Effect of olefin and aromatics on thiophene adsorption desulfurization over modified NiY zeolites by metal Pd

Abstract

NiY and NiPdY adsorbents were prepared by incipient wetness impregnation and characterized by X-ray diffraction (XRD), N₂ adsorption–desorption isotherms, scanning electron microscopy (SEM), and Fourier transform infrared spectroscopy (FT-IR). The effects of olefin and aromatics on desulfurization over NiY and NiPdY adsorbents were studied; moreover, the adsorption mechanisms of different compounds, including thiophene, olefin, and aromatics, were investigated by in situ FT-IR. The results showed that thiophene, olefin, and aromatics were adsorbed on HY through π complexation, which will lead to strong competitive adsorption. Significantly, after modifying HY zeolite by metal Ni, the adsorption desulfurization rate could be improved from 57.0% to 65.7% in the olefin–thiophene system (M-II) and from 21.8% to 73.1% in the aromatic–thiophene system (M-III) because S–metal bonds were formed directly when thiophene once touched the Ni surface. The breakthrough sulfur capacity of NiY was 0.68 mg g⁻¹ in M-II and 0.88 mg g⁻¹ in M-III. However, introduced Ni also increased the number of Brönsted acid sites, which could result in protonated reactions of thiophene and olefin and lead to blockage of pores. The metal Pd was chosen to further modify the acidity and pore structure of NiY. The results showed that not only was the adsorption desulfurization rate improved further from 65.7% to 85.8% in M-II and from 73.1% to 87.0% in M-III, but breakthrough sulfur capacity was also enhanced from 0.68 to 1.64 mg g⁻¹ in M-II from 0.88 to 2.01 mg g⁻¹ in M-III.

---

1. Introduction

More restrictive regulations are being implemented on the sulfur content in transportation fuels due to environmental problems. One of the common methods for removal of sulfur from gasoline is hydrodesulfurization technology (HDS).¹ However, hydrogenation will inevitably lead to the saturation of olefins and loss of octane number (RON). Thus, some non-HDS desulfurization technologies have been concerned such as oxidative desulfurization,² extraction desulfurization,³ biological desulfurization,³ and adsorption desulfurization (ADS).⁴

ADS has attracted considerable interest because of its mild conditions, no loss of RON, and ultra-low sulfur-containing products. The development of efficient adsorbents plays a vital role in the further progress of ADS technology. Research has proven that porous materials, such as zeolite,⁵ active carbon,^6,7 boron nitride,^8–10 and metal oxide,¹¹ could act as supporters for the adsorbents for sulfur storage. Among these, Y zeolite is believed to be a promising candidate for ADS due to the well-defined pore structure, large surface area, suitable acidity, and other good characteristics.^12–14 However, disputable opinions for the selection of active metal^15–17 within the adsorbent were held. Shan¹⁵ revealed Cu(I)Y to be a better adsorbent than NaY for desulfurization because the introduction of active metal improved the selectivity for sulfur; however, the performance of desulfurization of Cu(I)Y was greatly affected by the existing aromatics in gasoline. Shi¹⁶ reported that the ability of desulfurization of Ce(IV)Y was greater than that of NaY; however, it still cannot eliminate or decrease the negative influences of increasing toluene concentration during adsorption desulfurization. Wang¹⁷ investigated the selectivity of sulfur over Ce(IV)Y zeolites in thiophene/1-octane system and found that Ce(IV)Y had satisfied the selectivity of thiophene only when the concentration of 1-octane was less than 150 mg g⁻¹. Although active metals have the ability to improve the selectivity of sulfur compounds, coexisting olefins or aromatics with thiophene could generate strong competitive adsorption, which decrease the selectivity of thiophene and sulfur adsorption capacity of the adsorbents. Therefore, avoiding or decreasing these competition adsorptions becomes the bottleneck for developing effective adsorbents.

In competitive adsorption mechanisms, Xiong⁹ found that the Lewis acid–base interaction plays an important role in desulfurization. Shi¹⁸ employed FT-IR spectroscopy to determine the thiophene adsorption mode and indicated that thiophene was adsorbed mainly on NaY via π electron interactions, whereas it was adsorbed on LaNaY via π electron interaction combined with the La–S direct interaction. Yang¹⁹ reported that the desulfurization performance of Ce(IV)Y zeolites and proposed that thiophene was selected through π-complexation. However, Song²⁰ conducted the adsorptive desulfurization of model gasoline with zeolites modified by different metals (Cu, Ni, Zn, Pd, and Ce). They found that CeY zeolites had a better selectivity for thiophene than other adsorbents, and thiophene was adsorbed over CeY zeolites via a direct S–metal bond. Therefore, the adsorption mechanisms of thiophene on different metals were controversial and further research is required.

On the other hand, studies revealed that the acidity of adsorbents could affect the ADS mechanism. Subhan²¹ showed that Lewis acid sites played an important role in the removal of organic sulfur compounds over reduced 15% Ni-AlMCM-41. Zhang²² also studied the effects of the acid properties of Hβ zeolites for thiophene alkylation and hexene oligomerization. The results showed that olefin alkylation and thiophene oligomerization mainly occurred on Brönsted acid sites. Moreover, studies^17,18 also demonstrated that the protonation of thiophene and cyclohexene was also detected on the Brönsted acid sites of LaNaY or Ce(IV)Y. Thus, thiophene alkylation and olefin oligomerization could occur on the Brönsted acid sites, which inhibited thiophene adsorption.

Pd, as a modified metal, has been used widely in the hydrogenation of furfural over the Pd–Cu/Al₂O₃ catalysts²³ or liquid catalytic hydrogenation phenylacetylene reaction,²⁴ which modified the electronic effects of the catalysts. Although Pd has seldom been applied in adsorption desulfurization, the adsorbent containing Pd could improve the activity of the adsorbent to remove sulfur. Moreover, it seems that the adsorbents made by ion-exchange has been overemphasized, whereas the impregnation preparation has been ignored.

In this study, we investigated adsorptive desulfurization over Ni impregnated on Y zeolites, and further modified by Pd as a coaction in the Ni-containing Y zeolite on the adsorptive desulfurization of model gasoline. In addition, the mechanisms of thiophene, olefins, and aromatics on NiY and NiPdY adsorbents will be demonstrated by in situ FT-IR spectroscopy, and the effect of the acid properties of NiY and NiPdY adsorbents on the selectivity of thiophene adsorption is investigated.

2. Experimental

2.1 Preparation of adsorbents

HY zeolite was bought from the Catalyst Plant of Nankai University with a SiO₂ : Al₂O₃ ratio of 4 : 8. Ni(NO₃)₂·6H₂O and Pd(NO₃)₂·2H₂O were supported over HY zeolites by incipient wetness impregnation. The NiO and PdO loading were 15 and 1 wt%, respectively. In a typical impregnation, HY zeolite was dried overnight. Second, the weighed 5.96 g Ni(NO₃)₂·6H₂O was dissolved in 6.8 mL deionized water and the solution was added to the dried HY zeolite dropwise with constant stirring. After impregnation, the sample was kept in an oven at 120 °C for 6 h and calcined at 450 °C for 4 h. The prepared adsorbent was called NiY. The second metal Pd was loaded to the NiY by the same incipient wet impregnation procedures with weighed 0.11 g Pd(NO₃)₂·6H₂O. The sample was denoted as NiPdY.

2.2 Fuels

To understand the competitive adsorption between thiophene and olefin or aromatic, 1-hexene and toluene were selected as the representatives of olefin and aromatic, respectively. The HY, NiY, and NiPdY adsorbents were tested with different model gasolines, which were denoted as M-I, M-II, and M-III. The compositions of the model gasoline are described in Table 1.

Table 1 Compositions of model gasoline

Model gasoline	Sulfur content (mg L^−1)	Compositions
M-I	52.3	Thiophene/cyclohexane
M-II	51.7	5 vol% 1-hexene and thiophene/cyclohexane
M-III	49.8	5 vol% toluene and thiophene/cyclohexane

---

2.3 Characterization of adsorbent

X-ray diffraction of the adsorbents was carried out on a Bruker D8 advance X-ray diffractometer using Cu Kα radiation under the setting conditions of 40 kV, 30 mA, and a scan range from 5° to 80° at a rate of 10° min⁻¹. Micro-morphology was studied using a Cambridge S-360 scanning electron microscope (SEM). To identify the chemical state and surface compositions of the NiY and NiPdY adsorbents, X-ray photoelectron spectroscopy (XPS) of the adsorbents was performed using a Thermo Fisher K-Alpha analyzer.

The textural characterization was achieved using conventional N₂ adsorption–desorption method with a Micrometrics ASAP 2010 automatic analyzer. All the adsorbents were outgassed at 473 K until the vacuum pressure was 0.8 kPa. The adsorption isotherms for nitrogen were measured at 77 K. The t-plot method was used to calculate the external surface area. The specific surface area, pore volume and pore size distribution were measured by the BET (Brunauer–Emmett–Teller) nitrogen adsorption capacity method.

The acidic property (acid type and acid amount) of adsorbents was measured by FT-IR. The Brönsted acid and Lewis acid sites of the adsorbents were investigated using a Nicolet Avatar 360 spectrometer based on pyridine adsorption by 64 scans with a resolution of 4 cm⁻¹. It was reduced by H₂ at 360 °C for 5 h, then cooled to room temperature, pyridine adsorbed, heated to 200 °C, evacuated for 30 min, and then cooled to room temperature; the IR spectra were obtained. Similarly, the sample was heated to 350 °C and then cooled down; the IR spectra were acquired. The bands displayed were attributed to the interaction of pyridine with Lewis (L) and Brönsted (B) acid sites on the sample surfaces. The total amount of acid was calculated by the curve tested desorbed at 200 °C, strong acid amount was calculated by the curve tested desorbed at 350 °C.

Adsorption mechanisms of the adsorbents were measured by FT-IR on a Nicolet Avatar 360 spectrometer by 64 scans with a resolution of 4 cm⁻¹. The samples were placed in a quartz IR cell with CaF₂ windows. It was reduced by H₂ at 360 °C for 5 h and then cooled to room temperature. Thiophene, 1-hexene or toluene was adsorbed for 30 min, then the samples were vacuumed at room temperature for 20 min, and the spectra was obtained.

Adsorptive desulfurization for fuels over the adsorbents was carried out in a fixed-bed flow reactor. The diameter of the reactor was 1 cm and the length was 40 cm. Before each experiment, the adsorbent was heated to 360 °C under H₂ for 5 h to reduce Ni²⁺ and then cooled to room temperature. At room temperature and ambient pressure, fuel was pumped into the fixed-bed flow reactor with a flow rate of 0.5 mL min⁻¹. The sulfur content of the effluent was detected by a TCS-2000S Ultraviolet fluorescence sulfur analyzer.

where C₀ is the initial concentration of sulfur (mg L⁻¹) and C is the effluent concentration of sulfur (mg L⁻¹). Each experiment was repeated 5 times and the experiment data were the average values. Process flow diagram is shown in Fig. 1.

Fig. 1 Process flow diagram of fixed bed adsorption desulfurization.

3. Results and discussion

3.1 Adsorbent characterization

XRD patterns of HY, NiY, and NiPdY adsorbents are shown in Fig. 2. Eight characteristic diffraction peaks of Y zeolite at 6.31°, 10.31°, 12.10°, 15.92°, 20.71°, 24.06°, 27.52°, and 31.95° were detected for all displayed adsorbents, which correspond to the (1 1 1), (2 2 0), (3 1 1), (3 3 1), (4 4 0), (5 3 3), (6 4 2), and (5 5 5) planes, respectively. When modified by metals, no new diffraction peaks appeared, indicating that the active ingredient NiO and PdO were well dispersed. NiY and NiPdY maintained a good crystalline structure; only parts of the crystallites were slightly damaged by the acidic solution during impregnation. The morphologies of these two adsorbents could be found in Fig. 3.

Fig. 2 XRD patterns of HY, NiY, and NiPdY adsorbents.

Fig. 3 SEM images of adsorbents (a) HY, (b) NiY, (c) NiPdY.

To identify the chemical state and surface compositions of NiY and NiPdY adsorbents, X-ray Photoelectron Spectroscopy (XPS) was carried out and the results are shown in Table 2 and Fig. 4. Table 2 shows the XPS surface compositions of NiY and NiPdY adsorbents. Compared to NiY adsorbents, the surface Ni content of the NiPdY adsorbent increased by 24.6% and a higher Ni/Si ratio was observed in the NiPdY adsorbents, which indicated that Pd species was favorable to enrich and disperse the Ni species on the surface. It was worth noting that no Pd species were found on the surface of the NiPdY adsorbents, which was mainly due to the lower Pd loading. The X-ray Photoelectron Spectroscopy (XPS) spectra of Ni species on NiY and NiPdY adsorbents are shown in Fig. 4. Ni²⁺ can be confirmed by the Ni 2p_3/2 binding energy at around 855.5 eV and with a satellite centering at 862 eV.²⁵ Ni metal state can be assigned to the Ni 2p_3/2 binding energy at about 852.4 eV.²⁶ It was obvious that the content of the Ni metal state over NiPdY was lower than that over NiY, showing that the Pd species had an interaction with Ni species, which inhibited the reduction of Ni²⁺.

Table 2 XPS surface compositions of NiY and NiPdY adsorbents

Adsorbent	Surface composition (atom%)				Ni/Si atomic ratio
	Ni	Si	Al	O
NiY	4.44	16.26	13.15	66.15	0.27
NiPdY	5.16	15.30	12.78	66.76	0.34

---

Fig. 4 X-ray Photoelectron Spectroscopy (XPS) spectra of Ni species on NiY and NiPdY adsorbents.

The N₂ adsorption–desorption isotherms of the absorbents exhibited a type-I adsorption isotherms (Fig. 5), which indicated that the adsorbents HY had a regular pore structure and those pores were essentially composed of most microporous and few mesopores. However, the equilibrium adsorption capacity of the NiY and NiPdY adsorbents decreased greatly compared to that of HY because some pores were blocked by metal species. Data of the surface area and pore volume are given in Table 3. Surface area, pore volume and mesopore volume of HY were 629.6 m² g⁻¹, 0.34 cm³ g⁻¹, and 0.06 cm³ g⁻¹, respectively. The surface area and pore volume of NiY decreased by 310.7 m² g⁻¹ and 0.16 cm³ g⁻¹, respectively. This also indicated that parts of the pores were destroyed by the acidic solution during impregnation and some of the pores were also blocked, which decreased the surface area and pore volume greatly. However, the surface area, pore volume and mesopore volume of NiPdY increased compared with those of NiY. This demonstrated that trace Pd weakens the blocking of both microporous and mesoporous. As shown in Fig. 6, it can be seen that more mesopores appeared around 10 nm on the NiPdY adsorbent than that of NiY adsorbent.

Fig. 5 N₂ adsorption–desorption isotherms of HY, NiY, and NiPdY adsorbents.

Table 3 Texture properties of HY, NiY, and NiPdY adsorbents

Sample	Surface area A/(m^2 g^−1)			Pore volume V/(cm^3 g^−1)
	Total	Micropore	Mesopore	Total	Micropore	Mesopore
HY	629.6	573.7	55.9	0.34	0.28	0.06
NiY	318.9	289.1	29.8	0.18	0.14	0.04
NiPdY	345.5	300.4	45.1	0.20	0.15	0.05

---

Fig. 6 Pore size distribution of HY, NiY, and NiPdY adsorbents.

To explore the influence of acidities of the adsorbents on desulfurization performance, the type and amount of surface acidic sites on NiY and NiPdY were investigated. The acidities of different adsorbents are shown in Table 4. Lewis acid and Brönsted acid were detected on the adsorbents. Most of the Brönsted acid was weak Brönsted acid. The amount of Brönsted acid and Lewis acid increased on the NiY adsorbent due to the introduction of Ni. With metal Pd supported on NiY, the amount of strong Lewis acid on NiPdY increased, and the amount of weak Lewis acid was reduced by about 50%. The total amount of Brönsted acid reduced, and weak Brönsted acid increased, whereas strong Brönsted acid almost declined to zero. Brönsted acid was the active center of the C=C protonated reaction and oligomerization.¹⁷ The lone pair of electrons of thiophene was easily adsorbed on the Lewis acid sites to form a S–M bond.²⁷ The number of Brönsted acid sites (especially strong Brönsted acid) reduced by the introduction of Pd. Side reactions were weakened, and the desulfurization performance could be further improved.

Table 4 Acid properties of different adsorbents^a

	B (μmol g^−1)			L (μmol g^−1)
	TB	SB	WB	TL	SL	WL
a SL—strong Lewis acid; SB—strong Brönsted acid; WL—weak Lewis acid; WB—weak Brönsted acid.
HY	243.1	29.2	213.9	72.8	42.5	30.4
NiY	287.1	94.7	192.4	96.2	62.8	33.4
NiPdY	223.1	1.2	221.9	104.2	86.7	17.5

---

3.2 Adsorption mechanism

3.2.1 Adsorption mechanism of thiophene. FT-IR spectra of thiophene adsorbed on HY, NiY, and NiPdY are shown in Fig. 7. The bands at about 1401, 1418, 1446, 1455, and 1496 cm⁻¹ were observed for all the adsorbents. The band at 1401 cm⁻¹, which shifted 8 cm⁻¹ to a lower wavenumber, was assigned to the perturbed symmetric stretching vibration of C=C in the fundamental ring.²⁸ The red shift was caused by a decrease in the electron density of the entire thiophene ring, which indicated that thiophene was parallel to the surface of HY through π-complexation.²⁹

Fig. 7 FT-IR spectra of thiophene adsorption on HY, NiY, and NiPdY.

The band at 1418 cm⁻¹ could be ascribed to the stretching vibration of C=C in the fundamental ring of thiophene molecule shifting 9 cm⁻¹ to high wavenumber.¹⁷ The blue shift was caused by the increase in electron density of the entire thiophene ring by Ni or Pd metal directly interacting with the S atom in thiophene (S–metal). The bands at 1446 and 1455 cm⁻¹ were attributed to –CH₂ groups of *CH₂–(CH=CH₂) and *CH₂–S,³⁰ which indicated that a protonated reaction of thiophene occurred. The band at 1496 cm⁻¹ was related to a surface reaction of thiophene, involving protonation reactions, opening of the thiophene ring, or formation of new surface species.³¹ By the introduction of Pd, the oligomerization of thiophene molecules can be weakened. In conclusion, the adsorption mechanism of thiophene showed that thiophene mainly interacted with HY through π complexation. By modification, thiophene was adsorbed on NiY and NiPdY through π complexation and S–metal interaction. Moreover, oligomerization of thiophene molecules occurred on the center of the Brönsted acid sites.

3.2.2 Adsorption mechanism of 1-hexene. FT-IR spectra of 1-hexene adsorbed on HY, NiY, and NiPdY are shown in Fig. 8. The bands in the region of 1750–1300 cm⁻¹ were discussed. The characteristic absorption peaks of the C=C stretching vibration in 1-hexene were approximately 1645 cm⁻¹ (C=C stretching vibration).¹⁸ A peak at 1630 cm⁻¹ was detected on HY, which was 15 cm⁻¹ lower than that of 1-hexene gas. The low wavenumber shift was caused by a decrease in the electron density of C=C in 1-hexene, which indicated that C=C in 1-hexene interacted with the metal cations by π complexation.¹⁸ However, the bands at 1630 cm⁻¹ of 1-hexene adsorbed on the adsorbents were very weak, because protonated reaction of C=C of 1-hexene occurred. When 1-hexene was adsorbed on Brönsted acid sites on the adsorbents, 1-hexene was protonated to form alkenyl carbenium ions.¹⁸ Consequently, 1-hexene was adsorbed on the adsorbents. The protonated reaction of 1-hexene then occurred on the Brönsted acid sites. In addition, a small part of 1-hexene was adsorbed through π complexation.

Fig. 8 FT-IR spectra of 1-hexene adsorption on HY, NiY, and NiPdY.

3.2.3 Adsorption mechanism of toluene. FT-IR spectra of toluene adsorbed on HY, NiY, and NiPdY are shown in Fig. 9. The bands at about 1494 cm⁻¹ and 1598 cm⁻¹ were detected in the FT-IR spectra of toluene on HY, NiY, and NiPdY. The bands at about 1494 cm⁻¹ and 1598 cm⁻¹ can be assigned to the vibration bands of the ring skeleton of toluene adsorbed on the adsorbents.³² They were decreased separately by 4 cm⁻¹ and 7 cm⁻¹ compared to the band of gaseous toluene at 1498 cm⁻¹ and 1605 cm⁻¹. The red shifts were caused by a decrease in the electric density, which showed that toluene was adsorbed on the metal cation sites of HY, NiY, and NiPdY adsorbents by π complexation, implying that the mechanism of toluene adsorption did not change with modification by the introduction of the metal cations and was consistent with the literature.¹⁶ Summarily, the adsorption mechanisms of toluene on HY, NiY, and NiPdY were almost unanimous, namely, interaction by π complexation. The introduction of active metal Ni and Pd did not change the adsorption mode of toluene.

Fig. 9 FT-IR spectra of toluene adsorption on HY, NiY, and NiPdY.

Thus, adsorption mechanisms of thiophene, 1-hexene, and toluene on HY, NiY, and NiPdY have been manifested by in situ FT-IR. Thiophene, 1-hexene, and toluene adsorption on HY were through π complexation. In addition, the protonated reaction of both thiophene and 1-hexene occurred on Brönsted acid sites. Therefore, competitive adsorption can generate between olefin and thiophene, aromatics and thiophene on HY. To avoid this competitive adsorption, we can improve the selectivity of thiophene by optimizing the adsorbent. In this study, HY zeolite was modified by Ni, and then by noble metal Pd. To take advantage of the NiY adsorbent and obtain better adsorbents, Pd was supported on the NiY adsorbent (NiPdY) to alter the electronic effect or the acidity on the adsorbent surface. At room temperature and pressure, NiY and NiPdY will be evaluated by model gasoline, containing thiophene, 1-hexene or toluene.

3.3 Fixed-bed adsorption desulfurization experiments

To further examine the competition adsorption between olefin and thiophene, aromatics and thiophene, several experiments were designed. The desulfurization efficiency of different model fuels on the displayed adsorbents is given in Table 5.

Table 5 Desulfurization rates on HY, NiY, and NiPdY adsorbents in different model gasolines

Adsorbents	Desulfurization rate (%)
	M-I	M-II	M-III
HY	97.0	57.0	21.8
NiY	99.4	65.7	73.1
NiPdY	99.6	85.8	87.0

---

As shown in Fig. 10, HY, NiY, and NiPdY were researched to remove thiophene from cyclohexane using M-I. The sulfur adsorption performance of HY was demonstrated to be excellent, and the adsorption desulfurization rate of HY was 97%. After 1 hour in the fixed reactor, the sulfur content of the effluent was less than 10 mg L⁻¹, which was the reason why HY was selected as the support. The sulfur content of the fixed-bed effluent on HY arrived at penetration point (the breakthrough sulfur content was 10 mg L⁻¹) after 80 minutes. The sulfur content of fixed-bed effluent on NiY arrived at the penetration point after 120 minutes, namely, the breakthrough sulfur capacity of NiY was 3.44 mg g⁻¹. Similarly, the sulfur content of the fixed-bed effluent on NiPdY arrived at the penetration point after 140 minutes and the breakthrough sulfur capacity of NiPdY was 3.86 mg g⁻¹.

Fig. 10 Breakthrough curves for the adsorptive desulfurization at room temperature over HY, NiY, and NiPdY adsorbents in M-I.

For M-I only containing thiophene and cyclohexane, HY exerted a good adsorption desulfurization performance. Because competitive adsorption did not exist in M-I, the larger surface area and pore volume of HY provided a great advantage for adsorption desulfurization. Moreover, thiophene was adsorbed on HY by π complexation and a protonated reaction occurred on the center of Brönsted acid. The protonated reaction resulted in pore blockage and the coverage of adsorption active centers, which was unfavorable for adsorption desulfurization.³³ The adsorption modes of thiophene on NiY were S–M and π complexations. In particular, the S–M bond improved the selectivity of thiophene. Moreover, the specific surface area and pore volume of NiPdY increased compared to NiY during impregnation. The amount of Brönsted acid on the NiPdY adsorbent was reduced and the amount of Lewis acid on the NiPdY adsorbent was increased by the introduction of Pd, which promoted the thiophene adsorption.

3.3.1 Effect of olefin on thiophene adsorption. To study the effect of olefin on the sulfur adsorption performance, HY, NiY, and NiPdY adsorbents were implemented in the adsorption experiment. The results are shown in Fig. 11.

Fig. 11 Breakthrough curves for the adsorptive desulfurization at room temperature over HY, NiY, and NiPdY adsorbents in M-II.

In M-II, trace thiophene competed with a large number of olefin on the adsorbents. Obviously, breakthrough sulfur capacity decreased on all the adsorbents tested. The sulfur content of the fixed-bed effluent on HY arrived at the penetration point quickly. The breakthrough sulfur capacity of NiY and NiPdY was 0.68 and 1.64 mg g⁻¹, respectively. The adsorption desulfurization rates of HY, NiY, and NiPdY were 57.0%, 65.7%, and 85.8%, respectively. It could be seen that the impact of olefin on thiophene adsorption was high, which is consistent with the literature.³⁴ Although HY had great pore system, the protonated reaction of both thiophene and 1-hexene occurred on the Brönsted acid sites, which inhibited the thiophene adsorption.

The desulfurization performance of NiY was better than that of HY. The introduction of Ni increased the selectivity of thiophene among a number of olefins. Thiophene was adsorbed on NiY through two modes: π complexation and direct coordination via S atoms (S–metal). S–metal bond largely increased the thiophene adsorption capacity. The selectivity of thiophene was enhanced by the introduction of Pd on NiY. The number of Brönsted acid sites on NiPdY decreased, so the protonated reactions of thiophene and 1-hexene were weakened. The increase in the number of Lewis acid sites promoted thiophene adsorption, and the mesopore surface area and volume increased, which was beneficial to thiophene adsorption. In conclusion, the improvement in the thiophene removal can be attributed to both change in the number of the acid sites and the increase in the number of mesopores, which is consistent with the literature.³⁵

3.3.2 The effect of aromatic in thiophene adsorption. In M-III, trace thiophene competed with a large number of aromatics on the adsorbents (Fig. 12). Apparently, the presence of aromatic seriously affected thiophene adsorption on the adsorbents. HY almost lost the selectivity of thiophene. However, the breakthrough sulfur capacity of NiY and NiPdY was 0.88 and 2.01 mg g⁻¹, respectively. The adsorption desulfurization rates of HY, NiY, and NiPdY were 21.8%, 73.1%, and 87.0%, respectively.

Fig. 12 Breakthrough curves for the adsorptive desulfurization at room temperature over HY, NiY, and NiPdY adsorbents in M-III.

Both thiophene and toluene adsorptions on HY were mainly through π complexation, so strong competitive adsorption was generated. Therefore, the large amount of aromatics in gasoline occupied many active sites on HY. For NiY and NiPdY, the selectivity of thiophene was increased on the modified adsorbents, so the desulfurization performances of NiY and NiPdY were better than HY. Pd was supported on NiY, which caused the addition of some new actives, and the decrease in the number of Brönsted acid sites on NiPdY adsorbent weakened the protonated reaction of thiophene. An increase in the number of Lewis acid sites improved the selectivity of thiophene. Furthermore, the surface area and pore volume relatively increased compared with NiY.

In conclusion, olefin and aromatics have a great influence on the breakthrough sulfur capacity of thiophene; the protonated reaction of olefin and thiophene occurred on the Brönsted acid sites on the adsorbents, and the protonated species can then oligomerize on NiY and NiPdY, which inhibited thiophene adsorption. The strong competitive adsorption of aromatics was generated because aromatics and thiophene were adsorbed by π complexation.

4. Conclusions

Selective adsorption desulfurization on NiY and NiPdY were examined. The results showed that olefin and aromatic can largely reduce the breakthrough sulfur capacity of thiophene. Thiophene, 1-hexene, and toluene adsorption on HY were mainly through π complexation, so competition adsorption were obviously intense. Thiophene adsorbed on NiY and NiPdY was through two modes: π complexation and direct coordination via S atoms (S–metal). Therefore, the breakthrough sulfur capacity can be increased by the introduction of Ni and Pd. In addition, protonated reactions of olefin and thiophene at the Brönsted acid sites, and then the protonated species can oligomerize on NiY and NiPdY. Finally, the role of Pd was discussed in terms of the acidity and mesopores on the catalyst surface. Pd had greater selectivity for thiophene on the NiPdY adsorbent despite the olefin or aromatics.

Acknowledgements

This study was supported by the National Natural Science Foundation of China (21336011, 21476260, 21236009, and U1162204) and the Science Foundation of China University of Petroleum, Beijing (2462015YQ0316 and 2462015YQ0311).

References

1. Q. Huo, T. Dou, Z. Zhao and H. Pan, Appl. Catal., A, 2010, 381, 101–108.
2. W. Zhang, J. Xiao, X. Wang, G. Miao, F. Ye and Z. Li, Energy Fuels, 2014, 28, 5339–5344.
3. C. Song, Catal. Today, 2003, 86, 211–263.
4. K. K. Sarda, A. Bhandari, K. K. Pant and S. Jain, Fuel, 2012, 93, 86–91.
5. A. J. Hernández-Maldonado and R. T. Yang, Ind. Eng. Chem. Res., 2004, 43, 1081–1089.
6. C. Yu, J. S. Qiu, Y. F. Sun, X. H. Li, G. Chen and Z. B. Zhao, J. Porous Mater., 2008, 15, 151–157.
7. B. K. Jung and S. H. Jhung, Fuel, 2015, 145, 249–255.
8. J. Xiong, L. Yang, Y. Chao, J. Pang, P. Wu, M. Zhang, W. Zhu and H. Li, Green Chem., 2016, 18, 3040–3047.
9. J. Xiong, W. Zhu, H. Li, W. Ding, Y. Chao, P. Wu, S. Xun, M. Zhang and H. Li, Green Chem., 2015, 17, 1647–1656.
10. P. Wu, W. Zhu, Y. Chao, J. Zhang, P. Zhang, H. Zhu, C. Li, Z. Chen, H. Li and S. Dai, Chem. Commun., 2016, 52, 144–147.
11. A. H. M. Shahadat Hussain and B. J. Tatarchuk, Fuel, 2013, 107, 465–473.
12. L. Lv, J. Zhang, C. Huang, Z. Lei and B. Chen, Sep. Purif. Technol., 2014, 125, 247–255.
13. B. Shen, Z. Qin, X. Gao, F. Lin, S. Zhou, W. Shen, B. Wang, H. Zhao and H. Liu, Chin. J. Catal., 2012, 33, 152–163.
14. S. Huang, P. Chen, B. Yan, S. Wang, Y. Shen and X. Ma, Ind. Eng. Chem. Res., 2013, 52, 6349–6356.
15. J. Shan, X. Liu, L. Sun and R. Cui, Energy Fuels, 2008, 22, 3955–3959.
16. Y. Shi, X. Yang, F. Tian, C. Jia and Y. Chen, J. Nat. Gas Chem., 2012, 21, 421–425.
17. H. Wang, L. Song, H. Jiang, J. Xu, L. Jin, X. Zhang and Z. Sun, Fuel Process. Technol., 2009, 90, 835–838.
18. Y. Shi, W. Zhang, H. Zhang, F. Tian, C. Jia and Y. Chen, Fuel Process. Technol., 2013, 110, 24–32.
19. A. J. Hernández-Maldonado and R. T. Yang, Ind. Eng. Chem. Res., 2004, 43, 1081–1089.
20. S. Velu, X. Ma and C. Song, Ind. Eng. Chem. Res., 2003, 42, 5293–5304.
21. F. Subhan and B. S. Liu, Chem. Eng. J., 2011, 178, 69–77.
22. Z. Zhang, D. Liu, X. Zhu, H. Yu, S. Liu and L. Xu, J. Nat. Gas Chem., 2008, 17, 45–50.
23. M. Lesiak, M. Binczarski, S. Karski, W. Maniukiewicz, J. Rogowski, E. Szubiakiewicz, J. Berlowska, P. Dziugan and I. Witońska, J. Mol. Catal. A: Chem., 2014, 395, 337–348.
24. Z. Wang, L. Yang, R. Zhang, L. Li, Z. Cheng and Z. Zhou, Catal. Today, 2016, 264, 37–43.
25. G. Ertl, R. Hierl, H. Knözinger, N. Thiele and H. P. Urbach, Appl. Surf. Sci., 1980, 5, 49–64.
26. N. W. Cheung, P. J. Grunthaner, F. J. Grunthaner, J. W. Mayer and B. M. Ullrich, J. Vac. Sci. Technol., 1981, 18, 917.
27. D. Yi, H. Huang, X. Meng and L. Shi, Appl. Catal., B, 2014, 148–149, 377–386.
28. H. Wang, L. Song, H. Jiang, J. Xu, L. Jin, X. Zhang and Z. Sun, Fuel Process. Technol., 2009, 90, 835–838.
29. 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.
30. Y. Qin, Z. Mo, W. Yu, S. Dong, L. Duan, X. Gao and L. Song, Appl. Surf. Sci., 2014, 292, 5–15.
31. C. L. Garcia and J. A. Lercher, J. Phys. Chem. C, 1992, 96, 2669–2675.
32. L. Wang, B. Sun, F. H. Yang and R. T. Yang, Chem. Eng. Sci., 2012, 73, 208–217.
33. Y. Qin, Z. Mo, W. Yu, S. Dong, L. Duan, X. Gao and L. Song, Appl. Surf. Sci., 2014, 292, 5–15.
34. S. Velu, C. Song, M. H. Engelhard and Y. Chin, Ind. Eng. Chem. Res., 2005, 44, 5740–5749.
35. H. Sun, L. Sun, F. Li and L. Zhang, Fuel Process. Technol., 2015, 134, 284–289.

---

Footnote

† These authors contributed equally to this work.

---