Effect of CO2 on the Desulfurization of Sintering Flue Gas with Hydrated Lime

The effect of carbon dioxide (CO2) on the desulfurization of sintering flue gas with hydrate (Ca(OH)2) as an absorbent was investigated, and the formation of calcium carbonate (CaCO3) and its effect on the desulfurization was discussed. The competitive relationship between carbon dioxide (CO2) and sulfur dioxide (SO2) with the deacidification agent in sintering flue gas is discussed thermodynamically, showing that sulfates are more likely to be generated under high oxygen potential conditions and that SO2 reacts more preferentially than CO2 under a thermodynamic standard state. The amount of produced CaCO3 increases under the condition that the CO2 concentration is absolutely dominant to SO2 in the sintering flue gas atmosphere. The effect of temperature, humidity and CO2 concentration on the desulfurization of Ca(OH)2 are discussed experimentally. The increasing temperature is not conducive to desulfurization, and the humidity can promote desulfurization, while excessive humidity could inhibit desulfurization. The suitable relative humidity is 20%. In situ generated calcium carbonate has a certain desulfurization effect, but the desulfurization effect is not as good as Ca(OH)2. However, a large proportion of CaCO3 was produced in the desulfurization ash under the condition that CO2 concentration was absolutely dominant to SO2 in the sintering flue gas atmosphere.


Introduction
The sintering process is a major source of sulfur and nitrogen pollutant emissions of steel mills, and the proportion of SO 2 and NOx account for about 60% and 50% of the total emissions from the iron and steel industry, respectively [1,2]. In recent years, the Chinese government strengthened environmental protections, and a series of stricter environmental laws and regulations have been published. With the proposed new concept of "Ultralow emission", the SO 2 and NOx emission standards were changed to <35 mg/Nm 3 and <50 mg/Nm 3 , respectively [3], which were ≤200 mg/Nm 3 and ≤300 mg/Nm 3 in the 2015 national standard [2].
To match the emission requirements, many methods were applied in iron and steel factories, and the main methods include wet desulfurization technology [4][5][6][7][8], dry desulfurization technology [9,10] and semi-dry desulfurization technology [9]. The semi-dry desulfurization technology has drawn a lot of attention for its advantages of multi-pollutant synergistic control, water saving and low investment in equipment [11]. The mainstream denitrification technologies are SCR (selected catalytic reduction) [12], SNCR (selected non-catalytic reduction) [13], activated carbon adsorption [14] and ozone oxidation-adsorption [11]. Compared with SCR, SNCR and activated carbon methods, the ozone oxidation-absorption method exhibits the advantages of high oxidation selectivity, high oxidation efficiency, and no secondary by-products [15].
Shanghai Meishan Iron and Steel Co. adopted the semi-dry desulfurization technology for simultaneous desulfurization and denitrification from flue gas using a flue gas circulating fluidized bed. Ozone (O 3 ) was introduced to oxidize NO into dioxide (NO 2 ), which is soluble nitrogen and highly active [16]. In the absorption tower, the main component of hydrated lime is Ca(OH) 2 , the desulfurization rate of sintering flue gas reaches 98% and the denitrification rate reaches 88%, and those meet the requirements of ultra-low emissions for sintering flue gas pollutants.
However, there is a significant increase in the desulfurization ash production under the "ultra-low emission" working conditions, from the original 1349.8/t to 2101.6/t, an increase of about 55.7%. Meanwhile, the composition of CaCO 3 in desulfurization ash also increased from 31.17% to 51.50%, an increase of 65.2%, which not only increased the usage of hydrated lime but also imposed a burden on waste disposal and risked potential environmental pollution. Therefore, it is necessary to know the mechanism of CaCO 3 increase under a high oxide atmosphere.
Hydrated lime is an adsorbent commonly used in semi-dry desulfurization; NO 2 and SO 2 react with hydrated lime, curing to form nitrates and sulfates [17][18][19][20][21][22][23][24]. The main components of the sintering flue gas are SO 2 1026~1321 mg/Nm 3 , NOx 271.4~293.5 mg/Nm 3 , O 2 15.5~16.5 vol%, CO 2 6.1~7.0, H 2 O 10.1~10.9 vol% and dust 28.6~39.0 mg/Nm 3 . The components of CO 2 contribute to the generation of a large amount of CaCO 3 . For a long time, CO 2 has not received attention, and there are few studies on the change in CaCO 3 in the desulfurization ash. CaCO 3 as an adsorbent is mainly used in the high-temperature stage of 800~1200 • C, which decomposes to CaO to complete the desulfurization reaction [25,26]. For the sintering flue gas temperature conditions (80~120 • C), whether CaCO 3 has the effect of desulfurization in the absorption tower has not been clarified.
In this paper, hydrate was selected as the absorbent because it is widely used in the semi-dry desulphurization process in steel works. CO 2 contained in the sintering flue gas is a very important factor that needs to be studied. A study focusing on the effect of CO 2 in desulfurization was conducted. The competitive reaction of the gas components SO 2 and CO 2 in sintering flue gas was analyzed thermodynamically. Under the simulated sintering flue gas condition, the effect of CO 2 concentration, temperature and relative humidity on the desulfurization of calcium-based adsorbent were carried out. The mechanism of CaCO 3 formation and its effect of desulfurization under low temperatures were investigated. The research results will help to understand the mechanism of the formation of desulfurization ash under ultra-low emission conditions.

Experimental Materials
The raw material used in this paper is quicklime, provided by Meishan Iron and Steel Plant (Nanjing, China). The powdered raw material obtained after grinding and crushing was added to deionized water for digestion, and then the finished suspension was dried to constant weight, and finally the adsorbent was crushed and sieved. As shown in Figure 1a, the original quicklime was mainly composed of CaO, CaCO 3 and Ca(OH) 2 , of which the main ingredient was CaO. The main component of the digested sample was Ca(OH) 2 ( Figure 1b). The results of laser particle size analysis (Figure 1c) showed that the particle size distribution of original quicklime was 0.5~100 µm, and the average particle size was 7.8 µm, which was consistent with the SEM analysis results (Figure 1d), and the particle sizes were not uniform. The flow chart of making the samples is shown in Figure 1e. A certain amount of samples was taken, and it was ground into powder for XRF element analysis test. The test results are shown in Table 1 It can be seen from the table that raw quicklime contains elements such as Ca, Mg, Fe, Al, Si and K ( Table 1). As shown in Table 1, the total calcium content is 88.26 wt.%. The TG-DSC curve of the original quicklime sample is shown in Figure 2, which was obtained in an Ar atmosphere. It is clear that the mass loss process occurs in two different stages. The first stage is at 422.2 • C, which is attributed to the decomposition of Ca(OH) 2 , while the second stage, at 647.3 • C, shows the decomposition of CaCO 3 . According to the mass loss rates ∆w 1 and ∆w 2 of the two temperature segments on the thermogravimetric map, the mass fractions of Ca(OH) 2 and CaCO 3 can be roughly obtained; the calculation formula of Ca(OH) 2 content is w = ∆w 1 × 74/18, and the formula of CaCO 3 content is w = ∆w 2 ×100/44. Therefore, we found that the Ca(OH) 2 content is 75.44 wt.% while the CaCO 3 content is 12.43 wt.%. The specific surface area of raw quicklime was 3.3850 m 2 /g (Figure 3a), with a relatively uniform pore size distribution and an isothermal adsorption curve close to type IV with an obvious H3-type hysteresis loop.

Experimental Device
The schematic diagram of the experimental system is shown in Figure 4. The highpurity gases (N 2 , SO 2 and CO 2 ) from the high-pressure cylinder are mixed in the mixing tank. The specific purities of gases are listed in Table 2. The composition of the simulated blast furnace gas is controlled by mass flow controller (MFC). The simulated flue gas reacted with the calcium-based adsorbent in the tubular heating furnace. The gas composition was detected by a Testo 350 flue gas analyzer. The gas from the measuring system then passed through the tail gas absorption with the Na(OH) 2 solution to remove the acidic components of the tail gas.

Relative humidity calculation
Previous studies have shown that Ca(OH) 2 cannot react directly with SO 2 in an absolutely dry environment [27,28], so water is introduced into the desulfurization reaction by the water vapor carrying method.
The principle of this method is based on Dalton's law of partial pressure, and the relative humidity is controlled by adjusting the temperature of the constant temperature water bath. The water bath is heated to a set temperature, maintained for a certain time, and the inert gas N 2 was introduced into the water through the pipeline and bubbled in the water to saturate the water vapor, so as to obtain the humidity conditions required for the experiment. The specific calculation of relative humidity is shown in Equation (1).
H-Absolute humidity in grams per kilogram (g·kg −1 ). P-Saturation vapor pressure of water in Pa (Pa). ∆-Relative humidity in percent (%). P0-The total pressure of wet air in Pa (Pa).

Desulfurization ash composition analysis
Since there is no corresponding standard for the analysis of desulfurization ash components, the component analysis methods are mainly established based on GB176-1996 "Methods of Analysis of Cement" and GB/T5484 "Methods of Chemical Analysis of Gypsum". The calcium hydroxide content and calcium carbonate content are quantitatively analyzed by thermogravimetry, and the calcium sulfite content is determined by the iodometric method.

Evaluations
The evaluations of the desulfurization effect of adsorbent in the experiments are defined as follows.

1.
Desulfurization efficiency The specific calculation of desulfurization efficiency is shown in Equation (2).
η-The desulfurization efficiency (%). C SO 2 ,in -The concentration of SO 2 in the inlet reactor flue gas, mg/Nm 3 . C SO 2 ,out -The concentration of SO 2 in the outlet reactor flue gas, mg/Nm 3 .

Penetration time
When the SO 2 concentration of tail gas is higher than 35 mg/Nm 3 , which is the limitation of the national standard, the adsorbent is penetrated and ineffective, and the elapsed time is defined as the adsorbent penetration time.

Penetration sulfur capacity
The ratio of the absorbed sulfur mass by penetration to the original adsorbent mass is defined as the theoretical sulfur capacity. The specific calculation of theoretical sulfur capacity is shown in Equation (3).
P S -The penetration sulfur capacity, g/100 g adsorbent. Ms-The absorbed sulfur mass by penetration, g. Mt-The original adsorbent mass, g.

Characterization Methods
The particle size distribution (PSD) tests of raw materials were carried out by a laser particle size analyzer (Malvern PANalytical, Mastersizer 2000 MU, The Netherland). The phase composition was characterized by X-ray diffraction (XRD, Bruker-AXS D8 Advance, Billerica, MA, USA) using Cu Kα radiation (λ = 1.54056 Å) at a scan rate of 4 • min −1 from 10 to 80 • (2θ). The microscopic observations and analyses were performed by scanning electron microscope (SEM, FEI Nova Nano SEM 450, Hillsboro, OR, USA). Before being transferred into the test chamber, the specimens were sputtered with gold coating on the surface. Thermal analysis was conducted using thermogravimetric analysis combined with differential scanning calorimetry (TG-DSC, NETZSCH STA449 F3, Selb, Germany). The purge gas used for TG-DSC was N 2 , and the heating rate was set to 10 • C/min. The temperature rise range was set at 100~1100 • C.
Gas compositions were analyzed by a Testo 350 flue gas analyzer. The device is widely used in the analysis of plant flue gas for its small size, high sensitivity and portability. Meanwhile, it can measure a variety of complex gas components at the same time without interference. The analyzer adopted electrochemical principles, using the gas at the corresponding oxidation potential for potential electrolysis, and then calculated the concentration of the gas by the current consumed.

Thermodynamic Calculation
The main components of the sintering flue gas are N 2 , CO 2 and CO, and the main composition of the adsorbent is Ca(OH) 2 . In the temperature range of the sintering flue gas (80~120 • C), N 2 and CO do not react with hydrated lime and are treated as inert gases. The desulfurization reaction of the sintering flue gas is mainly the reaction of SO 2 with Ca(OH) 2 , as shown in reactions (4) and (5). In the flue gas, there is about 6~7 vol% CO 2 , which can also react with Ca(OH) 2 (reaction (6)), forming a competitive relationship with SO 2 . Figure 5 shows the ∆G θ -T relationship of these chemical reactions. In addition, the generated CaCO 3 has a removal effect on SO 2 , which occurs in reactions (7) and (8). Thermodynamically, the desulfurization capacity of Ca(OH) 2 is higher than that of CaCO 3 .
In the presence of high oxygen potential, it is easier to produce CaSO 4 , a stable sulfate, which is more favorable for SO 2 removal.

Effect of Temperature on Desulfurization Performance of Adsorbent
In the absorption tower, the temperatures of the sintering flue gas at the entrance and exit are about 120 and 80 • C, respectively. To simulate the actual temperature of the absorption tower, the desulfurization temperatures were set as 80 • C, 100 • C and 120 • C, respectively. The carrier gas N 2 flow rate was set to 1000 mL/min, the SO 2 flow rate was set to 500 mL/min, and the adsorbent loading was 2 g. Figure 6 shows the desulfurization curves of the adsorbent at different temperatures. It can be seen from Figure 6 that the increase in desulfurization temperature was unfavorable for the desulfurization process, and the penetration time became shorter with the increase in temperature. The penetration time was 573 s under 80 • C and 471 s at 120 • C, with a difference of 102 s. As the temperature increased, the desulfurization capacity was decreased, and the utilization rate of the adsorbent decreased. This is due to the fact that the increase in temperature reduces the solubility of SO 2 in water, which in turn reduces the concentration of SO 2 on the surface of the adsorbent, making the sulfur capacity lower.

Effect of CO 2 Concentration on Desulfurization Performance of Adsorbent
The effect of CO 2 concentration on the desulfurization penetration time was studied by varying the CO 2 content in the gas mixture to 0%, 2.5%, 5.5% and 7.5%, where the reaction temperature was set at 80 • C and the inlet SO 2 concentration was kept at 1200 mg/Nm 3 . The experimental results are shown in Figure 7a. With the increase in CO 2 content, the penetration time of the adsorbent decreased significantly. Under the condition without CO 2 , the penetration time of the adsorbent reached 2500 s. With the addition of 2.5% CO 2 , it plummeted to 600 s. Further increasing the CO 2 concentration, the penetration time did not drop proportional to the CO 2 content.  Figure 7b shows the penetration sulfur capacity of desulfurization agents with different CO 2 concentrations, and it can be seen that as the CO 2 concentration increases, the overall sulfur capacity shows a decreasing tendency. When the CO 2 content is 0%, the penetration sulfur capacity reached the maximum value of 6.5 g/100 g adsorbent, which means that the desulfurization effect of the adsorbent is the best at this time. When the CO 2 content was increased to 2.5%, it was obvious that the penetration sulfur capacity significantly decreased to only 3.3 g/100 g desulfurization agent, and the penetration time was also shortened, indicating that the CO 2 in the gas hinders the desulfurization process of the desulfurization agent and affects the desulfurization performance of it. When the CO 2 content was further increased to 7.5%, there was a small decrease in the penetrating sulfur capacity, but not as drastic as before, indicating that CO 2 has a decreasing effect on the absorption effect of desulfurization agent, and with the increase in CO 2 concentration, there is a plateau, and the decreasing effect is weakened.
With the increase in CO 2 concentration, the penetration time of the adsorbent was slightly shortened, the relative content of Ca(OH) 2 decreased, and the relative content of CaCO 3 increased, as shown in Figure 7c. The S content in the desulfurization ash was decreased, indicating that the desulfurization capacity of the adsorbent decreased. Figure 8a shows the effect of different relative humidities on the desulfurization process. When relative humidity was increased from 0 to 20%, the penetration time was increased significantly. When the relative humidity was further increased, the penetration time was decreased. Correspondingly, the sulfur capacity at 20% relative humidity was the highest (Figure 8b). It can be concluded that an appropriate humidity is helpful for increasing the desulfurization effect of sorbents, as the desulfurization effect will be weakened by excessive water. The 20% relative humidity is appropriate in this study: the sulfur capacity of the adsorbent reaches 29 g/100 g.

Desulfurization Performance of Adsorbent
Under the condition of 20% relative humidity, the penetration time decreased with the increase in CO 2 concentration (Figure 9a). The longest penetration time was 1260 s with 2.5% CO 2 concentration, and the shortest penetration time was 580 s with 7.5% CO 2 concentration. The overall sulfur penetration capacity tends to decrease as the CO 2 concentration increases (Figure 9b). This is because the increase in CO 2 concentration in the gas mixture consumes the adsorbent, resulting in a decrease in the adsorbent for desulfurization. This indicated that the CO 2 concentration weakened the desulfurization effect of the desulfurizing agent at 20% relative humidity, and the sulfur penetration capacity gradually decreased as the CO 2 concentration continued to increase.

Morphology and Pore Structure of Desulfurized Ash
As shown in Figure 10a, the surface of the original adsorbent was rough and porous, resulting in a relatively large specific surface area (as shown in Table 3) that was favorable for gas adsorption. After desulfurization in 2.5% CO 2 atmosphere (Figure 10b), the pores of the adsorbent became shallow, and there were many crystals growing on the surface. When the CO 2 concentration increased to 5.5% (Figure 10c), the surface of the adsorbent obviously became dense and rough. The pores were filled, and the surface was covered with products which had thin strips, flakes, columns, needles and other irregular crystal structures. When the CO 2 concentration was up to 7.5% (Figure 10d), the product crystal grew obviously on the adsorbent surface. As a result, the active points of the adsorbent were occupied and the pores were filled by the product, resulting in a decrease in BET and pore volume and pore diameter.  As shown in Figure 10a, the surface of the original adsorbent was rough and porous, resulting in a relatively large specific surface area (as shown in Table 3) that was favorable for gas adsorption. After desulfurization in 2.5% CO2 atmosphere (Figure 10b), the pores of the adsorbent became shallow, and there were many crystals growing on the surface. When the CO2 concentration increased to 5.5% (Figure 10c), the surface of the adsorbent obviously became dense and rough. The pores were filled, and the surface was covered with products which had thin strips, flakes, columns, needles and other irregular crystal structures. When the CO2 concentration was up to 7.5% (Figure 10d), the product crystal grew obviously on the adsorbent surface. As a result, the active points of the adsorbent were occupied and the pores were filled by the product, resulting in a decrease in BET and pore volume and pore diameter.

Composition Analysis of Desulfurization Ash
The composition analysis of desulfurization ash is shown in Figure 11a. Without adding CO 2 , Ca(OH) 2 accounted for about 76 wt.% of the desulfurization ash, indicating that the utilization rate of the adsorbent was not high. With the increase in CO 2 content in the flue gas, the content of CaCO 3 was increased, and the contents of Ca(OH) 2 and CaSO 4 and CaSO 3 were decreased. This is consistent with the results of the TG analysis, as shown in Figure 11b. The weightlessness stages I (330~450 • C) and II (545~750 • C) represented the decomposed Ca(OH) 2 and CaCO 3 . As we can see, with the increase in CO 2 content, Ca(OH) 2 was decreased and CaCO 3 was increased. This is because the concentration of CO 2 is much higher than that of SO 2 . Under the non-equilibrium condition, the formation trend of CaCO 3 becomes larger. From the above analysis, it can be seen that a large amount of CaCO 3 was generated with high CO 2 concentration in flue gas. To investigate whether in situ generated CaCO 3 has a desulfurization effect, a group of experiments was carried out. Firstly, the adsorbent was put in the flue gas containing 5.5% CO 2 atmosphere. When the CO 2 in the exhaust gas reached stability, it meant the reaction had reached equilibrium, and at this time the surface of the sample was covered with produced CaCO 3 . Secondly, SO 2 was introduced into the flue gas. In order to study the effect of CO 2 on desulfurization, two comparative experiments were designed: one for the CO 2 + SO 2 gas mixture to examine the effect of CO 2 on SO 2 removal, and the other for cutting off CO 2 gas and only passing SO 2 . The experimental results are shown in Figure 12. The penetration time of in situ generated calcium carbonate as the desulfurization agent under the condition of cutting off CO 2 is 405 s, which is higher than that of the CO 2 + SO 2 gas mixture of 342 s but lower than that of Ca(OH) 2 as the adsorbent. It can be seen that the in situ generated CaCO 3 has a certain desulfurization ability, but the desulfurization effect is lower than that of Ca(OH) 2 . It can be presumed that in the sintering flue gas with high CO 2 concentration and strong oxygen potential, CaCO 3 and CaSO 4 are more easily generated, and the reaction can be divided into three stages as shown in Figure 13. At the first stage, CO 2 and SO 2 react with Ca(OH) 2 at the same time to generate CaCO 3 and sulfate, and Ca(OH) 2 is continuously consumed as shown Equations (9)- (11). CaCO 3 is consumed by reacting with SO 2 as in Equation (12), while the generation rate is higher than consumption rate, so the amount of CaCO 3 is increasing. At the second stage, Ca(OH) 2 is consumed, the surface is wrapped by desulfurization production, and the desulfurization by Ca(OH) 2 is weakened. On the contrary, CaCO 3 concentration is increased, and CaCO 3 is the main desulfurization component as in Equation (12). After the peak point, the amount of CaCO 3 is decreased. At the third stage, the export SO 2 concentration exceeds the set value, and the adsorbent fails.
After the implementation of the ultra-low emission standard in the steel industry, the limited emitting concentration of SO 2 of sintering flue gas was decreased from the original 100 mg/Nm3 to 35 mg/Nm3. As shown in Figure 13, when the emission gas standards are reduced from S2 to S1, the penetration time of the absorbent is decreased from t2 to t1, so more fresh adsorbents are needed, and the corresponding CaCO 3 content in the desulfurization ash C1 is also higher than that of C2, which is produced by the original emission standar

Conclusions
A series of experiments were carried out to study the effect of CO 2 on the desulfurization of sintering flue gas with hydrated lime. The main conclusions can be summarized as follows: thermodynamic calculations showed that in the competitive relationship with Ca(OH) 2 , SO 2 reacts more preferentially than CO 2 under a thermodynamic standard state, and the desulfurization reaction is easier in the presence of oxygen. The penetration time of the adsorbent decreases with the increase in CO 2 concentration, and the decrease is more obvious from the penetration of sulfur capacity, which indicates that CO 2 is more likely to generate CaCO 3 under the condition of moisture, thus impeding the desulfurization process. In situ generated calcium carbonate may have a desulfurization effect, but the desulfurization effect is not as good as Ca(OH) 2 .
However, a large proportion of CaCO 3 was produced in the desulfurization ash under the condition that the CO 2 concentration was absolutely dominant to SO 2 in the sintering flue gas atmosphere.