Deep adsorption desulfurization of liquid petroleum gas by copper-modified bentonite

Abstract

The removal of sulfur compounds from liquid petroleum gas (LPG) was investigated using a fixed-bed flow sorption system. Copper-modified bentonite adsorbents significantly enhanced the desulfurization of LPG. Several factors that influence desulfurization, including the copper loading, the baking temperature, the valence state and the type of anion used, were investigated. The adsorbents were characterized by nitrogen adsorption, X-ray diffraction, inductively coupled plasma-atomic emission spectrometry, thermogravimetric analysis, Raman spectrometry and FTIR spectrometry. Optimum desulfurization with Cu(II)-modified bentonite adsorbents was obtained at a loading of 15 wt% Cu²⁺ and a calcination temperature of 150 °C. The Cu(I)-modified bentonite adsorbents were shown to be better than the Cu(II)-modified bentonite adsorbents in removing sulfur compounds from LPG; the anion used had no significant influence on the desulfurization ability of the Cu(II)-modified bentonite adsorbents. FTIR analyses showed that the surface Lewis acid sites contributed to the desulfurization process. The sulfur compounds were adsorbed over Cu(I)- and Cu(II)-modified bentonite by a direct sulfur–adsorbent interaction.

---

1. Introduction

Liquid petroleum gas (LPG) is a major product of the distillation of crude oil and mainly consists of light hydrocarbon compounds. LPG is the main household fuel in many countries, especially China, and is also an important petrochemical feedstock. The isobutylene from LPG is mainly used for the production of methyl tertiary butyl ether (MTBE), which is used as a high-quality, high-octane number additive to gasoline. The physical properties of MTBE and gasoline hydrocarbons are similar and therefore MTBE shows good miscibility with gasoline without phase separation. MTBE not only improves the octane number of gasoline, but also improves this fuel by indirectly reducing the amount of sulfur, alkenes and aromatic compounds in gasoline and by increasing the vapor pressure. The use of the C4 hydrocarbons obtained from LPG increases economic benefits and provides additional value. However, the LPG obtained by fluid catalytic cracking or the delayed coking process usually contains sulfur compounds, including carbonyl sulfide, mercaptan, carbon disulfide, dimethyl sulfide (DMS), and dimethyl disulfide (DMDS).^1,2 Among these organic sulfur compounds, DMS and DMDS are the most difficult to remove because of their poor reactivity and polarity.³ Therefore it is necessary to remove these sulfur compounds by a pretreatment process before the synthesis of MTBE because their presence will result in the products having a high sulfur content with the potential to poison catalysts.

The US Environmental Protection Agency has issued regulations that the total sulfur content of gasoline must be <10 ppmw.² Similar regulations were implemented in some cities in China in 2012, including Beijing and Shanghai.⁴ Sulfur-free fuel is now a goal in many countries and various desulfurization processes – for example, hydrodesulfurization (HDS), selective catalytic oxidation, and adsorption desulfurization (ADS) – are being used to remove sulfur compounds from commercial fuels. HDS processes are the most extensively applied method to reduce sulfur levels in commercial fuels; however, HDS is not suitable for the desulfurization of LPG for the following two reasons: (1) HDS processes use Co–Mo/Al₂O₃, Ni–Mo/Al₂O or Ni–W/Al₂O₃ catalysts and operate at high temperatures (300–400 °C) and pressures (20–100 atm H₂);^5,6 and (2) HDS processes do not meet the current requirements for levels of sulfur.⁷ In addition, during the desulfurization process, the loss of alkenes may be significant as a result of the conversion of paraffin.^8,9 The selective catalytic oxidation process also leads to a loss of alkenes as a result of oxidation and polymerization. Therefore alternative processes to remove sulfur without a significant decrease in the octane index are required.¹⁰ In contrast, the ADS process has the advantages that it does not require the addition of hydrogen and it operates under ambient conditions. It also has the advantages of flexibility, low energy requirements, a high degree of safety and low operating costs.^11,12

ADS is regarded as one of the most competitive methods for the desulfurization of fuels, especially ultra-deep desulfurization. Deep desulfurization on various adsorbents (e.g. activated carbon,^2,13–15 modified composite oxides,¹⁶ zeolites,^1,17,18 mesoporous silica,^19,20 and metal–organic frameworks^21,22) has been studied. Yang and co-workers^23–25 reported the ADS of transportation fuels over Ag⁺ and Cu⁺ ion-exchange Y zeolites. They attributed the good desulfurization performance of the adsorbents to π-complexation. Velu et al.²⁶ reported that sulfur compounds can be selectively removed through direct sulfur–adsorbent interactions. Although considerable research has been devoted to the desulfurization of liquid transportation fuels, few studies have reported the deep desulfurization of LPG.

Takatsu et al.¹⁷ have reported the removal of DMDS, DMS, carbonyl sulfide and TBM from LPG. They used activated carbon, ZnO/Al₂O₃, Ag-exchanged β-zeolite, and Ag/CeO₂. Among the tested adsorbents, the Ag-exchanged β-zeolite showed the best adsorption capacity for sulfur compounds such as DMDS, DMS, and TBM. Takatsu et al.¹⁷ deduced that the electrostatic attraction between the Ag-exchanged β-zeolite and sulfur compounds played a critical role in the adsorption of sulfur compounds by comparing the charges of the sulfur atom in each molecule.

Bentonite is increasingly attracting attention as a new type of mesoporous material for use as a separating agent or sorbent as a result of its excellent physical and chemical properties, including a high specific surface area, low cost, ordered structure, thermal stability, high exchange capacity, and adsorptive affinity for organic and inorganic ions.²⁷ The main objective of the research reported here was to study the efficiency of modified bentonite in the removal of various organic sulfur compounds from LPG using a dynamic adsorption method in a fixed bed. The adsorbents were characterized by nitrogen adsorption, X-ray diffraction (XRD), inductively coupled plasma-atomic emission spectrometry (ICP-AES), thermogravimetric analysis (TGA), Raman spectrometry, and FTIR spectrometry.

2. Experimental section

2.1. Materials and feedstocks

Raw activated bentonite was obtained from Zhejiang Province, China; the composition of the bentonite is given in Table 1. The LPG was obtained from the delayed coking process carried out at the Quanzhou oil refinery (Fujian Province, China); the composition of the LPG is given in Table 2.

Table 1 Composition of activated bentonite used in this study

Component	Content (wt%)
SiO2	67.8
Al2O3	16.3
Na2O	0.9
Fe2O3	4.0
CaO	1.1
K2O	1.6
MgO	2.1
Other	7.0

---

Table 2 Composition of LPG used in this study

Component	Content (wt%)	Concentration of sulfur (mg m^−3)
Alkanes	77.5
n-Butane	24.1
iso-Butane	14.2
Propane	39.2
Alkenes	14.2
Propene	3.9
Butene	10.3
Other	8.3
Total	100
Sulfur compounds		29.94
Carbonyl sulfide		7.32
Carbon disulfide		0.67
Dimethyl disulfide		7.45
Methanethiol		10.66
Ethanethiol		3.42
Hydrogen sulfide		0.42

---

2.2. Preparation of adsorbents

Activated bentonite and various concentrations of Cu(NO₃)₂ were thoroughly mixed in a kneader. A moderate amount of deionized water was then added to the mixture to give a slurry. After stirring for a period of time, the raw materials were passed through an extruder to give a series of strips of modified bentonite. The extrudate was then dried in an oven at 120 °C for 4 h. Finally, the extrudate was baked in air in a muffle furnace at 150 °C for 2 h, then crushed and screened to a size of 20–40 mesh. CuBr₂-bentonite and CuSO₄-bentonite were prepared using the same method. CuCl-bentonite was also prepared by the same method, but in this instance the final step was to bake the CuCl-bentonite sample at 150 °C for 2 h in pure N₂.

2.3. Adsorption desulfurization experiments

Evaluation of the ADS performance of the adsorbents was carried out in a fixed-bed flow sorption system with a stainless-steel column (8 mm i.d. × 400 mm long). The adsorbent particles were packed into the middle of the column and the remaining space was filled with quartz sand. The experiments were conducted at room temperature under 0.6 MPa pressure and a liquid hourly space velocity of 10.0 h⁻¹. The concentration of sulfur compounds was determined hourly after the start of each test at the inlet and outlet points. The total concentration of sulfur compounds in the outflow was determined by an ultraviolet fluorescence sulfur nitrogen analyzer. The breakthrough sulfur content was defined as a sulfur concentration in the effluent of 5 mg m⁻³. The desulfurization capacity was calculated by the following equation:
where S_t is the amount of sulfur adsorbed on the adsorbent (wt% mg S g⁻¹ of adsorbent), q is the flow velocity of the LPG (ml h⁻¹), C₀ is the inlet sulfur concentration of the LPG (mg m⁻³) and C_t is the outlet sulfur concentration of the LPG (mg m⁻³) at any time t (h), and m is the mass of the adsorbent (g).

2.4. Characterization of adsorbents

2.4.1. Powder XRD. The crystalline structure of the adsorbents was confirmed by powder XRD using a Siemens D-500 X-ray diffractometer equipped with Ni-filtered Cu K_α radiation (40 kV tube voltage, 100 mA tube current and λ = 0.154 nm). The range of the 2θ scanning angle was 5–80° in steps of 0.02° s⁻¹. The size of crystallites in the samples was calculated from the XRD patterns using the Scherrer formula:
where d is the crystallite size, λ is the wavelength of the X-rays and β is the width of the diffraction peak at half-maximum.

2.4.2. Nitrogen isotherms. The BET specific surface area of bentonite was measured by nitrogen adsorption at −196 °C using a Micromeritics JW-BK112 instrument. Prior to analysis, the samples were degassed for 2 h at 120 °C under a vacuum of P < 10⁻² Pa at a constant pressure. The specific BET surface areas were determined from N₂ adsorption at relative pressures of 0.05 < P/P₀ < 0.35. The pore size distribution (PSD) was obtained from the adsorption branch of the N₂ sorption isotherms using the Barret–Joyner–Halenda (BJH) formula.²⁸ The total pore volume (V_t) was estimated from the volume of N₂ (as liquid) held at a relative pressure (P/P₀) of 0.99.

2.4.3. ICP-AES. The actual copper loading in the Cu(NO₃)₂-bentonite adsorbents was determined by ICP-AES using an Agilent 725 ES instrument.

2.4.4. Thermal analysis. Thermogravimetric curves were recorded using a TA Instruments thermal analyzer. The samples were ground to 200 mesh and the curves were determined at atmospheric pressure at a heating rate of 10 °C min⁻¹ in the temperature range 25–700 °C.

2.4.5. FTIR spectrometry. The type and number of the surface acidity sites on the adsorbents were analyzed by Fourier transform infrared (FTIR) spectrometry using pyridine as the probe molecule. The sample loaded in the in-site cell was pretreated at 380 °C under vacuum conditions to remove any moisture. An excess of pyridine was then adsorbed after cooling to 80 °C. The adsorbed pyridine was then desorbed at 200 and 450 °C for 2 h.

2.4.6. Raman spectrometry. The Raman studies were conducted using a Renishaw System 100 Raman spectrometer with 514 nm red excitation from an Ar laser. The laser power was 3 mW at the sample position. The Raman scattered light was detected perpendicular to the laser beam with a Peltier-cooled CCD detector and the spectral resolution in all measurements was 1 cm⁻¹.

3. Results and discussion

3.1. XRD characterization of the adsorbents

Fig. 1 and 2 show the results of the XRD analyses carried out to identify the mineralogical structure of the modified bentonite adsorbents. The XRD pattern of the adsorbents showed characteristic reflections of SiO₂ at 2θ = 20.81 and 26.61°, suggesting that the structure of the bentonite remained intact after modification. However, the crystallinity of the modified bentonite adsorbents was slightly decreased compared with raw bentonite as a result of framework defects caused by the loading of the metal ions.

Fig. 1 XRD patterns of Cu-bentonite with different copper compounds: (a) raw bentonite; (b) CuBr₂-bentonite; (c) CuSO₄-bentonite; (d) CuCl₂-bentonite; and (e) CuCl-bentonite.

Fig. 2 XRD patterns of Cu-bentonite with different copper contents: (a) raw bentonite; (b) 5 wt% copper; (c) 10 wt% copper; (d) 15 wt% copper; and (e) 20 wt% copper.

Fig. 1 shows that the XRD patterns of the CuBr₂-bentonite samples examined at large angles showed the characteristic reflections of CuBr₂ at 2θ = 27.06, 44.97, and 53.27° corresponding to the (002), (110), and (112) planes, respectively, of cubic CuBr₂. The characteristic reflections of the CuSO₄·5H₂O species at 2θ = 16.09, 18.70, 22.17, and 23.93° corresponding to the (110), (011), (101), and (111) planes, respectively, indicate the cubic structure of CuSO₄·5H₂O. The XRD patterns of the CuCl₂-bentonite samples showed characteristic reflections of CuCl₂ at 2θ = 16.27, 22.96, and 40.89°, corresponding to the (101), (110), and (011) planes, respectively, indicative of a cubic phase. For the CuCl-bentonite samples, the Cu₂O diffraction peak at 2θ = 16.14° can be indexed to the (111) reflection of the cubic Cu₂O species. The mean particle sizes of the loaded CuBr₂, CuSO₄·5H₂O, CuCl₂ and Cu₂O particles calculated by the Scherrer equation were about 60, 59, 110, and 38 nm, corresponding to the strongest peaks of the (002), (011), (101), and (111) planes, respectively.

Fig. 2 shows the XRD patterns of the Cu(NO₃)₂-bentonite samples with the characteristic reflections of Cu₂(OH)₃NO₃ at 2θ = 12.74 and 25.76°, corresponding to the (001) and (002) planes of cubic Cu₂(OH)₃NO₃, respectively; this is attributed to the incomplete decomposition of trace amounts of Cu(NO₃)₂.¹⁸ The weak peak of CuO observed at a 2θ value of 33.52° can be indexed to the (211) reflection of cubic CuO resulting from the partial oxidation of Cu(NO₃)₂. The average crystallite sizes of Cu₂(OH)₃NO₃ and CuO were calculated to be 130 and 41 nm, respectively, using the Scherrer equation, which corresponds to the strongest peak of the (001) and (211) planes. The Cu₂(OH)₃NO₃ peaks of the modified bentonite adsorbents increased gradually as the copper loading increased.

3.2. Nitrogen isotherms

Fig. 3 and 4 show the N₂ adsorption–desorption isotherms and PSD curves of raw bentonite and Cu(NO₃)₂-bentonite with different copper loadings. All the samples exhibited type IV isotherms with a capillary condensation step and a strong peak on the PSD curves, characteristic of mesoporous materials according to the IUPAC classification.²⁹

Fig. 3 (a) N₂ adsorption–desorption isotherms and (b) pore size distribution curves of raw bentonite and Cu(NO₃)₂-bentonite with different copper loadings.

Fig. 4 (a) N₂ adsorption–desorption isotherms and (b) pore size distribution curves of raw bentonite and bentonite with different copper compounds.

Table 3 gives the structural parameters of the adsorbents obtained from the nitrogen adsorption isotherms. The BET surface area, pore size, and total pore volume decreased with increasing loading of copper ions for Cu(NO₃)₂-bentonite adsorbents. The raw bentonite adsorbent had the largest BET surface area, pore size, and total pore volume. The decrease in the BET surface area, pore size and total pore volume may be due to the deposition of metal clusters onto the surface of the bentonite channels.³⁰ The modified bentonite adsorbents did not have the same BET surface area and total pore volume, even though the same metal ion was loaded on the bentonite. The BET surface area was in the order: Cu(NO₃)₂-bentonite > CuCl-bentonite > CuBr₂-bentonite > CuCl₂-bentonite > CuSO₄-bentonite. The total pore volume was in the order: Cu(NO₃)₂-bentonite > CuCl-bentonite > CuCl₂-bentonite > CuBr₂-bentonite > CuSO₄-bentonite.

Table 3 Structural parameters of the adsorbents from nitrogen adsorption isotherms

Sample	Cu content (wt%)	BET surface area (m^2 g^−1)	Vtotal (cm^3 g^−1)	BJH pore size (nm)
Raw bentonite	0	179.212	0.294	8.615
5 wt% Cu(NO3)2-bentonite	4.6	138.681	0.262	6.584
10 wt% Cu(NO3)2-bentonite	9.6	127.106	0.243	6.341
15 wt% Cu(NO3)2-bentonite	14.5	107.754	0.223	5.915
20 wt% Cu(NO3)2-bentonite	19.2	108.037	0.232	5.854
15 wt% CuCl2-bentonite	14.1	77.037	0.203	7.762
15 wt% CuSO4-bentonite	14.5	70.526	0.165	7.212
15 wt% CuBr2-bentonite	14.6	93.987	0.192	6.446
15 wt% CuCl-bentonite	14.6	93.995	0.213	7.546

---

3.3. Effects of different copper compounds on the ADS of LPG

To examine the effect of different copper compounds on the ADS of LPG, four modified bentonite adsorbents were prepared using the mixing method. Fig. 5 shows that, apart from the CuCl₂-bentonite adsorbent, all the adsorbents had a very similar ability to remove sulfur and therefore the anion had no significant influence on the ability of the Cu(II)-modified bentonite adsorbents to desulfurize LPG. Combining these results with those in Table 3, the order of activity of the adsorbents was not related to the BET surface area nor the total pore volume. These results indicated that Cu²⁺ played the most critical role in the ADS of LPG rather than the anion. As a result of this similarity in activity between the adsorbents, copper nitrate was chosen for the subsequent investigations.

Fig. 5 Breakthrough curves and sulfur capacities of ADS at different loadings of copper compounds on bentonite.

3.4. Effects of Cu²⁺ loading on desulfurization performance

Dynamic tests were carried out to examine the effect of Cu²⁺ loading on the desulfurization performance. Four Cu(NO₃)₂-bentonite adsorbents with different copper loadings of 5, 10, 15, and 20 wt% were prepared using the mixed method. As shown in Fig. 6, compared with the raw bentonite adsorbent, the modified adsorbents were able to remove sulfur for longer time periods. In addition, at low Cu(NO₃)₂ concentrations, the amount of sulfur removed by the adsorbents increased with increasing Cu(NO₃)₂ concentration. The removal of sulfur was at an optimum level when the Cu(NO₃)₂ loading was 15 wt%. However, in the case of the 20 wt% Cu(NO₃)₂-bentonite sample, the amount of sulfur removed was obviously decreased. Table 3 shows that, compared with the raw bentonite adsorbent, the total pore volume and the BET surface area of the modified adsorbents were decreased. However, the performance of the modified adsorbents was still higher than that of raw bentonite. This result can be explained by the chemical interaction between the sulfur compounds and the Cu(NO₃)₂ content of the adsorbents. However, the amount of sulfur removed decreased when the Cu(NO₃)₂ concentration was above a certain threshold because the metal clusters were able to block the pore channels of the adsorbents and thus access to the active sites. This indicates that there are some trade-off effects between the Cu(NO₃)₂ loading and the amount of sulfur that can be removed. Therefore the amount of sulfur removed can be controlled by choosing an appropriate concentration of Cu(NO₃)₂.

Fig. 6 Breakthrough curves and sulfur capacities of ADS over Cu(NO₃)₂-bentonite sorbents with different copper loadings.

3.5. Effects of baking temperature on adsorption performance

The effect of different baking temperatures from 120 to 350 °C were measured using the 15 wt% Cu(NO₃)₂-bentonite adsorbent. Fig. 7 shows that the baking temperature had a strong effect on the ADS of LPG and that the ability of the adsorbents to remove sulfur was gradually reduced between 150 and 350 °C. The optimum baking temperature for the 15 wt% Cu(NO₃)₂-bentonite adsorbent was 150 °C.

Fig. 7 Breakthrough curves and sulfur capacities of ADS at different baking temperatures for the 15 wt% Cu(NO₃)₂-bentonite.

Thermogravimetric analysis was carried out to determine the thermal stability of the adsorbents. Fig. 8 shows the differential scanning calorimetry (DSC)–TGA curves for raw bentonite and the 15 wt% Cu(NO₃)₂-bentonite adsorbent. The peak in the range 25–150 °C on the TGA curve of the raw adsorbent represents the water adsorbed in the bentonite. This peak is also seen on the TGA curve of 15 wt% Cu(NO₃)₂-bentonite and the second peak on this curve at 200–300 °C is a result of the decomposition of Cu(NO₃)₂.¹⁸ This explains why the ADS performance of the absorbent baked at 150 °C was better than that baked at 120 °C; the water adsorbed in the bentonite blocks the pores of the adsorbent.

Fig. 8 DSC–TGA curves in air for raw bentonite and 15 wt% Cu(NO₃)₂-bentonite.

Powder XRD was carried out to determine the decomposition products of Cu(NO₃)₂ at different baking temperatures. Fig. 9 shows that the decomposition products of Cu(NO₃)₂ were Cu₂(OH)₃NO₃ and CuO at baking temperatures of 120 and 150 °C. However, the decomposition product of Cu(NO₃)₂ at baking temperatures of 250 and 350 °C was CuO. Combined with Fig. 6, which shows that the ADS performance of the absorbent baked at 150 °C was better than that baked at 250 and 350 °C; these results indicated that Cu₂(OH)₃NO₃ plays a more dominant role than CuO in the removal of sulfur compounds from LPG.

Fig. 9 XRD patterns of 15 wt% Cu(NO₃)₂-bentonite at different baking temperatures: (a) 120; (b) 150; (c) 250; and (d) 350 °C.

3.6. Effect of Cu(II) and Cu(I) on the ADS of LPG

We also investigated the effect of the valency of copper on the ADS of LPG. Fig. 10 shows that Cu(I)-bentonite had a greater ability to remove sulfur than Cu(II)-bentonite.

Fig. 10 Breakthrough curves and sulfur capacities of ADS over Cu(II)- and Cu(I)-modified bentonite adsorbents.

The surface acidic sites of the Cu(II) and Cu(I)-bentonite adsorbents were investigated by the adsorption of pyridine at 200 and 450 °C. The FTIR spectra in Fig. 11 and 12 show the presence of both Lewis and Brønsted acidity in these samples. The band at 1540 cm⁻¹ was assigned to the Brønsted acid sites, whereas the band at 1450 cm⁻¹ was assigned to the Lewis acid sites. Fig. 11 shows the characteristic peaks at 1450 cm⁻¹ arising from the vibrations of pyridine molecules adsorbed at Lewis acid sites on the modified bentonite adsorbents. It is evident that the Cu(I)- and Cu(II)-bentonite adsorbents show a higher Lewis acidity than the raw bentonite adsorbent because the metal cations act as electron acceptors. The total number of Lewis acid sites on the surface of the Cu(I)-bentonite was greater than on the Cu(II)-bentonite. Fig. 12 shows that strong Lewis acid sites were only present on the Cu(I)-bentonite adsorbent. This higher Lewis acidity facilitated the adsorption of sulfur compounds (electron donors) on the Cu(I)-bentonite. We concluded that the Lewis acid sites on the surface of the adsorbents facilitated the adsorption of sulfur compounds from LPG.

Fig. 11 Pyridine FTIR spectra for three adsorbents at 200 °C: (a) raw bentonite; (b) Cu(I)-bentonite; and (c) Cu(II)-bentonite.

Fig. 12 Pyridine FTIR spectra for three adsorbents at 450 °C: (a) raw bentonite; (b) Cu(I)-bentonite; and (c) Cu(II)-bentonite.

Yi et al.³¹ have reported that sulfur compounds can be removed through direct interaction between sulfur and the adsorbent. Fig. 13 shows the Raman spectra of the Cu(I)-bentonite and Cu(II)-bentonite adsorbents after the adsorption of sulfur compounds from LPG. Two peaks at 286 and 416 cm⁻¹ are observed on curve a, but only one peak at 280 cm⁻¹ on curve b. These regions are characteristic of the Cu–S stretching vibrations,³² indicating that the sulfur compounds are adsorbed over Cu(I)-bentonite and Cu(II)-bentonite by a direct sulfur–adsorbent interaction.

Fig. 13 Raman spectra of the adsorbents: (a) CuCl-bentonite after the adsorption of sulfur compounds from LPG; and (b) CuCl₂-bentonite after the adsorption of sulfur compounds from LPG.

4. Conclusions

Copper-modified bentonite has been shown to be an excellent adsorbent for the remove of sulfur compounds from LPG. The anion had no significant influence on the desulfurization ability of the Cu(II)-modified bentonite adsorbents. The bentonite loaded with 15 wt% copper and calcined at 150 °C showed the optimum ability to adsorb sulfur during the desulfurization of LPG containing about 29.4 mg m⁻³ sulfur. The Cu(I)-modified bentonite was shown to be better than the Cu(II)-modified bentonite in removing sulfur compounds from LPG. The contribution of the Lewis acid sites to the desulfurization process was determined by FTIR spectrometry. Raman spectrometry indicated that the sulfur compounds were adsorbed over Cu(I)-bentonite and Cu(II)-bentonite by a direct sulfur–adsorbent interaction.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (No. 21276086).

References

1. J. Lee, H. T. Beum, C. H. Ko, S. Y. Park, J. H. Park, J. N. Kim, B. H. Chun and S. H. Kim, Ind. Eng. Chem. Res., 2011, 50, 6382–6390.
2. K. S. Kim, S. H. Park, K. T. Park, B. H. Chun and S. H. Kim, Korean J. Chem. Eng., 2010, 27, 624–631.
3. D. Crespo, G. Qi, Y. Wang, F. H. Yang and R. T. Yang, Ind. Eng. Chem. Res., 2008, 47, 1238–1244.
4. D. Li, Chin. J. Catal., 2013, 34, 48–60.
5. A. J. Hernández-Maldonado and R. T. Yang, Catal. Rev., 2004, 46, 111–150.
6. A. J. Hernández-Maldonado and R. T. Yang, Ind. Eng. Chem. Res., 2004, 43, 1081–1089.
7. I. V. Babich and J. A. Moulijn, Fuel, 2003, 82, 607–631.
8. L. Wang, S. Li, H. Cai, Y. Xu, X. Wu and Y. Chen, Fuel, 2012, 94, 165–169.
9. Y. Wang, R. T. Yang and J. M. Heinzel, Chem. Eng. Sci., 2008, 63, 356–365.
10. C. Song, Catal. Today, 2003, 86, 211–263.
11. L. Duan, X. Gao, X. Meng, H. Zhang, Q. Wang, Y. Qin, X. Zhang and L. Song, J. Phys. Chem. C, 2012, 116, 25748–25756.
12. A. H. M. Shahadat Hussain and B. J. Tatarchuk, Fuel, 2012, 107, 465–473.
13. A. A. Yahia and S. B. Hisham, Fuel, 2009, 88, 87–94.
14. X. L. Tang, W. Qian, A. Hu, Y. M. Zhao, N. N. Fei and L. Shi, Ind. Eng. Chem. Res., 2011, 50, 9363–9367.
15. H. Cui, S. Q. Turn and M. A. Reese, Catal. Today, 2009, 4, 274–279.
16. C. Sentorun-Shalaby, S. K. Saha, X. L. Ma and C. S. Song, Appl. Catal., B, 2011, 101, 718–726.
17. K. Takatsu, G. Takegoshi, H. Katsuno, Y. Kawashima and H. Matsumoto, J. Jpn. Pet. Inst., 2007, 4, 200–207.
18. C. S. Le, T. B. Ting, J. Z. Li, L. Z. Yu, Z. X. Yu and H. X. Dao, J. Phys. Chem. C, 2014, 118, 9468–9476.
19. Y. Yin, D. M. Xue, X. Q. Liu, G. Xu, P. Ye, M. Y. Wu and L. B. Sun, Chem. Commun., 2012, 48, 9495–9497.
20. P. Tan, J. X. Qin, X. Q. Liu, X. Q. Yin and L. B. Sun, J. Mater. Chem. A, 2014, 2, 4698–4705.
21. Z. Hasan and S. H. Jhung, ACS Appl. Mater. Interfaces, 2015, 7, 10429–10435.
22. Y. X. Li, W. J. Jiang, P. Tan, X. Q. Liu, D. Y. Zhang and L. B. Sun, J. Phys. Chem. C, 2015, 119, 21969–21977.
23. R. T. Yang, A. J. Hernández-Maldonado and F. H. Yang, Science, 2003, 301, 79–81.
24. A. J. Hernández-Maldonado and R. T. Yang, J. Am. Chem. Soc., 2004, 126, 992–993.
25. A. J. Hernández-Maldonado and R. T. Yang, Ind. Eng. Chem. Res., 2003, 42, 123–129.
26. S. Velu, X. Ma and C. Song, Ind. Eng. Chem. Res., 2003, 42, 5293–5304.
27. X. L. Tang, X. Meng and L. Shi, Ind. Eng. Chem. Res., 2011, 50(12), 7527–7533.
28. E. P. Barrett, L. J. Joyner and P. P. Halenda, J. Am. Chem. Soc., 1951, 73, 373–380.
29. K. S. W. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierotti, J. Rouquérol and T. Siemieniewska, Pure Appl. Chem., 1985, 57, 603–619.
30. J. Kaur, K. Griffin, B. Harrison and I. V. Kozhevnikov, J. Catal., 2002, 208, 448–455.
31. D. Z. Yi, H. Huang, X. Meng and L. Shi, Appl. Catal., B, 2014, 148, 377–386.
32. R. A. Colin, Y. Hyeyeong, S. V. Joan, K. B. Gdran, B. Nicklas, V. P. Gertie, W. C. Gerard, M. L. Thomas and S. L. Joann, J. Am. Chem. Soc., 1994, 116, 11489–11498.

---