Removal of sulfur dioxide by carbon impregnated with triethylenediamine, using indigenously developed pilot scale setup

In order to provide protection against extremely toxic gases, activated carbon (AC) adsorption has long been regarded to be a useful technology in terms of gas removal. AC without chemical impregnation has been considerably less effective than impregnated ACs. AC in present use was modified with an organic amine, i.e., triethylenediamine (TEDA) to enhance the physical and chemical properties of AC in order to remove specific poisonous gases. With the rising concern on environmental pollution, there has been an increased curiosity in ACs as the means for eliminating pollutants from environment. Purpose of this study was to assess the TEDA impregnated AC in terms of adsorption capability for simulant gas like SO2. Analysis was done in a properly designed setup. By using the scheme reported here, significant adsorption of toxic gas was obtained. Maximum removal capability observed by AC-4 for SO2 gas was 374 mg/g-C and its breakthrough time was 264 min. Breakthrough time and adsorption capacity of AC-4 was found to be 25 times and 10 times greater as compared to raw AC. Different characterization techniques were also used to study impregnated AC. It was found that chemical adsorption was the crucial means by which TEDA-impregnated AC removed the simulant gas. Langmuir model was best to represent equilibrium, and adsorption kinetics follow second-order model. The process was endothermic, favorable, and spontaneous.


Introduction
Due to high adsorptive properties of AC, it has been employed in a huge range of applications as an easy and safe technique for eliminating contaminants from air stream and from water as well. Of the many carbonaceous materials, activated carbon has gained this special property. This is mainly because of porous nature and huge surface area of activated carbon which make it functional for removal of irritating and toxic gases from the environment. Impregnating activated carbon (IAC), with warily chosen materials, significantly increases its capacity of adsorption for a large number of gases that raw activated carbon is unable to filter. This marvelous property of activated carbon has been used in the manufacturing of canisters and gas mask filters for the last few decades. With the rising concern on environmental pollution, there has been an increased curiosity in ACs as the means for eliminating pollutants from environment (Abdulrasheed et al. 2018;Huve et al. 2018;Eskandari et al. 2020;Cai et al. 2021;Wen et al. 2021). Both the Occupational Safety and Health Administration (OSHA) and Environmental Protection Agency (EPA) have considered AC adsorption as the "gold standard" technology for optimum removal of contaminants on the priority list (Wu et al. 2007;Yang et al. 2020).
AC is a nice adsorbent for some organic vapors, but for polar and low molecular weight gases, it is a poor adsorbent. On the other hand, impregnated ACs which have been modified with chemical reagent react strongly with these kinds of gases, bind them on the surface and thus remove them from airstream. Activated carbon impregnated with different transition metals and organic amine Responsible Editor: Tito Roberto Cadaval Jr * Sidra Shaoor Kiani sidrachm5@gmail.com plays an important role in the removal of toxic gases such as SO 2 , NO 2 , HCN, CNCl, and H 2 S from the polluted atmosphere with maximum efficiency (Ho et al. 2019a;Kiani et al. 2020;Kiani et al. 2021). On the AC surface, pollutant gas molecules can be adsorbed by two approaches, physisorption and chemisorption. Physisorption being a surface phenomenon holds the adsorbate pollutant molecules on the surface of AC by Van der Waal's forces and classical electrostatic interactions. In chemisorption, adsorbent and adsorbate pollutant molecules are held together on AC interface by means of chemical bond. In case of unimpregnated AC, molecules attach to the surface of AC only by weak physical forces. Owing to the weakness of these interactions between adsorbate and AC, adsorbate can be released into atmosphere easily causing several ecological concerns (Mahle et al. 2010;Muzarpar et al. 2020).
Unimpregnated AC does not have the capability to remove contaminants from airstream to a greater extent. Therefore, processes have been developed to coat chemicals on the AC surface to provide essential filtering capabilities. In order to improve removal mechanism of gases from airstream, various treatments were used. Most important one of them is the use of organic coatings (Ho et al. 2019b;Khayan et al. 2019). Several impregnating materials, i.e., diisopropylamine (DIPA), triethylenediamine (TEDA), piperidine, di-N-propylamine (DNPA), citric acid, and tartaric acid have been used for contaminant removal. These "new generation" impregnated ACs can be used for several applications, most important one being used for toxic gases adsorption from air stream. In nuclear testing, most successful combinations to date consist of coal-based carbon impregnated with TEDA for the removal of radioactive methyl iodide (González-García et al. 2011;Farooq et al. 2012;Zhou et al. 2014;Lee et al. 2020). Various methods have been reported in the literature for the impregnation of TEDA on AC surface. Commonly used methods are soaking method or spraying method. In both of these methods, surface area of AC is reduced to a greater extent as the excess solution causes blocking of pores orifices present on AC surface. In the present work, various types of TEDA-impregnated activated carbon samples were prepared by using a very novel and properly developed setup which involves sublimation process for the impregnation of TEDA on activated carbon using Fluidized bed adsorbing tower (FBT). This method is very much significant in the sense that sublimed TEDA vapors are uniformly distributed on AC surface due to fluidization in FBT. As a result, a uniform layer of TEDA gets deposit on AC surface with a considerable surface area. Breakthrough times of IACs were quantified for SO 2 as challenge gases. The intention of present study was to assess the adsorption capacity of AC and TEDA-impregnated ACs for SO 2 gas. Concentration of gas was measured by using FTIR-based gas analyzer. In addition to IACs, an un-impregnated activated carbon was also tested for gas adsorption capacity.

Materials/chemicals
Coal-based granular AC having surface area 984 m 2 /g, total pore volume 0.423 cm 3 /g, density 0.41 g/cm 3 , moisture content 9.5%, and pH value of 7.8 was used. Chemical used for impregnation was triethylenediamine (C 6 H 12 N 2 ). Leaching agent used was acetonitrile.

Design and scheme of pilot scale impregnation setup
TEDA was impregnated on AC by using sublimation process. This sublimation process was carried out in a specially designed pilot scale impregnation setup which consists of a fluidized bed adsorbing tower (FBT), TEDA vessel, blower, and heat exchanger (HX). These are connected to an electronic control panel (Farooq et al. 2012). Design of setup is shown in Fig. 1. Fluidized bed-adsorbing tower is the most important part of this setup where TEDA in vapor form impregnates on AC. This technique has the advantage that a uniform impregnant layer can be achieved over all internal and external surfaces of AC.

GASMET DX-4040 FTIR gas analyzer
GASMET DX-4040 is a portable FTIR-based gas analyzer used for measuring concentration of gases involved in breakthrough testing. It is designed for onsite measurement at low concentrations. In standard configuration, concentration of 25 gases can be measured simultaneously. For SO 2 , its detection range is from 0 to 100ppmv. The analyzer is connected to a PC with CALCMET 4040 professional software for extended analysis. IR radiations produced by source is modulated by interferometer. The interferometer performs an inverse Fourier transform of the IR radiations emitted by the source. The modulated IR radiations passes through sample cell which absorbs certain wavelength of radiations. The transmitted IR radiation is detected by detector. The signal is digitized by A/D converter.
Before starting the analysis, first of all, zero check and background check were done. Then concentration of challenge gas (against which the adsorbent was to run) was adjusted by selecting the concentrations of toxic and diluent gases by means of mass flow controller. After attaining challenge gas concentration, adsorbent is exposed for adsorption of toxic gas. Adsorption of gas was found by measuring the change in concentration of gas at inlet and outlet with respect to time. This data is then used to plot breakthrough curve.

Sample preparation procedure
First of all, oil present in the HX is heated by means of heater till it attains 250 °C temperature. Blower is then switched on and air moves towards HX. This hot air then travels directly towards FBT. By this manner, air is recirculated till it attains 100 °C temperature within the FBT. AC bed present inside the FBT is also heated by this hot air. In the meantime, TEDA vessel heater is also operated to attain the temperature of 100 °C for TEDA sublimation. At this time, valve is opened, and 100 °C air flows into TEDA vessel, carries TEDA vapors from the vessel, and then moves to the FBT. As a result of fluidization of AC bed in FBT, each and every particle of AC from all sides comes in contact with TEDA-laden air. So, in this manner, TEDA in vapors form gets impregnated on external and internal surfaces of AC. This cycle is repeated for a specified time period until all the TEDA (in the TEDA vessel) is sublimed and gets adsorbed on AC. This method of fluidization and sublimation by using hot air has a number of advantages over other impregnation methods. Here, post drying process of impregnated AC is not needed, and this process also ensures uniform impregnation. The samples prepared with various concentrations of TEDA are listed in Table 1.

Sample characterizations
A Hitachi S-4800 SEM was used to image the samples. Beam extraction current of 15μA, accelerating voltage 20kV, and working distance of 12mm were the characteristic conditions employed. To attain stability and to facilitate imaging,  a conducting carbon paste for fixing the granules of AC was used. Magnification used was 30-500×.
Determination of breakthrough time was done by FTIR-based GASMET Analyzer. For gas filtration, samples were tested by using sulfur dioxide (SO 2 ) as a challenge gas. 0.5g of each sample was tested for SO 2 gas. These were exposed to 57-pmv challenge gas. Total flow rate comprising SO 2 and diluent N 2 was 3L/min. In dilution chamber, SO 2 gets diluted after mixing with N 2 and passed through AC bed for toxic gas adsorption. Maximum SO 2 (toxicity) limit established by OSHA (Occupational Safety and Health Administration) and EPA (Environmental Protection Agency) is 5 ppmv. So, breakthrough time (t b ) was measured as the time when challenge gas concentration downstream of AC bed reached a value of 5ppmv. Experimental setup flow chart for SO 2 testing is shown in Fig. 2.

ASTM standard tests for AC
ASTM standard tests were performed to calculate the particle size distribution, ball-pan hardness number, water solubles, apparent density, total ash content, moisture content, iodine number, and pH of raw and TEDA-impregnated activated carbon sample (ASTM 1996).

Adsorption capacity
Experimentally, adsorption capacity was calculated by using breakthrough time curve data by subtracting the area below the breakthrough time curve from the total area. This final area above the curve gives value of adsorption capacity. Area below breakthrough curve was calculated by integration method. Theoretically, breakthrough time was determined by using modified Wheeler Jonas equation and this equation is written as follows:

Adsorption isotherms
The adsorption data of SO 2 on AC was subjected to three adsorption models, i.e., Langmuir, Freundlich, and Dubinin-Radushkevich (D-R). Langmuir isotherm was applied in the following linearized form (Chen 2015): C e is the equilibrium concentration of gas (SO 2 ) in ppm, C ads is the adsorbed concentration of gas on activated carbon in ppm. Q is the constant which signifies maximum adsorption capacity (mg/g) at monolayer formation, and b is the characteristic Langmuir constant (l/mg) for the adsorption system.
Freundlich adsorption isotherm was applied in the following linearized form (Chen 2015): q e is the amount of gas adsorbed in mg/g, C e is the equilibrium concentration of gas, K f is the Freundlich rate constant (mg/g (l/mg) 1/n ), and 1 n is Freundlich adsorption constant strength.
D-R adsorption isotherm was applied in the following linearized form (Chen 2015): where q s is the constant in D-R isotherm which is related to adsorption capacity (mg/g), K D is constant related to mean free energy of adsorption (mol 2 /kJ 2 ), and ℇ is the Polanyi potential which is given by the following: where R is gas constant, T is absolute temperature and C e is equilibrium concentration of gas.

Adsorption kinetic models
Mass transfer phenomenon can be best understood by its thermodynamics and kinetics. Pseudo-first order and pseudo-second order models were applied to analyze the adsorption data for kinetic study. Pseudo-first order model is the simplest equation which relates dependence of adsorption rate on adsorption capability. Equation is written as (Sumathi et al. 2010) follows: Linearized form is given as follows: q e and q t are the amounts of gas adsorbed in mg/g at equilibrium and at time "t" respectively, K 1 is rate constant of pseudo-first order (min −1 ), and t is contact time.
Equation for pseudo-second order model is written as follows: Linearized form is given as follows: where t is time in min and q t is the amount of gas adsorbed at time t, K 2 is rate constant of pseudo-second order reaction (g mg −1 min −1 ), and q e is equilibrium adsorption capacity in mg/g.

Thermodynamic parameters
The primary thermodynamic parameters which are used to calculate the probability and spontaneity of adsorption process are change of enthalpy (∆H), change of entropy (∆S), and free energy change (∆G). Following equations can be used to predict these parameters (Myers 2002): R is gas constant, T is absolute temperature, and K d is distribution coefficient and is calculated by following equation.
Value of entropy change and enthalpy change was analyzed from intercept and slope of plot between lnK d versus 1 T respectively.

Results and discussion
Comparison of ASTM standard test results of raw AC and AC-4 was made and shown in Table 2. Results clearly indicate that AC-4 showed an enhancement in properties as compared to the raw AC. This is mainly because of the impregnant which imparts these improved properties (Al-Qodah and Shawabkah 2009). So, we can say that the (7) log q e − q t = log q e − k 1 ∕2.303 t prepared AC-4 sample might be a sustainable candidate for purification of noxious gases from atmosphere. SEM analysis was done to study the distribution of TEDA on AC surface. Fig. 3 a and b show SEM images of raw AC at different magnifications which represent cracks and cavities on surface representing a system of complex porous network. Impregnation results in blocking of cavities and cracks and clogging of pore openings (Kiani et al. 2020). Fig. 3 c, d, e, f, g, and h represent the distribution of different concentrations of TEDA on surface of AC-1, AC-3, AC-4, AC-5, AC-7, and AC-8. In all these images, it can be seen that the AC surface is immensely covered with TEDA impregnants. Better the distribution of TEDA on surface, better will be its capability to react with toxic gases in atmosphere by adsorbing them on the surface of AC by means of chemical bond with the impregnant. Apart from the chemical adsorption of gases on AC, some pores are also available on AC which are responsible for physical adsorption of gases (Wu et al. 2007;Arcibar-Orozco et al. 2013). As TEDA is basic in nature, it plays a significant role in the removal of acidic gases like SO 2 from air. pH of samples was measured by ASTM-D3838 method and results are presented in Fig. 4. Results clearly show that as we increase the amount of TEDA, pH value of ACs also increases. This is due to the fact that TEDA is basic in nature and its impregnation makes AC more basic. This increased basicity of adsorbent (by means of basic adsorbent) is highly desirable for the chemical adsorption of acidic gases as it leads to the filtration of acidic gases from contaminated environment more efficiently (Wu et al. 2007).
Gas filtration capability was tested for SO 2 gas. Comparison of breakthrough time was made between the raw and various TEDA-impregnated ACs and shown in Fig. 5a. These results indicated that raw AC showed less breakthrough time, i.e., 11.6 min as compared to other impregnated samples and so responsible for the purification of toxic gases to lesser extent. In case of TEDA-impregnated samples, breakthrough time increases initially with the increase in concentration of TEDA but after a certain limit of impregnation, breakthrough time decreases with further increasing the concentration. The reason behind this is the blocking of pore mouths. Impregnation beyond a certain limit results in the blockage of AC pores, and gas molecules do not find more available sites for adsorption; hence, its breakthrough time decreases considerably (Wu et al. 2007;Bobbitt et al. 2017). After 7% TEDA impregnation, a considerable decrease in breakthrough time was observed. Breakthrough time of AC-1, AC-2, AC-3, AC-4, AC-5, AC-6, AC-7, and AC-8 is 160,191,247,264,234,222,220, and 211 min respectively. Breakthrough time of AC-4 is almost 25 times greater than raw AC. SO 2 adsorption was supposed to occur through subsequent paths (Kiani et al. 2020): In case of raw AC, adsorption of SO 2 gas takes place on the surface of activated carbon by means of weak electrostatic forces of attraction. Owing to the weakness of these interactions between AC and SO 2 gas, SO 2 can be released into atmosphere easily causing several ecological concerns. After modification of AC surface by TEDA impregnation, TEDA molecules present on the surface of activated carbon react chemically with SO 2 gas, and chemisorption of gas occurs here. SO 2 gas gets attach on both the reactive sites of TEDA molecules and results in the formation of strong chemical bond.
Adsorption capacity of raw AC was found to be 4 mgSO 2 /g-C whereas all other impregnated samples showed higher adsorption capacities. Maximum adsorption capacity was observed for AC-4 which is about 10 times higher than raw AC as shown in Table 3. The order of adsorption capacities of samples was found as raw AC<AC-1<AC-2<AC-8<AC-7<AC-6<AC-3<AC-5<AC-4 for SO 2 gas. Impregnation of TEDA on AC surface might be responsible for a greater resistance or reaction sites. As TEDA is basic in nature, its impregnation makes AC surface more basic. This basic surface is responsible for providing sites to acidic gases. As a result, acidic gases come in contact with basic sites present on AC surface and convert them into non-toxic products. In this way, basic nature of TEDA plays an important role in the effective removal of SO 2 from contaminated atmosphere. (Mahle et al. 2010;Perera et al. 2012;Zhou et al. 2018;Hyuncheol Lee et al. 2019).  breakthrough times are slightly on higher side as compared to the experimental ones. Surface area of raw AC and TEDA-impregnated samples is shown in Table 5. Raw AC exhibit surface area of 984m 2 /g which decreases considerably after TEDA impregnation. Table 5 shows that as we increase the amount of TEDA, surface area decreases. Impregnation beyond a certain limit results in the blockage of AC pores, and gas molecules do not find more available sites for adsorption. This results in the reduction of surface area, and consequently we get less breakthrough time and adsorption capacity (Kiani et al. 2020).      Fig. 6 shows comparison of breakthrough time calculated by both methods. It can be seen that experimental and model results are linearly varying, but the model breakthrough time is slightly greater than experimental breakthrough time (Lodewyckx et al. 2004). The reason is that theoretical equations always overpredict physical quantities. The regression value is 0.97 which shows a good linear fit between model and experimental results. Table 6 shows the comparison of adsorption capacity of carbon adsorbents (impregnated with different types of impregnants) available in literature. Literature reveals that raw AC has less adsorption capacity than impregnated activated carbon therefore responsible for the filtration of toxic gases to a lesser extent. Table 6 shows that maximum adsorption capacity is shown when metals and TEDA both are impregnated on activated carbon. Each impregnant is responsible for the filtration of gas to some extent, and the combined effect is therefore enhanced, so we get greater adsorption capacity as we increase the number of impregnants. TEDA is specifically impregnated for the removal of radioactive iodine and methyl iodide from the nuclear power plant exhaust.

Adsorption isotherms
Langmuir adsorption isotherm Fig. 7 shows plot of C e /C ads vs C e for SO 2 adsorption on AC-4 with initial concentration 57 ppm and at 298 K. From the intercept and slope of this plot, values of Langmuir parameters are estimated for the system and its corresponding data is given in Table 7. Both Langmuir parameters increase with temperature. Value of "b" suggests that the affinity between SO 2 and activated carbon is highly favored. "Q" value suggests that the adsorption capacity is favored at 298 K. Maximum adsorption capacity found was 379 mg/g. This value can be considered satisfactory in comparison with the literature cited in this work. It was observed that Langmuir isotherm fits well for SO 2 adsorption on activated carbon surface as compared to Freundlich and D-R isotherm.
For Langmuir isotherm plot value of R 2 =1. Langmuir isotherm considers that the ions remain adsorbed on certain sites that are mono-energetic, and each site binds only one adsorbate molecule without any interaction with the neighboring ions. Moreover, it supports the monolayer adsorption of gas on substrate (Chen 2015;Langmuir 1918). Langmuir model suggests that adsorbate molecules first get adsorb on the external surface of activated carbon in the form of monolayer and then moves inward and get adsorb on the internal surface of activated carbon. Fig. 8 shows the plot of lnC e vs lnq e for the adsorption data of SO 2 gas on AC-4 at initial concentration of 57 ppm and 298 K. Slope and intercept were used to determine the Freundlich constant (Muzarpar et al. 2020). Value of R 2 has been determined and its corresponding data is given in Table 7. R 2 value suggests that this model does not fit well to the adsorption data of SO 2 as compared to Langmuir model. This model deals with the multilayer adsorption of the substance on the adsorbent. Freundlich type of adsorption isotherm is an indication of surface heterogeneity of the adsorbent and is responsible for the multilayer adsorption due to presence of energetically heterogeneous adsorption sites (Freundlich 1907).

D-R isotherm
D-R model is an empirical model formulated for the adsorption process following a pore filling mechanism. It is applied to express the adsorption process occurred on both homogenous and heterogeneous surfaces. Fig. 9 shows the plot of lnC ads vs ℇ 2 for adsorption data of toxic gas on AC-4 surface at initial concentration of 57 ppm and 298 K. From intercept and slope of plot, we get value of D-R parameters (Sumathi et al. 2010). R 2 value provides information about adsorption process. Fig. 9 shows D-R isotherm plot and its corresponding data is given in Table 7. R 2 value suggests that this model is least applicable to SO 2 adsorption on activated carbon surface as compared to Freundlich and Langmuir isotherm model. Adsorption potential is independent of temperature, and it depends upon the nature of adsorbent and adsorbate. Mean free energy of adsorption provides information about the nature of adsorption, whether it is physical, ion exchange, or chemical adsorption (Dubinin et al. 1947). The order of the best fit isotherm is Langmuir˃Freundlich˃Dubinin-Radushkevich.

Adsorption kinetics
The adsorption kinetic curve of SO 2 on activated carbon was obtained at different contact times. Adsorption kinetics show that adsorption sites are occupied gradually by adsorbate molecules. It can also be observed that higher adsorption kinetics were obtained in short times. For better understanding of adsorption process, SO 2 adsorption data was fitted with pseudo-first order and pseudo-second order models. Plot of log (q e -q t ) (mg/g) vs T (min.) for pseudo-first order reaction is shown in Fig. 10 and plot of t/q t (min.g/mg) vs T (min.) for pseudo-second order reaction is shown in Fig. 11 for the adsorption of SO 2 gas on AC-4. The adsorption kinetic parameters for pseudo-first  order and pseudo-second order model are calculated from the corresponding graphs. Pseudo-first order model parameters, q e and k 1 , were found to be 315 mg/g and 0.02 min −1 respectively whereas pseudo-second order parameters, q e and k 2 , were found to be 389 mg/g and 0.008 g mg −1 min −1 respectively. Kinetics of adsorption can be predicted from the corresponding value of R 2 (Yi et al. 2014). Value of R 2 for pseudo-first order reaction is 0.72 and for pseudo-second order is 0.99. This concludes that pseudo-second order reaction is the most favorable path for adsorption of SO 2 on AC-4 and fits best to adsorption process (Sumathi et al. 2010). Best fit of pseudo-second order reaction indicates the chemisorption of SO 2 on activated carbon surface.

Thermodynamic parameters
Plot of lnK d vs 1/T is shown in Fig. 12. From the slope and intercept of this graph, value of ∆H and ∆S have been computed respectively. Positive value of ∆S proposed an increased randomness at the interface in the adsorption process. Positive enthalpy value shows the endothermic reaction while negative free energy indicates the spontaneous nature of adsorption process respectively (Zhou, Yi et al. 2012).
Corresponding values are shown in Table 7. Comparison of three isotherm models shows that Langmuir adsorption isotherm model is the best fitted model as compared to Freundlich and D-R isotherms.

Conclusion
In this study, different TEDA impregnated ACs were prepared by using sublimation method. All samples displayed higher breakthrough times and gas adsorption capacities to a considerable amount as compared to the raw AC. Moreover, AC-4 prepared in this work is a sustainable candidate for the purification of toxic gases from contaminated air by chemisorption on porous carbon surface. Its breakthrough time and adsorption capacity were found to be 264 min and 374 mgSO 2 /g-C respectively for 57 ppmv SO 2 . Being basic in nature, TEDA played a dynamic role in the purification of acidic gases from air. Gas adsorption data are reliable with Langmuir, Freundlich, and D-R isotherms. Adsorption data strongly follow Langmuir adsorption isotherm and pseudosecond order kinetics. Thermodynamic parameters ∆H, ∆G, and ∆S have been calculated. However, the positive enthalpy