Microbial resource management of OLAND focused on sustainability by Haydée De Clippeleir

Promotors: Prof. dr. ir. Willy Verstraete Department of Biochemical and Microbial Technology, Faculty of Bioscience Engineering, Ghent University, Gent, Belgium Prof. dr. ir. Nico Boon Department of Biochemical and Microbial Technology, Faculty of Bioscience Engineering, Ghent University, Gent, Belgium

Members of the examination committee: Prof. dr. ir. Pascal Boeckx Department of Applied analytical and physical chemistry, Faculty of Bioscience Engineering, Ghent University, Gent, Belgium Prof. dr. Juan M. Lema Department of chemical engineering, School of engineering, University of Santiago de Compostela, Santiago de Compostela, Spain Dr. Bernhard Wett ARAconsult GmbH, Innsbruck, Austria Prof. dr. ir. Ingmar Nopens (Secretary) Department of Mathematical Modelling, Statistics and Bioinformatics, Faculty of Bioscience Engineering, Ghent University, Gent, Belgium Peter Bossier (Chairman) Departement of animal production, Faculty of Bioscience Engineering, Ghent University, Gent, Belgium

Dean: Prof. dr. ir. Guido Van Huylenbroeck Rector: Prof. dr. Paul Van Cauwenberge

ir. HaydĂŠe De Clippeleir

Microbial resource management of OLAND focused on sustainability

Thesis submitted in fulfillment of the requirements for the degree of Doctor (PhD) in Applied Biological Sciences

Dutch translation of title: Microbial resource management van OLAND met de focus op duurzaamheid

This work was supported by the Institute for the Promotion of Innovation by Science and Technology in Flanders (IWT-Vlaanderen, number SB-81068).

Cover illustration: "Energieweelde" made by Lutgarde Van Hoey, based on a design of Kantschool Artofil, Nadine Pauwels. Photography by Verne.

To refer to this thesis: De Clippeleir, H. (2012) Microbial resource management of OLAND focused on sustainability. PhD thesis, Ghent University, Belgium.

ISBN: 978-905989-551-5

The author and the promotors give the authorisation to consult and to copy parts of this work for personal use only. Every other use is subject to the copyright laws. Permission to reproduce any material contained in this work should be obtained from the author.

Notation index

Notation index A/B process

2-stage activated sludge system

anaerobic digestion

ADP

abiotic depletion potential

AerAOB

aerobic ammonium-oxidizing bacteria

AnAOB

anoxic ammonium-oxidizing bacteria

activated sludge

AX/B system

A/B process with X% COD removal in the A-stage

bCOD

biodegradable fraction of the chemical oxygen demand

BOD

biological oxygen demand

CAS

conventional activated sludge system

CHP

combined heat and power

COD

chemical oxygen demand

CSTR

continuous stirred tank reactor

DEMON

OLAND with pH controlled aeration

dry matter

dissolved oxygen

E-index

energy index, ratio produced over consumed electricity

EUP

eutrophication potential

free ammonia

FET

freshwater ecotoxicity

FISH

fluorescent in-situ hybridization

FNA

free nitrous acid

GHG

greenhouse gases

GWP

global warming potential

HRM

humane resource management

HRT

hydraulic retention time

inhabitant equivalent

LCA

life cycle assessment

MABR

membrane aerated bioreactor

MBBR

moving bed bioreactor

Notation index MBR

moving bed reactor

MRM

microbial resource management

narr

nitrite accumulation rate ratio

N/DN

nitrification/denitrification

NOB

nitrite-oxidizing bacteria

ozone depletion potential

OFMSW

organic fraction of municipal solid waste

OLAND

oxygen-limited autotrophic nitrification/denitrification

person equivalent

photochemical oxidation potential

RBC

rotating biological contactor

SBR

sequencing batch reactor

SRT

sludge retention time

suspended solids

TET

terrestrial ecotoxicity

TSS

total suspended solids

UASB

upflow anaerobic sludge blanket

VSS

volatile suspended solids

WWTP

wastewater treatment plant

Table of contents

Table of contents PART I: Introduction Chapter 1: Introduction.......................................................................................................... 3 1

Autotrophic nitrogen removal ...................................................................................................... 3

Cost and energy effectiveness of OLAND ................................................................................... 5

OLAND design parameters ............................................................................................................. 7

3.1

Choice of reactor technology................................................................................................................... 7

3.2

Key control mechanism to obtain stable performance ............................................................. 11

3.3

OLAND performance ............................................................................................................................... 12

Microbial resource management (MRM) ............................................................................... 14 4.1

Maximizing nitrogen removal efficiency ......................................................................................... 14

4.2

Minimizing harmful gas emissions .................................................................................................... 17

4.3

OLAND enabling energy positive sewage treatment ................................................................. 21

Objectives and outlines of this research ................................................................................ 24

PART II: MRM output optimization Chapter 2: A low volumetric exchange ratio allows high autotrophic nitrogen removal in a sequencing batch reactor ................................................................................................ 29 1

Introduction ..................................................................................................................................... 30

Materials and methods ................................................................................................................. 31

2.1

OLAND SBR.................................................................................................................................................. 31

2.2

SBR cycle....................................................................................................................................................... 32

2.3

Aerobic and anoxic batch tests ............................................................................................................ 32

2.4

Chemical analyses ..................................................................................................................................... 32

2.5

Physical aggregate characteristics ..................................................................................................... 33

Results ................................................................................................................................................ 33 3.1

OLAND SBR performance ...................................................................................................................... 33

3.2

Biomass morphology ............................................................................................................................... 34

3.3

Control of the microbial balance in the reactor ........................................................................... 36

Discussion ......................................................................................................................................... 37 4.1

OLAND SBR performance ...................................................................................................................... 37

4.2

Biomass morphology ............................................................................................................................... 38 iii

Table of contents 4.3

Control of the microbial balance in the reactor ........................................................................... 39

Conclusions ....................................................................................................................................... 39

Acknowledgements ........................................................................................................................ 39

Chapter 3: Interplay of intermediates in the formation of NO and N2O during full-scale partial nitritation/anammox .................................................................................................. 41 1

Introduction ..................................................................................................................................... 42

Materials and methods ................................................................................................................. 43 2.1

Reactor operation ..................................................................................................................................... 43

2.2

Emission measurments .......................................................................................................................... 43

Results and discussion .................................................................................................................. 45

Conclusions ....................................................................................................................................... 51

Acknowledgements ........................................................................................................................ 52

PART III: Exploration of new applications Chapter 4: OLAND maximizes net energy gain in technology schemes with anaerobic digestion................................................................................................................................... 55 1

Treatment of digestates by OLAND .......................................................................................... 55 1.1

Organic fraction of municipal solid waste (OFMSW) ................................................................. 57

1.2

Manure-based agricultural waste ...................................................................................................... 60

1.3

Sugar/starch-based agro-industrial waste .................................................................................... 63

1.4

Sewage-based organics .......................................................................................................................... 64

1.5

Treatment of digestates by OLAND: conclusions and perspectives .................................... 72

OLAND as mainstream treatment process ............................................................................ 73 2.1

Wastewater as an energy resource ................................................................................................... 74

2.2

Main stream OLAND application: conclusions.............................................................................. 76

General conclusions....................................................................................................................... 76

Acknowledgements ........................................................................................................................ 77

Chapter 5: Efficient total nitrogen removal in an ammonia gas biofilter through highrate OLAND............................................................................................................................ 79

Introduction ..................................................................................................................................... 80

Materials and methods ................................................................................................................. 83

Table of contents

2.1

Biofilter set-up and operation ............................................................................................................. 83

2.2

Profile measurements ............................................................................................................................. 83

2.3

Activity batch test ..................................................................................................................................... 83

2.4

Chemical analyses ..................................................................................................................................... 84

2.5

Quantification with real-time PCR ..................................................................................................... 85

Results ................................................................................................................................................ 85 3.1

Performance of the biofilter ................................................................................................................. 85

3.2

Vertical distribution of microbial activity ...................................................................................... 89

3.3

Vertical abundance of N species ......................................................................................................... 90

Discussion ......................................................................................................................................... 91 4.1

OLAND application for NH3 treatment ............................................................................................. 91

4.2

AnAOB niche in NH3 biofilters ............................................................................................................. 92

4.3

OLAND: gas versus water treatment ................................................................................................ 94

Conclusions ....................................................................................................................................... 94

Acknowledgements ........................................................................................................................ 94

Chapter 6: OLAND is feasible to treat sewage-like nitrogen concentrations at low hydraulic residence times ...................................................................................................... 95 1

Introduction ..................................................................................................................................... 96

Material and methods ................................................................................................................. 100

2.1

OLAND rotating biological contactor (RBC) ................................................................................100

2.2

Reactor operation ...................................................................................................................................100

2.3

Chemical analyses ...................................................................................................................................100

2.4

Fluorescent in-situ hybridization (FISH) ......................................................................................101

2.5

Denaturing Gradient Gel Electrophoresis (DGGE) ....................................................................101

Results ..............................................................................................................................................102 3.1

Treatment of high nitrogen levels....................................................................................................102

3.2

Treatment of low nitrogen levels .....................................................................................................102

3.3

Suppression of nitratation at low nitrogen levels .....................................................................102

Discussion .......................................................................................................................................106 4.1

OLAND removal rate and efficiency treating low nitrogen levels ......................................106

4.2

Role of DO levels in suppressing nitratation ...............................................................................106

4.3

OLAND operation at low HRT ............................................................................................................107

4.4

Implementation of OLAND in the main stream ..........................................................................108

Acknowledgements ...................................................................................................................... 108 v

Table of contents

Chapter 7: Cold OLAND on pretreated sewage: feasibility demonstration at lab-scale ................................................................................................................................. 109 1

Introduction ...................................................................................................................................110

Materials and methods ............................................................................................................... 111

2.1

OLAND rotating biological contactor (RBC) ................................................................................111

2.2

RBC operation ..........................................................................................................................................112

2.3

Detection of AerAOB, NOB and AnAOB with FISH and qPCR ...............................................112

2.4

Detailed reactor cycle balances.........................................................................................................113

2.5

Chemical analyses ...................................................................................................................................113

Results ..............................................................................................................................................114 3.1

Effect of temperature decrease .........................................................................................................114

3.2

Effect of COD/N increase .....................................................................................................................118

3.3

Nitratation and NO/N2O emissions .................................................................................................121

Discussion .......................................................................................................................................124 4.1

Effect of temperature decrease .........................................................................................................124

4.2

Effect of COD/N increase .....................................................................................................................125

4.3

NOB-AnAOB competition at mainstream conditions ...............................................................126

4.4

OLAND application in the main line ................................................................................................127

Conclusions .....................................................................................................................................127

Acknowledgements ...................................................................................................................... 128

Supplementary data .................................................................................................................... 128

Chapter 8: Environmental assessment of one-stage partial nitritation/anammox implementation in sewage treatment plants ...................................................................... 133 1

Introduction ...................................................................................................................................134

Materials and methods ............................................................................................................... 136

4 vi

2.1

Scope definition .......................................................................................................................................136

2.2

Plant description .....................................................................................................................................137

2.3

Data inventory..........................................................................................................................................141

2.4

Impact assessment .................................................................................................................................142

Results and discussion ................................................................................................................ 143 3.1

Impact of nitrogen removal process on process level .............................................................143

3.2

From energy-negative to energy-positive WWTP on system level ....................................145

3.3

Environmental impact of DEMON implementation on life cycle level ..............................147

Conclusions .....................................................................................................................................154

Table of contents 5

Acknowledgements ...................................................................................................................... 154

Supplementary data .................................................................................................................... 155

PART IV: General discussion Chapter 9: General discussion and perspectives .............................................................. 159 1

Main outcome and positioning of this work .......................................................................159

OLAND and sustainability ..........................................................................................................160

2.1

Balancing energy recovery with sustainability ..........................................................................160

2.2

Mitigation strategies based on chemical markers.....................................................................161

2.3

Mitigation strategies which minimize emission ........................................................................164

Energy positive WWTP: reality or fantasy?.........................................................................164 3.1

Water-energy nexus...............................................................................................................................164

3.2

Is OLAND an essential treatment step? .........................................................................................166

3.3

Decision making for the wastewater engineer ...........................................................................170

Nitrogen removal versus nitrogen recovery ......................................................................171

Future challenges and opportunities .................................................................................... 173

5.1

Future challenges for mainstream OLAND ..................................................................................173

5.2

OLAND biofilter application ...............................................................................................................174

5.3

What are the temperature limits of the OLAND process ........................................................175

Conclusions .....................................................................................................................................176

PART V: Appendices Abstract ................................................................................................................................. 181 Samenvatting ........................................................................................................................ 185 Bibliography ......................................................................................................................... 189 Curriculum vitae .................................................................................................................. 207 Dankwoord............................................................................................................................ 215

vii

Lab-scale OLAND rotating biological contactor (RBC senior, LabMET) 2

Chapter 1

Chapter 1: Introduction 1

Autotrophic nitrogen removal

Several new biological nitrogen removal processes have been developed to treat nitrogen-rich wastewaters devoid in carbon such as digestates (Table 1.1). These processes are mostly composed out of two main conversion steps; hence a one-step or two-step configuration of the processes is possible. Performing autotrophic nitrogen removal in two stages implies that both process steps should be optimized and controlled individually. In contrast, the investment cost and the difficulty to balance both steps are decreased when operating the process in one step. Moreover,

full-scale

application

studies

showed

that

for

the

one-step

partial

nitritation/anammox process harmful emission of NO and N2O could be decreased to 1 and 0.001%, respectively (Desloover et al., 2011b). Because of these advantages, the full-scale applications are all becoming one-step configurations. Table 1.1: Overview on terminology of one-stage and two-stage autotrophic nitrogen removal processes based on partial nitritation and anammox and indication of the amount of full-scale plants operational at this moment. MBBR: moving bed bioreactor, SBR: sequencing batch reactor; RBC: rotating biological contactor Process name Stages Patent Plants Reference SHARON-ANAMMOX 2 Yes 3 (van der Star et al., 2007) 速 ANAMMOX 1 No 12 (Abma et al., 2010) ANITATM MOX 1 Yes 2 (Christensson et al., 2011) Deammonification MBBR 1 No 2 (Beier and Schneider, 2008) Deammonification SBR 1 No 3 (Joss et al., 2009) TM Cleargreen 1 No 1 (Jeanningros et al., 2010) DEMON速 1 Yes 20 (Wett, 2006) OLAND RBC 1 No 1 (Kuai and Verstraete, 1998)

Chapter redrafted after: Vlaeminck, S.E., De Clippeleir, H., Verstraete, W., 2012. Microbial resource management of one-stage partial nitritation/anammox. Microbial Biotechnology. Microbial Biotechnology, 5, 433-488. 3

Introduction

Therefore, in this work we will focus on the one-step partial nitritation/anammox process, also known as deammonification but generally referred to as the oxygen-limited autotrophic nitrification/denitrification (OLAND) process in this thesis. An overview of the terminology used for pilot and full-scale applications is given in Table 1.1.

Oxygen-limited autotrophic nitrification/denitrification (OLAND) is a one-step nitrogen removal process based on partial nitritation, performed by aerobic ammonium-oxidizing bacteria (AerAOB) and anammox, performed by anoxic ammonium-oxidizing bacteria (AnAOB; Fig. 1.1). The AerAOB, mainly belonging to Nitrosomonas europaea eutropha and halophila (Vlaeminck et al., 2010), are set so that they oxidize half of the influent ammonium to nitrite in oxygen-limited conditions (Eq 1; Table 1.2). The AnAOB, mainly members of the Candidatus genera Kuenenia and Brocadia (van der Star et al., 2007; Vlaeminck et al., 2010), oxidize the residual ammonium with nitrite to dinitrogen gas under anoxic conditions (Eq. 3, Table 1.2). Consequently, in the OLAND process ammonium is converted mainly into nitrogen gas without the use of organic carbon in one reactor. The overall stoichiometry shows that if the AerAOB and AnAOB activity is well balanced, only 11% of the converted ammonium is converted to nitrate due to growth of the AnAOB. Higher nitrate formation (> 11%) implies that nitrite oxidation by nitrite-oxidizing bacteria (NOB) can take place, probably

due

excess

oxygen

(Fig.

1.1).

Lower

nitrate

production

(< 11%) can occur when denitrification can take place due to the presence of organic carbon. Overall nitrogen removal efficiencies obtained in full-scale one-step process are between 80 and 95% (Table 1.4), depending on the presence of COD.

Figure 1.1: Schematic overview of the balanced and imbalanced output caused by the oxic and anoxic reactions during OLAND by aerobic and anoxic ammonium-oxidizing bacteria (AerAOB and AnAOB) and nitrite-oxidizing bacteria (NOB).

Chapter 1

Cost and energy effectiveness of OLAND

Conventionally nitrogen is biologically removed by nitrification/denitrification (N/DN). This process converts first all ammonium to nitrate and thereafter denitrifies nitrate with organic carbon to dinitrogen gas. In case some COD is still present in the wastewater, this endogenous organic carbon source can be used for denitrification. The typical composition of wastewater COD is C5H9NO, and 1 mol can reduce 3.36 moles of nitrate (Mateju et al., 1992). Hence, a strictly anoxic biodegradable COD/N ratio of about 4 is needed for denitrification, or about 5 taking into account some aerobic COD conversion. For wastewaters with lower COD/N ratios, external addition of a carbon source such as methanol is needed to obtain sufficient nitrogen removal rates. A cost-saving alternative for the latter is the application of nitritation/denitritation, saving 40% of the operational costs. The latter is the result of the decrease of the methanol requirement, sludge production and aeration with 50, 40 and 24%, respectively (Table 1.2 and 1.3). Moreover, when the fully autotrophic OLAND process is applied, 84% of the operational costs are saved, with a 100, 89 and 57% decrease in methanol requirement, sludge production and aeration, respectively (Table 1.3). Note that savings on methanol might in practice still be somewhat higher because of some aerobic consumption in the presence of residual dissolved oxygen (DO). The main cause of the low sludge production is the low biomass yield of the AnAOB. This group of bacteria has a long doubling time of 1-2 weeks (Strous et al., 1998) compared to the AerAOB (1 day) and denitrifying bacteria (in the order of 1h). In view of energy recuperation by anaerobic digestion, OLAND can offer a higher net energy gain because it minimizes the energy cost for further digestate treatment. It should also be mentioned that the choice of reactor technology will also further determine the operational and investment costs. Reactors with passive aeration, such as rotating biological contactors for instance, have a 3 times lower energy requirement for aeration compared to reactors with bubble aeration.

Introduction

Table 1.2: Overall stoichiometry for nitrification/denitrification, nitritation/denitritation and OLAND which are based on conversions of AerAOB, NOB, AnAOB en denitrifiers (Barnes and Bliss, 1983; Mateju et al., 1992; Strous et al., 1998). Process Equation Subreaction Stoichiometry nr Nitritation (AerAOB) 1 Substrates NH4+ + 1.382 O2 + 0.091 HCO3Products 0.982 NO2- + 1.891 H+ + 0.091 CH1.4O0.5N0.2 + 1.036 H2O Nitratation (NOB) 2 Substrates NO2- + 0.488 O2 + 0.003 NH4+ + 0.013 HCO3Products NO3- + 0.013 CH1.4O0.5N0.2 + 0.008 H2O Anammox (AnAOB) 3 Substrates NH4+ + 1.32 NO2- + 0.066 HCO3- + 0.13 H+ Products 1.02 N2 + 0.26 NO3- + 0.066 CH2O0.5N0.15 + 2.03 H2O Denitrification (Denitrifiers) 4 Substrates NO3- + 1.080 CH3OH Products 0.476 N2 + OH- + 0.760 CO2 + 0.325 CH1.4O0.5N0.2 + 1.440 H2O Denitritation (Denitrifiers) 5 Substrates NO2- + 0.53 CH3OH Products 0.48 N2 + OH- + 0.33 CO2 + 0.20 CH1.4O0.5N0.2 + 0.56 H2O Nitrification/denitrification 1+2+4 Substrates NH4+ + 1.856 O2 + 1.058 CH3OH Products 0.457 N2 + 1.010 H+ + 0.641 CO2 + 0.421 CH1.4O0.5N0.2 + 2.349 H2O Nitritation/denitritation 1+5 Substrates NH4+ + 1.382 O2 + 0.52 CH3OH Products 0.47 N2 + 0.998 H+ + 0.235 CO2 + 0.057 CH1.4O0.5N0.2 + 1.497 H2O OLAND 1+3 Substrates NH4+ + 0.792 O2 + 0.080 HCO3Products 0.435 N2 + 1.029 H+ + 0.111 NO3- + 0.052 CH1.4O0.5N0.2 + 0.028 CH2O0.5N0.15 + 1.460 H2O Table 1.3: Approximation of operational costs of biological nitrogen removal. Calculation factors: 0.32 EUR kg-1 methanol (Mathanex, 2011), dosed at 120% of stoichiometric requirement to compensate for aerobic breakdown; 0.10 EUR kWhel-1 (Europeâ&#x20AC;&#x2122;s energy portal); 0.47 EUR kg-1 sludge dry weight (DW) (Paul et al., 2006); 2 kg O2 kWhel-1; personnel costs based on a medium-sized plant treating 450 kg N d-1, requiring 1/2 full-time equivalent staff (FTE) for operation, maintenance and repair (50 000 EUR FTE-1 yr-1). Process Aeration requirement Methanol addition Sludge cost Personnel Total cost Cost savings kWh EUR kg EUR kg EUR EUR EUR % kg-1 N kg-1 N kg-1 N kg-1 N kg-1 N kg-1 N kg-1 N kg-1 N Nitrification/denitrification 2.1 0.21 2.9 0.93 1 0.47 0.15 1.76 0 Nitritation/denitritation 1.6 0.16 1.4 0.46 0.6 0.28 0.15 1.05 40 OLAND 0.9 0.09 0 0 0.1 0.05 0.15 0.29 84

Chapter 1

3 3.1

OLAND design parameters Choice of reactor technology

The application criteria including the complexity of the wastewater, the available footprint area and the need for high level trained operators are dominating the choice of reactor technology. Also, the operational costs and particularly the energetic aspects can further influence the reactor type chosen. In the Table 1.5 a qualitative comparison between different possible reactor types for applying OLAND is given. Three main categories can be distinguished i.e. attached, immobilized and suspended growth systems. Due to the lower complexity and low energy usage (passive aeration), attached growth systems such as rotating biological contactors (RBC), are preferentially applied at smaller scale (Meulman et al., 2010) or for complex wastewaters such as landfill leachates (Siegrist et al., 1998). OLAND RBC are robust and can stably operate for years (LabMET experience). However, the flexibility of the loading rate is limited and the oxygen balance is hard to control in contrast to suspended growth systems or systems with carrier material in suspension where oxygen can be regulated by controlling the aeration rate. Most full-scale OLAND-type of reactors until now are gas-lift or sequencing batch reactors (SBR), offering efficient DO control mechanisms and high operational flexibility. Both reactor types (gas-lift and SBR) have high biomass retention based on well settling sludge allowing to separate the sludge at the top of the reactor with a three phase separator or during a settling phase, respectively. However, due to the complexity of the control mechanisms, qualified operators are needed to allow stable and highly efficient performances in the reactor types based on suspended biomass. The ease of inoculation of new reactors with cultivated sludge from suspended growth systems is an additional advantage is this type of reactors and can therefore accelerate the implementation rate of the anammox-based processes.

Introduction

Place Olburgen, NL1 China1 China1 China1 China1 China1 Poland1

Table 1.4: Full-scale one-stage partial nitritation/anammox applications treating digestates from industrial and municipal origin. Reactor Water N in COD/N Bv N-Rf Vol pH DO Temp Sludge type type (mg N L-1) in (g N L-1 d-1) (%) (m3) (mg O2 L-1) (째C) (g SS L-1) Airlift Potato 250-350 0.6-0.8 1.8 73 600 8.0 2-3 30-35 15 processing Airlift Glutamate 600 2.0 >80 5000 factory Airlift Glutamate <500 2.0 >80 4500 factory Airlift Glutamate <500 2.0 >80 5350 factory Airlift Yeast 300-800 2.0 500 factory Airlift Yeast 300-800 2.0 3500 factory Airlift Distillery 1000 2.0 600 -

Strass, Austria2 Heidelberg, D2 Glarnerland, CH2 Plettenberg D2 Apeldoorn, NL2 Thun, CH2

SBR

Niederglatt, CH3

SBR

SBR SBR SBR SBR SBR

Sludge filtrate Sludge filtrate Sludge filtrate Sludge filtrate Sludge filtrate Sludge filtrate Sludge reject water

1800

0.57

< 1.0

90-95

500

7.0

0-0.35

30-34

1300

0.7-1.0

0.60

90-95

800

7.0

0-0.35

25-35

1000

0.8

0.69

> 90

379

7.0

0-0.35

25-35

800

0.7-1.2

0.50

> 90

134

7.0

0-0.6

25-35

950

0.7-1.0

0.66

> 90

2914

7.0

0-0.35

25-35

1300

0.7-1.0

0.67

> 90

606

7.0

0-0.35

18-30

760

0.37

150

7.8

Sludge aggregate Granules Granules Granules Granules Granules Granules Granules Small granules Small granules Small granules Small granules Small granules Small granules Flocs

Chapter 1 Zurich, CH3

SBR

Sludge 650 0.45 2x 7.1 30 3.4-3.8 Flocs reject water 1400 Sint Gallen, SBR Sludge 890 0.36 2x 8.0 18-30 5.9-7.7 Flocs CH3 reject water 300 Hattingen, MBBR Sludge 503 0.55 63 171 7.8 3 30 13* Biofilm SE4 reject water Hattingen, MBBR Sludge 275 1.06 52 67 7.4 3.8 30 13.6* Biofilm SE4 reject water Himmerfjärd MBBR Sludge 776 0.29 74 699 8.0 27 6* Biofilm en, SE4 reject water / industrial (9/1) Himmerfjärd MBBR Sludge 497 0.24 59 699 7.1 31 5* Biofilm en, SE4 reject water / industrial (9/1) Sjölunda, MBBR Sludge 855 0.3 1.30 90 4 x 50 6.80.5-1.5 22-33 Biofilm 5 SE centrate 7.5 1 2 3 personal communication, Tim Hülsen; personal communication, Bernhard Wett; Joss et al. (2009); 4 Beier and Schneider (2008); 5 Christensson et al. (2011) * g TS L-1 Kaldness packing material

Introduction

Table 1.5: Qualitative comparison of OLAND reactor configurations (advantages indicated in bold). RBC: rotating biological contactor; SBR: sequencing batch reactor; CSTR: continuous stirred-tank reactor (Vlaeminck et al., 2012) Biomass growth Attached (biofilm) Immobilized Suspended (flocs and/or granules) RBC† Bed reactor Gas-lift CSTR Reactor configuration Trickling filter Upflow/SBR MBR SBR Fixed/moving Fixed/moving or with Overall costs Area requirement Aeration Ease of DO control Sludge content Ease of biomass retention Inoculation feasibility♯ Low HRT feasibility° Risk for mechanical failure Risk for clogging Operational flexibility Operational complexity †

Low Medium Passive Low Medium Medium Medium Yes Medium High Low Low

Low High Passive Medium¶ Medium Medium Low/Medium Yes High Low Low Low

Medium Low Active Medium/High Medium Medium Low/High Yes Low High/Low Low/Medium Medium

Medium High Medium upflow Medium Low/Medium Medium Medium Low Active Active Active Active High High High High Medium Low High High Medium Low Low High High High High High No No Yes Yes Medium Low Low Low High Low Low Low Medium/High Medium Medium High* Medium/High High Medium High

Medium settler‡ High Active High Low Low High No Low Low Medium Medium

Biofilm can grow on rotating discs (fixed), or on carrier material brought in rotating porous cages (moving); ‡ Similar configuration as conventionally used for activated sludge; ¶ Rotation speed can be controlled by bulk DO level (Meulman et al., 2010a); ♯ Assuming sufficient availability of enriched inoculum, attached to carrier material if applicable; ° Important for wastewaters with low nitrogen level. For SBR and CSTR, this largely increases required settling time or settler volume, whereas for MBR this largely increases the amount of membranes required; * Cycle duration can be adjusted to meet effluent requirements (Siegrist et al., 2008), allowing to respond to changes in wastewater composition

Chapter 1

3.2

Key control mechanism to obtain stable performance

3.2.1 Balancing oxygen budget Oxygen plays a key role in the OLAND process. One hand, enough oxygen should be present to allow aerobic ammonium oxidation to nitrite. However, if the oxygen input is too high, further oxidation to nitrate by NOB can take place, decreasing the overall removal efficiency. Since AnAOB can be reversibly inhibited by oxygen concentration levels of 0.02 â&#x20AC;&#x201C; 0.15 mg O2 L-1 (Strous et al., 1997), the two OLAND key players (AerAOB and AnAOB) have opposite oxygen needs which implies a three-dimensional stratification in the granule, floc or biofilm (Vlaeminck et al., 2010). The maximum dissolved oxygen (DO) level experienced by the biomass can be directly controlled in most reactor technologies, except for RBC and trickling filters. The DO can be kept at a certain setpoint or within a certain range, with either continuous or intermittent aeration. The effect of the aeration regime on the OLAND performance is not fully clear yet. Joss et al. (2009) showed that continuous aeration was preferred over intermittent aeration (75% of the time aerated), because of the lower nitrite accumulation for these conditions and the better monitoring due to the higher signal to noise ratio when the aerators were not continuously switched on and off. In contrast to the latter study at low DO set point (0.5 mg O2 L-1), Zubrowska-Sudol and co-authors (2011) suggested that an intermittent regime (66% of time aerated) was optimal at higher DO levels (2, 3, 4 mg O2 L-1) obtaining higher nitrogen removal rates but also higher nitrite accumulations. The optimal DO set point is dependent on the preferred quality of the effluent, mixing conditions in the reactor and type of biomass (oxygen gradient). For smaller granular (< 1mm, Wett, 2006) or floccular biomass (Joss et al., 2009), DO levels below 0.5 mg O2 L-1 are advisable to avoid nitrite accumulation and development of NOB. When larger granules (2-3 mm), allowing higher AnAOB concentrations, are used, higher DO set points can be applied up to 2 mg O2 L-1 (Abma et al., 2010). However, at these higher DO conditions, NOB can more easily compete with the AerAOB for oxygen and can therefore form a barrier between AerAOB and AnAOB in the granule (Vlaeminck et al., 2010). 3.2.2 pH control mechanism to obtain balanced performance In the DEMON process, based on the same microbial conversions as the OLAND process, the balance between AerAOB and AnAOB is obtained by a dedicated control mechanism based on pH measurements. As the aerobic ammonium oxidation by AerAOB produces 1.9 mol H+ per mol NH4+ converted, this first reaction causes a decrease in pH which can be correlated 11

Introduction

with nitrite production. The aeration control system in this process is therefore based on a very tight pH control interval of 0.01 units (Wett, 2006). When a pH decrease of 0.01 units is measured, aeration is stopped and this allows depletion of the formed nitrite by AnAOB and some recovery of alkalinity (Table 1.2). Additionally, alkaline influent water is continuously fed to the system increasing the pH value until the upper value is reached and aeration is switched on again. This control strategy leads to an intermittent aeration regime with DO concentrations between 0 and 0.3 mg O2 L-1 while constant feeding is applied (Wett, 2006). While the OLAND-type of processes are mainly applied at pH ranges between 7 and 8, this control strategy is applied at pH value between 7.0 and 7.1 (Table 1.4).

3.2.3 Retaining sufficient microbial biomass Since the doubling time of the AnAOB is 1-2 weeks (Strous et al., 1998), high microbial biomass retention is a crucial factor to maintain sufficient AnAOB activity in the process. The microbial biomass retention is most delicate in suspended growth systems where it mainly depends on the formation of well settling sludge. In a continuously stirred tank reactor (CSTR) or a SBR, the microbial biomass retention by settling occurs in a separate step divided in space or time, respectively. In a SBR, biomass loss occasionally occurred due to small N2 bubbles attached to the flocs (Joss et al., 2009) or due to foaming problems (Wett, 2006). Adjustments of the feeding strategy (Wett, 2006), the settling phase or addition of flocculants (Joss et al., 2009) could solve this problem. Formation of both well settling flocs and granules is possible in SBR systems (Wett, 2006; Joss et al., 2009). Formation of granules is of utmost importance in gas-lift reactors because they depend on the continued presence of well settling granules (Abma et al., 2010). In attached growth system, biofilm formation allows for high biomass retention. In general, a total sludge retention time (SRT) of at least 30-45 days is recommended (Wett et al., 2010b; Desloover et al., 2011a; Joss et al., 2011).

3.3

OLAND performance

According to the reported OLAND-type of applications, the size of the reactor can be dimensioned based on a volumetric loading rate of 0.4 to 2 g N L-1 d-1 (Table 1.4). If the nitrogen removal rate is monitored directly by an ion-selective ammonium probe or indirectly via conductivity measurements, the SBR cycle can be adjusted according to the obtained removal obtaining optimal effluent quality and stable nitrogen removal rates (Joss et al., 2009).

Chapter 1

Figure 1.2: Microbial resource management view on the OLAND process. AerAOB and AnAOB: aerobic and anoxic ammonium-oxidizing bacteria; NOB: nitrite-oxidizing bacteria; GHG: greenhouse gas; bCOD: biodegradable chemical oxygen demand; GHG: greenhouse gas; DO: dissolved oxygen; VSS: volitale suspended solids 13

Introduction

Microbial resource management (MRM)

The close interaction between the different microbial groups during the OLAND process is comparable with human beings working together in firms for a shared profit. In this sense, the concept of human resource management (HRM) can be translated to the microbial biotechnology as Microbial Resource Management (MRM) and will therefore strive after maintaining the best performing microbial community for a certain application (Verstraete et al., 2007). To properly manage complex microbial systems, the engineer needs welldocumented concepts, reliable tools and a set of default values (Verstraete, 2007).

A MRM OLAND framework was elaborated, showing how the OLAND engineer/operator (1: input) can design/steer the microbial community (2: biocatalyst) to obtain optimal functionality (3: output), depending on the application domain (0: wastewater) (Fig. 1.2). Taken this MRM framework in to account, the OLAND engineer can steer the OLAND process to obtain maximum efficiency (see section 4.1) and higher sustainability (see section 4.2) or to increase the impact of OLAND on the energy balance of wastewater treatment plants (WWTP) (see section 4.3).

4.1

Maximizing nitrogen removal efficiency

The maximum nitrogen removal efficiency that can be obtained in a balanced OLAND system without additional denitrification is 89% (Eq. 1, Table 1.2). Lower removal efficiencies are mainly caused by hampered nitritation resulting in residual ammonium, by an imbalance between nitritation and anammox resulting in nitrite accumulation, or by increased nitratation resulting in a higher nitrate production. For most OLAND applications treating high-strength nitrogenous wastewaters, a post-treatment is obligatory to meet discharge limits. For sewage sludge reject water treatments (Fux and Siegrist, 2004) or source-separated black/grey-water systems (Verstraete and Vlaeminck, 2011), the OLAND effluent is sent to the diluted treatment stream for polishing. For industrial applications, the effluent can be sent to a sewage treatment plant (Abma et al., 2010), or can be polished by an additional separate nitrification and denitrification stage (Desloover et al., 2011a; Tokutomi et al., 2011b). The latter techniques are also used to polish OLAND-treated landfill leachate, and can be complemented with an activated-carbon stage (Hippen et al., 2001; Denecke et al., 2007). A possibility which has not been explored so far, is the inclusion of an anoxic reaction phase in 14

Chapter 1

the OLAND reactor to denitrify the nitrate produced with either autochtonous or added COD to further increase the removal efficiency. Given the low COD/N required to remove the remaining 11% of the nitrogen load, it is anticipated that denitrifying bacteria would not outgrow the AnAOB.

AerAOB activity should be high enough to deliver nitrite to the AnAOB, otherwise residual ammonium prevails (Fig. 1.3). An increase in AerAOB activity can be obtained by adjustment of the oxygen supply and level, yet care should be taken not to use DO levels above 0.5 mg O2 L-1, since this will favour the development of NOB (Wett, 2006; Joss et al., 2009). It should be noted that in systems with larger aggregates (granules), higher DO setpoints can be applied (Volcke et al., 2010). Under more extreme conditions, high free ammonia (8 - 120 mg N L-1) could decrease AerAOB activity at high ammonium concentrations, high pH and elevated temperatures, or high nitrous acid concentrations (0.2 - 2.8 mg N L-1) could be inhibitory at high nitrite concentrations, low pH and low temperatures (Anthonisen et al., 1976; Fig. 1.3). However, these conditions are not likely for OLAND reactors.

Figure 1.3: OLAND MRM framework highlighting tools to obtain high nitrogen removal efficiency. FA: free ammonia; FNA: free nitrous acid; SRT: sludge retention time

Introduction

If the AnAOB are not able to consume the formed nitrite or AerAOB leave not enough ammonium to combine with nitrite, nitrite accumulation will occur, which in a more extreme case (> 100-250 mg NO2−-N/L; Strous et al., 1999; Egli et al., 2001; Dapena-Mora et al., 2007) can inhibit AnAOB. Besides lowering the AerAOB activity by operational parameters such as a lower oxygen supply and level, one of the main factors to discounter the difference in growth rate between AerAOB and AnAOB is the separation of the sludge retention of small flocs, containing mainly AerAOB, and larger denser biomass particles, containing mainly AnAOB (Vlaeminck et al., 2010). Different selection methods are available to decrease the aerobic activity, depending on the applied reactor technology. Typical critical settling velocities applied in SBR systems are 0.3 – 3 m h-1 (Chapter 2; Wett, 2006; Joss et al., 2009). Selection for larger, denser biomass particles can therefore be based on the selective removal of smaller particles, which have a lower density and hence lower settling velocity. In granular upflow systems, removal of smaller, nitrifying granules at the top of the sludge bed led to higher biomass specific conversion rates (Winkler et al., 2011). In floccular systems, the use of hydrocyclones has been initiated to selectively maintain AnAOB-containing granules (Wett et al., 2010b). As the AnAOB are the slowest growers in the OLAND system, they should be maximally maintained in the system and stimulated as much as possible. It has been shown in several studies that the AnAOB are sensitive for oxygen (Strous et al., 1997; Egli et al., 2001). The presence of anoxic zones can also be promoted by the use of suspended carrier material in a MBR (Beier and Schneider, 2008) or by biomass immobilization in a gel matrix. Moreover, depending on the reactor technology applied, anoxic reactor zones can be created in space or time. It should be noted that methanol, commonly used as exogenous carbon source for denitrification, is detrimental for anammox (Güven et al., 2005; DapenaMora et al., 2007). Besides prevention of anammox inhibition, anammox can also be stimulated with components such as hydrazine, and dodecanoyl homoserine lactone (De Clippeleir et al., 2011). Other operational conditions that selectively favor AnAOB activity are not clear yet.

Nitrate accumulation due to NOB should be avoided at all time. For high-strength wastewaters followed by a post-treatment, NOB can be suppressed in the OLAND system at high free ammonia concentrations (> 5 mg N L−1) and low oxygen concentrations (Vlaeminck et al., 2009b). In the latter case, the AerAOB will have a competitive advantage over the NOB for substrate and space. In the case of diluted wastewater systems which have to reach effluent quality standards, free ammonia levels will not be sufficient anymore to suppress 16

Chapter 1

NOB and other methods should be searched especially for application at low temperatures (Section 4.3). One option is the addition of compounds such as sulphide at concentrations of 20-80 mg S L-1 (Erguder et al., 2008) or chlorate at concentrations of 83-830 mg L-1 (Belser and Mays, 1980), which have been shown to inhibit NOB activity. However, as long-term addition of these compounds could result in adaptation and could also affect AerAOB or AnAOB, this should be avoided as much as possible. Although Nitrospira lacks the common protection mechanisms for reactive oxygen species (Lücker et al., 2010), the addition of peroxide (up to 1.0 g H2O2 L−1) had no influence on the nitratation rate of a nitrifying culture with Nitrospira. In contrast, already at 0.5 g H2O2 L−1, the nitritation rate was significantly inhibited, rendering peroxide addition as an useful strategy to suppress nitritation (Vanslambrouck, unpublished). A close interaction between AerAOB and AnAOB could also play a role in avoiding nitratation, as the affinity of the AnAOB for nitrite is higher than the affinity of NOB for nitrite (Lackner et al., 2008). It should be however noted that until now, only limited knowledge exists about the genus/species dependency of these inhibition factors and it is therefore not always straightforward to avoid nitratation.

In general, it is suggested that to obtain a balanced OLAND system with maximum nitrogen removal efficiency, sufficient DO limitation, and a separation between the SRT of small aerobic flocs and larger anoxic particles are desired (Fig. 1.3).

4.2

Minimizing harmful gas emissions

In terms of gaseous emissions, sustainability mainly includes minimal emissions of nitric oxide (NO), an ozone degrader, and nitrous oxide (N2O) and methane (CH4), two potent greenhouse gases (GHG).

Methane can be expected in the OLAND influent when treating anaerobic digestates (dissolved at 11 g m-3 at 35°C), and small quantities might be formed in a non-aerated phase if all oxygen and nitrate are consumed (Desloover et al., 2011a). Aeration causes stripping of this methane. Although this can have a non-negligible contribution to the overall carbon footprint of the process (Desloover et al., 2011a), it is difficult to prevent the emission of dissolved influent methane, unless bubbleless aeration would be used for OLAND, as for instance in a membrane aerated biofilm reactor (MABR; Pellicer-Nacher et al., 2010).

Introduction

In contrast to methane, the formation of N2O and NO occurs in situ (Fig. 1.4). As mentioned above, for three monitored full-scale OLAND-type of systems, 0.4-1.3% of the nitrogen load was emitted as N2O (Joss et al., 2009; Kampschreur et al., 2009a; Weissenbacher et al., 2010). These values can be considered acceptable, since they do not significantly exceed the N2O emission values from nitrification/denitrification (Kampschreur et al., 2009a). NO emissions are normally ranging from negligible to 0.01% of N load (Joss et al., 2009; Kampschreur et al., 2009a; Weissenbacher et al., 2010), but NO is due to its low water solubility easily emitted when formed. The formation of N2O and NO is complex and often difficult to predict due to the interplay of many parameters and contributors (Fig. 1.4).

Figure 1.4: OLAND MRM framework elaborated for the risk of N2O and NO emissions in OLAND systems. q: specific microbial activity 18

Chapter 1

AerAOB are probably the predominant responsibles for N2O/NO emissions in OLAND, through so-called â&#x20AC;&#x2DC;nitrifier denitrificationâ&#x20AC;&#x2122;. The dominant energy generation method by AerAOB is via the aerobic metabolic pathways (Chain et al., 2003). However, under oxygen limitation or anoxic conditions AerAOB, including Nitrosomonas europaea and N. eutropha, can use NO2- or N2O4 as electron acceptors and NH3 or H2 as electron donors to produce NO and N2O, but no N2 (Ritchie and Nicholas, 1972; Poth and Focht, 1985; Schmidt et al., 2004). The oxygen level and regime (i) have profound effects on N2O/NO emissions. At oxygen concentrations below 1 mg O2 L-1, N2O productions up to 10% of the nitrogen load were observed (Goreau et al., 1980). While NO can be produced under both aerobic and complete anoxic conditions (Ritchie and Nicholas, 1972; Yu et al., 2010), N2O formation by AerAOB was only detected at aerobic or microaerophilic conditions. The N2O production by AerAOB mainly occurs at the transition from anoxic to aerobic conditions and is coupled to the presence of accumulated ammonium (Yu et al., 2010). Besides oxygen, nitrite concentrations (ii) play an important role in AerAOB NO and N2O emission (Kampschreur et al., 2009b). Nitrite accumulation is a common malfunctioning in OLAND reactors (Section 4.1), and significantly increases AerAOB N2O emissions (Colliver and Stephenson, 2000). High N2O production is additionally linked to high specific activity or alternately high metabolic rates (iii) during periods with high nitrogen flux through the catabolic pathways (Yu et al., 2010). Imbalanced enzyme expression in AerAOB performing close to their maximum specific activity (Yu et al., 2010), would suggest that, according to the Monod kinetics, working with a AerAOB community with lower substrate affinities (higher Ks) would yield a bigger risk of N2O emission at lower substrate accumulations. Therefore, process configurations that work under constant specific activity values, which are related to uniform DO and ammonium concentrations in the reactor, are expected to produce less N2O. In this content, discontinuous technologies such as SBR systems have more potential for N2O formation due to more frequent transitions. Slow and long feeding during the reaction phase would result in more stable nitrogen concentrations in the liquid phase (Wett, 2006) and could therefore potentially lower the risk of N2O formation. Athough ammonium oxidizing archaea (AOA) have recently been shown to produce N2O (Santoro et al., in press), so far no AOA have been detected in OLAND systems, rendering their contribution to N2O emissions likely nihil.

Introduction

Chemical formation of NO/N2O is another, potentially important pathway. An important factor is the accumulation of the AerAOB intermediate hydroxylamine. If this compound accumulates, it can either biochemically by AerAOB (Yu et al., 2010) or purely chemically (van Cleemput, 1998) react with nitrite and form NO and N2O. Moreover, chemical nitrite reduction at neutral pH can occur with ferrous iron (van Cleemput, 1998), sulfide (Grossi, 2009) or organic compounds (van Cleemput, 1998) and will also result in the formation of NO and N2O. It should be noted that N2O/NO emissions can also be lowered by a decrease of stripping. It was described that NO and N2O emissions increased with the air flow rate because the concentration of both gases remained constant in the gas phase. Therefore NO and N2O emissions can be minimized by minimizing the airflow rate under optimal conditions (Kampschreur et al., 2008) or by using bubbleless aeration in a MABR (Pellicer-Nacher et al., 2010).

Although denitrification is limited in OLAND systems, typical OLAND conditions promote NO/N2O emissions by denitrifiers. A high nitrite concentration during denitrification suppresses the denitrification rate and therefore leads to NO and N2O accumulation (von Schulthess et al., 1995). Also COD limitation during denitrification is a known cause for NO or N2O accumulation (von Schulthess and Gujer, 1996; Chung and Chung, 2000). Moreover, as oxygen inhibits both the synthesis and activity of denitrifying enzymes and N2O reductase is the most oxygen-sensitive denitrifying enzyme (Otte et al., 1996), the low DO values typical for OLAND can lead to N2O emission by denitrifiers. Although NO is one of the AnAOB intermediates (Kartal et al., 2011), it is unlikely that AnAOB leak NO, and therefore AnAOB probably do not contribute to NO emissions. Due to the absence of N2O reductase in the AnAOB genome, N2O production is not expected during anammox.

Overall, stable conditions allowing for constant specific microbial activities and avoiding accumulation of nitrite and ammonium likely lead to lower NO and N 2O emissions from OLAND systems (Fig. 1.4). However, the oxygen-limited conditions needed to avoid NOB activity or caused by well settling sludge remain a risk factor. Note that preliminary measurements of intermittent versus continuous aeration could not point out lower N2O 20

Chapter 1

emissions for the latter (Joss et al., 2009). It is expected that future long-term, on-line measurements will reveal the best aeration level and regime to minimize NO/N2O emissions.

4.3

OLAND enabling energy positive sewage treatment

Until now, the OLAND process has been successfully applied for medium and high-strength nitrogen wastewaters (> 0.2 g N L−1) such as landfill leachate and digestates from sewage sludge, specific industrial streams and concentrated black water. For centralized domestic wastewater treatment, the inclusion of OLAND to treat sludge digestates in the side stream of a conventional wastewater treatment plant (WWTP) lowered the overall plant energy requirements with about 50% (Siegrist et al., 2008). Furthermore, Wett et al. (2007) demonstrated energy autarky by including OLAND in the sidestream of a two-stage activatedsludge (AS) process (‘AB Verfahren’). In the mainstream, the first AS unit (A stage) has a very high loading rate (SRT ≈ 0.5 d), and the second AS unit (B stage) has a low loading rate (SRT ≈ 10 d). Besides these energy saving options with OLAND in a side stream, a novel treatment scheme was recently proposed, bringing OLAND to the main treatment stream substituting the previous B stage (Wett et al., 2010b; Verstraete and Vlaeminck, 2011). This even allows the electrical energy recovery and savings to exceed the electrical energy input. Moreover, instead of a biological concentration of the sewage, an enhanced physico-chemical concentration step can be applied, involving enhanced sedimentation, dissolved air flotation and/or membrane filtration, separating more than 75% of the COD load from the main stream (Verstraete et al., 2009).

A first difference between treatment of the main or side stream is the lower nitrogen concentration to be treated by OLAND (Fig. 1.5). Domestic wastewater after advanced concentration will still contain most of the nitrogen while around 75% of the COD is removed and sent to the digester, resulting in main stream wastewater with around 30-100 mg N L-1 and 113-300 mg COD L-1 (Metcalf and Eddy, 2003; Tchobanoglous et al., 2003; Henze et al., 2008). Taking into account the affinity constant of the AerAOB and AnAOB for ammonium i.e. 2.4 and 0.07 mg N L-1 respectively and the AnAOB affinity constant for nitrite of 0.05 mg NO2--N L-1 (Lackner et al., 2008), these low concentrations as such should not be a problem. However, these low substrate conditions could imply that the microbial community will have to work at lower metabolic and lower growth rates compared to side stream processes, which allow higher concentrations in the reactor.

Introduction

To obtain high nitrogen removal rates at low concentrations, low hydraulic residence times are needed for main stream treatment, in the order of hours and hence about 24 times lower than for side stream treatment (Joss et al., 2009; Weissenbacher et al., 2010). Given the slow biomass growth of the AnAOB, good biomass retention is a prerequisite for OLAND activity under low HRT. Sufficient AnAOB retention can be obtained by separating the retention of small aerobic and larger anoxic particles, which selectively will favour the AnAOB retention (see section 4.1). On the other hand, by increasing the external settler volume, applying a granular technology (Abma et al., 2010) or using biofilm-based technology, the total SRT can be increased.

Besides the survival of the AnAOB under low hydraulic retention times, an important challenge is to obtain a good microbial balance and activity at low temperature. Some studies already described the effect of lower temperatures on the separate activity of AnAOB, AerAOB and NOB. However, limited information exists about the microbial balance of these three groups under OLAND conditions at low temperature. AerAOB activity decreased with 50% at a temperature interval from 27 to 15°C, yet only limited aerobic ammonium oxidation could be observed at 5°C (Guo et al., 2010). For AnAOB the critical temperature at which it was difficult to obtain AnAOB activity was 18°C (Dosta et al., 2008), although several AnAOB species are found in nature at -1 to 15°C (Dalsgaard et al., 2005). It is not clear whether other AnAOB species, more related to the cold-temperature marine genus “Candidatus Scalindua”, will take over from the WWTP types “Candidatus Kuenenia and Brocadia” at colder temperatures. For inoculation purposes it is important to elucidate if the same AerAOB and AnAOB species do the job at cold temperatures or other species take over. In the latter case, the first start-ups will be slower again due to the absence of appropriate inoculation sources. The possible loss of both AerAOB and AnAOB activities compared to higher temperatures will result in the accumulation of nitrite and a decrease in oxygen uptake (Wett et al., 2010a). It will therefore be important to adjust the oxygen regime to impose oxygen-limited conditions to the AerAOB and by this avoid inhibition of AnAOB by nitrite. However, due to the decreased total activity, longer HRT or higher biomass concentrations will be necessary to obtain the same volumetric nitrogen removal rates. Beside the microbial balance between AerAOB and AnAOB, the lower temperature will have an effect on the NOB-AnAOB balance. At temperatures lower than 15°C, the growth rate of NOB will become higher than the growth rate of AerAOB (Hellinga et al., 1998) and it will therefore not be possible to wash out NOB based on overall or even selective sludge retention. The 22

Chapter 1

main challenge in this application will therefore be the suppression of NOB at low temperature and low nitrogen concentration (low free ammonia and low nitrous acid).

Figure 1.5: OLAND MRM framework elaborated to elucidate challenges for application of OLAND in the main stream of a sewage treatment plant.

The last point of attention concerning new inputs in this application domain is the presence of organics, i.e. moderate levels of bCOD (90-240 mg L-1) in the wastewater. Depending on the strength of the raw sewage, COD/N ratios between 2.4 and 3 are expected after the concentration step, which is on the edge of the described limit for successful OLAND 23

Introduction

(Lackner et al., 2008). On the one hand, the presence of organics will facilitate DO control at low DO levels due to heterotrophic aerobic activity. On the other hand, competition for nitrite between heterotrophic denitrification and anammox will take place. These processes have already been demonstrated to successfully co-exist at a COD/N ratio of 2.2 (Desloover et al., 2011a). It is anticipated that higher nitrogen sewage levels together with the higher sewage temperature which will facilitate OLAND treatment in the main stream, will exist in the main stream due to further dilution preventions (Henze, 1997; Brombach et al., 2005).

Finally, according to this MRM approach (Fig. 1.5), to be able to apply OLAND in the main stream of the WWTP, the challenges of biomass retention at low HRT and NOB suppression at low temperature should be resolved first.

Objectives and outlines of this research

Altough the first OLAND applications have shown that this technology works in a stable and efficient way (Table 1.4), the implementation rate of this technology remains dependent on a few companies. Many potential users hold back because it seems that due to the long start-up periods for the first reactors and the reported sensitivities, a lot of experience is needed to keep this process running. To overcome this problem, the output box of the MRM framework was further studied in detail for high-strength nitrogen containing wastewaters (known application) in Part II of this work. In Chapter 2 the effect of the hydraulic conditions on the start-up of the OLAND SBR was studied. Furthermore, strategies to obtain a well-balanced OLAND system were proposed. As not only the effluent quality, but also the sustainability can be a competitive factor to choose an environmental technology, the N2O and NO emissions were studied in a full-scale OLAND reactor in Chapter 3. The relation between the N2O/NO emission and the accumulation of substrates/intermediates and changes in the operational conditions was elaborated.

In Part III of this work new opportunities were explored for the OLAND process (box 0 of MRM framework). In Chapter 4, the impact of OLAND on the total energy balance was calculated for industrial, communal and agricultural applications to elucidate where opportunities for OLAND can be found, regarding energy efficiency. From Chapter 4 it became clear that wastewater treatment of manure-based wastestreams is very complex and therefore OLAND implementation will depend on specific cases. However, this sector also 24

Chapter 1

has nitrogen-rich gaseous emissions, i.e. ammonia streams, which are mostly treated inefficiently. Therefore in Chapter 5, the possibility of OLAND to treat gaseous ammonia streams, instead of water streams, was tested in a biofilter. Another opportunity for OLAND, based on the energy calcutations of Chapter 4, was the implementation of OLAND in the mainstream of the municipal WWTP. This application domain was step by step elaborated. In Chapter 6, the OLAND performance at low nitrogen concentrations and low HRT was tested in a RBC at 34째C as a first preriquisite for mainstream OLAND. In Chapter 7, the same RBC was used and was adapted to lower tempatures (up to 15째C) and the presence of organics (COD/N ratio of 2) to simulate at lab-scale the OLAND performance at mainstream conditions. Based on a full-scale trial to implement OLAND in the mainstream at a WWTP in Strass (Austria), a life cycle analysis (LCA) was performed. This analysis was used to evaluate the effect of OLAND implementation in side and mainstream on a process, plant and life cycle level (Chapter 8).

In Chapter 9 (Part IV), the results obtained are discussed in the framework of the research objectives. Conclusions are drawn and perspectives of further research are presented.

Introduction

Floating hood for greenhouse gas emission measurements (WWTP Strass, Austria) 28

Chapter 2

Chapter 2: A low volumetric exchange ratio allows high autotrophic nitrogen removal in a sequencing batch reactor Abstract Sequencing batch reactors (SBRs) have several advantages, such as a lower footprint and a higher flexibility, compared to biofilm-based reactors, such as rotating biological contactors. However, the critical parameters for a fast start-up of the nitrogen removal by oxygen-limited autotrophic nitrification/denitrification (OLAND) in a SBR are not available. In this study, a low critical minimum settling velocity (0.7 m h−1) and a low volumetric exchange ratio (25%) were found to be essential to ensure a fast start-up. To prevent nitrite accumulation, two effective actions were found to restore the microbial activity balance between aerobic and anoxic ammonium-oxidizing bacteria (AerAOB and AnAOB). A daily biomass washout at a critical minimum settling velocity of 5 m h−1 removed small aggregates rich in AerAOB activity, and the inclusion of an anoxic phase enhanced the AnAOB to convert the excess nitrite. This study showed that stable physicochemical conditions were needed to obtain a competitive nitrogen removal rate of 1.1 g N L−1 d−1.

Chapter redrafted after: De Clippeleir, H., Vlaeminck, S.E., Carballa, M., Verstraete, W., 2009. A low volumetric exchange ratio allows high autotrophic nitrogen removal in a sequencing batch reactor. Bioresource Technology, 100, 5010-5015. 29

A low volumetric exchange ratio allows high autotrophic N removal in a SBR

Introduction

Despite the economical advantage of the AnAOB-based processes, such as OLAND, in comparison with the conventional nitrification/denitrification, these processes are hindered by the long start-up period due to the slow growth rate of the AnAOB, which have a doubling time of 7 to 14 days (Strous et al., 1998). Therefore, biomass washout has to be minimized, e.g. by biofilm formation or granulation. Several biofilm-based reactors, such as a rotating biological contactor (RBC; Siegrist et al., 1998; Pynaert et al., 2004), a moving bed reactor (Cema et al., 2006) or a fixed bed reactor (Furukawa et al., 2006) have already been successfully applied. High biomass retention can also be obtained in a sequencing batch reactor (SBR) operated at a critical minimum biomass settling velocity. The latter is defined as the ratio between the settling time and the vertical distance of the water volume decanted per cycle, and it can also be expressed as the volumetric exchange ratio, i.e. the ratio of the decanted to the total water volume. Reported minimum biomass settling velocities for OLAND type SBRs are in the range of 0.3-0.7 m h−1 (Third et al., 2001; Sliekers et al., 2002; Wett, 2006; Vlaeminck et al., 2009a). Although SBRs have advantages, such as a lower footprint and a higher flexibility, compared to biofilm based reactors, such as RBC, so far the nitrogen removal rates obtained in these reactors are almost five times lower (Table 2.1).

Not only efficient biomass retention is required for a successful OLAND process, a good balance between the AerAOB and AnAOB is needed as well. A higher activity of the AerAOB in comparison to the AnAOB results in nitrite accumulation in the reactor, which can inhibit the AnAOB activity at nitrite concentrations of 98 to 350 mg NO2−-N L−1 (Strous et al., 1999; Dapena-Mora et al., 2007). While in RBCs the microbial balance is equilibrated spontaneously due to the limited penetration depth of oxygen in the biofilm, the control of this microbial balance in SBRs is not straightforward. Two kinds of biomass morphologies, flocs and granules, were mainly present in suspended growth systems (Innerebner et al., 2007; Vlaeminck et al., 2010). Granules can be described as compact and dense aggregates with a high macroscopic circularity that do not coagulate under reduced hydrodynamic shear and settle significantly faster than flocs (Lemaire et al., 2008). Flocs were found to be enriched in AerAOB, while AnAOB were dominant in the granules (Nielsen et al., 2005; Vlaeminck et al., 2009a; Vlaeminck et al., 2010). Therefore, the overall balance between the AerAOB and AnAOB is dependent on the biomass morphology distribution in the reactor. Morphology selection on the basis of the settling velocity could therefore improve the microbial balance. 30

Chapter 2 Table 2.1: Overview of the volumetric nitrogen removal rates in OLAND type rotating biological contactors (RBC) and sequencing batch reactors (SBR). Reactor type

Volume (m³)

Nitrogen removal rate (kg N m−3 d−1)

Reference

RBC RBC RBC RBC SBR SBR SBR SBR SBR Gas-lift

33 0.044 240 0.005 0.002 0.002 0.002 500 0.002 0.002

0.4 1.1 1.7 1.8 0.1 0.3 0.5 0.6 1.1 1.5

Siegrist et al. (1998) Pynaert et al. (2003) Schmid et al. (2003) Pynaert et al. (2004) Third et al. (2001) Sliekers et al. (2002) Vlaeminck et al. (2009a) Wett (2006) This study Sliekers et al. (2003)

In this study, the microbial balance between the AerAOB and AnAOB was evaluated in an OLAND SBR. The critical parameters for a fast start-up were determined and strategies to control the microbial balance and enhance the biomass retention in the reactor were evaluated.

2 2.1

Materials and methods OLAND SBR

The lab-scale OLAND SBR consisted of a cylindrical vessel with an internal diameter of 14 cm (working volume of 2.5 L). The reactor was inoculated with OLAND biomass harvested from the reactor described by Pynaert et al. (2003) at an initial biomass concentration of 2.3 g VSS L−1. The reactor was fed with synthetic wastewater containing an initial ammonium concentration of 100 mg N L−1, 10 mg KH2PO4-P L−1 and 2 mL L−1 of a trace elements solution (Kuai and Verstraete, 1998). To provide both buffering capacity and inorganic carbon, 1 mole of bicarbonate was added per mole of nitrogen. If necessary, the latter ratio was increased temporarily to ensure that the reactor pH did not drop below 7.4. In addition, the influent ammonium concentration was gradually increased whenever the effluent concentration was below ca. 25 mg N L−1. The reactor was mixed with a magnetic stirrer at 245 rpm and aerated at an airflow rate of 40 L h−1. The temperature and the dissolved oxygen (DO) concentration were controlled automatically at 33 ± 1°C (temperature controlled room) and 0.3 to 0.7 mg O2 L−1 (Oxymax W COS31 probe with Liquisis M COM 223 controller; Endress & Hauser, Reinach, Switzerland), respectively. Three different phases of operation were carried out: phase 1 with high volumetric exchange ratio (40%) and high critical minimum settling velocity (2 m h−1); phase 2 with high volumetric exchange ratio (40%) and 31

A low volumetric exchange ratio allows high autotrophic N removal in a SBR

low critical minimum settling velocity (0.7 m h−1); and, phase 3 with low volumetric exchange ratio (25%) and low critical minimum settling velocity (0.7 m h−1). During the first and the second phases, the exchangeable volume was fixed at 1 L, resulting in a volumetric exchange ratio of 40% (1/2.5). During the third phase, the exchangeable volume was reduced to 0.5 L (by lowering the working volume to 2 L), and consequently, the volumetric exchange ratio decreased to 25% (0.5/2). The nitrogen compounds (ammonium, nitrite, nitrate), DO concentrations and pH were monitored during the whole experiment.

2.2

SBR cycle

The SBR was operated with 1h cycles during the whole experimental period. During the first phase, 1 L of synthetic medium was fed to the reactor during a 5 minutes filling period. The reactor was mixed and the DO was controlled both during the feeding and the reaction phase. Subsequently, the biomass was allowed to settle for 2 minutes, so that the minimum biomass settling velocity was 2 m h−1. Finally, an effluent pump removed the supernatant. During the second and third phases, the settling time was increased to 6 and 3 minutes, respectively, resulting in a lower selection pressure (critical minimum settling velocity of 0.7 m h−1).

2.3

Aerobic and anoxic batch tests

The specific activities of AerAOB and AnAOB were determined in aerobic and anoxic batch tests, respectively, as described in detail by Vlaeminck et al. (2007). Prior to the activity tests, the biomass was washed with a phosphate buffer (100 mg P L−1; pH 8) on a sieve (pore size 50 µm) to remove residual dissolved reactor compounds. The aerobic tests were performed in open Erlenmeyer with ammonium as substrate. For the anoxic tests, biomass incubation occurred in a gas-tight anoxic serum flask with ammonium and nitrite as substrates. Both tests were performed on a shaker at 34 ± 1°C.

2.4

Chemical analyses

Nitrite and nitrate were determined on a Metrohm 761 Compact Ion Chromatograph (Zofingen, Switzerland) equipped with a conductivity detector. Ammonium (Nessler method) was measured according to standard methods (Greenberg et al., 1992). The pH was measured with a Consort C532 pH meter (Turnhout, Belgium).

Chapter 2

2.5

Physical aggregate characteristics

A mixed liquor sample obtained in a Petri dish was photographed with a high resolution (10 megapixels) digital camera for particle analyses. The Feret diameter (largest diameter in irregular particle), the circularity of the biomass aggregates and settling velocity were determined as described by Vlaeminck et al. (2009a).

3 3.1

Results OLAND SBR performance

During phase 1 the strategy of the OLAND SBR operation was based on a high critical minimum settling velocity of 2 m h−1 to induce granulation. These conditions resulted in an average total nitrogen removal rate of only 20 mg N L−1 d−1, which was attributed to complete conversion of ammonium to nitrite (Fig. 2.1B). Moreover, no anammox activity was observed during this phase. Therefore, it was concluded that a critical minimum settling velocity of 2 m h−1 was too high. At the beginning of phase 2 (day 46), 1.6 g OLAND biofilm-VSS L−1 was added and the critical minimum settling velocity was decreased to 0.7 m h−1. These actions resulted in a higher anammox activity since the nitrite production was lower compared to the ammonium consumption (Fig. 2.1B), and consequently, a higher total nitrogen removal rate (around 135 mg N L−1 d−1) was obtained (Fig. 2.1A). However, the nitrite production rate was still high (around 88 mg N L−1 d−1) and no improvement of the total nitrogen removal over time was obtained.

In the subsequent phase 3 (day 95), the critical minimum settling velocity was kept constant (0.7 m h-1), but the volumetric exchange ratio was decreased from 40 to 25%. Similar to phase 2, extra biomass (1.6 g OLAND biofilm-VSS L−1) was added. These changes resulted in a steep and continuous increase of the total nitrogen removal rate and in a stable nitrate production (Fig. 2.1A). The fraction of nitrate produced in comparison with the net ammonium consumed was around 11%, which was in accordance with the expected nitrate production in the OLAND process (Strous et al., 1999). At the end of the experiment, a competitive removal rate of 1.1 g N L−1 d−1 was obtained.

A low volumetric exchange ratio allows high autotrophic N removal in a SBR

Figure 2.1: Performance of the OLAND SBR, subdivided in three experimental phases. Phase 1: high critical minimum settling velocity (2 m h−1) and high volumetric exchange ratio (40%). Phase 2: low critical minimum settling velocity (0.7 m h−1) and high volumetric exchange ratio (40%). Phase 3: low critical minimum settling velocity (0.7 m h−1) and low volumetric exchange ratio (25%). A Nitrogen removal rate and relative nitrate production. B Ammonium consumption and nitrite production.

3.2

Biomass morphology

The inoculum of the OLAND SBR consisted of biofilm pieces originating from an OLAND RBC. Due to mixing, these aggregates disintegrated resulting in a variety of biomass particles (Table 2.2). To improve the biomass retention in the SBR, granulation was pursued. 34

Chapter 2

Therefore, a stronger selection pressure (2 m h−1) was applied during phase 1, but this action did not result in enhanced granulation. In addition, the particle size distribution analyses did not show significant change during this phase (data not shown). From phase 3 on, the particle size distribution shifted progressively to the situation of day 135, when granules could be detected among other smaller biomass aggregates (Table 2.2). The fraction of granules increased during the operation period to an average of 20% by the end of the experiment.

Table 2.2: Distribution of biomass fractions at the start-up (day 1), when granules were present (day 135) and when nitrite accumulation occurred (day 161). Biomass fraction per size class (mm) Time < 0.5 0.5 – 1 1 – 1.5 >1.5 Day 1 (start-up)

0.10

0.59

0.11

0.20

Day 135 (granule formation)

0.11

0.45

0.24

0.20

Day 161 (nitrite accumulation)

0.22

0.32

0.20

0.25

Two kinds of granules (red and brown) with different characteristics could be distinguished (Table 2.3). Although the red granules were more uniformly distributed while the brown granules had a high variety of sizes (data not shown), the average Feret diameter was not significantly different. The red granules had a high circularity and good settling properties. Moreover, the red granules were perfectly balanced in activity in contrast with the brown granules which had an excess AerAOB activity (Fig. 2.2). The equilibration in aerobic and anoxic activity can also be represented by the nitrite accumulation rate ratio (narr), defined as the ratio of the net aerobic nitrite production rate to the anoxic nitrite consumption rate (Vlaeminck et al., 2010). The narr of the red granules, brown granules and the inoculum of SBR was 1.2, 3.8 and 1.4, respectively.

Table 2.3: Characteristics of the red and brown granules obtained in the OLAND SBR. The significantly different parameters are indicated with a star (p<0.01). Red granules Brown granules Feret diameter (mm) Circularity (-) Settling velocity (m h−1)

2.39 ± 0.32 0.75 ± 0.11

2.49 ± 0.62 0.65 ± 0.14 *

55 ± 10

35 ± 8 *

A low volumetric exchange ratio allows high autotrophic N removal in a SBR

3.3

Control of the microbial balance in the reactor

Parallel to the increase of the total nitrogen removal rate, an increase in nitrite production rate up to a maximum of 377 mg NO2−-N L−1 d−1 on day 116 was observed. As the AerAOB grow a factor 10 faster than the AnAOB (Jetten et al., 2001), the nitrite production can never be kept low if no external actions are taken. Particle size analyses indicated that the fraction of biomass particles smaller than 0.5 mm increased when nitrite accumulation was detected (Table 2.2), suggesting that these small particles were related to the AerAOB. Since AnAOB can loose activity at nitrite concentrations in the range of 98 to 350 mg NO2−-N L−1 (Strous et al., 1999; Dapena-Mora et al., 2007), a daily selection (every 24h, not every cycle) corresponding to a critical minimum settling velocity of 5 m h−1 (settling time of 23 seconds) and an anoxic phase of 10 minutes at the end of the reaction period were included from day 122 on. This occasional higher selection could wash the excess AerAOB and the anoxic phase could enhance the AnAOB activity. As a consequence, nitrite production decreased and remained low during the rest of the experimental run (Fig. 2.1).

Figure 2.2: Specific aerobic and anoxic activity and nitrite accumulation rate ratio (narr) of the SBR inoculum, SBR biomass mixture, the biomass washed out at a critical settling velocity of 10 m h−1 (n=3), the brown and red granules (n=1). Calculation and interpretation of narr is presented in section ‘Results’.

To test the individual effect of the daily selection on the suppression of the nitrite accumulation, both actions (daily selection and anoxic phase) were left out on day 156, resulting in an increase of nitrite in the reactor. On day 161 a selection corresponding to a critical minimum settling velocity of 10 m h−1 was performed and the AerAOB and AnAOB activity of the washed biomass was determined (Fig. 2.2). It could be observed that the 36

Chapter 2

washout fraction had a higher AerAOB activity (narr 4.1). Together with the small particles, a small fraction of AnAOB was removed from the reactor as well, since AnAOB activity was detected in the washed biomass (Fig. 2.2). The single effect of the anoxic phase could be determined by monitoring the concentrations of the nitrogen compounds during a SBR cycle (Fig. 2.3). The consumption of ammonium occurred during the complete cycle and nitrate was simultaneously produced. However, nitrite accumulated linearly (RÂ˛: 0.99) during the reaction period and it was only sufficiently consumed during the anoxic phase (23% of reaction period).

Figure 2.3: Evolution of ammonium, nitrite and nitrate concentrations during 1 cycle in the OLAND SBR with 5 periods (day 143): feeding phase (1), reaction phase (2), anoxic phase (3), settling period (4) and withdrawal (5).

4 4.1

Discussion OLAND SBR performance

During the first two phases no improvement of the nitrogen removal could be obtained. However, during phase 3 a steep increase of the nitrogen removal was detected. This increase in removal rate was attributed to the low volumetric exchange ratio of 25%, because the addition of extra biomass (1.6 g OLAND biofilm-VSS Lâ&#x2C6;&#x2019;1) at the beginning of phase 2 had no effect. To link low volumetric exchange ratio with good start-up, two hypotheses are formulated. Firstly, the hydraulic retention time was increased from 2.5 to 4 h when lowering the volumetric exchange ratio. However, these results contradict some results found in 37

A low volumetric exchange ratio allows high autotrophic N removal in a SBR

literature, showing that nitrogen removal rates did not increase by applying higher hydraulic retention times (Third et al., 2001; Sliekers et al., 2002; Tsushima et al., 2007b). However, the different operational and dimensional conditions between the different studies make the evaluation of the effect of the hydraulic retention time on the start-up and performance of OLAND type reactors rather difficult. Secondly, a lower volumetric exchange ratio yielded more stable hydraulic and chemical conditions in the reactor. The variation in the nitrogen concentrations, metabolic products and shear rates in the reactor between the beginning and the end of the feeding period were 1.3-fold instead of 1.7-fold in phase 1 and 2. Thus, the chemical and physical stress was lower, resulting in a more stable and continuous-like process. These stable physicochemical conditions could explain why higher nitrogen removal rates up to 2 g N Lâ&#x2C6;&#x2019;1 dâ&#x2C6;&#x2019;1 can be obtained in continuous reactors, such as RBC and airlift reactors (Sliekers et al., 2003; Pynaert et al., 2004). Although better performances have been reported in these continuous OLAND reactors, the nitrogen removal rate obtained in this study was exceptionally high compared with other lab- and full-scale OLAND-type SBRs (Table 2.1).

4.2

Biomass morphology

To granulate active or nitrifying sludge, a minimum settling velocity of 4.5 m h -1 is required (Liu et al., 2005). In this study, granulation was obtained at a critical minimum settling velocity of only 0.7 m h1. Other researchers detected granulation of OLAND type of biomass at similar critical settling velocities (Innerebner et al., 2007; Vlaeminck et al., 2009a) and, similar to Innerebner et al. (2007), granulation was only detected after a good nitrogen removal had been obtained. This fact suggests that the granulation process of OLAND biomass is different from the aerobic granulation, where a strong selective settling pressure is a prerequisite (Liu et al., 2005).

Red and brown granules were detected in the OLAND SBR. These two types of granules had similar sizes, but different physical and microbial properties. Red granules had a high circularity and settling velocity, resulting in efficient biomass retention. Moreover, these red granules had a narr between 0.6 and 1.4, indicating that these aggregates could perform the OLAND process autonomously (Vlaeminck et al., 2010). The brown granules had an excess AerAOB activity resulting in a high narr value (3.8). Thus, a high circularity and fast settling of the biomass aggregates in combination with narr values around 1 is preferable in the OLAND process. 38

Chapter 2

4.3

Control of the microbial balance in the reactor

Parallel with the steep increase in total nitrogen removal, nitrite accumulation occurred. In this study, a combination of a daily selection and an anoxic phase could restore the balance. The effect of the daily selection was confimed in an activity test. This test showed that the aerobic ammonium oxidizers were dominant in the small fraction (narr of 4.1), but also a small fraction of anammox bacteria was present. Calculated back to the effect on the reactor performance, one selection decreased the VSS content, the aerobic and the anoxic activity with 2, 3 and 0.3%, respectively. Although these percentages are small, the nitrite accumulation could be suppressed sufficiently. However, this daily selection resulted in sharp decreases of the nitrogen removal rates (Fig. 2.1), thus indicating that this effect must be modulated on a long-term basis.

The second action to avoid nitrite accumulation and, consequently the inhibition of the anammox bacteria was the insertion of an anoxic phase. The anammox reaction was predominant during the anoxic phase because the nitrate produced per ammonium consumed (23%) was similar to the relative nitrate production of the anammox reaction, i.e. 26% (Strous et al., 1999). Therefore, the anoxic phase was effective to control the nitrite concentrations, but the long-term effect of this action is difficult to asses.

Conclusions

A low selection pressure, corresponding to a critical minimum settling velocity of 0.7 m h −1, combined with a low volumetric exchange ratio of 25% and an equilibrated microbial activity were essential to obtain a competitive removal rate of 1.1 g N L−1 d−1. Besides the better settling properties, the red OLAND granules were well balanced in activity, and thus more suitable for a stable operation compared to the brown granules and the small aggregates. However, without a dominance of red granules, actions should be taken to avoid nitrite accumulation.

Acknowledgements

This research was funded by a PhD grant for Haydée De Clippeleir from the Institute for the Promotion of Innovation through Science and Technology in Flanders (IWT-Vlaanderen, SB81068), by a PhD grant (Aspirant) for Siegfried E. Vlaeminck from the Fund of Scientific 39

A low volumetric exchange ratio allows high autotrophic N removal in a SBR

Research-Flanders (Fonds voor Wetenschappelijk Onderzoek (FWO) Vlaanderen), and by a postdoctoral contract for Dr. Marta Carballa from the Xunta de Galicia (Isidro Parga Pondal program, IPP-08-37). The authors gratefully thank Bart De Gusseme and Peter Aelterman for the inspiring scientific discussions.

Chapter 3

Chapter 3: Interplay of intermediates in the formation of NO and N2O during full-scale partial nitritation/anammox Abstract Next to energy- and cost-efficiency, sustainability is evolving as a benchmark for wastewater treatment. Taking into account the high global warming potential of nitrous oxide (N2O), minimization of its emission is gaining attention. As the formation of N2O and nitric oxide (NO) is complex and relies on the interplay of different intermediates, such as nitrite (NO2-) and hydroxylamine (NH2OH), a detailed monitoring of all nitrogen species in both the gas and liquid phase was performed in this study. The aim was to find a link between measurable N components, operational conditions and the NO/N2O emissions from a full-scale OLANDtype reactor. High loading rates, resulting in highly dynamic cycles with rapid on/off aeration regimes, resulted in higher NO and N2O emissions, indicating that transient conditions favour both N2O and NO emission. Therefore, the beginning of a cycle during which most changes in operational conditions occurred was studied in detail. At the beginning of the cycle a lag phase in N2O and NO (30 and 15 min., respectively) emission was measured. Sudden peaks in ammonium oxidation rate up to 335 kg d-1 were accompanied with transient accumulations of NH2OH (up to 0.001% of NH4+ consumption) and/or NO2- (up to 0.2% of NH4+ consumption) and resulted in N2O and NO emission peaks. Despite the complex interplay of many factors, this study showed that NH2OH accumulation and NO/N2O emission can be correlated positively. Therefore, a better understanding of the conditions leading to NH2OH accumulation could help to find strategies to minimize N2O and NO emission. Chapter redrafted after: De Clippeleir H., Weissenbacher N., Boeckx P., Chandran K., Boon N. and Wett B. 2012. Interplay of intermediates in the formation of NO and N2O during fullscale partial nitritation/anammox. Ecotechnologies for wastewater treatment, Santiago de Compostela, Spain. 41

Interplay of intermediates in formation of N2O/NO during OLAND

Introduction

Besides cost- and energy-efficiency, the sustainability of the process is gaining more and more attention. Since 1 kg N2O has the global warming potential of 298 kg CO2 on a 100-yr time horizon (Solomon et al., 2007), the N2O emissions can have a huge impact on the CO2 footprint of a wastewater treatment plant (WWTP). Moreover, NO2 and NO emission contribute to the formation of tropospheric ozone and can cause acidification (Solomon et al., 2007). The formation of N2O and also NO occurs in situ. Recently, some studies have specifically addressed N2O emission from full-scale OLAND-type of systems, showing that 0.4-1.3% of the nitrogen load was emitted as N2O (Joss et al., 2009; Kampschreur et al., 2009a; Weissenbacher et al., 2010). These values can be considered acceptable, since they do not significantly exceed the N2O emission values from nitrification/denitrification (Kampschreur et al., 2009a). NO emissions during OLAND are normally ranging from negligible to 0.01% of N load (Joss et al., 2009; Kampschreur et al., 2009a; Weissenbacher et al., 2010). However, NO is due to its low water solubility easily emitted when formed. The formation of N2O and NO is complex and often difficult to predict due to the interplay of many parameters, contributors and mechanisms within the contributors (simplified overview in Fig. 3.1).

It is believed that the decrease in NO and N2O emission can be accomplished by optimization of the operational parameters. However to do so, a better understanding of the role of the different NO and N2O producing pathways in in situ conditions, characterized by an interplay of AerAOB, AnAOB, NOB and denitrifier activities under changing operational conditions, is needed. In this study, a detailed follow-up of all nitrogen species in liquid and gas phase was performed with the aim to link the presence of intermediates with NO and N2O formation. This study was performed on a full-scale OLAND-type of reactor, more specifically a DEMON SBR in the side line of the WWTP of Strass (Wett, 2006).

Chapter 3

N2 O

O2 NO3−

NO2−

N2 Anammox (AnAOB) Nitratation (NOB) Denitrification (HDN) Nitritation (AerAOB) Chemical reaction

NH2OH NH4

Figure 3.1: Nitrogen conversion in relation to NO and N2O formation.

N2 O

Materials and methods O2

2.1

Reactor operation

NO2−

NO3−

A full-scale DEMON sequencing batch reactor (SBR, 500 m ) treating sludge digestor Anammox (AnAOB) Nitratation (NOB)WWTP in Strass, Austria (Wett, 2006) was monitored in this supernatant at the municipal Denitrification (HDN)

study. The SBR was operated in cycles of 6 hours of which 75% of the time the oxygenNitritation (AerAOB)

NH2OH

limited reaction phase took place.NH During this phase the reactor was continuously fed and the + 4

balance between AerAOB and AnAOB activity was obtained by a dedicated control O2

mechanism based on pH measurements. As the aerobic ammonium oxidation by AerAOB produces 1.9 mol H+ per mol NH4+ converted, this first reaction causes a decrease in pH, NO The aeration control system in this process is which can be correlated N2O with nitrite production.

therefore based on a very tight pH control interval of 0.01 units (Wett, 2006). When a pH O2

decrease of 0.01 units is measured, aeration is stopped and this allows depletion of the formed − nitrite by AnAOB alkaline influent water is NO2− Additionally,NO N2 and some recovery of alkalinity. 3

(AnAOB) continuouslyAnammox fed to the system increasing the pH value until the upper value is reached and Nitratation (NOB)

aeration is switched on again. This control strategy leaded to an intermittent aeration regime Denitrification (HDN)

with DO concentrations between 0 and 0.7NH mg O2 L-1 while constant feeding is applied (Wett, Nitritation (AerAOB) 2OH 2006).

NH4+ O2

2.2

Emission measurments

To allow continuous off-gas measurements and to control foam formation, a cylinder (diameter 0.3 m, height 2 m) was vertically placed into the reactor and a defined air stream 43

Interplay of intermediates in formation of N2O/NO during OLAND

(0.7 ± 0.6 m s-1) was blown into the cylinder. Therefore, the in situ emitted gas concentrations were diluted with at least a factor 2 as during aeration periods maximum gas velocities of 1.8 m s-1 were detected. Gaseous N2O concentrations were measured online at a time interval of 3 minutes with a photo-acoustic infrared multi-gas monitor (Brüel & Kjær, Model 1302, Nærem, Denmark). NO was measured online using a chemiluminescense analyzer (APNA 350, Horiba, Japan) and recorded at one minute intervals. For dissolved N2O measurements, a 1 mL filtered (0.45 μm) sample was brought into a 7 mL vacutainer (-900 hPa) and measured afterwards by pressure adjustment with He and immediate injection at 21°C in a gas chromatograph equipped with an electron capture detector (Shimadzu GC-14B, Japan). Ammonium concentration (Nessler method) in the water phase was determined according to standard methods (Greenberg et al., 1992). Nitrite and nitrate were determined on an ion chromatograph equipped with a conductivity detector (Metrohm, 761 compact, Zofingen, Switzerland). Hydroxylamine was determined spectrophotometrically (Frear and Burrell, 1955). The N2O and NO fluxes of the full-scale reactor were based on the measured off-gas concentration corrected for the background concentration in the defined air stream and converted to molar concentration with the ideal gas law at the measured temperature and atmospheric pressure. Multiplication of the measured gas velocity of the air stream and the cross section area of the outlet of the cylinder (28 cm2) yielded the off-gas flow rate.

Figure 3.2: Picture of the set-up for greenhouse gas emissions in the DEMON reactor

Chapter 3

Results and discussion

The NO, NO2 and N2O emission from the full-scale OLAND-type SBR was measured at two different loading rates i.e. 247 kg N d-1 and 107 kg N d-1 (Fig. 3.3). In both cases, the airflow rate, mixing rate and operational conditions such as DO, pH and influent quality were kept constant, as no changes were made in set-points and control mechanisms and only the feeding rate was changed. The effluent quality in both cycles changed from COD, NO3--N, NO2--N and NH4+ concentration of 632, 49, 1 and 60 mg L-1 to 632, 48, 2 and 15 mg L-1 for the high and low loading conditions, respectively. As the digestate contained a considerable amount of inorganic carbon, a lower loading rate caused lower CO2 emissions and CO2 emissions rapidly followed the aeration regime (Fig. 3.3). To obtain a similar pH and DO pattern at higher loading rate, a more transient operation was imposed characterized by rapid on/off aeration regimes. The full-scale reactor emitted at the high loading rate 3.5 kg N2O-N d-1, 33 g NO-N d-1 and 6.7 g NO2-N d-1, which corresponded to 1.4, 0.02 and 0.003% of the nitrogen load, respectively. These emissions were in the expected range according to literature (Joss et al., 2009; Kampschreur et al., 2009a; Weissenbacher et al., 2010). As the effluent COD and nitrate concentration were constant at lower loading rate compared to the higher loading rate and an increased nitrite and decreased ammonium effluent concentration was observed, similar or higher relative N2O and NO emissions were expected. However, at the lower loading rate, characterized by longer anoxic periods, N2O emissions decreased until 0.37 kg N2O-N d-1 or 0.3% of N load. NO was more easily stripped out, but because longer anoxic phases were applied, the total emission was lowered until 6 g NO-N d-1 or 0.01% of the N load. As a constant aeration flow rate was maintained during the SBR cycles, the lower emission was caused by a lower concentration of N2O and NO concentration that was emitted. The aeration during the SBR cycle was controlled by measuring a decrease in pH during aerobic phases, which was linked to AerAOB activity and a similar increase in pH during anoxic phases caused by addition of digestate (Wett, 2006). As a lower amount of digestate was added during the 2nd cycle, the pH increase took longer and caused an increase of the anoxic periods during the cycle with 25% compared to the first cycle at high loading. In addition, the individual aeration phases were 50% longer at lower loading. Although NO and N2O are mainly emitted during the aeration phases, the decreased total aeration time could not fully explain the decrease of 79 and 50% for the N2O and NO emission, respectively. NO2 45

Interplay of intermediates in formation of N2O/NO during OLAND

emission was less dependent of the aeration regimes and the emission increased with almost a factor 2 until 0.005% of N load. As NO2 emission is strongly linked to nitrite concentration in the liquid phase (Weissenbacher et al., 2007), the lower emission of N2O and NO were probably not caused by lower nitrite fluctuations.

In both cases, a lag phase in the N2O and NO emission compared to CO2 emission was observed between the start of the aeration in the beginning of the cycle (Fig. 3.3). Therefore, the question arose whether this occurred because the formation of NO and N2O did not start from the beginning or because this was just a matter of stripping and the formation a gas/liquid equilibrium. Because of the higher N2O/NO dynamics at high loading, a detailed follow-up of the first two hours of this cycle, characterized by highly changing operational conditions, was performed to answer this question and to try to understand mechanisms responsible for the emission under transient conditions better (Fig. 3.4).

Chapter 3

Figure 3.3: Emission of CO2, N2O, NO and NO2 of a SBR cycle at high (top) and low loading rate (bottom).

Interplay of intermediates in formation of N2O/NO during OLAND

Figure 3.4: Top: Concentrations of CO2, N2O, NO and NO2 measured in the defined air stream. Middle: Intermediate (NO2- and NH2OH), NH4+ and N2O concentrations in the liquid phase. Bottom: Dissolved oxygen (DO), NO3- and COD concentration in liquid phase. 48

Chapter 3

During the settling phase (end of previous cycle), oxygen concentrations were depleted and emissions in the gas phase were limited (Fig. 3.4). However, ammonium consumption (1.2 kg N d-1) followed by a NH2OH peak (0.2% of NH4+ consumption) could take place due to the sudden presence of a limited amount of oxygen (0.2 mg O2 L-1). The NH2OH peak, together with a decrease in ammonium and nitrite concentration, was accompanied with an increase in N2O(l+g) and NO(g) concentration, representing 0.67 kg N2O-N d-1 and 0.001 g NO-N d-1 or 56 and 0.08% of the NH4+ consumptions rate, respectively (Fig. 3.4, Table 3.1). In this phase several mechanisms could have played a role i.e. nitrifier denitrification, biological or chemical reaction of NO2- with NH2OH or nitrification-dependent NO and N2O formation (Chandran et al., 2011). Moreover, a second actor could have been responsible as a COD removal rate of 3.1 kg COD d-1 was observed at that time, which could indicate that denitrification could occur at a maximum rate of 1 kg N d-1. The latter could be another cause of the N2O and NO emission as the small amount of oxygen present (Fig. 3.4) could probably inhibit the N2O reductase during heterotrophic denitrification (Otte et al., 1996). Based on the calculated denitrification rate, this would mean that 67 and 0.1% of denitrified nitrogen ended up as N2O and NO, respectively. In this context, it should also be mentioned that autotrophic NO to N2O conversion is not inhibited by oxygen, which is a major departure from known pathways of heterotrophic denitrification. Taking into account the small increase in nitrate (0.28 kg N d-1) and decrease in nitrite (0.08 kg N d-1), it is plausible that all consumed ammonium was oxidized to nitrate and further reduced during denitrification. It is hard to distinguish the mechanisms at this point because of the interplay of several actors (AerAOB, AnAOB and denitrifiers) during this phase and because the in situ oxygen availability was unclear. The sudden pulse of NH4+ together with the start of the aeration at the beginning of the cycle, resulted in an initial increase of the NO and N2O production. However, the N2O emission in the gas phase showed a lag phase of about 30 minutes while the lag phase for NO was only 15 minutes and no lag phase was detected for CO2 emission (Fig. 3.4). The difference in lag phase is on one hand caused by the difference in water solubility and on the other hand a result of the sequential formation of N2O from NO. A first sharp increase in the ammonium oxidation rate from 1.2 to 320 kg N d-1 was directly followed by NO emission and N2O formation of which 13% remained in the liquid phase (Table 3.1, Fig. 3.4). During the following minutes, nitrite and NH2OH accumulation was observed together with a 1.3 and 5.4 fold increase of the NO and N2O emission, respectively (Table 3.1). In literature, it was 49

Interplay of intermediates in formation of N2O/NO during OLAND

described that the imposition of excessive NH3 loads triggers a higher ammonium oxidation rate, and potentially also a higher amo gene expression (Chandran et al., 2011). The latter could in turn result in NH2OH accumulation, which is in agreement with our observations. The same effect was observed when going from anoxic to oxic conditions (Chandran et al., 2011). AerAOB potentially need to oxidize the accumulated NH2OH to NO in addition to NO2- to prevent self-inhibition and more effectively derive energy. The latter can both be done biologically or chemically. Thus, the steep increase in N2O formation during this phase could mainly be explained by higher specific AerAOB activities resulting in oxidative formation of NO out of NH2OH. During the anoxic phase where only feeding was supplied, nitrite consumption occurred while ammonium concentrations gradually increased. Ammonium consumption rate during this phase was comparable to the other phases in the cycle (on average 304 kg N d-1), but decreased until 290 kg N d-1 at the end of this phase (Table 3.1). A build-up of NH2OH, as a result of increasing DO concentrations (from 0 to 0.08 mg O2 L-1) was observed while NO and N2O(l,g) concentrations were stable around 0.8 and 0.01% of the NH4 oxidation rate, respectively. From the nitrite consumption, an anoxic ammonium oxidation rate of only 0.31 kg NH4+-N d-1 was estimated. However, as oxygen still seemed present, aerobic oxidation of ammonium should have taken place combined with AnAOB activity to explain the total nitrogen loss of around 300 kg N d-1.

Table 3.1: Accumulation rate of the different intermediates and anoxic products relatively based on the total ammonium oxidation rate under different conditions during the SBR cycle. The ammonium oxidation is a combination of AerAOB and AnAOB activity, which seemed very balanced and stable over the cycle as no substantial nitrite accumulation took place. Conditions NH4+ Rv Accumulation rate (% of NH4+ removal rate) Feed Aeration DO (kg N d-1) NO2- NH2OH N2O (l) N2O NO Total (g)

No No 0.2 1.2 0.2 29.4 26.7 0.08 * Yes Yes 0 320 0.1 0.7 0.03 Yes** Yes 0 312 0.2 0.001 4.3 0.04 Yes No 0 -0.1 304 0.001 0.8 0.01 Yes On/off 0 -0.7 308 0.0001 1.6 0.004 * First 10 minutes of feeding and aeration; ** Second 10 minutes of feeding and aeration

56 0.8 4.6 0.8 1.6

During the subsequent transient phase with rapidly on/off aeration regimes (on average 6 minutes aeration, 6 minutes without aeration), DO levels sharply increased for the first time (up to 0.7 mg O2 L-1). Moreover due to continuous NH4+ feeding, the aerobic oxidation had to 50

Chapter 3

start at higher NH4+ concentrations (170 mg N L-1 instead of 74 mg N L-1). From the moment oxygen was present, aerobic ammonium oxidation could start which resulted in a peak in the total ammonium oxidation of 335 mg N d-1, which was even higher than during the start of the cycle. The consequent NH2OH accumulation and NO and N2O emission could be the result of both oxidative and reductive formation of NO (Chandran et al., 2011). Both the ammonium peak and as a consequence specific activity peak and the transition from anoxic to oxic conditions favor oxidative formation of NO (Chandran et al., 2011), indicating that the NH2OH availability should have been important. Indeed, as all nitrogen compounds in the liquid phase remained constant expect for the NH2OH concentration, the latter could be linked to the increasing NO emission (Fig. 3.4). Due to the low water solubility of NO, the subsequent increased N2O formation was probably avoided. As already suggested (Kampschreur et al., 2009b), also in our study AerAOB seem to be the major contributor to the NO and N2O emission. Sudden pulses of O2 always resulted in increased NO and N2O formation (Fig. 3.4) and these emissions were accompanied with NH2OH accumulation. Moreover, peaks in the ammonium oxidation rate (start of the cycle and start of transition phase, Fig. 3.4) increased the NO and N2O emission. As in both cases NH2OH accumulation was observed, these measurements indicated a strong link between NH2OH concentration in aerobic conditions and emissions of these harmful gases. This could indicate that, under these highly dynamic conditions, nitrifier denitrification played only a minor role in the system and that the NO and N2O formation mainly followed the NH2OH route (either chemically or biologically, see Fig. 3.1). This suggests that a better understanding of the conditions that lead to transient NH2OH accumulation could help to develop operational strategies to reduce NO and N2O emission from one-stage partial nitritation/anammox systems.

Conclusions

Three main conclusions could be drawn from these measurements: ď&#x201A;ˇ Highly transient conditions, implying peaks in aerobic ammonium oxidation rates resulted in increased NO and N2O emissions. ď&#x201A;ˇ Peaks in NO and N2O emission were always accompanied with NH2OH accumulation. Therefore, it seemed that biological or chemical production of NO and N 2O from

Interplay of intermediates in formation of N2O/NO during OLAND

NH2OH is the most important cause for the emission during transient one-stage partial nitritation/anammox. ď&#x201A;ˇ Operation at more stable conditions and avoidance of NH2OH accumulation could be key parameters to decrease the NO/N2O emissions.

Acknowledgements

H.D.C. is a supported by a PhD grant from the Institute for the Promotion of Innovation by Science and Technology in Flanders (IWT-Vlaanderen, number SB-81068). The investigations at the Strass treatment plant were also supported by the Austrian Federal Ministry of Environment.

Sludge settler (WWTP Strass, Austria) 54

Chapter 4

Chapter 4: OLAND maximizes net energy gain in technology schemes with anaerobic digestion 1

Treatment of digestates by OLAND

Autotrophic nitrogen removal processes are the economically preferred method for nitrogen containing wastewaters low in organic carbon. Landfill leachate, urine and industrial wastewaters from coke-ovens (Toh and Ashbolt, 2002), tanneries (Abma et al., 2007), semiconductor plants (Tokutomi et al., 2011a) and the fertilizer industry (Alberta Environment, 1999) have these characteristics as such, while others obtain an optimal COD/N ratio after anaerobic digestion. Since this chapter focuses on the impact of OLAND on the energy balance of systems with energy recovery by anaerobic digestion, the above-mentioned streams are not covered in this chapter. In what follows, four important combinations of anaerobic digestion with subsequent OLAND-based nitrogen removal are discussed i.e. the treatment of (I) the organic fraction of municipal solid waste (OFMSW), (II) manure-based agricultural waste, (III) starch/sugar-based agro-industrial wastes and (IV) sewage-based organics. These four application domains in which OLAND minimizes energy use for digestate treatment, will be discussed in detail in a first section. In a second section, OLAND application as an active step in minimizing the energy usage of a wastewater treatment plant with anaerobic digestion as central treatment step is discussed.

Chapter redrafted after: De Clippeleir, H., Vlaeminck, S.E., Courtens, E., Verstraete, W., Boon, N., in press. Oxygen-limited autotrophic nitrification/denitrification maximizes net energy gain in technology schemes with anaerobic digestion Renewable Energy Sources. Academy Publish, Wyoming, U.S.A. 55

OLAND maximizes net energy gain in systems with anaerobic digestion

BOX 1: General assumptions for energy calculations The energy values expressed as kWh are considered electrical energy values. If total or thermal energy is considered the indices ‘tot’ and ‘th’ were used, respectively.

Discharge limits A discharge limit of 135 mg COD L-1 and 15 mg N L-1 for starch-based agro-industrial and OFMSW biogas plants, and 5 g COD IE-1 d-1 and 2.5 g N IE-1 d-1 for sewage plants (Siegrist et al., 2008), was taken into account. For manure-based agricultural waste, a specific case study from literature was considered (Karakashev et al., 2008), in which 4% of the original COD and 13% of the original nitrogen present in the digestate was send back to the land.

Energy production through anaerobic digestion For anaerobic digestion of OFMSW, starch-based agro-industrial waste, source separated black water and primary/A-stage and secondary sludge, COD removal efficiencies of 87, 90, 80, 60/60 and 35%, respectively, were taken into account (Vaz et al., 2008; Abma et al., 2010; de Graaff et al., 2010; Hernández Leal et al., 2010). Moreover in every case, a constant biogas production (0.5 m3 kg-1 COD removed) and CH4 content of the biogas (65%) was considered. For the electrical energy recovery, a total energy content 10 kWhtot m-3 CH4 and electrical recovery efficiency through combined heat and power (CHP) unit of 38% (Wett et al., 2007) was used. Together these factors lead to an electrical energy recovery of 1.24 kWh kg-1 COD removed. The composition of the liquid fraction of the digestate was calculated taken into account an anaerobic yield factor of 0.05 kg COD in biomass kg-1 COD removed and a COD/N ratio in the sludge of 15 (van Haandel and van der Lubbe, 2007). For each application, it was considered that nitrogen which was not assimilated, ended up in the digestate. For COD the assumption differed depending on the application. In sewage-based systems, COD which was not converted was considered to end up in the solid fraction of the digestate, while in the other applications, due to a lack of full-scale data, a worst case scenario was calculated in which all COD that was not converted to biogas or sludge ended up in the liquid fraction of the digestate.

Energy consumption for the different aerobic/anoxic treatment steps From

the

stoichiometry

Table

1.2

(Chapter

1),

the

oxygen

demand

for -1

nitrification/denitrification and OLAND could be calculated and was 4.34 kg O2 kg -1

removed and 1.81 kg O2 kg N removed, respectively. Furthermore assuming an oxygen 56

Chapter 4

demand for COD removal of 1 kg O2 kg-1 COD removed, an actual electrical oxygen transfer efficiency of 1 kg O2 kWh-1 (van Haandel and van der Lubbe, 2007), the resulting energy requirements for the different treatment schemes were calculated. In case post-denitrification was considered the energy consumption was calculated based on a volumetric nitrogen removal rate of 1 kg N m-3 d-1 and an energy demand for mixing of 10 W m-3 reactor. Postdenitrification could be performed by either adding an external carbon source or by using the raw waste stream as carbon source, which results in an additional cost or a lower energy recovery, respectively. The assimilation due to heterotrophic growth was calculated taken into account a yield factor of 0.5 kg COD in biomass kg-1 COD removed and assuming a COD/N ratio in the sludge of 15, the sludge production was calculated (van Haandel and van der Lubbe, 2007). For autotrophic conversions sludge production was neglected.

Additional assumptions for sewage plants A loading of 135 g COD IE-1 d-1 and 10 g N IE-1 d-1 was used in all sewage treatment schemes (Verstraete and Vlaeminck, 2011). It was considered that during primary settling 20% of the COD and 10% of the total N could be separated from the water flow. During enhanced primary settling the COD removal efficiency could be improved to 40%. For the harvesting of primary sludge, no additional energy demand was taken into account. For the application of a highly loaded activated sludge step (A-step) an electrical energy requirement of 0.11 kWh kg-1 COD removed was considered (SalomĂŠ, 1990) and a COD removal efficiency of conventionally 60 and 50% was considered for centralized and decentralized systems. Due to a lack of data for the other domains of application, sludge dewatering was only taken into account for the sewage-based applications at an electrical energy demand of 0.15 kWh kg-1 COD (Zessner et al., 2010).

1.1

Organic fraction of municipal solid waste (OFMSW)

1.1.1 State of the art The Food and Agriculture Organization (FAO) of the United Nations showed that one third of all food produced for human consumption, namely 1.3 billion tons wet biomass each year (84-115 g COD IE-1 d-1), ends up as municipal solid waste (Monson et al., 2007; Vaz et al., 2008; FAO et al., 2011). This waste stream consists of kitchen (food) waste (22%), paper and cardboard (23%) and garden waste (16%) (Burnley, 2007). These organic wastes are classified as organic fraction of municipal solid waste (OFMSW). If the OFMSW would be digested anaerobically with a methane yield of 0.3 mÂł CH4 kg-1 COD (Monson et al., 2007; 57

OLAND maximizes net energy gain in systems with anaerobic digestion

Vaz et al., 2008; FAO et al., 2011), an electrical energy production of 101-131 Wh IE-1 d-1 or 1.1-1.3 kWh kg-1 COD is expected. This represents about 2% of the total global electrical energy utilization in 2008, which amounted to 16 1012 kWh (IEA, 2010). Besides electrical energy recovery, digesting food waste lowers the amounts of solids send to landfills and thus reduces greenhouse gas emissions and transportation costs. The high potential of energy recovery out of those wastes together with the need of new waste management strategies gave rise to OFMSW specialized biogas producing plants. Energy production is the main objective of these plants, and one therefore wants to maximize the ‘energy index’, i.e. the ratio of the produced and consumed electrical energy.

Municipal solid waste generally has a higher risk to contain toxic and inhibitory compounds than wastewater. These compounds can upon entering the reactor, diffuse quickly in the diluted slurry and hereby negatively affect the microorganisms (Vandevivere et al., 2002). It is important to make the distinction between so-called ‘dry’ and ‘wet’ anaerobic digesters within the OFMSW biogas plants, treating respectively ‘semi-solid’ and ‘liquid’ waste streams. ‘Dry’ digesters are fed with OFMSW characterized by a low water content (< 15%; De Baere, 2006), and operate in a semi-solid way resulting in biogas and a solid digestate that generally is converting into high quality compost or directly applied for agricultural purposes. The DRANCO, Valorga and Kompogas processes are the most common technologies for this type of digestion (Six and Debaere, 1992; Wellinger et al., 1993; de Laclos et al., 1997). Semi-solid systems have become prevalent in Europe, making up 60% of the single-stage digester capacity installed (De Baere, 2006). On the other hand, ‘wet’ digesters deal with separately collected food waste with a high moisture content (74-90%, Zhang et al., 2007). For these streams, the upflow anaerobic sludge blanket (UASB) technology and the continuous stirred-tank reactors (CSTR) are the most applied technologies. The latter plants produce next to biogas a solid and a liquid digestate, that can be spread on agricultural land after pasteurization, in the simplest case. In most cases further treatment of the nitrogen-rich digestate is required, which has a big influence on the overall electrical energy balance. The electrical energy demand for pasteurizing is 12%, while 69% is needed to cool the digestate and subsequently treat by conventional nitrification/denitrification (De Sousa and Vaz, 2009). The deficit in organic carbon to allow full denitrification implies that less can be digested (lower energy recovery) or that methanol or another carbon source needs to be added to the liquid digestate (higher operational costs; Table 1.3 Chapter 1). Until now, full-scale OLAND-type reactors are not yet applied for the treatment of liquid digestates from OFMSW 58

Chapter 4

biogas plants. However, successful OLAND treatment has been demonstrated on digestates from co-digestion plants with food waste in centralized (Wett et al., 2007) or decentralized (Zeeman et al., 2008) sewage plants, suggesting the feasibility of OLAND implementation in OFMSW plants.

1.1.2 Implication of OLAND application on the energy balance The main goal of OFMSW biogas plants is to produce energy from biodegradable municipal solid waste. As mentioned above, nitrogen treatment will only have an impact on the total energy balance of â&#x20AC;&#x2DC;wetâ&#x20AC;&#x2122; digesters, which produce a liquid digestate that cannot be discharged as such.

BOX 2: Energy calculation from an OFMSW biogas plant Assuming a COD loading rate to the OFMSW digestor of 81 g COD IE-1 d-1 and a digestibility of 87% (Vaz et al 2008), a biogas production of 0.02 m 3 CH4 IE-1 d-1 and electricity gain of 87 Wh IE-1 d-1 or 1 kWh kg-1 COD can be achieved. The digestate still contains 7 g COD IE-1 d-1 (maximum value) and 6 g N IE-1 d-1, which should be treated biologically by OLAND/post-denitrification or by conventional nitrification/denitrification. Discharge limits are assumed at 0.0025 g N IE-1 d-1 and 0.005 g COD IE-1 d-1, and the volume at 0.2 Lsolid

municipal waste

IE-1 d-1. For the OLAND scenario, 11.7 Wh IE-1 d-1 is needed for

ammonium oxidation. For aerobic COD degradation, an electrical energy demand of 4.9 Wh IE-1 d-1 is expected. Because OLAND oxidizes 11% of the ammonium to nitrate, 0.1 Wh IE-1 d-1 is needed for mixing in the post-denitrification reactor to meet discharge limits. Hence, the electricity consumption of the OLAND scenario amounts to 17 Wh IE-1 d-1. Using conventional nitrification/denitrification, addition of an external carbon source is needed or raw waste should be used for denitrification (lower biogas recovery). In the latter case, the electrical energy recovery will decrease to 67 Wh IE -1 d-1 and a total energy consumption of 22 Wh IE-1 d-1 is needed to fully polish the digestate. As organic waste is highly digestible (efficiency around 80-90%, BOX 1), in general an electrical energy recovery of 1.1 kWh kg-1 CODin can be expected. The liquid digestate is rich in nitrogen and devoid in organic carbon (4800-9700 mg COD L-1, 1119-1500 mg N L-1, COD/N 4-11). For further digestate polishing by OLAND with post-denitrification and by conventional nitrification/denitrification, an electricity consumption of 17 and 22 Wh IE-1 d-1 or 0.19 and 0.25 kWh kg-1 COD, respectively, was estimated (BOX 2). It can be concluded 59

OLAND maximizes net energy gain in systems with anaerobic digestion

that by applying OLAND for this application a net electrical energy gain of 0.8 kWh kg-1 CODin compared to 0.5 kWh kg-1 CODin for conventional treatment can be obtained.

1.2

Manure-based agricultural waste

1.2.1 State of the art Animal waste streams are characterized by a high nitrogen and organic carbon content. Although the latter implies a high energy potential, particularly for pig manure the relatively high free ammonia concentrations can significantly disturb biogas formation (Chen et al., 2008). To avoid this, an acidifying digestion (pH 6-7; lower free ammonia levels) is sometimes performed, converting the more complex organics into volatile fatty acids (Chen et al., 2008). Biogas production is limited in this step. By separation of the solid and liquid fraction afterwards, the volatile fatty acids in the liquid fraction can be converted to biogas in an anaerobic digester (e.g. UASB). In this step around 57% of the present COD can be converted to biogas in the presence of about 4 g NH4+-N L-1 (Karakashev et al., 2008). Given the high organic carbon content of pig manure (around 70 g COD L-1), the digestate still contains a considerable amount of soluble organics (10 g COD L-1) (Karakashev et al., 2008). The latter levels and some specific compounds might be inhibitory for AnAOB (DapenaMora et al., 2007) and their removal will enhance the success of OLAND treatment. A separate oxidation step can decrease COD levels to about 3.5 g COD L-1 and COD/N ratios to about 2. Given the choice of a separate oxidation, most studies on autotrophic nitrogen removal on digested manure focused on two separate nitrogen removal steps (partial nitritation â&#x20AC;&#x201C; anammox) (Van Hulle et al., 2010), incorporating the COD oxidation in the partial nitritation stage. Removal efficiencies obtained with AnAOB-based technologies for digested manure are generally around 70% (Van Hulle et al., 2010), and thus lower than for less complex digestates such as sludge reject water (Table 1.4, Chapter 1). Research showed that the removal efficiency of AnAOB-based processes was not only dependent on the COD/N ratio obtained after digestion or post-digestion, but also on the absolute COD levels. COD concentrations above 142 and 242 mg COD L-1 for instance ceased the AnAOB activity for post-digested manure and partially oxidized digestate, respectively (Molinuevo et al., 2009). These values should however not be generalized, and are likely dependent on the test conditions and acclimatization.

Treatment schemes with autotrophic nitrogen removal in a one-stage (Karakashev et al., 2008) or two-stage (Hwang et al., 2006) setting have thus far only been tested in batch or in 60

Chapter 4

continuous lab-scale reactors. The lack of pilot- and full-scale investigations does not yet allow to discuss their environmental and economical sustainability (Karakashev et al., 2008). Due to the stringent environmental regulations concerning the application of the manure nutrients as direct fertilizer on agricultural land, treatment of digestates is more and more becoming a necessity. At this moment, the amount of added nitrogen may not exceed 170 kg N ha-1 yr-1 (Oenema, 2004), and this amount is even likely to be decreased in the future. This should provide a stimulus to validate schemes with autotrophic nitrogen removal on a larger scale. 1.2.2 Implication of OLAND application on the energy balance Because of the relatively low biodegradability of manure COD and the relatively high nitrogen content, electrical energy recovery through anaerobic digestion is more difficult than for other streams. For pig manure for instance, only 29 kWh m-3 raw manure or 0.4 kWh kg-1 CODmanure can be recovered as electricity (BOX 3). The implementation of OLAND in the treatment scheme requires an electrical input of 17.6 kWh m-3 raw pig manure or 0.25 kWh kg-1 CODmanure (BOX 3), taking into account the energy need for solid/liquid separation, COD oxidation and OLAND treatment. The electrical energy demand for conventional nitrification/denitrification is somewhat higher, i.e. 0.27 kWh kg-1 CODmanure (BOX 3). Therefore, the net electrical energy gains are 11.4 and 10.2 kWh m-3 raw pig manure, or 0.16 and 0.15 kWh kg-1 CODmanure for the treatment with OLAND and nitrification/denitrification, respectively (BOX 3). It should be mentioned that external carbon will be required for nitrification/denitrification, given the relatively low BOD/COD ratio of manure (Lemmens et al., 2007). So, although additional energy gain by implementation of OLAND seems minor (BOX 3), the cost for an external organic carbon source can be avoided in this way.

Direct biological treatment of the liquid manure fraction is often applied to avoid the cost of an external carbon source. This approach does not recover energy and has an average net energy consumption between 16 and 22 kWh m-3 raw manure (Lemmens et al., 2007), which is 26 and 33 kWh m-3 raw manure higher than the electricity consumption of the schemes with anaerobic digestion.

OLAND maximizes net energy gain in systems with anaerobic digestion

BOX 3: Energy calculation of a manure-based agricultural plant As an example, a treatment scheme was chosen for piggery manure with thermophilic anaerobic digestion followed by a centrifugation step. The liquid fraction (~90 vol% of the raw manure) is subsequently treated in a UASB reactor to produce biogas. The digestate is then treated in a separate COD oxidation step, followed by an OLAND reactor. The mass balances of COD and N are shown in Fig. 4.1. If conventional nitrification/denitrification would be applied, the COD oxidation step can be left out of the scheme. From the proposed treatment scheme 25 kg COD m-3 raw manure can be recovered as biogas, hence the electrical recovery is 29 kWh m-3 raw manure. A solid/liquid separation step, often using decanter centrifugation, consumes on average 4 kWh m-3 raw manure (Lemmens et al., 2007). An electricity consumption of 11.8 kWh m-3 raw manure was calculated to convert COD to CO2 and sludge in the oxidation stage. It was assumed that nitrogen losses were minimal, and mainly caused by volatilization of ammonia and nitrogen assimilation. The subsequent OLAND reactor has an electrical energy demand of 1.8 kWh m-3 raw manure. The total electrical energy requirement for this system is therefore 17.7 kWh m-3 raw manure. When conventional nitrification/denitrification is applied, the total electrical energy demand only slightly increases to 18.8 kWh m-3 raw manure due to the absence of the aerobic COD degradation. However, external organic carbon source addition will be needed in this scheme, increasing the operational costs. Addition of raw manure as carbon source for denitrification seems not a good option due to the low biodegradability and high nitrogen content.

Figure 4.1: Possible innovative treatment scheme for pig manure, with COD and N quantities expressed for 1 m3 of manure (Karakashev et al., 2008).

Chapter 4

1.3

Sugar/starch-based agro-industrial waste

1.3.1 State of the art Several full-scale OLAND-type processes are used to treat digestates in food industry, yeast factories and distilleries (Table 1.4, Chapter 1). For example wastewater from potato processing can be treated by anaerobic digestion to recover energy from the present organics (57 kg COD m-3), followed by struvite precipitation to recover the phosphorous and OLAND treatment to remove the residual nitrogen (0.7 kg N m-3 digestate) (Abma et al., 2010). Both one-step and two-step autotrophic nitrogen removal processes for potato processing plants are operational at full-scale (Abma et al., 2010; Desloover et al., 2011a). Another example can be found in Asian food culture, where monosodium glutamate is a popular flavor. It is produced by fermentation of rice, starch and molasses, which finally creates a wastewater rich in suspended solids (SS; 200-1000 mg SS L-1), COD (1500-60000 mg L-1), NH4+ (200-15000 mg N L-1) and sulfate (3000-70000 mg L-1) (Zhang et al., 2008). This wastewater is traditionally treated by physico-chemical and biological methods decreasing the wastewater content to around 200-270 mg SS L-1, 1000-1400 mg COD L-1 and 150-350 mg N L-1 (Zhang et al., 2008). In case of a low biodegradability of the residual COD, denitrifier overgrowth of AnAOB can be avoided, making the OLAND process feasible. Nitrogen removal efficiencies above 80% were obtained at full-scale OLAND plants treating effluent from monosodium glutamate wastewater in China (Table 1.4, Chapter 1).

In general, anaerobic digestion of carbon-rich, digestible industrial wastestreams can significantly decrease the BOD/N ratio below 3-4. The digestate then comes into the scope of subsequent OLAND treatment.

1.3.2 Implication of OLAND application on the energy balance Most industrial full-scale OLAND applications are situated in the food industry (Table 1.4, Chapter 1). As wastewater from several food processing companies is directly digestible, the electrical energy recovery from this waste stream is expected to be around 1.1 kWh kg-1 CODin. Nitrogen removal through OLAND and post-denitrification has an electricity demand of 0.11 kWh kg-1 CODin. This leads to an overall net electrical energy gain of 1.0 kWh kg-1 CODin (BOX 4). In the other scenario, with nitrification/denitrification and bypassing some organic carbon source to the denitrification reactor, the net electrical energy gain is reduced with 27% to 0.73 kWh kg-1 CODin (BOX 4). OLAND implementation on digestible industrial 63

OLAND maximizes net energy gain in systems with anaerobic digestion

wastewater can hence decrease the electricity consumption with a factor 2, having a significant impact on the overall energy balance.

BOX 4: Energy calculation of a starch-based agro-industrial plant A potato factory producing 17000 kg COD d-1 and 1000 kg N d-1 (at 3000 m3 d-1) obtained a COD removal efficiency during anaerobic digestion of 90% (Abma et al., 2010). The electrical energy recovery from biogas production is 19000 kWh d-1 or 1.1 kWh kg-1 CODin. In the worst-case assumption, the COD that is not converted to biogas ends up in the digestate, i.e. 316 mg COD L-1, next to 316 mg N L-1. The latter should be treated biologically by OLAND/post-denitrification or by nitrification/denitrification. Discharge limits of 15 mg N L-1 and 135 mg COD L-1 were used. For the scheme with OLAND, electrical energy demands of 1700 kWh d-1, 142 kWh d-1 and 11 kWh d-1 are needed for the OLAND reactor, aerobic COD removal and final denitrification, respectively. Thus, the total electrical energy need for digestate amounts to 1900 kWh d-1 or 0.11 kWh kg-1 CODin. In contrast to this scenario, a part of the raw waste can bypass the digestor and serve as carbon source for denitrification. In this case the electrical energy recovery decreases to 16000 kWh d-1 (0.9 kWh kg-1 COD) and the electrical energy consumption increases to 3500 kWh d-1 (0.20 kWh kg-1 COD). The overall electricity gain of this scenario is hence 0.73 kWh kg-1 COD.

1.4

Sewage-based organics

1.4.1 State of the art: centralized treatment As the main treatment step during sewage treatment is based on heterotrophic, aerobic conversions, sludge production during conventional activated sludge (CAS) treatment is about 0.5 kg COD converted to sludge biomass per kg COD converted. The daily specific sludge production varies between 40 and 60 g DM IE-1 d-1, with the lowest values for CAS systems with nitrogen treatment and the highest production for CAS systems with additional P removal (Zessner et al., 2010). Sludge treatment by anaerobic digestion in combination with land application is the most sustainable approach due to the low emission and low energy consumption (Suh and Rousseaux, 2002). The digestate formed as a result of sludge digestion, normally only accounts for 1% of the influent water flow but for 15-20% of the nitrogen load of the CAS system. Therefore, it should be further treated before discharge (Fux and Siegrist, 2004). 64

Chapter 4

Figure 4.2: Sewage treatment schemes based on CAS (A, B) and the A/B process (C, D) with and without OLAND implementation in the side stream.

OLAND maximizes net energy gain in systems with anaerobic digestion

Digestates from municipal sludge digestion are characterized by a high nitrogen content (0.4-1 g N L-1) while a high proportion of the biodegradable COD is already digested, obtaining in most cases biodegradable COD (bCOD)/N ratios of 1-2. Due to the optimal composition of the sludge reject water, and provided the presence of qualified operators at CAS systems and the possibility to treat the OLAND effluent in the existing treatment system, around 70% of the full-scale OLAND-type reactors are applied in this area. Long start-up periods (order of years) were needed for the first full-scale applications (van der Star et al., 2007). At present, due to the possibility to acquire active biomass from other installations, start-up periods are ranging from 1-2 months for suspended growth systems (personal communication, Bernhard Wett) and 3-6 months for systems based on attached growth (Christensson et al., 2011). Since similar total nitrogen removal efficiencies are obtained in the different reactor types (Table 1.4, Chapter 1), the choice of technology mainly depends on the footprint area availability and the importance of energy efficiency. Moving bed bioreactors (MBBR) consume almost 5 times more electrical energy to remove the same amount of nitrogen (5.63 compared to 1.13 kWh kg-1 N) and moreover require lower volumetric loading rates to obtain optimal performance (Wett et al., 2007; Christensson et al., 2011). In most cases however, the reactor choice is connected with the constructor choice for implementation of a full-scale OLAND reactor, as every constructor is known for his operation methodology and design parameters.

As the implementation of anaerobic digestion in municipal WWTP is still growing, with Sweden as one of the leading countries, digesting already 83% of the sludge (Lantz et al., 2007), the OLAND application in this area is immerging. It is slowly becoming a standard treatment method in municipal WWTP as high and stable nitrogen removal performance has been demonstrated and a positive influence on the energy balance of the WWTP has been established.

1.4.2 State of the art: decentralized treatment Since 70-80% of the costs of municipal sewage management derive from the sewerage system (Bieker et al., 2010), and since additionally during the long transport around 30% of the dissolved COD (potential energy) in wastewater is lost (Huisman et al., 2004), decentralized concepts are suggested, aiming at a maximum recovery of energy and nutrients. Source separation can prevent pollutant dilution, and hence renders recovery feasible. At the household level, three main streams can be separated (Table 4.1). Urine or yellow water 66

Chapter 4

contain most of the soluble nutrients and due to the high concentration, recovery of N and P is possible (Otterpohl, 2002). From brown water (faeces), energy can efficiently be recovered due to the concentrated organic carbon content by anaerobic digestion, which simultaneously allows hygienisation towards an organic soil improver. In some cases, urine and faeces are collected together as black water which can be even concentrated by the use of vacuum toilets (Zeeman et al., 2008). Grey water, a less concentrated stream from showers, bath, washing machines, kitchens etc can be treated with small efforts as its temperature is elevated and the nutrient content is low enough to require nutrient elimination to reach service water quality (Cornel et al., 2011). The grey water can for example by recycled as toilet flushing water, saving 30% of the potable water consumption (Cornel et al., 2011).

Table 4.1: Distribution of the daily COD and nitrogen loads in different wastewater streams (Henze, 1997; Otterpohl, 2002). Sewage Grey water Brown water Yellow water (faeces) (urine) Volume 25-100 25-100 0.05 0.5 (L IE-1 d-1) Compound (g IE-1 d-1) % % % COD 135 41 47 12 N 10 3-8 8-10 77-87 P 2.5 10-20 20-40 50-60

At present, few source-separated schemes incorporate a complete treatment scheme in operation. A semi-centralized approach (20 000 IE) including separate grey and black water treatment and biogas production from bio-waste was applied in Qingdao (China). Due to the incorporation of bio-waste digestion, this concept could provide the electric energy demand within this semi-centralized treatment process (Cornel et al., 2011). Another pioneering project in Sneek (The Netherlands), where grey water and concentrated black water are collected from 32 houses, has shown to be feasible and profitable (Zeeman et al., 2008; Verstraete and Vlaeminck, 2011). In the latter concept concentrated black water vacuum collection consumes in terms of electricity 10-27 Wh IE-1 d-1, but it uses 7 times less water (WRS, 2001), thus allowing to treat a very concentrated black water stream that is directly suitable for anaerobic digestion. Co-digestion of bio-waste and sludge from the grey water treatment (A/B system) can further increase the biogas output. However, due to the small scale of this project so far, electrical power production from biogas formation is not feasible and thus only heat is used by applying a co-combustion, which switches between biogas and natural gas. The digestate, containing 90% of the nitrogen load of a household is treated by an 67

OLAND maximizes net energy gain in systems with anaerobic digestion

energy-friendly OLAND RBC, decreasing the overall energy need of the plant. Compared to the centralized approach where around 15-20% of the nitrogen load ends up in the digestate, the importance of nitrogen removal technology choice increases significantly in the decentralized approach. Therefore, OLAND technology can drastically change the energy consumption and operational costs in the latter case. Due to the concentration of the nutrients in black water, recovery of phosphorous and part of the nitrogen can also be achieved by struvite precipitation.

1.4.3 Implication of OLAND application on the energy balance Current sewage treatment systems are mainly designed to remove organics from wastewater although the latter can be regarded as a source of energy. Nitrogen removal in the conventional activated sludge (CAS) system requires a lot of electrical energy to obtain full nitrification (Table 1.3, Chapter 1, Fig. 4.2). Moreover, it uses organic carbon to denitrify nitrate to nitrogen gas. Aeration constitutes over 60-70% of the electrical energy consumption of CAS systems with and without anaerobic digestion, respectively (Zessner et al., 2010). In an average CAS system the total electrical energy consumption is around 96 Wh IE -1 d-1, which can be covered for 44% by electrical energy recovery through anaerobic digestion of the primary and secondary sludge (Table 4.2, Fig. 4.2). Application of a two-stage activated sludge system or A/B Verfahren system, can increase the role of electrical energy recovery by anaerobic digestion to 90% of the electrical energy consumption (Table 4.2). Implementation of OLAND for digestate treatment in the side line of the CAS and A/B Verfahren system, without other changes, can not increase the total electrical energy gain (Table 4.2). The main reason for this is that the decrease in electrical energy requirement for nitrogen removal can not counteract the increase in aerobic COD removal due to the lower denitrification rate in the main line. Therefore on first side, the implementation of the OLAND process does not seem to have a significant effect on the total energy balance. However, OLAND in the side line allows higher organic carbon removal through sludge by enhanced primary settling (e.g. by addition of flocculants) or improved highly loaded activated sludge treatment (A-step of A/B system), because less organic carbon is needed for final nitrogen polishing through denitrification and can thus be recovered as biogas. OLAND implementation for digestate treatment in CAS systems can lower the overall plant energy requirements with about 50% (Table 4.2; Siegrist et al., 2008) due to a higher electrical energy recovery and a decrease in aerobic COD degradation. Furthermore, according to the theoretical calculations (Table 4.2) and the in practice experiments of Wett et al. (2007), energy autarky by including OLAND in 68

Chapter 4

the sidestream of a two-stage activated-sludge (AS) process (‘A/B Verfahren’) can be obtained if the A-step efficiency is high enough. In general, OLAND implementation in the side line of centralized WWTP can allow a higher net electrical energy gain if at the same time a higher organic carbon recovery through biogas production is applied to make profit from the lower organic carbon requirements for denitrification.

Table 4.2: Energy demand and gain of municipal WWTP schemes. CAS: conventional activated sludge treatment; OL: OLAND reactor in side line, treating sludge reject water; ° enhanced primary settling applied; AX/B: A/B system with a COD removal efficiency of X in the A-stage; OL: OLAND in the main stream; *corrected for COD removal through denitrification Oxygen and energy demand Electrical energy gain (Wh IE-1 d-1)

COD removal* N removal main line OLAND in side line Energy consumption A step Pumping/mixing Sludge dewatering Total energy consumption Biogas-based energy production Net energy gain

Case 1 CAS -37.6 -32.0 / / -20.0 -6.1 -95.7 +42.3 -53.4

Case 2 CASOL -41.9 -24.7 -3.1 / -20.0 -6.1 -95.8 +42.3 -53.5

Case 3 CASOL° -25.9 -28.6 -2.4 / -20.0 -6.2 -83.1 +56.4 -26.7

Case 3 A60/B -9.7 -33.2 / -8.9 -20.0 -6.4 -78.3 +70.6 -7.7

Case 4 A60/BOL -15.6 -23.2 -4.2 -8.9 -20.0 -6.4 -78.3 +70.6 -7.7

Case 5 A75/BOL -3.6 -26.2 -3.7 -11.1 -20.0 -6.5 -71.1 +81.2 +10.1

BOX 5: Sensitivity analysis of energy calculations The assumptions made regarding the energy calculations of the centralized wastewater treatment schemes, have a high impact on the obtained energy gain. The following parameters showed the highest impact on the total energy balance of the systems:  The actual oxygen transfer efficiency The actual oxygen transfer efficiency is dependent on the type of aerator (surface, diffusers), the reactor design and the wastewater and operational properties. A high oxygen transfer efficiency can drastically decrease the total electrical energy consumption (Fig. 4.3). Moreover, in A60/B systems at oxygen transfer efficiencies above 1.8 kg O2 kWh-1 a net energy production independent of the digestibility of primary sludge can be obtained. In CAS systems it is clear that a net electrical energy gain is hard to obtain without increasing the primary sludge production.  The digestibility of primary sludge Digestion efficiencies of mixtures of primary and secondary sludge are reported around 50%. However, a difference in digestion efficiencies between both types of sludge exists. In the 69

OLAND maximizes net energy gain in systems with anaerobic digestion

main calculations an efficiency of 60 and 35% was considered for primary and secondary sludge, respectively. Depending on the type of wastewater treated and the method of primary sludge production (settler, enhanced settler, A-stage), the digestibility of primary sludge can differ. Applying this deviation to the calculation, it could be shown that the net energy gain can be significantly influenced by this parameter (Fig. 4.3). For A60/B systems this influence starts earlier and is higher because of the higher proportion of primary sludge production compared to CAS systems.

Figure 4.3: Net electrical energy gain (Wh IE-1 d-1) of CAS system (top) and A60/B system (bottom) in function of the actual oxygen transfer efficiency and primary sludge digestibility. 70

Chapter 4

ď&#x201A;ˇ Energy for pumping In all schemes the electrical pumping energy was considered constant (20 Wh IE -1 d-1) and represented around 22-31% of the total electrical energy consumption. Depending on the treatment scheme, the amount of nitrogen that should be denitrified differs. The latter is mainly steered by applying a certain recirculation ratio in the activated sludge system in the main line, which means that a lower denitrification rate implies a lower recirculation ratio and thus a lower pumping cost. A lower electrical pumping energy due to the absence of recirculation with for example 25% (to 15 Wh IE-1 d-1; Kartal et al., 2010a), can decrease the total energy consumption with 5-8%. ď&#x201A;ˇ Aerobic yield of heterotrophs The aerobic COD yield of heterotrophs in the activated sludge system was considered 0.5 kg COD in cellular microbial biomass kg-1 COD removed in the calculation. A deviation of this factor with 20% to 0.6 or with 40% to 0.7 kg COD in biomass kg-1 COD removed could increase the net electrical energy gain in a CAS system with 15 and 65%, respectively. The latter indicated that for a CAS system and A60/B system an increase in the yield to 0.8 and 0.52 kg COD in biomass kg-1 COD removed, respectively, could result in an energy selfsufficient system in terms of electrical energy. On decentralized level, a size of 50 000 to 100 000 IE is recommended to obtain electrical energy recovery from biogas (Bieker et al., 2010). Current research gives reasons to believe that investment costs and income from energy are going to balance after about 15 to 20 years (so far integrated decentralized systems may be more expensive in investments) while the operation costs of these systems seem to be only a fraction of the costs of centralized systems. The main reason for the latter is the energetic use of solid waste and sewage sludge within the decentralized approach, while this is more difficult in the conventional centralized approach due to the high dilution (Bieker et al., 2010). Since more than 75% of the nitrogen load is present in the liquid fraction of digested concentrated black water, which was separated from grey water (Table 4.1), the application of OLAND vs. nitrification/denitrification has a very high impact. The implementation of OLAND in source-separated systems with black water (from vacuum toilets) and grey water, makes the crucial difference between energy negative and energy-positive treatment (Table 4.3). Decentralized schemes based on source separation with nitrification/denitrification consume 43 and 24 Wh IE-1 d-1 less than the classical centralized system, when digestion of only black water or also additional sludge of the A/B treatment of the grey water line and of the nitrogen treatment in the black water line is 71

OLAND maximizes net energy gain in systems with anaerobic digestion

considered, respectively. However, both schemes with nitrification/denitrification are energy negative. The implementation of OLAND makes both schemes energy positive (Table 4.3), due to a lower electrical energy consumption of 86 and 125 Wh IE-1 d-1 for both schemes respectively, compared to CAS systems. Therefore, an electrical energy saving with a factor 2-5, depending on the digestion options, can be established by implementation of OLAND in source separated treatment schemes. Table 4.3: Electricity consumption (Wh IE-1 d-1) and energy (E) index (-) comparing different options for decentral sewage treatment schemes with the central, conventional activated sludge (CAS). For ‘decentral 1’, only black water is digested, whereas also sludge from A and B stages (grey water line) and nitrogen removal (black water line, through OLAND or nitrification/denitrification; N/DN) is digested for ‘Decentral 2’. Calculation assumptions are mentioned in Box 1. Organic carbon for denitrification was provided by raw black water. CAS Decentral 1: Decentral 2: AD black water OLAND

N/DN

AD black water+all sludge OLAND

N/DN

COD -37.5 -15.9 -15.9 -15.9 -15.9 Biogas 42.3 78.7 56.5 112.1 34.6 removal* N removal -32.0 -13.1 -32.8 -11.2 -34.1 Pumping -20.0 -20.0 -20.0 -20.0 -20.0 Dewatering -6.4 -9.1 -10.2 -5.3 -6.1 Water* +12 +12 +12 +12 Total gain -53.4 32.3 -10.7 71.5 -29.6 E - index 0.4 1.6 0.7 2.4 0.6 * Incorporating savings aerobic COD oxidation in case of mainstream CAS treatment * The drinking water production of 25 L IE-1 d-1 is avoided, requiring 0.47 Wh L-1 (Verwin, 2006)

1.5

Treatment of digestates by OLAND: conclusions and perspectives

As OLAND decreases the energy need for nitrogen removal up to a factor 2, OLAND has the potential to decrease the overall electrical energy consumption significantly. However, during OLAND treatment, 11% of the converted nitrogen ends up in the effluent as nitrate, which especially for high-strength wastewaters such as digestates needs further polishing before discharge is permitted. The significance of OLAND implementation on the net electrical energy gain of a treatment system depends firstly on the proportion of the nitrogen load send to the OLAND reactor and secondly on the composition of the digestate itself (Table 4.4). Due to the suboptimal composition of liquid manure digestates, OLAND implementation has no strong effect on the electrical energy balance compared to conventional treatment. However, OLAND can offer a cost-effective treatment method because the cost of an external organic carbon source is avoided. In OFMSW plants and sugar/starch-based agro-industrial 72

Chapter 4

treatment plants, OLAND application could significantly enhance electrical energy recovery from waste by minimizing the energy consumption for digestate treatment mainly because around 95% of the nitrogen load is treated by OLAND and optimal BOD/N ratios are obtained in the digestate. Moreover, also in the latter applications, the cost for external organic carbon source addition is avoided in contrast to conventional treatment. In municipal treatment plants the effect of OLAND implementation in the side line, without other adjustments, is negligible because only 15-20% of the nitrogen load is treated through OLAND and the decrease in electrical energy demand for nitrogen removal can not counteract the increased demand for aerobic COD removal (Table 4.4). Despite the low nitrogen load treated by OLAND in municipal WWTP, energy autarky is possible through the implementation of enhanced primary sludge production and thus increased electrical energy recovery (Table 4.4, Wett et al., 2007; Siegrist et al., 2008). At decentralized level, due to source separation, OLAND can make the difference between energy-positive and energynegative operation.

Table 4.4: Comparison of energy index (energy production/energy consumption) of the treatment of anaerobic digestion digestates with the conventional treatment through nitrification/denitrification (N/DN) and the alternative treatment through OLAND. Application domain Energy index Energy index N/DN

OLAND

OFMSW plant* 3 6 Manure-based agricultural plant* 1.6 1.7 Starch-based agro-industrial plant* 5 10 Sewage with CAS 0.4 0.4/0.7° Sewage with A60/B 0.9 0.9/1.1°° Decentral 1 (AD black water) 0.9 1.6 Decentral 2 (AD black water + additional sludge) 0.6 2.4 *Pumping energy is considered negligible; °With enhanced primary settling; °° With an improved A-step (75%)

OLAND as mainstream treatment process

OLAND implementation combined with improved primary sludge production in municipal WWTP allows energy autarky as discussed in the previous section. However, if a higher proportion of the nitrogen present in sewage can be send to the OLAND system, for example by implementation of OLAND in the main line of the municipal WWTP, a higher decrease of the electrical energy consumption should be possible. Moreover, more organics could be recovered as electrical energy since no additional organic carbon is needed to meet nitrogen 73

OLAND maximizes net energy gain in systems with anaerobic digestion

discharge limits. The consequences of applying OLAND as a main treatment step in municipal WWTP are discussed in this section (Fig. 4.4).

Figure 4.4: Schematic overview of implementation of OLAND in mainline of WWTP (A/OL).

2.1

Wastewater as an energy resource

The potential energy in the form of organics available in the raw sewage exceeds the electricity requirements of the treatment process. Based on calorimetric measurements a specific energy input of 14.7 kJ per g COD can be calculated (Shizas and Bagley, 2004). In the conventional activated sludge (CAS) system, 38% of the incoming COD is aerobically converted to CO2 or anaerobically used for denitrification (Table 4.2, BOX 1). So, this means that 754 kJ IE-1 d-1 is spilled through metabolic reactions and not recovered as electrical energy resource. As a consequence the CAS systems can only produce the electrical equivalent of 42 Wh IE-1 d-1 which can not cover the electrical total energy costs 96 Wh IE-1 d-1 (Table 4.2). To fully recover the potential energy in the raw sewage, not only electrical energy usage minimization should be accomplished but more important, the electrical energy recovery by anaerobic digestion should be maximized. Therefore, the CAS system should be replaced by a first concentration step, bringing as much as COD as possible to the solid fraction, and a second biological step removing the residual nitrogen and COD with a minimal energy demand (Fig. 4.4).

A highly loaded activated sludge (A-step) compartment in the mainline can act as a concentration step. This A-step works at low hydraulic (15-30 minutes) and sludge retention 74

Chapter 4

(0.5 d) times, allowing the bacteria to work at their maximum growth yield. Therefore, organics are only incorporated in the biomass resulting in COD removal efficiencies to the sludge phase of 60-70% (SalomĂŠ, 1990). Higher efficiencies up to 80% can be accomplished by an increased loading rate, addition of flocculants, iron etc (Xia et al., 2005), allowing even higher biogas production rates. In Austria, an A-step is operated at COD removal efficiencies of 60% allowing conventional nitrification/denitrification in the main stream provided that the nitrogen in the side stream is separately treated in an OLAND step (Wett et al., 2007). This concept as such allows almost electrical energy autarky and can lead with co-digestion of kitchen waste to electrical energy neutral operation (Wett et al., 2007). An overall net electrical energy gain can be accomplished by applying a more efficient A-step (75% COD removal instead of 60%) (Table 4.5) allowing higher energy recovery through anaerobic digestion. Taking into account an average domestic wastewater composition of 30-100 mg N L-1 and 450 â&#x20AC;&#x201C; 1200 mg COD L-1 (Tchobanoglous et al., 2003; Henze et al., 2008), rendering COD/N ratios between 12 and 15, COD separation efficiencies in the A-step of 75-80% will result in COD/N ratios which are too low to allow full nitrification/denitrification. Therefore to obtain the same overall removal efficiencies in the WWTP, an external organic carbon source should be added to allow conventional nitrification/denitrification. This additional cost for an organic carbon source counteracts the advantages of the higher net electrical energy gain and therefore, in this case, the OLAND process could offer a cost-effective and energy friendly alternative. Theoretically, the implementation of OLAND in the main line at an Astep COD efficiency of 75%, does not further increase the net electrical energy gain (Table 4.5). Applying OLAND in the main line implies that the remaining COD should be removed aerobically and cannot be totally removed by denitrification, increasing the electrical energy need for COD removal with a factor 4. Therefore the decrease of the electrical energy requirement for nitrogen removal with 35%, can not counteract the increase in electrical energy demand for COD removal (Table 4.5). The main advantage of the implementation of OLAND in the main line compared to conventional nitrification/denitrification is the cost savings for external organic carbon addition. Since the improved CAS system with enhanced primary settling and OLAND in the side line requires an electrical input of 27 Wh IE -1 d-1 and this new concept can potentially gain an electrical equivalent of 10 Wh IE-1 d-1 (Table 4.5), the OLAND application in the main stream should be a further step towards a more energy friendly wastewater treatment. In the following sections the challenges (as indicated in Chapter 1) to accomplish this concept are studied in detail (Chapter 6-8).

OLAND maximizes net energy gain in systems with anaerobic digestion Table 4.5: Electrical energy demand and gain of municipal WWTP schemes. CAS: conventional activated sludge treatment; OL: OLAND reactor in side line, treating sludge reject water; 째 enhanced primary settling applied; AX/B: A/B system with a COD removal efficiency of X in the A-stage; OL: OLAND in the main stream; *corrected for COD removal through denitrification Oxygen and energy demand Electrical energy gain (Wh IE-1 d-1) COD removal* N removal main line OLAND in side line Energy consumption A step Pumping/mixing Sludge dewatering Total energy consumption Biogas-based energy production Net energy gain

2.2

CASOL째 -25.9 -28.6 -2.4 / -20.0 -6.2 -83.1 +56.4 -26.7

A60/B -9.7 -33.2 / -8.9 -20.0 -6.4 -78.3 +70.6 -7.7

A75/BOL -3.6 -26.2 -3.7 -11.1 -20.0 -6.5 -71.1 +81.2 +10.1

A75/OL -14.4 -19.5 / -11.1 -20.0 -6.5 -71.5 +81.2 +9.7

Main stream OLAND application: conclusions

The implementation of OLAND in the main line of the WWTP shows high potential to further increase the energy gain from sewage. A net electrical energy gain of 10 Wh IE-1 d-1 is expected to be possible. However, the challenges of microbial biomass retention at low HRT (< 1d) and NOB suppression at low temperature (10-20째C) should be first resolved before successful operation will be possible. Also the balance between energy gain and CO2 footprint of the WWTP should be considered when selecting the most sustainable solution. By this time, interest in this concept is rising and resulted in already a full-scale trial in Strass (Austria, in WERF project) and pilot set-up in Rotterdam (Paques).

General conclusions

The need to minimize the use of fossil fuel energy in the treatment of sewage, manures, agroindustrial wastes and municipal solid waste organics will continue to increase in the future. Anaerobic digestion allows to recover the chemical energy present as organic carbon and to convert it to electrical energy. The latter can make the overall treatment plant self-sufficient in electrical energy (energy index ranging from 1 to 5; Table 4.4) in the case of treatment of manures, agro-industrial wastes and municipal solid waste organics. Yet, problems with excess nitrogen and deficit of organic carbon in digestates to remove nitrogen by conventional nitrification/denitrification warrant the development of a new process design.

Chapter 4

OLAND, integrated with anaerobic digestion, avoids external organic carbon addition and allows to increase the energy index to 6-10 for the treatment of agro-industrial wastes and municipal solid waste organics (Table 4.4). OLAND application for sewage treatment only significantly lowers the electrical energy demand if the amount of organic carbon normally needed for denitrification is captured in the primary sludge by applying an enhanced primary settling or an improved A-stage. OLAND application in the main stream of sewage plants will allow an energy index above 1 (Table 4.5). OLAND, as a downstream process of anaerobic digestion is gradually becoming a mature technology which in different configuration (SBR, MBBR, RBC, airlift, etc) can be reliable operated at full-scale. Several full-scale cases are discussed here.

Acknowledgements

H.D.C. was recipient of a PhD grant from the Institute for the Promotion of Innovation by Science and Technology in Flanders (IWT-Vlaanderen, number SB-81068). E.C and S.E.V. were supported as a doctoral and postdoctoral fellow from the Research Foundation Flanders (FWO-Vlaanderen), respectively. The authors gratefully thank Bernhard Wett and Tim HĂźlsen for providing full-scale plant data and thank Tim Lacoere for technical support. This work was supported by Ghent University Multidisciplinary Research Partnership (MRP) â&#x20AC;&#x201C; Biotechnology for a sustainable economy (01 MRA 510W).

OLAND maximizes net energy gain in systems with anaerobic digestion

Chapter 5

Chapter 5: Efficient total nitrogen removal in an ammonia gas biofilter through high-rate OLAND Abstract Ammonia gas is conventionally treated in nitrifying biofilters, however addition of organic carbon to perform post-denitrification is required to obtain total nitrogen removal. Oxygenlimited autotrophic nitrification/denitrification (OLAND), applied in full-scale for wastewater treatment, can offer a cost-effective alternative for gas treatment. In this study, the OLAND application was broadened towards ammonia loaded gaseous streams. A down flow, oxygensaturated biofilter (height of 1.5 m; diameter of 0.11 m) was fed with an ammonia gas stream (248 ± 10 ppmv) at a loading rate of 0.86 ± 0.04 kg N m-3 biofilter d-1 and an empty bed residence time of 14 s. After 45 days of operation a stable nitrogen removal rate of 0.67 ± 0.06 kg N m-3 biofilter d-1, an ammonia removal efficiency of 99%, a removal of 75-80% of the total nitrogen and negligible NO/N2O productions were obtained at water flow rates of 1.3 ± 0.4 m3 m-2 biofilter section d-1. Profile measurements revealed that 91% of the total nitrogen activity was taking place in the top 36% of the filter. This study demonstrated for the first time highly effective and sustainable autotrophic ammonia removal in a gas biofilter and therefore shows the appealing potential of the OLAND process to treat ammonia containing gaseous streams.

Chapter redrafted after: De Clippeleir H., Courtens E., Mosquera M., Vlaeminck S.E., Smets B.F., Boon N. and Verstraete W. 2012. Efficient total nitrogen removal in an ammonia gas biofilter through high-rate OLAND. Environmental Science and Technology, Environmental Science and Technology, 46(16), 8826-8833.

Efficient total nitrogen removal in an ammonia gas biofilter through high-rate OLAND

Introduction

Ammonia (NH3) is a colorless and reactive air pollutant that is an important cause of acidification of soils and waters, and high levels of nitrate in surface and drinking waters. It is commonly emitted from both industrial and agricultural activities such as wastewater treatment plants, chemical and manufacturing industries, composting plants, and livestock farming (Chung et al., 1996; Busca and Pistarino, 2003; Kim et al., 2007). In contrast to the operational complexity and high costs of physico-chemical treatment processes, biological treatment can offer cost-effective and straightforward purification of gas streams. The latter biofiltration systems are mainly based on nitrification, transforming ammonia into nitrite and nitrate, and as a result end up with a highly loaded percolate mixture of ammonium (NH4+), nitrite (NO2-) and nitrate (NO3-) (Baquerizo et al., 2009). To obtain dischargeable effluent, post-denitrification with the addition of an external organic carbon source is applied or the effluent is send to a central wastewater treatment facility (Sakuma et al., 2008; Cabrol, 2010).

Anoxic autotrophic nitrogen removal by anoxic ammonium-oxidizing or anammox bacteria (AnAOB), able to combine nitrite with ammonium to N2 gas, can offer a solution in this nitrogen rich biofilter environment devoid in organic carbon. Oxygen-limited autotrophic nitrification/denitrification

(OLAND)

one-stage

realization

partial

nitritation/anammox, the economically preferred nitrogen removal technology for wastewaters with a biodegradable COD/N ratio below 3 (Kuai and Verstraete, 1998). This process is based on the cooperation between aerobic ammonium-oxidizing bacteria (AerAOB), which oxidize part of the ammonium to nitrite in the outer aerobic zones of the biofilm, and AnAOB, which subsequently convert nitrite and ammonium to nitrogen gas in the inner, anoxic zones. As a result, nitrogen is converted autotrophically in one step to nitrogen gas. This autotrophic nitrogen removal process has been established in full-scale for several wastewater treatment applications (Wett, 2006; Joss et al., 2009; Abma et al., 2010). However, this process was thus far not applied for the treatment of gaseous ammonia-rich streams.

Application of an OLAND biofilter would allow a total nitrogen removal, defined as a total nitrogen loss based on gas and water composition, in the biofilter itself due to N2 gas production by AnAOB. Although most ammonia gas biofilters are based on nitrification, a total nitrogen removal efficiency is commonly observed ranging from 10 to 50% (Table 5.1). 80

Chapter 5

Total nitrogen removal can occur in the inert form of N2 or in the unsustainable form of NO or N2O. However, the contribution of NO and N2O production to the total nitrogen removal and the operational factors inducing higher total nitrogen removal rates are unclear. Until now the total nitrogen removal was attributed to denitrification by heterotrophic and/or nitrifying bacteria, both needing oxygen limitation, or by bacterial growth. The contribution of AnAOB as a cause for total nitrogen removal in biofilters was not considered before despite the presence of ammonium and nitrite and the occurrence of anoxic activity in these filters. Heterotrophic denitrification is possible when organic compounds in the gas or water phase are available and can lead to both N2 and NO/N2O formation (Juhler et al., 2009). Nitrifier denitrification by aerobic ammonium-oxidizing bacteria (AerAOB) implicates nitrogen removal by NO and N2O formation instead of N2 production (Chandran et al., 2011), and hence negatively affects the sustainability of the technology. It was reported that almost 20% of the NH3 loading can be converted to N2O by autotrophic and/or heterotrophic denitrification (Maia et al., 2012). Finally, nitrogen can also be incorporated into the biomass and used for growth. It was estimated that the nitrogen incorporation in biomass accounts for 7% of the nitrogen input (Cabrol, 2010). The stimulation of AnAOB in the biofilter and thus application of the OLAND process for the treatment of ammonia containing gas streams could offer two advantages. Firstly, AerAOB inhibition by free ammonia or free nitrous acid is commonly observed in biofilters and results in ammonium to nitrate ratios in the percolate of around 1 (Smet et al., 2000; Chen et al., 2005; Baquerizo et al., 2009; Cabrol, 2010). The lower ammonium consumption rate by AerAOB can be compensated during the OLAND process by ammonium consumption by AnAOB. Secondly, the higher the AnAOB activity in the filter, the higher the nitrogen gas production rate, and thus the higher the total nitrogen removal rate in the filter will be. Together, these two facts will decrease the need for posttreatment of the percolate and consequently the cost for external organic carbon source addition. The goal of this study was to demonstrate the possibility to obtain fully autotrophic total nitrogen removal in an ammonia gas biofilter through a combination of AerAOB and AnAOB activity, also referred to as the OLAND process. This is the first study showing anammox as the main nitrogen removal process in ammonia gas biofilters.

Efficient total nitrogen removal in an ammonia gas biofilter through high-rate OLAND

Table 5.1: Overview of the operational parameters and nitrogen losses in ammonia gas biofilters. DF/UF: down flow/upflow reactors, H/D: height over diameter ratio; EBRT: empty bed residence time N loss DF/UF Packing Height H/D EBRT NH3 in Loading rate (++) Water Temperature Reference material flow rate(+) kg N m-3 kg N m-2 biofilter d biofilter 1 section 0 UF Slow release 1.0 10 20-36 90-260 0.1-0.6 0.1-0.6 d-1 2700-30 DF Slow release 0.3 3 14 1.3-3.0 0.4-0.9 700 15 DF Slow release 0.6 4 32-85 10-150 0.1-0.2 0.07-0.1 16-32 UF Slow release 1.0 7 30-35 50-200 0.1-0.6 0.1-0.6 52 DF Slow release 1.5 9 54 35 0.03 0.05 30-60 DF Slow release 1.4 14 50 35-170 0.1-0.3 0.1-0.5 10098* DF Inert 0.6 11 60 0.1-0.5 0.05-0.3 600± 250 75-80 DF Inert 1.6 14 14 0.9 ± 0.1 1.3 ± 0.1 10 nitrification/denitrification *External organic carbon source addition in filter to obtain simultaneous °The air flow was humidified before entering the biofilter The water to N ratio expressed as L water g-1 Nin can be calculated by dividing (+) by (++) %

ppm

m3 m-2 biofilter d1 0.1° 3.9 1.6 0° 0.4 0.08 72 1.2 ± 0.4

°C 24 22-25 30 25-30 20-25 20-30 20-25 20-25

(Baquerizo et al. 2009) (2009) (Sakuma et al. 2008) (Kim et al. 2007) (Chen et al. 2005) (Cabrol 2010) (Malthautier et al. 2003) (2003) (Moussavi et al. 2011) This study

Chapter 5

2 2.1

Materials and methods Biofilter set-up and operation

The biofilter consisted of a PVC cylindrical column with a height of 1.57 m and an internal diameter of 0.11 m. The section surface of the filter was thus 95 cm2. The column was packed with Kaldnes K1 packing material (AnoxKaldnes, Lund, Sweden) and on 50% of the carriers OLAND biomass from a stably working OLAND rotating contactor was added (Pynaert et al., 2003), resulting in an initial biomass concentration of 3.8 g VSS L-1 biofilter. The total contact surface based on the specific surface of the Kaldnes rings was estimated at 800 m2 m-3 total reactor. The inlet ammonia stream was supplied at the top as a mixture of compressed air and pure ammonia, and was controlled by two digital mass flow controllers (Bronkhorst, The Netherlands) to ensure a stable inlet concentration of 248 ± 10 ppmv, a gas velocity of 0.1 m s-1 and a gas empty bed residence time (EBRT) of 14 seconds. The biofilter was humidified by discontinuously spraying (1 second every 5 minutes) tap water at an initial flow rate of 0.8 m3 m-2 biofilter section d-1 on top of the filter. The filter was operated at room temperature (23±1°C). Daily, samples were taken from the gas in- and outlet (200 mL) and from the water phase (10 mL) to determine the ammonia, ammonium, nitrite and nitrate concentration. Nitrous oxide and nitric oxide concentration were only measured during the profile measurements.

2.2

Profile measurements

On days 90 and 99, gas and water samples were taken at 0, 7, 32, 57, 82, 107, 132 and 157 cm depth from the top for the detection of NH3, O2, NO, N2O and NH4+, NO2- and NO3-. In all water samples, the pH was also measured. These measurements allowed obtaining vertical activity profiles.

2.3

Activity batch test

On day 125, the specific activities of AerAOB, AnAOB and nitrite oxidizing bacteria (NOB) in the different zones of the biofilters (see profile measurements) was determined in separate activity tests in aqueous media at 22°C and at initial nitrogen concentration of 100 mg N L-1, as described by Vlaeminck et al. (2007).

Efficient total nitrogen removal in an ammonia gas biofilter through high-rate OLAND

2.4

Chemical analyses

NH3 was measured in the gas phase with colorimetric gas detection tubes (RAE, Hoogstraten, Belgium), using 100 mL of gas sample. The NH3 detection tubes had a detection limit of 1 ppmv NH3 (0.62 mg NH3-N L-1). The N2O and O2 concentrations in the gas phase were analyzed with a Compact GC (Global Analyser Solutions, Breda, The Netherlands), equipped with a Porabond precolumn and a Molsieve SA column. The thermal conductivity detector had a detection limit of 1 ppmv for each gas component. NO measurements were done based on the principle of chemiluminescence using Eco Physics CLD 77 AM (Eco Physics AG, Duernten, Switzerland) with a detection limit of 1 ppbv. Ammonium (Nessler method) and VSS (after removing the biomass from the carriers) were measured according to standard methods (Greenberg et al. 1992). Nitrite and nitrate were determined on a Metrohm 761 Compact Ion Chromatograph (Zofingen, Switzerland) equipped with a conductivity detector. Dissolved oxygen (DO) and p

were measured with, respectively, an

0d DO meter

( ach Lange, D sseldorf, Germany) and an electrode installed on a C833 meter (Consort, Turnhout, Belgium).

Table 5.2: Overview of the primers sets and conditions used for determination of the abundance of AerAOB, AOA, AnAOB and NOB with qPCR. Functional Target Primers Sequences (5´-3´) Melting Ref. group gene temp (ºC) AerAOB

GGGGTTTCTACTGGT 55 (1) GGT amoA-2R CCCCTCKGSAAAGCC TTCTTC AOA Creanarchaeal CrenamoA23f ATGGTCTGGCTWAG 56 (2) amoA gene ACG Creanamo A616r GCCATCCATCTGTAT GTCCA Nitrospira sp. 16S rRNA Nspra675f GCG GTG AAA TGC 67.2 (3) GTA GAK ATC G Nitrospira sp. 16S rRNA Nspra746r TCA GCG TCA GRW 65.3 (3) AYG TTC CAG AG AnAOB 16S rRNA Amx809f GCC GTA AAC GAT 67.1 (4) GGG CAC T AnAOB 16S rRNA Amx1066r ATG GGC ACT MRG 67.4 (4) TAG AGG GGT TT (1): Rotthauwe et al. (1997); (2) Tourna et al. (2008); (3) Graham et al. (2007); (4) Tsushima et al. (2007a)

amoA gene

amoA- 1F

Chapter 5

2.5

Quantification with real-time PCR

Biomass samples (approx. 5 g) for nucleic acid analysis were taken from the OLAND rotating contactor (inoculum of the biofilter) and at 7, 32, 57, 82, 107, 132 and 157 cm depth after 125 days of operation. DNA was extracted using FastDNA® SPIN Kit for Soil (MP Biomedicals, LLC), according to the manufacturer’s instructions. The obtained DNA was purified with the Wizard® DNA Clean-up System (Promega, USA) and its final concentration was measured spectrophometrically

using

NanoDrop

ND-1000

spectrophotometer

(Nanodrop

Technologies). The SYBR Green assay (Power SyBr Green, Applied Biosystems) was used to quantify the 16S rRNA of bacterial anammox bacteria and Nitrospira sp. and the functional amoA gene for AerAOB and ammonium-oxidizing archaea (AOA). The primers for quantitative polymerase chain reactions (qPCR) used in this study are listed in Table 5.2. Plasmid DNAs carrying AerAOB, AOA functional AmoA gene and Nitrospira and anammox 16SrRNA gene, respectively, were used as standards for qPCR.

3 3.1

Results Performance of the biofilter

In a biomass free control test with inert Kaldnes K1 packing material, all nitrogen inserted via the gas phase as NH3 could be found back in the effluent gas and water phase, excluding nitrogen removal by leakages. After inoculation of the biofilter with active OLAND biofilm on the Kaldnes K1 packing material, the biofilter was immediately fed with an ammonia gas stream, without acclimatization of the biomass by water recirculation, at a loading rate of 0.88 ± 0.04 kg N m-3 biofilter d-1. After 31 days of operation, the ammonia gas removal efficiency remained stable around 99 ± 0.7%, independent of the operational conditions (Fig. 5.1). Although a high pH value around 8.3 ± 0.6 was measured during the start-up period (phase I, Fig. 5.1), only 20 ± 5% of the nitrogen load was detected in the percolate as ammonium and the total nitrogen removal accounted already for 53 ± 11% of the total nitrogen load. During phase II (Fig. 5.1), an ammonium decrease and nitrite and nitrate increase in the percolate together with higher total nitrogen removal efficiencies up of 70 ± 5% were accompanied by a pH decrease from 8.3 ± 0.6 (Phase I) to 6.6  0.4 (Phase II). From day 45 onwards, the decrease in pH was stabilized by addition of coccolith lime on the biofilter top (on average 0.7 kg m-3 biofilter d-1) resulting in a pH value of 6.9 ± 0.3 during the following operation period (end of phase II and phase III). During phase III, the influence of the water flow rate

Efficient total nitrogen removal in an ammonia gas biofilter through high-rate OLAND

on the biofilter performance was tested as the latter influences the NH3 dissolution and the nitrogen concentration at which the bacteria are exposed to. A small increase in the water flow rate from 1.2 ± 0.6 to 1.7 ± 0.2 m3 m-2 biofilter section d-1 combined with the stable pH conditions allowed higher total nitrogen removal efficiencies of 79 ± 6% between day 73 and day 90. The latter was mainly due to higher ammonium removal efficiencies (Fig. 5.1). A decrease from day 91-105 of the water flow rate to 1.2 ± 0.4 m3 m-2 biofilter section d-1 did not have a significant effect on the removal performance. Moreover, during the increase of the water flow rate up to 2.4 ± 0.7 m3 m-2 biofilter section d-1 (day 106-125), the removal efficiency remained stable around 79 ± 7% and only the absolute concentration in the percolate decreased due to dilution. A NH3 gas inlet failure (day 114), resulting in 1 day without NH3 addition, had no significant influence on the performance afterwards.

Chapter 5

Figure 5.1: Nitrogen loading and removal rates (top) and corresponding contribution of the different nitrogen species in the emitted air and water flow, as a percentage of the incoming nitrogen (bottom). The nitrogen removal is considered to be due to N2 formation, since profile measurements showed negligible amounts of NO (0.5% of incoming N) and N2O (below detection limit) production. Three main periods are distinguished: a start-up period (phase I); a pH stabilization period (phase II) and a water flow rate optimization period (phase III).

Efficient total nitrogen removal in an ammonia gas biofilter through high-rate OLAND

Table 5.3: Operational conditions measured directly in the filter and microbial activities measured in separate aqueous activity tests (n=3) at different top down biofilter zones. n.d.: not detected; AerAOB: aerobic ammonium oxidizing bacteria; AnAOB: anoxic ammonium oxidizing bacteria; NOB: nitrite oxidizing bacteria

Water flow rate (m3 m-2 biofilter section d-1) pH (-) 1.4 1.1 Free ammonia (mg N L-1) 1.4 1.1 NO2- (mg N L-1) 1.4 1.1 Microbial group Total anoxic nitrogen removal rate AnAOB (mg N g-1 VSS d-1) Aerobic ammonium oxidation rate AerAOB (mg N g-1 VSS d-1) Aerobic nitrate production rate NOB (mg N g-1 VSS d-1)

Top down biofilter zone 57-82 cm 82-107 cm 107-132 cm

7-32 cm

32-57 cm

8.6 8.1 61 51 200 441

7.2 6.6 0.4 1.0 147 411

7.1 6.8 0.4 0.4 119 248

6.5 6.0 0.07 0.07 122 254

6.3 6.3 0.03 0.09 30 21

7.1 7.1 0.2 0.5 71 121

13 ± 3

9±3

13 ± 4

2 ± 2*

n.d.*

142 ± 52

252 ± 27

389 ± 10

226 ± 54

149 ± 59

244 ± 37

1±1

n.d.

19 ± 14

157 ± 63

140 ± 26

136 ± 46

*Anoxic nitrite consumption without anoxic ammonium consumption was observed, but this should not be considered as AnAOB activity.

132-157 cm

Chapter 5

3.2

Vertical distribution of microbial activity

The vertical profile measurement during phase III showed that the highest microbial activity occurred in the top 0.57 m of the biofilter (Fig. 5.2), while at all heights oxygen was saturated in the gas phase. Ammonia dissolved for 80-95%, depending on the water flow rate, in the first 32 cm of the biofilter (Fig. 5.2). In the first 7 cm, only dissolution and no microbial activity occurred. In the subsequent zones the highest total nitrogen removal rates were observed. In these zones, ammonium and nitrite were consumed without an equivalent nitrate production (Fig. 5.2). In this upper 57 cm of the filter, 91% of the total nitrogen removal was taking place (Fig. 5.2A) and according to the stoichiometry, the absence of organic carbon and the absence of NO or N2O production, this was mainly attributed to the AnAOB. In the lower two thirds (> 57 cm depth, Fig. 5.2A) some denitrification (9% of the total nitrogen removal), probably using organics from bacterial decay, occurred. Although the total nitrogen removal rate in the biofilter remained constant when the water flow rates was decreased from 1.4 to 1.1 m3 m-2 biofilter section d-1 (Fig. 5.2B), a downward shift of the OLAND activity from 0.07-0.57 m to 0.32-0.82 m was observed. This was probably attributed to the inhibition of the AnAOB activity by higher nitrite concentrations in the upper section of the filter and the slower dissolution of NH3 (Table 5.3). Biomass samples were taken at 6 biofilter zones and the AnAOB, AerAOB and NOB activity was tested in aqueous medium. Despite the 4-fold lower total nitrogen removal rate compared to the biofilter performance (0.2 compared to 0.8 kg N m-3 biofilter d-1), the vertical profile distribution of the AnAOB activity confirmed the direct biofilter profile measurements (Table 5.3). AnAOB activity was measured in the zone 7-82 cm and decreased rapidly in the lower compartments (Table 5.3). In the lower zones (> 82 cm), (nitrifier) denitrification could take place because anoxic nitrite consumption was taking place while no difference in ammonium concentration was observed (data not shown). AerAOB were active over the total depth of the biofilter, while NOB started to show activity at the lower part of the biofilter (> 82 cm). The total biomass concentration, measured after emptying the biofilter, increased during 125 days of operation from 3.8 g VSS L-1 biofilter to 19 g VSS L-1 biofilter, with the highest concentration at 7-32 cm depth.

Efficient total nitrogen removal in an ammonia gas biofilter through high-rate OLAND

Figure 5.2: Vertical profile measurement expressed as NH3, NH4+, NO2- and NO3- productions based on the gas and water phase analysis at day 90 (A) and day 99 (B) operated at water flow rates of 1.4 and 1.1 m3 m-2 biofilter section d-1, respectively. Total NO production was negligible (0.5% of nitrogen input) and N2O production was not detected.

3.3

Vertical abundance of N species

The biofilter was inoculated with biomass containing 2 102 AerAOB-AmoA copies, 9 103 AnAOB-16SrRNA copies and 2 102 Nitrospira-16SrRNA copies ng-1 DNA, which was homogeneously distributed over the filter. AOA were not detected in the inoculum. However after 125 days of operation, up to 2 102 AOA-AmoA copies ng-1 DNA were detected (Fig. 5.3). AnAOB abundance remained constant over the filter. AerAOB showed a peak concentration at a depth of 57-82 cm, which correlated well with the activity test (Table 5.3).

Chapter 5

The Nitrospira abundance increased significantly to 2 105 Nitrospira-16SrRNA copies ng-1 DNA below a depth of 82 cm. The observed decrease in inhibition factors such as free ammonia and the NOB activity measured at these zones (Table 5.3) confirmed the abundance measurements.

Figure 5.3: Abundance of AerAOB, AOA, AnAOB and Nitrospira, expressed as copies of AerAOBAmoA, AOA-amoA, AnAOB-16SrRNA and Nitrospira-16SrRNA ng-1 DNA, respectively, in the inoculum and in the different biofilter zones after 125 days of operation.

4 4.1

Discussion OLAND application for NH3 treatment

This study showed for the first time that AnAOB activity can be obtained in an oxygensaturated biofilter treating a NH3 gas stream. The application of the OLAND process instead of nitrification in the biofiltration technology would allow higher total nitrogen removal in the biofilter itself (up to 80%), significantly decreasing the cost for an external carbon source addition needed for post-denitrification of the percolate. Moreover, this study showed that by implementing the OLAND process, a sustainable total nitrogen removal can be obtained without NO and N2O formation. The unsustainable nitrogen removal during conventional NH3 treatment is in most studies neglected and not measured (Table 5.1), but is expected to be high (up to 20% of the nitrogen loading rate can be emitted as N2O; Maia et al., 2012).

Efficient total nitrogen removal in an ammonia gas biofilter through high-rate OLAND

4.2

AnAOB niche in NH3 biofilters

Total nitrogen removal in NH3 fed biofilters has been reported in several studies (Table 5.1). However, the total nitrogen removal efficiency was mainly lower than 60%, while in this study a total nitrogen removal efficiency of almost 80% was obtained (Table 5.1). The total nitrogen removal rates obtained in the NH3 fed OLAND biofilter were in the same range as OLAND application for wastewater treatment (Vlaeminck et al., 2012). Generally, the total nitrogen removal in ammonia gas biofilters can be attributed to several processes: (i) denitrification; (ii) nitrifier denitrification; (iii) nitrogen biomass incorporation, (iv) chemical reactions and as shown in this study (v) anammox. Because inert packing material was used in this study and no organic carbon source was present in the gas or water flow, the contribution of denitrification to the total nitrogen removal was considered to be minor. In contrast to several studies suggesting that AerAOB were responsible for the total nitrogen removal due to nitrifier denitrification (Chen et al., 2005; Kim et al., 2007), the latter pathway could be excluded in this study because no N2O and very low NO emissions (0.5% of N loading) were detected. Also chemical reactions leading to NO or N2O formation could be neglected in this study (Chandran et al., 2011; Vermeiren et al., 2012). Nitrogen incorporation in the biomass could probably explain for a part the 9-15% nitrogen loss that was measured during the profile measurements but that was, based on the stoichiometry not caused by AnAOB activity. The total nitrogen removal to N2 in the top part of the filter (> 82 cm) was attributed to AnAOB activity as 85-91% of the nitrogen removal took place at the biofilter zones where ammonium and nitrite consumption was observed (Fig. 5.2) and as the specific activity tests confirmed AnAOB activity in the top part of the filter (Table 5.3). Moreover, if denitrification had been responsible for the total nitrogen removal, at least 2 kg COD m-3 biofilter d-1 should have been consumed, corresponding with 1.5 kg VSS m-3 biofilter d-1, or around 40% of the inoculated biomass organics, allowing no biomass growth in the filter.

Data on the microbial community structure in NH3 biofilters are still relatively scarce, compared to other engineered systems such as bioreactors for wastewater treatment. Studies performed on NH3 fed biofilters discuss mainly overall diversity and dynamics (Cabrol et al., 2010) or focus only on the AerAOB (Juhler et al., 2009; Yin and Xu, 2009; Yasuda et al., 2010) or ammonium-oxidizing Archaea (Yasuda et al., 2010). The anaerobic ammonium oxidation was not considered before in this application domain. Moreover, due to a lack of information about the relation between the community structure and the total nitrogen balance 92

Chapter 5

in these biofilters (Table 5.1), there was no evidence of the presence of Planctomycetes and more specifically AnAOB in NH3 fed biofilters. However, this study showed that AerAOB in close relationship with AnAOB can cause high nitrogen removal rates in gaseous biofilters. Both substrates ammonia and nitrite are commonly present in biofilters due to the high AerAOB activity and lower NOB activity (Baquerizo et al., 2009), which indicate a niche environment for AnAOB, provided that anoxic conditions are created.

As the biofilter was fed under fully aerobic conditions, anoxic zones should have been present to allow AnAOB activity. It could be calculated that anoxic zones could be obtained in the biofilm itself when the thickness of an oxygen-consuming biofilm was higher than 84 Îźm (Perez et al., 2005) given oxygen saturation in the gas phase over the whole depth of the biofilter. On the other hand, preferential gas and water flow due to a low ratio between the biofilter reactor diameter and packing material diameter (11 < 12), could probably occur allowing oxygen gradients in the filter (Beavers et al., 1973).

Due to the high free ammonia concentration and consequently NOB inhibition (Anthonisen et al., 1976) at the top of the biofilters (Table 5.3), total nitrogen removal by AnAOB was mainly taking place between 7-57 cm depth despite the high nitrite levels (around 200 mg NO2--N L-1). AnAOB can irreversibly be inhibited by nitrite. However the reported inhibition range (100-350 mg N L-1) is broad and the effect depends on the AnAOB species (Strous et al., 1999; Egli et al., 2001; Dapena-Mora et al., 2007). In this study, inhibition of AnAOB was only observed at nitrite levels above 411 mg N L-1 (Table 5.3), and this effect seemed reversible in case the water flow rate was adjusted. Therefore, the water flow rate relative to the nitrogen gas loading of the system, calculated as the ratio between the water flow rate (m3 m-2 biofilter section d-1) and the nitrogen loading rate (kg N m-2 biofilter section d-1), determined the degree of AnAOB activity over NOB activity in the system as well as the AnAOB over NOB abundance. Water to N ratios lower than 1 L g-1 Nin resulted in both AnAOB and NOB inhibition in the top layers (Fig. 5.2B), while AnAOB were favored above NOB at higher water to N ratios (around 1 L g-1 Nin). In most studies reporting minor total nitrogen removal efficiencies (Table 5.1), this ratio was high (>>1 L g-1 Nin) decreasing free ammonia concentrations in the filter (higher dilution), and consequently allowing NOB activity. As a result, AnAOB probably did not have a competitive advantage compared to NOB in the top layers and could not significantly invade the filter in contrast to the OLAND biofilter in this study. So, to obtain high nitrogen losses without significant N2O emissions 93

Efficient total nitrogen removal in an ammonia gas biofilter through high-rate OLAND

and thus a niche for AnAOB, high nitrogen levels together with low water to N ratios (around 1 L g-1 Nin) are advised.

4.3

OLAND: gas versus water treatment

OLAND is considered an established technology for the treatment of digestates in several application domains (Chapter 4) as it can provide high and stable performance and decreased operational cost by decreasing the energy consumption and avoiding the addition of external organic carbon (Vlaeminck et al., 2012). Compared to these water applications, the OLAND biofilter for gas treatment could offer an additional advantage. The optimal balance between AerAOB, AnAOB and NOB activity is more easily obtained without complicated control systems as needed during wastewater treatment. In the latter application, NOB activity is avoided by a combination of DO control, free ammonia levels and specific SRT control of aerobic flocs (Joss et al., 2011), which significantly increases the operational complexity. Moreover, the control of the microbial balance becomes more difficult when treating low nitrogen concentration (Chapter 7). In the OLAND biofilter, the nitrogen gas flow, although containing low NH3 concentrations, is concentrated in the water film on top of the biofilm, allowing more easily NOB inhibition by free ammonia or even free nitrous acid. Therefore, besides saving costs for further percolate treatment, the OLAND biofilter can be stably operated at minimal operational complexity.

Conclusions

This study demonstrated for the first time highly effectient (up to 80%) and sustainable (negligible NO/N2O emission) autotrophic ammonia removal, based on AnAOB activity in a gas biofilter. Therefore, this study shows the appealing potential of the OLAND process to treat ammonia containing gaseous streams.

Acknowledgements

H.D.C. was a supported by a PhD grant from the Institute for the Promotion of Innovation by Science and Technology in Flanders (IWT-Vlaanderen, number SB-81068). E.C and S.E.V. were supported as doctoral candidate (Aspirant) and a postdoctoral fellow, respectively, from the Research Foundation Flanders (FWO-Vlaanderen). The authors thank Samuel BodĂŠ for kind assistance with NO analyses and Joachim Desloover, Tom Hennebel and FrederiekMaarten Kerckhof for inspiring scientific discussions. 94

Chapter 6

Chapter 6: OLAND is feasible to treat sewagelike nitrogen concentrations at low hydraulic residence times Abstract Energy-positive sewage treatment can in principle be obtained by maximizing energy recovery from concentrated organics, and by minimizing energy consumption for concentration and residual nitrogen removal in the main stream. To test the feasibility of the latter, sewage-like nitrogen influent concentrations were treated with oxygen-limited autotrophic nitrification/denitrification (OLAND) in a lab-scale rotating biological contactor (RBC) at 25°C. At influent ammonium concentrations of 66 and 29 mg N L−1 and a volumetric loading rate of 840 mg N L−1 d−1 yielding hydraulic residence times (HRT) of 2 and 1 h, respectively, relatively high nitrogen removal rates of 444 and 383 mg N L−1 d−1 were obtained, respectively. At low nitrogen levels, adapted nitritation and anammox communities were established. The decrease in nitrogen removal was due to decreased anammox and increased nitratation, with Nitrospira representing 6% of the biofilm. The latter likely occurred given the absence of dissolved oxygen (DO) control, since decreasing the DO concentration from 1.4 to 1.2 mg O2 L−1 decreased nitratation by 35% and increased anammox by 32%. Provided a sufficient suppression of nitratation, this study showed the feasibility of OLAND to treat low nitrogen levels at low HRT, a prerequisite to energypositive sewage treatment.

Chapter redrafted after: De Clippeleir, H., Yan, X., Verstraete, W., Vlaeminck, S.E., 2011. OLAND is feasible to treat sewage-like nitrogen concentrations at low hydraulic residence times. Applied Microbiology and Biotechnology, 90, 1537-1545.

OLAND is feasible to treat sewage-like nitrogen concentrations at low HRT

Introduction

Biological nitrogen removal is economically preferred above physicochemical nitrogen recovery for wastewaters containing less than 5 g N L−1 (Mulder, 2003). Furthermore, if the ratio of biodegradable chemical oxygen demand (bCOD) to nitrogen is relatively low (typically ≤ 2-3), nitrogen removal with partial nitritation and anammox saves about 60% of the aeration, 90% of the sludge handling and transport, and 100% of the organic carbon addition compared to conventional nitrification/denitrification (Mulder, 2003). Overall some 30-40% of the overall nitrogen removal costs can be saved (Fux and Siegrist, 2004). Oxygenlimited autotrophic nitrification/denitrification (OLAND) is a one-stage configuration of this process (Kuai and Verstraete, 1998), in which aerobic ammonium-oxidizing bacteria (AerAOB) oxidize about half of the ammonium to nitrite in the outer, aerobic zones of the biomass (partial nitritation), while the anoxic ammonium-oxidizing bacteria (AnAOB) subsequently convert nitrite and the residual ammonium to mainly nitrogen gas (89%) and some nitrate (11%) in the inner, anoxic zones (anammox; Pynaert et al., 2003; Vlaeminck et al., 2010). Oxygen plays a key role in balancing the microbial activities (Fig. 6.1A), with on the one hand an oxygen requirement of 1.8 g O2 g−1 N to achieve sufficient ammonium oxidation while avoiding excess nitrite production by AerAOB. On the other hand, sufficiently low dissolved oxygen (DO) levels (e.g. 0.3 mg O2 L−1) are needed to suppress excess nitrate production by nitrite-oxidizing bacteria (NOB) (Joss et al., 2009).

Conventional activated sludge (CAS) systems for sewage treatment have low volumetric carbon and nitrogen loading rates (around 1 g COD L−1 d−1 and 0.08 g N L−1 d−1) and are energy-negative. The aeration required for organic carbon and nitrogen removal constitutes about 60-70% of the total energy consumption of a sewage treatment plant (Zessner et al., 2010). However, if enhanced primary settling is applied to increase physico-chemical sludge production and if OLAND is used for nitrogen removal from the digestate of primary and secondary sludge (Fig. 6.1B), the aeration requirements of the CAS step can be decreased with 25% (Siegrist et al., 2008). Over the last five years several OLAND-type treatments were developed to treat sewage sludge digestates (Joss et al., 2009; Jeanningros et al., 2010). Furthermore, if primary settling is replaced by a highly loaded activated sludge step, where organic matter is converted to biomass at maximal yield, energy-neutrality is achievable given the even higher conversion of bCOD by anaerobic digestion into biogas and hence electricity (Wett et al., 2007). Given the high energetic content of the sewage bCOD, energy-positive 96

Chapter 6

sewage treatment should be possible (Siegrist et al., 2008; Verstraete et al., 2009; Kartal et al., 2010a). This requires an advanced biological or physicochemical bCOD concentration step to further increase energy recovery from anaerobic digestion of concentrated organics and a low energy demand for the concentration step and the removal of residual nitrogen (and some bCOD) in the main stream (Fig. 6.1B). The energy requirement for OLAND is influenced by the reactor configuration: active aeration in sequencing batch reactors requires 1.3 kWh kg−1 N (Wett et al., 2010b), whereas passive aeration in rotating biological contactors (RBC) requires down to 0.4 kWh kg−1 N (Mathure and Patwardhan, 2005). Depending on the dilution, sewage is typically composed of 30-100 mg N L−1 and 450-1200 mg COD L−1 rendering a COD/N ratio of about 12 to 15 (Metcalf and Eddy, 2003; Tchobanoglous et al., 2003; Henze et al., 2008). An advanced concentration step is expected to separate up to 7580% of the COD (Verstraete et al., 2009) and about 20% of the sewage nitrogen, mainly consisting of colloidal and particulate organic nitrogen, from which the anaerobically hydrolyzed part is returned to the main stream as ammonium (Fig. 6.1B). Hence, the OLAND stage would receive nitrogen as ammonium at a COD/N ratio below 4, which is theoretically low enough to avoid the risk that heterotrophs overgrow AnAOB (Lackner et al., 2008).

Until now, the OLAND process has been applied for medium and high-strength nitrogen wastewaters (> 0.2 g N L−1) such as landfill leachate and digestates from sewage sludge, specific industrial streams and concentrated black water at relatively high hydraulic residence times (HRT, Table 6.1). To obtain reasonably high nitrogen removal rates (400 mg N L−1 d−1), the treatment of low nitrogen levels (< 80 mg N L−1) has to occur at low HRT, in the order of some hours, rendering biomass retention an important requirement. In this study, a first and exploratory step towards the implementation of OLAND in the new sewage treatment scheme was tested, feeding sewage-like ammonium influent concentrations without COD addition. Given its low energy consumption for aeration, a RBC was chosen as lab-scale reactor, and operated at 25°C, simulating the maximum sewage temperatures in summer (Breda, NL; Mollen, personal communication). This is one of the first tests on the OLAND treatment of low nitrogen concentrations at such low HRT, a prerequisite to energy-positive sewage treatment.

OLAND is feasible to treat sewage-like nitrogen concentrations at low HRT

Figure 6.1: A. Conversion of nitrogen species, oxygen and protons in oxygen-limited autotrophic nitrification/denitrification (OLAND), showing balanced and imbalanced contributions of three bacterial groups, i.e. aerobic ammonium-oxidizing, nitrite oxidizing and anoxic ammonium-oxidizing bacteria (AerAOB, NOB and AnAOB, respectively); B. Conventional and redesigned sewage treatment schemes with OLAND in the side and main line, respectively. In the redesigned scheme, energy-positive sewage treatment can in principle be obtained by maximizing energy recovery through anaerobic digestion of concentrated organics in the side stream, and by minimizing energy consumption for the physicochemical and/or biological concentration step and the residual nitrogen removal step, applying OLAND. 98

Chapter 6

Table 6.1 Overview of typical average nitrogen concentrations, volumetric loading/removal rates and hydraulic residence times (HRT) for existing one-step partial nitritation/anammox processes and for the low nitrogen concentration and HRT application in this study. RBC: rotating biological contactor; SBR: sequencing batch reactor Wastewater Influent N loading rate N removal rate HRT Reactor Reference concentration (g N L−1 d−1) (g N L−1 d−1) (d) type (mg NH4+-N L−1) Digested black water

1023

0.94

0.71

1.33

RBC

(Vlaeminck et al., 2009b)

Sewage sludge digestate

800

0.74

0.67

0.93

SBR

(Jeanningros et al., 2010)

Sewage sludge digestate

650

0.54

0.51

1.20

SBR

(Joss et al., 2009)

Industrial digestate

300

2.0

1.17

0.18

Gas-lift

(Abma et al., 2010)

Landfill leachate

209

0.38

0.55

RBC

(Hippen et al., 1997)

Landfill leachate

250

0.67

0.41

0.51

RBC

(Siegrist et al., 1998)

Sewage-like nitrogen concentrations

0.86

0.44

0.08

RBC

This study

Sewage-like nitrogen concentrations

0.84

0.38

0.04

RBC

This study

OLAND is feasible to treat sewage-like nitrogen concentrations at low HRT

2 2.1

Material and methods OLAND rotating biological contactor (RBC)

The lab-scale RBC was based on an airwasher LW14 (Venta, Weingarten, Germany) with a rotor consisting of 40 discs interspaced at 3 mm, resulting in a disk contact surface of 1.32 m2. The reactor had a liquid volume of 3.6 L, immersing the discs for 64%. The reactor temperature was set at 25°C and the pH was adjusted to be higher than 7.3 by the addition of NaHCO3. The DO concentration was not directly controlled. For continuous rotation the rotation speed was fixed at 3 rpm and in the intermittent rotation mode, rotation at the same rotation speed occurred only 1/3 of the time, equally spread over time (1 min on, 2 min off).

2.2

Reactor operation

The influent of an OLAND lab-scale rotating biological contactor (RBC), as used by Vlaeminck et al. (2009b) to treat digested black water (Table 6.1), was switched to synthetic wastewater consisting of (NH4)2SO4, NaHCO3, KH2PO4 (10 mg P/L) and 2 mL L−1 of a trace element solution (Kuai and Verstraete, 1998). After a long term stable operation of the reactor treating 537 mg N L−1, the influent ammonium concentration was stepwise decreased to 278, 146, 66 and 31 mg N L−1 over 41, 48, 52 and 60 days, respectively, maintaining a constant loading rate (about 840 mg N L−1 d−1) by a stepwise decrease in hydraulic residence time (HRT) (Table 6.2). Each nitrogen influent concentration was applied for 1.5 to 2 months to obtain enough data points and stabilization for statistical comparison between the phases. Reactor pH, DO and temperature were daily monitored and influent and effluent samples were taken at least thrice a week for ammonium, nitrite and nitrate analyses.

2.3

Chemical analyses

Ammonium (Nessler method) was determined according to standard methods (Greenberg et al., 1992). Nitrite and nitrate were determined on a 761 compact ion chromatograph equipped with a conductivity detector (Metrohm, Zofingen, Switzerland). DO and pH were measured with respectively, an electrode installed on a C833 meter (Consort, Turnhout, Belgium) and a HQ30d DO meter (Hach Lange, Düsseldorf, Germany).

100

Chapter 6

2.4

Fluorescent in-situ hybridization (FISH)

At the start (537 mg N L−1) and at the end (29 mg N L−1) of experiment biomass samples were taken from the discs and bottom of the reactor for identification of the autotrophic nitrogen removing species present. At both time points FISH quantification of AerAOB, AnAOB and NOB was performed. A paraformaldehyde (4%) solution was used for biofilm fixation and FISH was performed according to Amann and coworkers (1990). Relevant target groups were gathered from a recent nitrogen cycle review (Vlaeminck et al., 2011), and probe sequences and formamide concentrations were applied according to (Lücker, 2010) for Nitrotoga and probeBase for the other targets (Loy et al., 2003): Amx820 for the AnAOB Kuenenia/Brocadia; a mixture of NSO1225 and NSO190 for the b-proteobacterial AerAOB Nitrosomonas/Nitrosospira; and NIT3 (+ competitor), Ntspa662 (+ competitor) and Ntoga221 for the NOB Nitrobacter, Nitrospira and Nitrotoga, respectively. The AnAOB, AerAOB and NOB abundance was evaluated by combining the specific probe with an equimolar mixture of EUB338I, II and III, targeting all bacteria, and 4'-6-diamidino-2-phenylindole (DAPI), targeting all DNA-containing cells. Image acquisition was done on a Zeiss Axioskop 2 Plus epifluorescence microscope (Carl Zeiss, Germany). For quantification, 20 randomly taken images were analyzed with ImageJ software, and the percentage of the specific group was calculated as the ratio of the specific area to the total DNA-containing area. The EUB338 signals served as a control.

2.5

Denaturing Gradient Gel Electrophoresis (DGGE)

At the start and at the end of experiment, biomass was harvested to compare the community structure (AerAOB, Planctomycetes and total bacteria) while treating high (537 mg N L−1) and low (29 mg N L−1) nitrogen concentrations, respectively. DNA extraction, nested PCR and DGGE were performed according to Pynaert et al. (2003), based on the primers CTO189ABf, CTO189Cf, and CTO653r for b-proteobacterial AerAOB; PLA40f and P518r for Planctomycetes, the bacterial phylum harbouring AnAOB; and GC338 and 518r for all bacteria. The obtained DGGE patterns were subsequently processed with BioNumerics software (Applied Maths, Sint-Martens-Latem, Belgium) and similarities were calculated as the Pearson correlation coefficient.

101

OLAND is feasible to treat sewage-like nitrogen concentrations at low HRT

3 3.1

Results Treatment of high nitrogen levels

Following the influent shift from digested black water (Vlaeminck et al., 2009b) to synthetic wastewater, the OLAND RBC was operated for 96 days at an influent concentration of 537 mg N L−1. Over the last 21 days of this period, the nitrogen removal rate was 642 mg N L−1 d−1 and the nitrogen removal efficiency was 79% (Table 6.2). The contributions of the different nitrogen pathways were quantified (Fig. 6.2), using the measured dissolved nitrogen species and assuming that (i) negligible denitrification occurred given the absence of bCOD in the influent, (ii) nitrogen gas was the product of the removed dissolved species and (iii) AnAOB produced 0.11 g NO3−-N per g NH4+-N converted to nitrogen gas. Initially, OLAND converted 90% of the influent nitrogen, and nitrite and nitrate accumulation were negligible at high nitrogen levels (Fig. 6.2). Indeed, no NOB could be detected in the biofilm with FISH. The average DO and free ammonia levels are relatively high at 1.4 mg O2 L−1 and 0.9 mg N L−1, respectively (Table 6.2). The AnAOB and AerAOB communities made up 5% and 23% of the biofilm, respectively (Table 6.3), and were composed of several species (Fig. 6.3).

3.2

Treatment of low nitrogen levels

The nitrogen influent concentration and HRT were gradually decreased over the phases II-Va while maintaining a constant nitrogen loading rate. Over these changes, the pH and DO levels remained in the same range, but the total nitrogen removal rate and hence the efficiency decreased significantly (p<0.05; Table 6.2). The proportion of ammonium oxidized remained relatively stable but nitrite and nitrate accumulated in the effluent (Table 6.2; Fig. 6.2), indicating that decreased anammox and increased nitratation were responsible for the decreased efficiency. Excess nitrate production by NOB significantly increased from period I to period II, remained constant for periods III and IV, and was followed by a significant increase in period Va (Table 6.2, Fig. 6.2). The NOB could be identified as Nitrospira, and composed 6% of the biofilm community at the lowest nitrogen concentration (Table 6.3). The free ammonia concentration decreased sharply over time whereas DO levels were stable, but relatively high (Table 6.2).

3.3

Suppression of nitratation at low nitrogen levels

In an attempt to decrease the DO level in phase Vb in order to decrease nitratation and increase the total nitrogen removal efficiency, discontinuous rotation was introduced. This 102

Chapter 6

resulted in a significant decrease of the oxygen concentration from 1.4 to 1.2 mg O2 L−1. Consequently, nitratation decreased with 35% and anammox increased with 32%, restoring the total nitrogen removal efficiency which was previously obtained in phase IV (Fig. 6.2; Table 6.2). The lower DO resulted in a total nitrogen effluent concentration of 20 mg N L−1 and a nitrogen removal rate of 383 mg N L−1 d−1, removing 46% of the influent nitrogen (Table 6.2). The used RBC set-up did not allow to further decrease of the DO levels since rotating more intermittently resulted in a strong decrease of the nitritation rate (data not shown). The relatively stable nitritation indicated few or no influence by the decreasing HRT (Fig. 6.2; Table 6.2). The decreasing AnAOB activity at lower HRT could be partly counteracted by a lower DO level and resulting lower nitratation in period Vb (Fig. 6.2), indicating the influence of DO level on NOB/AnAOB competition rather than a negative effect of low HRT on anammox

Figure 6.2: Contributions of microbial conversions to reactor nitrogen products: nitrogen gas production (nitrogen in – nitrogen out) by anoxic ammonium-oxidizing bacteria (AnAOB) from influent ammonium (upper part) and from nitrite produced by aerobic ammonium oxidizing bacteria (AerAOB; lower part); nitrate production by AnAOB (11% of ammonium converted by balanced OLAND); nitrate production by nitrite oxidizing bacteria (NOB: nitrate out – nitrate in – nitrate production by AnAOB); excess nitrite production by AerAOB (nitrite out – nitrite in); and residual ammonium (ammonium in – ammonium out). 103

OLAND is feasible to treat sewage-like nitrogen concentrations at low HRT Table 6.2: OLAND rotating biological contactor conditions and performance (average ± standard deviation) over the periods with stepwise decreases of the ammonium influent concentration and hydraulic residence time (HRT). In periods I-Va, rotation was continuous, whereas this was intermittent in period Vb. For the eight bottom rows, statistical analyses were performed and the phases that were not significantly different (p>0.05) are indicated with the number of the similar phase. d: days; h: hours; DO: dissolved oxygen level; prod.: production; cons.: consumption Period I II III IV Va Vb Duration (d)

Number of samples (-)

Influent NH4+ level (mg N L−1)

537  13

278 ± 11

146 ± 21

66 ± 5

29 ± 8

31  1

Influent flow rate (L d−1)

5.4  0.2

10.5  0.3

20.5  1.5

42.9  2.3

82.6  2.0

83.6  0.7

HRT (h)

16.0 ± 0.5

8.3 ± 0.3

4.2 ± 0.4

2.0 ± 0.1

1.0 ± 0.0

1.0  0.0

N loading rate (mg N L−1 d−1)

819 ± 30

840 ± 49

832 ± 68

855 ± 56

851 ± 66

840  20

DO level (mg O2 L−1)

1.4 ± 0.2III,Va

1.2 ± 0.2IV,Vb

1.4 ± 0.2I,Va

1.2 ± 0.1II,Va

1.4 ± 0.4I,III

1.2  0.1II,IV

pH (-)

7.6 ± 0.1

7.5 ± 0.1

7.3 ± 0.2IV,Va,Vb

7.4 ± 0.1III

7.3 ± 0.2III,Vb

7.3  0.0III,Va

Free ammonia (mg N L−1)*

0.91  1.58II,III

0.40  0.15III,I

0.40  0.17I,II

0.10  0.03

0.04  0.02Vb

0.04  0.01Va

N removal rate (mg N L−1 d−1)

642 ± 72

565 ± 42

471 ± 88IV

444 ± 84III,Vb

303 ± 75

383  52IV

N removal efficiency (%)

79 ± 9

67 ± 3

58 ± 9

51 ± 8Vb

35 ± 7

46  6IV

NH4+ removal efficiency (%)

94  10

91  3IV,Va,Vb

72  26

89  4II,Va,Vb

77  31II,IV,Vb

91  5II,IV,Va

NO3−prod./NH4+ cons. (%)**

12 ± 2

22 ± 2IV

18 ± 8IV

21 ± 6II,III

45 ± 11

32  6

−1

25 ± 10

I,III

29 ± 12

3.7 ± 1.5

3.0 ± 1.0Va

Effluent NH4 (mg N L )

19 ± 10

Effluent NO2- (mg N L−1)

19 ± 5III

10 ± 12

16 ± 6IV,I

14 ± 3III

5.2 ± 1.3Vb

5.1 ± 0.4Va

Effluent NO3- (mg N L−1)

65 ± 4

59 ± 5

21 ± 9

14 ± 2Va

15 ± 2IV

11 ± 0.9

* Calculated from the measured ammonium level, temperature and pH (Anthonisen et al. 1976) ** Values exceeding 11% indicate excess nitrate production by nitrite oxidizing bacteria (NOB) *** Sum of ammonium, nitrite and nitrate

104

7.4 ± 2.7

Chapter 6

While treating 29 mg N L−1, the AnAOB and AerAOB abundances in the biomass were 8 and 25%, respectively (Table 6.3), which was quite comparable to the abundance while treating high nitrogen concentrations. In the final operation period, part of the biomass was found at the bottom of the reactor, but neither FISH nor DGGE could detect important differences in the microbial composition of settled biomass versus biofilm on the discs (Table 6.3, Fig. 6.3), indicating that the biomass was probably the result of detachment of the biofilm from the discs. For the AnAOB communities, the DGGE profiles showed only small changes in the abundant species (88% similarity), while the AerAOB patterns changed more (23% similarity) (Fig. 6.3). The shift in the most abundant AerAOB could also be observed in the DDGE profiles of all bacteria.

Figure 6.3: DGGE gels for -proteobacterial aerobic ammonium-oxidizing bacteria (AerAOB) and Planctomycetes (Plancto), the phylum harbouring anoxic ammonium-oxidizing bacteria (AnAOB). Biomass samples were taken at the end of the treatment period at 537 mg N L−1 (high N; sample from the disc biofilm) and at 31 mg N L−1 (low N; sample from the disc biofilm and from settled biomass). Similarities were calculated as the Pearson correlation coefficient, and plus symbols highlight the three AerAOB and Planctomycetes bands with the highest intensity, indicating shifts of the most dominant species.

Table 6.3: Abundances of aerobic and anoxic ammonium-oxidizing bacteria (AerAOB and AnAOB) and nitrite oxidizing bacteria (NOB) in OLAND biomass, as determined from quantitative fluorescent in-situ hybridization (FISH). The NOB genera Nitrobacter and Nitrotoga could not be retrieved. ND: not detected; NM: not measured -1 Influent (mg L ) 537 29 Biomass sample Biofilm Biofilm Settled AerAOB Nitrosomonas/Nitrosospira (%) AnAOB Kuenenia/Brocadia (%) NOB Nitrospira (%)

23 ± 18 5±5

22 ± 12 7±4

30 ± 16 8±6

6±5

5±5

105

OLAND is feasible to treat sewage-like nitrogen concentrations at low HRT

4 4.1

Discussion OLAND removal rate and efficiency treating low nitrogen levels

In this study, operation of the OLAND RBC on sewage-like nitrogen concentrations (66 and 29 mg N L−1) at low HRT (2 and 1 h) resulted in nitrogen removal rates of 383-444 mg N L−1 d−1, which are reasonably high (Table 6.1). In the energy-positive sewage treatment scheme (Fig. 1.B), 20-25% of the sewage COD remains in the main line following an advanced concentration step (Verstraete et al., 2009) as well as about 80% of the sewage nitrogen, partly derived from returning the digestate to the main line. Assuming raw sewage compositions of 825 mg COD L−1 and 65 mg N L−1, the OLAND step would hence receive 165 mg COD L−1 and 52 mg N L−1. Given the overall required removal efficiencies of 5060% of COD and 75% of the nitrogen according to European standards (European Commision, 1991), the OLAND step should remove an additional 36 mg N L−1 and hence the desired OLAND removal efficiency should be around 70%. In this study, nitrogen removal efficiencies during treatment of low nitrogen levels were 46-51%, and hence not sufficiently high to comply with the required standards. The obtained nitrogen removal percentages were lower than previously reported for this type of reactors (Pynaert et al., 2003; Schmid et al., 2003; Pynaert et al., 2004), mainly due to additional nitratation. Also in absolute terms, the effluent nitrogen concentrations of around 20 mg N L−1 were slightly above the discharge requirements (> 15 mg N L−1; European Commision, 1991). Since AerAOB and AnAOB have high affinities for their nitrogen substrates, with half-saturation constants of 0.05-2.4 mg N L-1 (Lackner et al., 2008), the microbial capacity should allow further optimization.

4.2

Role of DO levels in suppressing nitratation

The DO levels in the RBC were 1.2-1.4 mg O2 L−1 (Table 6.1) and therefore not low enough to suppress NOB growth (Bernet et al., 2001; Joss et al., 2009), resulting in a substantial nitratation (Fig. 6.2). Indeed, NOB could not be detected treating 537 mg N L−1, but the NOB genus Nitrospira colonized the biomass at lower nitrogen influent levels, leading to a final abundance of 5-6%. In contrast to Nitrobacter and Nitrotoga, Nitrospira is typically found in systems under oxygen-limited conditions, relatively low nitrite levels and moderate temperature (Lücker et al., 2010). Free ammonia levels between 0.08-0.8 mg N L−1 can inhibit nitratation (Anthonisen et al., 1976), and could have played a role primarily in period I (0.9 mg NH3-N L−1). A DO decrease by 0.2 mg O2 L−1 during phase Vb lowered nitratation with 35% (Fig. 6.2), demonstrating the link between DO level and NOB activity. It is 106

Chapter 6

anticipated that controlled operation at a sufficiently low DO setpoint (e.g. 0.3 mg O 2 L−1) will effectively suppress NOB at long term, as demonstrated for treatment of higher strength OLAND applications (Joss et al., 2009). In an OLAND RBC, it is less straightforward to control the DO experienced by the biomass than in systems based on active aeration in which the biomass is either suspended (e.g. Joss et al., 2009) or attached to submerged carrier material (e.g. Szatkowska et al., 2007). Lower RBC DO levels can generally be obtained by decreasing the rotor speed (e.g. Meulman et al., 2010), which was not possible on the RBC in this study, or by increasing the immersion level of the disks. These two actions influence the biofilm exposure time to atmospheric oxygen and the input turbulence of air in the bulk liquid by rotation. An additional control of the oxygen level in gas phase of the RBC could further optimize the microbial balance. In practice, the higher oxygen demand in the presence of organics (165 mg COD L−1), will also yield lower bulk DO levels at a similar rotor speed as in the presence of ammonium only.

4.3

OLAND operation at low HRT

Compared to described OLAND systems, the applied HRT in this study were very low (Table 6.1) but this did not seem to have an adverse effect on AerAOB or AnAOB activity. It is not clear whether an expected higher biomass washout at lower HRT could have been responsible for shifts in the microbial community. The sequentially decreasing nitrogen concentrations in the reactor (Table 6.2) possibly had a stronger influence on establishing an adapted OLAND microbiome which was likely more oligotrophic.

Compared to treatment at high HRT (e.g. 24 h), applying lower HRT (e.g. 1 h) at high volumetric loading rates will have an influence on the design parameters, depending on the type of OLAND reactor. In suspended growth systems, biomass retention is based on settling. In case of an external settler, a lower HRT will result in a higher sludge surface load, and hence needs a relative increase of the settler volume compared to the reactor volume to maintain the same sludge residence time. In case of a sequencing batch reactor, decreased HRT will require an increased minimum biomass settling velocity from 1 m h−1 (Joss et al., 2009) to 24 m h−1 to maintain an acceptably low ratio of settling to reaction time, and therefore will require granules rather than flocs (Chapter 2; Vlaeminck et al., 2010). In contrast to suspended growth configurations, HRT in biofilm-based systems is expected to have only a minor influence on the biomass retention, allowing for compact reactors.

107

OLAND is feasible to treat sewage-like nitrogen concentrations at low HRT

4.4

Implementation of OLAND in the main stream

OLAND treatment of pretreated sewage should achieve sufficiently high nitrogen removal rates and efficiencies at low hydraulic residence times and nitrogen concentrations at minimal energy requirements, given the overall aim of energy-positive sewage treatment. Overall, several decision factors will determine the desirable reactor technology. Passive versus active aeration will determine energy requirements, but also the ease of controlling the microbial activity balance, and suspended versus attached biomass growth will determine the ease of maintaining a high biomass retention at low HRT.

The next research challenges for the implementation of OLAND in the main stream of the sewage treatment relate firstly to a decrease of the process temperatures from the maximum summer temperature (25°C) over the average year temperature (17°C) to the minimum winter temperature (8°C) (Breda, NL; Mollen, oral communication). This will elucidate whether OLAND requires a distinct oligotrophic and cold-tolerant autotrophic community and physiology. Secondly, the continued OLAND performance will have to be shown in the presence of moderate bCOD levels (90-240 mg L−1), with COD/N ratios between 2.4 and 3. The latter will likely facilitate DO control at low DO levels due to heterotrophic aerobic activity. However, also competition for nitrite will take place between heterotrophic denitrification and anammox. These processes have however been demonstrated already to successfully co-exist at a COD/N of 2.2 (Desloover et al., 2011). It is anticipated that due to future dilution preventions (Henze, 1997; Brombach et al., 2005), higher nitrogen sewage levels together with the higher sewage temperature will facilitate OLAND treatment in the main stream. Finally, high OLAND performance will have to be shown under realistic temporal variations in sewage composition and in performance of the preceding advanced concentration.

Acknowledgements

H.D.C. received a PhD grant from the Institute for the Promotion of Innovation by Science and Technology in Flanders (IWT-Vlaanderen, SB-81068) and S.E.V. was supported as a postdoctoral fellow from the Research Foundation Flanders (FWO-Vlaanderen). The authors gratefully thank Hans Mollen (Waterschap Brabantse Delta, NL) for sharing temperature data, Siska Maertens for molecular analyses, and Nico Boon, Tom Hennebel, Jan Arends, Yu Zhang, 108

Sebastià

Puig

and

Samik

Bagchi

for

inspiring

scientific

discussions.

Chapter 7

Chapter 7: Cold OLAND on pretreated sewage: feasibility demonstration at lab-scale Abstract Energy-positive sewage treatment can be achieved by implementation of oxygen-limited autotrophic nitrification/denitrification (OLAND) in the main water line, as the latter does not require organic carbon and therefore allows maximum energy recovery through anaerobic digestion of organics. To test the feasibility of mainstream OLAND, the effect of a gradual temperature decrease from 29°C to 15°C and a COD/N increase from 0 to 2 was tested in an OLAND rotating biological contactor (RBC) operating at 55-60 mg NH4+-N L-1 and a hydraulic retention time of 1 hour. Moreover, the effect of the operational conditions and feeding strategies on the reactor cycle balances, including NO/N2O emissions were studied in detail. At 15°C (9 months) high anoxic and aerobic ammonium oxidation activities were maintained. However, nitratation (NOB activity) occurred at temperatures below 20°C. Operation at COD/N ratios of 2 and 15°C (2 months) still allowed for high nitrogen removal rates of 0.5 g N L-1 d-1, which are in the same range as high temperature applications. The main challenge to allow high removal efficiencies in this application was the suppression of NOB at low free ammonia (< 0.25 mg N L-1), low free nitrous acid (<0.9 μg N L-1) and higher DO levels (3-4 mg O2 L-1). This study showed that high NO levels had the potential to favor anammox above NOB activity. It should be evaluated if the increased NO/N2O emission can be compensated with a decreased energy consumption to justify OLAND mainstream treatment.

Chapter redrafted after: De Clippeleir H., Vlaeminck S.E., De Wilde F., Daeninck K., Mosquera M., Boeckx P., Verstraete W. and Boon N. Cold one-stage partial nitritation/anammox on pretreated sewage: feasibility demonstration at lab-scale. Submitted. 109

Cold OLAND on preteated sewage: feasibility demonstration on lab-scale

Introduction

Currently, around 40 full-scale 1 stage partial nitritation/anammox plants are implemented to treat highly loaded nitrogen streams devoid in carbon (Chapter 1). This process, known under the acronyms OLAND (Kuai and Verstraete, 1998), DEMON (Wett, 2006), CANON (Third et al., 2001) etc, showed highly efficient and stable performance when treating digestates from sewage sludge treatment plants and industrial wastewaters (Wett, 2006; Abma et al., 2010; Jeanningros et al., 2010). From an energy point of view, the implementation of the OLAND process for the treatment of sewage sludge digestate decreased the net energy consumption of a municipal wastewater treatment plant (WWTP) with 50% (Siegrist et al., 2008). Moreover, when co-digestion of kitchen waste was applied, an energy neutral WWTP was achieved (Wett et al., 2007). To fully recover the potential energy present in wastewater, a ‘ZeroWasteWater’ concept was proposed which replaces the conventional activated sludge system by a highly loaded activated sludge step (A-step), bringing as much as organic carbon (COD) as possible to the solid fraction, and a second biological step (B-step) removing the residual nitrogen and COD with a minimal energy demand (Verstraete and Vlaeminck, 2011). Subsequently, energy is recovered via anaerobic digestion of the primary and secondary sludge. For the B-step in the main line, OLAND would potentially be the best choice as this process can work at low COD/N ratio, allowing maximum recovery of COD in the A-step. Moreover, it was calculated that if OLAND is implemented in the main water treatment line and a maximum COD recovery takes place in the A-step, a net energy gain of the WWTP of 10 Wh inhabitant equivalent (IE)-1 d-1 is feasible (Chapter 4).

To allow this energy-positive sewage treatment, OLAND has to face some challenges compared to the treatment of highly loaded nitrogen streams (> 250 mg N L-1). A first difference is the lower nitrogen concentration to be removed by OLAND. Domestic wastewater after advanced concentration will still contain around 30-100 mg N L-1 and 113300 mg COD L-1 (Tchobanoglous et al., 2003; Henze et al., 2008). High nitrogen conversion rate (around 400 mg N L-1 d-1) by the OLAND process can be obtained at nitrogen concentrations of 30-60 mg N L-1 and at low hydraulic retention times (HRT) of 1-2 hours (Chapter 6). A second challenge is the low temperature at which OLAND should be operated (10-15°C compared to 34°C). Several studies already described the effect of temperature on the activity of the separate microbial groups (Dosta et al., 2008; Guo et al., 2010; Hendrickx et al., 2012). However, limited knowledge exists about the microbial balances of these 110

Chapter 7

3 groups under OLAND conditions at low temperature (< 20°C). At temperatures around 15°C, maintaining the balance between nitrite-oxidizing bacteria (NOB) and anoxic ammonium-oxidizing bacteria (AnAOB) and the balance between NOB and aerobic ammonium-oxidizing bacteria (AerAOB) will get more challenging since the growth rate of NOB will become higher than the growth rate of AerAOB (Hellinga et al., 1998). Therefore, it will not be possible to wash out NOB based on overall or even selective sludge retention. The third and main challenge in this application will therefore be the suppression of NOB at temperature ranges of 10-20°C and at nitrogen concentration ranges of 30-60 mg N L-1 (low free ammonia and low nitrous acid). A final fourth challenge will include the higher input of organics at moderate levels of 90-240 mg bCOD L-1 in the wastewater. Depending on the raw sewage strength, COD/N ratios between 2 and 3 are expected after the concentration step, which is on the edge of the described limit for successful OLAND (Lackner et al., 2008). The presence of organics could result in an extra competition of heterotrophic denitrifiers with AerAOB for oxygen or with AnAOB for nitrite.

In this study an OLAND RBC at 29°C was gradually adapted over 24, 22 and 17°C to 15°C under synthetic wastewater conditions (60 mg N L-1, COD/N of 0). Additionally, the COD/N ratio of the influent was increased to 2 by supplementing NH4+ to diluted sewage to simulate pretreated sewage. The effect of the operational conditions and feeding strategies on the reactor cycle balances, including gas emissions and microbial activities were studied in detail. An alternative strategy to inhibit NOB activity and as a consequence increase AnAOB activity at low temperatures was proposed.

2 2.1

Materials and methods OLAND rotating biological contactor (RBC)

The lab-scale RBC described in Chapter 6 was further optimized at 29°C by an increase in the influent nitrogen concentration from 30 to 60 mg N L-1 and a limitation of the oxygen input through the atmosphere by covering the reactor before this test was started. The reactor was based on an air washer LW14 (Venta, Weingarten, Germany) with a rotor consisting of 40 discs interspaced at 3 mm, resulting in a disc contact surface of 1.32 m2. The reactor had a liquid volume of 2.5 L, immersing the discs for 55%. The latter was varied over the time of the experiment. The reactor was placed in a temperature-controlled room. The DO

111

Cold OLAND on preteated sewage: feasibility demonstration on lab-scale

concentration was not directly controlled. For continuous rotation the rotation speed was fixed at 3 rpm.

2.2

RBC operation

The RBC was fed with synthetic wastewater during phases I to VII. From phase VIII onwards, the COD/N was gradually increased (Phase VIII-X) to 2 (phase XI-XIII). The synthetic influent of an OLAND RBC, consisted of (NH4)2SO4 (55-60 mg N L-1), NaHCO3 (16 mg NaHCO3 mg-1 N) and KH2PO4 (10 mg P L-1). Pretreated sewage was simulated by diluting raw sewage of the communal WWTP of Gent, Belgium (Aquafin). The raw wastewater contained 23-46 mg NH4+-N L-1, 0.2-0.4 mg NO2--N L-1, 0.4-2.7 mg NO3--N L-1, 23-46 mg Kjeldahl-N L-1, 3.8-3.9 mg PO43--P L-1, 26-27 mg SO42--S L-1, 141-303 mg CODtot L-1 and 74-145 mg CODsol L-1. The raw sewage was diluted by a factor 2-3 to obtain COD values around 110 mg CODtot L-1 and by addition of (NH4)2SO4 to obtain final COD/N values around 2. The reactor was fed in a semi-continuous mode: 2 periods of around 10 minutes per hour for phases I-XI, 1 period of 20 minutes per hour for phases XII and XIII. Reactor pH, DO and temperature were daily monitored and influent and effluent samples were taken at least thrice a week for ammonium, nitrite, nitrate and COD analyses.

2.3

Detection of AerAOB, NOB and AnAOB with FISH and qPCR

For NOB and AnAOB, a first genus screening among the most commonly present organisms was performed by fluorescent in situ hybridization (FISH) on biomass of day 1 (high temperature) and day 435 (low temperature and COD presence). A paraformaldehyde (4%) solution was used for biofilm fixation, and FISH was performed according to Amann et al. (1990). The Sca1309 and Amx820 probes were used for the detection of Cand. Scalindua and Cand. Kuenenia & Brocadia, respectively, and the NIT3 and Ntspa662 probes and their competitors for Nitrobacter and Nitrospira, respectively. This showed the absence of Nitrobacter and Scalindua (Table S7.1). Biomass samples (approx. 5 g) for nucleic acid analysis were taken from the OLAND RBC at days 1, 60, 174, 202, 306, 385, 399 and 413 of the operation. DNA was extracted using FastDNA® SPIN Kit for Soil (MP Biomedicals, LLC), according to the manufacturer’s instructions. The obtained DNA was purified with the Wizard® DNA Clean-up System (Promega, USA) and its final concentration was measured spectrophotometrically

using

NanoDrop

ND-1000

spectrophotometer

(Nanodrop

Technologies). The SYBR Green assay (Power SyBr Green, Applied Biosystems) was used to quantify the 16S rRNA of AnAOB and Nitrospira sp. and the functional amoA gene for 112

Chapter 7

AerAOB. The primers for quantitative polymerase chain reactions (qPCR) for detection of AerAOB, NOB and AnAOB were amoA-1F – amoA-2R, NSR1113f-NSR1264r and Amx818f – Amx1066r, respectively. For bacterial amoA gene, PCR conditions were: 40 cycles of 94°C for 1 min, 55°C for 1 min and 60°C for 2 min. For the amplification of Nitrospira sp 16S rRNA gene 40 cycles of 95°C for 1 min, 50°C for 1 min and 60°C for 1 min were used while for AnAOB 16S rRNA the PCR temperature program was performed by 40 cycles of 15 seg at 94°C and 1 min at 60°C. Plasmid DNAs carrying Nitrospira and AnAOB 16SrRNA gene and AerAOB functional AmoA gene, respectively, were used as standards for qPCR. All the amplification reactions had a high correlation coefficient (R 2> 0.98) and slopes between -3.0 and -3.3. A paraformaldehyde (4%) solution was used for biofilm fixation, and FISH was performed according to Amann et al. (1990). The Bfu613 probe was used for the detection Brocadia fulgida (Kartal et al., 2008) and EUB I,II,II for detection of all bacteria.

2.4

Detailed reactor cycle balances

For the measurements of the total nitrogen balance, including the NO and N2O emissions, the OLAND RBC was placed in a vessel (34 L) which had a small opening at the top (5 cm2). In this vessel a constant upward air flow (around 1 m s-1) was generated to allow calculations of emission rates. On the top of the vessel (air outlet), the NO and N2O concentration was measured, off- and online, respectively. In the water phase, ammonium, nitrite, nitrate, hydroxylamine (NH2OH), N2O and COD concentrations were measured. Moreover, DO concentration and pH values were monitored. The air flow was measured with Testo 425 hand probe (Testo, Ternat, Belgium).

2.5

Chemical analyses

Ammonium (Nessler method) was determined according to standard methods (Greenberg et al., 1992). Nitrite and nitrate were determined on a 761 compact ion chromatograph equipped with a conductivity detector (Metrohm, Zofingen, Switzerland). Hydroxylamine was measured spectrophotometrically (Frear and Burrell, 1955). The chemical oxygen demand (COD) was determined with NANOCOLOR® COD 1500 en NANOCOLOR® COD 160 kits (Macherey-Nagel, Düren, Germany). The volumetric nitrogen conversion rates by AerAOB, NOB and AnAOB were calculated based on the measured influent and effluent compositions and the described stoichiometries (Vlaeminck et al., 2012). DO and pH were measured with respectively, a HQ30d DO meter (Hach Lange, Düsseldorf, Germany) and an electrode 113

Cold OLAND on preteated sewage: feasibility demonstration on lab-scale

installed on a C833 meter (Consort, Turnhout, Belgium). Gaseous N2O concentrations were measured online at a time interval of 3 minutes with a photo-acoustic infrared multi-gas monitor (Brüel & Kjær, Model 1302, Nærem, Denmark). Gas grab samples were taken during the detailed cycle balance tests for NO detection using Eco Physics CLD 77 AM (Eco Physics AG, Duernten, Switzerland), which is based on the principle of chemiluminescence. For dissolved N2O measurements, a 1 mL filtered (0.45 μm) sample was brought into a 7 mL vacutainer (-900 hPa) and measured afterwards by pressure adjustment with He and immediate injection at 21°C in a gas chromatograph equipped with an electron capture detector (Shimadzu GC-14B, Japan).

3 3.1

Results Effect of temperature decrease

During the reference period (29°C), a well-balanced OLAND performance (Fig. 7.1, Table 7.1) was reached with minimal nitrite accumulation (2%) and minimal nitrate production (7%). This was reflected in an AerAOB/AnAOB activity ratio of 0.6 (Table 7.1, Phase I). The total nitrogen removal rate was on average 470 mg N L-1 d-1 and the total nitrogen removal efficiency was 54%.

Decreasing the temperature from 29 to 24°C and further to 22°C over the following 40 days, did not result in any significant changes of the operational conditions (Table 7.1, Phases I-III), performance of the reactor (Fig. 7.1) or abundance of the bacterial groups (qPCR, Fig. S7.1). However at 17°C, a decrease in total nitrogen removal efficiency was observed (Table 7.1, Phase IV). As the ammonium consumption rate went down (from 501  23 to 383  80 mg N L-1 d-1) and the effluent nitrite and nitrate levels remained stable (Fig. 7.1), a higher relative nitrite and nitrate production indicated an imbalance between the AerAOB and the AnAOB. Moreover, NOB activity was for the first time detected while no difference in free ammonia (FA) or free nitrous acid (FNA) suppression on NOB was observed (Table 7.1, Phase IV). Moreover, no significant differences in abundance of NOB, AerAOB and AnAOB could be detected with qPCR (Fig. S7.1). However, DO concentrations started to increase during that period from 1.4 to 1.7 mg O2 L-1. To counteract the decrease in ammonium removal efficiency the immersion level was lowered to 55% to increase the availability of oxygen. Consequently the volumetric loading rate increased (factor 1.7) due to the decrease in reactor volume (day 210, Fig. 7.1). This action allowed higher ammonium removal efficiencies due 114

Chapter 7

to higher AerAOB activities (factor 3). AnAOB activity increased with a similar factor as the volumetric loading rate (1.8 compared to 1.7) consequently resulting in an increased imbalance between these two groups of bacteria (Table 7.1, Phase V). Moreover, although the FNA increased with a factor 2, the NOB activity increased with a factor 7, resulting in a relative nitrate production of 30% (Table 7.1, Phase V). As NOB activity prevented good total nitrogen removal efficiencies, the immersion level was increased again to 78% (day 263, Fig. 7.1). This resulted indeed in a lower NOB activity (Table 7.1, Phase VI). However, also the AerAOB activity decreased with the same factor, due to the lower availability of atmospheric oxygen. Therefore, the reactor was subsequently operated at this low immersion level (55%) to allow sufficient aerobic ammonium conversion. The latter allowed a stable removal efficiency of 42%. The AnAOB activity gradually increased to a stable anoxic ammonium conversion rate of 529 mg N L-1 d-1. During the synthetic phase, no changes in AerAOB, AnAOB and NOB abundance were measured with qPCR (Fig. S7.1). The effluent quality was however not optimal as still high nitrite (around 15 mg N L-1) and nitrate (around 13 mg N L-1) levels were detected.

115

Cold OLAND on preteated sewage: feasibility demonstration on lab-scale

Figure 7.1: Phases I-VII: Effect of temperature decrease on the volumetric rates (top) and nitrogen concentrations (bottom).

116

Chapter 7

Table 7.1: Effect of temperature decrease on the operational conditions and performance of OLAND RBC reactor. DO: dissolved oxygen; HRT: hydraulic retention time; FA: free ammonia; FNA: free nitrous acid; cons: consumption; tot: total Phase I II III IV V VI VII Period (d) 1-21 22-35 36-61 62-210 210-263 263-274 275-306 Immersion level (%) 78 78 78 78 55 78 55 Temperature (°C) 29 ± 2 24 ± 1 22 ± 0.6 17 ± 1.2 16 ± 0.9 15 ± 0.8 14 ± 0.4 Operational conditions: DO (mg O2 L-1) 1.1 ± 0.2 1.3 ± 0.2 1.4 ± 0.1 1.7 ± 0.3 2.8 ± 0.4 2.4 ± 0.2 3.1 ± 0.2 pH (-) 7.5 ± 0.1 7.5 ± 0.1 7.5 ± 0.1 7.6 ± 0.1 7.7 ± 0.1 7.7 ± 0.1 7.8 ± 0.1 HRT (h) 1.85 ± 0.04 1.84 ± 0.09 1.73 ± 0.04 1.86 ± 0.11 1.09 ± 0.02 1.57 ± 0.02 1.09 ± 0.02 FA (mg N L-1) 0.35 ± 0.18 0.36 ± 0.18 0.34 ± 0.14 0.36 ± 0.13 0.25 ± 0.16 0.33 ± 0.17 0.13 ± 0.04 FNA (μg N L-1) 0.3 ± 0.1 0.3 ± 0.2 0.4 ± 0.2 0.4 ± 0.1 0.9 ± 0.4 0.6 ± 0.1 0.9 ± 0.2 Performance: Total N removal efficiency (%) 54 ± 5 52 ± 5 49 ± 9 34 ± 9 36 ± 9 36 ± 9 42 ± 4 + Relative NO3 prod (% of NH4 cons*) 7±1 7±1 7±1 14 ± 6 18 ± 9 16 ± 3 21 ± 4 Relative NO2- prod (% of NH4+ cons) 2±4 3±4 5±5 15 ± 5 30 ± 8 26 ± 6 31 ± 5 + -1 -1 AerAOB activity (mg NH4 -N L d ) 267 ± 38 267 ± 49 260 ± 52 260 ± 53 811 ± 229 460 ± 44 986 ± 71 NOB activity (mg NO2—N L-1 d-1) 0±0 0±0 0±0 9 ± 12 60 ± 94 20 ± 5 85 ± 25 -1 -1 AnAOB activity (mg Ntot L d ) 412 ± 38 403 ± 37 368 ± 76 248 ± 67 448 ± 117 305 ± 74 529 ± 75 + *NH4 consumption is corrected for nitrite accumulation

117

Cold OLAND on preteated sewage: feasibility demonstration on lab-scale

3.2

Effect of COD/N increase

The synthetic feed was gradually changed into pretreated sewage by diluting raw sewage and adding additional nitrogen to obtain a certain COD/N ratio. During the first 3 weeks of this period (Fig. 7.2), the COD/N ratio was gradually increased from 0.5 to 2. Due to the short adaptation periods (1 week per COD/N regime), the performance was unstable (Fig. 7.2, Table 7.2, phase VIII-XI). Compared to the end of the synthetic period (phase VII), operation at a COD/N ratio of 2 (phase XI) resulted in a sharp decrease in nitrite accumulation (Fig. 7.2) and an increase in the ammonium and nitrate levels. This indicated increased NOB activity (factor 4), decreased AerAOB (factor 3) and decreased AnAOB (factor 2) activity (Table 7.1 and 7.2). To allow higher nitrogen removal rates, the HRT was increased from 0.94 to 1.1 h, by decreasing the influent flow rate. Moreover, the feeding regime was changed from 2 pulses of 10 minutes in 1 hour to 1 period of 20 minutes per hour. These actions did not significantly decrease the effluent nitrogen concentration (Fig. 7.2) and did not influence the microbial activities (Table 7.2, phase XII). Therefore the loading rate was again increased to the levels before phase XII. However the single-pulse feeding was maintained. This resulted in high ammonium removal efficiencies and therefore low ammonium effluent concentration around dischargeable level (4 ± 1 mg NH4+-N L-1; Fig. 7.2). Nitrate and nitrite accumulation were not counteracted by denitrification as only 0.02 mg COD L-1 d-1 was removed. Therefore nitrite and nitrate levels were still too high to allow effluent discharge. The total nitrogen removal efficiency (42%) and rate (549 ± 83 mg N L-1 d-1) at COD/N ratios of 2 was similar as during the synthetic period (phase VII). Compared to the reference period at 29°C, 22% of the removal efficiency was lost, while the total nitrogen removal rate did not changed significantly (470 ± 43 versus 549 ± 83 mg N L-1 d-1 at high and low temperature, respectively).

118

Chapter 7

Figure 7.2: Phases VIII-XIII: Effect of COD/N increase on the volumetric rates (top) and nitrogen concentrations (bottom) at 15째C and an immersion level of 55%.

119

Cold OLAND on preteated sewage: feasibility demonstration on lab-scale

Table 7.2: Effect of COD/N increase on the operational conditions and performance of OLAND RBC reactor at 15°C and an immersion level of 55%.. DO: dissolved oxygen; HRT: hydraulic retention time; FA: free ammonia; FNA: free nitrous acid; cons: consumption; tot: total Phase VIII IX X XI XII XIII Period (d) 355-361 362-369 370-374 375-406 407-421 422Immersion level (%) 55 55 55 55 55 55 COD/N 0.5 1 1.5 2 2 2 Feeding regime (pulses h-1) 2 2 2 2 1 1 Operational conditions: DO (mg O2 L-1) 2.9 ± 0.3 2.5 ± 0.6 2.4 ± 0.3 3.0 ± 0.7 3.6 ± 0.3 3.2 ± 0.3 pH (-) 7.8 ± 0.02 7.7 ± 0.1 7.6 ± 0.02 7.6 ± 0.1 7.6 ± 0.2 7.6 ± 0.1 HRT (h) 1.06 ± 0.11 1.03 ± 0.02 0.92 ± 0.02 0.94 ± 0.05 1.10 ± 0.05 1.06 ± 0.2 FA (mg N L-1) 0.10 ± 0.05 0.04 ± 0.05 0.15 ± 0.05 0.21 ± 0.10 0.23 ± 0.12 0.04 ± 0.02 FNA (μg N L-1) 0.4 ± 0.1 0.2 ± 0.2 0.2 ± 0.01 0.3 ± 0.1 0.2 ± 0.1 0.6 ± 0.2 Performance: Total N removal efficiency (%) 36 ± 5 45 ± 18 23 ± 3 28 ± 6 23 ± 13 42 ± 3 Relative NO3- prod (% of NH4+ cons*) 42 ± 5 43 ± 12 63 ± 2 50 ± 6 62 ± 18 46 ± 6 + Relative NO2 prod (% of NH4 cons) 20 ± 4 10 ± 10 5±1 8±3 7±4 13 ± 6 AerAOB activity (mg NH4+-N L-1 d-1) 592 ± 15 446 ± 31 238 ± 28 352 ± 73 289 ± 138 600 ± 204 NOB activity (mg NO2--N L-1 d-1) 257 ± 19 294 ± 81 465 ± 60 352 ± 84 427 ± 115 394 ± 76 AnAOB activity (mg Ntot L-1 d-1) 385 ± 86 452 ± 205 262 ± 39 355 ± 73 281 ± 159 481 ± 73 + *NH4 consumption is corrected for nitrite accumulation

120

Chapter 7

3.3

Nitratation and NO/N2O emissions

At the end of the synthetic phase (Phase VII) and the end of the experiment (Phase XIII) the total nitrogen balance of the reactor was measured. A total nitrogen balance was obtained by measuring all nitrogen species (NH4+, NO2-, NO3-, NH2OH, N2O) in the liquid phase and N2O and NO in the gas phase. A constant air flow, diluting the emitted N2O and NO concentrations was created over the reactor to measure gas fluxes over time. The effect of the loading rate, feeding pattern and concentration of nitrite and ammonium on the total nitrogen balance in the reactor were tested (Table 7.3). NH2OH measurement showed low concentrations (< 0.2 mg N L-1) in all tests, making it difficult to link the profiles with the N2O emission. Lowering the loading rate by increasing the HRT (Test B, Table 7.3) increased the DO values and allowed higher DO fluctuations over time at synthetic conditions. Moreover NOB activity increased significantly resulting in lower total nitrogen removal efficiencies and high levels of nitrate in the effluent (Table 7.3, Test B). The relative N2O emissions did not change and were relatively high (6% of N load). However, the concentration of N2O in the liquid and in the gas phase decreased with a factor 2 (Table 7.3).

When pretreated sewage was fed to the reactor, the OLAND RBC was operated at lower nitrite concentration, while similar ammonium and nitrate concentrations were obtained (Table 7.3, Test C). The latter however did not result in lower N2O emission rates. When the feeding regime was changed to a more continuous-like operation (4 pulses h-1), the N2O emission increased significantly, while NO emission remained constant (Table 7.3, test D). Due to the lower ammonium removal efficiency (65 compared to 81%), but similar relative nitrite and nitrate accumulation rate, the total nitrogen removal efficiency decreased. When a nitrite pulse was added just after feeding, about 20 mg NO2--N L-1 was obtained in the reactor. This did increase the NO and N2O emissions significantly (p<0.05) compared to the same feeding pattern (Table 7.3, Test C-E). Although similar constant total nitrogen removal efficiencies were obtained during this operation, a significant (p<0.05) decrease in the relative nitrate production was observed. The latter was mainly caused by a global increase in AnAOB activity. In the last test (F), the influent ammonium concentration was doubled, leading to higher ammonium and also free ammonia concentrations (1 Âą 0.4 mg N L-1 compared to 0.1 Âą 0.4 mg N L-1). Due to overloading of the system, the total nitrogen removal 121

Cold OLAND on preteated sewage: feasibility demonstration on lab-scale

efficiency decreased. However, at these conditions a lower relative nitrate production was obtained, due to a decrease in NOB and increase in AnAOB activity (Table 7.3, Test F). Together with this, increased NO and N2O emissions were observed. As the influence of the nitrogen loading and DO concentration could be considered minor in this test range (Figure S7.2), these tests show a relation between increased NO emissions and decreased relative nitrate productions (Table 7.3).

When the activity during the feeding cycle was studied in more detail, it could be concluded that the highest nitrogen conversion rates took place during the feeding period (Fig. 7.3). As the HRT is only 1 hour, the reactor volume is exchanged in 20 minutes. During this phase, ammonium increased, while nitrite and nitrate concentrations decreased due to dilution (Fig. S7.3-S7.5). The NOB/AnAOB ratio was around 1, which means that NOB were able to take twice as much nitrite than AnAOB did, as the latter also consumed ammonium (Fig. 7.3). After the feeding period, a lag phase of the ammonium increase was observed, because the reactor liquid was not homogenously mixed yet. After mixing (10 minutes after feeding) was established, a N2O peak was reached during every test (Fig. S7.3-S7.5). At this point, during the reference period with pretreated sewage (Test C) total activity decreased and a very low NOB activity was observed (Fig. 7.3). Moreover, the NOB/AnAOB ratio decreased to 0.4 (Test C, Fig. 7.3), which means that during these conditions nitrite consumption by AnAOB was higher than nitrite consumption by NOB. The increased relative AnAOB activity was more pronounced when a higher NO and N2O peak were present (Test E). The latter was caused by an increased nitrite concentration in the reactor. When N2O concentration started to decrease again (last 20 minutes of feeding regime), nitrite consumption by NOB was again higher than the nitrite consumption by AnAOB (Fig. 7.3).

122

Chapter 7

Table 7.3: Operational parameters and nitrogen conversion rates during the 6 different RBC operations which differ from feeding composition and feeding regime (volume 2.5 L and 50 % immersion of the discs, day 307-309 for synthetic feed, days 424-431 for pretreated sewage). Reactor phase VII: synthetic XIII: pretreated sewage Test A° B C° D EF Additive NO2 NH4+ Feeding regime (pulses/h) 2 2 1 4 1 1 -1 -1 Total N loading rate (mg N L d ) 1169 585 1340 1554 1737 2718 Temperature water (°C) 15 ± 0.3 16 ± 0.2* 14 ± 0.4 15 ± 0.1* 16 ± 0.1* 15 ± 0.4 DO (mg O2 L-1) 2.9 ± 0.1 3.7 ± 0.6* 4.0 ± 0.1 3.2 ± 0.1* 3.3 ± 0.1* 3.2 ± 0.1* pH 7.6 ± 0.06 7.6 ± 0.05 7.6 ± 0.04 7.6 ± 0.01 7.6 ± 0.02 7.8 ± 0.02* Ammonium out (mg N L-1) 9±1 1.4 ± 1* 11 ± 3 19 ± 3* 12 ± 1 58 ± 4* -1 Nitrite out (mg N L ) 14 ± 2 13 ± 1 6±1 6 ± 0.4 18 ± 2* 9 ± 0.3* Nitrate out (mg N L-1) 17 ± 3 37 ± 6* 18 ± 2 16 ± 1* 18 ± 0.4 20 ± 0.4 + -1 -1 NH4 oxidation rate (mg N L d ) 895 ± 22 509 ± 2* 1051 ± 73 957 ± 89 1053 ± 16 1285 ± 93* Relative nitrite accumulation (%) 25 ± 3 20 ± 1* 14 ± 3 15 ± 1 8 ± 4* 15 ± 1 Relative nitrate production (%) 36 ± 8 76 ±6* 48 ± 1 47 ± 3 42 ± 2* 34 ± 3* Total efficiency (%) 38 ± 4 17 ± 4* 35 ± 3 28 ± 4* 32 ± 2 27 ± 4* AerAOB activity (mg NH4+-N L-1 d-1) 658 ± 88 469 ± 17* 827 ± 44 781 ± 57 795 ± 30 938 ± 46* NOB activity (mg NO2--N L-1 d-1) 174 ± 59 299 ± 28* 375 ± 38 342 ± 24* 362 ± 13 277 ± 18* AnAOB activity (mg Ntot L-1 d-1) 205 ± 38 49 ± 13* 234 ± 20 218 ± 29 263 ± 15* 354 ± 49* -1 N2O in liquid (μg N L ) 64 ± 46 30 ± 22* 78 ± 12 104 ± 29* 61 ± 13 74 ± 4 -1 NO emission (mg N d ) 0.53 ± 0.03 n.d. 0.66 ± 0.06 0.74 ± 0.08 1.65 ± 0.18* 0.82 ± 0.1* N2O emission (mg N d-1) 151 ± 28 93 ± 23* 170 ± 19 179 ± 6* 274 ± 37* 202 ± 18* % N2O emission on loading 5.1 ± 1.0 6.4 ± 1.6* 5.0 ± 0.6 4.5 ± 0.2* 6.2 ± 0.8* 3.0 ± 0.3* °Reference period for synthetic and pretreated sewage *Significant differences (p < 0.05) compared to reference period

123

Cold OLAND on preteated sewage: feasibility demonstration on lab-scale

Figure 7.3: Detailed NO/N2O monitoring during the reference test (Test C, Table 7.3) and when nitrite was pulsed (Test E, Table 7.3) and effect on AerAOB, AnAOB and NOB activity during the different phases of the feeding cycle. Significant differences in AerAOB, AnAOB, NOB and NO/N2O concentration compared to the reference period are indicated with *, °, “ and +, respectively.

4 4.1

Discussion Effect of temperature decrease

Average temperatures of sewage in west European region are around 17°C, with a minimum of 8°C and a maximum of 29°C (Mollen, personal communication). Therefore, the temperature of the OLAND RBC was decreased from 29°C to 15°C. In contrast to the optimal microbial balance at temperatures > 20°C, excess nitrite and nitrate formation was observed at lower temperatures. Improved operational conditions (O2 availability) resulted in similar nitrogen conversion rates for AerAOB and AnAOB at lower temperature (< 20°C) compared to the reference period at 29°C. The gradual adaptation of the nitrogen converting community to low temperatures probably attributed to the lower temperature dependence of AerAOB and AnAOB activity compared to the temperature shocks described in literature (Dosta et al., 2008; Guo et al., 2010). Similar long-term effect of temperature on AerAOB activity (Guo et 124

Chapter 7

al., 2010) and AnAOB activity (Hu et al., 2011; Hendrickx et al., 2012) were observed before. Due to the higher DO concentration at lower temperatures, the oxygen penetration depth possibly increased causing a decrease in AnAOB activity. On the other hand, higher oxygen inputs were needed at lower temperatures to obtain the same AerAOB activity as at high temperature. The combination of these two factors could have been responsible for the increased nitrite accumulation from phase IV onwards. Therefore, at lower temperature the OLAND performance will be limited by AerAOB activity as their activity guarantees anoxic zones in the biofilm (Vazquez-Padin et al., 2011).

Increased AerAOB activities were obtained at high DO levels (3 times higher than at 29째C). This was on one hand caused by a better solubility of oxygen at lower temperatures and on the other hand by a decrease of the immersion level from 78 to 55%. Although the changes in immersion level did not always resulted in a significant DO change (phase IV to V; phase V to VI), the oxygen availability through contact with the atmosphere was changed drastically. This suggests that oxygen transfer through atmospheric oxygen is more important in this system compared to transfer from dissolved oxygen. Although oxygen concentrations in OLAND systems at high temperature conditions are controlled at levels below 1 mg O2 L-1 to avoid nitrate oxidation by NOB, at low temperatures 2-4 mg O2 L-1 is needed to allow sufficient AerAOB activity (Vazquez-Padin et al., 2011). As nitrite accumulated in the OLAND RBC, oxygen input was probably too high to allow a balanced performance between nitrite production and consumption. Therefore, in practice a bulk DO control system is advisable to obtain a better removal efficiency.

4.2

Effect of COD/N increase

COD addition did not result in a better nitrogen removal efficiency or lower AnAOB activities (Lackner et al., 2008) as almost no COD removal was observed. Therefore, OLAND performance was not affected by COD/N ratios of 2 and stable nitrogen removal rates were maintained. It has been already successfully demonstrated that anammox can co-exist with heterotrophic denitrifiers at COD/N ratios of 2.2 (Desloover et al., 2011). Therefore, it should be possible to obtain high nitrogen removal efficiencies without the loss of AnAOB activities at mainstream conditions.

125

Cold OLAND on preteated sewage: feasibility demonstration on lab-scale

4.3

NOB-AnAOB competition at mainstream conditions

Although Nitrospira sp. were present from day 0-375 (phase I-X) at a stable level of around 40 copies ng-1 DNA, at temperatures below 20Â°C (day 61, phase IV) NOB activity increased significantly (Fig. S7.1). Together with the increased NOB activity, nitrite accumulated in the system. Moreover, when COD was added and lower nitrogen concentrations (55 mg N L-1 instead of 60 mg N L-1) were fed (phase X to XIII), relative nitrate productions up to 62% were observed (Table 7.2, phase XII). Free ammonia (FA) and free nitrous acid (FNA) concentrations (Table 7.1 and 7.2) were in all phases too low to suppress nitratation (Anthonisen et al., 1976). Moreover, oxygen inputs (mainly through atmospheric contact) were rather high to allow sufficient nitritation, which could also have stimulated NOB growth and activity. Therefore, for mainstream treatment other strategies beside FA, FNA and oxygen limitation should be applied to suppress nitratation. Detailed nitrogen balance tests showed that the different feeding strategies did not affect the microbial balance in the system (Table 7.3). However, pulse feeding allowed higher AerAOB activities increasing the ammonium removal efficiency (phase XIII) and continuous-like feeding resulted in a decreased ammonium and as a consequence total nitrogen removal efficiency (Table 7.3). Moreover, a sufficient loading rate was needed to allow a good microbial balance and thus increased AnAOB activity and/or decreased NOB activity (Table 7.2 Phase XIII; Table 7.3 test F). Overloading of the system and therefore obtaining higher FA levels, could inhibit the NOB activity (Table 7.3 test F) in contrast to the long-term performance at lower FA concentrations. Therefore, the latter could not be responsible for the better microbial balance during reactor operation. However, nitrite accumulation, resulting in higher peaks in NO and N2O production (Table 7.3) occurred in all well performing periods at low temperature. High NO emissions, initiated by addition of nitrite could increase relative AnAOB activity and decrease NOB activity (Fig. 7.3). It is well known that NO is toxic to most of the bacteria (Mancinelli and McKay, 1983). It has been described before that NO2--dependent O2 uptake by NOB could reversibly be inhibited by NO at concentrations of 7-448 Îźg NO-N L-1 (Starkenburg et al., 2008). In contrast, NO is an intermediate for the AnAOB metabolism and high NO concentrations do not affect their activity (Kartal et al., 2010b). Therefore, at conditions of high NO concentration, AnAOB can have a competitive advantage compared to NOB. However, from the moment NO is depleted NOB activity can increase again (Fig. S7.3S7.5; Starkenburg et al. 2008). Although nitritation is stimulated by NO (Zart et al., 2000), it seemed that also AerAOB activity was affected by NO at the NO/N2O peak (Fig. 7.3). 126

Chapter 7

Therefore, a balance should be found between stimulating AnAOB above NOB activity and allowing sufficient nitrite production by AerAOB. The high volumetric loading rate applied, together with the pulse feeding and the nitrite accumulation led to high NO/N2O emissions compared to mesophilic OLAND applications (Kampschreur et al., 2009a; Weissenbacher et al., 2010). This could however be a prerequisite for obtaining low nitratation levels at these mainstream conditions.

4.4

OLAND application in the main line

At 15°C and a COD/N ratio of 2, high total nitrogen removal rates of 0.5 g N L-1 d-1 were obtained. However, the total nitrogen removal efficiency was too low to obtain dischargeable effluent (European Commision, 1991). As similar total nitrogen removal rates were obtained at 15°C compared to 29°C, the performance was not limited by the mainstream conditions but it was limited by the reactor configuration. Because the discs only had a spacing of 3 mm, regular perforations of the biofilm were needed to allow sufficient diffusion. A better RBC configuration (higher disc distance) or another reactor technology (suspended growth system) could probably allow higher efficiencies due to more efficient diffusion. On the other hand, by a combination of OLAND and conventional nitrification/denitrification in the B-step, a better removal efficiency could be obtained. This can be achieved by only partly replacing the activated sludge by OLAND biomass. To allow AnAOB retention in this system a selectively higher SRT of the OLAND biomass compared to the activated sludge should be maintained. For granules, the latter can be obtained by implementation of cyclones (Wett et al., 2010b), but it should also be possible by inoculation of OLAND biomass on packing material which can be kept in the system by a grid. In this way, OLAND should not be responsible for the total nitrogen removal efficiency of the system and nitrite or nitrate formation can be compensated by denitrification.

Conclusions

This study showed for the first time that total nitrogen removal rates of 0.5 g N L-1 d-1 can be maintained when decreasing the temperature from 29°C at 15°C and when low nitrogen concentration and moderate COD levels are treated. Nitrite accumulation together with elevated NO/N2O emissions was needed to allow competition of AnAOB against NOB for nitrite at low FA, low FNA and high DO levels. Further research should elucidate the mechanism and the level of NO/N2O emission needed to obtain a balanced performance. 127

Cold OLAND on preteated sewage: feasibility demonstration on lab-scale

Moreover, it should be further evaluated if the increased NO/N2O emission can be compensated with a decreased energy consumption to justify OLAND mainstream treatment.

Acknowledgements

H.D.C. received a PhD grant from the Institute for the Promotion of Innovation by Science and Technology in Flanders (IWT-Vlaanderen, SB-81068) and S.E.V. was supported as a postdoctoral fellow from the Research Foundation Flanders (FWO-Vlaanderen). The authors gratefully thank Aquafin for providing the sewage, Eva Spieck for providing the qPCR standards and Tom Hennebel, Joachim Desloover and Simon De Corte for inspiring scientific discussions.

Supplementary data Table S7.1: Overview of the presence of NOB and AnAOB species, confirmed with FISH

High temperature,

Low temperature,

COD/N of 0 (day 1) COD/N of 2 (day 435) Nitrobacter

Nitrospira

Cand. Scalindua

Cand. Kuenenia & Brocadia

128

Chapter 7

Figure S7.1: Abundance of functional Amo gene copies of AerAOB and 16SrRNA copies of AnAOB and Nitrospira sp. measured by qPCRin the OLAND RBC.

129

Cold OLAND on preteated sewage: feasibility demonstration on lab-scale

Figure S7.2: Scatter plot showing the influence of the nitrogen load (A) and dissolved oxygen concentration (B) on the AerAOB, AnAOB and NOB activity. The zone between the dashed lines represents the nitrogen load and DO range studies in the N balance test (Table 7.3)

130

Chapter 7

Figure S7.3: Phase VII, Test B (Table 7.3): Profiles of ammonium, nitrite, nitrate and oxygen concentration in the reactor water phase and N2O concentration in the defined air flow out of the reactor. Gray boxes mark the feeding periods.

Figure S7.4: Phase XIII, Test C (Table 7.3): Profiles of ammonium, nitrite, nitrate and oxygen concentration in the reactor water phase and N2O and NO concentration in the defined air flow out of the reactor. Gray boxes mark the feeding periods.

131

Cold OLAND on preteated sewage: feasibility demonstration on lab-scale

Figure S7.5: Phase XIII, Test E (Table 7.3): Profiles of ammonium, nitrite, nitrate and oxygen concentration in the reactor water phase and N2O and NO concentration in the defined air flow out of the reactor when extra nitrite pulses were added just after the feeding period. Gray boxes mark the feeding periods.

132

Chapter 8

Chapter 8: Environmental assessment of onestage partial nitritation/anammox implementation in sewage treatment plants Abstract Implementation of one-stage nitritation/anammox (e.g. DEMON速, OLAND) for the treatment of sludge digestates allowed energy autarky in the wastewater treatment plant (WWTP) in Strass (Austria). To further increase the overall energy production of the plant a first trial was performed to implement mainstream DEMON operation. To evaluate the environmental impact of DEMON implementation, life cycle assessment (LCA) was carried out based on on-site measurement campaigns for three scenarios: (1) without DEMON; (2) with DEMON in the side line; (3) with DEMON in the side and main lines. The results of these assessments showed that DEMON implementation in the side line, had a positive effect on the eutrophication potential, abiotic depletion potential and global warming potential. For the present situation with mainstream DEMON, 9% of the electrical needs could be saved at the moment, but control optimization is needed to decrease N2O emissions on the long-term. From the LCA, it could be concluded that the WWTP in Strass can be seen as one of the benchmark plants, not only based on energy efficiency with 153 and 167% energy coverage, but also based on overall environmental sustainability with a CO2 footprint of 7 and 36 kg CO2 PE-1 year-1 for side and mainstream DEMON, respectively.

Chapter redrafted after: De Clippeleir H., Schaubroeck S., Weisssenbacher N., Dewulf J., Boeckx P., Boon N. and Wett B. Environmental assessment of one-stage nitritation/anammox implementation in sewage treatment plants. Submitted. 133

Environmental assessment of the implementation of OLAND in sewage treatment plants

Introduction

Around 24000 sewage treatment plants (WWTP) are operational in Europe which together treat about 580 million person equivalents (PE) (UWWTD, 2011). Although the main aim of the WWTP is to decrease harmful emissions towards water bodies, recently more attention is paid on energy efficiency and overall environmental sustainability (Verstraete and Vlaeminck, 2011). The implementation of anaerobic digestion, which provides on-site production of renewable energy increased significantly over the last 5 years (Chapter 4). The latter has the potential to be implemented in around 85% of the WWTP in Europe as from a size of around 10 000 PE anaerobic digestion starts to be economically feasible (UWWTD, 2011). Depending on the primary sludge production, performance of the anaerobic digester, the efficiency of electrical energy production from biogas and the oxygen transfer efficiency for aeration (Wett et al., 2007; Nowak et al., 2011), energy self-sufficiency can be reached.

Besides energy recovery through anaerobic digestion, energy minimization for nutrient removal by implementation of one-stage partial nitritation/anammox, also known as DEMON (Wett, 2006), OLAND (Kuai and Verstraete, 1998) and CANON (Third et al., 2001) could positively contribute to the energy balance of the plant. At the moment around 30 one-stage autotrophic nitrogen removal plants are operational for the treatment of sludge liquors from anaerobic digestion, which account for 15-25% of the total nitrogen load of the WWTP (Vlaeminck et al., 2012). High and stable nitrogen removal rates for the treatment of sludge digestates are reported (Vlaeminck et al., 2012). Moreover, the implementation of DEMON in the side line of the WWTP, can result in a total decrease in energy consumption of the WWTP with more than 50% compared to conventional nitrification/denitrification (Siegrist et al., 2008).

Further decreasing the energy consumption by up-grading the activated sludge step in the main line by DEMON would offer two main advantages. The first advantage would be that the aeration need for nitrogen removal in the main wastewater line can decrease by almost 60% as DEMON only requires 1.8 kg O2 kg-1 N removed and conventional nitrification/denitrification requires 4.3 kg O2 kg-1 N removed. The second advantage of mainstream DEMON is that this process allows higher COD removal through primary sludge production, for example by a highly-loaded activated sludge step, as no carbon is needed in the DEMON process 134

obtain high (89%) nitrogen removal

contrast

Chapter 8

nitrification/denitrification which needs 3 kg COD kg-1 N. Theoretically, it was estimated that mainstream DEMON could allow energy-positive wastewater treatment (Siegrist et al., 2008; Verstraete and Vlaeminck, 2011; Chapter 4).

The challenges of mainstream DEMON (Vlaeminck et al., 2012) on the other hand, are the retention of the anoxic ammonium-oxidizing bacteria (AnAOB) in the activated system (sludge retention time (SRT) of around 10 days), as AnAOB have a doubling time of around 1-2 weeks. Moreover, nitratation suppression in the mainstream will be more challenging because of the lower selection pressure on nitrite-oxidizing bacteria compared to side stream conditions (Chapter 4). DEMON lowers the energy demand but, on the other hand emits on average around 1% of the N load as N2O-N, a powerful greenhouse gas (Joss et al., 2009; Kampschreur et al., 2009a; Weissenbacher et al., 2010). Therefore, greenhouse gas emission at mainstream conditions should be evaluated as they can offset energy efficiency and performance.

For a better environmental sustainability assessment, the environmental impact should not only be considered on a process and plant level, but also from a more broader life cycle perspective. Life cycle assessment (LCA) is an appropriate tool. It is a holistic tool increasingly used to evaluate environmental impacts associated with a product, process or activity (Iso, 2006a, b). LCA has been widely used to study WWTP configuration (Clauwaert et al., 2010; Hospido et al., 2005; Hospido et al., 2008; Foley et al., 2010). These studies showed that operational energy, direct greenhouse gas emissions and chemical consumption generally increase with increasing nitrogen removal (Foley et al., 2010). However, environmental impact assessment of the implementation of DEMON in WWTP has not been evaluated before. Since a strong point of DEMON is its higher environmental sustainability on process level due to its lower oxygen demand and its potential to allow higher energy recovery in other steps of the WWTP, this study evaluated this on a plant and life cycle level for different scenarios of one WWTP in Strass (Austria). The latter WWTP is based on a twostage activated sludge system (A/B system; Wett et al., 2007) as mainstream treatment and on sludge digestion with electricity and heat production via a combined heat and power (CHP) unit in the side line. Sludge digestate treatment (sidestream treatment) was in the reference period (2003, scenario 1) based on nitritation/denitritation with A-stage sludge as external carbon source, but has been replaced by DEMON (from 2004 onwards, scenario 2). Preliminary results of DEMON implementation in the mainstream, as an up-grader for the B 135

Environmental assessment of the implementation of OLAND in sewage treatment plants

stage, were also incorporated in this study (scenario 3) to elucidate the critical factors for final implementation important to obtain a higher environmental sustainability.

2 2.1

Materials and methods Scope definition

In this research, three scenarios applied on the WWTP of Strass (Austria) were studied and compared using the LCA framework according to the ISO 14040/14044 guidelines (Iso, 2006a, b). The three scenarios had a different degree of DEMON implementation level (from none, to side stream and further to mainstream). The system boundaries were the same for all scenarios. In general, the considered life cycle system included the WWTP itself, the part of the human industrial system responsible for products (mainly chemicals and electricity) needed in the WWTP and the part for the further processing of its waste products (composting of the dewatered digestate). The foreground system, main system of interest, consisted of the WWTP. The other parts of the life cycle were considered as the background system. The transportation of chemicals from the supplier to the WWTP, the transportation of the dewatered digestate to the composting facility and the infrastructure of the WWTP and composting facility were excluded from the life cycle systems. The infrastructure is left out since the focus was on the real-time operation of the WWTP. Not included in the life cycles were the upstream collection and transportation of the municipal wastewater to the WWTP and the usage/disposal on land of the compost processed out of the digestate.

The main system input was wastewater, which was considered as a waste product in LCA and therefore no environmental impact of its generation was allocated to it (Iso, 2006b). Likewise, the addition of co-substrate to the digester, which consisted out of kitchen waste and fat, was considered in this study as a waste product. The organic carbon present in the wastewater was assumed to be 100% biogenic, neglecting the amount of fossil carbon from detergents and soaps (Griffith et al., 2009). CO2 emissions from the oxidation of this biogenic carbon were by consequence considered as biogenic in compliance with the Intergovernmental Panel on Climate Change (IPCC) accounting guidelines (Doorn et al., 2006). Two products were formed in the studied life cycles namely electricity and compost. The conventional production of these products was avoided and thus also the total impacts of their production processes. Their impacts should therefore be subtracted from the total impact of the life cycle (Finnveden et al., 2009). Electricity produced at the plant displaced electricity provision by 136

Chapter 8

the grid and thus electricity production in Austria. The substitutability of the compost, which was intended for agricultural application, was assumed to be 50% for nitrogen and 70% for phosphorus (Bengtsson et al., 1997). Concerning industrial products avoided, fertilizer mixes with a similar nutrient composition were chosen (see data inventory). The functional unit (FU) was the treatment of 1 m3 of sewage. The effluents of the different scenarios (Table 8.1) were all in compliance with Austrian legislation regarding necessary water quality (BGBL, 1996). Additionally, the waste sludge after dewatering complied with the legal guidelines in terms of composition, especially the presence and quantity of heavy metals for agricultural application (BGBL, 1996).

2.2

Plant description

The plant in Strass (Austria) is based on a two-stage activated sludge system in the main water line, referred to as an A/B plant (A/B Verfahren, Wett et al., 2007). The first step is a high rate activated sludge step with short hydraulic (30 minutes) and sludge (0.5 days) retention time. About 50-60% of the COD is removed in the A-stage and due to the short retention time the organics are adsorbed or incorporated in the sludge and not emitted as CO2. Nitrogen and phosphorous removal in this step mainly occurs via the organic fraction and only accounts for on average 23 and 26%, respectively. The second activated sludge step (B-step) is a low loaded step with temperature dependent aerobic SRT of ca. 10 days. This step consists of a predenitrification and nitrification step with recycle from the second to the first step. The aeration is controlled by a combination of dissolved oxygen (DO) and NH4+ measurement to obtain optimal effluent quality without excess aeration. Sludge from the mainline (A and B-stage) is send to the digester. Organic co-substrate, which mainly consisted out of kitchen waste, was added to enhance the electrical energy recovery. Sludge waste after dewatering of the digestate is composted in an external facility. The liquid fraction from the digestate after filtration was initially treated in a separate nitritation denitritation reactor (Fig. 8.1A), but was currently replaced by a DEMON system (scenario 2, Fig. 8.1B). DEMON was further implemented in the B-stage by inoculation with DEMON granules and implementation of cyclones in the sludge recycle to maintain the DEMON granules in the system (scenario 3). The hydraulics and aeration control system was not changed (Fig. 8.1C).

137

Environmental assessment of the implementation of OLAND in sewage treatment plants Table 8.1: Monthly averages of the parameters of the inventory data for the different scenarios studied. All data are presented in function of 1 m3 sewage treated. N/DN: nitrification/denitrification; N/DN*: nitritation/denitritation; DEMON: deammonification; DM: dry matter Scenario 1a Scenario 1b Scenario 2 Scenario 3 Mainstream treatment

N/DN

DEMON

Sidestream treatment

N/DN*

DEMON

COD (g)

666

643

526

NH4 -N (g)

Organic N (g)

PO43- -P (g)

338

239

Electricity from the grid (Wh)

86.49

1.55

0.829

Sodium aluminate (g)

78.4

46.9

35.3

9.68

7.29

FeCl2 (g)

10.9

9.56

4.33

Polymer (g)

1.68

1.92

1.63

140/258

61/111

76/139

7.92

4.06

3.76

P fertilizer mix (g P)*

3.40

2.66

N fertilizer mix (g N)*

1.30

Inputs to foreground system Waste Water

Co-substrate (g DM) Products (External resources)

Flocculant (g)

Avoided products (resources) C fertilizer mix (g peat/g straw) N&P fertilizer mix (g P)*

138

Chapter 8 Electricity into the grid (Wh)

179

209

COD (g)

NH4+-N (g)

NO2--N (g)

0.13

NO3 -N (g)

N org â&#x20AC;&#x201C;N (g)

0.87

PO43- -P (g)

0.710

0.27

0.39

0.668

0.658

0.249

Emissions to water

Emissions to air CH4 (g) N2O (g)

0.325

1.59

0.520

2.45

NO (g)

0.00516

0.0318

0.0158

0.0134

NO2 (g)

1.59

1.269

1.275

CO (g)

0.659

0.799

0.803

357

547

518

0.122

0.0733

0.0737

CO2-biogenic (g) SO2 (g)

* the specific composition of the fertilizer mix can be found in Table S8.1

139

Environmental assessment of the implementation of OLAND in sewage treatment plants

Figure 8.1: Schematic overview of the 3 scenarios studied. Scenario 1 included a nitritation/denitritation (N/DN*) in the side line (A), while scenario 2 and 3 had a DEMON reactor for digestate treatment (B, C, respectively). Moreover, in scenario 3, the low loaded activated sludge step (B-step) is upgraded to a DEMON step. AD: anaerobic digestion; CHP: combined heat and power; AS: activated sludge 140

Chapter 8

2.3

Data inventory

Foreground data were collected directly from the WWTP itself. Operational data (water/sludge flows, water/sludge composition, energy demands, chemical usage etc) for scenarios 1, 2 and 3 were collected from daily logging results of April to July 2003, April 2011 and April 2012, respectively. Greenhouse gas emissions were measured on-site at all biological treatment steps during April 2011, April-May 2012 for scenario 2 and 3, respectively, following the method described for N2O emission measurements by Desloover et al. (2011a) and for NO and NO2 measurements by Weissenbacher et al. (2010). Additional gas measurement campaigns during July 2011 and November 2011, confirmed the emission factors used in the different scenarios. For scenario 1, similar emissions for the A and B-step in the mainline compared to the emissions measured for scenario 2 were considered. For the side line treatment in scenario 1, two different emission estimations were proposed: (a) the same emission as measured for a DEMON reactor, (b) an increased N2O and NO emission (factor 5) because of the high nitrite concentrations (Desloover et al., 2011a). An overview of all the in- and out-coming flows of the foreground system is given in Table 8.1.

Data for the processes of the background system were retrieved from the ecoinvent v2.2 database (Swiss Centre for Life Cycle, 2010), unless mentioned otherwise. The electricity production mix of Austria (last update: June 2010) originated from burning of fossil fuels (≈ 7%), nuclear energy (≈5%) and renewable energy (≈5 %) of which most part is hydropower generated in Austria itself (≈9 %). Most chemical products added to the WWTP were not present as such in the ecoinvent database. Their life cycle data and eventually their impact calculated in the impact assessment phase were replaced by those of other products available in the ecoinvent database, namely similar products or individual reagents needed for their production. Sodium aluminate (NaAl(OH)4) added in high quantity was substituted by stoichiometric quantities of its conventional reagents: sodium hydroxide (NaOH) and aluminium hydroxide (Al(OH)3). The iron(II)chloride (FeCl2) (32%) solution was replaced by iron(III)chloride (FeCl3) (40%). ‘S dflock K2’, a flocculent containing FeCl3 (3.00%) and aluminiumchloride (AlCl3) (9.50%), was replaced by a solution of only FeCl3 (12.50%). ‘Zetag’ is a polyacrylamide and it is a polymer of acrylamide, which is in its turn formed by hydratation of acrylonitril. ‘Zetag’ was substituted by acrylonitril. The amount of ‘Flockungsmittel M se K222L’ added was neglected. The composting process was custom made. For 1 ton dry matter (DM) of biosolids input a fixed assumed amount of 75.73 Wh 141

Environmental assessment of the implementation of OLAND in sewage treatment plants

electricity consumption and emissions of 163.3 g N2O, 63.90 g NH3, 1278 g CH4 were set for all scenarios based on figures of the direct composting of biodegradable municipal waste (Van Ewijk, 2008). It was assumed that 14% of carbon was removed during the composting process (personal communication Steven De Meester, May 2012). Out of this carbon removal, the amount of biogenic CO2 emissions per ton DM (scenario 1: 158 kg; scenario 2: 72 kg; scenario 3: 123 kg) and residual carbon (86% of original C) left in the compost were calculated. The final compost consisted of this residual carbon and the same amount as that of the dewatered digestates for the other nutrients, neglecting the small amounts of these removed by gaseous emission and via the leachate during the composting process. The compost of each scenario was displaced by a fertilizer mix using the substitutability factor of 50 and 70% for N and P, respectively (Bengtsson et al., 1997). This was specifically done by defining a mix with the representative C, N and P content as the displaced compost, based on the methodology described by Hermann et al. (2011). For the amount of carbon present in the compost, an amount of peat and straw in a ratio of 1:3 were calculated to replace the amount of humus carbon (51% of its carbon content) present in the compost (Hermann et al., 2011). The humus carbon present in peat and straw are 0.077 and 0.084 kg kg-1 fresh matter (FM), respectively (Hermann et al., 2011). Out of these figures, FM quantities of peat and straw were calculated. For the other nutrients, only nitrogen (N) and phosphorous (P) were taken into account. The amounts of P and N in the carbon fertilizers (Phyllis-Database, 2012), peat (P: 0.1%; N: 1%) and straw (P: 0.091%; N: 0.71%) were substracted from that in the compost, leading to the amount needed to be displaced by other fertilizers. Different fertilizers are on the market for these two nutrients. Out of the 2010 consumption figures of such fertilizers in Western and Central Europe (International Fertilizer Industry, 2012). P&N, N and P fertilizer mixes were constructed (Table S8.1). For the individual fertilizers, ecoinvent data was available. To determine the fertilizer mix for each scenario, first, an amount of P&N fertilizer mix was calculated. Thereafter, an extra quantity of N or P fertilizer mix was quantified for the leftover P or N, respectively, not covered by the P&N fertilizer mix.

2.4

Impact assessment

The impact assessment was done using Simapro version 7.2 software (with ecoinvent database version 2.2) and the selected method was the ‘CML 2001 method (all impact categories)’ version 2.05, normalized for West-Europe 1995, hereafter referred to as the ‘CML 2001’ method. A CML method was selected since these are most commonly applied in other LCAs of WWTPs (Hospido et al., 2008; Clauwaert et al., 2010; Foley et al., 2010). This 142

Chapter 8

method was developed by the Center of Environmental Science of Leiden University (CML), the Netherlands. The version in the Simapro software was based on the spreadsheet version 3.2 (December 2007) published on the CML web site (http://www.cml.leiden.edu/). Important to notice is that biogenic CO2 uptake and emissions were not accounted for in the global warming potential (GWP) 100a (100 years) category. Resources are accounted for using abiotic depletion. In the impact assessment, the emissions of the foreground system (the WWTP) were selected to end up in a low populated area.

3 3.1

Results and discussion Impact of nitrogen removal process on process level

The conventional process for nitrogen removal is nitrification/denitrification (N/DN) and was applied in the main line of the WWTP of Strass (B-stage). COD/N ratios send to the B-stage were 7.2, 8.2 and 9.2 for scenario 1, 2 and 3, respectively and therefore allowed full denitrification. A cost-saving alternative for the conventional nitrification/denitrification is the application of nitritation/denitritation, saving theoretically around 24% of aeration requirement as during this process nitrite oxidation (nitratation) is avoided (Vlaeminck et al., 2012). Moreover, the COD demand and sludge production can decrease with 50% and 40%, respectively (Vlaeminck et al., 2012). This process was applied for digestate treatment in the WWTP during scenario 1 (Fig. 8.1). Compared to the energy requirements for nitrification/denitrification in the mainstream of the WWTP in Strass, assuming that 50% of the available COD was denitrified and an oxygen transfer efficiency of 2 kg O2 kWh-1 was applicable, the nitritation/denitritation applied in the side line decreased the energy requirements from 4.30 to 2.65 kWh kg-1 N removed. So, this process has a potential to save around 38% of the energy needed for aeration during nitrogen removal. DEMON implementation in the side line of the WWTP could further decrease the energy requirement for nitrogen removal to 1.52 kWh kg-1 N removed for scenario 2. This additional 43% decrease in energy requirement for aeration is exactly what is theoretically expected for DEMON implementation compared to nitritation/denitritation (Vlaeminck et al., 2012).

Besides energy savings for aeration, the choice of nitrogen removal process has an influence on the energy recovery potential as nitritation/denitritation and DEMON save 52 and 100% of COD source needed compared to conventional nitrification/denitrification (Vlaeminck et al., 2012). Therefore, these processes can increase electricity production at the plant through 143

Environmental assessment of the implementation of OLAND in sewage treatment plants

anaerobic digestion of primary sludge, which was otherwise used as a COD source (scenario 1).

The choice of nitrogen removal process also determines the sludge production. Conventional nitrification/denitrification compared to nitritation/denitritation and DEMON produces around 1 compared to 0.6 and 0.1 kg sludge kg-1 N removed, respectively (Vlaeminck et al., 2012). The latter is the result of the avoidance of heterotrophic growth in the system. DEMON implementation can therefore lower production sludge and as a consequence emissions, chemicals, electricity and waste products related to sludge handling.

At this moment insufficient comparative data are available for N2O emissions in nitrogen removing processes, making it hard to evaluate the impact difference of these processes. Reported N2O emission in activated sludge systems based on nitrification/denitrification ranged from 0.001-25% of the N load and were mainly linked with nitritation activity (Desloover et al., 2011b). Moreover, the N2O emission is mainly linked with the operational conditions rather than the process itself (Kampschreur et al., 2009b; Chandran et al., 2011). For the treatment plant of Strass, N2O-N emissions in the nitrification/denitrification step (Bstage) were very low, e.g. 0.01% of the N load compared to 1.0-1.3% of the N load measured during side stream DEMON (scenario 2 and 3). As nitrite, a precursor for N2O production (Kampschreur et al., 2009b; Chandran et al., 2011), accumulated up to 96 mg N L-1 during nitritation/denitritation compared to 0 mg N L-1 and 1 mg N L-1 in the nitrification/ denitrification and DEMON step, respectively, increased levels of N2O were expected for nitritation/denitritation (scenario 1b). Due to possible nitrite accumulation in processes based on nitritation, an increased level of N2O emission in the mainline was therefore expected when DEMON was implemented in the B-stage initially based on nitrification/denitrification.

From the comparison of the different nitrogen removal processes, it could be concluded that DEMON can lower the electricity needs, the sludge production and the COD demand, but has the potential to increase N2O emission due the higher risk for accumulation of nitrite.

144

Chapter 8

3.2

From energy-negative to energy-positive WWTP on system level

On the WWTP level, scenario 2 including the implementation of DEMON in the side line caused a 24% decrease in total energy consumption from 0.45 to 0.34 kWh m-3 sewage treated (scenario 1 compared to 2). Although DEMON implementation could decrease the energy needs for nitrogen removal with 43% in the side line (scenario 2), the relative energy requirement of the DEMON reactor remained around 0.02 kWh m-3 raw sewage due to the low contribution of side line treatment on the total energy consumption (4%). The latter was also the result of the higher kitchen waste dosage, which increased the nitrogen load to the DEMON reactor in comparison to scenario 1. Assuming relative on the incoming sewage, the same nitrogen load to the DEMON reactor as in scenario 1, 0.01 kWh m-3 sewage could potentially be directly saved by the implementation of DEMON due to lower aeration requirements in the side line. Moreover, implementation of DEMON allowed a higher sludge load to the digester (0.14 kg TS m-3 sewage) as this process had no need for a carbon source. Therefore, DEMON implementation could directly increase the energy recovery and thus the electrical energy production with 0.06 kWh m-3 sewage (13% of energy requirement in scenario 1) without taking the co-substrate addition into account. DEMON implementation could result in an electrical energy production on-site of 93% of the needs instead of 80% in scenario 1. This increased electrical energy recovery due to DEMON implementation could have been at least a factor 2 higher when compared to systems, which send the digestate directly to the B-stage (Siegrist et al., 2008; Chapter 4). However, the observed energy consumption decrease of 0.11 kWh m-3 sewage in scenario 2 compared to scenario 1 was also attributed to further optimizations in the A-stage, B-stage and sludge handling which allowed an energy consumption decrease of 0.02, 0.04 and 0.05 kWh m-3 sewage treated, respectively. Due to an overall better energy efficiency both in the side line but as a consequence also in the A- and B-stage, the electrical input of 0.087 kWh m-3 sewage of scenario 1 was avoided, even without considering the higher biogas production by the increased sludge load and addition of co-substrate to the digester during scenario 2. Therefore, DEMON implementation together with energy optimizations in the other steps led to an energy self-sufficient system (Wett et al., 2007). The addition of co-substrate increased the energy net production further to 153% of the energy demand of the plant (scenario 2), instead of 109% assuming that the biogas production remained constant.

145

Environmental assessment of the implementation of OLAND in sewage treatment plants

To further increase the energy efficiency of the plant, the potential of DEMON implementation in the main water line (scenario 3) was tested at full-scale as the energy needs for aeration in the B-stage accounts for 40% of the total energy needs of the WWTP (scenario 1 and 2). DEMON granules from the side stream reactor were inoculated in the B-stage and the SRT of the DEMON granules was increased compared to the activated sludge flocs by the installation of cyclones in the sludge recycle (Wett et al., 2012). Although the operation of the B-stage remained the same in terms of hydraulics, oxygen profile and loading, a metabolic shift was observed characterized by a significant decrease in nitrate concentration and increase in nitrite concentration in the reactor. Especially in winter when the highest loading rates were supplied, nitrite concentrations were higher than nitrate concentrations in the effluent (Wett et al., 2012). A NO3--N over NO2--N ratio of 2 was observed in the effluent of scenario 3 (April 2012), compared to a ratio of 31 in scenario 2 (Table 8.1). The latter indicated a decrease in the nitrite oxidation activity (nitratation), which was the first prerequisite to allow anammox activity. Due to the higher COD/N ratio, the energy savings by nitrogen removal through anammox were counteracted by the increased aerobic COD removal in the B stage. It could be calculated that compared to nitrification/denitrification (scenario 2), the energy demand for DEMON (0.9 kWh kg-1 N) increased with 1.4 kWh kg-1 N due to the aerobic removal of the COD which was normally denitrified and with 0.5 kWh kg-1 N due to the higher incoming COD/N ratio. As a consequence, the expected lower oxygen demand and thus energy demand observed in the B-stage was only minor i.e. 0.12 instead of 0.14 kWh m-3 sewage. It is suggested that increasing the COD removal in the A-stage through primary sludge production and thus decreasing the COD/N ratios of the B-stage influent could increase the role of the anammox bacteria in the mainstream and would therefore allow higher energy saving in the B-stage and higher energy recovery by the plant itself. During scenario 3, electricity production remained constant, but the total energy consumption decreased from 0.34 to 0.31 kWh m-3 sewage, due to the poorly working A-stage (Table 8.2) and a decrease in the energy demand in the B-stage. Therefore, the electricity production increased to 167% of the electrical energy needs of the plant, showing the potential of scenario 3 to allow higher energy recovery.

146

Chapter 8 Table 8.2: Overview of the performance for the different step of the WWTP. AD: anaerobic digestion; TS: total solids; tot: total Scenario 1 Scenario 2 Scenario 3 Mainstream treatment

N/DN

DEMON

Sidestream treatment

N/DN*

DEMON

COD removal (%)

Ntot removal (%)

P removal (%)

COD removal (%)

Ntot removal (%)

P removal (%)

Co-substrate addition (% of AD feed)

TS to biogas (%)

0.377

0.283

0.357

COD removal (%)

85째

Ntot removal (%)

High loaded AS

Low loaded AS

Sludge digestion

-1

Biogas yield (m kg TS input) Reject water treatment

째primary sludge was added as carbon source

3.3

Environmental impact of DEMON implementation on life cycle level

3.3.1 Eutrophication The primary objective of a WWTP is to decrease COD, N and P concentrations in the water phase and to obtain dischargeable effluent qualities. In all scenarios the eutrophication potential (EUP) of the wastewater (0.05 kg PO43--eq m-3 sewage) decreased sharply with 91, 94 and 93% for scenario 1, 2 and 3, respectively, by implementation of the sewage treatment system as expected (Fig. 8.2A). The best effluent quality was obtained during scenario 2. The higher ammonium, phosphorus and COD effluent concentrations (Table 8.1) caused the increased EUP of the WWTP during scenario 3. However, the lower effluent nitrate concentration during scenario 3 resulted in a more equally distribution of the EUP over the different effluent compounds (Fig. 8.3). A 60% decrease of the ammonium effluent concentration to 1 mg N L-1 by a more stringent control of the aeration system would result in the same EUP as during scenario 2. It is therefore expected that optimization of the operational conditions (DO and ammonium set point) will limit the EUP. 147

Environmental assessment of the implementation of OLAND in sewage treatment plants

Figure 8.2: Contribution of the different elements of the LCA system to (A) the eutrophication potential, (B) the abiotic depletion potential and (C) the global warming potential for the different scenarios studied. The wastewater itself had an eutrophication potential of 0.05 kg PO43--eq m-3 sewage. As the GWP of the WWTP was dominated by N2O (>90%), two different N2O emission scenarios for scenario 1 were included. Negative values are related to impacts, which are avoided by recovery of products on-site.

148

Chapter 8

Figure 8.3: Contribution of the different substances emitted by the WWTP to eutrophication potential (kg PO43--eq m-3 sewage)

3.3.2 Abiotic depletion potential The abiotic depletion potential (ADP) takes into account the use of abiotic resources such as iron ore and crude oil (Guinee, 2001). Energy consumption and chemical addition are the main factors of the described system, which relies on the use of abiotic resources (Fig. 8.2B). The difference between the scenarios (1>2>3) in the use of sodium aluminate, needed for P removal, dominated the ADP of chemical addition (Table 8.1. Fig. 8.2B). However, it should be noted that the sodium aluminate used at this plant was considered as a product and not as a waste, although sodium aluminate was retrieved from the alum industry nearby. Also, the recovery of nitrogen and more importantly phosphorus through composting significantly influenced the ADP of the WWTP. The resource intensive processes of the production of the fertilizers are the main cause of this (Silva and Kulay, 2003). Therefore, this indicates that P and N recovery from a life cycle perspective can be an important factor in decreasing the resource needs and counteracting the need for chemicals and electricity. Besides nutrient recovery, also electricity production (scenario 2 and 3) further decreased the ADP. So, it can be concluded that for this impact category the implementation of DEMON in the side (scenario 2) and mainstream (scenario 3) is advantageous, because it allows a higher net electricity production. It should also be noted that depending of the electricity mix used, the effect of electricity production on-site can increase with a factor 2 depending on the country (Fig. S8.1). For example, energy recovery in countries that produce electricity from hard coal or natural gas (i.e. Netherlands, Poland, USA) will have a bigger effect on the abiotic

149

Environmental assessment of the implementation of OLAND in sewage treatment plants

depletion potential compared to countries based on hydropower or nuclear power (i.e. Norway, Austria, France, Belgium) based on the ecoinvent v2.2 database.

3.3.3 Global warming potential Global warming caused by an enhanced greenhouse effect is defined as the impact of human radiative active gas emission on the radiative forcing of the atmosphere, causing the temperature at the earthâ&#x20AC;&#x2122;s surface to rise (Guinee, 2001). It should be noted that biogenic formation of CO2 is not incorporated in the global warming potential (GWP) in contrast to fossil-based CO2, methane (CH4) and nitrous oxide (N2O) emission, which accounts for 1, 25 and 298 kg CO2-eq kg-1 emission, respectively. Figure 8.2C shows that the GWP is mainly dominated by the greenhouse gases emission of the WWTP itself, although also electrical energy consumption, composting and the production of chemicals contributed. On the other hand, the recovery of C, N and P through composting and production of electricity on-site saved greenhouse gas emissions produced during the production of the respective fertilizer mixes and electricity at the Austrian grid.

Due to the high GWP of N2O, the CO2 footprint of the WWTP was for 91, 98, 98 and 99% determined by the total N2O emissions measured for scenario 1a, 1b, 2 and 3, respectively. As the latter was the only difference between scenario 1a and 1b, a 5-fold increase in N2O emission in the side line, increased the total CO2 footprint of the plant from 0.16 to 0.53 kg CO2-eq m-3 sewage. As the emissions between scenario 2 and 1a were similar, the lower total GWP of the plant during scenario 2 was mainly caused by a net electricity production (Fig. 8.2C). The CO2 footprint of the plant with DEMON in the side line was 0.12 kg CO2 m-3 sewage or around 7 kg CO2-eq PE-1 year-1, which is low compared to the average reported operational CO2 footprints of WWTP ranging from 12-80 kg CO2-eq PE-1 year-1 (Hospido et al., 2008; Clauwaert et al., 2010). Moreover, dependent of the electricity mix provided by the grid, a CO2 neutral WWTP based on GWP is feasible as the GWP of electricity production can significantly differ per country (Fig. S8.1).

During the implementation of DEMON in the mainline of the WWTP in Strass, nitrite accumulation was observed increasing the B-stage N2O emission from negligible to 2.3% of its N-load (Fig. 8.4). This increase caused the higher CO2 footprint of the total system during scenario 3 compared to scenario 2: 0.66 compared to 0.12 kg CO2-eq m-3 sewage treated, respectively (Fig. 8.2C). As the mainstream DEMON operation was not stable yet and 150

Chapter 8

adaptation and improved process control could probably lower the N2O emissions (Ahn et al., 2011), the CO2 footprint can be further optimized. It could be estimated that one should aim for a maximum N2O emission in the mainstream DEMON reactor of 0.5 % of the N-load to maintain the same CO2 footprint as in scenario 2. However, it should also be noted that the CO2 footprint of scenario 3 (36 kg CO2-eq PE-1 y-1) still correlated well with the average CO2 footprints of WWTP (Hospido et al., 2008; Clauwaert et al., 2010). This indicated that the WWTP of Strass can be seen as a benchmark WWTP, not only based on energy efficiency but also based on GWP.

3.3.4 Impact categories of minor importance for DEMON implementation The acidification potential (AC) was mainly counteracted by the recovery of nitrogen and phosphorous (Table 8.3). For the plant itself, the SO2 emission of the CHP unit was the main factor that contributed to the AC potential. Therefore, DEMON application as such had no significant influence on this impact category. The same minor influence of DEMON implementation could be observed for the ozone depletion potential, which was mainly dominated by chemical usage, for the ecotoxicity, which was related with the amount of nutrient recovery and for the photochemical oxidation potential which was mainly influenced by the emissions (CO, NO2 and SO2) from the CHP unit (Table 8.3). It should however been noted that the ecotoxicity impact in this study was relatively low in contrast to reported LCA studies for WWTP (Hospido et al., 2008; Clauwaert et al., 2010; Foley et al., 2010), because the usage phase of the compost was excluded and thus also the environmental impact of its metal content.

151

Environmental assessment of the implementation of OLAND in sewage treatment plants

Figure 8.4: N2O, NO and CO2 emission data selection from the B-stage during scenario 2 (top) and 3 (bottom) which had a N load of 780 and 826 kg N d-1, respectively. Fluctuations in the CO2 emissions were strongly correlated with the aeration regime.

152

Chapter 8

Table 8.3: Results of the impact assessment for the acidification potential (AC), ozone depletion potential (OD), photochemical oxidation potential (PO), freshwater aquatic ecotoxicity (FET) and terrestrial ecotoxicity (TET). Negative values are related to impacts that are avoided for the production of C, N and P fertilizers. Impact category Unit Scenario 1 Scenario 2 Scenario 3 Total

WWTP

Total

WWTP

Total

WWTP

(kg SO2-eq m-3 sewage)

0.00033

0.00094

-5 10-5

0.00072

4.5 10-5

0.00073

(kg CFC-11-eq m-3 sewage)

2.1 10-8

1.4 10-8

3.0 10-9

-3

-5

(kg C2H4-eq m sewage)

5.3 10

FET

(kg 1.4-DB-eq m-3 sewage)

-0.0034

TET

(kg 1.4-DB-eq m-3 sewage)

MET

(kg 1.4-DB-eq m-3 sewage)

7.0 10

-5

3.6 10

-5

6.2 10-5

3.3 10

6.3 10

-0.0030

-0.0027

-7.9

-10

-8.7

0.00049

0.00010

-0.00018

153

Environmental assessment of the implementation of OLAND in sewage treatment plants

Conclusions

On plant level, DEMON implementation, which excluded the need for a COD sources in the side line, had the potential to save 13% of the electricity consumption through a higher electrical energy recovery. Besides the saving in resources, side stream DEMON implementation positively influenced the eutrophication and global warming potential, the most important categories of the LCA of the WWTP in Strass. First results of DEMON implementation in the mainstream of the WWTP showed the potential to further decrease the energy consumption and therefore also the abiotic depletion potential. However, the first tests also showed a higher risk for increased eutrophication potential and increased global warming potential due to increased N2O emissions. Therefore, further optimization of the operational conditions will be needed to obtain an environmental sustainable treatment plant with DEMON in the mainstream.

Acknowledgements

H.D.C. is a supported by a PhD grant from the Institute for the Promotion of Innovation by Science and Technology in Flanders (IWT-Vlaanderen, number SB-81068). T.S. is granted by a research project (number 3G092310) of the Research Foundation - Flanders (FWOVlaanderen). The investigations at the Strass treatment plant were also supported by the Austrian Federal Ministry of Environment. The authors gratefully thank Tim Lacoere for technical support, Martin Hell for providing operational data of the plant and Steven De Meester, Rodrigo Alvarenga, Siegfried E. Vlaeminck and Chris Callewaert for inspiring scientific discussions.

154

Chapter 8

Supplementary data

Table S8.1: P&N. N and P fertilizer mixes composition based on the consumption of the individual fertilizers in 2010 (International Fertilizer Industry, 2012). Compounds Consumption Amount per kg of (2010) fertilizer mix P&N fertilizer mix* ktonnes P2O5/yr %P Monoammonium phosphate NH4H2PO4 313 26.16 Diammonium phosphate (NH4)2HPO4 884 73.84 N fertilizer mix ktonnes N/yr %N Urea CO(NH2)2 4950 43.59 Ammonium nitrate NH4NO3 2714 23.90 Calcium ammonium nitrate 20-30% CaCO3 2928 25.79 and 70-80% NH4NO3 Ammonium sulphate (NH4)2SO4 763 6.72 P fertilizer mix ktonnes P2O5/yr %P Triple super phosphate Ca(H2PO4)2 251 100.00 *Amount of N per kg of P in P&N fertilizer mix is 0.7861

155

Environmental assessment of the implementation of OLAND in sewage treatment plants

Figure S8.1: Country-dependent impact of the electricity mix used on the global warming potential and abiotic depletion potential.

156

157

Lab-scale OLAND rotating biological contactor (RBC junior, LabMET) 158

Chapter 9

Chapter 9: General discussion and perspectives 1

Main outcome and positioning of this work

In this doctoral work, in a first phase, the output of the OLAND process for the treatment of digestates was optimized and studied in detail. Low volumetric exchange ratios, which assure stable hydraulic conditions, were needed to allow a fast start-up, granulation and high performance in SBR systems (Chapter 2). The sustainability of the process in terms of NO/N2O emissions, was mainly linked with accumulations of intermediates such as NO2- and NH2OH and the frequency of transient conditions (Chapter 3). Better understanding of the conditions which lead to the accumulation of intermediates and further optimization of the feeding pattern which determines the degree of fluctuations, will allow a further decrease of the N2O emission in these systems. In a second part of this work, new application domains for the OLAND process, which could improve the overall sustainability of the applied processes, were explored. Energy calculations revealed that OLAND could significantly increase the energy index of agroindustrial and OFMSW-based treatment system from 3-5 to 6-10 (Chapter 4). However, for manure-based digestate treatment, OLAND application seemed more difficult and therefore ammonia gas treatment by OLAND was suggested for this application domain. A pilot-scale OLAND biofilter fed with a flow of ammonia gas, obtained a high performance (0.7 g N L-1 d-1) and a high total nitrogen removal efficiency (75-80%). Although the filter was saturated with oxygen, the low relative water flow rate ratio (â&#x2030;&#x2C6;1 L g-1 Nin) ensured high FA concentration in the water phase, which resulted in a dominance of AnAOB compared to NOB activity at the top of the biofilter (Chapter 5).

A specific application domain, which could particularly improve the energy efficiency of sewage treatment plants was the implementation of OLAND in the mainstream of the system. This would allow a net electrical energy production, due to a higher carbon recovery and lower energy needs for aeration (Chapter 4). Four challenges to allow mainstream OLAND 159

General discussion and perspectives were encountered. A first challenge, namely the performance of OLAND at low nitrogen concentration and low hydraulic residence time was shown in an OLAND RBC (Chapter 6). The reactor obtained high nitrogen removal rates (0.4 g N L-1 d-1) treating nitrogen concentration of 30-60 mg N L-1 at a HRT of 1-2 hours. A second challenge, operation at low temperatures (15째C), was surmounted in the same RBC by gradually decreasing the temperature starting from 29째C. During operation at 15째C with synthetic feed (60 mg N L-1) and a HRT of 1h, a similar nitrogen removal rate as at high temperatures was obtained i.e. 0.5 g N L-1 d-1 (Chapter 7). Compared to higher temperatures only a decrease of the total removal efficiency of 22% was detected. The switch from synthetic feed to pretreated sewage with a COD/N ratio of 2 (challenge 3) did not significantly affect the performance (Chapter 7). However, during the low temperature performance of the RBC system, NOB activity started to increase, as well as competition between AnAOB and NOB for nitrite (challenge 4). It was shown that increased levels of NO selectively enhanced AnAOB over NOB activity (Chapter 7). Therefore, high peak loading rates together with nitrite accumulation, increasing the NO production, enhanced the overall removal efficiency. To evaluate the mainstream OLAND application in a broader context, a LCA was performed on full-scale data of the WWTP in Strass, in which an OLAND-type of process, called DEMON was implemented. Three scenarios were studied: (1) the WWTP without a DEMON system; (2) the WWTP with DEMON in the side line; (3) the WWTP with DEMON in the main line. For the latter scenario, data from a first full-scale trial were used. The LCA showed that implementation of DEMON in the side line of the WWTP positively influenced all impact categories and therefore resulted in a more sustainable WWTP. The first full-scale results ever of DEMON implementation in the mainstream of the WWTP in Strass (Austria) showed that to obtain the same degree of sustainability than the sidestream treatment, the N2O emission (around 2% of N load) in the main line should be decreased. As N2O emission is mainly related with operational conditions and not with the process itself, it should be possible to further optimize the emission to around 0.5% of the N load allowing the same CO2 footprint of the plant in comparison with sidestream DEMON implementation.

2 2.1

OLAND and sustainability Balancing energy recovery with sustainability

The overall CO2 footprint of a WWTP is dominated by the amount of N2O emitted (Chapter 8, Foley et al. 2010). Generally, it is accepted that the OLAND process for the treatment of 160

Chapter 9

highly N-loaded streams such as digestates, emits around 1% of the N load as N2O (Chapter 3, Kampschreur et al., 2009a; Weissenbacher et al., 2010). This is mainly steered by the degree of transient oxygen and N-loading conditions and the accumulation of NH2OH and NO2-. However, for mainstream OLAND it is not clear yet if the high N2O emissions (1-5% of N load, Chapter 7-8) are essential to allow AnAOB activity at low nitrogen concentration and low temperatures or if there is room for further optimization. Therefore, further research in this field is needed to achieve full-scale mainstream OLAND treatment.

At this moment, insufficient comparative data are available about N2O emissions in conventional activated sludge systems with nutrient removal, making it hard to set a critical level of acceptable N2O emissions for mainstream OLAND. Reported N2O emission in activated sludge systems ranged from 0.001-25% of the N load and were mainly related to nitritation activity (Desloover et al., 2011b). To evaluate if a certain level of N2O emission is acceptable for mainstream OLAND, it is proposed to evaluate the sustainability of the plant before and after implementation of mainstream OLAND (Chapter 8). It should be possible to counteract a certain increase in the N2O emission by the increase in energy recovery to maintain a constant CO2 footprint. For the treatment plant in Strass (Austria), it was estimated that a N2O emission in the mainstream of 0.5% of the N load would be acceptable compared to the current performance (Chapter 8).

As the degree of N2O emissions are strongly related to the operational conditions rather than the nitrification/denitrification or OLAND process, mitigation of N2O emission should be possible (Chandran et al., 2011; Desloover et al., 2011b). In the next section, mitigation strategies are proposed based on the control of the N2O production and control of the N2O emission.

2.2

Mitigation strategies based on chemical markers

In Chapter 3, a detailed analysis of the relation between accumulation of chemical intermediates and N2O emission was performed on a full-scale OLAND-type reactor treating sludge digestate. This study showed that NH2OH and NO2- were both precursors for increased N2O emission. These intermediates were mostly formed during transient conditions, regarding the input of oxygen and ammonium (Chapter 3). The uncoupling of AerAOB with NOB, AnAOB or heterotrophic denitrifiers can also occur in less transient conditions, for example by inhibition of one of the above groups. Thus, the inhibitory effect of NO towards NOB 161

General discussion and perspectives could have been responsible for the increased nitrite accumulation during mainstream treatment and therefore also for the increased N2O emission (Chapter 7). A first strategy to decrease N2O emission, as suggested in Chapter 1, could consist of performing OLAND at stable operational conditions, which allow constant specific microbial activities and therefore avoiding accumulation of NO2- and NH2OH. However, stable operational conditions, for example by applying constant aeration instead of intermittent aeration, do not always generate lower N2O emission (Joss et al., 2009). Moreover, it should be noted that a certain accumulation of nitrite can be needed to channel nitrite into the anammox or nitrite oxidation route as the corresponding microbial groups have affinity constants of 0.05 and 5.5 mg N L-1, respectively (Lackner et al., 2008). Therefore, monitoring of the precursors, NO2- and NH2OH, of N2O emission could be necessary. However, NH2OH concentrations detected in OLAND systems are always very low, which makes it difficult to evaluate changes. Moreover, for mainstream conditions which work at lower nitrogen concentrations, the monitoring of NH2OH will not be reliable enough. Compared to NH2OH, NO2- concentration differences are easier to detect as they occur in higher concentrations. However, the on-line NO2- probes on the market still need further optimization to allow reliable long-term measurement.

During a first full-scale trial to implement OLAND in the mainstream of the WWTP of Strass (Austria), an increased nitrite accumulation (up to 9 mg N L-1) was observed. This was especially the case in wintertime when the loading rate of the WWTP significantly increased due to the tourist season (Wett et al., 2012). A shift in the effluent value of the nitrite over nitrate levels was observed reaching values above 1, which however also lead to increased levels of NO (0.0004-0.03%) and N2O (2-9%). N2O emission in this full-scale system was mainly steered by the degree of nitrite accumulation and not by the transition from anoxic to oxic conditions. This was indicated by the increased levels of N2O emissions at higher DO concentrations from 1 to 3 mg O2 L-1 (Fig. 9.1) and concomitantly increased nitrite levels from 0.5 at the lowest DO set point up to 4 mg N L-1 at the highest set point. Due to the unreliable online measurement systems for nitrite in practice, control through an operational parameter, which is strongly linked with nitrite is advisable. The latter is for example done in the DEMON process, which is based on pH decrease caused by the oxidation of ammonium to nitrite.

162

Chapter 9

Another option is to control the aeration system based on the on-line measurement of gaseous NO2 as this parameter is directly determined by NO2- and is easy to measure in the gas phase. A drawback could be that depending on the type of wastewater or type of sludge used, the relation between the NO2 emitted and the NO2- concentration in the liquid phase could differ (Weissenbacher et al., 2007). Moreover, NO2 levels tend to be very dynamic (Fig. 9.1) and from the moment NO2 is emitted, already significant levels of NO2- are present in the system, which could have already caused increased N2O emissions (Weissenbacher et al., 2010). Therefore, a control strategy based on this parameter should be further explored.

NO can be seen as a universal N2O precursor as this compound is always formed before N2O is emitted. Moreover due to the low solubility of NO, this compound could give a faster indication of accumulation of intermediates, which are difficult to detect. As NO could control the microbial balance under mainstream conditions (Chapter 7), online measurement and aeration control through the measurement of NO could probably allow a good microbial balance and avoid excessive NO2- levels and as a consequence excessive N2O emissions. Further long-term measurement of NO at WWTP with mainstream OLAND is necessary to find a relation between NO emission, performance and N2O emission and to select threshold concentrations for proper control.

Figure 9.1: Emission of NO, N2O and NO2 in relation to the oxygen set point tested in the B-stage of the WWTP of Strass (Austria) inoculated with DEMON granules (preliminary results). Detection limit of the NO measurement was 1 ppm. Higher NO concentrations were therefore set at 1000 ppb.

163

General discussion and perspectives

2.3

Mitigation strategies which minimize emission

It is important to distinguish N2O formation from N2O emissions, which is a physical mechanism governed by stripping in aerated parts of the system and by passive diffusion, mixing and wind advection in non-aerated compartments. Because of the difference, limiting the transfer of the formed N2O to the atmosphere can lower the overall emissions of nitrogen removing plants. For systems with active aeration, minimization of the airflow rate could lower N2O emissions (Kampschreur et al., 2008; Kampschreur et al., 2009a). Other factors that could influence the physical transfer of N2O from the water to the gas phase are the aeration system itself (size of bubbles) and the aeration control (avoidance of overaeration). In Figure 3.3 (Chapter 3) an example has been given of the establishment of full stripping of N2O. A lag phase between N2O stripping and CO2 and NO emission was observed due to the difference in solubility and the stepwise formation of N2O from NO. The length of the aeration periods (N2O formation) as well as the length of the anoxic periods (N2O consumption) could therefore also play a role in the overall degree of N2O emission. Moreover, it was shown that bubbleless aeration systems such as membrane-aerated systems could lead to a 100-fold decrease in N2O emissions (Pellicer-Nacher et al., 2010). Similarly, systems based on passive aeration such as RBC systems, are believed to emit less N2O than systems with active aeration because of the lower kLa (Desloover et al., 2011b). However, data on biofilm-based systems are limited and the high N2O emission measured in the RBC of Chapter 7, did not confirm this assumption.

As NO and N2O will always be present at a certain level in systems based on nitritation, better understanding of the parameters influencing the mass transfer of N2O from the liquid to the gas phase will allow further optimization of the overall greenhouse gas emission of the WWTP.

3 3.1

Energy positive WWTP: reality or fantasy? Water-energy nexus

Water and energy are intertwined. Water is needed for energy production to power the turbines in hydro-electric facilities, for cooling in thermal or nuclear energy plants, and to extract oil from tar sands etc. Energy is needed to pump, treat and heat water, to generate steam for urban, industrial and agricultural use and to deal with the resulting wastes (Table 9.1). Moreover, the water-energy nexus is deeply connected with the climate change. Burning 164

Chapter 9

fossil fuels, water transport through sewers and wastewater treatment systems all contribute to the emission of greenhouse gases and therefore add their part to the global warming. As the latter causes an increased rate of evaporation, variability in precipitation and a greater demand for cooling, the climate change, which is mainly created by energy use, is strongly experienced through the water cycle (Lazarova et al., 2012).

In a time of climate change and global warming, a need for a holistic approach, which also integrates the growing urban development, is advisable to manage water and energy, along with nutrients. Therefore, closing the water and energy cycles could be a step forward in decreasing the use of resources. Advanced wastewater treatment methods, mainly based on micro- and ultrafiltration are needed to reuse water, but the latter also requires more energy (Table 9.1). Nevertheless, alternatives such as desalination are still not competitive enough in terms of energy and costs (Table 9.1). The most energy efficient desalination plant (Ashkelon, Israel) requires an energy consumption of 2.9 kWh m-3 water produced (Voutchkov, 2010), which is still 6 times higher than the application of advanced water reuse for the same treatment capacity (Mehul and Dunvin, 2010). Within the field of wastewater treatment, one can try to maximize water reuse and therefore treat the water only towards a specific purpose. It was already suggested to reuse for example grey water from the washing machines and bathing, after a limited treatment in a decentralized system as toilet flushing water (Bieker et al., 2010). On the other hand, production of potable water from conventional activated sludge treatment effluent is economically and technically feasible through a multiple barrier approach using microfiltration, reverse osmosis and UV disinfection methods. Several examples of the latter are operational in Singapore (PUB, 2010), Belgium (Dewettinck et al., 2001) and California (OCWD, 2009).

Table 9.1: Water footprint for energy production and energy footprint for water elements of the water cycle (Lazarova et al., 2012). Water for energy Energy for water Energy source m3 MWh-1 Elements of water cycle kWh m-3 Gas 0.38 Potable water treatment 0.2 -1.5 Nuclear 0.38 Potable water distribution 0.05 – 0.24 Coal 0.72 Preliminary treatment of wastewater 0.16 – 0.3 Solar thermal 1.1 Activated sludge system (AS) 0.25 – 0.6 Crude oil 4.0 AS with nitrification 0.3 – 1.4 Hydropower 250 Water reuse 0.2 – 2.5 Biogas from crops 600 Brackish water desalination 1 – 1.5 Biodiesel from crops 1130 Seawater desalination 2.5 - 5

165

General discussion and perspectives Besides water reuse from WWTP, also energy recovery can be obtained. As discussed in Chapter 4, wastewater has a high energy content in the form of heat and organic carbon. Enhanced reuse of the energy contained in wastewater is another possibility to improve the water-energy nexus. Maximization of the energy recovery by anaerobic digestion and minimization of the energy consumption by autotrophic nitrogen removal (Chapter 4), can therefore lead to energy-neutral or even energy-positive wastewater treatment plants. How one can reach energy self-sufficient systems is discussed in the following sections.

3.2

Is OLAND an essential treatment step?

3.2.1 Yes it is, to allow maximum energy recovery Implementation of OLAND in the municipal wastewater treatment chain, allows high nitrogen conversion rates without the need for organic carbon. Therefore, to allow maximum recovery of organics from sewage and obtain a dischargeable effluent quality, implementation of OLAND is needed. The application of OLAND in the sidestream for the treatment of the digestate, has already shown to be reliable, robust and highly efficient (Table 1.4, Chapter 1). Both by energy calculations (Chapter 4, Siegrist et al., 2008) and full-scale experiences (Wett et al., 2007), it was shown that for sidestream OLAND, around 50% of the aeration requirements of the plant could be saved compared to CAS. It should however be noted that the net energy decrease was mainly caused by the higher recovery of organic carbon by anaerobic digestion, due to the lower COD removal needed in the mainstream (Chapter 4). Conditions for sidestream OLAND provide easy control of the microbial balance due to the high temperatures and high nitrogen concentrations, which allow prevention of nitrate production by FA inhibition. Therefore, it is possible to guarantee high nitrogen removal efficiencies for the treatment of digestates (Table 1.4, Chapter 1). Also for industrial and decentral treatment of digestates with OLAND, the energy index significantly increased (Chapter 4). Moreover, stable and highly efficient performance is already demonstrated in practice. Therefore, the implementation of OLAND for the treatment of digestates is advisable to decrease the overall energy consumption and allow higher organic carbon recovery.

For mainstream conditions, our lab-scale reactor tests (Chapter 6 and 7) showed that high nitrogen removal rates could be obtained at mainstream conditions (low temperatures, low nitrogen, COD/N of 2). However, competition between AnAOB and NOB in these conditions is more difficult to control compared to the conventional techniques based on FA, FNA and 166

Chapter 9

oxygen used in sidestream treatment. It was suggested that NOB suppression by NO concentration could be an alternative strategy as this compound stimulated AnAOB over NOB activity (Chapter 7). However, NO also decreased the AerAOB activity, which indicated that a balance between obtaining enough nitritation without substantial nitratation is essential to allow high nitrogen removal efficiencies. A first full-scale approach (Strass, Austria; Chapter 8) in which a combination of OLAND and nitrification/denitrification was applied, has experimentally examined several operational conditions, mainly related to the input of oxygen (data not shown). These first primary tests showed that NOB easily adapted to lower DO conditions in the reactor, and therefore were strong competitors for nitrite compared to AnAOB. However, when high loading rates (winter time) and/or DO set points of 1-2 mg O2 L-1 were applied, high NO/N2O emission occurred together with nitrite accumulation, showing a decrease in NOB activity. These primary results at full-scale therefore correlate very well with the lab-scale reactor tests shown in this work (Chapter 7). Further research is needed to elucidate the NO concentration needed to suppress NOB and evaluate the sustainability of this application compared to conventional treatment. Moreover, mainstream OLAND should first show reliability, sustainability and energy efficiency at larger scale, before conclusions can be made about the need for implementation to allow energy-positive treatment.

3.2.2 No it is not, other adjustments can help Improving wastewater treatment performance is and should be the primary objective of a WWTP. After obtaining stable discharge limits for the effluent, the best available practices and technologies for enhanced energy efficiency and the best use of sludge for energy production and recovery can be investigated and implemented. OLAND allows for lower oxygen consumption and thus a lower aeration rate compared to conventional nitrification/denitrification (Kuai and Verstraete, 1998). However, energy calculations revealed that also oxygen transfer efficiency, digestibility of sludge and efficiency of anaerobic digestion could significantly influence the energy balance (Chapter 4, BOX 5).

As aeration accounts for 60-70% of the energy demand of a WWTP (Zessner et al., 2010), optimization in this area can save up to 20% of the energy consumption (Lazarova et al., 2012). Aeration systems can achieve oxygen transfer efficiencies above 2 kg O2 kWh-1 (Nowak et al., 2011) and the use of premium efficiency motors and variable frequency drivers for large pumps and aeration blowers can also limit the total energy consumption for the same 167

General discussion and perspectives amount of oxygen input. Moreover, efficient aeration control systems, for example based on the measurement of ammonium in the effluent, can further optimize energy input.

In addition to energy input for aeration, the efficiency of energy recovery from sludge can influence the energy balance significantly (Chapter 4, Box 5). As primary sludge is easier to digest than secondary sludge, increasing the primary sludge production can also improve the energy recovery. This can for example explain the difference between the CAS and A60/B system (Chapter 4, Table 4.2). Next to this, the primary and secondary sludge mixture can be pretreated or codigestion can be applied to further increase the methane yield (Carrere et al., 2011). The CHP-units operational at present have an electrical efficiency of 31-38% (Nowak et al., 2011). Therefore, the choice of the CHP unit or the choice of the method to use the biogas (through electricity, via gasification or through direct mechanical energy) can also influence the energy gain obtained. Together, these adjustments can lead to 20-40% increased energy recovery (Lazarova et al., 2012).

Furthermore, depending on the climate and transport distances, energy can also be gained from the sewage flows themselves by hydro-turbines and heat pumps (Verstraete and Vlaeminck, 2011). This can increase energy recovery up to 10% (Lazarova et al., 2012). Another 10% increase in energy recovery can be obtained by the production of renewable energy from external sources such as solar, wind or geothermal energy (Lazarova et al., 2012).

Together, these adjustments can also lead to energy autarky, even for CAS systems. This was for example shown in the Wolfgangsee-Ischl WWTP (Austria), which treats 40 000 IE and is based on a singly stage AS-system with primary sedimentation and anaerobic sludge digestion (Nowak et al., 2011). From 2009 onwards, energy self-sufficiency was reached, by optimization of the aeration system and control (2.3 kg O2 kWh-1), increasing primary sedimentation (37%), optimization of the digesters (2 in series with a SRT of 80 days) and implementation of a better CHP unit (electrical efficiency of 34%). Therefore, this full-scale example shows that OLAND is not essential in all cases to obtain energy self-sufficient WWTP.

168

Chapter 9

Figure 9.2: Overview of the degree of OLAND implementation and oxygen transfer efficiency needed (A: 1 kg O2 kWh-1; B: 2 kg O2 kWh-1) to allow energy-positive WWTP in function of the primary sludge production efficiency and COD/N of the incoming sewage. Grey: energy-negative; yellow: energy-positive without OLAND implementation; orange: energy-positive if OLAND is implemented in the side stream; red: energy-positive if OLAND is implemented in the meanstream. Primary sludge production higher than 75% is considered as technically not feasible at the moment (light grey boxes). Other parameters such as digestibility of the sludge, growth yield etc were kept at default values (Chapter 4, BOX 1).

169

General discussion and perspectives

3.3

Decision making for the wastewater engineer

It is clear that OLAND is not essential in all cases to obtain energy neutral or even energypositive wastewater treatment. Several application domains have other needs and some general guidelines are subsequently proposed to help decide if OLAND implementation can be necessary to achieve an energy-positive WWTP.

For the treatment of highly-loaded organic streams, which can be immediately subjected to biogas digestion, OLAND treatment of the resulting digestate can have a high impact on the energy and cost balance. Examples were shown in Chapter 4 for the treatment of the OFMSW, agro-industrial waste and manure-based organics. Due the high digestibility of the first two streams, high energy recoveries could be obtained and OLAND could increase the energy index with a factor 2. However, energy-positive treatment was also obtained when conventional nitrification/denitrification was applied. However, the latter had higher needs for external carbon addition to meet discharge limits. Therefore, in treatment schemes with anaerobic digestion of agro-industrial waste and OFMSW, OLAND implementation is advisable. OLAND implementation for the treatment of manure-based organics seemed more difficult and the effect on the energy balance was therefore minor. In this field of application, treatment of the gaseous ammonia streams by OLAND showed a better potential for application (Chapter 5).

For a municipal WWTP due to its complexity, it is more difficult to estimate if energypositive wastewater treatment is possible and which prerequisites will determine the energy balance of the WWTP. It was suggested by Nowak and colleagues (2011) that energy autarky should be achievable for WWTP removing at least 70% nitrogen and treating sewage with COD/N ratios > 10 (Nowak et al., 2011). To test this proposal, some additional calculations were made based on the assumptions used in Chapter 4 (Fig. 9.2). It was shown that indeed the COD/N ratio of the sewage together with the degree of primary sludge production could affect the degree of OLAND implementation needed to obtain energy autarky. The higher the COD/N ratio of the sewage, the more easily energy-positive treatment is obtained, because a higher proportion of primary sludge can be separated without influencing the efficiency of the conventional nitrification/denitrification (Fig. 9.2). OLAND implementation in the sidestream becomes more important at sewage COD/N ratios between above 8-14 (Fig. 9.2). At lower COD/N ratios, mainstream OLAND together with substantial primary sludge production is 170

Chapter 9

needed to decrease energy consumption and allow optimal recovery (Fig. 9.2B). Besides the sewage COD/N ratio, the oxygen transfer efficicieny for aeration in both side and mainstream treatment plays a crucial role. In case of inefficient aeration (1 kg O2 kWh-1), energy autarky is very difficult and can only be achieved with high primary sludge productions (>65%) and high COD/N ratios (>10-13) in the sewage (Fig. 9.2). The aeration systems with higher oxygen transfer efficiencies (2 kg O2 kWh-1), create a higher possibility to attain energy autarky (Fig. 9.2B). For municipal WWTP, energy-positive treatment is only possible when during primary settling more than 50% of incoming COD is removed with the primary sludge. It should be however noted that the latter assumes a default anaerobic digestion efficiency while further improvements in this step are still conceivable (Chapter 4). A higher methane yield can further shift the pattern towards energy-positive treatments even at lower primary sludge productions, which was for example the case in the Wolfgangsee-Ischl WWTP (Austria) (Nowak et al., 2011).

Nitrogen removal versus nitrogen recovery

Nowadays, the fertilizer industry is based on the Haber Bosh process, which catalytically combines hydrogen and nitrogen gas to ammonia (N2 + 3 H2 ď&#x192; 2 NH3) under high pressure (15-25 MPa) and high temperatures (300-550Â°C)(Chagas, 2007). The specific operational conditions make this process energy intensive (Table 9.2). In total, this process accounts for a nitrogen fixation rate of 120 106 tons N year-1 and is expected to increase up to 165 106 tons N year-1 by 2050 (Galloway et al., 2004). As a consequence, the occurrence in the environment of reactive nitrogen compounds such as ammonium, nitrite and nitrate is sharply increasing. It was estimated that the global population discharges around 20 106 tons N year-1 in wastewater of which 99% of this reactive nitrogen is not treated and released as such in the environment (Galloway et al., 2008). The increasing amount of regulations and the increasing number of WWTP with nutrient removal is a first step towards decreasing the environmental problems of this excess of reactive nitrogen compounds. During our research, the focus was always on nitrogen removal and the aim to remove nitrogen in a cost effective and energy friendly way. However as resources are decreasing, recovery of nutrients will become necessary.

Nitrogen recovery can be obtained by several physico-chemical methods of which ammonia stripping and struvite precipitation are the most common ones (Siegrist, 1996). When ammonia stripping is applied an acidic salt NH4(SO4)2 is formed which can be concentrated 171

General discussion and perspectives after drying. The nitrogen can also be recovered as (NH4)2CO3. To obtain high recovery efficiencies, high nitrogen concentration levels, high pH values and high temperatures are advisable as these factors shift the ammonium balance to ammonia. During struvite precipitation an insoluble magnesium ammonium phosphate (MgNH4PO4.6H2O) is formed under alkaline conditions (pH 8.5 â&#x20AC;&#x201C; 10). A N/P ratio above 1 is needed and enough Mg-ions have to present in the water to avoid external addition of phosphate or magnesiums salts, respectively. Both NH4(SO4)2 and struvite can be used as fertilizers and therefore have the potential to decrease the need to perform the energy intensive Haber-Bosh process. However, at this moment the price of fertilizers made through the Haber-Bosh process are too low to give enough driving force towards nitrogen recovery (Table 9.2). As energy prices are rising, the price of the fertilizers will follow this trend, and the economical discrepancy between nitrogen removal and nitrogen recovery will decrease in the future. At this moment nitrogen recovery is only a cost-efficient option if streams contain more than 5 g N L-1 (Mulder, 2003). The latter is for example the case for urine (Larsen and Gujer, 1996).

Table 9.2: Comparison of the maximum fertilizers market prices (Apodaca, 2007), converted at 1.4 USD EURâ&#x2C6;&#x2019;1, and the costs for nitrogen recovery (Siegrist, 1996; personal communication Colsen nv) and removal (Fux and Siegrist, 2004) from sludge digestates. Cost Energy (â&#x201A;Ź kg-1 N) (kWh kg-1 N) Nitrification/denitrification 1.79 2 Nitrogen removal OLAND 0.29 1 Ammonia stripping 2-10 6-25 Nitrogen recovery Struvite precipitation 13 28 Haber-Bosh process 0.37 9-12 Fertilizer production Anhydrous ammonia 0.59 Ammonium nitrate 0.87 Ammonium sulphate 0.93 Urea 0.83

A combination of nitrogen recovery and nitrogen removal can also be attractive. In this way the nitrogen recovery efficiency only goes to the point where it is still economically attractive. The last part of the nitrogen, which is the most difficult to recover could then be removed by OLAND or nitrification/denitrification, depending on the COD/N ratio. Especially, after a thermophilic anaerobic digestion step, nitrogen recovery through ammonia stripping could be attractive as in this way the heat is more efficiently used compared to direct biological treatment of thermophilic digestates which have to be cooled first. Full-scale tests revealed that the treatment cost for ammonia stripping of thermophilic digestates with nitrogen 172

Chapter 9

concentrations around 2 g N L-1, could already be decreased to 2 EUR kg-1 N recovered (personal communication Colsen nv). This is only a factor 2 difference compared to a final treatment scheme based on nitrogen removal, which can meet discharge limits. Therefore, further optimizations of the operational conditions for ammonia stripping, especially regarding the stripping mechanism itself, will allow a cost-effective combination of nitrogen recovery and removal in the future.

Besides the economical aspect, from a LCA point of view, nutrient recovery can positively affect impact categories such as the eutrophication potential, abiotic depletion potential, global warming potential and ecotoxicity potential (Chapter 8). Therefore, nutrient recovery can increase the overall sustainability of the WWTP. The latter can also be obtained by for example composting of waste sludge (Chapter 8). However, the latter does not allow the production of high purity products and therefore tends to decrease the value of the product.

5 5.1

Future challenges and opportunities Future challenges for mainstream OLAND

Decreasing the nitratation activity at mainstream conditions and thus allowing sufficient AnAOB activity will be the main challenge for the future large-scale implementation of OLAND. One of the research lines that could be followed, as already highlighted before (Chapter 9, section 2.2) is the optimization of a control system based on NO measurement or a parameter closely related with NO. This strategy will therefore profit from the higher sensitivity of NOB for NO, compared to AerAOB and AnAOB (Chapter 7). Another possibility could be to install a small breeding reactor, which treats a highly-loaded nitrogen stream (digestate), but at low temperature (for example max 20째C). By recirculating the OLAND biomass from the mainstream to the breeding reactor, inhibition of NOB can be established in this breeding reactor without complicated control strategies needed in the mainstream. It was for example shown that the NOB in the mainstream biomass, which were grown at low FA concentration, were easily inhibited by this compound (Chapter 7, Table 7.3). The main challenge in the latter case is to find an elegant way to easily circulate OLAND biomass without the activated sludge and to determine the SRT of the sludge needed in both steps to obtain an optimal mainstream performance without long-term adaptation to the inhibitory factors in the breeding reactor. As during mainstream treatement a separation between the SRT of AnAOB containing particles and aerobic flocs is advisable to maintain 173

General discussion and perspectives AnAOB activity in the system (Wett et al., 2010), these separation systems could be connected to the breeding reactor. Until now, one system based on cyclones was proposed (Wett et al., 2012). However, the inoculation of the B-stage with OLAND biomass on e.g. a floating carrier, which can easily be retained in the system by grids, could be another technical solution. Besides, optimization of control mechanisms and suppression of NOB externally, the performance of mainstream OLAND could also be improved by a better combination of OLAND with nitrification/denitrification (Chapter 7). Implementation of predenitrification â&#x20AC;&#x201C; OLAND or OLAND â&#x20AC;&#x201C; postdenitrification are some options to guarantee stable discharge limits and to better counteract fluctuations. The latter would for example allow retrofitting of existing activated sludge systems.

5.2

OLAND biofilter application

In Chapter 5, it was shown that nitrogen removal directly from the gas phase was possible and that high nitrogen volumetric removal rates and efficiencies were obtained in an OLAND biofilter. As in practice ammonia containing gaseous wastestreams will rarely contain only ammonia, in most cases a combination of ammonia with sulfide (H2S) (Malhautier et al., 2003) and/or ammonia with volatile organic compounds (VOC) (Cabrol et al., 2009) will have to be treated. Therefore, the combination of ammonia removal through OLAND with sulfide and VOC removal is the challenge for successful implementation in practice. Sulfide can be inhibitory for AnAOB at levels around 10 mg S L-1 (Dapena-Mora et al., 2007). However, in an autotrophic denitrifying reactor based on H2S oxidation AnAOB were detected (Mulder et al., 1995) and other batch tests also showed that AnAOB could resist H2S concentrations of at least 64 mg S L-1 (van de Graaf et al., 1996). Therefore, it should be possible to cultivate an AnAOB culture that can adapt to higher H2S concentrations. A shift of the nitrifying activity to lower sections of the biofilters was observed in a biofilter treating H2S and NH3 (Malhautier et al., 2003), probably making the interference of H2S with the OLAND process minor. Moreover, in these filters low nitratation activity was observed leading to nitrite accumulation, which could give an opportunity to the AnAOB to survive in the system (Malhautier et al., 2003). Besides the shift in activity, also the limitation of NO emission will be challenging as H2S can chemically react with HNO2 to NO and oxidized sulfur compounds (Vermeiren et al., 2012).

174

Chapter 9

Emission of VOCs such as volatile fatty acids, ketones, aldehydes and alcohols are frequently associated with composting (Cabrol and Malhautier, 2011). The VOC concentration observed in these systems is mostly below 1 mg m-3, compared to NH3-N and H2S-S concentrations of on average 30 mg m-3 (Cabrol et al., 2009). Moreover, the solubility of these compounds differs and this will determine the contact with the OLAND biomass. Stable simultaneous VOC removal and nitrification was already shown in biofilters (Sakano and Kerkhof, 1998; von Keitz et al., 1999; Friedrich et al., 2003; Cabrol et al., 2009). To increase the total nitrogen removal in these systems, and therefore to apply OLAND, the effect of the VOCs composition and concentration on the AnAOB activity should be examined carefully.

5.3

What are the temperature limits of the OLAND process

This doctoral research showed that OLAND can be performed at low temperature (15째C) in contrast to the main mesophilic (30-35째C) research domain (Chapter 7, Chapter 1, Table 1.4). As AnAOB species were found in thermophilic conditions in nature, such as hot springs (Byrne et al., 2009) and high temperature petroleum reservoirs (Li et al., 2011), AnAOB activity should also be possible at thermophilic conditions. Thermophilic environmental biotechnology is established for carbon treatment (Wiegel and Ljungdahl, 1986). However, thermophilic nitrogen removal processes are not developed yet, although several types of nitrogenous wastewaters have temperatures above the mesophilic range. Most progress has been obtained in thermophilic denitrification (Laurino and Sineriz, 1991). However, for nitritation, nitratation and anammox no successful reports are found for thermophilic conditions (>40째C).

From more fundamental work however, thermophilic ammonium-oxidizing Archaea (AOA) were isolated and cultured (Hatzenpichler et al., 2008). Moreover, enrichments of thermophilic AerAOB and NOB were obtained (Lebedeva et al., 2005; Lebedeva et al., 2011; Shimaya and Hashimoto, 2011). It should therefore be possible to cultivate a nitrifying culture (AerAOB-NOB or AOA-NOB) and couple this to the existing thermophilic denitrification process (Laurino and Sineriz, 1991). Moreover, also in this case combining nitritation by AerAOB or AOA with thermophilic adapted AnAOB would allow a new application domain for OLAND.

For this application domain, the same strategy as for low temperature application (Chapter 7) can be applied e.g. gradual temperature adaptation. However, it could be needed to cultivate 175

General discussion and perspectives some adapted species, to mix and inoculate afterwards, as high temperatures will probably require thermostable enzymes (Haki and Rakshit, 2003). The first reactor tests performed on nitrification showed that a mesophilc consortium could only survive for 1 week at 45째C, although stable performance at 40째C was feasible (Shore et al., 2012). Therefore, this could suggest that specialist groups of AerAOB, AOA and/or NOB are needed.

Until now, at temperatures above 40째C, nitrogen loss is mainly caused by ammonia stripping and not by ammonium conversions (Abeynayaka and Visvanathan, 2011), which is not sustainable. Thermophilic nitrogen removal could however offer several advantages. First of all energy costs for cooling can be saved in contrast to treatment of thermophilic effluents from several industries such as pulp and paper manufacturing (Suvilampi et al., 2001) or from effluents of thermophilic activated sludge or anaerobic digestion, which at this moment first needs a cooling step before treatment is applied. Moreover, because of the high FA concentration and lower oxygen solubility at these high temperatures, suppression of nitratation and therefore control of the OLAND process will be easier. The challenges for this application domain will mainly be related with the search for a good consortium and with limiting the ammonia stripping.

Conclusions

Autotrophic nitrogen removal based on partial nitritation/anammox, as performed in a onestage treatment during the OLAND process, can significantly decrease the energy consumption, CO2 emission, sludge production and the needs for an external organic carbon source compared to conventional nitrification/denitrification. Due to these main advantages, the application of OLAND for the treatment of digestates can be seen as an established technology. Recently, some 44 full-scale plants using this process are reported. The results of this work for the mesophilic application domain showed that for the rapid start-up and high performance of OLAND SBR systems, stable hydraulic conditions (low volumetric exchange ratio) were needed. Moreover, to allow a sustainable process and thus low N2O emission, accumulation of NH2OH and NO2- should be avoided. In addition, energy calculations showed that new potential domains for OLAND were located (1) in agricultural application requiring ammonia removal and (2) in municipal WWTP using mainstream treatment.

176

Chapter 9

This work constitutes an attempt to contribute to the application of OLAND by the following key accomplishments:  For gas treatment, containing ammonia: o High removal rates (0.7 kg N m-3 d-1) and high removal efficiencies (75-80%) were obtained at a pilot scale (height 1.6 m) OLAND biofilter. o AnAOB activity and presence was obtained in the OLAND biofilter demonstrating the contribution of AnAOB to the ammonia removal process.  For mainstream water treatment, containing ammonium: o High total nitrogen removal rates (0.5 kg N m-3 d-1) were obtained at 15°C, nitrogen concentrations of 55 mg N L-1 and COD/N ratios of 2. o An alternative strategy of nitratation suppression at mainstream conditions based on NO was proposed.

177

General discussion and perspectives

178

179

Sludge from B-stage (WWTP Strass, Austria)

180

Abstract

Abstract Several new biological nitrogen removal processes, which are based on partial nitritation/anammox, have been developed to treat nitrogen-rich wastewaters devoid in carbon such as digestates. Around 40 full-scale realizations of one-stage partial nitritation/anammox, in this work referred to as the oxygen-limited autotrophic nitrification/denitrification (OLAND) process, are operational at this moment for high strength nitrogen streams. OLAND is based on partial nitritation, performed by aerobic ammonium-oxidizing bacteria (AerAOB) and anammox, performed by anoxic ammonium-oxidizing bacteria (AnAOB). The AerAOB, mainly belonging to Nitrosomonas europaea eutropha and halophila, are set so that they oxidize half of the influent ammonium to nitrite in oxygen-limited conditions. The AnAOB, mainly members of the Candidatus genera Kuenenia and Brocadia, oxidize the residual ammonium with nitrite to dinitrogen gas under anoxic conditions. Consequently, in the OLAND process ammonium is converted mainly into nitrogen gas without the use of organic carbon in one reactor. Overall OLAND can save 84% of the operational costs, by a 100, 89 and 57% decrease in methanol requirement, sludge production and aeration, respectively.

The close interaction between the different microbial groups during the OLAND process is comparable with human beings working together in firms for a shared profit. In this sense, the concept of human resource management (HRM) was translated to the microbial biotechnology as Microbial Resource Management (MRM) and therefore strives after maintaining the best performing microbial community for a certain application. A MRM OLAND

framework

was

elaborated

(Chapter

1),

showing

how

the

OLAND

engineer/operator (1: input) can design/steer the microbial community (2: biocatalyst) to obtain optimal functionality (3: output), depending on the application domain (0: wastewater). Taken this MRM framework into account, the OLAND engineer can steer the OLAND process to obtain maximum efficiency and higher sustainability or to increase the impact of OLAND on the energy balance of wastewater treatment plants (WWTP).

Although the first OLAND applications have shown that this technology works in a stable and efficient way, the implementation rate of this technology remains dependent on a few 181

Abstract companies. Many potential users hold back because it seems that due to the long start-up periods for the first reactors and the reported sensitivities, a lot of experience is needed to keep this process running. To overcome this problem, the output box of the MRM framework was further studied in detail for high-strength nitrogen containing wastewaters (known application). Firstly, the effect of the hydraulic conditions on the start-up of the OLAND sequencing batch reactor (SBR) was examined. Low volumetric exchange ratios, which assure stable hydraulic conditions, were needed to allow a fast start-up, granulation and high performance in SBR systems (Chapter 2). Furthermore, strategies to obtain a well-balanced OLAND system, were proposed based on wash-out of nitrite oxidizing bacteria (NOB) through selection on settling velocity or by stimulation of AnAOB through the implementation of an anoxic phase. As not only the effluent quality, but also the sustainability can be a competitive factor to choose an environmental technology, the N2O and NO emissions were studied in a full-scale OLAND-type reactor (Chapter 3). The sustainability of the process in terms of NO/N2O emissions was mainly linked with accumulations of intermediates such as NO2- and NH2OH and the frequency of transient conditions. Better understanding of the conditions which lead to the accumulation of intermediates and further optimization of the feeding pattern which determines the degree of fluctuations, will allow a further decrease of the N2O emission in these systems. In a next part of this work, new opportunities for OLAND, which could improve the overall sustainability of the applied processes, were explored (box 0 of MRM framework). Energy calculations revealed that OLAND treatment of digestates could significantly increase the energy index of agro-industrial and organic fraction of municipal solid waste-based treatment system from 3-5 to 6-10 (Chapter 4). However, for manure-based digestate treatment, OLAND application seemed more difficult and therefore ammonia gas treatment by OLAND was suggested for this application domain. A pilot-scale OLAND biofilter fed with a flow of ammonia gas, obtained a high performance (0.7 g N L-1 d-1) and a high total nitrogen removal efficiency (75-80%; Chapter 5). Although the filter was saturated with oxygen, the low relative water flow rate ratio (â&#x2030;&#x2C6;1 L g-1 Nin) ensured high free ammonia concentration in the water phase, which resulted in a dominance of AnAOB compared to NOB activity at the top of the biofilter.

A specific application domain, which could particularly improve the energy efficiency of sewage treatment plants was the implementation of OLAND in the mainstream of the system. 182

Abstract

This would allow a net electrical energy production, due to a higher carbon recovery and lower energy needs for aeration (Chapter 4). Four challenges to allow mainstream OLAND were encountered. A first challenge, namely the performance of OLAND at low nitrogen concentration and low hydraulic residence time (HRT) was shown in an OLAND rotating biological contactor (RBC; Chapter 6). The reactor obtained high nitrogen removal rates (0.4 g N L-1 d-1) treating nitrogen concentration of 30-60 mg N L-1 at a HRT of 1-2 hours. A second challenge, operation at low temperatures (15째C), was surmounted in the same RBC by gradually decreasing the temperature starting from 29째C. During operation at 15째C with synthetic feed (60 mg N L-1) and a HRT of 1h, a similar nitrogen removal rate as at high temperatures was obtained i.e. 0.5 g N L-1 d-1 (Chapter 7). Compared to higher temperatures only a decrease of the total removal efficiency of 22% was detected. The switch from synthetic feed to pretreated sewage with a COD/N ratio of 2 (challenge 3) did not significantly affect the performance. However, during the low temperature performance of the RBC system, NOB activity started to increase, as well as competition between AnAOB and NOB for nitrite (challenge 4). It was shown that increased levels of NO selectively enhanced AnAOB over NOB activity (Chapter 7). Therefore, high peak loading rates together with nitrite accumulation, increasing the NO production, enhanced the overall removal efficiency. To evaluate the mainstream OLAND application in a broader context, a life cycle assessment (LCA) was performed on full-scale data of the WWTP in Strass, which applied an OLANDtype of system, referred to as DEMON. Three scenarios were studied: (1) the WWTP without a DEMON system; (2) the WWTP with DEMON in the side line; (3) the WWTP with DEMON in the side and main lines (Chapter 8). For the latter scenario, data from a first fullscale trial were used. The LCA showed that implementation of DEMON in the side line of the WWTP positively influenced the eutrophication potential, abiotic depletion potential and global warming potential and therefore resulted in a more sustainable WWTP. The first fullscale results of DEMON implementation in the mainstream of the WWTP in Strass (Austria) showed that to obtain the same degree of sustainability compared to the WWTP with sidestream treatment, the N2O emission (around 2% of N load) in the main line should be decreased as this compound dominated the global warming potential of the plant with 99%. N2O emission is mainly related with operational conditions and not with the process itself, it should therefore be possible to further optimize the emission to around 0.5% of the N load allowing the same CO2 footprint of the plant in comparison with sidestream OLAND implementation.

183

Abstract Generally, this work showed that new potential domains for OLAND were located in agricultural applications requiring ammonia gas removal and in municipal WWTP using mainstream treatment. Future tests in these domains will need to evaluate the performance and overall environmental sustainability at larger scale

184

Samenvatting

Samenvatting Er werden reeds verschillende nieuwe biologische processen, gebaseerd op partiële nitritratie/anmmox ontwikkeld voor de behandeling van stikstofrijk afvalwater zoals bijvoorbeeld digestaten. Op dit moment zijn er ongeveer 44 volle-schaal éénstaps partiële nitritatie/anammox reactoren, in dit werk ook wel oxygen-limited nitrification/denitrification (OLAND) genoemd, operationeel voor de behandeling van hoogbelaste stikstofstromen. OLAND is gebaseerd op partiële nitritatie, uitgevoerd door aerobe ammonium-oxiderende bacteriën (AerAOB) en anammox, uitgevoerd door anoxische ammonium-oxiderende bacteriën (AnAOB) in één reactor. De AerAOB, die meestal tot de groep van Nitrosomonas europaea eutropha en halophila behoren, oxideren de helft van de ammonium tot nitriet onder zuurstofgelimiteerde omstandigheden. De AnAOB, meestal behorende tot de Candidatus genera Kuenenia en Brocadia, oxideren de overblijvende ammonium met het gevormde nitriet tot stikstofgas onder anoxische omstandigheden. Dus, tijdens het OLAND proces wordt ammonium in één reactor omgezet naar stikstofgas zonder gebruik te maken van een organische koolstofbron. Hierdoor kan het OLAND proces 84% van de operationele kosten besparen aangezien de behoefte aan externe methanol toediening, de slibproductie en de energiekost voor beluchting, met respectivelijk 100, 89 en 57% dalen.

De nauwe interactie tussen de verschillende microbiële groepen in het OLAND proces kan men vergelijken met werknemers, elk met hun specifieke taken, die werken voor de algemene winst van een bedrijf. In dit perspectief, kan het concept van human resource management (HRM) ook doorgetrokken worden naar microbiële biotechnologie. Microbial resouce mangagement (MRM) zal daarom streven naar het onderhouden van de best presterende microbiële gemeenschap voor een bepaalde toepassing. Een OLAND MRM kader werd uitgewerkt (Hoofdstuk 1), waarbij getoond werd hoe de OLAND ingenieur/operator (1: input) een microbiële gemeenschap (2: biokatalysator) kan aansturen om zo een optimale functionaliteit (3: output) te bekomen, afhankelijk van het toepassingsgebied (0: afvalwater). Met dit MRM kader in het achterhoofd kan de OLAND ingenieur het OLAND proces aansturen zodat een maximale efficiëntie en hogere duurzaamheid of een grote impact van OLAND op de energiebalans van afvalwaterzuiveringssystemen bekomen kan worden.

185

Samenvatting

Hoewel de eerst volle-schaal OLAND toepassingen een stabiele en efficiënte performantie vertonen, blijft de snelheid waarmee dit proces wordt geïmplementeerd eerder beperkt en is afhankelijk van een handvol bedrijven. Potentiële gebruikers wachten af omdat het lijkt dat dit proces door de lange opstarttijden bij de eerste toepassingen en de gepubliceerde sensitiviteiten veel ervaring vergt. Om dit probleem gedeeltelijk te overbruggen werd de output box van het MRM kader verder in detail bestudeerd voor hoogbelaste stikstofstromen (gekende toepassing). Ten eerste werd het effect van de hydraulische condities op de opstart van OLAND sequencing batch reactoren (SBR) bestudeerd. Een kleine volumetrische uitwisselingsverhouding, welke stabiele hydraulische condities verzekerde, was essentieel om een snelle opstart, granulatie en een hoge performantie in SBR systemen te verkrijgen (Hoofdstuk 2). Verder werden er ook strategieën voorgesteld om een goede microbiële balans te behouden in het OLAND systeem die enerzijds gebaseerd waren op de uitwassing van nitriet-oxiderende bacteriën (NOB) door selectie op bezinkingssnelheid en anderzijds gebaseerd waren op de stimulatie van AnAOB door het invoeren van een anoxische fase. Naast het behalen van een goede effluent kwaliteit, kan ook de algemene duurzaamheid een competitieve factor worden tussen verschillende milieutechnologieën. In deze context werden N2O en NO emissies bestudeerd in een volle-schaal OLAND-type reactor (Hoofdstuk 3). De broeikasgasemissies waren vooral gelinkt aan accumulaties van intermediairen zoals nitriet en hydroxylamine en de frequentie van het opleggen van transiënte condities. Het verder bestuderen van de factoren die leiden tot de accumulatie van intermediairen en het verder optimaliseren van het voedingspatroon, wat de graad van fluctuaties in de reactor bepaalt, zal in de toekomst toelaten om de N2O emissies verder te onderdrukken. In een tweede deel van dit werk werden nieuwe toepassingsmogelijkheden voor OLAND onderzocht die een positieve invloed zouden hebben op de algemene duurzaamheid van systemen (box 0 van MRM kader). Energieberekeningen toonden aan dat de energie-index voor de behandeling van agro-industriële afvalstromen en groente-fruit en tuinafval verhoogd kan worden van 3-5 tot 6-10 door de implementatie van OLAND voor de behandeling van digestaten (Hoofdstuk 4). Echter, OLAND implementatie in de verwerking van mestgebaseerde afvalstromen lijkt een stuk moeilijker, waardoor voor dit toepassingsdomein de OLAND behandeling van ammoniakgasstromen werd voorgesteld. Een pilootschaal OLAND biofilter, gevoed met een ammoniakstroom, behaalde een hoge performantie (0.7 g N L-1 d-1) en een hoge stikstofverwijderingsefficiëntie (75-80%, Hoofdstuk 5). Hoewel de filter gesatureerd was met zuurstof kon de lage relatieve waterstroom (≈1 L g-1 Nin), hoge vrije 186

Samenvatting

ammoniak concentraties in de waterfase verzekeren, welke een dominante activiteit van de AnAOB over de NOB stimuleerden aan de top van de filter.

Uit verdere energieberekeningen bleek dat door de toepassing van OLAND in de hoofstroom van rioolwaterzuiveringsinstallaties (RWZI), een netto energiewinst kon bekomen worden (Hoofdstuk 4). De hoger koolstofrecuperatie en de lagere energiekosten waren hiervoor de belangrijkste oorzaken. Vier uitdagingen die gepaard gingen met de toepassing van OLAND in the hoofdstroom van RWZI werden in dit werk overwonnen. De eerste uitdaging, namelijk OLAND performantie bij lage stikstofconcentraties en lage hydraulische verblijftijd (HRT) werd aangetoond in een OLAND rotating biological contactor (RBC; Hoofdstuk 6). Er werden in deze reactor voor de behandeling van lage stikstofconcentraties (30-60 mg N L-1), hoge stikstofverwijderingssnelheden (0.4 g N L-1 d-1) behaald bij een HRT van 1-2 uur. Een tweede uitdaging, namelijk performantie bij lage temperaturen (15°C) werd overwonnen in dezelfde RBC door een geleidelijke daling van de temperatuur startende van 29°C. Bij operatie op 15°C, een HRT van 1 uur en voeding met synthetisch afvalwater (60 mg N L-1) werden gelijkaardige stikstofverwijderings-snelheden behaald, namelijk 0.5 g N L-1 d-1, ten opzichte van operatie bij hoge temperatuur (Hoofdstuk 7). In vergelijking met de hogere temperaturen, werd een daling in de stikstofverwijderingsefficiëntie van 22% gedetecteerd. Overschakeling van synthetisch naar voorbehandeld rioolwater met een COD/N verhouding van 2 (uitdaging 3) gaf geen significant verschil in de performantie. Echter, bij de lagere temperaturen werd een stijging van de NOB activiteit waargenomen en hierdoor dus ook een grotere competitie tussen AnAOB en NOB voor nitriet (uitdaging 4). Er werd aangetoond dat verhoogde NO concentraties selectief de AnAOB konden stimuleren over de NOB activiteit (Hoofdstuk 7). Piekbelastingen samen met nitrietaccumulatie, welke de NO productie deden stijgen, zorgden dan ook voor een verhoogde verwijderingsefficiëntie. Om de toepassing van hoofdstroom OLAND in een bredere context te beoordelen, werd een levenscyclusanalyse (LCA) uitgevoerd op volle-schaal data van de RWZI in Strass (Oostenrijk). In deze RWZI wordt een OLAND-type reactor, in dit specifieke geval DEMON genoemd, toegepast. Drie scenarios werden onderzocht: (1) de RWZI zonder DEMON; (2) de RWZI met DEMON in de zijstroom; (3) de RWZI met DEMON in zowel zij- als hoofdstroom (Hoofdstuk 8). Voor dit laatste scenario werden volle-schaal data gebruikt van een allereerste poging tot hoofdstroom DEMON in de RWZI in Strass. De LCA toonde aan dat DEMON in de zijstroom van de RWZI een positief effect had op de eutrophication potential, abiotic depletion potential en global warming potential en hierdoor kon zorgen voor een meer duurzame waterzuivering. 187

Samenvatting

Uit de eerste volle-schaal data van hoofstroom DEMON operatie in Strass kon worden geconcludeerd dat om een gelijkaardige graad van duurzaamheid te bekomen, de N2O emissies (nu ongeveer 2% van de N belasting) in de hoofdstoom verminderd dienen te worden tot 0.5% van de stikstofbelasting. Dit aangezien de N2O emissies de dominante factor was in de global warming potential en dus ook de CO2 footprint van de plant. Aangezien N2O emissies vooral gestuurd worden door de operationele condities en niet door het specifieke proces zelf, zou het mogelijk moeten zijn om de emissies verder te optimaliseren en te verminderen zodat eenzelfde CO2 footprint bekomen wordt als bij zijstroom toepassing. In het algemeen toonde dit werk aan dat nieuwe potentiĂŤle toepassingsdomeinen voor OLAND te vinden zijn in landbouw, waar ammoniakbehandeling nodig is en in huishoudelijke waterzuiveringssystemen, door een nieuwe hoofdstroombehandeling. Verdere testen in deze toepassingsdomeinen zijn echter nodig in de toekomst om de performantie en algemene duurzaamheid te evalueren op grotere schaal.

188

Bibliography

Bibliography Abeynayaka, A., Visvanathan, C., 2011. Performance comparison of mesophilic and thermophilic aerobic sidestream membrane bioreactors treating high strength wastewater. Bioresource Technology 102, 5345-5352. Abma, W., Schultz, C., Mulder, J.W., van Loosdrecht, M., van der star, W., Strous, M., Tokutomi, T., 2007. The advance of Anammox. Water21 february, 36-27. Abma, W.R., Driessen, W., Haarhuis, R., van Loosdrecht, M.C.M., 2010. Upgrading of sewage treatment plant by sustainable and cost-effective separate treatment of industrial wastewater. Water Sci. Technol. 61, 1715-1722. Ahn, J.H., Kwan, T., Chandran, K., 2011. Comparison of Partial and Full Nitrification Processes Applied for Treating High-Strength Nitrogen Wastewaters: Microbial Ecology through Nitrous Oxide Production. Environmental science & technology 45, 2734-2740. Alberta Environment, 1999. Wastewater management review for the fertilizer manufacturing sector. Alberta Environment - Environmental Sciences Division. Amann, R.I., Krumholz, L., Stahl, D.A., 1990. Fluorescent-Oligonucleotide Probing of Whole Cells for Determinative, Phylogenetic, and Environmental-Studies in Microbiology. Journal of Bacteriology 172, 762-770. Anthonisen, A.C., Loehr, R.C., Prakasam, T.B.S., Srinath, E.G., 1976. Inhibition of nitrification by ammonia and nitrous acid. Journal Water Pollution Control Federation 48, 835-852. Apodaca, L.E., 2007. Nitrogen. Minerals Yearbook U.S Department of the Interior, U.S. Geological survey, Reston. Baquerizo, G., Maestre, J.P., Machado, V.C., Gamisans, X., Gabriel, D., 2009. Long-term ammonia removal in a coconut fiber-packed biofilter: Analysis of N fractionation and reactor performance under steady-state and transient conditions. Water Research 43, 2293-2301. Barnes, D., Bliss, P.J., 1983. Biological control of nitrogen in wastewater treatment. E. & F.N. Spon, London. Beavers, G.S., Sparrow, E.M., Rodenz, D.E., 1973. Influence of bed size on the flow characteristics and porosity of randomly packed beds of spheres. Transactions of the ASME. Series E, Journal of Applied Mechanics 40. Beier, M., Schneider, Y., 2008. Abschlussbericht Entwickelung von Bilanzmodellen f端r die Prozesse Deammonifikation und Nitritation zur Abbildung gross- and halb-technischer Anlagen (Final report development of balanse models for deammonification and nitritation processes to illustrate full and half technical scale installations). Leibniz University Hannover, Hannover, p. 39. 189

Bibliography

Belser, L.W., Mays, E.L., 1980. Specific Inhibition of Nitrite Oxidation by Chlorate and Its Use in Assessing Nitrification in Soils and Sediments. Appl Environ Microbiol 39, 505-510. Bengtsson, M., Lundin, M., Molander, S., 1997. Life cycle assessment of wastewater systems: case studies of conventional treatment, urine sorting and liquid composting in three Swedish municipalities. Report 1997:9. Chalmers University of Technology, Gothenburg, Sweden. Bernet, N., Dangcong, P., Delgenes, J.P., Moletta, R., 2001. Nitrification at low oxygen concentration in biofilm reactor. Journal of Environmental Engineering-Asce 127, 266-271. BGBL, 1996. Bundesgesetzblatt f r die republik sterreich. Verordnung des Bundesministers f r Land- und Forstwirtschaft ber die Begrenzung von Abwasseremissionen aus Abwasserreinigungsanlagen f r Siedlungsgebiete. Wien, Austria. Bieker, S., Cornel, P., Wagner, M., 2010. Semicentralised supply and treatment systems: integrated infrastructure solutions for fast growing urban areas. Water Science and Technology 61, 2905-2913. Brombach, H., Weiss, G., Fuchs, S., 2005. A new database on urban runoff pollution: comparison of separate and combined sewer systems. Water Sci. Technol. 51, 119-128. Burnley, S.J., 2007. A review of municipal solid waste composition in the United Kingdom. Waste Management 27, 1274-1285. Busca, G., Pistarino, C., 2003. Abatement of ammonia and amines from waste gases: a summary. Journal of Loss Prevention in the Process Industries 16, 157-163. Byrne, N., Strous, M., Crepeau, V., Kartal, B., Birrien, J.-L., Schmid, M., Lesongeur, F., Schouten, S., Jaeschke, A., Jetten, M., Prieur, D., Godfroy, A., 2009. Presence and activity of anaerobic ammonium-oxidizing bacteria at deep-sea hydrothermal vents. Isme Journal 3, 117123. Cabrol, L., 2010. Evaluation de la robustesse d'un systĂ¨me de biofiltration d'effluent de compostage: Approche structurelle et fonctionnelle. University of Montpellier II. Cabrol, L., Malhautier, L., 2011. Integrating microbial ecology in bioprocess understanding: the case of gas biofiltration. Applied Microbiology and Biotechnology 90, 837-849. Cabrol, L., Malhautier, L., Poly, F., Lepeuple, A.-S., Fanlo, J.-L., 2010. Assessing the bias linked to DNA recovery from biofiltration woodchips for microbial community investigation by fingerprinting. Applied Microbiology and Biotechnology 85, 779-790. Cabrol, L., Malhautier, L., Poly, F., Lepeuple, A.S., Fanlo, J.L., 2009. Shock loading in biofilters: impact on biodegradation activity distribution and resilience capacity. Water Science and Technology 59, 1307-1314. Carrere, H., Monlau, F., Barakat, A., Dumas, C., Steyer, J., 2011. Biogas from lignocellulosic biomass: interest of pretreatments. Progress in biogas II, Stuttgart, Germany.

190

Bibliography

Cema, G., Szatkowska, B., Plaza, E., Trela, J., Surmacz-Gorska, J., 2006. Nitrogen removal rates at a technical-scale pilot plant with the one-stage partial nitritation/Anammox process. Water Science and Technology 54, 209-217. Chagas, A.P., 2007. The ammonia synthesis: Some historical aspects. Quimica Nova 30, 204247. Chain, P., Lamerdin, J., Larimer, F., Regala, W., Lao, V., Land, M., Hauser, L., Hooper, A., Klotz, M., Norton, J., Sayavedra-Soto, L., Arciero, D., Hommes, N., Whittaker, M., Arp, D., 2003. Complete genome sequence of the ammonia-oxidizing bacterium and obligate chemolithoautotroph Nitrosomonas europaea. J Bacteriol 185, 2759-2773. Chandran, K., Stein, L.Y., Klotz, M.G., van Loosdrecht, M.C.M., 2011. Nitrous oxide production by lithotrophic ammonia-oxidizing bacteria and implications for engineered nitrogen-removal systems. Biochemical Society Transactions 39, 1832-1837. Chen, Y., Cheng, J.J., Creamer, K.S., 2008. Inhibition of anaerobic digestion process: A review. Bioresource Technology 99, 4044-4064. Chen, Y.X., Yin, J., Wang, K.X., 2005. Long-term operation of biofilters for biological removal of ammonia. Chemosphere 58, 1023-1030. Christensson, M., EkstrĂśm, S., Lemaire, R., Le Vaillant, E., Bundgaard, E., Chauzy, J., Stalhandske, L., Hong, Z., Ekenberg, M., 2011. ANITA Mox - A BioFarm Solution for Fast Start-up of Deammonifying MBBRs. WEFTEC11, Los Angelos. Chung, Y.C., Chung, M.S., 2000. BNP test to evaluate the influence of C/N ratio on N2O production in biological denitrification. Water Sci Technol 42, 23-27. Chung, Y.C., Huang, C.P., Tseng, C.P., 1996. Reduction of H2S/NH3 production from pig feces by controlling environmental conditions. Journal of Environmental Science and Health Part a-Environmental Science and Engineering & Toxic and Hazardous Substance Control 31, 139-155. Clauwaert, P., Roels, J., Thoeye, C., De Gueldre, G., Van De Steene, B., 2010. Evaluatie van de milieuimpact van rioolafvalwaterzuivering met nutriĂŤntenverwijdering met behulp van levenscyclusanalyse (LCA) [Evaluation of the environmental impact of sewage treatment with nutrient removal by means of life cycle analysis (LCA)]. WT-Afvalwater 10, 186-195. Colliver, B.B., Stephenson, T., 2000. Production of nitrogen oxide and dinitrogen oxide by autotrophic nitrifiers. Biotechnol Adv 18, 219-232. Cornel, P., Meda, A., Bieker, S., 2011. Wastewater as a source of energy, nutrients and service water. Treatise on Water Science. Academic Press, Oxford, pp. 337-335. Dalsgaard, T., Thamdrup, B., Canfield, D.E., 2005. Anaerobic ammonium oxidation (anammox) in the marine environment. Res Microbiol 156, 457-464.

191

Bibliography

Dapena-Mora, A., Fernandez, I., Campos, J.L., Mosquera-Corral, A., Mendez, R., Jetten, M.S.M., 2007. Evaluation of activity and inhibition effects on Anammox process by batch tests based on the nitrogen gas production. Enzyme Microb Technol 40, 859-865. De Baere, L., 2006. Will anaerobic digestion of solid waste survive in the future? Water Science and Technology 53, 187-194. De Clippeleir, H., Defoirdt, T., Vanhaecke, L., Vlaeminck, S.E., Carballa, M., Verstraete, W., Boon, N., 2011. Long-chain acylhomoserine lactones increase the anoxic ammonium oxidation rate in an OLAND biofilm. Appl Microbiol Biotechnol 90, 1511-1519. de Graaff, M.S., Zeeman, G., Temmink, H., van Loosdrecht, M.C.M., Buisman, C.J.N., 2010. Long term partial nitritation of anaerobically treated black water and the emission of nitrous oxide. Water Research 44, 2171-2178. de Laclos, H.F., Desbois, S., Saint-Joly, C., 1997. Anaerobic digestion of municipal solid organic waste: Valorga full-scale plant in Tilburg, the Netherlands. Water Science and Technology 36, 457-462. De Sousa, F., Vaz, B., 2009. As caracteristicas da fracção organica dos rsu recolhidos selectivamente na area metropolitana de lisboa e a sua influência no comportamento do processo de digestão anaeróbia. Faculdade de Ciências e Tecnologia. University of Lisbon, Lisbon. Denecke, M., Rekers, V., Walter, U., 2007. Einsparpotenziale bei der biologischen Reinigung von Deponiesickerwasser (Cost saving potentials in the biological treatment of landfill leachate). Muell Abfall 39, 4-7. Desloover, J., De Clippeleir, H., Boeckx, P., Du Laing, G., Colsen, J., Verstraete, W., Vlaeminck, S.E., 2011a. Floc-based sequential partial nitritation and anammox at full scale with contrasting N2O emissions. Water Research 45, 2811-2821. Desloover, J., Vlaeminck, S.E., Verstraete, W., Boon, N., 2011b. Strategies to mitigate N2O emissions from biological nitrogen removal systems. Current Opinion in Biotechnology. Current Opinion in Biotechnology 23, 1-9. Dewettinck, T., Van Houte, E., Geenens, D., Van Hege, K., Verstraete, W., 2001. HACCP (hazard analysis and critical control points) to guarantee safe water reuse and drinking water production. Water Science and Technology 43, 31-38. Doorn, M.R.J., Towprayoon, S., Vieira, S.M.M., Irving, W., Palmer, C., Pipatti, R., Wang, C., 2006. Wastewater treatment and discharge. Institute for Global Environmental Strategies (IGES), Japan. Dosta, J., Fernandez, I., Vazquez-Padin, J.R., Mosquera-Corral, A., Campos, J.L., MataAlvarez, J., Mendez, R., 2008. Short- and long-term effects of temperature on the Anammox process. J Hazard Mater 154, 688-693.

192

Bibliography

Egli, K., Fanger, U., Alvarez, P.J.J., Siegrist, H., van der Meer, J.R., Zehnder, A.J.B., 2001. Enrichment and characterization of an anammox bacterium from a rotating biological contactor treating ammonium-rich leachate. Arch Microbiol 175, 198-207. Erguder, T.H., Boon, N., Vlaeminck, S.E., Verstraete, W., 2008. Partial Nitrification Achieved by Pulse Sulfide Doses in a Sequential Batch Reactor. Environ Sci Technol 42, 8715-8720. European Commision, 1991. Council Directive 91/271/EEC of 21 May 1991 concerning urban waste-water treatment. Official Journal L 135, 40â&#x20AC;&#x201C;52 FAO, Gustavsson, J., Cederberg, C., Sonesson, U., van Otterdijk, R., Meybeek, A., 2011. Global food losses and food waste. Extent, causes and prevention. Food and agriculture organization of the united nations, Rome. Finnveden, G., Hauschild, M.Z., Ekvall, T., Guinee, J., Heijungs, R., Hellweg, S., Koehler, A., Pennington, D., Suh, S., 2009. Recent developments in Life Cycle Assessment. Journal of Environmental Management 91, 1-21. Foley, J., De Haas, D., Hartley, K., Lant, P., 2010. Comprehensive life cycle inventories of alternative wastewater treatment systems. Water research 44, 1654-1666. Frear, D.S., Burrell, R.C., 1955. Spectrophotometric method for determining hydroxylamine reductase activity in higher plants Analytical Chemistry 27, 1664-1665. Friedrich, U., Van Langenhove, H., Altendorf, K., Lipski, A., 2003. Microbial community and physicochemical analysis of an industrial waste gas biofilter and design of 16S rRNAtargeting oligonucleotide probes. Environmental Microbiology 5, 183-201. Furukawa, K., Lieu, P.K., Tokitoh, H., 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., Siegrist, H., 2004. Nitrogen removal from sludge digester liquids by nitrification/denitrification or partial nitritation/anammox: environmental and economical considerations. Water Sci. Technol. 50, 19-26. Galloway, J.N., Dentener, F.J., Capone, D.G., Boyer, E.W., Howarth, R.W., Seitzinger, S.P., Asner, G.P., Cleveland, C.C., Green, P.A., Holland, E.A., Karl, D.M., Michaels, A.F., Porter, J.H., Townsend, A.R., Vorosmarty, C.J., 2004. Nitrogen cycles: past, present, and future. Biogeochemistry 70, 153-226. Galloway, J.N., Townsend, A.R., Erisman, J.W., Bekunda, M., Cai, Z., Freney, J.R., Martinelli, L.A., Seitzinger, S.P., Sutton, M.A., 2008. Transformation of the nitrogen cycle: Recent trends, questions, and potential solutions. Science 320, 889-892. Goreau, T.J., Kaplan, W.A., Wofsy, S.C., McElroy, M.B., Valois, F.W., Watson, S.W., 1980. Production of NO2- and N2O by Nitrifying Bacteria at Reduced Concentrations of Oxygen. Appl Environ Microbiol 40, 526-532.

193

Bibliography

Graham, D.W., Knapp, C.W., Van Vleck, E.S., Bloor, K., Lane, T.B., Graham, C.E., 2007. Experimental demonstration of chaotic instability in biological nitrification. Isme Journal 1, 385-393. Greenberg, A.E., Clesceri, L.S., Eaton, A.D., 1992. Standard Methods for the Examination of Water and Wastewater. American Public Health Association, Washington DC. Griffith, D.R., Barnes, R.T., Raymond, P.A., 2009. Inputs of fossil carbon from wastewater treatment plants to U.S. rivers and oceans. Environmental science & technology 43, 56475651. Grossi, L., 2009. Hydrogen sulfide induces nitric oxide release from nitrite. Bioorg Med Chem Lett 19, 6092-6094. Guinee, J., 2001. Handbook on life cycle assessment - Operational guide to the ISO standards. International Journal of Life Cycle Assessment 6, 255-255. Guo, J., Peng, Y., Huang, H., Wang, S., Ge, S., Zhang, J., Wang, Z., 2010. Short- and longterm effects of temperature on partial nitrification in a sequencing batch reactor treating domestic wastewater. Journal of Hazardous Materials 179, 471-479. G端ven, D., Dapena, A., Kartal, B., Schmid, M.C., Maas, B., van de Pas-Schoonen, K., Sozen, S., Mendez, R., Op den Camp, H.J.M., Jetten, M.S.M., Strous, M., Schmidt, I., 2005. Propionate oxidation by and methanol inhibition of anaerobic ammonium-oxidizing bacteria. Appl Environ Microbiol 71, 1066-1071. Haki, G.D., Rakshit, S.K., 2003. Developments in industrially important thermostable enzymes: a review. Bioresource Technology 89, 17-34. Hatzenpichler, R., Lebedeva, E.V., Spieck, E., Stoecker, K., Richter, A., Daims, H., Wagner, M., 2008. A moderately thermophilic ammonia-oxidizing crenarchaeote from a hot spring. Proceedings of the National Academy of Sciences of the United States of America 105, 21342139. Hellinga, C., Schellen, A., Mulder, J.W., van Loosdrecht, M.C.M., Heijnen, J.J., 1998. The SHARON process: An innovative method for nitrogen removal from ammonium-rich waste water. Water Sci Technol 37, 135-142. Hendrickx, T.L.G., Wang, Y., Kampman, C., Zeeman, G., Temmink, H., Buisman, C.J.N., 2012. Autotrophic nitrogen removal from low strength waste water at low temperature. Water Research 46, 2187-2193. Henze, M., 1997. Waste design for households with respect to water, organics and nutrients. Water Sci Technol 35, 113-120. Henze, M., Van Loosdrecht, M., Ekama, G.A., Brdjanovic, D., 2008. Biological wastewater treatment - Principles, Modelling and Design. IWA Publishing, London.

194

Bibliography

Hermann, B.G., Debeer, L., De Wilde, B., Blok, K., Patel, M.K., 2011. To compost or not to compost: Carbon and energy footprints of biodegradable materials‚Äô waste treatment. Polymer Degradation and Stability 96, 1159-1171. Hernández Leal, L., Temmink, H., Zeeman, G., Buisman, C.J.N., 2010. Bioflocculation of grey water for improved energy recovery within decentralized sanitation concepts. Bioresource Technology 101, 9065-9070. Hippen, A., Helmer, C., Kunst, S., Rosenwinkel, K.H., Seyfried, C.F., 2001. Six years' practical experience with aerobic/anoxic deammonification in biofilm systems. Water Sci Technol 44, 39-48. Hippen, A., Rosenwinkel, K.H., Baumgarten, G., Seyfried, C.F., 1997. Aerobic deammonification: A new experience in the treatment of wastewaters. Water Sci. Technol. 35, 111-120. Hospido, A., Moreira, M.T., Feijoo, G., 2008. A comparison of municipal wastewater treatment plants for big centres of population in Galicia (Spain). International Journal of Life Cycle Assessment 13, 57-64. Hospido, A., Moreira, M.T., Martin, M., Rigola, M., Feijoo, G., 2005. Environmental evaluation of different treatment processes for sludge from urban wastewater treatments: Anaerobic digestion versus thermal processes. International Journal of Life Cycle Assessment 10, 336-345. Hu, Z., Lotti, T., de Kreuk, M., Kleerbezem, R., van Loosdrecht, M., Kruit, J., Jetten, M.S.M., Kartal, B., 2011. Adaptation of anammox bacteria to low temperature in a lab-scale bioreactor. 16th European N-cycle Meeting - 2nd International Conference on Nitrification (ICoN), Nijmegen, The Netherlands. Huisman, J.L., Gasser, T., Gienal, C., Kuhni, M., Krebs, P., Gujer, W., 2004. Quantification of oxygen fluxes in a long gravity sewer. Water Research 38, 1237-1247. Hwang, I.S., Min, K.S., Choi, E., Yun, Z., 2006. Resource recovery and nitrogen removal from piggery waste using the combined anaerobic processes. Water Science and Technology 54, 229-236. IEA, 2010. Key World Energy Statistics. International energy agency. Innerebner, G., Insam, H., Franke-Whittle, I.H., Wett, B., 2007. Identification of anammox bacteria in a full-scale deammonification plant making use of anaerobic ammonia oxidation. Systematic and Applied Microbiology 30, 408-412. International Fertilizer Industry, A., 2012. Fertilizer production and trade statistics. Iso, 2006a. ISO 14040: Environmental Management - Life Cycle Assessment - Principles and Framework. International Organization for Standardization, Geneva, Switzerland. Iso, 2006b. ISO 14044: Environmental Management - Life Cycle Assessment - Requirements and guidelines. International Organization for Standardization, Geneva, Switzerland. 195

Bibliography

Jeanningros, Y., Vlaeminck, S.E., Kaldate, A., Verstraete, W., Graveleau, L., 2010. Fast startup of a pilot-scale deammonification sequencing batch reactor from an activated sludge inoculum. Water Science and Technology 61, 1393-1400. Jetten, M.S.M., Wagner, M., Fuerst, J., van Loosdrecht, M., Kuenen, G., Strous, M., 2001. Microbiology and application of the anaerobic ammonium oxidation ('anammox') process. Current Opinion in Biotechnology 12, 283-288. Joss, A., Derlon, N., Cyprien, C., Burger, S., Szivak, I., Traber, J., Siegrist, H., Morgenroth, E., 2011. Combined Nitritation-Anammox: Advances in Understanding Process Stability. Environmental Science & Technology 45, 9735-9742. Joss, A., Salzgeber, D., Eugster, J., Konig, R., Rottermann, K., Burger, S., Fabijan, P., Leumann, S., Mohn, J., Siegrist, H., 2009. Full-Scale Nitrogen Removal from Digester Liquid with Partial Nitritation and Anammox in One SBR. Environ Sci Technol 43, 5301-5306. Juhler, S., Revsbech, N.P., Schramm, A., Herrmann, M., Ottosen, L.D.M., Nielsen, L.P., 2009. Distribution and Rate of Microbial Processes in an Ammonia-Loaded Air Filter Biofilm. Applied and Environmental Microbiology 75, 3705-3713. Kampschreur, M.J., Poldermans, R., Kleerebezem, R., van der Star, W.R.L., Haarhuis, R., Abma, W.R., Jetten, M.S.M., van Loosdrecht, M.C.M., 2009a. Emission of nitrous oxide and nitric oxide from a full-scale single-stage nitritation-anammox reactor. Water Sci Technol 60, 3211-3217. Kampschreur, M.J., Temmink, H., Kleerebezem, R., Jetten, M.S.M., van Loosdrecht, M.C.M., 2009b. Nitrous oxide emission during wastewater treatment. Water Research 43, 4093-4103. Kampschreur, M.J., van der Star, W.R.L., Wielders, H.A., Mulder, J.W., Jetten, M.S.M., van Loosdrecht, M.C.M., 2008. Dynamics of nitric oxide and nitrous oxide emission during fullscale reject water treatment. Water Res 42, 812-826. Karakashev, D., Schmidt, J.E., Angelidaki, I., 2008. Innovative process scheme for removal of organic matter, phosphorus and nitrogen from pig manure. Water Research 42, 4083-4090. Kartal, B., Kuenen, J.G., van Loosdrecht, M.C.M., 2010a. Sewage Treatment with Anammox. Science 328, 702-703. Kartal, B., Maalcke, W.J., de Almeida, N.M., Cirpus, I., Gloerich, J., Geerts, W., den Camp, H.J.M.O., Harhangi, H.R., Janssen-Megens, E.M., Francoijs, K.-J., Stunnenberg, H.G., Keltjens, J.T., Jetten, M.S.M., Strous, M., 2011. Molecular mechanism of anaerobic ammonium oxidation. Nature 479, 127-U159. Kartal, B., Tan, N.C.G., Van de Biezen, E., Kampschreur, M.J., Van Loosdrecht, M.C.M., Jetten, M.S.M., 2010b. Effect of Nitric Oxide on Anammox Bacteria. Applied and Environmental Microbiology 76, 6304-6306.

196

Bibliography

Kartal, B., van Niftrik, L., Rattray, J., de Vossenberg, J., Schmid, M.C., Damste, J.S.S., Jetten, M.S.M., Strous, M., 2008. Candidatus 'Brocadia fulgida': an autofluorescent anaerobic ammonium oxidizing bacterium. Fems Microbiology Ecology 63, 46-55. Kim, J.H., Rene, E.R., Park, H.S., 2007. Performance of an immobilized cell biofilter for ammonia removal from contaminated air stream. Chemosphere 68, 274-280. Kuai, L.P., Verstraete, W., 1998. Ammonium removal by the oxygen-limited autotrophic nitrification-denitrification system. Applied and Environmental Microbiology 64, 4500-4506. Lackner, S., Terada, A., Smets, B.F., 2008. Heterotrophic activity compromises autotrophic nitrogen removal in membrane-aerated biofilms: Results of a modeling study. Water Res 42, 1102-1112. Lantz, M., Svensson, M., Bjornsson, L., Borjesson, P., 2007. The prospects for an expansion of biogas systems in Sweden - Incentives, barriers and potentials. Energy Policy 35, 18301843. Larsen, T.A., Gujer, W., 1996. Separate management of anthropogenic nutrient solutions (human urine). Water Science and Technology 34, 87-94. Laurino, C.N., Sineriz, F., 1991. Denitrification by thermophilic soil bacteria with ethanol as substrate in a UASB reactor. Biotechnology Letters 13, 299-304. Lazarova, V., Choo, K.-H., Cornel, P., 2012. Meeting the challenges of the water-energy nexus: the role of reuse and wastewater treatment. Water21 April, 12-17. Lebedeva, E.V., Alawi, M., Fiencke, C., Namsaraev, B., Bock, E., Spieck, E., 2005. Moderately thermophilic nitrifying bacteria from a hot spring of the Baikal rift zone. Fems Microbiology Ecology 54, 297-306. Lebedeva, E.V., Off, S., Zumbraegel, S., Kruse, M., Shagzhina, A., Luecker, S., Maixner, F., Lipski, A., Daims, H., Spieck, E., 2011. Isolation and characterization of a moderately thermophilic nitrite-oxidizing bacterium from a geothermal spring. Fems Microbiology Ecology 75, 195-204. Lemaire, R., Webb, R.I., Yuan, Z.G., 2008. Micro-scale observations of the structure of aerobic microbial granules used for the treatment of nutrient-rich industrial wastewater. Isme Journal 2, 528-541. Lemmens, B., Ceulemans, J., Elslander, H., Vanassche, S., Brauns, E., K., V., 2007. Best Beschikbare Technieken (BBT) voor mestverwerking derde editie. VITO. Li, H., Chen, S., Mu, B.-Z., Gu, J.-D., 2011. Molecular Detection of Anaerobic AmmoniumOxidizing (Anammox) Bacteria in High-Temperature Petroleum Reservoirs. Microbial Ecology 60, 771-783. Liu, Y., Wang, Z.W., Qin, L., Liu, Y.Q., Tay, J.H., 2005. Selection pressure-driven aerobic granulation in a sequencing batch reactor. Applied Microbiology and Biotechnology 67, 2632. 197

Bibliography

Loy, A., Horn, M., Wagner, M., 2003. probeBase: an online resource for rRNA-targeted oligonucleotide probes. Nucleic Acids Res. 31, 514-516. Lücker, S., 2010. Exploring the ecology and genomics of globally important nitrite-oxidizing bacteria. Universität Wien, Wien, p. 142. Lücker, S., Wagner, M., Maixner, F., Pelletier, E., Koch, H., Vacherie, B., Rattei, T., Damste, J.S.S., Spieck, E., Le Paslier, D., Daims, H., 2010. A Nitrospira metagenome illuminates the physiology and evolution of globally important nitrite-oxidizing bacteria. Proceedings of the National Academy of Sciences of the United States of America 107, 13479-13484. Maia, G.D.N., Day, G.B., Gates, R.S., Taraba, J.L., 2012. Ammonia biofiltration and nitrous oxide generation during the start-up of gas-phase compost biofilters. Atmospheric Environment 46, 659-664. Malhautier, L., Gracian, C., Roux, J.C., Fanlo, J.L., Le Cloirec, P., 2003. Biological treatment process of air loaded with an ammonia and hydrogen sulfide mixture. Chemosphere 50, 145153. Mancinelli, R.L., McKay, C.P., 1983. Effects of nitric-oxide and nitrogen-dioxide on bacterial-growth. Applied and Environmental Microbiology 46, 198-202. Mateju, V., Cizinska, S., Krejci, J., Janoch, T., 1992. BIOLOGICAL WATER DENITRIFICATION - A REVIEW. Enzyme and Microbial Technology 14, 170-183. Mathanex, 2011. Methanol price http://www.methanex.com/products/methanolprice.html (verified 15/12/2011). Mathure, P., Patwardhan, A., 2005. Comparison of mass transfer efficiency in horizontal rotating packed beds and rotating biological contactors. Journal of Chemical Technology and Biotechnology 80, 413-419. Mehul, P., Dunvin, W., 2010. The groundwater replenishment system: the largest indirect potable reuse plant in the United States. WEFTEC 2010, New Orleans, Louisiana, US. Metcalf, Eddy, 2003. Wastewater engineering: Treatment and reuse. McGraw-Hill, New York. Meulman, B., Elzinga, N., Gorter, K., Zeeman, G., Buisman, C.J.N., Vlaeminck, S.E., Verstraete, W., 2010. Pilot-scale demonstration of sustainable C and N removal from concentrated black water. IWA World Water Congress & Exhibition, Montréal, Canada. Molinuevo, B., Cruz Garcia, M., Karakashev, D., Angelidaki, I., 2009. Anammox for ammonia removal from pig manure effluents: Effect of organic matter content on process performance. Bioresource Technology 100, 2171-2175. Monson, K.D., Esteves, S.R., Guwy, A.J., Dinsdale, R.M., 2007. Ludlow (Greenfinch) Trial Scale Kitchen Waste Treatment Plant. University of Glamorgan.

198

Bibliography

Moussavi, G., Khavanin, A., Sharifi, A., 2011. Ammonia removal from a waste air stream using a biotrickling filter packed with polyurethane foam through the SND process. Bioresource Technology 102, 2517-2522. Mulder, A., 2003. The quest for sustainable nitrogen removal technologies. Water Science and Technology 48, 67-75. Mulder, A., Vandegraaf, A.A., Robertson, L.A., Kuenen, J.G., 1995. Anaerobic ammonium oxidation in a denitrifying fluidized-bed reactor. Fems Microbiology Ecology 16, 177-183. Nielsen, M., Bollmann, A., Sliekers, O., Jetten, M., Schmid, M., Strous, M., Schmidt, I., Larsen, L.H., Nielsen, L.P., Revsbech, N.P., 2005. Kinetics, diffusional limitation and microscale distribution of chemistry and organisms in a CANON reactor. Fems Microbiology Ecology 51, 247-256. Nowak, O., Keil, S., Fimml, C., 2011. Examples of energy self-sufficient municipal nutrient removal plants. Water Science and Technology 64, 1-6. OCWD, 2009. Orange County Water District. Groundwater management plan. Fountain Valley (CA). Oenema, O., 2004. Governmental policies and measures regulating nitrogen and phosphorous from animal manure in European agriculture. Journal of Animal Sciences 82, 196-206. Otte, S., Grobben, N.G., Robertson, L.A., Jetten, M.S.M., Kuenen, J.G., 1996. Nitrous oxide production by Alcaligenes faecalis under transient and dynamic aerobic and anaerobic conditions. Appl Environ Microbiol 62, 2421-2426. Otterpohl, R., 2002. Options for alternative types of sewerage and treatment systems directed to improvement of the overall performance. Water Science and Technology 45, 149-158. Paul, E., Camacho, P., Sperandio, M., Ginestet, P., 2006. Technical and economical evaluation of a thermal, and two oxidative techniques for the reduction of excess sludge production. Process Safety and Environmental Protection 84, 247-252. Pellicer-Nacher, C., Sun, S., Lackner, S., Terada, A., Schreiber, F., Zhou, Q., Smets, B.F., 2010a. Sequential Aeration of Membrane-Aerated Biofilm Reactors for High-Rate Autotrophic Nitrogen Removal: Experimental Demonstration. Environmental science & technology 44, 7628-7634. Perez, J., Picioreanu, C., van Loosdrecht, M., 2005. Modeling biofilm and floc diffusion processes based on analytical solution of reaction-diffusion equations. Water Research 39, 1311-1323. Phyllis-Database, 2012. Phyllis, Database for Biomass and Waste, Http://www.ecn.nl/phyllis Energy Research Centre of the Netherland, n.d. http://www.ecn.nl/phyllis/ (verfied 2012/05/26). Poth, M., Focht, D.D., 1985. 15N Kinetic Analysis of N2O Production by Nitrosomonas europaea: an Examination of Nitrifier Denitrification. Appl Environ Microbiol 49, 11341141. 199

Bibliography

PUB, 2010. Public Utilities Board. NEWater. http://www.pub.gov.sg/water/Pages/NEWater.aspx. (verified 2012/2006/2006).

Pynaert, K., Smets, B.F., Beheydt, D., Verstraete, W., 2004. Start-up of autotrophic nitrogen removal reactors via sequential biocatalyst addition. Environmental Science & Technology 38, 1228-1235. Pynaert, K., Smets, B.F., Wyffels, S., Beheydt, D., Siciliano, S.D., Verstraete, W., 2003. Characterization of an autotrophic nitrogen-removing biofilm from a highly loaded lab-scale rotating biological contactor. Applied and Environmental Microbiology 69, 3626-3635. Ritchie, G.A.F., Nicholas, D.J.D., 1972. Identification of the Sources of Nitrous Oxide Produced by Oxidative and Reductive Processes in Nitrosomonas europaea. Biochem J 126, 1181. Rotthauwe, J.H., Witzel, K.P., Liesack, W., 1997. The ammonia monooxygenase structural gene amoA as a functional marker: Molecular fine-scale analysis of natural ammoniaoxidizing populations. Applied and Environmental Microbiology 63, 4704-4712. Sakano, Y., Kerkhof, L., 1998. Assessment of changes in microbial community structure during operation of an ammonia biofilter with molecular tools. Applied and Environmental Microbiology 64, 4877-4882. Sakuma, T., Jinsiriwanit, S., Hattori, T., Deshusses, M.A., 2008. Removal of ammonia from contaminated air in a biotrickling filter - Denitrifying bioreactor combination system. Water Research 42, 4507-4513. SalomĂŠ, A.A., 1990. AB-systemen: een inventarisatie. Lelystad : Rijkswaterstaat, DBW/RIZA. Santoro, A.E., Buchwald, C., McIlvin, M.R., Casciotti, K.L., in press. Isotopic Signature of N2O Produced by Marine Ammonia-Oxidizing Archaea. Science DOI: 10.1126/science.1208239. Schmid, M., Walsh, K., Webb, R., Rijpstra, W.I.C., van de Pas-Schoonen, K., Verbruggen, M.J., Hill, T., Moffett, B., Fuerst, J., Schouten, S., Damste, J.S.S., Harris, J., Shaw, P., Jetten, M., 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. Schmidt, I., van Spanning, R.J.M., Jetten, M.S.M., 2004. Denitrification and ammonia oxidation by Nitrosomonas europaea wild-type, and NirK- and NorB-deficient mutants. Microbiology-Sgm 150, 4107-4114. Shimaya, C., Hashimoto, T., 2011. Isolation and characterization of novel thermophilic nitrifying Bacillus sp. from compost. Soil Science and Plant Nutrition 57, 150-156. Shizas, I., Bagley, D.M., 2004. Experimental determination of energy content of unknown organics in municipal wastewater streams. Journal of Energy Engineering-Asce 130, 45-53. 200

Bibliography

Shore, J.L., M'Coy, W.S., Gunsch, C.K., Deshusses, M.A., 2012. Application of a moving bed biofilm reactor for tertiary ammonia treatment in high temperature industrial wastewater. Bioresource Technology 112, 51-60. Siegrist, H., 1996. Nitrogen removal from digester supernatant - Comparison of chemical and biological methods. Water Science and Technology 34, 399-406. Siegrist, H., Reithaar, S., Koch, G., Lais, P., 1998. Nitrogen loss in a nitrifying rotating contactor treating ammonium-rich wastewater without organic carbon. Water Science and Technology 38, 241-248. Siegrist, H., Salzgeber, D., Eugster, J., Joss, A., 2008. Anammox brings WWTP closer to energy autarky due to increased biogas production and reduced aeration energy for Nremoval. Water Science and Technology 57, 383-388. Silva, G.A., Kulay, L.A., 2003. Application of life cycle assessment to the LCA case studies single superphosphate production. International Journal of Life Cycle Assessment 8, 209-214. Six, W., Debaere, L., 1992. Dry anaerobic conversion of municipal solid-waste by means of the DRANCO process. Water Science and Technology 25, 295-300. Sliekers, A.O., Derwort, N., Gomez, J.L.C., Strous, M., Kuenen, J.G., Jetten, M.S.M., 2002. Completely autotrophic nitrogen removal over nitrite in one single reactor. Water Research 36, 2475-2482. Sliekers, A.O., Third, K.A., Abma, W., Kuenen, J.G., Jetten, M.S.M., 2003. CANON and Anammox in a gas-lift reactor. Fems Microbiology Letters 218, 339-344. Smet, E., Van Langenhove, H., Maes, K., 2000. Abatement of high concentrated ammonia loaded waste gases in compost biofilters. Water Air and Soil Pollution 119, 177-190. Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L., 2007. Climate Change 2007: The Physical Science Basis. Cambridge University Press, Cambridge. Starkenburg, S.R., Arp, D.J., Bottomley, P.J., 2008. Expression of a putative nitrite reductase and the reversible inhibition of nitrite-dependent respiration by nitric oxide in Nitrobacter winogradskyi Nb-255. Environmental Microbiology 10, 3036-3042. Strous, M., Heijnen, J.J., Kuenen, J.G., 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., Jetten, M.S.M., 1999. Key physiology of anaerobic ammonium oxidation. Appl Environ Microbiol 65, 3248-3250. Strous, M., van Gerven, E., Kuenen, J.G., Jetten, M., 1997. Effects of aerobic and microaerobic conditions on anaerobic ammonium-oxidizing (Anammox) sludge. Appl Environ Microbiol 63, 2446-2448. 201

Bibliography

Suh, Y.J., Rousseaux, P., 2002. An LCA of alternative wastewater sludge treatment scenarios. Resources Conservation and Recycling 35, 191-200. Suvilampi, J., Lepisto, R., Rintala, J., 2001. Biological treatment of pulp and paper mill process and wastewaters under thermophilic conditions - a review. Paperi Ja Puu-Paper and Timber 83, 320-325. Swiss Centre for Life Cycle, I., 2010. Ecoinvent database version 2.2. www.ecoinvent.org. Szatkowska, B., Cema, G., Plaza, E., Trela, J., Hultman, B., 2007. A one-stage system with partial nitritation and Anammox processes in the moving-bed biofilm reactor. Water Sci. Technol. 55, 19-26. Tchobanoglous, G., Burton, F.L., Stensel, H.D., 2003. Wastewater engineering, treatment and reuse - Metcalf & Eddy. McGraw-Hill, New York. Third, K.A., Sliekers, A.O., Kuenen, J.G., Jetten, M.S.M., 2001. The CANON system (completely autotrophic nitrogen-removal over nitrite) under ammonium limitation: Interaction and competition between three groups of bacteria. Systematic and Applied Microbiology 24, 588-596. Toh, S.K., Ashbolt, N.J., 2002. Adaptation of anaerobic ammonium-oxidising consortium to synthetic coke-ovens wastewater. Applied Microbiology and Biotechnology 59, 344-352. Tokutomi, T., Shibayama, C., Soda, S., Ike, M., 2011a. A novel control method for nitritation: The domination of ammonia-oxidizing bacteria by high concentrations of inorganic carbon in an airlift-fluidized bed reactor. Water Research 44, 4195-4203. Tokutomi, T., Yamauchi, H., Nishimura, S., Yoda, M., Abma, W., 2011b. Application of the Nitritation and Anammox Process into Inorganic Nitrogenous Wastewater from Semiconductor Factory. Journal of Environmental Engineering-Asce 137, 146-154. Tourna, M., Freitag, T.E., Nicol, G.W., Prosser, J.I., 2008. Growth, activity and temperature responses of ammonia-oxidizing archaea and bacteria in soil microcosms. Environmental Microbiology 10, 1357-1364. Tsushima, I., Kindaichi, T., Okabe, S., 2007a. Quantification of anaerobic ammoniumoxidizing bacteria in enrichment cultures by real-time PCR. Water Research 41, 785-794. Tsushima, I., Ogasawara, Y., Kindaichi, T., Satoh, H., Okabe, S., 2007b. Development of high-rate anaerobic ammonium-oxidizing (anammox) biofilm reactors. Water Research 41, 1623-1634. UWWTD, 2011. Waterbase-UWWTD: Urban waste water treatment directive. In: Agency, E.E. (Ed.). European Environmental Agency. van Cleemput, O., 1998. Subsoils: chemo- and biological denitrification, N2O and N2 emissions. Nutr Cycl Agroecosyst 52, 187-194.

202

Bibliography

van de Graaf, A.A.V., deBruijn, P., Robertson, L.A., Jetten, M.S.M., Kuenen, J.G., 1996. Autotrophic growth of anaerobic ammonium-oxidizing micro-organisms in a fluidized bed reactor. Microbiology-Uk 142, 2187-2196. van der Star, W.R.L., Abma, W.R., Blommers, D., Mulder, J.W., Tokutomi, T., Strous, M., Picioreanu, C., Van Loosdrecht, M.C.M., 2007. Startup of reactors for anoxic ammonium oxidation: Experiences from the first full-scale anammox reactor in Rotterdam. Water Research 41, 4149-4163. Van Ewijk, H.A.L., 2008. Milieuanalyse vergisten GFT-afval. Vereniging Afvalbedrijven, Nederland. van Haandel, A., van der Lubbe, J., 2007. Handbook Biological waste water treatment Design and optimisation of activated sludge systems. Quist publishing Leidschendam, the Netherlands. Van Hulle, S.W.H., Vandeweyer, H.J.P., Meesschaert, B.D., Vanrolleghem, P.A., Dejans, P., Dumoulin, A., 2010. Engineering aspects and practical application of autotrophic nitrogen removal from nitrogen rich streams. Chemical Engineering Journal 162, 1-20. Vandevivere, P., De Baere, L., Verstraete, W., 2002. Types of anaerobic digester for solid waste. In: Mata-Alvarez, J. (Ed.). Biomethanization of the Organic Fraction of Municipal Solid Wastes. Iwa Publishing, London, pp. 336-367. Vaz, F., Torres, A., Correia, C.N., 2008. Case study: the characteristics of the biodegradable waste for the anaerobic digestion plant in Lisbon area. Water Science and Technology 58, 1563-1568. Vazquez-Padin, J.R., Fernandez, I., Morales, N., Campos, J.L., Mosquera-Corral, A., Mendez, R., 2011. Autotrophic nitrogen removal at low temperature. Water Science and Technology 63, 1282-1288. Vermeiren, J., Van de Wiele, T., Van Nieuwenhuyse, G., Boeckx, P., Verstraete, W., Boon, N., 2012. Sulfide- and Nitrite- Dependent Nitric Oxide Production in the Intestinal Tract. Microbial Biotechnology 5, 379-387. Verstraete, W., 2007. Microbial ecology and environmental biotechnology. Isme Journal 1, 48. Verstraete, W., de Caveye, P.V., Diamantis, V., 2009. Maximum use of resources present in domestic "used water". Bioresour Technol 100, 5537-5545. Verstraete, W., Vlaeminck, S.E., 2011. ZeroWasteWater: Short-cycling of Wastewater Resources for Sustainable Cities of the Future. International Journal of Sustainable Development & World Ecology 18, 253-264. Verstraete, W., Wittelbolle, L., Heylen, K., Vanparys, B., de Vos, P., van de Wiele, T., Boon, N., 2007. Microbial resource management: The road to go for environmental biotechnology. Engineering in Life Sciences 7, 117-126.

203

Bibliography

Verwin, 2006. Reflections on performance. Verwin no. 2007/80/6292. Den Haag, The Netherlands, pp. http://www.verwin.nl/SiteCollectionDocuments/Publicaties/Overige%20VerwinUitgaven/2007/Reflections%2020on%performance%202006.pdf. Vlaeminck, S.E., Cloetens, L.F.F., De Clippeleir, H., Carballa, M., Boon, N., Verstraete, W., 2009a. Granular biomass capable of partial nitritation and anammox. Water Science and Technology 59, 609-617. Vlaeminck, S.E., De Clippeleir, H., Verstraete, W., 2012. Microbial resource management of one-stage partial nitritation/anammox. Microbial Biotechnology 5, 433-488. Vlaeminck, S.E., Geets, J., Vervaeren, H., Boon, N., Verstraete, W., 2007. Reactivation of aerobic and anaerobic ammonium oxidizers in OLAND biomass after long-term storage. Applied Microbiology and Biotechnology 74, 1376-1384. Vlaeminck, S.E., Hay, A., Maignien, L., Verstraete, W., 2011. In quest of the nitrogen oxidizing prokaryotes of the early Earth. Environmental Microbiology 13, 283-295. Vlaeminck, S.E., Terada, A., Smets, B.F., De Clippeleir, H., Schaubroeck, T., Bolca, S., Demeestere, L., Mast, J., Boon, N., Carballa, M., Verstraete, W., 2010. Aggregate Size and Architecture Determine Microbial Activity Balance for One-Stage Partial Nitritation and Anammox. Applied and Environmental Microbiology 76, 900-909. Vlaeminck, S.E., Terada, A., Smets, B.F., Van der Linden, D., Boon, N., Verstraete, W., Carballa, M., 2009b. Nitrogen Removal from Digested Black Water by One-Stage Partial Nitritation and Anammox. Environ Sci Technol 43, 5035-5041. Volcke, E.I.P., Picioreanu, C., De Baets, B., van Loosdrecht, M.C.M., 2010. Effect of granule size on autotrophic nitrogen removal in a granular sludge reactor. Environ Technol 31, 12711280. von Keitz, V., Schramm, A., Altendorf, K., Lipski, A., 1999. Characterization of microbial communities of biofilters by phospholipid fatty acid analysis and rRNA targeted oligonucleotide probes. Systematic and Applied Microbiology 22, 626-634. von Schulthess, R., Gujer, W., 1996. Release of nitrous oxide (N2O) from denitrifying activated sludge: Verification and application of a mathematical model. Water Res 30, 521530. von Schulthess, R., K端hni, M., Gujer, W., 1995. Release of nitric and nitrous oxides from denitrifying activated sludge. Water Res 29, 215-226. Voutchkov, N., 2010. Seawater desalination: current trends and challenges. Desalination - a Filtration + Separation Publication 5, 4-7. Weissenbacher, N., Loderer, C., Lenz, K., Mahnik, S.N., Wett, B., Fuerhacker, M., 2007. NO(x), monitoring of a simultaneous nitrifying-denitrifying (SND) activated sludge plant at different oxidation reduction potentials. Water Research 41, 397-405.

204

Bibliography

Weissenbacher, N., Takacs, I., Murthy, S., Fuerhacker, M., Wett, B., 2010. Gaseous Nitrogen and Carbon Emissions from a Full-Scale Deammonification Plant. Water Environ Res 82, 169-175. Wellinger, A., Wyder, K., Metzler, A.E., 1993. KOMPAGAS - A new system for the anearobic treatment of source separated waste. Water Science and Technology 27, 153-158. Wett, B., 2006. Solved upscaling problems for implementing deammonification of rejection water. Water Sci Technol 53, 121-128. Wett, B., Buchauer, K., Fimml, C., 2007. Energy self-sufficiency as a feasible concept for wastewater treatment systems. IWA Leading Edge Technology conference, Singapore. Wett, B., Hell, M., Nyhuis, G., Puempel, T., Takacs, I., Murthy, S., 2010a. Syntrophy of aerobic and anaerobic ammonia oxidisers. Water Sci Technol 61, 1915-1922. Wett, B., Nyhuis, G., Hell, M., Takacs, I., Murthy, S., 2010b. Development of enhanced deammonification selector. WEFTEC10, New Orleans. Wett, B., Podmirseg, S.M., Hell, M., Nyhuis, G., Bott, C., Murthy, S., 2012. Expanding DEMON sidestream deammonification technology towards mainstream application. SIDISA, Milano, Italy. Wiegel, J., Ljungdahl, L.G., 1986. The importance of thermophilic bacteria in biotechnology Crc Critical Reviews in Biotechnology 3, 39-108. Winkler, M.K.H., Kleerebezem, R., Kuenen, J.G., Yang, J., van Loosdrecht, M.C.M., 2011. Segregation of biomass in cyclic anaerobic/aerobic granular sludge allows the enrichment of anaerobic ammonium oxidizing bacteria. 1st international symposium on Microbial resource management in biotechnology: Concepts & applications, Gent, Belgium. WRS, 2001. Water revival systems Uppsala AB. Market survey - Extremely low flush toilets, plus urine diverting toilets and urinals, for collection of black water and/or urine. SwedEnviro Report 20001:1. Available from http://www.swedenviro.se/pdf/engmarknadsoversikt/PDF. Xia, S.Q., Yang, D.H., Xu, B., Zhao, J.F., 2005. Chemical and biological flocculation process to treat municipal sewage and analysis of biological, function. Journal of Environmental Sciences-China 17, 163-167. Yasuda, T., Kuroda, K., Hanajima, D., Fukumoto, Y., Waki, M., Suzuki, K., 2010. Characteristics of the Microbial Community Associated with Ammonia Oxidation in a FullScale Rockwool Biofilter Treating Malodors from Livestock Manure Composting. Microbes and Environments 25, 111-119. Yin, J., Xu, W., 2009. Ammonia biofiltration and community analysis of ammonia-oxidizing bacteria in biofilters. Bioresource Technology 100, 3869-3876. Yu, R., Kampschreur, M.J., van Loosdrecht, M.C.M., Chandran, K., 2010. Mechanisms and Specific Directionality of Autotrophic Nitrous Oxide and Nitric Oxide Generation during Transient Anoxia. Environ Sci Technol 44, 1313-1319. 205

Bibliography

Zart, D., Schmidt, I., Bock, E., 2000. Significance of gaseous NO for ammonia oxidation by Nitrosomonas eutropha. Antonie Van Leeuwenhoek International Journal of General and Molecular Microbiology 77, 49-55. Zeeman, G., Kujawa, K., de Mes, T., Hernandez, L., de Graaff, M., Abu-Ghunmi, L., Mels, A., Meulman, B., Temmink, H., Buisman, C., van Lier, J., Lettinga, G., 2008. Anaerobic treatment as a core technology for energy, nutrients and water recovery from source-separated domestic waste(water). Water Science and Technology 57, 1207-1212. Zessner, M., Lampert, C., Kroiss, H., Lindtner, S., 2010. Cost comparison of wastewater in Danubian countries. Water Science and Technology 62, 223-230. Zhang, L., Zheng, P., Tang, C.-j., Jin, R.-c., 2008. Anaerobic ammonium oxidation for treatment of ammonium-rich wastewaters. Journal of Zhejiang University-Science B 9, 416426. Zhang, R., El-Mashad, H.M., Hartman, K., Wang, F., Liu, G., Choate, C., Gamble, P., 2007. Characterization of food waste as feedstock for anaerobic digestion. Bioresource Technology 98, 929-935. Zubrowska-Sudol, M., Yang, J., Trela, J., Plaza, E., 2011. Evaluation of deammonification process performance at different aeration strategies. Water Science and Technology 63, 11681176.

206

Curriculum vitae

Curriculum vitae Personal information

Full name:

Haydée De Clippeleir

Date of birth:

6th May 1985

Place of birth:

Sint-Niklaas, Belgium

Nationality:

Belgian

Adres:

Schaubeke 19a, 9220 Hamme, Belgium

Phone:

+32 473 68 65 88

Email:

Haydee.DeClippeleir@UGent.be Haydee.dc@telenet.be

Education

2008-now:

Ph.D. in Applied Biological Sciences (option environmental technologies) (LabMET, Ghent University) Doctoral schools of engineering – Ghent University  Funding: Institute for the Promotion of Innovation through Science and Technology in Flanders (IWT-Vlaanderen)  Ph.D. thesis: Microbial resource management of OLAND focused on sustainability  Promotors: Prof. dr. ir. Willy Verstraete and Prof. dr. ir. Nico Boon

2003-2008:

Bioscience engineer in Environmental technology (Master) Faculty of Bioscience engineering – Ghent University Graduated with great distinction  Master thesis: Technological and microbial aspects of the OLAND process  Promotor: Prof. dr. ir. Willy Verstraete  Training period: Svartsjöprojektet (Hultsfred, Zweden) for DEC nv.

1997-2003:

Science – Mathematics (8h) Sint-Vincentius instituut, Dendermonde

207

Curriculum vitae

Professional activities

2008-now:

Scientific collaborator at Laboratory for microbial ecology and technology (LabMET) Contact: Coupure Links 653, 9000 Gent; phone +32(0)92645976

Coordinator of PC exercises for the courses: ‘Environmental biotechnology’, ‘Microbial ecological processes’ and ‘Re-use Technologies’

Tutor of 7 master students

Collaborations and contributions to:  CO project: aerobic granular sludge technology (Paul Ockier, TNAV)  Project: ‘Analyse des Einflusses der auptstrom -Deammonifikation auf die flüssigen und gasförmigen Emissionen kommunaler Kläranlagen in Österreich’, in collaboration with Norbert Wiessenbacher (BOKU, Vienna, Austria) and Bernhard Wett (ARAconsult, Innsbruck, Austria)

Scientific contributions

A1 publications

De Clippeleir H., Schaubroeck S., Weisssenbacher N., Dewulf J., Boeckx P., Boon N. and Wett B. Environmental assessment of one-stage nitritation/anammox implementation in sewage treatment plants. Submitted.

De Clippeleir H., Vlaeminck S.E., De Wilde F., Daeninck K., Mosquera M., Boeckx P., Verstraete W. and Boon N.. Cold one-stage partial nitritation/anammox on pretreated sewage: feasibility demonstration at lab-scale. Submitted.

De Clippeleir H., Courtens E., Vlaeminck S.E., Boon N. and Verstraete W. Efficient total nitrogen removal in an ammonia gas biofilter through high-rate OLAND. Environmental Science and Technology, 46(16), 8826-8833.

208

Curriculum vitae Vlaeminck, S.E., De Clippeleir, H. and Verstraete, W. Microbial resource management of one-stage partial nitritation/anammox. Microbial Biotechnology, 5(3), 433-488.

Schaubroeck, T., Bagchi, S., De Clippeleir, H., Carballa, M., Boon, N., Verstraete, W. & Vlaeminck S.E. Successful hydraulic strategies to start up OLAND sequencing batch reactors at lab scale. Microbial Biotechnology, 5(3), 403-414.

De Clippeleir, H., Yan, X., Verstraete, W., Vlaeminck, S.E., 2011. OLAND is feasible to treat sewage-like nitrogen concentrations at low hydraulic residence times. Applied Microbiology and Biotechnology, 90, 1537-1545.

De Clippeleir, H., Defoirdt, T., Vanhaecke, L., Vlaeminck, S.E., Carballa, M., Verstraete, W., Boon, N., 2011. Long-chain acylhomoserine lactones increase the anoxic ammonium oxidation rate in an OLAND biofilm. Applied Microbiology and Biotechnology, 90, 1511-1519.

Desloover, J., De Clippeleir, H., Boeckx, P., Du Laing, G., Colsen, J., Verstraete, W., Vlaeminck, S.E., 2011. Floc-based sequential partial nitritation and anammox at full scale with contrasting N2O emissions. Water Research, 45, 2811-2821.

Vlaeminck, S.E., Terada, A., Smets, B.F., De Clippeleir, H., Schaubroeck, T., Bolca, S., Demeestere, L., Mast, J., Boon, N., Carballa, M., Verstraete, W., 2010. Aggregate size and architecture determine microbial activity balance for one-stage partial nitritation and anammox. Applied and Environmental Microbiology, 76, 900-909.

De Clippeleir, H., Vlaeminck, S.E., Carballa, M. & Verstraete, W. (2009). A low volumetric exchange ratio allows high autotrophic nitrogen removal in a sequencing batch reactor. Bioresource Technology, 100, 5010-5015.

Vlaeminck, S.E., Cloetens, L.F.F., De Clippeleir, H., Carballa, M. & Verstraete, W. (2009). Granular biomass capable of partial nitritation and anammox. Water Science and Technology, 59(3), 609-617.

209

Curriculum vitae Other publications

De Clippeleir H., Vlaeminck S.E., Courtens E., Verstraete W. and Boon N. (2012) Behandeling van anaerobe digestaten met OLAND maximaliseert de elektrische netto-energiewinst. WT-afvalwater, 2, 137-153.

Weissenbacher N., De Clippeleir H., Hell M. and Wett B. Lachgasemissionen bei der behandlung von prozesswĂ¤ssern im deammonificationsverfahren. Ă&#x2013;sterreichische Wasser-und Abfallwirtschaft, 64(12), 247-252.

De Clippeleir H., Vlaeminck S.E., Courtens E., Verstraete W. and Boon N. (2012) Oxygen-limited autotrophic nitrification/denitrification maximizes net energy gain in technology schemes with anaerobic digestion, In Renewable Energy Sources, Academy Publish: Wyoming, U.S.A., accepted.

Contributions to conferences, symposia, workshops and seminars

De Clippeleir H., Courtens E., Vlaeminck S.E., Boon N. and Verstraete W. Efficient total nitrogen removal in an ammonia gas biofilter through high rate OLAND. Ecotechnolgies for wastewater treatment, Santiago de Compostela, Spain, 25-27 June 2012. Oral presenation

De Clippeleir H., Weissenbacher N., Boeckx P., Chandran K., Boon N. and Wett B. 2012. Interplay of intermediates in the formation of NO and N2O during full-scale partial nitritation/anammox. Ecotechnologies for wastewater treatment, Santiago de Compostela, Spain, 25-27 June 2012. Oral Presentation

De Clippeleir H., Vlaeminck S.E., Courtens E., Verstraete W. and Boon N. Oxygen-limited autotrophic nitrification/denitrification maximizes net energy gain in technology schemes with anaerobic digestion Leading Edge Technology 2012, Brisbane, Australia, 4-7th June 2012. Poster presentation

De Clippeleir H., Courtens E., Vlaeminck S.E., Boon N. and Verstraete W. A high-rate ammonia gas biofilter based on partial nitritation/anammox removes total nitrogen at high efficiency. 17th PhD symposium, 10 February 2012. Oral presenation

210

Curriculum vitae Weissenbacher, N., De Clippeleir, H., Boeckx, P., Hell, M., Chandran, K., Murthy, S. and Wett, B. Control of N2O-emissions from Sidestream Treatment. WEFTEC, Los Angeles, 15-19 October 2011. Oral presentation (co-author)

De Clippeleir, H. Vlaeminck, S.E., Van Acker, J., Boon, N. and Verstraete, W. An oxygen-limited batch test as experimental model for OLAND application screening and scenario analysis. IWA young water professionals, 20-22 September 2011. Oral presentation

De Clippeleir, H., Yan, Y., Verstraete, W. and Vlaeminck, S.E. OLAND is feasible to treat sewagelike nitrogen concentrations at low hydraulic residence time. MRM symposium, Gent, 30the June 2011. Poster presentation

De Clippeleir, H., Yan, X., Verstraete, W. and Vlaeminck, S.E. Approaching energy-positive sewage treatment: OLAND removes nitrogen from low-strength wastewater. Nutrient Recovery and Management Conference, Inside and Outside the Fence, Miami, 9-12 January 2011. Oral presentation

De Clippeleir, H., Yan, Y., Verstraete, W. and Vlaeminck, S.E. OLAND is feasible to treat sewagelike nitrogen concentrations at low hydraulic residence time, 16th PhD symposium Applied biological sciences, 20 December 2010. Oral presentation

De Clippeleir, H., Defoirdt, T., Vanhaecke, L., Vlaeminck, S.E., Carballa, M., Verstraete, W., Boon, N. Long-chain acylhomoserine lactones increase the anoxic ammonium oxidation rate in an OLAND biofilm. 16th PhD symposium Applied biological sciences, 20 December 2010. Poster presentation

De Clippeleir, H., Defoirdt, T., Vanhaecke, L., Vlaeminck, S.E., Carballa, M., Verstraete, W. and Boon, N. Long-chain acylhomoserine lactones increase the anoxic ammonium oxidation rate in an OLAND biofilm. Conference ISME13: Microbe â&#x20AC;&#x201C; stewards of a changing planet, Seatlle, 22the August 2010, poster presentation

211

Curriculum vitae De Clippeleir H., Defoirdt T., Vanhaecke L., Vlaeminck S.E., Carballa M., Verstraete W. and Nico Boon (2010). Long-chain acylhomoserine lactones increase the anoxic ammonium oxidation rate in an OLAND biofilm. Workshop on bacterial and fungal biofilms, Leuven, Belgium, 25 May 2010. Oral presenatation

De Clippeleir, H., Vlaeminck, S.E., Carballa, M. and Verstraete W. High and stable autotrophic nitrogen removal in a sequencing batch reactor by applying a low volumetric exchange ratio. IWA 2nd Specialized Conference on Nutrient Management in Wastewater Treatment Processes, Krakow, 6-9 September 2009. Oral presentation

Vlaeminck, S.E., Carballa, M., De Clippeleir, H. and Verstraete, W. Biofilm and granule applications for one-stage autotrophic nitrogen removal. Seminar Nederlandse Biotechnologische Vereniging and UNESCO-ISHE on nitrogen removal and recovery from water and wastewater. Delft, 26 March 2009. Oral presentation (co-author)

Vlaeminck, S.E., Terada, A., Carballa, M., De Clippeleir, H., Boon, N., Smets, B.F. and Verstraete, W. Fluorescence in situ hybridization (FISH) to elucidate structure and diversity in granular biomass for the treatment of nitrogenous wastewater. 14the Symposium on Applied Biological Sciences, Ghent, 15 September 2008. Oral presentation (co-author)

Vlaeminck, S.E., De Clippeleir, H., Carballa, M., Terada, A., Smets, B.F. and Verstraete, W. Granular biomass capable of partial nitritation and anammox. IWA World Water Congress and Exhibition. Vienna, 7-12 September 2008. Oral presentation (co-author)

Vlaeminck, S.E., Cloetens, L.F.F., De Clippeleir, H., Carballa, M., Smets, B.F. and Verstraete, W. Granular biomass capable of combined aerobic and anoxic ammonium oxidation. Fall symposium of the ‘Nationale Vereniging voor Microbiologie (NVvM) - Microbiële Ecologie’. Amsterdam, 23 November 2007. Oral presentation (co-author)

Awards

Best platform presentation with “Interplay of intermediates in the formation of NO and N2O during full-scale partial nitritation/anammox.” Ecotechnologies for wastewater treatment, Santiago de Compostela, Spain, June 27th 2012.

212

Curriculum vitae Best short presentation with â&#x20AC;&#x153;Efficient total nitrogen removal in an ammonia gas biofilter through high rate OLAND.â&#x20AC;? Ecotechnolgies for wastewater treatment, Santiago de Compostela, Spain, June 27th 2012.

213

Curriculum vitae

214

Dankwoord

Dankwoord “Doctoren is zoals een olympische discipline” werd me bij de start van mijn onderzoek medegedeeld. Nu ik terugblik, kan ik toch menig gelijkenis erkennen. Zoals atleten die toewerken naar de olympische spelen, werd ook tijdens dit onderzoek met vallen en opstaan gezocht naar de beste techniek en tactiek om uiteindelijk dit ene einddoel, het afleggen van een doctoraat, te behalen. Er werd gezweet (letterlijk en figuurlijk), gejuicht maar er werden ook tegenslagen verwerkt. Echter, niets kon tot stand komen zonder een fantastisch team en een enthousiaste achterban.

First of all I would like to thank the jury members, it is really an honor to defend my work in front of such prominent group of scientists. Your thorough examination of this work and the doubtlessly critical questions are greatly appreciated.

Mijn coaches, promotoren, Willy Verstraete en Nico Boon ben ik dankbaar omdat ze me de kans gaven bij LabMET te werken. Prof. Verstraete, jouw energie en enthousiasme werkten steeds aanstekelijk. Ik wil je bedanken om me steeds verder te pushen, maar ook voor de vrijheid dit u me gegeven hebt om achter ideeën aan te gaan. Nico, ik wil je bedanken voor het vertrouwen dat je altijd in me had. De korte, maar efficiënte discussies die we hadden waren steeds leerrijk en brachten me steeds weer op het goede pad.

Het OLAND-team dat later werd uitgebreid naar het N-team heeft steeds een belangrijke rol gespeeld in mijn onderzoek en hoewel dit team jaarlijks wisselde was er steeds één vaste speler. Siegfried, a.k.a. doctor OLAND, jij was de dirigent van dit team. Ik had de ongelooflijke luxe om met jou als OLAND expert samen te werken. Ik heb genoten van onze brainstormsessies, discussies, uitstappen, bbq’s enz. Ik kon me geen betere begeleider wensen. Dank je. One of the people who inspired and motivated me to do research was Marta, which I admire for her nononsence approach and infectious motivation and energy. Marta, thanks for the good talks, advices and great time in Santiago de Compostela. Ons team werd elk jaar versterkt door thesisstudenten: Yan, Tijs, Katrien, Emilie, Jeroen, Fabian en Mariela. Jullie hebben door jullie briljante werk een grote hand gehad in dit werk. Emilie, met jou is de OLAND opvolging verzekerd. Jouw enthousiasme en gedrevenheid zal je nog ver brengen. 215

Dankwoord

I would also like to thank Bernhard, Norbert and Martin for the great cooperation during the measurement campaigns in Strass. Bernhard, I could never believe that you agreed in a cooperation on the mainstream treatment subject when I asked you in Miami (with some pushing of Sudhir, I have to admit). You gave me the chance to get a feeling with the fullscale application. I would like to thank you for the open discussions we had and for the good contacts, which opened doors for new challenges. Norbert, I would like to thank you for the enjoyable time in Strass, we formed a good team. Keep on practicing your lab-skills though! Martin, if every operator was that dedicated to his work as you are, DEMON or OLAND reactors were already operational in every wastewater treatment plant. Thank you for your help, enthusiasm and great atmosphere during work. Ook alle collega’s van LabMET wil ik bedanken voor de aangename werksfeer. De Rotonde, ook wel in de volksmond beter bekend als ‘het centrum van de kennis’, werd bevolkt door één voor één flamboyante figuren die zorgden voor een levendige sfeer en een goede afwisseling tussen het labowerk door. Joachim, medebewoner van het eiland, bedankt voor het tegengewicht aan al dat wielergeweld, de lachgasdiscussies en veel succes met het afwerken van je sprookjesboek. Simon, voorzitter van de frietcluster, hou de traditie hoog en sprokkel nog wat energie voor je laatste jaar. Willem, weetjesman van de rotonde, je flitsbezoeken aan de rotonde zijn legendarisch. Tom, voorzitter van de rotonde, veel succes daar aan de overkant van de grote plas. Als je in de buurt van DC komt, spring gerust even binnen. Aan het nieuw jonge geweld (Sam, Emilie, Joeri en Stephen): bedankt voor de nieuwe frisse wind. Ook de oud-rotondenaars (Ilse, Selin, Bart, Peter, Lois) waren één voor één kleppers die het aangenaam werken maakten.

Verder waren er nog mensen van binnen en buiten LabMET waar ik veel aan te danken heb. Greet, bedankt voor het helpen met de IC en bestellingen. Tim, je figuren maken dit werk af! Bedankt voor je tijd en precisie. Siska, bedankt voor je moleculaire ondersteuning. Ook een dikke merci aan het secretariaat (Kris, Regine). Samuel en Katja, bedankt voor de hulp bij de NO testen en metingen. Jan bedankt voor de hulp met de N2O metingen. Thomas bedankt voor je inzet bij het LCA werk.

Uiteraard zijn het niet alleen de werk-gerelateerde mensen maar des te meer de achterban van vrienden en familie die maken dat je je zinnen kan verzetten als het iets minder. Een heel groot deel van mijn vrije tijd ging uit naar volleybal. Mijn geweldige ploegmaats groeiden uit 216

Dankwoord

tot vrienden. De super sfeer, cava-momenten, verkleedtrainingen, maar ook het samen afzien en strijden tijdens de wedstrijden waren een belangrijke bron van ontspanning. Het was dan ook een eer jullie kapitein te zijn. Sophie, An, Marijke, Lore, Assunta, Tessa, Kimberly, Anke, Ruth, Joëlle, Sylvie en Philip doe dit nog goed dit seizoen en hou met op de hoogte. Ik hou me alvast klaar voor het kerstfeestje. Ook onze verlaters Evelien (mede orvalliefhebber), Kim (altijd klaar voor een frietkotstop) en Steven (onze reddende engel) hebben me een leuke tijd bezorgd. Naast dit fantastisch team was er ook nog het jeugdig geweld van onze miniemenploeg. Hun enthousiasme en speelsheid deden me terug beseffen waarom we dit spelletje zo leuk vinden. Naast het volleybalgeweld brachten de leuke babbels en etentjes en drinks een goed tegengewicht. Annabel en

endrik, hoewel onze agenda’s niet altijd even

gemakkelijk bij elkaar te leggen vielen, waren de momenten dat we samen waren altijd zeer tof en ontspannend. Dennis en Jessica, met jullie heb ik mijn eerste ski-ervaring opgedaan, mountainbike tochten georganiseerd, maar vooral leuke momenten beleefd. Zeker nu Yente ertussen loopt, brengt dit steeds leven in huis.

Als laatste en belangrijkste steun wil ik nog mijn ouders en broers bedanken. Mama en papa, bedankt voor jullie onvoorwaardelijke steun en vertrouwen. Door jullie ben ik kunnen uitgroeien tot wat ik nu ben. Johan en Maarten, bedankt voor de leuke ontspannende momenten, de toffe reizen en zoveel meer. Ik verwacht jullie dan ook in de zomer voor een Amerikaans avontuur!

Haydée, Oktober 2012

217