European Journal of Electrical Engineering
Vol. 24, No. 1, February, 2022, pp. 41-46
Journal homepage: http://iieta.org/journals/ejee
Comparative Study and Experimental Optimization of Ozone Generators by Cylindrical
Surface Dielectric Barrier Discharge (DBD)
Rachida Hedara1*, Said Nemmich2
1
Faculty of Technology, Aboubekr Belkaïd University of Tlemcen, Tlemcen 13000, Algeria
2
APELEC Laboratory, Djillali Liabes University of Sidi Bel-Abbes, Sidi Bel Abbes 22000, Algeria
https://doi.org/10.18280/ejee.240106 ABSTRACT
Received: 24 May 2021 The dielectric barrier discharge (DBD) is a more efficient medium of ozone production,
Accepted: 10 January 2022 which requires voltages of the order of several kV and frequencies of a few tens of kHz.
The principal objective of this study is to model and optimize a cylindrical-shaped ozone
Keywords: generator with a surface DBD that will be used in a water treatment plant. By playing on
ozone generators, DBD generator, surface the electrical parameters (voltage, oxygen flow) and geometric parameters (the nature and
discharge, experimental design, response diameter of the dielectric, the length of the electrode), to achieve high ozone efficiency
surface modeling and minimum energy consumption. It is recommended to use cylindrical surface discharge
reactors with reduced diameter and electrode length and with the smallest possible oxygen
circulation interval to achieve satisfactory ozone concentrations, with reduced power
consumption.
1. INTRODUCTION 2𝑂2 + 𝑂 ⇌ 𝑂3 + 𝑂2 (2)
Significant research effort has been made in this area to Ozone concentration in nature varies between 0.01 ppm to
carry out water treatment, new, high-performance techniques 0.05 ppm, depending on geographic location and season. The
have emerged from scientific researchers in diverse domains concentration of ozone increases with increasing area of
(chemistry, physics and electronics), increasingly interested in electrode and decrease with increasing flow gas rate and gap
the identification and elimination of polluting elements. space [6]. In addition, ozone concentrations increased with
Efficient generation of ozone by cold atmospheric plasmas is increasing AC high-voltage amplitude and with increasing
interesting for sterilization and decontamination of thermally- residence time [7].
sensitive surfaces [1]. The idea is to produce surface DBD ozone in order to
Processes (physical, chemical, electro-technical and compare the different configurations in terms of ozone
electronic, etc.) have been widely studied to eliminate dyes concentration and energy efficiency.
polluting wastewater at high concentrations, plasmas among Over and above, the work consists of experimentally
these treatments to have pure water is the treatment of water optimizing these types of surfaces DBD by indicating the
with ozone which is generated by DBD dielectric barrier results of each dielectric barrier discharge realized.
discharge, such a discharge occurs between two metal
electrodes separated by one or more thicknesses of dielectric
material and a gas space where the discharge occurs [2-4]. 2. GENERAL ON THE SURFACE DBD OZONE
The production of ozone O3 is made from air or O2. Initially, GENERATOR
an oxygen atom is dissociated in collisions between high-
energy electrons and oxygen molecules. Alongside the Among the different designs of ozone generator, we have
radicals combine with free molecules O2 and form O3. These chosen a configuration that is most often used in industry,
processes are described conceptualized in Figure 1 [5]. namely a cylindrical shape. This cylindrical shape has the
advantage of offering a larger discharge surface compared to
flat and pointed shapes; therefore, it allows to generate a
higher ozone yield.
Obtained results showed that although the majority of ozone
generators are of cylindrical surface discharge type, the
cylindrical surface DBD presents significant superiority in
terms of ozone generation and energy efficiency [8].
Types of surface-mounted DBD reactors have been
developed and studied, in terms of ozone concentration and
Figure 1. Schematic diagram of formation of ozone by energy consumed.
electron impact The surface reactor is made with the glass tube dielectric
barrier, in which the outer stainless steel electrode has been
𝑒 + 𝑂2 → 𝑂 + 𝑂 + 𝑒 (1) replaced by a 220 mm long adhesive aluminum strip glued to
41
the outer surface of the glass. The internal high voltage an external diameter of 40 mm, almost in contact with the high
electrode has been replaced by a 240 mm long stainless steel voltage electrode; the 2 mm gas flow interval between the two
grid is composed of several elementary meshes with an area of coaxial tubes.
1 mm2 each with a thickness of 0.75 mm, in contact with the
interior of the glass tube. Micro-discharge occurs on the 3.4 For a DBDS4
surface of the electrode in contact with the glass. Micro-
discharges are generally generated at the high voltage In this configuration, we played on the diameter of the
electrode in contact with the inner surface of the dielectric dielectric and the length of the ground electrode. For which we
barrier. have reduced the diameter of the glass tube by another glass
tube of the same nature of the dielectric with a reduced internal
diameter equal to 38 mm. In order to keep the same discharge
3. REALIZATION AND STUDY OF THE DIFFERENT surface, we have increased the length of the adhesive
VERSIONS OF THE SURFACE DBD REACTOR aluminum ground electrode to 230 mm.
In this part, we have made an experimental study to realize
ozone generators with a dielectric barrier "surface" in several 4. EXPERIMENTAL STUDY OF THE SURFACE DBD
configurations, with two different dielectric materials (glass REACTOR
and ceramic).
We are interested in this work in well-defined forms of: In these studies, colored water is used to estimate the level
Ground electrode: a 220 mm long aluminum adhesive strip of ozone production by calculating the rate of discoloration τ
mounted and glued to the outer surface of the glass tube, the measured from the absorbance before and after the treatment
latter, acting as a dielectric barrier with an outer diameter of by a visible spectrophotometer. The general estimate of the
50 mm and a length of 300 mm and a thickness of 1.5 mm. discoloration rate is given by the following equation:
High voltage electrode: a metal grid made of high voltage
stainless steel of cylindrical shape with an outer diameter of 𝐴𝑏𝑠0 − 𝐴𝑏𝑠1
43 mm and a length of 255 mm and a thickness of 0.75 mm. 𝜏= 𝑥100 (3)
𝐴𝑏𝑠0
This grid contacts the inner surface of the dielectric barrier.
where, Abs0: absorbance before treatment; Abs1: absorbance
3.1 For a DBDS1 after treatment for a determined wavelength.
The configuration of Figure 2 shows that the entire internal 4.1 Effect of oxygen flow on ozone
volume of the glass tube occupied by the gas.
Figure 4 gives the concentration of the variation in the rate
of discoloration as a function of the oxygen flow rate applied
for the ozone generators DBDS1, DBDS2, DBDS3, DBDS4.
We consider a fixed supply voltage V=6 Kv and we vary the
oxygen flow Do=1 to 5 L/min.
100
90
Figure 2. Surface DBD ozone reactor (DBDS1) 80
Taux de decoloration en (%)
3.2 For a DBDS2
70
60
Figure 3 illustrates the fact of the decrease in the internal 50
volume of the gas by inserting a small-diameter glass tube with 40
an external diameter of 32 mm inside the reactor in a coaxial DBDS1
manner, such that the gas flow interval between the two tubes 30 DBDS2
DBDS3
is 6 mm. 20 DBDS4
10
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5
Debit d'oxygene Do(L/min)
Figure 4. Evolution of the discoloration rate as a function of
the oxygen flow rate for the different configurations (V=6
kV)
We can first observe that the evolution of the discoloration
Figure 3. DBDS3 reactor with tube inside, 6 mm gap rate increases with the flow of oxygen injected for each
between glass and Teflon tubes configuration of DBD.
In fact, the water discoloration rate reaches its maximum
3.3 For a DBDS3 89.5% at an oxygen flow rate of 3 L/min with the
configuration at DBDS4. This can be justified by the increase
To further minimize the internal volume of the gas, we in the quantity of oxygen molecules subjected to electric
inserted a second Teflon tube inside the DBDS1 reactor with shocks transforming into ozone.
42
Furthermore, the observation of a decrease in the Moreover, the maximum discoloration rate achieved with
discoloration rate after the value 3 L/min, this is can be the DBDS4 configuration is 84.5% for an electrode length of
explained by the increase in oxygen flow beyond 3 L/min 200 mm. On the other hand, the use of the 120 mm length
makes it possible to increase the speed of the oxygen achieves a discoloration rate of around 71.2%.
molecules, therefore a short residence time in contact with the Consequently, the comparison between the two
discharge surface. This will cause some oxygen molecules not configurations DBDS3 and DBDS4 shows the advantage of
to be subjected to micro-discharge, and consequently a using the configuration of DBDS4 in ozone production as it
decrease in the ozone concentration (more and more oxygen carries a reduced dielectric diameter.
molecules compared to ozone molecules).
90
DBDS3
4.2 Effect of voltage on ozone concentration DBDS4
85
Figure 5 gives the variation in the rate of discoloration as a
Taux de decoloration en (%)
function of the voltage applied for the ozone generators 80
DBDS1, DBDS2, DBDS3, DBDS4. We consider a fixed
oxygen flow Do=5L/min and we vary the voltage from 3 to 6 75
Kv.
70
90
DBDS1
DBDS2
DBDS3
65
80
DBDS4
Taux de decoloration en (%)
70 60
110 120 130 140 150 160 170 180 190 200 210
Longueur d'electrode L(mm)
60
50
Figure 6. Evolution of the discoloration rate as a function of
the length of the ground electrode (V=6 kV, DO2=5 L/min)
40
4.4 Effect of voltage on power consumption
30
2.5 3 3.5 4 4.5 5 5.5 6 6.5
The power factor is a product of the voltage V delivered by
Tension appliquee V(Kv) the power supply applied to the ozone generator and the
current I were measured using a digital oscilloscope.
Figure 5. Evolution of the discoloration rate as a function of
the applied voltage, for the four configurations (DO2=5 1 𝑇
𝑃(𝑊) = ∫ 𝑈(𝑡). 𝐼(𝑡). 𝑑𝑡 (4)
L/min) 𝑇 0
The results obtained show that the rate of discoloration τ 75
DBDS3
increases linearly with the voltage applied for all 70 DBDS4
configurations of the reactor. 65
The increase in voltage increases the rate of dissociation of
Puissance consommée P(W)
60
oxygen molecules.
55
We have seen a great improvement with the DBDS4
50
configuration, which has about 84.2% higher discoloration
rate at 6 Kv voltage compared to DBDS1, DBDS2 and DBD3, 45
which are adopted by high discharge volume, corresponding 40
to larger quantity of oxygen passes through the generator 35
without being transformed into ozone. 30
To maintain an ozone concentration at higher values we 25
decrease the volume of gas flow that it produces an increase in
20
micro-discharges. 2.5 3 3.5 4 4.5 5 5.5 6 6.5
Tension appliquee V(Kv)
4.3 Influence of the electrode length
Figure 7. Evolution of the power consumed as a function of
We consider the DBD3 and DBDS4 configurations after the voltage applied for the DBDS3 and DBDS4 reactors
having their best previous experimental results in terms of (DO2=5 L/min)
ozone concentration.
Both configurations with a fixed supply voltage V=6 Kv, a Figure 7 illustrates a progressively increase in the power
constant oxygen flow rate 5 L/min and the electrode length consumed as a function of the voltage applied for the two
L=120 to 200 mm is varied. configurations.
In Figure 6, the results of these reactors clearly show us that We have remarked that applying the voltage of 3 and 6 kV
the rate of discoloration (the concentration of ozone molecules) on the DBDS4 consumes less power than DBDS3. Indeed, the
continuously increases with increasing electrode length, results confirm that the application of high voltage with the
resulting in increased micro-discharge area. DBDS4 configuration results in minimal power consumption.
43
5. IMPROVEMENT OF THE OZONE GENERATOR of influence on the responses. We have clearly recorded that
TO DBDS4 the applied voltage and the electrode length used have a
positive effect on the rate of discoloration and therefore the
When we reduced the diameter of DBDS4 its ozone yield generation of ozone on the other hand the flow of oxygen has
was very satisfactory, we observed a significant production of a less effect in comparison with the other factors on the surface
ozone and a reduction in energy consumption. DBD.
This result is the fruit of an improvement in the electrical In addition, it can be seen that the power consumed is
and geometric parameters of this generator, namely the applied influenced by two input factors, the applied voltage and the
voltage, oxygen flow rate and electrode length [9, 10]. In order length of the electrode used.
to increase its energy efficiency, we use the plan method Figure 8 represents the iso-response contour plots obtained.
experiences. Besides, the values of the input factors (voltage, mass
We have studied all kinds of DBDS structure in order to electrode, oxygen flow) are optimized by this software for the
capture the optimal values. In our case, we are evaluating the different mathematical models of two responses, in order to
values of the electrical and geometric input factors of the obtain a maximum ozone yield (the areas in orange) and also
DBDS so that they give us a maximum ozone generation rate the minimum energy consumed.
and a minimum power consumption achieved by different
experiments.
It is not easy to realize such a relationship between the
response and the factors considered and above all, we know
that there are interactions between the latter.
We have developed very suitable experimental studies
based on a methodical work of composite experiment hoists
[11, 12], which serve to optimize the organization of the trials.
Table 1 gives the results of the experiments of the CCF
experimental design for the two significant responses, the
discoloration rate (%) and the power consumed P (W).
Table 1. Results of the face-centered composite design in 17
experiments
Exp
Applied Oxygen Electrode
Discoloration
Consumed (a) Discoloration rate (Do=3 L/min)
voltage flow length power
No rate (%)
V(Kv) Do(L/min) L(mm) P(W)
1 4 2 160 65 30,32
2 6 2 160 74,5 43,69
3 4 4 160 64,36 30,32
4 6 4 160 73,2 43,62
5 4 2 200 69,43 41,6
6 6 2 200 77,93 57,03
7 4 4 200 78,48 41,67
8 6 4 200 83,01 57,13
9 4 3 180 77,3 31,4
10 6 3 180 86,14 53,42
11 5 2 180 66,59 42,18
12 5 4 180 71,11 42,18
13 5 3 160 70,2 38,37
14 5 3 200 78,06 48,83
15 5 3 180 77,3 42,18 (b) Discoloration rate (L=180 mm)
16 5 3 180 77,3 42,18
17 5 3 180 77,3 42,18
The mathematical models are shown by the following
equations for the rate of discoloration and the power consumed
respectively.
𝛕 = 76,59 + 3,96 L + 4,02 V + 1,67 Do + 5,65 V 2
− 7,21 Do2 − 1,93 L2 − 0,66 V ∗ L (5)
− 0,57 V ∗ Do + 2,00 Do∗ L
𝐏 = 42,32 + 5,99 L + 7,95 V + 0.009 Do
+ 0,02 V 2 − 0,25 Do2 + 1,16 L2
(6)
− 0,004 V ∗ Do + 0,52 V ∗ L
+ 0.03 L∗ Do
(c) Power consumption
From the results obtained by the simulation, the values of
the factors of the mathematical model can be taken as a degree Figure 8. Iso-response contours for ozone concentration
44
These simulation results allow us to conclude that the best explained by the difference in the values of the permittivity
way to control the ozone production and the power consumed between the two materials. Since the permittivity of glass is
is to apply a high voltage and a long enough electrode length. greater than that of ceramic, therefore the electric field
The program also has an optimization subroutine, which generated by the glass DBD is larger, this results in the
gives the optimal values of the factors to obtain the maximum dielectric barrier discharge being more intense and the ozone
ozone concentration, the optimal values proposed of iso- efficiency rate is bigger.
response contours for ozone concentration (a) and (b) are: Moreover, at high temperatures the resistivity of glass is
lower than that of ceramic. Indeed, several glass barriers of the
V=6 kV, L=200 mm and Do=3L/min for τ=87.7% DBD were broken by the high temperatures due to high
voltages. In continuous life we find glasses that withstand high
For a minimum power, the optimal values proposed of iso- temperatures but with a manufacturing cost is too high.
response contours for ozone concentration (c) are:
V=4 kV, L=160 mm and Do=3L/min for P=30 W.
6. CERAMIC DIELECTRIC MATERIAL
Another experimental study is proposed. This is devoted to
the DBDS4 configuration with a ceramic dielectric.
Table 2 gives the results of the experiments of the CCF
experimental device on the DBDS4 reactor with a ceramic
dielectric for the two significant responses, the discoloration
rate (%) and the power consumption P(W).
Table 2. Results of the face-centered composite design in 17
experiments
Applied Oxygen Flow Electrode Figure 9. Iso-response contours for Discoloration rate with
Exp Discoloration
voltage Do length ceramic dielectric (Do=3 L/min)
No rate (%)
V(Kv) (L/min) L(mm)
1 4 2 160 36
The optimal values proposed by Figure 9 about the iso-
2 6 2 160 46
3 4 4 160 35,4 response contours for ozone concentration are:
4 6 4 160 44,4
5 4 2 200 40,5 V=5.43 kV, L=200 mm and Do=3L/min for τ=50.4%
6 6 2 200 49
7 4 4 200 49,5
8 6 4 200 54,2 7. CONCLUSIONS
9 4 3 180 48,6
10 6 3 180 57,2 In this chapter we carried out experiments on the surface
11 5 2 180 37,6
dielectric barrier discharge by playing on the geometric
12 5 4 180 42,4
13 5 3 160 41,5 parameters (the diameter and the nature of the dielectric, the
14 5 3 200 50 electrode length, the gas flow volume) and on the electrical
15 5 3 180 48,5 parameters (Applied voltage, injected oxygen flow).
16 5 3 180 48,5 We have found that the best way to control ozone generation
17 5 3 180 48,5 is to use a reduced glass dielectric diameter with a long enough
ground electrode length in this high voltage powered process.
The mathematical models are shown by the following Furthermore, in terms of minimum power consumption, the
equations for the rate of discoloration. electrode length had to be minimized as well as the applied
voltage.
𝛕 = 47,88 + 3,99 L + 4,02 V + 4,08 Do − 7,42 V 2 The results obtained on the surface DBD in glass give us a
+ 5,47 Do2 − 1,67 L2 + 2,05 V ∗ L (7) satisfactory result comparable with those of the surface DBD
− 0,59 V ∗ Do − 0,72 Do∗ L in ceramic.
We conclude that this project is very beneficial for science
We have found that the rate of discoloration and therefore and technology and opens up unlimited horizons in this
the generation of ozone seems to be strongly influenced by the domain.
high voltages applied, the length of the electrode used and the
flow rate of oxygen injected.
Moreover, it has been remarked that the stress factor REFERENCES
remains a factor has a lower effect in comparison with the
other factors on the rate of discoloration. [1] Homola, T., Pongrác, B., Zemánek, M., Šimek, M.
These curves in Figure 9 clearly show that the ceramic (2019). Efficiency of ozone production in coplanar
dielectric barrier discharge is less intense in comparison with dielectric barrier discharge. Plasma Chemistry and
the glass dielectric barrier discharge. Indeed, this can be Plasma Processing, 39(5): 1227-1242.
45
https://doi.org/10.1007/s11090-019-09993-6 Chemistry and Plasma Processing, 30(5): 607-617.
[2] Chang, J.S., Kelly, A.J., Crowley, J.M. (1995). https://doi.org/10.1007/s11090-010-9245-4
Handbook of electrostatic processes. CRC Press. Ozone [8] Nassour, K., Brahami, M., Nemmich, S., Hammadi, N.,
Generation and Applications, 26: 581-605. Zouzou, N., Tilmatine, A. (2016). Comparative
[3] Nassour, K., Brahami, M., Nemmich, S., Hammadi, N., experimental study between surface and volume DBD
Zouzou, N., Tilmatine, A. (2017). New hybrid surface– ozone generator. Ozone: Science & Engineering, 38(1):
volume dielectric barrier discharge reactor for ozone 70-76. https://doi.org/10.1080/01919512.2015.1095632
generation. IEEE Transactions on Industry Applications, [9] Wei, L., Hu, Z., Zhang, Y., Wu, Q. (2010). Experimental
53(3): 2477-2484. study on ozone generation and ozone oxidation to
https://doi.org/10.1109/TIA.2017.2675978 removal multi-pollutant of flue gas. In 2010 4th
[4] Bouregba, N., Benmimoun, Y., Meddah, B., Tilmatine, International Conference on Bioinformatics and
A., Ouldmoumna, A. (2016). Ozonation of wastewater in Biomedical Engineering, Chengdu, China, pp. 1-4.
Algeria by dielectric barrier discharge. Desalination and https://doi.org/10.1109/ICBBE.2010.5515999
Water Treatment, 57(4): 1824-1835. [10] Delamézière, A., Naceur, H., Breitkopf, P., Knopf-
https://doi.org/10.1080/19443994.2014.979440 Lenoir, C., Batoz, J.L., Villon, P. (2002). Faisabilitéen
[5] Shrestha, R., Joshi, U.M., Subedi, D.P. (2015). emboutissage: Optimisation du matériau par surface de
Experimental study of ozone generation by atmospheric réponse. Mécanique & Industries, 3(2): 93-98.
pressure dielectric barrier discharge. International [11] Nemmich, S., Bellebna, Y., Bendaoud, A., Tilmatine, A.,
Journal of Recent Research and Review, 8(4): 24-29. Hammadi, N., Ouidir, R. (2013). Étude Expérimentale de
[6] Murbat, H.H. (2014). Effects of applied voltage and flow L’Influence des Paramètres sur l’Efficacité du Procède
rates of ozone generator fed by dry air and O 2 on the de Génération d’Ozone par Décharge à Barrière
coaxial wire-cylinder reactor by varying various Diélectrique.
electrodes parameters. International Open Access. [12] Remaoun, S.M., Tilmatine, A., Hammadi, N., Miloua, F.,
Journal of Modern Engineering Research (IJMER), 4(9): Medles, K. (2012). Optimisation of one stage
56-60. electrostatic precipitator for welding fume filtration.
[7] Šimek, M., Pekárek, S., Prukner, V. (2010). Influence of Scientia Iranica, 19(6): 1861-1864.
power modulation on ozone production using an AC https://doi.org/10.1016/j.scient.2012.10.016
surface dielectric barrier discharge in oxygen. Plasma
46
