- Email: [email protected]
The different poisoning behaviors of various alkali metal containing compounds on SCR catalyst
Applied Surface Science 392 (2017) 162–168
Contents lists available at ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
The different poisoning behaviors of various alkali metal containing compounds on SCR catalyst Xuesen Du a,b,∗ , Guangpeng Yang b , Yanrong Chen a,b , Jingyu Ran a,b , Li Zhang a,b a Key Laboratory of Low-grade Energy Utilization Technologies & Systems of Ministry of Education of China, College of Power Engineering, Chongqing University, Chongqing 400044, China b Institute of Energy and Environment of Chongqing University, Chongqing, 400044, China
a r t i c l e
i n f o
Article history: Received 13 May 2016 Received in revised form 9 September 2016 Accepted 9 September 2016 Available online 12 September 2016 Keywords: Selective catalytic reduction Poisoning Alkali metal Density functional theory Vanadia
a b s t r a c t Alkali metals are poisonous to the metal oxide catalyst for NO removal. The chemical configuration of alkali containing substance and interacting temperature can affect the poisoning profile. A computational method based on Frontier Molecular Orbital analysis was proposed to determine the reacting behavior of various alkali-containing substances with SCR catalyst. The results reveal that the poisoning reactivities of various substances can be ranked as: E (MOH) > E (M2 SO4 ) > E(MCl) > E(MNO3 ) > E(MHSO4 ). The experimental activity tests of the catalysts calcined at stepped temperatures show that NaOH can react with the catalyst below 200 ◦ C. NaCl and NaNO3 start to react with the catalyst at a temperature between 300 and 400 ◦ C. Unlike MOH, MCl and MNO3 , which can produce volatile or decomposable species for the anions after reacting with the catalyst, M2 SO4 and MHSO4 will leave both cations and anions on the catalyst surface. The sulfate ions left on the catalyst can generate active acid sites for NH3 adsorption. The experimental results also show that Na2 SO4 and NaHSO4 will not lower the NO conversion. The afterreaction influences of various alkali metals were studied using theoretical and experimental methods. The theoretical results show that the acidity decreases with doping of alkali metal. Experiments show a consistent result that the NO conversion decreases as undoped >LiCl > NaCl > KCl. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Nitric oxides (NOx ), mostly nitric monoxide (NO), emitted from stationary and mobile sources is thought to be one of the major pollutants in atmosphere. Selective catalytic reduction (SCR) is the most efficient and commonly used technologies to remove gas phase NO from flue gas. Catalyst systems consisting of vanadia supported on titania (TiO2 ) and promoted with tungsten oxide (WO3 ) or molybdenum oxide (MoO3 ) have been proven to be highly active and been commercially applied for decades [1]. However, several issues have emerged regarding the deactivation of the SCR catalyst during operation. One of the important issues is its chemical poisoning by substances in the reaction mixture. Alkali metals have been reported to be severely poisonous to vanadia based SCR catalyst. After Shikada et al. [2] found the poisoning effect of potassium on supported vanadia catalyst, numerous papers [3–13] have
∗ Corresponding author at: Key Laboratory of Low–grade Energy Utilization Technologies & Systems of Ministry of Education of China, College of Power Engineering, Chongqing University, Chongqing 400044, China. E-mail address: [email protected] (X. Du). http://dx.doi.org/10.1016/j.apsusc.2016.09.036 0169-4332/© 2016 Elsevier B.V. All rights reserved.
been published on the influence of alkali metal and its mechanism. Most of the literature suggests that the poisoning strength of the alkali metal is associated with the basicity [3,12], that is Cs >Rb >K >Na > Li. Neutralization of active acid sites was found to be the main reason for catalyst deactivation, as proved by NH3 adsorption characterizations and NO reduction tests [4,5,8]. Researchers also found that doping of alkali metal on the catalyst surface would weaken the strength of V O bond and thus lower the oxidizing ability of catalyst [6,14], which is also a reason for catalyst deactivation. Although lots of work has been done on the poisoning of SCR catalyst by alkali metals, there are still issues waiting to be studied. For instance, researchers have mostly conducted characterizations on the ‘poisoned catalyst’, which is prepared by impregnation method and calcined in the end. The reactions between various alkali species and the catalysts were ignored. Furthermore, most studies were focusing on one specific alkali metal or compound. Jensen et al. [15] have built a pilot platform for catalyst deactivation using aerosol deposit. They conducted a serial of experimental studies on the deactivation of catalyst by polyphosphoric acid [15], potassium phosphate [16], KCl and K2 SO4 [17]. They have focused on the distribution of poisons and the final physical and chemical influences on the catalyst. Their studies also indicate that alkali
X. Du et al. / Applied Surface Science 392 (2017) 162–168
163
Fig. 1. The HOMO energies of various alkali species (a) and HOMO snapshots of Na-containing molecules (b).
metals can exist in various compounds under different situations. Potassium is usually contained in biomass-firing flue gas [17] and sodium is rich in coals from some regions, such as Zhundong region in northern area of Xinjiang Province of China [18]. The alkali containing species in the flue gas of municipal solid waste combustion [19] can be more complicated due to the broad origins of waste. Other types of flue gas from glass furnaces, steel sintering furnaces, diesel engines and other combustion units also contain alkali metals in various forms. One another important factor is the temperature of the area where alkali species interacts with the catalyst, which has not been reported. In general, a study on the poisoning of SCR catalyst by alkali species concerning the poisoning strength of various species, interaction profiles and reaction temperature is still needed. Our study screened the poisoning reactivities of alkali metal hydroxide, chlorate, nitrate, sulfate and bisulfate using Density Functional Theory (DFT) calculations. Reactions between sodium compounds and the VOx -WOx /TiO2 catalyst at various temperatures were conducted to study the reacting kinetics of poisons with the catalyst. The ‘after poisoning’ effects of various alkali metals were also studied by both DFT calculations and experiments.
2. Experimental and theoretical All DFT calculations were performed using the gradient corrected Becke’s [20,21] three-parameter hybrid exchange functional in conjunction with the correlation functional of Lee, Yang, and Parr [22] (B3LYP). Throughout the theoretical study, V, K, Rb and Cs were treated by the Los Alamos set of double-zeta type basis set (LANL2DZ). For other atoms related in this study, 6-311 + (d,p) basis set was employed. Each stationary structure has been confirmed as a minimum-energy structure from the calculated vibrational frequencies. The Gaussian 03 [23] package was applied in the theoretical part. A commercial SCR catalyst composed by V2 O5 and WO3 supported on anatase TiO2 was used in this study. The loading rate of V2 O5 and WO3 are 0.8% and 3.5% based on TiO2 respectively. The catalyst was doped with various Na-containing compounds including NaOH, NaCl, NaNO3 , Na2 SO4 and NaHSO4 . These Na-containing species were added onto the catalyst surface by the impregnation method using the respective aqueous solution. The impregnation procedure is shown as follows: impregnated with respective aqueous solution, dried at 105 ◦ C for 12 h, and calcined at 500 ◦ C in air for 5 h or at stepped temperatures (200, 300, 400, 500 ◦ C) in simulated air (20% O2 and 80% N2 ). Different alkali loadings were used
in Figs. 4 and 6 to better present the differences of the data in their own figure. The catalytic activity tests for the reduction of NO by NH3 were carried out in a fixed bed micro-reactor with catalyst samples of 0.2 g. The simulated gas for these tests contained 1000 ppm NO, 5 vol% O2 , and 1000 ppm NH3 in N2 . The catalytic reactions were carried out at temperatures from 150 to 450 ◦ C under atmosphere pressure, with a total flow rate of 0.8 L min−1 and GHSV = 2.4 × 105 mL/(g*h). The NO concentrations before and after reaction were determined using an ECOM-J2KN gas analyzer manufactured in Germany. Brunauer-Emmett-Teller (BET) surface area was measured by N2 adsorption at 77 K using a Quantachrome Autosorb-1 instrument. 3. Results and discussion 3.1. Initial screening of the poisons Frontier molecular orbital (FMO) theory was firstly used to screen the reactivities of various poisons with the catalyst. As discussed in our previous studies [24,25], energy of the lowest unoccupied molecular orbital (LUMO) can be used to evaluate the acidity of catalyst site, which is crucial for the adsorption of NH3 . Lower LUMO energy indicates the easier to accept electrons from other molecules and also the stronger acidity. Alkali metals are well known for easily neutralizing the acidity of catalyst. In this study, energies of highest occupied molecular orbitals (HOMO) of the alkali-containing molecules were computed to analyze their abilities to donate electrons. The higher the HOMO energy is, the easier it is to donate its electrons to the unoccupied orbital of the acid sites of a catalyst. In another word, HOMO energy indicates the reactivity of a poison with the catalyst. In this study, the main active site, vanadia, was modeled and used to react with the poisons. The vanadia site was represented by the mono VO4 H3 cluster, which has been proven to well simulate the electronic and structural profiles of vanadia site. The structure validity has been discussed in the Supplementary material. One thing should be noticed is that LUMO energy was chosen as the electronic parameter for model validity discussion since LUMO energy is the crucial indicator to screen the poisoning strength of various substances. The calculation results show that the LUMO energy of VO4 H3 is −0.121 Hartree. Various alkali metal containing molecules in after-combustion flue gas including hydroxide, chlorate, nitrate, sulfate and bisulfate were studied. All the molecular structures and snapshots of HOMOs are shown in Supplementary material. Fig. 1 shows the HOMO energies of these molecules and HOMO snapshots of the Na-containing
164
X. Du et al. / Applied Surface Science 392 (2017) 162–168
Fig. 2. The adsorption of NH3 on the acid site and reactions between vanadia site and various alkali species. ‘M’ stands for alkali metal.
species. The HOMO snapshots (Fig. 1(b)) indicate that HOMOs are distributed around electronegative atoms like O and Cl. In these molecules, basic alkali metal atoms will donate electrons to O or Cl and thus transfer the basicity to O and Cl atoms. The out shell electrons of O and Cl atoms can interact with the LUMOs of acid sites on catalyst surface, which causes poisoning of catalyst. The HOMO energy profile in Fig. 1(a) shows that with the increase of basicity of alkali metal from Li to Cs, the corresponding HOMO energy increases. The increase from K to Cs is slower than that from Li to K. For a given alkali metal, the HOMO energies can be ranked as: hydroxide > sulfate > chlorate > nitrate > bisulfate. One exception is that HOMO of lithium chlorate is slightly higher than that of lithium sulfate. Since a molecule with higher HOMO energy means it is the easier to donate electrons to the catalyst and react with the active sites, the reactivities of these alkali substances with catalyst can be predicted. For different alkali metals, the reactivity with catalyst increases from Li to Cs. The anion effect also shows an identical trend. When combined with a same alkali metal, the reactivity can be predicted to follow the same trend with the above HOMO energies for different anions.
3.2. The reactions between poisons and catalyst The reactions between the poisons with the vanadia acid were also simulated. Fig. 2 shows the adsorption of NH3 on the vanadia acid site and the reactions between poisons and vanadia site. NH3 can be adsorbed by the Brønsted acid to form a NH3 -H- structure, with is highly active to react with gasphase NO [26]. Due to the basicities, the alkali metal containing species will neutralize the Brønsted acids of vanadia sites. For these neutralization reactions, respective acids (as shown in Fig. 2) will be produced. These reactions were simulated and the energies were calculated. Fig. 3 shows the heats of reactions between various poisons and VO4 H3 model. The reaction heat was calculated as E(products) – E(reactants), in which E is the zero-point energy. The structures of the products of these reactions are shown in the Supplementary material. As can be seen in Fig. 3, the reaction energy is more depending on anions than on cations. The cations can affect the reaction energy only in sulfate and hydroxide. For sulfate and hydroxide, the reaction energy decreases from Li to Cs. The lower the reaction energy is, the more the balance will move to the product side and the more favorable the reaction can happen. This indicates that Cs2 SO4 and CsOH can neutralize the vanadia sites more readily than Li2 SO4 and LiOH respectively. For MCl, MNO3 and MHSO4 (M stands for alkali metal), however, different alkali metal does not affect the reaction energy obviously.
Fig. 3. Energies profiles of the reactions between alkali species and VO4 H3 molecule. VO4 H3 and VO4 H2 M molecules are denoted as VOH and VOM respectively in this figure.
The influence of anion is remarkable, which is due to that the HOMOs are more concentrated around anions, as shown in Fig. 1(b). The overall energy sequence can be described as: E (MOH) < E (M2 SO4 ) < E(MCl) < E(MNO3 ) < E(MHSO4 ), in which for example, E (MOH) means the reaction heat of MOH with VO4 H3 model. Exceptions can only be found for Li2 SO4 and Na2 SO4 , which do not obey this order. This can possibly be ascribed to the tight bonding between sulfate anion and alkali cations. These data also indicate that besides the case of MHSO4 , other reactions concerning MOH, M2 SO4 , MCl and MNO3 are all exothermic. This indicates that the reactions of MHSO4 with the catalyst is thermodynamically unfavorable and others are favorable. Another noticeable fact is that the sequences of HOMO energies and the reaction heats are related. The HOMO energies were ranked as E(MOH) > E (M2 SO4 ) > E(MCl) > E(MNO3 ) > E(MHSO4 ), which is the opposite of the reaction energy sequence. This well proves our theory that a molecule with higher HOMO energy implies it is more favorable to react with the vanadia site. As shown in Fig. 2, the reactions B, C and D can generate volatile (HCl, H2 O) and decomposable (HNO3 ) products. Thus, the anions of MCl, MOH and MNO3 can easily escape from the catalyst surface after neutralization. While, for M2 SO4 and MHSO4 , the products are MHSO4 and H2 SO4 respectively, which are not highly volatile or decomposable. The reaction to produce escapable SO3 was also calculated and shown in Fig. 3. The reaction of MHSO4 with VO4 H3
X. Du et al. / Applied Surface Science 392 (2017) 162–168
165
Fig. 4. NO conversions of the catalyst doped with (a) deionized water, (b) NaOH, (c) NaCl, (d) NaNO3 , (d) Na2 SO4 , (f) NaHSO4 (loading rate of alkali metal is 40 mol% based on vanadium) at stepped temperatures (200, 300, 400, 500 ◦ C).
molecule to produce VO4 H2 M, H2 O and SO3 is endothermic by 0.75 ev, despite of different alkali cations. Thus, for sulfate and bisulfate poisons, the anions are hard to escape from the catalyst surface. As published, the SO4 2− and HSO4 − left on the catalyst surface can provide acid sites and benefit for NH3 adsorption [27]. Consequently, for M2 SO4 and MHSO4 , the neutralization of catalyst acidity and generation of new acid sites can happen simultaneously. Experiments were performed to further study the reactions between poisons and the catalyst sites. Sodium containing com-
pounds were used as the representatives for various alkali metal containing compounds. Reactions at elevated temperatures from 200 to 500 ◦ C were studied to analyze the reactivities of various alkali species with the catalyst. The VOx -WOx /TiO2 catalyst was first impregnated with respective sodium species and dried at 105 ◦ C. In Fig. 4, the catalyst was impregnated with deionized water (a), NaOH (b),NaCl (c), NaNO3 (d), Na2 SO4 (d) and NaHSO4 (f) solutions. The samples were then calcined at 200 ◦ C in simulated air (20% O2 and 80% N2 ) and tested for NO reduction at temperatures
166
X. Du et al. / Applied Surface Science 392 (2017) 162–168
Table 1 BET surface area, total pore volume and average pore size of the fresh catalyst and poisoned catalysts calcined at different temperatures.
Fresh Calcined at 300 ◦ C Calcined at 400 ◦ C Calcined at 500 ◦ C
BET surface area (m2 /g)
Total pore volume (cm3 /g)
Average pore size (Å)
51.05 49.79 51.37 47.03
0.2164 0.2115 0.2168 0.2068
169.58 169.96 168.84 175.92
up to 200 ◦ C. After that, the samples were calcined in-situ at lifted temperatures (300, 400 and 500 ◦ C) and SCR activities were tested at temperatures below the calcining temperature. The NO reduction activities of various samples calcined at stepped temperatures are shown in Fig. 4. In Fig. 4(a), the activity difference between the fresh sample calcined at 200 ◦ C and that calcined at 300 ◦ C is neglectable. Raising the calcining temperature to 400 ◦ C causes a slight decrease of NO conversion. Higher calcining temperature at 500 ◦ C leads to a larger activity decrease. Activity decrease can be caused by chemical or physical reasons. High calcining temperature can result in sintering of catalyst and then lead to decrease of surface area. In Table 1, the BET surface area and pore volume are listed for the catalyst calcined at different temperatures. Impregnating with deionized water and calcining at 300 ◦ C slightly lowers the surface area of the fresh catalyst. Calcining at 400 ◦ C recovers the surface area back to more than 51 m2 /g. While, as shown in Fig. 4(a), calcining at 400 ◦ C causes a slight decrease of NO conversion. This can be attributed to the chemical change of the catalyst surface, such as the dehydration of two Brønsted acid. Calcining at 500 ◦ C leads to a slight sintering of the catalyst. Thus, the decrease of NO conversion by calcining at 500 ◦ C can be due to a combination of chemical and physical transformation. For the samples impregnated with poisons, Fig. 4(b–f) gives the NO conversion profiles. For the NaOH doped sample, the calcining temperature does not influence the SCR performance evidently. The catalyst activity drops to a low level after being calcined at 200 ◦ C and maintains stably although the calcining temperature increases. This indicates that the reaction between NaOH and the catalyst happens at temperature lower than 200 ◦ C. The influences of NaCl and MNO3 are similar to each other. These two samples show noticeable drops of NO conversion after being calcined at 400 ◦ C. This is possibly due to the chemical reaction between NaCl/NaNO3 and the catalyst when the temperature increases to 400 ◦ C. The difference between the sample doped with NaCl (NaNO3 ) and NaOH is noticeable. NaOH reacts with the catalyst at much lower temperature and thus it is more active to the catalyst surface. This is identical with the theoretical calculation results that NaOH has a much higher HOMO energy (Fig. 1) and its reaction with the vanadia site is much more thermodynamically favorable (Fig. 3). The reaction of NaCl and NaNO3 with the SCR catalyst can also generate volatile HCl and decomposable HNO3 respectively. Energy profile of HNO3 decomposition can be found in the Supplementary material. But NaCl and NaNO3 possess lower theoretical HOMO energies than NaOH and thus they are less active to the catalyst than NaOH. Consequently, a noticeable activity drop can be found for NaCl or NaNO3 doped sample when calcined at a higher temperature of 400 ◦ C. The influences of Na2 SO4 and NaHSO4 are much different from the above three compounds. Doping of Na2 SO4 does not inhibit but even slightly enhances the SCR reaction of the catalyst. Based on the reaction stoichiometry, H2 O, HCl and HNO3 will be produced for MOH, MCl and MNO3 after reacting with V-OH. While for M2 SO4 and MHSO4 , the products are MHSO4 and H2 SO4 respectively. H2 O and HCl are gaseous species under SCR reaction condition. HNO3 and H2 SO4 need to decompose into acid anhydride and H2 O to be volatile. We have searched for possible decomposition paths for HNO3 and H2 SO4 . The direct decomposing products for H2 SO4
Fig. 5. LUMO energies and orbital snapshots of the fresh VO4 H3 and doped VO4 H2 M.
can be H2 O + SO3 or H2 O2 + SO2 based on stoichiometry. Two HNO3 can react with each other to produce H2 O and N2 O5 . The reaction paths are shown in Fig. S5. The easier path (both in reaction energy and energy barrier) for H2 SO4 to decompose is to split into H2 O and SO3 . This process is endothermic by 15 Kcal/mol and needs to cross a barrier of 34 Kcal/mol. While the reaction energy and barrier for decomposition of HNO3 are +11 and +21 Kcal/mol respectively. Thus, H2 SO4 need to consume 13 Kcal/mol more energy to escape from the catalyst surface. This is in accordance with our experimental XPS results (not shown here) that S element can be detected on the doped catalyst. Therefore, the sulfate ion can hardly escape from the catalyst surface after reaction. Both cation Na+ and anion SO4 2− can deposit on the catalyst surface. As reported, sulfate can generate active Brønsted acid sites [27]. Consequently, doping of Na2 SO4 causes a dual effect on the catalyst. On one hand, the cation Na+ neutralizes the Brønsted acid. On the other hand, the cation SO4 2− can generate new Brønsted acid. This is the reason why Na2 SO4 does not lower the activity of NO reduction of the VOx -WOx /TiO2 catalyst. The promotional effect of NaHSO4 is stronger than that of Na2 SO4 . The reason is similar as that of Na2 SO4 . The anion HSO4 − will deposit on the catalyst surface under our reaction condition, which can generate active Brønsted acid sites. The acidity of HSO4 − is higher than SO4 2− and thus promotional effect of NaHSO4 is stronger than that of NaSO4 . The surface sulfation and its influence was also modeled in this study. According to Chen et al. [12], TiO2 can easily be sulfated. Thus, the sulfation of VOx -WOx /TiO2 was modeled by the sulfation of TiO2 . A dual site TiOx cluster was cut from the surface to represent TiO2 . The structures and LUMOs are shown in Table S3 of the Supplementary material. The LUMOs are mainly composed of the Ti 3d orbitals before and after sulfation. The LUMO energy will be changed from −0.074 to −0.120 due to sulfation, which indicates that the acidity of TiOx will be promoted by sulfation. 3.3. Poisoning by various alkali metals The influences of various alkali metals are varied to each other. In Fig. 5, the LUMO energies and snapshots of the fresh and doped VO4 H3 molecules are shown. LUMO energy can be an indicator to analyze the ability to attract electrons from other molecules. The lower the energy of a LUMO is, the easier it is to attract electrons. For the VO4 H3 molecule, the LUMO are mainly composed of V 3d orbital. It has a low LUMO energy to easily accept electrons and thus possesses a high acidity. Doping of Li does not influence the location of LUMO largely. Distinguished difference of the LUMO
X. Du et al. / Applied Surface Science 392 (2017) 162–168
167
Table 2 Model characters before and after being doped with alkali metals. Catalyst model
Bond length of V O(Å)
Stretching Frequency of V O (cm−1 )
Mulliken Charge of V (e)
Mulliken Charge of alkali metal (e)
LUMOEnergy (Hartree)
Undoped Li-doped Na-doped K-doped Rb-doped Cs-doped
1.565 1.645 1.642 1.635 1.634 1.632
1128 974 976 990 992 996
+1.219 +1.125 +1.093 +1.079 +1.094 +1.090
– +0.501 +0.789 +0.976 +0.957 +0.936
−0.121 −0.087 −0.071 −0.062 −0.063 −0.059
Fig. 6. The NO conversions of the VOx -WOx /TiO2 catalyst poisoned by LiCl, NaCl and KCl (loading rate of alkali metal is 20 mol% based on vanadium).
snapshot can be found for the molecules doped with Na, K, Rb or Cs element. For these molecules, the LUMO shifts to the alkali cations. Meanwhile, the LUMO energy increases noticeably after being doped with alkali metal. The increases from Li to K are noteworthy. From K to Cs, the LUMO energy maintains stably at a high level. These results indicate that alkali metals will lower the acidity of vanadia in different degrees. The influence increases from Li to K. The model characters also change after being doped with alkali metals. As shown in Table 2, doping of alkali metal results in elongation of the V O bond length and redshift of the stretching frequency of V O. These data indicate that the V O bond is weakened due to these alkali metals. Charge analysis also shows that the V atom will be reduced by the doping of alkali metal. Reduction of V atom causes the more satisfaction of its outer shell and thus the ability to attract electrons decreases. This is in accordance with the change of LUMO energies. Experiments were performed to test the poisoning strengths of various alkali metals. The performances of the catalysts poisoned by LiCl, NaCl and KCl are shown in Fig. 6. The NO conversions at the testing temperatures decline with the doping of alkali chlorate. The NO conversions follow the order of fresh sample >LiCl-doped > NaCl-doped > KCl-doped. Thus, the poisoning strength order is KCl > NaCl > LiCl, which is in accordance of our theoretical results. 4. Conclusions Various alkali metal-containing substances can cause different influences to the VOx -WOx /TiO2 SCR catalyst. HOMO energies of these substances can be used to analyze their reactivities with the acid catalyst. The initial screen by HOMO energy shows that the reactivities with SCR catalyst can be ranked
as E (MOH) > E (M2 SO4 ) > E(MCl) > E(MNO3 ) > E(MHSO4 ) for the anion aspect. For different cations, the poisoning strength follows as Cs > Rb > K > Na > Li. The increase from Li to K is more obivious than that form K to Cs. The theoretical thermodynamics study on the poisoning reactions presents the same order. Unlike the MOH, MCl and MNO3 , which can produce volatile species after reaction with the catalyst, M2 SO4 and MHSO4 will leave both cations and anions on the catalyst surface after deposite. The experimental SCR activities after poisoning are NaHSO4 -doped > Na2 SO4 -doped > undoped > NaNO3 doped ≈ NaCl-doped ≈ NaOH-doped. This confirms that the sulfate ions which can generate active acid sits are left on the catalyst after deposite. The calcining of doped catalyst at various temperatures also indicates that NaOH can react with the catalyst at temperature below 200 ◦ C, while NaCl and NaNO3 will react with the catalyst at temperature between 300 and 400 ◦ C. The after reaction-influences of various alkali metals are different. The LUMO energy of the VO4 H3 molecule increases with the doping of alkali metal. The LUMO energy follows the order of (Cs, Rb, K-doped) > Na-doped > Li-doped, which is also the order of poisoning strengths of various alkali metals. The activity test shows that NO conversion decreases as Fresh > Li-doped > Na-doped > Kdoped, which implies that the experimental poisoning strength is in accordance with the theoretical result. Acknowledgements We gratefully acknowledge the financial support of the National Natural Science Foundation of China (51506015), Fundamental & Advanced Research Projects of Chongqing (cstc2015jcyjA20006), Chongqing social undertakings and livelihood security special projects (cstc2015shmszx20004), Fundamental Research Funds for the Central Universities (106112016CDJXY145503, 106112016CDJZR145507) and general financial grant from the China postdoctoral science foundation (2015M570768). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apsusc.2016.09. 036. References [1] G. Busca, L. Lietti, G. Ramis, F. Berti, Chemical and mechanistic aspects of the selective catalytic reduction of NOx by ammonia over oxide catalysts: a review, Appl. Catal. B 18 (1998) 1–36. [2] T. Shikada, K. Fujimoto, Effect of added alkali salts on the activities of supported vanadium-oxide catalysts for nitric-oxide reduction, Chem. Lett. 12 (1983) 77–80. [3] J.P. Chen, M.A. Buzanowski, R.T. Yang, J.E. Cichanowicz, Deactivation of the vanadia catalyst in the selective catalytic reduction process, J. Air Waste Manage. 40 (1990) 403–1409. [4] Y.J. Zheng, A.D. Jensen, J.E. Johnsson, Laboratory investigation of selective catalytic reduction catalysts: deactivation by potassium compounds and catalyst regeneration, Ind. Eng. Chem. Res. 43 (2004) 941–947.
168
X. Du et al. / Applied Surface Science 392 (2017) 162–168
[5] Y.J. Zheng, A.D. Jensen, J.E. Johnsson, Deactivation of V2 O5 -WO3 -TiO2 SCR catalyst at a biomass-fired combined heat and power plant, Appl. Catal. B 60 (2005) 253–264. [6] J. Due-Hansen, S. Boghosian, A. Kustov, P. Fristrup, G. Tsilomelekis, K. Stahl, C.H. Christensen, R. Fehrmann, Vanadia-based SCR catalysts supported on tungstated and sulfated zirconia: influence of doping with potassium, J. Catal. 251 (2007) 459–473. [7] Å. Kling, C. Andersson, Å. Myringer, D. Eskilsson, S.G. Järås, Alkali deactivation of high-dust SCR catalysts used for NOx reduction exposed to flue gas from 100 MW-scale biofuel and peat fired boilers: influence of flue gas composition, Appl. Catal. B 69 (2007) 240–251. [8] O. Krocher, M. Elsener, Chemical deactivation of V2 O5 /WO3 -TiO2 SCR catalysts by additives and impurities from fuels lubrication oils, and urea solution – I. Catalytic studies, Appl. Catal. B 77 (2008) 215–227. [9] A.E. Lewandowska, M. Calatayud, E. Lozano-Diz, C. Minot, M.A. Banares, Combining theoretical description with experimental in situ studies on the effect of alkali additives on the structure and reactivity of vanadium oxide supported catalysts, Catal. Today 139 (2008) 209–213. [10] D. Nicosia, I. Czekaj, O. Krocher, Chemical deactivation of V2 O5 /WO3 -TiO2 SCR catalysts by additives and impurities from fuels lubrication oils and urea solution – Part II. Characterization study of the effect of alkali and alkaline earth metals, Appl. Catal. B 77 (2008) 228–236. [11] Y.J. Zheng, A.D. Jensen, J.E. Johnsson, J.R. Thogersen, Deactivation of V2 O5 -WO3 -TiO2 SCR catalyst at biomass fired power plants: elucidation of mechanisms by lab- and pilot-scale experiments, Appl. Catal. B 83 (2008) 186–194. [12] L. Chen, J. Li, M. Ge, The poisoning effect of alkali metals doping over nano V2 O5 –WO3 /TiO2 catalysts on selective catalytic reduction of NOx by NH3, Chem. Eng. J. 170 (2011) 531–537. [13] F. Castellino, A.D. Jensen, J.E. Johnsson, R. Fehrmann, Influence of reaction products of K-getter fuel additives on commercial vanadia-based SCR catalysts part II. Simultaneous addition of KCl, Ca(OH)2 H3 PO4 and H2 SO4 in a hot flue gas at a SCR pilot-scale setup, Appl. Catal. B 86 (2009) 206–215. [14] H. Si-Ahmed, M. Calatayud, C. Minot, E.L. Diz, A.E. Lewandowska, M.A. Banares, Combining theoretical description with experimental in situ studies on the effect of potassium on the structure and reactivity of titania-supported vanadium oxide catalyst, Catal. Today 126 (2007) 96–102. [15] F. Castellino, S.B. Rasmussen, A.D. Jensen, J.E. Johnsson, R. Fehrmann, Deactivation of vanadia-based commercial SCR catalysts by polyphosphoric acids, Appl Catal. B 83 (2008) 110–122. [16] F. Castellino, A.D. Jensen, J.E. Johnsson, R. Fehrmann, Influence of reaction products of K-getter fuel additives on commercial vanadia-based SCR catalysts part I. Potassium phosphate, Appl. Catal. B 86 (2009) 196–205.
[17] Y.J. Zheng, A.D. Jensen, J.E. Johnsson, J.R. Thogersen, Deactivation of V2 O5 -WO3 -TiO2 SCR catalyst at biomass fired power plants: elucidation of mechanisms by lab- and pilot-scale experiments, Appl. Catal. B 83 (2008) 186–194. [18] Y. He, J.J. Zhu, B. Li, Z.H. Wang, Z.S. Li, M. Alden, K.F. Cen, In-situ measurement of sodium and potassium release during oxy-fuel combustion of lignite using laser-induced breakdown spectroscopy: effects of O−2 and CO2 concentration, Energy Fuel 27 (2013) 1123–1130. [19] L. Lisi, G. Lasorella, S. Malloggi, G. Russo, Single and combined deactivating effect of alkali metals and HCl on commercial SCR catalysts, Appl. Catal. B 50 (2004) 251–258. [20] A.D. Becke, Density-functional exchange-energy approximation with correct asymptotic-behavior, Phys. Rev. A 38 (1988) 3098–3100. [21] A.D. Becke, Density-functional thermochemistry. 3. The role of exact exchange, J. Chem. Phys. 98 (1993) 5648–5652. [22] C.T. Lee, W.T. Yang, R.G. Parr, Development of the colle-salvetti correlation-energy formula into a functional of the electron-density, Phys. Rev. B 37 (1988) 785–789. [23] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, J.A. Montgomery, Jr., T. Vreven, K.N. Kudin, J.C. Burant, J.M. Millam, S.S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox, H.P. Hratchian, J.B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, P.Y. Ayala, K. Morokuma, G.A. Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich, A.D. Daniels, M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J. Cioslowski, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, C. Gonzalez, J.A. Pople, Gaussian 03, Revision B.02, Gaussian Inc., Pittsburgh PA, 2003. [24] X. Du, X. Gao, W. Hu, J. Yu, Z. Luo, K. Cen, Catalyst design based on DFT calculations: metal oxide catalysts for gas phase NO reduction, J. Phys. Chem. C 118 (2014) 13617–13622. [25] X. Du, X. Gao, K. Qiu, Z. Luo, K. Cen, The reaction of poisonous alkali oxides with vanadia SCR catalyst and the afterward influence: a DFT and experimental study, J. Phys. Chem. C 119 (2015) 1905–1912. [26] M. Anstrom, N.Y. Topsoe, J.A. Dumesic, Density functional theory studies of mechanistic aspects of the SCR reaction on vanadium oxide catalysts, J. Catal. 213 (2003) 115–125. [27] J.P. Chen, R.T. Yang, Selective catalytic reduction of NO with NH3 on SO4 2 -/TiO2 superacid catalyst, J. Catal 139 (1993) 277–288.
Contents lists available at ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
The different poisoning behaviors of various alkali metal containing compounds on SCR catalyst Xuesen Du a,b,∗ , Guangpeng Yang b , Yanrong Chen a,b , Jingyu Ran a,b , Li Zhang a,b a Key Laboratory of Low-grade Energy Utilization Technologies & Systems of Ministry of Education of China, College of Power Engineering, Chongqing University, Chongqing 400044, China b Institute of Energy and Environment of Chongqing University, Chongqing, 400044, China
a r t i c l e
i n f o
Article history: Received 13 May 2016 Received in revised form 9 September 2016 Accepted 9 September 2016 Available online 12 September 2016 Keywords: Selective catalytic reduction Poisoning Alkali metal Density functional theory Vanadia
a b s t r a c t Alkali metals are poisonous to the metal oxide catalyst for NO removal. The chemical configuration of alkali containing substance and interacting temperature can affect the poisoning profile. A computational method based on Frontier Molecular Orbital analysis was proposed to determine the reacting behavior of various alkali-containing substances with SCR catalyst. The results reveal that the poisoning reactivities of various substances can be ranked as: E (MOH) > E (M2 SO4 ) > E(MCl) > E(MNO3 ) > E(MHSO4 ). The experimental activity tests of the catalysts calcined at stepped temperatures show that NaOH can react with the catalyst below 200 ◦ C. NaCl and NaNO3 start to react with the catalyst at a temperature between 300 and 400 ◦ C. Unlike MOH, MCl and MNO3 , which can produce volatile or decomposable species for the anions after reacting with the catalyst, M2 SO4 and MHSO4 will leave both cations and anions on the catalyst surface. The sulfate ions left on the catalyst can generate active acid sites for NH3 adsorption. The experimental results also show that Na2 SO4 and NaHSO4 will not lower the NO conversion. The afterreaction influences of various alkali metals were studied using theoretical and experimental methods. The theoretical results show that the acidity decreases with doping of alkali metal. Experiments show a consistent result that the NO conversion decreases as undoped >LiCl > NaCl > KCl. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Nitric oxides (NOx ), mostly nitric monoxide (NO), emitted from stationary and mobile sources is thought to be one of the major pollutants in atmosphere. Selective catalytic reduction (SCR) is the most efficient and commonly used technologies to remove gas phase NO from flue gas. Catalyst systems consisting of vanadia supported on titania (TiO2 ) and promoted with tungsten oxide (WO3 ) or molybdenum oxide (MoO3 ) have been proven to be highly active and been commercially applied for decades [1]. However, several issues have emerged regarding the deactivation of the SCR catalyst during operation. One of the important issues is its chemical poisoning by substances in the reaction mixture. Alkali metals have been reported to be severely poisonous to vanadia based SCR catalyst. After Shikada et al. [2] found the poisoning effect of potassium on supported vanadia catalyst, numerous papers [3–13] have
∗ Corresponding author at: Key Laboratory of Low–grade Energy Utilization Technologies & Systems of Ministry of Education of China, College of Power Engineering, Chongqing University, Chongqing 400044, China. E-mail address: [email protected] (X. Du). http://dx.doi.org/10.1016/j.apsusc.2016.09.036 0169-4332/© 2016 Elsevier B.V. All rights reserved.
been published on the influence of alkali metal and its mechanism. Most of the literature suggests that the poisoning strength of the alkali metal is associated with the basicity [3,12], that is Cs >Rb >K >Na > Li. Neutralization of active acid sites was found to be the main reason for catalyst deactivation, as proved by NH3 adsorption characterizations and NO reduction tests [4,5,8]. Researchers also found that doping of alkali metal on the catalyst surface would weaken the strength of V O bond and thus lower the oxidizing ability of catalyst [6,14], which is also a reason for catalyst deactivation. Although lots of work has been done on the poisoning of SCR catalyst by alkali metals, there are still issues waiting to be studied. For instance, researchers have mostly conducted characterizations on the ‘poisoned catalyst’, which is prepared by impregnation method and calcined in the end. The reactions between various alkali species and the catalysts were ignored. Furthermore, most studies were focusing on one specific alkali metal or compound. Jensen et al. [15] have built a pilot platform for catalyst deactivation using aerosol deposit. They conducted a serial of experimental studies on the deactivation of catalyst by polyphosphoric acid [15], potassium phosphate [16], KCl and K2 SO4 [17]. They have focused on the distribution of poisons and the final physical and chemical influences on the catalyst. Their studies also indicate that alkali
X. Du et al. / Applied Surface Science 392 (2017) 162–168
163
Fig. 1. The HOMO energies of various alkali species (a) and HOMO snapshots of Na-containing molecules (b).
metals can exist in various compounds under different situations. Potassium is usually contained in biomass-firing flue gas [17] and sodium is rich in coals from some regions, such as Zhundong region in northern area of Xinjiang Province of China [18]. The alkali containing species in the flue gas of municipal solid waste combustion [19] can be more complicated due to the broad origins of waste. Other types of flue gas from glass furnaces, steel sintering furnaces, diesel engines and other combustion units also contain alkali metals in various forms. One another important factor is the temperature of the area where alkali species interacts with the catalyst, which has not been reported. In general, a study on the poisoning of SCR catalyst by alkali species concerning the poisoning strength of various species, interaction profiles and reaction temperature is still needed. Our study screened the poisoning reactivities of alkali metal hydroxide, chlorate, nitrate, sulfate and bisulfate using Density Functional Theory (DFT) calculations. Reactions between sodium compounds and the VOx -WOx /TiO2 catalyst at various temperatures were conducted to study the reacting kinetics of poisons with the catalyst. The ‘after poisoning’ effects of various alkali metals were also studied by both DFT calculations and experiments.
2. Experimental and theoretical All DFT calculations were performed using the gradient corrected Becke’s [20,21] three-parameter hybrid exchange functional in conjunction with the correlation functional of Lee, Yang, and Parr [22] (B3LYP). Throughout the theoretical study, V, K, Rb and Cs were treated by the Los Alamos set of double-zeta type basis set (LANL2DZ). For other atoms related in this study, 6-311 + (d,p) basis set was employed. Each stationary structure has been confirmed as a minimum-energy structure from the calculated vibrational frequencies. The Gaussian 03 [23] package was applied in the theoretical part. A commercial SCR catalyst composed by V2 O5 and WO3 supported on anatase TiO2 was used in this study. The loading rate of V2 O5 and WO3 are 0.8% and 3.5% based on TiO2 respectively. The catalyst was doped with various Na-containing compounds including NaOH, NaCl, NaNO3 , Na2 SO4 and NaHSO4 . These Na-containing species were added onto the catalyst surface by the impregnation method using the respective aqueous solution. The impregnation procedure is shown as follows: impregnated with respective aqueous solution, dried at 105 ◦ C for 12 h, and calcined at 500 ◦ C in air for 5 h or at stepped temperatures (200, 300, 400, 500 ◦ C) in simulated air (20% O2 and 80% N2 ). Different alkali loadings were used
in Figs. 4 and 6 to better present the differences of the data in their own figure. The catalytic activity tests for the reduction of NO by NH3 were carried out in a fixed bed micro-reactor with catalyst samples of 0.2 g. The simulated gas for these tests contained 1000 ppm NO, 5 vol% O2 , and 1000 ppm NH3 in N2 . The catalytic reactions were carried out at temperatures from 150 to 450 ◦ C under atmosphere pressure, with a total flow rate of 0.8 L min−1 and GHSV = 2.4 × 105 mL/(g*h). The NO concentrations before and after reaction were determined using an ECOM-J2KN gas analyzer manufactured in Germany. Brunauer-Emmett-Teller (BET) surface area was measured by N2 adsorption at 77 K using a Quantachrome Autosorb-1 instrument. 3. Results and discussion 3.1. Initial screening of the poisons Frontier molecular orbital (FMO) theory was firstly used to screen the reactivities of various poisons with the catalyst. As discussed in our previous studies [24,25], energy of the lowest unoccupied molecular orbital (LUMO) can be used to evaluate the acidity of catalyst site, which is crucial for the adsorption of NH3 . Lower LUMO energy indicates the easier to accept electrons from other molecules and also the stronger acidity. Alkali metals are well known for easily neutralizing the acidity of catalyst. In this study, energies of highest occupied molecular orbitals (HOMO) of the alkali-containing molecules were computed to analyze their abilities to donate electrons. The higher the HOMO energy is, the easier it is to donate its electrons to the unoccupied orbital of the acid sites of a catalyst. In another word, HOMO energy indicates the reactivity of a poison with the catalyst. In this study, the main active site, vanadia, was modeled and used to react with the poisons. The vanadia site was represented by the mono VO4 H3 cluster, which has been proven to well simulate the electronic and structural profiles of vanadia site. The structure validity has been discussed in the Supplementary material. One thing should be noticed is that LUMO energy was chosen as the electronic parameter for model validity discussion since LUMO energy is the crucial indicator to screen the poisoning strength of various substances. The calculation results show that the LUMO energy of VO4 H3 is −0.121 Hartree. Various alkali metal containing molecules in after-combustion flue gas including hydroxide, chlorate, nitrate, sulfate and bisulfate were studied. All the molecular structures and snapshots of HOMOs are shown in Supplementary material. Fig. 1 shows the HOMO energies of these molecules and HOMO snapshots of the Na-containing
164
X. Du et al. / Applied Surface Science 392 (2017) 162–168
Fig. 2. The adsorption of NH3 on the acid site and reactions between vanadia site and various alkali species. ‘M’ stands for alkali metal.
species. The HOMO snapshots (Fig. 1(b)) indicate that HOMOs are distributed around electronegative atoms like O and Cl. In these molecules, basic alkali metal atoms will donate electrons to O or Cl and thus transfer the basicity to O and Cl atoms. The out shell electrons of O and Cl atoms can interact with the LUMOs of acid sites on catalyst surface, which causes poisoning of catalyst. The HOMO energy profile in Fig. 1(a) shows that with the increase of basicity of alkali metal from Li to Cs, the corresponding HOMO energy increases. The increase from K to Cs is slower than that from Li to K. For a given alkali metal, the HOMO energies can be ranked as: hydroxide > sulfate > chlorate > nitrate > bisulfate. One exception is that HOMO of lithium chlorate is slightly higher than that of lithium sulfate. Since a molecule with higher HOMO energy means it is the easier to donate electrons to the catalyst and react with the active sites, the reactivities of these alkali substances with catalyst can be predicted. For different alkali metals, the reactivity with catalyst increases from Li to Cs. The anion effect also shows an identical trend. When combined with a same alkali metal, the reactivity can be predicted to follow the same trend with the above HOMO energies for different anions.
3.2. The reactions between poisons and catalyst The reactions between the poisons with the vanadia acid were also simulated. Fig. 2 shows the adsorption of NH3 on the vanadia acid site and the reactions between poisons and vanadia site. NH3 can be adsorbed by the Brønsted acid to form a NH3 -H- structure, with is highly active to react with gasphase NO [26]. Due to the basicities, the alkali metal containing species will neutralize the Brønsted acids of vanadia sites. For these neutralization reactions, respective acids (as shown in Fig. 2) will be produced. These reactions were simulated and the energies were calculated. Fig. 3 shows the heats of reactions between various poisons and VO4 H3 model. The reaction heat was calculated as E(products) – E(reactants), in which E is the zero-point energy. The structures of the products of these reactions are shown in the Supplementary material. As can be seen in Fig. 3, the reaction energy is more depending on anions than on cations. The cations can affect the reaction energy only in sulfate and hydroxide. For sulfate and hydroxide, the reaction energy decreases from Li to Cs. The lower the reaction energy is, the more the balance will move to the product side and the more favorable the reaction can happen. This indicates that Cs2 SO4 and CsOH can neutralize the vanadia sites more readily than Li2 SO4 and LiOH respectively. For MCl, MNO3 and MHSO4 (M stands for alkali metal), however, different alkali metal does not affect the reaction energy obviously.
Fig. 3. Energies profiles of the reactions between alkali species and VO4 H3 molecule. VO4 H3 and VO4 H2 M molecules are denoted as VOH and VOM respectively in this figure.
The influence of anion is remarkable, which is due to that the HOMOs are more concentrated around anions, as shown in Fig. 1(b). The overall energy sequence can be described as: E (MOH) < E (M2 SO4 ) < E(MCl) < E(MNO3 ) < E(MHSO4 ), in which for example, E (MOH) means the reaction heat of MOH with VO4 H3 model. Exceptions can only be found for Li2 SO4 and Na2 SO4 , which do not obey this order. This can possibly be ascribed to the tight bonding between sulfate anion and alkali cations. These data also indicate that besides the case of MHSO4 , other reactions concerning MOH, M2 SO4 , MCl and MNO3 are all exothermic. This indicates that the reactions of MHSO4 with the catalyst is thermodynamically unfavorable and others are favorable. Another noticeable fact is that the sequences of HOMO energies and the reaction heats are related. The HOMO energies were ranked as E(MOH) > E (M2 SO4 ) > E(MCl) > E(MNO3 ) > E(MHSO4 ), which is the opposite of the reaction energy sequence. This well proves our theory that a molecule with higher HOMO energy implies it is more favorable to react with the vanadia site. As shown in Fig. 2, the reactions B, C and D can generate volatile (HCl, H2 O) and decomposable (HNO3 ) products. Thus, the anions of MCl, MOH and MNO3 can easily escape from the catalyst surface after neutralization. While, for M2 SO4 and MHSO4 , the products are MHSO4 and H2 SO4 respectively, which are not highly volatile or decomposable. The reaction to produce escapable SO3 was also calculated and shown in Fig. 3. The reaction of MHSO4 with VO4 H3
X. Du et al. / Applied Surface Science 392 (2017) 162–168
165
Fig. 4. NO conversions of the catalyst doped with (a) deionized water, (b) NaOH, (c) NaCl, (d) NaNO3 , (d) Na2 SO4 , (f) NaHSO4 (loading rate of alkali metal is 40 mol% based on vanadium) at stepped temperatures (200, 300, 400, 500 ◦ C).
molecule to produce VO4 H2 M, H2 O and SO3 is endothermic by 0.75 ev, despite of different alkali cations. Thus, for sulfate and bisulfate poisons, the anions are hard to escape from the catalyst surface. As published, the SO4 2− and HSO4 − left on the catalyst surface can provide acid sites and benefit for NH3 adsorption [27]. Consequently, for M2 SO4 and MHSO4 , the neutralization of catalyst acidity and generation of new acid sites can happen simultaneously. Experiments were performed to further study the reactions between poisons and the catalyst sites. Sodium containing com-
pounds were used as the representatives for various alkali metal containing compounds. Reactions at elevated temperatures from 200 to 500 ◦ C were studied to analyze the reactivities of various alkali species with the catalyst. The VOx -WOx /TiO2 catalyst was first impregnated with respective sodium species and dried at 105 ◦ C. In Fig. 4, the catalyst was impregnated with deionized water (a), NaOH (b),NaCl (c), NaNO3 (d), Na2 SO4 (d) and NaHSO4 (f) solutions. The samples were then calcined at 200 ◦ C in simulated air (20% O2 and 80% N2 ) and tested for NO reduction at temperatures
166
X. Du et al. / Applied Surface Science 392 (2017) 162–168
Table 1 BET surface area, total pore volume and average pore size of the fresh catalyst and poisoned catalysts calcined at different temperatures.
Fresh Calcined at 300 ◦ C Calcined at 400 ◦ C Calcined at 500 ◦ C
BET surface area (m2 /g)
Total pore volume (cm3 /g)
Average pore size (Å)
51.05 49.79 51.37 47.03
0.2164 0.2115 0.2168 0.2068
169.58 169.96 168.84 175.92
up to 200 ◦ C. After that, the samples were calcined in-situ at lifted temperatures (300, 400 and 500 ◦ C) and SCR activities were tested at temperatures below the calcining temperature. The NO reduction activities of various samples calcined at stepped temperatures are shown in Fig. 4. In Fig. 4(a), the activity difference between the fresh sample calcined at 200 ◦ C and that calcined at 300 ◦ C is neglectable. Raising the calcining temperature to 400 ◦ C causes a slight decrease of NO conversion. Higher calcining temperature at 500 ◦ C leads to a larger activity decrease. Activity decrease can be caused by chemical or physical reasons. High calcining temperature can result in sintering of catalyst and then lead to decrease of surface area. In Table 1, the BET surface area and pore volume are listed for the catalyst calcined at different temperatures. Impregnating with deionized water and calcining at 300 ◦ C slightly lowers the surface area of the fresh catalyst. Calcining at 400 ◦ C recovers the surface area back to more than 51 m2 /g. While, as shown in Fig. 4(a), calcining at 400 ◦ C causes a slight decrease of NO conversion. This can be attributed to the chemical change of the catalyst surface, such as the dehydration of two Brønsted acid. Calcining at 500 ◦ C leads to a slight sintering of the catalyst. Thus, the decrease of NO conversion by calcining at 500 ◦ C can be due to a combination of chemical and physical transformation. For the samples impregnated with poisons, Fig. 4(b–f) gives the NO conversion profiles. For the NaOH doped sample, the calcining temperature does not influence the SCR performance evidently. The catalyst activity drops to a low level after being calcined at 200 ◦ C and maintains stably although the calcining temperature increases. This indicates that the reaction between NaOH and the catalyst happens at temperature lower than 200 ◦ C. The influences of NaCl and MNO3 are similar to each other. These two samples show noticeable drops of NO conversion after being calcined at 400 ◦ C. This is possibly due to the chemical reaction between NaCl/NaNO3 and the catalyst when the temperature increases to 400 ◦ C. The difference between the sample doped with NaCl (NaNO3 ) and NaOH is noticeable. NaOH reacts with the catalyst at much lower temperature and thus it is more active to the catalyst surface. This is identical with the theoretical calculation results that NaOH has a much higher HOMO energy (Fig. 1) and its reaction with the vanadia site is much more thermodynamically favorable (Fig. 3). The reaction of NaCl and NaNO3 with the SCR catalyst can also generate volatile HCl and decomposable HNO3 respectively. Energy profile of HNO3 decomposition can be found in the Supplementary material. But NaCl and NaNO3 possess lower theoretical HOMO energies than NaOH and thus they are less active to the catalyst than NaOH. Consequently, a noticeable activity drop can be found for NaCl or NaNO3 doped sample when calcined at a higher temperature of 400 ◦ C. The influences of Na2 SO4 and NaHSO4 are much different from the above three compounds. Doping of Na2 SO4 does not inhibit but even slightly enhances the SCR reaction of the catalyst. Based on the reaction stoichiometry, H2 O, HCl and HNO3 will be produced for MOH, MCl and MNO3 after reacting with V-OH. While for M2 SO4 and MHSO4 , the products are MHSO4 and H2 SO4 respectively. H2 O and HCl are gaseous species under SCR reaction condition. HNO3 and H2 SO4 need to decompose into acid anhydride and H2 O to be volatile. We have searched for possible decomposition paths for HNO3 and H2 SO4 . The direct decomposing products for H2 SO4
Fig. 5. LUMO energies and orbital snapshots of the fresh VO4 H3 and doped VO4 H2 M.
can be H2 O + SO3 or H2 O2 + SO2 based on stoichiometry. Two HNO3 can react with each other to produce H2 O and N2 O5 . The reaction paths are shown in Fig. S5. The easier path (both in reaction energy and energy barrier) for H2 SO4 to decompose is to split into H2 O and SO3 . This process is endothermic by 15 Kcal/mol and needs to cross a barrier of 34 Kcal/mol. While the reaction energy and barrier for decomposition of HNO3 are +11 and +21 Kcal/mol respectively. Thus, H2 SO4 need to consume 13 Kcal/mol more energy to escape from the catalyst surface. This is in accordance with our experimental XPS results (not shown here) that S element can be detected on the doped catalyst. Therefore, the sulfate ion can hardly escape from the catalyst surface after reaction. Both cation Na+ and anion SO4 2− can deposit on the catalyst surface. As reported, sulfate can generate active Brønsted acid sites [27]. Consequently, doping of Na2 SO4 causes a dual effect on the catalyst. On one hand, the cation Na+ neutralizes the Brønsted acid. On the other hand, the cation SO4 2− can generate new Brønsted acid. This is the reason why Na2 SO4 does not lower the activity of NO reduction of the VOx -WOx /TiO2 catalyst. The promotional effect of NaHSO4 is stronger than that of Na2 SO4 . The reason is similar as that of Na2 SO4 . The anion HSO4 − will deposit on the catalyst surface under our reaction condition, which can generate active Brønsted acid sites. The acidity of HSO4 − is higher than SO4 2− and thus promotional effect of NaHSO4 is stronger than that of NaSO4 . The surface sulfation and its influence was also modeled in this study. According to Chen et al. [12], TiO2 can easily be sulfated. Thus, the sulfation of VOx -WOx /TiO2 was modeled by the sulfation of TiO2 . A dual site TiOx cluster was cut from the surface to represent TiO2 . The structures and LUMOs are shown in Table S3 of the Supplementary material. The LUMOs are mainly composed of the Ti 3d orbitals before and after sulfation. The LUMO energy will be changed from −0.074 to −0.120 due to sulfation, which indicates that the acidity of TiOx will be promoted by sulfation. 3.3. Poisoning by various alkali metals The influences of various alkali metals are varied to each other. In Fig. 5, the LUMO energies and snapshots of the fresh and doped VO4 H3 molecules are shown. LUMO energy can be an indicator to analyze the ability to attract electrons from other molecules. The lower the energy of a LUMO is, the easier it is to attract electrons. For the VO4 H3 molecule, the LUMO are mainly composed of V 3d orbital. It has a low LUMO energy to easily accept electrons and thus possesses a high acidity. Doping of Li does not influence the location of LUMO largely. Distinguished difference of the LUMO
X. Du et al. / Applied Surface Science 392 (2017) 162–168
167
Table 2 Model characters before and after being doped with alkali metals. Catalyst model
Bond length of V O(Å)
Stretching Frequency of V O (cm−1 )
Mulliken Charge of V (e)
Mulliken Charge of alkali metal (e)
LUMOEnergy (Hartree)
Undoped Li-doped Na-doped K-doped Rb-doped Cs-doped
1.565 1.645 1.642 1.635 1.634 1.632
1128 974 976 990 992 996
+1.219 +1.125 +1.093 +1.079 +1.094 +1.090
– +0.501 +0.789 +0.976 +0.957 +0.936
−0.121 −0.087 −0.071 −0.062 −0.063 −0.059
Fig. 6. The NO conversions of the VOx -WOx /TiO2 catalyst poisoned by LiCl, NaCl and KCl (loading rate of alkali metal is 20 mol% based on vanadium).
snapshot can be found for the molecules doped with Na, K, Rb or Cs element. For these molecules, the LUMO shifts to the alkali cations. Meanwhile, the LUMO energy increases noticeably after being doped with alkali metal. The increases from Li to K are noteworthy. From K to Cs, the LUMO energy maintains stably at a high level. These results indicate that alkali metals will lower the acidity of vanadia in different degrees. The influence increases from Li to K. The model characters also change after being doped with alkali metals. As shown in Table 2, doping of alkali metal results in elongation of the V O bond length and redshift of the stretching frequency of V O. These data indicate that the V O bond is weakened due to these alkali metals. Charge analysis also shows that the V atom will be reduced by the doping of alkali metal. Reduction of V atom causes the more satisfaction of its outer shell and thus the ability to attract electrons decreases. This is in accordance with the change of LUMO energies. Experiments were performed to test the poisoning strengths of various alkali metals. The performances of the catalysts poisoned by LiCl, NaCl and KCl are shown in Fig. 6. The NO conversions at the testing temperatures decline with the doping of alkali chlorate. The NO conversions follow the order of fresh sample >LiCl-doped > NaCl-doped > KCl-doped. Thus, the poisoning strength order is KCl > NaCl > LiCl, which is in accordance of our theoretical results. 4. Conclusions Various alkali metal-containing substances can cause different influences to the VOx -WOx /TiO2 SCR catalyst. HOMO energies of these substances can be used to analyze their reactivities with the acid catalyst. The initial screen by HOMO energy shows that the reactivities with SCR catalyst can be ranked
as E (MOH) > E (M2 SO4 ) > E(MCl) > E(MNO3 ) > E(MHSO4 ) for the anion aspect. For different cations, the poisoning strength follows as Cs > Rb > K > Na > Li. The increase from Li to K is more obivious than that form K to Cs. The theoretical thermodynamics study on the poisoning reactions presents the same order. Unlike the MOH, MCl and MNO3 , which can produce volatile species after reaction with the catalyst, M2 SO4 and MHSO4 will leave both cations and anions on the catalyst surface after deposite. The experimental SCR activities after poisoning are NaHSO4 -doped > Na2 SO4 -doped > undoped > NaNO3 doped ≈ NaCl-doped ≈ NaOH-doped. This confirms that the sulfate ions which can generate active acid sits are left on the catalyst after deposite. The calcining of doped catalyst at various temperatures also indicates that NaOH can react with the catalyst at temperature below 200 ◦ C, while NaCl and NaNO3 will react with the catalyst at temperature between 300 and 400 ◦ C. The after reaction-influences of various alkali metals are different. The LUMO energy of the VO4 H3 molecule increases with the doping of alkali metal. The LUMO energy follows the order of (Cs, Rb, K-doped) > Na-doped > Li-doped, which is also the order of poisoning strengths of various alkali metals. The activity test shows that NO conversion decreases as Fresh > Li-doped > Na-doped > Kdoped, which implies that the experimental poisoning strength is in accordance with the theoretical result. Acknowledgements We gratefully acknowledge the financial support of the National Natural Science Foundation of China (51506015), Fundamental & Advanced Research Projects of Chongqing (cstc2015jcyjA20006), Chongqing social undertakings and livelihood security special projects (cstc2015shmszx20004), Fundamental Research Funds for the Central Universities (106112016CDJXY145503, 106112016CDJZR145507) and general financial grant from the China postdoctoral science foundation (2015M570768). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apsusc.2016.09. 036. References [1] G. Busca, L. Lietti, G. Ramis, F. Berti, Chemical and mechanistic aspects of the selective catalytic reduction of NOx by ammonia over oxide catalysts: a review, Appl. Catal. B 18 (1998) 1–36. [2] T. Shikada, K. Fujimoto, Effect of added alkali salts on the activities of supported vanadium-oxide catalysts for nitric-oxide reduction, Chem. Lett. 12 (1983) 77–80. [3] J.P. Chen, M.A. Buzanowski, R.T. Yang, J.E. Cichanowicz, Deactivation of the vanadia catalyst in the selective catalytic reduction process, J. Air Waste Manage. 40 (1990) 403–1409. [4] Y.J. Zheng, A.D. Jensen, J.E. Johnsson, Laboratory investigation of selective catalytic reduction catalysts: deactivation by potassium compounds and catalyst regeneration, Ind. Eng. Chem. Res. 43 (2004) 941–947.
168
X. Du et al. / Applied Surface Science 392 (2017) 162–168
[5] Y.J. Zheng, A.D. Jensen, J.E. Johnsson, Deactivation of V2 O5 -WO3 -TiO2 SCR catalyst at a biomass-fired combined heat and power plant, Appl. Catal. B 60 (2005) 253–264. [6] J. Due-Hansen, S. Boghosian, A. Kustov, P. Fristrup, G. Tsilomelekis, K. Stahl, C.H. Christensen, R. Fehrmann, Vanadia-based SCR catalysts supported on tungstated and sulfated zirconia: influence of doping with potassium, J. Catal. 251 (2007) 459–473. [7] Å. Kling, C. Andersson, Å. Myringer, D. Eskilsson, S.G. Järås, Alkali deactivation of high-dust SCR catalysts used for NOx reduction exposed to flue gas from 100 MW-scale biofuel and peat fired boilers: influence of flue gas composition, Appl. Catal. B 69 (2007) 240–251. [8] O. Krocher, M. Elsener, Chemical deactivation of V2 O5 /WO3 -TiO2 SCR catalysts by additives and impurities from fuels lubrication oils, and urea solution – I. Catalytic studies, Appl. Catal. B 77 (2008) 215–227. [9] A.E. Lewandowska, M. Calatayud, E. Lozano-Diz, C. Minot, M.A. Banares, Combining theoretical description with experimental in situ studies on the effect of alkali additives on the structure and reactivity of vanadium oxide supported catalysts, Catal. Today 139 (2008) 209–213. [10] D. Nicosia, I. Czekaj, O. Krocher, Chemical deactivation of V2 O5 /WO3 -TiO2 SCR catalysts by additives and impurities from fuels lubrication oils and urea solution – Part II. Characterization study of the effect of alkali and alkaline earth metals, Appl. Catal. B 77 (2008) 228–236. [11] Y.J. Zheng, A.D. Jensen, J.E. Johnsson, J.R. Thogersen, Deactivation of V2 O5 -WO3 -TiO2 SCR catalyst at biomass fired power plants: elucidation of mechanisms by lab- and pilot-scale experiments, Appl. Catal. B 83 (2008) 186–194. [12] L. Chen, J. Li, M. Ge, The poisoning effect of alkali metals doping over nano V2 O5 –WO3 /TiO2 catalysts on selective catalytic reduction of NOx by NH3, Chem. Eng. J. 170 (2011) 531–537. [13] F. Castellino, A.D. Jensen, J.E. Johnsson, R. Fehrmann, Influence of reaction products of K-getter fuel additives on commercial vanadia-based SCR catalysts part II. Simultaneous addition of KCl, Ca(OH)2 H3 PO4 and H2 SO4 in a hot flue gas at a SCR pilot-scale setup, Appl. Catal. B 86 (2009) 206–215. [14] H. Si-Ahmed, M. Calatayud, C. Minot, E.L. Diz, A.E. Lewandowska, M.A. Banares, Combining theoretical description with experimental in situ studies on the effect of potassium on the structure and reactivity of titania-supported vanadium oxide catalyst, Catal. Today 126 (2007) 96–102. [15] F. Castellino, S.B. Rasmussen, A.D. Jensen, J.E. Johnsson, R. Fehrmann, Deactivation of vanadia-based commercial SCR catalysts by polyphosphoric acids, Appl Catal. B 83 (2008) 110–122. [16] F. Castellino, A.D. Jensen, J.E. Johnsson, R. Fehrmann, Influence of reaction products of K-getter fuel additives on commercial vanadia-based SCR catalysts part I. Potassium phosphate, Appl. Catal. B 86 (2009) 196–205.
[17] Y.J. Zheng, A.D. Jensen, J.E. Johnsson, J.R. Thogersen, Deactivation of V2 O5 -WO3 -TiO2 SCR catalyst at biomass fired power plants: elucidation of mechanisms by lab- and pilot-scale experiments, Appl. Catal. B 83 (2008) 186–194. [18] Y. He, J.J. Zhu, B. Li, Z.H. Wang, Z.S. Li, M. Alden, K.F. Cen, In-situ measurement of sodium and potassium release during oxy-fuel combustion of lignite using laser-induced breakdown spectroscopy: effects of O−2 and CO2 concentration, Energy Fuel 27 (2013) 1123–1130. [19] L. Lisi, G. Lasorella, S. Malloggi, G. Russo, Single and combined deactivating effect of alkali metals and HCl on commercial SCR catalysts, Appl. Catal. B 50 (2004) 251–258. [20] A.D. Becke, Density-functional exchange-energy approximation with correct asymptotic-behavior, Phys. Rev. A 38 (1988) 3098–3100. [21] A.D. Becke, Density-functional thermochemistry. 3. The role of exact exchange, J. Chem. Phys. 98 (1993) 5648–5652. [22] C.T. Lee, W.T. Yang, R.G. Parr, Development of the colle-salvetti correlation-energy formula into a functional of the electron-density, Phys. Rev. B 37 (1988) 785–789. [23] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, J.A. Montgomery, Jr., T. Vreven, K.N. Kudin, J.C. Burant, J.M. Millam, S.S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox, H.P. Hratchian, J.B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, P.Y. Ayala, K. Morokuma, G.A. Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich, A.D. Daniels, M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J. Cioslowski, B.B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham, C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen, M.W. Wong, C. Gonzalez, J.A. Pople, Gaussian 03, Revision B.02, Gaussian Inc., Pittsburgh PA, 2003. [24] X. Du, X. Gao, W. Hu, J. Yu, Z. Luo, K. Cen, Catalyst design based on DFT calculations: metal oxide catalysts for gas phase NO reduction, J. Phys. Chem. C 118 (2014) 13617–13622. [25] X. Du, X. Gao, K. Qiu, Z. Luo, K. Cen, The reaction of poisonous alkali oxides with vanadia SCR catalyst and the afterward influence: a DFT and experimental study, J. Phys. Chem. C 119 (2015) 1905–1912. [26] M. Anstrom, N.Y. Topsoe, J.A. Dumesic, Density functional theory studies of mechanistic aspects of the SCR reaction on vanadium oxide catalysts, J. Catal. 213 (2003) 115–125. [27] J.P. Chen, R.T. Yang, Selective catalytic reduction of NO with NH3 on SO4 2 -/TiO2 superacid catalyst, J. Catal 139 (1993) 277–288.