Water Research 84 (2015) 207e217

Contents lists available at ScienceDirect

Water Research
journal homepage: www.elsevier.com/locate/watres

Full scale performance of the aerobic granular sludge process for

sewage treatment
M. Pronk a, *, 1, M.K. de Kreuk a, b, 1, B. de Bruin c, P. Kamminga d, R. Kleerebezem a,
M.C.M. van Loosdrecht a
a

Department of Biotechnology, Delft University of Technology, The Netherlands

Department of Water Management, Delft University of Technology, Stevinweg 1, 2628CN, Delft, The Netherlands
Royal HaskoningDHV B.V., P.O Box 1132, 3800 BC Amersfoort, The Netherlands
d
Waterschap Noorderzijlvest, Stedumermaar 1, 9735 AC Groningen, The Netherlands
b
c

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 13 April 2015
Received in revised form
24 June 2015
Accepted 6 July 2015
Available online 9 July 2015

Recently, aerobic granular sludge technology has been scaled-up and implemented for industrial and
municipal wastewater treatment under the trade name Nereda. With full-scale references for industrial
treatment application since 2006 and domestic sewage since 2009 only limited operating data have been
presented in scientic literature so far. In this study performance, granulation and design considerations
of an aerobic granular sludge plant on domestic wastewater at the WWTP Garmerwolde, the Netherlands
were analysed. After a start-up period of approximately 5 months, a robust and stable granule bed
(>8 g L
1) was formed and could be maintained thereafter, with a sludge volume index after 5 min
settling of 45 mL g
1. The granular sludge consisted for more than 80% of granules larger than 0.2 mm
and more than 60% larger than 1 mm. Efuent requirements (7 mg N L
1 and 1 mg P L
1) were easily met
during summer and winter. Maximum volumetric conversion rates for nitrogen and phosphorus were
respectively 0.17 and 0.24 kg (m3 d)
1. The energy usage was 13.9 kWh (PE150$year)
1 which is 58e63 %
lower than the average conventional activated sludge treatment plant in the Netherlands. Finally, this
study demonstrated that aerobic granular sludge technology can effectively be implemented for the
treatment of domestic wastewater.
2015 Elsevier Ltd. All rights reserved.

Keywords:
Aerobic granular sludge
Full scale
Nutrient removal
Sludge volume index
Sewage treatment
Granulation
Energy

1. Introduction
Aerobic granular sludge (AGS) technology is an upcoming
technology for the treatment of domestic and industrial wastewater (Heijnen and Van Loosdrecht, 1998; Morgenroth et al., 1997;
de Bruin et al., 2004; de Kreuk et al., 2007; Coma et al., 2012; Show
et al., 2012; Morales et al., 2013). AGS technology for combined
carbon, nitrogen and phosphorous removal is based on a repeated
fed batch process and relies on microorganisms selected to grow in
granules rather than ocs. As a result of the high settling rate of the
sludge granules, separate settling tanks are not needed and an 80%
reduction in area use is possible (de Bruin et al., 2004).
Aerobic granules are characterised by a compact structure,

* Corresponding author. Department of Biotechnology, Delft University of Technology, Julianalaan 67, Delft, 2628 BC, The Netherlands.
E-mail address: [email protected] (M. Pronk).
1
Both authors contributed equally to this manuscript.
http://dx.doi.org/10.1016/j.watres.2015.07.011
0043-1354/ 2015 Elsevier Ltd. All rights reserved.

without the need for carrier material, resulting in high settling

velocities and a low sludge volume index (SVI). A good indication of
granulation is the limited difference between SVI after 5 and 30 min
(Etterer and Wilderer, 2001). Aerobic granules are also characterised by their layered structure. The presence of an aerobic outer
layer and an anaerobic or anoxic core, facilitates co-existence of
nitrifying organisms in the outer layers of the granules and denitrifying phosphate accumulating organisms (dPAO), as well as
(facultative) anaerobic organisms towards the centre of the granules (Gieseke et al., 2001; Winkler et al., 2012; Pronk et al., 2015).
Due to this structure, aerobic granular sludge can simultaneously
remove phosphorus, nitrogen and COD (chemical oxygen demand)
from the liquid (de Kreuk et al., 2005; Gonzalez-Gil and Holliger,
2011).
Aerobic granular sludge technology was developed during the
last decade at laboratory scale (Morgenroth et al., 1997; Beun et al.,
1999; Tay et al., 2002; Zeng et al., 2003), as well as pilot scale
(Morales et al., 2013; Liu et al., 2010; Isanta et al., 2012; Wei et al.,

208

M. Pronk et al. / Water Research 84 (2015) 207e217

2012; Li et al., 2014). Aerobic granular sludge is applied by RoyalHaskoningDHV in the Nereda technology which was rst adapted
for industrial applications and then further scaled-up for domestic
sewage treatment. Valuable scale-up experience gained from fullscale demonstration plants in Gansbaai, South Africa and Frielas,
Portugal were used for optimization of the process design and
construction in 2010 of the rst full-scale AGS wastewater treatment plant in Epe, The Netherlands (van der Roest et al., 2011;
Giesen et al., 2013). The AGS technology used relies on a
sequencing fed-batch process with a constant working volume.
This is possible due to simultaneous feeding and efuent discharge,
that relies on a plug-ow pattern for displacement of efuent from
the reactor (de Kreuk et al., 2005). In addition, as a result of an
oxygen gradient within the granular sludge particle during aeration, extensive biological phosphate removal and simultaneous
nitrogen removal can be achieved during one aeration step. The
absence of conventional recycle pumps, sludge return pumps and
mixers provides a signicant reduction in electricity consumption
compared to standard nutrient removal plants.
In July 2013, a full scale installation based on the aerobic granular sludge process was taken into operation in Garmerwolde, the
Netherlands.
Few papers have been published so far describing full scale
operation of the AGS process on domestic wastewater. Li et al.
(2014) showed the performance of an full-scale AGS plant fed
with 30% domestic and 70% industrial wastewater (BOD/COD
0.23). The full-scale installation in Epe, the Netherlands briey
described by Giesen et al. (2013) also treats wastewater that is
derived for a large part (35%) from industry (mainly slaughterhouses). Moreover, data provided in these studies are very general.
More detailed descriptions of the process, conversions, energy usage and design considerations when treating domestic wastewater
are lacking.
After start-up and more than one year of operation, this paper
reects not only on the performance, but also on granulation, COD,
nitrogen and phosphorus conversions and especially design considerations. The measured energy requirement of the AGS process
is compared to conventional activated sludge systems. Furthermore, differences between the full scale granular sludge process,
conventional activated sludge and laboratory reactors are discussed
in detail.

sludge loading rate was 0.10 kg COD (kg TSS d)
1 at an expected
sludge concentration of 8 kg m
3. The sludge-loading rate is
calculated by dividing the treated kg COD per day by the total
biomass present in the reactor (Table 2). The volumetric loading
rate of the AGS reactors is 1.5 m3 (m3 d)
1. Wastewater enters the
plant by a pressure main. After screening by 6 mm screens, the
wastewater goes to a grit removal plant and an inuent buffer
(4,000 m3) (Fig. 1). From the inuent buffer, the wastewater is fed to
two AGS reactors (height 7.5 m, volume 9,600 m3 each) that are
equipped with an internal recirculation system (top to bottom of
reactor) with a capacity of 2,500 m3 h
1 for each reactor. Treated
efuent is directly discharged from the reactors to the surface water
via static xed overow weirs.
Biological phosphate removal in the AGS process can be supplemented by metal salt addition directly in the bulk if necessary.
Surplus sludge is stored in a sludge buffer tank (400 m3). To prevent
anaerobic phosphorus release and to ensure continuous discharge
towards the mechanical belt thickeners, the retention time in the
surplus sludge buffer is kept to a minimum.
The AGS plant is operated as a sequencing fed batch process,
consisting of a simultaneous feeding and efuent withdrawal
period, a reaction period, and a settling/sludge withdrawal/idle
period. Nitrogen removal is predominantly established by simultaneous nitrication and denitrication, but for maximisation of
nitrogen removal (non-mixed) anoxic periods with a recycle from
top to bottom can be provided. The cycle can be adjusted to the
inuent characteristics (rain or dry weather conditions), the actual
sludge conversion rates, the desired efuent conditions and the
granular sludge selection pressure. During aeration periods, the
dissolved oxygen concentration was maintained between 1.8 and
2.5 mg L
1. The total operational cycle time of the reactors is 6.5 h at
dry weather conditions. During rainy weather, the cycle time is
shortened to 3 h by decreasing the aeration and increasing the
feeding time in order to treat the increased inuent ow (Fig. 2).
The reactor was seeded with surplus sludge from an existing fullscale AGS plant in Epe, the Netherlands, treating wastewater that
consist out of a large industrial part (slaughterhouses) to a concentration of 1 g L
1. In this surplus sludge, no granules were
present (SVI30 140 mL g
1).

2. Materials and methods

Each reactor is equipped with measurements for dissolved oxygen concentration, redox potential, temperature, water level, dry
matter and turbidity. Ammonium and phosphate are semi continuously measured (5e10 min interval) during the cycle by an automatic sampling and analysis device (Hach Lange; Filtrax, AMTAX
and PHOSPHAX). Sampling points for ammonium and phosphate
are located on 0.5 m under the water surface in the reactors. This
means that during feeding, when the reactor is not mixed the
concentrations of ammonium and phosphate in the efuent can be
followed. During the feeding, the liquid an S:CAN spectro:lysertm
probe from Interline was used to continuously monitor nitrate
concentrations at 0.5 m under the water surface. This means that
during feeding, when the reactor is not mixed the concentrations
measured are the efuent concentrations.

2.1. Description of plant

In 2012 water board Noorderzijlvest decided that the treatment
plant in Garmerwolde would be upgraded with the aerobic granular sludge process. The extension of the existing activated sludge
based sewage treatment process (STP) was necessary to meet the
efuent requirements. The STP in Garmerwolde treats approximately 27 million m3; of wastewater per year and 0.8 MW of
electricity is generated by the use of the biogas formed during
sludge digestion. The existing WWTP consists of a two-stage activated sludge plant (so-called AB-process), with chemical phosphorus removal and either glycerol or methanol for denitrication
hnke, 1978) and an SHARON reactor to treat side-streams from
(Bo
the plant's digester and sludge thickeners (Hellinga et al., 1998).
The AGS plant was designed by Royal HaskoningDHV and is
operated in parallel with the existing AB-plant. The AGS plant treats
41% (28,600 m3 d
1) of the total inuent received at Garmerwolde
WWTP during dry weather ow with a maximum of 4,200 m3 h
1.
The average ow received by AGS and the AB-plant totals approximately 70,000 m3 d
1 with a peak ow of 11,600 m3 h
1.
Wastewater characteristics are given in Table 1. The designed

2.2. Online measurements

2.3. Grab samples

Periodic sampling for the determination of the concentration of
the biomass present in the reactor is performed by a grab sampler.
Samples can be acquired pressureless at various depths to account
for possible segregation over the depth of the reactors due to
granulation. Normally, samples are only taken when the reactor is
well mixed during the reaction period.

M. Pronk et al. / Water Research 84 (2015) 207e217

209

Table 1
Inuent and efuent of the aerobic granular sludge plant in Garmerwolde, Netherlands.
Inuent

Efuent

Min.

Max.

1

Parameter

[mg L

Suspended Solids
COD
BOD5
Total Phosphorus
PO34eP dissolved
Total Nitrogen
NH
4 -N dissolved

101
146
60
1.9
1.5
14
13.4

Average

1

[mg L
465
715
420
9.7
6.8
81
56.5

Load

1

[mg L

Efuent requirements

1

[kg d

236
506
224
6.7
4.4
49.4
39.0

8,000
14,636
5,495
212
1,387

MarcheDecember 2014 average

1

[mg L

[mg L
1]

30
125
20
1
e
7
e

20
64
9.7
0.9
0.4
6.9
1.10

Table 2
Operational parameters of the Nereda granular sludge plant in Garmerwolde, Netherlands.
Data obtained between March and December 2014
Parameter

Unit

Value

Parameter

Unit

Value

Solid retention time

Dry weight
Ash content
Sludge loading totala
Sludge loading biologicalb
Sludge production
Hydraulic retention time
Volumetric load
Nitrogen Load

d
1
kg m
3
%
kg COD (kg TSS d)
1
kg COD (kg TSS d)
1
kg d
1
h
1
m3 (m3 d)
1
Kg Ntot (kg TSS d)
1

20e38
6.5e8.5
25
0.10
0.12
3900
17
1.45
0.011

Max Recycle ratio

Fe(III)/P ratiod
Volumetric phosphate uptake rateb
Max. volumetric phosphate uptake ratec
Volumetric ammonium uptake rateb
Max. volumetric ammonium uptake ratec
Energy
Energy
Energye

e
e
kg P (m3 d)
1
kg P (m3 d)
1
kg N (m3 d)
1
kg N (m3 d)
1
kWh (m3 year)
1
kWh (kg N)
1
KWh (PE150, rem. year)
1

0.3
0.18
0.011
0.24
0.058
0.17
0.17
3.6
13.9

a
b
c
d
e

Based on kg COD received per day over the total biomass present in the reactor.
Based on the aerated time.
Based on actual rates measured inside the cycle at 20  C.
Only during heavy rain weather events. Molar ratio is given.
P.E.: Population equivalent or 1 i.e. equals 150 mg TOD per day.

2.4. Physical characteristics of granules

Approximately 1,500 mL of reactor volume was washed over a
212 and 600 mm sieve. The sieved granules and 100 mL of un-sieved
sample were dried at 105  C until no weight change was measured.
The dried fractions were then used to calculate the percentage of
granules and the total biomass concentration in the reactor. Sludge
volume index was measured by pouring 1,000 mL of sample in a
graduated measuring cylinder. Volume of the biomass was consequently recorded after 5 and 30 min. Only undiluted samples were
used in determining the SVI.
2.5. Fluorescent in situ hybridisation
Sample handling, xing and staining was performed as
described in Bassin et al. (Bassin et al., 2011a). Probes used to stain
all bacteria consisted out of a mixture of EUB338, EUB338II and
EUB338III (Amann et al., 1990; Daims et al., 1999). To stain the PAOs
present, a mixture of PAO462, PAO651 and PAO846 described by
Crocetti et al. (2000) was used. Probes GAOQ431 and GAOQ989
were used to stain the glycogen accumulating organisms (GAO)
population (Crocetti R et al., 2002). The samples were examined
with a Zeiss Axioplan 2 epiuorescence microscope equipped with
lter set 26 (bp 575e625/FT645/bp 660e710), 20 (bp 546/12/
FT560/bp 575e640), 17 (bp 485/20/FT 510/bp 5515e565) for Cy5,
Cy3 and uos respectively.
3. Results
3.1. Description of start up
The rst year of operation of the treatment plant can be divided

into two periods, based on the extent of granulation and conversion

efciencies established. The rst period was characterised by a
gradual increase of the volumetric loading rate to the design
loading rate and the development of granules from the occulent
sludge (September 2013eFebruary 2014). Reaching design loading
rate took about 3 months with uctuating conversion efciencies.
In period two reactor operation and process efciencies were
optimized (MarcheDecember 2014).
Preceding period one, technical testing was performed in July
and August 2013. Both AGS plant and the existing conventional
activated sludge treatment plant were expected to meet nitrogen
and phosphorus efuent standards of 15 and 1 mg L
1 respectively.
To safeguard these efuent quality requirements, the loading rate
towards the AGS reactors was increased only as far as the efuent
quality allowed. This meant that on some occasions during the
start-up period, the loading rate had to be temporarily reduced.
Within 3 months the design ow rate was reached and 41% of the
wastewater received at the WWTP was treated by the aerobic
granular sludge plant (Fig. 3a). Nitrogen, phosphorus and COD
removal were already stable during a part of period 1, but were not
optimized yet (Fig. 5). Sludge volume indexes, decreased signicantly during start-up (Fig. 3b). Initially, the SVI5 and SVI30 were
145 and 90 mL g
1, but they decreased during the start-up period to
approximately 70 and 50 mL g
1 respectively. The difference in SVI5
between the samples from the reactor and those from the surplus
sludge continued in the second period. SVI values kept improving
until stable values were reached of 45 and 35 mL g
1 for SVI5 and
SVI30 respectively. Biomass concentration increased from 3 to
6.5 kg m
3 in the rst period (Fig. 4). At the end of the second
period it reached a concentration of >8 kg m
3. The percentage of
sludge present as granules increased from 30 to >80% over this
period.

210

M. Pronk et al. / Water Research 84 (2015) 207e217

Fig. 1. Garmerwolde WWTP in the Netherlands. The photograph displays the existing AB- treatment system in the background with the aerobic granular sludge plant extension in
front. The gure below shows the process scheme of the aerobic granular sludge plant.

3.2. Performance results during period 2

February 2014 was considered as the end of period 1. Nutrient
removal as well as a signicant granule bed was present from that
time onward. Performance results and operational parameters
obtained between March and December 2014 are summarized in
Tables 1 and 2 Monthly average nutrient concentrations are given
in Fig. 5. Efuent guidelines were met for all parameters, with an
average Ntot and Ptot of 6.9 and 0.9 mg L
1, respectively, in the
period MarcheDecember 2014. Elevated nitrogen and phosphorus
concentrations in the efuent always coincided with heavy rain
weather events. Due to presence of a large sewer system, heavy rain
events provoke a ow rate that can be 2e5 times as high (Fig. 3a)
with nutrient concentrations as occurring in dry weather ow (socalled rst ush). The combination of increased ow and loading
rates leads to shorter cycle times and resulted in elevated efuent
concentrations. At these times, in accordance with the design setup targeting an optimized installed reactor volume, metal salts
are used for the removal of phosphorus at the end of the cycle. The

Fe(III)/P molar ratio during the operational period was 0.18 with a
total P-removal efciency of 90%. During normal rain and dry
weather ow the removal of phosphorus is completely biological.
3.2.1. Granulation
As the dry weight increased, so did the percentage of granules in
the total biomass. In period 2 more than 80% of the TSS consisted of
granules larger than 0.2 mm with more than 60% above 1 mm. Since
selection pressure on settling behaviour is needed for the development of well settling granular sludge during start-up (Beun et al.,
1999; Qin et al., 2004), the smaller and less settling sludge should
be removed with the surplus sludge, while larger granules should
be maintained in the reactor. Barely any granules larger than
0.2 mm were detected in the surplus sludge and consisted mostly
out of relatively slow settling ocs, which indicates good separation
between slow and fast settling biomass (Fig. 7a, b, c). Protozoa,
mainly ciliated, are present on the outside of the granules. Sliced
granules examined under a stereo zoom microscope revealed a
layered structure with micro colonies present inside (Fig. 7d).

M. Pronk et al. / Water Research 84 (2015) 207e217

211

Dry weather cycle

Time [min]

15 30 45 60 75 90 105 120 135 150 165 180 195 210 225 240 255 270 285 300 315 330 345 360 375 390
REACTOR 1

Fill/ draw
Reaction/Aeration
Settling
REACTOR 2
Fill/ draw
Reaction/Aeration
Settling
Rain weather cycle
Time [min]

15 30 45 60 75 90 105 120 135 150 165 180 195 210 225 240 255 270 285 300 315 330 345 360 375 390
REACTOR 1

Fill/ draw
Reaction/Aeration
Settling
REACTOR 2
Fill/ draw
Reaction/Aeration
Settling
Fig. 2. Typical cycle timing during dry and rain weather operation of the aerobic granular sludge plant.

3.2.2. Fish
Fluorescent in situ hybridisation performed on pottered granules showed the presence of a signicant PAO population (Fig. 7e, f).
Most of the bacteria stained with the EUB probe (blue) (in the web
version), also stained with the PAO probe (red) resulting in the pink
colour. In fact, hardly any GAOs could be detected (<5%). Other
heterotrophs that are typically stained with the EUB probe were
present, but were not further investigated.
3.3. Conversions within a typical Nereda cycle
The cycle begins with simultaneous inuent feeding and
efuent discharge (Fig. 2 and Fig. 6). Inuent is pumped into the
reactor from the bottom with an upward velocity of 3e3.3 m h
1
without further mixing or aeration. Simultaneously, efuent is
pushed over the overow weirs at the top of the reactor. In these
anaerobic feeding conditions, easy biodegradable organic carbon
will be converted into storage polymers by glycogen (GAO) and
phosphorus (PAO) accumulating organisms in the granule bed. The
uptake of easy biodegradable substrates by PAOs (Fig. 7h) is
accompanied by the release of ortho-phosphate into the bulk liquid
as can be seen from the relative high phosphate peak at the
beginning of the aeration period. The volume exchange ratio in
each cycle is limited by dispersion of the inuent and thus the
imperfection in the plug ow pattern during feeding. Since the
sampling points for nitrogen and phosphorus are located 0.5 m
under the liquid surface (just below the efuent weirs), the inuent
feeding can and needs to be stopped as soon as the phosphate
concentration increases at this sampling point, since this indicates
that a small fraction of the inuent is reaching the efuent weir.
This occurs when the effective volume exchange ratio is 65%, while
during dry weather ow in accordance with the design this effective volume exchange ratio is typically between 30 and 40 %.
Ammonium present in wastewater is partially adsorbed by the
biomass during the feeding period (Bassin et al., 2011b). Adsorbed
ammonium will, however, slowly desorb from the biomass during
the aerobic period and consequently be converted. After the
feeding period, the reactor is aerated and thus mixed, while the
dissolved oxygen concentration is controlled at 1.9 mg L
1. At the
start of the aeration period ammonium and phosphate concentration peak due to the mixing of the clean water in the top (where the

measurement system is) and the inuent in the lower part of the
reactor. The concentrations after mixing, combined with the
amount of wastewater fed, can be used as an indication of the
loading rate of the cycle, but need to be corrected for phosphorus
release and ammonium adsorption.
In the aerated period ammonium and phosphorus uptake
occurred, while nitrate was produced. Once ammonium reached its
desired set-point the oxygen concentration in the bulk was lowered
to allow for maximum denitrication potential. Denitrication is
taking place inside the granule. Since the diffusion of oxygen into
the granule is slower than the conversion rate, an anoxic volume is
created inside the granules (de Kreuk et al., 2005; Osman et al.,
2001). Storage polymers of the GAOs and PAOs are delivering the
reducing power for the conversion of nitrate to dinitrogen gas. Once
both nitrogen and phosphorus have reached the required efuent
quality the reactor is ready to receive new inuent.
4. Discussion
4.1. Start-up
Generally, technological start-up times for nutrient removing
conventional activated sludge (CAS) plants range anywhere between 30 and 40 days, assuming the biomass inoculated is adapted
to the wastewater and adequate amounts are available. Otherwise
longer start-up times are needed to grow and adapt the biomass.
Start-up times with respect to efuent quality required for the AGS
plant were similar. In this case, the start-up period was a bit longer
due to the requirement to achieve full efuent quality standards at
all times, even during start-up conditions. This resulted in a 3
months stepwise increase in redirected inuent ow from the ABplant towards the aerobic granular sludge plant.
This study demonstrated that a good efuent quality can be
achieved long before granulation is complete. The required ow
and treatment capacity was reached in only 3 months, but the
intended biomass concentration of 8 g L
1 was only achieved after 9
months. The studied reactors were started with surplus sludge of
the full scale Nereda in Epe, The Netherlands. Although this surplus
sludge was well settling (SVI30 140 mL g
1), it hardly contained
any granular sludge. Development of granular sludge is expected to
be shortened drastically by starting with granular sludge from

212

M. Pronk et al. / Water Research 84 (2015) 207e217

Fig. 3. Flow and biomass characteristics during start-up of the aerobic granular sludge reactor: (a) percentage of the inuent ow towards the aerobic granular sludge plant (black
line; ve day moving average), total inuent ow received at the Garmerwolde plant (light grey line) and the design inuent percentage towards the aerobic granular sludge plant
(dotted line); (b) sludge volume index after 5 (lled cubes) and 30 (open circles) minutes of settling.

other plants (Pijuan et al., 2011). Also, when the efuent criterium is
less critical during start-up, the time to reach the design sludge
mass can be signicantly reduced by increasing the COD load. It
should be noted that a treatment plant is designed for the highest
load at the coldest temperature. For this reason the required
treatment capacity can be achieved during most of the year, even if
full granulation is not yet reached.
4.2. Granulation
Successful formation of AGS in a full scale plant treating domestic wastewater was demonstrated in this study. A high percentage (>80%) of the sludge consisted of aerobic granules larger
than 0.2 mm. Aerobic granulation on domestic wastewater was
shown before in lab-and pilot scale studies, but the full granulation
of biomass on full scale was not (Coma et al., 2012; Giesen et al.,

2013; de Kreuk and van Loosdrecht, 2006; Wagner and Da Costa,

2013).
As proposed previously in lab-scale and pilot-scale studies,
granule formation could also be monitored satisfactorily with the
SVI5/SVI30 ratio for this full-scale reactor start-up (de Kreuk et al.,
2007; Etterer and Wilderer, 2001). More complicated tests as
granular size distribution measurements give more insight, but are
not required for full scale operations. In full scale plants there will
always be a fraction of occulent sludge resulting in an SVI5/SVI30
ratio higher than 1 (Fig. 7 a, b). This occulent material is resulting
from inuent particulates, detached pieces of granules and growth
on polymeric substrates (De Kreuk et al., 2010).
Many factors have been mentioned as critical for the formation
of granular sludge, such as high shear, very short settling times and
high DO concentrations (Beun et al., 2000a; Liu et al., 2005; Lee
et al., 2010). Most of these are related to experiments with

M. Pronk et al. / Water Research 84 (2015) 207e217

213

aerobic or completely mixed feeding. In the case of anaerobic plug

ow feeding through the granule bed, most of these factors (like
high settling pressure, shear, high DO etc.) become less relevant,
although not irrelevant, during start-up (de Kreuk et al., 2005).
Anaerobic feeding selects for relatively slow growing bacteria such
as PAOs, that convert the easy degradable COD under anaerobic
conditions into storage polymers, resulting in smooth and compact
biolms (Van Loosdrecht et al., 1997). The plug ow feeding
furthermore ensures high substrate concentrations near the
granule surface allowing substrate to diffuse through the entire
granule, further contributing to the formation of stable, compact
and dense granular biomass.

4.3. Suspended solids in the efuent

Fig. 4. Evolution of biomass concentration with time (lled circles) and temperature of
the bulk liquid (dotted line) in the aerobic granular sludge installation.

Fig. 5. Monthly nutrient averages in the efuent of the aerobic granular sludge system
with ( ) total nitrogen, ( ) ammonium, ( ) nitrate, ( ) P-total, ( ) ortho-phosphate.
Total nitrogen and phosphorus measurements for the aerobic granular sludge plant
separately were only available from October 2013.

Suspended solids in the efuent can originate from the inuent,

as well as from detached biomass. In the present case the inuent
contained on average 236 mg SS L
1 while in the efuent on
average only 20 mg SS L
1 was present. The exact removal mechanism of suspended solids in aerobic granular sludge systems is not
fully understood yet, however a few relevant observations were
made. Suspended material occulates and can be retained in the
AGS plant despite the up ow velocity (3e3.3 m h
1), after which it
is removed with the excess sludge. Cellulose bers, which are
present in domestic wastewater (Ruiken et al., 2013), are predominately removed with the excess sludge (Fig. 7b), and do not
end up in the efuent. Particles can also be entrapped in the
granules themselves (Fig. 7h), where they are degraded or nally
end up in the excess sludge as well. Finer material and colloidal
particles that are not entrapped can be further removed by protozoa (Lemaire et al., 2008; Priya et al., 2007; Li et al., 2013). In
granular sludge, stalked protozoa such as Vorticella and Carchesiumlike organisms attach to the granules (Fig. 7c). Protozoa have also
been related to particulate starch removal in laboratory experiments (De Kreuk et al., 2010; Lemaire et al., 2008; Schwarzenbeck
et al., 2004) and in biolm plants fed with suspended bacteria (Lee,
1996), however, their exact contribution to the removal of suspended solids from municipal wastewater has not been quantied
yet. Although these protozoa slightly hinder settling, they do not

Fig. 6. Typical operational cycle in the aerobic granular sludge plant with concentration of ammonium (green solid line), phosphate (black solid line), nitrate (red dotted line),
oxygen (blue dashed line) and feed (yellow solid line). (For interpretation of the references to colour in this gure legend, the reader is referred to the web version of this article.)

214

M. Pronk et al. / Water Research 84 (2015) 207e217

Fig. 7. Aerobic granular sludge: (a, b) biomass on the top of the granular sludge bed, (c) mixed sample of the bulk liquid, (d) granules sieved and washed with tap water, (e) phase
contrast and (f) uorescent in situ hybridisation with blue EUB (Cy5) and red PAOmix (Cy3) with pottered granules. Overlay of blue and red gives the pink colour, (g) 1000
magnication of phosphate accumulating organisms found in pottered granules and (h) sliced granule. (For interpretation of the references to colour in this gure legend, the reader
is referred to the web version of this article.)

disturb the plant and should be seen as a sign of a healthy microbial

ecosystem in the treatment of domestic wastewater.
4.4. Nutrient removal
In sewage treatment, the nitrogen removal efciency depends
on the incorporation of nitrogen in the growing biomass, and the
nitrication and denitrication capacity of the microbial community. The efuent nitrogen concentration of conventional activated
sludge plants with pre-denitrication is furthermore limited by the

recycle ow from the aerobic to the anoxic tank. Through post

denitrication very low nitrogen concentrations can be established,
but this requires the supply of external carbon. External carbon
dosage is costly and can lead to elevated BOD concentrations in the
gy et al., 2008). One of the main
efuent when overdosed (Jobba
advantages of the aerobic granular sludge plants is that the biological nitrogen removal processes are not separated in different
anoxic and aerobic tank volumes, but can be optimised by a exible
aeration and denitrication period. During aeration, ammonium
will be converted to nitrate and due to the inherent oxygen

M. Pronk et al. / Water Research 84 (2015) 207e217

gradients present in the granules, a part of the nitrogen will be

simultaneously denitried. The ratio of nitrication to denitrication can therefore be regulated by changing the DO set-point.
When the nitrate concentration at the end of the nitrication
process is still too high for efuent discharge, a recycle from the
bulk liquid at the top of the reactor towards the settled granules
bed can be introduced for enhanced nitrate removal. As an alternative, an intermittent aeration strategy can be applied. The
effective recycle ow rate can be much smaller than is usually used
in activated sludge plants and in the present study, the average was
only 0.3 of the inuent ow rate to the reactors during the operational period. Compared to the AGS plant (0.3), a conventional predenitrication system would require a recycle ratio of roughly
10e15 times higher to achieve the same total nitrogen removal of
over 90% (excluding nitrogen in SS efuent). In the aerobic granules, internally stored polymers of the PAO and the GAO serve as
electron donor in the denitrication process. Simultaneously, the
PAOs store phosphate as poly-P, hereby fullling a double function
(denitrication and phosphate removal). The phosphate uptake is
therefore less sensitive to aeration than nitrication, making the
process control more exible. As long as anoxic conditions prevail,
the PAOs do not release phosphate in the bulk liquid, which results
in very low nitrogen and phosphorus concentrations in the efuent.
Due to the high amount of phosphorus stored inside the sludge
as poly-P, a large amount of phosphorus is released during the
anaerobic periods. This is especially true during heavy rain events,
when the increased ow rate is combined with concentrations
that initially remain as high as during dry weather ow (so-called
rst ush). The resulting increased COD loading rate triggers extra
phosphate release. The high hydraulic loading rate, however, enforces a short hydraulic residence time with a shorter cycle time as
a consequence (Fig. 2). This reduces the time for aerobic P-uptake
signicantly. In the design of the Garmerwolde WWTP, this
decreased biological phosphorous removal capacity under
extreme rain events is tackled by iron dosing, rather by tuning the
reactor size. This enables further reduction of the installed reactor
volume. During normal rain e and dry weather - events metal
dosing was not needed as phosphate was always removed
biologically.
4.5. Volumetric conversion rates
As a result of the higher biomass concentrations (>8 g L
1), the
batch wise operation and the parallel occurrence of conversions
that can be reached with granular sludge plants, volumetric conversion rates are higher than in conventional activated sludge
plants and thus volumetric loading rates can therefore be increased
accordingly. The simultaneous nitrication and denitrication in
granular sludge plants, can even lead to a more than proportional
increase of volumetric loading rate. The volumetric nitrogen conversion rate (based on the aerated time) of this plant was
0.058 kg N (m3 d)
1. The maximum volumetric ammonium conversion rate was 0.17 kg N (m3 d)
1 at an average temperature of
20  C (Table 2). This is faster than generally found in conventional
activated sludge plants. The obtained conversion rates underestimate the actual rate, since they are based on on-line ammonium
measurements, while a fraction of the inuent ammonium is
initially adsorbed or incorporated in struvite (Bassin et al., 2011b;
Lin et al., 2012). The granular sludge allows for a compact treatment process, which is reected in the volumes of the existing old
conventional plant, treating 59% of the sewage in 28,400 m3 of
activated sludge tanks plus 24,800 m3 of clariers, while the AGS
reactors treats 41% of the sewage in 19,200 m3 tanks plus 4,000 m3
of rainwater buffer. This amounts to 1.2 m3 d
1 of wastewater
treated per m3 volume needed for the AGS plant compared to

215

0.8 m3 d
1 for the existing treatment plant. Leading to a volume

reduction of 33% for the AGS plant compared to the existing plant.
Efuent nutrient concentrations were also higher for the conventional AB-plant (9.9 mg N L
1 and 0.9 mg P L
1) in the same
period. It should be noted that there is no biological phosphorus
removal in the AB-plant; hence most of the phosphorus is removed
chemically (metal dosing).
Generally, the ammonium and phosphorus removal rates are
such that both compounds can be easily removed well before the
end of the cycle, thus creating room for polishing the efuent
quality if desired (Fig. 6). However, in light of saving energy,
efuent concentrations are steered towards the demands instead of
the lowest concentration obtainable.
The oxygen gradients inside the granules, due to diffusion limitation, and the batch wise operation of the plant allow for
enhanced process control via DO (Lochmatter et al., 2013; Isanta
et al., 2013; Kagawa et al., 2015). An increase of the DO will lead
to a larger aerobic volume fraction of the biolm and thus an increase in volumetric nitrication rate (Aravinthan et al., 1998;
Wijffels and Tramper, 1995; Beun et al., 2000b). Likewise,
lowering the DO creates more anoxic volume leading to higher
denitrication rates. This opens up process control opportunities
that are difcult to achieve with activated sludge plants. A proper
DO control strategy will therefore enable accurate steering of
efuent parameters by manipulating the oxygen gradients inside
the granules.
Finally, transport characteristics are less temperature dependent than bacterial growth rates (Wijffels et al., 1995). This implies
that biolm systems are less sensitive to temperature changes than
suspended and occulated systems (Leenen et al., 1997). So the
effect of decreasing temperature (summer - winter) in AGS systems
on the conversion rates is expected to be less than that of activated
sludge systems.
4.6. Energy
We have estimated the electricity consumption of the AGS
plant in the total energy balance of the Garmerwolde WWTP. The
calculation includes the fraction of energy used in the inuent
pumping station for the AGS plant inuent, but dewatering of the
biomass and the energy generation by sludge digestion is not
taken into account. The reason for this is that the digester
(onsite) receives sludge from the AGS as well as the AB-plant plus
external sludge from neighbouring plants. Determining the exact
contribution of the waste aerobic granular sludge to the total
biogas production is therefore impossible. The energy consumption of the full-scale AGS plant is compared to the average
energy use of all Dutch WWTP's (Table 3). These are partitioned
in plants with a good efuent quality (N z 10 mg N L
1;
P z 1 mg L
1) and plants with an excellent efuent quality due
to post treatment (N 3e10 mg N L
1 and P 0.1e0.5 mg L
1)
(STOWA, 2013). These data for Dutch WWTP's include energy
generation by sludge digestion, which is present at most of the
Dutch WWTPs. The AGS process should be compared to the
higher efuent quality WWTP's. The AGS process shows very low
energy consumption. Due to the lack of mixers, conventional
recycle pumps, settlers and sludge return pumps, energy usage of
sequencing batch systems will be lower than that of conventional
activated sludge plants. Moreover, energy usage for aeration is

ndez-Alvarez
affected by the water depth of the tanks (Ferna
et al., 2014). The AGS tank studied had a height of 7.5 m, while
activated sludge tanks are often limited to 6 m. This potentially
adds 5e10 % extra to the aeration energy saving. Overall, the
studied AGS plant consumed 58e63 % less energy compared to
conventional activated sludge plants.

216

M. Pronk et al. / Water Research 84 (2015) 207e217

Table 3
Comparison between the average energy demand between various WWTP's in the
Netherlands.
Type

Specic energy demand

[kWh (PE150, removed year)
1]

WWTP with polishing

(Total N efuent < 10 mg N L
1)
WWTP without polishing
(Total N efuent  10 mg N L
1)
WWTP Nereda Garmerwolde

37.5
33.4
13.9

When comparing the energy usage of the existing conventional

AB-plant (0.33 kWh m
3) with the AGS plant (0.17 kWh m
3) at
WWTP Garmerwolde over the same period per m3 of treated
inuent, the AB-plant consumed 51% more energy. To make a
proper comparison between the two plants the energy required for
the sludge treatment and the SHARON plant were omitted in both
energy calculations.
Due to the low energy use of the AGS plant, the concept contributes to the development of energy neutral or energy producing
WWTP's. In addition, preliminary results showed that the total biomethane potential of aerobic granular sludge during digestion is
similar to the digestion of activated sludge (data not shown), but
more detailed studies on the digestion and biogas production of
aerobic granules grown on sewage are needed to be able to give a
detailed energy balance over the plant (Val Del Ro et al., 2014).
5. Conclusion
The present study described the start-up and operation of one of
the currently largest full scale aerobic granular sludge plants
treating domestic sewage. The results illustrate that the technology
has grown in a decade from a promising innovation to a mature
technology capable of competing with established wastewater
treatment technologies. The operational data show that in Dutch
climate conditions the efuent requirements (7 mg N L
1 and
1 mg P L
1) were easily reached, while maintaining a high granular
biomass concentration (>8 g L
1) with SVI5 values of 45 mL g
1 and
process stability for both summer and winter periods. Because of
the high biomass concentration, the volumetric loading rate could
be increased accordingly. Energy usage of the aerobic granular
sludge system was 13.9 kWh (PE150$year)
1. This is 58e63 % lower
than the average conventional activated sludge plant in the
Netherlands, with an efuent quality which is comparable or better.
The specic volume that was needed for the aerobic granular
sludge based system was 33% lower than the volume needed for the
existing conventional activated sludge plant. Finally, the results
show that aerobic granular sludge technology can be effectively
implemented for the treatment domestic wastewater. Further
enhancement, mainly based on improved process control, is still
possible. The full operational window with respect to inuent
variations has still to be established.
Acknowledgements
This research was partly supported by Stichting Toegepast
Onderzoek Waterbeheer (STOWA) (432.662 NNOP 2014), the Dutch
Foundation for Applied Water Research.
References
Amann, R.I., Binder, B.J., Olson, R.J., Chisholm, S.W., Devereux, R., Stahl, D.A., 1990.
Combination of 16S rRNA-targeted oligonucleotide probes with ow cytometry
for analyzing mixed microbial populations. Appl. Environ. Microbiol. 56 (6),
1919e1925.

Aravinthan, V., Takizawa, S., Fujita, K., Komatsu, K., 1998. Factors affecting nitrogen
removal from domestic wastewater using immobilized bacteria. Water Sci.
Technol. 38 (1 pt 1), 193e202.
Bassin, J.P., Pronk, M., Muyzer, G., Kleerebezem, R., Dezotti, M., van Loosdrecht, M.C.,
2011a. Effect of elevated salt concentrations on the aerobic granular sludge
process: linking microbial activity with microbial community structure. Appl.
Environ. Microbiol. 77 (22), 7942e7953.
Bassin, J.P., Pronk, M., Kraan, R., Kleerebezem, R., Van Loosdrecht, M.C., 2011b.
Ammonium adsorption in aerobic granular sludge, activated sludge and
anammox granules. Water Res. 45 (16), 5257e5265.
Beun, J.J., Hendriks, A., Van Loosdrecht, M.C.M., Morgenroth, E., Wilderer, P.A.,
Heijnen, J.J., 1999. Aerobic granulation in a sequencing batch reactor. Water Res.
33 (10), 2283e2290.
Beun, J.J., Van Loosdrecht, M.C.M., Heijnen, J.J., 2000a. Aerobic granulation. Water
Sci. Technol. 41, 41e48.
Beun, J.J., Verhoef, E.V., Van Loosdrecht, M.C.M., Heijnen, J.J., 2000b. Stoichiometry
and kinetics of poly-b-hydroxybutyrate metabolism under denitrifying conditions in activated sludge cultures. Biotechnol. Bioeng. 68 (5), 496e507.
hnke, B., 1978. Installation for the Treatment of Waste Water by the Activated
Bo
Sludge Process.
Coma, M., Verawaty, M., Pijuan, M., Yuan, Z., Bond, P.L., 2012. Enhancing aerobic
granulation for biological nutrient removal from domestic wastewater. Bioresour. Technol. 103 (1), 101e108.
Crocetti, G.R., Hugenholtz, P., Bond, P.L., Schuler, A., Keller, J., Jenkins, D.,
Blackall, L.L., 2000. Identication of polyphosphate-accumulating organisms
and design of 16S rRNA-directed probes for their detection and quantitation.
Appl. Environ. Microbiol. 66 (3), 1175e1182.
Crocetti, R.,G., Baneld, F.,J.F., Keller, J., Bond, P.L., Blackall, L.L., 2002. Glycogenaccumulating organisms in laboratory-scale and full-scale wastewater treatment processes. Microbiology 148 (11), 3353e3364.
Daims, H., Brhl, A., Amann, R., Schleifer, K.H., Wagner, M., 1999. The domainspecic probe EUB338 is insufcient for the detection of all bacteria: development and evaluation of a more comprehensive probe set. Syst. Appl.
Microbiol. 22 (3), 434e444.
de Bruin, L.M.M., de Kreuk, M.K., van der Roest, H.F.R., Uijterlinde, C., van
Loosdrecht, M.C.M., 2004. Aerobic granular sludge technology: an alternative to
activated sludge? Water Sci. Technol. 49, 1e7.
de Kreuk, M.K., van Loosdrecht, M.C., 2006. Formation of aerobic granules with
domestic sewage. J. Environ. Eng. 132 (6), 694e697.
de Kreuk, M.K., Heijnen, J.J., van Loosdrecht, M.C., 2005. Simultaneous COD, nitrogen, and phosphate removal by aerobic granular sludge. Biotechnol. Bioeng. 90
(6), 761e769.
de Kreuk, M.K., Kishida, N., van Loosdrecht, M.C.M., 2007. Aerobic granular sludge e
State of the art. Water Sci. Technol. 55, 75e81.
De Kreuk, M.K., Kishida, N., Tsuneda, S., van Loosdrecht, M.C.M., 2010. Behavior of
polymeric substrates in an aerobic granular sludge system. Water Res. 44 (20),
5929e5938.
Etterer, T., Wilderer, P.A., 2001. Generation and properties of aerobic granular
sludge. Water Sci. Technol. 43, 19e26.

rez, J., Go
mez, M.A., 2014. Optimization of reactor depth in
Fern
andez-Alvarez,
G., Pe
membrane bioreactors for municipal wastewater treatment. J. Environ. Eng.
(United States) 140 (7).
Gieseke, A., Purkhold, U., Wagner, M., Amann, R., Schramm, A., 2001. Community
structure and activity dynamics of nitrifying bacteria in a phosphate-removing
biolm. Appl. Environ. Microbiol. 67 (3), 1351e1362.
Giesen, A., de Bruin, L.M.M., Niermans, R.P., van der Roest, H.F., 2013. Advancements
in the application of aerobic granular biomass technology for sustainable
treatment of wastewater. Water Pract. Technol. 8 (1), 47e54.
Gonzalez-Gil, G., Holliger, C., 2011. Dynamics of microbial community structure of
and enhanced biological phosphorus removal by aerobic granules cultivated on
propionate or acetate. Appl. Environ. Microbiol. 77 (22), 8041e8051.
Heijnen, J.J., Van Loosdrecht, M.C.M., 1998. Method for Acquiring Grain-shaped
Growth of a Microorganism in a Reactor (Netherlands).
Hellinga, C., Schellen, A.A.J.C., 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 (9), 135e142.
 Morales, N., Pe
rez-Ojeda, M.E., Val del Ro, A.,
rez, J., Carrera, J., 2012.
Isanta, E., Sua
Long term operation of a granular sequencing batch reactor at pilot scale
treating a low-strength wastewater. Chem. Eng. J. 198e199, 163e170.
rez, J., 2013.
Isanta, E., Figueroa, M., Mosquera-Corral, A., Campos, L., Carrera, J., Pe
A novel control strategy for enhancing biological N-removal in a granular
sequencing batch reactor: a model-based study. Chem. Eng. J. 232, 468e477.
 , G., Bena
kov
, O., Wanner, J., 2008.
Jobb
agy, A., Tardy, G.M., Palko
a, A., Krhutkova
Savings with upgraded performance through improved activated sludge denitrication in the combined activated sludge-biolter system of the Southpest.
Wastewater Treatment Plant. Water Sci. Technol. 57 (8), 1287e1293.
Kagawa, Y., Tahata, J., Kishida, N., Matsumoto, S., Picioreanu, C., van
Loosdrecht, M.C.M., Tsuneda, S., 2015. Modeling the nutrient removal process in
aerobic granular sludge system by coupling the reactor- and granule-scale
models. Biotechnol. Bioeng. 112 (1), 53e64.
Lee, N.M., 1996. Use of protozoa and metazoa for decreasing sludge production in
aerobic wastewater treatment. Biotechnol. Lett. 18 (4), 429e434.
Lee, D.J., Chen, Y.Y., Show, K.Y., Whiteley, C.G., Tay, J.H., 2010. Advances in aerobic
granule formation and granule stability in the course of storage and reactor
operation. Biotechnol. Adv. 28 (6), 919e934.

M. Pronk et al. / Water Research 84 (2015) 207e217

Leenen, E.J.T.M., Van Boxtel, A.M.G.A., Englund, G., Tramper, J., Wijffels, R.H., 1997.
Reduced temperature sensitivity of immobilized Nitrobacter agilis cells caused
by diffusion limitation. Enzyme Microb. Techn 20 (8), 573e580.
Lemaire, R., Webb, R.I., Yuan, Z., 2008. Micro-scale observations of the structure of
aerobic microbial granules used for the treatment of nutrient-rich industrial
wastewater. ISME J. 2 (5), 528e541.
Li, J., Ma, L., Wei, S., Horn, H., 2013. Aerobic granules dwelling vorticella and rotifers
in an SBR fed with domestic wastewater. Sep. Purif. Technol. 110 (0), 127e131.
Li, J., Ding, L.B., Cai, A., Huang, G.X., Horn, H., 2014. Aerobic sludge granulation in a
full-scale sequencing batch reactor. BioMed Res. Int. 2014 http://dx.doi.org/
10.1155/2014/268789. Article ID 268789, 12 pages.
Lin, Y.M., Bassin, J.P., Van Loosdrecht, M.C.M., 2012. The contribution of exopolysaccharides induced struvites accumulation to ammonium adsorption in aerobic granular sludge. Water Res. 46 (4), 986e992.
Liu, Y., Wang, Z.W., Qin, L., Liu, Y.Q., Tay, J.H., 2005. Selection pressure-driven aerobic
granulation in a sequencing batch reactor. Appl. Microbiol. Biotechnol. 67 (1),
26e32.
Liu, Y.Q., Moy, B., Kong, Y.H., Tay, J.H., 2010. Formation, physical characteristics and
microbial community structure of aerobic granules in a pilot-scale sequencing
batch reactor for real wastewater treatment. Enzyme Microb. Technol. 46 (6),
520e525.
Lochmatter, S., Gonzalez-Gil, G., Holliger, C., 2013. Optimized aeration strategies for
nitrogen and phosphorus removal with aerobic granular sludge. Water Res. 47
(16), 6187e6197.
Morales, N., Figueroa, M., Fra-V
azquez, A., Val Del Ro, A., Campos, J.L., Mosquerandez, R., 2013. Operation of an aerobic granular pilot scale SBR
Corral, A., Me
plant to treat swine slurry. Process Biochem. 48 (8), 1216e1221.
Morgenroth, E., Sherden, T., Van Loosdrecht, M.C.M., Heijnen, J.J., Wilderer, P.A.,
1997. Aerobic granular sludge in a sequencing batch reactor. Water Res. 31 (12),
3191e3194.
Osman, J.J., Birch, J., Varley, J., 2001. N-removal in a granular sludge sequencing
batch airlift reactor. Biotechnol. Bioeng. 75 (1), 82e92.
Pijuan, M., Werner, U., Yuan, Z., 2011. Reducing the startup time of aerobic granular
sludge reactors through seeding occular sludge with crushed aerobic granules.
Water Res. 45 (16), 5075e5083.
Priya, M., Haridas, A., Manilal, V.B., 2007. Involvement of protozoa in anaerobic
wastewater treatment process. Water Res. 41 (20), 4639e4645.
Pronk, M., Abbas, B., Al-zuhairy, S.H.K., Kraan, R., Kleerebezem, R., van
Loosdrecht, M.C.M., 2015. Effect and behaviour of different substrates in relation to the formation of aerobic granular sludge. Appl. Microbiol. Biotechnol.
1e12.
Qin, L., Liu, Y., Tay, J.H., 2004. Effect of settling time on aerobic granulation in

217

sequencing batch reactor. Biochem. Eng. J. 21 (1), 47e52.

Ruiken, C.J., Breuer, G., Klaversma, E., Santiago, T., van Loosdrecht, M.C.M., 2013.
Sieving wastewater e Cellulose recovery, economic and energy evaluation.
Water Res. 47 (1), 43e48.
Schwarzenbeck, N., Erley, R., Wilderer, P.A., 2004. Aerobic granular sludge in an SBRsystem treating wastewater rich in particulate matter. Wat. Sci. Technol. 49
(11e12), 41e48.
Show, K.Y., Lee, D.J., Tay, J.H., 2012. Aerobic granulation: advances and challenges.
Appl. Biochem. Biotechnol. 167 (6), 1622e1640.
STOWA, 2013. Nereda Praktijkonderzoeken 2010e2012. ISBN 978-90-5778-604-9.
Tay, J.H., Liu, Q.S., Liu, Y., 2002. Characteristics of aerobic granules grown on glucose
and acetate in sequential aerobic sludge blanket reactors. Environ. Technol. 23
(8), 931e936.
Val Del Ro, A., Palmeiro-Sanchez, T., Figueroa, M., Mosquera-Corral, A., Campos, J.L.,
ndez, R., 2014. Anaerobic digestion of aerobic granular biomass: effects of
Me
thermal pre-treatment and addition of primary sludge. J. Chem. Technol. Biotechnol. 89 (5), 690e697.
van der Roest, H.F., de Bruin, L.M.M., Gademan, G., Coelho, F., 2011. Towards sustainable waste water treatment with Dutch Nereda technology. Water Pract.
Technol. 6 (3).
Van Loosdrecht, M.C.M., Picioreanu, C., Heijnen, J.J., 1997. A more unifying hypothesis for biolm structures. FEMS Microbiol. Ecol. 24 (2), 181e183.
Wagner, J., Da Costa, R.H.R., 2013. Aerobic granulation in a sequencing batch reactor
using real domestic wastewater. J. Environ. Eng. (United States) 139 (11),
1391e1396.
Wei, D., Si, W., Zhang, Y., Qiao, Z., Yao, Z., Zhao, W., Zhao, J., Chen, G., Wei, Q., Du, B.,
2012. Aerobic granulation and nitrogen removal with the efuent of internal
circulation reactor in start-up of a pilot-scale sequencing batch reactor. Bioprocess Biosyst. Eng. 35 (9), 1489e1496.
Wijffels, R.H., Tramper, J., 1995. Nitrication by immobilized cells. Enzyme Microb.
Technol. 17 (6), 482e492.
Wijffels, R.H., Englund, G., Hunik, J.H., Leenen, E.J.T.M., Bakketun, A., Gunther, A.,
Obon de Castro, J.M., Tramper, J., 1995. Effects of diffusion limitation on
immobilized nitrifying microorganisms at low temperatures. Biotechnol. Bioeng 45 (1), 1e9.
Winkler, M.K.H., Kleerebezem, R., Khunjar, W.O., de Bruin, B., van
Loosdrecht, M.C.M., 2012. Evaluating the solid retention time of bacteria in
occulent and granular sludge. Water Res. 46 (16), 4973e4980.
Zeng, R.J., Lemaire, R., Yuan, Z., Keller, J., 2003. Simultaneous nitrication, denitrication, and phosphorus removal in a lab-scale sequencing batch reactor.
Biotechnol. Bioeng 84 (2), 170e178.