Mechanistic evaluation of ferrite AFe2O4 (A = Co, Ni, Cu, and Zn) catalytic performance in oxalic acid ozonation

Article abstract of DOI:10.1016/j.apcata.2017.08.025

Spinel ferrites AFe₂O₄ (A = Co, Ni, Cu, and Zn) synthesized by citrate sol-gel method were tested for their catalytic performance in ozonation of aqueous oxalic acid. The objectives of this work include the evaluation of catalytic activity and ferrites stability in mineralization of oxalic acid together with exploring the activation mechanisms of ferrite catalysts in catalytic ozonation process. The results indicated that CoFe₂O₄ was the most active in catalyzing oxalic acid mineralization, removing 68.3% of TOC within 120 min in the reaction conditions of initial pH 2.3, oxalic acid concentration 5 mM, inlet ozone concentration 14 ± 1 mg L^?1, catalyst dose 1.0 g L^?1. All catalysts exhibited satisfactory composition stability in a batch experiment, although losing up to 4.67% of the catalyst mass in Fe³⁺ leaching from ZnFe₂O₄. H₂-Temperature programmed reduction and cyclic voltammetry scan analysis revealed that these spinel ferrites had the potential reducibility and the electron donating ability to ozone molecule, respectively. Additionally, X-ray photoelectron spectroscopy showed that the surface metal ions’ valence state and the surface oxygen species played important roles in catalytic ozonation process. The reaction rate constant of oxalic acid mineralization were well fitted to non-linearly depend on variable metal valence and linearly depend on surface hydroxyl group density. The radical-based mechanistic reaction pathways involving A²⁺-A³⁺-A²⁺ cycle and positively charged surface hydroxyl groups were proposed.

Full text of DOI:10.1016/j.apcata.2017.08.025

Accepted Manuscript
Title: Mechanistic evaluation of ferrite AFe O (A = Co, Ni,
2
4
Cu, and Zn) catalytic performance in oxalic acid ozonation
Authors: Fengzhen Zhang, Chaohai Wei, Kaiyi Wu, Hongtao
Zhou, Yun Hu, Sergei Preis
PII:
S0926-860X(17)30405-2
DOI:
Reference:
http://dx.doi.org/10.1016/j.apcata.2017.08.025
APCATA 16380
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
7-6-2017
8-8-2017
18-8-2017
Please cite this article as: Fengzhen Zhang, Chaohai Wei, Kaiyi Wu, Hongtao
Zhou, Yun Hu, Sergei Preis, Mechanistic evaluation of ferrite AFe2O4 (A=Co, Ni,
Cu, and Zn) catalytic performance in oxalic acid ozonation, Applied Catalysis A,
Generalhttp://dx.doi.org/10.1016/j.apcata.2017.08.025
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Mechanistic evaluation of ferrite AFe
performance in oxalic acid ozonation
2
4
O (A = Co, Ni, Cu, and Zn) catalytic
Fengzhen Zhang, Chaohai Wei*, Kaiyi Wu, Hongtao Zhou, Yun Hu, Sergei Preis
School of Environment and Energy, South China University of Technology, Guangzhou
5
10006, PR China
*
Corresponding author.
E-mail address: cechwei@scut.edu.cn (C. Wei).
Graphical Abstract
1

Highlights


Cobalt ferrite shows the most reactive performance in catalytic ozonation of oxalic acid.
The surface metal ions with variable valent states and positively charged surface
hydroxyl groups are involved in the course of catalytic ozonation.
Linear and non-linear dependence of surface properties on the catalytic activities are
obtained.


The activation mechanism of spinel ferrites in catalytic ozonation has been proposed.
2

Abstract
Spinel ferrites AFe
2
4
O (A=Co, Ni, Cu, and Zn) synthesized by citrate sol-gel method
were tested for their catalytic performance in ozonation of aqueous oxalic acid. The
objectives of this work include the evaluation of catalytic activity and ferrites stability
in mineralization of oxalic acid together with exploring the activation mechanisms of
ferrite catalysts in catalytic ozonation process. The results indicated that CoFe
the most active in catalyzing oxalic acid mineralization, removing 68.3% of TOC within
20 min in the reaction conditions of initial pH 2.3, oxalic acid concentration 5 mM,
2
4
O was
1
1
1
inlet ozone concentration 14 ± 1 mg Lˉ , catalyst dose 1.0 g Lˉ . All catalysts exhibited
satisfactory composition stability in a batch experiment, although losing up to 4.67%
of the catalyst mass in Fe3+ leaching from ZnFe
O
4
. H -Temperature programmed
2
2
reduction and cyclic voltammetry scan analysis revealed that these spinel ferrites had
the potential reducibility and the electron donating ability to ozone molecule,
respectively. Additionally, X-ray photoelectron spectroscopy showed that the surface
metal ions’ valence state and the surface oxygen species played important roles in
catalytic ozonation process. The reaction rate constant of oxalic acid mineralization
were well fitted to non-linearly depend on variable metal valence and linearly depend
on surface hydroxyl group density. The radical-based mechanistic reaction pathways
2
+
3+
2+
involving A -A -A cycle and positively charged surface hydroxyl groups were
proposed.
Keywords: Spinel ferrite; Reducibility; Electron transfer; Variable valence; Surface
oxygen species
3

1
. Introduction
Municipal and industrial wastewaters contain toxic organic pollutants, trace organic
compounds and carcinogenic constituents recognized as serious environmental threat.
Advanced oxidation processes (AOPs), particularly ozone-based technologies receive
immense attention for their high reactivity in removing hazardous organic compounds
[
1-3]. Owing to its electrophilic character, ozone molecule selectively reacts with
unsaturated organic compounds on their electron-rich functional sites, resulting in
saturated organic by-products such as aldehydes, ketones and carboxylic acids [4, 5].
These products are more refractory to ozone reaction than their precursors. Oxalic acid
is one of the most commonly identified electron-deficient ozonation by-products with
1
1
low rate constant of reaction with molecular ozone (k < 0.04 Mˉ sˉ at pH > 5 [6]).
This makes oxalate accumulating in the course of non-catalytic ozonation of organic
compounds, acidifying the treated wastewater.
Catalytic ozonation, a radical-based AOP attracts attention in wastewater treatment
due to its potential in producing hydroxyl radical (•OH) out of decomposed ozone.
Heterogeneous catalysts including metal oxides [7], carbon-derived materials [8],
mesoporous siliceous materials [9], metal oxides supported on activated carbon [10],
and homogenous catalysts such as metal ions [11], hydrogen peroxide [12], and metal
ions coordinated with α-carbonyl or α-hydroxy carboxylate anions [13] have been
studied in assisting ozonation. The reaction mechanism and the catalytic performance
of metal oxides have been addressed recently [4]. Generally, the mechanism of
ozonation catalyzed by metal oxides consists of three aspects: (1) •OH generation from
4

the reaction of ozone with surface metal ion featured by its variable valence, (2) direct
elimination of organic compounds via reaction of ozone with intermediate metal-
organic complexes, and (3) •OH generation from the reaction of ozone with surface
hydroxyl groups [4]. A considerable selection of researches have focused on the
electron transfer process from a metal ion with variable valence to ozone. For example,
Hou et al. [14] found that •OH is generated via ozone decomposition promoted by
2
+
3+
electron directly supplied by the ≡Fe and ≡Fe species on magnetite surface. Also,
2
+
the coordination of surface Cu on CuO with organic ligand (oxalate) was confirmed
3+
2+
θ
to be responsible for the reduced redox potential of Cu /Cu couple (E =2.3 V), which
was favorable for electron extraction by ozone, thus facilitating its decomposition and
•
OH generation [15]. Sun et al. [16] indicated that manganese oxide supported on SBA-
5 showed excellent catalytic activity in ozonation of oxalic acid ascribed to the
enhanced •OH generation. The radical was stimulated by electron transferred from
multivalent MnO (Mn(III)/Mn(IV)) to ozone, and the physical interaction of
protonated surface hydroxyl groups with molecular ozone.
Concerning the catalysts’ stability, spinel ferrites (AFe
1
x
2
4
O ), where A represents a
transition metal ion, have been attracted attention exhibiting excellent catalytic activity
and structural stability. Ferrites were described being reduced to a cation excess
composite and re-oxidized to their initial state, while conserving the spinel structure
[
17, 18]. The catalytic activity of spinel ferrites in ozonation of various contaminants
has been reported. ZnFe produced by hydrothermal method exhibited higher
2
O
4
catalytic activity in phenol ozonation compared with the ferrite synthesized from citrate
5

sol-gel, which was attributed to larger contact surface area and higher surface hydroxyl
group density at the former [7]. Zhao et al. [19] used NiFe nanomaterial as a catalyst
2
O
4
for phenol ozonation, and demonstrated that not only the surface Lewis acid sites but
also interfacial electron transfer resulted in its high catalytic activity. Qi et al. [20]
reported mixed spinel ferrites (MnFe
activity in phenacetin ozonation. The nickel-catalyzed ozone decomposition was found
to be critical for •OH as well as H generation. Yin et al. [21] described a synergistic
effect of photocatalysis and ozonation catalyzed by bismuth ferrite (BiFeO
2
O
4
and NiFe
2
O
4
) showing notable catalytic
2
O
2
3
)
nanoparticles. It was suggested that the photo-generated electrons produced on BiFeO
3
could be trapped by ozone, thus initiating ozone decomposition and •OH generation.
The given references point out that the catalytic activity of spinel ferrites in ozonation
is closely related to the surface hydroxyl groups as well as the electron transfer capacity.
It should be emphasized that the contributions of surface hydroxyl groups and metal
ions with variable valent states to the performance of catalytic ozonation present the
basic considerations in designing effective catalysts. The authors, however, failed to
find published studies dealing with the activation mechanisms of series spinel ferrites
in catalytic ozonation. The correlations of referred surface properties to the performance
of catalytic ozonation have not yet been sufficiently considered.
The objective of the present research was characterization and testing of a series of
spinel ferrites in catalytic ozonation of oxalic acid. The series was comprised of
CoFe
2
O
4
, CuFe
2
O
4
, NiFe
2
O
4
, and ZnFe
2
4
O prepared using the citrate sol-gel method.
The crystal structure, texture and chemical forms of surface metallic species were
6

analyzed. The catalytic reaction mechanisms were studied using cyclic voltammetry
scans (CVs) and X-ray photoelectron spectroscopy (XPS) analysis. The relationship of
catalytic activity with surface hydroxyl groups density and electron transfer capacity
was quantitatively described for the spinel ferrites under consideration. The reaction
mechanism of oxalic acid ozonation catalyzed by ferrites was described in detail.
2
. Experimental section
2
.1. Reagents
Ferric nitrate nonahydrate, citric acid, oxalic acid monohydrate (Sinopharm Chemical
Reagent Co., Ltd.), zinc nitrate hexahydrate, cobalt nitrate hexahydrate (Guangzhou
Chemical Reagent Factory), nickel nitrate hexahydrate, copper nitrate trihydrate
(Tianjin Damao Chemical Reagent Plant) of analytic reagent grade were used as
received without further purification. Milli-Q water (18.2 MΩ cm in resistivity,
Millipore) was used for preparing all solutions.
2
.2. Catalyst preparation and characterization
The spinel ferrites were prepared by citrate sol-gel method [7]. The ferrite precursor
was obtained by mixing aqueous solutions of metal nitrates and citric acid at a molar
ratio of A(II):Fe(III):citric acid = 1:2:3. The resulting solution was continuously stirred
at 60 ºC for 2 h, and allowed to evaporate in water bath at 80 ºC to form a viscous gel.
The gel was dried at 80 ºC overnight and annealed at 550 ºC for 5 h with a temperature
1
growing rate of 2 ºC minˉ (zinc, cobalt and nickel ferrites), during which the material
spontaneously ignited and burned forming a powder material. The dried copper gel was
annealed at the similar temperature growing rate to 650 °C. The obtained solid catalysts
7

were denoted as CoFO, NiFO, CuFO and ZnFO for cobalt, nickel, copper and zinc
ferrite, respectively.
The X-ray diffraction (XRD) patterns of the catalysts were recorded by using a
Bruker D8-Advance X-ray diffractometer (Bruker, Germany). The surface functional
groups on the powders were analyzed by means of Raman spectroscopy (LabRAM
Aramis, H.J.Y, France) and Fourier transform infrared spectrometry (FT-IR, Nicolet-
6
700, Thermo, USA). The surface morphology of catalysts were obtained using
scanning electron microscopy (SEM, S-3700N, Hitachi, Japan). Brunauer-Emmett-
Teller (BET) surface area of samples were measured with the NOVA4200e specific
surface area analyzer (Quantachrome, USA). Prior to the measurement, the samples
were degassed at 150 ºC for 4 h. The valence states of elements on the catalyst surface
were recorded by X-ray photoelectron spectroscopy (XPS, Axis Ultra DLD, Kratos,
United Kingdom). The binding energy of C 1s was shifted to 284.6 eV as an internal
reference, while the conventional Al Kα (1486.6 eV) anode radiation was used as
excitation source. The pH at the point of zero charge (PZC) was measured by zeta
potential analyser (Zetasizer Nano ZS, Malvern, United Kingdom). The catalysts’
reducibility was evaluated by temperature-programmed reduction in hydrogen flow
(H
2
-TPR) on Auto Chem II Chemisorption Analyzer 2920 (Micromeritics, USA) and
the cyclic voltammetry (CV) scan using CHI660 potentiostat (Chenhua Instrument
Company, China). The procedures of H
supplementary material Text S1.
2
-TPR and CV analyses were described in the
2
.3. Non-catalytic and catalytic ozonation experiments
8

Semi-batch experiments were carried out in a 500 mL three-necked glass flask used
as an ozonation reactor. The reactor was submerged to a thermostatic bath at 25 °C.
Gaseous ozone was produced from pure oxygen with a laboratory ozone generator (CH-
ZTW, Guangzhou Chuanghuan Ozone Electric Appliance Co. Ltd., China). The
concentration of ozone in the ozonized gas was controlled by gas flow rate and current
in the discharge cell. Prior to ozonation experiment, ozone generator was allowed to
stabilize the output ozone concentration for 5 min. The excess ozone in off-gas at the
reactor’s outlet was trapped by 20% KI solution as a safety precaution. In experiments,
oxygen containing ozone in amount of 14 ± 1 mg Lˉ at the flow rate of 1.0 L minˉ1
was dispersed using a porous porcelain pellet at the bottom of reactor containing 100
mL of 5 mM oxalic acid solution with initial pH 2.30 ± 0.02. The reaction solution was
simultaneously stirred with a magnetic stirrer for better mass transfer. In catalytic
ozonation experiments, the catalyst powder was added in concentration of 1.0 g Lˉ1
having other reaction conditions identical to the non-catalytic ozonation. Samples for
analysis were taken at regular time intervals in amount of 1 mL, being purged with
nitrogen for 5 min to remove residual aqueous ozone. The suspension samples were
subjected to filtration through 0.45 µm Nylon membrane filter. The identical
experimental procedures were conducted for adsorption evaluation by using oxygen
rather than ozone.
1
2
.4. Analytical procedures
The decay of oxalic acid concentration was expressed by its total organic carbon
(TOC), analyzed with the TOC-analyzer (TOC-VCPH, Shimadzu, Japan). The hydroxyl-
9

group density on the catalyst surface was determined by complete deprotonating [22].
.05 g of the catalyst was suspended in 50 mL of 0.1 M NaOH solution, which was
shaken for 4 h at 25 ºC. The suspension was filtered through 0.45 µm Nylon membrane
and excess NaOH in the filtrate was titrated by 0.05 M HNO . The amount of surface
0
3
hydroxyl groups is obtained from the consumed NaOH quantity. The solution pH was
measured with PHS-3C pH meter (Shanghai Precision & Scientific Instrument Co. Ltd.,
China). Gaseous ozone concentration was obtained by iodometric titration method. The
leaching concentration of metal ions in aqueous solution was measured by the PinAAcle
9
00T atomic absorption spectrometer (Perkin Elmer, USA). Each experiment was run
in triplicate showing deviations from the mean value not exceeding 10%.
3
. Results and discussion
3
.1. Characterization of the spinel ferrites
The XRD patterns of spinel ferrites are displayed in Fig. 1. Characteristic Bragg
planes of (111), (220), (311), (222), (400), (422), (511), (440) and (533) at 2θ= 18.4º,
0.2º, 35.6º, 37.2º, 43.2º, 53.8º, 57.1º, 62.7º and 74.3º suggest typical spinel structures
of CoFO and ZnFO, being in good agreement with JCPDS: 22-1086 and JCPDS: 22-
012, respectively. As for NiFO, all featured planes are identical with those of ZnFO
3
1
except plane (440) at 2θ= 43.2º, indicating the presence of nickel ferrite in accordance
with JCPDS: 44-1485. The crystal pattern of CuFO shown in the space group I41/amd
(141) demonstrates three special diffraction peaks at 2θ of 30.1º, 35.9º, 62.3º, well
matching JCPDS: 34-0425 for copper ferrite [23]. No secondary peaks are detected in
the XRD patterns, ensuring the phase purity of spinel ferrites. The crystallite size and
1
0

lattice constant of ferrites are summarized in Table 1. The former is obtained from the
Debye-Scherrer equation [24] based on measurement of full width at half-maximum of
the (311) peak, while the latter is calculated from Eq. (1):
2
2
2 1/2
a = d(h + k + l ) ,
(1)
where d is the inter-planar spacing and h, k and l are the Miller indices corresponding
to respective planes [25]. The packing factor p is calculated using Eq. (2):
p = D/d ,
(2)
where D is crystallite size obtained at the corresponding plane [26]. One can see that
the average crystallite size of spinel ferrite varies from 25 to 40 nm, whereas the lattice
constant is in the range of 6.1286 to 8.4341 nm, indicating changes of ionic radius and
stress in the interior crystal structure with different A sites. Importantly, as shown in
Fig. 1, all spinel ferrites display a similar diffraction peak of crystal plane (111) at about
2
θ = 18º. The dissociation of water molecule takes place at this specific crystal face,
resulting in adsorbed hydroxyl groups at the solid-liquid interface [27], presenting the
active sites in catalytic ozonation.
Fig. 1
Table 1
The lattice structural information of spinel ferrites characterized by Raman spectra is
shown in Fig. 2a. In the spectra, characteristic Raman active modes are clearly
identified, indicating the presence of tetrahedral (A-site) and octahedral (Fe-site) lattice
11

structures for spinel ferrites [28]. Typically, NiFO and CuFO exhibit five Raman active
modes (A1g, E
g
, 3F2g) at similar Raman shift. For example, in CuFO spectra, modes at
1
5
50 and 708 cmˉ are related to tetrahedral (Cu-O) sublattice, while modes at 225, 290
1
and 482 cmˉ correspond to octahedral (Fe-O) sublattice, which evidently prove the
spinel ferrite structure. The Raman spectrum of CoFO (amplification in Fig. S1) also
consists of five modes at 307 (E
1g(1)), suggesting an inverse cubic spinel structure. Raman modes at A1g(2) and A1g(1)
reflect the symmetrical vibration of Co-O in tetrahedral site, whereas Raman modes at
, T2g(2), T2g(3) originate from anti-symmetrical vibration of Fe-O in octahedral site
g
), 470 (T2g(2)), 572 (T2g(3)), 620 (A1g(2)) and 695 cmˉ1
(A
E
g
[
29]. As for ZnFO in Fig. 2a, there are four first-order Raman active modes A1g, E ,
g
1
F
2g(1) and F2g(2) at Raman shift of 650, 491, 351 and 290 cmˉ , respectively. The A1g
mode is associated with Zn-O vibration at tetrahedral site, whilst the modes E , F2g(1)
g
1
and F2g(2) below 600 cmˉ are attributable to vibrations of Fe-O at octahedral site [30].
It has been reported that different Raman modes for these ferrites are mainly influenced
by crystallite size and lattice constant (Table 1) [30, 31], demonstrating the changes of
lattice structure with different A sites in accordance with XRD results.
Fig. 2
The spinel structure comprising tetrahedral and octahedral groups were analyzed by
FT-IR as displayed in Fig. 2b. Both high frequency band (υ
1
) ranging from 540 to 575
1
1
cmˉ and low frequency band (υ
2
) ranging from 409 to 431 cmˉ , which are
characteristic bands of spinel ferrite, are identified in FT-IR spectra. The spectrum of
typical inverse CoFO, for example, consists of two IR absorption bands, υ
at 575 cmˉ1
1
1
2

1
and υ at 410 cmˉ , which represent the stretching vibration of tetrahedral group Co-O
2
and octahedral group Fe-O, respectively [32]. On the other hand, the bands observed at
1
1
630 and 3450 cmˉ correspond to -OH stretching vibration, proving the presence of
adsorbed water on the surface. The FT-IR spectra of the other catalysts show similar
absorption band features, indicating the presence of spinel ferrite structure. The relative
intensity of υ
1
is, generally, greater than that of υ , probably resulting from the bond
2
length of tetrahedral group shorter than the one of octahedral group [33].
The surface morphologies of spinel ferrites obtained by SEM are shown in Fig. 3 and
Fig. S2. Fig. 3a and 3b display the morphology of CoFO, revealing a rough surface with
uniformly dispersed micro-particles on the relatively smooth plane surface. The other
three spinel ferrites present similar surface morphology (Fig. S2) closely related to the
citric acid spontaneous ignition observed at temperatures above 250 ºC releasing CO
2
and water vapor thus forming loose as well as uniform surface [7]. The specific surface
areas of the spinel ferrites obtained from BET analysis are given in Table S1. These
2
1
samples have comparable surface areas with ZnFO having the largest 61.7 m gˉ , while
2
1
CuFO has the smallest 43.3 m gˉ . Therefore, the surface morphologies and crystal
structure (crystal plane at 2θ = 18º) ensure a sufficient solid-liquid interface area and
consequently a number of active sites for catalytic ozonation.
Fig. 3
3
.2. Catalytic activity and stability of the spinel ferrites
The activity and stability of the spinel ferrites were experimentally evaluated in oxalic
acid mineralization. Fig 4a shows the non-catalytic and catalytic ozonation of oxalic
1
3

acid with ferrites. It can be seen that CoFO demonstrates the highest catalytic activity,
mineralizing more than 68.3% of oxalic acid within 120 min. Other ferrites show from
8
.5% (ZnFO) to 15% (CuFO and NiFO) of TOC removal, while non-catalytic ozonation
under identical conditions results in 4.7% TOC degradation. As for the adsorption of
oxalic acid on ferrites’ surface (data not shown), CoFO surface adsorption leads to
1
0.3% of TOC reduction, significantly surpassing that of NiFO (2.01%), CuFO (1.87%)
and ZnFO (0.93%). It should be emphasized that, in the acidic reaction solution (initial
pH=2.30 ± 0.02), oxalic acid molecule will dissociate in the form of monoacid oxalate
(
HC
of these ferrites is positive, due to pH of reaction solution is lower than pHPZC for CoFO
7.31), NiFO (7.27), CuFO (7.09) and ZnFO (7.13). Thus, the observed adsorption is a
2
O
4
ˉ) according to its dissociation constants, pK
a
=4.27, 1.27. The surface charge
(
result of the electrostatic force between positively charged ferrite’s surface and
negatively charged monoacid oxalate.
Evolution of pH in catalytic and non-catalytic ozonation is illustrated in Fig. 4b. The
solution pH increases from 2.32 to 2.63 in ozone reaction catalyzed by CoFO, while it
only slightly rises in ozonation catalyzed by NiFO and CuFO, and remains practically
constant in non-catalytic and ZnFO-catalyzed ozonation. The result is expectedly
consistent since mineralization of oxalic acid ends up with formation of carbon dioxide
and water [16]. The solution pH growth indicates the extent of reaction, matching the
catalytic activities of spinel ferrites in TOC reduction (Fig. 4a).
Fig. 4
1
4

The stability of ferrites is evaluated following the metal leaching, the crystallite
structure and the surface morphology. Aqueous concentrations of metal ions eluted
from suspended solution after 120 min of reaction are shown in Table 2. It can be found
3
+
-1
out that the concentration of released Fe is high reaching, for example, 46.74 mg L
2
+
2+
2+
2+
in ZnFO/ozonation, while the concentrations of released Co , Ni , Cu and Zn are
2
+
-1
relatively low, e.g., the concentration of Ni as low as 0.84 mg L is observed. The
metal leaching can also be expressed as a percentage of the leached metal mass relative
3
+
to the total catalyst mass also shown in Table 2. The maximum Fe leaching is, for
example, 4.67% of the total ZnFO ferrite, which may be explained by the dissolution
of metal oxide in strong acidic media of 5 mM oxalic acid solution. Some literatures
have also reported the high leached metal ions from heterogeneous catalytic ozonation
2
+
of oxalic acid [16]. Since aqueous Co ion is a well-known catalyst for oxalic acid
ozonation [34], it is not surprising that the metal ions leaching from the CoFO in present
study have a significant contribution to oxalic acid oxidation (Fig. S3), even surpassing
that of heterogeneous catalytic ozonation (limited by surface mass transfer compared
with homogeneous reaction). Nevertheless, there is still around 20% of oxalic acid
mineralized by heterogeneous reaction, making this work worthwhile in elucidating the
involvement of surface active sites in catalytic ozonation. It should be emphasized that
the higher catalytic activity of CoFO compared to other ferrite catalysts is due to the
heterogeneous reaction, as the oxalic acid removed by heterogeneous reaction (CoFO)
is more than the overall oxalic acid removal (including adsorption, homogeneous and
heterogeneous reaction) catalyzed by other ferrite catalysts. There is a noticeable
1
5

decrease in specific surface area of catalysts after the first batch experiment (Table S1),
which could be due to the aggregation of these ferrite particles in aqueous solution.
However, compared to the freshly prepared ferrites, the ferrites after first batch
experiment remained the initial crystallite structure (Fig. S4), the surface functional
groups (Fig. S5), the surface morphology (Fig. 3) and the surface element composition
(Fig. S2d). Therefore, cobalt ferrite could be used for oxalic acid degradation in
catalytic ozonation with excellent catalytic activity and crystallite structural stability,
although cobalt ferrite together with other three ferrites are subjected to remarkable
metal ion leaching during the reaction. These results indicate that the stability of these
ferrites requires improvement, efforts that can be potentially reached via supporting
metal oxides on molecular sieves or conducting experiments in phosphate buffer
solution (5 mM, pH=7), being subjects of future work.
Table 2
3
.3. Isothermal H
2
reduction and electron-transfer properties of spinel ferrites
The catalytic activities of spinel ferrites with different A site metals are ascribed to
their redox properties, determining the electron transfer interaction with ozone, and thus
influencing OH-radical generation. In order to assess the reductive capacity and the
electron transfer activity of ferrites, the H
2
-TPR and CV analyses were carried out. H -
2
TPR profiles of ferrites are illustrated in Fig. 5 with well-defined reduction steps clearly
identified. It is obvious that CuFO and NiFO yield three peaks, while CoFO and ZnFO
produce only two peaks in the same temperature range. The CoFO shows two shoulder
peaks at around 500 and 630 ºC, which is different from the reduction steps based on
1
6

previous publications [35-37]. In this study, CoFO starts to be reduced at 500 ºC,
partially producing Fe and zero-valent Co. With increasing temperature, the
3
O
4
trivalent iron is reduced to bivalent iron, followed by the reduction of FeO to zero-
valent Fe at 630 ºC. The reduction of NiFO experiences nearly the same pathway as
CoFO: initially, bivalent nickel is reduced to metallic Ni at 265 ºC, having trivalent iron
partly reduced to bivalent state forming Fe
further reduced to FeO at 520 ºC, and then to metallic iron at 635 ºC [38, 39]. As for
the reduction of CuFO, metallic Cu and Fe are produced at 250 ºC with trivalent
iron in Fe subsequently reduced to bivalent state in the form of Fe at 350 ºC. Full
reduction of Fe to metallic Fe takes place at 575 ºC [35]. The reduction behavior of
ZnFO in the present study differs from the described elsewhere [40, 41]: the peak at
30 ºC marks the production of metallic Zn and Fe resulting from ferrite reduction.
3
O
4
. With growing temperature, Fe
3
O
4
is
2
O
3
2
O
3
3
O
4
3
O
4
3
2
O
3
Iron oxide is further reduced to metallic Fe at 573 ºC. Evidently, the distinct reduction
peaks identified in Fig. 5 demonstrate the comparable reduction capacity for these
ferrites. Concerning catalytic ozonation, in which ozone is decomposed and radical
species is generated, these spinel ferrites having the reduction capacity are prone to
provide electron to ozone molecule, accelerating its decomposition and OH-radical
generation [4], thus owning the catalytic activity in oxalic acid mineralization.
Fig. 5
The CV analysis of spinel ferrites was performed in order to characterize the electron
transfer potential (i.e., the redox properties) on solid-liquid interface. Fig. S6 shows the
qualitative CV scans of ferrites in 0.1 M Na
2
SO aqueous solution. Nearly no obvious
4
1
7

current is observed for ZnFO, while the current signals can be observed in the ascending
order of CuFO, NiFO and CoFO. This result suggests that CoFO owns the most
prominent electron transfer ability, followed by NiFO and CuFO, which is in
accordance with their catalytic activities in oxalic acid mineralization. The CV analysis
shows that the highest electron transfer capability of CoFO catalyst provides prompt
redox process on its solid-liquid interface in catalytic ozonation, being consistent with
previously published results [19, 35]. It should be emphasized that both reduction
capacity and interfacial electron transfer ability are essential properties in providing
electron to ozone molecule, which could ensure improving catalytic activity in ozone
reaction.
3
3
.4. Activation mechanism of ferrite in catalytic ozonation
.4.1. Evaluation of active surface elements
In a catalytic ozonation process, the surface metal ions and charged surface hydroxyl
groups on the solid-liquid interface are critical reactive sites, on which ozone is
decomposed. The Raman spectra of freshly synthesized and used spinel ferrites are
shown in Fig. S7. There have been observable changes in Raman shift for CoFO, NiFO
and CuFO, whereas for ZnFO this remains nearly the same. The change in Raman shift
could be due to the cation variation and its redistribution in the ferrite [33]. This
phenomenon indicates that there may be certain component changes, i.e., cation valence
state on ferrite surface might change during the course of catalytic ozonation.
The variation of A-metal valence before and after reaction was explored by means of
XPS as illustrated in Fig. 6 and Table 3. Fig. 6a demonstrates that the freshly
1
8

synthesized CoFO has two measurable Co 2p3/2 peaks, with binding energy locating at
2
+
7
78.2 and 782.4 eV [42], which are ascribed to Co ion in octahedral and tetrahedral
sites, respectively. No Co3+ is detected in the freshly synthesized CoFO. The
2
+
contributions of octahedral and tetrahedral Co relative to total Co intensity are of
4
4.8% and 55.2%, respectively. After catalytic ozonation, however, the Co 2p3/2 peaks
are composed of three distinct peaks appearing at 777.5, 780.7 and 784.1 eV,
contributing to overall Co intensity of 40.3%, 30.0% and 29.7%, respectively (Fig. 6a).
Importantly, the peak at 784.1 eV is corresponded to Co3+ [35], i.e., Co is solely
present before ozonation reaction, changing to trivalent state after catalytic ozonation.
This could be resulted from the oxidation and reduction of Co ion along with the process
of catalytic ozonation. Oxidation of bivalent Co is carried out along ozone
2+
2+
decomposition, driven by the electron transferred from Co to molecular ozone, thus
producing Co3+ and O
•ˉ, the intermediate in •OH generation [4]. In order to keep
charge balance on ferrite surface, Co accepts electrons from catalyst surface [35]. The
3
3
+
2
+
3+
2+
redox process thus consisted of Co -Co -Co transformation. Similarly for NiFO
Fig. 6b) and CuFO (Fig. 6c), the intensity of trivalent state Ni and Cu relative to the
overall intensity increase after ozonation reaction, although trivalent state Ni and Cu
(
2
+
3+
2+
are noticed before reaction. The cycle of A -A -A , where A represents Co, Ni and
Cu, takes place with the electron transfer between A-metal and ozone is involved in
catalytic ozonation. By contrast, the intensity of trivalent state Zn (Fig. 6d) relative to
the overall intensity has remained practically unchanged after exposure to catalytic
ozonation. This could be due to the inferior surface electron transfer ability of ZnFO as
1
9

revealed in Fig. S6. Noticeably, the extent of trivalent state increment for the A-metal
is consistent with their catalytic activities in the mineralization of oxalic acid. Similarly,
the valence state changes of Fe in these ferrites were examined by XPS study (Fig. S8),
indicating that the bivalent and trivalent state Fe coexist in ferrites. It can be revealed
3
+
that the ratio of Fe intensity relative to total Fe intensity increases after first batch
2
+
3+
2+
catalytic ozonation. This may demonstrate the participation of Fe -Fe -Fe circle in
catalytic ozonation process, as Fe2+ in ferrite can reduce both ozone molecule and
trivalent A-metal ion [43].
Fig. 6
Table 3
The high-resolution O 1s spectra of spinel ferrites before and after catalytic ozonation
displaying the evolution of oxygen containing species are shown in Fig. S9. For the
freshly synthesized CoFO catalyst, the spectra can be resolved into two individual peaks
2
at 529.7 and 531.4 eV, which are ascribed to surface lattice oxygen O ˉ denoted as O
A
,
and adsorbed oxygen or surface hydroxyl species denoted as O [44], respectively.
B
These two kinds of surface oxygen species have been reported as active sites involved
in catalytic ozonation process [45]. After batch treatment, the two peaks area ratios
change slightly with O
9.9%. This result demonstrates both oxygen species participating in catalytic
ozonation. The growth of O content is attributed to the formation of surface hydroxyl
groups or O adsorbed on catalyst surface [35]. The decline of O which is oxidized to
[35] may be due to the lattice oxygen oxidation by Co (or Fe ) concomitant with
A
reduced from 81.3 to 80.1%, while O
B
increased from 18.7 to
1
B
2
A
3
+
3+
O
2
2
0

2+
3+
its reduction to Co as described previously. In other words, Co ion produced by
2
+
ozone oxidation is reduced to its bivalent state by O , establishing the cycle of Co -
A
3
+
2+
Co -Co . Similar results was observed from XPS analysis for CuFO and ZnFO (Fig.
S9c and S9d). The oxygen species evolution in NiFO (Fig. S9b), however, shows a
significant reverse trend by contrast to that of CoFO, CuFO and ZnFO: the lattice
oxygen O
A
increase from 24.1 to 82.7%, whereas the adsorbed oxygen O decrease
B
from 38.8 to 17.3%. The reason for this irregular trend of oxygen species evolution is
unclear so far and needs further investigations. The detailed variation of oxygen species
with evolution of oxidizing species including O
3
, •OH and O •ˉ in aqueous solution has
2
to be studied. All known oxygen species should be summarized as the total surface
oxygen concentration, and the correlation between •OH exposure and total surface
oxygen concentration needs to be established [8].
3
.4.2. Correlation of surface properties with reactivity
Since metal ions with variable valence and surface oxygen species determine the
catalytic activity of spinel ferrites in ozone reaction, establishing the relationship of
these variables with mineralization efficiency of oxalic acid makes sense in
understanding the reactivity of catalysts. Herein, the valence variation of metal ions
after catalytic ozonation is quantitatively evaluated by the increment of valence state,
defined by the difference of the percentage of trivalent state after reaction relative to
that before reaction. The surface oxygen species are representatively expressed by the
density of surface hydroxyl groups. The oxidation rate of oxalic acid is described by its
1
apparent reaction rate constant k (minˉ ). The correlations between catalysts’ surface
2
1

properties and reaction rate coefficient are shown in Fig. 7. Apparently, the increased
content of oxidized form of A-metal and the surface hydroxyl group density are
positively correlated with reaction rate constant. It is observable that k increases non-
linearly with the increment of valence state, complying with Eq. (1). By contrast, the k
rises linearly with increasing surface hydroxyl group density, abiding by Eq. (2). These
results assume the great dependence of oxalic acid mineralization on these two surface
properties. Basically, this dependence is fundamentally explained by catalytic ozone
decomposition, proceeding via direct electron transfer from low valent metal ion to
molecular ozone, as well as physical interaction of positively charged surface hydroxyl
group with ozone. It is believed that the metal ions serving as Lewis acidic sites
combine with H O molecule acting as Lewis base, thus resulting in the generation of
2
surface hydroxyl groups [4, 35]. Therefore, the surface metal ion with variable valence
is the essential controlling factor, which should be taken advantage of in designing and
fabricating highly reactive catalysts for ozonation.
2
y = 47.42 + 3.71ln(x - 0.0008175)
y = 0.31 + 1010.09x
R = 0.9999
(1)
(2)
2
R = 0.9949
Fig. 7
.4.3. Catalytic mechanism of ferrites in catalytic ozonation
The radical-based mechanism of catalytic ozonation is established by the interaction
of ozone with surface active sites, on which ozone decomposition is promoted [16, 46,
7]. The authors’ previous studies also suggest the generation of •OH in phenol
3
4
ozonation catalyzed by zinc ferrite [7]. There are publications, however, reporting the
radical-free mechanism of oxalic acid mineralization involving formation of surface
2
2

metal-oxalate complexes [48-50]. The radical-free mechanism has been proposed due
to considerable adsorption of oxalic acid on catalysts’ surfaces. The present work
assumes both radical-based mechanism according to the experimental results and
radical-free mechanism on the basis of literature reviews.
The overall heterogeneous mechanism of oxalic acid ozonation catalyzed by spinel
ferrites is illustrated in Scheme 1, although homogeneous reactions make significant
contribution to oxalic acid removal falling out of the aim of this study. Generally, the
surface hydroxyl groups are produced as a result of interaction between A site metal
ion and water molecule [4]. In this work, the pH of oxalic acid (pH=2.3) is basically
lower than pHPZC of these four spinel ferrites, resulting in positively charged surface
hydroxyl groups. Specifically, the intermediate HO • radical is produced by the reaction
3
of ozone with positively charged hydroxyl group driven by electrostatic force and
hydrogen bond [51] resulting in the formation of •OH. Alternatively, an electron will
be transferred from bivalent A site metal ion to ozone, therefore producing O •ˉ and
3
+
trivalent A site metal ion. The intermediate O
producing HO
3
•ˉ reacts with H in the solution,
3
• and, ultimately, •OH [4]. At the same time, trivalent A site metal ion
is reduced to its bivalent state by the lattice oxygen. Finally, the •OH produced by two
described pathways oxidizes oxalic acid. It should be emphasized that the coexistence
of iron ion with multiple valences and A-metal ion may accelerate the process of ozone
decomposition.
On the other hand, oxalic acid dissociates in the form of monoacid oxalate (HC
2
4
O ˉ)
in pH 2.3, coordinating with bivalent A-metal ion, forming initial complex
2
3

+
≡A(II)(HC
2
O
4
)
2
. It will be then subjected to oxidation into ≡A(III)(HC
2
O
4
)
2
by
adsorbed ozone (Eq. (3)) [34, 52, 53]. The intermediate A(III) complex is highly
unstable, undergoing fast dissociation into superoxide ion radical and carbon dioxide
via a single intra-molecular electron transfer (Eq. (4)) [49]. The ≡A(II)(HC
then combine one monoacid oxalate to form ≡A(II)(HC (Eq. (5)). It needs to be
emphasized that, the formation of O ˉ in Eq. (3) is an essential intermediate responsible
2
O
)+ will
4
2
O )
4 2
3
for •OH generation in acidic medium [4]. Therefore, the radical-based and radical-free
mechanistic pathways may occur concomitantly for oxalic acid ozonation. This
hypothesis needs further verifications of surface complex characterization and hydroxyl
radical scavenging test, which will be elucidated in our future work.
Scheme 1
+
≡
A(II)(HC
A(III)(HC
A(II)(HC
2
O
4
)
2
+ O
3
→ ≡A(III)(HC
→ ≡A(II)(HC
)+ + HC
ˉ →≡A(II)(HC
2
O
4
)
2
+ O
3
ˉ
(3)
(4)
(5)
+
)+ + 2CO
+ H+
≡
2
O
4
)
2
2
O
4
2
≡
2
O
4
2
O
4
2
O
4
)
2
4
. Conclusions
Ozonation of oxalic acid catalyzed by spinel ferrites was conducted in order to
elucidate the activation mechanism of ferrites in catalytic ozonation. The physico-
chemical characterizations of CoFO, NiFO, CuFO and ZnFO show similar structural
features and surface morphology having, however, different contact surface area. The
cobalt-containing catalyst demonstrates superior catalytic performance in oxalic acid
mineralization, followed by the catalytic activity sequence of NiFO, CuFO and ZnFO.
After a batch catalytic ozonation, the ferrites preserve their structure, although certain
2
4

amount of metal ions leach to the solution. The CV evaluations suggest that CoFO is
the most active catalyst in electron transfer, the essential property responsible for
catalytic ozone decomposition. It is found out that surface metal ions (including A-site
and Fe) with variable valence and surface oxygen species participate in catalytic
ozonation as revealed by XPS studies. The radical-based mechanistic pathways
including the interaction of ozone with surface hydroxyl groups and surface metal ions
have been proposed.
Acknowledgments
The authors gratefully acknowledge the financial support from the National Natural
Science Foundation of China (No. 51278199), the Research Project of Production,
Education and Research of Guangdong Province, China (No. 2012B091100450) and
the Special Project of Applicable Scientific Research of Guangdong Province, China
(No. 2015B020235005).
References
[
1] G. Knopp, C. Prasse, T.A. Ternes, P. Cornel, Water Res. 100 (2015) 580-592.
[
2] F. Soltermann, C. Abegglen, C. Gꢀtz, U. von Gunten, Environ. Sci. Technol. 50 (2016)
9825-9834.
[
[
[
[
[
3] J. Chen, S.H. Tian, J. Liu, Y. Xiong, Appl. Catal. A Gen. 506 (2015) 118-125.
4] C.H. Wei, F.Z. Zhang, Y. Hu, C.H. Feng, H.Z. Wu, Rev. Chem. Eng. 33 (2016) 49-90.
5] P.C.C. Faria, J.J.M. Órfão, M.F.R. Pereira, Appl. Catal. B Environ. 79 (2008) 237-243.
6] J. Hoigné, H. Bader, Water Res. 17 (1983) 185-194.
7] F.Z. Zhang, C.H. Wei, Y. Hu, H.Z. Wu, Sep. Purif. Technol. 156 (2015) 625-635.
2
5

[
8] R. Oulton, J.P. Haase, S. Kaalberg, C.T. Redmond, M.J. Nalbandian, D.M. Cwiertny,
Environ. Sci. Technol. 49 (2015) 3687-3697.
[
[
9] W.W. Li, X.W. Lu, K. Xu, J.H. Qu, Z.M. Qiang, Water Res. 86 (2015) 2-8.
10] S.F. Tang, D.L. Yuan, Q. Zhang, Y.M. Liu, Q. Zhang, Z.Q. Liu, H.M. Huang, Environ.
Sci. Pollut. Res. 23 (2016) 18800-18808.
[
[
11] M.L. Wilde, S. Montipó, A.F. Martins, Water Res. 48 (2014) 280-295.
12] E.D. Torres-Socías, I. Fernández-Calderero, I. Oller, M.J. Trinidad-Lozano, F.J. Yuste, S.
Malato, Chem. Eng. J. 234 (2013) 232-239.
13] F.J. Beltrán, F.J. Rivas, R. Montero-de-Espinosa, Ind. Eng. Chem. Res. 42 (2003) 3210-
217.
[
3
[
[
[
14] L.W. Hou, H. Zhang, L.G. Wang, L. Chen, Chem. Eng. J. 229 (2013) 577-584.
15] T. Zhang, J.P. Croué, Appl. Catal. B Environ. 144 (2014) 831-839.
16] Q.Q. Sun, L.S. Li, H.H. Yan, X.T. Hong, K.S. Hui, Z.Q. Pan, Chem. Eng. J. 242 (2014)
348-356.
[
[
[
17] Y. Tamaura, M. Tahata, Nature 346 (1990) 255-256.
18] T. Zhang, H.B. Zhu, J.P. Croue, Environ. Sci. Technol. 47 (2013) 2784-2791.
19] H. Zhao, Y.M. Dong, G.L. Wang, P.P. Jiang, J.J. Zhang, L.N. Wu, K. Li, Chem. Eng. J.
219 (2013) 295-302.
[
[
20] F. Qi, B.B. Xu, W. Chu, J. Mol. Catal. A Chem. 396 (2015) 164-173.
21] J. Yin, G.Z. Liao, J.L. Zhou, C.M. Huang, Y. Ling, P. Lu, L.S. Li, Sep. Purif. Technol.
1
68 (2016) 134-140.
22] H. Tamura, A. Tanaka, K.Y. Mita, R. Furuichi, J. Colloid Interface Sci. 209 (1999) 225-
31.
23] P. Laokul, V. Amornkitbamrung, S. Seraphin, S. Maensiri, Curr. Appl. Phys. 11 (2011)
01-108.
24] R. Sharma, P. Thakur, M. Kumar, N. Thakur, N.S. Negi, P. Sharma, V. Sharma, J. Alloy.
Compd. 684 (2016) 569-581.
[
2
[
1
[
2
6

[
25] P.B. Belavi, G.N. Chavan, L.R. Naik, R. Somashekar, R.K. Kotnala, Mater. Chem. Phys.
32 (2012) 138-144.
26] T. Kavetskyy, O. Shpotyuk, M. Popescu, A. Lorinczi, F. Sava, Adv. Mater. 9 (2007) 3079-
081.
1
[
3
[
[
[
[
27] Y. Joseph, W. Ranke, W. Weiss, J. Phys. Chem. B 104 (2000) 3224-3236.
28] D. Harikishore Kumar Reddya, Y.S. Yuna, Coord. Chem. Rev. 315 (2016) 90-111.
29] D. Varshney, K. Verma, A. Kumar, J. Mol. Struct. 1006 (2011) 447-452.
30] Q.Y. Xiang, D. Wu, Y. Bai, K. Yan, W.Q. Yao, L. Zhang, J. Zhang, J.L. Cao, Ceram. Int.
42 (2016) 16882-16887.
[
(
[
31] J.P. Singh, G. Dixit, R.C. Srivastava, H.M. Agrawal, R. Kumar, J. Alloy. Compd. 551
2013) 370-375.
32] T. Dippong, O. Cadar, E.A. Levei, I. Bibicu, L. Diamandescu, C. Leostean, M. Lazar, G.
Borodi, L.B. Tudoran, Ceram. Int. 43 (2017) 2113-2122.
33] R.S. Yadav, J. Havlica, J. Masilko, L. Kalina, J. Wasserbauer, M. Hajdúchová, V. Enev,
I. Kuřitka, Z. Kožáková, J. Magn. Magn. Mater. 399 (2016) 109-117.
[
[
[
34] D.S. Pines, D.A. Reckhow, Environ. Sci. Technol. 36 (2002) 4046-4051.
35] Y.M. Ren, L.Q. Lin, J. Ma, J. Yang, J. Feng, Z.J. Fan, Appl. Catal. B Environ. 165 (2015)
5
72-578.
36] R. Dumitru, F. Papa, I. Balint, D.C. Culita, C. Munteanu, N. Stanica, A. Ianculescu, L.
Diamandescu, O. Carp, Appl. Catal. A Gen. 467 (2013) 178-186.
37] C.A. Chagas, E.F.de Souza, M.C.N.A. de Carvalho, R.L. Martins, M. Schmal, Appl. Catal.
A Gen. 519 (2016) 139-145.
38] R. Benrabaa, H. Boukhlouf, A. Löfberg, A. Rubbens, R.N. Vannier, E. Bordes-Richard,
A. Barama, J. Nat. Gas Chem. 21 (2012) 595-604.
39] R. Benrabaa, H. Boukhlouf, E. Bordes-Richard, R.N. Vannier, A. Barama, Stud. Surf. Sci.
Catal. 175 (2010) 301-304.
40] M.A. Valenzuela, P. Bosch, J. Jiménez-Becerrill, O. Quiroz, A.I. Páez, J. Photochem.
Photobiol. A Chem. 148 (2002) 177-182.
[
[
[
[
[
2
7

[
[
41] K. Winiarska, I. Szczygieł, R. Klimkiewicz, Ind. Eng. Chem. Res. 52 (2013) 353-361.
42] Z.P. Zhou, Y. Zhang, Z.Y. Wang, W. Wei, W.F. Tang, J. Shi, R. Xiong, Appl. Surf. Sci.
254 (2008) 6972-6975.
[
[
43] A.H. Lv, C. Hu, Y.L. Nie, J.H. Qu, Appl. Catal. B Environ. 117-118 (2012) 246-252.
44] I.M. Ismail, B. Abdallah, M. Abou-Kharroub, O. Mrad, Nucl. Instrum. Methods Phys. Res.
Sect. B Beam Interact. Mater. Atoms 271 (2012) 102-106.
45] A. Galtayries, R. Sporken, J. Riga, G. Blanchard, R. Caudano, J. Electron. Spectrosc. Relat.
Phenom. 88-91 (1998) 951-956.
46] F.J. Beltrán, F.J. Rivas, R. Montero-de-Espinosa, Appl. Catal. B Environ. 39 (2002) 221-
31.
[
[
2
[
[
[
[
[
[
[
47] M.H. Sui, L. Sheng, K.X. Lu, F. Tian, Appl. Catal. B Environ. 96 (2010) 94-100.
48] F.J. Beltrán, F.J. Rivas, R. Montero-de-Espinosa, Water Res. 39 (2005) 3553-3564.
49] T. Zhang, W.W. Li, J.P. Croué, Appl. Catal. B Environ. 121-122 (2012) 88-94.
50] T. Zhang, J.P. Croué, Appl. Catal. B Environ. 144 (2014) 831-839.
51] L. Zhao, Z.Z. Sun, J. Ma, Environ. Sci. Technol. 43 (2009) 4157-4163.
52] A.E. Martell, Pure Appl. Chem. 17 (1968) 129-178.
53] M. Orban, React. Kinet. Catal. Lett. 42 (1990) 343-353.
2
8

Figure Captions
Fig. 1. XRD patterns of spinel ferrites.
Fig. 2. Raman (a) and FT-IR (b) spectra of spinel ferrites.
Fig. 3. SEM micrographs of CoFO before (a) and after (c) first batch ozonation
reaction. Freshly synthesized catalyst is shown at higher magnification (b).
Fig. 4. TOC removal of oxalic acid (a) and evolution of solution pH (b) by non-catalytic
and catalytic ozonation with spinel ferrites.
Fig. 5. H -TPR spectra of spinel ferrites.
2
Fig. 6. XPS spectra of Co (a), Ni (b), Cu (c), and Zn (d) ferrites before and after first
batch catalytic ozonation.
Fig. 7. Correlation of ferrites’ catalytic reactivity with surface properties.
Scheme Captions
Scheme 1 Catalytic mechanism of oxalic acid ozonation catalyzed by spinel ferrites.
2
9