IOA-EA3G Conference & Exhibition, Toulouse, France
– June 4-6, 2012
Ozone disinfection efficiency of resistant microorganisms assessed at a drinking water
treatment plant by combining inactivation kinetics and computational fluid dynamics
Pauline Talbot 1, Laure Martinelli 1, Samuel Talvy 1, Eric Chauveheid 2, Benoît Haut 1
1. Transfers, Interfaces and Processes (TIPs)—Chemical Engineering Unit, Université Libre de
Bruxelles, av. F.D. Roosevelt 50, CP 165/67, 1050 Brussels, Belgium
2. Vivaqua, Chaussée de Waterloo 764, 1180 Brussels, Belgium
Abstract
In this work, a method to assess the efficiency of a drinking water plant to inactivate resistant
microorganisms using ozone is proposed. This method compares the duration of the lag phase of
the ozone inactivation of resistant microorganisms with the contact time distribution of these
microorganisms with the disinfectant in the plant. To evaluate the duration of the lag phase, an
experimental procedure is proposed and applied to Bacillus subtilis spores. These spores are used
as a surrogate for resistant microorganisms such as Cryptosporidium parvum oocysts. The
procedure aims at characterizing the ozone inactivation kinetics of Bacillus subtilis spores for
different temperature and ozone concentration conditions. From experimental data, the duration of
the lag phase of the ozone inactivation of Bacillus subtilis spores is identified for different
temperature and ozone concentration conditions. To evaluate the contact time distribution between
microorganisms and ozone in a disinfection step of a drinking water plant, a computational fluid
dynamics tool is used. The proposed method is applied to an existing drinking water plant located
in Belgium and operated by Vivaqua. This plant uses the ozone as disinfectant. Results show that
the durations of the lag phase and contact times are both in the same order of magnitude of a few
minutes. For a large range of temperatures and ozone concentrations in the plant ozonation
channel and for the highest hydraulic flow rate applied, a significant proportion of resistant
microorganisms similar to Bacillus subtilis spores is not inactivated.
Key-words: Disinfection, ozonation, Bacillus subtilis spores, inactivation kinetics, Computational
Fluid Dynamics.
Introduction
Until now, European standards for assessing the microbiological quality of drinking water are
focused only on E. Coli and Enterococcus [1]. New coming standards could probably also consider
the inactivation of pathogenic resistant microorganisms such as Cryptosporidium parvum oocysts
or the inactivation of their indicator species such as Bacillus subtilis spores. In order to anticipate
these new standards and ensure consumer’s health and safety, drinking water producers need
new methods to assess the efficiency of their disinfection step. Hitherto, few methods enabling the
producers to fully assess the capacity of a drinking water plant to inactivate resistant pathogenic
microorganisms have been proposed. Some previous studies have proposed Computational Fluid
Dynamics (CFD) tools to assess an ozonation step in a drinking water plant [2, 3, 4]. These tools
provide information such as the dynamics of the liquid and gas phases, the ozone concentration
field in the disinfection step and the contact time between the microorganisms and the disinfectant
in the process. Nevertheless, these tools do not allow an estimation of the concentration of
pathogenic resistant microorganisms at the outlet of the ozonation step. Indeed, few investigations
have been done to set up the kinetics of the ozone inactivation of resistant microorganisms.
Therefore, such an inactivation has never been included in a CFD tool. Hence, the inactivation
kinetics of resistant microorganisms needs to be established to complete these tools and get
information on the inactivation of pathogenic resistant microorganisms during an ozonation step.
Ozone resistant microorganisms present typical inactivation kinetics. When these microorganisms
get in contact with the disinfectant, their resistance results in an initial lag phase during which they
are not, or weakly, inactivated. The duration of this phase depends on several parameters upon
which the most important is the temperature [5]. The lag phase is followed by a decaying phase
that can be modeled by a pseudo-first order reaction [5, 6, 7].
IOA-EA3G Conference & Exhibition, Toulouse, France
– June 4-6, 2012
The general purpose of this paper is to propose a method to assess the efficiency of a drinking
water plant to inactivate resistant microorganisms.
The first goal of this work is to draw up and apply to Bacillus subtilis spores an experimental
protocol to characterize the ozone inactivation kinetics of ozone resistant microorganisms. From
experimental data, the duration of the lag phase of the ozone inactivation of Bacillus subtilis spores
is identified for different temperature and ozone concentration conditions.
The second goal of this paper is to assess the efficiency of an existing ozonation process to
inactivate resistant pathogenic microorganisms similar to Bacillus subtilis spores using the CFD
tool proposed by Talvy et al. (2011) and the durations of the lag phase of Bacillus subtilis spores
inactivation identified in this work. The studied process is the ozonation step of the Tailfer drinking
water plant operated by Vivaqua (Belgium). From the CFD tool proposed and validated by Talvy et
al. (2011), the contact time distribution between microorganisms and ozone in the disinfection step
of Tailfer plant is determined. Evaluation of the Tailfer plant disinfection step efficiency is obtained
by comparing the durations of the lag phase of the resistant microorganisms with the ozone
contact time distribution of these microorganisms. Indeed, if the duration of the lag phase of the
resistant microorganisms is larger than the contact time between these microorganisms and the
ozone in the disinfection step, these microorganisms are still in their resistance phase at the exit of
the step.
In this work, Bacillus subtilis spores are used as a surrogate for Cryptosporidium parvum oocysts.
Cryptosporidium parvum oocysts are known to be problematic in drinking water production
because of their persistence in the environment, their pathogenic properties and their resistance to
conventional disinfectants such as chlorine, chlorine dioxide and ozone [5, 7, 8, 9]. These
microorganisms are difficult to quantify. Current methods are expensive, time-consuming and
poorly reliable for the low concentrations found in the treated water [7]. Studies have shown a
correlation between the inactivation kinetics of Cryptosporidium parvum oocysts and the
inactivation kinetics of Bacillus subtilis spores [5, 6, 7, 8, 10]. Moreover, Bacillus subtilis spores are
easier to quantify than Cryptosporidium parvum oocysts. Therefore, Bacillus subtilis spores are
chosen as Cryptosporidium parvum oocysts surrogate, and our experiments are conducted with
these spores.
Material and methods
Experimental characterization of Bacillus subtilis spores inactivation kinetics
Bacillus subtilis spores inactivation experiments are conducted with a set of twenty different
experimental conditions. These experiments are performed for temperatures from 5 to 25 °C by
step of 5 °C. For each temperature, four initial ozone concentrations are tested, from 0.5 to
1.5 mg.L-1. These conditions correspond to the temperature and ozone concentration ranges
applied at many drinking water plants and more specifically at the Tailfer plant.
In this work, indigenous Bacillus subtilis cells are isolated from the river Meuse, and their
sporulation is induced by thermal treatment. The water samples used in the experiments are taken
from the Tailfer plant, just before the ozonation step. The pH of the water is measured before every
experiment and is 7.4 ± 0.2 (average ± standard deviation).
At the beginning of the experiments, a stirred reactor is filled with 2 liters of the Tailfer water. A
period of 30 to 60 minutes is needed to stabilize the water temperature. Ozone is then injected into
the reactor until a 10 % higher concentration than the desired initial concentration. After the
injection, the dissolved ozone concentration decreases and when the desired initial ozone
concentration is reached, 10 mL of a suspension of Bacillus subtilis spores are injected into the
reactor. The injected spore suspension is prepared in order to obtain a spore concentration in the
reactor close to 105 spores.L-1. The starting time (t = 0 min) of the experiments corresponds to the
spore injection. At this moment, records of ozone concentration and temperature are triggered. The
ozone sensor is calibrated by the indigo method [11]. In order to be sure that the reactor is wellmixed and that samples are homogeneous, the first water sample (about 10 mL) is taken after a
IOA-EA3G Conference & Exhibition, Toulouse, France
– June 4-6, 2012
minimum of 15 seconds. The following samples (about 10 mL) are taken at regular intervals during
a predefined period depending on the temperature and initial ozone concentration conditions. The
samples are stored at 4 °C and analyzed by filtration [8] within 24 hours after the sampling.
Establishment of the contact time distribution between microorganisms and ozone in the
disinfection step of the Tailfer plant
The disinfection step of the Tailfer plant is composed of two identical ozonation channels. The
geometry of an ozonation channel is shown in figure 1. Baffles divide the channel into successive
compartments crossed by the water. Air enriched in ozone is bubbled at the bottom of the fourth
compartment, called chamber 1, through porous diffusers.
Figure 1: Geometry of an ozonation channel of the Tailfer drinking water plant.
The determination of the contact time distribution between microorganisms and ozone in the
disinfection step of the Tailfer plant is carried out with CFD according to the method proposed and
validated by Talvy et al. (2011). To establish the contact time distribution, two stages are followed.
First, the gas-liquid two phases flow, taking place in an ozonation channel of the Tailfer plant, is
modeled and simulated. Second, contact time distribution between microorganisms and ozone in
an ozonation channel of the Tailfer plant is computed from the previous simulation.
The simulation of the gas-liquid two phases flow is computed by solving transport equations in
Ansys Fluent® 6.3 using a two phases Eulerian model. The gas-liquid two phases flow is simulated
for a gas flow of 24 Nm³.h-1 and the maximum hydraulic flow rate of 2900 m³.h-1 applied at the
Tailfer disinfection step. The boundary conditions and discretization schemes used are the same
than those proposed by Talvy et al. (2011).
To determine the contact time distribution between microorganisms and ozone in the Tailfer
ozonation channel, the transport equation of a tracer is solved. A fixed quantity of this tracer is
virtually and punctually injected in the region of the ozonation channel where the microorganisms
are first in contact with dissolved ozone. This region is a box situated at the top of the chamber 1
represented in figure 1. The evolution of the tracer concentration at the ozonation channel exit
gives, after standardization, the tracer residence time distribution. This residence time distribution
is considered to correspond to the contact time distribution between microorganisms and ozone.
Results and discussion
Bacillus spores inactivation kinetics
The experimental results of the Bacillus subtilis spores ozone inactivation at 5 °C are shown in
figure 2 for four initial ozone concentrations (0.449, 0.740, 1.00 and 1.25 mg.L-1). Inactivation
kinetics of Bacillus subtilis spores shows a lag phase followed by a pseudo-first order decreasing
phase. The results of the experiments at 10, 15, 20 and 25 °C present the same two phases. The
identification of these two phases agrees with the results found in the literature [5, 6, 7, 10]. The
IOA-EA3G Conference & Exhibition, Toulouse, France
– June 4-6, 2012
duration of the lag phase is identified from experimental data between 2 and 13 min. In a paper
submitted in Water Research, a model of the inactivation of Bacillus subtilis spores is proposed
[12]. In this model, the duration of the lag phase is identified as a parameter defined as the lag
time. This lag time is then correlated to the temperature and ozone concentration conditions of the
laboratory experiment in a correlation called the lag time correlation [12].
Figure 2: Time evolution of the natural logarithm of Bacillus subtilis spores concentration divided by the
calculated initial Bacillus subtilis spores concentration (2.46 x105 spores.L-1) for the Bacillus subtilis spores
inactivation experiments at 5 °C with four initial ozone concentrations (  = 0.449 mg.L-1,  = 0.740 mg.L-1, 
= 1.00 mg.L-1,  = 1.25 mg.L-1).
Contact time distribution between microorganisms and ozone in the Tailfer ozonation channels
tc  tc
The contact time distribution, E(t), is shown in figure 3 for the highest hydraulic flow rate.
 E(t )dt
tc
is the fraction of the microorganisms in contact with the ozone between times t c and t c  t c . The
minimum contact time obtained for the highest hydraulic flow rate applied at the Tailfer ozonation
channels is 1.37 minutes. The mean contact time, t c , is calculated at 8.5 min by the equation (1):

t c   t c E(t c )dt c
0
(1)
IOA-EA3G Conference & Exhibition, Toulouse, France
– June 4-6, 2012
Figure 3: Contact time distribution between microorganisms and ozone in a Tailfer ozonation channel
(continuous line), obtained for the highest hydraulic flow rate applied at the Tailfer disinfection step
(2900 m³.h-1) and duration of the lag phase of 11.4 minutes (dashed line).
Assessment of the Tailfer disinfection step efficiency
Curves of the same duration of the lag phase of Bacillus subtilis spores in a
plane ( S O3 ,chamber1 , Tchannel ), obtained with the lag time correlation [12] for the ozonation step of the
Tailfer plant, are presented in figure 4. Each duration of the lag phase can be compared with the
contact time distribution between microorganisms and ozone in the Tailfer ozonation channel in
order to evaluate the percentage of non-inactivation of Bacillus subtilis spores. The comparison is
illustrated in figure 3 with a duration of the lag phase of 11.4 min. This duration of the lag phase
(dashed vertical line) is superimposed on the contact time distribution between microorganisms
and ozone in the ozonation channel (continuous line). The percentage of non-inactivation of
Bacillus subtilis spores is obtained by computing the fraction of the microorganisms in contact with
the ozone during a period comprises between the time 0 and the time t  11.4 min with the
t 11.4 min
equation (2): %non  inactivation  100
 E(t )dt
(2)
0
Every duration of the lag phase can thus be associated with a percentage of non-inactivation of
Bacillus subtilis spores in the Tailfer ozonation channel. This association is presented in figure 4 for
five different durations of the lag phase. Therefore, the figure 4 gives a tool to adjust the dissolved
ozone concentration in the ozonation channel according to the temperature in order to reach a
fixed percentage of non-inactivation of Bacillus subtilis spores in the channel.
IOA-EA3G Conference & Exhibition, Toulouse, France
– June 4-6, 2012
Figure 4: Curves of the same duration of the lag phase/iso-percentage curves of non-inactivation of Bacillus
subtilis spores for the temperatures, the ozone concentrations, and the highest hydraulic flow rate applied at
the Tailfer disinfection step.
In figure 4, the only conditions which result in an inactivation of 100 % of the Bacillus subtilis
spores are grouped in the hatched area in the right corner of the figure. The conditions included in
this area are high temperatures and high ozone concentrations, rarely used in the plant. The
results highlight a high risk for incomplete disinfection of highly resistant microorganisms similar to
Bacillus subtilis spores for the highest hydraulic flow rate applied at the Tailfer ozonation channel.
This outcome tallies with the results obtained by Talvy et al. (2011).
Conclusions
In this work, a method to assess the efficiency of a drinking water plant to inactivate resistant
microorganisms is proposed. This method consists in comparing the duration of the lag phase of
ozone inactivation of pathogenic resistant microorganisms similar to Bacillus subtilis spores with
the contact time distribution of the microorganisms with the disinfectant in the disinfection step. The
proposed method is applied to an existing drinking water plant using the ozone and operated by
Vivaqua (Belgium).
To evaluate the duration of the lag phase of ozone resistant microorganisms, the ozone
inactivation kinetics of these microorganisms is experimentally established. From experimental
data, the duration of the lag phase of the ozone inactivation of Bacillus subtilis spores is identified
for different temperature and ozone concentration conditions.
The contact time distribution between the microorganisms and the ozone at the Tailfer disinfection
step is established by using the CFD tool proposed by Talvy et al., 2011.
Durations of the lag phase and contact times are both of the same order of magnitude of a few
minutes in the operating conditions of the Tailfer ozonation channel, for the highest hydraulic flow
rate applied. The comparison of the durations of the lag phase for different temperature and ozone
concentration conditions with the contact time distribution highlights the risk for incomplete
disinfection. For most temperature and ozone concentration conditions, the contact time of a large
proportion of the microorganisms entering in the ozonation channel is shorter than the duration of
the lag phase. These microorganisms are thus probably not inactivated at the exit of the drinking
water plant. By applying non extreme hydraulic flow rates at the ozonation channels, higher
disinfection efficiency will be reached.
IOA-EA3G Conference & Exhibition, Toulouse, France
– June 4-6, 2012
Further research is required to improve and complete the method developed in this paper. For
example, the proposed inactivation kinetics of ozone resistant microorganisms could be integrated
in the CFD tool to follow the resistant microorganisms inactivation through the ozonation channel.
Improvements of the ozonation channels could also be proposed and evaluated with the CFD tool.
Acknowledgements
The authors want to acknowledge Vivaqua for their industrial collaboration and the F.R.S.-FNRS
Fonds de la Recherche Scientifique for their financial support.
References
1.
European Union Council Directive 98/83/EC. 3 November 1998. The quality of water intended for
human consumption. Official Journal of the European Communities 330, 32–54 (1998).
2. Audenaert, W.T.M., Callewaert, M., Nopens, I., Cromphout, J., Vanhoucke, R., Dumoulin, A., Dejans,
P., Van Hulle, S.W.H., Full-scale modelling of an ozone reactor for drinking water treatment. Chemical
Engineering Journal 157, 551–557 (2010).
3. Cockx, A., Do-Quang, Z., Liné, A., Roustan, M., Use of computational fluid dynamics for simulating
hydrodynamics and mass transfer in industrial ozonation towers. Chemical Engineering Science 54,
5085–5090 (1999).
4. Talvy, S., Debaste, F., Martinelli, L., Chauveheid, E., Haut, B., Development of a tool, using CFD, for
the assessment of the disinfection process by ozonation in industrial scale drinking water treatment
plants. Chemical Engineering Science 66, 3185–3194 (2011).
5. Dow, S. M., Barbeau, B., von Gunten, U., Chandrakanth, M., Amy, G., Hernandez, M., The impact of
selected water quality parameters on the inactivation of Bacillus subtilis spores by monochloramine and
ozone. Water Research 40, 373–382 (2006).
6. Driedger, A., Staub, E., Pinkernell, U., Mariñas, B., Köster, W., von Gunten, U., Inactivation of Bacillus
subtilis spores and formation of bromate during ozonation. Water Research 35 (42), 2950–2960 (2001).
7. Larson, M.A., Mariñas, B.J., Inactivation of Bacillus subtilis spores with ozone and monochloramine.
Water Research 37, 833–844 (2003).
8. Mazoua, S., Chauveheid, E., Aerobic spore-forming bacteria for assessing quality of drinking water
produced from surface water. Water Research 39, 5186–5198 (2005).
9. Rennecker, J. L., Mariñas, B. J., Owens, J. H., Rice, E. W., Inactivation of Cryptosporidium parvum
oocysts with ozone. Water Research 33, 2481–2488 (1999).
10. Facile, N., Barbeau, B., Prévost, M., Koudjonou, B., Evaluating bacterial aerobic spore as a surrogate
for Giardia and Cryptosporidium inactivation by ozone. Water Research 34 (12), 3238–3246 (2000).
11. Bader, H., Hoigné, J., Determination of ozone in water by the indigo method. Water Research 15, 449–
456 (1981).
12. Talbot, P., Martinelli, L., Talvy, S., Chauveheid, E., Haut, B., Ozone inactivation of resistant
microorganisms: laboratory analysis and evaluation of the efficiency of plants. Submitted in Water
Research in April 2012.