Full-Scale Investigation of Dry Sorbent Injection for NO_x Emission Control and Mercury Retention

Abstract

An innovative dry SNCR method realized by a sorbent injection applied to a stoker furnace is presented. The process is based on urea powder admixed with halloysite, an aluminosilicate clay mineral. Field tests were performed at an industrial stoker hot water boiler of 30 MW_th capacity. A unique nozzle design for injecting powdery sorbents into the combustion zone was implemented. The base NO_x emission without SNCR was determined to be 365 mg/Nm³. During the reference test, the emission was reduced to avg. 175 mg/Nm³, which produces a NO_x reduction of 52%. NH₃ slip in the flue gas was stable and did not exceed 2 ppm. Combining urea and halloysite powders leads to a number of positive effects; not only is NO_x emission reduced to values typical for wet SNCR, but also a significant, over ten-fold increase in the concentration of adsorbed mercury in fly ash was observed. When confronted with wet SNCR, dry SNCR has no adverse effect on boiler efficiency because it does not increase the stack heat loss. The presented method can be used in any small- or medium-scale furnace, including waste-to-energy units or medical and hazardous waste incineration units.

1. Introduction

The reduction of nitrogen oxides emissions during combustion of solid fuels requires new, yet economically viable technological solutions. While large combustion plants successfully confronted the challenge of reducing pollutants, the small- and medium-sized plants (<50 MW_th) such as stoker boilers, when they became the subject of more stringent emission regulations, were particularly challenging in meeting satisfactory NO_x reduction efficiency.

Stoker boilers are commonly used in heating and waste-to-energy plants along all over Europe. They can consume low-quality fuels such as RDF (Refuse Derived Fuel) or waste, do not require advanced maintenance and are easy to operate. In the past, there was not enough focus on developing flue gas cleaning methods for stokers and frequently they have not been equipped with efficient systems to reduce nitrogen oxides emissions. Similarly, the emission of mercury from such small furnaces was not a subject of interest. Hence, the application of NO_x and mercury emission control methods dedicated to stoker furnaces is a relatively new issue to solve and requires research and development.

The combustion of alternative fuels, such as waste, RDF and SRF (Solid Recovered Fuel), due to unfavorable feedstock characteristics may be a source of elevated pollutant emissions. Wasilewski et al. investigated the NO_x and mercury emission from a stoker during the combustion of coal and co-combustion of SRF [1]. For coal combustion, the NO_x emission exceeded 400 mg/Nm³ and the mercury emission was 0.00014 400 mg/N m³_. During the co-combustion of 10% SRF the emission of mercury rises to 0.00080 mg/Nm³ while NO_x dropped to 356 mg/Nm³_.

Sefidari et al. [2] determined the concentrations of O₂, NO and CO at different sections of the grate furnace and found a higher NO emission tendency for greenery fuel as compared to coal. Liang et al. [3] used a fixed bed test rig to simulate the combustion of the grate boiler and investigated the effects of moisture in fuel on the combustion in the fixed bed. They found out that higher moisture content in fuel would lead to decreased overall NO emission.

Even relatively mature technologies such as typical Selective Non-Catalytic Reduction (SNCR), when applied in small furnaces of stoker boilers, may not guarantee desired results. There is very limited research concerning the application of secondary denitrification systems in stoker boilers and usually typical SNCR is investigated [4,5,6]. Numerous studies propose primary methods such as air staging, staged combustion or flue gas recirculation technology (FGRT) as a method for NO_x limitation in stokers and grate boilers [7,8,9]. However, most of that research is numerical investigation, while full-scale field tests are not common.

In the presented study, a potential of NO_x reduction by means of dry sorbent injection (DSI) is evaluated. It is believed that the DSI process is a suitable solution for small- and medium-capacity furnaces, where the implementation of primary reduction methods is extremely challenging and other secondary reduction methods usually are not economically justified, for example for waste incineration units.

The presented sorbent contains 25% of dry urea powder CO(NH₂)₂ and 75% of halloysite Al₂Si₂O₅(OH)₄ (weight basis) for simultaneous NO_x reduction and mercury adsorption. Very limited data can be found in the literature when it comes to the applicability of dry urea in the SNCR method. Kuropka [10] determined the influence of powdery urea on nitrogen oxides concentration for pulverized coal combustion and the molar ratio of CO(NH₂)₂/NO_x was investigated as well. As a result, the NO_x emission was reduced by 20% when the molar ratio was 1, and by a further 11.6% when the molar ratio was raised to 2.5. In further research [11] the dry urea was investigated together with calcium sorbents for simultaneous reduction of nitrogen and sulfur oxides. The mixture of CO(NH₂)₂ with Ca(OH)₂ or CaCO₃ was injected into the combustion zone within the temperature range of 1000 °C. In both cases, the NO_x emission was reduced from 300 mg/Nm³ to 180–210 mg/Nm³.

Dry urea is not commonly applied as SNCR agent due to the agglomeration tendency and problems with smooth injection into the combustion zone. The urea water solution, which is commonly applied and is easy to inject, contributes to lower the boiler efficiency, which is estimated to be 0.3–2.0 p.p. to evaporate water. The introduction of urea in the dry state does not cause any loss of boiler efficiency. The blending of powdery urea with aluminosilicate, such as halloysite, is expected to reduce the sorbent agglomeration tendencies and reduce problems with proper injection. According to the authors’ best knowledge, the use of a urea-aluminosilicate blend is a novelty and absolutely no studies on this field have been found.

Aluminosilicate minerals applied as additives for combustion minimize adverse effects such as slagging, fouling, ash deposition and chlorine corrosion. Additives such as halloysite, bentonite and kaolinite act via binding of potassium in the potassium aluminosilicates while Cl is released as HCl [12]:

$${\mathrm{Al}}_2{\mathrm{O}}_3·2\mathrm{Si}{\mathrm{O}}_2\left(\mathrm{s}\right)+2 \mathrm{KCl}\left(\mathrm{g}\right)+{\mathrm{H}}_2\mathrm{O}\left(\mathrm{g}\right) \to {\mathrm{K}}_2\mathrm{O}·\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3·2\mathrm{Si}{\mathrm{O}}_2+2\mathrm{HCl}\left(\mathrm{g}\right)$$

(1)

The studies of Hardy et al. [13] showed that among commonly used aluminosilicate sorbents, the most effective is halloysite. Positive properties of the halloysite enhancing combustion of solid fuels were studied by the authors elsewhere [14,15]. Halloysite is a porous alumina silicate with the structure of nanotubes and excellent characteristics for preventing the deposits formation on the heating surfaces of the boiler [16].

Among other minerals in its group, halloysite features a high value of the specific surface area (60–80 m²/g), compared to kaolin (3–12 m²/g), and a high porosity of structure (40–50%). The high porosity makes it possible for the vapors of alkaline compounds to penetrate the grain and bond with the surface inside forming kalsilite, leucite or potassium mullite—substances with a high melting point. These data concern raw halloysite. Activated halloysite may have a specific surface area of up to 500 m²/g, and porosity of up to 80%. All aluminosilicates are relatively inexpensive, commonly occurring minerals. Those features make halloysite a prospective sorbent for the dry flue gas cleaning process.

Dry sorbent injection was investigated as a syngas cleaning method by Szul et al. [17]. Various naturally occurring minerals were selected for testing, such as chalk, dolomite, kaolinite and halloysite. While kaolinite and chalk showed a tendency to agglomerate, halloysite was one of the easiest to flow and handle. The potential of heavy metals adsorption by halloysite was proven as well, however, among investigated metals, mercury was not included, what indicates the knowledge gap.

Recently, different technologies for mercury capture have been investigated, such as catalytic oxidation and sorbent injection. Among these available technologies, the active removal of mercury from the flue gas is realized mainly by the injection of powdered activated carbon (PAC) [18]. The sorbent promotes the adsorption and oxidation of elementary mercury Hg⁽⁰⁾, which increases the share of Hg²⁺ and Hg^(p) in the flue gas. Activated carbon particles with adsorbed mercury can be easily removed by de-dusting devices such as electrostatic precipitators or fabric filters [19].

The high price of activated carbon is the main barrier preventing its wide use in power engineering. In addition, the so-called “activated carbon paradox” is described in the literature. The process of activated carbon production and subsequent regeneration or disposal contributes to environmental pollution [20]. The other unwanted effect is relatively high whole life cycle carbon footprint of powdered activated carbon determined by Gazda-Grzywacz et al. [21]. Therefore, other cost-effective and environmentally friendly sorbents for mercury removal should be found and investigated.

Fly ash can be treated as a sorbent for mercury removal. Charpenteau C. et al. [22] proposed coal fly ash as a low-cost material for mercury removal. Similarly, Liu et al. [23] identified fly ash as a promising mercury adsorbent. Studies showed the larger potential of adsorption of mercury on fly ash with the presence of Fe acting as a catalyst [24,25]. Similarly, a cobalt–iron modified porous carbon sorbent was found to have an excellent Hg⁰ removal performance [26].

Halloysite ore is usually naturally contaminated with iron. In the presented study a halloysite with 20% content of iron oxides was used. The halloysite is expected to capture mercury by both adsorption and oxidation mechanisms. High content of Fe₂O₃ in the halloysite is expected to act as a catalyst and oxidize mercury, which can be adsorbed by the halloysite micropores as well. During presented tests, a significant, over ten-fold increase in the concentration of adsorbed mercury in fly ash was observed compared to the reference sample.

The main advantage of the presented technology is the safe and easy storage of sorbent, which is harmless to the environment [27]. Dry sorbent does not require special conditions of transport and storage, compared to methods using gaseous ammonia and ammonia water.

The novel dry SNCR method presented in the paper is offered under the trade name HALKORNOKS and this name is used throughout the text. The technology uses a combination of powdery urea and halloysite as NO_x reduction and mercury sorption agent. Stoker hot water boiler, WR-25 type, of 30 MW_th capacity from a polish District Heating Plant (DHP) was selected for the investigation. The distribution of sorbent into the combustion chamber is realized with the use of dedicated DSI system. Presented work reports complete tests for two boiler loads: 27.5 MW_th and 24.5 MW_th, and variable nozzle configuration.

The aim of the presented study is to determine:

(1)

The influence of dry urea–halloysite blend on NO_x emission;

(2)

The influence of dry urea–halloysite blend on ammonia slip in flue gas and fly ash;

(3)

The influence of dry urea–halloysite on mercury sorption.

2. Materials and Methods

2.1. Stoker Boiler and In-Furnace Temperature Measurement

The field tests were carried out in an industrial hot water stoker boiler of 30 MW_th capacity, WR-25 type, located in a district heating facility. The boiler scheme with ports of dry sorbent injection is presented in Figure 1. Powdery sorbents can be injected by ports located at three levels of 2480, 3580 and 4730 mm above the grate. Ports 1–4 were used for in-furnace temperature measurement to determine the optimal sorbent injection zone.

According to the most recent knowledge [28,29,30] an optimal temperature window for a selective non-catalytic reduction of nitrogen oxides with urea ranges from 950 to 1050 °C. Injection into a higher temperature zone leads to increased NO_x formation, and into a temperature lower than 950 °C leads to ammonia slip formation. In the case of urea injection in the right temperature zone, it is possible to significantly improve the reduction process; therefore, every full-scale research should be preceded by in-furnace temperature measurement in the combustion chamber to determine the appropriate temperature window.

An in-furnace flue gas temperature measurement for determination of the temperature zone at levels 2480 mm and 3580 mm using uncooled tube and thermocouple was performed. Usually, we conduct the in-furnace temperature measurements by a 5 m water-cooled high-velocity temperature (HVT) aspiration probe. The measurement error resulting from the radiation, which may reach 20%, is reduced to approx. 10% by using a single ceramic thermocouple shield. The use of the double shielded cover and high speed of flue gas aspiration (60–80 m/s) reduces this error to approx. 3% [31]. For the presented boiler, the HVT measurement was impossible because of insufficient space available around the boiler. Therefore, the results obtained from the uncooled probe were corrected based on previous comparative measurements. A series of temperature measurements using both HVT and uncooled probe were carried out and the results were compared. Based on those comparative measurements, the relationship between HVT and uncooled probe results was determined and the corrected temperature value was obtained by adding 75 °C. This is determined to be the corresponding (corrected) temperature value for uncooled thermocouple measurement and is presented in this study.

The research presented in this study was carried out in a municipal heating plant; therefore, the operating conditions were strictly dependent on the municipal heat demand. For presented tests, the in-furnace temperature was determined for maximum boiler output of 29 MW_th via ports no 1–4 according to Figure 1. The results of the corrected in-furnace temperature measurements are given in Figure 2.

Significant variability of the temperature profile can be seen at both 2480 and 3580 mm levels across and along the furnace which is expected in the case of the stoker boiler. At the beginning of the stoker (port no 4) temperature exceeds 1200 °C and the temperature profile do not vary significantly, since the flame is shaped by the ignition arch. Along the grate length, the flame becomes less voluminous. The temperature profile changes: the temperature by the walls lowers and stays high in the center of the furnace, as measured in port 1 and port 3. Moreover, some air enters the furnace through the outermost parts of the grate and thus the temperature near the walls is lower. The average in-furnace temperature in the middle of the stoker (ports no. 1 and 2) shows that for the maximum boiler load the flue gas temperature at the level 3580 mm is above 1020 °C and at 2480 mm it reaches 1050 °C. Based on the measurements, the expected temperature window for minimum and maximum boiler loads can be determined and the most optimal level of sorbent injection can be defined. HALKORNOKS tests took place at a high boiler load, so level 4730 mm was chosen. The temperature measurement at this level was not possible due to insufficient space around the boiler. However, the measurement at levels 2480 and 3580 mm indicated that the sorbent injection should be realized at a higher level and the level of 4730 was the only injection location available.

2.2. Sorbent Injection System

The presented technology can be successfully used for the injection of a wide range of powdery sorbents (combustion additives). The installation includes sorbent (additive) tanks, from which the sorbent is pneumatically transported through process lines to the furnace chamber or exhaust duct using a dosing nozzle. The nozzle has a cooling air duct to protect the material from overheating. Ports for dosing the sorbent to the combustion chamber are located in the existing openings of the boiler walls. The block diagram of the system is shown in Figure 3. The powdery sorbent was distributed into the combustion furnace by ejector nozzles (2 to 4 units) supplied with compressed air and illustrated in Figure 4 and Figure 5. For proper operation of the system, it is necessary to use dehumidified compressed air at a pressure of at least 3 bar.

During presented tests, two types of nozzles: straight nozzle (Figure 4) and splash nozzle (Figure 5) were used. Splash nozzle, compared to the straight nozzle, produces a wider but shorter stream of a sorbent. A wider stream allows better coverage of the combustion furnace volume, wherein its unwanted feature is the distribution of large quantities of sorbent near the walls where the NO_x concentration is lower than in the center (as a result of the lower temperature near the walls [15]). In order to achieve optimum process performance, splash nozzles were placed in the ports near the middle of boiler side walls while straight nozzles near the corners. During the HALKORNOX tests, the angle of nozzles inclination was varied in the range of ±20° in order to hit the right temperature window.

The sorbent used in the study was industrial-application dry urea admixed with halloysite powder in a mass ratio of 1:3. The halloysite used in this study contains 20% of naturally occurring iron oxides. The nominal grain sizes of both halloysite and urea powders were set to <350 µm. The grain size of halloysite was limited by the sieving method. The grain size of urea was standardized by the industrial manufacturer.

The sorbent was used in a dose of 30 kg/h, which approximately corresponded to the stoichiometric excess ratio SR = 2. The SR = 2 was chosen based on the initial research. That research included preliminary tests of the technology for various SR, ranging from 1 to 2. Such range of SR ratio was chosen based on results of Kuropka and Kuropka and Gostomaczyk [10,11]. The most optimal and stable NO_x reduction was achieved for SR = 2, hence it was chosen to be presented in the paper.

2.3. Fly Ash and Fuel Sampling and Chemical Analysis

During presented tests samples of fuel and fly ash were collected and analyzed. Coal samples were taken from the feeding system prior to grate with a sampling period before and after every individual test. The proximate and elemental analysis of coal is listed in

Table 1. Proximate and elemental analysis of fuel.

Parameter	Basis	Unit	Fuel
Moisture	a.r.	%	3.10
Ash	d.b.	%	19.80
Volatile matter	a.r.	%	26.10
HHV	a.r	J/g	26,307
LHV	a.r.	J/g	25,403
C	a.r.	%	64.28
H	a.r.	%	4.03
N	a.r.	%	1.09
S	a.r.	%	0.72
Ammonium compounds as NH3	d.b.	mg/kg	9.10
Hg	d.b.	mg/kg	0.138

. The analysis was conducted according to European standards: moisture content PN-EN ISO 18134-2:2017-03, ash content PN-EN ISO 18122:2016-01, Higher Calorific Value (HHV) and Lower Calorific Value (LHV) PN-EN ISO 18125:2017-07, carbon, hydrogen, nitrogen and sulfur by IR (Infrared) automatic analyzer according to PN-EN ISO 16948:2015-07, mercury content by EPA Method 7473 (spectrophotometric) and ammonia content by a spectrophotometric method according to VGB-B 401:1998.

The fly ash samples were taken from the right and left fly ash hopper (Fly ash sampling point in Figure 1) and an average sample was prepared to form a representative sample from the measurement period. After the tests, fly ash was examined to determine the mercury retention, sulfur content and ammonia slip according to standards listed above for the coal samples.

2.4. Flue Gas Sampling and Analysis

Operating and emission data from a standard boiler data acquisition system of the District Heating Plant (DHP) were collected and supplemented together with data from dedicated Silesian University of Technology (SUT) measurement equipment. The DHP monitored emissions at the collective stack (with other stokers) while SUT monitored emissions directly at the boiler output (Flue gas sampling point in Figure 1). Flue gas composition was measured on-line with multi-sensors Siemens Ultramat 23 analyzer (measuring accuracy 1 ppm for NO, CO, CO₂ and 0.1% ppm for O₂) and Fourier Transformed Infrared Spectroscopy (FTIR) Gasmet analyzer (measuring accuracy 0.1 mg/Nm³), both located at the boiler outlet. The following gas species were measured: NO, SO₂, O₂, CO, CO₂ and NH₃. NO concentration was recalculated into NO₂ concentration at 6% O₂. Such method of NO_x determination is in line with current law regulations, where emission standards are expressed at 6% O₂.

3. Results

Presented tests of dry sorbent injection were fully dependent on ambient air temperature since the District Heating Plant needed to cover the municipality demand. The actual load of the stoker (lower than the maximum load) during the test was caused by the need to maintain the required parameters of the heat distribution network and avoiding the danger of the network overheating. Therefore, the following series of HAKORNOKS tests were performed for two boiler load levels:

• Optimization tests at 27.5 MW_th; sorbent distribution at level +4730 mm.
• Reference test at 23.5 MW_th; sorbent distribution at level +4730 mm.

The optimal levels of injection and nozzles inclination can be determined based on the in-furnace temperature measurement. In general, for HALKORNOKS distribution, the level of 4730 mm is considered to be optimal and the following nozzle inclination is proposed:

• Low load—the level of 4730 mm—nozzles tilted down −20°.
• Average load—the level of 4730 mm—nozzle leveled +0°.
• High load—the level of 4730 mm—nozzle tilted up +20°.

The sorbent was fed into the combustion chamber in a dose of 30 kg/h, what corresponded to the stoichiometric excess ratio SR = 2. Operation parameters for the HALKORNOX technology reference test (the boiler load of 23.5 MW_th) are presented in

Table 2. Operation parameters for HALKORNOKS technology reference test.

Parameter	Value	Unit
Boiler load	23.5	MWth
Test duration	100	min.
Avg. NO2 concentration at the inlet (at 6% O2)	365	mg//Nm3
Avg. NO2 concentration at the exit (at 6% O2)	175	mg//Nm3
NOx conversion (max)	52	%
The percentage of urea in the sorbent	25	%
The stoichiometric excess ratio	2	-
Urea mass flow	7.5	kg of urea/h
Halloysite mass flow	22.5	kg of halloysite/h
Sorbent total mass flow	30.0	kg/h
NH3 slip in flue gas	<2.0	ppm
NH3 slip in fly ash (avg.)	279	mg/kg

.

The reference test confirmed the efficiency of the dry SNCR method. Before the reference test, optimization tests were performed to obtain the highest possible performance in terms of NO_x and ammonia slip reduction. In every case, the NO_x emission was monitored by both DHP and SUT equipment. The results of emission measurements during the optimization and reference tests are presented collectively in Figure 6. All emission data are given at 6% O₂.

The two optimization tests allow adjusting the installation before the reference test. The specified results of NO_x reduction performed at the reference tests are presented in Figure 7.

According to the authors’ industrial experience, the operation of a stoker boiler with optimal primary and secondary air streams and maintaining the proper excess air ratio can reduce the emission of nitrogen oxide emission by about 20% compared to typical operating conditions. This results in the NO_x emission in the range of 310–370 mg/Nm³. Based on the authors’ experience, previously performed measurements and literature data [2,3] it can be concluded that the initial NO_x concentration of 370 mg/Nm³ can be reduced by secondary methods to a stable level below 200 mg/Nm³ regardless of the boiler load.

For the presented tests, the base NO_x emission was determined to be 365 mg/Nm³. During the reference test of the HALKORNOKS the emission was reduced to avg. 175 mg/Nm³ what gives the NO_x reduction of 52%. The SR value of 2 used in the presented research can be considered as correctly chosen. The results are in line with the previous research of SNCR in stoker furnace made by Krawczyk [5] which indicates that with NSR value of 2, the concentration NO_x stabilizes at a satisfactory level and the concentration of N₂O does not increase significantly.

NH₃ slip in the flue gas was stable and did not exceed 2 ppm. This may be considered very satisfying, as increased ammonia slip is an unwanted issue. In flue gas rich in SO₃ and H₂SO_4, it favors the formation of ammonium salts such as ammonium bisulfate NH₄HSO₄ (ABS), a gummy and corrosive substance that is prone to deposit and can build up in various locations of the flue gas flow path [32]. Ammonia compounds in fly ash were determined as 59.6–325 mg/kg as shown in

Table 3. Sulphur, mercury and ammonia content in fly ash.

No	Sample	Sulfur as SO3 (%)	Mercury as Hg (mg/kg)	Ammonia as NH3 (mg/kg)
1	Reference coal ash sample	-	0.138	9.1
2	Fly ash 14:10 (27 MWth)	1.25	2.14	59.6
3	Fly ash 21:10 (23.5 MWth)	0.96	1.40	325
4	Fly ash 22:10 (23.5 MWth)	0.9	1.64	279

.

When confronted with wet SNCR, dry SNCR has no adverse effect on boiler efficiency because it does not increase the stack heat loss. Moreover, it minimizes the corrosion risk from urea since the contact of water and urea at the low temperature can prompt ammonium carbonate formulation, which is a strongly corrosive agent.

Sulfur and mercury concentration in fly ash were determined and are presented in Table 3. The reference coal ash sample (no. 1) was obtained by incineration of coal collected from the feeding system. Fly ash sample no. 2 was collected at 14:10 during previous test not included in this paper. Samples 3 and 4 were collected at 21:10 and 22:10, respectively. The time of samples collection corresponds to process performance in Figure 6 and Figure 7.

A significant, an over ten-fold increase in the concentration of adsorbed mercury in fly ash was observed. Mercury content in the reference sample was determined to be 0.138 mg/kg and in sample no. 4 to 1.64 mg/kg. Mercury was bound in the fly ash as the effect of halloysite addition.

The sorption behavior of fly ash was empowered by the Fe presence in halloysite. Cupric oxide (CuO) and ferric oxide (Fe₂O₃) that are normally present in coal fly ash, were identified as the active components for Hg⁰ oxidation [33] and the addition of halloysite strengthened those features. The presence of NO_x may play an important role in mercury capture as well, since it was proven to do so for model fly ash [34].

Halloysite is expected to adsorb mercury into its pores as well, such as it does for alkali metals in temperature below 815 °C [12,35]. Halloysite is expected to be the most promising aluminosilicate sorbent for mercury capture due to its high porosity and high sorption properties.

In the presented study, the dry sorbent was injected directly into the combustion chamber. That resulted in a longer contact between the flue gas and sorbent particles and is desired for units without fabric filters. In fabric filtration, flue gas penetrates a layer of fly ash with sorbent accumulated on the filtration unit. This contact provides the opportunity for mercury oxidation. This process does not occur in units with electrostatic precipitators, where flue gas does not pass through the accumulated layer.

Dry SNCR based on halloysite-urea blend is prospective to be applied in waste-to-energy units, including combustion of medical, industrial and hazardous waste. It can be potentially used in other stationary sources as well, such as the chemical and metallurgical industry. In those facilities, the simultaneous reduction of NO_x and mercury emission is a key factor promoting the possible application of urea and halloysite mixture.

4. Summary

The novel, dry SNCR method with simultaneous mercury adsorption was presented. This method involves injecting a blend of powdery urea and halloysite Al₂Si₂O₅·(OH)₄ into a combustion zone to reduce NO_x emission. Moreover, the mercury sorption capacity in fly ash was observed.

Bringing together the urea and halloysite powders into a unique dry SNCR injection system led to several positive effects: the NO_x emission was reduced from an average value of 365 to 175 mg/Nm³ and significant, over 10-fold mercury retention was observed in the fly ash. Ammonia slip in the flue gas did not exceed 2 ppm, which proves the safety and stability of the presented method.

By this method, stable NO_x reduction is possible to obtain reaching the emission level below 200 mg/Nm³. Moreover, the introduction of sorbent in the dry state does not cause any loss of boiler efficiency, which is usually 0.3–2.0 p.p. to evaporate water in the classic SNCR.

The research was carried out on a time-limited scale. A long-time continuation is required to determine the impact of the presented method on slagging, fouling, ash deposition and chlorine corrosion of the boiler heating surfaces since the positive impact of halloysite as a fuel additive on those unwanted issues has been already demonstrated for biomass and coal.

The obtained results are an introduction to a deeper analysis and further research on the correlation between the halloysite application and the concentration of gaseous pollutants as well as possible heavy metals retention.

The presented dry sorbent injection system is safe and easy to apply even in small combustion chambers, such as waste-to-energy units or medical and hazardous waste incineration units. Further research and modifications such as increasing the number of nozzles, injection of dry sorbent on several levels and process automation should enable further emission reduction. The presented dry SNCR method could be advantageous when the implementation of a typical installation may not be economically effective. The presented technology can be potentially attractive for units with moderate emission limits and with a strong focus on low/moderate investment costs.