Wastewater Disinfection by Ozone: Main Parameters For Process Design

Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

Water Research 36 (2002) 1043–1055

Wastewater disinfection by ozone: main parameters for


process design
Pei Xua, Marie-Laure Janexb, Philippe Savoyeb, Arnaud Cockxc,
Valentina Lazarovab,*
a
Laboratoire Hydrosciences, MSE, UMR no. 5569, Universite! Montpellier II, 34095 Montpellier cedex 5, France
b
Ondeo Services-CIRSEE, 38 Rue du President
! Wilson, 78230 Le Pecq, France
c
Laboratoire GPI, INSA, 135 avenue de Rangueil, 31077 Toulouse cedex 4, France
Received 1 March 2000; accepted 30 May 2001

Abstract

Wastewater disinfection by ozone was investigated at pilot scale on different wastewater effluents. Variations in
operating conditions showed that a very low hydraulic retention time (2 min) was sufficient for efficient fecal coliform
inactivation, provided a sufficient ozone dose was transferred to the effluent. Therefore, the transferred ozone dose
appeared to be the critical parameter for the design of wastewater disinfection. As a consequence, the ‘‘Ct’’ approach
commonly applied in drinking water treatment should not be used for wastewater ozonation. Design parameters of
ozonation were proposed for two types of regulations, and for effluents of different qualities. It was demonstrated that
only with an efficient filtration step one can meet stringent standards such as the California Title 22 criteria. In all cases,
viruses were totally inactivated; consequently, viruses do not constitute a limiting factor in wastewater disinfection by
ozone.
The standard drinking water model failed to match the experimental data obtained on real wastewater effluents.
A modified approach was successfully developed, based on the simultaneous consumption of ozone by the
microorganisms and the organic matrix. r 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Wastewater disinfection; Ozone; Process design; Water quality; Wastewater reuse; Disinfection modelling

1. Introduction reliability of wastewater disinfection, which is the most


important treatment process for public health protec-
Wastewater reuse has become an attractive option for tion. The health-related microbiological regulations [3]
protecting the environment and extending available and the more recent impetus of producing virus-free
water resources. In the last few years, there has been a effluents [4] require the development of highly effective
significant diversification of water reuse practices, such advanced disinfection processes. Chlorination is still the
as green space and crop irrigation, recreational im- most widely used means to inactivate pathogenic
poundment, various urban uses including toilet flushing, microorganisms in water and wastewater, but alterna-
industrial applications and water supply augmentation tive technologies have to be evaluated because of
through groundwater or reservoir recharge [1,2]. The increasing concern over undesirable byproducts after
safe operation of water reuse systems depends on the chlorination and its inefficiency in eliminating some
epidemic microorganisms at low chlorine doses [5,6].
*Corresponding author. Tel.: +33-134-802-251; fax: +33- Ozone has been proved to be one of the most effective
130-536-207. disinfectants and is widely used to inactivate pathogens
E-mail address: valentina.lazarova@lyonnaise-des-eaux.fr in drinking water, especially in Europe [7,8]. Design
(V. Lazarova). engineers in the US began to evaluate ozone for

0043-1354/02/$ -see front matter r 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 3 - 1 3 5 4 ( 0 1 ) 0 0 2 9 8 - 6
1044 P. Xu et al. / Water Research 36 (2002) 1043–1055

wastewater disinfection in the early 1970s. However, ozone disinfection performance on different target
because of operational and maintenance problems that microorganisms.
appeared in the first generation of facilities, it has been
considered to be a less attractive alternative to chlorine
than UV disinfection. Also, many researchers initially 2.1. Experimental set-up: continuous-flow ozone pilot
sought to achieve a measurable level of dissolved ozone plant
residual in treated wastewater, which resulted in high
ozone dosages that were not economically feasible [9]. Ozonation tests were conducted in two different pilots
Earlier studies pointed out the need for a thorough designed as bubble diffuser columns (Fig. 1), which were
investigation of wastewater ozone treatment in order to operated in continuous counter-current mode. Ozone
predict disinfection performance and design the disin- between 2% and 6% was generated from oxygen (95–
fection system for wastewater disinfection [10–12]. 98% purity), with generators provided by Ozonia. The
The present study investigates the main factors related specifications of the ozone generators and contactors
to ozone disinfection performance, for the purpose of used and the operating conditions are given in Table 1.
facilitating its design and application to wastewater Tracer test studies were performed by impulse
disinfection. injections of sodium chloride in the different operating
conditions on both pilots. This made it possible to
characterize the pilots as a series of two to five CSTRs.
2. Materials and methods Salt recovery showed that dead zones were between 5%
and 12% for the smaller hydraulic retention times
Experiments were performed in a continuous-flow (HRT), and up to 25–30% for the highest HRT (10 min
pilot plant with different types of effluents to evaluate in pilot 1, and 15 min in pilot 2). This was taken into

Fig. 1. Schematic diagram of the continuous-flow ozone pilot plant.


P. Xu et al. / Water Research 36 (2002) 1043–1055 1045

Table 1
Characteristics of the ozone pilots and operating conditions used in the study

Parameter Pilot 1 Pilot 2

(Evry, France/Washington, UK) (Indianapolis, USA)

Column height, m 2.6 3.6


Column diameter, m 0.15 0.30
Porous plate porosity, mm 100 50
Hydraulic retention time, min 2–10 3–15
Number of CSTRs in seriesa 2–5 2
Applied ozone dose, mg L1 3–16/4–50 1–35
Transferred ozone dose, mg L1 2–13/4–30 0.5–12
a
Obtained by tracer tests in the different operating conditions.

account to calculate real contact times when data were an infinite rate constant
expressed as a function of ‘‘Ct’’. d½O3 
The ozone concentrations in the influent gas and off- ¼ KL að½O3  *  ½O3 Þ  kD ½O3  if TOD > X;
dt
gas were measured by UV absorbance monitors.
Residual ozone concentration in water was analysed d½O3 
by the indigo carmine method of [13], using HACH DR/ ½O3  ¼ 0 and ¼0 if TODoX; ð2Þ
dt
2000 spectrophotometer. The mass transfer efficiency in
the columns was found to be consistently 55% and 30– where KL a is the mass transfer coefficient; [O3]* is the
50%, respectively, in the pilots 1 and 2. In order to equilibrium (maximum) concentration of dissolved
compare the data, the transferred ozone dose (TOD) ozone, [O3], corresponding to Henry’s law; kD is the
was used as a descriptive parameter throughout the first order decay constant, in min1. TOD is the
study. It is defined as follows: transferred ozone dose calculated by
Z t
TOD ¼ Qgas =Qliq ð½O3 g in  ½O3 g out Þ; ð1Þ TOD ¼ KL að½O3  *  ½O3 Þ dt: ð3Þ
0

where Qgas and Qliq are gas and water flow rates, Eq. (2) can easily be solved in an open, completely mixed
respectively, [O3]g in is the ozone concentration in the reactor (CSTR with a contact time t) to estimate the
feed gas to the column, [O3]g out is the ozone concentra- ozone concentration
tion in the off-gas leaving the column. For a given set of TOD  X
½O3  ¼ : ð4Þ
operating conditions, a time interval of three to four 1 þ kD t
times the mean hydraulic retention time was allowed to
reach steady state and take samples. In order to study
the effect of post-contact without additional ozone 2.3. Wastewater characterisation
introduction, water samples were taken from the outlet
of the column and were kept in a brown bottle without Effluents from three different wastewater treatment
light or air, to be analysed after a controlled contact plants (WWTP) were used for the studyFtwo second-
time. ary effluents and one tertiary effluent. The secondary
treatment trains in Evry, France (48,000 m3 d1) and
Washington, UK (90,000 m3 d1) are similar: pretreat-
2.2. Determination of ozone demand ment, primary clarification, activated sludge (extended
aeration and high rate activated sludge, respectively),
The immediate ozone demand of the effluents, X; was secondary clarification and discharge. The tertiary
evaluated from the plots of the ozone residual vs. the treatment train in Indianapolis, USA (300,000 m3 d1)
transferred ozone dose during disinfection tests, accord- consists of dual media filtration and chlorination after
ing to the method proposed by Roustan et al. [14]. X primary clarification, and coupled bio-roughing and
represents the minimum dose to be transferred to get activated sludge nitrification facilities. The effluent for
measurable ozone residual in the water. Mathematically, pilot testing was taken after the tertiary filtration. The
the concept is represented by the following equations, main characteristics of the effluents during the tests are
where the ‘‘instantaneous’’ demand would correspond to given in Table 2.
1046 P. Xu et al. / Water Research 36 (2002) 1043–1055

Table 2
Wastewater characteristics, average (min–max)

Parameter Tertiary effluent Secondary effluent

Indianapolis (USA) Evry (France) Washington (UK)


1
Suspended solids (mg L ) 2.3 (o1–4) 5 (3–6) 18 (7–33)
COD, (mg O2 L1) 30 (24–38) 36 (26–56) 71 (41–150)
TOC (mg L1) 8 (5.5–10.2) o10 (o10–14) 26 (o11–30)
UV 254 abs (m1) 15.5 (12.5–20.8) 22.2 (17.4–20.8) 34.9 (26.0–50.9)
pH 7 (6.9–7.2) 7.3 (7.3–7.4) 7.5 (7.4–8.0)
Fecal coliforms (log CFU per 100 mL) F 3.6–4.5 4.3–6.5
E. coli (log CFU per 100 mL) 2.7–4.3 F F
Clostridium (log CFU per 100 mL) F 3.0–4.5 3.6–5.5

Table 3
Analytical methods for microbiological parameters

Parameters Analytical methods

E. coli (Indianapolis/Evry) Standard method 9222 O-M/Enterolert, Idexx (CIRSEE)


Fecal coliforms (Indianapolis/Washington) Standard method 9222 D-M/membrane filtration: incubation on 0.45 mm membrane
lauryl sulphate Broth for 4 h at 371C then 14 h at 441C. Enumeration of presumptive
fecal coliforms and confirmation by subculture into lactose peptone water at 371C in
conjunction with an oxidase test
Enterococci Membrane filtration: incubation on Slanetz and Bartley Agar for 4 h at 371C
followed by 44 h at 441C. Count all maroon colonies, confirmation on bile aesculin
azide agar
Clostridium Membrane filtration: heat-treat the sample at 751C for 10 min. Serially dilute and
vacuum filter appropriate dilutions/volumes through 0.45 mm membrane. Incubate
on Perfringens OPSP medium anaerobically at 371C for 48 h. Count all black
colonies and confirm in crossley milk
Enterovirus Suspended cell plaque assay: adsorption onto a cellulose nitrate membrane at pH 3.5;
elution by a positively charged protein solution; flocculation of the protein solution
and centrifugation; virus numeration by tissue culture assay (recovery quoted>20%)
F-specific RNA bacteriophages Incubation with a host strain: direct plating using a semi-solid overlay technique,
with Salmonella typhimurium WG49 as host bacterium (MS2 bacteriophage used as
positive control)
Salmonellae Filter appropriate volumes of sample through a 0.45 mm filter (using filter-aid if
turbid). Pre-enrich the filter in buffered peptone water for 24 h at 371C. Enrich a
0.1 ml portion of the culture in RVS broth for 48 h at 411C subculturing onto XLD
Agar and Brilliant Green Agar after 24 and 48 h. Presumptive Salmonellae are
confirmed serologically and biochemically

Fecal Coliforms and E. coli were chosen as standard A number of physico-chemical parameters were
fecal indicators in the study, because they are usually monitored on the effluents before and after ozonation
regulated for wastewater discharge or reuse. Other using Standard methods for pH, TOC, Turbidity, SS,
microorganisms were also studied in the case of the total and filtered COD, TOC, BOD5, UV 254 abs. (total
Washington effluent: enterococci Clostridium (as surro- and after filtration), N–NO2, alkalinity, Mictotox tests
gates for more resistant organisms), Salmonellae, en- for toxicity and particle size distribution. To investigate
teroviruses and F-specific bacteriophages (considered a the effect of ozonation on colour abatement, absorbance
good model for virus disinfection [15]). The microbio- at 400 nm was used to characterise apparent colour.
logical methods used were based on the French and UK Water appears to be coloured when dissolved matter
Industry Standard Methods [16] with additional dilution absorbs visible light or when suspended particles scatter
of the samples to appropriate levels. The analytical light (Rayleigh scattering). Finally, to stop the effect of
methods are detailed in Table 3. ozone residual on microorganism concentration after
P. Xu et al. / Water Research 36 (2002) 1043–1055 1047

sampling, 4% (wt/vol) sodium thiosulfate was added to of the hydraulic retention time. The measured values of
the treated effluent samples. 7.4–9.6 mg L1 are significantly higher than those in the
To investigate the influence of particles on disinfection other two effluents, as expected from the water quality
performances, additional tests were run with the Evry data (see Table 2). The ozone demand of the Evry
effluent after filtration (75 mm Arkal prefilter, followed secondary effluent (extended aeration with nitrification)
by a 10 mm canvas filter), to lower the suspended solids of 3.1–4.2 mg L1 is similar to the values measured in the
concentration below 2 mg L1. Indianapolis tertiary effluent of 2.5–5.3 mg L1. These
results are in agreement with an earlier study performed
on other effluents, which showed organic content to be a
much more influential parameter than suspended solids
3. Results and discussion on the ozone demand [12]. These values are used
hereafter when discussing the disinfection performances.
3.1. Inactivation of fecal coliforms: impact of
operating conditions and wastewater quality
3.1.2. Influence of operating conditions on coliform
3.1.1. Influence of wastewater quality on ozone demand inactivation
Fig. 2 illustrates the determination of ozone demand Fig. 3 summarises the results from all the experiments,
in the Washington secondary effluent for different values where the residual concentration of bacteria after

Fig. 2. Determination of the immediate ozone demand according to the classical approach used for drinking water (effluent from
Washington, UK).

Fig. 3. Performances of ozone for FC inactivation on three different effluents: comparison of concentration level after ozonation with
reuse standards.
1048 P. Xu et al. / Water Research 36 (2002) 1043–1055

ozonation is plotted as a function of the transferred contact in the post-contactor chamber. Residual ozone
ozone dose (TOD). One important observation is that a decreased significantly in the post-contactor after 2 min
significant 1–3 log inactivation is already reached when and no significant increase in FC inactivation is
TOD approximates the immediate ozone demand, i.e. observed. The quick decay of ozone may be explained
with no measurable residual of ozone in solution. This is by the wastewater matrix-consuming ozone; therefore,
in agreement with earlier data on wastewater ozonation no further inactivation can be expected from a post-
[12]. More precisely, the higher the immediate ozone contactor without additional ozone injection. As a
demand, the higher the inactivation level reached at that consequence, no credit of additional inactivation can
dose. This phenomenon can be explained by the fact that be attributed to a storage reservoir or outfall that would
bacteria themselves participate in the immediate ozone be used for discharge of ozonated wastewater.
demand, because of their high kinetic rate with ozone.
One important consequence from these results is that 3.1.3. Effect of wastewater quality on ozone
design and control of ozonation for fecal coliform disinfection performances
disinfection in wastewater should not be based on the Despite similar values of immediate ozone demand
standard parameter of residual ozone or ‘‘Ct’’ factor. In and not very different initial concentrations of bacteria,
agreement with this, Rakness et al. [10] reported that the secondary and tertiary effluents of Evry and
direct measurement of ozone residual within the full- Indianapolis display different inactivation performances
scale contactors at Indianapolis had been unsuccessful (see Fig. 3). It appears that only the tertiary effluent is
although good disinfection had occurred. The ‘‘Ct’’ able to meet stringent standards for almost total bacteria
approach is applied to drinking water because more inactivation like the Californian Title 22 criteria. This
resistant microorganisms like Giardia are targeted. result led to a more in-depth investigation of the
Data from Fig. 3 show that the hydraulic retention influence of particles on disinfection.
time (HRT) has no impact on the performances of fecal Comparative tests were performed with the effluent
coliform (FC) or E. coli disinfectionFfor a given TOD, from Evry, and with the same effluent after a filtration
a 2 min HRT provides the same inactivation as 10 min step. These tests were performed simultaneously in order
HRT. These results have major consequence for the to minimise any fluctuation in water quality or operating
design of ozone wastewater disinfection, demonstrating conditions. The ozone demand of the effluent was found
that mass transfer is the critical step that has to be to be exactly the same. The effect of filtration on
optimised, and that no long-contact-time chamber is disinfection is shown in Fig. 5: an additional 1 log
necessary. inactivation was obtained by filtration. It must be
Fig. 4 illustrates this statement, showing the distribu- stressed that for a given bacteria concentration (log N0),
tion of ozone residual and FC inactivation along the a lower level of contamination was obtained after
ozonation column and after 2, 4 and 6 min additional filtration with the same ozone dose. This conclusion

Fig. 4. FC inactivation and residual ozone distribution vs. contact time in and after the ozone column (Washington secondary effluent,
HRT 4 min, TOD 13.1 mg L1).
P. Xu et al. / Water Research 36 (2002) 1043–1055 1049

Fig. 5. Impact of a 10 mm pre-filtration on the inactivation of total coliforms by ozone (secondary effluent in Evry, France).

explains the higher inactivation level obtained with the countries for landscape (Australia, 1 pfu/50 L) or agri-
tertiary effluent from Indianapolis, with its very low cultural irrigation (Arisona and Hawaii, USA, 1 pfu/
suspended solids content. 40 L).
Compared to all other microorganisms, the higher
3.2. Ozone disinfection of other microorganisms resistance of Clostridium was confirmed by the experi-
mental data. A thorough investigation was made with
In order to evaluate the disinfection efficiency of the effluents from Evry and Washington with that
ozone on other microorganisms, fecal streptococci (FS), indicator (Fig. 6). With a TOD approximately equal to
Clostridium, Salmonellae, Enterovirus, and F+-specific the immediate ozone demand of the effluents (3–
bacteriophages were chosen as target microorganisms 5 mg L1 for Evry and 8–10 mg L1 for Washington),
for additional tests with the Washington secondary less than 0.5 log inactivation of Clostridium was
effluent. The inactivation efficiencies are presented in achieved. The maximum inactivation level was less than
Table 4. It should be noted that the initial concentration 2 log for high TOD of 33 mg L1 (HRT 9.6 min,
of microorganisms limited the maximum inactivation Washington secondary effluent).
values that could be reached. The resistance of fecal
coliforms and enterococci to ozonation was similar, in 3.3. Impact of ozonation on effluent water quality
agreement with previous results [17,18]. It is important
to stress also that a relatively low ozone dose of Due to the high oxidative potential of ozone,
8.6 mg L1 totally inactivates Salmonellae. ozonation has a beneficial effect on effluent quality,
A higher efficiency of ozonation was observed for which argues in favour of its application for wastewater
virus inactivation: a transferred dose of 4.8 mg L1 with reuse (Table 5). The most significant effect of ozone was
4 min HRT was enough for total inactivation of on UV-254 absorbance and colour (Figs. 7a and b).
enteroviruses (>2.9 log inactivation). Total inactivation With an increase in transferred ozone dose from 2 to
of F+ coliphages (>2.2 log inactivation) was reached 13 mg L1 at HRT 4 min, the variation of UV absor-
with a slightly higher transferred dose of 8.6 mg L1. bance in the Evry secondary effluent increased from
The strong virucidal power of ozone for wastewater 28% to 55%. These results indicate that the ozone reacts
disinfection confirms its well-known performance in and oxidises the organic matter, in particular, the
drinking water [7,18]. Indigenous enteric viruses isolated compounds having double bonds and/or an aromatic
from wastewater effluents have been shown to be much structure that determine the value of the absorbance at
more resistant [18]. In comparison, coliphages have been 254 nm. The UV absorbance abatement was higher in
found to be very sensitive to ozone, which puts some the Evry secondary effluent than in Washington,
doubt on the validity of these coliphages as surrogates indicating the presence of refractory contaminants.
for enteric viruses [19,20]. Finally, no significant difference was observed between
This study shows that the ozone dose required to different contact times in the reactor (not shown),
satisfy WHO regulations (1000 FC per 100 mL) also revealing the fast kinetics of the reaction between ozone
provides total inactivation of indigenous enteric viruses. and unsaturated and aromatic compounds. The critical
This indicates that ozone would be highly recommended factor for water quality improvement is also the ozone
for the production of virus free water, required in several dose transferred into the water.
1050 P. Xu et al. / Water Research 36 (2002) 1043–1055

Total BOD5 increased up to 20% after ozonation of

F-coliphage (/mL)

1.68
>2.16
>2.09
the highly polluted secondary effluent in Washington,

Red
UK. This is typical with ozone, which can oxidise
recalcitrant compounds and thereby increase effluent
N

2
0
0
biodegradability. The total COD was not significantly
influenced by ozonation, while an increase of dissolved
96
144
122
N0

COD was observed. The difference between total and


dissolved COD behaviours could be related to the global
Enterovirus (pfu/10L)

1.66
>2.89

>2.81
>2.82
>2.89
>2.89
decrease of turbidity observed during the tests (Fig. 7-
Red

c)Fsome particles and high weight organic compounds


would be destroyed by ozone and converted into
dissolved compounds.
12
0

0
0
0
0
N

Toxicity was detected neither in the untreated


secondary effluent nor in the ozonated effluents. These
775
544
650
654
774
774
N0

results are in compliance with previous studies on urban


wastewater disinfection by ozone [17]. It is important to
Abs
Abs
Abs
Abs
Abs
Salmonella

stress that the presence of toxicity after ozonation


N

reported in the literature is usually related to the


Prst
Prst
Prst
Prst
Prst

presence of industrial wastewater [21–24].


F
N0

3.4. Design of wastewater ozonation for given regulations


0.88
1.64
2.50
2.26
3.56
3.33
Red
Enterococci (/100 mL)
Comparison of ozone disinfection on different microorganisms in the secondary effluent in Washington, UKa

Ozone disinfection results were compared for two


7100
980
173
400
14
16

different water reuse standards: (1) WHO stringent


N

guidelines for irrigation, fecal coliformso1000 cfu per


100 mL, and (2) Californian Title 22 standards, total
log N0

4.73
4.63
4.74
4.86
4.71
4.53

coliformso2.2 cfu per 100 mL (see Fig. 3). The TOD


required to meet WHO guidelines are 2, 4 and
15 mg L1, respectively, for the tertiary affluent and the
0.31
0.21
1.16
1.72
Red
Clostridium (/100 mL)

two secondary effluents in Evry and Washington for an


F
0

HRT of 2 min and an additional beneficial effect of 30%


21000
6200
4800
3100

2000
550

reduction in UV absorbance. Compliance with the


stringent Title 22 criteria of virtually total removal of
N

fecal coliforms can be reached only after tertiary


log N0

filtration in the Indianapolis effluent and with a TOD


3.56
3.67
3.81
4.53
4.46
4.46

of 8 mg L1 for 2 min HRT. A very low suspended solids


concentration (o5 mg L1) emerges as the most im-
HRT min TOD (mg L1) Fecal coliform (Cfu per 100 mL)

portant design requirement to meet very stringent


Note: RedFlogðN0 =NÞ; PrstFpresent; AbsFabsent.
0.34
1.82
2.48
2.52
4.42
4.06
Red

disinfection requirements.
In both scenarios, total inactivation of viruses is
achieved, which may be important if viruses are also
included in regulations.
4600
1320
300
840
14
14
N

3.5. Modelling approach of wastewater disinfection


by ozone
log N0

5.00
4.94
4.95
5.45
5.92
5.20

Disinfection is standardly described, for drinking


water, by the Chick–Watson model:
d½N 
¼ kN ½N½O3 : ð5Þ
dt
15.2
11.0
24.8
29.5
4.8
8.6

In the case of a CSTR, Eq. (5) enables the evaluation of


the number of microorganisms, N:
Table 4

½N 1
¼ ; ð6Þ
a
9.6
9.6
9.6

½N0  1 þ kN ½O3 t
4
4
4
P. Xu et al. / Water Research 36 (2002) 1043–1055 1051

Fig. 6. Performances of ozone for Clostridium inactivation. (secondary effluents of Washington, UK and Evry, France).

Table 5
Impact of ozonation on water quality in Washington, UKa

Parameters HRT 4.0 min HRT 2.0 min HRT 9.6 min HRT 4.0 min
TOD 9.2 mg L1 TOD 12.3 mg L1 TOD 15 mg L1 TOD 21.1 mg L1

C0 9 24 17 10
TBOD C 8 30 21 12
(mg L1) ðC  C0 Þ=C0 11% +25% +24% +20%
C0 73 94 93 70
TCOD C 66 92 94 69
(mg L1) ðC  C0 Þ=C0 10% 2% +1% 1%
C0 42 43 47 38
DCOD C 42 51 63 47
(mg L1) ðC  C0 Þ=C0 0% +19% +34% +24%
Total C0 0.44 0.48 0.51 0.43
UV Abs C 0.36 0.4 0.39 0.27
(Cm1) ðC  C0 Þ=C0 20% 17% 24% 38%
Dissolved C0 0.22 0.23 0.23 0.22
UV Abs C 0.17 0.17 0.17 0.12
(Cm1) ðC  C0 Þ=C0 25% 24% 27% 44%
a
C0 is the initial concentration of secondary effluent; C is the concentration of ozonated effluent.

where N is the density of viable microorganisms (N0 at kinetics, characterised by coefficient rates kX and kY
t ¼ 0), t is the hydraulic retention time in the reactor, [25]:
and kN is the inactivation rate constant.
In fact, the use of the Chick–Watson model implies d½O3 
¼ KL að½O3  *  ½O3 Þ  kX ½X½O3   kY ½Y½O3 ;
the presence of ozone residual to achieve inactivation of dt
microorganisms; there would be no disinfection before ð7Þ
the ozone demand is met, i.e. for transferred ozone doses
lower than 8 mg L1, the immediate ozone demand of d½Y 
¼ kY ½Y½O3 ; ð8Þ
wastewater. dt
In order to account for the 1–3 log inactivation
observed without measurable residual in wastewater d½X 
¼ kX ½X½O3 : ð9Þ
ozonation, a modified approach was therefore devel- dt
oped, considering the consumption of ozone by the The comparison with the previous equations used for
organic matter as a combination of a rapid and a slow drinking water shows that we can obtain the same model
1052 P. Xu et al. / Water Research 36 (2002) 1043–1055

ozone concentration and also concentrations of the


different species X and Y (rapidly consumed and the
slowly consumed organic matter):
DOT
½O3  ¼ ð10Þ
1 þ kX ½Xt þ kY ½Yt

½Xo 
½X ¼ ; ð11Þ
1 þ kX ½O3 t

½Yo 
½Y ¼ : ð12Þ
1 þ kY ½O3 t
The use of these equations should be interpreted as a
valid model for quantification but not necessarily as an
accurate representation of the more complex mass
transfer steps and heterogeneous reactions taking place
during the disinfection process. In particular, radical
mechanisms are ignored for simplification. Besides, in a
first step, hydraulic aspects are ignored in the inter-
pretation of the data and we assume that the bubble
column can be represented as a CSTR, which is close
enough to the experimental data. The focus is on the
kinetic modelling, which can be completed afterwards
with hydrodynamic information. The exact value of the
fitting parameters will then be slightly altered, but not
their order of magnitude. A very similar approach was
taken by Hunt and Marinas * [26] on the inactivation of
E. coli with ozone in synthetic waters.
This modified approach was applied to the experi-
mental data from Washington WWTP, with residual
ozone as a function of TOD. It led to the following
fitting parameters (Fig. 8):
kX ¼ 10 L mg1 min1 ;

kY ¼ 0:01 L mg1 min1 ;

½Xo  ¼ 8 mg L1 ;

½Yo  ¼ 60 mg L1 :
This represents a first improvement, but the major
difference between both models appears when consider-
Fig. 7. Influence of ozonation on wastewater quality. (a) Total ing the inactivation of fecal coliforms. For transferred
and dissolved UV-254 absorbance removal vs TOD (4 min ozone doses lower than the fast ozone demand,
HRT, secondary effluents of Washington, UK and Evry, integrated exposure to ozone in terms of Ct is equal to
France), (b) colour removal vs TOD at different contact times zero for the classical approach and reaches
(Evry secondary effluent) and (c) turbidity change versus 0.40 mg min L1 with the modified model for a TOD
transferred ozone dose (Washington secondary effluent). of 8 mg L1. When applying the Chick–Watson model
given by Eq. (5), this slight difference in the Ct makes it
by simplification if possible to account for the inactivation of fecal coli-
forms during X organic matter consumption. Fig. 9
kD ¼ kY ½Y; shows the strong impact of this low Ct in wastewater,
which inactivates 2 log of E. coli. The corresponding
kX ¼ N: inactivation rate constant kN is equal to
100 L mg1 min1 and the fitting curves are depicted in
Eqs. (7–9) can easily be solved in an open completely Figs. 9a and b. Besides, the model properly accounts for
mixed reactor (CSTR with a contact time t) to estimate the fact that the contact time has no effect on
P. Xu et al. / Water Research 36 (2002) 1043–1055 1053

Fig. 8. Modelling of the residual ozone with the modified approach developed for wastewater disinfection.

Fig. 9. Inactivation of E. coli with ozone for the classical (a) and modified (b) models applied to disinfection in wastewater effluent.

inactivation performances. Therefore, a significant im- major advantage of ozone for regulations that include
provement is observed when the apparent ozone virus removal. In agreement with previous data,
decomposition is represented as a mixed second order bacteriophages were found very sensitive to ozone,
rate expression depending on the fast ozone demand X laying doubt on the pertinence of such microorganisms
(Eq. (11)). as indicators for ozone treatment. By contrast, the
higher resistance of Clostridium confirms that they are
good candidates for resistant microorganism indicator.
4. Conclusions Ozonation also provides a significant reduction of UV
absorbance and colour, which can be an advantage for
The experimental results obtained at pilot scale on some reuse applications. More stringent regulations like
different wastewater effluents confirm the efficiency of Title 22 require the implementation of an efficient
ozone for wastewater disinfection. Depending on the tertiary filtration step.
quality of the effluent, a TOD of between 2 and From an operational viewpoint, transfer of ozone
15 mg L1 was necessary to meet the WHO standard from the gas phase to the water was found to be the
for irrigation (1000 FC per 100 mL). Such a dose was critical step for fecal coliform inactivation with ozone,
shown to provide total elimination of enteroviruses in because of the fast kinetics between ozone and coliform
the worst quality secondary effluent, which can be a bacteria. No difference in inactivation was found
1054 P. Xu et al. / Water Research 36 (2002) 1043–1055

between 2 and 10 min hydraulic retention time, for a MFA, Shayya WH, editors. Water management, purifica-
given ozone dose transferred to the effluent. As a tion and conservation in arid climates, Water Conserva-
consequence, the ‘‘Ct’’ approach commonly applied in tion, vol. 3. Technomic p. 171–98.
drinking water treatment should not be used for the [7] Langlais B, Reckhow D.A, Brink DR. Ozone in water
ozonation of wastewater. The new approach to waste- treatmentFapplication and engineering. Co-operative
Research Report. Lewis, 1991.
water ozone contactor design must be based on short
[8] Facile N, Barbeau B, Pr!evost M, Koudjonou B. Evaluat-
contact times and enhanced mass transfer. Further, no
ing bacterial aerobic spores as a surrogate for Giardia and
credit of inactivation can be attributed to a storage Cryptosporidium inactivation by ozone. Water Res
reservoir or outfall that would be used for wastewater 2000;34(12):3238–46.
discharge after ozonation, because of high ozone decay [9] Robson CM, Rice RG. Wastewater ozonation in the
in an effluent matrix. USAFhistory and current statusF1989. Ozone Sci Eng
Finally, a kinetic model was developed in order to 1991;13:23–40.
account for the 1–3 log inactivation of bacteria that was [10] Rakness KL, Corsaro KM, Hale G, Blank BD. Waste-
observed experimentally without a measurable concen- water disinfection with ozoneFprocess control and
tration of ozone in the bulk solution. In fact, a operating results. Ozone Sci Eng 1993;15:497–514.
combination of the classic Chick–Watson disinfection [11] Hunter GF, Rakness KL. Start-up and optimization of the
ozone disinfection process at the Sebago Lake water
model and the ‘‘instantaneous demand’’ model used in
treatment facility. Ozone Sci Eng 1997;19:255–72.
drinking water was insufficient. The modified model was [12] Janex ML, Savoye P, Roustan M, Do-Quang Z, La#ın!e JM,
based on the simultaneous consumption of ozone by the Lazarova V. Wastewater disinfection by ozone: influence
microorganisms and by the organic matrix. The of water quality and kinetics modelling. Ozone Sci Eng
apparent decomposition rate of dissolved ozone was 2000;22(2):113–22.
represented successfully by mixed second-order rate [13] Bader H, Hoigne J. Determination of ozone in water by
equations. indigo methods: a submitted standard method. Ozone Sci
Eng 1982;4:169–76.
[14] Roustan M, Debellefontaine H, Do-Quang Z, Duguet JP.
Acknowledgements Development of a method for the determination of ozone
demand of a water. Proceedings of IOA Congress, Kyoto,
1997. p. 589–94.
The authors would like to thank Luc Burtin
[15] Hall RM, Sobsey MD. Inactivation of hepatitis A virus
(CIRSEE) for technical assistance, Ozonia for the
and MS2 by ozone and ozone-hydrogen peroxide in
provision of an ozone generator, Evry and Washington buffered water. Water Sci Technol 1993;27(3–4):371–8.
wastewater treatment plants staff for field assistance, [16] HMSO Publication (1994). The microbiology of water
CIRSEE and Northumbian Water Group of Ondeo 1994FPart 1Fdrinking water. Report on public health
Services for cooperation and lab analysis. and medical subjects No. 71. Methods for the examination
of waters and associated materials.
[17] Mandra V, Lazarova V, Dumoutier N, Audic JM.
References Comparative study of urban sewage disinfection by
peracetic acid, UV and ozone (Etude comparative de la
[1] Lazarova V. Role of water reuse in the integrated resources desinfection des eaux r!esiduaires urbains par l’acide
management: costs, benefits and technological challenges perac!etique, l’irradiation UV et l’ozone). Journ!ees Infor-
# de la r!eutilisation des eaux us!ees pour la gestion
(Role mation Eaux, Poitiers, 18–20 September, 1996.
int!egr!ee des ressources: couts,
# b!en!efices et d!efis technolo- [18] Hartemann PH, Block JC, Joret JC, Foliguet JM, Richard
giques). Leau, Lindustrie, Les nuisances 1999;227:47–57. Y. Virological study of drinking and wastewater disinfec-
[2] Lazarova V, Cirelli G, Jeffrey P, Salgot M, Icekson N, tion by ozonation. Water Sci Technol 1983;15:145–54.
Brissaud F. Enhancement of integrated water management [19] Helmer RD, Finch GR. Use of MS2 coliphage as a
and water reuse in Europe and the Middle East. Water Sci surrogate for enteric viruses in surface waters disinfected
Technol 2000;42(1/2):193–202. with ozone. Ozone Sci Eng 1993;15:279–93.
[3] Levine B, Lazarova V, Manem J, Suffet IH. Wastewater [20] Harakeh MS, Butler M. Factors influencing the ozone
reuse standards: goals, status and guidelines. WEF inactivation of enteric viruses in effluent. Ozone Sci Eng
beneficial reuse of water and biosolids Conference 1985;6:235–43.
Proceedings, Marbella, 6–9 April 1997; 13/59–13/71. [21] Langlais B, Legube B, Beuffle H, Dor!e M. Study of the
[4] USEPA Manual (1992). Guidelines for water reuse. EPA/ nature of the by-products formed and risks of toxicity
625/R-92/004. when disinfecting a secondary effluent with ozone. Water
[5] Tyrrell SA, Rippey SR, Watking WD. Inactivation of Sci Technol 1992;25(12):135–43.
bacterial and viral indicators in secondary sewage efflu- [22] Paraskeva P, Lambert SD, Graham NJD. Influence of
ents, using chlorine and ozone. Water Res 1995;29(11): ozonation conditions on the treatability of secondary
2483–90. effluents. Ozone Sci Eng 1998;20:133–50.
[6] Lazarova V. Wastewater disinfection: assessment of the [23] Ito K, Jian W, Nishijima W, Base AU, Shoto E, Okada M.
available technologies for water reclamation. In: Goosen Comparison of ozonation and AOPs combined with
P. Xu et al. / Water Research 36 (2002) 1043–1055 1055

biodegradation for removal of THM precursors in [25] Cockx A, Janex ML, Lazarova V. Development of ozona-
treated sewage effluents. Water Sci Technol 1998; tion modelling for the disinfection of wastewater effluents.
38(7):179–86. Proceedings of the International Ozone Association
[24] Monarca S, Feretti D, Collivignarelli C, Guzzella L, Conference, Wasser Berlin, 23–26 October 2000. p. 79–93.
Zerbini I, Bertanza G, Pedrazzani R. The influence [26] Hunt NK, Marinas* B. Inactivation of Escherichia coli with
of different disinfectants on mutagenicity and toxicity ozone: chemical and inactivation kinetics. Water Res
of urban wastewater. Water Res 2000;34(17):4261–9. 1999;33(11):2633–41.

You might also like