Aerobic granular-type biomass development in a continuous stirred tank reactor

(This is a sample cover image for this issue. The actual cover is not yet available at this time.)

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Aerobic granular-type biomass development in a continuous stirred tank reactor

N. Morales ⇑, M. Figueroa, A. Mosquera-Corral, J.L. Campos, R. MéndezDepartment of Chemical Engineering, School of Engineering, University of Santiago de Compostela, rúa Lope Gómez de Marzoa s/n, E-15782 Santiago de Compostela, Spain

a r t i c l e i n f o

Article history:Received 20 July 2011Received in revised form 10 January 2012Accepted 10 January 2012Available online 17 January 2012

Keywords:Aerobic granulationContinuous operationOrganic matterSettling

a b s t r a c t

A continuous stirred tank reactor (CSTR) was operated in order to define the appropriate operational con-ditions to obtain aerobic biomass grown in the form of granules. A selection pressure of 10 m/h of solidssettling velocity was imposed while the hydraulic retention time (HRT) was gradually decreased from 6to 1 h to promote the wash-out of the suspended biomass. At HRT values of 6 and 3 h, filamentous bac-teria were dominant in the biomass. The formation of aerobic granular-type biomass was only achievedwhen the HRT was fixed at 1 h. The physical properties of these granules were: sludge volumetric indexof 127 mL/g TSS, density of 11 g VSS/Lgranule, settling velocity of 36–48 m/h and average diameter of6.8 mm. Abundance of filamentous species was different at the tested HRTs. When the HRT was fixedat 6 h filamentous shape bacteria appeared in large amounts, while at the HRT of 1 h the bacterial shapewas mainly bacillus and filamentous shape bacteria were absent. Microbial populations were mainlycomposed by members of Comamonadaceae and Rhodocyclaceae families within b-Proteobacteria, andmembers of the subclass c-Proteobacteria.

� 2012 Elsevier B.V. All rights reserved.

1. Introduction

The biological treatment of effluents in wastewater treatmentplants (WWTPs) is often accomplished by means of the applicationof conventional activated sludge systems where the aerobic organ-ic matter oxidation occurs in a continuous stirred tank reactor.These treatment systems generally require large surface areas forimplantation and the presence of biomass separation systemsdue to the usually poor settling properties of the sludge. The lowsettling velocity of biomass flocs also limits the maximum biomassconcentration inside the reactors to about 3 g VSS/L. Taking intoaccount a typical design value of food to microorganism ratio (F/M) of 1 g/g used in activated sludge processes, the maximumachievable reactor capacity is about 3 kg COD/m3 td.

One possible action to reduce the volume of the required units isthe development of systems based on the improvement of biomassretention, for example, by modifying the operational conditions inorder to obtain granular sludge. Granular growth is a particular caseof biofilm growth, in which no carrier material is required. Bacteriaare self immobilized forming microbial aggregates. Initial applica-tions of granular biomass were associated to anaerobic process: Up-flow Anaerobic Sludge Blanket (UASB), Expanded Granular SludgeBlanket (EGSB) or Internal Circulating (IC) Reactors.

First applications of granular biomass grown in aerobic systemswere based on the Aerobic Upflow Sludge Blanket (AUSB) reactor[1]. This system consisted of two different units: an AUSB reactorwhere high upflow velocities were maintained (hydraulic selectionpressure) to retain only granular biomass and to wash suspendedbiomass; and an external oxygenation chamber where pure oxygenwas added to the recirculated effluent. This kind of operationalstrategy promoted sludge aggregation and also allowed most ofthe produced biomass to be retained in the form of a well-settledsludge blanket [2]. However, this system required high recircula-tion rates in order to obtain suitable oxygen transfer rates to re-move pollutants, which increased the operational costs.

Nowadays, the development of aerobic granules is mostlyfocused on the application of column-type SBR reactors with alarge height to diameter ratio (H/D). This kind of reactors not onlyallows maintaining a high hydraulic selection pressure on thebiomass by applying short settling periods [3] but also other envi-ronmental conditions, such as alternative feast-famine periods andhigh shear forces, which promote the aggregation of biomass [4].Granular SBRs were successfully used to treat various kinds ofeffluents at lab scale [5–7] and, nowadays, there are several pilotscale reactors in operation [8–10]. Nevertheless, the applicationof the concepts of the granular SBR technology for the upgradingof existing WWTPs could be limited by the different operationalconditions and geometry of both column-type SBR andconventional activated sludge reactors. Transforming a continuoussystem into a SBR suitable to obtain aerobic granules is difficult.Then, the research on the feasibility of the production of aerobicgranules in a continuous reactor would open a new perspective

1383-5866/$ - see front matter � 2012 Elsevier B.V. All rights reserved.doi:10.1016/j.seppur.2012.01.024

⇑ Corresponding author. Tel.: +34 881816739; fax: +34 881816702.E-mail addresses: [email protected] (N. Morales), [email protected]

(M. Figueroa), [email protected] (A. Mosquera-Corral), [email protected] (J.L. Campos), [email protected] (R. Méndez).

Separation and Purification Technology 89 (2012) 199–205

Contents lists available at SciVerse ScienceDirect

Separation and Purification Technology

journal homepage: www.elsevier .com/ locate /seppur


to the application of the aerobic granular technology for theimprovement of already existing WWTP.

Therefore, the objective of the present work was to define thesuitable operational conditions to develop granular biomass incontinuous stirred tank reactors (H/D < 1), using a biomass selec-tion system based on the reactor hydrodynamics and progressiveshorting of the HRT applied to the reactor. The performance ofthe aerobic granular system was evaluated in terms of carbonand nitrogen removal efficiencies, and special attention was paidto the physical properties and bacterial populations of the granularbiomass.

2. Materials and methods

2.1. Reactor description

Two cylindrical CSTRs with a working volume of 6 and 3 L (R1and R2, respectively) were used in a continuous mode. Their heightto diameter ratios (H/D) were 0.7 and 0.5 m/m, respectively (Fig. 1).Dissolved oxygen (DO) was supplied continuously by means of airflow (2–6 L/min) through air spargers to promote the formation ofsmall bubbles. A stirrer provided with flat impellers was used insome experimental stages (Table 1). A set of two peristaltic pumpswas used to introduce in a continuous mode the concentratedfeeding solution (flow 1–6 mL/min) and the dilution water (flow10–100 mL/min) while a third peristaltic pump simultaneouslydischarged the effluent (flow 10–100 mL/min). To regulate thethree pumps to provide inflow equal to outflow was almost impos-sible (considering that pumps functioned 24 h per day with lowflows). To avoid unbalances between influent and effluent flows,a level controller turned off/on the effluent pump. It can be esti-mated that pumps were working simultaneously 99% of the time,so the flow per minute was essentially the same. The reactor wasoperated at room temperature (15–25 �C) and without pH control(7.1–8.3).

In order to select the biomass with a high settling velocity, theeffluent was discharged through a tube demi-submerged in the li-quid media which fixed the upflow velocity at 10 m/h (vup). Parti-cles with a settling velocity (vset) smaller than this fixed upflowvelocity will be washed out from the reactor. Biomass with goodsettling properties will be retained in the system, and eventuallyit will form granules, or at least it will become a biomass with bet-ter settling properties. Since the influent flow rate changed duringthe different operational stages, the diameter of the discharge tubewas changed to maintain a constant upflow velocity. Tubes with1.2, 1.6, 2.0 and 2.8 cm of diameter were used.

2.2. Feeding composition

The feeding supplied to the reactors was the liquid fraction ofpig slurry, which was diluted with tap water to fix the influent

COD concentration at the required values (Table 1). The slurrywas collected monthly from a pig farm located in San Marcos (San-tiago de Compostela) and it was sieved through 1 mm diametermesh and settled overnight. It was kept refrigerated at 4 �C to avoiddegradation before entering the reactor.

2.3. Inoculum

Reactor R1 was initially inoculated with flocculent activatedsludge collected from a municipal WWTP. A sample of 500 mL ofsettled sludge was added to the reactor. On day 205 of operationpig slurry was used as inoculum of the reactor. The initial biomassconcentration was of 0.75 g VSS/L. Slurry itself was used as inocu-lum for reactor R2.

2.4. Operational conditions

Reactors R1 and R2 were operated in a continuous mode underaerobic conditions during 260 and 120 days, respectively. ReactorR1 was operated in three stages differenced by the decreasing ofHRT with the consequent increase of OLR. In the case of R2, theoptimum obtained conditions to form aerobic granules in R1 wereapplied for its start-up (Table 1).

2.5. Analytical methods

Solids concentration (TSS, VSS), sludge volume index (SVI), set-tling velocity, pH value, nitrate, nitrite and ammonia concentra-tions were determined according to the Standard Methods [11].DO concentration was measured with a dissolved oxygen electrode(DurOx 325) connected to a meter (WTW Oxi 340i). Density of thegranules as g VSS/Lgranule was measured using the dextran bluemethod described by Beun et al. [12]. Chemical Oxygen Demand(COD) was determined by a semi-micro method [13]. Concentra-tions of total organic carbon (TOC) and inorganic carbon (IC) weremeasured with a Shimadzu analyzer (TOC-5000). The morphologyand size distribution of the granules were measured regularly byusing an image analysis procedure [14] with a stereomicroscope(Stemi 2000-C, Zeiss). The elemental analysis of the biomass wascarried out with a dried biomass sample by means of the ElementalThermo Finnigan model Flash 1112.

Microbial populations were followed by the Fluorescence inSitu Hybridization (FISH) technique. Biomass samples from thereactor were collected, disrupted and fixed, according to the proce-dure described by Amann et al. [15] with 4% paraformaldehydesolution. Hybridization was performed at 46 �C for 90 min. Bacte-rial cells were hybridized with the following FISH probes: EU-B338mix, for the domain Bacteria, ALF1b, BET42a and GAM42aprobes for the a-, b- and c-subclasses of the Proteobacteria. Morespecific probes were: Cte probe, for Comamonas spp., Acidovoraxspp., Hydrogenophaga spp. and Aquaspirillum spp., Rhoc-1425 forsome member of family Rhodocyclaceae, Pae997 probe, specificfor Pseudomonas spp., and NEU653 probe, that targets most halo-philic and halotolerant Nitrosomonas spp. Details on oligonucleo-tide probes are available at probeBase [16]. The probes were 50

labeled with the fluorochromes FITC and Cy3. Fluorescence signalswere recorded with an acquisition system coupled to an Axioskop2 epifluorescence microscope (Zeiss, Germany). In order to quan-tify the bacterial population, for each hybridization experimentat least 20 randomly chosen images were recorded, and the ratioof the area of those cells labeled by the specific probe to the areaof all bacteria stained by DAPI was determined by digital imageanalysis using Image ProPlus (Media Cybernetics) [17].

Fig. 1. Reactor layout. Feeding pumps (1 and 2), effluent discharging pump (3),stirrer (4), air pump (5), spargers (6), flat impeller (7), effluent discharge tube (8)and liquid level controller (9). Solids settling velocity (vset) and upflow liquidvelocity (vup).

200 N. Morales et al. / Separation and Purification Technology 89 (2012) 199–205


3. Results and discussion

3.1. Reactor performance

The organic matter removal efficiency, expressed in terms of sol-uble COD, ranged between 50–80% and 30–60% for R1 and R2,respectively (Fig. 2A and B). A similar maximum removal efficiencyof soluble COD close to 80% was also obtained by Vanotti et al. [18]and Karakashev et al. [19] during aerobic and anaerobic treatmentof pig slurry, respectively. From previous works it can be consideredthat the fraction of the slowly biodegradable or non-biodegradableorganic matter present in the pig slurry [20] can achieve values from10% to 55% [21,22]. Nitrogen removal efficiency in both reactors var-ied between 10% and 15% during the whole operational period. Nei-ther nitrite nor nitrate was measured in the effluent; therefore,nitrogen removal can be attributed only to biomass assimilation.Since the nitrogen removal efficiency was very low, a post-treat-ment of the effluent would be necessary to fulfill the disposalrequirements. The absence of nitrifying activity could be due tothe low solids retention time (SRT) values obtained. These valueswere always lower than 2 days even when granules with good set-tling properties were developed in the reactor.

No improvement in the biomass retention was observed dur-ing Stage I of R1 (Fig. 3A) since the biomass concentration inthe effluent was very similar to that present inside the reactor(0.4–0.6 g VSS/L). Tijhuis et al. [14] stated that, in order to devel-op biofilms on carrier materials in an airlift suspension reactor,the hydraulic retention time must be shorter than the inverseof the maximum growth rate of the suspended bacteria. Follow-ing this concept in the present work the HRT was reduced to3 h during Stage II although no positive effect on the biomassretention was observed. Then, mechanical stirring was switchedoff to decrease applied shear stress on the biomass and try to pro-mote the development of biomass aggregates. Under these condi-tions, biomass concentration rapidly increased inside the reactor,reaching values of 6 g VSS/L, while biomass concentration in the

effluent ranged between 0.2 and 0.4 g VSS/L. The biomass concen-tration in the reactor increased up to 10 g VSS/L in spite of therestoration of the mechanical stirring from day 132. As the phys-ical characteristics of biomass were unsuitable due to the pres-ence of filamentous bacteria, it was sieved and the liquidfraction obtained was used as inoculum on Stage III. However,the growth of filamentous bacteria took place again and, in only5 days, the biomass concentration reached again values of 10 gVSS/L. In addition to this, the dissolved oxygen concentration dur-ing this period was below 2 mg O2/L. In previous research works[23] it was identified the low dissolved oxygen concentration as apossible cause of filamentous growth in aerobic granules. On day205, this biomass was discarded and the pig slurry was directlyused as inoculum to re-start the reactor in the operational condi-tions of Stage III. The pig slurry was directly used as inoculumdue to the fact that in the initial operational days the inoculatedsludge was observed to be fully removed from the system makingit similar to seed or not the reactor. Having in mind that the pigslurry contained a certain amount of organic solids it was as-sumed that these solids would act as an inoculum. Under theseoperational conditions the appearance of aerobic granular-typebiomass took place on day 227 and reached a concentration inthe reactor around 1.0–1.5 g VSS/L at the end of this stage, whilethe solids concentration in the effluent was 0.15 g VSS/L. Similarresults were obtained during the operation of R2 (Fig. 3B), wherepig slurry was used as inoculum, and biomass concentrationreached values around 1.0–1.2 g VSS/L, while solids concentrationin the effluent did not exceed 0.15 g VSS/L. First granules in thisreactor were observed after 24 days of operation. Variations inthe VSS concentration and in the average size of the granulesare usual in aerobic granular reactors, especially during the startup period [4]. The application of low HRTs (lower than the in-verse of the maximum heterotrophic bacterial growth rate whichranges for heterotrophic bacteria around 3.0–13.2 d�1 [24]) al-lowed the wash out of microorganisms that did not form aggre-gates or biofilms [14].

Table 1Operational conditions of R1 and R2.

Reactor R1 R2

Stages S I S II S III S I

Days (d) 0–78 79–108 109–131 132–184 185–260 1–120HRT (h) 6 3 1 1Mechanical stirring Yes Yes No Yes Yes YesCOD (mg COD/L) 600 600 600 600 200–500 200–250OLR (g COD/L d) 2.4 4.8 4.8 4.8 4.8–12.0 4.8–6.0NLR (g N/L d) 0.5 0.8 1.2 1.2 0.9–2.2 1.4–1.7DO (mg/L) 5.0 3.8 3.0 2.5 6.1 8.6VSS influent (g/L) 0.06–0.19 0.14–0.26 0.07–0.11 0.06–0.10 0.02–0.10 0.02–0.06pH influent 7.1–8.1 7.1–8.1 7.9–8.1 7.5–8.1 6.9–7.3 7.3–8.3Effluent tube diameter (cm) 1.2 1.6 1.6 1.6 2.8 2.0

BAS I S II S III

0

200

400

600

800

1000

0 40 80 120 160 200 240 280time (d)

mg

CO

D/L

0

25

50

75

100

% C

OD

0

100

200

300

400

0 20 40 60 80 100 120time (d)

mg

CO

D/L

0

25

50

75

100

% C

OD

S I

Fig. 2. COD concentration in the influent (j) and effluent (s) and percentage of removed COD (—) in reactors R1 (A) and R2 (B).

N. Morales et al. / Separation and Purification Technology 89 (2012) 199–205 201


The average biomass yield coefficient estimated during Stages Iand II of R1 was about 0.55 g VSS/g CODremoved. This value agreeswith that of 0.4–0.6 g VSS/g CODremoved observed in continuousactivated sludge systems [25] and SBR systems fed during the reac-tion period [26]. In those stages when aerobic granular type bio-mass was present (Stage III of R1 and Stage I of R2), the averagebiomass yield coefficient calculated decreased down to 0.44 and0.37 g VSS/g CODremoved, respectively, which is higher than thatof 0.2 g VSS/g CODremoved obtained for aerobic granules cultivatedin SBRs [21,27–29]. The difference between both values could beexplained by the low SRT achieved in this work due to the deficientbiomass retention.

3.2. Biomass characteristics

The biomass present in reactor R1 during Stage I and the firstdays of Stage II was mainly filamentous biomass, with poor settlingproperties (Fig. 4A). Once the mechanical stirring was switched off,the biomass formed a gelatinous structure that occupied most ofthe reactor (Fig. 4B). This kind of biomass was again developedduring first days of Stage III until it was removed, and pig slurrywas used as inoculum. In Stage III (Fig. 4C) and during the opera-tion of R2 (Fig. 4D–F) a biomass aggregation process occurred,and aerobic granules dominated the reactor. Aggregates had awhite color, with a smooth surface, and a regular round shape.

A

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

0.0

2.0

4.0

6.0

8.0

10.0

12.0

0 40 80 120 160 200 240 280

Dia

met

er (m

m)

g V

SS

/L

Time (d)

S I S II S III

B

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0 20 40 60 80 100 120

Dia

met

er (m

m)

g V

SS

/L

Time (d)

S I

Fig. 3. Solids concentration inside the reactor (d) and effluent (s) and average granule diameter (j) in millimeters in R1 (A) and R2 (B).

Fig. 4. Biomass images from reactors R1 on days: (a) 44, (b) 185, (c) 247; and R2 on days: (d) 16, (e) 24, (f) 77.



The average diameter of the aggregates continuously increasedduring Stage III and during the operation of R2 (Fig. 3). The averagediameter reached a value of 4 mm in Stage III and of 6.8 mm duringthe operation of R2 with a granular size distribution similar to thatfound previously in granular SBRs fed with industrial wastewater[5]. Granules with a diameter larger than 6.0 mm in the final daysof operation of R2 represented about 96% of the biomass volume(Fig. 5B) while on day 42 represented only 15% of it (Fig. 5A). Thesize of the granules formed in R1 and R2 was larger than those cor-responding to aerobic granules formed in a SBR fed also with di-luted pig slurry [21].

During the formation of the aggregates, the SVI10 decreasedgradually from 700 to 127 mL/g TSS. This value was larger thanthe SVI10 value corresponding to aerobic granules cultivated inSBR reactors, where values around 30–40 mL/g TSS were reached[21,30]. Density of the aggregates in R2 varied between 7 and11 g VSS/Lgranule. These values are relatively low compared to thatof 43.5 g VSS/Lgranule corresponding to granules formed in SBR reac-tors treating pig slurry [21]. However, the settling velocity values ofthe granules formed in both reactors varied between 36 and 48 m/hwhich are similar to those of aerobic granules cultivated in SBRreactors [31,32] and are remarkably higher than the typical valuesof settleability of activated sludge [33]. The biomass produced inthis continuous system will be more easily separated from theliquid phase, reducing the need of big settlers, and it will be more

easily dewatered due to the high cell hydrophobicity of aerobicgranules [34], facilitating the management and disposal of thesludge produced during wastewater treatment.

Elemental composition of aerobic granules observed in R1 wasCH1.68N0.18O0.55S0.003, while in R2 was CH2.10N0.20O0.64S0.004. Both re-sults were similar to those obtained in aerobic granular reactorswith feeding media characterized by a C/N ratio of 5 g/g by Liuet al. [34] and Yang et al. [35].

The surface organic loading rates applied and treated by aerobicgranules ranged between 7–12 and 24–40 g COD/m2 d for R1 (aver-age diameter 4.0 mm) and R2 (average diameter 6.8 mm), respec-tively. Figueroa et al. [21] using pig slurry as wastewater in a SBR(granules with an average diameter of 5 mm) reached values about17 and 20 g COD/m2 d. Westerman et al., [36] using upflow aeratedbiofilms with an specific surface area of 140 m2/m3 for the treat-ment of flushed swine manure achieved values about 34 and 47 gCOD/m2 d.

3.3. Microbial populations

Samples of sludge from R1 (day 64, Stage I) and R2 (day 120)were collected. These samples were chosen because they representthe two different kinds of biomass formed in the continuous oper-ated systems (fluffy biomass in Stage I and granular biomass fromR2). A set of general probes was applied to the sample from R1 to

BA

0

5

10

15

20

25

0 1 2 3 4 5 6 7 8 9 10 11 >12

% v

olu

me

diameter (mm)

0

5

10

15

20

25

0 1 2 3 4 5 6 7 8 9 10 11 > 12

% v

olu

me

diameter (mm)

Fig. 5. Volumetric distribution of the granules in R2 on day 42 (A) and on day 102 (B).

Fig. 6. Images of biomass samples from R1 analyzed with FISH, Stage I. (a) Filamentous and bacillus shape bacteria of class b-Proteobacteria (Bet42a probe; Cy3, red) and DAPI(blue). (b) Bacteria of class c-Proteobacteria (GAM42a probe; Cy3, red) and DAPI (blue). (c) Filamentous bacteria of the family Comamonadaceae (Cte probe; Cy3, red) and allbacteria (EUBmix; FITC, green). (d) Bacteria of family Rhodocyclaceae (Rhoc-1425 probe; Cy3, red) and all bacteria (EUBmix; FITC, green). In all cases, the bar represents 25 lm.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)



detect the main classes of bacteria involved in the process (EU-B338mix, ALF1b, BET42a and GAM42a probes). In Stage I (Fig. 6),the hybridized bacteria belonged mainly to b- and c-Proteobacteriasubclasses and represented an important fraction in comparisonwith all the positives from EUB338I probe (30% in biovolume inboth cases) while no a-Proteobacteria were detected. An importantfraction of the b-Proteobacteria detected were filamentous shapebacteria, and also hybridized with Cte probe, specific for Comamon-as spp., Acidovorax spp., Hydrogenophaga spp. and Aquaspirillumspp. According to the phylogenetic affiliation of identified filamen-tous b-Proteobacteria, these bacteria can be classified within themorphotype Type 1701, Type 0041/0675, Sphaerotilus natans orLeptothrix [37], however, considering the probe coverage, these fil-amentous bacteria are probably organisms of genus Curvibacterwithin family Comamonadaceae. These bacteria are responsible offilamentous bulking episodes in activated sludge plants and de-crease of sludge settleability. This observation corresponds withthe settling properties of the biomass on Stage I when mainly acti-vated sludge flocs with poor settling properties (SVI30 above250 mL/g TSS) were produced.

Positive results were also obtained with Rhoc-1425 probe, de-signed for family Rhodocyclaceae within b-Proteobacteria (Fig. 6D).Positive results of these probe corresponded with the bacillus-shapebacteria that hybridized with probe Bet42a. This family comprisesbacteria belonging to the genera Azoarcus, Thauera, and Zoogloeathat are abundant and believed to be the main denitrifiers in waste-water treatment plants [38] although in this study the denitrifyingactivity was negligible.

The Pae997 probe, specific for Pseudomonas spp., was applied inorder to test if these bacteria were the dominant c-Proteobacteriain the sample; however, no positive results were obtained.

Results from the FISH analysis of the sample from R2 (biomassmainly in form of aerobic granules) were different from previousdetailed ones (Fig. 7). Although the hybridized bacteria belongedmainly to b- and c-Proteobacteria classes the bacterial shape wasmainly bacillus, and no filamentous shape bacteria were observed.An estimated biovolume percentage calculated for this sample was

20% for b-Proteobacteria and 15% for c-Proteobacteria. Positive re-sults for Cte and Rhoc-1425 probes were obtained, covering themajor part of observed b-Proteobacteria. In this case, only a smallfraction of bacteria hybridized with Pae997 probe (around 1%).

No positive results were obtained with specific probes forammonia oxidizing bacteria Nitrosomonas (AOB) (NEU653). Theseresults agree with the fact that no ammonia oxidation activitywas detected in the reactor.

4. Conclusions

Microbial aggregates were formed using a continuous stirredtank reactor with similar geometry to the activated sludge reactorsused in WWTPs. The formation of these aggregates was achievedwhen a HRT of 1 h and a hydraulic pressure of 10 m/h of settlingvelocity in the effluent tube were applied to the system. At thispoint, OLRs as high as 4.8–6.0 g COD/L d were treated with removalpercentages around 60%. Nitrogen removal varied between 10%and 15%, and can be attributed completely to biomass growthwhile nitrification and denitrification processes did not occur inthe reactor. In case nitrogen disposal requirements had to be ful-filled, a post-treatment would be necessary.

The aggregates formed in the system had a SVI10 of 127 mL/gTSS and a density of 11 g VSS/Laggregate, worse values comparedwith aerobic granules formed in SBR reactors. However, theseaggregates showed an average diameter as high as 7 mm and alarge settling velocity (36–48 m/h). These high values are compa-rable with those from aerobic granules formed in SBR reactorsand allowed the biomass to remain inside the reactor. The averagebiomass yield coefficient of these microbial aggregates was of0.37–0.44 g VSS/g COD. Besides the effect of hydrodynamic condi-tions, the origin of the inoculum should be taken into account toachieve microbial aggregation.

The presence of filamentous microorganisms could cause thefailure in the microbial aggregation in R1 (Stages I and II), sinceno filaments were observed in R2. In both cases, the microbial

Fig. 7. Images of biomass samples from R2 analyzed with FISH. (a) Bacillus shape bacteria of class b-Proteobacteria (Bet42a probe; Cy3, red) and DAPI (blue). (b) Bacteria ofclass c-Proteobacteria (GAM42a probe; Cy3, red) and DAPI (blue). (c) Bacteria of family Comamonadaceae (Cte probe; Cy3, red) and all bacteria (EUBmix; FITC, green). (d)Bacteria of family Rhodocyclaceae (Rhoc-1425 probe; Cy3, red) and all bacteria (EUBmix; FITC, green). In all cases, the bar represents 25 lm. (For interpretation of thereferences to color in this figure legend, the reader is referred to the web version of this article.)



population was mainly composed by members of Comamonadaceaeand Rhodocyclaceae families within b-Proteobacteria, and membersof class c-Proteobacteria.

More research is needed to establish the operational conditionswhich allow improving both physical properties of the biomassand the nitrogen removal efficiency.

Acknowledgments

This work was funded by the Spanish Government (TOGRANSYSCTQ2008-06792-C02-01 and NOVEDAR Consolider project CSD2007-00055). Authors want to thank Mar Orge, Mónica Dosil and MiriamVieites for their support in the analytical techniques.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.seppur.2012.01.024.

References

[1] K. Mishima, K. Nakamura, Self-immobilization of aerobic activated sludge – apilot study of the aerobic upflow sludge blanket process in municipal sewagetreatment, Water Science and Technology 23 (1991) 981–990.

[2] K.R. Sharma, J.C. Huang, High-rate wastewater treatment by aerobic upflowsludge blanket reactor: system performance and characteristics, Water Scienceand Technology 50 (2004) 193–200.

[3] Y. Liu, J.H. Tay, State of the art of biogranulation technology for wastewatertreatment, Biotechnology Advances 22 (2004) 533–563.

[4] J.J. Beun, A. Hendriks, M.C.M. van Loosdrecht, E. Morgenroth, P.A. Wilderer, J.J.Heijnen, Aerobic granulation in a sequencing batch reactor, Water Research 33(1999) 2283–2290.

[5] B. Arrojo, A. Mosquera-Corral, J.M. Garrido, R. Mendez, Aerobic granulationwith industrial wastewater in sequencing batch reactors, Water Research 38(2004) 3389–3399.

[6] N. Schwarzenbeck, J.M. Borges, P.A. Wilderer, Treatment of dairy effluents in anaerobic granular sludge sequencing batch reactor, Applied Microbiology andBiotechnology 66 (2005) 711–718.

[7] M. Figueroa, A. Mosquera-Corral, J.L. Campos, R. Mendez, Treatment of salinewastewater in SBR aerobic granular reactors, Water Science and Technology 58(2008) 479–485.

[8] L.M.M. de Bruin, H.F.R. van der Roest, M. de Kreuk, M.C.M. van Loosdrecht, S.Bathe, B.S. McSwain, N. Schwarzenbeck, Promising results pilot researchaerobic granular sludge technology at WWTP Ede, in: Aerobic Granular Sludge,IWA Publishing, London, 2005, pp. 135–142.

[9] M. Inizan, A. Freval, J. Cigana, J. Meinhold, Aerobic granulation in a sequencingbatch reactor (SBR) for industrial wastewater treatment, Water Science andTechnology 52 (2005) 335–343.

[10] B.J. Ni, W.M. Xie, S.G. Liu, H.Q. Yu, Y.Z. Wang, G. Wang, X.L. Dai, Granulation ofactivated sludge in a pilot-scale sequencing batch reactor for the treatment oflow-strength municipal wastewater, Water Research 43 (2009) 751–761.

[11] APHA, Standard methods for the examination of water and wastewater,American Public Health Association/American Water Works Association/Water Environment Federation, Washington DC, USA, 1998.

[12] J.J. Beun, J.J. Heijnen, M.C.M. van Loosdrecht, N-removal in a granular sludgesequencing batch airlift reactor, Biotechnology and Bioengineering 75 (2001)82–92.

[13] M. Soto, M.C. Veiga, R. Mendez, J.M. Lema, Semi-micro C.O.D. determinationmethod for high-salinity wastewater, Environmental Technology Letters 10(1989) 541–548.

[14] L. Tijhuis, M.C.M. van Loosdrecht, J.J. Heijnen, Formation and growth ofheterotrophic aerobic biofilms on small suspended particles in airlift reactors,Biotechnology and Bioengineering 44 (1994) 595–608.

[15] R. Amann, W. Ludwig, K.H. Schleifer, Phylogenetic identification and in-situdetection of individual microbial-cells without cultivation, MicrobiologicalReviews 59 (1995) 143–169.

[16] A. Loy, F. Maixner, M. Wagner, M. Horn, probeBase – an online resource forrRNA-targeted oligonucleotide probes: new features, Nucleic Acids Research35 (2007) (2007) D800–D804.

[17] G. Crocetti R, J.F. Banfield F, J. Keller, P.L. Bond, L.L. Blackall, Glycogen-accumulating organisms in laboratory-scale and full-scale wastewatertreatment processes, Microbiology 148 (2002) 3353–3364.

[18] M.B. Vanotti, A.A. Szogi, P.D. Millner, J.H. Loughrin, Development of a second-generation environmentally superior technology for treatment of swinemanure in the USA, Bioresource Technology 100 (2009) 5406–5416.

[19] D. Karakashev, J.E. Schmidt, I. Angelidaki, Innovative process scheme forremoval of organic matter, phosphorus and nitrogen from pig manure, WaterResearch 42 (2008) 4083–4090.

[20] J.H. Shin, S.M. Lee, J.Y. Jung, Y.C. Chung, S.H. Noh, Enhanced COD and nitrogenremovals for the treatment of swine wastewater by combining submergedmembrane bioreactor (MBR) and anaerobic upflow bed filter (AUBF) reactor,Process Biochemistry 40 (2005) 3769–3776.

[21] M. Figueroa, A. Val Del Río, J.L. Campos, A. Mosquera-Corral, R. Méndez,Treatment of high loaded swine slurry in an aerobic granular reactor, WaterScience and Technology 63 (2011) 1808–1814.

[22] A. Magrí, M. Givernau, G. Baquerizo, M. Viñas, F.X. Prenafeta-Boldú, X. Flotats,Batch treatment of liquid fraction of pig slurry by intermittent aeration:process simulation and microbial community analysis, Journal of ChemicalTechnology and Biotechnology 84 (2009) 1202–1210.

[23] Y. Liu, Q.S. Liu, Causes and control of filamentous growth in aerobic granularsludge sequencing batch reactors, Biotechnology Advances 24 (2006) 115–127.

[24] M. Henze, C.P.L. Grady Jr, W. Gujer, A general model for single-sludgewastewater treatment systems, Water Research 21 (1987) 505–515.

[25] R.L. Droste, Theory and practice of water and wastewater treatment, Wiley,New York, 1997.

[26] E. Klimiuk, D. Kulikowska, Organics removal from landfill leachate andactivated sludge production in SBR reactors, Waste Management 26 (2006)1140–1147.

[27] Y. Liu, Z.W. Wang, J.H. Tay, A unified theory for upscaling aerobic granularsludge sequencing batch reactors, Biotechnology Advances 23 (2005) 335–344.

[28] Y. Liu, Y.Q. Liu, Z.W. Wang, S.F. Yang, J.H. Tay, Influence of substrate surfaceloading on the kinetic behaviour of aerobic granules, Applied Microbiology andBiotechnology 67 (2005) 484–488.

[29] Y.Q. Liu, J.H. Tay, Influence of cycle time on kinetic behaviors of steady-stateaerobic granules in sequencing batch reactors, Enzyme and MicrobialTechnology 41 (2007) 516–522.

[30] N. Schwarzenbeck, R. Erley, B.S. Mc Swain, P.A. Wilderer, R.L. Irvine, Treatmentof malting wastewater in a granular sludge sequencing batch reactor (SBR),Acta Hydrochimica et Hydrobiologica 32 (2004) 16–24.

[31] S.S. Adav, D.J. Lee, K.Y. Show, J.H. Tay, Aerobic granular sludge: recentadvances, Biotechnology Advances 26 (2008) 411–423.

[32] A. Jang, Y.H. Yoon, I.S. Kim, K.S. Kim, P.L. Bishop, Characterization andevaluation of aerobic granules in sequencing batch reactor, Journal ofBiotechnology 105 (2003) 71–82.

[33] B. Jin, B.-M. Wilén, P. Lant, A comprehensive insight into floc characteristicsand their impact on compressibility and settleability of activated sludge,Chemical Engineering Journal 95 (2003) 221–234.

[34] Y. Liu, S.F. Yang, J.H. Tay, Elemental compositions and characteristics of aerobicgranules cultivated at different substrate N/C ratios, Applied Microbiology andBiotechnology 61 (2003) 556–561.

[35] S.-F. Yang, J.-H. Tay, Y. Liu, Respirometric activities of heterotrophic andnitrifying populations in aerobic granules developed at different substrate N/COD ratios, Current Microbiology 49 (2004) 42–46.

[36] P.W. Westerman, J.R. Bicudo, A. Kantardjieff, Upflow biological aerated filtersfor the treatment of flushed swine manure, Bioresource Technology 74 (2000)181–190.

[37] H. Nielsen, H. Daims, H. Lemmer, FISH handbook for biological wastewatertreatment, IWA Publishing, London, UK, 2009.

[38] M. Wagner, A. Loy, Bacterial community composition and function in sewagetreatment systems, Current Opinion in Biotechnology 13 (2002) 218–227.


Aerobic granular-type biomass development in a continuous stirred tank reactor

Documents

Transcript of Aerobic granular-type biomass development in a continuous stirred tank reactor