ArticlePDF Available

Ozone Application in Recirculating Aquaculture System: An Overview

Authors:

Abstract and Figures

In recirculating aquaculture systems (RAS), particulates (including feces, uneaten feed, bacteria, and algae) can cause several problems, in that they may harbor pathogens, can physically irritate the fish, and upon decomposition, release ammonia and consume oxygen. Mechanical filters, foam fractionators, and other engineered devices are used to remove particles quickly from aquaculture systems, in order to improve fish health and decrease the load on biofilters and oxygenators. Ozone is used in RAS as a disinfectant, to remove organic carbon, and also to remove turbidity, algae, color, odor and taste. Ozone can effectively inactivate a range of bacterial, viral, fungal and protozoan fish pathogens. But the effectiveness of ozone treatment depends on ozone concentration, length of ozone exposure (contact time), pathogen loads and levels of organic matter. In spite of ozone is a very effective oxidizing agent, higher ozone concentrations are a risk to cultured fish stocks causing gross tissue damage and stock mortalities, and also are a risk to bacterial films on the biofilter.
Content may be subject to copyright.
Ozone: Science & Engineering, 33: 345–367
Copyright © 2011 International Ozone Association
ISSN: 0191-9512 print / 1547-6545 online
DOI: 10.1080/01919512.2011.604595
Ozone Application in Recirculating Aquaculture System:
An Overview
Alex Augusto Gonçalves,1and Graham A. Gagnon2
1Federal University of Semi Arid, Animal Science Department, Mossoró, RN, Brazil
2Center of Water Resources Studies, Dalhousie University, Halifax, NS, Canada B3J 1Z1
In recirculating aquaculture systems (RAS), particulates (includ-
ing feces, uneaten feed, bacteria, and algae) can cause several
problems, in that they may harbor pathogens, can physically irri-
tate the fish, and upon decomposition, release ammonia and consume
oxygen. Mechanical filters, foam fractionators, and other engineered
devices are used to remove particles quickly from aquaculture sys-
tems, in order to improve fish health and decrease the load on
biofilters and oxygenators. Ozone is used in RAS as a disinfectant,
to remove organic carbon, and also to remove turbidity, algae, color,
odor and taste. Ozone can effectively inactivate a range of bacte-
rial, viral, fungal and protozoan fish pathogens. But the effectiveness
of ozone treatment depends on ozone concentration, length of ozone
exposure (contact time), pathogen loads and levels of organic mat-
ter. In spite of ozone is a very effective oxidizing agent, higher ozone
concentrations are a risk to cultured fish stocks causing gross tissue
damage and stock mortalities, and also are a risk to bacterial films on
the biofilter.
Keywords Ozone, Aquaculture, Re-circulating systems,
Disinfection, Water quality, Toxicity
INTRODUCTION
Aquaculture will be a critical component of future seafood
production to supply the ever-expanding human population
and is continuing to expand worldwide, but such growth is
dependent on the availability of high-quality water. The use
of recirculating culture systems is one means of using avail-
able water more efficiently (Bullock et al., 1997; Kim, 2000;
Bai, 2007), and provides potential advantages over pond or
cage-based forms of aquaculture. These include flexibility
in site selection, reduced water usage, lower effluent vol-
umes, better environmental control, and higher intensity of
Received 11/1/2010; Accepted 4/12/2011
Address correspondence to Dr. Alex Augusto Gonçalves, Post
doctoral fellow at Dalhousie University, (Present address) Federal
University of Semi Arid (UFERSA), Animal Science Department,
Av. Francisco Mota, 572, Costa e Silva 59625-900, Mossoró, RN,
Brazil. E-mail: alaugo@gmail.com
production. However, as stock densities and levels of water
re-use increase, wastes accumulate rapidly and environmental
control becomes more difficult. Sophisticated systems capa-
ble of removing both particulate and dissolved organic wastes
become necessary (Read, 2008; Pfeiffer et al., 2008).
Despite the advantages, recirculating aquaculture systems
(RAS) pose a latent disease and public health risk. Part of the
biological filtration necessary for removal of harmful toxins
involves the biofilm that forms on all components of a recir-
culating system (King, 2001). Because the water is reused,
pathogens introduced into the system could remain through
incorporation into the biofilm, leading to recurring expo-
sure of fish to pathogens and the presence of asymptomatic
carriers.
Maintaining healthy fish in a recirculating system involves
establishing adequate dissolved oxygen levels, removal of
solid wastes, sufficient ammonia nitrification (King, 2001;
Ebeling and Timmons, 2009) and also for fish sensorial qual-
ity, i.e., removal off-flavors from MIB and geosmin from
water (Westerhoff et al., 2006). Increasing the daily water
exchange rate in an RAS will remove accumulated colloidal
solids, refractory organics and nitrite, to the detriment of water
budgets and the cost of heating or cooling the system. The
alternative method of removal is to break down these organic
wastes using an oxidizing agent, such as ozone (Tango and
Gagnon, 2003; Sharrer and Summerfelt, 2007; Read, 2008).
Ozone has been shown to be efficient in eliminating
most pathogens affecting seafood in freshwater and sea-
water aquaculture systems (salmon, halibut, tilapia, shrimp,
etc.). Moreover, ozone can improve water quality by reduc-
ing biochemical oxygen demand (BOD), ammonia and nitrite
(Hunter, 2000; Meunpol et al., 2003; Summerfelt, 2003;
Tango and Gagnon, 2003; Buchan et al., 2005; Coman et al.,
2005; Ritar et al., 2006; Sharrer and Summerfelt, 2007), to
disinfect incoming hatchery water (Tipping, 1987; Crisp and
Bland, 1990), to disinfect hatchery wastewater (Majumdar
and Sproul, 1974; Conrad et al., 1975) and to disinfect fish
Overview of Ozone in Recirculating Aquaculture System September–October 2011 345
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
eggs (Mimura et al., 1999; Grotmol and Totland, 2000; Ben-
Atia et al., 2001; Su et al., 2001; Grotmol et al., 2003).
Although direct exposure of aquatic organisms to ozone
and the oxidants formed in ozonated seawater can be lethal
(Wedemeyer et al., 1979a, 1979b; Coler and Asbury, 1980;
Fisher et al., 1999); but fertilized fish eggs can tolerate vary-
ing levels of dissolved ozone. Specific exposure levels need to
be determined for each species (Grotmol et al., 2003), and
reliable methods to measure ozone in sea water are there-
fore needed to ensure that lethal limits are not exceeded.
According to this information, this review intends to provide
information of ozone use in aquaculture systems and also its
safety to cultured organisms.
OZONE
Chemistry of Ozone
Ozone is a very unstable molecule and, after injection into
raw water, it decomposes very rapidly. The pathway of the
ozone reaction depends on the properties and concentrations
of the other compounds, as well as the quality of the water,
including pH, bicarbonate level, the level of total organic
carbon and temperature. These factors affect the decompo-
sition of ozone and therefore the effective oxidation power
and oxidation rate of ozone (Lawson, 1995; Summerfelt and
Hochheimer, 1997; Rakness, 2005).
In the presence of organic impurities, the half-life of ozone
is reduced to minutes (Duvivier et al., 1996). The rate of
decomposition also increases with increasing pH; and at low
pH (<7), molecular ozone (O3) is the dominant species; as pH
increases, O3turns into the very short-lived (microseconds)
hydroxyl radicals (OH) (Rice and Wilkes, 1992; Rakness,
2005).
The chemistry of ozone in seawater is considerably
different than that in brackish and freshwater. This dif-
ference in chemistry has a profound effect on disinfec-
tion. The most important difference with ozone chemistry
in seawater compared to freshwater is due to the pres-
ence of bromide ion (Br) in seawater (Oemcke and van
Leeuwen, 1998). Bromide ion catalytically decomposes ozone
(Westerhoff et al., 1998; von Gunten 2003a, 2003b). In sea-
water, the primary brominated compounds formed by ozone
are hypobromous acid (HOBr), which is in equilibrium with
hypobromite (OBr). These compounds have disinfection
properties (Herwig et al., 2006; Perrins et al., 2006). Bromine
has disinfectant properties and is quantified as total residual
oxidant (TRO), usually in units of mg Br2L1or mg Cl2L1
(White, 1999).
Typical bromide ion concentrations of 60–70 mg L1in
seawater give a high formation potential of bromo-oxides,
even at a common oxidation ratio of 10%. Correspondingly,
chloro-oxy anions may be formed, but these are limited
by a higher oxidation-reduction potential barrier (Grguric
et al., 1994). The following reactions of ozone with bromide
ion (Br) and hypobromite ion (OBr) have been proposed
(Grguric et al., 1994; Bonacquisti, 2006):
O3+BrO2+OBr
O3+OBr2O2+Br
2O3+OBr2O2+BrO
3
The first reaction shows that ozone will oxidize the bro-
mide ion in seawater to hypobromite ion. The hypobromite ion
will hydrolyze into hypobromous acid, which is a weak acid
with a pKaof 8.8 at 20 C (HOBr OBr+H+). The sum
of HOBr and OBris the biocidal bromine. In seawater with
a typical pH of 8, hypobromous acid will predominate and
be the most important disinfectant with a half-life of hours to
days dependent on light conditions and water quality charac-
teristics (Westerhoff et al., 1998; Legube et al., 2004; Liltved
et al., 2006).
In freshwater, ozone decomposes rapidly to oxygen after
application. By introducing ozone in aquacultural seawater
systems, a series of redox-reactions take place and several
reactive intermediates are formed (Liltved et al., 2006). The
halogen ions in seawater are oxidized by ozone to halo-
oxy anions. The specific formation potential is highest for
iodine and bromine and somewhat lower for chlorine species.
Although the iodide ion may be completely oxidized, the
low iodide concentration in seawater (<1mgL
1) makes the
succeeding oxidation products less important.
System and Application
The design of the ozone reactor or contact vessel is very
important for safe, successful ozonation. There are a range
of reactors available using various designs to transfer ozone
to the water. Designs include fine bubble diffusers, turbine
contactors, injectors, deep u-tube reactors, packed columns,
static mixers and spray contact chambers. Some designs are
also used for oxygen transfer or aeration. Each design has
advantages and disadvantages not discussed here, but some
important considerations when choosing a reactor include:
i) ozone transfer efficiency; ii) leak-free design and con-
struction; and, iii) construction with ozone resistant materials
(Hunter, 2000; Rakness, 2005; Read, 2008).
Ozone can be applied continuously, as a series of treat-
ments per day or as a single batch treatment per day (Hunter,
2000; Rakness, 2005). Application in most situations can be
linked to the feeding strategy employed in the culture system.
Three to 4 h after feeding fish, the concentrations of ammonia,
dissolved organics and other wastes products reach a maxi-
mum. If fish are fed several times during the day, a series of
ozone treatments can be introduced after each feed to target
the associated rise in waste levels. If feed is introduced 24 h
per day, water quality degrades continuously and so ozone
application should be continuous. A single-batch ozone treat-
ment can be used to target rises in waste levels in the system
associated with a moderate feed event or to treat batches of
exchange or inlet water from the supply source. Continuous
346 A.A. Gonçalves and G.A. Gagnon September–October 2011
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
ozonation is beneficial when compared to batch and serial
treatments because water quality remains relatively stable.
However, the lower costs of serial and batch ozonation make
these treatments regimes viable management options (Hunter,
2000; Read, 2008).
The required amount of ozone for treatment in a RAS is
usually calculated according to the daily feed rate. Rates of
10–15 g ozone per kg feed are generally recommended to
reduce accumulated organics. Any background organic load-
ings of the source water used for the RAS should also be taken
into account (King, 2001).
Ozone Residual (in-Water) Measurement
Accurate measurements of ozone residual are required to
properly determine disinfection credit. The accuracy of the
ozone residual measurement is affected by the sampling sys-
tem layout for both grab sample tests and continuous readings
from on-line meters. Successful ozone residual measurement
is enhanced by sampling systems that: i) minimize deten-
tion time in the sample line; ii) allow for easy collection of
grab samples that can be taken without affecting the on-line
instrument measurement; iii) are properly configured such
that consistent and sufficient flow is available to an on-line
instrument; and, iv) provide flexibility for measuring ozone
residual at a variety of sampling locations within the ozone
contactor (Rakness and Hunter, 2000; Rakness, 2005; Lee
et al., 2008).
The successful use of ozone in the aquaculture indus-
try also requires a reliable, simple and fast method for its
measurement that can be conducted on-site at a hatchery,
under non-laboratory conditions. The test must allow multi-
ple readings during the course of a treatment to ensure that
constant concentrations are maintained to prevent an unneces-
sary overexposure and high mortalities (Buchan et al., 2005).
Although direct exposure of fish and other organisms to ozone
and the oxidants formed in ozonated sea water can be lethal
(Wedemeyer et al., 1979a, 1979b; Coler and Asbury, 1980;
Fisher et al., 1999), most of them can tolerate varying levels
of dissolved ozone. Specific exposure levels need to be deter-
mined for each species (Grotmol et al., 2003), and reliable
methods to measure ozone in sea water are therefore needed
to ensure that lethal limits are not exceeded. However, no
single method is consistently used by the scientific commu-
nity or industry, making comparisons and standardization of
exposures difficult (Buchan et al., 2005).
The most important reaction is the oxidation of bromide
ions (Br), forming hypobromite ions (OBr), which can then
either be reduced back to Bror further oxidized to form
bromate ions (BrO3) (Grguric et al., 1994; Buchan et al.,
2005; Liltved et al., 2006). The residual oxidant (OBr) inter-
feres with reagents used to measure ozone as it reacts with
the reagents as if it were ozone. Because of this, it is impor-
tant to be careful of the units used to report dissolved ozone
concentrations in seawater. The values obtained by measuring
“ozone” will not only include any ozone present, but also any
other oxidants in the sample. However, dissolved ozone con-
centrations in seawater can be expressed in mg L1of total
residual oxidants (TRO), or ozone produced oxidants (OPO),
or residual ozone concentration (ROC) and ozone (Buchan
et al., 2005).
Comparative ozone residual (in-water) measurement
methods. Five methods for measuring ozone are cited reg-
ularly in the literature, often using different measurement
units and with little to no explanation as to why a particular
method was chosen. These methods (summarized in Buchan
et al., 2005) are: (1) Neutral buffered iodometric method
(Shechter, 1973) where ozone concentrations is reported in
mg L1of TRO; (2) Modified neutral buffered iodometric
method (Sugita et al., 1992; 1996), reported in mg L1of
TRO; (3) Iodometric titration method (Franson, 1989), usu-
ally used to measure chlorine levels, and ozone concentrations
using this method are reported in mg L1of TRO or OPO; (4)
DPD (N,N-diethyl-p-phenylenediamine) colorimetric method
(Franson, 1976), ozone concentrations using this method are
reported in mg L1of ROC or TRO; (5) Oxidation/reduction
potential (ORP), an ORP probe attached to a meter is placed
in the water sample, giving the ORP in mV. This technique
was used by Tango and Gagnon (2003).
Also, ozone residual can be measured by gravimetric
method (Yates and Stenstrom, 2000), i.e., determining volume
based on weight. This procedure is easy to implement in the
field and is highly accurate when properly performed. Buchan
et al. (2005) compared commonly cited methods of ozone
measurement for their ability to measure dissolved ozone
and the ease with which they can be applied on-site in an
aquaculture facility. Among the different units for expressing
dissolved ozone concentrations examined, it is recommended
that under non-laboratory (i.e., hatchery) conditions, a DPD
“total chlorine test” be used to measure dissolved ozone lev-
els, and the results be reported in TRO (mg L1as Cl2).
The TRO concentrations (mg L1) were also calculated and
expressed as equivalent concentrations of bromine (Br2;1mol
Cl2=0.44 mol Br2) (Jones et al., 2006; Perrins et al., 2006;
Lee et al., 2008).
Similar to the TRO measurement, AccuVac®Ampoules
were used with freshly collected samples and analyzed using a
water quality laboratory spectrometer. This is a commercially
available modification of the indigo colorimetric method pro-
duced by Hach Chemical Company, Loveland, Colorado,
USA (Hunter and Rakness, 2002). Powdered indigo reagent
is enclosed in a small glass under vacuum. When the tip of the
ampoule is broken under the water surface, a specific volume
of the sample water is drawn into the ampoule and mixed
with a specific amount of indigo powder (i.e., low-, medium-,
and high-range measurement). Low-range ampoules are used
when the ozone residual is less than 0.25 mg L1; medium-
range ampoules are used to measure ozone residual between
zero and 0.75 mg L1; and high-range ampoules are used to
measure ozone residual between zero and 1.5 mg L1,with
Overview of Ozone in Recirculating Aquaculture System September–October 2011 347
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
a sensitivity of 0.1 mg L1(Rakness, 2005; Herwig et al.,
2006). This method is more simple and easy to use.
To compare the official Indigo Colorimetric Method for
ozone residue determination in water to Accuvac®test kit,
an experiment was performed using Milli-Q®water (Gradient
System A10, MilliporeTM, Billerica, MA, USA) as a con-
trol (Salinity 0‰), brackish water (Salinity 5‰) and seawater
(Salinity 25‰). A laboratory semi-bath apparatus method was
used for this study was carried out at the Center of Water
Resources Studies Laboratory, Dalhousie University (Halifax,
NS, Canada). Compressed air is passed into the ozone gener-
ator (VMUS-4, AZCO Industries Ltd., Langley, B.C. Canada)
where high voltage corona discharge causes the breakdown of
oxygen molecules into oxygen radicals which combine with
the oxygen molecules and form ozone. The produced ozone
is passed into the 10 L Glass through the tubing (rate of
0.3434 mgO3L1min1) for 60 minutes and turned off. For
the collection of residual ozone in the off gas system, a 2%
potassium iodide solution was used in a flask as shown in
Figure 1.
The temperature and pH during each experiment were
maintained constant and the average was for temperature
25.05±0.28 (Salinity 0‰ - control), 25.18±0.12 (Salinity
5‰) and 25.10±0.21 (Salinity 25‰); and for pH 5.48±0.79
(Salinity 0‰ - control), 7.41±0.49 (Salinity 5‰) and
8.05±0.56 (Salinity 25‰). Ozone concentrations, total resid-
ual oxidants (TRO), bromate and bromoform were monitored
each 5 min during 60 min of ozonation and 60 min after
stopped ozonation, and were performed in duplicate using
standard methods (APHA, 1995) as follows:
i. Residual aqueous ozone concentration (mgO3L1):
was measured using the official Indigo Colorimetric
Method for ozone determination in water (4500-O3:
Ozone residual) (APHA, 1995) and Accuvac®test
kit (Hach Co., Loveland, CO, USA). Freshly col-
lected samples were analyzed by HACH DR/4000U
Spectrophotometer (Method 8311). The ampoules had
a range of 0–1.5 mg L1, with a sensitivity of 0.1 mg
L1ozone;
ii. Total residual oxidant (TRO): ozone quickly reacts
with bromide ion in seawater, forming hypobro-
mous acid that is in equilibrium with hypobromite.
Together, these compounds are referred to as
bromines and they constitute TRO measure in
ozonation seawater (Herwig et al, 2006). TRO
was determined using a standard DPD colorimet-
ric analysis for bromine (measured as mgBr2L1),
analyzed by HACH DR/4000U Spectrophotometer
(Method 8016);
iii. Oxidation/reduction potential (ORP): was measured
in mV using a Fisher Scientific Accumet Excell
XL60 Meter;
iv. Bromate (mgBrO3L1): was measured using a
761 Compact Ion Chromatography and 788 IC Filtration
Sample Processor (Metrohm AG, Herisau, Switzerland).
Calibration verification standard was performed (R2=
0.9996) for all analyses in accordance with their strin-
gent QA/QC (Quality Assurance and Quality Control)
program. The percent recovery of each compound was
calculated and recorded on the quality control chart. The
FIGURE 1. Schematic of the laboratory scale semi-batch apparatus during ozone treatment.
348 A.A. Gonçalves and G.A. Gagnon September–October 2011
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
calibration curve was: y =3.7225x +0.2499 (R2=
0.9996);
v. Bromoform (µgCHBr3L1): was measured using
a CP3800 VARIAN Gas Chromatograph and
CP8400 Auto Sampler System (VARIAN Inc., Palo
Alto, CA, USA). The calibrations were done (R2 =
0.9923) for all analyses in accordance with their strin-
gent QA/QC (Quality Assurance and Quality Control)
program. The percent recovery of each compound was
calculated and recorded on the quality control chart.
The calibration curve was: y =0.5358x +0.6198
(R2=0.9923).
The results were presented in Figure 2 and showed sim-
ilar results using both methods suggesting that we can use
each one to check ozone concentration in water. This experi-
ment only reinforces the need to be cautious when comparing
different ozone concentrations made using different measur-
ing techniques. The Indigo colorimetric method must to be
done carefully because you need to prepare indigo stock
solutions, standardize indigo reagents, and when necessary,
diluted samples, which increase the variability in the final
results.
Interestingly, the residual ozone concentration was lower in
water with 25‰ of salinity when compared to 5‰ and Milli-
Q water, which has lower concentration of organic matter and
could react with ozone. But the main concern was the resid-
ual ozone concentration after 60 minutes of the ozonation was
stopped. Some authors (Summerfelt, 2003; Summerfelt et al.,
2004; Liltved et al., 2006) comment the importance of the
presence of a treatment unit for removal residual ozone and
others oxidants, which could be toxic for aquatic organisms.
This issue is discussed later.
A major encumbrance in the use of ozonation of seawa-
ter is that it is difficult to measure accurately the formation
of ozonation by-products (OBP), particularly inorganic OBP
such as bromate and other brominated compounds (Tango and
Gagnon, 2003). The measurement of OBP is further compli-
cated because the chemistry of seawater is variable especially
inshore and in estuaries, often subject to fluctuations in salin-
ity and turbidity due to wave action and terrestrial run-off. The
content of OBP may be estimated indirectly via oxidation–
reduction potential (ORP, or redox potential), which measures
the potential of the seawater to oxidize or reduce, and is
thus an indication of its ability to disinfect against micro-
organisms or kill aquaculture animals (Tango and Gagnon,
2003).
ORP may be used to control ozone addition to seawater but
is not necessarily equivalent to the ability of the treated water
to disinfect (Tango and Gagnon, 2003). Rather than measure
ozone directly, an ORP probe measures the total capacity,
in millivolts (mV), or various oxidants in a solution to oxi-
dize an electrode. By keeping ORP measurements within a
certain range, the levels of total oxidants can be controlled,
which gives indirect control over ozone. A safe ORP level of
FIGURE 2. Ozone measurements during and after ozonation
using the official indigo method (COLOR) and Accuvac®test kit
(AMP); MILQW (Milli-Q water), 5 PSU (5 practical salinity units or
5‰), 25 PSU (25 practical salinity units or 25‰).
freshwater fish culture is generally considered to be 300 mV
(Read, 2008).
A major encumbrance in the use of ozonation of seawa-
ter is that it is difficult to measure accurately the formation
of ozonation by-products (OBP), particularly inorganic OBP
Overview of Ozone in Recirculating Aquaculture System September–October 2011 349
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
such as bromate and other brominated compounds (Tango and
Gagnon, 2003). The measurement of OBP is further compli-
cated because the chemistry of seawater is variable especially
inshore and in estuaries, often subject to fluctuations in salin-
ity and turbidity due to wave action and terrestrial run-off. The
content of OBP may be estimated indirectly via oxidation–
reduction potential (ORP, or redox potential), which measures
the potential of the seawater to oxidize or reduce, and is
thus an indication of its ability to disinfect against micro-
organisms or kill aquaculture animals (Tango and Gagnon,
2003).
ORP may be used to control ozone addition to seawater but
is not necessarily equivalent to the ability of the treated water
to disinfect (Tango and Gagnon, 2003). Rather than measure
ozone directly, an ORP probe measures the total capacity, in
millivolts (mV), or various oxidants in a solution to oxidize
an electrode. By keeping ORP measurements within a certain
range, the levels of total oxidants can be controlled, which
gives indirect control over ozone. A safe ORP level of fresh-
water fish culture is generally considered to be 300 mV (Read,
2008).
Results presented in Figure 3 showed final ORP values
below to 250 mV and 150 mV in brackish water (salinity of
5‰) and Milli-Q water (0‰), respectively. However, in sea
water (25‰) the ORP values have remained above 500 mV,
which could be dangerous to aquatic organisms. According to
Tango and Gagnon (2003), this indirect measure of residual
ozone (ORP) could be an indication of process water’s poten-
tial to disinfect or to kill fish, and moreover, can be used to
control ozone addition and thus ensure the desired treatment
objective.
OZONE APPLICATION IN RECIRCULATING
AQUACULTURE SYSTEMS
Ozone application within aquaculture systems requires
ozone generation, ozone transfer into solution, contact time
for ozone to react, and possibly ozone destruction to ensure
that no ozone residual makes it into the culture tanks
(Summerfelt and Hochheimer, 1997; Summerfelt et al., 2001;
Summerfelt, 2003; Summerfelt et al., 2004; Liltved et al.,
2006; Sharrer and Summerfelt, 2007). Ozone off-gas needs
to be destroyed too.
Effective transfer of ozone into water is important because
the cost of producing ozone is not insignificant, especially if
the ozone is carried within purified oxygen feed gas that is
either purchased or produced on site. The rate of ozone trans-
fer and the subsequent rate of ozone decomposition depend
upon the contact system efficiency and the reaction rates of
ozone with constituents in the water. The ozone reaction rate
depends on the water temperature and on the concentration
and type of constituents contained in the water (Summerfelt,
2003; Summerfelt et al., 2004).
Ozone is used in aquaculture systems to improve water
quality and overall system performance (King, 2001).
Ozonation has proven useful in aquaculture systems in pro-
moting the removal of solid matter (Rueter and Johnson, 1995;
Tango and Gagnon, 2003), stabilization of water quality in
recirculating systems (Reid and Arnold, 1994; Summerfelt
et al., 1997), water clarification and dissolving nonbiodegrad-
able organic material (Paller and Lewis, 1988; Summerfelt
and Hochheimer, 1997; Summerfelt et al., 2004), is advanta-
geous in disease control (Liltved and Landfald, 1995; Liltved
FIGURE 3. ORP measurements during ozonation (60 min) and after stopping ozonation.
350 A.A. Gonçalves and G.A. Gagnon September–October 2011
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
et al., 1995), and ozonation of the water appeared to remediate
the fish health decline (Good et al., 2009). This indicates that
ozonation could find a place in the intensive culture, but also a
potential candidate for commercial aquaculture (Jobling et al.,
1993; Ritola et al., 2000).
In general, ozone is an effective bactericide, parasiticide,
and virucide (Lohr and Gratzek, 1984; Colberg and Lingg,
1978; Liltved et al., 1995; Bullock et al., 1997; Liltved, 2002)
by disrupting cell membrane function, entering the cell and
destroying the nuclear chemistry of the cell (Lawson, 1995;
Sharrer and Summerfelt, 2007); however, some viruses have
shown high resistance to ozonated seawater (Liltved et al.,
2006). This effectiveness is a function of dosage and con-
tact time (Lawson, 1995). The target organism and water
quality determine the required concentration of ozone and
the necessary contact time (Lawson, 1995; Summerfelt and
Hochheimer, 1997). Microbial reductions are limited by the
ability to maintain a specific ozone concentration for the time
needed (Summerfelt and Hochheimer, 1997).
In spite of the apparent advantages of ozone in aquaculture,
its physiological effects on the fish are still largely unknown
since unexpected deaths have occurred even at low O3con-
centrations (Bullock et al., 1997). This issue will be discussed
later in this paper.
Reduction of Organic Matter
Besides its primary use in disinfection, the ozonation pro-
cess has been observed to be associated with flocculation in
waters containing particulate matter. Moreover, an essential
ingredient for the success of any hatchery is a suitable water
supply having both sufficient flow rate and appropriate quality.
The presence of fine particulate matter can cause problems,
i.e., shielding of pathogens, a decreased oxygen transfer rate
to eggs and fry, and the deposition of sediment in rearing
zones. Rueter and Johnson (1995) found in their study that
the use of ozone prior to sedimentation or filtration improved
the removal of suspended solids and concluded that ozone’s
multiple uses for disinfection, water aeration, removal of
metabolic by-products, along with the improvements in sus-
pended solids removal shown in this and other research, make
it an especially appropriate treatment process for hatcheries.
Summerfelt et al. (1997) applied ozone to water in a recir-
culating rainbow trout (Oncorhynchus mykiss) culture system
just prior to the culture tanks in order to oxidize nitrite and
organic material, improve overall water quality, and assist
removal of solids across the microscreen filter. Adding ozone
(25–39 g ozone/kg feed fed) reduced the mean concentra-
tion of Total Suspended Solids (TSS) by 35%, Chemical
Oxygen Demand (COD) by 36%, Dissolved Organic Carbon
(DOC) by 17%, and color by 82% within the water enter-
ing the culture tanks. Additionally, ozone reduced the mean
nitrite concentration by 82% within the culture tanks; and,
reduced bacterial gill disease associated mortalities and chem-
ical treatments required to control bacterial gill disease (BGD)
epizootics.
Ozonation leads to coagulation of smaller particles into
larger ones, and the large particles are removed from the sys-
tem by the screen filter, the question remains as to the effect
of ozone on the particle size distribution of the remaining par-
ticles. With continual production and continuous breakdown,
aggregation, and removal of particles, it is difficult to predict
the net effect of ozone on the particle distribution in a recircu-
lating system. The study conducted by Krumins et al. (2001b)
addresses this question by quantifying the effect of ozona-
tion on the particle size distribution in RAS. Surprisingly,
regardless of ozone dose, there was no significant difference
in the slope of the power law fit for the particle size distri-
bution. In over one-half of the experimental trials when no
ozone was added, the particle size distributions were distinctly
bimodal.
Edwards et al. (1993) note that ozone can act to remove
carbon (Total Organic Carbon TOC) in two ways, by oxi-
dizing dissolved organic matter or by improving coagulation
of organic-containing particulates. They found that ozona-
tion improved particle removal for doses of less than about
0.7 mg ozone per mg TOC. Rosenthal (1981) reported that
ozone doses of approximately 7–10 mg L1(approximately
0.1–0.2 mg ozone per mg TOC, this author’s calculation) in a
RAS increased the biological oxygen demand (BOD5)ofthe
water.
According to Krumins et al. (2001a), recent work on
ozone application in RAS gives guidelines as to the daily
ozone dose as a function of feed rates, but does not indicate
whether the ozone should be added continuously throughout
the day or in shorter, more intense doses. Their study exam-
ined the effect of adding the same total amount of ozone
(15 g ozone/kg feed) 24, 12, and 6 h per day, compared
with a control (without ozone). The ozone treatments signif-
icantly (p<0.05) reduced TOC, turbidity, and total ammonia
nitrogen (TAN) compared with the control, but surprisingly
did not significantly reduce average nitrite concentrations.
They concluded that cycling ozone on and off throughout
the day can help to maintain a stable population of nitrifying
bacteria.
Water Disinfection
Recirculation of water is practiced when the water supply
is limited or when energy saving is demanded. Disinfection
seems to be required at high degrees of recirculation in order
to suppress the general microflora, and to prevent spread-
ing of pathogens with the effluent from special aquacultural
activities, like as scientific laboratories working with infected
fish and fish pathogens, and processing industry for farmed
fish (Liltved et al., 1995; Bullock et al., 1997; Buchan
et al., 2005).
Recirculating systems for salmonids can require excep-
tional water quality and tight biosecurity to reduce the likeli-
hood of restricted fish growth and increased mortality (Noble
and Summerfelt, 1996). To optimize water quality, recirculat-
ing systems will use water treatment processes that effectively
Overview of Ozone in Recirculating Aquaculture System September–October 2011 351
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
and rapidly remove fecal matter and waste feed, because rapid
removal of organic matter can minimize the amount of fine
particulates, soluble organic compounds, and ammonia that
they would release if given the opportunity to degrade within
the recirculating system (Summerfelt et al., 2004).
Bacterial reduction and viral inactivation may be desirable
within recirculating systems. However, to disinfect recirculat-
ing systems water with ozone could be very expensive due to:
(1) the much higher ozone loading required to overcome the
organic demand and to sustain a residual that would be suffi-
cient to achieve significant bacterial and viral reductions; and
(2) the need to strip any remaining residual ozone from the
water before it is returned to the culture tank (Bullock et al.,
1997).
According to Bullock et al. (1997) adding ozone at a lower
rate (25 g ozone/kg feed) to water in a recirculating rain-
bow trout (Oncorhynchus mykiss) culture system just before
it entered the culture tanks could provide about the same ben-
efits as a higher dosing rate (36–39 g ozone/kg feed fed)
e.g., reduced bacterial gill disease (BGD) associated mortali-
ties and no required use of non-approved chemical treatments
to control BGD epizootics. In the sequence of this study,
Summerfelt et al. (1997) showed that ozone can improve over-
all water quality at the same ozone rate. Hence, use of the
lower dose could provide all of the benefits but also reduce
capital and operating costs associated with the higher ozone
dosing rate.
Greater reductions in bacteria within the recirculating sys-
tem, with its high oxidation demand, would have required
ozone loading rates greater than those used by Bullock et al.
(1997) (i.e., >39 g ozone/kg feed), which would be diffi-
cult to achieve without: (1) wasting excess oxygen to carry
more ozone to the low-head oxygenators (LHO®) unit, and/or
(2) replacing the ozone generator with a larger unit that could
produce a higher ozone concentration in the oxygen feed
gas (6–l0% instead of 4–5%), and/or (3) installing an ozone
removal unit (air stripper, UV light, or large hydraulic reten-
tion chamber) to prevent the increased ozone residual from
reaching toxic levels in the culture tank.
Ozone has a unique characteristic that distinguishes it from
other disinfectants: it is highly volatile and disappears shortly
after contact with organic material suspended in water. Unlike
other disinfectants, ozone can be continuously applied to a
water stream within a culture system if there is sufficient res-
idence time for its dispersion between the treatment location
and the pond intake (USEPA, 1999; Schuur, 2003; Newman,
2006; Lee et al., 2008).
Studies from Meunpol et al. (2003) showed that degree
of bacterial inhibition also depend on the concentration of
ROC at first contact. Longer ozonation resulted only in longer
suppression period but not greater bacterial reduction.
The total dose of ozone is commonly expressed as a CT
value, which is the product of the ozone concentration (C) and
exposure time (T). Several studies have found that the toxic-
ity of ozone generally increases with ozone dose (CT value)
(Davis and Arnold, 1997; Theisen et al., 1998; Su et al., 2001).
However, the individual contributions of C and T to the toxic-
ity of the ozone dose can only be assessed in studies involving
factorial combinations of C and T (Rakness, 2005).
Ozone oxidation can kill microorganisms, but disinfect-
ing the water requires maintaining a certain dissolved ozone
concentration for a given contact time. Thus, disinfecting
efficiency depends on the product of the ozone residual con-
centration multiplied by its contact time. An ozone contact
vessel should provide the time necessary for the ozone resid-
ual to react with and inactivate the target microorganism(s).
Disinfecting water can require maintaining a residual ozone
concentration of 0.1–2.0 mg L1in a plug-flow type contact
vessel for periods of 1–30 min, depending upon the target
microorganism (Summerfelt, 2003). Ozone at a concentration
not exceeding 0.5 mg L1(to minimize bromate production)
can be used to treat seawater in batches for periods up to
10 minutes (Lee et al., 2008).
If disinfection is the primary goal of ozonation, the amount
of ozone necessary is largely dependent on the background
organic loading of the water to be treated (Summerfelt, 2003).
In pure water, residual concentrations of 0.01–0.1 mg L1
ozone for periods as short as 15 sec can be effective in reduc-
ing bacterial loads. However, in water with organic loadings
the residual ozone concentration and/or contact time of ozone
must be increased to produce significant disinfection. Natural
waters (seawater, brackish and freshwaters) generally require
residual concentrations of between 0.1–0.2 mg L1ozone and
contact times of 1–5 min for disinfection. Aquaculture effluent
generally requires between 0.2–0.4 mg L1residual ozone for
1–5 min for significant disinfection to occur after oxidation of
organics (Read, 2008).
There is growing awareness of the need to disinfect water
entering and leaving aquaculture systems. Improvement of the
microbial standard of inlet water and stringent microbiologi-
cal restrictions on effluent water are in many cases needed to
control diseases in the fish farming industry. Different types
and volumes of water to be disinfected and different target
pathogens may influence the choice of method, dose require-
ments and other design criteria for disinfection units (Liltved
et al., 1995; Bullock et al., 1997).
In freshwater hatcheries, the necessity of new methods is
mainly for fungi pathologies control. Ozone (O3)isavery
unstable allotrope state of oxygen producing O native char-
acterized by high bactericide activity (Forneris et al., 2003).
In some cases, almost the complete elimination of pathogenic
bacteria such as Aeromonas salmonicida,Vibrio anguillarum,
Vibrio salmonicida and Yersinia ruckeri has been obtained
(99.9%). The disappearance of viral diseases such as infec-
tious pancreatic necrosis virus (IPNV) in both fresh and
marine water fish (Colberg and Lingg, 1978; Liltved et al.,
1995) was also observed.
Nodavirus is the causative agent of viral nervous necrosis
(VNN), also known as viral encephalopathy and retinopathy
(VER), or fish encephalopathy. Nodavirus infections are a
worldwide problem affecting over 30 species of marine fin-
fish including Atlantic halibut (Hippoglossus hippoglossus),
352 A.A. Gonçalves and G.A. Gagnon September–October 2011
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
sevenband grouper (Epinephelus septemfasciatus), Japanese
flounder (Paralichthys olivaceus), red drum (Sciaenops ocel-
latus), winter flounder (Pseudopleuronectes americanus),
Atlantic cod (Gadus morhua) and haddock (Buchan et al.,
2006). Arimoto et al. (1996) assessed the effectiveness of
ozone (as a total residual oxidant) to inactivate striped jack
nervous necrosis virus (SJNNV) and 0.1 µgL
1was required
to inactivate SJNNV for 2.5 minutes. Also, washing fertilized
eggs and the treatment of sea water with ozone decrease the
rate of occurrence of viral nervous necrosis (VNN).
Due to few published data on virus inactivation in seawater,
Liltved et al. (2006) initiated a study to supply fish farm-
ers, consultants and manufacturers with more precise dose
requirements of ozone and UV irradiation to accomplish inac-
tivation of viruses important in aquacultural systems. Such
data are crucial to establishing a firm basis for the design of
disinfection systems and for better operation and control of
existing installations. The viruses studied were IPNV, Atlantic
halibut nodavirus (AHNV) and infectious salmon anaemia
virus (ISAV) and the results demonstrated a wide span in
UV and TRO resistance among them. ISAV was sensitive to
both methods, while high resistance in AHNV and IPNV were
experienced. The TRO resistance in IPNV and AHNV con-
tradict earlier published results and suggests reconsideration
of existing ozonation practice to inactivate these viruses in
seawater (Liltved et al., 2006).
Among the numerous viruses of penaeid crustacea, several,
notably Taura Syndrome Virus (TSV), White-spot Syndrome
Virus (WSSV), and Yellow-head Virus (YHV), are responsi-
ble for epizootic incidents in commercial shrimp ponds that
have resulted in high mortality and economic losses. WSS was
first reported in 1992 in Penaeus japonicus cultured in north-
eastern Taiwan and continues to cause substantial losses in the
aquaculture industry in many countries; any practical means to
eradicate or to inactivate the virus in the culture environment
would be of enormous practical benefit (Chang et al., 1998;
Schuur, 2003).
Chang et al. (1998) investigated the efficacy of some com-
monly used chemical and physical disinfectant methods for
inactivation of white spot syndrome bacilovirus (WSBV) in
juvenile black tiger prawn (Penaeus monodon) and concluded
that the effective concentration for ozone to reduce WSBVs
infectivity to zero was 0.5 mg L1as a total residual oxidant
for 10 min at 25 C.
In controlled bioassay conditions a concentration-time
(CT) product sufficient to deactivate most viruses including
WSSV is about 5.0 (e.g., 1.0 mg L1for5minor0.5mgL
1
for 10 min). A nominal design value for ozone disinfection is a
CT product of 1.5–2.0 times the bioassay value. A CT product
of 7.5–10 is therefore suggested as an estimate for crustacean
viruses (Schuur, 2003).
The causative virons have been detected in eggs, larvae
and broodstock of striped jack (Pseudocaranx dentex), indi-
cating that spawners can be a source of infection (Arimoto
et al., 1992). Nodavirus is durable and tolerant to various envi-
ronmental conditions (Frerichs et al., 2000). However, ozone
has been shown to inactivate nodavirus, although specific
VNN titers were not reported (Arimoto et al., 1996; Grotmol
and Totland, 2000). Fertilized eggs from different species of
fish can tolerate varying levels of dissolved ozone, so spe-
cific exposure levels need to be determined for each species
(Grotmol et al., 2003).
Considering these concerns, Buchan et al. (2006) inves-
tigated the tolerance of newly fertilized haddock eggs to
dissolved ozone and to determine if this exposure is sufficient
to disinfect against piscine nodavirus. Ozone can successfully
disinfect fertilized haddock eggs against nodavirus. It is rec-
ommended that fertilized haddock eggs be disinfected with an
ozone dose of 3.0 mg L1TRO for 3.3–6.7 min (10–20 CT
units).
Treatment with ozone has also shown a positive effect
against infection in abalone (Dixon et al., 1991) and in crus-
taceans, against viral pathologies (Chang et al., 1998) or
against viral pathologies of the pancreas in Atlantic salmon
(McLoughlin et al., 1996). The effect of ozone has also
resulted positive in the treatment of ceratomyxosis in rainbow
trout (Tipping, 1988). In adult fish, the use of ozone requires
particular care as it is the cause of death, even at low concen-
trations, following the alteration of the branchial epithelium
and the outermost layers of the epidermis (Wedemeyer et al.,
1979b; Paller and Heidinger, 1980; Richardson et al., 1983).
The effectiveness of ozone has instead proved to be limited
in particular farming conditions, for example, when there is a
high amount of suspended organic substances and even some-
times in recycling systems (Bullock et al., 1997; Summerfelt
et al., 2004).
Saprolegniasis is a widespread mycotic infection in fresh-
water aquaculture and represents a serious problem affecting
egg production in trout hatcheries (Forneris et al., 2003). The
damage results in an average annual lack of production of
about 20%, with peaks higher than 40%. Saprolegniasis is a
secondary manifestation of a pathology suffered by develop-
ing embryos. In order to reduce the presence of water moulds,
Forneris et al. (2003) studied the effectiveness of ozone as a
fungicide to control the incidence of saprolegniasis in trout
eggs incubation. From the results, it has emerged that the
treatment with ozone is effective and the hatching eggs range
from 42.6% to 49.1% dose of ozone from 0.01 to 0.2 mg L1.
Ozone is also a likely candidate for disease prevention
and water quality management in shrimp culture (Rosenthal,
1980; Menasveta, 1980; Matsumura et al., 1998; Sellars et al.,
2005). Despite ozone’s potential benefits, few shrimp cultur-
ists embrace its use (Matsumura et al., 1998), perhaps because
practical details are often lacking, and because of inconsistent
results or the limited documentation on ozone’s effectiveness
to inactivate shrimp viruses (Sellars et al., 2005).
Meunpol et al. (2003) evaluated the effects of ozone on
bacteria and black tiger shrimp postlarvae (Penaeus mon-
odon), which included using different ozone exposures on
shrimp as well as on harmful bacteria (Vibrio harveyi D331),
and beneficial probiotic bacteria (Bacillus S11). They con-
cluded that shrimp postlarvae exposed to 0.34–0.50 mg O3
Overview of Ozone in Recirculating Aquaculture System September–October 2011 353
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
L1(8-h ozonation) caused loss of balance, immobility and
destruction of gill lamellar epithelium, but damage of organ-
isms by ozone is not only related to ozone concentration but
exposure time.
Although ozonation has been proven non-toxic to post-
larval shrimp at concentrations capable of eliminating viruses
present in the culture systems (Blogoslawski et al., 1977;
Jiang et al., 2001; Meunpol et al., 2003), protocols are yet
to be established for using ozone as a virucidal treatment
for shrimp embryos. An assessment of embryo tolerance to
ozone is essential before investigations into ozone’s capac-
ity as a virucidal embryo treatment. Such studies highlight
the requirement for comprehensive toxicity trials to be per-
formed for each cultured species before ozone can be applied
to improve water quality or deactivate pathogens in a culture
system.
Sellars et al. (2005) investigated the tolerance of Penaeus
(Marsupenaeus)japonicus embryos to ozone disinfection.
The toxicity of ozone applied at different concentrations and
periods of time was assessed at three developmental stages
(post-spawning treatment times) for three separate families.
This study shows considerable potential for using ozone as
a virucidal agent for P. japonicus embryos as they can tol-
erate exposure to ozone at concentrations well above those
reported to inactivate viral and bacterial pathogens of other
marine species (CT 2 mg L1). Future studies investigating
the ability of ozone to inactivate viral pathogens of P. japoni-
cus embryos should use 2 mg L1ozone for 1 min applied at
either 120 or 480 min post-spawning detection.
Following from previous work examining the tolerance of
P. japonicus embryos to ozone disinfection (Sellars et al.,
2005), Coman et al. (2005) assessed the hatch rates of
P. japonicus embryos ozonated at several CT values, gener-
ated from different combinations of C and T, to determine
the relative effect that C and T has on ozone toxicity, and
suggested that the toxicity of ozone is more dependent on C
than T. This suggests that CT value is a redundant measure of
ozone toxicity. Protocols for ozonation in aquaculture should
quote the individual values of C and T that provide effective
disinfection for the cultured organism.
Ritar et al. (2006) examined different levels of ozonation
of culture water, as determined by ORP levels, on the sur-
vival, growth and bacteriology of phyllosoma larvae of the
southern rock lobster (Jasus edwardsii). The culture of phyl-
losoma to Stage IX using ozonated seawater was effective
in controlling pathogenic bacteria to improve larval survival.
However, excessive ozonation caused deformities and even-
tual death. Tolerance to ozonation declined at later stages
of larval development. It is clear that ozonation needs to
be monitored and controlled precisely to deliver an effective
disinfection for the benefit of larval health, while avoiding
the detrimental toxins causing deformities and death. Thus,
a better understanding is needed of the changes in water
chemistry during ozonation and improved technologies are
required to accurately measure the levels of OBP in ozonated
seawater.
Improvement of Taste and Odor
The occurrence of off-flavors in all types of aquaculture
products is costly and continues to be a detriment to the
growth of the aquaculture industry. Off-flavors also contribute
to losses to the industry due to consumer dissatisfaction with
the cultured product, which can result in the decreased like-
lihood of future purchases and inhibit expansion into new
markets. Producers using recirculation or partial recirculation
systems, in which off-flavored fish are present, have instituted
depuration practices where by fish are held in clean water
until any off-flavor disappears from the product (Tucker et al.,
2000; Schrader et al., 2005).
Taste and odor of water is a common source of cus-
tomer complaints to water utilities (Suffet et al., 1995). The
odorants 2-methylisoborneol (MIB) and trans-1,10-dimethyl-
trans-9-decalol (geosmin) often cause earthy/musty odors.
The production of fish in recirculating aquaculture systems
(RAS) continues to be hampered by problems with envi-
ronmentally derived “off-flavor” that is best described as an
“earthy” and/or “musty” taste of the fillet. Recent studies
have determined that the presence of the odorous compounds
geosmin and MIB in the flesh of RAS-cultured fish is respon-
sible for these off-flavors (Schrader et al., 2005; Guttman and
van Rijn, 2008; Schrader and Summerfelt, 2010).
Blue-green algae and actinomycetes present in surface
waters produce MIB and geosmin, resulting in part-per-
trillion odorant concentrations in water supplies (Schrader
et al., 2005; Westerhoff et al., 2005). This evidence of MIB
and geosmin accumulating is also common in recirculating
systems (Masser et al., 1999; Schrader et al., 2005). So far,
few studies have addressed the possible causes and prevention
of off-flavor compounds’ accumulation in these latter systems
(Schrader et al., 2005; Guttman and van Rijn, 2009). The
biofilters used in the water recirculating systems for cultur-
ing the white sturgeon may be a source of geosmin-producing
actinomycetes. Research is ongoing to help determine the
sources of geosmin and MIB in these recirculating-water
systems (Schrader et al., 2005).
Ozone is widely used to achieve multiple water qual-
ity benefits (e.g., disinfection, trace organic removal, natural
organic matter removal), also effectively oxidizes odorants
such as MIB and geosmin (Westerhoff et al., 2006; Guttman
and van Rijn, 2008). Over the past several decades, various
municipal drinking water facilities in the United States have
used ozonation to remove geosmin and MIB from the water
via oxidation (Schrader et al., 2010).
Much of the information on MIB and geosmin oxidation
by ozone is dose-response relationships, which are inherently
difficult to extrapolate from one pilot study or water to other
locations. A number of factors can impact the efficiency of
ozone addition in the removal of geosmin and MIB including
dosage (Koch et al., 1992; Westerhoff et al., 2006), water tem-
perature (Westerhoff et al., 2006), organic matter content of
the water (Bruce et al., 2002; Ho et al., 2004), pH (Westerhoff
et al., 2006), and alkalinity (Ho et al., 2004).
354 A.A. Gonçalves and G.A. Gagnon September–October 2011
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
Variable removal (35% to 95%) of both MIB and geosmin
was observed during ozonation (1.5 mg L1) of surface waters
spiked with 50 ng.L1of MIB and geosmin; a higher ozone
dose(7mgL
1) removed >95% of the odorants (Lundgren
et al., 1988). Ozonation of Colorado River water (4 mgO3
L1) removed 78% of the MIB and 89%of the geosmin (Glaze
et al., 1990).
High odorant removal during advanced oxidation pro-
cesses suggests that hydroxyl radicals (HO) play an impor-
tant role in MIB and geosmin oxidation. Based on advanced
oxidation process (AOP) studies, second-order rate constants
have been estimated between HOand MIB or Geosmin,
but rate constants are not available for direct reactions of
the odorants with ozone (O3). Thus, Westerhoff et al. (2006)
quantified the relative importance of O3and HOoxidation
pathways and rate constants for MIB and geosmin. Kinetics
was evaluated for MIB and geosmin oxidation by ozone in
surface waters with a range of water qualities. In addition,
geosmin oxidizes faster than MIB (Westerhoff et al., 2006).
The resistance of MIB to oxidation by O3may be due to
greater steric hindrance based upon its chemical structure
compared to geosmin (Ho et al., 2004). MIB and geosmin oxi-
dation increases with greater ozone dose, higher pH, higher
temperature or addition of H2O2.
The use of ozonation in RAS has recently been studied to
determine the requirements to achieve adequate disinfection
of recirculating water (Summerfelt et al., 2009). In addi-
tion, previous studies have determined that ozone addition in
RAS will improve water quality by inducing microfloccula-
tion of fine particles [i.e., improving total suspended solids
(TSS) capture and reducing TSS concentration] and oxidiz-
ing nitrite (i.e., reducing nitrite concentration) and undesir-
able organic molecules (e.g., reducing non-biodegradable and
refractory compounds that stain the water) (Chen et al., 1994;
Summerfelt et al., 1997, 2009; Krumins et al., 2001a, 2001b;
Summerfelt, 2003).
Ozonation has been demonstrated to be effective in reduc-
ing geosmin and MIB concentrations in water. For example,
ozone dosages of 1, 2, and 4 mg/L at a contact time of 12 min
reduced an initial MIB concentration of 100 ng/L, in river
water, by 58%, 65%, and 75%, respectively (Koch et al.,
1992). In another study, Glaze et al. (1990) determined that
0.1 mg/L of ozone with a contact time of 20 min reduced ini-
tial geosmin and MIB levels of 100 ng/L in aqueduct water by
35% and 40%, respectively, and 0.2 mg/L of ozone (20 min
contact) reduced 100 ng/L of geosmin and MIB by 86–92%
and 73–83%, respectively (Schrader et al., 2010).
In the current study (Schrader et al., 2010), ozone addition
was at significantly lower levels than those used in the above
mentioned studies. The concentration and flow of ozone in the
feed gas supplied to one of the three RAS receiving ozona-
tion were measured and used to quantify that approximately
20–25 g of ozone were added to the recirculating flow for
every 1 kg of feed fed daily, which was equivalent to an ozone
dose of approximately 0.25–0.28 mg/L. This ozone dose was
used to maintain an ORP of 248 mV, which extrapolates to
less than 1µg/L of dissolved ozone residual according to data
published by Summerfelt et al. (2009); in fact, an ozone resid-
ual concentration of 1µg/L is not expected until ORP reaches
approximately 350 mV.
A dissolved ozone concentration below 1µg/Lissafe
for rainbow trout in freshwater (Bullock et al., 1997). Even
though the recirculated water was subjected to the continu-
ous addition of ozone, the ozone addition had no significant
effect (p >0.05) in reducing levels of geosmin and MIB in
the water or trout fillets compared to RAS with no ozone
addition. These results indicate that “low-dose” ozone addi-
tion with the intended goal of improving certain water quality
parameters (e.g., TSS, color, etc.) will not provide benefits in
the management of off-flavor problems related to geosmin and
MIB.
Ozone in Shellfish Depuration
Filter-feeding molluscan shellfish accumulate microorgan-
isms, such as bacteria and human viruses, when grown in
sewage-polluted waters and can present a significant health
risk when consumed raw or lightly cooked (Sobsey and
Jaykus, 1991). Current regulations of shellfish and their grow-
ing waters are based on bacterial standards (fecal coliforms
and Escherichia coli) and have prevented bacterial gastroin-
testinal infections. However, they are believed to have limited
predictive value for viral pathogens such as enteroviruses
(Jofre, 1992), Norwalk-like virus (NLV), hepatitis A virus
(HAV) and hepatitis E virus (HEV) (Wanke and Guerrant,
1990; Desenclos et al., 1991; Holliman, 2005a, 2005b).
Ozone is very effective at inactivating both bacteria and
viruses, and at a concentration not exceeding 0.5 mg L1(to
minimize bromate production) can be used to treat seawater
in batches for periods up to 10 min. This is undertaken in a
separate tank to that used for depuration and then the residual
ozone has to be discharged from the seawater before use so
that it does not adversely affect the animals this is achieved
by aeration. There are two additional concerns with the use
of ozone the first is that bromates are formed when ozone
is in contact with seawater and these are regarded as potential
cancer forming compounds. The second is that residual levels
of ozone may cause the shellfish to reduce or stop activity,
thus reducing the effectiveness of the depuration process (Lee
et al., 2008).
Several studies have revealed the differential rates of reduc-
tion of bacteria and viruses in depurating shellfish (Power and
Collins, 1989; De Mesquita et al., 1991; Dore and Lees, 1995)
and there is an urgent need for indicators of human-specific
viral fecal pollution to improve the biological safety of shell-
fish. Then, to evaluate the period of depuration that ensures
the microbiological safety of shellfish, Muniain-Mujika et al.
(2002) analyzed the depuration rates of various parameters in
naturally polluted shellfish mussel (Mytilus galloprovincial-
lis) using highly efficient depuration equipment. Seawater of
the depuration tank was disinfected by UV irradiation, ozone
and passed through a skimmer and a biological filter; and in
Overview of Ozone in Recirculating Aquaculture System September–October 2011 355
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
this specific depuration system, 5 days may be necessary to
assess the sanitary quality of shellfish.
Certain microalgae are also relatively insensitive to resid-
ual oxidants. The growth of the microalgae Tetraselmis chuii
was found to be unaffected at levels up to 0.7 mg L1;at1mg
L1, growth was impacted negatively (Zheng et al., 2002).
The use of ozone to depurate shellfish was used in Europe
as early as 1929 (Violle, 1929) and has been shown to reduce
toxicity of K. brevis toxins (Blogoslawski et al., 1973, 1975).
The use of ozone to detoxify shellfish or minimally, its utiliza-
tion to treat incoming seawater to a depuration or wet storage
facility could reduce the risk of cross contamination in areas
where a bloom might have occurred (Schneider et al., 2003).
As for other organisms, the damage to the American oyster
(Crassostrea virginica) by residual oxidants varied with their
age. Even for adults, fecal matter accumulation was reduced
at TRO levels as low as 0.05 mg L1(Richardson et al., 1982).
Removal of Algae and its Toxins
There have been several reports regarding the poten-
tial of use of ozone to kill algal cells (Sengco, 2009).
Deeds et al. (2004) found that 0.4 mg L1led to lysis of
Karlodinium micrum (>80% removal) and a reduction in
hemolytic activity in fish farm operation in the Chesapeake
Bay (Maryland, USA). Similarly, cultures of Prorocentrium
triestinum,Scrippsiella trochoidea, and Karenia diginata
were killed within 15 min after exposure to1gO
3m3(Ho
and Wong, 2004). The treatment also reduced concentrations
of ammonium and total inorganic nitrogen while dissolved
oxygen levels remained within acceptable levels.
Ozone was also tested against Karenia brevis and its
toxins (Scheider et al., 2003). Direct treatment of K. bre-
vis culture with 25 mg of ozone resulted in an 80% loss
of cells within 10 s. All of the cells destroyed after 60 s.
Similarly, free brevetoxins introduced into seawater were sig-
nificantly reduced after a 10-min treatment. However, 135 mg
of ozone was needed. The survival of Cyprinodon varie-
gates in fish bioassays was inversely related to the time
after ozone treatment, indicating a reduction in toxicity over
time.
Other important issue that might be considered is the
occurrence of red tide blooms which interrupts the culture,
production, harvesting and subsequently the marketing of
seafood products (Schneider et al., 2003). Shellfish toxicity
has also been reported during the occurrence of Florida
red tides and brevetoxins have been cited as the cause of
sub-lethal human intoxications known as neurotoxic shellfish
poisoning (NSP). Toxin levels normally found in shellfish
during Karenia brevis blooms can be fatal to humans,
although no deaths have been attributed to the consumption
of affected shellfish (Baden et al., 1984). Ozonated seawater
has been shown to be effective in inactivating crude toxins
associated with dinoflagellate blooms, as well as in reduc-
ing the levels accumulated in shellfish (Thurberg, 1975;
Blogoslawski et al., 1979).
Schneider et al. (2003) examined the effectiveness of ozone
to reduce the numbers of Florida red tide organism (Karenia
brevis Davis) and its associated toxins in an artificial seawater
environment. The reduction in toxin concentration, as mea-
sured by high performance liquid chromatography (HPLC)
analysis, displayed a positive correlation with the reduction
of toxicity as determined by a fish (Cyprinodon variegatus)
bioassay. Despite large total doses of ozone applied, as com-
pared to levels that might be found at a commercial ozonation
facility, some toxins were still recoverable by HPLC after
ozone treatment.
Improvement of Quality of Live Food
The economic profitability of larval rearing in marine fin-
fish hatcheries depends to a large extent on the continuous
availability of high quality live food. In this respect, the
demand for rotifers has gradually increased over the last years
which can be explained by the relatively stagnating Artemia
supply for an increased aquaculture production (Suantika
et al., 2003).
Artemia, Brachionus plicatilis, is an excellent first food for
fish and crustacean larvae, but there are still some problems
related to its culture and use. The most stringent problems for
automation of the production cycle reside in the unpredictabil-
ity of the mass production and the variability in the quality of
the product (Walz et al., 1997). The unstable production and
low quality of rotifers produced in commercial hatcheries can
mainly be explained by the static culture a procedure (batch
cultures) in which water quality is degrading rapidly (Suantika
et al., 2001, 2003).
Moreover, water quality was a first prerequisite for healthy
rotifer cultures but no attention was made to the hygienic
quality of rotifers (Yoshimura et al., 1994). However, it is
assumed that rotifers, the first food administered to fish lar-
vae, are the major carriers of bacteria causing poor survival
and growth of fish larvae (Munro et al., 1994). For this rea-
son, the reduction of bacteria and/or a method of controlling
bacterial populations in rotifer cultures need to be considered
in rotifer production units.
Suantika et al. (2001) evaluated the use of ozone as a dis-
infectant in a recirculation rotifer culture system as a tool to
improve water quality and reduce the bloom of opportunistic
bacteria. In general terms, it can be stated that supplemen-
tation of ozone in a closed recirculation system for rotifers
considerably improves water quality (the ammonium levels
were reduced by 67%, nitrite levels by 85% and nitrate lev-
els by 67%), ensures stable and longer rotifer culture periods
and controls bacterial proliferation.
Preliminary work (Tolomei, unpublished data) using ozone
as a disinfectant has shown similar positive results, i.e.,
Artemia appear resilient to ozone, and bacterial levels in
Artemia culture water were reduced by 99.9% within min-
utes of exposure to 4 mg L1ozone. To confirm these results,
Tolomei et al. (2004) assessed the efficacy of various com-
mercial and algal enrichment diets and ozone treatments to
356 A.A. Gonçalves and G.A. Gagnon September–October 2011
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
reduce bacterial load in the intestine and on external sur-
faces of on-grown Artemia. Direct exposure to ozone at 4 mg
L1for 5 min provided further bacterial reduction, result-
ing in a combined bacterial load reduction of 99.5% without
compromising Artemia viability.
Risks
There are two additional concerns with the use of ozone
the first is that bromate and bromoform form when ozone
is in contact with seawater and these are regarded as potential
cancer forming compounds (Lee et al., 2008). Disinfection by-
product (DBP) formation is associated with all disinfectants
and oxidants; however, the major DBP of concern when using
ozone is bromate (BrO3), a DBP that forms from naturally
occurring bromide (Br) in raw water (Legube et al., 2004).
When sufficient ozone has been transferred to create disin-
fecting ozone residual concentration at the end of the contact
chamber, then that residual will need to be removed before the
water reaches aquatic organisms in the culture tanks. Residual
ozone can be lethal to fish at concentrations as low as 0.01 mg
L1, but the actual concentration depends upon species and
life stage (Summerfelt, 2003). The fact that ozone reacts with
bromide (Br) and chloride (Cl) ions to form oxidants in sea-
water strongly suggests that direct extrapolation of the results
obtained in freshwater to applications in seawater is dangerous
(Sugita at al., 1996).
One of the main reasons that ozone is not widely used in
aquaculture is its toxicity and a manager’s unwillingness to
risk losing fish to an accidental overdose. Residual ozone is
highly toxic to fish at low levels (Bullock et al., 1997). Ozone
is reported to be toxic to a wide range of fresh and salt-water
organisms at residual concentrations between 0.01 mg L1
and 0.1 mg L1. When deciding where to introduce ozone
the effect of residual concentrations from the reactor on either
the biofilter or fish stocks should be carefully considered.
Direct treatment of the culture tank is not recommended. This
method carries a high risk of exposing fish stocks to residual
ozone concentrations (Read, 2008).
Residual oxidants vs. RAS health. Ozonation of estu-
arine (brackish) or marine waters can produce different
by-products oxidants and significant amount of TRO which is
highly toxic for aquatic organisms, i.e., bromate, bromoform,
etc. (Holmes-Farley, 2006). Ozone reacts with bromide and
chloride ions in saltwater to produce relatively stable oxidants
that are toxic to aquatic organisms. Use of ozone in saltwa-
ter systems is usually restricted to batch treatment of water
separate to the main recirculating flow.
The maximum safe level of chronic ozone exposure for
salmonids is 0.002 mg L1(Wedemeyer et al., 1979a, 1979b).
A compilation of results from several studies indicates that
most fish exposed to ozone concentrations greater than
0.008–0.06 mg L1will develop severe gill damage that can
result in serum osmolality imbalances and can kill fish imme-
diately or leave them more susceptible to microbial infections
(Bullock et al., 1997). Wedemeyer et al. (1979b) reported gill
epithelial damage and death of rainbow trout (Oncorhynchus
mykiss) exposed to 0.0093 mg O3L1.
Study conducted by Bullock et al. (1997), ozone was
added to water in a recirculating rainbow trout (Oncorhynchus
mykiss) culture system just before it entered the culture tanks
in an attempt to reduce the numbers of heterotrophic bacte-
ria in system water and on trout gills, and to prevent bacterial
gill disease (BGD) in newly stocked fingerlings. Rationale for
ozone’s success at preventing BGD mortalities were not fully
understood but may in part be due to improved water qual-
ity. Use of the lower ozone dosing rate (25 kg ozone/kg feed)
appeared to provide the same benefits as the higher dosing
rate (36–39 kg ozone/kg feed fed); however, the lower ozone
dosing rate was less likely to produce a toxic ozone residual
in the culture tank and would also reduce ozone equipment
capital and operating costs.
Recent study of histological evaluations in rainbow trout
(Oncorhynchus mykiss) conducted by Good et al. (2009)
revealed that fish in ozonated systems (RAS received water
ozonation to an ORP of 250 mV) were significantly more
likely to exhibit subclinical gill pathology (epithelial hyper-
plasia and hypertrophy) as well as hepatic lipidosis. The
findings of this study indicate that, despite an increase in spe-
cific subclinical pathologies, water ozonation in low-exchange
RAS can improve water quality and overall rainbow trout
performance as they are reared to market size.
As discussed previously (ORP results, Figure 3) oxidation-
reduction reactions occur during the disinfection process
(after ozonation), and the residual oxidants remained after
ozonation should be considered for aquatic organisms in RAS.
According to Cooper et al. (2002) ozone toxicity tests have
been conducted for several marine taxa, including microalgae,
invertebrates and vertebrates; and analytical measurements
taken in most tests were not specific to ozone, but rather are
expressed as TRO, or “ozone-produced oxidants”. Ozone tox-
icity is thus most correctly expressed as a function of TRO,
rather than O3per se.
Figure 4 showed final TRO values below to 2 mg L1in
sea water (salinity of 5‰) and zero for Milli-Q water (salinity
of 0‰). However, in sea water (salinity of 25‰) the TRO
values have remained between 8 and 10 mg.L1, which could
be dangerous to aquatic organisms.
Substantial mortality (i.e., 50–100% mortality) was
observed for microalgae, crabs and lobster at concentration
from 0.14–1.0 mg L1of TRO. Ozone toxicity tests with
striped bass and white perch were conducted using flow-
through test systems to deliver more reliable and consistent
ozone exposures. For striped bass, LC50s (i.e., concentration
that kills 50 % of the organisms) ranged from 0.06–0.2 mg
L1, depending on the life stage tested and length of expo-
sure. Eggs were the most sensitive life stage when reared
in freshwater (LC50 =0.06 mg L1), but fingerlings were
most sensitive in seawater if the test was run for 96 h (LC50
Overview of Ozone in Recirculating Aquaculture System September–October 2011 357
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
FIGURE 4. TRO measurements during ozonation (60 min) and after stop ozonation: 5 PSU (5 practical salinity units or 5‰), 25 PSU
(25 practical salinity units or 25‰).
=0.08 mg L1). Slightly higher concentrations (0.15–0.4 mg
L1) induced 100% mortality (i.e., LC100) to striped bass
fingerlings. In contrast to striped bass, TRO was slightly
less toxic to white perch with LC50 values ranging from
0.2–0.38 mg L1, and an LC100 of 0.8 mg L1after a 6-h
exposure (Cooper et al., 2002).
In aquaculture the risk of transmission of fish pathogens
via eggs is reduced by disinfection in ozonated seawater, but
this treatment may delay or reduce hatching. Disinfection of
the egg surface may reduce the negative effects of the micro-
bial load and offer a barrier to the transfer of pathogens
both between broodstock and their offspring and between
geographical regions. Egg disinfection may thus contribute to
a more stable production (Grotmol et al., 2003).
Larvae are, in general, more sensitive to TRO than are eggs
(Asbury and Coler, 1980) adults or juveniles (Ozawa et al.,
1991). Japanese flounder eggs were found to be impacted by
residual oxidants to the extent that 50% did not hatch after
1 min of exposure to 2.2 mg L1TRO. Larvae aged 3–15 days
were killed to the extent of 50% in 24 h at 0.02–0.05 mg
L1TRO. Larvae aged 44 days were killed to the extent of
50% in 24 h at 0.15 mg L1TRO. In this case, the larvae
were shown to have damage to their branchial tissues (Mimura
et al., 1998).
The eggs and larvae of Japanese whiting (Sillago japonica)
also have been tested for toxicity by residual oxidants. In this
case, half of the eggs and larvae died in about 24 h when
exposed to 0.18 and 0.23 mg L1TRO, respectively (Isono
et al., 1993). The oxidants formed during sea water ozona-
tion may also react with compounds in the eggshell (chorion),
altering its functional properties and thus possibly influenc-
ing hatchability. Then, Grotmol et al. (2003) investigated the
effects of disinfection with ozonated seawater on the hatcha-
bility of eggs of Atlantic cod, turbot and Atlantic halibut. Two
milligrams O3per liter for 2 min and lower exposures ought
to be sufficient to ensure an excess of oxidants for efficient
inactivation of fish pathogens while avoiding negative effects
on the hatchability of halibut, cod and turbot eggs.
Toxicity tests of residual oxidants on shrimp show them to
be less sensitive than fish. Penaeus chinensis and Paralichthys
olivaceus were found to live up to 48 h at TRO concentrations
of more than 1 mg L1, while Bastard halibut (fish) in the
same study lived only 3 h at 1 mg L1and 48 hours at 0.13 mg
L1(Jiang et al., 2001).
The effect of residual oxidants on rotifers (Brachionus pli-
catilis) has also been determined (Davis and Arnold, 1997).
No effect on survival was seen at less than 0.22 mg L1,
but effects became significant above that level. The authors
point out those bacteria and other pathogens can be killed at
that level, so rotifer cultures can be used with that amount of
continuous ozone to reduce bacterial contamination.
Leynen et al. (1998) investigate the possible adverse impact
of discharged dissolved ozone in terms the acute toxicity
to fish larvae (Cyprinus carpio,Leuciscus idus and Clarias
gariepinus) and Daphnia magna. Results indicate that ozone
is very toxic to aquatic organisms. Neonates of D. magna are
358 A.A. Gonçalves and G.A. Gagnon September–October 2011
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
more susceptible to ozone than are fish larvae. No major dif-
ference in toxicity of ozone for daphnids was observed at the
different test temperatures, and all three species of fish larvae
are similarly sensitive to ozone. The 48-h LC50 for fish larvae
C. carpio (at 27 C), L. idus (at 27 C), and C. gariepinus (at
32 C) ranges between 30 and 45 µgL
1, and the 48-h No
Observed Effect Concentration (NOEC) for D. magna is 11
µgL
1at 21 C and 16 µgL
1at a test temperature of 27 C.
Bromate and bromoform formation (BrO3vs. ozona-
tion time). In recent years, bromate has become known as
a contaminant of potable water supplies and in aquariums
due to its formation from naturally occurring bromide during
ozonation. Evidence supports the view that bromate is a pos-
sible human carcinogen and is therefore strictly controlled in
drinking water (Krasner et al., 1993b; Weinberg et al., 1993;
Bull and Cotruvo, 2006). Ozonation has become increasingly
important in water treatment across the world as an oxidiz-
ing agent and disinfectant due to its strong oxidation potential
(von Gunten, 2003a, 2003b; Jarvis et al., 2007).
Disinfection by-product (DBP) formation is associated
with all disinfectants and oxidants (Sohn et al., 2004; Jarvis
et al., 2007); however, the major DBP of concern when using
ozone is bromate (BrO3), a DBP that forms from natu-
rally occurring bromide (Br) in raw water (Legube et al.,
2004). Toxicity testing on experimental animals has consis-
tently shown bromate to induce cancer in rats, mice and
hamsters through damage to genetic material (Chipman et al.,
1998; Bull and Cotruvo, 2006).
The bromide ion (Br) is defined as an inorganic ion found
in surface water and ground water and caused by (i) sea intru-
sion, (ii) the impact of connate, or (iii) industrial and oil-field
brine discharge (Symons, 1999). When oxidized by chlorine
(Cl2or HOCl) or ozone (O3), it can result in the forma-
tion of organic and inorganic bromine-substituted disinfection
by-products.
The bromate ion (BrO3) is the highest oxidation state
of the bromide ion. The bromate ion can be formed during
the ozonation of bromide-containing waters (Symons, 1999).
Bromate is formed through a complex web of pathways with
several bromine-containing intermediates that undergo reac-
tions with both molecular ozone and hydroxyl radicals (Glaze
and Weinberg, 1999). Ozone oxidizes bromide to form hypo-
bromite ion (OBr). Hypobromite continues to be oxidized
to form bromate or to form an unidentified species, possi-
bly BrO2that regenerates bromide ion (Glaze and Weinberg,
1999).
Brominated by-product formation in ozonated waters is
influenced by bromide ion concentration, the source and con-
centration of natural organic matter (NOM), pH, ozone dose,
and reaction time. It is important to note that ozonation under
higher pH conditions produces higher bromate concentra-
tions, such that with sufficient bromide and ozone applied to
meet an ozone residual for disinfection; tens of micrograms
per liter of bromate can be formed (Faust and Aly, 1998).
Ozone reacts with bromide ions in brackish and seawater
systems to form the oxidants hypobromous acid (HOBr) and
hypobromite ion (OBr), which is relatively stable and toxic
to fish and shellfish (Blogoslawski and Perez, 1992; Keaffaber
et al., 1992). Prolonged ozonation can further oxidize hypo-
bromite ion to bromate (BrO3), which is another persistent
and toxic compound (Marhaba and Bengraïne, 2003; Jarvis
et al., 2007).
The bromate ion cannot be further oxidized and will be
the final product of the oxidation of bromide ion in seawa-
ter. Bromate ion is a stable compound and not acutely toxic
to aquatic animals (Marhaba and Bengraïne, 2003; Liltved
et al., 2006; Jarvis et al., 2007). Unfortunately, the produc-
tion conditions and toxicity towards aquatic animals of these
ozonation by-products are not well understood (Summerfelt,
2003).
Bromate formation has been the major barrier in the use
of ozone for water treatment where the source water con-
tains bromide, particularly given the challenging targets set for
the maximum allowable bromate concentration (Magazinovic
et al., 2004). The formation of bromate during ozonation
is strongly dependent on the characteristics of the water to
be treated and the amount of ozone contacting the water.
The following are important variables for bromate formation:
bromide concentration, pH, applied ozone concentration and
contact time, DOC concentration, alkalinity, ammonia con-
centration, and temperature (Marhaba and Bengraïne, 2003;
Sohn et al., 2004; Jarvis et al., 2007).
The most stable by-products of seawater ozonation typ-
ically are bromate ion and bromoform and both may per-
sist long after ozone treatment is terminated (Cooper et al.,
2002). However, the limited available toxicity data set sug-
gests that these compounds are not acutely toxic with LC50
values 1–2 orders of magnitude higher than either TRO or
bromine.
The data for the bromate and bromoform in the three exper-
iments are presented in Figure 5. For sea water (25‰) the
concentration of bromate and bromoform increased with the
time of ozonation and remained at higher levels. For lower
water salinity (5‰) the bromate formation was lower and bro-
moform formation increase slowly during ozonation time and
remained constant to the end of the experiment (lower than
the salinity of 25‰). These results showed the importance of
the control of by-product formation during the ozonation for
aquatic organisms, and these results should be considered for
others studies and/or applications in RAS.
The most sensitive species to bromate ion is the mysid
shrimp Neomysis awatschensis with an acute LC50 of 176 mg
bromate ion L1, and the most sensitive species to bromoform
is the sheepshead minnow with 96-h LC50 values ranging from
7.1–18 mg bromoform L1. Therefore, even if bromate ion
and/or bromoform are produced as by-products of seawater
ozonation, they are not likely to be of toxicological concern
(Cooper et al., 2002).
Bromate formation and control has been the focus of inten-
sive research efforts since the early 1990s when bromate
Overview of Ozone in Recirculating Aquaculture System September–October 2011 359
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
FIGURE 5. Bromate (BrO3) and Bromoform (CHBr3) measurements during ozonation (60 min) and after stop ozonation: 5 PSU (5 practical
salinity units or 5‰), 25 PSU (25 practical salinity units or 25‰).
(BrO3) was implicated as a potential carcinogen (Buffle
et al., 2004). Whilst there is no data demonstrating that
bromate is carcinogenic to humans, it is plausible to assume
that the mechanisms resulting in tumor formation in labo-
ratory animals could occur in humans. For this reason the
World Health Organization (WHO) has set a provisional
guideline concentration of 10 µgL
1(0.01 mg L1)bromate
in drinking water (WHO, 2004). European Union law speci-
fies that all member states must enforce a maximum bromate
concentration of 10 µgL
1by 2008 (European Drinking
Water Directive, 1998). In the United Kingdom, the legislation
enforcing this standard came into effect in 2003. In the United
360 A.A. Gonçalves and G.A. Gagnon September–October 2011
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
States (United States Environmental Protection Agency),
regulations also specify a maximum value of 10 µgL
1
(USEPA, 1998).
Few studies have examined the toxicity of excess bromate
itself to marine organisms (Hutchinson and van Wijk, 1998).
Bromate is classified as carcinogenic to human health by the
IARC (International Agency for the Research on Cancer) and
the USEPA (United States Environmental Protection Agency)
and is a known toxin to fish and other aquatic life, causing
respiratory and osmoregulatory dysfunction (Grguric et al.,
1994). It is also probably toxic to crustaceans such as phyl-
losoma larvae (Ritar et al., 2006).
Bromate toxicity tests on marine animals indicate the levels
of bromate produced by chlorination or ozonation of power
plant cooling waters are not acutely toxic. The LC50 ranged
from 30 mg L1bromate for Pacific oyster (Crassostrea
gigas) larva to several hundred mg L1for fish, shrimp and
clams (Lugo-Fernandez and Roscigno, 1999). Toxicity studies
carried by Burton and Richardson (1981) showed that the con-
centrations of bromate which theoretically could be formed in
an ozonated discharge were not toxic to the early life stages of
striped bass (Morone saxatilis) and juvenile spot (Leiostomus
xanthurus).
An individual study showed that Pacific oysters
(Crassostrea gigas) had abnormal larval development at
bromate levels of 30–300 mg L1(Hirtle and Mann, 1978).
Fertilized eggs of the oyster Crassostrea virginica were killed
at 1 mg L1(Stewart et al., 1979). The clams Protothaca
staminea (littleneck) and Macoma inquinata (bent-nosed)
were killed by 880 mg L1(Hirtle and Mann, 1978).
The marine dinoflagellate Glenodinium halli showed
changes in population growth at 16 mg L1; the marine
microalgae Isochrysis galbana showed changes in popula-
tiongrowthat8mgL
1; the marine diatom (Skeletonema
costatum) showed changes in population growth at 0.125 to
16 mg L1; and the marine diatom Thalassiosira pseudo-
nana showed changes in population growth at 16 mg L1
(Erickson and Freeman 1978). The salmon Oncorhynchus
keta was killed at 500 mg L1and the perch Cymatogaster
aggregata at 880 mg L1and shrimps (Pandalus danae and
Neomysis awatschensis) were killed at 880 and 176 mg L1,
respectively (Hirtle and Mann, 1978).
Although there are limited data for several important tax-
onomic groups, the majority of available data suggest that
bromate is non-toxic to many aquatic organisms with E(L)C50
values being generally greater than 100 mg BrO3L1.
Although the toxicity of bromate has been addressed using
a relatively wide range of organisms (4 algae, 8 invertebrates,
and 4 fish species), the most sensitive organisms appear to
be marine fish larvae, with a 96-h LC50 of 31 mg BrO3
L1(Richardson et al., 1981b). For toxicity data from this
range of aquatic species, it is customary to apply a factor
of 10 to extrapolate from an acute to a safe chronic level
of a given substance. On this basis, the most conservative
data available (marine fish larvae) suggest that to protect
aquatic organisms from long-term adverse effects, surface
water concentrations should not exceed approximately 3.0 mg
BrO3L1(Hutchinson et al., 1997).
Removal of disinfection by-products (DBP) residuals.
The minimization of DBPs can be approached differently.
First, it would be logical to remove the precursors and second,
to control the BrO3formation (Grecelius, 1977; Richardson
et al., 1981a; Wajon and Morris, 1982; Haag and Hoigné,
1983, 1984; Krasner et al., 1991; Siddiqui and Amy, 1993;
von Gunten and Hoigné, 1992, 1994; von Gunten et al.,
1993). This can be done by pH depression, ammonia addi-
tion, hydrogen peroxide addition, or modifications to ozone
contactor design and operation. However, these strategies are
not effective especially in the presence of natural organic mat-
ter (NOM). The alternative would be to eliminate BrO3once
formed (Marhaba and Bengraïne, 2003).
Legube et al. (1995) showed that pH, Brconcentration,
and O3dosage control the formation of BrO3. Higher pH and
temperature favor the formation of BrO3due to the decrease
in the pKa of HOBr and OBr. Therefore, O3concentra-
tion should be kept low and contact time extended (Marhaba
and Medlar, 1993). Krasner et al. (1993b) concluded that the
appropriate staging of O3through two or three chambers has
the potential to minimize O3residual and BrO3formation
while still meeting the CT criterion. Further, as the pH of
ozonation was lowered, the O3dosage necessary to meet the
CT criterion dropped and less BrO3was produced (Krasner
et al., 1993a).
Due to the acute toxicity of residual ozone and others oxi-
dants (i.e., bromine) to aquatic organisms, a treatment unit for
removal of residual bromine has to be included when seawater
is ozonated in aquacultural systems. Activated carbon filtra-
tion, addition of a reducing agent, UV radiation, or passing
through a sand filter or biofilter or by air stripping prior to
fish cultural systems or discharge to surface waters will reduce
or eliminate these residues (Liltved et al., 1995; Summerfelt,
2003; Summerfelt et al., 2004; Liltved et al., 2006; Read,
2008).
Summerfelt et al. (2004) determined the ultraviolet (UV)
irradiation dosages required o destroy dissolved ozone in a
commercial-scale recirculating salmonid culture system oper-
ated at a constant 13–15 C. The results showed that dissolved
O3removal across the UV irradiation unit could be modeled
using first-order kinetics and was dependent upon the inlet
O3concentration and the retention time within the irradiation
chamber. At a temperature of 13–15 C, UV irradiation doses
of 80.4±2.6 mW s cm2and 153.3±2.1 mW s cm2consis-
tently removed 100% of the dissolved O3when the inlet O3
concentration was 0.30 mg L1.
Sharrer and Summerfelt (2007) assessed the degree of total
heterotrophic and total coliform bacteria inactivation using
ozone alone (at several ozone dosages) and to determine if
a synergistic effect is seen in the disinfection of microorgan-
isms from process water in a fully recirculating fish culture
system when UV irradiation is applied directly after ozona-
tion. Combining ozone dosages of only 0.1–0.2 min mg L1
Overview of Ozone in Recirculating Aquaculture System September–October 2011 361
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
with a UV irradiation dosage of 50 mJ cm2provides an
advanced oxidation process that could consistently produce
post-treatment water nearly free from total coliform and total
heterotrophic bacteria colony forming units.
CONCLUSION
Ozone is a powerful oxidizing agent that has seen wide use
in aquaculture applications for achieving both disinfection and
water quality improvements. Ozone is added to aquaculture
system waters to inactivate fish pathogens, oxidize organic
wastes (including color) and nitrite, or supplement the effec-
tiveness of other water treatment units like UV light, filters
and biofilters.
Care must be used when determining the effective ozone
dose that must be supplied to achieve disinfection. Certain
pathogens may require much higher ozone cxtvalues in order
to achieve inactivation. Applying ozone to disinfect aquacul-
ture system influents or effluents can be quite complex and
costly, yet disinfection is still necessary in many situations to
control pathogen introduction.
For maximum benefit from ozonation, the user have to
first know the purpose of applying ozone so as to deter-
mine the correct techniques and ozone dosages. It is critical
that all ozone disinfection systems are pilot tested and cal-
ibrated prior to installation to ensure the true output and
concentration.
Further researches are required into the potential
disinfectant-by-product risks associated with ozone and
their impacts on aquatic health.
ACKNOWLEDGMENTS
We thank Ms. Heather Daurie for her kind laboratory
assistance. The present study was carried out with the sup-
port of Foreign Affairs and International Trade Canada
(DFAIT), which are also acknowledged. The authors also
acknowledge funding support from NSERC to support these
activities.
LIST OF ABBREVIATIONS
AHNV Atlantic Halibut Nodavirus
AOP Advanced Oxidation Process
BGD Bacterial Gill Disease
BOD5Biological Oxygen Demand
BrBromide ion
BrO3Bromate ions
COD Chemical Oxygen Demand
CT value ozone concentration (C), exposure
time (T)
DBP Disinfection By-Product
DOC Dissolved Organic Carbon
DPD N,N-diethyl-p-phenylenediamine
FE Fish Encephalopathy
GEOSMIN Trans-1,10-dimethyl-trans-9-decalol
HAV Hepatitis A Virus
HEV Hepatitis E Virus
HOBr Hypobromous acid
HPLC High Performance Liquid Chromatography
IPNV Infectious Pancreatic Necrosis Virus
ISAV Infectious Salmon Anaemia Virus
LC100 100% lethal concentration
LC50 50% lethal concentration
MIB 2-methylisoborneol
NLV Norwalk Like Virus
NOM Natural Organic Matter
NSP Neurotoxic Shellfish Poisoning
OBP Ozonation By-Products
OBrHypobromite ion
OPO Ozone Produced Oxidants
ORP Oxidation/Reduction Potential
RAS Recirculating Aquaculture System
ROC Residual Ozone Concentration
SJNNV Striped Jack Nervous Necrosis Virus
TAN Total Ammonia Nitrogen
TOC Total Organic Carbon
TRO Total Residual Oxidants
TSS Total Suspended Solids
TSV Taura Syndrome Virus
USEPA U.S. Environmental Protection Agency
VER Viral Encephalopathy and Retinopathy
VNN Viral Nervous Necrosis
WHO World Health Organization
WSBV White Spot Syndrome Bacilovirus
WSSV White Spot Syndrome Virus
YHV Yellow Head Virus
REFERENCES
American Public Health Association (APHA), Standard Methods for
Examination of Water and Wastewater”, 19th Ed. (APHA: Washington,
DC, 1995).
Arimoto, M., K. Mushiake, Y. Mizuta, T. Nakai, K. Muroga, and I. Furusawa,
“Detection of Striped Jack Nervous Necrosis Virus (SJNNV) by Enzyme-
Linked Immunosorbant Assay (ELISA)”. Fish Pathol., 27:191–195
(1992).
Arimoto, M., J. Sato, K. Maruyama, G. Mimura and I. Furusawa, “Effect
of Chemical and Physical Treatments on the Inactivation of Striped Jack
Nervous Necrosis Virus (SJNNV)”, Aquaculture, 143:15–22 (1996).
Asbury, C. and R. Coler, “Toxicity of Dissolved Ozone to Fish Eggs and
Larvae”, Water Poll. Cont. Feder., 52(7):1990–1996 (1980).
Baden, D.G., G. Bikhazi, S.J. Decker, F.F. Foldes, and I. Leung,
“Neuromuscular Blocking Action of Two Brevetoxins from the
Florida Red Tide Organism Ptychodiscus brevis”, Toxicon, 22(1):75–84
(1984).
Bai, C., “President’s Column”, World Aquacult. Mag., 38(4):3 (2007).
Ben-Atia, I., S. Lutzky, Y. Barr, K. Gamsiz, Y. Shtupler, B. Koven, and A.
Tandler, “The Effect of Ozone Treatment on Egg and Larvae Performance
in the Gilthead Seabream (Sparus aurata) and Other Marine Species”,
Euro. Aquacult. Soc. Spec. Publ., 30:58–59 (2001).
362 A.A. Gonçalves and G.A. Gagnon September–October 2011
Downloaded by [Alex Augusto Gonçalves] at 08:57 22 September 2011
Blogoslawski, W., C. Brown, D. Lewis, and J. Leong, “Use of Ozone for
Crustacean Disease Prevention: A Brief Review”, Proceedings of the
Second Biennial Crustacean Health Workshop, Galveston, TX, USA, pp.
220–230 (1977).
Blogoslawski, W.J. and C. Perez, “Ozone Treatment of Seawater to
Control Vibriosis in Mariculture of Penaeid Shrimp, Penaeus vannameii”,
In: Blogoslawski, W.J. (ed.), Proceedings of the Third International
Symposium on the Use of Ozone in Aquatic Systems. Pan American Group
of the International Ozone Association, Stamford, CT, pp. 123–133
(1992).
Blogoslawski, W.J., M.E. Stewart, J.W. Hurst Jr., and F.G. Kern III, “Ozone
Detoxification of Paralytic Shellfish Poison in the Softshell Clam (Mya
arenaria)”. Toxicon, 17:650–654 (1979).
Blogoslawski, W.J., F.P. Thurberg, and M.A. Dawson, “Ozone Inactivation of
aGymnodinium breve Toxin”, Water Res., 7:1701–1703 (1973).
Blogoslawski, W.J., F.P. Thurberg, M.A. Dawson, and M.J. Beckage, “Field
Studies on Ozone Inactivation of a Gymnodinium breve Toxin”, Environ.
Lett., 9(2):209–215 (1975).
Bonacquisti, T.P., “A Drinking Water Utility’s Perspective on Bromide,
Bromate, and Ozonation”, Toxicology, 221(2–3):145–148 (2006).
Bruce, D., P. Westerhoff, and A. Brawley-Chesworth, “Removal of 2-
Methylisoborneol and Geosmin in Surface Water Treatment Plants in
Arizona”, J. Water Supply Res. Technol.- AQUA, 51:183–197 (2002).
Buchan, K.A.H., D.J. Martin-Robichaud, and T.J. Benfey, “Measurement of
Dissolved Ozone in Sea Water: A Comparison of Methods”, Aquacult.
Eng., 33(3):225–231 (2005).
Buchan, K.A.H., D.J. Martin-Robichaud, T.J. Benfey, A.M. MacKinnon,
and L. Boston, “The Efficacy of Ozonated Seawater for Surface
Disinfection of Haddock (Melanogrammus aeglefinus) Eggs Against
Piscine Nodavirus”. Aquacult. Eng., 35:102–107 (2006).
Buffle, M.O., S. Galli, and U. von Guten, “Enhanced Bromate Control During
Ozonation: The Chlorine-Ammonia Process”, Environ. Sci. Technol.,
38(19):5187–5195 (2004).
Bull, R.J. and J.A. Cotruvo, “Research Strategy for Developing Key
Information on Bromate’s Mode of Action”, Toxicology, 221(2–3):135–
144 (2006).
Bullock, G.L., S.T. Summerfelt, A.C. Noble, A.L. Weber, M.D. Durant,
and J.A. Hankins, “Ozonation of a Recirculating Rainbow Trout Culture
System. I. Effects on Bacterial Gill Disease and Heterotrophic Bacteria”,
Aquaculture, 158:43–55 (1997).
Burton, D.T. and L.B. Richardson, “An Investigation of Chemistry and
Toxicity of Ozone-Produced Oxidants and Bromate to Selected Estuarine
Species”, EPA-600/4-81-040. U.S. Environmental Protection Agency,
Environmental Research Laboratory, Gulf Breeze, FL. 88 p., Avail. from
NTIS, Springfield, VA. Project number: PB82–116351, (1981).
Chang, P.S., L.J. Chen, and Y.C. Wang, “The Effect Of Ultraviolet
Irradiation, Heat, pH, Ozone, Salinity and Chemical Disinfectants on the
Infectivity of White Spot Syndrome Bacilovirus”, Aquaculture, 166:1–17
(1998).
Chen, S., D. Stechey, and R.F. Malone, “Suspended Solids Control in
Recirculating Aquaculture Systems”, In: Timmons, M.B. and Losordo,
T.M . (E ds .) , Aquaculture Water Reuse Systems: Engineering Design and
Management, pp. 61–100 (Elsevier, New York, NY, 1994).
Chipman, J.K., J.E. Davies, J.L. Parsons, J. Nair, G. O’Neill, and J.K. Fawell,
“DNA Oxidation by Potassium Bromate: A Direct Mechanism or Linked
to Lipid Perioxidation”, Toxicology 126: 93–102 (1998).
Colberg, P.J. and A.J. Lingg, “Effects of Ozonation on Microbial Fish
Pathogens, Ammonia, Nitrate, Nitrite and BOD in Simulated Reuse
Hatchery Water”, J.Fish.Res.Brd.Can., 35:1290–1296 (1978).
Coler, R.A. and C. Asbury, “Acute Toxicity of Dissolved Ozone to Eggs and
Larvae of Selected Freshwater Fish Species”, Ozone: Sci. Eng., 2:177–
182 (1980).
Coman, G.J., M.J. Sellars, and D.T. Morehea, “Toxicity of Ozone Generated
From Different Combinations of Ozone Concentration (C) and Exposure
Time (T): A Comparison of the Relative Effect of C and T on Hatch Rates
of Penaeus (Marsupenaeus)japonicus Embryos”, Aquaculture, 244:141–
145 (2005).
Conrad, J.F., R.A. Holt, and T.D. Kreps, “Ozone Disinfection of Flowing
Water”, Progress. Fish-Cultur., 37:134–136 (1975).
Cooper, W.J., G.M. Dethloff, R.W. Gensemer, M.L. House, R.A. Mueller,
G.M. Ruiz, W.A. Stubblefield, J.R. Cordell, P.A. Dinnel, R.P. Herwig,
J.A. Kopp, J.C. Perrins, G.M. Snevil, and E.V. Wende, “Full-Scale
Ozone Ballast Water Treatment for Removal of Marine Invasive Species”,
University of North Carolina, Department of Chemistry and Center for
Marine Science June 13, 2002. Available at http://www.nutech-o3.com/
files/2002june15_finalreport.pdf - Access 14.04.08 (2002).
Crisp, L.M. and C.E. Bland, “Potential Use of Ozone to Disinfect Sea Water
of Fungi Causing Diseases of Cultured Marine Crustacean”, J. Invert.
Pat hol., 55:380–386 (1990).
Davis, D.A. and C.R. Arnold, “Tolerance of the Rotifer Brachionus plicatilis
to Ozone and Total Oxidative Residuals”, Ozone: Sci. Eng., 19(5), 457–
469 (1997).
Deeds, J.R., A.P. Mazzaccaro, D.E. Terlizzi, and A.R. Place, “Treatment
Options for the Control of the Icthyotoxic Dinoflagellate Karlodinium
micrum in an Estuarine Aquaculture Facility: A Case Study”, In: Hall,
S. et al. (eds), Harmful Algae Management and Mitigation, Asia Pacific
Economic Cooperation, Singapore, p. 177–181 (2004).
De Mesquita, M.M.F., L.M. Evison, and P.A. West, “Removal of Faecal
Indicator Bacteria And Bacteriophages from the Common Mussel
(Mytilus edulis) under Artificial Depuration Conditions”, J. Appl.
Bacteriol., 70:495–501 (1991).
Desenclos, J.C., K.C. Klontz, M.H. Wilder, O.V. Naiman, H.S. Margolis,
and R.A. Gunn, “A Multistate Outbreak of Hepatitis A Caused by the
Consumption of Raw Oysters”, Amer. J. Public Health, 81:1268–1272
(1991).
Dixon, M.G., T. Hecht, and C.R. Brandt, “Identification and Treatment of a
Clostridium and Vibrio infection in South Africa Abalone, Haliotis midae
L.” J. Fish Dis., 14:693–695 (1991).
Dore, W.J. and D. Lees, “Behaviour of Escherichia coli and Malespecific
Bacteriophage in Environmentally Contaminated Bivalve Molluscs
Before and After Depuration”, Appl. Environ. Microbiol., 61:2830–2834
(1995).
Duvivier, L., M. Leynen, A. van Damme,