Available online at www.sciencedirect.

com

Bioresource Technology 100 (2009) 5058

Aerobic selectors in slaughterhouse activated sludge

systems: A preliminary investigation
Nayef Z. Al-Mutairi *
Civil Engineering Department, Kuwait University, P.O. Box 5969, Safat 13060, Kuwait
Received 25 September 2006; received in revised form 2 December 2007; accepted 3 December 2007
Available online 15 July 2008

Abstract
Filamentous bulking at a slaughterhouse activated sludge treatment plant signicantly reduced mixed liquor settling properties, which
caused many operational problems and worsening in euent quality. The main cause of this condition was attributed to signicant levels
of inuent readily biodegradable COD, which was present primarily in the form of organic acids. An aerobic selector was chosen to
eradicate the usual bulking incidents of slaughterhouse wastewater treatment plants. Other plant enhancements included increased aeration batch reactor volume, and provision of step feed capability.
Comparison of data before and after aerobic selector installation showed a signicant improvement in mixed liquor settleability,
which eradicated the need for chemicals that had been used to control laments and to control euent solids loss. The additional volume
of the aeration and chemicals eliminations from the activated sludge system also served to eliminate aquatic toxicity in the treated
euent.
2007 Elsevier Ltd. All rights reserved.
Keywords: Slaughterhouse wastewater; Selector; Filamentous bulking; Volatile fatty acids; Toxicity

1. Introduction
The activated sludge process has found vast application
as an ecient means of treating wastewater. The purpose
of the activated sludge process is to remove soluble and
insoluble organics from the wastewater and to change this
material into a occulent microbial suspension that sinks
well in a conventional tank. In most cases, the nature of
the wastewater will impose the preferred process modications, mainly for the purpose of maintaining appropriate
mixed liquor settling properties (Eckenfelder and Musterman, 1995).
Basically, activated sludge comprises a microbiological
enrichment culture consisting of a mixed, and largely
uncontrolled, consortium of micro- and macro-organisms
(Richard, 1989). It is important to maintain the growth
of oc-forming bacteria on wastewater organics, which will
*

Tel.: +965 4817524.

E-mail address: nayef@civil.kuniv.edu.kw

0960-8524/$ - see front matter 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2007.12.030

settle easily under gravity in the nal clarier to obtain a

claried supernatant and a thickened return sludge. However, not all bacteria in the activated sludge process are
oc-formers. Many dierent types of lamentous bacteria
have been identied in activated sludge and play important
roles in wastewater treatment. Filamentous bacteria
directly aect sludge settling as they make provision for
the rigid support network or backbone upon which ocforming bacteria can adhere and grow into suitable activated sludge ocs (Richard, 1989). Filamentous bacteria
may be considered detrimental to wastewater treatment
when they occur in excessive quantities, but are just as
important in the development of activated sludge ocs with
proper settling and clarication properties (Gerardimh
et al., 1990).
Filamentous organisms grow out of the activated sludge
ocs and lead to a diuse and very irregular shape of the
activated sludge ocs. This phenomenon can be either classied as bulking sludge, oating sludge or foam. Bulking
sludge is dened as sludge with a sludge volume index

N.Z. Al-Mutairi / Bioresource Technology 100 (2009) 5058

(SVI) of more than 150 mL/g. Floating sludge or foam

builds a stable layer at the surface of the aeration basin
or nal clarication basin. Filamentous microorganisms
can also be good indicators of conditions prevailing in an
activated sludge system on a microbiological level. Low
organic sludge loads (<0.1 kg BOD5/(kgTSS d)), low
wastewater temperatures (1215 C) and high sludge ages
are considered favoring factors to the abundant development of Microthrix parvicella, which is the most frequently
occurring organism (Knoop and Kunst, 1998). The indications given by the lamentous bacteria could also be of low
dissolved oxygen (DO) (e.g. Sphaerotilus natans), presence
of septic waste (e.g. Thiothrix spp.), nutrient deciency (e.g.
Haliscomenobacter hydrossis) and low pH in the system
(e.g. fungi) (Jenkins et al., 1993).
Furthermore, the existence of volatile fatty acids (VFAs)
in wastewater has long been linked with lamentous phenomena (Richard and Kampfer, 1997). VFAs are the primary constituent of readily biodegradable COD in
inuent slaughterhouse wastewaters. Direct measurements
of changes in specic organic pools in slaughterhouse
wastewater have not been performed so far. Richard
and Kampfer (1997) measured protein, carbohydrate,
and VFAs in two types of wastewater of dierent ages
and found signicant dierences in composition. Unfortunately, little research has focused on the production of
VFAs from slaughterhouse wastewater. In addition, practical design criteria for the purication process specically
oriented toward slaughterhouse WWTPs are not generally
available.
Since the introduction of conventional activated sludge
systems, sludge bulking has been one of the major problems aecting every biological process (Bitton, 1999).
Two toxicants, chlorine and hydrogen peroxide, have been
used to selectively kill lamentous organisms in conventional activated sludge systems and therefore alleviate
symptoms of activated sludge bulking (Jenkins et al.,
1993). Chlorination is however, a last alternative in the
control of bulking. It should only be contemplated when
euent from the bulking plant is likely to cause environmental damage to the receiving water. If used correctly it
provides a rapid alleviation of bulking, but if used incorrectly, it may completely inhibit all treatment (Horan,
1990). On the other hand, Hwang and Tanaka (1998)
showed that the application of chlorination had no eect
on M. parvicella reduction or foaming suppression and that
a more eective and economical method for the control
thereof still needs to be established.
Often, non-specic methods are used to combat bulking
or oating sludge. The use of alum and iron salts or the
application of strong oxidizing agents belongs to these
methods. One disadvantage of using chemicals is that
oc-forming bacteria are also inuenced negatively. This
can lead to the failure of the treatment process. Additionally, the use of chemicals results in an increase in toxicity
and in sludge mass, which has to be disposed of or recycled
(Al-Mutairi, 2006).

51

Theoretically, the massive growth of lamentous organisms could be avoided by increasing the sludge load, but a
higher sludge load leads to a decrease of the sludge age.
The cause for the excessive growth of lamentous organisms can be controlled by specic measures. A more
detailed understanding of the physiology and biochemistry
of lamentous bacteria is still required for eective longterm control of bulking and foaming. For example, by
improving wastewater quality (e.g. the prevention of H2S
formation) or by applying modications in the operation
mode (e.g. aerobic or anaerobic selector) lamentous
organisms are taken out of the system. Moreover, plant
conguration provides operational conditions that suppress lamentous bacteria and favor growth of oc-formers, leads to high sludge settleability characteristics
(Vaiopoulou et al., 2007). Mueller et al. (2000) successfully
tested a side-by-side full-scale aeration of a modied contact stabilization process incorporating an anaerobic selector to control lamentous bacteria in wastewater.
The application of selector reactor technology has
become the promoted method for control of lament proliferation to enhance sludge settleability in activated sludge
systems (Pujol and Canler, 1994; Kruit et al., 1993; Eikelboom, 1994). A successful application of selectors relies
on detailed knowledge of: (a) physiology and substrate
requirement of the lamentous microorganisms, (b) wastewater composition and (c) substrate removal kinetic in the
selector system (Andreasen et al., 1999). Madoni and
Davoli (1997) noticed a complete suppression of the
growth of nocardio forms was obtained where returned
activated sludge and wastewater were mixed together under
both anoxic conditions and high sludge loading. However,
under moderate sludge loading, the selector suppressed the
growth of nocardio forms but was not able to improve settlement of the sludge and prevent the growth of other lamentous microorganisms responsible for foaming, such as
M. parvicella.
A principal components analysis of patterns and level of
microbial activity suggested that microbial communities
statistically dierentiate between the selector and conventional activated sludge system. There was seasonal variation in the structure and function of the microbial
community in conventional samples while, for the selector
system, there were no identiable dierences between the
data communities (Al-Mutairi, 2007).
Aerobic selectors are an engineered system that uses differential growth kinetics to promote the development of
oc-forming bacteria rather than lamentous bacteria.
This is accomplished because most oc-forming bacteria
grow faster than lamentous bacteria at higher BOD loading rates. Conversely, most lamentous organisms grow
faster than oc-forming organisms at lower BOD loading
rates. Basically, the process goal is to assimilate as much
of the soluble organics as fast as possible, thereby denying
the lamentous microbes this food source. At high BOD
concentrations, the oc-forming bacteria have a higher
BOD uptake rate than the lamentous bacteria and so

52

N.Z. Al-Mutairi / Bioresource Technology 100 (2009) 5058

the oc-forming bacteria are more competitive for the food

source. This action enhances the growth and formation of
better settling, oc-forming microbes, and limits the
growth rate of the lamentous organisms.
Furthermore, biological treatment of slaughterhouse
wastewaters may not be straightforward due to high variations in ow and chemical characteristics. Those factors,
combined with low temperatures in the months of DecemberFebruary and high temperatures in the months of
MayAugust may make consistent biological treatment
dicult (Al-Mutairi et al., 2003). The selector is also
employed to smooth out uctuations in inuent composition and allowing the wastewater treatment plant to run
more uniformly. Inuent wastewater characteristics can
be impacted by uctuations in temperature and pH, as well
as ow surges. Proper design of the selector includes providing the correct retention time under aeration, with the
exibility to change this as wastewater characteristics
change. Adjustable dissolved oxygen targets, avoidance
of prolonged over-aeration, variable return activated
sludge (RAS) rates, control of nal tank blanket levels,
etc. are all useful tools to reduce the eects of foaming
and bulking.
Low F/M bulking (Microthrix parvicella, type 0092,
type 0041 and Nostocida I) and foaming (Microthrix parvicella, and Nocardia sp.) can be controlled in a selector by
inducing a high substrate gradient in the selector zone
which stimulates oc forming microorganisms with high
substrate uptake and storage rates which allows them to
out compete over the lamentous microorganisms through
the resulting feedstarve cycle.
The experimental plan was divided into two parts. In the
rst part, experiments were conducted using a conventional
activated sludge unit for two years and in the second, using
a contact/aeration unit (selector) for another year. The
objective of the present work was to compare the selector
process with the conventional activated sludge process in
terms of COD treatment capacity, oxygen uptake rate
and toxicity. In addition, assess the selector process performance in terms of process capability to enhance nal clarication and settleability. Furthermore, the study results
are expected to be used to adjust euent treatment technologies or regulations, so that the potential threat to the environment by such materials will be reduced.
2. Plant description
Hawally wastewater treatment plant (WWTP) used for
this study, a relatively small plant (360 m3/day), was
designed to use activated sludge process to achieve secondary treatment. The existing biological reactors are congured as a complete mix system and are limited to MLSS
concentrations of 18002000 mg/L because of the deleterious eects of sludge bulking and foaming at higher concentrations. This system was prone to prolonged bulking
periods, with Thiothrix I and II, Nostacoida limicola, S.
natans, H. hydrossis, and Types 0411, 1701, 0041, 0803,

0914, and 0961 being identied in the mixed liquors. In

light of the new permit limits, the WWTPs desire to
expand the facility by 25%, and given the extreme space
limitations on the existing site, increased process control
is required to allow higher operating MLSS concentrations
to achieve the desired treatment.
Furthermore, Nakhla and Lugowski (2003) tested various measures at a food-processing wastewater treatment
plant to prevent sludge bulking by type 0041 and 0675 lamentous microorganisms. The aerobic selector was the
most successful system used, reducing SVIs to 79 cm 3 g 1
in 2 weeks. Sludge settleablity was found to be inversely
proportional to the aerobic selector food-to-microorganism ratio. The optimal aerobic selector loading was found
to be 1.82.7 kg BOD5 kg MLVSS d 1.
Hawally WWTP was evaluated to dene potential operational changes and upgrades that would allow achievement of the permit limits consistently. Finally, the
WWTP was upgraded with an aerobic biological selector
to improve sludge settling characteristics. The primary
treatment in Hawally WWTP is achieved by a rotary screen
and a dissolved air otation unit (DAF). Fat, oil, grease
(FOG), settleable solids, and sand in the slaughteroor
euent are removed in the otation unit. The pretreated
water will ow to the aeration tank of the biological treatment system. After settling, the supernatant is decanted
from the tank. A commercial occulant known as Praestol
is dosed in the feed pipe to the second stage otation unit
to concentrate the colloidal and suspended pollutants in a
oc-like agglomerate which can be separated from the
water by otation. Flocs, solids, and FOG are removed
by the second stage DAF process. The WWTPs original
design-rated monthly average capacity: ow 310 m3/day
and COD 1962 kg/day. The original design predicted
95% BOD and TSS removal eciencies with resultant euent concentrations of less than 30 mg/L. However, the
treatment system experienced frequent equipment and process limitations and was unable to meet its design goals.
3. Methods
Crude and settled sewage of the conventional activated
sludge plant and the contact stabilization plant were examined. Two hundred fty ml sterile bottle samples were used
(3) of mixed liquor, obtained from the aerobic zones of
the aeration and the selector tank. The mixed liquor samples were stored half-full in the bottles so as to maintain
aerobic conditions for lamentous survival during sample
transit. The procedures employed for identication of
laments were those used by Jenkins et al. (1993). Samples
were evaluated microscopically (1000, oil immersion) so
as to determine the eects of the operation of the conventional activated sludge systems and a selector system
on the integrity of the laments and oc structure. Respiration measurements were generally obtained under periods of load. The VFAs (formate, acetate, propionate,
n-butyrate, and isobutyrate) were measured on a Dionex

N.Z. Al-Mutairi / Bioresource Technology 100 (2009) 5058

ion-chromatograph with a suppressed conductivity detector. The VFA concentrations lower than 1 mg/L could be
detected. The sludge index was also observed during the
time of the experiment. Other parameters were measured
according to Standard Methods for the Examination of
Water and Wastewater (1995).
Finally, acute bioassay is a widely used tool for adding
toxicity data to biological/chemical/physical eld surveys
of wastewater euents. Mean 15-min EC50 values (a dose
that produces a 50% decrease in a sublethal response) for
the wastewater euent before and after the selector installation. The Microtox assay is based on the inhibition of
light production of the bioluminescent microbe, Vibrio scheri, in the presence of toxicants (Microbics, 1992). A photometer is used to provide temperature control and
measurement of light emission. An M500 MicrotoxTM system as well as reagents and protocols followed were
obtained from Microbics Corporation (Microbics, 1992).
Data reduction and statistical analysis were performed by
Microbics statistical software.
4. Results and discussion
4.1. Analysis of bulking problems
A microbiological assessment of the slaughterhouse
wastewater before the upgrade of the plant showed that
many lamentous bacteria were present in the mixed liquor
including Thiothrix I and II (predominant), N. limicola, S.
natans, H. hydrossis, and Types 0411, 1701, 0041, 0803,
0914, and 0961. The Thiothrix spp. was so concentrated
that the bacteria imparted a bright orange color to the
mixed liquor.
Consequently, the use of chemicals to limit bulking was
followed in concomitantly with more vital corrective measures. Once control is obtained, sucient time was allowed
for changes in the H2O2 dose to take eect. With lower
doses of H2O2 (e.g., <100 mg/L), several days may pass
before the eects are evidenced in the activated sludge oc.
Furthermore, related operating data and wastewater
parameters should be collected before, during and after
the trial to assist in interpreting the results.
On the other hand, lamentous bulking in Hawalli
WWTPs created a sludge that was light, had a high sludge
volume index, and was being washed out of the secondary
clariers even with very high polymer dosing rates. Solids
that were washed out were captured in part by the tertiary
DAF clarier. However, the recovered sludge could not initially be returned to the aeration tanks, so any biomass that
washed out of the secondary clariers was lost from the
activated sludge process. As a result, the F/M increased
and new high F/M laments was established.
The returned activated sludge was dosed with hydrogen
peroxide, which was only partially successful in killing laments. Higher doses resulted in creating conditions that
were favorable to a lament (Type 0803) that is normally
only found in oxygen activated sludge systems. Polymers

53

were added at high doses to keep euent TSS under control throughout the two years of operation (Al-Mutairi
et al., 2004). Hydrogen peroxide was used to correct a serious lamentous bulking situation or, preferably, to prevent
one from occurring until adjustments can be made to
remove the cause. When applied to the return activated
sludge, H2O2 supplies dissolved oxygen which helps restore
the microbial activity necessary for eective operation,
while selectively oxidizing the laments which retard
settling.
Furthermore, healthy activated sludge systems normally
contain about 15% polysaccharides on a dry weight basis.
Hawally biomass before the installation of the selector contained polysaccharide concentrations as high as 60%,
because of the cells inability to metabolize high concentrations of readily biodegradable COD (RBCOD) (present as
VFAs in the inuent). This condition is known as zoogleal
bulking and is induced by inadequate availability of nutrients. In conditions where high RBCOD levels are present,
cells respond by trying to metabolize the substrate at high
rates and, in order to do so, require high concentrations of
biologically available nutrients and oxygen. In this case,
the biomass retain a gelatinous texture, which impeded
mixed liquor settleability, elevated SVI and euent TSS
concentrations, caused foaming on the aeration tanks
and a thick oating scum layer in the secondary clariers
that was very dicult to break up and remove. RAS chlorination did not correct this condition either.
The high doses of cationic and anionic polymers that
were used to trap solids and make them heavy enough to
settle, at times created a sticky sludge that the suction
sludge withdrawal mechanism was unable to remove. High
doses of defoamers were also used to contain the foaming
in the aeration basins. The high rates of chemicals used
to keep the slaughterhouse WWTP in compliance drove
up operational costs extensively.
Fermentation of starch in the wastewater to VFAs also
caused a reduction in pH in the equalization tank. VFA
levels of up to 400 mg/L as measured by distillation, were
observed in the tank. Table 1, summarizes limited VFA
speciation data during both the summer and winter
months. It is signicant that VFAs production at elevated
temperatures during summer months is nearly double that
observed at lower temperatures during winter months.
Consumption of readily fermented substrate may have
occurred during winter months causing lower concentration of VFAs.
Table 1
Speciation of VFAs in the equalization euent
VFA species

Acetate
Propionate
n-Butyrate
Isobutyrate
Formate

Concentration (mg/l)
Winter months

Summer months

91 22
62
51
3 0.3
0.6 0.1

166 32
29 5
10 2
81
7 0.1

54

N.Z. Al-Mutairi / Bioresource Technology 100 (2009) 5058

Sodium hydroxide was added to the raw wastewater to

control the septic conditions before it was pumped to the
treatment system to suppress the growth of lamentous
bacteria and to retain H2S in the dissolved form until it
could be oxidized in the aeration tanks. This procedure
was partially successful, but was discontinued because of
its high cost and because of the heavy scale deposition in
the equalization euent pipes resulting from the change
in wastewater chemistry. The impact of the termination
of pH correction was to elevate the level of VFA production in the primary clarier.
In addition, a remarkable increase in VFA was observed
in the equalization euents, as represented by the increase
from a mean value of 138 62 mg/L in the equalization
inuent to 217 119 mg/L in equalization euent and
then to 351 112 mg/L in the equalization euent after
pH termination. One-way analysis of variance within the
three means of VFA concentrations of the equalization
tank showed that the dierence between them were statistically signicant at a = 0.05 (F = 307 and p-value = 0.00).
The installation of biological selector at other treatment
plant facilities showed promise for creating conditions that
would favor the growth of oc forming bacteria over lamentous bacteria in wastewater containing high RBCOD
levels (Marshall and Richard, 1999). However, there was
little information in the literature that related directly to
slaughterhouse wastewater, and the slaughterhouse management had mixed success with the conventional activated
sludge technology. The existing aeration basins of the
Hawally slaughterhouse had not prevented the proliferation
of the lamentous bacteria because they had not created a
properly sized selector zone, and so they were removed. It
was also evident that the existing aeration system was
unable to maintain a steady acceptable DO concentration
at the inuent end of the aeration tanks. Also, excess air
entrainment in the mixed liquor contributed to settling
problems in the clariers. A temporary step feed system
was installed to distribute the loading along the length of
the aeration basins. This improved system operability, lent
credibility to the concept of adopting step feed as an alternative operating strategy.
4.2. Aerobic selector performance
The method selected to treat slaughterhouse wastewater
is inuenced by its characteristics. The slaughterhouse
wastewater contains starch-like, low molecular weight
organic compounds that are very biodegradable. Oxygen
uptake rate (OUR) tests were conducted to assess the level
of slaughterhouse wastewater biodegradability. Preliminary OUR measurements of the inuent, gave very high
readings but tended towards endogenous respiration rates
(in the range, 106240 mg/g h). In addition, the results of
mixed liquor respiration tests were in the range of 4
105 mg/g h. The high values observed in the inuent are
likely due to the physiological stress experienced by the
bacteria upon the exposure to an environment having

markedly changed conditions of temperature, dissolved

oxygen and pH.
The respirometry conducted, showed that the wastewater RBCOD was 8085% of the primary euent COD.
Thus, the wastewater was readily biodegradable which
would, in the presence of microorganisms, exert a high
demand for oxygen and nutrients. Comparable RBCOD
values for sewage are 116% and 3050% for other industrial wastewaters (Sperandio et al., 2000). The results of
this test indicated that the slaughterhouse wastewater was
amenable to treatment in an aerobic selector-assisted activated sludge system. A selector is a portion of an activated
sludge system that precedes the main bioreactor, receives
the inuent wastewater and RAS, and has a low residence
time (1530 min). This environment favors oc formers
over lamentous bulking organisms, which results in mixed
liquor with good settling properties.
The WWTPs upgrade design-loading monthly average
capacity: ow 650 m3/day and COD 1900 kg/day. Furthermore, there was a concern over the frequency of periods of sustained COD loadings that might not be
adequately addressed by a reactor volume design based
on a daily average COD loading. Review of inuent
COD loading data led to a maximum week condition being
selected as the design basis for reactor volume.
The design of aerobic selectors is controversial, with
some researchers advocating multiple steps to induce an
incremental F/M loading through the selector (Jenkins
et al., 1993). Others base their approach on a COD oc
loading (Al-Mutairi et al., 2003), and others yet base their
designs on complete COD uptake in the selector (Still et al.,
1996).
A two-cell aerobic selector was chosen over other
designs to provide operational exibility given the uncertainty of using selectors to treat slaughterhouse wastewaters. The selector volume is approximately 10% of the
overall volume under aeration. The design incorporates a
primary euent bypass capability to permit diversion of
excessively high loads around the rst stage of the selector
and to the main aeration basins. There was also a need for
two new aeration tanks of the same volume as the two
existing tanks to satisfy projected COD loadings. A provision was made to allow step feeding of primary euent to
four locations along the length of the aeration basin. Also,
a review of two years of COD loading rate data revealed
periods of sustained elevated loadings that lasted as long
as 6 h. This peak loading condition was used as the basis
of the design of the aeration system.
Following the startup of the modied treatment facility
in May 2003, performance has been excellent. Fig. 1ad
summarize monthly average euent characteristics for
2003/2004. They show an average soluble COD removal
of 79% across the selector. Soluble COD is typically
slightly higher than RBCOD, which suggests that the selector is functioning close to the predicted level of performance. The graphical summary of the data includes four
graphs: histogram of data with an overlaid normal curve,

N.Z. Al-Mutairi / Bioresource Technology 100 (2009) 5058

55

Fig. 1. (ad) Show histograms of monthly average inuent/euent characteristics for 2000/2001 data with an overlaid normal curve.

boxplot, 95% condence intervals for l, and 95% condence intervals for the median. This tool, along with normality assessment statistics generated using minitab
univariate procedure, alerts the investigator to non-normal
data distributions and the presence of outliers. The histogram clearly portrays information on location, spread,
and shape that enables the user to perceive subtleties
regarding the functioning of the chemical process that is
generating the data.
It can also help suggest both the nature of, and possible
improvements for, the physical mechanisms at the treatment plant. The mean of the inuent COD concentration
is 1144 mg/L (95% condence intervals of 1108 and
1182 mg/L). The standard deviation is 58 mg/L (95% condence intervals of 41 and 99 mg/L). On the other hand,
the mean of the selector COD concentration is 240 mg/L
(95% condence intervals of 217 and 263 mg/L). The standard deviation is 36 mg/L (95% condence intervals of 25
and 61 mg/L). The nal euent COD quality of the selector is very good with a monthly average total COD ranging
from 217 to 263 mg/L. A one-way analysis of variance of
all the means of soluble COD showed that the dierence
between all the euents were statistically signicant at
a = 0.05 (F = 1075.73 and p-value = 0.00).
The reliability of the selector can be assessed using principal component analysis (PCA) in terms of its ability to
produce consistently acceptable reclaimed wastewater.
The PCA technique is one of the most powerful and common techniques used for reducing the dimensionality of
large sets of data without loss of information. It implies a
mathematical procedure that transforms the overall set of
original variables into principal components (PCs).

The statistical results from Table 2 show, the unrotated

loadings and communalities which explains 100% of the
data variability and the communality values indicate that
all variables are well represented by these two factors.
The percent of total variability represented by the factors
does not change with rotation, but after rotating, these factors are more evenly balanced in the percent of variability
that they represent, being 44.7% and 35.0%, respectively.
Factor 1 has large negative loadings on inuent COD concentration ( 0.798), selector euent concentration
( 0.390), and nal DAF euent ( 0.196), and a small
positive loading on rst DAF euent COD concentration
(0.332). Factor 2 has a large positive loading on rst DAF
euent of 0.754 and small positive loadings of 0.347 and
0.434, respectively, on primary and selector euents, and
a negative loading on nal DAF euent. The rst factor
negatively loads on primary and on two variables, selector

Table 2
Principal component factor analysis of the correlation matrix showing the
unrotated factor loadings and communalities
Variable
Primary
euent
First DAF
euent
Selector
euent
Final DAF
euent
Variance
% Variance

Factor 1

Factor 2

Factor 3

Factor 4

Communality

0.798

0.347

0.244

0.428

1.000

0.332

0.754

0.524

0.216

1.000

0.390

0.434

0.468

0.342

1.000

0.196

0.547

0.506

0.298

1.000

1.5793
0.395

1.1761
0.294

0.8095
0.202

0.4352
0.109

4.000
1.000

56

N.Z. Al-Mutairi / Bioresource Technology 100 (2009) 5058

and nal DAF euents, that generally decrease when the

primary euent is reduced. We might consider factor 1
to be primary, selector, and nal DAF euents factors.
The second factor might be considered a rst DAF euent factor.
Table 2 clearly demonstrates that, in-spite of signicant
variations in the quality of primary euent, the nal DAF
euent remained essentially unchanged. This indicates that
the selector combined with the nal DAF system played an
important role in the stability of euent quality so as to
dampen variations in the quality of the euent. The nal
DAF euent consistently satised the water quality
requirements for irrigation. Variability in the primary euent quality may be taken as an indication of an inherent
diurnal or seasonal variations in inuent wastewater ow
and characteristics as well as process control practices.
Selector operational targets vary according to change in
wastewater characteristics. When the system before modication and the selector response were compared in terms of
SVI, as a result of dierent factors impacting the plant over
the past three years of operation. SVI values were highly
variable before treatment system modications. Values as
high as 600 mL/g were common with peak values as high
as 1100 mL/g. Furthermore, SVI signicantly dropped to
250 mL/g when chlorine was added to the RAS to rapidly
kill laments rather than wait for them to die o due to
changes in selector operation. However, when pH correction was terminated, SVI values began to increase dramatically to about 900 mL/g. In addition, the installation of
the selector showed an excellent SVI values in the range
of 80140 mL/g.
At startup on May 2003, the selector was operated with
the rst cell utilized for RAS reaeration during which oc is
saturated with oxygen before the inuent is fed to the selector, preparing the oc-forming bacteria for rapid COD
uptake and utilization. When RAS and inuent were both
fed to the second cell, the SVI gave excellent results with
SVI < 120 mL/g without RAS chlorination. However, a
test of a two-cell selector operation was initiated in September 2003 by feeding RAS and inuent to the rst selector
cell in which the performance was not as good, with SVI
in the range, 250300 mL/g.
To further test operational exibility part of the inuent
was step fed around the selector in November 2003. Minimal changes were made to step feed ow during this period.
In March 2004, the system reverted to RAS re-aeration. In
June 2004, SVI values started to decline to a low of 37 mL/
g on July, 2004 and euent turbidity increased. Step feed
ow was increased to grow some laments to serve as a oc
backbone. SVI values stabilized and slowly began to
increase until August, 2004 when step feed was terminated.
No step feed was used until September, 2004. The step feed
ow is now adjusted based on microscopic examination of
MLSS and daily SVI values. There have been times when
step feed has been completely turned o but this almost
always results in too few laments to support oc development and excessive globular zooglea growth.

Table 3
Microtox toxicity of wastewater treatment plant euents
TM

Location

EC50 (15 min test)

Activated sludge system

Selector system

Inuent
Selector euent
Final euent

15% (1020)
NA
3% (35)

13% (725)
70% (5985)
>100%

Values in parentheses are as % condence limits calculated by Microtox

software.

TM

5. Toxicity assessment
Results of acute toxicity tests on whole euents are
reported as EC50 (for eective concentration). Microtox
suggested mild acute toxicity from the slaughterhouse
wastewater inuent. The slaughterhouse wastewater euent before the selector installation severely depressed light
production in the Microtox assay. However, the selector
euent exhibited no acute toxicity (did not inhibit bacterial
light production) at the highest concentrations tested
(Table 3). The nal euent of the activated sludge system
indicated an increase in toxicity from 15% to 3% of EC50
during the tertiary chemical process, presumably due to
the overdose of polymer/alum and chemicals used. However, after the installation of the selector, the nal euent
achieved a toxicity reduction from 13% to >100% of EC50.
There are a number of potentially toxic constituents in
the slaughterhouse euent, but only three, alum/polymer,
chlorine and ammonia, are present in suciently high concentrations to produce the observed acute toxicity. Aluminum and polymer were present at relatively high
concentrations for the coagulation process and are almost
certainly the chief toxicants beside ammonia (Al-Mutairi,
2006). While aluminum concentrations, appear high in
absolute terms (0.70.9 mg/L) toxicity of this metal is mitigated at neutral pH. Although the total aluminum concentration appears high, it would be almost entirely particulate
and would therefore have very low biological activity.
Given the importance of alkalinity and hardness to toxicity
of many contaminants, it is unfortunate that these parameters were not measured. Generally, toxicity agreed with
expectations based on the chemical composition of each
euent. The euents diered both in the strength and
the nature of their toxicities.
6. Implications for selector design and operation
With a successful selector operation, several operational
issues have been identied and appropriate operator interventions have been developed. The selector has proven
eective in controlling lamentous bacteria levels and polysaccharide levels have stayed within normal mixed liquor
concentrations. Sludge bulking is well controlled and euent quality has been consistently good. Process adjustments
were made based on regular wastewater characterization,
microscopic examination of mixed liquor, (presence of
laments, globular zooglea and polysaccharides) and visual

N.Z. Al-Mutairi / Bioresource Technology 100 (2009) 5058

observations such as aeration tank foam and euent

quality.
F/M and SRT together dictate the required selector size.
Distributing the feed across the selector entrance was
needed to lower F/M in the selector. Operating at a
2 < F/M < 5 on COD basis, on the other hand, requires
proportionally higher reactor volume for a given SRT.
Therefore, results indicating slow gradual changes because
of long response times at SRT as high as 18 days, are
achievable. Solids deposition prevention in sumps and the
equalization tank just before the selector stage may
improve settleability and reduce VFAs production without
commensurate selector tankage requirements. However,
minimization before the selector would not aect any
potential reaction-limiting eects inside the selector. Temperature is an important design consideration for the selector process which may be critical depending on the regional
climate. Increased VFAs production at a high temperature
during summer season may require additional tankage,
higher SRT, and lower solids concentration to reduce
VFA production.
An important observation, was the tendency of solids to
oat in the equalization section of the selector system. Flotation appeared to be caused by gases produced within the
sludge blanket. Floating solids would accumulate until they
either plugged the submerged launder or escape into the
euent where they occasionally plugged the submersible
pump that conveyed supernatant to the selector unit. A
provision of an adequate oat solids collection and
removal-recycling system would be required. The rapid
buildup of grease in inuent sludge piping, was another
noteworthy operational problem observed in the plant.
During piping repairs, grease buildup that had reduced
the eective sludge piping diameter was observed. Provisions to facilitates regular pipeline cleaning and eective
dissolve air otation systems, should be considered at
slaughterhouses with heavy grease loadings.
Four additional considerations became evident during
the full-scale-experiments include, DO, organic loading,
toxicity, and wastewater characteristics variability. DO
should be maintained at 46 mg/L residual in the selector
to adequately match COD utilization rate. A taper aeration
system to provide DO residual gradient from 4 at head end
to 2 mg/L at discharge was installed. Excess organic loadings should be step fed around the selector to minimize
polysaccharide and globular zooglea growth. Adjustments
should be made gradually and in small increments. Toxic
chemicals (solvents, chlorination, alum, polymer, etc.) can
kill or inhibit the selector biomass. The additional volume
of the aeration and chemicals eliminations from the selector were needed to eliminate aquatic toxicity in the treated
euent. Good communication between manufacturing and
wastewater personnel to manage toxic chemical releases
and prevent upsets are needed.
The selector designs should be conservative and exible
to accommodate wide variations in wastewater characteristics. The key factor that the full-scale demonstration units

57

focused on was the liquidsolid separation step because of

the limited underow concentrations achievable using
gravity separation and the desire to reduce VFAs production. Equalization solids are undesirable because these solids include degradable solids that increase the loading on
the selector system and inert solids that accumulate in the
biological system and displace active biomass, thereby
reducing capacity.
7. Conclusions
The combination of aerobic selector, step feed, increased
aeration tank capacity, and tapered aeration has proven to
be eective in controlling lamentous bacteria at Hawally
slaughterhouse wastewater treatment plant. It was an interesting approach to overcome problems with oating suspended matter discharged from equalization tank, which
could be eliminated without using chemicals. Furthermore,
proper design of the selector includes providing the correct
retention time under aeration, with the exibility to change
this as wastewater characteristics change. In this regard, a
high aeration capacity is required so that DO concentration is high enough to penetrate the oc structure and
match COD utilization rate. In addition, minimum DO
residuals of 2 mg/L are recommended for both, and
4 mg/L preferred when VFAs are present.
RAS reaeration is a valuable tool to ensure complete
digestion of stored food in the mixed liquor solids. A
multi-cell selector can provide this function while also providing operating exibility and ecient maintenance capability. Step feed should be used to maintain a low lament
population to serve as a backbone for the bacteria oc and
prevent euent turbidity. Likewise, step feed maybe used
to shunt excess COD to the aeration tanks to minimize
the formation of globular zooglea and excess polysaccharides. It is also useful in controlling excessive nitrication
in the aeration tanks. Tapered aeration permits setting up
appropriate DO residual gradients along the length of aeration tanks. The process must be properly monitored and
controlled. Otherwise, the process can eventually become
unstable and euent quality deteriorates.
Finally, the Microtox acute toxicity test has been successfully used to measure the toxicity of wastewater in
slaughterhouse wastewater euent samples. Results indicated that initially the wastewater caused inhibitory eects
to the system because the toxicity of many chemicals is
more severe under low oxygen tensions. While the selector
conguration helped eliminate euent aquatic toxicity.
The toxicity of the three euents varied in strength and
character and was generally explicable based on chemical
data. However, toxicity tests revealed strong chronic toxicity in the conventional activated sludge system euent.
Future chemical analyses should include some measure of
ionic strength and ion balance. Electrical conductance is
the minimum requirement, however total dissolved solids,
hardness, alkalinity and major ions should ideally be
included. Future measurement of color should be added

58

N.Z. Al-Mutairi / Bioresource Technology 100 (2009) 5058

to the analyses because toxicity testing can be confused or

complicated by highly colored euents.
Acknowledgement
This research was funded by Grant No. EV07/00 from
the research administration of Kuwait University.
References
Al-Mutairi, N., 2006. Coagulant toxicity and eectiveness in a slaughterhouse wastewater treatment plant. Ecotoxicol. Environ. Safety 65 (1),
7483.
Al-Mutairi, N., 2007. Functional biodiversity of microbial communities in
aerobic selector slaughterhouse wastewater. Water Environ. Res. 79
(6), 660666.
Al-Mutairi, N., Hamoda, M., Al-Ghusain, I., 2003. Performance-based
characterization of a contact stabilization process for slaughterhouse
wastewater. J. Environ. Sci. Health, Part A 38 (10), 22872300.
Al-Mutairi, N., Hamoda, M., Al-Ghusain, I., 2004. Coagulant selection
and sludge conditioning in a slaughterhouse wastewater treatment
plant. Biores. Technol. 95 (2), 115119.
Andreasen, K., Agertved, J., Petersen, J., Skaarup, H., 1999. Improvement
of sludge settleability in activated sludge plants treating euent from
pulp and paper industries. Water Sci. Technol. 40 (1112), 215222.
Bitton, G., 1999. Wastewater Microbiology, second ed. Wiley-Liss,
Canada.
Eckenfelder, W., Musterman, J., 1995. Activated Sludge Treatment of
Industrial Wastewater. Technomic Publishing Company Inc.,
England.
Eikelboom, D., 1994. The Microthrix parvicella puzzle. Water Sci.
Technol. 29 (7), 273279.
Gerardimh, M., Horsfall, F., Barron, T., Baumgardner, M., Boyt, L.,
1990. Wastewater Biology: The Microlife. The Water Pollution
Control Federation, USA.
Horan, N., 1990. Biological Wastewater Treatment Systems: Theory and
Operation. Wiley, England.
Hwang, Y., Tanaka, T., 1998. Control of Microthrix parvicella foaming in
activated sludge. Water Res. 32 (5), 16781686.
Jenkins, D., Richard, M., Daigger, G., 1993. Manual on the Causes and
Control of Activated Sludge Bulking and Foaming, second ed. Lewis
Publishers Inc., Chelsea, MI.

Knoop, S., Kunst, S., 1998. Inuence of temperature and sludge loading
on activated sludge settling, especially on Microthrix parvicella. Water
Sci. Technol. 37, 2735.
Kruit, J., Boley, F., Jacobs, L., Wouda, T., 1993. Prediction of O2
conditions in the selector. In: Proceedings of the 1st International
Conference on Microorganisms in Activated Sludge and Biolm
Processes, 2728 September.
Madoni, P., Davoli, D., 1997. Testing the control of lamentous
microorganisms responsible for foaming in a full-scale activatedsludge plant running with initial aerobic or anoxic contact zones.
Biores. Technol. 60 (1), 4349.
Marshall, R., Richard, M., 1999. Selectors in pulp and papermill activated
sludge operations do they work? In: Tappi International Environmental Conference, pp. 1125.
Microbics, 1992. Microtox Procedures Manual. Microbics Corporation,
Carlsbad, CA.
Mueller, J., Kim, Y., Krupa, J., Shkreli, F., Nasr, S., Fitzpatrick, B., 2000.
Full-scale demonstration of improvement in aeration eciency. J.
Environ. Eng. 126 (6), 549555.
Nakhla, G., Lugowski, A., 2003. Control of lamentous organisms in
food-processing wastewater treatment by intermittent aeration and
selectors. J. Chem. Technol. Biotechnol. 78 (4), 420430.
Pujol, R., Canler, J., 1994. Contact zone: French practice with low F/M
bulking control. Water Sci. Technol. 29 (7), 221228.
Richard, M., 1989. Activated Sludge Microbiology. The Water Pollution
Control Federation, USA.
Richard, M., Kampfer, P., 1997. Detection and cultivation of lamentous bacteria from activated sludge. FEMS Microbiol. Ecol. 23,
169181.
Sperandio, M., Urbain, V., Ginestet, P., Audic, M., Paul, E., 2000.
Application of COD fractionation by a new combined technique:
comparison of various wastewaters and sources of variability. Water
Sci. Technol. 43 (1), 181190.
Standard Methods for the Examination of Water and Wastewater, 1995.
American Public Health Association/American Water Works Association/Water Environment Federation, 19th ed., Washington, DC,
USA.
Still, D., Ekama, G., Wentzel, M., Casey, T., Marais, G., 1996.
Filamentous organism bulking in nutrient removal activated sludge
systems, Paper 2: stimulation of the selector eect under aerobic
conditions. Water SA 22 (2), 97118.
Vaiopoulou, E., Melidis, P., Aivasidis, A., 2007. An activated sludge
treatment plant for integrated removal of carbon, nitrogen and
phosphorus. Desalination 211 (13), 192199.
TM