Performance of gamma-Al₂O₃ decorated with potassium salts in the removal of CS₂ from C₅ cracked distillate

Abstract

Deep desulfurization is a key process for the production of high value-added products from C₅ distillates. In this work, different potassium salt modified gamma-Al₂O₃ adsorbents were prepared by an incipient-wetness impregnation method and characterized by N₂ adsorption–desorption, SEM-EDS, TEM, CO₂-TPD, XRD, FT-IR, and IC. The C₅ distillate with a 1200 μg mL⁻¹ sulfur content is desulfurized to less than 10 μg mL⁻¹ within 24 hours by the static adsorption method. For the desulfurization in the fix-bed reactor, the breakthrough sulfur capacity of K₂CO₃-decorated gamma-Al₂O₃ reaches 0.76 wt% under the optimized conditions, viz., at 30 °C, with a sulfur content of 50 μg mL⁻¹ in the raw oil, and a liquid hourly space velocity of 1 h⁻¹. The desulfurization activity of the exhausted adsorbent can be recovered after regeneration. Selective adsorption of CS₂ includes three processes: adsorption, hydrolysis, and oxidation. CS₂ is first adsorbed on the adsorbent and hydrolyzed to form H₂S. H₂S is further oxidized to form S/SO₄²⁻, and then deposits on the surface of the adsorbent. Adsorption, hydrolysis, and oxidation all play essential roles in the removal process of CS₂.

---

1 Introduction

C₅ distillate is one of the by-products of high-temperature cracking of naphtha to ethylene.¹ Its yield is generally about 14–20% that of ethylene. C₅ contains nearly 20 kinds of compounds, including isoprene, cyclopentadiene, and piperylene.² Their unique molecular structures can be used to synthesize other chemical products, which have high economic value.³ In the petrochemical industry, two technologies are commonly used to treat the C₅: one method is to separate it into monomers; another way is to use it as a raw material for producing low-carbon olefins after hydrogenation.

The sulfur composition of C₅ is different compared with other distillates. When naphtha is cracked to ethylene, it is easy to coke due to the high temperature during the pyrolysis process, which affects the product quality. Salari et al.⁴ found that different types of sulfides have different effects on coking performance. At a certain sulfur content, the coking rate is as follows: dimethyl disulfide (DMDS) < disulfide oil < dimethyl sulfide < CS₂. To reduce the amount of coke and prolong the cracking furnace operating cycle, DMDS and CS₂ are selected as coking inhibitors to reduce the coking rate in the industrial production process. Therefore, the main sulfide in the C₅ is CS_2, with a small number of other sulfides such as thioether, mercaptan, and thiophene. The existence of sulfur compounds greatly impacts the physical properties of catalysts, additives, and products, as well as the processing stability and the value of the products. Although the sulfur content of C₅ varies (often in the range of 40–100 μg mL⁻¹) from different manufacturers, the main sulfide is always CS₂.

In traditional desulfurization technology, the more mature hydrodesulfurization technology results in unnecessary olefin saturation, excessive sulfur species,⁵ which is not suitable for the C₅. Compared with traditional hydrogen desulfurization, adsorption desulfurization has the advantages of a short process, low energy consumption, and no hydrogen required.^6–8 However, the C₅ adsorption desulfurization has not achieved large-scale industrialization due to the insufficient experimental data and unclear desulfurization mechanism. Patent CN101450303A⁹ reported a method of adsorbents preparation, they mixed group 8 and group 11 metal salts with sodium hydroxide as the precipitant in an aprotic solvent, then mixed with the Al₂O₃ dry glue to obtain the adsorbent. Under certain conditions, the adsorbent could remove almost all sulfide in C₅ conjugated diene raw materials. However, this desulfurization process requires hydrogen and a higher operating temperature, which leads to the saturation and polymerization of C₅ olefin. Patent CN103182291A¹⁰ used the pseudo-boehmite powder to prepare gamma-Al₂O₃ support and then impregnated a mixed solution of Zn²⁺, Cu²⁺, group 1, or group 2 metal to obtain a desulfurization adsorbent. The desulfurization of the cracked C₅ using a fixed bed can reduce the sulfur content to 1 μg g⁻¹ or lower. However, the adsorbent sulfur capacity is low, and regeneration is difficult.

The methane reaction with sulfur and hydrogen sulfide is considered the leading cause of undesirable CS₂ in the modified Claus process.¹¹ A highly efficient hydrolysis catalyst is required to remove CS₂ and COS. The reactions of CS₂ hydrolysis are as follows:¹² CS₂ + H₂O → COS + H₂S, COS + H₂O → CO₂ + H₂S, CS₂ + 2H₂O → CO₂ + 2H₂S. The C₅ contains a trace amount of water, and the hydrolysis reaction of CS₂ may occur. Therefore, using a hydrolysis catalyst to remove CS₂ from the C₅ would be an effective method.^13,14 In this paper, the gamma-Al₂O₃ support is loaded with different potassium salts by an incipient-wetness impregnation method. The C₅ distillate is used as raw material. Static adsorption experiments and fixed bed dynamic experiments are carried out to study the adsorption performance of the modified adsorbent on CS₂. Furthermore, the desulfurization mechanism is investigated by N₂ adsorption–desorption, SEM-EDS, TEM, XRD, CO₂-TPD, FT-IR, and IC.

2 Results and discussion

2.1 Performance of catalysts in static experiments

Fig. 1 shows the desulfurization performance of gamma-Al₂O₃ adsorbents decorated by three different potassium salt solutions. The Al₂O₃ can remove the CS₂ in C₅, and the sulfur content in C₅ is 989 μg mL⁻¹ after 3 hours. After the calcination at 500 °C, the KNO₃ in the KNO₃/Al₂O₃ catalyst decomposes into K₂O and form an Al–O–K structure with Al₂O₃.¹⁵ However, the adsorption desulfurization ability of KNO₃/Al₂O₃ is slightly improved, only 16 μg mL⁻¹ higher after 3 hours. The KOH/Al₂O₃ and K₂CO₃/Al₂O₃ improve the adsorption and desulfurization ability of the Al₂O₃ support significantly, indicating that the alkaline has good selective adsorption for CS₂. The modified Al₂O₃ is more alkaline than Al₂O₃, which corresponds to the subsequent CO₂-TPD characterization. The desulfurization effect of K₂CO₃/Al₂O₃ is better than that of KOH/Al₂O₃. When the adsorption time is 3 hours, the sulfur content in the product is 412 μg mL⁻¹ and 496 μg mL⁻¹, respectively. The cause of this phenomenon is analyzed in the next section.

Fig. 1 Removal performance of CS₂ over Al₂O₃ modified by different precursors.

Fig. 2 shows the effect of adsorbents prepared by different concentrations of K₂CO₃ on the removal performance of CS₂. The desulfurization capacity gradually increases with the increase of K₂CO₃ impregnation concentration. When the impregnation concentration of K₂CO₃ is 0.3 g mL⁻¹, the static desulfurization performance is the greatest, and the sulfur content in C₅ can be reduced to less than 10 μg mL⁻¹ from 1200 μg mL; however, continue increasing the impregnation concentration K₂CO₃ results in decreasing of the desulfurization capacity. This indicates that the amount of alkali has a significant influence on the desulfurization activity. The loading of K₂CO₃ affects the alkalinity of the adsorbent surface and the distribution of basic sites in the desulfurizer. The weak base center on the surface of adsorbent is more critical, and the strong alkalinity is not conducive to the adsorption of CS₂. Excessive alkalinity causes irreversible adsorption of CS₂ and intermediate formation (i.e., H₂S and S) in the adsorbent,¹⁶ covering the active sites, which is why the same loading amount of KOH is not as good as K₂CO₃. Consequently, the experimental K₂CO₃ immersion concentration is determined as 0.3 g mL⁻¹. These results indicate that OH⁻ on the adsorbent plays a vital role in the adsorption activity.

Fig. 2 Removal performance of CS₂ over Al₂O₃ modified by different K₂CO₃ concentration.

The C₅ component contains round 80 μg mL⁻¹ of water, so we used the 0.3 g mL⁻¹ K₂CO₃/Al₂O₃ catalyst to investigate the influence of water. In Fig. 3, as the concentration of added water increases, the initial desulfurization rate gradually increases, which shows that the addition of water promotes the hydrolysis rate of CS₂. However, with the further progress of desulfurization, CS₂ the final desulfurization result decreases with added water. The water may compete with CS₂ for adsorption, resulting in the weakening of the adsorption capacity of CS₂.¹⁷ The increase in water is also beneficial to the oxidation rate. Primavera et al.¹⁸ believed that water plays an essential role in the oxidation of H₂S. The increase of water content can promote the deposition of sulfur products on the catalyst, resulting in an increasing deactivation rate. As the addition of water increases, the water cannot dissolve thoroughly in the C₅, resulting in the separation of oil and water. Therefore, the addition of water cannot improve the performance of the catalyst but reduce the quality of the products.

Fig. 3 Effect of water on the removal performance.

2.2 Performance of catalysts in fixed-bed reactor

In order to investigate the effect of reaction conditions on the adsorption performance of modified Al₂O₃, the fixed bed was used to study the dynamic adsorption performance of 0.3 g mL⁻¹ K₂CO₃/Al₂O₃ catalyst. Fig. 4 shows the effect of (a) sulfur contents, (b) LSHVs, and (c) reaction temperature on the adsorption desulfurization performance in the dynamic desulfurization process of C₅ distillate.

Fig. 4 Effect of (a) sulfur content of raw materials, (b) LHSV, and (c) temperature on removal performance of CS₂.

Sulfur content is one of the most important factors affecting the desulfurization capacity of the adsorbent. When the raw material sulfur content is 50, 75, and 100 μg mL⁻¹, the breakthrough sulfur capacity is 0.76 wt%, 0.60 wt%, and 0.46 wt%, respectively. The increase in the sulfur content of C₅ means that H₂S generation increases, which also causes an increase in oxidation products of S and sulfate. The corresponding sulfur products deposit on the surface of the adsorbent, cover the active sites, and hinder the conversion of CS₂. Therefore, the increase of sulfur content in C₅ leads to the decrease of the adsorption desulfurization performance of the adsorbent.

The prepared K₂CO₃/Al₂O₃ adsorbent is used to investigate the effect of different LHSVs on the desulfurization performance of the adsorbents at room temperature. Fig. 4(b) shows that the adsorption capacity of the adsorbent for CS₂ gradually decreases as the space velocity increases. When the space velocity is 0.5, 1.0, and 1.5 h⁻¹, the breakthrough sulfur content is 0.86 wt%, 0.76 wt%, and 0.50 wt%, respectively. When the feedstock feed amount is low, the CS₂ molecule in the C₅ oil has a longer residence time on the surface of the adsorbent, which can promote its adsorption and diffusion process. So, the adsorption performance is better at low space velocity. However, when the space velocity is 0.5 h⁻¹, the amount of raw material oil processed per unit time is too small to meet industrial demand. When the space velocity is too high, the contact time between CS₂ and the surface of the adsorbent is shortened, which is not conducive to the adsorption desulfurization process of the adsorbent. Therefore, it is most suitable for industrial applications when the space velocity is 1.0 h⁻¹.

The prepared 0.3 g mL⁻¹ K₂CO₃/Al₂O₃ adsorbent is applied to examine the effect of different adsorption temperatures on the desulfurization performance of the adsorbent at an adsorption LSHV of 1.0 h⁻¹. The influence of temperature on the desulfurization effect is shown in Fig. 4(c). This experiment selects a temperature range of 5–55 °C. At 5 °C, the breakthrough sulfur capacity is 0.57 wt%. It is because the hydrolysis reaction rate of CS₂ in C₅ is low at a relatively lower temperature, resulting in poor desulfurization activity of the adsorbent. With the increase of temperature, the hydrolysis rate increases, and the reaction is more likely to occur. Besides, high temperature also favors the oxidation of H₂S, facilitating the next reaction. When the temperature is 30 °C, the breakthrough sulfur capacity is 0.76 wt%, and the desulfurization effect is best. The increase in the temperature leads to a gradual rise of SO₄²⁻/S ratio.¹⁹ Sulfate is formed more rapidly, causing catalyst poisoning and inhibiting the progress of the hydrolysis reaction. When the temperature is 55 °C, the breakthrough sulfur capacity of the adsorbent is 0.47 wt%, which is not suitable for the desulfurization of C₅. Besides, the product begins to turn yellow at 55 °C, indicating that the product has been aggregated, thus affecting desulfurization activity and product quality.

2.3 Evaluation of regeneration performance of adsorbents

The performance evaluation of regenerated adsorbents is carried out in a fixed bed reactor, and its results are shown in Table 1. The evaluation conditions are 30 °C, with a sulfur content of 50 μg mL⁻¹ in the raw oil, and an LHSV of 1 h⁻¹. The sulfur capacity of the adsorbent is 0.74 wt% for the first regeneration and 0.71 wt% for the second regeneration. The adsorbent shows good regeneration performance.

Table 1 Breakthrough sulfur capacity of adsorbents with different regeneration times

Regeneration times	0	1	2
Sulfur capacity (wt%)	0.76	0.74	0.71

---

2.4 Characterization of the adsorbents

Table 2 shows the pore structure parameters of the adsorbents prepared via different concentrations of K₂CO₃. It can be seen from Table 2 that as the K₂CO₃ loading increases, the average pore size increases slightly, while the specific surface area and total pore volume decrease significantly. After loading K₂CO₃, the average pore size increase is due to the pores being filled with K₂CO₃, rather than the pore size becoming larger due to impregnation. The decrease in specific surface area and total pore volume is due to the filling of K₂CO₃, the small pore size disappears, and the large pore size becomes small. When the loading of K₂CO₃ increases from 0.3 g mL⁻¹ to 0.4 g mL⁻¹, a large amount of K₂CO₃ is deposited on the surface, the specific surface area decreased from 211.79 m² g⁻¹ to 158.54 m² g⁻¹, and the total pore volume decreased from 0.37 cm³ g⁻¹ to 0.31 cm³ g⁻¹.

Table 2 Structural characteristics of prepared K₂CO₃/Al₂O₃ adsorbents

K2CO3 concentration (g mL^−1)	0.0	0.2	0.3	0.4
BET surface area (m^2 g^−1)	272.09	229.27	211.79	158.54
Total pore volume (cm^3 g^−1)	0.47	0.39	0.37	0.31
Average pore diameter (nm)	6.02	6.07	6.08	6.89

---

Fig. 5(a) is an isotherm adsorption–desorption curve of different concentrations of K₂CO₃/Al₂O₃ adsorbent. In Fig. 5, the isothermal adsorption–desorption curves of different K₂CO₃/Al₂O₃ adsorbents are type IV and have the same H hysteresis loop in medium pressure and high-pressure parts (0.4 < P/P₀ < 1), which is generally considered to be interconnected by the pore size of the mesoporous material.²⁰ In the low-pressure section, the adsorption amount is gently increased, and N₂ molecules are adsorbed on the mesopores inner surface in a single layer to a plurality of layers. There is a sudden increase in the adsorption amount at P/P₀ = 0.4–0.8, which means that the mesoporous structure is relatively uniform. The pore size distribution of the adsorbent is obtained by the Barrett–Joyner–Halenda method (BJH), which is shown in Fig. 5(b). Interaction between the K₂CO₃ with the support Al₂O₃ does not significantly change the structure.²¹ The pore size distribution of the adsorbent loaded with different concentrations of K₂CO₃ is similar, mainly concentrated at 4–5 nm. When the K₂CO₃ loading is 0.3 g mL⁻¹, the pore diameter is the smallest, which is mainly about 4.3 nm. The pore size of this range is suitable for the adsorption of CS₂, and the adsorption desulfurization performance is best.

Fig. 5 (a) Adsorption isotherm and (b) PSD of gamma-Al₂O₃ and supported K₂CO₃ based adsorbents.

Fig. 6 shows the SEM image of virgin Al₂O₃, fresh 0.2 g mL⁻¹, 0.3 g mL⁻¹, 0.4 g mL⁻¹ K₂CO₃/Al₂O₃ adsorbent and deactivated 0.3 g mL⁻¹ K₂CO₃/Al₂O₃ adsorbent. When K₂CO₃ is not loaded, the surface of Al₂O₃ is very rough with a large number of holes, which is an irregular block structure. After the loading of K₂CO₃, the pores of Al₂O₃ are filled with K₂CO₃, and the surface becomes smooth gradually. The unfilled K₂CO₃ is loaded on the surface of the support in the form of small particles. As the loading increased, filamentous K₂CO₃ crystals appeared on the surface of the 0.4 g mL⁻¹ K₂CO₃/Al₂O₃ adsorbent.

Fig. 6 SEM of (a) the virgin-Al₂O₃, (b–d) the fresh 0.2 g mL⁻¹, 0.3 g mL⁻¹, 0.4 g mL⁻¹ K₂CO₃/Al₂O₃, (e) the exhausted 0.3 g mL⁻¹ K₂CO₃/Al₂O₃.

When 0.3 g mL⁻¹ K₂CO₃/Al₂O₃ adsorbent adsorbs CS₂, the product produced by the adsorption process will block the pores and accumulate on the surface, one of the reasons for the decreased activity after adsorption of CS₂.²² Among them, many kinds of filamentous or banded crystals appear on the surface of the adsorbed Al₂O₃ support. Since the loading of K did not increase, the crystals that appeared are associated with the adsorbed products, and the specific results are further analyzed by TEM.

Table 3 shows the EDS characterization result corresponding to the SEM. Virgin-Al₂O₃ contains only Al and O elements, and other elements are almost zero. After loading K₂CO₃, the content of the K element gradually increases as the amount of load increases. When the load concentration of K₂CO₃ is increased from 0.3 g mL⁻¹ to 0.4 g mL⁻¹, the mass increase of K is small, probably because the pores are already filled with K₂CO₃, and excessive K species blocks the pores and enriches the surface, resulting in a decrease in the loading mass. When CS₂ is adsorbed, the presence of the S element can be detected, and the mass is 1.23 wt%, which means that the S element can be well adsorbed on the K₂CO₃/Al₂O₃ adsorbent.

Table 3 EDS analysis of the surface of fresh and exhausted K₂CO₃/Al₂O₃

		Al	O	K	S
Virgin Al2O3	at.%	39.62	60.28	0.10	0.01
	wt%	47.33	52.47	0.19	0.01
Fresh 0.2 g mL^−1 K2CO3/Al2O3	at.%	36.18	60.25	3.52	0.05
	wt%	46.95	46.35	6.62	0.07
Fresh 0.3 g mL^−1 K2CO3/Al2O3	at.%	29.68	62.54	7.74	0.03
	wt%	38.04	47.52	14.38	0.05
Fresh 0.4 g mL^−1 K2CO3/Al2O3	at.%	30.47	60.24	9.28	0.01
	wt%	38.25	44.85	16.89	0.01
Exhausted 0.3 g mL^−1 K2CO3/Al2O3	at.%	32.34	61.25	5.61	0.80
	wt%	41.60	46.72	10.46	1.23

---

Since a newly formed crystal are found in the SEM image after desulfurization, TEM analysis is used. Fig. 7 shows a TEM image of the 0.3 g mL⁻¹ K₂CO₃/Al₂O₃ adsorbent adsorbed by CS₂ and a diffraction pattern of the adsorbed product. After the adsorption of CS₂, a wafer appears around the adsorbent, and three crystal rings appear in the diffraction pattern, which is analyzed and calculated. Comparing the calculated Di with the PDF card, it is found that the three Di are 5.756, 1.988, and 1.374, respectively, corresponding to the three diffractive crystal surface (113), (408), and (288) of the S element, which conforms to the S elemental crystal (JCPDS No. 83-2283), demonstrating the formation of elemental S in the products.

Fig. 7 (a) TEM images of exhausted 0.3 g mL⁻¹ K₂CO₃/Al₂O₃ adsorbent and (b) diffraction pattern taken from region c marked in (a).

Fig. 8 is the CO₂-TPD spectra of the 0.3 g mL⁻¹ KNO₃/Al₂O₃, K₂CO₃/Al₂O₃, and KOH/Al₂O₃ adsorbents. For both K₂CO₃/Al₂O₃ and KOH/Al₂O₃ adsorbents, there are two CO₂ desorption peaks at a lower temperature (∼100 °C) and higher temperature (∼600 °C), which belong to the weak base and strong base centers, respectively. Besides, both peak strength and temperature of KOH adsorbent are higher than those of K₂CO₃, indicating that the alkalinity of KOH/Al₂O₃ is stronger than that of K₂CO₃/Al₂O₃ adsorbent. KNO₃/Al₂O₃ adsorbent has only one weak base and one medium strong base adsorption center, and the desulfurization effect is only improved a little compared with pure Al₂O₃ adsorbent. In the desulfurization process, the hydrolysis center is OH⁻, and water plays a role in supplementing OH⁻.^23,24 However, high alkalinity is not favorable for the hydrolysis of CS₂ because CS₂ and its final hydrolyzed products (H₂S and CO₂) are easily irreversibly adsorbed on the surface,¹⁶ inhibiting the catalytic hydrolysis reaction. Therefore, the adsorption and hydrolysis capacity of KOH with strong alkalinity is weaker than that of K₂CO₃.

Fig. 8 CO₂-TPD spectra of 0.3 g mL⁻¹ KNO₃/Al₂O₃, K₂CO₃/Al₂O₃, and KOH/Al₂O₃ adsorbents.

Fig. 9 is an XRD spectrum of K₂CO₃, Al₂O₃ support, fresh 0.3 g mL⁻¹ K₂CO₃/Al₂O₃ adsorbent, and exhausted 0.3 g mL⁻¹ K₂CO₃/Al₂O₃ adsorbent. The main peaks of pure K₂CO₃ XRD images are at 2θ = 26.3, 30.1, 32.2, 34.3, and 42.9°, which are consistent with PDF cards (JCPDS No. 71-1466).²⁵ The peaks of the XRD images of the support are mainly at 2θ = 19.5, 37.6, 39.4, 45.8, 60.8, and 67.0°, which is consistent with the position of the main peak of gamma-Al₂O₃ (JCPDS No. 10-0425). Compared with gamma-Al₂O₃, the strength of the main peak after K₂CO₃ modification is lower, and a new K₂CO₃ peak is generated at 2θ = 32.2 and 34.3°. The peak intensity is lower than pure K₂CO₃, which means that K₂CO₃ has a good load on the Al₂O₃ and has good dispersion.²⁵ Liu et al.²⁶ found that the K₂CO₃/Al₂O₃ adsorbent after calcining at 500 °C contained K₂O, which may be overlaid in the XRD spectra due to lower content. Therefore, the composition of K₂CO₃/Al₂O₃ adsorbent is K₂CO₃, Al₂O₃, and a small amount of K₂O.

Fig. 9 XRD spectra of the K₂CO₃, the virgin adsorbent, the fresh adsorbent, and the exhausted adsorbent.

When the adsorbent adsorbs CS₂, no peak of elemental S was found because the generated elemental S is easily sublimated in the air, resulting in too little amount on the surface of the adsorbent being difficult to detect. At 2θ = 26.4 and 29.8°, there is a new peak generation, and the peak intensity at 2θ = 30.8° increases, which is the characteristic peak of K₂SO₄ (JCPDS No. 73-1674), which means the production of SO₄²⁻ after adsorption of CS₂. The peak intensity of Al₂O₃ and K₂CO₃ after adsorption is reduced due to the deposition of sulfur products formed by adsorption of CS₂ on the adsorbents.

Fig. 10 is a Fourier infrared spectrum of the 0.3 g mL⁻¹ K₂CO₃/Al₂O₃ adsorbent before and after the adsorption of CS₂. It can be seen from Fig. 10 that the peak appearing around 3558 cm⁻¹ is a characteristic peak of –OH,^27–29 which is derived from the –OH in K₂CO₃/Al₂O₃ adsorbent during catalyst preparation. The peaks at 1400 cm⁻¹ and 1580 cm⁻¹ are characteristic peaks of –COO and –C=O, which are related to the supported potassium carbonate CO₂.²⁶ –COO and –C=O promote the oxidation of H₂S, which accelerates the reaction rate of H₂S by O₂ to S, further improving the conversion ability of CS₂. The peaks at 787 cm⁻¹ (AlO₄) and 582 cm⁻¹ (AlO₆) are related to the Al–O–Al structure.³⁰

Fig. 10 FT-IR spectra of (a) the fresh and exhausted K₂CO₃/Al₂O₃ adsorbents and (b) the enlarged area c marked in (a).

After adsorption of CS₂, new peaks at 2329 cm⁻¹ and 2370 cm⁻¹ are related to CO₂,³¹ indicating that the CO₂ produced by hydrolysis is chemically adsorbed to the adsorbent. The peaks at 2919 cm⁻¹ and 2850 cm⁻¹ are related to –CH₃ and –CH₂, indicating the residual of the adsorbed C₅ on the adsorbent. The peaks at 1383 cm⁻¹ and 1129 cm⁻¹ are related to SO₄²⁻, which means that sulfite or sulfate may form. The peak at 617 cm⁻¹ represents the formation of elemental S,³² and the adsorbed CS₂ produces elemental sulfur on the surface of the K₂CO₃/Al₂CO₃ adsorbent. This result is consistent with that reported in the literature.³³

Although we have used XRD and FT-IR to find sulfate group, due to its small content, the peak is not obvious, and it is not possible to distinguish sulfate group from sulfite group. So, the contents of sulfate and sulfite in the fresh and the spent catalysts are analyzed by IC to better explore the sulfur species on the adsorbent. The IC results in Table 4 show that there are no sulfates and sulfites in the fresh adsorbents. The exhausted adsorbent contains 2.62 wt% sulfate but no sulfite. The accumulation of sulfate on the catalyst is the main reason for the decrease of CS₂ adsorption performance.

Table 4 Sulfate and sulfite contents of the fresh and the spent catalysts

	The fresh adsorbent	The spent adsorbent
SO4^2− (wt%)	0.00	2.62
SO3^2− (wt%)	0.00	0.00

---

2.5 Discussion on the reaction mechanism

When the K₂CO₃/Al₂O₃ adsorbent adsorbs CS₂ in C₅, the adsorbent gradually turns yellow due to the accumulation of sulfur on the adsorbent.³⁴ After adsorbed K₂CO₃/Al₂O₃ is taken out, and it is a yellow granule after grinding. The sulfur smell is dissipated, and the yellow color gradually becomes lighter as time lapses. The adsorbent is further ground to powder. The powder completely turned white after 2 hours, as shown in Fig. 11. This is because CS₂ adsorbs on the adsorbent and is converted into elemental S, and elemental S is disappeared when exposed to the air.

Fig. 11 The macro morphology of exhausted adsorbent.

The removal process of CS₂ in C₅ distillate includes three processes of adsorption, hydrolysis, and oxidation. The adsorption is related to the pore structure of Al₂O₃ and active components. The OH⁻ on the support is mainly derived from K₂CO₃ which is essential for the CS₂ conversion.^35,36 During the hydrolysis process, trace water in the C₅ distillate oil and a large number of OH⁻ on the adsorbent can hydrolyze CS₂ to H₂S. The C₅ contains approximately 50–70 μg mL⁻¹ of dissolved O₂; the O₂ then oxidizes H₂S to elemental S or SO₄²⁻, covering the active site of the adsorbent, resulting in a decrease in adsorption capacity.³⁷ During this period, OH⁻, –COO and C=O act to promote oxidation.³⁸ The specific process is shown in Fig. 12.

Fig. 12 The mechanism of removal of CS₂ by K₂CO₃/Al₂O₃.

3 Methods

3.1 Materials

The pseudo-boehmite powder and sesbania powder were purchased from Shandong Zibo Hengyi Chemical Technology Co., Ltd. K₂CO₃ was purchased from Tianjin Reagent Factory no. 3 Factory (A. R.), and potassium nitrate was purchased from Beijing Chemical Plant (A. R.).

3.2 Preparation of adsorbent

A certain amount of pseudo-boehmite powder was mixed with an appropriate amount of sesbania powder. 2 wt% dilute nitric acid solution was added to the mixed powder and mixed well. After kneading in a squeezer for 90 min, the template was extruded and dried at 120 °C for 12 h followed by calcination in air at 500 °C for 4 hours to prepare strip-shaped gamma-Al₂O₃ support.

The same concentration (0.3 g mL⁻¹) of KNO₃, KOH, K₂CO₃ solution and different concentrations (0.1 g mL⁻¹, 0.2 g mL⁻¹, 0.3 g mL⁻¹, and 0.4 g mL⁻¹) of K₂CO₃ solution were prepared. The gamma-Al₂O₃ support was impregnated with these solutions, respectively, by an incipient-wetness impregnation method, dried at 120 °C for 12 hours, and then calcined at 500 °C for 4 hours to obtain the corresponding potassium salt modified gamma-Al₂O₃ desulfurization adsorbents. They were denoted as 0.3 g mL⁻¹ KNO₃/Al₂O₃, 0.3 g mL⁻¹ KOH/Al₂O₃, 0.1 g mL⁻¹ K₂CO₃/Al₂O₃, 0.2 g mL⁻¹ K₂CO₃/Al₂O₃, 0.3 g mL⁻¹ K₂CO₃/Al₂O₃, and 0.4 g mL⁻¹ K₂CO₃/Al₂O₃. Before the experiments, the catalyst was calcined at 400 °C to remove the H₂O and CO₂.

The regeneration performance of adsorbents was also investigated. The deactivated adsorbents were calcined at 500 °C in the air atmosphere for 3 hours and then cleaned with deionized water for 3 times to remove the deposited sulfate. After that, 0.1 g mL⁻¹ K₂CO₃ solution was used for reimpregnation to obtain the regenerated adsorbent.

3.3 Desulfurization performance tests

The cracked C₅ distillate was used as raw material, and 0.12 g and 0.005 g of CS₂ was added to 100 mL of cracked C₅ distillate, respectively. The mixture was stirred until CS₂ was completely dissolved. The sulfur contents in the as-prepared C₅ oil were 1200 μg mL⁻¹ and 50 μg mL⁻¹, respectively. 0.001 g, 0.01 g, and 0.1 g of water were added to 100 mL of prepared C₅ distillate, respectively. The final water content of the C₅ oil is 10 μg mL⁻¹, 100 μg mL⁻¹, and 1000 μg mL⁻¹, respectively.

The static adsorption method was applied to investigate the adsorption performance of CS₂. The experimental conditions were as follows: the adsorption temperature was 30 °C, and 5 g of the modified desulfurization adsorbent was added to 50 mL of C₅. The adsorption performance is determined by evaluating the sulfur content of the C₅ distillate after adsorption.

The dynamic desulfurization experiment was carried out in a fixed bed reactor at a pressure of 0.4 MPa. The volume of the reaction tube was 30 mL. The sulfur-containing C₅ distillate oil is passed through a micro-injection pump from the bottom to the top through a fixed bed with 21 g of modified adsorbent. The effects of temperatures (5, 30 and, 55 °C), different sulfur contents of raw material (50, 75, and 100 μg mL⁻¹), and different liquid hourly space velocities (LHSV) (0.5, 1.0, and 1.5 h⁻¹) on desulfurization performance of the adsorbent were investigated. During the experiment, C₅ was collected at the fixed bed outlet at intervals, and the sulfur content was analyzed. When the sulfur content of the C₅ was higher than 5 μg mL⁻¹, it was considered that the adsorbent was exhausted, and the experiment was stopped. The calculation formula of the adsorbent breakthrough sulfur capacity S_C (wt%) is shown in eqn (1),

where V is the inlet flow rate (mL h⁻¹) of C₅ distillate, S_in and S_out are the sulfur content (μg mL⁻¹) of C₅ at the inlet and outlet of the reaction tube, respectively. And t is the breakthrough time (h), and m is the mass of the adsorbent (g).

Total sulfur content in the liquid of C₅ before and after the reaction was analyzed using an RPP-2000S UV fluorescence sulfur analyzer.

3.4 Characterization of catalysts

The Brunauer–Emmett–Teller (BET) surface area and total pore volume of the adsorbent were measured using an N₂ adsorption–desorption apparatus (TriStar II 3020). The total pore volume was calculated by the amount of N₂ adsorbed at a relative pressure (P/P₀) = 0.99. The total area was calculated using a multipoint BET surface area calculation method. The morphology of the surface of the adsorbent particles was observed by a field emission scanning electron microscope (SEM, Zeiss Sigma 500). The surface elemental analysis of the adsorbents was measured by an energy spectrometer (EDS, Bruker XFlash 6/30). Transmission electron microscopy image (TEM) was measured using F-20 microscopy (FEI, USA). CO₂-TPD was characterized using a Micro Chemitics Instrument's Auto Chem II Model 2920 Multi-Function Adsorber. The catalyst was first purged in helium at a temperature of 400 °C for 1 h, cooled to room temperature, and saturated with pure CO₂ at a flow rate of 50 mL min⁻¹ for 30 min. After that, the TPD experiment was started, the heating rate was 10 °C min⁻¹, the helium flow rate was 50 mL min⁻¹, and the temperature was raised to 800 °C. XRD was characterized by a Panaco Sharp Xpert Pro MPD instrument for fresh K₂CO₃/Al₂O₃ adsorbents and exhausted adsorbents. The test conditions were: Cu Kα ray, voltage 36 kV, current 30 mA, scanning range 10–80°, step size 0.02° s⁻¹. The FT-IR was tested using a Nicoleti S50 infrared spectrometer from PerkinElmer. The sample was mixed and milled with KBr before testing, and the infrared wavelength ranged from 4000 to 400 cm⁻¹. Ionic chromatography (IC) was tested by Thermo Dionex ICS-1100. Samples were prepared using 0.10 N NaOH to extract sulfate and sulfite from the spent catalysts using sonication.

4 Conclusions

In this paper, we prepared K₂CO₃/Al₂O₃ adsorbents to remove CS₂ from C₅ distillate. When the K₂CO₃ loading concentration is 0.3 g mL⁻¹, the desulfurization capacity is the best, and the C₅ fraction oil with a sulfur content of 1200 μg mL⁻¹ can be desulfurized to below 10 μg mL⁻¹. The use of the adsorbent in a fixed bed reactor also exhibits good adsorptive desulfurization capabilities. When the sulfur content is 50 μg mL⁻¹, the space velocity is 1 h⁻¹, and the temperature is 30 °C, the sulfur content can be reduced to 5 μg mL⁻¹ or less and maintained for 108 hours. Under the optimized operating conditions, the breakthrough sulfur capacity is 0.76 wt%. The second and third regeneration capacities of the adsorbent were 74 wt% and 71 wt%, respectively. Based on the evidence of surface chemical characterization data, the removal mechanism of CS₂ is proposed. The process of desulfurization of C₅ distillate oil includes adsorption, hydrolysis, and oxidation. The adsorption is related to the pore structure of the adsorbents and the active components. The hydrolysis and oxidation processes are consistent with the removal of CS₂ in the Claus plant. Further work would be to explore the regeneration mechanism of the deactivated adsorbent.

Author contributions

Xiance Zhang: methodology, investigation, formal analysis, data curation, writing-original draft, Guanglin Zhou: conceptualization, writing-review & editing, project administration, resources, supervision, Mengying Wang: formal analysis, writing-review & editing, Xiaosheng Wang: writing-review & editing, resources, Weili Jiang: writing-review & editing, Hongjun Zhou: supervision.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We are grateful to the financial supports from the Science Foundation of the China University of Petroleum, Beijing (No. 2462018YJRC028), China Postdoctoral Science Foundation (2019M660931).

References

1. J. Y. Hou, F. Guo, Q. Hu, Y. Li and Z. M. Hou, Chin. J. Polym. Sci., 2019, 37, 674–680.
2. H. C. Hsu, S. J. Wang, J. D. Y. Ou and D. S. H. Wong, Ind. Eng. Chem. Res., 2015, 54, 9798–9804.
3. H. Miki, J. Jpn. Pet. Inst., 2019, 62, 245–254.
4. D. Salari, A. Niaei, J. Towfighi and P. Panahi, Iran. J. Chem. Chem. Eng., 2006, 2, 40–51.
5. S. Brunet, D. Mey, G. Pérot, C. Bouchy and F. Diehl, Appl. Catal., A, 2005, 278, 143–172.
6. W. Jiang, X. Gao, L. Dong, J. Xiao, L. H. Zhu, G. Y. Chen, S. H. Xun, C. Peng, W. S. Zhu and H. M. Li, Pet. Sci., 2020, 17, 1422–1431.
7. Y. Liu, Y. Pan, H. Wang, Y. Liu and C. Liu, Chin. J. Catal., 2018, 39, 1543–1551.
8. M. A. Larrubia, A. d. Gutièrrez-Alejandre, J. Ramìrez and G. Busca, Appl. Catal., A, 2002, 224, 167–178.
9. C. Li, Z. Jiang, Y. Zhang, Y. Yang, M. Yang and L. Wang, Chinese Pat., CN101450303A, 2009.
10. H. Yu, J. Nan, J. Zhang, Y. Zhang, X. Qu, S. Geng and Y. Shi, Chinese Pat., CN103182291A, 2013.
11. K. Karan and L. A. Behie, Ind. Eng. Chem. Res., 2004, 43, 3304–3313.
12. C. Deng, X. P. Wu, X. M. Sun, Y. Ren and Y. H. Sheng, J. Comput. Chem., 2009, 30, 285–294.
13. P. D. Clark, N. I. Dowling and M. Huang, Appl. Catal., B, 2001, 31, 107–112.
14. F. I. Khalili, M. Sultan, C. Robl and M. A. Al-Ghouti, J. Ind. Eng. Chem., 2015, 28, 282–293.
15. M. Song, X. Yang and G. Wang, RSC Adv., 2018, 8, 35014–35022.
16. H. Yi, D. He, X. Tang, H. Wang, S. Zhao and K. Li, Fuel, 2012, 97, 337–343.
17. N. Pechler and G. Emig, Gas Sep. Purif., 1991, 5, 247–251.
18. A. Primavera, A. Trovarelli, P. Andreussi and G. Dolcetti, Appl. Catal., A, 1998, 173, 185–192.
19. L. Wang, D. Wu, S. Wang and Q. Yuan, J. Environ. Sci., 2008, 20, 436–440.
20. S. Naumov, R. Valiullin and J. Kärger, Diffus. Fundam., 2007, 6, 67.61–67.62.
21. A. E. Aksoylu, A. N. Akin, S. G. Sunol and Z. İ. Önsan, Turk. J. Chem., 1996, 20, 88–94.
22. X. Song, K. Li, C. Wang, X. Sun, P. Ning and L. Tang, Chem. Eng. J., 2017, 330, 727–735.
23. Q. Li, H. Yi, X. Tang, S. Zhao, B. Zhao, D. Liu and F. Gao, Chem. Eng. J., 2016, 284, 103–111.
24. J. West, B. P. Williams, N. Young, C. Rhodes and G. J. Hutchings, Catal. Lett., 2001, 74, 111–114.
25. S. C. Lee, Y. M. Kwon, H. J. Chae, S. Y. Jung, J. B. Lee, C. K. Ryu, C. K. Yi and J. C. Kim, Fuel, 2013, 104, 882–885.
26. H. Liu, L. Su, F. Liu, C. Li and U. U. Solomon, Appl. Catal., B, 2011, 106, 550–558.
27. C. Murugan, H. C. Bajaj and R. V. Jasra, Catal. Lett., 2010, 137, 224–231.
28. S. U. Rege and R. T. Yang, Chem. Eng. Sci., 2001, 56, 3781–3796.
29. L. Fernández-Carrasco and E. Vázquez, Fuel, 2009, 88, 1533–1538.
30. A. Boumaza, L. Favaro, J. Lédion, G. Sattonnay, J. B. Brubach, P. Berthet, A. M. Huntz, P. Roy and R. Tétot, J. Solid State Chem., 2009, 182, 1171–1176.
31. K. Li, G. Liu, T. Y. Gao, F. Lu, L. Tang, S. Liu and P. Ning, Appl. Catal., A, 2016, 527, 171–181.
32. M. Liang, C. Li and H. Guo, J. Fuel Chem. Technol., 2002, 30, 347–352.
33. E. He, G. Huang, H. Fan, C. Yang, H. Wang, Z. Tian, L. Wang and Y. Zhao, Fuel, 2019, 246, 277–284.
34. G. Zhou, Y. Fu and H. Zhou, Petrochemical Technol., 2001, 30, 602–604.
35. R. Sui, C. B. Lavery, D. Li, C. E. Deering, N. Chou, N. I. Dowling and R. A. Marriott, Appl. Catal., B, 2019, 241, 217–226.
36. Y. Yue, X. Zhao, W. Hua and Z. Gao, Appl. Catal., B, 2003, 46, 561–572.
37. H. Yi, S. Zhao, X. Tang, C. Song, F. Gao, z. Wang, B. Zhang and Y. Zuo, Catal. Commun., 2014, 56, 106–109.
38. X. Song, P. Ning, C. Wang, K. Li, L. Tang, X. Sun and H. Ruan, Chem. Eng. J., 2017, 314, 418–433.

---