Molecular Sieves - Science and Technology - Vol. 6 - Characterization II / 06-Isomorphous Substitution in Zeolites

Table 29 compares the relative intensities of the 480 nm band (octahedral Co(III)), the sum of 530, 590 and 640 nm bands (tetrahedral Co(II)) and the 340 nm band (Co(III)). The values are in arbitrary units, and only their variations are meaningful for the state of Co ions in the ZSM-5 samples. The octahedral Co(II) ions should be extra-framework ions, the tetrahedral Co(II) ions could be due to both framework and extra-framework species and the Co(III) ions are considered to be essentially extra-framework species. It can be seen that tetrahedral Co(II) ions are present in all studied samples in roughly similar amounts. Only sample 5 shows a higher amount of tetrahedral Co(II) ions. The amount of extra-framework Co(II) ions are similar in samples 1, 5, and 9. The amount of extra-framework Co(III) ions are similar in samples 5 and 9. No systematic variations could be detected as a function of the formal Co/u.c. content (Table 28). XPS or electron spectroscopy for chemical analysis (ESCA) is used for characterizing surface species. For Co2+, the 2p electron transitions are observed which, after ionization, become separated into two levels: 2p3/2 and 2p1/2. The 2p1/2 peak is shifted upwards in energy by about 15.7 eV. Each peak has one, more or less intense, shoulder, a so-called satellite. Thus, we do not see two, but rather four peaks (three peaks and one strong shoulder) in the XP spectrum.

The appearance of these “excess” peaks is due to multiplet splitting of the respective electron orbitals. It would have been very informative to take the XP spectra of each sample, however, the well-known general trend of such syntheses in alkaline media (resulting only in partial substitution), and the registered UV-Vis spectra (see above) revealed unequivocally that it is impossible to attain complete incorporation of Co2+, thus, only three samples (1, 8 and 12) have been chosen for closer examination. In a previous study we succeeded in detecting a slight, but unmistakable 1.9 eV difference in the locations of 2p3/2 XP signals of FW and EFW Fe3+ ions in heat-treated [Fe]- ZSM-5 samples [225]. Even though the topic was not treated in the literature so far, it was deemed to be interesting to try finding a small shift for FW and EFW Co2+, too. This would have been indicative of mixed coordination, i.e., the presence of both tetrahedral and octahedral Co ions. The XP spectra of the Co2+ photoelectron region for the [Co]-ZSM-5 samples 1, 8, and 12 are given in Fig. 35. Table 30 provides information on the binding energies of the Co(2p3/2) and Co(2p1/2) core levels (with respect to Si(2p)) and on the location of the satellites. If we add a fourth [Co]-ZSM-5 sample to the selected ones (synthesized from magadiite in the presence of Co-pyrocatecholate by P. Fejes) (in Table 30 marked as reference), it is perhaps not an overstatement that the increase of the Co2+ content (cf. Table 29) brings about a little upward shift in the location of the 2p3/2 peaks, which do not attain 781.5 eV, typical of free Co2+, even for sample 8, having the largest Co2+ content (12 Co2+/u.c., corresponding to Si/Co = 6.56). It is believed that the binding energies of core level electrons for Co2+ in tetrahedral coordination (similar to the respective Fe3+ levels) are smaller than those in octahedral symmetry and

Fig. 35 XPS spectra of Co(2p3/2) and Co(2p1/2) of samples Nos. 1, 8 and 12 (see Table 30)

Table 30 XPS data of selected [Co]-ZSM-5 samples

provided by P. Fejes, see text

in mixed coordination. The spectrometer—not able to resolve the two peaks (separated by a few tenths of eV)—registers only one peak at the weighted average of the two binding energies. This manifests itself as an upward energy shift when the percentage of octahedral Co2+ increases with the cobalt content.

3.10

Calcination Using Ozone: Preservation of Framework Elements

The [Co]-ZSM-5, [Co,Al]-ZSM-5 and [Ga]-MCM-22 samples were synthesized in alkaline media, while the B- and [Ga]-ZSM-5 samples were obtained in fluorine-containing media [269]. The ozone treatment was carried out as follows. The sample dried at room temperature was placed in a glass boat in thin layer and placed into a glass tube and heated up to 373 K for dehydra-

448 J. B.Nagy et al.

Table 31 Visible diffuse reflectance spectra of as-made (a), oxygen- (b) and ozone-treated

(c) [Co]-ZSM-5 and [Co,Al]-ZSM-5 samples

creasing, do not change much, suggesting that possible framework species remain essentially in the framework. We would like to emphasize that these measurements are not quantitative and only some trends of band intensities can be shown. Note also that a definite proof for tetrahedral Co(II) species in the zeolite framework is still missing, although XPS, XRD and diffuse reflectance spectra suggest that part of the Co(II) ions occupy tetrahedral positions [197]. The diffuse reflectance spectra of the O2-treated samples are more controversial. Indeed, the intensities of the Co(II) tetrahedral lines are increasing in almost all samples but sample 18. Note also that the increase is much higher when aluminum is not present in the structure (Fig. 36a and b).

More quantitative data are necessary to understand these phenomena. The 27Al NMR spectra show the presence of both tetrahedral framework (in large amount) and octahedral extra-framework (in small amount) aluminum species (Table 32). Oppositely to the changes observed, in the Co visible dif-

fuse reflectance spectra, the relative amounts of tetrahedral and octahedral species do not change during either O2 or O3 treatment. Only the highest Al-content sample shows some decrease in the framework tetrahedral aluminum species. The 29Si NMR data are reported in Table 33, and Fig. 37

Table 32 27Al NMR data of as-made, oxygenand ozone-treated [Co,Al]-ZSM-5 samples

AlT: Al in tetrahedral coordination

AlO: Al in octahedral coordination

Fig. 37 29Si NMR spectra of [Co]-ZSM-5 sample No. 1 (see Table 31) as-made (a), O2- treated (b), O3-treated (c) and of [Co,Al]-ZSM-5 sample No. 21 as-made (d), O2-treated (e), O3-treated (f)

illustrates the spectra of samples 1 ([Co]-ZSM-5) and 21 ([Co,Al]-ZSM-5) in the as-made (a), O2-treated (b), O3-treated (c) forms. The O2-treated samples show completely different spectra with respect to the as-made samples when no aluminum is present in the samples (samples 1, 5 and 9). Indeed, in these cases, the – 102 ppm line is absent in the O2-treated samples, and the spectrum shows high resolution (Fig. 37b). The latter is characteristic of completely siliceous materials, that means no Co is found in the calcined samples. In addition, due to the presence of extra-framework cobalt oxide species, no cross-polarized spectra could be taken. The O3-treated samples always show the presence of the – 102 ppm line even in the absence of aluminum. In the Al-containing samples (18, 20 and 21) the presence of tetrahedral aluminum is always detected confirming the 27Al NMR results (Fig. 37c).

The 71Ga NMR of the [Ga]-MCM-22 samples shows clearly the advantage of the ozone treatment over the calcinations in oxygen. Indeed, in the O3- treated samples, only the tetrahedral framework gallium is detected at ca. 150 ppm as in the as-made sample (Fig. 38).

Oppositely, during the O2 treatment some 22% of the gallium is found as octahedral extra-framework species at ca – 10 ppm. The 29Si NMR spectra are illustrated in Fig. 39. The spectrum of the as-made [Ga]-MCM-22 (Si/Ga = 16) is quite similar to the previously published spectrum of aluminosilicate

Fig. 38 71Ga NMR spectra of as-made (a), O2-treated (b), O3-treated (c) [Ga]-MCM-22 sample

MCM-22 [270]. The lines at – 94 ppm and at – 100 ppm are due to =Si(OM)2 and =Si(OM) defect groups, respectively (M = H, Na and hexamethylendiamine). The number of the other NMR lines is eight if the decomposition is made following a P6/mmm model of the structure, or thirteen if it is made following the C/mmm model [270]. It is clearly seen that the as-made and the O3-treated samples are similar and the O3 treatment leaves the structure quasi intact (Fig. 39a and b). In contrast, the sample calcined in air shows a very different 29Si NMR spectrum (Fig. 39c), where the = Si(OM)2 and ≡ SiOM groups are drastically reduced. Work is in progress to interpret more deeply the decomposed 29Si NMR spectra.

The 11B NMR results are also very clear concerning the state of boron in the various samples. In the as-made samples, boron is essentially at tetrahedral framework positions characterized by an NMR line at – 3.7 ppm. The O3-treated samples basically preserve the tetrahedral framework boron. Sometimes a supplementary NMR line is also found at – 2 ppm what we interpret as stemming from deformed tetrahedral boron species, where one B–O–Si bond is broken: (SiO)3BOH–. When the samples are calcined in air, some 50% of the boron is found as trigonal extra-framework species (δ = ca. 6 ppm, quite a broad line). Fig. 40 illustrates the 11B NMR spectra of (NH4,

TPA)-[B]-ZSM-5 samples (Si/B = 20.8). The as-made and O3-treated samples only show the tetrahedral framework boron at – 3.7 ppm. Oppositely, the sample calcined in air shows the presence of a large amount of trigonal extra-framework boron at ca. 6 ppm.

3.11 Cu-TON

The main characteristics of the Cu-TON samples, chemical analysis and EPR results are reported in Table 34. It can be observed that the amount of copper introduced into the zeolite depends on the exchange procedure. A small content of Cu is introduced during double ionic exchange (Na-TON → NH4+ – TON → H+ – TON → Cu-TON). Following Wichterlova et al. [271] and due to the low Al-content of the samples, the Cu2+ ions occupy essentially cation exchange sites close to individual (Al – O – Si)– negative charges. These Cu species were readily reduced as it was shown elsewhere [272]. On the contrary, higher amount of copper is detected when the exchange process is carried out by the solid-state reaction method. In both cases, i.e., in the direct exchange and the solid-state reaction, we observed “over-exchange”. Moreover, by applying the direct exchange procedure not all sodium ions were removed, thus, they block the openings of the TON pore system. This

is confirmed by nitrogen adsorption measurements of the micropore volume: Na-TON = 0.06 ml g–1, H+ – TON = 0.91 ml g–1. The presence of sodium is

also proved by 23Na NMR spectroscopy. Data of Table 34 show that the solidstate reaction is one of the best techniques for introducing the transition metals. The EPR results indicate that the Cu2+ ions have octahedral symmetry in the Cu-TON zeolite (g|| = 2.34, g = 2.09 and A|| = 12.7 mT) according to the interpretation of the data found by Kucherov et al. [273].

The spectrum of the as-made sample is shown in Fig. 41. It can be seen that it is composed of three lines: the – 111.5 ppm is attributed to T1 sites (eight sites per u.c.), the – 113.2 ppm line to T2 sites (eight sites per u.c.) and

Table 34 Chemical analysis of the Cu-TON samples (Si/Al = 50) prepared in three different ways (Cu = direct exchange, NH4+ – Cu = double exchange and CuCl2 = solid-state reaction) and results of EPR (A|| is given in mT)