Schluesseltech_39 (1)

In the case of nuclear scattering of neutrons the structure factor has the dimension of a length, as has the scattering length bj(;) = bj = const. of nucleus j. Tj(;) is the DebyeWaller factor which takes into account dynamical and static displacements of the nucleus j from its average position rj in the unit cell. With the fractional coordinates xj, yj and zj, the scalar product in the exponential function can be written as

In a diffraction experiment normally only relative Bragg intensities are measured. A scale factor SCALE takes into account all parameters which are constant for a given set of diffraction intensities. Additional corrections have to be applied, which are a function of the scattering angle. For nuclear neutron diffraction from single crystals the integrated relative intensities are given by

The Lorentz factor L is instrument specific. The absorption correction A depends on the geometry and linear absorption coefficient of the sample and the extinction coefficient E takes into account a possible violation of the assumed conditions for the application of the kinematical diffraction theory.

Information on the crystal system, the Bravais lattice type and the basis vectors a1, a2, a3 of the unit cell (lattice parameters a, b, c, 86 , 7) may be directly deduced from the reciprocal lattice. The |F(;)|2 values associated as weights to the nodes of the reciprocal lattice give the diffraction symbol and hence valuable information on the space-group symmetry (see chapter 3). Here, systematic absences (zero structure factors) can be used to determine non-primitive Bravais lattices or detect the presence of non-symmorphic symmetry operations (symmetry operations with translation components).

As an example, consider a body centered cubic lattice with atoms at 0,0,0 and ½,½,½. Using eqn. 4.31 and dropping the Debye-Waller factor for the moment, eqn. 4.30 may be rewritten as:

For a centrosymmetric structure, F is a real quantity (instead of complex), the exponentials in (4.33) reduce to cosines and the phase factor assumes only the values + or -. For this simple structure, index j just runs over the two equivalent atoms with scattering length b within the unit cell. Thus we get:

If h+k+l is even, the cosine term is +1, otherwise it is -1.

Reflections with h+k+l=2n+1 are therefore systematically absent.

These statements apply equally well to x-ray and neutron diffraction and to powder as well as to single crystal diffraction data.

In the case of a powder sample, orientational averaging leads to a reduction of the dimensionality of the intensity information from 3D to 1D: Diffraction intensity I is recorded as a function | ; | = 1/dhkl or, by making use of Bragg’s law, of sin( )/B or just as a function of 2 . For powders, two additional corrections (M and P in eqn. 4.36) need to be applied in order to convert between the measured intensities I and the squared structure factor magnitudes F2:

M is the multiplicity of the individual reflections and takes into account how many symmetrically equivalent sets of lattice planes correspond to a given hkl. In the cubic crystal system, for instance, M111=8 (octahedron) while M100=6 (cube). P is the socalled preferred orientation parameter which corrects the intensities for deviations from the assumption of randomly oriented crystals in the powder sample.

4.5 Diffractometers

Single Crystal Neutron Diffractometry

Hot source

Collimator

Source

Collimator

Eulerian Detector cradle

Sample

Monochromator

2

Detector

Fig. 4.18: Principle components of a constant wavelength single crystal diffractometer.

Monochromator and collimator

For constant wavelength diffraction, the energy (wavelength) and direction (collimation) of the incident neutron beam needs to be adjusted. For that purpose, the diffractometer is equipped with a crystal monochromator to select a particular wavelength band (B ? EBJB) out of the “white” beam according to the Bragg condition for its scattering plane (hkl)

with the interplanar spacing dhkl and the monochromator scattering angle 2 hkl = 2 M. The width of the wavelengths band EBJB, which is important for the Q-resolution, depends on the divergences of the beam before and after the monochromator (collimations 81 and 82), on the mosaic spread of the monochromator crystal EM, and on the monochromator angle 2 M. In order to increase the intensity of the monochromatic beam at the sample position the monochromator crystal is often bent in vertical direction perpendicular to the diffraction plane of the experiment. In this way the vertical beam divergence is increased leading to a loss of resolution in reciprocal space. The diffracted intensity from the sample is measured as a function of the scattering angle 2 and the sample orientation (especially in case of a single crystal). 2 is again defined by collimators. As there is no analysis of the energy of the scattered beam behind the sample, the energy resolution EE/E of such a 2-axes diffractometer is not well defined (typically of the order of some %). In addition to the dominant elastic scattering also quasi-elastic and some inelastic scattering contributions are collected by the detector.

Neutron filters and the problem of /2 contamination

Unfortunately, the monochromator crystals not only “reflect” the desired wavelength by diffraction from the set of lattice planes (hkl) but also the higher orders of /2 or /3 etc. from 2h,2k,2l or 3h,3k,3l to the same diffraction angle:

The only requirement is, that the higher order reflection (2h,2k,2l) or (3h,3k,3l) has a reasonably large structure factor (see chapter 4). Higher order contamination causes measurable reflection intensities at “forbidden” reflection positions and in addition to that can modify intensities at allowed positions. Thus it can very much affect the correct determination of the unit cell as well of the symmetry (from systematically absent reflections). The solution to this problem is to minimize the /2 contamination by using filters which suppress the higher orders stronger than the desired wavelength. One such type of filters uses resonance absorption effects - completely analogous to the suppression of the K line in x-ray diffractometers. Another way to attenuate short wavelengths is to use the scattering from materials like beryllium or graphite. These filters use the fact that there is no Bragg diffraction if B > 2dmax, where dmax is the largest interplanar spacing of the unit cell. As we have shown above, for such long wavelengths the Ewald sphere is too small to be touched by any reciprocal lattice point. Below this critical wavelength, the neutron beam is attenuated by diffraction and this can be used to suppress higher order reflections very effectively. Frequently used materials are polycrystalline beryllium and graphite. Due to their unit cell dimensions, they are particularly suitable for experiments with cold neutrons because they block wavelengths smaller than about 3.5 A and 6 A respectively.

Resolution function:

An important characteristic of any diffractometer is its angular resolution. Fig. 4.19 shows (on the right) the resolution function (reflection half width as a function of scattering angle) for the four circle single crystal neutron diffractometer HEiDi at FRM II shown on the left. The resolution depends on a number of factors, among them the

collimation, the monochromator type and quality, the 2 and (hkl) of the reflection used for monochromatization etc.

Fig. 4.19: Left: Experimental setup of the four circle single crystal diffractometer HEiDi at FRM II. Right: Resolution function of HEiDi for different collimations, monochromator: Cu (220), 2 Mono = 40° = 0.873 Å.

Powder Neutron Diffractometry:

Fig. 4.20: Left: Typical setup of a (constant wavelength) powder neutron diffractometer with position sensitive detector (PSD). Right: Neutron powder diffractometer SPODI at FRM II

Neutron Rietveld analysis:

The conversion from 3Dto 1D-intensity data caused by the averaging over all crystallite orientations in a powder sample severely restricts the informative value of

powder neutron (or x-ray) diffraction experiments and makes the resolution function of the instrument even more important than in the single crystal case. Even with optimized resolution, the severe overlap of reflections on the 2 -axis often prohibits the extraction of reliable integrated intensities from the experiment. Instead, the Rietveld method, also referred to as full pattern refinement, is used to refine a given structural model against powder diffraction data. The method, which is widely used in powder x-ray diffraction, has actually been invented by Hugo Rietveld in 1966 for the structural analysis from powder neutron data. Full pattern refinement means that along with the structural parameters (atomic coordinates, thermal displacements, site occupations) which are also optimized in a single crystal structure refinement, additional parameters like the shape and width of the reflection profiles and their 2 -dependence, background parameters, lattice parameters etc. need to be refined.

Fig. 4.21: Results of a Rietveld refinement at the magnetic phase transition of CoGeO3 [5], red: measured intensity, black: calculated from model, blue: difference, green: tick-marks at allowed reflection positions. The figure shows the low-angle part of two diffractograms measured at SPODI at 35K and 30K. Note the strong magnetic reflection appearing below the magnetic ordering transition (in the inset).

References

[1]J. Strempfer et al. Eur. Phys. J B 14, 63 – 72, (2000). J. Strempfer et al. Physica B 267 - 268, 56 – 59, (1999).

[2]A. Dianoux, G. Lander (Eds.) "Neutron Data Booklet", Institute Laue-Langevin (2002)

[3]Reproduced from: Braggs world, talk by Th. Proffen http://ebookbrowse.com/proffen-talk-bragg-pdf-d59740269

[4]Courtesy of Kevin Cowtan , http://www.ysbl.york.ac.uk/~cowtan/

[5]G. Redhammer et al. Phys. Chem. Min. 37, 311-332, (2010) .

Exercises

E4.1 Types of Scattering Experiments

Discuss/define the following terms:

A. Elastic scattering, B. Inelastic scattering, C. Coherent scattering, D. Incoherent scattering

What is the major source of incoherent elastic scattering that is specific to neutrons?

E4.2 Energy and Wavelength

Give orders of magnitudes for the energy [eV] and the wavelength [Å] of the following types of radiation which are being used for diffraction experiments:

A. Thermal neutrons, B. x-ray photons, C. Electrons

E4.3 Scattering Length

Discuss the terms (units, similarities, differences):

A. Elastic scattering length, B. Elastic scattering cross section,

C. Atomic form factor (for x-rays)

E4.4 The Phase Problem

Describe, in simple terms, the “phase problem of crystallography”

A.Formulate the diffraction experiment in terms of the Fourier transform with subsequent squaring of the modulus of the Fourier coefficients

B.Discuss in how far these operations may be inverted.

C.Describe qualitatively how the phase problem is solved.

E4.5 Ewald Construction

Sketch the Ewald-construction for a single crystal experiment.

What is this geometric construction useful for?

E4.6 Intensity Corrections

The experimental Bragg-reflection intensity I(;) and the squared modulus of the calculated structure factor |F(;)|2 (from the structure factor formula) are proportional to each other.

A number of corrections have to be made to get from I(;) to |F(;)|2 or vice versa.

A. Recall (from your experience with x-rays) and discuss the physical origin of these intensity corrections:

SCALE: Scalefactor; L: Lorentz factor; A: Absorption correction, E: Extinction correction. For powder methods also: M: Multiplicity, P: Preferred orientation.

B.Discuss the relative importance of these factors for neutrons and x-rays.

C.The polarisation correction, which is important in x-ray scattering, is missing in the neutron case: Discuss this fact in terms of the different physical meaning of “polarization” for x-rays and neutrons.

E4.7 Fourier Transform

A.Define the terms “Fourier-analysis” and “Fourier-synthesis” in the context of a diffraction experiment (formula and description)

B.What is the purpose of calculating a Fourier synthesis in crystallography?

E4.8 Filtering

A.How does a beryllium filter work? What is it used for?

B.Discuss why filters are also used in laboratory x-ray diffraction.

E4.9 Systematic absences

Calculate (from the structure factor formula) the systematic absences of reflections for an orthorhombic C-centered lattice.