Chemical Engineering Journal 481 (2024) 148488

Contents lists available at ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier.com/locate/cej

Ozone-assisted degradation of 2-methoxyethanol in a prototype plug flow

photocatalytic reactor
Kristen Altof a, *, Marina Krichevskaya a, Sergei Preis a, Toivo Tähemaa b, Juri Bolobajev a, *
a
Laboratory of Environmental Technology, Department of Materials and Environmental Technology, Tallinn University of Technology, Ehitajate tee 5, Tallinn, Estonia
b
Department of Mechanical and Industrial Engineering, Tallinn University of Technology, Ehitajate tee 5, Tallinn, Estonia

A R T I C L E I N F O A B S T R A C T

Keywords: Airborne volatile organic compounds (VOCs) present an increasingly relevant health hazard addressed by the EU
VOCs directive 2016/2284 requiring 40-% reduction in emissions of non-methane VOCs by 2030 in comparison to
Methyl cellosolve 2005. Advanced oxidation processes (AOPs) provide a solution for the VOC problem in industry. Among AOPs,
Advanced oxidation process
pulsed corona discharge (PCD) and photocatalytic oxidation (PCO) are of particular interest in their synergy
Air treatment
Scale up
complementing each other’s strengths in energy-efficient manner. In the present study, air polluted with 2-
Photocatalysis methoxyethanol (2ME) was treated by using a prototype 21.3 L photocatalytic reactor for the pollutant degra­
dation at its variable concentrations (6–50 ppm), irradiances (65–119 W m− 2) and residence times (12–77 sec).
Further combination of PCO reactor with forthcoming PCD requires studies in ozone photocatalytic degradation
as a standalone byproduct and in combination with 2ME. The analysis of variable conditions resulted in singling
out the PCO major restrictive parameters, also demonstrating the degradation enhancement, when both 2ME and
ozone were present in air. The PCO treatment combined with ozone resulted in degradation of 2ME and ozone for
40 % and 95 %, respectively.

1. Introduction pollution with VOCs is bound to worsen.

Various exhaust air treatment strategies have been developed to
Volatile organic compounds (VOCs) of toxic character in- and out­ reduce air pollution [7,8] including filtration systems using adsorbents
doors present an increasingly relevant health hazard to be addressed for with traditional drawbacks of VOCs transfer to the solid phase at a
proper elimination. It has been found that the indoor air pollutants limited capacity. To truly rid the air from pollutants, these must be
originate from cleaning products, dwelling construction materials, degraded to less harmful substances, ideally to carbon dioxide and
kitchen activities and furniture, whereas outdoor sources include traffic water.
and neighboring industries [1–4]. Keeping in mind that VOCs are Oxidation approaches used for air purification include advanced
generally hazardous, the countries of the European Union took an oxidation processes (AOPs) - ozonation, photocatalytic oxidation (PCO),
obligation to reduce the emissions of non-methane volatile organic and electric discharge plasma [9–11]. Among them, pulsed corona
compounds (NMVOCs) for 40 % by 2030 as stated in the directive (EU) discharge (PCD) [12,13] and photocatalytic oxidation (PCO) [14]
2016/2284 of the European Parliament and council states [5]. This puts deserve particular attention for showing promising results in degrading
the European Union under pressure to apply innovative technologies VOCs at high energy efficiency, demonstrating the best performance
reducing the NMVOCs emissions to achieve the ambitious goal. when used in combination [15]. This explains the keen interest to test
Industrial air pollutant emissions reported by Estonia in 2020 to the the performance of combined PCD and PCO in a twostep combination
EU Environmental Board reached 1,260 tons of VOCs [6], demonstrating for air purification.
the industrial exhaust as a relevant source of atmospheric pollution. Pulsed corona discharge technology studied for oxidation of water­
Wood fabrication and furniture production industries stick out with borne organic pollutants has been showing remarkable energy efficiency
using solvents and paint bases, e.g., 2-ethoxyethanol (2EE). Twenty-two [16,17]. One of the main problems of PCD application in air treatment,
Estonian enterprises declared emissions of 2EE in a total amount of 13.7 however, is residual ozone, a side product of air ionization considered as
t. Considering the ever-growing demand in goods, the problem with air the secondary air pollution. Photocatalytic processes are known to

* Corresponding authors.
E-mail addresses: [email protected] (K. Altof), [email protected] (J. Bolobajev).

https://doi.org/10.1016/j.cej.2023.148488
Received 1 September 2023; Received in revised form 30 November 2023; Accepted 30 December 2023
Available online 31 December 2023
1385-8947/© 2024 Elsevier B.V. All rights reserved.
K. Altof et al. Chemical Engineering Journal 481 (2024) 148488

degrade ozone thus solving the problem, usefully extending the photo­ 2.2. Air pollution
catalyst lifetime [18–21]. Further understanding the PCO mechanism in
simultaneous VOCs and ozone abatement with spreading the knowledge The 2ME vaporization setup was operated in continuous mode
to a wide spectrum of volatiles is a pre-requisite for the progress in the (Fig. 1). For the generation of polluted air, 2ME was vaporized using a
combined PCD/PCO applications. bubble column 3. The Eco Air Pump PA200 (Jecod Co. Ltd., People’s
A prototype PCO reactor was built and tested with ozone and 2- Republic of China) with flow rate varying from 1 to 6 m3 h− 1 fed air
methoxyethanol (2ME) in combinations. 2ME was chosen as a more through the main line to the reactors. A bypass controlled by valve 5
hazardous model VOC rather than similar glycol ether 2EE for both directed a 1/400 to 1/100 vol part of the main air flow through a bubble
substances being used in solvents and carriers in paints [22], with the column 3 (H = 700 mm; di = 50 mm) filled with liquid 2ME (750 ml). As
occupational exposure limit of 1 ppm [23–25]. The refractory character the air passed through the column, it became saturated with vapor and
of 2ME brings to front the experimental effort targeting the abatement of was mixed with the main flow through the valve 7 and subsequently
a recalcitrant VOC in a pilot-size PCO reactor thus making a good step directed to the PCO reactor. Ozone was generated onsite in the ozone
forward in the technology implementation. This gives incentive to study generator with the flow controlled by valve 6 and mixed with the main
the energy efficient AOP technologies helping to avoid health problems air flow. The VOC concentration was regulated and maintained with air/
and damage to the biosphere caused by 2ME emissions applying a larger bypass flow ratio. All experiments were performed at the room tem­
scale reactor. In the study, the 2ME industrial pollution was simulated perature, 22 ◦ C, atmospheric pressure, and relative humidity of 30 ± 3
with the PCO in a reactor varying light intensity, the pollutant and ozone % RH at 22 ◦ C.
concentrations at high levels, and the air flow rates in a wide span of
values.
2.3. Reactor

2. Materials and methods

The plug flow PCO reactor is flat bedded with a transparent cover (2
mm polyethylene terephthalate, PET) sealing in the gas passing through
2.1. Chemicals and reagents
a maze-like chamber where glass plates covered with TiO2 lay on the
chamber floor as shown in Fig. 2. Entering the PCO reactor, polluted air
Analytical grade 2ME (C3H8O2 ≥ 99 %) was purchased from Fisher
flows through a series of chambers coming into contact with the surface
Chemicals (USA) and used without further treatment. Ozone was
of the photocatalyst. The internal size of the reactor is, mm: L = 996, W
generated onsite from air using 10G Lab Ozone Generator (A2Z Ozone,
= 534, and H = 40, with a total volume of 21.3 L. The reactor is divided
USA). Titanium dioxide P25 was purchased from Evonik Industries AG
into five sections sized 4.26 L with stainless-steel plates with holes in the
(Germany) and ethanol (C2H5OH ≥ 99 %) - from Sigma-Aldrich (USA).
end of each section providing diagonal air flow along the sections. The
air flow rates in the reactor were varied from 1.0 to 6.0 m3 h− 1 corre­
sponding to the residence times from 77 to 16 s, respectively; the

Fig. 1. Experimental setup: 1 – main flow control valve, 2 - main flow rotameter, 3 – 2ME column, 4 – air dryer filled with silica gel for the generation of ozone, 5 –
rotameter and control valve for the 2ME-saturation bypass, 6 – rotameter and control valve for ozone generator, 7 – shut off valve for 2ME bypass, 8 - shut off valve
for ozone generator, 9 – control valve for main airflow sampling before the PCO reactor, 10 - control valve for main airflow sampling after the PCO reactor, 11 –
sampling line rotameter.

2
K. Altof et al. Chemical Engineering Journal 481 (2024) 148488

the catalyst with the pollutant and exclude adsorption from the overall
pollutant’s removal. Initial gaseous samples were collected after the air
had passed through the dark reactor ensuring the adsorption equilib­
rium. After that, the UV-A TL-D 15 W lamps (Philips, Germany) were
switched on and air samples were withdrawn until the outlet 2ME
concentration stabilized. The 2ME concentrations and the formation of
gaseous byproducts were registered with FT-IR spectroscopy using
Interspec 200-X spectroscope (Interspectrum OÜ, Estonia) and analyzed
with Essential FTIR (Operant LCC) software with (FDM, HiRes VPFTIR
for Quant) database. The 2ME content in the air was measured using a
20 m metal body gas cell (Specac, United Kingdom) at a bandwidth
between 2921 and 2838 cm− 1 having the least interference from other
substances. Gaseous ozone content was measured with ozone analyzer
(Anseros Klaus Nonnenmacher GmbH, Germany).

2.5. Preparation of photocatalytic coating

Powdery P25 titanium dioxide AEROXIDE® TiO2 P25 from Evonik ≥

99.5 % TiO2 [26], crystalline with predominantly anatase structure
(rutile fraction 13.4 ± 7 wt%), was used as photocatalytic material. TiO2
was coated on 115 x 90 x 2 mm glass plates by spraying as described by
Fig. 2. Photocatalytic oxidation reactor (The picture is for illustration purpose
Kask et al. [18] and shown in Fig. 4. Forty glass plates covered with 1.53
only; for experiments the lamps were lowered to 7 cm above the reactor).
± 0.11 mg cm− 2 TiO2 were fitted into the reactor, making the overall
catalyst surface area equal to 4,140 cm2. The thickness of the photo­
average air velocity at the reactor inlet was 0.2 – 1.3 m s− 1, and 0.003 –
catalytic coating was measured using the surface profiler TENCOR P-10
0.2 m s− 1 in the reactor section. To assure the turbulent contact of
showing a range of 1 μm. The image of the coating with specific weight
polluted air with the surface of the catalyst, three additional stainless-
of 1.4 mg cm− 2 TiO2 P25 was obtained by the field emission scanning
steel partition walls were placed across the channels providing the
electron microscopy (FE SEM, Dual-Beam Helios Nanolab 600, FEI) and
zigzag trajectory of the gas flow along the section. In order to check the
is presented in supplementary material (Figure S1). The UV-A radiation
airflow distribution in the reactor, a software simulation was performed
was provided by twenty TL-D 15 W lamps positioned at 7.0 cm above the
with SolidWorks (Fig. 3) demonstrating flow trajectory within the
catalyst surface (Fig. 5). The intensity of light was regulated by the
reactor chambers. The experimental runs for PCO of 2ME and decom­
number of switched-on lamps in sets of five distributed evenly along the
position of ozone were carried out with and without the partition walls
reactor.
in each section. No difference in the reactor’s ability to degrade pol­
Spectrometry parameters of UV-A radiation within 315 – 400 nm
lutants, however, was detected at the average flow rate of 5.0 m3 h− 1
were measured using a USB 2000 + UV–VIS spectrometer (Ocean Op­
confirming sufficient turbulence of air passing through the reactor with
tics, USA). The dependence of the average irradiance at the photo­
no additional mass transfer intensification.
catalyst surface on the nominal power is seen in Fig. 6, ranging from 29
to 119 W m− 2 provided by the number of lamps from 5 to 20,
respectively.
2.4. Experimental procedure

Prior to each experimental run, polluted air was fed to the reactor in
a continuous mode for at least 30 min in the absence of light to saturate

Fig. 3. Air flow trajectories simulated in photocatalytic oxidation reactor at the

gas flow of 5.0 m3 h-1. Fig. 4. Glass plates covered with TiO2 by spraying.

3
K. Altof et al. Chemical Engineering Journal 481 (2024) 148488

VOCs and the maintenance of the photocatalyst activity was studied for
its photocatalytic conversion. Degradation of ozone in air with varia­
tions in irradiance and residence time may be seen in Fig. 7. The pho­
tocatalyst irradiance has a clear effect on conversion as quadrupling the
energy input from 29 to 119 W m− 2 increases the conversion of ozone by
the factor of 1.5–1.9. Small impact of irradiance is consistent with the
kinetic regime of ozone decomposition at the photocatalyst surface with
the rate limited by chemical reaction. The air residence time controlled
by the flow rate varied between 7 and 77 sec at the flow rates between
166 and 17 L min− 1, respectively, and also showed conversion reason­
ably improved with extended contact: the residence time increased
almost tenfold provides only doubled ozone conversion confirming the
kinetic limitation of the reaction with excess amounts of ozone.
The role of residence time in the photocatalytic degradation of
substrate could be explained in terms of the mass balance for ideal plug
flow reactor. With respect to a substrate, the mass-balance is written in
general word statement Eq. (1) with corresponding mathematical
expression Eq. (2) as follows:
Accumulation = inflow − outflow + generation (1)

∂C
ΔV = QC|L − QC|L+ΔL + rc ΔV, (2)
∂t

where ΔV, Q, C, L, and rc are differential volume element in reactor

Fig. 5. UV-A lamps above the PCO reactor (The picture is for illustration section, volumetric flow, concentration of substrate in air, flow path,
purpose only; for experiments the lamps were lowered to 7 cm above and reaction rate for heterogeneous photocatalysis, respectively.
the reactor). Eq. (2) undergoes a series of transformations, which are omitted here
for the sake of brevity. The transformations end up by taking the limit as
ΔL approaches zero, where A is the surface area of the reactor Eq. (3):
∂C Q ∂C
= − + rc (3)
∂t A ∂L

With respect to rc, the most common decay model observed in PCO is
Langmuir-Hinshelwood (LH) kinetics of heterogeneous catalysis, Eq. (4)
[29–33]:
kr KC
rc = − (4)
1 + KC

where kr, K, and C are limiting rate constant of reaction, adsorption

equilibrium constant, and substrate concentration, respectively.
Fig. 6. Irradiance of the photocatalyst vs nominal power consumed by the UV- Considering the steady-state conditions (∂C/∂t = 0) and LH reaction rate
A lamps. Eq. (4), the equation Eq. (5) for a plug flow reactor becomes as follows:
Q dC kr KC
2.6. Reference experiments = − (5)
A dL 1 + KC
To exclude the possible role of photolysis induced by UV-A, the ex­ Under condition KC ≪ 1, Eq. (5) could be approximated to the first order
periments were conducted in the absence of photocatalyst. Air mixtures kinetics. The zero order is observed at surface saturation when KC ≫ 1, i.
containing 2ME and ozone at various concentrations were led through e., zero order suggests occupation of all reaction sites by substrate, and
the reactor under UV-A radiation. No changes in 2ME concentration the LH reaction rate approximates to a constant (kobs), Eq. (6),
with or without ozone were observed. Small, below 10 % conversion of
ozone was detected with the residence time as long as 77 sec. The effect Q dC
= − kobs (6)
of ozone photolysis was thus found to be neglectable in the current A dL
study. Even though it has been suggested that ozone degradation in gas
Integrating Eq. (6) between the limits C = C0 and C = C and corre­
phase could play a role in the oxidation rate of VOCs [19] no changes of
sponding L = 0 and L = L yields the linear dependence between resi­
2ME concentration in the presence of ozone and the absence of photo­
dence time and concentration of substrate Eq. (7):
catalyst could be detected. Thus, it leads to the conclusion that gas phase
oxidation is negligible at the residence times applied in the experiments, C = C0 −
A
• L•kobs = C0 − τ • kobs (7)
as was also observed in other studies [18,27,28]. Q

3. Results and discussion A pseudo-zero order reaction with the substrate initial concentration
having no influence on the reaction kinetics has also been observed
3.1. Photocatalytic degradation of ozone earlier in similar flow reactors [34,35]. With respect to ozone degra­
dation, it is evident that linear paths (Fig. 7) have cross-sections with Y-
In anticipation of combining the PCD and PCO reactors to a two-step axis significantly below the initial concentration (Cin = 22.6 ppm),
air purification system, ozone with its important role in oxidation of which is inconsistent with Eq. (7). Besides, clear deviations from the

4
K. Altof et al. Chemical Engineering Journal 481 (2024) 148488

Fig. 7. Photocatalytic degradation of ozone dependent on UV-A irradiance and residence time: initial concentration of ozone (in) 22.6 ppm, inset – ozone degra­
dation at residence time from 7 to 26 sec.

linearity are seen at residence time of 7 sec (Fig. 7, see inset). The reason
could lay in a diffusion-controlled process at short residence times,
where the reaction rate is limited by the substrate transport to the active
sites of the photocatalyst.

3.2. Photocatalytic degradation of 2ME

The effects of 2ME inlet concentration and irradiance on the PCO rate
were studied. The results obtained at the residence time of 12.8 sec are
seen in Fig. 8. The irradiance doubled from 65 to 119 W m− 2 resulted in
only slight change in degradation rate confirming the diffusion-
controlled process. It is reasonable to presume that solar light, the UV-
A component of which does not exceed 90 W m− 2 at the Earth’s sur­ Fig. 9. Photocatalytic degradation of 2ME dependent on the air residence time:
face [36], will not substantially improve the PCO performance if 2ME inlet concentration 19.7 ± 0.2 ppm, irradiance 119 W m− 2.
compared with the studied light irradiances.
2ME degraded in the PCO reactor for about 4 ± 1 ppm within the respectively. Thus, the contact time increased more than six times pro­
tested residence time regardless of its input concentration thus following vides degradation of 2ME accelerated for only about 36 %, indicating
the pseudo-zero order kinetic model. It has also been stated that at very equilibrium established at the photocatalyst surface densely occupied
high light intensities, the experimental data are approximated to the with the VOC molecules attributable to the high pollutant concentra­
zero-order kinetics with respect to the light intensity because of diffu­ tions. Weak dependence of oxidation rate on residence time is consistent
sion limitation for the transport of reagent molecules to the photo­ with the LH description of descendent rate curve approximating a con­
catalyst surface [37]. The substrate degradation rate becomes invariant stant manifesting the kinetic regime of chemisorption Eq. (6). Higher
to the UV-A irradiance. conversions may be achieved by increasing the residence time on ac­
The effect of residence time on PCO of 2ME with its initial concen­ count of substantially decreased flow rate or increased reactor size.
tration 20 ± 2 ppm at the highest irradiance of 119 W m− 2 is seen in
Fig. 9. The residence time varied between 12.8 and 77.0 sec was pro­
vided with the air flow rate from 6.0 to 1.0 m3 h− 1, respectively, 3.3. Photocatalytic degradation of 2 ME in ozonized air
showing a moderate effect on oxidation performance: conversions of
2ME varied from 3.6 to 4.9 ppm at residence times of 12.8 and 77.0 sec, The results of simultaneous photocatalytic degradation of 2ME and

Fig. 8. Photocatalytic degradation of 2ME dependent on its inlet concentration and photocatalyst irradiance at the residence time 12.8 sec.

5
K. Altof et al. Chemical Engineering Journal 481 (2024) 148488

ozone are shown in Fig. 10. The experimental conditions were chosen
imitating PCD and PCO combined action, in which ozone concentration
remains more or less constant determined by the air flow rate and
discharge parameters, whereas the VOC concentration may vary
dependent on external parameters at the point of air pollution. In the
study, the inlet ozone concentration was kept at 33 ± 3 ppm, and con­
centrations of 2ME were chosen as 8 ± 2, 20 ± 5, and 40 ± 10 ppm at
maximum available irradiance and rather short residence time. One can
see a significant difference in 2ME oxidation rates promoted with ozone
compared to unassisted PCO: maximum 2ME degradation observed in
non-assisted PCO did not exceed 5 ppm regardless of its inlet concen­
tration, whereas the one achieved in PCO with ozone tripled the number
at the highest 2ME concentration. Ozone mixed with 2ME at the highest Fig. 11. Photocatalytic degradation of 2ME and ozone in their mixture with
inlet concentrations 19 ± 1 and 26 ± 1 ppm, respectively, irradiance 119
content of the latter also decomposed faster showing 80 % conversion
W m− 2.
(Fig. 7 and Fig. 10): similar ozone conversion rates in absence of 2ME
may only be seen at residence time of 77 sec.
In simultaneous degradation of 2ME and ozone, one can see the Irradiance was kept constant at the highest available level of 119 W m− 2.
conversion of both increased with growing concentration of 2ME. This Conversions of 2ME and ozone seem to follow a pattern of deeper
phenomenon was observed and explained earlier by formation of addi­ ones at longer residence times. The maximum conversion of 2ME at 77
tional oxidative species, e.g., ozonides [38,39]. Characteristic for un­ sec residence time reached 40 %, i.e., 9.5 ppm oxidized out of inlet 24 ±
saturated and aromatic hydrocarbons oxidized by ozone with formation 1 ppm. At the same time, over 95 % of ozone was converted resulting in
of unstable ozonides, these are less probable to form when the 2ME the highest VOC removal in the current study. This confirms the as­
molecule reacts having no moieties attractive for the electrophilic at­ sumptions of residence time and ozone having the strongest impact on
tacks and possibilities for ozonide formation or peroxo bridging. This the reactors performance in VOCs removal. Besides, noteworthy is the
makes ozonides formation mechanism less relevant to the observed degradations of 2ME and ozone not following the pseudo zero-order
oxidation acceleration. Another study pointed out that ozone degrada­ reaction rate when combined. The linearization result suggests a sec­
tion in gas phase could play a role in the oxidation rate of VOCs [19]. Yet ond order reaction kinetics in respect to both reactants with their
photolysis of ozone was found to be neglectable in the current study. It degradation being directly and/or indirectly affected by each other’s
has also been stated that ozone can prohibit the recombination of concentration, thus suggesting PCO not being the only oxidation
electron-hole pairs on the surface of TiO2 [40,41]. Possible pathways of pathway in presence of ozone.
oxidants formation in a PCO reactor in presence of ozone were sum­
marized by Huang et al (2011) [42]. In the presence of TiO2 under UV-A 3.4. Degradation byproducts of 2ME and ozone
irradiation, ozone acts as an electron acceptor and as a source of reactive
oxygen species maintaining the photocatalyst activity [18,27,28,43]. Ideally, the ultimate target of PCO reactors following the PCD ones is
The recombination of electron-hole pairs on TiO2 surface is slowed down complete mineralization of VOCs rendering mute the question of
freeing up electron holes for the oxidation of 2ME. The higher conver­ byproducts. Yet as the complete mineralization of 2ME is not observed,
sion rate of ozone in the presence of 2ME could be explained by the the question of byproducts requires analysis. Three byproducts were
formation of intermediates and overall higher concentration of organic detected in the gas phase, formic acid, methyl formate, and formalde­
material reacting with ozone and ozone-derived radicals freeing up even hyde, having the two formers registered in nearly every experiment
more active sites on the catalyst surface. except for PCO at 65 W m− 2 irradiance. Formic acid and methyl formate
The impact of residence time on 2ME conversion in the presence of reached their maximum concentrations, 10 and 15 ppm, respectively, in
ozone can be seen in Fig. 11. A gaseous mixture of 2ME with ozone at experiments, where 2ME at its inlet concentration of 40 ± 10 ppm was
constant inlet concentrations of 24 ± 1 and 26 ± 1 ppm, respectively, degraded in the presence of ozone regardless of irradiance (Supple­
was treated in the PCO reactor with residence times varied between 77.0 mentary material, Fig S1). Formaldehyde could only be detected at 2ME
and 12.8 sec, i.e., flow rates between 1.0 and 6.0 m3 h− 1, respectively. degradation in the presence of ozone, reaching its concentration up to

Fig. 10. Photocatalytic degradation of 2ME and ozone in their mixtures dependent on 2ME concentration: ozone inlet concentration 33 ± 3 ppm, residence time 12.8
sec, irradiance 119 W m− 2.

6
K. Altof et al. Chemical Engineering Journal 481 (2024) 148488

12 ppm (Supplementary material, Fig S2). irradiance, increasing the UV-A irradiance from 65 to 119 W m− 2
Figure S2 (Supplementary material) shows the amount of interme­ resulted in slight change. Residence time plays an important role in PCO
diate products unaffected by the increased residence time. Since the as degradation of 20 ppm 2ME varied from 3.5 to 5.0 ppm at residence
amount of degraded 2ME increases with residence time, but the content times of 12.8 and 77.0 sec, respectively. At rather high 2ME concen­
of intermediates in gaseous phase changes negligibly, an outrunning trations exceeding 10 ppm, PCO turned out to follow the pseudo-zero-
degradation of intermediates is seen. Thus, further increase in the resi­ order reaction kinetics, i.e., the efficacy of pollutant removal appears
dence time in a reactor or extending of the reactors’ line should lead to invariant to its concentration.
the decomposition of intermediate products. In this perspective, the performance of standalone PCO reactor could
The simplified outline of 2ME PCO is suggested in Fig. 12. The be improved in three ways. Firstly, the activity of photocatalyst needs
degradation of 2ME employing HO• and •O–2 as probable oxidation improvement, i.e., the state-of-the-art materials or coating techniques
species generated by photocatalysis involves primarily hydrogen atom would accelerate PCO. Photocatalytic materials activated by a broader
abstraction, ether bond cleavage, addition, hydroxylation, and decar­ light spectrum than anatase TiO2 could prove useful using more energy
boxylation reactions. The present degradation pathway has priori pre­ from the light sources. Secondly, increasing the photocatalyst surface
dictions and was based on the mechanisms of the gas-phase reactions of should also result in faster PCO [18], e.g., the inner walls and the cover
HO• with ethers, alcohols, aldehydes, and carboxylic acids published by of the reactor could be coated with the photocatalyst. This question
Atkinson [44]. deserves studies for unclear effect of indirect illumination or the coated
Hydroxyl radical attack of hydroxyl group of 2ME (1) is associated cover screening-off the light.
with H atom abstraction at C-H adjacent to hydroxyl group producing Thirdly, residence time is a key factor in photocatalytic oxidation.
the corresponding radical. The radical reacts with atmospheric oxygen Therefore, increasing the residence time by increasing the reactor size or
(2) to yield 2-methoxyacetaldehyde and HO2•. The following series of connecting multiple reactors in series can improve the overall efficiency
reactions including hydrogen abstraction followed by HO• addition (3) of the process as studies have shown before [18,45]. It should be noted
leads to the formation of 2-methoxyacetic acid, which undergoes the that an increase in the reactor size should go along with an increase in
decarboxylation transformation (4) producing carbon dioxide and the surface area of the photocatalytic coating.
dimethylether. The HO• attack proceeds via H atom abstraction from C- However, the true potential of the PCO reactor is revealed when used
H bond of ether (5) producing methoxymethyl radical. The formation of with ozone as a complementary oxidant nearly doubling 2ME degra­
methyl formate from this radical is well described by Atkinson (1986). dation from 5.0 to 9.5 ppm. Ozone extends the lifetime of electron-hole
Instead of ether bond decomposition, the reaction with reactive oxygen pairs on the photocatalyst surface, adding also extra oxidant species
species dominates (6) yielding methyl formate. The formation of more improving the effectiveness of the PCO reactor. This potential is to be
toxic intermediate compounds formaldehyde and formic acid is associ­ disclosed in upcoming research of PCO reactor paired with PCD plasma.
ated with either 2ME ether bond cleavage or further decomposition of It is also important to stress that the PCO should not be used for merely
methyl formate. degrading ozone as pure ozone degradation follows the zeroth order
All formed intermediate products can be successfully mineralized, reaction kinetics and is only influenced by residence time. The PCO
which requires a targeted study of operating conditions for the practical reactor should be used with ozone and small amounts of VOCs which
application of a specific reactor modification. It would be necessary to reveals second order reaction kinetic with the ability to degrade 95 % of
determine the maximum concentrations of VOCs and the minimum ozone and accelerate oxidation of residual VOCs.
residence time at which nearly complete mineralization of the initial
pollutants can be achieved. CRediT authorship contribution statement

4. Conclusions Kristen Altof: Conceptualization, Data curation, Formal analysis,

Investigation, Methodology, Validation, Visualization, Writing – orig­
An upscaled PCO reactor was tested in removal of 2ME and ozone inal draft, Writing – review & editing. Marina Krichevskaya: Concep­
from air being considered as a simulation of PCD and PCO reactors tualization, Data curation, Formal analysis, Methodology, Supervision,
combined in a sequence. The reactor operates continuously and can be Validation, Visualization, Writing – original draft, Writing – review &
used outdoors using the solar radiation to breach the band gap of TiO2, editing. Sergei Preis: Formal analysis, Funding acquisition, Investiga­
thus providing an energy-efficient way to remove VOCs and ozone from tion, Project administration, Resources, Writing – original draft, Writing
the airstreams. However, the results of the present research show the – review & editing. Toivo Tähemaa: Formal analysis, Resources,
reactor being effective only at rather long retention times (77 s) and low Writing – review & editing. Juri Bolobajev: Conceptualization, Data
pollutant concentrations (<10 ppm). For higher pollutant’s concentra­ curation, Formal analysis, Investigation, Methodology, Supervision,
tion exceeding 10 ppm, the PCO reactor is presumed to function effec­ Validation, Visualization, Writing – original draft, Writing – review &
tively in a tandem with PCD partially oxidizing pollutants and providing editing.
ozone. Gaseous intermediate products of 2ME degradation were also
studied. Formaldehyde, formic acid and methyl format could be detec­ Declaration of competing interest
ted. Considering the toxicity of mentioned by-products it is crucial that
in future studies the operating parameters for PCO reactors will be The authors declare that they have no known competing financial
optimized to completely oxidize all intermediate products. interests or personal relationships that could have appeared to influence
The PCO reactor’s performance in degrading 2ME is controlled by the work reported in this paper.

Fig. 12. Rationale of gaseous 2ME degradation induced by reactive oxygen species.

7
K. Altof et al. Chemical Engineering Journal 481 (2024) 148488

Data availability [19] P. Zhang, J. Liu, Photocatalytic degradation of trace hexane in the gas phase with
and without ozone addition: kinetic study, J Photochem Photobiol A Chem. 167
(2004) 87–94, https://doi.org/10.1016/j.jphotochem.2004.05.015.
Data will be made available on request. [20] J. Kim, P. Zhang, J. Li, J. Wang, P. Fu, Photocatalytic degradation of gaseous
toluene and ozone under UV254+185 nm irradiation using a Pd-deposited TiO2
Acknowledgements film, Chem. Eng. J. 252 (2014) 337–345, https://doi.org/10.1016/j.
cej.2014.05.015.
[21] P. Fu, P. Zhang, J. Li, Photocatalytic degradation of low concentration
This work was supported by the Institutional Development Program formaldehyde and simultaneous elimination of ozone by-product using palladium
of Tallinn University of Technology for 2016–2022, project 2014- modified TiO2 films under UV254+185nm irradiation, Appl Catal b. 105 (2011)
220–228, https://doi.org/10.1016/j.apcatb.2011.04.021.
2020.4.01.16-0032 from EU Regional Development Fund. [22] I.N. Sax, Dangerous Properties of Idustrial Materials, Sixth Edition, Van Nostrand
Reinhold Company Inc., New York, 1984.
Appendix A. Supplementary data [23] The European Chemicals Agency, European Chemicals Agency, Https://Echa.
Europa.Eu/Substance-Information/-/Substanceinfo/100.003.377. (2023).
[24] R.R. Miller, E.-A. Hermann, J.T. Young, T.D. Landry, L.L. Calhoun, Ethylene glycol
Supplementary data to this article can be found online at https://doi. monomethyl ether and propylene glycol monomethyl ether: metabolism,
org/10.1016/j.cej.2023.148488. disposition, and subchronic inhalation toxicity studies, Environ Health Perspect. 57
(1984) 233–239.
[25] G. Bagchi, D.J. Waxman, Toxicity of ethylene glycol monomethyl ether: Impact on
References testicular gene expression, Int J Androl. 31 (2008) 269–274, https://doi.org/
10.1111/j.1365-2605.2007.00846.x.
[1] U. Schlink, A. Thiem, T. Kohajda, M. Richter, K. Strebel, Quantile regression of [26] Evonik Industries AG, AEROXIDE® TiO₂ P 25, Https://Www.Silica-Specialist.Com/
indoor air concentrations of volatile organic compounds (VOC), Sci. Total Environ. En/Product/PR_52000356. (2023).
408 (2010) 3840–3851, https://doi.org/10.1016/j.scitotenv.2009.12.002. [27] Z. Pengyi, L. Fuyan, Y. Gang, C. Qing, Z. Wanpeng, A comparative study on
[2] R.D. Edwards, J. Jurvelin, K. Koistinen, K. Saarela, M. Jantunen, VOC source decomposition of gaseous toluene by O3/UV, TiO2/UV and O3/TiO2/UV,
identification from personal and residential indoor, outdoor and workplace J Photochem Photobiol A Chem. 156 (2003) 189–194, https://doi.org/10.1016/
microenvironment samples in EXPOLIS-Helsinki, Finland, Atmos Environ. 35 S1010-6030(02)00432-X.
(2001) 4829–4841. [28] M. Krichevskaya, S. Preis, A. Moiseev, N. Pronina, J. Deubener, Gas-phase
[3] H.L. Wang, L. Nie, J. Li, Y.F. Wang, G. Wang, J.H. Wang, Z.P. Hao, Characterization photocatalytic oxidation of refractory VOCs mixtures: through the net of process
and assessment of volatile organic compounds (VOCs) emissions from typical limitations, Catal Today. 280 (2017) 93–98, https://doi.org/10.1016/j.
industries, Chin. Sci. Bull. 58 (2013) 724–730, https://doi.org/10.1007/s11434- cattod.2016.03.041.
012-5345-2. [29] I. Konstantinou, Photocatalytic transformation of pesticides in aqueous titanium
[4] J.K. Pearson, European solvent VOC emission inventories based on industry-wide dioxide suspensions using artificial and solar light: intermediates and degradation
information, Atmos Environ. 204 (2019) 118–124, https://doi.org/10.1016/j. pathways, Appl Catal b. 42 (2003) 319–335, https://doi.org/10.1016/S0926-3373
atmosenv.2019.02.014. (02)00266-7.
[5] European Parliament and of the Council of the European Union, Directive (EU) [30] V. Iliev, D. Tomova, L. Bilyarska, A. Eliyas, L. Petrov, Photocatalytic properties of
2016/2284 of the European Parliament and of the Council of 14 December 2016 on TiO2 modified with platinum and silver nanoparticles in the degradation of oxalic
the reduction of national emissions of certain atmospheric pollutants, amending acid in aqueous solution, Appl Catal b. 63 (2006) 266–271, https://doi.org/
Directive 2003/35/EC and repealing Directive 2001/81/EC, 2016. 10.1016/j.apcatb.2005.10.014.
[6] E.B. Republic of Estonia, https://kotkas.envir.ee/annual_reports_registry, [31] S. Alofi, C. O’Rourke, A. Mills, Kinetics of stearic acid destruction on TiO2 ‘self-
Keskkonnaotsuste infosüsteem, (2022). cleaning’ films revisited, Photochem. Photobiol. Sci. 21 (2022) 2061–2069,
[7] Joint Research Centre Directorate B – Growth and Innovation Circular Economy https://doi.org/10.1007/s43630-022-00278-0.
and Industrial Leadership Unit European IPPC Bureau, Best Available Techniques [32] J. Bolobajev, M. Kask, K. Kreek, M. Kulp, M. Koel, A. Goi, Metal-doped organic
(BAT) Reference Document for Common Waste Gas Management and Treatment aerogels for photocatalytic degradation of trimethoprim, Chem. Eng. J. 357 (2019)
Systems in the Chemical Sector, Brussels, 2019. 120–128, https://doi.org/10.1016/j.cej.2018.09.127.
[8] W. Zou, B. Gao, Y.S. Ok, L. Dong, Integrated adsorption and photocatalytic [33] J. Alvarez-Ramirez, R. Femat, M. Meraz, C. Ibarra-Valdez, Some remarks on the
degradation of volatile organic compounds (VOCs) using carbon-based Langmuir-Hinshelwood kinetics, J Math Chem. 54 (2016) 375–392, https://doi.
nanocomposites: a critical review, Chemosphere 218 (2019) 845–859, https://doi. org/10.1007/s10910-015-0566-7.
org/10.1016/j.chemosphere.2018.11.175. [34] P.F. Biard, A. Bouzaza, D. Wolbert, Photocatalytic degradation of two volatile fatty
[9] S. Almaie, V. Vatanpour, M.H. Rasoulifard, I. Koyuncu, Volatile organic acids in an annular plug-flow reactor; kinetic modeling and contribution of mass
compounds (VOCs) removal by photocatalysts: a review, Chemosphere 306 (2022), transfer rate, Environ Sci Technol. 41 (2007) 2908–2914, https://doi.org/
https://doi.org/10.1016/j.chemosphere.2022.135655. 10.1021/es062368n.
[10] Y. Zhang, Y. Wang, R. Xie, H. Huang, M.K.H. Leung, J. Li, D.Y.C. Leung, [35] B. Boulinguiez, A. Bouzaza, S. Merabet, D. Wolbert, Photocatalytic degradation of
Photocatalytic oxidation for volatile organic compounds elimination: from ammonia and butyric acid in plug-flow reactor: degradation kinetic modeling with
fundamental research to practical applications, Environ Sci Technol. 56 (2022) contribution of mass transfer, J Photochem Photobiol A Chem. 200 (2008)
16582–16601, https://doi.org/10.1021/acs.est.2c05444. 254–261, https://doi.org/10.1016/j.jphotochem.2008.08.005.
[11] Y. Huang, S.S.H. Ho, R. Niu, L. Xu, Y. Lu, J. Cao, S. Lee, Removal of indoor volatile [36] M. Rivas, G.M. Calaf, D. Laroze, E. Rojas, J. Mendez, J. Honeyman, M.C. Araya,
organic compounds via photocatalytic oxidation: a short review and prospect, Solar ultraviolet A radiation and nonmelanoma skin cancer in Arica, Chile,
Molecules 21 (2016), https://doi.org/10.3390/molecules21010056. J Photochem Photobiol b. 212 (2020) 112047, https://doi.org/10.1016/j.
[12] C. Du, X. Gong, Y. Lin, Decomposition of volatile organic compounds using corona jphotobiol.2020.112047.
discharge plasma technology, J Air Waste Manage Assoc. 69 (2019) 879–899, [37] E. Pelizzetti, M. Schiavello, Photochemical Conversion and Storage of Solar Energy,
https://doi.org/10.1080/10962247.2019.1582441. Springer Netherlands, Dordrecht, 1991. https://doi.org/10.1007/978-94-011-
[13] I. Kornev, S. Preis, Aqueous benzene oxidation in low-temperature plasma of 3396-8.
pulsed corona discharge, Journal of Advanced Oxidation Technologies. 19 (2016). [38] F.J. Rivas, F.J. Beltrán, A. Encinas, Removal of emergent contaminants: Integration
[14] M. Kask, Combination of Advanced Oxidation Methods for the Energy-Efficient of ozone and photocatalysis, J Environ Manage. 100 (2012) 10–15, https://doi.
Abatement of Aqueous and Gaseous Hazardous Pollutants, Doctoral Thesis, Tallinn org/10.1016/j.jenvman.2012.01.025.
University of Technology, 2021. [39] F.J. Rivas, F.J. Beltrán, O. Gimeno, M. Carbajo, Fluorene oxidation by coupling of
[15] L. Khezami, P. Nguyen-Tri, W.A. Saoud, A. Bouzaza, A. El Jery, D. Duc Nguyen, V. ozone, radiation, and semiconductors: a mathematical approach to the kinetics, Ind
K. Gupta, A.A. Assadi, Recent progress in air treatment with combined Eng Chem Res. 45 (2006) 166–174, https://doi.org/10.1021/ie050781i.
photocatalytic/plasma processes: a review, J Environ Manage. 299 (2021), https:// [40] H. Qi, S. De-Zhi, C. Guo-Qing, Formaldehyde degradation by UV/TiO2/O3 process
doi.org/10.1016/j.jenvman.2021.113588. using continuous flow mode, J. Environ. Sci. 19 (2007) 1136–1140.
[16] L. Onga, E. Kattel-Salusoo, S. Preis, N. Dulova, Degradation of anti-inflammatory [41] K.P. Yu, G.W.M. Lee, Decomposition of gas-phase toluene by the combination of
drug dexamethasone by pulsed corona discharge: the effect of peroxycompounds ozone and photocatalytic oxidation process (TiO2/UV, TiO2/UV/O3, and UV/O3),
addition, J Environ Chem Eng. 10 (2022), https://doi.org/10.1016/j. Appl Catal b. 75 (2007) 29–38, https://doi.org/10.1016/j.apcatb.2007.03.006.
jece.2022.108042. [42] H. Huang, W. Li, Destruction of toluene by ozone-enhanced photocatalysis:
[17] P. Tikker, D. Nikitin, S. Preis, Oxidation of aqueous bisphenols A and S by pulsed Performance and mechanism, Appl Catal b. 102 (2011) 449–453, https://doi.org/
corona discharge: Impacts of process control parameters and oxidation products 10.1016/j.apcatb.2010.12.025.
identification, Chem. Eng. J. 438 (2022), https://doi.org/10.1016/j.
cej.2022.135602.
[18] M. Kask, J. Bolobajev, M. Krichevskaya, Gas-phase photocatalytic degradation of
acetone and toluene, and their mixture in the presence of ozone in continuous
multi-section reactor as possible air post-treatment for exhaust from pulsed corona
discharge, Chem. Eng. J. 399 (2020), https://doi.org/10.1016/j.cej.2020.125815.

8
K. Altof et al. Chemical Engineering Journal 481 (2024) 148488

[43] J. Yuan, X. Huang, M. Chen, J. Shi, W. Shangguan, Ozone-assisted photocatalytic [45] J. Sydorenko, A. Mere, M. Krunks, M. Krichevskaya, I.O. Acik, Transparent TiO2
degradation of gaseous acetaldehyde on TiO2/M-ZSM-5 (M = Zn, Cu, Mn), Catal thin films with high photocatalytic activity for indoor air purification, RSC Adv. 12
Today. 201 (2013) 182–188, https://doi.org/10.1016/j.cattod.2012.06.003. (2022) 35531–35542, https://doi.org/10.1039/d2ra06488j.
[44] R. Atkinson, Kinetics and mechanisms of the gas-phase reactions of the hydroxyl
radical with organic compounds under atmospheric conditions, Chem Rev. 86
(1986) 69–201, https://doi.org/10.1021/cr00071a004.