In recent decades, and especially in the last 10 years, gold catalysts have become a vital new force in the field of green chemistry [1] owing to the explosive growth of research and development in the fields of pollution control, fine chemical synthesis and energy.From simple reactions, such as CO oxidation and epoxidation of propylene [2, 3], the studied reaction systems have been extended to hydrogenation, carbonylation, condensation and other kinds of reactions in organic synthesis [4-6].The research category has expanded from heterogeneous catalysis [7] to homogeneous catalysis [8] and photocatalysis [9].Studies of many different reaction systems over Au catalysts have shown that compared with other precious metals, Au catalysts were characterized by some particular features, such as a relatively high catalytic activity at low temperature [2, 4], high selectivity of the target product [3-6] and a simple route for the green synthesis of chemicals [4-6].However, most of the studied reactions have been performed under mild reaction conditions; the investigation of gold catalysts in relatively harsh catalytic reaction systems, such as catalytic cracking of heavy oil at temperatures higher than 400 ℃ with numerous reactants, has rarely been reported.
Propylene, as an important organic synthetic raw material, can be used to produce polypropylene, acrylonitrile, propylene oxide, isopropyl benzene, isopropyl alcohol and other industrial raw materials.The consumption of propylene worldwide has increased dramatically owing to market demands for the downstream derivatives (mainly polypropylene).The rate at which the consumption of propylene has grown in recent years is faster than that of ethylene.Propylene is mostly produced by steam cracking and fluidized catalytic cracking (FCC) as well as dehydrogenation of propane.Because of lower power consumption, cheaper materials and adaptable devices, the FCC process has played an increasingly important role in the production of propylene, whereas the ratio of propylene made from steam cracking has gradually decreased [10-12].The propylene yield obtained with the FCC process can be increased significantly by using high quality raw materials, novel selective catalysts/additives and by optimizing the process [13].
ZSM-5 are the most widely used active components in catalytic cracking catalysts to increase the propylene selectivity owing to their special pore structure with shape selectivity, strong acidity, low hydrogen transfer activity and hydrothermal stability [10, 11, 14].The modification of ZSM-5 is one of the main methods used to develop more active catalysts and increase the propylene yield.To further increase the accessibility of the active sites of ZSM-5, the study on nanometer ZSM-5 [15, 16], single [17] or meso-and microporous composite ZSM-5 [18], and mesoporous ZSM-5 [19] has attracted attention.However, the major method for modification of ZSM-5 is through the addition of different elements including electronegative elements such as P and B, alkali and alkaline earth elements such as Mg, Ca, Ba, K and Na, other rare earth elements such as La and Ce, and transition metal elements such as Cu, W, Pt, Fe and Ti [17, 20].
The modified ZSM-5 that is currently available has been effective for the production of propylene by catalytic cracking of C4 hydrocarbons and pyrolysis of naphtha and as an additive in the FCC process.However, the working temperatures are relatively high ( > 510 ℃) and the capability to produce propylene is greatly reduced if the reaction temperature is reduced [10, 14, 17, 19, 21, 22].Nanoparticles, clusters and complexes of gold have widely been used to catalyze different types of reactions and have significant catalytic activity under mild reaction conditions [2-9].Can gold be used to reduce the reaction temperature of FCC and maintain the propylene yield?
Incipient wetness impregnation [23], ion exchange [24, 25], chemical vapor deposition [26] and deposition precipitation (DP) [27] have been used to deposit gold into ZSM-5.Good thermal stability of Au/ZSM-5 catalyst at temperatures as high as 800 ℃ has also been reported [28], indicating the possible application of Au/ZSM-5 in the FCC process.In the present study, Au/ZSM-5 catalyst samples were prepared using a modified DP method described previously [29, 30].We investigated their selectivity for the synthesis of propylene and the micro-activity index in the catalytic cracking of light diesel oil (235-337 ℃) at a relatively low temperature of 460 ℃.The modification of ZSM-5 by a small amount of gold had a positive effect on the production of propylene and maintained or even enhanced the micro-activity.
The received ZSM-5 zeolite (Si/Al=35, surface area 281.9 m2/g) was calcined at 540 ℃ in air for 2 h prior to further application.
First, to investigate the impact of the preparation parameters, four Au/ZSM-5 catalysts, denoted as Au/ZSM-5-a, Au/ZSM-5-b, Au/ZSM-5-c and Au/ZSM-5-d, were prepared by using our modified DP method described previously [29, 30].Briefly, ZSM-5 was first impregnated using a given concentration of HAuCl4 solution at pH=9, which was adjusted using a KOH solution.The nominal Au loading was 1.0 wt%.Au/ZSM-5-a was obtained by reducing the impregnated precursor in-situ using NaBH4.Au/ZSM-5-b was obtained by soaking the impregnated precursor in an ammonia solution at pH=10 for 24 h, washing three times followed by reduction in situ using NaBH4.Au/ZSM-5-c was obtained by only drying the impregnated precursor at 110 ℃ for 2 h.Au/ZSM-5-d was obtained by soaking the impregnated precursor in an ammonia solution at pH=10 for 24 h followed by filtration and washing three times with water, as reported previously [29, 30].All four catalysts precursors were calcined at 300 ℃ for 2 h.
Secondly, to investigate the impact of Au loading, the Au/ZSM-5-d catalysts with theoretical Au contents of 0.5, 0.8, 1.2 and 1.5 wt% except for 1.0 wt% were also prepared by above described procedure.The obtained catalysts were designated as x% Au /ZSM-5-d, in which x% is the nominal quantity percentage of Au.
Because the micro-activity of fresh FCC catalysts is high and unstable, which would not reflect the real situation in a practical application, hydrothermal treatment is usually adopted prior to micro-activity testing to deactivate the fresh catalyst.Therefore, all the samples including the parent ZSM-5 were aged at 800 ℃ in flowing water steam for 4 h and then calcined at 300 ℃ for 1 h.
A Shimadzu XRD-6100 diffractometer using Cu Kαradiation (40 kV, 30 mA) was employed to obtain X-ray diffraction (XRD) patterns of all studied samples with a scanning range of 2θ from 10 to 80° at a scanning rate of 6°/min.A NOVA-3000E (Quantachrome Corp.) was used to obtain N2 adsorption-desorption isotherms at −196 ℃.Prior to analysis, each sample was evacuated at 120 ℃ for 4 h.
A Tp-5080 multi-functional automatic adsorption instrument (Tianjin Xianquan Industry and Trade Development Co., LTD, Tianjin, China) was used to take temperature- programmed desorption of NH3(NH3-TPD) curves for all studied catalysts.The sample was loaded into a stainless U-shaped microreactor and pretreated at 600 ℃ in flowing He stream (20 mL/min) for 0.5 h before it was saturated with NH3 gas at low temperature.
A transmission electron microscope (TEM, JEOL JEM-2100, JEOL Company, Tokyo, Japan) equipped with an Oxford INCA spectrometer for X-ray energy dispersive (EDX) spectroscopy analysis was used at 200 kV accelerating voltage to evaluate the dispersion properties of two representative Au/ZSM-5 samples.
An IRIS IIXSP inductively coupled plasma spectrometer (ICP-AES, Thermo Electron Corporation, USA) was used to measure the quantity of gold in the samples at working conditions of 27 MHz and 112 kW.A hydrofluoric acid solution and aqua regia were used to dissolve the samples.
Micro-activity tests of the catalysts were obtained using Dagang light diesel (235-337 ℃) as the feedstock in a small confined fluidized bed at 460 ℃.Three grams of catalyst was placed into a self-constructed reaction device (Fig. 1) and the oil was injected into the furnace by a syringe when the temperature reached a preset temperature.The weight ratio of the catalyst to oil and the contact time were 3.2 and 70 s, respectively.The products were purged by using high-purity nitrogen after the completion of the feed and separated through an ice-water bath.
The gaseous products were determined by an on-line GC-920X with a flame ionization detector equipped with an OV-101 capillary column.The liquid products were analyzed by manual sampling after collection in a glass container.The composition of gasoline and diesel in the liquid phase was determined at the retention time of n-dodecane (Fig. 2).
The area normalization method was applied to collect the data.Propylene selectivity was defined by the content of propylene in the gaseous products.The micro-activity index (MAT) and conversion of the oil were calculated according to the RIPP92-90 method using equations (1) and (2):
where D0% refers to the proportion of diesel components in the light diesel oil feed with hydrocarbons heavier than n-dodecane; W and W1 refer to the quality of the liquids before and after reaction; D% and G% refer to relative content of diesel and gasoline, respectively, in the liquid product.At least, two tests were performed to evaluate the catalysts and the mean values were taken with the absolute difference between the two parallel results less than 2%.
The results of the catalytic cracking of light diesel over ZSM-5 with or without gold modification are listed in Table 1, including the actual Au loading and the ratio of weak acid sites to strong acid sites, which were based on our NH3-TPD measurements.In comparison with the parent ZSM-5 catalyst, it is clear that the introduction of gold into ZSM-5 improved the propylene selectivity with the exception of the Au/ZSM-5-a catalyst.However, the catalytic cracking activity represented by a micro-activity index was only enhanced over the three catalysts of Au/ZSM-5-d with theoretical Au loadings of 0.5, 0.8 and 1.0 wt%.The corresponding increase in propylene selectivity over these three catalysts was 8.8%, 2.9% and 23.2%, whereas the increase in their micro-activity index (MAT) was 7.1%, 4.3% and 4.5%, respectively.The results suggest that a small amount of gold has a positive effect on the catalytic cracking of light diesel oil to increase the production of propylene at a lower reaction temperature of 460 ℃.
Recently, Guo et al.[31, 32] found that a series of Si/Al molecular sieves deposited with gold showed higher activity and a lower reaction temperature than those without gold modification for aromatization, isomerization and alkylation reactions and catalytic cracking of C3-C9 hydrocarbons.This further supports the potential application of gold in the field of petroleum chemistry.
Fig. 3 shows the XRD patterns of ZSM-5 and the Au/ZSM-5 catalysts.In comparison with the parent ZSM-5, the patterns of Au/ZSM-5-b, c, and d seem to maintain the same structure, but the intensity decreases gradually from Au/ZSM-5-d to Au/ZSM-5-c and then Au/ZSM-5-b.Conversely, Au-ZSM-a completely loses the structure of the original pattern.The Au/ZSM-5-d sample exhibits the typical peaks of ZSM-5 with similar intensities, indicating that the structure remains intact after the modification.The XRD patterns of Au/ZSM-5-d with various Au loadings of 0.5%-1.5% (not shown) further support the maintenance of the ZSM-5 structure through the d-type preparation procedure.
For all samples with gold modification, the characteristic peaks corresponding to Au (111) and (200) crystal planes are detected at 2θ =38.2° and 44.4°, which means that the Au particles are not as small as those that we previously prepared over an alumina support [29, 30].The mean size of the Au particles on alumina was in the range of 2-3 nm.The diameters of the Au particles estimated by the Scherer formula are approximately 20 nm.This value is quite rough because the intensities of Au peaks are not very visible and vary with different samples.
To investigate the dispersion of gold particles in the ZSM-5, two samples of 0.5 and 1.5 wt% Au/ZSM-5-d were selected to make TEM observations.The representative TEM images of the 1.5 wt% Au/ZSM-5-d sample are displayed in Fig. 4.The size distribution of the Au particles varied in a rather large range from < 10 to < 200 nm regardless of the Au loadings.The deposition of Au particles was very inhomogeneous, with a few Au particles on some parts of the ZSM-5 substrate whereas on other parts tacking numbers of Au particles.A few small Au particles, which are not easy to distinguish, are indicated by red arrows in the images.
From these randomly dispersed Au particles, we can see that some Au particles form a flat shape and show a stronger interaction with the substrate.Furthermore, some smaller Au particles may be embedded in slit-pores between the flaky ZSM-5 particles.Clearer images of this are displayed in Fig. 5, which shows two TEM images of 1.5 wt% Au/ZSM-5-d before and after it is aged in steam at 800 ℃.These cases are surely not popular for the studied samples with gold modification; however, it may account for the contribution of these small Au particles on the pore size distribution of the parent ZSM-5 substrate and hence exert some impact on the catalytic cracking performance.
Fig. 6 shows the N2 adsorption-desorption isotherms of ZSM-5 and the Au/ZSM-5-a, b, c, d catalysts.The isotherms of ZSM-5 and Au/ZSM-5-c and Au/ZSM-5-d present a type I isotherm with a H4 hysteresis loop, which is typical for microporous materials with some slit-like pores and mesopores.The capillary condensation phenomenon at lower relative pressure range of 0.2-0.3 for the sample of Au/ZSM-5-c indicates the presence some disordered pores.The corresponding pore size distributions confirmed that the majority of these pores were approximately 2 nm, as shown in Fig. 7 for the pore size distributions of ZSM-5 and 1.0 wt% Au/ZSM-5-d as well as other samples.
Conversely, the adsorbed amount of liquid nitrogen on Au/ZSM-5-a and Au/ZSM-5-b is quite low, showing a type III isotherm with a H3 hysteresis loop.These are nonporous materials that have larger slit-like pores among stacked flaky particles.The XRD patterns indicate that Au/ZSM-5-a and Au/ZSM-5-b almost lose the original MFI structure of the ZSM-5.This may explain their lower selectivity for propylene production, as shown in Table 1.
N2 adsorption-desorption isotherms and pore size distribution curves of ZSM-5 and five Au/ZSM-5-d samples with various Au loadings are displayed in Fig. 7.Similar to that of the parent ZSM-5, the isotherms of all the samples containing Au are typical type I isotherms with a H4 hysteresis loop, which means that these microporous materials have a combination of some slit-like pores and mesopores.The difference between them is reflected, to some extent, in their pore size distribution curves.The Au/ZSM-5-d sample with 0.5 wt% nominal Au loading showed the narrowest pore size distribution, followed by the sample with 1.0 wt% Au loading.The sample with 0.8 wt% Au showed a similar pore size distribution range to that of the parent ZSM-5, slightly wider than the sample with 1.0 wt% Au and narrower than the sample with 1.2 wt% Au.The sample with 1.5 wt% Au had the widest pore size distribution.The width of the pore size distribution could be sorted as follows: The sample with 1.5 wt% Au > > the sample with 1.2 wt% Au > the parent ZSM-5 ≈ the sample with 0.8 wt% Au > the sample with 1.0 wt% Au > the sample with 0.5 wt% Au.Notably, the three samples that have a narrower pore size distribution than the parent ZSM-5 show an enhancement in both propylene selectivity and micro-activity.
The acid sites of zeolites are often characterized by NH3-TPD.In general, there are two well resolved peaks for zeolites, which are referred to as the low temperature (LT) peak and the high temperature (HT) peak, corresponding to the total number of weak acid sites and strong acid sites, respectively.The strong acid sites contribute to the decomposition of NH4+ ions formed over strong Br nsted acid sites.As shown in Fig. 8, one clear LT peak and one small HT peak are observed.Except for Au/ZSM-5-a and Au/ZSM-5-b in which the structure is destroyed, the Au modification dramatically increases the total acid density, and especially a significant increase in the weak acid intensity with an increase in gold loading.This could be correlated to the increase in selectivity for the synthesis of propylene.However, only the three catalysts 0.5, 0.8 and 1.0 wt% Au/ZSM-5-d exhibited good performance in both propylene selectivity and micro-activity index, as shown in Table 1.It has been reported that relatively mild acidity strength and a reasonable number of strong acid sites favor propylene production [33].Taking into account the pore size distribution, the ratio of weak acid sites to strong acid sites calculated from the NH3-TPD curves (Table 1) suggests that there is a synergy between the acid property and pore selectivity in propylene production.There are no big differences in the ratio of weak to strong acid sites for the catalysts that maintain the ZSM-5 structure.
In general, some SiO2 and Al2O3 in the zeolite can be dissolved in a basic solution, leading to the loss of the strong acid sites and the partial destruction of the structure.A weak base could dissolve a small amount of silica and alumina, forming some mesopores by extending the micropores.A strong base could destroy the crystal structure of the molecular sieve and turn the molecular sieve into an amorphous non-acidic material with big holes [34].The destruction of the structure of Au/ZSM-5-a and Au/ZSM-5-b is possibly related to the use of NaBH4 in the preparation procedure, which is in agreement with the results in Ref.[35].
In this study, a KOH solution was used to adjust the pH of a HAuCl4 solution to 9 and then the dried precursor was soaked in ammonia solution at pH=10 for 24 h before the final step to prepare five Au/ZSM-5-d catalysts with nominal Au loadings of 0.5, 0.8, 1.0, 1.2 and 1.5 wt%.To evaluate the influence of the two bases used in the preparation of the catalyst on the catalytic cracking reaction, two new samples were prepared without introducing gold and tested under the same reaction conditions.The first sample was prepared by replacing the HAuCl4 solution with a HCl solution (same amount of Cl-) and adjusting the pH of the HCl solution to 9 by using a KOH solution before the impregnation of the parent ZSM-5.The dried precursor was then soaked in ammonia solution (pH=10) for 24 h followed by washing, drying and calcination.The second sample was directly soaked in the parent ZSM-5 in an ammonia solution (pH=10) for 24 h and then washed, dried and calcined.The catalytic cracking testing results at 460 ℃ together with the corresponding results on the parent ZSM-5 performed at the same time are shown in Table 2.The parent ZSM-5 used here was from another batch, which was purchased from the same company and had the same Si/Al ratio and a similar surface area to the one in Table 1.Small differences in testing results can be found over two parents of ZSM-5.
It is clear that similar propylene selectivity and micro-activity are obtained for the two base-treated samples, with an increase in the propylene selectivity and a decrease in the micro-activity in comparison with those of the parent ZSM-5.This implies that any difference in the amounts of residual K and Cl in the five Au/ZSM-5-d catalysts with various Au loadings does not significantly influence the catalytic performance under the tested conditions.Taking into account the data shown in Table 1, we can conclude that gold enhances the propylene selectivity and micro-activity of the catalytic cracking reaction of light diesel oil, although it cannot simply take a subtraction of bases effect which was involved in preparation process.
The TEM observations show that the Au particles in the studied Au/ZSM-5 catalyst are predominantly distributed on the outer surface of the molecular sieve with an uneven dispersion and size.However, the embedding of a few small Au particles inside the molecular sieve pores cannot be ruled out.As evidenced by the N2 adsorption-desorption measurement, three samples with 0.5, 0.8, and 1.0 wt% Au loadings, which have narrower pore size distribution than the parent ZSM-5, exhibit an enhancement in both propylene selectivity and micro-activity.
According to the literature, there are two processes involved in the formation of propylene: β-scission and a proteolysis mechanism on the acid sites of the zeolite [36].Generally, the metal sites affect the second forming carbenium species that can produce light olefins on Br nsted acid sites.Therefore, the Br nsted acid sites are very important for the catalytic cracking reaction.However, reports about how to increase the number of Br nsted acid sites of ZSM-5 by metal modification to promote light olefins production are rare [37].
The NH3-TPD results showed that Au modification of ZSM-5 resulted in more acid sites than in ZSM-5 itself after the process of high temperature steam treatment, indicating that gold suppressed the de-alumination to some extent.Although the Au particles dispersed randomly on the ZSM-5 substrate, some of the Au particles strongly interacted with ZSM-5 substrate by forming a flat shape.This may lead to a partial electron transfer from metallic Au to (SiAl)O(OH)m, increasing the mobility of the proton in the hydroxyl groups and thus enhancing the acid intensity.The electron transfer will benefit the carbenium ion mechanism.In addition, the Au/ZSM-5 sample has stronger total weak acid sites (LT region).The stronger acid sites give the carbenium ions a longer lifetime to undergo secondary transformations [38].
For the catalytic cracking reaction, acid catalysis is dominant.Further research is needed to confirm the above conclusions and gain a better understand of the role of nanogold in such harsh reaction conditions.
The catalytic cracking reaction of light diesel over ZSM-5 with or without gold modification was investigated.The use of NaBH4 in the preparation process destroys the MFI structure of the ZSM-5 substrate and results in worse catalytic performance.Gold modification enhances the propylene selectivity, although the contribution from the bases (KOH and ammonia) used during the preparation cannot be neglected.The simultaneous enhancement of the propylene selectivity and micro-activity index is only obtained over the Au/ZSM-5 catalysts that maintained their MFI structure and had small Au particles with a narrow pore size distribution, relatively mild weak acidity strength and a reasonable number of strong acid sites.These results suggest that a small amount of gold has a positive effect on the catalytic cracking of light diesel oil and increased propylene production at a lower reaction temperature of 460 ℃.