Results in Engineering 19 (2023) 101348

Contents lists available at ScienceDirect

Results in Engineering
journal homepage: www.sciencedirect.com/journal/results-in-engineering

Solar photocatalytic degradation of carbendazim in water using TiO2

particle- and sol-gel dip-coating filters
Chalita Mungsuk a, Suriyakit Yommee a, Sitthisuntorn Supothina b, Paradee Chuaybamroong a, *
a
Department of Environmental Science, Faculty of Science and Technology, Thammasat University, Pathumthani, 12120, Thailand
b
National Metal and Materials Technology Center, National Science and Technology Development Agency, Thailand

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

Keywords: Carbendazim degradation with TiO2 photocatalysis under sunlight irradiation using particle- and sol-gel coating
Carbendazim filters was studied. For particle coating, an appropriate concentration of Degussa P25, namely 0.3, 0.5, and 1%
TiO2 w/v, and binder types, namely Duramax B1000, PEG1000, PEG2000, PEG4000, and PEG6000 were investigated.
Photocatalysis
It was found that 0.5% w/v of P25 and 1 wt% of PEG6000 yielded the highest efficiency of 89–91% from
Sol-gel
Degussa P25
triplicate runs. The best kinetic rate constant was 0.048 min− 1. For sol-gel coating, different anatase to rutile (A/
Binder R) ratio were compared for efficiency. The synergistic effect from photocatalysis and sorption due to the
honeycomb-like coating surface was observed. The efficiency of 98–99.5% in 10 min were obtained from
PEG4000-6000. Zeta potential analysis revealed that the TiO2 from both coatings and carbendazim had negative
charge, which was the same as that of the TiO2, then the sorption was not due to the opposite charge attraction.

1. Introduction low pressure mercury lamp, and >90% efficiency was observed. Rajes­
wari and Kanmani [5] used TiO2 photocatalysis from Degussa P25 slurry
Carbendazim (methyl N-(1H-benzimidazol-2-yl)carbamate: combined with ozonation under a 15-W medium pressure mercury lamp,
C6H9N3O2) is a popular fungicide to control Ascomycetes, Fungi and 97% efficiency was reported. Costa et al. [6] used photo-Fenton
Imperfecti, and Basisiomycetes on various crops such as cereals, cotton, from a 80-W low pressure mercury lamp and solar photo-Fenton, and
mushrooms, bananas, grapes, fruits, ornamentals, soybeans, sugarbeet, >95% efficiencies were obtained from both light sources. Sebastian
peanuts, tobacco, and vegetables. Its residues and metabolites are et al. [7] used carbon nanofibers functionalized with β-cyclodextrin and
strongly bound to the soil organic matter, resulting in the high Koc (i.e., decorated with hematite nanoparticles through a sonochemical method
the ratio of carbendazim concentration per g of soil organic carbon, to to detect and degrade carbendazim via photocatalysis, and 93.5% effi­
carbendazim concentration per mL of water) approximately of 2000. A ciency was pronounced. Kuar et al. [8] used clay beads which were
high value implies that carbendazim strongly binds or sorbs to the soil dip-coated with 2 wt% Degussa P25 and irradiated under UV lamps,
and it is less likely to leach or mobile with water. Thus it remains in soil intensity of 30 W/m2, and found the average efficiency of 77 ± 3.85%.
surface layer for up to 3 years with the half-life of 3–12 months [1]. The In order to improve the efficiency to as high as 93 ± 3.85% under
solubility in water at 24 ◦ C and pH 4 is 29 mg/L and reduces to 8 mg/L at sunlight irradiation, Fe-doping was needed. Marcelino et al. [9] depos­
pH 7 and 1.49 mg/L at pH 8 [2]. At ambient temperature (22–25 ◦ C) and ited TiO2 thin film onto polyethylene terephthalate (PET) sheet by
pH 5–7, carbendazim is hydrolytically stable, while at pH 9 the hydro­ reactive magnetron sputtering to degrade carbendazim in water under a
lysis half-life is 22–124 days [3], indicating persistence in the environ­ 9-W UVA irradiation. With the presence of turmeric (curcumin source
ment. Moreover, it is very toxic to aquatic life with long lasting effects for enhancing photon absorption) in water, the efficiency of 35% and a
and may cause genetic defects and fertility damage in human. In April pseudo first-order kinetic rate of 0.0025 min− 1 were obtained.
2006, the US EPA classified carbendazim as group C: possible human From the literature reviews above, most procedures use mercury
carcinogen [2]. lamps which require unavoidably electricity costs. These are a big
Many efforts have been made for decades to degrade carbendazim in burden for developing countries and rather impractical for the field
water. For example, Mazellier et al. [4] used UV/H2O2 photolysis with treatment. Solar irradiation is more appropriate with no cost for

* Corresponding author.
E-mail address: paradee@tu.ac.th (P. Chuaybamroong).

https://doi.org/10.1016/j.rineng.2023.101348
Received 15 June 2023; Received in revised form 29 July 2023; Accepted 3 August 2023
Available online 4 August 2023
2590-1230/© 2023 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
C. Mungsuk et al. Results in Engineering 19 (2023) 101348

Fig. 1. A schematic diagram of the experiment.

countries that have plenty of sunlight all year. Moreover, advanced and B1000 was purchased from Dow Chemical. Ethanol (99.9%) was pur­
sophisticated treatment procedures are yet unaffordable. This work, chased from RCI Labscan. Nitric acid (65%) was purchased from Carlo
therefore, employs a simple but efficient TiO2 photocatalysis from par­ Erba. AgNO3 (99.9%) was purchased from Fisher. Commercial carben­
ticle- and sol-gel coating filters to degrade carbendazim residue in water. dazim was carbendazim50 (SCI), having 50% of methyl benzimidazol-2-
One of the most popular TiO2 catalyst is Degussa P25 powder, which ylcarbamate. Standard carbendazim (98.6%) was purchased from Dr.
comprises 73–85% anatase, 14–17% rutile, and 0–13% amorphous [10]. Ehrenstorfer. Sodium dodecyl sulfate was purchased from Loba Chemie.
However, it is very difficult to separate such small powder, 30 nm [11],
from the treated water. Then colloidal coating onto substrates is more 2.2. Coating procedures
practical, but a suitable binder is necessary. This study investigated 5
different binders: polyethylene glycol (PEG), MW of 1000, 2000, 4000, A piece of 4 × 4 cm glass-fiber pre-filters were pre-treated with 0.08
6000 and Duramax B1000 acrylic latex emulsion for Degussa P25 par­ M sodium dodecyl sulfate to remove water-resistant wax. For TiO2
ticle coating onto the inexpensive glass-fiber pre-filter. Different con­ particle coating, each PEG was mixed with deionized water using
centrations of TiO2 particle and binder were also studied. The PEG magnetic stirrer, in a concentration of 0.3 wt%, for 30 min. Then
(H-(O–CH2–CH2)n-OH) is selected because it is well known for Degussa P25 (will be referred to as “P25” from now on), namely 0.3%,
exothermic adsorption onto a TiO2 oligomer via hydrogen bond [12]. Its 0.5% or 1% w/v was slowly added and stirred for another 60 min. A
long chain can stabilize TiO2 particles with steric hindrance, resulting in filter was then dipped into the suspension for 1 min, dried, and heated in
homogeneous coating. Meanwhile, Duramax B1000 can absorb onto a hot air oven at 100 ◦ C for 1 h, and repeated for a total of 10 rounds.
TiO2 powder surface and performs bridging between TiO2 particles, Finally, the filter was heated at 420 ◦ C in the furnace for 1 h to remove
providing inter-particle flocculation and binding action [13]. any residual organic compounds in the binder.
Sol-gel coating is another method that gains popularity from its For sol-gel coating, 26-mL TTiP was mixed with 390-mL ethanol,
simplicity and efficiency. The PEG in this method plays a role in pore- using a magnetic stirrer for 30 min. A 40-mL nitric acid was added
forming reagent. The porous structure of TiO2 film depends on its con­ dropwise, followed by 7.6-g AgNO3 and stirred for another 30 min. Then
tent and molecular weight [14]. Thus, different PEG molecular weights 3.4 g of polymer was added and stirred for another 12 h at room tem­
of 1000, 2000, 4000, and 6000 were examined in this study. TiO2 has perature. A pretreated filter was dip-coated in the same manner as above
many phases; for example, anatase, rutile, brookite, but anatase and for 8–10 rounds and finally calcined at 450–650 ◦ C for 1 h at the heating
rutile are more responsible for photocatalysis. Anatase is more active rate of 10 ◦ C/min (Carbolite, model 301) to create anatase and rutile
than rutile due to the larger surface area and longer lifetime of photo­ structures in different A/R ratio.
induced electron-hole pairs [15]. However, mixing anatase with rutile
allows the excited electrons from the valence band to the conduction
2.3. TiO2 characterization
band transferring to the rutile particles, resulting in reducing
electron-hole recombination in the anatase [16]. Thus, suitable anata­
Surface characteristic of TiO2-coated filters were examined using a
se/rutile (A/R) ratio is another factor that needs to be considered. In this
field emission scanning electron microscope (FE-SEM) (JEOL, model
study, titanium (IV) isopropoxide and PEG1000-6000 were used as
JSM-7800 F). Surface area, pore size, and pore volume were analyzed
sol-gel precursors for glass-fiber pre-filter coating. The coated filters
with Brunauer–Emmett–Teller (BET) analysis with N2 sorption (Micro­
were calcined at 450–650 ◦ C in order to find a suitable A/R ratio for
meritics, models Smart VacPrep and 3Flex) using a degassing condition
carbendazim degradation. Carbendazim sorption onto the coated filters
of 110 ◦ C for 8 h. X-ray diffraction patterns of each A/R ratio were ob­
in the dark condition was also studied. Our works should be useful to
tained from an x-ray diffractometer (Bruker, model D2 Phaser). Peak
anyone who seeks an easy, inexpensive, and sound method to solve the
positioned at 2θ = 25.5◦ was assigned to d (101) of anatase, while that at
environmental problem due to the pesticide residues.
27.5◦ was assigned to d (110) of rutile. The A/R ratio were calculated
using equations of Spurr and Myers, while the crystallite size was from
2. Materials and methods
Scherrer equation. Zeta potential was conducted with Particle Size
Analyzer, a dynamic light scattering method (Malvern Instruments Ltd.,
2.1. Chemicals
model ZetaSizer Nano ZS).

Degussa P25 was purchased from Evonik Industries. Titanium (IV)

2.4. Photocatalytic degradation and analysis procedures
isopropoxide (TTiP) (97%) and polyethylene glycol (PEG) with MW of
1000, 2000, 4000, 6000 were purchased from Sigma-Aldrich. Duramax
Commercial carbendazim was dissolved in deionized water in a

2
C. Mungsuk et al. Results in Engineering 19 (2023) 101348

Fig. 2. Kinetic studies from TiO2 particle-coating filters (0.3/0.5/1% P25 + 0.3% binders)Note: 1st run: UV 3.5–4.2 mW/cm2, Temp. 31–40 ◦ C; 2nd run: UV 3.2–3.8
mW/cm2, Temp. 34–39 ◦ C; 3rd run: UV 3.9–4.5 mW/cm2, Temp. 31–35 ◦ C.

the remaining carbendazim concentration (C). UV intensities and water

Table 1
temperature were monitored using a radiometer (Analytic Jena, model
Kinetic rate constants from particle-coating filters (0.3/0.5/1% P25 +
UVP UVX), and a thermometer, respectively.
0.3% binders).

2.5. Carbendazim sorption procedure

The experiments were conducted in the same manner as those in the

photocatalytic degradation except they were in the laboratory with the
white-light fluorescent irradiation (UVA, UVB, UVC intensities = 0 mW/
cm2).
A schematic diagram of the experiment procedure is shown in Fig. 1.

3. Results and discussion

3.1. TiO2 particle coating

The study started with fixing the concentration of all binders at 0.3
wt%, but varying P25 concentrations from 0.3% to 0.5% and 1% w/v.
The particle-coating filters were compared for kinetic study in triplicate
runs and the results are shown in Fig. 2. Their symbols are in the order
from high to low rates for easy consideration and their rate values are in
Table 1. TiO2 loading in all beakers was 0.5 g/L. Total UV intensities
(UVA + UVB + UVC) in the 1st, 2nd, and 3rd run could be averaged as
3.9, 3.6, and 4.2 mW/cm2, respectively. Although the highest UV in­
concentration of 10 mg/L (equivalent to about 5 mg/L of standard
tensity was observed in the 3rd run, the reaction rates were lower than
carbendazim due to its 50% purity). About 500 mL of carbendazim so­
the others. Water temperature seems to affect the reaction rates since the
lution was allocated to each beaker. Water sample was withdrawn from
average temperatures from those 3 runs were 37.5, 36.7, and 33.8 ◦ C,
each one and filtered through a 0.22-μm syringe filter to analyze for the
respectively and the reaction rate constants were in the order concurrent
initial concentration (C0) using a UV–Vis spectrophotometer by scan­
with the water temperature. This agrees well with the Stokes-Einstein
ning from 190 nm to 390 nm. Standard carbendazim peak at 285 nm was
equation [17] that the diffusion coefficient (D) of colloidal spherical
used for a calibration curve. One piece of coated filter was immersed
particles is directly proportional to the temperature (D = kBT/3πηd),
into the solution and placed under sunlight irradiation for a duration of
where kB is Boltzmann’s constant, T is temperature, η is media viscosity,
3 h. Every 10 min in the first 50 min (for kinetic study) and every 30 min
and d is particle diameter. Thus higher water temperature enhances
after that, the water sample was withdrawn and filtered to analyze for
better diffusion for free radicals toward the carbendazim, resulting in

3
C. Mungsuk et al. Results in Engineering 19 (2023) 101348

Fig. 3. SEM images of particle-coating filters (0.3/0.5/1% P25 + 0.3% binders).

better chemical reaction rates in the heterogeneous treatment. Similar P25 + 0.3% B1000 shows the highest rate in the 1st run too. Among
results are found from the work of Mehrjouei et al. [18]. They reported 0.3–1% w/v of P25, the highest rate constant is mostly found from 0.5%
the reaction rates for oxalic acid degradation from TiO2 photocatalysis w/v. This is probably due to the coating surface. The 0.5% w/v of P25
plus ozonation increased from 0.046 mM/min at 10 ◦ C to 0.138 seems to cover the fiber better than 0.3% w/v, which is rather thin and
mM/min at 40 ◦ C and 0.213 mM/min at 50 ◦ C. Later, they used the same sparse, while 1% w/v of P25 is somewhat compact (Fig. 3). BET analyses
treatment to degrade dichloroacetic acid in water and found that for PEG4000 and PEG6000 were also performed, as shown in Table 2. It
increasing water temperature from 10 ◦ C to 70 ◦ C could increase the was found that increasing P25 concentration did not increase the surface
degradation rate from 0.29 mM/h to 0.63 mM/h as well [19]. area, pore volume, nor pore diameter. Since P25 powder is nonporous,
It can be seen in Table 1 that 0.5% P25 + 0.3% PEG4000 and 0.5% consisting of agglomerates of crystallites, its surface area is considered
P25 + 0.3% PEG6000 provide the best rate constants although 0.5% from the size of the clump and the pore volume is measured between

4
C. Mungsuk et al. Results in Engineering 19 (2023) 101348

Table 2
BET analyses from particle-coating filters (0.3/0.5/1% P25 + 0.3% binders).
Particle-coating filters Surface area, Pore volume, Pore diameter,
m2/g cm3/g nm Table 3
Kinetic rate constants from particle-coating filters (0.5% P25 + 0.3/
0.3% P25 + 0.3% 5.5 0.04 24.7
PEG4000
0.5/1% binders).
0.5% P25 + 0.3% 3.2 0.02 25.6
PEG4000
1% P25 + 0.3% 5.6 0.04 25.2
PEG4000
0.3% P25 + 0.3% 6.7 0.04 19.5
PEG6000
0.5% P25 + 0.3% 4.9 0.04 30.5
PEG6000
1% P25 + 0.3% 5.1 0.03 19.8
PEG6000

these agglomerates [20]. Because a binder plays a major role in the P25
aggregation; with the same binder amount, similar cluster size should be
obtained. In the other word, with different binder amount, the different Table 4
cluster size and hence different surface area should be yielded and this BET analyses from particle-coating filters (0.5% P25 + + 0.3/0.5/1% binders).
will be discussed later. Particle-coating filters Surface area, Pore volume, Pore diameter,
Since PEG4000 and PEG6000 showed higher rate constants than the m2/g cm3/g nm
others in Table 1, more concentrations of these 2 binders were further 0.5% P25 + 0.3% 3.2 0.02 25.6
studied, namely 0.3, 0.5, and 1 wt%, while P25 was fixed at 0.5% w/v. PEG4000
The results from triplicate runs are shown in Fig. 4 for the reaction time 0.5% P25 + 0.5% 5.4 0.04 30.7
of 3 h, comparing with P25 suspension for the same TiO2 loadings of 0.5 PEG4000
0.5% P25 + 1% 8.1 0.06 28.5
g/L. It should be noted that in the 2nd run, the power supply was failed
PEG4000
after 60 min, then the complete results from the spectrophotometer 0.5% P25 + 0.3% 4.9 0.04 30.5
analysis cannot be shown. Despite this, it can be seen that 1 wt% PEG6000
PEG6000 yielded the highest efficiency rate with the average rate con­ 0.5% P25 + 0.5% 5.6 0.05 30.0
stant of 0.022 min− 1 from 3 runs (Table 3). However, in the 2nd run, the PEG6000
0.5% P25 + 1% 7.2 0.05 29.0
same rate constants were found from 1 wt% PEG4000 and 1 wt% PEG6000
PEG6000. BET analyses from different PEG concentrations are shown in

Fig. 4. Photocatalytic degradation of carbendazim from particle-coating filters (0.5% P25 + 0.3/0.5/1% PEG4000 and 6000)
Note: duration time for the 2nd run was 60 min due to the power supply failure.1st run: UV 3.8–4.9 mW/cm2, Temp. 36–40 ◦ C; 2nd run: UV 3.1–3.7 mW/cm2, Temp.
28–31 ◦ C; 3rd run: UV 3.6–4.1 mW/cm2, Temp. 29–37 ◦ C.

5
C. Mungsuk et al. Results in Engineering 19 (2023) 101348

Table 5 and PEG8000 were the optimum binders for tape casting TiO2 film due
Calcination temperatures and crystallite sizes of anatase and rutile from sol-gel to their homogeneous, non-agglomerated, and porous microstructures.
coating filters. In contrast, they found that PEG1000 yielded very thin film and large
Polymers Temp., ◦ C A/R ratio Crystallite size, nm agglomerates after sintering. This is because the length of the PEG1000
A R Ag
chain is insufficient to provide stable viscous polymeric template, if
compared with the longer chains of PEG4000-8000. Similarly, Karimian
PEG1000 480 100:0 81.4 51.9
–
and Babaluo [23] reported that PEG6000, among PEG2000/6000/10,
580 80:20 19.4 50.5 39.8
581 60:40 22.0 75.7 58.8 000/35,000, was a suitable binder to stabilize colloidal alumina sus­
600 40:60 28.4 77.9 51.9 pension with sufficient strength in the ceramic green bodies. Because its
587 20:80 32.3 62.9 54.3 chain can adsorb on the surface of alumina particles and stabilizes them
PEG2000 575 100:0 22.0 – 47.2 with the steric mechanism, the green bodies can be molded and retained
570 80:20 20.6 81.8 55.3
585 60:40 22.0 81.8 50.0
in a desired shape with no damage during sintering. In the case of
590 40:60 81.4 68.7 51.6 PEG2000, they found that its short chains could not produce significant
605 20:80 28.6 68.7 48.6 steric hindrance between colloidal particles, resulting in instability of
PEG4000 580 100:0 20.1 – 52.9 alumina suspension. Since longer chain provides more bridging and
590 80:20 21.7 60.6 54.2
steric hindrance between TiO2 particles, thus particle aggregation is
587 60:40 22.0 78.7 53.2
592 40:60 24.8 71.7 51.6 prevented and colloidal stability is enhanced [13]. Thus PEG 6000 in
615 20:80 32.8 81.8 54.6 this study is superior to other binders. Similar results were reported by
PEG6000 480 100:0 19.0 – 47.5 Jampawal et al. [24]. They degraded carbaryl in water using solar
530 80:20 21.4 43.8 50.4 photocatalysis from P25-coated filters and found that 1 wt% of PEG6000
570 60:40 20.4 42.8 40.0
580 40:60 24.6 69.3 50.3
was the best binder. With 1% w/v of P25 + 1 wt% of PEG6000, their rate
610 20:80 24.7 81.8 51.9 constants were 0.022–0.025 min− 1 which were lower than this study,
depending on the solar intensity and water temperature on each day.
Note: A = anatase, R = rutile.
It should be noted that the rate constants in Table 3 are lower than
those in Table 1. It is probably due to the lower water temperatures. The
Table 4. This time, it is obvious that increasing the binder concentrations average total UV in Table 3 are 4.5, 3.5, and 3.7 mW/cm2, respectively,
results in the increases of surface area and pore volume for both which are quite close to 3.9, 3.6, and 4.2 mW/cm2 in Table 1. However,
PEG4000 and PEG6000. Although 1 wt% PEG4000 yields higher surface water temperature in Table 3 are 37.6, 29.5, and 29.9 ◦ C, respectively,
area and pore volume than those of PEG6000, the higher efficiency rates which are lower than those in Table 1 (37.5, 36.7, and 33.8 ◦ C). As
were still found from PEG6000. discussed before that higher temperature enhances the photocatalytic
PEG is a polymer consisting of repeating ethyleneoxy units with a reactions by more Brownian motion or diffusion rate, the higher rate
structure of H-(O–CH2–CH2)n-OH [21]. Higher molecular weight means constant is thus observed. Another point to mention is that the P25
longer polymeric chain. Tasić et al. [22] found that PEG4000, PEG6000, suspension yielded superior rate constant to the particle coated filters.

Fig. 5. Photocatalytic degradation of carbendazim from sol-gel coating filters with different A/R ratio.
Note: PEG1000: UV 2.6–6.8 mW/cm2, Temp. 32–39.5 ◦ C; PEG2000: UV 2.2–6.9 mW/cm2, Temp. 29–36 ◦ C; PEG4000: UV 4.3–5.9 mW/cm2, Temp. 32–35 ◦ C;
PEG6000: UV 3.3–4.2 mW/cm2, Temp. 28–36 ◦ C.

6
C. Mungsuk et al. Results in Engineering 19 (2023) 101348

Table 6 3.2. TiO2 sol-gel coating

BET analyses from sol-gel coating filters.
Polymers A/R Surface area, Pore volume, Pore diameter, Calcination temperatures and crystallite sizes of anatase and rutile
ratio m2/g cm3/g nm from the A/R ratio of 100:0, 80:20, 60:40, 40:60, and 20:80 are shown in
PEG2000 100:0 1.7 0.01 26.7 Table 5. It can be seen that increasing the temperature promotes the
80:20 3.1 0.02 27.9 proportion of rutile and crystallite sizes due to the shrinkage and coa­
60:40 2.3 0.02 24.6 lescence of the grains although some order may switch. The uncon­
40:60 2.9 0.02 25.7 trolled film thickness might affect the calcination temperature for
20:80 3.1 0.02 25.3
PEG6000 100:0 3.3 0.02 21.6
anatase-to-rutile transformation and the crystallite size. Simonenko
80:20 2.7 0.02 24.9 et al. [26] reported that with 1 layer of sol-gel dip-coating titanium
60:40 2.3 0.02 36.5 butoxide onto silica substrate and calcined at 500 ◦ C, a 9-nm crystallite
40:60 2.4 0.02 34.2 size of anatase was obtained. However, with 2 layers for the same
20:80 2.5 0.02 29.9
condition, a crystallite size increased to 21 nm. Moreover, the anatase
Note: A = anatase, R = rutile. crystallite sizes were different between 1- and 2-layer coatings when the
temperature rose to 800 ◦ C. With 1 layer the size was 20 nm, while with
This is because immobilizing P25 onto a glass filter would reduce its 2 layers the size was 53 nm. In this study, TiO2 loading on each filter was
surface area and the mass transfer of surrounding carbendazim mole­ controlled at approximately 1 g/piece, then 8–10 layers deposition were
cules, compared to the colloidal particles. However, the water turbidity conducted. Consequently, the temperatures and crystallite sizes in
from the P25 suspension was very high, 2426 ± 50 NTU, while that from Table 5 were not in the consecutive order since the number of layer was
the particle coated filter was only 20 ± 5 NTU. Hence, it is impractical to not strictly fixed.
use P25 suspension in the real environment. Regarding carbendazim degradation efficiency, the photocatalysis
In terms of the rate constant, this study obtained higher rate values from sol-gel coating filters using different PEG molecular weights and A/
than the work of Rajeswari and Kanmani [5] who simultaneously used 1 R ratio are shown in Fig. 5. It is obvious that all PEG molecular weights
g/L of P25 suspension under a 15-W mercury lamp (254 nm) and 0.48 with A/R ratio of 100:0 yielded the best efficiency, equivalent to P25
g/h ozone dose to treat 40 mg/L of carbendazim. They reported the suspension, and even higher in the cases of PEG4000 and PEG6000. TiO2
highest rate constant of 0.0197 min− 1. Similarly, El Madani et al. [25] loading was fixed at 2 g/L in all beakers. Within 10 min, 99.5% and 98%
used 0.6 g/L of P25 suspension to degrade 85.8 μmol/L imazethapyr efficiencies were achieved from PEG4000 and PEG6000, respectively.
herbicide under a 125-W HPK lamp (365 nm) and reported the rate Since the reduction rate was very fast within the first 10 min, the kinetic
constant of 0.0273 min− 1, which was very close to this study. This im­ rate constant was not performed. However, BET analyses from the sol-
plies that sunlight irradiation is compatible with the mercury lamps, gel coating filters using PEG2000 and PEG6000 in Table 6 show less
then the electricity cost can be eliminated. surface area and pore volumes than those from P25-coating filters in
Tables 2 and 4 This study used alcoholysis (ethanol mixed with nitric
acid) instead of hydrolysis (no water was added) in the sol-gel process.
Through hydrolysis, traditional growth of particles starts from nuclei

Fig. 6. Sorption of carbendazim onto particle-coating filters (0.5% P25) with all binders and PEG6000 only
Note: UV = 0 mW/cm2. Temp = 26–29 ◦ C.

7
C. Mungsuk et al. Results in Engineering 19 (2023) 101348

Fig. 7. Sorption of carbendazim onto sol-gel coating filters with different A/R ratio.

(2–10 nm) and later develops to primary particle (50–100 nm) via porosity of 10–70% to degrade malic acid via photocatalysis and
polycondensation, and then secondary particles (300–1000 nm) occur observed that there was no correlation of the film porosity to the effi­
via aggregation [27]. However, ethanol causes esterification, depoly­ ciency. Too high porosity even reduces the inter-granular transfer and
merization, and solvation of the siloxane polymers, leading to a pore size limits the photocatalytic activity. They also studied the porosity of
reduction and narrowing the pore size distribution [28]. Moreover, high 1-layer versus 3-layer coatings and found that less porosity occurred
amount of alcohol reduces the number concentration of primary particle from 3 layers (38%) comparing to 1 layer (60%) due to the inhibition of
by dilution, then the particle aggregation rate is reduced, resulting in a the TiO2 particle growth from PEG. Thus, the less specific surface area
dense packing of the primary particles within the secondary particles. and pore volume of sol-gel coated filters in this study do not reflect their
Hence, diameter of intraparticle pores and specific surface area are capability of carbendazim degradation.
decreased [27]. Ageing the sol prior to film deposition is another factor
that affects the vol% porosity, pore size and surface area. With no 3.3. Carbendazim sorption onto the coating filters
ageing, Brinker et al. [28] found that the polymers are small which is
leading to small pore opening in the film and less accessibility of N2 as Carbendazim sorption onto the particle-coating filters, with TiO2
indicated by Type II N2-adsorption-condensation isotherms in BET loading of 0.5 g/L, are shown in Fig. 6. In the first 2 runs, all binders
analysis. They reported 0% porosity, <0.4 nm pore diameter, and were tested, while in the 3rd run, only PEG6000 with its different con­
1.2–1.9 m2/g surface area in this case. In contrast, sol particle size in­ centrations was tested. The P25 concentration in all runs was fixed at
creases with ageing time due to particle aggregation and more 0.5% w/v. It can be seen that no reduction of carbendazim occurs in a
condensation, leading to porosity of the coating. Nevertheless, although duration of 180 min, implying that there is no sorption onto the particle
unaged sol-gel was used in this study, PEG was added to and dispersed coated filters. However, different results are seen from the sol-gel
throughout the sol. Its ether oxygens could strongly hydrogen-bond with coating filters in Fig. 7. Significant carbendazim reduction happened
the isolated silanol (Si–OH) groups of the silica particles [29]. During from the coating sol-gel filters. Since the reductions were similarly
calcination, it was decomposed and left the pores behind. Thus, the pore observed from all PEG, only PEG2000 and PEG6000 were shown. With
sizes and surface area of the sol-gel coating onto the glass filters in our the A/R ratio of 100:0, a 98% reduction was found at 90 min in the case
study (Table 4) were higher than the above mentioned values. Wong­ of PEG2000, while a 97% reduction at 60 min was found from PEG6000.
charee et al. [29] also insisted that adding PEG as a pore creator was With the synergistic effect from sorption, the degradation efficiency of
more effective than ageing for controlling the film porosity. PEG6000 in Fig. 5 was as high as 98% after 10 min of reaction. Pore
In terms of the vol% porosity, Guillard et al. [30] found that as far as morphology of the TiO2 coating revealed by SEM images in Figs. 8 and 9
the porosity of the film is sufficient to allow the diffusion of organic should be the main contribution for the observed results.
compounds in the bulk of the film, an inter-granular transfer of e− /h+ is The best efficiencies from particle-coating filters in Fig. 4 were
unnecessary. They used TTiP/PEG sol-gel dip-coating onto glass with PEG6000 and PEG4000 (in the 2nd run). Their SEM images are shown in

Fig. 8. SEM images of particle-coating filters ( × 50,000 magnification).

8
C. Mungsuk et al. Results in Engineering 19 (2023) 101348

Fig. 9. SEM images of sol-gel coating filters (PEG6000) with different A/R ratio ( × 50,000 magnification).

Fig. 8. Although porosities on their coating surfaces can be seen, the surface microgaps were increased with the high molecular weight
cavities seem to be shallow compared to those of sol-gel coating filters in polymer.
Fig. 9. A honeycomb-like surface is observed from all A/R ratio of To further understand carbendazim adsorption mechanism, we also
PEG6000 (Fig. 9), which could enhance their sorption capability. Mo­ analyzed the zeta potential of particle- and sol-gel coating filters from
lecular weight of PEG also affects the film porosity. Wongcharee et al. pH 3–9 as shown in Fig. 10. Zeta potential is a good index for considering
[29] used PEG600, PEG3400, and PEG10000 as pore creators on their particle interactions in terms of dispersion stability. High negative/
sol-gel dip-coating onto glass slides. They found that higher molecular positive zeta potential implies strong repulsive force between colloidal
weight of PEG related to more adsorption onto silica particles via particles, while zero potential charge implies no repulsive force, then
hydrogen bonding; consequently, higher film porosity was achieved colloidal agglomeration can occur. At pH < pHpzc, TiO2 surface is pro­
after calcination. With alcohol-based solvent, solubility of PEG tonated and has a positive charge (TiOH + H+ → TiOH+ 2 ), whereas pH >
decreased as its molecular weight increased, enhancing macroscopic pHpzc, TiO2 surface is hydroxylated and has negative charge (TiOH +
domain formation. Thus, it was observed that numbers and sizes of OH− → TiO− + H2O) [31]. In Fig. 10, pH of carbendazim solution under

9
C. Mungsuk et al. Results in Engineering 19 (2023) 101348

Fig. 10. Zeta potential of the particle- and sol-gel coating filters and pH of carbendazim solution.

sunlight irradiation was 6.7 ± 0.4 and point of zero charge (pzc) of the Research Fund through Contract No. TUGR 2/23/2562. Chalita Mung­
coated filters was observed at pH ~5.2–5.4. Thus TiO2 on both particle- suk is grateful for the scholarship for talent student to study graduate
and sol-gel coating filters had negative charges as TiO− . Carbendazim program in Faculty of Science and Technology Thammasat University
has an acid dissociation constant (pKA) of 4.2–4.3 [2], meaning that at via Contract No. TB. 14/2562.
pH 6.7 it is deprotonated or in a basic form (pH > pKA). Both TiO− and
carbendazim solution had the same charges, then the opposite charge References
attraction would not occur. Therefore, carbendazim reduction in Fig. 7
was mainly from sorption of the coating surface. [1] WHO, Environmental Health Criteria 149, Carbendazim, 1993. https://apps.who.
int/iris/bitstream/handle/10665/37443/9241571497-eng.pdf?sequence=1.
(Accessed 7 May 2023).
4. Conclusion [2] PubChem, Carbendazim. https://pubchem.ncbi.nlm.nih.gov/compound/25429, n.
d (accessed 7 May 2023).
[3] IUPAC (international union of pure and applied chemistry), Carbendazim (Ref: BAS
Carbendazim residue in water was successfully degraded by TiO2 346F) (2022). http://sitem.herts.ac.uk/aeru/iupac/Reports/116.htm. (Accessed 7
photocatalysis under sunlight irradiation. In the case of particle-coating May 2023).
filter, PEG6000 was the most suitable binder. With the best coating [4] P. Mazellier, É. Leroy, J. De Laat, B. Legube, Degradation of carbendazim by UV/
H2O2 investigated by kinetic modelling, Environ. Chem. Lett. 1 (2003) 68–72.
concentration, namely 0.5%P25 + 1%PEG6000, 89–91% degradation [5] R. Rajeswari, S. Kanmani, TiO2-based heterogeneous photocatalytic treatment
efficiency was achieved after 180 min of reaction. The reaction rate combined with ozonation for carbendazim degradation, Iran. J. Environ. Health
constant could vary from 0.02 to 0.05 min− 1, depending on UV intensity Sci. Eng. 6 (2009) 61–66.
[6] E. Costa, S. Bottrel, M. Starling, M. Leão, C. Amorin, Degradation of carbendazim in
and water temperature on each day. Higher amount of PEG yielded
water via photo-Fenton in raceway pond reactor: assessment of acute toxicity and
higher surface area and pore volumes of the coating substance. No transformation products, Environ. Sci. Poll. Res. 26 (2019) 4324–4336.
sorption onto the particle-coating filter could be pronounced. [7] N. Sebastian, W.-C. Yu, D. Balram, F. Al-Mubaddel, M.T. Noman, Functionalization
In the case of sol-gel coating filter, less surface area and pore volume of CNFs surface with β-cyclodextrin and decoration of hematite nanoparticles for
detection and degradation of toxic fungicide carbendazim, Appl. Surf. Sci. 586
than those of the particle-coating were observed. The honeycomb-like (2022), 152666.
surface enhanced the carbendazim sorption onto the filter, thus a [8] T. Kaur, A. Sraw, R.K. Wanchoo, A.P. Toor, Solar assisted degradation of
reduction of 97–98% in 60–90 min was obtained. However, synergy carbendazim in water using clay beads immobilized with TiO2 & Fe doped TiO2,
Sol. Energy 162 (2018) 45–56.
with photocatalysis, the efficiency of 98–99.5% was quickly gained in [9] R.B.P. Marcelino, C.C. Amorim, M. Ratova, B. Delfour-Peyrethon, P. Kelly, Novel
10 min when using PEG4000-PEG6000 as a polymer. Zeta potential and versatile TiO2 thin films on PET for photocatalytic removal of contaminants of
analysis revealed that TiO2 and carbendazim were both negative emerging concern from water, Chem. Eng. J. 370 (2019) 1251–1261.
[10] B. Ohtani, O.O. Prieto-Mahaney, D. Li, R. Abe, What is Degussa (Evonik) P25?
charges, then the opposite charge attraction was less likely to occur. Crystalline composition analysis, reconstruction from isolated pure particles and
photocatalytic activity test, Photochem. Photobiol.: Chem 216 (2010) 179–182.
Author contribution [11] Y. Chen, D.D. Dionysiou, A comparative study on physicochemical properties and
photocatalytic behavior of microporous TiO2-P25 composite films and
macroporous TiO2 films coated on stainless steel substrate, Appl. Catal. A: Gen. 317
Chalita Mungsuk: Investigation. Suriyakit Yommee: Funding (2007) 129–137.
acquisition, Visualization. Sitthisuntorn Supothina: Methodology, [12] X. Liu, Z. Jin, S. Bu, T. Yin, Influences of solvent on properties of TiO2 porous films
prepared by a sol-gel method from the system containing PEG, J. Sol. Gel Sci.
Validation. Paradee Chuaybamroong: Conceptualization, Methodol­
Technol. 36 (2005) 103–111.
ogy, Formal analysis, Project administration, Writing-Review & Editing. [13] K. Sungkajuntranon, P. Sribenjalux, S. Supothina, P. Chuaybamroong, Effect of
binders on airborne microorganism inactivation using TiO2 photocatalytic
fluorescent lamps, J. Photochem. Photobiol. B Biol. 138 (2014) 160–171.
Declaration of competing interest
[14] S. Bu, Z. Jin, X. Liu, H. Du, Z. Cheng, Preparation and formation mechanism of
porous TiO2 films using PEG and alcohol solvent as double-template, J. Sol. Gel Sci.
The authors declare that they have no known competing financial Technol. 30 (2004) 239–248.
interests or personal relationships that could have appeared to influence [15] Y. Kakuma, A.Y. Nosaka, Y. Nosaka, Difference in TiO2 photocatalytic mechanism
between rutile and anatase studied by the detection of active oxygen and surface
the work reported in this paper. species in water, Phy. Chem. Chem. Phys. 17 (2015) 18691–18698.
[16] H. Nakajima, T. Mori, Q. Shen, T. Toyoda, Photoluminescence study of mixtures of
Data availability anatase and rutile TiO2 nanoparticles: influence of charge transfer between the
nanoparticles on their photoluminescence excitation bands, Chem. Phys. Lett. 409
(2005) 81–84.
Data will be made available on request. [17] J. Zmpitas, J. Gross, Modified Stokes-Einstein equation for molecular self-diffusion
based on entropy scaling, Ind. Eng. Chem. Res. 60 (2021) 4453–4459.
[18] M. Mehrjouei, S. Müller, D. Möller, Degradation of oxalic acid in a photocatalytic
Acknowledgements ozonation system by means of Pilkington ActiveTM glass, J. Photochem. Photobiol.,
A: Chem 217 (2011) 417–424.
This study was financially supported by Thammasat University

10
C. Mungsuk et al. Results in Engineering 19 (2023) 101348

[19] M. Mehrjouei, S. Müller, D. Möller, Synergistic effect of the combination of [25] M. El Madani, M. Harir, A. Zrineh, M. El Azzouzi, Photodegradation of imazethapyr
immobilized TiO2, UVA and ozone on the decomposition of dichloroacetic acid, herbicide by using slurry and supported TiO2: efficiency comparison, Arab. J.
J. Environ. Sci. Health A. 47 (2012) 1073–1081. Chem. 8 (2015) 181–185.
[20] M.A. Nawi, S.M. Zain, Enhancing the surface properties of the immobilized [26] N.P. Simonenko, V.A. Nikolaev, E.P. Simonenko, N.B. Generalova, V.
Degussa P-25 TiO2 for the efficient photocatalytic removal of methylene blue from G. Sevastyanov, N.T. Kuznetsov, Preparation of nanostructured titania thin films by
aqueous solution, Appl. Surf. Sci. 258 (2012) 6148–6157. sol-gel technology, Russ. J. Inorg. Chem. 61 (2016) 1505–1511.
[21] Wikipedia, Polyethylene glycol (2023). https://en.wikipedia.org/wiki/Polyeth [27] K.C. Song, S.E. Pratsinis, The effect of alcohol solvents on the porosity and phase
ylene_glycol#. (Accessed 28 May 2023). composition of titania, J. Colloid Interface Sci. 231 (2000) 289–298.
[22] N. Tasić, Z. Branković, Z. Marinković-Stanojević, G. Branković, Effect of binder [28] C.J. Brinker, D.M. Smith, R. Deshpande, P.M. Davis, S. Hietala, G.C. Frye, et al.,
molecular weight on morphology of TiO2 films prepared by tape casting and their Sol-gel processing of controlled pore oxides, Catal. Today 14 (1992) 155–163.
photovoltaic performance, Sci. Sinter. 44 (2012) 365–372. [29] K. Wongcharee, M. Brungs, R. Chaplin, Y.J. Hong, R. Pillar, E. Sizgek, Sol-gel
[23] H. Karimian, A.A. Babaluo, Effect of polymeric binder and dispersant on the processing by aging and pore creator addition for porous silica antireflective
stability of colloidal alumina suspensions, Iran. Polym. J. (Engl. Ed.) 15 (2006) coatings, J. Sol. Gel Sci. Technol. 25 (2002) 215–221.
879–889. [30] C. Guillard, S. Debayle, A. Gagnaire, H. Jaffrezic, J.-M. Herrmann, Physical
[24] J. Jampawal, S. Supothina, P. Chuaybamroong, Solar photocatalytic degradation of properties and photocatalytic efficiencies of TiO2 films prepared by PECVD and sol-
carbaryl in water using TiO2-coated filters with different binders and effect of the gel methods, Mater. Res. Bull. 39 (2004) 1445–1458.
operating conditions, Environ. Sci. Pollut. Res. 29 (2022) 88027–88040. [31] M.G. Alalm, A. Tawfik, S. Ookawara, Comparison of solar TiO2 photocatalysis and
solar photo-Fenton for treatment of pesticides industry wastewater: operational
conditions, kinetics, and costs, J. Water Process Eng. 8 (2015) 55–63.

11