Chemical Engineering Journal 373 (2019) 1329–1337

Contents lists available at ScienceDirect

Chemical Engineering Journal

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

Enhanced degradation of paracetamol in water using sulfate radical-based T

advanced oxidation processes catalyzed by 3-dimensional Co3O4 nanoflower
Wan-Chu Yuna, Kun-Yi Andrew Lina, , Wai-Chi Tonga, Yi-Feng Linb, , Yunchen Duc,
⁎ ⁎ ⁎

a
Department of Environmental Engineering, National Chung Hsing University, 250 Kuo-Kuang Road, Taichung, Taiwan
b
Department of Chemical Engineering and R&D Center for Membrane Technology, Chung Yuan Christian University, 200 Chung Pei Rd., Chungli, Taoyuan, Taiwan
c
School of Chemistry and Chemical Engineering, Harbin Institute of Technology, Harbin, China

HIGHLIGHTS GRAPHICAL ABSTRACT

• Adeveloped
facile but convenient protocol is
to prepare a Co O nano-
3 4
flower (CNF).
• CNF, with needle-like petals, exhibits
tremendous higher surface area and
pore volume.
• CNF shows significantly higher cata-
lytic activities than the commercial
Co3O4 NPs.
• APAP degradation is confirmed to in-
volve with sulfate radicals via EPR
analysis.
• The APAP degradation pathway by
this CNF-activated Oxone is also in-
vestigated.

ARTICLE INFO ABSTRACT

Keywords: Paracetamol (APAP) represents a critical emerging contaminant in various water bodies and sulfate radical
Paracetamol (SO4%−)-involved Advanced Oxidation Processes (AOPs) using Oxone are proven as promising approaches for
Cobalt degrading APAP. While Co3O4 is the benchmark catalyst for activating Oxone, Co3O4 nanoparticles are usually
Oxone found to aggregate severely, losing their catalytic activities. In this study, a facile but convenient protocol is
Sulfate radical
developed to prepare a unique Co3O4 material, which exhibits a morphology of nanoflower. More importantly,
Nanoflower
the flower petals of this Co3O4 nanoflower (CNF) are needle-shaped and comprised of numerous fine Co3O4
particles which are separated, enabling this CNF to exhibit high surface areas and large pore volumes. CNF also
shows significantly higher catalytic activities than the commercial Co3O4 NPs for activating Oxone to degrade
APAP. APAP degradation is confirmed to involve with sulfate radicals via examining effects of radical probe
reagents and EPR analysis. CNF is reusable over several continuous cycles. The APAP degradation pathway is
also investigated according to the degradation products determined using LC-MS/MS. The findings of this study
are valuable to advance treatments of APAP in water by sulfate-radical-based processes, and also insightful to
determine catalytic characteristics of Co3O4 with such a unique morphology of nanoflower.

⁎
Corresponding authors.
E-mail addresses: linky@nchu.edu.tw (K.-Y.A. Lin), yflin@cycu.edu.tw (Y.-F. Lin), yunchendu@hit.edu.cn (Y. Du).

https://doi.org/10.1016/j.cej.2019.05.142
Received 10 January 2019; Received in revised form 22 April 2019; Accepted 21 May 2019
Available online 22 May 2019
1385-8947/ © 2019 Elsevier B.V. All rights reserved.
W.-C. Yun, et al. Chemical Engineering Journal 373 (2019) 1329–1337

1. Introduction In this study, we particularly propose a facile but convenient pro-

tocol to prepare a Co3O4 material, which exhibits a unique morphology
Paracetamol (also known as APAP) is considered as one of the most of nanoflower. More importantly, the flower petals of this Co3O4 na-
consumed non-prescription medicines to treat pain and fever [1], and noflower (CNF) are needle-shaped and comprise of numerous fine
thus APAP has been extensively detected in various water bodies as Co3O4 particles which are specially configured to be separated rather
listed in Table S1 (see the supporting information), including influents than being aggregated, enabling this CNF to exhibit high surface areas
and effluents of wastewater treatment plants, hospital effluents, and large pore volumes. Therefore, CNF should be a promising and
groundwater, surface water and so on, ranged from 10 to 134,000 ng/L advantageous Co3O4-based heterogeneous catalyst for activating Oxone
[2–7]. Even though removal of APAP from water bodies is crucial, ty- to degrade APAP. Especially, as no similar materials have been reported
pical water and wastewater treatments seem impotent for degrading for APAP degradation, investigation of this study would be valuable to
APAP [8]. Accordingly, several advanced techniques have been devel- advance treatments of APAP, and also insightful to probe into catalytic
oped to degrade APAP, including Fenton’s reactions [9–11], ozonation characteristics of Co3O4 with such a unique morphology of nanoflower.
[12], and UV irradiation [13]. Apart from these approaches, recently Additionally, commercial Co3O4 NPs are directly used here for
SO4%−-involved techniques receive increasing attention for degrading comparisons with CNF for activating Oxone to degrade APAP, and the
APAP, as SO4%− possesses several merits, such as high oxidation po- mechanism of APAP degradation by Oxone activated by CNF is in-
tentials (i.e., 2.5–3.1 V), high selectivities for unsaturated and aromatic vestigated and the corresponding degradation pathway is also de-
structures [14,15], and long half-life (∼40 μs) [14,16]. termined.
To obtain SO4%−, Oxone reagent is frequently-adopted as Oxone is
environmentally benign [17]. Nonetheless, Oxone requires “activation”
2. Experimental
to facilitate generation of SO4%− from Oxone as Oxone is difficult to
self-decompose. While numerous approaches have been proposed to
2.1. Materials
“activate” Oxone, utilization of metal catalysts for activating Oxone has
been deemed as practical and efficient approaches [18–20]. Hence,
All chemicals in this study were purchased from fine chemical
several transition metals have been studied for activating Oxone to
suppliers and used directly without purification. Paracetamol (APAP),
degrade APAP, such as Mn [21,22], Co [22] and Fe [8], etc.
cobalt sulfate (Co(SO4)2), Oxone, tert-butyl alcohol (TBA) and 5,5-di-
Despite the fact that Mn and Fe are able to activate Oxone, Co has
methyl-1-pyrroline N-oxide (DMPO) were received from Sigma-Aldrich
been validated as the most effective transition metal to activate Oxone
(USA). Co3O4 nanoparticles (50–80 nm) and urea were purchased from
[23,24]. Thus, many efforts have been made to develop Co-based cat-
Alfa Aesar (USA). Ethanol and methanol were obtained from Echo
alysts, especially heterogeneous ones, for activating Oxone to degrade
Chemicals (Taiwan). Deionized (DI) water was prepared to less than
APAP. To date, Co3O4 has been a longstanding popular Co-based het-
18 MΩ-cm.
erogeneous catalyst for Oxone activation [18,25]. Many studies even
fabricated Co3O4 nanoparticles (NPs) and anticipated that Co3O4 NPs
may exhibit enhanced catalytic activities for Oxone activation owing to 2.2. Preparation and characterization of CNF
the nanoscale dimensions of Co3O4 NPs [26,27]. Unfortunately, NPs
agglomerate severely, especially in water, losing their contacting sur- The fabrication of CNF can be illustrated as Fig. 1. A precursor
faces and diminishing catalytic activities [26]. Therefore, it is critical to mixture was firstly prepared by mixing 30 mmol of Co(SO4)2 and
develop Co3O4 NPs-based catalysts, which, however, exhibit high sur- 150 mmol of urea in 150 mL of DI water. Next, the mixture was trans-
faces and porosities for enhancing their catalytic activities. While Co3O4 ferred to a Teflon-lined stainless steel autoclave for a hydrothermal
can be fabricated into various configurations to achieve high surface process at 100 °C for 12 h. The resulting precipitate was then collected
areas and pore volumes, many techniques involve complicated proto- and washed thoroughly with ethanol and DI water, and dried at 65 °C
cols and reagents, such as electrospinning [28], hard-templated and for 12 h, followed by calcination at 300 °C for 5 h at a heating rate of
soft-templated methods [29,30], etc. Thus, it would be even more ad- 1 °C/min to afford CNF.
vantageous to employ simple processes and reagents to prepare such The morphology of CNF was measured by scanning and transmis-
Co3O4 materials. sion electronic microscopies (SEM/TEM) (JEOL, Japan), whereas the
chemical composition of CNF was determined using Energy-dispersive

Fig. 1. The preparation scheme for 3-D Co3O4 nanoflower (CNF).

1330
W.-C. Yun, et al. Chemical Engineering Journal 373 (2019) 1329–1337

Fig. 2. The morphology of CNF: (a) SEM; (b) and (c) TEM images and (d) HRTEM.

X-ray spectroscopy (EDS) (Oxford Instrument, UK). The crystalline emission spectrometry (ICP-OES) (Perkin Elmer Optima, USA).
structure of CNF was characterized by an X-ray diffractometer (Bruker,
USA). The surface chemistry of CNF was then analyzed using X-ray 3. Results and discussion
photoelectron spectroscopy (XPS) (PHI 5000, ULVAC-PHI, Japan).
Textural properties of CNF and the commercial Co3O4 NPs were mea- 3.1. Characterization of CNF
sured by a volumetric gas adsorption analyzer (Quantachrome Auto IQ,
USA). The morphology of CNF was first visualized using SEM in Fig. 2(a),
in which several flower-shaped spheres were bundled together, vali-
2.3. Degradation of paracetamol (APAP) by Oxone activated using CNF dating that the resulting CNF certainly exhibited the flower-shaped
appearance. The size of single CNF was ranging from 1 to 3 µm. On the
Degradation of APAP using Oxone activated by CNF was in- other hand, Fig. 2(b) shows the TEM image of the edge of CNF, re-
vestigated by batch-type degradation experiments. In a typical experi- vealing that CNF had very fine and thin flower petals. A closer view
ment, Oxone (e.g., 40 mg) was firstly dissolved in 0.2 L of APAP solu- (Fig. 2(c)) further validates that these petals of CNFs exhibited the
tion with an initial concentration C0 of 5 mg/L at a desired temperature needle-shaped morphology and, more interestingly, these petals were
(e.g., 30 °C). Subsequently, CNF (e.g., 10 mg) was immediately added to comprised of nanoscale spherical particles. Fig. 2(d) further displays a
the APAP solution. At a pre-set time, a sample aliquot was filtrated HRTEM image of the needle-shaped petals of CNF and the observed
through a PVDF syringe filter to separate CNF from the APAP solution. lattice spacing of 0.244 nm and 0.286 nm, corresponding to the (3 1 1)
The filtrate was then analyzed by a UV–Vis spectrophotometer (Jasco and (2 2 2) planes of Co3O4, confirming that CNF was assuredly com-
V670, Japan) at 243 nm to determine the remaining concentration of prised of Co3O4.
APAP (Ct). Furthermore, electron paramagnetic resonance (EPR) Furthermore, EDS was employed to characterize CNF as shown in
(Bruker EMX-P, Germany) was employed to specify radical species Fig. 3(a) and only Co and O were detected, confirming that these
generated from CNF-activated Oxone using DMPO as a radical-trapping needle-shaped petals and CNF were certainly made of cobalt oxide.
agent. The recyclability of CNF for activating Oxone to degrade APAP Fig. 3(b) also validates that CNF was made of Co3O4 as its crystalline
was evaluated by reusing CNF over multiple cycles of APAP degrada- structure could be well-indexed to the typical XRD pattern of Co3O4
tion experiments. The used CNF was collected by centrifugation and based on JCPDS #43-1003.
washed thoroughly by ethanol and water to remove impurities and Moreover, the surface chemistry of CNF was analyzed using XPS and
residues, followed by drying before reuse. The resulting CNF was then its full-spectrum survey can be seen in Fig. 4(a), in which only Co and O
used as the pristine CNF in the new batch of APAP solution with Oxone signals can be detected, validating no other significant elements in CNF.
for a new experiment of APAP degradation. Degradation products of Fig. 4(b) further reveals the Co2p core-level spectrum of CNF, which
APAP were then identified by liquid chromatography tandem-mass was deconvoluted to display several noticeable peaks. The peaks at
spectrometry (Thermo TSQ Quantum, USA). The concentration of Co 780.2 and 795.1 eV were derived from Co3+, whereas the peaks at
ion in solutions were analyzed by inductively coupled plasma-optical 781.8 and 796.9 eV could be corresponding to Co2+. These specific Co

1331
W.-C. Yun, et al. Chemical Engineering Journal 373 (2019) 1329–1337

Fig. 3. Characteristics of CNF (a) EDS and (b) XPS pattern.

species were attributed to Co3O4 [31], confirming the content of Co3O4 3.2. Paracetamol degradation by Oxone activated using CNF
in CNF.
Since textural properties are crucial parameters for catalytic reac- Prior to evaluating APAP degradation by Oxone activated using
tions, N2 sorption/desorption isotherms of CNF were then analyzed CNF, it was essential to examine whether APAP could be removed due
(Fig. 5(a)). The isotherm of CNF appeared to be a combination of the to adsorption to CNF. Fig. 6 displays that the concentration of APAP
IUPAC type III and IV isotherms, suggesting that CNF was a porous remained almost the same after 30 min, demonstrating that CNF was
material. Fig. 5(b) confirms that CNF was comprised of pores (ranging unable to remove APAP via adsorption. On the other hand, Oxone alone
from 2 to 150 nm) with a high pore volume of 0.41 cm3/g and the was also examined for APAP degradation; however APAP was only
surface area was determined as 121 m2/g. In this study, the commercial slightly eliminated after 30 min. This indicates that Oxone (without
Co3O4 NPs were employed for comparisons with CNF; thus the textural activation) could not efficiently degrade APAP. However, when CNF
properties of Co3O4 NPs were then characterized and shown in and Oxone were both added to the APAP solution, the concentration of
Fig. 5(a). However, the N2 adsorption amount to the Co3O4 NPs seemed APAP rapidly diminished and Ct/C0 approached zero in 30 min,
very low and the corresponding pores were also very few, leading to a showing that CNF combined with Oxone could degrade APAP. Fig. S2
significantly small pore volume of 0.007 cm3/g and a very low surface also demonstrates that APAP was certainly mineralized by the combi-
area of 2.8 m2/g. These comparisons indicate that CNF was significantly nation of CNF and Oxone, validating that APAP was decomposed. As
more porous than the commercial Co3O4 NPs. This is probably because CNF and Oxone were unable to degrade APAP individually, this effec-
the Co3O4 NPs, even though their size was a few tens nanometers, tive degradation of APAP by CNF and Oxone demonstrated that Oxone
tended to aggregate as shown in Fig. S1. On the contrary, those Co3O4 was activated by CNF.
particles of CNF were specially arranged and separated owing to the Since CNF consisted of cobaltic species of Co2+ and Co3+ from
needle-shaped floral configuration, thus leading to the much higher Co3O4, these Co species can react with Oxone for generating SO4%− as
pore volume and surface area. follows [32,33]:

Co2+@CNF + HSO5− → Co3+@CNF + SO4%− + OH− (1)

3+ 2+
Co @ CNF → Co @CNF + SO5%− + H+ (2)

Fig. 4. XPS analyses of CNF: (a) Full-survey spectrum; (b) Co2p core-level spectrum.

1332
W.-C. Yun, et al. Chemical Engineering Journal 373 (2019) 1329–1337

Fig. 5. CNF and the commercial Co3O4 NPs: (a) N2 sorption/desorption isotherms of CNF and (b) Pore size distributions.

Furthermore, the resulting SO5%− can evolve to SO4%− as follows: it with the Co3O4 NPs. As displayed in Fig. 6, APAP was certainly de-
graded in the presence of Co3O4 NPs and Oxone, suggesting that Co3O4
SO5%− → 2 SO4%− + O2 (3)
NPs were capable of activating Oxone to degrade APAP. Nevertheless,
Consequently, SO4%− can be continuously generated through these the degradation extent was significantly lower than that by CNF. This
catalytic cyclic reactions to degrade APAP. validates that even though Co3O4 NPs consisted of Co2+ and Co3+, the
Even though CNF was capable of activating Oxone to degrade APAP, corresponding catalytic activity was much lower than that of CNF. This
Oxone could be just activated by Co ion leached out from CNF. significant difference between CNF and Co3O4 NPs could be attributed
Therefore, Co ion leached out from CNF was then studied. Nonetheless, to the noticeably distinct textural properties between CNF and Co3O4
the concentration of Co ion leached out from CNF in the end of the NPs. As revealed in the earlier section, CNF exhibited a tremendous
APAP degradation experiment was only 0.1 mg/L, which was sig- larger surface area and pore volume than the Co3O4 NPs because the
nificantly less than the concentration of CNF (i.e., 50 mg/L) in water. Co3O4 particles were specially-configured in the flower-like CNF. In
Fig. 6 also reveals APAP degradation using Oxone activated by 0.1 mg/ contrast, the conventional Co3O4 NPs were highly agglomerated, and
L of Co ion (Co2+) to evaluate homogeneous activation of Oxone for thus their pore volume and surface were low, leading to ineffective
APAP degradation. Nevertheless, the corresponding degradation extent activation of Oxone.
of APAP was very limited, demonstrating that APAP degradation by
CNF-activated Oxone was mainly attributed to heterogeneous activa- 3.3. Effects of CNF and Oxone dosages on APAP degradation
tion of Oxone.
Although CNF was able to activate Oxone, it was crucial to compare Even though the combination of CNF and Oxone was capable of
degrading APAP, it was essential to determine their respective con-
tributions to APAP degradation. Thus, effects of CNF and Oxone do-
sages on APAP degradation were then distinguished. Fig. 7(a) reveals
APAP degradation by CNF at different dosages and CNF at the four
dosages degraded APAP efficiently within 30 min.
This result demonstrates that CNF should exhibit a very high cata-
lytic activity to activate Oxone as a low dosage of CNF at 25 mg/L could
still activate Oxone for almost complete degradation of APAP.
Nonetheless, the degradation kinetics seemed different at various cat-
alyst dosages. To quantitatively compare the degradation kinetics, the
pseudo first order equation, which has been extensively employed for
determining kinetics of pollutant degradation by Oxone, was then
employed as follows [34]:

Ct = C0 exp( kobs t ) (4)

where kobs is the observed pseudo first order rate constant. When CNF
increased from 25 to 50 mg/L, kobs increased from 0.050 to 0.103 min−1
(Table S2, see the Supplementary material) and further rose up to 0.127
and 0.312 min−1 at CNF = 100 and 200 mg/L, respectively. This vali-
dates that a higher CNF dosage accelerated APAP degradation by pro-
moting generation of sulfate radicals from Oxone in the presence of
more active sites.
Furthermore, as CNF dosage varied and their degradation extents
were quite comparable in Fig. 7(a), their rate constants were then
Fig. 6. Comparison of APAP degradation by Oxone alone, adsorption to CNF, normalized to surface area and thus ratios of rate constants over their
CNF-activated Oxone Co3O4 NP-activated Oxone and Co2+-activated Oxone surface areas for CNF at 25, 50, 100, 200 mg/L were calculated as
(Co2+ = 0.1 mg/L) (catalyst = 50 mg/L, APAP = 5 mg/L, Oxone = 200 mg/L, 0.117, 0.085, 0.052 and 0.064 min−1 m−2, respectively. This indicates
and T = 30 °C). that a relatively low CNF dosage led to a higher degradation ratio per

1333
W.-C. Yun, et al. Chemical Engineering Journal 373 (2019) 1329–1337

Fig. 7. Effects of (a) CNF dosage (APAP = 5 mg/L), Oxone = 200 mg/L and (b) Oxone dosage (catalyst = 50 mg/L) (APAP = 5 mg/L and T = 30 °C) on APAP
degradation.

unit of surface area, possibly because the surface of CNF at a low dosage 45 and 60 °C. As the solution temperature increased, the APAP de-
could be more exposed to reactants, leading to a higher-degree of uti- gradation was substantially expedited. For instance, when temperature
lization of CNF surface. increased from 30 to 45 °C, kobs accordingly increased from 0.103 to
APAP degradation by Oxone at different dosages was also examined 0.215 min−1 and then further changed to 0.371 min−1 at 60 °C, con-
in Fig. 7(b). When Oxone dosage increased from 200 to 400 mg/L, firming that higher temperatures considerably accelerated APAP de-
APAP degradation was accelerated as kobs increased from 0.103 to gradation by CNF-activated Oxone.
0.145 min−1. Nevertheless, as Oxone dosage decreased to 100 mg/L, As higher kobs values were acquired at higher temperatures, kobs
the degradation kinetics was lowered to 0.039 min−1 and more im- values were further correlated to temperatures via the following
portantly, the degradation extent of APAP also became 70%. This de- Arrhenius equation:
monstrates that the lower dosage of Oxone led to ineffective APAP
degradation probably due to insufficient SO4%− from Oxone. Ln kobs = ln A Ea/ RT (5)
As shown in Fig. 7(a), when CNF dosage was varied, even though
where A represents the temperature-independent factor; R denotes the
degradation kinetics was altered noticeably, the degradation extent was
universal gas constant; Ea is the activation energy (Ea, kJ/mol) of APAP
consistent as APAP was completely eliminated at 30 min. Nevertheless,
degradation; and T means the solution temperature in Kelvin (K). Based
when Oxone dosage was changed (Fig. 7(b)), not only degradation ki-
on Eq. (5), a plot of 1/T vs. Ln kobs was then drawn as the inset in
netics was altered but also was the degradation extent. This demon-
Fig. 8(a) and the data points were properly fit by the linear regression
strates that the degradation extent of APAP seemed more involved with
with R2 = 0.997 and then the corresponding Ea was calculated as 41
Oxone dosages. On the contrary, CNF should act as a catalyst to ma-
kJ/mol.
nipulate the kinetics. Several previous studies have also observed si-
In addition, since APAP degradation by CNF-activated Oxone was
milar respective contributions of catalysts and Oxone dosage to de-
an aqueous reaction, solution pH represents a critical parameter.
gradation of pollutants [35,36].
Fig. 8(b) displays APAP degradation by CNF-activated Oxone at dif-
ferent pH values. When pH was decreased to 5, the APAP degradation
3.4. Effects of temperature and initial pH on APAP degradation was unfavorably affected as kobs decreased to 0.036 min−1 and Ct/C0
approached 0.38. The unfavorable effect became even more pro-
Fig. 8(a) reveals APAP degradation by CNF-activated Oxone at 30, nounced when the pH was even lower to 3 as kobs diminished to 0.012

Fig. 8. Effects of (a) temperature and (b) pH (T = 30 °C) on APAP degradation (catalyst = 50 mg/L, Oxone = 200 mg/L, APAP = 5 mg/L).

1334
W.-C. Yun, et al. Chemical Engineering Journal 373 (2019) 1329–1337

more intense [8], restricting contact between CNF and SO52−. Hence,
the APAP degradation became less effective under highly alkaline
conditions.
On the other hand, when SO4%− was present under alkaline con-
ditions, it could also react with OH− to result in OH% [8,39]. Never-
theless, OH% might be scavenged by anionic by-products derived from
Oxone [39], leading to less effective degradation [8,39]. Therefore, it
was consistent to observe that APAP degradation became less effective
under highly alkaline conditions [8,39].

3.5. The recyclability of CNF to activate Oxone for APAP degradation

Since CNF was proposed to be a heterogeneous catalyst for acti-

vating Oxone to degrade APAP, it was essential to evaluate whether
CNF could be reused to activate Oxone to degrade APAP. Fig. 9 displays
that APAP was still effectively degraded using Oxone activated by used
CNF over the multiple cycles. This indicates that CNF was a reusable,
durable and efficient catalyst for activating Oxone to APAP.
Fig. S4(a) further displays a SEM image of used CNF, which still
exhibited the floral morphology with petals, and its crystalline structure
was also comparable to that of pristine CNF (Fig. S4(b)), demonstrating
that CNF remained intact without significant destruction after Oxone
Fig. 9. The recyclability test (catalyst = 50 mg/L, Oxone = 200 mg/L, activation. Besides, the surface chemistry of CNF (Fig. S4(c) and (d))
APAP = 5 mg/L, and T = 30 °C). further indicates that Co species (Co2+ and Co3+) in used CNF were
consistent to those in pristine CNF. However the fractions of Co2+ and
min−1 and Ct/C0 only approached 0.7. This feature was possibly due to Co3+ of used CNF seemed to slightly change in comparison to those of
the fact that Oxone becomes much more stable under acidic conditions pristine, suggesting that these species participated in Oxone reaction as
[25]; thus Oxone becomes relatively difficult to be activated. discussed previously.
In contrast, when the aqueous solution became alkaline to pH = 9,
APAP degradation was slightly decreased as kobs diminished to 3.6. A proposed mechanism of APAP degradation by CNF-activated Oxone
0.084 min−1. This adverse effect become pronounced as pH was in-
creased to 11 and the corresponding kobs decreased to 0.027 min−1. Even though CNF was proven as a promising and useful catalyst for
This was because Oxone tends to decompose without generation of Oxone activation to degrade APAP, it was also important to probe into
SO4%− under alkaline conditions [25,37] and thus SO4%− could be not the APAP degradation mechanism by CNF-activated Oxone. In parti-
produced rapidly to degrade APAP. In addition, as the pKa of APAP is cular, when SO4%− is present in water, it may further evolve into OH%
9.38, the strongly alkaline condition (e.g., pH = 11) caused the de- through the following equation [40] (Eq. (6)):
protonation of APAP. Simultaneously, the CNF surface became even
more negatively-charged as pHzpc of CNF was 6.5 (as show in Fig. S3) SO4%− + H2O → SO42− + OH% + H+ (6)
and thus the electrostatic repulsion between CNF and the resulting Hence, it was essential to determine whether these potential species
deprotonated APAP (APAP−) might increase, hindering with the con- (i.e., OH% and SO4%−) contributed to APAP degradation by CNF-acti-
tact between APAP and CNF. On the other hand, since the pKa of vated Oxone. Herein, two special radical probe reagents were adopted:
HSO5− was 9.4 [38], HSO5− would also become SO52− under strongly TBA and methanol. Since TBA does not contain α-hydrogen and it can
alkaline conditions and thus the electrostatic repulsion might also grow react with OH% very rapidly, TBA has been adopted as a probe reagent

Fig. 10. (a) effect of radical probe reagents

on APAP degradation (catalyst = 50 mg/L,
Oxone = 200 mg/L, APAP = 5 mg/L, TBA
= methanol = 0.5 M, and T = 30 °C); (b)
EPR spectra of Oxone alone and Oxone in
the presence of CNF (□: DMPO-OH%; ○:
DMPO- SO4%−).

1335
W.-C. Yun, et al. Chemical Engineering Journal 373 (2019) 1329–1337

Fig. 11. A possible degradation pathway of APAP by CNF-activated Oxone.

for OH%. Fig. 10(a) reveals that APAP degradation by CNF-activated shaped petals, enabling Co3O4 particles separated. As a result, CNF
Oxone in the presence of TBA was slightly affected. This suggests that exhibited the tremendous higher surface area and pore volume than the
OH% might participate in APAP degradation. On the other hand, another commercial Co3O4 NPs. CNF also showed significantly higher catalytic
radical probe reagent, methanol, was also adopted as methanol, con- activities than the commercial Co3O4 NPs for activating Oxone to de-
taining α-hydrogen, is able to react with both SO4%− and OH% at grade APAP. CNF was also reusable for activating Oxone to degrade
comparable rates [41]. Fig. 10(a) shows that APAP degradation was APAP over several continuous cycles. The APAP degradation pathway
noticeably hindered in the presence of methanol, demonstrating that by this CNF-activated Oxone was also investigated according to the
APAP degradation was certainly attributed to SO4%−. degradation products determined using LC-MS/MS. The findings of this
To further identify radical species generated from CNF-activated study validate that CNF is a promising and advantageous heterogeneous
Oxone, EPR analyses were performed as displayed in Fig. 10(b). When catalyst for effectively activating Oxone to degrade APAP.
DMPO was used as a spin-trapping agent, there was almost no signal
present in the presence of Oxone alone, indicating that no significant Acknowledgements
amounts of radicals were produced from Oxone without activation.
Nevertheless, as CNF was used to activate Oxone, a distinct signal of the This work is supported by the Ministry of Science and Technology
triplet pattern was detected and ascribed to the hyperfine splitting of (MOST), Taiwan, and financially supported by the “Innovation and
oxidation adduct products of DMPO-SO4 and DMPO-OH [42–46]. This Development Center of Sustainable Agriculture” from The Featured
suggests that APAP degradation was attributed to both SO4%− and OH%, Areas Research Center Program within the framework of the Higher
generated from Oxone activated by CNF. Education Sprout Project by the Ministry of Education (MOE), Taiwan.

3.7. A potential degradation pathway for APAP by CNF-activated Oxone Appendix A. Supplementary data

While APAP was efficiently degraded by CNF-activated Oxone, it Supplementary data to this article can be found online at https://
was important to probe into the degradation pathway of APAP. Fig. S5 doi.org/10.1016/j.cej.2019.05.142.
displays spectra of degradation intermediate products determined by
LC-MS/MS. In view of these detected intermediate compounds from C1 References
to C6, a potential decomposition route of APAP by CNF-activated
Oxone can be illustrated in Fig. 11 [47,48]. [1] Y. Dao, H. Tran, T. Tran-Lam, T. Pham, G. Le, Degradation of paracetamol by an
UV/chlorine advanced oxidation process: influencing factors factorial design, and
Initially, APAP could be oxidized, leading to the formation of a intermediates identification, Int. J. Environ. Res. Public Health 15 (2018) 2637.
product, C1 [48], which was then oxidized to generate C2. Next, APAP, [2] M.J. Gomez, M.J. Martinez Bueno, S. Lacorte, A.R. Fernandez-Alba, A. Aguera, Pilot
C1 and C2 might be further oxidized to result in C3 [47], which might survey monitoring pharmaceuticals and related compounds in a sewage treatment
plant located on the Mediterranean coast, Chemosphere 66 (2007) 993–1002.
evolve into C4 and C5. Subsequently, C4 and C5 might further trans-
[3] M.J. Gomez, M. Petrovic, A.R. Fernandez-Alba, D. Barcelo, Determination of
formed to analogues of C6 [47]. The relatively simple aromatic pro- pharmaceuticals of various therapeutic classes by solid-phase extraction and liquid
ducts (C6) could be continuously oxidized to open rings and decom- chromatography-tandem mass spectrometry analysis in hospital effluent waste-
posed, leading to low-molecular weight compounds. waters, J. Chromatogr. A 1114 (2006) 224–233.
[4] K.K. Barnes, D.W. Kolpin, E.T. Furlong, S.D. Zaugg, M.T. Meyer, L.B. Barber, A
national reconnaissance of pharmaceuticals and other organic wastewater con-
4. Conclusions taminants in the United States–I) groundwater, Sci. Total Environ. 402 (2008)
192–200.
[5] K. Choi, Y. Kim, J. Park, C.K. Park, M. Kim, H.S. Kim, P. Kim, Seasonal variations of
In this study, a unique cobalt oxide nanoflower (CNF), consisting of several pharmaceutical residues in surface water and sewage treatment plants of
fine rounded Co3O4 particles, was facilely fabricated via a simple hy- Han River Korea, Sci. Total Environ. 405 (2008) 120–128.
[6] P.H. Roberts, K.V. Thomas, The occurrence of selected pharmaceuticals in waste-
drothermal process. In comparison to commercial Co3O4 NPs, which are
water effluent and surface waters of the lower Tyne catchment, Sci. Total Environ.
highly aggregated, Co3O4 particles were configured as the needle-

1336
W.-C. Yun, et al. Chemical Engineering Journal 373 (2019) 1329–1337

356 (2006) 143–153. (2010) 542–548.

[7] M.P. Meritxell Gros, Damià Barcelo, Tracing pharmaceutical residues of different [29] N. Dahal, I.A. Ibarra, S.M. Humphrey, High surface area mesoporous Co3O4 from a
therapeutic classes in environmental waters by using liquid chromatography/ direct soft template route, J. Mater. Chem. 22 (2012) 12675–12681.
quadrupole-linear ion trap mass spectrometry and automated library searching, [30] G. Kwak, J. Hwang, J.-Y. Cheon, M.H. Woo, K.-W. Jun, J. Lee, K.-S. Ha, Preparation
Anal. Chem. (2009). Method of Co3O4 nanoparticles using ordered mesoporous carbons as a template
[8] C. Tan, N. Gao, Y. Deng, J. Deng, S. Zhou, J. Li, X. Xin, Radical induced degradation and their application for fischer-tropsch synthesis, J. Phys. Chem. C 117 (2013)
of acetaminophen with Fe3O4 magnetic nanoparticles as heterogeneous activator of 1773–1779.
peroxymonosulfate, J. Hazard. Mater. 276 (2014) 452–460. [31] L.-L. Feng, M. Fan, Y. Wu, Y. Liu, G.-D. Li, H. Chen, W. Chen, D. Wang, X. Zou,
[9] C.-C. Su, A.-T. Chang, L.M. Bellotindos, M.-C. Lu, Degradation of acetaminophen by Metallic Co9S8 nanosheets grown on carbon cloth as efficient binder-free electro-
Fenton and electro-Fenton processes in aerator reactor, Separation Purif. Technol. catalysts for the hydrogen evolution reaction in neutral media, J. Mater. Chem. A 4
99 (2012) 8–13. (2016) 6860–6867.
[10] M.D.G. de Luna, M.L. Veciana, J.I. Colades, C.-C. Su, M.-C. Lu, Factors that influence [32] S. Muhammad, E. Saputra, H. Sun, J.D.C. Izidoro, D.A. Fungaro, H.M. Ang,
degradation of acetaminophen by Fenton processes, J. Taiwan Inst. Chem. Eng. 45 M.O. Tade, S. Wang, Coal fly ash supported Co3O4 catalysts for phenol degradation
(2014) 565–570. using peroxymonosulfate, RSC Adv. 2 (2012) 5645–5650.
[11] L.M. Bellotindos, A.-T. Chang, M.-C. Lu, Degradation of acetaminophen by different [33] Y. Yao, Z. Yang, H. Sun, S. Wang, Hydrothermal synthesis of Co3O4–graphene for
Fenton processes, Desalination Water Treatment 56 (2015) 1372–1378. heterogeneous activation of peroxymonosulfate for decomposition of phenol, Ind.
[12] J. Gomes, R. Costa, R.M. Quinta-Ferreira, R.C. Martins, Application of ozonation for Eng. Chem. Res. 51 (2012) 14958–14965.
pharmaceuticals and personal care products removal from water, Sci. Total [34] Q. Wang, Y. Shao, N. Gao, W. Chu, J. Chen, X. Lu, Y. Zhu, N. An, Activation of
Environ. 586 (2017) 265–283. peroxymonosulfate by Al2O3-based CoFe2O4 for the degradation of sulfa-
[13] Y. Zhu, M. Wu, N. Gao, W. Chu, K. Li, S. Chen, Degradation of phenacetin by the chloropyridazine sodium: Kinetics and mechanism, Sep. Purif. Technol. 189 (2017)
UV/chlorine advanced oxidation process: kinetics, pathways, and toxicity evalua- 176–185.
tion, Chem. Eng. J. 335 (2018) 520–529. [35] C.-B. Chen, F. Zhang, C.-X. Li, J.-Y. Lu, S. Cui, H.-Q. Liu, W.-W. Li, A magnetic
[14] P. Neta, R.E. Huie, A.B. Ross, Rate constants for reactions of inorganic radicals in CoFe2O4-CNS nanocomposite as an efficient, recyclable catalyst for perox-
aqueous solution, J. Phys. Chem. Ref. Data 17 (1988) 1027–1284. ymonosulfate activation and pollutant degradation, RSC Adv. 7 (2017)
[15] M.G. Antoniou, A.A. de la Cruz, D.D. Dionysiou, Degradation of microcystin-LR 55020–55025.
using sulfate radicals generated through photolysis, thermolysis and e-transfer [36] Y. Pi, L. Ma, P. Zhao, Y. Cao, H. Gao, C. Wang, Q. Li, S. Dong, J. Sun, Facile green
mechanisms, Appl. Catal. B 96 (2010) 290–298. synthetic graphene-based Co-Fe Prussian blue analogues as an activator of perox-
[16] E.G. Janzen, Y. Kotake, H.D. Randall, Stabilities of hydroxyl radical spin adducts of ymonosulfate for the degradation of levofloxacin hydrochloride, J. Colloid Interface
PBN-type spin traps, Free Rad. Biol. Med. 12 (1992) 169–173. Sci. 526 (2018) 18–27.
[17] P. Hu, M. Long, Cobalt-catalyzed sulfate radical-based advanced oxidation: a review [37] A. Rastogi, S.R. Al-Abed, D.D. Dionysiou, Sulfate radical-based ferrous–perox-
on heterogeneous catalysts and applications, Appl. Catal. B: Environ. 181 (2016) ymonosulfate oxidative system for PCBs degradation in aqueous and sediment
103–117. systems, Appl. Catal. B 85 (2009) 171–179.
[18] G.P. Anipsitakis, E. Stathatos, D.D. Dionysiou, Heterogeneous activation of oxone [38] Y.-H. Guan, J. Ma, X.-C. Li, J.-Y. Fang, L.-W. Chen, Influence of pH on the formation
using Co3O4, J. Phys. Chem. B 109 (2005) 13052–13055. of sulfate and hydroxyl radicals in the UV/peroxymonosulfate system, Environ. Sci.
[19] G. Wei, X. Liang, Z. He, Y. Liao, Z. Xie, P. Liu, S. Ji, H. He, D. Li, J. Zhang, Technol. 45 (2011) 9308–9314.
Heterogeneous activation of Oxone by substituted magnetites Fe3−xMxO4 (Cr, Mn [39] J. Deng, Y. Shao, N. Gao, Y. Deng, C. Tan, S. Zhou, Zero-valent iron/persulfate(Fe0/
Co, Ni) for degradation of Acid Orange II at neutral pH, J. Mol. Catal. A: Chem. 398 PS) oxidation acetaminophen in water, Int. J. Environ. Sci. Technol. 11 (2014)
(2015) 86–94. 881–890.
[20] K.-Y.A. Lin, H.-A. Chang, R.-C. Chen, MOF-derived magnetic carbonaceous nano- [40] L.J. Xu, W. Chu, L. Gan, Environmental application of graphene-based CoFe2O4 as
composite as a heterogeneous catalyst to activate oxone for decolorization of an activator of peroxymonosulfate for the degradation of a plasticizer, Chem. Eng.
Rhodamine B in water, Chemosphere 130 (2015) 66–72. J. 263 (2015) 435–443.
[21] J. Fan, H. Qin, S. Jiang, Mn-doped g-C3N4 composite to activate peroxymonosulfate [41] Z. Huang, H. Bao, Y. Yao, W. Lu, W. Chen, Novel green activation processes and
for acetaminophen degradation: the role of superoxide anion and singlet oxygen, mechanism of peroxymonosulfate based on supported cobalt phthalocyanine cata-
Chem. Eng. J. 359 (2019) 723–732. lyst, Appl. Catal. B 154–155 (2014) 36–43.
[22] C. Tan, N. Gao, D. Fu, J. Deng, L. Deng, Efficient degradation of paracetamol with [42] S. Indrawirawan, H. Sun, X. Duan, S. Wang, Nanocarbons in different structural
nanoscaled magnetic CoFe2O4 and MnFe2O4 as a heterogeneous catalyst of per- dimensions (0–3D) for phenol adsorption and metal-free catalytic oxidation, Appl.
oxymonosulfate, Sep. Purif. Technol. 175 (2017) 47–57. Catal. B 179 (2015) 352–362.
[23] Q. Yang, H. Choi, S.R. Al-Abed, D.D. Dionysiou, Iron–cobalt mixed oxide nanoca- [43] A. Khan, S. Zou, T. Wang, J. Ifthikar, A. Jawad, Z. Liao, A. Shahzad, A. Ngambia,
talysts: heterogeneous peroxymonosulfate activation, cobalt leaching, and ferro- Z. Chen, Facile synthesis of yolk shell Mn2O3@Mn5O8 as an effective catalyst for
magnetic properties for environmental applications, Appl. Catal. B 88 (2009) peroxymonosulfate activation, PCCP 20 (2018) 13909–13919.
462–469. [44] L. Peng, X. Gong, X. Wang, Z. Yang, Y. Liu, In situ growth of ZIF-67 on a nickel foam
[24] C. Cai, H. Zhang, X. Zhong, L. Hou, Ultrasound enhanced heterogeneous activation as a three-dimensional heterogeneous catalyst for peroxymonosulfate activation,
of peroxymonosulfate by a bimetallic Fe–Co/SBA-15 catalyst for the degradation of RSC Adv. 8 (2018) 26377–26382.
Orange II in water, J. Hazard. Mater. 283 (2015) 70–79. [45] D. Xia, R. Yin, J. Sun, T. An, G. Li, W. Wang, H. Zhao, P.K. Wong, Natural magnetic
[25] W. Guo, S. Su, C. Yi, Z. Ma, Degradation of antibiotics amoxicillin by Co3O4-cat- pyrrhotite as a high-Efficient persulfate activator for micropollutants degradation:
alyzed peroxymonosulfate system, Environ. Prog. Sustainable Energy 32 (2013) radicals identification and toxicity evaluation, J. Hazard. Mater. (2017).
193–197. [46] C. Tan, X. Jian, Y. Dong, X. Lu, X. Liu, H. Xiang, X. Cui, J. Deng, H. Gao, Activation
[26] X. Chen, J. Chen, X. Qiao, D. Wang, X. Cai, Performance of nano-Co3O4/perox- of peroxymonosulfate by a novel EGCE@Fe3O4 nanocomposite: free radical reac-
ymonosulfate system: kinetics and mechanism study using Acid Orange 7 as a tions and implication for the degradation of sulfadiazine, Chem. Eng. J. 359 (2019)
model compound, Appl. Catal. B 80 (2008) 116–121. 594–603.
[27] C. Wang, P. Shi, X. Cai, Q. Xu, X. Zhou, X. Zhou, D. Yang, J. Fan, Y. Min, H. Ge, [47] N. Villota, J.M. Lomas, L.M. Camarero, Study of the paracetamol degradation
W. Yao, Synergistic effect of Co3O4 nanoparticles and graphene as catalysts for pathway that generates color and turbidity in oxidized wastewaters by photo-
peroxymonosulfate-based orange II degradation with high oxidant utilization effi- Fenton technology, J. Photochem. Photobiol., A 329 (2016) 113–119.
ciency, J. Phys. Chem. C 120 (2016) 336–344. [48] Y. Chen, X. Zhang, L. Mao, Z. Yang, Dependence of kinetics and pathway of acet-
[28] Y. Ding, Y. Wang, L. Su, M. Bellagamba, H. Zhang, Y. Lei, Electrospun Co3O4 na- aminophen photocatalytic degradation on irradiation photon energy and TiO2
nofibers for sensitive and selective glucose detection, Biosens. Bioelectron. 26 crystalline, Chem. Eng. J. 330 (2017) 1091–1099.

1337