Preparation of M/γ-Al₂O₃ sorbents and their desulfurization performance in hydrocarbons

Abstract

M/γ-Al₂O₃ sorbents with different metals (Ag, Cu, Ni, Zn) as the active component loaded on a γ-Al₂O₃ support were prepared by the incipient wetness impregnation method, and their adsorption behavior for thiophene was investigated. The results show that all these metals can obviously promote the desulfurization activity of the prepared sorbents, and Ag is the best one. Then silver was selected to modify γ-Al₂O₃ with a different loading amount, and the desulfurization behavior of Ag/γ-Al₂O₃ series sorbents in a thiophene–benzene solution was evaluated. It was found that the silver content has a significant impact on desulfurization efficiency, and the A15 sorbent with 13.7 wt% silver has the best adsorption desulfurization performance. XRD results show that the simple Ag⁰ is the main active component in Ag/γ-Al₂O₃ sorbent. SEM/EDS and BET characterization show that the specific surface area and pore volume decrease obviously when the silver loading amount is more than 13.7 wt%, because of the agglomeration of silver. The desulfurization mechanism of the Ag/γ-Al₂O₃ sorbent was explored by using thiophene which has both a conjugated pi bond and sulfur, tetrahydrothiophene which has sulfur but no conjugated pi bond, benzene which has a conjugated pi bond but no sulfur, cyclohexane which has no conjugated pi bonds or sulfur as the model compounds. The desulfurization efficiencies of A15 sorbent in thiophene–benzene, thiophene–cyclohexane, thiophene–tetrahydrothiophene–benzene and thiophene–tetrahydrothiophene–cyclohexane solutions were compared. The results indicate that the thiophene adsorption on Ag/γ-Al₂O₃ sorbent is mainly dominated by two kinds of connection between thiophene and silver. One is the connection between the conjugated pi bond and silver (π-complexation), and the other one is the connection between sulfur and silver (S–metal bond). This is also the main reason that benzene has the competitive adsorption behavior on thiophene.

---

1. Introduction

Benzene is an important chemical feedstock, which is widely used for synthesizing polymers,¹ resins,² fibers,³ pharmaceuticals, pesticides, explosives, dyes and so on. But as its main source, coking benzene contains a certain content of sulfur-containing compounds, especially thiophene, which will have bad effects on further uses of benzene.⁴ Accordingly, attempts have been made for the purification of coking benzene and some methods to chemically remove thiophene impurities from benzene have been presented. Nowadays, the methods industrialized for benzene desulfurization purification mainly include sulfuric acid washing, catalytic hydrogenation and extractive rectification. However, the sulfuric acid refining method has the disadvantages of equipment corrosion and environmental pollution,⁵ the catalytic hydrogenation method consumes hydrogen and has high operating costs, the extractive rectification method consumes large amounts of energy and the desulfurization efficiency is not very high.⁶ In contrast, the selective adsorption method has the great potential with the advantages of smaller equipment corrosion, less environmental pollution, lower operating costs, and no hydrogen consumed.

However, lots of literature about selective adsorption method are all almost focused on the desulfurization in the transportation fuels,^7–15 only a few on the purification of benzene.^16–19 Some researchers reported that the transition metal-based sorbent is effective for selectively adsorbing the sulfur compounds in the liquid hydrocarbon fuels at ambient temperature under atmospheric pressure with low investment and operating cost.²⁰ Transition metal (Cu, Zn, Ni or Ag) could enhance adsorption properties for thiophene and toluene on some sorbents, such as Y-zeolites,^21,22 because Cu and Ag in sorbents are likely combine S-containing compounds by π-complexation,²³ Ni and Zn in the S-sorb desulfurization catalyst invented by Sinopec and Philips companies play a very important role on the desulfurization process.^24,25 Nevertheless, a common problem is that the aromatic compounds, e.g. benzene, have the competitive adsorption with thiophene,^8,26,27 because benzene has very similar physical and chemical properties as that of thiophene, and thus have an obvious competitive effect on thiophene adsorption. Therefore, most of these sorbents reported for desulfurization of liquid fuel could not remove enough thiophene from benzene to give the required low sulfur concentration because the adsorptive removal of thiophene from liquid fuel and benzene is quite different. For liquid fuel, its main components are saturated hydrocarbons, the competitive effect of which is much less and can be ignored, whereas for conking benzene, the main component is benzene. Even though there is the study about the competitive adsorption of thiophene and benzene,²⁸ it is only the removal of thiophene from fuel containing the similar level of benzene, in which the competitive adsorption is much smaller comparing with benzene solvent. Therefore, it is difficult to remove thiophene with the ppm level from benzene because of the competitive adsorption of aromatic compound with very similar physical and chemical properties. In order to improve the sulfur capacity and efficiency of sorbents removing thiophene from benzene, it is necessary to study their adsorption mechanism and optimize the technological parameters.

In this paper, γ-Al₂O₃ was selected as the support and different metal nitrates were chosen as the precursors of active components because of their high water solubility and the easy decomposition during heating. A series M/γ-Al₂O₃ sorbents were prepared by incipient wetness impregnation method. And then the removal of thiophene from benzene was investigated. In order to investigate the adsorption mechanism of thiophene on M/γ-Al₂O₃ sorbent, the adsorption behaviors of the aromatic hydrocarbon (benzene) and the unsaturated cyclic sulfur organic-compound (thiophene) and their saturated analogs (cyclohexane and tetrahydrothiophene, respectively) on M/γ-Al₂O₃ sorbent were compared.

2. Experimental

2.1. Preparation of sorbents

γ-Al₂O₃ powder (0.177–0.125 mm), a meso-pore material, which was purchased from Shandong Kunpeng New Materials Co. Ltd, was used as the support of sorbent. Before use, it was dried at 120 °C for 4 h to remove the water. Then, it was respectively incipient wetness impregnated in the aqueous solution of Cu(NO₃)₂, Ni(NO₃)₂, Zn(NO₃)₂ and AgNO₃ with a predicted volume for 4 h. The samples prepared were dried at 80 °C for 2 h, calcined in Muffle furnace at 500 °C for 4 h to obtain the experimental sorbents.

The actual silver loading amount of the prepared samples was quantified using the Volhard method. The samples were first dried overnight at 220 °C, and dissolved with excess HNO₃ and titrated with a standard thiocyanate solution with an Fe(III) indicator. And the amount of copper, nickel and zinc loaded on γ-Al₂O₃ was analyzed by atomic absorption spectrophotometer (AAS). The series sorbents prepared and their metal loaded amounts were shown in Table 1. A0 denotes the blank γ-Al₂O₃ support, which is the original γ-Al₂O₃ following the sorbent preparation procedure without adding any metal. A5, A10, A15 and A20 respectively denote the sorbent with different Ag content. C5, N5, Z5 denote the sorbent with Cu, Ni and Zn as the active component respectively.

Table 1 The metal content of series prepared sorbents

Sorbents	A0	A5	A10	A15	A20	C5	N5	Z5
Metal content (wt%)	0.0	4.6	9.5	13.8	17.6	4.7	4.4	4.5

---

2.2. Adsorption experiments

In order to choose the suitable active component for thiophene adsorption from benzene, 1.00 g of A0, A5, C5, N5 and Z5 sorbent was respectively soaked in 4.00 mL thiophene–benzene solution with the thiophene concentration of 559.8 mg L⁻¹ in an ampoule bottle at room temperature for 24 h, which was considered that it reached the adsorption quasi-equilibrium and also was proved by the data stated in this paper. After adsorption, the concentration of thiophene in the solution was determined by gas chromatography GC-950 (Shanghai Haixin chromatography Co., Ltd) coupled with flame photometric detector (FPD) and TCEP-7 column (2.5 m in length and 4 mm in diameter), and the column temperature, inlet temperature and detector temperature are 80 °C, 120 °C and 160 °C, respectively. The carrier gas is high purity nitrogen, and the column pressure is 0.1 Mpa. The flow rate of high purity hydrogen and air for FPD is 40 mL min⁻¹ and 50 mL min⁻¹, respectively.

In order to investigate the influence of active metal component on thiophene adsorption and the corresponding adsorption mechanism, the sorbents with different active component and different content were put in 4.00 mL different organic sulfur-containing solutions shown in Table 2 at room temperature for 24 h. The reason why these sulfur-containing solutions were used is: thiophene molecule has a conjugated pi bond and a sulfur atom, tetrahydrothiophene molecule has a sulfur atom but does not have a conjugated pi bond, benzene molecule has a conjugated pi bond but does not have a sulfur atom, cyclohexane has no conjugated pi bonds and sulfur, and the structure diagram of those four compounds was shown in Fig. 1. The desulfurization efficiency for the sulfur-containing compounds in different solutions on the selected sorbent was measured. The effect of the conjugated pi bond on adsorption can be explored by comparing the desulfurization efficiency in thiophene–benzene solution (TH–B) and thiophene–cyclohexane solution (TH–C). The effect of the sulfur atom on adsorption can be explored by comparing the desulfurization efficiency in thiophene–benzene solution (TH–B) and tetrahydrothiophene–benzene solution (THT–B). While, tetrahydrothiophene–thiophene–benzene solution (TH–THT–B) and tetrahydrothiophene–thiophene–cyclohexane solution (TH–THT–C) were used to further explore the adsorption mechanism.

Table 2 Solutions used in this study

Solutions	Sulfur-containing compounds	Solvent	Cthiophene (mg L^−1)	Ctetrahydrothiophene (mg L^−1)
TH–B	Thiophene	Benzene	559.8	0.0
THT–B	Tetrahydrothiophene	Benzene	0.0	592.1
TH–C	Thiophene	Cyclohexane	592.5	0.0
TH–THT–B	Thiophene/tetrahydrothiophene	Benzene	545.2	592.1
TH–THT–C	Thiophene/tetrahydrothiophene	Cyclohexane	543.9	577.6

---

Fig. 1 Structure diagram of benzene, thiophene, cyclohexane and tetrahydrothiophene.

The desulfurization efficiency (η, %) and thiophene adsorption capacity (Q, mg g⁻¹ sorbent) were calculated as eqn (1) and (2).

where, c₀ is the initial concentration of thiophene or tetrahydrothiophene in the solution (mg L⁻¹); c is the concentration of thiophene or tetrahydrothiophene in the solution (mg L⁻¹) when it reached the adsorption quasi-equilibrium.

The saturation adsorption capacities, measured by weight method, for benzene, thiophene and hydrotetrathiophene of A15 sorbent are 19.36 mg g⁻¹, 24.66 mg g⁻¹ and 36.51 mg g⁻¹. The best sorbent was put into 4.00 mL thiophene, tetrahydrothiophene and benzene with the analytically pure at room temperature for 24 h respectively, then filtrated and dried in a vacuum oven at 50 °C for 12 h, and the samples prepared were denoted by A15TH, A15THT and A15B respectively.

2.3. Characterization of sorbents

The structure of sorbents was characterized by X-ray diffraction (XRD) apparatus of D/max-rB made in Japan, at a scanning rate of 5° min⁻¹ from 30° to 80° with Cu-Kα radial, λ = 0.15046 nm X-ray, 40 kV tube voltage and 100 mA tube current. The surface area, pore volume and pore size distribution of the samples were measured by nitrogen adsorption isotherms using a Micromeritics/ASAP2000 apparatus according to the BET and BJH models.

FT-IR spectra of samples were recorded on an Equinox55 spectrometer (Bruker, Germany) in the wavenumber range of 400–4000 cm⁻¹, 16 scans were taken at a resolution of 4 cm⁻¹. To prepare pellets, 1 mg samples were first ground to powder in an agate mortar and then mixed with 100 mg KBr. A hydraulic press was used to press the resulting mixtures to discs of 10 mm in diameter at 10 MPa for 2 min.

3. Results and discussion

3.1. Selection of active component for desulfurization over prepared M/γ-Al₂O₃ sorbents

Fig. 2 shows the desulfurization efficiency and thiophene adsorption capacity of sorbents with different metal components in thiophene–benzene solution. It can be seen that A0 sorbent which is the blank sample, γ-Al₂O₃ support without metal loaded actually, can only adsorb a little thiophene (0.006 mg g⁻¹) from benzene with the desulfurization efficiency of 2.9%, whereas the desulfurization efficiency of other sorbents with metal loaded (C5, N5, Z5 and A5) is significantly higher. This indicates that the metals (Cu, Ni, Zn and Ag) loaded on γ-Al₂O₃ are the main active components, which play a very important role on thiophene adsorption, and the activity of silver is the highest. It was reported in the literature that after modifying γ-Al₂O₃ in Cu(NO₃)₂, Ni(NO₃)₂ or Zn(NO₃)₂ aqueous solution followed by calcination in air, these metal elements will exist as their oxidates,²⁹ which belong to Lewis acid. In contrast, thiophene is soft Lewis base,³⁰ thus it can combine with such metal oxidates. Moreover, NiO, ZnO, and CuO are intermediate Lewis acids, whereas Ag and Ag₂O are both soft Lewis acid. So Ag and Ag₂O has a stronger force with thiophene according to the principle of “Hard and Soft Acid and Base”-hard likes the hard and soft likes the soft.³¹ This could explain Ag component is more suitable for the preparation of M/γ-Al₂O₃ sorbent for the desulfurization from coking benzene, compared with Cu, Ni and Zn.

Fig. 2 Desulfurization efficiency (a) and thiophene adsorption capacity (b) of sorbents with different metal loaded in TH–B solution.

3.2. Effect of loading amount and distribution of metal component on desulfurization activity of sorbent

From above experimental results, the desulfurization activity of Ag/γ-Al₂O₃ sorbent is higher than others. So it is necessary to investigate the effect of loading amount and distribution of silver on desulfurization activity of Ag/γ-Al₂O₃ sorbent. Fig. 3 illustrates the desulfurization efficiency and thiophene adsorption capacity of Ag/γ-Al₂O₃ sorbents with different silver loaded amount in thiophene–benzene solution. It can be found that from A0 sorbent to A15 sorbent, the desulfurization efficiency and thiophene adsorption capacity increase with the increase of silver content (shown in Table 1). However, compared with A15 sorbent, the silver content of A20 sorbent is higher, but its desulfurization efficiency (30.6%) is much lower than that of A15 sorbent (96.9%). In order to reveal this phenomenon, XRD, SEM/EDS and N₂ adsorption were carried out to characterize sliver existing form and distribution and the physical structure of A0, A5, A10, A15 and A20 sorbents, which are the typical properties affecting the desulfurization activity of sorbent.

Fig. 3 Desulfurization efficiency (a) and thiophene adsorption capacity (b) of Ag/γ-Al₂O₃ sorbents with different silver content in TH–B solution.

XRD patterns of Ag/γ-Al₂O₃ sorbents with different Ag content are displayed in Fig. 4. Three diffraction peaks at 37.28°, 46.28° and 67.03° corresponding to γ-Al₂O₃, indicate that the crystal structure of γ-Al₂O₃ is retained after AgNO₃ loading and followed heat treatment. Other peaks at 38.12°, 44.30°, 64.44° and 77.40°, corresponding to the simply Ag⁰ crystals,³² show that the active component mainly appears in the form of Ag⁰ after calcination step during preparation. Fig. 5 shows the thermal decomposition process of AgNO₃ in air. The decomposition begins at about 450 °C, which is the same as the result reported by literature,²⁹ and when the decomposition finished, about 64% of the total weight left, which is just the silver content in pure AgNO₃. So from this point, AgNO₃ loaded on sorbent should be decomposed under the experimental condition (calcination temperature was 500 °C) and the chemical reaction follows the equation of 2AgNO₃ = 2Ag + O₂ + 2NO₂.^33,34 In order to further investigate the silver existing form, XPS was carried out to characterize A15 sorbent. As shown in Fig. 6, for A15 sorbent, its XPS spectral of the core level of both Ag 3d_3/2 (a) and O 1s all could be deconvoluted into three peaks. As demonstrated in Fig. 6(a), in the XPS spectral of Ag 3d3/2, the peaks at 373.7 eV, 374.6 eV and 375.5 eV are corresponding to Ag⁰,^35,36 Ag₂O and Ag–O–Al species, respectively. As listed in Fig. 6(b), on the XPS spectral of O 1s the peak at 530.6 eV is attributed to Al₂O₃, and the peaks at 532.2 eV and 533.0 eV should belong to Ag₂O and Ag–O–Al species, respectively. As listed in Fig. 7, in the spectrum of A0, the peaks at 3438 cm⁻¹ and 1636 cm⁻¹ are attributed to –OH groups, the peaks at 760 cm⁻¹ and 574 cm⁻¹ belong to O–Al bond, whereas in the spectrum of A15, the intensity of peaks at 2955 cm⁻¹, 2926 cm⁻¹ and 2855 cm⁻¹ are obviously increases, and some new peaks, 1385 cm⁻¹, 1385 cm⁻¹ and 1116 cm⁻¹ appear, which are attributed to Ag₂O, and the peak at 1743 cm⁻¹ may be corresponding to Ag–O–Al. The XRD, XPS and FT-IR results all suggest that silver exists in Ag⁰, Ag₂O and Ag–O–Al forms on γ-Al₂O₃ support, and among of them Ag is the main component.

Fig. 4 XRD patterns of Ag/γ-Al₂O₃ series sorbents prepared ((a) Ag, (b) γ-Al₂O₃).

Fig. 5 TG and DTG curves of AgNO₃ in air.

Fig. 6 XPS spectral of the core level of Ag 3d3/2 (a) and O 1s region recorded with A15 sorbent.

Fig. 7 FT-IR spectral of A0 and A15 sorbents.

Silver distribution on the surface of Ag/γ-Al₂O₃ sorbents was measured by SEM/EDS, and the results are shown in Fig. 8. On the surface of A10 sorbent, silver distributes homogeneously, and there is no sliver agglomeration in the surface EDS image, which could explain the relative high desulfurization efficiency of A10 sorbent. On A15 sorbent's surface, most of sliver is still distributed very well which should explain why the desulfurization efficiency of A15 sorbent is higher than that of A10 sorbent. Moreover, a few small particles, which are the sliver agglomeration verified from the EDS results, appear on A15 sorbent could explain why the desulfurization efficiency increment from A10 to A15 is not so obvious as that from A5 to A10. This should explain why the desulfurization efficiency of A15 sorbent is slightly higher than that of A10 sorbent. However, for A20 sorbent, most of its surface was covered by sliver agglomeration particles, which could aromatically decrease its desulfurization efficiency, and this is why the desulfurization activity of A20 sorbent is much lower than that of A15 sorbent.

Fig. 8 SEM and EDS results of A10, A15 and A20 sorbents.

Fig. 9 provides the nitrogen adsorption and desorption isotherms of sorbents. The adsorption isotherms of all the sorbents belong to the typical IV type, and all sorbents have mesoporous adsorption behavior. In adsorption isotherms, as the relative pressure increasing from 0.4 to 0.7, the curves rises up sharply, which can be attributed to the occurrence of capillary condensation. And in adsorption/desorption branches, a hysteresis loop is appeared, which indicates the mesopore characteristic of the adsorption–desorption isotherms. The corresponding surface area and pore size distribution of Ag/γ-Al₂O₃ sorbents are listed in Table 3. The pore volume and BET specific area (S_BET) of Ag/γ-Al₂O₃ sorbents all decreases with the silver content increasing. The decrease of S_BET for the A20 sorbent with 20% Ag content should be due to the agglomeration of the active component, which can be observed in its SEM/EDS image (Fig. 8), and this is also why the total pore volume and the average pore diameter (shown in Table 3) of A20 sorbent is much less than that of A15 sorbent.

Fig. 9 N₂ adsorption and desorption isotherms of Ag/γ-Al₂O₃ sorbents.

Table 3 Surface area and pore volume of different sorbents

Sorbents	Specific surface area (m^2 g^−1)	Pore volume (cm^3 g^−1)	Average pore diameter (nm)
A0	156	0.208	36.8
A5	146	0.180	35.0
A10	132	0.170	33.5
A15	132	0.154	31.5
A20	120	0.145	27.9

---

3.3. Adsorption desulfurization mechanisms of Ag/γ-Al₂O₃ sorbent

Adsorption results of Ag/γ-Al₂O₃ sorbents in TH–B and TH–C solutions are shown in Fig. 10. In TH–C solution, the desulfurization efficiency of all Ag/γ-Al₂O₃ sorbents is about 100%, whereas, in TH–B solution, it is lower for A10 and A15 sorbents, and much lower for A5 and A20 sorbents. This suggests that benzene presents the competitive adsorption with thiophene over Ag/γ-Al₂O₃ sorbent. This competitive adsorption effect should be attributed to the connection between the active component and the conjugated pi bond as well as the π-complexation, because the structures of thiophene and benzene are very similar to each other with the unsaturated organic compounds, and the similar composition of six electron conjugated pi bond, which could connect to silver on these sorbents by π-complexation. However, for cyclohexane, it does not have conjugated pi bond, and thus does not have competitive adsorption effect, which is why the desulfurization efficiency of all the Ag/γ-Al₂O₃ sorbents is close to 100%.

Fig. 10 Comparison of desulfurization efficiency of different Ag/γ-Al₂O₃ sorbents in TH–B and TH–C solutions.

Fig. 11 shows the desulfurization efficiency of Ag/γ-Al₂O₃ sorbents in TH–B and THT–B solutions. It can be found that the desulfurization efficiency of all the Ag/γ-Al₂O₃ sorbents in THT–B solution is higher than that in TH–B solution, which means that the removal of tetrahydrothiophene from benzene is much easier than that of thiophene from benzene, and the removal efficiency of tetrahydrothiophene from benzene over Ag/γ-Al₂O₃ sorbents is around 99%. It indicates that benzene almost has no competitive adsorption effect on the adsorption of tetrahydrothiophene on Ag/γ-Al₂O₃ sorbents' surface even benzene could connect to silver by π-complexation, which means that tetrahydrothiophene molecule could be adsorbed on Ag/γ-Al₂O₃ sorbents by a different way. And thus, in addition to the π-complexation adsorption mechanism discussed above, there should be another adsorption mechanism between thiophenes and silver. If a comparison is made between tetrahydrothiophene and thiophene molecules, it is very easy to find that both of them all have sulfur atoms, which have lone-pair electrons. So it is probable that tetrahydrothiophene and thiophene can act as the n-type donors by donating the lone-pair electrons that lie in the plane of the ring to the silver on Ag/γ-Al₂O₃ sorbents to form S–Ag bond (S–M mechanism³⁷), because silver has unoccupied orbital, which could accept this donation.

Fig. 11 Desulfurization efficiency of Ag/γ-Al₂O₃ sorbents in TH–B and THT–B solutions.

Fig. 12 show the FT-IR spectral of different samples in the range of 4000 cm⁻¹–400 cm⁻¹ (a), 3100 cm⁻¹–2700 cm⁻¹ (b) and 1800 cm⁻¹–1000 cm⁻¹ (c). It can be seen that, after benzene, thiophene and tetrahydrothiophene absorbed on A15 sorbent, the intensity of the peaks at 2925 cm⁻¹, 2926 cm⁻¹, 2855 cm⁻¹ and 1385 cm⁻¹ all decreased significantly, and the peaks at 1743 cm⁻¹, 1462 cm⁻¹, 1116 cm⁻¹ disappeared, which suggest that the Ag₂O and Ag–O–Al spices all can combine to benzene, thiophene and tetrahydrothiophene.

Fig. 12 FT-IR spectral of different samples in the range of 4000 cm⁻¹–400 cm⁻¹ (a), 3100 cm⁻¹–2700 cm⁻¹ (b) and 1800 cm⁻¹–1000 cm⁻¹ (c).

In order to further understand the π-complexation and S–M mechanisms proposed above, the desulfurization behavior in TH–THT–C and TH–THT–B solutions over Ag/γ-Al₂O₃ sorbents was studied, and the results were shown in Fig. 13. The desulfurization efficiency for tetrahydrothiophene and thiophene in TH–THT–C solution is 100%, whereas that for thiophene in TH–THT–B solution (95.8%) is relatively lower. It indicates that thiophene almost has no competitive effect on tetrahydrothiophene adsorption, which may because the S–Ag bond formed between tetrahydrothiophene and silver is stronger than that formed between thiophene and silver, for the sulfur atom in tetrahydrothiophene molecule has two lone-pair electrons, whereas the sulfur atom in thiophene molecule only has one lone-pair electrons. It could also explain why the adsorption rate of tetrahydrothiophene (shown in Fig. 14) in these two solutions is much higher than that of thiophene.

Fig. 13 Comparison of thiophene adsorption over A15 sorbent in TH–THT–B and TH–THT–C solution.

Fig. 14 Adsorption curves of thiophene and tetrahydrothiophene over A15 sorbent in thiophene–tetrahydrothiophene–cyclohexane and thiophene–tetrahydrothiophene–benzene solutions.

4. Conclusions

Copper, nickel, zinc and silver elements loaded on γ-Al₂O₃ could significantly improve the desulfurization activity of M/γ-Al₂O₃ sorbents, and silver is the best. On Ag/γ-Al₂O₃ sorbent, the loading amount and distribution of silver are the main factors influencing the desulfurization capacity of sorbent. With the increase of silver content in sorbent from 0 to 13.8%, the desulfurization efficiency increases, whereas when it is increased to 17.6%, the sliver agglomeration appears very seriously, and thus its desulfurization efficiency decreases obviously. There are two adsorption desulfurization mechanisms on Ag/γ-Al₂O₃ sorbent, π-complexation and S–M bond between thiophene and silver. The competitive adsorption of benzene with thiophene over Ag/γ-Al₂O₃ sorbent is caused by the π-complexation mechanism.

Acknowledgements

The authors gratefully acknowledge Mr Yanjun Zhang and Ms Yanna Han for the samples preparation and SEM/EDS characterization, respectively. The financial support of National Natural Science Foundation of China (51372161, 21406151), Research Fund for Doctoral Program of High Education (20131402110010) and Shanxi Scholarship Council of China (2012-0039) were highly appreciated.

References

1. P. F. McMillan, Nat. Mater., 2007, 6, 7–8.
2. R. Ruhela, A. Rao, N. Iyer, A. Das, P. Kumar, A. K. Singh, B. S. Tomar and R. C. Hubli, RSC Adv., 2014, 4, 62654–62661.
3. T. Fukuda, Y. Hayasaki, T. Hasumura, Y. Katsube, R. L. D. Whitby and T. Maekawa, RSC Adv., 2015, 5, 12671–12677.
4. J. Liao, Y. Wang, L. Chang and W. Bao, Green Chem., 2015, 17, 3164–3175.
5. C. A. Johnson and S. C. Schuman, US Pat., 2707699, 1955.
6. K. Wehner, W. Kisan and G. Kunz, US Pat., 3879268, 1975.
7. Y. Shen, X. Xu and P. Li, RSC Adv., 2012, 2, 6155–6160.
8. F. Duan, C. Chen, G. Wang, Y. Yang, X. Liu and Y. Qin, RSC Adv., 2014, 4, 1469–1475.
9. P. S. Kulkarni and C. A. M. Afonso, Green Chem., 2010, 12, 1139–1149.
10. X. Li, X. Zhang and L. Lei, Sep. Purif. Technol., 2009, 64, 326–331.
11. G. Bai, X. Lan, X. Liu, C. Liu, L. Shi, Q. Chen and G. Chen, Green Chem., 2014, 16, 3160–3168.
12. A. Samokhvalov, E. C. Duin, S. Nair, M. Bowman, Z. Davis and B. J. Tatarchuk, J. Phys. Chem. C, 2010, 114, 4075–4085.
13. C. Shen, Y. J. Wang, J. H. Xu, Y. C. Lu and G. S. Luo, Green Chem., 2012, 14, 1009–1015.
14. C. Li, D. Li, S. Zou, Z. Li, J. Yin, A. Wang, Y. Cui, Z. Yao and Q. Zhao, Green Chem., 2013, 15, 2793–2799.
15. T. Jin, Q. Yang, C. Meng, J. Xu, H. Liu, J. Hu and H. Ling, RSC Adv., 2014, 4, 41902–41909.
16. Y. Zeng and S. Ju, Sep. Purif. Technol., 2009, 67, 71–78.
17. L. Song, T. Bu, L. Zhu, Y. Zhou, Y. Xiang and D. Xia, J. Phys. Chem. C, 2014, 118, 9468–9476.
18. J. Liao, W. Bao, Y. Chen, Y. Zhang and L. Chang, Energy Sources, Part A, 2012, 34, 618–625.
19. J. Liao, W. Wang, Y. Xie, Y. Zhang, L. Chang and W. Bao, Sep. Sci. Technol., 2012, 47, 1880–1885.
20. 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.
21. A. H. M. S. Hussain, M. L. McKee, J. M. Heinzel, X. Sun and B. J. Tatarchuk, J. Phys. Chem. C, 2014, 118, 14938–14947.
22. L. Wu, J. Xiao, Y. Wu, S. Xian, G. Miao, H. Wang and Z. Li, Langmuir, 2014, 30, 1080–1088.
23. D. Lee, J. Kim, H. C. Lee, K. H. Lee, E. D. Park and H. C. Woo, J. Phys. Chem. C, 2008, 112, 18955–18962.
24. J. Gislason, Oil Gas J., 2001, 99, 72.
25. C. Song, Catal. Today, 2003, 86, 211–263.
26. J. Guo, M. J. Janik and C. Song, J. Phys. Chem. C, 2012, 116, 3457–3466.
27. S. Sato, F. Nozaki, S.-J. Zhang and P. Cheng, Appl. Catal., A, 1996, 143, 271–281.
28. A. J. Hernández-Maldonado and R. T. Yang, Ind. Eng. Chem. Res., 2003, 42, 123–129.
29. S. Yuvaraj, L. Fan-Yuan, C. Tsong-Huei and Y. Chuin-Tih, J. Phys. Chem. B, 2003, 107, 1044–1047.
30. M. Xue, R. Chitrakar, K. Sakane, T. Hirotsu, K. Ooi, Y. Yoshimura, Q. Feng and N. Sumida, J. Colloid Interface Sci., 2005, 285, 487–492.
31. R. G. Pearson, J. Am. Chem. Soc., 1963, 85, 3533–3539.
32. S. Ribeiro, C. M. Granadeiro, P. Silva, F. A. Almeida Paz, F. F. de Biani, L. Cunha-Silva and S. S. Balula, Catal. Sci. Technol., 2013, 3, 2404–2414.
33. J.-W. Kwon, S. H. Yoon, S. S. Lee, K. W. Seo and I.-W. Shim, Bull. Korean Chem. Soc., 2005, 26, 837–840.
34. K. Otto, I. Oja Acik, M. Krunks, K. Tõnsuaadu and A. Mere, J. Therm. Anal. Calorim., 2014, 118, 1065–1072.
35. S. M. Hosseini, I. A. Sarsari, P. Kameli and H. Salamati, J. Alloys Compd., 2015, 640, 408–415.
36. S. W. Han, Y. Kim and K. Kim, J. Colloid Interface Sci., 1998, 208, 272–278.
37. X. Yan, P. Mei, L. Xiong, L. Gao, Q. Yang and L. Gong, Catal. Sci. Technol., 2013, 3, 1985–1992.

---