Simultaneous removal of SO2 and NOx from flue gas by low-temperature adsorption over activated carbon

An exceptional phenomenon has been observed that SO2 and NOx in flue gas can be effectively adsorbed over activated carbon with a surprising capacity at cold temperatures with the presence of oxygen. In this study, the adsorption characteristics of NO and SO2 over activated carbon at 80, 20, 0, and − 20 is experimentally investigated. Without the presence of oxygen, adsorption of NO is negligible. In the presence of oxygen, NO can be oxidized to NO2 over activated carbon which leads to the co-adsorption of NO/NO2 within the adsorption bed. Catalytic oxidation of NO over activated carbon can be significantly enhanced at cold temperatures, leading to an extraordinary increase of adsorption capacity of NO. With an initial concentration of NO = 200 ppmv and a space velocity of 5000 h−1, the average specific capacity increases from 3.8 to 169.1 mg/g when the temperature decreases from 80 to – 20 ℃. For NO–O2 co-adsorption, the specific capacity increases along the adsorption bed due to the increasing NO2 concentrations. The adsorption capacity of SO2 is also significantly enhanced at cold temperatures. With an initial concentration of SO2 = 1000 ppmv, the specific capacity increases from 12.9 to 123.1 mg/g when the temperature decreases from 80 to – 20 ℃. A novel low-temperature adsorption (LAS) process is developed to simultaneously remove SO2 and NOx from flue gas with a target of near-zero emission. A pilot-scale testing platform with a flue gas flowrate of 3600 Nm3/h is developed and tested. Emission of both SO2 and NOx is less than 1 ppmv, and the predicted energy penalty is about 3% of the net generation.

and NO x in flue gas are major air pollutants responsible for acid rain and photochemical smog. SO 2 is an acidic gas and can be scrubbed by alkalic solutions such as lime, sodium carbonate, ammonia, etc 1,2 . Seawater is also considered as a scrubbing agent for desulfurization 3,4 . NO x is composed of various forms of nitrogen oxides such as NO, NO 2 , N 2 O, etc. The dominant species is nitrogen monoxide (NO) which can be either reduced to N 2 by selective catalytic reduction or oxidized to NO 2 which can be scrubbed by alkalic solutions 5 . Wet flue gas desulfurization (WFGD) and SCR dinitrification are the dominant technologies in power plants nowadays.
In addition, adsorption technology has also been widely used for gas cleanup. Simultaneous removal of SO 2 and NO x by activated carbon or coke has been successfully demonstrated in flue gas treatment 6,7 . An schematic drawing of the process is shown in Fig. 1.
Activated coke technology can remove SO 2 , NO x , Hg and other adsorbable pollutants simultaneously 8 . The removal of SO 2 over activated carbon in the presence of oxygen and moisture involves a series of reactions that leads to the formation of sulfuric acid. The used carbon is regenerated through heating to recover their adsorbing activity. The desorbed SO 2 is recycled as elemental sulfur, sulfuric acid or liquid SO 2 , as shown in Fig. 1. And the desorbed Hg and other trace contaminant species can be separated and collected during the sulfuric acid or sulfur production process. The typical operating temperature is 80-150 ℃ in the adsorber and 350-450 ℃ in the regenerator. The overall adsorption and desorption reactions are as follows 9 : (1) www.nature.com/scientificreports/ For activated coke technology, NO x is not removed by adsorption. It is removed through catalytic reduction over activated carbon by reacting with injecting NH 3 , which converts NO x to non-toxic gas N 2 . The overall reactions are as follows 10 : Activated coke technology has been widely used for sintering flue gas treatment in iron and steel industry 11 , but only a few applications in power plants have been reported 12 . Power plant flue gas normally has a much higher concentration of SO 2 and larger flowrate than sintering flue gas, the specific capacity of SO 2 is not sufficient enough to treat the power plant flue gas. To improve the adsorption capacity, researchers have also investigated the possibility of using other adsorbents such as modified activated carbon 13,14 , activated carbon fibers 15 , molecular sieve 16,17 , alumina substrate impregnated with sodium carbonate 18 , copper oxide 19 , etc. But none of these adsorbents has been successfully commercialized.
Another defective feature of traditional activated coke technology is that NO x is not able to be removed effectively through adsorption and the injection of ammonia is required to improve the denitrification rate. The removal rare of NO x is less than 20% without NH 3 and can be increased to around 70% if sufficient NH 3 is injected 7,20 .
This study proposed a novel low-temperature adsorption (LAS) technology which is able to remove both SO 2 and NO x through adsorption with extraordinary adsorption capacity and high efficiency. The development of LAS technology is inspired by an interesting phenomenon observed accidentally that NO x can be adsorbed by activated carbon effectively with an astonishing capacity when flue gas is cooled to cold temperatures. The fundamental behaviors of NO and SO 2 adsorption at cold temperatures over activated carbon is investigated in this study. Furthermore, a pilot-scale testing platform with a flue gas treatment capacity of 3600 Nm 3 /h is designed and built to validate the performance of LAS technology. A brief introduction and first-hand data from the pilot-scale testing facility is also shared in this study. But the detail results from the pilot tests will be discussed in the future.
A comparison between the LAS technology and traditional activated coke technology is given in Table 1. The specific capacity is obtained in this study. The removal rate of traditional activated coke technology is from reported literature 7 and the removal rate of LAS technology is obtained from the pilot platform.

Method
Material. Commercial coconut activated carbon (CAC) with granular size between 26 and 30 mesh is used in this study. The surface physical properties of CAC is characterized by BET method (Quantachrome QUDRA-SORB SI). The specific surface area of CAC is 1314.5 m 2 /g, the pore size is mainly smaller than 2 nm, as shown in Fig. 2, in which DFT method is used to determine the pore size distribution. CAC is pretreated by heating to 300 ℃ in an vacuum tube for 2 h before each test. Mass of CAC is measured after the pretreatment. The loading density of granular CAC is 0.5 g/cm 3 . Figure 3 is a schematic drawing of the experimental setup for investigating the adsorption behaviors of SO 2 and NO at cold temperatures. The dry flue gas has a volume flow rate of Q = 1 L/  The is because the extra moisture will be removed from the flue gas during cooling process. The flue gas in pre-heated or pre-cooled to adsorption temperature by a coiled copper pipe immersing in a thermostatic bath with a temperature range of -40 to    www.nature.com/scientificreports/ 100 ℃. The granular activated carbon is loaded in a glass tube with an inner diameter of 5 mm which is also immersed in the thermostatic both. The difference between bath temperature and the gas temperature at the exit of adsorption bed is less than 2 ℃. During each test, 6 g of CAC is loaded and the space velocity of adsorption bed is 5000 h −1 . The gas composition of flue gas leaving the adsorption tube is measured by flue gas analyzer testo 350. The gas analyzer is capable of measuring O 2 , CO 2 , SO 2 , NO and NO 2 . The method, accuracy and resolution of the analyzer for each component is given in Table 2.

Pilot platform.
A pilot-scale test platform is designed to simultaneously remove SO 2 , NO x and other adsorbable pollutants based on the novel low-temperature adsorption technology. The designed flue gas flowrate is 3600 Nm 3 /h. The adsorption temperature is between − 20 to 5 ℃. The pollutants control target is near-zero emission: SO 2 and NO x ≦ 1 ppmv. A schematic flowchart of the pilot-scale test platform is shown in Fig. 4. Flue gas is extracted from the inlet duct before SCR denitrification system. Hot flue gas is cooled to around 120 ℃ by an air preheater and the dust is then removed by a bag-type dust remover. Flue gas is further cooled to around 70 ℃ by a residue heat recovery exchanger (HX1) which can generate usable hot water. Flue gas is cooled to cold temperatures by a direct contact cooling (DCC) tower which has three cooling stages. In the lower stage (stage 1), flue gas is cooled close to room temperature by water scrubbing and the cooling load is provided by cooling water; in the second stage (stage 2), flue gas is cooled to 2-5 ℃ by cold water scrubbing and the cooling load is provided by an industrial chiller; in the upper stage (stage 3), flue gas is cooled to below freezing point by calcium chloride solution scrubbing and the cooling load is provided by an industrial refrigerator. The cold flue gas enters the adsorber in which SO 2 and NO x are removed by activated carbon through adsorption. The cold energy of clean flue gas is recovered by cooling the scrubbing water in the DCC cooling tower through HX2. The saturated activated carbon is heated in the regenerator to recover its adsorption activity for repeatable utilization. The desorbed SO 2 can be recovered as elemental sulfur or sulfuric acid, and the desorbed NO x can be introduced to the boiler to form a stable thermal balance of NO x -N 2 -O 2 , which has been successfully demonstrated in the NO x SO process 18 . In this pilot study however, the post-treatment of desorbed gas is not considered.

Results and discussion
Phenomenon. This section is to introduce an interesting observation on the adsorption of SO 2 and NO at cold temperatures, which to our knowledge has not yet been reported in open literature. This phenomenon is the origin of this study as well as the development of LAS technology.  www.nature.com/scientificreports/ The adsorption characteristics of SO 2 and NO at 80, 20 and − 20 ℃ is investigated by the experimental setup shown in Fig. 3. The simulated flue gas has a composition described in section "Experimental setup". During each test, 6 g of CAC is loaded in the tube and the space velocity is 5000 h −1 . The concentration of SO 2 and NO is 1000 and 200 ppmv, respectively. The dry flue gas has a volume flow rate of 1 L/min. The water content is 10 vol.% at 80 ℃, 2.3 vol.% (saturated) at 20 ℃, 0.1 vol.% (saturated) at -20 ℃, respectively. Figure 5 is the breakthrough curve of SO 2 (blue) and NO (red) of the three experiments. The concentrations of SO 2 and NO leaving the adsorption bed (C out ) are measured and plotted. Since the gas analyzer has a resolution of 1 ppmv for both SO 2 and NO, the breakthrough time (t b , min) is defined as the time when C out (SO 2 ) = 1 ppmv and C out (NO) = 1 ppmv. The breakthrough adsorption capacity (A b , mg/g) is defined as the adsorption capacity which can maintain near-zero emission (less than 1 ppmv). The saturated adsorption capacity (A s , mg/g) is defined as the maximum adsorption capacity when the bed stops adsorption. These two specific capacities are key parameters which determine the loading space velocity of adsorption bed and the frequency of bed regeneration.
The breakthrough and saturated adsorption capacity are calculated by the following equations: where A b and A s are the breakthrough and saturated adsorption capacity (mg/g), respectively. t b (min) is the breakthrough time. Q (L/min) is the flow rate of simulated flue gas, which is 1 L/min. M(g/mol) is the molecular weight of SO 2 or NO. m(g) is the mass of loaded CAC, which is 6 g when the space velocity is 5000 h −1 . C in (ppmv) is the inlet concentration, which is 1000 ppmv for SO 2 and 200 ppmv for NO. C out (t) is the outlet concentration of SO 2 or NO at time t(min). In the time integral of Eq. (6), t 0 is the starting time and t e is ending time when the bed stops adsorption (C out = C in ). As shown in Fig. 5, NO breaks through the adsorption bed almost instantaneously at an adsorption temperature of 80 ℃ which is close to the operating temperature of traditional activated coke technology. The breakthrough time of NO is about 5 min at 20 ℃ and 225 min at -20 ℃. The breakthrough time of SO 2 is 20, 68 and 295 min at 80, 20 and -20 ℃, respectively. The specific capacity of SO 2 and NO calculated by Eqs. (5) and (6) is given in Table 3.  www.nature.com/scientificreports/ Based on the observation from Fig. 5 and Table 3, it is found that the specific capacity of both NO and SO 2 increases substantially with decreasing temperature. At above room temperature, the specific capacity of NO is less than 1 mg/g, which is insufficient for adsorption removal. At -20 ℃, the specific capacity of NO increases to 13.24 mg/g, and the removal of NO x from flue gas through adsorption becomes possible. At above room temperatures, the breakthrough time of NO is much shorter than that of SO 2 . At -20 ℃, although the specific capacity of NO is far smaller than that of SO 2 , the breakthrough time of NO is close to that of SO 2 . This is because the concentration of SO 2 is much higher than NO in the flue gas. Therefore, it is technically feasible to remove SO 2 and NO x by adsorption at cold temperatures.
The above experiments reveal the phenomenon of simultaneous adsorption of NO and SO 2 in simulated flue gas. To understand the mechanism of NO and SO 2 adsorption at cold temperatures, the adsorption of NO and SO 2 are studied and discussed separately in sections "Adsorption of NO" and "Adsorption of SO 2 ". Figure 6 shows the breakthrough curves of the NO adsorption over CAC at 80, 20, 0 and − 20 ℃. 6 g of pretreated CAC is loaded in the adsorption tube and the space velocity is 5000 h −1 . The inlet gas has flow rate of 1 L/min, and is composed of NO and N 2 , with an concentration of C in (NO) = 200 ppmv. The concentration of NO at the outlet of adsorption bed is measured and plotted in Fig. 6. The breakthrough time and adsorption capacity of NO adsorption is given in Table 4. The physisorption (Van der Waals adsorption) of NO over activated carbon is enhanced by decreasing the adsorption temperature since physisorption is an exothermic process 21 . Without the presence of oxygen, the specific capacity of NO physisorption is less than 0.5 mg/g even at -20 ℃.   www.nature.com/scientificreports/ Figure 7 shows the breakthrough curves of the NO-O 2 co-adsorption over CAC at 80, 20, 0 and − 20 ℃. 6 g of pretreated CAC is loaded in the adsorption tube and the space velocity is 5000 h −1 . The inlet gas has flow rate of 1 L/min, and is composed of NO (200 ppmv), O 2 (6 vol.%) and N 2 . With the presence of oxygen, NO can be oxidized to NO 2 over activated carbon 22,23 , a steady NO-NO 2 equilibrium will be formed at the exit of activated carbon bed 24,25 . Therefore, the concentration of both NO and NO 2 at the outlet of adsorption bed is measured and plotted in Fig. 7. At 80 ℃, NO (red) breakthrough the adsorption bed within a few minutes and reaches a steady concentration after about 400 min, and about 2% of NO is oxidized to NO 2 . At 20 ℃, NO (red) breakthrough the bed after about 56 min, and NO 2 (blue) is detected after about 1600 min. After about 3500 min, both NO and NO 2 reaches a steady concentration and the oxidation rate of NO is about 45%. When adsorption temperature is lowered to 0 ℃ and − 20℃, the breakthrough time of NO increases to 392 and 1591 min, respectively. Meanwhile, the oxidation rate of NO increases to 69% and 91.5%, respectively. For each test, the sum of NO and NO 2 concentration is 200 ppmv when the steady state is reached.

Adsorption of NO.
The adsorption capacity of NO-O 2 co-adsorption at various temperatures is given in Table 4. Both NO and NO 2 has a breakthrough time during NO-O 2 co-adsorption, as shown in Fig. 7. The breakthrough time for NO-O 2 co-adsorption is defined as the breakthrough of NO which always occurs first. The breakthrough adsorption capacity A b is calculated by Eq. (5) and the saturated adsorption capacity A s is calculated by Eq. (6), where C out (t) = C out (NO x ) = C out (NO) + C out (NO 2 ) in the case of NO-O 2 co-adsorption.
As shown in Table 4, the presence of oxygen can significantly increase the breakthrough time and specific capacity of NO. With presence of oxygen, NO can be oxidized to NO 2 which is a much more adsorbable species over activated carbon 26,27 . The oxidation rate η(NO 2 ) is calculated by the following equation: where C out (NO 2 ) is the concentration of NO 2 when it reaches steady state. C in (NO) is the inlet NO concentration. The oxidation rate of NO at 80, 20, 0 and -20 ℃ is given in Table 4 as well. With the presence of oxygen, the catalytic oxidation of NO is significantly enhanced by at sub-zero temperatures. The mechanism of NO oxidation over activated carbon is complicated, involving both surface reactions and gaseous reactions. The oxidation of gaseous NO by adsorbed oxygen over the activate surface site is believed to be the dominant pathway 26 . www.nature.com/scientificreports/ where C( ) represents the activated carbon with active surface site. Based on reaction (9), the oxidation reaction rate can be calculated by the following equation: Temperature can impact the oxidation reactions in many ways. First of all, the physisorption of oxygen over activated carbon is enhanced and the concentration of C(O) is increased by reducing adsorbing temperature 28 . Secondly, the rate constant k(T) of NO oxidation increases with decreasing temperature 29 .
As shown in Table 4, temperature has a significant impact on the breakthrough time and specific capacity of NO-O 2 co-adsorption. When the adsorption temperature decreases from 80 to -20 ℃, the breakthrough time increases from 3.5 to 1591 min, and the saturated capacity increases from 3.8 to 169 mg/g. This extraordinary increase of specific capacity is due to the NO 2 adsorption. Figure 8 shows the breakthrough curves of the NO-O 2 co-adsorption over CAC with various loading space velocity. In Fig. 8a The purpose of conducting these six independent experiments is to mimic the experiment of simultaneously motoring NO and NO 2 at six different cross sections of a single adsorption bed which is difficult to achieve in our experimental setup.
The average adsorption capacity and oxidation rate along the axial direction of the CAC bed are given in Fig. 9 and Table 5. As shown in Fig. 9, the NO 2 concentration increases along the bed, indicating that the oxidation rate increases along the axial direction of the adsorption bed. At each cross section of the adsorption bed, a stable NO-NO 2 equilibrium is formed both in the gas phase and the adsorption surface. Therefore the NO-O 2 co-adsorption mechanism involves the adsorption of both NO and NO 2 over activated carbon. Since NO 2 is a much more adsorbable species than NO 30 , the adsorption capacity increases with the increasing NO 2 concentration along the axial direction of activated carbon bed, as shown in Fig. 9. This is quite different with the adsorption of SO 2 which has an uniform adsorption capacity along the adsorption bed. Due to this distinct characteristics, it should be noted that the adsorption capacity given in Fig. 9 and Table 5 should be defined as average adsorption capacity of a specific CAC bed. The real adsorption capacity at each cross section should be larger than the average value.
Based on the above analysis, it is now quite clear why the breakthrough time and adsorption capacity of NO-O 2 co-adsorption increases dramatically when the adsorption temperature is lowered to below room temperatures. At cold temperatures, the catalytic oxidation of NO is fastened, therefore the breakthrough time and breakthrough adsorption capacity are largely increased. In addition, the adsorption capacity of NO 2 increases significantly at cold temperatures, leading to prominent increase of saturated adsorption capacity.
Adsorption of SO 2 . Figure 10 shows the breakthrough curves of the SO 2 adsorption over CAC at 80, 20, 0 and -20 ℃. 6 g of pretreated CAC is loaded in the adsorption tube and the space velocity is 5000 h −1 . The inlet gas has flow rate of 1 L/min, and is composed of SO 2 (1000 ppmv), O 2 (6 vol.%) and N 2 . The concentration of SO 2 at the outlet of adsorption bed is measured and plotted in Fig. 10. Results indicate that the breakthrough time and specific capacity of SO 2 adsorption increases with decreasing temperature. When the temperature decreases from 80 to -20 ℃, the breakthrough time and adsorption capacity increase by about 13 and 10 times, respectively. SO 2 adsorption over activated carbon is fast and the breakthrough adsorption capacity is quite close to the saturated adsorption capacity. The breakthrough curve has a sharp slope which is quite different with that of NO adsorption.
The impact of O 2 , CO 2 and H 2 O on the adsorption of SO 2 at various temperatures are also investigated. As shown in Table 6, the presence of oxygen has a slight improvement of SO 2 adsorption due to the catalytic oxidation over activated carbon. The presence of CO 2 in the opposite has a negative impact on the SO 2 adsorption due to the occupation of active carbon surface. The presence of H 2 O and O 2 can enhance the adsorption of SO 2 through H 2 SO 4 adsorption 9 .
The impact of space velocity of CAC load on the adsorption of SO 2 is shown in Fig. 11. The breakthrough time is doubled when the CAC load is doubled. This indicates that the adsorption capacity (mg/g CAC) is irrelevant with space velocity and is a constant value at certain temperature and partial pressure of SO 2 . This is also quite different with the adsorption capacity of NO (with the presence of O 2 ) which increases along the adsorption bed as shown in Fig. 9.
The adsorption process of SO 2 and NO with the presence of oxygen over activated carbon is illustrated in Fig. 12. There are three regions during the SO 2 adsorption process (upleft). In the saturated region, an equilibrium of [SO 2 (g), SO 2 (a)] is established. Where (g) present gaseous phase and (a) represent adsorbed phase. In the adsorption region, SO 2 (g) is being adsorbed and converted to SO 2 (a). Since the adsorption of SO 2 is fast, the adsorption region is within a narrow region. In the fresh carbon region, both SO 2 (g) and SO 2 (a) are zero. When the adsorption bed reaches the saturated status (upright), a homogeneous equilibrium of SO 2 (g) and SO 2 (a) is established cross the entire adsorption bed: Since the adsorption of NO is almost neglectable compared with NO 2 , the total adsorption capacity increases along the adsorption bed in the saturated region. In the adsorption region, the remaining NO is further oxidized to NO 2 and adsorbed. Since the adsorption rate is limited by oxidation rate, the adsorption of NO + O 2 is much slower than SO 2 and the adsorption region is much wider. When it is reaches saturated status, equilibrium of [NO(g), NO(a)] with decreasing partial pressure and equilibrium of [NO 2 (g), NO 2 (a)] with increasing partial pressure are established along the adsorption bed. At each cross section of the bed, the following equilibriums co-exist:        www.nature.com/scientificreports/ platform is shown in Fig. 13. A preliminary low-temperature adsorption test is conducted in October. The flue gas flow rate is 3600 Nm 3 /h and the operating temperature is − 15 to − 20 ℃. Figure 14 is simplified process diagram of the pilot system. Figure 15 shows the inlet and outlet flue gas composition monitored by online CEMS system. Figures 16 and 17 shows some of the operating data of a successive 72 h test. The inlet flue gas has an SO 2 concentration of around 500-1000 ppmv and NO x concentration of around 70-100 ppmv. The concentrations of both SO 2 and NO x are reduced to below 1 ppmv when leaving the adsorber. More tests are undergoing and will be shared as soon as the data are unclassified. In addition, the performance and energy penalty are evaluated by conducting Aspen Plus modeling, and the energy penalty is about 2-3% of the total net power generation depending on the ambient temperatures. The detail data of the pilot test and the modeling work will be presented and discussed shortly.  www.nature.com/scientificreports/

Conclusion
In this study, the oxidation and adsorption characteristics of NO over activated carbon at cold temperatures is investigated. With the presence of oxygen, both the oxidation rate and adsorption of NO over activated carbon is enhanced significantly at cold temperatures. The breakthrough time of NO increases from 3.45 to 1591.75 min when the adsorption temperature decreases from 80 to −20 ℃. At each cross sections along the adsorption bed,   www.nature.com/scientificreports/ NO and NO 2 equilibrium is established with decreasing NO and increasing NO 2 along the bed, which leads to a increasing adsorption capacity along the bed. The adsorption of SO 2 also increases significantly at cold temperatures. The adsorption capacity increases from 12.87 to 123.11 mg/g when the temperature decreases from 80 to − 20 ℃. A novel low-temperature adsorption process is developed to simultaneously remove SO 2 and NO x from flue gas. A pilot scale test platform is built and the low-temperature adsorption process is tested. Near-zero emission of both SO 2 and NO x is achieved during a 72 h performance validation test.