North American Journal of Aquaculture 66:198–207, 2004

q Copyright by the American Fisheries Society 2004

Application of Chemical Coagulation Aids for the Removal of

Suspended Solids (TSS) and Phosphorus from the Microscreen
Effluent Discharge of an Intensive Recirculating
Aquaculture System
JAMES M. EBELING* AND SARAH R. OGDEN
The Conservation Fund Freshwater Institute, Shepherdstown, West Virginia 25443, USA

PHILIP L. SIBRELL
U.S. Geological Survey, Leetown Science Center,
Kearneysville, West Virginia 25430, USA

KATA L. RISHEL
The Conservation Fund Freshwater Institute,
Shepherdstown, West Virginia 25443, USA

Abstract.—An evaluation of two commonly used coagulation–flocculation aids (alum and ferric
chloride) was conducted to determine optimum conditions for treating the backwash effluent from
microscreen filters in an intensive recirculating aquaculture system. Tests were carried out to
evaluate the dosages and conditions (mixing and flocculation stirring speeds, durations, and settling
times) required to achieve optimum waste capture. The orthophosphate removal efficiency for
alum and ferric chloride were greater than 90% at a dosage of 60 mg/L. Optimum turbidity removal
was achieved with a 60-mg/L dosage for both alum and ferric chloride. Both alum and ferric
chloride demonstrated excellent removal of suspended solids from initial total suspended solid
values of approximately 320 mg/L to approximately 10 mg/L at a dosage of 60 mg/L. Flocculation
and mixing speed and duration played only a minor role in the removal efficiencies for both
orthophosphates and suspended solids. Both coagulation–flocculation aids also exhibited excellent
settling characteristics, with the majority of the floc quickly settling out in the first 5 min.

Because of its potential negative impact on
freshwater systems, phosphorus is one of the most
scrutinized nutrients discharged by aquaculture
systems. Phosphorus is often the limiting nutrient
in natural ecosystems, and excessive algae blooms
can occur if discharge concentrations exceed the
absorption capacity of the receiving body of water.
Both state and federal regulatory agencies have
imposed limitations on phosphorus discharge to
high-quality receiving waters. Because of these
limitations and the potential impact of phosphorus
on the environment, numerous research projects
have been conducted over the last few years on
ways to reduce phosphorus in discharges from
aquaculture systems. Much of the work in this area
has focused on decreasing the total phosphorus in
the feed or increasing dietary phosphorus avail-
ability. Recirculating aquaculture systems, how-
ever, provide excellent opportunities for phospho-
rus control. Because recirculating systems provide
a small but concentrated waste stream, they pro-
* Corresponding author: j.ebeling@freshwaterinstitute.org
Received October 30, 2003; accepted February 14, 2004
vide opportunities for more efficient and econom-
ical phosphorus control. In contrast, in systems
such as raceways, the equivalent treatment of the
dilute effluent flow stream would be extremely dif-
ficult both from an engineering and an economic
standpoint.
Microscreen filters have become very popular
for suspended solids removal because they require
minimal labor and floor space and can treat large
flow rates of water with very little head loss.
Screen filters remove solids by virtue of physical
restrictions (or straining) on a media when the
mesh size of the screen is smaller than the particles
in the wastewater. Microscreen filters, though,
generate a separate solids waste stream that must
be further processed before final discharge. The
backwash flow will vary in volume, and the solids
content will vary based on several factors (e.g.,
screen opening size, type of backwash control em-
ployed, frequency of backwash, and influent total
suspended solid [TSS] load on the filter). Back-
wash flow is generally expressed as a percentage
of the flow the filter treats, with reported backwash
flows ranging from 0.2% to 1.5% of the treated
flow (Ebeling and Summerfelt 2002).

198
 COAGULATION AIDS REMOVAL OF TSS AND PHOSPHORUS
Several chemical and biological processes have
been investigated for the removal of phosphorus
from aquaculture effluent water. Adler and Sibrell
(2003) investigated the use of neutralized acid
mine drainage to reduce the loss of soluble P from
agricultural fields and animal wastewater. A bio-
logical means of phosphate removal by Barak and
van Rijn (2000) demonstrated that some denitri-
fiers were capable of phosphate uptake in excess
of their metabolic requirements. Kioussis et al.
(1999) developed a polymeric hydrogel, which de-
creased phosphorus in aquaculture wastewater ef-
fluents by more than 99% to less than 0.01 ppm.
However, it has been determined that the majority
of the phosphorus discharged from intensive aqua-
culture systems (50–85%) is contained in the fil-
terable or settleable solids fraction (Bergheim et
al. 1993; Heinen et al. 1996). Thus, any mecha-
nism that could enhance solids removal would also
contribute to a reduction in the overall level of
phosphorus discharge. Coagulation and floccula-
tion processes, with aids such as alum or ferric
chloride, are standard techniques in the wastewater
and drinking water industry for the removal of
suspended solids. These aids have not been exten-
sively applied in the aquaculture industry primar-
ily because of the dilute nature of most aquaculture
waste streams. However, the concentrated waste
stream from recirculating systems, especially the
backwash from microscreen filters, makes this op-
tion feasible from both an engineering and an eco-
nomic standpoint.
Over the past several years, the Conservation
Fund’s Freshwater Institute in Shepherdstown,
West Virginia, has demonstrated several technol-
ogies and strategies to manage and reduce the
wastes generated during aquaculture production,
including improved feed and feeding strategies
(Tsukuda et al. 2000), technologies to minimize
water use and concentrate waste streams (Tim-
mons and Summerfelt 1997; Summerfelt et al.
2000), and overall waste management and treat-
ment reviews (Summerfelt 1998; Summerfelt et al.
1999; Ebeling and Summerfelt 2002). In related
research, several coagulation aids were evaluated
for the physical–chemical treatment of the super-
natant discharge from manure-thickening tanks
(Ebeling et al. 2003). Future research at the Fresh-
water Institute will include the application of mi-
crofiltration technology using semipermeable
membranes and the construction of a demonstra-
tion pilot-scale compost facility.
The objectives of this research were to evaluate
the physical–chemical treatment of a recirculating
199
system’s microscreen backwash effluent using sev-
eral common coagulation–flocculation aids em-
ployed in the drinking and wastewater treatment
industry (i.e., alum and ferric chloride). In addition
to determining their effectiveness in removing
both suspended solids and phosphorus, a system-
atic evaluation of the variables encountered in the
coagulation–flocculation process (mixing and floc-
culation stirring speeds and durations, and settling
times) was conducted.
Background
One of the most commonly used methods for
the removal of suspended solids in drinking water
is the addition of coagulant and flocculation aids,
such as alum, ferric chloride, and long-chain poly-
mers (AWWA 1997). Coagulation is the process
of decreasing or neutralizing the electric charge
on suspended particles or zeta potential. Similar
electric charges on small particles in water cause
the particles to naturally repel each other and hold
the small, colloidal particles apart and keep them
in suspension. The coagulation–flocculation pro-
cess neutralizes or reduces the negative charge on
the particles. This allows the van der Waals force
of attraction to encourage the initial aggregation
of colloidal and fine suspended materials to form
microfloc. Flocculation is the process of bringing
together the microfloc particles to form large ag-
glomerations by physically mixing them or
through the binding action of flocculants (such as
long-chain polymers). A classical coagulation–
flocculation unit process (Metcalf and Eddy, Inc.
1991) consists of three separate steps:
(1) Rapid or flash mixing—the suitable chemicals
(coagulants, flocculants, and, if required, pH
adjusters) are added to the wastewater stream,
which is stirred and intensively mixed at high
speed;
(2) Slow mixing (coagulation and flocculation)—
the wastewater is only moderately stirred in
order to form large flocs, which are easily set-
tled out; and
(3) Sedimentation—the floc formed during floc-
culation is allowed to settle out and is sepa-
rated from the effluent stream.
Numerous substances have been used as coag-
ulant and flocculation aids, including alum (Al2
[SO4]3 · 18H2O), ferric chloride (FeCl3 · 6H2O), ferric
sulfate (Fe2[SO4]3), ferrous sulfate (FeSO4 · 7H2O),
and lime (Ca[OH]2) (Metcalf and Eddy, Inc. 1991).
Aluminum sulfate (alum) is the most commonly
200 EBELING ET AL.
used coagulant, is easy to handle and apply, and
produces less sludge than lime. Its primary dis-
advantage is that it is most effective over a limited
pH range of 6.5–7.5. Ferric chloride is also a com-
monly used coagulant and is effective over a wider
pH range of 4–11. The ferric hydroxide floc is also
heavier than the alum floc, improving its settling
characteristics and reducing the size of the clari-
fier. Neither ferric sulfate nor ferrous sulfate is
commonly used today, but ferric sulfate is slowly
replacing ferric chloride because it is easier to
store and handle. Lime is commonly used and is
effective, but is quite pH-dependent and produces
a large quantity of sludge requiring disposal.
When alum is added to a wastewater, the fol-
lowing reaction takes place:
A12 (SO4 )3 · 18 H2O 1 3 Ca(HCO3 )2 ⇔ 3 CaSO4
1 2 Al(OH)3 1 6 CO2 1 18 H2O.
The insoluble aluminum hydroxide, Al(OH)3, is a
gelatinous floc that settles slowly through the
wastewater, sweeping out the suspended material.
Alkalinity is required for the reaction and, if not
available, must be added at the rate of 0.45 mg/L
as CaCO3 for every 1 mg/L alum.
Similarly, for ferric chloride:
2 FeCl3 · 6 H2O 1 3 Ca(HCO3 )2 ⇔ 3 CaCl2
1 2 Fe(OH)3 1 6 CO2 1 12 H2O.
The insoluble ferric hydroxide, Fe(OH)3, is also a
gelatinous floc that settles through the wastewater,
sweeping out the suspended material. Alkalinity is
required for the reaction and, if not available, must
be added at the rate of 0.55 mg/L CaCO3 for every
1 mg/L ferric chloride.
In addition, both aluminum and iron salts can
also be used for the chemical precipitation of phos-
phorus. The basic reactions involved are
Al13 1 PO24 3 ⇔ AlPO4 and
Fi13 1 PO24 3 ⇔ FePO4 .
The above equations are the simplest forms of the
reaction (Metcalf and Eddy, Inc. 1991; Lee and
Lin 2000). Due to the many other competing re-
actions, the effects of alkalinity, pH, trace ele-
ments, and other compounds in the wastewater, the
actual chemical dosage required to remove a given
quantity of phosphorus is usually established on
the basis of a bench-scale test or sometimes pilot-
scale tests.
Mixing disperses precipitating agents, coagu-
lants, and coagulant aids throughout the waste-
water to ensure the most rapid precipitation re-
actions and subsequent settling of precipitates pos-
sible. For engineering design, the degree of mixing
is dependent, in turn, upon the amount of energy
supplied, the mixing residence time, and the re-
lated turbulence effect (which depends on the size
and shape of the mixing tank). Mixing can be sub-
divided into two types: (1) flash or rapid mixing
of chemicals and (2) continuous mixing in reactors
or holding tanks. Both are employed during the
coagulation–flocculation process. In flash or rapid
mixing, the principal objective is to provide com-
plete and uniform dispersion of a chemical added
to the water (i.e., alum in wastewater). In contin-
uous mixing, the principal objective is to maintain
the contents of a reactor in a completely mixed
state that promotes the aggregation of particles
(e.g., flocculation). Types of mixing include pro-
peller, turbine, paddle, pneumatic, and hydraulic
mixers. The power input per unit volume of liquid
can be used as a rough measure of mixing effec-
tiveness (Metcalf and Eddy, Inc. 1991). The fol-
lowing equation is normally used to estimate the
velocity gradients in coagulation tanks and is used
for the design and operation of mixing systems
(Metcalf and Eddy, Inc. 1991):
G 5 (P/mV )1/2 ,
where
G 5 mean velocity gradient (s21 ),
P 5 power applied (W or ft lb/s),
m 5 dynamic viscosity (N · s/m 2 or lb · s/ft 2 ), and
V 5 tank volume (m 3 or ft 3 ).
The actual values for G are usually provided by
the equipment manufacturers, although they can
be estimated from the area of the paddle, fluid
density, and paddle velocity in the fluid. Rapid or
flash mixing residence times typically range from
30 s to 2 min, with 1 min being the most common.
The intensity and duration of the mixing of the
coagulants must be controlled to prevent the break-
up the microfloc or the uneven dosing of coagulant.
The typical range of velocity gradient values for
rapid mixing is 250–1,500/s (Metcalf and Eddy,
Inc. 1991). Flocculation mixing is much slower to
allow the maximum interaction of the floc to ag-
gregate together. Typical values of the velocity
gradient for flocculation range from 20 to 80/s,
with a retention time of 20–30 min (Lee and Lin
2000). In our previous coagulation/flocculation
testing on the supernatant discharge from fish ma-
 COAGULATION AIDS REMOVAL OF TSS AND PHOSPHORUS 201

TABLE 1.—Water quality characteristics of the micro- Jar tests.—For over 50 years, the jar test has
screen backwash effluent in nine samples. Abbreviations been the standard technique used to optimize the
are as follows: NTU 5 nephelometric turbidity units; TP
addition of coagulants and flocculants used in the
5 total phosphorus; SRP 5 soluble reactive phosphorus;
TSS 5 total suspended solids; TVS 5 total volatile solids; wastewater and drinking water treatment industry.
TN 5 total nitrogen; TAN 5 total ammonia nitrogen; and The Standard Practice for Coagulation–Floccula-
cBOD5, 5 carbonaceous biochemical oxygen demand. tion Jar Test of Water, ASTM-D-2035, was first
approved in 1980 and reapproved in 1999 (ASTM
Parameter Mean SD Range
1995). The scope of this practice ‘‘covers a general
pH 7.43 0.26 6.97–7.78 procedure for the evaluation of a treatment to re-
Temperature (8C) 19.4 1.4 18–21
Alkalinity (mg/L) 292 21 260–324
duce dissolved, suspended, colloidal, and nonset-
Turbidity (NTU) Over range tleable matter from water by chemical coagula-
TP (mg/L) 12.5 6.3 5.8–21 tion–flocculation, followed by gravity settling.’’
SRP (mg/L) 4.0 1.9 2–7.2 This standard was utilized to provide a technique
TSS (mg/L) 1015 401 517–1540
TVS (mg/L) 753 206 489–1100 to systematically evaluate the variables normally
TN (mg/L) 77.8 89.6 8–236 encountered in the coagulation–flocculation pro-
TAN (mg/L) 14.8 24.5 3.4–92 cess.
NO2 (mg N/L) 0.43 0.34 0.23–1.36
NO3 (mg N/L) 38.8 9.2 25.5–48.6 Since coagulant interactions are very complex,
cBOD5 (mg/L) 548 190 281–947 laboratory studies are used to determine the op-
timal dosage, duration, and intensity of mixing and
flocculation. The coagulation–flocculation process
nure thickening tanks, optimum conditions were a consists of three distinct steps. First, the coagulant
mixing speed of 165/s for 1 min, a 14/s flocculation is added to the effluent water and a rapid and high
speed for 10 min, and a settling time of 10–20 intensity mixing is initiated. The objective is to
min. obtain a complete mixing of the coagulant with
the wastewater to maximize the effectiveness of
Methods destabilization of colloidal particles and initiate
The coagulation–flocculation tests were carried coagulation. Critical parameters for this step are
out following the standard practice for coagula- the duration and the paddle speed or mixing in-
tion–flocculation testing of wastewater to evaluate tensity (velocity gradient, G). Second, the suspen-
the chemicals, dosages, and conditions required to sion is slowly stirred to increase contact between
achieve optimum results (ASTM 1995). Jar tests coagulating particles and to facilitate the devel-
provide insight into the overall process effective- opment of large flocs. Again, the flocculation du-
ness, particularly with regard to mixing intensity ration and intensity are critical parameters (e.g.,
and duration as it affects floc size and density (Lee too high an intensity can break up the aggregate
and Lin 1999). Samples for jar tests were taken floc). Third, mixing is terminated and the floc is
directly from the holding tanks receiving the back- allowed to settle. The velocity gradient is a mea-
wash water from several rotating microscreen fil- sure of the mixing energy and allows an engineer
ters used for suspended solids removal in two, to scale the test results to proportionally larger
commercial-size recirculating production systems system sizes.
growing Arctic char Salvelinus alpinus. The first A standard jar test apparatus, the Phipps and
of these is a pilot-scale, partial-reuse system con- Bird six-paddle stirrer with illuminated base (Fig-
sisting of three 3.66-m-diameter 3 1.1-m-deep cir- ure 1), was employed for the tests, with six 2-L
cular ‘‘Cornell-type’’ dual-drain culture tanks with square B-Ker2 Plexiglas jars, sometimes called ga-
a maximum feed loading rate of 45–50 kg feed/d tor jars. The jars are provided with a sampling port,
(Summerfelt et al. 2002). The second system is a 10 cm below the water line, which allows for re-
fully recirculating system consisting of a 150-m 3 petitive sampling with minimal impact on the test.
circular production tank with a maximum daily This type of jar has several advantages over the
feed rate of 200 kg feed/d (Summerfelt et al. 2003). more traditional 1-L circular jars, including a larg-
Water quality characteristics of the microscreen er volume for reduced errors in mixing and a larger
backwash effluent are summarized in Table 1. Be- volume of supernatant for analysis. In addition,
cause of the excess alkalinity of the water at this the square walls reduce water rotation, making baf-
location (i.e., approximately 260 mg/L as CaCO 3), fles or stators unnecessary. Finally, the thick Plex-
no alkalinity additions were required in conjunc- iglas walls offer sufficient thermal insulation to
tion with alum and ferric chloride treatments. minimize temperature changes during the testing
202 EBELING ET AL.
FIGURE 1.—Phipps and Bird six-paddle stirrer with illuminated base.
period. The six flat paddles are all driven by a
single variable speed motor from 0 to 300 revo-
lutions per minute (rpm). Velocity gradient curves
(G; s21) versus agitator paddle speed (rpm) are
provided by the manufacturer. An illuminated base
helps in the observation of the floc formation and
settling characteristics.
Stock solutions of the coagulants and flocculants
were used to improve the ease of handling and
measuring and to ensure good mixing in the jars.
Simple dilutions of alum and ferric chloride with
distilled water to a 0.2% solution by weight were
employed. Normally, the actual test procedures are
representative of an existing treatment system (for
example, a wastewater treatment plant’s mixing,
flocculation, and settling tanks) in terms of the
duration of mixing and flocculation and the ve-
locity gradients, as well as the settling time em-
ployed. In contrast, in this engineering design
study (and in future work), a wide range of chem-
icals, dosages, and conditions (mixing and floc-
culation stirring speeds and durations, and settling
times) were examined to achieve optimum removal
of suspended solids and phosphorus. Based on
these tests, recommendations can then be made as
to the engineering design and operation of both a
pilot-scale study and large-scale treatment sys-
tems.
For each jar test, the following procedure was
followed (ASTM 1995). Each jar was filled with
2 L of sample measured with a graduated cylinder,
and the initial temperature was recorded. The co-
agulant or flocculant dose destined for each jar was
carefully measured into 150-mL beakers, and dis-
tilled water was added to yield equal volumes in
all the beakers. The multiple stirrer speed was set
to the ‘‘flash mix’’ values of 50, 75, 150, and 225
rpm (corresponding to G values of 42, 70, 165,
and 272/s, respectively), and the test solutions
were added. After the predetermined flash mix du-
ration, the mixing speed was reduced to the floc-
culation or ‘‘slow mix’’ values of 10, 20, 30, or
40 rpm (corresponding to G values of 6, 14, 22,
and 32/s, respectively) for a specified duration.
After this time period, the paddles were withdrawn
and the floc allowed to settle for specified time (0,
5, 10, 15, 20, 30, or 45 min). Samples were then
withdrawn from the sampling ports located 10 cm
below the water level for analysis. A baseline set
of mixing and flocculation speeds and durations
were used for comparison purposes (150 rpm for
1 min mixing, 20 rpm for 20 min flocculation, and
a settling time of 30 min).
Analysis.—For all of the jar tests, pH, turbidity,
and soluble reactive phosphorus (SRP, orthophos-
phate) were measured. For the purpose of com-
paring the effect of various operating parameters
such as mixing and flocculation speed, turbidity
was used as an indicator of suspended solids and
orthophosphate for phosphorus content. Table 2
 COAGULATION AIDS REMOVAL OF TSS AND PHOSPHORUS 203

TABLE 2.—Laboratory methods used for analysis via a Hach DR/2010 colorimeter.

Parameter Method; range

Alkalinitya Standard Methods 2320B
Reactive phosphorusb Hach Method 8048 (orthophosphate; 0–2.50 mg PO423/L expressed as P
Total suspended solidsa Standard Methods 2540D
Turbiditya Hach Model 2100P turbidimeter; nephelometric turbidity units
a Adapted from Standard Methods for the Examination of Water and Wastewater (APHA 1989).
b Approved for reporting by the U.S. Environmental Protection Agency.

shows the methods used for each analysis. When
appropriate, reagent standards and blanks were an-
alyzed along with the samples to ensure quality
control. The comparison of alum or ferric chloride
treatments were statistically analyzed using a sin-
gle-factor analysis of variance (ANOVA) test
(with triplicate samples at each dosage and at a
5 0.01). For tests on the effect of mixing and
flocculation variables using alum, the results at
each dosage level were lumped, averaged, and then
tested using ANOVA for significant differences.
At concentrations above 90 mg/L, orthophosphate
concentrations were reduced to about 0.12 mg P/
L for both alum and ferric chloride. At concentra-
tions of 60 mg/L, the turbidity approached its low-
est value. The results show that both coagulants
had similar effects despite the different chemicals
involved. A comparison of the molecular weights
of these compounds indicates that the ferric chlo-
ride should be slightly more effective on a weight
basis given its lower formula weight. However,
this difference was not observed in these tests.

Results and Discussion Effect of Mixing Intensity

Comparison of Alum and Ferric Chloride
The two coagulant aids, alum and ferric chlo-
ride, were tested. Figure 2 shows a comparison of
the effectiveness of alum and ferric chloride in
removing orthophosphate from the microscreen
backwash effluent as compared with the control of
settling only. The turbidity data showed a similar
relationship (Figure 3). For orthophosphate per-
cent removal and normalized turbidity remaining,
no significant difference was seen between the
alum and ferric chloride (P 5 0.01) using a single-
factor ANOVA test of triplicate samples at each
dosage. Also no significant difference (P 5 0.01)
was seen in percent phosphorus removal for both
coagulant aids above a concentration of 60 mg/L.
Since there was no practical difference between
the two coagulation agents, a series of tests was
conducted with alum to examine the effect of the
initial mixing intensity (paddle speed or velocity
gradient) and the flocculation intensity on the re-
moval efficiency of suspended solids and ortho-
phosphate. The results of these jar tests are shown
in Figures 4 and 5 over a range of alum dosages
from 0 to 120 mg/L and for mixing speeds of 50,
75, 150, and 225 rpm. As can be seen in Table 1,
the effluent varied slightly from day to day due to
changes in, for example, feed addition, biomass,
and makeup water addition. Therefore, the results
in Figure 4 are expressed in terms of percent re-

FIGURE 3.—Comparison of alum and ferric chloride

FIGURE 2.—Comparison of alum and ferric chloride in terms of normalized turbidity (nephelometric turbidity
in terms of the percent removal of orthophosphate under units [NTU]) under standard conditions (mixing at 150
standard conditions (mixing at 150 rpm for 1 min, floc- rpm for 1 min, flocculation at 20 rpm for 20 min, and
culation at 20 rpm for 20 min, and settling for 30 min). settling for 30 min).
204 EBELING ET AL.
FIGURE 4.—Effect of mixing intensity (rpm) and alum
dosage on the percent removal of orthophosphate using
the standard jar test with mixing for 1 min, flocculation
at 20 rpm for 20 min, and settling for 30 min.
moved, and turbidity in Figure 5 has been nor-
malized to minimize the effect of different initial
starting concentrations of orthophosphate and tur-
bidities (although normally very small).
For the removal of orthophosphorus, a signifi-
cant difference was seen between the dosage levels
(P 5 0.01) between each treatment level except
for the 90 and 120 mg/L dosage. For the normal-
ized turbidity data, no significant difference was
seen between the treatment levels for dosages of
alum above 40 mg/L (P 5 0.01). As can be seen
in Figure 5, for alum dosages above 60 mg/L, there
is a practical effect of mixing speed on turbidity.
The lower mixing speed may improve the removal
of turbidity at high concentrations due to reduced
shearing of the floc during initial formation.
Effect of Flocculation Intensity
Figures 6 and 7 shows the effect of flocculation
speed on the removal of orthophosphorus and nor-
FIGURE 5.—Effect of mixing intensity (rpm) and alum
dosage on normalized turbidity (nephelometric turbidity
units [NTU]) using the standard jar test with mixing for
1 min, flocculation at 20 rpm for 20 min, and settling
for 30 min.
FIGURE 6.—Effect of flocculation intensity (rpm) and
alum dosage on the percent removal of orthophosphate
using the standard jar test with mixing at 150 rpm for
1 min, flocculation for 20 min, and settling for 30 min.
malized turbidity as a function of alum dosage.
Similar to previous results, no significant differ-
ence was seen between dosage levels in the re-
moval of phosphorus (P 5 0.01) above a 90-mg/
L alum dosage. Although statistically significant
differences were found at the lower dosages, these
were both small in magnitude and outside the rec-
ommended operational range of dosages. For the
normalized turbidity data, no significant difference
was seen between any of the dosage levels. The
lower mixing speed may improve the removal of
turbidity at lower concentrations of alum due to
reduced shearing during the flocculation phase.
Effect of Flocculation Time
Figures 8 and 9 show the effect of flocculation
mixing time on the orthophosphate concentration
and turbidity and total suspended solids concen-
tration. Based on the previous results, a minimum
alum dosage of 60 mg/L was chosen as a good
operating point. Flocculation mixing times of 0,
FIGURE 7.—Effect of flocculation intensity (rpm) and
alum dosage on normalized turbidity (nephelometric tur-
bidity units [NTU]) removal using the standard jar test
with mixing at 150 rpm for 1 min, flocculation for 20
min, and settling for 30 min.
 COAGULATION AIDS REMOVAL OF TSS AND PHOSPHORUS
FIGURE 8.—Effect of flocculation time on phosphorus
concentration using the standard jar test with mixing at
150 rpm for 1 min, flocculation at 20 rpm, and settling
for 30 min at an alum concentration of 60 mg/L.
5, 10, 15, 20, and 25 min were used. As can be
seen from the bar charts in Figures 8 and 9, after
the first 20 min, there was very little change in the
concentration of orthophosphate, and after 20 min
both turbidity and TSS concentration are at their
minimum values. Total suspended solids concen-
tration data were included here to demonstrate the
close relationship between turbidity and TSS for
this particular effluent at these relatively low con-
centrations.
Effect of Settling Time
The effect of settling time on removal efficien-
cies was also examined. A standardized set of mix-
ing and flocculation speeds and durations were
used for comparison purposes: 150 rpm for 1 min
mixing; 20 rpm for 20 min flocculation; and a set-
tling time of 5, 10, 15, 20, and 30 min at an alum
dosage of 60 mg/L. Figures 10 and 11 show the
FIGURE 9.—Effect of flocculation time on the turbidity
(nephelometric turbidity units [NTU]) and total sus-
pended solid (TSS) concentration using the standard jar
test with mixing at 150 rpm for 1 min, flocculation at
20 rpm, and settling for 30 min at an alum concentration
of 60 mg/L.
205
FIGURE 10.—Effect of settling time on the ortho-
phosphate concentration using the standard jar test with
mixing at 150 rpm for 1 min and flocculation at 20 rpm
for 20 min and settling at an alum concentration of 60
mg/L.
effect of settling time on the removal of soluble
reactive phosphorus and on turbidity and TSS with
alum. As both graphs show, the floc quickly settles
out within the first 5 min, with little change in final
values after 10 min.
Conclusions
An evaluation of two commonly used coagu-
lation–flocculation aids (alum and ferric chloride)
was conducted to determine the optimum condi-
tions for treating the backwash effluent from mi-
croscreen filters in an intensive recirculating aqua-
culture system using standard jar test procedures.
The backwash effluent contains very high concen-
trations of total suspended solids (ranging from
517 to 1,540 mg/L) due to the nature of the mi-
croscreen operation. The orthophosphate removal
efficiencies for alum and ferric chloride were
greater than 90% at a dosage of 60 mg/L, with
final concentrations of SRP approaching 0.15 mg
FIGURE 11.—Effect of settling time on turbidity (neph-
elometric turbidity units [NTU]) and total suspended
solid (TSS) concentration using the standard jar test with
mixing at 150 rpm for 1 min, flocculation at 20 rpm for
20 min, and settling at an alum concentration of 60 mg/L.
206 EBELING ET AL.
P/L. Optimal turbidity removal was achieved with
a 60 mg/L dosage for both alum and ferric chlo-
ride. Both alum and ferric chloride demonstrated
an excellent removal of suspended solids from ini-
tial TSS values of approximately 320 mg/L to ap-
proximately 10 mg/L at a dosage of 60 mg/L. Floc-
culation and mixing intensity and duration played
only a minor role in the removal efficiencies for
both orthophosphates and suspended solids. Both
coagulation–flocculation aids also exhibit excel-
lent settling characteristics, with the majority of
the floc quickly settling out in the first 5 min.
It is important that mixing and flocculation
speeds and times be determined by jar test for each
effluent, but good results were obtained for this
aquaculture effluent with a mixing speed of 150
rpm (G 5 165/s) for 1 min, 20 rpm (G 5 14/s)
flocculation speed for 20 min, and a settling time
of 10–20 min. Based on these initial studies, for
maximum orthophosphate removal the dosages of
both alum and ferric chloride applied should be
greater than 60 mg/L. For maximum turbidity re-
moval, a dosage of 60 mg/L or more of both alum
and ferric chloride is required. Future studies will
include additional tests with varying mixing, floc-
culation, and settling times at various dosages for
other potential coagulation–flocculation aids, in-
cluding a screening of potential polymer additions.
Acknowledgments
This work was supported by the U.S. Depart-
ment of Agriculture, Agricultural Research Ser-
vice, under Cooperative Agreement number 59-
1930-1-130. The experimental protocol and meth-
ods used in this study were in compliance with
Animal Welfare Act requirements and were ap-
proved by the Freshwater Institute Institutional
Animal Care and Use Committee.
References
Adler, P. R., and P. L. Sibrell. 2003. Sequestration of
phosphorus by acid mine drainage floc. Journal of
Environmental Quality 32:1122–1129.
APHA (American Public Health Association). 1989.
Standard methods for the examination of water and
wastewater, 17th edition. American Public Health
Association, American Water Works Association,
and Water Pollution and Control Federation, Wash-
ington, D.C.
ASTM (American Society for Testing and Materials). 1995.
Standard practice for coagulation2flocculation jar test
of water E1-1994 R(1995), D 2035-80. Pages in An-
nual book of ASTM standards, volume 11.02.
AWWA (American Water Works Association). 1997.
Water treatment plant design. American Water
Works Association and American Society of Civil
Engineers, 3rd edition. McGraw-Hill, New York.
Barak, Y., and J. van Rijn. 2000. Biological phosphate
removal in a prototype recirculating aquaculture
treatment system. Aquaculture Engineering 22:121–
136.
Bergheim, A., R. Kristiansen, and L. Kelly. 1993. Treat-
ment and utilization of sludge from land-based
farms for salmon. Pages 486–495 in J.-K. Wang,
editor. Techniques for modern aquaculture. ASAE,
St. Joseph, Michigan.
Ebeling, J. M., P. L. Sibrell, S. Ogden, and S. T. Sum-
merfelt. 2003. Evaluation of chemical coagula-
tion2flocculation aids for the removal of phospho-
rus from recirculating aquaculture effluent. Aqua-
culture Engineering 29:23–42.
Ebeling, J. M., and T. S. Summerfelt. 2002. Performance
evaluation of a full-scale intensive recirculating
aquaculture system’s waste discharge treatment sys-
tem. Pages 506–515 in T. T. Rakestraw, L. S. Doug-
las and G. J. Flick, editors. Proceedings of the
Fourth International Conference on Recirculating
Aquaculture. Virginia Polytechnic Institute and
State University, Blacksburg.
Heinen, J. M., J. A. Hankins, and P. R. Adler. 1996.
Water quality and waste production in a recirculat-
ing trout culture system with feeding of a higher-
energy or lower-energy diet. Aquaculture Research
27:699–710.
Kioussis, D. R., F. W. Wheaton, and P. Kofinas. 1999.
Phosphate-binding polymeric hydrogels for aqua-
culture wastewater remediation. Aquaculture En-
gineering 19:163–178.
Lee, C. C., and S. D. Lin, 2000. Handbook of environ-
mental engineering calculations. McGraw-Hill,
New York.
Metcalf and Eddy, Inc., 1991. Wastewater engineering:
treatment, disposal, and reuse, 3rd edition.
McGraw-Hill, Boston.
Summerfelt, S. T. 1998. An integrated approach to aqua-
culture waste management in flowing-water tank
culture systems. Pages 87–97 in G. S. Libey and M.
B. Timmons, editors. Proceedings of the Second
International Conference on Recirculating Aqua-
culture. Virginia Polytechnic Institute and State
University, Blacksburg.
Summerfelt, S. T., P. R. Adler, D. M. Glenn, and R. N.
Kretschmann. 1999. Aquaculture sludge removal
and stabilization within created wetlands. Aqua-
cultural Engineering 19(2):81–92.
Summerfelt, S. T., J. Davidson, and M. B. Timmons.
2000. Hydrodynamics in the ‘‘Cornell-type’’ dual-
drain tank. Pages 160–166 in G. S. Libey, M. B.
Timmons, G. J. Flick, and T. T. Rakestraw, editors.
Proceedings of the Third International Conference
on Recirculating Aquaculture. Virginia Polytechnic
Institute and State University, Blacksburg.
Summerfelt, S. T., G. Wilton, D. Roberts, T. Rimmer,
and K. Fonkalsrud. 2003. Developments in recir-
culating systems for arctic char culture in North
America. Aquacultural Engineering 30:31–71.
Timmons, M. B., and S. T. Summerfelt. 1997. Advances
 COAGULATION AIDS REMOVAL OF TSS AND PHOSPHORUS 207

in circular culture tank engineering to enhance hy-
draulics, solids removal, and fish management. Pag-
es 66–84 in M. B. Timmons and T. Losordo, editors.
Recent advances in aquacultural engineering.
Northeast Regional Agricultural Engineering Ser-
vice, Ithaca, New York.
Tsukuda, S., R. Wallace, S. T. Summerfelt, and J. Han-
kins. 2000. Development of a third-generation
acoustic waste feed monitor. Pages 105–108 in G.
S. Libey, M. B. Timmons, G. J. Flick, and T. T.
Rakestraw, editors. Proceedings of the Third Inter-
national Conference on Recirculating Aquaculture.
Virginia Polytechnic Institute and State University,
Blacksburg.