X-band (Ç9.3 GHz) pulsed ENDOR measurements were carous studies due to the potential new catalytic properties of ried out on 57 Fe-substituted sodalite (FeSOD) which contains only the modified materials (1-9). In these Fe-substituted materione type of Fe(III) (S Å 5
2 ) located at a framework site. The als, the iron does not necessarily exclusively occupy frame-ENDOR spectrum recorded at g Å 2 shows three doublets corrework sites. Fe cations can also be found at extra-framework sponding to the six M S manifolds. The assignment of these signals sites, balancing the negative charge of the framework, or as was confirmed by hyperfine-selective and triple ENDOR experian interstitial phase of small particles located either within ments. The components of each of the doublets had different intenthe molecular sieve cavities or between the crystallites (10sities, reflecting the different populations of the EPR energy levels 12). In order to control and optimize the catalytic activity at the measurement temperature, 1.8 K. ENDOR spectra were of the iron-substituted materials, it is necessary to know recorded at magnetic fields within the EPR powder pattern, and which type of iron is responsible for the activity, and therethe field dependence observed showed an anisotropic behavior, fore an unambiguous characterization of the various iron unexpected from the isotropic character of the 57 Fe(III) hyperfine sites and types is essential.
coupling. This dependence was attributed to the high-order effects of the zero-field splitting (ZFS) interaction on the ENDOR fre-One of the methods commonly applied to characterize quencies. Three different theoretical approaches were used to ac-Fe(III) sites in molecular sieves is EPR spectroscopy. EPR count for the dependence of the ENDOR spectrum on the ZFS can provide information regarding the number of the differinteraction. The first involves the exact diagonalization of the total ent types of Fe(III) present (lower limit only), and it gives spin Hamiltonian, the second uses third-order perturbation apgeneral information regarding their symmetry. However, beproximations, and the third employs an effective nuclear Hamiltocause of its low resolution, arising from the inhomogeneous nian for each of the M S manifolds. The simulations showed that broadening caused by the zero-field splitting (ZFS) interacthe ENDOR signals of the M S Å {5/2 (n {5/2 ) manifold are the tion, the EPR spectrum usually does not give detailed inforleast sensitive to the magnitude of the ZFS parameter D and are mation on the local environment of the Fe(III) (13). Additherefore the most appropriate for the determination of a iso . It is tional characterization information can be obtained from the shown that at X band and a iso values of about 30 MHz, the pertur-NMR frequencies of 57 Fe(III) and of other nuclear spins bation approach is valid up to D values of 500 MHz if all three doublets are concerned. However, if only the n {5/2 doublet is con-coupled to the iron. These, in turn, give the corresponding sidered, then this approach is valid for D õ 1000 MHz. The third hyperfine interactions which provide information such as the approach was found inappropriate for a iso values of Ç30 MHz. identity of neighboring atoms, the degree of covalency, and Using the method of exact diagonalization together with orientabonding distances. The NMR frequencies are best detertion selectivity, the trends observed in the experimental spectra mined from electron-nuclear double-resonance (ENDOR) could be reproduced. The ENDOR spectra of the 57 Fe-substituted spectroscopy (14, 15) and/or electron spin-echo-envelope zeolites ZSM5, L, and mazzite showed broad and ill-defined peaks modulation (ESEEM) techniques (16, 17). Mo ¨ssbauer specsince the ZFS of Fe(III) in these zeolites is significantly larger troscopy can also provide the hyperfine interaction of than that of FeSOD. Because this broadening is a high-order effect, 57 Fe(III) but ENDOR spectroscopy may be advantageous it can be significantly reduced at higher spectrometer frequencies.
due to the additional resolution provided by the EPR spec-᭧ 1997 Academic Press trum and the broad scope of experiments offered by the pulsed methodology.