NEW RADIATION KINDS

In recent times, Mineralogy underwent an increasing inflow of new analytical methods derived from Physics which make use of high-energy sources other than x-rays, namely neutrons, electrons and synchrotron radiation.

Neutron diffraction was experimentally proven by W.M. Elsasser in 1936, and the first diffractometer built by W.H. Zinn in 1947. However, there are serious limitations in its use, both intrinsic (the wavelenght of the neutron beam is much longer than the diameter of the atom which is expected to diffract; the interaction with matter is rather weak; the diffusion power of atoms varies irregularly, and irrespectively upon atomic number; etc.) and practical (reactor sources are few and mostly weak, specimens need to be large; etc.). Consequently, neutron diffraction began to give useful data only in the late '60s, especially after H.M. Rietveld conceived (1967) and improved (1969) his structure refinement method based on fitting the full diffraction profile that made working on powders possible. The first significant results, on zeolites, are as late as the years '70s. Neutron diffraction is complementary to x-ray diffraction in that it is at its best for light elements; thus, it is usually applied to locate the H atom in OH-bearing minerals (occasionally also after deuterization), or to differentially locate the distribution in a structure of two atoms close in their atomic weights, both light (e.g., Al vs. Si) and heavy (e.g., Fe vs. Mn or Ni, Ti vs. Mg or Nb, etc.). Furthermore, neutrons have a dipole momentum and interact with atoms having unpaired electrons; consequently, they are of use to define magnetic ordering in minerals showing the ferromagnetic to antiferromagnetic transition, such as the magnesioilmenites, which are important on Earth because they carry residual paleomagnetism.

Electron diffraction is widespread because it does not require a reactor as a source, but an electron microscope working in the transmission mode, TEM. In its basic principles it completely differs from both x-ray and neutron diffractions; the accelerated electron beam has an extremely short wavelenght (<0.03 Å) but it is strongly attenuated by the atoms it impinges: thus specimens have to be prepared very thin. Nevertheless, this is the only currently available method to collect information on crystals in extremely small areas, down to few unit-cells in diameter, and to evaluate defects such as stacking faults, dislocations, microtwins, etc. This is why electron diffraction is so important for crystal growers. When TEM is implemented with a few accessories, a spectroscopic study can be made either on the surface or in deeper strata of the material.

Ever since G. Bathow, E. Freytag and R. Haensel (1966) showed that a portion of the synchrotron radiation spectrum is emitted in the energy range of x-rays while maintaining its outstanding special properties (very high intensity and polarization, very low divergence, pulsate structure), synchrotron radiation is being used to gather diffraction effects in very short times (milliseconds) or from very small crystals (microns) with a resolution and a peak to background ratio that make crystal structure solving a very precise procedure. Synchrotron radiation is useful especially in: (i) very high pressure crystallography with diamond anvil cells: e.g., good diffraction effects have been obtained from ice that permitted detecting a new structure type via the Rietveld method; (ii) in high-resolution diffraction studies, where reflections splitted by few hundreths of degree become index of deviations from the theoretical structure due to the presence of impurities or internal tensions in the sample; (iii) in following the reaction kinetics, when a phase changes into another with a rapid re-ordering of the structure. The "white" character of synchrotron light permits a new diffraction strategy to be followed in determining crystal structures: such an "anomalous diffusion" (L.M. Maroney, P. Thompson & D.E. Cox, 1988) consists in three or more recordings of the full diffraction spectrum with the impinging radiation monochromatized at or near the characteristic wavelenghts of the atoms present in the sample. This information makes it easy to separately refine the thermal factors of the various atoms, and reduces their positional disorder in the crystal structure.

The "white" character of synchrotron radiation also revived Laue's original method: it is now possible to record thousands of diffraction spots from very small crystals (<0.05 mm) in short times (min) on position- sensitive image plates (PSD) and solve structures containing hundreds of atoms. To this achievement, the increased power of modern computers has given a definite contribution.

Where, however, synchrotron radiation plays best is in the study of minerals by spectroscopic methods, which are complementary to diffraction but did not play a major role in Mineralogy for many years because the available x-ray sources were too weak to allow meaningful signals: for a long time spectroscopy was limited to spotting foreign atoms present in a sample, mostly REE. Synchrotron radiation made x-ray absorption spectroscopy (XAS) an efficient method only in the late '60s. The oscillations above the absorption edge of an atom discovered by L.D. Kroning in 1922 can now be recorded independently upon those of other atoms present; they can then be interpreted (F.W. Lytle & E.A. Stern, 1979) to determine the oxidation state and the coordination number, and even the shape and orientation of the coordination polyhedron around the absorbing atom; thus an atomic short-range order distribution can be determined that compares usefully with long-range order best determined by x-ray diffraction; local defects down to an individual site in the unit cell can be appreciated and, in particular, all this can be obtained even for atoms in dilute amounts in the sample, since the very high intensity of the synchrotron beam enhances them enough even in the presence of abundant other components. Total amount (SR-XRF), local environment (EXAFS), valence and coordination (XANES) can be separately determined for all atoms present in a compound, and used to re-construct its bulk crystal-chemistry. A special technique (SEXAFS) allows studying the surface of the specimen, where most reactions with the external environment occur; this contributes to understanding of industrial and/or natural pollution through minerals.

Gamma-rays resonance spectroscopy is based on an effect discovered by R.L. Mössbauer in 1958 and became the standard method of Mineralogy (G.M. Bancroft & R.G. Burns, 1967) to determine the oxidation state of Fe and to locate its distribution among the various sites of a structure, provided these have different crystal field energies. However, atoms displaying the Mössbauer effect are few (Fe, Sb, Sn, Dy etc., among which Au), and the relevant minerals have been by now all investigated. Recently, however, this method received a new impulse by the development of a microprobe (C. McCammon, V. Chaskar & G. Richards, 1991) suitable to extract information from very small samples such as those in the diamond anvil cell.

Nuclear magnetic resonance (NMR) spectroscopy was discovered by E. Zavoisky in 1945, but entered Mineralogy only when the magic angle spinning (MAS) technique was developed and permitted recording effects on solids that are as well resolved as those previously obtained on liquids. E. Lippmaa. M. Mägi, A. Samoson, M. Tarmak & G. Engelhardt (1980) showed the importance of NMR by systematically surveying the 29Si behaviour in silicates, and J. Klinowski, S. Ramdas, J.M. Thomas, C.A. Fyfe and J.S. Hartman (1982) showed how important could be to selectively study 27Al and 29Si in aluminosilicate minerals such as zeolites, where NMR is the only possible method producing meaningful information. NMR is used also for very light elements such as H, B, P, F and O: all this makes NMR one of the most promising methods to study real minerals, in contrast with the average distribution determined by XRD, and to kinetically study the transition from solid to liquid, since MNR, like all other spectroscopies, gives just as good information whatever the state of the sample may be.