Modeling and simulation of an improved ammonia-based desulfurization process for Claus tail gas treatment

Abstract

An improved ammonia-based desulfurization technology was proposed, in which the produced excess NH₄HSO₃ was concentrated by using the heat from the Claus tail gas and then SO₂ was recovered by decomposing the concentrated NH₄HSO₃ in the Claus furnace. Furthermore, the process was modeled and simulated using the commercial software Aspen Plus. The characteristics of the ammonia-based desulfurization model were validated and compared with experimental results. Moreover, the water balance in the system was maintained, therefore, no extra water was required. Under the water balance condition, the SO₂ removal performance in the system was predicted with various operating conditions. The model exhibited a reliable prediction of the desulfurization performance of the innovative process, which will be favorable for actual industrial application.

---

1 Introduction

Nowadays, coal chemical industries are playing a more and more important role due to economic development and the increasingly serious energy crisis. However, during the coal chemical production process, a large amount of H₂S, which is one of the most toxic environmental pollutants, is produced.^1,2 Therefore, it must be appropriately treated and removed prior to emission. Currently, the Claus process is one of the most important and widely used technologies to recover elemental sulfur from H₂S-containing gases.^3,4 Nevertheless, the sulfur recovery efficiency is only 90–96% for two-stage Claus reactors and 95–98% for three stage Claus reactors,^5–8 due to the thermodynamic limitations of the Claus equilibrium reaction. Thus, there are still 2–3% of sulfur compounds (elemental sulfur vapor, COS, CS₂, SO₂ and H₂S) left in Claus tail gas. From the viewpoint of environmental protection, the Claus tail gas must be seriously considered and treated before discharge into the atmosphere.

As is known, the remaining H₂S in Claus tail gas can be eliminated to a great extent by various additional purification technologies (such as low temperature Claus technology,^9,10 reduction–absorption technology¹¹ and H₂S selective oxidation technology^12–15). Even though additional purification technologies are applied, the sulfur recovery can only reached as high as 99%. On the other hand, SO₂ emission restriction has become increasingly stringent. The Chinese government had enforced the new emission standards of pollutants for petroleum chemical industry in 2015, wherein SO₂ emission has been restricted for 100 mg m⁻³ and 50 mg m⁻³ for special regions. It is significantly indicated that the emission cannot meet the new standard even the Claus tail gas is treated by the mentioned additional process. On this occasion, various absorption technologies, such as CaO,^16–18 MgO^19–21 and NH₃ (ref. 22–24) based desulfurization processes, are seriously considered as the potential candidate for Claus tail gas treatment due to the higher sulfur removal efficiency (can reach almost 100%) and low cost. In these absorption technologies, all sulfur compounds are firstly transformed into SO₂ after Claus reaction. Subsequently, SO₂ is further absorbed by alkaline absorbents and transformed into useful byproducts. Among them, NH₃ based desulfurization process is the most promising one, taking into account that NH₃ can be easily obtained in coal chemical industries. However, in the case of traditional ammonia-based desulfurization process, a tail gas cooling system is required to cool the fuel gas, which wastes a lot of water and energy. Furthermore, the produced excess (NH₄)₂SO₃ and NH₄HSO₃ are oxidized and crystallized to (NH₄)₂SO₄ chemical fertilizer by a long and complex process along with a large consumption of heat. Meanwhile, a large amount of extra water is consumed during the SO₂ absorption reaction.

Therefore, from the viewpoint of practice, the production of (NH₄)₂SO₄ byproducts makes the tail gas treatment process rather complex and further increases the operating costs. Moreover, the storage, transportation and sale of byproducts formed in absorption process are still serious problems.

On basis of this, an improved ammonia-based desulfurization technology is tentatively proposed taking into account of the tail gas composition in coal chemical industry, usually 25% H₂O, 24% CO₂, 50.5% N₂ and 0.5% SO₂. In the process, the produced excess NH₄HSO₃ is concentrated using the heat from the Claus tail gas, which also plays the role of cooling system. And then, the NH₄HSO₃ solution is recycled to the Claus furnace. Wherein, NH₄HSO₃ is decomposed into SO₂, N₂ and H₂O. Subsequently, SO₂ participates in the Claus reaction. Therefore, the process is shorter, heat saving and has no byproduct. It is worth mentioning that extra water for absorption reaction is not required due to the large amount of water in tail gas. More importantly, the purified gas must be maintained at a special temperature in order to take away the excess water in tail gas after absorption reaction, i.e., maintained the water balance of the system. Otherwise, the sulfur removal efficiency will decrease drastically. Additionally, a water spray column is designed to remove the escaped NH₃ as well as the trace SO₂ in purified gas.

In the present paper, the SO₂ absorption and NH₄HSO₃ concentrating under the water balance condition was modeled and simulated by the commercial software Aspen Plus. Moreover, the characteristics of ammonia-based desulfurization process was also studied and compared with experimental results. Additionally, the impact of various operating parameters on SO₂ removal efficiency was examined. All in all, the desulfurization performance of the innovated process can be predicted reliably, which is favorable for the actual industrial application.

2 Computational details

2.1 Process description

A second-stage ammonia-based desulfurization system with a prewashing step was developed. The prewashing step was designed to cool the tail gas as well as to concentrate the produced excess NH₄HSO₃. The process is shown in Fig. 1, including a SO₂ absorbing column, a washing column, a prewashing column, a mixer and a stream splitter. To simplify the process, the fluid transportation units are omitted, and the pressure of the system is considered stable at 100 kPa and without any pressure drop.

Fig. 1 Process for desulfurization system with a prewashing system.

In the present process, the tail gas (TAILGAS) containing a mixture of 25% H₂O, 24% CO₂, 50.5% N₂ and 0.5% SO₂ was firstly introduced to the prewashing column from the bottom, while the rich NH₄HSO₃ solution (RICH) entered from the top of the column. In this heat transfer process, the tail gas was cooled and partial water in the rich liquid was evaporated into the gas phase, i.e., the rich liquid NH₄HSO₃ was concentrated. And then, the concentrated NH₄HSO₃ absorption liquid (PRODUCT) was transferred to Claus furnace, while the cooled tail gas (GASIN) sequentially entered the absorbing column (ABSORBER) from the bottom. SO₂ was absorbed by the (NH₄)₂SO₃-rich solution (LIQIN), which was sprayed from the top of the absorbing column. The NH₄HSO₃-rich solution (LIQOUT) was collected at the bottom of the column. Partial rich solution was regenerated by the mixture of the production of the washing column and ammonia gas, and recycled back to the absorbing column. Meanwhile, the excess rich solution was transferred to the prewashing column and participated in the mentioned heat transfer process. However, the process must be operated at a relatively high temperature due to the requirement of water balance, which facilitated the ammonia escape. Thus, a washing column was designed to inhibit the ammonia escape. Escaped ammonia was absorbed by the sprayed water in the washing column, and the produced ammoniated water was reused in the above mentioned regeneration step associate with ammonia. A mixer (MIXER) was included in the desulfurization model to control the flow rate and composition of the lean solution in the SO₂ absorber.

2.2 Absorption reaction model of SO₂

The reactions that are mainly occurred during the SO₂ absorption reaction are shown in Table 1. The chemical absorption process of SO₂ by ammonia-based solution was similar to physical absorption process, and conformed to two-film theory. SO₂ was firstly transferred from the gas phase into the liquid-film, and then into the liquid phase. And then, SO₂ reacted with the absorbent instantaneously when it entered the liquid phase, which enlarged the solubility of SO₂ in liquid phase as well as enhanced the absorption driving force. Consequently, SO₂ concentration dropped sharply in the liquid phase due to the reactions, which decreased the mass transfer resistance, increased the coefficient of absorption and ultimately improved the absorption efficiency.

Table 1 Reaction model for the NH₃–SO₂–CO₂–H₂O system

Reaction no.	Type	Chemical equation
1	Equilibrium	H2O ↔ OH^− + H^+
2	Equilibrium	H2O + NH3 ↔ NH4^+ + OH^−
3	Equilibrium	H2O + SO2 ↔ H^++ HSO3^−
4	Equilibrium	HSO3^− ↔ H^+ + SO3^2−
5	Equilibrium	H2O + CO2 ↔ H^+ + HCO3^−
6	Equilibrium	HCO3^− ↔ H^++ CO3^2−
7	Salt	NH4HSO3 ↔ NH4^+ + HSO3^−
8	Salt	(NH4)2SO3 + H2O ↔ H2O + 2NH4^+ + SO3^2−
9	Salt	(NH4)2SO3 ↔ 2NH4^+ + SO3^2−
10	Salt	NH4HCO3 ↔ NH4^+ + HCO3^−

---

According to the theory of chemical absorption, SO₂ reacted with the absorbent instantaneously when it entered the liquid phase, therefore the partial pressure of SO₂ on the gas–liquid interface was infinitely close to 0. The diffusion resistance of gas film was the rate-determining step of absorption reaction.

2.3 Rate kinetics of the simulation

During the simulation, SO₂ and CO₂ reacted with the ammonia solution and formed NH₄HSO₃, (NH₄)₂SO₃, NH₄HCO₃, (NH₄)₂CO₃ and (NH₄)₂SO₃·H₂O. The gas reaction between NH₃, SO₂ and CO₂ was ignored during the simulation. The equilibrium constants (K) used for the reactions were calculated by using the following rate equation:where T stands for absolute temperature (K). The model specifications and parameters A, B, C and D are available in the Aspen databank, and are listed in Table 2.

Table 2 Reaction parameters for the NH₃–SO₂–CO₂–H₂O system

Parameter	A	B	C	D
Reaction 1	132.9	−1.345 × 10^4	−22.48	0
Reaction 2	−1.257	−3335	1.497	−0.03706
Reaction 3	−5.979	637.4	0	−0.01513
Reaction 4	−25.29	1333	0	0
Reaction 5	231.5	−1.209 × 10^4	−36.78	0
Reaction 6	216.1	−1.243 × 10^4	−35.48	0
Reaction 7	−1.962	637.4	0	−0.01513
Reaction 8	−1297	3.347 × 10^4	224.2	−0.3516
Reaction 9	920.4	−4.450 × 10^4	−139.3	0.03619
Reaction 10	554.8	−2.244 × 10^4	−89.01	0.06473

---

The rate-based RadFrac model was adopted for the absorbing, prewashing and washing column. The rate-based approach uses the Maxwell–Stefan model to compute the multicomponent mass transfer, including the use of binary coefficients to evaluate the mass transfer rates between the vapor and liquid phases.²⁵ The absorption is based on the two-film theory,²⁶ wherein the mass transfer resistance was concentrated in a thin layer adjacent to the bulk phase boundary. The discretized film equations are combined with the heat and mass balance at each stage to obtain an accurate temperature and concentration profile over the entire column.²⁷

The mixed flow model was selected for the three columns, in which the bulk properties of each phase are assumed to be the same as the outlet conditions for that phase. Wegstein method was employed to solve the system of algebraic equations. The electrolyte NRTL model was selected as the thermo physical properties of the system, by which the thermodynamic properties of the liquid phase were determined. Henry's law was applied on N₂, NH₃, CO₂ and SO₂. The Henry's law constants of SO₂, NH₃ and CO₂ in water were taken from Meyer et al.,²⁸ Que et al.²⁹ and Yan et al.³⁰

The coefficients used for temperature-dependent physical properties of model components were taken from the Design Institute for Physical Properties and the Aspen Plus data system. The molar volume of the electrolyte solution was calculated basing on the Clarke aqueous electrolyte volume model. The liquid viscosity was determined by the Jones–Dole model. The Riedel model was adopted to calculate the thermal conductivity coefficients of the electrolyte solution. The diffusion coefficients were computed basing on the Nernst–Hartley model, and the surface tension of the aqueous electrolyte solution was calculated using the Onsager–Samaras model.

3 Results and discussion

3.1 Model validation

The validation experiments were carried out on special manufactured equipment. The experimental procedure was consistent with the established model. Additionally, pumps and fans were added in the experiments to transport liquid and gas phases. In the experiments, the composition and concentration of the absorbent was adjusted by adding ammonia gas, SO₂ and (NH₄)₂SO₃ to the solution. The experiments were conducted under the same conditions (pH, absorbent temperature, initial total salt concentration, gas volumetric flow, SO₂ concentration) with the simulation. SO₂ concentration, the pH, composition, and concentration of the absorbent solution were determined after the system was in equilibrium for at least 5 min. In the experiment, the concentration of the absorbent (NH₄HSO₃, (NH₄)₂SO₃) was determined by an iodometry method and acid–base neutralization titration. Each sample was titrated in triplicate. The SO₂ concentration was measured by an Infrared Flue Gas Analyzer. The experiments were carried out under the condition of 5000 ppm SO₂, initial total salt concentration 0.6 mol L⁻¹, gas volumetric flow 8 m³ h⁻¹, liquid–gas ratio 8 L m⁻³, absorbent pH 6.7 and absorbent temperature 40 °C.

The simulated SO₂ removal results were verified at different operating conditions (pH, liquid–gas ratio (L/G), SO₂ concentration, absorbent temperature and absorbent concentration) against experimental results. The operating parameters of the Aspen Plus model were set as the same as the experimental conditions. Moreover, the water balance of the system was not considered in this situation.

Fig. 2 shows the relationship between SO₂ removal efficiency and the absorbent pH. The absorbent pH during the experiment was adjusted by adding SO₂ or NH₃ into the solution, while the absorbent pH during the simulation was adjusted by changing the concentration of (NH₄)₂SO₃ and NH₄HSO₃ of the inlet streams. It can be seen that the simulation results agreed well with the experimental data. Both results revealed that the SO₂ removal efficiency increased with the rise of absorbent pH. The SO₂ removal efficiency of experimental data and simulation result around pH = 4 was 27.3% and 20.6%, respectively. And then, the SO₂ removal efficiency rose sharply when the pH reached to 6. The SO₂ removal efficiency could attain as high as 98% at pH = 6.4. Furthermore, the SO₂ removal efficiency kept almost steady when pH was higher than 6.5. The pH of absorbent reflects the composition of the absorbent. A lower pH indicates a lower concentration of (NH₄)₂SO₃ and a higher concentration of NH₄HSO₃, which not only obstructs the SO₂ absorption reaction, but ionizes H⁺ ion, inhibiting the dissolution of SO₂ into the water. The combined affection leads to a lower SO₂ removal efficiency at low pH levels. When pH is higher than 6.5, the absorbent is meanly composed of (NH₄)₂SO₃, and the SO₂ absorption rate is mainly controlled by the mass transfer rate instead of solution composition. It means that the continually increase of pH had little impact on the SO₂ removal efficiency, but greatly increased the operating cost. The pH of the absorbent should be kept between 6 and 7. As a buffer solution, the composition and the concentration of (NH₄)₂SO₃–NH₄HSO₃ system varies even with the same solution pH. Due to the difficulty of controlling both the absorbent concentration and solution pH at the same time, the experimental results showed some variations in absorbing capacities.

Fig. 2 Simulation and experimental results of SO₂ removal efficiency with different absorbent pH.

Liquid–gas ratio (L/G) is an influencing factor that can decide the size of the absorption equipment and influence the operating cost. With a larger L/G ratio, the driving force of absorption enhances, which is in favor of absorption progress, but at the same time increases the operating cost. Fig. 3 illustrated the effect of liquid–gas ratio (L/G) on SO₂ removal efficiency. Both curves exhibited consistent regime. It is significantly noted that the SO₂ removal efficiency increased with the increase of liquid–gas ratio. The SO₂ removal efficiency was only 26% when the liquid–gas ratio was 2. However, the SO₂ removal efficiency increased sharply to 83.6% with the rise of liquid–gas ratio to 4. The growth rate decreased when the liquid–gas ratio exceeded 4. It can be explained by the fact that the absorbent was initiatorily insufficient. Therefore, the increase of liquid–gas ratio implied the increase of the liquid flow rate entering the absorbing system, which increased the contact of gas–liquid. Simultaneously, the mass transfer force increased, which was beneficial to more SO₂ in gas phase entering liquid phase.

Fig. 3 Simulation and experimental results of SO₂ removal efficiency with different liquid–gas ratio.

The influence of different SO₂ inlet concentration on SO₂ removal efficiency is depicted in Fig. 4. It is noteworthy that the simulation results and experimental results exhibit remarkably different variation trend with respect to the rise of SO₂ concentration. In the simulation results, the SO₂ removal efficiency kept almost stable at 98% when the SO₂ concentration rose from 2000 to 15 000 ppm. Whereas, the SO₂ removal efficiency dropped sharply to 86% with a further rise of SO₂ concentration. Conversely, in experimental results, the SO₂ removal efficiency gradually dropped from 100% to 86% with a rise of SO₂ concentration. This difference can be attributed to the fact that the simulated process was conducted in a rather ideal condition, i.e., the two phases in tower are uniform distributing and well intermixed, which assures a better reaction completion. Moreover, the SO₂-absorption process is a gas-film controlled process, in which the gas film resistance decides the reaction rate and degree. Therefore, the initiate increase of SO₂ inlet concentration improves the mass transfer driving force, leading to the rise of SO₂ absorbing capacity. Hence, the SO₂ removal efficiency can be maintained in the simulated process. The gradual decrease of SO₂ removal efficiency in experimental data can be attributed to the insufficiency of the gas–liquid contact, actually. SO₂ molecules cannot reach the gas–liquid interface before they were carried out by the gas stream, causing the gradual decrease of SO₂ removal efficiency.

Fig. 4 Simulation and experimental results of SO₂ removal efficiency with different SO₂ inlet concentration.

Fig. 5 shows the effect of the absorbent (NH₄)₂SO₃ concentration on the SO₂ removal efficiency. It can be noted that the simulation results agreed well with the experimental data. Both results revealed that the SO₂ removal efficiency was almost maintained at 97% with the rise of (NH₄)₂SO₃ concentration. It can be explained that the absorbent was (NH₄)₂SO₃-excess in the process. Therefore, the absorption rate was mainly determined by the mass transfer rate between gas and liquid. Thus the increasing (NH₄)₂SO₃ concentration can hardly improve the SO₂ removal efficiency.

Fig. 5 Simulation and experimental results of SO₂ removal efficiency with different absorbent concentration.

In the experiments, the existence of the fluid transportation units can act as fans and pumps. Namely, the experiments are operated under a micro positive pressure, i.e., the experimental pressure is always a little higher than the simulated process. As known a higher operating pressure is beneficial to the absorbing process. Therefore, under the same operating conditions, the desulfurization efficiencies of experiments are higher than simulated results.

Particularly, almost all the single factor simulation results are extremely consistent with the experimental results. Therefore, it can be considered that the developed model can predict the desulfurization performance of the innovated process reliably.

3.2 Desulfurization performance simulation

The desulfurization performance of the innovated process was simulated by the developed and validated model. The simulation condition and results were collected in Tables 3 and 4, respectively. Especially, all the simulations were conducted based on the water balance of the system (the purified gas was controlled at 70.4 °C). Moreover, the SO₂ removal efficiency can attain as high as 96%.

Table 3 Model parameters and operating conditions

Units	Specifications	Operating conditions
Prewashing column	Calculation type: RadFrac model	Operating pressure: 100 kPa
	Flow model: mixed	Enrichment stream flow rate: 0.71 L h^−1
	Height: 500 mm	Enrichment stream temperature: 70 °C
	Diameter: 100 mm	Tail gas flow rate: 8 m^3 h^−1
	Packing type: Raschig rings 13 mm	Tail gas temperature: 160 °C
		CO2 concentration in tail gas (v/v%): 24
		Water vapor concentration (v/v%): 25.5
		SO2 volume fraction (ppm): 5000
SO2 absorber	Calculation type: RadFrac model	Operating pressure: 100 kPa
	Flow model: mixed	Lean solution flow rate: 7.13 L h^−1
	Height: 1000 mm	Lean solution temperature: 70.2 °C
	Diameter: 100 mm
	Packing type: Raschig rings 13 mm
Washing column	Calculation type: RadFrac model	Operating pressure: 100 kPa
	Flow model: mixed	Inlet temperature of wash water: 25 °C
	Height: 500 mm	Wash water flow rate: 0.75 L h^−1
	Diameter: 100 mm
	Packing type: Raschig rings 13 mm
Mixer	Type: Stream mixer	Operating pressure: 100 kPa
	Valid phase: vapor–liquid	NH3 flow rate: 1.28 mol h^−1
		NH3 temperature: 25 °C
Splitter	Valid phase: vapor–liquid	Operating pressure: 100 kPa
	Recycle flow split ratio: 0.9

---

Table 4 Simulation results with water balance

Stream name
Simulation results	TAILGAS	NH3	WATER	OUT	PRODUCT	RECYCLE	ENRICH
Temperature/°C	160.0	25.00	25.00	70.55	70.29	70.21	70.21
Pressure/kPa	101.0	101.0	101.0	100.0	100.0	100.0	100
Vapor/frac	1.000	1.000	0	1.000	0	0	0
Mole flow/mol h^−1	224.6	1.280	41.51	239.6	26.76	319.9	35.54
Mass flow/kg h^−1	6.633	0.02180	0.7479	6.854	0.5494	6.285	0.6985
Volume flow/L h^−1	8000	31.19	0.7500	6828	0.5734	6.377	0.7085
Enthalpy/Gcal h^−1	−8.160 × 10^−3	−1.400 × 10^−5	−2.800 × 10^−3	−9.200 × 10^−3	−1.85 × 10^−3	−0.02188	−2.400 × 10^−3
pH			7.00		5.12	5.90	5.90
Mole/frac
NH3	0	1	0	5.740 × 10^−4	2.620 × 10^−5	9.340 × 10^−5	9.340 × 10^−5
SO2	5.000 × 10^−3	0	0	2.300 × 10^−4	1.770 × 10^−5	1.520 × 10^−6	1.520 × 10^−6
CO2	0.2400	0	0	0.2250	5.030 × 10^−5	4.890 × 10^−5	4.890 × 10^−5
N2	0.5050	0	0	0.4734	2.950 × 10^−6	2.990 × 10^−6	2.990 × 10^−6
WATER	0.2500	0	1	0.3008	0.9173	0.9411	0.9411
NH4^+	0	0	0	0	4.266 × 10^−2	3.277 × 10^−2	3.277 × 10^−2
H^+	0	0	1.810 × 10^−9	0	2.300 × 10^−7	4.070 × 10^−8	4.070 × 10^−8
HCO3^−	0	0	0	0	1.260 × 10^−5	7.320 × 10^−5	7.320 × 10^−5
OH^−	0	0	1.810 × 10^−9	0	9.640 × 10^−10	5.530 × 10^−9	5.530 × 10^−9
HSO3^−	0	0	0	0	3.713 × 10^−2	1.921 × 10^−2	1.921 × 10^−2
CO3^2−	0	0	0	0	1.240 × 10^−9	3.770 × 10^−8	3.770 × 10^−8
SO3^2−	0	0	0	0	2.759 × 10^−3	6.743 × 10^−3	6.743 × 10^−3

---

3.2.1 Water balance of the system. As mentioned above, the process must be operated under the condition of water balance. Otherwise, the SO₂ removal efficiency will decrease drastically and further increase the operation cost. For the present simulation, water in the system can be classified into three kinds: water of entering the system, water of leaving the system and reacted water. Water of entering the system is mainly originated from the water scrubbing step (using to catch the escaped ammonia) and the TAILGAS (containing about 24% H₂O), while it leaves the system by the outlet streams PRODUCT and OUT. The amount of water participated in the reactions can be calculated by the changes of the concentration of HSO₃⁻, SO₃²⁻, HCO₃⁻, and CO₃²⁻ in the liquid phase entering (WATER) and leaving (PRODUCT) the system according to reactions (1) to (10). The mole flow of water in the inlet and outlet streams are shown in Table 5, and the water consumption of the reactions is shown in Table 6.

Table 5 The mole flow of water entering and leaving the system

Stream	Inlet streams		Outlet streams
	TAILGAS	WATER	OUT	PRODUCT
H2O mole flow/mol h^−1	56.15	41.51	72.06	24.53

---

Table 6 Water consumption of reactions in the system

Mole flow/mol h^−1	WATER stream	PRODUCT stream	Increase of the ion
SO3^2−	0	0.07393	0.07393
HSO3^−	0	0.9937	0.9937
CO3^2−	0	3.330 × 10^−8	3.300 × 10^−8
HCO3^−	0	3.400 × 10^−4	3.400 × 10^−4

---

The consumption of water can be calculated as:

n(H₂O) = Δn(SO₃²⁻) + Δn(HSO₃⁻) + Δn(CO₃²⁻) + Δn(HCO₃⁻) = 1.068 mol h⁻¹

The amount of water entering the system should keep balance with the amount of water leaving the system and participating in the reactions. The water balance equation can be presented as:

n(H₂O-WATER) + n(H₂O-TAILGAS) = n(H₂O-PRODUCT) + n(H₂O-GASOUT) + Δn(H₂O-consumption)

Δn(H₂O) = Δn(H₂O-WATER) + Δn(H₂O-TAILGAS) − Δn(H₂O-PRODUCT) − Δn(H₂O-GASOUT) − Δn(H₂O-consumption) = 0.0109 mol

It positively means that the increased amount of water in the system was only 0.0109 mol h⁻¹, which can be neglected. Therefore, it is considered that the water balance of the system is reached.

3.2.2 Effect of tail gas temperature on SO₂ removal efficiency. Due to the easy volatility of NH₃ and easy decomposition of (NH₄)₂SO₃ and NH₄HSO₃, most of the ammonia-based desulfurization processes are carried out at low temperatures and the temperature of the inlet gas is usually controlled at 60–80 °C. In the present process, the tail gas is cooled by the rich liquid. The impact of tail gas temperature on SO₂ removal efficiency in the ranges of 90 to 170 °C is shown in Fig. 6. It can be clearly observed that the SO₂ removal efficiency increased from 89 to 95.6% along with the tail gas temperature increased from 90 °C to 170 °C.

Fig. 6 Simulation results of SO₂ removal efficiency with different tail gas temperature.

During the course of absorption process, the rise of inlet tail gas temperature can hardly influence the liquid phase temperature due to the notably difference of specific heat capacity. Thus, the differences in temperature between gas and liquid phase were increased. Subsequently, the mass transfer driving force was increased. On the other hand, the composition of the absorption solution varied with different inlet tail gas temperatures. A higher inlet tail gas temperature can induce the increase of reaction rate, therefore the SO₂ absorption capacity increased. Notably, the SO₂ absorption capacity was the rate-determining step in the absorption reaction according to the two-film theory. Therefore, the increase of tail gas temperature induced the increase of SO₂ remove efficiency.

3.2.3 Effect of introduced water temperature on SO₂ removal efficiency. As is known, the introduced water was used to catch the escaped ammonia and further used to regenerate the absorption solution. Therefore, the absorbent temperature varies with the rise of introduced water temperature. On the other hand, the absorbent temperature is an influential parameter that affects the diffusion rates, mass transfer and reaction kinetics. Fig. 7 shows the effect of introduced water temperature on SO₂ removal efficiency. It can be seen that the SO₂ removal efficiency was decreased slightly from 95.5 to 95% when the temperature of introduced water was rose up to 50 °C. Thus, it indicates that the introduced water temperature can hardly influence the SO₂ removal efficiency, which be explained by the fact that the flow rate of the introduced water was rather small compared with the flow rate of lean solution stream, i.e. the temperature of introduced water has little impact on the absorbent temperature.

Fig. 7 Simulation results of SO₂ removal efficiency with different water temperature.

3.2.4 Effect of the volume flow rate of NH₃ on SO₂ removal efficiency. Fig. 8 shows the effect of the volume flow rate of NH₃ added to the system on SO₂ removal efficiency. During the course of experiment, the SO₂ removal efficiency was only 43.61% when the volume flow of NH₃ was 12.2 L h⁻¹. And the concentration of SO₂ and NH₃ in the outlet gas was 2656 and 15 ppm, respectively. However, the SO₂ removal efficiency was raised sharply to 93.75% with the volume flow of NH₃ increased to 29.2 L h⁻¹. Meanwhile, the concentration of SO₂ in the outlet gas exhibited a remarkably decline to 293 ppm, and the concentration of NH₃ rose slightly to 334 ppm. After that, the increase of ammonia volume flow had little positive impact on the SO₂ removal efficiency and the concentration of SO₂ in the outlet gas, which moved from 93.75 to 97.3%. Thus the volume flow rate of ammonia should be kept at the ranges of 30 to 35 L h⁻¹. The continue increase of ammonia flow rate cannot improve the SO₂ removal efficiency but would result in an increase in the ammonia waste and operating cost.

Fig. 8 Simulation results of SO₂ removal efficiency with different inlet NH₃ volume.

3.2.5 Effect of CO₂ fraction on SO₂ removal efficiency. The tail gas contains large quantities of CO₂, which can also be absorbed by the ammonia-based solution, and worsen the SO₂ removal efficiency. Therefore, the effect of CO₂ fraction on SO₂ removal efficiency was studied. The results are shown in Fig. 9. It is significant to observe that the SO₂ removal efficiency was decreased slightly from 95.45% to 95.35% when the CO₂ fraction increased from 8 to 36%. Meanwhile, the CO₂ removal efficiency was maintained at a rather low level. Therefore, it is illustrated that the CO₂ fraction has little effect on SO₂ removal.

Fig. 9 Simulation results of CO₂ fraction on SO₂ removal efficiency.

3.2.6 Ammonia escape. The escape of ammonia in the system is depended on several operating parameters, i.e. inlet tail gas temperature, introduced water temperature, introduced water flow rate and inlet NH₃ flow rate. The influence of these parameters on ammonia escape is shown in Fig. 10. It can be noted that the ammonia escape from the column decreased slightly from 576 to 559 ppm when the introduced water temperature rose from 20 to 50 °C. And the ammonia escape increased from 115 to 661 ppm when the gas temperature was raised from 90 to 170 °C. Meanwhile, by increasing the introduced water flow rate from 0.5 to 4 L h⁻¹, the ammonia escape in the outlet gas increased linearly from 514 to 671 ppm. However, the ammonia outlet concentration increased exponentially from 15 to 7347 ppm when the inlet NH₃ flow rate increased from 12 to 73 mol L⁻¹. Therefore, the ammonia escape decreased with the rise of introduced water temperature. Moreover, it increased with the rise of tail gas temperature, introduced water flow rate and inlet NH₃ flow rate.

Fig. 10 Effect of operating conditions on ammonia escape.

4 Conclusions

An improved ammonia-based desulfurization technology was tentatively proposed, wherein the produced excess NH₄HSO₃ was concentrated using the heat from the Claus tail gas and SO₂ was recovered by decomposing the concentrated NH₄HSO₃ in the Claus furnace. Furthermore, the process was modeled and simulated using the commercial software Aspen Plus. The characteristics of ammonia-based desulfurization model was validated and compared with experimental results. Moreover, the water balance in the system must be maintained, therefore, no extra water was required. Under the water balance condition, the SO₂ removal performance of the system was predicted with various operating conditions. It revealed that the SO₂ removal efficiency increased slightly with the rise of tail gas temperature and introduced water temperature. Oppositely, it increased remarkably with the rise of NH₃ flow rate. Particularly, the effect of CO₂ fraction on SO₂ removal efficiency can be ignored. And the ammonia escape of the system was appropriately inhibited. The model exhibited a reliable prediction on the desulfurization performance of the improved process, which was favorable for the actual industrial application.

The main starting point of the modified process was the realization of elemental sulfur recovery. Therefore, the excess concentrated NH₄HSO₃ was transferred to the Claus furnace and further decomposed to SO₂ and H₂O. However, the Claus reaction was thermodynamic equilibrium limited (2H₂S + SO₂ → 3S + H₂O). Thus, the large extra amount H₂O entering the furnace can play a negative effect to the reaction, causing the decrease of Claus reaction efficiency. Hence, realizing the rational integration of the modified process and Claus process was the most crucial challenge for the further application of this process.

Acknowledgements

This work was financially supported by the National Natural Science Foundation (21577158, 21507148, 21327011) and the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB05050200).

References

1. J. Abbasian and R. B. Slimane, Ind. Eng. Chem. Res., 1998, 37, 2775–2782.
2. J. S. Chung, S. C. Paik, H. S. Kim, D. S. Lee and I. S. Nam, Catal. Today, 1997, 35, 37–43.
3. V. Shinkarev, A. Glushenkov, D. Kuvshinov and G. Kuvshinov, Carbon, 2010, 48, 2004–2012.
4. B. ZareNezhad, J. Ind. Eng. Chem., 2009, 15, 143–147.
5. X. Zhang, G. Dou, Z. Wang, J. Cheng, H. Wang, C. Ma and Z. Hao, Catal. Sci. Technol., 2013, 3, 2778–2785.
6. X. Zhang, G. Dou, Z. Wang, L. Li, Y. Wang, H. Wang and Z. Hao, J. Hazard. Mater., 2013, 260, 104–111.
7. J. S. Eow, Environ. Prog., 2002, 21, 143–162.
8. P. Van Nisselrooya and J. Lagasb, Catal. Today, 1993, 16, 263–271.
9. X. Wang, T. Sun, J. Yang, L. Zhao and J. Jia, Chem. Eng. J., 2008, 142, 48–55.
10. E. Alvarez, S. Mendioroz, V. Mun and J. M. Palacios, Appl. Catal., B, 1996, 9, 179–199.
11. A. N. Zagoruiko, V. V. Shinkarev, S. V. Vanag and G. A. Bukhtiyarova, Catal. Ind., 2010, 2, 343–352.
12. X. Zhang, Z. Wang, N. Qiao, S. Qu and Z. Hao, ACS Catal., 2014, 4, 1500–1510.
13. X. Zhang, Y. Tang, S. Qu, J. Da and Z. Hao, ACS Catal., 2015, 5, 1053–1067.
14. X. Zhang, Y. Tang, N. Qiao, Y. Li, S. Qu and Z. Hao, Appl. Catal., B, 2015, 176, 130–138.
15. X. Zhang, Z. Wang, Y. Tang, N. Qiao, Y. Li, S. Qu and Z. Hao, Catal. Sci. Technol., 2015, 5, 4991–4999.
16. J. Abbasian, A. Rehmat, D. Leppin and D. D. Banerjee, Fuel Process. Technol., 1990, 25, 1–15.
17. H. Zhang, H. Tong, S. Wang, Y. Zhuo, C. Chen and X. Xu, Ind. Eng. Chem. Res., 2006, 45, 6099–6103.
18. A. Lancia, D. Musmarra and F. Pepe, Ind. Eng. Chem. Res., 1997, 36, 197–203.
19. S. J. Lee, S. Y. Jung, S. C. Lee, H. K. Jun, C. K. Ryu and J. C. Kim, Ind. Eng. Chem. Res., 2009, 48, 2691–2696.
20. Z. Shen, X. Chen, M. Tong, S. Guo, M. Ni and J. Lu, Fuel, 2013, 105, 578–584.
21. Z. Shen, S. Guo, W. Kang, K. Zeng, M. Yin, J. Tian and J. Lu, Ind. Eng. Chem. Res., 2012, 51, 4192–4198.
22. Y. Jia, Q. Zhong, X. Fan and X. Wang, Chem. Eng. J., 2010, 164, 132–138.
23. X. Gao, H. Ding, Z. Du, Z. Wu, M. Fang, Z. Luo and K. Cen, Appl. Energy, 2010, 87, 2647–2651.
24. X. Liu, Q. Zhong, J. Wang, X. Ji, Y. Jia, Y. Xu and L. Li, Energy Sources, Part A, 2011, 33, 2027–2035.
25. V. Alopaeus, J. Aittamaa and H. V. Nordén, Chem. Eng. Sci., 1999, 54, 4267–4271.
26. W. K. Lewis and W. G. Whitman, Ind. Eng. Chem., 1924, 16, 1215–1220.
27. R. Schneider, E. Kenig and A. Górak, Chem. Eng. Res. Des., 1999, 77, 633–638.
28. E. Meyer, V. Ermatchkov, Á. P. S. Kamps and G. Maurer, Fluid Phase Equilib., 2006, 244, 137–152.
29. H. Que and C. C. Chen, Ind. Eng. Chem. Res., 2011, 50, 11406–11421.
30. Y. Yan and C. Chen, J. Supercrit. Fluids, 2010, 55, 623–634.

---