2O, 0.20 MgSO4·7H2O, 0.00625 FeSO4, 0.00625 EDTA, 0.15 mL L-1 H2SO4 and 1.25 mL L-1 of a trace solution. The composition of the trace solution was in g L-1: 15 EDTA, 0.43 ZnSO4·7 H2O, 0.24 CoCl2·6 H2O, 0.99 MnCl2·4 H2O, 0.25 CuSO4·5 H2O, 0.22 NaMoO4·2 H2O, 0.19 NiCl2·6 H2O, 0.21 NaSeO4·10 H2O, 0.014 H3BO3 and 0.05 NaWO4·2 H2O. 5.3.3. Inocula The reactor was inoculated with enriched Anammox sludge from a laboratory scale SBR operated at the University of Santiago of Compostela (Dapena-Mora et al., 2004b). The initial concentration of biomass was 1.0 g VSS L-1. The initial specific Anammox activity of the biomass was 0.4 g N (g VSS)-1 d-1. 5.3.4. Analytical methods The pH, ammonia, volatile suspended solids (VSS), inorganic suspended solids (ISS) and sludge volumetric index (SVI5) were determined according to the Standard Methods (APHA-AWWA-WPCF, 1998). Nitrite and nitrate concentrations were determined by capillary electrophoresis (Vilas-Cruz et al., 1994). DO concentration was measured with a dissolved oxygen probe (AQUALITYC, model OXI-921) connected to a meter (M-Design Instruments TM-3659). The distribution of particle size was measured using an Image Analysis procedure (Tijhuis et al., 1994). The morphology and size distribution of the granules were measured regularly by using an image analysis procedure with a stereomicroscope (Stemi 2000-C, Zeiss) provided with a digital camera (Coolsnap, Roper Sicientific Photometrics). For the digital image analysis the programme Image Pro Plus was used. The elemental analysis of the surface of the biomass aggregates was made using a transmission electron microscope (TEM) (PHILIPS CM12). Further information about the analytical methods is provided in Chapter 2. 5.3.5. Specific Anammox activity tests and specific nitrogen loading rate Batch experiments to determine the specific Anammox activity (SAA) were performed according to the methodology described by Dapena-Mora et al., (2007), based on the measurement along time of the overpressure generated in closed vials by the nitrogen gas produced as detailed in Chapter 2. For stages II and III where NaCl was present in the influent, two different tests were performed, one using a medium without NaCl and the other one using the medium composition in terms of NaCl concentration corresponding to each stage. The Specific Nitrogen Loading Rate (SNLR) was calculated dividing the Nitrogen Loading Rate (NLR) by the concentration of VSS in the reactor at the moment of the calculation. 102 Granular systems to improve Anammox biomass retention 5.4. Results and discussion 5.4.1. Reactor operation The experiments in the SBR were performed during 72 days. The nitrite (which is the limiting substrate since the feeding media contained ammonium and nitrite in a ratio 1:1) was almost fully depleted along the cycle and the concentration of ammonium in the effluent was practically constant at a value of 100 mg NH4+-N L-1 (Fig. 5.3). The NLR applied during the whole operation of the SBR was around 0.4 g N L-1 d-1. No significant effect was observed in the effluent composition due to the presence of the different NaCl concentrations (5 and 10 g L-1) tested in the feeding media. 0 20 40 60 80 100 0 2040608 Tim e (d) NH 4+ , NO 2- (mg N L -1 ) III III 0 Figure 5.3. Ammonium (○) and nitrite (●) concentrations in the liquid phase of the SBR. In order to prove that the system was not overloaded the SNLR applied to the reactor and the SAA of the biomass obtained from batch tests were compared following the same procedure as Dapena-Mora et al., (2004c). The SNLR applied to the SBR during its operation was always below the maximum specific activity of the biomass which ranged between 0.35 and 0.45 g N (g VSS)-1 d-1 (Fig. 5.4). This fact indicates that the system was operated close to be overloaded only during stage I. During the rest of the operational stages not only no biomass wash-out was observed but also an important increase of the biomass concentration was measured which explains the decrease in the SNLR while the NLR was kept constant. The SAA of the biomass decreased from 0.4 to 0.35 g N (g VSS)-1 d-1 when 5 g L-1 of NaCl were added but when the concentration of salt was increased up to 10 g L-1 the SAA increased to 0.45 g N (g VSS)-1 d-1 which is slightly higher than the initial value (Fig. 5.4). These changes of the SAA values are not correlated to the salinity of the medium and could be due to the necessity of an adaptation period. This adaptation process is needed when adding salt to biomass non-adapted but also when removing salt of salt-adapted biomass. The batch activity assays to determine the SAA showed no difference with or without salt addition in the medium during stage II (5 g NaCl L-1). However, during stage III the batch assays performed without salt addition in the medium showed a decrease in the SAA of 20% corroborating the need of an adaptation period. The adaptation of Anammox biomass to high salt concentrations observed in the present work agrees with the results obtained in previous works. Kartal et al., (2006) observed that adapted Anammox biomass 103 Chapter 5 maintain some activity at salt concentrations up to 30 g L-1. Dapena-Mora et al., (2007) performed batch tests to study the effect of different salts on the specific activity of Anammox biomass non-adapted to high concentrations of salts. They observed that 3 mg L-1 of NaCl enhanced the Anammox activity; an increase in the SAA of 20% was obtained. However, a NaCl concentration of 10 mg L-1 decreased slightly the Anammox activity to 90% of its initial value in absence of salt. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0204060 Time (d) SNLR, SAA (g N (g VSS) -1 d -1 ) 80 III III Figure 5.4. Specific Nitrogen Loading Rate (SNLR) applied (○) and Specific Anammox Activity of the biomass (SAA) (). The values of SAA obtained in the present work were similar to those found by Dapena-Mora et al., (2004c) operating a SBR with granular biomass. In the same work, Dapena-Mora et al. (2004c) used a gas-lift reactor and observed values as high as 0.9 g N L-1 d-1, however the stability of their system was not very good and both, overloads and biomass flotation events took place. 5.4.2. Biomass retention The biomass concentration was constant in the SBR (Fig. 5.5) during stage I which was an indication of the existing balance between production and wash-out of biomass. 0.0 0.5 1.0 1.5 2.0 0 2040608 Time (d) Reactor (g VSS L -1 ) III III 0 Figure 5.5. Concentration of biomass in the reactor (g VSS L-1) (●). 104 Granular systems to improve Anammox biomass retention Once the NaCl was added in the feeding (stages II and III) the biomass concentration increased steadily in 0.6 g VSS L-1 within 54 days. The productivity of Anammox sludge obtained from the stoichiometry of the process given by Strous et al. (1998) was 0.038 g VSS produced per g NH4+-N consumed. Therefore, 0.82 g VSS would have been produced during this interval of operation of the reactor. Assuming that the decay of the biomass is neglected the amount of biomass retained in the reactor was the 73 % of the stoichiometrically produced amount. After NaCl addition, the concentration of solids in the effluent decreased significantly and, as a consequence, SRT values of 50 days were obtained (Fig. 5.6). 0 10 20 30 40 50 60 020406080 Time (d) Efluent (mg VSS L -1 ) 0 10 20 30 40 50 60 SRT (d) I II III Figure 5.6. Biomass concentration in the effluent (○) and sludge retention time (SRT) (▲). 5.4.3. Reasons for the improvement of biomass retention In order to establish the different phenomena involved in the improvement of the biomass in the reactor, the characteristics of the obtained biomass aggregates were analysed. The percentage of inorganic fraction in the biomass, as the ratio between ISS and TSS, increased from the 10% in the stage I until 21% in stage III. This inorganic fraction is attributed to the precipitation of inorganic salts due to the NaCl addition. The precipitation was achieved increasing the salinity of the media and therefore varying the equilibrium and the solubility of the products. Elemental analysis of the precipitates present in the surface of the granules indicated that they were mainly composed by Ca3PO4 which is characterized by its very low solubility. These precipitates could easily act as precursors for the formation of Anammox granules, being a support material where the biomass could attach, or this precipitation could also occur directly on the surface of already formed aggregates. Thus the improvement of the settleability of the granules can be directly related to this inorganic precipitation which caused an increase of the density of the aggregates producing sludge with good SVI5 values: 52 mL (g VSS)-1 at the end of stage III (Fig. 5.7). Although the precipitates play an important role as support material, in the conditions described it can also represent a problem if it occurs in a large extension on the surface of already formed aggregates. In such a way, if the amount of formed precipitates is too high the whole surface of the aggregates can be covered and the activity of the biomass reduced due to the increase of diffusional resistance (Trigo et al., 2006). 105 Chapter 5 0 30 60 90 120 0 20406080 Time (days) SVI 5 (mL/g VSS) 0.2 0.3 0.4 0.5 0.6 Diameter (mm) I II III Figure 5.7. Sludge volumetric index (○) and average feret diameter (●). Concerning the biomass physical characteristics, the feret diameter increased from 0.47 to 0.58 mm (24% of increase) and the SVI5 decreased from 120 to 50 mL (g VSS)-1 indicating the compacting of the biomass (Fig. 5.7). Therefore, the capacity of the system to retain biomass increased. Visual observation of the biomass samples indicated that during stage I the biomass forming small flocs and in suspension was predominant (Fig. 5.8). During stages II and III no biomass in suspension was observed and the small flocs became aggregates. This behaviour was already observed by Campos et al., (2002) working with nitrifying sludge. 5.4.4. Applications The external addition of salt to enhance granulation of Anammox biomass might not be feasible at industrial scale due the cost of reagents. But the use of wastewater containing high salt concentrations is a possibility. In the industry effluents with high salt content are relatively common, as those produced in the fish canning industry, due to the use of sea water during the manufacturing processes. These effluents are firstly treated in an anaerobic digestion stage and then a treatment for nitrogen removal is necessary. Therefore, the application of the Anammox process for their treatment could be advisable together with the use of the salt content to promote precipitates and to produce good settling aggregates. Although the promotion of precipitates must be carefully controlled to avoid the precipitation of extremely high amounts of salts, which could diminish the activity of the Anammox biomass as it was previously observed in the case of anaerobic sludge (Van Langerak et al., 1998). 106 Granular systems to improve Anammox biomass retention a) Day 1 b) Day 39 c) Day 61 d) Day 72 Figure 5.7. Stereomicroscope images (zoom: 10x) of biomass in the SBR in different operational days. 5.5. Conclusions ● Improvement of the Anammox biomass retention, indicated by an increase of biomass concentration inside SBR via dissolved salt addition up to 10 g NaCl L-1 for precipitate production. ● The increase of biomass retention was correlated to an improvement of biomass settling properties (lower SVI5 values of 50 mL (g VSS)-1 and bigger average diameters of 0.58 mm) in SBR1. Reduction of biomass growth in suspension was observed. ● The SAA of the biomass experienced slight variations in the presence of NaCl concentrations and was maintained around 0.40―0.45 g N (g VSS)-1 d-1. ● The choice of the application of precipitation to improve Anammox biomass retention depends upon the characteristics of the water to be treated. 107 Chapter 5 108 5.6. References APHA-AWWA-WPCF (1998) Standard methods for the examination of water and wastewater. Washington DC, USA: American Public Health Association/American Water Works Association/Water Environment Federation. Arrojo, B., Mosquera-Corral, A., Garrido, J.M., and Mendez, R. (2004) Aerobic granulation with industrial wastewater in sequencing batch reactors. Water Research 38: 3389-3399. Campos, J.L., Mosquera-Corral, A., Sanchez, M., Mendez, R., and Lema, J.M. (2002) Nitrification in saline wastewater with high ammonia concentration in an activated sludge unit. Water Research 36: 2555-2560. Dapena-Mora, A., Arrojo, B., Campos, J.L., Mosquera-Corral, A., and Mendez, R. (2004a) Improvement of the settling properties of Anammox sludge in an SBR. Journal of Chemical Technology and Biotechnology 79: 1417-1420. Dapena-Mora, A., Van Hulle, S.W.H., Campos, J.L., Mendez, R., Vanrolleghem, P.A., and Jetten, M. (2004b) Enrichment of Anammox biomass from municipal activated sludge: experimental and modelling results. Journal of Chemical Technology and Biotechnology 79: 1421-1428. Dapena-Mora, A., Campos, J.L., Mosquera-Corral, A., Jetten, M.S.M., and Mendez, R. (2004c) Stability of the ANAMMOX process in a gas-lift reactor and a SBR. Journal of Biotechnology 110: 159-170. Dapena-Mora, A., Fernandez, I., Campos, J.L., Mosquera-Corral, A., Mendez, R., and Jetten, M.S.M. (2007) Evaluation of activity and inhibition effects on Anammox process by batch tests based on the nitrogen gas production. Enzyme and Microbial Technology 40: 859-865. Jetten, M.S.M., Horn, S.J., and vanLoosdrecht, M.C.M. (1997) Towards a more sustainable municipal wastewater treatment system. Water Science and Technology 35: 171-180. Kartal, B., Koleva, M., Arsov, R., van der Star, W., Jetten, M.S.M., and Strous, M. (2006) Adaptation of a freshwater anammox population to high salinity wastewater. Journal of Biotechnology 126: 546-553. Maximova, N., and Dahl, O. (2006) Environmental implications of aggregation phenomena: Current understanding. Current Opinion in Colloid & Interface Science 11: 246-266. Strous, M., Heijnen, J.J., Kuenen, J.G., and Jetten, M.S.M. (1998) The sequencing batch reactor as a powerful tool for the study of slowly growing anaerobic ammonium-oxidizing microorganisms. Applied Microbiology and Biotechnology 50: 589-596. Strous, M., Kuenen, J.G., and Jetten, M.S.M. (1999) Key physiology of anaerobic ammonium oxidation. Applied and Environmental Microbiology 65: 3248-3250. Tijhuis, L., Vanbenthum, W.A.J., Vanloosdrecht, M.C.M., and Heijnen, J.J. (1994) Solids Retention Time in Spherical Biofilms in a Biofilm Airlift Suspension Reactor. Biotechnology and Bioengineering 44: 867-879. Trigo, C., Campos, J.L., Garrido, J.M., and Mendez, R. (2006) Start-up of the Anammox process in a membrane bioreactor. Journal of Biotechnology 126: 475-487. Van Langerak, E.P.A., Gonzalez-Gil, G., Van Aelst, A., Van Lier, J.B., Hamelers, H.V.M., and Lettinga, G. (1998) Effects of high calcium concentrations on the development of methanogenic sludge in upflow anaerobic sludge bed (UASB) reactors. Water Research 32: 1255-1263. Vilas-Cruz, M., Gómez, J., Méndez, R., and Lema, J.M. (1994) Simultaneous determination of NO2and NO3in wastewater by capillary electrophoresis. In International Symposium of Analytical Methodology for the Environment, pp. 1-50. Chapter 6 Development of the CANON process in a continuously aerated granular SBR and an air pulsing SBR to treat anaerobic digester supernatants at room temperature1,2,3 Summary The CANON (Completely Autotrophic Nitrogen removal Over Nitrite) process was successfully developed in two sequencing batch reactors (SBR): a granular and an air pulsing reactor (SBR and SBRP) fed with the supernatant from an anaerobic sludge digester and operated at moderately low temperatures (18-24 ºC). Both units were started up as nitrifying reactors where the dissolved oxygen concentration was decreased until partial nitrification was achieved. In the granular SBR, while partial nitrification occurred, nitrogen losses due to the spontaneous growth of Anammox bacteria inside the reactor occurred. Once the stable CANON process was established, a mean nitrogen removal rate of 0.8 ± 0.1 g N L-1 d-1 was registered. The settling velocities of the granules ranged from 70 to 150 m h-1 with sludge volumetric index (SVI5) values lower than 50 mL (g VSS)-1 during the whole operation. The SBRP was inoculated with sludge containing Anammox biomass once stable partial nitrification was achieved. The maximal nitrogen removal rate obtained was of 0.45 g N L-1 d -1. By working at a dissolved oxygen concentration of 0.5 mg O2 L -1 in the bulk liquid, nitrogen removal percentages up to 85% were achieved. The biomass content in the reactor was of 4.5 g VSS L-1. The ammonium oxidizing bacteria (AOB) were in the form of flocs and granules (average diameter of 1.6 mm). These granules contained Anammox bacteria (present in the inner core of the granules) surrounded by AOB (present in the external layers of the granules). 1Vázquez-Padín J., Fernández I. Figueroa M., Mosquera-Corral A., Campos J.L. and Méndez R. (2009) Applications of Anammox based processes to treat anaerobic digester supernatant at room temperature. Bioresource Technology 100: 2988-2994. 2Vázquez-Padín J., Pozo M.J. Jarpa M., Figueroa M., Franco A., Mosquera-Corral A., Campos J.L. and Méndez R. (2009) Treatment of anaerobic sludge digester effluents by the CANON process in an air pulsing SBR. Journal of Hazardous Materials 166: 336-341. 3Vázquez-Padín J., Figueroa M., Fernández I., Mosquera-Corral A., Campos J.L. and Méndez R. (2009) Post-treatment of effluents from anaerobic digesters by the Anammox process. Water Science and Technology 60(5): 1135-1143. Chapter 6 6.1. Introduction Nitrogen removal has gained in attention over the last decades due to the problem that its presence originates in aquatic systems, such as oxygen depletion, fish toxicity and media eutrophication. The conventional process is based on a sequence of aerobic and anoxic conditions where the ammonium is firstly oxidized to nitrate and then reduced using an organic carbon source as electron donor during the denitrification process. Efficient and sustainable processes like anaerobic digestion in many cases produce effluents highly contaminated in ammonium and poorly concentrated in biodegradable chemical oxygen demand (COD). In those cases, to carry out the conventional nitrification-denitrification processes the addition of an external carbon source is needed to reach a complete nitrogen removal. When this is the case an alternative strategy consists of making a shortcut in the nitrogen cycle, the “nitrite route” avoiding the nitrite oxidation to nitrate and its further denitrification. This route can be carried out in two ways. One alternative consists of the partial nitrification of all the ammonium to nitrite and the subsequent denitrification with an organic carbon source. This alternative involves the saving of 25% of aeration costs, and 40% of addition of external carbon source. Furthermore lower sludge production is registered (Van Loosdrecht and Jetten, 1998). The application of the second option, the combination of a partial nitrification followed by the Anammox process, is recommended when the treated wastewater has no biodegradable carbon source. The Anammox process consists of the anaerobic oxidation of ammonium (Van de Graaf et al., 1995; Van de Graaf et al., 1996) using nitrite as electron acceptor according to the stoichiometry described by Strous et al., (1999) (Eq. 6.1). NH4+ + 1.32 NO2- + 0.066 HCO3- + 0.13 H+  1.02 N2 + 0.26 NO3- + 0.066 CH2O0.5N0.15 + 2 H2O (6.1) In general it is not common to find effluents with the required composition to be treated via the Anammox process. For this reason several alternatives have been studied to obtain an effluent containing both ammonium and nitrite. To reach this objective, half of the ammonium fed to the system has to be converted into nitrite by the Ammonium oxidizing bacteria (AOB) and therefore, the oxidation of nitrite to nitrate carried out by Nitrite Oxidizing Bacteria (NOB) has to be avoided. The AOB and NOB are two phylogenetically unrelated groups whose different growth rates and the way this growth rates are affected by parameters like temperature, pH, dissolved oxygen (DO), etc. can be used to outcompete NOB and to uncouple both reaction rates. Nitrite accumulations have been reported in several systems, for biomass growing in suspension (Blackburne et al., 2008), in biofilm (Garrido et al., 1997) or in aggregates/granules (Kim and Seo, 2006). Several strategies have been used to reach partial nitrification (Ahn, 2006): 1) Increasing free ammonia concentration working at high pH values and limiting the growth of NOB due to their higher sensitivity to free ammonia inhibition than AOB (Anthonisen et al., 1976). 2) Decreasing the dissolved oxygen concentration due to the lower oxygen affinity of the NOB compared to AOB (Wiesmann, 1994). 3) Operating at temperatures above 25 ºC since the maximum specific growth rate of the AOB will be higher than that of NOB at these conditions. In fact this is the basis of the SHARON technology which consists 110 Development of the CANON process to treat anaerobic digester supernatants at room temperature of a Continuous Stirring Tank Reactor (CSTR) operated a hydraulic retention time (HRT) of around 1 day and 30 ºC to favor the growth of AOB and the washout of the NOB (Hellinga et al., 1998). In order to apply the partial nitrification and the Anammox processes it is important to take into account that they can be performed in two different units (Sharon-Anammox processes) or in a single one, called by different names, CANON: Completely Autotrophic Nitrogen removal Over Nitrite process (Third et al., 2001), OLAND: Oxygen-Limited Autotrophic Nitrification-Denitrification (Kuai and Verstraete, 1998; Pynaert et al., 2004) or aerobic/anoxic deammonification (Hippen et al., 1997; Helmer et al., 2001). Under oxygen-limited conditions a co-culture of aerobic and anaerobic ammonium-oxidizing bacteria can be established in a single unit in a CANON system. In those systems, NOB compete for oxygen with the aerobic AOB and for the nitrite with Anammox bacteria, and thus its growth (and subsequent nitrate production) is prevented. Another reason to maintain low oxygen concentrations is that Anammox bacteria are reversibly inhibited by dissolved oxygen concentration higher than 0.5% of air saturation (Strous et al., 1997). Several strategies have been tested to optimize the start-up of reactors where the CANON process occurred. In most of the cases the CANON process was developed from an Anammox reactor where nitrifying biomass was inoculated and low oxygen concentrations were maintained inside the system (Sliekers et al., 2003; Third et al., 2005). Another possible strategy was to start-up the reactor as a partial nitrification system under oxygen limited conditions and then to inoculate Anammox biomass as it was carried out by Pynaert et al., (2004) and Gong et al., (2007) in a rotating disk contactor or a membrane aerated reactor, respectively. The main disadvantage of these processes relies on the low growth rate of AOB and Anammox bacteria with doubling times of 1 day (Wiesmann, 1994) and 11 days (Strous et al., 1998), respectively. To enhance the performance of reactors involving slow growing bacteria, high sludge retention times are mandatory and therefore the attachment of bacteria on a carrier material to develop biofilms or the self-aggregation concept in granules are recommended. Aerobic granulation presents the advantage of excellent settleability of the granules, high biomass retention and high resistance to toxic compounds. The applicability of the Anammox process at full-scale has already been demonstrated. Full-scale SHARON-Anammox plants are operating in the Netherlands and Japan treating wastewaters coming from: municipal reject water, tanneries, potato processing and semi conductor industry (Abma et al., 2007; van der Star et al., 2008; Lamsam et al., 2008). The one single reactor configuration has also been implemented at full scale in Austria, Germany and Switzerland to treat anaerobic digester supernatants and demonstrated being economically viable (Wett, 2006). 6.2. Objectives ● The aim of this work relies on the development of the CANON process in a continuously aerated granular sequencing batch reactor (SBR) and an air pulsing sequencing batch reactor (SBRP) operated both at room temperature. ● The suitability of both reactors will be analyzed according to the start-up duration to achieve significant CANON capacity, the maximal nitrogen removal capacity reached and the stability of the process operation. 111 Chapter 6 From day 120 on, increasing nitrogen losses were registered in the granular SBR with a simultaneous change in the colour of the biomass to reddish. After an exponential increase in the nitrogen removal rate during Stage III (day 190, Fig. 6.2b), a problem with the feeding pump caused an important decrease in the ANR down to 0.2 g N L-1 d-1. One month later, the system recovered its efficiency and an ANR of 0.6 g N L-1 d-1 was achieved at the end of this stage. Applied ALR to the system was increased from 1.2 to 1.5 g N L-1 d-1 during Stage IV. Nitrite was almost not detected in the effluent but ammonium concentrations between 30 and 120 mg NH4+-N L-1 still remained in the effluent. In spite of the increase of the DO up to 3.5 mg O2 L-1 during this stage the ammonium oxidation was the limited step of the CANON process. This result agrees with the results reported by Sliekers et al., (2003) working with the CANON process in an airlift reactor. An ANR of 1.0 g N L-1 d-1 was achieved and the nitrogen removal percentage was around 60%. During Stage V the applied ALR and the DO concentration were progressively increased up to 2.2 g N L-1 d-1 and 4.6 mg O2 L-1, respectively. Both AOR and ANR remained in similar values to those obtained during Stage IV and only an increase of the NOR from 0.1 to 0.3 g NO3--N L-1 d-1 was observed. This fact caused that ammonium concentration in the effluent increased significantly. Finally, DO concentration was restored to 2.8 mg O2 L-1 in order to avoid nitrite oxidation (Stage VI). This action completely stopped nitrite oxidation but also caused a strong decrease of the AOR to 0.4 g N L-1 d-1. 6.4.1.2. Air pulsing SBRP The SBRP reactor was operated for 400 days (Belmonte et al., 2009) previously to the 350 operation days of the present work. A synthetic media was used as feeding media in order to better control the operation of the pulsing reactor since no information is available about nitrification in aggregates formed in this kind of reactors. During the first 54 days (Stage I, Table 6.2), 0.25 g N L-1 d-1 of ammonium were oxidized to nitrate operating the reactor at a mean DO concentration in the liquid media of 3.1 mg O2 L-1 (Fig. 6.3a). From this point and in order to achieve partial nitrification in the system, the DO concentration was diminished by stepwise decreasing the air flow supplied in each pulse. From days 55 to day 79 the mean DO concentration was fixed at 2.4 mg O2 L -1. Under these conditions, no complete nitrification was achieved and nitrite accumulated in the liquid media due to DO limitation of NOB which present lower oxygen affinities for oxygen than AOB. From day 80 onwards the DO concentration was gradually decreased reaching a value of 0.5 g O2 L-1 at the end of Stage III. During this Stage III, on day 95 of operation, the synthetic media was substituted by the supernatant of an anaerobic digester in order to prove that the system was able to operate in stable conditions treating this effluent. At the end of the Stage III, partial nitrification with 1:1 molar ratio of NH4+/NO2was achieved while nitrate was absent due to NOB oxygen limitation. On day 125 of operation (Stage IV), once suitable conditions to grow Anammox bacteria (low DO concentration and equal concentrations of ammonium and nitrite) were achieved, 100 mL of sludge containing Anammox bacteria were inoculated to the reactor. One month after inoculation, significant decrease of ammonium and nitrite concentrations together with the slight increase of nitrate concentration and disappearance of nitrogen from the balance were measured in the reactor indicating the occurrence of an incipient Anammox activity. At this point it was considered that the nitrogen removal via the CANON process was established. To favour the nitrogen removal the nitrite concentration in the liquid media was maintained close to zero, in order to avoid nitrite inhibition of Anammox bacteria, while the volume of air injected in each pulse was slightly increased to augment the ammonium oxidation to nitrite. The increase of air volume pulsed did not revert on an increase of DO concentration in the liquid media because of its fast consumption by the ammonium oxidizing bacteria. At the end of Stage IV, a stable ANR of 0.25 g N L-1 d-1 was achieved (Fig. 6.3b). 118 Development of the CANON process to treat anaerobic digester supernatants at room temperature On day 180 (Stage V), the feeding pump did not work properly and provoked the destabilization of the system due to ammonium limitation which caused DO increase and subsequent nitrite peak that reached 50 mg N L-1. After 20 days the previous operational conditions were restored in the reactor and the nitrogen removal capacity increased continuously until the end of the experiment. From days 295 to 350, a stable operational period for the CANON process was obtained with an average ANR of 0.36 g N-1 L-1 d-1 (Fig. 6.3b) and nitrogen removal percentage of 77%. The assumption made in the balances (Eq. 6.3 to 6.5) not considering heterotrophic denitrification was justified by the low TOC removal of only 82 mg TOC L-1 d-1 which did not justify the amount of nitrogen removed from the reactor. Once nitrogen removal percentages close to the theoretical maximum value of 89% (according to the Anammox stoichiometry, Eq. 6.1) were reached, no further study of the maximal ANR reachable in this system was performed. The obtained nitrogen removal efficiency demonstrates that the growth of NOB is avoided due to nitrite and oxygen limitations. 0 50 100 150 200 250 300 350 0 50 100 150 200 250 300 350 Time ( d) NH 4+ , NO 2- , NO 3- (mg N L -1 ) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 DO (mg L -1 ) III IV III V a) 0.0 0.1 0.2 0.3 0.4 0.5 0 50 100 150 200 250 300 350 Time (d) AOR, NOR, ANR (g N L -1 d -1 ) III III IV V b) Figure 6.3. SBRP operation a) Concentrations of nitrogen compounds, in the influent: ammonium (▬) and in the effluent: ammonium (○), nitrite (●) and nitrate (▬).DO concentration in the bulk liquid (•••). b) Ammonium oxidation rate (AOR, ○), nitrite oxidation rate (NOR, ▬) and, nitrogen removal rate (ANR, ▬). 119 Chapter 6 6.4.2. Biomass physical properties 6.4.2.1. Granular SBR The biomass concentration in the granular SBR reactor ranged during the whole operational period between 5 and 8 g VSS L-1 (Fig. 6.4). During partial nitrification stages (Stages I and II), the biomass concentration was maintained around 5 g VSS L-1. This value increased once the CANON process developed (from day 120 on) presumably due to the growth of Anammox biomass inside the granules (Stages III to VI), reaching a maximum concentration of 8 g VSS L-1 which remained stable after day 300 of operation. A similar trend was registered in the evolution of the solids concentration in the effluent which increased from a mean value of 40 mg VSS L-1 during the partial nitrification operation to values up to 80 mg VSS L-1 during the CANON operation. The SRT in the SBR varied in the range between 20 and 80 d. A slight decrease in the SRT for the last stages was associated to the increase in the VSS concentration of the effluent (Fig. 6.5d). 0 2 4 6 8 10 0 50 100 150 200 250 300 350 400 Time (d) Reactor (g VSS L -1 ) 0.00 0.02 0.04 0.06 0.08 0.10 Effluent (g VSS L -1 ) III III IV VVI Figure 6.4. Concentrations of biomass in the granular SBR (●) and in the effluent (○). An increase in the mean diameter of the granules during the reactor operation from 2.3 to 3.2 mm was registered. This evolution was also associated to the development of Anammox biomass inside the granules (Fig. 6.5). These values are higher than those reported for nitrifying granular systems (Campos et al., 2000; Tsuneda et al., 2003; Kim and Seo, 2006) and similar to those reported for aerobic granules grown on carbon sources (de Kreuk et al., 2005). The colour of the granules changed once the Anammox biomass developed in their inner core. The typical reddish colour of Anammox bacteria could be easily recognized when observing the granules on a stereomicroscope (Fig 6.5b). The SVI5 value was kept always under 50 mL (g VSS)-1 and the settling velocity of the granular sludge ranged from 70 to 150 m h-1 during the whole operation, maintaining, therefore excellent settling properties. 120 Development of the CANON process to treat anaerobic digester supernatants at room temperature a) b) c) 0 5 10 15 20 0123456 Diameter (mm) % in volume ■ Day 44 □ Day 302 d) 0 20 40 60 80 100 0 100 200 300 400 Time (d) SRT (d) I II III IV VVI Figure 6.5. Pictures of the granular biomass of the SBR (zoom: 6.5) taken a) on day 44 and b) on day 302 and c) size distribution by volume on the cited days and d) Evolution of the SRT in the SBR. 6.4.2.2. Air pulsing SBRP In the case of the air pulsing SBPR during the two first stages, the biomass concentration in the reactor was of 1.5 g VSS L-1 and solids concentration in the effluent was of 5 mg VSS L-1 (Fig. 6.6) which allowed working at a solids retention time (SRT) of 150 d (Fig. 6.7d). Solids concentration in the effluent increased up to 20 mg VSS L-1 during Stage III whereas the biomass concentration in the reactor remained constant, reducing the SRT down to 40 d. This concentration in the effluent remained constant until the end of the operation and due to the increase of biomass concentration up to 4.5 g VSS L-1 (Stage V) the initial value of SRT was restored (Fig. 6.7d). Biomass production during Stage V was calculated taking into account the nitrogen consumption and the observed yield coefficients under these operational conditions (Eq. 6.6). The estimated amounts of biomass produced corresponding to AOB, NOB and Anammox bacteria were 1.27; 0.01 and 3.64 g VSS, respectively for the whole stage, which involved a retention capacity of 87% of generated biomass. A high biomass retention capacity is a key factor for a fast start-up of bioreactors working with slow growing biomass such as ammonium oxidizing and Anammox bacteria. A significant change in size, colour and aspect of the biomass was observed from Stage I to Stage V (Fig. 6.7). During the two first stages, the biomass was composed by granules with a mean diameter of 0.8 mm. After the decrease in the DO concentration, AOB formed fluffy structures in order to overcome diffusion limitation into the granules. As the ANR increased after the inoculation, an increase in the average feret diameter of the granules was also observed reaching values around 1.6 mm, doubling the initial size, with 121 Chapter 6 some granules reaching 3.6 mm of diameter. This increase could be attributed to the development of granular Anammox biomass during this stage. 0 1 2 3 4 5 0 50 100 150 200 250 300 350 Time (d) Reactor (g VSS L -1 ) 0.00 0.01 0.02 0.03 0.04 0.05 Effluent (g VSS L -1 ) I II III IV V Figure 6.6. Concentrations of biomass in the SBRP (●) and in the effluent (○). Therefore, according to Nielsen et al., (2005), two different biomass types are expected to perform the CANON process, a fluffy structure performing mainly partial nitrification and granules where a thin layer of AOB consumes oxygen protecting and providing nitrite to Anammox bacteria responsible of the autotrophic nitrogen removal. a) b) c) 0 10 20 30 0 0.8 1.6 2.4 3.2 4 Diameter (mm) % in volume ■ Day 74 □ Day 340 d) 0 50 100 150 200 0 50 100 150 200 250 300 350 Time (d) SRT (d) III III IV V Figure 6.7. Pictures of the biomass of the SBRP (zoom: 6.5) taken a) on day 74 and b) on day 340 and c) size distribution by volume on the cited days and d) Evolution of the SRT in the SBRP. 122 Development of the CANON process to treat anaerobic digester supernatants at room temperature 6.4.3. Identification of bacteria populations by FISH By the application of the FISH technique to a biomass sample collected from the granular SBR during Stage II (day 100), bacteria belonging to the genus Nitrosomonas were identified as the dominant AOB population in the samples. No positive results were obtained when the probes NIT3 and Ntspa712 were applied indicating the absence of nitrite oxidizing bacteria. In order to confirm the results obtained applying mass balances during Stage VI a sample of granules was collected on day 380. The granules were sliced and pictures after applying the FISH technique were taken to confirm the presence of AOB and Anammox bacteria. Bacteria belonging to the genus Nitrosomonas (NEU653) were identified as the dominant AOB population in the samples and they were located in the outermost layers of the granules. Anammox bacteria gave positive results with Amx820 probes, indicating the presence of Candidatus Brocadia anammoxidans and/or Candidatus Kuenenia stuttgartiensis located in more internal layers of the granules, where oxygen was consumed by AOB. In Fig. 6.8, the zone where both bacteria coexist is shown. This is a zone where the DO concentration would be close to zero. Since AOB are producing nitrite which is consumed by Anammox bacteria, a straight relation between them is established. This layers configuration could explain the robustness of this process. As it was already reported by Pynaert et al., (2004), the autotrophic nitrogen removal in one stage could be developed in a rotating disk contactor or in a granular SBR which confer to the Anammox process high resistance to temperature, pH or oxygen changes. This effect can be caused by the protection provided by the oxic layer with a high density of Nitrosomonas hindering and lowering the variations registered in the bulk liquid. b ) a ) Figure 6.8. a) Image of the biomass in the granular SBR on day 380 with the applied FISH probes: NEU653 (green, FLUOS) and Amx820 (blue, Cy5). b) Image of the biomass in the SBRP on day 262 with the applied FISH probes: NEU653 (green, FLUOS) and Amx820 (red, Cy3). The bar corresponds to 15 µm. Very similar results were obtained for the SBRP. In order to identify the main bacteria populations present in the reactor, the FISH technique was applied to biomass samples collected on day 262. By the application of FISH probes, bacteria belonging to the genus Nitrosomonas (NEU653) were identified as the dominant population in the samples accounting for the 75% ± 10% of the total biomass. Anammox bacteria gave positive results of both PLA46 and Amx820 probes, indicating the presence of Candidatus Brocadia 123 Chapter 6 anammoxidans and/or Candidatus Kuenenia stuttgartiensis. The coexistence of ammonium oxidizing and Anammox bacteria was well-supported by the results obtained with CANON process inside the pulsing SBR. Positive results were obtained with probes ALF1b and GAM42a, indicating the presence of some αand γ-Proteobacteria whereas no positive results were obtained when the probes NIT3 and Nitspa712 indicating the absence of NOB. 6.4.4. Partial nitrification and Anammox growth in the reactors As it is shown in Fig. 6.2 (Stage II of operation in the SBR) and Fig. 6.3 (at the end of the Stage III of operation in the SBRP) partial nitrification was achieved in both systems, SBR and SBRP, with the regulation of the DO concentration in the bulk liquid. The oxygen mass transfer limitation in the granules could explain the occurrence of the partial nitrification with an efficiency of 50% (as aforementioned in Chapter 3) and this fact will be analyzed in this section. No significant effect of the large SRT values measured in both reactors (Fig 6.5b and 6.7b) was observed even though these values were long enough to establish the NOB activity when neither inhibition nor substrate limitation was present. The SHARON process is one of the technologies proposed to achieve the partial nitrification previously to the application of the Anammox process and it was successfully used at full scale to treat the effluent from sludge digesters (van Dongen et al., 2001; van der Star et al., 2007). However when the temperature of the treated effluent is lower than 24 ºC the maximal growth rate of AOB turns lower than that of NOB and ammonium is fully oxidized into nitrate (Fux et al., 2002). Therefore, to achieve partial nitrification at temperatures lower than 24 ºC other strategies, such as inhibition of NOB by NH3 (FA) or HNO2 (FNA) (Kim and Seo, 2006) or operation at low DO concentrations (Blackburne et al., 2008), should be applied. Results from the granular SBR indicated that nitrite was not oxidized into nitrate until day 215 (Fig. 6.2b) in spite of the concentration of FA inside the system remained close to zero (Fig. 6.9a) while the highest rate of nitrite oxidation was registered during Stage V when FA concentrations were between 5 and 10 mg NH3-N L-1, values that would inhibit completely NOB activity according to Anthonisen et al., (1976). Therefore, there is not a direct correlation between the presence of free ammonia and the accumulation of nitrite in the system. The appearance of the activity of the NOB occurred simultaneously with an increase of the total surface of the granules due to both factors, the increase of the diameter of the granules and biomass concentration at the end of Stage III (Fig. 6.9a). This fact could indicate that partial nitrification was due to DO limitation by the external mass transfer resistance which provoked that the DO concentration on the surface of the granule was lower than that in the bulk liquid. This mass transfer resistance together with the internal one would explain that partial nitrification in granular and biofilm systems was observed at higher DO concentrations than in the case of activated sludge systems (Ruiz et al., 2003; Yun and Kim, 2003; Wyffels et al., 2004). The oxygen limitation was even more pronounced in the SBRP since once the DO concentration in the bulk liquid decreased under 1 mg O2 L-1 (after day 100) no more NOB activity was registered. At the end of Stage V (from day 270 to the end of the operation), even with free ammonia concentrations as low as 0.5 ± 0.5 mg N L-1 and the high surface available for oxygen transfer into the granule no nitrite oxidation took place (Fig. 6.9b). Mass transfer resistance caused that oxygen could only penetrate around 100-400 μm into the granule (Tsuneda et al., 2003; Chiu et al., 2007) meaning that the majority of the volume of the granule was not penetrated by oxygen and, therefore, its specific aerobic activity was lower in comparison with the activated sludge systems. This drawback is compensated by the large biomass amounts retained inside the system due 124 Development of the CANON process to treat anaerobic digester supernatants at room temperature to its excellent settling properties. The low biomass concentrations in the effluent of the nitrifying granular reactor would minimize the presence of solids in the influent of the Anammox SBR that could enhance heterotrophic activity compromising the process efficiency (Lackner et al., 2008). Partial nitrification with biomass in suspension controlling the DO concentration in the bulk liquid could turn unstable due to the AOB wash out and the NOB growth (Blackburne et al., 2008). In this aspect, the granular systems showed a stable value of the AOR during whole operational period and NOB activity could be avoided by decreasing DO concentration of the liquid bulk. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 50 100 150 200 250 300 350 400 Time (d) Granules surface (m 2 ) 0 5 10 15 20 25 30 35 mg NH 3 -N L -1 III III IV V VI a) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0 50 100 150 200 250 300 350 Time (d) Granules surface (m 2 ) 0 5 10 15 20 25 30 35 mg NH 3 -N L -1 IIIIIIIV V b) Figure 6.9. Concentrations of free ammonia (○) and total surface of the granules with trend line (●) in a) the SBR and b) the SBRP. After day 120, a progressive increase of nitrogen loss was measured in the effluent of the granular SBR indicating the growth of Anammox biomass (and, therefore, the development of the CANON process), in spite of the existence of adverse environmental conditions for these microorganisms: presence of oxygen, high nitrite concentrations and low temperature. The same phenomenon was already reported by Siegrist et al. (1998) in a nitrifying rotating contactor. This is also in accordance to the simulation results of Hao and van Loosdrecht (2004) who stated that the natural development of the CANON process in stable nitrifying biofilms operated at low temperatures is possible. ´ 125 Chapter 6 Since Anammox developed spontaneously in the SBR without any inoculation, it is difficult to compare the start up of the autotrophic nitrogen removal process with other systems; however, this comparison can be performed with the data obtained from the SBRP. 6.4.5. Comparison of the start up strategies of CANON reactors Several strategies have been tested to start-up and optimize the performance of reactors where autotrophic nitrogen removal takes place (Table 6.4). Among them, two can be pointed out: 1) to inoculate an Anammox reactor with nitrifying biomass and to supply air to maintain microaerobic conditions (Sliekers et al., 2002) or 2) to operate a nitrifying reactor under oxygen limited conditions to obtain the desired ammonium to nitrite molar ratio inside the system and then inoculate Anammox biomass (Pynaert et al., 2004; Gong et al., 2007). Table 6.4. Comparison of several strategies to start up the autotrophic nitrogen removal process. Background Inoculation Initial ANR (g N L-1 d-1) Final ANR (g N L-1 d-1) Start-up time Referencea Anammox reactor Nitrifying sludge 0.2 g VSS L-1 8.9 1.5 0 d [1] Nitrifying granules Anammox 0.12 g VSS L-1 0.05 0.25 35 d This study SBRP Nitrifying Biofilm Anammox 1L 0.04 0.63 40 d [2] Glanerland (400 m3) Nitritation/denitritation Anammox from Strass WWTP 1.25 g VSS L-1 0.15 0.63 55 d [3] 0.4 100 d Nitrifying Biofilm Anaerobic granular sludge 10 g L-1 ~0 1.7 200 d [4] Nitrifying/denitrifying activated sludge - ~0 0.36 1 year [5] Strass (500m3): Nitritation/denitritation Stepwise enrichment 4 L – 0.3 m3 – 2.4 m3 – 500 m3 ~0 0.7 2.5 years [3] a[1] Sliekers et al., (2003) [2] Gong et al., (2007) [3] Wett, (2007) [4] Pynaert et al., (2004) [5] Gaul et al., (2005) As represented in Table 6.4, the second strategy seems to be more suitable because an important decrease of the Anammox activity is observed when the first strategy is applied (Sliekers et al., 2003). The inoculation of Anammox enriched biomass in a nitrifying reactor accelerates the start-up process and allows registering important increases in the ANR after one or two months instead of several months or even years without inoculation. The full-scale example is very illustrative since in Strass, 2.5 years were necessary to startup the plant in front of the 55 days registered in Glanerland once an enriched inoculum was available (Wett, 2007). Gong et al., (2007) multiplied by 16 the ANR in 40 days whereas in our study this increase was three times lower due to the lower temperature of operation. Most of the CANON systems reported in literature were operated at temperatures around 30 ºC. It must be taken into account that, according to the activation energy of the Anammox process, the activity at 24 ºC would be only 40% of that at 35 ºC. The feasibility of Anammox systems at temperatures around 20 ºC was already reported by Isaka et al., (2007), Dosta et al., (2008) in Anammox reactors and Pynaert et al., (2004) in one stage systems. 126 Development of the CANON process to treat anaerobic digester supernatants at room temperature 6.4.6. Comparison of nitrogen removal performances in different reactor configurations As Anammox microorganisms have a low growth rate, a lot of effort was focused on developing systems with biomass retention capacity in order to shorten the start up period of the process (Trigo et al., 2006) and, for this reason, most of the works were carried out under optimum conditions for the Anammox biomass (pH between 7 and 8 and temperature higher than 30 ºC). However, up to now, limited information is available on the efficiency of this process operated under low temperatures. Isaka et al., (2008) and Dosta et al., (2008) adapted respective Anammox systems by gradually decreasing the temperature and, in both cases, a strong decrease of the nitrogen conversion rate was observed. Generally, CANON systems reported in the literature were operated at 30-35 ºC (Table 6.5). In the present study, the granular SBR and the pulsing SBRP were operated at temperatures around 20 ºC, reaching rates of nitrogen removal up to 1.1 g N L-1 d-1 and 0.45 g N L1 d-1 respectively, which are in the range of 0.075 to 1.5 g N L-1 d-1 reported for CANON systems operated at higher temperatures. Different technologies have been used to carry out autotrophic nitrogen removal in one or two units (Table 6.5). Nitrogen removal rates achieved with both configurations are in the same range when the systems are operated at a temperature close to 30 ºC. However, in this study, the use of one single unit allowed the achievement of higher treatment capacities compared to the two units configuration, basically due to the negative effects of low temperatures on Anammox biomass and its necessity of an adaptation period. In general, the use of one single reactor could represent some advantages with respect to the two units configuration such as, lower capital costs and less oxygen consumption by AOB, which prevent possible negative effects on the Anammox bacteria. Nevertheless, the application of the two units configuration would be appropriated when toxic or organic biodegradable compounds are present in the feeding, since these compounds will be degraded in the nitrifying unit avoiding its entrance in the Anammox reactor. Comparing now the granular SBR and the pulsing SBRP, the granular SBR presented higher ANR but also higher DO concentration in the bulk liquid and higher VSS in the reactor. The specific Anammox activity of the biomass was calculated dividing the ANR by the amount of VSS in the reactor, its maximal value was around 0.1 g N (g VSS)-1 d-1 in both systems despite the lower DO concentration in the SBRP. The explanation is that the granular SBR was composed almost exclusively by granules whereas in the SBRP a mixture of granules and flocs was present. Since AOB activity is the limiting step in the reactor, despite the DO concentration is higher in the SBR, its lower specific surface availability due to the higher diameter of the granules difficult the oxygen mass transfer as aforementioned. Therefore, comparing both reactors, it is inferred that the aeration costs in the granular SBR would be higher to achieve the same Anammox specific activity than in the pulsing SBR. On the other hand, the granular biomass of the SBR presented better settling properties than the biomass of the SBRP. From the results obtained, it is inferred that both reactors (the granular SBR and the pulsing SBRP) could be a suitable technology to carry out autotrophic nitrogen removal at moderately low temperatures. The good retention capacity of both reactors and the low VSS concentration in the effluent will minimize the size or even the need of a posterior settler. Finally, the high H/D ratio of the reactors would reduce the surface requirement and would make the technology promising under an economical point of view. More studies are necessary to study the feasibility of a pulsing device at an industrial scale. 127 Chapter 7 7.1. Introduction Nowadays, the removal of nutrients from wastewater is a major concern due to the strengthening of disposal normatives. Wastewaters characterized by low organic matter content and high nitrogen concentrations are difficult to be treated by conventional processes like nitrification-denitrification. In these cases the Anammox process arose as an interesting alternative to treat this type of wastewaters since these bacteria are able to perform the anaerobic oxidation of ammonium with nitrite as electron donor. As a previous step, part of the ammonium has to be oxidized into nitrite. This step can be carried out in a previous reactor, e.g., a Sharon reactor (Mosquera-Corral et al., 2005) or a granular nitrifying reactor (Vázquez-Padín et al., 2009a). Another possibility consists of performing the partial nitrification and the Anammox processes in one single reactor. This process has been given different names: CANON (Third et al., 2001), OLAND (Kuai and Verstraete, 1998) and deammonification (Hippen et al., 1997) processes. Under microaerobic conditions, ammonia oxidizing bacteria (AOB) oxidize ammonium into nitrite consuming the dissolved oxygen (DO) and creating anoxic niches where Anammox bacteria can exist and convert both ammonia and nitrite into nitrogen gas and produce small amounts of nitrate. The optimization of the performance of the autotrophic nitrogen removal in the CANON process requires the control of DO and NO2concentrations. To establish an adequate control of the DO concentration in the liquid media is necessary: 1) to avoid the inhibition of Anammox bacteria caused by DO concentrations higher than 0.5% of air saturation (Strous et al., 1997), 2) to prevent the growth of nitrite oxidizing bacteria (NOB) which have lower affinity for oxygen compared to AOB. The control of NO2concentration is necessary since this compound inhibits Anammox activity although variable ranges of concentrations are provided in the literature for this inhibitory effect. Dapena-Mora et al., (2007) reported that concentrations of nitrite of 350 mg NO2--N L-1 corresponded to 50% inhibition of Anammox bacteria. Due to the slow growth of both nitrifying and Anammox bacteria involved in the CANON process, the use of good biomass retention systems is mandatory to reach significant nitrogen removal rates. In this sense, the development of granular biomass allows accumulating large biomass concentrations in the reactors without the need of carrier material. Moreover, the use of granular biomass allows the existence of substrates gradients, in such way that the external part of granule could be under aerobic conditions while anoxic conditions would maintain in the core of the granule. Therefore, several biological processes can be carried out in the same granule: partial nitrification in the outer part and Anammox process in the inner part. The potential of the granular technology to carry out autotrophic nitrogen removal has been demonstrated (Vlaeminck et al., 2008; Vázquez-Padín et al., 2009a); allowing the treatment of nitrogen loads similar to systems with two different units for partial nitrification and Anammox processes, respectively. Generally, the operational strategy of granular CANON systems is only based on the oxygen and nitrite concentrations measured in the bulk liquid. In this sense, the knowledge of the concentration profiles of these compounds and the distribution of bacterial populations inside the granule would be very useful in order to a better understanding and control of the CANON process. Microsensors, due to their very small dimensions, can be used for the determination of substrate profiles while the distribution of bacterial populations could be determined by microbiological techniques (e.g. FISH, PCR). The combination of microbiological techniques (e.g. FISH, PCR) and measurements using microelectrodes has been used to obtain detailed analysis of the in situ structure and function of nitrifying biofilms. Some examples are: 1) To estimate kinetic parameters which can be further used in mathematical models (Schramm et al., 1999; Kindaichi et al., 2006). 134 Microbial community distribution and activity dynamics of granular biomass in a CANON reactor 2) To study the mass transport of substrates through biofilms or aggregates: determination of the limiting substrate, the active zone of the biofilm and the activities under different substrate concentrations (de Beer et al., 1993; Gieseke et al., 2003). 3) To determine the consumption rates of substrates in non homogeneous biofilm reactors (Schramm et al., 1999; Kindaichi et al., 2007). 4) To corroborate the environmental conditions (e.g., pH) inside the biofilm (Gieseke et al., 2006). 5) To observe the stratification of biomass (Okabe et al., 1999; Kindaichi et al., 2006). Therefore, combining microelectrode measurements and FISH analysis allows gaining information about the substrate concentrations to which the different layers of the biofilm/aggregate are exposed and the microbial populations involved in the different biological processes. All this microscopic information can further be transposed to a macroscopic level, giving valuable information about the possible control strategies of CANON reactors. 7.2. Objectives ● Since little information is known from a microscopic point of view about granules performing complete autotrophic nitrogen removal and taking into account the relevance of DO and NO2concentrations, the objectives of this study were: the identification of the main bacteria populations present in the granules, the determination of their distribution inside the granules and the estimation of their activities by combining the concentrations profiles measured with microelectrodes and FISH images taken from cryosectioned slices of granules. From the combination of the previously obtained results a better insight about the performance of the completely autotrophic nitrogen removal granules was obtained. 7.3. Materials and Methods 7.3.1. Reactor description The growth of biomass in the form of granules performing the CANON process was described elsewhere (Chapter 6, Vázquez-Padín et al., 2009a). Complete nitrification to nitrate and later partial nitrification to nitrite were carried out by regulation of the DO concentration in the bulk liquid. Finally, Anammox bacteria were grown in the anoxic core of the granules to form the CANON granules. Five months after the macroscopic evidence of Anammox activity in CANON granules (day 330, chapter 6), 4 g VSS of granular biomass from the reactor operated at the University of Santiago de Compostela were inoculated in a sequencing batch reactor (SBR) with a working volume of 0.7 L at the University of Aarhus. This new reactor was operated in cycles of 3 h distributed as: 175 min of aeration and feeding, 1 min of settling and 4 min of effluent withdrawal. The hydraulic retention time (HRT) was fixed at 0.25 d and the exchange volume was fixed at 30%. The reactor was operated at room temperature which ranged between 19 and 22 ºC. The pH value was not controlled and ranged from 7.0 to 8.1 with a mean value of 7.6 ± 0.2. Air was supplied through a diffuser at the bottom of the reactor by using an air pump to promote the transfer of oxygen into the bulk liquid and to reach a suitable mixing. The average DO concentration in the reactor was 6.6 ± 0.5 mg O2 L-1 and represented the main difference with the operational conditions from the reactor operated at the laboratory in Santiago de Compostela where the DO remained around 3.5 mg O2 L-1. 135 Chapter 7 The CANON SBR was fed with the supernatant of the anaerobic sludge digester of the WWTP of Aarhus (Denmark) which was stored in a cold room (4 ºC). The composition of the supernatant was: pH of 7.6―8.3; ammonium concentration of 256―666 mg N L-1; inorganic carbon (IC) of 232―374 mg IC L-1 and total organic carbon (TOC) of 103―150 mg TOC L-1. Figure 7.1. Photographs of the granular SBR. 7.3.2. kLa measurement An experimental estimation of the oxygen gas-liquid transfer coefficient (kLa) was carried out by means of a dynamic method (described in chapter 2), registering the increments of DO concentrations in the SBR after the reestablishment of the aeration (without biomass in the reactor). The value of the kLa for dissolved oxygen obtained in the reactor was of 1 min-1, a similar value to the one obtained in the SBR installed in Santiago de Compostela. 7.3.3. Microscale experiments The microelectrodes were used to measure the concentration of nitrite and dissolved oxygen at different depth positions inside the granules performing the CANON process and to obtain the corresponding microprofiles. The granules were collected directly from the CANON SBR and inserted on a needle supported on a metal grid inside the experimental chamber. In order to simulate the hydrodynamic conditions from the reactor, the aeration volume in the chamber was regulated to a value which maintained the kLa (gas-liquid mass transfer coefficient) at 1 min-1 which was the value obtained in the SBR. All the microprofiles were measured in granules with average diameters of 5 mm. The mean temperature of the aerated chamber was 20 ± 1 ºC. Granules were kept for 1 h inside the chamber as a pre-incubation period to create pseudo steady state conditions. Concentration profiles were recorded by introducing the sensors into the granules at different depth positions using a manual micromanipulator. A dissection microscope was used to visually estimate the position of the granule/water interface by visual observation. For each granule and experimental condition tested several microprofiles were measured (the number of microprofiles performed is described in the legend of figures as n). Microprofiles of dissolved oxygen were performed by measuring its concentration at depth intervals of 25 µm while in the case of the microprofiles of NO2the measurements were performed at 50 or 100 µm due to the slower response time of this sensor. The sensor signal was continuously recorded on a strip-chart recorder. 136 Microbial community distribution and activity dynamics of granular biomass in a CANON reactor The liquid media inside the chamber was the anaerobic digester supernatant diluted with tap water. The concentration of ammonium was maintained constant in all experiments at 140 mg N L-1 in order to avoid ammonium limitation in the experiments and the nitrite concentrations varied from 0.7 mg NO2--N L-1 to 42 mg NO2--N L-1 by means of NaNO2 addition to the diluted supernatant. The microsensors were calibrated previous to each experiment using different nitrite concentrations in the medium. Dissolved oxygen concentrations in bulk liquid were varied between 1.5 and 35.2 mg O2 L -1. These different DO concentrations were achieved by flushing a mixture of pure O2 and air for DO values in the aeration chamber above air saturation or a mixture of air with N2 for values under air saturation. In order to maintain a constant kLa, in all the experiments the total gas flow (N2/air/O2) was kept constant. 7.3.4. NO2and O2 microsensors Microelectrodes were used to determine concentrations of the measured compounds inside the granules due to their small sizes. A Clark-type O2 microsensor equipped with a guard cathode was used for microscale analysis of DO (Revsbech, 1989). These sensors were constructed with tip diameters of 10 µm and they had 90% response time lower than 1 s. The oxygen microsensor was calibrated at two points, the zero was established dropping the microelectrode in an anoxic alkaline ascorbic acid solution and the saturation was obtained or with an air saturated solution or with pure oxygen saturated solution depending on the DO analyzed in the aeration chamber. The NO2microelectrode was a biosensor, i.e., it was based on a combination of biological reactions catalyzed by bacteria and an electrochemical N2O microsensor. The NO2microsensor consisted of an immobilized pure culture of Stenotrophomonas nitritireducens, which reduced NO2to N2O, coupled to a Clark type N2O microsensor. The 90% response time was 30-50 s. Due to the broad range of NO2concentrations in the bulk liquid used in the performed experiments from 0.7 mg NO2--N L-1 to 42 mg NO2--N L-1 it was necessary to add tungstate inside the biosensor in order to measure the highest NO2concentrations. The tungstate increased the stability of the microsensor and slowed down the nitrite consumption rate of the bacteria (Nielsen et al., 2005). 7.3.5. Analytical methods The pH and the concentrations of DO, ammonia, volatile suspended solids (VSS), and sludge volumetric index (SVI5) were determined according to the Standard Methods (APHA-AWWA-WPCF, 1998). Nitrite and nitrate concentrations were determined by capillary electrophoresis (Vilas-Cruz et al., 1994). Concentrations of TOC and IC were measured with a Shimadzu analyser (TOC-5000). Density of the granules was measured using the dextran blue method described by Beun et al., (2002). The morphology and size distribution of the granules were measured regularly by using an image analysis procedure with a stereomicroscope (Stemi 2000-C, Zeiss) provided with a digital camera (Coolsnap, Roper Sicientific Photometrics). For the digital image analysis the programme Image Pro Plus was used. In order to identify bacterial populations of ammonia oxidizing bacteria (AOB), nitrite oxidizing bacteria (NOB) and Anammox bacteria, granules from the reactor were collected, kept in their aggregated form or disaggregated, and fixed according to Amann et al., (1995) with 4% paraformaldehyde solution. Entire granules were embedded in OCT reagent (Tissue-Tek; Miles, Ind.) prior to their cryosectioning at -35°C. Slides with a thickness of 14 µm were cut at -16 °C, and these single sections were placed on the surface of poly-L-lysine 137 Chapter 7 coated microscopic slides. Hybridization was performed at 46 ºC for 90 minutes adjusting formamide concentrations at the percentages shown in Table 7.1. The used probes for in situ hybridization were 5’ labelled with the fluorochromes FLUOS, Cy3 or Cy5. A TCS-SP2 confocal laser scanning microscope (Leica, Germany), equipped with a HeNe laser for detection of Cy3 and Cy5 and one Ar ion laser for detection of FLUOS, was used with the sliced samples. The method to quantify bacterial populations was based on the one published by Crocetti et al., (2002). The digital image analysis program used was Image ProPlus. For the quantification of bacteria populations, cryosectioned slices of granules were used to perform a triple hybridization with EUBmix (a mixture of EUB338, EUB338 II and EUB338 III) labelled with Cy5, Amx820 labelled with Cy3 and NEU653 labelled with FLUOS. Several pictures were taken subsequently from the granule surface throughout the active layers with the confocal microscope at a magnification of 630 times. The pictures were superimposed to represent the whole active layers and then discretized in layers of 100 µm with the help of Adobe Photoshop® software. Form each image three different colour components corresponding to each fluorochrome were separated generating three different images. The area corresponding to the fluorescence of each FISH probe was obtained as the area of all pixels above a manually determined minimum pixel intensity. In order to be able to compare the areas occupied by the different populations, the maximal area occupied in a discretized picture was taken as reference and the values obtained with the different probes in the different layers were obtained as normalized area values. Further information about the analytical methods is provided in Chapter 2. Table 7.1. Targeted organisms and the corresponding formamide (FA) percentages for the used oligonucleotide probes. Probea Probe sequence (5’3’) % FA Targeted organisms EUB338 GCT GCC TCC CGT AGG AGT 0-50 Domain bacteria EUB338 II GCA GCC ACC CGT AGG TGT 0-50 Planctomycetales EUB338 III GCT GCC ACC CGT AGG TGT 0-50 Verrucomicrobiales Nso190 CGA TCC CCT GCT TTT CTC C 55 Ammonia-oxidizing - Proteobacteria NEU653b CCC CTC TGC TGC ACT CTA 40 Most of the halophilic and halotolerant Nitrosomonas spp. Ntspa712b CGC CTT CGC CAC CGG CCT TCC 50 Most members of the phylum Nitrospira NIT3b CCT GTG CTC CAT GCT CCG 40 Nitrobacter spp. Amx820 AAA ACC CCT CTA CTT AGT GCC C 40 Anaerobic ammonium-oxidizing bacteria Candidatus Brocadia anammoxidans and Candidatus Kuenenia stuttgartiensis Kst157 GTT CCG ATT GCT CGA AAC 25 Candidatus Kuenenia stuttgartiensis Ban162 CGG TAG CCC CAA TTG CTT 40 Candidatus Brocadia anammoxidans a Details on oligonucleotide probes are available at probeBase (Loy et al., 2007). b Used with an equimolar amount of corresponding unlabeled competitor oligonucleotide probe. 7.3.6. Calculations 7.3.6.1. Estimation of oxygen and nitrite consumption and production rates Mass balances were calculated for oxygen and nitrite in the granules assuming a flat geometry was assumed to describe the shape of the granules. The activities of the processes involved in the nitrogen removal inside the granules were restricted to an external layer of 1 mm (as it will be further discussed) which was smaller than the mean radius of the granules used to record the microprofiles (mean radius of 2.5 mm). 138 Microbial community distribution and activity dynamics of granular biomass in a CANON reactor The balance of component for a flat geometry in a solid matrix can be written according to Eq. 7.1 assuming that the compound mass transfer is carried out only by diffusion and that the granule has an homogeneous structure since the diffusivity (D) is considered constant (Lorenzen et al., 1998).   R(z) z tz,C D t tz,C 2 2       (7.1) where C is the concentration of compound (g L-1), z the depth coordinate in the granule (dm), t the time (d) and R the reaction rate (g (Lgranule)-1 d-1). Considering steady state conditions the Eq. 7.2 is obtained.  R(z) z zC D2 2    (7.2) making A(z) = -R(z)/D and using Euler’s formula for numeric integration the following Eq. 7.3 is obtained. n n1n Ah z C z C       (7.3) where h represents the step size used for numerical integration. This discretization parameter was of 25 µm for oxygen profiles and of 50 µm for nitrite concentrations in the bulk liquid smaller than 2.8 mg NO2--N L-1 and of 100 µm for higher NO2concentrations. After further integration Eq. 7.4 is obtained. n n1n z C hCC     (7.4) Substituting Eq. 7.3 in Eq. 7.4, the Eq. 7.5 is obtained.              1n 1n n1n Ah z C hCC (7.5) Using the Solver tool (available in Microsoft Excel® software) the values of A were iterated in order to minimize the error between the concentration calculated with Eq. 7.2 and that one measured with the microelectrode. By multiplying the value of A obtained in a discretization point by the diffusivity of the substrate desired, the value of R(z), i.e., the local volumetric consumption rate can be obtained. The diffusion coefficients of NO2and O2 in water at 20 ºC were chosen as 1.5·10-4 and 1.7·10-4 m2 d-1, respectively (Picioreanu et al., 1997). To calculate the net fluxes of substrates (J, g N m-2 d-1) through the diffusive boundary layer (DBL) separating the surface of the granule and the bulk liquid, the Fick’s first law was used (Eq. 7.6). h sb Wδ CC DJ   (7.6) being Dw the molecular diffusion coefficient in water (m2 d-1), Cb is the bulk liquid concentration (g m-3), CS is the concentration (g m-3) at the surface of the granule and δh is the hypothetical (also called effective) diffusive boundary layer thickness (m) which is defined by extrapolating the radial oxygen gradient at the granule-water interface to the bulk water phase concentration (Ploug et al., 1997). 139 Chapter 7 7.3.6.2. Nitrogen removal rates Ammonia oxidation rates (AOR) and nitrogen removal rates by Anammox bacteria (ANR) of the CANON granular reactor were estimated as g N L-1 d -1 based on nitrogen balances and the stoichiometry of the Anammox process (1.02 moles of dinitrogen gas produced per mole of ammonium reacted). ))NNO()NNO()NNH(()NNH(N eff3eff2eff4inf4        (7.7) HRT 04.2 N )NNH()NNH( AOR eff4inf4     (7.8) HRT N ANR   (7.9) being, N the difference between total nitrogen concentration in the influent and effluent (g N L-1), NH4+-Ninf the ammonium concentration in the influent (g N L-1) and NH4+-Neff, NO2--Neff, NO3--Neff the ammonium, nitrite and nitrate concentrations in the effluent (g N L-1), respectively. 7.3.6.3. Estimation of the number of granules in the SBR The number of granules in the reactor was calculated as follows: Vgranule = VR·XR / ρgranule (7.10) nT = Vgranule / (4/3·π·Rm3) (7.11) being Vgranule the volume of granules (L), VR the reactor volume (L), XR the VSS concentration in the reactor (g VSS L-1), ρgranule the granules density (g VSS (Lgranule)-1), nT the number of granules, Rm the average radius of the granules (dm). In order to estimate either the AOR or the ANR from the microsopic observations, Eq. 7.12 was used to the zone were AOB or Anammox bacteria were located.  n 33 Tz z=1 R 4 n R(z) πR-R 3 Rate = V    z-1 (7.12) Where the rate is calculated in g (Lreactor)-1 d-1, R(z) is the local reaction rate (g (Lgranule)-1 d-1), VR is the reactor volume (L), and R corresponds to the radius (dm). 7.4. Results and discussion 7.4.1. Operation of the CANON SBR The SBR was operated to keep the granular biomass performing the partial nitrification and the Anammox processes active in order to be used for the determination of the O2 and NO2profiles with the microelectrodes. The SBR reactor was operated at DO concentrations around 6.6 mg O2 L -1 and at a temperature of 20 ºC. The mean NO2concentration in the effluent was 25 mg N L-1. The total nitrogen removal rate (Fig. 7.2) ranged between 0.35 and 0.91 g N L-1 d -1. Those values are between the highest ones registered for autotrophic nitrogen removal in one reactor despite the low temperature of operation (in Chapter 6, a comparison of the nitrogen removal rates obtained in CANON systems can be found, Table 6.5). 140 Microbial community distribution and activity dynamics of granular biomass in a CANON reactor According to Dosta et al., (2008) the specific activity of Anammox biomass at 37 ºC is 3.6 times higher than at 20 ºC, therefore, the operation at higher temperature would allow achieving larger biomass activities. 0.0 0.3 0.6 0.9 1.2 1.5 0 102030405060 Time (d) NLR, AOR, ANR (g N L -1 d -1 ) ison to the autotrophic nitroge velocity of the granular sludge was of 110 m h-1. The number of granules in the reactor was estimated as 2233 (Eq. 7.11) using a value of density of 29 g VSS (Lgranule)-1 and the aforementioned biomass concentration. Figure 7.2. Nitrogen loading rate applied (▬) Ammonia oxidation rate (○), nitrogen removal rate (▲). The removal of total organic carbon accounted for 0.11 g TOC L-1 d-1. Since this value is low and the main part of this carbon would be removed aerobically or it is expected to be recalcitrant, the amount of nitrogen removed by denitrifying heterotrophic bacteria can be neglected in compar n removal. For this reason, heterotrophic denitrification was not considered in the nitrite balance (Eq. 7.8) and the totality of the removed nitrogen was considered due to Anammox activity. The biomass concentration inside the reactor remained almost constant at 5.7 g VSS L-1 during the 60 days of operation. The average diameter of the granules in the reactor was of 5 mm (Fig. 7.3). The value of the SVI5 was 25 mL (g VSS)-1 and the settling Figure 7.3. a) Photograph of several granules in a Petri plate b) Size comparison of some granules with a 1 cent of € (Diameter of the 1 cent of € = 1.6 cm). 141 Chapter 7 The biomass concentration inside the reactor remained almost constant at 7.5 g VSS L-1 during the 60 days of operation. The average diameter and density of the granules were 5 mm and 36 g VSS (Lgranule)-1, d using Eq. 7.11 giving the following result: the settling velocity of the granular sludge was of 110 es NEU653 and Nso190. Bacteria belonging to the genus Nitrosomonas spp. were still present until a depth of 600 µm but their proportion decreased along the depth (Fig. 7.5). Nitrosomonas spp. were determined as the main AOB community as expected due to conditions of ammonium excess (Schramm et al., 1998). respectively. With these data, the number of granules was estimate 2230 granules. The value of the SVI was 25 mL (g VSS)-1 and m h-1. 7.4.2. Identification of bacteria populations by FISH The stratification of the AOB and the Anammox bacteria in depth inside the granule can be observed in Fig. 7.4. In the outermost layer of the granule of 200 µm almost all the biomass was composed exclusively by Nitrosomonas spp. which gave positive signals with prob Figure 7.4. Image of a cryosectioned slice of a granule with a triple hybridization of FISH probes: NEU653 (light g No significant fluorescence signal was detected with probe NIT3 specific for Nitrobacter spp. and neither with probe Ntspa712 specific for Nitrospira phylum. Therefore, no NOB activity was expected in the granules despite the high DO concentration in the liquid bulk. reen, FLUOS); AMX820 (pink, Cy3) and EUBmix (blue, Cy5).The right part of the picture corresponds to the surface of the granule (the bar corresponds to 75 µm and the total width of the image corresponds to 1100 µm). -1.0 -0.8 -0.6 -0.4 -0.2 0.0 Normalized area 0.0 0.2 0.4 0.6 0.8 1.0 Depth (mm) Amx AOB EUB -1.2 Figure 7.5. Depth distribution of AOB populations (hybridized with probe NEU653), Anammox bacteria (hybridized with probe Amx820) and all bacteria (hybridized with EUBmix) inside the granule. The value of depth equal to 0 mm corresponds to the granule surface. 142 Microbial community distribution and activity dynamics of granular biomass in a CANON reactor Anammox bacteria were mainly located between 400 and 1000 µm of depth inside the granule where absence of dissolved oxygen is ensured. Bacteria belonging to the genus Candidatus Kuenenia stuttgartiensis were identified as the main Anammox bacteria in the reactor through positive results with probe Kst157. No positiv with NEU653 and g deeper inside the granule and they could correspond to those e metabolic compounds (Kindaichi et al., 2004). 7.4.3.1. Oxygen microprofiles in the partial nitrification zone e results with probe Ban162 were obtained indicating absence of Candidatus Brocardia anammoxidans. Therefore, in the depths ranging from 400 µm to 600 µm AOB and Anammox bacteria coexist. It is also interesting to point out that the area corresponding to EUBmix probe which represent all bacteria decreased along the granule, therefore, the activity of the granule is mainly located in the first 1000 µm which is in contact with the bulk liquid. Other bacteria (marked with EUBmix but not Amx820) increased in proportion by goin heterotrophs which possibly grew on solubl 7.4.3. Microprofiles measurements DO cons. rate (g O 2 (L granule ) -1 d -1 ) 0255075100 Depth (mm) -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 Oxygen (mg O 2 L -1 ) 0.0 0.5 1.0 1.5 2.0 DO cons. rate (g O 2 (L granule ) -1 d -1 ) 0255075100 Depth (mm) -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 Oxygen (mg O 2 L -1 ) 0.0 0.6 1.2 1.8 2.4 DO cons. rate (g O 2 (L granule ) -1 d -1 ) 0 255075100 Depth (mm) -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 Oxygen (mg O 2 L -1 ) 0.0 1.5 3.0 4.5 6.0 DO cons. rate (g O 2 (L granule ) -1 d -1 ) 0 25 50 75 100 Depth (mm) -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 Oxygen (mg O 2 L -1 ) 0.0 2.5 5.0 7.5 10.0 Oxygen (mg O 2 L -1 ) DO cons. rate (g O 2 (L granule ) -1 d -1 ) 0 255075100 Depth (mm) -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 0 6 12 18 24 Oxygen (mg O 2 L -1 ) 0 10203040 DO cons. rate (g O 2 (L granule ) -1 d -1 ) 0 255075100 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3 Depth (mm) Figure 7.6. DO concentration profiles (●) and local consumption rates (■) under different DO concentrations in the bulk liquid (number of microprofiles (n): n=3 for all cases except for DO=8 mg O2 L-1 where n=20). 143