Regeneration of Potassium Poisoned Catalysts for the Selective Catalytic Reduction of NO with NH 3

In this study, we investigated the effect of potassium on the activity and regeneration of potassium-poisoned SCR catalysts. With the addition of potassium species, the NO conversion rate of the catalysts continuously decreased. After washing the poisoned catalysts with an H2SO4 solution or doping them with CeO2, the activity of the catalysts was improved to different extents. Acid washing almost completely removed the surface potassium species, freeing acidic sites to adsorb NH3, but it also potentially removed some of the active components, such as vanadia. CeO2 doping, on the other hand, added active components. Combining these two methods, the poisoned catalysts were washed with an H2SO4 solution and then doped with 5 wt.% CeO2. It was found that the level of activity could be restored to that of a fresh catalyst, and a conversion rate of over 90% was observed for NO between 300°C and 450°C, as the added CeO2 compensated for the active components lost during SCR reactions. Consequently, the above hybrid method shows high potential for regenerating commercial SCR catalysts.


INTRODUCTION
The selective catalytic reduction (SCR) of NO x with NH 3 has been regarded as an effective method to control NO x emissions from stationary and mobile sources (Busca et al., 1998).The commercial SCR catalysts consist of TiO 2 as support and V 2 O 5 -WO 3 or V 2 O 5 -MoO 3 as active components.They are shaped into honeycomb matrix because of the advantages such as low pressure drop, high geometric surface area, and resistance to deposition of dust (Lei et al., 2009).
The (co-)firing of biomass is a significant way to reduce the net CO 2 emissions.However, high levels of alkali and alkaline earth metals, especially potassium, are present in the fly ash of the biomass fired systems.And potassium has been demonstrated to do harm to the SCR catalysts (Kamata et al., 1999;Moradi et al., 2003;Zheng et al., 2004Zheng et al., , 2005;;Due-Hansen et al., 2007;Castellino et al., 2009;Klimczak et al., 2010).Doping with alkali and alkaline earth metals results in a strong catalyst deactivation and the poisoning effect of alkali and alkaline earth metals is related to their basicity.The order is listed as K > Na > Ca > Mg (Klimczak et al., 2010).It is concluded that potassium preferentially coordinates to the Brønsted acid sites, which are responsible for the ammonia adsorption, thus decreasing their number and strength of the Brønsted acid sites.As a result, catalyst activity would be adversely affected.A deactivation of about 1% per day was found over monolith catalysts, which were exposed in a high-dust flue gas produced from straw-fired grate boiler (Zheng et al., 2005).
The cost of the catalysts is a major part of the total expense in an SCR system.Hence, it is important and necessary to regenerate the catalysts (Du et al., 2018;Wang et al., 2018a, b).Regeneration by washing with water followed by sulfation was not an optimal regeneration method due to the insufficient removal of the poison (Zheng et al., 2004).Washing with sulfuric acid can remove potassium accumulated on the surface and recover the catalytic activity (Khodayari and Odenbrand, 2001a, b;Zheng et al., 2004).However, active components such as vanadium and tungsten can also be removed through washing with sulfuric acid.So, it is necessary to compensate the loss of the active components.Recently, ceria-based catalysts have been investigated for SCR reactions because of the high oxygen storage capacity and excellent redox properties of CeO 2 (Li et al., 2012;Hu et al., 2017;Huang et al., 2017;Jiang et al., 2017;Li et al., 2017;Yao et al., 2017;Chen et al., 2018;Jiang et al., 2018).We could also add ceria on the catalyst to compensate the loss of the active components.
In this work, potassium ions were doped on the commercial V 2 O 5 -WO 3 /TiO 2 catalysts to simulate the poison effect.Then, a study of the regeneration of K-poisoned catalysts was presented.Specifically, loading different amount of ceria on the deactivated catalysts with or without sulfuric acid and deionized water washing have been investigated to show the commercial potential of ceriainvolved regeneration method.

Catalyst Preparation
The commercial honeycomb-type catalysts used in this investigation were obtained from RAGA Technology Co., Ltd.The catalysts contained about 1% V 2 O 5 and 5% WO 3 as active phase doped on a TiO 2 support.The catalyst had a wall thickness of 1 mm and a channel pitch of 6 mm.The catalysts were cut into 25 mm × 20 mm × 20 mm blocks for poisoning of potassium and regeneration method tests.

Deactivation of Catalysts
The catalysts doped with potassium were prepared by wet-impregnation method with aqueous solutions of KNO 3 .The prepared monolithic catalysts were immersed into 25 mL aqueous solutions with different concentration of KNO 3 .Then the samples were first dried at 110°C overnight and then calcined at 500°C for 5 h in a muffle furnace to form potassium oxide.

Regeneration of Deactivated Catalysts
Several blocks of deactivated catalysts (25 mm long monolithic segment) were washed in 1000 mL 0.5 M sulfuric acid (SA) solution for 120 min and then washed in 1000 mL deionized water for 15 min.In some cases, different concentrations of sulfuric acid (0.05 M, 0.1 M, 0.3 M, 0.5 M, 0.7 M) were used to optimize the washing process.The washing process was under continuous stirring.The temperature of the washing solution was kept at 50°C.After washing, different contents of CeO 2 were loaded on the catalysts using the same method as potassium doping.Another regeneration method was that CeO 2 was added on the deactivated catalysts without washing process.The catalysts were donated as xK-ySA-zCeO 2 .x represents the loading of potassium element (wt.%), y represents the concentration of SA (sulfuric acid) used for catalysts washing (M) and z represents the loading of CeO 2 (wt.%) on the catalysts.

Activity Tests
The activity tests for the reduction of NO by NH 3 were carried out in a fixed bed quartz micro-reactor (inner diameter: 4 mm) with 0.2 g catalyst power of 250-380 µm in diameter.The feed gas mixture contained 1000 ppm NO, 1000 ppm NH 3 , 5 vol.%O 2 and N 2 as the balance gas.The total flow rate of the feed gas was 1000 mL min -1 and the GHSV was 183,000 h -1 .The catalytic reaction was carried out with temperature ranging from 150°C to 450°C.The concentrations of NO and N 2 O before and after reaction were measured by an FTIR gas analyzer, the Gasmet DX4000.The NO conversion is defined as where NO in and NO out stand for the NO concentration at the inlet and outlet, respectively.

Catalyst Characterization
The specific surface area and the textural properties (i.e., pore volume and average pore diameter) were measured by N 2 adsorption and desorption experiments at liquid nitrogen temperature (-196°C) with Autosorb-1-C instrument (Quantachrome Instrument Corp.).The specific surface area was calculated by the Brunauer-Emmett-Teller (BET) method while the average pore diameter was calculated from the surface area and BET pore volume.
The X-ray diffraction (XRD) patterns were collected using a Panalytical X'pert Pro diffractometer equipped with Cu Ka radiation.The X-ray tube was operated at 40 kV and 40 mA.
The chemical composition of catalyst samples was analyzed by inductively coupled plasma-mass spectroscopy (ICP-MS; 7500a, Agilent) after the catalysts dissolved completely.
NH 3 -TPD tests were carried out on an AutoChem 2920 instrument provided by Micromeritics Corporation.NH 3 signal was detected using a Hiden QIC20 mass spectrum instrument.0.2 g catalyst powder was pretreated in He at 500°C for 30 min.After that, the sample was cooled to 100°C and exposed to a gas mixture of 5% NH 3 in He for 30 min.Then the sample was flushed with pure He until signal was stabilized.The sample was then heated up to 700°C at a rate of 10°C min -1 .
H 2 -TPR was also carried out in a quartz-tube reactor.0.10 g catalyst powder was pre-treated at 200°C in N 2 for 30 min.After that, the sample was cooled to room temperature and then heated to 700°C at a rate of 10°C min -1 in a gas mixture of 5 vol.%H 2 and Ar.The consumption of H 2 was detected by a thermal conductivity detector (TCD).

Activity and Selectivity Effect of Potassium Poisoning
Fig. 1 shows the activity of the fresh and potassiumloaded catalysts.For the fresh catalyst, the NO conversion increased as the temperature increased, which almost reached 100% at 350°C.Given that potassium was introduced, the NO conversion decreased.With the increase in the doping amount of potassium, catalyst activity continuously decreased.When the potassium loading amount reached 1%, the NO conversion of the catalyst was always below 30%.This is consistent with the result of Chen et al. (2011).

Effect of Acid Washing
To regenerate the poisoned catalyst, the first step is to remove the accumulated K. Sulfuric acid was used as detergent combined with fresh water rinsing after the process.Different concentrations of sulfuric acid solutions were chosen and effect on NO conversion is depicted in Fig. 2. It is clear that with concentrated acid, poisoned catalyst is easily recovered, approaching its initial state.However, this effect was gradually weakened when the concentration increased from 0.05 M to 0.5 M. Specially, no obvious difference was observed for the catalysts treated with 0.5 M and 0.7 M H 2 SO 4 .As a result, 0.5 M H 2 SO 4 was used in later section for acid washing.Moreover, the washing process also contributed to loss of active components, such as V and W. The negative effect of acid washing is shown in Fig. S1 in supporting information.The effect of washing time was also evaluated.It is evident that using diluted acid and short time could remove most of K and reduce the loss of V.However, as shown in Fig. 2, the diluted acid washing couldn't totally recover the activity of the catalyst.On the other hand, after 2 h washing with 0.5 M H 2 SO 4 , 40% V of the catalyst was dissolved into the solution, resulting in the reduction of the activity, especially in lower temperatures (< 350°C).For this reason, it is necessary to load active components to compensate the loss.

Effect of CeO 2 Loading
The NO conversion of the K-poisoned and regenerated catalysts is presented in Fig. 3.The direct loading of 10% CeO 2 to the deactivated catalyst could improve the activity, which was, however, much lower than that of the fresh catalyst.This is because potassium still remained on the catalyst and exerted a negative effect on the catalyst.Thus, removing potassium species constituted the first step to recover the activity of the deactivated catalyst.After washing with H 2 SO 4 solution, an obvious enhancement in the high-temperature activity could be obtained, while the activity below 300°C remained almost unchanged.It should be noted that the activity of the regenerated catalyst washed with H 2 SO 4 solution was still lower than that of the fresh one.That might be due to the loss of active components during the H 2 SO 4 solution washing process.Therefore, various amounts of CeO 2 were added to the deactivated catalysts after treatment by 0.5 M H 2 SO 4 and deionized water.The NO conversion of 0.5K-0.5SA-3CeO 2 catalyst is almost the same as that of the 0.5K-10CeO 2 catalysts below 300°C.With temperature increasing, 0.5K-0.5SA-3CeO 2 catalyst exhibited a better activity than that of 0.5K-10CeO 2 catalyst.When the loading amount of CeO 2 exceeded 5 wt.%, the activity of the regenerated catalysts almost restored to the level of the fresh one.Thus, it seems that washing with 0.5 M sulfuric acid solution and deionized water and then doping with 5 wt.%CeO 2 is a good method to regenerate the K-poisoned catalysts.
In order to investigate the adaptability of this regeneration method, catalyst with a higher potassium loading amount were chosen and tested.As illustrated in Fig. 4, it can be concluded that the amount of potassium had no obvious effect on the activity of the regenerated catalysts in the investigated range and the activity of the regenerated catalysts is almost the same as that of the fresh one.This indicated that 0.5 M H 2 SO 4 and deionized water treatment could remove potassium species that interacted with the active sites, and 5 wt.%CeO 2 additives could supply extra active sites to some extent.

Selectivity of Regenerated Catalysts
In addition to NO conversion, N 2 O formation is also an important parameter to evaluate catalyst performance.Fig. 5 displays N 2 O formation in the SCR reactions.For the fresh and regenerated catalysts, the N 2 O concentration were fairly low (< 5 ppm) below 350°C.As the temperature increased, N 2 O concentration of the fresh catalyst rapidly increased up to 47 ppm at 450°C, while the N 2 O formation of the regenerated catalysts stayed below 17 ppm, suggesting a better selectivity than the fresh one.This phenomenon could be explained by the reduction of V and the addition of Ce, confirmed by Fig. S1 and Table 1.Since V 2 O 5 has a strong ability to oxidize NH 3 into N 2 O at higher temperatures (Chen et al., 2009), decrease of V 2 O 5 could weaken this side reaction and improve the selectivity of SCR reaction.

Characterization of the Catalysts Chemical Composition and BET Analysis
Table 1 shows the ICP-MS and BET results of the fresh and regenerated catalysts.It seems that no obvious variations could be observed in these catalysts, partly ruling out the possibility that physical properties mainly determined the activity of the catalysts.Instead, the variations in the chemical properties including acidity and redox constituted the main reason for enhanced activity of the catalysts after regeneration.
The ICP-MS results showed that the washing process could drastically remove potassium species over catalyst surface.At the same time, vanadia, as the active component, was also removed by a third with the content of tungsten remaining almost unchanged after the washing process.Additionally, the measured Ce content was lower than the calculated value.This meant that there was loss of Ce during impregnation process.

XRD Analysis
XRD patterns of the series catalysts are shown in Fig. 6.All the catalysts presented anatase TiO 2 phase with the absence of V 2 O 5 and WO 3 crystallites, indicating that V 2 O 5 and WO 3 were amorphous in structure on the TiO 2 support (Lisi et al., 2004;Kustov et al., 2005;Zhang et al., 2009).Given the addition of 3% CeO 2 , no cubic CeO 2 phase could be detect.This result indicated that ceria were highly dispersed and existed as an amorphous state.Further increasing the CeO 2 loading amount to 5% or more, cubic CeO 2 phase began to be observed, suggesting that CeO 2 loading was beyond the theoretical monolayer coverage on the TiO 2 support.

NH 3 -TPD
According to previous studies, acidity plays an important role in the SCR reactions, since the first step is the adsorption of NH 3 on the surface acidic sites of catalyst (Busca et al., 1998;Forzatti, 2001;Ye et al., 2018).NH 3 -TPD tests were carried out to investigate surface acidity of the series samples, and the results of which are illustrated in Fig. 7.The 0.5K and 0.5K-10CeO 2 catalysts had several NH 3 desorption peaks in the temperature region of 100-350°C, while the other catalysts possessed broad desorption peaks between 100°C and 450°C.The peaks near 170°C (Peak I) could be assigned to the desorption of physisorbed NH 3 , and the peaks around 270°C (Peak II) were attributed to NH 3 on weakly acidic sites, while the peaks centering at 610°C (Peak III) linked to the strongly acidic sites (Guan et al., 2011;Li et al., 2012;Li et al., 2017;Yao et al., 2017).The quantity analysis of NH 3 -TPD is summarized in Table 2.For the 0.5K and 0.5K-10CeO 2 catalysts, the presence of potassium caused the reduction of Peak II, which correlated with acidic sites, responsible for SCR activity (Forzatti, 2001).After washing process, abundant acidic sites were present, and NH 3 could be adsorbed on  (e) 0.5K-0.5SA-3CeO 2 ; (f) 0.5K-0.5SA-5CeO 2 ; (g) 0.5K-0.5SA-10CeO 2 .
the catalyst surface, reacting with NO.Besides, doping CeO 2 could also in part enhance the acidity of the catalysts, which made 0.5K-0.5SA-5CeO 2 catalyst had the same activity with the fresh one (Shi et al., 2017).That is consistent with SCR activity results in Fig. 3.Note that after washing and CeO 2 modification, Peak I increased almost up to two times while Peak II reached about 80% of that of the fresh catalyst.It may be related to promoted SCR activity at low temperatures and inhibited NH 3 oxidation at high temperatures.
H 2 -TPR The redox properties of the catalysts play an important role in the catalytic cycle of the SCR reactions (Topsoe, 1994;Topsoe et al., 1995;Putluru et al., 2009;Wang et al., 2018).The H 2 -TPR profiles and quantity analysis are shown in Fig. 8 and Table 2, respectively.The fresh catalyst showed a reduction peak located at 500°C, while the 0.5K sample presented a peak at 523°C.These peaks could be explained by the reduction of V 5+ to V 3+ (Tang et al., 2010;Chen et al., 2011;Guan et al., 2011).And it should be noted that the reduction peak shifted to higher temperatures after doping potassium, which is consistent with the results of Chen et al. (2011), showing that potassium doping exerted a negative effect on the catalyst redox properties.After H 2 SO 4 washing process, a reduction peak at 485°C came out.This meant that washing with sulfuric acid solution and deionized water could remove potassium species and recover the redox ability of the catalysts.The reduction peaks shifted to higher temperature with the increasing loading amount of CeO 2 .Moreover, the adding of CeO 2 slightly increased the consumption of H 2 , compared to 0.5K-0.5SA.This indicated that loading more Ce had a negative effect on the catalyst reducibility.Besides, XPS results shown in Fig. S2 confirmed that no obvious change in chemical state of V, W and Ti after acid washing or CeO 2 modification.Combined with aforementioned results, it implied that improved SCR activity after acid washing and CeO 2 adding could be ascribed to optimized acidic sites, which facilitate the low temperature SCR activity and inhibit high temperature NH 3 oxidation.
1. Potassium doping had a negative effect on the activity of the catalysts.The increase in the potassium loading amount continuously decreased the NO conversion rate of the catalysts.After washing the catalysts with an H 2 SO 4 solution and deionized water, the SCR activity was partially restored.The further addition of CeO 2 caused the regenerated catalysts to exhibit almost as much activity as the fresh ones.2. After washing the catalysts with an H 2 SO 4 solution and deionized water, the potassium species were almost completely removed, benefical to NH 3 adsorption.CeO 2 was then added to compensate for the lost active species.These factors were primarily responsible for the recovery of the catalyst activity after regeneration.

Fig. 1 .
Fig. 1.NO conversion of the catalysts poisoned with different K loading amounts.

Fig. 2 .Fig. 3 .
Fig. 2. NO conversion of the catalysts regenerated with different concentrations of H 2 SO 4 solution.

Fig. 4 .
Fig. 4. Effect of K content on the catalysts regenerated by washing and CeO 2 modification.

Fig. 5 .
Fig. 5. N 2 O formation of the catalysts regenerated by washing and CeO 2 modification.

Table 1 .
Chemical composition, BET surface area, total pore volume, and average pore diameter of different samples.

Table 2 .
Quantitative analysis of NH 3 -TPD and H 2 -TPR over the samples * .Normalized by the values of the fresh catalyst. *