Korean J. Chem. Eng.

, 36(6), 965-974 (2019) pISSN: 0256-1115

DOI: 10.1007/s11814-018-0230-1 eISSN: 1975-7220
INVITED REVIEW PAPER INVITED REVIEW PAPER

Synergistic decomposition of imidacloprid by TiO2-Fe3O4 nanocomposite conjugated

with persulfate in a photovoltaic-powered UV-LED photoreactor
Mohammad Reza Eskandarian*,**,***, Mohammad Hossein Rasoulifard**,†,‡,
Mostafa Fazli*, Leila Ghalamchi**, and Hyeok Choi***,†,‡

*Department of Applied Chemistry, Faculty of Chemistry, Semnan University, Semnan 363-35196, Iran
**Water & Wastewater Treatment Research Laboratory, Department of Chemistry,
University of Zanjan, Zanjan 313-45195, Iran
***Department of Civil Engineering, The University of Texas at Arlington, Arlington, TX 76019-0308, United States
(Received 9 October 2018 • accepted 17 January 2019)

AbstractTo facilitate decomposition of imidacloprid (IMD), as a persistent probe insecticide, TiO2-Fe3O4 (TF) nano-
composite was synthesized and characterized. TF particles in size of 50-60 nm with band-gap of 2.8 eV were immobi-
lized onto glass tubes and utilized as a photocatalyst irradiated with ultraviolet-light emitting diode (UV-LED) powered
by photovoltaics. Synergistic decomposition of IMD in the photocatalytic reactor injected with persulfate (PS) was
investigated. Along with various control and reference tests, parametric studies to evaluate the effects of PS concentra-
tion, IMD concentration, and circulation rate on IMD decomposition kinetics and electrical energy consumption were
performed. The contribution of various physical and chemical mechanisms to IMD removal was discussed, including
self-decomposition, direct photolytic decomposition, chemical oxidation by PS, photolysis of PS to produce sulfate radi-
cals, Fenton-like reaction to produce sulfate radicals, photocatalysis to generate hydroxyl radicals, and adsorption onto
catalysts. TF conjugated with PS under UV-LED synergistically decomposed IMD. Additionally, results demonstrated
the synergy index factor of 75%, 65%, and 60% for IMD degradation by UV-LED/TF/PS, UV-LED/Fe3O4/PS, and UV-
LED/TiO2/PS routes, respectively. Outcomes also showed that utilizing TF can greatly reduce electrical energy con-
sumption. Since all devices used in this study, including UV-LED, were powered solely by a photovoltaic module, the
immobilized TF photoreactor was proposed as a sustainable, self-powered, energy-saving, and practical point-of-use
decontamination system to remove organic contaminants in water under solar radiation.
Keywords: Imidacloprid, Persulfate, Photocatalysis, Photovoltaics, TiO2-Fe3O4, UV-LED

INTRODUCTION water splitting, hydrogen evolution, and other environmental reme-

diation approaches [18,19]. AOPs also include UV-based chemical
Around 1-2.5 million tons of active pesticide ingredients are used oxidation processes such as O3/UV, H2O2/UV, and O3/H2O2/UV,
annually, mainly in agriculture [1-6]. Based on the report by the and their combination with TiO2 photocatalysis such as H2O2/UV/
United States Environmental Protection Agency, imidacloprid (IMD) TiO2 [20-27].
is considered a highly toxic insecticide [7]. IMD is made to mimic In those AOPs, O3 and H2O2 have been commonly used as oxi-
nicotine and it acts as an insect neurotoxin. Leaching of IMD often dants to produce mostly robust hydroxyl radicals (E0=2.8 eV) [28-
causes groundwater contamination because of its high solubility 32]. In recent years, there has been also huge research effort to
and mobility. Among many treatment processes to decompose such introduce persulfate (PS) to substitute O3 and H2O2 in the AOPs
persistent and non-biodegradable contaminants in water, advanced [31,32]. PS is a more cheap, stable, and safe oxidant than others,
oxidation processes (AOPs) have recently gained significant atten- and is decomposed under UV radiation to produce sulfate radicals
tion [8-10]. Highly oxidizing transient radical species such as hy- (E0=2.6 eV) [33,34]. Sulfate radicals are as strong as hydroxyl radi-
droxyl radicals are generated in AOPs and readily and non-selec- cals and they are relatively more selective than hydroxyl radicals.
tively attack organic chemicals [10-13]. As one of the most promis- Sulfate radicals might work in a wide range of pH conditions. They
ing AOPs, photocatalysis using titanium dioxide (TiO2) is significant are known to be more reactive for some organic chemical groups
due to its effectiveness in generating radicals under ultraviolet (UV) than hydroxyl radicals [33,34].
radiation and many other advantages such as low cost and chemi- Consequently, the combination of PS/UV system with TiO2 pho-
cal stability [14-17]. In addition, many investigations have been con- tocatalysis might have high potential to synergistically decompose
ducted with TiO2 as photocatalyst because of its unique features for persistent organic contaminants such as IMD. In this present study
for decomposition of IMD by PS/UV/TiO2, two important techni-
†
To whom correspondence should be addressed. cal aspects were addressed. First, TiO2-Fe3O4 (TF) nanocomposite
E-mail: m_h_rasoulifard@[Link], hchoi@[Link] was examined instead of pure TiO2 [35-38]. Due to the magnetic
‡
The two authors contributed equally to this work. property of Fe3O4, TF composite can be easily removed by apply-
Copyright by The Korean Institute of Chemical Engineers. ing a magnetic field after its applications for water treatment [39-

965
966 M. R. Eskandarian et al.

43]. According to the energy level theory, the presence of Fe3O4 can solution was continuously sonicated at 60 oC for 30 min. Then,
also enhance the photocatalytic activity of TiO2 [44-46]. Eventually, only magnetic Fe3O4 particles were separated by applying a mag-
TF composite will be immobilized onto glass tubes to build a more netic field with 1.2 T field strength supermagnet (Kanetec, Japan)
sustainable and practical photocatalytic reactor composed of PS/ [39-41].
UV/TF. To prepare TF nanocomposite particles, i.e., TiO2-Fe3O4, equal
Second, light emitting diodes (LEDs) instead of mercury-based amounts of TiO2 nanoparticles and Fe3O4 nanoparticles were added
UV lamps were used to activate PS and TF, so-called UV-LEDs into C2H5OH to achieve total solid concentration of 33 g/L and
[47-50]. Recently, UV-LEDs have been widely introduced as greener the suspension was sonicated to segregate particles at 100 W and
light sources in various photochemical and photocatalytic processes 60 oC for 40 min. Before particle immobilization, the outer surface
for degradation of water contaminants. UV-LEDs show many advan- of glass tubes with length of 150 mm and diameter of 15 mm was
tages such as better energy efficiency, longer lifetime, no warm-up scratched with carbide tungsten grinding stone (Shimi Azma, Iran)
time, and no mercury disposal problem [50,51]. In particular, UV- and washed with diluted HF followed by NaOH solutions. TF sus-
LEDs exhibit higher quantum yields [51,52]. In this study, UV- pension was carefully poured and uniformly descended down onto
LEDs were also be powered by photovoltaics (PV). Considering glass tubes. After drying to remove C2H5OH at room temperature,
that most of the developing countries have limited energy sources TF-immobilized tubes were calcined in a furnace at 480 oC for
and demand point-of-use applications, such a PV-based photochemi- 3 hr to enhance mechanical adherence strength. Then, they were
cal and photocatalytic process might be essential for in-situ small- finally washed with water several times to remove un-integrated
scale water treatment and disinfection [53-63]. TiO2 and Fe3O4 particles. This cycle was repeated three times to
As a result, the overall objective of this study was to evaluate the ensure enough TF particles were immobilized onto glass tubes. As
synergistic decomposition of IMD by immobilized TF nanocom- references, only TiO2 particles and only Fe3O4 particles were also
posite conjugated with PS in a PV-powered UV-LED photoreactor. immobilized onto glass tubes by following the exact same proce-
Physical and chemical mechanisms for the removal of IMD were dure above for TF.
elucidated, including self-decomposition, direct photolytic decom- 3. Characterization of TiO2/Fe3O4/TF Catalysts
position, chemical oxidation by PS, photolysis of PS to produce The amounts of the catalysts loaded onto glass tubes were deter-
sulfate radicals, Fenton-like reaction to produce sulfate radicals, mined by monitoring weight change before and after coating. Visual
photocatalysis to generate hydroxyl radicals, and adsorption onto morphology of the catalysts was investigated with field emission
catalysts. Along with various control and reference tests, paramet- scanning electron microscopy (SEM/TESCAN MIRA3-FEG, United
ric studies to evaluate the effects of PS concentration, IMD con- States). Surface functional groups were analyzed with Fourier trans-
centration, and circulation rate on IMD decomposition kinetics form infrared spectroscopy (FT-IR, iS 10Thermo Nicolet, England).
and electrical energy consumption were performed. The TF pho- X-ray diffraction (XRD) pattern was collected with XRD instru-
toreactor is proposed as a sustainable, self-powered, energy-saving, ment (XRD-D8 ADVANCED-BRUCHERS AXS, Germany) to
and practical point-of-use decontamination system to remove organic investigate the crystallographic properties. Energy dispersive X-ray
contaminants in water under solar radiation. (EDX) spectroscopy connected to SEM was used to conduct ele-
mental analysis and eventually elemental mapping. To find the
EXPERIMENTAL synergistic effect of co-catalyst Fe3O4 on the photocatalytic activity
of TiO2, the band-gaps of TF and TiO2 were measured by UV-visi-
1. Chemicals and Materials ble diffuse reflectance spectra (DRS) analysis (Cary 500, Varian Co.,
IMD was purchased from Aria Shimi Co. (Iran) and used as Agilent Technologies, Australia) [39-44].
received. The chemical properties of IMD are reported in Table S1. 4. UV-LED Reactor Setup and IMD Decomposition
Potassium persulfate (K2S2O8), sodium hydroxide (NaOH), and The UV-LED photoreactor used in this study is shown in Fig. 1
hydrofluoric acid (HF) were purchased from Merck (Germany). (also note Fig. S1). The reactor had two chambers with internal circu-
Ethanol (C2H5OH) was purchased from Zakariya Razi Co. (Iran). lation; one with UV-LED (the first chamber in the figure) and the
Double distilled water was utilized for all experiments (GFL, type other with UV-LED and catalysts (the second chamber). The first
2008, Germany). TiO2 nanoparticles (Degussa P-25) with surface chamber with diameter of 15 cm and height of 28 cm had effec-
area of 55 m2/g and in size of 30 nm were supplied by Fluka (Ger- tive volume of 5,000 mL while the second chamber with diameter
many). Ferrous chloride (FeCl2·6H2O) and ferric chloride (FeCl3· of 15 cm and height of 14 cm had effective volume of 2,500 mL.
6H2O) (Merck, Germany) were used for synthesis of Fe3O4 nanopar- The chambers had double-walls. Six glass tubes loaded with total
ticles. Glass tubes used as a substrate for catalyst immobilization around 1,875 mg of catalysts (TiO2, Fe3O4, or T-F) were installed
were obtained from a local glass company (Zanjan, Iran). between the walls of only the second chamber. Inside of both cham-
2. Synthesis of Magnetic TiO2-Fe3O4 Nanocomposite Immo- bers, 66 UV-LED lamps with 1 W power each (Optodevice, South
bilized onto Glass Tubes Korea) irradiated UVA at max=390 nm measured by ultrafast fiber
To synthesize Fe3O4 nanoparticles, 50 mL of 1 mol/L FeCl3·6H2O optic spectrometer (AvaSpec-128, Avantes, Netherlands). PV mod-
and 50 mL of 2 mol/L FeCl2·6H2O were poured into 100 ml beaker ule with power range of 225-250 W and efficiency range of 13.8-
and sonicated in a bath sonicator at 100 W for 20 min. Then, 25 15.3% (TN-P225-250, Energy Technology, China) was used to har-
mL of 1 mol/L NaOH was added into Fe solution, forming black vest electrical power from solar radiation and thus to operate all
iron oxide precipitates. To complete the precipitation process, the devices used in this study, including UV-LEDs and pumps. IMD

June, 2019
 Synergistic decomposition of IMD by TF nanocomposite conjugated with PS in a photovoltaic-powered UV-LED photoreactor 967

Fig. 1. Photovoltaic-powered UV-LED photoreactor with two chambers: first one with UV-LED and second one with UV-LED and catalysts.
The reactor setupwas considered as batch with internal circulation. The reactor setup made it possible to conduct various types of
photolytic, photocatalytic, oxidative, and enzymatic experiments in the reactor not only for this study but also for many other studies.
All devices used in this study were powered by a photovoltaic module.

solution was first provided to the second chamber for photocataly-

sis and then the effluent continuously passed through the first cham-
ber for UV-LED irradiation. Finally, the effluent was recirculated
within the whole system. The reactor setup is more or less complex,
but such a setup makes it possible to conduct various types of photo-
lytic, photocatalytic, oxidative, and enzymatic experiments in the
reactor not only for this study but also for many other studies.
The reactor setup is considered as batch with internal circula-
tion at 100 mL/min. All reactions proceeded for 4 hr in the batch
reactor, and thus overall hydraulic residence time (HRT) was also
4 hr. Considering that the amount of solution in the tubes, pumps,
and auxiliary containers was negligible, the first and second cham-
bers had 160 min and 80 min of HRT, respectively. Standard oper-
ational conditions were 250 mg/L of catalyst, 20 mg/L of IMD, 20
mM of PS, 100 mL/min of circulation rate, 66 W of UV-LEDs, and
25 oC of temperature. For parametric studies, IMD concentration
at 10-40 mg/L, PS concentration at 0-120 mM, and circulation rate
at 100-375 mL/min were varied. Note that changing the circulation
rate did not affect HRT in each chamber. All experiments were per-
formed without controlling pH and using buffer. Initial pH at around
6 decreased slightly over reaction. A UV-Visible spectrophotometer
(DR 5000, Hach, USA) was employed to measure IMD concentra-
tion. UV absorption intensity at max=270 nm was correlated with Fig. 2. FT-IR spectra for (a) TiO2, (b) Fe3O4, and (c) TF composite.
IMD concentration as demonstrated elsewhere [64].

RESULTS AND DISCUSSIONS The results are important to explain the observed reactivity of var-
ious systems later. FT-IR analysis of TF is shown in Fig. 2. The ab-
1. Properties of TiO2, Fe3O4, and TF sorption peaks represented the presence of TiO2 and Fe3O4 and shift
The properties of TiO2, Fe3O4, and TF were briefly characterized. of some peaks and formation of new peaks were also observed.

Korean J. Chem. Eng.(Vol. 36, No. 6)

968 M. R. Eskandarian et al.

Fig. 4. Band-gap analysis for TiO2, Fe3O4, and of TF composite.

erogeneous mixture of TiO2 and Fe3O4. Fig. 3 shows XRD pattern

of TF. All peaks well indicated the crystal structures of TiO2 and
Fe3O4 [65-67]. Based on the Scherrer equation, the crystal sizes of
TiO2 and Fe3O4 were at around 30 nm and 23 nm, respectively [67].
Fig. 4 shows DRS analysis result to determine the band-gaps of
TiO2 and TF. TiO2 showed band-gap at around 3.17 eV. Since this
study used P-25, 7 : 3 mixture of anatase (3.2 eV) and rutile (3.0
eV), the band-gap at 3.17 eV was well expected. Also, in the pres-
ent study a 1 : 1 loading ratio for TiO2 : Fe3O4 in TF nanocompos-
ite was considered. Meanwhile, TF showed much reduced band-
gap at 2.78 eV. Adding co-catalyst Fe3O4 to TiO2 was beneficial to
reduction of the band-gap of TiO2 in addition to magnetic removal
of TiO2 [44-46]. The reduced band-gap of TF might lead to better
reactivity in UV-LED photocatalytic systems [40-45].
Fig. 5 shows SEM images of Fe3O4, TiO2, and TF supported onto
Fig. 3. XRD pattern for TiO2, Fe3O4, and of TF composite. glass tubes. Formation of Fe3O4 nanoparticles is clear in Fig. 5(a).
Fe3O4 nanoparticles are uniformly dispersed onto glass substrate.
The similar result was also observed for TiO2 in Fig. 5(b). TiO2 parti-
OH bending vibration band for TF appeared at 3,430 cm1, which cles were slightly larger than Fe3O4, which agrees with the crystal
is close to that for TiO2 typically at 3,440 cm1 but far from that for sizes of the particles (30 nm and 23 nm). TF exhibited aggregated
Fe3O4 at 3,413 cm1, suggesting that Fe3O4 was somehow surrounded mixture of TiO2 and Fe3O4. TF composite was still at nanoscale
by TiO2. The observation was further supported by the disappear- much less than 100 nm. All the results proved uniformly dispersed
ance of a peak at 580 cm1, which corresponds to Fe-O stretching and stably immobilized nanoscale catalysts. EDX result shown in
vibration in Fe3O4 [40]. Fe-O vibration absorption band presum- Fig. S2 supports the presence of Ti and Fe in TF composite. Ti and
ably in Ti-O-Fe bridge structure was observed at 497 cm1, indicat- Fe content was 23.2% and 10.3% (as weight), respectively. Elemen-
ing substitution of Ti in Ti-O-Ti lattice with Fe [40-42]. The result tal mapping of TF composite shown in Fig. S3 also indicates that
indicated formation of new TF nanocomposite, rather than het- Ti and Fe elements were uniformly distributed. Ti was more fre-

Fig. 5. SEM images (a) Fe3O4, (b) TiO2, and (c) TF composite. TF image was obtaniend at higher magnification to distinguish Fe3O4 and TiO2.

June, 2019
 Synergistic decomposition of IMD by TF nanocomposite conjugated with PS in a photovoltaic-powered UV-LED photoreactor 969

quently mapped than Fe, supporting the previous observation that droxyl radicals (R6) were considered to explain the removal of IMD
Fe3O4 was somehow covered by TiO2. in each process. The removal was also partly ascribed to adsorp-
2. Comparative Reactivity of Various Combinational Systems tion of IMD onto catalysts (A1). Note that when IMD decomposi-
Various photochemical and photocatalytic processes of UV-LED, tion occurred in this photoreactor, its concentration kept decreasing
catalysts, and PS alone and in combination were performed as con- in some cases, implying that the reaction did not reach steady-state.
trols and references. As summarized in Table 1, several decompo- 2-1. Single Component
sition mechanisms, including self-decomposition (R1), direct pho- Results on single component are shown in Fig. 6(a). IMD itself
tolytic decomposition (R2), chemical oxidation by PS (R3), pho- (indicated as control) was stable and thus not decomposed, i.e., no
tolysis of PS to produce sulfate radicals (R4), Fenton-like reaction R1 mechanism. There was also negligible photolytic decomposi-
to produce sulfate radicals (R5), and photocatalysis to generate hy- tion of IMD under UV-LED, i.e., no R2 mechanism. The absence

Table 1. Possible removal mechanisms for IMD

Index Mechanism Required conditions
a
R1 Self-decomposition None
R2a Direct photolytic decomposition UV-LED
R3 Chemical oxidation by PS PS
R4 Photolysis of PS to produce sulfate radicals UV-LED and PS
R5b Fenton-like reaction to produce sulfate radicals PS and catalystsc
R6 Photocatalysis to generate hydroxyl radicals UV-LED and catalysts
A1 Adsorption onto catalysts Catalysts
a
R1 and R2 were negligible under given conditions and thus their contributions to IMD removal were not taken into consideration later
b
Fenton-like reaction is likely to occur when catalysts contain small amounts of transition metals such as Fe
c
TiO2, Fe3O4, and TF

Fig. 6. IMD degradation in various control and reference tests: (a) Single control, (b) UV-LED/catalyst, (c) PS/catalyst, and (d) UV-LED/PS/
catalyst (UV-LED=66 W, I=1.40 A, pHinitial~6, T=25 oC, [TF]=250 mg/L, [IMD]=20 mg/L, [PS]=120 mM, and circulation rate=100
mL/min). Please note that different Y axis scales were used in each figure to visualize differences in IMD degradation kinetics. Also, in
order to quickly compare the reactivity of various combinational systems, the highest concentration of PS at 120 mM was used in this
test, not 20 mM which is standard condition.

Korean J. Chem. Eng.(Vol. 36, No. 6)

970 M. R. Eskandarian et al.

of R1 and R2 mechanisms makes it easy to quantify the contribu- Fe ions (Fe2+) dissolved in water from Fe3O4 and TF and those
tions of other removal mechanisms. R1 and R2 mechanisms were available in solid Fe3O4 and TF can activate PS. PS/TiO2 also showed
not taken into consideration later. TiO2, Fe3O4, and TF alone (A1 relatively high reactivity because a trace level of transition metals
mechanism) removed at most around 6%, 7%, and 8%, respec- present in TiO2 as impurities might have triggered a Fenton-like
tively, due to adsorption of IMD. Moreover, Zeta potential analysis reaction. It was reported that even small amounts of transition
was conducted for better understanding of the adsorption mecha- metals, much less than 2 mg/L (as a threshold concentration), are
nism. BET Analysis illustrated values as follows: (Fe3O4=58.98 effective to activate oxidants for the generation of radicals [68-70].
m2g1), (TiO2=50 m2g1) and (TF=62.7 m2g1). 2-4. UV-LED/PS/Catalysts
2-2. UV-LED/Catalysts Removal of IMD by the most complex combinational systems
Photocatalytic decomposition of IMD by TiO2, Fe3O4, and TF is shown in Fig. 6(d). UV-LED/PS (R3 and R4) was very effective.
under UV-LED is shown in Fig. 6(b). All experiments were con- Compared with PS only (R3 shown in Fig. 6(c)), the high reactiv-
ducted based on IMD UV-Vis spectra (Fig. S4). Since R1 and R2 ity of UV-LED/PS could be explained by photolysis of PS to gen-
are negligible, any removal of IMD can be explained by R6 and erate sulfate radicals (SO4• ) for the fast oxidation of IMD, i.e., R4
A1. TiO2/UV-LED and Fe3O4/UV-LED exhibited IMD removal at mechanism. Interestingly, UV-LED/TiO2/PS combining R3, R4,
8% and 10%, respectively. Considering the magnitude of A1 in the R6, and A1 (presumably R5 as well) and UV-LED/Fe3O4/PS com-
previous case, the contribution of photocatalysis to IMD removal, bining R3, R4, R5, and A1 (presumably R6 as well) were slightly
R6, was very low at less than 2-3%. TiO2/UV-LED and Fe3O4/UV- more effective than UV-LED/PS combining R3 and R4. Consider-
LED under the given condition seemed to be ineffective to gener- ing adsorption of IMD to TiO2 and Fe3O4 (A1), contribution of R5
ate hydroxyl radicals for the decomposition of IMD. This is because and R6 to IMD decomposition was not as significant as for persul-
UV-LED used in this study with max=390 nm did not effectively fate. In addition, it is crystal clear that because of the more powerful
activate TiO2 with band-gap of 3.17 eV. In general, most TiO2 pho- oxidative nature of PS and relatively high concentration of that (in
tocatalysts require UV at max=365 nm. Meanwhile, TF/UV-LED this test only), it acts much better than individual catalysts. Actu-
was able to significantly remove IMD via photocatalytic decompo- ally, there is no doubt that individual catalysts did not perform as
sition mechanism, R6. The high reactivity of TF can be ascribed to good as oxidative or photo-oxidative procedure. Meanwhile, UV-
its narrow band-gap at 2.78 eV corresponding to 445 nm. UV-LED LED/TF/PS combining R3, R4, R5, R6 and A1 showed the best
at max=390 was effective enough to activate TF. Other unique prop- reactivity to remove IMD. TF was consistently superior to TiO2 and
erties of TF mentioned previously might have impacted the reac- Fe3O4 in all cases combining it with any photochemical and pho-
tivity [68]. tocatalytic processes. Only UV-LED/TF/PS achieved pseudo-stead
2-3. PS/Catalysts state after 4 hr of reaction.
Reaction between PS and catalysts was evaluated to remove 3. Parametric Studies for UV-LED/TF/PS
IMD via R3, R5, and A1 mechanisms, as shown in Fig. 6(c). IMD Parametric studies to evaluate the effects of PS concentration,
was recalcitrant enough to resist chemical decomposition by PS IMD concentration, and circulation rate on IMD decomposition
alone as a strong oxidant. Only 16% of IMD was decomposed by kinetics were performed for the best combinational system, UV-
PS (R3). There was significant decomposition of IMD when PS LED/TF/PS.
was coupled with TiO2<Fe3O4<TF due to mainly additional R5 The effect of PS concentration on IMD decomposition kinetics
mechanism and partly A1 mechanism. The reason for the increased in the absence of TF (i.e., R3+R4) and presence of TF (i.e., R3+
reactivity of PS/Fe3O4 and PS/TF seems to be obvious. It is known R4+R5+R6+A1) is shown in Fig. 7. It was confirmed that the
that PS can be activated by transition metals to produce strong sul- absence of both PS and TF (R1 and R2 mechanisms only) did not
fate radicals, called Fenton-like reaction (i.e., R5) [69]. In this case, show decomposition of IMD. TF without PS (R6 and A1 mecha-

Fig. 7. Effect of PS concentration at 0-120 mM on IMD decomposition kinetics: (a) In the absence of TF and (b) in the presence of TF (UV-
LED=66 W, I=1.40 A, pHinitial~6, T=25 oC, [TF]=250 mg/L, [IMD]=20 mg/L, and circulation rate=100 mL/min).

June, 2019
 Synergistic decomposition of IMD by TF nanocomposite conjugated with PS in a photovoltaic-powered UV-LED photoreactor 971

Fig. 8. Effect of PS concentration at 0-120 mM on IMD degradation

rate constant, k: (a) In the absence of TF and (b) in the pres-
ence of TF (UV-LED=66 W, I=1.40 A, pHinitial~6, T=25 oC,
Fig. 10. Effect of IMD concentration on IMD degradation rate con-
[TF]=250 mg/L, [IMD]=20 mg/L, and circulation rate=100
stant, k (UV-LED=66W, I=1.40A, pHinitial~6, T=25 oC, [TF]=
mL/min).
250 mg/L, [PS]=20 mM, and circulation rate=100 mL/min).

nisms only) also showed no decomposition of IMD. As PS con-

centration increased, which affected R3, R4 and R5 mechanisms, involved in the observed results. Note that the y-axis is normalized
IMD decomposed faster. Similar pattern was observed regardless concentration of IMD. The total amount of IMD removed over time
of the presence of TF. The presence of TF slightly improved IMD
removal due to the addition of R5, R6, and A1 mechanisms. How-
ever, as observed previously, the contribution of R5 and R6 was
not significant, considering that IMD adsorption onto TF (A1) was
also in place. R3 and R4 mechanisms already significantly contrib-
uted to IMD removal. IMD removal followed first-order kinetics
and thus first-order reaction rate constants were summarized in
Fig. 8. Increase in PS concentration resulted in higher reaction
rates. Impact of PS was more significant in the presence of TF.
The effect of IMD concentration on its decomposition kinetics
is shown in Figs. 9 and 10. IMD decomposition kinetics signifi-
cantly decreased almost linearly over initial IMD concentration.
There was competition for UV photon energy among direct pho-
tolysis of IMD (R2), photolysis of PS (R4), and photocatalysis (R6),
which was influenced by initial IMD concentration. At higher
Fig. 11. Effect of circulation rate at 100-375 mL/min on IMD decom-
IMD concentration, adsorption of IMD to TF (A1) might be more position kinetics (UV-LED=66 W, I=1.40 A, pHinitial~6, T=
significant. Many factors alone and in combination might have been 25 oC, [TF]=250 mg/L, [IMD]=20 mg/L, and [PS]=20 mM).

Fig. 9. Effect of IMD concentration on IMD decomposition kinet- Fig. 12. Effect of circulation rate on IMD degradation rate constant,
ics (UV-LED=66 W, I=1.40 A, pHinitial~6, T=25 oC, [TF]=250 k (UV-LED=66 W, I=1.40 A, pHinitial~6, T=25 oC, [TF]=250
mg/L, [PS]=20 mM, and circulation rate=100 mL/min). mg/L, [IMD]=20 mg/L, and [PS]=20 mM).

Korean J. Chem. Eng.(Vol. 36, No. 6)

972 M. R. Eskandarian et al.

Fig. 13. Effect of PS concentration at 0-120 mM on electrical energy consumption during degradation of IMD (a) in absence of TF and (b) in
the presence of TF (UV-LED=66 W, I=1.40 A, pHinitial~6, T=25 oC, [TF]=250 mg/L, [IMD]=20 mg/L, and circulation rate=100 mL/min).

was more or less similar regardless of initial IMD concentration. 1000  P  t

EE0  ------------------------------------------- (1)
The effect of internal circulation rate in UV-LED photoreactor 60  V  log Ci/Cf
on IMD decomposition kinetics is shown in Figs. 11 and 12. Increas-
The effect of PS concentration is shown in Fig. 13. Apparently,
ing circulation rate from 100 mL/min to 375 mL/min greatly reduced
increase in PS dose significantly decreased electrical energy con-
IMD decomposition. Decomposition rate constants also almost lin-
sumption. The presence of TF significantly reduced electrical energy
early decreased over circulation rate. This result is interesting because
consumption. The observation was more remarkable in the range
the reactor is batch-type. IMD and PS were first introduced to the
of low PS concentrations because even a small amount of PS could
second chamber (2,500 mL) with TF under UV-LED irradiation,
significantly trigger R3, R4 and R5 mechanisms. As a result, use of
continuously moved to the first chamber (5,000 mL) under UV-
TF than TiO2 or Fe3O4 as well as introduction of PS was crucial to
LED irradiation without TF, and then circulated for 4 hr. All chemi-
enhance reaction kinetics and thus to reduce energy consump-
cals stayed in the first reactor for average 160 min and in the sec-
tion. Similarly to PS concentration, the effects of IMD concentra-
ond reactor for average 80 min regardless of circulation, resulting
tion and circulation rate on electrical energy consumption are shown
in HRT or total reaction time of 4 hr. R3, R4, R5, R6 and A1 mecha-
in Figs. 14 and 15, respectively. As expected from the previous results,
nisms occurred in the second chamber, while only R3 and R4
increasing IMD concentration and circulation rate resulted in more
mechanisms occurred in the first chamber. Although circulation
energy consumption.
rate does not affect HRT in each chamber, circulation can speed
5. Reusability Assessment
up movement of chemicals between the chambers, i.e., chamber
Reusability of TF was briefly assessed for six cycles, as shown in
2chamber 1chamber 2chamber 1… If there is no circu-
Fig. 16. The experiment was conducted in UV-LED/TF/PS system
lation, ⅓-rd of IMD molecules will stay only in the second cham-
under standard conditions. Once pseudo-steady state of IMD de-
ber and ⅔-rds of them will stay only in the first chamber. If there
composition for 4 hr was achieved, the reaction solution in the
is high circulation, all IMD molecules will stay in the second cham-
ber shortly, then continuously move to/stay in the first chamber
shortly and recirculate quickly. The difference might have impacted
IMD decomposition kinetics somehow, which should be thor-
oughly investigated in a follow-up study.
4. Electrical Energy Consumption
In addition to IMD decomposition kinetics of the various UV-
LED/PS/TF systems, their electrical energy consumption is an im-
portant factor which impacts their scale-up later and practical appli-
cations under solar radiation. The results on the effects of PS con-
centration, IMD concentration, and circulation rate (shown in Figs.
7, 9, and 11) were reconstructed to determine their effects on elec-
trical energy consumption. Energy consumption was measured using
Eq. (1) proposed by International Union of Pure and Applied Chem-
istry, where EE0 is electrical energy cost or consumption (KWh/
m3) and thus electrical energy efficiency, P is power of irradiation
source (kW), t is time of irradiation (min), V is volume of water Fig. 14. Effect of IMD concentration at 10-40 mg/L on electrical en-
(L), and Ci and Cf are the initial and final concentrations of IMD ergy consumption during degradation of IMD (UV-LED=
(mg/L) [71]. In particular, t/log(Ci/Cf ) in the equation represents 66 W, I=1.40 A, pHinitial~6, T=25 oC, [TF]=250 mg/L, [PS]=20
the inverse of the first-order rate constant of the reaction. mM, and circulation rate=100 mL/min).

June, 2019
 Synergistic decomposition of IMD by TF nanocomposite conjugated with PS in a photovoltaic-powered UV-LED photoreactor 973

ered by PV. Various control and reference tests and parametric

studies enabled us to explain different IMD removal mechanisms
including self-decomposition, direct photolytic decomposition,
chemical oxidation by PS, photolysis of PS to produce sulfate radi-
cals, Fenton-like reaction to produce sulfate radicals, photocataly-
sis to generate hydroxyl radicals, and adsorption onto catalysts.
The reactivity increased in order of UV-LED<TiO2<Fe3O4TF<
UV-LED/Fe3O4UV-LED/TiO2<UV-LED/TF<PS<<TiO2/PS<Fe3O4/
PS<PS/TF<UV-LED/PS<<UV-LED/TiO2/PS<UV-LED/Fe3O4/PS<
UV-LED/TF/PS. TF conjugated with PS under UV-LED was able
to most synergistically remove IMD by combining all the removal
mechanisms. Use of TF than TiO2 or Fe3O4 and introduction of
even a small amount of PS significantly contributed to decreasing
energy consumption. Consequences illustrated that the synergy index
Fig. 15. Effect of circulation rate at 100-375 mL/min on electrical en- factor of 75%, 65%, and 60% for IMD removal was achieved by
ergy consumption during degradation of IMD (UV-LED= UV-LED/TF/PS, UV-LED/Fe3O4/PS, and UV-LED/TiO2/PS routes,
66 W, I=1.40 A, pHinitial~6, T=25 oC, [TF]=250 mg/L, [IMD]= respectively. Since all devices used in this study were powered solely
20 mg/L, and [PS]=20 mM). by a PV module, the photocatalytic reactor is proposed as a sus-
tainable, self-powered, energy-saving, and practical point-of-use
decontamination system to remove organic contaminants in water
under UV-LED powered by solar radiation.

ACKNOWLEDGEMENTS

Dr. Eskandarian thanks the University of Zanjan and Semnan

University for providing financial support for this work. Dr. Choi
also acknowledges partial financial support for this research by the
University of Texas at Arlington.

SUPPORTING INFORMATION

Additional information as noted in the text. This information is

available via the Internet at [Link]
Fig. 16. Reactivity change of TF immobilized onto glass tubes upon journal/11814.
repeated use (6 cycles) with respect to degradation of IMD
(UV-LED=66 W, I=1.40 A, pHinitial~6, T=25 oC, [TF]=250 REFERENCES
mg/L, [IMD]=20 mg/L, [PS]=20 mM, and circulation rate=
100 mL/min). 1. M. A. Shannon, P. W. Bohn, M. Elimelech, J. G. Georgiadis, B. J.
Mariñas and A. M. Mayes, Nature, 452, 301 (2008).
reactor was decanted and TF was dried under ambient condition. 2. M. A. Montgomery and M. Elimelech, Environ. Sci. Technol., 41,
Then, the reactor was recharged with new IMD solution. For six 17 (2007).
cycles, almost the same decomposition pattern was observed. The 3. M. H. Rasoulifard, M. Akrami and M. R. Eskandarian, J. Taiwan
result indicated that TF catalyst was firmly attached to the glass Inst. Chem. Eng., 57, 77 (2015).
tubes and it was not deactivated over the cycles. Furthermore, spent 4. J.-L. Liu and M.-H. Wong, Environ. Int., 59, 208 (2013).
solution after each cycle was analyzed using an atomic absorption 5. R. Loos, G. Locoro, S. Comero, S. Contini, D. Schwesig, F. Werres,
spectrometer (Varian 240AA, Agilent, Australia) to detect any proba- P. Balsaa, O. Gans, S. Weiss and L. Blaha, Water Res., 44, 4115
ble released or un-attached particles from the substrate of the uti- (2010).
lized catalysts. Analysis did not show any important amounts for 6. K. Fenner, S. Canonica, L. P. Wackett and M. Elsner, Science, 341,
toxic heavy metals such as Fe or Ti in final solution. 752 (2013).
7. Label Review Manual for Pesticides, U.S. Environmental Protec-
CONCLUSIONS tion Agency, 2015, [Link]
07.
TF nanocomposite combining photocatalytic TiO2 and mag- 8. E. Mathioulakis, V. Belessiotis and E. Delyannis, Desalination, 203,
netic Fe3O4 was successfully synthesized and immobilized onto 346 (2007).
glass tubes to build a photocatalytic reactor for the removal of IMD. 9. H. Choi, S. R. Al-Abed, D. D. Dionysiou, E. Stathatos and P. Lia-
PS was injected to the reactor under irradiation with UV-LED pow- nos, Sustain. Sci. Eng., 2, 229 (2010).

Korean J. Chem. Eng.(Vol. 36, No. 6)

974 M. R. Eskandarian et al.

10. A. R. Ribeiro, O. C. Nunes, M. F. R. Pereira and A. M. T. Silva, Envi- Appl. Surf. Sci., 288, 51 (2014).
ron. Int., 75, 33 (2015). 42. J. Zhan, H. Zhang and G. Zhu, Ceram. Int., 40, 8547 (2014).
11. H. Park, H. Kim, G. Moon and W. Choi, Energy Environ. Sci., 9, 43. P. Zhang, Z. Mo, L. Han, X. Zhu, B. Wang and C. Zhang, Ind. Eng.
411 (2016). Chem. Res., 53, 8057 (2014).
12. A. Dhakshinamoorthy, S. Navalon, A. Corma and H. Garcia, Energy 44. Q. He, Z. Zhang, J. Xiong, Y. Xiong and H. Xiao, Opt. Mater.
Environ. Sci., 5, 9217 (2012). (Amst), 31, 380 (2008).
13. R. E. Galian and J. Perez-Prieto, Energy Environ. Sci., 3, 1488 (2010). 45. T. Harifi and M. Montazer, Sep. Purif. Technol., 134, 210 (2014).
14. R. T. Koodali and D. Zhao, Energy Environ. Sci., 3, 608 (2010). 46. F. Deng, Y. Li, X. Luo, L. Yang and X. Tu, Colloids Surfaces A Physi-
15. W.-W. Li, H.-Q. Yu and Z. He, Energy Environ. Sci., 7, 911 (2014). cochem. Eng. Asp., 395, 183 (2012).
16. S. J. A. Moniz, S. A. Shevlin, D. J. Martin, Z.-X. Guo and J. Tang, 47. S. Vilhunen and M. Sillanpää, Rev. Environ. Sci. Bio/Technology, 9,
Energy Environ. Sci., 8, 731 (2015). 323 (2010).
17. J. Zhang, Y. Wu, M. Xing, S. A. K. Leghari and S. Sajjad, Energy 48. W. Jo, G. T. Park and R. J. Tayade, J. Chem. Technol. Biotechnol., 90,
Environ. Sci., 3, 715 (2010). 2280 (2015).
18. T. Su, Q. Shao, Z. Qin, Z. Guo and Z. Wu, Acs Catal., 8, 2253 (2018). 49. E. Korovin, D. Selishchev, A. Besov and D. Kozlov, Appl. Catal. B
19. R. Ma, L. Dong, B. Li, T. Su, X. Luo, Z. Qin and H. Ji, Chemistry- Environ., 163, 143 (2015).
Select., 3, 5891 (2018). 50. Q. Zhang, C.-F. Wang, L.-T. Ling and S. Chen, J. Mater. Chem. C,
20. O. Legrini, E. Oliveros and A. M. Braun, Chem. Rev., 93, 671 (1993). 2, 4358 (2014).
21. J. Schneider, M. Matsuoka, M. Takeuchi, J. Zhang, Y. Horiuchi, M. 51. M. R. Eskandarian, M. Fazli, M. H. Rasoulifard and H. Choi, Appl.
Anpo and D. W. Bahnemann, Chem. Rev., 114, 9919 (2014). Catal. B Environ., 183, 407 (2016), DOI:10.1016/[Link].2015.11.004.
22. E. Brillas, I. Sirés and M. A. Oturan, Chem. Rev., 109, 6570 (2009). 52. M. R. Eskandarian, H. Choi, M. Fazli and M. H. Rasoulifard,
23. M. Pelaez, P. Falaras, A. G. Kontos, A. Armah, K. O’shea, P. S. M. Chem. Eng. J., 300, 414 (2016), DOI:10.1016/[Link].2016.05.049.
Dunlop, J. A. Byrne and D. D. Dionysiou, Appl. Catal. B Environ., 53. M. D. Sobsey, Managing water in the home: accelerated health
121, 30 (2012). gains from improved water supply, World Health Organization
24. J. Deng, Y. Shao, N. Gao, S. Xia, C. Tan, S. Zhou and X. Hu, Chem. (2002).
Eng. J., 222, 150 (2013). 54. S. Gundry, M. Sobsey and J. Wright, Waterlines., 23, 14 (2005).
25. X. Van Doorslaer, K. Demeestere, P. M. Heynderickx, H. Van Lan- 55. J. Brown and M. D. Sobsey, Am. J. Trop. Med. Hyg., 87, 394 (2012).
genhove and J. Dewulf, Appl. Catal. B Environ., 101, 540 (2011). 56. D. Mäusezahl, A. Christen, G. D. Pacheco, F. A. Tellez, M. Iriarte,
26. A. O. Kondrakov, A. N. Ignatev, F. H. Frimmel, S. Bräse, H. Horn M. E. Zapata, M. Cevallos, J. Hattendorf, M. D. Cattaneo and B.
and A. I. Revelsky, Appl. Catal. B Environ., 160, 106 (2014). Arnold, PLoS Med., 6, e1000125 (2009).
27. T. Alapi and A. Dombi, J. Photochem. Photobiol. A Chem., 188, 409 57. B. Parida, S. Iniyan and R. Goic, Renew. Sustain. Energy Rev., 15,
(2007). 1625 (2011).
28. R. F. Dantas, O. Rossiter, A. K. R. Teixeira, A. S. M. Simões and V. L. 58. O. Autin, C. Romelot, L. Rust, J. Hart, P. Jarvis, J. MacAdam, S. A.
da Silva, Chem. Eng. J., 158, 143 (2010). Parsons and B. Jefferson, Chemosphere, 92, 745 (2013).
29. S. Sanches, M. T. B. Crespo and V. J. Pereira, Water Res., 44, 1809 59. M. A. Würtele, T. Kolbe, M. Lipsz, A. Külberg, M. Weyers, M.
(2010). Kneissl and M. Jekel, Water Res., 45, 1481 (2011).
30. S. Bounty, R. A. Rodriguez and K. G. Linden, Water Res., 46, 6273 60. M. A. S. Ibrahim, J. MacAdam, O. Autin and B. Jefferson, Environ.
(2012). Technol., 35, 400 (2014).
31. C. Pablos, J. Marugán, R. van Grieken and E. Serrano, Water Res., 61. C. Chatterley and K. Linden, J. Water Health, 8, 479 (2010).
47, 1237 (2013). 62. G. Y. Lui, D. Roser, R. Corkish, N. Ashbolt, P. Jagals and R. Stuetz,
32. A. Adak, K. P. Mangalgiri, J. Lee and L. Blaney, Water Res., 70, 74 Sci. Total Environ., 493, 185 (2014).
(2015). 63. G. Y. Lui, D. Roser, R. Corkish, N. J. Ashbolt and R. Stuetz, Sci.
33. S. Chakma, S. Praneeth and V. S. Moholkar, Ultrason. Sonochem., Total Environ., 553, 626 (2016).
38, 652 (2017). 64. O. Iglesias, J. Gómez, M. Pazos and M. Á. Sanromán, Appl. Catal.
34. S. Chakma and V. S. Moholkar, Indian Chem. Eng., 57, 359 (2015). B Environ., 144, 416 (2014).
35. M. M. Khin, A. S. Nair, V. J. Babu, R. Murugan and S. Ramakrishna, 65. S. Wu, A. Sun, F. Zhai, J. Wang, W. Xu, Q. Zhang and A. A. Volin-
Energy Environ. Sci., 5, 8075 (2012). sky, Mater. Lett., 65, 1882 (2011).
36. M. Wang, J. Ioccozia, L. Sun, C. Lin and Z. Lin, Energy Environ. 66. Y. Eom, M. Abbas, H. Noh and C. Kim, RSC Adv., 6, 15861 (2016).
Sci., 7, 2182 (2014). 67. S. Liu, K. Yao, L.-H. Fu and M.-G. Ma, RSC Adv., 6, 2135 (2016).
37. I. Ali, Chem. Rev., 112, 5073 (2012). 68. G. P. Anipsitakis and D. D. Dionysiou, Environ. Sci. Technol., 37,
38. M. Rasoulifard, M. Fazli and M. Eskandarian, J. Ind. Eng. Chem., 4790 (2003).
20, 3695 (2014). 69. G. P. Anipsitakis and D. D. Dionysiou, Environ. Sci. Technol., 38,
39. L. Tan, X. Zhang, Q. Liu, X. Jing, J. Liu, D. Song, S. Hu, L. Liu and J. 3705 (2004).
Wang, Colloids Surfaces A Physicochem. Eng. Asp., 469, 279 (2015). 70. Z. Xu, C. Shan, B. Xie, Y. Liu and B. Pan, Appl. Catal. B Environ.,
40. M. A. Habila, Z. A. ALOthman, A. M. El-Toni, J. P. Labis and M. 200, 439 (2017).
Soylak, Talanta, 154, 539 (2016). 71. J. R. Bolton, K. G. Bircher, W. Tumas and C. A. Tolman, Pure Appl.
41. T. Xin, M. Ma, H. Zhang, J. Gu, S. Wang, M. Liu and Q. Zhang, Chem., 73, 627 (2001).

June, 2019