catalysts

Article
Comparision on the Low-Temperature NH3-SCR
Performance of γ-Fe2O3 Catalysts Prepared by
Two Different Methods
Naveed Husnain , Enlu Wang * , Shagufta Fareed and Muhammad Tuoqeer Anwar
Institute of Thermal Energy Engineering, School of Mechanical Engineering, Shanghai Jiao Tong University,
Shanghai 200240, China; naveed69@sjtu.edu.cn (N.H.); shagufta91@sjtu.edu.cn (S.F.);
engr.tauqeer137@gmail.com (M.T.A.)
* Correspondence: elwang@sjtu.edu.cn; Tel.: +86-21-34205683




Received: 28 October 2019; Accepted: 30 November 2019; Published: 3 December 2019


Abstract: Maghemite (γ-Fe2 O3 ) catalysts were prepared by two different methods, and their activities
and selectivities for selective catalytic reduction of NO with NH3 were investigated. The methods of
X-ray powder diffraction (XRD), Brunauer–Emmett–Teller (BET), X-ray photoelectron spectroscopy
(XPS), hydrogen temperature-programmed reduction (H2 -TPR), ammonia temperature-programmed
desorption (NH3 -TPD), transmission electron microscopy (TEM), Energy-dispersive X-ray spectroscopy
(EDS), and in situ diffuse reflectance infrared Fourier transform spectroscopy (in situ DRIFTS) were
used to characterize the catalysts. The resulted demonstrated that the γ-Fe2 O3 nanoparticles prepared
by the facile method (γ-Fe2 O3 –FM) not only exhibited better NH3 -SCR activity and selectivity than the
catalyst prepared by the coprecipitation method but also showed improved SO2 tolerance. This superior
NH3 -SCR performance was credited to the existence of the larger surface area, better pore structure,
a high concentration of lattice oxygen and surface-adsorbed oxygen, good reducibility, a lot of acid
sites, lower activation energy, adsorption of the reactants, and the existence of unstable nitrates on the
surface of the γ-Fe2 O3 –FM.

Keywords: low-temperature; γ-Fe2 O3 ; nanoparticles; NO reduction; selective catalytic reduction;

facile one-step synthesis method; coprecipitation method

1. Introduction
Nitrogen oxides (NOx , x = 1, 2) are the main sources of global environmental concerns like acid
rain, fine particle pollution, smog, and ozone depletion [1]. The demand for reducing fuel consumption
and making the environment clean has increased with increased awareness in society about protecting
the global environment. Due to this persistent demand, NOx legislation for both point and mobile
sources has become more and more strict [2]. Therefore, in the recent era, a lot of focus is given to
the removal of NOx by the researchers. To meet these strict regulations for NOx abatement, selective
catalytic reduction (SCR) of NOx with ammonia (NH3 ) has become the most efficient and widely used
NOx abatement technology, for point sources like coal-fired power plants [3].
In recent years, Fe2 O3 has been extensively studied as a primary catalyst due to its naturally
environmentally benign features, low toxicity, and noticeable thermal stability [2]. For example,
Shen et al. [4] used a sol–gel method to dope Mn-Ce/TiO2 with iron to make Fe-Mn-Ce/TiO2 catalyst.
It was found that the addition of Fe has enhanced the low-temperature NH3 -SCR activity of the catalyst.
Yang et al. [5] substituted low-cost Fe2 O3 to replace WO3 in the V2 O5 /WO3 -TiO2 catalyst to improve
the N2 selectivity of the catalyst. He found that the support (Fe2 O3 -TiO2 ) mainly resulted in the acid
sites in the catalysts, so the adsorbed NH3 was favored to be activated by Fe3+ rather than by V5+ .

Catalysts 2019, 9, 1018; doi:10.3390/catal9121018 www.mdpi.com/journal/catalysts

Catalysts 2019, 9, 1018 2 of 17

Catalysts
Liu et al. [6]2019,
used9, x the
FOR hydrothermal
PEER REVIEW method to prepare an environmentally benevolent Fe-Ce-Ti 2 of 17mixed

oxide catalyst. It was found that the addition of Fe significantly achieved excellent NO reduction (over
sites in the catalysts, so the ◦ adsorbed NH3 was favored to be activated by Fe3+ rather than by V5+. Liu
90%etNO conversion at 200 C), N2 selectivity
al. [6] used the hydrothermal method
(nearly 100% N2 selectivity),
to prepare an environmentally
and SO2 tolerance
benevolent Fe-Ce-Ti mixed
of the
catalysts. Wang et al. [7] prepared
oxide catalyst. It was found that the Fe O particles
2 addition
3 of Fe significantly achieved excellent NO reduction that
with aqueous precipitation method and found
the Fe O
(over particles exhibited excellent catalytic
2 3 90% NO conversion at 200 °C), N2 selectivity (nearly activity reaching
100% almost 95% in and
N2 selectivity), a low-temperature
SO2 tolerance of range
of 150 ◦
thetocatalysts.
270 C.Wang However, theprepared
et al. [7] SO2 tolerance of the Fe
Fe2O3 particles 2 O3aqueous
with catalystprecipitation
was not very good.and
method The literature
found
showsthatthe
thepotential of Feexhibited
Fe2O3 particles O
2 3 catalysts as
excellent an active
catalytic component
activity reaching to selectively
almost 95% in catalyze
a the
low-temperature reduction
of NOrangeto of 150 to 270
nontoxic N2 °C. However, the SO2 tolerance
at low-temperatures. In most of the Fe2O3 works,
of these catalyst the
was active
not very good. Thein the
ingredient
literature shows the potential of Fe 2O3 catalysts as an active component to selectively catalyze the
Fe2 O3 was α-Fe2 O3 (hematite). However, Liu et al. [8] prepared α-Fe2 O3 and γ-Fe2 O3 particles by the
reduction of NO to nontoxic N2 at low-temperatures. In most of these works, the active ingredient in
coprecipitation method and compared the catalytic activity and selectivity of both these phases of
the Fe2O3 was α-Fe2O3 (hematite). However, Liu et al. [8] prepared α-Fe2O3 and γ-Fe2O3 particles by
ferric oxide under the same conditions. The results demonstrated that the γ-Fe2 O3 catalyst’s surface
the coprecipitation method and compared the catalytic activity and selectivity of both these phases
adsorbed the nitric oxide and ammonia more easily and efficiently than then α-Fe2 O3 catalyst’s
of ferric oxide under the same conditions. The results demonstrated that the γ-Fe2O3 catalyst’s surface
surface.
Thisadsorbed
enhanced the nitric oxide and ammonia more easily and efficiently than2 then α-Fe2O3 catalyst’s 2 O3
adsorption resulted in superior catalytic activity and N selectivity of the γ-Fe
catalyst at low temperatures ◦ C). γ-Fe O is nonstoichiometric iron oxide. It can be considered
(100–300resulted
surface. This enhanced adsorption in2superior
3 catalytic activity and N 2 selectivity of the γ-
as a Fe 2+
Fe2O3-deficient 3+ 3+ V
catalyst atFe low O
3 4 with a
temperaturesunit cell representation
(100–300 °C). γ-Fe 2 O 3of
is (Fe tet )8 [Feoct 5/6
nonstoichiometric ironFe1/6 ]16 OIt32can
oxide. , where
be the
considered as a Fe 2+-deficient Fe3O4 with a unit cell representation of (Fetet3+)8[Feoct3+5/6V 1/6]16O32, where
brackets () represents tetrahedral sites, VFe represents a cations vacancy, and [] represents Fe octahedral
sites.the brackets
Some portion of Fe3+ can
() represents tetrahedral
be easilysites, VFe to
reduced Fe2+ in magnetite
represents a cations vacancy,
because and of its[]metastability
represents [9].
octahedral
Figure 1 shows sites.
the Some portion
structure of Fenonstoichiometric
of the
3+ can be easily reduced to Fe2+ in magnetite because of its
iron oxide (γ-Fe2 O3 ). The octahedral and
metastability [9]. Figure 1 shows the structure of the nonstoichiometric iron oxide (γ-Fe2O3). The
tetrahedral interstitial sites in the γ-Fe2 O3 structure along with octahedral vacancies could be seen.
octahedral and tetrahedral interstitial sites in the γ-Fe2O3 structure along with octahedral vacancies
Apart from this, γ-Fe2 O3 can also facilitate CO oxidation [10] desulfurization [11] and mercury
could be seen. Apart from this, γ-Fe2O3 can also facilitate CO oxidation [10] desulfurization [11] and
oxidation [12].
mercury oxidation [12].

(A) (B)

Figure 1. Structure
Figure of of
1. Structure γ-Fe 2O
γ-Fe 3 3(A)
2O (A)octahedral (Feoctoct
octahedral (Fe ) and
) and tetrahedral
tetrahedral (Fe(Fe tet ) interstitial
tet) interstitial
sites along
sites along with with
octahedral
octahedral vacancies; (B) spinel structure of γ-Fe22O3 3with randomly distributed Feoct vacancies [13]. [13].
vacancies; (B) spinel structure of γ-Fe O with randomly distributed Fe oct vacancies

It isItreported
is reportedthat ininthe
that theSCR
SCRreaction, themethod
reaction, the methodofofpreparation
preparation of aof a catalyst
catalyst strongly
strongly affectsaffects
the catalytic activity
the catalytic ofof
activity catalyst
catalyst[14].
[14]. For example,Chen
For example, Chenet et
al.al. [15]
[15] used
used three
three different
different methods
methods to to
synthesized a series of Fe-Mn mixed oxide catalysts. The methods were the coprecipitation
synthesized a series of Fe-Mn mixed oxide catalysts. The methods were the coprecipitation method, method,
citriccitric
acidacid method,
method, and
and solidreaction
solid reaction method.
method. The Thecatalyst
catalystsynthesized
synthesized by the
by citric acid acid
the citric methodmethod
exhibited better activity (reaching 98.8% NO x conversion) at 120 °C ◦than other methods. Previously,
exhibited better activity (reaching 98.8% NOx conversion) at 120 C than other methods. Previously,
Wang et al. [16] synthesized a series of γ-Fe2O3 catalysts with microwave-assisted coprecipitation
Wang et al. [16] synthesized a series of γ-Fe2 O3 catalysts with microwave-assisted coprecipitation
method (MACP). The catalytic activity of γ-Fe2O3 catalysts synthesized by using different
method (MACP). The catalytic activity of γ-Fe2 O3 catalysts synthesized by using different precipitation
precipitation techniques as well as by using different precipitants was compared. The results
techniques as wellthat
demonstrated as by usingalldifferent
among precipitants
the catalysts, the γ-Fewas compared. The results demonstrated that
2O3 catalyst synthesized by direct titration
among precipitation technique and NH4OH as precipitant showedby
all the catalysts, the γ-Fe 2 O3 catalyst synthesized direct
better SCRtitration precipitation
performance. However,technique
the
and researchers
NH4 OH aswere precipitant
unable to achieve good surface chemistry of the γ-Fe2O3 catalyst due were
showed better SCR performance. However, the researchers to theunable
to achieve good surface chemistry of the γ-Fe2 O3 catalyst due to the conventional preparation
method. As a result, NO reduction activity was not good (approximately 52% at 200 ◦ C) at low
Catalysts 2019, 9, 1018 3 of 17
Catalysts 2019, 9, x FOR PEER REVIEW 3 of 17

conventional In
temperatures. preparation
addition, method. As a result,
the conventional NO reduction
synthesis processactivity was not
to prepare γ-Fe good (approximately
2 O3 particles, such as
52% at 200 °C)[7],
coprecipitation at microemulsions
low temperatures.[17], In addition, the conventional[18–20],
thermal decomposition synthesis process
sol–gel to prepare
[21–23], γ-
and some
Fe 2O3 particles, such as coprecipitation [7], microemulsions [17], thermal decomposition [18–20], sol–
other chemical processes [24,25], also requires long reaction time, suitable pH value, and surfactants or
gel [21–23],
certain additives andfor some other chemical
synthesizing pure processes [24,25], also
γ-Fe2 O3 particles withrequires long reaction
controllable time, suitable
morphology. pH
Furthermore,
value, and surfactants or certain additives for synthesizing pure γ-Fe2O3 particles with controllable
in order to gain monodisperse and single-phase products, purification and centrifugation are also
morphology. Furthermore, in order to gain monodisperse and single-phase products, purification
required. All this makes these processes very cumbersome and complicated procedures for synthesizing
and centrifugation are also required. All this makes these processes very cumbersome and
monodisperse γ-Fe2 O3 .
complicated procedures for synthesizing monodisperse γ-Fe2O3.
Motivated by the idea to find an efficient, environmentally benign, and low-cost γ-Fe2 O3 catalyst
Motivated by the idea to find an efficient, environmentally benign, and low-cost γ-Fe2O3 catalyst
with improved NH3 -SCR performance, we prepared γ-Fe2 O3 nanoparticles by a facile one-step
with improved NH3-SCR performance, we prepared γ-Fe2O3 nanoparticles by a facile one-step
synthesis
synthesis method.
method.The TheNH NH 3 -SCR performance of the γ-Fe O catalyst prepared by the facile method
3-SCR performance of the γ-Fe22O3 3catalyst prepared by the facile method
(γ-Fe 2 O23O–FM)
(γ-Fe was also compared with the catalyst prepared by the conventional coprecipitation
3–FM) was also compared with the catalyst prepared by the conventional coprecipitation
method (γ-Fe
method (γ-Fe O
2 2O3 –CP). The methods of XRD, BET, XPS, H -TPR, NH3 -TPD,
3–CP). The methods of XRD, BET, XPS, H22-TPR, NH3-TPD,
TEM,
TEM, EDS,EDS, and
and in in situ
situ
diffuse
diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) were used to characterize thethe
reflectance infrared Fourier transform spectroscopy (DRIFTS) were used to characterize
catalysts.
catalysts.The Theresult
resultdemonstrated
demonstratedthat thatthe γ-Fe22O
theγ-Fe -FM exhibited
O3-FM exhibitedsuperior
superiorNH NH 3 -SCR
3-SCR performance
performance
than γ-Fe
than γ-Fe O
2 2O –CP. The
3 3–CP. The γ-Feγ-Fe O
2 2O33-FM exhibited regular spherical particles, smaller particle
exhibited regular spherical particles, smaller particle diameter, diameter,
larger
largersurface
surface area,
area, and
andbetter
betterpore
poreconnections.
connections. The facile
facile method
methodalsoalsoinhibited
inhibitedthe the formation
formation of of
thethe
α-Fe
α-Fe 2 O23Oin3 in thecatalyst,
the catalyst,which
whichpositively
positivelyinfluenced
influencedthethelow-temperature
low-temperature SCR SCR activity
activity and selectivity of
theofcatalyst.
the catalyst.
γ-Fe2 O γ-Fe 2O3-FM
3 -FM also contains
also contains a higher
a higher concentration
concentration of lattice
of lattice and surface-adsorbed
and surface-adsorbed oxygen,
oxygen, good reducibility, many acid sites, lower activation energy, more adsorption
good reducibility, many acid sites, lower activation energy, more adsorption of the reactants, and the of the reactants,
and theof
existence existence
unstable ofnitrates
unstableon nitrates on the surface.
the surface.

2. 2. Results
Results andDiscussion
and Discussion

2.1.2.1. Characterization
Characterization
Figure
Figure 2 shows
2 shows the the
X-rayX-ray diffraction
diffraction patterns
patterns of theofcatalysts
the catalysts synthesized
synthesized by twobydifferent
two different
methods
(FM and CP). It can be seen, from the sharp diffraction peaks at 2θ◦
(FM and CP). It can be seen, from the sharp diffraction peaks at 2θ equal 18.46 , 26.35 , 30.3626.35°,
methods equal ◦
18.46°, ◦ , 35.76◦ ,
30.36°,
37.2 ◦ 35.76°,
, 43.49 ◦ 37.2°,
, 53.93 ◦ 43.49°,, 63.17
, 57.49 ◦ 53.93°, (JCPDS
◦ 57.49°, 63.17° (JCPDS
39-1346), that39-1346),
the mainthat the main
phase phase
present present
in the γ-Fein the
2 O3 -CP
γ-Fe 2O3-CP was maghemite (γ-Fe2O3) with good crystallinity. Whereas, 2θ ◦ equal
was maghemite (γ-Fe2 O3 ) with good crystallinity. Whereas, 2θ equal 24.2 , 33.2 , and 41.2 coincide◦ 24.2°, 33.2°,
◦ and
41.2°
well with coincide well with the
the characteristics characteristics
peaks of hematitepeaks
(α-Fe2of
O3hematite
) (JCPDS (α-Fe 2O3) (JCPDS 89-8104), which
89-8104), which suggests that some
suggests that some spots of α-Fe2O3 are present in the sample of γ-Fe2O3-CP. In comparison, all the
spots of α-Fe2 O3 are present in the sample of γ-Fe2 O3 -CP. In comparison, all the diffraction peaks
diffraction peaks of the γ-Fe2O3-FM corresponded well with the maghemite (γ-Fe2O3) (JCPDS 39-
of the γ-Fe2 O3 -FM corresponded well with the maghemite (γ-Fe2 O3 ) (JCPDS 39-1346), avoiding the
1346), avoiding the formation of the α-Fe2O3, which can be beneficial to promote the NH3-SCR
formation of the α-Fe2 O3, which can be beneficial to promote the NH3 -SCR denitration performance.
denitration performance.

Figure 2. XRD spectra of the γ-Fe2 O3 catalysts.

Figure 2. XRD spectra of the γ-Fe2O3 catalysts.
Catalysts 2019, 9, 1018 4 of 17

The diffraction peak intensities of the γ-Fe2 O3 -FM are higher than the γ-Fe2 O3 -CP, also the
shapes of the peaks were sharper. It is evident from the Scherrer formula (Equation (6)) that the
crystallite size of a catalyst and half-width of diffraction peak is inversely proportional to each other.
Therefore, the crystallite size, as well as the crystallinity of the γ-Fe2 O3 -CP was higher than that of
γ-Fe2 O3 -FM, which might be due to an increase in the formation of α-Fe2 O3 as well as the difference of
the preparation method. Thus, the facile method using the DMF solvent can avoid the creation of the
crystalline phase of α-Fe2 O3 , which resulted in a higher BET surface area of the sample (Table 1).
Table 1 shows the values of specific surface areas, pore volume, and pore size of the catalysts.
The results demonstrated that the specific surface area and the specific pore volume of the γ-Fe2 O3 -FM
were 3.45 and 1.39 times larger than that of γ-Fe2 O3 -CP, respectively. Meanwhile, a difference of around
7.30 nm could be seen in the average pore diameter of the two samples, γ-Fe2 O3 -FM having a lower
value (5.419 nm). The number of active sites could be achieved by attaining a higher surface area and
pore volume of the catalyst. It can be concluded that this facile method using DMF as a solvent can
aid to achieve a decrease in average pore diameter as well as the particle size of the γ-Fe2 O3 catalyst.
These results could easily be verified by the TEM images.

Table 1. Textural parameters of the catalysts.

Pore Volume Pore Diameter

Samples SBET (m2 .g−1 ) a Particle Size (nm)
(cm3 .g−1 ) b (nm) c
γ-Fe2 O3 -FM 194.91 0.271 5.419 10.30
γ-Fe2 O3 -CP 56.4 0.195 12.72 84.8
γ-Fe2 O3 -MACP [16] 29.4 0.13 17.1 92.6
a BET surface area. b BJH desorption pore volume. c BJH desorption pore diameter.

N2 adsorption-desorption curves of the two catalysts measure at the temperature of liquid N2 are
shown in Figure 3A. When comparing these isotherms with the IUPAC classification, the isotherm of
the γ-Fe2 O3 -CP resembled type II isotherm [26] dedicated to macroporous materials. Additionally,
the type of hysteresis loop resembled with H3 type, which demonstrated that a slit-shaped pore
structure could be present in the catalyst [27]. However, the isotherm of the γ-Fe2 O3 -FM resembled
type IV isotherms [28]. The type IV isotherms were characterized as mesoporous materials ranging
from 2 to 50 nm. The type of hysteresis loop resembled with H2 type, which demonstrated that the
pore structure could be present in an ink bottle-shaped structure [29].
If we compare the closure points of the hysteresis loop for the catalysts, we can see that the closure
points of the γ-Fe2 O3 -CP are at higher P/Po as compare to that of γ-Fe2 O3 -FM, which indicate that
number of micropores (<2 nm) are present in the γ-Fe2 O3 -FM. Moreover, lower P/Po closure points of
the hysteresis loops for the γ-Fe2 O3 -FM indicated that the average pore size of the γ-Fe2 O3 -FM was
lower than the γ-Fe2 O3 -CP. Meanwhile, it can be seen from Table 1 that the particle sizes and surface
areas of the two catalysts were inversely proportional to each other. Therefore, more pore volume and
inner surface area of the catalyst can help NH3 -SCR reaction.
Figure 3B shows the pore size distribution of the catalysts. The results showed that the γ-Fe2 O3 -FM
pore size was distributed from 2 to 8 nm, it also has a well-developed pore structure. However,
the pore size of the γ-Fe2 O3 -CP has shown no peaks between 2 and 8 nm. The first peak could be
seen at 12.20 nm, and then two small peaks are seen from 15 to 50 nm, indicating a reduction in its
pore structure. It can be determined that the coprecipitation method was playing a part in destroying
the pore structure of the catalyst, especially between 2 and 8 nm. From the results demonstrated in
Figure 3, it can be determined that the NH3 -SCR activity is enhanced when the quantity of the pores
increases between 2 to 8 nm. γ-Fe2 O3 -FM containing pores 2–8 nm in size and a well-developed
structure exhibited better NH3 -SCR activity (approximately 10% higher at 230 ◦ C) than γ-Fe2 O3 -CP
containing more pores within 15 to 50 nm. As well, in γ-Fe2 O3 -CP with the shifting of pore volume
peak right word and increase in the pores within 15 to 50 nm, the number of available active sites
decreases due to the collapse of the pore structure which adversely affected the NH3 -SCR activity.
Catalysts 2019, 9, 9,
Catalysts 1018 55ofof517
of 17
Catalysts2019,
2019, 9,xxFOR
FORPEER
PEERREVIEW
REVIEW 17

Figure 3. γ-Fe2 O3 catalyst (A) N2 adsorption–desorption; (B) distribution of pore size.

Figure
Figure3.3.γ-Fe
γ-Fe2O 3 catalyst (A) N2 adsorption–desorption; (B) distribution of pore size.
2O3 catalyst (A) N2 adsorption–desorption; (B) distribution of pore size.

XPS analysis was carried out to characterize the oxidation state of the surface elements on the
XPS
XPS analysis
analysis was
wascarried
carriedout out to characterize the
theoxidation state of the
thesurface
surfaceelements on
onthe
iron (III) oxide catalysts (Figure 4). Ittocan characterize
be seen from oxidation
Figure 4Astate thatof electron-binding elements
energies the
of Fe
iron
iron (III)
(III)oxide
oxide catalysts
catalysts (Figure
(Figure 4).
4). It
It can
can be
be seen
seen from
from Figure
Figure 4A
4A that
that electron-binding
electron-binding energies
energies of
ofFe
Fe
2p3/2 (710.6–710.7 eV) and Fe 2p1/2 (724.3–724.4 eV) correspond well to Fe (III) [30], and the binding
2p
2p3/2 (710.6–710.7 eV) and Fe 2p1/2 (724.3–724.4 eV) correspond well to Fe (III) [30], and the binding
3/2 (710.6–710.7 eV) and Fe 2p1/2 (724.3–724.4 eV) correspond well to Fe (III) [30], and the binding
energy
energy values ofofFeFe2p for γ-Fe 2 O33-FM
-FM werewere slightly
slightlyless lessthan
thanγ-Fe
γ-Fe 2 3O3 -CP, maybe due to the existence
energyvalues
values of Fe2p 2pfor
forγ-Fe
γ-Fe2O 2O -CP, maybe
2O3-FM were slightly less than γ-Fe2O3-CP, maybe due to the existence
due to the existence
ofof
hematite phases (α-Fe 2 O33)) inin the
the γ-Fe
γ-Fe22O O33-CP
-CPcatalyst.
catalyst.FigureFigure 4B4B shows thethe deconvolution of the O
ofhematite
hematitephases phases(α-Fe
(α-Fe2O shows
2O3) in the γ-Fe2O3-CP catalyst. Figure 4B shows the deconvolution
deconvolution of
ofthe
theOO
2− (denoted
1s1s
spectrum. The binding energy at 529.9–530.0 eV corresponds to the lattice oxygen O by
1sspectrum.
spectrum.The Thebinding
bindingenergy
energyat at529.9–530.0
529.9–530.0eV eVcorresponds
correspondsto tothe
thelattice
latticeoxygen
oxygenOO2−(denoted
(denotedby
2−
by
Oβ),
Oβ),whereas,
whereas, thethe
sub-bands
sub-bands at higher
at higherbinding energy
binding (531.1–531.4)
energy (531.1–531.4) were assigned
were to surface
assigned to adsorbed
surface
Oβ), whereas, –the sub-bands at higher binding energy (531.1–531.4) were assigned to surface
oxygen
adsorbed such as O or O 2– which correspond to defect oxide/hydroxyl-like group (denoted(denoted as Oα) [31].
adsorbedoxygen oxygensuch
such2asasOO–or orOO222–which
whichcorrespond
correspondto todefect
defectoxide/hydroxyl-like
oxide/hydroxyl-likegroup
– 2–
group (denoted
Surface
as Oα) chemisorbed
[31]. Surface oxygen
chemisorbed (Oα) is very
oxygen active
(Oα) inis oxidation
very active
as Oα) [31]. Surface chemisorbed oxygen (Oα) is very active in oxidation reactions because reactions
in because
oxidation its mobility
reactions becauseis higher
its
its
than the
mobility lattice
is oxygen
higher than(Oβ)
the [32],
lattice and
oxygenthe high
(Oβ) SCR
[32], activity
and thecould
high
mobility is higher than the lattice oxygen (Oβ) [32], and the high SCR activity could be correlated be
SCR correlated
activity with
could the
be high content
correlated
ofwith
withthe thehigh
chemisorbed highcontent
contentof
oxygen onchemisorbed
of the catalyst’s
chemisorbed oxygen
surface.
oxygen on
onthethecatalyst’s
The ratios of surface.
catalyst’s surface. The
Oα/(Oα + Oβ)
The ratios of
over
ratios Oα/(Oα
ofγ-Fe ++Oβ)
2 O3 -FM
Oα/(Oα Oβ)and
over
γ-Fe
overO γ-Fe
-CP 2O 3-FM
are and
33.59% γ-Fe
and 2O 3-CP are
31.86%, 33.59%
respectively. and 31.86%,
This respectively.
indicated the
2 3γ-Fe2O3-FM and γ-Fe2O3-CP are 33.59% and 31.86%, respectively. This indicated the existence of This
existence indicated
of more the existence
hydroxyls/defect of
more
oxide
more onhydroxyls/defect
the γ-Fe2 O3 -FMoxide
hydroxyls/defect oxide on
catalyst’son the γ-Fe
γ-Fe2O
surface.
the 3-FM
In
2O the catalyst’s
3-FM SCR surface.
reaction,
catalyst’s these
surface. In the
the SCR
SCR reaction,
Inhydroxyls/defect
reaction, these
oxides
theseare
hydroxyls/defect
supposed oxides
to offer acid
hydroxyls/defect oxides are
sites
aretosupposed
the ammonia
supposed to offer acid
to offermolecule sites
acid sitesto to the ammonia
to adsorb
the ammonia molecule
in themolecule to
to adsorb inthe
adsorb
form of coordinated in NH3
the
form
and NH of4+coordinated
. NH
form of coordinated NH3 and NH . 3 and NH4+4+ .

Figure 4.4.XPS
Figure4.
Figure XPSspectra
XPS spectraof
spectra of(A)
(A)Fe
Fe2p
Fe 2pand
2p and(B)
and (B)OOO
(B) 1s.
1s.
1s.

InIn
Inorder
order
order toto have
have
to further
havefurther insight
furtherinsight into
intothe
insightinto theNO
the NOconversion
NO conversionof
conversion ofthe
of thecatalysts,
the catalysts,
catalysts, the H
the
the -TPR
H2H experiments
2 -TPR
2-TPR experiments
experiments
of γ-Fe
of γ-Fe
of γ-Fe O2O3-FM and γ-Fe2O3-CP catalysts were conducted. Generally, the hydrogen consumption
2 O
-FM
3 -FMand andγ-Feγ-FeO 2 O
-CP
3 -CP catalysts
catalysts were
were conducted.
conducted. Generally,
Generally, the the hydrogen
hydrogen consumption
consumption peaks
2 3 2 3
peaks
peaksat
located located
located
250–500 at
at◦250–500
250–500 °C
°C are
C are attributed are attributed
to surfaceto
attributed to surface
surface
lattice lattice oxygen,
latticeand
oxygen, peaksand
oxygen, and peaks
peaks at
at 500–800 ◦C
at 500–800
500–800 °C
°C are
are to
are attributed
attributed
attributed toto bulk
bulk lattice
lattice oxygen
oxygen [33].
[33]. ItItcan
canbe
be seen
seen in
inFigure
Figure 5A
5A that
that two
two well-separated
well-separated reduction
reduction
bulk lattice oxygen [33]. It can be seen in Figure 5A that two well-separated reduction peaks were
peaks
peakswere
werepresent
presentin inthe
theH H2-TPR profiles of the catalysts. Surface lattice oxygen was credited with
2-TPR profiles of the catalysts. Surface lattice oxygen was credited with
present in the H2 -TPR profiles of the catalysts. Surface lattice oxygen was credited with the reduction
process of maghemite to magnetite (γ-Fe2 O3 → Fe3 O4 ) in the low-temperature peak, and the bulk
Catalysts 2019, 9, 1018 6 of 17

Catalysts 2019, 9, x FOR PEER REVIEW 6 of 17

lattice oxygen was responsible for the reduction of magnetite to metallic iron (Fe3 O4 → Fe) in the
the reduction process of maghemite to magnetite (γ-Fe2O3 → Fe3O4) in the low-temperature peak, and
◦ C and higher, two heavily
high-temperature peak [33]. In actuality, with a sample heating rate of 5.5
the bulk lattice oxygen was responsible for the reduction of magnetite to metallic iron (Fe 3O4 → Fe)
overlapped peaks could be peak
in the high-temperature observed in actuality,
[33]. In the high-temperature
with a sampleregionheating(asrate
shown
of 5.5 inoCFigure 5A). That
and higher, twois
whyheavily
the reduction of magnetite was believed to happen in a two-step magnetite
overlapped peaks could be observed in the high-temperature region (as shown in Figure 5A). reduction sequence,
i.e., That
2Fe3 O → 6FeO
is 4why → 6Fe [34].
the reduction γ-Fe2 O3 -CP
of magnetite was was characterized
believed to happen by in athree different
two-step reduction
magnetite peaks
reduction
at 355 ◦ ◦
C, 624i.e.,C,2Fe
and ◦ ◦
sequence, 3O772
4 → 6FeOC, corresponding
→ 6Fe [34]. γ-Fe to2Othe reduction
3-CP of Fe2 O3 to
was characterized by Fe 3 O4different
three (355 C),reduction
Fe3 O4 to
(624 ◦atC),
FeOpeaks and FeO to Fe (772 ◦ C) [33,34]. However, the results exhibited that the corresponding
355 °C, 624 °C, and 772 °C, corresponding to the reduction of Fe2O3 to Fe3O4 (355 °C), Fe3O4
temperatures
to FeO (624to°C),theandreduction
FeO to peaks
Fe (772have shiftedHowever,
°C) [33,34]. to lower values for the
the results γ-Fe2 Othat
exhibited 3 -FM,thewhich indicated
corresponding
higher mobility of
temperatures tothe
the oxygen
reduction species
peaks inhavetheshifted
γ-Fe2 O -FM. As
to3lower the reducibility
values for the γ-Fe2O can
3-FM,be which
indicated by the
indicated
higher peak
reduction mobility of the oxygen
temperature, thespecies in the γ-Fe2peak
lower reduction O3-FM. As the reducibility
temperature indicated canstronger
be indicated by the
reducibility.
reduction
Therefore, thepeak temperature,
catalytic activity of theγ-Fe
lower
2 O3reduction
-FM was betterpeak temperature
than that ofindicated
γ-Fe2 O3 -CP. stronger reducibility.
Therefore,
One of thethe catalytic
main activity
processes of of
theγ-Fe
NH2O 3-FM was better than that of γ-Fe2O3-CP.
3 -SCR reaction is the adsorption and activation of ammonia
on the acid Onesites
of thethat
main existprocesses of the NHsurface.
on a catalyst’s 3-SCR reaction
In orderis the
to adsorption
examine this, andNH activation of ammonia
3 -TPD experiments
on the acid sites that exist on a catalyst’s surface. In order to examine
were conducted. It can be seen from Figure 5B that the samples have three NH3 desorption peaks this, NH 3-TPD experiments

from were conducted.

50–500 ◦ C. It isIt can be seen that
suggested fromionic
FigureNH 5B4+that
arethe samplestohave
ascribed three NH
Brønsted acid3 desorption peaks
sites, which are
from 50–500 °C. It is suggested that ionic NH4+ are ascribed to Brønsted acid sites, which are
considered thermally less stable than the coordinated NH3 molecules linked to the Lewis acid sites [33].
considered thermally less stable than the coordinated NH3 molecules linked to the Lewis acid sites
Consequently, it can be concluded that the adsorption peaks at lower temperatures (118, and 134.8 ◦ C)
[33]. Consequently, it can be concluded that the adsorption peaks at lower temperatures (118, and
are assigned to weak acid sites (weakly bonded NH3 ), the medium temperature peaks (197, 231.7 ◦ C)
134.8 °C) are assigned to weak acid sites (weakly bonded NH3), the medium temperature peaks (197,
can be assigned to both Lewis and Bronsted acid sites, and the adsorption peaks at higher temperatures
231.7 °C) can be assigned to both Lewis and Bronsted acid sites, and the adsorption peaks at higher
(294 and 343.2 ◦ C) can be ascribed to Lewis acid sites. Therefore, NH3 -TPD results indicated the
temperatures (294 and 343.2 °C) can be ascribed to Lewis acid sites. Therefore, NH3-TPD results
presence of both
indicated Lewis and
the presence Brønsted
of both Lewis acid sites in the
and Brønsted acidcatalysts forcatalysts
sites in the ammonia foradsorption. Note that
ammonia adsorption.
the peak area of γ-Fe 2 O 3 -FM Lewis acid sites and weak Brønsted acid
Note that the peak area of γ-Fe2O3-FM Lewis acid sites and weak Brønsted acid sites is obviouslysites is obviously greater than
those of γ-Fe
greater than O -CP. The peak area implies the amount of ammonia adsorption
2 3those of γ-Fe2O3-CP. The peak area implies the amount of ammonia adsorption in the in the sample [34], this
indicated
samplethat γ-Fe
[34], 2 Oindicated
this 3 -CP has lessthatnumber
γ-Fe2O3-CP of acid
hassites
less than
numberthe γ-Fe 2 O3 -FM.
of acid sites Therefore,
than the γ-Fe the 2catalytic
O3-FM.
activity and selectivity
Therefore, the catalytic of the 2 O3 -FM
γ-Feand
activity were higher
selectivity than2Othe
of the γ-Fe 3-FMγ-Fe 2 O3higher
were -CP. than the γ-Fe2O3-CP.

Figure 5. (A) H2-TPR. (B) NH3-TPD of the catalysts.

Figure 5. (A) H2 -TPR. (B) NH3 -TPD of the catalysts.

Figure 6 shows the TEM images of the γ-Fe2O3 samples prepared by two different methods, it
Figure 6 shows the TEM images of the γ-Fe2 O3 samples prepared by two different methods, it can
can be seen that the γ-Fe2O3 catalysts synthesized by different methods contain different
be seen that the γ-Fe2 O3 catalysts synthesized by different methods contain different microstructures.
microstructures. γ-Fe2O3-FM showed fine spherical particles with relatively uniform particle
γ-Fe2 O3 -FM showed fine spherical particles with relatively uniform particle diameter (Figure 6A–C,
diameter (Figure 6A–C, respectively). HRTEM characterizations demonstrated lattice fringes of the
respectively). HRTEM characterizations demonstrated lattice fringes of the obtained ferrite (Figure 6)
obtained ferrite (Figure 6) and the interfringe distance shown in Figure 6E is 0.252 nm, which
and the interfringe distance shown in Figure 6E is 0.252 nm, which corresponds well to {311} planar
corresponds well to {311} planar space of γ-Fe2O3 nanoparticles. Meanwhile, (Figure 6I) gives EDS
space of of
data γ-Fe 2 O23Onanoparticles.
γ-Fe Meanwhile, (Figure 6I) gives EDS data of γ-Fe2 Omay
3-FM nanoparticles and the Cu peaks were shown in the results 3 -FMbenanoparticles
due to the
andcopper
the Cu net
peaks
usedwere shown
during the in the results The
experiments. mayelemental
be due toratio
the copper
of Fe:Onet used duringtothe
is determined be experiments.
32.89:61.50,
Thewhich
elemental ratio
is very of Fe:O
near to theisstoichiometry
determined toofbeγ-Fe
32.89:61.50, which is very near to the stoichiometry of
2O3, which further approves that the structure and
O , which
γ-Fecomposition
2 3 further approves that the structure and composition
are coincident with the chemical formulation of γ-Fe2O are coincident
3. Figure with thethe
6J exhibited chemical
same
formulation of γ-Fe
particle size 2 O3 . Figure
distribution 6J exhibited
obtained from the the
TEM same particle size
micrograph, distribution
and it can be seenobtained from the
that the uniform
Catalysts 2019, 9, 1018 7 of 17

TEM micrograph,
Catalysts and
2019, 9, x FOR it can
PEER be seen that the uniform particle distribution is shown by the histogram.
REVIEW 7 of 17
The Gaussian fit was used to find mean particle sizes, which comes out near 10 ± 1 nm, which is
particle distribution
comparable with XRDislinewidth
shown byresults.
the histogram. The Gaussian fit was used to find mean particle sizes,
which comes out near 10 ± 1
In addition, the particles of γ-Fenm, which is comparable with XRD linewidth results.
2 O3 -FM are relatively uniform and regular spherical particles;
In addition, the particles of γ-Fe 2O3-FM are relatively uniform and regular spherical particles; an
an obvious decrease in particle diameter and better pore connection could be observed (Figure 6 A–C,
obvious decrease in particle diameter and better pore connection could be observed (Figure 6 A–C,
respectively). However, γ-Fe2 O3 -CP demonstrated irregular particle morphology, due to the sintering
respectively). However, γ-Fe2O3-CP demonstrated irregular particle morphology, due to the sintering
phenomenon, the existence of the needle-like particles, slice particles, as well as aggregated particles in
phenomenon, the existence of the needle-like particles, slice particles, as well as aggregated particles
the TEM of γ-Fe2 O3 -CP. In addition during the sintering phenomenon, the phenomenon of interparticle
in the TEM of γ-Fe2O3-CP. In addition during the sintering phenomenon, the phenomenon of
conglutination will become worse, and in the meantime, the pore voids are completely plugged due
interparticle conglutination will become worse, and in the meantime, the pore voids are completely
to plugged
the breakdown of the pore structure. Spherical particles can offer a larger specific surface area as
due to the breakdown of the pore structure. Spherical particles can offer a larger specific
compared
surface areato theasneedle
compared particles
to theand sliceparticles
needle particlesand
[16],slice
which will increase
particles the adsorption
[16], which sites
will increase theon
theadsorption
surface ofsitesthe catalyst. In addition, mass transfer and diffusion can be ensured by the better
on the surface of the catalyst. In addition, mass transfer and diffusion can be ensured pore
connections, which will influence the NH -SCR reaction positively.
by the better pore connections, which will3 influence the NH3-SCR reaction positively.

Figure TEM
6. 6.
Figure images
TEM images(A–C)
(A–C)γ-Fe 2 O23O
γ-Fe –FM. (F–H) γ-Fe2 O32-CP.
3–FM. (F–H) γ-Fe
(D)(D)
O3-CP. SAED. (E) HRTEM
SAED. image.image.
(E) HRTEM (I) EDX.
(J) (I)
Grain size distribution of γ-Fe O -FM nanoparticles.
EDX. (J) Grain size distribution of γ-Fe2O3-FM nanoparticles.
2 3

Table 2 shows the EDS analysis of the two samples. As expected, the major elements in the ferric
Table 2 shows the EDS analysis of the two samples. As expected, the major elements in the ferric
oxide samples are Fe and O (some proportion of Cu was present because it was used to guarantee
oxide samples are Fe and O (some proportion of Cu was present because it was used to guarantee
thethe
electrical
electrical conductivity
conductivityofofthese
thesesamples,
samples, and
and CC was presentfrom
was present fromthe
thegelatinous
gelatinoussubstrate).
substrate). It can
It can
be be
seen
seen from Table 2 that the O/Fe atomicity ratio over γ-Fe22O33-FM surface was 1.86, whereas, in thethe
from Table 2 that the O/Fe atomicity ratio over γ-Fe O -FM surface was 1.86, whereas, in
γ-Fe 2 O23O-CP
γ-Fe 3-CP, ,the
theratio
ratiowas
was1.45.
1.45. This
This indicated that the
indicated that the crystal
crystalsurface
surfaceofofthe
theγ-Fe
γ-Fe2O2O 3 -FM
3-FM
could
could be be
exposed by more number of lattice oxygen sites because of the plentiful adsorbed
exposed by more number of lattice oxygen sites because of the plentiful adsorbed oxygen and latticeoxygen and lattice
oxygen
oxygen onon thethecatalyst’s
catalyst’ssurface,
surface,the
thecatalyst’s
catalyst’s surface oxidizabilitywill
surface oxidizability willenhance,
enhance,which
which in in turn
turn willwill
increase
increase thetheNH NH3 -SCR
3-SCR activity.
activity.
Catalysts 2019, 9, 1018 8 of 17

Catalysts 2019, 9, x FOR PEER REVIEW 8 of 17

Table 2. The EDS analysis of the catalysts.
Table 2. The EDS analysis of the catalysts.
Percentage by Atomicity (at%) Percentage by Weight (wt%) O/Fe
Sample Percentage by Atomicity Percentage by Weight
O/Fe Weight
Atomicity
Fe O (at%)C Cu Fe O(wt%) C Cu
Sample Ratio Ratio
Atomicity Weight
γ-Fe2 O3 -FM 32.89Fe 61.50O C
1.38 Cu
1.21 Fe
61.15 O
32.76 C
0.55 Cu
2.57 1.86 0.53
Ratio Ratio
γ-Fe2 O3 -CP 39.79 57.68 1.39 0.84 66.69 27.56 0.52 2.56 1.45 0.42
γ-Fe2O3-FM 32.89 61.50 1.38 1.21 61.15 32.76 0.55 2.57 1.86 0.53
γ-Fe2O3-CP 39.79 57.68 1.39 0.84 66.69 27.56 0.52 2.56 1.45 0.42
2.2. Low-Temperature NH3 -SCR Performance
2.2.The
Low-Temperature NH3-SCR
low-temperature Performance
SCR removal activity and N2 selectivity of the two maghemite samples are
shown Thein Figure 7. The NHSCR
low-temperature 3 -SCR performance
removal of γ-Fe
activity and 2 O3 -FM isof
N2 selectivity obviously better than
the two maghemite γ-Fe2 Oare
samples 3 -CP.
The NO conversion of γ-Fe O -FM reached approximately 96% at 230 ◦ C, and the N selectivity
shown in Figure 7. The NH23-SCR 3 performance of γ-Fe2O3-FM is obviously better than γ-Fe 2 2O3-CP.
remained
The NOabove 90% up
conversion 2902O◦ C
oftoγ-Fe (Figure
3-FM 7A,B).
reached Dong et al. previously
approximately 96% at 230reported
°C, and the
the catalytic activity
N2 selectivity
of remained
the maghemite
abovecatalyst
90% up prepared by a microwave-assisted
to 290 °C (Figure coprecipitation
7A,B). Dong et al. previously method
reported (γ-Fe2 Oactivity
the catalytic 3 -MACP),
andofthe
theresults
maghemite
are alsocatalyst
plottedprepared
in Figure by 7A aformicrowave-assisted
comparison. In thecoprecipitation
study of Dong et method
al. [16],(γ-Fe O3-
the 2GHSV
wasMACP),
lower and
thanthe
ourresults are also plotted
experiment, still theinNOFigure 7A for comparison.
conversion was lower In the that
than studyofofthe
Dong
γ-Feet2 O
al.3 -FM
[16], at
lowthetemperatures.
GHSV was lower than our experiment, still the NO conversion was lower than that of the γ-Fe2O3-
FM at low temperatures.

Figure 7. 7.(A)
Figure (A)Comparison
Comparison of theNO
of the NOconversion
conversion ofof
thethe catalysts.
catalysts. (B)
(B) N N2 selectivity
2 selectivity of theof the catalysts.
catalysts. (C)
(C)Influence
InfluenceofofHH O and SO on SCR activity of the catalysts at 230 ◦ C. (D) Arrhenius plots of NO
2O2 and SO2 on 2 SCR activity of the catalysts at 230 °C. (D) Arrhenius plots of NO
oxidation
oxidationrates.
rates.Reaction
Reaction conditions
conditions of ofused
usedexperiments:
experiments: 500500
ppmppmNO,NO,500 500
ppmppmNH3NH 3 , 3 vol%
, 3 vol% O2, N2O2 ,
N2balance,
balance,300300ppm
ppm −1
SOSO 2 (when
2 (when used),
used), and/or
and/or 5 vol%
5 vol% H2OH(when
2 O (when
used),used), total rate
total flow flow150rate 150 mL·min
mL.min −1, and ,
and GHSV
GHSV = 40,000
= 40,000 −1 −1
h .hReaction
. Reaction conditions
conditions of Dong
of Dong et al.:et1000
al.: 1000
ppm ppm NO, 1000
NO, 1000 ppm ppm
NH3, NH 3, 3 O
3 vol% vol%
2, N2O2 ,

N2balance,
balance,andandGHSV
GHSV = 30,000
= 30,000 −1 [16].
h−1 h[16].
Catalysts 2019, 9, 1018 9 of 17

The influence of SO2 and H2 O on the SCR activity of both samples was investigated at 230 ◦ C
(Figure 7C). To test SO2 tolerance, 300 ppm SO2 was added to the simulated flue gas after 40 min of
stable reaction. It can be seen that the NO conversion of the γ-Fe2 O3 -FM showed a negligible decrease
and the NO conversion still remained above 90% after 100 min. However, the NO conversion decreased
from 94% to 79% in 200 min, after that the NO conversion became almost stable. When SO2 feed was
stopped after 400 min, the NO conversion was restored to some extent. Compared with the previous
reports [35,36], our experiments exhibited that the γ-Fe2 O3 -FM had better tolerance to SO2 at 230 ◦ C in
one hour of added SO2 . However, the coexistence of H2 O and SO2 caused more catalyst deactivation
to the addition of a single gas (SO2 ), because H2 O can compete with the gaseous NH3 for the active
sites [35]. When 5 vol% H2 O and 300 ppm SO2 was added to the simulated flue gas after 40 min of
stable reaction, the synergy effects of the H2 O and SO2 could be seen on the NO conversion of the
γ-Fe2 O3 –FM. Similar variation trends in the decrease of NO conversion for γ-Fe2 O3 –CP could be seen
after the addition of H2 O and SO2 in the feed gas. However, the decrease in the catalytic activity was
more may be due to a decrease in the surface area of the sample. The results showed that the decrease
in activity by H2 O and SO2 are due to their competing adsorption with the reactant over the surface of
the catalyst.
Figure 7D shows the Arrhenius plot of NO oxidation over the two samples in the temperature
range of 140 to 290 ◦ C. Equation (5) was used to find out the NO oxidation rate of reaction in
this temperature range. The slop of the plot was used to find out the activation energy for NO
oxidation. With γ-Fe2 O3 -FM the activation energy (17 KJ·mol−1 ) is much lower than that of γ-Fe2 O3 -CP
(29 KJ·mol−1 ). Note that the value of the activation energy of γ-Fe2 O3 -FM is much lower not only
from γ-Fe2 O3 -CP (29 KJ·mol−1 ), but also some other catalysts like Ti0.9 Fe0.1 O2−δ , (27 KJ·mol−1 ) [37],
and MnOx /TiO2 (38 KJ·mol−1 ) [38]. It is also well known that activation energy is an important
parameter in the NH3 -SCR reaction, because lower the activation energy means to lower the reaction
temperature required for the reaction to take place. Therefore, γ-Fe2 O3 -FM shows its superior
NH3 -SCR performance for low temperatures. This result showed that under these conditions the
active sites generated on the catalysts may have changed; however, the NO2 produced in large
quantities at higher temperatures (300–350 ◦ C) did not play its part in the NH3 -SCR activity due to its
deactivation mechanism.

2.3. NO and NH3 Oxidation Activity

It was reported that, with the increase in the NO2 /NO molar ratio in the feeding gas, the fast
SCR reaction, NO+ NO2 +2NH3 →2N2 +3H2 O, could be promoted, which can significantly boost the
low-temperature SCR activity of the catalyst [15]. That is why separate NO oxidation activities of
the different samples were tested and results are demonstrated in Figure 8A. It can be seen that with
the rise in temperature the NO oxidation activities increased first up to 300 ◦ C and then decreased.
Some evident differences could be seen between the NO oxidation activity and NH3 -SCR activity of
the samples which could be ascribed to the role of ammonia during NH3 -SCR reaction; in real SCR
reaction, the presence of NH3 would raise the NO to NO2 conversion in comparison to the pure NO
oxidation. NO conversion to NO2 was verified to be a slow step for the SCR reaction, and the NO2
generated was quickly consumed in the existence of ammonia [39,40]. Besides, Figure 8B shows that
the NH3 oxidation of the samples increased with the rise in temperature. This increase in the NH3
oxidation at high temperatures is considered a side reaction, aiding in the formation of NO in the
NH3 -SCR system and also causing the deactivation of the catalysts at high temperatures [41]. Similar
to the NO oxidation ability, the NH3 oxidation activity of the γ-Fe2 O3 -FM was higher than γ-Fe2 O3 -CP.
Note that the catalytic activity and selectivity of two samples also increase in the same manner.
Catalysts 2019, 9, 1018 10 of 17
Catalysts 2019, 9, x FOR PEER REVIEW 10 of 17

Figure 8. 8.(A)
Figure (A)NO NOoxidation
oxidationactivity.
activity. (B)
(B) NH33 oxidation
oxidationactivity.
activity.Reaction
Reactionconditions:
conditions:
500500
ppm ppm
NONO
or or
500
500 ppm
ppm NH NH , 3vol%
3 ,33 vol%OO22,, N
N22 balance,
balance,total
totalflow
flowrate
rate150
150mL.min
mL·min −1,−1
and GHSV
, and = 40,000
= 40,000
GHSV h−1. h−1 .

2.4. InIn
2.4. Situ
SituDRIFTS Studyofofγ-Fe
DRIFTSStudy γ-Fe22O -FM
O33-FM

2.4.1. Ammonia
2.4.1. AmmoniaAdsorbents
Adsorbentsover
over γ-Fe O33-FM
γ-Fe22O -FM
The The reaction
reaction mechanism
mechanismover overSCRSCRcatalysts
catalystshashasalways
alwaysbeen
beena atopic
topicofofinterest
interestfor
forresearchers
researchersand
different
and different hypotheses were proposed by the researchers for the SCR mechanism. Tothe
hypotheses were proposed by the researchers for the SCR mechanism. To find findreaction
the
mechanism of the γ-Feof2 O
reaction mechanism 3 -FM
the γ-Fein2Ositu
3-FMstudies
in situ have been
studies haveperformed.
been performed.
It It
is is
known
known thatthat adsorption
adsorptionofofNH NH3 3ononLewis
Lewisacid
acidororBrønsted
Brønstedacid
acidsites
sitesforms
forms coordinated
coordinated NH33 or
NH or4+
NH in the
in4+the NHNH 3 -SCR
3-SCR reaction
reaction [42].
[42]. Then
Then the
the gaseous
gaseous or
or adsorbed
adsorbed NO
NO reacts
reacts with these
with thesecoordinated
coordinated
NH NH 3 or 4+
NH to produce
4+ produce nitrogen
nitrogenand andwater
waterasasa product. Therefore, thethe
catalyst’s surface acidity is a is
3 or NH a product. Therefore, catalyst’s surface acidity
vital factor
a vital factor for an NH33-SCR -SCR reaction.
reaction.The TheDRIFTS
DRIFTSspectra
spectraforfor
adsorption
adsorption of of
ammonia
ammonia at different
at different
temperatures
temperatures over
over γ-Fe
γ-Fe 2O3-FM are shown in Figure 9A. The band at 1745 and 1600 cm−1 −1
2 O3 -FM are shown in Figure 9A. The band at 1745 and 1600 cm
were assigned
were assigned
to coordinated NH 3 linked to the Lewis acid sites [43], and the band at approximately 1310 cm
to coordinated NH3 linked to the Lewis acid sites [43], and the band at approximately 1310 cm−1 can
−1 can
be
be linked to the symmetric deformation of NH 3 coordinatively linked to one type of Lewis acid sites
linked to the symmetric deformation of NH3 coordinatively linked to one type of Lewis acid sites [43].
[43]. Moreover,
Moreover, the band the band
at 1423 at 1423
and and
12601260
cm−1 cmare
−1 are ascribed to ionic NH4+ bound to the Brønsted acid
ascribed to ionic NH4+ bound to the Brønsted acid
sites [44].
sites [44].
The results demonstrated that NH3 species adsorbed over γ-Fe2O3-FM showed a slow decrease
The results demonstrated that NH3 species adsorbed over γ-Fe2 O3 -FM showed a slow decrease
and less amount of oxidized intermediates were present over the γ-Fe2O3 catalyst with the rise in
and less amount of oxidized intermediates were present over the γ-Fe2 O3 catalyst with the rise in
temperature. Some ammonia absorbents were present on the surface of the γ-Fe2O3 catalyst even at
temperature. Some ammonia absorbents were present on the surface of the γ-Fe2 O3 catalyst even at
higher temperatures (200–300◦°C). By taking into account the XPS and NH3-TPD results, we can say
higher temperatures (200–300 C). By taking into account the XPS and NH3 -TPDas
that the γ-Fe2O3-FM catalyst may have a number of defect oxides on its surface,
results, we can say
well as higher
that the γ-Fe 2 O 3 -FM catalyst may have a number of defect oxides
ammonia adsorption, which could be the reason of the better NH 3-SCR catalytic activity. on its surface, as well as higher
ammonia adsorption, which could be the reason of the better NH3 -SCR catalytic activity.
2.4.2. Nitric Oxide plus Oxygen Adsorption over γ-Fe2O3-FM
2.4.2. Nitric Oxide Plus Oxygen Adsorption over γ-Fe2 O3 -FM
In situ DRIFTS spectra of co-adsorption of NO + O2 over the catalysts are displayed in Figure 9B.
In situ
It can DRIFTS
be seen spectra
that the bandsofappeared
co-adsorption of NO
were less + O2 over
thermally theAcatalysts
stable. are displayed
band at 1630 in Figure 9B.
cm−1 was attributed
It can be seen that the bands appeared were less thermally stable. A band at 1630 cm −1 was attributed
to gaseous NO2 [45]. A broad band centered at 1230 cm can be linked to bridging nitrate species.
−1

to The
gaseous −1 can be linked to bridging nitrate species.
bandsNO 2 [45].1520,
at 1570, A broad band
and 1360 cmcentered at 1230 cm
−1 were attributed to bidentate nitrate, monodentate nitrate, and
The bands at 1570, 1520, and respectively−1
1360 cm were attributed totheirbidentate nitrate, monodentate
monodentate nitrite species, [46]. Additionally, intensities were reduced withnitrate,
the
and monodentate
rise in temperature nitrite
andspecies,
the peaksrespectively [46]. Additionally,
almost disappeared after 200 °Ctheir intensities
[41,45]. wereinto
By taking reduced with
account
thethese
rise results,
in temperature and the peaks
we can conclude almost disappeared
that adequate NO2 ad species after
were200 ◦ C [41,45].
formed by theBy taking into
oxidation of NOaccount
ad
species
these overwe
results, thecan
γ-Fe 2O3-FM that
conclude catalyst, accompanied
adequate NO2 adby the transformation
species were formed by of the
some species of
oxidation likeNO
ad species over the γ-Fe2 O3 -FM catalyst, accompanied by the transformation of some speciesbelike
bidentate nitrate into other species which were active in the reaction. These characteristics could
the possible
bidentate reasons
nitrate for the
into other excellent
species NHwere
which 3-SCRactive
activity
in of
thethe γ-Fe2O3These
reaction. -FM catalyst.
characteristics could be the
possible reasons for the excellent NH3 -SCR activity of the γ-Fe2 O3 -FM catalyst.
Catalysts 2019, 9, 1018 11 of 17
Catalysts 2019, 9, x FOR PEER REVIEW 11 of 17

Figure 9. 9.In
Figure Insitu diffuse reflectance
situ diffuse reflectanceinfrared
infrared Fourier
Fourier transform
transform spectroscopy
spectroscopy (DRIFTS)
(DRIFTS) spectraspectra
of (A) of
(A) NH3 adsorption. (B) NO + O
NH3 adsorption.
2 adsorptions over γ-Fe O
(B) NO + O2 adsorptions-FM.
2 3 over γ-Fe2O3-FM.

2.4.3. Ammonia
2.4.3. AmmoniaAd
AdSpecies
SpeciesReaction
Reaction with Nitrogen Oxides
with Nitrogen Oxides
ToTo examine
examinethe theadsorbed
adsorbedammonia
ammonia reaction with nitrogen
reaction with nitrogenoxideoxideover
overthetheγ-Fe γ-Fe2O23O 3 -FM
-FM catalyst,
catalyst,
thethe
NO + O flow was introduced to the ammonia pre-exposed catalysts
NO + O2 flow was introduced to the ammonia pre-exposed catalysts at 200 °C and then the
2 at 200 ◦ C and then the changes

in changes
the DRIFTS spectra
in the DRIFTS were recorded
spectra were with respect
recorded withtorespect
time (Figure
to time10A).
(FigureAfter treating
10A). the γ-Fethe
After treating 2 O3γ-
-FM
catalyst
Fe2O3-FM with NH3 /N
catalyst 2 gases,
with NH3/N a 2series
gases,of bandsoflinked
a series bands to Brønsted
linked acid sites
to Brønsted acid(1423, and 1260
sites (1423, cm−1 ),
and 1260
and −1
cmLewis
−1), and acid sitesacid
Lewis (3357, 1745,
sites 1600,
(3357, 1310
1745, cm 1310
1600, ) appeared in the DRIFTS
cm−1) appeared spectra spectra
in the DRIFTS of the γ-Fe 2 O3γ-
of the -FM
catalyst
Fe2O3-FM (Figure 10A)
catalyst [42]. The
(Figure 10A)intensities of these bands
[42]. The intensities reduced
of these bands quickly
reduced after
quickly passing the gases
after passing theNO
+O 2 /N2NO
gases and+ vanished
O2/N2 andcompletely
vanished completely
after 5 min after
of5exposure
min of exposure of the gases.
of the gases. And the Andbidentate
the bidentate
nitrate
nitrate(1570
species cm−1
species (1570
), trans-N O2 2− species
cm−1), 2trans-N 2O22- species cm−1 ),cm
(1450 (1450 −1), gaseous NO2 (1632−1
gaseous NO2 (1632 cm cm ), nitro
), −1nitro species
species (M- 2 )
(M-NO
NO ) −1
(1348 cm −1), nitrite species (1213 −1 cm −1) were formed [47]. This result implies that the NH3 ad
(1348 cm ), nitrite species (1213 cm ) were formed [47]. This result implies that the NH3 ad species
2

species
could react could reactwith
quickly quickly
the with the gasNO
gas phase phase NOx on the γ-Fe2O3-FM catalyst’s surface at◦200 oC.
x on the γ-Fe2 O3 -FM catalyst’s surface at 200 C.

2.4.4.
2.4.4. NitrogenOxides
Nitrogen OxidesAd
AdSpecies
Species Reaction
Reaction with
with Ammonia
Ammonia
TheThe transientstudies
transient studieswere
wereconducted
conducted toto examine
examinethetheadsorbed
adsorbednitrogen
nitrogenoxide species
oxide reactivity
species reactivity
in the NH 3-SCR reaction over the γ-Fe2O3-FM catalyst (Figure 10B). It is apparent that after turning
in the NH3 -SCR reaction over the γ-Fe2 O3 -FM catalyst (Figure 10B). It is apparent that after turning
ononthethe flowofofammonia
ammoniaover overthe
thecatalyst
catalyst peaks
peaks linked
linked to
to gaseous
gaseous NO
NO22 (1630
(1630cm
−1), monodentate
flow cm−1 ), monodentate
nitrate (1287 cm −1 −1 ), M-NO2 nitro compounds (1360 cm −1 −1) disappeared within 2 min, demonstrating
nitrate (1287 cm ), M-NO2 nitro compounds (1360 cm ) disappeared within 2 min, demonstrating
that these species were active in the reaction [48]. Afterward, ionic NH4+4+bound to the Brønsted acid
that these species were active in the reaction [48]. Afterward, ionic NH bound to the Brønsted acid
sites (1423 cm −1) and coordinated NH3 on Lewis acid sites (3347, 1313, 1602 cm −1) formed on the
sites (1423 cm ) and coordinated NH3 on Lewis acid sites (3347, 1313, 1602 cm−1 ) formed on the
−1
catalyst’s surface.
catalyst’s surface.
It can be seen from the in situ DRIFTS results that both Brønsted acid and Lewis acid sites were
present over the γ-Fe2 O3 -FM catalyst’s surface. However, a lesser amount of intermediates was found
on the γ-Fe2 O3 -FM catalyst’s surface, indicating that the adsorbed ammonia species were steady on
γ-Fe2 O3 –FM. In addition, unstable nitrates were present over the γ-Fe2 O3 -FM catalyst and more NO2 ad
species were formed by the oxidation of NO ad species, which promoted the SCR reaction. The DRIFTS
results exhibited that both NO and NH3 can be adsorbed on the γ-Fe2 O3 -FM catalyst’s surface. Adsorbed
NH3 species can react with gaseous NO+O2 or/and adsorbed NOx species. Therefore, over γ-Fe2 O3 -FM
NO reduction followed both the Eley-Rideal and Langmuir-Hinshelwood mechanisms.

Figure 10. In situ DRIFTS (A) in a flow of NO + O2 on adsorbed ammonia species and (B) a flow of
ammonia on adsorbed NO + O2 species at 200 °C over γ-Fe2O3-FM.
in the NH3-SCR reaction over the γ-Fe2O3-FM catalyst (Figure 10B). It is apparent that after turning
on the flow of ammonia over the catalyst peaks linked to gaseous NO2 (1630 cm−1), monodentate
nitrate (1287 cm−1), M-NO2 nitro compounds (1360 cm−1) disappeared within 2 min, demonstrating
that these species were active in the reaction [48]. Afterward, ionic NH4+ bound to the Brønsted acid
sites (1423 cm−1) and coordinated NH3 on Lewis acid sites (3347, 1313, 1602 cm−1) formed on the
Catalysts 2019, 9, 1018 12 of 17
catalyst’s surface.

Catalysts 2019, 9, x FOR PEER REVIEW 12 of 17

It can be seen from the in situ DRIFTS results that both Brønsted acid and Lewis acid sites were
present over the γ-Fe2O3-FM catalyst’s surface. However, a lesser amount of intermediates was found
on the γ-Fe2O3-FM catalyst’s surface, indicating that the adsorbed ammonia species were steady on
γ-Fe2O3–FM. In addition, unstable nitrates were present over the γ-Fe2O3-FM catalyst and more NO2
ad species were formed by the oxidation of NO ad species, which promoted the SCR reaction. The
DRIFTS results exhibited that both NO and NH3 can be adsorbed on the γ-Fe2O3-FM catalyst’s
surface. Adsorbed NH3 species can react with gaseous NO+O2 or/and adsorbed NOx species.
Therefore,
Figure 10. over
In situγ-Fe 2O3-FM
DRIFTS (A) inNO reduction
a flow followed
of NO + O both the Eley-Rideal and Langmuir-
2 on adsorbed ammonia species and (B) a flow of
Hinshelwood mechanisms.
ammonia on adsorbed NO + O2 species at 200 C over γ-Fe2 O3 -FM.
◦

3. Materials and
3. Materials Methods
and Methods

3.1.3.1.
Materials and
Materials Reagents
and Reagents
ToTo prepare
prepare thethe
γ-Fe 2 O2O
γ-Fe 3 3catalysts
catalystsby bycoprecipitation
coprecipitation method,
method, Iron Iron (II)
(II) Sulfate
Sulfate Heptahydrate
Heptahydrate
(FeSO
(FeSO 4 .7H 2 O)
4.7H (99.5%,
2O) (99.5%, Sigma
Sigma Aldrich)
Aldrich)was wasused
usedasasa aFeFeprecursor.
precursor.InInthe thefirst
firststep,
step,a acertain
certainamount
amountof
FeSO
of FeSO
4 .7H 4
2 O
.7H was
2 O dissolved
was in
dissolved the
in deionized
the deionized water
water totomake
make a a0.2
0.2M M solution.
solution. The
The solution
solution obtained
obtained
waswasthen stirred
then stirredforfor6060min
minatat2525◦ C.
°C.Next,
Next,aacertain
certain amount
amount of precipitator
precipitator was wasadded
addedinto intothe
thesolution
solution
dropwise
dropwise with with continuous
continuous stirring
stirring for h2–3
for 2–3 by h by maintaining
maintaining the pHthe pH solution
of the of the solution
around around 9–10.
9–10. Without
Without
aging, aging, theprecipitated
the resulting resulting precipitated
particles particles
are filtered areout
filtered
and outthen and then washed
washed thoroughly
thoroughly with the with DI
the DI water. Last, the precipitate was desiccated ◦ at 105 °C for 10 h
water. Last, the precipitate was desiccated at 105 C for 10 h and calcined at 220 C for 3 h in air. and calcined ◦ at 220 °C for 3 h in
air.To prepare the γ-Fe2 O3 nanoparticles with a facile method, Dimethyl Formamide (DMF) (99.5%,
To prepare
Sigma Aldrich) was the γ-Feas
used 2Oa3 nanoparticles
solvent to dissolve with athefacile
Ironmethod, Dimethyl
(III) Nitrate Formamide
Nonahydrate (Fe(DMF)
(NO3 )(99.5%,
3 ·9H2 O)
Sigma Aldrich) was used as a solvent to dissolve the Iron (III) Nitrate
(98.5%, Sigma Aldrich) to make a 0.6 molar solution. This solution was calcined in the presence of Nonahydrate (Fe (NO₃)₃·9H₂O)
air(98.5%,
with a Sigma
suitable Aldrich)
heatingtorate
make fora 20.6h molar
at 220solution. This γ-Fe
◦ C to obtain solution was calcined in the presence of air
2 O3 nanoparticles [49]. The schematic
with a suitable
representation heating
for the rate for of
preparation 2 the
h atγ-Fe
220 °C to obtain γ-Fe2O3 nanoparticles [49]. The schematic
2 O3 -FM by the facile one-step synthesis method is shown
representation for the preparation of the γ-Fe2O3-FM by the facile one-step synthesis method is shown
in Figure 11. Both the samples were crushed and sieved to 40–60 mesh size to examine their catalytic
in Figure 11. Both the samples were crushed and sieved to 40–60 mesh size to examine their catalytic
activities. DMF, as a polar aprotic solvent, aids the contact and diffusion of the reactant molecules in the
activities. DMF, as a polar aprotic solvent, aids the contact and diffusion of the reactant molecules in
way of volatilization. When the precursors are calcined at moderate temperature (200 ◦ C in our case),
the way of volatilization. When the precursors are calcined at moderate temperature (200 °C in our
a large area of γ-Fe2 O3 nucleation is crystallized to the uniform nanoparticles, which can be confirmed
case), a large area of γ-Fe2O3 nucleation is crystallized to the uniform nanoparticles, which can be
by TEM image and XRD spectrum of γ-Fe2 O3 –FM. DMF is a commendable solvent in comparison to
confirmed by TEM image and XRD spectrum of γ-Fe2O3–FM. DMF is a commendable solvent in
water, and DMF is associated well with the cations [50]. As a solvent, DMF could disperse the ions,
comparison to water, and DMF is associated well with the cations [50]. As a solvent, DMF could
and coats each
disperse the ion
ions,during the calcination
and coats each ion process
during the [51].calcination
Cao et al. prepared
process [51].maghemite
Cao et nanoparticles
al. prepared
with
maghemite nanoparticles with both water and DMF as a solvent. He also found that underconditions,
both water and DMF as a solvent. He also found that under the same experimental the same
α-Fe 2 O3 particles
experimental are produced
conditions, α-Fe2with water asare
O3 particles a solvent
produced butwith
it cannot
watergenerate in DMF
as a solvent but (also could
it cannot
be generate
seen in Figure 2) [49]. This illustrates that DMF may slower the reaction
in DMF (also could be seen in Figure 2) [49]. This illustrates that DMF may slower the process of calcination,
which restrains
reaction processthe oftransformation
calcination, which process of γ-Fethe
restrains 2 Otransformation
3 to α-Fe2 O3 , and this was
process better
of γ-Fe 2O3attopromoting
α-Fe2O3, and the
NHthis -SCR denitration performance sequentially.
3 was better at promoting the NH3-SCR denitration performance sequentially.

Figure 11. Schematic for the preparation of γ-Fe2 O3 -FM by facile one-step synthesis method.
Figure 11. Schematic for the preparation of γ-Fe2O3-FM by facile one-step synthesis method.

3.2. Experimental Setup and Governing Equations

The experimental setup included a simulated flue gas system, an electrically heated test rig, a
fixed-bed quartz reactor, and a flue gas analyzer system. The reactor was made of quartz glass (6mm
Catalysts 2019, 9, 1018 13 of 17

3.2. Experimental Setup and Governing Equations

The experimental setup included a simulated flue gas system, an electrically heated test rig,
a fixed-bed quartz reactor, and a flue gas analyzer system. The reactor was made of quartz glass
(6mm i.d × 500 mm length). A K-type thermocouple of 2 mm diameter with an accuracy of 2.5 ◦ C
was inserted into the reactor to obtain the data of the flue gas temperature from the reactor inlet and
outlet. The reaction gas consisted of 500 ppm NO, 500 ppm NH3 , 3 vol% O2 , 300 ppm SO2 (when used),
and/or 5 vol% H2 O (when used), N2 balance, the total mass flow rate of 150 mL·min−1 , and with
a corresponding gas hourly space velocity (GHSV) of 40,000 h−1 . Water vapors were generated by
passing N2 through a heated gas-wash bottle containing deionized water. Mass flow controllers
(MFCs) (CS200A, CS200D, Sevenstar, Beijing, China) were used to control the flow of simulated flue
gas. Due to the density difference, for the NH3 -SCR activity test 0.2 g of γ-Fe2 O3 -FM powder, and 0.1 g
of γ-Fe2 O3 -CP powder were used to obtain the consistent GHSV.
A flue gas analyzer (Testo 350, Lenzkirch, Germany) was used to constantly monitor the
concentrations of NO, NO2 , O2 . To test N2 selectivity, the outlet gas compositions were detected
by GC-14C with Porapak Q column (Shimadzu, Kyoto, Japan) and FT-IR spectrometer (Vertex70v,
Billerica, USA) with a scanning range of 4000 to 400 cm−1 , 0.15 cm−1 resolution, and an average of
32 scans for each spectrum. Using the concentration of gases at steady state, NO conversion and N2
selectivity were calculated according to the following equations.

[NO]in − [NO]out
NO conversion (%) = × 100% (1)
[NO]in
[NO]in + [NH3 ]in − [NO]out − [NH3 ]out − 2[N2 O]out
N2 selectivity (%) = × 100% (2)
[NO]in + [NH3 ]in − [NO]out − [NH3 ]out
where [NO]in represents the NO concentration at the inlet of the reactor (ppm) and [NO]out represents
the NO concentration at the outlet of the reactor (ppm).
Both NO and NH3 oxidation tests were conducted in the same reactor. For NO oxidation, the feed
gas consisted of 500 ppm NO, 3 vol% O2 , and balance N2 . For NH3 oxidation, the feed gas consisted
of 500 ppm NH3 , 3vol% O2 , and balance N2 . The total flow rate was 150 mL·min−1 . The NO to NO2
conversion percentage was calculated using the following equation.

[NH3 ]out
NO to NO2 conversion (%) = × 100% (3)
[NH3 ]in
The NH3 conversion percentage was calculated using the following equation.

[NH3 ]in − [NH3 ]out

NH3 conversion (%) = × 100% (4)
[NH3 ]in
where [NH3 ]in and [NH3 ]out represent the concentration of NH3 in inlet and outlet flue gases,
respectively (ppm).
With the assumption that the components of the reaction were free of limitations for diffusion,
the equation used to calculate the NO oxidation rate normalized by the specific surface area of the
catalyst is given as follows,
  XNO QCf
Rate mmol. m−2 . h−1 = (5)
Vm WSBET
where XNO represents the NO-to-NO2 conversion, Cf denotes the inlet concentration of NO (500 ppm),
Q represents the volumetric flow rate (mL·h−1 ), Vm is the molar volume of gas (22.4 mL·mmol−1 ),
SBET denotes the value of a specific area of the catalyst (m2 ·g−1 ), and W represents the weight of the
catalyst (g) [43].
The Scherrer formula is
D = 0.94 × λ/(β × Cosθ) (6)
Catalysts 2019, 9, 1018 14 of 17

where D represents the average diameter of particles, λ denotes the wavelength of incident X-ray, β is
the half-width of diffraction peak, and θ is the diffraction angle.

3.3. Characterization Used

X-ray powder diffraction (XRD) measurements were carried out to determine the crystalline
structures of the catalysts with CuKα radiation on a D8 Advance X-ray diffractometer (Bruker, Billerica,
USA). The scan rate of the diffraction pattern was 1◦ min−1 with a resolution of 0.02◦ , and the diffraction
pattern was taken in a 2θ range of 10 to 90◦ .
An Autosorb-IQ3 (Quantachrome; Anton Paar, Austria) analyzer was used to determine the
Brunauer–Emmett–Teller (BET) surface properties of the catalysts. To examine the surface characteristics
of the catalysts, the samples were undergone by N2 adsorption at 77 K; next, the samples were degassed
under vacuum for 12 h at 180 ◦ C.
X-ray photoelectron spectroscopy (XPS) analysis was performed on an AXIS Ultra DLD (Shimadzu
Kratos, Kyoto, Japan) X-ray photoelectron spectrometer with a spherical mirror and concentric
hemispherical detector operating at constant pass energy (PE = 46.95 eV). All binding energies (BE)
were referenced to the C1s line at 284.6 eV.
An AutoChem II 2920 (Micrometrics, Norcross, USA) instrument was used to conduct
temperature-programmed reduction (H2 -TPR) experiments. Each catalyst (100 mg) was placed
in a quartz U-tube reactor to conduct the experiment. The samples were pretreated in He at 100 ◦ C for
1 h before reduction and then cooled to a temperature of 50 ◦ C. Then, the samples were heated from
50–700 ◦ C with a heating rate of 5.5 ◦ C·min−1 and were simultaneously introduced to a mixture of
gases (90% Ar and 10% H2 ) with a flow rate of 0.03 L·min−1 . A thermal conductivity detector (TCD)
was used to determine the content of H2 in the effluent gas.
An AutoChem II 2920 (Micrometrics, Norcross, USA) instrument was used to conduct
temperature-programmed desorption (NH3 -TPD) experiments. Each catalyst (100 mg) was placed
in a quartz U-tube reactor to conduct the experiment. The samples were pretreated at 100 ◦ C in He
gas atmosphere with a flow rate of 0.03 L·min−1 for 2 h. After that, the samples were cooled to a
temperature of 50 ◦ C, and at this temperature, the samples were fed with a mixture of gases (10% NH3
and 90% He) until saturation. Then, He gas was used for cleaning the samples. After this, the samples
were heated to 550 ◦ C with a heating rate of 10 ◦ C·min−1 .
The morphology of the catalysts was examined by the transmission electron microscope (TEM,
Tecnai G2 spirit Biotwin with accelerating voltage of 120 kV) (FEI, Hillsboro, OR, USA), whereas
scanning electron microscope (SEM) Sirion 200 (Thermo Fisher Scientific, Waltham, MA, USA) was
used to obtain the energy-dispersive spectrum analysis (EDS) of the catalysts.
FTIR spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) (Nicolet 6700) (with a Harrick
DRIFTS cell having an MCT detector and ZnSe windows) was used to conduct in situ diffuse reflectance
infrared Fourier transform spectroscopy (DRIFTS) experiments. Liquid nitrogen was used to cool
the MCT detector. To simulate the real flue gas conditions, mass flow controllers (CS200A, CS200D,
Sevenstar, Beijing, China) and a temperature controller were used. The first step in each experiment
includes the degasification of the ground samples (placed in a ceramic crucible) with a flow of nitrogen
at 300 ◦ C for 60 min, then the background spectrum was recorded in the flow of nitrogen. Next,
this background spectrum was automatically subtracted from all the final spectrums of the particular
catalyst. To understand the adsorption studies of NH3 (or NO + O2 ), the catalysts were exposed to
a flow of NH3 (or NO + O2 ), the catalysts were purged with N2 for 30 min, and the spectra were
collected. Now, with the rise in temperature the spectra were collected at the desired temperatures,
whereas in the transient studies, after exposing the pretreated catalysts to a flow of NH3 (or NO +
O2 ) for 60 min and purging it with nitrogen for 30 min, the gas flow was switched to NO + O2 (or
NH3 ) to achieve the changes in the DRIFTS spectra with in real-time. Feeding gases flow rate was
kept at 150 mL·min−1 and the spectral resolution was 4 cm−1 with collecting 100 scans for the spectra,
with reaction conditions of 500 ppm NO, 500 ppm NH3 , 3 vol% O2 , and N2 balance.
Catalysts 2019, 9, 1018 15 of 17

4. Conclusions
This study compares the NH3 -SCR performance of the maghemite catalysts prepared by a facile
method and the conventional coprecipitation method. The resulted demonstrated that the γ-Fe2 O3
nanoparticles prepared by the facile method (γ-Fe2 O3 -FM) not only exhibited better NH3 -SCR activity
and selectivity than the catalyst prepared by the coprecipitation method but also showed improved
SO2 tolerance. The γ-Fe2 O3 -FM exhibited regular spherical particles, smaller particle diameter, larger
surface area, and better pore connections. The facile method also inhibited the formation of the α-Fe2 O3
in the catalyst, which positively influenced the low-temperature SCR activity and selectivity of the
catalyst. In addition, a higher concentration of lattice and surface-adsorbed oxygen, good reducibility,
a lot of acid sites, lower activation energy, better adsorption of the reactants, and the existence of
unstable nitrates on the surface of the γ-Fe2 O3 -FM catalyst were responsible for the excellent NH3 -SCR
performance of the γ-Fe2 O3 -FM catalyst.

Author Contributions: Conceptualization, E.W.; methodology, N.H. and E.W.; Characterization, N.H., S.F.;
and M.T.A.; investigation, N.H. and E.W.; writing—original draft preparation, N.H. and E.W.; writing—review
and editing, E.W., N.H, and M.T.A.; visualization, S.F. and N.H.; supervision, E.W; funding acquisition, E.W.
Funding: This work was supported by the National Natural Science Foundation of China (grant no. 50676057).
Acknowledgments: The authors would like to thank the National Natural Science Foundation of China for the
research funding.
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Liu, F.; Asakura, K.; He, H.; Liu, Y.; Shan, W.; Shi, X. Influence of calcination temperature on iron titanate
catalyst for the selective catalytic reduction of NOx with NH3 . Catal. Today 2011, 164, 520–527. [CrossRef]
2. Husnain, N.; Wang, E.; Li, K.; Anwar, M.T.; Mehmood, A.; Gul, M.; Li, D.; Jinda, M. Iron oxide-based catalysts
for low-temperature selective catalytic reduction of NOx with NH3 . Rev. Chem. Eng. 2019, 35, 239–264.
[CrossRef]
3. Husnain, N.; Wang, E.; Fareed, S. Low-temperature selective catalytic reduction of NO with NH3 over
natural iron ore catalyst. Catalysts 2019, 9, 956. [CrossRef]
4. Shen, B.; Liu, T.; Zhao, N.; Yang, X.; Deng, L. Iron-doped Mn-Ce/TiO2 catalyst for low temperature selective
catalytic reduction of NO with NH3 . J. Environ. Sci. 2010, 22, 1447–1454. [CrossRef]
5. Yang, S.J.; Wang, C.Z.; Ma, L.; Peng, Y.; Qu, Z.; Yan, N.Q. Substitution of WO3 in V2 O5 /WO3 -TiO2 by Fe2 O3
for selective catalytic reduction of NO with NH3 . Catal. Sci. Technol. 2013, 3, 161–168. [CrossRef]
6. Liu, Z.; Liu, Y.; Chen, B.; Zhu, T.; Ma, L. Novel Fe-Ce-Ti catalyst with remarkable performance for the
selective catalytic reduction of NOx by NH3 . Catal. Sci. Technol. 2016, 6, 6688–6696. [CrossRef]
7. Wang, X.; Gui, K. Fe2 O3 particles as superior catalysts for low temperature selective catalytic reduction of
NO with NH3 . J. Environ. Sci. 2013, 25, 2469–2475. [CrossRef]
8. Liu, C.; Yang, S.; Ma, L.; Peng, Y.; Hamidreza, A.; Chang, H. Comparison on the performance of α-Fe2 O3 and
γ-Fe2 O3 for selective catalytic reduction of nitrogen oxides with ammonia. Catal. Lett. 2013, 143, 697–704.
[CrossRef]
9. Lu, C.; Li, J.; Peng, Y.; Wang, D.; Gan, L.; Ma, Y. De-reducibility mechanism of titanium on maghemite
catalysts for the SCR reaction: An in situ DRIFTS and quantitative kinetics study. Appl. Catal. B Environ.
2017, 221, 556–564.
10. Xu, H.; Ni, K.; Li, X.; Fang, G.; Fan, G. Structural transformation of Pd-α-Fe2 O3 and Pd-γ-Fe2 O3 catalysts
and application in the CO oxidation reaction. RSC Adv. 2017, 7, 51403–51410. [CrossRef]
11. Ishaq, M.; Sultan, S.; Ahmad, I.; Ullah, H.; Yaseen, M.; Amir, A. Adsorptive desulfurization of model oil
using untreated, acid-activated and magnetite nanoparticle loaded bentonite as adsorbent. J. Saudi. Chem.
Soc. 2017, 21, 143–151. [CrossRef]
12. Lin, C.; Pehkonen, S.O. Aqueous free radial chemistry of mercury in the presence of iron oxides and ambient
aerosol. Atmos. Environ. 1997, 31, 4125–4137. [CrossRef]
Catalysts 2019, 9, 1018 16 of 17

13. Stevin, M.; Wagner, M.; Schmid, M.; Parkinson, G.S.; Diebodd, U. Surface point defects on bulk oxides:
Atomically-resolved scanning probe microscopy. Chem. Soc. Rev. 2017, 46, 1772–1784. [CrossRef] [PubMed]
14. Shi, X.; Liu, F.; Xie, L.; Shan, W.; He, H. NH3 -SCR performance of fresh and hydrothermally aged Fe-ZSM-5 in
standard and fast selective catalytic reduction reactions. Environ. Sci. Technol. 2013, 47, 3293–3298. [CrossRef]
15. Chen, Z.; Wang, F.; Li, H.; Yang, Q.; Wang, L.; Li, X. Low-Temperature selective catalytic reduction of NOx
with NH3 over Fe–Mn mixed-oxide catalysts containing Fe3 Mn3 O8 phase. Ind. Eng. Chem. Res. 2011, 51,
202–212. [CrossRef]
16. Wang, D.; Yang, Q.; Li, X.; Peng, Y.; Li, B.; Si, W. Preparation of γ-Fe2 O3 catalysts and their deNOx
performance: Effects of precipitation conditions. Chem. Eng. Technol. 2018, 41, 1019–1026. [CrossRef]
17. Vidal-Vidal, J.; Rivas, J.; López-Quintela, M.A. Synthesis of monodisperse maghemite nanoparticles by the
microemulsion method. Colloids Surf. A Physicochem. Eng. Asp. 2006, 288, 44–51. [CrossRef]
18. Zare, N.; Zabardasti, A.; Mohammadi, A.; Azarbani, F. Synthesis of spherical Fe3 O4 nanoparticles from the
thermal decomposition of iron (III) nano-structure complex: DFT studies and evaluation of the biological
activity. Bioorg. Chem. 2018, 80, 334–346. [CrossRef]
19. Aliahmad, M.; Nasiri, M.N. Synthesis of maghemite (γ-Fe2 O3 ) nanoparticles by thermal-decomposition of
magnetite (Fe3 O4) nanoparticles. Mater. Sci. Pol. 2013, 31, 264–268. [CrossRef]
20. Al-Gaashani, R.; Radiman, S.; Tabet, N.; Daud, A.R. Rapid synthesis and optical properties of hematite
(α-Fe2 O3 ) nanostructures using a simple thermal decomposition method. J. Alloys. Compd. 2013, 550, 395–401.
[CrossRef]
21. Ismail, A.A. Synthesis and characterization of Y2 O3 /Fe2 O3 /TiO2 nanoparticles by sol-gel method. Appl. Catal.
B Environ. 2005, 58, 115–121. [CrossRef]
22. Alagiri, M.; Hamid, S.B.A. Sol-gel synthesis of α-Fe2 O3 nanoparticles and its photocatalytic application.
J. Sol.-Gel. Sci. Technol. 2015, 74, 783–789. [CrossRef]
23. Reda, S.M. Synthesis of ZnO and Fe2 O3 nanoparticles by sol-gel method and their application in dye-sensitized
solar cells. Mater. Sci. Semicond. Process. 2010, 13, 417–425. [CrossRef]
24. Lassoued, A.; Dkhil, B.; Gadri, A.; Ammar, S. Control of the shape and size of iron oxide (α-Fe2 O3 )
nanoparticles synthesized through the chemical precipitation method. Results. Phys. 2017, 7, 3007–3015.
[CrossRef]
25. Khodadadi-Moghaddam, M.; Salimi, F.; Sahebalzamani, H.; Jafari, N. Synthesis of Fe2 O3 nanoparticles via
various methods and coating with silica for drug immobilization. Synth. React. Inorg. Met. Nano-Metal.
Chem. 2013, 43, 1224–1227. [CrossRef]
26. Zhang, G.; Huang, X.; Tang, Z. Enhancing Water resistance of a Mn-based catalyst for low temperature
selective catalytic reduction reaction by modifying super hydrophobic layers. ACS Appl. Mater. Interfaces
2019, 11, 36598–36606. [CrossRef]
27. Duan, Z.; Chi, K.; Liu, J.; Shi, J.; Zhao, Z.; Wei, Y. The catalytic performances and reaction mechanism of
nanoparticle Cd/Ce-Ti oxide catalysts for NH3 -SCR reaction. RSC Adv. 2017, 7, 50127–50134. [CrossRef]
28. Jiang, M.; Wang, J.; Wang, J.; Shen, M. The influence of Si/Al ratios on adsorption and desorption
characterizations of Pd/Beta served as cold-start catalysts. Materials 2019, 12, 1045. [CrossRef]
29. Li, Y.; Wan, Y.; Li, Y.; Zhan, S.; Guan, Q.; Tian, Y. Low-temperature selective catalytic reduction of NO with
NH3 over Mn2 O3 -doped Fe2 O3 hexagonal microsheets. ACS Appl. Mater. Interfaces 2016, 8, 5224–5233.
[CrossRef]
30. Shan, H.; Liu, C.; Liu, L.; Zhang, J.; Li, H.; Liu, Z. Excellent toluene sensing properties of SnO2 -Fe2 O3
interconnected nanotubes. ACS Appl. Mater. Interfaces 2013, 5, 6376–6380. [CrossRef]
31. Ning, R.; Chen, L.; Li, E.; Liu, X.; Zhu, T. Applicability of V2 O5 -WO3 /TiO2 catalysts for the SCR denitrification
of alumina calcining flue gas. Catalysts 2019, 9, 220. [CrossRef]
32. Wu, Z.; Jin, R.; Liu, Y.; Wang, H. Ceria modified MnOx /TiO2 as a superior catalyst for NO reduction with
NH3 at low-temperature. Catal. Commun. 2008, 9, 2217–2220. [CrossRef]
33. Yu, J.; Si, Z.; Chen, L.; Wu, X.; Weng, D. Selective catalytic reduction of NOx by ammonia over
phosphate-containing Ce0.75 Zr0.25 O2 solids. Appl. Catal. B Environ. 2015, 163, 223–232. [CrossRef]
34. Xu, L.; Yang, Q.; Hu, L.; Wang, D.; Peng, Y.; Shao, Z.; Lu, C.; Li, J. Insights over titanium modified FeMgOx
catalysts for selective catalytic reduction of NOx with NH3 : Influence of precursors and crystalline structures.
Catalysts 2019, 9, 560. [CrossRef]
Catalysts 2019, 9, 1018 17 of 17

35. Li, Q.; Yang, H.; Ma, Z.; Zhang, X. Selective catalytic reduction of NO with NH3 over CuOx -carbonaceous
materials. Catal. Commun. 2012, 17, 8–12. [CrossRef]
36. Xie, G.; Liu, Z.; Zhu, Z.; Liu, Q.; Ge, J.; Huang, Z. Simultaneous removal of SO2 and NOx from flue gas using
a CuO/Al2 O3 catalyst sorbent: II. Promotion of SCR activity by SO2 at high temperatures. J. Catal. 2004, 224,
42–49. [CrossRef]
37. Roy, S.; Viswanath, B.; Hegde, M.S.; Madras, G. Low-temperature selective catalytic reduction of NO with
NH3 over Ti0.9 M0.1 O2−δ (M = Cr, Cr, Mn, Fe, Co, Cu). J. Phys. Chem. C 2008, 112, 6002–6012. [CrossRef]
38. Wu, Z.; Jiang, B.; Liu, Y.; Zhao, W.; Guan, B. Experimental study on a low-temperature SCR catalyst based on
MnOx /TiO2 prepared by sol-gel method. J. Hazard. Mater. 2007, 145, 488–494. [CrossRef]
39. Madia, G.; Koebel, M.; Elsener, M.; Wokaun, A. The effect of an oxidation precatalyst on the NOx reduction
by ammonia SCR. Ind. Eng. Chem. Res. 2002, 41, 3512–3517. [CrossRef]
40. Koebel, M.; Madia, G.; Elsener, M. Selective catalytic reduction of NO and NO2 at low temperatures.
Catal. Today 2002, 73, 239–247. [CrossRef]
41. Madia, G.; Koebel, M.; Elsener, M.; Wokaun, A. Side reactions in the selective catalytic reduction of NOx
with various NO2 fractions. Ind. Eng. Chem. Res. 2002, 41, 4008–4015. [CrossRef]
42. Liu, F.; He, H.; Ding, Y.; Zhang, C. Effect of manganese substitution on the structure and activity of iron
titanate catalyst for the selective catalytic reduction of NO with NH3 . Appl. Catal. B Environ. 2009, 93,
194–204. [CrossRef]
43. Cai, S.; Shi, L.; Huang, L.; Zhang, D.; Hu, H.; Li, H. Mechanistic aspects of deNOx processing over TiO2
supported Co–Mn oxide catalysts: Structure–activity relationships and In Situ DRIFTs analysis. ACS Catal.
2015, 5, 6069–6077.
44. Cai, S.; Zhang, D.; Shi, L.; Li, H.; Huang, L.; Hu, H. In Situ DRIFTs investigation of the low-temperature
reaction mechanism over Mn-doped Co3 O4 for the selective catalytic reduction of NOx with NH3 . J. Phys.
Chem. C. 2015, 119, 22924–22933.
45. Kijlstra, W.S.; Brands, D.S.; Poels, E.K.; Bliek, A. Mechanism of the selective catalytic reduction of NO by
NH3 over MnOx /Al2 O3 . J. Catal. 1997, 218, 208–218. [CrossRef]
46. Gu, T.; Liu, Y.; Weng, X.; Wang, H.; Wu, Z. The enhanced performance of ceria with surface sulfation for
selective catalytic reduction of NO by NH3 . Catal. Commun. 2010, 12, 310–313. [CrossRef]
47. Cataluna, R.; Arcoya, A.; Martínez-Arias, A.; Conesa, J.C.; Seoane, X.L.; Soria, J. NO reaction at surface
oxygen vacancies generated in cerium oxide. J. Chem. Soc. Farad. Trans. 2004, 91, 1679–1687.
48. Underwood, G.M.; Miller, T.M.; Grassian, V.H. Transmission FT-IR and knudsen cell study of the
heterogeneous reactivity of gaseous nitrogen dioxide on mineral oxide particles. J. Phys. Chem. A
1999, 103, 6184–6190. [CrossRef]
49. Cao, D.; Li, H.; Pan, L.; Li, J.; Wang, X.; Jing, P. High saturation magnetization of γ-Fe2 O3 nano-particles by a
facile one-step synthesis approach. Sci. Rep. 2016, 6, 1–9. [CrossRef]
50. Santos, P.; Marzan, L.M. N, N-Diamethylformamide as a reaction medium for metal nanoparticle synthesis.
Adv. Funct. Mater. 2009, 19, 679–688. [CrossRef]
51. Muzart, J. N, N-Diamethylformamide: Much more than a solvent. Tetrahedron 2009, 65, 8313–8323. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).