Bioresource Technology 186 (2015) 276–285

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Rheological and fractal hydrodynamics of aerobic granules H.I. Tijani a, N. Abdullah b, A. Yuzir c,⇑, Zaini Ujang d a

Faculty of Biosciences and Medical Engineering, Universiti Teknologi Malaysia, 81310 Johor Bahru, Malaysia Palm Oil Research Center, Faculty of Biosciences and Medical Engineering, Universiti Teknologi Malaysia, 81310 Johor Bahru, Malaysia c Centre for Environmental Sustainability and Water Security (IPASA), Faculty of Civil Engineering, Universiti Teknologi Malaysia, 81310 Johor Bahru, Malaysia d Ministry of Education Malaysia, Blok E8, Kompleks E, Pusat Pentadbiran Kerajaan Persekutuan, 62604 Putrajaya, Malaysia b

h i g h l i g h t s  Microbial bioactivity are maintained during granulation.  Granules constitutes significant structural framework.  Scaling relationships based on fractal geometry are vital in granulation.

a r t i c l e

i n f o

Article history: Received 31 December 2014 Received in revised form 26 February 2015 Accepted 27 February 2015 Available online 17 March 2015 Keywords: Settling Fractal dimension Aggregation Aerobic granulation Palm oil mill effluent

a b s t r a c t The structural and hydrodynamic features for granules were characterized using settling experiments, predefined mathematical simulations and ImageJ-particle analyses. This study describes the rheological characterization of these biologically immobilized aggregates under non-Newtonian flows. The second order dimensional analysis defined as D2 = 1.795 for native clusters and D2 = 1.099 for dewatered clusters and a characteristic three-dimensional fractal dimension of 2.46 depicts that these relatively porous and differentially permeable fractals had a structural configuration in close proximity with that described for a compact sphere formed via cluster–cluster aggregation. The three-dimensional fractal dimension calculated via settling-fractal correlation, U / lD to characterize immobilized granules validates the quantitative measurements used for describing its structural integrity and aggregate complexity. These results suggest that scaling relationships based on fractal geometry are vital for quantifying the effects of different laminar conditions on the aggregates’ morphology and characteristics such as density, porosity, and projected surface area. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction Different safe biological systems have been employed in recent years at minimizing the impact of toxic waste on the ecosystem. Aerobic granulation is a unique microbial immobilization technology developed in recent years based on the activated sludge system to produce aerobic granules for wastewater treatments. These granules constitute a specialized microbial biofilm composed of densely packed aggregates of self-immobilized cells which may degrade organic and inorganic components of the

⇑ Corresponding author. Tel.: +60 7 5538687 (A. Yuzir). E-mail addresses: [email protected] (H.I. Tijani), [email protected]. my (N. Abdullah), [email protected] (A. Yuzir), [email protected] (Z. Ujang). http://dx.doi.org/10.1016/j.biortech.2015.02.107 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.

waste. Previous researches indicated that biofilm is one of the most efficient techniques for wastewater treatment as compared to its conventional suspended activated sludge systems. This is due to its compact structure developed under the sequencing batch reactor (SBR) operations (Tay et al., 2004; Yang et al., 2007; Abdullah et al., 2011). Aerobic granulation system fed with palm oil mill effluent (POME) as substrate previously developed by Abdullah et al. (2011) in an SBR was proven efficient for treating high strength effluent discharges. Excellent reactor performances was observed in settling ability, mass transfer efficiency, differential sludgeeffluent phase separation, biomass retention, resilience to shock loadings and bioactivity in stable granules formation using POME. Thus, the use of activated sludge reactors for treatment applications of industrial and pharmaceutical wastes under cyclic aerobic systems serves as succeeding alternative for effluent treatments.

277

H.I. Tijani et al. / Bioresource Technology 186 (2015) 276–285

The adhesive principle of microorganisms to defined surfaces, suspended particles and other microorganisms is pervasive in nature and is required for the proper operation of bioreactors. Within these suspended-growth bioreactors, microbial self-induced flocculation occurs through cell-to-cell attachment, resulting in the formation of rapid-settling granules which enables efficient removal of biomass from the fluid suspension. The irregularity in shape and compressibility features common to these biological aggregates, makes it reprehensible to compute granule geometrical characteristics such as density, porosity and fluid collection properties. These granules’ characteristics are contributive to efficient design of granulation system aimed for single-column operation. For an accurate computations, fractal dimensional geometry has been established to define these highly frail structures so as to promote the understanding of the influence of a wide array of granulation phenomena such as turbulence, shear abrasion and fluid kinetics. The fractal geometry of granular aggregates can be determined via scaling relationship. For Euclidean aggregates, the three dimensional fractal dimension (D3) is 3; however, granular aggregates exhibit significantly lower fractal dimensions. Simulation models developed to describe the random processes of aggregate formation is related to the magnitude of their fractal dimensions. Granules cultivated through particle-to-cluster and clusterto-cluster aggregation have three-dimensional fractal dimensions in the range of 2.5–3.0 (Logan and Wilkinson, 1991; Maggi, 2007). The weak structure of immobilized aggregates can be attributed to the random mobility of its adhesive particles due to their inability to penetrate the aggregates resulting in particle collision and adherence on the aggregate exterior. This mode of aggregation is termed as substrate-limited granular immobilization (SLGI) and results in formation of well-defined aggregates with D3 values greater than 2. For granular aggregates cultivated under laminar flow systems, the presence of two distinct regimes of flocculation indicates that fractal scaling calculations can be used to describe the formation mechanism and structural compactness. Lin et al. (2003) proposed that fractal dimensions are proximally identical for aggregates developed under similar reactor conditions. However, the simulation models defined for fractal calculations have varied with different forms of microbial aggregates. This indicates that such generalization are yet to be established because bioflocs aggregation via particle collisions experienced varying flocculation mechanisms, such as sub-mechanical and turbulent shear as well as different sedimentation. According to Mohammed et al. (2013), the microbial networks that constitute the granules formation are formed by intra- and inter-linked bridges of the cellular surface appendages. Although the adhesive strength of these immobilized microbial consortium is not uniform throughout the granules’ layers, specific internal and external forces tends to promote their coagulation into welldefined structures. Thus, the prospect for other flocculation mechanisms, the unpredictability in the microbes’ consortium and population generate variations in the aerobic granulation systems is characterized by fractal analyses. This research aims to determine if a universal fractal equation can be developed for in-stored granules cultivated from palm oil mill effluent (POME). Recent studies have proven that these microbial aggregates are efficient in for successful treatment of agrobased wastewater such as POME (Abdullah et al., 2013, 2011). In this study, additional characterization of the rheological hydrodynamics of granules will be provided. The fractal relationships are also compared to its Euclidean replicas under similar hydraulic systems. We calculate the granules’ three dimensional geometries and discuss the transcendence of fractal calculations via settling-fractal simulations.

2. Methods 2.1. Bio-flocculation characteristics of activated sludge Bio-flocculation was established by exerting a low centrifugal force which slowly mixes the sludge sample at 15 rpm for 10 min. The fluid viscosity of the sub-mechanical granulation system was also evaluated to reflect the settling ability, compressibility and structural compactness of the microbial association within granules. The concentrated mass (mo) of the fractal aggregates was determined by using electronic microbalance (AEM-5200, Shimadzu, Japan). The spatial volume (vo) was determined from its diameter sizes (do) and their projected surface area as visualized by the image analysis system. The density of the aerobic granules (qo) were calculated via the mathematical expression as given in Eq. (1):

mo

qo ¼

ð1Þ

vo

The mass of an emptied graduated measuring cylinder (mc), volume of the mixed-liquor (vl) and the coupled mass after filling the measuring cylinder (mc+l) were measured to determine the density of the mixed-liquor (ql) in the SBRs following Eq. (2):

ql ¼

  mcþl  mc

ð2Þ

vl

According to the well-known Stokes equation, the viscosity (l) of the fluid media in a laminar flow system of the SBR was calculated as described by Li and Yuan (2002) in Eq. (3):

"

2

gð1  eÞðqc  ql Þd l¼ 18U act

# ð3Þ

e, Uact and do are the degree of porousness, actual settling velocity and diameter of a specific granular aggregate, respectively. 2.2. Hydrodynamic features For a porous bacterial aggregate, it is assumed that there are N identical cells. Each cell has a dry solid mass (md) when dewatered and a wet mass (mo) and density (qo) when the immobilized aggregates remained native. Theoretically, the porosity of an individualized granule of mass, mo with an impermeable Euclidean volume, Va and an enclosed spatial volume, vo are defined by the cumulative number of granular aggregates, N constituting the granulation system as described in Eq. (4):

  Nv o Porosity ðeÞ ¼ 1  Va

ð4Þ

Deducing a volume-based equation from Xiao et al. (2008) Nv o Va

d ¼ p6fm thus, Eq. (4) becomes q d3

"

o

e¼ 1

6fmd

# ð5Þ

pqo d3

f is a dimensionless ratio between the wet mass and dry mass of each granular aggregate; f = mc/md. The overall density difference between the bacterial aggregates and water is a function of the porosity and calculated as by Eq. (6):



qc ¼ ql þ

ðqo  ql Þ 1e

 ð6Þ

qc is the density of the bacterial cells; qo and ql are the densities of the granules and the mixed-liquor, respectively (Li and Yuan, 2002). The granules were observed as porous and highly fractal, a

278

H.I. Tijani et al. / Bioresource Technology 186 (2015) 276–285

structural feature that creates an intra-flow via the pores of the flocs influencing the viscous drag coefficient to attain faster settling.

Cd is the empirical drag coefficient. However, Cd is a function of the Reynolds number (Re) calculated as Eq. (11):

2.3. Settling experiments

Re ¼

The settling experiments were conducted on individual granules in a settling column following previous techniques described by Johnson et al. (1996), Li and Logan (1997), Li and Yuan (2002) and Zhang et al. (2004). The experimental column used for characterizing the settling rates of granules comprised of a glass burette with a valve. The column had a settling height of 650 mm to ensure that the actual settling velocity could be established, and an internal diameter of 13 mm to minimize the influence of viscous drag exerted by the column walls on aggregate settling as previously explained by Li and Yuan (2002). The system was assembled in single-column configuration and filled completely with double distilled water (qw = 0.997 g cm3 at 22 °C). For each run, a granular aggregate was placed lightly at the top of the settling column while the valves were closed and the duration of time for the granule to settle through the midpoint at 628 mm distance was recorded. The granule was recovered for subsequent analyses. Prior to settling, the floc size of these individual granules placed in a petri dish (95 mm) were determined. The fractal image of each aggregates was captured using a Digital Single Lens camera (GC100, Samsung, France), and processed with the aid of a computer-based image analysis program (ImageJ, Rasband, NIH). Granule size, do was calculated in terms of equivalent diameter from the projected surface area analyzed by ImageJ. The actual settling velocity of the aggregates was determined as the rate of settling motion of the fractal aggregates through the fluid system. A known amount of granules was put in motion to transverse from the surface of the fluid through a designated graduation towards the bottom of the settling column while the duration of motion was being monitored with the aid of a microtiming stopwatch. Thus, the actual velocity of granular motion (Uact), was calculated following Eq. (7):

U act ¼



 Dist: travelled by the granular aggregate Duration of viscous motion

ð7Þ

After the settling test, the retrieved granules were dried at 30 °C for 24 h on a pre-weighted polycarbonate membrane filter (0.4 lm), and their dry masses, md was determined using an electronic microbalance (AEM-5200, Shimadzu, Japan). To account for the influence of porosity, permeability and other hydrodynamic functions on the settling ability of granules, the terminal settling velocity of an individualized impermeable aggregate under gravity was predicted from the exerting force balance on the granules to generate the generalized Stokes’ law as in Eq. (8):

Us ¼

    fmd q g 1 l 3p qc ldo

ð8Þ

do is the diameter of the granule; qc and ql are the densities of the microbial aggregates and liquid, respectively; g is the gravitational constant; and Cd is the empirical drag coefficient. However, Cd is a function of the Reynolds number (Re) calculated following Eq. (9):

 Re ¼

ql U act d l o

 ð9Þ

The actual settling velocity (Uact), aggregate density (qc) and its surface area (A) are derivatives that describes the empirical drag coefficient (Cd) of a particle as Eq. (10):

" #   2fmd qc ql 1 Cd ¼  A ql qc U 2act



ql U act d l o



ð11Þ

2.4. Rheology of aerobic granules Rheology is the science describing the deformation of a body under the influence of stresses. The rheological feature of the granules is a significant means of describing the formation mechanism of an aggregate under the influence of enhanced stresses. The Herschel–Bulkley equation was used to define these functions (Seyssiecq et al., 2003) as in Eqs. (12) and (13)

s ¼ sY þ K cn c¼

ð12Þ

s l

ð13Þ

s, sY, c, l are the shear stress, yield stress, shear rate and fluid viscosity, respectively; K and n are rheological constants. Series of mathematical formulations have been described for calculating the area of fluid permeability of an aggregate (j); the correlation (d) as proposed by Happel provides a more realistic estimate of the intra-aggregate permeability for immobilized biological flocs (Li and Logan, 2001) as described in Eq. (14); "

2

d 3  4:5d þ 4:5d5  3d6 j¼ o 18 d3 ð3 þ 2d5 Þ

!# ð14Þ

pffiffiffiffiffiffiffiffiffiffiffi whereas d ¼ 3 1  e The porous structure of granule allows fluid substrate to flow through its interior to an extent that its internal permeation reduces the drag force exerted on the aggregate which results in a faster velocity than predicted for impermeable aggregates. The internal permeation (n) of these aggregates can thus be estimated from the j-relationship via the correlation (Zhang et al., 2004) as presented in Eq. (15);

do n ¼ pffiffiffiffi 2 j

ð15Þ

From the n value, the fluid collection efficiency of granule (g), which is defined as the ratio of the interior flow passing through the aggregate to the flow approaching it, can be calculated by (Xiao et al., 2008) as described in Eq. (16);

"

g¼

9ðn  tanhðnÞÞ

#

2n3 þ 3ðn  tanhðnÞÞ

ð16Þ

2.5. Calculations of fractal dimensional analysis The basis for calculating granules’ dimension is based on settling experiments and processed fractal images using ImageJ analyses. Several algorithms and flocculation models have been proposed to calculate the fractal dimension (Logan and Kilps, 1995; Jarvis et al., 2005; Li and Yang, 2007; Xiao et al., 2008; Gao and Li, 2011). Dimensional analysis of granules were determined for its second-order and third-order fractal dimensions. The twodimensional fractal dimension was estimated via the power-law function of ImageJ-fractal box counts. According to Johnson et al. (1996), the settling correlation (U) of the granules to their diameter sizes (l) was mathematically represented by a scaling function as described in Eq. (17):

ð10Þ U/l

Df 1

ð17Þ

279

H.I. Tijani et al. / Bioresource Technology 186 (2015) 276–285 0.8

Calculations of Df via this equation is only valid since D2 < 2 as depicted by the fractal box count; if D2 was not included in this scaling relationship, the fractal dimension of these aggregates will be underestimated. For an accurate fractal geometry relationship to be established, the aggregates should have the equal fractal dimensions and developed from the same primary particles under identical reaction processes (Jiang and Logan, 1991). This accounts for need to fit the empirical drag factor (b) of granules into the settling correlation. Thus, Eq. (14) was modified to incorporate D2 and b factors with a proportionality constant (v) that gives Eq. (18):

R² = 0.9755

0.5

1

0.3 0.2 0.1

In the experimental application for in-stored granules, the dimensionless drag factor (b) based on a finite number was calculated from the Cd  Re relationship based on Eq. (19):

0.0

Cd ¼

1.5

0.4

ð18Þ

aRb e

2

0.6

R² = 0.9913

1.8

2.3

2.8

3.3

Aggregate size (mm)

ð19Þ

3.1. Morphology of the aerobic granules The matured cultivated aerobic granules has been in-storage at 4 °C for 2 years with its sludge volume index (SVI = 31.3 mL gSS1) remaining relatively unaltered after storage; however, they had slight variations in the granular sizes’ (do) and the biomass concentration of cellular aggregates (qc) biomass and settling property. With the mean granular size decreasing from 3 mm to 2.77 mm, the biomass concentration of the cellular aggregates increased from 7600 mg/L to 7807 mg/L. This phenomena changes is probably due to the discharge of soluble organic material and cell hydrolysis as described by Tay et al. (2002). These fully-fledged aggregates were reactivated after washing three (3) times with distilled water to remove the fermentation products and the residual nutrient substances; then it was acclimatized to SBR operational cycle with synthetic wastewater components as the primary influent. The temperature of the reactivation system was kept at ambient temperature, and influent pH was adjusted to 7.0. The palm oil mill sludge attained successful treatment in RA after the inoculation of seeding granules. The in-stored granules employed as secondary inoculum were biologically active, with a mean aggregate size of 2768.7 ± 48.77 lm, whereas those dewatered were smaller, with a mean size of

B

______ 10 mm

|

0

η

к

1345.8 ± 32.57 lm as shown in Fig. 1. Upon granulation, the morphology of the palm oil mill sludge in the SBRs transformed completely into denser and fast-settling granules as compared to the seed sludge flocs. The differential microbial population, particularly the differing screens of bacteria strains, generated two forms of granular sludge with distinct color differences. It was observed that the dark granules had more compact structures of evenly shaped aggregates with faster settling properties than the light-colored granules. It is also significant to note that both granular morphologies seems to exercise a specific threshold of electromagnetic charges that tend to establish a magnetic field of attraction within the liquor. The resultant effect of repulsive charges between the surface charges of the aggregates and the effluents was deduced from the calculated Happel coefficient function (d = 0.667) which signifies that the immobilized aggregates establishes a well-defined adventive flow through the aggregate cortex to attain a significant level of fluid collection. This result reveals that the intra-particle charges, a function of its fluid collection efficiency of the granules (g = 0.27 ± 0.02) were comparable with those described for hydrogen-producing aggregates (g = 0.26 ± 0.09) and substantially higher than those for characterized for activated sludge aggregates and latex microspheres (Li and Yuan, 2002; Zhang et al., 2004). As illustrated in Fig. 2, the rate of substrate transport within the immobilized aggregates via internal permeation is relatively proportional to the rate of molecular substrate diffusion. Thus, the intra-particle charges within the aggregate enhances substrate mass transport into the granules by specific order of magnitude which increases the propensity for the synthesized EPS to form a

3. Results and discussion

|

3.8

Fig. 2. Fluid collection efficiency (g) and fluid permeable area (j).

In understanding of the self-similarity feature of fractal aggregates, the granules’ structural complexity as well as their fractal dimension were not subjected to variations under settling experiments or other external stimulus that may establish granular breakage; thus, a, b, and Df was characterized for these granules.

A

0.5

Area of fluid permeability (mm2)

0.7

Fluid collection efficiency

Df þD2  1 U ¼ vl b

2.5

|

______ 10 mm

|

Fig. 1. Morphology of the digitized granular bioparticles for immobilized granules; (A) – native granules, (B) – dewatered granules.

280

H.I. Tijani et al. / Bioresource Technology 186 (2015) 276–285

gel-matrix that clogs the interior pores of the granules and limits its permeability (Li et al., 2003; Zhang et al., 2004). Fig. 3 describes the influence of shearing and dewatering effects on the superficial morphology of the native granules via the surface plot images as depicted by ImageJ-particle analyzer. The impact of shear stress on the native aggregates under laminar shear rates (c) of 0 s1, 10 s1, 30 s1, 41.67 s1 as represented in Fig. 3(a–d) respectively indicates that the in-stored granules had a more compact structure with a defined superficial coat pattern which may be a result of the structural organization of cohabiting microbial aggregates, thus, influencing the stability of the granules. The visual observation in Fig. 3e describes the impact of shear stress on the dewatered aggregates under laminar shear rates (c) of 0 s1. It is obvious that dewatering enables the central anoxic core which is constricted in microbial activities to become apparently observable as the structural skeleton that aids the granules’ resiliency against complete disintegration. This implies that the water content of the granule’s anoxic core may be significant in evaluating the impact of metabolic heat on the microbial community. Failure to promote evaporation and enhance the high heat removal potential of evaporative cooling in the reactors can potentially decrease the bioavailability of substrates to the central anoxic core and incapacitate the anaerobic consortia. These visualized images depicts that native granules exhibit specific degree of mass transfer limitation of nutrients (substrates/metabolites) to its anoxic core. While the superficial layers seemed to be differentially permeable and easily sheared under laminar conditions. Based on the observed residual effects of dewatering and shearing at c = 41.67 s1, EPS molecules may constitute the primary buffering layer for the microbial cells against shock loadings which could also serve as a viable source of carbon and energy during famine periods. As a result, dense consortia of microbes immobilized by EPS may assume a defined architectural organization that contributes to the hydrodynamic features of the granules. It is, however, expected that the granules’ surface becomes smoother when the cohabiting biomass concentration decreases due to oxygen/substrate deficiency (de Kreuk et al., 2005). According to Wang et al. (2005), EPS are metabolic products composed of a variety of macromolecules such as exopolysaccharides, exoproteins, DNA, humic acid and uronic acids which are extracellularly secreted to promote microbial biofilm formation.

3.2. Rheological characteristics of granules The non-Newtonian behavior of granules studied via rheology describes their tendencies to structural deformation under the influence of enhanced stress and the nature of their deformation were linked to the aggregates’ hydrodynamic properties and mass oxygen transfer ability. The viscosity of the dispersion liquor, the granular sludge concentration, size and shape, and the cluster– cluster interactions were the predominant factors influenced by the increasing sludge viscosity during the shearing test. The rheological characteristics of the in-stored granules are described in Table 1. The non-linear rheological correlation of the responsive shear stress to the exerted shear rates for the granules are described in Fig. 4. Enhanced granular abrasions established during extensive shearing analysis transformed the particle size distribution (PSD) of the sludge aggregates as previously described in Fig. 3, such that larger microbial aggregates of the sludge mass were sheared by vigorous fluid turbulence to smaller aggregates. This denotes that the turbulent breakage applied attained a threshold capable of destabilizing the sludge PSD due to the complex structure and deformation effects exhibited by the immobilized microbial aggregates.

The Herchel–Bulkley equation used to simulate the shearing behavior of these aggregates generated a high correlation coefficient in accordance with observed results. This suggests that the granular aggregates assume a non-Newtonian motion under laminar flow systems and as well, exhibit defined yield stress. As delineated, the apparent fluid solids released in the suspension exponentially declined with successive increase in the stress rate (c) which implies that the fluid suspension of the granular sludge was shear-thinning. Seyssiecq et al. (2003) described the rheology of classified activated sludge suspensions as non-Newtonian fluids with a non-linear relationship of shear rate to shear stress. The yield stress of the sludge flocs is proportional to its physicochemical features, such as the fluid solid suspension, entrapped water content and surface charge. Thus, the rheological profile of a sludge can influence its aeration mix, settling and mass transfer abilities. Thus, the bioaugmentation of in-stored aggregates in granulation systems develops an aerobic granular sludge with denser structures and faster settling abilities than its sludge flocs. The settling values are significantly higher than the settling velocities of aerobic granules reported by Liu and Tay (2004) and Xiao et al. (2008). The settling behavior of the cultivated granules was invariantly expressed with a faster settling rate of 3.14 ± 0.058 cm s1 than stokes’ predictions for its homogenously permeable and impermeable Euclidean replica with settling velocities of 0.12 ± 0.002 cm s1 and 0.42 ± 0.010 cm s1, respectively. This implies that fully-fledged granules settled 86% faster than the equally sized Euclidean aggregates. 3.3. Fractal differentiation analysis of granules The fractal nature of microbial granular aggregates is a complex function used to define the structural organization of the consortia of microorganisms and the influence of varying reactor conditions on the immobilized aggregates. For sludge aggregates, the presence of external forces acting on an aggregate may assist aggregate reassembling in which particles and clusters were interlinked by relatively weak Van der Waals interactions with few polymeric bridges. These connections can be disrupted by thermal fluctuations, shear stress, or through collision momentum that results in particle re-arrangement and matrix restructuring of the aggregates. In order to standardize measurement for sludge bio-flocculation of microbial aggregates, an additional application of fractal dimension may be as a tool to understand the variations in sludge morphology and aggregate properties under different reactor conditions as it may be possible to use the fractal dimension characteristics of different aggregates to distinctively describe the influence of particle nature, shear rate, biofilm prototype and bioflocculation mechanism on their granulation processes. 3.3.1. Mass-based particle size distribution With the advent of fractal geometry, the mathematical characterization of the flocculation processes for particle aggregates has been vastly analyzed (Li and Ganczarczyk, 1989; Jiang and Logan, 1991; Logan and Kilps, 1995; Jackson, 1998; Zhang and Li, 2003). The fact that granules having much larger sizes than other forms of coalesced spherical particles is the primary factor that enhances the flocculation between its bioparticles as well as the breakage of its immobilized clusters. Although mass-based PSD analysis is a geometric feature that is conserved within the particle aggregation and cluster breakage of the granules, the actual floc sizes distinctly describe its particles’ property as fractals. Fig. 5 illustrates the mass-based particle size distribution of native granules and its dewatered aggregates. Approximately 230 granules were tallied to generate the

281

H.I. Tijani et al. / Bioresource Technology 186 (2015) 276–285

A

B

C

D

E

Fig. 3. Superficial morphology of granules via surface plot. Native aggregates, (A) – c = 0 s1, (B) – c = 10 s1, (C) – c = 30 s1, (D) – c = 41.67 s1; dewatered aggregates, (E) – c = 0.

282

H.I. Tijani et al. / Bioresource Technology 186 (2015) 276–285

Table 1 The rheological characteristics of the in-stored granules. Rheological parameters of in-stored aggregate 2768.7 lm 8.48 mg 3.14 cm s1 0.68 0.236 Pa s 7.807 kg/m3 0.003 cm2 5.034 0.270 0.030 0.667

Diameter size (do) Wet mass (fmd) Settling velocity (Uact) Porosity (e) Fluid viscosity (l) Density (qo) Area of fluid permeability (j) Permeability coefficient (n) Fluid collection coefficient (g) Empirical drag coefficient (Cd) Happel coefficient (d)

mass-based PSD for the in-stored aggregates. In contrast to the observations of Chaignon et al. (2002) and Li and Leung (2005) which recorded much smaller aggregates than larger aggregates, the mass-based PSD of granules featured a relatively invariant aggregates sizes of 8.48 ± 0.093 mg with a mean diameter size of 2.77 ± 0.049 mm. Mathematical analysis of the transformation plots reveals a relative series of linear correlation for mass-based differentiation. It is significant to note that no single proportionality constant can be functional to describe the entire size distributions due to the non-rectilinear behavior of the larger aggregates. Even though the dimensional analysis and simulation models of Li et al. (2003) and Zhang and Li (2003) generated a curvilinear PSD correlation signifying increased propensity for particle breakage via aeration and shearing rates; yet, no slope could be extrapolated to define the entire PSD of aerobic granular sludge and its fractal dimension (Df) was solely dependent on the power law PSD function. 3.3.2. Fractal dimension The fractal dimension of granules were described from its second-order (D2) and third-order (D3) dimensional analysis. The second order dimensional analysis of these aggregates was determined via the power-law function of ImageJ fractal box counting technique to define their two-dimensional fractal dimension as D2 = 1.795 for native clusters and D2 = 1.099 for dewatered clusters. Lower fractal feature has been reported for the bacterial bioflocs generated in a laboratory SBR for synthetic wastewater treatment (Li and Yuan, 2002). The log–log transformation plot of the fractal counts to their aggregate sizes for native and dewatered clusters as predicted by ImageJ-particle analysis is illustrated in Fig. 6. The aerobic granules characterized in this study were much denser than other bio-aggregates, as proven by their lower

porosities. It has been reported that activated granular sludge constituting larger-sized flocs have a porosity greater than 0.95 (Li and Yuan, 2002; Li et al., 2003). In contrast, the individualized granules were analyzed to exhibit porosities in the range, 0.29 6 e 6 0.88; a porosity value which are comparable to that of biohydrogen producing granules (Zhang et al., 2004). Matured granules had fractal dimensions within the range projected for biological aggregates (1.4 6 Df 6 2.8) produced in wastewater treatment processes as defined by Xiao et al. (2008), although the different fractal characterization approaches may undermine the significance of a proportional correlation of the Df values described under varying fractal simulation models. As previously derived, the discrete Df value in correlation to D3 is the fractal dimension of the granular microbial aggregates. With the aid of a pre-derived settling-fractal simulations, the three-dimension fractal dimension for granules was defined by the power rank correlation from the transformation plot. The dimensionless drag factor (b) determined from the Cd  Re calculations as 2.131 was inserted in the U  lD simulation. Recalling the pre-derived settling-fractal simulation which incorporated D2 and b factors with a proportionality constant (v) in Eq. (15) as:

U ¼ vl

b

1

The three-dimensional fractal dimension of granules (D3) was calculated as 2.46 via the extrapolated power plot equation. These data calculations are described in Table 2. Contrasting the fractal dimensioning analysis for granules and its Euclidean replica, it can be relatively deduced that the immobilized granules were porous, differentially permeable fractals which possess a defined particle-sized distribution behavior of their fractal dimensions and aggregate sizes. These fully-fledged granules used as secondary inoculum exhibited a structural feature that shows a rectilinear plot relationship between particle sizes and fractal dimension. The granules were relatively symmetrical over a wide array of granular sizes, and a clear increase in surface regularity was prominent; this granular feature provides creative insights for describing its flocculation mechanism. The calculated three-dimensional fractal dimensions of granules is in conformity with the fractal dimensioning simulations of Johnson et al. (1996), providing the basis for describing the proportional correlation between the fractal dimension and settling factor of individualized aggregates. The fractal dimensioning analysis and the hydrodynamic characteristics seems to depict that microbial aggregates of activated sludge were structured via cluster–cluster aggregation rather than particle-cluster connections. This implies

12

R² = 0.9845 R² = 0.9906

10

Shear Stress (mPa)

Df þD2 

R² = 0.9939

γ = 3.33 /s

R² = 0.9967

8

γ = 10.00 /s γ = 16.67 /s

R² = 0.9955

6

γ = 23.33 /s γ = 30.00 /s

R² = 0.9947

γ = 36.67 /s

4

0

γ = 41.67 /s

R² = 0.9811

2

0

50

100

150

200

250

300

Shear Time (s) Fig. 4. Rheological correlation of the responsive shear stress at apparent shear rates for the granular aggregates under laminar flow.

283

H.I. Tijani et al. / Bioresource Technology 186 (2015) 276–285 10000 R² = 0.5299

Aggregate mass (μg)

8000

6000

4000 R² = 0.0159

2000

0

1.8

2.3

2.8

3.3

3.8

f

Aggregate size (mm) Fig. 5. Mass-based particle size distribution of granules. [ – dewatered clusters.

] – native clusters, [

]

18 16 14

R² = 1 D2 = 1.795 a

Log (Count)

12 10

D2 = 1.099 R² = 0.9986 a

8 6 4 2 0

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Log (Size) Fig. 6. Two-dimensional fractal dimension (D2) of granules. (d) – native clusters, (s) – dewatered clusters.

that granular aggregates formed in SBRs were extensively restructured to assume higher fractal dimensions than those of the unprocessed sludge. The hydrodynamic characteristics as well as substrate transfer mechanisms of aerobic granular sludge have been relatively explored in recent times. Chiu et al. (2007a,b) described the heterogeneous nature of the structural matrix and the differentiation of microbial networks into distinct zones within the granules was understood to be associated with the secretion of bacteria EPS into the extracellular matrix. This biofilm feature is in conformity with the current findings for the in-stored granules to synthesize higher concentrates of gel-matrix EPS which stimulated the granulation system to establish efficient treatments than the activated sludge flocs in RC, noting that the EPS embedding the aggregates

significantly limits their permeability and porosity to achieve faster settling. It is well known that the hydrodynamic behavior of activated granular sludge flows plays key importance in the optimization of the aerobic sludge granulation process parameters. These granules with an aggregate size of 2768.7 ± 48.77 lm were relatively influenced by enhanced shearing and resilient to complete disintegration at a shear rate (c) of 41.67 s1 thus, granular abrasions established during the strength analysis were shear-thinning. The granules’ layers had different permeability and the morphological effect of dewatering depicts that EPS forms the primary buffering layer for the microbial cells. These granules had a unique hydrodynamic feature with a faster settling rate of 3.14 ± 0.058 cm s1 than stokes’ calculations for its homogenous permeable and impermeable Euclidean replica. With a porosity (e) of 0.68 ± 0.014, granules experience a decreased fluid influx through its clusters’ interior than other homogeneously dispersed permeable aggregates featuring the aggregates as a more compactly packed clusters. The SBR processing conditions which exert laminar effects, by mixing and aeration establishes particle breakage and restructuring to relatively promote structural organization of the granular matrix and increase its fractal dimension. This is feature is more expressed when increasing folds of EPS are synthesized to improve close-proximity contacts among bacterial cells which results in a less porous matrix. Li et al. (2003) observed that the increasing stability contacts between microbial cells during matrix restructuring is the primary factor in bio-flocculation. Since the settling rates of these fractal aggregates were higher than those predicted for permeable and impermeable Euclidean aggregates, it can be deduced that the configuration structures of granules does not constitute homogeneously dispersed particles; a factor that characterize them as being differentially permeable. This implies that there exist a differentiation of non-homogeneous particle distribution within the fractal aggregates which is due to the flocculation of small densely packed clusters into less dense larger aggregates of immobilized biofilm. The analytic technique employed for the determination of Df via settling analysis is an improvement over the box-counting method and used by Wu and Lee (1998) to describe fractal aggregates, although, it requires the power law functions for determining the second-order fractal dimension. However, this approach also proffers superiority over the particle concentration technique (PCT) and the mass-size relationship (Li and Yuan, 2002; Li et al., 2003) as it does not necessitate the use of a Coulter particle counter which determine the solid volume of aggregates for PSD analysis, noting that granules were relatively porous and frail. As proposed by Gorczyca and Ganczarczyk (2001), the direct observation of thin microtome slices of an aggregate immobilized in paraffin is an alternatively potent method for fractal analysis but the applicability of this sectioning approach has remained unexplored for characterizing highly friable aggregates due to their compressibility factor. As a matter of fact, the technique for calculating Df values

Table 2 Plot characteristic of the U  lD simulation. Equation parameters

Granules

Power equation Proportionality constant (v) b-factor D2 Calculations: 3rd order fractal dimension

Error analysis

Euclidean replica 0:997

D3 = Df

1:3

U ¼ 1:16 l 1.16 2.131 1.795

Df þ1:795  1 ¼ 0:997 2:131

U ¼ 0:11 l 0.11 2.131 1.795

Df þ1:795  1 ¼ 1:3 2:131

Df ¼ ½2:131ð0:997 þ 1Þ  1:795 Df ¼ 2:46 0.10

Df ¼ ½2:131ð1:3 þ 1Þ  1:795 Df ¼ 3:10

284

H.I. Tijani et al. / Bioresource Technology 186 (2015) 276–285

of different aggregates cannot be relatively generalized due to the differences in their structural organization and flocculation mechanisms. Thus, the determination of Df via settling simulations is theoretically accurate, easily accessible and consistently precise for granules than other alternative techniques. Analysis of fractal dimensions and settling measurements obtained for granules cultivated under varying SBR engineered systems lacked universality as it has not incorporated a shape factor into Stokes’ law. Thus, the decrease in the calculated drag coefficient used to describe the fractal nature of granules is limited by its shape factor. This depicts that its fractal calculations may not be applicable to all categories of activated sludge flocs in engineered reactors due to differences in their structural organization of polymeric secretions, filaments and other biomaterials capable of clogging pores and influencing SBR flow conditions. From the observations of the mechanistic approach of granulation system, it can be deduced that the fractal dimensional analysis can be employed as an evaluation tool to characterize immobilized microbial aggregates to provide quantitative measurements in describing its structural integrity and complexity. At present, the theoretical simulation models and experimental designs used in this study to describe aerobic granulation systems have been focused on exploiting the colloidal flocculation properties of the in-stored aggregates; however, the bioreactor configurations, shear flocculation, laminar flow processes and enhanced turbulent effects have been noted to play dominant roles in the treatment capabilities of the granular sludge. It is significant to note that the analysis of granules’ settling behavior via differential sedimentation was vital to describe the influence of mass transfer rates on the immobilized biofilm structure. The particle quantification and structural differentiation of the surface morphology of granules using the fractal dimension concept revealed a linear trend between its fractal dimension and mass-based particle size distribution. This proves the hypothesis that these methods and mathematical derivations can be standardized to monitor aggregates’ formation in granulation systems. Thus, in-stored aggregates cultivated under effective operating protocols would favorably suit the pilot-scale implementation of granulation treatment for discharges. Future research should focus on characterizing the fractal dimensions for granules cultivated under different physicochemical environments and define the propensities of the individual microbial consortia to form a stable immobilized biofilm structure from the floc formations in the SBRs. 4. Conclusion Enhanced shearing effects influences the aggregates’ skeletons and the microbial screens of the aerobic granules which agrees with earlier reports that defined specific clusters of protozoans and metazoans enmeshed within the sludge. Based on their resilience to shear stress, the dominant microbial culture were able to synthesize substantial amounts of EPS with excellent adsorption for the stabilization of the aerobic granular sludge. Fractal dimensioning analysis and hydrodynamic characteristics depicts that microbial aggregates were structured via cluster–cluster aggregation rather than particle-cluster connections resulting in extensive restructuring to assume higher fractal dimensions than those of the unprocessed sludge. Acknowledgements We wish to thank Universiti Teknologi Malaysia and Ministry of Education Malaysia for funding this research under Grants No. 08H00 and FRGS 4F198.

References Abdullah, N., Ujang, Z., Yahya, A., 2011. Aerobic granular sludge formation for high strength agro-based wastewater treatment. Bioresour. Technol. 102, 6778– 6781. Abdullah, N., Yuzir, A., Curtis, T.P., Yahya, A., Ujang, Z., 2013. Characterization of aerobic granular sludge treating high strength agro-based wastewater at different volumetric loadings. Bioresour. Technol. 127, 181–187. Chaignon, V., Lartiges, B.S., Samrani, A.E., Mustin, C., 2002. Evolution of size distribution and transfer of mineral particles between flocs in activated sludges: an insight into floc exchange dynamics. Water Res. 36, 676–684. Chiu, Z.C., Chen, M.Y., Lee, D.J., Wang, C.H., Lai, J.Y., 2007a. Oxygen diffusion and consumption in active aerobic granules of heterogeneous structure. Appl. Microbiol. Biotechnol. 75, 685. Chiu, Z.C., Chen, M.Y., Lee, D.J., Wang, C.H., Lai, J.Y., 2007b. Oxygen diffusion in active layer of aerobic granule with step change in surrounding oxygen levels. Water Res. 41, 884. de Kreuk, M.K., Pronk, M., van Loosdrecht, M.C.M., 2005. Formation of aerobic granules and conversion processes in an aerobic granular sludge reactor at moderate and low temperatures. Wat. Res. 39, 4476–4484. Gao, Z.N., Li, Y.M. (2011). Impact of Size Fractal Distribution on Flow Behavior of Granular Flow System. Computer Distributed Control and Intelligent Environmental Monitoring (CDCIEM). 2011 International Conference, 1545– 1548. Gorczyca, B., Ganczarczyk, J.J., 2001. Fractal analysis of pore distributions in alum coagulation and activated sludge flocs. Water Qual. Res. J. 36, 687–700. Jackson, G.A., 1998. Using fractal scaling and two-dimensional particle size spectra to calculate coagulation rates for heterogeneous systems. J. Colloid Interface Sci. 20, 202. Jarvis, P., Jefferson, B., Gregory, J., Parsons, S.A., 2005. A review of floc strength and breakage. Water Res. 39, 3121–3137. Jiang, Q., Logan, B.E., 1991. Fractal dimension of aggregates determined from steady-state size distributions. Environ. Sci. Technol. 25, 2031–2038. Johnson, C.P., Li, X.Y., Logan, B.E., 1996. Settling velocities of fractal aggregates. Environ. Sci. Technol. 30, 1911–1918. Li, D.H., Ganczarczyk, J.J., 1989. Fractal geometry of particle aggregates generated in water and wastewater treatment processes. Environ. Sci. Technol. 25, 1385– 1389. Li, X.Y., Leung, R.P.C., 2005. Determination of the fractal dimension of microbial flocs from the change in their size distribution after breakage. Environ. Sci. Technol. 39 (8), 2731–2735. Li, X.Y., Logan, B.E., 1997. Collision frequencies between fractal aggregates and small particles by differential sedimentation. Environ. Sci. Technol. 31, 1229– 1236. Li, X.Y., Logan, B.E., 2001. Permeability of fractal aggregates. Wat. Res. 35, 3373– 3380. Li, X.Y., Yang, S.F., 2007. Influence of loosely bound extracellular polymeric substances (EPS) on the flocculation, sedimentation and dewaterability of activated sludge. Water Res. 41 (5), 1022–1030. Li, X.Y., Yuan, Y., 2002. Settling velocities and permeabilities of microbial aggregates. Water Res. 36, 3110–3120. Li, X.Y., Yuan, Y., Wang, H.W., 2003. The hydrodynamics of biological aggregates of different sludge ages: an insight into the mass transport mechanisms of bioaggregates. Environ. Sci. Technol. 37, 292–299. Lin, Y.M., Liu, Y., Tay, J.H., 2003. Development and characteristics of phosphorousaccumulating granules in sequencing batch reactor. Appl. Microbiol. Biotechnol. 62, 430–435. Liu, Y., Tay, J.H., 2004. State of the art of biogranulation technology for wastewater treatment. Biotechnol. Adv. 22 (7), 533–563. Logan, B.E., Kilps, J.R., 1995. Fractal dimensions of aggregates formed in different fluid mechanical environments. Water Res. 29, 443–453. Logan, B.E., Wilkinson, D.B., 1991. Fractal dimensions and porosities of Zoogloea ramigera and Saccharomyces cerevisae aggregates. Biotechnol. Bioeng. 38 (4), 389–396. Maggi, F., 2007. Variable fractal dimension: a major control for floc structure and flocculation kinematics of suspended cohesive sediment. J. Geophys. Res. 112, C07012. http://dx.doi.org/10.1029/2006JC003951. Mohammed, J.N., Abubakar, B.M., Yusuf, H., Sulaiman, M., Saidu, H., Idris, A., Tijani, H.I., 2013. Bacterial biofilm: a major challenge of catheterization. J. Microbiol. Res. 3 (6), 213–223. http://dx.doi.org/10.5923/j.microbiology.20130306.04. Seyssiecq, I., Ferrasse, J.H., Roche, N., 2003. State-of-the-art: rheological characterization of wastewater treatment sludge. Biochem. Eng. J. 16, 41–56. Tay, S.T.L., Ivanov, V., Yi, S., Zhuang, W.Q., Tay, J.H., 2002. Presence of anaerobic Bacteroides in aerobically grown microbial granules. Microb. Ecol. 44 (3), 278– 285. Tay, J.H., Pan, S.H., He, Y., Tay, S.T.L., 2004. Effect of organic loading rate on aerobic granulation II – characteristics of aerobic granules. J. Environ. Eng. 130 (10), 1102–1109. Wang, Z.W., Liu, Y., Tay, J.H., 2005. Distribution of EPS and cell surface hydrophobicity in aerobic granules. Appl. Microbiol. Biotechnol. 69, 469–473. Wu, R.M., Lee, D.J., 1998. Hydrodynamic drag force exerted on a moving floc and its implication to free-settling tests. Water Res. 32, 760–768.

H.I. Tijani et al. / Bioresource Technology 186 (2015) 276–285 Xiao, F., Yang, S.F., Li, X.Y., 2008. Physical and hydrodynamic properties of aerobic granules produced in sequencing batch reactors. Sep. Purif. Technol. 63 (3), 634–641. Yang, Z., Peng, X.F., Lee, D.J., Ay, S., 2007. Advective flow in spherical floc. J. Colloid Interface Sci. 308, 451–459.

285

Zhang, J.J., Li, X.Y., 2003. Modeling particle-size distribution dynamics in a flocculation system. AIChE J. 49, 1870–1882. Zhang, J.J., Li, X.Y., Oh, S.E., Logan, B.E., 2004. Physical and hydrodynamic properties of flocs produced during biological hydrogen production. Biotechnol. Bioeng. 88, 854–860.