Cation-induced speciation of port-size during mordenite zeolite synthesis

Mordenite (MOR) zeolite, an important industrial catalyst exists in two, isostructural variants defined by their port-size, small and large-port. Here we show for the first time how a systematic, single-parameter variation influences the synthesis out-come on the final MOR material leading to distinctly different catalysts. The cation identity has a direct impact on the synthesis mechanism with potassium cations generating the more constrained, small-port MOR variant compared to the large-port obtained with sodium cations. This was expressed by different degrees of accessibility ascertained with a combination of toluene breakthrough and temperature programmed desorption (TPD), propylamine TPD, as well as sterically sensitive isobutane conversion. Rietveld refinement of the X-ray diffractograms elucidated the preferential siting of the smaller sodium cations in the constricted 8-ring, from which differences in Al distribution follow. Note, there are no organic structure directing agents utilized in this synthesis pointing at the important role of inorganic structure directing agents (ISDA).


Supporting Note 1. Preliminary synthesis studies
We previously synthesized MOR zeolites with the help of precipitated silica (Hi-Sil 233) and various Al-sources and NaOH as mineralizer.To obtain a potassium pure K-MOR, however, we need to perform the synthesis with as little Na impurities as possible.Consequently we used Al(OH) 3 xH 2 O as the Al source and initially investigated our previously reported synthesis method, solely substituting KOH for NaOH.However, even after extended periods of time no crystallinity was observed requiring us to attempt the synthesis using a method reported by Chi and Sand. 1 The main difference with this method is the elevated crystallization temperature of 185 °C compared to 170 °C which is expected to significantly increase crystallization rate.As the Na content of the Hi-Sil 233 is too high (SiO 2 /Na 2 O >75) we were prompted to try the synthesis of 1K-MOR with Hi-Sil 915 (SiO 2 /Na 2 O >270) as well as Ludox AS-40 (SiO 2 /Na 2 O >470).We also investigated the effect of aging which typically influences particle size. 2 These preliminary experiments concluded that aging is a crucial step ensuring phase-purity.The SEM images show pristine cubic crystals for the aged Ludox AS-40 whereas in the absence of aging, the crystals are decorated with small platelike agglomerates (Figure S1).Comparing the crystallization curves for the two Si sources highlights the identical induction periods followed by diverging crystallization rates, reflecting the difference in reactivity of the Si source (Figure S2).It is likely that the slower crystallization rate results in the formation of impurities observed with both XRD and SEM (Figure S1).Therefore, we opted to use Ludox AS-40 as the Si source.If we assume the site A/B ratio proportionality across the series, presumed Al occupancies can be calculated by a simple system of equations, given that T1/T3 ratios (0.419) and T2/T4 (0.345) remain constant within the relative occupancies.Values obtained are described in the table and forced in the final refinement models with negligible effect in the fittings:  The structural parameter differences became less significant for the protonic forms (Table S2 & Figure S8).The most significant difference between dry H-MOR samples of both extremes is b lattice parameter, which suggest that something must be different between their empty frameworks (See 020 reflection shift in Figure S5).However we are talking about a minimal difference in the unit cell volume; 2777 vs 2774 Å 3 for K-and Na-MOR, respectively.It has been suggested that the unit cell volume can be correlated to the Si/Al ratio. 3However, we see quite a large variation within our measured samples of similar Si/Al ratios and it should be emphasized that water has an enormous effect so care must be taken when reporting and comparing these values.

Toluene TPD
On select samples we investigated the uptake behavior as well as desorption behavior as a function of the temperature.The alkali exchanged forms of the listed mordenite samples were exposed to toluene vapors (see main text) followed by a stepwise TPD sequence with an isothermal step at 100 °C.When performing this procedure on the protonic forms, the peaks overlap as seen in Figure S10.In addition to the above experiments, we also performed consecutive propylamine-TPD runs on the H-MOR (K) and H-MOR (Na) samples to investigate whether the cation nature had any impact on the structural integrity of the materials (Table S7). 6,7During the TPD sequence the material is exposed several times to high temperatures (670 °C) in an oxidative atmosphere which results in the loss of some acid sites.Using the NH 3 concentration as a more reliable measure we find that 31% and 35% of acid sites are lost for the H-MOR (K) and H-MOR (Na), respectively.As the loss is comparable for both samples we conclude that the cation speciation in the synthesis did not have an impact on the structural integrity of the framework.

Figure S1 .
Figure S1.(left) Scanning electron micrographs of the fully crystalline (96+ h) K-MOR synthesized with two different silica sources in the presence and absence of aging (overnight).Only in the case of the aged gel, using Ludox AS-40 did we obtain cubic crystallites.(right) The particle size distribution suggests smaller and more uniform crystallites for Ludox AS-40 compared to using Hi-Sil 915 as Si source.

Figure
Figure S2.X-ray diffractograms of MOR crystallized with different Si sources (top).In case of using Hi-Sil 915, there are some phase impurities present.(right) Crystallization curve for K-MOR synthesized with the two different Si sources.

Figure S6 .
Figure S6.Rietveld refinements of the two end-members of the sample series in the three relevant states; M + -MOR (top), H-MOR (middle) and H-MOR hydrated (bottom).Red lines are the calculated model, the black circles represent the experiment and the green is error line, both panels show the same refinement but the right highlights the high angle fitting.

Figure S7 .
Figure S7.Description of M + sites A and B with the pairs of T atoms near each of them; Site A = T2 & T4, Site B = T1 & T3.

Figure S8 .
Figure S8.(Left) Evolution of b lattice parameter, (right) evolution of unit cell volume, across the samples, form and hydration.

Figure S11 .
Figure S11.Toluene stepwise TPD performed on the protonic forms of various mordenites.In contrast to the alkali-form, the peaks assigned to siliceous pore wall and acid site overlap here.

Figure S12 .
Figure S12.VT-DRIFS experiments of the protonic form H-MOR ion exchanged from Na-MOR and K-MOR respectively.Last panel shows the comparison in the region of interest of both samples of the high temperature spectra.The bands at 3255 cm -1 correspond to hydrogen-bonded silanol nests.4

Table S1 . Physicochemical characterization data for the synthesized MOR with varying amounts of Na and K in the gel. The solid yield is calculated with respect to the amount of SiO 2 utilized. Si/Al and Al content were determined from MP-AES and error margins are reported from performing multiple analyses across different batches. The average error margin for M + /Al content is ca. 5%.
Unrefined, capillary XRD patterns of select M + -MOR samples (left).Unrefined XRD patterns of H-MOR of series end-members where the reflections showing intensities differences are highlighted (right).
Figure S4.Particle size distributions for select mordenite samples and the derived crystal parameters such as crystal volume and aspect ratio.For representative SEM images see main text Figure 3.