Planificación

Figure 2.5 shows a schematic of a Philips high-resolution single-crystal diffractometer that uses a four-circle goniometer to position the detector and sample relative to the incident x-ray beam. The two circles to and 20 lie in the plane of diffraction, in which the x-ray beam is already conditioned by a monochromator (<J> and y circles are found at ninety degrees to this plane, and allow for accessibility to different reflections, though chi is limited to ±10°). Rotations about these axes allow positioning of sample and detector, with respect to the incident x-ray beam, to allow access to large sections of reciprocal space. The advantage of such a diffractometer over 2-circle diffractometers is that for a single sample, a greater area of the diffraction space can be covered in one experiment, and this flexibility is advantageous where samples are small, difficult to produce or prepare, or are otherwise valuable.

Figure 2.5 Schematic 8-bounce, 4-crystal diffractometer (after Fewster4)

Here, the co-circle controls the rotation of the sample (and hence the points of the reciprocal lattice) around the origin of reciprocal space, as in Figure 2.6.

Figure 2.6 also shows the effect of 20 rotation on the detector. As the detector arm rotates, the detector itself is exposed to x-rays scattered into different directions by the crystal. These two rotations can be coupled, such that for any given angle-sweep of to, 20 sweeps twice the angle, and the product of such combined rotations is a relative motion of RLP and detector along the vector - the scattering vector for the sample, geometrically perpendicular to the scattering planes themselves. In the high-resolution diffractometer described by Fewster1 \ the angular precision of the to and 20-circles is made high enough that the small arcs described in reciprocal space by rotations around them are approximately flat.

Figure 2.6 Diagram showing motion of detector and diffracting

planes to give omega and 2-theta rotations, combined inlo_r*.

By composing a series of uV20 scans, which move the detector across the RLP, with stepped rotations in to (moving the RLP across the Bwald sphere) a “map" of the reciprocal space in the region of a single RLP can be created, as in Figure

2.7.

Figure 2.7 Diagram showing mapping of area in region o f

RLP by composing parallel r* scans offset by 10 steps.

If an RLP is contained in this area, then all of its diffracted intensity can be recorded, and any of its structural features examined. There are, however, certain conditions which have to be met in order that this can be usefully achieved.

2.3.2.2 Beam conditioning.

To probe reciprocal space with high-resolution requires a probe design that is primarily smaller than the RLP it is examining. Since this probe is moved around in reciprocal space, it must also be of predictable shape, and introduce no artefacts which would obscure or confuse information from the RLP. These are the criteria described by Fewster4, and they are achieved by firstly modifying the x-ray beam before it meets the sample, and then controlling the diffracted x-rays as they meet the detector.

A conventional sealed-tube x-ray source has a band of wavelengths in its output, centred around its characteristic wavelengths, which are superposed on a broad, white, Bremsstrahlung background. These emerge from the fine-focus tube as a divergent cone of radiation. Both these conditions have to be altered in order to avoid the Ewald-sphcre modifications seen in Figures 2.3 and 2.4 - i.e. to achieve as near as possible an infinitely thin sphere. In the 8-bounce, 4-crystal high-

resolution diffractometer (8B4CHRXRD), the modifications to the incident x-ray beam are achieved by a 4-bounce Bartels monochromator, which sits at the emergence point of the x-rays from the tube. Two channel-cut Ge crystals (with the 110 plane as the face of the cut) are set as in Figure 2.8, modifying the x-rays before they proceed to the sample.

Figure 2.8 Representation of monochromator reflections for Bartels

monochromator.

All bounces from the monochromator remove x-rays that fall outside the reflectivity profile for the Ge 220 reflection. By arranging the Ge crystals such that four bounces take place, both divergence and monochromaticity are minimised'1. The primary reflection coarsely selects the wavelength range that fits under the reflectivity curve. This wavelength group passes onto the second bounce (set non-dispcrsively) and makes the beams of different wavelength more parallel by cutting the tails of the distribution. The third bounce is set dispersively, and disposes of beams arriving at large wavelength deviation, since the angular difference is doubled. It also serves to remove more intensity from the tails. The final bounce narrows the peak still more, and also returns the beam to its initial direction of propagation. By thus using the intrinsic-reflectivity profile of the Gc 220 reflection as a “filter”, the beam that proceeds to the crystal has a virtually tail-less profile, and a tight wavelength spread. In fact the divergence in the diffraction plane is only 12” or less, and the wavelength spread (for Gc 220), AXA, is 6.9 * 10 ’ - also around 12” equivalent4. The beam is also

almost completely polarised, in the vertical plane, after the four reflections each with an angle of reflection near 45°.

The monochromator thus produces an Ewald sphere which has near-ideal geometry. Though it must be remembered that if an RLP is moved through a probe with such a shape, any fine-structure the RLP has will be lost in convolution with the shape function of a coarse detector.

A typical 2-dimensional RLP image from such a diffractometer for an ideal, strongly scattering crystal, has two streak-like distortions in reciprocal space. The first is the crystal truncation rod (or surface-streak) and is caused by the sampling of a finite number of planes at the surface of the crystal. The second is an instrumental function, caused by the angular resolution of the detector, and is the analyser streak.

Thus for true high-resolution analysis of reciprocal space, beam conditioning must also be performed at the detector. If the detector is to have a fine angular acceptance, in order to discriminate the spatial distribution of x-rays coming from the sample, then another multiple-bounce technique can be used. For extremely high angular discrimination such as in the 8B4CHRD4, a 3-bounce channel-cut Ge crystal is used, as in Figure 2.5, and carries the benefits of not reducing the intensity of the signal to the extent that a slit would. The triple­ bounce causes an extremely fine angular selectivity, as in the monochromator.