Journal of Hazardous Materials 171 (2009) 948–953

Contents lists available at ScienceDirect

Journal of Hazardous Materials

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

Enhancing biodegradation of wastewater by microbial consortia

with fractional factorial design
Yuancai Chen a , Che-Jen Lin b,c,∗ , Gavin Jones d , Shiyu Fu a , Huaiyu Zhan a
a
State Key Laboratory of Pulp and Paper Engineering, South China University of Technology, Guangzhou, 510640, Guangdong, China
b
Department of Civil Engineering, Lamar University, Beaumont, TX 77710-0024, USA
c
School of Environmental Science & Engineering, South China University of Technology, Guangzhou, 510006, Guangdong, China
d
Texas Research Institute for Environmental Studies, Sam Houston State University, Huntsville, TX 77341-2506, USA

a r t i c l e i n f o a b s t r a c t

Article history: Batch experiments were conducted on the degradation of synthetic and municipal wastewater by six
Received 25 January 2009 different strains, i.e., Agrobacterium sp., Bacillus sp., Enterobacter cloacae, Gordonia, Pseudomonas stutzeri,
Received in revised form 18 June 2009 Pseudomonas putida. By applying a fractional factorial design (FFD) of experiments, the influence of each
Accepted 18 June 2009
strain and their interactions were quantified. An empirical model predicting the treatment efficiency was
Available online 25 June 2009
built based on the results of the FFD experiments with an R2 value of 99.39%. For single strain, Enterobacter
cloacae, Gordonia and P. putida (p = 0.008, 0.009 and 0.023, respectively) showed significant enhancement
Keywords:
on organic removal in the synthetic wastewater. Positive interaction from Enterobacter cloacae, Gordonia
Bioaugmentation
Fractional factorial design
(p = 0.046) was found, indicating that syntrophic interaction existed, and their coexistence can improve
Microbial consortia total organic carbon (TOC) degradation. Verification experiments were performed to evaluate the effect of
Municipal wastewater bioaugmentation by introducing three selected strains into an activated sludge reactor for treating munic-
ipal wastewater. The removal efficiency of TOC with the bioaugmentation was increased from 67–72% to
80–84% at an influent TOC concentration of 200 mg/L. The results derived from this study indicate that the
FFD is a useful screening tool for optimizing the microbial community to enhance treatment efficiency.
© 2009 Elsevier B.V. All rights reserved.

1. Introduction A number of researches have reported the characteristics of

Municipal wastewater is typically treated by biological pro-
cesses. The performance of a biological system is often enhanced
through bioaugmentation of one or more species of microor-
ganisms. One factor hampering the bioaugmentation is the
disappearance of introduced strains after along period of opera-
tion. Bioaugmentation of a group of microorganisms may improve
reliability compared to bioaugmentation using a single strain [1].
Hisashi et al. [2] established a specific consortium using immo-
bilized photosynthetic bacteria, Rhodobacter sphaeroides S, Rb.
sphaeroides NR-3 and Rhodopseudomonas palustris, fixed on ceramic
media for wastewater treatment. They found that mixture of
the three bacteria was more effective for the removal of COD,
nitrate, phosphate and odor compared to using a single strain. It
is important for the introduced strains to establish stable microbial
communities with the indigenous bacterial consortium for treat-
ment improvement [3,4].
∗ Corresponding author at: Department of Civil Engineering, Lamar University,
Beaumont, TX 77710-0024, USA. Tel.: +1 409 880 8761.
E-mail address: Jerry.Lin@lamar.edu (C.-J. Lin).
microbial community during biodegradation of organic pollutants
[5,6]. Some used rRNA/rDNA molecular techniques to determine
the diversity of microbial populations during the degradation pro-
cess [8–10]. Although these methods are accurate, there are some
drawbacks such as long runtime, intensive instrumental effort and
sometimes neglecting the interplays among the microbial species.
This makes the molecular techniques unsuitable or infeasible in
some applications. In fact, it is difficult to evaluate the role of indi-
vidual microorganisms and their interactions during treatment, as
the survival of inoculants is affected by multiple factors including
phenotypic characteristics of the selected strains, complex micro-
bial interactions and habitat environments [1,7].
In contrast, statistical experimental design can analyze the
interactions among multiple microorganisms, and optimize the
composition of inoculum to improve treatment efficiency. Such
experimental design techniques have been widely used in various
fields such as in biochemistry for optimizing culture medium and
conditions [11,12], and in a microextraction for determining ben-
zene, toluene, ethylbenzene and xylenesmaterial [13]. However, to
our knowledge, there have not been reports on using the statistical
methods to investigate the effect of a mixed-culture inoculum on
the performance of wastewater treatment.

0304-3894/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jhazmat.2009.06.100
 Y. Chen et al. / Journal of Hazardous Materials 171 (2009) 948–953
In this study, six strains, i.e., Agrobacterium sp., Bacillus sp.,
Enterobacter cloacae, Gordonia, Pseudomonas stutzeri, Pseudomonas
putida, were selected to understand their roles in organic removal
and how they interact with each other. P. putida strain is frequently
used for wastewater treatment and has the ability of degrading
aromatic hydrocarbons such as toluene and phenol, which are com-
mon municipal contaminants [15]. Bacillus subtilis is a common soil
bacterium that can function as a secondary degrader of aromatic
hydrocarbons [15]. E. cloacae is a gram-negative, rod-shaped, fac-
ultative bacterium capable of degrading fatty acids [16]. Gordonia
is a petroleum degrading strain, it can utilize n-hexadecane, ben-
zene, naphthaline, anthracene, phenanthrene and malvasylvestris
as a carbon and an energy source [17]. Although the interactions
among different strains might be expected in a mixed-culture treat-
ment system, few studies investigate the combined effects of the
selected six species on treatment performance.
Municipal wastewater primarily contains carbonaceous pollu-
tants. Total organic carbon (TOC) is an ideal surrogate to reflect the
degree of organic pollution in wastewater, and serves as one of the
most important parameters for characterizing wastewater. The aim
of this work is to apply a fractional factorial design (FFD) of experi-
ments to investigate the interplays of six bacterial strains in treating
wastewater, and to construct a mixed-culture inoculum consisting
of specific microorganisms for enhanced wastewater treatment. We
expect our experiments will improve the selection of microorgan-
isms for bioaugmentaion in wastewater treatment.
2. Materials and methods
2.1. Microorganisms
Six strains of bacteria, i.e., Agrobacterium sp. (R1 ), Bacillus sp.
(R2 ), E. cloacae (R3 ), Gordonia (R4 ), P. stutzeri (R5 ), and P. putida (R6 )
were used to prepare the inoculum for degrading synthetic and
municipal wastewater. These strains were isolated from municipal
and industrial wastewater treatments [14].
2.2. Nutrient and feed medium
The nutrient medium contained 3 g/L of beef extract, 5 g/L of
peptone, and mineral salt medium (MSM) at pH 7. The composi-
tion of MSM (g/L) was: KH2 PO4 (0.42), K2 HPO4 (0.375), (NH4 )2 SO4
(0.244), NaCl (0.015), CaCl2 ·2H2 O (0.015), MgSO4 ·7H2 O (0.05) and
FeCl3 ·6H2 O (0.054). A phosphate buffer solution (PBS) (pH 7) was
prepared by dissolving NaCl (8 g/L), KCl (0.2 g/L), K2 HPO4 (1.15 g/L),
and KH2 PO4 (0.2 g/L) in deionized water (Millipore, Milli-Q). The
PBS was used for diluting the cell concentration in solutions.
To represent the organic substrates and nutrient contents of
municipal wastewater realistically, a stock solution was prepared
with the following composition (g/L): glucose (7.5), NH4 Cl (1.5),
K2 HPO4 (0.2), KH2 PO4 (0.2), MgSO4 ·7H2 O (0.25), CaCl2 ·2H2 O (0.3).
The solution yielded a total COD of 8 g/L, a TN (total nitrogen) of
0.4 g/L, and a total P of 0.08 g/L. The solution was kept at 4 ◦ C in a
refrigerator, then diluted to a desirable concentration (ca. 135 mg/L
TOC) and used as the synthetic wastewater in the single strain and
FFD experiments. To each liter of wastewater, 0.5 mL of a trace ele-
ment solution was added with the following composition (mg/L):
H3 BO3 (50), ZnCl2 (50), CuCl2 (30), MnSO4 ·H2 O (500), CoCl2 ·H2 O
(50), (NH4 )6 MoO24 ·H2 O (50). The pH of the synthetic wastewater
was controlled at 7.0 ± 0.1 by the addition of NaHCO3 . The water
used for dilution was from a water purification system (Millipore,
Milli-Q). Prior to use, all the media and solutions were sterilized at
121 ◦ C for 15 min.
949
2.3. Bacteria cultivation and biodegradation experiments
To recover the activity of the stock culture, one loop of each
of the six bacteria from the culture-contained agar was separately
transferred to 20 mL of the nutrient medium in a glass flask. Each
bacterium was activated at 30 ◦ C. These activated cells in the late
exponential phase were harvested as inocula, respectively. The
cells collected after centrifugation (6000 rpm for 10 min) were re-
suspended in the PBS and then centrifuged again. After cleaning,
for biodegradation experiments using a single bacterium (Section
3.1), the activated cells were individually inoculated into the syn-
thetic wastewater (120 mL) in Erlenmeyer flasks (250 mL) to give an
initial optical density at 600 nm (OD600 ) of 0.12–0.13. For the FFD
experiments (Section 3.2), six inocula were separately prepared by
inoculating the six activated strains into the synthetic wastewater
to give an initial OD600 of 0.20 ± 0.01. Then each 20 mL inoculum
containing 0 or 0.2 OD600 of bacteria (according to the fractional
experimental design detailed in Section 3.2) was added aseptically
to Erlenmeyer flasks (250 mL) yielding a final volume of 120 mL.
After inoculation, the Erlenmeyer flasks were capped with cotton
plugs and placed in a shaker controlled at 100 rpm and 30 ◦ C. The
pH of the mixture was adjusted to 7 by adding NaHCO3 . The pH
remained at 7.0 during the course of experiments. Samples were
withdrawn every hour, and the OD600 of cells and TOC were mea-
sured as described below. Each experiment was stopped when there
was no further increase of OD600 (stationary phase).
2.4. Determination of bacterial growth
The dry cell weight of bacteria was determined by centrifuga-
tion and drying. The cells in the cultures (40 mL) with different
cell concentrations were harvested by centrifugation (6000 rpm)
at 4 ◦ C for 10 min several times and washed with deionized water.
The cells were then dried at 80 ◦ C over 24 h until a constant weight
was obtained.
The concentration of cells in the sample, X, was analyzed by
measuring OD600 using an UV/vis spectrophotometer (Pharmaspec
UV-1700, SHIMADZU) with the synthetic wastewater as the blank.
In our earlier experiments, we found that 0.1 OD600 of the samples
corresponds to volatile suspended solids (VSS) of about 0.039 g/L.
In the single bacterium experiments (Section 3.1), X was estimated
by measuring OD600 then converted into VSS. The samples exceed-
ing 0.8 OD were diluted accordingly using the culture medium so
that the Beer–Lambert law applies.
2.5. TOC determination
TOC was analyzed on filtered samples (0.45 ␮m Magna nylon
membrane filters, SELECTRON, USA) with a Model 100 Bench-
top TOC Analyzer (Thermo Electron Corporation, USA) using the
combustion/non-dispersive infrared gas analytical method. Tripli-
cate analyzes were performed on all assays.
2.6. Kinetic analysis
The specific growth rate of cell,  (h−1 ), was determined
according to first-order growth kinetics with respect to biomass
concentration at a small initial cell concentration and a large initial
substrate concentration (S0 ) in a batch reactor. The value of  can
be calculated as:
 1   dX  X
= = d ln (1)
X dt dt
where X is the biomass concentration (mg VSS/L).  was deter-
mined using the data at the exponential phase of growth curve.
950 Y. Chen et al. / Journal of Hazardous Materials 171 (2009) 948–953

The specific substrate utilization rate, q (g TOC/(d–gVSS)), can be

expressed as:
 dS   1   1   dX   dS   d ln X   1  
q= = = = (2)
dt X X dt dX dt Y Y
where Y is the yield defined as the biomass produced per unit mass
of substrate consumed:
X − X0
Y= (3)
S0 − S
where X0 and S0 represent the initial biomass and substrate
concentrations, respectively. S is the residual TOC concentration
(mg TOC/L) in wastewater.

2.7. Statistical methods

A fractional factorial design of experiments was used to ana-

lyze the TOC degradation data, and to build an empirical model to
explain the observed degradation. The effects on TOC degradation
Fig. 1. Cell growth of each bacterial strain in synthetic wastewater.
caused by individual strains or by their combinations were evalu-
ated by T statistics and analysis of variance (ANOVA). A two-level,
1/4 fractional design with sixteen runs (26−2 ) was employed to eval-
uate the individual and combined effects of the six different strains
using the initial cell concentrations as the factors and the maxi-
mum TOC removal efficiency (rmax ) as the response. The levels of
experimental factors and the selected runs are discussed in Section
3.2. A significant level of 5% was used as the criterion to reject the
null hypothesis. All analyzes were performed using the MINITAB
Software (Release 14, Minitab Inc., State College, PA).

3. Results and discussion

3.1. Biodegradation by a single bacterium

The cell growth of an individual microorganism and biodegra-

dation of TOC by each of the six strains were examined in batch
reactors. All batch experiments started with the same initial optical
density (0.12–0.13 OD600) and TOC concentration (ca. 135 mg/L).
Figs. 1 and 2 show the growth of six strains of bacteria, i.e., Agrobac-
terium sp. (R1 ), Bacillus sp. (R2 ), E. cloacae (R3 ), Gordonia (R4 ), P.
stutzeri (R5 ), and P. putida (R6 ), and the residual TOC concentration
of the synthetic wastewater as a function of time. All the bacterial
growth (OD600 vs. time) followed a typical three-phase pattern:
a lag period, an exponential phase and a stationary phase. The
observed TOC consumption closely mirrored the cell concentration.
When the cell concentration began to increase, the TOC degradation
started, and the TOC degradation ceased when the bacteria reached
their stationary phase. These well-defined growth phases demon-
strated that these bacteria can appropriately utilize the substrate in
the synthetic wastewater to support their growth in the suspended
media.
To evaluate the performance of each of the six strains in degrad-
ing TOC in the synthetic wastewater, the kinetic parameters of
Fig. 2. Biodegradation of synthetic wastewater by each bacterial strain.
biomass growth and substrate utilization are compared in Table 1.
The parameters include the lag time for cell growth (Tlag ), the ini-
tial cell optical density (Linitial ), the maximum cell optical density
(Lmax ), the time to reach stationary phase (i.e., at stable TOC con-
centration, Ts ), the maximum TOC removal efficiency (rmax ), the
maximum specific growth rate of cell max (h−1 ), and the specific
substrate utilization rate q (g TOC/(d–gVSS)).
From the kinetic parameters in Table 1, the growth pattern of
E. cloacae (R3 ) and Gordonia (R4 ) are comparable. They had rela-
tively higher TOC degradation ability with the greatest rmax (88%)
and the shortest Tlag (1.1 h). Moreover, Enterobacter had a slightly

Table 1
Characterization of cell growth and substrate degradation for the six strains of inocula.

Parametersa Agrobacteria Bacillus sp. Enterobacter Gordonia P. Stutzeri P. putida

Tlag (h) 4.8 2.5 1.1 1.1 5.8 2.2

Linitial 0.126 0.130 0.130 0.123 0.126 0.130
Lmax 0.935 0.644 0.594 0.662 0.578 0.830
Ts (h) 20 10 8 10 13 18
max (h−1 ) 0.506 0.385 0.772 0.588 0.366 0.358
rmax (%) 77 55 88 88 80 84
qmax (g TOC/(d–gVSS)) 3.73 2.08 4.83 4.74 4.11 2.59
a
Tlag : the lag time for cell growth, Linitial : the initial cell density, Lmax : the maximum cell density, Ts : the time required for getting stable phase of TOC degradation, rmax :
the maximum TOC removal efficiency, : the specific growth rate of cell, and qmax : the maximum specific substrate utilization rate.
 Y. Chen et al. / Journal of Hazardous Materials 171 (2009) 948–953 951

Table 2
Applied levels of independent variables in the fractional factorial design.

Variable Bacteria species Applied level of cells

density (OD600 )
−(Low) +(High)

R1 Agrobacterium 0 0.2
R2 Bacillus sp. 0 0.2
R3 E. cloacae 0 0.2
R4 Gordonia sp. 0 0.2
R5 P. stutzeri 0 0.2
R6 P. putida 0 0.2

greater qmax (4.83 g TOC/(d–gVSS)) and max (0.77 h−1 ) and shorter
Ts (8 h). These results revealed that Enterobacter was the most
effective strain for degrading the organic substrates in the syn-
thetic wastewater. Furthermore, with their short Tlag and Ts , E.
cloacae (R3 ) and Gordonia (R4 ) may evolve into the predominant
species in the mixed-culture if all six strains were introduced.
Their relatively higher rmax also indicate that the two strains
can resist low-substrate conditions better than the other strains.
Bacillus sp. had the lowest TOC degradation ability (rmax = 55%,
qmax = 2.08 g TOC/(d–gVSS)), thus may not play a role in a mixed-
culture environment. Agrobacteria took significantly longer time to
reach the stationary phase compared to other strains, although it
has the highest yield among the tested strains (greatest Lmax ). P.
Putida could achieve reasonably high rmax (84%). However, its small
qmax and max as well as long Ts indicate that it is a slow grower.
3.2. Optimization of inoculum
A fractional factorial design of experiments was conducted to
identify the predominant strains or strain combinations affecting
the substrate degradation in the synthetic wastewater. The maxi-
mum TOC removal efficiency, rmax , was measured after inoculated
with different strains or strain combinations. The fractional design
experiments were performed at two levels of inoculated concen-
trations. A low level serves as the reference point for determining
rmax enhancement at the higher level. The low level (OD600 = 0)
was denoted as −1 and high level (OD600 = 0.2) was denoted as +1
(Table 2). In this set of experiments, we mainly focused on screen-
ing out the bacterial strains that can enhance the removal of TOC in
wastewater. Two-level experimental design is effective for this pur-
pose [18]. A middle level (0) experiment was not performed since
such experiments are typically to test the hypothesis of linearity,
which is not within the scope of the current study.
Fig. 3. Normal probability plot of the effects of different strain combinations on the
TOC removal rate.
A full factorial design for the six bacteria (factors) that requires
64 (26 ) experiments provides one with sufficient information to
evaluate the whole set of main effects as well as interaction effects.
The main effects and the lower-order interactions, however, are
usually the most significant terms [18]. In this work, a 26−2 frac-
tional factorial design consisting of 16 factorial runs was performed
to reduce the experimental efforts. This allows all experiments to
proceed in parallel to avoid possible impact caused by different
experimental blocks. The order in which the experiments were per-
formed was randomized. The selected set of experiments and the
measured rmax are listed in Table 3. All experiments were performed
for three replicates.
Fig. 3 is a normal probability plot of the effects of different
strain combinations on rmax . The effects that lie along the normal
probability line are negligible, whereas significant effects are those
square points far from the normal probability line. For single strain,
R3 , R4 and R6 show significant enhancement on the value of rmax
(p = 0.008, 0.009 and 0.023, respectively). The other strains did not
produce significant effects on their own within the levels tested.
Our results are consistent with the results of growth kinetics
experiments using single strains (Table 1), which showed E. cloacae
(R3 ), Gordonia (R4 ) and P. putida (R6 ) can degrade TOC in the syn-

Table 3
Results of the 3-replicate FFD experiments on the maximum TOC removal efficiency (rmax ).a .

Run Coded variables

X1 Agrobacterium X2 Bacillus X3 E. cloacae X4 Gordonia X5 P. Stutzeri X6 P. putida rmax (%)

1 −1 −1 −1 −1 −1 −1 18.2
2 1 −1 −1 −1 1 −1 12.0
3 −1 1 −1 −1 1 1 37.7
4 1 1 −1 −1 −1 1 27.2
5 −1 −1 1 −1 1 1 55.2
6 1 −1 1 −1 −1 1 47.0
7 −1 1 1 −1 −1 −1 27.0
8 1 1 1 −1 1 −1 22.6
9 −1 −1 −1 1 −1 1 51.2
10 1 −1 −1 1 1 1 34.5
11 −1 1 −1 1 1 −1 37.7
12 1 1 −1 1 −1 −1 38.2
13 −1 −1 1 1 1 −1 60.2
14 1 −1 1 1 −1 −1 78.0
15 −1 1 1 1 −1 1 85.0
16 1 1 1 1 1 1 82.4
a
Xi = (Li − 0.1)/0.1; i = 1,2,3,4,5,6, indicate six strains respectively; Li is the actual level of cells concentration (OD600 ).
952 Y. Chen et al. / Journal of Hazardous Materials 171 (2009) 948–953

Table 4 Table 5
Estimated effects and coefficients of the empirical model for rmax . Analysis of variance for the maximum TOC removal efficiency (rmax ).

Term Effect Coefficient T p Source DF SS Adj MS F p

Constant 44.631 36.81 0.001 Main effects 6 6659.68 1109.95 47.20 0.021
X1 −3.787 −1.894 −1.56 0.259 2-Way interactions 7 970.61 138.66 5.90 0.153
X2 0.188 0.094 0.08 0.945 Residual error 2 47.04 23.52
X3 25.087 12.544 10.35 0.009
Total 15 7677.3
X4 27.537 13.769 11.36 0.008
X5 −3.688 −1.844 −1.52 0.268 2 2
S = 4.84955; R = 99.39%; R (adj) = 95.41%; SS: sum of squares; DF: degree of free-
X6 15.787 7.894 6.51 0.023 dom; MS: mean sum of square.
X1 × X2 −0.463 −0.231 −0.19 0.866
X1 × X3 4.438 2.219 1.83 0.209
X1 × X4 3.558 1.779 1.47 0.280 of the two strains. Table 5 shows the analysis of variance (ANOVA) of
X1 × X5 −6.038 −3.019 −2.49 0.130
X1 × X6 −5.713 −2.856 −2.36 0.143
the empirical model. The ANOVA evaluates the significance of the
X2 × X4 4.683 2.341 1.93 0.193 main effects and interaction terms on rmax . The results showed a
X3 × X4 10.913 5.456 4.50 0.046 coefficient of determination (R2 ) of 99.39%, indicating the accuracy
of the empirical model to explain experimental data. The F statistics
was calculated as ratio of adjusted mean square of regression to
thetic wastewater more effectively. In addition, there was only one the adjusted mean square of residual. The high F and low p values
significant interaction term in the model from R3 R4 (p = 0.046), indi- (0.021) of the main effects indicate that certain single strains (R3 ,
cating that coexistence of R3 and R4 produced a synergistic effect. R4 , and R6 ) contributed more significantly to the response than the
The coefficients of the empirical model, T statistics and the asso- significant two-way interaction (R3 R4 ). The enhancement on TOC
ciated p values for the linear and interaction terms are shown in removal was mainly from the single strain inoculation rather than
Table 4. By applying multiple regression analysis on the experimen- the interaction term.
tal data, a second-order polynomial model in coded unit explains
the role of each variable and their second-order interactions.
3.3. Bioaugmentation of the selected strains on municipal
max = 44.631 − 1.894X1 + 0.094X2 + 12.544X3 + 13.769X4 wastewater treatment

− 1.844X5 + 7.894X6 − 0.231X1 X2 + 2.219X1 X3

+ 1.779X1 X4 − 3.019X1 X5 − 2.856X1 X6 + 2.341X2 X4
+ 5.456X3 X4
where Xi is cell concentration in coded unit (+1 or −1), i = 1–6 for
the six strains, respectively. After neglecting the insignificant terms
based on the 5% level of significance, the empirical model becomes:
max = 44.631 + 12.544X3 + 13.769X4 + 7.894X6 + 5.456X3 X4
Eq. (5) demonstrates that R3 , R4 and R6 were responsible for
most of the observed TOC degradation. There is only one significant
interaction term (R3 R4 ). This indicates additional synergistic effect
Considering the synergistic interaction E. cloacae (R3 ) and Gor-
donia (R4 ) and the significant effect from P. putida (R6 ) for degrading
TOC in the synthetic wastewater, the three strains were selected for
(4) additional TOC degradation experiments using municipal wastew-
ater. Agrobacterium sp. (R1 ), Bacillus sp. (R2 ) and P. stutzeri (R5 ),
which showed insignificant effects on enhancing TOC removal,
were excluded. It is hypothesized that the three selected strains
will improve the treatment performance of municipal wastewater.
To verify this hypothesis, three parallel experiments were carried
out in an activated sludge process treating municipal wastewa-
(5)
ter primarily from domestic sources. One is bioaugmented with
600 mL of the selected strains (R3 R4 R6 ) (each strain has 200 mL
inoculum at a concentration of OD600 = 2.5), the other is bioaug-
mented with 600 mL of six strains (R1 R2 R3 R4 R5 R6 ) (each strain has
100 mL inoculum of OD600 = 2.5). The controlled experiment was

Fig. 4. Residual TOC concentration and TOC removal efficiency in bioaugmented and control systems.
 Y. Chen et al. / Journal of Hazardous Materials 171 (2009) 948–953
operated without bioaugmentation. The experiments were per-
formed in three 2-L laboratory-scale activated sludge bioreactors.
Aeration was applied using porous stone sparger to keep the dis-
solved oxygen concentration greater than 3 mg/L. The mixed liquor
suspended solids (i.e., biomass) concentration and hydraulic reten-
tion time (HRT) were kept at 3 g/L and 4 h, respectively.
The treatment was operated for 42 days under the above opera-
tion conditions. The results are shown in Fig. 4. When influent TOC
concentration was below 92 mg/L before Day 17, there were not
significant TOC removal differences among the three experiments.
The TOC removal efficiencies were in the range of 72–80%. With
the influent TOC concentration increasing, the removal efficiency
in the reactor inoculated with the three selected strains (R3 R4 R6 )
was substantially higher than the other two reactors. In addition, a
removal efficiency decline was observed in the controlled experi-
ment. This indicates that the selected strains are capable of handling
a greater organic loading during the treatment. When influent TOC
concentration was further increased to above 200 mg/L, the TOC
removal efficiencies in the bioaugumented reactors were clearly
better (10–15% higher) than the controlled reactor. Furthermore,
the reactor inoculated with the selected three strains also yielded a
better treatment performance compared to the reactor inoculated
with six strains (77–79% vs. 80–84% for TOC removal).
Hisashi et al. [2] used the mixture of three photosynthetic bacte-
ria to established microbial consortia, they achieved simultaneous
removal of COD (89%), phosphate (77%), nitrate (99%) and H2 S
(99.8%). In their experiments, COD reduction (89%) was relatively
higher. However, the removal efficiency was achieved after 48 h of
operation. In this study, reasonably high TOC removal efficiency was
attained in 4 h, which is a relatively short hydraulic residence time.
In addition, it was demonstrated that the introduction of the three
selected strains (R3 R4 R6 ) into the activated sludge system achieved
higher TOC removal at higher organic loadings. This suggests that
the three introduced R3 R4 R6 were compatible with each other and
with the indigenous microbial populations. In other words, the
‘right’ inocula (e.g., R3 R4 R6 ) are more advantageous to inoculation
without selection (i.e., R1 R2 R3 R4 R5 R6 ). Since the community com-
position is mainly determined by strain-specific maximum growth
rate [19], the composition of indigenous culture of the activated
sludge may be shifted by attenuating or increasing certain strain
after the selected strains are introduced into the reactor. After a
period of continuous operation, the ultimate biodegradation effi-
ciency is mainly contributed by the growth of predominant strains
having higher specific maximum growth rate.
4. Conclusions
Six strains of microorganisms, i.e., Agrobacterium sp., Bacillus sp.,
E. cloacae, Gordonia, P. stutzeri, P. putida, were investigated for their
roles in TOC removal in both synthetic and municipal wastewa-
ter. Pure culture growth kinetic investigation was conducted to
understand the degradation rates of the studied strains. A fractional
factorial design of experiment was employed to select members
in an effort to construct a mixed-culture community that yields
optimal treatment performance for TOC removal in municipal
wastewater. Verified experiment has demonstrated that bioaug-
menting an activated sludge system with three selected genera,
i.e., E. cloacae, Gordonia and P. putida, led to improve TOC removal
efficiency of municipal wastewater under high organic loadings.
The fact that TOC removal efficiencies in the bioaugumented
reactors were significantly greater (10–15%) than the controlled
experiments indicates the effectiveness of implementing exper-
imental design approaches for biological screening of treatment
microorganisms. The observed improvement in TOC removal effi-
953
ciency may be due to syntrophic interactions among inoculated
strains and with the indigenous microorganisms. These results
reveal that the FFD is useful for biological screening and for char-
acterizing the interplays among different strains. A wastewater
bioreactor containing a population of highly specialized microor-
ganisms can enhance the performance of wastewater treatment.
Acknowledgements
This research was funded by the Engineer Research and Develop-
ment Center of the US Army Corps of Engineers under a subcontract
through Sam Houston State University (Subcontract No. W912HZ-
07-2-0010) and Natural Science Foundation of China (20676045).
The authors would also like to thank Dr. Jeffrey Davis, Mr. Scott
Waisner and Mr. Sabin Holland for their assistance in this research
work.
References
[1] H.L. Jiang, A.M. Maszenan, J.H. Tay, Bioaugmentation and coexistence of two
functionally similar bacterial strains in aerobic granules, Appl. Microbiol.
Biotechnol. 75 (2007) 1191–1200.
[2] N. Hisashi, K. Tomohiro, W. Masanori, S. Ken, Simultaneous removal of chem-
ical oxygen demand (COD), phosphate, nitrate and H2 S in the synthetic
sewage wastewater using porous ceramic immobilized photosynthetic bacteria,
Biotechnol. Lett. 22 (2000) 1369–1374.
[3] A. Briones, L. Raskin, Diversity and dynamics of microbial communities in engi-
neered environments and their implications for process stability, Curr. Opin.
Biotechnol. 14 (2003) 270–276.
[4] N.R. Smith, Z. Yu, W.W. Mohn, Stability of the bacterial community in a pulp mill
effluent treatment system during normal operation and a system shutdown,
Water Res. 37 (2003) 4873–4884.
[5] S. Thomas, S. Sarfaraz, L.C. Mishra, L. Iyengar, Degradation of phenol and phe-
nolic compounds by a defined denitrifying bacterial culture, World J. Microbiol.
Biotechnol. 18 (2002) 57–63.
[6] M.F. Carvalho, C.C.T. Alves, M.I.M. Ferreira, P.D. Marco, P.M.L. Castro, Isolation
and initial characterization of a bacterial consortium able to mineralize fluo-
robenzene, Appl. Environ. Microbiol. 68 (2002) 102–105.
[7] J.A. van Veen, L.S. van Overbeek, J.D. Van Elsas, Fate and activity of microorgan-
isms introduced into soil, Microbiol. Mol. Biol. Rev. 61 (1997) 121–135.
[8] B. Calli, B. Mertoglu, K. Roest, B. Inanc, Comparison of long-term performances
and final microbial compositions of anaerobic reactors treating landfill leachate,
Bioresour. Technol. 97 (2006) 641–647.
[9] K. McMahon, P.G. Stroot, R.I. Mackie, L. Raskin, Anaerobic codigestion of munic-
ipal solid waste and biosolids under various mixing conditions-II: microbial
population dynamics, Water Res. 35 (2001) 1817–1827.
[10] Y. Sekiguchi, Y. Kamagata, K. Nakamura, A. Ohashi, H. Harada, Fluorescence
in situ using 16S rRNA-targetted oligonucleotides reveals localization of
methanogens and selected uncultured bacteria in mesophilic and thermophilic
sludge granules, Appl. Environ. Microbiol. 65 (1999) 1280–1288.
[11] G. Annadurai, L.Y. Ling, J.-F. Lee, Statistical optimization of medium components
and growth conditions by response surface methodology to enhance phenol
degradation by Pseudomonas putida, J. Hazard. Mater. 151 (2008) 171–178.
[12] H.C. Chen, Optimizing the concentrations of carbon, nitrogen and phosphorous
in a citric acid fermentation with response surface method, Food Biotechnol. 10
(1996) 13–27.
[13] A. Gaujaca, S.E. Elissandro, S. Navickieneb, S.L.C. Ferreirac, H.S. Dóreab, Multi-
variate optimization of a solid phase microextraction-headspace procedure for
the determination of benzene, toluene, ethylbenzene and xylenes in effluent
samples from a waste treatment plant, Chromatogr. A 1203 (2008) 99–104.
[14] D. Chen, C.-J. Lin, R. G. Jones, et al., Deployable decentralized biofilm system to
degrade organic carbon, nutrients and benzene in wastewater, J. Environ. Eng.
(submitted for publication).
[15] M. Bramucci, H. Kane, M. Chen, V. Nagarajan, Bacterial diversity in an industrial
wastewater bioreactor, Appl. Environ. Microbiol. 62 (2003) 594–600.
[16] J.L. Tisinger, D.J. Drahos, Use of respirometry to evaluate the efficancy of bio-
logical products for the biodegradation of wastewater grease and oil, in: School
of Civil Engineering, Purdue University (Ed.), Proceedings of the 51th Industrial
Waste Conference, Ann Arbor Press, Chelsea, MI, 1997, pp. 472–482.
[17] L. Liu, X.W. Li, S.J. Liu, Z.P. Liu, Isolation and identification of a PAHs-degrading
Strain Gordonia sp., He4 and Its dynamics during bioremediation of phenan-
threne polluted soil, Chin. J. Environ. Sci. 28 (2007) 617–622.
[18] D.C. Montgomery, G.C. Runger, N.F. Hubele, Engineering Statistics, fourth ed.,
John Wiley & Sons, Hoboken, NJ, 2007, pp. 341–413.
[19] B. Kiesel, G.U. Balcke, J. Dietrich, C. Vogt, R. Geyer, Microbial community shifts as
a response to efficient degradation of chlorobenzene under hypoxic conditions,
Biodegradation 19 (2008) 435–446.