The role of cobalt oxide or magnesium oxide in ozonation of ammonia nitrogen in water

Article abstract of DOI:10.1016/j.apcata.2020.117515

In this study, the reaction mechanisms for ozonation of ammonia nitrogen in the presence of Co₃O₄ or MgO were investigated. For the reaction over Co₃O₄, Cl^– in the reaction solution was indispensable and ClO^– was formed by a non-catalytic oxidation of Cl^–. Co₃O₄ promoted the reaction of NH₄⁺ with ClO^– to give the products including NO₃^–, chloramines and gaseous products. In contrast, Cl^– was unnecessary for the reaction with MgO. pH of the reaction solution was maintained at around 9 throughout the reaction owing to partial dissolution of MgO. Ammonia nitrogen was decomposed to mainly NO₃^– by non-catalytic radical reaction involving OH·, which was formed by the reaction of OH^– with O₃ in weakly basic solution. To keep the reaction solution weakly basic, H⁺ formed with the decomposition of NH₄⁺ was neutralized. As a result, about the same amount of Mg²⁺ as that of decomposed ammonia nitrogen was dissolved.

Full text of DOI:10.1016/j.apcata.2020.117515

Journal Pre-proof
The role of cobalt oxide or magnesium oxide in ozonation of ammonia
nitrogen in water
Philip Anggo Krisbiantoro (Investigation) (Writing - original draft)
(Visualization), Tomokazu Togawa (Validation) (Investigation), Lina
Mahardiani (Investigation), Haruka Aihara (Investigation), Ryoichi
Otomo<ce:contributor-role>Writing – review and editing), Yuichi
Kamiya (Conceptualization) (Writing - review and editing)
(Supervision)
PII:
S0926-860X(20)30108-3
DOI:
https://doi.org/10.1016/j.apcata.2020.117515
APCATA 117515
Reference:
To appear in:
Applied Catalysis A, General
Received Date:
Revised Date:
Accepted Date:
23 October 2019
3 March 2020
8 March 2020
Please cite this article as: Anggo Krisbiantoro P, Togawa T, Mahardiani L, Aihara H, Otomo R,
Kamiya Y, The role of cobalt oxide or magnesium oxide in ozonation of ammonia nitrogen in
water, Applied Catalysis A, General (2020), doi: https://doi.org/10.1016/j.apcata.2020.117515

This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. 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.
©
2020 Published by Elsevier.

Revised Manuscript, APCATA-D-19-02066
The role of cobalt oxide or magnesium oxide in ozonation of ammonia nitrogen in water
a
a
b
a
Philip Anggo Krisbiantoro , Tomokazu Togawa , Lina Mahardiani , Haruka Aihara , Ryoichi
c
c,
Otomo , Yuichi Kamiya *
a
Graduate School of Environmental Science, Hokkaido University, Nishi 5, Kita 10, Kita-ku
Sapporo 060-0810, Japan
b
Department of Chemistry Education, Faculty of Teacher Training and Education, Sebelas
Maret University, Jl. Ir. Sutami 36A Surakarta, Central Java 57126, Indonesia
c
Faculty of Environmental Earth Science, Hokkaido University, Nishi 5, Kita 10, Kita-ku
Sapporo 060-0810, Japan
*
Corresponding author: Prof. Yuichi Kamiya
Email: kamiya@ees.hokudai.ac.jp
TEL: +81 11 706 2217
Graphical abstract
1

Highlight

Co O and MgO have different roles for ozonation of ammonia nitrogen in water.
3
4
–
Chloramines are formed over Co O in addition to NO and gaseous nitrogen

3
4
3
products.
–
–



ClO formed by the reaction of Cl with O oxidizes ammonia nitrogen over Co O .
3
3
4
Non-catalytic radical oxidation occurs in the presence of MgO.
–
MgO is not a catalyst, but a reactant to produce OH by dissolution
Abstract
In this study, the reaction mechanisms for ozonation of ammonia nitrogen in the presence of
–
Co
3
O
4
or MgO were investigated. For the reaction over Co
3
O
4
, Cl in the reaction solution
2

–
–
was indispensable and ClO was formed by a non-catalytic oxidation of Cl . Co
3
4
O promoted
the reaction of NH +
with ClO to give the products including NO
–
–, chloramines and gaseous
4
3
–
products. In contrast, Cl was unnecessary for the reaction with MgO. pH of the reaction
solution was maintained at around 9 throughout the reaction owing to partial dissolution of
MgO. Ammonia nitrogen was decomposed to mainly NO –
by non-catalytic radical reaction
3
–
involving OH·, which was formed by the reaction of OH with O
3
in weakly basic solution.
+
+
To keep the reaction solution weakly basic, H formed with the decomposition of NH
4
was
2
+
neutralized. As a result, about the same amount of Mg as that of decomposed ammonia
nitrogen was dissolved.
Keywords: ozonation; ammonia nitrogen; cobalt oxide; magnesium oxide; wastewater
treatment
3

1
. Introduction
While ammonia (NH
) and ammonium ion (NH +
), which are collectively called
3
4
ammonia nitrogen, are indispensable nutrient for plants, excess ammonia nitrogen in water
often causes eutrophication of rivers, lakes, and inland seas. In addition, ammonia nitrogen is
–
1
toxic to fish species even if the concentration in water is less than 10 mg L [1]. High
concentration ammonia nitrogen in natural environment comes from inadequate treatments of
industrial wastewater, household sewage, and livestock manure and runoff water from
farmland with excess fertilization. Nitrogen-containing organic matters, such as urea, amino
acids, polypeptides and proteins can be a source for ammonia nitrogen in wastewater with
microbial action [2]. For environmental protection, ammonia nitrogen in environmental water
–
1
should be below 0.5 mg L [3,4]. To achieve this criteria, highly efficient and feasible
processes for the purification of water containing ammonia nitrogen are absolutely necessary.
So far, many treatment processes have been studied and some of them including
adsorption [5], ion exchange [6], activated sludge [7], stripping [8] and breakpoint
chlorination [2] are commercially used for the treatment of wastewater containing ammonia
nitrogen. However, each process has some limitations and disadvantages from technical and
economic standpoints. For example, because ion exchange is the processes only removing
ammonia nitrogen, waste brine containing high concentration ammonia nitrogen is generated,
which should be further treated. Activated sludge process consumes a lot of energy and
oxygen, and burdensome treatment of produced generated sludge is necessary [7]. Ammonia
stripping technique requires a large stripping tower, strict control of CaCO
ineffective in winter due to the low volatility of NH under low temperature environment [8].
In addition, it potentially causes air pollution if the stripped NH is not appropriately treated.
3
dose, and is
3
3
4

Catalytic wet air oxidation (CWAO) is a process to oxidatively decompose
contaminants in water by reaction with air (O ) in the presence of a catalyst. CWAO is
2
considered to be applicable for purification of wastewater containing ammonia nitrogen. So
far, several precious metals supported on carriers like metal oxides have been investigated for
the reaction [9-14]. Taguchi and Okuhara reported that Pt/TiO
2
[9] and Pd/ZrO
2
[10]
effectively and selectively promoted the oxidation of ammonia nitrogen into N
2
at relatively
low temperature (433 K) and low pressure (0.5 MPa) in mild acidic solution (pH 5.5). These
previous studies showed high potential of CWAO for the purification of wastewater
containing ammonia nitrogen. However, harsh reaction conditions are still indispensable for
achieving high decomposition rate to make up for low oxidizing ability of O
highly active catalysts [14].
2
even by using
Catalytic ozonation is the reaction for oxidative decomposition of contaminants with
. Because O is a strong oxidant comparing with O , the reaction often proceeds even under
O
3
3
2
mild reaction conditions, e.g., low temperature and ambient pressure [15-17]. While non-
catalytic ozonation of ammonia nitrogen in water [18] and catalytic one with a homogeneous
catalyst [19] were reported in 1975 and 1984, respectively, the first paper on the latter over
heterogeneous catalysts was reported in 2014 [20]. After that, Chen et al. reported Mg-Co
mixed metal oxide catalyst with different Mg to Co ratios for the reaction [21] and found that
MgO-Co
decomposition rate of ammonia nitrogen and selectivity to gaseous nitrogen compounds.
In our previous paper [20], eight metal oxides (MO , M = Co, Ni, Fe, Sn, Mn, Cu,
3
O
4
with Mg/Co = 4/1 calcined at 773 K was the best catalyst in terms of
x
Mg and Al) were tested as a catalyst for the ozonation of ammonia nitrogen in mild-acidic
–
solution under atmospheric pressure at 333 K in the presence of Cl . It was found that the
highest decomposition rate of ammonia nitrogen was obtained in the presence of MgO, but
5

–
undesirable NO was predominantly formed. In addition, a part of MgO was dissolved during
3
the reaction and pH of the reaction solution was increased up to around 9. The increase in pH
of the reaction solution for the same reaction in the presence of MgO was also reported by Liu
et al [22]. In stark contrast, Co
3
4
O gave the highest selectivity to gaseous nitrogen products
with more than 85% selectivity and was insoluble in the reaction solution, while the catalytic
activity was the third best among them. Furthermore, we found that the reaction proceeded
–
–
only in the presence of Cl , demonstrating that Cl was indispensable for the reaction over
–
Co
3
O
4
. However, it is not known how Cl is involved in the reaction over Co
3
4
O and why the
selectivity is so different for the reactions between in the presence of MgO and that of Co
3
4
O .
In the present study, we conducted an in-depth investigation to provide a clear
overview on the reaction mechanism for the ozonation of ammonia nitrogen in water in the
–
presence of Co
3
O
4
or MgO. For the former, we focused on the behavior of Cl in the reaction
solution during the reaction and tried to identify gaseous nitrogen products. The results clearly
demonstrated that Co effectively promoted the formation of chloramines (NH3-xCl , x = 1,
, and 3) as gaseous nitrogen products. Meanwhile, we found that the reaction proceeded with
3
O
4
x
2
a non-catalytic radical reaction in the presence of MgO and that MgO was just a reactant for
the reaction.
2
. Experimental
2
.1. Materials
Analytical grade reagents of Mg(OH)
SO , NH Cl, NaClO, NaCl and aqueous ammonia were purchased from Wako Pure
Chem. Ind. Ltd. and used without further purification.
2
, Ca(OH)
2
, Co(NO
3
)
2
·6H
2
O, KOH, HNO
3
,
(
NH
4
)
2
4
4
6

2
.2. Preparation of Co
Cobalt oxide (Co
aqueous ammonia (1.0 mol L ) was added to a solution of Co(NO
3
O
4
and MgO
3
O
4
) was prepared by a conventional precipitation method [20]. An
–
1
–1
3
) (1.0 mol L ) until pH of
2
the solution was stably at 8.1. The suspension was stirred for 10 min at room temperature. The
formed blue-green precipitate was separated by filtration, washed with distilled water, and
dried at 373 K for 24 h. The solid was ground with a mortar and then was calcined in air at
7
73 K for 4 h.
Magnesium oxide (MgO) was prepared by calcining Mg(OH)
Calcium oxide (CaO) was also prepared with the same manner to that of MgO except for
using Ca(OH)
2
at 773 K for 5 h [20].
2
.
2
.3. Characterization
Powder X-ray diffraction (XRD) patterns were recorded on an X-ray diffractometer
Rigaku, Miniflex) with Cu Kα radiation (λ= 0.154 nm, 30 kV, 15 mA) at room temperature.
Specific surface area was estimated by applying the Brunauer–Emmett–Teller (BET) equation
to an adsorption isotherm of N at 77 K, which was acquired on a Belsorp-mini instrument
BEL Japan Inc.). The sample was pretreated in N flow at 473 K for 3 h before the
measurement.
(
2
(
2
2
.4. Ozonation of ammonia nitrogen in water
Ozonation of ammonia nitrogen in water was performed by using a semi-batch
reactor connected to a gas flow line and traps (Fig. 1). Two reaction solutions containing 10
+, which was present as NH
mM NH
NH
4
3
when the solution was alkaline, were prepared from
4
Cl and (NH
4
)
2
SO
4
and are denoted as Sol(Cl) and Sol(SO ), respectively. Typically, 0.1 g
4
7

of Co
suspension was vigorously stirred in a stream of O
suspension reached to predetermined one (298 or 333 K), the gas was changed to a mixture of
3
4
O or MgO was added to the reactor containing 100 mL of the reaction solution and the
3
–1
2
(100 cm min ). After temperature of the
–
1
3
–1
O
3
/O
ammonia nitrogen. O
3UN-OX).
2
(1.88 mmol L as O
3
, total flow rate = 100 cm min ) to start the ozonation of
3
was generated from O
2
using an ozone generator (Tokyu Car Co., SO-
0
A small portion of the reaction solution was periodically withdrawn and analyzed to
determine the concentrations of NO −, NO − and NH +
3
2
4
by using two ion chromatographs
(Tosoh Co. Ltd., IC-2001). A column containing an anion-exchange resin (TSK gel Super IC-
−
1
CO
(3.1 mmol L−1)
AZ, Tosoh) and an aqueous solution of NaHCO
3
(2.9 mmol L ) and Na
2
3
were used as stationary and mobile phases, respectively, for anion analysis. For cation
analysis, a column containing a cation-exchange resin (TSK gel IC-Cation 1/2 HR, Tosoh)
−
1
and an aqueous solution of methanesulfonic acid (2.2 mmol L ) and 18-crown-6 (1.0 mmol
−
1
L ) were used as stationary and mobile phases, respectively. Since the eluting solution used
for the cation analysis was highly acidic, all ammonia nitrogen in the solution was detected as
+
was present in the solution. The yield of and selectivity to NO –
were
NH
4
even if NH
3
3
calculated by Eqs. 1 and 2, respectively.
-
Formed NO3-
Initial NH4+
Yield of NO =
(1)
(2)
3
-
Formed NO3-
Consumed NH4+
Selectivity to NO =
3
Because basically gaseous nitrogen products were not analyzed in this study, the selectivity to
them was calculated by subtracting the selectivity to NO –
3
from 100%.
8

While pH of both Sol(SO
4
) and Sol(Cl) was around 5.5, it was adjusted to 7 and 9 for
some reactions by the addition of aqueous solution of KOH. For the reaction in the presence
2
+
of MgO, the concentration of Mg in the reaction solution was determined by the ion
chromatograph in a similar manner to that of NH +
.
4
–
ClO
x
in solutions were determined by using another ion chromatograph (Thermo
Scientific, Dionex ICS-900) equipped with a column of Dionex IonPac AS16 RFIC. An
–
1
aqueous solution of NaOH (35 mmol L ) was used as a mobile phase.
Total chlorine in the reaction solution was quantified by using an inductively coupled
plasma mass spectrometer (ICP-MS, Agilent Technology, 8800 ICP-QQQ). Gas
chromatography mass spectroscopy (GC-MS) analysis was performed by using GCMS-
QP2010 SE (Shimadzu) equipped with a capillary column (SH-Rxi-5 Sil MS, 30 m × 0.25
mm × 0.25 μm, Shimadzu). Mass spectrum was obtained with electron impact mode, and the
interface and ion source were heated at 333 and 423 K, respectively.
+
–
2
.5. Reaction of NH
4
with ClO without O
3
Reaction of NH +
with ClO without O
–
was carried out in a similar manner to that for
was fed into the reactor instead of
mixture. The reaction solution was prepared by adding NaClO to Sol(Cl) and the
4
3
the ozonation of ammonia nitrogen in water, while O
/O
concentrations of NH +
pH of the reaction solution was adjusted to 5.5 with dilute nitric acid.
2
O
3
2
–
−1
4
and ClO in the solution were 10 and 20 mmol L , respectively. The
3
. Results and discussion
3
.1. Physical properties of Co
3
4
O and MgO
9

Fig. 2 shows the XRD patterns of as-prepared Co
diffraction lines due to spinel type Co were observed at 2θ = 19.08, 31.32, 36.90, 38,68,
4.82, 55.82, 59.34, 65.26 and 77.42°, which were ascribed to (111), (220), (311), (222),
400), (422), (511), (440) and (533), respectively. MgO gave the diffraction lines at 2θ 
6.91, 42.88, 62.27, 74.65 and 78.58°, which were assignable to (111), (200), (220), (311) and
222), respectively, for cubic MgO. The sizes of the crystallites estimated by applying a
Scherrer’s equation to the diffraction lines at 36.90 and 42.88° of Co and MgO were 26
3
O
4
and MgO. The typical
3
O
4
4
(
3
(
3
O
4
2
–
and 36 nm, respectively. The specific surface areas of Co
1,
3
O and MgO were 12 and 252 m g
4
respectively.
–
3
.2 Involvement of Cl for ozonation of ammonia nitrogen
Fig. 3 shows the conversion and selectivity at 6 h for the ozonation of ammonia
nitrogen in Sol(Cl) and Sol(SO
4
) at 333 K in the presence of Co
3
4
O and MgO. In the absence
of neither solid materials, only a small amount of NH +
was decomposed in Sol(Cl) and no
, NH +
4
reaction occurred in Sol(SO
4
). In the presence of Co
3
O
4
4
was decomposed only in
Sol(Cl), forming gaseous nitrogen products with ca. 83% selectivity, while no reaction
–
occurred in Sol(SO
4
), suggesting involvement of Cl in the reaction. To further confirm the
–
–
participation of Cl , we examined influence of the concentration of Cl on the conversion of
+
NH
4
and selectivities by the addition of NaCl into Sol(SO
4
) (Fig. 4). With an increase in the
–
+
concentration of Cl , the conversion of NH
4
increased. The selectivity to gaseous products
–
1
–
with 10 mmol L Cl was almost the same as that for the reaction in Sol(Cl). Therefore, it
–
was concluded that Cl played a critical role for the reaction over Co
3
4
O .
In contrast to the reactions over Co
3
4
O , there was little difference in the reaction results
with Sol(Cl) and Sol(SO
4
) in the presence of MgO. Almost all NH +
4
was decomposed at 6 h
10

–
in the presence of MgO and NO was a main product regardless of the reaction solutions.
3
–
These results indicate that Cl in the solution was not involved in the reaction with MgO. This
significant difference in the reaction behavior between Co
3
4
O and MgO strongly suggest a
difference in reaction mechanism in the presence of each solid.
3
3
.3. Reaction mechanism over Co
3
O
4
–
.3.1. Behavior of Cl during the reaction
–
As mentioned in 3.2, Cl was indispensable for the reaction over Co
3
4
O . To
–
–
investigate how Cl was involved in the reaction, time course in the concentration of Cl was
examined in Sol(Cl) in the presence of Co
3
O
4
(Fig. S1). It should be noted that the
–
–
concentration of Cl was decreased with reaction time and more than 85% of Cl in the
reaction solution was lost at 6 h.
–
by the oxidation of Cl–
/O mixture.
One possibility for the loss of Cl was the formation of Cl
2
with O
3
(Eqs. 3 and 4) and then it went out from the reactor with the flow of O
3
2
–
+
2
2
Cl + O
3
+ 2H → Cl
2
+ O
2
+ H
2
O
(3)
(4)
–
+ 2OH–
Cl + O
3
+ H O → Cl
2
2
+ O
2
In fact, no formation of Cl
2
occurred as was confirmed by the following experiment. The
and the outlet gas from the reactor
reaction was performed in Sol(Cl) in the presence of Co
3
O
4
was absorbed by pure water in the three traps connected in series. If Cl
was formed, Cl–
2
–
should be detected in the water along with ClO because Cl
2
is easily reacted with water to
–
–
form Cl and ClO (Eq. 5).
11

–
–
+
Cl
2
+ H
2
O → Cl + ClO + 2H
(5)
–
–
Neither Cl nor ClO was detected in water of the traps. To further investigatr the possibility
–
of the formation of Cl
2
by the oxidation of Cl with O
3
, we carried out the reaction in an
aqueous solution of NaCl with a similar manner to that in Sol(Cl). However, no formation of
–
–
Cl nor ClO was confirmed. Thus, the formation of Cl was improbable.
2
–
–
Another possibility for the loss of Cl was the formation of ClO
x
(x = 1–4) by the
during the reaction in Sol(Cl) in the
, we applied analytical conditions for ion-chromatography analysis being
reaction with O
presence of Co
able to detect them in the solution (see the experimental section). Unfortunately, ClO and
3
. To check the formation of ClO –
x
3
O
4
–
ClO –
cannot be analyzed individually even with the analytical conditions if they exist
3
together in the solution because unavoidable reduction of ClO –
to ClO occurs in the column.
–
3
–
–. The ion-chromatography analysis
Thus, we only knew the total amount of ClO and ClO
3
–
and ClO –
in the reaction solution at any reaction time. On the
showed no formation of ClO
2
4
–
–, was found in the solution, but the
other hand, ClO , some of which could come from ClO
3
–
–
amount of ClO at 6 h corresponded to only 5% or less of that of the lost Cl . Thus, the
–
–
–
formation of ClO or ClO
3
was not the main cause for the loss of Cl during the reaction in
Sol(Cl).
3
.3.2. Formation of chloramines
To investigate the involvement of NH +
–
4
in the loss of Cl , we evaluated the
–
consumption rate of Cl in an aqueous solution of NaCl in the presence of Co
3
4
O . It was
–
found that the consumption rate of Cl in the aqueous solution of NaCl was about only one-
–1
tenths of that in Sol(Cl), which were 0.087 and 0.92 mmol h , respectively. These results
12

–
+
strongly suggest that the reaction of Cl with NH gave some chlorinated nitrogen
4
compounds that were undetected by ion chromatography, namely non-ionic compounds.
To confirm the formation of such compounds, the reaction solution of Sol(Cl) in the
presence of Co
3
O
4
was analyzed by an inductively coupled plasma-mass spectrometry (ICP-
MS). The analysis with ICP-MS gives total concentration of chlorine in the solution. At 3 h of
–
1
the reaction, the total concentration of chlorine in the solution was 9.5 mmol L , while the
–
–1
concentration of Cl in the solution analyzed by ion chromatography was 7.0 mmol L (Fig.
–
S1). The fact that the total concentration of chlorine was higher than that of Cl indicated the
presence of non-ionic chlorinated nitrogen compounds in the reaction solution.
It is known that monochloramine (NH
2
Cl) is formed by the reaction of NH +
with
4
–
hypochlorite (ClO ) in weakly acidic aqueous solution even without any catalyst [23].
Monochloramine is further transformed into dichloramine (NHCl
2
) and trichloramine (NCl )
3
by the reaction with hypochloric acid (HClO) if pH of the solution is around 4 or less [24].
Initial pH of Sol(Cl) was 5.5 and pH was decreased to 3 at 6 h in the presence of Co
3
4
O due to
the oxidation of NH + into neutral products as well as anionic NO –
. Thus, the formation of
4
3
chloramines was possible for the reaction in Sol(Cl).
Chloramines easily undergo decomposition in strongly basic solution, giving Cl–
25]. Thus, we took out an aliquot of the reaction solution and pH of the solution was
[
increased to 11 by the addition of an aqueous KOH solution. If chloramines were present in
–
the solution, the concentration of Cl was increased after the pH increase. Furthermore, the
outlet gas of the reactor was absorbed into a strongly basic KOH solution (pH 11), in which
fine powder of MgO was further dispersed, in traps during the reaction in Sol(Cl) in the
presence of Co
3
4
O . Because chloramines especially trichloramine are less soluble in water,
chloramines formed in the reaction solution could go out from the reactor with a flow of
13

O
3
/O mixture. To avoid slipping chloramines from the trap solution, three traps containing
2
the strongly basic KOH solution with MgO fine powder were connected in series at the outlet
of the reactor (Fig. 1) and we confirmed that there was nothing detected in the third trap. The
reaction solution and trap solutions were analyzed by ion chromatography to determine the
concentration of Cl–.
–
Fig. 5 displays time-course changes in the concentrations of Cl in the reaction
solution as it was and in the reaction solution after the pH increase, and the concentration of
–
–
Cl in the trap solution. For easy understanding, amounts of Cl found in the three trap
solutions are combined and is displayed as if those exist in the reaction solution. It should be
–
noted that the concentration of Cl in the reaction solution was increased after the pH
increase. Especially at 3 h, the increment was significant. This behavior strongly suggested
the formation of chloramines. The amount of chlorine being incorporated in chloramines,
–
which was estimated from the difference in the concentrations of Cl before and after the pH
–
1
+
increase, corresponded to 2.5 mmol L at 3 h. At 6 h, at which almost all NH disappeared
4
–
in the reaction solution (Fig. S1), the increase in the concentration of Cl by the pH increase
was only little, suggesting that the formed chloramines went out from the reactor. In fact, the
–
concentration of Cl in the trap solution was drastically increased at 6 h. However, chlorine
balance defined as Eq. 6 at 6 h was about 85%, suggesting that some chlorine-containing
compounds that were stable even in the strong alkaline solution were formed and/or were
slipped out of the traps.
[
Cl-]reaction solution after pH increaseꢀ[Cl-]trap
Chlorine balance (%)=
× 100
(6)
[
Cl-]0
14

Unfortunately, we do not have any evidence to explain it and now it is still under
investigation.
Next, we tried to detect chloramines directly by using GC-MS. The reaction solution at
3
h was analyzed, but chloramines were not detected probably because the concentrations
were below the detection limits. Thus, we applied a head-space GC-MS method to
demonstrate the formation of chloramines during the reaction. A head-space GC-MS method
is suitable for analyzing minute amounts of volatile compounds present in solution, because
such compounds are enriched in a gas phase of a vial container with the solution, namely head
space, when the solution is allowed to stand in the sealed container [26]. Unfortunately, no
standard solution of chloramines is available for analysis, because chloramines are unstable
and are easy to strip into air. Thus, we performed only a qualitative analysis here.
According to Fig. 5, it seemed that the concentration of chloramines in the reaction
solution was the highest at 3 h for the reaction in Sol(Cl) in the presence of Co
3
4
O . Thus, we
analyzed that reaction solution with the head-space GC-MS method. However, no peak due to
chloramines was found on the gas chromatogram, probably because their concentrations in the
solution were below detection limits for the head-space GC-MS analysis.
Thus, we collected chloramines formed during the reaction in Sol(Cl) in the presence
of Co
3
4
O by making the outlet gas of the reactor absorbed into pure water in a trap for 6 h and
analyzed the trap solution. Fig. 6a gives the chromatogram of the trap solution obtained by the
head-space GC-MS method. The large peak at 1.1 min on the chromatogram was due to air.
The chromatogram had a large peak at 2.4 min, while very small ones were observed at 1.4
and 1.6 min. Mass spectrum for the peak at 2.4 min showed the peaks at m/z = 119 and 121
3
5
35 35
(
Fig. 6b), which was assignable to the molecular ions of trichloramine with N Cl Cl Cl and
3
5
35 37
35 37 37
37 37 37
N Cl Cl Cl, respectively. Since the abundance ratios of N Cl Cl Cl and N Cl Cl Cl
15

among trichloramine were small, being 0.07 and 0.02, respectively, the peaks at m/z = 123
and 125 assignable to their molecular ions were not observed. In addition to the molecular ion
peaks, various fragment ion peaks were observed at 84, 86 and 88 (NCl +
) and 49 and 51
2
+
(
NCl ) with reasonable intensity ratios. Therefore, it was concluded that the formation of
trichloramine was successfully confirmed by using the head-space GC-MS method. The small
peaks at 1.4 and 1.6 min on the chromatogram seemed to be monochloramine and
dichloramine, but the identifications were impossible due to low intensities.
3
.3.3. Re-evaluation of the selectivity for the reaction over Co
In our previous paper, only NO –
3
O
4
3
was analyzed as a product and the rest of the
products was regarded as gaseous nitrogen products, which were presumed to N
20]. However, as was demonstrated in 3.3.2, the formation of chloramines was confirmed,
though it is still unknown which chloramine and how much amount of them were formed.
2
and N
2
O
[
Thus, we re-evaluated the selectivity for the reaction in Sol(Cl) in the presence of Co
taking the formation of chloramines into account. At 6 h, conversion of NH +
3
O by
4
4
was 97%. At the
–
–
time, selectivity to NO
3
was 17%. The combined concentration of Cl derived from the
–
1
–1
formed chloramines in the reaction solution (1.9 mmol L ) and trap solutions (5.8 mmol L )
–
1
was 7.7 mmol L at 6 h. If all the chloramines were trichloramine (NCl
was 26%. Thus, the selectivity to other gaseous nitrogen products, which might be N
NO, and NO , was 57%. On the other hand, if all the chloramines were monochloramine
NH Cl), the selectivity to it was 79% and that to other gaseous nitrogen products was 4%.
3
), the selectivity to it
2
, N O,
2
2
(
2
According to the head space GC-MS analysis, it was presumed that trichloramine was
predominant among three chloramines. Thus, the selectivity to chloramines is thought to be
16

close value to the former estimation. However, to obtain accurate selectivity, the gaseous
nitrogen products should be qualitatively analyzed, and it is still in progress.
3
.3.4. Reaction pathway for the reaction over Co
3
O
4
As mentioned before, it is reported that chloramines are formed by the reaction of
with hypochlorite ion (or hypochlorous acid) in acid solution [23]. Under the reaction
+
NH
4
conditions in Sol(Cl) in this study, hypochlorite ion was formed. Thus, it was plausible that
monochloramine was formed by the reaction of NH +
with ClO , and dichloramine and
–
4
trichloramine were successively formed by the reaction of mono- and dichloramine,
–
+
respectively, with ClO (or HClO). As Fig. 3 showed, the decomposition rate of NH
4
in the
presence of Co
3
4
O was higher than that in its absence. There are two possibilities for the
–
promotion by Co
3
O
4
: Co
3
O
4
accelerated (i) the formation of ClO and (ii) the reaction of
+
–
–
NH
4
with ClO . However, the former was improbable, because the formation of ClO by the
–
reaction of Cl with O
3
was rather inhibited by the presence of Co
3
O
4
for the reaction in an
aqueous solution of NaCl (Fig. S2).
To confirm the possibility of the latter, we carried out the reactions of NH
+
4
with
–
ClO under O
2
flow without O
3
in the presence and absence of Co
3
O
4
(Fig. 7). As Fig. 7
was much higher than
clearly demonstrates, the conversion of NH +
in the presence of Co
O
4
3
4
that in its absence. Furthermore, we confirmed the formation of chloramines found in the
reaction solution and trap solutions by similar manners to those performed for the reaction
shown in Fig. 5. Therefore, it was concluded that Co
ClO–.
3
O
4
accelerated the reaction of NH +
with
4
Based on the results, we propose the reaction pathway for the reaction over Co
3
O
4
as
–
–
Fig. 8. In the first step, Cl is non-catalytically oxidized with O
3
to form ClO . Co
3
O
4
rather
17

+
–
to form NO –
,
inhibits this reaction. Thereafter, NH
4
is reacted with ClO on Co
3
O
4
3
chloramines, and other gaseous nitrogen compounds.
3
3
.4 Reaction mechanism for ozonation of NH +
4
in the presence of MgO
.4.1. Role of MgO for the ozonation of ammonia nitrogen
Fig. 9 shows time-course changes in the concentrations of ammonia nitrogen as well
–
2–
as Cl and SO
4
for the reactions in Sol(Cl) and Sol(SO ) in the presence of MgO. In contrast
4
to Co
3
O
4
, the ozonation of ammonia nitrogen proceeded even in Sol(SO
4
) in the presence of
–
–
MgO, and the presence of Cl did not affect the reaction rate and selectivity. In fact, Cl was
not lost over the reaction time in Sol(Cl), suggesting that the reaction did not follow the
reaction mechanism as it occurred over Co
A stark difference between Co
behavior. In contrast to insoluble Co
3
O .
4
3
O
4
and MgO for the reaction was their dissolution
3
O
4
, about 40% of MgO was dissolved in Sol(SO ) for 6
4
h, though MgO is known to be poorly soluble in water. Interestingly, pH of the reaction
solution rose from 5.5 to around 9 just by the addition of MgO even before the flowing of O
3
and the pH was maintained throughout the reaction. This pH increase could be caused by the
forward reaction of Eq. 7.
2
+
–
MgO(s) + H
2
O ⇄ Mg + 2OH
(7)
Ideally, a small amount of MgO less than 0.1% of the amount added in the reaction solution is
sufficient to cause the pH increase. Liu et al. also proposed a similar mechanism for the pH
increase of the solution by the addition of MgO [22]. Since the pH was quickly increased just
18

by the addition of MgO and was maintained near constant throughout the reaction, the
reaction of Eq. 7 must reach equilibrium quickly under the reaction conditions with MgO.
To get an insight into the role of MgO, firstly, a reaction in the absence of MgO at
(= [NH +
][OH ]/[NH
−
]) is 10–4.8, ammonia nitrogen
pH 9 was carried out. Because K
b
of NH
3
4
3
existed predominantly as NH
3
in the reaction solution. As shown in Fig. S3, 16% of ammonia
nitrogen was decomposed at 1 h and concomitantly NO –
that was equivalent to the consumed
3
ammonia nitrogen was formed. After that, however, the reaction was completely stopped. At
the time, the solution turned to acidic because of charge compensation for the formation of
anionic product (NO
3
–) from neutral reactants (NH
3
and O ) (Eq. 8).
3
–
+
NH
3
+ 4O
3
→ NO
3
+ H
2
O + H + 4O
2
(8)
We further carried out a similar reaction, but pH of the reaction solution was repeatedly
restored to 9 at each sampling time (Fig. S4). For that reaction, the decomposition of
ammonia nitrogen continually proceeded. Under the reaction conditions, stripping of NH
3
from the reaction solution was only little. These results clearly demonstrated that in mildly
alkaline solution, ammonia nitrogen can be oxidized into mainly NO –
3
even without any
catalysts.
As mentioned above, to increase pH to around 9, only less than 0.1% of MgO added
to the reaction solution was sufficient, but the fact was that much amount of MgO was
dissolved with the reaction, implying the presence of another mechanism for the dissolution
of MgO. To get a clue for it, we carried out the ozonation of ammonia nitrogen with different
initial concentrations of ammonia nitrogen in the presence of MgO and investigated the
relationship between the amount of decomposed ammonia nitrogen and that of dissolved
19

2
+
Mg at each reaction time with different initial concentrations of ammonia nitrogen. The
2
+
results are shown in Fig. 10. Notably, the amount of Mg dissolved was linearly increase
with increase in the amount of ammonia nitrogen decomposed. It should be noted that the
slope of the line in Fig. 10 was nearly unity. Because ammonia nitrogen exists originally as
+
NH
4
in the as-prepared reaction solution before the addition of MgO, the stoichiometric
is as follows.
equation for the oxidation of ammonia nitrogen to NO –
3
NH +
+ 4O
→ NO
–
+ H
O + 2H + 4O
+
(9)
4
3
3
2
2
Since the pH of the solution was maintained at around 9 throughout the reaction in the
+
–
presence of MgO, H formed by the reaction of Eq. 9 was neutralized with OH that was
formed by the forward reaction of Eq. 7, inducing further dissolution of MgO. Consequently,
2
+
a single decomposition reaction of ammonia nitrogen yielded the dissolution of one Mg as
represented by Eq. 10.
NH +
+ 4O
+ MgO → NO
–
+ 4O
+ Mg + 2H
2+
(10)
4
3
3
2
2
O
In short, MgO was a reactant for the ozonation of ammonia nitrogen in water.
In addition to MgO, CaO, which is another alkaline-earth metal oxide, acted in the
same way as MgO did for the reaction. The reaction solution turned into strong basic (pH
around 11) immediately by the addition of CaO to it. The ozonation of ammonia nitrogen
proceeded smoothly in the presence of CaO and the decomposition rate of ammonia nitrogen
was about 1.5-times faster than that in the presence of MgO. This supports our conclusion
above mentioned in the case of MgO.
20

3
.4.2. Possibility of involvement in homogeneous catalysis by Mg2+
2
+
Because Mg was present in the reaction solution for the reaction with MgO, it
2
+
could be that Mg acted as a homogeneous catalyst. Thus, we performed the ozonation of
–
1
2+
ammonia nitrogen in Sol(SO
4
) with 10 mmol L of Mg , which was prepared by adding
Mg(NO ·6H O to the reaction solution. The initial pH of the reaction solution was adjusted
3
)
2
2
to 9 by the addition of KOH. The result showed that the conversion of ammonia nitrogen was
only 12% at 1 h and after that the reaction stopped, which was the same as that for the blank
test in the absence of MgO (Fig. S3). Therefore, it was concluded that there was no
2
+
contribution of homogeneous catalysis by Mg to the ozonation of ammonia nitrogen.
3
.4.3. Reaction pathway for the reaction in the presence of MgO
The effect of radical scavenger was investigated for the reaction in Sol(SO ) in the
4
–
1
presence of both MgO and tert-butyl alcohol (TBA, 100 mmol L ) and the result was
compared to that in the absence of TBA (Fig. 11). Since TBA acts as a radical scavenger, the
addition of TBA decreases the concentration of radicals in the solution [27]. Thus, if the
reaction follows radical mechanism, the decomposition rate of ammonia nitrogen is decreased
along with decrease in the leaching amount of MgO. As Fig. 11 clearly shows, the
decomposition of ammonia nitrogen was inhibited by the presence of TBA and the leaching
amount of MgO was also decreased, indicating the involvement of a radical mechanism. Liu
et al. also investigated the effect of TBA for ozonation of ammonia nitrogen in the presence of
MgO and reached a totally opposite conclusion that TBA did not interfere the reaction [22]. In
the reaction in Fig. 11, tenfold amount of TBA to ammonia nitrogen was used, while only
one-tenth of it was added to the reaction solution in the study by Liu et al. [22]. Such
21

difference in the amount of TBA added may bring about the discrepancy of the reaction
mechanism.
Based on the results mentioned so far and the fact that OH· radical is formed by the
–
reaction of OH with O
3
at pH more than 9 [28], we propose the reaction pathway for the
reaction in the presence of MgO as Fig. 12. The reaction of solid MgO with H
O giving Mg2+
2
–
and OH is in equilibrium and thus pH of the solution is maintained at mildly alkaline pH.
Ammonia nitrogen is decomposed to mainly NO –
with the decomposition of ammonia nitrogen is neutralized, leading to further dissolution of
by the reaction with OH·. Proton formed
3
2
+
MgO. Thus, the same amount of Mg as that of decomposed ammonia nitrogen is dissolved
in the reaction solution.
4
. Conclusion
In the present study, we investigated the reaction mechanisms for ozonation of
–
ammonia nitrogen in the presence of Co
3
O
4
or MgO. For the reaction over Co
3
O , Cl in the
4
solution was indispensable and chloramines (NH3-xCl
NCl
and other gaseous nitrogen products (N
x
), being predominantly trichloramine
–, chloramines
(
3
), were formed as a gaseous nitrogen product. The selectivities to NO
3
2
, N
2
O, and NO
x
) were estimated to 17, 26, and 57%,
–
if only trichloramine was formed in chloramines. Hypochlorite ion (ClO ) that was formed by
–
acted as an oxidizer for NH +
a non-catalytic reaction of Cl with O
3
4
and Co
3
4
O accelerated
the reaction of NH +
with ClO to give the products.
–
4
–
In contrast to the reaction with Co
3
O
4
, Cl was not involved in the reaction with
MgO. During the reaction in the presence of MgO, pH of the reaction solution was maintained
–
2+
at around 9 owing to the reaction of MgO with H
2
O to give OH (MgO(s) + H
2
O ⇄ Mg +
–
–
2
OH ). Ammonia nitrogen is decomposed to form mainly NO
3
by a non-catalytic radical
22

–
reaction with OH·, which was formed by the reaction of OH with O
3
in mildly alkaline
+ 4O → NO
) were neutralized with OH to keep the pH of the reaction solution at
solution. Protons formed with the decomposition of ammonia nitrogen (NH +
–
4
3
3
+
–
+
H
2
O + 2H + 4O
2
2
+
around 9. As a result, much the same amount of Mg as that of decomposed ammonia
nitrogen was dissolved in the reaction solution. In short, MgO was a reactant for the ozonation
of ammonia nitrogen in water.
Acknowledgements
The first author of the paper (P. A. Krisbiantoro) received a scholarship from
Indonesia Endowment Fund for Education (LPDP scholarship) from the Ministry of Finance,
Republic of Indonesia.
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the online version, at
doi:https://doi.org./10.1016/j.apcata.?????????.
References
[
[
[
[
[
[
[
1] D.J. Randall, T.K.N. Tsui, Mar. Pollut. Bull. 45 (2002) 17–23.
2] T.A. Pressley, D.F. Bishop, S.G. Roan, Environ. Sci. Technol. 6 (1972) 622–628.
3] H.B. Perry, R. Zulliger, M.M. Rogers, Annu. Rev. Public Heal. 35 (2014) 399–421.
4] M. Owens, G. Wood, Water Res. 2 (1968) 151–159.
5] Y. Tu, P. Feng, Y. Ren, Z. Cao, R. Wang, Z. Xu, Fuel 238 (2019) 34–43.
6] A. Thornton, P. Pearce, S.A. Parsons, J. Hazard. Mater. 147 (2007) 883–889.
7] Y. Wei, M. Ji, R. Li, F. Qin, Waste Manag. 32 (2012) 448–455.
23

[
[
[
[
[
[
[
8] P.H. Liao, A. Chen, K.V. Lo, Bioresour. Technol. 54 (1995) 17–20.
9] J. Taguchi, T. Okuhara, Appl. Catal. A 194-195 (2000) 89–97.
10] J. Taguchi, Y. Yoshinaga, T. Okuhara, Chem. Lett. 30 (2001) 112–113.
11] J. Barbier, L. Oliviero, B. Renard, D. Duprez, Catal. Today 75 (2002) 29–34.
12] D.K. Lee, S.C. Jeong, L.Y. Wang, Chemosphere 61 (2005) 573–578.
13] C. Lousteau, M. Besson, C. Descorme, Catal. Today 241 (2015) 80–85.
14] S.K. Bhargava, J. Tardio, J. Prasad, K. Föger, D.B. Akolekar, S.C. Grocott, Ind. Eng.
Chem. Res. 45 (2006) 1221–1258.
[
[
15] S.T. Oyama, Catal. Rev. Sci. Eng. 42 (2000) 279–322.
16] B. Kasprzyk-Hordern, M. Ziółek, J. Nawrocki, Appl. Catal. B Environ. 46 (2003) 639–
6
69.
[
[
[
[
[
17] J. Nawrocki, B. Kasprzyk-Hordern, Appl. Catal. B 99 (2010) 27–42.
18] P.C. Singer, W.B. Zilli, Water Res. 9 (1975) 127–134.
19] W.R. Haag, J. Hoigné, H. Bader, Water Res. 18 (1984) 1125–1128.
20] S. Ichikawa, L. Mahardiani, Y. Kamiya, Catal. Today 232 (2014) 192–197.
21] Y. Chen, Y. Wu, C. Liu, L. Guo, J. Nie, Y. Chen, T. Qiu, J. Environ. Sci. 66 (2018) 265–
2
73.
[
[
[
[
[
22] H. Liu, L. Chen, L. Ji, J. Hazard. Mater. 376 (2019) 125–132.
23] P.J. Vikesland, K. Ozekin, R.L. Valentine, Water Res. 35 (2001) 1766–1776.
24] R.M. Chapin, J. Am. Chem. Soc. 51 (1929) 2112–2117.
25] V.C. Hand, D.W. Margerum, Inorg. Chem. 22 (1983) 1449–1456.
26] K. Kosaka, K. Seki, N. Kimura, Y. Kobayashi, M. Asami, Water Sci. Technol.: Water
Supply 23 (2010) 23–29.
24

[
[
27] A. Ikhlaq, D.R. Brown, B. Kasprzyk-Hordern, Appl. Catal. B Environ. 165 (2015) 408–
18.
28] J. Hoigne, H. Bader, Environ. Sci. Technol. 12 (1978) 79–84.
4
25

Fig. 1. Illustration of the semi-batch reactor used for ozonation of ammonia nitrogen in
water. Pure water or an aqueous KOH solution (pH 11) with fine powder of MgO was put in
three traps.
26

Fig. 2. XRD patterns of Co
3
4
O and MgO.
27

NO3-
+
Conversion of NH4
Selectivity to
gaseous N compound
100
80
60
40
20
0
no reaction
no reaction
Sol(Cl) Sol(SO4)
Sol(Cl) Sol(SO4)
Sol(Cl) Sol(SO4)
Blank
Co O
MgO
3
4
Fig. 3. Conversions and selectivity for ozonation of ammonia nitrogen in Sol(Cl) and
Sol(SO ) in the presence of Co or MgO. Sol(Cl) and Sol(SO ) represent the reaction
solutions prepared from NH , respectively. Reaction conditions: Co
/O mixture, 100 mL
4
3
O
4
4
4
Cl and (NH
4
)
2
SO
4
3
4
O or
–
1
+
MgO, 0.1 g; 10 mmol L NH
4
solution, 100 mL; total flow rate of O
3
2
–1
min ; concentration of O
3
in the mixture, 0.7 vol%; reaction temperature, 333 K; reaction
time, 6 h.
28