The industrial feasibility of low temperature DeNOx in the presence of SOx: a project case in a medium coking plant

Catalyst poisoning by SOx has hindered the industrial application of selective catalytic reduction (SCR) technology for the DeNOx of low temperature flue gas for many decades. The current engineering process of placing the SCR unit after the desulphurization and dedusting units can lead to high project and operational costs owing to low DeNOx efficiency at low temperatures. Based on our previous pilot results, a DeNOx project case was built before the desulfurization unit in a medium coking plant to explore the industrial feasibility of low temperature DeNOx in the presence of SOx. A new engineering process was considered and designed to overcome the problem of SOx, including the elimination of SO3 with NH3 before the SCR reactor, the filtration of ash by a foam metal plate and in situ regeneration technology. The project case could run continuously for six months with above 70% DeNOx efficiency and less than 10 ppm NH3 slip at 250 °C and a space velocity of 4000 h−1 in the presence of 260–300 mg m−3 SOx. The activity loss for the catalyst itself was not obvious after it had been running for six months, but blocking of the honeycomb channels by the sedimentary ash on the honeycomb catalyst modules owing to the low linear velocity resulted in decreased DeNOx efficiency and an increased pressure drop. A improved DeNOx process with gravitational dust collectors was also proposed to upgrade the present DeNOx project case for further continuous and stable operation.


Introduction
The DeNO x of low temperature ue gas from medium/small boilers is an urgent task in China because of increasingly strict NO x emission regulations in recent years. [1][2][3][4] Unlike the well-developed engineering application of selective catalytic reduction (SCR) technology for high temperature ue gas from power plants, great obstacles still restrict its wide industrial application in medium/small boilers because of the unsatisfactory catalytic activity and stability resulting from the unavoidable formation of ammonium sulfate species in the presence of SO x at low temperatures. [5][6][7][8] The ue gas has to undergo desulphurization and dedusting before the SCR units to avoid sulfur poisoning of the SCR catalyst, but the temperature of the ue gas may fall from more than 250 C to less than 200 C. 9,10 The low temperature ue gas has to be reheated, as well as more catalyst used, with low space velocity for sufficient DeNO x efficiency, which signicantly increases the project and operation costs. As far as we know, however, all of the established SCR units are placed aer the desulphurization and dedusting units for the purication of low temperature ue gas from medium/small boilers in China.
Our previous results from a SCR pilot test in a coking plant revealed that the deactivation of the honeycomb catalyst was mainly caused by severe coverage of the catalyst surface with ferric sulfate and ammonium sulfate species in the presence of SO x below 270 C. 11 The ferric sulfate was from the corrosion of iron pipe by SO 3 and ammonium sulfate was the product of a reaction between the injected NH 3 and SO 3 (but not SO 2 ). But aer SO 3 and the ash with these sulfate species were eliminated using a dry desulfurization-dedusting process, the catalyst could even work with stable activity at 220 C in the presence of high-concentration SO 2 . This previous work signies that low temperature DeNO x in the presence of SO x may be industrially feasible if the problems of SO 3 removal and pipeline corrosion are resolved by new engineering technology.
In this work, a DeNO x project case was built before the desulfurization unit in a medium coking plant. A long ue of more than 50 meters was reserved between the locations of the injecting NH 3 and SCR reactors, in which SO 3 could be eliminated by the injected NH 3 to form ammonium sulfate species through the gas-phase reaction between SO 3 and NH 3 with enough residence time. All the ues underwent antiseptic treatment using heat resistant paint. A replaceable foam metal plate was placed above the honeycomb catalyst in the SCR reactors to trap the ash and homogenize the ue gas. Moreover, an in situ regeneration process was also considered and designed to regenerate the deactivated catalyst with hot gas. In this way, this DeNO x project case is expected to overcome the adverse effects of sulfate covering and has attractive technological advantages such as higher DeNO x efficiency, lower amounts of catalyst used, and lower operation costs.

Technological process of the project case
In the project case, the SCR unit was placed before the desulphurization and dedusting units; the detailed technological process of the SCR unit is shown in Fig. 1. The SCR unit mainly included an aqua ammonia system, a heating system, a compressed air system and SCR reactors. 20 wt% aqua ammonia was used to provide NH 3 for the SCR reaction. The aqua ammonia must be sufficiently gasied at a temperature above 200 C in an ammonia evaporator before it can be injected into the main pipe, because unevaporated water droplets have been conrmed to facilitate the rapid formation of ammonium sulfates, in a similar way to the ammonia desulfurization process taking place through the main reactions in eqn (1) and (2). 12 Moreover, the pipe length between the entrance for injecting NH 3 and the SCR reactors was more than 50 meters, which provided enough residence time for the gasphase reaction between NH 3 and a small amount of SO 3 through the reaction in eqn (3). This design could eliminate SO 3 using NH 3 before the ue gas was introduced into the SCR reactors and avoid the direct formation of ammonium sulfate species on the catalytic surface between the adsorbed ammonia and gaseous SO 3 . The produced ammonium sulfate dust would be trapped by the replaceable foam metal plate placed above the honeycomb catalyst in SCR reactors.
The high-temperature gas (>450 C) from the hot-blast stove was used to heat and vaporise aqua ammonia, or in situ regenerate the honeycomb catalyst in the SCR reactors through the decomposition of the ammonium sulfate species above 450 C. 13,14 Compressed air was supplied for the ammonia evaporator to dilute the vaporized NH 3 and for the soot blower to clean the ash deposited on the catalytic surface. Four SCR reactors were designed for use alternately on a line. Two of the SCR reactors were kept spare under normal operating conditions, and the spare reactors would start being used when the catalyst in the used reactors needed to be regenerated by being heated. Thus, the SCR unit can operate continuously even if the stability of the catalyst is retained for only several months. In addition, a replaceable foam metal plate was placed above the honeycomb catalyst in each SCR reactor, which was used to trap the ammonium sulfate species and other dust to avoid coverage and blocking of the honeycomb catalyst. In this way, the design of the project case is expected to overcome the disadvantage of catalyst deactivation resulting from sulfur poisoning at low temperatures.
In the project case, a V based honeycomb catalyst was used, and the catalyst composition, as well as information on the specic surface and pore structure, is provided in Table 1. The V based catalyst was chosen by considering that the used catalyst poisoned by ammonium sulfate species can easily undergo thermal regeneration below 450 C, which could be realized in the in situ regeneration process in the project case.
The parameters of the ue gas and designed cra in the DeNO x project case are listed in Table 2. The concentration of NO x in the original ue gas was in the range 400-500 mg m À3 , and it needed to decrease to below 150 mg m À3 through the SCR reactors. The operation temperature of the DeNO x reactor was in the range 240-270 C, normally about 250 C. The ue gas Fig. 1 The technological process of the DeNO x project case together with the details of a single SCR reactor: (1) ammonia tank, (2) motorized valve, (3) ammonia evaporator, (4) draught fan, (5) hot-blast stove, (6) air reservoir, (7) SCR reactor, (8) hand valve, (9) air compressor, (10) soot blower, (11) foam metal plate and (12) honeycomb catalyst. also contained 200-300 mg m À3 SO 2 and a small amount of SO 3 (below 5 ppm). In accordance with previous results from our pilot test, 11 the gaseous hourly space velocity (GHSV) was designed to be 4000 h À1 , and the desired DeNO x efficiency would have been more than 70% with less than 150 mg m À3 NO x in the exit of the SCR reactors. The ow of ammonia was tuned according to the NH 3 slip in the exit of the SCR reactors, which was controlled below 10 ppm.
3. Operating data for the DeNO x project case Fig. 2 shows the operating data for the DeNO x project case in the past ten months. The conditions of the ue gas were relatively stable during the ten months. The average NO x concentration was about 450 mg m À3 , the average SO 2 concentration was from 260-300 mg m À3 , and the average temperature was from 240-260 C. In the rst three months, the NO x concentration in the exit of the SCR reactors was below 50 mg m À3 with about 90% DeNO x efficiency. Meanwhile, the NH 3 slip and pressure drop were also stable at low levels. However, values for these data (the outlet NO x concentration, NH 3 slip and pressure drop) began to increase from the fourth month. The NO x concentration for the sixth month was close to 150 mg m À3 together with a NH 3 slip of more than 10 ppm, and even the pressure drop increased rapidly. The high pressure drop with an decreased DeNO x efficiency implied that the honeycomb catalysts in SCR the reactors were probably blocked up. On the other hand, an attempt to clean the catalytic bed using the soot blower could not decrease the pressure drop. Thus, the spare reactors were used temporarily, and the used reactors were opened aer having run for six months. Indeed, it was found that both the foam metal plates and honeycomb catalysts were severely blocked up by ash (the details are introduced in the next part). Aer replacing the blocked foam metal plate and dredging the blocked channels of the honeycomb catalysts, function of the used reactors was recovered, as shown, from the seventh to the tenth month. In the seventh month, the DeNO x efficiency recovered close to the initial value, and the values for NH 3 slip and pressure drop were a little higher than their initial values. However, the outlet NO x concentration, NH 3 slip and pressure drop increased gradually in the next three months (until now), which might also have resulted from the blocking of the SCR reactors.

Characterization of the unloaded catalyst sample
To nd out the reason for the decreased DeNO x efficiency and increased pressure drop aer the DeNO x project case had been operating for six months, the used SCR reactors were opened and samples of both the foam metal plate and the honeycomb catalyst were also taken for further study in the lab. Fig. 3 shows the digital photographs of the used foam metal plate and catalyst. It can be seen that the foam metal plate was severely blocked by ash (Fig. 3A). Even worse, a thick layer of dust had been deposited and covered the upper side of the honeycomb catalyst modules (Fig. 3B). For a single honeycomb, it can be seen that about a quarter of the channels were blocked up (Fig. 3C). The coverage and blocking of the honeycomb channels by deposited ash would observably obstruct the normal ow of ue gas through the honeycomb catalysts, resulting in the increased pressure drop and decreased DeNO x efficiency. However, the inner channels of the honeycomb catalyst were not blocked at all. Only some gray ash had been deposited on partial channels of the honeycomb. By close examination, interestingly, the ash was found to be deposited only on the channels with defects (derived from their manufacture, gray channel part "a" in Fig. 3D) or the channels with their upper side blocked up (gray channel part "b" in Fig. 3D). In contrast, nearly no ash was deposited on the channels without any blocking (yellow channel part "c" in Fig. 3D). These results  Fig. 2 The results of the DeNO x project case for ten months: (A) the average inlet/outlet NO x concentration for each month, (B) the average NH 3 slip for each month, (C) the average pressure drop through the SCR reactor for each month, and (D) the average outlet SO 2 concentration and temperature for each month. The reactors were opened after having run for six months, and were reused after replacing the blocked foam metal plate and dredging the blocked channels of the honeycomb catalysts.
indicated that the linear velocity of the ue gas through the honeycomb channels had a great impact on the deposition of ash on the inner surfaces of the honeycomb channels. Defective or blocked channels with low linear velocities facilitated the deposition of ash. More detailed characterization was carried out as shown in Fig. 4 to evaluate the inuence of the deposited ash on the DeNO x efficiency of catalyst itself. It can be seen from the gray channel that the catalyst particles were partially covered (Fig. 4A), but no covering was found on the surface of the yellow channel (Fig. 4B). Moreover, the thermogravimetric (TG) analysis of gray ash shown in Fig. 4C indicated that the ash contained about a 50 wt% carbon component (the weightlessness step from 480-580 C in air) and about 40 wt% ammonium sulfate species (the weightlessness step from 250-480 C in air or N 2 ). The carbon component was from the y ash from the coke oven, and the ammonium sulfate species were from the gas-phase reaction between SO 3 and NH 3 according to the previous results of our pilot test. 11 No obvious weightlessness steps in air from 500-650 C for iron sulfate species were observed, indicating that few iron sulfate species were produced aer the pipeline underwent antiseptic treatment. 15 Moreover, compared with the fresh catalyst, the yellow channel part of the used catalyst even had better catalytic DeNO x activity owing to the sulfurization of V based catalysts. 16,17 However, the activity only decreased slightly for the gray channel part of the used catalyst, which resulted from partial coverage of the catalytic surface by ash. Combined with the operating data in Fig. 2, it can be deduced that the main reason for the decreased DeNO x efficiency and increased pressure drop was the blocking of the honeycomb channels and the foam metal plate but not obvious deactivation of the catalyst itself.

Problems and solutions
According to the above results, the ash on the catalyst was mainly composed of carbon and ammonium sulfate species. The carbon component was from the y ash in the original ue gas. The ammonium sulfate species were the gas-phase reaction products of the injected NH 3 and SO 3 , which were formed and then bonded with y ash in the ue before the SCR reactors. Aer the ue gas entered the SCR reactors, the linear velocity of the ue gas decreased suddenly owing to the markedly larger inner diameter of the SCR reactors than the pipe diameter of the ue. The ash was then sedimentary and partly trapped by the foam metal plate. However, a lot of ash still dropped down and covered the upper side of the honeycomb catalyst modules, which blocked up the honeycomb channels. On the other hand, the inner channels of the honeycomb catalysts were relatively clear, and only a small amount of ash was mainly deposited on the defective or blocked channels. The activity of the catalyst itself could still be maintained even aer having run for six months. Therefore, the decreased DeNO x efficiency and increased pressure drop of the DeNO x project case were caused by the blocking up of the honeycomb channels with deposited ash.
To solve these problems, an improved technological process was proposed, as shown in Fig. 5. In this process, gravitational dust collectors were added before where the ue gas was introduced into the SCR reactors, which were used to remove the ash of both y ash and the ammonium sulfate species.  Moreover, the ue gas owed through the SCR reactor from the bottom to the top with an ash bucket in the entrance. The foam metal plate was placed below the honeycomb catalyst in the SCR reactors to trap the remaining ash and homogenize the ue gas. In this way, the deposition of ash on the catalyst would be alleviated to avoid the blocking up of the channels of the honeycomb catalyst. According to the operating data in Fig. 2, the lifetime of the catalyst would be more than one year if the problem of ash blocking was solved by the new process. On the other hand, the slow deactivation of the catalyst, which was caused by the covering on the catalytic surface as shown in Fig. 4, can also not be neglected. In addition to the carbon component, the covering also contained ammonium sulfate species. Therefore, the in situ regeneration cra was also necessary for the continuous and stable operation of the DeNO x project case. The high-temperature gas from the hot-blast stove would burn out the covering to clear the catalyst surface and recover its intrinsic activity, in case the DeNO x efficiency fell below the designed value. It is well known that a DeNO x catalyst may be poisoned by ammonium sulfate species in the presence of SO x at low temperatures. 18 Although the poisoning by SO x may be unavoidable in terms of the chemistry and the catalyst itself, the problem can be solved by advanced process engineering such as the elimination of SO 3 using NH 3 , gravitational dust collection and in situ regeneration technology. The results of the project case in this work demonstrated that the engineering application of low temperature DeNO x in the presence of SO x is feasible in industry.
Moreover, the project case also has great cost advantages. In contrast, our other SCR unit was placed aer the dry desulfurization and dedusting units for the purication of ue gas from steelworks, in which the ue gas temperature in the SCR reactors fell down from 260 C to 200 C. A much lower space velocity of 3000 h À1 as well as much higher content of active V component (5 wt%) had to be adopted for enough catalytic DeNO x efficiency at 200 C. The catalyst cost of the post-DeNO x process per cubic ue gas was increased by about 40% compared with that of the pre-DeNO x process reported in this work.

Conclusions
In summary, a DeNO x project case was built and successfully run for the purication of low temperature ue gas in the presence of SO x from a medium coking plant. New process engineering was considered and designed to overcome the problem of SO x , including the elimination of SO 3 using NH 3 before the SCR reactor, the ltration of ash using a foam metal plate and in situ regeneration technology. The project case could run continuously for six months with above 70% DeNO x efficiency and less than 10 ppm for NH 3 slip at 250 C and a space velocity of 4000 h À1 in the presence of 260-300 mg m À3 SO x . The main problem of the project case was the blocking of the honeycomb channels by the sedimentary ash on the honeycomb catalyst modules, which resulted in decreased DeNO x efficiency and an increased pressure drop. The ash was mainly composed of 50 wt% carbon component from the inherent y ash and 40 wt% ammonium sulfate species from the gas-phase reaction between SO 3 and NH 3 in the ue aer the injection of NH 3 . The sudden decrease in linear velocity from the pipeline to the SCR reactor resulted in severe deposition of ash on the upper side of the honeycomb catalyst modules, and the ash could also be deposited on the defective or blocked channel surfaces with a low linear velocity. Compared with the adverse effect of the ash, the effect of SO 2 on the catalyst was not prominent for the low temperature SCR process, and the activity loss of the catalyst itself was not obvious even aer having run for six months. A possible improved DeNO x process was proposed with gravitational dust collectors to solve the ash blocking problems of the present project case, and the proposed process with the assistance of in situ regeneration technology is expected to realize continuous and stable operation of the industrial DeNO x project for low temperature ue gas with high-concentration SO x .

Conflicts of interest
There are no conicts to declare.