1 Introduction

Serving as important chemical intermediates, toluene (T) and para-xylene (PX) have been broadly employed in industries [1,2,3,4], especially T has not kept up with demand in northwest of China. However, the traditional way to produce T and PX, including catalytic reforming and petroleum cracking, greatly depends on the consumption of petroleum and has been called into question due to the shortage of petroleum resources [5, 6]. Currently, the alkylation of benzene is receiving increased attention in the petrochemical industry due to its nature as a good alternative to produce toluene and xylene from coal and natural gas [1, 7]. Meanwhile, employing conventional ZSM-5 zeolites as the catalyst for the alkylation of benzene with methanol that could inevitably form OX, TMB and EB and the difficult separation of OX and EB from C8 aromatics due to the close boiling points between EB and xylene remain major challenges [8]. The process of benzene alkylation with methanol would be more commercially feasible if high suppression of the formation of OX, TMB and EB can be achieved.

Numerous efforts have been made to suppress the formation of TMB, OX and EB by optimizing the reaction conditions or changing catalysts. Hu et al. reported that EB formation could be highly suppressed by changing the Si/Al ratio of the catalyst [9]. Tan et al. reported that the multiple modification of ZSM-5 with SiO2, P2O5, and MgO in a suitable sequence could efficiently eliminate external surface acid sites and enhance the selectivity of PX [10]. Adebajo et al. introduced Na into ZSM-5 by ion exchange and observed that ZSM-5 with low Brønsted acidity could minimize the occurrence of such a reaction [11]. Meanwhile, great efforts have been made to suppress the side reaction of MTO [12, 13]. Unfortunately, few studies have demonstrated a significant decrease in the selectivity of OX, EB and TMB. Other researchers [14, 15] report that ethylene is the key reactant for the formation of EB and decreasing its content is a possible way to tackle the problem of EB.

Recently, researchers found that Pt-modified ZSM-5 could exhibit high suppression capability toward the formation of EB in benzene alkylation with methanol, which effectively avoided the difficulty of separating EB from xylene [16]. Nevertheless, the price of Pt is eight times than the price of Ru and the selectivity of OX and TMB wasn’t showed. In this study, we showed that a new modified agent of zeolites catalyst overturned these weaknesses.

In this paper, the catalytic performance of Ru-modified hierarchical ZSM-5 in benzene alkylation with methanol was investigated. The structural and textural properties of the catalysts were systematically characterized by different techniques, and the selectivity of OX, EB and TMB products versus Ru/HZSM were also investigated. We observed that Ru-modified hierarchical ZSM-5 catalysts could retain excellent catalytic activity of hierarchical ZSM-5, extend the live time of the catalyst, reduce the selectivity of OX and TMB and significantly suppress the formation of EB.

2 Experimental Procedure

2.1 Catalyst Preparation

Hierarchical ZSM-5 was purchased from the Catalyst Plant of Lanzhou (China). (NH4)2CO3, RuCl3·3H2O, benzene and methanol were purchased from Lanzhou Chemical Reagents Co. and used without further purification. The catalysts were prepared by the following procedures:

  1. (1)

    ZSM-5 sample was first crushed and then calcined in air at 550 °C for 4 h. Then, 10.0 g ZSM-5 calcined sample was disposed with 1.0 mol/L (NH4)2CO3 at ambient temperature for 4 h before drying at 110 °C overnight (repeat the above steps three times). Finally solid phase was calcined again at 550 °C for 4 h in air to obtain HZSM-5.

  2. (2)

    ZSM-5 sample was initially crushed and then calcined in air at 550 °C for 4 h. ZSM-5 sample was disposed by applying the incipient impregnation using 11.7 mmol/L RuCl3·3H2O aqueous solution at room temperature for 1 h. Then 1.0 mol/L (NH4)2CO3 solution was dropwise added. After stirring for 24 h, the resulting slurry was filtered to obtain a grey cake and colourless filtrate, confirming the complete precipitation of Ru as Ru2(CO3)3. The grey cake was washed with distilled water, and dried overnight at 383 K. The sample was heated directly in a N2 stream to 773 K and reduced in flowing H2 at 773 K for 3 h with a fixed-bed reactor to obtain Ru/HZSM-5 (2.5 wt%) by one-step method (Ion exchange of ZSM-5 and preparation of Ru/HZSM-5 Catalysts).

2.2 Catalyst Characterization

The X-ray diffraction (XRD) patterns of catalysts were obtained by a D/MAX-2400 X-ray diffractometer equipped with atmosphere and temperature control stages and using Cu-Kα radiation (λ = 1.542 Å) operated at 40 kV and 50 mA. The catalysts were characterized using transmission electron microscopy (TEM) (JEM-6700) and scanning electron microscopy (SEM) (JSM-6701F). FT-IR of the materials were acquired on a Nicolet Nexus 670 FT-IR spectrometer (Thermo Nicolet, USA) at room temperature in KBr pellets over the range of 7400–350 cm−1 under the atmospheric conditions. The Zeta potential of samples was characterized using Powereach JS94H miniature electrophoresis apparatus.

2.3 Thermodynamics Analysis

The process of alkylation reaction of benzene/methanol was studied thermodynamically based on functional group contribution method of Benson, and the reaction heat ΔHT, Gibbs energy ΔGT, Entropy Changes ΔST and equilibrium constant Kp about major and minor reactions were calculated. The thermodynamic properties of the material involved are derived from data manual [17].

The reaction heat ΔHT, Gibbs energy ΔGT, Entropy Changes ΔST and equilibrium constant Kp are calculated from below.

$$C_{p}^{0}=a+bT+c{T^2}+d{T^3}$$
(1)
$$\Delta {C_p}=\Delta a+\Delta bT+\Delta c{T^2}+\Delta d{T^3}$$
(2)
$$\Delta {H_T}=\Delta H_{{298}}^{\theta }+\mathop \smallint \limits_{{298}}^{T} \Delta {C_p}dT$$
(3)
$$\Delta {S_T}=\Delta S_{{298}}^{\theta }+\mathop \smallint \limits_{{298}}^{T} \frac{{\Delta {C_p}}}{T}dT$$
(4)
$$\Delta {G_T}=\Delta {H_T} - T\Delta {S_T}$$
(5)
$${K_p}=\exp \left( { - {\raise0.7ex\hbox{${\Delta {G_T}}$} \!\mathord{\left/ {\vphantom {{\Delta {G_T}} {RT}}}\right.\kern-0pt}\!\lower0.7ex\hbox{${RT}$}}} \right)$$
(6)
$$R=8.314J/({\text{mol}} \cdot K)$$

2.4 Catalytic Activity Tests

The catalytic activity of the catalysts was measured in a continuous flow jet bubbling reactor with a glass tube (8 mm i.d.) at atmospheric pressure. The catalyst (0.4 g) was heat-treated in situ from ambient temperature to 673 K at 10 K/min and maintained at 673 K for 15 min in N2 flow before the reaction. Then, benzene and methanol mixture (B/M molar ratio = 1:1) was fed into the reactor (WHSV = 2.0 h−1), and the collected products from the reactor were analyzed by gas chromatography (Fuli GC9790). To ensure all the products were in the gas phase, the temperature of the effluent line was maintained at 200 °C by heating belt.

3 Results and Discussion

3.1 Catalyst Characterizations

Figure 1 presents the XRD patterns of HZSM-5 and Ru-HZSM-5 catalysts. Both unmodified and modified catalysts displayed two distinct diffraction peaks in 8°–10° and 20°–25° (2θ) ranges, which corresponded to the typical characteristic pattern of MFI structure [18,19,20]. The XRD patterns of Ru/HZSM-5 catalysts reduced by H2 did not show any characteristic diffraction peak of Ru particles, which indicated that Ru particles were well dispersed on the surface of ZSM-5.

Fig. 1
figure 1

XRD patterns of hierarchical HZSM-5 and Ru/HZSM-5 catalysts

Figure 2 shows the SEM images of HZSM-5, Ru/HZSM-5, and the catalyst after the reaction (written as C-Ru/HZSM-5). ZSM-5 zeolites had rectangular brick-type crystals with smooth angular surfaces, a long-axis length of approximately 6 µm, and a short-axis length of 2 µm. Both Ru/HZSM-5 catalyst and C-Ru/HZSM-5 exhibited an anomalistic rod-like shape with rough surfaces and ill-defined particles. This morphology originated from the (NH4)2CO3 treatment to achieve partial desilication [21], which decreased the crystallinity. This finding agreed well with the XRD patterns of Ru/HZSM-5. A high-magnification TEM image (Fig. 2 d, e) indicated that the particle size of Ru was made of nanoparticles ranging from 0.5 to 2 nm, which were dispersed faultlessly on HZSM-5. However, black mulch was discovered in the TEM image of C-Ru/HZSM-5, which was regarded as coke.

Fig. 2
figure 2

In order, SEM image of HZSM-5 (a), Ru/HZSM-5 (b), and C-Ru/HZSM-5 (c); TEM image of Ru/HZSM-5 (d) and C-Ru/HZSM-5 (e)

Figure 3 shows the FTIR spectra ranging from 400 to 4000 cm−1 of HZSM-5, Ru/HZSM-5 and C-Ru/HZSM-5. The comparison and explanation of HZSM-5, Ru/HZSM-5 and C-Ru/HZSM-5 samples about FTIR spectra are shown in Table 1.

Fig. 3
figure 3

The FT-IR spectra of HZSM-5, Ru/HZSM-5 and C-Ru/HZSM-5 catalysts

Table 1 Comparison and explanation of HZSM-5, Ru/HZSM-5 and C-Ru/HZSM-5 samples about FT-IR spectra

Skeleton changes of HZSM-5, Ru/HZSM-5 and C-Ru/HZSM-5 are no significant change, where are mainly the vibration strength of skeleton. In contrast, the intensity of the (Si, Al)–OH tetrahedron at 1111 cm−1, the stretching vibration of –OH groups at 1634 cm−1 and 3420 cm−1 for Ru/HZSM-5 are increased and that of C-Ru/HZSM-5 are decreased to a large extent due to coke. Meanwhile, the main band (832 cm−1) is slightly shifted to the low band (804 cm−1), which are caused by (NH4)2CO3 treatment to achieve partial desilication [21]. As is well-known the length of the Si–OH (0.161 nm) is less than Al–OH (0.175 nm) which changes the structure parameters of HZSM-5. Therefore, the Si atoms of HZSM-5 are gradually removed which result in the decrease of the force constant and the (Si, Al)–OH groups are shifted to the low bands. In conclusion, the above analysis shows that the skeletal structure of HZSM-5 has not much changed.

3.2 Thermodynamics Analysis for Alkylation of Benzene by Methanol

There may be many reactions in the process of benzene and methanol alkylation. The main reaction and possible secondary reaction is as follows.

  1. (1)

    Main alkylation reaction:

    figure b
  2. (2)

    Methanol self reaction:

    figure c
  3. (3)

    Other alkylation reactions:

    figure d
  4. (4)

    Carbon deposition reaction

    figure e

The thermodynamic data are shown in Table 2.

Table 2 The reaction heat ΔHT, Gibbs energy ΔGT, Entropy Changes ΔST and equilibrium constant Kp

Table 2 presents that enthalpy changes ΔHT of each reaction is small in the range of 623–723 K, and even considered as constant to some extent. Among the reaction of benzene/methanol alkylation, all the reactions are exothermic except for the coke deposition reaction. The heat releottom of four main alkylation reactions [(1)–(4)] are remarkably close, but the formation reaction of TMB, tetra-methyl-benzene (TTMB), propylene and EB were strongly exothermic. The greater the carbon number of olefins produced by methanol itself, the greater the heat releottom. Therefore, the temperature of alkylation reaction can’t be too high. High temperature will be unfavorable to the main reactions and the coke reaction rate will be greatly accelerated. Four main reactions, the formation reaction of EB and MTO are entropy increasing reaction, the reaction of the rest are entropy decreasing.

The four primary reactions can occur spontaneously in the range of 623–723 K because Gibbs energy ΔGT < 0. EB, TMB, TTMB can also occur spontaneously. Olefin especially ethylene (10) is more likely to happen than main reactions because Gibbs energy ΔGT is smaller, which illustrate the formation reaction rate of ethylene is higher than the main reactions. Therefore, MTO is mainly competing reaction. EB, TMB and TTMB have a lower degree of reaction than the main reaction. And the formation reaction of propyl-benzene, isopropyl benzene and ethyl-4-methylbenzene can’t be proceeded spontaneously at this temperature because Gibbs energy ΔGT > 0.

According to the relationship between Gibbs energy ΔGT and equilibrium constant Kp, the value of the reactions [(1)–(4)] is very close, indicating that the main reaction is irreversible. The equilibrium constant of propyl-benzene, isopropyl benzene and ethyl-4-methylbenzene is < 1.5 or even negative, indicating that high temperature can inhibit their generation. The equilibrium constant Kp of EB is more than 10. The equilibrium constant Kp of TMB, TTMB is < 10, but is large relatively compared with others reaction. Results show that TMB, EB, TTMB are the main byproducts in benzene/methanol alkylation. The equilibrium constant Kp of ethylene constant from self-methanol reaction is much larger than 0, indicating that methanol can spontaneously generate alkene. Other researchers [14, 15] report that ethylene is the key reactant for the formation of EB and decreasing its content is a possible way to tackle the problem of EB.

In conclusion, the selectivity of T and PX are enhanced, the critical point lies in the fact that:

  1. (1)

    MTO is mainly competing reaction and must be suppressed.

  2. (2)

    DTMB>DMX>DOX>DPX: limited by catalyst hole size.

  3. (3)

    The temperature can’t be too high.

The experiment indicates that Ru/HZSM-5 will be the best choice, and the evaluation of experimental performance is described below.

3.3 Catalytic Performance of Ru Catalysts

As shown in Table 3, the conversion of benzene and the selectivity of T, PX, OX, EB, TMB, etc. over hierarchical Ru/HZSM-5 (48.3% and 79.2%, 16.9%, 1.1%, 0%, 0.9%, respectively) are apparently different from HZSM-5 catalysts (47.6% and 76.3%, 15.8%, 2.3%, 1.3%, 2.1%, respectively), thereby indicating that two kinds of catalyst activity have obviously change before and after Ru loaded. For Pt/ZSM-5 catalyst [16], the contents of EB in N2 atmosphere was 3.4%, and the selectivity of OX was not divided from the total selectivity of xylene. The decrease of OX, TMB selectivity indicated that the existence of Ru nanoparticles could hinder the generation of methyl over the ortho-position of toluene may be because the existence of Ru nanoparticles diminished the average pore size of the catalyst (molecular dimension: DMX>DOX>DPX) (Fig. 4), the methyl group was o-position and p-position orienting group and the lnKp value (lnKp(PX) = 12.66 ≈ lnKp(OX) = 12.66 > lnKp(MX) = 11.25) calculated by the thermodynamics at 673 K on Table 2. The selectivity of OX dropped from 2.3 to 1.1%, the selectivity of TMB dropped from 2.1 to 0.9% and the selectivity to EB dropped from 1.34 to 0%. Thus OX, TMB, EB were suppressed indirectly by Ru nanoparticles.

Table 3 Product content of benzene alkylation with methanol over different catalysts
Fig. 4
figure 4

Size of main organic molecules and the pore distribution of catalysts

The reasons can be explained as follows:

It is well known that HZSM-5 had weak adsorption capacity for aromatic and olefin compared with other acidic zeolites. Thus, it had high stability and low coking rate. Ru addition increased the adsorption properties of zeolites. In addition, the more the gas chemical properties were active, the easier gas was adsorbed [22]. In the environment of acid zeolites and N2, the active sequence was methanol > ethylene > benzene. Methanol was activated to become CH3+OH, and benzene was activated by the acid sites that meet to produce toluene. Coke deposition was mainly caused by the side effects of self-methanol. This conclusion can be evaluated by testing with only methanol or only benzene as raw material. (1) With methanol as a reactant, the coke deposition rate over HZSM-5 was faster than the rate over Ru/HZSM-5. Based on HZSM-5 catalyst, methanol to olefins and olefins stayed a very short time on the surface of HZSM-5, which generated more ethylene. Furthermore, EB and coke deposition are increased indirectly at the same time. Ethylene formed a strong adsorption bond over Ru/HZSM-5, which inhibited the generation of ethylene because leaving from micro-volume was difficult. Meanwhile, the coke rate of olefins was much higher than that of olefin alkylation. Therefore, EB was not produced. (2) Benzene as a reactant: the coke rate of Ru/HZSM-5 was less than that of the catalytic system of HZSM-5. However, the coke rate of only methanol as a reactant was much higher than that of only benzene.

According to the theory of valence bond [23], the greater d% of metal bonds and have more electrons, thus adsorption heat is lower and easier adsorption. Since d% (Pt) = 44 and d% (Ru) = 50, Pt/ZSM-5 had weak adsorption capacity for methanol, olefin and benzene compared with Ru/ZSM-5. This conclusion can be evaluated by testing with Powereach JS94H (Fig. 5; Table 4). The order of Zeta potential was Ru/HZSM-5 (-57.27) < C-Ru/HZSM-5 (-38.83) < HZSM-5 (-32.78) in Table 4 and declare that the catalyst after the reaction (150 h) still has good catalytic performance (Fig. 5). Whereas olefin could cause more EBs and coke (Table 5) because leaving from Pt-micro-reactor to acid active sites, which was easier further reaction over Pt/ZSM-5 [2]. In a word, the existence of Ru would help to suppress efficiently the formation of EB and reduce OX, TMB. The results are interesting and lead to future scientific work.

Fig. 5
figure 5

Surface electrical behavior of HZSM-5 and Ru/HZSM-5

Table 4 Zeta potential of HZSM-5, Ru/HZSM-5 and C-Ru/HZSM-5
Table 5 Deactivation parameter (λ) of HZSM-5 and Ru/ZSM-5 catalysts

3.4 Catalyst Stability

Catalyst stability was shown clearly in Fig. 6. The conversion of benzene over HZSM-5 catalyst increased gradually to 55.7% during the first 10 h. Moreover, the conversion of benzene over HZSM-5 catalyst dropped down to 39.4% after 30 h. However, the conversion of benzene over Ru/HZSM-5 decreased to 47.6% after 150 h. Therefore, Ru/ZSM-5 catalyst was more stable than HZSM-5 in benzene alkylation with methanol. As long as H elements in organic matter, the metal surface retains its catalytic activity. However, catalytic activity becomes inactive when the surface undergoes graphitization. In addition, the amount of coke deposit (removed within 200–800 °C) of C-Ru/HZSM-5 is calculated. The deactivation parameter (λ) along the reaction time was defined as: 100 × (χe − χs)/χs and Ru/HZSM-5 catalyst had a much smaller λ value than that of HZSM-5 catalyst. However, Pt/ZSM-5 catalyst had a little larger λ value than that of Ru/HZSM-5 catalyst and the inference above was further proved.

Fig. 6
figure 6

Stability of HZSM-5 and Ru/HZSM-5 in benzene alkylation with methanol

It is well known that coke deposition was mainly caused by the side effects of self-methanol (olefin). By analyzing the gas phase products of the alkylation reaction over HZSM-5 and Ru/HZSM-5 catalysts in Fig. 7, complete reaction of residual methanol form certainly gaseous substances (mainly ethylene and propylene). During the reaction test, the amount of ethylene and propylene were plentiful over HZSM-5 catalyst. By contrast, the amount of ethylene and propylene were almost none over Ru/HZSM-5 catalyst. And as the reaction went on, the amount of ethylene and propylene was not detected, but the quantity of methanol was increased. Therefore, the stability of the alkylation catalyst was enhanced after the modification.

Fig. 7
figure 7

Gas phase composition of HZSM-5 and Ru/HZSM-5

4 Conclusions

Results show that OX, EB and TMB are the main byproduct in alkylation of benzene by methanol. We have suited the remedy to the case. Hierarchical porous ZSM-5 catalysts with Ru were prepared via chemical precipitation and H2 reduction. High-magnification TEM image indicated that Ru particles were well dispersed on the surface of HZSM-5. Experimental data displayed that Ru/ZSM-5 catalyst could perfectly suppress EB formation by suppressing the side reaction of methanol. Meanwhile, the decrease of OX and TMB selectivity indicated that the existence of Ru nanoparticles could hinder the generation of methyl over the ortho-position of toluene. Most importantly, Ru/ZSM-5 catalyst was more stable than HZSM-5 in benzene alkylation with methanol possibly due to the presence of H elements in organic matter, which allows metal surface to retain its catalytic activity.