[2022 report] Naonobu Katada, Yuya Kawatani, Kana Yamamoto, Satoshi Suganuma and Etsushi Tsuji
[2026 report] Naonobu Katada, Shinichiro Matsushita, Nao Tsunoji and Etsushi Tsuji
Center for Research on Green Sustainable Chemistry, Tottori University, 4-101 Koyama-cho Minami, Tottori 680-8552, Japan
katada@tottori-u.ac.jp
When the distribution of a sample of the International Standard Zeolite, IRZ-MFI003, began in May 2022, the results of ammonia IRMS-TPD measurements for the sample were opened to public on the web site of Catalysis Commission, International Zeolite Association. However, evidence suggests that those results contained significant experimental error. Consequently, ammonia IRMS-TPD measurements were repeated in September 2026 using the same sample (IRZ-MFI003). The results presented here are reliable. The samples were taken from the same batch of zeolite powder, and the discrepancy in results is attributed to experimental error during the 2022 measurements.
This document is a revised version in which the data regarding IRZ-MFI003 have been completely updated. Appendix illustrating the differences between the old and new data is attached at the end of document for reference.
Here the ammonia IRMS (infrared-mass spectroscopy)-TPD (temperature-programmed desorption)1 was applied to the samples. The method was developed by combining the conventional ammonia TPD and IR techniques to measure the number and strength of Brønsted and Lewis acid sites on a solid, and recently the accuracy has been improved by developing an automatic analyzer2.
As shown in our previous paper3, ca. 6 mg of the powder sample was compressed into a self-supporting thin wafer with 1 cm of the diameter. In an operando analyzer (MicrotracBEL, IRMS-TPD), it was pre-treated in flowing oxygen (102 kPa, 33 μmol s−1) at 823 K for 90 min. The IR spectra of thus pre-treated zeolite were measured with 1 K of the interval in flowing inert gas (helium or argon, 5.5 kPa, 81 μmol s−1) with elevating the temperature from 313 to 813 K at 0.033 K s−1. Then, ammonia (13 kPa) was admitted contacting to the sample at 313 K for 30 min. After stabilization in the inert gas flow (5.5 kPa, 81 μmol s−1), the sample was again heated from 313 to 813 K at 0.033 K s−1 with recording the IR spectra at 1 K interval and MS ion currents. After the measurements, the intensity of MS was calibrated with ammonia / helium (or argon) mixture with known concentration.
Thus, MS-TPD profile, i.e., plots of concentration of ammonia in gas phase (Cg) against the temperature (T) was obtained, as well as the difference IR spectra of adsorbed ammonia (spectrum after the adsorption of ammonia − that before the adsorption of ammonia measured at the same temperature) at elevated temperatures. Differentiation of the peak intensities (peak areas) of symmetric deformation of NH4 held by Brønsted acid site (1450 cm−1) and NH3→Al on Lewis acid site (1300 cm−1) gave the IR-TPD, namely, the plots of −dAi/dT against T, where Ai shows the absorbance of species i (i = NH4 or NH3→Al) in the difference IR spectrum. It is here assumed that the sum of desorption rates of ammonia from Brønsted and Lewis acid sites equals to the total desorption rate reflected by Cg. Based on this assumption, the curve fitting between kNH4(−dANH4/dT)+kNH3→Al(−dANH3→Al/dT) and Cg to find the pair of kNH4 and kNH3→Al giving the best fitting over the experimental temperature range was carried out. The thus calculated ki of the species i shows πd2β/4εiF, where π, d, β and F show the circle ratio (3.14), diameter of the wafer (10 mm), ramp rate (0.033 K s−1) and apparent flow rate of carrier gas (1.50×10−5 m3 s−1). The parameter εi is the molar extinction coefficient of the discussed species i, which has been unknown before the measurement, but here thus estimated. Simultaneously, the plots of kNH4(−dANH4/dT) and kNH3→Al(−dANH3→Al/dT) against T are obtained. These should be the plots of Cg desorbed from each of Brønsted and Lewis acid sites as functions of T, i.e., the ammonia IRMS-TPD profiles from each of Brønsted and Lewis acid sites3.
The acid amount (A0) is calculated by ∫FCgdt/W [= ∫FCgdT/(βW)], where t and W are the time and amount the zeolite wafer, measured after the experiment. As an index of acid strength, the molar standard enthalpy of ammonia desorption (ΔH∘) and its distribution was calculated by a curve fitting method4, 5, according to the theoretical equation Cg = −A0W/F(dθ/dt) = θP∘e−ΔH∘/(RT)eΔS∘/R/{(1−θ)RT}, where e, θ, P∘, R and ΔS∘ are the Napier's constant (2.72), coverage of acid site by ammonia, pressure in the standard conditions (105 Pa), gas constant (8.314 J K−1 mol−1) and molar standard entropy of ammonia desorption (assumed to be 155 J K−1 mol−1 based on the experiments)6, 7, 8.
As shown in Figure 1, IRZ-FAU001, IRZ-MFI001 and IRZ-MFI002 revealed IR spectra of typical zeolites which consisted of bands at 3745 cm−1 due to isolated SiOH on the external surface and those at 2000, 1860, 1765, 1540 and ca. 1100 cm−1 (very strong) ascribable to SiO framework, as well as some noises due to H2O and CO2. In addition, IRZ-MFI001 and IRZ-MFI002, the MFI samples with high Al concentration showed 3610 cm−1-band due to the Brønsted acidic AlOHSi, whereas IRZ-FAU001, the FAU sample showed those at 3625 and 3580 cm−1, attributed to the AlOHSi in supercage and sodalite cage9, respectively. IRZ-MFI003, the MFI sample with low Al concentration showed a broad absorption due to OH stretching in 2500−3800 cm−1 region, but no clear band ascribable to the Brønsted acidic AlOHSi was observed.
As shown in Figure 2, the IR spectra of adsorbed ammonia revealed the symmetric deformation of NH4 bonded to Brønsted acid site (1450 cm−1) on all samples, while the symmetric deformation of NH3→Al as a Lewis acid site (1300 cm−1) was small, especially on MFI (IRZ-MFI001, IRZ-MFI002 and IRZ-MFI003). These bands were deconvoluted as Figure 3, and the temperature dependences of band intensities were analyzed. On the other hand, the mass spectrometer gave the change in NH3 concentration in the gas phase. They were totally analyzed to provide the TPD profiles of Brønsted and Lewis acid sites as shown in Figure 4. The amounts of Brønsted and Lewis acid sites were estimated from the peak intensities of these profiles as shown in Table 1. Distribution of ammonia desorption enthalpy (close to energy), an index of acid strength, was analyzed by means of an analytical method5 based on the curve-fittings in Figure 5 and Figure 6 to give the distribution curves shown in Figure 7.
Figure 7 shows that the enthalpy of ammonia desorption of most Brønsted acid sites on MFI (IRZ-MFI001, IRZ-MFI002 and IRZ-MFI003) was 130−150 kJ mol−1, as typically observed on MFI type aluminosilicates without extra-framework cations2,4. On the other hand, the FAU sample (IRZ-FAU001) had weak (enthalpy = 110−115 kJ mol−1) and strong (120−145 kJ mol-1) fractions of Brønsted acid strength. The former is attributed to the Brønsted acid sites generated by isomorphous substitution of Si with Al in the FAU framework9, while the latter is speculated to be the Brønsted acid site enhanced by extra-framework AlOH2+ cation located near to the site10, 11, 12, 13, 14. Figure 7 also shows the enthalpy of ammonia desorption of Lewis acid sites. Again the relatively large amount of Lewis acid sites on the FAU sample (IRZ-FAU001) is found.
The amounts of Brønsted acid sites are summarized in Table 2 as well as the chemical compositions. The NH4-MFI sample with high Al concentration (IRZ-MFI002) had a large amount of Brønsted acid sites, corresponding to 0.982, close to 1, of [Al] − [Na] contents, indicating generation of Brønsted acid site by the isomorphous substitution of Si with Al. On the other hand, the ratio [Brønsted acid] / ([Al] − [Na]) was lower, ca. 0.8, on the H-MFI (IRZ-MFI001 and IRZ-MFI003). IRZ-MFI001 was prepared by calcination to remove NH3 from an NH4-MFI equivalent to IRZ-MFI002. Calcination and storage in atmosphere of H-MFI with high Al concentration was thus found to decrease the ion exchange capacity, presumably due to dealumination from Al-rich proton-form zeolite in humid atmosphere even at room temperature, while NH4-form was stable as reported15, 16, 17. The FAU sample (IRZ-FAU001) also showed low the [Brønsted acid] / ([Al] − [Na]) ratio, 0.30. In addition to the detection of Brønsted acid site enhanced by the extra-framework Al species as above, this points out the presence of considerable amount of extra-framework Al on IRZ-FAU001.
Figure 8 compares the ammonia desorption enthalpy distributions of Brønsted acid sites on the present samples and Beta, MWW, YFI and MOR zeolites. This tells us that the Brønsted acid strength was different among zeolites with different framework topologies. The reason has been discussed18.