Process Biochemistry 40 (2005) 645–650

Physical and chemical characteristics of granular activated

sludge from a sequencing batch airlift reactor

Yu-Ming Zheng, Han-Qing Yu

∗

, Guo-Ping Sheng

Laboratory of Environmental Engineering, Department of Chemistry, University of Science and Technology of China, Hefei, Anhui 230026, China

Received 14 September 2003; accepted 28 January 2004

Abstract

Granular activated sludge was cultivated in a sequencing batch airlift reactor fed with a synthetic sucrose-rich wastewater for 68 days.

The formation of granular activated sludge and its physical and chemical characteristics were investigated. With the granulation, the specific
gravity and surface hydrophobicity of sludge increased, but the specific surface area and the concentrations of carbohydrate and protein in the
extracellular polymeric substances decreased. Microscopic examination showed that the morphology of the mature granular sludge was nearly
spherical (0.5–1.2 mm in diameter) with a clear outline, and had a strong structure. The granular sludge had good settleability, as indicated
by a low sludge volume index of 23 ml/g and a high settling velocity of 18–31 m/h. Compared with activated sludge flocs, the granules were
regular and dense and exhibited high biomass retention.
© 2004 Elsevier Ltd. All rights reserved.

Keywords: Activated sludge; Characterization; Granulation; Sequencing batch airlift reactor; Settling ability

1. Introduction

Since the 1980s, studies regarding sludge granulation have

focused on upflow anaerobic sludge blanket (UASB) reac-
tors

[1]

. Granulation of anaerobes has been well documented

for decades. In recent years, research have moved interests
into the development of granular activated sludge, as the
feasibility and efficiency of an activated sludge process de-
pends highly on the quality of the sludge formed in the re-
actor. When compared with conventional activated sludge
flocs, the advantages of granular activated sludge are com-
pactness and strength of the structure. It also has good set-
tleability, high capacity for biomass retention and is able to
withstand high organic loading rates

[2]

.

With molasses as the sole carbon source, round-shaped

aerobic granules with an average diameter of 0.6 mm started
to appear in a sequencing batch reactor (SBR) after around
40 days of operation. The aerobic granules taken from the
SBR were stored for 6 weeks without disintegrating

[2]

. In

an aerobic SBR fed with sodium acetate, the granules with
a diameter of 0.3–0.5 mm were formed after 1 month of op-

∗

Corresponding author. Tel.:

[phone];

fax:

[phone].

E-mail address: [e-mail] (H.-Q. Yu).

eration

[3]

. Large-sized aerobic granules with an average di-

ameter of 3.3 mm were produced in an SBR fed with ethanol
and operated at a short hydraulic retention time of 6.75 h
and influent chemical oxygen demand (COD) of 830 mg/l.
A relatively high shear was found to be favourable for gran-
ulation

[4]

. Tay et al.

[5]

found that periodical aerobic star-

vation was an effective trigger for microbial granulation of
activated sludge. Aerobic granules could be formed within
3 weeks in two SBRs fed, respectively, with glucose and
acetate.

However, so far, information about the physical and

chemical characteristics of aerobic granules is still sparse.
Therefore, the main objective of this work was to explore
the physical and chemical properties of granular activated
sludge. It is expected that the information derived from this
work would be useful for the cultivation of granular activated
sludge and its further application for wastewater treatment.

2. Materials and methods

2.1. Experimental set-up and operation

The sequencing batch airlift reactor had a working volume

of 2.2 l (

Fig. 1

). The internal diameter of the down-comer

0032-9592/$ – see front matter © 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.procbio.2004.01.056

646

Y.-M. Zheng et al. / Process Biochemistry 40 (2005) 645–650

effluent

influent

5

downcomer

riser

1

2

3

4

5

6

1. refrigerator

2. peristaltic pump

3. electromagnetic valve

4. time controller

5. air pump

6. mass-flow controller

Fig. 1. Schematic diagram of the SBR.

was 6.0 cm. The riser had a height of 90 cm and an internal
diameter of 4.0 cm. It was positioned at a distance of 2.0 cm
from the bottom of the down-comer. Air was introduced
through an air diffuser by an air pump at the bottom of the
reactor at a superficial air velocity of 2.2 cm/s. The airflow
rate was controlled by a gas-flow controller. The temperature
of the reactor was maintained at 25

◦

C using a ribbon heater

and a temperature controller.

The reactor was operated in a fill-draw mode. The hy-

draulic retention time was 8 h and the influent COD was
kept at 1250 mg/l. The reactor was operated in succes-
sive cycles of 4 h each. One cycle consisted of 10 min
of influent addition, 220 min of aeration, 5 min of set-
tling and 5 min of effluent withdrawal. Effluent was with-
drawn from the port at 50 cm from the bottom of the
reactor.

Activated sludge from a local municipal wastewater

treatment plant was used as inoculum. Six hundred fifty
milliliter of inoculum was seeded to the reactor, resulting in
an initial mixed liquor volatile suspended solids (MLVSS)
of 3000 mg/l in the reactor.

2.2. Media

A synthetic wastewater was used in the study. The com-

position of the synthetic wastewater was as follows: sucrose,
830 mg/l; peptone, 250 mg/l; beef extract, 160 mg/l; NH

4

Cl,

125 mg/l; K

2

HPO

4

, 30 mg/l; CaCl

2

, 20 mg/l; MgSO

4

,

15 mg/l; FeSO

4

·7H

2

O, 15 mg/l and trace element solution,

1.0 ml/l. This gave a total COD of 1250 mg/l. The trace el-
ement solution contained (in mg/l): H

3

BO

3

, 50; ZnCl

2

, 50;

CuCl

2

, 30; MnSO

4

·H

2

O, 50; (NH

4

)

6

Mo

7

O

24

·4H

2

O, 50;

AlCl

3

, 50; CoCl

2

·6H

2

O, 50 and NiCl

2

, 50

[6]

. The influent

pH value was adjusted to 7.0 by the addition of NaHCO

3

and H

2

SO

4

.

2.3. Analytical methods

2.3.1. Microscopy and image analysis

Microbial observation was conducted using an opti-

cal microscope (Olympus CX31). The granular activated
sludge size was measured by an image analysis system
(Image-pro Express 4.0, Media Cybernetics) with an Olym-
pus CX31 microscope and a digital camera (Olympus C5050
Zoom).

2.3.2. Specific gravity

Sucrose was used to make a series of solutions with den-

sities of 1.05, 1.06, 1.07, 1.08, 1.09, 1.10 and 1.11 g/cm

3

.

Ten granules were added into each of the 10 ml tubes
filled with the sucrose solutions of different densities.
Under a quiescent condition, the granules moved up or
down in the tubes depending on the solution density. Thus,
the wet specific gravity of granular activated sludge was
measured.

2.3.3. Specific surface area

The surface area of granular activated sludge was de-

termined using a method based on the principle of dye
adsorption

[7]

. In adsorption of a solute it was assumed

that complete monolayer coverage had occurred when the
isotherm reached a plateau. Thus, the specific surface area
of granular activated sludge can be calculated with the
following equation

[8]

:

S

o

= χN

A

σ

where S

o

is the specific surface area of the adsorbent (m

2

/g

VSS),

χ is the amount of dye adsorbed (mol/g VSS), N

A

is

Avogardro’s number (6

.02 × 10

23

molecules per mol) and

σ

is the area of adsorbent covered by one molecule of adsor-
bate (m

2

per molecule). Rhodamine B was used for mea-

suring the specific surface area of the granular sludge. The
σ value of rhodamine B is 1.85 × 10

−18

m

2

per molecule

[9]

.

2.3.4. Determination of extracellular polymeric
substances (EPS)

The extraction of EPS from granular activated sludge fol-

lowed the procedure described by Sponza

[10]

. Granular

activated sludge was harvested by centrifugation (8100 g,
15 min) and washed with 0.9% NaCl solution twice prior to
extraction. It was suspended in 25 ml of 2% (m/v) EDTA
solution. The sludge suspension was incubated overnight at
4

◦

C. Each sample was centrifuged at 8100 g for 15 min. The

resulting supernatant of the extraction were dialysed using a
membrane of 14,000 molecular weight cut off against ultra
pure water for 2 days at 4

◦

C.

The carbohydrate concentration in EPS was determined

as glucose equivalent using Dubois’ method

[11]

. Protein

concentration was measured as bovine albumin equivalent
using the Lowry method

[12]

.

Y.-M. Zheng et al. / Process Biochemistry 40 (2005) 645–650

647

2.3.5. Hydrophobicity

The hydrophobic nature of the activated sludge particles

was determined by measuring contact angle by axisymmet-
ric drop shape analysis following the method proposed by
Duncan-Hewitt et al.

[13]

. A suspension of sludge contain-

ing biomass was deposited on a cellulosic membrane fil-
ter. Samples were washed three times with deionised wa-
ter, and residual water was removed by filtration. The drop
shape of a sessile distilled water droplet placed on the layer
of biomass was determined using a contact angle measur-
ing apparatus (Powereach JC2000A, Zhongchen Company,
Shanghai).

2.3.6. Other analyses

Measurement of COD, mixed liquor suspended solids

(MLSS), MLVSS and sludge volume index (SVI) were per-
formed using standard methods

[14]

. The settling velocity

was measured by recording the time taken for an individual
granule to fall from a certain height in a measuring cylinder.

3. Results and discussion

Seed sludge and granular sludge (after 60 days of oper-

ation) were obtained from the reactor. Their characteristics
were determined in terms of MLVSS, SVI, and granule size,
density, settling velocity, hydrophobicity, EPS. These data
are summarised in

Table 1

.

3.1. Overall performance of the reactor

Fig. 2A

illustrates that after seeding, the biomass con-

centration in the reactor increased slightly. However, on day
20, the MLVSS values of the sludge decreased steeply at-
tributed to the disorder of the time controller. After recovery
from this failure, the MLVSS of the activated sludge slightly
rose again, and reached 6.0 g/l as the experiment terminated
on day 68. The ratio of MLVSS to MLSS of inoculum was
about 66%. As shown in

Fig. 2B

, after seeding, the ratio in-

creased sharply and reached 90% on day 10. Thereafter, they
were stable between 92 and 94%. The high MLSS/MLVSS
ratio might be attributed to the fact that the inorganic salts

Table 1
Characteristics of seed sludge and granular sludge

Seed sludge

Granular sludge

MLVSS (g/l)

1.0

6.0

MLVSS/MLSS (%)

66

94

SVI (ml/g)

51

23

Average diameter (mm)

<0.1

1.0

Settling velocity (m/h)

<10

18–31

Specific gravity (g/cm

3

)

1.05

1.10

Specific surface area (m

2

/g VSS)

53.3

10.3

Carbohydrate in EPS (mg/g VSS)

15.3

5.9

Protein in EPS (mg/g VSS)

126.7

51.4

Contact angle (

◦

)

35.0

46.3

0

10

20

30

40

50

60

70

0

10

20

30

40

50

60

70

0

10

20

30

40

50

60

70

0

1500

3000

4500

6000

(A)

MLVSS (m

g/l)

60

70

80

90

100

(B)

MLVSS / MLSS (%

)

88

92

96

100

(C)

Operating time (d)

COD re

moval effi

ci

enc

y

(%

)

Fig. 2. Performance of the reactor in 68-day operation: (A) biomass; (B)
ratio of MLVSS/MLSS; (C) COD removal efficiency.

concentration in the synthetic wastewater was less than the
actual municipal wastewater.

As illustrated in

Fig. 2C

, at the initial operating stage the

COD removal efficiency of the reactor was about 88%, as
the seed sludge had a high bioactivity. With the granula-
tion, the COD removal efficiency increased slightly. It even
reached 97% at the end of the experiment.

3.2. Formation and morphology of granules

Seeding sludge had a fluffy, irregular and loose-structure

morphology, as shown in

Figs. 3A and 4A

. The colour of

activated sludge changed from brown to yellow gradually
with the process of the experiment.

648

Y.-M. Zheng et al. / Process Biochemistry 40 (2005) 645–650

Fig. 3. Microscopic observation of (A) seed sludge; (B) granular sludge.

A settling period of 8 min was applied in the initial 3

days to prevent severe wash out of the biomass. From
day 4 to the end, the settling period was kept at 5 min.
The activated sludge had no significant change in the first

Fig. 4. Photographs of (A) seed sludge; (B) granular activated sludge.

Fig. 5. The rod-shaped bacteria in the granular sludge.

week (day 1–7). Tiny granular activated sludge was ob-
served on the wall of the reactor on day 10. From day 18
to day 23, because of the disorder of the time controller,
the sludge worsened and showed bad settling ability. As
a result, a low biomass concentration was maintained and
tiny granules vanished. After recovery from this failure,
the properties of sludge resumed. In the fifth week, tiny
granular activated sludge appeared again. From then on,
the number of granular sludge increased, and its size in-
creased gradually as well. After 60-day operation, as show
in

Fig. 3B

, granular sludge was matured in the reactor.

At this stage, the reactor was dominated by the mature
granular sludge, and little flocs could be observed in the
reactor.

Fig. 4

illustrates the photographs of the seed sludge and

the granular sludge on day 60. The inoculum consisted
of sludge flocs only. It was fluffy and irregular. In addi-
tion, filamentous bacteria were present in the seed sludge.
On the other hand, the granular sludge showed a regu-
lar round-shaped structure, and had an average diameter of
1.0 mm with a range of 0.5–1.2 mm. The granules exhib-
ited a compact structure.

Fig. 5

shows that rod-shaped bac-

teria were predominant and that few filamentous forms was
present in the granules.

3.3. SVI and settling velocity

The SVI of seed sludge was 51 ml/g. After 20-day op-

eration, the SVI increased sharply to 170 ml/g due to the
disorder of the time controller (

Fig. 6

). After the recov-

ery of the reactor, the SVI dropped remarkably to 60 ml/g
in 10 days. Along with the formation of granules, the SVI
decreased gradually. At the termination of the experiment,
the SVI of the sludge decreased to only 23 ml/g, suggest-
ing that the mature granular sludge had an excellent settling
capacity.

The average settling velocities of the granular sludge was

in the range of 18–31 m/h. These values are comparable
with those of anaerobic granules in UASB

[5]

, and are at

least two times greater than those of activated sludge flocs
(lower than 9 m/h)

[15]

. It was observed that after 2 min

Y.-M. Zheng et al. / Process Biochemistry 40 (2005) 645–650

649

0

10

20

30

40

50

60

70

0

30

60

90

120

150

180

SVI (ml/g)

Operating time (d)

Fig. 6. Changing patterns of SVI in 68-day operation.

of settling, the mature granules were well settled, leaving a
clear supernatant in the reactor.

3.4. Specific gravity and specific surface area

As illustrated in

Table 1

, the specific gravity of activated

sludge increased after granulation. It was 1.05 g/cm

3

at the

beginning of the experiment, and increased to 1.10 g/cm

3

af-

ter the granules formed. The significant improvement of spe-
cific gravity of the sludge indicated the granules had highly
dense and compact structure. It benefits for solid–liquid sep-
aration and biomass retaining in the reactor.

Table 1

shows that the specific surface area values were

53.3 and 10.3 m

2

/g VSS for seed sludge and granular sludge,

respectively. This indicates that the formation of granules
reduced the specific surface area of sludge.

3.5. EPS and cell surface hydrophobicity

Polysaccharides can mediate both cohesion and adhesion

of cells, and play a crucial role in maintaining structural in-
tegrity in a community of immobilised cells

[16]

. However,

the precise role of EPS in relation to the formation of gran-
ular activated sludge is not well known.

Table 1

shows that

the concentration of carbohydrate and protein in granular
sludge were 5.9 and 51.4 mg/g VSS, respectively, and that
the corresponding values in the seed sludge were 15.3 and
126.7 mg/g VSS, respectively. The comparison indicates that
the decrease of both carbohydrate and protein concentrations
was concurrent with the formation of granular sludge.

Hydrophobicity of cell surface plays an important role

in the self-immobilisation and attachment of cells to a sur-
face

[17]

. In this study, a significant difference in cell hy-

drophobicity was observed before and after the formation of
granular sludge. The mean contact angle values were 35.0

◦

and 46.3

◦

for seed sludge and granular sludge, respectively.

This suggests that the formation of aerobic granular sludge
is coupled to an increase in the hydrophobicity of the cell
surface. This is in good agreement with the results reported
by Tay et al.

[18]

.

4. Conclusions

This study demonstrated that granular activated sludge

could be cultivated in a sequencing batch airlift reactor fed
with sucrose. After 60-day operation, stable granules with an
average size of 1.0 mm were formed. With the granulation,
the SVI value gradually decreased from 52 to 23 ml/g, while
the specific gravity increased from 1.05 to 1.10 g/cm. The
granular sludge had an excellent settling ability.

Compared with the fluffy, loose and irregular activated

sludge flocs, the granular activated sludge showed a smooth,
compact and round shape structure. The mature granular
sludge had a higher hydrophobicity than the seed sludge,
but the concentration of EPS and specific surface area
were much lower than the corresponding values in the seed
sludge.

Acknowledgements

Authors wish to thank the Trans-Century Training Pro-

gramme Foundation for the Talents, Ministry of Educa-
tion, China and the Anhui Foundation for Excellent Talents,
China, for the partial support of this study.

References

[1] Schmidt JE, Ahring BK. Granular sludge formation in upflow

anaerobic sludge blanket (UASB) reactor. Biotechnol Bioeng 1996;
49(3):229–46.

[2] Morgenroth E, Sherden T, van Loosdrecht MCM, Heijnen JJ,

Wilderer PA. Aerobic granular sludge in a sequencing batch reactor.
Water Res 1997;31(12):3191–4.

[3] Peng DC, Bernet N, Delgenes JP, Moletta R. Aerobic granular

sludge—a case report. Water Res 1999;33(3):890–3.

[4] Beun JJ, Hendriks A, Van Loosdrecht MCM, Morgenroth E, Wilderer

P, Heijnen JJ. Aerobic granulation in a sequencing batch reactor.
Water Res 1999;33(10):2283–90.

[5] Tay JH, Liu QS, Liu Y. Microscopic observation of aerobic granu-

lation in sequential aerobic sludge blanket reactor. J Appl Microbiol
2001;91(1):168–75.

[6] Tay JH, Yan YG. Influence of substrate concentration on microbial

selection and granulation during start-up of upflow anaerobic sludge
blanket reactors. Water Environ Res 1996;68(7):1140–50.

[7] Smith PG, Coackley P. A method for determining specific surface

area of activated sludge by dye adsorption. Water Res 1983;17:595–8.

[8] Allen T. Particle size measurement. London: Chapman & Hall; 1990.
[9] Sorensen BL, Wakeman RJ. Filtration characterisation and specific

surface area measurement of activated sludge by rhodamine B ad-
sorption. Water Res 1996;30(1):115–21.

[10] Sponza DT. Extracellular polymer substances and physicochemical

properties of flocs in steady- and unsteady-state activated sludge
systems. Process Biochem 2002;37(9):983–98.

[11] Dubois M, Gilles KA, Hamilton JK, Rebers PA, Smith F. Calorimetric

method for determination of sugars and related substances. Anal
Chem 1956;28(3):350–6.

[12] Lowrey OH, Resebrough NJ, Farr AL, Randall RJ. Protein measure-

ment with the folin phenol reagent. J Biol Chem 1951;193(2):265–
75.

650

Y.-M. Zheng et al. / Process Biochemistry 40 (2005) 645–650

[13] Duncan-Hewitt WC, Policova Z, Cheng P, Vargha-Butler EI, Neu-

mann AW. Semi-automatic measurement of contact angles on cell
layers by a modified axis symmetric drop shape analysis. Colloids
Surf 1989;42:391–402.

[14] APHA. Standard methods for the examination of water and wastew-

ater. 19th ed. Washington, DC: American Public Health Association;
1995.

[15] Campos JL, Garrido-Fernandez JM, Mendez R, Lema JM. Nitrifi-

cation at high ammonia loading rates in an activated sludge unit.
Bioresour Technol 1999;68:141–8.

[16] Liu Y, Tay JH. Detachment forces and their influence on the struc-

ture and metabolic behavior of biofilms. World J Microb Biot
2001;17(2):111–7.

[17] Zita A, Hermansson M. Effects of bacterial cell surface structures

and hydrophobicity on attachment to activated sludge flocs. Appl
Environ Microbiol 1997;63:1168–70.

[18] Tay JH, Liu QS, Liu Y. Characteristics of aerobic granules grown

on glucose and acetate in sequential aerobic sludge blanket reactors.
Environ Technol 2002;23:931–6.