Pore Size Expansion Accelerates Ammonium BisulfateDecomposition for Improved Sulfur Resistance in Low-TemperatureNH3‑SCRKai Guo,† Gaofeng Fan,‡ Di Gu,† Shuohan Yu,† Kaili Ma,† Annai Liu,† Wei Tan,† Jiaming Wang,†

Xiangze Du,∥ Weixin Zou,† Changjin Tang,*,†,§ and Lin Dong*,†,‡,§

†School of Chemistry and Chemical Engineering and ‡School of Environment, Nanjing University, Nanjing 210023, China§Jiangsu Key Laboratory of Vehicle Emissions Control, Nanjing 210023, China∥School of Chemistry, Sichuan University, Chengdu 610000, China

*S Supporting Information

ABSTRACT: Sulfur poisoning has long been recognized as a bottleneck for thedevelopment of long-lived NH3-selective catalytic reduction (SCR) catalysts. Ammoniumbisulfate (ABS) deposition on active sites is the major cause of sulfur poisoning at lowtemperatures, and activating ABS decomposition is regarded as the ultimate way toalleviate sulfur poisoning. In the present study, we reported an interesting finding that ABSdecomposition can be simply tailored via adjusting the pore size of the material itdeposited. We initiated this study from the preparation of mesoporous silica SBA-15 withuniform one-dimensional pore structure but different pore sizes, followed by ABS loadingto investigate the effect. The results showed that ABS decomposition proceeded moreeasily on SBA-15 with larger pores, and the decomposition temperature declined as largeas 40 °C with increasing pore size of SBA-15 from 4.8 to 11.8 nm. To further ascertain thereal effect in NH3-SCR reaction, the Fe2O3/SBA-15 probe catalyst was prepared. It wasfound that the catalyst with larger mesopores exhibited much improved sulfur resistance,and quantitative analysis results obtained from Fourier transform infrared and ion chromatograph further proved that thedeposited sulfates were greatly alleviated. The result of the present study demonstrates for the first time the vital role of poresize engineering in ABS decomposition and may open up new opportunities for designing NH3-SCR catalysts with excellentsulfur resistance.

KEYWORDS: NH3-SCR, low temperature, sulfur poisoning, ammonium bisulfate decomposition, pore size effect

1. INTRODUCTION

Selective catalytic reduction (SCR) by ammonia has beenproved to be the most effective technology for removing NOx

from diesel engines and power plants, where the NH3-SCRcatalyst is the core of the SCR system.1,2 It is well accepted thatthe key to developing practically used NH3-SCR catalysts notonly relies on excellent denitration efficiency but also highresistance to sulfur oxides, due to the fact that most exhaustsfrom mobile sources and stationary sources contain not onlynitrogen oxides but also quite a good amount of sulfurdioxide.3,4 For commercial V2O5−WO3(MoO3)/TiO2 cata-lysts, excellent sulfur resistance is shown at 300−420 °C, whichcoincides with the temperature window of coal-fired powerplants.5,6 However, when the reaction temperature declines tolower values, e.g., 200−250 °C, catalyst deactivation oftentakes place. This is mainly attributed to the cover of stickysulfates like ammonium bisulfate (ABS) on the catalystsurface.7,8 It was reported that these ammonium salts couldnot decompose completely until the temperature exceeds 400°C.9 Consequently, they deposit gradually on the catalystsurface, causing pore plugging and active site blocking.10

Conventionally, deposited ABS on the NH3-SCR catalystsurface can be removed through water washing or hightemperature calcination.11−13 Chang et al. recently showed thatthe V2O5−WO3/TiO2 catalyst poisoned by ABS can beregenerated under elevated temperature treatment.14 However,both water washing and thermal regeneration display somenegative effects. For example, washing treatment will inevitablygenerate plenty of wastewater and treating the catalyst at hightemperatures is an energy-required process, which also caninduce significant reduction of the surface area and severesulfation of active components.15−19 From a practical point ofview, it is highly expected that the thermal drivendecomposition of ABS can be accelerated so that it can rivalwith ABS deposition. In an ideal case, negligible ABS can bedeposited on the catalyst surface when the formation anddecomposition rates of ABS reach a dynamic balance.

Received: September 10, 2018Accepted: January 17, 2019Published: January 17, 2019

Research Article

www.acsami.orgCite This: ACS Appl. Mater. Interfaces 2019, 11, 4900−4907

© 2019 American Chemical Society 4900 DOI: 10.1021/acsami.8b15688ACS Appl. Mater. Interfaces 2019, 11, 4900−4907

Dow

nloa

ded

via

NA

NJI

NG

UN

IV o

n M

ay 2

2, 2

019

at 0

6:50

:17

(UT

C).

Se

e ht

tps:

//pub

s.ac

s.or

g/sh

arin

ggui

delin

es f

or o

ptio

ns o

n ho

w to

legi

timat

ely

shar

e pu

blis

hed

artic

les.

There are some of the recent studies focused on the thermalstability of ABS on a catalyst surface. For example, Zhu et al.carried out a study on ABS decomposition over the V2O5catalyst supported on activated carbon (AC) and found that incomparison with TiO2 support, ABS decomposition waspromoted due to the reactivity between ABS and AC.20 Shenet al. investigated the mechanism of ammonium bisulfatedecomposition over V/WTi catalysts and revealed thatelectronic interactions between ABS and metal oxides (titaniaand tungsta) could weaken the thermal stability of ABS.7,21

Recently, Tang et al. reported a novel cation−anion separationstrategy against ABS deposition. They found that byintercalating NH4

+ into the layer structure MoO3, the strongelectrostatic interactions between NH4

+ and HSO4− were

broken, which was advantageous to promote ABS decom-position.22 Notably, in addition to the impacts from chemicalaspects, Xu et al. and co-workers proposed that the mesoporechannels in the MnO2−Fe2O3−CeO2−TiO2 catalyst played animportant role in ABS decomposition, based on theexperimental result that a balance between ABS formationand decomposition in SCR reaction was achieved on thecatalyst with abundant mesopores.23 Nevertheless, in theirstudy, due to the multiple components and coupled changes ofchemical properties (e.g., redox, acid−base) by constructingthe mesoporous catalyst, the complex influence from chemicalmodification cannot be discriminated, and therefore thedetailed role of the pore structure in ABS decomposition isstill not clear. In consideration that most of the preparedcatalysts are porous, it is imperative to discover the unique roleof pore parameters in heterogenous catalysis, particularly forthe development of long-lived SCR catalysts. In the presentstudy, we report a novel finding that irrespective of chemicalmodifications, ABS decomposition can be simply tailored viaadjusting the textual properties of the material it deposited.Specifically, ordered mesoporous silica SBA-15 with uniformone-dimensional pore structure and tunable pore size isemployed as a supporting material, and ABS decomposition isfound to be promoted in SBA-15 with a larger pore size. Aplausible explanation for the accelerated decomposition isproposed based on the Kelvin equation. Finally, to explore thereal effect on sulfur resistance during NH3-SCR reaction, amodel Fe2O3/SBA-15 catalyst was constructed and it wasobserved that the catalyst with a larger pore size showed muchbetter sulfur tolerance. We expect that the informationobtained in the present research could provide useful guidancefor the design of NH3-SCR catalysts with superior antisulfurpoisoning performance.

2. EXPERIMENTAL SECTION2.1. Sample Preparation. SBA-15 with different pore sizes were

synthesized by a hydrothermal method at different temperatures.24 Ina typical preparation, 4.0 g of P123 was dissolved in 30 mL ofdeionized water and 120 mL of 2.0 M HCl. Tetraethoxysilane (9.0 g)was added into the solution dropwise under stirring at 40 °C for 24 h.The milky slurry was transferred into a Teflon-lined autoclave forhydrothermal treatment at different temperatures (40, 100, 130, and150 °C). After that, the obtained white product was collected byfiltration, followed by washing with deionized water and ethanolseveral times. The cake was dried at 110 °C overnight and air-calcinedat 550 °C for 5 h. The obtained powder was labeled SBA-x, where xrepresents the operated hydrothermal temperature.To realize ABS loading on SBA-15, the calculated amount of

NH4HSO4 (15, 30 wt %) was dissolved in 30 mL of deionized waterwith 0.3 g of SBA-x support added. After vigorous stirring for 2 h,

water was evaporated at 100 °C and the obtained sample wassubsequently dried at 100 °C overnight. The sample was labeled mS/SBA-x, where m represents ABS loading in terms of weightpercentage.

The introduction of the active Fe2O3 component on SBA-15 wasperformed by a simple wet impregnation method. The calculatedamount (10 wt %) of Fe(NO3)3·9H2O was dissolved in 30 mL ofdeionized water and 0.5 g of support (SBA-40 and SBA-150) wasadded. The suspension was vigorously stirred for 2 h and driedovernight at 100 °C. Finally, the catalyst was obtained by calcinationat 450 °C for 4 h in a muffle furnace. For simplicity, the catalysts weredenoted as Fe2O3/SBA-40 and Fe2O3/SBA-150.

2.2. Sample Characterization. The structure of synthesizedSBA-15 was determined by small-angle X-ray diffraction (XRD),transmission electron microscopy (TEM), and N2-sorption. A PhilipsX’pert Pro diffractometer with Ni-filtered Cu Kα1 radiation (0.15408nm) was used to collect X-ray powder diffraction (XRD) patternswith a step width of 0.06°/s from 0.6 to 5°. The X-ray tube wasoperated at 40 kV and 40 mA. The specific surface area and pore sizedistribution (PSD) of samples were measured by N2-physisorption at−196 °C on a Micromeritics ASAP-3020 analyzer. Before measure-ment, the sample was vacuum-pretreated at 300 °C (120 °C for ABSdeposited samples) for 3 h and pore size distribution results wereobtained by the Barrett−Joyner−Halenda method. Transmissionelectron microscopy (TEM) images were taken on a JEM-1011instrument at an acceleration voltage of 200 kV. The sample wasdispersed in A.R. grade ethanol with ultrasonic treatment and theresulting suspension was allowed to dry on a carbon film supported oncopper grids. Fourier transform infrared (FT-IR) spectra of sampleswere collected on a Nicolet Is10 FT-IR spectrometer from 400 to4000 cm−1. The background was collected first and then subtractedfor every sample. The number of scans was 32 at a resolution of 4cm−1.

The decomposition behaviors and products of NH4HSO4 onvarious mesoporous silica were measured by simultaneous thermalanalysis and mass spectroscopy (STA-MS). The measurement wascarried out in a flowing nitrogen atmosphere (60 mL/min) on aNETZSCH STA-449-F5 combined with a QMS-403D instrumentand the heating rate was set at 5 °C/min. The reference curve forDSC was operated on an empty crucible.

To obtain the quantitative data of surface species (sulfates andsoluble metal ions) after sulfur resistance test, 50 mg of the spentcatalyst was dispersed in 20 mL of deionized water and ultrasonicatedfor 2 h, then the solid was filtered out and the liquid phase was keptfor the following tests. Inductively coupled plasma atomic emissionspectroscopy (ICP-AES) was used for analyzing soluble cations in thesample after the sulfur resistance test on a PerkinElmer Optima 5300DV instrument with a radio frequency power of 1300 W. The sulfateanions in the liquid phase were analyzed by ion chromatography(DIONEX ICS-1100AR, Thermofisher scientific, India). The systemconsisted of one IC pump, an injection valve, column heater, an ion-exchange analytical column (AS19), and a guard column (AG19).Samples were loaded onto a 100 μL sample loop with a syringe. Allcolumns, tubing, and fittings were made of poly-ether-ether-ketone.During the IC measurement, 25 mM KOH eluent, 0.38 mL/min flowrate, and 24 mA applied current were maintained.

2.3. Activity and Sulfur Resistance Test on Model Catalysts.The activity performance of the model catalyst was evaluated in afixed-bed quartz reactor. The feed stream was fixed with 500 ppmNO, 500 ppm NH3, 5% O2, 800 ppm SO2 (when used), and Ar inbalance, and total gas flow was controlled at 200 mL/min. Thecatalyst (200 mg) was sieved to 40−60 mesh. Before the test, it waspretreated in purified Ar stream at 150 °C for 1 h. Then, the mixedgases were switched on and the activity data were collected at everytarget temperature after stabilizing for 1 h in the range of 50−400 °C.The sulfur resistance experiment was carried out in a similar way.After Ar pretreatment, the SCR reaction was performed by holdingthe reaction temperature at 250 °C. SO2 was introduced after thereaction steady state was reached. The concentration of effluent gaseswas continuously analyzed on an IS10 FT-IR spectrometer equipped

ACS Applied Materials & Interfaces Research Article

DOI: 10.1021/acsami.8b15688ACS Appl. Mater. Interfaces 2019, 11, 4900−4907

4901

with a 2 m path-length gas cell (2 L volume). The NO conversion andN2 selectivity were calculated based on the following equations

NO conversion(%)NO NO

NO100%in out

in=

[ ] − [ ][ ]

×(1)

N selectivity(%)NO NO NH NH NO 2 N O

NO NO NH NH

100%

2

in out 3 in 3 out 2 out 2 out

in out 3 in 3 out=

[ ] − [ ] + [ ] − [ ] − [ ] − [ ][ ] − [ ] + [ ] − [ ]

× (2)

3. RESULTS AND DISCUSSION3.1. Structure Properties of SBA-15 Synthesized from

Different Hydrothermal Temperatures. Mesoporous silicaSBA-15 exhibits chemical inertness, uniform one-dimensionalpore structure, and tunable pore size, and thus is an idealsupporting material to investigate the pore size effect on ABSdecomposition. SBA-15 synthesized at different hydrothermaltemperatures was first characterized by TEM, and the obtainedimages are displayed in Figure S1. Ordered channel arrays wereobserved along the direction of pore arrangement, indicatingthe formation of uniform one-dimensional pore structures. Thelarge-scale periodicity of mesopores was identified by theappearance of a family of diffraction peaks in small-angle XRDpatterns (Figure S2).25 As well, the structural details wererevealed by N2-physisorption isotherms and pore sizedistribution (PSD) curves (Figures S3 and 1). It can be seen

that all samples display type IV isotherms according to IUPACclassification with H1 hysteresis loops, indicative of theformation of cylindrical mesoporous materials.26 As expected,PSD curves revealed clear differences in pore size. At thehydrothermal temperatures of 40, 100, 130, and 150 °C, SBA-15 with pore sizes of 4.84, 6.24, 8.43, and 11.8 nm wasobtained. Taken together, we can safely conclude that uniformone-dimensional mesopores with varying pore sizes areestablished and the influence of pore size on ABSdecomposition can thus be readily investigated.3.2. Location and Chemical State of ABS. The

investigation system of ABS deposited on mesoporous silicawas constructed via the wetness impregnation method. TableS1 lists the textual properties of the samples before and after

ABS impregnation. It can be seen that ABS deposition resultsin an obvious decrease of the surface area and pore volume ofthe carrier, indicating that the deposited ABS is filled in thepore channels of SBA-15. Previously, it was reported that thechemical state of ABS on the catalyst surface can affect theirdecomposition behaviors.7 To explore any influence of poresize on the chemical state of impregnated ABS in themesopores, FT-IR analysis is conducted on 15S/SBA-40 and15S/SBA-150 samples. In general, for sulfur-contained species,their bands are centered at the wavenumber range of 800−1500 cm−1.7,18,277,18,27 However, this range is entangled withthe Si−O vibrations from mesoporous silica.28,29 Thus, todiscriminate the signal of sulfate from that of silica material,differential FT-IR spectra obtained by deducting the spectra ofpure SBA-40 and SBA-150 supports are plotted and the resultsare displayed in Figure 2. In the differential spectra of 15S/

SBA-40 and 15S/SBA-150, well-resolved vibration bands dueto ABS deposition are present. The broad band at 3500−3000cm−1 and peaks at 1641 and 1446 cm−1 represent thevibrations of N−H in NH4

+,30,31 whereas the bands at 1172,1033, 882, and 588 cm−1 are characteristic of differentvibration modes of sulfate and bisulfate species.32 Bycomparing the two differential spectra, no evident differencesin band intensity and peak position are observed, suggesting anegligible influence on the deposition state of ABS by differentpore sizes.

3.3. Decomposition of ABS Supported on SBA-15. Atypical 15S/SBA-40 sample was employed to approach thedecomposition details of ABS on SBA-15. Thermogravimetry−differential scanning calorimetry (TG−DSC) and MStechniques were jointly used to collect the signals during thedecomposition process and the results are displayed in Figure3. Two weight losses are visible from TG curves in the testedrange, and no further weight loss is detected at highertemperatures (data were not shown). The weight loss below150 °C can be attributed to the evaporation of absorbed water,whereas the weight loss between 300 and 400 °C is due to thedecomposition of ABS. Correspondingly, two endothermicpeaks are shown in the DSC curves. The attribution is furtherconfirmed by qualitative analysis from mass spectroscopy. As

Figure 1. Pore size distribution curves of SBA-15 synthesized atdifferent hydrothermal temperatures.

Figure 2. Differential spectra obtained by subtracting the spectra ofpure supports from 15S/SBA-40 and 15S/SBA-150.

ACS Applied Materials & Interfaces Research Article

DOI: 10.1021/acsami.8b15688ACS Appl. Mater. Interfaces 2019, 11, 4900−4907

4902

can be seen from MS profiles, the signal of water starts toappear at very low temperature and obtains a peak at 69.2 °C,and then declines to baseline at a temperature below 150 °C.For the SO2 signal, a much broad temperature window isdisplayed. SO2 begins to release at a temperature of 212 °C,and the maximal decomposition rate is reached at atemperature of ca. 384 °C.To explore the possible loading amount effect, the

decomposition experiment was also carried out on the 30S/SBA-40 sample and the result is presented in Figure S4. It isobserved that the TG−DSC and MS signals of 30S/SBA-40display similar trends to 15S/SBA-40, revealing that thedecomposition behavior of ABS is not much affected by itsloading. Nevertheless, the detailed decomposition temperaturechanges with ABS loading. The onset and peak decompositiontemperatures of 30S/SBA-40 are observed at 255.7 and 392.5°C, which are much higher than those of 15S/SBA-40 (212.0and 384.1 °C), suggesting that higher loading of ABS leads todifficult decomposition.The influence of pore size on ABS decomposition was

investigated and detailed TG−DSC curves are shown in FigureS5. For convenience, the peak temperature of the secondendothermic DSC peak is taken as an indication of thedecomposition temperature of ABS on SBA-15. It is found thatfor the set of 15S/SBA-x samples, both TG and DSC curvesshow a similar trend with the curves of 15S/SBA-40. Allsamples reveal two main weight losses and two endothermicpeaks. Moreover, it is observable that the decompositiontemperature displays a monotonic reduction with increasingsupport pore size (Figure 4), suggesting accelerated decom-position of ABS on SBA-15 with a larger pore size.Remarkably, an almost 40 °C decrease in ABS decompositiontemperature is reached with the increase of the support poresize from 4.84 to 11.8 nm.With the employment of inert mesoporous silica as a

supporting material, the influence of chemical factors (redoxand/or acid−base properties) on ABS decomposition isinsignificant. As such, we can focus on the impact of physicalfactors, specifically, the pore width of SBA-15 in the presentstudy. Previously, Taira et al. reported that the Pt/TiO2catalyst with a high ratio of larger pores exhibited higher COoxidation activity under the influence of SO2. They proposed

that the enhanced performance was related to reducedblockage of the catalyst surface from the condensation of thebyproduct, sulfuric acid.33 According to the Kelvin equation,the condensation of sulfuric acid occurs in smaller pores firstand that larger pores are less vulnerable to condensation. In thepresent study, we suppose the similar rule may be held for ABSdecomposition.Regarding the status of ABS during thermal treating, it was

reported that ABS would become sticky and melt at around150 °C during temperature programming.9 This is evidencedby the STA-MS characterization of pure ABS. As can be seenfrom Figure 5, two endothermic peaks centered at 159 and 513°C are shown in the DSC curve. In contrast, no obvious weightloss is detected around 160 °C from the TG curve. Therefore,we can speculate that the phase transition of ABS from solid toliquid takes place at this temperature range and ABS actuallyexperiences a liquid phase before subsequent decomposition athigher temperatures.It is known that for a liquid phase with a curved surface, such

as the surface of a droplet, the equilibrium vapor pressure canbe different from vapor pressure at a flat surface. The change ofvapor pressure due to curved liquid−vapor interface can be

Figure 3. STA-MS patterns of 15S/SBA-40.

Figure 4. Decomposition temperature of 15 wt % ABS supported onSBA-15 with different pore sizes.

ACS Applied Materials & Interfaces Research Article

DOI: 10.1021/acsami.8b15688ACS Appl. Mater. Interfaces 2019, 11, 4900−4907

4903

well described by the Kelvin equation (eq 3), where p is theactual vapor pressure, p0 is the vapor pressure when the surfaceis flat, VL is the molar liquid volume, γ is the surface tension, θis the contact angle, rp is the curvature radius, R is the gasconstant, and T is the temperature.

pp

Vr RT

ln2 cos

0

L

p

γ θ=

(3)

For liquid phase ABS in SBA-15, it displays a concave surfaceand the vapor pressure (p) is lower than p0. Thus, the highervapor pressure of the melted ABS is attained in larger porechannels. With the increase of temperature, liquid phase ABSin larger pores of SBA-15 is more liable to vaporize and thedecomposition is easier to take place by taking intoconsideration the negligible intermolecular interaction forcesof gaseous ABS.34 As a result, the gas−liquid equilibrium willbe prompted to the gas side, which further alleviates thecondensation and deposition of ABS on larger pores. Scheme 1illustrates the state of ABS on SBA-15 with different pore sizes.This can well explain the accelerated decomposition of ABS onSBA-15 with larger pores. We anticipate that this uniquecharacter can be used to promote the sulfur resistance ability ofthe SCR catalyst operated at low temperatures.3.4. Pore Size Effect on Sulfur Resistance during NH3-

SCR Reaction. A model catalyst with iron oxide supported onSBA-15 is constructed to explore the pore size effect on ABSdecomposition in practical NH3-SCR reaction. For thecomparison purpose, two SBA-15 supports (SBA-40 andSBA-150) are used and the obtained catalysts are denoted as

Fe2O3/SBA-40 and Fe2O3/SBA-150, respectively. Figure S6shows the pore size distributions of the two catalysts. Asexpected, after Fe doping, both catalysts display a decrease inpore size. Nevertheless, uniform pore distribution is stillpresent and distinct pore sizes are apparently displayed, whichis helpful for us to investigate the pore size effect. The chemicalstate of mesoporous silica after iron oxide deposition wasstudied by FT-IR spectroscopy. Figure S7 presents the IRspectra of SBA-15 and Fe2O3/SBA-15 samples. The vibrationsat 1054 and 803 cm−1 are attributed to asymmetric andsymmetric stretching of the Si−O−Si bond, and the band at964 cm−1 is ascribed to the stretching vibration of Si−O−H.35,36 After introducing Fe oxide, the vibrations of Si−Orelated bands remain almost the same as that of pure SBA-15.It proves that SBA-15 maintained its chemical properties afterimpregnating with iron oxides. The significant disparity in thepore size comparable chemical state of two model catalysts wasalso much convenient for us to investigate the pore size effecton ABS decomposition in the practical catalytic process.The NO conversion and N2 selectivity results of Fe2O3/

SBA-40 and Fe2O3/SBA-150 are shown in Figure S8. It isobvious that they have almost identical activity in the range of50−400 °C, suggesting that the chemical properties of activecomponents (Fe2O3) were also barely influenced by differentsilica supports. Figure 6 illustrates the NH3-SCR performanceof the two catalysts operated in the presence of 800 ppm SO2.The reaction temperature is controlled at 250 °C to avoidsignificant decomposition of ABS. The initial NO conversionsare comparable for both samples, which is in line with theabove activity results. However, a notable difference in reactionefficiency can be observed from the longtime stability test. ForFe2O3/SBA-150, NO conversion reveals no evidence of declinein the entire test range, demonstrating that the catalyst isbarely deactivated after SO2 injection. In contrast, for Fe2O3/SBA-40, NO conversion decreases instantly when SO2 isintroduced into the feeding gas. After little recovery of activity,NO conversion shows a continuous recession to 36% after the16 h test. Additionally, for the universal concern of the poresize effect, Fe2O3/MCM-41 was also prepared. MCM-41 is akind of mesoporous silica with a similar cylindrical porechannel to SBA-15, but it shows a much smaller pore size (2−3nm). The result in Figure 6 shows that when MCM-41 wasemployed, a more rapid decrease of activity was obtained. Afterthe 16 h test, the NO conversion of Fe2O3/MCM-41 was

Figure 5. STA-MS patterns of pure ABS.

Scheme 1. Illustration of ABS Decomposition Behaviors on SBA-15 with Different Pore Sizes

ACS Applied Materials & Interfaces Research Article

DOI: 10.1021/acsami.8b15688ACS Appl. Mater. Interfaces 2019, 11, 4900−4907

4904

finally declined to ca. 25%. Apparently, the SO2 tolerance testtells us that the catalyst with a larger pore size can restrain SO2poisoning more effectively, and the pore size effect on ABSdecomposition could also be verified on various supports.It is well acknowledged that deactivation of NH3-SCR

catalysts in the sulfur-contained stream can be attributed to theformation of sulfate-related species.8,37−39 To know moredetails about the deactivation, FT-IR spectra were employed toidentify the change of surface species, and the results arepresented in Figure 7a. It is obvious that the spectra of thespent catalysts display some differences compared to those ofthe fresh ones. Several new peaks appear after the sulfurresistance test and could be reasonably ascribed to the reactionbyproducts. The broad band emerged in the range of 3000−3500 cm−1 and an intense peak at 1450 cm−1 on the spentcatalysts were characteristic of the N−H stretching mode andNH4

+ on Brønsted acid sites,40 implying that ammoniumspecies are formed on the catalyst surface after the sulfur

resistance test. For sulfate signals, it can be seen from Figure 7bthat Fe2O3/SBA-40-spent reveals more intense bands thanFe2O3/SBA-150-spent. The bands at about 1094, 1011, 941,and 787 cm−1 are clear on Fe2O3/SBA-40 after the sulfurresistance test. However, the differential spectrum of Fe2O3/SBA-150-spent does not show these significant peaks, implyingthat sulfate species was slightly deposited. Based on FT-IRresults, the cover of ammonium sulfate species on the catalystsurface is established. Moreover, by comparing the differentialspectra, it can be deduced that the catalyst with larger pore sizecan greatly reduce the deposition of surface sulfate salts. This isin line with the experimental result operated on bare SBA-15and suggests that the rule is also applicable to the practicalcatalytic process.Finally, to acquire quantitative information of the formed

sulfate salts (Fe2(SO4)3, NH4HSO4) on the spent catalyst,ICP-AES and ion chromatography experiments were per-formed. Table 1 shows the concentration of various ions on the

two spent samples. It can be seen that quite a good amount ofsulfate ions is generated on the spent samples but only a traceamount of Fe3+, indicating that the negligible metal componentis sulfated during the sulfur resistance test. Thus, under currentreaction conditions, the contribution of deactivation fromchemical poisoning of the active component should be minor.On the other hand, the deposition of sulfate salts on Fe2O3/SBA-150 is greatly alleviated, as can be evidenced by the factthat the amount of SO4

2− on Fe2O3/SBA-40-spent is almostthree times that on Fe2O3/SBA-150-spent. This can wellaccount for the distinct sulfur resistance performance betweenFe2O3/SBA-40 and Fe2O3/SBA-150. Based on the aboveresults, we could safely conclude that the textual properties ofthe catalyst have a significant impact on ABS decomposition,

Figure 6. Sulfur resistance tests on Fe2O3/SBA-40, Fe2O3/SBA-150,and Fe2O3/MCM-41 catalysts. Reaction conditions: 500 ppm NO,500 ppm NH3, 5 vol % O2, 800 ppm SO2, GHSV of ca.15 000 h−1.

Figure 7. (a) FT-IR spectra of fresh and spent Fe2O3/SBA-15 catalysts and (b) the differential spectra obtained by subtracting the spectra beforereaction from those after reaction.

Table 1. Soluble Ions on the Spent Catalyst Surfacea

sample SO42− (mg/L) Fe3+ (mg/L)

Fe2O3/SBA-40-spent 33.814 0.14Fe2O3/SBA-150-spent 11.288 0.21

aThe concentrations of SO42− and Fe3+ were measured by ion

chromatography and ICP-AES, respectively.

ACS Applied Materials & Interfaces Research Article

DOI: 10.1021/acsami.8b15688ACS Appl. Mater. Interfaces 2019, 11, 4900−4907

4905

and the catalyst with larger pores can effectively prevent thedeposition of ammonium bisulfate, which is in favor of goodsulfur resistance performance in practical SCR reaction.

4. CONCLUSIONSIn summary, we reported in the present study the pore-size-dependent property of ABS decomposition for low-temper-ature NH3-SCR reaction. It was found that ABS deposition canbe greatly alleviated by adjusting the pore size of the materialsit deposited, with no need for chemical modification. By usingmesoporous silica SBA-15 as a supporting material, an obviousacceleration effect of ABS decomposition is observed with anenlargement of pore size. This decomposition promoting effectwas mainly related to the fact that larger pore size generateshigher vapor pressure, which is helpful for ABS vaporizationand decomposition. Furthermore, the sulfur resistance test onthe model Fe2O3/SBA-15 catalyst showed the catalyst withlarger pore sizes exhibited much better durability than thesmaller one, validating that the promotion effect was alsoapplicable for the real catalytic process. The present studybrings us a brief cognition on how the catalyst pore structureaffects ammonium bisulfate decomposition and is expected tobe helpful for us in designing novel NH3-SCR catalysts withexcellent sulfur resistance ability.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acsami.8b15688.

Small-angle XRD, TEM, N2-sorption, TG−DSC pat-terns of 15S/SBA-15 series, TG−MS of 30S/SBA-40,pore parameter of ABS or Fe2O3 deposited samples, andSCR activity with N2 selectivity (PDF)

■ AUTHOR INFORMATIONCorresponding Authors*E-mail: tangcj@nju.edu.cn (C.T.).*E-mail: donglin@nju.edu.cn (L.D.).ORCIDLin Dong: 0000-0002-8393-6669NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSFinancial support from the National Natural ScienceFoundation of China (21773106, 21707066, 21573105) andEnvironmental Protection Department of Jiangsu Province(2016048) is gratefully acknowledged.

■ REFERENCES(1) Busca, G.; Lietti, L.; Ramis, G.; Berti, F. Chemical AndMechanistic Aspects of the Selective Catalytic Reduction of NOx byAmmonia over Oxide Catalysts: A Review. Appl. Catal., B 1998, 18,1−36.(2) Kamasamudram, K.; Currier, N. W.; Chen, X.; Yezerets, A.Overview of the Practically Important Behaviors of Zeolite-basedUrea-SCR Catalysts, Using Compact Experimental Protocol. Catal.Today 2010, 151, 212−222.(3) Li, J.; Chang, H.; Ma, L.; Hao, J.; Yang, R. T. Low-temperatureSelective Catalytic Reduction of NOx with NH3 over Metal Oxide andZeolite CatalystsA review. Catal. Today 2011, 175, 147−156.

(4) Liu, Z.; Li, J.; Junaid, A. S. M. Knowledge and Know-how inImproving the Sulfur Tolerance of DeNOx Catalysts. Catal. Today2010, 153, 95−102.(5) Lietti, L.; Nova, I.; Ramis, G.; Dall’Acqua, L.; Busca, G.;Giamello, E.; Forzatti, P.; Bregani, F. Characterization and Reactivityof V2O5−MoO3/TiO2 De-NOx SCR Catalysts. J. Catal. 1999, 187,419−435.(6) Gao, R.; Zhang, D.; Liu, X.; Shi, L.; Maitarad, P.; Li, H.; Zhang,J.; Cao, W. Enhanced Catalytic Performance of V2O5-WO3/Fe2O3/TiO2 Microspheres for Selective Catalytic Reduction of NO by NH3.Catal. Sci. Technol. 2013, 3, 191−199.(7) Li, C.; Shen, M.; Yu, T.; Wang, J.; Wang, J.; Zhai, Y. TheMechanism of Ammonium Bisulfate Formation and Decompositionover V/WTi Catalysts for NH3-selective Catalytic Reduction atVarious Temperatures. Phys. Chem. Chem. Phys. 2017, 19, 15194−15206.(8) Xi, Y.; Ottinger, N. A.; Liu, Z. G. New Insights into SulfurPoisoning on a Vanadia SCR Catalyst under Simulated Diesel EngineOperating Conditions. Appl. Catal., B 2014, 160-161, 1−9.(9) Kiyoura, R.; Urano, K. Mechanism, Kinetics, and Equilibrium ofThermal Decomposition of Ammonium Sulfate. Ind. Eng. Chem.Process Des. Dev. 1970, 9, 489−494.(10) Guo, X.; Bartholomew, C.; Hecker, W.; Baxter, L. L. Effects ofSulfate Species on V2O5/TiO2 SCR Catalysts in Coal and Biomass-fired Systems. Appl. Catal., B 2009, 92, 30−40.(11) Jin, R.; Liu, Y.; Wu, Z.; Wang, H.; Gu, T. Relationship BetweenSO2 Poisoning Effects and Reaction Temperature for SelectiveCatalytic Reduction of NO over Mn−Ce/TiO2 Catalyst. Catal. Today2010, 153, 84−89.(12) Chang, H.; Li, J.; Yuan, J.; Chen, L.; Dai, Y.; Arandiyan, H.; Xu,J.; Hao, J. Ge, Mn-doped CeO2−WO3 Catalysts for NH3−SCR ofNOx: Effects of SO2 and H2 Regeneration. Catal. Today 2013, 201,139−144.(13) Sheng, Z.; Hu, Y.; Xue, J.; Wang, X.; Liao, W. SO2 Poisoningand Regeneration of Mn-Ce/TiO2 Catalyst for Low TemperatureNOx Reduction with NH3. J. Rare Earths 2012, 30, 676−682.(14) Xu, L.; Wang, C.; Chang, H.; Wu, Q.; Zhang, T.; Li, J. NewInsight into SO2 Poisoning and Regeneration of CeO2−WO3/TiO2and V2O5−WO3/TiO2 Catalysts for Low-Temperature NH3−SCR.Environ. Sci. Technol. 2018, 52, 7064−7071.(15) Zhang, L.; Zou, W.; Ma, K.; Cao, Y.; Xiong, Y.; Wu, S.; Tang,C.; Gao, F.; Dong, L. Sulfated Temperature Effects on the CatalyticActivity of CeO2 in NH3-Selective Catalytic Reduction Conditions. J.Phys. Chem. C 2015, 119, 1155−1163.(16) Xu, W.; He, H.; Yu, Y. Deactivation of a Ce/TiO2 Catalyst bySO2 in the Selective Catalytic Reduction of NO by NH3. J. Phys.Chem. C 2009, 113, 4426−4432.(17) Zhang, L.; Li, L.; Cao, Y.; Yao, X.; Ge, C.; Gao, F.; Deng, Y.;Tang, C.; Dong, L. Getting Insight into the Influence of SO2 onTiO2/CeO2 for the Selective Catalytic Reduction of NO by NH3.Appl. Catal., B 2015, 165, 589−598.(18) Song, L.; Chao, J.; Fang, Y.; He, H.; Li, J.; Qiu, W.; Zhang, G.Promotion of Ceria for Decomposition of Ammonia Bisulfate overV2O5-MoO3/TiO2 Catalyst for Selective Catalytic Reduction. Chem.Eng. J. 2016, 303, 275−281.(19) Yang, H.; Wei, G.; Wang, X.; Lin, F.; Wang, J.; Shen, M.Regeneration of SO2-poisoned Diesel Oxidation Pd/CeO2 Catalyst.Catal. Commun. 2013, 36, 5−9.(20) Zhu, Z.; Niu, H.; Liu, Z.; Liu, S. Decomposition and Reactivityof NH4HSO4 on V2O5/AC Catalysts Used for NO Reduction withAmmonia. J. Catal. 2000, 195, 268−278.(21) Ye, D.; Qu, R.; Zheng, C.; Cen, K.; Gao, X. MechanisticInvestigation of Enhanced Reactivity of NH4HSO4 and NO on Nb-and Sb-doped VW/Ti SCR Catalysts. Appl. Catal., A 2018, 549, 310−319.(22) Chen, Y.; Li, C.; Chen, J.; Tang, X. Self-Prevention of Well-Defined-Facet Fe2O3/MoO3 against Deposition of AmmoniumBisulfate in Low-Temperature NH3−SCR. Environ. Sci. Technol.2018, 52, 11796−11802.

ACS Applied Materials & Interfaces Research Article

DOI: 10.1021/acsami.8b15688ACS Appl. Mater. Interfaces 2019, 11, 4900−4907

4906

(23) Yu, J.; Guo, F.; Wang, Y.; Zhu, J.; Liu, Y.; Su, F.; Gao, S.; Xu, G.Sulfur Poisoning Resistant Mesoporous Mn-base Catalyst for Low-temperature SCR of NO with NH3. Appl. Catal., B 2010, 95, 160−168.(24) Zhao, D.; Huo, Q.; Feng, J.; Chmelka, B. F.; Stucky, G. D.Nonionic Triblock and Star Diblock Copolymer and OligomericSurfactant Syntheses of Highly Ordered, Hydrothermally Stable,Mesoporous Silica Structures. J. Am. Chem. Soc. 1998, 120, 6024−6036.(25) Kruk, M.; Jaroniec, M.; Ko, C. H.; Ryoo, R. Characterization ofthe Porous Structure of SBA-15. Chem. Mater. 2000, 12, 1961−1968.(26) Zhao, D.; Sun, J.; Li, Q.; Stucky, G. D. Morphological Controlof Highly Ordered Mesoporous Silica SBA-15. Chem. Mater. 2000, 12,275−279.(27) Kantcheva, M.; Cayirtepe, I.; Naydenov, A.; Ivanov, G. FT-IRSpectroscopic Investigation of the Effect of SO2 on the SCR of NOxwith Propene over ZrO2−Nb2O5 Catalyst. Catal. Today 2011, 176,437−440.(28) Kang, T.; Park, Y.; Choi, K.; Lee, J. S.; Yi, J. OrderedMesoporous Silica (SBA-15) Derivatized with Imidazole-containingFunctionalities as a Selective Adsorbent of Precious Metal Ions. J.Mater. Chem. 2004, 14, 1043−1049.(29) Nguyen, T. P. B.; Lee, J.-W.; Shim, W. G.; Moon, H. Synthesisof Functionalized SBA-15 with Ordered Large Pore Size and ItsAdsorption Properties of BSA. Microporous Mesoporous Mater. 2008,110, 560−569.(30) Peng, Y.; Li, K.; Li, J. Identification of the Active Sites onCeO2−WO3 Catalysts for SCR of NOx with NH3: An in situ IR andRaman Spectroscopy Study. Appl. Catal., B 2013, 140−141, 483−492.(31) Ramis, G.; Yi, L.; Busca, G. Ammonia Activation over Catalystsfor the Selective Catalytic Reduction of NOx and the SelectiveCatalytic Oxidation of NH3. An FT-IR Study. Catal. Today 1996, 28,373−380.(32) Miller, F. A.; Wilkins, C. H. Infrared Spectra and CharacteristicFrequencies of Inorganic Ions. Anal. Chem. 1952, 24, 1253−1294.(33) Taira, K.; Nakao, K.; Suzuki, K.; Einaga, H. SOx Tolerant Pt/TiO2 Catalysts for CO Oxidation and the Effect of TiO2 Supports onCatalytic Activity. Environ. Sci. Technol. 2016, 50, 9773−9780.(34) Matsuda, S.; Kamo, T.; Kato, A.; Nakajima, F.; Kumura, T.;Kuroda, H. Deposition of Ammonium Bisulfate in the SelectiveCatalytic Reduction of Nitrogen Oxides with Ammonia. Ind. Eng.Chem. Prod. Res. Dev. 1982, 21, 48−52.(35) Yang, J.; Zhang, J.; Zhu, L.; Chen, S.; Zhang, Y.; Tang, Y.; Zhu,Y.; Li, Y. Synthesis of Nano Titania Particles Embedded inMesoporous SBA-15: Characterization and Photocatalytic Activity.J. Hazard. Mater. 2006, 137, 952−958.(36) Izquierdo-Barba, I.; Ruiz-Gonzalez, L.; Doadrio, J. C.;Gonzalez-Calbet, J. M.; Vallet-Regí, M. Tissue Regeneration: ANew Property of Mesoporous Materials. Solid State Sci. 2005, 7, 983−989.(37) Jangjou, Y.; Wang, D.; Kumar, A.; Li, J.; Epling, W. S. SO2Poisoning of the NH3-SCR Reaction over Cu-SAPO-34: Effect ofAmmonium Sulfate versus Other S-Containing Species. ACS Catal.2016, 6, 6612−6622.(38) Zhang, L.; Wang, D.; Liu, Y.; Kamasamudram, K.; Li, J.; Epling,W. SO2 Poisoning Impact on the NH3-SCR Reaction Over aCommercial Cu-SAPO-34 SCR Catalyst. Appl. Catal., B 2014, 156−157, 371−377.(39) Liu, Y.; Cen, W.; Wu, Z.; Weng, X.; Wang, H. SO2 PoisoningStructures and the Effects on Pure and Mn Doped CeO2: A FirstPrinciples Investigation. J. Phys. Chem. C 2012, 116, 22930−22937.(40) Larrubia, M. A.; Ramis, G.; Busca, G. An FT-IR Study of theAdsorption of Urea and Ammonia over V2O5−MoO3−TiO2 SCRCatalysts. Appl. Catal., B 2000, 27, L145−L151.

ACS Applied Materials & Interfaces Research Article

DOI: 10.1021/acsami.8b15688ACS Appl. Mater. Interfaces 2019, 11, 4900−4907

4907