3.1 Characterization of the catalysts
Adsorption isotherms (Fig. 1) for MFI-based catalysts with the introduced nickel by ion exchange and impregnation are traditional ones as for microporous zeolite samples with some mesoporosity.13–17 The latter caused by partial zeolite destruction during ion exchange procedure and calcination. The obtained isotherms characterized by very narrow hysteresises. The introduction of nickel into catalysts by ion exchange, due to the presence of a wider hysteresis loop, has a greater effect on the porous structure of the sample compared to impregnation. Besides micropores up to 2 nm both samples have pores of 2.8 and 5.6 nm (Fig. 2).
Figures 3 and 4 show adsorption data for catalysts with ion-exchanged nickel, reduced under different conditions. We can see insignificant influence of the reduction method on the porous characteristics of the samples. Only a larger volume of pores with a radius of 1.2 nm can be noted in some samples with the exception of the HNa-IE1.5Ni-450 catalyst, in which these pores disappeared and instead a small number of pores with a radius of 1.8 nm appeared. Table 1 summarized calculated porous characteristics of the investigated samples. It is well known that specific surface area is very important characteristic for the effective proceeding of a range of heterogeneous processes.13,18 For all synthesized samples this parameter values are appropriate from 300 to 370 m2/g. Part of micropores in all samples are near 70–87%.
Table 1
Porous properties of Ni-containing zeolite catalysts of different preparing method
Sample
|
SBET, m2/g
|
St, m2/g
|
Smiсro, m2/g
|
VΣ, cm3/g
|
Vtmicro, cm3/g
|
Vtmicro/VΣ, %
|
RDFT, nm
|
R, nm
|
HMFI
|
371
|
14.5
|
356
|
0.18
|
0.15
|
83
|
1.0
|
1.0
|
H-Im1.5Ni
|
335
|
5.17
|
330
|
0.16
|
0.14
|
87
|
0.9
|
1.0
|
H-IE1.5Ni
|
365
|
29.9
|
336
|
0.21
|
0.15
|
71
|
0.9
|
1.2
|
HNa-IE1.5Ni
|
319
|
16.5
|
302
|
0.16
|
0.13
|
81
|
1.0
|
1.0
|
HNa-IE1.5Ni-380
|
319
|
12.2
|
307
|
0.16
|
0.13
|
81
|
1.2
|
1.0
|
HNa-IE1.5Ni-450
|
301
|
15.8
|
286
|
0.16
|
0.13
|
81
|
1.1
|
1.1
|
HNa-IE1.5Ni-500
|
317
|
14.8
|
302
|
0.16
|
0.13
|
81
|
1.1
|
1.0
|
According to the results of the analysis of XRD patterns (Fig. 5), the crystal structure of the base of the catalysts, in which nickel was introduced by the ion exchange method, corresponds to the MFI structure (JCPDS Сard #42 − 24). In contrast to catalysts obtained early by impregnation11 with greater amount of nickel, these samples contained only 1.5% of nickel and therefore reflections of crystalline Ni are not shown on diffractograms. Fragments of diffractograms recorded with a smaller step and greater exposure in the area of nickel reflexes demonstrate this more clearly (Fig. 6). It can be caused by the fact that nickel is more dispersed in samples of ion-exchange metal introduction, and at the same content, its particles are smaller, which cannot be determined using X-ray diffraction.
Transmission electron microscopy was additionally used for nickel particles identification (Fig. 7). One can see that it is difficult to recognize nickel particles on the ion-exchange sample, but they are more visible on the impregnated samples surface.
3.2 Catalytic properties
Six samples with a nickel content of 1.5% by weight were studied in n-hexane hydroisomerization. Four of them, which were reduced in different modes with the production of a partially hydrogen form, and contained residual sodium in the amount of 35% of the ion exchange capacity (HNa-IE1.5Ni, HNa-IE1.5Ni-380, HNa-IE1.5Ni-450, and HNa-IE1.5Ni-500), and two in the completely hydrogen form (H-IE1.5Ni and H-Im1.5Ni).
All samples in the partially hydrogen form did not differ among themselves in terms of catalytic properties, so the Fig. 8 shows the results only for the sample of the conventional method of reduction (HNa-IE1.5Ni). These samples turned out to be almost inactive in the target reaction (Fig. 8, a), instead showed a high yield of cracking products (Fig. 8, b). Their high cracking activity can be connected with nickel localization in the zeolite pores, where contact with Brønsted acid sites are more close.19 The Brønsted sites in these samples appeared in the result of nickel reduction and bridge OH-groups generation. In the result, Brønsted acidity can increase significantly,20–22 which leads to intensification of cracking reaction. Performance of full hydrogen form in comparison with partially hydrogen one is really better. In both cases metal particles localized in zeolite micropores, but, firstly, the ratio between metal sites and Brønsted sites may not be optimal in the case of partially hydrogen form as it was stated in some papers.2 Due to authors conclusions, the optimal catalyst must exhibit near equal amounts of metal and acid sites, and it facilitates rapid mass transfer. But in our case, from one ion exchanged Ni2+ cation two OH-groups can be generated. Secondly, the localization of Ni in the zeolite micropores can be also different, because of different kinetic diameter of sodium cations (0.099–0.139 nm) and ammonium cations (0.14–0.167 nm). The latter are bigger and nickel in the case of introduction in ammonium form can be localized in the MFI zeolite channel intersections only. These intersections have diameter of 0.8 nm which is better for formation of branched hydrocarbon structures. These assumptions confirm by porous characteristics, which demonstrate bigger drop in SBET for partially hydrogen form samples. Nickel nanoparticles located in channels leads to decrease of microporosity of the zeolite samples of partially hydrogen forms.
Samples in fully hydrogen form with the same amount of nickel, regardless of the method of introduction (impregnation sample H-Im1.5Ni or ion exchange sample H-IE1.5Ni), are more selective for branched C6 hydrocarbons (Fig. 8, c). In their presence, a slightly increased conversion of hexane is observed (Fig. 8, d) due to a decrease in the yield of cracking products (Fig. 8, b). However, the maximum yield of hexane isomers (25% by weight) does not reach the value of this indicator for the catalyst with 1% of nickel on the hydrogen form of MFI obtained early by impregnation (35%),11 and it is observed at a bit higher temperature. It is obvious that for MFI zeolite with a silica-to-alumina ratio of 41, the number of "metallic" centers, which corresponds to a nickel content of 1.5% by weight, might be too large. Comparing H-Im1.5Ni and H-IE1.5Ni samples, one can see similar yields of isohexanes and selectivity at a little higher temperatures for the latter, but ion exchanged sample demonstrates higher selectivity for iC6 at 250 oC. It seems that for the H-Im1.5Ni sample optimum temperature is 240 oC, at which its selectivity is near 70%.
On the other hand, for zeolites with a lower SiO2/Al2O3 ratio, the optimal amount of the metal component may be different than utilized in this study. Therefore, the use of ion exchange, in turn, can help to introduce metal particles with significantly better distribution in the sample volume. However, using ion exchange it is more difficult to control accurately the amount of introduced metal.